the eye · 100 scientific american august 1998 reviews and commentaries of this view. several...

Martin Gardneron Math Games!

SPACE DEBRIS • DNA COMPUTERS • SEAWATER IRRIGATION

THEZMACHINE:

PINCHING PLASMA

FOR FUSION ENERGY

New Thinkingabout

Back Pain

AUGUST 1998 $4.95

Copyright 1998 Scientific American, Inc.

A u g u s t 1 9 9 8 V o l u m e 2 7 9 N u m b e r 2

FROM THE EDITORS

7

LETTERS TO THE EDITORS

8

50, 100 AND 150 YEARS AGO

10

NEWS AND

ANALYSIS

IN FOCUS

How India and Pakistan got the bomb.

15

SCIENCE AND THE CITIZEN

Neutrino mass detected.. . .

The second-biggest bang?.. . The

hungry among us.. . . Science and

theology.. . . Mutant athletes.

18

PROFILE

J. Craig Venter of the Institute

for Genomic Research races

to sequence human DNA.

30

TECHNOLOGY AND BUSINESS

How trade treaties weaken environ-

mental laws.. . . The euro bugs com-

puters.. . . Unions and productivity.

33

CYBER VIEW

Women exit computer science.

38

Even proponents of nuclear fusion research have grown weary of perennial pre-

dictions, going back for decades, that fusion power is just 10 years away. But the

Z machine, a new device for generating intense pulses of x-rays at Sandia Nation-

al Laboratories, might finally give that claim credibility again.

4

Fusion and the Z-PinchGerold Yonas

Electrons moving through silicon do not have a

monopoly on computing. Snippets of DNA in a test

tube, by combining, growing and recombining, can

also solve computational problems. The inventor of

DNA computing reminisces about his discovery and

discusses its future.

CA

RL M

ILLE

R Pe

ter A

rnol

d, In

c.

In a regulatory soup(page 33)

Computing with DNALeonard M. Adleman

Up to 80 percent of all adults eventually suffer back pain. Its

possible causes are multifarious and mysterious: why some

people experience it is as hard to understand as why many

others don’t. Fortunately, treatment options are improving,

and they usually involve neither surgery nor bed rest.

Low-Back PainRichard A. Deyo

40

48

54


Scientific American (ISSN 0036-8733), published monthly by Scientific American, Inc., 415 Madison Avenue, New York,N.Y. 10017-1111. Copyright © 1998 by Scientific American, Inc. All rights reserved. No part of this issue may be repro-duced by any mechanical, photographic or electronic process, or in the form of a phonographic recording, nor mayit be stored in a retrieval system, transmitted or otherwise copied for public or private use without written permissionof the publisher. Periodicals postage paid at New York, N.Y., and at additional mailing offices. Canada Post Internation-al Publications Mail (Canadian Distribution) Sales Agreement No. 242764. Canadian BN No. 127387652RT; QST No.Q1015332537. Subscription rates: one year $34.97 (outside U.S. $49). Institutional price: one year $39.95 (outside U.S.$50.95). Postmaster : Send address changes to Scientific American, Box 3187, Harlan, Iowa 51537. Reprints available:write Reprint Department, Scientific American, Inc., 415 Madison Avenue, New York, N.Y. 10017-1111; fax: (212) 355-0408or send e-mail to [email protected] Subscription inquiries: U.S. and Canada (800) 333-1199; other (515) 247-7631.

Monitoring and ControllingDebris in SpaceNicholas L. Johnson

Dead satellites, dismembered rockets, drifting

bolts, flecks of paint and more dangerous junk cir-

cle the earth. How can spacefarers avoid making

the problem worse, and can the existing mess in

valuable orbits be cleaned up?

“The line between entertaining math and serious

math is a blurry one,” reflects the author of this

magazine’s “Mathematical Games” column for 25

years. Here he recalls some intriguing puzzles and

makes the case for playing math games in schools.

Secrets behind the revival of the U.S.

industrial economy... . The bottom-

line value of biodiversity.

Wonders, by the MorrisonsMirror molecules.

Connections, by James BurkeClockmakers, mapmakers

and Wanamaker’s.

98

WORKING KNOWLEDGEHow to breathe underwater.

104

About the Cover

The causes of low-back pain remain a

great puzzle: although rates of manual

labor have fallen, the incidence of back-

ache has risen. The Thinker statue by

Auguste Rodin; photograph courtesy of

the Bridgeman Art Library. Digital ma-

nipulation by Jana Brenning.

A Quarter-Century of Recreational MathematicsMartin Gardner

62

68

76

82

88

THE AMATEUR SCIENTISTDream a little stream with me.

94

MATHEMATICALRECREATIONS

If he knows that she knows

that he knows.. . .

96

5

Seawater is lethal to conventional crops. Yet salt-

tolerant plants—suitable for food, animal forage

and oilseed—can flourish in it. Agriculture using

seawater irrigation could transform 15 percent of

the world’s coastal and inland deserts.

Irrigating Crops with SeawaterEdward P. Glenn, J. Jed Brown and James W. O’Leary

Pestilence and ignorance about the cause of the

disease contributed to this outbreak that devastat-

ed America’s early capital. Disturbingly, a repeat

of the disaster is not out of the question.

The Philadelphia Yellow Fever Epidemic of 1793Kenneth R. Foster, Mary F. Jenkins and Anna Coxe Toogood

At 0.01 carat, these microscopic sparklers would

make disappointing rings, but they have impor-

tant industrial applications. The Lilliputian crys-

tals could also help mineralogists better under-

stand how all diamonds form and grow.

MicrodiamondsRachael Trautman, Brendan J. Griffin and David Scharf

Visit the Scientific American Web site

(http://www.sciam.com) for more informa-

tion on articles and other on-line features.

REVIEWSAND

COMMENTARIES


Scientific American August 1998 7

This is the sort of thing I’d expect of the supermarket tabloids, not

of Scientific American,” one letter writer growled. The target

of his ire was the cover of our May issue, which portrays astro-

naut Shannon W. Lucid gazing out a porthole of the Mir space station. He

was annoyed at the realism of the picture and at what he mistook to be

our intention to fool readers that it was a photograph.

Sorry, but innocent as charged. The “About the Cover” description on

that issue’s table of contents states that the image is an artist’s concep-

tion—a fact nearly all readers, I believe, understood. But I understand that

writer’s confusion, and it gives me the opportunity to discuss the covenant

that Scientific American keeps with its readers.

Computers gave artists a new canvas and palette. With scans of pho-

tographs, modeling software and an array of digital techniques, a

skilled artist can create amazingly lifelike images. For the illustrative pur-

poses of a magazine like this one, digital artistry can be a godsend. It can

reimagine unwitnessed scenes (as when Lucid looked out of Mir, or our

sunset panorama of Mars from the July issue, both reproduced above); it

can display inventions still on the drawing board; it can show individual

molecules and other objects that evade regular photography. Convention-

al art techniques can do the same, but sometimes the digital ones do the

job more accurately and realistically. The only catch is that all the realism

can mislead readers into accepting fabulous images too literally.

These pros and cons vex editors and art directors at all publications

these days, and they all wrestle with the problem in their own way. Scien-tific American’s policy is and has been to try scrupulously to avoid mis-

leading anyone. When photolike images are computerized creations, we

note that fact in the captions and credits appearing alongside. We value

our readers’ trust too highly to abuse it.

Believing Your Eyes

®

Established 1845

F R O M T H E E D I T O R S

John Rennie, EDITOR IN CHIEF

Board of EditorsMichelle Press, MANAGING EDITOR

Philip M. Yam, NEWS EDITOR

Ricki L. Rusting, ASSOCIATE EDITOR

Timothy M. Beardsley, ASSOCIATE EDITOR

Gary Stix, ASSOCIATE EDITOR

W. Wayt Gibbs, SENIOR WRITER

Kristin Leutwyler, ON-LINE EDITOR

Mark Alpert; Carol Ezzell; Alden M. Hayashi; Madhusree Mukerjee; George Musser; Sasha Nemecek; David A. Schneider;

Glenn ZorpetteCONTRIBUTING EDITORS: Marguerite Holloway,

Steve Mirsky, Paul Wallich

ArtEdward Bell, ART DIRECTOR

Jana Brenning, SENIOR ASSOCIATE ART DIRECTOR

Johnny Johnson, ASSISTANT ART DIRECTOR

Bryan Christie, ASSISTANT ART DIRECTOR

Bridget Gerety, PHOTOGRAPHY EDITOR

Lisa Burnett, PRODUCTION EDITOR

CopyMaria-Christina Keller, COPY CHIEF

Molly K. Frances; Daniel C. Schlenoff; Katherine A. Wong; Stephanie J. Arthur;

Eugene Raikhel; Myles McDonnell

AdministrationRob Gaines, EDITORIAL ADMINISTRATOR

David Wildermuth

ProductionRichard Sasso, ASSOCIATE PUBLISHER/

VICE PRESIDENT, PRODUCTION

William Sherman, DIRECTOR, PRODUCTION

Janet Cermak, MANUFACTURING MANAGER

Tanya Goetz, DIGITAL IMAGING MANAGER

Silvia Di Placido, PREPRESS AND QUALITY MANAGER

Madelyn Keyes, CUSTOM PUBLISHING MANAGER

Norma Jones, ASSISTANT PROJECT MANAGER

Carl Cherebin, AD TRAFFIC

CirculationLorraine Leib Terlecki, ASSOCIATE PUBLISHER/

CIRCULATION DIRECTOR

Katherine Robold, CIRCULATION MANAGER

Joanne Guralnick, CIRCULATION

PROMOTION MANAGER

Rosa Davis, FULFILLMENT MANAGER

Business AdministrationMarie M. Beaumonte, GENERAL MANAGER

Alyson M. Lane, BUSINESS MANAGER

Constance Holmes, MANAGER,

ADVERTISING ACCOUNTING AND COORDINATION

Electronic PublishingMartin O. K. Paul, DIRECTOR

Ancillary ProductsDiane McGarvey, DIRECTOR

Chairman and Chief Executive OfficerJohn J. Hanley

Co-ChairmanRolf Grisebach

PresidentJoachim P. Rosler

Vice PresidentFrances Newburg

Scientific American, Inc. 415 Madison Avenue

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

PRINTED IN U.S.A.

JOHN RENNIE, Editor in [email protected]

DO

N D

IXO

N (t

op);

U.S

. GEO

LOG

ICA

L SU

RV

EY A

ND

NA

SA/J

ET P

ROPU

LSIO

N L

AB

ORA

TORY

; SU

NSE

T SI

MU

LATI

ON

BY

LAU

RIE

GR

AC

E (b

otto

m)

DIGITAL IMAGES of Shan-non Lucid on board Mir (left)

and a Martian sunset (below)

are artists’ interpretations.That fact was noted in the Mayand July issues, respectively.Drawing on real photographs,these images evoke scenes thatno one has witnessed.


TRADE SECRETS

Ihave to admit that the April Fools’

page, “Self-Operating Napkin,” by

Rube Goldberg [Working Knowledge,

April], completely took me in. I must

have misread the heading as “self oper-

ating system.” I believed that the illus-

tration really did divulge part of the

workings of the Windows 95 operating

system. On second thought, are you sure

it doesn’t?

A. PETER BLICHERPrinceton, N.J.

BRAVEHEART

How Females Choose Their Mates,”

by Lee Alan Dugatkin and Jean-

Guy J. Godin [April], reports that fe-

male guppies prefer to mate with the

male expressing the greatest courage by

swimming closest to a predator. The se-

lection advantage of this rash behavior

in human male survivors was noted in

Iolanthe, by Sir William S. Gilbert, in

his couplets:

In for a penny, in for a pound—

It’s Love that makes the

world go round.. . .

Beard the lion in his lair—

None but the brave deserve

the fair!

DAVID MAGEResearch Triangle Park, N.C.

ACROSS THE SPECTRUM

As a cellular engineer, I enjoyed the

special section in the April issue on

“Wireless Technologies.” In the article

“Spread-Spectrum Radio,” by David R.

Hughes and Dewayne Hendricks, the

introduction implies that advances in

spread-spectrum radio obviate the need

for spectrum conservation and a sepa-

rate cellular service provider (the “mid-

dleman”). But to send signals to one of

today’s cellular phones, for instance, you

need to know its code. The codes are

fairly long, so the memory required to

store the codes for a large metropolitan

area is prohibitive for a single phone.

The solution is to store these codes in a

central server to which the cellular

phone has access. The “middleman” is

still a necessity.

TONY DEANMotorola, Inc.

Hughes and Hendricks reply:Dean’s comments are quite true for

the devices used in today’s cellular tele-

phone networks. What we were refer-

ring to in our introduction was the tech-

nology used in today’s Internet, which

allows routing information to reside at

many different points in a network (rath-

er than on one centralized server), wher-

ever it is necessary and appropriate.

Another good example of this ap-

proach would be the use of TCP/IP pro-

tocols in amateur radio (a.k.a. ham ra-

dio) packet networks, which allow a

metropolitan-area network to be de-

ployed and operated without the need

of a centralized server, or “middleman.”

THE HIGH-TECH LIFE

Marguerite Holloway’s profile of

Sherry Turkle [“An Ethnologist

in Cyberspace,” News and Analysis,

April] contains the following astonish-

ing sentence: “Turkle says her own child-

hood was relatively technology free.”

As she was growing up in Brooklyn in

the 1950s and ’60s, surely Turkle had

electricity, telephones, radio, television,

paved streets, automobiles, subway

trains and skyscrapers. She probably

even crossed some of New York’s mag-

nificent bridges on occasion. Today there

is a tendency to think of “technology”

as anything that is built around an inte-

grated circuit on a chip, like a comput-

er, pager or cell phone, and to take all

the rest for granted as though no tech-

nology were involved.

S. J. DEITCHMANChevy Chase, Md.

POLIO AFTERMATH

As I was reading Lauro S. Halstead’s

“Post-Polio Syndrome” [April], his

reminder that perhaps only one in 50 of

those originally infected with the virus

suffered contemporaneous paralysis led

me to wonder whether the other 49

might also later be candidates for some

variety of post-polio syndrome, albeit

not recognized by that name. For in-

stance, do sufferers of chronic fatigue

syndrome have their spinal fluid checked

for RNA fragments reminiscent of the

poliovirus? There is a large group of

people who suffered polio, but an even

larger group has since received the at-

tenuated virus in the Sabin oral vaccine.

Could some ailments in these people be

related to post-polio syndrome?

C. G. MADSENvia e-mail

Halstead replies:There has been considerable interest

over the years in the possible role of po-

liovirus associated with other diseases.

Chronic fatigue syndrome (CFS) is the

latest. Although the cause of CFS is still

unknown, some researchers believe a

virus may be involved, in particular the

enteroviruses, which include poliovirus.

Unfortunately, most studies of CFS do

Letters to the Editors8 Scientific American August 1998

L E T T E R S T O T H E E D I T O R S

Readers of the April issue were certainly in the April Fools’ spirit. Our deliberate-ly funny tribute to Rube Goldberg received considerable mail, as did John Rennie’seditorial on how SCIENTIFIC AMERICAN really works. Other readers found humorin unexpected places, including a photograph of what seemed to be a guy with light-bulbs in his pants (page 21), and in the article about how females choose their mates.

COMMUNICATIONS SATELLITEis part of today’s wireless network.

SLIM

FIL

MS


not include analyses of patients’ spinal

fluid. One recent investigation did ex-

amine the cerebrospinal fluid, along

with other bodily tissues and fluids, and

found no evidence of a persistent en-

terovirus infection.

We continue to see a number of peo-

ple every year in the clinic with no his-

tory of paralysis or polio who develop

classic post-polio syndrome. It may

turn out that post-polio syndrome is

the most common unexpected condi-

tion experienced by the 49 individuals

who did not suffer paralysis—but un-

knowingly did sustain a critical thresh-

old of spinal cord damage.

OOPS!

There was an error in the Errata in

May: you wrote that the “correct

number of platelets in human blood is

between 150,000 and 400,000 per cu-

bic centimeter of blood.” The correct

unit is cubic millimeter of blood.

PHRIXOS OVIDIU XENAKISSan Antonio, Tex.

Editors’ note:We certainly goofed on this one. To

set the record straight: in the article “The

Search for Blood Substitutes” [Febru-

ary], the chart on page 74 does not con-

tain an error. The article, however, does.

On page 73, center column, the text

should refer to a cubic millimeter of

blood, not a cubic centimeter. We apol-

ogize for the errors.

Letters to the editors should be sentby e-mail to [email protected] or bypost to Scientific American, 415 Madi-son Ave., New York, NY 10017. Let-ters may be edited for length and clari-ty. Because of the considerable volumeof mail received, we cannot answer allcorrespondence.

Letters to the Editors Scientific American August 1998 9

ERRATA

In “Nanolasers” [March], one of

the researchers who proposed

“thresholdless” lasers was misidenti-

fied: he is Tetsuro Kobayashi. In “La-

ser Scissors and Tweezers” [April],

the two photographs labeled mouse

eggs on page 64 actually show bo-

vine eggs and were taken by Cell Ro-

botics in Albuquerque, N.M.

OTHER EDITIONS OF SCIENTIFIC AMERICAN

Le ScienzePiazza della Repubblica, 8

20121 Milano, ITALYtel: +39-2-655-4335

[email protected]

Spektrum der Wissenschaft Verlagsgesellschaft mbH

Vangerowstrasse 2069115 Heidelberg, GERMANY

tel: [email protected]

Investigacion y CienciaPrensa Científica, S.A.

Muntaner, 339 pral. 1.a

08021 Barcelona, SPAINtel: +34-3-4143344

[email protected]

Pour la Science Éditions Belin8, rue Férou

75006 Paris, FRANCEtel: +33-1-55-42-84-00

Majallat Al-Oloom Kuwait Foundation for

the Advancement of SciencesP.O. Box 20856

Safat 13069, KUWAITtel: +965-2428186

Swiat NaukiProszynski i Ska S.A.

ul. Garazowa 7 02-651 Warszawa, POLAND


Nikkei Science, Inc. 1-9-5 Otemachi, Chiyoda-Ku

Tokyo 100-66, JAPANtel: +813-5255-2800

Svit NaukyLviv State Medical University

69 Pekarska Street290010, Lviv, UKRAINE


Ke XueInstitute of Scientific and

Technical Information of ChinaP.O. Box 2104

Chongqing, Sichuan PEOPLE’S REPUBLIC OF CHINA

tel: +86-236-3863170

NEW YORKKate Dobson, PUBLISHER

tel: 212-451-8522, [email protected] Madison Avenue

New York, NY 10017fax: 212-754-1138

Thomas Potratz, EASTERN SALES DIRECTOR

tel: 212-451-8561, [email protected]

Kevin Gentzeltel: 212-451-8820, [email protected]

Randy Jamestel: 212-451-8528, [email protected]

Stuart M. Keatingtel: 212-451-8525, [email protected]

Wanda R. Knoxtel: 212-451-8530, [email protected]

Laura Salant, MARKETING DIRECTOR

tel: 212-451-8590, [email protected] Schube, PROMOTION MANAGER

tel: 212-451-8592, [email protected] Spirakis, RESEARCH MANAGER

tel: 212-451-8529, [email protected] Mongelli, PROMOTION DESIGN MANAGER

tel: 212-451-8532, [email protected]

ASSISTANTS: May Jung, Beth O’Keeffe

DETROITEdward A. Bartley, MIDWEST MANAGER

3000 Town Center, Suite 1435Southfield, MI 48075

tel: 248-353-4411, fax: [email protected]

OFFICE MANAGER: Kathy McDonald

CHICAGORandy James, CHICAGO REGIONAL MANAGER


LOS ANGELESLisa K. Carden, WEST COAST MANAGER

1554 South Sepulveda Blvd., Suite 212Los Angeles, CA 90025


ASSISTANT: Stacy Slossy

SAN FRANCISCODebra Silver, SAN FRANCISCO MANAGER

225 Bush Street, Suite 1453San Francisco, CA 94104


ASSISTANT: Rosemary Nocera

DALLASThe Griffith Group

16990 Dallas Parkway, Suite 201Dallas, TX 75248


International Advertising Contacts

CANADAFenn Company, Inc.

2130 King Road, Box 1060King City, OntarioL7B 1B1 Canada


EUROPERoy Edwards, INTERNATIONAL

ADVERTISING DIRECTOR

Thavies Inn House, 3/4, Holborn CircusLondon EC1N 2HB, England

tel: +44 171 842-4343, fax: +44 171 [email protected]

BENELUXReginald Hoe Europa S.A.

Rue des Confédérés 291040 Bruxelles, Belgium

tel: +32-2/735-2150, fax: +32-2/735-7310

MIDDLE EASTPeter Smith Media & Marketing

Moor Orchard, Payhembury, HonitonDevon EX14 OJU, England

tel: +44 140 484-1321, fax: +44 140 484-1320

JAPAN Tsuneo Kai

Nikkei International Ltd.CRC Kita Otemachi Building, 1-4-13 Uchikanda

Chiyoda-Ku, Tokyo 101, Japan tel: +813-3293-2796, fax: +813-3293-2759

KOREAJo, Young Sang

Biscom, Inc.Kwangwhamun, P.O. Box 1916

Seoul, Koreatel: +822 739-7840, fax: +822 732-3662

HONG KONGStephen Hutton

Hutton Media LimitedSuite 2102, Fook Lee

Commercial Centre Town Place33 Lockhart Road, Wanchai, Hong Kong

tel: +852 2528 9135, fax: +852 2528 9281

Advertising and Marketing Contacts


AUGUST 1948SPELLING BEE—“Karl von Frisch, in his studies of honey

bees, discovered that bees returning from a rich source of

food perform special movements for other bees, which he

called dancing. He distinguished between two types of dance:

the circling dance (Rundtanz) where the bee circles, and the

wagging dance (Schwänzeltanz) when it moves forward,

wagging its abdomen, and turns. Von Frisch showed that the

circling dance is used when the food source is closer than

about 100 meters. In the wagging dance the frequency of turns

indicates the distance to the food source. When the feeding

place was 100 meters away, the bee made about 10 short

turns in 15 seconds. To indicate a distance of 3,000 meters, it

made only three long ones in the same time.”

HYPERTENSION—“There is a growing concern over vul-

nerability to high blood pressure and hardening of the arter-

ies. Death certificates show that these associated conditions

kill some 600,000 people in this country annually. Since the

18th century, life expectancy in the U.S. has increased from

39 to 57 years, largely because of the conquest of diseases

such as smallpox, typhoid fever, tuberculosis, plague, diph-

theria and, more recently, pneumonia and streptococcic in-

fections. The reduction of these diseases has permitted people

to live to the age when hypertension and arteriosclerosis take

their greatest toll.”

AUGUST 1898DAILY GRIND—“It has been considered a quite sufficient

educational training for the young to cram and overload their

brains with a quantity of matter difficult to digest, and of lit-

tle use in after life. Numbers of delicate, highly strung children

have broken down under the strain, and the dreary daily grind

has developed many of the nervous diseases to which the pres-

ent generation is so peculiarly susceptible. However, Ameri-

cans are becoming alive to the pernicious effects of develop-

ing the mind at the expense of the body, and in the ten years

since German gymnastics were introduced, physical training

holds a place in the curriculum of most larger schools.”

EARLIEST MULTIPLEXING?—“Experiments are being

conducted on the Paris-Bordeaux line by the inventor M.

Mercadier. With some interesting instruments, called duode-

caplex, twelve Morse transmitters can work simultaneously

on a single wire, each sending its signals to the proper receiv-

er at the end of the line. Each transmitter receives a current

through a tuning fork having a special note, its vibrations be-

ing electrically maintained. These vibrations cause resonance

at the proper receiving circuit. The sifting out of signals, it

seems, is very perfect, each receiver giving no evidence of

those signals not intended for it.” [Editors’ note: This devicepresages the multiplexing technologies used in analog anddigital voice systems.]

BETTER OPTICS—“The new Zeiss binocular field glasses,

now being manufactured in this country by the Bausch &

Lomb Optical Company, Rochester, N.Y., are the invention

of Prof. Ernst Abbe, of Jena, to whom optical science owes

so many recent improvements. The three principal defects of

the ordinary field glass are overcome by the use of two pairs

of prisms, which erect the inverted image formed by the ob-

ject glass, shorten the telescope by two-thirds and place the

object glasses farther apart than the eyepieces are, thus in-

creasing the stereoscopic effect.”

A BIRD-PLANE IDEA—“Is the ‘Avion,’ an apparatus de-

vised and constructed by M. Clément Ader, a French engineer,

finally to permit man to realize the legendary dream of Icarus?

Perhaps so. M. Ader has abandoned the plane surfaces of the

Maxim and Richet apparatus and substi-

tuted in their place incurved surfaces. The

wings, jointed in all their parts, serve for sustentation and do

not flap. In the version shown in our illustration, the wing

span is 483/4 feet. The motive power is furnished by steam;

the fuel employed is alcohol.”

AUGUST 1848ABORIGINAL BOOMERANG—“The Boomering is a curi-

ous instrument used as an offensive weapon by the blacks of

Australia, and in their hands, it performs most wonderful and

magic actions. A late resident of that strange country published

a description of some of the feats performed by the Boomer-

ing. The instrument itself is a thin curved piece of wood up to

three feet in length and two inches broad—one side is slightly

rounded, the other quite flat. An Australian black can throw

this whimsical weapon so as to cause it to describe a com-

plete circle in the air; the whole circumference of the circle is

frequently not less than two hundred and fifty yards.”

50, 100 and 150 Years Ago

5 0 , 1 0 0 A N D 1 5 0 Y E A R S A G O

10 Scientific American August 1998

A magnificent man and his flying machine


News and Analysis Scientific American August 1998 15

Although observers have presumed for years that

both India and Pakistan had nuclear arsenals, the

11 nuclear detonations on the Indian subconti-

nent in May heralded a new level of tension between foes

that have fought three conventional wars in their 51-year his-

tory as independent nations. More broadly, however, the de-

velopment is illustrative of the challenges facing the interna-

tional community as it seeks to halt global nuclear prolifera-

tion. India and Pakistan, which took widely divergent routes

to their bombs, have shown how hard it is to deny simple but

effective nuclear weaponry even to relatively poor countries,

how easy it is to rend the fragile de facto moratorium on the

use of these munitions and how difficult it becomes to help a

developing nation build nuclear reactors while ensuring that

none of the aid is put to military use.

The Indian and Pakistani tests should not have been a sur-

prise. They have roots, in fact, extending back to the early

years of the nuclear era. In the 1950s, 1960s and into the

1970s, both India and Pakistan participated in nuclear coop-

eration efforts, most notably the Atoms for Peace program

begun by the U.S. Under this program, the U.S. provided crit-

ical blueprints, know-how and components for a plant at the

Bhabha Atomic Research Center in Trombay, just north of

Bombay, for reprocessing the spent fuel from nuclear reactors.

Reprocessing uses a series of complex industrial-chemical

processes to remove from spent fuel variant forms of ele-

ments, called isotopes, that can be reused in reactors. Essen-

tially the same processes, though, can be used to extract the

plutonium isotope 239—the key ingredient in the core of the

most common types of nuclear bombs. In fact, the Trombay

reprocessing plant, which was built in the late 1950s and ear-

ly 1960s, was based on a process called Purex, originally de-

veloped in the U.S. to separate weapons plutonium.

Before nuclear materials can be reprocessed, they must first

be irradiated in a reactor to create plutonium. And here, too,

India got assistance from the West. In the late 1950s Canada

provided a research reactor called Cirus. Significantly, the

NEWS AND ANALYSIS18

SCIENCE

AND THE

CITIZEN

30PROFILE

J. Craig Venter

33TECHNOLOGY

AND

BUSINESS

IN FOCUS

DEADLY SECRETS

As India and Pakistan demonstrated, military nuclear know-how is spreading with frightening ease

38CYBER VIEW

“WE WILL DESTROY INDIA,”reads the headline on this Pakistani newspaper,

after the country’s nuclear tests.

22 IN BRIEF

22 ANTI GRAVITY

27 BY THE NUMBERS

B. K

. BA

NG

ASH

AP

Phot

o


PAKISTAN

INDIA

BOMBAY

CHAGAI HILLS(TEST SITE)

POKHARAN(TEST SITE)

INDIRA GANDHI CENTER FOR ATOMIC RESEARCH

KHAN RESEARCHLABORATORIES (KAHUTA)

PAKISTAN INSTITUTE OFNUCLEAR SCIENCE AND TECHNOLOGY

BHABHA ATOMICRESEARCH CENTER

ISLAMABAD

CALCUTTA

NEW DELHI

MADRASMYSORE

reactor was a type that used so-called heavy water (enrichedin the hydrogen isotope deuterium) rather than ordinary (or“light”) water to lessen the energy of neutrons in the reac-tor’s core. According to John C. Courtney, professor emeritusof nuclear engineering at Louisiana State University, this fea-ture suited it to producing bomb-grade plutonium 239 fromuranium 238, which is abundant in easily produced uraniumoxide fuel. Using the Cirus reactor and the Trombay facility,Indian scientists and engineers accumulated enough plutoni-um to build an atomic bomb, which was detonated in 1974.

After that test, India lost all external support for its nucle-ar program, leaving the country on its own, for the most part,in its nuclear efforts. Indian engineers and scientists built theirown heavy-water reactors, modeled on the Canadian design,as well as their own reprocessing facilities. And in 1985 a rel-atively large plutonium-produc-ing reactor called Dhruva beganoperating at Trombay.

The Indian program did notstop there. According to prolif-eration experts, the country hasalso built a plant near Mysore toproduce tritium—used in thermo-nuclear reactions (which exploitthe nuclear fusion of tritium anddeuterium) and in “boosting”the yields of less powerful fissionbombs. India also reportedly con-structed a plant to produce ura-nium highly enriched in the 235isotope, which can also be usedin concocting nuclear devices.

But the focus of the country’snuclear program rests on pluto-nium. Peter L. Heydemann, thewell-connected former scienceattaché at the U.S. embassy inNew Delhi, estimates that by us-ing Dhruva and other reactors India accumulated some 1.2tons of plutonium, which could yield dozens if not hundredsof bombs, depending on the program’s technical sophistica-tion. With lavish government funding and a large pool of topscientific and engineering talent, India’s nuclear establish-ment swelled in the 1980s to become the only one in the de-veloping world that encompasses all aspects of what isknown as the nuclear fuel cycle, from the mining of uraniumores to the construction of reactors.

Although its nuclear talent pool is by all accounts large andimpressive, India’s claim that one of the three nuclear devicestested on May 11 was a true thermonuclear device was metwith skepticism. Such a device exploits nuclear fusion toachieve explosive yields that can be thousands of times greaterthan those of fission devices. Indian officials said that thepurported thermonuclear device had an explosive yield of 43kilotons. But outside experts, using seismic data, have esti-mated that the yield may have been as low as 25 kilotons. Bythermonuclear standards, these values are rather small.

Given the history of deceit in nuclear weapons programs,many observers have suggested that the Indians tested aboosted-fission device rather than a true thermonuclear bombor that they tested a thermonuclear weapon that turned outto be a dud. Yet Theodore B. Taylor, a former thermonuclearweapons designer at Los Alamos National Laboratory, says

it is possible the Indians opted for a low yield in part because“it would be a more severe test. It’s harder to be accurate inpredicting the behavior of a smaller weapon.” He also notesthat the U.S. has produced tactical thermonuclear weaponsin the past with yields in the tens of kilotons.

In contrast to India’s heavy reliance on plutonium-basedweapons, Pakistan’s nuclear program is—for now—built en-tirely around the use of uranium 235. The isotope, which ac-counts for 0.72 percent of natural uranium, is separatedfrom the useless 238 form in high-performance centrifuges atthe Kahuta research center near Islamabad. The man nowhailed as the father of Pakistan’s bomb, Abdul Qadeer Khan,smuggled plans for the centrifuges in the 1970s from a Dutchcompany, Ultra-Centrifuge Nederland.

Compared with plutonium, uranium has a key disadvan-tage as a bomb material: thecritical mass needed is substan-tially larger, making it harder tofashion into warheads that canbe launched on missiles. U.S.intelligence indicates that Pak-istan’s bombs were relativelysimple Chinese pure-fission de-signs. Such bombs would needmore than 15 kilograms ofweapons-grade uranium (mate-rial containing at least 93 per-cent of the 235 isotope). Pluto-nium-based bombs, in contrast,can make do with as little asthree to six kilograms of fissilematerial. The Institute for Sci-ence and International Security(ISIS) in Washington, D.C., esti-mates that Pakistan could bynow have produced enoughweapons-grade uranium at Ka-huta for up to 29 bombs, with

the total growing by five each year. Pakistan apparently choseuranium because during the 1970s international safeguardsthwarted its attempts to separate plutonium covertly from ir-radiated fuel. This past April, however, Pakistan said it hadcommissioned a new reactor at Khushab in the province ofPunjab that is not subject to safeguards. According to ISIS,Pakistan should within a few years be able to produce 10 to15 kilograms of plutonium a year. The country already hasreprocessing capability, so it could in the future supplementits uranium bombs with lighter plutonium ones.

Diplomatic efforts to avert a catastrophe in the Indian sub-continent will now focus on persuading both of the newlydeclared nuclear powers to refrain from manufacturing war-heads and especially from placing them on missiles, whichcould strike each other’s cities within minutes. Missile tech-nology is well advanced in the region—thanks in part to U.S.supercomputers that IBM and Digital Equipment Corpora-tion supplied to India, according to Gary Milhollin of the Wis-consin Project on Nuclear Arms Control. The best possibleoutcome, Milhollin says, would be one in which India andPakistan are persuaded to join the Nuclear Non-ProliferationTreaty and give up the bomb. But in the near future, the chanc-es are not great. The world should hope both belligerentshave established effective controls on their deadly new toys.—Glenn Zorpette, with Tim Beardsley in Washington, D.C.

News and Analysis16 Scientific American August 1998

NUCLEAR TEST SITES AND FACILITIES enabled India and Pakistan to create the bomb.

LAU

RIE

GR

AC

E


Future historians may look backon 1998 as the year that parti-cle physics got interesting again.

For decades, the search for the funda-mental nature of matter has been re-duced to a jigsaw puzzle. The StandardModel of particle physics provided theframe, with outlines of each of the twodozen elementary particles sketched intheir proper places. When an army ofalmost 1,000 physicists discovered thetop quark in 1995, the puzzle seemedto be complete. Only a bit of bookkeep-ing remained: to confirm that the threelightest particles—the electron-, muon-and tau-neutrinos—indeed weigh exactly

nothing, as the Standard Model predicts.But in June the 120 Japanese and

American physicists of the Super-Kamio-kande Collaboration presented strongevidence that at least one of the neutri-nos (and probably all of them) weighssomething. That neutrinos have a smallmass is no small matter. It could helpexplain how our sun shines, how otherstars explode into brilliant supernovaeand why galaxies cluster in the patternsthat they do. Perhaps most important,explains Lincoln Wolfenstein, a physicistat Carnegie Mellon University, “onceyou accept that one neutrino has mass,you realize that the truth is somethingbeyond the Standard Model.”

Nearly all neutrino physicists haveaccepted the conclusion, because thenew data are supported by several yearsof similar observations at other detec-tors. John N. Bahcall of the Institute forAdvanced Study in Princeton, N.J., saysthe evidence “seems completely con-vincing to me. It is simply beautiful!”

The neutrino is certainly sublime inits subtlety. Quarks andelectrons are impossibleto miss; we and ourworld are made fromthem. The muon andtau, cousins of the elec-tron, are unfamiliar be-cause they die almost atbirth. But neutrinossurround us perpetual-ly, yet invisibly. Tril-lions zip through yourbody as you read this.Created by the bigbang, by stars and bythe collision of cosmicrays with the earth’s atmosphere, neutrinosoutnumber electronsand protons by 600million to one. “If theyhave a mass of just onetenth of an electron volt[an electron, in com-parison, weighs about500,000 eV], then neu-trinos would accountfor about as much massas the entire visible uni-verse,” says Joel R. Pri-mack, a cosmologist atthe University of Cali-fornia at Santa Cruz.

About 0.1 eV now

seems to many physicists a likely massfor the muon-neutrino. They can’t becertain yet, because the only way toweigh particles that can zoom almostunhindered through the earth at nearlythe speed of light is to do so indirectly.By patiently watching a high-tech cis-tern buried 2,000 feet underneath a Jap-anese mountain, physicists working onthe Super-Kamiokande project couldrecord faint flashes emitted on the ex-ceedingly rare occasions when a muon-or electron-neutrino collided with oneout of the 50,000 tons of water mole-cules in the tank. Over time, traces fromthose neutrinos that had been created inthe atmosphere started to reveal a pat-tern. Those arriving from above camein the expected proportion and number.“We even saw a hot spot toward the eastcaused by a well-known asymmetry inthe earth’s magnetic field” that createsmore cosmic-ray collisions in that di-rection, says Todd J. Haines of Los Ala-mos National Laboratory. But too fewmuon-neutrinos arrived from below.

Two large groups of physicists workedindependently to explain why. Both ruledout all explanations save one: the threekinds of neutrinos are not different par-ticles in the way that electrons andmuons are. Each neutrino is in fact amixture of three mass states. The mix-ture can change as the neutrino travels,transforming muon-neutrinos createdabove South America into heavier tau-neutrinos by the time they reach the de-tector in Japan. That is why too fewmuon-neutrinos appeared in the Super-Kamiokande tank; some had metamor-phosed into another, undetectable type.

Theorists figure that there is no way tohave mass states without having mass.But so far all that Henry W. Sobel, aphysicist at the University of Californiaat Irvine and spokesman for the collab-oration, can say is that the differencebetween the mass of muon-neutrinosand whatever they are changing into isbetween 0.1 and 0.01 eV—definitelynot zero.

Wolfenstein points out that “this doesnot solve the solar-neutrino problems,”the most baffling of which is the fact thatonly half the electron-neutrinos thattheoretically should fall from the sun tothe earth are actually detected here. ButBahcall adds that it does “strengthen theconviction of nearly everyone involvedin the subject that the explanation of


SCIENCE AND THE CITIZEN

A MASSIVE DISCOVERY

The weight of neutrinos offers clues to stars, galaxies and the

fate of the universe

PHYSICS

PHOTOMULTIPLIER TUBESlining the Super-Kamiokande detector record the

collision of neutrinos with water molecules.

KYO

DO

NEW

S IN

TERN

ATIO

NA

L, IN

C.


Modern science, like everysuccessful philosophy, hasaxioms that it takes on

faith to be true. Allan R. Sandage, oneof the fathers of modern astronomy,has just slid one of these precepts ontoan overhead projector. In letters toolarge to ignore, it hangs before the eyesof several hundred scientists, theolo-gians and others gathered here at theUniversity of California at Berkeley todiscuss the points of conflict and con-vergence between science and religion.The axiom is called Clifford’s dictum:“It is wrong always, everywhere andfor everyone to believe anything on in-sufficient evidence.”

Is there sufficient evidence to supporta belief in a Judeo-Christian God? Al-though many scientists working in theU.S. would doubtless agree with Sand-age that “you have to answer the ques-

the solar-neutrino problems is oscilla-tions” of neutrinos from one variety toanother on their journey to the earth.

“A little hot dark matter in the formof massive neutrinos may be just whatis needed” to help reconcile another as-trophysical accounting discrepancy, Pri-mack says. Many lines of evidence sug-gest that there is about 10 times morematter in the universe than human in-struments can see. Neutrinos will nowfill in some of that missing matter.

Whether neutrinos weighenough to make a significantdifference in the fate of the uni-verse and the composition ofmatter remains a mystery. “Wecan’t build theories on thiswithout firmer data about themasses and transition ampli-tudes,” says Steven Weinberg,one of the architects of theStandard Model and a profes-sor at the University of Texas.“We’re still far from that. Butthere are some very importantexperiments in the wings thatmay answer those questions.”

In January physicists will cre-ate a beam of muon-neutrinosin an accelerator near Tokyo

and aim it at the Super-Kamiokande de-tector. Within a few years, scientists atFermi National Accelerator Laboratoryin Batavia, Ill., hope to send swarms ofthe particles flitting toward a detectordeep in a Minnesota mine shaft. Thesecontrolled experiments may finally fillthe last holes in the Standard Modeland, if we are lucky, reveal it to be onlya part of a much larger, more beautifulpicture.

—W. Wayt Gibbs in San Francisco

SUPE

R-K

AM

IOK

AN

DE

CO

LLA

BO

RAT

ION

SUN SEEN IN “NEUTRINO LIGHT”should theoretically shine brighter than it does.

