Scientific American-June 2007

Home , Bootherium, Equus conversidens, Equus occidentalis, Third Voice

BREAKING NETWORK LOGJAMS • TRULY 3-D IMAGES

Bring Back America’s Prehistoric Beasts

JUNE 2007 WWW.SCIAM.COM Did this molecule start

life

The Mysteries of Anesthesia

When It Pays to Be Irrational

How Particles Shape the Cosmos

BIOLOGY 46 IA Simpler Origin for Life BY ROBERT SHAPIRO 46 Molecular Energy-driven networks of small molecules may be more likely seed of life? ﬁ rst steps for life than the commonly held idea of the sudden emergence of large self-replicating molecules such as RNA.

MEDICINE 54 ILifting the Fog around Anesthesia BY BEVERLEY A. ORSER Learning why current anesthetics are so potent and sometimes dangerous will lead to a new generation of safer targeted drugs.

PARTICLE COSMOLOGY 62 IWhen Fields Collide BY DAVID KAISER The history of particle cosmology shows that science can beneﬁ t from wrenching changes.

ECOSYSTEMS 70 IRestoring America’s Big, Wild Animals BY C. JOSH DONLAN Pleistocene rewilding—a proposal to bring back animals that disappeared from North America 13,000 years ago—offers an optimistic agenda for 21st-century conservation.

INFORMATION TECHNOLOGY 78 IBreaking Network Logjams BY MICHELLE EFFROS, RALF KOETTER AND MURIEL MÉDARD Network coding could dramatically enhance the efﬁ ciency of communications networks.

INNOVATIONS 86 ISeeing Triple BY STUART F. BROWN Anticipated for decades, machines are ﬁ nally displaying objects in three true dimensions.

GAME THEORY 90 IThe Traveler’s Dilemma BY KAUSHIK BASU People playing this simple game consistently reject the rational choice. In fact, by acting illogically, they end up reaping a larger reward—an outcome that demands a new kind of formal reasoning.

www.sciam.com SCIENTIFIC AMERICAN 5 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more.. SCIENTIFIC AMERICAN Volume 296 Number 6 departments 8 SA Perspectives Serengeti in the Dakotas. 10 How to Contact Us 10 On the Web 20 104 12 Letters 18 50, 100 & 150 Years Ago 20 News Scan 40 Insights New genetic studies, Eric Vilain says, should ■ Planting a better biodiesel. force a rethinking about mixed-sex babies and ■ Fingerprinting individual atoms. gender identity. ■ Meet the Memjet. ■ Less lethal turbines for dams. 96 Working Knowledge ■ Solid-state laser blaster? Optical character recognition fi nds the write type. ■ New fi sh food: psoriasis patients. 98 Reviews ■ Computers can fi nally pass Go. Books on dirt, fl owers, uncertainty, ■ Data Points: Large Hadron Collider. perishable artifacts.

Eric Vilain, 40 University of California, Los Angeles columns 39 Skeptic BY MICHAEL SHERMER The inverse square law trumps the law of attraction. 43 Sustainable Developments BY JEFFREY D. SACHS As global warming tightens the availability of water, prepare for forced mass migrations. 102 Anti Gravity BY STEVE MIRSKY Coffee cup full of beans. 104 Ask the Experts How do itches come about? Why is the sun in the middle of the solar system?

Cover image and illustration of molecule on preceding page by Ken Eward, BioGrafx; cover photograph of cheetah by Shani; photograph at left by Brad Swonetz, Redux.

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.

Some people love science for the crazy ideas, Whether or not cheetahs ever chase pronghorn the ones that transport you beyond the everyday across the continent again, the rewilding concept has grind: black holes, alien life, anything with the word drawn attention to the fact that the loss of biodiver- “quantum.” Others prefer the not-crazy ideas, the sity is not just a problem for the rain forest; it affects practical solutions: zippier computers, 100-mpg less exotic locations, too. The demise of large ani- cars, cures for cureless diseases. mals has thrown entire ecosystems out of balance. So what do you make of an idea like Pleistocene Even if humans decided now to leave these ecosys- rewilding? It manages to be both crazy and not crazy tems alone, they are too far gone to recover on their at the same time. As the article by C. Josh Donlan own. The prairie would revert not to its Pleistocene beginning on page 70 describes, glory but to a scraggly weedland. a team of biologists has pro- Instead of merely bemoaning nature’s plight, the posed a decades-long project to proponents of rewilding are doing something about restock North America with it. The reintroduction of wolves in Yellowstone a de- large mammal species like those cade ago brought the population of moose and elk that roamed the continent be- under control, with a cascade of benefi ts for vegeta- fore humans crossed the Bering tion, birds and beavers. Other projects have reintro- Strait—species such as camels, duced tortoises, bison and falcons into their old lions and elephants (the nearest haunts. Wildlife-stocked private ranches let you go thing to mammoths). The un- on safari within half an hour’s drive of the Alamo. dertaking would culminate More broadly, biologists are also working to restore in a vast national park—1,000 fi sheries, forests and wetlands. square miles or more—stretch- Most of these efforts are too piecemeal to nurse ing across the Great Plains. The whole ecosystems back to health. They need a broad- NOT QUITE a woolly mammoth, plains states are depopulating er framework, and a grand rewilding project can pro- but close. anyway, whereas Africa and vide it. Visionary schemes have a checkered history. Asia are fi lling up. So the project The ones that work combine big ideas with baby would transplant wildlife from where it gets in the steps. They need both: for lack of attention to detail, way to where it would have plenty of room. a grand project can easily fail; equally so, for lack of To be sure, Midwesterners might not see it that an overarching plan, incremental steps can seem like way. Elephant families running free under big skies spitting in the sea. Recent history offers many ex- sounds romantic—unless you have to dodge them on amples of enterprises that, deprived of a real goal, lost your morning commute. Lion cubs are so very cute— their way. One that comes to mind is the human space except when they wander into your backyard. Farm- program. After the moon landings, NASA became ers worry about rampaging rogues, cattle ranchers little more than a delivery van service. about novel diseases. Proponents have addressed A thriving ecosystem can underpin economic some of the concerns but clearly have a lot more prosperity and enhance our quality of life, but it won’t Peter Arnold, Inc.

work to do. happen on its own. We have to make it happen. DONA

THE EDITORS [email protected]

8 SCIENTIFIC AMERICAN JUNE 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more.. 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 Gene Makes Small Dogs Small and mailing address, West/Northwest and cite the article 818-827-7138 fax: 818-827-7139 In big news for small dogs and the issue in Detroit which it appeared. Karen Teegarden & Associates everywhere, researchers have Letters may be edited 248-642-1773 found a tie that binds the small fax: 248-642-6138 for length and clarity. Midwest breeds, from Chihuahua to We regret that we cannot Derr Media Group Pomeranian to Pekingese: they all 847-615-1921 answer all correspondence. share the same version of a gene For general inquiries: fax: 847-735-1457 Southeast and Southwest for a growth hormone called Scientifi c American Publicitas North America, Inc. 415 Madison Ave. 972-386-6186 insulinlike growth factor 1 (IGF1). New York, NY 10017-1111 fax: 972-233-9819 Direct Response 212-451-8200 Special Aditions Advertising, LLC fax: 212-755-1976 914-461-3269 or fax: 914-461-3433 SPECIAL REPORT Australia [email protected] IMR Pty Ltd. +612-8850-2220 The Poisoning of Our Pets SUBSCRIPTIONS fax: +612-8850-0454 Scientists and government agencies home in on the For new subscriptions, Belgium Publicitas Media S.A. cause of more than 100 pet deaths from tainted food. renewals, gifts, payments, +32-(0)2-639-8420 and changes of address: fax: +32-(0)2-639-8430 U.S. and Canada Canada Derr Media Group FACT OR FICTION? 800-333-1199 847-615-1921 Outside North America fax: 847-735-1457 Waking Sleepwalkers May Kill Them 515-248-7684 France and Switzerland On the contrary, rousing a sleepwalker could save or PEM-PEMA +33-1-46-37-2117 the person’s life. www.sciam.com fax: +33-1-47-38-6329 or Germany Scientifi c American Publicitas Germany GmbH +49-211-862-092-0 BLOG Box 3187 fax: +49-211-862-092-21 Harlan, IA 51537 Hong Kong When the Sky Falls, Where Will NASA Be? Hutton Media Limited REPRINTS +852-2528-9135 Everybody duck and cover. In March, NASA shrugged fax: +852-2528-9281 and told Congress that it neither has the funding nor the To order reprints of articles: India Reprint Department Convergence Media resources to meet its goal of identifying 90 percent Scientifi c American +91-22-2414-4808 fax: +91-22-2414-5594 of near-Earth asteroids 150 yards or more in diameter 415 Madison Ave. Japan by the year 2020. New York, NY 10017-1111 Pacifi c Business, Inc. 212-451-8877 +813-3661-6138 fax: +813-3661-6139 fax: 212-355-0408 Korea ASK THE EXPERTS [email protected] Biscom, Inc. Why is there an ozone hole in the atmosphere, +822-739-7840 PERMISSIONS fax: +822-732-3662 whereas there is too much ozone at ground level? Middle East Why doesn’t ground level ozone rise to fill the hole? For permission to copy or reuse Peter Smith Media & Marketing material from SA: +44-140-484-1321 Ross J. Salawitch, a senior research scientist at the NASA www.sciam.com/permissions fax: +44-140-484-1320 Jet Propulsion Laboratory in Pasadena, Calif., explains. The Netherlands or Insight Publicitas BV 212-451-8546 for procedures +31-35-539-5111 or fax: +31-35-531-0572 NEW: VIDEO Permissions Department Scandinavia and Finland M&M International Media AB SciAm.com now features the latest video science news Scientifi c American +46-8-24-5401 every weekday. 415 Madison Ave. fax: +46-8-24-5402 New York, NY 10017-1111 U.K. The Powers Turner Group Please allow three to six weeks +44-207-592-8331 Subscribe to Scientifi c American Digital

for processing. fax: +44-207-630-9922 Visit www.sciamdigital.com FITZMAURICE DEANNE ©

® Letters EDITORS@ SCIAM.COM Established 1845 EDITOR IN CHIEF: John Rennie UNSURPRISINGLY, Frank Keppler and Thomas Röckmann’s EXECUTIVE EDITOR: Mariette DiChristina MANAGING EDITOR: Ricki L. Rusting February article on the discovery of methane emissions in CHIEF NEWS EDITOR: Philip M. Yam plants elicited strong reactions. Although some correspon- SPECIAL PROJECTS EDITOR: Gary Stix SENIOR EDITOR: Michelle Press dents seemed eager to jump to the exact global warming– EDITORS: Mark Alpert, Steven Ashley, denying conclusions Keppler and Röckmann had cautioned Graham P. Collins, Mark Fischetti, Steve Mirsky, George Musser, Christine Soares against, others expressed a more thoughtful attitude as to CONTRIBUTING EDITORS: W. Wayt Gibbs, how this discovery might affect our understanding of global Marguerite Holloway, Michael Shermer, climate change. Readers were also drawn to the more cosmic Sarah Simpson implications of Christopher J. Conselice’s overview of how sci- EDITORIAL DIRECTOR, ONLINE: Kate Wong entists believe dark energy shapes the universe. Many were NEWS EDITOR, ONLINE: Lisa Stein ASSOCIATE EDITORS, ONLINE: David Biello, unsatisfied with our inability to detect such a force from any- Christopher Mims thing but its effects and offered their own theories as to what NEWS REPORTER, ONLINE: JR Minkel dark energy might tangibly represent or how the universe could function without it. ART DIRECTOR: Edward Bell SENIOR ASSOCIATE ART DIRECTOR: Jana Brenning ASSOCIATE ART DIRECTOR: Mark Clemens ASSISTANT ART DIRECTOR: Johnny Johnson DENTAL DELIMITATION minutes after we make this switch (ef- PHOTOGRAPHY EDITOR: Emily Harrison PRODUCTION EDITOR: Richard Hunt “License to Work” [News Scan], by Rod- fectively reversing the principle of fair ger Doyle, suggests that the reason the use forged in the early days of video- COPY DIRECTOR: Maria-Christina Keller COPY CHIEF: Molly K. Frances number of dentists in the U.S. has not tape)? Or that we could be charged ev- COPY AND RESEARCH: Daniel C. Schlenoff, grown substantially compared with ery time we view their programming Michael Battaglia, Smitha Alampur, Michelle Wright, John Matson, Aaron Shattuck other professions is restrictive licensing with a premium for every different de- practices. Doyle has the cart before the vice we view it on? EDITORIAL ADMINISTRATOR: Jacob Lasky SENIOR SECRETARY: Maya Harty horse. The license to practice dentistry Thomas Phelan

ASSOCIATE PUBLISHER, PRODUCTION: William Sherman is obtained after the completion of edu- Hightstown, N.J. MANUFACTURING MANAGER: Janet Cermak cational requirements and is typically ADVERTISING PRODUCTION MANAGER: Carl Cherebin passed by most dentists, although it ANTONOFF REPLIES: Digital TV potentially PREPRESS AND QUALITY MANAGER: Silvia De Santis PRODUCTION MANAGER: Christina Hippeli sometimes calls for more than one at- affords more control by broadcasters and CUSTOM PUBLISHING MANAGER: Madelyn Keyes-Milch tempt. The restriction on numbers is at studios over their products and can restrict ASSOCIATE PUBLISHER, CIRCULATION: Simon Aronin the beginning of the road, where the what consumers are allowed to do with pro- CIRCULATION DIRECTOR: Christian Dorbandt RENEWALS MANAGER: Karen Singer educational system has not changed the grams once they enter their homes. But to FULFILLMENT AND DISTRIBUTION MANAGER: Rosa Davis number of dentists it is capable of train- do this subject justice would have required VICE PRESIDENT AND PUBLISHER: Bruce Brandfon ing on average since the 1970s. This rean in-depth discussion of the various as- SALES DEVELOPMENT MANAGER: David Tirpack striction is not caused by the licensing pects of DRM (many of which may or may not SALES REPRESENTATIVES: Jeffrey Crennan, Stephen Dudley, Stan Schmidt board but by the cutting of direct and be implemented). Recording rights, home

ASSOCIATE PUBLISHER, STRATEGIC PLANNING: indirect federal and state support for networking, down conversion of digital sig- Laura Salant dental education (number of schools, nals to analog devices and legal fair-use is- PROMOTION MANAGER: Diane Schube RESEARCH MANAGER: Aida Dadurian class size, faculty numbers, student sues would all need to be examined in detail, PROMOTION DESIGN MANAGER: Nancy Mongelli loans, and so on). requiring a full article unto itself. GENERAL MANAGER: Michael Florek BUSINESS MANAGER: Marie Maher Keith J. Lemmerman MANAGER, ADVERTISING ACCOUNTING Graduate Periodontics ILLUSORY IQ AND COORDINATION: Constance Holmes University of Kentucky In “Unsettled Scores: Has the Black- DIRECTOR, SPECIAL PROJECTS: Barth David Schwartz College of Dentistry White IQ Gap Narrowed?” [News Scan], MANAGING DIRECTOR AND VICE PRESIDENT, ONLINE: Marina Krakovsky cites so-called he- Mina C. Lux DIRECTOR, WEB TECHNOLOGIES, ONLINE: Vincent Ma DIGITAL DOMINION reditarians who believe that lower IQ SALES REPRESENTATIVE, ONLINE: Gary Bronson In “Digital TV at Last?” Michael An- test scores among blacks are the result DIRECTOR, ANCILLARY PRODUCTS: Diane McGarvey tonoff writes about consumer issues re- of inherent genetic factors. The presup- PERMISSIONS MANAGER: Linda Hertz lated to switching from analog broad- position that what is being measured by CHAIRMAN EMERITUS: John J. Hanley casts to digital with no mention of digi- such tests is an innate intelligence that CHAIRMAN: Brian Napack PRESIDENT AND CHIEF EXECUTIVE OFFICER: tal-rights management (DRM). How can be described as biological and uni- Gretchen G. Teichgraeber many know that there is a serious move ﬁ ed is one that has never been proved. VICE PRESIDENT AND MANAGING DIRECTOR, INTERNATIONAL: Dean Sanderson afoot to allow content owners to limit Psychologist Alfred Binet, who created VICE PRESIDENT: Frances Newburg time shifting of their programs to 90 the ﬁ rst standard for testing, insisted

himself that his tests merely identifi ed preindustrial times with those of today. Shermer cites the wrong Newtonian where below-average performers lacked Ruminants are listed as a major source concept. The correct equation would be skills in the classroom and could best be of current emissions but are not included P = FV (Power equals Force times Veloc- tutored to bring them up to normal per- in the preindustrial chart. Did as many ity). Assuming the Power (the rate at formance standards. The American as 70 million bison really produce that which the rider converts internal energy twist on IQ testing, which included the much less methane than today’s cattle? into mechanical energy to propel the bi- conception of a biological and unitary Ed Miller cycle) remains the same with or without intelligence, was based on assumptions Falls City, Ore. the extra 10 pounds, the increase in that all character traits are Mendelian Force required to “lift” that weight must in nature. KEPPLER REPLIES: Although wildlife cer- result in a decrease in the Velocity of the Aaron Method tainly produced methane in preindustrial bicycle and rider. The increase in Force Indianapolis times, this output was just a minor fraction is not equal to the full 10 pounds, be- of the 233 million metric tons of yearly glob- cause the rider is on an inclined plane, RUMINANT RUMINATION al methane emissions. According to esti- but this is a separate issue. In “Methane, Plants and Climate mates made by environmental scientist Su- Allen Zimmerman Change,” by Frank Keppler and Thomas san Subak and her colleagues in a 1994 ar- Ohio State University Röckmann, two graphs compare sourc- ticle for Chemosphere, the total production es of methane in the atmosphere during of methane by wild ruminants in that period ERRATA In “The Universe’s Invisible Hand,” was no more than 10 million by Christopher J. Conselice, a graph in the metric tons a year—a fi gure that box “Dark Energy Takes Charge” represents takes into account the North the fraction of galaxies taking a spiral shape American bison population with a pink line and those settling into an (which Subak estimates to have elliptical shape with a yellow line. The labels comprised 60 million animals) on both of these should be swapped. and the natural ruminants of Af- “Making Silicon Lase,” by Bahram Jalali, rica and other continents. An es- incorrectly states that a data-transfer rate timated 1.4 billion head of cattle of 10 gigabits a second would constitute a populate the world today—far 10,000-fold improvement over a rate of one more ruminants than existed in megabyte a second. Because one byte is preindustrial times. equal to eight bits, the improvement would Furthermore, modern cattle be 1,250-fold. are bred for productivity, which probably leads them to emit CLARIFICATION “Methane Flatline,” by Da- more methane than their wild vid Biello [News Scan], reports that meth- relatives did. Estimates put ane traps 23 times more heat per molecule their methane production at than carbon dioxide. This fi gure refers to the 115 million metric tons a year. average amount of infrared radiation absorbed by a methane molecule during its 10- PHYS ED PHYSICS year life span in the atmosphere as com- In “Eat, Drink and Be Merry” pared with the amount absorbed by a CO2 [Skeptic], Michael Shermer molecule over a century. It is a separate fi g- suggests that the mechanics ure from the Intergovernmental Panel on of a cyclist taking longer to Climate Change’s estimates of the global- climb a 5 percent grade with warming potential (GWP) of both substanc- 10 pounds of extra weight es (methane’s GWP is 23 times that of CO2 can be described in terms of per kilogram). Newton’s law of F = MA (Force equals Mass times Send letters to [email protected] or to Acceleration): “The Force Scientifi c American, 415 Madison Ave., GROWING CONCERN: Estimates of the relative contributions needed to turn the pedals New York, NY 10017. Letters become to global methane levels from natural sources now need to take emissions from living plants into consideration. equals Acceleration times the property of Scientifi c American

that Mass on the saddle.” and may be edited for space and clarity. BROWN KENN

50, 100 & 150 YearsSCIENTIFIC AMERICAN Ago FROM

Lunar Idea ■ Milk Agenda ■ Balloon Dismay

JUNE 1957 JUNE 1907 have come to realize this fact, and in up- MOON DUST—“The possibility of actu- GOT BOTTLED MILK?—“Milk, the article to-date institutions the old-fashioned ally bringing back some of the moon’s of food most susceptible to contamina- fi nger alphabet is now unknown. Every material is a scientifi c bonanza so allur- tion, is served in bottles which are used child is taught to speak in the natural ing that ingenious schemes have been again and again. This is vitally wrong. way by means of the vocal organs. The proposed to accomplish it, even without The solution is the abandonment of it, four or fi ve years of the primary course landing on the moon. We might, for ex- and the substitution of the single-service are devoted almost exclusively to the ample, send a pair of rockets, one trail- paper package of milk. These are now acquirement of language and numbers ing the other closely by means of a hom- just beginning to be available. One of the [see illustration].” ing device. The fi rst rocket would drop fi rst paper bottles to be placed on the a small atomic bomb on the moon. Since market is a plain paper cylinder, made of JUNE 1857 the moon has no atmosphere and com- new spruce-wood paper in clean, sani- ELECTRIC MOTOR—“We admit that no paratively little gravity, the bomb cloud tary surroundings. After the bottom is explosion can take place in the batteries would rise very high. The sec- of Prof. Vergnes’ electro-mag- ond rocket could dive into the netic engine, as with a steam cloud, collect some of the spray boiler; but his engine is neither as and emerge from its dive by simple nor compact as a steam means of an auxiliary jet. Of engine, taking the latter with all course, such a maneuver would its appurtenances. His large elec- require a miracle of electronic tro-magnetic engine in the Crys- guidance.—Krafft A. Ehricke tal Palace, which is not claimed and George Gamow” to exert more than 10 horse power (and which, we believe, from JAWS—“The plain fact is that the mere inspection of its operation, ferociousness of the shark is too is less than fi ve), with 128 cups well documented by long ex- of electric battery, occupies more perience to be dismissed as a leg- space—engine and battery—than end. Jacques-Yves Cousteau, the many steam engines working up greatest living expert on modern to 20 horse power.” deep-sea diving, has reported harrowing encounters with LOOSE GAS—“The miniature toy sharks. Off the Cape Verde Is- balloons introduced with such lands of Portugal, Cousteau and success into Paris during the his fellow diver Frédéric Dumas holidays of last winter [made found themselves closely pressed from ox intestine membranes], by an eight-foot gray shark, later we suggested as pleasing and joined by two fi ve-footers and beautiful toys to delight Young several blue sharks. The pack re- America. Quite recently they sisted all attempts to drive them TEACHING a deaf pupil how to count, 1907 have come into pretty general off: they would shy away only to circulation, and may be seen in come back almost at once. The two men put in, the bottle is dipped in hot paraf- various store windows in our city. A used every method in the diver’s book. fin, the same paraffin the housewife number of our boys who have purchased The sharks were not scared off by any of pours over her jellies to keep out the air, such balloons have been rather aston- these tricks, or by the shark ‘repellent’ moisture, and dust.” ished to fi nd them daily growing beauti- (copper acetate) strapped to the divers’ fully less in size, and prematurely old in legs or by a blow Cousteau gave one ap- FORCING SPEECH—“The loss of the sense wrinkles, and at last ceasing to be bal- proaching shark on the nose with his of hearing should not necessarily mean loons at all. The cause of this is the per- heavy undersea camera. The two back- deprivation of the power of speech also. colation of the infl ating gas (hydrogen) tracking divers got out just in time.” It is only within recent years that we through the pores of the balloon.”

ENERGY

I Green Gold in a Shrub ENTREPRENEURS TARGET THE JATROPHA PLANT AS THE NEXT BIG BIOFUEL BY REBECCA RENNER

woody shrub with big oily seeds could gue, will be better for the global environ- be the ideal source for biofuel. For ment and economy than conventional bio- A hundreds of years, Africans in places fuel crops grown in temperate climates. such as Tanzania and Mali have used Jatro- Ethanol from corn or sugarcane and pha curcas (jatropha) as a living fence. Now biodiesel from canola, soy or palm oil have biodiesel entrepreneurs in tropical zones in become major players in renewable energy. Africa and India are buying up land, start- In principle, biofuels do not increase the ing plantations and looking forward to amount of carbon dioxide in the air, because making fuel from the seeds, which, they ar- as the plants grow they trap the CO2 that is released when the biofuels are burned. Still, biofuels face a great deal of criti- cism. Food commodities such as corn, canola and soy all yield oil, but they are expensive, require intensive agricultural practices and threaten food supplies. Jatropha seems to offer the beneﬁ ts of biofuels without the pitfalls. The plants favor hot, dry conditions and hence are unlikely to threaten rain forests. There is no trade-off between food and fuel either, because the oil is poisonous. John Mathews, a professor of strategic management at Macquarie University in Austra- lia, notes that many tropical developing countries have huge swaths of degraded and semiarid land that can be utilized for fuel crops. The cost of labor there is cheap, too. Biofuels made from plants such as jatropha, he argues, “represent the best bet for a last- ditch effort to industrialize the poor south and end poverty.” He advocates large-scale JATROPHA SEEDLINGS are planted in Zambia for U.K. biodiesel ﬁ rm D1 Oils—part of an plantings to aid energy independence in ex- increasing effort to harvest the shrub, which favors hot, dry climates, as a source of biofuel.

panding economies such as China and India OILSD1 PLC

and to boost exports in the less developed crops, then apply fi rst-rate farming practic- news countries of Africa. es: prune branches, apply manure and pro- SCAN Mathews’s vision may be coming true. vide water. U.K. biodiesel company D1 Oils has planted But the realization that successful large- 150,000 hectares of jatropha in Swaziland, scale operations have to func- Zambia and South Africa, as well as in India, tion like well-run farms raises where it is part of a joint venture. The fi rm the issue of competition with plans to double its crop sizes this year. Dutch food crops for water and land, biodiesel equipment manufacturer BioKing says agronomist Raymond is developing plantings in Senegal, and the Jongschaap of Wageningen government of China has embarked on a University in the Netherlands. massive project. “People aren’t making much Jongschaap is spearheading jatropha oil right now, because everyone one of the research projects wants seeds for planting,” says Reinhard looking for different types of Henning, a German technology transfer jatropha with the goal of NUTTY GOODNESS: Jatropha fruit may contain consultant and expert in jatropha. matching plants to growing up to 40 percent oil, depending on the variety. In addition to establishing plantations, conditions and maximizing oil jatropha boosters are starting to identify, se- yields. He has the most faith in lect and propagate the best varieties for small-scale efforts based on hedges or inter- biodiesel production. Henning has found cropping jatropha with other plants— NEED TO KNOW: Brazilian jatropha seeds that contain 40 a method used in projects in Kenya and BIOFUEL BLUES percent oil—about the same as canola and Madagascar, where jatropha is planted The jatropha plant has excited more than twice the 18 percent contained in alongside vanilla. some energy experts because it soybeans. Indonesia has a dwarf variety Henning agrees that it is smart for jatro- does not have the same negative impacts as conventional biofuel that is especially easy to harvest. pha growers to start small. Biodiesel cannot crops. Touted as a renewable Finding the variety best suited to par- compete with current petroleum prices, resource and a means to promote ticular growing conditions is crucial, ex- which are relatively low, so jatropha would energy independence, today’s plains D1 Oils agronomy director Henk be better suited for local projects that im- biofuel plants are expensive, Joos, because right now not much hard sci- prove rural livelihoods and basic energy require intensive farming and threaten food supplies. The recent entifi c information exists about jatropha— services. These small projects have already U.S. push for ethanol has already just lots of stories. “We know that this plant started to build a framework of familiarity contributed to rising corn prices. is environmentally elastic and drought-tol- and expertise—in parts of Tanzania, kids Biofuel plantations can also harm erant. But the aura that this is a wonder crop learn about jatropha in school. Then, as fuel the environment. In Europe a that you can plant in the desert and harvest prices increase, jatropha cultivation can go biofuel quota backfi red when it gold” is a dangerous notion that threatens to a larger scale. The wild shrub could then increased demand for palm oil social and economic sustainability, Joos become a “sustainable cash crop,” Joos be- from Southeast Asia; as a result, says, adding that jatropha needs to be man- lieves, and a fuel for the future. farmers carved new plantations out of dwindling rain forest and aged like any other crop. He notes that at D1 released millions of tons of Oils plantations, farmers plant in land that Rebecca Renner is an environmental carbon dioxide when they cut into

D1 OILSD1 PLC is as good as possible without replacing food writer based in Williamsport, Pa. carbon-rich peat soils. PHYSICS

I Atomic Fingerprinting MICROSCOPE DISCERNS AN ATOM’S CHEMICAL IDENTITY BY LUIS MIGUEL ARIZA

eciding whether a substance is, say, team of physicists has developed a method steel, brick, wood or plastic is easy— of atomic “ﬁ ngerprinting” that can deter- D but not on the atomic scale, which mine the chemical identity of individual at- lacks information about such everyday char- oms on a surface mixed with all of them. acteristics. Using an atomic-force micro- “Until now, there was not any technique scope (AFM), however, an international that would allow us to identify atom by atom

In contrast, the fi ngerprinting AFM technique works for conductor and in- sulator alike. Like a phonograph, the AFM employs an ultrafine needle mounted on a fl exible cantilever. As the needle gets dragged across a surface, it jogs up and down as it encounters atoms on that surface. This oscillation actually occurs because of the attractive forces associated with the onset of chemical bonding between the silicon in the tip and the atoms on the surface. The Japanese-Spanish team showed SEEING CHEMICALS: Physicists used an atomic-force microscope to identify the that the oscillation frequency depends surface composition of an alloy made of silicon on the atom’s chemical nature. It thus (red), tin (blue) and lead (green) atoms. enabled the researchers to identify different atomic species even if they exist and see them at the same time,” says in equivalent abundance, Peréz says, Rubén Peréz of the Autonomous Univer- “like distinguishing a tree in a noisy, sity in Madrid. Using their AFM ap- fuzzy forest.” proach, Óscar Custance and his collab- In previous work, Custance and his orators at Osaka University, along with team had demonstrated that they could Peréz and his colleagues and others, use the AFM to move tin atoms strong- could discern tin, silicon and lead, which ly attached to a germanium surface, are all chemically similar. The resulting writing the letters “Sn” (tin’s chemical image of the atoms resembles a granu- symbol). Combining the method with lated painting, where the “grains”—the atomic fi ngerprinting opens exciting individual atoms—are distinguishable possibilities for the AFM—researchers in false color. might be able to “visualize reactions The ability to identify and manipu- with atomic resolution,” Custance re- late atoms fi rst gained prominence in marks. And, he adds, as microelec- 1989, when IBM scientists spelled out tronics shrink into the nanoscale their company logo with xenon atoms. realm—2,000 of today’s transistors Back then, the physicists relied on a can fi t across the width of a human scanning tunneling microscope (STM), hair—then “just by arranging a few at-

which detects atoms by a slight ﬂ ow of oms in predeﬁ ned patterns, it could be University Osaka electrons between the microscope tip possible to enhance the performance and an atom. The STM, however, can of the devices.” identify only atoms of materials that

conduct electricity. Luis Miguel Ariza is based in Madrid. CUSTANCE ÓSCAR PRINTERS

