e e.

The G E & Science Prize for Young Life Scientists. Because brilliant ideas build better realities.

Imagine standing on the podium at the Grand Hotel in Stockholm. making your acceptance speech for the GE & Science Prize for Young Life Scientists. Imagine having your essay read by your peers around the world. Imaging discussing your work ina sem i nor with other prize wi n ners and Nobel La ureotes. I magi ne what you cou Id do with the $ 25,000 prize money. Now stop imagining_ If you were awarded your Ph.D. in molecular biology in .2009, then submit your 1000-word essay by August 1. and you can make it reality

Wont to build a better reality? Go to wwwgescienceprize_org

GE & Science Prize for Young Life Scientists

... For the P urpose of this prize, melee ulor b iolOCl...,Y is delined tIS ·-1l101 port "I' bioloW which oueropts to itllerprel bialagieel events it) terms or the ph)tsOoochemiccl properties. of molecules in CI CGII·

(M<Grow-Hill Oiclionory ofSoiemificond Technical Term,. 4,h EdiUon).

GE H..,", sio-scences AB,

Bjork(lOWn 30. 7,)18~ Upps:o'o. Swede.n.

© .2010 G. nero' ~I ectric Company

- All rfghts re se rved


im.agination at work

R&D Systems Bioactive Proteins High quality proteins aren't a luxury, they are a necessity.__'"

WHAT'S THE RISK? '" Missed opportunities '" Non-specific results

'" Experiments that can't be repeated .I Weeks or months of wasted time

LISA &: Canada R&D Systems, Inc. Tel: (800) 343·7475 info@RnDSystems . com

Eutope R&D Systems Europe, Ltd. Tel: +44 (0)1235529449 info@RnOSy'tem,.( C;hin3 R&D Systems China co, ltd. Tel: (800) 988-1270,CIl








For resear{h use only. Not for use in diagnostic procedures.

R&D Systems has spent almost 25 years buBding its reputation as a source for high quality proteins.

Every stage of protein development takes place in R&D Systems' laboratories, from cloning of the gene, to protei n purification and testing for bioactivity. Because we control all aspects of protein rnanutactu ring, R&D Systems can better control the quality of ou r products and the technical assistance we offer. Please visit our website at for more information.

Ca!lcer Development Endocrinology Glycobiology Immunology Neuroscience Proteases Signal Transduction Stem Cells

R&D Systems Tools for Cell Biology Research 1M



R&D Systems recombinant mouse Wnt-3a (Catalog # 1324-WN) and Wnt-5a (Catalog # 645- WN) promote stress fiber formation in NIH-3T3 cells, while only WnHa promotes nuclear IHatenin accumulation. Please visit our website for information about our new high purity human Wnt-3a (Catalog # S036-WNPl.lmages (ourtesyofDr. Raymond Habas, Robert Wood Johnson School of Medidne.

Selection expanding weekly-visit;uest to sign up for weekly new product updates.

GE Healthcare Life Sciences

Cell analysis just got easier

INtroducing the new IN Cell Analyzer 2000

The new IN Cell Analyzer 2000 makes previously challenging cell analysis a thing

of the past by offering significant throughput gains compared to typical lamp based imaging systems, while still delivering enabling features for your research assays.

From organelles to cells, to tissues and whole organisms - from fixed end-point assays to extended live cell studies - from manual microscopy to automated high-content screening, IN Cell Analyzer 2000 offers you the performance you need for all your screening and research ~ making your cell analysis faster and easier.

At GE Healthcare Life Sciences, our focus is on helping scientists achieve even more, faster. It's a commitment we have in our genes. And all this is backed by the service, support, and investment for the future that being part of GE can bring.

Cell analysis just got easier. Find out more. Visit www.gelifesciences.comlincell

I AKTA I Amersham I Biacore I IN Cell.Analysi.s I Whatman I GE Service I

imagination at work

AKTA, Amer.hom, Biocore ond whotmon ore trode morks 01 G E Heol thcore com pMie<.

© ~009 Gene rol ,Ioo(rio Com po ny - All rights reserved. Firs( pubfshed Moy 2009

GE t-teo I(hea re Bio·Science. AB. Bj5rkgoton 30. 7518~ Upp'QIQ, Sweden



, 429 S dance-Based Hea lth Ca re ChmZ11lI


, 434 Budget Red Tape in Europe Brings New Delay to ITER

1435 Silicon Mystery End ures

in Solved Anth rax Case

1437 Beyond DSM: Seeking a Brai n- Based Classification of Mental Illness

1438 Budget Shortfall Could Derail Plans for Underground Lab

1438 MIT Engineering Dean Tapped to Head NSF 1439 Hardy Cotton·Mu nrhin g Pests

Are Latest Blow toGM Crops

1440 China Amasses War Chestto Confront Its Environmental Nightmares

Fresh Momentum for China's Science Juggernaut

1441 From Science's Online Daily N'ews Site 1442 Polish Science Reforms Bring Fear and Hope 1443 Ma le Rivalry Extends to Sperm in

Female Reproductive Tract

1443 From the Science Policy Blog

Chilean Scientists Rally After Qua ke G. Dumenil

Salmon Swimming Against Multiple Threats 8. Wright

PIa nni ng for Biodiversity in Future Climates L. P_ Shoo


0. Venter et al,

15 the Message from Athens Being Heard? R. Merken et al.



1455 The Fourth Paradigm

I. Hey, 5. rall5/ey, K. Tolle, tas, reviewed by l- P_ Collin>

1456 Delete

lC Mayer·Schonberger, reviewed by W. Dutton


1457 China, India, and the Environment K: S. Bowa et al,


1460 SeW n9 the Ira p for Reactive Resonances S C A/t/wrpe

»Report p. 1501

1461 Pheflology Under Global Warming

pag.e 1444

NEWS FOCUS C Kiuner andD. Basler

, 444 The Nile Delta's Sinking Future > > Sci ence Podcast

1448 Lunar and Planeta ry Science Conte ranee Lucky Glimpses of a Weirdly Welter Moon Coaxinq Out Another Taste 01 the Sun

Spirit I, Willing, Though Weak

Sna p shots From the Mee ti I1g


1451 U nlockjng the Door to Better Cybersecurity A. K.Y. Wong


w. A. Wulf and A. K. tones


1462 Controlling Turbulence B.]. McKeon

»z-Repor: p. 1491

1463 AMPA Receptors-Another Twist?

M. Farrant and 5. G_ (ul/·Candy »Report p_ 1518

1466 Toward Organic AIL·Optical Switching S A. Haque and ,. Nelson

» Report p. 1485

1467 Fairness in Modern Society K. Hoff

»s-Researd: Article p. 1480

1468 EXPil nding the Repe rtoire of Shape Memory Alloys

1- Mo and I. Kamman »Reportp.1488

page. 1461

CONTENTS continued»


1425 This Week in Science 1430 Edit.ors' Choice

1432 Science- Staff

1433 Random Samples 1527 New Products

1528 Science Careers

Saturn and its rings as seen by the Cassini orbiter. Saturn's rings contain innumerable chunks of water ice polluted by material

of varyi I1g color and composition. The massive B ring (bottom) and the tenuo us C ring (closest to the planet) are d omi nated

by unexplained structu re. Satu rn 's yellowish appeara nee

sig nifies a deep cloud layer beneath seasonal haze.

See pages 1470 and 1476.


www.sciencernaq.orq SCIEN CE VOL 327 19 MARCH 2010


Detection platforms, assays, and analysis software


Rotor-Gene Q

Use QIAGEN® solutions from sample to result,

and benefit from sensitive and reliable detection systems:

• Quantitative, real-time peR detection

• Automated analysis of DNA fragments and RNA

• Pyrosequenting® sequence-based DNA detection and quantification

• Optimized, reedy-to-use assays and reagents

Making improvements in life possible -

Sample & Assay Technologies


1470 An Evolving View of Satu rn' s Dynamic Rings

). N. Clizzi et al.

:» Revfew p. 1476

1476 Saturn: Atmosphere, Ionosphere, and Magnetosphere

T. I. Gotnbosi and A. P. Illgerso{{ » Review p. 1470


1480 Malkets, Religion, Community Size, and the Evolution of Fai rness and Punish ment ]. Henrich et at.

The origins of mad ern soda.1 norms and bell avior s rna y be foun dill the evolution of institutiens,

» Perspecti'le p. 1467


1485 Desig n 01 Polymethi ne Dyes with Large Third-Order Optical Nonlinearities and Loss Figures of Me rit

J. M. Hale, et al.

NonliJle<lr optkal materials are designed and characterized tor poten.tial applications in all-optical ~v.itchillg.

:>:> Perspective p. 1466

Ferrous PolycrystaLline Shape.-Memory ALLoy Showi ng Huge S uperelasfidty . ~ ral)aka et at.

A shape-memory attoy has been prepared with high mec.hanical energy absorpfioncapability and reversi blemall n etizati 011 ch an 9 e.

:» Per!>perUve p. 1468

w 1491
~ Eliminating Turbulence in Spatia tty I ntermitrent Flows 8. Holel al

" z

"'" E


12 9


" o

~ 1495

z -c ,,; o


" 5

o ur e: -c


1:1 1497 The CLimatic $ignatuJe 01

t;; Incised River Meanders

d 'i:



is Typ h 0 on Ire q lie ney and bed rotk stre n gth

i' inftllence river meandering in mountain

"'i environments.

'" w





::< o ~


'nj ectio not j e Is of wa ler is use d to centro 1 the transition from laminar 10 turbulent 1Iow in pipes.

» Perrpertive p. 1462

Microcavity Lase r OsciLLating in a Cirruit-B ased Resonator

C. Walther et al,

An ullrasmaillaser is labricilted from

conventional electronic components combined with an amplifying medium.

C. P. Starkel a I.

1501 Transition-State Spectroscopy of Pa rtial Wave Resonances in the F + HO Reaction w. Dong et al.

Spec trosce py ca n di sti ng u i sll Ih e rea ction paths in a ccllision between an atom and a diatomic sys te m.

>:> Perspective p. 1460

1502 Mechanosensitiv'e Self- Replication Driven by Self-Organization

J. M. A. Cat nail et al,

Tne type of mechanical agitatiDn applied to a solution influences which of two mole(uiar products dominate.

1506 Semina! FLuid Mediates Eja culata Competition in Socia L Insects

s. P. A. den Boer et a I.

Substances produced by male social insects destroy sperm., and females produce compounds 10 counteract sperm loss.

1509 Patterns of Diversity in Marine PhytopLa nkton A. D. Borton et al.

Highest diversity occurs in pl1ysicaUy dynamic mid-latitude zones, and lowest diversity and highest biomass occur toward tile poles.

1512 Unicellular Cyanobacterial Distributions Broade n the Oceanic N2 Fixation Domai n P H. Moi5ander et al

Nitrogen fi;i;ation in the South Pacific Ocean is partitioned among several microbespecies with distinct ecophysiologies.

1515 A Critical RoLe for a..4B5 GABAA Receptors in Shaping l.earni ng Deficits at Puberty in Mice

H. Shen et al,

Learning incapacity observed during puberty is related to receptor tocatlon ill the hippocampus.

» Science ?odcast

1518 CKAMP44: A B rain-Spedfic Protein Attenuating' Short-Term Synaptic PLasticity in the Dentate Gyrus

]. von Engel/)ardt et ~L

A synaptic protein thallegulates postsynaptic MAPA receptor responses has been cloned and functionally characterized.

,.,. Perspective p. 1463

1522 Circa dian Gating of the CeLL Cycle Revea Led in SingLe Cyanobacterial Cells o. Yilllg et al.

Modeling and observation of cy.anobacteria show e ntr a in me nt ofth e celi cycle by thei r biologkal clock,

CONTENTS continued»

www.sciencernaq.orq SCIEN CE VOL 327 19 MARCH 2010


pages 1462 & 1491

page 1495



page 1509



Take the direct route from sample to results

Finnzymes' Direct peR approach saves you. time and cost by allowing amplification of DNA directly from the source material. No DNA purification is needed. Direct peR is suitable for various kinds of sample materials such as plant and animal tissues, blood, and FFPE tissue samples.

Direct peR is based on Finnzymes' unique peR enzymes, Phusion'" High-Fidelity and Phire'" Hot Start DNA Polymerases. These polymerases are exceptionally tolerant of many peR inhibitors. To achieve the shortest possible protocols, combine the Direct peR approach with Finnzymes' Piko? Thermal Cyclers and UTW'" reaction vessels.

Dilution protocol


Oi I utlon Buffer




Direct protocol


peR reaction


See the latest results, application notes and product updates at

Reagent' dlstributedln US and Canada by New England Blolabs ",,-,'..L..II.A,..;.u Int: For other products and countries visit "",~" ... m:ll'r~lalll~"f<!tI<I>I ..


Ot' .

P Imrzed kits available:

Phire® Plant O·

. . .. Ireet peR Kit

for plant material

Phire® Anima .. ' r· .

ISSue O·

for various a . 'feet PCR Kif

·n1mal tissues

Phusion® BlOod O·

Ireel PCR Kif for blood







Functional Hierarchy and Reversibility Within the Murine Spermatogenic Stem Cell Compartment

I. Nakagawa et at.

Developmental flexihility within a

s tern cell sys !em u nde rpin 5 the robu 51 m .. i n tena nee of spe em ala !lene sis,

10 _1126Iscience_U82868

Resolving Mechanisms of Competitive Fertilization Success in Drosophila melanogaster

M. K. Monier et al.

Huorescently labeled sperm allow direct visualizatiun ul their activit.y within the female reproductive tract ot flies. 10.1126Iscieoce.1187096

Va rialion in Transcription Factor Binding Among Humans

M. Kosowski et al,

Trail sc ripti on fa etc r bi nd i IIg ~ tes val)' arne ng individuals and are correlated with differences in expression.


H erila ble Individ ual -Spadfic and Allele-S pecific Chromati n Signatu res in Humans

R. McDal1ie/l et at.

An appreciable amount of variation in chromatin status and transcription fador binding has .. genetic basi 5 .•


Three-Dimensional Invisibility Cloak

at Optical Wavelengths

r Elgil1 et al,

A structured photonic uystaL can be used to do ak an 0 bj ect a t aptic~ I wave leng tlls and over a wide viewing a.ngle. 10.1126/science.1186351

>:> Science Podcost


Highlights From Our Daily News Coverage

Carbon-Capture Method Could Poison Oceans Fertilizing algae with iron boosts levels

of harmful neurotoxin,

csr, Latest Clue------B acteria

Unique skin microbes might allGw identilicatiol> of criminals.

Psychopaths Keep Their Eyes on the Prize ~ The brain's overreaction to rewards may lead ~ to i mputsive, en tis 0 cia! beha vio r;







"' l:




"' u


The Signal Transduction Knowledge Environment

RESEARCH ARTKLE: Cholinergic Augmentation

of Insulin Release Requires Ankyrin-B

J. A. Healy et at.

SCIENCESIGNALING Pi ckin g the ri ght path.

RESEARCH ARTICLE: Mutations in Fibrillin-l Cause Congenital Srleroderrna+Stilt Skin Syndrome

B .. L. ioeys et at

PERSPEGIVE: Stiff and Tight Skin-A Rear Window into Fibrosis Without Inflammation E. V. Avvedlm~nto and A. Ga/Jri~1/i

Mutations il' thl!nbrillin· 1 gene that cause a rara congellital form of scleroderma shed light on

.. cornmcn fibroti( condition.


v. Bennett and A. M. VanHook

Redu clio n of an ky r in· B acti vi ty 1m pa i rs maxi mal insulin release and isa risk fador for diabetes.

RESEARCH ARTICLE: !nferring Signaling Pathway Topologies from Multiple Perturbatio n Measurements of Specific Bioche mical Species [,R. Xu et al.

B aye§;a 11 i flfen nc e- based mod e H IIg i del'tifi es the most likely p .. ths fuwugh a sigllaling network.

RESEARCH ARTICLE: A Lcss-of-Fu nction Screen Reveals Ras- and Raj-Independent MEK·ERK Signaling During Chlamydia troctiomatis Infection R. K. Gututnurthy et at.

A siRNA-based screen of host tactors thai influence in Ie etten by Chlam ytfia revea Is the dec ou pli ng ot the canonic a l Ras- ERK signalillg pathway.

SCI E NC ECAREE RS Free Career Resources for Scientists

Mind Matters: Back to Work After a Baby 1.5. Levine

Dro ppin!l 0 ff <1[1 i nf a nl an d retumin!l to work is stre ssf lit fur new mothers,

Tooling Up: I'm Special Aren't You? D. Jensen

Your parents may say you're sped a l.

but the job market requir~ some humility.

Going Home 10 a Changing Poland E. Pain

Reform~ in Potalld gille molecutar Agllieszka Cllacinska hopes for tile future, :>:>News story p. j 442

SCIENCETRANSLATIONAL MEDICINE www.sciencetranslationalmedicine.ofg Integrati n9 Medicine and Science

PERSPEGIVE: Finding New Cures for Neurological Disorders-A Possible Fring.e Benefil of Biodefense Research?

D. A. Jett

Drugs de veto pe d to trea t ell p !lSU re to I oxi r ch",mica ls might help cure condition, such as stroh, epilepsy, a nd neurudege!lerative diseases.

RESEARCH ARTICLE: Targeting Rob04·Dependent Slit Signaling to Survive the Cytokine Storm in Sepsis and Influenza

N. R. London et al,

Bluoting ihe increased Vascular permeability caused by ;m rnune res pon ses t o S lit li !land i ncreas es survi v~l ; n models of 5 epsis all d vi ra I infect; on.


[)OWIl toad Hie 19 March S ctence Pod cast to Ilea r about tile biology behind leaming defidts al puberty, a three-dimensional invisibility cloak, and U1 e N ile della's 5; nki ng future.

SCI EN CEI NSIDER Science Policy News and Analysis

S-C1ENU {lSSN (UUHU1:J) 1:5: pubUshed weekJy on Frica)lj- '~(~1 tht I.rn ItIIHk in D ecemtrer; by th~ AIt'i~r[Gl.r.. A.uor~atl.on for the A IiV a ncemene 01:11 5d enee, lZl10 Ni!)1I'f vork "''iI,~ II!?, ~W. W~~hitl Ilton, I) ( :1:(11)0 5. Periodk~l$ ~ I pa$1;il9~ (Pii blil:atio II No. 4.a .... 1I00) ~id at Wa:s:hill stc II, bC, and add itio na I ~i li 1'!9 oU-i~~_ CoP1r1ght O~Ql0 b1theArner~1I ~i:J,tioii forllieltd\!cll1~lMrr!; cf Science. 'the rille SCIE NC~ h: .a. re.g i$'t;::~ lr~;::mark a f the AAAS. Domertii;. il1di'Liid ~l membershi p and SIJb!<tiptlol'l {Sol iS$lIes): 51116 U1t1 allaw=.t«l to :$I,!bSi;.r1ptioll)_ Dame~ii;: il1$titl,ll·iaiial SIIMcrip'ioll t51 iz;1I~s): :;"910; F-oroigll postage carJ: Maico. C;aribbNll (sllrf.ocC I"Mil) :S.5S·; cjhcr eeu mncs t1i r esstst dclivcry) $0&5, Fitst cross. .1irm.1i I. student, end emeritus rates 0 n request G1llild lan rates wil h GST .1Vi:IilJble upon re-quest. GST ft2 511 8,911_2 .. P'tJ bliwtiOils M_.)iI Agre-cmCllt NumlJ:.c.r 10:6'9'6211, Pnl'itei1 ill the U.S..A_.

O! iU)Q e of add(·eS1: Allow 11 ~Ks, givillg old end Ilew ood~s and .a-digit ecccent PMtrnaner: 'S6Id dl.1l1ge of oodross t:o MAS, P.O. Bo:( '9'-617S, ~"VilstlillglGi\ DC 200'9"-0=6173. 5ingl~,o{{Jopy 1~·[~1: $010,00 cnrcnt Issue, :S15,()0 back issue prepaid mdudes SlJrfLiCC porng.~; bulk: rates all eqeest,"il.atim to photoco:py material (Of 'O[ perwl1<;JllI~ tinder dl1)J~~nal fulling wlthill the fair lise pro""i5iQns ar 1f1~ (OP'fright Act is granted by MAS to libraries end ather I,!$ft"$ ~~~ wi1h the Co Wrig In (l~ra!1c:e (el1l:er (('(0 lral1~diOI1.!!1 Repqortillg Service, provided that 5-;W.(lO ~r ~rtide is po;iid directl~ to (C(, 22~ RQ~ Dn~, Deneers, 1M 01"94)_ Ibe iIIenlifi(Jtiorl (oo.c. rOI" Sdma" i:s. 0036-0075, Sd~ce is illdc-lftI ill the Reatk!"'s GuUk 10 PmOdt61 Utef(JWre and irl several spcdalirod illde:x1!S.




Reagents for Sample Preparation from New England Biolabs

lnrroducing NEBNext''', a series of highly pure reagems that facilitate sample preparation for downstream applications such as next generation sequencing and expression library construction. Available in sets, master mixes and modules, these robust reagents undergo stringent quality co Ilt rols and fu nctioual validation, ensu ri ng maximum y iel d, conven ience and value.

Sequencing coverage m.ap af the E. coli gena me after using NEBNext'M DNA Sample Prep Reagent Set I for Sample Preparation





Genome Posltl on

E. roli sfrJin MG 1665 gDNA IWS p!epareiJ 'llilll NE8NexI DNA Sample Prep Reagffll Set I and Se!]UerJa!iJ on an /IIumina Genome At1illyzer /I.

for more information about NEBNext, including customized solutions, please contact


J & rca

Now available:

NEBNext'~ dsDNA Fragmentase"" an enzyme· based solution for the fragmentation of DNA



Saturn IS Secrets Probed

The Cassin i spa tee raft Was La u nch ed on 15

Octo ber 1997. It to ok it almost 7 years to

reach Sat-urn, the seco nd-largest plan et j n th e so la r system. After almost 6 yea rs of observations of the serit;s of interactin g moons, ri nqs, and magnetospheric plasmas, known as the Kronian system, Cuzzi et ot. (p. 1470) review our current understandmq of Saturn', rings-the most extensiVe and complex in the solar system-and draw pa rallels with circu rnstelLa r dish. Gombosi and Ingersoll (p. 1476; see the cover) review what is known about Saturn's atmosphere, ionosphere, and magnetosphere.

A Fair Society

Many of the social interactions of everyday lite, esp eoa lty those i nvol vi ng econo mic exchan ge,. take place between individuals who are unrelated to each other and often do not know each other. Countless laboratory experiments have dow rnented the prope nsity of subjects to behave fairly in theseinte raction s and to punish those participants deemed to have beh aved unfai rly. Henrich et of. (p, 1480, see the Perspective by Hoff) measured fairness in thousands of individuals from 15 contern porary, small-s cale societies t a gain an u nderstandmq of the evolu tton of trustworthy exchange among human societies. FaIrness was qua ntitated using th ree economic games. Various societal parameters, such as the extent to which food was purchased versus produced, were a150 collected. Institutions, as represented by markets, community size, and adherence to a world religion all predict a greater exercise of fairness in social exchange.

At Sixes and Sevens

Motecu la r synt hesis an d mac roscopic ag 9 reqation have often been reg ard ed as entirely separate processes. From the reseaJ'(her's standpoint, once reagents have been mixed, synthe-

..; sis is Largely passive, whe reas processes such

~ .

I,;; as crystallization can be more actively manipu-

~ lated. Carnall et ai, (p, 1502) characteriaed an g -r

'" ..


0( ..


~ .-




~ unusual system in which the formation of

g aggregated cyclic macromolecules (macrog cycles) from small peptide-based building

~ blocks was governed by intimately interdepeng dent factors ra ngin g fro m the sea le of covalent ~ bond formation all the way to micron scale

5 fiber growth. As the macrocydes stacked

- .



Messy Mountain Mea nderi ng »

Predicti ng the influence of !

eli mate sn la ndsca pes is sometimes straiqhtforward: l~H exam pte, river deposits might !JfOW with inc reased rainfall beca use erosion rates and sediment transport increase. However, lonq-term tectonic p rocesses com plicate the geomorphic signatures of more gradual climate -related ph en am en a that reconfiqure landscapes. By co rrelating a d eca d sslong record of typhoon rainfall in Japan with digital elevation models, Stafk Elt at. (p. 1497) show that climate directly inftuences the extent of river meandering. When expa nded to a la rger region of th e western North Pacific, this analysis revea led it 5 trong climatic imprint on the landscape of humid mountainous areas. The region-wide analy-

sis alse revealed that underlying bedrock strength, as opposed to tectonic uplift, acts

as a secondary control,

ag ain st one an oth er to fo r m th e fi b ers, th ey remained loosely bonded enough internally to incorporate or expel individual building blocks. Varying the type of mechanical force applied to the growing fibers (either through shaking or stirring the solution), alternately favored formation of either 6-membered or 7-membered covalent macrocydes.

Ferrous Shape Memory Alloy

So-called shape memory alloys "remember" the sh ape they a re processed into, and ca n retu rn to that sha pe after bei ng deformed by heat. A lirnitation for most metal-based shape memory i!lloys 15 the extent to wh ich th ey ca n be

deformed el<Jslkalty. Tanaka et at. (p, 1488; see the Perspective by Ma and Karaman) demon strate a n iron -based a lloy that shows much higher levels of superelastic strain, surpassing the performance of nickel-titanium alloys. I n addition to high superetastic strain, this ferrous shape memory alloy has much

h ig her sire ngt h than NiTi a nd co pper-based shape memory alloys and, consequently, a highenergy absorption capability. These properties may alLow shape memory alloys to be exploited as strain sen sors or energy da rnpers.

Taming Turbulence

When fluid flows through a pipe, if the inertial

f 0 rc es are in creased 0 r the vi SC05 ity is d ecrea S ed, the fLow will become increasing noisy and will shift Irom being laminar to turbulent. Turbulence can be triggered by roughness in the pipe or other irreg ulantiss, which cause local eddies that grow into fu ll-scale disru ption of the otherwise smooth flow. Hof et al. (p, 1491; see the Perspective by McKeon) show that a continuous turbulent eddy, downstream, eliminates the growth of upstream distu rbanGes and ca n prevent the overall flow from becoming turbulent. Unlike many other controL methods, the energy cost for implementing this strategy is leSS than the benefit gained by maintaining a laminar flow.

Small Is Beautiful

Sh rin king the size of lasers is attractive beca use it generally leads to a reduction in power requirements, an increase in switching 5 peed, and possibLy a cleaner output. Walther et of. (p, 1495) combined patterned electronic components (inductor and capacitor) and an active gain material to develop a submillimeter laser that emits in the microwave reqirne at low temperature. The use of established patterning techniques and tunabLe

Continued on page 1427 SCIENCE VOL327 19 MARCH 2010


Taking Your Research to the Next-Level. ..

Next-Generation Genomics from Roche NimbleGen

The future of genomics is available today with tools flexible enough to conform to your specific research needs, Tools that let you discover your data, your way, and provide you with the most complete picture of your disease-associated research,

Only Roche NimbleGen provides you with the genomic content, high data quality, and array format flexibility you need to make next-generation genomics a reality,

Discover next-generation genomics

.. _


For life science research only, Not for use in diagnostic procedures, NI M BLEGE N is a trade mark of Roche

o ZQ 10 Roche N i rnt,leGen, I nc, All rig h ts reserved,

Roche NimbleGen, Inc, Madison, WI USA

Cvntinurd tram page 1425

This Week in Science

superlattice strurtu res offer the prospect of s hr\nking the Slle slill further. as well as providing a rou te to designer laser output for hiqh -speed intorrnation transport and optical processing.

Sperm Wars

Some female insects mate on only one day of their life, but then they may mate with multiple males and store the sperm, sometimes for years. But as the mates compete for mates, so their sperm compete fo r ova, an d competition between ejaru lares (an result in th £I destruction of sperm inside multiply mated females. But females need to select the sperm they wa,nt and to maintain stores of viable sperm to ensure a lifetime's fertility. Den Boer et al. (p, 1S06) compared species of bees and ants with queens that either mate once or mate multiple times, and found that sperm competition has driven the evolution 01 compounds in the male accessory gland that protect a male's own sperm while damaging another male's sperm. To counteract the male effect. queens produce compounds that mitigate sperm destruction and maximize the number of her offspring.

Diversity Gradients

latitudtnel gradients in species abundance, with relatively feW occu rring at the poles and many at the eq uster, are well known for macroorqanisrns. It is a matter of controversy, fueled by a lack of observational data, whether such gradients 01150 occur among microorganisms. Ba rton et 01. (p, 1509, published online 25 February) have built on a global marine qrculiltion model to predict the dynarntcs of phytoplankton populations. In silico, they obtain patterns of latitudinal gradation for plankton that are interspersed with hotspots of amplified diversi,ty, which point to plausible natu ral explanations for the phenomenon that can be tested in the future by "ystematic metagenomlc surveys.

Puberty Impairs Plasticity

While the existence of a period of reduced Learning coin cidi ng with th £I onset of pu berty in mice is well ch aracterized, the u nderlyin g cellula rand molecular mechanisms remain unclear. Shen et aL (p, 1515) assessed the role of specific y-

amin obutyric acid type A (GABAA) receptors fo r restrictin 9 hip pocampa I plasticity durin g puberty. At puberty, but not in adults or the very young, GABA receptors containing the a4 andS subunits we re targeted pe risyn aptically to excitatory synapses, shuntinq the depolarizing current necessary for N·methyl+aspartate (NMDA) receptor activation. As a conseque nee, 519 nat trans mission was affected and spatia I learnmq red uced,

Dancing with AMPARs

A type of transmembrane receptor for glutamate, known as AMPAR, mediates most of the fast excitatory transmission in the mammalian central nervous system. Their function is regulated by the (om' position of their su bu nits, posttra nslarion al mod ili(atl0n$, a nd protcin-p rot ei n i nterartions. Rece ntly, several protei ns that interact with AMPARs have been identified that affect t hei r su bcellular loca lilation, synaptic stabilization, and kinetics. Using proteornir analysis, immunohistochemistry, and elsetrophysiology,. von Engelhardt etal. (p, 1518, published online 25 february; see the Perspective by Farrant and Cull-Candy) identified a protein. CKAMP44, which modulates postsynaptic AMPA receptor gating, deactivation, and desentization,

Cycle Entrainment

Cells manage many cyclic processes that must coordinate with each other for best cellular performa nee. Ya ng et 01. (p. 1522) present a gen eral theo retical f ra mewo rk th at qu antitatively describes coupled cyclic processes and then apply this to the interaction between the circadian and celldivision cycles j n sin 9 le cyanobacteria. 51 mu ltan eou sly trackinq in divid ual ce IL division 5. and d rcadia n phases a nd fitting the data with the model su ggesl that cell-cycle p rog res sion slows down d rarnatica lly du rm 9 a speCific dread! a n in te rval, wh ereas (ell-cycle prog ression is in depen dent of th e cell-cycle phase. SCIENCE VOL 327 19 MARCH 2010


live cell vlsuallzatlon 'medium DMEMgfp

Photobleachlng of fluorescent orotelns in response to prolonged exposure 1.0 exc1!in,g radialia n sign i fi cantly affec ts tn. e ir lIti! it)' as .in vivo labels, New Evrogen DMEMI1.1p I,ive cell vlsual[zation medlul)1 slgnlflcently increases photostability of .green fluorescent proteins. Replacing the culture m edi um Wi lh DMEMgftJ for the period of yisualilatlol1 results in up to a 9·fold increase of photostebltrtyot EG,FP, a ·3.3- fold lnceease of pMtostablilty of TagGFP2 and more than a 4·[0Id increase of photostability of activated forms of pnotoacnvatabre PA·GFP and PS-CFP2

t.eng-term fluorescent microscopy of EOFP In lIye HEK293T cells matrualned in DMEM 0' DMEMl1P.




~ O,Q ~M



n ~-_--

o .w<l ~M



(I,u o.~


0.2 '---

G __ -----"=~-

o ~ .9Il 120

Influence of cell medium on PhOt05tilMi~Qf EGfP (A) and TegGfP2 .(8). Graphs SMW normal[~M bleach I ng cu rvas 0 r fluore see III prote I Min Bye HEK'293 ceus IllalnlalneClIn DM:EM or DMEMtll'.

Evrogen J5C, Moscow, Russia Till: + 7{ 495J 988 40 84 fa~: +7(495198840 S5 Web site:


Chen Zhu is the Minister of Hea lth nf the People's Repu blie of Chino.

Science-Based Health (are

THE PROVISION OF ADEQUATE HEALTH (ARE IS ONE OF THE MAJOR CHALLENGES FOR MODERN societies. It is an especially tough task in developing countries with limited resources and insufficient capacity. Obstacles a re even encountered at the conceptual level: For example, a traditional miscoucep tion is that spending on health is a social burden, instead of being a strategic investment essential for each nation's socioeconomic development. According to economic analyses, health system innovations will not Oldy improve livelihoods but also boost internal consumption and job opportunities .. But these innovations cannot succeed without tb.e strong support of science.

Both the biomedical and social sciences will be critical f01" developing sound policies that reshape health care systems. In China, with its fast-aging populace and a disease burden increase associated with urbanization and industrialization, science must drive \Ll1 evidencebased analysis of the cost-effectiveness of drugs and medical {cchnologies to enable effective and affordable prevention, diagnoses, and treatments. Science also facilitates the evaluation of the perf ormance of health care institutions to ensure quality services, And science can drive a nations 1 innovation strategy. Thus, the education of medical professionals, trai ning of regulatory teams, ·and fosteri ng of biotechnology talents can be leveraged through Internet-based platforms that r-each remote aIMS. Moreover, scienoe education raises public awareness of issues such as food and environmental safety, and healthy lifestyles and behaviors. Finally, through worldwide collaborations, science can meet critical global challenges such as HIY I AIDS and pandemic infl uenza, with many opportunities for South-North and South-South cooperation.

To fulfill these missions, visionary policies are needed to support capacity-building in science and encourage translational research III multidisciplinary clinical studies, health system innovation, and health industry growth. It was in this context thai China launched its campaign for "Deepening the Health Care Reform" in April 2009, with an additional budget 0£$124 billion for 3 years (200g,...20 11) despite the international financial crisis.s This ref mID is aimed at enhancing equity and accessibility through five targets: create a. universal medical insurance system; establish a national essential drug system (NEDS) that meets everyone's primary need for medicines; improve grassroots medica I and health care; make pub] ic health service available to a II; and promote pi lot reform projects in management and financing as experiments, to' be scientifica lly analyzed ill publ ic hospita Is.

Major progress has already been achieved. Today, life. expectancy in China is 73 years (as compared to 35 years in 1949), and over 90% of Chinese people are. covered by medical insurance, although still at only a basic level for farmers and urban children and some elderly residents. A NEDS has been implemented ill 30% of urban and rural areas, thus ensuring people's inexpensive. access to quality drugs, More than half of the county and township hospitals in rural areas have been renovated, and pilot reform programs in public hospitals have been initiated in16 cities. More than 1 rniiliongrassroots-Ievel health professionals are now in training, and services for vaccination and the prevention of some chronic diseases are available for most people. 1n addition, a successful control and mitigation of pandemic H IN 1 influenza has given impetus to China's health care reforms.j

Coincidentally, US. President Barack Obama and his congressional allies are attempting to undertake health care reform, Because all countries seek innovative ways to improve health care, nations should share their experiences in using science to drive policies that increase life expectancy and health literacy, thereby providing benefit to all peoples

- Chen Zhu

10.112611 d ence.rt B S 96 5

.' See http://tJ ~ws. X in h UB n et, (omflii de tlI2 00 9- 04/0 7lw~te nLll14!5 3"62 . h trn and (ont~nt_1279122 .htrn. tR. Stone, Science 3.25, 1482 (2Q09). SCIENCE VOL 327 19 MARCH 2010




forming amide linkages directly, they Interpose an azuidine-bound aldehyde and an Isocyanide. In the ens uin 9 reaction, th e isocya nide bridges the negative carboxylate at 0 ne end of the peptide strand and a positive irninium (formed by amine attack on the aldehyde) at the other end. The reactio n to lerates the fu II

range of natural amino acids, with

no observed disruption of chirality.

Cydes comprising one to five amino

acids, in addition to th e aziridi ne

coupling unit, were isolated in good

yield after simple precipitation, and

the aziridine's reactivity allows for

fu rt her fu ncti ana hzation of the rin 9

perimeter. - JSY

). Am. Chem. 50(. 132,2889 (ZOIO).

Corralling Peptides

There' 5 an inherent entropic cha llenqe in p repa ri ng cyclic melee ules from linear precursors, because the two ends must be coaxed together before they can farm a bon d. In zwitt erio nic peptides, the negative carboxylate and positive ammonium groups at either end attract one

a not her electrostatically, and so, in prin ci pte, favor a conformation poised to dose the loop. U nfortu nately, tradition al (0 upli ng agen ts

used to faciLitate formation of the amide bond eliminate the charged motifs, and with them the convenient conformationaL biasing. Hili et ai. present a n alternative cou piing scheme that conserves the zwitterionic attraction. Instead of



Weathering lnfluenza

One sneeze and influenza is d rifting through the air, plastered across palms of hands and arou nd door handles, poised for its next victim. How Long can the virus survive outside a living host? The answer to this question depends on ambient environmental conditions. Shaman and Kahn showed experimentally that low absolute humidity (grams of water per cubic meter of air), which tends to prevail during temperate winters, improves the airborne su rvival of influenza viruses within aerosolized drops and favors transmission. Shama n

et al. modeled how changes in absolute humidity have driven the seasonal peaks an d troughs of influenza in the U nlted States du ring a 30-year period. Epidemics were correlated with the onset of anomalously low absolute humidity, and variatiers in absolute humidity affected the occurrence of outbreaks during anyone season. Th us, It may be just as feaSIble to forecast short-term influenza risk as it is the weather. - CA

Proc. Natl. Acad. Sci. u.s.A. 106, 3243 (2009); nos Bioi. 8, eI0003I6 (20I0).


Polar Bear, Polar Bear

Polar bears are adapted to living i,n one of the harshest environments on earth, their range being determined by the extent of Arctic polar sea ice. They arose from the brown bear lineage and are most closely related to 11 group of gen ehcally distinct brown bears that inhabit the Admiralty, Baranof, and Chichagof Islands of Alaska's Alexan'


der Archipelago (known as the ABC brown bears). 0


Time estimates for the brown bear-polar bear 0


divergence vary considerably. Stratigraphy and ~

datlng of a rare polar hesr fossil Jaw bone, found ~


on the Svalbard Archipelago in Norway, suggest ,,'

that iti, 130,000 to 110,000 yeil'rs old and reveal ~ that polar bears were a distinct spedes at this time. ~ Lindqvist ei al. extracted DNA from this fossil and i)

sequenced its mitochondrial genome (mtONA). ~

Comparison of the ancient polar bear mtDNA ~

'" '" ""


" t:;:


'" '"





~ ~

o ....






t_ ~ ~~======~~~ ~


sequence with a n umber of mtDNA qenornes from modern polar hears and brown bears revealed

that the ancient polar bear lies almost directly at the branching point of polar bears and ABC brown 'bears, indicating that this bear was very dose fa the common ancestor of both. Time estimates for the divergence suggest a range 01177,000 to 131,000 years ago. Furthermore, stable isotope analysis of carbon and nitrogen in the fossil indicated that the ancient polar bear was a marine predator at the top of the Arctic marine food chain, like its modern counterpart {and unlike brown bears}. Thus, polar bears appear to have adapted very rapidly to their Arctic sea ice environ rnent, perhaps within 10,000 to 30,000 years alter the split from the brown bear precursor. - GR

Proc. Not], /lead. Sci. U.S.A. 10.10731 pnas.0914266107 (2010).


Grounding Mosquitoes for Dengue Control

Dengue fever affects 50 to 100 million people each year and can be accompanied by symptoms so severe that it has been called "breakbone fever." It is ca used by a virus ca rri I'd by th I' mosq ulto Aedes aegypti. There is no vaccine or specific therapeutic drug for dengue fever, and control efforts are complicated by the fact that the f<>male mosquitoes

I (which bite) are active during the entire day. An

g; active area of investigation in fighting insect vectors ~ h as been th e possibility of g en era tin 9 ste rHe 0 r con~ ditionallethal varieties. Now, Fu et al. have created ~ flightless female A. aegypti by linking the promoter .~ of a gene, Adin-4 (which is found in the indirect

e flight muscles of femaleA. aegypti), to a tetracycline::: repressible transactivalor construct. Because of difd ferential alternative splicing, males are unaffected.

d The flightLess females should be more susceptible


':l to predators and unable to attract males with their

~ wing oscillation sounds for mating. It should be

~ possibLe to distribute engineered eggs, which might ~ then control or eradicate infected mosquito popula~ tions by a release of transgenic males-whose

~ female progeny would be flightless. - BJ

Proc. Natl. Aead. Sci. USA. 10.10731 pnas.l000251107 (2010).



Familial Ties in the Nucleus

Inherited mutations in tu mar sup pressor genes cause an increased risk of developing familial cancer syndromes. Many of these familial turner suppressor genes are also treq uently mutated in somatic cancers .. The tumor suppressor gene NFZ is mutated in the tamilial cancer syndrome neurofibro matosis type 2, which causes multiple brain tumors such as schwannomas and meningiomas. NF2 encodes the. protein Merlin, which appears

to lin k celL adhesion receptors at the cell surface to the actin cytos keleton and is th us poised to Inhibit mitogenic signaling downstream of mtegrins and adhesins. Now, U et al. have identified a very different function for Merlin, this time in the nucleus. Endogenous Merlin WaS observed in the nucleus of mu [tiple (ell types by virtue of its binding to an E3 ubiquinn ligase, CRL4D(AF1. The binding of CRL4D(AF110 Merlin inhibited

the u biquirin Ligase activity a nd suppressed cell proliferation. Tumor-derived mutations in NF2 prevented Merlin from inhibiting CRL4DCAF1 activity, and CRL4DCAFl was required lor the malignant properties of primary human tumor cells derived from NF2 patients, th us providing a possible drug target. - HP

Cell 140, 477 (2010).


In a crystalline material, plastic (or permanent) deformation involves the movement of dislocatio ns in a series of eLeme ntary glide steps. Dislocations are defects in crystalllne ordering, and their motion in simple materials is well understood. However, some complex metal alloys can have hundreds of atoms in their unit cell, and it is not at all dear what intricate series

of steps guides the plastic delormation that is known to occur in such systems. Heggen et 01. used aberration-corrected tran smission electron microscopy 10 track the rearrangement of atoms in the T phase of an Al-Mn-Pd alloy-a lattice with 156 atoms in its u nit (I'll and structu ra I subunits that a ppear as hexagonal tiling with alternating orientation. In deformed regions, a mix of stackinq f au Its w.,s observed along with neigh boring regions of the urthorh ornbk R phase, which is closely related to the T ph ase but with para tlel hexaqonal t\ling. 5 urrou nding the dislocatton core were defect reqio ns, known as phason defects, which did not possess a strain field. During defo rrna tion, the phasons escorted the dislocation co re and locally transformed the material, thus allowlng the core to move. - MSL

Nat. Maler. 9, 10.1038Inmat2713 (2010). SCIENCE VOL 327 19 MARCH 2010

Science Careers in Translation

Want to build relationships with clinical

or basic scie ntists? Get advice on the best way to conduct a c lin ica I and tra nslatlona I science career? There's no better place to explore these ideas, and to build new scientific relationshi ps, than CTSdN et,

the new online community from Science, Science Careers, and MAS made possible by the Burroughs Wellcome Fund.

The ra's no charge for [ol III ng; and you'll enjoy access to:

~ Practical and specific information 011 navigating a career in clinical or

Ira ns lat ional research

Opportu nities to cormect with other scientists including peers, mentors, and mentees

i) Access to the resources of the world's leading multidisciplinary professional society a nd those of our partner

orga nlzatlons

Connect with CTSciNel now at:

Communltv.SdenceCareers .. org/CTSdNet

l'AAAS pt!§ll!