BEYOND PHYSICS

Renowned scientists contemplate the evidence for God

SCIENCE AND RELIGION


tion of what is ‘sufficient’ for yourself,”recent polls suggest that most of themwould nonetheless answer no. But theprogram for this conference includessome two dozen scientists, nearly all ofthem at the top of their fields, who havearrived at a different conclusion.

“There is a huge amount of data sup-porting the existence of God,” assertsGeorge Ellis, a cosmologist at the Uni-versity of Cape Town and an activeQuaker. “The question is how to evalu-ate it,” he says, because the data onlyrarely yield to scientific analysis.

Item one on Ellis’s list of evidence isthe so-called Anthropic Principle. In re-cent decades, explains John D. Barrow,an astronomer at the University of Sus-sex, physicists have noticed that manyof the fundamental constants of nature—

from the energy levels in the carbon atomto the rate at which the universe is ex-panding to the four perceptible dimen-sions of space-time—are just right for life.Tweak any of them more than a tad,and life as we know it could not exist.

John Polkinghorne, a particle physi-cist turned Anglican priest, points outother curiosities. “How is it that hu-mans’ cognitive abilities greatly exceedthe demands imposed by evolutionary

pressures, so that we canperceive the quantumnature of the universeand map its cosmic fea-tures?” he asks. Andwhy is mathematics sosurprisingly effective atdescribing the physicalworld? One possible ex-planation, Polkinghorne,Ellis and other well-re-spected physicists ar-gue, is that the universewas designed.

“Certainly more andmore top-level scientistsare considering the An-thropic Principle seri-ously in their work,”concedes Andrei D.Linde, a Stanford Uni-versity cosmologist. But he disputesthat the coincidences point to God. As-tronomical observations have so farsupported a so-called inflationary theo-ry of creation that Linde helped to de-velop. If the theory is correct, then ouruniverse is just one bubble in a muchlarger, eternal foam of universes. Theconstants and laws of physics may welldiffer in each bubble. Our universe may

be tuned for carbon-based life not be-cause it was set up that way, Barrowadds, but because even such a delicatearrangement was bound to happen inone of the myriad bubbles.

Science will probably never be able todetermine which case is true. “We arehitting the boundaries of what is evergoing to be testable,” Ellis says. And notjust in cosmology. “The science of the

OF ALL POSSIBLE UNIVERSES,only those similar to our own could support carbon-

based life. Was ours custom-made, or are we just lucky?

1,000

APPROXIMATE AGE OF THE UNIVERSE (BILLION YEARS)

15 (NOW)BIGBANG

LIFE CAN EVOLVE HERE

FLAT UNIVERSEUNIVERSES THATEXPAND FOREVER

UNIVERSES THATEVENTUALLY COLLAPSE

BIGCRUNCH

SIZ

E O

F TH

E U

NIV

ERSE

2,000

JOH

NN

Y JO

HN

SON


20th century is showing us, if anything,what is unknowable using the scientificmethod [and] what is reserved for reli-gious beliefs,” argues Mitchell P. Mar-cus, chairman of computer science atthe University of Pennsylvania. “Inmathematics and information theory,we can now guarantee that there aretruths out there that we cannot find.”

“The inability of science to provide abasis for meaning, purpose, value andethics is evidence of the necessity of reli-gion,” Sandage says—evidence strongenough to persuade him to give up hisatheism late in life. Ellis, who similarlyturned to religion only after he was wellestablished in science, raises other mys-teries that cannot be solved by logicalone: “the reasons for the existence ofthe universe, the existence of any physi-cal laws at all and the nature of the phys-ical laws that do hold—science takes allof these for granted, and so it cannotinvestigate them.”

“Religion is very important for an-swering these questions,” Sandage con-cludes. But how exactly? Pressing thescientists on the details of their beliefs re-

veals that most have carved a few coreprinciples out of one of the major reli-gious traditions and discarded the rest.“When you start pushing on the dog-ma, most scientists tend to part compa-ny,” observes Henry S. Thompson of theUniversity of Edinburgh.

Indeed, for Ellis, “religion is an exper-imental endeavor just like science: alldoctrine is a model to be tested, and noproof is possible.” Sandage confessesthat, like many other theoretical physi-cists, “I am a Platonist,” believing theequations of fundamental physics areall that is real and that “we see onlyshadows on the wall.” And Pauline M.Rudd, a biochemist at the University ofOxford, observes: “I have experiencesthat cannot be expressed in any lan-guage other than that of religion. Wheth-er the myths are historically true or falseis not so important.”

There seem only two points on whichall the religious scientists agree. ThatGod exists. And that, as Albert Einsteinonce put it, “science without religion islame; religion without science is blind.”

—W. Wayt Gibbs in Berkeley, Calif.


New Planet? A California astronomer has snappedwhat appears to be the first image of aplanet outside our solar system. Using

the Hubble SpaceTelescope, SusanTerebey photo-graphed the object,believed to be twoto three times themass of Jupiter, es-caping from a pairof young binarystars 450 light-yearsaway. The supposedplanet is connectedto the stars by a fila-

ment of light, probably an artifact of itstrajectory. Life on the new body is un-likely, as its surface temperature is sev-eral thousands of degrees. Even so, thediscovery could change ideas of howplanets usually form.

Polarized VisionResearchers have figured out why squiddon’t squint: like several other color-blind animals (but unlike people), theseleggy ocean dwellers are visually sensi-tive to polarized light. Roger T. Hanlon,director of the Marine Resources Centerat the Marine Biological Laboratory inWoods Hole, Mass., and his colleaguesdetermined that polarization helps toenhance the contrast of the squid’sblack-and-white vision—enabling it tobetter detect prey, such as plankton,that have evolved transparent bodiesfor protection from predators.

Nintendo Neurology Video games play with your brain. PaulM. Grasby and his colleagues at Ham-mersmith Hospital in London gave adrug called raclopride, which binds toreceptors for the neurotransmitterdopamine, to eight male volunteers.The drug was tagged with low-level ra-diation. Next, they monitored brain ac-tivity using positron emission tomogra-phy (PET) as the subjects played videogames or stared at blank screens. Dur-ing the games, raclopride binding de-creased in the striatum, indicating asurge of dopamine secretion there.Moreover, the increase in dopaminewas as large as that seen when subjectsare injected with amphetamines or thestimulant Ritalin.

IN BRIEF

More “In Brief” on page 24

A N T I G R AV I T Y

Hoop Genes

You know how close I came to playingprofessional basketball? About 17

inches. Seventeen inches taller, and I wouldhave been an even seven feet, which atleast would have given me a shot at seeingjust how close a shave Michael gets on thetop of his head. Having immense strength,acrobatic agility, catlike coordination, unbri-dled desire and a soft touch on my fall-away jumper would also have come inhandy, but the height thing couldn’t havehurt. All of which brings us to genes.

Although the Human Genome Projectwill probably fail to uncover a DNA sequencegoverning three-point shooting, British re-searchers have indeed found a jock gene.The gene in question, which comes in twoforms called I and D (for “insertion” and“deletion”), is for angiotensin-convertingenzyme (ACE), a key player in modulatingsalt and water balance, blood vessel dila-tion and maybe more. People carrying theD form are perfectly normal but will proba-bly be at home watching television reportsof people who have the I form plantingflags on the top of Mount Everest.

“We got into this system because it wasthought to control heart growth,” says

Hugh E. Montgomery of the University Col-lege London Center for Cardiovascular Re-search, who is the lead author of the ACEgene study, which appeared in the May 21issue of Nature. “We started looking at exer-cising athletes purely because their heartsgrow when they exercise.” Assuming that acrucial factor in endurance would be effi-cient use of oxygen, Montgomery and hiscolleagues tested “extreme endurance ath-letes exposed to low oxygen concentra-tions,” namely mountaineers, all of whomwere also British. (Why did the British re-searchers test British mountain climbers?Because they were there.)

Montgomery’s group found a dispropor-tionately high representation of the I formamong elite mountaineers, some of whomhad repeatedly climbed to 8,000 meters(over 26,000 feet) without supplementaloxygen. In another study, they followed 78army recruits through a 10-week trainingprogram. Those with I and D forms testedequally before the workouts began, butsoon the Is had it. All the exercisers increasedtheir endurance, as measured by curls of a15-kilogram (33-pound) barbell, but I folksimproved 11 times more than the Ds.

The studies bring to mind the eerie caseof Eero Maentyranta, winner of three cross-country skiing gold medals at the 1964Winter Olympics in Innsbruck. In 1993 re-

NA

SA A

ND

SU

SAN

TER

EBEY

Ext

raso

lar R

esea

rch

Cor

pora

tion



If diabetics don’t eat, they stop tak-ing insulin and can develop ke-toacidosis, a potentially life-threat-

ening complication of diabetes that canbe severe enough to send someone intoa coma. So when Nicole Lurie noticedthat many of her diabetic patients werecoming to Hennepin County MedicalCenter in Minneapolis with ketoacido-sis because they could not afford to eat,she was understandably alarmed. Lurie,a professor of medicine and publichealth at the University of Minnesota,along with her colleagues Karin Nelsonand Margaret E. Brown, decided to in-vestigate just how prevalent hungerwas among all their patients at the hos-pital. Now Lurie describes the diabetics

as “canaries in the coal mine,” provid-ing advance warning of hunger’s wide-spread threat to health.

The Minnesota team reported its find-ings in a recent issue of the Journal ofthe American Medical Association: ofthe 567 patients interviewed in early1997, 13 percent reported that duringthe previous year they had, on severaloccasions, not eaten for an entire day,because they could not afford food. Aseparate survey of 170 diabetics re-vealed that almost 19 percent had suf-fered complications that resulted fromnot having enough money to eat. Hen-nepin County Medical Center, which isa public hospital in an urban setting,treats many low-income patients. In-deed, Lurie and her colleagues note thatamong patients reporting hunger or un-certainty about when they might eatnext, many had annual incomes of lessthan $10,000. Many also shared an-other characteristic—recent reductionsin their food stamp benefits.

The 1996 welfare reform bill trimmed$27 billion over six years from the foodstamp program, which is run by the U.S.

Department of Agriculture (USDA). Theprogram provides food vouchers topeople with gross incomes of less than130 percent of federal poverty guide-lines. (That is, the income of a family ofthree must be less than around $1,450per month.) Nearly all recipients havefelt the effects of the cutbacks, but thereduction hit two groups particularlyhard. The new law denied food stampsto all legal immigrants (illegal immi-grants have never been eligible). And itput new financial pressure on unem-ployed recipients, by limiting the maxi-mum duration of benefits. Now unem-ployed able-bodied adults younger than50 with no dependents may receivefood stamps for no longer than threemonths during any three-year period.

Congress recently voted to restorefood stamp benefits to certain categor-ies of legal immigrants—such as chil-dren, the elderly and the disabled—afterresearchers documented that thesegroups were being harmed by the cuts.One disturbing study by Physicians forHuman Rights, for example, found that79 percent of legal immigrant house-holds surveyed in March were “food in-secure”—meaning they often went hun-gry and worried about when their nextmeal would be. The study coordinator,Jennifer Kasper of the Boston Universi-ty School of Medicine, notes that be-cause hunger has been linked to asth-ma, diarrhea and anemia in children,“policymakers should consider the truecost of cutting food stamps.”

Lurie agrees. “Cuts in food stampsmay not be benign,” she says. “Squeezefood stamps, and people can end up inthe hospital.” She also emphasizes thatphysicians should talk to their patientsabout hunger. “Doctors need to under-stand when patients are making trade-offs, buying food instead of medicine.People don’t volunteer this informa-tion—doctors have to ask,” she states.

When Lurie and her colleagues asked,they found that 8 percent of their dia-betic patients had decreased or stoppedtaking their insulin because they had noteaten enough to need the normal dose.Getting enough food and enough medi-cine can be a challenge not just for dia-betics: a survey by Second Harvest, thenation’s largest charitable hunger-relieforganization, found that 28 percent ofpeople visiting food banks have at somepoint been forced to choose betweengetting medical care and buying food.

Researchers say it is notable that thesealarming new findings were obtained

searchers found that Maentyranta and a lotof his family have a genetic mutation affect-ing red blood cell production. As a result,Maentyranta’s blood carries up to 50 per-cent more hemoglobin than an averagemale’s. And whereas having, say, a thirdlung might get you disqualified from com-petition, a set of genes that grants you ac-cess to half again as much oxygen as thecompetition is still legal.

Biology’s newfound ability to spot indi-vidual genes associated with athleticism isintriguing but, on contemplation, not sur-prising. “Athletics, by already selecting forextreme phenotypes, must be se-lecting for a significant genetic influ-ence,” says Philip Reilly, executive di-rector of the Shriver Center for Men-tal Retardation in Waltham, Mass., aclinical geneticist and an attorneywho has spent the past 25 yearspondering genetics and its implica-tions for society. “In one sense, thisis old news,” he notes. “There is nophysical trait for which we have bet-ter evidence of substantial geneticcontribution to ultimate expressionof phenotype than height.” And formost athletes, jockeys being the ri-diculously obvious exception, beingbig is advantageous. (The late, greatsportswriter Shirley Povich actually

bemoaned all the tallness. “[Basketball] lostthis particular patron back when it went ver-tical and put the accent on carnival freaks,”he wrote. “They don’t shoot baskets any-more, they stuff them, like taxidermists.”)

So genes for robust physical traits helpmake for a good athlete. But the basis forbehavior, which also contributes to athleti-cism, remains more mysterious. Another 45army recruits, with the same proportion of Iand D forms as the 78 who trained for 10weeks, dropped out entirely to go home.“Sometimes,” Montgomery explains, “theyjust miss their mums.” —Steve Mirsky

MIC

HA

EL C

RA

WFO

RD

FOOD FOR THOUGHT

Cutbacks in the food stamp program are proving hazardous

to the nation’s health

PUBLIC HEALTH


Hardly an astronomical an-nouncement makes the frontpages without being said to

overturn all existing theories. Gamma-ray bursts are one of the few things forwhich this has actually been true. Theseseemingly random flashes—which, ifyou had gamma-ray vision and didn’tblink at the wrong time, would outshinethe rest of the sky—were long thoughtto originate in our galaxy. But that the-ory foundered in 1991, when the Comp-ton Gamma Ray Observatory satellitedetected bursts all over the sky, not onlyin the Milky Way. Another hypothesis,in which bursts occur just outside ourgalaxy, crumbled last year when the firstdistance measurement of a burst put ittoo far away. Although astronomers nowagree that the bursts are some kind ofmegaexplosion in distant galaxies, they

still appear on most top-10 lists of cos-mic mysteries.

This past spring astronomers report-ed the distances to two new bursts—

which, at first glance, seemed to scuttlesome of the few remaining plausible ex-planations. The first of these bursts,spotted last December 14 by the Beppo-SAX satellite, occurred in a galaxy 12billion light-years away, according toobservations by Shrinivas R. Kulkarniand S. George Djorgovski of the Cali-fornia Institute of Technology. To be sobright at such a distance, the burst musthave shone more brilliantly than anyobject previously recorded.

Just as astronomers were reconcilingthemselves to the unexpected distanceand brilliance of this burst, along cameanother on April 25 that was unexpect-edly near and dim: nearly 100 timescloser and 100,000 times dimmer thanthe December event. Even stranger, thisburst was followed not by the usual af-terglow of less energetic radiation butby a supernova—the first time an ex-ploding star and gamma burst have beenseen together. The supernova appearedin visible-light images made by Titus J.Galama of the University of Amsterdam


In Brief, continued from page 22

Spinach Power Popeye had better join the bombsquad. Scientists at the Department ofEnergy’s Pacific Northwest National Lab-

oratory have found thatspinach enzymes can

neutralize danger-ous explosives,such as TNT. Nitro-reductase, foundin sundry leafy

greens, can eat, di-gest and transform

these compounds intoless toxic by-products,

which can then be used by industry orbe further reduced to carbon dioxideand water.

Hot Finding on MarsPhilip R. Christensen of Arizona StateUniversity made a startling announce-ment at the American GeophysicalUnion’s spring meeting in Boston: datafrom the Thermal Emission Spectrome-ter (TES) on board the Mars Global Sur-veyor—a spacecraft now in orbit aroundthe Red Planet—indicated the presenceof a large deposit of coarse-grainedhematite. This rock is formed by thermalactivity or bodies of water, and so geol-ogists take the TES finding as further ev-idence that Mars was once warmer andmay have supported life. They hope tostudy the deposit—which measuressome 300 miles in diameter—moreclosely during the 2001 Lander mission.

Fast ExtinctionAt the end of the Permian period, morethan 85 percent of all ocean species and70 percent of terrestrial vertebrate gen-era died out. A study in Science on May15 shows that this vanishing act hap-pened fast—in less than a million years.Douglas H. Erwin of the National Muse-um of Natural History and his co-work-ers examined 172 samples of volcanicash from various levels at three loca-tions in southern China and one inTexas. The samples were readily datedbecause they contained zircon. Thismineral incorporates uranium into itscrystal lattice when it forms; over time,the element decays into lead and tellsthe crystal’s age. Erwin found that mostspecies had disappeared between 252.3million and 251.4 million years ago. Al-though the extinction’s cause is still un-clear, its speed does rule out an earliertheory implicating plate tectonics. (SeeJuly 1996, page 72.)

More “In Brief” on page 26

within two years after the1996 welfare bill was passed.As the law stands now, manyof the cuts in food stampswill continue for anotherfour years. J. Larry Brown,director of the Tufts Universi-ty Center on Hunger, Povertyand Nutrition Policy, saysthe USDA has only recentlystarted monitoring hungerand food insecurity moresystematically.

Because it takes a coupleof years to obtain a clear pic-ture of hunger across the na-tion, Brown explains that heand others will be conducting smaller-scale surveys. Results should come inwithin several months, “enabling us toshow trends faster,” he notes. Brownpoints out the importance of identifyinghunger and food insecurity before peo-ple wind up in the hospital, as they did

in the Minnesota study. “By the time[hunger] shows up clinically, these peo-ple have been without food for a longtime,” he says. “I can separate my hu-man side from my scientific side, andon the human side the situation is veryfrightening.” —Sasha Nemecek

BRIGHT LIGHTS,

BIG MYSTERY

The brightest flash ever seen has revealed exotic celestial objects

ASTRONOMY

B. D

AEM

MRI

CH

Imag

e W

orks

FEEDING THE HUNGRYoften falls to food banks, such as this branch of the

Salvation Army in San Antonio, Tex.

DAV

IES

& S

TARR

, IN

C. L

iais

on In

tern

atio

nal


and in radio observations by Mark H.Wieringa of the Australia Telescope Na-tional Facility. This was no normal su-pernova, either. It dimmed more slowlythan others with its spectral character-istics, and it spewed out more radiopower than any supernova seen before.

Amid these events, a third burst wentoff—the burst of hype. A widely quotedNational Aeronautics and Space Ad-ministration press release called the De-cember event “the most powerful ex-plosion since the creation of the uni-verse,” which slighted any upheavalsthat release energy in forms we cannotyet detect. The hyperbole also madegamma bursts out to be something rad-ically different from supernovae, whenin fact both probably entail the death ofa star and the conversion of much of itsmass into pure energy.

In a supernova, the energy conversioninvolves the thermonuclear detonationof a white dwarf star or the implosionof a massive star. But such cataclysmscould not produce gamma bursts: thestellar debris would choke off the gam-ma rays. Bursts must come from low-density sources. And yet their energymust ultimately originate in high-den-sity volumes 100 kilometers across atmost, or else the bursts would not flashand flicker as they do. These two re-quirements imply that the productionof energy and the generation of gam-mas occur in separate locations.

The second of these steps has beenexplained by Martin J. Rees of the Uni-versity of Cambridge and Peter I.Mészáros of Pennsylvania State Univer-sity. A jet of radiation and electronssquirting out at near light speed wouldoutrun any interfering debris. After thisfireball had ballooned to several hun-dred million kilometers in size, shock

waves within it would give off the gam-ma flash; later, collisions with surround-ing matter could emit an afterglow. Ra-dio observations by Dale A. Frail of theNational Radio Astronomy Observato-ry have supported this explanation.

But what would produce enough en-ergy to create such a jet? In one sce-nario, proposed in 1986 by BohdanPaczynski of Princeton University, twoneutron stars or a neutron star and ablack hole orbit ever more tightly andeventually unite in a passionate but ru-inous embrace. Stan E. Woosley of theUniversity of California at Santa Cruzsuggested another possibility in 1993.Perhaps “hypernovae,” also called “col-lapsars”—souped-up supernovae thatoccur when stars are too massive to un-dergo normal supernovae—power thecosmic flashes.

Despite some observers’ initial claims,the brightness of the bursts may not beenough to decide between these models.Both lead to the same kind of object: ablack hole surrounded by a ring of de-bris. And both leave the same questionsunanswered: What triggers the fireball?Is the radiation focused into beams oremitted uniformly? Pinpointing thebursts’ locations within their galaxiesmay supply clues. Neutron stars wan-der far from their places of birth beforemerging, whereas the stars that under-go hypernovae stay put. So if bursts tendto occur in star-forming regions—as afew shreds of evidence now suggest—hypernovae seem the more likely source.

But there is a third possibility. The di-versity of these latest bursts, Woosleysays, suggests that each model accountsfor some bursts. This time, rather thanoverturning all the theories, observersmay have done the opposite: confirmedthem, all of them. —George Musser


Practical ChaosResearchers have at last drawn a credi-ble link between chaos theory andepileptic seizures. Klaus Lehnertz andChristian E. Elger of the University ofBonn collected 68 electroencephalo-grams from 16 patients. They thenmade use of a mathematical propertyfrom chaos theory called dimension,which is essentially a measure of com-plexity. The pair found that 11 minutesbefore the onset of seizure, there is acharacteristic loss of complexity in thefiring of neurons near the epileptic fo-cus in the brain. The finding may helpscientists develop more sensitive instru-ments to detect oncoming seizures—and better ways to prevent them.

Stained-Glass Myth Investigators have smoothed the lumpsout of an old theory, finding that me-dieval stained glass is not thicker at thebottom because it has slowly flowed

downward. EdgarDutra Zanotto of theFederal University ofSão Carlos in Brazilcalculated how longit would take for aviscous liquid toflow enough to

change the visible thickness of glass.His conclusion, published in May in theAmerican Journal of Physics, was thatcathedral glass would need a period“well beyond the age of the universe.”In fact, medieval stained glass is proba-bly thicker at the bottom because of12th-century manufacturing methods.

MagnetarsSix years ago Robert C. Duncan of theUniversity of Texas at Austin and Chris-topher A. Thompson of the University ofNorth Carolina at Chapel Hill predictedthe existence of magnetars, ultradensestars with tremendous magnetic fields.Now they have been found. In May inNature, Chryssa Kouveliotou of NASA’sMarshall Space Flight Center and an in-ternational team described the x-raymeasurements of SGR 1806-20. Themeasurements showed that the objectis most likely spinning at a rate charac-teristic of a heavy, supermagnetic star.The team estimated SGR 1806-20’smagnetic field to be a whopping 800trillion gauss—or 100 times strongerthan the magnetic fields around anyother known stars. —Kristin Leutwyler

In Brief, continued from page 24

SA

UNUSUAL SUPERNOVA BECAME VISIBLEshortly after a gamma-ray burst this spring, linking the two phenomena.

TITU

S J.

GA

LAM

A E

urop

ean

Sout

hern

Obs

erva

tory

RIEG

ER B

ERTR

AN

D G

amm

a Li

aiso

n



Some 5,000 species of amphibians inhabit the world,mostly frogs, toads and salamanders, and they seem to

be dying at unprecedented rates. This situation has raisedalarm because amphibians are widely regarded as uniquelysensitive indicators of the planet’s health. Much of the dam-age to amphibians comes from habitat destruction, particu-larly the draining of wetlands, but what has scientists particu-larly worried are the declines and apparent extinctions in ar-eas far removed from obvious human intrusion, such as thecloud forest at Monteverde, Costa Rica, where the goldentoad, once abundant, hasnot been seen since 1989.

Do reports such as theseindicate a worldwide am-phibian crisis? Not neces-sarily, according to JosephPechmann of Florida Inter-national University, whosework suggests that report-ed declines and extinctionsin near-pristine environ-ments could simply be nat-ural year-to-year variations:a drought, for instance, thataffects egg laying and lar-vae survival. Because ofsuch fluctuations, it is oftenimpossible, in the absenceof more complete histori-cal information, to judgewhether a reported declineis natural or a reaction tohuman activity. On the oth-er hand, many researchers,including Andrew R. Blau-stein of Oregon State Uni-versity, believe the reportedrates of decline and extinc-tion are so extraordinarythat they cannot be a partof the natural cycle.

If the reported declinesand extinctions are indeedhighly abnormal—and thisseems to be the majoritypoint of view—what mighthave led to them? A varie-ty of causes is probably atwork here. Among the lead-ing culprits are acid rain, synthetic chemicals, metallic con-taminants and infectious diseases. Excessive ultraviolet radia-tion (presumably caused by the thinning of the ozone layer)working synergistically with fungal disease may explain someof the declines. Another possible cause is global warming–re-lated droughts, a particular threat to amphibians, which gen-erally require high humidity or an aquatic environment.

There has also been a recent surge of reported amphibianabnormalities, such as missing limbs and eyes. One of the fac-

tors could be a substance like retinoic acid, which may bepresent in water naturally or as a residue from pesticides; itproduces birth defects in vertebrates. It is not clear, however,whether the increase in reports stems from greater humanactivity or is simply the result of more surveys: reports on am-phibian abnormalities are not new and have fluctuated innumber since 1700.

One of the few areas with reasonably complete informationis the U.S. The Nature Conservancy and the Natural HeritageNetwork have identified 242 native amphibian species that

have inhabited the 50states since the beginningof European settlement.Three of these species arepresumed to be extinct; anadditional three are classi-fied as possibly extinct.Moreover, 38 percent areclassified as imperiled orvulnerable, a level that sug-gests that amphibians aremore affected by humanactivity than birds, mam-mals and reptiles, which areat far lower levels of risk inthe U.S. Freshwater fish areat about the same level ofrisk as amphibians; crayfishand freshwater mussels areat substantially higher lev-els of risk.

As the top map shows,the number of amphibianspecies in the U.S. variesconsiderably, with Califor-nia and the South havingthe largest number. Thebottom map shows thatamphibians tend to be atgreater proportional risk inthe Far West. In much of theWest, habitats are discon-tinuous, making it more dif-ficult for locally threatenedgroups to be recolonizedfrom other areas. In Califor-nia’s Central Valley, pesti-cides and herbicides aresuspected of contributing

to amphibian declines. Amphibians in certain parts of theWest may be adversely affected by the introduction of nonna-tive species such as the American bullfrog, which competewith native amphibians for food, or by the introduction ofsalmon and trout, which eat amphibian eggs, tadpoles andeven adults. The Southeast—which has more areas of contin-uous suitable habitat (such as the southern Appalachians)and a warm, moist climate—is better suited for recolonization.

—Rodger Doyle ([email protected])

TOTAL NUMBER OF AMPHIBIAN SPECIES

LESS THAN 20 20 TO 29 30 TO 49 50 OR MORE

SPECIES AT RISK (PERCENT)

0 1 TO 9 10 TO 19 20 OR MORE

SOURCE: The Nature Conservancy and Natural Heritage Network in cooperation with theAssociation for Biodiversity Information. All data are for 1997. Excluded are amphibianspecies not native to their areas.

B Y T H E N U M B E R S

Amphibians at Risk

ROD

GER

DO

YLE


JCraig Venter, the voluble directorof the Institute for Genomic Re-search (TIGR) in Rockville, Md.,is much in demand these days. A

tireless self-promoter, Venter set offshock waves in the world of human ge-netics in May by announc-ing, via the front page of theNew York Times, a privatelyfunded $300-million, three-year initiative to determinethe sequence of almost all thethree billion chemical unitsthat make up human DNA,otherwise known as the ge-nome. The claim promptedincredulous responses frommainstream scientists en-gaged in the internationalHuman Genome Project,which was started in 1990and aims to learn the com-plete sequence by 2005. Thispublicly funded effort wouldcost about 10 times as muchas Venter’s scheme. But Ven-ter’s credentials mean thatgenome scientists have totake his plan seriously.

In 1995 Venter surprisedgeneticists by publishing thefirst complete DNA sequenceof a free-living organism, thebacterium Haemophilus in-fluenzae, which can causemeningitis and deafness. Thisachievement made use of athen novel technique knownas whole-genome shotguncloning and “changed all theconcepts” in the field, Venterdeclares: “You could see thepower of having 100 percentof every gene. It’s going to be the futureof biology and medicine and our spe-cies.” He followed up over the next twoand a half years with complete or par-tial DNA sequences of several more mi-crobes, including agents that cause Lymedisease, stomach ulcers and malaria.

The new private human genome ini-tiative will be conducted by a company

(yet to be named) that will be owned byTIGR, Perkin-Elmer Corporation (theleading manufacturer of DNA sequenc-ers) and Venter himself, who will be itspresident. He plans to warm up for thehuman sequence next year by knocking

off the genome of the fruit fly Drosoph-ila melanogaster, an organism usedwidely for research in genetics.

Venter, 51, has a history of lurchinginto controversy. As an employee of theNational Institutes of Health in the ear-ly 1990s, he became embroiled in a dis-pute over an ultimately unsuccessful at-tempt by the agency to patent hundreds

of partial human gene sequences. Ven-ter had uncovered the partial sequences,which he called expressed sequence tags(ESTs), with a technique he developedin his NIH laboratory for identifying ac-tive genes in hard-to-interpret DNA.“The realization I had was that each ofour cells can do that better than thebest supercomputers can,” Venter states.

Many prominent scientists, includingthe head of the NIH’s human genomeprogram at the time, James D. Watson,opposed the attempt to patent ESTs,saying it could imperil cooperationamong researchers. (Venter says the NIH

talked him into seeking the patents onlywith difficulty.) And at a congressional

hearing, Watson memorablydescribed Venter’s automat-ed gene-hunting technique assomething that could be “runby monkeys.” An NIH col-league of Venter’s respondedlater by publicly donning amonkey suit.

Venter left the NIH in 1992feeling that he was beingtreated “like a pariah.” Andhe does not conceal his irri-tation that his peers wereslow to recognize the meritsof his proposal to sequenceH. influenzae. After failingto secure NIH funding for theproject, Venter says he turneddown several tempting invi-tations to head biotechnolo-gy companies before finallyaccepting a $70-million grantfrom HealthCare InvestmentCorporation to establishTIGR, where he continuedhis sequencing work. Today,when not dreaming up auda-cious research projects, Ven-ter is able to relax by sailinghis oceangoing yacht, theSorcerer.

His assault on the humangenome will employ thewhole-genome shotgun clon-ing technique he used on H.influenzae and other mi-crobes. The scheme almost

seems designed to make the HumanGenome Project look slow by compari-son. To date, that effort has devotedmost of its resources to “mapping” thegenome—defining molecular landmarksthat will allow sequence data to be as-sembled correctly. But whole-genomeshotgun cloning ignores mapping. In-stead it breaks up the genome into mil-


PROFILEAn Express Route to the Genome?

In his race to beat the Human Genome Project, J. Craig Venter has riled geneticists everywhere

AHEAD OF THE PACK: J. Craig Venter admires a model of his racing yacht.

JEA

N-C

HRI

STIA

N B

OU

RC

AR

T G

amm

a Li

aiso

n

.


lions of overlapping random fragments,then determines part of the sequence ofchemical units within each fragment.Finally, the technique employs power-ful computers to piece together the re-sulting morass of data to re-create thesequence of the genome.

Predictably, Venter’s move promptedsome members of Congress to questionwhy government funding of a genomeprogram was needed if the job could bedone with private money. Yet if the goalof the Human Genome Project is to pro-duce a complete and reliable sequence ofall human DNA, says Francis S. Collins,director of the U.S. part of the project,Venter’s techniques alone cannot meetit. Researchers insist that his “cream-skimming” approach will furnish infor-mation containing thousands of gapsand errors, even though it will haveshort-term value. Venter accepts thatthere will be some gaps but expects ac-curacy to meet the 99.99 percent targetof the existing genome program.

Shortly after Venter’s proposed schemehit the headlines, publicly funded re-searchers started discussing a plan tospeed up their own sequencing timeta-ble in order to provide a “rough draft”of the human genome sooner than orig-

inally planned. Collins says this propos-al, which would require additional fund-ing, would have surfaced even withoutthe new competition. Other scientiststhink Venter’s plan will spur the publicresearchers forward.

Venter has always had an iconoclas-tic bent. He barely graduated from highschool and in the 1960s was happilysurfing in southern California until hewas drafted. Early hopes of training forthe Olympics on the navy swim teamwere dashed when President Lyndon B.Johnson escalated the war in Vietnam.But Venter says he scored top marks outof 35,000 of his navy peers in an intelli-gence test, which enabled him to pursuean interest in medicine. He patched upcasualties for five days and nights with-out a break in the main receiving hospi-tal at Da Nang during the Tet offensive,and he also worked with children in aVietnamese orphanage.

Working near so much needless death,Venter says, prompted him to pursueThird World medicine. Then, while tak-ing premed classes at the University ofCalifornia at San Diego, he was bittenby the research bug and took up bio-chemistry. He met his wife, Claire M.Fraser, now a TIGR board member,

during a stint at the Roswell Park Can-cer Institute in Buffalo, N.Y., and tookhis research group to the NIH in 1984.

His painstaking attempts to isolateand sequence genes for proteins in thebrain known as receptors started tomove more quickly after he volunteeredhis cramped laboratory as the first testsite for an automated DNA sequencermade by Applied Biosystems Interna-tional, now a division within Perkin-Elmer. Until then, he had sequenced justone receptor gene in more than a decadeof work, so he felt he had to be “farmore clever” than scientists with biggerlaboratories. Venter was soon employ-ing automated sequencers to find moregenes; he then turned to testing proto-cols for the Human Genome Project,which was in the discussion phase.

After leaving the government andmoving to TIGR, Venter entered a con-troversial partnership with Human Ge-nome Sciences, a biotechnology compa-ny in Rockville established to exploitESTs for finding genes. The relationship,never easy, foundered last year. Accord-ing to Venter, William A. Haseltine, thecompany’s chief executive, became in-creasingly reluctant to let him publishdata promptly. Haseltine replies that he



often waived delays he couldhave required.

The divorce from Human Ge-nome Sciences cost TIGR $38million in guaranteed funding.The day after the split was an-nounced, however, TIGR start-ed to rehabilitate itself with asuspicious scientific communityby posting on its World WideWeb site data on thousands ofbacterial gene sequences.

The sequencing building forVenter’s new human genomecompany, now under construc-tion adjacent to TIGR, shouldbe a technological mecca. It willproduce more DNA data thanthe rest of the world’s output com-bined, employing 230 Perkin-ElmerApplied Biosystems 3700 machines,which are now in final development.These sophisticated robots, which willsell for $300,000 apiece, should requiremuch less human intervention thanstate-of-the-art devices. Venter says thenew venture will release all the humangenome sequence data it obtains, atthree-month intervals. It plans to makea profit by selling access to a databasethat will interpret the raw sequence

data as well as crucial information onvariations between individuals thatshould allow physicians to tailor treat-ments to patients’ individual geneticmakeups. Most of the federal sequenc-ing centers do not look at the data theyproduce, Venter thinks, but just put itout as if they were “making candy barsor auto parts.”

The unnamed new TIGR–Perkin-El-mer operation will also patent severalhundred of the interesting genes it ex-pects to find embedded within the hu-

man genome sequence. Venterdefends patents on genes, sayingthey pose no threat to scientificprogress. Rather, he notes, theyguarantee that data are avail-able to other researchers, be-cause patents are public docu-ments. His new venture will notpatent the human genome se-quence itself, Venter states.

The plan, if it comes to pass,should be a boon for biomedicalresearch. Yet more than 100 sci-entists who met recently tomake further plans for the exist-ing, more complete and thor-ough international Human Ge-nome Project were unanimous

that it, too, should go forward. Collinswonders whether the entrepreneur atTIGR will be able to stick to his statedpolicy of releasing all data, given finan-cial exigencies. But he and Venter haveboth pledged to cooperate.

Venter is unfazed. “We will make thehuman genome unpatentable” by plac-ing it in the public domain, he pro-claims. For the next three years, all eyeswill be on Venter to see how closely hecan approach that goal.

—Tim Beardsley in Washington, D.C.

READOUT OF GENETIC DATA comes from a DNA sequencing machine. More

advanced units will tackle the entire human genome.

PERK

IN-E

LMER

CO

RPO

RAT

ION


The 1990s have seen the emer-gence of a conciliatory credo:business and environmental

interests are not just compatible, theyare inextricable. A healthy environmentand natural-resource base are prerequi-sites for a healthy economy; a smoothlyfunctioning world market, in turn, pro-duces the resources and mind-set need-ed to protect the environment. Conceptssuch as “sustainable development” and“eco-labeling” have accordingly prom-ised consumers a hand in making themarket environmentally accountable.

Although the rhetoric seems reason-able to many on both sides of the divide,it is proving difficult to implement. Asan April decision by the World TradeOrganization (WTO) indicates, the in-terests of one nation’s consumers maybe consistently forced to yield to the in-terests of free trade. The ruling statedthat a U.S. law prohibiting the importof shrimp caught in nets that can en-trap sea turtles was a barrier to trade.According to the WTO, the U.S. mustimport shrimp from Thailand, Malay-sia, India and Pakistan—regardless ofwhether the harvests endanger sea tur-tles—or face large fines.

The WTO ruling—which the U.S. saysit will appeal—is not the first to subsumea nation’s environmental laws or its con-sumers’ desires to the demands of glob-al markets. Europeans are legally re-quired to import U.S. hormone-treatedbeef, Americans must stomach tunafrom Mexico that endangers dolphins,and the U.S. Environmental ProtectionAgency lowered air-quality standardsto allow imports of reformulated gaso-line. The shrimp ruling, however, hasmobilized the environmental communi-ty in an unprecedented fashion.

Part of the reason is that all sevenspecies of sea turtle are endangered. Pop-ulations of these mysterious, long-livedcreatures have plummeted in the past

50 years. But the other stimulus for theoutcry is a growing concern over theWTO’s environmental mores and therole of so-called processing and produc-tion methods. Greatly simplified, tradelaw states that one country cannot ex-clude another’s product if it is “like”most such products. Therefore, the U.S.cannot ban Asian shrimp if it looks andtastes like other shrimp. The productionmethod is not a consideration. “That isa distinction that the trade system doesnot want to recognize,” explains DanielA. Seligman of the Sierra Club. And itis on precisely that distinction that eco-labeling and consumer choice hinge:Was wood harvested sustainably? Didshrimpers harm an endangered species?

“Industry has increasingly been argu-ing that labeling is a trade barrier be-cause it establishes different standards,”Seligman continues. In other words, ifrulings such as this one continue to hold,trade could lose some of its power toenact environmental change.

At the same time, the WTO rulinghardly represents a black-and-white caseof trade trumping environment. Just asAmericans resist being force-fed a prod-uct, the Asian countries are resisting U.S.efforts to force its domestic law downtheir throats. Further compli-cating the issue is the murkyhistory of the U.S. position:the Clinton administrationfinds itself defending a legalinterpretation it did not coun-tenance in the first place.

The shrimp conflict has itsroots in Public Law 101-162,Section 609, which the federalgovernment enacted in 1989.The statute stated that theU.S. would not buy shrimpcaught in nets without turtle-excluder devices, or TEDs.These grills sit in the necks ofshrimpers’ nets, allowingsmall crustaceans throughbut stopping larger creaturesand diverting them out of thetrap. By 1993 the govern-ment had effectively appliedthe law to fishermen plyingthe Caribbean and westernAtlantic. (U.S. shrimpers hadbeen required to use TEDssince 1987.)

As applied, the law did littleto protect sea turtles world-

wide. Shrimp caught in Asia, for exam-ple, was not prohibited from U.S. mar-kets. So several organizations, includingthe Earth Island Institute, challengedthe government, claiming that catchesof all wild shrimp come under the law’sjurisdiction. In December 1995 the U.S.Court of International Trade agreed: theU.S. must embargo all shrimp caught inTED-less nets.

The government then interpreted thelaw to mean that foreign imports shouldbe checked on a shipment-by-shipmentbasis. But the plaintiffs argued that suchchecking does not protect sea turtles,because TED-less boats could hand offtheir catch to TED-certified trawlers.