I “Memjet” Momentum INK-JET PRINTING AT RAMJET SPEED BY CHARLES Q. CHOI

tealthily, over more than a de- yielded an ink-jet printer, dubbed the cade, a new kind of printer has Memjet, that can print color photo- S been under development in Aus- graphs up to 30 times faster than any tralia. The original vision was to create other printer. a printer small enough to ﬁ t inside a Ink-jet printers, which dominate digital camera. Instead the research has the desktop market, work by spraying

drops of ink from nozzles onto a sheet. Conventional ink-jets shuffle their nozzles back and forth across paper like a shuttle would between threads in a loom. The Memjet, from Silverbrook Re- search in Sydney, has fi xed rows of nozzles stretching from one edge of the page to the other that all fi re simultaneously. Besides cutting down on moving parts, noise and vibration, the approach enables the device to fi nish one high-quality color page a second, roughly fi ve times faster than consumer ink-jets, and to generate 30 four-by- six-inch color photographs a minute. “Every other printer seems glacial by comparison,” says Steve Hoffenberg, an analyst for Lyra Research in New- tonville, Mass., which tracks the imaging industry. Like many ink-jet devices, the Memjet (a name derived from micro- electromechanical systems, or MEMS) relies on heating ink to a boil, which generates steam explosions that fi re ink droplets out through nozzles. But whereas conventional ink-jets heat ink by using electronics located in the nozzle walls, the Memjet suspends its heaters in the fl uid. This setup cools the heaters off, which in turn allows the Memjet to pack nozzles closer together, amount- ing to 320 nozzles per square millime- ter. That density is 17 times more than what leading printer makers can do now. Closer nozzle packing means greater efficiency in the amount of material used and lower costs. The aim is for Memjet printers to cost

NOVEL NOZZLES: The Memjet uses ﬁ xed rows of nozzles that span the entire width of a page, enabling rapid printing. COURTESY OF SILVERBROOK RESEARCH SILVERBROOK OF COURTESY

news roughly as much as other consumer-grade have plenty of speed,” he adds. The company SCAN ink-jets. also asserts that its device should be no more The Memjet fi res droplets one-fi fth the prone to clogging than existing printers. size of those sprayed in most ink-jet printers. Silverbrook unveiled its new technology Such picoliter volumes enable it to print in March 21 at the Global Ink Jet Printing sharp detail at 1,600 dots per inch, better Conference in Prague. It should be available than the eye can see and comparable with in photograph and label printers later this leading printers. It would not be hard to de- year and in home and offi ce printers in 2008. sign conventional ink-jets to print with drops And as for a printer that can be embedded as small as the Memjet’s, but they would inside a camera, the company says that, too, then have fi ve times as many droplets to fi re still remains a possibility. and thus “would be fi ve times slower,” explains lead researcher Kia Silverbrook. “We Charles Q. Choi is a frequent contributor. ECOLOGY

I A Good Turn A FISH-FRIENDLY HYDROELECTRIC TURBINE GETS NEW LIFE BY MADELINE BODIN

early two years ago Alden Research Laboratory in Holden, Mass., hauled N the scale model of a promising hydropower turbine out of its massive test fl ume and set it in a dim corner of the company’s hydraulics laboratory building. As an innovation developed in the 1990s, the device proved quite promising in reducing one of hydropower’s drawbacks: the turbines kill creatures that pass through them. The novel design enabled at least 98 percent of fi sh A TURBINE to survive. But orphaned by federal budget DUSTED OFF cuts, it has sat gathering dust. Now a new The scale model of the fi sh- push has begun to retool the turbine for po- friendly Alden/Concepts NREC tential commercial use. turbine had been mothballed until Conventional turbines, which resemble this past February, when the blades of an electric fan, kill as many as Canadian energy fi rm Brookfi eld Power began installing the device 40 percent of the fi sh that are swept through at its School Street hydropower them. Working from U.S. Department of facility in Cohoes, N.Y. There the Energy funds fi rst granted in 1994, Alden turbine will be located in what was Lab teamed up with Concepts NREC in once a spillway that allowed fi sh White River Junction, Vt., to develop a fi sh- safe passage around the dam. friendly turbine. The design features three It probably will not generate a lot of power, but nonetheless rotor blades wrapped around a conical hub Brookfi eld considers the to create a kind of helix. A rotating case cov- 60 installation a win-win situation ers the rotor blades, so that only a fraction 50 for the fi rm and the fi sh. of their edges are exposed. The turbine has 40 The hardest part, says Dave no gaps between the blades, fewer blades FISH-FRIENDLY TURBINE relies on Culligan, manager of licensing for and a slower spin rate. All these features sheathed rotor blades that wrap 30 Brookfi eld’s U.S. Development lower the chance of a fi sh being injured by around the hub to form a kind of helix Group, was getting state and 20 moving parts. Moreover, the fl ow of water (top). The design keeps water fl owing federal wildlife agencies “over the smoothly and at a constant rate 10 philosophical hurdle of letting through the turbine is smooth, creating less through most of the turbine (bottom). second) Water per (feet Velocity 0 fi s h g o t h r o u g h a t u r b i n e . ” potentially harmful shearing force. INC. LABORATORY, RESEARCH ALDEN

The design is particularly friendly will be tough. “As a rule, what’s good to eel and sturgeon— long, thin swim- for engineering is not good for fi sh, and mers whose populations are declining vice versa,” says Ned Taft, president of and for whom traditional hydropower Alden Lab. The solution, Alden engi- turbines are decidedly deadly. Accord- neers believe, is to increase the amount ing to initial tests, all the eel and stur- of water fl owing through the turbine. geon that pass through the new turbine They recently fi gured out a way to dou- can be expected to survive. ble the volume of the spiral pipe that But about three years ago federal feeds water into the blades while in- funding for the turbine was cut com- creasing the turbine’s diameter by only pletely. Research stalled, and eventu- 2 percent. Besides reshaping and rean- ally the scale model was put in storage. gling the blades to handle the increased The Electric Power Research Insti- fl ow, Concepts NREC plans to broad- tute (EPRI), based in Palo Alto, Calif., en the leading edges of the blade to a has taken up the fi sh-friendly turbine’s width that is closer to the length of the cause. It has raised about $300,000 of typical fi sh moving through the tur- the $500,000 needed to make the turbine—a size correlation that seems to bine commercially feasible and to re- boost survival. place conventional models as hydro- The design is on the brink of com- electric plants are renovated. The fi rst mercial viability, Dixon believes. “But step will be to increase the new tur- we’re struggling to fi nd the resources to bine’s power output so that it can com- make it happen.” Supporters say that pete with existing turbines, comments more is at stake than the success of this Doug Dixon, senior project manager at particular turbine design. “A lot of EPRI. The original design generated people don’t see hydro as green power,” half the power of a commercial turbine Taft states. “This turbine could change of the same size. “The more effi cient it that.” is, the more attractive it will be to industry,” Dixon says. Madeline Bodin is based Boosting the power output, though, in Andover, Vt.

Beam Weapons DEFENSE SYSTEMS

I Get Real SOLID-STATE LASERS NEAR BATTLEFIELD DEPLOYMENT BY STEVEN ASHLEY

he ray gun, long a mainstay of sci- hundreds to thousands of kilowatts— ence-fiction tales, may actually offer several advantages over conven- T enter the American war-fi ghting tional projectile weapons, according to arsenal in a few years. Engineers at sev- Mark Neice, director of the Depart- eral defense fi rms have conducted suc- ment of Defense’s High-Energy Laser cessful tests of key prototype compo- Joint Technology Offi ce in Albuquer- nents of truck-size “laser cannon” sys- que, N.M. “They could provide ultra- tems capable of firing a beam from precise, speed-of-light strike capabili- aircraft, naval ships or armored vehi- ties that could leave little or no collat- cles to zap targets many kilometers eral damage,” he says. away—even through intervening dust Although overly optimistic predic- or fog. tions in the past have led skeptics to High-power lasers—measured in quip that “lasers are the weapons of the

future, and they always will be,” this time watts, the so-called entry-level average pow- news beam weapons fi nally look real. “Industry er benchmark for military laser applications. SCAN is just on the cusp of delivering practical di- Other key performance targets, Neice says, rected-energy weapons for offensive and are an operational run time of 300 seconds defensive military purposes,” Neice states. (suffi cient for multiple laser shots), an electri- In the past year or so, with funding from cal-to-optical system energy conversion ef- the U.S. Air Force, Army and Navy, research- fi ciency of 17 percent or more, and, crucially, ers at Northrop Grumman, Textron, Ray- adequate “beam quality” (essentially, focus- theon and Lawrence Livermore National ing) to ensure enough photons reach the Laboratory have achieved marked progress target to heat its exterior sufficiently to on the solid-state bulk laser, which runs directly on electricity. Coupled to a ground vehicle’s electric generator, fuel cell or battery bank, a solid-state laser with an average power capacity greater than 100 kilowatts would feature a nearly “infi nite magazine” of low-cost shots with which to blast incoming Kuwait University mortars, artillery shells, rockets and Faculty of Science missiles in fl ight from fi ve to eight kilo- Kuwait meters away. Such a system could also blind electro-optical and infrared bat- The Department of Biological Sciences at the Faculty of Science of tlefield sensors and enable troops to Kuwait University invites applications for appointment at the rank of neutralize mines and improvised explo- Full Professor, starting September 2007 in the following area: sive devices from a safe distance. The key to this high-energy device Molecular Biology (Human genetics with special interest in is the gain medium, the material that cytogenetics and/or molecular genetics) amplifi es the laser photons. In laser diodes used in DVD players and other Required Qualifi cations: consumer electronics, semiconductor • Ph.D. degree in the area of specialization from a reputable university layers amplify light after a jump start • The applicant’s bachelor’s degree GPA should be 3 out of 4 or equivalent • Research experience and publications in refereed international journals from an electrical charge. In the solid- • Full command of teaching in English state bulk laser, the gain medium takes • University teaching experience in the specifi ed fi eld the form of a few-centimeter square (or rectangular) slab, explains Jackie Benefi ts include attractive tax-free salary according to rank and teaching Gish, Northrop Grumman Space Tech- experience (Professor’s monthly salary varies from 2950 to 3192 KD., 1 KD. = $3.40), annual air tickets for the faculty member and his/her family nology’s director of directed-energy (spouse and up to three children under the age of 20), a one time settling-in technology and products. The slabs allowance, housing allowance, free national health medical care, paid mid- consist of a tough ceramic material, term holidays and summer vacations, and end-of-contract gratuity. The such as yttrium-aluminum-garnet University also offers an excellent academic environment and fi nancial doped with the rare-earth element support for research projects. neodymium. But instead of an electri- To apply send by express mail/courier service or e-mail, within six weeks of cal priming, stacks of laser diodes op- the date of announcement, a completed application form, updated curriculum tically pump, or excite, the gain me- vitae (including mailing address, phone and fax numbers, e-mail address, dium. In general, the larger the slab, academic qualifi cations, teaching and research experience, and a list of the greater the power output. publications in professional journals), three copies of Ph.D., Master and Bachelor certifi cates and transcripts (an English translation of all documents in Each research team has different other languages should be enclosed), a copy of your passport, three letters of ways to link several slabs together to recommendation, and names and addresses of three persons well-acquainted create laser “chains” that produce with your academic and professional work to the following address: high power levels in the tens of kilowatts, says John Boness, vice president The Dean, For inquiries: Faculty of Science Fax: +965 4836127 of applied technology at Textron Sys- Kuwait University, P.O. Box 5969, E-mail: [email protected] tems. In the coming year, engineers ex- Safat, 13060, Kuwait pect to gang these chains coherently in series or in parallel to achieve 100 kilo- www.sciam.com SCIENTIFIC AMERICAN 29 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more..

news destroy it, detonate it or send it off course. SCAN Assuming that slab lasers meet these goals, their making it into actual combat depends on their successful integration into a functioning weapons system that is compact LETHAL enough to ﬁ t in a vehicle, Boness explains. REACTIONS Besides a renewable electrical power source of 1,000 kilowatts or more, a solid-state The U.S. military has already developed one kind of mighty bulk laser weapon needs a thermal chiller to laser—a directed-energy device keep the slabs from heating up enough to powered by chemical reactions. distort the beam. Much more potent than their solid- It would also require a beam director to state cousins, these megawatt- place the photons on target—probably a class devices are called chemical oxygen iodine lasers (COILs). But large, mobile mirror equipped with adaptive they are large and can run only as or deformable optics to compensate for at- long as their stores of reactants mospheric distortion, which would be de- hold out. Nevertheless, defense tected by a low-power “sensing” laser beam. contractors are now preparing to Finally, aiming such a system would rely on install COILs on aircraft. A Boeing 747 airliner will host a COIL a radar-based or optical cueing system to system—called the YAL-1A ﬁ nd and track the intended target. Airborne Laser—for standoff Effective beam weapons would initiate strikes agaitnst air vehicles, SOLID-STATE LASER is tested at Northrop Grumman nothing short of a revolution in warfare. especially ballistic missiles, and Space Technology. A high-power version of such a But packing all that technology into some- an AC-130 cargo plane-cum- device—capable of generating 100 kilowatts or thing the size of Captain Kirk’s handheld gunship will wield the Advanced more—could enable ground troops to destroy Tactical Laser, which is intended incoming mortars or rockets. phaser still clearly lies in the realm of sci-

for precision air-to-ground attacks. ence ﬁ ction. PHOTO GRUMMAN NORTHROP

I Fish That Go Skin-Deep SCAN TRAPPED FISH ADAPT TO A LIFE OF NIBBLING ON HUMANS BY MATT MOSSMAN

angal, Turkey—Tucked between walled off the spring from the creek in the brown hills in central Turkey is a nat- 1950s to preserve a captive school. A Turk- K ural hot spring where, for a fee, you ish family has now constructed a hotel, vil- can become fi sh food. Dip in a hand or foot, las and a playground and markets the resort and within seconds small fi sh will swarm, to psoriasis patients. Some 3,000 people ev- bump and nibble it. Stand above the pools, ery year pay for the privilege of sitting in the and the fi sh will gather below, waiting. The spring and allowing these omnivores to eat scaly swimmers—the “Doctor Fish of Kan- their dead skin, a process that may stimulate gal’’—supposedly have curative powers. But new skin growth or relax patients and therein this unusual case of adaptive ecology, the by ease stress-triggered psoriasis. human visitors may be helping the fi sh more Unquestionably for the fi sh, “the human than themselves. skin is a big help,” remarks Fevzi Bardakci, a These fi sh have acquired a taste for hu- biologist at Turkey’s Adnan Menderes Uni- mans largely because they have little choice. versity. “It’s like meat for them.” In 2000 The spring is too hot to sustain enough algae Bardakci published a paper in the World and plankton to feed them all. In the past, Wide Web Journal of Biology on Garra rufa, the fish were able to move between the one of the two species found in the hot spring. spring and a creek that runs nearby. But af- He discovered that members of the same spe- ter learning of a story about a local shepherd cies that swim in a nearby creek grow to an whose wounded leg healed after being average 97 millimeters and about 11 grams. dipped into the spring in 1917, builders In the hot spring, the fi sh are three-quarters

BITE ME: Fish in hot-spring pools at a Turkish resort survive by feasting on human skin. The water does ity, says Richard Londraville, a biologist at not sustain enough plankton and algae that would the University of Akron. He adds that those otherwise serve as food. in the hot spring may eventually evolve into a separate species in a few thousand years. the length and weigh one quarter as much. Other fi sh survive in waters as hot as or Moreover, during the summer spawning sea- hotter than this spring, which hovers near 34 son, the trapped females grow fewer and degrees Celsius. None are widely known to smaller oocytes, the cells that develop into feed on skin, which may explain why G. rufa eggs. In the creek, the gonads balloon from 3 are catching on elsewhere. A Chinese spa- percent of body weight to almost 8 percent. building company, which claims to have in- In the hot spring, organs increase from 1 per- vented the concept, says on its Web site that it cent of body weight to 2 percent. They would trained its own doctor fi sh and built 10 spas grow even less without submerged skin to in China, including in Beijing. Some of the nibble, Bardakci concluded. Some 90 percent Turkish fi sh were scooped up and enlisted in of visitors arrive in the summer, providing a springs in Japan, where at several spas they nutritional supplement at the perfect time. are now also performing fi sh pedicures. The fi sh come from the carp and min-

now family, which is known for adaptabil- Matt Mossman is based in Istanbul. MOSSMAN MATT COMPUTING

I Silicon Smackdown NEW GO ALGORITHM AIMS TO DEPOSE HUMANS BY KAREN A. FRENKEL

decade ago IBM’s chess program, an opponent by placing black or white stones Deep Blue, beat world champion Gar- on the intersections of a nine-by-nine or 19- A ry Kasparov in a six-game match. The by-19 line grid. Especially on the large board, event marked a milestone, forcing humans the number of possible moves per turn is to yield dominance of yet another strategic huge—200 on average for each midgame po- diversion. Only the Asian board game Go sition compared with the several dozen pos- seemed to be computer science’s Achilles’ sible in chess. There are also enormous heel: humans could soundly beat the ma- branching factors. Given N positions on the chines. A new algorithm can now take on board, the total number of possible game strong human players—and win. positions is 3N, because every position can Go has proved enormously difﬁ cult for be occupied by a black or white piece, or it computer programmers because of the can be empty. The total number of legal po- game’s deceptive complexity. The objective sitions on the small board is about 1038; on of Go is to stake out territory and surround the large board, about 10170. Additionally,

news more stones do not ensure victory, and play- the number of relevant potential moves. SCAN ers must be able to consider local positions UCT takes Monte Carlo further by fo- and the board as a whole. cusing the search on the most promising To cope with such an enormous number moves. “The main idea is to sample actions of options, artificial-intelligence experts selectively,” Kocsis says. The algorithm have designed algorithms to limit searches, must strike a balance, testing alternatives but the programs have not been able to beat that look the best at the moment to fi nd pos- the better human players on large boards. sible weaknesses and exploring “less opti- Last fall two Hungarian researchers report- mal-looking alternatives, to ensure that no ed that their algorithm outdid the win rates good alternatives are missed because of ear- of the best Go programs by 5 percent and ly estimation errors,” he explains. could compete with professional Go players UCT calculates indices for moves and se- on small boards. Levente Kocsis of the Com- lects the move that has the highest index. puter and Automation Research Institute at The algorithm computes the index from the the Hungarian Academy of Sciences in Buda- win rate, which describes how often that po- pest and Csaba Szepesvári, now at the Uni- sition leads to a win, as well as the number versity of Alberta in Edmonton, developed of times the position has been visited but not the algorithm, called UCT (for upper confi - played. UCT grows a decision tree in mem- dence bounds applied to trees). It extends the ory and uses it to track these statistics. When well-known Monte Carlo method. UCT encounters a move that has not been First incorporated into Go programs in visited previously, it adds the move to the tree the 1970s, Monte Carlo works like a po- and plays the rest of the game randomly. litical poll: it performs statistical sampling UCT decides if the fi nished random game to predict the behavior or characteristics of is a win or loss, then updates the statistics of a large group. When applied to Go, the al- all moves made during the game. If the index gorithm evaluates and ranks candidate equals the win rate of the move, the algo- moves by playing a large number of random rithm quickly focuses on the most promising games. But playing the move with the high- path. But nothing guarantees that an initial- est score in each position does not guaran- ly successful path will eventually yield a win- tee that the player will win the game. In- ning move. So when selecting moves, UCT stead this type of search merely restricts infl ates win rates by weighting less visited candidate moves more heavily. The researchers borrowed this idea from bandit problems—selective weighting yields the maximum gain for a gambler playing several slot machines with unknown average payoffs. Mathematicians Sylvain Gelly of the University of Paris-South and Yizao Wang of the Polytechnic School outside Paris have GO BEYOND incorporated UCT into a program they call GAMES MoGo. It has a 95 percent better win rate as compared with a previous state-of-the-art The new Go-playing algorithm, called upper confi dence bounds Monte-Carlo extension algorithm. Now the applied to trees (UCT), is not top-ranked Go algorithm, MoGo demon- limited to games. It applies to any strated its abilities this past spring, van- problem that involves choosing quishing strong amateur players on nine-by-

the best option, as long as nine boards and beating weaker ones on alternatives have an internal treelike structure (that is, a large boards. Gelly says that UCT is simple cascading set of choices) and to implement and can be improved. So turn- their values can be recursively ing the ultimate corner—ending the reign of computed. UCT may prove useful professional human Go players—could oc- for targeting advertisements cur in 10 years, Kocsis states. Daily PrincetonianThe on the Web, ﬁ nding the best settings for an industrial plant GO ATTACK: Humans still rule when it comes to Go, or optimizing channel allocation but a new algorithm can topple strong players. Karen A. Frenkel is based

in cellular systems. in New York City. SHIN JAE WON

news NEUROSCIENCE SCAN Brain Damage for Easier Moral Choices

How long would you hesitate before push- many. In a similar scenario, VMPC patients ing someone in front of a runaway train to were five times more likely to advocate keep it from killing fi ve other people? The smothering one’s baby to save others. Senior answer may be no time at all, if you have author Antonio Damasio of the University damage to the ventromedial prefrontal cor- of Southern California says that the patients tex (VMPC)—a region in the forebrain as- are not amoral but seem to lack the natural sociated with emotional response. Re- confl ict between emotion and reason. The searchers confronted volunteers with such study, in the March 22 Nature, also shows scenarios and found that those with VMPC that such decisions result not from a single injury were three times more likely than moral faculty but from two different pro- healthy people to advocate throwing the cesses that can compete with each other. person to certain death for the good of the —Nikhil Swaminathan DATA POINTS: COLLISION PLANETARY SCIENCE DECISION Martian Cave Dwellings Design fl aws caused a support structure for a magnet to fail Seven football fi eld–size caves may have been discov- during a stress test of the Large Hadron Collider (LHC) on March 27. ered on Mars. Analysis of photographs from NASA’s Built to be the world’s most Mars Odyssey orbiter revealed black spots near the powerful particle accelerator, the massive Martian volcano Arsia Mons that do not look CRATERS serve as skylights to underground LHC will smash lead ions into one like impact craters because they lack blast patterns and caverns on Mars, as seen in this artist’s another with energies of trillions raised rims. Scientists at Northern Arizona University rendering based on photographs. of electron volts (TeV). Repairs may delay the LHC’s target and their colleagues say the possible caverns range start date of November. from 330 to 825 feet wide and are 425 feet deep and have named them after their loved ones: Dena, Chloe, Wendy, Annie, Abbey, Nikki and Jeanne. Caves would serve as havens from Circumference of LHC radiation on the surface and so would be the most likely areas to harbor life. They could also in meters: 26,659 accumulate ice, which could help to support future human exploration. NASA’s Mars Re- Energy imparted to each connaissance Orbiter could take sidelong glances at the putative caves, a view that might show proton in TeV: 7 whether wider chambers exist underneath. The fi ndings were unveiled during a March meeting of the Lunar and Planetary Science Conference in League City, Tex. —Charles Q. Choi Total collision energy of lead-ion beams (consisting of many protons) in TeV: 1,150 EPIDEMICS

Number of particles in each beam: 300 trillion Stick It to the Kids

Odds that two particles will When flu epidemics loom, the long-standing recommendation of the U.S. Centers for Dis- actually collide: 1 in 10 billion ease Control and Prevention is to vaccinate the elderly ﬁ rst, because they are at greater risk of dying if they contract the virus. New evidence

Approximate power consumption COLLINS MATT BY ILLUSTRATION ); suggests that youngsters should take priority. in watts of: Work at Yale University and Rutgers University

A household microwave oven: bottom 1,100 underlines that children are the group most re- The LHC: 120 million sponsible for spreading the ﬂ u: they carry the virus into the home and infect adults, who then Time it takes to defrost a pizza in: A household microwave oven: bring the ﬂ u into the workplace. Vaccinating most young people would virtually eliminate ( GETTY IMAGES );

6 minutes top The LHC: 7.6 femtoseconds the ﬂ u, the researchers calculate, thereby cut- (7.6 ϫ 10–15 second) ting down on the mortality of the elderly, the SOURCES: CERN; DiGiorno’s Microwave young and people overall. The details are in the Rising Crust Four-Cheese Pizza. Pizza TARGETED for ﬂ u shots: children are the main calculation is theoretical and assumes March 27 Proceedings of the National Acade- that the collision energy can be evenly vectors of the illness.

distributed over the pizza. my of Sciences USA. —Charles Q. Choi MICHAEL CARROLL (

SCAN PALEONTOLOGY Collagen from T. Rex

Researchers have extracted collagen protein from a 68-million-year-old Tyrannosaurus rex femur, which two years ago was revealed to have soft tissue. Chem- ical analysis of the protein yielded seven sequences of about 10 to 20 amino acids in length. Three sequences matched collagen peptide scripts from chickens, one matched a frog and another a salamander; the other two matched multiple organisms, including chickens and salamanders. The results strengthen the bird-dinosaur connection and jettison the belief that ancient fossils could not provide protein samples for study (sorry, Jurassic Park fans—any genetic material degraded long ago). The ﬁ ndings and the techniques used to uncover them should clarify the relations between extinct species and modern-day animals and reveal more about “patterns of molecular change and the rates and directions of molecular evolution,” BRIEF says Mary Schweitzer of North Carolina State University, who reports the work POINTS with her colleagues in the April 13 Science. —Nikhil Swaminathan

■ Blood for all: researchers have discovered two bacterial BEHAVIOR VISION enzymes that can efﬁ ciently remove immune-triggering sugar molecules from red blood cells, Brain Brakes Rodent Roy G. Biv thereby turning A, B and AB blood types into the universal type O. So what stops you from pressing that send Mice, like most mammals, normally view button for an e-mail that tells off your boss? the world in colors limited to yellows, blues Nature Biotechnology, April Three distant brain regions connected by and grays, similar to what people with red- ■ New car smell—a stew of “hyperdirect cables,” believe scientists at the green color blindness see. By introducing a volatile organic compounds—is not toxic, at least not to human University of California, San Diego. They single human gene into mice, scientists en- cells in culture. The chemicals did asked participants to plan an action, listen dowed the animals with full-color vision. aggravate the cells’ immune for a stop signal and decide whether to obey Humans and closely related primates pos- response, suggesting that people or continue as planned. Brain scans revealed sess an extra light-sensitive pigment permit- with allergies should beware. that a neural braking network kicks in for a ting them to see red. (Color-seeing mam- Environmental Science & Technology, few milliseconds—just long enough for par- mals have at least two pigments, for blue April 1 ticipants to make a decision. The inferior and green.) Researchers at the University of ■ Lower-dose chemo: a fast- frontal cortex sends the braking signal to California, Santa Barbara, and their col- screening process using small the midbrain’s subthalamic nucleus, which leagues inserted the gene for this extra pig- interfering RNA molecules has uncovered 87 genes that affect stops motor movement; a third region, the ment into the mouse X chromosome. Even chemotherapy sensitivity; presupplementary motor area, initiates the though the rodent brains had not evolved to silencing some of those genes, plan to halt or continue the action. No syn- use these signals, they were able to rewire for instance, made lung cancer apses lie between the areas, enabling direct themselves to handle the upgrade, correctly cells 10,000 times more and fast communication. Understanding discriminating between colored lights to sensitive to the drug Taxol. this network could explain neurological dis- win soy-milk rewards. The investigators, Nature, April 12 orders such as stuttering, which may arise who detail the work in the March 23 Sci- ■ Planets of other solar systems from the brain’s inability to coordinate stop ence, say this ﬁ nding could help explain might have plants whose signals. The work appears in the April 4 how color vision evolved in humans. dominant colors are yellow, red or Journal of Neuroscience. —Thania Benios —Charles Q. Choi

even “infrared.” NASA scientists Researchers, Inc. Photo say they can predict the foliage color by the type of light emitted News Scan briefs with this icon have extended coverage at by the parent star. sa www.sciam.com/ontheweb

Astrobiology, March HARRIS ROGER

The (Other) Secret The inverse square law trumps the law of attraction By MICHAEL SHERMER

An old yarn about a classic marketing con game on the secret and in accordance with Maxwell’s equations any electric cur- of wealth instructs you to write a book about how to make a rent produces a magnetic fi eld. But as neuroscientist Russell lot of money and sell it through the mail. When your marks A. Poldrack of the University of California, Los Angeles, ex- receive the book, they discover the secret—write a book about plained to me, these fi elds are minuscule and can be measured how to make a lot of money and sell it through the mail. only by using an extremely sensitive superconducting quan- A confi dence scheme similar to this can be found in The tum interference device (SQUID) in a room heavily shielded Secret (Simon & Schuster, 2006), a book and DVD by Rhonda against outside magnetic sources. Plus, remember the inverse Byrne and a cadre of self-help gurus that, thanks to Oprah square law: the intensity of an energy wave radiating from a Winfrey’s endorsement, have now sold more than three million source is inversely proportional to the square of the distance copies combined. The secret is the so-called law of attraction. from that source. An object twice as far away from the source Like attracts like. Positive thoughts sally forth from your body of energy as another object of the same size receives only one- as magnetic energy, then return in the form of whatever it was fourth the energy that the closer object receives. The brain’s you were thinking about. Such as money. “The only reason any magnetic fi eld of 10–15 tesla quickly dissipates from the skull person does not have enough money is be- and is promptly swamped by other magnet- cause they are blocking money from coming A pantheon of shiny, ic sources, not to mention the earth’s mag- to them with their thoughts,” we are told. happy people netic fi eld of 10–5 tesla, which overpowers it Damn those poor Kenyans. If only they by 10 orders of magnitude! weren’t such pessimistic sourpusses. The assures viewers that Ceteris paribus, it is undoubtedly better fi lm’s promotional trailer is fi lled with such The Secret is to think positive thoughts than negative vainglorious money mantras as “Everything grounded in science. ones. But in the real world, all other things I touch turns to gold,” “I am a money mag- are never equal, no matter how sanguine net,” and, my favorite, “There is more money being printed for your outlook. Just ask the survivors of Auschwitz. If the law of me right now.” Where? Kinko’s? attraction is true, then the Jews—along with the butchered A pantheon of shiny, happy people assures viewers that The Turkish-Armenians, the raped Nanking Chinese, the massa- Secret is grounded in science: “It has been proven scientifi cally cred Native Americans and the enslaved African-Americans— that a positive thought is hundreds of times more powerful than had it coming. The latter exemplar is especially poignant given a negative thought.” No, it hasn’t. “Our physiology creates Oprah’s backing of The Secret on her Web site: “The energy disease to give us feedback, to let us know we have an imbal- you put into the world—both good and bad—is exactly what anced perspective, and we’re not loving and we’re not grateful.” comes back to you. This means you create the circumstances Those ungrateful cancer patients. “You’ve got enough power of your life with the choices you make every day.” Africans cre- in your body to illuminate a whole city for nearly a week.” Sure, ated the circumstances for Europeans to enslave them? if you convert your body’s hydrogen into energy through nu- Oprah, please, withdraw your support of this risible twad- clear fi ssion. “Thoughts are sending out that magnetic signal dle—as you did when you discovered that James Frey’s mem- that is drawing the parallel back to you.” But in magnets, op- oir was a million little lies—and tell your vast following that posites attract—positive is attracted to negative. “Every thought prosperity comes from a good dollop of hard work and cre- has a frequency.... If you are thinking that thought over and ative thinking, the way you did it. over again you are emitting that frequency.” The brain does produce electrical activity from the ion cur- Michael Shermer is publisher of Skeptic (www.skeptic.com).

