Methane Point to Bacteria on Mars and Titan?

DO-IT-YOURSELF QUANTUM MECHANICS • LOST MAMMALS REWRITE HISTORY

CANCER AND CHROMOSOMES SCIENTIFIC A Radical New Theory

MAY 2007 AMERICAN WWW.SCIAM.COM Hazy Hints of Alien Life Does methane point to bacteria on Mars and Titan?

Brains beyond Coma Mysteries of the vegetative state

Prevent Blackouts Smart power grids reroute themselves

Nanotube Nets Better than wires for ﬂ exible electronics

COVER STORY: SPACE SCIENCE 42| The Mystery of Methane on Mars and Titan BY SUSHIL K. ATREYA Could the methane in the atmospheres of Mars and Titan be caused by unusual geologic activity—or life?

MEDICINE 52IChromosomal Chaos and Cancer BY PETER DUESBERG Current wisdom on the role of genes in malignancy may not explain some of the features of cancer, but stepping back to look at the bigger picture inside cells reveals a view that just might.

ENERGY 60IPreventing Blackouts BY MASSOUD AMIN AND PHILLIP F. SCHEWE A smarter power grid that automatically responds to problems could reduce the rising number of debilitating blackouts.

PALEONTOLOGY 68ISouth America’s Missing Mammals BY JOHN J. FLYNN, ANDRÉ R. WYSS AND REYNALDO CHARRIER An unexpected menagerie of unique mammal fossils is overturning long-held ideas about South America’s geologic history. 52 Do faulty chromosomes INFORMATION TECHNOLOGY trigger cancer? 76ICarbon Nanonets Spark New Electronics BY GEORGE GRUNER Random networks of tiny carbon tubes could make possible low-cost, ﬂ exible devices such as “electronic paper” and printable solar cells.

NEUROSCIENCE 84IEyes Open, Brain Shut BY STEVEN LAUREYS Brain-imaging techniques yield a better understanding of patients in the vegetative state.

PHYSICS 90IA Do-It-Yourself Quantum Eraser BY RACHEL HILLMER AND PAUL KWIAT Carry out a home experiment to illustrate one of the weirdest effects in quantum mechanics.

www.sciam.com SCIENTIFIC AMERICAN 5 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. SCIENTIFIC AMERICAN Volume 296 Number 5 departments 10 SA Perspectives Even ﬂ awed scientists have good ideas. 12 How to Contact Us 12 On the Web 20 104 14 Letters 18 50, 100 & 150 Years Ago 20 News Scan 40 Insights Saving Amazonia’s indigenous peoples means ■ Three additions to nanotech’s tool kit. not meeting them, insists Sydney Possuelo— ■ Protein folding and “silent mutations.” a policy of noninterference he hopes to extend, ■ Live streaming mobile video. even if others hate it. ■ Ionization via electron tunneling observed. ■ Text-free cell phones for China, Africa and India. 96 Working Knowledge ■ Heart cells power micropumps. Beams of sound. ■ By the Numbers: Agricultural productivity. 98 Reviews ■ Data Points: Data explosion. In two new books, University of Oxford art historian Martin Kemp explores, often brilliantly, the interplay of art and science.

40 Sydney Possuelo, advocate of Brazil’s indigenous people columns

38 Skeptic BY MICHAEL SHERMER How self-deception proves to be more powerful than deception. 39 Sustainable Developments BY JEFFREY D. SACHS Nurturing new technologies is the key to preventing climate change. 102 Anti Gravity BY STEVE MIRSKY Ale’s well with the world. 104 Ask the Experts Where is the universe expanding to? What is “junk” DNA, and what is it worth?

Cover image by Ron Miller; photograph at left courtesy of Sydney Possuelo; illustration on page 5 by Kenn Brown.

Scientifi c American (ISSN 0036-8733), published monthly by Scientifi c American, Inc., 415 Madison Avenue, New York, N.Y. 10017-1111. Copyright © 2007 by Scientifi c American, Inc. All rights reserved. No part of this issue may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording for public or private use, or by any information storage or retrieval system, without the prior written permission of the publisher. Periodicals postage paid at New York, N.Y., and at additional mailing offi ces. Canada Post International Publications Mail (Canadian Distribution) Sales Agreement No. 40012504. Canadian BN No. 127387652RT; QST No. Q1015332537. Publication Mail Agreement #40012504. Return undeliverable mail to Scientifi c American, P.O. Box 819, Stn Main, Markham, ON L3P 8A2. Subscription rates: one year $34.97, Canada $49 USD, International $55 USD. Postmaster: Send address changes to Scientifi c American, Box 3187, Harlan, Iowa 51537. Reprints available: write Reprint Department, Scientifi c American, Inc., 415 Madison Avenue, New York, N.Y. 10017-1111; (212) 451-8877; fax: (212) 355-0408. Subscription inquiries: U.S. and Canada (800) 333-1199; other (515) 248-7684. Send e-mail to [email protected] Printed in U.S.A.

Even mentioning the name Peter Duesberg in- have dramatic relevance to human affairs: Is there fl ames strong feelings, both pro and con. After gain- strong new hope for curing a disease, transforming the ing fame in 1970 as the virologist who fi rst identifi ed economy, building a better mousetrap? Alternatively, a cancer-causing gene, in the 1980s he became the reporters and editors may gravitate toward new sci- leading scientifi c torchbearer for the so-called AIDS ence that easily inspires the public’s sense of wonder, dissidents who dispute that HIV causes the immuno- as so many astronomy stories do. And reports that ap- defi ciency disorder. To the dissidents, Duesberg is pear in certain major scientifi c journals tend to get Galileo, oppressed for proclaiming scientifi c truth more play because those publications have a self-ful- against biomedical dogma. A far larger number of fi lling reputation for releasing the most noteworthy AIDS activists, physicians and papers. (It doesn’t hurt that those journals have par- researchers, however, think ticularly strong public relations departments, too.) Duesberg has become a crank When we look at submitted manuscripts from sci- who refuses to accept abundant entists, we consider it a reassuring sign when the au- proof that he is wrong. To them, thors forthrightly acknowledge both their collabora- he is at best a nuisance and at tors and their competitors and note potential con- worst a source of dangerous dis- fl icts of interest before we ask. If we see that they are information on public health. describing the science of their rivals fairly, we can Readers may therefore be have more confi dence that they are being similarly shocked to see Duesberg as an candid about their own work. (Still, the old nuclear author in this month’s issue. He disarmament treaty maxim applies: trust, but verify.) is not here because we have mis- We typically steer away from controversial ideas too givings about the HIV-AIDS new to have much supporting evidence. Those that UNORTHODOX cancer theory link. Rather Duesberg has also have lasted for years and accumulated some substan- bears consideration. developed a novel theory about tiation have earned consideration. Our judgments are the origins of cancer, one that imperfect, but they tend to mirror those of the scien- supposes a derangement of the chromosomes, rather tifi c community. than of individual genes, is the spark that ignites ma- Blots on a researcher’s history often should bear lignant changes in cells. That concept is still on the on regard for his or her new work. Scientists who have fringe of cancer research, but laboratories are inves- intentionally published fraudulent papers, as the stem tigating it seriously. Thus, as wrong as Duesberg cell researcher Woo Suk Hwang so notoriously did surely is about HIV, there is at least a chance that he two years ago, may be irredeemably tainted. But to is signifi cantly right about cancer. We consider the dismiss a scientist solely for holding some wrong or Researchers, Inc. Photo case worthy of bringing to your attention, with the controversial views risks sweeping away valuable nug- article beginning on page 52. gets of truth. We respect the opinions of any readers Thousands of scientifi c papers appear in technical who may criticize our choice to publish Duesberg in journals every month; why do some rate more fame this case but hope they will nonetheless evaluate his

and journalistic attention? It helps for science news to ideas about cancer on their own merits. STEVE GSCHMEISSNER

THE EDITORS [email protected]

10 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. How to Contact Us On the Web WWW.SCIAM.COM EDITORIAL ADVERTISING For Letters to the Editors: www.sciam.com has electronic contact UPDATED EVERY WEEKDAY Letters to the Editors information for sales representatives of Scientifi c American in all regions of Visit www.sciam.com/ontheweb Scientifi c American the U.S. and in other countries. 415 Madison Ave. New York to find these recent additions to the site: New York, NY 10017-1111 Scientifi c American or 415 Madison Ave. NEWS [email protected] New York, NY 10017-1111 212-451-8893 Please include your name fax: 212-754-1138 First “Commercial” Quantum and mailing address, West/Northwest and cite the article 818-827-7138 Computer Solves Sudoku Puzzles fax: 818-827-7139 and the issue in Detroit which it appeared. Karen Teegarden & Associates Canadian quantum Letters may be edited 248-642-1773 fax: 248-642-6138 computing company for length and clarity. Midwest We regret that we cannot Derr Media Group banks on a long-shot answer all correspondence. 847-615-1921 form of quantum fax: 847-735-1457 For general inquiries: Southeast and Southwest computation: Scientifi c American Publicitas North America, Inc. adiabatic. 415 Madison Ave. 972-386-6186 New York, NY 10017-1111 fax: 972-233-9819 Direct Response 212-451-8200 Special Aditions Advertising, LLC BLOG fax: 212-755-1976 914-461-3269 or fax: 914-461-3433 Australia The One Thing You Need to Know [email protected] IMR Pty Ltd. +612-8850-2220 about Quantum Computers SUBSCRIPTIONS fax: +612-8850-0454 Everything you may think you know about how For new subscriptions, Belgium Publicitas Media S.A. quantum computers work—and what many science renewals, gifts, payments, +32-(0)2-639-8420 and changes of address: fax: +32-(0)2-639-8430 journalists explain in their reporting—is wrong. U.S. and Canada Canada Derr Media Group 800-333-1199 847-615-1921 SPECIAL REPORT Outside North America fax: 847-735-1457 515-248-7684 France and Switzerland Happy International Polar Year! PEM-PEMA or +33-1-46-37-2117 Scientists from around the globe embarked on www.sciam.com fax: +33-1-47-38-6329 a two-year effort to explore and boost understanding or Germany Scientifi c American Publicitas Germany GmbH of the earth’s poles on March 1. +49-211-862-092-0 Box 3187 fax: +49-211-862-092-21 Harlan, IA 51537 Hong Kong FACT OR FICTION? Hutton Media Limited REPRINTS +852-2528-9135 fax: +852-2528-9281 Living People Outnumber the Dead To order reprints of articles: India Booming population growth among the living, Reprint Department Convergence Media Scientifi c American +91-22-2414-4808 according to one rumor, outpaces the dead. fax: +91-22-2414-5594 415 Madison Ave. Japan New York, NY 10017-1111 Pacifi c Business, Inc. PUZZLE 212-451-8877 +813-3661-6138 fax: +813-3661-6139 fax: 212-355-0408 Korea Sci-Doku [email protected] Biscom, Inc. +822-739-7840 Try your hand at Sci-Doku, a Sudoku puzzle that uses PERMISSIONS fax: +822-732-3662 letters instead of numbers, with an added twist: a science- Middle East For permission to copy or reuse Peter Smith Media & Marketing related clue accompanies each puzzle, and the answer is

material from SA: +44-140-484-1321 spelled out in one row or column of the solution. Inc. Systems, D-Wave www.sciam.com/permissions fax: +44-140-484-1320 The Netherlands Only at www.sciam.com or Insight Publicitas BV 212-451-8546 for procedures +31-35-539-5111 or fax: +31-35-531-0572 Permissions Department Scandinavia and Finland 4New Podcasts M&M International Media AB Scientiﬁ c American +46-8-24-5401 Every weekday: 60-Second Science 415 Madison Ave. fax: +46-8-24-5402 New York, NY 10017-1111 U.K. The Powers Turner Group Every Wednesday: Science Talk Please allow three to six weeks +44-207-592-8331

for processing. fax: +44-207-630-9922 Listen at www.sciam.com/podcast JASPAUL CHUNG AND MURRAY THOM

12 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. ® Letters EDITORS@ SCIAM.COM Established 1845 EDITOR IN CHIEF: John Rennie READERS RESPONDED with great passion to our January EXECUTIVE EDITOR: Mariette DiChristina MANAGING EDITOR: Ricki L. Rusting issue. Although Bill Gates’s vision of the developing robotics NEWS EDITOR: Philip M. Yam industry instigated much debate, the majority of letters cen- SPECIAL PROJECTS EDITOR: Gary Stix tered on Matthew L. Wald’s article and its pessimistic view of SENIOR EDITOR: Michelle Press EDITORS: Mark Alpert, Steven Ashley, corn-based ethanol fuel. Correspondents argued from all sides Graham P. Collins, Mark Fischetti, Steve Mirsky, of this issue, although a large number echoed Joyce Ferree of George Musser, Christine Soares CONTRIBUTING EDITORS: W. Wayt Gibbs, Salida, Colo.: “Ethanol may not be the perfect fuel, but it is Marguerite Holloway, Michael Shermer, showing Americans that there are alternatives to foreign oil Sarah Simpson that stimulate rural economies across the U.S., improve our for- EDITORIAL DIRECTOR, ONLINE: Kate Wong eign trade balance and don’t require our military to keep the NEWS EDITOR, ONLINE: Lisa Stein ASSOCIATE EDITOR, ONLINE: David Biello supply coming.... Corn ethanol is not a ‘bridge to nowhere’ if NEWS REPORTER, ONLINE: JR Minkel we look at it as a solid foundation for a new industry to build on.”

ART DIRECTOR: Edward Bell SENIOR ASSOCIATE ART DIRECTOR: Jana Brenning ASSOCIATE ART DIRECTOR: Mark Clemens CORN-FED BEEF izer, the fertilization process itself is far ASSISTANT ART DIRECTOR: Johnny Johnson PHOTOGRAPHY EDITOR: Emily Harrison In “Is Ethanol for the Long Haul?” from 100 percent effi cient, especially PRODUCTION EDITOR: Richard Hunt Matthew L. Wald speculates that etha- when animal waste is used. The result- COPY DIRECTOR: Maria-Christina Keller nol producers may turn to coal or even ing pollution from the nitrogen and COPY CHIEF: Molly K. Frances electricity for process heat instead of phosphorus not assimilated by the crop COPY AND RESEARCH: Daniel C. Schlenoff, Michael Battaglia, Smitha Alampur, natural gas because that fuel has be- has economic and societal consequences Michelle Wright, John Matson, Aaron Shattuck come expensive. Although a small mi- that are real. EDITORIAL ADMINISTRATOR: Jacob Lasky nority of new ethanol plants use coal Lynton S. Land SENIOR SECRETARY: Maya Harty instead of natural gas, an encouraging Ophelia, Va. ASSOCIATE PUBLISHER, PRODUCTION: William Sherman trend appears to be a switch to renew- MANUFACTURING MANAGER: Janet Cermak able biomass for processing, which Living near three ethanol plants, I know ADVERTISING PRODUCTION MANAGER: Carl Cherebin PREPRESS AND QUALITY MANAGER: Silvia De Santis greatly reduces fossil-fuel use. Several that it takes more than a gallon of water PRODUCTION MANAGER: Christina Hippeli new and existing ethanol plants have to produce a gallon of ethanol. Though CUSTOM PUBLISHING MANAGER: Madelyn Keyes-Milch taken this path, yet Wald makes no touted as a “green” fuel in the Midwest, ASSOCIATE PUBLISHER, CIRCULATION: Simon Aronin CIRCULATION DIRECTOR: Christian Dorbandt mention of it. ethanol’s effect on the further reduction RENEWALS MANAGER: Karen Singer Ryan Katofsky of our water tables, lakes and rivers, FULFILLMENT AND DISTRIBUTION MANAGER: Rosa Davis Arlington, Mass. when much of the country has been ex- VICE PRESIDENT AND PUBLISHER: Bruce Brandfon periencing droughtlike conditions over SALES DEVELOPMENT MANAGER: David Tirpack SALES REPRESENTATIVES: Jeffrey Crennan, Naysayers love to jab away at ethanol’s the past several years, does not make Stephen Dudley, Stan Schmidt alleged negative energy balance. The the fuel green in my book. ASSOCIATE PUBLISHER, STRATEGIC PLANNING: truth of the matter is that all energy con- Joy Ophoben Laura Salant versions have a negative energy balance. Rockville, Minn. PROMOTION MANAGER: Diane Schube RESEARCH MANAGER: Aida Dadurian All electric motors, for example, use PROMOTION DESIGN MANAGER: Nancy Mongelli more energy in electricity than they de- ROBOT RUMBLE GENERAL MANAGER: Michael Florek BUSINESS MANAGER: Marie Maher liver in horsepower. In developing the “A Robot in Every Home,” by Bill Gates, MANAGER, ADVERTISING ACCOUNTING energy balance fi gures cited by Wald, was little more than a sales pitch for Mi- AND COORDINATION: Constance Holmes the study’s author, Cornell University’s crosoft. Gates already controls how DIRECTOR, SPECIAL PROJECTS: Barth David Schwartz David Pimentel, piled on extraneous most of the computers on the planet op- MANAGING DIRECTOR AND VICE PRESIDENT, ONLINE: bloated fi gures that other researchers erate and clearly wants to do the same Mina C. Lux DIRECTOR, WEB TECHNOLOGIES, ONLINE: Vincent Ma have deemed insignifi cant. for the home robotics market. The di- SALES REPRESENTATIVE, ONLINE: Gary Bronson Stan L. Simon versity of robot software providers that DIRECTOR, ANCILLARY PRODUCTS: Diane McGarvey Kerkhoven, Minn. he fears will enhance competition and PERMISSIONS MANAGER: Linda Hertz result in better products, whereas a mo- CHAIRMAN EMERITUS: John J. Hanley Most critics and proponents of ethanol nopoly will yield the same security and CHAIRMAN: Brian Napack PRESIDENT AND CHIEF EXECUTIVE OFFICER: grown from corn ignore one major prob- reliability issues we currently experi- Gretchen G. Teichgraeber lem: corn is a nitrogen-hungry crop. In ence with personal computer software. VICE PRESIDENT AND MANAGING DIRECTOR, INTERNATIONAL: Dean Sanderson addition to the costs of manufacturing Dan Gravatt VICE PRESIDENT: Frances Newburg and applying chemical nitrogen fertil- Lawrence, Kan.

As someone who has been working on ture of such competences nor what prob- defi nition of a planet really arbitrary?” requirements for robots for more than lems need to be solved for a robot to The answer should be a simple yes. Ex- 30 years, I believe that Gates has made achieve them. Discerning what is being act defi nitions are undoubtedly neces- a mistake (as have most in the fi eld) in achieved, even in apparently simple ac- sary but are rarely perfect refl ections of thinking that the main obstacle to prog- tions, often requires deep interdisciplin- reality. Any classifi cation or categoriza- ress is a lack of solutions to problems. In ary analysis based in part on experience tion of reality imposes arbitrary separa- truth, the key obstacle is a lack of under- in artifi cial intelligence and philosophy. tions on spectra of experience or objects. standing of what the problems are, how Aaron Sloman What nonarbitrary sense can be read they differ and how they are related. Honorary Professor of Artifi cial from the defi nition adopted by the In- Most robotics engineers are attempt- Intelligence and Cognitive Science ternational Astronomical Union: “[a ing to design systems without thorough- University of Birmingham, England body that] has cleared the neighborhood ly analyzing the requirements necessary around its orbit”? to make them functional. One example RUNAWAY RANGE? Matt Craver is the need to characterize the many I have recently seen a documentary via e-mail functions of vision. Too many people that describes a theory not mentioned in think of vision as simply the recognition “The Mississippi’s Curious Origins,” by SOTER REPLIES: Many defi nitions involve im- Roy B. Van Arsdale and Randel T. Cox. precise or arbitrary boundaries. But when In the fi lm the missing piece of the Oua- nature reveals a large gap in some parameter chita Mountains was believed to have that might have been continuous, it may be been found ... in Argentina of all plac- telling us something. By any measure of dyes! (The Argentine Precordillera region namical dominance, we find an enormous would constitute the missing piece.) Has gap between eight bodies and everything this theory been discredited? else that orbits the sun. This gap provides a Mike Lehrter physically signifi cant criterion and not an ar- Fort Thomas, Ky. bitrary place to draw the boundary defi ning planets. We may have to revise or reject the VAN ARSDALE AND COX REPLY: The idea that new defi nition when we learn more—particu- the southern part of North America rifted larly about other planetary systems—but for away and ended up as the Argentine Precor- now it appears to describe what we know. dillera has been around for at least a couple of decades. It is largely accepted, based on ERRATUM On page 46 of “Is Ethanol for the fossil and lithologic affi nities of sediments Long Haul?” by Matthew L. Wald, it is stated dated to the Cambrian period (0.5 billion that a standard barrel (42 gallons) of ethanol years ago) found in Argentina and the Oua- contains 80,000 British thermal units (Btu) chita Mountains. These similarities suggest of energy. The text should have read that a the two terranes (bodies of rock that exhibit gallon of ethanol contains this amount. distinctive geologic histories) were close together in Cambrian times (that is, they had CLARIFICATION “Airborne Baloney,” by Mi- just rifted apart). Subsequent plate motions chael Shermer [Skeptic], states that it is un- ARE ROBOTS ready for the big time? put the rifted block in Argentina. safe to consume more than 10,000 units of This idea is not in competition with the vitamin A daily. Per the National Institutes of of objects and ignore its role in control- theory that the Ouachita Mountains were Health, this limit applies to adults consum- ling intricate actions, perceiving and breached when the North American conti- ing preformed vitamin A, which is derived understanding processes, seeing possi- nent moved over the Bermuda hot spot. The from animal products. Provitamin A carot- bilities for future processes, and com- Cambrian rifting episode created the Missis- enoids, derived from plants, are considered prehending causal relations. sippi Valley graben (a depressed segment of safe in higher units. We can deepen our understanding of the valley’s crust bordered by faults) shown requirements by examining in far more in our article. Send letters to [email protected] or to detail than roboticists typically do the Scientifi c American, 415 Madison Ave., actual variety of competences displayed PLANETARY PLACEMENT New York, NY 10017. Letters become by humans and other animals. But sim- In “What Is a Planet?” Steven Soter dodg- the property of Scientifi c American and

Automation Phobia ■ Horseless Hearse ■ Clean Weeds

MAY 1957 creasing the stresses in the frail hull NAOSAURUS—“In presenting to public MACHINE BRAWN—“The word ‘automa- structure of torpedo boats. An English view the fi rst and only complete mount- tion,’ a journalistic coinage, has unde- naval architect proved that in the case of ed skeleton in the world of a remarkable servedly become ‘a source of fear,’ ac- the British torpedo boat whose back was fossil wonder, the carnivorous reptile cording to the Earl of Halsbury, an Eng- broken when she was plunging heavily Naosaurus, from the Permian of Texas, lish authority on industrial technology. into a head sea, the gyroscopic resistance Prof. Henry F. Osborn, curator of the In the British journal Impact he at- to a change of plane of the revolving Department of Vertebrate Paleontology tempts to correct the impression created parts of the turbine may have amounted in the Museum of Natural History, New by journalists that automatic processes to several tons, and that these stresses, York, has made a noteworthy contri- in industry ‘will cause widespread un- being unrecognized at the time the boat bution to science. The reader’s special employment.’ Machines require trained was designed, may have carried the total attention is called to the splendid and men to maintain them. Displacements bending and wrenching stress beyond ideal drawing, by Mr. Charles R. Knight. occur mainly in the ranks of the un- the limit of strength of the hull.” —Walter L . Beasley” [Editors’ note: The skilled, who have a high rate of turnover model was an incorrect composite of anyway. Lord Halsbury is concerned for HEARSE SAY—“A lugubrious automobile Edaphosaurus and Dimetrodon fossils.] the man who is not going to be affected novelty in the form of a motor-driven by automation—the coal miner, the ste- hearse has recently made its appearance MAY 1857 vedore and others ‘who do the heavy in the streets of Paris. A nation that re- FOUL FUEL—“We believe that no particu- laboring work for a society which does joices in the grim delights of trolley-car lar use is made of the fl uid petroleum, not know how to lighten their task.’” funerals is hardly justifi ed in shaking its from the ‘tar springs’ of California, ex- head at this latest application of the elec- cept as a lotion for bruises and rheumat- MAY 1907 tric auto. No good reason can be ad- ic affections. It has a pungent odor, and TURN FOR THE WORSE—“The Scientifi c vanced why a man should not be buried although it can be made to burn with a American recently discussed the gyro- in a ceremonious and somber motor pretty good light, its smell is offensive. scopic action of steam turbines in in- hearse, in these piping motor times.” This, perhaps, may be obviated by distill- ing it with some acid; we believe that this is not impossible in this age of advanced chemistry. If the offensive odor could be removed, a valuable and profi table business might be carried on in manufacturing burning fl uid from it.”

IT’S CLEANER GREENER—“‘The tanks for water in India are covered with a green weed,’ says the India Annals of Medical Science, ‘and this, at the same time that it imparts a greenish hue to the water, possesses a remarkable power of fi ltering it, and rendering it wholesome, for where you have this green weed you also fi nd small fi sh and infusoria, which preserve the water also. Sir Charles Napier, in- specting the hill districts of the Panjaub, observing the Bheestees drawing water from one of these tanks, ordered it to be immediately cleaned. The water soon became offensive, and was unfi t for use NAOSAURUS—from a display mounted in 1907 until the same weed was replanted and had again covered the tanks.’”

SCIENTIFIC AMERICAN Digital news SCAN

NANOTECH

I Structured Settings NOVEL TECHNIQUES EXPAND THE NANOTECH TOOL KIT BY CHARLES Q. CHOI

anoparticles are tantalizing con- In the January 19 Science, materials sci- struction blocks for researchers, ca- entist Francesco Stellacci of the Massachu- N pable of displaying properties of both setts Institute of Technology and his col- tiny atoms and far bulkier conventional ma- leagues revealed a way to make nanoparti- terials. They generally behave only like balls, cles act like links in a chain, capable of however, which makes it hard to assemble hooking together into strings of beads. The them into solid structures other than those strategy was to take advantage of the so- resembling displays of oranges in a grocery called hairy ball theorem, which states that store. Now researchers have taken big steps if a sphere is covered in hair, attempts to in creating and using nanostructures that brush those strands to make them all lie fl at have eluded manipulation in the past. will always leave behind two hairs standing up straight, each at opposite poles. (Imagine fl attening hair on a globe just along the latitude lines; in the end, the hair on the poles will stick out.) The investigators covered gold nanoparticles with a mix of two kinds of sulfurous molecular hairs. The points where hairs were supposed to stand up essentially became unstable defects on the nanoparticles’ surfaces, making it easy to replace those hairs. Stellacci and his colleagues substitut- ed these standouts with chemicals that behaved as handles, enabling the nanoparticles to hold onto one another. “This really makes nanoparticles analogous to atoms—specifi cally, divalent atoms that have two chemical bonds. You can now form really interesting structures with them, Yale University just like you can make molecules from at- FAST DIAGNOSES could be done via etched nanowires (blue rods), with bound antibodies (blue and red Y shapes) that detect other antibodies in solution. Rectangles are electrical contacts. oms,” says Stanford University materials

scientist Yi Cui. Stellacci mentions that his MARK REED

20 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. group is now pursuing ways for nanoparti- conductor processes, which would help in- news cles to each have four bonds. tegrate nanowires into electronics. SCAN These nanostructures could be connect- These nanowires proved extraordinarily ed together with nanowires to help form ad- sensitive to their environment, changing in MOLECULAR vanced electronics. Researchers can make voltage on contact with molecules. They can CARRIERS nanowires either by growing them from the detect the activation of a T cell immune re- Ludwig Bartels of the University bottom up or by etching them from the top sponse to a foreign compound within rough- of California, Riverside, and his down, akin to whittling toothpicks from a ly 10 seconds by sensing the acid released by colleagues were the fi rst to log. Bottom-up approaches face the chal- the cells. Detecting such activation usually design a molecule that could lenge of integrating each tiny, fl oppy and takes several minutes, if not hours, to com- move in a straight line on a fl at surface. Now the investigators often randomly dispersed strand into deplete via conventional diagnostic assays us- can make such walkers haul cargo. vices, whereas top-down methods, encom- ing tagged antibodies. The investigators They coaxed the organic compound passing more traditional industrial tech- also show that nanowires bound with anti- anthraquinone to carry and niques, “are like hacksaws,” explains Yale bodies could detect as few as 60 cancer- release two carbon dioxide University biomedical engineer Eric Stern. linked molecules per cubic millimeter, as molecules, as if it had a shopping bag in each hand. Understanding They leave behind rough-edged nanowires good as any current state-of-the-art sensor. how to transport atoms and with poor electrical properties. “You can imagine diagnosing patients molecules in this manner could In the February 1 Nature, Yale research- right at the point of care with devices based help engineers assemble objects ers revealed an etching method that creates off these nanowires, at the ER, the offi ce or or deliver materials. high-quality, smooth-walled nanowires. the battlefi eld,” says Jonathan Schneck, an Next, the researchers plan to The trick was using the ammonium salt immunologist at Johns Hopkins University. make molecular carriers go TMAH. It can etch silicon more slowly and “This represents a much more powerful tool around corners, manipulate their more smoothly than other solvents used than anything I’ve seen coming down the cargo or emit photons to convey their positions. The team may also thus far. Yale biomedical engineer Tarek road in terms of speed of response.” “build ‘muscles’ into them that Fahmy adds that the technique is readily may be triggered by light,” compatible with standard industrial semi- Charles Q. Choi is a frequent contributor. Bartels says.

GENETICS

I Codon Spell Check SILENT MUTATIONS ARE NOT SO SILENT AFTER ALL BY CHRISTINE SOARES

s any diplomat will attest, nuances in single-nucleotide polymorphisms (SNPs) in language can dramatically affect the MDR1. Some did not change the amino acid A success of communication. But be- sequence encoded by the gene, so the scien- cause the idiom of human cells was assumed tists assumed they could have no effect on to be much more literal than that of whole the fi nal protein. Still, those synonymous people, a recent discovery surprised even the SNPs kept showing up in studies of cancer scientists who made it. Researchers at the patients, sometimes seeming to infl uence National Cancer Institute (NCI) identifi ed a drug response. “It drew our attention be- human protein that takes two distinct shapes cause there were so many reports, and often depending on a subtle difference in its gene really contradicting each other,” Kimchi- that should not make any difference at all. Sarfaty says, “so we started to explore.” Their fi nding may explain variations in che- Zeroing in on three particular synony- motherapy response among certain cancer mous mutations, the NCI investigators in- patients and raises new questions concern- serted versions of MDR1 containing those ing so-called silent mutations. SNPs into human and monkey cells. Because Chava Kimchi-Sarfaty and Michael M. the gene encodes a cell membrane pump Gottesman were examining a gene called called a P-glycoprotein, which acts to siphon CHANGES IN CODONS, represented MDR1, which is named for its association chemotherapies out of tumor cells, the team by three letters specifying nucleotides, affect the body more with multiple-drug resistance in tumor cells. measured how well the pump was working than previously thought.

SLIM FILMS FILMS SLIM Speciﬁ cally, they looked at mutations called by adding drugs to the test cells. They no-

www.sciam.com SCIENTIFIC AMERICAN 23 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. news ticed reduced function in cells carrying one don appears to slow protein manufacture in SCAN of the SNPs and then found that the P-glyco- a way that maximizes efficient folding. protein made by those cells was abnormally Whether those examples reveal a new level of shaped. Further experiments revealed that information about the correct fi nal protein the protein was also being manufactured structure encoded in the gene sequence re- more slowly than usual in the cells, which mains an open question. David Baker, who led the researchers to suspect that a phenom- studies protein-folding dynamics at the Uni- enon previously seen only in microbes might versity of Washington, thinks such rare situ- be at work. “It came to us that maybe we ations are more likely to be just accidents that have a problem of codon usage, and we’re evolution has ignored. “There’s probably no really exhausting the system by expressing deeper level of meaning that would be select- lots of MDR1,” Kimchi-Sarfaty recalls. ed for,” he says. For instance, human P-glyco- Codons are three-nucleotide sequences protein is a large protein with multiple do- SUBTLETIES IN within genes that specify amino acids. With mains, Baker explains, “and for a membrane GENE-SPEAK 64 possible triplet combinations of the four protein that folds and is inserted as it gets Making experimental or industrial DNA bases to encode 20 amino acids, multi- made, you can imagine that if you slow it amounts of drugs and other novel ple codons can share the same amino acid down, the folding won’t happen correctly.” proteins often means inserting meaning. Different organisms use certain co- Genome analyses tend to confi rm that foreign genes into microorganisms such as bacteria dons more often than alternative synonyms, there is little evolutionary selection of codon and yeast. Designers of those and such favored usage is refl ected in the rela- bias in mammals, according to Laurence D. genes defi nitely need to take tive abundance of cellular translators called Hurst of the University of Bath in England, codon biases into account, says transfer RNA (tRNA). Each tRNA molecule but codon selection is still sometimes detect- Stephen J. Freeland of the recognizes a particular codon in a gene tran- able. Hence, codon usage could affect the University of Maryland. Because codon usage can affect the script and delivers a corresponding amino accuracy of gene mutation rates estimated microbes’ effi ciency, Freeland’s acid to the ribosome assembling a protein with the assumption that synonymous co- group is developing a database of chain. When a rarer codon appears, fewer dons are “silent” in protein terms. Acknowl- synthetic genes and related matching tRNAs are available in the cell, and edging the various mechanisms by which software to aid in adapting the ribosome’s operation can slow down. synonymous mutations may not always be designer genes to a target organism’s preferences. The synonymous mutation in MDR1, the neutral, Hurst wrote in a recent review, “will NCI investigators found, created a less com- be important for understanding and poten- The discovery that synonymous mon codon. Because the P-glycoprotein’s as- tially combating genetic disease.” codons could disrupt protein folding in humans illustrates why sembly slowed, it folded improperly. The Kimchi-Sarfaty, now at the U.S. Food and thinking of the genetic language same problem can occur in microorganisms, Drug Administration, has already moved on as a “digital” one is too simplistic, such as Escherichia coli and yeast, when they to exploring how polymorphisms and careful Freeland says. “The idea that an attempt to produce a protein from an insert- gene design could lead to better versions of RNA codon such as UUU ‘means’ ed gene containing codons the organisms human blood coagulation proteins. “People phenylalanine just like 010101 ‘means’ the letter A in ASCII,” he rarely employ. were defi nitely ignoring ‘silent’ mutations,” adds, is one that 21st-century But researchers have also documented she says, but she will no longer assume that biologists “have to unlearn.” cases in which the use of an uncommon co- “synonymous” translates into silence.