So, in October 1996, the trade courtruled against the shipment-by-shipmentinterpretation. (The future of the rulingis unclear, as it was recently vacated, orannulled, during appeal.) Thus, Ameri-can law mandated that other countriesset up and enforce regulations requiringTEDs if they wanted access to the U.S.market. Countries were given a fewmonths to do so.

Things quickly became dicey. Thai-land—which already used TEDs—joinedIndia, Malaysia and Pakistan in oppos-ing the ban. These nations argued that


TECHNOLOGY AND BUSINESS

TRADE RULES

A World Trade Organization decision about sea turtles raises

doubts about reconciling economicsand the environment

ENVIRONMENTAL POLICY

LOGGERHEAD TURTLE WAS CAPTURED LIVEon a shrimp trawler with other bycatch. Many other sea turtles aren’t so lucky.

MIK

E W

EBER

Cen

ter f

or M

arin

e C

onse

rvat

ion


In May the European Union movedone step closer to economic super-power when it officially adopted

the euro as the single currency for 11 ofits member nations. But as politicianscelebrated, computer managers aroundthe world issued a collective moan.They must now refit every system thatdeals in marks, francs,guilders and othercoins of the continentto handle the new cur-rency, which will bephased in starting inJanuary 1999. Facedwith both the euro andwidespread year 2000bugs, “a lot of com-panies in Europe willfind themselves in direstraits,” predicts AchiRacov, the chief infor-mation officer of theLondon-based NatWestGroup. He expects thathis firm will spend inexcess of £200 million

(about $320 million) to enable its sys-tems to handle both the euro and theturn of the millennium.

Adding a new currency to financialsystems sounds straightforward, but itis not. The European Commission inBrussels has published strict rules thatcomputers must follow. Converting fromfrancs to marks, for example, could soonrequire “triangulation,” in which francsare changed first into euros, then intomarks. Some software will need to dis-play both the local currency and theeuro during the transition period, from1999 to 2002. And terabytes of histori-cal financial data must eventually beconverted into euros.


the U.S. had no right to mandate theirdomestic policy and that the U.S. hadnot approached them to negotiate anagreement. “To me, this is perfectly rea-sonable for these four countries to saythis is something that we do not buy andwe will not do it,” says Jagdish N. Bhag-wati, a professor of economics at Co-lumbia University. “The great advantageof rulings like this is that it forces theU.S. to go talk to countries and get intosome discussion with the countries.”

Such discussion did take place withCaribbean and Latin American nations.An Inter-American Convention on theProtection and Conservation of Sea Tur-tles was drafted in 1996; as of this spring,only six of the original 25 had signed it,however. “Multilateral agreement is theway to go,” agrees Deborah T. Crouse,formerly of the Center for Marine Con-servation. “But it is also much slowerand more labor intensive. It took twoyears to get the Inter-American Conven-

tion, and the U.S. has not ratified or im-plemented it.”

None of the four Asian countries con-test that sea turtles are endangered. Sev-eral have conservation efforts in place,such as closing beaches to allow nesting,and have sought out TED trainingcourses. Each has also signed the Con-vention on International Trade in En-dangered Species (CITES), which listsall sea turtles as threatened.

But it is the muddy legal area wherethe WTO and CITES—as well as otherenvironmental treaties—intersect thatconcerns environmentalists. CITES pro-hibits the trade of endangered species;the WTO prohibits barriers to tradesuch as objections to processing and pro-duction methods. What should happenwhen a production method threatensan endangered species? The WTO offersnothing but ambiguity on this front.

Formed in 1995 to replace the GeneralAgreement on Tariffs and Trade (GATT),

THE OTHER COMPUTER PROBLEM

Will computers be ready forthe euro and the year 2000?

COMPUTER SCIENCE

YV

ES L

OG

GH

E A

P Ph

oto

EURO—THE NEW, SINGLE CURRENCY for 11 European nations—will require extensive

software modifications to financial computer systems.


The euro will also have a widespreadsecondary effect on computers. No long-er vulnerable to fluctuations in mone-tary exchange rates, companies will in-crease their use of cheaper supplies andworkers in other countries. Such opera-tional changes will require revised ornew software. “The euro touches justabout every part of British Airwaysfrom a business point of view,” assertsNigel T. Barrow, euro computer man-ager for the airline. All told, the cost ofmodifying systems in Europe (includingthose owned by U.S. multinational cor-porations) will run between $150 bil-lion and $400 billion, states the GartnerGroup, a consultancy based in Stam-ford, Conn.

Euro conversion projects at this timewill cost more, and will progress moreslowly, than they otherwise would be-cause they will be competing for quali-fied programmers against the year 2000effort. The U.S. is already short morethan 100,000 information technologyprofessionals, according to industry es-timates. That number is expected to bal-loon in the near future.

One obvious way to solve the prob-lem of the two largest software projectsin history occurring almost simultane-ously would be to push back the dead-line for euro compliance from 2002 to2005, advises Capers Jones, chairmanof Software Productivity Research inBurlington, Mass. Otherwise, he claims,


the WTO does make environmental pro-visions. The preamble to its charter men-tions sustainable development; ArticleXX of GATT, which still holds for theWTO, offers exceptions to the rule offree trade; and the organization set upthe Committee on Trade and the Envi-ronment. But that committee has notclarified the relation between the WTOand international environmental treaties.

Some lawyers point out that CITES,the Basel Convention (which prohibitsthe export of hazardous waste) and theMontreal Protocol (which limits chloro-fluorocarbon sales) could be interpretedas illegal under the WTO: each presentsbarriers to trade. And the legality ofconsumer choice—particularly as ex-pressed through eco-labeling—remainsjust as unresolved. Even given its idio-syncratic intricacies, the WTO shrimpruling suggests that trade and the envi-ronment are still far from being happilyunited. —Marguerite Holloway


After nearly a century of union-management warfare in the

U.S., a series of nationwidesurveys showing that union shops dom-inate the ranks of the country’s mostproductive workplaces may come as asurprise. In fact, according to Lisa M.Lynch of Tufts University and SandraE. Black of the Federal Reserve Bank ofNew York, economic Darwinism—thesurvival of the fittest championed bygenerations of hard-nosed tycoons—maybe doing what legions of organizerscould not: putting an end to autocraticbosses and regimented workplaces.

American industry has been trying toreinvent itself for more than 20 years.Management gurus have proclaimedTheories X, Y and Z, not to mentionQuality Circles, Total Quality Manage-ment (TQM) and High-Output Man-agement. Only in the past few years,

however, have any solid data becomeavailable on which techniques workand which don’t. Businesses do not al-ways respond to surveys, and previousattempts to collect data ran into re-sponse rates of as low as 6 percent, mak-ing their results unrepresentative. Enterthe U.S. Census’s Educational Qualityof the Workforce National EmployerSurvey, first conducted in 1994, whichcollected data on business practices froma nationally representative sample ofmore than 1,500 workplaces.

Lynch and Black correlated the sur-vey data with other statistics that de-tailed the productivity of each businessin the sample. They took as their “typi-cal” establishment a nonunion compa-ny with limited profit sharing and with-out TQM or other formal quality-en-hancing methods. (Unionized firmsconstituted about 20 percent of the sam-ple, consistent with the waning reach oforganized labor in the U.S.)

The average unionized establishmentrecorded productivity levels 16 percenthigher than the baseline firm, whereasaverage nonunion ones scored 11 per-cent lower. One reason: most of theunion shops had adopted so-called for-mal quality programs, in which up tohalf the workers meet regularly to dis-

cuss workplace issues. Moreover, pro-duction workers at these establishmentsshared in the firms’ profits, and morethan a quarter did their jobs in self-man-aged teams. Productivity in such unionshops was 20 percent above baseline.That small minority of unionized work-places still following the adversarial linerecorded productivity 15 percent lowerthan the baseline—even worse than thenonunion average.

Are these productivity gains the re-sult of high-performance managementtechniques rather than unionization?No, Lynch and Black say. Adoption ofthe same methods in nonunion estab-lishments yielded only a 10 percent im-provement in productivity over thebaseline. The doubled gains in well-rununion shops, Lynch contends, may re-sult from the greater stake unionizedworkers have in their place of employ-ment: they can accept or even proposelarge changes in job practice withoutworrying that they are cutting their ownthroats in doing so. (Lynch tells the op-posing story of a high-tech companythat paid its janitors a small bonus forsuggesting a simple measure to speednightly office cleaning—and then laidoff a third of them.)

Even if a union cannot guarantee jobsecurity, she says, it enables workers tonegotiate on a more or less equal foot-ing. Especially in manufacturing, Lynchnotes, unionized workplaces tend tohave lower turnover. Consequently, theyalso reap more benefit from company-specific on-the-job training.

These documented productivity gainscast a different light on the declining per-centage of unionized workers through-out the U.S. Are employers actingagainst their own interests when they


more than 35 percent of the software inwestern Europe that needs millenniumrepairs will not be ready in time, lead-ing to computer failures.

Political concerns, however, have sofar outweighed the technical ones. “It’snot politically feasible to push back thedeadline,” states Pieter Dekker of theEuropean Commission. “It would belike trying to postpone the presidential

elections in the U.S.,” he concludes.The danger, of course, is that untreat-

ed computers will crash, forcing peopleto enter orders and to write invoices byhand. “Unfortunately,” says the GartnerGroup’s Nick Jones, “the politiciansneither understand nor care about thishuge and horrible burden they are plac-ing on Europe’s information technologyindustry.” —Alden M. Hayashi

LOOK FOR THE

UNION LABEL

New analysis of economic data shows that unionization could

maximize productivity

ECONOMICS

HIGH-PRODUCTIVITY WORKERS,UNIONIZED COMPANIES

HIGH-PRODUCTIVITY WORKERS,NONUNIONIZED COMPANIES

PERCENT CHANGE*

CHANGES IN LABOR PRODUCTIVITY

* Compared with a nonunionized company with no quality management program.

SOURCE: Lisa M. Lynch and Sandra E. Black

0 5 10 15 20–5–10–15–20

NO EMPLOYEE INVOLVEMENT,UNIONIZED COMPANIES

JAC

K KU

RTZ

Impa

ct V

isua

ls

JOH

NN

Y JO

HN

SON

UNIONIZED EMPLOYEES are the most efficient workers if they are part of a

“high-productivity” system, which includes qualitymanagement programs, profit sharing and regular

discussions of workplace issues.


With all the advances to-ward equality for wom-en, you’ve got to assume

there are lots of women programmers,right? After all, it’s not as though com-puter science is a discipline that requiresmore brawn than brains. But no: ac-cording to statistics compiled by TracyK. Camp, then an assistant professor ofcomputer science at the University ofAlabama, the number of undergraduatedegrees in computer science that isawarded to women has been shrinkingsteadily, both in real numbers and as apercentage of degrees awarded.

For instance, in the academic year1980–81, women obtained 32.5 percentof the bachelor’s degrees in computerscience; the figure for 1993–94 was 28.4percent, a drop of 12.6 percent. Calcu-lated from the peak year in 1983–84,when 37.1 percent of the degrees wentto women (representing 32,172 B.A.and B.S. degrees), the total decline is analarming 23.5 percent.

This decline, Camp notes, is true eventhough other science and engineeringfields boosted their recruitment of wom-en. From 1980–81 to 1993–94, the per-centage of physics degrees earned bywomen rose 36.6 percent (from 24.6 to33.6 percent of recipients) and those inengineering by 44.7 percent (from 10.3to 14.9 percent).

As in many other disciplines, there hasbeen a “shrinking-pipeline” effect forwomen in computer science, who makeup half of the high school computer sci-ence classes but who in 1993–94 con-stituted only 5.7 percent of full profes-sorships in Ph.D.-granting departments.The percentage of M.S. degrees (logical-ly) peaked in 1985–86, two years afterthe B.A. and B.S. degrees peaked; wom-en earned 29.9 percent of the master’sdegrees awarded that year. Ph.D. degreespeaked at 15.4 percent in 1988–89 andagain in 1993–94. Those figures decline

further through the ranks of facultymembers.

But Camp’s research highlights a dif-ferent problem: women are progressive-ly staying away from computer scienceas a profession. Even though the abso-lute numbers of M.S. and Ph.D. degreesawarded in computer science have con-tinued to increase, these numbers arelikely to drop soon, reflecting the fewerundergraduate degrees awarded.

The pattern is not unique to the U.S.In Britain, says Rachel Burnett, an in-formation technology (IT) lawyer andpartner with Masons Solicitors in Lon-don, the decline of women in computerscience is even more marked. “The in-take of women into university IT cours-es has declined,” she says. “It used to bea third—it’s now about 5 percent.”

The question is why. To find out,Camp, who works on the Committeeon Women in Computing for the Asso-ciation for Computing Machinery

(ACM), conducted a survey of ACMmembers. This is not, as she herself saidwhen she presented the survey results atthe ACM Policy ’98 conference inWashington, D.C., this past May, exact-ly the right sample. To be realistic, youneed to ask the people who didn’t jointhe profession.

Still, the ACM respondents providedsimilar answers that plausibly explainthe drop in women. They agreed that in

the 1980s, computer games,which tend to be male-oriented,gave boys more computer experi-ence. The nonstop long hours typ-ical of programming jobs, thegender discrimination, the lack ofrole models and the antisocial im-age of the hacker have also tend-ed to steer women to other areas.One issue that was not on the sur-vey but that was brought up fre-quently by many conference par-ticipants was sexual harassment,which is perceived to be higher inthe computer industry than inother fields.

Camp argues that this shrink-ing pipeline does matter. There iscurrently such a severe shortageof computer scientists that the in-dustry recently lobbied Congressto grant more visas to would-beimmigrants. And the quality andvariety of software could onlybenefit from having a less homo-geneous group of people creatingit. Proposed solutions to raise thepercentage of women who enter

the profession include making morevisible the female role models that exist,improving mentoring and encouragingequal access to computers for girls fromkindergarten through the 12th grade.Perhaps then the existing pool of talentwon’t inadvertently seep away.

—Wendy M. Grossmanin Washington, D.C.

WENDY M. GROSSMAN, a free-lance writer based in London, is the au-thor of net.wars.


work to block unionization? Lynch be-lieves that a follow-up survey, with ini-tial analyses due out this winter, mayhelp answer that question and others.Economists will be able to see howmany of the previously sampled firmsthat have traditional management-la-

bor relations managed to stay in busi-ness and to what extent the “corporatereengineering” mania of the past fewyears has paid off. Most serious reengi-neering efforts—the ones that aren’t justdownsizing by another name—lead toincreased worker involvement, Lynch

argues, if only because they requirefinding out how people actually dotheir jobs. Armed with that knowl-edge—and with the willing cooperationof their employees—firms may yet beable to break out of the productivitydoldrums. —Paul Wallich

CYBER VIEWAccess Denied

DIC

K B

LUM

E Im

age

Wor

ks

WOMEN MAKE UP HALF THE STUDENTS in high school computer science classes

but often avoid the field later in life.


Fusion and the Z-Pinch40 Scientific American August 1998

Some things never change—or dothey? In 1978 fusion research hadbeen under way almost 30 years,

and ignition had been achieved only inthe hydrogen bomb. Nevertheless, I de-clared in Scientific American at the timethat a proof of principle of laboratoryfusion was less than 10 years away andthat, with this accomplished, we couldmove on to fusion power plants [see “Fu-sion Power with Particle Beams,” Sci-entific American, November 1978].Our motivation, then as now, was the

knowledge that a thimbleful of liquidheavy-hydrogen fuel could produce asmuch energy as 20 tons of coal.

Today researchers have been pursu-ing the Holy Grail of fusion for almost50 years. Ignition, they say, is still “10years away.” The 1970s energy crisis islong forgotten, and the patience of oursupporters is strained, to say the least.Less than three years ago I thought aboutpulling the plug on work at Sandia Na-tional Laboratories that was still a factorof 50 away from the power required to

FIRING OF Z MACHINE producesa spectacular display. Z’s primarytransmission lines are submerged inwater for insulation. A very smallpercentage of the huge energy usedin firing the device escapes to thesurface of the water in the form oflarge electrical discharges. As in alightning strike, the high voltagesbreak down the air-water interface,creating the visual discharges. Theevent lasts only microseconds. Anautomatic camera, with its shutteropen, records the rapid discharges asa filigree of brilliant tracks.

RA

ND

Y M

ON

TOYA

San

dia

Nat

ion

al L

abor

ator

ies


light the fusion fire. Since then, however,our success in generating powerful x-raypulses using a new kind of device calledthe Z machine has restored my beliefthat triggering fusion in the laboratorymay indeed be feasible in 10 years.

The hydrogen bomb provides theproof that fusion can be made to hap-pen. In an H-bomb, radiation from anatomic fission explosion acts as a trig-ger, heating and compressing a fuel con-tainer to ignite and burn the hydrogeninside. That sounds simple, but causing

a fusion reaction to ignite and burnmeans forcing together the nuclei oftwo forms of hydrogen, deuterium andtritium so that they fuse to form heliumnuclei, giving off enormous energy. Thecompression must be done with almostperfect symmetry so that the hydrogenis squeezed uniformly to high density.

In the early decades of fusion research,the prospect of making a laboratory ver-sion of the H-bomb seemed remote. In-stead efforts to control fusion relied onthe principle of magnetic confinement,

in which a powerful magnetic field trapsa hot deuterium-tritium plasma longenough for fusion to begin. In 1991 deu-terium-tritium fusion was achieved inthis way by the Joint European Torusand later by Princeton University’s Tok-amak Fusion Test Reactor (TFTR). Thenext step on this road is the Interna-tional Thermonuclear Experimental Re-actor (ITER), a project involving theU.S., Europe, Japan and Russia. ITER’santicipated expense and technical diffi-culty, however, as well as disagreement

Fusion and the Z-Pinch Scientific American August 1998 41

Fusion and the

ZPinch

A device calledthe Z machine

has led to a new way of triggering

controlled fusion with intensenanosecond

bursts of x-rays

by Gerold Yonas


over where it should be built, have al-ready slowed progress on the engineer-ing design phase.

As far back as the early 1970s, re-searchers at Los Alamos, Lawrence Liv-ermore and Sandia national laboratoriesturned their attention to another way ofachieving fusion. In the inertial confine-ment approach, typified by the H-bomb,the idea is to use radiation to compress apellet of hydrogen fuel. Whereas the H-bomb relies on radiation from an atomicbomb, the first attempts at laboratory in-ertial fusion made use of intense laser orelectron beams to implode a fuel pellet.

The power then thought necessary toachieve ignition was much smaller thanwe now know it to be. By 1978 (as dis-cussed in my earlier Scientific Americanarticle), the estimated requirement hadrisen to one million joules delivered in10 nanoseconds to the outside of a pep-percorn-size fuel pellet—a power de-mand of 100 terawatts and the equiva-lent of condensing several hours’ worthof electricity use by half a dozen homesinto a fraction of a second. To meet thisneed, we at Sandia and scientists in theSoviet Union began research with anovel technology called pulsed power.

In a pulsed power system, electricalenergy is stored in capacitors and thendischarged as brief pulses, which aremade briefer still to increase the power

in each pulse and compressed in spaceto increase the power density. Thesebursts of electromagnetic energy arethen converted into intense pulses ofcharged particles or used to drive otherdevices. Laser fusion systems, in con-trast, start with much longer electricalpulses, which are amplified and shapedwithin the lasing system itself. Pulsedpower was seen as an attractive alterna-tive to lasers because of its proved effi-ciency and low cost.

The technology began in 1964 at theU.K. Atomic Energy Authority and de-veloped through the mid-1960s in theSoviet Union, the U.K. and the U.S.with support from the Energy and De-fense departments. But the technique hadlimited power output, making it a darkhorse in the race to fusion. Instead itspreferred purpose was to simulate, inthe laboratory, the effects of radiationon weapon components.

In 1973 funded programs in inertialconfinement fusion began in the U.S. atSandia under my direction and at othernational laboratories and in the U.S.S.R.at the Kurchatov Institute under LeonidRudakov. Since then, we have learned atremendous amount about the technol-ogy for creating the power to reach ig-nition and the ignition requirementsthemselves, whether with lasers or withpulsed power. Decades of carefully di-

agnosed experiments using powerfullasers have improved and validated thecomputer codes that design fuel pellets.Today these simulations indicate thatalmost 500 terawatts and two millionjoules of radiation at a temperature ofthree million degrees for four nanosec-onds are required to ignite the fuel.

Lasers can do this. After 13 years ofresearch using the 30-kilojoule Nova la-ser, Lawrence Livermore is now build-ing a much more powerful laser as theheart of the National Ignition Facility(NIF). If successful, the NIF will pro-duce at least as much energy from fu-sion as the laser delivers to the pellet,but that will still not come close to pro-ducing the several 100-fold greater en-ergy required to power the laser itself.That goal requires high yield—that is,fusion energy output much greater thanthe energy put into the laser. The NIFwill take the next step toward highyield, but present laser technology istoo expensive to go further.

Reviving the Z-Pinch

Just a few years ago, despite 25 yearsof effort, we at Sandia were still farfrom achieving fusion through

pulsed power technology. The decisionto continue the program was not easy,but our perseverance has recently beenrewarded. For one, power output hasgrown enormously: pulsed power wasproducing one thousandth of a tera-watt of radiation in the mid-1960s; re-cently we have reached 290 terawattsin experiments on the Z machine. Weare confident that we can achieve high-yield fusion with radiation pulses of1,000 terawatts, and we have comewithin a factor of three of that goal.


INERTIAL CONFINEMENT FUSION uses theprinciples of the hydrogen bomb (left), in which radi-ation from a fission bomb (called the primary) com-presses and heats the fusion fuel, which is containedin the secondary. The minuscule laboratory equiva-lent (right) aims to bathe the peppercorn-size fuel pel-let symmetrically in radiation and to concentrate thepower into the pellet so that it implodes uniformly.

MIC

HA

EL G

OO

DM

AN

X-RAYS are generated when a plasmafrom many fine tungsten wires collapsesonto a carbon-deuterium straw placed onthe axis of a Z-pinch. End-on views runfrom left to right at intervals of threenanoseconds. The top series shows x-raysthat have energy greater than 800 elec-tron volts; the bottom series shows x-raysaround 200 electron volts in energy. SA

ND

IA N

ATIO

NA

L LA

BO

RAT

ORI

ES

PRIMARY (FISSION DETONATOR)

PRIMARY HOHLRAUM

PROPOSED FUSION REACTION CHAMBERHYDROGEN BOMB

FOAM

PRIMARY HOHLRAUM

SECONDARY HOHLRAUM

SECONDARY (FUSION FUEL)

RADIATION

RADIATIONCASE

DEUTERIUM-TRITIUM PELLET

TUNGSTENWIRE ARRAY


Equally important has been progressin concentrating the intense energy ontoa tiny fuel pellet. In the 1970s we beganwith electron beams and, in the 1980s,switched to beams of ions, which shouldheat a target to higher temperatures. Butcharged particles are hard to steer andto focus tightly into beams. X-rays have,in principle, a much more promisingcharacteristic: they can uniformly fill thespace around a fuel container, just asheat in an oven envelops a turkey. Theprospect of initiating fusion by usingpulsed power systems to create intensebursts of x-rays within a small reactionchamber has now emerged from re-search on a concept called the Z-pinch,which dates back to the beginnings ofmagnetic confinement fusion researchin the 1950s.

Originally, the Z-pinch was an at-tempt to initiate fusion by passing astrong electric current through deuteri-um gas. The current both ionizes thegas and generates a magnetic field that“pinches” the resulting plasma to hightemperature and density along the cur-rent path, conventionally labeled the zaxis. But the technique proved unableto compress a plasma uniformly: fluidinstabilities break the plasma into blobs,making the Z-pinch inadequate to sup-port fusion. The compression of the plas-ma, however, also generates x-rays, withenergies up to 1,000 electron volts. For30 years, research on Z-pinches in theU.S., the U.K. and the U.S.S.R. focusedon optimizing the subkiloelectron-voltx-ray output, using those x-rays to testthe response of materials and electronicsto radiation from nuclear weapons.

The Z-pinch has now acquired newlife as a way of initiating inertial fusion.In the past three years we have shownthat by combining the efficiency andlow cost of fast pulsed power with thesimplicity and efficiency of the Z-pinchas a radiation source, we should reachignition by using subkiloelectron-voltx-rays to compress a fusion fuel pellet.Moreover, the affordability of a pulsedpower x-ray source should allow us togo beyond that, to efficient burn-up ofthe fuel and high yield.

To trigger fusion, the Z-pinch mustbe enclosed in a radiation chamber (orhohlraum, German for “cavity” or “hol-low”) that traps the x-rays. In one sys-tem we have explored, the Z-pinchwould be placed in a primary hohlraum,with the fuel contained in a smaller, sec-ondary hohlraum. In another method,the pellet would sit in low-density plas-

tic foam at the center of the implodingpinch inside the primary hohlraum. Thekey is that the x-rays generated as thepinch crashes onto itself, either onto thez axis or onto the foam, are containedby the hohlraum so that they uniformlybathe the fuel pellet, just as the casingof an H-bomb traps the radiation fromthe atomic trigger. Experiments over thepast three years show that both meth-ods should work, because we can nowmake a Z-pinch that remains uniformand intact long enough to do the job.

Finding the Secret

What is different now comparedwith the long, earlier period of

slow progress? Like Thomas Edison,who tried a thousand materials beforefinding the secret to the lightbulb, wediscovered that the instability of the Z-pinch can be greatly reduced by ex-tracting energy quickly, in the form of ashort burst of x-rays, before instabilitiesdestroy the geometry. In effect, the en-ergy of the pinch is removed before ittransforms into rapid and small-scalemotions of plasma.

The instability that afflicts the pinchis the same one that makes a layer ofvinegar poured carefully on top of lessdense salad oil drop irregularly to thebottom of the jar. The instability isbeneficial for salads when the jar isshaken, but it was a barrier to achiev-ing fusion. From computer simulationsby Darrell Peterson of Los Alamos andMelissa R. Douglas of Sandia, however,we knew that the more uniform the ini-tial plasma, the more uniform and reg-ular the pinch would be when it stag-nated on axis and produced x-rays.

Researchers had tried many methodsto make the plasma more uniform, suchas using thin metal shells or hollow

puffs of gas to conduct the electric cur-rent, but none met with great success.The breakthrough came at Sandia in1995, when Thomas W. L. Sanford, us-ing many fine aluminum wires, andthen Christopher Deeney and Rick B.Spielman, using up to 400 fine tungstenwires, achieved the needed uniformity.Wire-array Z-pinches were first devisedin the late 1970s at Physics Internation-al—a private company interested in gen-erating x-rays as a source for testing ra-diation effects in the laboratory—for en-hancing the energy output of one to fivekiloelectron-volt x-rays. But the low-cur-rent accelerators then available could notdeliver enough electric power to implodelarge numbers of many small wires.

After the 1995 experiments Barry M.Marder of Sandia suggested that thekey was to have a number of wires ar-ranged so that as they explode with thepassage of the current, they merge tocreate a nearly uniform, implodingcylindrical plasma shell. Subsequent ex-periments at Sandia have shown thatthe desired hot central core is producedafter the entire shell implodes onto afoam cylinder positioned on the z axis.Also, experiments at Cornell Universityindicate that each wire may not turncompletely to plasma early on, as Mar-der’s simulations suggest. Instead a coldcore of wire may remain, surroundedby plasma, allowing the current flow tocontinue for a time and increasing theefficiency of the pinch.

These breakthroughs began three yearsago at Sandia on the 10-million-ampereSaturn accelerator and, since October1996, have continued on the 20-million-ampere Z machine, which now produc-es the world’s most powerful and ener-getic x-ray pulses. In a typical experi-ment, we generate nearly two millionjoules of x-rays in a few nanoseconds,


POWER from wire-ar-ray Z-pinches increasedslowly from 1970 to1995. The rapid ad-vances of the past twoyears in particular arethe result of evolutionto complex multiple-wire experiments usinghigh-current accelera-tors and of improvedcomputer modeling ofplasmas. Red dots andscale show the increasingcomputer power avail-able for simulations.

MIC

HA

EL G

OO

DM

AN

0

50

100

150

200

250

300

Z–P

INC

H P

OW

ER (

TER

AW

AT

TS)

YEAR

CO

MP

UTE

R P

OW

ER (

TER

AFL

OP

S)

0

50

100

150

200

250

1970 1975 1980 1985 1990 1995 2000 2005

300


for a power of more than 200 terawatts.In a series of experiments beginning

in November 1997, we increased the x-ray power by 45 percent, to 290 tera-watts, by using a double-nested arrayof wires. The current vaporizes the out-er array, and the magnetic field pushesthe vaporized material inward. The fast-er-moving parts strike the inner arrayand are slowed, allowing the slowerparts to catch up and sweep materialinto the inner array. This geometry re-duces the instabilities in the implosion,and when the vaporized materials col-lide at the z axis they create a shorterpulse of x-rays than a single array can.The nested array has produced a radia-tion temperature of 1.8 million degrees.

In other experiments on Z, led byArthur Toor of Livermore, foam layerssurrounding a beryllium tube within asingle array provide a slower, more sym-metrical implosion of the Z-pinch plas-ma, also resulting in increased hohlraumtemperatures. It took 40 years to get 40terawatts of x-ray power from a Z-pinch. Now, in the past three years, wehave moved much closer to our finalgoal of 1,000 terawatts and 16 millionjoules of x-rays, which should producethe three-million-degree hohlraum tem-peratures required for high-yield fusion.We knew that pulsed power should bemore efficient and less costly than thelaser approach, and indeed our Z accel-

erator now produces a total x-ray energyoutput equal to 15 percent of its electri-cal energy input; for Lawrence Liver-more’s Nova laser the equivalent effi-ciency is 0.1 percent. Design improve-ments could increase that figure to 0.5percent at the NIF, but the inherent in-efficiency of the laser process preventssuch devices from achieving any higherefficiencies.

The Final Factor of Three

These intense x-ray pulses have manyapplications. The energies and pow-

ers attained with Z already allow labo-ratory measurements of material prop-erties and studies of radiation transportat densities and temperatures that couldpreviously be achieved only in under-ground nuclear explosions. These labo-ratory experiments, and the fusion yieldsthat a higher-current device would per-mit, are part of the Department of En-ergy’s stockpile stewardship program toguarantee the safety and reliability ofaging nuclear weapons if the U.S. mustuse them in the future.

There are even astrophysical applica-tions, because the x-ray sources pow-ered by the Z machine produce plasmassimilar to those in the outermost layersof a star. The light output from a type ofpulsating star, the Cepheid variable, isnow better understood because of data

obtained by Paul T. Springer of Law-rence Livermore from Z-pinch experi-ments on the Saturn accelerator in 1996.We expect other data to lessen the mys-tery of stellar events such as supernovae.Laboratory plasmas also offer the po-tential for new studies in atomic phys-ics and x-ray lasers.

As an x-ray source, the Z-pinch is re-markably efficient and reproducible; re-peated experiments yield the same mag-nitude of x-ray energy and power, eventhough we cannot predict in detail whathappens. What we can forecast is scale:every time we double the current, the x-ray energy increases fourfold, followinga simple square law. And as theoretical-ly expected for thermal radiation, thepinch temperature increases as the squareroot of the current. If this physics holdstrue, another factor-of-three increase incurrent—to 60 million amperes—shouldallow us to achieve the energy, powerand temperature needed to trigger fu-sion and reach high yield.

What questions must be resolved be-fore that next step? The first is whetherwe can cram a factor-of-three-highercurrent into the same container. The rea-son that the enormous concentration ofpower into small cavities in a Z-pinch ispossible at all was discovered almost 30years ago at the Kurchatov Institute, atPhysics International and at Sandia.Ordinarily, electric fields tend to disrupt


SEQUENTIAL CONCENTRATIONof power in a pulsed-power facility be-gins in a circular array of 36 Marxgenerators, in which 90,000 voltscharge a 5,000-cubic-meter capacitorbank in two minutes. Electrical pulsesfrom the 36 modules enter a water-in-sulated section of intermediate storagecapacitors, where they are compressedto a duration of 100 nanoseconds.Passing through a laser-triggered gasswitch that synchronizes the 36 pulsesto within one nanosecond, the com-bined pulse travels to the wire array(far left) along four magnetically insu-lated transmission lines that minimizeloss of energy. The Z-pinch (left) formsas thousands of amps of electric currenttravel through wires one tenth the di-ameter of an average human hair. Theillustrations on the opposite page dem-onstrate how x-rays from the pinchwould implode a fuel pellet and initiatefusion in a schematic hohlraum design.In the proposed X-1 accelerator, theculmination would be the implosion,in about 10 nanoseconds, of a pepper-corn-size fuel pellet to the size of theperiod at the end of this sentence.

PULSED-POWER GENERATOR

CURRENT

MAGNETIC FIELD

RADIATION

Z-PINCH

MARX GENERATOR

INTERMEDIATESTORAGE

CAPACITOR

MIC

HA

EL G

OO

DM

AN


the current that generates them, but aphenomenon called magnetic insulationallows short, powerful electrical pulsesto be transmitted along a channel be-tween two metal surfaces without break-ing down and with almost no energyloss. The magnetic field of the powerfulpulse acts to contain the pulse itself,overcoming the electric field that wouldotherwise cause breakdown. In Z ex-periments by John L. Porter of Sandiain April 1998, a gap of 1.5 millimetersbetween a wire array and the surround-ing stationary hohlraum wall remainedopen in the face of the intense radiationbecause of magnetic insulation, allow-ing the hohlraum temperature to reach1.7 million degrees.

Fifty terawatts are transmitted intothe tiny space between the wires and thehohlraum wall, and the power densityreaches 25 terawatts per square centi-meter. If we increase the power to 150terawatts at 60 million amps, the pow-er density would rise to 75 terawattsper square centimeter. That increaseraises new questions, because the mate-rial pressure in the metal wall rises to1.5 to three million atmospheres. Other

questions are whether the efficiency ofconversion to x-rays remains at the 15percent level seen at 20 million amps,whether instabilities stay under control,and whether we can achieve the sym-metry and shape of the radiation pulseonto the pellet that computer calcula-tions suggest are needed.

Another important step is developingpredictive models to scale the complexphysics. The two-dimensional simula-tions available today have provided agreat deal of insight into the physics ofpinches but, even though restricted totwo dimensions, require tremendouscomputer power. Simulation of the fullthree-dimensional magnetic, hydrody-namic and radiative character of thepinch is beyond our current capability,but advances in high-performance com-puting and diagnostics for x-ray imag-ing are rapidly catching up with our ad-vances in radiated power. In 1998 wehave been operating the Janus comput-er at 1.8 trillion floating-point (multi-plication or division) operations persecond (teraflops). Colleagues at Sandiaand Los Alamos are developing a com-puter model to simulate the pinch phys-

ics and the transport of radiation to thepellet. With the advent of these tools onthe new generation of supercomputers,we expect to continue our rapid prog-ress in developing Z-pinches for fusion.

We at Sandia now hope to begin de-signing the next big step. At the end ofMarch we requested approval from theDepartment of Energy to begin the con-ceptual design for the successor to Z,the X-1. This machine should give us 16megajoules of radiation and, we believe,allow us to achieve high yield. Quoting acost is premature, but we expect it willbe in the neighborhood of $400 million.It is important to remember that Z, X-1and the NIF are still research tools. Zmay achieve fusion conditions; the NIFshould achieve ignition; and X-1, build-ing on the lessons of the NIF, shouldachieve high yield—but none of theseexperiments is expected to provide acommercial source of electrical power.

As Yogi Berra pointed out, “It is toughto make predictions, especially aboutthe future,” but if we can get startedsoon on design and construction of thenext big step, we really think we can dothe job in 10 years.


The Author

GEROLD YONAS is vice president of systems, scienceand technology at Sandia National Laboratories. His ca-reer began in 1962 at the Jet Propulsion Laboratory inPasadena, Calif., while he was studying at the CaliforniaInstitute of Technology for his doctorate in engineering sci-ence and physics. Yonas joined Sandia in 1972. From 1984to 1986 he provided technical management to the StrategicDefense Initiative Organization, serving as its first chief sci-entist. He was president of Titan Technologies in La Jolla,Calif., from 1986 to 1989, when he rejoined Sandia. Hehas published extensively in the fields of intense particlebeams, inertial confinement fusion, strategic defense tech-nologies and technology transfer.

Further Reading

Neutron Production in Linear Deuterium Pinches. O. A. Anderson,W. R. Baker, S. A. Colgate, J. Ise and R. V. Pyle in Physical Review, Vol.110, No. 6, pages 1375–1387; 1958.

X-rays from Z-Pinches on Relativistic Electron Beam Generators.N. R. Pereira and J. Davis in Journal of Applied Physics, Vol. 64, No. 3,pages R1–R27; 1988.

Fast Liners for Inertial Fusion. V. P. Smirnov in Plasma Physics andControlled Fusion, Vol. 33, No. 13, pages 1697–1714; November 1991.

Dark Sun: The Making of the Hydrogen Bomb. Richard Rhodes. Si-mon & Schuster, 1995.

Z-Pinches as Intense X-Ray Sources for High-Energy DensityPhysics Applications. M. Keith Matzen in Physics of Plasmas, Vol. 4,No. 5, Part 2, pages 1519–1527; May 1997.

X-rays vaporize outer layer of fuelpellet, which bursts outward andimparts inward momentum tothe hydrogen fuel.

Pellet implodes to 1/1,000 to 1/10,000of its original volume, and fusionbegins as temperature reaches120 million degrees or more. Wavefront of burning hydrogen

expands outward and cools untilfusion ceases.

SA

MIC

HA

EL G

OO

DM

AN


JOH

N W

. KA

RA

PELO

U

LOWER BACK consists of numer-ous structures, any of which maybe responsible for pain. The mostobvious are the powerful musclesthat surround the spine. Other po-tential sources of pain include thestrong ligaments that connect ver-tebrae; the disks that lie betweenvertebrae, providing cushioning;the facet joints, which help to en-sure smooth alignment and stabili-ty of the spine; the vertebral bonesthemselves; blood vessels; and thenerves that emerge from the spine.


The catalogue of life’s certain-ties is usually limited to deathand taxes. A more realistic list

would include low-back pain. Up to 80percent of all adults will eventually ex-perience back pain, and it is a leadingreason for physician office visits, for hos-pitalization and surgery, and for workdisability. The annual combined cost ofback pain–related medical care and dis-ability compensation may reach $50billion in the U.S. Clearly, back pain isone of society’s most significant non-lethal medical conditions. And yet theprevalence of back pain is perhapsmatched in degree only by the lingeringmystery accompanying it.

Consider the following paradox. TheAmerican economy is increasingly post-industrial, with less heavy labor, moreautomation and more robotics, andmedicine has consistently improved di-agnostic imaging of the spine and de-veloped new forms of surgical and non-surgical therapy. But work disabilitycaused by back pain has steadily risen.Calling a physician a back-pain expert,therefore, is perhaps faint praise—med-icine has at best a limited understand-ing of the condition. In fact, medicine’sreliance on outdated ideas may have ac-tually contributed to the problem. Oldconcepts were supported only by weakevidence such as physiological inferencesand case reports, rather than by clinicalfindings from rigorous controlled trials.

The good news is that most back-painpatients will substantially and rapidlyrecover, even when their pain is severe.This prognosis holds true regardless oftreatment method or even without treat-ment. Only a minority of patients withback pain will miss work because of it.Most patients who do leave work re-

turn within six weeks, and only a smallpercentage never return to their jobs.(At a given time, about 1 percent of thework force is chronically disabled be-cause of back problems.) Overall, then,prospects for patients with acute backpain are quite good. The bad news isthat recurrences are common; a majori-ty of patients will experience them. For-tunately, these recurrences tend to playout much as the original incidents did,and most patients recover again quicklyand spontaneously.

Sources of Pain

Low-back pain is a symptom that may signal various conditions affecting

structures in the low back. Part of themystery of back pain comes from thediagnostic challenge of determining itscause in a mechanical and biochemicalsystem of multiple parts, all of which aresubject to insult. Injuries to the musclesand ligaments may contribute, as mayarthritis in the facet joints or disks. Aherniated (or “slipped”) disk, in whichthe soft inner cushioning material pro-trudes through the disk’s outer rim andirritates an adjacent nerve root, can bethe source of pain. Or the culprit mightbe spinal stenosis, a narrowing of thespinal canal that can cause a pinchednerve; stenosis usually accompanies ag-ing and wear of the disks, the facet jointsand the ligaments in the spinal canal.