BRAD HINES rents ﬂ owing among neurons during synaptic transmission, His new book is Why Darwin Matters.

Going beyond X and Y Babies born with mixed sex organs often get immediate surgery. New genetic studies, Eric Vilain says, should force a rethinking about sex assignment and gender identity By SALLY LEHRMAN

When Eric Vilain began his medical school rotation By coincidence, Vilain was also reading the jour- two decades ago, he was assigned to France’s refer- nals of Herculine Barbin, a 19th-century hermaphro- ence center for babies with ambiguous genitalia. He dite. Her story of love and woe, edited by famed social watched as doctors at the Paris hospital would check constructionist Michel Foucault, sharpened his ques- an infant’s endowment and quickly decide: boy or tions. He set on a path to fi nd out what sexual “nor- girl. Their own discomfort and social beliefs seemed mality” really meant—and to fi nd answers to the ba- to drive the choice, the young Vilain observed with sic biology of sex differences. shock. “I kept asking, ‘How do you know?’ ” he re- Today the 40-year-old French native is one of a calls. After all, a baby’s genitals might not match the handful of geneticists on whom parents and doctors reproductive organs inside. rely to explain how and why sex determination in an infant may have taken an unusual route. In his genetics laboratory at the University of California, Los Angeles, Vilain’s fi ndings have pushed the fi eld toward not only improved technical understanding but more thoughtful treatment as well. “What really matters is what people feel they are in terms of gender, not what their family or doctors think they should be,” Vilain says. Genital ambiguity occurs in an estimated one in 4,500 births, and problems such as undescended testes happen in one in 100. Altogether, hospitals across the U.S. perform about fi ve sex-assignment surgeries every day. Some of Vilain’s work has helped topple ancient ideas about sex determination that lingered until very recently. Students have long learned in developmental biology that the male path of sex development is “active,” driven by the presence of a Y chromosome. In contrast, the female pathway is passive, a default route. French physiologist Alfred Jost seemed to prove this idea in experiments done in the 1940s, in which castrated rabbit embryos developed into females. In 1990, while at the University of Cambridge, Peter Goodfellow discovered SRY, a gene on the Y chromosome hailed as the “master switch.” Just one

base pair change in this sequence would produce ERIC VILAIN: SEX BIOLOGY IN THE MIX a female instead of a male. And when researchers Redux ■ Studies genetics of sex determination; advises on intersex diagnoses. integrated SRY into a mouse that was otherwise chro- ■ Birth rate in which genitals are ambiguous: 1 in 4,500. mosomally female, an XX fetus developed as a male. ■ Discourages hasty sex-assignment surgery based on distress of parents But studies by Vilain and others have shaped a

or physicians: “That should appropriately be treated by a psychologist.” more complex picture. Instead of turning on male de- BRAD SWONETZ

velopment directly, SRY works by blocking an “antitestis” gene, colleagues to let the explosion of new genetic fi ndings point he proposes. For one, males who have SRY but two female toward a more scientifi c approach. Rather than using “her- chromosomes range in characteristics from normal male to an maphrodite,” for instance, he recommended referring to a ambiguous mix. In addition, test-tube studies have found that “disorder of sexual development” (DSD) and applying the SRY can repress gene transcription, indicating that it operates more precise term of “ovatesticular DSD.” through interference. Finally, in 1994, Vilain’s group showed Although the attendees eventually concurred, not everyone that a male could develop without the gene. Vilain offers a likes the new terminology. Some who prefer “intersex” feel model in which sex emerges out of a delicate dance between a that a “disorder” is demeaning. Milton Diamond, who studies variety of promale, antimale, and possibly profemale genes. sex identity at the University of Hawaii, complains that it stig- Because researchers have long viewed the development of matizes people who have nothing wrong with their bodies. females as a default pathway, the study of profemale genes has But the decision to change nomenclature realizes a 15-year taken a backseat. Over the past few years, though, geneticists dream for Cheryl Chase, executive director of the Intersex have uncovered evidence for active female determination. Society of North America (ISNA). Chase has fought for years DAX1, on the X chromosome, seems to start up the female against secret, rushed surgeries intended to comfort parents pathway while inhibiting testis formation—unless the gene and adjust anatomy to match an assigned social gender. Re- has already been blocked by SRY. With too much DAX1, a calling how a doctor once called her “formerly intersex,” she person with the XY complement is born a female. Vilain’s hopes physicians will begin to see mixed sex characteristics as group found that another gene, WNT4, operates in a similar a lifelong medical condition instead of a problem to be quick- way to promote the formation of a female. ly fi xed. “Now that we’ve accomplished The researchers discovered that these two Terms such as the name change, culture can accomplish work together against SRY and other pro- “hermaphrodite” and a little magic for us,” she predicts. male factors. “Ovary formation may be “intersex” are vague For her, Vilain has been a valued ally just as coordinated as testis determination, and hurtful, Vilain says. in the process as a member of ISNA’s consistent with the existence of an ‘ovari- medical advisory board. The job, he ad- an switch,’ ” report geneticist David Schlessinger and his col- mits, forces him to listen to patients, a practice he considers laborators in a 2006 review in the journal Bioessays. unusual for the fi eld. He expects the new, medicalized termi- Lately Vilain has been exploring molecular determinants nology for DSD to have what he describes wryly as “an inter- of sex within the brain and whether they may be linked to gen- esting side effect,” in which “medical science should apply” der identity. Despite classic dogma, he is certain that sex hor- to clinical decisions about ambiguous sex. mones do not drive neural development and behavioral differ- Indeed, the new consensus statement on managing intersex ences on their own. SRY is expressed in the brain, he points out, disorders encourages physicians to see beyond a patient’s sex suggesting that genes infl uence brain sexual differentiation di- organs, agrees conference co-organizer Peter A. Lee. The rectly. His lab has identifi ed in mice 50 new gene candidates statement, released last fall, recommends speedy gender as- on multiple chromosomes for differential sex expression. Sev- signment but a more cautious approach to surgery. The fam- en of them begin operating differently in the brain before go- ily should participate in decision making, along with a multi- nads form. Vilain’s group is testing these fi ndings using mice disciplinary team of caregivers in specialties that include psy- and is collaborating with a clinic in Australia to study expres- chology and ethics. But Lee, a pediatric endocrinologist at the sion patterns of the sex-specifi c genes in transsexual people. Penn State College of Medicine, cautions that much more work This work, like much of Vilain’s efforts, treads on fairly lies ahead to fi ll in data gaps. For instance, physicians have not touchy ground. He copes by sticking to his fi ndings conserva- measured how their choices affect patients over a lifetime. tively. “You also have to be aware of the social sensibilities,” On one Friday afternoon Vilain’s white coat and stetho- he explains. Accordingly, he has come to agree with some scope lay tossed amid the papers on his desk, a reminder that gender activists that it is time to revamp the vocabulary used his discoveries have more than philosophical meaning. He sees to describe ambiguously sexed babies. six to eight patients in the U.C.L.A. intersex clinic every month, At the 2005 Intersex Consensus Meeting in Chicago, he and in his on-call capacity, he receives two calls about babies stood before a group of 50 geneticists, surgeons, psycholo- in the hospital within the space of a couple of hours. Even gists and other specialists and argued that terms such as “her- while immersed in the workings of DNA transcription, Vilain maphrodite,” male or female “pseudohermaphrodite” and stays grounded in what his fi ndings mean for people’s lives. “intersex” were vague and hurtful. Instead of focusing on a newborn’s confusing mix of genitals and gonads, he urged his Sally Lehrman is based in the San Francisco Bay Area. www.sciam.com SCIENTIFIC AMERICAN 41 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more.. Sustainable Developments

Climate Change Refugees As global warming tightens the availability of water, prepare for a torrent of forced migrations By JEFFREY D. SACHS

Human-induced climate and hydrological change is likely to Plain and in China’s Yellow River Basin will most likely face make many parts of the world uninhabitable, or at least un- severe disruptions in water availability. Yet those regions are economic. Over the course of a few decades, if not sooner, already experiencing profound water stress because of unsus- hundreds of millions of people may be compelled to relocate tainable rates of groundwater pumping performed to irrigate because of environmental pressures. large expanses of northern China and northern India. To a signifi cant extent, water will be the most important In Africa, all signs suggest currently subhumid and arid determinant of these population movements. Dramatic altera- areas will dry further, deepening the food crisis for many of the tions in the relation between water and society will be wide- world’s poorest and most vulnerable people. The severe decline spread, as emphasized in the new report from Working Group in precipitation in the African Sahel during the past 30 years II of the Intergovernmental Panel on Climate Change. These seems to be related to both anthropogenic warming and aero- shifts may include rising sea levels, stronger tropical cyclones, sol pollutants. The violence in Darfur and Somalia is funda- the loss of soil moisture under higher tem- mentally related to food and water insecu- peratures, more intense precipitation and Hundreds of millions rity. Ivory Coast’s civil war stems, at least fl ooding, more frequent droughts, the melt- of people may be in part, from ethnic clashes after people fl ed ing of glaciers and the changing seasonality compelled to relocate. the northern drylands of Burkina Faso for of snowmelt. the coast. Worse chaos could easily arise. Impacts will vary widely across the world. It will be impor- Each El Niño cycle brings drying to thousands of islands tant to keep our eye on at least four zones: low-lying coastal in the Indonesian archipelago, with attendant crop failures, settlements, farm regions dependent on rivers fed by snowmelt famine and peat fi res. Some climatologists hypothesize that and glacier melt, subhumid and arid regions, and humid areas global warming could induce a more persistent El Niño state; in Southeast Asia vulnerable to changes in monsoon patterns. if so, the 200 million people in Indonesia and neighboring A signifi cant rise in sea levels, even by a fraction of a meter, areas could experience lasting drought conditions. could wreak havoc on tens or even hundreds of millions of Until now, the climate debate has focused on the basic sci- people. One study found that although coastal areas less than ence and the costs and benefi ts of reducing greenhouse gas 10 meters above sea level constitute only 2 percent of the world’s emissions. Attention will now increasingly turn to the urgent land, they contain 10 percent of its population. These coastal challenge of adapting to the changes and helping those who zones are vulnerable to storm surges and increased intensity of are most affected. tropical cyclones—call it the New Orleans Effect. Some hard-hit places will be salvaged by better infrastruc- Regions much farther inland will wither. Hundreds of mil- ture that protects against storm surges or economizes on water lions of people, including many of the poorest farm house- for agriculture. Others will shift successfully from agriculture holds, live in river valleys where irrigation is fed by melting to industry and services. Yet some places will be unable to glaciers and snow. The annual snowmelt is coming earlier ev- adjust altogether, and suffering populations will most likely ery year, synchronizing it less and less well with the summer move. We are just beginning to understand these phenomena growing season, and the glaciers are disappearing altogether. in quantitative terms. Economists, hydrologists, agronomists Thus, the vast numbers of farmers in the Indo-Gangetic and climatologists will have to join forces to take the next steps in scientifi c understanding of this human crisis.

An expanded version of this essay is available at sa www.sciam.com/ontheweb Jeffrey D. Sachs is director of the Earth Institute at

BRUCE GILBERT/EARTH INSTITUTE GILBERT/EARTH BRUCE Columbia University (www.earth.columbia.edu).

OARIGIN Simpler forLIFE

BY ROBERT SHAPIRO

The sudden appearance of a large self-copying molecule such as RNA was exceedingly improbable. Energy-driven networks of small molecules afford better odds as the initiators of life

xtraordinary discoveries inspire extraor- part in life.) Many different defi nitions of life have dinary claims. Thus, James Watson re- been proposed. Muller’s remark would be in accord ported that immediately after he and with what has been called the NASA defi nition: life Francis Crick uncovered the structure of is a self-sustained chemical system capable of un- EDNA, Crick “winged into the Eagle (pub) to tell dergoing Darwinian evolution. everyone within hearing that we had discovered the Richard Dawkins elaborated on this image of secret of life.” Their structure—an elegant double the earliest living entity in his book The Selfi sh helix—almost merited such enthusiasm. Its propor- Gene: “At some point a particularly remarkable tions permitted information storage in a language molecule was formed by accident. We will call it the in which four chemicals, called bases, played the Replicator. It may not have been the biggest or the same role as 26 letters do in the English language. most complex molecule around, but it had the ex- Further, the information was stored in two long traordinary property of being able to create copies chains, each of which specifi ed the contents of its of itself.” When Dawkins wrote these words 30 partner. This arrangement suggested a mechanism years ago, DNA was the most likely candidate for for reproduction: The two strands of the DNA dou- this role. Later, researchers turned to other possible ble helix parted company, and new DNA building molecules as the earliest replicator, but I and others blocks that carry the bases, called nucleotides, lined think that this replicator-fi rst model of the origin up along the separated strands and linked up. Two of life is fundamentally fl awed. We prefer an alter- double helices now existed in place of one, each a native idea that seems much more plausible. replica of the original. The Watson-Crick structure triggered an ava- When RNA Ruled the World lanche of discoveries about the way living cells complications to the DNA-fi rst theory soon function today. These insights also stimulated spec- set in. DNA replication cannot proceed without the ulations about life’s origins. Nobel laureate H. J. assistance of a number of proteins—members of a Muller wrote that the gene material was “living ma- family of large molecules that are chemically very terial, the present-day representative of the fi rst different from DNA. Both are constructed by link- life,” which Carl Sagan visualized as “a primitive ing subunits together to form a long chain, but free-living naked gene situated in a dilute solution whereas DNA is made of nucleotides, proteins are of organic matter.” (In this context, “organic” made of amino acids. Proteins are the handymen of specifi es compounds containing bound carbon at- the living cell. Enzymes, proteins’ most famous sub- oms, both those present in life and those playing no class, act as expediters, speeding up chemical pro-

An earlier, longer version of this story was posted on www.sciam.com. Feedback about that version helped to shape the article that appears here. BioGrafx KEN EWARD EWARD KEN

cesses that would otherwise take place plate an RNA world, containing only tinguish it from the assertion that RNA too slowly to be of use to life. Proteins RNA molecules that serve to catalyze merely arose before DNA and proteins. used by cells today are built following in- the synthesis of themselves…. The fi rst structions encoded in DNA. step of evolution proceeds then by RNA The Soup Kettle Is Empty The above account brings to mind the molecules performing the catalytic ac- the rna-first hypothesis faces old riddle: Which came fi rst, the chicken tivities necessary to assemble themselves a tremendously challenging question: or the egg? DNA holds the recipe for pro- from a nucleotide soup.” In this vision, How did that fi rst self-replicating RNA tein construction. Yet that information the first self-replicating RNA that arise? Enormous obstacles block Gil- cannot be retrieved or copied without emerged from nonliving matter carried bert’s picture of RNA forming in a non- the assistance of proteins. Which large out the various functions now executed living nucleotide soup. molecule, then, appeared fi rst—proteins by RNA, DNA and proteins. RNA’s building blocks, nucleotides, (the chicken) or DNA (the egg)? A number of additional clues support are complex substances as organic molecules go. Each contains a sugar, a phosphate and one of four nitrogen-containing bases as sub-subunits. Thus, each Inanimate nature provides us with RNA nucleotide contains nine or 10 carbon atoms, numerous nitrogen and oxy- a variety of mixtures of small molecules gen atoms and the phosphate group, all connected in a precise three-dimension- as potential incubators for life. al pattern. Many alternative ways exist for making those connections, yielding A possible solution appeared when the idea that RNA appeared before pro- thousands of plausible nucleotides that attention shifted to a new champion— teins and DNA in the evolution of life. could readily join in place of the stan- RNA. This versatile class of molecule is, For example, many small molecules, dard ones but that are not represented in like DNA, assembled of nucleotide called co-factors, play a role in enzyme- RNA. That number is itself dwarfed by building blocks but plays many roles in catalyzed reactions. These co-factors the hundreds of thousands to millions of our cells. Certain RNAs ferry informa- often carry an attached RNA nucleotide stable organic molecules of similar size tion from DNA to ribosomes, structures with no obvious function. Such struc- that are not nucleotides. (which themselves are largely built of tures have been considered “molecular The idea that suitable nucleotides other kinds of RNA) that construct pro- fossils,” relics descended from the time might nonetheless form draws inspira- teins. In carrying out its various duties, when RNA alone, without DNA or pro- tion from a well-known experiment pub- RNA can take the form of a double helix teins, ruled the biochemical world. lished in 1953 by Stanley L. Miller. He that resembles DNA or of a folded single This clue and others, however, sup- applied a spark discharge to a mixture of strand, much like a protein. port only the conclusion that RNA pre- simple gases that were then thought to In the early 1980s scientists discov- ceded DNA and proteins; they provide represent the atmosphere of the early ered ribozymes, enzymelike substances no information about the origin of life, earth and saw that amino acids formed. made of RNA. A simple solution to the which may have involved stages prior to Amino acids have also been identifi ed in chicken-and-egg riddle now appeared to the RNA world in which other living en- the Murchison meteorite, which fell in fall into place: life began with the ap- tities ruled supreme. Confusingly, re- Australia in 1969. Nature has apparently pearance of the fi rst self-copying RNA searchers use the term “RNA world” to been generous in providing a supply of molecule. In a germinal 1986 article, refer to both notions. Here I will use the these particular building blocks. By ex- Nobel laureate Walter Gilbert wrote in term “RNA fi rst” for the claim that RNA trapolation of these results, some writers the journal Nature: “One can contem- was involved in the origin of life, to dis- have presumed that all life’s building blocks could be formed with ease in Origin of Life Miller-type experiments and were pres- Overview/ ent in meteorites. This is not the case. ■ Theories of how life fi rst originated from nonliving matter fall into two broad Amino acids, such as those produced classes—replicator fi rst, in which a large molecule capable of replicating (such in experiments like Miller’s, are far less as RNA) formed by chance, and metabolism fi rst, in which small molecules complex than nucleotides. Their defi ning formed an evolving network of reactions driven by an energy source. features are an amino group (a nitrogen ■ Replicator-fi rst theorists must explain how such a complicated molecule and two hydrogens) and a carboxylic acid could have formed before the process of evolution was under way. group (a carbon, two oxygens and a hy- ■ Metabolism-fi rst proponents must show that reaction networks capable drogen), both attached to the same car- of growing and evolving could have formed when the earth was young. bon. The simplest of the 20 amino acids used to build natural proteins contains

only two carbon atoms. Seventeen of the In contrast, no nucleotides of any kind ecules made of fewer rather than greater set contain six or fewer carbons. The have been reported as products of spark- numbers of carbon atoms and thus shows amino acids and other substances that discharge experiments or in studies of no partiality in favor of creating the nu- were prominent in the Miller experiment meteorites. Apparently inanimate nature cleotides required by our kind of life. contained two and three carbon atoms. has a bias toward the formation of mol- To rescue the RNA-first concept from this otherwise lethal defect, its ad- REPLICATOR VS. METABOLISM vocates have created a discipline called prebiotic synthesis. They have attempted Scientifi c theories of the origin of life largely fall into two rival camps: replicator fi rst and metabolism fi rst. Both models must start from molecules formed by nonbiologi- to show that RNA and its components cal chemical processes, represented here by balls labeled with symbols (1). can be prepared in their laboratories in In the replicator-fi rst model, some of these compounds join together in a chain, a sequence of carefully controlled reac- by chance forming a molecule—perhaps some kind of RNA—capable of reproducing tions, using what they consider to be rel- itself (2). The molecule makes many copies of itself (3), sometimes forming mutant evant conditions and starting materials. versions that are also capable of replicating (4). Mutant replicators that are better The Web version of this article, avail- adapted to the conditions supplant earlier versions (5). Eventually this evolutionary able at www.sciam.com/ontheweb, goes process must lead to the development of compartments (like cells) and metabolism, into more detail about the shortcomings in which smaller molecules use energy to perform useful processes (6). of prebiotic synthesis research. The prob- 7 Metabolism fi rst starts off with the spontaneous formation of compartments ( ). lems bring the following analogy to Some compartments contain mixtures of the starting compounds that undergo mind: Consider a golfer who, having cycles of reactions (8), which over time become more complicated (9). Finally, the system must make the leap to storing information in polymers (10). played a ball through an 18-hole course, then assumes that the ball could also play itself around the course in his absence. REPLICATOR FIRST A S 1 ! T © METABOLISM FIRST * ¢ He had demonstrated the possibility of Replicator ¿ • % L ¤ Compartments molecule F ± « the event; it was only necessary to pre- § # ‡ + @ C R sume that some combination of natural Chance T £ 7 Chance 2 ¥ forces (earthquakes, winds, tornadoes B and fl oods, for example) could produce S A R L I F ‡ + T T E A D the same result, given enough time. No C R % @ # £ physical law need be broken for sponta- Replication 3 § ¥ neous RNA formation to happen, but the chances against it are immense. A S T R T L I F E Reaction cycles 8 Selection Some chemists have suggested that a S T A R T L I F E simpler replicator molecule similar to RNA arose fi rst and governed life in a B 4 Mutation C “pre-RNA world.” Presumably this fi rst E replicator would also have the catalytic A R I A # S T T L F E D capabilities of RNA. Because no trace of this hypothetical primal replicator and S M A R T L I F E B C catalyst has been recognized so far in E 5 Selection 9 More complex modern biology, RNA must have com- cycles A # pletely taken over all its functions at G D L some point after its emergence. S M A R T L I F E F Yet even if nature could have provid- R T I S M A L F E M C ed a primordial soup of suitable building

Compartments form, and 6 10 Information is stored ROBERT SHAPIRO metabolism develops in polymers is professor emeritus of chemistry and senior research sci-

E E entist at New York University. He is au- L F T I F thor or co-author of more than 125 pub- R I T L E B A R G Metabolism lications, primarily in the area of DNA M A D C THE AUTHOR chemistry. In particular, he has studied S F M A the ways in which environmental chem- # S M icals can damage our hereditary mate- L Protocell C rial, causing changes that can lead to

JEN CHRISTIANSEN CHRISTIANSEN JEN mutations and cancer.

blocks, whether nucleotides or a simpler that fall outside the scope of scientifi c tion to heat. The released heat increases substitute, their spontaneous assembly inquiry.” DNA, RNA, proteins and oth- the entropy of the environment, com- into a replicator involves implausibilities er elaborate large molecules must then pensating for the decrease in living sys- that dwarf those required for the prepa- be set aside as participants in the origin tems. The boundary maintains this divi- ration of the soup. Let us presume that of life. Inanimate nature instead pro- sion of the world into pockets of life and the soup of building blocks has somehow vides us with a variety of mixtures of the nonliving environment in which they been assembled, under conditions that small molecules with which to work. must sustain themselves. favor their connection into chains. They Fortunately, an alternative group of Today sophisticated double-layered would be accompanied by hordes of de- theories that can employ these materials cell membranes, made of chemicals clas- fective units, the inclusion of which in a has existed for decades. The theories use sifi ed as lipids, separate living cells from nascent chain would ruin its ability to act a thermodynamic, rather than a genetic, their environment. When life began, as a replicator. The simplest kind of definition of life, under a scheme put some natural feature probably served the fl awed unit would have only one “arm” forth by Sagan in the Encyclopedia Bri- same purpose. In support of this idea, available for connection to a building tannica: a localized region that increases David W. Deamer of the University of block, rather than the two needed to sup- in order (decreases in entropy) through California, Santa Cruz, has observed port further growth of the chain. cycles driven by an energy fl ow would be membranelike structures in meteorites. An indifferent nature would theoreti- considered alive. This small-molecule ap- Other proposals have suggested natural cally combine units at random, produc- proach is rooted in the ideas of Soviet bio- boundaries not used by life today, such ing an immense variety of short, termi- chemist Alexander Oparin. Origin-of-life as iron sulfi de membranes, rock surfaces nated chains, rather than the much longer proposals of this type differ in specifi c de- (in which electrostatic interactions segre- one of uniform backbone geometry need- tails; here I will list fi ve common require- gate selected molecules from their envi- ed to support replicator and catalytic ments (and add some ideas of my own). ronment), small ponds and aerosols. functions. The probability of this latter 1. A boundary is needed to separate 2. An energy source is needed to process succeeding is so vanishingly small life from nonlife. Life is distinguished by drive the organization process. We con- that its happening even once anywhere in its great degree of organization, yet the sume carbohydrates and fats, combining the visible universe would count as a piece second law of thermodynamics requires them with oxygen that we inhale, to of exceptional good luck. that the universe move in a direction in keep ourselves alive. Microorganisms which disorder, or entropy, increases. A are more versatile and can use minerals Life with Small Molecules loophole, however, allows entropy to de- in place of the food or the oxygen. In nobel laureate Christian de Duve crease in a limited area, provided that a either case, the transformations that are has called for “a rejection of improbabil- greater increase occurs outside the area. involved are called redox reactions. They ities so incommensurably high that they When living cells grow and multiply, entail the transfer of electrons from an can only be called miracles, phenomena they convert chemical energy or radia- electron-rich (or reduced) substance to FIVE REQUIREMENTS FOR METABOLISM FIRST

At least fi ve processes must occur for small molecules to must drive a chemical reaction (3). A network of chemical achieve a kind of life—here defi ned as the creation of greater reactions must form and increase in complexity to permit order in localized regions by chemical cycles driven by an energy adaptation and evolution (4). Finally, the network of reactions fl ow. First, something must create a boundary to separate the must draw material into itself faster than it loses material, and living region from the nonliving environment (1). A source of the compartments must reproduce (5). No information-storing energy must be available, here depicted as a mineral (blue) molecule (such as RNA or DNA) is required; heredity is stored in undergoing a heat-producing reaction (2). The released energy the identity and concentration of the compounds in the network.

Molecule Mineral 1 2 3

Energy 5 4 JEN CHRISTIANSEN CHRISTIANSEN JEN

an electron-poor (or oxidized) one. What Readers Want to Know Plants can capture solar energy directly In Scientifi c American’s blog, Robert Shapiro answered questions raised by and adapt it for the functions of life. readers of the Web version of this article. An edited selection follows. Other forms of energy are used by cells in specialized circumstances—for ex- Does the metabolism-fi rst hypothesis point to a single origin or ample, differences in acidity on opposite multiple independent origins of life? —JR sides of a membrane. Yet others, such as A: Multiple origins seem more viable with the metabolism-fi rst scenario. Gerald radioactivity and abrupt temperature Feinberg and I discussed the possibility of alien life (life not based on RNA, DNA and differences, might be used by life else- other biochemistry familiar to us) in our 1980 book, Life beyond Earth. Researchers where in the universe. at a conference hosted by Paul Davies at Arizona State University in December 2006 3. A coupling mechanism must link concluded that alien life may even exist, undetected, on this planet. The great the release of energy to the organization majority of microorganisms that can be observed under a microscope cannot be process that produces and sustains life. grown in conventional culture media and remain uncharacterized. Alien microbes The release of energy does not necessar- may also exist in habitats on the earth that are too extreme for even the hardiest forms of our familiar life. ily produce a useful result. Chemical energy is released when gasoline is burned Why do we have to demonstrate metabolism fi rst in a reaction within the cylinders of an automobile, —Dave Evanoff but the vehicle will not move unless that vessel? Can’t we simulate it in software? A: Stuart Kauffman, Doron Lancet and others have used computer simulations to energy is used to turn the wheels. A me- illustrate the feasibility of self-sustaining reaction cycles. Such simulations have not chanical connection, or coupling, is re- specifi ed the exact chemical mixtures and reaction conditions needed to establish quired. Every day, in our own cells, each self-sustaining chemical networks. We do not yet know all the reaction pathways of us degrades pounds of a nucleotide open to mixtures of simple organic compounds, let alone their thermodynamic called ATP. The energy released by this constants. Even if such data were available, most chemists would not be convinced by reaction serves to drive processes neces- a computer simulation but would demand an experimental demonstration. sary for our biochemistry that would otherwise proceed too slowly or not at The fact that all biological molecules are of one handedness needs all. Linkage is achieved when the reac- some explanation. —John Holt tions share a common intermediate, and A: If the mineral transformation that powered the reaction cycle I discuss in my the process is sped up by the intervention article were selective for only one mirror-image form of chemical A, then the product of an enzyme. One assumption of the B and other members of the cycle might also occur in only one mirror-image form. small-molecule approach is that coupled Control of handedness, or chirality, becomes crucial when small chiral molecules are reactions and primitive catalysts suffi - linked together to form larger ones. A modern enzyme may contain 100 linked amino cient to get life started exist in nature. acids, all of the same handedness (so-called L-amino acids). If a D-amino acid were substituted for its mirror-image L-form at a sensitive site within the enzyme, then 4. A chemical network must be the enzyme’s shape would change and its function might be lost. formed to permit adaptation and evolution. We come now to the heart of the matter. Imagine, for example, that an Branch reactions will occur as well, pathway from B to A, material would energetically favorable redox reaction of such as molecules converting back and back up until another route was found. a mineral drives the conversion of an or- forth between D and another chemical, Additional changes of this type would ganic chemical, A, to another one, B, E, that lies outside the ABCD cycle. Be- convert the original cycle into a network. within a compartment. I call this key cause the cycle is driven, the E-to-D reac- This trial-and-error exploration of the transformation a driver reaction, be- tion is favored, moving material into the chemical “landscape” might also turn cause it serves as the engine that mobi- cycle and maximizing the energy release up compounds that could catalyze im- lizes the organization process. If B simply that accompanies the driver reaction. portant steps in the cycle, increasing the reconverts back to A or escapes from the The cycle could also adapt to chang- effi ciency with which the network used compartment, we would not be on a path ing circumstances. As a child, I was fas- the energy source. that leads to increased organization. In cinated by the way in which water, re- 5. The network must grow and re- contrast, if a multistep chemical path- leased from a leaky hydrant, would fi nd produce. To survive and grow, the net- way—say, B to C to D to A—reconverts a path downhill to the nearest sewer. If work must gain material faster than it B to A, then the steps in that circular pro- falling leaves or dropped refuse blocked loses it. Diffusion of network materials cess (or cycle) would be favored to con- that path, the water would back up until out of the compartment into the exter- tinue operating because they replenish another route was found around the ob- nal world is favored by entropy and will the supply of A, allowing the contin- stacle. In the same way, if a change in the occur to some extent. Some side reac- uing useful discharge of energy by the acidity or in some other environmental tions may produce gases, which escape, mineral reaction [see box on page 53]. circumstance should hinder a step in the or form tars, which will drop out of so-

lution. If these processes together should functions in a compartment within a nism for heredity. In other words, they exceed the rate at which the network rock may overfl ow into adjacent com- contain no obvious molecule or structure gains material, then it would be extin- partments. Whatever the mechanism that allows the information stored in guished. Exhaustion of the external fuel may be, this dispersal into separated them (their heredity) to be duplicated and would have the same effect. We can units protects the system from total ex- passed on to their descendants. Yet a col- imagine, on the early earth, a situation tinction by a local destructive event. lection of small items holds the same in- where many start-ups of this type occur, Once independent units were established, formation as a list that describes the items. involving many alternative driver reac- they could evolve in different ways and For example, my wife gives me a shop- tions and external energy sources. Fi- compete with one another for raw mate- ping list for the supermarket; the collec- nally, a particularly hardy one would rials; we would have made the transition tion of grocery items that I return with take root and sustain itself. from life that emerges from nonliving contains the same information as the list. A system of reproduction must even- matter through the action of an available Doron Lancet of the Weizmann Institute tually develop. If our network is housed energy source to life that adapts to its of Science in Rehovot, Israel, has given in a lipid membrane, physical forces may environment by Darwinian evolution. the name “compositional genome” to he- split it after it has grown enough. (Free- redity stored in small molecules, rather man Dyson of the Institute for Advanced Changing the Paradigm than a list such as DNA or RNA. Study in Princeton, N.J., has described systems of the type i have de- The small-molecule approach to the such a system as a “garbage bag world” scribed usually have been classifi ed under origin of life makes several demands on in contrast to the “neat and beautiful the heading “metabolism fi rst,” which nature (a compartment, an external en- scene” of the RNA world.) A system that implies that they do not contain a mecha- ergy supply, a driver reaction coupled to that supply, a chemical network that includes that reaction, and a simple mech- An RNA-First Researcher Replies anism of reproduction). These requirements are general in nature, however, Steven A. Benner of the Westheimer Institute for Science and Technology and are immensely more probable than in Gainesville, Fla., argues that RNA-fi rst models are alive and well. the elaborate multistep pathways needed Even as some declare that the RNA-fi rst model of life’s origin is dead because RNA arising to form a molecule that is a replicator. spontaneously is fantastically improbable, research is lending support to the model. Over the years, many theoretical pa- Let me fi rst acknowledge that most organic molecules when hit with energy pers have advanced particular metabo- (such as lightning or heat from volcanoes) become something resembling asphalt, lism-fi rst schemes, but relatively little more suitable for paving roads than sparking life. But metabolism-fi rst models, to the experimental work has been presented extent that they have been supported with any real chemicals, must also deal with in support of them. In those cases where this paradox: molecules reactive enough to participate in metabolism are also reac- experiments have been published, they tive enough to decompose. There are no easy solutions. have usually served to demonstrate the Like many others, my research group has returned to the scientifi c imperative: plausibility of individual steps in a pro- actually do laboratory research to learn about how RNA might have arisen on the earth. posed cycle. The greatest amount of new The sugar ribose, the “R” in RNA, provides an object lesson in how a problem data has perhaps come from Günter declared “unsolvable” may instead merely be “not yet solved.” Ribose long remained Wächtershäuser and his colleagues at “impossible” to make by prebiotic synthesis (reactions among mixtures of molecules Munich Technical University. They have that could plausibly have existed on a prebiotic earth) because it contains a carbonyl demonstrated parts of a cycle involving group—a carbon atom twice bonded to an oxygen atom. The carbonyl group confers the combination and separation of ami- both good reactivity (the ability to participate in metabolism) and bad reactivity (the no acids in the presence of metal sulfi de ability to form asphalt). A decade ago Stanley L. Miller concluded that the instability catalysts. The energetic driving force for of ribose stemming from its carbonyl group “preclude[s] the use of ribose and other the transformations is supplied by the sugars as prebiotic reagents.... It follows that ribose and other sugars were not oxidation of carbon monoxide to carbon components of the fi rst genetic material.” dioxide. The researchers have not yet But prebiotic soups need soup bowls made of appropriate minerals, not Pyrex beakers. demonstrated the operation of a com- One attractive “bowl” is found today in Death Valley. In a primordial Death Valley, the envi- plete cycle or its ability to sustain itself ronment was alternately wet and dry, rich in organic molecules from planetary accretion and undergo further evolution. A “smok- and (most important) full of minerals containing boron. Why care about boron? Because ing gun” experiment displaying those boron stabilizes carbohydrates such as ribose. Further, if borate (an oxide of boron) and three features is needed to establish the organic compounds abundant in meteorites are mixed and hit with lightning, good quanti- validity of the small-molecule approach. ties of ribose are formed from formaldehyde and the ribose does not decompose. The principal initial task is the iden- The fact that such a simple solution can be found for a problem declared “unsolv- tifi cation of candidate driver reactions— able” does not mean that the fi rst form of life defi nitely used RNA to do genetics. But small-molecule transformations (A to B it should give us pause when advised to discard avenues of research simply because 52some SCIENTIFIC of their problematic AMERICAN pieces have not yet been solved. JUNE 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more..