IlilWW. sci en (e!ljag,org



12.00 New York Avenue, NW Wa.hingmn,D( znees Editori.>l; 20Z-326r6550, FAX 202-28.907562 N~w" 202-326.6581. fAX l02-, 11-9l27 8atem~n Hou""" 82-88 Hilh Road Cambridge, UK (82 llQ

+44 (01 122 ~ 3.26500, fAX +44 {OJ 12,23 326 SOl

S'\JlI~rnlflHl~1 Si:iHilCf..5 For dJan9~ ol ildd~e5Si'j miSi5'ing hSlJIZS1 niltiw erde 1'0 'ond re m!W~15, and pilyment. questi 0 11,: B~6-43~ -MAS (nn I 0, <,02-326·64 1;, rJl).:,z02·842·lo6S, Mai!ing addre;,'" AM5., P.O, Box 96178" Wl,hingtlln, DC JOO.90-6178 or MAS M_f1ibd SO"","" 1200 N~wYork M~nuQ, NW, Wil,hlngton, DC 2QOOS

I.'nl"j,o~., 5", LI«"'" please QU 102· 326·6755 rQr al1Y ~ U.5~O ns or infolma t ion

R, .. ,~'" AUlhor Inquirie$ AOO-63 ,-71B1

Citmnlerdalll1qllllicl BOBS9·4S7<1

P"'.\"~,,., 20Z·5'26-707Q. FAX 202·682·0016

M" .... B"",,,,, AAAS/!!am es &Noblo,colii bookstore ~._."i1" s, oiWl:m; AAf!.!; Qlltine 'ilo"~ ..-ww.aphOUrIQ.<om/ailal/ <ode MKB6: MAS Iravels: Bel.harr E.l.pedition. IlOQ-.252-~910; AprM Slore "'''~. iJpplo . ..,m'opp>lor""""ils;' fIilnk 01 Ameri". Milstoraord 1-800·8H·6262 prtoruy code FAA,VU; (old Spring Harber labor~t';fY P,e"," Publ tcations www.t'hlpro.l",omWf1li~l~;I~iI~.i1lm; G~lm A"lO Insurance' .>lW'.9"imcoWlonoingp"golgoSl.htm!logo-176.24:

Hem ~OO·6J4·2200 CDP~3~3~5 7.: Office D~pot hllPI;l/bs.:1. orfi(~d .pol{omlpor'!;'lILogin .do; 5mbury '& S rtrith Uf. Ins Uta n(e 800- ~;>~-.9B83; Sulk11'u VIP Prllgw" 202·326,6417; VIP Moving 5 .. \'100,"iayllow"'Jilml; Olil,; B.neht>: AAAS Member ~lvic5 202-326·6417 01 \wN/.<lil"'!Mmbcr.or~.

~ciet1w.Jdiforl@-"W.ol!J !lQr gene",1 edit~rial querill5l

,Ci"n(2_I.lwrs@ilo",.0I9 (for qurui., <1boOI !.Iters)

,Citn",,,.org (for r~tuming rymnllSrnpl iWi"y,;1

,denc,,---UookreVl@o".,llfg (for book r"wow que"",,1

Published 11Y dle Ameri(llfi i\l.\ociF<litJI1 fOf 1he I\d"ol\Celil~1 01 5cton,e (AAA5.I, Sdeo,' '"<'10,0; its fead." ~5 J toru rn lor th~ p~'" .n!:l!'",n lind di,ai5~(m ",!imparl"nl i$5u"~, .. m. advancement of >0;00<". ihdueling the w'~nlilrloo 01 ml"onty 'Qf ,mnlll(llng PQhll< of vifW, rJtilef lhan by publishing only m.l~ria! 011 ,,11im ~ COII;~IIIUI has bren ,Q;1cI1NI. Ac<:o,diI19~, all al1ld~1 publi>h~d il1 ~den01-intl"tlin9 ~diloria~, news ~nd tolJlmel1~ "lid booR l'evi(!\'f.;------f<l~ ';911&1 and reftw the Illclividllill view> of Ihe;llJl hOI'< ~l1d ,wi Qffk;;,l ~Qirtls of vi""" adoptllci by AAM 0, 111. fn,HIUlion< willi 11l1j," _t11~ a'llhofl ar,~ Jflilj;)led,

MAS ,"'" f<lUn~"d in U48 an~ iocmpOIill.d in 1a7~. Its miliior'l i • .lo "dV"n(" sri,,",e, .ngin '-"I\g, and inoo,,"fian ih,cughounh. w,ld 1m \lIe b "rlenl of.U l)eoplf. The gQill:; Qf lf1e .,wei.lion "r~ ro~ 'nh~",,, (ommllnfcalio. 11 alwng 1£1<> n~ru. • eI1gfnCC'I:;,. "ndlile pUblici, promOle <Jmj def<'n d th (' mt~ghty 01 "ICQ((l and lIS UIll; 51l1lng then supporl lor tlw science and lecl1holiJ9"l en(el'prlle; prQV\de .~ ~ok~ for s&llte 011 5Oq~lill 1"u05; p~oliiole tile r"'i"'n'iible Usc 01 ,ci;n'e in public JI"llfcy; ,nen,9lhen .• nd divOI',ify th,; ;;cH>n(~ 000 l£dioolo9Y ~""kfor(J'; to!I.' ed UGltion in ,den,. and lechnoL09'/ for everyone: in(re<IIE. pub Ii( engngemeni with Kim", Jlld todlJlOlogy; Jnd ildV1l0"'- inlemalilnal lXJOf>'1@tion in 9Oel1( •.


See pages 352 .and 3S:~ 01 t~e_ 15 January 20:10 Issue or access www-!Cie",emag.orgl~bouvauthQrs

SENIOR EDlTOIUAt80ARD k~~!d"';=,~~~-B,,1Q'dlh1"'.

Liod. Portri;;D~, U/W ,aI!i!'g«O/ldon MichOH!l ~.1\B1:I8j Ui'

[OI;QK I~ (IilEF lI,u(~ Atb"efts

D:tWlJl'( wnOR ~[I'I, EOIIll R

Mo n iea M. Bradford Cot ill No rman

M~'~G'~" (arm. ~(.L\'CIIJGU~~'lS Katr;n..- L Kelner

DEPUTY fDITOIt\ R. BrOD" Han.op, Barbor-3 R. Jasny, Andr~w M. Sugden

EllIlQ.RfAl !i':;IIIIOR aJII1Q:JBJ{"~9ftillR.'f Lisa D_ (flOl\~t Br:rlld WIble: !ia,tI"R """"'" GillH>~ J. Cliin. 1- Hines. Paul" A. ~iile,-;;ri~ (llffi(onJ, Ma,( S. ~vio~ (fOroI1tO). !>ev~rl y A. Pu [11 ell, L Blyan Ra;<, GlIY Rid9 ibough, 1-1,] "''''' 'inj;!h, ~hlllip D. S'urami(Teno"5~'I, v"Wo \fin""" J~ke ~_ Ye$n: .,'0',"1£ '""0"' K,istQn L Muoller, Jolon. S(;)jic, NichOl"! S. Wlgginion, l<1ur.a M. ZIolm; """"'-'" .;"",itrr AI~i> Wyn nQ Mo~CI~ "" U"" WlIO. SI,w{lF1 Will>; ""0'",,, 0"',1.' snnu .. ~o""rl Fredelilli. T"ro S. M¥;,tlle: .... """"" 0"",-"","' Mill1'/n Green, Al1dr~ iNhilelell;..,,,,, ~iV1B'j EDf:J~, SMrm~t'I J, 5lJt~r; mQaMii lfllIRS, mrioFl')etmjler s.nls~, rnnaliJ~ ~ car~ lat<': '"",OIl ""'" """"" Jeffrey ~. (001:, Cyolfii. Ho'\'<'; H~rry l~(b. e"fb~ro P. Ordw"y, Trisn, Wil>JOn"r; <0'" , .. ,,"" Chris filWrcoU, la II ren Kill 11(; m~o","" (0'00",_00> CJ '·01 yn K"yk, B~v~,1 Y Shiold,; "",",,,,,,OIl' M~"'"" R!lm,~lo~IJye Diop, Jo; S. Grange" jeffrey ~ ""m, llla jobnson, SCOli MiUel'. Jerry Rich.rdson, Jennl1ei A. Stibert, Brion Wl1ile, Mila Wyno; ... lofII'" .... ,.,.." tinily G"I,O,. Midlild Hi""", Poll;,;. M, MilO(~; Mi~am W"'nb~rg; £lit""", .. •• " Ali"," C",Wfo~d, Sylvia S, KiI1""~: '"",",,,,,,TI,,,, ''''''N'' M"''jI1lI.e Madrid;, .. nO!IJ"'- f!lJ.O" Mllli>Sa It M((a~n'i'

N'~' "" "'V ~"" '",,0., Robe,! Coo n I>. EHot M;",ha It- Jell to t Me <'IiI, teslle Robem; <"''' ... ''''.~ """"" f[iUlIH>th eli 10 II~, PQlLy Sfiul,""n; ~""!"'I'!'! Yudhijit Bha~a(h~Ii~~, Adrian (1)0, lennijer Counn, David Gii mrn, Consla nee Holden, )o<oIyn KiI i!;e" Al(h~!d A. Kerr, Eli Kinfism, GrEg Miller. E1i<obr!!h Pmni$i,. !\Obert f. SCl'jj~ (J><o:\n( NIM. [ri k S toiU;I;lO, lu, 1'II~l1g: 1"""" Lauren 5t1\enkmn 11; ro~.'" lIjO co •• "",,"n,",, Jon (oh@\1.,n Dje_go, Gi\.I, D.ooiel fe~be(, Ann Glbbonl, 5.1m K~"I!, Robort 1(oolll9, Alldrow t.,Wllel; Milch U'51ie, Cllarl5 (_ Manne Virginia M9r.II, !;ary raubfS; <01'1' "'''"R< linda~. fel¥9, IMll'in Gii~i<llJ; MelisSa Rilimon{Jt 'OM' ","' .. ,,"" ".,,,,, .. &"e~IlIi"'IM<:k; ",,"<A," 5Jn Dl~o, CA: 760·g12·3252, FAX 760·9~;>·1979: P,1citlr Nonh ... "", 5 D3"'16l'19~ 0

P'OOUC1\O.' Dr.,,'", J~ll1es liIndry; "'" 0. "'",,0,. Wendy K" S 11 '10k; M5t~lANi' MANAr.a. Rc bece,[) D05:hJ; satmR' j,rE.ctf;ml~ Steve! F orrCSI-c:r, en ri~ Reowood: .. ,w,v" Mtt>o,W Rosen; ~~'rn"", ".'<;IOR /II. TQI"~~in~ ~ ...... MM,U,'Spiegler. ,",0,,-,'" j.l(Il1lliUman

,A'Rl DIREtio~ Yael KaB; "~'D~t"lE".AU Lll!il'EOgfli lallr~ C:1'(1'1;.1~lll'tlj .'5E:NI[II~ 1"-""'''''0'''' (oris steRe!, t\alh~rin. Sut(lff; 1,,-""'",0. Varl~ (ire'hlnon; "'"'0' 'JIl.""".m HoH~ Bilholl, Pre,\ol1 Huey. Nayomi Keviiiy.~.g.I~; "'" .,,,,,,,/liE> Kay I1!g InJl1, /I>'I<IUh~w l\I'Q/11bly: rIOmJ ",,,,,, ll1S1 if: BUZdrd

S(J~a IPJill~ill~ONr'd

E.uFt~ff (Si(.ieou~@:s(jl2'ntG-int.w.ukj ttl11i0ftlAl! ~'rnDn~1I!nCI'PfIj,,, ~rulF«) ElIn"'" Andrew M. S l,gd~n; "'"'''' ''''''''''W~M_' Mia Fail rcn~a",p· tJppenbrini(; "H'o, <OllO'~ (omlloe /Il;h, Stella M. Hurtle'j, '1M ,. Osbornel Petar s.tmn~ M:5.0[J.lUE alll0f{· MlJrid Cruz; tOm"", EnnOR Hr:l~n Pld\CrlgiU; ., "0''", '""'0"' DcborJh Dcnn is<ln. Rach.l RobertI" Al it.(' Whalfy: "'~'~"T"'l '" <"""0" Iphn C~nn.U, )0 net CI~ment', lQ~is" r.rr.l1YrIell~ ~EWS~ E~ii!:DP~ "EW~ nHl9R' John TrnVi$; Oa"UT'i' Ihms amo~ Dnl,iel (I.ery; (~"lF""';'" ,o.",,,o.~,.j, 'Mirll,lCI Baltc, (PMll, John B"oh"nnon (V;el1J\ill, Me11]n Er.<enok (/Irn<lerdom ~nd pi,lrio;l, Gr~ld)1'n \Io<JeL !Be~inl; I", ... n Tim Wogi1~

lPLTI~ Arf.i'iitr~ -t,ONlfi!~i1l1':11W!i' tOR~F/OffD~f Ant¢hi;o Rcg(,llado

A,,'A Jop~n 01110", I\sca tOI"pora~on, 1011101;0 Furmav,'il, RUltic BI(Ig, 7 r. 77 r ~nji n-(h 0, Shinju ku<ku, To kyo 162-0808, Jil pan; + 81 , 6802 ~ 616, FAX ,,81 3 680Z ~ 615, ioqWil)'@ld~I1'em~g._jI1; .,,~ ~'1" ~."9' R ichanl S tQne (Be;j1 n9: nlane@"""'; roOT"" \IT'NO {o •• mo."",,, Denni; Normil" U"p1ln: +111 (0) 3 B91 06-30, fA)( +31 (eI) l $·9% 3S31, dnorlnile@gol.comJ; Hao Xin [Chjn~: +86 (OJ to 6307 4~ 39 0' 6,01 %) 0., fAA +86 (Ql 10 B01 ~ 3'\8; tindyh""@gl"r\<liLmmJ; PJI1i>Vil Baglil ISoulh /Il;,", +91 (0) 11 2271 28%; pl}agfa@VSI)Lcom)

em:, liI'i E PIJ ~11;11 tR A~ n I .. le~h,",r p~ 'lI~11 ,. B~lh R",n~r

FUII.ifil :LIM'E:NT SVniE.M:5- .IHfO 0 rnA 1 f'[J,N;'5 (m e mb ~ r sh i P@,1 u i~ s.o fa) ~ DUIHTOPl: WZlylon Buller; ("'!iTOM~a !iH~~j,H S.UPiE.,H1UOA Pill Ru ll~r; '.'''''m,. l a I oyo (asteel, [aVo nd a U~\'lfo rd, Vicki UI1tPJl, April M"rllial~ 0"' .•• ,rI, ," •••• 1'0. Cyntil fa John'oo; ~"1'''''''' ~hi~ene Hall, Tarri~" Hill. WiUloQ\ Jane>

BU~tN E;:50~ 0 i'.E.lU!.l~'DN:i ANj) A:fI'HUN~otl~~llJlIII [I ~ftE{li~R Dtborah R iV-E::ra· W1en hold; "'1',""". "' ...... "'. "!Wi""- '.Al",a< ""'''J>If f\dndyYi: """"_ au~Ii!!EfS, '~,"JWllk5- Elic K'nott; MAft:A{:£H,. B-U5tN~ OfEA~~I'II~ JessiUl Tj~rn~y; FtI'U!;xUAt ".'rot.trm PrlH PtUfUl'd01J t!2'leste: 1 roxl11r; ~u'.I"n:s N!iD ~t~1C1~ ..... "' ... , .. ~milie Dovill; """".,, ~11lil bel h 50ndll'r, .."",,,,,,," ... m .. I~n King; "" .. m,.. .",.,,,,,,,, AUi'OIl PMt,lwd, AIi.<OI1 Chand~r. I ulMnne Wld9a; ... _,." .,;,xlA,o> A1mt~'ApoOI~, MellY EIl"'1 (rowll"j, Welldy Wise; 1Jt~10ift: MAimE'lilliil6 [j(R\J'UVE J~nnUer Re~Vr,5~ ~ill:EO~, :!i~fiE UCffiitNt! Tom RI"'n;",.m ... ~" ..... ;r"."Eil~en aero.delle Moranj .. ,."" ..... RWl.ll ~~, ~«E5'OO ~CE~ 5PE(lJ;il5T }{j kl f(j~lhr; 5£Hl~ i'Uili..!'iHBiI RaAil [JN5 """' ........ "'I herine Hollilnd; .. n"","", ...... M.., ..., <""" Pliillip 5mll h, PtI!lu:r~~ 'Il~lj!Lf!IJIN.S;k WEST {OoB.Sl Philip l:Wlak,ldi.s.; RJ~ftllill'lENl '!i'u~,",!iitl~ lquc Ed iln;F!A.r"'"-,", ~OO""'~.'O!! Girlie M.l,[fo,,,,ld: ..... _om1 ........ G<l!.

til ri;lina S~h I.thl; flJ;c,"""" ~'" '"' _ If..,,bel h Harmsn; '",,1m

..... "'" Ti-i5til 'Snyder: ss. ""'''''' 1M Li-lil Sto1nlQrd~ ".",. ,""""""".

_UiU"" Ryall A!~in~, ehrislQP~er Colfm'!n, m"Y",," ,,~'" Wal!", Jim.~ "'00""''''' <PEO",-,,,,, t.JidTele10hnlton, ~imbQrly Olt.r; " .. m"", ..... _ .... """.WiIiCoIlfn,

lfI\tQtl,I!i1 N'6 WrOA:!,DWH1E" AD ~,Q],[,5 6i1l M~n "'''~at1" "'''0' S.on 5.Jrulef>:, 202 -3l6>'64 30

P1'II D}fr'1 mfiI!10'l", QlfJM.A{~ Bni:H'IOa Bla5er

ho.un (-,de n {e._~d'iel Ii s i ng@~.,".n rgl: ii' ow.,, Rick Ii0I19iov~oI11' 310-4·05-70$0, FAX 130'40:'-7081; •• " (0,,-,/ E. ,".M.' l."ne F"f~(I~y: 508·7~ ,·9395. FAX 611-507·8189: \If"" co.",/". w"""lyone Stkk(o~: ~).5·'~!"918l, f~X ~!5·5?O- 6940; u,.:/"'.o.".,,. ~oger GOI1(~lve': TEUFAX +~ 1 43 24.1 1153; , ..... A'>CA CorporaUoo, Nall~~O Ide +81 (0) J ';80.~ ~ 616. fAl( +81 (0) 3 6802 ~615; ad'@'tief)_(em~g_ip: ,~~,o. ' .... " ''ijO'"'' Oei\lndr~ S41111l1l

,",",b,,",o •• "i><I .... P .. ;t, o. 'Of .5"" ... ", .. Co. •• "', fI'~cy HoI m .1: +~ ~ (D) 12<1 HQ~.2~, FlU' +~A (01 lin 3265'12

CtA";5UUO: (aiJV~rl is-Ql@sdenL12{ u.s.: :5,,IIotE!i iI!IIl!o~Ii.:G(H:

DM_y1 A h<1 e r SOl;: 202 - n 6-654 3; M' oW", Tin ~ 8urk ,: .202' 3 Z 6- 6517; EA;' c ••• , l02·c326·6513; .. ,,-r/,m"" ,".tR.l N'cI101.s Hlolibid,.: 202-326-6513; .. m ,0<>'"'.",0'" Roh~n Edtnon,on. Shir~y YO'ung; "" .. 'Susanne Klw",~, Di!1l P~nhlf<_g(Oh, Alex Palmer; ">"" ml<Wil I.i5a P~tl~r~n: "''''~ MeA Corporation, lie Chin +81(0) 3 6802 ~616, JAX +Bl (0) 3 6802 ~615; (A!Q.r'1Iru@;cien(; AO' .... '~~ .. '0",,'0"' """" .. " F<lotc: 202-326-6740; .0 ... "~nG P1!!OOUt'HClN, j):PE~lltJPilS. 1U:fflld,ER O~bor2lh i"omplrln:i~ !iEN~OR Pf!OOULJ[Q8 5PEW!'U5"G~ffi_I[ DESI6NoI:A, ArJlY HurdGtl~tle-; 5miOlil; ffiODUnlOoNi sf'ruAIJ'n RDbeoi BlId{; ,,~'o. ""If'< '"01''''' {hM5~OC H~l

MAS 8{"\ftD tI f DU~HifIlRj 'R.E1i~_1 NG" f'Ri:5i[l~!II L ~H.p.l R r.~~(!:r c. A9r~; PJI,!=!iIDEl!'iIl Al.k~ Huang; rJl(!,mr,'ft~rtH"l Nina 'FooofotJ; 'A.EA!ilU{.E&I' OCivid ,E. Show; ,..... ., 'Lull"' om WI AI" n I, losh ner; an.Rn li hdil P. B, Kil I ~Il t Nanty KoO\~IIOIl, 5teplll'n M.~o. (hl'ny It Murr<w, JUIlJ M. Phillip:;, David O. S;,tb"\i~i; lhomo:< A.W",,~



'19 MARCH 2010 va L 327 5 CI EN CE www.sciencemag.Qrg

Elephant (amp Swept Away

Ea rlier this month, a flash flood from the Ewa50 Ng'iro River sent a wall of water through the Save the Elep hants research cam p in Sam bu rll NationaL Reserve, Kenya. The camp, fou nded in 1996 by elephant expert lain Douglas-Hamilton, is known for its studies of elephant behavior and cognition. "We expect it will cost hundreds of thousands of dollars" to replace tents, roofing,

a nd co rnputers as well as electrical and p lu rnbing systems, says camp operations manager Lucy King, a grad uate student from th e University of Oxford in the United Kingdom. The scientists

n, " o



~ have launched a fundraising campaign {see

~ www.savetheelephants.orgl. But money won't ~ make up for the 1055 of their field notebooks ~ with their penciled records about individual

~ elephants. "These are i.rreplacea ble, " says Ki ng. ~ No elepha nts were hurt; they all took to the hills is when the rains began.

ur o


~ Big Prize for Gene


~ Sequencers


~ The 10th ann ual Albany Medical Center Prize-

~ the United Slates's biggest prize in biomedicine-


:0; will go to three scientists who ccnceptualized the

~. Human Genome Project: Francis Collins, director



Three Q '8 > > Museumgoers outside China

. will get their first Look at early

European and western Asian migrants to China on 27 March, when the Bowers Museum in Santa Ana, California, opens a major exhibit, Secrets of the Silk Road. It features naturally preserved Caucasoid mummies, one as old as 3800 years, and 150 artifacts discovered in the arid Tarim Basin. SiAOlogist Victor Mair, left, of th e University of Pen nsylva nia h as studied th e finds for nearly 2 0 years.

Q: Where did these BrOrlze Age migrants come from?

I think from somewhere in the steppes north of the Black Sea and southwest of the Ural Mountains in Russia ... , These early people were buried with little baskets of wheat, and they were also cattle herders and sheep and goat herders. They were very good at making woolen clothing .... I think that special kind of aqropastoralist culture is related to the steppes.

Q: How did they survive in one of the most arid places on Ea.rth?

There were rivers extending out in the desert, and there's evidence of old poplar groves near some sites. So there would have been little strips of pastu re fo r the an irnals,

Q: How did the opening of the Sill( ROad 2.200 years ago change the population?

More people started coming from Southwest Asia or even the Mediterra nean. Yin 9 pan Man [a [most 200 centi meters ta lO (rig ht) is th e on ly mummy whose limbs were wrapped in doth, similar to the Egyptian practice. He has this elaborate caftan with classical Greek and Roman motifs. I think he wa.s probably a we.althy trader from around Sarnarkand in Uzbekistan.

of the Nahonallnstilutes of Health (NIH); Eric Lander, founding director of the Broad Institute of Harvard University and the Massachusetts Institute ofTechnology; and David Botstein, director of Princeton University's Lewis-Sigler Institute for Integrative Genomics. The three

share the $500,000 prize. At the announcement last week, Coli ins said he was dedi ning the cash for conflict-of-interest reasons-"so you could say I'm a cheap date, but I'm having a great time." NIH had no information on what will be done with Collins's share of the money.

A total ban on fishing in the S44,OOO-square-kilometer lone, an area the size of France, would make it an even larger protected area than the current record-holder, the 360,OOO-km2 Papahanaumokuakea Marine National

Monument in the northwestern Hawaiian Islands. The Pew Environment Group has spearheaded a 3-year campaig n for creation of a Chagos reserve. It would be "litera lly an

tndonesta island of abundance ina sea of depletion,"

__ Proposed reserve

S"ilys Pew's Jay Nelson.

The isla nd s are u nin ha bited except for the U .. S. Navy base on Diego Garcia. Some 1500 Chaqossians were deported to Mauritius in the 19705 for military secu rity.


Sri Lanka



~ Britain has taken another step toward designating the world's largest marine


~ reserve a rou nd th e Chagos Islands, a 9 rou P of S 5 cora I protrusio ns in the

~ Indian Ocean. The government announced the end of a 4-month public 'i co mment period on 5 March an d is expected to

~ reach a final decision by May.

LI Atrle<l

~ The (hagos contain half of the Indian Ocean's

~ remai,ning healthy reefs. The waters are said to g be among the cleanest on Earth, allowing corals 5 to grow in deep water less vulnerable to global ~ warming. The islands are located in the equatorial g "tuna belt," which hosts what a Royal Zoological 5 Sobety of london report (ailed one of the "most

w .

(5 exploited, badly enforced fisheries in the world."

chagosl Archi~lago


www.screncernaq.orq SCIENCE VOL 327 19 MARCH 2010



Budget Red Tape in Europe Brings New Delay to ITER

The projected start of the ITER fusion reactor in France looks set to slip by another 10 months, Although the new completion date of November 2Q19 for construction may seem like a minor schedule a Iteration for the multibillion-dollar research reactor. it is the tip of a large iceberg of negotiations over management structures, design changes, cost increases, and risk mitigation that has gripped the proj ect for th e past 2 years (Scfence, [3 November 2009, p. 932).

The latest proposed delay was revealed last week by "William Brinkman, director of the U.S. Department of Energy's Office of Science, at a meeting of scientific advisers to the government's fusion research program. Science has learned that it stems in part from budgetary problems in approving contracts by the European Union. "Europe turned out [to be] a weak link, ... [It] must fulfill its commitments," Russian researcher Yevgeny Velikhov, who chairs the ITER Council, told Russian Prime Minister Vladimir Putin in February, according to an official transcript of the meeting. Still, fusion researchers who spoke with Science say they are confident that this and other hurdles can be worked out in time for a meeting of tile lTER Council in June, removing perhaps the fina I obsta c les to starting ITER construction.

ITER aims to harness fusiors-ethe process that powers the sun and stars-to produce usable amounts of energy. Its site, in Cadaracbe, France, has been ready for about a year, and components are being built by all of the lTER partners-China, the European Union, India, Japan, South Korea, Russia, and the United States. But the partners still have not. agreed to ITER's detailed design, schedule, and joint costs, together known as the project baseline.

Most 0 f the reactor components ate manufactured by each partner's home industry and delivered to Cadarache as contributions in-kind, The manufacturingjobs were divvied up in an agreement drafted in 2005~06, but 'most partners found out later that they had underestimated the. cost of what they had


agreed to. The United States, which had estimated a $1 'billion bill, is now looking at somewhere between $1.4 billion and $2.2 billion. The European Union as: host, will build 45% of the reactor, and sorts cost increases are the largest

In June 2008, the JTER Council pushed back the date for the reactor's first plasma by 2 years, to 2018. Last November, the European Union requested further delays, but

the other partners wanted to push ahead. Part of the problem lies in the European Union's desire to build prototypes ofiwo key elements-its parts of the vacuum vessel that surrounds the plasma, and half of the toroidal field coils that hold the plasma in placebefore making the final items. A 2018 startup would require the European Union to begin building the manufacturing facility before the prototypes were finished and tested, running the risk of requiring expensive changes to a partly built facility.

E. U. offi cials have insisted that they asked for more time to mitigate such risks. But fusion researchers who spoke with Science 0]1 the condition of anonymity say there is another reason. According to E.U. financial rules, 0 fficia Is cannot si gil a deal wi til a COl1- tractor un I ess funding for the whole contract is available. Yet E.U. unclear research funding is approved in 4-year chunks, and the curreu t budget, which runs throu gh 201 1, doesn't cover the now-inflated costs. "The E. U. cannot promise money it doesn't have," says one researcher.

Some contracts for ITER's components may be worth hundreds 0 f millions 0 f euros over 7 or 8 years, and tbe European Union simply doesn't have enougb money in its current coffer. Insiders say fjleTC was talk of a loaf) from the European Investment Bank-an E. U. institntiou=to guarantee the contracts, but some E.U member states rejected that idea. The European Union would not confirm any budgetary problems, saying that. the current schedule "woul d have

Assembly required. Europe wants to make prototypes of its parts for the ITER reactor (above).

entailed extraordinary measures to accelerate activities in the building phase ... and would increase unnecessarily risk and associated costs."

The E. U.'s proposal to la st N ovem ber's ITER Council meeting was for a 2020 start, but the other partners refused. Last month in Paris, delegates provisionally compromised on November 2019. Vel ikhov 's conversation with Putin is one of several signs that HER partners are exasperated by the delays. The 201 J budget request to Congress by President Barack Obama would cut ITER spending next year by $55 million, to $80 million. The lower figure is «a reflection of the pace of ITER construction as of the end of2009," US. Energy Secretary Steven Chu told a Senate spending panel last month.

The ITER organization is now adjusting the baseline document. to the new schedule for review by two advisory panels before seeking final approval by the ITER Council

iJ1 June. Researchers remain hopeful that agreement on the baseline will remove the logjam slowing ITER constnrction. "There's ~ a sense that ITER is reaching a turning point, ~ moving in a positive direction," says Steward ~ Prager, director of tile Princeton Plasma ffi Physics Laboratory. Positive, yes; but it's not ~ moving there fast enough for some. ~




Silicon Mystery Endures in Solved Anthrax Case

What about the silicon?

That question has confounded investigators throughout the probe into the 2001 anthrax letter attacks, which the U.S. government formally concluded in February Seientists inside and outside the government say there is clear evidence that the higb levels. of silicon found in the anthrax came not from anything added to "weaponize" the anthrax spores-c-as researchers bad suggested early in the probe-but from tIH! culture in which the spores were grown, 111at evidence may have settled the issue of whether theanthrax was weaponized, at least for scientists familiar with the case. But it raises a different question:

Wby did the mailed anthrax have such a high proportion of spores with a silicon signature in comparison to most other anthrax samples?

The answer, according to academic scientists who helped with the case, probably would not change the FBI's conclusion that the attacks were the sale handiwork of now-deceased U.S. Army researcher Bruce Ivins. But it could help illuminate exactly how the attack material wa s prepared, Reso I vin g the mystery might also pave the way for new techniques using trace elements in a bioterrorism agent to link it to its source.

&<There's tremendous interest in usi I1.g metal signatures <I.S a forensic tool," says Adam Driks, an anthrax researcher at Loyola University Chicago in Illinois. But the science to do that is lacking: "We know very, very little about the diversity of elemental composition within spores when they are produced in different ways."

The FBI's scientific case against.Ivins rests 01] DNA tests showing; that the mailed anthrax came from a flask under Ivins's control at the U.S. Army Medical Research Institute of Infectious Diseases in Frederick, Maryland, Investigators also had the attack material chemicallyaualyzed, first at the Armed Forces Instinlte of Pathology (AFIP) ill Washingtou, D.C., within weeks of the attack. Examining

the spores nuder a scanning electron microscope, AFlP scientists detected silicon and oxygen and concluded that the spores had been coated with silica to make them float easily, enhancing their power to kilL

A more detailed analysis by Joseph Michael and Paul Kotula of Sandia National Laboratories in Albuquerque, New Mexico, contrad.icted that conclusion. Studying individual spores with a transmission electron microscope, they found that the silicon Was located within the spore coat, well inside the cell's exosporium (outermost covering). By contrast, when they looked at surrogate spores weaponized with silica, tile. silicon was clearly outside the exosporium,

But the Sandia study, presented last September to a National Academies panel

(I.oser look. Mi(ha~l (above) saw silicon well inside i ndivi d-

ual spores, visible as the 9 fee n Ii ning ill the top rig hI image.

reviewing the science behind the investigation, still leaves questions. Out of 124 spores from a letter mailed to Senator Patrick Leahy of Vermont, Michael found the silicon-andoxygen signature in 97-78% of the sample. the signanue was present in 66% ofa sample from a letter to former Senator Tom Daschle and in 65% of spores from a letter sent to the New York Post,

Out of nearly 200 other antllrax samp les from different labs, none carne close. to displaying such a. prominent silicon signature, The highest, in a sample from Dugway Proving Ground in Utah, was 29%. The researchers

couldn't find silicon tn the coat of a sing le spore out of some 300 taken from R.;_tvIR- 1029, the flask in Ivins's lab identified as the source oftbe bacteria used in the attacks; they concluded thai all the silicon had come from the culture,

The unusually high percentage of siliconbearing spores in the attack material "is a bit of a strange thing," says M !chael. "We have no way ofkncwing howthey were really grown," An anthrax researcher who did not wish to be named calls it "awfully weird" and "a particuIarly inconvenient ex ception" because it Leaves a gap in the case. However, neither scientist thinks the anomaly casts doubt on the broader investigation

The key to the mystery likely lies in the culture medium the perpetrator used to grow the anthrax spores., says Michael, In a recent study, Japanese researchers grew colonies of Bacillus cerellsa close relative of the anthrax bacteri lim, B. anthracis-e-in culture media with and without added si I icate, Spores grown in the silicate-containing ell lture showed si I icon withi n the spore coat. Til the absence of silicate, there was no silicon, the grmlp reported in January in the Journal 0/ Bacteriology.

One of the study'S authors, microbiologist Akio Kuroda of Hiroshima University in Japan, says the precise amount of silicon in individual spores from the anthrax letters c01.11d offer clues about the medium. "If the anthrax spores contained .<1 high amount, tIle suspect must have, used a medium that was supplemented with silicon or (hat mtrinsically contained a lot of si licon "Kuroda says. "If a thorough testing of various media sold in the U. S. identifies a. few that contain higher 8TTJ ounts of si licon, those coul d become an investigative clue."



Eppendorf Et Science Prize for Neurobiology

2009 Winner

Richard Benton, Ph.D. Assistant Professor University of Lausanne SWitzerland

Deadline for entries

June 15, 2010

It's easy to apply! Learn more at www.eppendorf .. com/prize


In touch with life

US$ 25,000 Prize

Congratulations to Dr. Richard Benton on winning the 2009 Eppendorf & Science Prize for his studies on odor detection in the fruit fly, Drosophila. His findings have revealed unexpected evolutionary parallels between insect chemosensation, immune recognition and synaptic transmission.

The annual international Eppendorf & Science Prize for Neurobiology honors young scientists for their outstanding contributions to neurobiology research based on methods of molecular and cell biology. The winner and finalists are selected by a committee of independent scientists, chaired by Science's Senior Editor, Dr. Peter Stern.

To be eligible, you must be 35 years of age or younger. If you're selected as this year's winner, you will receive US$ 25,000, have your work published in the prestigious journal Science and be invited to visit Eppendorf in Hamburg, Germany.


Beyond OSM: Seeking a Brain-Based Classification of M'entaliliness

if a frigh: at despondency lasts for" long time. it is" melancholic (lUeCl.ion. =Hippocraies, Aphorisms, 400 B. C. E.

Since the time ofthe ancient Greeks, mental disorders have been classified according to their outward signs. But even in Hippocrates' dayattentioo was paid to the underlying causes. The word "melancholy" derives from the Greek word fOT black bile, an excess of which was thought to cause prolonged sadness.

Modern research in neuroscience and genetics has provided a. more sophisticated understanding of mental illness, ~1fI,d harnessing this knowledge to improve the diagnosis 'Of psychiatric disorders was a 'major impetus for undertaking a revision of the Diagnostic and Sta1!:~/'ical Manual of Menial Disorders (fJSM) (Science, 12 February, p. 770). B\lt even some Of fhose leading the revision say therc's still too IittJe known about the biological basis of men - tal illness, and as a result DSM continues to be based on symptoms ruther than causes. "We just don't know enough to do a lot better," says psychiatrist Steven Hyman, the provost ofHarvard University and a member ofthe committee in charge ofthenew edition, DSM-V

A new initiative by-the US, National InstiMe of Mental Health (MMH) aims to foster the research needed to dose this knowledge gap. 'What we are doing is trying to develop new ways to classify disorders that are based 011 identifiable neural circuits," says Bruce. Cuthbert, an NlMH psychophysiologist lcadi ng the effort, call ed Research Dom ain Criteria (RDoC), NIMH expects to start roll ing 0 lit the project in ea rnest next month, beginning with II draft document that sketches out five "domains" of mental function

_ that correspond-with varying g degrees of confidence-e-tc spe,S cif'ic brain regions or neuro~ chemical signaling pathways or ~ both. This new classification g i sn 't in tended to compete wi th ~ DSM anytime soon, Cuthbert ~ says, but it is intended to change ~ the way researchers study mental disorders.

c .. -

~ The Impetus for RDoe dates back to

f Hyman's tenure as NIMH" director it .. the late g 19905. Hyman says he became. concemed t11at ~ the DSM classification of mental disorders


5 was hampering research. Investigators used


last In a se des

DSM criteria to frame their research questions, study sections used them to evaluate grant.applications, journal editors used them to judge papers, and pharmaceutical compa.nies used them to design clinical trials. AUhe same time, Hyman says, "it was clear that DSMwas a poor mirror of nature."

By way of Illustration, he notes in a recent review article that the DSM-JV diagnosis of' major depression requires that a patient have. at least five of nine possible syrnptoms.In this scenario, its possible for two patie:nts to receive the same diagnosis ·with Drily one symptom .in common. Their inner turrnoj] and its biological roots might differ substantially, but they could easily be. lumped together in a study on "major depression.vvWe needed some way to break out of the cognitive box and encourage scientists to do research that disregards the current disease boundaries," Hyman says.

Hyman's successor at NIMH, Thomas> Insel, made this a priority for the institute, and RDoC is the product. The draft document, to be posted on the RDoC Web site" next month, describes five broad mental domains thatare present in everyone but whose extremes correspond. to mental illness: negative emotionality, positive emotionality, cognitive processes. soc~a[ processes, and arousal/regulatory systems. It further divides each of these domains into Individual entries linked to parricular neura I circuits. Under negative emotionality, fot example, are entries for three spec; fie subtypes: fear (hypothesized to result from dysfunction in the amygdala and connected brain regions), stress and anxiety (linked to abnormalities in the bypothalamic-pitnitaryadrenal system and stress hormones), and aggression (involving the amygdala and hypothalamus, as well as hormones such as testosterone and vasopressin).

The current DSM diagnoses don't necessarily map neatly onto the RDoC entries, Cuthbert says, arid that's partly the point "Exactly the problem with the DSM disorders is that they're very heterogeneous and

*hllp;/IWl\IW. nl mh.nih.gQvlresearth-I~l1dj nglrdOLshlml


may invelve multiple brain systems," he says,

Beginning this summer, workshops wilI bring together groups of experts to refine the RDoC entries and to identify gaps in the current understanding of the genetic risks, neural dysfunction, and behavioral problems associated with eacl1 one. Over the next 2 to 3 years, Nll\1H will encourage researchers to shift from using DSM criteria in their grant proposals to using the RDoC categories.

One goal ofRDoC is to change the current practice of selecting research subjects based on their DSM diagnosis and. to encourage studies that use biological indicators instead, A study on anxiety disorders, for example, might examine people who show a heightened amygdala

Faulty circuit. The! arnydna la (red), the hu b of the brain's lea r ci rtuitry, (OU ld Ieatu re prominently ina brain-based system to classify mental disorders.

response to frightillli(lg pictures, regardless of whether their DSM diagnosis is panic disorder or social phobia. Another study might e11I"O I! people with a particular variation offhe DISC1 gene, regardless of whether they have a diagnosis of bipolar disorder or schizophrenia (both of which have been linked to DISCl), Such studies are esse ntia l for connecting the dots between biological abnormalities and the symptoms of mental illness, Cuthbert says.

Ultimately, NIMH hopes RDoe will inform clinical practice. In the future, a psyc chiatnst mightexamine Hyman's two hypothetical depression patients and diagnose one with anhedonia resulting from a glitch in the mesolimbic dopamine system; and diagnose the otl1(':r With ci disruption in serotonin signaling, coupled with anxiety caused by dysregulation of cortjcotropin-releasing factor in the hypothalamus, A11d tli.eri, witli luek, the psychiatrist would know just what treatment each oue requires. -GR.EG MILLER SCI EN CE VOL 327 19 MA RCH 2010




Budget Shortfall Could Derail Plans for Underground Lab


Depth of field. 0 usn wou ld carry out experiments by ~everill disciplines at va ryi ng depths.

A tight budget and lUHlI1' ticlpated safety problems are threaten ing to kill plans to convert an abandoned gold mille in South Dakota into a $750 million deep underground science and engineering laboratory (DUSEL),

Since 2007, the U. S.

National Science Foundation (NSF) has been supporting a team of scieutists and engineers developing plans to convert the Homestake mine ncar the town of Lead into an enormous Jab for experiments iii fields including partide and nuclear physics and geology and microbiology, Although NSF has not yet agreed to build DUSEL~scienti.sts are hoping for final approval as earlyas spring 2011 and the start of'constrnction in 20 I3~ NSF is spending $36 million tills year on the effort, But its 20 II budget, released last month, requests only $19 million to continue design work, half of what scientists and NSF program staff say is needed to keep the project on track.

To make matters worse, a receul review concluded that an additional $73 million is needed in the next 3 years to shore up aging mine shafts and make the space usable for research. That leaves the 20 II request some $40 million short ofwhat's needed next year, Joseph Dehmer, director of NS P's division of physics, said at a meeting last week of the federal government's High-Energy Physics Advisory Panel. Without a significant boost ill funding. Dehmer warned the panel, "the

project will be phased down in, 20 1.] and 2012."

Edward Seidel, acting assistant director of NSF's mathematical and pbysica 1 sciences directorate, declined to say bow the projected shortfall came about. But it does not signify a lack of enthusiasm for the lab, he emphasized:

"No one should interpret this as a sign that. we 'te not. committed to the project We.ll1lderstaud its fantastic scientific potential."

Although $40 million is a drop ill. the bucket compared with the project's tQt~tl cost, the money is critical to keeping people together, says Kevin Lesko a physicist at the Lawrence Berkeley National Laboratory, who heads the design team. "What's in the budget is nowhere near adequate to support that team through 2011," Lesko says. "To dismantle that team and say, 'Come back next year,' that's 11.0 way to treat professionals."

Part of the problem, according to Dehmer, is that much of the $19 million is already committed to private contractors


MIT Engineering Dean Tapped to Head NSF

The dean of engineering at the Massachusetts Institute of Technology (MIT) in Cambridge is in line to become the next director of the National Sc ience Fo un dati on (NSF).

Science has learned that Subra Suresh, 53, has been tapped to succeed Arden Bement, who announced last mouth that he would be stepping down before his 6cyear term ends in November. The National Science Board, NSF's oversight body, was informed last month that Suresh was the White House's pick. He is currently undergoing vetting, a process tbat could take weeks or longer,and if no minated he would need to be approved by the Senate.

The. Indian-born Suresh is urn well know'Ji


in U.S, science-policy circles. But those who have worked with him rave about his sci enti fi c accomplishments in the emergin,g field of nanobiomechanics as well as his administrative talents. "Subra represents the new style of science and engineering," says former MIT President. Charles Vest, now president of the National Academy of Engineering, to which Suresh was elected in 2003. "He has done amazing work on the mechanics. ofindividual cells, and he's used to working w:ith interdisciplinary teams. As a dean, he has been a leader In helping create. powerful interdisciplinary entities such as the new Koch Institute for Integrative Cance .r Research, tbe Center for Computational

Engineering, and Tnmsportation@MIT;" an initiative by the engineering, business, and architecture schools to examine the global environmental impact of an increasingly

mobile society. ~


Suresh declined to comment on whether ~


he was under consideration. But there are u

<I; indications he's already in step wit.h the z ~ Administration. In a January 2009 ~


blog for Technology Review about a "dream ~

stimulus package," Suresh proposed high- ~ speed railways as a way to "stimulate worker ~ productivity and the economy while reduc- .~ ing our damage to the environment." The N

~ American Recovery and Reinvestment Act Q

included $8 bi \Ii on for such purposes, Smesb£ also called on the new president to reverse ~ Bush Administration policies that he said g have undermined the nation's "ecosystem of § scholarship arid innovation ... [bringing it] ~


and DUSEL-related SCiCJ1Ce projects. ~lt different universities. A~ H result, he told panel members, "the amount left for the project office to do the P Ianni ng does not all ew for a vi abl e project,"

The death of D US EL, or a serie us delay, would have effects that would ripple across the scientific community, says Milind Diwan, a physicist at Brookhaven National Laboratory in Upton, New York. The blow would fall particularly bard on particle physicists at Fermi National Accelerator Laboratory (Fermilah) in Batavia, Illinois, wb ich ho p es m 20 18 to begin shooting a beam of particles called neutrinos through Earth to an enormous particle detector hunkered deep within DUSEL

Researchers at Fermilab currently shoota neutrino beam to a detectorin the Soudan mine in Minnesota. But the next generation of experiments requires a much longer baseline to accentuate the effects physicists hope to see. The distance from Ferrnilab to Homestake is twice that to Soudan, and Hornestake is also twice as deep, which means that.background radiation would be extremely Jow."If we didn't open DUSEL, we would be 'in trollble," says Pier Odd one, director of Fermilab,

However, Oddone is optimistic that NSF willremedy the problem before work grinds to a halt "From what Lunderstand, people in the administration are working very hard to fix this," he says.


perilously close to a tipping point"

A graduate of the Indian Institute of'Technology, Chennai, Suresh received his doctoral degree from MIT ill 1981 and taught at Brown University before joining the MlT faculty in 1991. He was chair of the department 0 f materia Is sci ence and engineering before becoming dean of engineering, the largest ofMTT's five schools, in 2007.

The choice 0 f an active researcher would be a departure from the 1l0Lm at NSF, which has traditionally been led by senior administrators whose days in the laboratory are mostly a memory. That trait is believed to have put him at the top 0 f tb.e Jj st of candidates to' succeed the 77-year-old Bement,

~ who is stepping down on 1 June to lead a new 8 global policy research institute at Purdue 3 University, where he has been on leave.