Back pain also may be the result ofcongenital abnormalities of the spine.These odd structures are often asymp-tomatic but may cause trouble if severeenough. Diseases of other parts of theanatomy, such as the kidneys, pancreas,aorta or sex organs, can be responsibleas well. Finally, back pain may be a


Low-Back PainLow-back pain is at epidemic levels. Althoughits causes are still poorly understood, treatment

choices have improved, with the body’s ownhealing power often the most reliable remedy

by Richard A. Deyo


symptom of serious underlying diseasessuch as cancer, bone infections or rareforms of arthritis. Fortunately, suchcritical causes are extremely rare; about98 percent of back-pain patients sufferfrom injury, usually temporary, to themuscles, ligaments, bones or disks.

The physical complexity of the lowerback combines with another vexing re-ality to hinder diagnosis of the cause ofpain: only a weak association exists be-tween symptoms, imaging results, andanatomic or physiological changes. Un-der these circumstances, most diagnos-tic evaluations focus on excluding ex-treme causes of pain—such as cancer orinfection—that can be more preciselyidentified or on determining whether apatient has a pinched or irritated nerve.Up to 85 percent of patients with low-back pain are then left without a defini-tive diagnosis, a nuts-and-bolts reasonfor their pain. Most patients cannot re-call a specific incident that brought ontheir suffering, and heavy lifting or in-juries, though risk factors, do not ac-count for most episodes. Back pain oftenjust seems to happen, and the medicalcommunity, reflecting this vagueness,has by no means reached a consensusas to the causes of garden-variety cases.

Some commonplace back pain isprobably related to stress. A study pub-lished in May by Astrid Lampe and hercolleagues at the University of Innsbruckrevealed a connection between stressfullife events and occurrences of back pain.Previous work by Lampe found that pa-tients without a definite physical reasonfor low-back pain perceived life as morestressful than a control group of back-pain patients who had definite physical

damage. John E. Sarno of the Rusk In-stitute of Rehabilitation Medicine atNew York University Medical Centerhas concluded that unresolved emo-tionally charged states produce physicaltension that in turn causes pain. In fact,he asserts that this variety of back painactually serves to distract patients fromthe potential distress of confronting theirpsychological conflicts; Sarno has suc-cessfully treated selected patients withpsychological counseling.

Simple muscle soreness from physicalactivity very likely causes some backpain, as does the natural wear and tearon disks and ligaments that creates mi-crotraumas to those structures, especial-ly with age. Determining the cause of agiven individual’s pain, however, oftenremains more art than science. Withspontaneous recovery the rule—once se-rious disease is eliminated as a factor—

pinpointing an exact cause may not evenbe necessary in most cases.

Diagnostic Challenges

The inadequacy of definitively diag-nosing the cause of back pain led

my colleague Daniel C. Cherkin ofGroup Health Cooperative of PugetSound and my research group at theUniversity of Washington to conduct anational survey of physicians from dif-ferent specialties. We offered standard-ized patient descriptions and asked oursubjects how they would manage thesehypothetical patients. Reflecting the un-certainty in the state of the art, recom-mendations varied enormously. The re-sults can be summed up by the subtitleof our publication of the survey results:

“Who You See Is What YouGet.” For example, rheuma-tologists were twice as likelyas physicians of other special-ties to order laboratory testsin a search for arthritic condi-tions. Neurosurgeons weretwice as likely to ask for imag-ing tests that would uncoverherniated disks. And neurolo-gists were three times moreinclined to seek the results ofelectromyograms that wouldimplicate the nerves. If pa-tients are confused, they arenot alone.

Until recently, doctors re-lied on spine x-rays, oftenperforming one on every pa-tient with low-back pain.Various studies have revealed

multiple problems with this approach.First, a 10-year Swedish research effortdemonstrated that at least for adultsunder age 50, x-rays added little of di-agnostic value to office examinations,with unexpected findings in only aboutone of every 2,500 patients x-rayed.

Second, epidemiological research re-vealed that many conditions of the spinethat often received blame for pain wereactually unrelated to symptoms. Largenumbers of pain-free people have beenx-rayed in preemployment medical ex-ams and for military induction in somecountries, and multiple studies deter-mined that many spine abnormalitieswere as common in asymptomatic peo-ple as in those with pain. X-rays cantherefore be quite misleading.

Third, low-back x-rays unavoidablyinvolve exposing sex organs to largedoses of ionizing radiation, more than1,000 times greater than that associatedwith a chest x-ray. Last, even highly ex-perienced radiologists interpret the samex-rays differently, leading to uncertaintyand even inappropriate treatment. Thelatest clinical guidelines for evaluatingback pain thus recommend that x-raysbe limited to specific patients, such asthose who have suffered major injuriesin a fall or automobile accident.

Medical experts hoped that improveddiagnostic imaging instrumentation,such as computed tomographic (CT)scanning and magnetic resonance imag-ing (MRI), would make possible moreprecise diagnoses for most back-painpatients. This promise has been illusory.One important reason is that, as in thex-ray studies, alarming abnormalitiesare found in pain-free people.

A 1990 study by Scott D. Boden of theGeorge Washington University MedicalCenter and his colleagues looked at 67individuals who said they had never hadany back pain or sciatica (leg pain fromlow-back conditions). Herniated disksoften get cited as the reason for a pa-tient’s pain, but MRI found them in onefifth of pain-free study subjects underage 60. Half that group had a bulgingdisk, a less severe condition also oftenblamed for pain. Of adults older than60, more than a third have a herniateddisk visible with MRI, nearly 80 percenthave a bulging disk and nearly every-one shows some age-related disk degen-eration. Spinal stenosis, rare in youngeradults, occurred in about one fifth ofthe over-60, pain-free group. A similarstudy of 98 pain-free people, publishedin 1994 by Michael N. Brant-Zawad-

Low-Back Pain50 Scientific American August 1998

CH

EL D

ON

G

Primary Reasons for AdultVisits to Office-Based Physicians

RANK

SOURCE: National Ambulatory Medical Care Survey, 1980–90

HYPERTENSION (HIGH BLOOD PRESSURE)

PREGNANCY CARE

CHECKUPS, WELL CARE

UPPER RESPIRATORY INFECTIONS (COLDS)

LOW-BACK PAIN

DEPRESSION AND ANXIETY

DIABETES

1

2

3

4

5

6

7


zki of Hoag Memorial Hospital in New-port Beach, Calif., and his colleagues,revealed that about two thirds had ab-normal disks. Detecting a herniated diskon an imaging test therefore provesonly one thing conclusively: the patienthas a herniated disk.

These findings suggest that many redherrings confuse imaging interpretationand that at least for some, spine abnor-malities are purely coincidental and donot cause pain. Moreover, even the bestimaging tests fail to identify the simplemuscle spasm or injured ligament prob-ably responsible for pain in a substan-tial percentage of back patients. All thisimaging perplexity caused one orthope-dic surgeon to remark, “A diagnosisbased on MRI in the absence of objec-tive clinical findings may not be thecause of a patient’s pain, and an at-tempt at operative correction could bethe first step toward disaster.” In otherwords, the office examination is at leastas important as the imaging test, andsurgery for patients whose back pain isassociated only with abnormal imagingresults can be unnecessary if not down-right detrimental. Many physiciansnow advocate CT scans and MRI onlyfor those patients who are already sur-gical candidates for other reasons.

Complicating the situation still fur-ther is the fact that most patients withacute low-back pain simply get better—

and quickly. A study comparing treat-ment outcomes found no differences infunctional recovery times among pa-tients who saw chiropractors, familydoctors or orthopedic surgeons. Cost,on the other hand, varied substantially,with family doctors costing least andsurgeons most. The Hippocratic admo-nition “First, do no harm” may be themost important counsel with regard tothis condition—the favorable naturalhistory of acute low-back pain is hardto beat.

Extended bed rest was once regardedas the standard therapy. This approachwas based on the rationale that somepatients experience at least transient re-lief when lying down, as well as on thephysiological observation that pressuresin the intervertebral disks are lowestwhen patients are prone. But a guilty-looking disk may be innocent, and mostpatients improve naturally. Neverthe-less, recommendations of one to twoweeks of strict bed rest were the normuntil about 10 years ago. Bed rest’s fallfrom favor has been almost as dramaticas the reversal in status suffered by thatformer favorite of primary care, blood-letting. Extended bed rest is now consid-ered anathema, and resuming normalactivities as much as possible may bethe best option for patients with acuteback pain.

Watchful Waiting as Treatment

When bed rest was still the stan-dard, my group tested it by com-

paring seven days of bed rest with justtwo days. The results were striking. Af-ter three weeks and three months, therewere no differences in pain relief, indays of limited activity, in daily func-tioning or in satisfaction with care. Theonly difference was that, obviously, pa-tients with longer bed rest missed morework. Severity of a patient’s pain, dura-tion of pain, and abnormalities found inthe office examination offered no pre-dictive value for how long the patientwould be off the job. In fact, data anal-ysis showed that the only factor thatpredicted the duration of the patient’sabsence from work was our recommen-dation for how long to stay in bed.

Other studies have confirmed and ex-tended these findings. Four days of bedrest turns out to be no more effectivethan two days—or even no bed rest atall. The fear that activity would exacer-bate the situation and delay recoveryproved to be unfounded. In fact, pa-

Low-Back Pain Scientific American August 1998 51

SCOTLAND

ENGLAND (1985)

MANITOBA, CANADA

SWEDEN

NEW ZEALAND

SOUTH AUSTRALIA

ONTARIO, CANADA

NORWAY

FINLAND

DENMARK

NETHERLANDS

UNITED STATES

0 0.2 0.4 0.6 0.8 1

SURGERY RATES for back pain vary widely from country to country, with the U.S.having more than five times as many back operations performed as England. Varyingcultural attitudes play a subtle role in the experience and the treatment of back pain, ofwhich surgery is but an obvious example. In Oman, for example, back pain was com-mon, but disability from back pain was rare until the introduction of Western medi-cine, according to a 1986 report to the minister of health.

ABNORMAL MRI SCAN from pain-freesubject illustrates one of the great pitfallsin diagnosing a cause for low-back pain.Numerous studies have shown that largenumbers of asymptomatic people alsohave disk bulges or herniations. This sub-ject has two herniated disks (arrows) andadditional disk abnormalities.

CH

EL D

ON

G; S

OU

RC

E: S

pine

Ratio of Back-Surgery Rates in Selected Places, 1988–89C

OU

RTE

SY O

F RI

CH

ARD

A. D

EYO

; SO

URC

E: J

EFFR

EY J

AR

VIK


tients with acute back pain who contin-ue routine activities as normally as pos-sible do better than those who try eitherbed rest or immediate exercise. Studieshave shown that people who remainedactive despite acute pain experiencedless future chronic pain (defined as painlasting three months or more) and usedfewer health care services than patientswho rested and waited for the pain todiminish. (The fact that bed rest is inef-fective does not mean that everyone canreturn to their normal jobs immediate-ly, however. Some people with physical-ly demanding jobs may be unable to goback to their normal work as quickly aspeople with more sedentary occupa-tions. Nevertheless, it is often useful tohave patients with back pain return tosome form of light work until they haverecovered more fully.)

Recent research has also challenged

the effectiveness of other types of passivetreatment. For example, several studiesconcluded that traction for the low backsimply does not work. More controver-sially, there is growing evidence thattranscutaneous electrical nerve stimula-tion (TENS), which delivers mild elec-tric current to the painful area, has littleif any long-term benefit. Similarly, injec-tions of the facet joints with cortisone-like drugs appear to be no more effec-tive than injections with saline solution.

In contrast, there is growing evidencefor exercise as an important part of theprevention and treatment of back prob-lems for those suffering from eitherchronic or acute back pain. No singleexercise is best, and effective programscombine aerobics for general fitness withspecific training to improve the strengthand endurance of the back muscles.

An exhaustive review of clinical stud-ies of exercise and back pain found thatstructured exercise programs preventedrecurrences and reduced work absencesin patients with acute pain who regu-larly took part soon after an episode ofback pain had subsided. The preventivepower of exercise was stronger than the

effect of education (for example, howto lift) or of abdominal belts that limitspine motion. Patients experiencingchronic pain also benefited from exer-cise. In contrast to acute back-pain suf-ferers, who did better during a pain epi-sode by resuming normal activities thanthrough exercise, chronic back-pain pa-tients substantially improved by exer-cising even with their pain.

The inability of conventional medicalpractice to “cure” a large percentage ofback-pain patients has no doubt led thecondition to be a major reason patientsseek various forms of alternative treat-ment, including chiropractic care andacupuncture. Chiropractic is the mostcommon choice, and evidence accumu-lates that spinal manipulation may in-deed be an effective short-term pain rem-edy for patients with recent back prob-lems. Whether chiropractic or otheralternative treatments can impart long-term pain relief remains unclear. Thenormal recovery from back pain mostlikely leads to a belief in whatever treat-ment is employed and probably ac-counts for the large number of therapeu-tic options with passionate advocates.

At the other end of the strategic spec-trum is surgery. Most specialists agreethat disk surgery is appropriate onlywhen there is a combination of a defi-nite disk hernia on an imaging test, acorresponding pain syndrome, signs ofnerve root irritation and failure to re-spond to six weeks of nonsurgical treat-ment. For patients with these findings,surgery can offer faster pain relief. Un-fortunately, patients who do not meet

52 Scientific American August 1998 Low-Back Pain

MYTH 5:

Back pain is usually disabling.

ILLU

STR

ATIO

NS

BY

DU

SAN

PET

RIC

IC

MYTH 2:

X-ray and newer imaging tests (CT and MRI scans) can always identify the cause of pain.

Myths about Low-Back PainMYTH 6:

Everyone with back painshould have a spine x-ray.

MYTH 3:

If your back hurts,you should take it easyuntil the pain goes away.

MYTH 1:

If you have a slipped disk (also known as a herniatedor ruptured disk) you must have surgery. Surgeons agree about exactly who should have surgery.

MYTH 4:

Most back pain is caused by injuries or heavy lifting.

MYTH 7:

Bed rest is the mainstay of therapy.


all these standards also often go underthe knife, and there is extensive litera-ture on failed low-back surgery. Indeed,if the pain is not actually from disk her-niation, surgical repair of a disk cannotbe expected to end it.

Surgical Interventions

The scapegoating of the herniateddisk deserves further reflection. Her-

niated disks are most common in adultsbetween ages 30 and 50, and most pa-tients whose pain is actually caused bya disk herniation have leg pain withnumbness and tingling as the primarysymptom; their back pain is often lesssevere. A positive MRI should only sup-port a physical examination that inves-tigates a constellation of effects—suchas nerve root irritation, reflex abnor-malities and limited sensation, musclestrength and leg mobility—to implicatethe disk definitively as the factor in pain.

Recent studies show that even for pa-tients with a herniated disk, spontane-ous recovery is the rule. Studies usingrepeated MRI revealed that the herniat-ed part of the disk often shrinks natu-rally over time, and about 90 percent ofpatients will experience gradual improve-ment over a period of six weeks. Thus,only about 10 percent of patients witha symptomatic disk herniation wouldappear to require surgery. And becausemost back pain is not caused by herni-ated disks, the actual proportion of allback-pain patients who are surgical can-didates is only about 2 percent.

Herniated disks nonetheless remainthe most common reason for back sur-gery. A long-term follow-up study of280 patients, performed by Henrik We-ber of Ullevaal Hospital in Oslo andpublished in 1983, raises serious ques-

tions about the enthusiasm for surgicalintervention. Although patients whohad surgery had faster pain relief thandid patients treated conservatively, thedifferences evaporated over time. At thefour- and 10-year follow-ups, the twogroups of patients were virtually indis-tinguishable. Thus, reasonable peoplemight have preferences for different med-ical interventions, and there is growingrecognition that these preferences shouldbe an important consideration in treat-ment decisions.

Spinal stenosis is the most commonreason for back surgery in those overage 65. National hospital survey datashow stenosis correction to be the mostrapidly increasing form of back surgery.Surgery for herniated disks increased39 percent between 1979 and 1990;stenosis surgeries increased 343 percent.Reasons for this rapid rise are unclearbut may simply reflect the ability of thenew CT and MRI scans to reveal steno-sis. Unfortunately, the indications forsurgery in this condition are even lessclear-cut than they are for herniateddisks. As a result, there are enormousvariations, even within the U.S., in ratesof surgery for spinal stenosis. For ex-ample, by analyzing Medicare claims,my group found approximately 30stenosis surgeries in Rhode Island forevery 100,000 people older than 65 but132 in Utah.

Surgery for this condition is morecomplex than simple disk surgery. Spi-nal stenosis tends to occur at multiplelevels within the spine rather than at asingle level, as is usually true for herni-ated disks. Furthermore, these patientsare older and therefore more suscepti-ble to complications of surgery. In addi-tion, we know less about the long-termeffectiveness of surgical and nonsurgi-

cal approaches for treating spinal steno-sis than we do about management ofherniated disks. Because symptoms ofspinal stenosis often remain stable foryears at a time, decisions are rarely ur-gent, and the preferences of the patientshould again play an important role.

Classifying as trivial a condition thatannually drives millions of Americansto their knees and drains $50 billionfrom the economy would be a mistake.A collective shrug at the condition, how-ever, may be the most appropriate, albeitunsatisfying, societal attitude. Nearlyeveryone will have back pain, and weshould perhaps simply accept it as partof normal life. Once serious conditionsget ruled out, a sufferer is usually bestserved by simply attempting to cope aswell as possible with a condition thatwill almost certainly improve in days ora few weeks. The wide variability insurgical recommendations should makeall back-pain experts circumspect, andthe patient’s wishes should carry con-siderable weight in treatment choice.

The mysterious nature and economiccost of back pain are driving a growinginterest in research, and the comingyears may reveal the fundamental as-pects of this problem in more detail. Inthe meantime, for most back-pain pa-tients the stereotypical physician advi-sory to “take two aspirin and call me inthe morning” comes to mind. A richerand better course of action might be totake pain relievers as needed, stay ingood overall physical condition, keepactive through an acute attack if at allpossible and monitor the condition forchanges over a few days or a week. Backpain’s power to inflict misery is great,but that power is usually transient. Inmost cases, time and perseverance willcarry a patient through to recovery.


Further Reading

A New Clinical Model for the Treatment of Low-Back Pain. Gordon Waddell inSpine, Vol. 12, No. 7, pages 632–644; 1987.

Cost, Controversy, Crisis: Low Back Pain and the Health of the Public. Richard A.Deyo, Daniel Cherkin, Douglas Conrad and Ernest Volinn in Annual Review of PublicHealth, Vol. 12, pages 141–156; 1991.

Physician Variation in Diagnostic Testing for Low Back Pain. Daniel C. Cherkin,Richard A. Deyo, Kimberly Wheeler and Marcia A. Ciol in Arthritis & Rheumatism, Vol.37, No. 1, pages 15–22; January 1994.

Magnetic Resonance Imaging of the Lumbar Spine in People without Back Pain.Maureen C. Jensen, Michael N. Brant-Zawadzki, Nancy Obuchowski, Michael T. Modic,Dennis Malkasian and Jeffrey S. Ross in New England Journal of Medicine, Vol. 331, No. 2,pages 69–73; July 14, 1994.

The Mindbody Prescription: Healing the Body, Healing the Pain. John E. Sarno.Warner Books, 1998.

The Author

RICHARD A. DEYO is a general in-ternist with a long-standing interest inclinical research on low-back problemsand is a professor in the departments ofmedicine and of health services at theUniversity of Washington. He received hisM.P.H. from its School of Public Healthand Community Medicine in 1981 andhis M.D. from the Pennsylvania StateUniversity School of Medicine in 1975.Deyo has studied various therapies forlow-back pain, including bed rest, exer-cise regimens and transcutaneous electri-cal nerve stimulation (TENS).

SA

Low-Back Pain


Computer. The word conjuresup images of keyboards andmonitors. Terms like “ROM,”

“RAM,” “gigabyte” and “megahertz”come to mind. We have grown accus-tomed to the idea that computationtakes place using electronic compo-nents on a silicon substrate.

But must it be this way? The comput-er that you are using to read thesewords bears little resemblance to a PC.Perhaps our view of computation is toolimited. What if computers were ubiq-uitous and could be found in manyforms? Could a liquid computer exist inwhich interacting molecules performcomputations? The answer is yes. Thisis the story of the DNA computer.

Rediscovering Biology

My involvement in this story beganin 1993, when I walked into a

molecular biology lab for the first time.Although I am a mathematician andcomputer scientist, I had done a bit ofAIDS research, which I believed and stillbelieve to be of importance [see “Bal-anced Immunity,” by John Rennie; Sci-entific American, May 1993]. Unfor-

tunately, I had been remarkably unsuc-cessful in communicating my ideas tothe AIDS research community. So, in aneffort to become a more persuasive ad-vocate, I decided to acquire a deeperunderstanding of the biology of HIV.Hence, the molecular biology lab. There,under the guidance of Nickolas Chelya-pov (now chief scientist in my own lab-oratory), I began to learn the methodsof modern biology.

I was fascinated. With my own hands,I was creating DNA that did not existin nature. And I was introducing it intobacteria, where it acted as a blueprint forproducing proteins that would changethe very nature of the organism.

During this period of intense learn-ing, I began reading the classic text TheMolecular Biology of the Gene, co-au-thored by James D. Watson of Watson-Crick fame. My concept of biology wasbeing dramatically transformed. Biolo-gy was no longer the science of thingsthat smelled funny in refrigerators (myview from undergraduate days in the1960s at the University of California atBerkeley). The field was undergoing arevolution and was rapidly acquiringthe depth and power previously associ-

ated exclusively with the physical sci-ences. Biology was now the study of in-formation stored in DNA—strings offour letters: A, T, G and C for the basesadenine, thymine, guanine and cyto-sine—and of the transformations thatinformation undergoes in the cell. Therewas mathematics here!

Late one evening, while lying in bedreading Watson’s text, I came to a de-scription of DNA polymerase. This isthe king of enzymes—the maker of life.Under appropriate conditions, given astrand of DNA, DNA polymerase pro-duces a second “Watson-Crick” com-plementary strand, in which every C isreplaced by a G, every G by a C, everyA by a T and every T by an A. For ex-ample, given a molecule with the se-quence CATGTC, DNA polymerasewill produce a new molecule with thesequence GTACAG. The polymeraseenables DNA to reproduce, which inturn allows cells to reproduce and ulti-mately allows you to reproduce. For astrict reductionist, the replication ofDNA by DNA polymerase is what lifeis all about.

DNA polymerase is an amazing littlenanomachine, a single molecule that

Computing with DNA54 Scientific American August 1998

Computing with DNAThe manipulation of DNA to solve mathematical problems is redefining

what is meant by “computation”

by Leonard M. Adleman


“hops” onto a strand of DNA and slidesalong it, “reading” each base it passesand “writing” its complement onto anew, growing DNA strand. While lyingthere admiring this amazing enzyme, Iwas struck by its similarity to somethingdescribed in 1936 by Alan M. Turing,the famous British mathematician. Tur-ing—and, independently, Kurt Gödel,Alonzo Church and S. C. Kleene—hadbegun a rigorous study of the notion of“computability.” This purely theoreti-cal work preceded the advent of actualcomputers by about a decade and led tosome of the major mathematical resultsof the 20th century [see “UnsolvedProblems in Arithmetic,” by HowardDeLong; Scientific American, March1971; and “Randomness in Arithme-tic,” by Gregory J. Chaitin; ScientificAmerican, July 1988].

For his study, Turing had invented a“toy” computer, now referred to as aTuring machine. This device was notintended to be real but rather to be con-ceptual, suitable for mathematical in-vestigation. For this purpose, it had tobe extremely simple—and Turing suc-ceeded brilliantly. One version of hismachine consisted of a pair of tapes

and a mechanism called a finite control,which moved along the “input” tapereading data while simultaneously mov-ing along the “output” tape reading andwriting other data. The finite controlwas programmable with very simple in-structions, and one could easily write aprogram that would read a string of A,T, C and G on the input tape and writethe Watson-Crick complementary stringon the output tape. The similarities withDNA polymerase could hardly havebeen more obvious.

But there was one important piece ofinformation that made this similaritytruly striking: Turing’s toy computerhad turned out to be universal—simpleas it was, it could be programmed tocompute anything that was computableat all. (This notion is essentially the con-tent of the well-known “Church’s the-sis.”) In other words, one could programa Turing machine to produce Watson-Crick complementary strings, factornumbers, play chess and so on. This re-alization caused me to sit up in bed andremark to my wife, Lori, “Jeez, thesethings could compute.” I did not sleepthe rest of the night, trying to figure outa way to get DNA to solve problems.

My initial thinking was to make aDNA computer in the image of a Tur-ing machine, with the finite control re-placed by an enzyme. Remarkably, es-sentially the same idea had been sug-gested almost a decade earlier by CharlesH. Bennet and Rolf Landauer of IBM[see “The Fundamental Physical Limitsof Computation”; Scientific Ameri-can, July 1985]. Unfortunately, while anenzyme (DNA polymerase) was knownthat would make Watson-Crick comple-ments, it seemed unlikely that any exist-ed for other important roles, such asfactoring numbers.

This brings up an important factabout biotechnologists: we are a com-munity of thieves. We steal from thecell. We are a long way from being able

Computing with DNA

TOM

O N

AR

ASH

IMA

DNA MOLECULES—with their sequences ofadenine, thymine, guanine and cytosine (repre-sented by the letters A, T, G and C)—can beused to store information and to perform com-putations. The molecule shown here in color,GCAGTCGGACTGGGCTATGTCCGA, en-codes the solution to the sample HamiltonianPath Problem on the next page.



to create de novo miraculous molecularmachines such as DNA polymerase.Fortunately, three or four billion yearsof evolution have resulted in cells thatare full of wondrous little machines. Itis these machines, stolen from the cell,that make modern biotechnology possi-ble. But a molecular machine that wouldplay chess has apparently never evolved.So if I were to build a DNA computerthat could do something interesting, Iwould have to do it with the tools that

were at hand. These tools were essen-tially the following:

1. Watson-Crick pairing. As statedearlier, every strand of DNA has its Wat-son-Crick complement. As it happens,if a molecule of DNA in solution meetsits Watson-Crick complement, then thetwo strands will anneal—that is, twistaround each other to form the famousdouble helix. The strands are not cova-lently bound but are held together by

weak forces such as hydrogen bonds. Ifa molecule of DNA in solution meets aDNA molecule to which it is not com-plementary (and has no long stretchesof complementarity), then the two mol-ecules will not anneal.

2. Polymerases. Polymerases copy in-formation from one molecule into an-other. For example, DNA polymerasewill make a Watson-Crick complemen-tary DNA strand from a DNA template.In fact, DNA polymerase needs a “start


Consider a map of cities connected by certain nonstopflights (top right). For instance, in the example shown here,

it is possible to travel directly from Boston to Detroit but not viceversa. The goal is to determine whether a path exists that willcommence at the start city (Atlanta), finish at the end city (De-troit) and pass through each of the remaining cities exactly once.In DNA computation, each city is assigned a DNA sequence(ACTTGCAG for Atlanta) that can be thought of as a first name(ACTT) followed by a last name (GCAG). DNA flight numbers canthen be defined by concatenating the last name of the city oforigin with the first name of the city of destination (bottom right).

The complementary DNA city names are the Watson-Crickcomplements of the DNA city names in which every C is replacedby a G, every G by a C, every A by a T, and every T by an A. (To sim-plify the discussion here, details of the 3′ versus 5′ ends of theDNA molecules have been omitted.) For this particular problem,

only one Hamiltonian path exists,and it passes through Atlanta,Boston, Chicago and Detroit in thatorder. In the computation, thispath is represented by GCAGTCG-GACTGGGCTATGTCCGA, a DNA se-quence of length 24. Shown at theleft is the map with seven citiesand 14 nonstop flights used in theactual experiment. —L.M.A. SL

IM F

ILM

SSTART

END

ATLANTA TO BOSTONBOSTON TO CHICAGO

CHICAGO TO DETROIT

COMPLEMENT OF BOSTON

COMPLEMENT OF CHICAGO

G

CA

A

A

A

A

A

A

A

A

G

T

T

T

T

T

T

T

C

C C

C

C

C

C

C

C

C

C

G

G

G

G

G

G

GG

G

G

Hamiltonian Path Problem

CITY

ATLANTABOSTONCHICAGODETROIT

FLIGHT

ATLANTA - BOSTONATLANTA - DETROITBOSTON - CHICAGOBOSTON - DETROITBOSTON - ATLANTACHICAGO - DETROIT

DNA NAME

A C T T G C A GT C G G A C T GG G C TAT G TC C G A G C A A

DNA FLIGHT NUMBER

G C A G T C G GG C A G C C G AA C T G G G C TA C T G C C G AA C T G A C T TAT G T C C G A

COMPLEMENT

T G A A C G T CA G C C T G A CC C G ATA C AG G C T C G T T

Boston

Atlanta

Detroit

Chicago

WATSON-CRICKANNEALING,in which Cs pairwith Gs and As joinwith Ts, will result inDNA flight-numberstrands (shown here: At-lanta to Boston, Bostonto Chicago, and Chicagoto Detroit) being held end-to-end by strands encodingthe complementary DNA citynames (shown here: Bostonand Chicago).


signal” to tell it where to begin makingthe complementary copy. This signal isprovided by a primer—a (possibly short)piece of DNA that is annealed to thetemplate by Watson-Crick complemen-tarity. Wherever such a primer-templatepair is found, DNA polymerase will be-gin adding bases to the primer to createa complementary copy of the template.

3. Ligases. Ligases bind molecules to-gether. For example, DNA ligase willtake two strands of DNA in proximityand covalently bond them into a singlestrand. DNA ligase is used by the cell torepair breaks in DNA strands that oc-cur, for instance, after skin cells are ex-posed to ultraviolet light.

4. Nucleases. Nucleases cut nucleicacids. For example, restriction endonu-cleases will “search” a strand of DNAfor a predetermined sequence of basesand, when found, will cut the moleculeinto two pieces. EcoRI (from Escheri-chia coli) is a restriction enzyme that willcut DNA after the G in the sequenceGAATTC—it will almost never cut astrand of DNA anywhere else. It hasbeen suggested that restriction enzymesevolved to protect bacteria from viruses(yes, even bacteria have viruses!). Forexample, E. coli has a means (methyla-tion) of protecting its own DNA fromEcoRI, but an invading virus with thedeadly GAATTC sequence will be cutto pieces. My DNA computer did notuse restriction enzymes, but they havebeen used in subsequent experimentsby many other research groups.

5. Gel electrophoresis. This is notstolen from the cell. A solution of het-erogeneous DNA molecules is placed inone end of a slab of gel, and a current isapplied. The negatively charged DNAmolecules move toward the anode, with

shorter strands moving more quicklythan longer ones. Hence, this processseparates DNA by length. With specialchemicals and ultraviolet light, it is pos-sible to see bands in the gel where theDNA molecules of various lengths havecome to rest.

6. DNA synthesis. It is now possibleto write a DNA sequence on a piece ofpaper, send it to a commercial synthesisfacility and in a few days receive a testtube containing approximately 1018

molecules of DNA, all (or at least most)of which have the described sequence.Currently sequences of length approxi-mately 100 can be reliably handled inthis manner. For a sequence of length20, the cost is about $25. The moleculesare delivered dry in a small tube and ap-pear as a small, white, amorphous lump.

None of these appeared likely to helpplay chess, but there was another im-portant fact that the great logicians ofthe 1930s taught us: computation iseasy. To build a computer, only twothings are really necessary—a methodof storing information and a few simpleoperations for acting on that informa-tion. The Turing machine stores infor-mation as sequences of letters on tapeand manipulates that information withthe simple instructions in the finite con-trol. An electronic computer stores in-formation as sequences of zeros andones in memory and manipulates thatinformation with the operations avail-able on the processor chip. The remark-able thing is that just about any methodof storing information and any set ofoperations to act on that informationare good enough.

Good enough for what? For univer-sal computation—computing anything

that can be computed. To get your com-puter to make Watson-Crick comple-ments or to play chess, you need onlystart with the correct input informationand apply the right sequence of opera-tions—that is, run a program. DNA is agreat way to store information. In fact,the cell has been using this method tostore the “blueprint for life” for billionsof years. Further, enzymes such as poly-merases and ligases have been used tooperate on this information. Was thereenough to build a universal computer?Because of the lessons of the 1930s, Iwas sure that the answer was yes.

The Hamiltonian Path Problem

The next task was choosing a prob-lem to solve. It should not appear

to be contrived to fit the machine, andit should demonstrate the potential ofthis novel method of computation. Theproblem I chose was the HamiltonianPath Problem.

William Rowan Hamilton was As-tronomer Royal of Ireland in the mid-19th century. The problem that hascome to bear his name is illustrated inthe box on the opposite page. Let thearrows (directed edges) represent thenonstop flights between the cities (ver-tices) in the map (graph). For example,you can fly nonstop from Boston toChicago but not from Chicago to Bos-ton. Your job (the Hamiltonian PathProblem) is to determine if a sequenceof connecting flights (a path) exists thatstarts in Atlanta (the start vertex) andends in Detroit (the end vertex), whilepassing through each of the remainingcities (Boston and Chicago) exactlyonce. Such a path is called a Hamilton-ian path. In the example shown on

Computing with DNA Scientific American August 1998 57

LIGASES connectthe splinted mole-cules. Wherever theprotein finds twostrands of DNA inproximity, it willcovalently bondthem into a singlestrand.

TOM

O N

AR

ASH

IMA

LIGASE

GC

A

A

A

A

A

A

A

AA

G

T

T

T

T

T

T

T

CC

C

C

C

C

C

C

C

C

C

G

G

GG

GG

GGG

G


, it is easy to see that a uniqueHamiltonian path exists, and it passesthrough the cities in this order: Atlanta,Boston, Chicago, Detroit. If the startcity were changed to Detroit and theend city to Atlanta, then clearly therewould be no Hamiltonian path.

More generally, given a graph withdirected edges and a specified start ver-tex and end vertex, one says there is aHamiltonian path if and only if there is apath that starts at the start vertex, endsat the end vertex and passes though eachremaining vertex exactly once. TheHamiltonian Path Problem is to decidefor any given graph with specified startand end vertices whether a Hamiltoni-an path exists or not.

The Hamiltonian Path Problem hasbeen extensively studied by computerscientists. No efficient (that is, fast) al-gorithm to solve it has ever emerged. Infact, it seems likely that even using thebest currently available algorithms andcomputers, there are some graphs offewer than 100 vertices for which de-termining whether a Hamiltonian pathexists would require hundreds of years.

In the early 1970s the Hamiltonian

Path Problem was shown to be “NP-complete.” Without going into the the-ory of NP-completeness, suffice it tosay that this finding convinced mosttheoretical computer scientists that noefficient algorithm for the problem ispossible at all (though proving this re-mains the most important open prob-lem in theoretical computer science, theso-called NP = P? problem [see “TuringMachines,” by John E. Hopcroft; Scien-tific American, May 1984]). This isnot to say that no algorithms exist forthe Hamiltonian Path Problem, just noefficient ones. For example, consider thefollowing algorithm:

Given a graph with n vertices,1. Generate a set of random

paths through the graph.2. For each path in the set:

a. Check whether that path starts at the start vertex and ends with the end vertex. If not, remove that path from the set.

b. Check if that path passes through exactly n vertices. If not, remove that path

from the set.c. For each vertex, check if

that path passes through that vertex. If not, remove that path from the set.

3. If the set is not empty, then report that there is a Hamil-tonian path. If the set is empty, report that there is no Hamiltonian path.

This is not a perfect algorithm; never-theless, if the generation of paths is ran-dom enough and the resulting set largeenough, then there is a high probabilitythat it will give the correct answer. It isthis algorithm that I implemented in thefirst DNA computation.

Seven Days in a Lab

For my experiment, I sought a Ham-iltonian Path Problem small enough

to be solved readily in the lab, yet largeenough to provide a clear “proof ofprinciple” for DNA computation. I chosethe seven-city, 14-flight map shown inthe inset on page 56. A nonscientificstudy has shown that it takes about 54


POLYMERASE CHAIN REACTION, or PCR, is used to replicate DNA molecules that begin withthe start city (Atlanta) and terminate with the end city (Detroit). In this example, a primer—GGCT,representing the complement of the first name of Detroit—is annealed to the right end of a DNAstrand. The primer signals the polymerase to begin making a Watson-Crick complement of the strand.After the polymerase is done, the double helix is split into two strands so that Watson-Crick comple-ments of each half can be made. This process is repeated to obtain a large number of copies of mole-cules that have the correct start and end cities. Gel electrophoresis is then used to isolate those moleculesthat have the right sequence length of 24.

DNA POLYMERASE

G C

A

AA

A

A

G

T

T

TT

T

C

C

C

C

CC

G

G G

G

G

G


seconds on average to find the uniqueHamiltonian path in this graph. (Youmay begin now... .)

To simplify the discussion here, con-sider the map on page 56, which con-tains just four cities—Atlanta, Boston,Chicago and Detroit—linked by sixflights. The problem is to determine theexistence of a Hamiltonian path start-ing in Atlanta and ending in Detroit.

I began by assigning a random DNAsequence to each city. In our example,Atlanta becomes ACTTGCAG, BostonTCGGACTG and so on. It was conve-nient to think of the first half of theDNA sequence as the first name of thecity and the second half as the lastname. So Atlanta’s last name is GCAG,whereas Boston’s first name is TCGG.Next, I gave each nonstop flight a DNA“flight number,” obtained by concate-nating the last name of the city of originwith the first name of the city of desti-nation. In the example on page 56, theAtlanta-to-Boston flight number be-comes GCAGTCGG.

Recall that each strand of DNA hasits Watson-Crick complement. Thus,each city has its complementary DNA

name. Atlanta’s complementary namebecomes, for instance, TGAACGTC.

After working out these encodings, Ihad the complementary DNA city namesand the DNA flight numbers synthesized.(As it turned out, the DNA city namesthemselves were largely unnecessary.) Itook a pinch (about 1014 molecules) ofeach of the different sequences and putthem into a common test tube. To be-gin the computation, I simply addedwater—plus ligase, salt and a few otheringredients to approximate the condi-tions inside a cell. Altogether only aboutone fiftieth of a teaspoon of solutionwas used. Within about one second, Iheld the answer to the HamiltonianPath Problem in my hand.

To see how, consider what transpiresin the tube. For example, the Atlanta-to-Boston flight number (GCAGTCGG)and the complementary name of Boston(AGCCTGAC) might meet by chance.By design, the former sequence ends

with TCGG, and the latter starts withAGCC. Because these sequences arecomplementary, they will stick togeth-er. If the resulting complex now en-counters the Boston-to-Chicago flightnumber (ACTGGGCT), it, too, willjoin the complex because the end ofthe former (TGAC) is complementaryto the beginning of the latter (ACTG).In this manner, complexes will grow inlength, with DNA flight numbers splint-ed together by complementary DNAcity names. The ligase in the mixturewill then permanently concatenate thechains of DNA flight numbers. Hence,the test tube contains molecules thatencode random paths through thedifferent cities (as required in the firststep of the algorithm).

Because I began with such a largenumber of DNA molecules and the

problem contained just a handful ofcities, there was a virtual certainty thatat least one of the molecules formedwould encode the Hamiltonian path. Itwas amazing to think that the solutionto a mathematical problem could bestored in a single molecule!

Notice also that all the paths werecreated at once by the simultaneous in-teractions of literally hundreds of tril-lions of molecules. This biochemical re-action represents enormous parallelprocessing.

For the map on page 56, there is onlyone Hamiltonian path, and it goesthrough Atlanta, Boston, Chicago andDetroit, in that order. Thus, the mole-cule encoding the solution will have thesequence GCAGTCGGACTGGGCT-ATGTCCGA.

Unfortunately, although I held the so-lution in my hand, I also held about 100trillion molecules that encoded pathsthat were not Hamiltonian. These hadto be eliminated. To weed out moleculesthat did not both begin with the startcity and terminate with the end city, Irelied on the polymerase chain reaction(PCR). This important technique re-quires many copies of two short piecesof DNA as primers to signal the DNApolymerase to start its Watson-Crickreplication. The primers used were thelast name of the start city (GCAG for


TOM

O N

AR

ASH

IMA

PRIMER

A

A

A

T

T

T

C

C

C

C

GG

G

G


Atlanta) and the Watson-Crick comple-ment of the first name of the end city(GGCT for Detroit). These two prim-ers worked in concert: the first alertedDNA polymerase to copy complementsof sequences that had the right start city,and the second initiated the duplicationof molecules that encoded the correctend city.