EVOLUTION OF CHEMICAL NETWORKS steps need not duplicate those that took place on the early earth. It is more im- The metabolism-fi rst hypothesis requires the formation of a network of chemical portant that the general principle be reactions that increases in complexity and adapts to changes in the environment. demonstrated and made available for further investigation. Many potential CYCLE FORMATION: An energy source (here the so-called redox reaction converting mineral X to mineral Y) couples to a reaction that converts the organic molecule A to molecule B. paths to life may exist, with the choice Further reactions (B to C, C to D....) form a cycle back to A. Reactions involving molecular dictated by the local environment. species outside the cycle (E) will tend to draw more material into the cycle. An understanding of the initial steps leading to life would not reveal the spe- INCREASING COMPLEXITY: If a change in conditions inhibits a reaction in the cycle (for example, C to D), then other paths can be explored. Here a bypass has been found by which C cifi c events that led to the familiar DNA- is converted to D through intermediates F, G and H. Another solution would be the RNA-protein-based organisms of today. incorporation into the reaction network of a catalyst (I) whose action (dotted line) unblocks Still, because we know that evolution the C to D transformation. To survive, the evolving network must draw in carbon-containing materials from the environment more rapidly than it loses them by diffusion and side does not anticipate future events, we reactions, such as the formation of tars that settle out of the solution. can presume that nucleotides fi rst appeared in metabolism to serve some CYCLE FORMATION INCREASING COMPLEXITY other purpose, perhaps as catalysts or as containers for the storage of chemical Y Y Redox Redox energy (the nucleotide ATP continues to X reaction X reaction serve this function today). Some chance Heat Heat event or circumstance may have led to the connection of nucleotides to form Driver Driver reaction B reaction B RNA. The most obvious function of A A Catalyst modern RNA is to serve as a structural action element that assists in the formation of bonds between amino acids in the syn- C C thesis of proteins. The fi rst RNAs may D D I have served the same purpose, but without any preference for specifi c amino E acids. Many further steps in evolution Mineral F would be needed to “invent” the elabo- E Organic H rate mechanisms for replication and molecule specifi c protein synthesis that we ob- Chemical Tar G serve in life today. reaction If the general small-molecule para- Carbon supply digm were confi rmed, then our expecta- tions of the place of life in the universe in the preceding example) that are cou- crease in quantity. The reactor would would change. A highly improbable start pled to an abundant external energy need an input device (to allow replenish- for life, as in the RNA-fi rst scenario, im- source (such as the oxidation of carbon ment of the energy supply and raw mate- plies a universe in which we are alone. In monoxide or a mineral). Once a plausible rials) and an outlet (to permit removal of the words of biochemist Jacques Monod, driver reaction has been identifi ed, there waste products and chemicals that were “the universe was not pregnant with life should be no need to specify the rest of not part of the network). nor the biosphere with man. Our num- the system in advance. The selected com- In such experiments, failures would ber came up in the Monte Carlo game.” ponents (including the energy source), be easily identifi ed. The energy might The small-molecule alternative, plus a mixture of other small molecules be dissipated without producing any however, is in harmony with the views normally produced by natural processes signifi cant changes in the concentra- of biologist Stuart Kauffman: “If this is (and likely to have been abundant on the tions of the other chemicals, or the all true, life is vastly more probable than early earth), could be combined in a suit- chemicals might be converted to a tar, we have supposed. Not only are we at able reaction vessel. If an evolving net- which would clog the apparatus. A suc- home in the universe, but we are far work were established, we would expect cess might demonstrate the initial steps more likely to share it with as yet un- the concentration of the participants in on the road to life. These known companions.” the network to increase and alter with time. New catalysts that increased the Additional coverage—including commentaries, answers to questions, rate of key reactions might appear, sa links to further reading and the opportunity to post your own comments—can be found at www.sciam.com/ontheweb

JEN CHRISTIANSEN CHRISTIANSEN JEN whereas irrelevant materials would de-

Learning why current anesthetics are so potent and sometimes dangerous will lead to a new generation of safer targeted drugs without unwanted side effects

By Beverley A. Orser

A Hollywood thriller due out this year centers on a young man who awakens while undergoing open-heart surgery but is unable to move or cry out. The fi lm’s plot will undoubtedly take many more dramatic turns from there, but its early premise is, sadly, not entirely far-fetched. Episodes of intraoperative awareness while under general anesthesia are reported by one or two of every 1,000 patients. In reality, such incidents are usually brief and generally do not involve pain or distress, but they do highlight one of several ways that even the newest generation of anesthetic drugs can sometimes leave much to be desired. Indeed, the medical specialty of anesthesiology has evolved into a sophisticated art form because scientifi c understanding of how anesthetic drugs actually work, and how to make them better, has lagged behind most other areas of drug development. Many of the modern anesthetics, in fact, share structural properties and clinical effects with ether, whose application as an anesthetic was fi rst successfully demon- Corbis strated in public by Boston dentist William Morton in 1846. Since then, the use of general anesthesia has expanded to 40 million patients each year in North America alone. Yet advances in anesthetic care since Morton’s day have come largely from the development of complex drug delivery systems and strategies for managing anesthe-

sia’s dangers and side effects. HOLLINGSWORTH JACK

Today’s general anesthetics are the death during an operation than the sur- thesia’s main components are sedation, most potent depressors of nervous sys- gical procedure itself. And because an- unconsciousness (also sometimes called tem activity used in medicine. They even esthesia-related mortality has plateaued hypnosis), immobility, absence of pain affect regulation of breathing and heart at a rate of approximately one patient in (analgesia) and absence of memory for function. As a result, the drugs have a 13,000 for the past 15 years, it appears the anesthetized period (amnesia). By fairly narrow margin of safety, which is that anesthesiologists may have reached studying the mechanisms through the difference between the therapeutic the limits of our ability to deliver these which anesthetics achieve these end dose and a dose that is toxic, even lethal. toxins safely. Moreover, severe side ef- points, many groups, including my own That is one reason why individuals fects—ranging from loss of airway con- at the University of Toronto, are begin- whose lung or cardiovascular function trol to memory and cognitive problems ning to tease those effects apart. Such is already unstable—such as trauma vic- after general anesthesia—may also stem studies are revealing that the activity of tims undergoing emergency operations from the broad yet poorly understood these potent drugs involves highly spe- or patients in the midst of heart sur- infl uence that current anesthetics exert cifi c interactions with subpopulations gery—must receive a lighter than nor- on the central nervous system. of nervous system cells to create each of mal dose of anesthesia, which could Science should be able to do much the separate properties of anesthesia. make them susceptible to brief aware- better, and very recent research is begin- Armed with this knowledge, we will ness incidents, as in the movie. ning to reveal how it can. be poised to fi nally move beyond the Although radical improvements in ether era and develop a new generation the care of people under general anes- Pulling Out the Plugs of highly specifi c drugs that can be used thesia have laid the foundation for com- all of today’s general anesthetic in combinations to deliver only the de- plicated procedures such as organ trans- drugs were developed empirically, sired results without the dangers. As a plants and open-heart surgery, the pow- which is to say, they were tested for their bonus, this research is also yielding in- erful neurodepressive effects of these ability to produce the desirable effects sights that can improve related therapies, drugs make them more likely to cause that defi ne the anesthetized state. Anes- such as sedatives and sleep aids, that www.sciam.com SCIENTIFIC AMERICAN 55 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more..

share some of anesthesia’s mechanisms. ing anesthesia. One leading theory holds contain subtly different versions, how- Anesthetics fall into two main cate- that it is simply the result of “cognitive ever, which tend to predominate in dif- gories based on whether they are deliv- unbinding”—a severing of communica- ferent areas of the central nervous sys- ered by inhalation, such as isoﬂ urane, or tion between the many brain regions tem. The presence of particular receptor intravenously, such as propofol. These that usually cooperate in higher cogni- subtypes on only certain subpopulations drugs may appear to induce a deep sleep, tive processing. Even at the local level, if of cells will thus determine which cells but the state produced by most modern one imagines groups of neurons as form- are inﬂ uenced by an anesthetic. general anesthetics is more of a pharma- ing lines in a vast telephone network, the Contemporary studies are therefore cological coma. In a step toward clarify- effect of general anesthesia is analogous focusing on identifying which receptor ing the mechanisms underlying their to pulling out plugs at the switchboard. variants are the targets of current anes- effects, technologies such as magnetic Researchers are, however, making en- thetic drugs, understanding how the he effect of anesthesia is analogous T to pulling out plugs at the switchboard.

resonance imaging (MRI) and positron- couraging progress in discovering de- drugs interact with the receptors to emission tomography (PET) have helped tails about the ways that anesthetic change the cell’s function and how those identify some of the discrete brain re- drugs physically act on the individual cellular changes produce both the gions and neural circuits involved in spe- cells of the nervous system to block their desired and unwanted “symptoms” of cific components of the anesthetized transmissions. anesthesia. state. For instance, anesthetic action on During most of the 20th century, an- the spinal cord accounts for the immobil- esthetics were widely thought to work by Signaling Silence ity produced by the drugs, whereas drug- disrupting the lipid components of cell many categories of receptor pro- induced changes to the hippocampus, a membranes. Most anesthetics are highly teins are found on the surface of neu- brain structure involved in memory for- fat-soluble compounds with widely dif- rons, but those activated by natural neu- mation, have been linked to amnesia. fering chemical structures that range rotransmitter chemicals have garnered Chronic memory impairment following from simple inert gases to complex ste- the most interest in anesthesia research surgery, one of the undesirable side ef- roids. Their great physical and chemical because they critically regulate commu- fects suffered by some patients, may also diversity supported the idea that anes- nication along the neural “telephone represent a leftover infl uence of the drugs thetics must work in some nonspecifi c lines.” As their name implies, neuro- on the hippocampus. way to depress neuronal functioning. transmitter molecules transmit messag- Because consciousness is a complex Modern research has shown, however, es between neurons at points of contact experience whose defi ning properties that anesthetics actually interact with called synapses. They do so by traveling are still hotly debated by neuroscientists, multiple varieties of specifi c proteins, from the so-called presynaptic neuron it is not as easy to pinpoint a single ana- known as receptors, found on the sur- across a tiny gap to bind to receptors on tomical source of unconsciousness dur- face of nerve cells. Families of receptors the postsynaptic neuron’s cell membrane. When enough neurotransmitter Refi ning a Blunt Instrument molecules trigger the appropriate recep- Overview/ tors, the postsynaptic cell’s membrane ■ General anesthetics are powerful nervous system suppressors, but how generates an electrical potential that the drugs produce their broad effects throughout the brain and body travels down its length to the next neu- is poorly understood. ron in its network. Glutamate, sero- ■ Investigation of anesthetics’ underlying mechanisms is revealing that tonin, norepinephrine and acetylcholine individual aspects of the anesthetized state are attributable to different sets are just a few of the neurotransmitters of nerve cells, which are themselves distinguished by specifi c surface widely studied for their role in promot- proteins that interact with the drugs. ing such signaling throughout the cen- ■ New compounds designed to target just those proteins, and hence only tral nervous system. specifi c cell types, could be combined to selectively produce the desirable In anesthesia research, however, an- effects of anesthetics—as well as sedatives, sleep aids and memory drugs— other neurotransmitter called gamma- with fewer risks and side effects. aminobutyric acid (GABA) has gained the most attention because of its ability

to block neural communication. GABA is an inhibitory neurotransmitter: it Anesthesia’s Broad Impact helps to maintain overall balance in the Both the desirable and unwanted heart rate and breathing. Ongoing nervous system by dampening neurons’ effects of anesthetic drugs stem from research is attempting to pinpoint ability to respond to excitatory messag- their power to suppress neuronal the neural structures and regions es from other cells. For that reason, activity throughout the central whose changed functioning produces GABA is thought to play a central part nervous system, which encompasses each of the defining properties of the in the actions of anesthetic drugs. the brain and spinal cord and controls anesthetized state. Most postsynaptic receptors on cells that interact with GABA belong to a Motor cortex Somatosensory cortex class termed ligand-gated ion channels. When GABA (the ligand) binds to the receptor, the receptor changes its con- Prefrontal cortex formation, temporarily opening a chan- Thalamus nel that admits negatively charged ions Amygdala Hippocampus into the cell. The increased ion concen- Brain stem respiratory centers tration generates a negative potential, Cerebellum preventing the cell from being able to produce an excitatory electrical pulse. Spinal cord The receptor that is believed to be a primary target for anesthetics is the GABA subtype A, or GABAA, receptor, which is also known to underlie the therapeutic effects of other classes of sedative and hypnotic drugs, most notably benzodiazepines such as Valium. Very low concentrations of benzodiazepines increase GABAA receptor function, a relation that is easily confirmed because reversal agents that impede benzodiazepine from binding to the GABAA receptor rapidly blunt the effects of those drugs. Components of the Anesthetized State Unfortunately, no such reversal agents exist for general anesthetics that Sedation might provide clues to their receptor Reduced arousability, as evidenced by longer response times, slurred speech targets. Nevertheless, studies using slic- and decreased movement. Neuronal activity across brain cortical areas drops. es from many different brain regions and neurons grown in tissue culture Unconsciousness (also called hypnosis) Impaired perception of, and response to, stimuli. Cortical depression is deeper have shown that both intravenous and than in sedation. Activity in the thalamus, an area important for integrating brain inhaled anesthetics prolong the dura- processes, also falls signifi cantly. tion of postsynaptic electric currents generated by GABAA receptors. Immobility Anesthetics are believed to increase Lack of movement in response to stimulation such as shaking or heat. the function of GABAA receptors by in- Suppression of spinal cord neuronal activity is the main cause of this temporary teracting at discrete binding cavities or paralysis, although the cerebellum, a motor control area, may also contribute. attaching to specifi c amino acids in the receptors themselves and prolonging the Amnesia channel opening, which extends the in- Lack of recall for the anesthetized period. Many brain structures involved in hibitory effects of GABA molecules memory formation, including the hippocampus, amygdala, prefrontal cortex and sensory and motor areas, exhibit anesthetic-induced changes. bound to the receptor. At high enough concentrations, anesthetics may even Others trigger the GABA receptors alone. Muscle relaxation and lack of pain (analgesia) are sometimes included in The vast majority of neurons contain defi nitions of the anesthetized state and are largely attributed to depression GABAA receptors, however, so scientists of spinal cord activity.

CHRISTINE GRALAPP could not understand how anesthetics

could selectively infl uence different brain regions until research breakthroughs Jamming Transmission over the past decade revealed that not Anesthetic drugs have been found to dampen neuronal communication, in part, all GABAA receptors are structurally by enhancing the effects of the neurotransmitter GABA, a signaling molecule that or pharmacologically the same. The inhibits nerve cells from fi ring. Current research is focused on how the drugs GABAA receptor is a protein complex interact with cellular GABA receptors to block neural activity. composed of fi ve subunit parts, which can be mixed and matched in various Signal to Be Silent combinations. At least 19 different An electrical pulse in the membrane of one neuron provokes release of GABA GABA receptor subunits exist in mam- into the synapse, a juncture with another neuron. The molecules cross a A mals, and most of those have variant sub- small gap and bind to GABA-specifi c receptors on the postsynaptic cell. In many brain areas, GABA receptors are also found outside the synapse, along types, so the possible number of combi- the nerve cell body, and are activated by GABA that spills out of the synapse. nations is high. The subunits most commonly seen in neurons, however, are the ones designated alpha, beta and gamma. In fact, most GABAA receptors are composed of two alpha subunits, two betas and one gamma, although sometimes a Presynaptic neuron delta or epsilon subunit replaces the gamma, depending on the brain region. But the key discovery was that the recep- Vesicle tor’s subunit composition dramatically GABA alters its pharmacological properties: just one subunit difference within a GABAA receptor’s structure can deter- Synapse mine whether and how it will respond to a particular anesthetic drug. Postsynaptic neuron Because different GABAA receptor Extrasynaptic subtypes predominate in different brain GABA receptor GABA receptor regions, researchers are increasingly able to pinpoint how anesthetics produce specifi c effects in various parts of the Receptor central nervous system by examining subunits how the drugs interact in those regions Anesthesia molecule with their particular target receptors. GABA GABA Narrowing Down Targets my colleagues and I decided to fo- GABAA receptor cus on identifying the receptors that infl uence the memory-impairing proper- Cell membrane Hyperpolarized cell membrane ties of anesthetics, so we concentrated our studies on GABAA receptors in the Negative hippocampus. Anesthetics are known ions to cause amnesia at doses considerably lower than those required for unconsciousness or immobility, an effect that Anesthetics and GABA: Changing Charge is clearly evident to anesthesiologists, A receptor subtype called GABAA is a channel into the postsynaptic cell for example, because patients rarely re- composed of five protein subunits. When GABA binds to it, the receptor member their own animated conversa- opens to admit negatively charged ions, which increases polarization of the tions as they go under or emerge from cell’s membrane and prevents the neuron from generating an electrical anesthesia. Yet, for unknown reasons, pulse (left). Anesthetics are thought to act by binding to clefts in the GABAA receptor and prolonging the channel opening, which causes some patients experience unexpected hyperpolarization of the cell membrane (right). recall of events during the surgery itself. Thus, by fi nding the correct target re-

ceptors for the amnesia-inducing effects CHRISTINE GRALAPP

of anesthesia, it may become possible to Rather than provoking a response at the on other research questions had also in- identify patients at risk for intraopera- “switchboard,” the drugs were acting to dicated that GABAA receptors contain- tive awareness because they lack those boost a kind of static or inhibitory buzz ing the alpha-5 subunit are involved in receptors. Alternatively, drug strategies in the telephone line itself that interfered normal hippocampal-dependent memo- to prevent awareness or at least its recol- with communication. ry processes, supporting our theory that lection could also be developed. We found that the injectable anes- the extrasynaptic alpha-5 receptors were In the course of this work, we discov- thetics propofol and etomidate, and even responsible for the memory effects of an ered to our surprise that even receptors the inhaled anesthetic isoflurane, in- anesthetic. To test our hypothesis fur- outside the synapse could play a role in creased the amplitude of this current by ther, we turned to experimenting with anesthetic action. If the synapse serves as much as 35-fold at concentrations sev- genetically modiﬁ ed mice that lacked as a switchboard at the junction between eral times lower than those required to the alpha-5 subunit and wild-type mice two cells, then receptors at the synapse cause immobility. Other investigators, that had the normal receptor. As expect- periphery or scattered along the nerve including Stephen G. Brickley, Mark ed, in behavioral tests the wild-type rug strategies to prevent awareness or D at least its recollection could be developed.

cell body can be imagined as residing on Farrant and their colleagues at University mice were sensitive to amnesia-causing the telephone line itself. Such extrasyn- College London, had already described doses of etomidate, whereas the alpha-5- aptic GABAA receptors are activated by this steady low current even in the ab- deficient mice failed to manifest the even the very low concentrations of sence of anesthetics. But what surprised drug’s effects on memory. GABA that are naturally present in the our group was the extrasynaptic recep- We also established that the loss of extracellular space or that spill over tors’ remarkable sensitivity to tiny alpha-5 GABAA receptors had no conse- from nearby synapses. As it turns out, amounts of both inhaled and intrave- quences for any of the other anesthesia high numbers of extrasynaptic receptors nous anesthetics, whereas the low con- end points: sedation, immobility, hypno- are found in certain brain regions, such centrations of anesthetics caused only sis and response to a painful stimulus as the hippocampus and the thalamus negligible changes in postsynaptic cur- were the same in both groups of mice. (an area involved in consciousness and rents. Previous studies, such as our own, These results demonstrated that the pain processing), as well as parts of the had apparently focused on the right fam- memory-impairing effects of etomidate cortex and the cerebellum. ily of receptor proteins but were looking could be dissociated from the drug’s oth- We serendipitously learned the rele- in the wrong location. er properties based on the pharmacology vance of extrasynaptic GABAA recep- Eventually our experiments deter- of specifi c receptor subunits. They also tors as anesthetic targets after struggling mined that the extrasynaptic GABAA re- provided the fi rst animal model for re- unsuccessfully for quite some time to ceptors were structurally slightly differ- ceptor variations that might occur in hu- identify postsynaptic receptors that were ent from the populations of receptors mans and could explain some cases of sensitive to the very low amnesia-induc- within the synapse in that they predomi- resistance to an anesthetic’s ability to in- ing concentrations of anesthetics. We nantly contained an alpha-5 subunit, duce amnesia. Ongoing studies will de- had also searched for populations of which the postsynaptic receptors gener- termine whether other general anesthet- postsynaptic receptors that were syner- ally lacked. That single change seemed to ic drugs also preferentially target alpha-5 gistically modulated by midazolam and account for their sensitivity to even tiny GABAA receptors to produce amnesia. propofol, two of the most commonly amounts of anesthetics. These results At the same time, laboratories in Eu- used intravenous neurodepressive drugs, were exciting to us because accumulating rope and the U.S. have been employing and had not found any of those either. evidence from neuroscientists working similar experimental techniques to ex- Our work, however, was based on taking electrophysiological recordings of BEVERLEY A. ORSER is professor of anesthesiology and physiology at the University of currents generated in hippocampal neu- Toronto and a practicing anesthesiologist at the university’s hospital, the Sunnybrook rons in tissue culture, and we did notice Health Sciences Center, where she is also Canada Research Chair in anesthesia. As both that amnesia-producing concentrations a clinician and researcher, Orser focuses on improving patient safety. By studying the of anesthetics signifi cantly increased a molecular mechanisms underlying anesthetic drugs, she hopes to advance development

persistent low-amplitude current gener- THE AUTHOR of new agents and related therapies with more ﬁ nely controlled effects. Orser also con- ated by extrasynaptic GABAA receptors. sults for the pharmaceutical company Merck, developer of the sleep aid Gaboxadol. www.sciam.com SCIENTIFIC AMERICAN 59 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more..

Risk-Management Tools Breathing circuit Delivers oxygen and air mixed Before surgery with general anesthesia, a with inhaled anesthetics through patient will see and be attached to an array one tube; removes exhaled gases through the other tube of monitoring devices. Many of them are there to guard against side effects of the anesthetics, which depress breathing and heart function, lowering blood pressure and body temperature. An anesthesiologist must constantly calibrate drug delivery to achieve the desired depth of anesthesia without subduing the patient’s cardiac and respiratory functions enough to risk death. Electrocardiograph electrodes Applied to chest and limbs to track heart electrical activity and rate

Warming blanket Not shown: Provides forced hot air to maintain body Temperature probe temperature. Blood and Applied to the skin or intravenous ﬂ uid inserted into the esophagus warmers are also or rectum, depending on commonly used type and length of surgery

Arterial catheter Blood Inserted into artery in wrist pressure or groin. Provides beat-to- cuff beat measurement of blood Pulse oximeter probe pressure and allows Clipped to ﬁ nger or ear to Intravenous line frequent blood sampling measure blood oxygen level Delivers intravenous anesthetics and ﬂ uids

plore the hypnotic and immobilizing ef- turned the old notion that because anes- that etomidate was less effective at pro- fects of anesthetics. Gregg E. Homanics thetics are so chemically different from ducing unconsciousness in the animals; of the University of Pittsburgh School of one another they must produce their the drug’s immobilizing properties re- Medicine, for example, developed a multiple effects by some general mecha- mained, however. Uwe Rudolph, while mouse lacking the GABAA receptor del- nism. Instead empirical development of at the University of Zurich, also gener- ta subunit, which is known to confer anesthetic drugs seems to have stumbled ated transgenic mice with the same mu- high sensitivity to neurosteroids. His on chemicals that produce similar end tation in the beta-3 subunit and found group’s investigation found that the del- points, though each by its own unique that it greatly diminished the effective- ta-deﬁ cient mice were, predictably, less mechanisms. ness of both etomidate and propofol in sensitive to the steroid-based anesthetic Etomidate, for example, is the only producing unconsciousness and analge- alphaxalone in tests of the drug’s power anesthetic in clinical use that is selective sia in the animals. In contrast, he to induce unconsciousness. The mutant for GABAA receptors containing the showed that alphaxalone was equally mice, however, displayed no differences beta-2 or beta-3, but not the beta-1, sub- effective in wild-type mice and those in their responses to propofol, etomi- unit. Indeed, the differences between carrying the mutation, which indicates date and other nonsteroidal anesthetics the beta-subunit variants that respond that these receptor subunits are proba- as compared with wild-type controls. to etomidate and those that do not in- bly not important targets of that drug. Steroid anesthetics are not commonly volve a single amino acid change at a Whether the point mutations in beta- used today, but these results also dem- speciﬁ c point in the subunit’s protein 2 and beta-3 receptor subunits also in- onstrated the principle that different structure. The pharmaceutical compa- fluence the drugs’ amnesia-inducing classes of anesthetics target discrete ny Merck developed transgenic mice properties has not yet been established. subpopulations of GABAA receptors. with a mutation in that amino acid loca- And which central nervous system re-

Such experiments have truly over- tion within the beta-2 subunit and found gions in the transgenic mice are affected AGUILERA-HELLWEG MAX

is now being studied as a sleep-promoting drug. It targets GABAA receptors con- Capnograph monitor taining the delta subunit, primarily Displays measurements of found in the thalamus and cerebellum, exhaled carbon dioxide to ensure ventilation is appropriate and therefore may also avoid affecting memory. The memory-blocking potential of similar compounds that do inter- Gas sampler act with alpha-5 receptors could also Monitors oxygen, carbon dioxide and anesthetic concentrations in prove very useful in the surgical setting, both inhaled and exhaled air where drugs that cause profound amnesia without depressing respiration, air- Vaporizers way refl exes or the cardiovascular system Store and blend the inhaled might be highly desirable. In combina- anesthetics in exact desired concentrations tion with other anesthetics, a potent memory blocker could be used to prevent Anesthesia delivery monitor intraoperative awareness episodes, for example. Alone, such a drug might be helpful in the treatment of patients suf- Anesthesia fl ow controls fering post-traumatic stress disorder by inhibiting certain distressing memories. Management of anesthesia’s effects on memory is just one example of a new Anesthesia delivery unit approach to anesthesiology that will be- Combines inhaled anesthetics come possible with such targeted drugs. with oxygen and room air and controls their fl ow. Alarm system In many situations, the broad and pro- detects disconnection of the found neurodepression of current anes- breathing circuit or changes in levels of inhaled oxygen. Also thetics is unnecessary and undesirable scavenges exhaled carbon anyway. With a cocktail of compounds, dioxide and anesthetic gases so they do not contaminate the each of which produces only one desir- operating room able end point, the future version of anesthesia care could leave a patient conversant but pain-free while having a broken limb repaired or immobile and by the mutations is another unknown, In fact, such compounds are already sedated but aware while having a hip although some evidence suggests that ex- in development for other uses. Sedative- replaced. This polypharmaceutical ap- trasynaptic GABAA receptors in the hypnotic drugs that do not act on the al- proach is already widely used for other thalamus may be critical. Taken together, pha-5 subunit, and hence should lack the aspects of surgery-related care, most however, such studies are confi rming the memory-fuzzing effects of benzodiaze- notably in the treatment of postopera- central role of GABAA receptors in the pine sedatives and certain sleeping pills, tive pain. actions of anesthetics. The next step is are in the preclinical pipeline. And clini- Today anesthesia has never been saf- to begin translating this knowledge cal trials of Gaboxadol, the fi rst drug to er, but it is certainly not without risk. A gained from current general anesthetics selectively target extrasynaptic GABAA tremendous opportunity now exists for into drugs that are anything but general. receptors to enhance their function, are the fi eld to move beyond the ether era currently under way. Gaboxadol was ini- and toward a truly modern model of an- Tailored Treatment tially developed as an anticonvulsant but esthesia care. as the work of my research group and others has shown, extrasynaptic al- MORE TO EXPLORE pha-5 GABAA receptors in the hippo- Anesthesia Safety: Model or Myth? Robert S. Lagasse in Anesthesiology, Vol. 97, campus are vital to the amnesia-induc- pages 1609–1617; December 2002. ing effects of etomidate and possibly to Molecular and Neuronal Substrates for General Anaesthetics. Uwe Rudolph and Bernd other general anesthetics currently in Antkowiak in Nature Reviews Neuroscience, Vol. 5, pages 709–720; September 2004. use. These results suggest that drugs Emerging Molecular Mechanisms of General Anesthetic Action. Hugh C. Hemmings et al. in Trends in Pharmacological Sciences, Vol. 6, No. 10, pages 503–510; October 2005. that avoid or target that particular re- ␣5GABAA Receptors Mediate the Amnestic but Not Sedative-Hypnotic Effects of the General ceptor could selectively spare or block Anesthetic Etomidate. Victor Y. Cheng et al. in Journal of Neuroscience, Vol. 26, No. 14, memory formation as needed. pages 3713–3720; April 5, 2006. www.sciam.com SCIENTIFIC AMERICAN 61 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more..