~ Suresh's colleague, Science Dean Marc ~ Kastner, says th<~t his departure "would leave ~ a tremendous bole for MIT to fill. BI,[t the 13 scientific communi ty would be lucky to have 5 him at NSF;" -JEFfREY MERVlS



Hardy Cotton-Munching Pests Are Latest Blow to GM Crops

cotton farmers in Gnjarat state chose Bollgard Il in 2009, and bookings indicate thai more than 90% are expected to plant it this year:

Experts agree with Monsanto's statement that "resistance is natural and expected." But some dispute Monsanto's claim that this is the first report of Cry l Ac resistance. Bruce Tabashnik, an entomologist at the University of Tucson, says that uonindustry scientists had earlier reported res i stance to Bt cmps .in South A fricaand the United States.

One prominent researcher questions whether the Gujarat bollworms truly are resistant. MQlls~Il1I.o's conclusions and methodology are "fl<lwed" charges CICR Director KesJlav Raj Kranthi, an entomologist. In 8 years of monitoring Bt cotton, be says, CICR has "not

Bitter harvest. Monsanto claims that in GOjarat, large numbers of pi nk bollworms supped on 61 cotton=and su rvived.

found any resistance." Kranthi argues that Monsanto "should have analyzed tens of thousands of specimens before making this claim. ... It's a mystery Monsanto is hying to kill its own technology." Monsanto disputes that charge; it says its resistance tests were "standard practice" but declined to elaborate on its methodology

Assuming the phenomenon is real, it's a wonder it took so long for substantial Bt resistance to evolve, some scientists say. "I hope that this episode will cut down on the belief .. , that Bt has some magical immunity to re si stanc e," says Go ul d. I f as a re sul t regulatory agencies promote better rules for managing resistance for a range of pests and crops, be says, Monsanto 's findings "could be a blessing ill disguise," -PALLAVA BAGLA SCI ENCE VOL 327 19 MA RCH 2010


NEW DELHI-Monsanto has revealed that a common insect pest has developed resistance to its flagship genetically modified (GM) product in India .. The agricultural biotechnology leader says it "detected unusual survival" of pink bollworms that fed 011 cotton containing the Cry l Ac gene from the bacterium Bacillus thuringiensis (Bt), which codes for a protem that's toxic to many insect pests. In a statement to Science, Monsanto claims that the finding from western India "is the first case of field-relevant resistance to Cry J Ac products, anywhere in the world,"

The announcement hands OM critics a new cudgel It "certainly results in the anti-OM lobby having extra ammunition," says Fred Gould, an entomologist at North Carolina Sf<l-te Uni vers ity in Raleigh. "Thi s should be an eye opener." says Pushpa M, Bhargava, fonner director of the Centre for Cellular and Molecular Biology in Hyderabad and an archopponent cfUM crops who helped disallow the Indian government to prohibit commercial planting of G M eggp lant (Science, 12 February . .p. 767) . "India should immediately put a 1 O-yeannoratoriWll onthe lise and I cultivation ofGM organisms."

In a 5 March statement, Mensanto said that during field moni- I

toting of the 2009 cotton crop in Gujarat state, its scientists collected "large numbers" of pink bollworms from Bongard cotton, a first-ge.nerarion OM hybrid expressing a single Bt protein, (Newer Bollgard U hybrids produce two Bt proteins.) Back in the lab, the insects were fed Bt toxins at 1l00IDa lly lethal concentrations-and survived, the problem appears to be isolated but Monsanto says it reported its findings "to key stakeholders so appropriate decisions can be made." Some say the company aims ID shift customers to the pricier Bollgard 11

India is the second-largest cotton producer after China. Palmers first sowed OM cotton in India in 2002, and by last year they were cultivating it 011 8.3 million hectares, or 83% of the. country's total cotton crop, estimates the Central Institnte for Cotton Research (CICR) in Nagpur. EVen before the resistance revelation, Indian farmers were- adopting Bollgard U cotton.According to Monsanto, more than 65% of



China Amasses War Chest to Confront Its Environmental Nightmares

BEIJING-China's third-largest freshwater lake, Taihu, is a microcosm of what is going right-and wroug-e-ia the world's economic dynamo. Buoyed by manufacturing, the two provinces surrounding the lake, Jiangsu and Zhejiang, are enjoying sizzling growth .. And Taihu, which provides drinking water for more than 2 million people, sustains one of China's most important fisheries for crabs, carp, and eels. But it is ailing. Nutrient-rich sewage and industrial runoff have turned Taihu into a toxic soupand fueled vast algal blooms in recent summers.

Taihu and other ecological wrecks are now squarely in tile government's crosshairs. Ina nod to rising public expectations, China's government work plan for 2010, rolled out

last week at the country's two major annual political powwows, puts the environment front and center. At the National People's Congress (NPC), officials announced that science priorities includ e new en ergy sources, en.ergy conservation, environmental protection, and marine technology. Plans call for $20.7 billion to be spent mostly on engineering solutions for environmental woes. New initiatives are planned in health and food safety as welJ (see sidebar).

Cleaning up China's Augean stables is critical to the new strategy. At NPC and the Chinese People's Politi cal Consultative COl1~ ference (CPPCC), together known as Liang Hui, delegates outlined projects to transform cities arid provinces into incubators of a "low-carbon economy." "We will work hard to develop lowcarbon technologies, promote application of highly efficient, energy-conserv mg technologies, and develop new and renewable energies," Premier Wen Jiabao declared in a report to me.

Some sci enti sts hail this as a defining

Green GDP? CIrino plans agg ressive rneasu res to dean up messes like Taihu lake.

Fresh Momentum for China's Science Juggernaut

BEIJING-Pressing social problems such as the rich-poor gal'! and the rights 01 migrant workers eclipsed science at China's an nual exercise in socialist democracy: last week's meetings of the National People's Congress: (NPC) and the Chinese People's Political Consultative Conference (CPPCc), or Liang Hui. But science won hig h prai se and a hefty bu d get boost. Centra I government spending on science and technology is slated to rise 8% to $24 billion in 2'010. And science features prominently in several (PP(C proposals.

"We need to I" ma nci pate au r minds and


boldLy make breakthroughs and innovations," Premier Wen liabao said jn presenting the work plan at NpC. He said his government "will make fa rsi g h ted a rra ng em e nts" to s u ppo rt seve r a I research areas, including nancscience, climate change, aerospace, and oceanography. Wen also pledqed to "enerqetically attract hiqh-caliber person neL from overseas."

AGcording to figures released last month by the State Bureau of Statistics, China's total R&O spending in 2009-from central and local govern merrts an d in d u stry-was $ 79 billion, includIng $4 bi ILion for basic resea rch, This

moment for China. "I'm optimistic that we can start to bring development and environmental protection into harmony,' says Lu Yong!ong, an euvironmental management professor at the Research Center for Eco-Envirenmental Sciences of the Chinese Academy of Sciences (CAS) in. Beijing.

Public disaffection over China's myriad environmental ills is rising. One response, from the powerful National Development and Reform Commission (NDRC), is a blueprint to clean up bot spots such as Taihn and other polluted sonrces for the South-to- North Water Diversion Project. This $75 billion plan. would redirect water to the arid and heavily populated northeast, Meanwhile, 30 CPPCC proposals promote sustainable development of Poyang, China's largest freshwater lake (Science, 23 October 2009; p. 508). Unlike'Iaihu, Poyang is fairly clean and offers a testing ground for water-protection measures that are "win-win for ecology and economy," NPC Vice Chair Hua Jianrnin said on the eve of Liang Hui.

In addition, as part of an effort to "cornprehensively improve the nITaJ environment," NDRC said it would bring under control soil erosion on the Loess Plateau northwest of Beijing and protect tile Qinghai-Tibetan plateau's fragile ecology. Toward those ends, Wen vowed in 2010 to "accelerate afforestation, increase forest carbon sinks, and expand our forests by at least 5.92 milJion hectares." Tree planting last year hi! a target set by the State Forestry Administration to have one-fifth of' China's land area forested by 20 I 0, up from 18.2% in 2006. Another program aims to improve "ecological zones around the sources of the Yrulgtze, Ye.lI QW, and Lancang livers."

A more slippery subject is China's aspirations for a "low-carbon economy't'Tbat expression cameup over and OVe.r ill CPPCC meeting sessions and in proposals to the central government. "Low carbon is a hot topic nowadays,"

year, the main basic research funding agency, th e N ati 0 n a L N atu fa I S (ien"Ce Fo u n d atio n of Chin a, iss La ted to receive a 30% budget j ncrease to $1.2 billie n.

Althoug h the pu blic ha s little input into the workings of govern ment, one forum that can effect change is CPPCC, a n a ppoi nted body whose ranks include sports heroes and other glitterati, scientists, and Communist Party officials. At this yea r 's CPPCC plena ry session,

S 43 0 proposals were su b mitted. Most are :;: ex p ecte d to be she lved, I n reel" n t yea rs. fo r 8 example, proposals to create better conditions ~ for pcstdocs and a top scientific [ourna I for ~ China sank without a trace. But a few of this 5


On the horizon. Wind power and other renewable enNgy sources cou ld be the key to reduti n9 China's energy intensity.

says Chen JtIDWU, a chemist with China Petrochemical Corp . in Louyang. What those buzzwords mean is another question .. "There is alot of talk but not many specifics," Chen says.

At Liang Hui, low carbon appeared to encompass My activity that reduces energy intensity, or the amount of energy consumed per unit of gross domestic product (GDP). In the run-up to the Copenhagen Climate Stunmit, China pledged that by 2020 it wou.ld 'reduce energy intensity 40% to 45% from 2005 levels. Meeting fhat goal will be a stretch, At an NPC news conference on 10 March, NDRC Deputy Director Xie Zhenhua noted that efforts to meet a previous goal' cut energy intensity 14.38% from 2006 to 2009. But an analysis of GDr growth rates published by China's State Bureau of Statistics and total energy consumption data pegs the decrease at 8.2%, says Chen. The discrepancy may lie in how China's GOP is tallied. Several Hang HIli delegates questioned why the sum of provincial GDPs has been higher than the national GOP for years; in 2009, the differ-

=> ence was more than 8%.

8 No matter the precise figure, "increasing § energy efficiency by such a. large amount is o

{ not technically feasible," asserts He Zuoxin, a

~ pbys i cist at CA S's Institute of Theoretical ~ Physics in Beijing. Much of Chinas effit<


g yea r's preposa Is have passed party vettin 9 and


~ are bei n g ta ken sariou sly, in dudi ng i~ eas fo r

o Iorqinq a low-carbon economy (see rnam text). ~ Re llecti n 9 p e rs iste n teo ncerns .ove r food ~ safety, several highly rated eppee proposals ~ call for a regulatory overhaul and improved ~ surveilla n ce and ris k a ssessrn e nt, Th ese ji be ~ with the plans of the government, which last ~ month established a food security committee ~ chaired by Vice Premier Li Keqiang. ~lt'5 diffi~ cult to change the regulatory system g overnight," says health minister Chen Zhu, a g cppec delegate. But the new committee will ~ improve coordmatum, he says. And increased ~ (a padty fo r su rveilla n (e a nd risk assessment,

ciency savings so fa r have come from shuttering energy-chugging, high-polluting facto- 11e8. But China is running out of such soft targets, He says.

Further gains could be achieved by moving from a manufacturing-driven t.o a servicedriven economy, as developed nations have done. But the structure of China's economy has shifted little in recent years. A more promising approach could be to slash fossil fuel consumption and increase renewable and. nuclear energy use. PresidentHu Jintao has said China would strive to increase nonfossil fuel use to about 15% of total energy consumption by 2020. (It now stands at about 8%.)

Moving to a low-carbon economy and stemming pollution are immense challengesand Taihu sums up the complexities, Lus team recently completed a study for China's National Audit Office of the effectiveness of past cleanup efforts. Although municipalities have made strides in clamping down on industria] effluents into the lake, they have largely failed to tackle pollutants from homes and small businesses. Each province expects the other to take the lead. "It's a tragedy of the commons," says Lu, president of the International Council for Science's Scientific Committee on Problems of the Environment. "There's a long way to go."

To speed up progress, Lu advocates installing real-time monitoring systems to identify "who is doing the polluting." A second idea is for the central government to appoint directors of provincial environmental offices; too many these days are beholden to local interests, including the paramount interest of raising GDP at any cost "The central government needs to make such reforms," Lu says. He and others believe the time for action has finally arrived.

-HAO XIN AND RICHARD STONE With reporting by Li [iao,

Ch e n says, "sh ou ld be en h an (I'd wi tho ut any delay."

One CPPCC delegate has urqed the governmen t to pay more atten tio n to ra re .d iseases. The proposal from ti Ding-Guo of 5hang hai Second Medical University urges the health min istry 10 im prove tra cki n 9 of ra re d isea ses an d establish a n office to oversee orp han drug procu re men tan d ensu re ins u ran ce cover ag e. The proposal, which has also passed vetting, "is a bsolutely fig ht, " says Ch en. "5 pecia I policies: wiLl be made to en su re the availa bility of essential orp han drugs."


With reporting by Hao Xin,




From Science's

Online Daily News Site

Psychopaths Keep Th eir Eyes Of] the Prize

Whether it involves gambling away one's hfe savings or committing one murder alter an other, a psycho path in evita bly leaves the rest of us wondering: What was going on in his head? Now researchers report that part of th e answer may be hyperse nsitivity to rewards. which may create a pathological drive for money, sex', and status.

Carbon-Capture Method Could Poison Oceans

To help cool a warming world, some scientists have $uggested fertilizing the oceans with iron. The idea is to stirnu late vast bluo ms of phytoplan kton, which sequester carbon dioxide. But 5U ch an approach could have deadly consequences. hperime nts in th I' su b,Arctic show that phytoplan kton in some regions of the Pacific produce domoic acid, a neurotoxin that has killed wi Idlife and peo pie in coasta I areas. http://bit. Iy/oceaniron

CSI's Latest Clue-Bacteria Crimina Is a lrea dy have to be carefu I not to leave fingerprints 0 rONA that cou ld incrirnin ate them. But they might want to carry hand san itizer, too, at-cordmq to a new study that suggests a new way to finger perpetrators fro m their ski n bacteria.

Pardon, Your Thoughts Are 5 hawing Your thoughts leave a trace-and it's visible. Researchers. have successfully

id entified th e memo ry a person is reca lling by analyzing their brain activity. The result offers new insights into how and where the brain records memories and may help scientists understand memory impairments caused by injuries, aging, and neuro -Icqica I co nditi on s such as a stroke. http://biUy/rnemorytrace

Read the full postinqs, comments, and more on sdencencw.sriencemaq.orq. SCI ENCE VOL 327 19 MARCH 2010




Polish Science Reforms Bring Fear' and Hope

J fat first yo u don't succeed, tty, try aga i n. That's. a goodmotto for those now seeking to reform how Poland funds science. III the 19908, the country set up all independent agency that would hand out grants to individual scientists selected by peer review and would evaluate Poland's universities and research institutes. In 2002, Poland suddenly dissolved the body. "Itwas a step backwards," apparently motivated by the government's desire to exert more control over who got money, says Maciej Zylicz, executive director of the Foundation for Polish Science (FNP) in Warsaw.

Now it's back to the future . .After lobbying by PNP, the Polish Parliament this month began voting on legislation creating a new national agancy charged with distributing competitive grants for frontier research, The proposed 'Nationa 1 Center tor Science (N eM, to be: located ill, Krakow, is meant to be free from political pressures and would use an internarional peer-review systern modeled on those of the European Research Council and the U.S, National Science Foundation. NCN would also earmark at least 20% of its budget to grants for scientists under age 35. Michal Kleiber, president of the Polish Academy of Sciences (PAN), sees in NCN the type of reform the country's scientific community needs, "No doubt the system in Poland is. in spite of many attempts in the past, still not competitive enough," he says.

NCN is expected to soon be. approved. but the agency a nd other science-related reformsunder consideration have met with some resistance and skepticism. Money for research should first "be

"No doubt the system in

Poland is, in spite 0/ many attempts in the past, still not competitive enough. II

increased two, three, four times for 2 to 3 years, then yen can regulate who can usc those rcascnabl e [amounts of] money in a good way," says Julian Srebrny, a nuclear physicist at Warsaw University who is a member of the

Reform movement, Maciej Zylicz lobbied for changes in Poland' 5 sriente-fu ndi ng system,

'what little money isavaila ble goes to research instittltes and universities as block grants following an evaluation procedure that fails to adequately reflect differences In performance, says Zyliez.In 2008, less than 12% ofthe41 01 million zloty (about $1400 million) Poland spent on research was available to researchers submitting proposals outside of predetermined national research priorities, which is far from sufficient, adds.ZYlicz ..

The Budujemy IW wiedzy (Building Upon Knowledge) refQI1T1 package 'aims to change that: It pledges to distribute by 20lS about ha I f of the nationa 1 research budget to ind ividual scientists via competitive awards made by NCN and the Naticnal Centre for Research and Development (NCBiR), a funding agency set. up in 2007 to handle applied research proposa Is respon ding to national strategic priorities. Another signi ficant change is the plan to regularly submit governmentfunded research institutions to a more efficient, independent, peer-review evaluation with an aim 'to vromote 111e best. The reforms put underperforming institutions tinder "real threat ... to be closed down," ZyUC7 says. They will receive funding for Mother 6 months arid then "get a limited time to restructure," adds Poland's undersecretary of state, Jerzy Szwed, who bas overseen the reform bills.

For now, this threat applies mainly to Poland's 200 or so state-owned R&D units,


Committee forthe Development of Science in Poland and the Polish trade union Solidarity,

Twenty years after the fall of the Iron Curtain, Poland is still grapp ling with economic problems stemming' fromits communist past. According to Eurostat, the PolisH government's R&D budget in 2007 'was 0.32% of its GDp, representing just 1.03% of the overall budget and placing Poland among the lowest science nll'idersin the European Union, And most of


most tracing back to the communist era, although the 70 or so generallywell-regarded PAN research institutes will also be evaluated. As part of an even more controversial higher education bill currently under debate, underperforming universities would. probably have funding withdrawn too, ZYlicz says. Poland. would also label the best-performing university departments, as selected by international exp efts, as centers of excel! ence. The se Nationa] Leading Scientific Centers would for 5 years receive more than 10 mi Ilionzloty (about $3.5 million) annually for research, faculty's salaries, and Ph, D, scholarships.

Many ill the Polish scientific cornrounity are concerned with how the reforms will be implemented. "Some careful attention must be paid to research Which has a long-term nature and cannot be c~lfljcd out .yithOllt ,si~· ble financing extending over many yean," Kleiber says. "It is important that the evaluation teams consist of independent, wel.!skilled, and objective experts," adds Adam Hamrol, rector of Poznan University of Technology and vice-president of the Conference of Rectors of Academic Schools in Poland.

A big P31t of getting the scienti fie communitytowelcorne deep reforms was the Polish government's commitment in 2008 to increase science funding, The global financial crisis has slightly eroded this promise, however. Starting with 5202 mi llion i:!oty (about $1800 mill ion) this year, the govermnent has pledged

to increase funding for research by 13% every year through 20 l. 3. extra money that may now go to both NCN and NCBiR. Many scientists still find this too good to be true. It wouldn 't be the first time the government promised new funds for science and. "nothing happened,' Srebruy says. University of Warsaw plant biologist Stanislaw Karpi ']1 ski, a recently returned Polish scientist, adds that an insufficient increase in the science budget could delay and dilute necessary changes in ~ the Polish science system. ~

Many, including Zylicz and Kleiber, ~ advocate implementing the reforms step by ~ step. "There is a gena-fa I fear of change," ~ Hamrol notes. But hopes. too, are running ~

~ . '<

high. "We are SOre that changes, if wel] ~

introduced and accepted, will bring tremen- .2 dous contribution til the further develop- ~ ment of Polish science," says Hamrol. ~



Elfsa beth Pai n also profi les Polish scientist Agnie~2ka Chadnska at Scienc(!, for which she is a contributi ng editor.



Male Rivalry Extends to Sperm in Female Reproductive Tract

For males of some species, mating is Just the first step toward winning the battle to pass. on their genes. Females sometimes mate more than once in quick succession, filling their reproductive tract with rival sperm that must compete for access to the unfertilized eggs. Two groups now show details of what life. must be like for those sperm, with one offering unprecedented movies 0 f this sperm competition. On page 1506, Susanne .P. A. dell Boer of the University of Copenhagen demonstrates that such rivalries in SOUle ants and bees have led to the evohaion of seminal fluids containing toxins that impede rival sperm and to female: fluids that counter these toxins. Another team, reporting ani inc U1 S;;iencll ( c g i/e au t e n t/a b s t ra c t/ science, 1187096), followed redor green-glowing sperm as they jockeyed their way through the reproductive tracts of fruit flies. Both papers drive home the point that "the competition between males continues in a very fierce way" inside the female. says Tomrnaso Pizzari, an evolutionary biologist at the University of Oxford in the United Kingdom.

The papers provide a glimpse of where evo lutionary bi ology research is going: The femllie reproductive tract is "one of tile great, unexplored frontiers for the fiel ds of sexua I selection and speciation," says Scott Pitnick, an evolutionary biologist at Syracuse University ill New York state.

Pitnick's Syracuse colleague. John Belote entered this frontier by developing two fruit fly lines that produce different fluorescent pro reins ill the sp erm head one gre en and tile other red. After allowing female fruit flies to mate with one strain and then the other a few days later; Pitnicks postdoctoral fellow Mol-

~ lie Manier videotaped the streams of red and :i': green sperm, tracking their interactions in real ~ time. The first sperm in the reproductive tract ~ swim to the fly's sperm-storage organ, but ~ many are displaced by the second wave of ~ sperm, she found. However, once both males' ::;;:

~ sperm Were settled, they al L Seemed to have ,Ui

~ equal chance (If fertilizingan egg.

~ "This is one of the most exciting devel~ opmentsin evolutionaryand reproductive

biology=end will revolutionize the field," says Tim Birkhead, an evolutionary biologist at the University of Sheffield, United Kingdom. "After seeing these videos, researchers will now think of sperm competition in a new way."

Den Boer, University of'Copenhagen colleague Jacobus Boomsma, and Boris Baer, now at tbe University of Perth in Australia, find sperm i11 Some bees and ants do more than physically displace rivals. The team compared spenn dynamics within ant and bee queens t1Hl.t mate only once with ones in which females mate multiple times. duringa single courtsh ip flight and store sperm for years. For tbe multiple mating species stud-

Close quarters .. Riva I spe rm, la baled red a no green, jostle 0 ne another in the Dro50phila female's seminaLreceptade.

ied, two leafcuttcr ants and the honey bee, seminal fluid from a given ill <I Ie enhanced the survival time of its own sperm ill alab dish but damaged ll11n:latcd sperm and even sperm from il brother. Adding sperrnathecal fluid that ant queens make within their reproductive tract countered these effects, says Boomsma, In contrast, seminal fluids from singly mated bumble bees and ants showed none ofthese negative effects.

Sperm facing competition have evolved some as-yet-to-be-defined seminal fluid components that somehow recognize and thwart rivals, says Boomsma. But once the sperm reach their destination for long-term storage, the female apparently wants to keep all the sperm healthy and has evolved ways to counter the seminal flu id. Thi sshldy "beauti fully revealsjust how nuanced reprodnction can be," says Pitnick. "TI1c["('l will he much to gain from combining om' respective approaches." -[ELIZABETH PENNISI


. S'c£encelnsider --_

From the Science PolicyBlog

Conqression al su pporters of stem cell research have introduced legis latio n to endify President Barack Obama's 2009 executive 0 rder, which lifted restrictio ns on I he number of human ernbryonic stem cell lin es availa ble to federa lly fu nde.d

resear ch er 5 .. h ttp :llbit .Iy/b d ka 5 n

A University of MiChigan, Ann Arbor,anthropologist has criticized the u.s. government for not rna king better use of social. scientists in fighting terrorism.

http://b it ly IbZ6z R P

The InterAcademy Council, comprised of national science academies', wilL Lead a 6- month review of procedu res of the Intergovernrnental PaneL on Climate Chan,ge. The tevi~w will Look at transparency, conflicts of interest, a nd rules pertaining to (he qug,Lity of data.

The lsraeli government has launched a 5350 million .effa rt to .ture back I sraeli scientisb. worki ng abroad .. The effo rt wi II involv€ 30 new centers of academic excellence, funded by the govern rnent, academic institution s, and charities, http://biUy/ctAuVV

A veteran undersea robot operated by the Woods Hole Oceanographic Institution was lost during a research expedition off the Chilean coast. The Autonomous Benthic Explorer (ABE) had reached a depth of 3 kilometers in the first stages of its 222nd dive. http://b it .ly 19 pTYB 0

A lonq-runninq battle between the U.S. govern ment and il grou p of 29 scien Lists and enqi neers of the Jet Prop ulsion La boratorv (]PL) over privacy rights has now reached the S~preme Court.

http://bi Uy/b6H3 6T

The RoyalSoclety has released a report 0 n the tutu re of sci entific research ill the United Kingdom that calls for broad funding in creases in line with those of other countries that have included scientific research in their economic stimulus packages. 2r

For th e fu II posting sand rna re, go to nIMs.sciellcemag.orgfscienceillsider. SCI EN CE VOL 327 19 MA RCH 2010


The Ni,leDelta's Sinking Future

Climate change and damming the Nile thro8ten Egypt's8gricultufa.1 aasls





QALYUBJA, EGYPT -After fighting his way through Cairo's apocalyptic traffic to [eave the city, Atef Abd El-Rahman's progress is stalled again, first by a. donkey cart ambling slowly ahead and later by ashouting match among severa! men ill the middle of the street Finally reaching his destination, the engineer at Egypt's Ministry of Water Resources and Irrigation (MWRI) leaves the car behind and walks down a path into the quilted fields of a farm .. The owner of the land, sipping tea in the shade of a tree, introduces himself as Fathy Abdelaleem, Abd El-Ralnnan praises the fanner's irrigation system, a network of troughs as intricate as a computer circuit, delivering water to his plants from one of the many canals feeding off the Nile River. Abdelaleem says he dug it and connected it himself to the canal: "Now r have no problem with water,"

Here in the Nile Delta, starting just a few miles north of Cairo, the world's longest liver divides into a fractal

pattern of ever smaller livers, canals, and irrigation ditches. Before the Nile reaches the Mediterranean, every drop ofwater is put to work by ingenious

delta fanners like Abdelaleem, collectively known as the fellahin, Like most, Abdelaleem may have skirted a few laws while constructing Ius irrigation masterpiece=-but then, few fellahin are even aware of the dizzying array of water regulations issued by Carro. Environmental policing is "more difficult than you might think," says Essam Khalifa, a deputy director of MWRL For example, he says, "two-thirds of the wells in the delta are illegal." A new law will mandate stiff fines for unsanctioned pumping, though similar efforts have had little effect.

The only aspect of the Nile Delta over which the government exercises finn control is the input of water. The Nile was first dammed by the British ll11902 at Aswan. The Egyptians began construction on the far more massive Aswan High Dam in 1960 with the help of the Soviet Union. The hydroelectric dam changed life dramatically in the delta. It provided electricity and flood control to one of the most densely populated areas in the world. "Most people agree that it was for the better," says Farouk EI-Baz, an Egyptian earth scientist based at Boston University.

Yet as Egypt celebrates the 50th anniversary



N Pod cast interview I!.C WIth author

loh n Boha nnon.

... Fragile oases. Only 5% of Egypt's land area is inhabited. and some 50 million people pack into the delta north of Cairo that draws life from the waters of the Nile River.

VOL 327 SCIENCE www .

of the start of the High Dam's construction-e'f took 10 years to cornplete-e-some scientists say that this wonder of engineering is contributing to an environmental catastrophe that could force millions of fellahin to abandon the ILlSh, fertile delta. "It is now clear that [the High Dam] is having negative impacts," says Wahid Moufaddal, a remote-sensing scientist at the National Institute of Oceanography & Fisheries in Alexandria, Egypt.

The worst of these is coastal erosion and subsidence, the compacting of the delta soil. For millennia, the untamed Nile compensated for these natural processes by delivering fresh sediments along with its fresh water. The dam, however, now blocks the sediments far upstream of Cairo. As a result, the delta is sinking. Today, 30% of the land is less than a meter above sea level, and in some areas close to the Mediterranean coast, it is sinking by nearly a centimeter per year.

At the same. time, the Mediterranean Seals expected to rise. as a result of global warming, If the sea level increases by a meter by 2050, which is in the range of mainstream predictions, one-third of the delta could be lost. M eanwhile, the population here. is growing by a million people per year-the delta is already home to 50 million, most crammed into an area no bigger than the state of Delaware. Because of'these perfect.storm conditions, the Intergovernmental Panel on Climate Change in 2007 named the Nile Delta among the three areas most vulnerable to climate change. "If we continue with business as \1S11a1, the impact on the delta will be devastating," says Mounir Tabet, director of the United Nations Dovelopmeut Programme (UNDP) in Egypt.

Deciding on a course of action is easier said than done, because the rates at which the sea is rising and the delta is SUlking are subjects of fierce debate. As a first, cautious step, Egypt and the United Nations are this year Iannching a 5~year study of the options for protecting the

tl delta from the encroaching sea. "We do not ~ have enough data to advise policy yet," says ~ Taber, But as Egyptian scientists rush to pro~ vide those data, the government is steaming


g ahead with a series of"megaprojects" to boost

:.; the country's habitable area. In the most ambi~. tious of these, the largest pump in the world is g; diverting 10% of the Nile into an uninhabited ~ region offhe desert to create a new deJta.



~Tami"g the Nile

~ Because the river is fed by the variable rains ~ of the vast eastern arid central African high~ lands, the Nile's pulse wries with the seasons ~ arid even Longer cycles. "The Bible describes ~ a 7 -year cycle of flood and drought," says 5 EJ -Baz, "And we now know that there:rca lly is


Every drop used. By ma ppi ng the canal network 01 the Nile Delta Unset), geologist [ea 11- Daniel Sta nley discovered how farmers use the water SQ efficiently that the river 110 longer reaches the sea.

a cycle that varies between 4 and II years." That makes life on the delta a gamble. Before the High Dam, floods wiped out homes and farms, while droughts brought famine and disease. But with their hands firmly on the faucet, Egyptians have tamed the Nile. An entire generation has never seen a flood. Instead, steady irrigation is now possible throughout the year. Ratherthanjustone crop, delta farms can grow as many as three. Abdelaleem says he rotates wheat, taro, and clover on his.

At the time, the damming of the Nile raised few concerns, "There was 110 discussion" about the merits of such a potent source of pride for the newly independent nation, says Monfaddal. "It was a giant experiment," yet, he notes, there was no plan fOJ collecting environmental data.

Moufaddal recalls that the first "whisperings" about a negative impact began in the 198 Os. "Off the coast of the delta, 0 ur very

important fishery for sardine and anchovy started dying." Pointing to a graph of the annual fishery harvests, he says, "you can see that it later made a big recovery." These fluctuations remain a riddle. "The best theory is that the fish crashed when the dam stopped the sediments and then recovered because of plankton blooming from sewage," says Moufaddal. But the gathering of environmental data has been so scant, be adds, that it is difficult to know "what is a direct cause of the dam and what is natural variation, We baseline data from before 1964." EI-Baz puts it more bluntly: "No one gave a damn before the dam about science."

Complaints about the government's envirorunental research policies are a common refrain here. "What frustrates me as a scientist. is that our government has blocked us from studying the problems," says Moufaddal, "If I want to attend a conference abroad, or even to take my research ship out to do sampling, it is sometimes a year before 1 have permission," Moufadda! concedes that some caution in doing fieldwork off the coast is reasonable, considering the animosity between Egypt and neighboring Israel. "But r am often not allowed even to take soil samples right here in the delta. How is that a security issue'?"

Master of water. Delta tanner Fathy Abdelaleem looks over his irrigation ditches.

www.sciencem ag .org SCIEN CE VOL 327 19 MARC H 201 0



Now that climate change is considered a major threat to the country's security, attitudes, may be changing, "The government is relying au Egyptian scientists," says Tabet, The problem is, that scientists do not have simple advice to give.

Rising seas, sinking land

<'We used to relax an day on the beach right there," says Moufaddal, pointing to a rough patch of surf From his perch on Alexandria's new seawall, the coast of the Nil e Delta stretches to the east as a seemingly endless row of hotels and apartment buildings. He scrambles down to the city's original seawall, now crumbled and licked by the Mediterranean waves. "Here it is obvious that we are losing our coast to erosion," be says. "And if the water keeps rising, we have a problem."

Daniel Stanley, a geologist at the Smithsonian Institution ill Washington, D.C.

Stanley became the original authority on this. issue after he undertook the first comprehensive study of the delta's geologic history in the 19808. The delta is defined by the Nile's tW(J main branches, which split just 1101th of Cairo-c-one heading northwest to Rosetta and the other northeast to Damietta. During his fieldwork, Stanley noticed marine barnacles thriving kilometers inland within tile river's outlets. "I was shocked to realize that the Nile was no Ionger flowing into the sea," he says, "Where was ail. that water going?"

Stanley requested a detailed map of all the waterways between Cairo and the Mediterranean, To his surprise, no such map existed, So with the help of several colleagues, bema de (me himself. The picture that emerged was like

nutrients for the crops-now replaced by chemical fertilizers=-but the annual thickening of the soil countered subsidence, «AlI river deltas subside;" says Stanley. "The sediment beneath the surface is supercharged with water," As the water leaches out, the sediment compacts,

Stanley was the first to measure how much the delta naturally subsides, and he used that measurement to extrapolate the delta's future topography. He and colleagues dug 87 cores across the northern delta and determined the age of the layers going back 7500 years with radioisotope dating. The results revealed that the subsidence rates across the delta have been far from uniform over the ages (Science, 22 April 1988, p. 497). Aroundthe Nile'snortheastern outlet, Stanley estimated that the delta has subsided at an astounding rate of half a meter per century; in otherareas, the sediments have barely moved.

"There's a binge hue that runs along the northern delta," s ays Stanley. "Between that line and the sea, subsidence rates are the highest" In a second paper in Science (30 A pril 1993, p. 628), Stanley warned that even a conservative forecast of the rising sea sand subsiding land "augers poorly for the delta." For example, a relative sea level increase of I meter would flood more than 30% of the region's land

surface, he says. "But predictin!!; exactly where it will flood is difficult," says Moufaddal, because different parts of the delta are subsiding at different rates,

Adding to the complexity, the historic subsidence rates Stanley calculated contain various uncertainties related to the radioisotope dating method, As a reality cheek on the rates, Stanley is now dating sediment layers with ~ archaeological material such as pottery II shards. But. relia bIe traces, of ancient human ~ settlement are hard to come by here. ~.


Some scientists are trying a completely t


new approach to gauging bow fast the delta w


sinks. Rather than averaging the subsidence r5 G rates over the past millennia, Richard Becker ~

and Mohamed Sultan geologists at Western 'i


Michigan University jn Kalamazoo, CalC(I- ~


lated the current subsidence in the delta. They ~

used satellite-based radar interferometry to ~ mea HIre the rate at which We surface sank: ~


between 1992 and 1999. The technique, which r-

uses multiple radar images to measure. ~ changes in the altitudeof thousands of land- ~ marks such CiS buildings arid L1tility poles, is (5

The cost of control. Now that the Nile no longer floods the della, polluti on in {On nected cana Ls (left) and wasta l erosion (right) are growing problems,

The scientific consensus is that the Mediterranean will indeed rise as part of a global increase in sea levels, but how fat and how fast is another matter. "What we do know is that the temperature of the ocean is increaaing," says El-Baa.and water expands as it warms. "So this part is easy to predict." M ore worrying to cl imate scientists than this thermal expansion is whether global warming will trigger the ice on top of Greenland and Antarctica to break up and slip into the water. That is why forecasts for sea-level rise by the end of this century have ranged between 02 and. 2 meters (SCience, 5 September 2008, p. 1340). The former would be a manageable nuisance for the delta, whereas the latter would be catastrophic, all agree.

Predicting the future location of Egypt's coastline is especially complex because of the uneven sinking of the land, Yet whjlc scientists may disagree about the rate of sinking, few are optimistic. "The Nile Delta is l10W in its destruction phase," says Jea 1)-

the dense fi ligree of neurons in a brain, with more than! 0,000 ki lometers of canals and diversions branching into every corner of the delta. The reason the Nile no longer reaches the Mediterranean, he says, is the efficiency of the fellahin. With the flooding stopped, says Stanley, "they use every last drop,"

The Nile's flooding may have been destructive, but it was also crucial for the health of the delta. Before 1960, the Nile delivered an average of 100 million tons of sediment each year to. the Mediterranean. Although the prevailing eastward sea current constantly nibbles away at Egypt's coastline, the sediment from tile Nile easily compensated for that loss. But by 1970, when the Nile had formed Lake Nasser upstream of the Aswan dams, "that process stepped completely," says. El·Baz."The. sediment enters the lake and it's like hitting a brick wall. The water slows and the sediment dumps,"

Before the Nile was dammed, theannual fl ood depo si ted arnill i 111 ete r 0 f fresh s ilt onto the delta's surface, Not only did it provide



sensitive enough to detect movement as small as 0,1 millimeter per year. The results, published in the September 2009 issue of The Holocene, found higher subsidence rates than Stanley calculated, suggesting that some areas could be submerged sooner than already feared. In particular, the fastest rates were in the youngest sediments, such as those beneath the coastal city of'Damietta,

Even if one accepts these current subsidence rates, predicting the future topography of tb.e delta may be impossible, says Stanley. Besides the steady leaching ofwater from the top layers of sediments, he says, "there are things happening deep below." Stanley SlISpects that sudden subsidence can be triggered when seismic jolts reorganize sediments laid down by the Nile long ago, "The della we see is only the tip of a massive structure,"

GoiOg Dutch

1.'11 spite of the uncertainty, Egyptians are already planning for the future, "There are many options on the table," says Moufaddal, Some seem inevitable, such as the relocation of delta residents affected by coastal flooding. Others seem like science fiction. "One ideals to stop sea level rise by blocking the Mediterranean at Gibraltar," Moufaddal says; shaking his head. Moufaddal does not support even moderate plans for flood prevention, such as tidal gates, sea walls, and sand dunes. "You can't stop the sea," he says.

However, stopping the sea, or at least hampering its landward thrust, is what the Egyptian government is now pursuing. A $16 million pilot proj ect, launched this year by Egypt and UNDP,. will test some strategies for elimale change "adaptation" ill the low-Iying della. The 5-yea1" project includes "strengthening of sand dune systems, beach nourishment, [and the] establishment of engineered wetlands," says Mohamed Bayoumi, the project's coordinator,

According to Stanley, saving the delta will require far grander schemes. "The Netherlands faced exactly the same situation," he says. One-fifth of the Netherlands is below sea level Mer catastrophic flooding in 1 953, the nation spent billions on the Delta Works, an elaborate system of dikes and storm-surge barriers designed to stop the NOIth Sea from devastating the Dutch coast. "What Egypt needs is a Great Delta Works," Stanley says. But considering the price tag, "there's no way Egypt cart afford it."

§ Even if the sea can be, stopped, Dotes


~ Stanley, the delta is facing a crisis in water

~ quality, Without the annual fl cods flushing ~ the delta clean, sewage, fertilizers, and 5 industrial waste "go nowhere,"

Problems also lurk below the delta. The Nile water that does not evaporate seeps dOWJ1 fhrough the delta soil where it heads gradually for the Mediterranean like a sluggish mirror image of the river above. Because the canals are polluted, the fellahin use thisnaturally filtered water for drinking. But iftoo much water is sucked up for consumption, saltwater from the Mediterranean intrudes into the aquifer. A "salt wedge" is now creeping in, rendering well water as far as 30 kilometers inland too salty to mink. Wi thou t potable water above or below ground, says Sta:nley,"how are people going to live there?" In an effort to push back the salt wedge, Khalifa says that his ministry


30 meters between tbe river and the wafer's final destination, the government built the largest pump in the world to lift the Nile's water up to canals leading to Toshka. Since the pump fired lip in 2005, more than 2000 square kilometers of desert have been irrigated. According t.o Moufadda1, however, Egypt hasn't seriously thought about. the environmental impact. 0 f the project, which he fears could destroy desert habitats and hasten the demise of the Nile Delta by siphoning away its water. "We are repeating the mistake of Aswan," he says,

Egyptian officials view things differently.

"Bringing life to the desert" is the goa 1 of the

is encouraging delta farmers along the coast to switch to rice cultivation. Rice paddies are hardly an efficient use of limited fresh water, it soaks down, the fresh water C,Ul help block the sea where it intrudes.

Returning the Nile Delta. to its natural state could solve these water problems, but "removing the dam is not an option," says Tarek Hussein, president oftbe Egyptian Academy of Scientific Research and Technology in Cairo. "One idea we are discussing is to divert sediments around the dam," Moufaddal says even that is futile: "Building lip sediment will never keep up with rising sea levels,"

Although the Egyptian government has never publicly said that the Nile Delta is doomed, it is creating a new delta in the desert. Upstream ofthe darns, 10% of Egypt's share of the Nile-vamounting to 5 billion cubic [Deters of water per year-s-is being diverted southwest of Aswan to a desolate area known as Toshka. Because. the land rises

Toshka "megaproject," says Khalifa It is seen as crucial for meeting the government's goal of a 50% increase in the country's farmland by 2017. And by that date, according to a government brochure, 2 mi II ion peop Ie wi I I be living at Toshka,

"That's not going to happen," warns EIBaz. Besides temperatures. that can reach 50be, b.e says, "no one wants to live out in the middle of nowhere." The government is now promising free plots of land as an incentive for people to relocate to Toshka, but, says EIBaz, "no one is taking the offer." To increase the area's lure, he has proposed a "development corridor" of highways and electrical lines from Toshka to Cairo and.Alexandria, It would cost $24 billion over 10 years. "ConuectingToshka to civilization is the only way to convince people to move," EI-Baz says. Then again, if the worst predictions for the Nile Delta come true, millions of fellahin will be in need of a new home.






Lucky Glimpses of a Weirdly Wetter Moon

Water is turning up on tile moon in unoxpected places and quantities. but scientists almost missed the big splash. Without a bit of good fortune when NASA blasted the moon last October, they would have gotten "zippo," says Lunar Crater Observation and Sensing Satellite (LCROSS) team member Peter Schultz of Brown University.

LCROSS scientists and engineers aimed a spent rocket upper stage at the moon's frigid, shadowed Cabeus crater .. hoping to blast some rock and buried water icc into view (Science, 20 November 2009, p. 1046). But when the impactor hit its target, scores of telescopic observers back on Earth SAW no dusty debris. Luckily, the LCROSS spacecraft following close behind returned abundant evidence of long-sought water,

The near-emptiness of tile spent rocket stage made all the difference. To team rnembers' surprise, being hollow allowed the impactor to splash debris not only low and to the sides-as modeling of solid impactors had predicted-but also nearly straight up, Schultz says, Earth-bound observers couldn't see the low-angle ejecta behind an intervening mountain, and viewed side-on the higbangle plume was too meager fOT them to detect so far.

LCROSS, on the other hand, \WJs looking straight dO\1JIl the high-angle plume, giving its sensors a strong water signal. At the meeting, LCROSS principal investigator Anthony Colaprete of NASA's Ames Research Center

in Mountain View. California, confirmed the team's estimate of at least 2% water by weight in the top couple- of meters of Cabeus soil. He also reported the first detection ofwater ice in tile plurne.notjust water vapor. The ice means the water is not simply locked up in minerals.

Elsewhere 011 the moon, water is showing up through Jess violent means. Planetary scientist Igor Mitrofanov of the Institute for Space Research ill Moscow reported another find by the Lunar Exploration Neutron Detector (LEND) instrument all board NASA's LUJ1<:n· Reconnaissance Orbiter (LRO), which guided the LCROSS team to Cabeuss water. Now LEND is detecting signs of water-in the form ofenhanced subsurface hydrogenin two other areas ncar the south pole. That's remarkable, said Mitrofanov, because unlike the floor or Cabeus, these areas are not in permanent shadow. The sun shines on them and warms them. Scientists looking for water had ignored such areas, assuming they would be too warm to retain water ice.

But Mitrofanov says that planetary scientist David Paige of the University of ca1i fornia , Los Angeles, a member ofLRO's Diviner ternperature-sensing instrumentteam, has calculated that buried ice can survive outside pennanently shadowed craters. In the newly identified hydn)g;eI1-r.ich areas, Paige finds, some teIiS of centimeters of lunar soil. are enough to insulate and preserve deeper water rce, That would be gOOdl16WS for astronauts; they could mine the key lunar resource without venturing

.... Dry vs, wet? A radar sig.nature confined toa "II1II lu na r crater (right) s u 99 est, bu ried ice r ath er tha n rough terrai J1 (left).

into same of the coldest spots in the solar system, whenever:they get to the moon again.

Still, the real water bonanza may lie in the north .. Planetary scientist Paul Spudis of the Lunar and Planetary Institute in Houston, Texas, and colleagues recently flew a groundpenetrating radar on India's Chandrayaau-I lunar orbiter; At the meeting, they reported a distinctive. radar 51 gnanrre from the interiots but not the surroundings ofrnorethan 40 small craters in the north polar region. Almost all of the cratersare pennanondy shadowed.

That combination, Spudis argues, suggests ice, perhaps as much as 80% or 90% pure, But the Chandrayaan- 1 radar is not confuming long-suspect reports of radar detection of massive ice in the south (Science, 13 March 1 998. p. 1628), so most researchers are reserving judgment They do agree on one thing: Someone. Of something, needs to return to the moon to touch its water.