CONSUMER CONSUMER ELECTRONICS

I I Motion Pictures HDTV BROADCASTERS PLAN TO GET INTO MOBILE VIDEO BY MICHAEL ANTONOFF

ven as high-definition televisions are sidered supplying video to nomadic subscrib- eclipsing entire walls in homes, tiny ers its next big opportunity after voice calls E screens are proliferating on cell phones, and text messaging, so wireless carriers have pocket PCs and portable media players. The been upgrading their systems to third-gen- telecommunications industry has long con- eration, high-speed networks for that pur-

24 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. problem in stationary receivers, but recep- news tion on a TV in a vehicle zipping down the SCAN left lane presents a greater diffi culty. As a solution, engineers have proposed an advanced form of VSB called A-VSB that relies on two key components. One is the supplementary reference sequence, which helps the receiver stay locked onto the signal. The second is turbo coding, a mathematical technique for wrapping data in multiple layers of redundant error correction so that if the re- TV TO GO: Mobile video systems, such as Verizon’s ceiver misses a data bit, three or four others V Cast, may face competition from broadcasters. are in line to replace it. Although A-VSB does not specify a particular resolution, engineers pose. But thanks to new transmission stan- have been experimenting with an image size dards, another industry plans to enter the of 320 by 240 pixels—good enough to profi eld, one with much more experience in vid- vide moderate-to-sharp pictures on a small eo delivery: digital TV (DTV) broadcasters. window on a notebook computer. Currently the market for mobile video In January, Samsung, German transmit- seems largely untapped, as consumers have ter manufacturer Rohde & Schwarz and a yet to become enamored with existing sys- local Sinclair Broadcasting station demon- tems. Forrester Research, a technology and strated A-VSB during the 2007 Consumer market research fi rm, found that only 1 to 3 Electronics Show in Las Vegas. The group percent of people surveyed in 2006 said they fi tted a bus with various TVs, including some would care to access video on a handheld prototype handheld models. Even as the bus device. The chief reason is cost. Verizon approached speeds of 50 miles an hour, re- Wireless subscribers, for example, pay an ception of the A-VSB picture remained unin- additional $15 per month for its so-called terrupted; in contrast, a concurrently airing V Cast service. Moreover, a quality-of-ser- DTV program repeatedly disappeared. THE DISH ON vice issue plagues conventional cellular sys- Delivering A-VSB does not come without MOBILE VIDEO tems because video hogs bandwidth: frame cost—in the demonstration, the reference se- Besides broadcast stations and rates drop when too many callers are on at quence and turbo coding ate up about 28 wireless carriers, satellite TV and once. This problem has forced cellular pro- percent of the digital transmission’s data radio operators are also well viders to build higher-power transmitters stream of 19.39 megabits a second. For the positioned to deliver mobile video. At this year’s Consumer and to use additional frequencies. broadcaster, that might mean stealing bits Electronics Show in Las Vegas, DTV stations may have some advantages. from its other program streams and possibly EchoStar Communications First, each station already has a six-mega- reducing high-defi nition quality. introduced MobileDISH, an in-car hertz chunk of spectrum for transmitting John Godfrey, vice president of govern- system that uses a special, roof- digital programs. In that chunk, broadcast- ment and public affairs for Samsung Infor- mounted antenna. Satellite radio operators Sirius and XM have ers can squeeze in a high-defi nition channel, mation Systems America, expects ATSC-led talked about providing video to plus at least one other channel at a lesser res- fi eld tests to begin in the spring. He believes their customers for several years. olution. Second, commercial TV is free, so that the cost to broadcasters of implement- A mobile video service geared subscriber resistance is a nonissue. ing A-VSB is incremental; a completed stan- toward children in the backseat The big challenge for broadcasters will dard with rollouts of TV-capable cell phones will likely be available soon: the 2008 Dodge Magnum lineup, for be adapting federally mandated DTV stan- or media players could begin in 2008. instance, sports a factory- dards for mobile applications. In the U.S., As for what would be shown in mini form, installed Sirius satellite radio the Advanced Television Systems Commit- a Sinclair offi cial said it would likely be ded- with a streaming video option. tee (ATSC) has deemed that broadcast an- icated programming, such as national cable At the Las Vegas show, XM tennas must beam signals in a format called channels unleashed from their cables. Or it demonstrated how it could deliver

vestigial sideband (VSB). But over-the-air could be local trafﬁ c reports—the better to video to a vehicle using its reception has fared poorly in environments facilitate your commute home to watch the existing satellites and ground repeaters. Even so, XM

AP Photo AP where the VSB signals reﬂ ect off buildings next episode of Lost in your living room. spokesman David Butler and other obstacles, producing multiple cop- acknowledges the challenges: ies that compromise reception. Sophisticated Michael Antonoff writes about technology “It’s expensive, and it uses a

JAE C. HONG JAE HONG C. electronics can cope with this multipath from New York City. signiﬁ cant amount of bandwidth.”

PHYSICS

SCAN I Escape from the Nucleus IONIZATION VIA QUANTUM TUNNELING OBSERVED BY ALEXANDER HELLEMANS

o break free from an atom, the negative- atto second is a billionth of a billionth of a ly charged electron typically has to ab- second.) Attosecond laser pulses have al- T sorb a high-energy photon, such as that ready made it possible to probe the motion from the ultraviolet (UV) or x-ray spectrum. of electrons in atoms and molecules, and im- The electron then gets excited enough to over- proved versions will allow researchers to come the electrostatic attraction holding it to track electron movements that occur, for in- the positively charged nucleus and escapes, a stance, during chemical reactions. process called ionization. A German-Dutch Ferenc Krausz of the Max Planck Institute team has for the fi rst time provided direct for Quantum Optics in Garching, Germany, proof of an alternative mechanism. Powerful and his team describe their ionization exper- electric fi elds from a laser pulse can momen- iment in the April 5 Nature. Targeting a gas tarily weaken the electrostatic bonds and en- of neon atoms, the group fi rst used a 250-at- able the electron to quantum-mechanically tosecond UV laser pulse to nudge one elec- tunnel away from the atom. tron farther away from the nucleus. Almost Leonid Keldysh, now at the Lebedev simultaneously the physicists fi red an infra- Physics Institute in Moscow, predicted the red pulse 5,000 attoseconds long whose elec- effect in 1964, and experiments have already tric fi eld oscillates only a few cycles. The fi eld proved that such unusual ionization can oc- weakened the electrostatic force and enabled cur. But only with the advent of laser pulses the loosened electron to tunnel out, as quan- lasting just a few hundred attoseconds can tum particles can do when confronted with a physicists observe the phenomenon. (One thin barrier. By increasing the time between

26 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Infrared 12pulse As an example, Krausz cites the “shake news Electron up” process in atoms that occurs when an SCAN energetic x-ray photon kicks out an electron close to the nucleus. While fl ying away, this Electrostatic Nucleus fi e l d electron could impart some of its energy to another electron, which would become ex- TUNNEL TO FREEDOM: When an infrared photon strikes an atom (1), its electric fi eld can distort and cited and move farther away from the nucle- thereby weaken the atom’s electrostatic fi eld on us. Hence, a small delay might exist between one side, enabling the electron to tunnel away (2). the absorption of the x-ray photon by the ejected electron and the repositioning of the the UV and infrared pulses in small steps, the second electron. The delay, Krausz remarks, researchers found that the number of neon “could be as little as 50 attoseconds; nobody ions formed rose in parallel, clearly indicating really knows.” The length of the delay is not that whenever the electric fi eld of the infrared exciting, he explains—rather the point would laser pulse reached a maximum, the rate of be whether a delay existed at all. A delay produced ions increased as well. would mean the second electron got energy Keldysh’s theory of strong-fi eld ionization from the fi rst and was not coincidentally and has become part of many other theories, and simultaneously excited by the x-ray photon. the result is “not really a dramatic surprise,” Krausz claims he has now achieved 100- Krausz admits. But critically, “the team has attosecond UV pulses, so he may soon solve shown a new way to make measurements” of that puzzle. As the lasers improve, answers to electron dynamics, comments physicist Paul other questions are sure to follow in the com- Corkum of Canada’s National Research ing years, if not in the coming atto seconds. Council in Ottawa. And the technique could probe poorly understood processes in which Alexander Hellemans is a science writer

LUCY READING-IKKANDA electrons exchange energy with one another. based in Bergerac, France.

DESIGN

I Dialing In MOBILE PHONES TARGET THE WORLD’S NONREADING POOR BY SALLY LEHRMAN

hose days are long gone when placing use cell phones. But selling to poor rural a telephone call meant simply picking areas is not likely to happen with a market- up the receiver and asking the operator ing version of “plug and play.” Most poten- Tto patch you through. Modern cell phones tial buyers have little exposure to anything require users to navigate a series of menus other than simple electronics. Reading to fi nd numbers, place calls or check mes- through a series of hierarchical menus and sages. Even the most tech-savvy pushing buttons for multiple purposes may take weeks to discover would be new concepts for such customers. some of the more arcane multi- To come up with a suitable device, Mo- media functions. Imagine the torola relied on a team of anthropologists, diffi culty for someone unable psychologists and designers to study how to read. textually illiterate villagers use their aging That is the challenge for mo- televisions, tape players and phones. The bile communications compa- researchers noticed that their subjects nies aiming to branch out into would learn each button’s dedicated func- developing countries. The pros- tion. With something more complicated, pects are alluring: according to such as an automated teller machine, users the GSM Association, a global would memorize a set of behaviors in order, Photo AP UNTAPPED POTENTIAL: Whereas some customers trade group, only about one which allowed them to move through the in India may have no problem with cell phones, nonliterate people may benefi t from a new design. third of China’s vast population machine’s basic hierarchy without having to

and about one tenth of India’s read the menu. LAKSHMAN M.

30 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. The research, which lasted three years, representing just one activity. A navigation news led Motorola to craft a cellular handset dial enables users to select the icons. SCAN slimmed down to three essential activities: Not all cell phone companies believe that calling, managing numbers and simple text a design for nonliterate users should start messaging. “A lot of the functions in a cell from scratch. Nokia’s behavioral researchers phone are not useful to anyone,” points out noticed that “newbies” rely on friends and Gabriel White, who headed the interactive relatives to help them with basic functions. design team. The icon-based interface also Rather than confronting the challenge of a required thought. Although some images completely new interface, Nokia chose to worked ﬁ ne—a clock with ringers to indicate provide some audio menus in its popular the alarm-setting function—others had to 1100 model and a preview mode so that peo- be redrawn. A stack of coins was intended to ple could try out functions without the risk indicate the remaining prepaid call value, of changing anything important. but potential users mistook the original de- Mobile phones may even become tools sign as an illustration of a pile of food. for literacy, predicts BJ Fogg, who studies The culmination of Motorola’s efforts is computer-human interaction at Stanford the low-cost Motofone, which the company University. Phones might teach the alphabet introduced to ﬁ rst-time mobile users in In- or tell a story as users read along. “Imagine dia and Pakistan late last year. At first if it eventually could understand your weak glance, the nine-millimeter-thick phone points and drill you on those,” Fogg propos- looks much like other mobiles, except with es. And soon enough, he declares, designs fewer keys and no wording. It has a numer- for illiterate users will lead to more straight- ic keypad and separate buttons for call, end, forward, elegant phones for everyone. menu and phone book. Instead of layer on layer of function lists, the menu button pulls Sally Lehrman is based in the up a set of icons on the screen, with each one San Francisco Bay Area.

MICROFLUIDICS

I Microchips with Heart NO BATTERIES NEEDED: PUMPS DRIVEN BY HEART CELLS BY TIM HORNYAK

ioengineers have eagerly ex- five millimeters in diameter that re- plored tiny mechanisms that quires no external power, wires or B move microscopic or nanoscopic stimuli. The simple device consists of a amounts of fl uid. Already used in ink- hollow, fl exible silicone sphere with jet printers, microfl uidic devices hold tubes 400 microns wide at opposite out the promise of effi cient and rapid sides—in structure, it resembles an medical assays requiring only minus- earthworm’s one-chamber heart. Rat cule sample volumes for results. Poten- heart muscle cells (cardiomyocytes) tial applications include lab-on-a-chip surround the sphere, and when they devices for on-site chemical analyses, pulsate in sync, they drive fl uid through engineered vital organs, and in vivo the tubes and the sphere. Cell nutrients pharmacies that secrete drugs when in the surrounding medium power needed. In terms of implants, one of the the cells. main problems is power. Microfl uidics “In the future, this kind of microsys- typically rely on large external gear tem may be applied to medical, clinical and electricity supplies to run actuators and biological applications,” says proj- that drive fl uids through microchan- ect leader Takehiko Kitamori of the nels. Now a group of researchers has University of Tokyo. As examples, he innovated a concept that instead relies points to “implanted chemical systems, on an old-fashioned kind of engine—a such as drug delivery systems, or an in- living heart. sulin monitoring and delivery system The team, based in Japan, has trans- on a wristwatch.” formed cultured rat heart muscle cells Kitamori has spent four years build- into a spherical, pulsating micropump ing the micropump. He collaborated with Teruo Okano of Tokyo Women’s Medical University, who has developed cellular sheets that can bind to an im- paired heart to support pumping. Forming a sphere out of the rat cells is a key technology of the research. First, the cells are cultured in a polystyrene dish coated with a heat-sensitive polymer that, when cooled, allows the cells to be harvested as a single sheet. The cell sheet, pulsating continuously, is

Heart cells Silicone polymer sphere

Glue

Capillary ) right 5 millimeters LUCY ); left

GETTING PUMPED: University of Tokyo’s Takehiko Kitamori holds a kind of “lab on a chip” (left) that could be powered by his micropump, which is fashioned from rat heart cells grown over a flexible, hollow silicone polymer ball (right). TIM HORNYAK ( HORNYAK TIM READING-IKKANDA (

32 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. then placed on a silicone sphere formed on a ball of sugar (the sugar is dissolved with water once the silicone solidifi es). The cells need only an hour to stick to the silicone because it is coated with fi bronectin, an adhesive glycoprotein involved in tissue repair and blood clotting. Epoxy glue attaches the two capillaries to the sphere. The researchers do not know exactly why the heart muscle cells start beating simultaneously once one gets going, but the harmonized motion continued in a test of the pump. The team watched polystyrene tracking particles move through the pump for fi ve days, with expected fl ow rate falling only slightly. Because replacing the culture medium seemed to cause beat- frequency fl uctuations, future applications would necessitate a constant environment. Kitamori would like to make the micropump more powerful and sophisticated—more like a fi sh heart—by adding cell layers and valves. Multiple chambers could also be incorporated by reshaping the silicone sphere. “This is chemical-energy-to-mechanical-energy conversion that can be used in the fi eld, with no electrical power,” says Kitamori, who wants to fashion chip- based model circulatory systems for direct observation of vascular and cardiac disease. He also sees humans of the future having health-monitoring chip labs inside them performing dozens of functions. The pumps running those labs would be formed from the person’s own myocardial cells, thereby elimi- nating immune rejection. In essence, such bioengineered humans would have not one heart but many. Of course, the miniature hearts would not replace the real thing. “We only need one heart for our lifetimes, and cardiac muscle is very tough,” Kitamori laughs. “That’s why we’re applying this robust actua- tor to a liquid-control system.”

BY THE NUMBERS THE BY

I Food Boom AGRICULTURAL PRODUCTIVITY SUSTAINS THE U.S. ECONOMY BY RODGER DOYLE

ne of the least appreciated but most use of fertilizers and pesticides. Later came remarkable developments of the past the adoption of hybrid seeds, genetic engi- O 60 years is the extraordinary growth of neering of plants and improved livestock American agriculture. Farming now accounts breeding. A key element was the U.S. De- AMERICAN for about one tenth of the gross domestic partment of Agriculture’s extension service. product yet employs less than 1 percent of all Operating through land-grant universities MONOCULTURE workers. It has accomplished this feat through and other organizations, it educated farm- The Food and Agriculture exceptionally high growth in productivity, ers on biotechnology, pest management and Organization (FAO) of the United which has kept prices of food low and thereby conservation. Nations advocates the contributed to rising standards of living. Fur- For many years, critics have claimed that maintenance of a permanent or semipermanent organic soil cover. thermore, the exportable surplus has kept modern agriculture is not sustainable, one To achieve this condition, it the trade deﬁ cit from reaching unsupport- of the major assertions being that it encour- recommends farming practices able levels. Agriculture not only has one of ages erosion, which will eventually wash that include little or no tillage and the highest rates of productivity growth of all away most of the topsoil. Lost topsoil, the varied crop rotation—measures industries, but this growth appears to have argument goes, is virtually irreplaceable be- that the U.S. follows to a limited extent. The FAO also insists that accelerated during the past two decades. cause it takes up to 300 years for one inch the future of agriculture depends Over the period 1948 to 2004, total farm of soil to form. Indeed, the fear of topsoil on biodiversity. The U.S., however, production went up by 166 percent. But as loss has dogged agriculture since the days of has largely continued to practice the 1930s Dust Bowl. monoculture: 82 percent of 200 But a detailed study of two large areas, harvested acreage in 2005 was 175 in four crops (corn, soybeans, Overall productivity the Southern Piedmont and the Northern hay and wheat), a rise from 150 of U.S. agriculture Mississippi Valley Loess Hills, showed that 78 percent in 1985. based on 1982 data, soil loss has dropped 125 sharply from the very high rates of the Materials costs 100 1930s. The conclusion of the analysis, which appeared in the July 14, 2000, issue of Sci- 75 FURTHER ence, was that if 1982 erosion rates contin- READING 50 ued for 100 years, crop yields would decline Capital by only 2 to 4 percent. The study attributed Agricultural Productivity in 25 expenditures the United States. Mary Ahearn, the decrease in soil erosion to the USDA, Jet Yee, Eldon Ball and Rich 0 Land use which urged farmers after World War II to Nehring. U.S. Department of adopt conservation practices such as strip Agriculture Bulletin No. 740, –25 cropping, whereby alternating rows are

January 1998. (percent) 1948 since Use or in Cost Change Labor costs –50 planted, and leaving plant residues in the U.S. Soil Erosion Rates— –75 ﬁ elds year-round to inhibit water runoff. Myth and Reality. Stanley W. 1950 1960 1970 1980 1990 2000 Despite being a robust contributor to the Trimble and Pierre Crosson Year in Science, Vol. 289, U.S. economy, modern agriculture is not pages 248–250; July 14, 2000. the chart shows, productivity per person without a dark side. Runoff of fertilizers, Productivity and Output improved so much that only one quarter as antibiotics and hormones degrade the envi- Growth in U.S. Agriculture. many hands were needed in 2004 as in 1948. ronment and can upset the local ecology. If Eldon Ball in Agricultural Furthermore, the arable land used for farm- not grown properly, genetically modiﬁ ed Resources and Environmental ing dropped by one quarter over the 56-year crops could spread their DNA to conven- Indicators. Economic Research period, and investment in heavy farm equip- tional species. And the industrial approach Service, U.S. Department of Agriculture, 2006. Available at ment and other capital expenditures de- to food has contributed to America’s obe- www.ers.usda.gov/ creased by 12 percent. sity epidemic as well as to sporadic but publications/arei/eib16/ Several developments drove these chang- widespread Escherichia coli outbreaks. Food and Agriculture Organization es, beginning with the replacement of the of the United Nations: remaining horses by tractors immediately Rodger Doyle can be reached at

www.fao.org after World War II and with the expanding [email protected] DOYLE; OVERALL SOURCE: AGRICULTURE; RODGER DEPARTMENT U.S. OF PRODUCTIVITY TAKES ACCOUNT INTO MATERIALS AND ENERGY LAND, CAPITAL, LABOR, OF CONTRIBUTION THE

Rats have a complex form of thought once learned to classify as short (lasting two to known only in humans and other primates— 3.6 seconds) or long (4.4 to eight seconds). specifi cally, metacognition, or the ability to The more diffi cult it was to gauge the sound mull over what one knows. Scientists at the length—for instance, if the noise lasted 4.4 University of Georgia gave rats a choice to seconds—the more often rodents bailed out either take a test or not. Passing the test re- from the test. The fi ndings, in the March 20 sulted in a large food reward, failure yielded Current Biology, suggest the rats could judge nothing, and opting out altogether led to a whether they would pass or fail, hinting that small reward. The test played noises of vari- metacognition may be more widespread ous durations that the rats had previously than thought before. — Charles Q. Choi

PATTERNS Medieval Quasicrystals

In designing architectural structures, medieval Islamic artisans hit on so-called quasicrystals, a complex pattern DATA POINTS: made famous by renowned mathematician Roger Penrose BIG BANG OF BITS in the 1970s. After seeing hints of the pattern while traveling in Uzbekistan, Peter J. Lu, a physics graduate student at Har- The expansion of the digital vard University, pored over photographs from Iran, Iraq, Tur- universe is accelerating, key and Afghanistan. Some of the ornate tile work, called according to a study by IDC, an information and technology girih, could have only been accurately constructed using a analysis fi rm, and EMC set of fi ve tiles—a bow tie, pentagon, diamond, elongated Corporation, a data management hexagon and large decagon—conclude Lu and co-author company. Powering this growth is Paul J. Steinhardt of Princeton University. The tiling the conversion of analog images, practice reached great sophistication with the Darb-i television broadcasts and voice calls to digital formats, Imam shrine in Iran, which dates to 1453; it displays a as well as the storing and symmetric pattern of pentagons and 10-sided stars. If sharing of such fi les. extended indefi nitely in all directions, this pattern would never repeat itself—the hallmark of a quasi- TILE WORK on this 15th-century Turkish Number of bits in a byte: 8 mosque exhibits a quasicrystal nature, in crystal. The researchers describe their conclusions Number of gigabytes which the pattern never repeats. in the February 23 Science. —JR Minkel in an exabyte: 1 billion Exabytes of data created, PHYSICS captured and copied in 2006: 161 Equivalent data in stacks of books stretching from the earth Spooky Action at a Greater Distance to the sun: 12 The quantum link called entanglement keeps of La Palma and then fi red one member of Exabytes expected to be made in: getting longer—in the latest demonstration, each pair to a telescope on Tenerife, 144 kilo- 2007: 255 the distance spanned two of Spain’s Canary meters away. That distance is 10 times farther 2010: 988 Islands. When two photons are entangled, than entangled photons have ever flown what happens to one instantaneously deter- through the air. Such photons might prove Available storage capacity today, mines the fate of the other, no matter how far suitable for sending scrambled messages that in exabytes: 246 apart they are. Using a laser, Anton Zeilinger cannot be decoded or secretly read. The team Number of bits stored on one of the University of Vienna and his team cre- discussed its feat at the March meeting of the square inch of a disk drive in: ated entangled pairs of photons on the island American Physical Society. —JR Minkel 1956: 2,000 2007: 100 billion Extended coverage of these News Scan briefs can be found at SOURCE: The Expanding Digital Universe, sa www.sciam.com/ontheweb

IDC White Paper, March 2007 COLLINS MATT BY ILLUSTRATION DENNY; B. W. OF COURTESY

Researchers have created a coating that refl ects almost no light, thanks to a refractive index close to that of air. A material’s index of refraction refers to the speed of light in that substance, which in turn governs how much light bends when it passes into the substance. The new material consists of angled nanorods laid on top of a transparent semiconductor wafer. By stacking fi ve layers of nanorods, each one less porous than the one below it, the researchers could vary the coating’s refractive index in a series of steps, going from 2.03 at the bottom (closely matching the wafer) to 1.05 at the top (compared with 1.0 for air). Total refl ection dropped to as little as 0.1 percent, note E. Fred Schubert and his group at Rens- LAYERED NANORODS, each with a selaer Polytechnic Institute in the March Nature Photo- different refractive index, nearly nics. The coating should prove useful in boosting the eliminate glare at several wavelengths. effi ciency of LEDs and solar cells. —JR Minkel

MEMORY BRIEF Sweet (Scented) Dreams POINTS

A night on a bed of roses could sharpen your to the rose fragrance, because odors are well ■ Prolactin, a hormone that memory. The purpose of sleep remains a known for bringing back memories. Volun- increases during pregnancy, mystery, but one idea is that slumber replays teers performed better at the game when regenerates myelin destroyed in mouse models of multiple new experiences to facilitate their inclusion given the scent during slow-wave sleep, often sclerosis (MS). The results may in long-term memory storage. To explore considered the deepest stage of sleep. Brain explain why symptoms of this notion, volunteers in a rose-scented imaging during slow-wave sleep showed that MS sometimes abate room played a computer version of the card the scent cue activated the hippocampus, a during pregnancy. game Memory, where they had to remember region linked with memory. Scientists at the Journal of Neuroscience, locations of paired cards that ﬂ ashed on the University of Lübeck in Germany and their February 21

screen for a few seconds at a time. Later, colleagues present their findings in the ■ T. rex and other meat-eating while they slept, researchers exposed them March 9 Science. —Charles Q. Choi dinosaurs had small genomes, on par with modern birds, which METEOROLOGY have about 1.45 billion bases (humans have about three

) billion). The fi nding counters a Drought-Making Pollutants hypothesis that birds evolved bottom and maintained smaller genomes Tiny particles resulting from fuel burning appear to be inter- to conserve energy for fl ight. fering with the formation of rain, snow and ice on one of Chi- Nature, March 8 ); KAI LIU ( KAI LIU ); na’s mountains. As aerosols increase in the atmosphere, fewer top ■ Out-of-body sensations may ( droplets collide with one another to form raindrops. Atmo- stem from a disturbance in the spheric scientists studied records from the past 50 years at a brain’s arousal system, in which meteor ological observatory that sits at the peak of sacred it would fail to make a proper transition from REM sleep Mount Hua. They not only found that visibility from the two- (when dreams often occur) kilometer-high summit has, on average, declined from rough- to wakefulness. ly 30 to 10 kilometers, but they also showed that rainfall has Neurology, March 6 dropped by as much as 17 percent compared with the precipitation in neighboring areas. The fi ndings explain the widely ■ Asteroid surfaces that have

Rensselaer Polytechnic Institute Institute Polytechnic Rensselaer absorbed warming sunlight can — — observed trend from Canada to South Africa to Israel of a reemit the heat when they move decrease in highland precipitation compared with that in ad- into the shade, thereby creating jacent lowlands. Pollution is not just obscuring the view; it is DRY TIMES: Haze not only recoil and affecting an shrouds the view on Mount also choking the mountain streams. The researchers report asteroid’s rotation. Hua but also discourages rain.