When Fields Collide The history of particle cosmology, a new branch of physics that has shed light on the origins of the universe, shows that science can sometimes beneﬁ t from wrenching changes

By David Kaiser

article cosmology, which investigates this branch of research did not even exist how the smallest units of matter 30 years ago. Before 1975, particle physics P have determined the shape and fate and cosmology were treated as separate of the universe, is one of the hottest topics fi elds of study (especially in the U.S.), and in physics today. In recent years the fi eld few scientists considered how discoveries has received as much as half a billion dol- in one specialty could enhance research in lars in funding from the U.S. Department the other. of Energy, the National Science Founda- So why did particle cosmology arise? tion and NASA. Scientists have made great During the mid-1970s, researchers real- strides in understanding the high-energy ized that studies of the early universe of- particle interactions that roiled the uni- fered a unique window for investigating verse in the fi rst moments of its history and high-energy phenomena that cannot be re- infl uenced cosmic evolution in the billions created in the laboratory. But a series of 4 FIRST EXPLOSIVE MOMENTS of cosmic of years afterward. changes in the funding and teaching of history have been reconstructed by The dramatic success of particle cos- physics also helped to push cosmological particle cosmologists, who study

mology is all the more striking given that questions to the forefront. The rapid emer- the birth of the universe. FILMS CREDIT SLIM

gence of particle cosmology illustrates The success of space and time, as in ordinary general how government budgets, educational relativity, and to variations in the local institutions and even the publication of particle cosmology strength of gravity [see top illustration textbooks can radically alter the direc- in box on opposite page]. The ␸ fi eld tion of research. The history of that era is all the more permeates all of space, and its behavior also shows that science can reap tremen- helps to determine how matter moves dous benefi ts when researchers move striking given that through space and time. Any measure- away from familiar subjects to tackle this branch of ment of an object’s mass therefore de- new challenges. pends on the local value of ␸. This A good way to tell the story is to fo- research did not theory was so compelling that members cus on the fortunes of two sets of ideas: of Kip Thorne’s gravity group at the the Brans-Dicke field, introduced by even exist California Institute of Technology used gravitational specialists, and the Higgs to joke that they believed in Einstein’s fi eld, puzzled over by particle physicists. 30 years ago. general relativity on Mondays, Wednes- Both groups created these concepts in days and Fridays and in Brans-Dicke response to a problem that exercised though this principle intrigued Einstein gravity on Tuesdays, Thursdays and many scientists during the late 1950s and spurred his thinking, his general Saturdays. (They remained agnostic on and early 1960s: why objects have mass. theory of relativity ultimately departed Sundays.) Although these two theories did not from it. Meanwhile, within the much larger drive the union of particle physics and To incorporate Mach’s principle into community of particle physicists, the cosmology, the course of their develop- gravitational theory, scientists postulat- problem of mass arose in a different ment demonstrates how the two branch- ed the existence of a new scalar fi eld form. Beginning in the 1950s, theorists es of research converged. that interacts with all types of matter. (A found that they could represent the ef- scalar fi eld has one value for each point fects of nuclear forces by imposing spe- A Tale of Two ␸’s in space and time.) In 1961 Princeton cial classes of symmetries on the equa- mass seems like such an obvious University graduate student Carl Brans tions governing the behavior of sub- property of matter that one might not and his thesis adviser, Robert H. Dicke, atomic particles. Yet the terms they think it requires an explanation. Yet pointed out that in Einstein’s general would ordinarily include in these equa- fi nding descriptions of mass that were relativity, the strength of gravity is fi xed tions to represent particle masses vio- compatible with other ideas from mod- by Newton’s constant, G. According to lated the special symmetries. In particu- ern physics proved no easy feat. Experts Einstein, G has the same value on Earth lar, this impasse affected the W and Z on gravitation and cosmology framed as it does in the most distant galaxies bosons—the particles that give rise to the problem in terms of Mach’s princi- and does not change over time. Offering the weak nuclear force, which is respon- ple, named for Austrian physicist and an alternative, Brans and Dicke suggest- sible for radioactive decay. If these force- philosopher Ernst Mach, a famed critic ed that Mach’s principle could be satis- carrying particles were truly massless, of Newton and an inspiration to the fi ed if Newton’s constant varied over as the symmetries seemed to require, young Albert Einstein. A good approxi- time and space. They introduced a fi eld then the range of nuclear forces should mation of Mach’s principle might be called ␸ that was inversely proportional have been infi nite—two protons, for ex- phrased this way: an object’s mass—a to Newton’s constant and swapped 1/␸ ample, should have been able to exert a measure of its resistance to changes in for G throughout Einstein’s gravitation- nuclear force on each other from across its motion—ultimately derives from that al equations. the galaxy. Such a long range fl agrantly object’s gravitational interactions with According to the Brans-Dicke theo- contradicted the observed behavior of all the other matter in the universe. Al- ry, matter responds to the curvature of nuclear forces, which fall off rapidly for distances larger than the size of atomic Overview/A Revolution in Physics nuclei. Only if the force-carrying particles had some mass would the theo- ■ Until the 1970s researchers considered particle physics and cosmology to be retically predicted range come into line completely separate fi elds of study. with observations. ■ Sharp cutbacks in particle physics starting in the late 1960s prompted Many physicists focused on this co- scientists to expand their horizons and explore topics in gravitation nundrum, trying to formulate a theory and cosmology. that would represent the symmetry ■ By the 1980s researchers had found that studying the early universe offered properties of subatomic forces while a new way to explore high-energy phenomena. Since then, the hybrid fi eld also incorporating massive particles. In of particle cosmology has become one of the most fruitful in physics. 1961 Jeffrey Goldstone, then at the Uni- versity of Cambridge, noted that the so-

lutions to the equations need not obey the same symmetries that the equations themselves do. As a simple illustration Separate Concepts of Mass he introduced a scalar fi eld, coinciden- The barriers that divided physicists in the early 1960s are illustrated by their tally labeled ␸, whose potential energy parallel attempts to explain why objects have mass. Although cosmologists and density, V(␸), bottoms out at two points: particle physicists proposed similar theories, few scientists saw the connection. when ␸ has the values of –v and +v [see bottom illustration in box at right]. Be- cause the energy of the system is lowest From Cosmology: Brans-Dicke Gravity at these minima, the fi eld will eventually settle into one of them. The poten- ▼ In 1961 Carl Brans and Robert Dicke of Princeton tial energy is exactly the same for both University proposed a fi eld called ␸ that allows values, but because the fi eld must even- Newton’s gravitational constant to vary over time tually land at just one value—either –v and space. An object at a point in space where the or +v—the solution to the equations constant is small (left) will be less massive—and spontaneously breaks their symmetry. warp the local spacetime less—than an identical In 1964 Peter W. Higgs of the Uni- Carl Brans Robert Dicke object at a point where the constant is large (right). versity of Edinburgh revisited Gold- stone’s work and found that a theory Deformation of spacetime with spontaneous symmetry breaking would allow for the existence of massive particles. The mass arises from interactions between the ␸ fi eld and all types of particles, including those that generate Identical spherical objects the weak nuclear force. The equations governing these interactions, Higgs Local value of demonstrated, obey all the requisite gravitational symmetries. Before ␸ settles into one of constant the minima of its potential energy, the particles skip lightly along, merrily un- ) encumbered. Once ␸ arrives at either +v From Particle Physics: The Higgs Field

Higgs or –v, however, the newly anchored ﬁ eld ( ); AIP EMILIO SEGRÈ SEGRÈ EMILIO AIP ); exerts a drag on anything coupled to ▼ In 1961 Jeffrey Goldstone, then at the University

Dicke it—the subatomic equivalent of being of Cambridge, introduced a fi eld—also called ␸, mired in molasses. In other words, the coincidentally—whose potential energy density V(␸) force-carrying particles (as well as gar- bottoms out at two points, –v and +v. Three years den-variety matter such as electrons) later Peter Higgs of the University of Edinburgh used Edinburgh University start to behave as if they have a nonzero this fi eld to explain mass. Particles were massless at Jeffrey Peter Higgs fi r s t w h e n ␸ varied (left); they acquired mass only ); BETTMANN/CORBIS ( BETTMANN/CORBIS ); mass, and any measurements of their Goldstone ␸ mass depend on the local value of ␸. after settled into one of its minima (right). Brans ( The Brans-Dicke and Higgs papers ); PETER TUFFY TUFFY PETER );

Luno were published at about the same time in the same journal, Physical Review. Goldstone Both articles quickly became well known; to this day, both are among the ); R. CRESSON CRESSON R. ); most cited physics articles of all time. Each proposed to explain the origin of mass by introducing a new scalar ﬁ eld field diagrams field that interacted with all types of matter. Given the similarity of the proposals and the quick attention they both received, one might have expected physicists to consider them alongside each other. Yet this pairing rarely happened. LUCY READING-IKKANDA ( VISUAL ARCHIVES, PHYSICS TODAY COLLECTION ( COLLECTION TODAY PHYSICS ARCHIVES, VISUAL Of the 1,083 articles that cited either

the U.S. when Brans, Dicke, Goldstone and Higgs introduced their respective A Change of Topic ␸’s. The Physics Survey Committee of the National Academy of Sciences, for ▼ Government support for physics boomed in the 1950s and 1960s but fell example, issued a policy report in 1966 sharply in the late 1960s and 1970s. The number of new Ph.D.s plummeted as well that recommended doubling the fund- (top). Many particle physicists, who were hit hardest, shifted their attention to ing and Ph.D.-level personnel for parti- cosmology. Research in that ﬁ eld blossomed (bottom). cle physics over the next few years but 1,800 called for virtually no expansion in the already small areas of gravitation, cos- 1,500 mology and astrophysics. Furthermore, even though some of the Soviet textbooks published on gravity during that 1,200 era included speculations about nuclear forces, such mixing of genres was absent 900 from American textbooks. These research patterns, however, would change radically by the late

Granted in the U.S. in the Granted 600 1970s. Looking back on the swift rise of Number of Physics Ph.D.s Ph.D.s Physics of Number 300 particle cosmology, physicists almost always point to two important developments that spurred the merger: the dis- 0 covery of asymptotic freedom in 1973 500 and the construction of the fi rst grand unifi ed theories, or GUTs, in 1973 and 450 1974. Asymptotic freedom refers to an 400 unexpected phenomenon in certain 350 classes of theories governing particle interactions: the strength of the interac- 300 tion decreases as the energy of the par- 250 ticles goes up, rather than increasing the 200 way most other forces do. For the fi rst time, particle theorists were able to 150 make accurate and reliable calculations Cosmology (worldwide) Cosmology 100 of phenomena such as the strong nucle- Number of Articles Published on on Published Articles of Number 50 ar force—which keeps quarks bound within nuclear particles such as protons 0 and neutrons—as long as they restricted 1955 1960 1965 1970 1975 1980 Year their calculations to very high energy realms, far beyond anything that had been probed experimentally. the Brans-Dicke or Higgs paper be- To the experts in gravitation and cos- The introduction of GUTs likewise tween 1961 and 1981, only six—less mology, the Brans-Dicke fi eld (␸BD) was directed attention toward very high en- than 0.6 percent—included both arti- exciting because it offered an alterna- ergies. Particle physicists realized that cles in their references. (The earliest in- tive to Einstein’s general relativity. To the strengths of three of the fundamen- stance was in 1972, and the rest came the particle physicists, the Higgs fi eld tal forces— electromagnetism and the after 1975.) This mutual ignorance (␸H) was exciting because it offered hope weak and strong nuclear forces—might highlights the stark boundaries that ex- that their theories might be able to ex- converge as particle energies increased. isted at the time between the particle plain the behavior of nuclear forces Theorists hypothesized that once the en- physicists and the specialists in gravita- among massive particles. Before the ergies rose high enough, the three forces tion and cosmology. mid-1970s, nobody suggested that ␸BD would act as a single undifferentiated and ␸H might be physically similar or force. The energy scale at which this Pushes, Pulls and Pedagogy even worth examining side by side. grand unifi cation would set in was liter- clearly, the two communities saw The divide between particle physics ally astronomical: about 1024 electron ␸ different things in their respective ’s. and cosmology was especially sharp in volts, or more than one trillion times LUCY READING-IKKANDA

higher than the top energies physicists Beginning in the rise. Spurred in part by a series of break- had been able to probe using particle ac- throughs during the mid-1960s, includ- celerators. GUT-scale energies could late 1960s, drastic ing the discovery of quasars, pulsars never be achieved in Earth-bound labo- and the cosmic microwave background ratories, but some researchers realized cutbacks wreaked radiation, the number of new Ph.D.s in that if the entire universe had begun in this area grew by 60 percent between a hot big bang, then the average energy havoc on physics 1968 and 1970 and by another 33 per- of particles in the universe would have in the U.S. cent between 1971 and 1976—even as been extraordinarily high during early the total number of physics Ph.D.s fell periods in cosmic history. sharply. With the advent of asymptotic free- Defense spending on basic research, Surveying the wreckage in 1972, the dom and GUTs, particle physicists had wreaked havoc on physics in the U.S. National Academy’s Physics Survey an obvious reason to begin studying the Nearly all fi elds of science and engineer- Committee released a new report that early universe: the fi rst moments of the ing went into decline, but physics fell highlighted the troubles in particle phys- big bang would provide them with “the faster and deeper than any other fi eld. ics. Many young theorists in that fi eld, poor man’s accelerator,” allowing them The number of new physics Ph.D.s the committee noted, were having dif- to observe high-energy interactions that plummeted, falling nearly as fast from fi culty switching their research efforts were impossible to re-create on Earth. 1970 to 1975 as it had risen during the elsewhere because of their “narrow spe- Scores of scientists, journalists, philoso- years after Sputnik. cialization.” The report urged the na- phers and historians have pointed to Federal funding for physics also tion’s physics departments to revamp this development to explain the emer- plunged, dropping by more than one how particle theorists were trained: gence of particle cosmology. third (in constant dollars) between 1967 “University groups have a responsibility But is it the whole story? Although and 1976. From the 1950s to the mid- to expose their most brilliant and able the advances in particle theory were cer- 1960s, the number of available jobs had students to the opportunities in all sub- tainly important, they are not suffi cient always been greater than the number of fi elds of physics.” Changes in university to explain the growth of this new sub- physics students looking for work at the curricula quickly followed, aimed to fi eld. For one thing, the timing is a bit placement service meetings held by the broaden graduate students’ exposure to off. Publications on cosmology (world- American Institute of Physics. But em- other areas of physics—including more wide as well as in the U.S.) began a steep ployment prospects quickly turned emphasis on gravitation and cosmology. rise before 1973, and the rate of increase grim: 989 applicants competed for 253 Across the country, physics programs was completely unaffected by the ap- jobs in 1968, and 1,053 students scram- began to offer new courses on the sub- pearance of the papers on asymptotic bled for 53 positions in 1971. ject. After ignoring gravitation and cos- freedom and GUTs [see box on opposite Particle physics was hardest-hit by mology for decades, American publish- page]. Moreover, GUTs did not receive far, with federal spending on the fi eld ers pumped out scores of textbooks on much attention, even from particle the- falling by 50 percent between 1970 and the topic to meet the sudden demand. orists, until the late 1970s and early 1974. A swift exodus of talent began: 1980s. Three of the earliest review ar- between 1968 and 1970, twice as many Inflating the Ranks ticles on the emerging fi eld of particle U.S. researchers left particle physics as these abrupt changes left their cosmology, published between 1978 entered the fi eld. The number of new mark on the way physicists viewed con- and 1980, ignored asymptotic freedom Ph.D.s in particle physics fell by 44 per- cepts such as the Brans-Dicke and Higgs and GUTs altogether. cent between 1969 and 1975—the fast- fi elds. In 1979, after nearly two decades New ideas alone were not enough to est decline of any branch of physics. At in which virtually no one had even men- pave the way for particle cosmology; the same time, however, the fortunes of tioned the two fi elds in the same paper, governmental and educational changes astrophysics and gravitation began to let alone considered them to be physi- played major roles as well. Until the mid- 1960s, U.S. physicists had benefited DAVID KAISER is both a physicist and a historian. He received Ph.D.s in theoretical phys- from a “cold war bubble,” a period ics and the history of science from Harvard University and is now an associate professor when the federal government lavished in the Program in Science, Technology and Society at the Massachusetts Institute of funds on education, defense and scien- Technology and a lecturer in M.I.T.’s physics department. His recent book, Drawing The- tific research. Beginning in the late ories Apart: The Dispersion of Feynman Diagrams in Postwar Physics (University of Chi-

1960s, though, drastic cutbacks trig- THE AUTHOR cago Press, 2005), traces how Richard Feynman’s idiosyncratic approach to quantum gered by anti–Vietnam War protests, a physics entered the mainstream. He is completing a new book on physics during the thawing of the cold war and the intro- cold war, looking in particular at changes in the training of graduate students. His cur- duction of the Mansﬁ eld Amendment, rent physics research focuses on particle cosmology, working on ways that cosmic which heavily restricted Department of inﬂ ation might be made compatible with superstring-inspired large extra dimensions. www.sciam.com SCIENTIFIC AMERICAN 67 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more..

In this model, the local strength of gravity initially varied over space and Making the Connection time, with G proportional to 1/␸2, but its present-day constant value emerged af- By the late 1970s a new generation of physicists, conversant with both particle ter the ␸ fi eld settled into a minimum of theory and cosmology, explored possible links between Brans-Dicke gravity its symmetry-breaking potential, which and the Higgs fi eld. presumably occurred in the first mo- 3ANTHONY ZEE ments of the big bang. In this way, Zee As an undergraduate, Zee worked with gravitation expert and Smolin offered an explanation of John Wheeler at Princeton University, then pursued a Ph.D. why the gravitational force is so weak in particle theory. He renewed his interest in cosmology compared with other forces: when the while on sabbatical in Paris in 1974. fi eld settled into its fi nal state, ␸ = ±v, it anchored ␸ to some large, nonzero val- 3 LEE SMOLIN ue, pushing G (which is inversely pro- In the 1970s Smolin studied particle theory and portional to v2) to a small value. cosmology as a graduate student at Harvard University. He also worked with Stanley Deser of Brandeis University, The career paths of Zee and Smolin one of the pioneers of quantum gravity. illustrate the ways in which physicists focused their attention on cosmology after 3ALAN GUTH the collapse of the cold war bubble. Zee Guth earned his Ph.D. in particle physics from the had worked with gravitation expert Massachusetts Institute of Technology in 1972. He John A. Wheeler as an undergraduate at became interested in cosmology after attending a Princeton in the mid-1960s before pur- lecture by Dicke in the late 1970s. suing his Ph.D. in particle theory at Har- vard. He earned his degree in 1970, at Cosmic microwave Dark energy the same time as the biggest declines in background Dark ages accelerated expansion that area began. As he later recalled, cos- Infl ation mology had never even been mentioned while he was in graduate school. After postdoctoral work, Zee began teaching at Princeton. He rented an apartment from a French physicist while on sabbatical in Paris in 1974, and in his borrowed quarters he stumbled on a stack of papers by European theorists that tried to use ideas from particle theory to explain various cosmological features (such as why the observable universe contains

First stars more matter than anti matter). Although ); ) he found the par ticular ideas in the pa- Zee Guth

13.7 billion years pers unconvincing, the chance encounter ( reignited Zee’s earlier interest in gravitation. Returning from his sabbatical and ▲ In separate papers published in 1979, Zee and Smolin combined the Brans- back in touch with Wheeler, Zee began Dicke gravitational equations with a Goldstone-Higgs symmetry-breaking to redirect his research interests toward potential. In 1981 Guth introduced another fi eld, based on the Higgs, called the particle cosmology. infl aton. This fi eld provided the driving force behind a postulated period of

Lee Smolin, in contrast, entered grad- Researchers, Inc. Photo superfast expansion—or inﬂ ation—during the universe’s ﬁ rst moments. ( ZEE ANTHONY OF COURTESY ); uate school at Harvard in 1975, just as the curricular changes began to take ef- big bang big cally similar, two American theorists combining the Brans-Dicke gravitation- fect. Smolin studied gravitation and cos-

independently suggested that ␸BD and al equations with a Goldstone-Higgs mology there alongside his course work PARKER DAVID ); ␸H might be one and the same ﬁ eld. In symmetry-breaking potential. (Other in particle theory and worked closely separate papers, Anthony Zee, then at theorists, working outside the U.S., had with Stanley Deser (based at nearby Smolin the University of Pennsylvania, and Lee tentatively broached similar ideas be- Brandeis University), who was visiting Smolin, then at Harvard University, tween 1974 and 1978, but they received Harvard’s physics department at the ␸ DINA GRASER ( glued the two key pieces of together by little attention at the time.) time. Deser was one of the few American ( TEAM NASA/WMAP SCIENCE

theorists who had taken an interest in Perhaps today’s drei Linde, then at the Lebedev Physical quantum gravity in the 1960s, attempt- Institute in Moscow, was likewise poised ing to formulate a description of gravita- passionate debates to explore inflationary ideas: having tion that would be compatible with quan- studied in Russia, where particle phys- tum mechanics. He was also the very fi rst on the direction ics and gravitation had long fl ourished physicist to publish an article that cited side by side, Linde was quick to intro- both the Brans-Dicke work and the of theoretical duce improvements to the theory. Higgs work (although he treated the two physics are Since then, it has become routine for fi elds rather differently and in separate particle cosmologists to combine the parts of his 1972 paper). Smolin, who a symptom of the Brans-Dicke, Higgs and infl aton fi elds, worked on topics in quantum gravity, freely adapting the equations to explain suggested that ␸BD and ␸H might be the same kinds a variety of phenomena. This concep- same fi eld as he was fi nishing his disser- tual leap moved from unthinkable to tation in 1979. of growing pains unnoticeable in only a few academic Smolin’s experiences marked the that shook the generations. The shift in attitude illus- new routine for his generation of theo- trates the power of pedagogy and the rists, trained during the mid- to late discipline after immense influence that institutional 1970s. Physicists such as Paul J. Stein- changes can have on scientifi c thought. hardt, Michael S. Turner and Edward the last crash. Might history repeat itself? Particle “Rocky” Kolb studied gravitation as physics was hit hard again in the 1990s well as particle theory in graduate Guth’s career path was similar to (especially with the cancellation of the school. Soon Smolin, Turner, Kolb, Zee’s; he completed his Ph.D. in particle Superconducting Super Collider, a huge Steinhardt and others were training theory at M.I.T. in 1972, before the particle accelerator that was under con- their own graduate students to work in widespread curricular reforms that struction in Texas), and funding in the the new hybrid area of particle cosmol- brought gravitation back into American U.S. has continued to slide since then. ogy. For these young theorists and their classrooms. Hit hard by the collapse of Perhaps today’s passionate debates on growing numbers of students, it was particle physics, Guth toiled in a series the direction of theoretical physics, pit- natural to associate ␸BD and ␸H. Turner, of postdoctoral positions for several ting the advocates of string theory Kolb and Steinhardt each led research years. By chance he attended a lecture against the proponents of alternative groups that pursued further links be- by Dicke on the fl atness problem in the approaches, are a symptom of the same tween ␸BD and ␸H during the 1980s. late 1970s, which planted the idea in kinds of growing pains that shook the Building on his 1979 paper, Zee not- Guth’s head that cosmology might prove discipline after the last crash. ed in 1980 that standard cosmological interesting for thinking about puzzles in Physicists are now looking forward theories, such as the big bang model, re- particle theory. While immersed in the to new results from projects scheduled mained unable to account for the ex- new physics of GUTs and working hard to come online over the next year: the traordinary smoothness of the observ- to retool himself with some basic back- Large Hadron Collider in Switzerland, able universe (at least when viewed on ground in gravitation and cosmology, the Gamma-Ray Large Area Space Tele- the largest scales). Separately, Dicke con- he hit on infl ation. Most of the physi- scope and the Planck satellite, which cluded that the big bang also could not cists who advanced the idea, however, will measure the cosmic microwave explain the observed fl atness of the uni- were younger theorists—people such as background with unprecedented accu- verse, whose shape could in principle Steinhardt, Kolb and Turner and their racy. With any luck, high-energy phys- depart quite far from the minimal cur- students—who had been pedagogically ics will emerge just as vibrant and excit- vature that astronomers observed. In primed for just such a development. An- ing as it did 30 years ago. 1981 Alan H. Guth—then a postdoctoral fellow at Stanford University and now MORE TO EXPLORE a professor at the Massachusetts Insti- Was Einstein Right? Putting General Relativity to the Test. Second edition. Clifford M. Will. tute of Technology—introduced infl a- Basic Books, 1993. tionary cosmology to address both of The Inflationary Universe: The Quest for a New Theory of Cosmic Origins. Alan H. Guth. these problems. At the heart of Guth’s Addison-Wesley, 1997. model was another scalar fi eld, modeled Einstein’s Universe: Gravity at Work and Play. Anthony Zee. Oxford University Press, 2001. on the Higgs. Dubbed the infl aton, this Three Roads to Quantum Gravity. Lee Smolin. Basic Books, 2001. fi eld provided the driving force behind a Cold War Requisitions, Scientific Manpower, and the Production of American Physicists after World War II. David Kaiser in Historical Studies in the Physical and Biological Sciences, Vol. 33, postulated period of superfast expan- pages 131–159; 2002. sion—or infl ation—during the universe’s Inflationary Cosmology: Exploring the Universe from the Smallest to the Largest Scales. fi rst moments. Alan H. Guth and David Kaiser in Science, Vol. 307, pages 884–890; February 11, 2005. www.sciam.com SCIENTIFIC AMERICAN 69 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. RestoringVisit http://avaxhome.ru/blogs/vietmagz America’s for more..

n the fall of 2004 a dozen conservation biologists gath- Although the cheetahs, lions and mammoths that once ered on a ranch in New Mexico to ponder a bold plan. roamed North America are extinct, the same species or close The scientists, trained in a variety of disciplines, ranged relatives have survived elsewhere, and our discussions fo- from the grand old men of the fi eld to those of us earlier cused on introducing these substitutes to North American I in our careers. The idea we were mulling over was the reintro- ecosystems. We believe that these efforts hold the potential to duction of large vertebrates—megafauna—to North America. partially restore important ecological processes, such as pre- Most of these animals, such as mammoths and cheetahs, dation and browsing, to ecosystems where they have been died out roughly 13,000 years ago, when humans from Eur- absent for millennia. The substitutes would also bring eco- asia began migrating to the continent. The theory—propound- nomic and cultural benefi ts. Not surprisingly, the published ed 40 years ago by Paul Martin of the University of Arizona— proposal evoked strong reactions. Those reactions are wel- is that overhunting by the new arrivals reduced the numbers come, because debate about the conservation issues that un- of large vertebrates so severely that the populations could not derlie Pleistocene rewilding merit thorough discussion. recover. Called Pleistocene overkill, the concept was highly controversial at the time, but the general thesis that humans Why Big Animals Are Important played a signifi cant role is now widely accepted. Martin was our approach concentrates on large animals because present at the meeting in New Mexico, and his ideas on the they exercise a disproportionate effect on the environment. loss of these animals, the ecological consequences, and what For tens of millions of years, megafauna dominated the globe, we should do about it formed the foundation of the proposal strongly interacting and co-evolving with other species and that emerged, which we dubbed Pleistocene rewilding. infl uencing entire ecosystems. Horses, camels, lions, ele-

Pleistocene rewilding—a proposal to bring back animals that disappeared from North America 13,000 years ago—offers an optimistic agenda for 21st-century conservation