PCR proceeds through thermocycling,repeatedly raising and lowering the tem-perature of the mixture in the test tube.Warm conditions encourage the DNApolymerase to begin duplicating; a hot-ter environment causes the resulting an-nealed strands to split from their dou-ble-helix structure, enabling subsequentreplication of the individual pieces.

The result was that molecules withboth the right start and end cities werereproduced at an exponential rate. Incontrast, molecules that encoded theright start city but an incorrect end city,or vice versa, were duplicated in a muchslower, linear fashion. DNA sequencesthat had neither the right start nor endwere not duplicated at all. Thus, by tak-ing a small amount of the mixture afterthe PCR was completed, I obtained asolution containing many copies of themolecules that had both the right startand end cities, but few if any moleculesthat did not meet this criterion. Thus,step 2a of the algorithm was complete.

Next, I used gel electrophoresis toidentify those molecules that had the

right length (in the exam-ple on page 56, a length of24). All other molecules werediscarded. This completed step2b of the algorithm.

To check the remaining se-quences for whether their pathspassed through all the intermediarycities, I took advantage of Watson-Crick annealing in a procedure calledaffinity separation. This process usesmultiple copies of a DNA “probe” mol-ecule that encodes the complementaryname of a particular city (for example,Boston). These probes are attached tomicroscopic iron balls, each approxi-mately one micron in diameter.

I suspended the balls in the tube con-taining the remaining molecules underconditions that encouraged Watson-Crick pairing. Only those molecules thatcontained the desired city’s name (Bos-ton) would anneal to the probes. Then Iplaced a magnet against the wall of the


PROBE MOLECULES are used to locate DNA strands encoding paths that passthrough the intermediate cities (Boston and Chicago). Probe molecules containing thecomplementary DNA name of Boston (AGCCTGAC) are attached to an iron ball sus-pended in liquid. Because of Watson-Crick affinity, the probes capture DNA strandsthat contain Boston’s name (TCGGACTG). Strands missing Boston’s name are thendiscarded. The process is repeated with probe molecules encoding the complementaryDNA name of Chicago. When all the computational steps are completed, the strandsleft will be those that encode the solution GCAGTCGGACTGGGCTATGTCCGA.

TOM

O N

AR

ASH

IMA

IRON BALL

PROBE MOLECULES

C

A

A

A

AA

A

A

G

G

G

T

T

T

T

T

C

C

C

C

C

C C

C

C

G

GG

G G

G

G

G

C

A

A

G

T

C

G


test tube to attract and hold the metalballs to the side while I poured out theliquid phase containing molecules thatdid not have the desired city’s name.

I then added new solvent and removedthe magnet in order to resuspend theballs. Raising the temperature of themixture caused the molecules to breakfree from the probes and redissolve inthe liquid. Next, I reapplied the magnetto attract the balls again to the side ofthe test tube, but this time without anymolecules attached. The liquid, whichnow contained the desired DNA strands(in the example, encoding paths thatwent through Boston), could then bepoured into a new tube for furtherscreening. The process was repeated forthe remaining intermediary cities (Chi-cago, in this case). This iterative proce-dure, which took me an entire day tocomplete in the lab, was the most te-dious part of the experiment.

At the conclusion of the affinity sepa-rations, step 2c of the algorithm wasover, and I knew that the DNA mole-

cules left in the tube should be preciselythose encoding Hamiltonian paths.Hence, if the tube contained any DNAat all, I could conclude that a Hamilto-nian path existed in the graph. NoDNA would indicate that no such pathexisted. Fortunately, to make this deter-mination I could use an additional PCRstep, followed by another gel-elec-trophoresis operation. To my delight,the final analysis revealed that the mol-ecules that remained did indeed encodethe desired Hamiltonian path. Afterseven days in the lab, the first DNAcomputation was complete.

A New Field Emerges

What about the future? It is clearthat molecular computers have

many attractive properties. They pro-vide extremely dense information stor-age. For example, one gram of DNA,which when dry would occupy a volumeof approximately one cubic centimeter,can store as much information as ap-proximately one trillion CDs. They pro-vide enormous parallelism. Even in thetiny experiment carried out in one fifti-eth of a teaspoon of solution, approxi-mately 1014 DNA flight numbers weresimultaneously concatenated in aboutone second. It is not clear whether thefastest supercomputer available todaycould accomplish such a task so quickly.

Molecular computers also have thepotential for extraordinary energy effi-ciency. In principle, one joule is sufficientfor approximately 2 × 1019 ligation op-erations. This is remarkable consider-ing that the second law of thermody-namics dictates a theoretical maximumof 34 × 1019 (irreversible) operations perjoule (at room temperature). Existingsupercomputers are far less efficient, ex-ecuting at most 109 operations per joule.

Experimental and theoretical scien-tists around the world are working toexploit these properties. Will they suc-ceed in creating molecular computers

that can compete with electronic com-puters? That remains to be seen. Hugefinancial and intellectual investmentsover half a century have made electron-ic computers the marvels of our age—

they will be hard to beat.But it would be shortsighted to view

this research only in such practicalterms. My experiment can be viewed asa manifestation of an emerging newarea of science made possible by ourrapidly developing ability to control themolecular world. Evidence of this new“molecular science” can be found inmany places. For example, Gerald F.Joyce of Scripps Research Institute inLa Jolla, Calif., “breeds” trillions ofRNA molecules, generation after gener-ation, until “champion” moleculesevolve that have the catalytic propertieshe seeks [see “Directed Molecular Evo-lution,” by Gerald F. Joyce; ScientificAmerican, December 1992]. Julius Re-bek, Jr., of the Massachusetts Instituteof Technology creates molecules thatcan reproduce—informing us about howlife on the earth may have arisen [see“Synthetic Self-Replicating Molecules,”by Julius Rebek, Jr.; Scientific Ameri-can, July 1994]. Stimulated by researchon DNA computation, Erik Winfree ofthe California Institute of Technologysynthesizes “intelligent” molecular com-plexes that can be “programmed” toassemble themselves into predeterminedstructures of arbitrary complexity. Thereare many other examples. It is the enor-mous potential of this new area that weshould focus on and nurture.

For me, it is enough just to know thatcomputation with DNA is possible. Inthe past half-century, biology and com-puter science have blossomed, andthere can be little doubt that they willbe central to our scientific and econom-ic progress in the new millennium. Butbiology and computer science—life andcomputation—are related. I am confi-dent that at their interface great discov-eries await those who seek them.


The Author

LEONARD M. ADLEMAN received a Ph.D. in computer sci-ence in 1976 from the University of California, Berkeley. In 1977 hejoined the faculty in the mathematics department at the Mas-sachusetts Institute of Technology, where he specialized in algorith-mic number theory and was one of the inventors of the RSA public-key cryptosystem. (The “A” in RSA stands for “Adleman.”) Soonafter joining the computer science faculty at the University of South-ern California, he was “implicated” in the emergence of computerviruses. He is a member of the National Academy of Engineering.

Further Reading

Molecular Computation of Solutions to CombinatorialProblems. Leonard M. Adleman in Science, Vol. 266, pages1021–1024; November 11, 1994.

On the Path to Computation with DNA. David K. Gifford inScience, Vol. 266, pages 993–994; November 11, 1994.

DNA Solution of Hard Computational Problems. Richard J.Lipton in Science, Vol. 268, pages 542–545; April 28, 1995.

Additional information on DNA computing can be found at http:// users.aol.com/ibrandt/dna_computer.html on the World Wide Web.

A

T

T

C C

G

G

C

SA


Since the space age began four de-cades ago, rockets have liftedmore than 20,000 metric tons of

material into orbit. Today 4,500 tons re-main in the form of nearly 10,000 “res-ident space objects,” only 5 percent ofwhich are functioning spacecraft. Theseobjects are just the large ones that mili-tary radars and telescopes can track. Ofincreasing interest to spacecraft opera-tors are the millions of smaller, untrack-able scraps scattered into orbits through-out near-Earth space, from only a fewhundred kilometers to more than 40,000kilometers (25,000 miles) above the sur-face of the planet.

If Earth’s tiny attendants moved likethe hordes of miniature moons aroundJupiter or Saturn, they would be a thingof beauty. The rings of the giant planetsare finely orchestrated; their constituentrocks and chunks of ice orbit in well-be-haved patterns, and collisions betweenthem occur at gentle velocities. ButEarth’s artificial satellites resemble an-gry bees around a beehive, seeming tomove randomly in all directions. Thepopulation density of satellites is fairlylow; the region around Earth is still avacuum by any terrestrial standard. Butthe haphazard motions of the swarmlead to huge relative velocities when ob-jects accidentally collide. A collision witha one-centimeter pebble can destroy aspacecraft. Even a single one-millimetergrain could wreck a mission.

The extraterrestrial refuse comes inmany forms: dead spacecraft, discardedrocket bodies, launch- and mission-relat-ed castoffs, remnants of satellite break-ups, solid-rocket exhaust, frayed sur-face materials and even droplets fromleaking nuclear reactors [see a in top il-lustration on page 64].

Although more than 4,800 spacecrafthave been placed into orbit, only about2,400 remain there; the rest have reen-tered Earth’s atmosphere. Of the surviv-ing spacecraft, 75 percent have complet-ed their missions and been abandoned.Most range in mass from one kilogramto 20 tons, although the Russian Mirspace station now exceeds 115 tons.The oldest and one of the smallest isAmerica’s second satellite, Vanguard 1,launched on March 17, 1958, but func-tional for only six years.

In addition to launching a spacecraft,most space missions also leave one ormore empty rocket stages in orbit. Forexample, Japan’s weather satellite, Hi-mawari 3, launched in 1984, discardedthree stages: one in a low Earth orbitthat varied from 170 to 535 kilometersin altitude; one near the satellite’s desti-nation, the circular geosynchronous or-bit at 35,785 kilometers; and one in anintermediate, highly elliptical orbit from175 to 36,720 kilometers. Two of theserocket bodies have since fallen back toEarth, one in 1984 and the other in1994. But by 1998 about 1,500 uselessupper stages were still circling overhead.

For the first quarter century of thespace age, spacecraft designers paid lit-tle attention to the environmental con-sequences of their actions. In addition tothe abandoned spacecraft and leftoverrockets, small components were routine-ly ejected into space. Explosive bolts,clamp bands and springs were releasedwhen satellites separated from the rock-ets that launched them. Many space-craft also threw off sensor covers or at-titude-control devices. Some Russianspace missions spawned more than 60distinct objects in various orbits. Perhapsthe largest piece of discarded equipment,

Monitoring and ControllingDebris in Spaceby Nicholas L. Johnson


SLIM

FIL

MS

Copyright 1998 Scientific American, Inc.Copyright 1998 Scientific American, Inc.

with a diameter of four meters and amass of 300 kilograms, is the upper halfof the European Space Agency’s SPEL-DA (Structure Porteuse Externe pourLancements Doubles Ariane), a deviceused for launching multiple satelliteswith a single Ariane rocket.

Human spaceflights, too, have throwngarbage overboard. In 1965, during thefirst space walk by an American, Gem-ini 4 astronaut Edward White lost aglove; going into its own orbit at 28,000kilometers per hour, it was potentiallythe most lethal article of clothing evermade by (and worn by) human hands.More than 200 objects, most in the formof garbage bags, drifted away from Mirduring its first decade in orbit. Fortu-nately, because astronauts and cosmo-nauts orbit at fairly low altitudes, from250 to 500 kilometers, the things theymisplace soon fall into the atmosphereand burn up. The famous glove reen-tered the atmosphere within a month.

Breaking Up Isn’t Hard to Do

The bigger problem comes from ro-botic missions at higher altitudes,

where debris dawdles. Mission-relatedrubbish accounts for more than 1,000known objects. Among them: 80 clumpsof needles released in May 1963 as partof a U.S. Department of Defense tele-communications experiment. The radi-ation pressure exerted by sunlight wasto have pushed the tiny needles—all 400million of them—out of orbit, but a de-ployment malfunction caused them to


The path from Sputnik to the International Space Station has been

littered with high-tech refuse,creating an environmental problem

in outer space

NUTS, BOLTS, CLAMPS AND WIRESare among the orbital flotsam from thebreakup of an old rocket, as depicted inthis artist’s conception. Over time, the de-bris will scatter, reducing its density. Mili-tary radars and telescopes track the largerpieces (inset, above). White dots representthese objects (not to scale with Earth).

ITT

KA

MA

N S

CIE

NC

ES C

ORP

OR

ATIO

N


bunch together and loiter in orbits up to6,000 kilometers above Earth’s surface.

By far the greatest source of space debris larger than 0.1 millimeter is thebreakup of satellites and rockets. Since1961 more than 150 satellites haveblown up or fallen apart, either acciden-tally or deliberately—scattering morethan 10,000 fragments large enough tobe tracked. The most environmentallydamaging fragmentations have beenexplosions of derelict rocket bodies withfuel left on board. Detonation, presum-ably caused by either overpressuriza-tion or ignition of the residual propel-lant, has occurred as soon as a few hoursor as late as 23 years after launch, and

nearly every booster type is vulnerable.The upper stage of a Pegasus rocket

launched in 1994 broke up on June 3,1996, creating the largest debris cloudon record—more than 700 objects largeenough to be tracked, in orbits from250 to 2,500 kilometers in altitude [seeillustration below]. This single event in-stantly doubled the official collisionhazard to the Hubble Space Telescope,which orbits just 25 kilometers below.Further radar observations revealed anestimated 300,000 pieces of debris larg-er than four millimeters, big enough todamage most spacecraft. During the sec-ond Hubble servicing mission in Febru-ary 1997, the space shuttle Discovery

had to maneuver away from a piece ofPegasus debris that was projected tocome within 1.5 kilometers. The astro-nauts also noticed pits on Hubble equip-ment and a hole in one of its antennas,which had apparently been caused by acollision with space debris before 1993.

Fifty satellites pose a special threat:they contain radioactive materials, ei-ther in nuclear reactors or in radioiso-tope thermoelectric generators. In 1978a nuclear-powered satellite, the SovietKosmos 954, accidentally crash-landedin northern Canada with its 30 kilo-grams of enriched uranium. Soviet plan-ners designed subsequent spacecraft toeject their fuel cores at the end of their

SIZE RANGE (METERS)

1,500

1,800

ESTIMATEDPOPULATION

1,000

TRACKED BYSURVEILLANCE

SYSTEMS

>100,000

500,000

ROCKET BODIES

OPERATIONALSPACECRAFT

DEFUNCT SPACECRAFT

OPERATIONAL DEBRIS

10–6 10–5 10–4 10–3 10–2 10–1 1 10 100

>1,000,000

>1,000,000

BREAKUP DEBRIS

SOLID-ROCKET-MOTOR SLAG

SODIUM-POTASSIUMPARTICLES

PAINT FLAKES

550

0

NU

MBE

R O

F A

NN

UA

L M

ISSI

ON

S

30

60

90

120

150

0CU

MU

LATI

VE M

ASS

IN O

RBIT

(MET

RIC

TO

NS)

1,000

2,000

3,000

4,000

5,000

1980 1990 199719701957YEAR

64 Scientific American August 1998 Monitoring and Controlling Debris in Space

DEBRIS ORBITS disperse soon after a satellite breakup. Threeweeks after a Pegasus rocket fell apart in June 1996, the debriswas still concentrated in a tight band (left). After three months,the slight asymmetry of Earth’s gravitational field had alreadystarted to amplify differences in the initial orbits (center). Afternine months, the debris was spread out in essentially random or-

der (right). In addition, though not depicted here, the objects ini-tially orbited in a pack and slowly fell out of sync. Debris spe-cialists must take this process into account when assessing therisk to spacecraft such as the shuttle. Initially the shuttle mayavoid the concentrated debris. Later, the risk of impact from dif-ferent directions must be evaluated statistically.

LAU

RIE

GR

AC

E

LAU

RIE

GR

AC

E

EARTH’S ARTIFICIAL SATELLITES, broadly defined as any-thing put into orbit either intentionally or not, come in varioustypes and size ranges, from paint particles a thousandth of a mil-limeter across to the Mir space station, 30 meters long (a). Fewernew space missions are launched today than in the past, but thecumulative mass of satellites has continued to climb because mod-ern spacecraft are larger (b). Military surveillance systems can

spot only those satellites larger than 10 centimeters to one meter(depending on the orbit). Their counts have risen every year ex-cept when solar activity has been increasing (c). The density ofthese trackable objects has three closely spaced peaks at 850-,1,000- and 1,500-kilometer altitudes, as well as smaller peaks atsemisynchronous orbit (20,000 kilometers) and geosynchronousorbit (36,000 kilometers)—the most popular destinations (d ).

a b

JUNE 1996 SEPTEMBER 1996 MARCH 1997


missions. Ejecting the fuel allows it toburn up if it reenters the atmosphere,rather than hit the ground as one lump.

This program ended altogether in1988, and no nuclear reactors have beenoperated in orbit since then, but theproblem has not gone away. During a1989 experiment sponsored by the Na-tional Aeronautics and Space Adminis-tration, the Jet Propulsion Laboratory’sGoldstone radar in southern Californiadetected a large cloud of sodium-potas-sium droplets (leaking reactor coolantfrom an earlier ejection of a fuel core).Later observations by the MassachusettsInstitute of Technology’s Haystack ra-dar confirmed a huge number of sodium-potassium spheres (perhaps 70,000),each about one centimeter across, nearthe 900-kilometer altitude where the re-actors and cores were retired.

Objects smaller than 0.1 millimeterare not as hazardous as the larger junk,but their sheer quantity regularly inflictsminor damage on spacecraft. Numeri-cally the most abundant debris is solid-rocket exhaust. Even the proper opera-tion of a solid-rocket motor can yield acolossal number of micron-size alumin-um oxide particles (up to 1020 of them),as well as centimeter-size slag. Althoughthe use of solid-rocket motors in spacehas declined during the past 10 years,the total amount of debris they gener-ate annually has actually risen becauseheavy, modern spacecraft require largermotors, which expel more exhaust.

Small particles in orbit also result fromthe long-term, natural degradation ofspacecraft materials. Millions of tinypaint particles now litter near-Earthspace. Space scientists have found thattrails of minute flakes accompany many,if not most, aging spacecraft and rocketbodies. (Because of their relative orbitalmotions, these trails precede, rather thanfollow, the spacecraft: as the debris fallstoward Earth, its speed increases.) Most

of the detritus is invisible to remote sen-sors, but occasionally fragments fromthermal blankets and carbon-basedcomponents are large enough to be de-tected. NASA’s Cosmic Background Ex-plorer satellite, for example, has inex-plicably discarded at least 80 objects.

Exposing Satellites

Although some futurists in the 1960s and 1970s speculated on the rising

number of objects in orbit, not until theearly 1980s did a scientific disciplineemerge to address the issue. NASA, asearly as 1966, evaluated the orbital col-lision hazards for human spaceflight,but these calculations considered onlythe objects that could be tracked. Nomeans were then available to ascertainthe number of smaller particles.

The larger pieces of space debris aremonitored by the same tracking systemsthe superpowers built during the coldwar to watch for missile attacks and en-emy spy satellites. The Space SurveillanceNetwork in the U.S. and the SpaceSurveillance System in the former Sovi-et Union maintain the official catalogueof some 10,000 resident space objects.Fifty radar, optical and electro-opticalsensors collect an average of 150,000observations a day to track these ob-jects. They can spot scraps as small as10 centimeters at lower altitudes andone meter at geosynchronous altitudes.

The spatial density of this debris de-pends on altitude, with peaks near 850,1,000 and 1,500 kilometers. At thesealtitudes, there is an average of one ob-ject per 100 million cubic kilometers.Above 1,500 kilometers the density de-creases with altitude, except for sharppeaks near semisynchronous (20,000kilometers) and geosynchronous (36,000kilometers) altitudes [see d in illustrationabove]. The clustering in particular or-bits reflects the designs of the various

families of satellites and rockets. Be-cause the junk comes from satellites, itis densest in the regions where satellitestend to go, thereby amplifying the dan-ger to functioning spacecraft.

Objects smaller than 10 centimetersescape the attention of the satellite track-ers. They are too dim for the telescopesand too small for the radar wavelengthused by the surveillance systems. Until1984, these smaller objects went unde-tected. Since then, scientists have estimat-ed their numbers by statistically sam-pling the sky over hundreds of hoursevery year. Researchers have used scien-tific radars, such as Goldstone, a Germanradio telescope and the huge Arecibodish in Puerto Rico, in bistatic mode—

transmitting signals from one dish andreceiving the satellite reflections on anearby antenna—to detect objects assmall as two millimeters. The Haystackand Haystack Auxiliary radars, parts ofa joint NASA–Department of Defenseproject, can detect objects nearly assmall. To calibrate the sensors used todetect small debris, in 1994 and 1995the space shuttle deployed special tar-gets in the form of spheres and needles.

To study even smaller objects, re-searchers must inspect the battered sur-faces of spacecraft that astronauts havepulled into the shuttle cargo bay andbrought back to Earth. NASA’s LongDuration Exposure Facility deliberatelyturned the other cheek, as it were, tospace debris from 1984 to 1990; it wasstruck by tens of thousands of artificialshards, as well as natural meteoroidsfrom comets and asteroids. Other sittingducks have been the European Retriev-able Carrier, Japan’s Space Flyer Unit,components from Hubble and the SolarMaximum Mission and, of course, thespace shuttle itself. Because of the shut-tle’s limited range, all these targets wererestricted to altitudes below 620 kilo-meters. For higher altitudes, researchers

Monitoring and Controlling Debris in Space Scientific American August 1998 65

01980

YEAR1990 199819701957

100

200

300

OFF

ICIA

L EA

RTH

SAT

ELLI

TE P

OPU

LATI

ON

400

500

0

2,000

4,000

6,000

SOLA

R RA

DIO

FLU

X (S

TAN

DA

RD U

NIT

S)

8,000

10,000

POPULATION DECREASESAS SOLAR FLUX INCREASES

0 10,000 20,000 30,000 40,000 50,000

10–7

10–8

10–9

10–10

10–11

10–12

10–13

10–14

10–15

10–16

ALTITUDE (KILOMETERS)

SPAT

IAL

DEN

SITY

PER

CU

BIC

KIL

OM

ETER

c d


must rely on theoretical models. Theshort orbital lifetime of small particlesat low altitudes implies that large reser-voirs of them exist at higher altitudes.

In all, researchers estimate there aremore than 100,000 objects between oneand 10 centimeters in Earth orbit, alongwith tens of millions between one milli-meter and one centimeter. In these sizeranges and larger, artificial objects faroutnumber natural meteoroids. Themeteoroid and orbital debris popula-tions are comparable in the 0.01- to1.0-millimeter size regime; at very smallsizes, orbital debris dominates again.

Births and Deaths

For the past two decades, the trackedsatellite population has grown at an

average rate of roughly 175 additionalobjects each year, for an overall increaseof 70 percent. More than one quarterof this growth has been due to satellitebreakups and the remainder to new mis-sions. Since 1957 spacefaring nationshave launched an average of 120 newspacecraft each year. Over the past de-

cade, however, the activity has calmeddown. It peaked at 129 missions in 1984and plummeted to 73 in 1996, the low-est figure since 1963 [see b in top illus-tration on page 64]. The launch rate re-bounded to 86 last year, largely becauseof the 10 launches for new communica-tions satellite constellations (Iridiumand Orbcom).

Much of the general decrease has oc-curred in the former Soviet Union. Rus-sia had 28 space missions in 1997: 21for domestic purposes, seven for for-eign commercial firms. A decade earlierthe U.S.S.R. launched 95 missions, allto support domestic programs. Becausethe shrinkage of Russian space activityhas mostly affected low-altitude flights,however, it has had little effect on thelong-term satellite population.

On the other side of the equation,three courses of action remove satellites.Satellite operators can deliberately steertheir spacecraft into the atmosphere;the space shuttle can pluck satellites outof orbit; and satellite orbits can spiraldownward naturally. The first two ac-tions, though helpful, have little effect

overall, because they typically apply onlyto objects in low orbits, which wouldhave been removed naturally anyway.

The third—natural orbital decay—ex-erts the primary cleansing force. Al-though we usually think of outer spaceas airless, Earth’s atmosphere actuallypeters out fairly slowly. There is stillenough air resistance in low Earth orbitto generate friction against fast-movingspacecraft. This drag force is self-rein-forcing: as satellites lose energy to drag,they fall inward and speed up, which in-creases the drag, causing them to lose en-ergy faster. Eventually, satellites enter thedenser parts of the atmosphere and burnup partially or wholly in a fiery streak.

This natural decay is more pro-nounced below 600 kilometers’ altitudebut is discernible as high as 1,500 kilo-meters. At the lower altitudes, satelliteslast only a few years unless they firetheir rockets to compensate for thedrag. During periods of higher-energyemissions from the sun, which recur onan 11-year cycle, Earth’s atmospherereceives more heat and expands, in-creasing the atmospheric drag. Aroundthe time of the last solar maximum in1989–1990, a record number of cata-logued satellites fell back to Earth—threea day, three times the average rate—

sweeping more than 560 tons from or-bit in a year’s time [see c in illustrationon page 65]. The Salyut 7 space station,the predecessor of Mir, fell victim tothis activity, reentering in early 1991.The previous solar maximum in 1979– 1980 brought down the U.S. Skylabspace station ahead of schedule. Wel-come though the natural garbage col-lection has been, it has not prevented thetotal satellite population from increas-ing. And atmospheric drag is negligiblefor satellites in higher orbits. Therefore,we cannot rely entirely on natural pro-cesses to solve the debris problem.

The consequences of the growth inEarth’s satellite population are varied.The space shuttle occasionally executesan evasive maneuver to dodge large,derelict satellites, and an average of onein eight of its windows must be replacedafter each mission because of pits gougedby hypervelocity dust. Increasingly, mis-sion planners must consider the risk ofimpacts when they choose the orienta-tion in which the shuttle flies. Debris ismost likely to strike in the orbital plane30 to 45 degrees left and right of the di-rection of travel. If there is no compel-ling need to fly in a different orienta-tion, the astronauts point the most sen-

Monitoring and Controlling Debris in Space66 Scientific American August 1998

SHIELDS should protect key components of the International Space Station from mostobjects too small to track but large enough to puncture station walls. These shields,unlike the “force fields” of science fiction, are barriers mounted on the spacecraft. Asthe object approaches, it first encounters a sheet of aluminum (typically two millime-ters thick), known as the Whipple bumper, which causes the projectile to shatter (a).The fragments are slowed by one or more layers of Kevlar (b). Finally, the fragmentsbounce off the spacecraft wall (c).

NAT

ION

AL

AER

ON

AU

TIC

S A

ND

SPA

CE

AD

MIN

ISTR

ATIO

N; L

AU

RIE

GR

AC

E


sitive shuttle surfaces away from thesethreatening directions.

In July 1996 the first recognized acci-dental collision of two catalogued satel-lites occurred. The functioning Frenchmilitary spacecraft Cerise was disabledwhen a fragment from a European rock-et body, which had exploded 10 yearsearlier, struck the spacecraft’s attitude-control boom at a speed of nearly 15kilometers per second (nearly 33,400miles per hour). Though damaged, thespacecraft was able to continue its mis-sion after heroic efforts by its controllers.

Unfortunately, all spacecraft remainvulnerable to objects from one to 10centimeters in size—too small for thesurveillance systems to see but largeenough to puncture spacecraft walls.The Department of Defense is slowlymoving to improve its space surveillancenetwork to cover this range of sizes, butthese systems are intended for humanspaceflight; they offer little protectionfor most satellites.

Vacuum Cleaners

Most spacecraft are still at little riskof impacts during their opera-

tional lifetimes, but the environment inthe future is likely to be less benign. TheCerise collision generated little extradebris, but such a collision could createthousands of fragments large enough toshatter other satellites, which would cre-ate their own debris, and so on. This cas-cading phenomenon might become thedominant factor in the far distant evo-lution of Earth’s satellite population.

To prevent such a scenario, spacefar-ing nations have begun working togeth-er over the past decade. The Inter-Agen-cy Space Debris Coordination Commit-tee currently includes representativesfrom the U.S., Russia, China, Japan, In-

dia, the European Space Agency, France,Britain and Germany. In 1994 the Unit-ed Nations Scientific and Technical Sub-committee on the Peaceful Uses of Out-er Space finally placed orbital debris onits agenda, with the goal of completingan assessment by 1999. Although ev-eryone now recognizes the problem, thecommittees have yet to decide whatshould be done, who should police thesituation, and how a balance should bestruck between the risks of damage andthe costs of mitigation and cleaning up.

Meanwhile individual agencies arestriving to generate less garbage. NASA

and Japan’s principal space agency,NASDA, have used bolt catchers andspecial tethers to curtail the release ofoperational debris; dumped fuel andturned off electrical systems on deadspacecraft to prevent breakups; andrecommended that new low-altitudesatellites and rocket bodies be “de-or-bited”—that is, steered into the atmo-sphere for incineration—no more than25 years after completing their missions.

After the devastating breakup of thePegasus in 1996, the designer and oper-ator of the rocket, Orbital SciencesCorporation, redesigned the upper stageand implemented new preventive mea-sures before missions resumed last De-cember. Some of the new satellite-tele-phone networks plan to de-orbit satel-lites on retirement. At higher altitudes,the collisional dangers are fewer, butthere is another problem: the lack ofroom in geosynchronous orbit. To keepdead satellites from occupying slotsneeded for working ones, satellite oper-ators are now asked to give their space-craft a proper burial in less crowded“graveyard” orbits. In January the U.S.government presented draft debris-mit-igation standards to the aerospace in-dustry for comment.

These procedures, however, do noth-ing about the junk already in space.Spacecraft designers are now encouragedto protect against collisions, particularlyby the many objects up to one centime-ter in size. Shields—that is, exterior bar-riers—can protect most spacecraft com-ponents. Not only do they increase reli-ability, but they reduce the amount ofsecondary debris in the event a collisiondoes occur. The International Space Sta-tion will contain advanced-technologyshields around habitable compartments,fuel lines, control gyroscopes and othersensitive areas [see illustration on oppo-site page]. Other components, such assolar panels, are impractical to shield;spacecraft planners must assume thatthey will slowly deteriorate over timebecause of small collisions.

Cleaning up the debris remains a tech-nological and economic challenge. Us-ing the shuttle to salvage debris is dan-gerous and impractical. Several research-ers have devised imaginative schemes.NASA, the Department of Defense andthe Department of Energy have studiedone proposal, Project Orion, that woulduse ground-based lasers to remove smalldebris. These lasers are nothing like thosein science fiction; even if they couldblast a satellite into pieces, that wouldsimply create more debris. Instead thelasers vaporize some satellite material,nudging it off course and eventually intothe atmosphere. Other mechanisms arestill on the drawing boards, includinggiant foam balls. A particle penetratingthe foam would lose energy and fallback to Earth sooner.

Whereas future generations may beable to undo the shortsighted practicesof the past, for now we must strive toprevent the uncontrolled growth of thesatellite population—or else prepareourselves for a lack of space in space.

Monitoring and Controlling Debris in Space Scientific American August 1998 67

The Author

NICHOLAS L. JOHNSON serves as the NationalAeronautics and Space Administration’s chief scientistfor orbital debris at the Johnson Space Center inHouston. He is head of the space agency’s delegationto the Inter-Agency Space Debris Coordination Com-mittee, a co-chair of the International Academy of As-tronautics’s Subcommittee on Space Debris and amember of the American Institute of Aeronautics andAstronautics’s Orbital Debris Committee on Stan-dards. Johnson was a member of the National Re-search Council’s space debris committee (1993–1994)and the International Space Station meteoroid/orbitaldebris risk management committee (1995–1996).

Further Reading

Artificial Space Debris. N. L. Johnson and D. S. McKnight. Orbit Books, 1991.Preservation of Near Earth Space for Future Generations. Edited by J. A.Simpson. Cambridge University Press, 1994.

Interagency Report on Orbital Debris. Office of Science and Technology Pol-icy, National Science Technology Council, November 1995. Available athttp://www-sn.jsc.nasa.gov/debris/report95.html on the World Wide Web.

Orbital Debris: A Technological Assessment. Committee on Space Debris,Aeronautics and Space Engineering Board, National Research Council. NationalAcademy Press, 1995.

Protecting the Space Station from Meteoroids and Orbital Debris.Committee on International Space Station Meteoroid/Orbital Debris Risk Man-agement, Aeronautics and Space Engineering Board, National Research Council.National Academy Press, 1997.

SA


“Amusement is one of thefields of applied math.”

—William F. White, A Scrapbook of

Elementary Mathematics

My “Mathemati-cal Games” col-umn began in

the December 1956 issue ofScientific American with anarticle on hexaflexagons.These curious structures, cre-ated by folding an ordinarystrip of paper into a hexagonand then gluing the ends to-gether, could be turned insideout repeatedly, revealing oneor more hidden faces. Thestructures were invented in1939 by a group of PrincetonUniversity graduate students.Hexaflexagons are fun toplay with, but more impor-tant, they show the link be-tween recreational puzzlesand “serious” mathematics:one of their inventors was Richard Feyn-man, who went on to become one of themost famous theoretical physicists ofthe century.

At the time I started my column, onlya few books on recreational mathemat-ics were in print. The classic of thegenre—Mathematical Recreations andEssays, written by the eminent Englishmathematician W. W. Rouse Ball in1892—was available in a version up-dated by another legendary figure, theCanadian geometer H.S.M. Coxeter.Dover Publications had put out a trans-

lation from the French of La Mathéma-tique des Jeux (Mathematical Recrea-tions), by Belgian number theorist Mau-rice Kraitchik. But aside from a few otherpuzzle collections, that was about it.

Since then, there has been a remark-able explosion of books on the subject,many written by distinguished mathe-maticians. The authors include Ian Stew-art, who now writes Scientific Ameri-can’s “Mathematical Recreations” col-umn; John H. Conway of PrincetonUniversity; Richard K. Guy of the Uni-versity of Calgary; and Elwyn R. Berle-

kamp of the University ofCalifornia at Berkeley. Arti-cles on recreational mathe-matics also appear with in-creasing frequency in mathe-matical periodicals. Thequarterly Journal of Recrea-tional Mathematics beganpublication in 1968.

The line between entertain-ing math and serious math isa blurry one. Many profes-sional mathematicians regardtheir work as a form of play,in the same way professionalgolfers or basketball starsmight. In general, math isconsidered recreational if ithas a playful aspect that canbe understood and appreci-ated by nonmathematicians.Recreational math includeselementary problems withelegant, and at times surpris-ing, solutions. It also encom-passes mind-bending para-doxes, ingenious games, be-wildering magic tricks andtopological curiosities such

as Möbius bands and Klein bottles. Infact, almost every branch of mathemat-ics simpler than calculus has areas thatcan be considered recreational. (Someamusing examples are shown on theopposite page.)

Ticktacktoe in the Classroom

The monthly magazine published bythe National Council of Teachers

of Mathematics, Mathematics Teacher,often carries articles on recreational top-ics. Most teachers, however, continue to

A Quarter-Century of Recreational Mathematics68 Scientific American August 1998

A Quarter-Century of Recreational Mathematics

The author of Scientific American’s column “Mathematical Games” from 1956 to 1981 recounts 25 years of amusing puzzles and serious discoveries

by Martin Gardner

MARTIN GARDNER continues to tackle mathematical puz-zles at his home in Hendersonville, N.C. The 83-year-old writ-er poses next to a Klein bottle, an object that has just one sur-face: the bottle’s inside and outside connect seamlessly.

DO

NN

AB

ISE

Gam

ma

Liai

son


A Quarter-Century of Recreational Mathematics Scientific American August 1998 69

1 2

7 8 9 10

11

17

24

33

40 41 42 43 44 45 46

50494847

34 35 36 37 38 39

25 26 27 28 29 30 31 32

18 19 20 21 22 23

12 13 14 15 16

3 4 5 6

CUT THEDECK

RIFFLE-SHUFFLE

13

13

13

? ?

Printed above are the first three verses of Genesis in the KingJames Bible. Select any of the 10 words in the first verse: “In the

beginning God created the heaven and the earth.” Count the num-ber of letters in the chosen word and call this number x. Then go tothe word that is x words ahead. (For example, if you picked “in,” go to“beginning.”) Now count the number of letters in this word—call itn—then jump ahead another n words. Continue in this manner untilyour chain of words enters the third verse of Genesis.

On what word does your count end? Is the answer happenstanceor part of a divine plan?

Amagician arranges a deck of cards so that the black and red cards alternate. She cuts the deck about in half, making sure

that the bottom cards of each half are not the same color. Then sheallows you to riffle-shuffle the two halves together, as thoroughly orcarelessly as you please. When you’re done, she picks the first twocards from the top of the deck. They are a black card and a red card(not necessarily in that order). The next two are also a black card anda red card. In fact, every succeeding pair of cards will include one ofeach color. How does she do it? Why doesn’t shuffling the deck pro-duce a random sequence?

The matrix of numbers above is a curious type of magic square.Circle any number in the matrix, then cross out all the numbers

in the same column and row. Next, circle any number that has notbeen crossed out and again cross out the row and column containingthat number. Continue in this way until you have circled six numbers.

Clearly, each number has been randomly selected. But no matterwhich numbers you pick, they always add up to the same sum. Whatis this sum? And, more important, why does this trick always work?

Mr. Jones, a cardsharp, puts three cards face down on a table.One of the cards is an ace; the other two are face cards. You

place a finger on one of the cards, betting that this card is the ace.The probability that you’ve picked the ace is clearly 1/3. Jones now se-cretly peeks at each card. Because there is only one ace among thethree cards, at least one of the cards you didn’t choose must be a facecard. Jones turns over this card and shows it to you. What is the prob-ability that your finger is now on the ace?

Four Puzzles from Martin Gardner(The answers are on page 75.)

1 2

3 4

ILLU

STRA

TIO

NS

BY

IAN

WO

RPO

LE


ignore such material. For 40 years Ihave done my best to convince educa-tors that recreational math should beincorporated into the standard curricu-lum. It should be regularly introducedas a way to interest young students inthe wonders of mathematics. So far,though, movement in this direction hasbeen glacial.

I have often told a story from my ownhigh school years that illustrates the di-lemma. One day during math study pe-riod, after I’d finished my regular as-signment, I took out a fresh sheet of pa-per and tried to solve a problem thathad intrigued me: whether the first play-er in a game of ticktacktoe can alwayswin, given the right strategy. When myteacher saw me scribbling, she snatchedthe sheet away from me and said, “Mr.Gardner, when you’re in my class I ex-pect you to work on mathematics andnothing else.”

The ticktacktoe problem would makea wonderful classroom exercise. It is asuperb way to introduce students tocombinatorial mathematics, game theo-ry, symmetry and probability. More-over, the game is part of every student’sexperience: Who has not, as a child,played ticktacktoe? Yet I know fewmathematics teachers who have includ-ed such games in their lessons.

According to the 1997 yearbook ofthe mathematics teachers’ council, thelatest trend in math education is called“the new new math” to distinguish itfrom “the new math,” which floppedso disastrously several decades ago. Thenewest teaching system involves divid-ing classes into small groups of studentsand instructing the groups to solve prob-lems through cooperative reasoning.

“Interactive learning,” as it is called, issubstituted for lecturing. Although thereare some positive aspects of the newnew math, I was struck by the fact thatthe yearbook had nothing to say aboutthe value of recreational mathematics,which lends itself so well to cooperativeproblem solving.

Let me propose to teachers the follow-ing experiment. Ask each group of stu-dents to think of any three-digit num-ber—let’s call it ABC. Then ask the stu-dents to enter the sequence of digitstwice into their calculators, forming thenumber ABCABC. For example, if thestudents thought of the number 237,they’d punch in the number 237,237.Tell the students that you have the psy-chic power to predict that if they divideABCABC by 13 there will be no remain-der. This will prove to be true. Now askthem to divide the result by 11. Again,there will be no remainder. Finally, askthem to divide by 7. Lo and behold, theoriginal number ABC is now in the cal-culator’s readout. The secret to the trickis simple: ABCABC = ABC × 1,001 =ABC × 7 × 11 × 13. (Like every otherinteger, 1,001 can be factored into aunique set of prime numbers.) I know ofno better introduction to number theoryand the properties of primes than ask-ing students to explain why this trickalways works.

Polyominoes and Penrose Tiles

One of the great joys of writing theScientific American column over

25 years was getting to know so manyauthentic mathematicians. I myself amlittle more than a journalist who lovesmathematics and can write about it glib-

ly. I took no math courses in college. Mycolumns grew increasingly sophisticat-ed as I learned more, but the key to thecolumn’s popularity was the fascinatingmaterial I was able to coax from someof the world’s best mathematicians.