E. FRED SCHUBERT SCHUBERT FRED E. their work in the March 9 Science. —David Biello Science Express, March 8

SCIENTIFIC AMERICAN Digital Skeptic

Bush’s Mistake and Kennedy’s Error Self-deception proves itself to be more powerful than deception By MICHAEL SHERMER

The war in Iraq is now four years old. It has cost more than 3,000 Wrongly convicting people and sentencing them to death American lives and has run up a tab of $200 million a day, or is a supreme source of cognitive dissonance. Since 1992 the $73 billion a year, since it began. That’s a substantial invest- Innocence Project has exonerated 192 people total, 14 from ment. No wonder most members of Congress from both parties, death row. “If we reviewed prison sentences with the same along with President George W. Bush, believe that we have to level of care that we devote to death sentences,” says Univer- “stay the course” and not just “cut and run.” As Bush explained sity of Michigan law professor Samuel R. Gross, “there would in a speech delivered on July 4, 2006, at Fort Bragg, N.C.: “I’m have been over 28,500 non-death-row exonerations in the not going to allow the sacrifi ce of 2,527 troops who have died past 15 years....” What is the self-justifi cation for reducing in Iraq to be in vain by pulling out before the job is done.” this form of dissonance? “You get in the system, and you be- We all make similarly irrational arguments about deci- come very cynical,” explains Northwestern University legal sions in our lives: we hang on to losing journalist Rob Warden. “People are lying to stocks, unprofitable investments, failing What happens when you all over the place. Then you develop a businesses and unsuccessful relationships. If theory of the crime, and it leads to what we we were rational, we would just compute the someone says, call tunnel vision. Years later overwhelming odds of succeeding from this point forward “I was wrong”? evidence comes out that the guy was inno- and then decide if the investment warrants cent. And you’re sitting there thinking, the potential payoff. But we are not rational—not in love or ‘Wait a minute. Either this overwhelming evidence is wrong, war or business—and this particular irrationality is what or I was wrong—and I couldn’t have been wrong, because I’m economists call the “sunk-cost fallacy.” a good guy.’ That’s a psychological phenomenon I have seen The psychology underneath this and other cognitive falla- over and over.” cies is brilliantly illuminated by psychologist Carol Tavris and What happens in those rare instances when someone says, University of California, Santa Cruz, psychology professor “I was wrong”? Surprisingly, forgiveness is granted and re- Elliot Aronson in their book Mistakes Were Made (But Not spect is elevated. Imagine what would happen if George W. by Me) (Harcourt, 2007). Tavris and Aronson focus on so- Bush delivered the following speech: called self-justifi cation, which “allows people to convince This administration intends to be candid about its errors. themselves that what they did was the best thing they could For as a wise man once said, “An error does not become have done.” The passive voice of the telling phrase “mistakes a mistake until you refuse to correct it.” We intend to ac- were made” shows the rationalization process at work. “Mis- cept full responsibility for our errors.... We’re not going takes were quite possibly made by the administrations in to have any search for scapegoats ... the fi nal responsi- which I served,” confessed Henry Kissinger about Vietnam, bilities of any failure are mine, and mine alone. Cambodia and South America. The engine driving self-justifi cation is cognitive disso- Bush’s popularity would skyrocket, and respect for his nance: “a state of tension that occurs whenever a person holds ability as a thoughtful leader willing to change his mind in the two cognitions (ideas, attitudes, beliefs, opinions) that are teeth of new evidence would soar. That is precisely what hap- psychologically inconsistent,” Tavris and Aronson explain. pened to President John F. Kennedy after the botched Bay of “Dissonance produces mental discomfort, ranging from minor Pigs invasion of Cuba, when he spoke these very words. pangs to deep anguish; people don’t rest easy until they fi nd a way to reduce it.” It is in that process of reducing dissonance Michael Shermer is publisher of Skeptic (www.skeptic.com).

that the self-justiﬁ cation accelerator is throttled up. His latest book is Why Darwin Matters. BRAD HINES

The Road to Clean Energy Starts Here Realizing crucial energy technologies will take more than just research and development By JEFFREY D. SACHS

The key to solving the climate change crisis is technology. To however, will require an aggressive global technology policy. accommodate the economic aspirations of the more than fi ve First, there will have to be market incentives to avoid emissions, billion people in the developing countries, the size of the world in the form of either tradable permits or levies. A reasonable economy should increase by a factor of four to six by 2050; at levy might be $25 per ton of emitted carbon dioxide, intro- the same time, global emissions of greenhouse gases will have duced gradually over the next 10 to 20 years. Second, there will to remain steady or decline to prevent dangerous changes to the have to be ample government support for rapid technological climate. After 2050, emissions will have to drop further, near- change. Patents can help spur private market research and de- ly to zero, for greenhouse gas concentrations to stabilize. velopment (R&D), but public funding is required for basic sci- The overarching challenge is to make that transition at ence as well as for the public demonstration and the global minimum cost and without economic disruption. Energy- diffusion of new technologies. In sum, we need a strategy saving technologies will play a pivotal role. sometimes described as RDD&D. Buildings can save energy at low capital cost, Some will impose In the past two years, the Earth Institute at and often net overall savings, through im- real added costs; Columbia University has hosted a Global proved insulation, effi cient illumination and others will pay Roundtable on Climate Change, involving lead- the use of heat pumps rather than home fur- for themselves. ing corporations from around the world. These naces. Automobiles could, over time, reach 100 companies, including many of the largest pow- miles per gallon by a shift to plug-in hybrids, better batteries, er producers, are ready to reduce carbon emissions. They know lighter frames and other strategies. Of course, technologies that CCS must be a high priority. A new Global Roundtable such as heat pumps and plug-in hybrids partly reduce direct Task Force on CCS seeks to promote the required RDD&D. emissions by shifting from on-site combustion to electricity, so Fortunately, the European Union has already pledged to that low-emission power plants become paramount. build at least a dozen CCS demonstration projects in Europe Low-emission electricity generation will be achieved in by 2015. But we will also need such centers in the U.S., China, part through niche sources such as wind and biofuels. Larger- India, Australia, Indonesia and other highly signifi cant coal- scale solutions will come from nuclear and solar power. Yet power producers. In the low-income countries, this will re- clean coal will be essential. New combustion techniques, com- quire a few billion dollars; that is where the RDD&D invest- bined with carbon capture and sequestration (CCS), offer the ments of the high-income countries will be essential. prospect of low- or zero-emission coal-fi red thermal plants. The CCS Task Force aims to break ground on one or more The incremental costs of CCS may well be as low as one to demonstration plants by 2010 in every major coal region. By three cents per kilowatt-hour. 2015 this crucial technology can be proved and added to the All these technologies are achievable. Some will impose bid to avert climate disaster. This model of RDD&D won’t real added costs; others will pay for themselves as lower en- stop there. Harnessing technology to achieve sustainable energy bills offset higher capital outlays. Some estimates suggest ergy will involve much of the global economy for decades. that, as of 2050, the world will have to negate around 30 billion tons of carbon dioxide emissions a year at a cost of rough- Jeffrey D. Sachs is director of the Earth Institute at ly $25 per ton, or $750 billion annually. But with a world Columbia University (www.earth.columbia.edu). economy by then of perhaps $200 trillion, the cost would be well under 1 percent of world income and perhaps under 0.5 An expanded version of this essay is available at percent, a true bargain compared with the costs of inaction. sa www.sciam.com/ontheweb

BRUCE GILBERT/EARTH INSTITUTE GILBERT/EARTH BRUCE Achieving these technological solutions on a large scale,

Prime Directive for the Last Americans Saving Amazonia’s indigenous peoples means not meeting them, insists Sydney Possuelo— a policy of noninterference he hopes to extend, even if others hate it BY CLAUDIO ANGELO

The gray-bearded, balding man sips coffee in the kitch- Counting Possuelo, only five members remain. en of his apartment in Brasília, Brazil. With a blank They travel to the remotest corners of the Amazon rain stare, he ponders the future of the three things to which forest to track down isolated tribes and contact them he has dedicated his life: “Everything dies at its own before miners, cattle ranchers, loggers and others do, time. The forest dies, with it die the Indians, with them usually at gunpoint. Their job is to divert development die the sertanistas.” But at 67, Sydney Ferreira Pos- around those tribes, so as to keep them from sharing suelo acknowledges that the sertanistas, men who the usual bleak fate of indigenous Brazilians. But it has make a living out of protecting the isolated indigenous been 20 years since the last sertanista was hired. And peoples in the Brazilian jungles, may be the fi rst to go. most of the old school—that is, people who have actual contact experience—have either died or retired. Possuelo has done neither, although he was fi red from his position with Brazil’s National Foundation for Indigenous Peoples (FUNAI) last year after pub- licly criticizing the remarks of its president, who stated that there might be too much land already in the hands of native Brazilians. Today he coordinates efforts through the Instituto Indigenista Interamerica- no, a nongovernmental organization that is part of a multinational alliance dedicated to the protection of isolated tribes. Possuelo had helped initiate a sea change in federal policy toward indigenous tribes by persuading the government that the natives should be left alone, to live exactly as they have since prehistoric times. The decision to make contact should be left to them. The prevailing wisdom throughout much of the 20th century and before had been “integration” into society. “The idea was that we should go fetch those peoples and share with them the benefi ts of civilization, since civilization is a good of all humankind that belongs to non-Indians and Indians alike,” Possuelo says. But such a bright positivist view confl icted with SYDNEY POSSUELO: NATIVE PROTECTION reality. After contact, many natives lost their lives to ■ Works to protect the way of life of isolated indigenous groups in the infl uenza and other diseases to which they had no Amazon rain forest. (Possuelo is pictured above with a Korubo woman.) natural defense. Not a few were expelled from their National Geographic Image Collection ■ Perhaps four million indigenous people in 1,000 ethnic groups once lived traditional territories to make room for highways, in Brazil; today they number fewer than 734,000 in 220 tribes. dams and cattle ranches. Many were murdered, in ■ On being fired from his state job: “I have never been one of those nice some cases with approval of government offi cials. But

workers who agree with everything that the government does.” perhaps worst of all was the cultural dislocation. REYNARD NICOLAS

40 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. When meeting a new group, Pos- rubo group deeper in the jungle, feel- suelo recounts, “you must keep an eye ing that they are remote enough that on the folks who are there. The guys they will not meet outsiders. fi ght you with bows and arrows, they The noninterference strategy has kill you, they speak up to you, they as- gained worldwide approval. The Unit- sault you.” Contact, however, changes LAST OF A KIND: Possuelo meets in peace with the ed Nations Declaration on the Rights all that: “One year later they are slack, Korubo, who have been known to kill intruders. of Indigenous Peoples, adopted in emaciated, bowing their heads and 2006, incorporates the right to nonas- begging for food and money by the roadside.” A case in point: similation. Despite that, the policy faces critics in Brazil, who the Arara people, who had been called the scourge of the Trans- see such protection as an additional burden to taxpayers that amazônica Highway for their attacks on work crews but today only postpones inevitable integration. “Of course, you’re buy- live as poor peasants fully dependent on federal aid. “You ing time,” says Márcio Santilli, head of Brasília’s Socio-Envi- break down their health, their mythical universe, their work ronmental Institute and a former president of FUNAI. “But and their education system. They become outcasts, and many buying time is central to mitigating the impacts of contact.” of them have been outcasts for 500 years,” Possuelo says. And time is running out for those Indians. Brazil has demar- The veteran sertanista’s career began in 1959, as a disciple cated less than a dozen lands with isolated peoples. Possuelo of Orlando and Cláudio Villas Bôas, two of the three brothers and his colleagues have identifi ed 22 isolated groups in the Am- credited with saving 15 tribes by creating the Xingu Indige- azon, although FUNAI believes that the true number might be nous Park, Brazil’s fi rst megareservation. With them, he jour- closer to 68, based on aerial surveys and witness accounts. neyed into the Amazon for the fi rst time in the early 1960s. In Some live in previously protected indigenous lands, some in the 1970s he joined FUNAI and made his fi rst solo contact areas that are right in the range of agricultural expansion. with an isolated group, the Awá-Guajá of Maranhão. In the other six South American countries with isolated By 1987 Possuelo decided to take the Villas Bôas strategy peoples—Ecuador, Peru, Bolivia, Paraguay, Colombia and further, because even peaceful contact with such groups often Venezuela—the prospect is even bleaker. “The issue is all but destroyed their native culture and self-suffi ciency. His idea: to unknown to them,” Possuelo states. “And it is no use to un- avoid contact altogether. He convinced the government that it dertake protection policies here if indigenous peoples are not was more practical to demarcate indigenous lands and to protected across the border and get killed on the other side.” guard the borders with armed agents than to provide for them In 2005, before leaving FUNAI, Possuelo organized a indefi nitely. “We do it for threatened animal species, why can’t meeting in the Amazon city of Belém that gave birth to an we do it for a unique ethnic group that has been there for thou- alliance for the protection of isolated peoples in the seven sands of years?” he asks. “That’s good for the Indians, who countries. The alliance, consisting of state attorneys, environ- will see their land and their tribal traditions respected, and mental defense groups, anthropologists and indigenous orga- good for the state, who will no longer be charged with geno- nizations, called on their governments to identify and protect cide.” As an added bonus, a study found that legally recog- their indigenous charges. “The idea here is to raise public nized indigenous lands in the Amazon had lower average de- consciousness in the fi rst place,” he says. “Then we’ll be able forestation rates (1.14 percent) than federal conservation units to talk about public policy.” The alliance has been facing an (1.47 percent). unexpected challenge, though—from indigenous peoples Executing Possuelo’s plan, however, means identifying the themselves. “In Ecuador last year, contacted Indians killed indigenous groups before miners, loggers and others fi nd 20 isolated Indians,” says Marcelo dos Santos, who took over them—encounters that could lead to long-lasting, deadly skir- Possuelo’s old FUNAI post. mishes, as in the case of the Korubo people of the Javari Valley, Possuelo agrees that the isolated groups may not last much which lies near the Peruvian border. First contacted in 1972, longer. But pointing at several fi gurines of Don Quixote that they violently resisted intruders and over the decades had he keeps in his living room, he says he has not given up. “The clubbed to death many who ventured into their territories. In fate of those peoples will depend ultimately on our choices,” 1996 Possuelo managed to establish peaceful encounters with he argues. “Those peoples are the last Americans. We are in- one Korubo clan, in part by singing aloud (enemies come in debted to them.” silence) and by offering gifts. Despite having known the clan for many years now, Possuelo still considers them unpredict- Claudio Angelo is science editor of the Brazilian daily able; he has kept anthropologists away, he has said, for their newspaper Folha de S. Paulo. A Q&A version of his own safety. He himself will not bother to contact a larger Ko- interview with Possuelo is at www.sciam.com/ontheweb www.sciam.com SCIENTIFIC AMERICAN 41 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. The Mystery of Methane on Mars & Titan By Sushil K. Atreya

MARS has long been thought of as a possible abode of life. The discovery of methane in its atmosphere has rekindled those visions. The visible face of Mars looks nearly static, apart from a few wispy clouds (white). But the methane hints at a beehive of biological or geochemical activity underground.

Of all the planets in the solar system other than Earth, own way, revealing either that we are not alone in the universe Mars has arguably the greatest potential for life, either extinct or that both Mars and Titan harbor large underground bodies or extant. It resembles Earth in so many ways: its formation of water together with unexpected levels of geochemical activ- process, its early climate history, its reservoirs of water, its vol- ity. Understanding the origin and fate of methane on these bod- canoes and other geologic processes. Microorganisms would ﬁ t ies will provide crucial clues to the processes that shape the right in. Another planetary body, Saturn’s largest moon Titan, formation, evolution and habitability of terrestrial worlds in also routinely comes up in discussions of extraterrestrial biology. this solar system and possibly in others.

In its primordial past, Titan possessed conditions conducive to Methane (CH4) is abundant on the giant planets—Jupiter, the formation of molecular precursors of life, and some scientists Saturn, Uranus and Neptune—where it was the product of chem- believe it may have been alive then and might even be alive now. ical processing of primordial solar nebula material. On Earth, To add intrigue to these possibilities, astronomers studying though, methane is special. Of the 1,750 parts per billion by vol- both these worlds have detected a gas that is often associated ume (ppbv) of methane in Earth’s atmosphere, 90 to 95 percent with living things: methane. It exists in small but signifi cant is biological in origin. Grass-eating ungulates such as cows, goats quantities on Mars, and Titan is literally awash with it. A bio- and yaks belch out one fi fth of the annual global methane release; logical source is at least as plausible as a geologic one, for Mars the gas is a metabolic by-product of the bacteria in their guts. if not for Titan. Either explanation would be fascinating in its Other signifi cant sources include termites, rice paddies, swamps, COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. It might mean life, it might mean unusual geologic activity; whichever it is, the presence of methane in the atmospheres of Mars and Titan is one of the most tantalizing puzzles in our solar system

TITAN, technically a satellite of Saturn but for all intents and purposes a full-ﬂ edged planet, has a nitrogen atmosphere denser than Earth’s and a surface sculpted by tectonic activity and rivers of liquid methane. Where the methane comes from, no one knows for sure. The Cassini space probe took this composite infrared image last year.

COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. SNAPSHOTS FROM TITAN AND MARS leakage of natural gas (itself a result of past life) and photosyn- thetic plants [see “Methane, Plants and Climate Change,” by Astronomers discovered methane in Titan’s atmosphere in Frank Keppler and Thomas Röckmann; Scientiﬁ c Ameri- the 1940s, but dense haze blocked their view of the surface. can, February 2007]. Volcanoes contribute less than 0.2 per- The Cassini-Huygens mission has now seen the extent to which methane shapes the terrain. cent of the total methane budget on Earth, and even they may simply be venting methane produced by organisms in the past. TITAN Abiotic sources such as industrial processes are comparatively minor. Thus, detection of methane on another Earth-like object naturally raises the prospect of life on that body. Methane In the Air lakes that is what happened with Mars in 2003 and 2004, when three independent groups of scientists an- nounced the discovery of methane in the atmosphere of that planet. Using a high-resolution spectrograph at the In- frared Telescope Facility in Hawaii and at the Gemini South Telescope in Chile, a team led by Michael Mumma of the NASA Goddard Space Flight Center detected methane concentrations in excess of 250 ppbv, varying over the planet and perhaps over time. Vittorio Formisano of the Institute of Physics and Inter- planetary Science in Rome and his colleagues (including me) analyzed thousands of infrared spectra collected by the Mars Express orbiter. We found methane to be much less abundant, ranging from zero to about 35 ppbv, with a planetary average of approximately 10 ppbv. Finally, Vladimir Krasnopolsky of the Catholic University of America and his colleagues, using the Canada-France-Hawaii Telescope, measured a planetary average of about 10 ppbv. They could not determine the variation 140 KILOMETERS over the planet because of poor signal and spatial resolution. LARGE BODIES of liquid, probably methane, show up in these Cassini radar images of Titan’s far northern hemisphere. Liquid appears dark Mumma’s team is now reanalyzing its data to try to deter- (colorized blue) for the same reason that a wet road looks dark when mine why its value is the outlier. For now, I will take the 10 you are driving at night: the smooth liquid surface reﬂ ects the ppbv value as the most likely. It corresponds to a concen- headlight beams away from your eye. Conversely, dry, rough terrain appears bright (tan). This image has a resolution of 500 meters. tration of methane (in molecules per unit volume) that is only

RIVER CHANNELS may have been Overview/Methane carved by liquid methane fl owing from a series of ridges (about 200 ■ Astronomers have been talking about life on Mars for meters high) down to a lakebed (now a century or longer, but seldom with the benefi t of hard dry). The pattern of tributaries suggests that the methane came data. That situation changed in 2003 with the discovery of from rainfall. The Huygens probe River methane in the atmosphere. Some ongoing process must captured this image from an altitude channel pump it out to offset its steady destruction by sunlight. of 6.5 kilometers as it descended through the atmosphere. ■ Researchers have narrowed the possibilities to two. The fi rst is Martians—specifi cally, methane-belching bacteria like those in cows’ guts on Earth. The second is a rock- water reaction called serpentinization, which occurs in 40 millionths of the concentration in Earth’s atmosphere. the black smokers on Earth’s seafl oor. The latter possibil- Nevertheless, even the barest presence of the gas demands an ity may seem like a letdown but would be a dramatic dis- explanation. covery in its own right. A new rover scheduled for launch Although astronomers detected methane on Titan as early in 2009 may be able to settle the issue. as 1944, it was only the additional discovery of nitrogen 36 ■ A similar debate swirls around Saturn’s largest satellite, years later that generated the immense interest in this cold and Titan. In 2005 the Huygens space probe showed that distant moon [see “Titan,” by Tobias Owen; Scientific methane plays much the same role there as water does American, February 1982]. Nitrogen is a key constituent of on Earth. The methane may come from geochemical biological molecules such as amino acids and nucleic acids. A reactions in a vast hidden ocean. body with a nitrogen-methane atmosphere, where the ground- level pressure is one and a half times that of our home planet,

44 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. On the Red Planet, methane exists only at the parts-per-billion- by-volume (ppbv) level, so it cannot be seen as directly as on Titan. Various observed processes may destroy or create it.

Haze ) haze

HAZE in the upper atmosphere of Titan consists of hydrocarbons created when sunlight strikes methane. The ); ESA/NASA/JPL/UNIVERSITY OF ARIZONA ARIZONA OF ESA/NASA/JPL/UNIVERSITY ); haze resembles urban smog. This

image has a resolution of 700 meters. lakes

MARS ); NASA/JPL/SPACE); SCIENCE INSTITUTE ( DUST DEVILS, such as this one

spied by the Mars Global NASA/JPL/USGS); (

TITAN’S SURFACE had never been Surveyor on May 21, 2002, rub devil dust

Dust devil ridge

dust grains together and thus and glimpsed before the Huygens probe ascending the create strong static-electric and landed in January 2005. It may look like side of a crater a boring fi eld of rocks, but the “rocks” are fi e l d s —which can tear apart in fact fi st-size chunks of ice and, water molecules and lead to the on close examination, show signs of production of methane- erosion by fl owing liquid—probably destroying peroxides. methane. As the probe warmed the soil, methane oozed out. 3.1 KILOMETERS

UNDERGROUND WATER FLOW METHANE DATA EARTH MARS TITAN in the Martian past would explain this whitish ridge Atmospheric concentration 1,750 ppbv 10 ppbv 5 percent (arrow) recently seen by the Mars Reconnaissance Molecular lifetime in pages preceding on and page this on globe Mars 10 600 10 million Orbiter. Water ﬂ owing atmosphere (years) through a rock fracture would have deposited Titan globe on opposite page and on preceding pages Production rate needed to 515 million 125 25 million minerals, much as hard sustain constant amount water causes buildup in (tons per year) a household pipe. When surrounding rock eroded ( NASA/JPL/MSSS ); Main sources Cattle, Bacteria? Rock-water away, the minerals termites, Rock-water reactions remained, leaving a ridge. swamps, reactions in under- Underground water might

rice paddies, in aquifer? ground also have facilitated surface Titan natural gas ocean

1 KILOMETER methane production. and river river NASA/JPL/UNIVERSITY OF ARIZONA ( ARIZONA NASA/JPL/UNIVERSITY OF (

may have the right ingredients for molecular precursors of life dissipate. The methane rain that seems to have carved its sur- and, some have speculated, even life itself to form. face would stop. Lakes, puddles and streams would dry up. Methane plays a central, controlling role in maintaining And, with its veil lifted, Titan’s stark surface would lay bare Titan’s thick nitrogen atmosphere. It is the source of hydrocar- and readily accessible to telescopes on Earth. Titan would lose bon hazes, which absorb solar infrared radiation and warm the its mystique and turn into just another satellite with thin air. stratosphere by approximately 100 degrees Celsius, and of hy- Could it be that methane on Mars and Titan has a biolog- drogen, whose molecular collisions result in a 20-degree warm- ical origin, as on Earth, or does it have another explanation, ing in the troposphere. If the methane ever ran out, tempera- such as volcanoes or impacts of comets and meteorites? Our tures would drop, nitrogen gas would condense into liquid understanding of geophysical, chemical and biological pro- droplets and the atmosphere would collapse. Titan’s special cesses has helped narrow the ﬁ eld of possible sources on Mars, character would change forever. Its smog and clouds would and many of the same arguments apply to Titan as well.

www.sciam.com SCIENTIFIC AMERICAN 45 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Split by Sunlight METHANE ON MARS the first step to answering the question is to determine the rate at which methane must be produced or delivered. That, By all rights, Mars should have zero methane. The gas is quickly in turn, depends on how fast the gas is being removed from the cleansed from the air by chemical reactions driven by sunlight or atmosphere. At altitudes of 60 kilometers and higher above the weather patterns, and known geologic and astronomical processes Martian surface, solar ultraviolet radiation splits methane mol- cannot replenish it fast enough. Thus, the methane hints at unseen ecules apart. Lower in the atmosphere, oxygen atoms and hy- activity, such as black smokers or methane-creating microbes droxyl radicals (OH), which form when water molecules are swimming in underground bodies of water. broken apart by ultraviolet photons, oxidize methane. Without being resupplied, methane would gradually disappear from the SPACE atmosphere. The “lifetime” of methane—deﬁ ned as the time it METHANE DESTRUCTION takes for the gas concentration to drop by a factor of the mathematical constant e, or roughly three—is 300 to 600 years, depending on the amount of water vapor, which undergoes season- Ultraviolet photon al changes, and on the strength of solar radiation, which varies during the solar cycle. On Earth, similar processes give methane a lifetime of about 10 years. On Titan, where solar ultraviolet radiation is much weaker and oxygen-bearing molecules are Methane Ethane substantially less abundant, methane can last 10 million to 100 Photochemical reactions million years (which is still a short time in geologic terms). occur mainly above 60 kilometers Methane’s lifetime on Mars is long enough for winds and diffusion to mix the gas into the atmosphere fairly uniformly. Thus, the observed variations of methane levels over the planet ATMOSPHERE are puzzling. They may be a sign that the gas comes from localized sources or disappears into localized sinks. One possible sink is chemically reactive soil, which could accelerate the loss of methane. If such additional sinks operated, it would take Water Methane Formaldehyde an even larger source to maintain the observed abundance. The next step is to consider potential scenarios for forming Oxidation occurs in lower atmosphere methane. The Red Planet is a good place to start because its methane abundance is so low. If a mechanism cannot explain even this small amount, it would be unlikely to account for Ti- tan’s much greater quantity. For a 600-year lifetime, a little over 100 metric tons of methane would have to be produced each year to maintain a constant global average of 10 ppbv. That is about a quarter-millionth the production rate on Earth. As on Earth, volcanoes are most likely not responsible. Martian volcanoes have been extinct for hundreds of millions Dust Hydrogen devil peroxide Methane Formaldehyde of years. Furthermore, if a volcano had been responsible for Electrochemical reactions the methane, it would also have pumped out enormous quan- are driven by dust devils and wind tities of sulfur dioxide, and Mars’s atmosphere is devoid of SURFACE sulfur compounds. Extraplanetary contributions also appear minimal. Some 2,000 tons of micrometeoritic dust are esti- SUBSURFACE mated to reach the Martian surface every year. Less than 1 percent of their mass is carbon, and even this material is largely oxidized and hence an insigniﬁ cant source of methane. Comets are about 1 percent methane by weight, but they strike Mars only once every 60 million years on average. Thus, the amount of methane delivered would be about one ton a year, AQUIFER or less than 1 percent of the required amount. Could it be that a comet struck Mars in the recent past? It could have delivered a large amount of methane, and over time the abundance in the atmosphere would have declined to its present value. An impact of a comet 200 meters in diameter 100 years ago, or a comet 500 meters in diameter 2,000 years DEEP CRUST/ MANTLE

Winds should mix methane Meteoritic dust uniformly throughout contributes a negligible atmosphere, so amount of methane observed variations remain puzzling

Comet impacts contribute a negligible amount of methane

Volcanoes could vent methane if they erupted but currently appear to be dormant or extinct

Rover )

POSSIBLE METHANE SOURCES

Olivine rock Water Hydrogen chemical reactions insets reactions chemical

Carbon Watermonoxide Methane Methane clathrate Hydrogen Carbon Methane Microbes monoxide could store methane produced by microbes or smokers and may produce methane by combining Hydrothermal vents gradually release it to the water with carbon-bearing molecules may produce methane in a two-stage surface through cracks process involving water and rock RON MILLER; MELISSA THOMAS ( THOMAS MELISSA MILLER; RON

www.sciam.com SCIENTIFIC AMERICAN 47 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. ago, could have supplied suffi cient methane to account for the METHANE ON TITAN currently observed global average value of 10 ppbv. But this idea runs into a problem: the distribution of methane is not Methane is to Titan what water is to Earth: a substance that cuts uniform over the planet. The time it takes to distribute meth- rivulets into the surface, pools in standing bodies, evaporates into the ane uniformly vertically and horizontally is at most several air and rains back down. As on Mars, chemical reactions deplete Titan’s months. Thus, a cometary source would result in a uniform methane, so some geologic or biological activity must replenish it. methane distribution over Mars, contrary to observations. Smoke in the Waters that leaves us with two possible sources: hydrogeo- UPPER ATMOSPHERE chemical and microbial. Either one would be fascinating. Hy- drothermal vents, known as black smokers, were fi rst discovered on Earth in 1977 on the Galápagos Rift [see “The Crest of the East Pacifi c Rise,” by Ken C. Macdonald and Bruce P. Luyendyk; Scientifi c American, May 1981]. Since then, oceanographers have found them along many other mid- oceanic ridges. Laboratory experiments show that under the conditions prevailing at these vents, ultramafi c silicates— rocks rich in iron or magnesium, such as olivine and pyrox- ene—can react to produce hydrogen in a process commonly referred to as serpentinization. In turn, reaction of hydrogen with carbon grains, carbon dioxide, carbon monoxide or carbonaceous minerals can produce methane. METHANE CYCLE The keys to this process are hydrogen, carbon, metals (which act as catalysts), and heat and pressure. All are available on Mars, too. The process of serpentinization can occur ATMOSPHERE either at high temperatures (350 to 400 degrees C) or at mild- Methane clouds er ones (30 to 90 degrees C). These lower temperatures are estimated to occur within purported aquifers on Mars. Although low-temperature serpentinization may be capable of producing the Martian methane, biology remains a serious possibility. On Earth, microorganisms known as methanogens Evaporation Precipitation produce methane as a by-product of consuming hydrogen, carbon dioxide or carbon monoxide. If such organisms lived on Mars, they would fi nd a ready supply of nutrients: hydrogen Methane lake (either produced in the serpentinization process or diffusing into the soil from the atmosphere) plus carbon dioxide and carbon monoxide (in the rocks or from the atmosphere). Once formed by either serpentinization or microbes, meth- Huygens probe ane could be stored as a stable clathrate hydrate—a chemical SURFACE ICE SUSHIL K. ATREYA began his space career on the science teams of the Voyager missions to the giant planets, continuing with Galileo, Cassini-Huygens, Venus Express, Mars Express, Mars Science Laboratory (scheduled for launch in 2009) and Juno- UNDERGROUND Jupiter Polar Orbiter (2011). His research focuses on the origin OCEAN

THE AUTHOR and evolution of atmospheres and the formation of planetary systems. He is a professor at the University of Michigan at Ann Arbor, a Fellow of the American Association for the Advance- ment of Science, and a Distinguished Visiting Scientist at the Jet Propulsion Laboratory. When not exploring other worlds, ICE Atreya likes to go hiking in the mountains and cycling. He ex- presses his gratitude to Chloé Atreya, Wesley Huntress, Paul Mahaffy, Inge ten Kate, Henry Pollack and Elena Adams for dis- ROCKY CORE cussions and comments on the drafts of this article.

Ultraviolet photon

Methane Benzene Photochemical reactions destroy methane, creating thick haze

Haze

POSSIBLE METHANE SOURCES

Acetylene Methane Microbes may produce methane but only in negligible amounts Cryovolcanoes spew water-ammonia ice grains and can Methane river vent methane )

Methane clathrate in ice layer could store methane produced in Titan’s past and gradually release it to the surface through cracks chemical reactions insets reactions chemical

Hydrothermal vents may have operated in the distant past, when the ocean extended to the rocky core RON MILLER; MELISSA THOMAS ( THOMAS MELISSA MILLER; RON www.sciam.com SCIENTIFIC AMERICAN 49 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. structure that traps methane molecules like animals in a cage— grains or other carbonaceous material—producing methane. I for later release to the atmosphere, perhaps by gradual outgas- estimate that this process would have been capable of explain- sing through cracks and fi ssures or by episodic bursts triggered ing Titan’s observed methane abundance. Once produced, by volcanism. No one is sure how effi ciently the clathrates methane could have been stored as a stable clathrate hydrate would form or how readily they would be destabilized. and released to the atmosphere either gradually, through vol- The Mars Express observations hint at greater methane canism, or in bursts, triggered by impacts. concentrations over areas containing subsurface water ice. Ei- An intriguing clue is the argon 40 gas detected by Huy- ther the geologic or biological scenario would explain this cor- gens as it descended through Titan’s atmosphere. This iso- relation. Aquifers below the ice would provide a habitat for tope forms by the radioactive decay of potassium 40, which creatures or a venue for the hydrogeochemical production of is sequestered in the rocks deep in Titan’s core. Because the methane. Without more data, the biological and geologic pos- radioactive half-life of potassium 40 is 1.3 billion years, the sibilities appear equally likely. small amount of argon 40 in the atmosphere is evidence for slow release of gases from the interior. In addition, optical A Titanic Ocean at first glance, one might NASA’S NEXT ROVER think that Titan’s methane would be easier to understand: TheMARS next stepSCIENCE to LABORATORY Mast Spectrograph (to analyze composition) the moon formed in the sub- solving the problem of methane on Mars cameras Radioisotope nebula of Saturn, whose atmo- is NASA’s Mars thermoelectric generator sphere contains huge amounts UHF antenna High-gain Science Laboratory, of the gas. Yet the data argue Weather antenna scheduled for launch station Radiation for production of methane on Neutron detector in 2009. Its detector (to scan for Titan rather than delivery of instruments should underground methane to Titan. The Huy- be able to measure water) Sample gens probe of the joint NASA the methane’s processing and European Space Agency’s isotopic makeup (a station Cassini-Huygens Mission clue to its origin) and found no xenon or krypton in sift through soil the moon’s atmosphere. Had looking for organic Sample analysis compounds (the the planetesimals that formed instrument suite composition of which Descent imager (to search Titan brought methane, they for organic would indicate which compounds) would have brought these chemical or heavy noble gases as well. The biological processes X-ray diffraction absence of such gases indicates have created or 2 meters instrument (to that methane most likely destroyed methane). analyze minerals) formed on Titan. Therefore, the presence of Rock and soil analysis methane on Titan is as mysteri- Close-up instruments camera ous as it is on Mars—in some (simplifi ed depiction) respects more so because of its sheer quantity (5 percent by volume). A plausible source, as on Mars, is serpentinization at and radar images of the surface show signs of cryovolcanism— relatively low temperatures. Christophe Sotin of the Univer- geyserlike eruptions of ammonia-water ice—which also indi- sity of Nantes in France and his colleagues have argued that cates that material wells up from the interior. The surface Titan might sustain an underground ocean of liquid water [see appears relatively young and free of craters, which is a sign of box on two preceding pages]. Dissolved ammonia, acting as resurfacing by material from the interior. The estimated re- an antifreeze, would help to keep it from freezing solid. In their surfacing rate would release methane from the interior quick- model, the ocean is 100 kilometers underneath Titan’s surface ly enough to balance the photochemical loss. Methane on and 300 to 400 kilometers deep. In the past, the decay of ra- Titan plays the role of water on Earth, complete with liquid dioactive elements and the leftover heat from Titan’s forma- surface reservoirs, clouds and rain—a full-fl edged methalog- tion might have melted nearly all the body’s ice—so the ocean ical cycle. Thus, a substantial body of evidence exists, even might have extended all the way down to the rocky core. more so than for Mars, that methane stored in the interior Under those conditions, reactions between the water and the would have no diffi culty getting out to the surface and subse- rock would have liberated hydrogen gas, which in turn would quently evaporating into the atmosphere. have reacted with carbon dioxide, carbon monoxide, carbon Might biology also play a role in creating Titan’s methane? NASA/JPL SOURCE: FOLEY; DON