By C. Josh Donlan

phants and other large creatures were Questions from Readers everywhere: megafauna were the norm. A summary of this article on our Web site invited readers to But starting roughly 50,000 years ago, ask questions about rewilding. Many of the questions helped to the overwhelming majority went ex- shape the article, and answers appear throughout the text. tinct. Today megafauna inhabit less The author replies to a few others here. than 10 percent of the globe. Over the past decade, ecologist John Won’t African elephants and lions freeze to death in the Montana Terborgh of Duke University has ob- winter? Will these animals be kept indoors in served directly how critical large animals the winter, as they are in zoos? —Jason Raschka are to the health of ecosystems and how The reader brings up an important point—that many their loss adversely affects the natural questions, including climate suitability, would have world. When a hydroelectric dam fl ood- to be answered by sound scientifi c studies during the ed thousands of acres in Venezuela, Ter- process of Pleistocene rewilding. One could imagine borgh saw the water create dozens of is- scenarios in which animals would be free-living (albeit lands—a fragmentation akin to the vir- intensively managed) year-round in expansive reserves in the southwestern U.S. and perhaps other tual islands created around the world as scenarios in locales farther north, where animals humans cut down trees, build shopping would be housed indoors during the coldest months. ELEPHANT at the Copenhagen zoo. malls, and sprawl from urban centers. Even the latter alternative is arguably better than a The islands in Venezuela were too small zoo. Many zoos are struggling to fi nd appropriate space for elephants and are to support the creatures at the top of the choosing to abandon their elephant programs. food chain—predators such as jaguars, pumas and eagles. Their disappearance Aren’t there big-game hunting ranches in Texas? Could we learn sparked a chain of reactions. Animals anything from them? —Foster such as monkeys, leaf-cutter ants and There are many big-game hunting ranches throughout Texas, and some of them other herbivores, whose populations hold animals such as cheetahs that we know, from the fossil record, once lived were no longer kept in check by predation, in North America. To my knowledge, no conservation biologists have studied the ranch-held animals, but these ranches might present excellent research thrived and subsequently destroyed veg- opportunities if they are willing to collaborate. Other ranches, however, have etation—the ecosystems collapsed, with animals that are not supported by the fossil record and as such offer no biodiversity being the ultimate loser. potential for study. Similar ecological disasters have occurred on other continents. Degraded Hasn’t this general idea been around for a while? —Kevin N. ecosystems are not only bad for biodi- Both the concept of rewilding and the term itself have been around for some time. versity; they are bad for human econo- Rewilding is the practice of reintroducing species to places from which they have mies. In Central America, for instance, been extirpated in the past few hundred years. Pleistocene rewilding, in contrast, involves introducing species descended from creatures that went extinct some researchers have shown that intact trop- 13,000 years ago or using similar species as proxies. ical ecosystems are worth at least $60,000 a year to a single coffee farm because of the services they provide, ecosystems by keeping herbivores that habitat in which certain plants and ani- such as the pollination of coffee crops. eat kelp, such as sea urchins, in check. In mals can fl ourish. Lions and other preda- Where large predators and herbivores Africa, elephants are keystone players; as tors control the populations of African still remain, they play pivotal roles. In they move through an area, their knock- herbivores, which in turn infl uence the Alaska, sea otters maintain kelp forest ing down trees and trampling create a distribution of plants and soil nutrients. In Pleistocene America, large preda- Rewilding Our Vision tors and herbivores played similar roles. Overview/ Today most of that vital infl uence is ab- Peter Arnold, Inc. Arnold, Peter ■ A group of conservationists has proposed reintroducing to North America sent. For example, the American chee- large animals that went extinct 13,000 years ago. tah (a relative of the African cheetah) ); DONA DONA ); ■ Close relatives of these animals—elephants, camels, lions, cheetahs—survived dashed across the grasslands in pursuit elsewhere; returning them to America would reestablish key ecological of pronghorn antelopes for millions of processes that once thrived there, as well as providing a refuge for endangered years. These chases shaped the prong- species from Africa and Asia and creating opportunities for ecotourism. horn’s astounding speed and other bio- preceding pages preceding ■ The proposal has, understandably, evoked strong reactions, but the bold logical aspects of one of the fastest ani- suggestion has spurred debate and put a positive spin on conservation mals alive. In the absence of the cheetah, biology, whose role has been mainly a struggle to slow the loss of biodiversity. the pronghorn appears “overbuilt” for

its environment today. ( FELDER LARRY

Pleistocene rewilding is not about re- When the West Was Really Wild creating exactly some past state. Rather Soon after humans crossed the Bering land bridge into North America some it is about restoring the kinds of species 13,000 years ago, almost 75 percent of the continent’s large mammals (those interactions that sustain thriving eco- weighing more than 45 kilograms) disappeared (color). One of the goals of systems. Giant tortoises, horses, camels, Pleistocene rewilding is to restore some of these species or close proxies to the cheetahs, elephants and lions: they were American West. For example, the same species of lion and cheetah that once lived in North America survive today in Africa; the African or Asian elephant all here, and they helped to shape North could substitute for the extinct mammoth; and Bactrian camels might stand American ecosystems. Either the same in for the extinct Camelops. species or closely related species are available for introduction as proxies, Large Mammals of Pleistocene North America and many are already in captivity in the U.S. In essence, Pleistocene rewilding Xenarthra would help change the underlying prem- Glyptodont (Glyptotherium fl oridanum) ise of conservation biology from limit- Harlan’s ground sloth (Paramylodon harlani) ing extinction to actively restoring natu- Jefferson’s ground sloth (Megalonyx jeffersonii) Shasta ground sloth (Nothrotheriops shastensis) ral processes. At fi rst, our proposal may seem outra- Carnivores (Carnivora) geous—lions in Montana? But the plan Dire wolf (Canis dirus) deserves serious debate for several rea- Gray wolf (Canis lupus) sons. First, nowhere on Earth is pristine, Black bear (Ursus americanus) at least in terms of being substantially Brown bear (Ursus arctos) free of human infl uence. Our demograph- Giant short-faced bear (Arctodus simus) ics, chemicals, economics and politics Saber-toothed cat (Smilodon fatalis) pervade every part of the planet. Even in American lion (Panthera leo) our largest national parks, species go ex- Jaguar (Panthera onca) tinct without active intervention. And American cheetah (Miracinonyx trumani) Mountain lion (Puma concolor) human encroachment shows alarming signs of worsening. Bold actions, rather Elephants (Proboscidea) than business as usual, will be needed to American mastodon (Mammut americanum) reverse such negative infl uences. Second, Columbian mammoth (Mammuthus columbi) since conservation biology emerged as a Dwarf mammoth (Mammuthus exilis) discipline more than three decades ago, Woolly mammoth (Mammuthus primigenius) it has been mainly a business of doom and gloom, a struggle merely to slow the Horses (Perissodactyla) loss of biodiversity. But conservation Mexican horse (Equus conversidens) need not be only reactive. A proactive ap- Western horse (Equus occidentalis) proach would include restoring natural Other extinct horses and asses (Equus spp.) processes, starting with ones we know Even-Toed Ungulates (Artiodactyla) are disproportionately important, such Western camel (Camelops hesternus) as those infl uenced by megafauna. Long-legged llama (Hemiauchenia macrocephala) Third, land in North America is Long-nosed peccary (Mylohyus nasutus) available for the reintroduction of mega- Flat-headed peccary (Platygonus compressus) fauna. Although the patterns of human Mule deer (Odocoileus hemionus) land use are always shifting, in some ar- White-tailed deer (Odocoileus virginianus) eas, such as parts of the Great Plains Mountain deer (Navahoceros fricki) and the Southwest, large private and Woodland caribou (Rangifer tarandus) public lands with low or declining hu- Moose (Alces alces) man population densities might be used Wapiti (Cervus elaphus) Pronghorn (Antilocapra americana) for the project. Fourth, bringing mega- Harrington’s mountain goat (Oreamnos harringtoni) fauna back to America would also bring Mountain goat (Oreamnos americanus) tourist and other dollars into nearby Bighorn sheep (Ovis canadensis) communities and enhance the public’s Shrub ox (Euceratherium collinum) appreciation of the natural world. More Bonnet-headed musk ox (Bootherium bombifrons) than 1.5 million people visit San Diego’s Bison (Bison bison)

SHAWN GOULD GOULD SHAWN Wild Animal Park every year to catch a Extinct bison (Bison spp.)

Past, Present and Future glimpse of large mammals. Only a vate lands and could begin immediately. These artist’s conceptions illustrate some handful of U.S. national parks receive They will be critical in answering the of the ways that large animals infl uence the that many visitors. Last, the loss of some many questions about the reintroduc- ecosystem and how Pleistocene rewilding of the remaining species of megafauna tions and would help lay out the costs could be benefi cial. David Burney of the National Tropical Botanical Garden, who is a in Africa and Asia within this century and benefi ts of rewilding. member of the Pleistocene rewilding group, seems likely— Pleistocene rewilding One of these experiments is already worked with artist Larry Felder in conceiving could help reverse that. under way. Spurred by our 2004 meeting, the illustrations. biologists recently reintroduced Bolson How It Might Be Done tortoises to a private ranch in New Mex- we are not talking about backing ico. Bolson tortoises, some weighing up a van and kicking some cheetahs out more than 100 pounds, once grazed parts LATE APRIL DURING THE into your backyard. Nor are we talking of the southwestern U.S. before disap- PLEISTOCENE, 14,000 YEARS AGO about doing it tomorrow. We conceive of pearing around 10,000 years ago, vic- Humans have not yet appeared on the landscape, but glaciers have receded and the Pleistocene rewilding as a series of staged, tims of human hunting. This endangered climate of what is now the western U.S. is carefully managed ecosystem manipula- tortoise now clings to survival, restricted similar to modern conditions, though a little tions. What we are offering here is a vi- to a single small area in central Mexico. cooler. Dense snow covers the Rocky Mountains. The river, fl ushed by snowmelt, is sion—not a blueprint—of how this might Thus, the reintroduction not only repatri- fairly high and a milky blue because of the fi ne be accomplished. And by no means are ates the tortoise to the U.S., it increases particles carried from the young, postglacial we suggesting that rewilding should be the species’ chance for survival. Similar soils. The banks are grazed down short but are generally green. A Columbian mammoth, a a priority over current conservation pro- experiments are also occurring outside Camelops, a Shasta ground sloth, long-horned grams in North America or Africa. Pleis- North America [see box on page 77]. bison, horses and pronghorn antelope tocene rewilding could proceed along- The reintroduction of wild horses browse, while an American lion and two wolves watch from the opposite bank and a side such conservation efforts, and it and camels would be a logical part of bald eagle glides over the river. would likely generate conservation dol- these early experiments. Horses and lars from new funding sources, rather camels originated on this continent, and than competing for funds with existing many species were present in the late conservation efforts. Pleistocene. Today’s feral horses and The long-term vision includes a vast, asses that live in some areas throughout securely fenced ecological history park, the West are plausible substitutes for ex- APRIL 2007 encompassing thousands of square tinct American species. Because most of The Rocky Mountains have snow on their miles, where horses, camels, elephants the surviving Eurasian and African spe- upper summits only. The river, muddied by spring rains and by too many cattle on the and large carnivores would roam. As cies are now critically endangered [see banks, has widened, and the banks are happens now in Africa and regions sur- “Endangered Wild Equids,” by Patricia eroded and broken down by hoofprints and rounding some U.S. national parks, the D. Moehlman; Scientifi c American, cattle wallows. Grass is short, patches of soil are bare, thorny shrubs predominate. Bison ecological history park would not only March 2005], establishing Asian asses graze in the background, and a wary coyote attract ecotourists but would also pro- and Przewalski’s horse in North Amer- stands on the opposite shore. vide jobs related both to park manage- ica might help prevent the extinction of ment and to tourism. these animals. Bactrian camels, which To get to that distant point, we are critically endangered in the Gobi would need to start modestly, with rela- Desert, could provide a modern proxy tively small-scale experiments that as- for Camelops, a late Pleistocene camel. sess the impacts of megafauna on North Camels, introduced from captive or do- APRIL 2027 American landscapes. These controlled mesticated populations, might benefi t Vegetation is more summerlike—global experiments, guided by sound science U.S. ecosystems by browsing on woody warming conditions are in full swing. The mountains have no snow on the peaks. and by the fossil record, which indicates plants that today are overtaking arid Drought, low humidity and absence of what animals actually lived here, could grasslands in the Southwest, an ecosys- mountain snow have lowered the level of the occur fi rst on donated or purchased pri- tem that is increasingly endangered. river dramatically; the water is clear and dark. Erosion is less apparent; grass looks manicured from diversifi ed grazing and C. JOSH DONLAN holds a Ph.D. in ecology and evolutionary biology from Cornell University, browsing. An Asian eletphant, Bactrian where he is a research biologist. Founder and director of Advanced Conservation Strategies, camel, bison, antelope and Przewalski’s horses graze under the watchful eye of an he serves as an adviser to the Galápagos National Park and to Island Conservation and is a African lion; a cheetah approaches from the senior fellow at the Robert and Patricia Switzer Foundation and Environmental Leadership distance. In the background a monorail Program. He was highlighted in the New York Times Magazine’s “Big Ideas of 2005” issue and carries tourists. A high-tech electrifi ed fence with solar panels is out of view. THE AUTHOR named one of 25 “all-star” innovators for 2005 by Outside magazine. He spends much of

LARRY FELDER FELDER LARRY his time in Tasmania, Australia and Santa Cruz, Calif., trying to halt extinctions on islands.

Another prong of the project would pronghorn, facilitating ecotourism as A real-world example of how the re- likely be more controversial but could an economic alternative for ranchers, introduction of a top predator might also begin immediately. It would estab- many of whom are struggling fi nancial- work comes from the wolves of Yellow- lish small numbers of elephants, chee- ly, and helping to save the world’s fastest stone National Park [see “Lessons from tahs and lions on private property. carnivore from extinction. the Wolf,” by Jim Robbins; Scientifi c Introducing elephants could prove Lions are increasingly threatened, American, June 2004]. The gray wolf valuable to nearby human populations with populations in Asia and some parts became extinct in and around Yellow- by attracting tourists and maintaining of Africa critically endangered. Bring- stone during the 1920s. The loss led to grasslands useful to ranchers (elephants ing back lions, which are the same spe- increases in their prey—moose and elk— could suppress the woody plants that cies that once lived in North America, which in turn reduced the distribution threaten southwestern grasslands). In presents daunting challenges as well as of aspens and other trees they eat. Lack the late Pleistocene, at least four ele- many potential benefi ts. But private re- of vegetation destroyed habitat for mi- phant species lived in North America. serves in southern Africa where lions gratory birds and for beavers. Thus, the Under a scientifi c framework, captive el- and other large animals have been suc- disappearance of the wolves propagated ephants in the U.S. could be introduced cessfully reintroduced offer a model— a trophic cascade from predators to her- as proxies for these extinct animals. The and these reserves are smaller than some bivores to plants to birds and beavers. biggest cost involved would be fencing, private ranches in the Southwest. Scientists have started to document the which has helped reduce conflict be- If these early experiments with large ecosystem changes as reintroduced tween elephants and humans in Africa. herbivores and predators show promis- wolves regain the ecological role they Many cheetahs are already in captiv- ing results, more could be undertaken, played in Yellowstone for millennia. An ity in the U.S. The greatest challenge moving toward the long-term goal of a additional insight researchers are learn- would be to provide them with large, se- huge ecological history park. What we ing from putting wolves back into Yel- curely fenced areas that have appropri- need now are panels of experts who, for lowstone is that they may be helping the ate habitat and prey animals. Offsetting each species, could assess, advise and park cope with climate change. As win- these costs are benefi ts—restoring what cautiously lead efforts in restoring ters grow milder, fewer elk die, which must have been strong interactions with megafauna to North America. means less carrion for scavengers such What the Critics Say Since we proposed the rewilding idea in print, in Nature in 2005, the U.S. and Canada where DDT had wiped it out. Those birds were commentators have pointed out a number of concerns, some certainly different genetically from the ones that once soared legitimate, some not. “We all remember ‘Jurassic Park,’” wrote over the Midwest, yet they have done well in their new homes. Dustin Rubenstein and his colleagues in the journal Biological “One can only imagine ... farmers coping with crop destruction Conservation. “Pleistocene re-wilding of North America is by herds of elephants, or lions and cheetah attacking cattle, or only a slightly less sensational proposal.” We disagree even children,” Rubenstein and his colleagues warn. One with this assessment because the majority of need not merely imagine the challenges of coexisting dinosaurs went extinct 65 million years ago, whereas with large predators and herbivores. Africa and Asia many of North America’s original megafauna or very have been struggling with them for centuries, and close relatives are alive today elsewhere in the world substantial progress has been made; in our plan the and can be both studied and saved. animals will not be unrestrained. Rubenstein and colleagues go on to say, “Modern “Global climate change since the Pleistocene day proxies species are wrong ... different genetically extinctions makes the restoration of vanished from the species that occurred in North America during the ecosystems through large-mammal introduction quite Pleistocene.” True, but not that different. Available evidence unlikely,” wrote Christopher Smith in a letter to Nature in the weeks indicates that the lions in Africa and Asia today are the same following the initial proposal. Many have expressed concern about species, albeit of smaller stature, as the lions that prowled North the fact that North America’s ecosystems are not the same today as America 13 millennia ago. Recent studies of ancient DNA have they were 13,000 years ago and that reintroduced animals might elucidated close relationships between extinct elephant and therefore be unable to survive in them. Whereas habitats are and horse species and those alive today. Further, introduction of will continue to be dynamic on a timescale of thousands of years, proxies for now extinct species has proved successful in other very few plants or small mammals went extinct during the late experiments. Hundreds of peregrine falcons from Australia, Pleistocene. The major missing component of North American Europe and South America were used, for example, in captive- ecosystems today compared with the Pleistocene is the megafauna,

breeding programs to reintroduce the peregrine falcon to parts of which we can infer are critical cogs in the wheels. —C.J.D. CHOWDHURY TINA

76 SCIENTIFIC AMERICAN JUNE 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more.. Elsewhere in the World as coyotes, ravens and bald eagles. Wolves provide carcasses throughout In other parts of the world, preliminary efforts have begun to reintroduce the winter for the scavengers to feed on, species to places from which they have long been absent. bestowing a certain degree of stability. ■ In April 2006 a team of Canadian and Russian biologists flew 30 wood bison from Canada’s Elk Island National Park to Pleistocene Park reserve in the The Challenges Ahead Republic of Sakha, Russia, where the closely related steppe bison vanished as our group on the ranch in New 5,000 years ago. Mexico discussed how Pleistocene rewil- ■ At the 15,000-acre nature reserve Oostvaardersplassen in the Netherlands, ding might work, we foresaw many chal- conservationists are restoring horses, roe deer and Heck cattle. lenges that would have to be addressed ■ Beavers are being reintroduced throughout Europe, in some cases in areas and overcome. These include the possibil- where they have been absent for thousands of years. ity that introduced animals could bring ■ In the tropical Pacific, endangered birds from the Marquesas and Tongan islands have been reintroduced to nearby islands where fossils indicate novel diseases with them or that they they once lived. might be unusually susceptible to diseas- ■ In the Indian Ocean, scientists from the Mauritian Wildlife Foundation are es already present in the ecosystem; the using giant tortoises from Aldabra Island to replace two extinct species of fact that habitats have changed over the tortoise on the Mascarene Islands. Scientists have already documented millennia and that reintroduced animals increased seed dispersal for many of the island plants. The tortoises have might not fare well in these altered envi- also brought increased tourism. —C.J.D. ronments; and the likelihood of unantic- ipated ecological consequences and unexpected reactions from neighboring human communities. Establishing programs that monitor the interactions among species and their con sequences for the well-being of the ecosystem will require patience and expertise. And, of course, it will not be easy to convince the public to accept predation as an important natural process that actually nour- ishes the land and enables ecosystems to thrive. Other colleagues have raised additional concerns, albeit none that seems fatal [see box on opposite page]. KONIK HORSES stand in for the extinct tarpan horse at the Oostvaardersplassen Many people will claim that the con- reserve in the Netherlands. cept of Pleistocene rewilding is simply not feasible in the world we live in today. to go back to the past; it is about using tion carries risks as well. In the face of I urge these people to look to Africa for the past to inform society about how to tremendous uncertainty, science and so- inspiration. The year after the creation maintain the functional fabric of nature. ciety must weigh the costs and benefi ts of Kruger National Park was announced, The potential scientifi c, conservation of bold, aggressive actions like Pleisto- the site was hardly the celebrated main- and cultural benefi ts of restoring mega- cene rewilding against those of business stay of southern African biodiversity it fauna are clear, as are the costs. Al- as usual, which has risks, uncertainties is today. In 1903 zero elephants, nine though sound science can help mitigate and costs that are often unacknowl- lions, eight buffalo and very few chee- the potential costs, these ideas will edged. We have a tendency to think that tahs lived within its boundaries. Thanks make many uneasy. Yet given the appar- if we maintain the status quo, things

to the vision and dedication of African ent dysfunction of North American eco- will be ﬁ ne. All the available informa- conservationists, 7,300 elephants, 2,300 systems and Earth’s overall state, inac- tion suggests the opposite. lions, 28,000 buffalo and 250 cheetahs roamed Kruger 100 years later—as did MORE TO EXPLORE 700,000 tourists, bringing with them The Ghosts of Evolution: Nonsensical Fruit, Missing Partners, and Other Ecological tens of millions of dollars. Anachronisms. Connie Barlow. Basic Books, 2000. In the coming century, humanity will Twilight of the Mammoths: Ice Age Extinctions and the Rewilding of America. Paul S. Martin. Foto Natura/Minden Pictures Natura/Minden Foto decide, by default or design, the extent University of California Press, 2005. to which it will tolerate other species and Pleistocene Rewilding: An Optimistic Agenda for Twenty-First Century Conservation. C. J. Donlan, J. Berger, C. E. Bock, J. H. Bock, D. A. Burney, J. A. Estes, D. Foreman, P. S. Martin, thus how much biodiversity will endure. G. W. Roemer, F. A. Smith, M. E. Soulé and H. W. Greene in American Naturalist, Vol. 168, No. 5,

STEVEN RUITER RUITER STEVEN Pleistocene rewilding is not about trying pages 660–681; November 2006.

Breaking Network Logjams

An approach called network coding could By Michelle Effros, dramatically enhance the efﬁ ciency and Ralf Koetter and reliability of communications networks. At its Muriel Médard core is the strange notion that transmitting evidence about messages can be more useful than conveying the messages themselves MATT VINCENT MATT

USERS SHARING A NETWORK based on network coding would enjoy many beneﬁ ts, including seeing fewer delays when downloading videos and accessing Web sites.

The history of modern communications systems has been marked by fl ashes of startling insight. Claude E. Shannon, mathematician and engineer, launched The challenge, then, is to fi nd ways to make the sharing go one such revolution almost 60 years ago by laying the founda- smoothly; parents of toddlers will recognize the problem. Net- tion of a new mathematical theory of communications —now work operators frequently try to solve the challenge by increas- known as information theory. Practical outgrowths of his ing resources, but that strategy is often insuffi cient. Copper work, which dealt with the compression and reliable trans- wires, cables or fi ber optics, for instance, can now provide mission of data, can be seen today in the Internet, in landline high bandwidth for commercial and residential users yet are and wireless telephone systems, and in storage devices, from expensive to lay and diffi cult to modify and expand. Ultra- hard drives to CDs, DVDs and fl ash memory sticks. wideband and multiple-antenna transmission systems can ex- Shannon tackled communications over phone lines dedi- pand the number of customers served by wireless networks but cated to individual calls. These days, information increas- may still fail to meet ever increasing demand. ingly travels over shared networks (such as the Internet), in Techniques for improving effi ciency are therefore needed which multiple users simultaneously communicate through as well. On the Internet and other shared networks, informa- the same medium—be it a cable, an optical fi ber or, in a wire- tion currently gets relayed by routers—switches that operate less system, air. Shared networks can potentially improve the at nodes where signaling pathways, or links, intersect. The usefulness and effi ciency of communications systems, but routers shunt incoming messages to links heading toward the they also create competition for communal resources. Many messages’ fi nal destinations. But if one wants effi ciency, are people must vie for access to, say, a server offering download- routers the best devices for these intersections? Is switching able songs or to a wireless hot spot. even the right operation to perform?

Until seven years ago, few thought to ask such questions. THE BASIC IDEA But then Rudolf Ahlswede of the University of Bielefeld in Germany, along with Ning Cai, Shuo-Yen Robert Li and Ray- A simple six-node network, in which messages travel along links mond W. Yeung, all then at the University of Hong Kong, (numbered) at one bit per second, can illustrate the basic idea published groundbreaking work that introduced a new ap- of network coding. Nodes are labeled with letters. proach to distributing information across shared networks. In this approach, called network coding, routers are replaced Dana, by coders, which transmit evidence about messages instead of Meeting Friday @ 3. sending the messages themselves. When receivers collect the Carl, - Amy Amy The stock evidence, they deduce the original information from the as- went up! - Ben Amy’s message A sembled clues. Ben 2 1 Although this method may sound counterintuitive, net- Ben’s message B 3 work coding, which is still under study, has the potential to E 5 C dramatically speed up and improve the reliability of all man- Bottleneck F 6 Carl ner of communications systems and may well spark the next 4 revolution in the ﬁ eld. Investigators are, of course, also ex- 7 ploring additional avenues for improving efﬁ ciency; as far as D we know, though, those other approaches generally extend existing methods.

Bits Are Not Cars Dana ahlswede and his colleagues built their proposal in part on the idea, introduced by Shannon, that transmitting WHAT HAPPENS IN A STANDARD NETWORK evidence about data can actually be more useful than convey- If Amy at node A sends a message to Dana at node D just when Ben at node B ing the data directly. They also realized that a receiver would transmits to Carl at node C, the messages hit a bottleneck at node E, where a be able to deduce the original data once enough clues had router would have to relay one message at a time down link 5, causing delays. been gathered but that the receiver would not need to obtain all of the evidence emitted. One kind of clue could be replaced the cars that travel those roads. But an understanding of the by another, and all that was important was receiving some transportation model of communications is useful for grasp- combination of clues that, together, would reveal the original ing how the new scheme works and why it has such promise. message. (Receivers would be able to make sense of the evi- Shannon proved mathematically that every channel has a dence if they were informed in advance about the rules ap- capacity—an amount of information it can relay during any plied to generate it or if instructions on how to use the evi- given time frame—and that communications can proceed re- dence were included in the evidence itself.) liably as long as the channel’s capacity is not exceeded. In the Network coding breaks with the classic view that commu- transportation analogy, a road’s capacity is the number of nications channels are analogous to roads and that bits are like cars per second it can handle safely. If trafﬁ c stays below capacity, a car entering the road at one end can generally be Network Coding guaranteed to exit at the other end unchanged (barring the Overview/ rare accident). Engineers have built increasingly complex ■ In 2000 investigators proposed a seemingly crazy idea communications systems based on the transportation model. for limiting logjams in communications networks. For example, the phone systems Shannon pondered dedicate Called network coding, this potentially revolutionary a distinct “road” to every conversation; two calls over tradi- approach replaces routers, which simply relay tional phone lines never share a single line at the same time messages at intersections, with network coders, which and frequency. send evidence about the incoming messages instead of Computer networks—and the Internet in particular—are the messages themselves. essentially a maze of merging, branching and intersecting ■ Network coding is faring well in experiments, most of roads. Information traveling from one computer to another which so far focus on sending data across multicast typically traverses several roads en route to its destination. networks, where all receivers need to get the same Bits from a single message are grouped into packets (the car- information simultaneously. pools or buses of the information superhighway), each of ■ It promises to make the operations of many networks which is labeled with its intended destination. Routers sit at more efﬁ cient (increasing capacity without having to the intersections of the roads, examine each packet’s header add hardware or bandwidth) as well as faster, more and forward that packet toward its destination. reliable and better resistant to attack. Ironically, the very transportation model that fueled to-

day’s sophisticated communications systems now stands in VINCENT MATT

Amy A Carl, The stock KEY Ben went up! 2 1 - Ben Router B 3 E 5 C Network coder

Evidence about Amy’s F Carl Link and Ben’s messages 6 Link’s 4 directionality 7

D Dana, Meeting Amy Friday @ 3. 1 - Amy A 1 + 0 = 1 1 + 1 = 0 Ben 0 2 1 1 B 3 E 5 C 0 1 1 Dana F 6 Carl 1 4 0 7 D WHAT HAPPENS WITH NETWORK CODING 0 + 1 = 1 Amy sends her message down link 2 as before and also, counterintuitively, down link 1 (which would otherwise be unused) to Carl. And Ben transmits RULE FOR NODES E, C AND D down links 3 and 4. The coder at node E processes the two inputs from If sum is: Result is: links 2 and 3 and sends evidence about Amy’s and Ben’s messages down Dana Even (0 + 0 or 1 + 1) 0 link 5, with no delay. This evidence then gets relayed down links 6 and 7. Odd (0 + 1 or 1 + 0) 1 Carl’s computer deciphers Ben’s message by examining Amy’s message and the evidence received from link 6, and Dana’s computer deciphers Amy’s message analogously. How can that be? Read on. HOW DECODING WORKS Imagine that Amy’s message is the number 1 and that Ben’s is a 0 and that the coder at node E, as well as Carl’s and Dana’s comput- the way of progress. After all, bits are not cars. When two ers, processes two incoming data streams according to the code, vehicles converge on the same narrow bridge, they must take or rule, in the table. Node E would send a 1 down link 5, and Carl’s computer would combine the 1 from link 1 and the 1 from link 6 to turns traversing the bottleneck. When two bits arrive at a produce a 0, thus recovering Ben’s message. Meanwhile Dana’s bottleneck, however, more options are possible—which is computer would combine the 1 from link 7 and the 0 from link 4 to where network coding comes in. produce a 1—Amy’s message. In reality, if nodes C and D were not preprogrammed with the rule, the coder at node E would How It Works transmit it to them along with the evidence it sent down link 5. the hypothetical six-node digital network depicted in the box on these two pages can help clarify those options. link 5 starts, receives a total of two bits per second (one from Recall that in computers, all messages take the form of a string link 2 and one from link 3), but because link 5’s capacity is of binary code. Imagine that each link, or road, in this network one, the router can send only one bit per second along it. In can carry one bit—be it a 0 or a 1—per second and only in the the transportation model, such bottlenecks cause nightmare direction designated by the corresponding arrow. Amy, a net- traffi c jams, with more and more bits piling up over time, work user at node A, hopes to send information at one bit per waiting their turn. second to Dana at node D. Meanwhile Ben at node B hopes to In the new approach [see illustrations above], though, the send, at exactly the same time and rate, information to Carl at plain router would be replaced by a coder, which would have node C. Can both Amy’s and Ben’s demands be satisfi ed simul- more options than would be open to a traffi c cop. Instead of taneously without exceeding any of the links’ capacities? relaying the actual bit streams collected at the bottleneck, the In a router system [see leftmost illustration], the outlook coder could send quite different information. It could, for ex- seems bleak. Both paths, from Amy to Dana and from Ben to ample, add up the number of 1s that arrive during any given Carl, require traversing link 5. This link becomes the equiva- second and transmit a 0 if that sum is even. If the sum is odd, lent of a narrow, one-lane bridge. The router at node E , where the device could transmit a 1. So, if link 5 simultaneously re-

A CODE DESIGN EXAMPLE

To gain more of a taste for how coders would process information and to see why designers of multicast networks can easily select the rules that coders will use, consider our six-node network again. This time, though, Amy and Ben want to send information to both Carl and Dana. This is a multicast problem because both receivers hope to gather all the data. FIVE-BIT BINARY STRINGS Guys, STRING INTEGER VALUE IN Pix of the m = 5 EXAMPLE CODER IN ACTION summer house! 00000 0 XOXO Amy Amy 00001 1 Carl, Dana, 00010 2 Summer lease A 00011 3 a is attached. 00100 4 - Ben Ben 2 1 00101 5 00110 6 00111 7 B Bit string1 3 E X = 2 01000 8 5 C Output = 23 01001 9 Bit string 01010 10 Y = 30 F 6 Carl 01011 11 01100 12 01101 13 4 01110 14 7 01111 15 10000 16 D 10001 17 10010 18 10011 19 COMPUTATION AT NODE E 10100 20 b 10101 21 X ϫ ϫ Remainder after 32 is added or Function: a X + b Y = subtracted repeatedly to yield a 10110 22 value in the 0 through 31 range 10111 23 11000 24 Calculation: 3 ϫ 21 + 20 ϫ 30 = 663 11001 25 Dana 663 – 32 = 631 11010 26 631 – 32 = 599 11011 27 : : 11100 28 55 – 32 = 23 11101 29 OUTPUT = 23 11110 30 Y 11111 31