Coaxing Out Another Taste of the Sun

Five years ago, the out-of-control Genesis sample-return capsule bored into the Utah desert at 360 kilometers per hour (SC'ience, 17 September 2004,. p. 1(89). But things have been looking up fOI" the space mission, which returned atomic bits of the sun and thus Samples of the solar system's primordial material. "WIth that big splash at the end, it was an in teresti 11g little mi ssion ," reca II S COSlTl 0- chem j st Roger Wi ens of Los A I amos National Laboratory in New Mexico.

At the meeting) Genesis team members confirmed their measurement of the isotopic composition of the solar wind's oxygen (SC[eI1ce, 28. March 2008, p. 1756) and reported an isotopic composition for solarwind nitrogen. Compared with the solar wind, "nitrogen on Earth is heavy," pro- 0

lii uounced cosmcchemist Antti Kallio oftb.e ~

Uuiversity of California, Los Angeles ~ tV CLA) e By such threads will bang the story 5 of'how the stuff of stars became planets, 5"


The $264 million Genesis mission went _,

S\wim·m·inglY~ l~ to its rough Ianding, The Gen- ~ esis spacecraft cruised the solar system for 3 ~ years, exposing collection plates to the oJ1l11sJi~ .5


19 MARCH 2010 VOL 327 SCIENCE www.sciencemag.orq

Down, not out. Genesis had a hard landing but is yielding long ·soug ht >am ples of pri mordial matter.

ing solar wind. The samp Ie capsule reentered Earth's atmosphere loaded with embedded atomic particles from the Still. Then, its parachute failed to open tSeience, 22 October2004, p. 587). But team members picked up the pieces-all 10,000 of them-s-and proceeded to clea n and analyze them At UCLA, a team led by cosmochemist Kevin McKeegan had built the 8-ton MegaSTh1S, "a custom-made hybrid secondary ionization/accelerator mass spectremeter"-that is, ananalytical mstrument for extracting and "weighing" each isotope.

Before the: latest M egaSIMSresnlts measurements of the ratio of nitrogen's rare heavier isotope, nitrogen- I 5, to nitrogen-Ld had been all OVe_( the map. The UCLA value for the SlID (-470 ± 100 per lILiJ) looks "fairly robust," Kallio said, and roughly equals the isotopic ratio for Jupiter retrieved by spacecraft as well as the value for a primordial component of meteorites. I f the number is correct, £(111.h's nitrogen bas relatively more of nitrogen's Hire heavy isotope than the solar system's starting material does.

Snapshots From the Meetin:g »

"Row did it get that way'?" asks Wiens.

Drawing on the isotopic compositions of other elements, researchers, will be testing some possibilities: ~U1 infall of comets, solar wind erosion of the cady atmosphere, blasting of' the primordia! nebula by the sun, or some combination of these. In the meanfmc, Genesis analysts are taking up their next element, carbon.

Spirit Is Willing, Though Weak

Spirit's roving days are over, right? Back in January, NASA announced that after 6 years on Mars, the plucky rover was being consigned to purely stationary duty right where it had become stuck in a sort of dry quicksand. Rover obituaries proliferated, Butat the meeting, the talk was of Spirit's resurrection as a rover, or at least a realistic prospect of a return to limited mobility. NASA missions are indeed hard to kill (Science, 22 May 2009, p, 998).

NASA retired Spiritafter the rover became stuck last May and failed to extricate itself. Trundling toward its next target, a I ikely volcanic vent 100 meters away, Spirit broke through a surface crust and bogged down in fluffy sulfate sand. Repeated attempts to escape forward-s-guided by testing of a mockup in the Mars "sandbox" at the Jet Propulsion Laboratory in Pasadena, California=-yielded millimeters of progress and not mi;lny of

Tiny find, The Stardust mission ennounced the discollery of its first candidate bits of mineral born around distant stars. The Sta rdust spacecraft reru rned pleiiity of easily identified pa rtides from its close pass by comet Wild 2 (S.dem:e, 25 lanuary ZQOS, p. 401), bwt the rater a nd sma llerioterstellar d llst was hOI rder to spoLAn a utornatic microscope scan ned 1.6 million differe nt locations on the ~pa(ecraft' s collector at 40 diffe~enl depths at each location. But no one could leach a computer to recoq nize the tracks left by speeding stard ust as it imbed ded itself in the airy-lig hi collecto r material.

Enter the "dusters." Humans. it turns out, are easier to train for this

Striking it nch. A massive search ef a 5tard ust cQllettor tu r.n;ed up this rnierensize bit of what seems to tle the flr'>! p,ri~tJne "interstellar dust.


those, says Raymond Arvidson of Washington University in St. Louis, Missouri, deputy principal investigator of'thc M ars Exploration Rover mission. That's when NASA threw in the towel.

But then Spirit's operators figured, what the heck, let's try reverse. They also threw in a little wiggling -the-wheels maneuver intended to move some soil out of the way beforeeach attempt at driving. It worked, a bit. Instead of'a millimeter or two of progress per driving session, Spirit made several centimeters, Before it bad to be settled in for a long winter's nap as the available solar power dwindled, Spirit covered 34 centimeters backward, according to Arvidson. "I really think we can continue this kind of breaststroke ann extricate it," he says,

First, Spirit must survive the winter. Its handlers are guardedly optimistic that its power levels will not drop so far that the rover's electronics fieeze, one is making any promises. Next.September or October, as spring [eU-lITIS to Mars, Spirit will likely serve 6 months' duty as an essentially stationary geodetic marker-in a study (lfthe. planet's core.

All fl,e while, "there'll bea discussion of what Spirit beCOI!-1eS," says Arvidson. "There's a lot to do," even if Spirit-s-now down from six operable wheels to four-s-can only hobble along centimeter by centimeter. But in the end, it will be NASA headquarters that decides whether a rover with "limited mobility" is worth $27,000 a day in operating costs. If not, it will have to pull the ping.


sort of thi nf:J than cornp uters, T\'Venty-seven thousa Ad people volu nteered. Aftef 'Z1 milLion searches conducted on homecomputers, two micrometersize par~jdes along a sin'gle track~dubbed Orion (the m¥~hicil! Hl!lnter) and 5i ri us (his dog) by t~eir duster d i~colI!Her~h ave resea reh ers ~ (autiously exeited" that stardust is in band, says Stardust team member Andrew Westphal of the University of Califomia. Berkeley. If the particle, are the real McCoy. researchers estimate that. scores mere like them remain to be found.

Tilt-a-moon. OUf nearest neigbbor in space has always looked a little en, and now researchers think they lTiay know why. At the meeting, planetary scientists Lissa Ong of the University of Arizona. Tucson, and lay MeLos h of Puroue University in West Lafayette. Indiana, reported considerable progress ill making sense of the odd shape of the moon. Tidal forces should have [eft it more flattened side' to side than it 15 now. Bl!lt Ong and Meilosh fOURd that the huge: impact billions of years ago that left the great S'ol!lthPol'e-Aitken basin would have blanketed the moon with tens to hun d reds of mete rs of ~ocky debris. If that impact were i r;I th e center 'of the mOOR's: leadinf:J face-the most Likely plate fotit-e-that redistribution of mass could have given the moon its present shape, they calculate. The reshapinq WQl!Ild also have spun the moon around by as much as 90°. So, all eil~ly violent resurfacing may have turned the gaze of the Woman in ti'le MOOR from LlPward to forward. ~R.A.K. SCIENCE VOL 327 19 MARCH 2010




edited bV Jennifer Si/{s

Unlocking the Door' to Better Cyb ersecu rity

authenticated session (1). Using multiple security protocols leads to the same problem. Moreover, using the public key for aU objects in cyberspace is practically infeasible. Finally, cyber attacks. include denial of service, and preventi on of'this type of attack is still very difficult, irrespective of how secure the end-to-end protocols are.

1 believe that to ensure security in the cyberspace, multiple mechanisms should be used, including perimeter defense. intrusion detection, and application-defined security.

This concept is analogous to a treasury with ~1 locked, gate, surveillance equipment, and a guard who can identify authorized persons.


IN THE1R PERSPECTIVE ("REFlEGlONS ON CYBERSECURITY," 13 November 2009, p, 943), w. A. Wulf and A. K.. Jones claim that the perimeter defense model does not work. 1 agree. with them that a perimeter has limitations, However, just as a lock on

a frQllt door is useful even though it can be broken, perimeters provide good, albeit imperfect, protection.

"\Jill I f and J ones propose an a lternative approach in which a minimal mechanism ena bles multiple end-to-end security protocols. However, the reLi ance of the security of a ccmmuuicati on solely on the end-to-end security protocols is dangerous, because if the protocols were to be broken, the communication would also be broken. The Transport Layer Security (TLS) protocol is one such example; it has been used for over a decade but was just recently discovered to have security flaws, permitting an attacker to inject data in an

M~ rao PolytechniC I nstitute, Rua de tu is Gonzaga Gom es, M¥30. E-mail: kywong@ iprn.edo .. mo


1. U.s. Deportment of Homeland Security, United States Computer Emerg~m:y Reodin~s5

T €;J m, Vuln era bi I ity N ole VU Itl20S 41, • 5S L an d Tl5 pro to tu Is renegotia ti on vu I nerab it i t\l' (www,kb.(eft.orgNuWidI120541).


WE COMPLETELY AGREE THAT MULTIPLE MECHAnisms should be used. to ensure security; we would only go further to say that we should not force everyone to use the same multiple mechanisms. The essence of our proposa lis that what should be "built in" to the Internet is a mechanism to allow user-defined end-toend. security protocols.

Our approach allows the use of perirrreter defense, but also allows other approaches to be used if they are desired, It also allows for creating security appropriate to applications not currently envisioned. Finally, it permits multiple implementations of the same secu-

Letters to the Editor

Leiters (-300 word,.) discuss material pLI blished in Science in the p revious 3 mo nths 0 r issue, of general interest. They ca n be submitted I h rough the Web (www, Of by reg ular mail (1200 New York AVe., NW, Washingtoll, DC 20005, USA). Leiters are not acknowledged upon receipt, nor a re authors generally consulted before pu blicaticn. Whether published in fu II 01 in pari, leiters ale su bject to editinq for cia rily and space.

rity policy so that if there is a bug in one of the implementations, it doesn't open the door to all systems,

By all means, if you want a lock 011 your door, have one, Have three if that makes you feel. better, But your town iSI1't secure ifall the locks open to the same master key! Having a single implementation of a single security policy has precisely this v urn erabil ity,

WI LLI.I\M A. WULF* AND ANITA K. JONES Department of Computer Science. School of Engineering and Applied $cle,nce. University 01 Virginia. Ch~rlottesville, VA 229'04, USA.

'To whom correspondence should be addre~~ed. E-mail: wulf@'

Chilean Scientists Rally After Quake

SCIENTISTS, AS OTHERS IN CHilE, WERE HITHARO by the 27 February earthquake. Two days after the earthquake, I traveled to Santiago and visited the University of CM,e, (Universidad de Chile), one. of the two largest state. naiversities ill Chile. My friend and collaborator Maria Rosa Bono gave me, a tour of the sci-

ence campus, which had suffered substantial damage. Severa! walls were down, and people were using shovels and wheelbarrows to carry rubble out of buildings. There was neither water nor electricity, and an eerie atmosphere enveloped the campus.

As we walked by the animal facility, we saw water seeping below the door. We used a flashlight to find our way through ankle-deep Welte •. · to Iocate the source of the fioodi ng: a broken pipe with water pouring out. After some effort, we managed to stop the water.

The last major earthquake in Chile was ill 1985. Malia Rosa experienced that event firsthand. and learned li felong lessons about lab preservation as a resultIn her lab, all the solvents were stored in rigid plastic boxes on the floor and shelves were screwed to the walls. For 25 years, she has told her students to push the microscope against the wall at the end of the day. Generations of students probably felt this-was a useless precaution, but not today,

The scientists I spoke to in the immediate aftenuath of'this earthquake were universally calm and determined in the face of daunting setbacks, If there is anything positive to take mrvay from thi s earth qn.ake,it is renewed SCI ENCE VOL 327 19 MARCH 201 0




em temperatures ( I, 2). Thus, forests that are important to biodiversity now will not necessarily be so in the future (3). This complicates any assessment of the longer-term biodiversity value of carbon-rich forests at finer scales.

Funding allocations that simultaneously address protecti on 0 f biod iversity under current and future climates will ultimately be more effective and less costly (4), Analytical tools are available to identify areas important for future conservation (e.g: migration corridors and refuges) (4-6). Application of these too Is should be rapidly expanded across, the global tropics to inform within-country allocation of REDO funds. LU KE. P. SHOO Ccntre lor Irnpical. BiodiverS'ity and Climate (hang~, S,hoot of Marine and lropi cal Biology, lames Cook University 01 North 0 u een sla nd, 1 ownsville. 0 l D 4811, Au, tr alia, E-mail : lu~~"h~o@icu .edu. au


1, t-e Chen er ol; Ptoc. Norl, Acvd Sci', USA 1,06, 1479 (2009).

2. A. J. Pounds, M, P. L, Fogden, i. H_ Cilmpbell, No/ure 398,611(1999).

3. M. B. Ar~lijo, M. Cabeza, W. Thuiller, l. Hannah, P_ H.

Wil(j~h15, Global Change Bioi. 10, 1618 (2004).

4. L. H;mnah et ai; Ftomien Eeal. Environ. 5,131 (2007),

5. 5. L Phillips, P. William,. G, Midgl ... v, A. A,(h~t, Ecol.

App/.1B, izoe (2008).

6. 1)- KiIl~en, M. Douglas, T, C{)~)j9Iio, P. WI. Jm'g~nsen, 1- Mejia, 1- Biogeogr. 34, 1357 (2007).

appreciation of our courageous Chilean colleagues as they move forward to confront these challenges. GUilLAUME DUMENIL P ~ris-(a rdi ov a~cu ta r R e,ea.rch Cen ter, I n se rm U 970, 75737 Paris (ed~x 15, France. E-m~il: g,~dlloulTle.dum~nil@

the Bristol BRY watershed will be the straw that breaks the salmon '5 back.

BROCIi WRIGHT Conservation Science I nsti tu te, 5anto Cruz, CA 95061, USA. E.mail:


1. transportattoe R5eaf(h soard of the Nationill ACol,d-

emi es, • Ri ~~ 0 f vesse I at ti d~~ ts and spills i~ the A leo ~6ri lslands: DeSigning a comprehensive nsk assessment" (Spe cial Repo rt 293, 2008).

2. V. Mo,e~, SCience 326. U4Q (2009)

Salmon Swimming Against Multiple Threats

THE NEWS FOCUS STORY, "FISHING FOR GOLD IN the last frontier state" (S. Kean, 15 January, p. 263) highlights the importance of'habitat and the qualities of landscape, whether marine 01" terrestrial, to protecting diversity and abundance of animals, in this case the salmon of Bristol Bay ani! the adjacent watersheds.

The News story discusses mining but does not mention the many other threats to this ecosystem Bristol Bay is scheduled for oil and gas exploration in. tile near future; large tanker traffic passing through Unimak Pass and tile edge of Bristol Bay is increasing (1); and industrial fishing in the Bering Seawith high rates ofbycatch is common (.:?l We should be asking not Wh<~t mining will do to Bristol Bay salmon, but whether mining in

Planning for Biodiversity in Future Climates

0_ VENTER ET AL. ("HARNESSING (ARBON PAYments to protect biodiversity," Brevia, 4 December 2009, p, 1368) demonstrate that careful targeting of reduced emissions from deforestation and degradation (REDO) funds call double biodiversity benefits while IDCl1Iring only a small reduction (4 to 8%) in carbon benefits. However, they do notaddress Whetherthis win-win solution would still be tenable if objectives were extended to include protection of biodiversity under future climates.

Tropical species. are already shifting their distributions In. response to increasing mod-

Learn how current events are impacting your worl(.

Sciencelnsider. the new policy blog from the journal Science. is your source for breaking news and instant analysis from the nexus of politics and science.

Produced by an international team of science journalists. Sciencelnsider offers hard-hitting coverage on a range of issuesinci u ding climate chan ge, bi oterro risrn, researc h funding, and more.

Before research happens at the bench, science policy Is formulated in the halls of government. Make sure you understand how current events are impacting your work. Read Sciencelnsid e r today.


Breaking news and analysis from the W Drld 01 science pol icy




SHOO QUESTIONS WHETHER AREAS WE IDENTIFY as global priorities for reducing expected carbon emissions and species extinctions-such as Madagascar, Indonesia, and the Philippineswill still sustain high levels of biodiversity after a century of climate change.

We believe that present centers of tropical endemism and diversity are, broadly speaking, likely to remain important in the future, Such centers tend to occur in regions that have remained climatically stable over long periods, such as hyper-wet, cloudy areas in the Andes that have withstood Pleistocene climatic fluctuations (1). Moreover, temperature is predicted to shift at a global average velocity of 0.42 km per year, or 42 km this century (2), whereas our scheme focuses on developing countries that average more than 700,000 krn' in area. Hence, the Large scale of our analysis Tel ative to the pace of climate change makes it doubtfu J that priori ti es will shi ft much this century

Although we agree that climate-induced shi fts in speci es" ranges could theoretically be incorporated into our work and conservat] on planning more broadly, such efforts are

plagued by a dearth of data, methodological shortcomings, and uncertainty (3, 4). The choice of modeling approach, in particular, can strongly affect one's results. For instance, nine different bioclimatic models yielded wildly varying predictions of future range size fOI" a South American plant species, ranging from a loss of 92% to a gain of 322% (5). Hence, we stand by our priority areas for using REDD to protect both forest carbon stocks and biodiversity. Shoo raises an interesting point, but not one that is likely to be of practical relevance now, when forests are under siege and conservation planners must make immediate decisions.


'The Ecology (ell tre, U 11 iversi ty 0 f Q u ee ns la nd, Bri s ba ne, OLD 4072, Australia. 'School of Marine and Tropical Biology,. james Cook University, Cairns, OLD 4370, Australia.

'To whom correspondence should be addressed. E-mail: .. au


1. J. Fjeld sa, E. Lambi n, B. M~rt€~5, Ewgrophy 7. 2, 63 (1999). 2.. s R. lo~rie et al., Natur~ 462,1052 (2009).

3. c. F. Dormann, Ba5ic Appl. f(o/. 8, 387 (2007).

4. M. B.llraujo, C. Rahbek, Science 3l.l, 1396 (2006).

5. R. G. Pearson et a/.,) Biogeogr. 33, 1704 (2006).

LETIERS Is the Message from Athens Being Heard?

IN THEI'R 2009 POLICY FORUM (1), B. WALKER et al. caned for greaterinteraction among existing institutions to tackle global-scale challenges. In April 2009, leading scientists, politicians, economists, and academics, as well as representatives from European Union Member States. civil society. and business put forward the Message from Athens (2): eight key priorities in EU biodiversity policy. Key priority 7 emphasizes the need for integration of the EU nature directives into other policy areas such as agriculture and regional development. The EU directives are an excellent example of the kind of cooperation that Walker et at. were calling for. However, it appears that the message is not being heard by the public at large nor by some leaders.

Natura 2000 is the European Union's main policy instrument to address the loss of biodiversity (3). It is a network of nature protection areas established underthe 1992 Habitats Directive and 1979 Birds Directive, Most of the protected land is expected to continue to be privately owned, and ecologically and economically sustainable development is a goal, SCIENCE VOL327 19 MARCH 2010




In Western Europe, planned dredging or the Scheldt River estuary in the Netherlands for the benefit of Antwerp harbor 111 Belgium puts at risk surrounding foraging and breeding grounds for birds protected under EU J egislation, Under the Directives, loss of these Natura 2000 sites requires compensation measures, such as restoration of tidal mudfiats. This implies flooding of Dutch agricultural polder land ... Debate about the conflicting issues of bird protection and food production has begun to surface in the Dutch press [e.g .. , (4)]. The situation is also being seen as a conflict between a distant EU interest and local sacrifice (4). The resigning Dutch Prime Minister Jan-Peter Balkenende stated, in an attempt to weaken the EU nature directives (5), that "Natura 2000 ... overshoots the mark" and that public support f01" nature and biodiversity policy in the Netherlands (all EU ttl ember state) \V'J.S diminishing. Jose Manuel Barroso (President of the EU Commission) replied to the Dutch leader that a review of the directives would not be justified because "economic operators benefit from a stable and predictable legislative framework" and "the inevitable legal uncertainty that wonld be caused by a review would be. likely to slow

down the. development of existing plans and projects" (5).

The. EU nature directives are supranational and enforceable, working toward a global view 011 issues that transcend loca I priorities. Unfortunately, these efforts are 011Sunderstood, unappreciated, or feared, and the economic crisis offers further incentives to brush them aside. However, this directive tool is unique in an international context, as it oven-ides local interests when global issues are at stake. The ED nature directives strive to integrate conservation with human activity, wherever needed, and a confrontation should not be necessary, Although the strength of the directive tool is clear, the Message from Athens has apparently not been COlTIIDt1l1icated effectively, not even to a political leader in a member state, and tile European Union must think hard about how to bridge the divide.


!\Jrije Universiteit Brussel, Departernent of Biology, Plant Biology and Nature Management, Brussels, Belgium. 'EC joi nt R~sea rrh Centre, I nsti lute lor E nvi ronm en tan d S"U sta i nability, Rural. Water and Ecosystem Resources lspra, Italy.

'10 whom correspondence should be addressed. E"mail: rmef~el1@vub.~(. be


1. B. Walker et 0/., Science 325, 1345 (2009).

2. European Commission, "T~~ Mes;age from I\th~n5" (2009); http://ec.eumpa.eulenvironmentlnature/

bi od iller, i rylcon leren (eli nd~~3 n.h tm.

3. Eumpeao Commission, Natura 2000 Network; http1! e c, europa. eu/envi ronrnent/na t ur eina tura 2 OOOlin dex, en.htrn.

4. L. O. Fr5w, "Het ontpolderinqsmodel," NRC Hande/s· btad, 13 October 2009.

5. C]a nssen, M. van L i eshcut, • Sa lke nend e witd e EU· natuurWel "frwakk~n," Volkskmnt, 8]anuary 2010; wwI'I.vol kskra nt. n Vb in nen la n dla rti de 133 S 67 8. en:} Sa lken e nd e., witde_E U· no tu urweLa Izwa kken,


Reports:. "Dsletion 01 Atohl disrupts Sonic Hedgehog signal i og in th e develo pi n 9 ce reb ell u m and p rev en ts m edu lloblastoma" by A. Flora et at, (4 December 2009, p. 1424). Reference 3 should instead rite: M. C. Thompson et ai., J. ((in. 011[0/. 24,1924 (2006).

News Focus: "Sex and social structure" by E. Pennisi (23 October 2009, p. 518). Mary lane West·Eberhard should have been listed as affiliated with the Smithsonian 'Iropical Re,e.3r(h Institute in Panama.

Research Ar tides: "Macrover(ebr ate paleontology and the pliocene habitat oiArdipithec!l5 ramidus" byl D. White et al. (2 October 2009, p. 6 n The beginning of the description 01 panel (C) in the Fig. 1 caption on p. 88 should read, "(e) Total NISP. The NtSP value reflects all collected specirncns identified to taxon and excludes bu lk specimens. IIssO( la t ed de n ta L.. "




Sailing on an Ocean of Os and ts

James P. Collins

When the developmentof'theory outpfl ce~ data, scie.n tists often fi. 11 d th .. <I.t Hew 1 deas cannot be tested for lack oftools or technology. Researchers in genomics, astronomy, and many other active areas of science face a different challenge: Gathering data is so easy and quick that it exceeds am capacity to validate, analyze, visualize, store,

and curate the information. The Fourth Paradigm addresses this chaJlenge-and the opportunity it presents.

TOliY Hey, Stewart'Iansley, and Kristin Tolle (computer scienti sts at Mi crosoft Research) have grouped the essays into four sections:

Earth and Environment, Health and

Wellbeing, Scientific Infrastructure, and Scholarly Communication. Examples from health care and the environment dominate but are leavened with some cases from the physical and social sciences. The editors' thesis is that although empirical, analytical, and simulation methods have provided answers to many questions, a new scientific methodology driven by dataintensive problems is now ernergjng-e-the "fourth paradigm."

Some contributors discuss software and hardware advances that would help scientists cope with the data deluge. Most consider how research practices will be transformed by computational thinking-e-scientific instruments powered by computers in ways that make them a universal intellectual amplifier. Hardware and, especially, software will facilitate "cross-domain ventures that accelerate discovery, highlight new connections,

g and suggest unforeseen links that will speed t:; science forward." For example, while analyze

l2 ing the relationship between variables x and

~ y, background software will seamJ.essly link ,~ to diverse databases. A prompt may then sug~ gest looking at how variable z relates to x, to y, ~ or to x and y A word of catrtioIi.: in this work ~ flow, scientists may fall into the. trap of under- 3

t The reviewer is a! the School of Life S.tience~, Arfzo~a State

Ii:l University, Post Office Box 874501, Tempe, AZ 85287- 5 4501, USA. E-mail;

The Fourth Paradigm Data-l ntsnsiva Scientific Discovery

TOllY Hey, Stewart Tallsler, and I(ristin Tolle, .Eds ..

M ie resort Reseil rch, Redl11o~d, WA, Z009.

Z8'6 pp. Pa per, $46.

ISBN 978098Z544Z04. PDFat http:// r ese a re h. mi c ro soft

c om/en-u s/colla bo ratio nl tounhpa radigm/

standing fewer details of an analysis if they accept uncritically such prompts as offered.

Backing away from a paradigm shift in the sense of Thomas Kuhn, John Wilbanks characterizes cybermfrastrucmre advances as new tools cutting across the sciences. It will be up to historians and philosophers of science to wrestle with the questions: In what if any sense will computational thinking transform the process of scientific discovery? Will data-intensive science cause an incremental change in how science is done? Or will the changes be a true step function?

Many of the authors see a big new step in networking: researcher to researcher and al so lab notebooks to archived databases and published results. A network of investigators is, of course, uot novel-s-think of Darwin and his correspondents, But until recently, even scientists ina network usually conducted experiments and wrote papers alone or with only a few others. Now dozens to hundreds of researchers may participate in projects.

At its best, this environment will facilitate what we might call open source innovation, in which advances have a sociotechnical component. Individual labs will routinely reach beyond their walls, and organizations will strive to create and sustain collaborative, distributed networks of investigators. Buthnrdles will appear. The toughest changes ahead involve training, institutional organization, and the social context ofresearch. A radical restructuring of the culture of scientific work will be needed to integrate biological, physical, and social sciences and engineering; move across the sciencetechnology interface; foster systems thinking; support flexible and interdisciplinary approaches to problem-solving; integrate knowledge creation and knowledge use; and balance individual with group achievement, Fostering: innovation will need to become a core institutional value.

The editors' collective enthusiasm about data-intensive science leads 1.0 an occasional odd statement, such as "Scienoe is becoming increasingly dependent on data." More precisely, solving some modern science. prob-


lems requires very large data sets. A justifiable. excitement about large reference data collecti ens has caused some to ta lk about the end of theory because extensive databases will support "hypothesis-neutral research," Paul Ginsparg counters this argument nicely in his chapter on a data-centric wor.ld: "Science aims to produce fa.r more than a simpie mechanical prediction of correlations; instead, its goal is to employ those regularities extracted from data to construct a unified means of understanding them a priori" Data mining simply to predict trends confuses the goals of phenomenological modeling and theory development.

Although most authors foresee .<1 bright future, even the visionaries admit we are some distance from computer systems that seamlessly link large num bers of related but disjointinformation sources. The Fourth Paradigm also offers a vision including few caveats associated with ethics, privacy, or cybersecurity. No lessons are drawn from Aldous Huxley's Bra ve New /¥odd or George Orwell's 1984. Breathtaking advances in the sciences

must be placed in a larger societal context by drawing on the law, humanities, and arts.

The text has a: few rough patches. Some are tile inevitable oonsequence of melding 71 contributors and 36 chapters. Others are interesting, as the unevenness conveys a sense of excitement and creation. Even the term "fourth paradigm" is defined in several ways. The authors provide. a view limn the leading edge, and tlle front is often. ragged, inccmpleteand in constrtwtion. We. get a bottom-up view of change in progress. Ma.I:I.Y will enjoy this book: historians of science. interested in the dyuamies of change, philosophers wrestling with the nature of science, social scientists studying how disciplines evolve, and natural and physical scientists who want fresh ideas. SCIENCE VOL 327 19 MARCH 2010



771f! Fourth Paradigm. is dedicated to And reflects the vision of the late Jim Gray of Microsoft Research, who envisioned=a world of scholarly resources-text, databases, and any other associated materials-that were seamlessly navigable and interoperable." Gray loved sailing. Sailors, of coarse, guide a vessel by reacting to the nearest swell and wave. But the ocean also affords a chance to scan the horizon in anticipation of the future, to see what's ahead and imagine what's just mit of view. The individual essays-and The Fourth Paradigm as a whole-glve readers a glimpse of the horizon for 21 st-century research and, at their best. a peek at what lies beyond. It'sajourney well worth taking.



Programming to Forget

WilHam DuHon

The ~Iow~g cel1~·~.~it>: of the .. Inter" net IS leading to uutiatrves to archive, curate, and otherwise preserve digital collections. While experts and resources are focused on these problems of remembering, Viktor MayerSchonberger tells us that there is also a virtue in forgetting,

Delete begins with an anecdote about a student, Stacy, who is denied her te-aching certificate because a eel league discovered an old photo of Stacy wearing a pirate bat and drinking alcohol-D11C she had posted on a socialnetworking Web site. As

the author put i:t: "The Internet remembered what Stacy wanted to have forgotten." Similar stories of individua Is compromised by information stored on the Internet or related devices arc numerous. Mayer-Schonberger (a legal scholar at the National University of Singapore) diagnoses the problem. explains its growing importance. and. answers the question "What can be done?"

The book's central argument is that in the analog world of yesterday. forgetting was tile default position. It was somewhat. harder to remember than to forget, so unless we put effort into it, such as in taking notes or storing text, information disappeared. "Not any-

more." Tn tomorrow's digital world, the default will be remembering. The efficiency of remembering is gaining ground because of the lower costs efmemory devices and the accuracy of digital technologies, which can replicate content endless ly, creating the potential for a future of "perfeet remembering."

Taken to its logical COI]elusion, this capability could create a dystopiaa scenario of self-censorship that moves beyond contemporary C011- ceptions of a surveillance society. Building on Bentham's notion of the panopticon, digital memory is extending tiie-"mcchanism of panoptic control" into the past. However, Mayer-Schonberger argues, this problem call be addressed through a variety of legal and technical initiatives, such as creating a means for users to place anexpiration date on infermati 0 Jl they post.

The book offers a provocative counter to prevailing neologisms about information wanting to be shared, Om circumstances are far more complicated, with many not wanting

all information to be remem-

The memory of th e I nternet, (0 llective ly, serve r farms hou se hundreds of exabytes of information.


The Viltue of forgetting in the Digital Age

by ViktorMayerSchOnberger

Pril1ceton University Press, Prillceton, NJ, 2009.

251 pp. $24.95, £16.95.

ISBN 9780691138619.

The reviewer is at the Oxford Internet Institute, UniVersity of Oxford, 1 51. Giles', Oxford OXl 3J5, UK. E-mail: william. ciu(tol1@(lli,(lKac,uk

bered, In developing a clear line of reasoning behind his argument, Mayer-Schonberger draws evidence from multiple disciplines, bringing together considerations from the neurosciences, computation, and networking technology as well as from law, policy, and literature. It is rare but wonderful for an expert 01l digital technology to

glean from works of major literary figures with the same ease as he discusses shared memory devices and Vmmcvar Bush's "rnernex," His book also stands out in being truly international, anchored in European and Asian exampl es as firmly as North American legal and policy cases.

Mayer-Schonberger's focus on a sing!e iss lie-remembering and forgett:ing as enabled by digital technology-enables him to address some familiar subjects, such as the history of the communications revolution. in a fresh and engaging way. Moreover, his style is accessib [e and clearly targeted beyond his' academic peers to reach arraudience engaged by the issue. rather than the technology or the law. That said, readers will team about technology, law, and other fields as his narrative unfolds,

Most important, Mayer-Schonberger's focus illustrates a major turn ill debates about

the Internet. Since the dot-com bubble of the late J 990s, most research on the Internet has dealt primarily "vit,h its use and impact in the broadest sense. Wi.ll the technology become a. routine aspect 0 f everyday life and work? This book takes the Internet's role III society as a given and concentrates on the critical design features that make it easier for machines to remember than to forget. Refocusing on key design issues, as the author does, will enable social and policy research to contribute more to shaping the future Internet,

The author's dia.gnosis of the problem raises questions. Hasn't memory always been long-term for some people, such as those forced out of a community that will not forget a major transgression? In such cases; the Internet is transforming the geography of memory more than extending its longevity. A lso, if individuals can delete the) r past, will we face Orwellian issues over the rewriting ofhistory?

Mayer-Schonberger's discussion of potential remedies is less convincing than his exposition of the problem, He admits as much in suggesting that his solutions are less than perfect. For example, expiration dates wi!! be difficult to realize, given tile distributed nature of the Web. and might cause other problems, Since the book's publication, new applications have been released that enable a text message, for example. to vanish after being read or on a specified date. However, employing software to erase messages can create unwarranted suspicion, or we might lose information we later want to retrieve.

Even in am not completely convinced of the problem orsolutions, Delete is well placed

to accompl ish the author's aim: "to commence '"

~ a wide-ranging, open, and intense discussion 8.

about forgetting, and how we can ensure that 1 we'll remember its importancein om: digital ~ future," There, isno better SO-UI"ce for fostering ~ an informed debate on this issue, ~

'" u

10. 1 lQ6Isden(e.1187723




China, India, and the Environment

Kamaljit S. Bawa.1J.J Uan Pill Koh} lien Ming Lee,S,6 Jia;l1guo Uu,! P: S .. Ramakrishnan,! Douglas w. Yu,~,lU Va-ping Zhang,~ Peter H. Raven11*

Ch.ina,. m.l~ India,. tl.le emergi.l1g econorruc giants of the world, Will play It significant, perhaps a dominant, role in shaping the environmental outcomes forour planet in the 21st century Both conn tries are expected to maintain an 8 to 9% rate of economic growth over the next several years (1). Even when much of the world is experiencing a recession, China and India in 2009 are projected 10 achieve high Gross Domestic Product growth (China, 8.4%; India, 6.2%; world, ~2.2%) (1). Development in China and India is fueled not only by the natural capital within the countries, but also increasingly by imports of raw materials, particularly from southeast Asia. Both nations import ~9 million tons of crude palm oil annually+almost one-quarter of global production-crnostly from Malaysia and Indonesia (2). The degree to which China and India consume natural resources within their boundaries, and beyond, will largely determine future environmental, social, and economic outcomes. The two countries have been engaged in a border dispute that included a bitter, but brief, war in 1962. We propose that much more earnest cooperation between the world's two most POPUIOlIS countries will be vital for mitigating biodiversity loss, global warming, and deforestation.

Biodiversity, Water, Climate, and Forests Nowhere are these issues better exemplified than in the Himalayas, the 2400-km-long chain of mountain ranges stretching between the Indus and the Brahmaputra River valleys (see the picture, above). The unique biodiversity of this region faces multiple pressures, including those associated with the military presence of both countries along the international border (3), This situation is particularly damaging to fragile subalpine and alpine ecosystems, which recover slowly from disturbance. Furthermore, resources are so scarce in these cold, high-altitude areas that the armies often use rare species for fuel,

Both countries are already facing severe surface and groundwater shortages (4, 5).


Cooperation between China ;;lod India (an curtail bi cdive rsity (ass, mitlga tiO! climate rha nge, and reduce deforestation.

A Himalayan landscape-lin king environmental integ rity with peace and secmity.

Given the increased snow and glacier melt in the Himalaya-Hindu Kush region (6), the hydrology of major Asian rivers will be severely affected in the near future, The rush to harness hydroelectric power by building hundreds of dams on both sides of the earthquake-prone Himalayas will further accelerate changes in hydrology and the effects of climate change. For instance, India plans to increase hydropower projects in the Himalayas from 74 to 355 over the next 15 years, increasing the capacity from 15,208 to 126,588 Tv.fW (7); China is planning 750 pro} ects in Tibet alone (8). Further, China may also divert water from the Tsangpo-Brahmaputra, one of the largest rivers ill Asia, before it en te rs Ind j a (9). T f true, dispute over po liti cal boundaries could extend to water resources ..

Exploitation of Himalayan resources is likely to be accelerated as energy consumption rises more rapidly in Asia than anywhere else. Cbina and India, respectively, are already the world's top and the fifth-largest emitters of greenhouse gases. Snow melt from Himalayan glaciers constitutes a principal water resource

i 'University 01 Massachusetts, Boston, MA 02125, USA. 'Su>tainabilily Srience Program, Harvard University, Cambridge, MA '" 0213 8, USA. 'A,hoka Trust lor Research in Ecology and the Environment, Bangatore, 560024, India. '1IlStitut~ oj Terrestrial ~ Erosystems, ETH Ziirich, 8092 Zurich, Switzerland. 'Univer5ity 01 California, San Diego, l.alolla, CA 92093, USA. 'Yale Uni~ v€rsi'ty, N~w Hawn, a 06520, USA. 'Michigan State Uiliver~ily, East l(ll1sing, MI 48824, USA. 'Jawaharlal Nehru UniVersity, ~ Delhi, 110067, India. 'Kunming Institute of Zoology, Kunming, 650223 runnan, China. lOUl1iv~rsity of East Anglia, Norwich '" N R471], UK. "Missouri Botaniral Garden, 51. Louis, MO 63166, USA.


~ • Author for correspondence, E·maiL peter.raven@ri1obo!.org

during the summer months for at least half of the world's population (10). The synergistic effects of decreasing water resources, loss of biodiversity, increased pollution, and climate change may have negative social and economic consequences and, even worse, escalate conflicts within and between the two countries,

Beyond the Himalayas, both countries are enlarging their ecological footprint in Asia:

If current trends continue to 2020 (2), China and India may be importing 45.8 and 18.8%, respectively, of their roundwocd demand, or 64% of all currently produced roundwood in Asia. These imports will contribute to deferestaticn-driven greenhouse gas emissions and biodiversity loss in Asia.

Benefits and Barriers to Cooperation

CJ1!l1a and India have much to ream from each other. For instance, India's energy efficiency is higher than China's (11). However, China has much to offer to India in poverty reduction, health care, and large conservation programs, including the Natural Forest Conservation Program and the Grain-to-Green Program (1Z).

Cooperation between the two countries could help mitigate climate change, environmental damage, and biodiversity loss both regionally and globally. Both countries, but especially China, have experience with microhydropower projects. Local communities, especially in tlJe Himalayas, have a tremendous

www.sciencemag.or9 SCI ENCE VO L 327 19 MARCH 2010


knowledge of biodiversity, hydrology, and elimate change (13). Sharing this knowledge could enhance the development of appropriate approaches and technologies. Both countries have recently increased forest cover (China, 157.1 million ha in 1990 to L 97.1 million ha ill 2005; India, 1'13.9 million ha to 677 million ha) (2), Sharing expertise in afforestation and in the development of policy frameworks for sustainable use offorest resources with the rest of Asia could contribute to reduction in deforestation and to au Asia-wide effort to mitigate climate change through both the preserve- 60n and enhancement offorest carbon stocks. Cooperation between the two countries can also have a huge impact on trade in tiger body parts and on conservation ofthe last remaining and the largest population of all icon species.

So far. collaboration aud cnordination between the two countries in dealing with environmental challenges has been limited, although there have been several signed agreements since- 1988 (14). III I 993, a S inoIndia collaboration agreement.on the environment was signed (l5). In the- PaSt few years, there have been research coLlabora.tions on terrestrial ecosystems, atmosphere-oceanbiosphere interactions. and related modeling (J 4) but the level of overall scientific collaboration has> been lower than expected (16). In 2009, an international symposium on biodiversity and environmental changes in the Himalayan- region was held in Chinaand was attended by Chinese and Indian scientists (17). More recently; India and China have signed an agreement to cornonitor glaciers (l8) and to col 111 borate in the 111'Ca5 of energy, agriculture, and a fforestation ( I P),

Despite the presence of these frameworks for cooperation, majorbarriers to collaboration between China and India persist. Linguistic and cultural differences perhaps tend to keep Indian and Chinese academics apart even in international meetings. Unresolved border disputes also deter cooperation. Furthcnncre, the severity of environmental cha llenges, and the opportunitiesfor and benefits of coordinated and collaborative actions remain underappreciated.

Steps Toward Deepeni n9 Cooperation

First, both countries need to overcome cultural and linguistic barriers. Although China: has been actively promoting Chinese language and culture internationally in recent years. only 2 out of 523 Confucius Institutes that have been established worldwide to support local teaching of Chinese ate in India: (.10), Similarly, India needs to promote. Chinese stlldies in India arid Indian studies in China.

Second, with sparse or no human populations, some of the 'dlspilted areas or fhe areas

where armies have been amassed, which are alpine zones, arc ideal fer COil version to transboundary protected areas or peace parks. similar to the proposed K-2-Siachen Peace Park between India and Pakistan (21). The creation of'peace parks=for which the United Nations Educational, Scientific, and Cultural Organization (UNESCO) could develop guidelines=endjointly run research stations could not only protect biodiversity, provide key ecosystem services, help mitigate climate change, and foster scientific collaboration, but also promote cooperative conservation actions and peaceful resolutiou to the border dispute (21).

Third, 2009 marked the i su, anniversary of agreements. signed between the Chinese and Indian Academies of Sciences. The academies could' exercise srrcng leadership in highlighting I):l1\'i1'O[ll11eI11<11 issues, initiating joint actions, ,and fostering scientific exchanges, The Indian academies are gearing up to be mom acrive in policy-making and could accord high priority to policy-oriented initiatives. Exchange and joint supervision of graduate and postdoctoral students could build on ex isting programs, such as those run by the Academy of Sciences for the Developing World (J2). The governments could also establish a joint grants program for Himalayan research, as China and India have taken preliminary steps to establish a framework for scientific cooperation (23). The ecological societies of China, Japan, and South Korea have established formal agreements for cooperation and have held regular joint meetings. 'Such meetings could include India and place more emphasis on the environment,

Fourth, the United Nations and other organizations could create rneCh~1l1isj'J1S for regional cooperation and governance of natura! resources, Considering the number of heritage. sites in the Himalayas, UNESCO and the U.N. Environment Programme, in particular, could be key in bringing the two countries together and facilitating such cooperation. Moreover, international foundations and nongovernmental organizations that are engaged in environmental issues on both sides 0 f the Himalayas could develop transnational programs.

Fifth, the existing political and economic forums, such as the East Asia Summit and the Association of SoutheastAsian Nations with China, Japan, and South Korea, which prirnarily foster trade, could pay more. attention to the environment, which sustains most economic endeavors.

Finally, the United States, the Eu!'opeml 'Union, and other developed ceuntries could play pivotal roles in facilitating and encour-


aging multinational talks on climate cbange and related environrnenfal problems that transcend political boundaries: The United States has good relations with China and India, but is engaged in separate discussions with each country over economic, energy, and. environmental issues. Ultimately, as China and India begin to build more confidence and consensus from within, they will settle their differences bilaterally, particularly when environmental security starts to override-cencem for political boundaries,

Refefeli~es and Notes

1. World Ban~, Global feonQmi( Pf(J5pea~ lOll1: Crisis, nnonce; and Growth (World Baok, W"~hiOglOO, DC, 2010); 1'fWW.WO tid ba nk. org/gep2 01 O.

l. FAa, FAO$TAT Online Statistiral ServifE (~ood and Agriculture Organization of thauriitod No tlens, Rome:

Italy, 2009).

'l. Con.~ri"tiDn I~!em"tiona~ WW\V.~iQd[."r,ilyhotspDts. o rgl~plHots po !SIn i malaya/.

4, ]. Briscoe, R. P. S. Malik, India's Water EconomyBtrk; ng far a TUri)I,1/enl Future (Oxford UniV Pre! s, New York, 2006).

5, T. M. ]o~o son, F, U u, R. Newf armer, Clear Water, 8/ Uf!

Ski6S: China's Envitol1ment in the New (enrmy (World Ball k, Wa ,hi ~g ton, DC, 1997).

0. J. Xu et al., Cllnserv. Blol.Z3, 5 ZO (2 ii 09).

7. J.'W. van Gelder, C. Scheira, 11. Kroes. New trends in the Fif]anrifr.g 91 DalJ]s-a Re,earch Pap~r Prepa,ed!gf /nternaitona'i Rivers, BankTmcK, and WWf Germany (f>rofundo, C<l5lriculH, Netherlands, .20081.

8. M. Moore, reiegraph. 14 October 2008; www.ieL~graph. (o,U kin eW'!\'I<!fld neW51~5i~m bi'trl19 3 7 90/C~j na, pia n~damsacross-Tibeth Iml.

9, K. Pomsranz. New Lstt Rev. 58.5 (2009).