50 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Christopher McKay of the NASA Ames Research Center and Mars Science Laboratory (MSL) rover, scheduled to arrive at Heather Smith of the International Space University in Stras- Mars in 2010, should be able to carry out precise measure- bourg, France, as well as Dirk Schulze-Makuch of Washington ments of the carbon isotopes in methane and possibly other State University and David Grinspoon of the Denver Museum organic materials. It will also study solid and gaseous samples of Nature and Science, have suggested that acetylene and hydro- for other chemical signs of past or present life, such as a very gen could serve as nutrients for methanogens even in the extreme high abundance ratio of methane to heavier hydrocarbons cold of Titan’s surface (–179 degrees C). This biogenic process (ethane, propane, butane) and chirality (a preference for either differs from that employed by methanogens on Earth and their left-handed or right-handed organic molecules). cousins, if any, on Mars in that no water is needed. Instead liq- Tied up with these issues is the question of why organics uid hydrocarbons on Titan’s surface serve as the medium. seem to be missing from the surface of Mars. Even in the ab- Yet this hypothesis has a shortcoming. Huygens data rule sence of life, meteorites, comets and interplanetary dust par- out an underground source of acetylene; this compound must ticles should have delivered organics over the past four and a ultimately come from methane in the atmosphere. Thus, it half billion years. Perhaps the answer lies in Martian dust dev- seems like a circular argument: to produce methane (by mi- ils and storms and ordinary saltation (the hopscotching of crobes), one needs methane. Moreover, the sheer abundance windblown dust grains). These processes generate strong stat- of methane on Titan is so immense that methanogens would ic electric fi elds, which can trigger the chemical synthesis of have to work in overdrive to produce it, severely depleting the hydrogen peroxide. Being a potent antiseptic, hydrogen perox- available nutrients. ide would quickly sterilize the surface and scrub out the organ- In view of these obstacles, a biological explanation for ics. The oxidant would also accelerate the loss of methane lo- methane is much less attractive on Titan than on Mars. Nev- cally from the atmosphere, thus requiring a larger source to ertheless, the hypothesis of habitability bears investigating. explain the abundances observed in the Martian atmosphere. Some scientists argue that this moon might have been or still In summary, methane serves as the glue that holds Titan be habitable. It receives enough sunlight to turn nitrogen and together in some mysterious ways. The presence of methane on methane into molecules that are the precursors to biology. An Mars is equally intriguing, not the least because it evokes visions of life on that planet. Fu- ture exploration of both The Mars Science Laboratory rover will study solid and bodies will seek to determine whether they were gaseous samples for chemical signs of past or present life. ever habitable. Although life as we know it can pro- underground water-ammonia brine, with some methane and duce methane, the presence of methane does not necessarily other hydrocarbons thrown in, could be a friendly environ- signify the existence of life. So planetary scientists must inves- ment for complex molecules or even living organisms. In the tigate thoroughly the sources, sinks and isotopic composition distant past, when the young Titan was still cooling off, liquid of this gas, along with other organic molecules and trace con- water may even have fl owed on the surface. stituents in both gaseous and solid samples. Even if methane is found to have no connection to life, studying it will reveal some Organic Food of the most fundamental aspects of the formation, climate his- one crucial measurement that could help determine tories, geology and evolution of Mars and Titan. the sources of methane on Mars and Titan is the carbon isotope ratio. Life on Earth has evolved to prefer carbon 12, MORE TO EXPLORE which requires less energy for bonding than carbon 13 does. Detection of Methane in the Atmosphere of Mars. Vittorio Formisano, When amino acids combine, the resulting proteins show a Sushil Atreya, Thérèse Encrenaz, Nikolai Ignatiev and Marco Giuranna in Science, Vol. 306, pages 1758–1761; October 28, 2004. marked defi ciency in the heavier isotope. Living organisms on A Sensitive Search for SO2 in the Martian Atmosphere: Implications Earth contain 92 to 97 times as much carbon 12 as carbon 13; for Seepage and Origin of Methane. Vladimir A. Krasnopolsky in Icarus, for inorganic matter, the standard ratio is 89.4. Vol. 178, No. 2, pages 487–492; November 2005. On Titan, however, the Huygens probe measured a ratio Episodic Outgassing as the Origin of Atmospheric Methane on Titan. of 82.3 in methane, which is smaller, not larger, than the ter- Gabriel Tobie, Jonathan I. Lunine and Christophe Sotin in Nature, Vol. 440, pages 61–64; March 2, 2006. restrial inorganic standard value. This fi nding argues strongly Titan’s Methane Cycle. Sushil K. Atreya, Elena Y. Adams, Hasso B. against the presence of life as we know it. To be sure, some Niemann, Jaime E. Demick-Montelara, Tobias C. Owen, Marcello scientists suggest that life could have evolved differently on Fulchignoni, Francesca Ferri and Eric H. Wilson in Planetary and Space Titan than on Earth or that the inorganic isotope ratio may be Science, Vol. 54, No. 12, pages 1177–1187; October 2006. different there. Methane and Related Trace Species on Mars: Origin, Loss, Implications for Life, and Habitability. Sushil K. Atreya, Paul R. No one has yet determined the carbon isotope ratio for Mahaffy and Ah-San Wong in Planetary and Space Science, Vol. 55, Mars. This measurement is challenging when the concentra- No. 3, pages 358–369; February 2007. tion of the gas is so low (one billionth of that on Titan). NASA’s Sushil K. Atreya’s Web page: www.umich.edu/ãtreya

www.sciam.com SCIENTIFIC AMERICAN 51 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Chromosomal Chaos and Cancer Current wisdom on the role of genes in malignancy may not explain some features of cancer, but stepping back to look at the bigger picture inside cells reveals a view that just might

By Peter Duesberg

When I first began to study cancer as a young postdoctoral fellow in the early 1960s, it looked to leading scientists as though viruses could be the cause of most, if not all, malignancies. That idea was based on the discovery of several tumor- and leukemia-producing viruses that could infect a host cell and insert their own genetic material into its genome, sparking a cancerous transformation and proliferation of the cell. I was optimistic and naive enough to hope that if researchers could understand the exact molecular mechanisms by which such viruses caused cancer, we could develop vaccines to eliminate one of humanity’s most dreaded diseases. My own contribution to that pursuit came in 1970, when my colleagues, Mi- chael Lai and Peter Vogt, and I managed to isolate a specifi c gene, src, which was suspected to be the tumor-initiating culprit in avian Rous sarcoma virus. Within a few years, more creative scientifi c minds than mine had followed this lead to a realization that a closely related gene was already present in the normal DNA of animals, including humans. And a new cancer model was born: it proposed that some triggering event, such as a mutation in a human cell’s own version of src, could ignite tumorigenic powers like those possessed by its viral counterpart. The cancer-promoting potential of such a time bomb buried in our personal genomes earned it the title of “proto-oncogene.” Once the mutation occurred, it would become a full-fl edged oncogene. The theory that mutations in certain key human genes are at the root of all cancers has dominated research for the past 30 years. Yet despite all the attempts of investigators, including myself, during that time to demonstrate that a handful of such oncogenes alone can transform normal cells into malignant ones, none JEN CHRISTIANSEN JEN

www.sciam.com SCIENTIFIC AMERICAN 53 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. PETER DUESBERG is a professor of molecular and cell biology at the Univer- Editors’ note: The author, Peter Duesberg, a sity of California, Berkeley, where he arrived from Germany in 1964 as a pioneering virologist, may be well known to readers research virologist. Within six years he had isolated the ﬁ rst true onco- for his assertion that HIV is not the cause of AIDS. gene, from within the Rous sarcoma virus, and mapped the genetic struc- The biomedical community has roundly rebutted ture of the entire virus. He proceeded to do the same for 10 more mouse that claim many times. Duesberg’s ideas about THE AUTHOR and avian sarcoma and leukemia viruses and was elected to the National chromosomal abnormality as a root cause for can- Academy of Sciences in 1986. By 1987 his work with retroviruses led him cer, in contrast, are controversial but are being ac- to conclude that HIV is merely a bystander and AIDS results from chemical tively investigated by mainstream science. We exposures and malnutrition. His ongoing work with cancer viruses also have therefore asked Duesberg to explain that work persuaded him that mutations in individual genes are insufﬁ cient to cause here. This article is in no sense an endorsement by the malignant transformations seen in cancer. SCIENTIFIC AMERICAN of his AIDS theories.

have succeeded. The oncogene model also ignores what a ca- Invading Species sual observer might perceive as a rather large elephant in the our research group arrived at a chromosomal theory room: in every known instance of cancer, individual genes may of cancer in part by thinking about the basic biological features well contain mutations, but entire chromosomes, which carry that make a normal human cell “normal,” or even “human.” thousands of genes, are also severely scrambled—duplicated, Nature is extremely conservative with regard to chromo- broken, structurally rearranged or missing entirely. Growing somes—the bound volumes of the encyclopedia of life—and the By looking at exceptions to the current rule, we hoped to ﬁ nd A BETTER RULE.

evidence suggests that this chaos on the chromosomal level is specifi c chromosome complements, or karyotypes, that defi ne not just a side effect of malignancy, as the prevailing model each species are fi nal and stable for the duration of that species. holds, but the direct cause and driving force of cancer. Sexual reproduction also enforces conservation of a species- With several colleagues in the U.S. and Europe, I have specifi c karyotype because embryonic development depends on been investigating this possibility for more than a decade, and absolute chromosomal conformity—cells containing altered or the recent work of many other researchers is also pointing to misallotted chromosomes are almost never viable. A rare ex- the conclusion that changes to the number and structure of ception, Down syndrome, illustrates the systemic damage that entire chromosomes, rather than single genes, are suffi cient results from having just one extra copy of a relatively small to initiate and sustain malignancy. This view has important chromosome, number 21, added to human cells. implications for cancer treatment and prevention, as well as Individual genes, in contrast, can be quite variable within for diagnosis of precancerous lesions when there may still be a species. Single-nucleotide polymorphisms (SNPs), for ex- time to intervene. It also fi nally explains some characteristics ample, which are mostly benign changes in the DNA sequence, of cancer cells and whole tumors that the gene mutation hy- are found scattered throughout every person’s genome and pothesis leaves unresolved. passed from parent to child. Cancer cells, and their less conspicuous precursors, are notorious violators of nature’s karyotype stability laws, how- Overview/Vicious Cell Cycles ever. Whereas normal human cells are described as diploid, ■ Prevailing cancer theory blames mutations in important because they all have two copies of the chromosomes that de- regulatory genes for upsetting normal controls on cells. fi ne our species, the karyotypes of cells in solid tumors are But it does not give a primary role in carcinogenesis to always aneuploid—meaning that they have gained or lost the gross changes to whole chromosomes that are seen whole chromosomes or segments of chromosomes [see illus- in all cancer cells. tration on opposite page]. As a result of this aneuploidy, the ■ The author argues that these chromosomal mutations, total DNA content of a cancer cell can rise to more than twice which unbalance thousands of genes en masse, are or fall to nearly half that of a normal diploid cell. This situa- suffi cient to trigger cellular instability that leads to tion would cause the cell to produce wildly skewed dosages of further chromosome disruption and to account for the proteins encoded by the thousands of genes that were mul- properties of malignant cells that cannot be explained tiplied or lost. Such gross imbalances would inevitably disrupt by the activity of specifi c genes. the work of critical teams of enzymes, including those involved in repair or disposal of damaged DNA, and would destabilize

54 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. cellular structures and regulatory circuits. Indeed, the most such entities with single drugs is like going up against an entire complex coordinated squadron of proteins in a cell, and thus zoo of wild animals with a trap built only for crocodiles. the most vulnerable, is the mitotic spindle apparatus that seg- Yet scientists do know that in every case of cancer, the en- regates chromosomes during cell division. Once aneuploidy is tire menagerie of malignant cells has arisen from a single un- established, therefore, additional derangement of the chromo- stable mother cell. This so-called clonal origin of a tumor is somes is likely. demonstrated by the presence in all or most of its cells of chro- This could explain why even cancer cells within the same mosomes whose rearrangements are so unique that they could tumor can exhibit different combinations and alterations of not have arisen independently and so must all have been inher- their chromosomes, making each cell a kind of new species ited from a common source. The challenge, therefore, is to fi nd unto itself. Their inherent instability also sets individual can- a theory that explains how one normal cell, out of the trillions cer cells free to evolve new traits and behaviors (phenotypes), that make up a human body, becomes chromosomally and unlike the normal cells in an organism, which are destined to phenotypically so unstable that it gives rise to a lethal cancer. develop predetermined characteristics depending on the organ Until about 50 years ago, most cancer researchers saw the or tissue to which they belong and according to strict species- origins of cancer in aberrant chromosomes themselves. This specifi c programming. With this license to change, an aneu- idea had been advanced in the late 19th and early 20th centu- ploid cell can gradually drop more and more of its normal ries by two German scientists, David von Hansemann and social responsibilities in a multicellular organism, multiplying Theodor Boveri. Von Hansemann, while studying cancer in ever more autonomously at the expense of other cells. Berlin, had discovered that all cancers contain abnormal Collectively, cancerous cells tend to evolve from bad to chromosomes. Boveri, a biologist in Würzburg, had deduced worse. During this process, described as progression of carci- that chromosomes are the vehicles for heritable information nogenesis, the cells start developing their own exotic sizes, within cells by closely observing the steps of sea urchin em- shapes, metabolisms and growth rates. Indeed, the very defi ni- bryo development. He reached his conclusion after witnessing tion of malignancy includes the cells’ acquisition of an un- the cellular degeneration that would follow if something went natural ability to invade neighboring tissues and travel to dis- wrong during cell division and a single chromosome was bro- tant organs, known as metastasis. The extreme variability of ken or whole copies were unevenly distributed between two cancer cells and the enormous diversity of their phenotypes are daughter cells. Boveri likened some of the resulting mal- primary reasons cancer has remained an intractable problem, formed embryos to tumors and, in 1914, predicted that gains scientifi cally and from the standpoint of treatment. Soon after or losses of specifi c chromosomes would initiate cancer. a toxic drug is found to kill tumor cells, for instance, drug- That theory fell out of favor half a century later only be- resistant variant cells grow up in their place. Trying to tackle cause such patterns could not be detected in the chaotic and

NORMAL CELL TUMOR CELL

1 2 3 4 5 1 2 3 4 5

6 7 8 9 10 11 12

6 7 8 9 10 11 12 13 14 15 16 17 18

19 20 21 22 X Y 13 14 15 16 17 18 Marker chromosomes

2 8 5 7 5 7 4 20 1 8 22 18 5 16 X 20 18 9 11 12 14 22 17 Fragment 19 20 21 22 X Y + 12 1 14 9 4 5 5 6 19 1 3 X 15 2 15 8 – – – – 9 – 2 5 15 1 16 2 sources VOL. 27; 2005, REPRINTED WITH PERMISSION OF IOS PRESS IOS OF PERMISSION WITH REPRINTED 2005, 27; VOL.

ORDER AND ANARCHY: A normal human cell’s chromosome set, or fragments. Among these are so-called marker chromo somes, whose karyotype (left), includes pairs of 23 standard chromosomes, distinctive structural rearrangements are seen in all the tumor’s cells, whereas a tumor cell, like all cancers, exhibits the irregular karyotype indicating changes that originated in the cell that spawned the cancer. (right) described as aneuploid: some whole chromosomes are Numbers under each one specify the sources of its fragments; plus

CELLULAR ONCOLOGY, missing, extra copies of others are present, and many have traded and minus signs identify those that are larger or smaller than usual. FROM “THE CHROMOSOMAL BASIS OF CANCER,” BY PETER DUESBERG ET AL., AL., ET DUESBERG PETER BY CANCER,” OF BASIS CHROMOSOMAL “THE FROM IN

Abnormal chromosome numbers in a cell create conditions that lead to further chromosome damage and disarray. With each new generation,

Mitotic spindle Carcinogen Because of a random accident Most resulting cells are not viable 1 or damage by carcinogens to 2 and die, but a surviving cell may chromosomes or mitotic machinery, a continue proliferating. Low aneuploidy in dividing cell distributes its chromosomes its offspring begins to compromise their unevenly between two daughter cells, internal functioning, but they are not yet leaving both aneuploid. multiplying excessively.

Dying cell

Dividing cell

Normal Precancerous Low aneuploidy heterogeneous karyotypes of cancer cells, especially the most Effects and Causes malignant ones, with the technology available at the time. In our strategy was to collect and analyze the features of the absence of consistent chromosomal changes, the bewil- carcinogenesis that are most inconsistent with the prevailing dering cancer karyotypes were interpreted as mere conse- gene mutation theory—by looking at the exceptions to the cur- quences of malignancy, which must itself be caused by an- rent rule, we hoped to fi nd a better rule. In the end, we identi- other unseen factor. fi ed six features of cancer that are inexplicable by gene muta- All along, however, some evidence that was largely ignored tion alone but that can be explained by chromosomal changes, continued to support an important role for aneuploidy in the and a coherent theory did emerge. genesis and progression of cancer. In fact, degree of aneuploi- Cancer risk grows with age. Lamentably common, cancer dy is a feature that is still sometimes used by pathologists and affl icts about one in three people at some point in their lives, doctors to assess the cancerous potential of abnormal cells bi- but mostly after the age of 50, which is when chances for ma- opsied from the cervix, prostate, liver, stomach, throat, breast lignancy soar. Thus, cancer is, by and large, a disease of old and other tissues. Once an abnormal growth is defi ned as a age. The gene mutation theory of cancer’s origins, however, cancer, or neoplasm, high cellular aneuploidy is also consid- predicts that the disease should be quite common in newborns. ered by some to be a sign of aggressiveness. If, as that hypothesis holds, about half a dozen mutations to Moreover, new technologies have enabled researchers to critical genes were necessary to ignite malignancy, certainly reopen the question of whether specifi c chromosomal changes some of those mutations would accumulate like SNPs over the can be detected in cancer cells, and many investigators are be- course of generations in the genomes of many individuals. A ginning to fi nd those long-sought patterns. Even oncogene en- baby could thus inherit three of six hypothetical colon cancer thusiasts studying the effects of aneuploidy in existing tumors mutations from her mother, for example, and two from her have generated evidence that points to chromosome-level al- father and be at extremely high risk of cancer from picking up terations, rather than discrete gene mutations, as the driver of the missing sixth mutation in any one of her billions of colon malignancy [see “Untangling the Roots of Cancer,” by W. cells. Some babies might even be born with colon cancer from Wayt Gibbs; Scientifi c American, July 2003]. Intrigued by inheriting all six hypothetical colon cancer mutations from the emergence of these data over the past decade, as well as by their parents. But colon cancer is never seen in children. In- the enormous potential of chromosomal alteration to generate deed, even laboratory mice intentionally engineered to carry new cellular phenotypes—much as it produces new species in an assortment of ostensibly carcinogenic mutations from birth nature—my colleagues and I set ourselves the challenge of ex- can live and propagate happily, with no higher risk of develop- plaining the instability inherent in cancer cells, which is the ing tumors than normal lab mice. source of the chaos that became the nemesis of the original Some proponents of the mutation theory maintain that ex-

chromosomal cancer theory. cept in rare cases of genetic predisposition to cancer—by which CHRISTIANSEN JEN

Skewed dosages of proteins generated by the Cells begin exhibiting progressively Malignant features, such as the ability 3 cells’ irregular chromosome complements 4 more deviant traits as aneuploidy 5 to invade neighboring tissue or cause instability that further disrupts regulatory increases and their protein production metastasize to distant locations and and DNA-maintenance processes. Additional grows more aberrant. These changes include intrinsic resistance to drugs, may also arise chromosome breakage, structural rearrangements atypical appearance and hyperproliferation, as random effects of the internal chaos and duplication errors arise. leading to formation of a tumor. caused by the cells’ escalating aneuploidy.

Tumor

Cancerous Malignant High aneuploidy

they mean the presence of inherited oncogenic mutations—the A gene mutation scenario therefore does not explain why cells gene changes believed to cause malignancies must all be ac- exposed to carcinogenic agents become cancer cells, much the quired after birth. That assumption implies a natural gene way an undersea volcano becomes an island that appears only mutation rate over an individual’s lifetime much higher than after many invisible eruptions. the norm, which is one change to a given gene in one out of Carcinogens, whether or not they cause gene mutations, every one million to 10 million cells (that is, once every 106 to induce aneuploidy. Scientists have looked for the immediate 107 cell generations). genetic effects of carcinogens on cells, expecting to see muta- Interestingly, among the rare exceptions to cancer’s age tions in many crucial genes, but instead have found that some bias are children with congenital aneuploidy, as in Down syn- of the most potent carcinogens known induce no mutations at drome, or with inherited chromosome instability syndromes, all. Examples include asbestos, tar, aromatic hydrocarbons, such as the disease known as mosaic variegated aneuploidy nickel, arsenic, lead, plastic and metallic prosthetic implants, (MVA), which also causes severe mental retardation. Defects certain dyes, urethane and dioxin. Moreover, the dose of car- of mitotic spindle assembly in the cells of children with MVA cinogen needed to initiate the process that forms malignant produce random aneuploidies throughout their bodies, and tumors years later was found to be less than one-thousandth nearly one third develop leukemia or unusual solid cancers. the dose required to mutate any specifi c gene. In all cases, how- Being born aneuploid, or prone to aneuploidy, clearly ac- ever, the chromosomes of cells treated with cancer-causing celerates processes that lead to cancer. Indeed, the inherent doses of carcinogens were unstable—that is, displaying higher instability of aneuploid cells would explain why most aneu- than usual rates of breakage and disruption. ploid embryos, as Boveri observed 100 years ago, would not These fi ndings suggest that carcinogens function as “aneu- be viable at all and thus why newborns are cancer-free and ploidogens” rather than as mutagens. The gene mutation the- cancer is not heritable. ory has never been able to explain how nonmutagenic carcin- Carcinogens take a very long time to cause cancer. Numer- ogens cause cancer. In fact, even the mutagenic carcinogens ous chemicals and forms of radiation have been shown to be may actually work to trigger cancer by inducing aneuploidy carcinogenic in animals or established as the source of occu- through direct destruction or fragmentation of chromosomes. pational or accidental cancers in humans. But even the stron- (Radiation, for example, induces mutations indirectly by fi rst gest carcinogens at the highest survivable doses never cause breaking the DNA strand. Cellular repair proteins attempt to cancer right away. Instead the disease emerges only after defi x the damage but may introduce errors or rearrangements in lays lasting years or even decades. In contrast, when substanc- the nucleotide sequence.) Nonmutagenic carcinogens, such as es known to cause gene mutations are administered to bacte- aromatic hydrocarbons, can cause aneuploidy through a dif- ria, the cells begin displaying new phenotypes within hours; ferent mechanism. Those chemicals are known to destroy the in larger organisms such as fl ies, the effect is seen within days. microtubule polymers within a cell that normally pull dupli-

www.sciam.com SCIENTIFIC AMERICAN 57 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. cate chromosomes apart symmetrically during mitosis [see creased, and so did the cells’ cancerous progression. This ex- box on two preceding pages]. Thus, the common denomina- ample strongly supports the idea that by raising or reducing a tor of all carcinogens seems to be their initiation of random cell’s normal dosage of thousands of genes at once, aneuploidy aneuploidy. produces malignant phenotypes. Patterns of aneuploidy are seen in different tumors. If an- Gratuitous traits do not contribute to the cancer’s survival. euploidy is only a side effect of cancer, then chromosomal Some of the most common and dreadful characteristics of can- changes in the cancers of different people should be random. cer do not offer any competitive survival advantages to a tu- But a pair of chromosome-painting technologies, known as mor. Examples of these include intrinsic resistance to drugs comparative genomic hybridization and fl uorescent in situ hy- the tumor has never encountered before and metastasis, which bridization, have begun to reveal signature patterns amid the does not help tumor cells compete with normal cells at their chaos in cancer cells. These techniques enable scientists to tag site of origin. Individual gene mutations, which are rare to be- Aneuploid cells reshuffl e their chromosomes MUCH FASTER than mutation can alter their genes. ) and track bits and pieces of chromosomes with colored DNA- gin with, would only be selectively conserved in tumor cells if specifi c probes to build a picture of all the chromosomal seg- the mutation were advantageous, so the chances of an untreat- ments gained, lost or rearranged in a given cell. ed tumor becoming drug-resistant through random gene mu-

Many researchers have begun to fi nd evidence of “nonran- tations in its cells are practically zero. Because chromosomes drawings Boveri dom” aneuploidies—specifi c chromosomal changes shared by are much larger and can harbor thousands of genes, however, most cancers of a certain kind, such as neoplasms of the breast they could be selectively retained for their contribution of or cervix, as predicted by the original chromosomal cancer some cancer-specifi c phenotype, and numerous unselected theory. In just one example reported last year, scientists at traits would be carried along as well. Indeed, the evidence of Karolinska University Hospital in Sweden examined cells specifi c chromosomal changes associated with drug resistance from 10 patients with Burkitt’s lymphoma and found frequent or metastasis supports this possibility. And collectively, cancer

translocations (swapping of chromosome fragments) involv- cells can evolve all kinds of new traits very rapidly. ( STUEBER KURT OF COURTESY ing chromosomes 3, 13 and 17, as well as specifi c losses or gains in segments of chromosomes 7 and 20. Because such patterns in cancer cells of different individuals ARGUMENT FOR ANEUPLOIDY? are specifi c to the type of tissue in which the cancer originates, In the early 20th century German they may well represent essential chromosome changes needed biologist Theodor Boveri advanced a to overcome programmed constraints on the original mother theory that cancer stems from cell’s development. These changes would thus represent the chromosomal disruption based on his minimum alterations needed for a viable aneuploid cell of that studies of sea urchin development (left). kind to start down the road toward becoming neoplastic. Boveri predicted that most aneuploid In addition to such cancer-type-specifi c patterns, chromo- cells would not be viable, but certain some changes indicative of cancer stage, metastatic potential, chromosome losses, gains or and even drug resistance have also been identifi ed by several rearrangements could give a cell and its research teams. For instance, the Karolinska group noted that descendants the ability to transform and multiply in ways that would lead to translocations in one segment of chromosome 17 and gains on cancer. Those signature alterations would parts of 7 and 20 were associated with drug resistance. therefore be seen in most cancers of the As scientists have continued to work out the exact func- same tissue type. Technology that tional effects of such cancer-specifi c aneuploidies, many anal- permits researchers to track specifi c yses of the amounts of different gene transcripts and proteins pieces of DNA is beginning to reveal manufactured by cancer cells have also shown that the pro- UNEQUAL CHROMOSOME distribution during chromosome changes common to tumors teins encoded by specifi c purported oncogenes are actually division of a sea urchin and even precancerous cells of the same often generated at the same levels as in normal cells. Among embryo cell is depicted in kind in different people (right). Because several interesting recent studies of this kind, one by a team of 1888 drawings by Boveri, such patterns are unlikely to arise U.S. and Israeli researchers set out to assess the protein levels who attributed the randomly if aneuploidy is only an effect, in colon cancer cells and did fi nd that large numbers of pro- problem observed in this sequence to the rather than a cause, of cancerous cell teins were being overproduced or underproduced—in amounts “defective” action of transformation, they could support the corresponding to the total DNA content changes within the “spindle fi bers.” new chromosomal theory of cancer. cell. With greater aneuploidy, the protein imbalances in-

58 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Cancer cells morph much faster than genes. Cancer cells is yet another hallmark of cancer that cannot be explained by generate new phenotypes and lose old ones exceedingly fast. the activity or inactivity of specifi c genes, which would be ex- Given a normal mutation rate—and many studies have dem- pected to have consistent effects each time and in each cell. onstrated that in more than 90 percent of cancers, the cellular Because every one of the extraordinary features of carci- gene mutation rate is not accelerated—the odds are very long nogenesis that cannot be explained by the mutation theory is once again that a particular gene will mutate to generate a new associated with chromosomal alterations, we have proposed phenotype as rapidly as is observed in cancer cells. a revised chromosomal theory of cancer that takes this inher- To test the hypothesis that whole-karyotype alterations ent instability into account. could achieve such speedy change, our group recently examined the chromosomal variations present in highly aneuploid Deadly Dynamic human breast and colon cancer cells, as well as the speed with when ca ncer is viewed as a chromosomal disease, car- which the cells acquired resistance to toxic drugs. Karyotype cinogens, rare genetic syndromes and accidental mitotic errors VOL. 19, NO. 4; 1997, changes were observed in these cells at rates of at least one in can initiate carcinogenesis by inducing random aneuploidies. 100 generations, and drug-resistance-specifi c changes at a rate Because aneuploidy unbalances thousands of genes and their of one in 1,000 to one in 100,000 cell generations. In other protein output, it sets the stage for still more aneuploidy. This words, aneuploid cells reshuffl e their chromosomes and phe- dynamic self-catalyzing condition becomes a steady source of notypes much faster than mutation can alter their genes. variation from which, in classical Darwinian terms, selection of These and similar experiments also demonstrated that the selfi sh chromosome combinations eventually gives rise to viable more aneuploid the cancer cells were, the quicker new chro- and competitive neoplastic cells. These are effectively new cell mosomal alterations appeared. This pattern strongly supports species, albeit parasitic ones, with unstable karyotypes.

GENES, CHROMOSOMES AND CANCER, AND CHROMOSOMES GENES, a conclusion that the chromosomal instability seen in cancer Once cancer progression is under way, random chromo- cells is catalyzed by aneuploidy itself. Once this vicious cycle some reshuffl ing can rapidly generate gratuitous traits that is under way, the fact that every cell would be randomly gen- include lethal properties such as drug resistance and metasta- erating its own new phenotypes could explain an observation sis. Thus, the prospects for success in treating tumors with made decades ago by Leslie Foulds of the Royal Cancer Hos- individual drugs, particularly those targeted to single genes, pital in London that “no two tumors are exactly alike ... even are not good. Some investigators have recently proposed fi ght- when they originate from the same tissue ... and have been ing fi re with fi re in the form of treatments that accelerate chro-

FROM “ADVANCED-STAGE CERVICAL CARCINOMAS ARE DEFINED BY A RECURRENT PATTERN OF CHROMOSOMAL CHROMOSOMAL OF PATTERN RECURRENT A BY DEFINED ARE CARCINOMAS CERVICAL “ADVANCED-STAGE FROM 3Q,” ARM CHROMOSOME OF GAIN CONSISTENT A AND INSTABILITY GENETIC HIGH REVEALING ABERRATIONS BY K. HESELMEYER INC. SONS, & WILEY JOHN OF SUBSIDIARY A INC., WILEY-LISS, ET OF AL., PERMISSION WITH REPRINTED IN induced experimentally in the same way.” Such individuality mosomal DNA damage and aneuploidy, with the goal of making tumor cells so unstable that they are no longer viable. This might work in a very small, well-contained tumor, but it is a CHROMOSOME CHANGES in cervical cancer cells from 30 patients showed technique that could prove diffi cult to control. some consistent patterns when areas where genetic material was lost (red) The slow progression from early aneuploidy to aggressive green or added ( ) were highlighted, as in the image from one case shown here. neoplasm does, however, offer a wide window of opportu- Specifi c gains in chromosome 3 were seen in 23 cases and appeared in tumors progressing to invasive carcinoma. Additions to chromosome 1 and nity for doctors to detect and surgically remove a potential losses in chromosomes 2, 3 and 4 were found in a third to half of all the cases. cancer before it enters a neoplastic stage. Early cancers could also be distinguished from morphologically identical benign tumors by checking for aneuploidy. In more advanced tumors, chromosome patterns associated with drug resistance or metastatic potential could be used to guide treatment choices. Finally, screening for chromosome-damaging substances 1 2 3 4 5 in foods, drugs and the environment could signifi cantly improve cancer prevention by identifying aneuploidy-inducing potential carcinogens. Today I am still optimistic enough to hope that by coming full circle, cancer scientists will eventu- 6 7 8 9 10 11 12 ally arrive at a basic understanding of this fearsome disease, yielding effective prevention, management and even cures. MORE TO EXPLORE 13 14 15 16 17 18 Debate Surges over the Origins of Genomic Defects in Cancer. Jean Marx in Science, Vol. 297, pages 544–546; July 26, 2002. The Sigmoidal Curve of Cancer. Roberto P. Stock and Harvey Bialy in Nature Biotechnology, Vol. 21, pages 13–14; January 2003. 19 20 21 22 X The Chromosomal Basis of Cancer. Peter Duesberg et al. in Cellular Oncology, Vol. 27, Nos. 5–6, pages 293–318; 2005. Cancer Drug Resistance: The Central Role of the Karyotype. Peter Duesberg et al. in Drug Resistance Updates (in press).