The coding challenge here basically boils down to selecting a an integer from 0 through 31 (table at right). We can also combine coding rule, or mathematical function, for node E, which receives the strings at node E using the function a ϫ X + b ϫ Y, where X and Y information from links 2 and 3 and transmits the output of the are the values between 0 and 31 that correspond to the strings function on link 5 (above). Many options would achieve our node E receives from links 2 and 3 (say, 21 and 30 at one instant) communication objective, but let us focus on linear functions, to and where a and b are also binary strings of length m (set arbitrarily show how simple the codes can be. In addition, because network here at 3 and 20). We also ensure that the output of this linear coding is typically applied to blocks of bits, we will design the combination is itself a member of the ﬁ nite set 0 through 31. This function for link 5 as a random combination of some set number (m) we do (as shown in computation box) by plugging the values of a, b, of bits from link 2 and an equal number of bits from link 3. X and Y into the formula, then repeatedly adding or subtracting 32 One approach we can take is to make m = 5. There are 32 from the initial result until we reach a number in the range 0 possible binary strings of length 5, and we represent each one by through 31. That last number becomes the output. ceives a 1 and a 0 from links 2 and 3, it carries a 1. If either sources (links 1 and 4) that the routing system could not use- two 0s or two 1s are received from links 2 and 3, link 5 car- fully employ for meeting Amy’s and Ben’s demands. Carl’s ries a 0. The result then gets sent by router F down links 6 and node receives Amy’s transmission and knows for each instant 7 to Carl and Dana, respectively. (from link 6) whether the number of 1s in the pair of mes- This approach replaces each pair of bits at node E with a sages issued by Amy and Ben is even or odd. If Carl’s node is hybrid of the two. Such a bit stream seems ridiculous. Our programmed to also “know” the rule used by the coders at proposed coder has done the equivalent of combining one the start of link 5 or if it can infer the rule from the evidence phone conversation with another in a way that obscures both. itself, the collected evidence will enable it to decipher the mes- The apparent absurdity of the approach is precisely why it sage sent by Ben. And Dana’s node will similarly uncover went uninvestigated for so long. Amy’s message. But sometimes apparent madness is true innovation. A hybrid bit stream may describe neither transmission perfectly, Clear Benefits yet it can supply evidence about both. Suppose we addition- this strategy accomplishes two goals that were ally send Amy’s missive to Carl along link 1 and Ben’s to Dana unthinkable given the limitations of the transportation mod-

along link 4. Sending these two messages uses network re- el. First, it enables the bit leaving a node to travel two paths VINCENT MATT

WHY SELECTING CODES IS EASY At fi rst, selecting codes randomly might Guys, seem like a bad idea. Because a and b can Pix of the each take any value in the set 0 through END RESULT summer house! ϫ Amy XOXO Amy 31, there are 32 32, or 1,024, possible functions of the form a ϫ X + b ϫ Y that we Carl, Dana, A Summer lease can use at node E, and not all of those Ben is attached. functions will ensure that both receivers 2 1 - Ben can determine the original messages X B 3 and Y. For example, if we use a code in E 5 C which a = 0, Dana gets Y on link 4 and b ϫ Y on link 7, and this information will not F 6 Carl suffi ce for obtaining message X, no matter what the value of b is. 4 7 Still, when neither a nor b is 0, both Guys, receivers can retrieve the proper D Pix of the messages successfully. As a result, all but summer house! 63 of the 1,024 possible functions XOXO Amy available to node E are good. Even Carl, Dana, Summer lease choosing the coeffi cients at random is attached. yields a good code roughly 94 percent of - Ben the time. The probability of success Dana approaches 100 percent very rapidly as we increase m, so all that programmers need do to ensure reliability is select an m that is suffi ciently high. If Amy sends her message down link 1 to Carl (in addition to sending it down link 2 to node E) and if Ben sends his message down link 4 as well as down link 3 (above), Carl’s and Dana’s computers will be able to plug the output from E into the same function applied by E stantial networks, because the interchangeable nature of evi- and deduce Y and X, respectively. Repeating the same process for dence means that some packets of evidence can be lost with- each bit string ultimately reveals both messages. (To see exactly how out creating problems. the computers uncover the values of X and Y and acquire the values for a and b, visit www.sciam.com/ontheweb) Lessons from Multicast Networks so far much of the research into implementing network simultaneously, something a car cannot do. Second, it allows coding has focused on multicast networks—in which all receiv- a pair of bit streams arriving at the head of a bottleneck to ers need to get the same information. Internet video games rely combine into a single stream, whereas two cars converging on multicast systems to update every player each time one makes on one narrow bridge cannot become a single entity; one a move. Webcasts of videos or live sporting events and new soft- would have to wait for the other to pass before it could pro- ware released electronically to a large group of customers also ceed across the bridge. travel over multicast networks. Today such networks still use The data-handling approach exemplifi ed by our six-node routers, and a return to the transportation analogy helps to model (a minor variation on one fi rst given by Ahlswede and explain why designing them is usually quite diffi cult. his colleagues in 2000) can potentially increase the capacity of Imagine the country’s highways teeming with cars. Each a network without requiring the addition of extra conduits be- router is like a police offi cer directing traffi c at a single inter- cause it avoids logjams. Using routing alone, our six-node net- section. Incoming cars join the queue behind vehicles that work could sustain simultaneous transmissions averaging one arrived before them. The offi cer reads each car’s destination half of a bit per second. (Because the two competing transmissions would have to share link 5, the effective data rate would MICHELLE EFFROS, RALF KOETTER and MURIEL MÉDARD are long- be one bit per two seconds, or one half of a bit per second, for standing collaborators and friends. Effros is professor of elec- each of the competing demands.) With network coding, the trical engineering at the California Institute of Technology same system supports simultaneous transmissions at one bit (Caltech). Technology Review named her among the top 100 per second. So, here, network coding doubles capacity. young innovators in 2002. Koetter is professor and head of the Sometimes network coding could yield even bigger capac- Institute for Communications Engineering at the Technical Uni- THE AUTHORS ity gains, sometimes none. But the approach would never de- versity of Munich. Médard is associate professor of electrical crease the capacity of a network because, at worst, it would engineering and computer science at the Massachusetts Insti- precisely mimic the actions of router systems. It should also tute of Technology and associate director of the Laboratory for increase reliability and resistance to attacks in relatively sub- Information and Decision Systems there.

A Brief History in turn and directs it on its way. The goal in system design is for each router to direct traffi c in a way that not only speeds of Network Coding each subsequent car to its intended destination but also allows the nation’s transportation system as a whole to satisfy as many drivers as possible. The entries below include some highlights; see “More to Even a central designer with a complete map of all the na- Explore” for related references. A fuller bibliography is at tion’s roads in hand would be hard put to determine the best www.ifp.uiuc.edu/˜koetter/NWC/ —M.E., R.K. and M.M. possible strategy for every router to follow. The diffi culty increases as the network changes over time: rush hours, road 2000: Concept introduced. In a landmark repairs, accidents and sporting events mean the roadways and paper, Rudolf Ahlswede, Ning Cai, Shuo- the demands placed on them change constantly. Yen Robert Li and Raymond W. Yeung Intuition might suggest that designing a system reliant on showed the potential power of network coding should be even harder, because there are network coding in multicast networks, where all receivers get more options to consider. A node could forward data un- identical information. They proved changed, thereby mimicking a router. But it might also mix that good (informative) codes exist, two or more incoming data streams before sending them on, although they did not describe a method and how it mixes them might also be open to consideration; for designing them. further, different nodes might use different algorithms. Luckily, this logic is fl awed. Sometimes adding more op- 2003: Important steps taken toward practical tions actually simplifi es things. Without coding, architects of implementation. Li, Yeung and Cai showed a multicast system would need to enumerate as many paths network coding for multicast networks can as possible from the transmitter to each receiver and then rely on mathematical functions involving determine how many of those paths the network could only addition and multiplication, which support simultaneously. Even for simple networks, fi nd- reduces the complexity of designing ing and testing all combinations of paths would be a diz- codes. And two of us (Koetter and zying task. Médard) introduced a powerful algebraic In contrast, a multicast system using network coding framework for analyzing coding would be rather easy to design. The startling truth is that approaches and simplifying code design. addition and multiplication are the only mathematical func- 2005–2006: Valuable design algorithms published. tions that coded networks need apply. Also, even if the func- Sidharth Jaggi, then at Caltech, with Peter Sanders of the tion, or rule, programmed into each coder in a network is University of Karlsruhe in Germany, one of us (Effros) chosen independently of the message and the other coding and collaborators and, separately, Tracey functions and without any knowledge of the network layout, Ho of Caltech, with the three of us and the system as a whole will, with extremely high probability, others, published low-complexity operate at peak performance. Even if the system changes over algorithms for designing the functions time, as can happen in mobile or reconfi gurable networks, the used by each node in a multicast network will continue to perform optimally without requir- network. The fi rst paper gave a ing redesign. To learn why, see the box on page 82. systematic approach for designing functions; the second showed that Tomorrow’s Networks choosing functions randomly and the operation of networks, then, will be very differ- independently for each node should work ent if coders replace routers. The way our messages traverse just as well. (Early versions of both contributions were networks will change: they will not only share “the road” with presented at a conference in 2003.) other transmissions but may become intimately entangled with traffi c from a variety of other sources. Some might fear that 2006: Applications for wireless networks explored. such entanglement would compromise the security of the mes- At a conference in 2006, Christina Fragouli sages. More likely, though, traffi c traversing networks would of the École Polytechnique Fédérale de become a locally undecipherable algebraic stream. Users on Lausanne in Switzerland and the network would unwittingly collaborate to one another’s collaborators demonstrated the potential benefi ts of network coding mutual advantage, allowing not just higher rates or faster for wireless applications and downloads of data but also, in the case of wireless net- characterized scenarios in which the works, an improvement in energy effi ciency. (Because each approach would be particularly helpful. wireless transmission consumes energy, a node can reduce consumption by mixing together the information intended

for several neighbors and sending only a single transmission.) VINCENT MATT

By changing how networks node. It also allows enormous fl exibility, because the network travels with the users, constantly reconfi guring and reestab- function, network coding lishing connections as needed. By changing how networks function, network coding may may influence society in infl uence society in ways we cannot yet imagine. In the mean- ways we cannot yet imagine. time, though, those of us who are studying it are considering the obstacles to implementation. Transitioning from our router-based system to a network-coded one will actually be one Moreover, delays in downloading videos and lost cell of the more minor hurdles. That conversion can be handled phone calls will be far less common. On the Internet, routers by a gradual change rather than a sudden overhaul; some fail or are taken down for maintenance and data packets are routers could just be reprogrammed, and others not built to dropped all the time. That is why people must sometimes re- perform coding operations would be replaced little by little. request Web pages and why a site sometimes comes up slowly. A bigger challenge will be coping with issues beyond re- Reliability will increase with network coding, because it does placing routers with coders. For instance, mixing information not require every single piece of evidence to get through. is a good strategy when the receiving node will gather enough And network managers will provide such benefi ts without evidence to recover what it desires from the mixture. This con- having to add new communications channels, because better dition is always met in multicast networks but may not be the use will be made of existing channels. Network coding will case in general. Moreover, in some circumstances, such as thereby complement other communications technologies, al- when multiple multicasts are transmitted, mixing information lowing users to get as much as possible out of them. can make it diffi cult or impossible for users to extract the Sometimes users will know that network coding is operat- proper output. How, then, can nodes decide which informa- ing, because it may modify how some common applications, tion can and cannot be mixed when multiple connections such as peer-to-peer downloads, function. Today someone share the same network? In what ways must network coding seeking to download a fi le searches for a collaborating user in wireless networks differ from its use in wired ones? What on whose machine the fi le resides. In a system using network are the security advantages and implications of network cod- coding, the fi le would no longer be stored as a whole or in ing? How will people be charged for communications services recognizable pieces. when one person’s data are necessarily mixed with those of But users would not personally have to fi gure out how to other users? In collaborations that span the globe, we and oth- fi nd the evidence needed to obtain the desired fi les. A request ers are pondering how to unravel such knots even as we strive sent into a network from a user’s computer or phone would to enhance the capabilities of the communications networks cause either that individual’s computer or a local server to that have become such an integral part of so many lives. scavenge through the network for pieces of evidence related to a fi le of interest. The gathered evidence, consisting of alge- For more on network coding, including braically mixed pieces of information relating to the desired sa a fuller explanation of how it can vastly fi le, would help recover that fi le. Instead of putting together a improve transmission rates, visit puzzle whose pieces are recognizable fragments of a whole, www.sciam.com/ontheweb the server or an individual’s computer would solve a collection of algebraic equations. And, all the while, most people MORE TO EXPLORE would remain blissfully unaware of these operations—just as A Mathematical Theory of Communication. C. E. Shannon in Bell most of us are ignorant of the complicated error-correction System Technical Journal, Vol. 27, pages 379–423 and 623–656; operations in our cell phones. July and October 1948. Available at http://cm.bell-labs.com/cm/ms/ what/shannonday/paper.html The military has recognized the robustness of network Network Information Flow. R. Ahlswede, N. Cai, S.-Y. R. Li and R. W. coding and is now funding research into its use in mobile ad Yeung in IEEE Transactions on Information Theory, Vol. 46, No. 4, hoc networks, which can form on the fl y. Such networks are pages 1204–1216; July 2000. valuable in highly changeable environments, such as on the Linear Network Coding. S.-Y. R. Li, R. W. Yeung and N. Cai in IEEE Transactions on Information Theory, Vol. 49, No. 2, pages 371–381; battlefi eld, where reliable communications are essential and February 2003. establishing and maintaining an infrastructure of fi ber-optic An Algebraic Approach to Network Coding. R. Koetter and M. Médard in cables or cell towers is diffi cult. In an ad hoc network, every IEEE/ACM Transactions on Networking, Vol. 11, No. 5, pages 782–795; soldier’s radio becomes a node in a communications system, October 2003. and each node seeks out and establishes connections to neigh- Polynomial Time Algorithms for Multicast Network Code Construction. S. Jaggi, P. Sanders, P. A. Chou, M. Effros, S. Egner, K. boring nodes; together these connections establish a net- Jain and L.M.G.M. Tolhuizen in IEEE Transactions on Information Theory, work’s links. Every node can both send and receive messages Vol. 51, No. 6, pages 1973–1982; June 2005. and serve as an intermediary to pass along messages intended A Random Linear Network Coding Approach to Multicast. T. Ho, M. Médard, R. Koetter, D. R. Karger, M. Effros, J. Shi and B. Leong in IEEE for other receivers. This technique extends communications Transactions on Information Theory, Vol. 52, No. 10, pages 4413–4430; capabilities far beyond the transmission range of a single October 2006. www.sciam.com SCIENTIFIC AMERICAN 85 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more..

INNOVATIONS Seeing Anticipated for decades, machines Triple are ﬁ nally Triplele displaying Trip real objects

in three true Inventors have struggled for years to create displays that can conjure vivid three-dimensional images that users can manipulate dimensions and interact with. Chemists could exploit such marvels to design new drug molecules. Oil and gas explorers could see exactly where to aim their drills. Surgeons could pass probes or radiation beams through

the collated slices of diagnostic data pro- products are just now transitioning from duced by magnetic resonance imaging the laboratory to commercial models. (MRI) and computed tomography (CT) machines to test procedures before per- Spinning Algorithms forming an operation. But shortcomings wait a minute. Aren’t holograms such as a fl ickering image, a narrow an- three-dimensional and viewable without gle of view or the need to wear special funny glasses? Yes, but they are record- glasses have bedeviled the devices. ed once as a fi nal image and thus do not Two companies have recently mixed allow interactivity. Engineers have also their own technology with off-the-shelf knit together cubes and spinning arrays components, including the Digital Light of light-emitting diodes to give a full- Processor (DLP) chip from Texas Instru- bodied view, but the resolution is coarse, By ments, to create interactive systems restricted by the connections among the called 3-D volumetric displays that over- diodes. Other contenders appear to be Stuart F. Brown come these limitations. The two fi rms’ 3-D but really are not; the Heliodisplay

ANGLE OF ATTACK: Data from a CT scan is pro- jected in 3-D in PerspectaRad, revealing a brain tumor core (yellow diamond, center) and possible paths for radiation treatment (green rays).

data,” says Gregg E. Favalora, Actuali- ty’s chief technology officer. “For instance, we have a patent on how you draw a straight line on a rotating screen — because it is not obvious what dot to pick on that grid as it spins around.” Perspecta creates a glowing, semi- translucent image. Every volumetric pixel (voxel) that seems to be in a certain point in space really is there, but it becomes visible only when the screen sweeps through that point—hence the fast rotation rate. The screen is made from a plastic that looks like taut tissue paper and is 50 percent refl ective and 50 percent transmissive, allowing the imagery to be seen from any angle by observ- ers gathered around it. The unit provides both vertical and horizontal parallax; if a viewer moves his or her head up and down or left and right, background objects that were previously obscured by foreground objects will come into view, as they are perceived in the real world. A user can also wield a penlike mouse to zoom into or out of the image, rotate and fl ip it, or change colors. This feature has only recently been made possible by from IO2 Technology in San Francisco by a CT, MRI or PET (positron-emission rapid advances in computer graphics. projects fl oating images onto a vertical tomography) scanner and mathematical- “In our fi rst demo in 2002,” Favalora re- plane of fi ne mist suspended above the ly divides the information into 198 radi- calls, “it took 45 minutes of processing instrument that seem to possess depth, ally disposed segments, like an apple time just to perform a click-and-drag. but the illusion comes from an absence thinly sliced about its core. Held in a Now we get a video card off the shelf for of depth cues, not actual imagery in the frame-buffer memory, the data slices are a few hundred dollars, and it computes third dimension. Users who want to load fed to three DLP chips. DLPs are arrays the problem just like that.” 3-D medical data into a machine one day of hundreds of thousands of tiny mirrors, It was also a while before Favalora and 3-D military scenes the next and each of which can be individually tilted and his colleagues realized that perfect- then turn, prod or alter them while on by onboard circuitry. They form the ing the core technologies would not be view can exploit two inventions worthy heart of projection televisions and new enough to build a business; they had to of the term “volumetric”: Perspecta and slide projectors, as well as digital movie identify an initial market and develop a DepthCube. projectors that may replace fi lm reels in ready-to-use system tailored to it. That Perspecta, developed by Actuality theaters. In Perspecta, each DLP is as- niche turned out to be radiation therapy Systems in Bedford, Mass., might best be signed a color and projects its light described as a crystal ball for looking in- through a prism onto the rapidly spin- STUART F. BROWN was formerly a side objects. A transparent polycarbon- ning screen, creating a 3-D apparition. staff writer at Popular Science and ate dome houses a fl at, disk-shaped screen A lot of mathematical heavy lifting Fortune. He calls his now freelance 10 inches in diameter that rotates on a was needed to make Perspecta work. “It beat “the man-made world,” which in- vertical spindle at 900 revolutions per took us three or four years to invent the cludes aerospace, transportation and THE AUTHOR

CHRISTOPHER HARTING CHRISTOPHER minute. The system takes data generated algorithms that let us slice the image biotechnology.

for cancer tumors. Doctors need to care- see exactly where the beams will strike which method had been used. The pro- fully plan the paths along which they the tumor, which healthy brain tissues tocols developed with PerspectaRad aim radiation beams, trying to maxi- they will pass through, and the dose turned out better in six cases, equivalent mize the killing effect where the rays cloud—the volume of tissue that will be in four cases, and worse in two cases. In converge on a tumor while minimizing affected by the radiation. This imagery one patient, PerspectaRad made it clear damage to nearby healthy tissue. Be- helps doctors adjust the beams to im- how to reduce incidental damage to the cause oncologists must work with 2-D prove treatment or reduce damage. The optic nerve. Calling the results “inter- slices of scanner data, planning the fi rst PerspectaRad systems cost about esting,” Chu is planning a larger study beam paths for a treatment can take sev- $90,000. Greater production could low- that will include patients with tumors in eral hours. Actuality developed its Per- er the price to $65,000, according to Fa- other parts of the body. “When working spectaRad system as an add-on to exist- valora, but the displays are unlikely to just with CT data,” he says, “you have ing radiation therapy equipment manu- reach consumer markets. to look at individual slices and some- factured by Philips Medical Systems. Nevertheless, treatment stands to how integrate them all in your head to PerspectaRad sports the 3-D display gain. James Chu, head of the medical get a 3-D picture. With Perspecta, you plus software that connects the device to physics department at Rush University see the 3-D picture directly.” the Philips systems. When a doctor Medical Center in Chicago, recently Chu is also excited by Perspecta’s pushes a button, an image of the CT studied 12 patients with brain tumors ability to display moving images of inter- data for, say, a brain tumor appears in for whom treatment plans were devel- nal body parts. Because internal organs 3-D. Another button adds the radiation oped using both PerspectaRad and con- and tissue move as the heart beats and pathways chosen by a dosimetrist, who ventional methods. The plans were re- the lungs fi ll and empty, it is very useful plans the treatment. The physician can viewed by doctors who were unaware of to be able to discern the axis of motion of a tumor. With this information, a doctor PERSPECTA: CRYSTAL BALL could direct a lower-energy radiation A transparent dome, translucent screen and optics all rotate at 900 rpm to beam along the motion axis, rather than create a 3-D image. A computer sends graphics data to electronics below the a more intense beam across it, lessening pedestal, which instruct three light-processor chips to focus light from an arc collateral damage. Chu says Perspecta lamp through a projector lens. It reflects a beam up through the spinning shaft could also allow more precise implanting and across relay mirrors onto the screen. A second, larger dome (not shown) of radioactive “seeds” in the prostate encases the spinning parts for safety. gland to treat cancer there, by allowing a physician to better compensate for motion of tissue that occurs when the needle that delivers the seeds is inserted. Voxels in Glass Plates depthcube, the other interactive vol- Dome umetric display, was developed by Light- Space Technologies in Norwalk, Conn. It is a rear-projection monitor shaped somewhat like a computer terminal, Screen with a face that measures 16 by 12 inches. The “screen” is four inches thick and made of 20 vertical, transparent glass Mirrors plates sandwiched together, yet the imagery looks about 12 inches deep. Col- laborators standing randomly in front of the screen will each see objects with the Projector lens Arc lamp corresponding perspective. Internal structures appear and disappear as a Light-processor chips viewer’s angle changes. The system Prism would be handy for, say, a team of prod- Pedestal uct engineers studying how parts draft- ed on a computer-aided design system would—or would not—fi t together. When LightSpace president Alan

Sullivan built his ﬁ rst prototype eight 5W INFOGRAPHICS; SOURCE: ACTUALITY SYSTEMS

DEPTHCUBE: THICK-SCREEN IMAGING theater stage, standing in front of and behind one another. That was when Sul- An arc lamp shines light through optics and a prism, breaking it into blue, red and livan, who formerly studied ultrahigh- green beams that are reﬂ ected by light-processor chips through a projector energy lasers at Lawrence Livermore toward relay mirrors. The mirrors direct the beams onto one of 20 screens. Each National Laboratory, had a smart attack screen consists of two glass plates that contain conducting layers and a liquid- that won him a patent. It occurred to crystal mixture that scatters light. By illuminating the four-inch-deep set of him that the so-called antialiasing algo- screens in succession, the system creates a 3-D image that appears 12 inches deep. The sample image of half a human rithms used to smooth jagged edges in head allows particular structures to be seen 2-D images could also be applied to (sinuses are bright yellow; cartilage is smoothing the transitions between the orange), whereas others can be hidden. DepthCube’s 20 planes. This innovation makes the display’s 15.3 million physical voxels look like a whopping 465 million virtual voxels. “We produce 31 subplanes between the physical planes, Set of screens so the perceived resolution is much higher,” Sullivan explains. As a result, to the human brain, the images can appear to be as much as 12 inches deep. The image data that are fed to the Glass plate Liquid crystal chips can come from nearly any 3-D soft- Air Conducting layer ware that runs the OpenGL application Light-processor programming interface, a common pro- chips tocol used by computer-aided design and Arc lamp engineering programs such as Catia or ProEngineer. LightSpace has sold a hand- Prism ful of DepthCubes to research institu- Projector lens tions, including the U.S. Air Force Re- search Laboratory and Hokkaido Uni- Mirrors versity in Japan, for about $50,000 apiece. Sullivan acknowledges that the market is limited at this price but says he can see the years ago, he managed to coax a trio of free.” The information drives the 20 path to a product that would cost about Texas Instruments DLP chips into pro- plates, known as liquid-crystal scatter- $5,000. “There’s nothing in our architec- jecting depth-related imagery onto the ing shutters, which can rapidly change ture that’s different from what’s in a rear- 20 plates, which are separated by thin from a transparent state to a scattering projection TV, except for the liquid-crys- air gaps. In DepthCube, each DLP con- state. That trait allows a plate to let pix- tal shutters,” he says, “and those could be ) tains 786,432 mirrors that tile an area els pass through to other plates as need- produced quite cheaply in volume.” similar to that of a ﬁ ngernail. ed yet also enables it to display a pixel. The products developed by these two

screen shot screen Sullivan still needed a convenient At any moment, all the plates are blank young companies are garnering respect way to generate depth information, and except for one, but the processors pro- from boffi ns in the 3-D world, and more he was elated when he realized that an ject coordinated image slices 50 times a applications will follow. Optical scien- affordable, commercial 3-D graphics second onto every plate, creating full tist Steve Hines, owner of HinesLab in card could suffi ce. Graphics cards use a depth, height and width. Glendale, Calif., states that “both these color buffer—a morsel of memory—to The prototype DepthCube conveyed groups are doing extremely hard things assign the appropriate color to every pix- three-dimensionality to a viewer nicely and have pulled them off.” The natural el on a two-dimensional screen. But the but only within the four-inch depth of places to sell the technology, he adds, ); LIGHTSPACE TECHNOLOGIES ( TECHNOLOGIES LIGHTSPACE ); cards also have a hidden component, the screen; items in an image appeared will be where the money is: “medicine, called the depth buffer, which describes almost like fl at scenery elements on a the military and the movies.” the depth of every pixel. In a normal ap- illustration plication, the depth buffer is largely un- MORE TO EXPLORE tapped, because only the frontmost layer Volumetric 3D Displays and Application Infrastructure. Gregg E. Favalora in Computer, Vol. 38, of a pixel must be defi ned to create a 2-D No. 8, pages 37–44; August 2005. A Method for the Real-Time Construction of a Full Parallax Light Field. K. Tanaka and S. Aoki picture. So a place for Sullivan’s depth in Stereoscopic Displays and Virtual Reality Systems XIII. Edited by A. J. Woods et al. Proceedings

5W INFOGRAPHICS ( INFOGRAPHICS 5W information, he muses, “was in there for of the SPIE, Vol. 6055, Article 605516; January 30, 2006.

By Kaushik Basu THE TRAVELER’S

Lucy and Pete, returning from a remote Pacifi c island, fi nd that the airline has damaged the identical antiques that What numbers will Lucy and Pete write? What number each had purchased. An airline manager says that he is happy would you write? to compensate them but is handicapped by being clueless about Scenarios of this kind, in which one or more individuals the value of these strange objects. Simply asking the travelers have choices to make and will be rewarded according to those for the price is hopeless, he fi gures, for they will infl ate it. choices, are known as games by the people who study them Instead he devises a more complicated scheme. He asks (game theorists). I crafted this game, “Traveler’s Dilemma,” each of them to write down the price of the antique as any in 1994 with several objectives in mind: to contest the narrow dollar integer between 2 and 100 without conferring togeth- view of rational behavior and cognitive processes taken by er. If both write the same number, he will take that to be the economists and many political scientists, to challenge the lib- true price, and he will pay each of them that amount. But if ertarian presumptions of traditional economics and to high- they write different numbers, he will assume that the lower light a logical paradox of rationality. one is the actual price and that the person writing the higher Traveler’s Dilemma (TD) achieves those goals because the number is cheating. In that case, he will pay both of them the game’s logic dictates that 2 is the best option, yet most people lower number along with a bonus and a penalty—the person pick 100 or a number close to 100—both those who have not who wrote the lower number will get $2 more as a reward for thought through the logic and those who fully understand honesty and the one who wrote the higher number will get $2 that they are deviating markedly from the “rational” choice. less as a punishment. For instance, if Lucy writes 46 and Pete Furthermore, players reap a greater reward by not adhering

writes 100, Lucy will get $48 and Pete will get $44. to reason in this way. Thus, there is something rational about COLLINS MATT

When playing this simple game, people consistently reject the rational choice. In fact, by acting illogically, they end up reaping a larger reward— an outcome that demands a new kind of formal reasoning

choosing not to be rational when playing Traveler’s Dilemma. on. Continuing with this line of reasoning would take the trav- In the years since I devised the game, TD has taken on a elers spiraling down to the smallest permissible number, name- life of its own, with researchers extending it and reporting ly, 2. It may seem highly implausible that Lucy would really go ﬁ ndings from laboratory experiments. These studies have all the way down to 2 in this fashion. That does not matter (and produced insights into human decision making. Nevertheless, is, in fact, the whole point)—this is where the logic leads us. open questions remain about how logic and reasoning can be Game theorists commonly use this style of analysis, called applied to TD. backward induction. Backward induction predicts that each player will write 2 and that they will end up getting $2 each (a Common Sense and Nash result that might explain why the airline manager has done so to see why 2 is the logical choice, consider a plausible line well in his corporate career). Virtually all models used by game of thought that Lucy might pursue: her ﬁ rst idea is that she theorists predict this outcome for TD—the two players earn should write the largest possible number, 100, which will earn $98 less than they would if they each naively chose 100 without her $100 if Pete is similarly greedy. (If the antique actually cost thinking through the advantages of picking a smaller number. her much less than $100, she would now be happily thinking Traveler’s Dilemma is related to the more popular Prison- about the foolishness of the airline manager’s scheme.) er’s Dilemma, in which two suspects who have been arrested Soon, however, it strikes her that if she wrote 99 instead, for a serious crime are interrogated separately and each has she would make a little more money, because in that case she the choice of incriminating the other (in return for leniency would get $101. But surely this insight will also occur to Pete, by the authorities) or maintaining silence (which will leave and if both wrote 99, Lucy would get $99. If Pete wrote 99, the police with inadequate evidence for a case, if the other then she could do better by writing 98, in which case she would prisoner also stays silent). The story sounds very different get $100. Yet the same logic would lead Pete to choose 98 as from our tale of two travelers with damaged souvenirs, but well. In that case, she could deviate to 97 and earn $99. And so the mathematics of the rewards for each option in Prisoner’s

Dilemma is identical to that of a variant of TD in which each $101. Because Lucy is better off by this change, the outcome player has the choice of only 2 or 3 instead of every integer (100, 100) is not a Nash equilibrium. from 2 to 100. TD has only one Nash equilibrium—the outcome (2, 2), Game theorists analyze games without all the trappings whereby Lucy and Pete both choose 2. The pervasive use of of the colorful narratives by studying each one’s so-called the Nash equilibrium is the main reason why so many formal payoff matrix—a square grid containing all the relevant in- analyses predict this outcome for TD. formation about the potential choices and payoffs for each Game theorists do have other equilibrium concepts—strict player [see box on opposite page]. Lucy’s choice corresponds equilibrium, the rationalizable solution, perfect equilibrium, to a row of the grid and Pete’s choice to a column; the two the strong equilibrium and more. Each of these concepts leads numbers in the selected square specify their rewards. to the prediction (2, 2) for TD. And therein lies the trouble. Despite their names, Prisoner’s Dilemma and the two- Most of us, on introspection, feel that we would play a much choice version of Traveler’s Dilemma present players with no larger number and would, on average, make much more than real dilemma. Each participant sees an unequivocal correct $2. Our intuition seems to contradict all of game theory. choice, to wit, 2 (or, in the terms of the prisoner story line, incriminate the other person). That choice is called the dom- Implications for Economics inant choice because it is the best thing to do no matter what the game and our intuitive prediction of its outcome the other player does. By choosing 2 instead of 3, Lucy will also contradict economists’ ideas. Early economics was ﬁ rmly receive $4 instead of $3 if Pete chooses 3, and she will receive tethered to the libertarian presumption that individuals should $2 instead of nothing if Pete chooses 2. be left to their own devices because their selﬁ sh choices will re- Game theory predicts that the Nash equilibrium will occur when Traveler’s

John Nash Dilemma is played rationally.