10. T. P. Ba me It. J. C. Ad~ n I > D. P. Le'\!~ i\n'Iaj~r c N ature4·~ 8, 30}~ZOOS).

n, I. G. Liu, J. Di~J110~d, NaMe'43~, 1179 (2005)

12. J. G. Lfu, S. U, Z. Ouy~ng, C. Tam,.x. Chen, rroc. Nail.

A(Qd. Sri. U. S, 11. 105, 941'7 (2008)_

13, K. S. Bawa, G. Joseph, S. Setty, Agrl!;, Eco.ysr Envirol!. 121, 237 {2.007).

14_ [).l\broL p_ Rup~~ Sci Dev, Netc 14 May 2008; w\VW.sddev. 0 et,

15, ]ian gSu En~ iron rn en to I Pro leeti on I n dum)' N e twoi~, ~pa9a.miili' www.~pa98.mml\>p?id~ 1039.

10. E. Hand, NO/UfE 46J. 282 (lOW).

17, Meeting, (nina.!uantaQhuil 1300S0.htrnl.

lB. Huanqju.corn, hllp1Idjina.~ua~qiu.[Omll,y"s_Dr'l_[lijr'lal en\fironment·te(hnology/200~·o!l!53 ':>769 .hlml.

19. "India, Chin~ [nl:,p~(t to light climate change together," Etottomit Tin)e" 21 O(\Q!)~( 2fi09; nt!p;lleclmon'lic!imes. tndtatimss.rom.

2Q., 'Confucius In,lllUtes," H~ob"n; htlp:l/ellglish-.hanban. o tglmy. ph P

21. S, H. Ali, Eil., Peace Paf~'" [onserl'atr'on and (onflict Resolution (Mil Pre». Cambridqe, MA, 1D07).

22. Th e A til a emy of S,jen ce for the Dev~lopl n 9 World, https;lltWiis.i[lp.itt.

n. "Th e Nab onal Naturol 5deoce FOUll dation of Chi na a nd the IndioO Ministry of Science and Technology signed il Memorandum oj Underlt<loding O~ CoopenNon;" 2 My 2003; wWlv,mi, gO" .CnIPDrt:I[OJI nfdModu le_ 407/1 osn. 111m.

24, Support~d in part by NS, (u.s, and].l,), NatfoMI Mronaulico; il.nd Space Admini5tration (l.U. the Blue MaDn Fund (K.S.B:), W. [etz (1.1\\.1.-). ETH Fel[ow)!lip ~L.P.KJ. 973 Program ~Z(jQ7C BU16GG, I'>p.z.). National Natural Scieoce Found~tion of (hin;1 (Y,-p,Z.), and ~l'IiS5 National Sdence Found.Jllon\l.P.K.I·]' M. Cl1u((h, 6. Billa ch "ad e r. and t hre e a no,nym OU! mvi@wers ntlered con5tructlv€ comments.

www" SCIENCE VOL327 1'9 MARCH 2010



Setting the Trap for Re'active Resonances

Stuart C.,Althorpe

chemical reactio,n occurs when. molecules collide, resultmg in the 'earrangement of the atomic nuclei from reactant. to product configurations. The rearrangement dynamics takes place on the blindingly fast rime scale of molecular vibrations (10-]4 to 10-13 s). Sometimes, however, the nuclei can become temporarily trapped during a reactive collision, resulting ill a metastable quantum state that typically lasts for to I~ to 10' 11 S before it decays into reactants or products, These short-lived states, ca lled reactive resonances, are extremely difflcult to observe. On page 1 50 l 0 f this issue, Dong et al. (1) report the detection of three individual reactive resonances in tile F + HD --} HF + D reaction,

Reactive resonances are an example of a more general :type of quantum phenomenon called a scattering resonance U), A typical signature of such a resonance is a sharp peak as a function of collision energy in the scattering cross section (the flux of scattered particles divided by the flux pcr unit area of the incoming particles, and thus a measure of the fraction of particles scattered by the colIision). Such resonances have, for example, been observed in the scattering of electrons from atoms and molecules, the scattering of atomic nuclei, and the scattering of subatomic patti cles, where scattering resonances indicate the formati on of unstable particles.

In contrast to these other types of scattering resonance, reactive resonances are elusive, with only a few definite sightings reported to date (3-6). Reactive resonances 11 re not present in a ll reactions; when they do occur, their contribution to the cross section is often obscured by a much bigger contribution from products formed by direct reaction paths thatdo not invol ve reactive resonances. Another difficulty is that the temporarily trapped atomic nucleican rotate, giving rise to a progression of rotational quantum states, each ofwhich is an individual reactive resonance. Typically, many such resonances- contribute to a scattering cross section, where tb.ey tend to average out. (7).

It is thus rernarkabl e th~tt Dong el a I, were abl e to d irec 11 y obse I've '111 ree i rid iv i dual

Department of Chemistry, Unjv~L,ity 01 Cambridge, Cambridge CB2 lEW, UK, E-mail: scal0@cam.acuk


A combined quantum mechanical catculation and molecular beam study has succeeded in detecting elusive metastable quantum states.

on the product side of the barrier and SlITvive for about 300 fs t 1 fs = 10-15 s). They are beld together because most ofthe energy is trapped in three quanta of vibrational exciration in the PH bond. The leakage of this energy into the other degrees of freedom causes the resonance to decay int.o prodnets or back to reactants. The total angular momentum of the three atomic nuclei. is conserved during the entire reactive collision, allowing the same quantum number J to be associated with overall rotation of the three nuclei ar aU stages of reaction (see- the figm-e). The resonances observed by Dong et at. correspond to J = 12, 13, and 14. Resonances corresponding to other- values of J also contribute to thl} cross sections, and further studies mayresolve some of these.

The assignment of these resonances required close collaboration between experimental and theoretical groups. The resonances appeared as peaks in part of the cross section as a function of collision energy, However, similar progressions of peaks observed in the cross sections of other reactions (7-9) were caused by an entirely different sort of process; namely the opening of thresholds on fop of the reaction barrier, To s how that the F + HD peaks are rea ctive resonances, Dong 13'1 al. solved the Schrodinger equation describing the motion of the atomic nuclei during the rcactiou, which showed tliat the wave functions responsible for the peaks correspond to reactive resonances,

Theoretical predictions of this type must be interpreted with caution, because they rely on potential energy surfaces (describing the forces between the nuclei) that are obtained by solving the electronic Schrodinger equation, to reproduce features in the cross sections, these surfaces must be highly accurate (to within 0,1 kcal moll). This makes them computationally very expensive to calculate, even for "simple" reactions such as F + HD. It is thus essential to compare theory with experiment.

Dong et al.'s theoretical predictions match the experimental data almost perfectly [see Fig. 1 iii (1)]. The anthers used a version of the erossed-beam technique that was origi- 118 lly used to study the F + H2 reaction (I fJ). Over the years, th is and rr;:lated (' 1 I) 11101 ec-



300 Is


Fnrmatlon and decay of individualreactive resonances. The D H rnoletu IE (b lue and black) 11 nd F atom (red) form a metastable DHF complex by temporarily trapping the e flergy, v lin three vi brationa I qua nta of ~eitati on of the H F bond. The floW of one qua ntl! m of energy 0 u t of the bond causes the comp lex to break Up into products. The overa Il ,;In 9 ular momentum of the three atoms i .. quantized and gives ri se to a progress] on of rotati ona I qua ntu m states, labeled by}; earn rotati onal state of the camplex is an i ndivtdua L reactive resona flee.

reactive resonances in the F + HD reaction, corresponding to different rotatioualquantum states of the temporarily trapped FHD complex. Skodje et al. previously identified and. interpreted this particular set of resonances after observing the averaged contribution from a sequence of the resonances in the cross sections (4). Dong el al. are, however, the first to pick out individua I reso[lances from this series,

These resonances are formed by conisions at stIch low energies that the nuclei can traverse the reaction barrier only by qrla.ntum tunneling. The. resonances are formed

ular beam techniques have been steadily refined. Dong et al.'s experiment W!IS fully "state-to-state": The reactants were cooled to prepare the HD in its ground rovibrational SU11c, and the cross sections were measured for the HF products in specific rovibrational quantum states as a simultaneous function of collision energy and scattering angle (of tile HF products with respect to the F + ED approach vector).

The theoretical calculations used. a potential surface obtained at. a very high level of theory, which correctly describes electron correlation and spin-orbit coupling (12, 13). Earlier studies on this system used a potential surface (14) that was able to reproduce the transition-state spectrum of

PHD (15) and to predict qualitatively the features in the cross section produced by the rotationa liy averaged resonances (4). The surface used by Dong et al. describes much better the interactions that hold together the nuclei within the resonances, This is why the individual resonance peaks obtained from it agree so closely with the experimental data.

The study by Dong ei al. is a major benchmark in the understanding of chemica! reactions from tbe point of view of first". principles quantum mechanics. The ability to detect and perhaps target individual reactive resonances may also allow chemists to better control chemical reactions, especially at very low temperatures.



1 W_ Dong er al., Science lZ7, 1501 (20101,

2. R. G. Newlon, Scattering meory aj Waves and Particles (Sp ri ng~r. N~\Y Yoll, 1982)

3. L M. Wa(ler, T. N, Ki 15 opo ulcs, D, M. N~u m a rk, ). phys,

Chem. 9Q, 2240 (19.90).

4_ R. T. S~odj" ~I or, Phys_ R~V- lett 85, 1206 (2000)_ 5. K. Liu e/ ol., P.hys(hem(amm 5, n (2002).

Ii M_ Qiu er 01., Science ~ 11. 1440 (2006,"

7 W H MilLer, J. Z. H. Zhang. J. Phys. (hem. 95, 12 (199 I)- 8. S. c..l\hllorpe et al., Nature 416, 67 (2002).

9_ O. Dai et 01.. sctence 300,1730 (2003).

10_ D_ M. Neum.f~ et oi., J- (hem. pnys. 82. 3045 (1985). 11_ II, I\oszi now,ki ~t ai,,). Ch~m _ Phys. 125, 1~ 350'3 (2006).

12. 2. !len d QI., {lroc. NQtl. Aead. 5d., 105, lzMz (20Dll)_

13. B. Fu. X. Xu. D. H. Zhang,]. Chem. Pllys. 129,011103 (ZOO!!).

1ll. K. StOlrk, 1:1 •• ). Werner,)~ (hem. phys. 104, 6S 15 (1996)_ 15. D. E:. ManolopouloHt 111., 5tiem;e 261. 1852 (1993).



Phenology Under Global Warming

Christian Ki:irnerand David Basler

PhenOlOgiC. al events such a .. s bud bUrst.. flowering, and senescence have received increased interest in the light of global warming (1-3), Spring events at temperate latitudes have advanced by 2.5 days per decade since 1971 (4). As global warming progresses, how w-ill it affect the arrival of spring and the length ofthe growing season?

In humid extratropical areas, the three most important factors controlling phenology ill dominant forest tree species arc the degree of winter chilling, photoperiod (d!ly length relative to night length). and temperature (5-7) (sec the figure), Becausethe seasonal course of temperattire va r ies strongly from year to year. sensitivity to photoperiod protects

plants from the potentially fatal consequences of simply tracki Ilg tempera-

tures at the "WI'Ol1g" time of'the year. Photoperiod controls the induction (formation ofwintet buds, leaf abscis-

sion rneristems, and freezing resistance) (8-10) and release from dorrnancy, the onset cf'growth, and reprodnctive events, including synchronous flowering (11. l1). 'Iernperature plays a modulating role and triggers the visible progress of phenology, such as leaf coloration, inmany species.

I nstitute of Botany, University of Basel, 4056 Basel, Swillerland, E·mai(: (h.koern~r@\lnjbas,dl, david,

Because the photoperiod is equally long in autumn find spring, dormancy release in spring requires the information that winter has passed, obtained from the dose of low temperatnres experienced by the plant. When this chilling requirement is fulfilled, plants become receptive to photoperiod signals, Once a critical photoperiod has passed, actual bud break isa matter of concurrent temperature. A lack of 51! fficient chill ing ill mild win-

• Chilling

• Photoperiod

• Iemperatu re

• chilli ng

• Iemperatu re

• Temperatu re

In most temperate tre.!' speries, phenological events such as flowering and autumnal cessation of growtn a (I' not pri mil rily controlled by temperature.

ters delays bud break U 3) but may be partially replaced by long photoperiods and/or velY high temperatures (14).

Not all tree species are sensitive to photoperiod, but the long-lived, late successional species that become dominant in mature forests commonly are. The genetic controls of plant development by photoperiod even remain in action when these temperate tree species are transplanted to subtropical parks, where

bud break in hackberry (Celtis), beech (Fagus), and oak (Quercus) species was never found to

occur before early March, despite ex ceptionally high temperatures ill this exotic environment (15). Tt is thus amisconception to I inearly extrapolate a few days advance of leafing during warm years into a proportional lengthening of the growing season in climate warming scenarios (16. J 7).

Shorter-lived, early successional species adopt a more risky life strategy (6). Many

phenological observations in the literature come from such pioneer species as hazel, poplars, or birch, ,),fliich are opportunisti c (photopeti od - i usens i~ tive in spring). Other opporturustic species include weeds, as well as ornamental plants from warmer climates. SCIENCE VOL 327 19 MARCH 2010




Not just temperature. Spring development in many orrra mental plants from warm regions, surh as lilac (Syringa), is primarily controlled by temperature, whereas early successione I species native to tern perate ls titud es, SUch as hornbeam (Carpinus). only become temperature-sensitive once their chilling demand has been fulfi lied. late successienal taxa, such as beech (Fagus), are photoperiod controlled. with temperature only exerting a limited modUlating effect onte the critical day length has passed. This mechanism prevents such taxa from sprouti ng at the "wrong " Ii me.



For instance, the farnou s phenol ogi eal time series for horse chestnut in the streets of Geneva (18), showing clear advances in leafing, is for an exotic species from a subMediterranean setting, Another prominent time series shows early flowering of domestic cherty trees (/8), which exhibit adaptive traits frorn central Asia, from where the cultivars originate. 1n these contiueatal regions, the advent of spring is rather invariable, presumably due to the great distance from the sea, and phenological tracking oftemperature bears no risk, In fact, trees in these regions should be more likely to keep tracking climatic warming than those in climates with more unpredictable weather systems, an interesting question to be explored in future work. Many ornamental plants in temperate gardens arc photoperiod-insensitive, arid their spring phenology tracks temperature with only very minor chilling requirements, <Ls exemplified by lilac (S)lI'ir1ga} (19),

Phenology in late successional species will thus not continue to trac k climatic warming (the. leugthening of the potential growing season) but will increasingly become 0011-

strained by internal controls .. as the photoperiod threshold (set by genes) is approached. FOI most extra tropical trees, seasons will not become substantially longer until new genotypes emerge, which will take H few tree generatiens (a few hundred years) (20).

Opperrunistie taxa may profit from a warmer climate and may thus gain a competitive advantage over photoperiod-sensitive taxa, Rapid climatic warming may also wive C,t1lTCl1t tree genotypes into a disparity between their insura nee against "misleading" (too early in the season) warm temperatures and concur - rent temperature-sensitive soil processes snell as mineralization. Ecosystem nutrient losses are apotential consequence of trees getting ontof'phase with the climate system. Climatic warming should thus 'Dot be seenas a seJf-eyident cause for more tree growth.

Refere 11 CI',S all d No res

1. E, E. Cleland, I. C~uine, A, Men1~1, H.A. IV\OOll~Y, M. D.

Schwdrtl .• rrend5 ecol. evol. zz, 357 {2 0071.

2, X MDrin et ai" Glob. [hong~ BioI. 1,5, 961 (2()09l.

3. V. f. Khandu~, C. M. Sharma. S. P. Singh, Environmentalist ~B. 143 (2Q08).

4, A, Mel"l1~lef al., Glob. Change Bioi. 12, 1969 (2~06). ~, I. C~uine, I'. (GUr, Neloi Phyl:vL 14~, 339 (1999).

6 (. KO mer, in Plant Growth and Climate (;honge. J. I. L.

Morison, M. D. Morenoft. Eds, (Bliid::weli, Oxford, 20(7). pp.48-69.

7, R. K. M. Hay, New Phytol. 116, 233 (1990).

S, C. u O. junttila, A. Ern.l';e~,. P. Heino, E.1. Pillva, Phys· ;01. Plant. 117/ .206 (2,003),

9. R. F. Wareing,Anml.lleV. P/QUtPhysiol. I'/O{1tll1o/. Bioi. 7,191 (1956).

10. j. K5kili!lo, G., BergqUist, P. Giltd~,trorn, 5. Jansson, II/ant rhy;iol; 13 9, 163 S (200 S).

11. f. Keller, C Komer, Am. AMoret. Alp. Res. 35, 361 (.2003).

12. s. D. lilck-on. /'Jew P/lyfoI. 181, 517 (2009).

B. M. II. Murray, M, G, R. Connell, R. l. Smilh,.j. Appl. tcol. ~6, 693 (1989).

14. Q. M. H.,jde, Physiol. Plan~ 8S, 531(1993)

15. R_ Bormen. K. il.ob~rlSOn, M, D. 5mWilrll. G. WillianlsU nera, /'nr }. BiQmeteoroi. SO. 5 7 (20051..

16, J. Pen uelas, 1. RUliihil user, I. ,ilella, S£ienre 324, 887


17. l. Moser [>'J or, liee Physiol. 30, 22.5 (2010).

18. C Deli!;!, B. C1Qt, Int. 1. B;ometeorol, ~ 5, J 03 1200 u, ~9, W larcher, Sitzber~Anftig~rMatli.-Nat. k{ Jg. 2006

2lZ, 3 (2007).

20. This e,timate is billed 0" what is known lor weeds, Which nee d a bDU t five. ge ner at io ns to BVDl ve new 10 Ii tudeIp~cifk photoperiod genotype:; (11).

ai. o, langl~t, TaxOll 20. 653 (1971).

22. funded by VelUlI"Fllundationand Nationill0~llter or Compet"I(" in R~e~frn (N(CR) Clin'lil!!'. of the 5w;" 5den(f' f'oumlatlon.

1O.1126/Icienre.113M H


Controlling Turbulence

BaverlllY J. McKelln

Pipes feature strongly in the infrastructure of everyday life. from domestic water pipes to oil and natural gas conduits. A primary consequence of the onset of ntrbnlenee in the fluid flowing through the pipes is the dramatically increased power required to pump stuff at the same rate, Th us, the incentives to understand and control the transition process are strong, However, more than 100 years. after Osborne Reynolds's seminal experiments 011 the transition of flow through 1] pipe from a laminar (s111ooth) to a turbulent state, the exact physical mechanism that drives this phenomenon still vexes the fluid mechanics community. On page 1491 of this issue, Hofel al. (1) describe a mechanism that feeds energy into a turbulent flow system, allowing the onset 0 f the transition to be manipulated and even the suppression of the turbulence.

Reynolds's 1883 paper (see tile figure, left panel) (2) initiated an enduring framework with which to understand the flow of fluid,

GradUalE' Aeruspace taboratori es, California I nstitute of lechno(ogy, 1200 East California Boulevard, Pasadena, CA 9112. 5, USA. ~-milil:

and particularly the range of conditions under which the transition from an ordered laminar state to 11 three-dimensional turbulent one will occur, The ratio of inertial to viscous forces is typically expressed in terms of the Reynolds number, ReD =- u7)/v in pipe flow (where U is the average velocity ill the pipe Ci"{)SS section, D is the pipe diameter; and v is the kinematic. vi scosity of the f uid). If the ReD is identica I between two idealized, incompressible flows in similar geometries, then similar flow behaviorwill occur, Thus, an experiment ill which air compressed to a pressure of 200 atmospheres flows through a 12-cm-diameter pipe, or liquid helium through a 0.47 -em pipe, can accurately mimic flow through a transcontinental natural gas pipeline of diameter larger than one meter (seethe figure, light panel) (3).

For all Reynolds numbers, laminar pipe flow is linearly stable .. Yet a transition to tnrbulence still occurs at Reynolds numbers 011 the order of a' few thousands, with the flow displaying the characteristic, spatially intarIII itten t structure 0 f turbulent "pu ffs" followed by extents of laminar flow (4) (see also figure 4 in Hof et 0/.), A lthough considerable

I njectlng a fluid jet into a pipe- at an optimized location can control the development of turbulent flow.

progress has been made in using unstable, traveling wave solutions of the Navier-Stokes equations, which govern fluid flow to define the state-space boundary between laminar and turbulent flow (5), these concepts still do not expla in the origin of the observed puffs.

Hof et al. introduce centro lied disturbances using small jets at an upstream location in a frilly developed laminar pipe flow experiment, which permits the study of "designer puffs:" The physics of their observations is cl~g"1llt. In the reference frame of the disturbance, the transition is a, local phenomenon-the streamwise velocity gradients associated with a change from a laminar to a turbulent velocity distribn- 60u lead to a local inflectional instability in the radial profile.of streamwise velocity. This instability is capable of driving turbulent dynamics in the puff When a second identical disturbance is introduced at an optimal point upstream, the leading edge ofthe second puff can reduce the local Velocity graciis!itat the trailing edge oftlle first puff s uppressin g the in fl ecti ona I instability that feeds the turbulence.

The centro I of turbulence has Tong been a "hoTy grail" of fluid mechanics; and the field




Pipe flows. From the 19th·centllry laboratory to the pressf ng ene rgy questi nns of today. (lett) Osborne Reynolds's experimental. demonstration of the transition to tu rbu lence in pi pe flows a nd (right) the Yarnal-Europe natural gas pipeline.

is highly active, The term "flow control" can be something of a misnomer, in that the complexity of fluid systems and the constraints on sensing and actuation me often too difficult for successful application of control theory, However, "flow manipulation" is possible with some active and passive techniques. Hof et al. exploit their understanding of a particular transition mechanism to identify an elegantly simple, physically realizable active approach. Most important, this control

is obtained with a net gain: The reduction in pumping power associated with the elimination of'turbulence outweighs the energy input 'required to generate the control disturbances, an essential element of practical flow control.

For large-scale pipelines, tile impact of preventing the transition to turbulence could be expressed in terms of more than a 100- fold decrease in the friction drag acting on the fluid for the same flow rate, a gain that would be directly reflected in the reduction in required pumping power. The economic impact and energy implications of controlling the transition to turbulence are appal'ent, However, the continued effectiveness of the control strategy for turbulent Reynolds numbers suggests that this approach could

also give insight into fundamental physics of fully deve loped turbulence, the multiscale nature of which provides an equally challenging problem to fluid mechanicians. There are important differences between pipe flow and other internal and external flows, such as the flow over a wing, but there is perhaps potential to develop the approach to address a broader class of flows.


1. B. Kof el ol., Science 327, 1491 (2010).

2. O. Reynold" Phi/as. Trans. R. Soc. Lond. 174,935 (1883).

3. B, J. McKeon et 0/.,1- Fluid Mech. 511., 41 (200a).

4. I.J, Wygnan)ki, F. H, Charnpaqne, 1- FlUid Mech_ 59, 281 (1973).

5. B. E(klmdt et 01., Annu Rev. FlUid Mech. 39,447 (2007L



AMPA Receptors-Another Twist?

Mark Farrant and Stuart G. Cull-C.andy

Neurons in the brain can alter their responsiveness to signals from other neurons, a flexibility that contributes to the richness of neuronal communication and underlies the fundamental processes of information transfer, teaming, and memory. The most important receptive elements that allow neurons to "listen" to one another are

g ligand-gated transmembrane ion channels, iii and those that enable fast excitatorv commuI<.J

8 nication belong to the AMPA receptor sub-


OJ Department of Nell rnscienre, Physiology. and Pharma-

~ cology, University Colleqe london, Gower Street, London o WClE 661, UK. E"mail:; s.(ull"CiIl1dy@ 5

type. When the neurctransmitter glutamate is released from a. presynaptic neuron, it activates postsynaptic Al\t1PA receptors, a llowing cations to enter, causing depolarization that triggers an action potential in the postsynaptic neuron. On page 1518 of this issue, von Engelhardt et al. (1) use a proteomic approach to identify an auxiliary protein that regulates Al\l1PA receptor activity.

AMPA receptors are homo- or heterotetramers assembled from subunits GluAI t.o 4. AMPA receptor-mediated excitation is regulated by numerous processes that influence biophysical properties of the receptors (i ncluding affinity for glutamate, ioui c selec-

A protei n expressed in brain controls the pia sticity of synaptic trans missio n by regUlating the- properties of a neurotr ansmitter recepto r.

tivity, conductance, and gating) or their location and stability within the cell membrane. These changes arise through developmental or activity-driven alteration in A.i\1PAreceptor subunit composition, and by posttranscriptional or posttranslational modifications such as alternative RNA splicing, RNA editing, and protein phosphorylation, glycosylation, or palmitoylation. The discovery that transmembrane AMP A receptor regulatory proteins (TARPs; y-2, -3, -4, -5, -7, and ~8) act as auxiliary subunits t1.1<lt affect receptor trafficking and function (2-6) revealed even greater capacity for variation in receptor regulation. TARP-hke moleculesalso exist SCIENCE VOL 327 19 MARCH 2010



Three dassesof AMPA receptor-interacting proteins. Vo n Enge lha rdt et al. identified TARPs (1-2 and -8) and CKAMP44, but not CNIH-.2 Dr -3, in AMPA receptor complexes (contai liing GluM receptor SU bu nits). I n the schematic, no particu lar stokhiom€try of association with AMPA receptor, IS implied.

in invertebrates, indicating evolutionarily conserved roles, for such proteins (7), Recently, cornichon proteins (CN1R~2 and -3) were identified as another distinct class of AMPA receptor regulatory proteins (8).

Using mass spectrometry to analyze Glua l-eoatainjag AMPA receptor complexes isolated fwm mouse forebrain, von Engelhardt et al. identified a protein they call cystine-knot AMPA receptor modulating protein (CKAMP, with a predicted mass of 44 kD; also known as mouse shisa homolog 9) (see the figllre), CKAMP44 has a single transmembrane domainand au LJ1trap·ell uJ ar PDZ motif that could anchor the molecule at the membrane. Intriguingly, like certain other proteins or polypeptideneurotoxins that interact with ion channels, the extracellular domain of CKAMP44 has a 'cysteine-rich region. This is similar to the cystine-knot motifs found in cone snail toxins fhat affect certain voltage-gated channels (9) and in a eODOtoxin that modifies AMPA receptor function (J 0'). Related motifs are present in the snake toxin o-bungarotoxin and the Ly-6 protein lynx I, both of which interact with nicotin.ic acetylcholine receptors (1 J). A Ly-6 protein a lso modifies Shaker-type K' channels in the fly Drosophila melanogaster (J 1).

CKAMP44 is most abundant in the hippocampus. specifically in the granule cell layer of the dentate gyrus, Subcellular fractionation of mouse forebrain identified CKAMP44 within membranes of postsynaptic neurons (the postsynaptic density region that contains AMPA receptors), When expressed 111 hippocampal neurons in culture, CKAMP44 local ized to the surfacernembrane of'dendritic spine heads, opposite presynaptic release sites. These observationssuggested a role for CKA1VIP44 as a modulator of synaptic AlVIPA receptors. However, von Engelhardt et £II. found that unlike TARP proteins, such as stargazin (TARP y-2), that enhance steadystate AMPA receptor responses, CK.A.MP44

'" decreased such responses. TARPs increase ~the surface expression of Al\.1PA receptors, ~ slow receptor deactivatiOll (delay channel is closure afterglutamate 'is removed), decrease ~ their desensitization (reduce the deCline in g response seen in the continued presence of )5 glutamate), and increasechannel conduc-

tance (2~4), Although CKAMP44 reduces steady-stare currents, it does not affed J\.MPA receptor surface expression, Instead it increases, and slows recovery from, desensitization. Exactly how CKAMP44 modifies AJ\I1PA receptor behavior is not clea.r, but by analogy with effects caused by receptor sub. unit mutations (13), the authors suggest that CKAMP44 might stabilize the closed conformation of the glutamate-binding cleft, most likely by interacting with the extracellular domain offhe AMPA receptor.

Whatdocs this mean for synaptic function?

Von Engelhardt et at examined miniature excitatory postsynaptic currents (mEPSCs) that result from the release.of'glutarnare frQm individua I vesicles' in the presynaptic neuron. Whereas TARPs shape both the time course and amplitude 'of 111 EPSCs (14), CKAl'vfP44 did not. Thus, ill CA 1 pyramidal cells of the mouse hippocampus, neither the removal nor overexpression of CKAMP44 affected mEP SCs. A modulatory action of CKANlP44 became apparent only during high-frequency transmission. When evoking pairs of EPSCs in CAl pyramidal cells in which CKAMP44 was. overexpressed the normally observed increase ill EPSC amplitude from the second of two closely timed stimuli (a form of synaptic plasticity termed short-term facilitation) was eliminated for Al\.1PA receptor-mediated responses, but responses from a different glutamate receptor type f)V-luclhY!-D-aSpattate (Nl\IIDA) receptor] were, unchanged. In CAl pyramidal cells, CKAMP44 abundance is relatively low, whereas in dentate gyrus granule cells, it Is high. In the latter, increasing CKAMP44

expression had no effect, but removal of CKAMP44 enhanced short-term facilitation. Thus, CKAMP44 may playa role ill setting the extent of short-term plasticity

at different synapses. .

How does this alter our understanding of such plasticity? At "facilitating" synapses, the increase in EPSC amplitude reflects a short-lived enhancement of neurotransmitter release brought about, in part, by an increase in the eoncentration of calcium in the presynaptic terminal, Ordinarily, desensitization of postsynaptic AMPA receptors, wh_ich opposes such facilitatiol1 (by depressing postsynClptic responsiveness), plays

little part, as recovery is rapid, Von Engelhardt et ol, demonstrate that CKJ\.MP44 ati.el'lwtes facilitation of tbe: post. synaptic response by slowing this reeevery from desensitization. Likewise, synapses that normally show short-term depression exhibit facilitation in the absence ofCKAMp44.

At hippocampal synapses, CKAMP44 is differentially expressed-but is this expression dynamically or developmentally regulated? Indeed, why is Cl(AMP44 necessary? Recovery from desensitization, and, by implication, short-term facilitation, might equally well depend on AMPA receptor subunit composition and the nature. of any associated TARP Understanding these issues will require i nvestigation of the stoich iometryof the interaction between CKA"MP44, the various AMPA receptor subunits, and TARPs.

References and Note s

1. 1. von Engelhardt et al., Science 3n, I5:Lll (2010).

2. A. Priel et ai; j. NeurQs(i. ~5,26S2 (2ee5).

3. 5. Tcrnits, V. Stein, T. J. Sloeter, R. A. Nicoll, D. 5. Bted~ NeiJron 45. 269 (?005)_

4. D. Ture15ky, E. G,minger, D. K. Pot.nNu, j. /lJeuroKi. 25.,

74~.s (ZOOS).

5. A. $. K.lo ~f (I/.J Nell($ti. 27, 4'769 (l007).

6. D. Sow et aI.., Nat Neurosd. 12, 277 (21)09).

7. R. Wang et a/., NeurOll 59, 997 (2008).

S, J. Schwenk .~I al., SCience :123, 1313 (2009).

9. S. H, H~in~mano, ~le'ipold, (elf- Mol. UreSti. 64,

13.29 (2001).

ro. c. S. Walker ef ot., CUrro Bioi. B, 900 (2009) 11 t. Ibo~ez·Tilllon et aLI Neurol] 33, 893 (2QQ2).

12. M. N. Wu ef al., Nai, /lJe(l{OsrJ. IS, 69 (2010).

13. M. C. W~wn, C, Gerller, M. L Milyer, (.,RQs€omund, J. Neurosci. 26, 71150 (2006).

14. C. H. (~Q, f. SI-GelOi5. w. Zh~r\g, S. Torruta, J. R H~, Neuron 55, 890 (Zee7).

15. Th e au Ih Of, are S upp orted by Ih e Wellm m e trust.

10, 1126/science, 1187920 SCIENCE VOL 327 19 MARCH 2010




Toward Organic All-Optical Switching

SaifA. Haque' and Jenny NelsonZ

I· n the past deca~e, tl1er~ has b~en a 11 explosron of academic and industrial interest ill the use 0 f organic conducting ("pi-conjugated', materials in optoelectronics, flexible logic circuits, light emission, energy conversion, optical communications (1-3). Compared with inorga nic materia Is, pi-conjugated materials can more easily be timed through molecular design and cost less to manufacture. Advances in physical understanding, synthetic control oftbe chemical product, and computational molecular design now allow pi-conjugated" materi-

als to be tailored for a specific purpose or f,mctiolL On page 1485 of this issue, Hales et (If, (4) report a promising new approach to tile design of piconjugated materials for alloptical commnnicarions,

In optical communica-

tions, information is' passed via guided optical beams with wavelengths typically between 13 and 1.5 ~llTI, where optical fiber transmission losses are minimized. All-optical switches deflect or delay the signal beam through a. change

in the refractive index of a

nonlinear optical material that is stimulated by a second light beam (see the figure, pane! A), In a 11011 I inear optical material, refractive. index and propertics depend on light intensity at strong intensities. The high speed «100 ps) of optical switching relative to electro-optica I switching 0 ffers the possibility of dramatic increases. in the speed and efficiency of a variety of applications.

Currently, optical switching is achieved by means of rare earth-doped optical fiber amplifiers or semiconductor optical ampli£ers. These components are costly to manufacture, restricting the technology to special applications such as long-distance telecommunications. Pi-conjugated molecular mate-


Inputslgna\ -


rials have strong optical nonlineanties and could. therefore perform effectively in a ll-optical switches. Moreover, the active materials cost little and should be easy to integrate into waveguide devices. Organic optical switching devices could thus massively reduce-the cost of'all-optical signal processing. Possible applications rauge from everyday data communications and data screening ill health care ~o fast optical computing.

In pi-conjugated materials, polarization of the molecule by the incidentIiglrtfield causes a



Control bea m

Switthed signa L

No nli nea r optical material


A sophisticated molecular design method yields an organic material for possible application in high-s peed, a ll-nptical comm u nication.

to achieve high 'Y is therefore to Increase the length of the molecule. However, long polyene-type molecules tend to distort, thereby losing the bond length symmetry that enables ground-state delocalizarion and thus limiting the value ofy.

A fllrtlwf nnportant compromise in the design of organic nonlinear optical materials results from the need to keep all absorption losses low while keeping y high. 'Y can be en harmed by using a material in which the firstlioear optical transition, Es' lies just above


E. E





J Two-p.helon transitiom



Linear absorption

Nonlinear a b sorptio n



Self nopyry li u m-ba sed end 9 rou P

Designing materials for orga nic att-cprical switching. (A) I n an a ll-optical switch, a control beam 0 r pu LsI" modulates the dirertion of an input signal and thereby switches, between an on and an off state. (II) In the compound reported by Haleser 0/" use 01 selenopyryliurn "end groups enables the conjugation Leflgth to be extended withoLit compromising bond length symmetry. As a result, the compound has a large ylabsilrption ratio. (C) AbsorptiOn [0551'5 are minimized by choosing a photon energy 17m such that no linea r or non If near transitions are excited at ejther '11m or 2'/100. If, in addition, tiff) lies close to the fi rst lin ear transi non Eg, then 'Y is enha need through a nonresonant effect.

1 D epartrn e n t of Chern i stry, Imperia I (a II eg~, London 5W 7 2AZ, UK. 'Deparlment of PhYSICS, Cenlre for Plastic. Hectronics, Imperial College, London SW7 2Al, UK. E-mail: s .. a.liaque@impenal.a(


change in its refractive index, Sn. This change is proportional to the real partofthe third-order coefficient, 'Y, in the expansion of polarizability in terms of electromagnetic field. strength. For high-performauce all-optics I switches, the molecular material should thus offer a: large r, as well as low absorption losses and-for fast switching-a short lifetime of the nonlinear excited state (5), Large 'Yare found in polyene-type rt)o.lecllles designed so that the C--C bon.d lengths are similar and that the electroni« ground state is delocalized along, the whole molecule (sec the. figure, panel B). FOI a given operating wavelength, 'Y increases V\'~Jy rapidly with molecular length, L. One strategy

the operating photon energy, tun (the nonresonant effect), but at the high field sttcngths used inno.nlinear optics, electronic transitions involving two photons and with twice the photon energy 211(1) can also become important. Therefore, the electronic structure must be designed so that no (linear or nonlinear) optical transitions occur at 2fiw. One approach is to choose a large-gap molecule with E > 2hro (for telecommunications, that means c~oosing E g> 1.65 e \I), but the size of"y is then limited , An a ltcmativeis to find a low-gap molecule fOT which 2"hoo lies in a gap between different rNO, photon transitions (see the figure, panel C). In this configuration, ris enhanced bytheuonres-

19 MARCH 2010 VOL 327 SCIENCE wwwscienr;emag"org

ouant effect and molecularlength is less critical. However, It is challenging to design molecules with the required electronic structure.

Hales ef at. now report a breakthrough in the design and synthesis of a po lymethine-type dye molecule capped with aromatic selenium (Se}-co.ntaillillg end groups. The aromatic end groups allow the frontier molecular orbitals to extend beyond the polymethine chain; they thuseffectively increase the conjugation Jength and reduce E withOllt the symmetry-breaking problem enc~lmtered in long chains. Moreover, the Se substitution increases the dipole moment above the value expectedfor the given polymethine chain length .. These factors result . in exceptionally high 'Y values. In addition, two-photon absorption losses at telecommunications wavelengths are minimized in these m.olecules because of a gap in the two-photon absorption spectrum at 2hw. The combination of high Yal:;d low absorption losses results in an unprecedented ratio ofr to absorption. If these results for solution can be reproduced

in solid films, the molecules could find application as some of the. first high-performance, low-cost all-optical switches.

The work of Hales et al. demonstrates that the design and synthesis of'functional pi-conjugated materials for specific purposes is reachiug anew maturity. Similarly striking achievements have been made recently in other fields. For example, use of conjugated polymers in electrically pumped lasers requires high charge mobilities as well as high Iuminesceat efficiency. A recent study showed that polyfiuoreue polymers with a small fraction of short side' chains can achieve both goals, because the side chains increase interchain hopping and raise charge mobility by almost two orders of magnitude, but do 110t reduce the luminescence efficiency (6). Similarly, in organic solar cells, a free-energy difference between the photoexcited donor or acceptor molecule aud the separated charged state is required to drive efficient charge separation, This free-energy difference limjts the electrochemical potential


energy that can be harnessed. Recent studies (7.8) have shown rhat charge separation can be achieved at lower energetic cost through use of donor polymers with in-chain donor-acceptor character. possibly by polarizing the excited state during the charge-transfer step. Given these advances, tcchnologieal applications. of these sophisticated. materials should not take long to be realized.

Refere~ces and Notes

1. 5. Gun~ el 01., Chern. Rev. 107, 1324 (200n.

2. S. R. Forrest, Nature 4i8, 911 (2004).

3. R. H. Frie~eI et al., NaMe 397, 121 (1999),

4. i. M. Hales et 01., Science .327, 1485 (2010); published

online 18 FebrUory 2010 (101l26h;oen(e.1185117).

5. l- l. Bred~j; rtf ot., Chern. Rev. 94, 243 (1994) .

6. B. K. Yap el ol., Nat. Marer. 7, J76 (zOoa).

7. Y liong er at., Ildv. Moter: 10. 1 OO.2/adl1ld .2 00903528 (201.0).

B. 1- M. ~Lil t~~, J. R, Oummt (hew REV: 10 10 21"r900.27 1> (2010).

9. S.A.H- th~l1ks the Roy"l Society for a RS·Univer.sUy Re>earch FeUmYlhip.J.N. thal1~ the ~a\'~l Sodety·for 0 ilS In dUltry F ellowlh ip, We 'tho nk P. Stwrin ou lor he lpl ul



Fairness in Modern Society

Ki!rla Hoff

E· xperiments in psychology and economics have demonstrated that in industrialized societies all over the world, a substantial fraction of individuals will be fair ill anonymous Interactions and will punish unfairness (1, 2). However, it has 110t been clear whether this benevolent, prosocial behavior depends 011 innate human psychology or norms peculiar to industrialized societies. Henrich et al. explored the. motivation for fairness ill anonymous inter- 11 ctions across dramatically diverse societies and (}J1 page 1480 of th is issue (3), they report that this behavior increases with the level of the society's market integration, measured as households' average percentage of calories that are purchased.

A game used to study how people behave toward others who are not linked to them by kinship or friendship is the Dictator Game, in which an individual (the "dictator") is matched with an anonymous person. The

z pair is allocateda stake of lO monetary units


~ (equal to 1 day's wage in the study by Hen-

~ rich ei al.). The dictator decides how much ~ 1)( the sta ke to keep and how m uch to send



g W(lrl~ B~1111. l8~S H Street, NW, MC3-301. W~~~jnglon, DC G 20433, U5A.E-mail:

to the other player, who is passive. Pure selfinterest would lead the dictator to send zero to the other player. Henrich et CIl, now show a strong and robust positive relationship between the mean amounts sent In 15 societies, including foraging and nomadic huntergatherer bands. and the level of the society's market integration. This is convincing evidence that societal standards of behavior in anonym ous i nteracrions have ccevolved with market institutions.

What features of a society motivate individuals to behave fairly?

By varying the rules of the Dictator Game, studies have shown that one motivation for sharing is the desire not to violate standards of expected behavior. For example, in one variant of the Dictator GIUTie, the dictator, after making an a II ocation decision, is given the option to exit the game and keep the full stake less 11 small amount, The exit option leaves the other player wjth zero but also ensures that he never knows that a Dictator Game was to be played. One-third of the dictators ta ke the exitoption (4). Thus, some participants are willing to pay a price to avoid a situation in which they ate expected to share became tb.ey dislike not doing so in that situation. In another valiant of the Dictator Game, the dictator's choice set is enlarged to include taking money from the other player. If the dictator's choice set ranges from -$5 to +$5 instead of from zero to $5, theproportion of'positive offers falls from n to 10% (5). This suggests that another motivation for sharing is a desire to avoid the most selfish feasible action. This motive would lead dictators to share when the choice set ranges from zero

Economic game. A "Third-Party Punishment Game" VIlas used by Henrith et at. to assess preferences across difh~rent societies. An experimenter is shown dernonstratinq such a game in a remote region of Papua New Guinea. SCIENCE VOL 327 19 MARCH 2010



to $5 but not to share when the choice set includes negative values.

It has been argued that Britain's] eadership in the Industrial Revoluriou-e-ttre onset of modem economic growtJl-<lepended on the unusual strength <'!11l011g European countries of its informal norms against opportunism in business. Although markets were high ly competitive, businessmen displayed a high degree of class solidarity, defined. as "sufficient trust rn one another so that pairwise cooperative behavior was expected and maintained" (6, 7).1n this secure environment, unprecedented levels of cooperation occurred between individuals with commercial acumen and those with technical skills. The exceptional cheaters risked punishment in the form of the exclusion from social groups.

In many settings, maintaining cooperation when interactions are impersonal is greatly enhanced by "altruistic punishment' ofnorm violations. Garnes in which disinterested parties have the opportunity, at a cost, to punish norm violators provide a measure of such alnuism (see thefigure), An open question, however, is how specific features of a society shape tlw willingness ofindividuals to eugage 'in altmistic. norm enforcement. Experimental evidence indicates that selective social status is accorded to those who altruistically contribute to group welfare and that such status enhances individuals' willingness to comribnte to the grollp in the future (8). Henrich 121 til. report evidence that group solidarity rituals have co evolved with

social complexity, These findings predict tlll11 denying members ofa group the possibility to enjoy social status and participate in COJlln1Unity rituals and religion will interfere with the emergence of altruistic norm enforcement. An experiment in India (9) examined the effect of caste status on th.c willingness to punish violations ofthe norm to reciprocate cooperation. The nOIID was beld by both caste groups that participated in the experiment-the low castes, wbo had been subject to the practice of exclusion (so-called "Untouchability"), and the high castes. Although there were controls for individual wealth, education, and political participation, low-caste individuals exhibited a much lower willingness to punish norm violations tJlat hurt members of their Own caste, suggesting a cultural difference across caste status in the concern for members of one's 0\11111 community. Low-caste illdividuals adopted an attitude toward norm enforcement that was closer to pure sel f- interesttha n did. indi vi dual s at the top of the caste hierarchy, There was, however, :110 caste di ffercnce in norm enforcement when the v ictim was not a member of oue's OWl1 cormnu[Jity~both low- and high-caste members pUDished little in that case. Because lowcastes were traditionally denied the possibility of any social status and entry to temples, these results support the idea that aU:rnistie norm enforcement is learned not innate. The findings also suggest that gtmlps denied free cultural expression are at a disadvantage with respect to norm enforcement and collective action,

A society is not just a random group of people with a shared territory. It is a group that shares cognitive frames and social norms ({ 0, J 1). We cannotknow for certain how fairly our ancestors ill foraging bands behaved ill situations lacking relationship in formation, but Henrich til al. bring lIS a closer understanding by studying people in simple societies that may be very like tbose of our early a ncestors, These findings call into question the standard assumption in ecoacmics thatpreferences are innate and stable, and. suggest instead that cultural conditioning of the expression of buman selfishness is a part of'rhe process of economic development.


1. t; Ca~'€[.fr, BehQI'l'ora/ Game Tbeol)': l'Xperiflll1!lB if' Strategic Interaction (Prln,ellm U niv. ~ri!>5, Pnnceton, N], 2003).