By Massoud Amin and Phillip F. Schewe

60 SCIENTIFIC AMERICAN COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. A smarter power grid that automatically responds to problems could reduce the rising number of PREVENTING BLACKOUTS debilitating blackouts

ugust 14, 2003, was a typical warm day in the Midwest. But shortly after 2:00 P.M. several power lines in northern Ohio, sagging under the high current they were carrying, A brushed against some overgrown trees and shut down. Such a disturbance usually sets off alarms in a local utility’s control room, where human operators work with controllers in neighboring regions to reroute power fl ows around the injury site. On this day, however, the alarm software failed, leaving The 2003 disaster was a harbinger, too. Within two local operators unaware of the problem. Other controllers who months, major blackouts occurred in the U.K., Denmark, were relaying, or “wheeling,” large amounts of power hun- Sweden and Italy. In September 2003 some 57 million Ital- dreds of miles across Ohio, Michigan, the northeastern U.S. ians were left in the dark because of complications in trans- and Ontario, Canada, were oblivious, too. Transmission lines mitting power from France into Switzerland and then into surrounding the failure spot, already fully taxed, were forced Italy. In the U.S., the annual number of outages affecting to shoulder more than their safe quota of electricity. 50,000 or more customers has risen for more than a decade. To make matters worse, utilities were not generating In addition to inconvenience, blackouts are causing major enough “reactive power”—an attribute of the magnetic and economic losses. The troubles will get worse until the entire electric fi elds that move current along a wire. Without suffi - transmission system that moves power from generating plants cient reactive power to support the suddenly shifting fl ows, to neighborhood substations is overhauled. More high-volt- overburdened lines in Ohio cut out by 4:05 P.M. In response, a age lines must be built to catch up with the rising demand power plant shut down, destabilizing the system’s equilibrium. imposed by ever more air conditioners, computers and re- More lines and more plants dropped out. The cascade contin- chargeable gadgets. ued, faster than operators could track with the decades-old But perhaps even more important, the power grid must be monitoring equipment that dots most of the North American made smarter. Most of the equipment that minds the fl ow of

power grid, and certainly much faster than they could control. electricity dates back to the 1970s. This control system is not Within eight minutes 50 million people across eight states and good enough to track disturbances in real time—as they hap- two Canadian provinces had been blacked out. The event was pen—or to respond automatically to isolate problems before

RANDY LYHUS RANDY the largest power loss in North American history. they snowball. Every node in the power grid should be awake,

SCIENTIFIC AMERICAN 61 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. responsive and in communication with THE PROBLEM: TOO MANY PLAYERS, TOO LITTLE INVESTMENT every other node. Furthermore, the information that operators receive at cen- Restructuring of the utility industry (below) because of deregulation is a major reason why tral control stations is sparse and at least U.S. blackouts are increasing (opposite page, bottom graph). With a single company no longer 30 seconds old, making it impossible for in charge of the action in a given region, the power grid is not being upgraded or expanded to them to react fast enough to stop the keep pace with growing demand (opposite page, top panels). large cascades that do start. A self-healing smart grid—one that is aware of na- DEREGULATION HAS FRAGMENTED CONTROL OF THE POWER SYSTEM scent trouble and can reconfi gure itself BEFORE RESTRUCTURING Home to resolve the problem— could reduce blackouts dramatically, as well as contain the chaos that could be triggered by Power line Power terrorist sabotage. It would also allow plant Substation more effi cient wheeling of power, saving Tower utilities and their customers millions of dollars during routine operation. The Industry technology to build this smart grid largely exists, and recent demonstration proj- GENERATION TRANSMISSION DISTRIBUTION ects are proving its worth. Overwhelmed by Progress POSSIBLE ADDITIONAL CONFIGURATIONS TODAY Home Distant the transmission system has utility become vulnerable to blackouts because plant of a century-long effort to reduce power Substation losses. As power moves through a wire, some of it is wasted in the form of heat. Power The loss is proportional to the amount plant Industry of current being carried, so utilities keep GENERATION TRANSMISSION DISTRIBUTION the current low and compensate by raising the voltage. They have also built progressively longer, higher-voltage OPERATOR Local utility Independent transmitter Industrial co-generator lines to more effi ciently deliver power Distant utility Residential generator from generation plants to customers located far away. These high-voltage lines also allow neighboring utilities to link eration supply and customer demand. sibility that a shutdown in one sector their grids, thereby helping one another Such interconnectedness entails cer- could rapidly propagate to others. A huge sustain a critical balance between gen- tain dangers, however, including the pos- 1965 blackout in the Northeast prompted utilities to create the North American Smart Grid Electric Reliability Council—now called Overview/ the North American Electric Reliability ■ Demand for electricity has increased steadily for decades, yet transmission Corporation (NERC)—to coordinate ef- lines that transport power from generation plants to customers have not been forts to improve system reliability. Simi- added or upgraded at the same pace. As a result, the grid has become overloaded, lar bodies, such as Europe’s Union for making it more prone to blackouts, which have risen in number and severity the Coordination of Transmission of and cost the U.S. more than $70 billion in annual economic losses. Electricity, exist around the world. ■ Even with more lines, a self-healing smart grid that can sense local problems Why, then, had the U.S. grid become early, and automatically fi x or isolate them before they snowball, is needed to vulnerable enough to fail massively in prevent the cascading power failures that cause blackouts. 2003? One big reason is that investment ■ Digital controllers and real-time communications devices must be placed on in upgrading the transmission system every transmission line, substation, power plant and utility operations center. has been lacking. Sharply rising fuel ■ The operations centers also need updated computers and software that prices in the 1970s and a growing disen- enable human controllers to manually take over the automated smart grid chantment with nuclear power prompt- if a blackout does somehow begin. And controllers require better training ed Congress to pass legislation intended to know how to react quickly. to allow market competition to drive ef-

ﬁ ciency improvements. Subsequent laws INSTITUTE, ELECTRIC EDISON MINNESOTA, OF UNIVERSITY SOURCES: CHRISTIANSEN; JEN ENERGY OF DEPARTMENT U.S. CORP., RELIABILITY ELECTRIC AMERICAN NORTH

62 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Investors have found generation, now and distribution lines. The system sends largely deregulated, to be attractive. But information or alarms back to operators because the transmission system has over telecommunications channels. MONEY SPENT ON U.S. TRANSMISSION LINES been only partially deregulated, uncer- SCADA technology goes back 40 IS DECLINING tainty over its fate makes investors wary. years, however. Much of it is too slow for 50 (Deregulation of distribution is still in its today’s challenges and does not sense or Generation infancy.) Meanwhile, even though wheel- control nearly enough of the components 40 Transmission ing occurred in the past, since the 1990s around the grid. And although it enables Distribution 30 much larger amounts of power have been some coordination of transmission 20 moved over great distances. As a result, among utilities, that process is extremely 10 massive transfers are ﬂ owing over trans- sluggish, much of it still based on tele-

of electricity revenue) electricity of 0 mission lines built mostly by utilities for phone calls between human operators at Capital Invested (percent Invested Capital 1965 1975 1985 1995 2005 local use decades ago. the utility control centers, especially dur- Year Proposed federal legislation might ing emergencies. What is more, most TRANSMISSION CAPACITY IS FALLING BEHIND encourage more investment, but even if programmable logic controllers and re- GROWTH IN DEMAND transmission capacity is added, black- mote terminal units were developed be- outs will still occur. The entire power fore industry-wide standards for interop- grid has to be refurbished, because the erability were established; hence, neigh- 20% 7% existing control technology—the key to boring utilities often use incompatible INCREASE IN INCREASE IN quickly sensing a small line failure or control protocols. Utilities are operating DEMAND TRANSMISSION the possibility of a large instability—is ever closer to the edge of the stability en- FOR ELECTRICITY CAPACITY antiquated. To remain reliable, the grid velope using 1960s-era controls. (1999–2009) (1999–2009) will have to operate more like a fi ghter plane, fl own in large part by autono- The Self-Healing Smart Grid RESULT: LARGE BLACKOUTS ARE GROWING mous systems that human controllers the result is that no single operator IN NUMBER AND SEVERITY can take over if needed to avert disaster. or utility can stabilize or isolate a trans- 150 mission failure. Managing a modern Number of significant outages A Need for Speed grid in real time requires much more au- 90 Outages affecting modern warplanes are so packed tomatic monitoring and far greater in- 50,000 or more with sophisticated gear that pilots rely on teraction among human operators, com- customers 30 a network of sensors and automatic con- puter systems, communications net- 1991–1995 1996–2000 2001–2005 trols that quickly gather information and works and data-gathering sensors that Year act accordingly. Fortunately, the soft- need to be deployed everywhere in pow- ware and hardware innovations required er plants and substations. Reliable op- have instigated a sweeping change in the to fl y the power grid in a similar fashion eration also requires multiple, high-data- industry that has come to be called re- and to instantly reroute power fl ows and rate, two-way communications links structuring. Before restructuring began shut down generation plants are at hand. among all these nodes, which do not ex- in earnest in the 1990s, most utilities Reconfi guring a widely interconnect- ist today, plus powerful computing fa- conducted all three principal functions ed system is a daunting challenge, though. cilities at the control center. And intelli- in their region: generating power with Most power plants and transmission gent processors—able to automatically large plants, transmitting it over high- lines are overseen by a supervisory con- reconfi gure power fl ows when precur- voltage lines to substations, then dis- trol and data acquisition (SCADA) sys- sors to blackouts are sensed—must be tributing it from there to customers over tem. This system of simple sensors and distributed across the network. lower-voltage lines. Today many inde- controllers provides three critical func- Flying the grid begins with a differ- pendent producers sell power near and tions—data acquisition, control of pow- ent kind of system design. Recent re- far over transmission lines they do not er plants, and alarm display—and al- search from a variety of fi elds, including own. At the same time, utilities have lows operators who sit at central control nonlinear dynamical systems, artifi cial been selling off parts of their companies, stations to perform certain tasks, such intelligence, game theory and software encouraged by the Federal Energy Reg- as opening or closing a circuit breaker. engineering, has led to a general theory ulatory Commission to further promote SCADA monitors the switches, trans- of how to design complex systems that competition. Gradually the transmis- formers and pieces of small hardware, adapt to changing conditions. Mathe- sion business has become a confusing known as programmable logic control- matical and computational techniques mixture of regulated and unregulated lers and remote terminal units, that are developed for this young discipline are services, with various companies con- installed at power plants, substations, providing new tools for grid engineers. trolling fragmented pieces. and the intersections of transmission Industry working groups, including a

www.sciam.com SCIENTIFIC AMERICAN 63 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. group run by one of us (Amin) while at THE SOLUTION: A SMART GRID THAT HEALS ITSELF the Electric Power Research Institute (EPRI) in Palo Alto, Calif., have pro- Imagine that a thunderstorm knocks out power lines L5 and L6. This occurrence would posed complex adaptive systems for typically cause a chain reaction of line faults that would black out Area 1. But a smart grid would isolate and correct the problem as depicted below. The action large regional power grids. Several util- begins as a look-ahead computer at the control center simulates ities have now deployed, at a demon- corrective actions in less than half a second and sends stration scale, smart remote terminal instructions to control computers around the grid. Power plant G2 units and programmable controllers that can autonomously execute simple processes without fi rst checking with a human controller, or that can be repro- L9 grammed at a distance by operators. Much wider implementation is needed. L8 A self-healing smart grid can best be A r e a 3 L6 Area 2 built if its architects try to fulfi ll three Control center primary objectives. The most funda- Power mental is real-time monitoring and re- plant G2 L5 action. An array of sensors would mon- L3 L5 Factory itor electrical parameters such as volt- L2 age and current, as well as the condition B1 L3 L1 Power plant G1 of critical components. These measure- Substation C ments would enable the system to con- B2 B3 stantly tune itself to an optimal state. Substation A Home The second goal is anticipation. The A r e a 1 system must constantly look for potential REACT problems that could trigger larger distur- QUICKLY bances, such as a transformer that is over- 0.04 second later heating. Computers would assess trou- The loss of L5 and L6 causes a fault in line L1. Control computers tell circuit breakers B1 and B2 ble signs and possible consequences. They to open to isolate the fault, but B2 becomes stuck in the closed position. would then identify corrective actions, 0.1 second simulate the effectiveness of each action, Power generator G1 automatically accelerates to meet demand from the loss of G2 caused by and present the most useful responses to problems on lines L5 and L1. G1 also accelerates to attempt to keep line voltage throughout Area 1 human operators, who could then quick- at the required 60 hertz (cycles per second). ly implement corrective action by dis- 0.4 second patching the grid’s many automated The control computer-simulator in substation A tells breaker B3 to open to protect the substation control features. The industry calls this against damage from excessive current fl ow through it. B3 opens, shutting down line L2. G1 capability fast look-ahead simulation. accelerates further to compensate. The third objective is isolation. If 0.5 second failures were to occur, the whole net- The control center shuts down generator G1 to prevent damage to it from excessive acceleration. work would break into isolated “islands,” each of which must fend for itself. Each island would reorganize its paired the failures, human controllers power lines themselves. As soon as pow- power plants and transmission fl ows as would prepare each island to smoothly er fl ows were restored, the system would best it could. Although this might cause rejoin the larger grid. The controllers again start to self-optimize. voltage fl uctuations or even small out- and their computers would function as To transform our current infrastruc- ages, it would prevent the cascades that a distributed network, communicating ture into this kind of self-healing smart cause major blackouts. As line crews re- via microwaves, optical fibers or the grid, several technologies must be deployed and integrated. The fi rst step is MASSOUD AMIN and PHILLIP F. SCHEWE have extolled the virtues of a smart grid for years. to build a processor into each switch, Amin is professor of electrical and computer engineering at the University of Minnesota circuit breaker, transformer and bus and is director of the school’s Center for the Development of Technological Leadership. bar—the huge conductors carrying elec- When he was with the Electric Power Research Institute in Palo Alto, Calif., he led the tricity away from generators. Each development of more than 20 advanced technologies and devised the foundations of a transmission line should then be fi tted “self-healing grid,” a term he coined. Schewe is chief science writer at the American In- with a processor that can communicate THE AUTHORS

stitute of Physics and edits the institute’s weekly newsletter, Physics News Update. with the other processors, all of which DON FOLEY; SOURCE: ELECTRIC POWER INSTITUTE RESEARCH

64 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Using power Breaker closed Making power Control center Substation C C o n t r o l c e n t e r Breaker open Alert condition Control Substation B Section down computer

L7 Factory L5 L9 L4 L2 Power L3 plant G1 L8 B1 Area 2 B4 L1 B1 Area 3 L6 B2 Control center B3 S u b s t a t i o n A L7 L5 A r e a 1 L4 L2 L3 B1 L 1

MAKE DECISIONS A r e a 1 0.6 second The control computer in substation B would typically shut down line L3 to reduce demand if generator G1 were accidentally lost, but because it was RETURN stopped deliberately, computers across Area 1 communicate and decide instead to shut down a big factory, lowering demand considerably. This TO NORMAL action reduces the mismatch between generation and demand so critical functions such as streetlights and hospitals can stay powered. 60 seconds Lines L3, L4 and L7 have been spared, but L4 is becoming over- 10 seconds loaded. Human operators at the control center communicate via After several seconds, however, the substation B computer detects that satellite to operators in the Area 2 control center, asking for the voltage there is beginning to oscillate beyond safe tolerances because help. Operators in Area 2 send power over line L8; they also the mismatch is still signifi cant, threatening to damage equipment on lines instruct the control computers in their sector to modify power L3, L4 and L7. Rather than shutting down those lines (the old-fashioned fl ows slightly to compensate for the sudden export. Once road response), the area computers change control of generator G2 to manual, crews fi x damaged lines L5 and L6, the computers will bring L1 advising human operators at the Area 1 control center to raise generation and power plant G1 back into service. Power in the three areas or reduce load. They do some of both. returns to normal fl ow.

would track the activity of their particu- oring and self-healing can be carried out. sense very local disturbances, which can lar piece of the puzzle by monitoring A complete transition also requires provide an earlier warning of problems sensors built into their systems. digitization of the small, low-voltage that may be mounting, thereby improv- Once each piece of equipment is be- distribution lines that feed each home ing look-ahead simulation. And it will ing monitored, the millions of electro- and business. A key element is to replace allow utilities to offer customers hour- mechanical switches currently in use the decades-old power meter, which re- by-hour rates, including incentives to should be replaced with solid-state, lies on turning gears, with a digital me- run appliances and machines during off- power-electronic circuits, which them- ter that can not only track the current peak times that might vary day to day, selves must be beefed up to handle the going into a building but also track cur- reducing demand spikes that can desta- highest transmission voltages: 345 kilo- rent sent back out. This will allow utili- bilize a grid. Unlike a meter, this digital volts and beyond. This upgrade from ties to much better assess how much energy portal would allow network in- analog to digital devices will allow power and reactive power is flowing telligence to ﬂ ow back and forth, with the entire network to be digitally con- from independent producers back into consumers responding to variations in trolled, the only way real-time self-monit- the grid. It will also allow a utility to pricing. The portal is a tool for moving

www.sciam.com SCIENTIFIC AMERICAN 65 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. THE HUMAN FACTOR If a local blackout begins to escalate beyond a smart grid’s ability to automatically keep beyond the commodity model of elec- it in check, human operators in regional control rooms could attempt to cut off the chain tricity delivery into a new era of energy reaction. To do so, they would need complete and up-to-the-second information about the services as diverse as those in today’s dy- network, consistent computer protocols, predetermined response procedures and solid namic telecommunications market. training. Each of these prerequisites was lacking when the massive 2003 U.S. blackout The EPRI project to design a proto- began to snowball, as dialogue during the fi rst minutes of the event shows (portions type smart grid, called the Complex In- appear below). The recorded conversations, published by the North American Electric teractive Networks/Systems Initiative, Reliability Council, took place between reliability controllers in neighboring regions who was conducted from 1998 to 2002 and were trying to help one another balance power fl ows that were heading out of control. involved six university research consor- tia, two power companies and the U.S. Department of Defense. It kicked off several subsequent, ongoing efforts at the U.S. Department of Energy, the Na- tional Science Foundation, the DOD and EPRI itself to develop a central nervous system for the power grid. Collectively, the work shows that the grid can be operated close to the limit of stability, as long as operators constantly have detailed knowledge of what is happening SATELLITE IMAGES show the Northeast the night before (left) and after the 2003 blackout (right). everywhere. An operator would moni- POOR OPERATOR TRAINING, LACK OF REAL-TIME DATA tor how the system is changing, as well AEP operator: “What do you have on the Sammis-Star [line]?” as how the weather is affecting it, and PJM operator: “I’m sorry? Sammis-Star, okay, I’m showing 960 on it and it’s have a solid sense of how to best main- highlighted in blue.... Tell me what that means on your machine.” tain a second-by-second balance be- AEP: “Blue? Normal.... I mean—that’s what’s on it?” tween load (demand) and generation. PJM: “960, that’s what it says.” As an example, one aspect of the AEP: “That circuit just tripped. South Canton-Star.” EPRI’s Intelligrid program is to give op- PJM: “Did it?” AEP: “It tripped and reclosed ...” erators greater ability to foresee large- scale instabilities. Current SCADA sys- INCONSISTENT COMPUTER PROTOCOLS, LACK OF CONTINGENCY PLANS tems have a 30-second delay or more in PJM: “I’m still seeing fl ow on both those lines. Am I looking at state- assessing the isolated bits of system be- estimated data?” havior that they can detect—analogous AEP: “Probably.” to fl ying a plane by looking into a foggy PJM: “Yeah, it’s behind, okay. You’re able to see raw data?” rearview mirror instead of the clear air- AEP: “Yeah; it’s open. South Canton-Star is open.... We have more trouble ... more things are tripping. East Lima and New Liberty tripped out. space ahead. At EPRI, the Fast Simula- Look at that.... Oh, my gosh, I’m in deep ...” tion and Modeling project is developing PJM: “You and me both, brother. What are we going to do?” faster-than-real-time, look-ahead simu- lations to anticipate problems—analo- INCOMPLETE INFORMATION, LACK OF REAL-TIME DATA gous to a master chess player evaluating PJM: “... it looks like they lost South Canton-Star 345 line. I was wondering if his or her options several moves ahead. you could verify fl ows on the Sammis-Star line for me.” This kind of grid self-modeling, or self- MISO operator: “Well, let’s see what I’ve got. I know that First Energy lost their Juniper line, too.” consciousness, would avoid disturbanc- PJM: “Did they?” es by performing what-if analyses. It MISO: “And they recently have got that under control here.” would also help a grid self-repair—adapt PJM: “And when did that trip? That might have ...” to new conditions after an outage, or an MISO: “I don’t know yet....” attack, the way a fi ghter plane reconfi g- PJM: “And right now I am seeing AEP systems saying Sammis to Star ures its systems to stay aloft even after is at 1378....” being damaged. MISO: “Let me see. I have got to try and fi nd it here, if it is possible.... I see South Canton-Star is open, but now we are getting data of 1199, and Who Should Pay I am wondering if it just came after.” technologically, the self-healing PJM: “Maybe it did.” smart grid is no longer a distant dream. Finding the money to build it, however, PJM: Pennsylvania–New Jersey–Maryland

is another matter. NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION AEP: American Electric Power MISO: Midwest Independent System operator MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. a major obstacle. States’ rights and state- level public utility commission regula- tions essentially kill the motivation of any utility or utility group to lead a na- tionwide effort. Unless collaboration can be created across all states, the forced nationalization of the industry is the only way to achieve a smart grid. At stake is whether the country’s critical infrastructures can continue to function reliably and securely. At the very least, a self-healing transmission system would minimize the impact of any kind of terrorist attempt to “take out” the power grid. Blackouts can be avoided or minimized, sabotage can be contained, outages can be reduced, and electricity can be delivered to everyone more effi ciently. Had a self-healing smart grid been in place when Ohio’s local line failed in Au- gust 2003, events might have unfolded very differently. Fault anticipators locat- RESEARCHERS at the Pacific Northwest National Laboratory sit in a simulated regional control ed at one end of the sagging transmission center and test prototype software that would provide human operators with real-time grid line would have detected abnormal sig- information, necessary to stop a nascent blackout before it expands. nals and redirected the power fl owing through and around the line to isolate The grid would be costly, though not What terrorism threat level is the indus- the disturbance several hours before the prohibitively so given historic invest- try responsible for and what should gov- line would have failed. Look-ahead sim- ments. EPRI estimates that testing and ernment cover? If rate increases are not ulators would have identifi ed the line as installation across the entire U.S. trans- palatable, then how will a utility be al- having a higher-than-normal probability mission and distribution system could lowed to raise money? Improving the of failure, and self-conscious software run $13 billion a year for 10 years— 65 energy infrastructure requires long- along the grid and in control centers percent more than the industry is cur- term commitments from patient inves- would have run failure scenarios to de- rently investing annually. Other studies tors, and all pertinent public and private termine the ideal corrective response. predict $10 billion a year for a decade or sectors must work together. Operators would have approved and im- more. Money will also have to be spent Government may be recognizing the plemented the recommended changes. If to train human operators. The costs need for action. The White House Of- the line somehow failed later anyway, the sound high, but estimates peg the eco- fi ce of Science and Technology Policy sensor network would have detected the nomic loss from all U.S. outages at $70 to and the U.S. Department of Homeland voltage fl uctuation and communicated it $120 billion a year. Although a big black- Security recently declared a “self-heal- to processors at nearby substations. The out occurs about once a decade, on any ing infrastructure” as one of three stra- processors would have rerouted power given day 500,000 U.S. customers are tegic thrusts in their National Plan for through other parts of the grid. The most without power for two hours or more. R&D in Support of Critical Infrastruc- a customer in the wider area would have Unfortunately, research and devel- ture Protection. National oversight may seen would have been a brief fl icker of opment funding in the electric utility in- well be needed, because the current ab- the lights. Many would not have been dustry is at an all-time low, the lowest sence of coordinated decision making is aware of any problem at all. of any major industrial sector except for pulp and paper. Funding is a huge chal- MORE TO EXPLORE lenge because utilities must meet com- Technical Analysis of the August 14, 2003, Blackout: What Happened, Why, and What peting demands from customers and Did We Learn? North American Electric Reliability Council, 2004. Available at www.nerc.com/ filez/blackout.html regulators while being responsive to ˜ Toward a Smart Grid: Power Delivery for the 21st Century. Massoud Amin and Bruce F. Wollenberg their stakeholders, who tend to limit in- in IEEE Power and Energy Magazine, Vol. 3, No. 5, pages 34–41; September/October 2005. vestments to short-term returns. The Grid: A Journey through the Heart of Our Electrified World. Phillip F. Schewe. Joseph Henry

PACIFIC NORTHWEST NATIONAL LABORATORY NATIONAL NORTHWEST PACIFIC Other factors must be considered: Press, 2007.

www.sciam.com SCIENTIFIC AMERICAN 67 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. SOUTH AMERICA’S MISSING MAMMALS Startling fossil discoveries in the Chilean Andes reveal an unexpected menagerie of unique mammals that once roamed South America. The ﬁ nds also overturn long-held wisdom about the continent’s geologic history.

By John J. Flynn, André R. Wyss and Reynaldo Charrier

t the edge of a sprawling grassland, a pair of hoofed grazers resembling horses, an antelopelike notoungulate and a ground sloth feed quietly, oblivious to their impending doom. Equally unaware are the chinchilla Aand the tiny, mouselike marsupial nibbling seeds nearby. Suddenly, one of the jagged, snow-covered volcanoes on the horizon explodes cata- strophically, sending a fl ood of muddy ash down its steep slopes. Soon after, this roiling slurry bursts across the fl atter lowlands, entombing the unsuspecting animals in its path. As devastating as this volcanic torrent was for the creatures it buried, it would become a boon for paleontology. Tens of millions of years after the mammals’ untimely deaths, the exhuming forces of mountain building and subsequent erosion exposed remnants of their fossilized skeletons to the UNUSUAL MAMMALS that once light of day high in the Andes Mountains of cen- populated South America go tral Chile. Our team discovered the fi rst of these about their business in this artist’s reconstruction, unaware bones in 1988 while searching for dinosaur re- that a violent fl ood of muddy ash mains in an alpine valley of the Tinguiririca River, from a nearby volcano will soon

near the border with Argentina. The initial ﬁ nd- spell their doom. MARTIN RAÚL

4 1

COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. ing of mammal bones proved so fruitful that we have returned to the region near- Peru Brazil ly every year since. So far we have uncov- Bolivia ered more than 1,500 fossils of ancient 25 kilometers El Volcan mammals from dozens of sites in the Chile central Chilean Andes.

Argentina

Painstaking laboratory analysis of M a our growing collection has yielded ma- ip o Ri jor revelations about the history of ver South America’s ancient mammals. To

C our astonishment, the Chilean fossils a c h Atlantic range from 40 million to 10 million ap oa Ocean l Ri years old—much younger than anything ver we expected to fi nd there. Indeed, many ILE Patagonia H of the specimens represent the only C San Fernando NTINA mammal remains from segments of that E T ARG ing time interval found anywhere in South uir iric a R America. Some of these unique fossils iv er FINDING FOSSILS illuminate a previously opaque period in More than a dozen locations in the Andes of the history of the continent’s native central Chile, including those indicated at mammal lineages; others help to resolve the left (dots), have together yielded Termas hundreds of mammal fossils since the long-standing debates about the origins del Flaco T authors initially discovered a cache of them eno of key immigrant groups. Together they River Tinguiririca in the Tinguiririca Valley in 1988. Between have revised understanding of when cer- 40 million and 10 million years old, these ancient mammalian remains are the fi rst tain ecosystems—and the mountains ever found in this region of South America; themselves—appeared in this part of most of the continent’s fossil mammals the world. derive from farther south in Patagonia. The remains are preserved in volcanic sediments of the Abanico Formation Tantalizing Discovery (brown), which are exposed across most of what scientists know about Lake Laja thousands of square kilometers South America’s ancient mammals is (250 kilometers south) of steep, alpine terrain. based on clues unearthed in the continent’s far southern reaches, mainly Pata- and violent to preserve organic remains.) contain bony remnants of the track mak- gonia. Those regions have abundant We decided to take a chance that the ers. If we were extremely lucky and kept outcrops of typical fossil-bearing Tinguiririca Valley might harbor fossils our eyes close to the ground, we might rocks—shale, sandstone and other hard- when we learned of a report about dino- even fi nd a fossil of one of the dinosaurs’ ened sediments from rivers and their saur footprints. The rocks were the right tiny mammalian contemporaries, which floodplains. Before our first visit to age—geologists then assumed that most were no bigger than shrews. Chile, researchers had not searched rocks along the main spine of the Chil- On the last day of a one-week recon- systematically for land animal fossils ean Andes dated back at least 65 million naissance trip in 1988, our team of four in that country’s mountainous areas, to 100 million years, to the latter part of split up to prospect the precipitous because most rocks there are volcanic. the Mesozoic era, when dinosaurs slopes flanking each side of the Tin- (The standard assumption is that lava reigned supreme. We knew that any sed- guiririca River. Almost immediately, and other erupted materials are too hot iments preserving footprints might also the pair working north of the river reached the layer of ancient sediments that bore the dinosaur tracks, then con- Overview/Fossils Galore tinued up the valley in search of more ■ The more than 1,500 fossil mammal specimens discovered in the Andes potential fossil-bearing deposits. To of central Chile include a remarkable array of new species as well as the their dismay, however, the only fossils continent’s oldest known rodent remains. they recovered were from fi sh, ammo- ■ From 40 million to 10 million years old, these new fossils bridge gaps in nites and other ocean-dwelling crea- the known history of South America’s unusual mammals. tures—no reptiles or mammals. Mean- ■ Some of the fossils document the existence of a 32-million-year-old grassland while the team members working on the that predates such ecosystems elsewhere in the world by some 15 million years. south side of the river were having a

similarly frustrating day. Late in the af- LUCY READING-IKKANDA

70 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. ternoon, though, their spirits soared lian, as indicated by, among other char- right away that we were onto something when they spied a few fossilized scraps acteristics, its single pair of lower jaw- big. These new fossils clearly had much of bone and teeth eroding out of a large bones. (The lower jaws of reptiles are to say about the history of South Amer- patch of reddish-brown volcanic sedi- made up of many separate bones.) Later, ica’s distinctive living mammals— ments nearly 1,000 meters above the we would formally describe that crea- sloths, monkeys, anteaters and chinchil- valley fl oor. A closer look revealed that ture as a new species of notoungulate, las among them. The ancestors of these the fossils were land-dwelling verte- a diverse group of hoofed herbivores animals, including many of the new brates about the size of a small horse. that ranged from rabbit- to hippo-size creatures we were fi nding in the Tin- At fi rst, we tried to shoehorn these and became extinct less than one mil- guiririca Valley, were among those that fragments into the prevailing view about lion years ago. This new species proba- evolved when South America was an is- the age of the rocks—animals this size bly looked a bit like an antelope. The land continent. For most of the past 80 must be peculiar dinosaurs or other odd teeth we found the previous year turned million years, following the breakup of Mesozoic beasts. But the complex, dif- out to be from a rhinolike notoungulate. the supercontinent Pangaea and its ferentiated teeth with the high-crowned, In all, during our first three seasons southern portion, Gondwana, tectonic flat-topped and multifaceted molars of fi eldwork in Tinguiririca, we would plate motions kept South America sepa- that are unique to some mammals told turn up more than 300 specimens, includ- rated from other landmasses. This pe- a different story. These mammals were ing marsupials, an early sloth, arm- riod of isolation fostered the evolution clearly too large—and too advanced—to adillos and a chinchillalike rodent. of native mammals uniquely adapted to have lived anytime before about 50 mil- The full significance of our finds that island’s conditions and every bit as lion years ago. Apparently geologists would take years to unfold, but we knew bizarre as those indigenous to modern had been a long way off in their estimate of the age of these rocks. Indeed, later analyses confi rmed that the new fossils came from well within the Cenozoic Cast of Characters era, the ongoing period of the earth’s During five seasons of exploring the Tinguiririca Valley of central Chile, our team history that began when the nonavian discovered fossils of 25 species of mammals. These ancient animals, most of dinosaurs went extinct 65 million years them new to science, lived about 32 million years ago. Known collectively as the ago. (Birds are now known to be thero- Tinguiririca fauna, the more than 300 individual specimens feature the oldest pods, thereby representing a living South American rodents, including an ancient chinchillalike form that is group of dinosaurs.) currently unnamed (1). Fossils of a new species of horselike Eomorphippus (2) and Santiagorothia chiliensis (3) are the oldest representatives of two Island Menagerie subgroups of notoungulates, hoofed herbivores that became extinct in the past finding fossils of any kind would 20,000 years. The fauna also includes South America’s greatest diversity of the notoungulates known as archaeohyracids—Archaeotypotherium have been great news for us. That they tinguiriricaense (4) among them. Unique to Tinguiririca are Pseudoglyptodon were mammals—and unexpectedly re- chilensis (5), the closest known relative of sloths, and shrewlike Klohnia cent ones at that—was more than enough charrieri (6), the only marsupial of its kind. —J.J.F., A.R.W. and R.C. motivation to focus our next fi eld season on that single area. We returned to the Tinguiririca Valley in the austral sum- mer of 1989, after the high mountain snows melted enough for local authori- ties to reconstruct the small access road, which washes out almost every spring. This time we reached the fossil site on a clear, sunny January morning with a crew of seven scientists and complete ●2 ●5 expedition gear. Quickly unloading the ●3 pack animals, we set up camp near a small creek and started hunting. ●4 To our delight, exquisite shards of ●6 bones and teeth popped into view with- ●1 in minutes of our beginning to comb the hillside. Protruding out both ends of one potato-size nodule of rock was a