Solomon W. Golomb of the Universi-ty of Southern California was one ofthe first to supply grist for the column.In the May 1957 issue I introduced hisstudies of polyominoes, shapes formedby joining identical squares along theiredges. The domino—created from twosuch squares—can take only one shape,but the tromino, tetromino and pento-mino can assume a variety of forms: Ls,Ts, squares and so forth. One of Gol-omb’s early problems was to determinewhether a specified set of polyominoes,snugly fitted together, could cover acheckerboard without missing anysquares. The study of polyominoessoon evolved into a flourishing branchof recreational mathematics. Arthur C.Clarke, the science-fiction author, con-fessed that he had become a “pentomi-no addict” after he started playing withthe deceptively simple figures.

Golomb also drew my attention to aclass of figures he called “rep-tiles”—

identical polygons that fit together toform larger replicas of themselves. Oneof them is the sphinx, an irregular pen-tagon whose shape is somewhat similarto that of the ancient Egyptian monu-ment. When four identical sphinxes arejoined in the right manner, they form alarger sphinx with the same shape as itscomponents. The pattern of rep-tilescan expand infinitely: they tile the planeby making larger and larger replicas.

The late Piet Hein, Denmark’s illus-trious inventor and poet, became a


√3

21

a b e

c d

LOW-ORDER REP-TILES fit together to make larger replicasof themselves. The isosceles right triangle (a) is a rep-2 figure:two such triangles form a larger triangle with the same shape. Arep-3 triangle (b) has angles of 30, 60 and 90 degrees. Other rep-tiles include a rep-4 quadrilateral (c) and a rep-4 hexagon (d).The sphinx (e) is the only known rep-4 pentagon.

IAN

WO

RPO

LE


1

2

3

4

5

6

7

good friend through his contributionsto “Mathematical Games.” In the July1957 issue, I wrote about a topologicalgame he invented called Hex, which isplayed on a diamond-shaped boardmade of hexagons. Players place theirmarkers on the hexagons and try to bethe first to complete an unbroken chainfrom one side of the board to the other.The game has often been called John be-cause it can be played on the hexagonaltiles of a bathroom floor.

Hein also invented the Soma cube,which was the subject of several columns(September 1958, July 1969 and Sep-tember 1972). The Soma cube consistsof seven different polycubes, the three-dimensional analogues of polyominoes.They are created by joining identicalcubes at their faces. The polycubes canbe fitted together to form the Somacube—in 240 ways, no less—as well asa whole panoply of Soma shapes: thepyramid, the bathtub, the dog and so on.

In 1970 the mathematician John Con-way—one of the world’s undisputed ge-niuses—came to see me and asked if Ihad a board for the ancient Orientalgame of go. I did. Conway then dem-onstrated his now famous simulationgame called Life. He placed severalcounters on the board’s grid, then re-moved or added new counters accord-ing to three simple rules: each counterwith two or three neighboring countersis allowed to remain; each counter withone or no neighbors, or four or moreneighbors, is removed; and a new coun-ter is added to each empty space adja-

cent to exactly three counters. By ap-plying these rules repeatedly, an aston-ishing variety of forms can be created,including some that move across theboard like insects. I described Life in theOctober 1970 column, and it became aninstant hit among computer buffs. Formany weeks afterward, business firmsand research laboratories were almostshut down while Life enthusiasts exper-imented with Life forms on their com-puter screens.

Conway later collaborated with fel-low mathematicians Richard Guy andElwyn Berlekamp on what I considerthe greatest contribution to recreationalmathematics in this century, a two-vol-ume work called Winning Ways (1982).One of its hundreds of gems is a two-person game called Phutball, which canalso be played on a go board. The Phut-ball is positioned at the center of theboard, and the players take turns plac-ing counters on the intersections of thegrid lines. Players can move the Phutballby jumping it over the counters, whichare removed from the board after theyhave been leapfrogged. The object ofthe game is to get the Phutball past theopposing side’s goal line by building achain of counters across the board.What makes the game distinctive is that,unlike checkers, chess, go or Hex, Phut-ball does not assign different game piec-es to each side: the players use the samecounters to build their chains. Conse-quently, any move made by one Phut-ball player can also be made by his orher opponent.


SOMA PIECES are irregular shapes formed by joining unit cubes at their faces (above).The seven pieces can be arranged in 240 ways to build the 3-by-3-by-3 Soma cube. Thepieces can also be assembled to form all but one of the structures pictured at the right.Can you determine which structure is impossible to build? The answer is on page 75.

IAN

WO

RPO

LE

DOG

STAIRS

STEAMER

CASTLE

SKYSCRAPER

BATHTUB

TUNNEL

WELL

SOFA

WALL

CHAIR

PYRAMID


Other mathematicians who contrib-uted ideas for the column include FrankHarary, now at New Mexico State Uni-versity, who generalized the game ofticktacktoe. In Harary’s version of thegame, presented in the April 1979 is-sue, the goal was not to form a straightline of Xs or Os; instead players tried tobe the first to arrange their Xs or Os ina specified polyomino, such as an L ora square. Ronald L. Rivest of the Mas-sachusetts Institute of Technology al-lowed me to be the first to reveal—inthe August 1977 column—the “public-key” cipher system that he co-invented.It was the first of a series of ciphers thatrevolutionized the field of cryptology. Ialso had the pleasure of presenting the

mathematical art of Maurits C. Escher,which appeared on the cover of theApril 1961 issue of Scientific American,as well as the nonperiodic tiling discov-ered by Roger Penrose, the British math-ematical physicist famous for his workon relativity and black holes.

Penrose tiles are a marvelous exam-ple of how a discovery made solely forthe fun of it can turn out to have an un-expected practical use. Penrose devisedtwo kinds of shapes, “kites” and“darts,” that cover the plane only in anonperiodic way: no fundamental partof the pattern repeats itself. I explainedthe significance of the discovery in theJanuary 1977 issue, which featured apattern of Penrose tiles on its cover. A

few years later a 3-D form of Penrosetiling became the basis for constructinga previously unknown type of molecu-lar structure called a quasicrystal. Sincethen, physicists have written hundredsof research papers on quasicrystals andtheir unique thermal and vibrationalproperties. Although Penrose’s ideastarted as a strictly recreational pursuit,it paved the way for an entirely newbranch of solid-state physics.

Leonardo’s Flush Toilet

The two columns that generated thegreatest number of letters were my

April Fools’ Day column and the oneon Newcomb’s paradox. The hoax col-umn, which appeared in the April 1975issue, purported to cover great break-throughs in science and math. The start-ling discoveries included a refutation ofrelativity theory and the disclosure thatLeonardo da Vinci had invented the flushtoilet. The column also announced thatthe opening chess move of pawn toking’s rook 4 was a certain game win-ner and that e raised to the power of π × √163 was exactly equal to the inte-ger 262,537,412,640,768,744. To myamazement, thousands of readers failedto recognize the column as a joke. Ac-companying the text was a complicatedmap that I said required five colors toensure that no two neighboring regionswere colored the same. Hundreds ofreaders sent me copies of the map col-ored with only four colors, thus up-holding the four-color theorem. Manyreaders said the task had taken days.

Newcomb’s paradox is named afterphysicist William A. Newcomb, whooriginated the idea, but it was first de-scribed in a technical paper by HarvardUniversity philosopher Robert Nozick.The paradox involves two closed box-es, A and B. Box A contains $1,000.Box B contains either nothing or $1million. You have two choices: takeonly Box B or take both boxes. Takingboth obviously seems to be the betterchoice, but there is a catch: a superbe-ing—God, if you like—has the power of


a

b

c

d

e

BLOCK

BEEHIVE

BEEHIVE

BEEHIVE

IN THE GAME OF LIFE, forms evolve by following rulesset by mathematician John H. Conway. If four “organ-isms” are initially arranged in a square block of cells (a),the Life form does not change. Three other initial patterns(b, c and d) evolve into the stable “beehive” form. The fifthpattern (e) evolves into the oscillating “traffic lights” figure,which alternates between vertical and horizontal rows.

IAN

WO

RPO

LE


72°

72°

36°

36°

36°36°

36°36°

φ

φ

φ

φ

φ

√

1

11

= 1 + 52

knowing in advance how you willchoose. If he predicts that out ofgreed you will take both boxes, heleaves B empty, and you will getonly the $1,000 in A. But if he pre-dicts you will take only Box B, heputs $1 million in it. You have watchedthis game played many times with oth-ers, and in every case when the playerchose both boxes, he or she found thatB was empty. And every time a playerchose only Box B, he or she became amillionaire.

How should you choose? The prag-matic argument is that because of theprevious games you have witnessed,you can assume that the superbeingdoes indeed have the power to makeaccurate predictions. You should there-fore take only Box B to guarantee thatyou will get the $1 million. But wait!The superbeing makes his predictionbefore you play the game and has nopower to alter it. At the moment youmake your choice, Box B is either emp-ty, or it contains $1 million. If it is emp-ty, you’ll get nothing if you choose only

Box B. But if you choose both boxes, atleast you’ll get the $1,000 in A. And ifB contains $1 million, you’ll get the mil-lion plus another thousand. So how canyou lose by choosing both boxes?

Each argument seems unassailable.Yet both cannot be the best strategy. No-zick concluded that the paradox, whichbelongs to a branch of mathematicscalled decision theory, remains unre-

solved. My personal opinion is that theparadox proves, by leading to a logicalcontradiction, the impossibility of a su-perbeing’s ability to predict decisions. Iwrote about the paradox in the July1973 column and received so many let-ters afterward that I packed them into acarton and personally delivered them toNozick. He analyzed the letters in aguest column in the March 1974 issue.

Magic squares have long been a pop-ular part of recreational math. Whatmakes these squares magical is the ar-rangement of numbers inside them: thenumbers in every column, row and di-agonal add up to the same sum. Thenumbers in the magic square are usual-ly required to be distinct and run in


TRAFFIC LIGHTS

PENROSE TILES can be constructed by dividinga rhombus into a “kite” and a “dart” such thatthe ratio of their diagonals is phi (φ), the goldenratio (above). Arranging five of the darts arounda vertex creates a star. Placing 10 kites aroundthe star and extending the tiling symmetricallygenerate the infinite star pattern (right). Othertilings around a vertex include the deuce, jackand queen, which can also generate infinite pat-terns of kites and darts (below right).

IAN

WO

RPO

LE

DEUCE

JACK QUEEN


consecutive order, starting with one.There exists only one order-3 magicsquare, which arranges the digits onethrough nine in a three-by-three grid.(Variations made by rotating or reflect-ing the square are considered trivial.) Incontrast, there are 880 order-4 magicsquares, and the number of arrange-ments increases rapidly for higher orders.

Surprisingly, this is not the case withmagic hexagons. In 1963 I received inthe mail an order-3 magic hexagon de-vised by Clifford W. Adams, a retiredclerk for the Reading Railroad. I sentthe magic hexagon to Charles W. Trigg,a mathematician at Los Angeles CityCollege, who proved that this elegantpattern was the only possible order-3

magic hexagon—and that no magic hex-agons of any other size are possible!

What if the numbers in a magic squareare not required to run in consecutiveorder? If the only requirement is thatthe numbers be distinct, a wide varietyof order-3 magic squares can be con-structed. For example, there is an infin-ite number of such squares that containdistinct prime numbers. Can an order-3magic square be made with nine distinctsquare numbers? Two years ago in anarticle in Quantum, I offered $100 forsuch a pattern. So far no one has comeforward with a “square of squares”—

but no one has proved its impossibilityeither. If it exists, its numbers would behuge, perhaps beyond the reach of to-

day’s fastest supercomputers. Such amagic square would probably not haveany practical use. Why then are mathe-maticians trying to find it? Because itmight be there.

The Amazing Dr. Matrix

Every year or so during my tenure atScientific American, I would devote

a column to an imaginary interviewwith a numerologist I called Dr. IrvingJoshua Matrix (note the “666” provid-ed by the number of letters in his first,middle and last names). The good doc-tor would expound on the unusual prop-erties of numbers and on bizarre formsof wordplay. Many readers thought Dr.Matrix and his beautiful, half-Japanesedaughter, Iva Toshiyori, were real. I re-call a letter from a puzzled Japanesereader who told me that Toshiyori wasa most peculiar surname in Japan. I hadtaken it from a map of Tokyo. My in-formant said that in Japanese the wordmeans “street of old men.”

I regret that I never asked Dr. Matrixfor his opinion on the preposterous1997 best-seller The Bible Code, whichclaims to find predictions of the futurein the arrangement of Hebrew letters inthe Old Testament. The book employsa cipher system that would have madeDr. Matrix proud. By selectively apply-ing this system to certain blocks of text,inquisitive readers can find hidden pre-dictions not only in the Old Testamentbut also in the New Testament, the Ko-ran, the Wall Street Journal—and evenin the pages of The Bible Code itself.

The last time I heard from Dr. Matrix,he was in Hong Kong, investigating theaccidental appearance of π in well-known works of fiction. He cited, forexample, the following sentence frag-ment in chapter nine of book two of H. G. Wells’s The War of the Worlds:“For a time I stood regarding. . .” Theletters in the words give π to six digits!


The Author

MARTIN GARDNER wrote the “Mathematical Games” col-umn for Scientific American from 1956 to 1981 and continued tocontribute columns on an occasional basis for several years after-ward. These columns are collected in 15 books, ending with TheLast Recreations (Springer-Verlag, 1997). He is also the author ofThe Annotated Alice, The Whys of a Philosophical Scrivener, TheAmbidextrous Universe, Relativity Simply Explained and TheFlight of Peter Fromm, the last a novel. His more than 70 otherbooks are about science, mathematics, philosophy, literature andhis principal hobby, conjuring.

Further Reading

Recreations in the Theory of Numbers. Albert H. Beiler. DoverPublications, 1964.

Mathematics: Problem Solving through RecreationalMathematics. Bonnie Averbach and Orin Chein. W. H. Freemanand Company, 1986.

Mathematical Recreations and Essays. 13th edition. W. W.Rouse Ball and H.S.M. Coxeter. Dover Publications, 1987.

Penguin Edition of Curious and Interesting Geometry.David Wells. Penguin, 1991.

Mazes of the Mind. Clifford Pickover. St. Martin’s Press, 1992.

The Vanishing Area Paradox

Consider the figures shown below. Each pattern is made with the same 16pieces: four large right triangles, four small right triangles, four eight-sided

pieces and four small squares. In the pattern on the left, the pieces fit togethersnugly, but the pattern on the right has a square hole in its center! Where did thisextra bit of area come from? And why does it vanish in the pattern on the left?

The secret to this paradox—which I devised for the “Mathematical Games” col-umn in the May 1961 issue of Scientific American—will be revealed in the Lettersto the Editors section of next month’s issue. Impatient readers can find the an-swer at www.sciam.com on the World Wide Web. —M.G.

IAN

WO

RPO

LE

SA



1. Most people guess that the probability has risen from 1/3 to1/2. After all, only two cards are face down, and one must bethe ace. Actually, the probability remains 1/3. The probabilitythat you didn’t pick the ace remains 2/3, but Jones has elimi-nated some of the uncertainty by showing that one of thetwo unpicked cards is not the ace. So there is a 2/3 probabilitythat the other unpicked card is the ace. If Jones gives you theoption to change your bet to that card, you should take it (un-less he’s slipping cards up his sleeve, of course).

I introduced this problem in my October 1959 column in aslightly different form—instead of three cards, the probleminvolved three prisoners, one of whom had been pardonedby the governor. In 1990 Marilyn vos Savant, the author of apopular column in Parade magazine, presented still anotherversion of the same problem, involving three doors and a carbehind one of them. She gave the correct answer but re-ceived thousands of angry letters—many from mathemati-cians—accusing her of ignorance of probability theory! Thefracas generated a front-page story in the New York Times.

2. The sum is 111. The trick always works because the matrix ofnumbers is nothing more than an old-fashioned additiontable (below). The table is generated by two sets of numbers:(3, 1, 5, 2, 4, 0) and (25, 31, 13, 1, 7, 19). Each number in the ma-trix is the sum of a pair of numbers in the two sets. When you

choose the six circlednumbers, you are se-lecting six pairs thattogether include all12 of the generatingnumbers. So the sumof the circled num-bers is always equal tothe sum of the 12generating numbers.These special magicsquares were the sub-ject of my January1957 column.

3. Each chain of words ends on “God.” This answer may seemprovidential, but it is actually the result of the Kruskal Count, amathematical principle first noted by mathematician MartinKruskal in the 1970s. When the total number of words in a textis significantly greater than the number of letters in the long-

est word, it is likely that any twoarbitrarily started word chainswill intersect at a keyword. Af-ter that point, of course, thechains become identical. As thetext lengthens, the likelihoodof intersection increases.

I discussed Kruskal’s principle in my February 1978 column.Mathematician John Allen Paulos applies the principle to wordchains in his upcoming book Once upon a Number.

4. For simplicity’s sake, imagine a deck of only 10 cards, with theblack and red cards alternating like so: BRBRBRBRBR. Cuttingthis deck in half will produce two five-card decks: BRBRB andRBRBR. At the start of the shuffle, the bottom card of one deckis black, and the bottom card of the other deck is red. If thered card hits the table first, the bottom cards of both deckswill then be black, so the next card to fall will create a black-red pair on the table. And if the black card drops first, the bot-tom cards of both decks will be red, so the next card to fall willcreate a red-black pair. After the first two cards drop—no mat-ter which deck they came from—the situation will be thesame as it was in the beginning: the bottom cards of thedecks will be different colors. The process then repeats, guar-anteeing a black and red card in each successive pair, even ifsome of the cards stick together (below).

This phenomenon isknown as the Gilbreathprinciple after its discover-er, Norman Gilbreath, aCalifornia magician. I firstexplained it in my column

in August 1960 and discussed it again in July 1972. Magicianshave invented more than 100 card tricks based on this princi-ple and its generalizations. —M.G.

Answers to the Four Gardner Puzzles(The puzzles are on page 69.)

19

7

1

13

31

25

3 1 5 2 4 0

23

13

10 12 16

815 35 7

9 11 18

13 4 2 19

14 6 1 17

SKYSCRAPERcannot be built from Soma pieces.

(The puzzle is on page 71.)

THOROUGH SHUFFLE

OR

STICKY SHUFFLE

IAN

WO

RPO

LE

MAGIC HEXAGONhas a unique property: every straight

row of cells adds up to 38.


Earth may be the Ocean Planet,but most terrestrial creatures—

including humans—depend forfood on plants irrigated by freshwaterfrom rainfall, rivers, lakes, springs andstreams. None of the top five plants eat-en by people—wheat, corn, rice, pota-toes and soybeans—can tolerate salt: ex-pose them to seawater, and they droop,shrivel and die within days.

One of the most urgent global prob-lems is finding enough water and landto support the world’s food needs. TheUnited Nations Food and AgricultureOrganization estimates that an addition-al 200 million hectares (494.2 millionacres) of new cropland—an areathe size of Arizona, New Mex-ico, Utah, Colorado, Idaho,Wyoming and Montana


Irrigating Crops with SeawaterAs the world’s population grows and freshwater stores

become more precious, researchers are looking to the sea for the water to irrigate selected crops

by Edward P. Glenn, J. Jed Brown and James W. O’Leary

GLASSWORT SPECIES (here, Salicorniabigelovii) usually grow in coastal marsh-es. Because of their ability to flourish insaltwater, glasswort plants are the mostpromising crop to be grown so far usingseawater irrigation along coastal deserts.They can be eaten by livestock, and theirseeds yield a nutty-tasting oil.


combined—will be needed over the next30 years just to feed the burgeoning pop-ulations of the tropics and subtropics.Yet only 93 million hectares are avail-able in these nations for farms to ex-pand—and much of that land is forest-ed and should be preserved. Clearly, weneed alternative sources of water andland on which to grow crops.

With help from our colleagues, wehave tested the feasibility of seawateragriculture and have found that it workswell in the sandy soils of desert environ-ments. Seawater agriculture is defined asgrowing salt-tolerant crops on land us-ing water pumped from the ocean forirrigation. There is no shortage of sea-water: 97 percent of the water on earth

is in the oceans. Desert land is alsoplentiful: 43 percent of the earth’s to-tal land surface is arid or semiarid,but only a small fraction is closeenough to the sea to make seawater

farming feasible. We estimate that 15percent of undeveloped land in theworld’s coastal and inland salt desertscould be suitable for growing crops us-ing saltwater agriculture. This amountsto 130 million hectares of new crop-land that could be brought into humanor animal food production—withoutcutting down forests or diverting morescarce freshwater for use in agriculture.

Seawater agriculture is an old idea thatwas first taken seriously after WorldWar II. In 1949 ecologist Hugo Boykoand horticulturalist Elisabeth Boykowent to the Red Sea town of Eilat dur-ing the formation of the state of Israelto create landscaping that would attractsettlers. Lacking freshwater, the Boykosused a brackish well and seawaterpumped directly from the ocean andshowed that many plants would growbeyond their normal salinity limits insandy soil [see “Salt-Water Agriculture,”by Hugo Boyko; Scientific American,March 1967]. Although many of theBoykos’ ideas of how plants toleratesalts have not stood the test of time,their work stimulated widespread inter-est, including our own, in extending thesalinity constraints of traditional irri-gated agriculture.

Seawater agriculture must fulfill tworequirements to be cost-effective. First,it must produce useful crops at yieldshigh enough to justify the expense ofpumping irrigation water from the sea.Second, researchers must develop agro-nomic techniques for growing seawa-ter-irrigated crops in a sustainable man-ner—one that does not damage the en-

vironment. Clearing these hurdles hasproved a daunting task, but we havehad some success.

Salty Crops

The development of seawater agri-culture has taken two directions.

Some investigators have attempted tobreed salt tolerance into conventionalcrops, such as barley and wheat. For ex-ample, Emanuel Epstein’s research teamat the University of California at Davisshowed as early as 1979 that strains ofbarley propagated for generations in thepresence of low levels of salt could pro-duce small amounts of grain when irri-gated by comparatively saltier seawater.Unfortunately, subsequent efforts to in-crease the salt tolerance of conventionalcrops through selective breeding andgenetic engineering—in which genes forsalt tolerance were added directly to theplants—have not produced good candi-dates for seawater irrigation. The uppersalinity limit for the long-term irriga-tion of even the most salt-tolerant crops,such as the date palm, is still less thanfive parts per 1,000 (ppt)—less than 15percent of the salt content of seawater.Normal seawater is 35 ppt salt, but inwaters along coastal deserts, such asthe Red Sea, the northern Gulf of Cali-fornia (between the western coast ofSonora in Mexico and Baja California)and the Persian Gulf, it is usually closerto 40 ppt. (Sodium chloride, or tablesalt, is the most prevalent salt in seawa-ter and the one that is most harmful toplant growth.)

Our approach has been to domesti-cate wild, salt-tolerant plants, calledhalophytes, for use as food, forage andoilseed crops. We reasoned that chang-ing the basic physiology of a traditionalcrop plant from salt-sensitive to salt-tol-erant would be difficult and that it mightbe more feasible to domesticate a wild,salt-tolerant plant. After all, our moderncrops started out as wild plants. Indeed,some halophytes—such as grain from thesaltgrass Distichlis palmeri (Palmer’sgrass)—were eaten for generations bynative peoples, including the Cocopah,who live where the Colorado River emp-ties into the Gulf of California.

We began our seawater agriculture ef-forts by collecting several hundred halo-phytes from throughout the world andscreening them for salt tolerance andnutritional content in the laboratory.There are between 2,000 and 3,000 spe-cies of halophytes, from grasses and


RIC

HA

RD J

ON

ES


shrubs to trees such as mangroves; theyoccupy a wide range of habitats—fromwet, seacoast marshes to dry, inland sa-line deserts. In collaboration with DovPasternak’s research team at Ben Guri-on University of the Negev in Israel andethnobotanists Richard S. Felger andNicholas P. Yensen—who were then atthe University of Arizona—we foundroughly a dozen halophytes that showedsufficient promise to be grown underagronomic conditions in field trials.

In 1978 we began trials of the mostpromising plants in the coastal desert atPuerto Peñasco, on the western coast ofMexico. We irrigated the plants daily byflooding the fields with high-saline (40ppt) seawater from the Gulf of Califor-nia. Because the rainfall at Puerto Peñas-co averages only 90 millimeters a year—

and we flooded our plots with an annu-al total depth of 20 meters or more ofseawater—we were certain the plantswere growing almost solely on seawater.(We calculate rainfall and irrigation ac-cording to the depth in meters that fallson the fields rather than in cubic me-ters, which is a measure of volume.)

Although the yields varied amongspecies, the most productive halophytesproduced between one and two kilo-grams per square meter of dry biomass—roughly the yield of alfalfa grown usingfreshwater irrigation. Some of the mostproductive and salt-tolerant halophyteswere shrubby species of Salicornia (glass-wort), Suaeda (sea blite) and Atriplex(saltbush) from the family Chenopodi-aceae, which contains about 20 percent

of all halophyte species. Salt grasses suchas Distichlis and viny, succulent-leavedground covers such as Batis were alsohighly productive. (These plants are notChenopodiaceae, though; they are mem-bers of the Poaceae and Batidaceae fam-ilies, respectively.)

But to fulfill the first cost-effectivenessrequirement for seawater agriculture,we had to show that halophytes couldreplace conventional crops for a specificuse. Accordingly, we tested whetherhalophytes could be used to feed live-stock. Finding enough forage for cattle,sheep and goat herds is one of the mostchallenging agricultural problems in theworld’s drylands, 46 percent of whichhave been degraded through overgraz-ing, according to the U.N. EnvironmentProgram. Many halophytes have highlevels of protein and digestible carbohy-drates. Unfortunately, the plants alsocontain large amounts of salt; accumu-lating salt is one of the ways they adjustto a saline environment [see illustrationon page 80]. Because salt has no caloriesyet takes up space, the high salt contentof halophytes dilutes their nutritionalvalue. The high salinity of halophytesalso limits the amount an animal can eat.In open grazing situations, halophytesare usually considered “reserve-browseplants,” to which animals turn onlywhen more palatable plants are gone.

Our strategy was to incorporate halo-phytes as part of a mixed diet for live-stock, replacing conventional hay foragewith halophytes to make up between30 and 50 percent of the total food in-

take of sheep and goats. (These percent-ages are the typical forage levels used infattening animals for slaughter.) Wefound that animals fed diets containingSalicornia, Suaeda and Atriplex gainedas much weight as those whose diets in-cluded hay. Moreover, the quality of thetest animals’ meat was unaffected bytheir eating a diet rich in halophytes.Contrary to our initial fears, the animalshad no aversion to eating halophytes inmixed diets; they actually seemed to beattracted by the salty taste. But the ani-mals that ate a halophyte-rich diet drankmore water than those that ate hay, tocompensate for the extra salt intake. Inaddition, the feed conversion ratio of thetest animals (the amount of meat theyproduced per kilogram of feed) was 10percent lower than that of animals eat-ing a traditional diet.

Farming for Oil

The most promising halophyte wehave found thus far is Salicornia

bigelovii. It is a leafless, succulent, annu-al salt-marsh plant that colonizes newareas of mud flat through prolific seedproduction. The seeds contain high lev-els of oil (30 percent) and protein (35percent), much like soybeans and otheroilseed crops, and the salt content is lessthan 3 percent. The oil is highly poly-unsaturated and similar to safflower oilin fatty-acid composition. It can be ex-tracted from the seed and refined usingconventional oilseed equipment; it isalso edible, with a pleasant, nutlike taste

Irrigating Crops with Seawater78 Scientific American August 1998

SEAWATER AGRICULTURE can require different agronomictechniques than freshwater agriculture. To grow saltbush, orAtriplex—a salt-tolerant plant that can be used to feed livestock—

seawater farmers must flood their fields frequently (left). In ad-

dition, irrigation booms (center) must be lined with plastic pipingto protect them from rusting when in contact with the salty wa-ter. But some techniques can remain the same: standard com-bines are used to harvest Salicornia seeds (right), for example.

UN

IVER

SITY

OF

ARI

ZON

A

DA

N M

URP

HY


and a texture similar to olive oil. A smalldrawback is that the seed contains sapo-nins, bitter compounds that make theraw seeds inedible. These do not contam-inate the oil, but they can remain in themeal after oil extraction. The saponinsthus restrict the amount of meal thatcan be used in chicken diets, but feedingtrials have shown that Salicornia seedmeal can replace conventional seedmeals at the levels normally used as aprotein supplement in livestock diets.Hence, every part of the plant is usable.

We have participated in building sev-eral prototype Salicornia farms of up to250 hectares in Mexico, the United ArabEmirates, Saudi Arabia and India. Dur-ing six years of field trials in Mexico,Salicornia produced an average annualcrop of 1.7 kilograms per square meterof total biomass and 0.2 kilogram persquare meter of oilseed. These yieldsequal or exceed the yields of soybeans

and other oilseeds grown using freshwa-ter irrigation. We have also shown thatnormal farm and irrigation equipmentcan be modified so that it is protectedfrom salt damage from the seawater. Al-though the irrigation strategies for han-dling seawater are different from thoseused for freshwater crops, we have notencountered any insurmountable engi-neering problems in scaling up from fieldtests to prototype farms.

Normally, crops are irrigated onlywhen the soil dries to about 50 percentof its field capacity, the amount of wa-ter it is capable of holding. In addition,in freshwater irrigation, farmers addonly enough water to replace what theplants have used. In contrast, seawaterirrigation requires copious and fre-quent—even daily—irrigation to pre-vent salt from building up in the rootzone to a level that inhibits growth.

Our first field trials used far more wa-ter (20 meters a year) than could be ap-plied economically, so in 1992 we be-gan experiments to determine the mini-mum amount of seawater irrigationneeded to produce a good yield. Wegrew test plantings of Salicornia in soilboxes buried within open, irrigated plotsof the same crop during two years offield trials. The boxes, called lysimeters,had bottom drains that conveyed excesswater to several collection points outsidethe plots, allowing us to measure thevolume and salinity of the drain water.Using them, we calculated the waterand salt balances required for a seawa-ter-irrigated crop for the first time. Wefound that the amount of biomass a

seawater-irrigated crop yields dependson the amount of seawater used. Al-though Salicornia can thrive when thesalinity of the water bathing its roots ex-ceeds 100 ppt—roughly three times thenormal saltiness of the ocean—it needsapproximately 35 percent more irriga-tion when grown using seawater thanconventional crops grown using fresh-water. Salicornia requires this extra wa-ter because as it selectively absorbs wa-ter from the seawater, it quickly rendersthe remaining seawater too salty for use.

Making It Pay

Can seawater agriculture be econom-ical? The greatest expense in irri-

gated agriculture is in pumping the wa-ter. The pumping costs are directly pro-portional to the amount of waterpumped and the height to which it islifted. Although halophytes require morewater than conventional crops, seawa-ter farms near sea level require less wa-ter lifting than conventional farms,which often lift water from wells deep-er than 100 meters. Because pumpingseawater at sea level is cheaper thanpumping freshwater from wells, seawa-ter agriculture should be cost-effectivein desert regions—even though its yieldsare smaller than traditional, freshwateragriculture.

Seawater irrigation does not requirespecial equipment. The large test farmswe have helped build have used eitherflood irrigation of large basins or mov-ing-boom sprinkler irrigation. Movingbooms are used in many types of crop

Irrigating Crops with Seawater Scientific American August 1998 79

200

150

100

50

0

Saltbush

Sea blite

Glasswort

DAILY WEIGHT

GAIN

PERC

ENT

OF

CO

NTR

OL

SHEE

P

USDA MEAT

QUALITYFEED

EFFICIENCYWATER INTAKE

2.0

1.5

1.0

0.5

0

ALF

ALF

A H

AY

QU

AIL

BU

SH

PIC

KLE

WEE

D

SEA

BLI

TE

FOU

R-W

ING

SA

LTB

USH

GLA

SSW

ORT

SEA

PU

RSLA

NE

PALM

ER’S

GRA

SS

SUD

AN

GRA

SS H

AY

CO

AST

SA

LTB

USHBIO

MA

SS Y

IELD

(K

ILO

GR

AM

S P

ER M

ETER

SQ

UA

RED

PER

YEA

R)

SLIM

FIL

MS

DA

N M

URP

HY

YIELDS of salt-tolerant crops grown using seawater agricultureare comparable to those of two freshwater-irrigated plants oftenused for livestock forage: alfalfa hay and Sudan grass hay (left,blue bars). Sheep raised on a diet supplemented with salt-tolerant

plants such as saltbush, sea blite and glasswort gain at least asmuch weight and yield meat of the same quality as control sheepfed conventional grass hay, although they convert less of thefeed to meat and must drink almost twice as much water (right).


production. For seawater use, a plasticpipe is inserted in the boom so the sea-water does not contact metal. Salicorniaseeds have also been successfully har-vested using ordinary combines set tomaximize retention of the very smallseeds, which are only roughly one milli-gram in weight.

Yet Salicornia, our top success storyso far, is not a perfect crop. The plantstend to lodge (lie flat in the field) as har-vest approaches, and the seeds may shat-ter (release before harvest). In addition,seed recoveries are only about 75 per-cent for Salicornia, compared withgreater than 90 percent for most crops.Further, to support high seed yields Sal-icornia must grow for approximately100 days at cool temperatures beforeflowering. Currently production of thiscrop is restricted to the subtropics,which have cool winters and hot sum-mers; however, some of the largest ar-eas of coastal desert in the world are inthe comparatively hotter tropics.

The second cost-effectiveness require-ment of seawater agriculture is sustain-ability over the long term. But sustain-ability is not a problem limited to irri-

gation using seawater: in fact, many ir-rigation projects that use freshwatercannot pass the sustainability test. Inarid regions, freshwater irrigation is of-ten practiced in inland basins with re-stricted drainage, resulting in the build-up of salt in the water tables underneaththe fields. Between 20 and 24 percent ofthe world’s freshwater-irrigated landssuffer from salt and water buildup in theroot zone. When the problem becomessevere, farmers must install expensivesubsurface drainage systems; disposingof the collected drain water creates ad-ditional problems. In California’s SanJoaquin Valley, for example, wastewaterthat had drained into a wetland causeddeath and deformity in waterfowl be-cause of the toxic effects of selenium,an element that typically occurs in manywestern U.S. soils but had built up tohigh concentrations in the drain water.

Seawater agriculture is not necessarilyexempt from such problems, but it doesoffer some advantages. First, coastal des-ert farms on sandy soils generally haveunimpeded drainage back to the sea.We have continuously irrigated the samefields with seawater for over 10 years

with no buildup of water or salts in theroot zone. Second, coastal and inlandsalt desert aquifers often already haveelevated concentrations of salt and soshould not be damaged by seawater.Third, the salt-affected soils that wepropose for seawater agriculture are of-ten barren—or nearly so—to start with,so installing a seawater farm may havefar less effect on sensitive ecosystemsthan conventional agriculture does.

No farming activity is completely benign, however. Large-scale coastalshrimp farms, for example, have causedalgal blooms and disease problems inrivers or bays that receive their nutri-ent-rich effluent [see “Shrimp Aquacul-ture and the Environment,” by ClaudeE. Boyd and Jason W. Clay; ScientificAmerican, June]. A similar problemcan be anticipated from large-scale halo-phyte farms, caused by the large vol-ume of high-salt drainage water con-taining unused fertilizer, which will ulti-mately be discharged back to the sea.On the other hand, seawater farms canalso be part of a solution to this prob-lem if shrimp-farm effluent is recycledonto a halophyte farm instead of dis-

RIC

HA

RD J

ON

ES

Anatomy of a Halophyte

H2O VAPOR

H2OVAPOR

H2OVAPOR

H2O

Na+Cl–

Na+

Cl–

LEAF CELL

TURGORPRESSURE

NaCl

SALT BLADDERS

H2O Na+Cl–

H2O

ATRIPLEX (SALTBUSH)

Some salt-tolerant plants, or halophytes, have evolved mechanisms at the root, leaf andcell levels for thriving in the presence of seawater.The cells that make up the outer layer, or epidermis, of each rootlet are nearly imperviousto salt (NaCl). In addition, the inner layer, or

endodermis, has a waxy layer between each cell

that forces water to passthrough the cells, whichfilter out more salt.

H2O

Na+Cl–

WAXY LAYER

LEAF CROSSSECTION

ROOTLET


charged directly to the sea: thehalophyte crop would recovermany of the nutrients in the efflu-ent and reduce the volume. Thefirst halophyte test farm we builtin Mexico was installed to recycleshrimp-farm effluent, and furtherresearch linking marine aquacul-ture effluent with halophyte farmsis under way.

Halophyte farms have also beenproposed as a way to recycle theselenium-rich agricultural drainwater generated in the San Joa-quin Valley of California. Seleni-um is an essential nutrient at low levelsbut becomes toxic at high levels. Halo-phytes grown on drain water in the val-ley take up enough selenium to makethem useful as animal-feed supplements

but not enough to make them toxic.Will seawater agriculture ever be prac-

ticed on a large scale? Our goal in thelate 1970s was to establish the feasibili-ty of seawater agriculture; we expectedto see commercial farming within 10years. Twenty years later seawater agri-culture is still at the prototype stage ofcommercial development. Several com-panies have established halophyte testfarms of Salicornia or Atriplex in Cali-

fornia, Mexico, Saudi Arabia, Egypt,Pakistan and India; however, to ourknowledge, none have entered large-scale production. Our research experi-ence convinces us of the feasibility ofseawater agriculture. Whether the worldultimately turns to this alternative willdepend on future food needs, economicsand the extent to which freshwaterecosystems are withheld from furtheragricultural development.


SLIM

FIL

MS

SA

Some halophytes, such as this Atriplex (saltbush),have specialized cells called salt bladders on their leaves for storing excess NaCl. When full, the salt bladders burst, releasing salt in a silvery layer that reflects light and cools the leaf. Water vapor escapes throughpores on the undersides of the leaves.

Cells inside each leaf are speciallyequipped to handle any salt that is absorbed by the plant. The central vacuole, or storage area, of each cellbears molecules that specifically importsodium ions (Na+), and chloride ions (Cl–) follow. The high concentration ofNa+ and Cl– attracts water, maintainingthe turgor pressure of the cell.

SUNLIGHT

The Authors

EDWARD P. GLENN, J. JED BROWN and JAMES W. O’LEARY have a combined total of45 years of experience studying the feasibility of seawater agriculture in desert environ-ments. Glenn began his research career as a self-described “marine agronomist” in 1978 af-ter receiving his Ph.D. from the University of Hawaii; he is now a professor in the depart-ment of soil, water and environmental science at the University of Arizona at Tucson. Brownreceived his Ph.D. from the University of Arizona’s Wildlife and Fisheries Program in May.O’Leary is a professor in the University of Arizona’s department of plant sciences. He re-ceived his Ph.D. from Duke University in 1963. Author of more than 60 publications onplant sciences, O’Leary served in 1990 on a National Research Council panel that examinedthe prospects of seawater agriculture for developing countries.

Further Reading

Saline Culture of Crops: A Genetic Approach. Emanuel Epstein et al. in Science, Vol.210, pages 399–404; October 24, 1980.

Saline Agriculture: Salt Tolerant Plants for Developing Countries. NationalAcademy Press, 1990.

SALICORNIA BIGELOVII Torr.: An Oilseed Halophyte for Seawater Irrigation. E. P.Glenn, J. W. O’Leary, M. C. Watson, T. L. Thompson and R. O. Kuehl in Science, Vol.251, pages 1065–1067; March 1, 1991.

Towards the Rational Use of High Salinity Tolerant Plants. H. Lieth and A. A. AlMasoom. Series: Tasks for Vegetation Science, Vol. 28. Kluwer Academic Publishers,1993.

Halophytes. E. P. Glenn in Encyclopedia of Environmental Biology. Academic Press, 1995.

LOCATIONS in coastal deserts andinland salt deserts (green areas)could be used for seawater agricul-ture—or agriculture using irrigationfrom salty underground aquifers—

to grow a variety of salt-tolerantcrops for food or animal forage.


Microdiamonds82 Scientific American August 1998

Diamond has been a highly valued min-

eral for 3,000 years. In ancient India

the gemstone was thought to possess

magical powers conferred by the gods. Medieval

European nobles sometimes wore diamond rings

into battle, thinking the jewelry would bring

them courage and make them fearless.

More recently, diamond has proved to be the

ideal material for a number of industrial uses.