2. E. ~e~r. S. Go(bter. Am. fron. Rev. 90. 980 (2000).

3. I· Henrich et ol., Science 3 27, 1480 (2010).

4. J. Q. fain, R. Dowes, Qrgan. Beha". Hum, Ded!.

Proc!'5s.100. 193 (2006).

S. I· A. ust, j Po/it. ECQIi, 115, 482 (ZOO71.

6. E. Posner, Low end Social NWOldHarva,d Univ. POl'S'>.

C" mb ridg~, M A, 2()!)O),

7 . .1. Mokyr, fhe Enlightened Economy: An Economic His/ory oj Britain, 1700-18.50 (Yale !In]v. Pre,l, New Haven. cr. 2010k

a, It Willer, Am. So6ol. Rev. 74. zs (2009).

9. K. Hoff, M. I(s hetta 01" de, !;. Fehr. World BIl 11k PoliO; RES ea ,ro Watkin 9 Pa per M. 5 ~ 40 (2 (09).

10. M, Dougl"",, How InsM~tiQns Thillli: (SYl'lIcu,e Univ. Pre)s.

Syra(ucs€, NY, 1986).

U D. North, Ullder;tandi~g the PhJeess Df Economic Change (p fin ret on U ni v; Pr"", Prim et on. N J, 2(05).



Expanding the Repertoire of Shape Memory Alloys

Ji Ma and Ibrahim Kafaman

The exceptional properties of many materials often come at the expense of I irnited performance in other areas, For example, conventional. metals and their alloys are sttong=-they are good at. resisting stress (i.e, an applied load)-but. they tolerate only a very small amount of strain (i.e., deformation) before they are irreversibly deformed, R.11 bber can easily return to its original shape, even after large deformations, but is much weaker than conventional metals. However, some metal alloys. exhibit "shape

Ma rerials Sden(e and Engineering I flterdisc.ip[inary Pro!j1a1ll and D~part!Mnt of Mgcil<llljl;al .E.o,g'i n~~rjl1g, T!)xa'i A.&M Uniyer~ity. College Station, TK 77.B43. USA, E'111<li!: ikar<lrnan@\<lI"flu.edLI; jm@oeo.\

memory"; they are strong but can recover from being deformed when heated. This process seems counterintuitive, but these alloys take advantage of solid-to-solid "diffusionless" phase transitions: The atoms rearrange how they pack into crystals in an orderly fashion, and this process changes the materi al '.'I macroscopic shape. Few other materials possess this combination ofstrength and flexibility (see the figure), and clever engineering has exploited these properties=-for example, in implanted medlcal devices ~"lloh asstents. On page 1488 of this issue, Tanaka ~I al. (1') report on a superel astic alloy tIlat a lrnost doubles the useful range of deformation that can bcindttced in such alloys.

An iron alloy may open up new applications for strong materia Is that a re a lsoca pa ble of la rge reversible shape changes.

Superelasticity is theterm used to describe a particular mechanical response of shape memmy alloys, or S!v1As. When a stress is applied to an S.MA-for example, through pulling or bending-a phase transition is induced, most commonly from a high-symmetry solid called austenite to a solid with lower symmetry, called martensite, When this applied stress is removedthe material fully recovers its original shape. Became superelasticity is stress-driven, it: must compete witll other deformation mechaniSnis suoh as dislocation motion and deformahan twinning, The victor will be the process tlmt is triggered by the least amount of stress.

Superelasticity appears only when the temperature is high enough for austenite to


19 MARCH 20·10 VOL 327 SCIENCE

become more sta hle than ITlfUtensite, but any further increase

above this temperature will also raise the stress needed to trigger the transfcrmaficn. With a sufficient increase ill deformation tern pera ture, p Iasti city sets in an d full strain recovery is no longer possible. Thus, superelasticity is restricted to a tenrperature range or "superelastic window" (2). in reality, large "and fully recoverable supereiasticity remains ra:re.

Nonetheless, Tanaka et al, discovered an iron-based SMA showing almost fl111 recovery

of shape change around 15% strain. Such remarkably high recoverable strain has never been observed in polycrystalline SMAs. It dwat'fs the 8% r-ecoverable stram of the most successful superelastic alloys: the nickeltitanium, or Nitinol, SMAs (3). Furthermore, the strength. 0 f the new alloy exceeds 1 0(><1,., which is stronger-than Nitinol and is 011 par with high-strength industrial alloys (see the figure). Its greater strength allows the USe ofsmaller cross sections. an advantage both in cost find design flexibility. It is also extremely duetile, so a roll mill can rednce its thickness by more than 90% at room temperature without cracking, a feature that conventional Nitinol and copper-based Sl\rtAs lack.

Engineers have exploited shape memory and superelasticity to execute assembly steps in remote locations, especially in implanted medical del/ices. Parts arc fa: bri cntcd and then stored. at low temperatures; the final shape is recoveredwhen the part is warmed by the body. Examples of tile utility of superelasticity include stents in. arteries that can open to a larger size, and medical guide wires that unfold. The iron alloy developed by'Tanaka at al. may allow for smaller devices or greater ranges of motion.

Another use for Sl'vfAs is in energy damping-that is, dissipating energy through thermal and mechanical cycling. The movement of atoms during the phase transition of SMAs generates friction, so during each superelastic cycle, some energy is lost as heat or acoustic emissions. Many SMAs can dissipate energy better titan ordinary metals and polymers can, and hence they are used in blast protection, shock absorption, vibration isolation, a..od noise reduction. Application areas are as diverse as aerospace and defense systems, seismic hazard mitigation, mid passive and

Til"niuMl ollloy.

Kevlil,' ~lli"ilIJ".5

1000.., ;SleeL

_ N~w·irOl,·balc<l I h ~pe 'f1e!11ory allO)l

- PreviQUs ·,I,ape

memory ~UQYI

Aluminum "Iloy~



E~'t()n\.rs ((Ubb.~s)


ihemlQpl..,tic PQly~\~r'


10" Woud

-I'ol~m~ric foams

1 0.1



Strong and shapely. Different rna teria Is ca n be cnarecterized by their strength and their recovery from deforrnatio n. 5trength is plotted as the stress limit, the maxirnu m pu lling force per unit area that a material can withstand without permanently rha nging shape. Deformation resista nee is plotted as the strain=the per(entage change in length of the material as it is pulled aparttha t a material C11 n ta ke and still recover its origi nal shape. Most strong materia 15, such as steels, do not easi Iy recover their shape; mast materials that rerever shape readi Iy, such as rub hers jelas· tomers), are not that strong. Shape memory alloys can combine both properties, and the i ron alloy destrl bed by lana ka er ot. has pushed their limit, even fu rther, The values for meta Is and cerarnits a re for eith€'r polycryStillline or arncrpho us structu res; 50 me datil were taken lrom (12).

active control of civil structures (4, 5). The iron-based SMA possesses energy-damping capacity several times that of'Nitinol, making it a parricnlarly attractive damping material, especially for large displacements.

Finally, this truly mu I Ii functi onal all oy is also a strong ferromagnet in its martensite phase, making it a particularly capable ferromagnetic SMA. Ferromagnetic SMAs (6, 7) couple mechanical energy to magnetic energy and enable a. number of applications, They can quickly change their shape. as a response to a change in the external magnetic field, or they can be. used to sense. mechanical deformation or harvest energy from mechanical motion, such as human motion during walking (8-10). Applying and removing a load to this material switches its phase between the weakly magnetized austenite and the strongly magnetized martensite. Changes in the materials' magnetization track tile extent ofthe phase transformation and provide a measure of the strain on the alloy. The result is a noncontact, sojid,state straiu sensor that is highly compliant and can detect even large displacements. Alternatively,. if ~\ pickup coil is WTa pped arou nd the a lIoy, the changes in field from defonnatiol1 generate electrical current in tnl;; coil (I tJ). This Iron-based SMA may also be able to harvest waste


mechanical vibration energy at muchlower frequencies than what can be achieved with piezoe I ectric IT1 aterials,

The alloy developed by Tanaka el al. COITlbines high strength and superelastie strain, good ductility, high damping capacity, and mechanioal-magaetic coupling. Each of these properties would be a valuable addition to the metallurgists toolbox; together, they fonn a.metallurgist's "Swiss army knife" that could allow functions to work ill conceit For example, a stent might be able to report that it is starting to fail to keep the artery open, A damper on a bridgem an earthquake-prone area could report the structura] deflection as well <1$ ils own condition.

Althougb these potential applications are exciting, several technical and economic challenges must be met before this new alloy ~eaches commercialization. High strength, dtwtiiity, and recoverable strain require precise control of microstructure and crystallcgrapbic texture, It is not possible to make use of damping capacity unless the component undergoes relatively large deformation, and high compliance and reversible strain are needed such as for blast protection or shock absorption or in civil structures for seismic hazard mitigation. Sensing and energy harvesting in ferromagnetic SMAs often require a high biasing magnetic field (9, J 0).

Fortunately, there axe ways to' further optimize this .sMA. Microstructure C'lJ1 be controlled completely by conventional coldworki ng processes (such as rolling, extrudi ng, or wire drawing) and heat treatments. Some properties, such as damping, may already be (I deqnare, Energy harvesting may be low, but self-powered devices often have minimal requirements, The iron-based SM A reported by Tanaka et al, is another example in which materials arc 110 longer just a passive support ( 11). By becoming multifunctional, "they can take en more (and even all) of'the active roles in complex systems.


1. Y. tanaka et ai .• Science 321, 148B (2010).

2 V.liu, S. P. Galvio, Acta Mo/~r 45,4431 i19.97).

3. Ie Ot5UKll, x, Reo. Prog. MD/er: Sci. 50, 511 (2005).

4. R. Desroches, B. Smitn,)c fQrthri~a~e Eng. 8. 415 (2004).

5. R. FQ>oie,l;. Y. Ketema.], H. 'Ill. If/I. I SDlids Stnut: 35,

403 (199B).

6. R. kainuma ec at; Narum 439,957 (2006).

7. I. Takeuc~i I:t al .• Nat. Mater. 2, 180 (200,).

8. M. Chmielul et oi., Nat. Maler. 8, 863 (2009).

9. N. N. S~r.wale, M. i. Dapino, IIp!,l. phys. Lett, 93,

062501 (200S).

10. I. Killama" et. ai., At'"I. Phys. Lett. 90. 112505 (2007). n. L. ~hl'1,tndauloull. D. Venable', 10M 55, ~9 (2003).

12. M. F. !l5hby. Materials Selection in Me'hani<a/ Design (Butterworth· Helnem;1nn, Arnsterdam, 200S).

iO.1126ISrien'l'_lIB6766 SCIENCE VOL 327 '19 MARCH 2010


An Evolving View of Saturn's Dynamic Rings

J. N. CuzzV~t). A. Bl.lrns,2t S. CharnOl,l R.. N. ctark,4 J. E .. (olwell,s L DOlles,6 L W. Esposito/ G. Filacchione,1' R. G. French.' M. M. Hedman,~ S. Kempf,lO E. A. Marouf,n C. 0, Murray,ll P. D. NichoLson,2 C. C. .Porco,H J. Schmidt/II M, R .. Showalter,:I-S L J. Spj!ker/~ J. N. SpitaLe,J.J R, Srama,l° M. Sremcevif/ M. S. Ti'stareno,' j. WeissU,17

We review our understanding of Saturn's rings after nearly 6 years 'If obs.ervations by the Cassini spacecraft. Saturn's ring~ are composed mostly of water iCe but also contain an undetermined reddish contaminant. The rings exhibit a range of structure across many spatial sGiles; some of this involves the lnterpLay of the fll)td nature and the ·self·gravity of innumerable orbiting centimeter- to meter-sized particles, and "the effects. of several peripheral and embedded rnoonlets, but much remains unexplained. A few aspects of ring structure change on lime scales as short

as days. It remains unclear whether the vigorous evoLutionary processes to which the rings are subject imply a much younger age than that of the solar system. Processes on view at Saturn have para llels in cireurnstellar disks.

S atum Is encircled by ·eo extensive ring sys:em that, like the lings ~IToun~i~g Jupiter, Uranus, and Neptune, resides within a region where tides from the parent planet frustrate aggregation of IDe ring particles into larger bodies (1). Several 1- to I OO"km moons are interspersed witbin, and along, the peripheries of each of the four systems. Saturn's rings are distinguished by [heir fur greater mass and by the purity of their icy particles, which is inconsistent witIl the unprocessed primordial mixture of ice, TO ck, and carbon-rich organics thai make up the other Illig ~'Ysterns. A bewildering diversity of snueture permeates Saturn's main rings (Figs. 1 through 3) (2), which include the A ring, separated from the massive B ring by the Cassini DivisiDn-ltself a ring-inward through the C ring and the nearly

lAm~s R~Il"r(h (~nter, NASA, M;ril Stop 245·3, Moffett Field, CA '94Q35-1000, USA. 'Department of Astronomy, Cornell Univep. ~ty, Ilh~!;,l, NY 14853, USA. 3L:lborJtojre ~trtJphy,fqll@ lnstrumentation Mbde\i'>dtion, Paris DiderotiCOfijm],saJ"ij1 ill'Energie Atomiq\Ie/CNAS, 91191 Gil sur YveUe Cedex, Fnmce. ~'U.S. Goologicat Su!Vey, D!>IWN, CD 80225, USA. ~Department oj PIl~L4, UniVersity of Central FIG/idil, Ori,)ndo, FL 3281,6, USA. 6 Depd rtmenr of Sp<1(~ 51\1 di es, SoutliWe'it R~5earth Institute, BOllld~r, CO 8Q302, USA 'lilbaratorylor Atll1o;plle~rand Space Physics, University of Colorado, Boulder, co 8()3Q~392, USA. sl~litllto di Astrofisi(J Sp.lli~le e l'isi(a Cosrnica, Rome 00133, ItalY. 9.AslfIJ\lOIi"J)Wepartni@IiL W~ll@s!~Y'(olleg", W@lle~ley, 1M 02481., UsA. l°Max·Plao[j{·ln$t1tu\ fijr Kernphysik,. Sauplerch· eckweg 1, Heidelberg 69H.7, GermarlY. 1JElectrical Engil1!l'eril1g Department, S,mJo,~ Sta\ll'University, SJnJo,~, CA 9519Z, uSA. 12 Astronomy Unil, Queen Milry U~ iVer>ity of LIJnd on, London E14NS, UK. l.'Ul,s'i~i lmagiilg Centr<ll L~boratory for Operations ((I CLOPS), Space 5ti~Il(·e lnstitute, Boulder, CO 80~O1. USA. Jo1lnmlulI> lor Ph~cs and A,tf!lllOll1Y, UnlVer!dt;y 01 Potsdam. Germ~ny, "'SETl lnstitute, MO(lnt<Jin View, CA 94043, USA. l"let Propulsion laboratory, CalilQrni~ l n sliMe ~I 1 ~ chnology, Pa sa dena, CA 91109, U5A. Ilearl Non (0 tleg e, NM!1fi~ld, MN 55057, USA.

"'10 whom correspondente should be addressed. E-mail: jef!r~Y,CU"l1:i @na~~ .gllv

tThe first two author, led lh.e manuscript's preparation: all

others ~re listed alphabetkally. .


transparent D ring. A transition region beyond the A ring contains the complex. multisranded Fling, and arrayed yet farther outside the main rings are several diffuse rings composed of'minute amounts of mbh'le and microscopic. "dust," Withiil this diversity can be found snuetnres that haveanalogs in the other three ring systems, Furtherrnore, the physics driving the evolution ofSatum's rings JDd determmi 11 g their fcrm has P<U'l1l e Is with th e processes active in protoplan etary dis ks,

The VoyagcN)l"a (1980s) perspective was that today's planetary ling systems cannot 00 primordial but must be continuously regenerated ficin their local arrays of moonlets, througb vigorous evclutionary processes (3, 4), To create Uranus' narrow rings. or the diffuse rings of Jupiter 01" Neptune, merely requires. destroying a. [. IO 10- km-diametcr moonlet by impact with a beliecentric interloper. The ongoing evolution suggests that Saturn 's rings, or parts thereof, might be only one-tenth the solar system's age, a greater ella] lenge given their large mass, JUst how Saturn's rings funned. and when, remain the most basic questions. driving their exploration by [be ongoing Cassini-Huygens mission [see Supporting Online Materia! (SOM) text I]. An emerging perspective, after almost 6 years of study, is that Saturn 's rings show dramatic variability on much Sl1011:cr time scales- decades, years, even weeks.


The story of ring structure begins with dynamics at the smallest level: Interactions between individual ling particles, Voyaget· and Earth-based occultations (5) revealed iii broad ring-particlesize distribution extending from centimeters to meters in radius, well modeled by a power law baving equal particle area per decade in radius (SOM text 2) Cassini's three-frequency radio occultations disclose rich radial variabi Ilty in the abundance of the centimeter -size particles across

the system (6). Rings B and inner A appear relatively devoid ofthese small particles compared with the C and outer A rings. Their abundance in outer ring A increases dramatically with ling radius .. The very short lifetimes of particles in thissize Tapge to various evolutionary proeesses suggest that sizes are determined by an active acererion-desnucrion OIe1e (7, 8) and are 110t primordial; thus, any radial variations indicate ongoing dynamics (9).

Particles closer to Saturn experience stronger gravity and move. taster than those funhee out, generating Keplerian velocity shear across the ring. Co tlisions be:t\veen p.:rrti 01 es, from a few to hundreds of times per orbit, ere basicto local ring dynamics, (9). Although ihe ring particles orbit Saturn at -20 km/s, impacts occur at merely O.O!, to 0.1 cm/s, These collisions are inelastic, damping relative motions of the ring particles: this c ireularizes their orb i ts and f] artees the system toward the planet's equator plane. Meanwhile, these small random motions are repienished by collisions and gravitational encounters with large parti cl es and dUlTI ps of parriel es, ultimately deriving energy from the overall orbital motion. The vertical excursions of pap tides out of the plane arising from this small random velocity establish !I ring thickness ofa few tens of$ at most (SOM text 2) (9), In regions 01' low-to-medium optical depth, the ring kinetically behaves like a dense g<J;S, of macroscopic particles, with the random velocity corresponding to gas "temperature; pressure and v is cosi ty can be assigned to ttl e ling material a S well (see SOM lext] to 4 for examples of'liquid, or even sefid, behavior), Most observed ring structure is created by the interplay between ring fluid dynamics and gravitational furces. Compared with other astrophysical disks (galaxies or protoplauetary disks}. Sauun's rings are extremeIy tbin tor dynamically "cold") owing to frequent inelastic collisions. Accordingly, the ring's selfgravity can be sufficiently strong compared to pressure forces to foster widespread, small-scale. gravitational instability.

Self·Gravity Wakes

The A ring's brightness bas been observed to vary systematically with longitude (10, JJ). Motivated by studies of galactic disks, the underlying structure was explained by gravitational instabilities, where ring particles clump under their mutual self-gravity. In the A ring. Keplerian shear then stretches these clumps into elongated self-gravity wakes having a characteristic cant angle of:Wo to 30"to theIecal orbital marion (Fig. 4 and SOM text J). This clumping of particles, 00"11 scrutinized in Saturn's rings with Cassini observations, is analogous to planetesimal formation through gravitational instabilities in the proroplanetary disk (12), However, in [he protoplanetary case, the surrcunding nebular gas. (missing in tho rings} bas much grearer iaffuence, whereas tidal forces have a lesser effect.





1'00000 loSooa 110000 i I~OOO 120000 , 25000

Ring Plane Radius (km)



Fig. 1. An overview of Saturn's main ring system. (Top) Radial profile of ring optical depth from stellar occultations (S). (Bottom) True color image. The B ring---especiaUy its dense central and ouler portion-cis filted with irregular structure that remains punling. The outer C ring contains a series of plateau

Cassmi's stellar and radio occultations and images (J 1) have revealed self-gravity wakes to be ubiquitous throughout the A and B rings (but app arently absent elsewhere). 111m is, the rings consist of dense self-gravity wakes packed with particles, alternating with less densely populated gaps containing isolated particles, The light transmitted through the- rings is thus controlled primarily by th.e gap sizes relative to the wakes, and sec, oudarily by the optical depth of material within the gaps (13--17). In optically thick regions of the B ring, for example, the opaque wakes cover --80% of the Ilng surface area, separated by gaps with a normal optical depth of --i12. Analysis

of occultation data, using simple models fur the wake's geometry, suggests a wake height of less fhan 10m and in some regions below 5 m,indicating that the wakes are flanened relati ve to the i:r lateral extent and consistent with direct measurements of ring edge thicknesses by occultations. Occultation data have been widely used to derive the ring's surface mass density, but the bulk of the ring mass may be concealed in these ubiquitous, opaque wakes (J 8) {SOM text 3}.

features that are also not understood. The dark gap in the outer A ring is the Encke gap (fig. 2), and the very narrow gap near the A ring's outer edge is !he Keeler ga p (Fig. 3). The identifiable brightness features in the A ring are spiral der]sity and bending waves (Fig. 2). [Figure from (11).)

lings (14, ]9,.10. These features appear to be periodic and, in contrast to self-gravity wakes, show uo measurable cant angle relative to the orbital direction. This axisymmetri c structure may ari se spontaneously from an oscillatory .instability or "overstability" (21) if the ring's viscosity increases rapidly enough with its surface mass density (SOM text 4). Because the requisite density is present across most of the B ring, overstabilities were predicted throughout it (21 23). However, althougll Cassini radio occultation data (211) identify candidate structures routinely in the Bring, overstabilities are not always apparent .. Thus, it remains

unclear what could make parts of this ring overstable and others (with otherwise similar properties) not Perhaps strong self-gravity wakes locally prevent overstability (9, 22). In any case, the SO~ called "irregular strucnu·e"that permeates the entire Bring (11) (Fig. 1) has a radisl scale<::JOO km, far too large to be expl ained by these oversta bilities, Its cause remains unknown (see, however, owdiscussion of Ring Origin and Evolution below).

Spiral Density and Bending Waves

Saturn's satellites, orbiting beyond the rings or within ring gaps, can excite spiral waves at


Organized, axisymmetric wavelike structures having only a few hundred meters radial length scale have been detected in the A and B

Fig.Z., A montage made from Cassini images, .showing part of the outer A ring, induding a prominent spiral bending wave and density waves (waves labeled by resonance, driven by Mimas, Pandora, Prometheus, a nd Janus), as well a~ the 320-km wid\' Encke gap {right),lIIIhkh contains severa 1 ringLet5, one associated with the lO-krn-radius embedded mooniet Pan that orbit5 in the gap's center. The wavy in net edge and the fan-shaped disturbance inside the edge result from the gravitational perturbations exerted on nearby tin 9 material by Pan. By comparison with spi ral wa.ves in galaxies, the spiral density and bending waves are very tightly wrapped, like walch springs, [Figure from (74).J

www.sciencernaq.orq SCI ENCE VOL 327 19 MARC H 201 0



locations where the orbital frequencies of ring particles are cororneosurate with those of the perturbing moons (SOM text 5). At these socalled resonances, nrrg particle orbits cart be perturbed ci tiler within or perpcndi cular to their orbit planes, resulting in compression (density) or transverse (bending) disturbances, respective Iy. These disturbaoces are transmitted by the ring's local self-gravity, propagating as spiral waves until damped by viscous effects (24) (SOM text 5}. Spiral bending waves (25) due to Mimas (Fig. 2) produce vertical corrugations in the ring with amplitudes as large as I. km After their-prediction by analogy with galactic features, numerous. spiro] waves' were detected by Voyager and Cassini, especially in the A ring (which, being closer to the perturbing moons, contains abundant resonant locations), but also in the Band C rings. Spiral waves in rings are more tightly wrapped than their gal actio COt! nterparts because th e rin gs , mass is small compared to the central planet's mass,

The local surface mass density-a critical property fur understanding ring evolution-is directly inferred from fhe wavelengths of spiral density and bending waves, TIle inner-to-mid-A ring is cba racterized by densities ~40 g/ crn' , whereas densities in the Cassini division are only a few g/cm2 (11). Comparing the mass. densities with the corresponding optical depths reveals substantial regional variations in the mean particle size (e.g., more. smaU particles in the C ring and Cassini division), consistent with radio occultation results (6, ll]. The damping ofspiral density waves measures the rings viscosity, which arises from imerparti cl e collisions plus Keplerian shear ami increases outward in the A ring (1] 24),. which suggests a gradually increasing contribution of self-gravity wakes to the rings' total viscosity (26) (SaM text 2). The value of viscosity also constrains the' rings' vertical thickness to 3 106m in tbc Cassini division (11) and <10 to 15 In in the inner A ring (24).

The self-gravity wakes (Fig, 4 and SOM text 3) have radial wavelengths of 40 to 60 In, much less than those of the spiral density waves propagating through regions where they are common, so all material in the wakes. should contribute to the surface mass densities calcu lated from the waves. Moreover, the wake lengthscale itself may be used to infer tho local surface mass density, providing an iudepeudent check. The central B ring contains regions that arc 110t sampled by spiral density waves. Thus the local surface mass density in the central B ring is essentially unconsirained and could be twice historical estimates (whk,h are ~IOO glcm2) or even more (18).

Spiral density waves transfsrangular momentum between the rings and the forcing moons; thus, the orbirs of the perturbing moons evolve outward, while those of the dog particles gCGay inward, at rates that limit the possible age of the ring-moon system The magnitude of this effect (2 n suggests that neither the A ring nor the dose"


in ling-moons could have retained their current separation for the solar system 's age- {Joe of two indications of youthful main rings. Cassini observations bays validated the gravitationa I torque theory in the context of two. embedded moonlers, where resonances merge (see below), even while recent work (28) has illuminated lcngsrandmg questions abeut how the moen Mirnas constrains tbe B ring edge at its isolated 2: I resonance (SOM text 6). Direct measuremerits of orbital evehstion of the ring-moons under gravitational torques have been fiustrated by dynamical. chaos (29, 30).

EmbeQd!!d Moonlets

Elubeddeo moons can open complete circumferential gaps in the surrounding nearby ring

mass objects can create cavities in circurnstellar disks (32).

However, despite substantial campaigns by Cassini, moons have not yet been found inhabiting andclearing the other [3 named gaps in Saturn's lings. Five ofthe regularly spaced gaps in the Cassin) division may be responding to subharmonics -assoeiated with the B ring's distorted edge (2S); that edge, which oscillates inand out by as much as 75 krn, appeal's to undergo unanticipated large angular llbrations or even circulations (33) relative to Mimes's longitude. Here the moving, nooaxisynunetric ring edge itself m\g11t play the role of a perturbing moon let. Even if this exp lams the Cassmi divistoU .gaps, the clearing of other nonresonant, apparently moonfiee gaps, most in the C ring, remains baffling,

---- .- ~

F~g. 3. Cass1nj image of the 30-km-wid€' Keeler gap at the A ring's, very outer edge, showing its 4-kmc radius embedded rnoonlet Daphnis 'lIang with the wake it creates in ring material at the gap edges. The image wa~ I'<Iken very dose to Saturn's equinox, with the sun at very low elevation, 50 !hat Daphnis and the wake's vertical reLief cast shadow'S, The bright, narrow feature toward the top of the image is the Mima5 8:5 bending wave, whereas the other horizontal features include spiral density waves induced by Prometheus (most of the brighter, evenLy spaced features) and Pandora; many density waves' also have some verticalwmponent as indicated by their br:ight}dark appearance. (Fig!Jre from (34) with permission of the American Astronomical Society.)

material by virtue of the gravitational torques transmitted tit theirclosely spaced res onances (27), (SOM text 6). This was first demonstrated when Voyager data revealed tho 14-kro radius moonlet Pan in the Encke gap. and further validated by Cassini's sighting of the 4-km radius mconlet Oaphnis in the Keeler gap (SaM text 6). 111e equilibrium widtb of a moon letcaused gap is obtained by balancing the moon's graviiatioual torque with thering's viscous torque (9). The Boeke gap's measured width is close to that predicted, and the relative scaling between the widths of the Keeler and Encke gaps is also roughly correct given the masses of Pan and Daphnia (3n These results support and constrain the widespread belief that Jupiter-

Particles moving near a gap's edge are tugged as tbey pass the perrurbing moon; their Keplerian shear, combined with the induced eccentric motion, produces a radial oscillation downstream of the moon whose period reflects the moon's distance and whose peak-to-peak amplitude measures the moon '8 mass. If the moonlet's orbit is sufficiently eccentric, or very close to the gap edge, nonlinear effects modify this result slightly (34,35). Collisions should Cause the wavy edges to decay downstream, but the Encke edge's undulations persist around the full circumference .. exhibiting the expected period and several others, too [see the figure in (36), taken from (37)]. Numerical simulations suggest that synchronization of orbit shapes: in the densely packed

19 MARCH 201'0 VOL 327 SCIENCE WV\


Fig. 4. This mod~t Cdku\ation i!1l1strat~s a "propeller" structure (the clark, mainly empty regions on either side of a 20-m-size object located at (O,~O)l. The slanted bright structures aU around are self-gravity wakes (SOM Text 3).

moonlet wakes associated with these edge phenomena might forestall the expected decay (3S) (SOM text 6). Cassiui observations near Saturn equinox have SIlOWll that Daphnis' inclination drives wavy edges that oscillate vertically by 1 to l.5 kin along (he Keeler gap'S perimeters (Fig. J) (34), whereas the Encke gap edges have undetectable vertical relief; consistent with Pan's kick of a measurable inclinaiion.

Most of the moons that lie within or close to the rings- Pall to Pandora display appreciably Il on spheriea I forms, surprisingly low densities (substantially smaller than the density of solid water ice). and shapes and sizes that approximately match those of their associated Rocbe lobes (38), which suggests accretion of loose rubble onto a core substantially denser than the ambient ring material (31) (SOM text 6). This is reminisceot of nurnerical simulations of local gravitational aggregation of material in Saturn's rings (39) and is apparent in the ellipsoidal shapes of'Phobos and Amalthea, close-in moons of Mars and Jupiter

Packing in Compressed Regions

As ring materia I gets thrust together; in either the crowded crests of resonantly forced density waves, the wakes of passing moonlets, or perhaps the narrowed periapse regions of eccemric ringlets (SOM text 7), changes may occur in the particles' orbits and perhaps even their physical structures. Thefinire volumes of ling particles run also cause the ring material to "splash' vertically (SQM text 4) when compressed. Diverse particle orbits can he jammed into synchronized trajectories such that limited radial regions may orbit as units rather than with the Donna! Keplerian shear, reducing viscous dissipation and differential precession, and perhaps even creating large, clumpy structures (35) (SOM text 7). Disaggregation by disruptive collisions or tidal shedding (7) may follow, lmagesaud occultations show broad swaths of , 'maw" in the innermost troughs between crests of strong spiral density waves (SOM text 7) and adjacent 10 the Encke gap edge (/9,35). These clumps of' "straw," probably formed by packing in the dense wave crests, are kilometers to tens of

Objects ('1 using propeller structures a re too small to detect directly, but statistics on their sizes and distribution can be determined from detections of the disturbed regions on either side. [Figure from (43).]

kilometers in extent. Whether this precess leads to accretion of-objects having some pennanence remains unknown; propeller objects. (see below) arc absent from the regions SUlTOU11dil1g tbe strongest density \¥!I ves (40),


Moonlets with sizes much smaller than Daphnis are unable to clear- a complete circumferential _gap,. because their gravitafional torques are too feeble to overcome viscous diffusion. However, they do create local disturbances that can be observed (Fig. 4 and ,sOM text 8). Such disrurbances, shaped like propellers due to Keplerian shear, were predicted theoretically (4J) and subsequently observed. by Cassini (40, 42, 43). The central moonlets causing the disturbances remain unseen, but their sizes can be inferred from models of 0"0 azimuthally aligned lobes, wi tb U,I} leading (trailing) one offset slightly closer to (farther frem) Sa111111. The radial separation between the two lobes, is a few times the central moonlet's diameter (9). Although the precise photometric and dynamical interpretations of the observations are controversial (SOM text 8), propeller moonlets appear to have radii from tens of'meters to I km, with a much steeper size distribunon than that of the centimeter- to few meter-sized particles that domlaate the main rings (40, 42, 43). The total mass in these bodies is therefore relatively small.

Propellers seem 10 be largely confined to a 3000-kru-wide band in the mid-A ring (43) that isdivided inti) three sob-belts (40). Perhaps each sub-belt was produced by the local breakup of II larger object (41, 43), or the propeller-rich belts ate regi on s where. acc.reti on is enhanced an d! or erosion is decreased (4(1), As inferred for Pan and Atlas t3 I), propeller mconlets may have grown to the ir current sizes by acoreti« n of p orous material onto II solid seed 'Until the moonlet filled its own Roche lobe; the ultimate origin ofthese "seeds" remains unknown. Rarer and much larger propellers have been ideutified in the outer A ring, allowing individual objects to be tracked over extended times where. some display evolving mbits (44) .. Continued monitoring of the orbital evolution of these propellers holds the promise of

directly observing processes analogous to tile .complex evohrtion of a protoplanet through a: circurnstellar disk (32). A small (300 m) rnoonlet has been found in the outer Bring (45) but is missing its di agnostic propell er sid e lobes.

Th.e F Ring

A. dusty band ofrubble orbiting 3000 km beyond Saturn's main lings, the F ling contains a longlived core and several narrow peripheral strands, tens of lou wide, that vary on lime. scales of hours to decades (1 J) (SOM text 9). A fainter dust bell spanning -1500 km (19, 46) ~1;Irfounds the strands. Nearby Prometheus causes the primary perturbations, distorting the ring by tens of km at each passage (46, 47). The phe- 1l0menOIJ is analogous to the wakes-produced by Pan and Daphnis but is complicated by the large variations in closest approach distance resulting from the orbita I eccentricities of (be ring and Prometheus (SOM text 9). For example, as Prometheus approaches and retreats from the ring each orbital period of 14.7 hOUl'S, its gravity repeatedly draws material out from the core to fonn a streamer, while leaving behind all emptier channel (46) (SOM text 9). The cycle recurs every 3.2" of'longirude (i. C., the Keplerian shear over 14.7 hours), prnducmg an obvious quasiperiodic pattem trailing Prometheus (Fig, 5.). The strength of these perturbarions peaks every -19 years as differential precession brings the orbits of'Protnetheus and the F ring into anti alignment; the closest approach between the pair OCCUlTed in late 2009.

Occasioually, more extraordinary events are observed. \Viihin a few days, a ring sector's brightness can double or triple after a sudden injection of dust (48), Cassini images show thai. these feamres subsequently shear" out 10 funn kinematic spirals and "jets" (47, 49) (Fig. 5). Even larger dumps have appeared. in Hubble images (50), with orbits that apparently differ slighdy finm the F ring's core, The nearby object 5/2004 S6 (19)-peJ'haps a ~S-.Ion moonlet enshrouded in dust-vis representative of several bodies tbat seemingly pass lhrough the F ring semiregularly, and collisionally trigger these events (47, 49, 51). A particularly bright and dense structure appeared SCIENCE VOL327 19 MARCH 2010



in late 2007 with properties much like those of the main F-ring core but more than 100 km aw.ty in places (52) (SOM text 9).

Tb6 primary core of the F ring bas an eccentric, inclined orbit that precesses smoothly (52,. 53), maintaining its integrity in seeming defiance of the large distortions and variations present, and. like Uranus's lings, avoiding differential precession as well. Because of the proximity of massive Prometheus and Pandora, which havenurnerous overlapping resonances, the dynamics of the Fling and nearby objects arc more likely chaotic than shepherded (54). Stellar occultations have revealed opaque (or nearly opaque) bodies present throughout the ring's core, from 30 to J 200111. in diameter (55) These may be members ef a previously unseen

influential (59) (SOM text 10). These taint rings, and their analogs in the Jupiter, Uranus, and 'Neptune systems, may have parallels if) circumstellar debris belts, whose apparently COlJfined edges ate considered to signify unseen p lanets (150).

Cassi ni observations have clarified the origins of many- !'bitlt rings. Plumes ofmloon-sizedgfains emerging from warm fissures near Enceladus's south pole likely supply the extensive E ring (58,61,62). More commonly, dusty rings are fed by mutual collisions among, or meteoroid erosion ot: various small parent bodies (59). Most of the moons interior to En eel adus 's orbit (including Pan, Jfl11us!Epimetheus,. Palleoe, Methene, and Amhe) generate faiot rings or resonantly cootined ares of material in their orbits. (58). The G

infrared spectra and radio/radar observations (6) (SOM text II). Particles in the C ring and Cassini division are known to be dirtier, compatihl e with models 0 f extrinsi C po lluti on by carbon- and silicate-rich meteoroids over the rings' lifetimes (6R). Cassini near-infrared observations have ruled out any COz., CH4, Of NH3 ices at the percent abundance level, yet all of fhese species have been detected on Saturn '5 moons (SOM text II). At wavdel1g):hs<520 TUn, the. A and B rings arc much redder than any of Saturn's icy moons; the Itltl'a_violCi(UV) absorber responsible for this remains a puzzling clue to the lings' origin, Cassini identifies no near-infrared C-H spectral feature in the ring.") which might preclude some large, reddish, organic tholins as possible absorbers (69). Two new candidates

'have been suggested: small clusters of carbon rings {polycyclic aromatic hydrocarbons} aodlor Ft?' compounds. such as nanoparti cl es 0 f iron 0 Iti de, which gives Mars its ruddy color. The idea of "rusty rings" was inspired by Cassini's identi f cati on of the rings' oxygen atmosphere and their spectra! d issi III i lari ty to supposedly organic-rich reddish icy solar system objects (6). The degree of visual redness is highly correlated with Water-ice baudstreogths as a function of radius (lO)-with redness and ice band depth increasing together in the more massive ring regions, suggesting that the distributed intrinsically, within the ice grains in the regolith of the ring particles, rather than as a distinct, or extrins i c, com poneo t,

Ring Origin and Evolution

Arguments that the rings may be just one-tenth as old as the solar system are (il mutual repulsive density-wave torques between (primarily) the A ring and the nearby ring-moons and (ii} meteoroid restructuring ami pollution of-the ring material (SOM text 12). These short Lifetimes are problematic because the generation of the entire ring: through disruption of a Mimassize (or larger) ~an~nt is unlikely on this. time scale (3. '11, 72).

However, loopholes remain in the young-ring arguments. The gravitational torque theories on which (i) depends have now been validated by observations of mconlets clearing gaps. Some flexibillty in their implicatiens for ring age may emerge if ring-moons periodically interact and perhaps temporarily destroy eaoh other (73) or are held IIp by much-sought-for, but as-yetunideutified, resonances with exterior massive moons (27). Pollution contaminates tbe entire system, but models rely on the poorly koOIMJ I nCOl11 log mas s flux and ri o:g IJ13,SS (SOM text 12). Any substantial increase in the rings' mass could make them better able to withstand the effects of meteoroid bombsrdmem. Finn mass

Fig. 5. A mosaic of reprojected Cassin; [55 narrow-angle images of the F ring obtained at low phase angle. The radial offset is relative to ~ prece5'5ing empticaL model of the F ring (5.2,53), and the horizontal axis is the longitude (in degrees) at the epoch of 12:00 lIre on 1 J,muary 2007. The mosaic is annotated to show the more prominent jets and spirals, thou g hrto be due to recent collisions with anum ber of different crossi ng bo di es, an d th e radia lty extended cha n nels, which are cau.sed by Prometheus and Pandora. /ls seen in the inset (expanded version of 750 to ~25° longitude), most of the short-wavelength featurE's in the bright [Ore are due to perturbations from Prometheus (11, 47,49). (Figure adapted

from (1J)J . .

population of larger bodies that serve as dust sources and tbat provide the mass needed to stabilize the ring's orbit (55,57),

The F ring dramatically documents the difficulty of living near the edge of the Roche zone, where accretion and disruption are in continual t:Ombat (J. 39), Understanding tile evehnion of tile ring bodies, and their interactions with Prometheus, should provide a better grasp on the more genera! problem of protoplanets perturbing a disk of bodies from which they are also growing.

Diffuse Rings

Saturn possesses several other low-oprical-depth rings primarily containing micron-sized grains (58)- Collisions happen infrequently in :;UGh systems" allowing nongravitational forces to be


ring is supplied from a resonantly trapped population of Obj6Cts (including the 500"01 Aegaeon) located near its inner edge (63, 64). These dynamica I configurations testify to the ubiquity of rescnant trapping in fuint debris disks. A Saturnsystem-encircling dust ring has been detected by the Spitzer infrared telescope (65), with radial and vertical dimensions matching Phoebe's orbit; ic too is a debris disk.

Some faint rings have changed appreciably since Voyager's visit (61l} Both the 0 ring and inner C ring display a vertical corrugation that may have been generated only 25 years ago (67).

Ring Composition

The.A and B ring particles are composed of >90 to 95% water ice, based on decades-old near-

19 MARCH 201'0 VOL 327 SCIENCE

measurements from density waves now blanket most of the rings, bot the murky depths of the B ring may contain considerably more material than previously believed (SOM tcKt 12).

Weremairt unsure whether the propeller Objects (or even the visible gap-moons Pan and Daphnis) are residual shards from a creation event or locally grown. Are-the large peripheral ring-moons (at least Prometheus and Pandora) examples of ring precursors, or were they grown within the A ring and repelled outwards by their gravitational interaction with the ring'? Is the 17 ringthe detritusof'somemore recently destroyed member of this moe, and/or does accretion continue there as well?

The composition or the maio rings, <JJ\d its variation with radius, rmgbt yet be the best due as to the provenance of their predecessons); .however, one must first unravel the various evolutionary processes affecting composition (tile ring atmosphere, meteoritic pollution, and the like) and structure. Yet. much of the ring's snucturc=the irregular ~tIUCture covering the B ring; the crisp, sym menl cal, banding in the C ring; and the Cassini division itself-remains unexplained (see, however, SOM text 12).

We have learned a great deal about the rings in the decades since VOYf\&Cf, limn ground-based observations and theoretical modeling, and in particular during Cassini's nearly 6 years at Saturn. Fat more remains to be done. By mission-g. end, Cassin] will return hundreds of times more data than Voyager, and careful examination .of this dan. set is still in its early stages. Explanations fur the origin of Saturn's rings will remain unconvincing until we have understood the powerful dynamical processes that have fcrmed, and continue to shape. these elegant structures on time scales reaching from yesterda;y to billions of years.

Referen(es and Note,s

1. R, M. Ca~up, L w. E5POO5ito, tcatus 113, sai (199 S),

2. The 'optical thickness" or "optlcal depth" 1: ofa ring 01 Jandomly distributed partides 15 defined as

• t ~ f /1(1') 11. r' (Jr, wn. te Ii (j') 15 t~e vef~ CiI ~y i n\~g;~ ted nu m be r de nsi ty per ra di us i ncrernen t for p am des of radius t. and we have assumed that the ring particles ore m urh Iilrg~r thiln tl1 e wdvele~ 9 til of ng~t. For sma II numll@r densi ti~, 1111, h eJlectively rhe pmjected. 5Urf~{:e area of lYartidl!5 per uni t riog are~. U IU81ly ! hiS i s expres5~d as tlJ e "norm" I' opt; ca I depth, correspondi ~ g to r" ys "rrM ng pet pendi cu(a r to th e mea Fl P I)m e of lh!! Ii ngl.

3. A. W. Ha r li5, i 0 Plan'<i¢ry Ring" R. Greenberg, A. Sr ~h i c, .ds, (Univ. Arizona Pres9, Thc,on, 19!!4), pp. 641-659.

4. 5. Chama" L. Dones, l. W. Espooito, P. R .. Estrad~, M. M. Hedm.n, lh Sal1Jm ftDm CI155illH1uygens, M. I( D{)l(g~h"rty, l. W. ~po,itDc T. Kriniigis, Ed,. (Sp ri og ~f Ni' the ria ~d >, Z 009), pp. 535-573.

5. In 0" Q,culta t1on, the '; igna I hom a it..! r tCa SI in; obl~rvf!l o({ultaijol15 1~ ultraViDlet and infrared Illdiatlon) or th~ ,paci'craft (its microwave transmi55ion) IS "Q((lJlt~d" when the ,nurce pal;~' behind thE rin9> "5 ,~el1 by the observer. The ,our~e's Ipatial fQotprint on the ring' is t~e ,nleorei'l Fr~nel-zcme ,i,~ 1ge{)['n"tii~ mEan 01 the wa vel en 9 th and di5ta.nce lrom th ~ ri ng > to th e ob,€ IW r)· im d i5 genera Ily tens of me t~ 1"5., compar il ble tQ the ,I.e III thE Iypic~l l~rge5t ring par~cl!'s. ThLIS, stell" r ,md radlu OCCUltil ti on' are key to 01< in a.\.ertaining ring mkrostructure.

~, J. ill. C~):ri et oi., in Saturn fram (o"inl·fiuygen5.

M.. 1(_ Dougb€r)y, L W. Esposito, T. Krimlgjl, td5. 1Springer N~therl,lOds, 2009l, pp. 459-509.

7. D, Ji. Davis, 5. J. WeldenschiUing. C. R. ~hapmal1.

I'\. Gte.~n berg, Science 224. 744 q 9a 4). !!. p_·V_ lonqaretti, IC(!HI5 81 .. 51 (1989).