RAÚL MARTIN MARTIN RAÚL skull that was unmistakably mamma-

www.sciam.com SCIENTIFIC AMERICAN 71 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. to science [see box on preceding page]. The era also saw the arrival of rodents African Immigrants and primates, neither of which were New World monkeys and caviomorph rodents (the group encompassing today’s among South America’s original mam- capybaras, chinchillas and their kin) were not original inhabitants of South mal inhabitants. America. Rather they arrived before 25 million years ago, while South America was an island. Some researchers have argued that the animals’ ancestors made the Seafaring Rodents relatively short water crossing from North America, but new fossils discovered in one of the most signifi cant of our central Chile strongly suggest that both groups are more closely related to Tinguiririca discoveries was a fossil of predecessors in Africa. Presumably the original colonists traveled from Africa to the earliest known South American ro- South America on fl oating islands of vegetation or by some other rare dispersal dent, a fi nd that lends powerful evidence mechanism that scientists may never be able to identify. to a debate over the origins of today’s living capybaras and chinchillas. Known as caviomorph rodents, these creatures and their immediate kin make up South ▼ MONKEY: Chilecebus carrascoensis America’s most ancient rodent lineage (and are distinct from the younger ro- TINY SKULL (fi ve centimeters long dent line of rats, mice and related crea- and 20 million years old) belonged to a small monkey that weighed a tures that arrived from the north about kilogram or less and may have 3.5 million years ago, when the Isthmus resembled a modern marmoset (right). of Panama first reconnected the two Americas). Paleontologists agreed that the first caviomorph rodents arrived sometime within the broad time span ▼ RODENT: Unnamed new species between 55 million and 25 million years ago, while South America was still an JAW FRAGMENT (two centimeters island. A few younger caviomorph fos- long) comes from a 32-million- sils hinted that the predecessors came year-old rodent and is one of the two oldest rodent from Africa, but many researchers found fossils ever found in it easier to imagine that the immigrant South America. This rodents made the shorter trip from one may have looked like North America, possibly via a chain of

a modern agouti (right). ); Caribbean islands. To help settle the debate, we compared the anatomical details of the Tin- monkey skull monkey

guiririca animal to rodent remains ( islands such as Australia (like the platy- now extinct (including certain marsupi- found elsewhere in the world. Most in- pus and koala) and Madagascar (fa- als and notoungulates). But no fossils formative was the shape of the tiny teeth mous for its lemurs). Unusual predeces- representing the transition to the second still rooted in the lower jaw (the upper sors of South America’s modern groups phase of South American mammal his- jaw and molars have yet to be found). ) include hopping marsupials; saber- tory—from about 40 million to 30 mil- That shape implied that the Tinguiririca

toothed marsupial “pseudocats”; arma- lion years ago—had ever been discov- animal’s upper molars had ﬁ ve distinct jaw rodent ( dillo cousins equipped with massive, ered. Most exciting for us during those crests—as did the upper molars of Afri- spike-studded tail clubs; bear-size ro- ﬁ rst few years, then, was the growing re- can rodents from the same period. In Field Museum of Natural History dents; sloths as big as elephants; and alization that the animals we found at contrast, North America’s ancient ro- sloths that swam in the sea. Tinguiririca lived during this period of dent species had only four crests on their Knowledge of the ancestors of South previously unknown history. upper molars. These comparisons America’s living mammals had been Paleontologists had suspected that strongly suggest that the Tinguiririca gleaned from earlier fossil discoveries in during this mysterious gap in the fossil rodent is more closely related to animals WEINSTEIN JOHN );

Patagonia and elsewhere, but critical in- record many of the continent’s unique in Africa. The absence of plausible cav- Field Museum of Natural History formation about many of these forebears lineages underwent explosive diversiﬁ - iomorph forebears in older North remained elusive. Paleontologists knew, cation. Indeed, our collection includes American fossil beds also supports the illustrations for instance, that sloths and anteaters got the earliest record of several groups of out-of-Africa theory. their start prior to 40 million years ago, notoungulates and represents at least 25 Presumably the original caviomorph RAÚL MARTIN ( MARTIN RAÚL as did several exotic lineages that are mammal species, nearly all of them new colonists traveled from Africa to South MARLENE DONNELLY

72 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. America on fl oating logs or other rafts JOHN J. FLYNN, ANDRÉ R. WYSS and REYNALDO CHARRIER together have explored the of vegetation—scientists’ best guess for fossil history entombed in the Chilean Andes for the past 20 years. Flynn is chairman how various unusual plants and animals and Frick curator of the division of paleontology at the American Museum of Natural made their way to many geographically History in New York City and dean of the museum’s new Richard Gilder Graduate School. isolated regions [see “Madagascar’s Wyss is professor of earth science at the University of California, Santa Barbara. Char- Mesozoic Secrets,” by John J. Flynn and rier is professor of geology at the University of Chile in Santiago. The authors thank the THE AUTHORS André R. Wyss; Scientifi c American, many dedicated fossil preparators, the National Museum of Natural History in Santiago, February 2002]. The idea of such an in- the National Monuments Council of Chile, the U.S. National Science Foundation, NASA, credible transoceanic journey may seem and the National Commission for Scientific and Technological Research of Chile for their far-fetched, but it is more plausible in exceptional support of this research. the context of the global environment before about 32 million years ago. At South American mammals known from crowned teeth with a thin cap of protec- that time, the South Atlantic was only about 65 million to 34 million years ago tive enamel covering each tooth only to about 1,400 kilometers at its narrowest were herbivores that browsed on typical the gum line. point—half as wide as it is today—and forest greenery, such as tree leaves and In sharp contrast, most of the Tin- east-to-west ocean currents in the trop- herbs. (Indeed, plant fossils do confi rm guiririca herbivores possessed extreme- ics were strengthening sporadically. that lush forests probably covered much ly high crowned teeth with enamel ex- These conditions would have per- of the continent during that period.) tending past the gum line nearly to the mitted a journey in about two weeks, Mammals that eat these soft foods, in- root tips, a condition known as hypso- and the animals may have gone into tor- cluding humans, typically have low- donty. The additional enamel (much por (inactivity and greatly reduced metabolism during times of stress). More- over, sea level was dropping at the time Teeth Tell All (because of the formation of ice sheets Most of the 32-million-year-old fossil on and around Antarctica), so one or herbivores discovered in Chile’s more volcanic “stepping stone” islands, Tinguiririca Valley have high-crowned, or now submerged, may have made the hypsodont, cheek teeth like those modern crossing easier. cattle use to grind gritty, fi brous plants. These specialized molars suggest that the Emerging Ecosystems Tinguiririca region was a dry grassland ▲ HIGH-CROWNED TEETH using a new and extremely precise when those animals lived. Earlier, much of The broken right lower jaw of a dating method that analyzes trace South America was heavily forested, and a sheep-size Tinguiririca herbivore amounts of argon gas trapped inside much smaller proportion of its herbivore displays the abrasion-resistant crystals of the fossil-bearing rocks, we species were hypsodont (red line in graph). teeth typical of grassland grazers. determined that the rodents and other Global climate change (dark blue line) to The crowns, which are coated with Tinguiririca mammals were between 33 cooler and thus drier conditions around 34 protective enamel, extend well ); ); million and 31.5 million years old. million years ago (light blue bar) may explain below the gum line toward the root

) Growing Antarctic ice sheets and other why the lush forests dried out. tips—providing more material to phenomena indicate that global climate wear down as the teeth grow. at that time was becoming cooler and 20

high-crowned teeth high-crowned drier. Knowing that a major shift in climate occurred just as the Tinguiririca low-crowned teeth low-crowned 4

( Average mammals were thriving led us to test 25 4Fraction of global herbivores sea-surface whether the animals and their environ- that are temperature ment might have been responding to hypsodont those changes. 30 ▲ LOW-CROWNED TEETH Several lines of indirect evidence In the right lower jaw of this as yet unnamed ancient Chilean herbi- have enabled us to reconstruct the habi- 35 tat of the Tinguiririca mammals even Cooler vore, protective enamel covers though we have never found plant fos- Ago Years of Millions only the portions of the crowns that sils from the same rocks. Our early anal- 40 extend above the gum line. Such American Museum of Natural History History Natural of Museum American ysis of their teeth told us that the Tin- teeth are ill suited to a gritty diet; guiririca animals must have lived in a Greater they usually belong to animals that 45 very different ecosystem than their im- browse on leaves or other soft food. ALLISON SMITH SMITH ALLISON JOHN J. FLYNN, ANDRÉ R. WYSS R. REYNALDO ANDRÉ AND ( CHARRIER FLYNN, J. JOHN mediate forebears did. Most of the older

www.sciam.com SCIENTIFIC AMERICAN 73 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. harder than the dentin in tooth interi- dental types generally increases with the vided two other independent deductions ors) makes hypsodont teeth more wear- amount of open habitat, and the Tin- about the annual rainfall and vegetation resistant than low-crowned teeth are. guiririca level of hypsodonty surpasses that dominated the ancient Tinguiririca The Tinguiririca herbivores almost cer- even that observed for mammals living ecosystem. Statistical analyses of the tainly evolved such teeth in response to in modern, open habitats such as the number of species within different body- abrasive particles within the foods they Great Plains of central North America. size categories and their ecological at- ate, as did cows, antelopes, horses and The fi ndings implied that the Tin- tributes—so-called cenogram and mac- other animals that eat gritty grasses in guiririca herbivores grazed in open roniche analyses—revealed that the Tin- open prairies and savannas elsewhere in grasslands rather than in forests, but the guiririca animals most closely resembled the world. It is also signifi cant that two teeth are not alone in pointing to this modern fauna that live in dry grasslands thirds of species of the entire Tinguirir- conclusion. Our former graduate stu- with patchy woodlands, such as parts of ica fauna were hypsodont. The propor- dent Darin Croft, now a professor at the savanna of Africa or the Caatingas tion of hypsodont taxa relative to other Case Western Reserve University, pro- and Chaco habitats of South America. Our conclusion that the ancient Tin- guiririca habitat was open, relatively dry and contained abundant grasses came as a bit of a surprise, considering that all previous evidence suggested that the fi rst open grasslands on other continents did not appear until 18 million years ago. The apparent emergence of the Tin- guiririca grasslands some 15 million years earlier could have resulted from the global trend toward aridity and cooling around that time—perhaps accentu- ated by the rain shadow cast by the rising Andes Mountains. Grasslands would have been better suited to a cooler, drier climate than the lush forests of previous millennia were. So far, though, suspi- cion that grasslands resulted from global cooling needs more scrutiny. Further testing a direct causal connection is an ▲ FOSSIL-CONTAINING LAYERS, pushed upward and tilted almost vertically, are avenue of future research. spectacular evidence of the tectonic forces that have been squeezing central Chile for millions of years. These steep slopes are located near Lake Laja, about 300 kilometers No Monkeying Around south of the Tinguiririca Valley. having uncovered such a rich trove of paleontological and environ- Late to Rise mental information at several sites in the Tinguiririca Valley, we began to consid- Making the fossil mammals of the Chilean Andes all the more signifi cant are the er whether this part of the Andes was a ways they have substantially advanced understanding of the geologic history of paleontological “one-hit wonder.” Soon this segment of the Andes. The unexpectedly young age of the ancient mammals after, an unexpected obstacle led us to overturns prevailing assumptions about the age of their host rocks, indicating an answer. In the spring of 1994 recon- that this part of the Andes formed at least 70 million years later than previously struction of the road leading into the thought. Most of the rocks making up the spine of the main range of the central Chilean Andes, fi rst named and described by Charles Darwin, were long believed to Tinguiririca Valley was especially slow, date back at least 100 million years, and the initial phases of uplift were thought but we did not learn of this impediment to be similarly ancient. Dating these fossil-bearing deposits enabled our team to until arriving on the scene. Turning frus- calculate, for the fi rst time, precisely when this part of the Andes began rising: tration into an opportunity, we began between 15 million and 18 million years ago. The rise continued episodically until explorations outside the valley. the present and still goes on today. It is now clear that extensive, sediment-fi lling We searched other major valleys basins formed during volcanic phases in the range’s history, both shaping the sporting outcrops of the same volcani- Barbara Santa California, of University area’s unique ancient ecosystems and providing a means of preserving its cally derived sedimentary rocks, which spectacular fossil mammal record. —J.J.F., A.R.W. and R.C. are exposed across thousands of square

kilo meters of mountainous terrain. GANS PHILLIP

74 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. HERBIVORE SKULLS, as fi rst seen in the fi eld, are preserved in a known as notoungulates. The high-crowned teeth typical of grass- coarse matrix of fossilized ash and mud from ancient volcanic eaters are particularly notable on the specimen at the left. The other eruptions. Even before researchers removed the bones from the rock, skull, which comes from an older leaf-eating notoungulate called a they identifi ed both specimens as types of extinct hoofed herbivores notopithecine, was recently discovered near the Teno River.

Over the ensuing years of fi eldwork, we rins. As with the caviomorph rodents, cipitous topography and the remoteness determined that fossil mammals were experts had long debated whether New of many of the localities. Some sites lie indeed not restricted to the Tinguiririca World monkeys originated in North within a few kilometers of gravel roads Valley and that the volcanic torrent that America or Africa. But anatomical de- or dirt trails, but most can be reached inundated that ancient landscape was tails of the Chilecebus skull and teeth only by long approaches on foot, horse or not a one-time, isolated cataclysm. Rath- argue for its common heritage with a even helicopter. We jokingly refer to er, seen over a time frame of millions of group of primates originating in Africa. “Andy’s Rule” (dubbed in honor of years, such devastating events occurred Like the caviomorph rodents, it seems Wyss), which says that the diffi culty of with some frequency. Each time, count- the ancestors of Chilecebus somehow getting there is proportional to the quan- less layers of material from additional made the Atlantic crossing from Africa. tity and quality of fossils we will fi nd. eruptions buried the older deposits (and Beginning with the Tinguiririca fau- Together the Cenozoic fossils of the the bones they contained) ever more na and culminating with the New World Chilean Andes are helping to elucidate deeply. Ultimately, this layered stack of monkey and other ongoing discoveries mammalian evolution and environmen- sediment (now transformed into rock) across central Chile, volcanic deposits tal transformations in South America, a and lavas measured upward of three ki- once ignored for fossils have turned out continent whose long history of isola- lometers thick. Later, collision of con- to contain superbly preserved bones and tion represents a splendid natural ex-

) verging tectonic plates squeezed this are now recognized as a premier archive periment for investigating large-scale unruly pile skyward. of South American mammal evolution. evolutionary phenomena. right ( Our ongoing analyses of the numer- Over the years, we have developed a keen ous faunas, which range between 10 mil- sense of what auspicious rocks look like, lion and 40 million years in age, are re- sometimes even being able to spot them Learn the story behind “Andy’s sa Rule”—photographs included— vealing fresh insights into the region’s from many kilometers away. These fos- at www.sciam.com/ontheweb history. One of our most signifi cant new- sils are hard-won, though, given the pre- er fi nds—from a site about 100 kilometers north of Tinguiririca, in the drain- MORE TO EXPLORE age basin of the Cachapoal River—is the Splendid Isolation: The Curious History of South American Mammals. George Gaylord Simpson. Case Western Reserve University University Reserve Western Case most complete skull of an early New Yale University Press, 1980. World monkey yet discovered. The fi ve- Cenozoic Environmental Change in South America as Indicated by Mammalian Body Size centimeter-long skull, with both eye Distributions (Cenograms). Darin A. Croft in Diversity and Distributions, Vol. 7, No. 6, sockets and every tooth in the upper jaw pages 271–287; November 2001. ); DARIN CROFT CROFT DARIN ); The Tinguiririca Fauna, Chile: Biochronology, Paleoecology, Biogeography, and a New

left intact, came from a petite monkey weigh- Earliest Oligocene South American Land Mammal “Age.” J. J. Flynn, A. R. Wyss, D. A. Croft and ing about one kilogram at the most. R. Charrier in Palaeogeography, Palaeoclimatology, Palaeoecology, Vol. 195, No. 3/4, Named Chilecebus carrascoensis, this pages 229–259; June 15, 2003. creature resembled modern New World For more details about the mammal lineages that inhabited South America when it became

ANDRÉ R. WYSS ( WYSS R. ANDRÉ monkeys, such as marmosets and tama- geographically isolated, visit www.dcpaleo.org/Research/SAMammals/SAMammals.html

CARBON NANOTUBES in a weblike mesh ensure multiple alternative pathways for electrons (pink highlights), providing sureﬁ re electrical conduction. The entire ﬁ eld of view is 0.7 micron in diameter. ; AND AND APL MicroscopicAPL Carl Zeiss SMT COLORIZED BY JEAN-FRANCOIS PODEVIN JEAN-FRANCOIS BY COLORIZED ANDREW PAUL LEONARD LEONARD PAUL ANDREW JOHN YORSTON JOHN

76 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Random networks of tiny carbon tubes could make possible low-cost, flexible devices such as “electronic paper” and printable solar cells

By George Gruner

In many classic science-fiction stories, alien life is based on silicon—the substance at the core of modern electronics technology—rather than on carbon, the fundamental building block of earth- ly biology. Scientists have even speculated that they might someday create silicon life-forms. Instead the opposite is starting to happen: carbon is serving as the foundation for electronic devices—and in the process is breathing new life into the quest for inexpensive, fl exible products that offer a broad range of capabilities. These developments may surprise those of us who learned in high school that carbon, in its familiar incarnations of diamond and graphite, does not conduct electricity well. During the past 15 years, however, researchers have discovered new forms of carbon: very small structures comprising a few hundred to 1,000 atoms, through which electrons travel with ease. Of particular interest is the carbon nanotube, a molecule that resembles rolled-up chicken wire, only the “wire” is a sheet of carbon atoms that is 100 million times as small as the version used for chicken coops. Investigators have found that random networks of carbon nanotubes —called nanonets—can perform a variety of basic electronic functions. Using novel chemistry, researchers can make such networks mimic the conductive properties of metals such as copper or the less conductive characteristics of semiconductors such as silicon. These innovations have paved the way for this single material to assume different roles in electronic devices. Further, engineers can construct such carbon-based devices by employing simple fabrication methods. Researchers can dissolve the tubes in a liquid and www.sciam.com SCIENTIFIC AMERICAN 77 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. spray the resulting solution to form thin layers on, say, fl exible FROM NANOTUBES TO NANONETS plastic sheets. They can also lay or print these materials on other layers that have various electronic functions; for ex- Thin networks of carbon nanotubes, or nanonets, can serve ample, substances that emit light when a voltage is applied. as electronic devices. Each nanotube is a one-atom-thick It takes little imagination to see how this kind of straight- sheet of carbon, called graphite, rolled into a cylinder with forward system could form the basis of many extremely cheap a diameter of about a nanometer (inset)—approximately but handy products: “electronic paper” that could display 50,000 times as small as the width of a human hair. Electric information on sheets that roll up like conventional news- current passes through the interconnected tubes from one print; chemical sensors; wearable electronic devices; solar electrode to another. cells that could be printed onto rooftop tiles; or scads of sim- Nanonet ple radio-frequency identifi cation (RFID) sensors for moni- Electrode toring warehouse or store inventories. For such applications, the expensive, lightning-fast processing power of integrated chips like Intel’s Pentium processors or Samsung’s video displays is not needed; rather R&D laboratories and start-up fi rms are racing to fi nd technologies that can do the job well enough at low cost [see table on page 83]. Such exciting applications would make severe demands on today’s electronic materials: they would Electrode need to be conductive, fl exible, lightweight, transpar- Carbon nanotube Current pathways ent (at least for some applications, such as solar cells and displays) and inexpensive. But most conductors are metals, the majority of which are not transparent, Carbon sheet (graphite) whereas, as a rule, thin fi lms of materials that are transparent, such as diamond, are insulators (they do not transmit electricity). Light can pass through one spe- can easily take on multiple forms and are, of cial class of metals, called metal oxides, however. The best course, dirt cheap [see “Next Stretch for Plastic known, indium tin oxide, is frequently used where engineers Electronics,” by Graham P. Collins; Scientifi c American, need see-through electrodes. But metal oxides are costly. They August 2004]. On the downside, weak bonds hold together are also heavy and brittle, and their manufacture requires the atoms in most plastics. The bonds can break rather read- high processing temperatures and multibillion-dollar fabrica- ily, which leads most polymers to degrade over time. Con- tion facilities. sider just how useful a solar cell would be if it failed after only Another alternative is an unusual category of plastics a few warm, sunny days. known as conducting polymers. Although common plastic substances are insulators, chemists have in recent decades A Better Wire managed to convert some polymers into semiconductors and enter carbon-based nanowires. Carbon nano- even full-fl edged conductors. Polymers can be produced using tubes were fi rst discovered several decades ago, but no one room-temperature techniques. Lightweight and fl exible, they realized their value at the time. Then, in 1991, Japanese chemist Sumio Ijima of NEC Corporation rediscovered them. These tiny tubes of carbon have a diameter of around one Overview/Nanonet Electronics nanometer—about the same as a strand of a DNA molecule ■ Carbon nanotubes—minuscule cylinders of rolled-up [see box above]. The electrical conductivity of the tubes is carbon sheets—conduct electricity well, which could comparable to that of copper and surpasses that of any poly- make them useful for many exciting electronic mer by several orders of magnitude. They can also carry more applications. But manufacturing products that use than 100 times more current than the best metals. Carbon single tubes is expensive and suffers from signifi cant nanotubes are, in addition, physically robust: they can be bent reliability problems. easily, they do not react with most chemicals and they resist ■ Random networks of many carbon nanotubes, called damage from day-to-day use. nanonets, enable numerous basic electronic functions Manufacturers make nanotubes by reducing coal into its at low cost. The durable nature of nanonets also makes component atoms using the heat of an electric arc or a laser, them suitable for portable devices. which creates a so-called carbon plume. They then add cata- ■ Carbon nanonets should fi nd use in sensors, solar cells, lysts to the plume, which promotes the formation of various electronic paper, and fl exible touch screens and types of carbon molecules. This relatively straightforward displays within a few years. procedure produces what is essentially soot—carbon mole-

cules in many forms, including spherical structures called JEAN-FRANCOIS PODEVIN

78 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. buckyballs as well as other “fullerenes” and carbon nanotubes. a few sticks, the chances of fi nding a connected pathway (by Fabricators must then laboriously separate the nanotubes out going from one stick to the next) from one end of the table to of the mixture. Techniques focus on separating out only the the other are slim. In fact, below a certain critical density of long, nearly perfect specimens that have a single “chicken sticks, the odds drop to zero. But as the number of sticks in- wire” wall (rather than multiple, concentric walls). Suitable creases, the pile will eventually surpass that critical density, nanotubes are thus currently quite pricey, but makers are con- the so-called percolation transition, where at fi rst one and then fi dent that costs will drop signifi cantly if market demand rises more and more pathways form. If the pickup-stick approach enough to justify high-volume manufacturing facilities. were applied to copper wires on the tabletop, at some point the When a single nanotube is employed to build a transistor, network would achieve electrical conduction across the table the voltage-activated switch that is the workhorse of modern as well—with the current dependent on the density of the cop- electronics [see “Nanotubes for Electronics,” by Philip G. per wires. Theorists studied this concept some time ago, and Collins and Phaedon Avouris; Scientifi c American, De- my group at the University of California, Los Angeles, was cember 2000], the resulting device can easily outperform the able to map out such a transition in networks of nanotubes. transistors on the silicon chips in today’s computers. But sin- Nanonets can be, in addition, highly transparent—an ad- gle carbon nanotubes will not replace silicon and copper in vantage in applications that require light transmission. Just as the foreseeable future. The main obstacle lies in their manu- the freeway pavement covers only a small fraction of the natu- facturability, which is one of the most vexing problems affl ict- ral terrain, a web of long and skinny wires allows passage of ing nanotechnology’s commercialization. Current devices most of the incident light—a fraction that approaches 100 per- based on a single nanotube can take days to make, because cent for what can be considered one-dimensional nanowires. they often must be assembled by hand. Another diffi culty is Finally, much like a spiderweb, a network of nanowires performance variability. Nanotubes come in slightly different typically is more robust than the same material in undifferen- shapes and forms, which affects their electrical attributes. tiated bulk, which often tends to break when bent. These From One Wire to a Network CHEAPER SOLAR CELLS although individual tubes differ from one to the next, researchers realized that this variation could be aver- Transparent and fl exible carbon nanonet electrodes would aged out by using many tubes together—any shortcomings be signifi cantly cheaper than the rigid indium tin oxide present in some of the tubes could be compensated for by bet- electrodes used in today’s solar cells (shown in exploded ter-performing counterparts. The simplest example is a ran- view). The carbon nanonets would conduct electric charges (negative electrons and positive “holes”) dislodged by dom network of nanotubes [see box on opposite page]. Just light photons from an active, charge-creating, as an interstate highway system can offer alternative layer of a semiconductor to an external routes when you encounter a traffi c jam on one roadway, circuit. A sample printed nanonet solar cell so, too, can a random assembly of electrically conduct- shows its bendability (inset). ing nanotubes—a nanonet—speed the transmission of electrons by providing alternative pathways. Investi- Sun

) gators soon established that these nearly two-dimensional random networks offered interesting properties in their own right. First, the nanonet’s many pathways and connections Electrical circuit guarantee good electrical conductance between one electrode Light and another, despite possible manufacturing ﬂ aws. A good photon analogy is the freeway system that serves the Los Angeles metropolitan area. No one would want to attempt to traverse solar cell, Stanford University Stanford cell, solar ( the City of Angels by hiking cross-country or driving the slow,

Redux Redux stoplight-strewn surface roads; instead travelers take the freeway. The same concept applies to the nanonet, which allows electrons to jump on the tubes and move around on what is essentially a nanoscale freeway system. The multiple avenues provided by these networks also afford a considerable resis- Transparent tance to failure, or fault tolerance; if one route breaks from nanonet use, others are there to take up the slack. electrode A conductive nanonet is in fact a simple example of the Electric current Active layer concept of percolation, which describes how objects, materials Hole Bottom electrode or electric currents move through a random medium. Imagine Electron

JEAN-FRANCOIS PODEVIN; BRAD SWONETZ SWONETZ BRAD PODEVIN; JEAN-FRANCOIS dropping pickup sticks on a tabletop one at a time. With only

www.sciam.com SCIENTIFIC AMERICAN 79 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. characteristics make the nanonet architecture eminently suit- them apart. Some agents, called surfactants (soaps), accom- ed for applications in which resistance to day-to-day use and plish this job by completely surrounding the tubes. But sur- misuse is important. Think about how many times you have factants, if they remain on the tubes after they are sprayed dropped your cell phone or iPod. onto a surface, impede the fl ow of electrons between tubes (blocking the freeway ramps, so to speak). Through steady Weaving a Nanonet trial-and-error efforts with innumerable solvents, surfactants these performance benefits augur well for the and processing procedures, however, researchers have created technology’s potential in real-world applications, but any new simple (room-temperature) avenues to make such thin fi lms replacement material must, of course, be more than competi- of nanotube networks. At the moment, a method pioneered tive in terms of function and cost with current materials. by my team and a group led by chemist Andrew Rinzler at the Nanotube fi lms initially made a couple of years ago—by my University of Florida yields fi lms that have the lowest electri- team, by a group led by physicist Siegmar Roth at the Max cal resistance and thus the best operating performance to date Planck Institute for Solid State Research in Stuttgart, Ger- among nanonet-based devices. many, and by one at the University of Texas at Austin—were As researchers experimented with the conductivity of the not up to the task. Finding the optimal processing routes and tubes, they learned that the material could be transparent, a the most advantageous way to deposit the tubes onto surfaces property that is important for applications such as displays were not trivial problems. and solar cells. The discovery that carbon nanonets are trans- Clearly, one cannot fabricate thin fi lms of such networks parent to light came about as a by-product of research on their by merely throwing down a tube at a time like playing pickup conductivity. The fi rst indication that nanonets could be clear sticks; another strategy is needed. One might, for instance, arose in 2001, when my former postdoctoral associate Leon- dissolve the tubes in a solvent (water, alcohol, organic liquids) ardo Degiorgi and his group at the Swiss Federal Institute of and then spray the resulting fl uid onto a surface, but that is Technology in Zurich, along with physicist David Tanner and not as easy as it sounds. When mixed in a liquid, the tubes his co-workers at the University of Florida, studied their opti- tend to bundle together, requiring a chemical additive to keep cal characteristics. To measure nanonet conductivity precisely, they fabricated thick fi lms: these were too deep to transmit TRANSPARENT TRANSISTORS light, but their data led them to conclude that a thinner fi lm would be both transparent and a good conductor. After these Carbon nanonet fi lms, tailored to perform like semi- groups made this determination, Rinzler’s team (collaborat- conductors, can serve as the basis for fi eld-effect ing with Katalin Kamarás and her colleagues at the Central transistors—the building blocks of computers, cellular Institute of Physics in Budapest) and mine at U.C.L.A. fol- phones and other digital devices. This switchlike lowed up with direct measurements of a nanonet fi lm’s optical mechanism (shown in exploded view) uses a small electric voltage provided by the gate electrode to greatly boost the transparency. Today scientists can fabricate tailor-made fi lms current in the source-drain circuit. In the inset, a with different levels of transparency and electrical conductiv- technician bends a plastic sheet onto which an array of ity by changing the thickness of the fi lms. see-through nanonet transistors has been printed. Nanonet Transistors Source-drain circuit researchers soon turned their attention from making nanonet conductors to nanonet semiconductors that could Source serve as the basis for transistors. A transistor requires materi- ) electrode Nanonet als whose conductivity changes greatly in response to only semiconductor Drain small incremental inputs, such as altering an electric fi eld [see electrode box at left]. The notion that carbon nanotube networks could serve as transistor, Unidym transistor, the backbone of thin-fi lm fi eld-effect transistors emerged (

around seven years ago. Thereafter, progress was relatively Redux rapid—with advances in creating nanonets on ﬂ exible substrates, demonstrating the transparency of the devices, coming Gate in short order. Working in parallel, my R&D group at Nano- electrode mix, a start-up ﬁ rm in Emeryville, Calif., where I served as Gate circuit chief scientist, and a research team at the Naval Research Current pathway Laboratory in Washington, D.C., led by materials scientist Insulator Eric Snow, produced nanonet transistors in 2003. But these devices were formed on rigid glass substrates at processing temperatures of 900 degrees Celsius—too

hot for use with ﬂ exible plastic substrates that melt at SWONETZ BRAD PODEVIN; JEAN-FRANCOIS

80 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. LOW-COST SENSORS organic or polymeric metals and some semiconductors. For these uses, electronic materials must exhibit good electrical A nanonet device can become a sensor with the addition of conductance (otherwise, applied current heats them up, re- “recognition molecules” that react with a target chemical or sulting in power losses) and high optical transparency (be- biological molecule. When the target binds to a recognition cause the viewer of a display, for example, needs to see the molecule, it alters the sensor’s electrical output. Such layers that lie underneath). devices can detect many chemicals, including a blood-borne Such substances will enable the development of what peo- cancer marker called prostate specifi c antigen (PSA) and, ple variously call printed, plastic, disposable or macroelectron- soon, microorganisms such as the anthrax bacterium, a ic products. One example is the photovoltaic cell. Typical solar potential bioweapon. Arrays of nanonet sensors, each with cells made of single-crystal silicon have excellent performance different recognition molecules, could cheaply detect (they convert as much as 18 percent of incoming light into elec- specifi c genes or proteins for medical purposes. The inset shows a nanonet-based detector chip tricity) but are bulky, heavy and costly to manufacture. (black) on a printed circuit board. Instead imagine a razor-thin solar cell that, though less effi cient (converting only 5 or 6 percent of in- Sample Target coming light), is signifi cantly cheaper to fabricate droplet chemical or and offers the potential for easy mass production biological Recognition molecule of large-area systems, both of which can compen- molecule sate for the material’s lower performance levels [see box on page 79]. Nanonet In a solar cell, incoming sunlight dislodges electrons transistor and their positively charged counterparts, called holes, in the middle layer of the device. The electrons then migrate to one electrode, power some electrical load and return to the holes via another electrode to complete the circuit. Several companies are working to perfect a cell’s active (charge-creating) layers using advanced polymers and other substances that are transparent and fl exible. Together with Michael McGehee’s Other materials science group at Stanford University and physical chemicals or chemist Niyazi Serdar Sariciftci of the University of Linz in biological Electrical circuit molecules Austria, my U.C.L.A. team has produced fl exible, proof-of- concept solar cells with nanonet electrodes that exhibit performance comparable to that of indium tin oxide electrodes. 120 degrees C. Nanomix researchers Keith Bradley and Jean- Also under consideration are nanonet-based fi lms that Christophe Gabriel, in collaboration with my U.C.L.A. team, would lie at the heart of an inexpensive, fl exible and light- manufactured the fi rst fl exible nanotube network transistors weight touch screen or visual display. A touch screen, for in- on plastic in 2003. Soon afterward, my colleagues and I at stance, consists of two sheets of electrodes separated by insu- U.C.L.A., working with Roth’s group at the Max Planck In- lating spacers. When a fi nger touches the top sheet at some stitute, managed to fabricate devices that were also transpar- point, the electrodes there meet, completing an electrical cir- ent, making them suitable for applications such as portable cuit specifi c for that location that is formed by a pattern of visual displays. Physicist John Rogers and his colleagues at the smooth, thin layers of conductive materials that have been University of Illinois achieved similar success only a few imprinted on the bottom sheet. In collaboration with Richard ) months later. Although these fi eld-effect transistors operated Kaner’s group at U.C.L.A., my team has fabricated and tested

sensor at fast rates—the key metric for such devices—other necessary proof-of-concept devices based on nanonets. characteristics, such as low-voltage function, were lacking. Nanonets also work in light-emitting diodes, which re- The goal was to run the devices at voltages less than those stan- semble photovoltaics that run in reverse so that they create dard batteries can provide to save power, but this feat was at- tained only recently by Rogers and by chemist Tobin Marks GEORGE GRUNER, a Distinguished Professor of Physics at the of Northwestern University, who employed specially made University of California, Los Angeles, investigates fundamental polymers to insulate the devices’ conductive parts. issues relating to nanoscale phenomena. He has served as the chief technology ofﬁ cer at Nanomix in Emeryville, Calif., and he Nanonets in Action is founder of Unidym in Menlo Park, Calif.; both companies con-

carbon nanonets can offer distinct advantages in THE AUTHOR centrate on developing novel applications of nanotechnology. many portable products, a conclusion that becomes more ob- Gruner’s interests include skiing, classical music and technol- vious when one compares them with some of the current con- ogy assessment—not necessarily in that order. His research is

JEAN-FRANCOIS PODEVIN; NANOMIX “SENSATION” ( “SENSATION” NANOMIX PODEVIN; JEAN-FRANCOIS tenders for these applications, including ﬁ lms composed of supported by the National Science Foundation.

www.sciam.com SCIENTIFIC AMERICAN 81 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. light when electricity passes between the electrodes. In col- must run rapidly so that the images can be readily refreshed. laboration with Marks’s group at Northwestern, my team has Of course, the kind of portable devices that will use these recently demonstrated proof-of-concept light-emitting diodes displays will need power sources as well—cheap, lightweight, with excellent performance (suffi cient to meet the require- razor-thin and disposable batteries and supercapacitors. ments for use in televisions, for example), as has a research Nanonets could also play an important role in such power group at the University of Montreal led by Richard Martel. devices, serving not only as electrodes but as high-surface- Transistors made from nanonets will, in addition, fi nd use area components for collecting electric charge to store it for in printed electronics. Tests indicate that the operating speed later discharge. of carbon nanonets lags somewhat behind that of crystalline silicon, from which most integrated chips are fabricated, but Many Pathways to Take their conductivity and durability advantages over polymers the nascent carbon nanonet industry has only just make them attractive to device manufacturers. Although begun to perfect this fl edgling technology. There is little doubt nanotube fi lms cannot yet work in laptop computers or televi- that the recent feasibility studies that I have described will sion sets, they are competitive in many other products—espe- soon be followed by working prototypes and eventually prod- cially those that require a material that is cheap, fl exible, light- ucts based on those new devices. Today the industry is at the weight, environmentally friendly and resistant to abuse. The stage where the silicon chip business was half a century ago. fi rst such application is expected to be large-area visual dis- The nanotubes are improving steadily, and researchers are plays, called active-matrix displays. The transistors in a display successfully sorting those that conduct electricity as well as PRINTING A NANONET

Prospective makers of products based on carbon nanonets surface. The primed stamp presses down on the surface are developing several inexpensive ways to “print” an of a substrate, printing the nanonet pattern onto it. engineered pattern of the material onto a flexible polymer Manufacturers are also working on two mass-production surface to produce, for example, an electronic circuit. The techniques, including the use of standard ink-jets (bottom simplest method resembles using an ink pad and a stamp left) to spray a liquid containing dispersed nanotubes (top). A patterned stamp comes into contact with a nanonet onto substrates, and a variant of offset printing, in which layer, parts of which stick to the stamp's bottommost a nanonet solution substitutes for ink (bottom right).