In contrast, the full version of TD has no dominant choice. sult in the economy running effi ciently. The rise of game-theo- If Pete chooses 2 or 3, Lucy does best by choosing 2. But if retic methods has already done much to cut economics free from Pete chooses any number from 4 to 100, Lucy would be better this assumption. Yet those methods have long been based on the off choosing a number larger than 2. axiom that people will make selfi sh rational choices that game When studying a payoff matrix, game theorists rely most theory can predict. TD undermines both the libertarian idea often on the Nash equilibrium, named after John F. Nash, Jr., that unrestrained selfi shness is good for the economy and the of Princeton University. (Russell Crowe portrayed Nash in the game-theoretic tenet that people will be selfi sh and rational. movie A Beautiful Mind.) A Nash equilibrium is an outcome In TD, the “effi cient” outcome is for both travelers to from which no player can do better by deviating unilaterally. choose 100 because that results in the maximum total earn- Consider the outcome (100, 100) in TD (the fi rst number is ings by the two players. Libertarian selfi shness would cause Lucy’s choice, and the second is Pete’s). If Lucy alters her se- people to move away from 100 to lower numbers with less ef- lection to 99, the outcome will be (99, 100), and she will earn fi ciency in the hope of gaining more individually. And if people do not play the Nash equilibrium strategy Sensible Irrationality (2), economists’ assumptions about rational behavior should Overview/ be revised. Of course, TD is not the only game to challenge ■ In the game Traveler’s Dilemma, two people separately the belief that people always make selfi sh rational choices [see choose a whole number from 2 to 100, and the player “The Economics of Fair Play,” by Karl Sigmund, Ernst Fehr whose number is smaller is rewarded with a larger sum. and Martin A. Nowak; Scientifi c American, January Game theory insists that rationality should lead players 2002]. But it makes the more puzzling point that even if play- to select 2, but most people pick an integer closer to 100. ers have no concern other than their own profi t, it is not ra- ■ A new kind of reasoning is needed to gain a rigorous tional for them to play the way formal analysis predicts. understanding of this rational choice not to be rational. TD has other implications for our understanding of real- ■ The results of Traveler’s Dilemma contradict economists’ world situations. The game sheds light on how the arms race Office University Communications, Princeton of assumption that standard game theory can predict how acts as a gradual process, taking us in small steps to ever supposedly selfi sh rational people will behave. They also worsening outcomes. Theorists have also tried to extend TD show how selfi shness is not always good economics. to understand how two competing fi rms may undercut each

other’s price to their own detriment (though in this case to MATTHEWSBOB

PAYOFF MATRIX OF THE TRAVELER’S DILEMMA

This payoff matrix summarizes everything game theorists The outcome in which both players choose 2 and each need to know about Traveler’s Dilemma. Lucy’s possible earns $2 (gold) is called the Nash equilibrium: Lucy does choices are shown in the leftmost column; Pete’s run across worse (earning $0) if she chooses any other number and Pete the top row. The ﬁ rst number in the square at the intersection still chooses 2. Similarly, Pete does worse if he alone chooses of the chosen row and column is Lucy’s payoff, and the second something other than 2. number is Pete’s payoff. For example, if Lucy chooses 98 and When restricted to choices of only 2 and 3 (black outline), Pete 99, then Lucy receives $100 and Pete receives $96. the game becomes equivalent to Prisoner’s Dilemma. PETE’S CHOICE (dollars) 234... 98 99 100 22 2 4 0 4 0 ... 4 0 4 0 4 0 30 4 3 3 5 1 ... 5 1 5 1 5 1 40 4 1 5 4 4 ... 6 2 6 2 6 2 ......

98 0 4 1 5 2 6 ... 98 98 100 96 100 96

LUCY’S CHOICE (dollars) LUCY’S CHOICE 99 0 4 1 5 2 6 ... 96 100 99 99 101 97 100 0 4 1 5 2 6 ... 96 100 97 101 100 100

the advantage of the consumers who buy goods from them). Would they learn to play the Nash equilibrium, even if that All these considerations lead to two questions: How do was not their fi rst instinct? Sure enough, when the reward was people actually play this game? And if most people choose a large the play converged, over time, down toward the Nash number much larger than 2, can we explain why game theory outcome of 80. Intriguingly, however, for small rewards the fails to predict that? On the former question, we now know a play increased toward the opposite extreme, 200. lot; on the latter, little. The fact that people mostly do not play the Nash equilibrium received further confi rmation from a Web-based experi- How People Actually Behave ment with no actual payments that was carried out by Ariel over the past decade researchers have conducted Rubinstein of Tel Aviv University and New York University many experiments with TD, yielding several insights. A cel- from 2002 to 2004. The game asked players, who were going ebrated lab experiment using real money with economics stu- to attend one of Rubinstein’s lectures on game theory and Nash, dents as the players was carried out at the University of Vir- to choose an integer between 180 and 300, which they were to ginia by C. Monica Capra, Jacob K. Goeree, Rosario Gomez think of as dollar amounts. The reward/penalty was set at $5. and Charles A. Holt. The students were paid $6 for partici- Around 2,500 people from seven countries responded, pating and kept whatever additional money they earned in the giving a cross-sectional view and sample size infeasible in a game. To keep the budget manageable, the choices were val- laboratory. Fewer than one in seven players chose the scenar- ued in cents instead of dollars. The range of choices was made io’s Nash equilibrium, 180. Most (55 percent) chose the max- 80 to 200, and the value of the penalty and reward was varied imum number, 300 [see box on next page]. Surprisingly, the for different runs of the game, going as low as 5 cents and as data were very similar for different subgroups, such as people high as 80 cents. The experimenters wanted to see if varying from different countries. the magnitude of the penalty and reward would make a dif- The thought processes that produce this pattern of choic- ference in how the game was played. Altering the size of the es remain mysterious, however. In particular, the most popu- reward and penalty does not change any of the formal analy- lar response (300) is the only strategy in the game that is sis: backward induction always leads to the outcome (80, 80), which is the Nash equilibrium in every case. KAUSHIK BASU is professor of economics, Carl Marks Professor The experiment confi rmed the intuitive expectation that of International Studies and director of the Center for Analytic the average player would not play the Nash equilibrium strat- Economics at Cornell University. He has written extensively in egy of 80. With a reward of 5 cents, the players’ average choice academic journals on development economics, welfare eco- was 180, falling to 120 when the reward rose to 80 cents. nomics, game theory and industrial organization. He also

Capra and her colleagues also studied how the players’ THE AUTHOR writes for the popular media, including a monthly column in

MATT COLLINS MATT behavior might alter as a result of playing TD repeatedly. BBC News online. He is a fellow of the Econometric Society.

“dominated”—which means there is another strategy (299) earn only an additional dollar, and so we choose 100 even that never does worse and sometimes does better. though we fully understand that, rationally, 99 is a better Rubinstein divided the possible choices into four sets of choice for us as individuals. numbers and hypothesized that a different cognitive process To go further and explain more of the behaviors seen in lies behind each one: 300 is a spontaneous emotional response. experiments such as these, some economists have made strong Picking a number between 295 and 299 involves strategic rea- and not too realistic assumptions and then churned out the soning (some amount of backward induction, for instance). observed behavior from complicated models. I do not believe Anything from 181 to 294 is pretty much a random choice. that we learn much from this approach. As these models and And fi nally, standard game theory accounts for the choice of assumptions become more convoluted to fi t the data, they 180, but players might have worked that out for themselves or provide less and less insight. may have had prior knowledge about the game. A test of Rubinstein’s conjecture for the fi rst three groups Unsolved Problem would be to see how long each player took to make a decision. the challenge that remains, however, is not ex- Indeed, those who chose 295 to 299 took the longest time on plaining the real behavior of typical people presented with average (96 seconds), whereas both 181 to 294 and 300 took TD. Thanks in part to the experiments, it seems likely that about 70 seconds—a pattern that is consistent with his hy- altruism, socialization and faulty reasoning guide most indi- pothesis that people who chose 295 to 299 thought more than viduals’ choices. Yet I do not expect that many would select 2 those who made other choices. if those three factors were all eliminated from the picture. Game theorists have made a number of attempts to ex- How can we explain it if indeed most people continue to plain why a lot of players do not choose the Nash equilibrium choose large numbers, perhaps in the 90s, even when they in TD experiments. Some analysts have argued that many have no dearth of deductive ability, and they suppress their people are unable to do the necessary deductive reasoning and normal altruism and social behavior to play ruthlessly to try therefore make irrational choices unwittingly. This explana- to make as much money as possible? Unlike the bulk of mod- tion must be true in some cases, but it does not account for all ern game theory, which may involve a lot of mathematics but the results, such as those obtained in 2002 by Tilman Becker, is straightforward once one knows the techniques, this ques- Michael Carter and Jörg Naeve, all then at the University of tion is a hard one that requires creative thinking. Hohenheim in Germany. In their experiment, 51 members of the Game Theory Society, virtually all of whom are profes- WHAT WERE THEY THINKING? sional game theorists, played the original 2-to-100 version of TD. They played against each of their 50 opponents by select- Researchers hypothesized that different thought processes ing a strategy and sending it to the researchers. The strategy lay behind different types of choices that people made could be a single number to use in every game or a selection playing a version of Traveler’s Dilemma with the options pie chart of numbers and how often to use each of them. The game had ranging from 180 to 300 ( ): a spontaneous emotional response (choosing 300), a strategically a real-money reward system: the experimenters would select reasoned choice (295–299) or a random one (181–294). one player at random to win $20 multiplied by that player’s Players making the formal rational choice (180) might have average payoff in the game. As it turned out, the winner, who deduced it or known about it in advance. As expected, had an average payoff of $85, earned $1,700. people making “spontaneous” or “random” selections took Of the 51 players, 45 chose a single number to use in every the least time to choose (bar graph). game (the other six specifi ed more than one number). Among those 45, only three chose the Nash equilibrium (2), 10 chose CHOICES MADE RESPONSE TIMES the dominated strategy (100) and 23 chose numbers ranging 100 from 95 to 99. Presumably game theorists know how to rea- 180 181–294 son deductively, but even they by and large did not follow the 13% 14% rational choice dictated by formal theory. 80 Superfi cially, their choices might seem simple to explain: 295–299 most of the participants accurately judged that their peers 17% 60 would choose numbers mainly in the high 90s, and so choos- 300 ing a similarly high number would earn the maximum aver- 55% age return. But why did everyone expect everyone else to Seconds 40 choose a high number? Perhaps altruism is hardwired into our psyches alongside Formal rational choice 20 selfi shness, and our behavior results from a tussle between Random the two. We know that the airline manager will pay out the Strategic reasoning largest amount of money if we both choose 100. Many of us Spontaneous emotional response 0

do not feel like “letting down” our fellow traveler to try to CHRISTIANSEN JEN

Game Theory vs. Ordinary Decision Theory: I Know that You Know that I Know ... mysteriously, none of the monkeys threw intuition and that, in the case of Traveler’s any hats, and only one monkey came down. Dilemma, the intuition is right and It took the hat on the ground fi rmly in hand, awaiting validation by a better logic. The walked up to the hat seller, gave him a slap problem is akin to what happened in early and said, “You think only you have a set theory. At that time, mathematicians grandfather?” took for granted the existence of a This story illustrates an important universal set—a set that contained distinction between ordinary decision everything. The universal set seemed theory and game theory. In the latter, what extremely natural and obvious, yet is rational for one player may depend on ultimately several paradoxes of set theory what is rational for the other player. For were traced to the assumption that it Lucy to get her decision right, she must put existed, which mathematicians now know herself in Pete’s shoes and think about is fl awed. In my opinion, the common what he must be thinking. But he will be knowledge of rationality assumed by game thinking about what she is thinking, theorists faces a similar demise. —K.B. leading to an infi nite regression. Game theorists describe this situation by saying that “rationality is common knowledge I heard this tale in India. A hat seller, on among the players.” In other words, Lucy waking from a nap under a tree, found that and Pete are rational, they each know that a group of monkeys had taken all his hats the other is rational, they each know that to the top of the tree. In exasperation he the other knows, and so on. took off his own hat and fl ung it to the The assumption that rationality is ground. The monkeys, known for their common knowledge is so pervasive in imitative urge, hurled down the hats, game theory that it is rarely stated which the hat seller promptly collected. explicitly. Yet it can run us into problems. In Half a century later his grandson, also some games that are played over time, a hat seller, set down his wares under the such as repeated rounds of Prisoner’s same tree for a nap. On waking, he was Dilemma, players can make moves that are dismayed to discover that monkeys had incompatible with this assumption. taken all his hats to the treetop. Then he I believe that the assumption that remembered his grandfather’s story, so he rationality is common knowledge is the threw his own hat to the ground. But, source of the confl ict between logic and

Suppose you and I are two of these smart, ruthless players. to it. If both players follow this meta-rational course, both What might go through our minds? I expect you to play a large will do well. The idea of behavior generated by rationally re- number—say, one in the range from 90 to 99. Then I should not jecting rational behavior is a hard one to formalize. But in it play 99, because whichever of those numbers you play, my lies the step that will have to be taken in the future to solve choosing 98 would be as good or better for me. But if you are the paradoxes of rationality that plague game theory and are working from the same knowledge of ruthless human behavior codifi ed in Traveler’s Dilemma. as I am and following the same logic, you will also scratch 99 as a choice—and by the kind of reasoning that would have made MORE TO EXPLORE Lucy and Pete choose 2, we quickly eliminate every number On the Nonexistence of a Rationality Definition for Extensive from 90 to 99. So it is not possible to make the set of “large num- Games. Kaushik Basu in International Journal of Game Theory, Vol. 19, pages 33–44; 1990. bers that ruthless people might logically choose” a well-defi ned The Traveler’s Dilemma: Paradoxes of Rationality in Game Theory. one, and we have entered the philosophically hard terrain of Kaushik Basu in American Economic Review, Vol. 84, No. 2, trying to apply reason to inherently ill-defi ned premises. pages 391–395; May 1994. If I were to play this game, I would say to myself: “Forget Anomalous Behavior in a Traveler’s Dilemma? C. Monica Capra et al. in game-theoretic logic. I will play a large number (perhaps 95), American Economic Review, Vol. 89, No. 3, pages 678–690; June 1999. and I know my opponent will play something similar and The Logic of Backwards Inductions. G. Priest in Economics and Philosophy, Vol. 16, No. 2, pages 267–285; 2000. both of us will ignore the rational argument that the next Experts Playing the Traveler’s Dilemma. Tilman Becker et al. Working smaller number would be better than whatever number we Paper 252, Institute for Economics, Hohenheim University, 2005. choose.” What is interesting is that this rejection of formal Instinctive and Cognitive Reasoning. Ariel Rubinstein. Available at

MATT COLLINS MATT rationality and logic has a kind of meta-rationality attached arielrubinstein.tau.ac.il/papers/Response.pdf

CHARACTER RECOGNITION The Write Type

Electronically scan a book to import its content into EXISTING TYPE is recognized by software that masks out stray marks, a word-processing program. Save a snippet handwrit- frames a word in a grid and segments it into letters. An algorithm then ten on a personal digital assistant (PDA) screen into a analyzes the features of each letter (common features are shown at spreadsheet. Decipher a scrawled form or the zip code right). In each case, the software compares the feature vector—a set of on an envelope. In all these cases, software translates numbers—against a table of values that indicate a likely character, punctuation mark or clutter (such as a smudge). This “offl ine” software, typed or handwritten characters into digital text that as it is called, weighs the results of different feature vectors to can be edited, e-mailed, stored or used to tell a high- determine the fi nal letter. Postprocessing reassembles characters into speed machine which direction to route a letter. words in a word-processing program. That software was originally known as optical character recognition; today the term refers just to rec- ognizing text from a typeset page. Analyzing printed or cursive handwriting is called intelligent character recognition. Regardless of labels, the programs rely on similar algorithms to assess the features of an inkblot [see illustrations]. The programs then compare the blot’s features against mathematical models to determine which letter or number it most closely resembles. Determining characters handwritten with a pen on a PDA is perhaps the easiest task, because the pen or screen can track the stylus’s movement. Analyzing type or handwriting on a printed page is tougher “because you must extract a signal from a static image, clean clutter, then discern letters,” says David Doermann, co- director of the University of Maryland’s Laboratory for Language and Media Processing. Translating uncon- WORDS are segmented strained cursive remains most elusive, and, Doermann into letters. says, “it is not being done commercially.” Years ago character-recognition techniques relied on one or two algorithms that compared simple patterns; current software exploits multiple algorithms and then weighs, or votes on, their results to make a fi nal determination. Furthermore, “the old software took several minutes to convert a page; new software takes only a few seconds,” says Allan Stratton, a technical director at Nuance Communications in Burling- ton, Mass., which makes OmniPage, a document conversion program. Extensive research is under way to devise systems for Arabic, Japanese and Chinese characters, in which strokes can represent whole words, not just letters. Recognition software is also beginning to be included SCANNER captures a in digital cameras and cell phones. “Just snap a photo page of typed text. of a document,” Stratton imagines, “then stream the text as an e-mail or text message.” —Mark Fischetti

96 SCIENTIFIC AMERICAN JUNE 2007 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. 5W INFOGRAPHICS; INFOGRAPHICS; 5W www.sciam.com 2 0 6 0%6 0% 6% 100% 76% 60% 12% %50% 0% 75% 50% 0% 0% 0% 75% 0% 25% 50% 100% 18% 0% 0% 12% 80% 0% 56% 0% 0% 0% 0% 0% 0% 6% LETTER features are determined. SOURCES: NUANCE IMAGING ( IMAGING NUANCE SOURCES: DID YOU KNOW... automate code breaking. In 1965 the U.S. Postal Service began to to began Service Postal U.S. the 1965 In breaking. code automate Newton PDA and the IBM ThinkPad notebook computer brought brought computer notebook ThinkPad IBM the and PDA Newton to 1950s early the in technology the tried Agency Security Forces handwriting recognition to the masses, although commercial suc- commercial although masses, the to recognition handwriting Apple the 1990s early the In way. this forms of millions process es cess was marginal. Bureau Census today the U.S. zip scan codes; electronically Armed ﬁ U.S. been The have 1929. since led recognition character on a small electronic pad by a cash register may think their signa- their think may register acash by pad electronic asmall on NOT AT HOME DEPOT: Customers who sign for a credit-card purchase acredit-card for sign who Customers DEPOT: HOME AT NOT 6 88% 76% LONG RECOGNIZED: European and American patents for optical optical for patents and American European RECOGNIZED: LONG letter n ); VISION OBJECTS ( Measure curvatures Measure within a cell pixels white and black of ratio Calculate METHOD: ALGORITHM: FEATURE: METHOD: ALGORITHM: FEATURE: METHOD: METHOD: ALGORITHM: facing are Direction curves FEATURE: inﬂ ection points and end points Plot of curves andNumber length Color Contour tracing matching Matrix Contour tracing Visit http://avaxhome.ru/blogs/vietmagzformore.. COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. AMERICAN, SCIENTIFIC 2007 COPYRIGHT

handwritten “ink” Pen down ) Pen up

reference tables that indicate the most likely character. likely most the indicate that tables reference against strings the compare and numbers of a string into data the translate can algorithms time, of a function as recorded are strokes and points the Because lift. pen-up subsequent the and contact apen-down between points of aset defi as that astroke nes software online so-called LETTERS HANDWRITTEN against reference tables. One mainstream classifi One tables. mainstream reference er against is NN (nearest to fi nd the closest match. Another classifi er is HMM (hidden Markov Markov classifi (hidden HMM is er Another match. fito closest the nd their findings intoclassifi the results compare that er programs the University of Maryland. Next time, he proposes, draw something simply retailers Most not. probably is it but analyzed, being is ture probability distribution of variables that make that up variables of a feature. distribution probability model), fi appliedrst to speech recognition, which the assesses table entry every against the values compares neighbor), which store the whole image “just to save paper,” says David Doermann of of Doermann David says paper,” save to “just image whole the store completely different and see if a clerk or machine questions you. questions machine or aclerk if see and different completely CLASS: The algorithms that analyze character features feed feed features character analyze that algorithms The CLASS: Send [email protected] Topic suggestedbyreaderMorton Nadler. are recognized in near real time by SCIENTIFIC AMERICAN SCIENTIFIC movement of the of stylus. movement and horizontal vertical PEN or tablet records the the records tablet or Time Stroke Point

97 ADVERTISEMENT Visit http://avaxhome.ru/blogs/vietmagz for more.. REVIEWS

The Editors Recommend

good old dirt has always been necessary UNCERTAINTY: EINSTEIN, HEISENBERG, to sustain civilizations, from ancient BOHR, AND THE STRUGGLE FOR THE times right on through today’s digital soci- SOUL OF SCIENCE ety. This natural resource is being by David Lindley exhausted at a faster rate than it is being Doubleday, 2007 ($26) replenished, however. In the past century Lindley, an astrophysicist- the effects of long-term soil erosion were turned-writer, charts the course masked by bringing new land under culti- of Werner Heisenberg’s uncer- vation and developing fertilizers, pesti- tainty principle. The culmination cides and crop varieties. But as the popu- of Heisenberg’s equally perplex- lation continues to grow and the arable ing quantum theory, the uncer- land base continues to shrink, Montgom- tainty principle posited that in ery believes that society must rethink its many physical measurements, one can relationship with the land. Governments extract one bit of information only at the and farmers must rely not only on techno- price of losing another. Heisenberg’s men- logical sophistication to protect the soil tor Niels Bohr agreed with the basic prem- but on intergenerational land stewardship ises of his startling insights but saw the and conservation. “Civiliza- need to “make sense of the new quantum tion’s survival depends on physics without throwing overboard the treating soil as an invest- hard-won successes of the previous era.” ment,” he says, “as a valu- The third voice in this argument was Albert able inheritance rather than Einstein, to whom Heisenberg’s ideas were a commodity—as something a “monstrous misrepresentation ... the other than dirt.” very idea of a true fact seemed to crumble into an assortment of irreconcilable points FLOWER CONFIDENTIAL: THE GOOD, of view.” Eventually, and reluctantly, Ein- THE BAD, AND THE BEAUTIFUL stein conceded the technical correctness by Amy Stewart of the system that Heisenberg and Bohr Algonquin Books of Chapel Hill, 2007 had laid out, but he never fully accepted it. ($23.95) This story has been told before but seldom with such clarity and elegance. Rose breeders can spend seven years developing a new variety; one team of sci- DIRT: THE EROSION OF CIVILIZATIONS entists has been working for 10 years on by David R. Montgomery a blue rose, using a pigment gene from University of California Press, 2007 petunias. Making flowers flawless is a ($24.95) $40-billion industry. Per capita spending on cut ﬂ owers in the U.S. is about $25 a Montgomery, a geomorphologist at the year; the Swiss have the highest con- University of Washington, argues that sumption, at $100 per person annually.

The facts are surpris- ing and intriguing. But it is the way nature writer Stew- art packages them that makes Flower Confi dential the rare nonfiction book that keeps you turning pages. Her own passion for fl owers, along with her adventures in the fi elds, greenhouses, auction houses and laboratories, gives the facts life. An almost perfect book—its only blemish being the off-putting title that seems to promise more muck than this intelligent, evenhanded account rakes.

THE INVISIBLE SEX: UNCOVERING THE TRUE ROLES OF WOMEN IN PREHISTORY by J. M. Adovasio, Olga Soffer and Jake Page Smithsonian Books, 2007 ($26.95)

Funny how no one has to think twice about which sex is invisible. Anthropologists Adovasio and Soffer and science writer Page take the fi eld of archaeology to task for its traditional focus on hard artifacts such as stone tools and weapons, pointing out that archaeologists are not even trained to look for evidence of women’s use of more perishable artifacts such as string and netting. They argue for the central importance of the “fi ber revolution,” which began some 26,000 years ago in Eurasia. In dry caves and other rare places where artifacts do not deteriorate, fi ber and wood objects can account for 95 percent of all artifacts recovered. The influence of fi ber and its use in nets, baskets and clothing had “profound effects on human destiny—probably more profound than any advance in the technique of making spear points, knives, scrapers and other tools of stone.” www.sciam.com SCIENTIFIC AMERICAN 99 COPYRIGHT 2007 SCIENTIFIC AMERICAN, INC. Visit http://avaxhome.ru/blogs/vietmagz for more.. ANTIGRAVITY

Dumb Cup RECIPE FOR A STEAMING CUP OF SOMETHING BY STEVE MIRSKY

On a chilly, late March day I was happily sipping a Starbucks The Way I See It #13.5 billion “On one corner, there was a half-caf when I caught a glimpse of a friend’s cup and nar- Starbucks. And across the street from that Starbucks … was a rowly avoided performing a Danny Thomas–style spit take. Starbucks. At fi rst I thought the sun was playing tricks on my On the side of the paper cup was printed: eyes. But no. There was a Starbucks across from a Starbucks. And that, my friends, is the end of the universe.”—Lewis Black, The Way I See It #224 “Darwinism’s impact on traditional so- comedian and philosopher cial values has not been as benign as its advocates would like us ␲ to believe. Despite the efforts of its modern defenders to dis- The Way I See It #e i + 1 “The ratio of the diameter and cir- tance themselves from its baleful social consequences, Darwin- cumference is as fi ve fourths to four.”—Indiana House Bill ism’s connection with eugenics, abortion and racism is a matter No. 246, which would have legally declared the value of ␲ to of historical record. And the record is not pretty.”—Dr. Jona- be 3.2 in 1897 than Wells, biologist and author of The Politically Incorrect Guide to Darwinism and Intelligent Design The Way I See It #Venti “Would you drink a Quarter Pounder with Cheese? If you order a venti (20-oz.) Starbucks Caffè I knew that Starbucks roasted the hell out of their beans, Mocha, you might as well be sipping that 500-calorie burger but I didn’t realize they published half-baked ideas. through a straw. And a venti Starbucks Java Chip Frappuc- A visit to the Starbucks Web site turned cino, with 650 calories and nearly a day’s saturated fat, is a up an explanation: “To get people talk- McDonald’s coffee plus 11 creamers and 29 packets of sug- ing, ‘The Way I See It’ is a collection ar.”—Center for Science in the Public Interest press release, of thoughts, opinions and expres- September 5, 2006 sions provided by notable fi gures that now appear on our widely The Way I See It #34 Proof “I’ll quit coffee. It won’t be easy shared cups.” Further, the cups drinking my Baileys straight, but I’ll get used to it. It’ll still be are supposed to extend “the the best part of waking up.”—Megan Mullally as Karen Walk- coffeehouse culture—a way to er on the TV series Will & Grace promote open, respectful conversation among a wide The Way I See It #1962 “Mr. Kroger: two C’s, two D’s and an variety of individuals.” F. That’s a 1.2 grade average. Congratulations, Kroger. You’re Fair enough, although an at the top of the Delta pledge class.”—Dean Vernon Wormer open, respectful conversation initiated by a closed, disrespectful The Way I See It #MVD “This coffee tastes like mud!” “It assertion is going to be a challenge, should—it was ground this morning.”—encounter between especially without any context. (To fi nd Ricochet Rabbit and Deputy Droop-a-long in the 1960s, al- some context, see, among myriad sources, though no doubt in existence in some prior form Ernst Mayr’s essay “Darwin’s Infl uence on Modern Thought” in the July 2000 Scientifi c American and Michael Shermer’s The Way I See It #Too “Popular, palatable views of the world 2006 book Why Darwin Matters.) and how it came to be do not constitute science or truth. But Nevertheless, I’d like to suggest some other quotes for decent science education requires that we share the truth we Starbucks cups in the hopes that they, too, may stimulate pip- fi n d —whether or not we like it.”—Lynn Margulis, Distin-

ing-hot conversations. guished Professor, University of Massachusetts Amherst VERONSKY FRANK BY PHOTOGRAPH COLLINS; MATT BY ILLUSTRATION

How do itches come about, and why does it feel good to scratch them? Q —B. Ericsson, Sweden Mark A. W. Andrews, associate professor of physiology at tor fi ring “turns off” the itch sensation for a short period. the Lake Erie College of Osteopathic Medicine, replies: Although it is helpful in relieving an itch, scratching offers An itch, also known as pruritus, arises from the irritation only temporary relief and may cause the skin to become fur- of nerve cells associated with the skin. Pruritus serves as an ther irritated and possibly even to tear. important sensory and self-protective mechanism—as do Despite approximately a century of research, no single ef- many other skin sensations—by alerting us to harmful exter- fective antipruritic treatment exists. Several topical and oral- nal agents. Constant itching, however, can become unbear- ly administered agents can help suppress itching, however. able if the underlying condition is not treated. These include lotions and creams (such as calamine and hy- Pruritus is a dominant symptom of many skin diseases and drocortisone), antihistamines, opioid antagonists (such as also occurs in some ailments that affect the entire body. An itch- naltrexone, a drug used to treat narcotic or alcohol depen- ing sensation results from the stimulation of pruriceptors —itch- dence), aspirin and ultraviolet light therapy. sensing nerve endings—by certain mediators. These stimulating agents include chemicals for immune response (such as histamines) and pain relief (such as opioids); neuropeptides, which How did the sun wind up in the include endorphins and other middle of the solar system? pain-regulating messengers re- Q A. Sommers, East Brunswick, N.J. leased within the brain; the neu- — rotransmitters acetylcholine and Michael A. Jura, an astrophysicist at the University of serotonin, which relay impulses California, Los Angeles, explains: between nerve cells; and prosta- The best model of our solar system’s history states that the glandins—lipids that, among planets formed from a spinning disk of dust encircling the sun, other functions, create the sensa- leaving a collection of bodies with the sun at its center. tion of pain in spinal nerve cells. According to this model, the solar system formed from the Stimulation by any of these agents collapse and fl attening of an interstellar cloud. The cloud may is typically related to infl amma- initially have measured as much as a light-year across—10 tion, dryness or other damage to million times wider than the sun. As the cloud compacted and the skin, mucous membranes or conjunctiva of the eye. cooled, its own gravity overpowered any forces acting to sta- Itching generally involves activation of the pruriceptors of bilize the system, causing it to further contract dramatically. specialized nerve cells called C-fi bers. These C-fi bers are Before this collapse, the original cloud probably began identical to those associated with the sensation of pain, but with a fi xed mass and a slight, random rotation relative to they are functionally distinct and only convey the itch sensa- some central axis. Most of the cloud’s mass helped to form tion. When stimulated on the skin, the C-fi bers carry signals the sun, but some of it fl attened out and remained as a disk along the nerve to the spinal cord and on to the brain. encircling that newly created star. Observations from else- Scratching and rubbing interfere with the sensations aris- where in the galaxy indicate that this disk most likely gave ing from pruriceptors by stimulating various pain and touch birth to Earth and the other planets, leaving them in natu- receptors in the same areas. Like many sensory systems in the rally heliocentric orbits. body, activation of one signal, in this case that of the pain and touch receptors, causes “surround inhibition” of another sig- For a complete text of these and other answers from

nal, that coming from the pruriceptor. This lack of pruricep- scientists in diverse ﬁ elds, visit www.sciam.com/askexpert COLLINS MATT