9. J. Schmidt, K. Ohtsuki, N. Rappoport, H. S~lo, c. Spahn, In So/urn tram C05.sinf-Huygens, M. K. Dougherty,

L.. W, Esposito, T. Kilmigis Eds, (5pri~ger N"!h~rl",ntb, 2009), pp, '113-4511.

10. "Ill e A, ri n g qu ad ril nts p r ~oedf og the two Ion gi ludE5 al.lgll@d with .11 observer are up to tens or percent di<tufl" than the other tw{). The "'2(t amplitudes depend 00 wavelength and the eleva lion! of th~ observer and ·01 th e S un as seen from Saturn.

11. J.E. Colwell er of., in Saturn tram C,mfl1i·f!uygens, M. K. [)ough~r(y, L W ~po,it{), T. Krimigi'l Ed" (Sp lin gel N eths t lan ds, 2(09), pp, nS'~412.

12. E. Chiil,"g, A. Youdtn, 1111i1l1. Rev. farth pldne/. Sri. (2l}1 0); h ttp:lla rjOUtnill>.~ flnu a I[evi eW.I.orgldoi!ahil 10.11.461:. !'Iou re"~alth-a 40S09·1.5? S 13

11. J. f. COlwell, L W, Bp05ito, M .. Srfm[~vi~ Goophys. iles. lEtt-. 33, l07201 (2(06).

14. J. E. Colw@l~ t. W. Espo,ito, M. S'~"'tevii', G. R. S~ew<irt, W. E. McCljntod::, Icarus 190, 127 (ZOO 7).

15. M. M. Hedmane! vi .• il511'011. J. 1.33, 2624 (200]).

16. P. D. Nkholso n, M. M. Hedm.o .. «arus, in press, Publi!hed oriline 28 i~jy 2009; 101016~i,o'~u,. 2009.07.0.2!t.

17. M. S .. TT~carenQ et G/., Astmn. ).139,492 (2010).

18. S. ]. Robbinl. G. R. S~ewart, M. C lewis, J. E. folweil.

M, Sr~m(eVi(. icaru" in pr<:.>s. P~bti)h~d onlln~ 26 Se ptern be / ZOO?; 10.10 161j. icarus. 2009.0 \),.012

19. C. C. Porco er ai, Science 307, 1226 (2005).

20. F. 5. Thomson, E. A. Marouf, G. 6. Tyler, R. G. Hench,

[II. R~ ppil port, Ge~p"ys. Res. I ett. 34. L2 4 203 .(2 eo 7).

21. \,J. Schmit, iN. M. Ts{hamuter, Icarus T15, 304 (1995-).

22. H. Solo, J. SchmTdt, F. Spahn, /Wrl/S 153, 295 tW(1).

23. J. Schm1dt, H,. S~Lo, F. Sp"h~, 0 .. PmzlmlTHmll. teotu: 153, 316 (2001).

24. M, S. ll,{areno, j. A. Bums, P. 0_ Nicho15oo,

M. M. Hedman, C. C. Porco, teon» 189, 14 (2007).

25. F. H. Shu., ]. N. Omj, I. J' .. tissauer, Icarus 53, 185 (1983).

26. H. Daisak<l, H. Tanaka, S. I~a, tcaru: 154, 2'16 (Z001).

27. P. Goldr.€lch, S. rrem~i~e, I1nml. Rev. Astron. Ilstropnys. 20. 2q 9 (1<)82).

28. M, M. Hedrnan eta/, A~frorJ 139, ~za (40W.

29. R. 1I. Fremh, C. A. McGhee,. L Done" J. J. tissauer; /canls 162, 143 (2003).

30. P_ Goldreich, N .. RapPJport, tcoa« 166, 120 (2003).

31. r. (,. Porco, P C lhomas, ]. W. W;,i'~, D. r; RiChilrd,{)n, S{ief)te 318, 1602' (2007),

32. J. E. Chamber.;, MlllI. Rev. Earth planet. sci. 37, 3'21 (;>'009) .

33. T'hi). '1~r'n1i!,lOt{)gy f{)ltow, t~~( for" vertieally hapgi~g pen dulum, wher e 'u bratie n" If I€ IS to 05d Ita tl ons a bowL lhe stable (tn1I,jmum ellergy) Location and 'ctrcul"tion" describes a (~nlpl"li' tycle lhtough Ihe Sbbl'l poin!.

34. ). W. Weill, C. C. p()f{O, M. S. Ti"arf~o, IlJi'ro~. J. 138,

272 (2009).

35. M. C. Lewis, G. R. Stewart, Icarus 118, 124 .(2005). 3~, J. A. Burns,]. tl. CUll;, Science 31~, 17S.3 (2006).

17. M. S, ll"ar~no ef at" Bull. 1111/. il5troll. 50(, 3], 767 (2005).

3 B. Th~ Rot he lob e 1 s the r~g ion 01 5pMe th ill is gr'ill1tati~niltly bound !o " body. rnl',"Dng the elfect:; Dr itl orbital and rorntionol motion5 and the porent pl,met's Ii des.

~9. R. Karjalolil"lfm, H. Salo, !caM 172, 328 (2004) 40, M. S. Ti"ilreno, I. A. Bum,. M. M.. Hedman,

C. C. Porm, Astl'on. J. 13 S, 1083 (2008).

41. F. Spahn, M. SremC"vi';' As/ron. I15lrophys. 356, l68


42. Ii!I, S, l'i,,~(~no E't ~I., NatUre 440, M e l200~).

43. M. Sfem~evle el at .. NalUrff 449, 1019 (2007).

44. M; S. TIscareno et al., Bull. Am lis/roil. Soc 41. '>59 (2.009).

45. (:. (,. P{)rcQ, Cas(]ni Im?gihg Team, II1U (ire ~()91 12009).


46 .. C. 0, MutroY et ai" Ndiure. 437, 1126 (20.os)_ 47. c. D. Murray et ot., Nature 453, 7}9 (zoo8). 4:8. M. R Show~ltet, teams In, 356 120041.

q 9. S. (nil! '101 e/ 01., Science 3-10. 13 00 (2.o0S).

SO. c. A. Mruh"~, P. D. Ni,Mlson. R G. f'len[~, 1'. 1. H~U.

touu: IS.l, 282 (2001).

51. S. Charnoz, I,ar!! ;201, 191 (2.(09).

5~. N. Albet, et 01., 8l1l1. Am. IIsIr.on. Soc 40, 430 (2008),

53. 1\. S. Bosh, C. B. Olkinl R. G. Frenth,. P. D. Ni[hols{)n, IWf!lS 157, 57 (2002),

54. O. C. Win~er, O. (. MoUr~o, S. M. Giu(iaur'Wi~ter,

F. Spahn, C d. Cruz, Molt Not. R. Aslton. §oc. ]So, 154 (20(1).

5-5_ L. W_ t.pOlito, e. K, M~inke, J. f. Cblwell.

P. D Nicholson, M. M. Hedman, !cams 194., 278 (2008).

56. J. N. OJl1i, 1. A. Burns, I,aro, 74, 2li4 (19liB),

57. I. M. Barbara, L W. Esposito, icarus 16{), 161 (2002),

58. M. Hor<loyi, J. A. Bums, M. M. HEdman, G ] ones., 5, Kempl, in Saturn from (a,sini·Huygen<.

M. K. Doug herty, l. W. E! posi 10, T. IQi m igis, Ed,. 15priOller Nelherlands, 2009), pp, 511-53'6.

59. I. A. Burns., D. P. Hamilton, M. R, Showalter. in /n~erptQ~r;jary DlEt, E. Gruil, B. A. S Gu,r..I'Dn,

S. F. Dermott, H. Fe fhtig , Ed,. (5pring~r-Verlag, Berlin. 2(01), pp. 641-7'25.

60:. I. M. C. rpen ter~1 al., Asfrophys. 1- 181, (Sup~.J, 197


61, c. C POft 0 e/ al., Science 311, 13·9} (2006).

62. F. P05tll~rg et 01., Nature 459., 1098 (2009).

63. M. M. Hedman rt al.. !i(filrJce 317. 653 (2(07).

64. c. r PQlm, Cjssii\1 If{I;Iging Team, II</) Ore. 9[J21 (20M).

6~. A. J. V~rbi~er, M. F. Skryt5kie, D. P. Ha milton, Nal,m! 461, 1098 (2.009).

66. M. M. Hedman E'I 0/., /c~rtI, 18!l, 89 (2007).

67. M. M. Hedman, ), A. Eums, M. M. 1iscareno, C. C. I'orw, 8(111. Am. ASt'fO/!. Soc. 41, 1042 12009!.

68. j. N. CUlil, P. R. Estr.da. tcotu: 13 2, 1 (1998).

69. F. P{)ule!. O. P. CflIi~,h;rnki ]. N. CUlli, T. l, RQulh, R. G. Fr en rh, IlslrMc Il'Itropnys. 412. 305 (2 [IO}).

70. P. D Nkhol>on eI ot; lccnts 193, 182 (2 OOS).

71. I. J. ussauer, S. W. 5qU\'r~5, W. K .. HartmoM, }. Gwphys. /?cr. 93 (Bll). 13 776 (1983).

n. L_ pones, /wru5 92, 194 (1~9.1J.

73. F. Poulet, a 5if<lrdy, Mon. Not. R. Aslron. 50(. 322, 343 (2001).

74. l., Lovett. I. Horvath, J. Cufli. sorum; A N"w Vi~w (H. N. Abrams, New York, 2006).

75. We a pJl 1,1 U d with gr. ti t~d e the (assini spacecre ft en gine~ri ng an d oil era tions teams. We I ha r'lk til e

other merilb ers of the Ca~>i ni Ifi n go Di,r:ip Ii ~e WOf~i n" g roup for their many co~ trib u rt ens, not ;3 II 01 w~ i ch could be described hBf!1'. We thank our Ring ScieMe Planning Team leader .. at IPt -B. Walli', K. p~(!y,

C. Rou m~lioti., R. La[lgi', an d 5. B rool&--il rid 1~ ~

obse rva Ii on pia on in g and desi go spe ci ahs ts on aU th" teams represented here for their ~>lential

contrlb utionj to !lie tTij)~iirn'5 smtess, We ",linOl'l(ed ge M. Lewis, H. Salo, and G. Stewart tor the simulil~om and v\sualllatio~1 described in the SOM ~nd tor helpful wmmeotl.. We allo thon~ three anonymoul reViewers,

R. E. lohll5o~. R, PdPp~lilrdQ, P. K;Jl<!s, and ( NieblH fo. helplul ,ugge5lioos on presentatio)1, Tllf V,5. iluthor,

we re ~u pported by NASA through tn e (~SSl ni ~roj ec t .nd (assin; D. t a An.1 ysi! P 109 ram. Oth et 5Up po rt was

p Will ded by ~h e tta\ ia r'I 51i" te /'Igeng(./l51),

Del.ltsrn e, Zen tm milk Lu It un d Ra ~mta h r t (DlR) I

the N erno" I "stiMe, Uo iverl i Ie Pa 14, Di derot ~ n d

(EA· S~ cl il~, ~ nd tM \1. K. Sci e nee .~n d T f,h Mtogy I';jdlitie, Coun r:i t.

Supporting Onlifle Matenil\

\VWW .,ciencernag,orgicgilco ntentlfu i113 n IS 9 7 2114 7 III Del SOM TeX\

Fig,. 51 to 5~

Table 51


Mavi", 51 In 57

1O.1126~de~{i'.1 1791 UI SCIENCE VOL 327 19' MARCH 2010



an estimate of 10 hours and 32 min 35 l_ 13 s for the in teriorpe Ii 0 d (1 ).. Tbep oren rial vorti ci Iy tPV) gradient, which is related to the curvature of the wind pro me with respe ct. to latitude, imp lies an internal period of iO hours mid]4 min 13 i. 20 s (2). These methods work reasonably well fur Jupiter, for which we know the internal period, but there is no guarantee th<lt !hey work fur other planets, For instance, Venus has an atmospheric period at the I OO-mba\" level of 4 days, even ihougb the planet's internal rotation period is 243 days.

Cassini and Voyager datu show that outside the equatorial region, the winds are remarkably steady by Earth standards (Fig. 1). The winds at the equator vary by -200 m s ", which is to times larger than the variation of Banh's jet streams. 'Whether this is II variation with time or a variation with altitude is still an open question (3), Nevertheless, some slowing seems to have occurred over the 25 years between Voyager and Cassini, Winds ill the stratosphere, above the 10-

rnbar level, seem to vary by:L 75m ~-'l with altitude and time in ,a manner reminiscent of Earth's quasibiennial oscillation (QRO) (4,5). Saturn's large

differeotial rotation, which is larger thsn that of Jupiter, is still unexplained,

A key question is, what maintains the jets against friction? Eddies carry momentum north and south between neighboring eastward and westward jets, and they can either add momentum to the jets or subtract it. If the eddies are tilted rronheast-soudrwcst (NE,SW), then they ate transporting eastward momentum to the north. If they are tUted NW-SE, then they are transporting. westward momentum to the north. In the NW-SE case, if there is a westward jet ill the north and au eastward jet to the south, then the jets are gaining energy TI.'om the eddies. Such eddY-1D-lnean,flow enGtgy tran_'Ife!' has been observed on Earth, Jupiter (6), and Saturn (7). Where the eddies get their energy is still an open question.

Polar vortex and he,xagon The east\ vard jet at 88" to 89cS is' a cyclonic vortex 'With 1mv clouds and high atrno-

8pheric tCl'i1pemtures at its center (8). The clouds .at 88" to 89~S tower SO to 70 lou above the clouds' at the pole (9). The combination of clockwise flow. bigh centro I temperarures, and a ring of high douds surreunding a central "eye" are features of hurricanes (tropical cyclones) in Earth's souihem helll~here. However, the Saturn ey3\vaU clouds. are ~4 times bigher fhan those on Earth (9). Also, the eye on Sanan h as. a diamerer of 2000 ian, which is 20 to 4{l times that of a terrestrial hurricane. Ofher eli iferences are that the structure is fixed at d1C south pole and operates

Saturn: Atmosphere, Ionosphere, and M'agnetosphere

Tamas 1. Gombosil~ and Andrew P. IngersollZ

The Cassini spacecraft has been in orbit around Saturn since 30 June 2004., yielding a wealth of data about the Saturn system. This review focuses on the atmosphere and magnetosphere and briefly Qutlines the state of our knowledge after the Cassini prime mission. The mission has addresseda host of fundamental questions: What processes control the physics, chemistry, and dynamics of the atmosphere? WhNe does the magnetospheric plasma come from? What are the physics I processes cou pHng the ionosphere a nd mag netosph ere? And, what are the rotation rates of Saturn's atmosphere and magnetosphere?

S atum is .the 0 .. 11.1Y. Plallet .. for w.hich we don't know the wind speed. We can seethe planet

, rotate by tracking.clouds in its atmosphere, but the douds move relative to each other; and their average speed is not necessarily that of the planetary interior, The intrinsic magnetic field is no help because it is symmetric about Saturn's spin axis, The relative speeds are large, up to an order of'magnirude larger than on Earth, which is a mystery because the incident sunlight at Sarurn is only 1% that at Earth.

There are. multiple condensates and multiple cloud layers ill Saturn's atmosphere. Hydrocarbon cbemiS1ly dominates because methane is Btl abundant _gaseous constituent Disequilibrium species arise because of absorption of energetic photons, rapid vauCIll transport, electrostatic discharges, and bombardment by charged particles from Saturn's magnetosphere. Sorting these processes out is a major challenge that teaches Us about atmospheric chemisny and evolution,

Most of the plasma in Saturn's magnetosphere originates from the rings, Enceladus, and other icy satellites, In this respect. the Saturn system is similar to Jupiter and differenr from Earth, where the solar wind is the primary plasma source, This difference arises because the solar wind is two orders of magnitude less dense at Saturn and the interplanetary magnetic field is an order of magnitude weaker than It is at Barth. Also, the rotation rates for Jupiter and Swum are -2.5 times faster than Ea nh' s, and their radi i a re -to times I arger, so centrifugal forces play a much Im:ger role.


Winds and planetary rotation. On a fluid planet, the wind is measured relative to the interior; Cloud features, which are rough ly at the l-bar level, move from west [Q east, circling the planet at periods ranging from 10 hours and 10 min to 10 hours and 40 min, Each latitude band has its, own emulation period, indicating that different Ianuide bands have different wind speeds. The question is

1 Depa rtmen t 01 A tmo spheric, 0 csa n ic, a nd Spa ce Scie f1 ces Ufliversi!}l of Michigan. Mn Arbor, MI 48109. USA. IOiVisio 11 01 Geol o~i ca t 0] n d PI <1 n elmy Sci ences, Glli f om la I n5titu1:e 01 Te!:hnol~y, P~sadell", tA 91125. USA.

'1 D whDm corr espon den re ,11 nul d be ad d ressed, E- rnai I:


whether observations of the atmosphere can be used to infer the rotation rate of the interior Cloud-tracked wind datil from VoyagCl'showed that the a:vcmge period is between 10 hours and 31 min and I 0 hours and 32 min. A least-squares 1'it to the oblate shape oftbe 100 mbar surface gives

19 MARCH 201'0 VOL 327 SCIENCE WV\


~oo aoo 400

Eastward velocity (mfs)

Fi g. 1. Eastw:a rd velocity 0 bta i n ed by track) n g dou ds in sequences. of images. The. winds are shown relative to a uniformly rotating cQordinate 5y~tem that WilS defined after the 'Voy.ager encounters from the SKR radio emissions. A. Vasavada created this figure from published Voyager and Cassirri data (9, 35, 36). The velocities are obtained by tracking douds in sequences of images. The crosses and diamonds are from images at 727 om (which senses reliltively high altitud es) -iI n d 750 n m (wtl it h se 0 ses lowe t a: ltl tude 5), respective I y. The thin dark line is a fit to the 750 nm data, and the thick gray line is a fit 10 Voyager data in a broadband tilter that sensE>slower altitudes. The rings o bscu red the northern hemisphere when the Cassini data were taken, a nd they obs'cured' the band from 2° to 11°5 when the Voyager data were taken.

without an ocean below it. Saturn bas no Ioog-lived anticyclonic ovals that rival Jupiter's Great Red Spot.

Saturn's northern hemisphere also has a warm eye at 88" to 89°N with a hurricane-like vorlex (Fig. 2). The next jet to the south, at 75"N, is part of the north polar hexagon. This pattern in the clouds has been in existence since Voyager detected. it in 1980. Small clouds move counterclockwise around the hexagon, following its outer boundmy at a speed of-IOO m S~l. The hexagonal pattern is probably a wave that just happens to fit six times around the globe at this latitude. The difficult questions are, why is tbe wave stable and w bat is special about this jet that causes it to develop waves?

SI(JnnS~ Everyfew years, Saturn has a lightning storm. We have 110t seen the flash es be-

cause light from the rings makes the night side of the planet too bright, But we "hear" the lightning on Cassini's short-wave radio, the Radio and Plasma Wave Science (RPWS) insmnnent (10). The storms last for weeks or months, and usually there is only one. A storm in 2004 produced three active centers, c, ~. and 'Y,. over a 24-day period (Fig. 3). Each center spent a few days in the active phase, producing high dense clouds, and then it became a stable dark spot that drifted off to the west and lasted for weeks or more. The radio emissions were hundreds of times brighter than those from a terrestrial storm. AJJ of the storms observed since 2004 have been in the same westward jet at 35°S planetocemric latitude (-40() planetographic). Great equatorial storms, which sometimes encircle the planet, erupt at ! 5- to 20- year intervals, The last such storm occurred in 1990~

Upper atmosphere, Diffusive separation, in which each gas follows its own scale height, occurs above a certain altitude called the bomopause. Atmospberic turbulence, on the other hand, tends to maintain the gases ina well-mixed state. The turbulence is usually maintained by breaking Waves tbat propagate up from below, but the sources and mechanisms are often mysterious. Wave-breaking can also heat the upper atmosphere. The heated region is called the tbermosphere, which merges into the exosphere, where the molecules follow ballistic trajectories before falling back into the gaseous atmosphere.

Above the homopause, the heavier gases like C2H2 and Ci-LJ separate out, and He, Hz, and H eventually dominate, On Saturn, the bomopanse height is highly variable, suggesting large variability ill the propagation of waves from below The cause of this variability is not known, naris the dominant source of heat for the thermosphere,



the average low-latitude observations, which is consistent with the presence of molecular ions, which decay rapidly at sunset. TIle large variability of electron densities observed at similar latitudes and times suggests that dynamical and/or electrodynamical processes control Saturn's ionospheric structure. TIle main daytime ionization source is photoionization. The ion lifetime is reduced by water, which comes from the rings and icy moons.

On ly H3 -+ ions have been directly observed, by means of infrared remote sensing CJ 3), H3 has three strong emission Lines near 3.67 !lID, and they constitute tbe infrared aurora shown in Fig. 2. Rt and R3'" are the dominant ions in Saturn's ionosphere, Hz + is scarce because it is rapidly converted to H3'.

The fast dissociative recombination of H3 I means that it has a strong diurnal variation, with a minimum just before sunrise,

The ionospheric temperature profile bas also been estimated by using radio occultation data and theoretical models. These calculations obtained plasma. temperatures in the 1500 to 3000 K range, with peak ionospheric densities of _104 cm-3 (iZ). TIle plasma outflow rate from tbe highlatimde ionosphere to the magnetosphere was estimated to be ~l kg s I, which is somewhat lower than the source rate of magnetospheric plasma from the rings but larger than the Titan-associated plasma source U 4).

Fig. 2. North polar hexagon (red) and aurora (blue) in the thermal infrared (PIAl1396). Red shows 5 urn emission coming up through holes in the douds. Blue shows 3.67 pm emission from H3'" ions in Saturn's auroral zone.

Interaction with the magnetosphere through ion drag is a. candidate, Other candidates are photolysis and wave-breaking. Thennospberic temperatures are in the range of300 to 450 K (11). Cooling is by conduction to deeper layers that radiate the energy to space.

Ionosphere and Aurora

lonoephere. Electron density profiles from Cassini (Fig. 4) (11) show that both the average electron density and the peak altitude increase with latitude. A decrease in the mean peak density and increase of its height from dUSK to dawn was also seen in

Fig. 3. False-color image showing depth variations of the 2004 lighining storm with time (l7). The red, green, and blue color planes of the image show the douds at 750, 727, and 889 nm, and are sensitive to deep, intermediate, and high clouds, respectively. The white spots are light, thick douds suggestive of moist convection (35). The blue spots are likely high, thin clouds, but thl;'y (auld also be clouds with a dark (oating on the particles. Carbon released from methane by lightning is one such coating material Ge). The figure shows the birth of !heseactive centers, a, ~, and y, which start out as high thick clouds and gradually turn dark. The dark spots drift off to the west and (an last for weeks or more (35).

www .. sciencemag . org SCIENCE VOL 327 19 MARC H 201 0



Aurora. As on Eartl1, the aurora is caused by energetic particle prccipitafiou The ultraviolet (U V) radiation is fiom excited atoms and molecules tbat are not in thermal equilibrium. Tbe infrared radiation is from the healed atmosphere and is essentially thermal radiation (Fig. 2). The bright ring 'had been seen from Earth, but the bright auroral emission \vithin the polar cap had not been detected until Cassini, The auroral emission arises hundreds of kilometers above the clouds thatare seem at 5 101m. The main auroral oval is ofteo incompicte, forming a spiral that docs not close upon itself. Often there are auroral bright arcs near dawn that rotate at a traction of the rotation rate of the magnetosphere, With the bright spots moving poleward in the afternoon as they fade. Disturbed conditions, triggered by solar wind compression regions, rcsult in the poleward expansion of the main oval. particularly oa the dawn side (] 5). The balance between solar wind control and the effect of planetary rotation is still under


dicities very close to the SKR period, Energetic neutral Hand 0 atoms also show modulations near the SKR frequency (19), Far out in the mag, netotail, there are magnetic field periodicities that m-e close to the rotation period of Sutton (20).

Plasma sources. Observations from the first close Cassini flyby revealed the surprising fact th<1t Eneeladus is actively venting gas· and ice grains (;11-24). These observations also showed that H30 I ions were dominant around Enceladus, The primary gas emitted is Water Vapor, potentially accounring fur the observed vast cloud of water vapor and water products. UV stellar OCc cultationsand a detailed analysis of the 'in situ

ably attributable to the inability of flux tubes at Titan's large orbital distance to execute complete drift orbits around Saturn, so that Titan-originating plasma cannot build up to substantial densities. Whatever the reason, there is no evidence in the inner magnetosphere fur appreciable amounts of plasma of Titan origin.

COm'I?CtiOf! ami plasma drainage. There are two types. of magnerospheres, these like Jupiter and those like Barth. Sannu is in berween. The terrestrial magm:tosph ere is so Jar wind contmlled, and the magnetospheric circulation is referred to as the "Dungey cycle" (31). The process struts at the dayside magnetopause, in which planetary magnetic field lines. and the southward component of the interplanetroy magnetic field (IMF) reconnect. creating open magnetic "field lines originating from the fugh-Iatinrde ionosphere and extending to the free-flowing solar wind. These open field lines fonn the open flux magnetospherio tail lobes (North and Sout11). 111e open flux tubes close by reconnecting behind Earth on the night side, all cI the process repeats,

Jupiter is a fast rotator wetll a strong surface magnetic field" so inc ternal processes dominate the magnetosphere out to ~ I 00 times, Jupiter's radius, and the Dungey cycle is only marginally important, To produces plasma deep inside Jupiter's magnetosphere at a tate of -io' kg s-l. and this adds eausiderable "new mass" to the corotating magnetic flux tubes, The magnetie tield lines remain attsebed to the corotating

ionosphere and are stretched outward by ibe heavy equatorial plasma. The current sy stern intensifies in ord er to balan ce the me, chanical stresses. The centrifugal force acting On the newly produced plasma at 10 exceeds the gravitational force by a large factor and only magnetic forces can confine the plasma. This magnetically confined, centrifugally outwarddriven plasma and the corresponding highly stretched elesed magnetic field lines form what is called the magnetodisk,

If the equatori a I quesi-di polar field ca nnot maintain stress balance with th e p lasma stres ses, the field will become more and more stretched. Qua siperiodica lly, the field lines become $0 stretched and thm that they "break," and a magnetic-loop with high plasma content (a plasmold) is forrned that can now freely move down the magnetotail, The shortened empty field Lines return to the inner magnetosphere via the interchange process- (32). This circulation process is called the "Vasylifmas cycle" (33).

Cassini showed that Sanan's magnerodisk is bowl-shaped and bent upward from the equator (toward ecliptic north) (34), During the season. Saturn's magnetic south pole was tilted toward the Sun (Saturn's magnetic dipole is oppositely

Intrinsic magnetic field and periodiciues. Saturn's internal magnetic field is generated inth (;1 el ectricall 'I conducting region that begins spproximately one third of the wa.y down to the center efthe planet. Par most planets, includi:ug Barth and Jupiter, the magnetic" and rotation a xes are suffic iently different so that periodiciti es in the etosp here provide mfbrmation about planetary rotation. Because Saturn's inI em aJ 111 agn etic fi e 1 d 1 s nearly ax ia 11):, symmetric around the rotation axis, the internal rotation rate cannot be inferred from magnetospheric: periodicities,

One suehperiodieity is the SKR, or Saturn kilemetric radiation. The two Voyager spacecraft measured the SKR period tn 1980 and 198 land found inc be 10 hours, 391nin, and 24 s. This was slower (hr,J11 most of the visible atmospbere bUI within the range of atmospheric periods and was taken to be that of'Satum 's interior. However, the Cassini spacecrafthas shown that the SKR period is variable (16) (Fig, 5) Because Saturn's period cannot W:1I}' appreciably over these short time scales, it appears that the SKR is not a reliable measure of the internal rotation tale,

Periodicities at or near a common rotation period of-'IO hours and 40 min are ubiquitous ill the magnetosphere of Saturn. Because the rotanon and magnetic axis are very closely aligned, the origin of these periodic modulations is still not understood, even Though a variety of processes were suggested for its causes (17, 18). Periodicities have long been recognized in the charged particles of Sa tum's. magnetosphere. hi particular, cbarged panicles with energies above 20 lreV display porto,



Fig. 4, Averaged near-equatorial dusk and dawn electron-density profiles observed by Cassini (12).

particle observations y.ielded a total plasma production rate of -300 ~g s'" (22, 25).

Enceladus is also the source of ions for the ring current, which circles the planet in the equatorial plane and is generated by the Iongitudinal drift of energetic charged particles trapped on dosed magneticfield lines, The ring current plays a major role in detennining the msgneric field configuration. It is primarily composed of accelerated water group ions (26,27), which strongly points to Enceladus as the main plasma source,

Deep inside the inner plasma source region, Cassini measurements revealed the existence of'a tenuous plasma layer in the vicinity of Sanarr's main lings (26, 28, 29). The ion composition of this "ring ionosphere" consists of 0+ and OJ·These ions are likely produced by UV photosputtering (particle emission caused by photon absorption) of the icy rings, with subsequent photoionization of the- O2 (26, 29).

The radial dependence of the W density and the energy of'this population point to a Source hi the inner magnetosphere: most likely Eneeladus (30). TI1is is surprising because before Cassini, Titan was expected to be a major source of magnetospheric N ions, This lack of'nitrogen is prob-

19 MARCH 201'0 VOL 327 SCIENCE WWW.SClencemag.Qrg


Fig. 5. Main magnetospheric results of the Cassi ni mission that are discussed in the text [background figure is courtesy of th e Cassi ni Magnetospheric Imaging Instrument (MIMI) team].

oriented that of Earth). This is different from Jupiter, where the magneto disk warps dowoward, and from Earth, where the magnetodisk is absent (there is no magnetospheric plasma source). Causes for this warping are still being debated.

Saturn falls somewhere between Earth and Jupiter, and is therefore mote complex, It is a fast rotator, but the equatorial magnetic field is comparable with that of Earth, The magnetospheric mass source (Enceladus) is a factor of three smaller than that of Jupiter (10). As a result of this intermediate parameter range, Saturn's magnetosphere exhibits both a Dungey cycle and a Vasyliilnas cycle at the same time (19).

Referen(es and Notes

1, J. b. Andarson, G. Schubert Science 317, 1384 (2007).

2. P. l. Read, T. E. Dowling, G Schubert, Naiure460, 60B (2009)"

]. A. Sanche;-Lavega, R. Hueso, S. Perez-Hoyos, icarus 187 ,

S10 (2007),

4. G. S. Orton et at; Nature 453, 196 (2003). S. T. FQuch~t et ol., NOWre 453, 200 (lOOaL

6. C Salyk, A. Ing~rlOll, 1. Lone, A. Vasa.vada, A. Delgenio, icarus 185, 430 (2006).

7. A. D. Del Genio et 0/" Icarus 189, 479 (2007).

8. L. N. Fletcher et al., Science 319, 79 (2008).

9. u A. Dyudina et al., Icarus 202, 240 (2009),

10. G. Fischer et ol., Icarus 190, 528 (2007),

11. A. F. Nagy et ot., in Sa/urn from Ca5sini-Huygens,

M. Dou.g he rty, l. Es pcsito, T. K" m igis, Ed>. (Sp ri ngE r, New York, 2009), pp, 181-201.

1Z. A. j. Kliore el ai, J Geoph}'5. Re,'. 114 lM), A04315 (Z009).

13. T. R. Gcllall~. M. F. 1;3 gad, 1 ala, A,ltrophft,· J. 408, L109 (1993).

14. A. Glo(er et al.,]. Geophys. Res. 112 (All, A01304 (ZOO?).

15. J.. C. Gerard el oi., J. Geopnys. Res Space Phy'S, 111, AI2210 (2006).

16. D. A GUmetl" et ol., Geophys. Res lett 36, 4 (2009).

17. D. G, Mitchell et ol., in Saturn from Cassini"Huygens, 11ft. Doug herty, l. Es pnsi to, T. K ri m igis, Ed,. (S p ri nge r, N~w York, 2009), pp, 257-280.

18. T. I. Gombosi et a/., in Sa/urn from Ca5silli-Huygens, M, Doug herty, l. E>posi to, T. K ri migis, Ed>. (Sp ri ~gE r, New York, 2009), pp. 203-256.

19. S. v. Bodman, S. W. H. (owley, Ann. GeophY5. 25, 941 (lOO?).

20. D. J. Andrew' et al., 1- Geophy.I. Res~ Space Phys. 113, 19


21. M, K. DaLlgherty et oi., Science 307, 1266 (200S). az, c. J. Hansen et al., Science 311, 1422 (2006),

23. c. C. Porco e( ol., Science 311, 1393 (2006).

24. J. H. Waite )f. et al., Science 311, 1419 (2006).

ZS. R. E. lohnscn el 01, Astrophys. J 644, L13 7 (2006).

26. D .. T. Young et al., Science 301, 1262 (2005).

27. s. M. Krin1igis, ill. Sergis, D, G. Mit(h~l~ n C Hamilton,

N. Krupp, Nature 450, 1050 (200?).

28. D, A. Gurnett et ai., SCience 307, 1255 (2005).

29. J. H, Waite jr, €I al., Science 307, 1260 (2005),

30. H, T. Smith el al., tcaru. 188, 356 (200n 3 L J. W. Dungey, Phys. Rev. Len. 6, 47 (196t).

32. B. H. M;,lU~ et aL in Saturn from [G5.sini·Huyg~n'i M. Dougherty, L, Esposito, T. Krimigil, Eds. (Springer, N~w York, 2009), pp. 281-332.

33. V. M. Va\yliunal, in Physics of the Jovian Magnetosphere, A I Dessl er, E,:L (G! rn bridge U ni v Press, Co robri dqe, 1983), pp. 395-454.

34. C S Arridg€ ef (/1., J. Geophys .. Res. Space Phy,. 113, 13 (2008).

35. c. c. Porco et ol., Science 307, 1243 (200S).

36. A. R. Vo.\ovado et al., j. Geoph}'5. Res Pianets 111 (E,), 13 (2006).

37. U. A Dyudino el al., tcarus 190, 545 (Z007).

33. K. H. Baine, et al., Planet. Space Sci. 57, 1150 (2009J.

39. 'Tl1is work was supported by the CaSlin; mission under contracts I PL 13 58124 (to Uti) and JPL 12 28972 (to A.P.I.).

10. 1126/1oence.1179119

www .. SCIENCE VOL 327 19 MARC H 201 0


Markets. Religion, Com:munity Size, and the Evolution of Fairness

and Punishment

Joseph Henlich,l* Jean .Ensminger,2 Richard Mc.E!reath,3 Abigail Barr," Clark .BarreU/' Alexa nder Bnlya natz, 6 Juan Ca nri 10 Cardenas/ Mi chae! Gurvert,8 Edwi ns Gwako,9 Natalie Henrich,tO Carolyn Lesorogol,H frank Marlowe,12 Davtd Tr<!cer,13 John Ziker14

Large-scale sodetie~ in which strangers regularly engage in mutually beneficial transections are puzzling. The evolutionary mechanisms 'associated with kinship and reciprocity, which underpin much of primate sociality, do not readily extend to large unrelated groups. Theory suggests that the evolution 01 such societies may have required norms and 'institutions that sustain fairness in ephemeral exchanqes, If that is. true, then engagement in larger-scale institutions, such as markets and world religions, 'should be assodated with greater fairness, a nd larger comrnuntties should punish u!i.fairn.ess more. Usin.!j three experiments. administered across 15 diverSe populations, we show that market integf<1tion (measured as the percentage of purchased {<Ilories) positively covanes, with fi1trne~s while community size positively eovaries with punishment. Participation in iI world reli9ion is asscdated with fairness, although not across all measures, These results suggest that modern prosociality is not solely the product of a n innate psychology, but also reflects norms and institutions that have emerged over the course of human history,

At the OIl set of the Holoceoe, the- 'stabilization of ,globsl climates created possibilities for the emergence of largerscale sedentary human societies (l), Over Ole next m millemua, the scale of some human communities dramatically expanded from kinbased foraging bands to complex, intensely c,oop6rativesocietles in Which strangers frequently engage in murually beneficial transactions (2), Consistent with life in these large-scale societies, behavioral experiments performed with people from these populations reveal fair,

'O~pafim~n15 01 P.syc,hology and E(Qllomks, UnlYi'ISiIy 01 B riti III Col urn bia, 213 6 We~t. Mo It, Vancouver, Brili sh Columbia Vel 12.4, Canada. 'Division 01 the Humanilfes all d S oda I S~i eo eel, Cl!i [omia lnstitu Ie of 1 ec!mo!ogy, Pasadena, CA 91125, USA ;OePdrtment of Anthropolqgy, University of California, Davii, CA 95616, USA, 'Centre- for tile Stu dy 01 AI rica n Eco nornies, Departm en! of Eronom ics, University 01 Oxford, 'Manor R.Cl<ld, O~ford 0)(1 3 UQ, UK. ~Departmen[ or Anthropology, Ulliv@l'iity of California, 341 Ha illes H" It, Box 951553, Los An,gete5, CA 90095, \J SA. • Oepa rtmeot 'Of Anthropology, Co!il'g e of O,u Pa ge, <ll@n Ellyn, Il 60137, USA. lFaclJ[lad de Ecol1omiol, Unjve~dad de Los I\ndl1'i, Kl Number IM-JO, Bogot<l, Colombia. UD@part~l~~t of Anthropologt UI;iver;;ty of Cali:oroia, Santa B1nbow, CA 93106, USA. DepJrtment of SOCiology and AntIJrupology, GUit!ord CoUege, 5800: w~.! f1l@11dly Ave~ue, Gre~nsbOTo, NC 27410., USA_ ]oCentre lor Health fVJlllatian and Outcome Sciences, Providen(e Henlth &Ire Rese~rch lrsntute, Vancouver, British (ort!mbi~ V6l 1Y6, GJnada. llGoo@e Warren BroWll Sdinol of Social Work, W~shil1gton UJ1iv~r>ity! 1 Brooll:ing5 Drive, SI. LOllis, MO 63130', USA, 120ep!lftn'lent 01 I\n~uopology, Plorida State Uiliversity, T~llah""ee, H 3.2306, USA, 13DepilrUTlent of Antl1ropology, Univ~rsity of Colorado, POI! om t~ Sox 173364, C~rn pu I B ax 10 3.. D~l1ver, co B 0217 , USA, ]~De paron em of Anthropolo gy, Boise '>Ill te UOivefSity, 1910 Univ~rsity DriV'!, Boise, I D 83725, USA,

'TD whom correspondence shnuld be addressed. E-mait: llenricb@p5y~b.ubc.GJ


trusting, and cooperative behavior among strangers, even in one-shot encounters (3),

Two major theoretical approaches have sought to explain both {he relatively rapid expansion of lnnnan societal scales and ihe puzzlingly prosocial behavior observed in experiments. The first approach proposes that humans possess an innate social psychology calibrated to life in the small-scale societies of our Paleolithic ancestors. Rooted in the evolutionary logic of kinship and reciprocity, these heuristics were mistakenly extended to nonkln and ephemeral interaetams, !IS societies expanded "with the emergence of agriculture, From this view, the prosocial behavior absented in experiments directly reflects the operation of

these ancient beuristies (4, 5), .

An alternative approach proposes that a crucial ingredient in the rise of more-complex societies was tl,e development of new social norms and informal institutions that are capable of domesticating out innate psychology for life in ever-expanding populations (6). Larger and more-complex societies prospered and spread to the degree t'h at their norms and insti ruti ons effi::ctivel y sust;:!! n ed succes 001 interaction in ever-widening socioeconomic spheres, well beyond individuals' local networks of kin and long-term relationships (n It is these particular norms and their gradual intcrnalizatien as proxim ate mati vations (8) th at reealibrate our innate psychoiogy fsr life ill small-scalesocieties in a maimer that permits successful larger-seale cooperation and exchange in vast eommuoities,

Much research suggests that flOUl1S arise because humans use evolved learning mechanisms to calibrate their behavior; motivations,

and beliefs to variable circumstances (7, 9), Modeling work shows that when these learning mecbanisrns 211). applied to different kinds of social interactions, such as large-scale cooperstion or ephemeral exchange, individually costly behaviors can be sustained by punishment, signaltng, and reputationa] mechanisms (10--J3), By sustaining such behaviors. nonns can facilitate trust, fairness, and cooperation in a diverse army of interactions, thereby allowing the most productive use of unevenly distributed skills. knowledge, and resources, as well as increasing cooperation in exchange. public goods, and warfare, More-effective noons and mstitutions can spread amoog societies by a variety of theoretically and empirically grounded mechanisms. including conquest and assimilation, preferential imitation of more-successfirl societiesor forwardlooking decision making by leaders or high-status coalitions (14, j 5).

Norms that enhance fairness among strangers are likely causally interconnected with the diffusion of several kinds of institutions. Here we focus OIl tWO: (i) the expansion of both the breadth and intensityof market exchange (l6, J 7), and (ii) the spread of world religions (l8). The efficiency of'marketexcbange involving infrequent or anonymous :interactions improves with an increasingly shared set of motiverions and expectations related to trust. fairness, and cooperation. This lowers transaction costs, raises the frequency of successful transactions, and increases long-term rewards (l6). Although frequent and efficient exchanges among stmng6!'S are now counnonplace, studies of nonhuman primates and small-scale societies suggest that during most of our evolutionary history, traasaetions beyond the local group! and certainly beyond the ethnolinguistic unit, were fraugb; with danger, mistrust, and exploitation (2, 16, J 9)., Thus, we propose that such "market norms" may have evolved as part of an oV6I:aU process of societal evolution to sustain mutually ben en ci al ex c han ges i n oont exts wb ere esrablished social (for example, kin, reciprocity, and status} were insufficient. If our theory is correct, then measures of fairness in situations lacking relationship infbsmarion (fur example, anonymous others} should positively covary with market integration,

Recent work bas also tentatively proposed that certain religious institutions, beliefs, and rituals may have coevolved with the norms that support large-scale societies and broad exchange (I8, 20,21). Intersocietal competition may have favored those religious systems that galvanize prosceial behavior in broader communities, perhaps using botb supernatural incentives (for example, bell) and rectnrent rituals that intensify group solidarity (20, 22), Consistent with this view, analyses of ethnographic data show that the emergence of moralizing religions increases with greater sooietal size and cOlnplcxltY (J 8; 23)Arehaeologioally, regular rituals and the eon-

19 MARCH 201'0 VOL 327 SCIENCE WV\

strucnon of mormmental religious architecture co-emerge 'with societal size and complexity (14). In experiments, unconsciously priming the fJithful with religious concepts favors greater fairness toward anonymous others al). This suggests that. in contrast to the religions. that li .. kely dominated our evolutionary history, modern world religions such as Christianity and Islam may be unusual in w<lys that buttress the nOl1DS and institutions tbat sustain larger-seale interaction [supporting online material (SOrvt) tcxr], If this theory is correct, then greater fairness toward anonymous others should be -aSBO' ciared with adherence to a world religion,

Evolutionary approaches to norms also afforda. prediction about tbewiUlngpess of individuals across populations to engage in the costly punishment of TI01m violations" Theoretical mode ling reveals a,t least two differen t kin ds of norm-stabilizing mechanisms: On6 involves rep utati {mal effects in which norm viola tors arc sanctioned in another Interaction by, fur example, om receiving aid in a dyadic helping situation (J 0), and a second involves the use of diffu.~e (15) costly punishment (J 2), Because Ole effectiveness of reputational systems in sustaining norms degrades rapidly as ooromunities expand (26, 27), fairness in larger communities must increasingly be maintained by diffuse punishment; tpat is, larger cormnunities should punish more.

Experiments. The evidence presented below derives from a seco n d rOU1~d 0 f cross-cultural experiments that were designed to illuminate findings from our first project (28, 19). This round replaces 10 populations from our earlier

effert with 10 new ones (swapping several researchers as well), willie resampling from five of the same populations used previously (Table I). This analysis complements prior work fiom Our team, which focused on the distributional patterns of punishment (30), by seeking 10 explain the variation in our experimental measures of both fairness and punishment, OUf analysis converges with recent work comparing diverse industrialized populations (31).

1 f markets and world religions are linked to the norms fun t su stain ex ell ange in large- scale societies, we expect that experimental measures of fairness in anonymous Interactions wiD positively covary with measures of 111volv.-:ment in these two iJ1sthmiolls. To rest this, we used three experiments that were designed to measure individuals' propensities for fairness and their willingness to punish unfairness across 15 populations that vary io their degree of market integration and their participation in world religions. Our three experunenrs ate the Dictator, Ultimatum, ~lld Third-Party Punishment Games (j]}

In the Dictator Game (DG), two anonymous players are allotted a sum of money (the stake) in a one-shot interaction (3). Player 1 must decide how to divide thissum between himself or herselfand Player 2. Player 2 receives rhe allocation (offer). and the game ends. Player 1 's offer to' Player 2 provides a measure of Player l's behavioral fairness in this context,

10 the Ultimatum Game (UG), two anonymous players areagain a lloued .1.1 stun III a oneshot interaction (3). Player I can offer a portion of this to Player 2. Player 2, before healing tbc actual offer from Player 1, must decide whether

Table 1. Summary information for the populations studied. The column Ecunomi( Base classifies the production systems (for example, hortirultura lists rely pri ma rily on slash-a nd-burn agriculture; pastoralists rely on herding), ReSIdence classifies the- nature and frequency of each population's movements. Mea,n MI is the average percenta_ge of total household calories that are purchased in the market. Mean WR give~ \he perm[ltage of the



to accept or rejeet each of the possible offers (in 10'1.. increments), Decisions are binding, If Player 2 specified that he or she would accept the amount of the actual offer, then Player 2 receives the offered amountand Player 1 gets the remainder. If Player 2 specified that he or she would reject the offered amount, then both players reeeive zero. If people are motivated purely by money maximization, Player 2s will always accept any positive offer; realizing this, Player 1 should offer the smallest nonzero amount. Because this is a one-shot anonymous interaction, Player 2'8 willingness to reject provides II measure of punishment, Player 1 's offer measures a combination of social motivations and an aS$eSS,l'tlent of the likelihood of rejection; ibis gives us a second behavioral measure of fairness.