STAMP PRINTING

Patterned stamp

Nanotube fi lm Patterned nanotube fi lm Flexible substrate Patterned fi lm on substrate (negative)

INK-JET PRINTING OFFSET PRINTING Plate cylinder Negative image Printed pattern

Impression cylinder Substrate Electrostatic plates (for directing the Coated plate Flexible droplets of nanonets) substrate Spreader Printed pattern blade Inking roll

Nanonet solution Ink-jets reservoir JEAN-FRANCOIS PODEVIN

82 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Carbon Nanonet Research and Product Development Many R&D organizations are producing or working to develop carbon nanotube materials, carbon nanonet fi lms and the electronic devices that incorporate them. A new technology typically passes through the following sequence of developmental stages: concept, R&D, proof of concept, prototype, product development and production. ORGANIZATION PRODUCT FOCUS STATUS HIGH-GRADE MATERIALS FOR ELECTRONICS CarboLex, Lexington, Ky. (www.carbolex.com) Electric arc– and chemical vapor deposition (CVD)–based Production fabrication Carbon Nanotechnologies, Houston CVD- and carbon monoxide–based fabrication Production (www.cnanotech.com) Carbon Solutions, Riverside, Calif. Electric arc–based fabrication Production (www.carbonsolution.com) SouthWest NanoTechnologies, Norman, Okla. CVD-produced specialty nanotubes Production (www.swnano.com) Thomas Swan, Consett, England High-volume CVD-based fabrication Production (www.thomas-swan.co.uk) TRANSPARENT FILMS Battelle Memorial Institute, Columbus, Ohio Transparent coatings R&D (www.battelle.org) Eikos, Franklin, Mass. (www.eikos.com) Conducting ink Product development Eastman Kodak, Rochester, N.Y. Transparent optical coatings R&D, prototype (www.kodak.com) Unidym, Menlo Park, Calif. (www.unidym.com) Films for touch screens, solar cells, light-emitting diodes Product development DEVICES DuPont, Wilmington, Del. Transparent electronics R&D (www.dupont.com) IBM, Armonk, N.Y. (www.ibm.com) Computer-compatible transistors and interconnects R&D Intel, Santa Clara, Calif. (www.intel.com) Interconnects R&D Motorola, Schaumburg, Ill. Biological and chemical sensors Prototype (www.motorola.com) Nanomix, Emeryville, Calif. Chemical and biological sensors Product development, (www.nano.com) R&D Nantero, Woburn, Mass. Novel memory technology Prototype (www.nantero.com) Samsung, Seoul, South Korea Displays R&D (www.samsung.com) Unidym (see above) Printed electronics for displays Proof of concept metals do from those that are semiconductors, which will MORE TO EXPLORE further better device performance. Meanwhile investigators Carbon Nanotubes. Mildred Dresselhaus, Gene Dresselhaus, have made progress on a process that resembles silicon dop- Peter Eklund and Riichiro Saito in Physics World, Vol. 11, pages 33–38; ing, in which special molecules are attached to the tubes to January 1998. Carbon Nanotubes and Related Structures. Peter J. F. Harris. fi nely alter their electrical properties. Many observers believe Cambridge University Press, 2001. that it is only a matter of time before such fi lms exceed the Nanotubes for Electronics. Philip G. Collins and Phaedon Avouris performance of traditional metals and start to make inroads in Scientifi c American, Vol. 283, No. 6, pages 62–69; December 2000. into silicon-based digital electronics technology. Carbon Nanotubes—The Route Toward Applications. R. H. Baughman, A. A. Zakhidov and W. A. de Heer in Science, Vol. 297, pages 787–792; Carbon nanonets have just recently left the realm of sci- August 2, 2002. ence fi ction and entered that of practical reality. Like silicon, Carbon Nanotube Films for Transparent and Plastic Electronics. this budding technology is highly unlikely to lead to artifi cial George Gruner in Journal of Materials Chemistry, Vol. 16, No. 35, life anytime soon, but it has every chance of enabling innova- pages 3533–3539; 2006. Carbon Nanotube Transistors for Biosensing Applications. tive products that in the not too distant future will improve G. Gruner in Analytical and Bioanalytical Chemistry, Vol. 384, our everyday lives. pages 322–335; 2006. www.sciam.com SCIENTIFIC AMERICAN 83 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Eyes Open, Brain R At other times they keep their eyes shut and appear to be asleep, asleep, be to appear and shut eyes their keep they times other At will show some refl viduals in this condition do not open their eyes; at best, they they at best, eyes; condition do not their open this in viduals feelings patients will not fi patients will riod, but in rare instances they may they briefl instances but rare riod, in refl

if the damage is severe, the victim will slip into Indi- acoma. will victim issevere, the damage the if lasts longer than two to two longer ﬁ lasts than tance and can make a variety of spontaneous movements of spontaneous avariety make can and tance assis- breathe without technical can patients usually to. These wander. sometimes eyes and open awake, to be their seem they trauma to the brain (often after aroad accident) after (often or alack brain to of the trauma mains intact, but awareness intact, mains object or turn ﬂ turn or object although they may open them and stir when touched or spoken spoken or touched when stir and them open may they although when a vegetative times Atthe cycles. state have sleep/wake as grinding teeth, swallowing, crying, smiling, grasping anoth- grasping smiling, crying, swallowing, teeth, grinding as oxygen (for example, after a cardiac arrest or drowning), but or drowning), arrest acardiac oxygen (for after example, others will awaken from their coma but remain unconscious, unconscious, remain but coma their from awaken will others er’s hand, grunting or groaning or grunting hand, er’s entering what is called the vegetative state. vegetative the called is what entering sciousness can become completely dissociated: wakefulness re- still and die, will Others days. within so do typically sciousness condition. It illustrates how the two main components of con- how main two condition. the Itillustrates exive and not the result of purposeful behavior. Typically ofexive not purposeful result and the Even for experts, the vegetative state is a very disturbing disturbing vegetative isavery state the Even for experts, Doctors can save can lives of the Doctors patients many who suffer — number of people who survive acute brain damage. damage. brain acute survive who people of number ecent progress in medical care has greatly increased the increased greatly has care medical in progress ecent is abolished. By wakefulness, I mean that patientsis abolished. Imean in By wakefulness, By Steven Laureys eetingly toward a loud sound. x their eyes on anything for a sustained pe- for asustained eyes onanything x their COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. AMERICAN, SCIENTIFIC 2007 COPYRIGHT ex movements of the limbs. Coma rarely Coma exmovements limbs. of the ve weeks. Those who regain con- who regain ve Those weeks. — encompassing all thoughts and and thoughts all encompassing — but these motions are always are motions but these y follow amoving — such such

CREDIT New brain-imaging techniques are giving researchers Shut a better understanding of patients in the vegetative state

HALFWAY WORLD: People in the vegetative state are awake without being aware. They can open their eyes and breathe without assistance, but they cannot obey commands or make purposeful movements.

COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Many people in a vegetative state re- damaged patients who have a chance of AWAKE BUT NOT AWARE gain consciousness in the fi rst month af- recovery and those who face bleaker After brain injury that leads to coma, a patient’s ter a brain injury. After a month, how- prospects. At the same time, this re- progress may follow one of several paths (left). If ever, the patient is said to be in a persis- search may shine a new light on the na- the patient does not die or quickly recover, he or tent vegetative state (PVS) and the ture of consciousness itself. she will most likely transition to the vegetative probability of recovery diminishes as state. (In rare cases, a person may develop locked- more time passes. Every year in the U.S. A Difficult Diagnosis PATIENT PATHWAYS at least 14,000 new victims of acute in patients who recover from Fast brain damage remain vegetative one the vegetative state, the fi rst signs of con- Recovery month after their injury. In 1994 the sciousness are often minimal and appear Multi-Society Task Force on PVS (a gradually. The patient may start making team of 11 researchers from various in- deliberate, nonrefl exive movements but Vegetative stitutions) concluded that the chances of remain unable to express and communi- State recovery are close to zero if the patient cate thoughts and feelings. To classify Acute Brain Coma shows no signs of awareness one year af- these cases, doctors have created a clini- Injury ter a traumatic brain injury or three cal category called the minimally con- months (later revised to six) after brain scious state. Like the vegetative state, Locked-In Syndrome damage from lack of oxygen or other the minimally conscious state may be a causes. For these very long-standing cas- transient condition on the way to further es, the task force proposed the name recovery of consciousness, or it may be “permanent vegetative state.” chronic and sometimes permanent. An Brain The study of PVS was thrust into the important difference, though, is that pa- Death public spotlight in 2005 as politicians tients who have remained in the mini- debated the case of Terri Schiavo, a Flor- mally conscious state for years still have ida woman who had been in a vegetative a chance of recovery. In a much publi- she has his or her eyes open), repeatedly state since 1990. (Her parents and hus- cized case, Terry Wallis, an Arkansas fails to obey commands (such as “pinch band disagreed over whether she could man who had been in a minimally con- my hand” or “look downward”) and ever recover; the courts ultimately al- scious state since a road accident in 1984, makes only refl exive movements, the lowed doctors to remove Schiavo’s feed- started talking in 2003. Wallis also re- doctor will conclude that the patient is ing tube, and she died of dehydration 13 gained some ability to move his limbs, in a vegetative state. days later.) The controversy highlighted although he cannot walk and still needs In the early 1990s, however, studies the importance of developing more ef- around-the-clock care. led by Nancy Childs of the Healthcare fective ways to determine whether a pa- Making the distinction between the Rehabilitation Center in Austin, Tex., tient is permanently vegetative or in a vegetative and minimally conscious and Keith Andrews of the Royal Hospi- more hopeful condition. Scientists have states is challenging and requires re- tal for Neurodisability in London showed recently studied the use of brain-imag- peated examinations by well-trained that more than one third of the patients ing techniques to identify signs of aware- physicians who have experience with initially diagnosed as being in a vegeta- ness that could be hidden in an unre- such patients. Clinicians making the di- tive state in fact showed some signs of sponsive patient. If physicians had ac- agnosis of a vegetative state will base awareness when carefully examined. To cess to reliable methods of detecting their judgment on the absence of behav- obtain a more reliable diagnosis, doctors conscious awareness, they might be able ioral signs of consciousness. Put simply, need to employ standardized clinical to better distinguish between brain- if a patient looks awake (meaning he or tests that gauge a patient’s responses to a wide variety of auditory, visual and tac- The Vegetative State tile stimuli. Examples of such tests in- Overview/ clude the Coma Recovery Scale, devel- ) ■ Every year thousands of Americans who have suffered brain injuries oped by Joseph Giacino of the JFK John- transition from coma to a vegetative state. If patients remain in this state for son Rehabilitation Institute in Edison, more than a year, their chances of regaining consciousness are close to zero. N.J., and the Sensory Modality Assess- ■ Researchers are trying to develop brain-imaging techniques for diagnosing ment Rehabilitation Technique, created pages preceding the vegetative state, which would allow doctors to better determine which by Helen Gill-Thwaites, also at the Roy- brain-damaged patients have a chance of recovery. al Hospital for Neurodisability. The di- ■ Functional neuroimaging of vegetative patients has revealed new clues to the agnostic superiority of these specialized mechanisms of consciousness, but more research is needed before consciousness scales is beyond doubt, physicians can use this tool for diagnosis and prognosis. but they are much more time-consuming

than a routine neurological examination JEAN-FRANCOIS PODEVIN (

86 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. ther confi rm their diagnoses nor predict in syndrome, a complete paralysis of the body’s voluntary muscles.) The patient may then evolve their chances of recovery. My colleagues to the minimally conscious state and often to further recovery of consciousness or remain in the and I at the University of Liège in Bel- vegetative state permanently. Compared with people in other mental states (right), vegetative gium have demonstrated that minimally patients have a high level of wakefulness—unlike comatose individuals, they have sleep/wake conscious patients show an electrical cycles—but none of the awareness that characterizes normal conscious wakefulness. brain response called the P300 potential when hearing their own names, but not COMPARING STATES when hearing other fi rst names. Some Conscious patients in a chronic vegetative state, Wakefulness Minimally Recovery of Lucid however, demonstrated similar P300 re- Conscious Consciousness Dreaming Locked-In Syndrome sponses, so the technique appears to State Drowsiness have no diagnostic use. REM Sleep Light Sleep A Consciousness Region? Permanent Vegetative Death perhaps the most promising meth- State od for investigating the vegetative state Minimally Conscious State Deep Sleep is functional neuroimaging. Studies employing positron-emission tomography General Anesthesia (PET) have shown that the metabolic activity of the brain—as measured by its Coma Vegetative State consumption of glucose— decreases in Content of Consciousness: Awareness Consciousness: of Content Level of Consciousness: Wakefulness the vegetative state to less than half of normal values. These experiments were fi rst performed in the late 1980s by a or a simpler test such as the Glasgow neuronal axons that carry nerve impuls- group led by Fred Plum of Cornell and Coma Scale. es—have increased understanding of later repeatedly confi rmed by several But conscious awareness is a subjec- the mechanisms underlying recovery European groups, including ours. But tive experience that is inherently diffi cult from the vegetative state. A team led by some patients who recover from the veg- to measure in another being. Might even Nicholas Schiff of Cornell University, etative state do so without substantial the most careful assessment miss signs of for instance, recently used diffusion ten- changes in overall brain metabolism, as awareness in acutely brain-damaged, sor imaging to show the regrowth of reported by our laboratory in the late noncommunicative patients? Over the axons in the brain of Wallis, the patient 1990s. past decade investigators have struggled who emerged from a minimally con- Moreover, we observed that some to fi nd an objective test that could conscious state after 19 years of silence. fully conscious and healthy volunteers fi rm or dispute a clinical diagnosis of the Another widely used tool is the elec- have global brain metabolism values vegetative state. Structural imaging of troencephalogram (EEG), which mea- comparable to those seen in some pa- the brain— employing either magnetic sures the brain’s electrical activity. EEG tients in a vegetative state. And Schiff has resonance imaging (MRI) or x-ray com- results can reveal a patient’s state of reported that a few vegetative patients puted tomographic (CT) scanning—can wakefulness because electrical activity had a close to normal cortical metabo- help doctors visualize the extent of brain slows during nondreaming sleep. For lism. Hence, measuring overall levels of damage, but such techniques cannot de- patients in a coma, the instrument can energy consumption in the brain cannot tect signs of consciousness. Recent stud- confi rm the clinical diagnosis of brain indicate the presence of awareness. ies, though, indicate that examining MRI death (when the EEG becomes isoelec- Our group could, however, identify images of traumatic brain injuries may tric—that is, a fl at line). The EEG is not areas in the brain that appear to be par- help physicians predict whether a patient as reliable, though, for measuring ticularly important for the emergence of will emerge from the vegetative state. For changes in awareness. For patients in awareness. Comparing patients in a veg- example, patients with damage to certain the vegetative state, the device can nei- etative state with a large cohort of healthy regions, such as the brain stem and the corpus callosum (the band of tissue con- STEVEN LAUREYS leads the coma research group at the Cyclotron Research Center at the necting the cerebral hemispheres), appear University of Liège in Belgium and is head of clinics at the department of neurology at to have a poorer chance of recovery. Sart Tilman Liège University Hospital. He obtained his M.D. from the Free University of Furthermore, investigations using a Brussels (Flemish) in 1993 and his Ph.D. from the University of Liège in 2000. His work new technique called MRI diffusion is supported by the Belgian National Fund for Scientifi c Research, the European Com-

tensor imaging—which gauges the in- THE AUTHOR mission and the Mind Science Foundation. Laureys recently published The Boundaries

MELISSA THOMAS MELISSA tegrity of the brain’s white matter, the of Consciousness: Neurobiology and Neuropathology (Elsevier, 2006).

www.sciam.com SCIENTIFIC AMERICAN 87 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. volunteers, we found a signifi cant lack of tients. Just as for somatosensory stimu- power (such as infant cries and the pa- metabolic activity in the widespread net- lation, the activity was limited to the tient’s own name) induce a much more work of polymodal associative cortices lowest cortical centers (in this case, the widespread activation in minimally con- (located in the frontal and parietal lobes primary auditory cortices) while higher- scious patients than meaningless noise of the brain) that are involved in the cog- order polymodal areas failed to respond does. These results indicate that content nitive processing of sensory information and remained functionally disconnect- does matter when talking to a minimal- [see box below]. We showed that aware- ed. This level of cerebral processing is ly conscious patient. But to develop this ness is also related to the so-called cross considered insufficient for auditory technique into a diagnostic tool, we talk, or functional connectivity, within awareness. In minimally conscious pa- knew we had to show that complex au- this frontoparietal network and with cen- tients, however, auditory stimuli can ditory stimuli never activate large-scale ters deeper in the brain, notably the thal- trigger large-scale, higher-order corti- networks in vegetative patients. amus. In our vegetative patients, the long- cal activity normally not observed in the distance connections between one cortex vegetative state. Schiff was the fi rst to Tennis in the Brain and another, as well as the links between use functional MRI (fMRI) on patients this hypothesis faced its tough- the cortices and the thalamus, appeared in the minimally conscious state and est test last year when a team led by to be disrupted. Moreover, recovery from showed that their language networks Adrian Owen of the University of Cam- the vegetative state is paralleled by a were activated during the presentation bridge, in collaboration with Melanie functional restoration of the frontopari- of a personally meaningful story read by Boly from our group, studied a 23-year- etal network and its connections. a familiar voice. When played back- old woman who had suffered a traumat- Unfortunately, patients in a mini- ward, the story did not provoke such a ic injury to the front of her brain in a mally conscious state seem to show a response, whereas it did so in healthy traffi c accident. She remained comatose rather similar cerebral dysfunction. As control subjects. for more than a week and then evolved a result, PET measurements of brain Similarly, in 2004 our group report- to a vegetative state. She spontaneously metabolism cannot distinguish between ed that auditory stimuli with emotional opened her eyes but never responded the vegetative and minimally conscious states when the patient is at rest. The VIEWING THE DAMAGE technique did reveal differences, though, when analyzing changes in brain function induced by external stimuli such as pain and spoken words. We studied pain perception by electrically stimulating the hand (which was Parietal cortex experienced as painful by healthy control subjects) and using PET to gauge cerebral blood fl ow, which is another marker for neural activity. In both the vegetative patients and the controls, the results indicated activity in the brain stem, the thalamus and the primary somatosensory cortex, which receives sensory information from the peripheral nerves. In the vegetative patients, however, the rest of the brain failed to re- Thalamus spond. The small cortical region that Prefrontal cortex did show activity (the primary somatosensory cortex) was isolated and dis- connected from the rest of the brain, in Brains of vegetative patients particular from the networks believed show distinctive losses in to be critical for conscious pain percep- metabolic activity, as measured by the consumption of glucose in tion. (These results can reassure family the tissues. Regions in the prefrontal and members and caregivers that patients in parietal cortices (purple) are signifi cantly less a vegetative state do not perceive pain in active in vegetative patients. The dysfunction could be a result the same way that healthy people do.) of cortical damage or disruption of the links (red arrows) between the cortices and The PET studies found a similar pat- the thalamus. These connections appear to be crucial to conscious awareness.

tern when we talked to vegetative pa- CHEN Y. ALICE

88 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. SIGNS OF RECOVERY physicians to search for other signs of Functional neuroimaging of a 23-year-old brain-damaged woman (top images) awareness. During another examina- showed activity in a motor area when she was asked to imagine playing tennis and tion about six months after the study, activation of a spatial navigation network when she was asked to envision walking the patient was able to fi x her eyes on an through her house. Similar responses were seen in healthy control subjects object for a sustained period (more than (bottom images), indicating that the woman was beginning the transition from the fi ve seconds) and could track her own vegetative to the minimally conscious state. mirror image; both these signs herald a TENNIS IMAGERY SPATIAL NAVIGATION CORTICES IN transition to the minimally conscious NAVIGATION state. Currently the patient is still mini- Patient NETWORK Supplementary mally conscious, sometimes obeying motor area Premotor commands but not communicating. Parietal Given her young age and the cause and duration of the vegetative state, we Parahippocampal knew from the start that the patient’s chances of recovery were not zero but about one in fi ve. Thus, the results of the study should not be misinterpreted as Controls evidence that all patients in a chronic vegetative state may actually be conscious. In fact, we have not observed any similar signs of awareness in functional scans of more than 60 other vegetative patients studied at the University of Liège. The most likely explanation of the results is that our 23-year-old patient to any verbal or nonverbal commands. was asked to imagine walking through was already beginning the transition to Five months after the accident Owen the rooms of her house, the scans showed the minimally conscious state at the time and his colleagues studied the woman’s activation of the network involved in of the experiment. Indeed, a recent study brain using fMRI. During the scanning, spatial navigation—the premotor, pari- by Di Haibo of Zhejiang University in the researchers played recordings of etal and parahippocampal cortices. China and his colleagues confi rmed that spoken sentences—for example, “There Again, the response was indistinguish- the activation of higher-level brain re- was milk and sugar in his coffee”— able from that seen in the healthy sub- gions during fMRI tests can predict re- along with noise sequences that acousti- jects. Despite the clinical diagnosis that covery to the minimally conscious state. cally matched those sentences. The spo- the patient was in a vegetative state, she These findings underline the point ken sentences triggered activity in her understood the tasks and repeatedly that ascertaining consciousness is a tricky superior and middle temporal gyri, performed them and hence must have business. We have learned much from brain regions involved in understanding been conscious. new imaging techniques that measure speech and the meaning of words; the The first question raised by these neural activity in brain-damaged patients, same pattern was observed in healthy spectacular results was whether the pa- but more research is needed before scien- control subjects. These results possibly tient was misdiagnosed. Although re- tists can use functional neuroimaging to indicated conscious linguistic process- peated expert assessments had con- confi rm a diagnosis of the vegetative state ing in the vegetative woman, but not fi rmed that she was in the vegetative and to help in the prognosis and treat- necessarily. Studies of healthy patients state at the time of the study, the exam- ment of this devastating medical condi- have shown that such processing may inations revealed that her eyes briefl y tion. For the time being, doctors must also occur during sleep and even under fi x e d o n o b j e c t s . T h i s fi nding is some- continue to rely on thorough clinical ex- general anesthesia. times observed in the vegetative state, aminations when making their diffi cult VOL. 313; SEPTEMBER 8, 2006, REPRINTED WITH PERMISSION OF AAAS OF PERMISSION WITH REPRINTED 2006, 8, SEPTEMBER 313; VOL. To clarify whether the patient was but it is atypical and should prompt therapeutic decisions. consciously responding to language, the

SCIENCE, SCIENCE, investigators conducted a second study MORE TO EXPLORE in which they asked her to perform men- The Vegetative State: Medical Facts, Ethical and Legal Dilemmas. B. Jennett. Cambridge tal imagery tasks. When she was asked University Press, 2002. to imagine playing a game of tennis, the Science and Society: Death, Unconsciousness and the Brain. Steven Laureys in Nature fMRI scans showed activity in the sup- Reviews Neuroscience, Vol. 6, No. 11, pages 899–909; November 2005. plementary motor area of her brain, just Detecting Awareness in the Vegetative State. A. M. Owen, M. R. Coleman, M. Boly, M. H. Davis,

FROM A. M. OWEN ET AL. IN as it did in the control subjects. When she S. Laureys and J. D. Pickard in Science, Vol. 313, page 1402; September 8, 2006.

Using readily available equipment, you can carry out a home experiment that illustrates one of the weirdest effects in quantum mechanics

BY REMOVING INFORMATION about things that have happened, a quantum eraser seemingly inﬂ uences past events. In a fanciful example, a cat may have scampered around both sides of a tree at once if information about which way it went is later erased.

BY RACHEL HILLMER AND PAUL KWIAT

otoriously, the theory of quantum mechanics reveals Yet all this strangeness still seems remote from ordinary a fundamental weirdness in the way the world works. life. Quantum effects are most evident when tiny systems are Commonsense notions at the very heart of our every- involved, such as electrons held within the conﬁ nes of an atom. day perceptions of reality turn out to be violated: con- You might know in the abstract that quantum phenomena un- N tradictory alternatives can coexist, such as an object following derlie most modern technologies and that various quantum two different paths at the same time; objects do not simultane- oddities can be demonstrated in laboratories, but the only way ously have precise positions and velocities; and the properties to see them in the home is on science shows on television. of objects and events we observe can be subject to an ineradi- Right? Not quite. cable randomness that has nothing to do with the imperfection On pages 92 and 93, we will show you how to set up an of our tools or our eyesight. experiment that illustrates what is known as quantum erasure. Gone is the reliable world in which atoms and other par- This effect involves one of the oddest features of quantum me- ticles travel around like well-behaved billiard balls on the chanics—the ability to take actions that change our basic in- green baize of reality. Instead they behave (sometimes) like terpretation of what happened in past events. waves, becoming dispersed over a region and capable of criss- Before we explain what we mean by that and outline the

crossing to form interference patterns. experiment itself, we do have to emphasize one caveat in the COLLINS MATT

90 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. interest of truth in advertising. The light patterns that you will tent random and unpredictable, but as thousands of them ac- see if you conduct the experiment successfully can be account- cumulate, the spots build up into a defi nite, predictable pat- ed for by considering the light to be a classical wave, with no tern. When the conditions are right for the particles to behave quantum mechanics involved. So in that respect the experi- as waves, the result is an interference pattern—in this case a ment is a cheat and falls short of fully demonstrating the quan- series of fuzzy bars, called fringes, where most of the particles tum nature of the effect. land, with very few hitting the gaps between them. Nevertheless, the individual photons that make up the light The particles will generate the interference pattern only if wave are indeed doing the full quantum dance with all its each particle could have traveled through either of the two weirdness intact, although you could only truly prove that by slits, and there is no way of ascertaining which slit each one sending the photons through the apparatus and detecting them passed through. The two pathways are then said to be indis- one at a time. Such a procedure, unfortunately, remains be- tinguishable and each particle acts as if it actually traveled yond the average home experimenter. Still, by observing the through both slits. According to the modern understanding patterns in your experiment and by thinking about what they of quantum mechanics, interference occurs when indistin- mean in terms of the individual photons, you can get a fi rst- guishable alternatives are combined in this way. hand glimpse of the bizarre quantum world. When two or more alternatives coexist, the situation is called a superposition. Erwin Schrödinger highlighted the If you want to go straight to the home experiment, it is oddity of quantum superpositions in 1935, when he proposed detailed on the next two pages. The discussion that fol- his now infamous concept of a cat that is simultaneously alive lows here (and continues on page 94) delves into the sci- and dead, sealed inside a hermetic box where it cannot be ob- ence of quantum erasers in general. This explanation will served. When quantum interference happens, something in the help you understand what the do-it-yourself eraser dem- experiment is like a kind of Schrödinger’s cat. But instead of onstrates, but you might want to come back to it after being alive and dead at the same time, the cat may be walking seeing what that specifi c kind of eraser does. by a tree, passing on both sides of it simultaneously. Schrödinger’s cat ceases to be in a superposition as soon What a Quantum Eraser Erases as we look inside its box: we always see it to be either alive or one of the strange features of quantum mechan- dead, not both (although some interpretations of quantum ics is that the behavior that something exhibits can depend on mechanics have it that we become in a superposition of hav- what we try to fi nd out about it. Thus, an electron can behave ing seen a dead or a live cat). If a spotlight is shining near the like a particle or like a wave, depending on which experimen- tree, we see the quantum cat go one way or the other. Simi- tal setup we subject it to. For example, in some situations par- larly, we can add a measurement tool to watch each particle ticlelike behavior emerges if we ascertain the specifi c trajec- as it passes the slits. One could imagine having a light shining tory that an electron has followed and wavelike behavior on the slits so that as each particle comes through we can see transpires if we do not. a fl ash of light scatter from where the particle went. The fl ash A standard demonstration of this duality relies on what is makes the two alternative pathways distinguishable, which called a two-slit experiment (your do-it-yourself quantum destroys the superposition, and the particles arrive at the fi nal eraser is similar to this experiment in that it involves two path- screen not in a pattern of fringes but in one featureless blob. ways, but not two slits). A source emits particles, such as elec- Experiments analogous to this scenario have been conducted, trons, toward a screen that has two slits they can pass through. and, as predicted by quantum mechanics, no interference pat- The particles ultimately arrive at a second screen where each tern builds up. one produces a spot. Where each particle lands is to some ex- We need not actually “do the looking.” We do not have to What you will need for the experiment

■ A very dark room. Use a rubber band to keep the laser turned on. ■ Polarizing film. Plain gray, high-quality film (“experimental ■ A thin, straight piece of wire, such as from an unused twist grade”) gives the best results; avoid film tinted with a tie or a straightened staple. The thinner the better. color (see www.sciam.com/ontheweb for some places that ■ Some tinfoil and a pin to poke a hole in it. The light that goes sell film). You need to cut it into six squares, each about through the pinhole will expand outward, forming a narrow, two inches on a side. The box on page 94 describes what conical beam. The pinhole makes the patterns dimmer but polarizers do to photons. may improve the results if the room is dark enough. ■ A laser, such as a laser pointer. If yours emits polarized ■ Some stands to hold the laser and polarizers in place. These light, align its polarization at 45 degrees from the could be as low-tech as cereal boxes. vertical. If your laser is not polarized, include a polarizer ■ A screen to display the final patterns. The bare wall will do if at 45 degrees immediately after the laser at every step. it is plain enough; otherwise use a sheet of paper.

www.sciam.com SCIENTIFIC AMERICAN 91 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Quantum Erasing in the Home The steps presented here outline how to see quantum erasure in action. See www.sciam.com/ ontheweb for a fuller description and additional information, such as the basics of how waves interfere and produce fringes.