The mineral, an incredibly pure composition of

more than 99 percent carbon, is the hardest sub-

stance known. It is capable of scratching almost

anything, making it suitable for use in abrasives

and in cutting, grinding and polishing tools. Di-

amond also has high thermal conductivity—more

than three times that of copper—and is thus op-

timal for spreading and dissipating heat in elec-

tronic devices such as semiconductor lasers. Be-

cause most of these applications can be accom-

plished with tiny crystals, both scientific and

technological interest have begun to focus on

microdiamonds, samples that measure less than

half a millimeter in any dimension.

Previously, microdiamonds were ignored,

mainly because of their minimal size and the

complex and expensive procedure required to

mine the tiny particles. But as extraction tech-

niques improve and industrial demand grows,

microdiamonds are becoming an increasingly at-

tractive source of the mineral.

Mysterious Origins

After decades of research, scientists now be-

lieve that commercial-size stones, or mac-

rodiamonds, form in the earth’s mantle and are

transported to the surface by volcanoes. The ori-

gin of microdiamonds, however, remains a mys-

tery. They may simply be younger crystals that

had little time to grow before being brought to

the earth’s surface. Or they may have formed in

an environment where carbon was limited, there-

by stunting their growth. Some researchers be-

lieve, however, that microdiamonds are pro-

duced by different, though related, processes.

Microdiamonds are occasionally found by

themselves, in the absence of bigger crystals. Of-

ten, though, they occur along with macrodia-

monds. Surprisingly, microdiamonds have even

been discovered in the same deposits with larger

samples that have experienced resorption—a

broad term encompassing any process, such as

dissolution or corrosion, that reduces the size of

a diamond, frequently resulting in the rounding

of corners or edges. Because the ratio of surface

area to volume is much higher for micro speci-

mens than for macro sizes, scientists would have

expected the resorption that reduced the size of

the macrodiamonds to have consumed any

small samples.

This paradoxical evidence suggests that mi-

crodiamonds might have origins distinct from

those of macrodiamonds. One hypothesis as-

serts that some microdiamonds form within as-

cending magma, where distinct growth condi-

tions and processes limit the size to which the

carbon crystals can grow. But an alternative the-

ory contends that the magma is merely the deliv-

ery mechanism and that microdiamonds, like

macrodiamonds, are products of the mantle.

Complicating this issue, researchers have dis-

covered microdiamonds that they believe were

formed in the earth’s crust from the collision of

tectonic plates. And tiny diamonds have also

been found in meteorites.

With continuing work, scientists may one day

resolve how microdiamonds are related to their

larger siblings. This knowledge will lead to a

greater understanding of the carbon substance,

perhaps resulting in more effective techniques

for mining the mineral economically and for

growing it synthetically.

MicrodiamondsThese tiny, enigmatic crystals hold promise both

for industry and for the study of how diamond grows

by Rachael Trautman, Brendan J. Griffin and David Scharf


Microdiamonds Scientific American August 1998 83

MICRODIAMOND is smaller than half a millimeter (500 microns) in any dimension, corresponding to a mass of less than a hundredth of a carat.

(In comparison, the famous Hope diamond has a mass of 45 carats.) This particular specimen is about 400 microns tall.

ALL

MIC

RO

GR

AP

HS

BY

DA

VID

SC

HA

RF


Mineralogists believe that a diamond gets

larger by adding new atoms to its surface,

with the speed of growth influencing the shape of

the crystal. If the temperature, pressure and oxygen

content are favorable, then the growth rate—and

final crystal form—will depend on how much car-

bon is available.

Microdiamonds occur as single crystals of octa-

hedrons, dodecahedroids, cubes, macles (twinned,

flat triangular plates) and irregular lumps. In addi-

tion, octahedral and cubic faces can sometimes de-

velop concurrently in the same crystal, resulting in

a cubo-octahedral form. Frequently, two or more

microdiamonds will fuse during their growth, lead-

ing to unusual and often striking configurations

such as aggregates of octahedrons, dodecahedroids

or cubes; and amorphous clusters.

Octahedrons (left) are the most common growth

form. In fact, in most microdiamond populations,

they usually make up more than half the observed

crystal shapes.

Less frequently found are cubes (center), which

rarely have a crystalline or ordered atomic structure.

Instead many cubic microdiamonds have a fibrous

composition with radial growth perpendicular to

the cubic faces. This distinctive form of growth, as

indicated in the textured faces, is the result of su-

persaturation of carbon in the source medium,

which leads to extremely rapid crystallization.

Aggregates (right) also imply an environment

high in carbon, with growth occurring simultane-

ously at several neighboring points. In such prox-

imity, the crystals may meld to form a single unit of

complex intergrowth.

84 Scientific American August 1998 Microdiamonds

LIQUID

SOLID

CRUST

MANTLE

CORE

OCTAHEDRON CUBE

After much research, scientists have concluded that diamond forms in the earth’s mantle (left), at depths of 150 to 200 kilometers (90 to 120 miles). The valuable

mineral is then carried upward by kimberlites and lamproites, special types of igneousrock that began as molten material. At the earth’s surface (near right), the rock is foundas “pipes,” or carrot-shaped deposits, predominantly within very old landmasses knownas cratons. Such deposits may produce as little as one carat per 100 metric tons of rockmined. Interestingly, some researchers believe that microdiamonds—unlike macrodia-monds—do not form in the mantle but rather in the kimberlitic and lamproitic magma.

Whatever the case, tremendous pressure is required for diamond to form. At a tem-perature of 1,600 kelvins (2,400 degrees Fahrenheit), a pressure of about 50,000 atmo-spheres is needed (center right)—roughly equivalent to the pressure of an upside-downEiffel Tower (all 7,000 metric tons of it) balanced on a 12-centimeter-square plate (farright). Without that extreme compression, the carbon may instead form graphite.

The Origin of Diamond

A Variety of Shapes

ILLU

STR

ATI

ON

S B

Y L

AU

RIE

GR

AC

E


BARREN PIPE

DIAMOND-BEARING

KIMBERLITE PIPE

DIAMOND-BEARING

LAMPROITE PIPE

GRAPHITE

DIAMONDGRAPHITE

DIAMOND

CRATON

DIAMONDDEPOSIT

90,000

70,000

50,000

30,000

10,000

DIAMOND

GRAPHITE

TEMPERATURE (KELVINS)

0 1,000 2,000 3,000

PRES

SURE

(ATM

OSP

HER

ES)

AGGREGATE

MicrodiamondsCopyright 1998 Scientific American, Inc.

Microdiamonds86 Scientific American August 1998

“Seeing” the Microscopic World

The microdiamond images presented here were recorded with a scanning electron micro-scope (SEM), by micrographer David Scharf (right). The SEM works by illuminating a subject

with electrons rather than light. Magnetic lenses focus a beam of electrons that is scanned acrossthe specimen in a raster pattern (similar to that used in television screens). When the beam hitsthe specimen, it knocks other (secondary) electrons from the surface of the object. These nega-tively charged particles are then attracted to the positive polarity of a detector, where they areamplified and converted into a video signal.

The microdiamonds appear to be opaque in the images because the crystals are not transpar-ent to electrons. The color comes from a system developed by Scharf—the Multi-Detector ColorSynthesizer (U.S. Patent No. 5,212,383)—which uses three electron detectors (instead of the usu-al one) spaced apart to capture three distinct illumination angles. A different (arbitrary) color isassigned to each detector to produce the various colors seen.

Resorption processes typically alter the shape of a mi-

crodiamond. Dissolution, for example, converts

plane crystal faces into curved or rounded surfaces. Specif-

ically, microdiamonds do not grow as dodecahedroids

(left); the 12-sided shape is instead the end product of oc-

tahedral dissolution. With sharp-edged octahedrons at one

end of the spectrum and rounded dodecahedroids at the

other, microdiamonds have exhibited a complete grada-

tion between those two extremes. Because dodecahedroids

are almost always found in microdiamond populations,

dissolution is believed to be common.

When microdiamonds are dissolved in the presence of a

corrosive agent, strongly developed etching or unusually

rough surfaces may develop. The erosion can often be ex-

treme, producing diamond morphologies with almost no

resemblance to their original forms, such as the transfor-

mation of the smooth, planar faces of an octahedral aggre-

gate into etch-sculptured surfaces (center).Sometimes the corrosion will result in faces with distinc-

tive structures, such as those resembling steps. Particularly

at the corners and edges, the etching can be severe enough

to alter a crystal’s overall shape (right).

DODECAHEDROID formed by octahedral dissolution OCTAHEDRAL AGGREGATE etched by corrosion

David Scharf

Changing Faces

KEV

IN E

LLSW

OR

TH


Microdiamonds Scientific American August 1998 87

Further Reading

DIAMONDS. Second edition. Eric Bruton. NAG Press, London,1978.

DIAMOND. Gordon Davies. Adam Hilger, Bristol, 1984.A COMPARISON OF THE MICRODIAMONDS FROM KIMBERLITE

AND LAMPROITE OF YAKUTIA AND AUSTRALIA. R. L. Trautman,B. J. Griffin, W. R. Taylor, Z. V. Spetsius, C. B. Smith and D. C. Lee in Proceedings of the Sixth International KimberliteConference, Vol. 2: Russian Geology and Geophysics, No. 38.Allerton Press, 1997.

THE NATURE OF DIAMONDS. Edited by George E. Harlow.Cambridge University Press, 1998.

The Authors

RACHAEL TRAUTMAN, BRENDAN J. GRIFFIN and DAVIDSCHARF met through their interest in microdiamonds and electron mi-croscopy. Trautman is a postgraduate with the Center for Strategic Miner-al Deposits in the department of geology and geophysics at the Universityof Western Australia, where she is studying the origin and genesis of micro-diamonds. Griffin is a senior lecturer at the Center for Microscopy and Mi-croanalysis at the University of Western Australia; his research interests in-clude diamond genesis, mineral microanalysis and new techniques in scan-ning electron microscopy. Scharf is a professional scanning electronmicrographer whose work has appeared in encyclopedias, museums andmagazines such as Scientific American.

OCTAHEDRON etched into near cubic form


The Philadelphia Yellow FeverEpidemic of 1793

One of the first major epidemics of the disease in the U.S., it devastated America’s early capital. It also had lasting

repercussions for the city and country

by Kenneth R. Foster, Mary F. Jenkins and Anna Coxe Toogood


Epidemics that quickly wipe outlarge segments of communitiesare rare in the U.S. these days.

But before the 20th century, such disas-ters occurred quite frequently, ravagingbewildered populations who usually hadlittle understanding of the causes. Apartfrom the human tragedies these episodesproduced, some had lasting political ef-fects on the nation. A dramatic examplearose in 1793, when one of the earliestsignificant epidemics of yellow fever inthe U.S. struck Philadelphia, then thenation’s capital.

At the time, Philadelphia was thecountry’s largest and most cosmopoli-tan city. Yet prominence and prosperity

offered little protection. Over the sum-mer and fall, roughly a tenth of its pop-ulation, some 5,000 people, perished.

The trouble began soon after Frenchrefugees from a bloody slave rebellion inSaint Domingue (now Haiti) landed onthe banks of the Delaware River, whichbounds the east side of the city. In addi-tion to news of the fighting, the refugeesrelayed tales of a mysterious pestilentialfever decimating several islands in theWest Indies. In July the same diseasebroke out in Philadelphia.

The first to fall were working-classpeople living along the Delaware. Theysuffered high fevers and hemorrhages;their eyes and skin turned yellow; andthey brought up black vomit. Manydied from internal bleeding within daysof becoming ill.

These early victims escaped notice byleaders of the medical establishment,perhaps because of their location (bothgeographic and social) on the marginsof the city. By August 19, however, aprominent physician named BenjaminRush had seen several similar cases andconcluded that the patients sufferedfrom “bilious remitting yellow fever.”Rush, then in his late 40s, had encoun-tered this malady once before, during hisapprenticeship. He immediately warnedthat an epidemic was under way andtook steps to combat it.

Rush stood in good position to influ-ence the city’s response to the disaster.He was a renowned professor at theUniversity of Pennsylvania and a found-er of the prestigious College of Physi-cians of Philadelphia. He was famousas a patriot and a signer of the Declara-tion of Independence, in addition to be-ing a philanthropist and teacher. Hewas also possessed of enormous energyand a forceful personality.

At Rush’s urging, Philadelphia’s may-or, Matthew Clarkson, asked the Col-lege of Physicians to recommend publichealth measures against the fever. It is-sued its report, drafted by Rush, on Au-

gust 26. Following theprevailing theory that thedisease was contagiousand spread by putrid vapors, the col-lege told citizens to avoid people strick-en with the disease, to breathe throughcloths soaked with camphor or vinegarand to burn gunpowder to clear the air.It also proposed establishing a hospitalto treat indigent victims, who were toopoor to pay for the home care mostpeople preferred. To avoid alarming thepublic further, the college called for thesilencing of church bells, which hadbeen ringing incessantly to announcethe many funerals in the city.

Yellow fever is now known to be a vi-ral infection transmitted by female mos-quitoes of the species Aedes aegypti. Ev-idently, Philadelphia’s troubles stemmedfrom virus-carrying A. aegypti stow-aways that arrived by ship with the ref-ugees from Saint Domingue. After theinfected insects spread the virus to peo-ple, uninfected A. aegypti picked up theviral agent from the blood of diseasedpatients and injected it into new victims.

Nevertheless, concern that yellow fe-ver was spread through the air was rea-sonable in the 18th century, given thestate of science at the time. Physiciansknew nothing about microorganismsand the transmission of disease, and notuntil 1900 did Walter Reed and his col-leagues finally prove the proposal, orig-inally put forward by Cuban physicianCarlos Finlay, that yellow fever wastransmitted to humans by mosquitoes.Accordingly, many doctors, includingRush, clung to the medieval theory thatdiseases were caused by impurities inthe air, particularly vapors from decay-ing vegetable matter.

Proponents of environmental theoriesof illness had a lot to worry about inPhiladelphia. The city lacked an efficientsewage system, and outhouses and in-dustrial waste polluted the water sup-ply. Noxious fumes filled the air fromtanneries, distilleries, soap manufactur-ers and other industries. Animal carcass-es lay rotting on the banks of the Dela-ware and in public streets, particularlyin the market along High Street (nowcalled Market Street). Rush himself at-tributed the disease to vapors from ashipload of coffee that had spoiled enroute from the West Indies and lay aban-doned and rotting on a wharf. At hisurging, Mayor Clarkson republished thecity’s unenforced laws mandating twice-weekly trash collection and ordered a


YELLOW FEVER VICTIM is helped to acarriage collecting the dead and dying inan undated woodcut recalling Philadel-phia’s 1793 epidemic. His aide is a busi-nessman, Stephen Girard. In the back-ground, a man covers his mouth to avoidcontracting the disease, then thought to bespread through the air. At the top of thispage is Aedes aegypti, the mosquito nowknown to transmit the yellow fever virus.

CO

RBIS

-BET

TMA

NN

HA

RRY

RO

GER

S Ph

oto

Rese

arch

ers,

Inc.


cleanup of the city’s streets and markets.Other doctors, notably William Cur-

rie, believed sick refugees from SaintDomingue brought the fever and spreadit by physical contact. Those physiciansargued for better quarantine regulations.The mayor tried to comply by orderingthat arriving passengers and goods bekept isolated for two to three weeks,but during most of the epidemic the citylacked the ability to enforce the order.Even Currie, though, suggested that acombination of very dry weather thatsummer and putrid vapors had nurturedthe disease by destroying the city’s “oxy-gen gas or pure air.”

Panic Starts

In correspondence, Rush estimatedthat 325 people died in August alone.

By the end of that month, many peoplewho could afford to leave Philadelphiadid so, including some of the city’s 80doctors. (Among the professionals whostayed was John Todd, a lawyer whohandled the surge of legal matters aris-ing from the many deaths. Todd caughtthe fever and died. Within a year, hiswidow, Dolley, married CongressmanJames Madison. Madison later becamethe fourth U.S. president; his wife be-came the First Lady, Dolley Madison.)

The residents who re-mained—mostly the poor—fol-lowed the advice of the collegeand applied other tactics to

protect themselves. They wore cam-phor bags or tarred rope around theirnecks, stuffed garlic into their pocketsand shoes, doused themselves with vin-egar, shot off guns in their living roomsand lit fires in the streets.

Soon people began abandoning theirdying spouses in the streets. Hungry andfrightened children roamed the city, theirparents dead from the fever. The HighStreet market stood empty; other com-mercial activity closed down; and thechurches and Quaker meeting houseslost their congregations. Pennsylvaniagovernor Thomas Mifflin departed onSeptember 6, leaving the crisis in May-or Clarkson’s hands. Within a week ofMifflin’s exit, nearly all in authority hadfled, and the economy had collapsed.

Among Clarkson’s top priorities be-fore and after the exodus was providingcare for the stricken, although his choic-es were limited by public fear of conta-gion. By late August many institutions—among them the Pennsylvania Hospital,the city’s almshouse and the Quakeralmshouse—had stopped admitting yel-low fever patients, to protect currentresidents. In response the Overseers andGuardians of the Poor (a city agency)tried to appropriate for a hospital JohnRickett’s circus on the outskirts oftown, but angry neighbors threatened

to burn the structure down.On August 31 the agencyturned to Bush Hill, an unoc-cupied building tenanted just

a year earlier by Vice President JohnAdams and his wife. Bush Hill stood amile from the center of town, too dis-tant to elicit resistance from the city res-idents but also too far for beleaguereddoctors to visit regularly.

At first, conditions at the makeshifthospital were abysmal. As one observernoted, “It exhibited as wretched a pic-ture of human misery as ever existed.

A profligate, abandoned set of nurs-es and attendants. . .rioted on theprovisions and comforts preparedfor the sick, who.. . were left almostentirely destitute of every assistance.

The dying and dead were indiscrimi-nately mingled together. The ordureand other evacuations of the sick wereallowed to remain in the most offensivestate imaginable.”

On September 10 Mayor Clarksonappealed for volunteers to organize aresponse to the hospital’s and the city’sdifficulties in caring for the sick. Abouta dozen men then worked for 46 consec-utive days to make the needed arrange-ments. They borrowed money, boughthospital supplies and other items neededby the poor, rented quarters for an or-phanage and hired a staff. And theywent door-to-door to check for the deadand dying and to rescue orphaned chil-dren. It was an incredible and brave ef-fort in the face of a disease that mostdoctors believed could be contractedeasily by breathing air or by personalcontact. Three of those dedicated vol-unteers eventually died of the fever.

Many African-Americans risked theirlives as well. Shortly before the mayor’sappeal, Rush had pleaded for nursingassistance from the city’s African-Amer-ican community. Most of its 2,500 mem-bers were free (albeit living in poverty)because Pennsylvania was phasing outslavery. Six years earlier two communi-ty leaders, Absalom Jones and RichardAllen, had formed the Free African So-ciety, the first self-help group in the U.S.organized by blacks. Rush, an outspo-ken abolitionist, had close associationswith both men and was a strong sup-porter of the society. From accounts hehad read, Rush believed blacks wereimmune to yellow fever and thereforein a position to help the sick. (In theend, he was proved wrong; more than300 African-Americans would die inthe epidemic, in a proportion close tothat of the rest of the population.)

The society agreed to provide work-ers. Jones, Allen and society memberWilliam Gray organized the hiring and

The Philadelphia Yellow Fever Epidemic of 1793

..

BENJAMIN RUSH, a leading physician, fomented controversy during the epidemic bystrongly advocating aggressive bloodletting as a “cure” for yellow fever. The instrumentsshown, from about 1790, are types Rush and his followers might have used. Rush’sopponents took a gentler approach, involving rest, cleanliness, wine and Peruvian bark.

CH

ARL

ES W

ILLS

ON

PEA

LE In

depe

nden

ce N

atio

nal H

isto

rica

l Par

k (le

ft);

BRI

DG

EMA

N A

RT

LIB

RA

RY (r

ight

)



dispatching of nurses to homes. Mem-bers of the black community alsobrought people to the hospital and bur-ied the dead. At the society’s request,the mayor released black convicts fromthe Walnut Street prison to work in the“contagious hospital.”

On September 18 Rush wrote to hiswife (who had left the city with thecouple’s children), “Parents desert theirchildren as soon as they are infected, andin every room you enter you see no per-son but a solitary black man or womannear the sick. Many people thrust theirparents into the streets as soon as theycomplain of a headache.”

As the nurses recruited by the societybegan caring for sick people around thecity, two exceptional members of themayor’s committee—Stephen Girard andPeter Helm—took charge of Bush Hill.Girard, then a little-known merchant,was a French immigrant who later be-came famous as the financier of the Warof 1812 and the founder of Girard Bankand Girard College for orphan boys.Helm was a cooper of German descent.

Girard accepted responsibility for op-erating the inside of the building, Helmthe exterior and outbuildings. Bothshowed up every day until the epidemicfinally waned later in the year. Girardestablished order and cleanliness andmade sure the patients were given care-ful attention. Helm restored the watersupply to the grounds, built coffins andmore hospital space and supervised the

reception of patients. This reception in-cluded placing the patients in holes inthe ground to protect them from vaporswhile they waited for hospital beds.

Girard was familiar with yellow feverthrough trading in the West Indies. Hepersuaded the mayor’s committee tohire the physician Jean Devèze, one ofthe refugees from Saint Domingue. De-vèze had treated the disease while in theFrench army and would become theworld’s leading authority on the condi-tion. He advocated the so-called Frenchmethod of therapy: bed rest, cleanli-ness, wine and treatment with Peruvianbark (a source of quinine, which is nowused to treat malaria, not yellow fever).At Bush Hill, Girard generously sup-plied the wine from his own cellar.

Controversy over Treatments

American physicians objected to De-vèze’s appointment, however, and

four resigned from the hospital. In part,they may have been angry at losing med-ical authority over Bush Hill. Also, theywere trained in the Scottish and Englishtradition, which was quite different fromthat of the French. Indeed, Devèze stoodat the opposite pole from Rush and hissupporters in what became a raging con-troversy over the best therapy.

In accordance with traditional teach-ings, Rush and many of his contempo-raries believed the body contained fourhumors (blood, phlegm, black bile andyellow bile). They therefore aimed to

treat most illnesses by restoring balanceto the body—through giving laxativesand emetics (to cause vomiting), draw-ing blood and inducing sweating.

Early on, Rush had become convincedthat a combination of these treatmentscould effect a cure—specifically, he fa-vored bleeding combined with deliveryof a mercury-based mixture of calomeland jalap. “I preferred frequent andsmall, to large bleedings in the begin-ning of September,” he wrote in a 1794account of the epidemic, “but towardthe height and close of the epidemic, Isaw no inconvenience from the loss of apint, and even 20 ounces of blood at atime. I drew from many persons 70 and80 ounces in five days, and from a few,a much larger quantity.”

Over time, Rush became increasinglydogmatic. He claimed he “did not losea single patient” who had been bledseven times or more. He also mounteda strenuous campaign, in private corre-spondence and in letters to the editors,urging doctors not to spare the lancet.His letters fiercely criticized physicianswho seemed to pay lip service to thebenefits of venisection but were morerestrained than him in its practice. Heaimed particularly strong invective atDevèze and the few other doctors whothought that radical bloodletting wasdangerous and said so openly.

As Rush’s treatment became progres-sively more radical, his medical col-leagues began to drop their loyalty. Butprominent Philadelphians still often

The Philadelphia Yellow Fever Epidemic of 1793 Scientific American August 1998 91

ABSALOM JONES, founder with Richard Allen of the Free African Society, workedwith Allen and William Gray of the society to recruit African-Americans as nurses forthe epidemic. The recruits won praise. But later Mathew Carey, the main historian ofthe episode, accused some of profiteering and theft. Infuriated, Jones and Allen wrote arebuttal (left) thought to be the first African-American political publication.

LATE SIGNS of yellow fever, such as aprofusely bloody nose and black vomit(seen on pillow), were depicted in 1820as part of a series illustrating the course ofthe disorder. The series, from a Frenchpublication, includes some of the earliestcolor images of the disease’s stages.

LIB

RARY

CO

MPA

NY

OF

PHIL

AD

ELPH

IA

RAPH

AEL

LE P

EALE

Del

awar

e A

rt M

useu

m

CO

LLEG

E O

F PH

YSIC

IAN

S O

F PH

ILA

DEL

PHIA


turned to Rush for care.John Redman, Rush’smentor and still at 71years the reigning deanof medicine in Philadel-phia, submitted to Rush’stherapy and lived. CasparWistar, a respected pro-fessor of anatomy, recov-ered, too. In Rush’s immediate circle,three of his apprentices died, as did hisown sister. Rush and two other appren-tices contracted yellow fever but recov-ered after taking his treatment.

Historians know that neither Rush’smethod nor Devèze’s could have curedpatients. With or without intervention,yellow fever runs a variable course, kill-ing some but sparing others. As is trueof most viral diseases, it remains essen-tially untreatable today, and therapymainly consists of keeping patients com-fortable and preventing dehydration.Devèze’s gentle approach certainly didless harm than Rush’s and may havehelped patients remain strong enough tocombat the virus. But it would be unfairto judge Rush for his decisions. His pro-tocol was simply an aggressive versionof a conventional therapy based on themedical theories of the day.

The epidemic died out on its own inNovember, when frosts killed the mos-quitoes. The city kept no accurate healthrecords, so the precise number of deathscannot be known. By counting freshgraves in the city’s cemeteries and ex-amining church records and othersources, Mathew Carey, who becamethe official historian of the epidemic, re-ported the toll as 4,041 out of a totalpopulation of 45,000. The actual num-ber was surely higher, however. In anaccount published within months afterthe epidemic, even Carey suggested thatthe figure exceeded 5,000.

Of the physicians who remained intown, 10 paid for their dedication withtheir lives. Overall, as many as 20,000residents fled, a number of them endingup quarantined or jeered by frightenedmobs along the roads out of town.

Several yellow fever epidemics were to

follow in Philadelphia, in 1797, 1798,1799, 1802, 1803 and 1805. Epidemicsalso hit other parts of the U.S. at varioustimes, especially in the South, where A.aegypti is common. The last major out-break occurred in New Orleans in 1905.

For Rush, the 1793 event representedboth the apogee and nadir of his career.Initially, his dedication to patients madehim a popular hero; he reportedly saw120 patients a day. But he was too sureof himself, intolerant of alternative ther-apies and unable to bear criticism. Byimposing his will and politicizing themedical community, he became a for-midable obstacle to more enlightenedtreatment by other physicians. “Thedifferent opinions of treatment excitegreat inquietude,” wrote Henry Knox,secretary of war, to President GeorgeWashington, on September 15, 1793.“But Rush bears down all before him.”

Ultimately, Rush’s divisiveness andradical treatment tarnished his reputa-tion. Enraged by perceived slights fromother doctors, he resigned from theCollege of Physicians shortly after theepidemic. Unbowed, however, he dedi-cated himself to treating victims of theyellow fever epidemics that struck Phil-adelphia subsequently, practicing evenmore radical bloodletting than before.

In 1797 his loyal formerstudent, Philip Physick,survived 22 bleedingsand reportedly lost 176ounces of blood.

For Philadelphia, the1793 epidemic led in thefollowing year to creationof the city’s first healthdepartment, among theearliest in the nation. Andpartly as a result of theepidemic, in 1801 Phila-delphia completed thefirst municipal water sys-tem in a major Americancity. Health authoritiesassumed eliminating filthwould help prevent yel-

low fever and other diseases; building awater system was simply part of thatgeneral cleanup.

The efforts of the African-Americancommunity improved relations betweenblacks and whites in Philadelphia whenracial tension in much of the countrywas mounting. During the epidemic, thesteadfastness of the blacks as nurses andhandlers of the dead won widespreadrecognition. In 1794 gratitude led cityleaders and an initially resistant whiteclergy to cooperate with the African-American community as it founded thefirst black-owned and -operated church-es in the nation. Those churches, Afri-can Episcopal Church of St. Thomasand Mother Bethel A.M.E. Church, stillhave active congregations today.

Broad Repercussions

All was not smooth, though. Histori-an Carey accused some blacks of

exploiting the situation, charging goug-ing prices for their services and pillag-ing the homes of patients they hadcared for. In what some experts think isthe nation’s earliest African-Americanpolitical publication, Jones and Allenpublished an angry but restrained re-buttal, vividly describing the horribleconditions under which the nursesworked, unassisted and sometimes re-fusing payment for their services.

The repercussions of the 1793 epi-demic extended beyond Philadelphia.During that episode, Ann Parish, aQuaker woman of independent means,established an institution that enabledwomen widowed by the disease to keeptheir families intact. (While the womenspun wool at the House of Industry, oth-ers watched and taught their children.)

The Philadelphia Yellow Fever Epidemic of 179392 Scientific American August 1998

LIB

RARY

CO

MPA

NY

OF

PHIL

AD

ELPH

IA

BUSH HILL (top), whereVice President John Adamsand his wife had lived fora time, became a hospitalfor indigent yellow fevervictims. To aid the hospi-tal, the mayor published acall requesting donatedsupplies (bottom).

LIB

RARY

CO

MPA

NY

OF

PHIL

AD

ELPH

IA


Her program became a model of inno-vative philanthropy for other cities thatlater suffered yellow fever epidemics.Meanwhile, as a result of the contro-versy stirred up by Rush, French meth-ods of treatment became better knownamong doctors elsewhere, many ofwhom now debated the merit of ex-treme bloodletting.

Along the seaboard, cities developedmeasures for public health, includingestablishing quarantine stations for for-eign ships. Moreover, a sanitation move-ment began that carried into the 19thcentury. For example, various citiesbuilt water and sanitation systems andcleaned the streets more often.

The epidemic—occurring as it did inthe nation’s capital at a crucial time inthe country’s history—had importantpolitical ramifications as well. Beforeyellow fever arrived, public sentimentin the capital leaned strongly towardsupporting the revolutionary govern-ment of France in the war it had de-clared against Great Britain, the Neth-erlands and Spain. An envoy sent byFrance to stir up popular support forthe war, Citizen Edmond Charles Ge-nêt, found enthusiastic crowds in Phila-delphia but received a cool receptionfrom President Washington.

John Adams, who would become sec-ond president of the U.S., later remem-bered that “ten thousand people in thestreets of Philadelphia day after day,threatened to drag Washington out ofhis house, and effect a revolution in thegovernment, or compel it to declarewar in favor of the French Revolution,and against England.. . . Nothing butthe yellow fever.. . could have saved theUnited States from a total revolution ofgovernment.”

The medical catastrophe defused thepolitical excitement. By the time the epi-demic ended, news of the excesses of theFrench government had reached Amer-

ica and public opinion had tilted againstFrance, toward neutrality.

After reappearing many times, yellowfever gradually subsided in the U.S. latein the 19th century. It nonetheless con-tinues to threaten certain tropical areas,particularly in South America and sub-Saharan Africa.

The decline in the U.S. began long be-fore an effective vaccine was introducedin the 1930s. (The vaccine is availablefor soldiers and travelers; unfortunate-ly, it costs too much for populations liv-ing in the areas most prone to epidem-ics today.) Quarantine measures andmunicipal water supplies that reduced

breeding areas for mosquitos probablyplayed a role in ending the U.S. epi-demics, but other factors, such as im-proved sanitation, must have been atwork as well. The drop-off is part of alarger pattern in which death ratesfrom many infectious diseases fell in thecourse of the 1800s.

In tropical regions, the virus finds anatural reservoir in monkeys, and thereis no hope of eliminating it by destroy-ing its natural hosts. Scientists may even-tually be able to control the disease bydisplacing the native mosquito popula-tion with genetically altered forms thatcannot spread the disease, but that ap-proach is far from realization. Mean-while a yellow fever epidemic remains aworrisome possibility in the U.S., par-ticularly in the Deep South, where largepopulations of A. aegypti are found.

In spite of modern knowledge and theexistence of a vaccine, such an epidemiccould grow quickly. Once it began, pub-lic health workers would need time toimmunize populations at risk and to es-tablish effective mosquito-control andquarantine measures. The U.S. Instituteof Medicine has estimated that an out-break of yellow fever in a major city,such as New Orleans, would potential-ly result in as many as 100,000 casesand 10,000 deaths.

The institute has additionally predict-ed that yet unknown viruses could alsobecome major health threats. (HIV, thevirus that causes AIDS, is an exampleof a once obscure virus that triggered amajor epidemic.) It has therefore urgedthe government to adopt better methodsfor identifying new outbreaks of infec-tious diseases, to stop them early. Aftertheir experience with the 1793 yellowfever epidemic in Philadelphia, Ben-jamin Rush and hismedical colleagueswould have firmlyagreed.

The Philadelphia Yellow Fever Epidemic of 1793 Scientific American August 1998 93

The Authors

KENNETH R. FOSTER, MARY F.JENKINS and ANNA COXE TOO-GOOD all share an intense interest inPhiladelphia’s history. Foster is associateprofessor of bioengineering at the Univer-sity of Pennsylvania. Jenkins is superviso-ry park ranger at Independence NationalHistorical Park in Philadelphia. Toogoodhas been a historian in the Park Servicefor more than 30 years. Foster and Jenk-ins are married to each other.

Further Reading

A Narrative of the Proceedings of the Black People during the Late Awful Calami-ty in Philadelphia in the Year 1793, and a Refutation of Some Censures Thrownupon Them in some Late Publications. Absalom Jones. Historic Publications, Philadelphia,1969. Reprinted by Independence National Historical Park, Philadelphia.

Forging Freedom: The Formation of Philadelphia’s Black Community 1720–1840.Gary B. Nash. Harvard University Press, 1988.

Bring out Your Dead: The Great Plague of Yellow Fever in Philadelphia in 1793. J. H. Powell. Reprint edition. University of Pennsylvania Press, 1993.

A Melancholy Scene of Devastation: The Public Response to the 1793 PhiladelphiaYellow Fever Epidemic. Edited by J. Worth Estes and Billy G. Smith. Science History Pub-lications/USA, Philadelphia, 1997.

HISTORIAN MATHEW CAREY calcu-lated lives lost to yellow fever in variousways, such as by counting new graves incemeteries (graph). Such methods sug-gested a total of 4,041. Later estimates,however, put the toll closer to 5,000.

SAM

UEL

SA

RTA

IN, A

FTER

A P

AIN

TIN

G B

Y JO

HN

NEA

GLE

Cul

ver P

ictu

res;

SLI

M F

ILM

S (g

raph

)

120

100

80

60

40

20

0

BU

RIA

LS

AUG. SEPT. OCT. NOV.

SA


Okay, I’ll confess: my scientif-

ic specialty is nuclear phys-

ics, but I would much rather

trace the course of a quiet stream than

find my way along a linear accelerator

any day. And ecology, especially stream

ecology, affords more opportunities for

meaningful amateur research. Indeed,

the techniques used are so straightfor-

ward and the results so immediate that

you can easily craft wonderful scientific

experiences for your whole family with

just a little preparation. And you can

forge ties with others, too—with the

rise of the environmental movement,

thousands of part-time naturalists have

banded together into nearly 800 inde-

pendent organizations that work to as-

sess the health of our nation’s water-

sheds. These associations can help you

and your family jump into this rich and

rewarding area of research.

Some of these groups have just a few

members and operate on a shoestring.

Others have hundreds of participants

and annual budgets of tens of thousands

of dollars. Some limit their activities to

physical and biological surveys: they re-

cord the temperature, rate of flow and

clarity of the water, or they take a regu-

lar census of the seasonal inhabitants of

a stream. Other groups have the equip-

ment to measure dissolved oxygen and

pH levels, check for fecal coliform bac-

teria and test for various pollutants. All

these volunteers make important con-

tributions to bettering the environment,

albeit one stream at a time.

These organizations report their find-

ings to local, state and federal agencies

that have an interest in water quality.

The Environmental Protection Agency

encourages this burgeoning torrent of

amateur talent with a number of free

publications. The EPA also helps to pro-

duce an outstanding semiannual news-

letter called The Volunteer Monitor. This

periodical knits independent environ-

mental groups into a nationwide com-

munity by providing detailed instruc-

tions, some that will enlighten even the

most experienced field researcher. It’s

easy to get your feet wet: just check out

the EPA’s listings on the World Wide

Web to find a watershed volunteer or-

ganization near you.

As with any diverse ecosystem,

streams afford many opportunities for

amateur research. Personally, I love to

investigate the little critters that cling to

rocks and branches or bury themselves

in the sediment. Studying the diversity of

these easily visible “macroinvertebrates”

provides an excellent way to judge the

health of the stream, because pollution

affects certain types more than others.

For example, finding stone-fly nymphs

(relatively sensitive creatures) inhabiting

a brook in one of the Middle Atlantic

states would signal that the level of pol-

lution is probably not too severe. But the

rules change from place to place. So you

will need to learn from people working

in your area which of the local organ-

isms are most sensitive to pollution.

To examine such invertebrates, you

may find that some of the methods used

for standing bodies of water are helpful

The Amateur Scientist94 Scientific American August 1998

by Shawn Carlson

T H E A M A T E U R S C I E N T I S T

SAMPLING FAUNAprovides a simple test for pollution.

Building a Consciousness of StreamsBuilding a Consciousness of Streams


[see “The Pleasures of Exploring Ponds,”

The Amateur Scientist, September 1996].

But the rush of the relatively shallow

water in streams makes the technique

described here particularly well suited

to collecting biological specimens.

To start, you’ll need a good net. Affix a

one-meter square of fine-mesh netting

between a pair of strong supports. (Stout

wooden dowels would work well.) Envi-

ronmental groups in your area may use a

standard-size mesh for such studies—500

microns is common—in which case you

should do the same. Otherwise you can

just use plastic window screening. Then

go out and select three representative

sites from a section of your stream that is

no more than 100 meters long. Your aim

is to sample a one-meter-square patch in

each spot. To do so, you and a partner

should put on rubber hip boots and

wade in, approaching from downstream

so that anything kicked up by your

steps will not affect your investigation.

When you reach your first site, open

the net, dig the ends of the dowels into

the bottom so that the net tilts down-

stream at a 45-degree angle. Make sure

the base of the net lies flush against the

streambed. Have your helper step one

meter in front of the net and pick up all

the large rocks from the sampling area,

gently rubbing them underwater to dis-

lodge any organisms living on them. The

flow of the stream will carry anything

rubbed off into the net. (Alternatively,

rubbing the rocks over a bucket may be

a more direct way for you to collect

these organisms.) Your partner should

then stir up the bottom in the designat-

ed patch by kicking things around for a

couple of minutes so that the organisms

living between the smaller rocks are

swept into the net. Finally, have your as-

sistant grasp the bottom of the net and

scoop it forward as both of you remove

it from the stream and return to shore.

Use a magnifying glass to find and col-

lect creatures from twigs and whatever

other debris you’ve captured before you

discard them back into the stream. Care-

fully dislodge the larger creatures from

the mesh with tweezers; flush off the rest

into a bucket with clean stream water.

Let the contents of the pail settle and

scoop out most of the excess water with

a cup. Strain this water through a nylon

stocking to make sure that nothing es-

capes. Then repeat this procedure at

your other two survey sites, combining

the specimens in your bucket.

Pour the water remaining in the buck-

et through your stocking strainers. Next,

turn them inside out and use a little rub-

bing alcohol to wash the contents into

an empty glass canning jar. Use a sugar

scoop and tweezers to place the rest of

the pail’s living contents into the jar and

fill it with alcohol to kill and preserve

all these invertebrates before screwing

on the top. Finally, place a large label

on the jar and pencil onto it (alcohol

causes ink to run) the location, date, time

and the names of everyone on your crew.

Later you can go through the collec-

tion and record the types and quantities

of organisms present. Environmental

groups working in your area can help

you learn how to identify the sundry

invertebrates. Or, at the very least, they

should be able to point you in the right

direction for gaining these skills. With

this knowledge, you can revisit these

same sites periodically to document en-

vironmental changes from season to sea-

son and, eventually, from year to year.

Although you’ll probably want to gain

some practice with streams close to your

home, don’t overlook the bigger picture.

You can see how your stream fits into

your regional river system by getting a

7.5-minute (scale 1:24,000) topographic

map by the U.S. Geological Survey. You

can purchase such maps at hiking sup-

ply stores or directly from the USGS. By

checking with other groups and plotting

the existing survey sites on your map,

you’ll quickly discover those places in

need of study. Then grab your net, your

bucket, your boots and a few family

members and go to it.

For more information about this andother projects for amateurs, visit theForum section of the Society for Ama-teur Scientists at http://web2.thesphere. com/SAS/WebX.cgi on the World WideWeb. You may also write the Society forAmateur Scientists, 4735 ClairemontSquare, Suite 179, San Diego, CA 92117,or call 619-239-8807.

The Amateur Scientist Scientific American August 1998 95

RESOURCES

The pamphlet “EPA’s Volunteer Moni-toring Program” is available from USEPA,Volunteer Monitoring, 401 M Street SW(4503F), Washington, D.C. 20460. For infor-mation, check out http://www.epa.gov/ owow/monitoring/vol.html on the WorldWide Web.

To obtain the national newsletter, sendyour name and address to The VolunteerMonitor, 1318 Masonic Avenue, San Francis-co, CA 94117, or call 415-255-8049.

To learn how to purchase a map from theUSGS, call 800-USA-MAPS.

Seniors looking for volunteer opportuni-ties should contact the Environmental Al-liance for Senior Involvement (EASI). Writeto EASI, 8733 Old Dumfries Road, Catlett,VA 20119, or consult http://www.easi.orgon the World Wide Web.

RIC

K JO

NES

SA


The extremely polite monks of

the Perplexian order like to

play logical tricks on one an-

other. One night, while Brothers Archi-

bald and Benedict are asleep, Brother

Jonah sneaks into their room and paints

a blue blob on each of their shaved

heads. When they wake up, each notic-

es the blob on the head of the other but,

being polite, says nothing. Each vaguely

wonders if he, too, has a blob but is too

polite to ask. Then Brother Zeno, who

has never quite learned the art of tact,

enters the cell and begins giggling. On

being questioned, he says only, “At least

one of you has a blue blob on his head.”

Of course, both monks already know

that. But then Archibald starts think-

ing: “I know Benedict has a blob, but he

doesn’t know that. Do I have a blob?

Well, suppose I don’t have a blob. Then

Benedict will be able to see I don’t have

a blob and will immediately deduce

from Zeno’s remark that he must have

a blob. But he hasn’t shown any sign of

embarrassment—oops, that means I

must have a blob!” At which point, he

blushes bright red. Benedict also blush-

es at about the same instant, for much

the same reason. Without Zeno’s re-

mark, neither train of thought could

have been set in motion, yet Zeno tells

them nothing—apparently—that they

do not know already.

When you start to analyze what’s go-

ing on, however, it becomes clear that

Zeno’s announcement—“At least one of

you has a blue blob on his head”—does

indeed convey new information. What

do the monks actually know? Well,

Archibald knows that Benedict has a

blob, and Benedict knows that Archi-

bald has a blob. Zeno’s statement does

more than just inform Archibald that

someone has a blob—it also tells Archi-

bald that Benedict now knows that

someone has a blob.

There are many puzzles of this kind,

including versions involving children

with dirty faces and partygoers wearing

silly hats. These logical conundrums are

called common knowledge puzzles be-

cause they all rely on a statement known

to everyone in the group. It is not the

content of the statement that matters: it

is the fact that everyone knows that ev-

eryone else knows it. Once that fact has

become common knowledge, it becomes

possible to reason about other people’s

responses to it.

This effect gets even more puzzling

when we try it with three monks. Now

Brothers Archibald, Benedict and Cyril

are asleep in their room, and Jonah

paints a blue blob on each of their heads.

Again, when they wake up, each of them

notices the blobs on the others but says

nothing. Then Zeno drops his bomb-

shell: “At least one of you has a blue

blob on his head.”

Well, that starts Archibald thinking:

“Suppose I don’t have a blob. Then Ben-

edict sees a blob on Cyril but nothing

on me, and he can ask himself whether

he has a blob. And he can reason like

by Ian Stewart

M AT H E M AT I C A L R E C R E AT I O N S

Mathematical Recreations

Monks, Blobs and Common Knowledge


ONE MONK

FIRSTRING

FIRSTRING

SECONDRING

TWO MONKS

ILLU

STRA

TIO

NS

BY

MAT

T C

OLL

INS

At least one of you hasa blob on his head. After each ring of the

bell, you must raise your hand if youknow you have a blob.

I’m the only onehere, so I must

have a blob.

My brother has ablob, so I may or may

not have one.

My brother didn’t raise his hand, so he

must’ve seen a blob onmy head.


this: ‘If I, Benedict, do not have a blob,

then Cyril sees that neither Archibald

nor Benedict has a blob and can deduce

immediately that he himself has a blob.

Because Cyril has had plenty of time to

work this out but remains unembar-

rassed, then I, Benedict, must have a

blob.’ Now, because Benedict has also

had plenty of time to work this out but

remains unembarrassed, then it follows

that I, Archibald, must have a blob.” At

this point, Archibald turns bright red—

as do Benedict and Cyril, who have fol-

lowed a similar line of reasoning.

The same logic works with four, five

or more monks, although their deduc-

tions become progressively more con-

voluted. Suppose there are 100 monks,

each bearing a blob, each unaware of

that fact and each an amazingly rapid

logician. To synchronize their thoughts,

suppose that the monastery’s abbot has

a bell. “Every 10 seconds,” the abbot

says, “I will ring this bell. Immediately

after each ring, all of you who can de-

duce that you have a blob must put

your hands up. At least one of you has

a blob.” Nothing happens for 99 rings,

and then all 100 monks simultaneously

raise their hands after the 100th ring.

The logic goes like this: If there is only

one monk in the group, he immediately

deduces that the blob must be on his

head, so he raises his hand after the first

ring of the bell. If there are two monks,

each assumes he doesn’t have a blob, so

neither raises his hand after the first ring.

But then each monk deduces from the

other’s reaction that his assumption is

wrong. “If there was no blob on my

head,” each thinks, “my brother would

have raised his hand. Because he didn’t,

I must have a blob.” So they both raise

their hands after the second ring. The

pattern is the same for any number of

monks—it’s an instance of mathemati-

cal induction, which says that if some

property of numbers holds when n = 1

and if its validity for n implies its validi-

ty for n + 1, then it must be valid for all

n. If there are 100 monks in the group,

each assumes that he has no blob and

expects the 99 other monks to raise their

hands after the 99th ring. When that

doesn’t happen, each monk realizes his

assumption was wrong, and after the

next ring everyone raises his hand.

Another intriguing common knowl-

edge puzzle was invented by John H.

Conway of Princeton University and

Michael S. Paterson of the University of

Warwick in England. Imagine a Mad

Mathematicians’ tea party. Each party-

goer wears a hat with a number written

on it. That number must be greater than

or equal to 0, but it need not be an inte-

ger; moreover, at least one of the players’

numbers must be nonzero. Arrange the

hats so no player can see his own num-

ber but each can see everyone else’s.

Now for the common knowledge.

Pinned to the wall is a list of numbers,

one of which is the correct total of all the

numbers on the players’ hats. Assume

that the number of possibilities on the

list is less than or equal to the number

of players. Every 10 seconds a bell rings,

and anyone who knows his number—

or equivalently knows the correct total,

because he can see everyone else’s num-

bers—must announce that fact. Conway

and Paterson proved that with perfect

logic, eventually some player will make

such an announcement.

Consider just two players, with the

numbers x and y on their hats, and sup-

pose that the list pinned to the wall has

the numbers 6 and 7. Both players know

that either x + y = 6 or x + y = 7. Now

for some geometry. The pairs (x,y) that

satisfy these two conditions are the co-

ordinates of two line segments in the

positive quadrant of the plane [see illus-tration above]. If x is greater than 6, the

y player will end the game after the first

ring of the bell, because he can see im-

mediately that a total of 6 is impossible.

Similarly, the x player will end the game

if y is greater than 6. If neither player

responds after the first ring, these possi-

bilities are eliminated. The game then

terminates on the second ring if either xor y is less than 1. Why? One of the

players can see the hat with a number

less than 1 and knows that his own

number must be 6 or less; therefore, the

total of 7 is ruled out.

With each ring of the bell, the pairs

(x,y) that are eliminated form successive

diagonal segments of the two original

line segments, thereby quickly exhaust-

ing all the possibilities. The game must

end by the eighth ring if x and y both

equal 3. Every other possibility requires

seven or fewer rings. A similar argument

can be applied to games with three or

more players, although the proof is more

mathematically sophisticated.

Mathematical Recreations Scientific American August 1998 97

JOH

NN

Y JO

HN

SON

8

y

7

6

5

4

3

2

1

00 1

1

1

2

2

2

3

3

3

4

4

4

5

5

5

6

6

68

7

7

8 x

NUMBER OF RINGSBEFORE GAME ENDS

JOH

NN

Y JO

HN

SON

FEEDBACK

Feedback in the July 1997 column discussed a problem suggested by RobertT. Wainwright of New Rochelle, N.Y., asking for the largest possible area that

can be covered by equilateral triangles whose sides are integers that have nocommon divisor. As several readers pointed out,there is no upper limit if the number of trian-gles can be made as large as required, but thepoint of the problem is to maximize the areafor a given number of triangles. Wainwrightreached an area of 496 for 11 triangles—where, for convenience in counting,the unit of area is that of an equilat-eral triangle with a side of 1. RobertReid of London found tilings with 12through 17 triangles. The areas are: 860(12 triangles); 1,559 (13); 2,831 (14); 4,751(15); 8,559 (16); and 14,279 (17). — I.S.

LARGEST TILING for 17 triangles

14 6

16

27

27

47

47

51

50

44

27

1111

5

SA

LINE SEGMENTS show how long the two-player game lasts.


Reviews and Commentaries98 Scientific American August 1998

One decade ago the Massa-

chusetts Institute of Technol-

ogy assembled a Commis-

sion on Industrial Productivity to ex-

amine the reasons American dominance

of high-technology industry was experi-

encing a serious competitive decline.

The commission’s 1989 book, Made inAmerica: Regaining the Productive Edge,documented the problems in a number

of industries—problems that were evi-

denced by poor product quality, non-

competitive production costs and ex-

cessively long product cycles, as com-

pared with best practice in Asia. The

most painful part of this realization was

that the decline affected not only steel,

textiles and automobiles but also the

high-tech microelectronics, computer

and communications industries. This

book was a call to arms. What were

Americans doing wrong? How could

we fix it?

Now the leader of the team that wrote

Made in America has come out with a

sequel, borrowing the title from the sub-

title of the earlier study. But the ques-

tion now being asked is not what did we

do wrong, but what are we doing right?

Lester examines the easy answers and

finds some truth in all of them: our most

threatening competitor, Japan, was

struggling with the rupture of the “bub-

ble” economy and no longer had almost

unlimited access to capital. After the

profligate public borrowing of the Rea-

gan administration, politicians of both

parties committed themselves to bring-

ing the federal budget into balance. Ex-

change rates had altered dramatically

throughout the 1980s, with the U.S. dol-

lar losing half its value against the yen.

And most important—and clearly a mat-

ter of serious concern for the future—

U.S. wages had reversed their traditional

course. Union membership in manufac-

turing was declining; firms were reduc-

ing costs by laying off middle managers

who could find only lower-paying jobs.

But these answers, which imply that

the favorable conditions industry now

enjoys may not last forever and may not

be acceptable to the population at large,

failed to explain the steps managers had

taken to change corporate culture. Les-

ter looks into corporate experience with

the “marginal manifestos” offered by

an explosion of business management

books. He evaluates two approaches

that did attempt to address the total per-

formance of the enterprise: Total Qual-

ity Management (TQM) and Corporate

Reengineering. About two thirds of some

500 firms surveyed by Arthur D. Little

were disappointed in these approaches.

Lester notes that many firms did not

appreciate how profound change had

to be and how long it would take for

the benefits to be evident.

The truth can be best understood in

the strengths and weaknesses of Ameri-

can culture: our tendency to compla-

cency and hubris (my word, not Les-

ter’s) and our ability to accept challeng-

es when they finally get our attention

and to come up with innovative solu-

tions. American managers seemed slow

to react when high-tech trade started to

go south around 1970. But once the

erosion of U.S. market shares got their

attention, many managers brought a

level of flexibility and innovation to

both organization and strategy that al-

lowed them to respond much more

quickly to competitive threats.

The key elements of success are basic

and easy to understand. They include

loyalty, focus, innovation and a willing-

ness to take risks. But they must be

deeply and comprehensively embedded

in corporate culture. This explains, at

least in part, why it took so long for the

turnaround to produce visible results at

the macroeconomic level (although some

firms, of course, were able to transform

their behavior earlier). Once American

industry had broken away from the

mass-production model and

given up hope that exposing

everyone to a TQM class

would do the trick, the rate

of recovery, assisted by fi-

nancial troubles in Japan, has

astonished almost everyone.

The groundwork for the re-

covery was in large measure

already coming into place be-

fore the 1990s. Many firms

had placed their priorities

where they saw the most en-

trepreneurial opportunity,

whereas Asian strategies

aimed at commoditizing products to

leverage their superior handling of pro-

cess technology, factory management

and cost. While Japan and Korea were

placing their bets on commodity mem-

ory chips and were failing to master the

software business, American entrepre-

neurship was alive and well. Engineers

were leaving semiconductor companies

and starting application-driven busi-

nesses that could add more value to the

electronic hardware.

The U.S. software industry not only

grew in size but began to take from

hardware the role of defining and deliv-

ering function for the user (witness the

extraordinary position now occupied

R E V I E W S A N D C O M M E N T A R I E S

AMERICA’S INDUSTRIAL COMEBACKReview by Lewis M. Branscomb

The Productive Edge: How U.S. Industries Are Pointing

the Way to a New Era of Economic Growth

BY RICHARD K. LESTER

W. W. Norton and Company, New York, 1998 ($29.95)

FRO

M T

HE

PRO

DU

CTI

VE E

DG

E

MARKET SHAREof global semiconductor industry

for Japan and the U.S., 1982–2000


by Microsoft). Businesses paid new at-

tention to total customer satisfaction by

delivering a complete system of hard-

ware, software and service. Like IBM,

many discovered that offering service

may bring more profit to the firm and

more value to the customer than simply

selling hardware whose price continues

to fall. But until companies began to pay

attention to their product quality and

cost and recognized that a high-tech

product built in a low-tech factory could

not compete, the more value-added strat-

egy of American firms could not show

through in the economic statistics.

Another important factor in the re-

birth of manufacturing competitiveness

is the power of information technology,

which has allowed everyone in the firm

to seize the initiative from management

and permitted flatter, more flexible or-

ganizations. But the Internet, like TQM

and Reengineering, proved to be no sil-

ver bullet. Its contribution to manage-

ment and to new business innovation is

real and highly significant. But the long

time required to realize the benefits of

information technology simply points

up the rightness of Lester’s focus on cul-

ture change. A Harvard University study

of supply-chain integration in the auto-

mobile industry revealed that the auto-

maker and its primary suppliers had

great difficulty using well-known com-

puter design and networking technolo-

gy to integrate their product develop-

ment until they were able to create a

project-driven, collaborative culture.

In short, there is no “one size fits all”

solution for U.S. manufacturing. Lester

notes in six industry case studies that

each industry, and often individual firms,

found unique solutions to the risks pro-

duced by more open global markets

and ever more sophisticated and deter-

mined competition. Furthermore, he is

not fully satisfied that industry’s prob-

lems are all solved or that Americans

would be satisfied with the solutions.

He leaves us with two problems.

Lester dares to ask the hard question:

Is the turnaround an illusion? He ob-

serves that manufacturing productivity

stopped growing at about 2 percent a

year around 1973 and has been stuck at

less than 0.5 percent ever since. Is this a

reflection of the difficulty economists

face in measuring productivity? the re-

sult of the admixture of services and

manufacturing as firms become more

responsive to customer needs? or the re-

sult of a more competitive economy in

which margins continue to be pressed

by the most productive performers, and

the workers laid off by one firm strug-

gle to find productive roles in another?

The optimists among us might feel that

as long as the stock market soars, high-

tech market shares are growing again

and the federal budget is in balance, we

can tolerate a lagging productivity rate.

But Lester doesn’t leave us much

room for a rebirth of complacency. He

finds that much of our business success

has been bought at the expense of an

inexorable growth in income disparity

in the American workforce and a deep

sense of personal vulnerability to the

uncertainties of a market-driven, global

economy. He worries that the massive

research investments made by govern-

ment in the past, which provided the in-

tellectual resources for American inno-

vation, may be curtailed. He recognizes

that the firms exhibiting best practice

are not typical. He is especially con-

cerned about low savings rates and rates

of investment. But underlying all the

threats and uncertainties about the fu-

ture is a pervasive empowerment of in-

dividuals, together with a great increase

in the demand for personal responsibil-

ity. Lester asks us to think about how

this call for a “new economic citizen-

ship” can enable a continuation of eco-

nomic success and at the same time re-

duce the income disparities and sense of

vulnerability that deprive many Ameri-

cans of the full fruits of that success.

LEWIS M. BRANSCOMB, formervice president and chief scientist of IBM,is the Aetna Professor, emeritus, inPublic Policy and Corporate Manage-ment at the John F. Kenne-dy School of Governmentat Harvard University.

Reviews and Commentaries Scientific American August 1998 99

No institution in the U.S. has

been more influential in in-

spiring people of all ages

with the wonders of life on Earth, past

and present, than the American Muse-

um of Natural History (AMNH) in

New York City. Most appropriately, the

museum has now turned its attention to

biodiversity and opened a splendid new

exhibit on biodiversity and the ways in

which we are affecting it. Simultaneous-

ly, noted AMNH scientist Niles Eldredge

has presented a book to enhance the

experience and to make more informa-

tion available. This is a book for the

general reader who wants to understand

what the mysterious notion of “biodi-

versity” comprises and to appreciate the

damage we are doing to our life-sup-

port systems on a daily basis.

Eldredge is at his best in telling his

readers about the rich web of life that

exists in the Okavango Delta in Botswa-

na (a landlocked country just north of

South Africa). This remarkable area is

one that he knows well and loves deep-

ly. It shows us, he points out, “precisely

what. . . the environment in which we

evolved...was like.” Furthermore, Egypt

was like this at the time of the pharaohs,

and our civilization grew and became

more complex and varied under such

conditions. Mummified remains of sa-

cred ibises and inscriptions on the walls

of the pyramids attest to the accuracy

LESSONS FROM PARADISEReview by Peter H. Raven

Life in the Balance: Humanity and the Biodiversity Crisis

BY NILES ELDREDGE

ILLUSTRATIONS BY PATRICIA WYNNE

A Peter N. Nevraumont Book/Princeton University Press,

Princeton, N.J., 1998 ($24.95)

PATR

ICIA

WYN

NE,

USE

D B

Y PE

RMIS

SIO

N O

F N

EVR

AU

MO

NT

PUB

LISH

ING

CO

MPA

NY


Reviews and Commentaries100 Scientific American August 1998

of this view. Several thousand years ago

droves of wildlife and all the interwoven

ecological systems of which they were

part existed throughout the East Afri-

can Rift Valley and, though sometimes

in diminished expression, far beyond.

Now they exist only in the Okavan-

go. There our ancestors lived in balance

with and exploited the productivity of

the ecosystems that still occupy this huge

depression. Eldredge describes it as a

fan-shaped system of waterways, grass-

lands and riverine forest, some 170 kilo-

meters (106 miles) wide and 140 kilo-

meters (87 miles) long—a gigantic oasis

where the water remains fresh and life

plentiful. The diversity of the verte-

brates that coexist here is very great

and that of invertebrates even more so.

Dominant in the region’s productivi-

ty are mound-building termites, Macro-termes michaelseni. A mature mound is

often three meters high—a dominant fea-

ture of the landscape. It contains some

five million individual termites, which

grow and harvest specialized fungi that

occur nowhere else, nourish plants that

are abundant only in the vicinity of the

mound (some of which form the princi-

pal food of the termites) and feed doz-

ens of species of predators large and

small. In various stages of maturation

and abandonment, the mounds provide

homes for many of these animals. In the

mounds, an elaborate system of vents

maintains the temperature at 25 to 30

degrees Celsius (77 to 86 degrees Fahren-

heit) and the humidity between 88 and

96 percent: neither the termites nor the

fungi can survive outside these ranges.

Building on this vision of an early and

relatively unspoiled paradise, Eldredge

shows how humans, with their cattle

and fences, are encroaching on the Oka-

vango Delta today. The cattle are barred

from the tsetse fly–infested swamps, but

the fences halt the native antelopes and

other grazing mammals in their normal

migratory tracks, and they die in droves

along the extensive fence lines. The de-

velopment of a more comprehensive

system, involving both humans and na-

ture, has been undertaken by the Camp-

fire Movement in East Africa and prac-

ticed well and beneficially by the Kenya

Wildlife Service under the direction of

Jonah Western. These approaches point

the way to greater success in preserving

what Eldredge calls “a fantastic rem-

nant of our own ancestral climes.”

Beyond the delta, similar ecosystems

have not fared well at the hands of our

forebears and are not faring well now.

With the cooling of the waters off Afri-

ca for the past several million years, the

continent itself began to cool and dry

out. Savanna vegetation became wide-

spread and largely replaced the rain for-

ests and other dense forests that former-

ly occupied vast areas. Our own ances-

tors adopted their distinctive ways of

life in these newly formed climes and,

Eldredge argues, ultimately came to live

in a degree of harmony with the “big

hairies” (the large African mammals)

that formed such an important feature

of their lives. As people spread through-

out the world, however, they came into

contact with many other big hairies and

quickly decimated the populations of

these creatures, a ready source of food

and sometimes a danger as well.

At the time Homo sapiens developed

agriculture in a number of widely scat-

tered centers, some 10,000 years ago,

there were very probably fewer than

five million of us throughout the world.

Now we number nearly six billion. We

consume, waste or divert about half the

total primary net photosynthetic pro-

ductivity on land; we use about half the

total supplies of freshwater and affect

directly some two thirds of the planet’s

surface. Over the past 50 years, while

our total population has grown by 3.5

billion people, we have lost a quarter of

our topsoil and a fifth of our agricultur-

al lands, changed the characteristics of

the atmosphere in important ways and

cut down a major part of the forests. It

is no wonder that we are driving, and

will drive, such a high proportion of the

other organisms that live here with us

to extinction over the course of the next

century, thus limiting our own material

and spiritual prospects substantially.

In his concluding chapter, Eldredge,

having shown the ways in which hu-

mans are destroying the very fabric of

life on Earth, offers a number of practi-

cal suggestions about how we can help

ensure its survival in as rich and diverse

a form as possible. The first step, he ar-

gues persuasively, is to acknowledge the

problem. Once we have done so, the at-

tainment of a stable human population

at whatever level we choose—thus setting

the standard of living that we collec-

tively find acceptable—is one key ele-

ment in making the transition to a sus-

tainable world. This transition consti-

tutes the most critical task we face in the

years to come. We must reformulate our

economic principles, employ our exist-

ing expertise in conservation and strike

a balance between what we perceive as

human needs and the continued exis-

tence of healthy ecosystems. Perhaps

most difficult of all, it means we must

organize the political will to confront

this problem for the sake of our chil-

dren and grandchildren—and for the

harmonious and productive

continuation of life on Earth.

The U.S., wealthiest coun-

try the world has ever

known, is home to some 4.5

percent of the human popu-

lation, yet we control about a quarter

of the world’s economy and cause some

25 to 30 percent of its pollution and

ecological damage. This simple relation

ought to make us aware that we de-

pend on global stability and on the

peace and prosperity of nations every-

where—but we are not. As I write these

words, we are selling ourselves gasoline

at the lowest prices since 1920. We are

deeply in arrears to the United Nations,

demanding that weaker and poorer na-

tions take the lead in controlling global

climate change. And we have joined a

handful of small and politically impo-

tent states in refusing to ratify the Con-

vention on Biological Diversity, already

signed by nearly 175 nations and the

principal legal instrument for protect-

ing the biodiversity on which we de-

pend on a global basis.

These alarming facts point clearly to

the unreality of our worldview and the

dangers it poses for us and all other in-

habitants of this planet. One can only

fervently hope that both this book and

the magnificent AMNH exhibit will in-

spire us, as so many have been inspired

in that institution before, to take action

on an individual and collective basis to

deal with the world as it is rather than to

follow blindly the imperfect model that

now so dangerously guides our actions.

PETER H. RAVEN is director of theMissouri Botanical Garden in St. Louis.

“As I write these words, we areselling ourselves gasoline at the

lowest prices since 1920.”



In 1848 the king of France abdicat-

ed his constitutional post—not that

he was the only king in Europe to

be replaced by republics in those times.

Yet the older regime came back; even

emperors remained through World War

I. That same spring of 1848 an eloquent

and enduring pamphlet appeared, the

Communist Manifesto of Karl Marx

and Friedrich Engels. Its radical appeal

would sound throughout the entire 20th

century; its critique still defines political

Right from Left, even though the demise

it prophesied seems farther off than ever

after 150 years. The metaphor of right

and left is powerful among us still, but

at the molecular level, that literal distinc-

tion is a rigorous and subtle property of

three-dimensional space. And it was in

1848 that its consequences for matter

first became clear.

No familiar molecular formula that

accounts for the kinds and numbers of

atoms present, such as H2O, is enough

to fix its shape fully. All atom-to-atom

links must be described geometrically.

In a plane that is easy, but in real 3-D it

is trickier. A right glove is so similar to

the left glove of its pair that a basic de-

scription often fits both, although we

know that the two gloves cannot equal-

ly fit either hand. Each glove matches

exactly the image of its partner in an

ideal mirror. Mirror-image molecular

pairs possess what we call handedness;

the chemists term them enantiomers,

from the Greek for “opposites.”

From grapes and from old wine casks,

chemists had by the 1840s separated a

number of compounds in crystalline

purity. Two similar acids held a mystery.

The salts of the two acids—tartaric and

racemic—were alike in their chemical

properties, yet a solution of tartaric acid’s

salt rotated the plane of polarized light

rightward as a beam passed through,

whereas its chemical match, the corre-

sponding salt of racemic acid, was opti-

cally inactive, causing no rotation. One

postdoctoral chemist gifted with the

dexterous hands and attentive eyes of

an artist examined the small crystals

formed as the inactive solution dried.

Then he patiently sorted the crystals by

minor differences in shape into two dis-

tinct crystal forms, until each of his two

piles held enough to be dissolved and

measured for rotation. His hand-sort-

ing of the grains was imperfect, yet the

rotations he found were opposite in

sign and quite similar in amount.

It was this young postdoc, Louis Pas-

teur himself, who disclosed in May 1848

the strangely dual world of biomole-

cules. The inactive racemic salt was a

mix at the molecular level of two forms:

one the expected right-handed salt, the

other its lefty antipode. Their rotations

canceled in the inactive sample. Pasteur

had seen right gloves, left gloves and

their 50–50 mix. A recent historian’s

close study of Pasteur’s laboratory note-

books has revealed that his celebrated

separation was not at all part of his

original experiment design but was

more likely a follow-up of visual hints

from his own observations.

We ourselves own two quartz mineral

crystals, each about the size of a quarter-

pound of butter; one rotates the plane-

polarized light clockwise, the other the

other way. Careful inspection of telltale

facets and close comparison are con-

vincing: the two big crystals are distinct

mirror pairs. But we are dazzled by

Pasteur’s hundreds of choices by eye

and hand at millimeter scale, at once so

bold and so tedious.

Early in 1998 we heard a witty and

compelling lecture by the chemist K.

Barry Sharpless of the Scripps Research

Institute. One astonishing accompany-

ing snapshot showed a robust Scripps

postdoc holding in each arm a 10-

pound bottle of white tartaric crystals.

He had prepared the two distinct forms

in bulk in a day’s work, using catalytic

chemical mixtures. The contrast be-

tween the postdocs of 1848

and of 1998 at work on the

same task offers a thrilling

view of the power of today’s

organic synthesis.

The Sharpless school has

found its way over 20 years

to a variety of practical catalytic syn-

theses of handed molecules, usually

starting with straight chains of a dozen

or so carbon atoms and a couple of side

groups, to add a number of wanted

groups with the chosen handedness. The

necessary information of which molec-

ular glove to sort out is supplied by

choosing a natural one-handed deriva-

tive of quinine. The catalyst is a metal

ion, often highly charged and multiva-

lent. Osmium is the best, a fierce oxidiz-

er for organics. Each expensive osmium-

bearing molecule draws around it the

handed starter and a few other adher-

ents, forming and unforming that mol-

ecular jig tens of thousands of times.

The enzymes of life, all larger protein

catalysts, are able to recycle millions of

times. They deliver a product scrupulous-

ly one-handed to seven decimal places;

lab syntheses admit the wrong hand a

COMMENTARY

A certain molecule induces the odor of caraway; in its mirrored

form, that of spearmint.

Continued on page 103

DU

SAN

PET

RIC

IC

WONDERSby Philip and Phylis Morrison

The Left and the Right


Walking through Parliament

Square in London the oth-

er day, I suddenly remem-

bered that when I was a very small child

during World War II, the sound of Big

Ben ringing the hour on the radio was a

comforting indication that it hadn’t been

hit by incoming German V-1 missiles.

In 1859 the new Big Ben clock became

the miracle of harrumph high-tech pre-

cision (as had been promised by Prime

Minister George Canning) when the

gravity escapement system created by

E. Beckett Denison was installed. Deni-

son’s trick for isolating the movement

of the pendulum from that of the gears

worked in such a way that no matter

how much dirt or ice accumulated on

the four sets of clock hands, Big Ben

would keep time accurate to the second.

(Thank you, Frederickton, New Bruns-

wick, on whose often icy cathedral clock

Denison had done a wet run.)

Now, your basic deadbeat escapement,

on which Denison improved, links the

movement of the pendulum to a couple

of blades. As these move from side to

side and turn about, each one catches on

the clock drive-wheel gear teeth and con-

trols its turn on a shaft wound with cord

attached to a weight. Perhaps the most

famous version of deadbeat escapement

was the brainchild of George Graham,

the greatest clockmaker of the 18th

century. In the summer of 1736 one of

his astronomically accurate clocks was

taken to Sweden by a particularly ar-

rogant French scientist, Pierre-Louis

Moreau de Maupertuis, who was

intent on proving the English

right about the earth’s shape.

That geodesy should have

been at issue was not just an-

other example of the Gallic

chip on the shoulder that had

been there ever since Newton.

It was also the small matter of

multiple shipwreck. Get the

shape of the earth wrong, and you also

get wrong the measure of a degree across

its surface. Which means that ships hit

rocks that shouldn’t be there, because

the ships aren’t where their navigators

think they are. Entire fleets were blun-

dering about in this fatal way and far

too often taking to the bottom large

quantities of gold and guns. Mauper-

tuis offered a solution: using a one-de-

gree difference in the position of a star

(hence the need for precise timing), he

calculated exactly how far across the

planetary surface represented one degree

of latitude. This took him into several

months of mosquito hell in the middle

of Swedish nowhere. And at the end of

it, sacre bleu, Newton had been right!

The earth was an oblate sphere. No

wonder Maupertuis made enemies back

home. Pro-English, arrogant and right.

Maupertuis became the target of sav-

age attacks by Voltaire and was effec-

tively hounded out of France. He later

died at the home of a Swiss mathemati-

cal pal, Johann Bernoulli, whose math-

ematical brother, Jacob, was deeply into

why hanging chains assumed the posi-

tion they did. This arcane catenary

matter had originally been the ob-

session of a Dutch noodler named

Simon Stevin, who in 1585 also

had the novel idea of decimal

coinage. No takers, until a one-

legged American aristocrat,

Gouverneur Morris, tossed it

to Thomas Jefferson, who

promptly made it his own.

In any case, in 1794 Morris

successfully promoted an-

other scheme (this also snitched, by an-

other politician, DeWitt Clinton): build

a canal from Lake Erie to the Hudson.

A project famed in song and story until

somebody built a railroad across New

York State in 1851, which eventually

doomed the canal.

By 1855 the New York & Erie Rail-

road had 4,000 employees and lo-

gistical problems so complex that its

general superintendent, Daniel McCal-

lum, was obliged to invent modern busi-

ness administration to deal with them:

autonomous heads of division; profes-

sional middle management; daily, weekly

and monthly reports—all that stuff. And

because one of the things regular and

high-volume railroad deliveries of goods

made possible was the high-turnover de-

partment store, places like Wanamaker’s

and Macy’s were among the first busi-

nesses to adopt the administrative rail-

road pyramid structure.

Efficient delivery of an unprecedented

range of items—from candelabra to ti-

aras to gloves—and production line–in-

spired sales procedures left the new

mass-market retailers with only one lit-

tle problem: how to persuade the cus-

tomer to cart purchases out the front

door as fast as they were being hauled

in the back. So the stores tried trans-

forming their downtown emporia into

something between a Victorian boudoir

and the palace of Ramses II. With ex-

tras thrown in, including musicians,

beauty salons, post offices and nurs-

eries. For the first time, shop-

ping became a luxury

COMMENTARY

CONNECTIONSby James Burke

Tick Tock

Sacre bleu,

Newton had been right!

DU

SAN

PET

RIC

IC

Reviews and Commentaries102 Scientific American


experience. Once, that is, management

had convinced its lady clients (the men

wouldn’t) to turn up and shop till they

dropped.

Wanamaker’s gave this crucial public-

relations exercise to N. W. Ayer, the first

fully fledged advertising agency, and near

the end of the 19th century a new breed

of thinkers known as psychologists were

being recruited to delve deeper into the

consumer’s emotional motivations to see

if further profitable manipulation might

be in the cards. Harvard professor of

psychology Walter B. Cannon (the only

psychologist, I believe, to have had a

mountain named after him) advanced

matters even more. With the use of the

amazing new x-rays and a barium meal

(which he invented), Cannon was able

to study the peristaltic waves that ac-

companied digestion and hunger, and he

saw that these stopped suddenly if the

subject was emotionally affected. After

years of experiment, in 1932 Cannon’s

masterwork, The Wisdom of the Body,introduced the concept of homeostasis:

the maintenance of balance in the body’s

state via chemical feedback mechanisms.

Collaborating with Cannon was Mex-

ican neurophysiologist Arturo Rosen-

bleuth, who in the early 1940s began to

have conversations about feedback with

a mathematical prodigy, the irascible

Norbert Wiener of M.I.T. He was inter-

ested in such phenomena as the way

feedback works to ensure that when you

pick up a glass of water and drink from

it, your mouth and glass manage to

meet. Wiener was concerned about hits

and misses because he was working on

the kind of equations that would facili-

tate antiaircraft artillery in potting their

targets more frequently than once every

2,500 shells (the average score on the

south coast of England at the time—and

no way to win a war).

Wiener’s system, called the M-9 Pre-

dictor, used feedback from radar data

to help forecast where the enemy would

be up in the sky when the next shell ar-

rived there. On that basis, his invention

directed how the gun should be point-

ed. The gizmo worked so well that in

the last week of large-scale German V-1

attacks, of 104 missiles launched against

London only four made it through.

Which is why, thanks to Wiener, Big

Ben chimed its way, undamaged, right

through World War II.

High time I was out of here.


ON SALE AUGUST 25

Also in September. . .

Attention-Deficit Hyperactivity Disorder

The Oort Cloud

A. B

ASS

, © J

. PA

UL

GET

TY T

RUST

NA

SA

A LAST LOOK AT The LaetoliFootprints

Human Healthin Space

COMING IN THE SEPTEMBER ISSUE. . .

MakingSuperheavy Elements

A LAST LOOK AT The LaetoliFootprints

Human Healthin Space

few percent of the time. Biological strict-

ness is based on a long-evolved molecu-

lar fit for the desired product, a precision

key and lock. But the price of life’s high

precision is a strictly limited start and

end. Synthetic catalysis exploits a high-

energy intervention instead of the many

controlled low-energy steps of protein-

led biochemistry, so it can be both selec-

tive and versatile.

For nonchemists, the insight is that a

chemical reaction is not at all only what

the brief equation of the reaction shows.

Those two balanced sides are merely the

endpoints, as though on a transcontinen-

tal road trip you saw only Cambridge,

Mass., and San Diego, Calif., and noth-

ing of heavy traffic and steep grades be-

tween. If we could watch, we would see

the heavy-ion catalyst draw its compan-

ions, bind and unbind in a sequence of

steps, and recycle. Material enough is

provided to supply the groups added in

the synthesis. A catalytic mix with its

auxiliaries includes understandably half

a dozen compounds, rationally chosen

from a complex lore of reactions and

tested by many trials. These are then

added in sequence to a big pot of cooled

alcohol solvent that is easy to evapo-

rate: no grain-by-grain sorting here!

Almost all hormones, pheromones

and other neurological drugs—indeed

most specific pharmaceutical or agrobi-

ological compounds—are functional in

single-handed forms. But synthetics

come as a mix, unless they have been

made through the use of enzymes chosen

from life or by one of the new asym-

metric syntheses. The mirrored molecule

can surprise, too: a certain molecule in-

duces the odor of caraway; in its mir-

rored form, that of spearmint. The

wrong hand of Ritalin, a molecular

species used or overused to control the

hyperactivity of children, seems to be

completely inert. It is said that the trag-

ic congenital injuries from thalidomide

were in fact a side effect of the worse

than useless wrong form. Thus, at best,

what is sold at high price is 50 percent

inactive.

But the synthetics industry is taking

swift steps to deliver life products only

in the hand desired. Within a decade

most handed products will be in your

drugstore in one form only; the Food

and Drug Administration will be

watching its mirrors carefully. SASA

Wonders, continued from page 101


INTERMEDIATE-PRESSURE CHAMBER

HIGH-PRESSURE AIR

VALVE

INHALING

AIRFROMTANK

TO SECONDSTAGE

EXHALING

WATERPRESSURE

HIGH-PRESSURE DIAPHRAGM

WATERPRESSURE DIAPHRAGM

VALVE

MOUTHPIECE

INHALING

EXHALING

DIAPHRAGM

CHAMBER INTERMEDIATE-PRESSURE

AIR

Working Knowledge

REGULATOR FIRST STAGE maintains an intermediate pressure of 90 to 140pounds per square inch in the hose leading to the second stage. There are sev-eral different designs; perhaps the most successful is the balanced diaphragm.Whenever the diver takes a breath, the pressure falls in the hose and intermedi-ate-pressure chamber, causing the high-pressure diaphragm to flex inward. Thisflexing opens a valve and allows high-pressure air to flow from the tank into thechamber and hose. The surrounding (ambient) water pressure against the di-aphragm also ensures that the intermediate pressure in the chamber and hosestays at a certain level above the ambient pressure.

W O R K I N G K N O W L E D G ESCUBA REGULATORS


by William A. Bowden

Chief Operating Officer, Dacor Corporation

The demand regulator, which

connects a high-pressure

scuba tank to a breathing

mouthpiece, opened the vast world

beneath the waves to hundreds of mil-

lions of people. It is not much to look

at, but this remarkable device enables

anyone in reasonably good health to

see what people could only imagine

or speculate about for thousands of

years. In so doing, it transformed per-

ceptions of the undersea domain

from that of a mysterious place hav-

ing dark and often fearsome connota-

tions to a more familiar realm known

mainly for its beauty and wonder.

A regulator enables a diver to

breathe comfortably underwater by

regulating the tank’s high-pressure

breathing gas, matching the gas pres-

sure to the ambient pressure at what-

ever depth the diver happens to be.

This regulation takes place in two

stages. The first stage reduces the tank

pressure, which is typically about

3,000 pounds per square inch (psi) in

a full tank, to an intermediate pres-

sure between 90 and 140 psi. The sec-

ond stage then reduces the intermedi-

ate pressure to the ambient pressure—

44 psi at a depth of 66 feet in seawater,

for example, or 59 psi at 100 feet.

Regulators first became available to

the general public in the U.S. in 1953,

when Dacor introduced the C1 and

U.S. Divers came out with its model

of the Aqua-Lung. Regulators meta-

morphosed significantly in the mid-

1960s, when double-hose designs were

replaced by single-hose types in which

the exhaled gas was discharged near

the mouthpiece. Another period of

rapid improvement occurred in the

late 1970s, when natural rubber and

brass components were replaced with

more durable and impervious syn-

thetic materials (such as silicone and

polyurethane) and stainless steel. Also,

there were ergonomic improvements

to mouthpieces and exhaust manifolds.

ILLU

STRA

TIO

NS

BY

KA

RL G

UD

E; S

CH

EMAT

ICS

BY

BRY

AN

CH

RIST

IE

REGULATOR SECONDSTAGE reduces the inter-mediate pressure in thehose to the ambient pres-sure at whatever depth ithappens to be, allowingthe diver to breathe com-fortably at that depth. Likethe first stage, it relies on adiaphragm. Breathing inthrough the mouthpiececauses the diaphragm toflex inward, pushing a leverthat opens a valve to admitintermediate-pressure airfrom the hose. Waterpressing against the outerside of the diaphragmmatches air pressure in thechamber to the surround-ing water pressure.


the eye · 100 scientific american august 1998 reviews and commentaries of this view. several...

Documents