Straightened Polarizing staple ﬁ l m

SEEING INTERFERENCE 1 ■ Wrap the tinfoil around the business end of the laser and put wire is combining, or “interfering,” with light passing on the right- a pinhole in it to let through some of the light beam. hand side. If you hold a piece of paper just after the wire, you will see ■ Set up the laser so it shines on the screen from at least six feet a lobe of light on each side of the shadow of the wire. The lobes away. It should produce a circular spot of light on the screen. expand and largely overlap by the time they reach the screen. For each individual photon arriving at the screen in the overlap region, it ■ Position the wire vertically and centered in the light. is impossible to tell whether it went on the left or the right side of the WHAT HAPPENS: As shown, you should see an interference pattern wire, and the combination of the two ways it went causes the fringes. consisting of a row of fringes (bright and dark bands). The Although you are looking at trillions of photons, each of them is interference pattern arises because light passing on the left of the interfering only with itself.

Wire Interference Laser pattern INTERFERENCE seen is captured in ); this photograph. The size and other features of the patterns depicted in Tinfoil

the diagrams are exaggerated. setup laser

LABELING THE PATH 2 ■ Take two polarizers and rotate one of them so that their axes are ■ Position the labeler in the beam so that its join is right behind the perpendicular; you have done this correctly if when you overlap the wire. Attaching the wire to the labeler might be easiest. Wire and film temporarily, no light goes through the overlap region. labeler will not be moving for the rest of the experiment. We will say ■ Tape them together side by side with no gap or overlap. Do the that the left-hand polarizer produces vertically polarized light (V), taping along the top and bottom so the tape will not block the light. and the right-hand one horizontally polarized (H). It does not We will call this the path labeler. matter if we have these labels reversed. ); BRIAN MARANAN PINEDA ( PINEDA MARANAN BRIAN ); WHAT HAPPENS: Even though the light is again passing on both sides

of the wire, the fringes should be gone. If a photon )

reaches the screen by passing to the left of the cat wire, it arrives V-polarized; if to the right of the wire, H-polarized. Thus, the labeler has made and available the information about which way each photon went, which prevents the interference. Wire photographs of light patterns drawings of setups setups of drawings

Path labeler RACHEL HILLMER ( HILLMER RACHEL MATT COLLINS ( COLLINS MATT COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. SELECTING THE LEFT-PASSING PHOTONS SELECTING THE RIGHT-PASSING PHOTONS 3 ■ Position a third polarizer (the “analyzer”) between the labeler 4 ■ Put the analyzer in the H orientation. and the screen in the V orientation. WHAT HAPPENS: The H analyzer blocks the left-hand lobe of WHAT HAPPENS: The analyzer will block all the right-passing light and lets through only the right-hand lobe. If you could photons (which became H-polarized at the labeler) and will let measure intensities of light (or numbers of photons) at the through all the left-passing ones. The pattern will be nearly the screen, you would find that the light in step 2 was just the same as in the previous step—just dimmer and not extending sum of the light in steps 3 and 4. Notice that the fringes were quite so far on the right, because it is only the left lobe of light. missing from step 2 even though you were not ascertaining With the analyzer, you are accessing the information that the the polarization of the photons; it was enough that you could labeler made available: you know that all the photons hitting the have done so, as in steps 3 and 4. screen passed to the left of the wire.

Analyzer Analyzer (V polarizer) (H polarizer)

ERASING THE PATH INFORMATION THE ANTI-ERASER 5 ■ Rotate the polarizer 45 degrees clockwise from V, an 6 ■ Rotate the polarizer 45 degrees counterclockwise from V orientation we call diagonal (D). (“antidiagonal” or “A”). WHAT HAPPENS: The fringes reappear! Why? The polarizer is WHAT HAPPENS: Again there are fringes—everything said in erasing the information about which side each photon used. Now step 5 applies to an A-polarized eraser as well. But if you look each left-passing V photon has a 50 percent chance of getting very closely, you will see that the fringes are shifted slightly through it to the screen, as does each right-passing H photon. In in the two cases. The A fringes are bright where the D ones are both cases, the photons that get through become D-polarized, so dark, and vice versa. If you could add up the intensities, or there is no way to tell which way each photon went. Once again, numbers of photons, for the D and A erasers, the sum would each photon apparently goes both ways at once and interferes again be the shape from step 2, with no interference visible. with itself.

Eraser Eraser (D polarizer) (A polarizer)

BOTH ERASERS AT ONCE CONCLUSION 7 ■ Cut in half horizontally a D-oriented and an A-oriented polarizer. Think about what the photons ■ Join the top half of the D with the bottom half of the A. were doing in each of the steps. ■ Put the hybrid analyzer in place. ■ In some steps (3 and 4), each WHAT HAPPENS: D fringes appear in the top half of the light and photon went on one side or A fringes in the bottom half. The pattern looks a bit like misaligned the other of the wire (no teeth and makes clearer how the dark and bright fringes of interference), but in others 1, 5, 6 7 each eraser correspond. ( and ), they seemingly went on both sides at once (producing interference). Hybrid eraser ■ Our interpretation of (A and D what the photons did polarizers) at the wire depends on what they encountered later on in the setup—be it an analyzer or an eraser or nothing but the screen. ■ Steps 6 and 7 revealed that the “which way?” information can be erased in more than one way, to produce either the original interference pattern or the inverse of it.

COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. What polarizers do to photons olarizing film has an axis (in our diagrams we depict its Light can also be direction with lines on the film), and the film allows unpolarized, which P passage of light that is oscillating parallel to the axis. means the photons You can think of light as being like a wave on a rope held making up the light have between two people; the wave can make the rope move up random polarizations. and down or side to side or at any angle in between. The That is another case in angle of the oscillation is the polarization of the wave. which half the photons Polarizing film is like a screen of parallel bars that the will get through a polarizer, and, as always, those that do so rope passes through: it lets through waves polarized parallel become polarized parallel with the polarizer. to it unhindered, blocks perpendicular ones completely and You can see how polarizers work by putting two of them allows waves on other angles to get through with reduced together. As you rotate one of the polarizers, you can see amplitude. Most important, the wave (if any) that comes out through them clearly when their axes are aligned, barely at the other side of a polarizer is polarized parallel with the all when they are perpendicular and to some extent at other polarizer’s transmission axis. angles. Photons that make it through the first polarizer are The quantum description of what happens to light going polarized by it, and then their probability of getting through through a polarizing film sounds only slightly different: The the second one depends on the angle between their light is made up of individual particles called photons, and polarization and the second polarizer’s axis. like a wave, the photons can each have a direction of An interesting effect happens if two polarizers are oscillation. A photon will get through every time when it hits perpendicular and a third one is inserted between them on a polarizer with the transmission axis parallel to the an angle (45 degrees is best): adding the third polarizer photon’s polarization. A perpendicular polarizer blocks the allows some light to get through, even though you might photon every time. At a 45-degree angle, the photon has a 50 expect it to be an additional obstacle for the light. See if percent chance of getting through (the exact probability you can explain why that happens (the answer is at varies as the angle is varied). Most important, when a www.sciam.com/ontheweb). The do-it-yourself quantum photon does go through a polarizer, on the other side it will eraser also relies on a polarizer at 45 degrees changing be polarized parallel with the polarizer’s transmission axis. what the light does. detect the light fl ashes and ascertain which way each particle able, the information about their positions is erased. When went. It suffi ces that the information is available in the fl ashes that happens, the two paths the particles can follow are again and could have been observed in that way. indistinguishable and interference is restored. Now we fi nally get to the quantum eraser. The eraser is We have omitted one last tricky detail, but we will come something that can erase the information indicating which back to that. First, stop and think a bit more about what is hap- path each particle has followed, thereby restoring the indistin- pening in the erasing process we just described, because that guishability of the alternatives and restoring interference. is where the weirdness lies. When we detect the position where How might an eraser do that? Imagine that the “fl ash of one of the photons scattered, we learn which slit its corre- light” that scatters from each particle is a single photon. For sponding particle went through, which means the particle did the photon to reveal the “which path?” information of the go through one slit or the other, not both. If we instead detect particle, it must be possible (even if only in principle) to tell the photon’s momentum, however, we cannot know which slit which slit the photon came from. That means we must be able the particle went through. What is more, when we do many to measure the position of where each photon scattered accu- momentum measurements and see an interference pattern, we rately enough to tell the slits apart. Heisenberg’s uncertainty infer that in those cases the particles went through both slits principle, however, tells us that if we instead measure the mo- (interference would be impossible otherwise). mentum of each photon with great accuracy, then the photons’ In other words, the answer to the question, “Did the par- positions become less well defi ned. So if we pass the photons ticle go through one slit or both slits?” depends on what we do through a lens that makes their momentum information avail- with its corresponding photon long after the particle has gone through. It is almost as if our actions with the photons infl u- RACHEL HILLMER and PAUL KWIAT are both at the University of ence what has happened in past events. We can fi nd out which Illinois, where Hillmer is an undergraduate student in Kwiat’s slit the particle went through, or with our quantum eraser we laboratory. Kwiat is Bardeen Chair in Physics, and his research can delete that information from the universe. includes the phenomena of quantum interrogation and quan- Strangest of all, we can decide which measurement to make tum erasure, optical implementations of quantum information after the particle has passed through the slits—we can have the protocols and nonquantum studies of swing dancing. Hillmer is apparatus for both alternative measurements in place, with a THE AUTHORS

working on novel ways to encode quantum information in light. switch that we ﬂ ick one way or the other just before each pho- PINEDA MARANAN BRIAN

94 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. ton arrives. Physicists call this variation a delayed-choice experiment, an idea introduced by John A. Wheeler of the Uni- How a Quantum Eraser Works versity of Texas at Austin in 1978 that extends a scenario that How quantum particles behave can depend on what Niels Bohr and Albert Einstein used in their arguments about information about them can possibly be accessed. quantum mechanics and the nature of reality in 1935. A quantum eraser eliminates some information and At this point, some particularly clever readers will be wor- thereby restores the phenomenon of interference. The rying about a fundamental problem that seems to undermine eraser’s action is most easily understood by considering what we have just described: Why can’t we delay the choice a “double-slit” experiment (below). of our photon measurement until after we have seen if the particles form an interference pattern? We could, in fact, ar- CREATING QUANTUM INTERFERENCE range to do just that by having the fi nal screen not too far Particles sent through two from the slits and the photon detector much farther away. So Fringes slits generate bands what would happen if we saw the particles form fringes but (called fringes) on a Particle detector screen when then chose to do photon position measurements that should source large numbers arrive at prevent such fringes from forming? Wouldn’t we have created some regions (blue) and a paradox? Surely we would not expect the already registered very few arrive at other regions (white). This interference pattern to vanish! Similar reasoning suggests we interference pattern could use the delayed-choice effect to transmit messages in- arises only if each particle stantaneously over arbitrary distances (thereby circumvent- Particle could have traveled Slits trajectory through both slits to arrive ing the speed of light). at the screen (arrows). That tricky detail that we omitted earlier is what saves the day: to see the interference of the particles after applying the PREVENTING INTERFERENCE quantum eraser, we fi rst have to divide them into two groups The fringes do not appear if the particles interact and observe the groups separately. One group will display the Light with something that could original pattern of fringes; the other will display the inverse of source thereby be used to that pattern, with particles landing on what were originally ascertain each particle’s location at the slits. For the dark bands and avoiding the places where the bright fring- example, a photon of light es were. The two groups combined fi ll in all the gaps, hiding (yellow line)might scatter the interference. from the particle and reveal that it went The paradox is avoided because we need data from the pho- through the right-hand ton measurement to know which group each particle belongs slit. The photon need not to. Thus, we cannot observe the fringes until after we have done be detected—all that matters is that the the photon measurements, because only then do we know how Scattered “which slit?” information to split the particles into groups. In the home experiment, divid- photon in principle could be ing particles into groups is done for you automatically because determined if it were to be detected. one group gets blocked by a polarizing fi lter, and you can there- Photon fore see the interference pattern of the group that gets through detector with your own eyes. In the fi nal step you can see the interference ERASER RESTORES INTERFERENCE patterns of the two groups right next to each other. A quantum eraser erases From a practical standpoint, the inability to send messages the “which slit?” informa- faster than the speed of light and create a paradox is perhaps tion. If the particle disappointing, but physicists and logicians consider it to be a scatters a photon, a lens could make it impossible very good feature. to ascertain which slit the photon came from. In that case, the corresponding For more discussion about quantum erasers, go to particle apparently goes sa www.sciam.com/ontheweb, where you will fi nd: through both slits, as before, and fringes can be ■ A list of cutting-edge interference and quantum eraser observed. The strangest experiments carried out in recent years. feature of this quantum erasing is that the ■ A short discussion of what quantum erasers have to do behavior of the particle at with how the ordinary world we are familiar with the slits seemingly emerges from the weird underlying quantum reality. Lens depends on what the after ■ More information about delayed-choice experiments photon encounters and the impossibility of superluminal messages. the particle has passed through the slit(s). ■ A few other related experiments you can do at home. MATT COLLINS COLLINS MATT

DIRECTIONAL SOUND Psst ... Hey, You

You are walking down a quiet grocery store aisle when DIRECTIONAL-SOUND speaker suddenly a voice says: “Thirsty? Buy me.” You stop in emits a naturally focused beam of front of the soda display, but no one is next to you, and ultrasound, which produces Conventional audible tones in air within the loudspeaker shoppers a few feet away do not seem to hear a thing. sound At that moment, you are standing in a cylinder of beam. A conventional loudspeaker broadcasts audio in all directions. sound. Whereas a loudspeaker broadcasts sound in all directions, the way a lightbulb radiates light, a directional speaker shines a beam of waves akin to a spot- Amplifi er light. The beam consists of ultrasound waves, which humans cannot hear, but which can emit audible tones as they interact with air. By describing these interac- Ultrasound Voltage speaker tions mathematically, engineers can coax a beam to from amplifi er exude voice, music or any other sound. Undistorted Military and sonar researchers tried to harness the Back cover ultrasound phenomenon as far back as the 1960s but only managed to generate highly distorted audible signals. In 1998 Joseph Pompei, then at the Massachusetts Insti- Thin fi lm tute of Technology, published algorithms that cut the distortion to only a few percent. He then designed an amplifi er, electronics and speakers to produce ultra- Mesh grille sound “that is clean enough to generate clean audio,” Plastic Pompei says. He trademarked the technology Audio support Spotlight and started Holosonics, Inc., in Watertown, Mass., in 1999. Rival inventor Woody Norris markets a competing product called HyperSonic Sound from his American Technology Corporation in San Diego. Pompei’s speakers are installed in company lobbies, Cloth cover and above exhibits at the Boston Museum of Fine Arts and Walt Disney World’s Epcot Center, among other ) locations. Narrations inform visitors standing in front SPEAKER, only half an inch thick, is driven by an amplifi er of artifacts or video screens without fi lling the rooms that sends voltage to a metal-polymer fi lm. The fi lm photograph with noise. Department stores have tried the arrange- vibrates at 60,000 cycles per second or more, generating ultrasound waves. ment for retail displays, and automakers are experimenting with them so passengers can hear only their own music or movies. A speaker above a recliner in the living room would allow Dad to hear the television while

); HOLOSONICS, INC. ( INC. HOLOSONICS, ); Undistorted ultrasound Distorted ultrasound Audible wave other family members read on the couch in peace. Detractors say that in certain situations headphones can provide similar benefi ts, and note random problems, illustration such as unwanted refl ections off a car seat. But the primary obstacle to wider deployment is cost: systems run from $600 to $1,000 or more. If the price drops, consumers are more likely to consider buying the gear ... UNIFORM ULTRASOUND waves (left) produce density fl uctuations in air,

or encounter it while shopping. —Mark Fischetti ( INFOGRAPHICS 5W which distort the waves (center), causing them to emit an audible tone (right).

DID YOU KNOW... Dogs can hear up to 40,000 Hz or so, mice up to 90,000, and bats, bats, and 90,000, to up mice so, or Hz 40,000 to up hear can Dogs Hz. 20,000 and 20 between frequencies hear typically Humans quencies 40,000 from to 80,000 cycles a second, (Hz). or hertz beam off a building at the end of an alley or off a distant window window adistant off or alley an of end the at abuilding off beam a bounce could teams Police surface. smooth refl ahard, off ect porpoises and beluga whales up to 100,000 Hz or higher. or Hz 100,000 to up whales beluga and porpoises behind audible directed sound. The ultrasound speakers emit fre- emit speakers ultrasound The sound. behind audible directed from the sound inside a warehouse to fl ush out suspects, who would run away away run would who fl to awarehouse suspects, inside out ush BATS NOT DOGS: Certain animals can detect the ultrasound noise noise ultrasound the detect can animals Certain DOGS: NOT BATS BOUNCED: Ultrasound waves remain in a tight column when they in remain when a column tight waves Ultrasound BOUNCED: — and right into the offi cers’ waiting arms. arms. offi waiting the into cers’ right and sound Audible COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. speakers when they are in the “spotlight.” All other space remains quiet. remains space other All “spotlight.” the in are they when speakers INDIVIDUALS Send [email protected] about murder? Committing the ultimate crime? I do. All the time. I time. the All Ido. crime? ultimate the Committing murder? about the skull may be distorting vibrations that reach the cochlea. reach that vibrations the skull may be distorting sounds; “hearing” report whom of some volunteers, of skulls the Hz. But researchers have applied ultrasound up to 200,000 Hz to to Hz 200,000 to up ultrasound applied have researchers But Hz. get paid to think about it. I’m a best-selling crime writer ... ” ... writer crime abest-selling I’m it. about think to paid get ries. When a patron, milling about, unwittingly stepped near apost- near stepped unwittingly about, milling apatron, When ries. over here. Don’t turn around. Do you hear me? Do you ever think think ever me? you Do hear you Do around. turn Don’t here. over er about the show, a voice from nowhere whispered, “Hey, you “Hey, whispered, nowhere from avoice show, the about er jor bookstores to market its new new its market to bookstores jor MURDERER: BONUS: Middle ear bones limit human hearing to below 20,000 20,000 below to hearing human limit bones ear Middle BONUS: at the New York Public Library can hear audio from overhead overhead from audio hear can Library Public York New the at

Last November, Court TV used Audio Spotlight in main Spotlight Audio used TV Court November, Last Have atopicidea? Murder by the Book SCIENTIFIC AMERICAN SCIENTIFIC television se- television

97 — REVIEWS

The Interplay of Art and Science TWO WAYS OF VIEWING THE WORLD MEET IN THE VISUAL REALM BY BETTYANN HOLTZMANN KEVLES

LEONARDO DA VINCI: EXPERIENCE, EXPERIMENT AND DESIGN by Martin Kemp Princeton University Press, 2006 ($60) SEEN/UNSEEN: ART, SCIENCE, AND INTUITION FROM LEONARDO TO THE HUBBLE TELESCOPE by Martin Kemp Oxford University Press, 2006 ($45)

Published almost simultaneously, these very different books present a double view of Martin Kemp’s original and often brilliant approach to the connection between science and art. Leonardo focuses on a single genius; Seen/Unseen pulls back the lens to investigate the nature of creativity thematically, using proﬁ les of extraordinary artists/scientists over a span of 500 years. Kemp is intrigued by visual works that combine the skills of artist and scientist, often, but not always, in the same person; he calls himself a “historian of the visual.” Kemp’s ideas may be familiar to readers of Nature who have followed his THROUGH HIS STUDIES and drawings, Leonardo became aware of the joining and subsequent science and culture column. Currently divergence of the nerves from the eye in the optic synapse. (Circa 1500; pen, ink and red chalk.) professor of art history at the University of Oxford, he has produced a veritable jottings and notes, providing insight ment of the possible discovery of a miss- library of books and articles about into the artist’s philosophy, mastery of ing Leonardo masterpiece, The Battle of Leonardo. Away from his desk he orga- geometry, and compassion. Anghiari—walled off but retrievable in nizes exhibitions, most recently one at Leonardo da Vinci has fascinated the Palazzo Vecchio—the reporter gave the Victoria and Albert Museum in generations of artists, art critics and weight to the claim by noting that Kemp London; the beautiful, coffee table–size novelists. But few have been as absorbed strongly supported the investigation. Leonardo da Vinci: Experience, Ex- as Kemp, who has walked in the mas- Leonardo is based on a “theme periment and Design accompanied this ter’s footsteps through Italy and France. sheet,” a miraculously recovered piece exhibition. Kemp’s text dovetails seam- Kemp’s name is synonymous with Leon- of folded paper (reproduced across two lessly with the often life-size reproduc- ardo scholarship. In a February New pages in the book) on which the artist

tions of Leonardo’s drawings, sketches, York Times account of the announce- began sketching and writing in 1490. II ELIZABETH MAJESTY QUEEN HER 2007 COLLECTION,© ROYAL THE

Kemp explains the value of paper half a the future. Ahead of his time in a visual George Stubbs saw “an animate nature millennium ago—which meant that ev- way, however, Leonardo observed na- in continual fl ux,” and the other foot in ery inch was used—and he interprets the ture closely and drew “notes” from the Victorian era, in which “Natural Se- apparent chaos of sketches as a record what he saw, from an autopsied heart to lection itself represents the emergence of of Leonardo’s “laboratory for think- fl oodwaters and debris, believing that the ‘fi ttest’ theory into a very specifi c ing.” He tells us that there are about all knowledge had to be confi rmed by environment” that was familiar with 6,000 surviving pages of the artist’s observing natural phenomena. He must the earlier worldview and thus receptive notes and sketches. have had remarkable eyesight to capture to his much more radical theory. Yet even for the keenly interested, motion as he did, as well as a rare abil- In this same chapter Kemp leaps to Leonardo’s notes, written in a mirror ity to see mentally in three dimensions, the 20th century and Richard Dawkins’s script, are diffi cult to decipher, and their one of the themes of Kemp’s second “selfi sh genes” and James Lovelock’s dispersion on this theme sheet may not book, Seen/Unseen. Gaia hypothesis and wonders about a be deliberate. A fi nely drawn portrait of Although Leonardo is not the heart self-regulating biosphere. Kemp likes to an old man merges with a delicate pair of Seen/Unseen, he is the touchstone for move back and forth across the centu- of trees, and the rest of the drawings, much of Kemp’s thinking and one of the ries, explaining how scientific ideas notes and designs bear no apparent rela- “nodal” fi gures, men (mostly) who as morph and mutate in different historical tion to these sketches or to one another. both artists and scientists are the focus eras—a sort of dialogue across time that The jottings, combined with references of each of its chapters. In “Wholes and he relishes. In “Growth and Form,” he to the well-known paintings and models Parts,” for instance, Kemp identifies pays homage to D’Arcy Wentworth for machines, reveal a man of his time at Charles Darwin as a man with one foot Thompson’s classic 1917 book On home with geometry, though probably in the dynamic view of nature of the ear- Growth and Form, explaining Thomp- not mathematically prepared for the sci- ly 19th century, where men such as his son’s references to Albrecht Durer’s entifi c revolution that was a century in grandfather Erasmus Darwin and artist method of proportional transformation

100 SCIENTIFIC AMERICAN MAY 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. and connecting this approach to form to contemporary planetary scientists who to promise a non-human precision.” the visual mathematics of fractals, cha- capture images of planets, such as Ve- And it would be a mistake to put the tool os theory and machine-made images. nus, while taking the liberty of coloring makers in the privileged position of de- Sometime in the 19th century, ac- them with familiar, Earth-like tones. ciding how their tools should be used. cording to Kemp, the invention of the Why, he asks, bother to map Venus, Kemp’s thoughts on art are self-con- camera and its offspring technologies, where we will never go? Because, he re- sciously different in kind from other re- cinema and radiography, liberated art- plies, we are explorers, and explorers cent efforts to link art and science that ists from having to reproduce nature want to map their discoveries. either concentrate on individual artists with conventional perspective, freeing Kemp acknowledges that his recon- and scientists or focus on a particular sci- them to move in different directions. sideration of scientists and artists across entifi c breakthrough. None approaches Scientists, in contrast, embraced pho- the centuries is sometimes diffi cult, and the world as Kemp does. Seen/Unseen is tography and happily wed cameras to he asks his readers to bear with him. a glimpse into his own thought process their microscopes and telescopes. Some may be left sensing that this is not the way his Leonardo is a glimpse into Kemp is still in the process of explor- his fi nal effort to explain visual culture, the mind of an almost mythical genius. ing the explosion of visual culture in the which is not the study of art or of science Kemp offers us a way of considering how late 20th and early 21st centuries. We but an inquiry into human creativity artists and scientists have intuited visual humans are, as he illustrates in various that can be seen. In his own ongoing in- truths in the past, reminds us that the ways throughout this rich discussion, vi- quiry, Kemp is ambitious and enlighten- past and the present are connected, and sual creatures. We need to see in order to ing but occasionally diffi cult to follow. warns us against the potential tyranny understand. Biologists, he reminds us, He questions the validity, interpretation of the newest digitized images that, needed an electron microscope to see and ultimately the use of computerized, though often beautiful and beguiling, DNA and, from that glimpse, to con- machine-made images extracted, for ex- are still man-made and not infallible. struct a model. He lauds the contribu- ample, from PET and fMRI brain scans. tion of physicist Richard Feynman, who Suddenly he fears the technology he has Bettyann Holtzmann Kevles, author of devised the “Feynman diagrams” (which been describing. “The more technologi- Naked to the Bone: Medical Imaging in are not included in the text), because he cal the image looks, the more it exudes ... the Twentieth Century, teaches a course thought, and taught, with images. Kemp authority,” he writes, but a computer is, on the history of science, invention and applauds the visual reconstructions of nonetheless, a man-made tool that “seems the visual arts at Yale University. www.sciam.com SCIENTIFIC AMERICAN 101 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. ANTIGRAVITY

Ale’s Well with the World A YEASTY DISCUSSION OF THE SCIENCE OF MALT AND HOPS BY STEVE MIRSKY

Free beer. And thus I found my way to converted to two molecules of ethanol conspicuously absent from beer making. a lecture in late February at the New and two of carbon dioxide (along with But fetid feet pale when an ale be- York Academy of Sciences by renowned some energy), Bamforth noted that “if comes stale: tinted bottles or talented beer maven Charlie Bamforth. A man the sugars are from grapes, you produce chemists must be employed to keep beer for whom the word “avuncular” was a fi ne beverage called wine. If the sugars from becoming “sunstruck.” Light con- coined, the British Bamforth has three are from grain, it’s a superior beverage verts certain beer bitter acids to the decades of brewing expertise under his called beer.” dreaded 3-methyl-2-butene-1-thiol belt. Over his belt is what he insists is “a Bamforth decried beer’s sometimes (MBT) compound, a close relative of the sausage belly, not a beer belly.” dicey image: “Beer is perceived as a bad- malodorous thiols produced by Mephitis “Beer is the basis of modern static civ- boy drink, and beer has been too often mephitis, the striped skunk. In his book ilization,” began Bamforth, Anheuser- marketed in all sorts of strange ways— Beer: Tap into the Art and Science of Busch Endowed Professor of Brewing fl atulent horses [an infamous ad during Brewing (Oxford University Press, 2003), Science at the University of California, the 2005 Super Bowl, overshadowed by Bamforth notes that some people can Davis. “Because before beer was discov- the even more infamous Janet Jackson smell MBT at levels as low as 0.4 part per ered, people used to wander around and “wardrobe malfunction”] and men be- trillion. “These poor people,” he writes, follow goats from place to place. And having badly. And wine is perceived “would have been able to detect a tenth of then they realized that this grain [bar- somehow as being superior and speaking a gram of MBT distributed throughout ley] could be grown and sprouted and to a higher quality of life. Really, it’s un- the balloon of the airship Graf Zeppelin made into a bread and crumbled and fair. Beer is more consistent; it’s pro- II.” Eau, the humanity. converted into a liquid which gave a duced with more devotion and care; it’s So, does the brewski master have a nice, warm, cozy feeling. So gone were at least as healthy.” Plus, human feet are favorite beer? “It depends on where I the days that they followed goats around. am,” Bamforth explains. “If I’m in an They stayed put while the grain grew old pub with a ceiling about my height and while the beer was brewed. And and there’s a roaring log fi re, cask ale they made villages out of their tents. from England is sublime. I wouldn’t have And those villages became towns, and an American-style lager. If I’m at a Sac- those towns became cities. And so here ramento River Cats AAA baseball game we are in New York, thanks to beer.” and it’s 100 degrees outside, I could kill Another syllogism ended his address: for a Bud. I’m not going to drink a Guin- “He who drinks beer sleeps well. He ness. So it’s horses for courses.” As long who sleeps well cannot sin. He who as those horses aren’t fl atulent. does not sin goes to heaven. The logic is Answering the charge that beer is impeccable.” empty calories, Bamforth points out that In between came numerous nuggets beer is actually rich in B vitamins, except about what Bamforth insists is “the for thiamine. “A famous doctor came up world’s favorite beverage,” some of to me,” he remembers, “and said, ‘Is it which were at the expense of another true that if we could just boost the levels popular adult drink. While discussing of thiamine, beer would be a meal in it- the simple chemical equation of fermen- self?’ I answered, ‘Even I wouldn’t say

tation, by which a molecule of sugar is that. You need a few pretzels.’” VERONSKY FRANK BY PHOTOGRAPH COLLINS; MATT BY ILLUSTRATION

Where is the universe expanding to? Q —A. Kenny, Canisbay, Scotland Astrophysicist Alexander Kashlinsky of the NASA Goddard ordinates (latitude and longitude) on the surface of the balloon, Space Flight Center offers this explanation: and it is the fabric of the balloon that is actually expanding. The evolution of the universe is described by the physics of In the framework of general relativity with only four di- general relativity discovered by Albert Einstein during the ear- mensions, the question posed here does not have an answer, ly 20th century. In general relativity, space and time are merged because it implies some other coordinate grid outside spacetime. into one continuum and the universe can be represented as a Because spacetime is linked to matter, there is no outside to the four-dimensional spacetime grid. When viewed from this per- surface of the balloon—it is all the spacetime that is available. spective, the universe’s expansion does not push it into new territory—rather the spacetime grid itself is expanding. In prerelativistic, Newtonian What is “junk” DNA, and what physics (which describes celestial is it worth? —A. Khajavinia, Isfahan, Iran bodies as moving according to the Q laws discovered by Isaac Newton), Wojciech Makalowski, a Pennsylvania State University space and time are absolute, with biology professor and genomics researcher, replies: time no more than a parameter in All animals have a large excess of junk DNA—genetic ma- the equations of motion. Gravity, terial that does not code for the proteins used to build bodies meanwhile, is seen as a force of at- and catalyze chemical reactions within cells. In our genetic traction between massive bodies, blueprint, for instance, only about 2 percent of DNA actually but its source is a mystery. codes for proteins. The physics of general relativ- In 1972 the late geneticist Susumu Ohno coined the term ity is conceptually distinct—even if its equations of motion can “junk DNA” to describe all noncoding DNA sections, most of be reduced to Newtonian equations in many practical cases. which consist of repeated segments scattered randomly through- In general relativity, the properties of the spacetime grid are out the genome. Typically sections of junk DNA come about uniquely specifi ed (via gravity) by the bodies inhabiting it. through transposition, or movement of sections of genetic ma- Gravity curves the spacetime continuum, and general relativ- terial to different positions in the genome. As a result, most of ity thus describes gravitational interactions as manifestations these regions contain multiple copies of so-called transpo- of that curvature. Objects under gravity’s infl uence “fall” sons—sequences that literally copy or cut themselves out of one from less curved parts of spacetime to more curved parts. part of the genome and reinsert themselves somewhere else. According to Einstein’s general relativity equations, the In the early 1990s interest in junk DNA, and especially in space time containing matter cannot remain stationary and repetitive elements, began to grow; many biologists now re- must either expand or contract. Galaxies, then, are not strictly gard such repetitions as genomic treasures. It appears that moving away from one another but rather are attached to the these transposable sequences increase the ability of a species fi xed grid on the expanding fabric of spacetime, thereby giving to evolve by serving as hot spots for genetic recombination the impression of moving away from one another. As an anal- and by providing important signals for regulating gene ex- ogy, imagine placing dots on the surface of a balloon, then pression. As such, repetitive elements are hardly “junk” but infl ating it. The distances between the dots—which represent rather are integral components of our genomes. galaxies—will increase, so if you live in one of these dots, you will interpret the others as receding from you. In reality the For a complete text of these and other answers from

dots remain in the same positions, with respect to the two co- scientists in diverse ﬁ elds, visit www.sciam.com/askexpert COLLINS MATT