1n the Third-Party Punishment Grone (TPG), two players are again allotted a stake, but new a third player also receives the equivalent of onehalf of'the stake (33). Player I must decide how much to. allocate to Player 2. who bas no choices, Player J, before hearing the actual amount that Player 1 allocated to Player 2, has to decide whether topay 20% of his 01' her allocatiou to punish Player 1 for each of the possible offers (in I Q% increments). If punished, Player 1 loses triple the amount paid by Player 3, Suppose the stake IS $100; if Player I offers $10 to Player 2 (keeping $90) .. and Player 3- wants. to punish this offer amount, {hen Player I takes home $60 ($90 $30), Player 2 gets $10, and Player 3 gets $40 ($50 $10'). If Player 3 bad instead decided not to punish offers of $10, then the take-borne amounts would be $.90,

sample that reported adhering to a world religion. Mean (S gives the average C5 for the populations studied. These (55 are lor villages, except among the Hadza (~ho live in carnpsl arid Accra. Com Sam (DG/UGfTPG) gives the number of different communities from which participants were drawn for the DG, UG, and TPG. N gives the number of pairs. or trios for each experim!;'nt (table 54).

Population Lo(ation Environs Economic base Residence Mean Mean Mean (om Sam N
Accra City Ghana Urban Wage work Sedentary 100 97 44t 3/3/11 30/30/39
A!J.* Papua New G.uinea Me tlntai nOU5 forest H C1T1icu lJure ant! for.agi ng Sedentary 1 100 309 3/3f1 30130/30
Dolg<lfl/ Ng. Siberia Tundra-taiga Hunting, fishing, anti wages Se m l-s ede ilia ry 63 59 612 111/0 30/30/0
Gusii Kenya Fertile high plains Farming and wage work Sed8nt~JY 28 100 4063 212)1 25/25/30
Hadza= Tanearria Savanna -wocdlarrds FOf<1ging Nomadic 0 0 43 4/4/3 31131127
lsa ngi! village Tanzania Mounli\i nous forest Fanning and wage work Sedentary to 99 1500 111/1 30/30/20
Marag.oli Kenya Fertile pl.,3ins Farming and wage work Sedentary 43 lQO 3843 212fl 25/25/30
Orma" Kenya Semi-arid savanna Pastoralism Seml-tiomadic 72 100 125 4/0/0 261010
Sarnburu Kenya Semi-arid savanna Psstora lism Semi-nomadic 69 66 .2000 1/1/1 31131/30
Sanquianqa Colombia Mangrove forest f.isheries Sedentary 82 84 1931 111/2 30/30/32
Shuar E(UCl.dor Tropical forest H orticu l!Llfe Sedentary 22 76 498 212/1 21/21/15
Sursurunga Papua New Guinea Island H 0 rticu ltu re Sedenta.ry 24 100 186 3/3/1 30130132
Tsimane~ Bolivia T ropi cal fo re st Horrict! [tllre and' foragjng Semi-nomadic 7 100 314 21.7)1 3 Bl36J27
U nited Sta!es~ Missouri Prairie Wage work Sedentary 100 tOO 1813 lJI/0 15/26/0
Yasilwa Fiji Coastal tropical. Horticu [tllre and marine Sedentary 21 tOO 109 Zl21Z 35134/30
pacific foraging
"Thl!5li'. populatlon5 were at50 lilmpled from during our first project (lB). tPartidpantl were sampled from small enterpri:.eS"in .the dty of Accra, The valuB 44 is t~e average number 01 employees- in
the enterprises, Be'ao~e of the anomalous nature 01 this situation with regar~ to a (S comp~fable with our other papulation,_ we did not include Accra in analy'e> involving (5, SCIENCE VOL 327 19' MARCH 2010



$10. and $50. respectively, Because a moneymaximizing Player 3 would never pay to punish, a similarly motivated Player I should ahvays give zero to Player 2. Thus, Player J's willingness-to-pay provides another measure of punishment. Player 1 's offer measures a mixture of his or her social motivations in this context and an assessment of'the punishment threat; this, provides a third behavioral measure of fairness.

These ex periments, with their galien t eoutextual cues of cashand anonymity; seem wellsuited to tapping the particular norms that support eo. operation and exchange among ephemeral interacsants in market transactions. Cash is closely associated with market transactions (34) and often signals a desire to avoid a longer-term nonmarket relationship (35). Anonymity in our games means that players lack the cues or information neecssary to apply the expectations and motivations associated with other kinds of relationships, sud] as those based 00 kinship, reciprocity, OJ status differences, forcing players to default to local norms for dealing wiill people outside durable relationships. However, an important concern i11 interpreting such experiments involves the degree to which participants accept the experimental situation, believe the anonym" ity, and WOI1Y about the experimenter's judgments of them (SOM text).

Our standardized protocols and methods included tbe following: (j) random adult samples: from OUT communities (with little attrition in mest sites), (ti) stakes set arl day's local. wage, (iii) a show-up fee of 20% of the stake, (iv) backtranslatien of scripts, (v) one-en-one iusnuetion and pregame testing for comprehension, (vi) steps to precludecollusion and contamination, and (vii) no deception (32), Ow' pregame tests combtnedWith analyses of the relationships between expemnental decisions and measures of'eompreheusion, iodicate no measurable impact of differences in comproocL'lsion on behavior (30),

We performed these experiments with 2148 individuals fJ{,\1"OSS populations from Africa, NOJ1b and South A merica, Oceania, New Guinea; and Asia that included small-scale societies of bunter-gatherers, marine foragers, pastoralists, horticulturalists, and wage laborers.

Table 1 provides the location (mapped in. fig. 53), eaviroameot, economic base, residence pattern, and sampling information fur each population, as well as averages fur three key variables.

Results, The theory outlinedabove predicts a positive relationship between ow' three measures of fairness-offers in each game-with market integration and adherence to a World teligion. Individuals' offers are measuretl as a percentage of fhc total stake tI day's local wage). Market lnregratinn (MJ) is measured at the household level by calculating the percentage of a bouse bold's tom! calories that were pure chased from the market, as opposed to homegrown, hunted~ or fished, and then averaged to obtain a community-level measure" WfJ use the community average for MI both to remain consistent with our definition ofnonn.s (as local equilibria) and to remove day-to-day stochastic variation (3ll Table. I shows that the population means for MI range from 0 to 100%, with a mean of 57.3%, World Religion (WR) was assessed by a,skiog participents what religion they practiced, and coding these as a binary Variable, 'with "1 " indicating Islam or Christia,nity; and "0" marking the practice of a tribal religion or "no religion." Table I provides the percentage of each population that adheres to a world religion. The mean value ofWR is 89%.

To analyze these data, we regressed offers on MJ and WR, as. well as seven control variables: age, sex, education, income, wealth, household size.and eommunity size (C5). Exceptfor wealth and household size, which are both measured at the household level, and CS, these control variables are individual-level measures. We used ]0 (DG), 26 (UO). and 16 (TPG) couununity-Ievel means forMI to predict 4 16 (DG). 398 (DOl,and 272 (TPG) individual offers with a minima,l set of eontrels. or 336 (DG), 319 (UO). and 2.65 (TPG) individual offers with the full set of controls,

Table 2 shows four regression models using M1 and WR to predict all ofi'e!·S together and offers £'0111 the DO, UO, and TPG, separately. Independent of the other soeiodernographic variables, Ole coefl1dents on MT are large and s.ignifi.eant at conventional levels across' all four models. A 20""4)ercentage point increase in MI is

associated with an increase in percentage offered ranging from roughly 2 to 3.4. The same qualitative results are obtained fur MI whether one 'Uses bousehold measures, community averages, or population averages.

The coefficients on WR are also large across all offers and in both the UG and DG, though not well estimated in (he DG (P < 0,10). However, WR's coefficients are signlfieant tP < O.OS) across all other specifications (t'dbles S5, S8, and S 11). Participating in a world religion is associated with an increase in percentage offered of between about 6 and 10 (36).

Taken together, these data indicate that going from a fully subsistence-based society (NIT = 0) with a local religion to a fully marker-incorporaied S013lCty (M] = 100%) with a world religion predicts an increase in percentage offered of roughly 23, 20, and 11 in the 00, UG, and Tl'O; respectively. This spans most of the range of variation aCl"OSS our populations: DO means range from 26 to 47%, UG from 25 to 51 %,. and TPG from 20 to 43%"

For the seven other socieeconomic variables. none of their coefficients is significant (P < 0.05) across alloffers or in the DG (tables 55 and S8). For UG offers, toe coefficient on age is also positive (P < 0.05, table S 11). For TPG offers. Table 2\-vs that the coefficients on income, wealth, and bousehold size are sign ifiC?11t (P < 0.05, table 514).

Figure 1 plots mean offers versus mean Ml values for our 15 populations. Population mean Ml values account frn' 52% of the variation in mean DG offers [correlation p = Q . .72, 95% bootstrap confidence interval (Cl) = 0.4 to 1.0 , P < 0.01, rI = 15 populations]. In designing ibis second round of experiments, we sought out an additional New Guinea population (Snrsurunga), because in the first round, the Au of'New Guinea revealed h.igbly unusual behavior, including relatively high offers with little marker integration. The Au petteen replicated, and now extends to the Sursurunga. However, because wc targeted a second population that skews our world sample unrepresematively toward New Guinea, we also examined this relatinnship with either the Au or ihe Surlil.ll\mga dropped Dropping the Sursuruegs

lable 2. tinear regression models for offers. These ordinary least-squares models indude 'four additional control variables (sex, ageJ community size, and education). Coefficients are followed by standard errors, indicated with ±; P values are given in parentheses.


AU offers*


lnrorne (per U;S. $1000) WeaLth (per U.S. S WOO) HousehoLd size Observations: R"

0.12 + 0.023 «0.001) 5.96 t 2.0.4 (0.0036) 0 .. 096 :l' 0.089 (0.2.8) 0.0012 j: 0.006 (0.83) -0.2.4 ± 0.21 (0.24) 920~ 0.084

DG offerst

0..17 ±- 0.035 «0.001) 6.4 .L 3.61 (0.079) --'0.012 '" 0.15 (0.93) 0.0013 ± 0.008 (0.88) --'0.13 ::i: 0.31 (0.61) 336; 0.10

UG offerst

TPG offerst

0.098 + 0.035 (0.005) 10.,4 .L 2.67 «0.001) 0.16 ± 0.10 (0.13) -0.0056 ± 0.008 (0.43)

-0.24 ± 0.2 (0.37)

319; 0.14

0.11 ..,. 0.044 (0.044) 0.45 1" 3.06 (0.879) -2.25 ± 0.94 (0.017) 1.2 ± 0.25 «0.001) -1.0 ± 0.43 (0.019) 265; 0.10

'Clu~lerM robust 51i1odard errors adj!l51 for repealed ob.lervatiom 01 the same people, S96 individuals (clusters), Indicator varisbtes for each eIIperiment lUG, lPG) were included to rcntrol for di!leterice, .. among offers in the three game. (table 55), tRobu,t standard errors used, See (3l) for models with clUstered robust standard errors [clustering ON population. nable 1) to control for t~ e no oin depe nd en ce 01 individ uetsl, (;0 no n entst-l eve I ec n trc ls to adlj ress any sha red cut ture history (Model 1 i~ table; S B, $11, and S 14), and a !Iilr]et)' 0 J cth ~ r ro ti.u5ln ess {h&k3, whlen are summariZed in the text, tWR'i. an lndiVidu.Hevel dummy variable with "1" indicating Islam or (~rlltianity, and "(l" marking the practice 01 " tribal religion or "no religion." We have lew "no religioi1 ~ responses, and ethnography 5Ugg~}15 that in th~ wake of mis-;iimary actioitie. people associate "religioo" With "Christianity.' Thus, r€5PQose, of "no religioo·' p,ob;;bly imply belief in th~ 10 (a I Q r tradl!la nat rei i g ion,


19 MARCH 201'0 VOL 327 SCIENCE

leads to mean Ml accounting for .58% of the variation Cp = 0.76, 95'1.1 bootstrap Cl '" 0.44 to 0.95., P < 0.00 I). Dropping the Au instead lead,'! to mean M1 capturing 7! % of the variation (p = 0.84,95% bootstrap 0: = 0.59 to 0.96, P< 0.0(1),

On the punishment side, we looked for II relationship between our individual-level measures of punishment, from the ua and TPG. and our theoretically important variable, CS, while contl"o!ling for our other eight variables, For both the UO mid TPO, we reduced the vector of "punish or not punish" responses to a single 'number caned a minimum acceptable offer (MA.O). Thls is the lowest offer (:0:;50%) for which an ind-ividual no longer punishes. Thus, if a player accepts all offil";l, bis or her MAO is zero. If the player punisbes offers of 0 and 10% but accepts all higher offers, his or her NlAO is set at 20. MAO measures an individual's willingness. to punish low often; (32).

To analyze MADs, we used an ordered logistic regression (DLR) because MAOs are both discrete and birncdally distributed. This model assumes that the dependent variable is discrete and rank ordered. but that tbe distance between ranks is not meaningful, Because the oU!pU1 of <In OLR is nonintuitive (tables S20A and S:!3A), we have captured the effects of the highly significant (P < 0.001) coefficients for CS LTI Fig. 2, A au!;l B. Holding the other (';ight variables at their mean values, the figure shows 'how the distribution of TvlAOs for the TPG and VG shift with increases in CS. Small communities, the size of foraging bands, are the least willing 10 punish. A$ CS increases fiom 50 to 5000, there is a dramatic shift from a modill tvlAO of 0 to 50. Our communities' sizes range from 20 10 4600 people.

Theoretical arguments suggest that punishment (MAO) should be related more directly to the natura! logarithm of es, because tbe effeotiveness of reputarionalsystems decays in rough proportion to fhisvariable (26:, 37)- Consisteot With this, the natural logarithm of CS is a better predictor of l\l[AO in both experiments than CS itself (tables S20B to S258), and this amplifies the effects illustrated in Fig. :4 (fig, S4),

Our theoretical approach makes no, predictions about the relationship between punishment

and Mr. because, for small communities, there are oumerous repurational mechanisms (not involving costly punishment) that Cl:l1J sustain equality norms. Similarly, because religion should ga lvsnize the existing rnechanisms fur tlO!11.1 stabilization, whether participants ill a world religion punish more depends entirely on {be local stabilization mechanisms. which themselves depend on CS (3St Nevertheless, for M)l.Os to the TPG, we find that WR is associated wid] significantly (p < 0.05) more punishment, although Ml reveals no such relationship. None of the other predictors is. significant (P < 0,05, table S20A)" For MAO in the UO (table S23,A), income, wealth, and Ml all significantly predict lower MAOs (P < 0_05:), The effect for MI, however. is contingent on having CS in the equation. If CS is dropped. the effect of Mf becomes nonsignificant. Thus, unlike. CS, none of these oVler effects is consistent for both MAO measures or aCl-OSS alternative specifications.

Our analyses for both offers and MADs are robust to. a variety of checks, including alternative model specifications. and adjustments to our wealth and income variables to account for local differences in purchasing power (tables S6, S9, S12, SIS, S21,and 524). We also included contineutal-lcvel dummies (Africa, Oceania, South America" and Eurasia) to address concerns about shared cultural phylogenies, and we used clustered robust standard errors (clustering on site) to control for the potential nonlndependence of indivi dual 0 bservati ons wi thin our sites (tables. .S5 to S25)_ The findings for MI and CS are robust te these cheeks, However. when continental coutrols are applied, the effects of VVR on offers disappears because of the bighly U11(:"!Vetl distribution of populations containing individuals with "\Ill '" 0 (tables S5, S8, and S 11. Modell). We reran our models rot Africa only and generally reconfirmed the above findings (32). Nevertheless, because of tbc rather small sample of individuals with WR = 0, any conclusions about the effect of WR rnust remain quite tentative (39).

Discussion. Our results on the relationship between MT and offers extend the findings of our previous project (29)[n several ways, including using fresh samples, better mea-


Fig. 1. Mean DG offers for each p opula tio n plotted against mean value of MI. Error bars a re be otstra P ped sta nderd errors (bias corrected a nd accelerated) on the population mean.

HaO za

• Tsir'I'lane

SanquTan;ga •



o 11) 20 ,30 40 5060 70 BO 90 100 Ma rke,. Integ rati 0" (PWcelll"g" 01 P u !'<:.ha S'ed ,,~I orje5 in di el)



sures, new experiments, and additional controls. Despite swapping in LO new sites and using a different protocol, we hove replicated our earlier va findings at botb the level of individualsites (in fCfW- sites) and ino braining a positive relationship between UG offers and market integration with a better measure. This relarionship Is now demonstrated for DO .and TPG offers, each of which reveals similarly large effects and is robust to the inclusion of a f'dll_ge of socioeconomic variables not previoualy measured and toa suite of statistical checks (tables S5 to. S25).

These findings also delineate two additional lines of research" First, our inclusion of pruneipation in a world religion oonvcrges with other recent findings (21) and tentatively supports the notion that religion may have coevolved with comp lex so c i eti 6S to facilnate larger-scale interactions. Second, our analyses (tables SIS and 819) open up the question of wily, when a thirdparty punisher is added to a DG to create a IPG. do mean offers usually decrease, the predictive effects of WR disappear (T~ble 2), and economic variables emerge as potent correlates of offers (tables S 14 to S 16)1

On the punishment side,our analyses of MADs are eonsistern with the idea that as rep utatio nal systems break down in larger populations, increasing levels of diffuse costly punishment are required to sustain large harmo- 11 iou s comm un iti es. TI1is exten cis earl i er b ivariate analyses using population means in the IPa (411) and converges with ethnographic data, suggesting that large oommunities, H!:cking sufficient and effec .. tive punishing mechanisms, fiagment into smaller greups as. their community sizes reach about 300 people (4J).

Although our primary interest here is behavioral fairness, which includes whatever combination of motivations and expectations yields more equal divisions, two features of these results suggest that they may stand when considered as measures of'imriosic (internalized) rnotivations. Pirst, because 110 punishment opportunities exist in fue DO, this experiment likely provides a purer measure of intrinsic motivations for equal offers. Second, we reran tbe above offer regressions for UG and TPO offers including each population's meat] lVlAO as an additional predictor variable. If individuals are well calibrated to the threat of punishmern posed by tb e ir local communities, then ad ding this variable should (at least partially) remove the impact of any threat of'punishmenr 011 offers. The inclusion of moon MAO as a predictor variable does not change the coefficients 011 Ml or WR. (table S 17). Finally, anonymity concerns regarding our experiments and results are taken up ill the SOM text where we discuss our doubleblind treatments and related analyses which indicate that differences in perceived anonymity are unlikely to explain our major findings,

These findings. indicate that people living in small communities lacking market integration or world religions=-absences that likely charaetcr- SCIENCE VOL327 19 MARCH 2010


Fig. 2. Effect of (5 on MAOs. Each .s€'t of bars shOWS the distribution of MADs, tanging discretely from .0 1.0 so, for different (55. The coefficients used 1.0 crea te the plot a re fa r the variable (5 in all ordered 10 gistk regression containing all eight of our other variables. All variablesexcept (S a re held consta nt at t,hei r mean values. The (Ilefficients on (S represented graphically here are .0 . .052 ± .0.0085 (robust standard error, P < 0.001, n =; 227) for the TPG and 0.087 ± 0.011 (robust standard error, P <: 0.001, n = 297) for the UG. (A) TPG (see table 52DA for full reqression). (8) UG (see table 523A).


A 0.-50 .(;1



:l 11.70
0 0.60
:;; 0.111





ized all societies until the Holocene display relatively little. eonceru with faimessor punishing unfairness in transactions involving strangers or anonymous others. This result challenges the hyp othe sis that successful sod B] interaction 1 n large-scale socleties=-and the corresponding experimental finding~'-arise directly from an evolved psychology that mistakenly applies kin and reciprocity-based heuristics to strangers in vast populations (4, 5), without any of the "psychological wnrkarounds' (42) that are ereated by norms ancli!1'Jt(rutions, Moreover, jt is not clear how this, hypothesis can explain why we find so, much variation among pcpulatiousin om experimental measures and why this variation is so strongly related to MI, WR, and CS. The mere fact that the largest and most anonymous communities engage in substantially greater punishment relative to the smallest-scale societies, who punish very little, challenges this interpretation.

Methodologically, our findings suggest caution in interpreting behavioral experiments from tndustrialized populations as providing direct insights into human nature. Combining our findings with work on the links between behavioral experiments andrea! life (3f, 43,44) suggests that such experiments elicit norms, o:r reflectinstitutions, that have evolved 10 facilitate interactions among indlvidnals not engaged in durable re I ation sb ips. Given this, much current work using behavioral games appears to be studying the interaction between a particular set of norms and ow' evolved psyehology vnot tapping this psychology directly.

Overall; these flfl:djn~ lend SUpP011 to the idea that the evolution of societal compiexiry, especially as it has occurred over the last 10 millennia, involved the selective spread of those norms and instinnions that best facilitated the successful exchange and interaction in socioeoonornic spheres well beyond local networks of durable kin and reciprocity-based relationships. Although differences in environmental affurd;mces. probably had a profom,d ·impam on the emergeflce of (:OlllpJex societies across the globe (2), the rate-detennining skj) in 'Societal evolution ma;y bave involved ibe. ilssemb!yof the nounS a.nd instinrtions t11at arc

o Commun~.y Size 0' ~o

tI Commum~.y Si'~ 0' ~oo II COrTl m unity S iz e 01 2500


(l Oornrnu nit)' Size 01 SO

~ Cemmu nily Size 01 saO Il Cornrnu nity Siz.e n125(lQ • CommuniI}' Size of sooo

• COrTl m unity S iz.e ol5QOO

]! 0.20





10 20 3Q ~O

Minimum ·.Acceptable Offers (MAO) (Willingness 10 enga,ge in f"i.d party punisllment)


capable of harnessing and extending our evolved social psychology to acoorrunodate life in large, intcnscly cooperative communities, Recent experimental work among diverse industrialized populations suggests tbar the gradual honing of these normsand institutions continues in modern societies (31),

References and Notes

1. P. I, Ri(herson, R BQyd. R. t, Bellinger. Am, Ant/q. 66, 387 (2001).

2 _ ]. M. 0 i ~ m ond, GUM. Gefm~ ol/d Steel: The Fate:; oj Humop Societies (Norton, New Vo!1c,. 1997),

3. (. Gam~rer, Behavior Game Theory: experiments in SI.rategJc Intemctio« (Princeton Unlv. Press, P1in(eton, Nj, 20~3).

4. T. C. BUT!lh1lm, D. D. Johm~n, Anal. J(ritjk 27, 113 /200S).

5. /III, A. NGwa~., K M Page. K. Si g m un d. Sc fence 289, 1773 (2000).

6, D. (_ Nnrth, Instituboll.'; iflStitutianol Change, and Lconom«: ?f!f/onTlon(e (cambMdge Univ. Pili!;. New Yorli. 1990).

7. S. BDwl~5, Microeconomics: Behavior. II15titutI0m, and Eva/UtiDIl (Pri~wton U~iv. Pr~>, Prin~eton, NI. 2004).

8. E. ""~t, C. r. camerer, Tr>;nd" Cog(" sri. 11. 419 (2007)·

9. ~. Rirnerson, R Boyd, NOf by Gene, Alone: /'loW Culture WrlllS/armed Human Evolution lUFliv. of Chicago Press, Chicago, 2005).

10, ~. P,mthan~rh;3n, R Boyd, Na(ure432, 499 (2004,).

11. H, (jio~l. E. A. Smitll, S.BowI1'5, J. tbeor: Bioi .. 213, 103 12001).

12. J. Henrich, R. Boyd,1. (Mor. BiD/. 208. 79 (2001).

B. K. Sjgolund, , 1i!1U~r1;. M_ A. ~"/lak, PrO(. N(1rl. Awd,

Sci. U.s.A 98, 10757 W)01).

14. R, Boyd, p, J. Rich€r~Qn. l. Iheor. 8fo/.no;, 287 (2Q02),

15. J. E~5nl in ge r; J. ~i1ight. [u(r, Anthrt>po/. 38. 1 U ~ 9 7).

16. S, B{)wt~'r 1. h/Jn. lit. 36, 75 (~99aL

17. A. Smit~, rile Tlleory of Moral Sentiments (Prom(!th~ul.

New Ym~ 2000).

lB. It Wright .The Evoliltfoll oj God (Little •. Brown. Boston, 2Q09).

19. E_ F.~r. 1 Henri,h, in (je,,~lic ond Cvlwral I!~O/Jjtioll of Coopeatio«, P. Hammerstein, or!. (MIT Press, (;1mbridge, MA, 2003).

20. A. Sh~riff, A~ ~olenza\,an, 1 Henl'i~h .• in &O(utIO".

Cli/ture and the Human Mind. M. S~ballef.

A. ill Orgol"ya n, 5_ H ei~ e, T. Yo rna gu i~ hi. T !Cd mild a , Eds. (P>YL~oloI}Y Presl, New y(i[~, 2009), pp, 119-136.

2 L A ND{enlOy'm, It F. Shariff, Sd~l/(e 322. 58: (200S) 22. ]. Henrk~. Evol. Hum. Benov. 30, 244 (;W(9).

2~. P 1. Roes, M. Raymond, /!vq/. Hum. 8ehev. z'\, 126 120(3),

24. J. M~fGUI, K. V. Flannery. Pmc. Non Acad. 5ei. U.S.A. 101. 1825 7 (2004)

25. Ditlu><, pUlli,hmen! impl\~, thotthe responsibility for

pu nishing is Ipread over. 1;1 rge ~Eg mEn! .Qf thE populii (Ion.

26. K. Pi!ncfl~ no~li1 n, R. Boyd, J. Theor. SioL U4, 115 12003).

27. A. K, Oi ~ II' [aWll$nes~ and [COMmie!· Ai temrJIiVe Mod", of Govef/llmte (PM n[eluTt Univ. Pres>. Princ€to n, N], 20(4).

0 .. 30



la 2.0 30 40

MinImum Acceptable Offers (MAO) (Wlllingn~!<S \0 pUlllsh in thl! "llim~tUlI!)

?B. I. Henrirh ~t 01.. Ed, .• FOUl/do/fDI15 oj H!lmM Sociality:

Ecollomic Experimel1ts and Elhl11JgroplTic Evirjel1te from Fiftem Small·Stale Societies (O~furd Univ. Press, Oxford, 2004),

2'9. 1- H~n M ~h ef ~I, Btha v. Bmi n Sui. 28, 7 95, disc ussion 815, W)Osl.

~Q. I. lieonthet al .• Science 3\~, 176, (2006).

31.. B. Herrmenn, C. ThOni, S. Garhter, Sdente 319. 136, (20081.

32. Ba{kgrou~d, methods, and luppi€ID€flllli allalr>E'! ore ava ita b Ie a, 5 upporfi n 9 rna le~ at Q n Science On II ne,

D. f. F ehr, U. FhLhb. ehe), £voL num. Beimv, 2 5, 63 (201l4).

~4. Ga.n has value in 011. aur 500€rie,. Ewryone waHl, steel, satt, ~nd sug~r. if nothing else ..

35. K. O. V~h5, N. l. Mead. M. R. Goode, Sdel1(~ 314, 1154 (201l6).

36, We JoulJd no robust d[tf€remes among different Iormsol Olristlanlty ((;)toolic, II€fSU, Pr'ot€Sldn1'», or between Cnri.lianlly and i>la'n.

37 _ R F_ I. (a~(hD; R. v_ SDl~; R. i(o~iBf, Phys. R~v, [,69. 051915 {20041.

~g Traditional r.-elig[onl moy sometimelg~lv~nile 1(l(~1 norms (SOM lext), but they do net g~lVimizE "mar~et norms,' which are for i nlerat!jng with str:Ingerl·or ;3 n noyrnou) otli~r'_

Wh en religio~5 dQ uud ~rvl~te Il arms. t~ey InU,t Op~fate in conceit with t~e local Mmr-stabillZi n g mecha oisrns.

3~. We did not use lixed effect> for communiHes Of populaiiol15 i~ any of ou f ~""li;j1i 5. b~use l)Uf til ~o"'~(ill tG(u~ i) on tiifteren(e5 in ,0001 norm,. lie.cou5€ norm, ole group-Iei'd properties (different Ioc,;lo/, stable equilibnd), Oluch of the relevan t. norm-related va Ii anon occur> b erween, not wiUll n, wmmun i~, Of pop uIa~ no, (SOM textL

40. F, W. Marlowe et 01 .. rtoc. 8io/. Sci. n 5. 587 (2008). 41, A .. Forge, in Man, Se.ttlement and Urban/so/ian, P. Ucko ..

R. Trfnghilri1, G. Dimb@lby. Ed!. (OmRwo(th, London, 197n pp. 363~376

42. P. Rkherson, R. Boyd. Hum. Nat. 1(1,25j (1999), 4J, M. Poppe, J li'eon. PsyrhoL .26, 4 H (ZOOS).

44. M. (jurv~ n, I. Winking. Am. Ill/thropoL l1e, 17 9 (2008). 4.5. We espelially !h,m~ the communities ihat participated

in our research. WE also thank the National Sden(e Foundation, the Madlithur Norm, and Pr.~f"rence5

[II e !Work. the Mil x Pia nc k 1 Mli lUI e for Socia l I\nlfl(opology. and t~1' RU),ell Sag" foundation for funding this project, 1hank> also 10 the manyaudi€oce, il n d rE ~d ers who wntrtb uted to imp rov; n 9 our efforts,

Sl!pporting Online Material

WW\\I.5deonO:fm aq, orgf{g ~{O nteu t/fu IV3ll/5 9 72114 8 01 DC 1 Ma teri ~Is OJ n d Metho ds

SOM Text

Fig'. ,1 ill <;4

Tabl€! 51 to S2S


21 September 2009.; iI(~~tited Z2 Jonuary 2010 10, 11l61,{j ence.llS 2 B$


19 MARCH 201'0 VOL 327 SCIENCE

Design of Polymethine Dyes with Large Third-Order Optical Nonlinearities and Loss Fi.gures of Merit

Joel M'. Hales," Jonathan Matichak/ Stephen Barlow, Shino ahir.a, Kada Yesudas, jean-rue Bredas, Joseph W_. Perry,t Seth R. Marderi'

All-optical switching applications require materials with [arg!" third-order nonlineanties and

IQW nonlinear Qptkallo~ses. We pre~ent a design approach that involve5 enl'ltltlcing the rea!

part of the third-order polarizability (y) of cyanine-Iike molecules through incorporation of polarizable chalcogen atoms into terminal groups, while controlling the molecular length to obtain f.avorable one- and two-photon absorption resonances that lead to suitably Low optical loss and appreciable dispersion enhancement of the real part of 'Y' We implemented this strategy in a soluble bis(selenopyrylium) heptamethine dye that exhibits a real part of y that is exceptionally large throughout the wavelength filnge used for telecommunications, and an imaginary part of "'{, a measure of nonlinear loss, that is smaller by Iwo orders of magnitude. This combination is crlttcalin enabling low-power, high-contrast optical switching.

The development of organic chromophores for nonlinear optical (NLO) applications bas yielded advances in our understanding of how to optimize microscopic nonlinearities as a function of structure (J, 2), to use cbromophores with very large second-order polarizabilities for electro-optic switching (3), and to achieve very large twe-photon absorption (2PA) cross sections (4). In contrast; third-order NLO chrernephores witb t}10 characteristics necessary for low-power all-optical switching (AOS) in the telecommunications bands (1300 to 1550 run) remain elusive (5). With recent developments in silicon-organic hybrid waveguide geometries {6, 7}, which provide a robust.plsrform to make these applications feasible, me identification of optimal third-order N LO materials has become even more critical. Although structure-property guidelines have aided in. t110 development of mol" ecules witb reasonably large third-order polarizabJ I iti es (8), perhaps a greater obstacle is that most molecular materials examined to date exhibit unacceptable optical losses (9-JJ) resulting from linear and/or nonlinear absorption (in particular, 2PA) as well as from 'scattering, It is relatively straightforward to avoid linear losses by working at energies well below that of the lowest-lying one-photon absorption OPAl transition. However, addressing the issue of nonlinear loss presents much greater challenges, because 2PA can be. appreciable at telecornmuruoation wavelengths fur materials with large third-order optical nonlinear-

School or Chemistry and Biochemistry and Center tor Organic PhOlonicl and Electronics, Georgia Institute 01 techno logy, Atl3M<l, GA 30332, USA.

"TIJ ese a uth or S COil tri b utec eq ua Ily to t hi" work.

tT 0 whom correspondence ~holiid ~e addressad, ~-majl; seth, ma J d er@lhem i stry .qa tech ,ed u {S. R, M.J; j oe, perry@ e hemistry.g a te~h. edtl (].w. P.)

ities (10,. 11). Consequently, II challenge fur AOS is to identify materials that possess good 1\yOphoton figures of merit (FOMs) (9).. At the moleeular level, this. implies that IRe(,y)(Im("Y)1 > 12 (12), where Re{'Y) and IruCY} are the Te<J.1 and imaginary parts of the third-order polarizBbi!ity, respectively,

Although 2PA can be 1111nil11.i7..ed in a similar manner to lPA [i.e., one can work <It photon energies (flw) less than half the energy of the lowest-lying 2PA transition (7) (Fig. lA)), the magnitude of Re(y) at this wavelength cannot reach its full potential This is a consequence of iii large energy-detunlng term, A = (E - tiro), ihe reduction of which can lead to preresonant enhancement of 1. by '\VeU over an order of magnitude (13). Although the detuning energy cannot be reduced indefinitely, because resonam excitation would eventually cause large linear losses, an alternative approach (Fig. IE) involves operating at photon energies near the lowest-IY1Tlgl PA transjtjon, such that y can be preresonantly enhanced while twice the photon energy falls between the energies that would excite potentially detrimental 2PA This latter approach requires a more careful molecular design strategy to achieve favorably spaced states and sufficiently narrow absorption features to minimize losses. We describe a series of efficient NLO molecules that saris1}' tbese conditions and consequently possess way large nonresonant values of Re(y) and for one member of the 'series> a ratio of IR6('Y) that exceeds \00 ttu·ougbout the. telecommunications. spectral window from 1300 to L550 run.

The linear and nonlinear pelarizabilities of polyene and polyrnethme dyes have been shown to be highly correlated with the bond length alternation (I3LA) between adjacent C--C bonds in the conjugated chain {J} Tbe'third-orderpolarizability shows two positive maxnna as a function of BLA,

but the largest magnitudes arc associated with a negative peak ill Y at zero I3LA (1). Cyaniue dyes possess vanishingly small ELA (1), leading to y values that are dominated by the so-called Naerm in the simplified perturbation theory expression for Of U> 2, 12, 13), which suggests that they should possess the Illfgest magnitude of r for -a discrete polymethine molecule of a given chain lengih, Furthermore, this zero BLA and the assoelated broadened electronic poreutial energy 'surfilet: (4) allow cyanine-like dyes to be reasonably well described by simple ftee-eleetron theory (15), giving rise to values of M!!,,, and IIE'$ (the 1rans~tion dipole moment and energy difference between the ground and excited stare, respectively} that 10- crease linearly with increasing oorlJugation length. 1. (15), Given the strong d:pendence of'the N-term OD these values (YN DC.M geIE!e), Y is expected to exhibit a very steep power-law dependence on conjugation length. It bas been showuexperimentally that y evolves as t: where n = 7 to II, for different cyanine and oyanine-like systems (5, 16}, i111 plying that they are eff ct ent NLO chromophores, particularly so when focusing on memes such as nonlinearity per wit mass (17) or number of n-electrons (J8) ..

However, it has been shown both tbeoretically (J9) and experimentally (20) that cyaainetype dyes are subject to a Peierls-type symmetry breaking at long chain lengths. Such a distortion leads to large deviations of"BIA from zero, placing practical restrictions on the length of molecules in which one can achieve vet)' large values of y (11). Recently, we found thata bis(dioXaborine)terminated anionic nonamefhine (lJOB-9C, Fig. 2) resists symmetry breaking despite the reasonably long conjugation length and low-energy absorption (10). Consequently, this ehromophore exhibited

. . .' -n <; .1

sizable values of Re{y)(~ -5 x 10 - - em statvolr

or esu) while maintaining a reasonable ratio of IRe('y)!Im("I)1 (4 to 12) throughout the tclecornrmmicstions spectral region.

1! A

I! :8

Eu --p~,

Fig. 1. 5tate en~rgy schematics depicting conditions to avoid 2PA in organic chromophores, (A) Twice the exdtation photon energy (Zhw) is less than the 2PA state energy (£02). (B) lim lies below EOland 2firo lies between two-photon state energie$ (fo? and foi). SCIENCE VOL 327 19' MARCH 2010



Fig. 2. Selenopyryliu m- and bis(dioxa.borine)-termi nated polym ethi ne dyes dis(lJssed i n th i s study.








OOB~ l' .3,. ~-"

S1!--Kt n .. 0. ft = Rl. - a ..,

S!-K. n~ t ,,= R2. l( :: BAr _,

Se-1C: n 2 f'I: 'In. l( I.'

Table 1. Summary of linear and nonlinear optical properties of selenopyrylium polymethine chromoph ores shown in Fig. 2. linear and nonlinear optical properties were measured using solutions of !he chromophores in chloroform.

Molecule Methine units J...nm,,(nm) :Ema>< Mg. (0)" Reh) (e~u)t IRe(y)lIm(o,)11
(M~l cm~l)
Se-3C 3 750 3.'0'0 x 1'05 12.7 -2.8 x l'O-J1 3.9
Se-5C 5 946 3.6'0 'X 105 16.4 -2.3 x 10~32 95
Se-7C 7 1'081 3.9'0 x 10" 19.4 -2.2 x 1'0-31 190 'Obtai~~d from int~gralioo: ct me absorption spectra ~il"l debyes, 0) ~5M." = O_9584{j.: dl'IV"".l~., wlJere € i5in M~1 em l and v is in Ql'I ". t MeiJ'u rements performed using th e remto""w nd-pulsed Z:sc;,-n ted:! n ique at ) -e se '" 13 Q{) 11 m. fxpenmEli tsl uncertainties were estimated to be U 0%. Additfon al results lor tile spectral ra n ge 880 to 1940 n m are given in tlble 52,

"A " '·!UI
,._ I
llil I
• -c
• ... OOlJ '/I H
1M •
g- ilt
- 1!i

e.: I., .... -=-------""'"~ ..... ~-----,,_-"""'I'_


I.!~' I' III'


, tI ,.onu 11" I'I~,' I~SlI I,.·,

\ 1!\ ... II.,..:llIlnml



_· ··C

-- ,.

~. 1"lt.JJ'_'''''


\ ~,-

t, ; .._. '"

'~I liN I II~ I "UI 'u"

\\o:l.:!I1:th jllml

Fig. 3. Plots of the dispersion of IRe('l')/lm(y)1 for compounds listed in Fig. 2. (A) Values experimentally determined from femtosecond-pulsed Z-scan measurements (experimental uncertainties estimated .10 be J.lO%l. (8) Vall!es calculated from simplified perturbation the.Gry expression fQr y with experimentally determined·1PA and 2PA data. The shaded region in (B) highlights the spedral region shown in (Al.

Another route for a chieving large nonlinearities while avoiding symmetry breaking in poly. meth i11e8 is to adapt the energy terms to maximize Re(y) fur a fixed c-onjugation length, This involves identifying cyanine-type dyes wjth absorptions that occur at low energy (small Ege),even <It moderate chain lengths, and are very strong (large Mge), Although, to a first approximation, many polymethines meet these requirements, modification of the terminal groups in such systems can still profoundly affect the optical properties. Arrappropriare choice for a terminal group should provide good energy matching of terminal-group orbitals 10 those of both tbe highest occupied and lowest unoccupied 1< -orbitals of the polymethine bridge, as wellas sufficient orbital overlap such tbat sub,


stantial mixing with the bridge n-electron wave function is possible. Thus, the terminal group could (i) participate effectively in extending the overall conjugation length, and (i.i) increase the iransirion densities toward the p::riphety.ihereby enhancing the transition dipole while maintaining a bridge of modest length that could help to prevent die onset of synuneny breaking. Heterocyclic end groups ineorporating'heavier chaleogens could enhance these properties, in part because of tbeir modeeae electronegativities and high polarizabilities; indeed, such end groups have been found to produce very low energy absorptions. for agiven chain leagth (2h Accordingly .. we synthesized (J 1) selenopyrylhuntGtlnmatad polymethines with varying bridgc l8ngth, solubilizing 4-n-but)llphooyl groups and tetm:kis(3;5,

bis( triffuoromerby \)pbenyl)borate (BAr' 4) co unterions (Fig. 2), and characterized their third-order NLO properties ..

To assess the extent to which this choice of terminal group can afford large values of Re(y) at modest chain lengths, we compared the progressian of ~ and M PI' as a fimction of L for cyaninelike systems with various terminal units (fig. S5). Despite the variety of terminal groups sampled, the selenopyrylium-based systems were found to exhibit superior linear optical properties, particularly f01" Sh0l1 conjugation lengths. As shown in Table 1, the selenopyrylium-based dyes exhibit a strong bathochromic shift of their absorption bands with increasing length, typical of cyarrines, and possess both yel)' I arge 1J1lUSition dipol e rnomel1ls and low transition energies. These trends give no evidence of the onset of symmetry breaking, in contrast to other eyanine-like systems (20) with longer polymethine chains but less favorable h~. and Mw- values.

To further investigate the role Se plays ill the linear and NLO properties. of these systems, we performed quantum chemical calcalations [symmetry adapted cluster/symmetry adapted cluster configuration interaction (SAC/SAC-G) method] on model polyrnethines with selenium incorporated into the terminal pyrylium rings. Placing the polarizable cbalcogens in the 4- position of the heterocycle was found to result in appreciable electron density in the frontier molecular orbitals mfthe terminal groups (fig . S3), indicating a sizable electronic mixing of thechalccgen p-orbitals with the relevant molecular orbitals of the n-system, The calculated magnitudes of Ege !lDd Mgc are comparable to the corresponding experimental values, and their evolution with molecular length (three .. five. and seven rnethines) is also consistent with experiment. Finally. the calculated values of y (1..ero"ooqueocy Limit) scale in the expected manncr, given the progression of the calculated line ear optical properties and the dominance of the N-tenn discussed above, The observations (tahle SJ) are consistent with the notions outlined above for the use of selenopyry lium termin81 groups to increase y fur a fixed conjugation length.

We investigated the third-order ·NLO optical properties of the selenopyrylium-based dyes ill chloroform solution, usingthe ferrnosecond-pulsed Z,SC!lO technique [to deteni1ine-lm(,,() and Re(y)] and degenerate four-wave mixing CDFWM, to determine II'I and its temporal response) at several near-i.n.frared wavelengths. The nonlinear optical properties of the compeunds at 1300 mn are summarized to Table I. Note that Re(y) increases by all order of magnitude with the addition of each vinylenegroup, leading to an exceptioaally large negative value of 2.2 x 10-31 esu fur Se-7C. cal culations using the simplified expression fur 'Y and the linear optical properties ill Table 1 indicate that these very large negative contributions to R<l(y) can be predictedand result from strong preresooance enhancement (table S3). Despite

19 MARCH 201'0 VOL 327 SCIENCE

Fig. 4. Nondegenerate 2PA spectra faT solutions of compounds listed in Fig. 2 dissolved in chloroform: (A) Se-3C, (8) Se-5C, (C) Se-7C, and (D) D08-9C. Different pump wavele ngth s (a s indica ted in ea ch pa net) were used to observe the full NO- 2 PA spectra .. Solid drdes are degenerate 2PA (D-2PA) (rOS:; sections derived from femtosecond-pulsed I-scan measurements. 1 GM is defined as 1 x 10--50 cm4 s photon-1• Experimental uncertainties in the values were estimated to be. .Ll0%. Linea r absorption sp ectra are sh own as refere n ce s,

I ~

I :

this enhancement, the temporal response of y remains ultrafast as indicated by time-resolved DPWM measurements (fig. 87), a critical char" aeteristic for performance in AOS applications, These magnitudes of Re(y) compare \11'1)' favorably to those reported for other organic systems in this spectral region (8), including those for DOB-9C (10), polarized carotenoid" (13), and a series of donor-acceptor substituted molecules shown to be highly efficient as measured by nonlinemity per unit 111;:!SS (17). A bis(porphyrin-ethynyl) carboeation (14) possesses a comparable Re(y) ar 1.550 nrn; however, this is accompanied by a relatively }l00I' IRe(y)/lm{ 'Y)I ~ 4, due to large 2PA th III leads In appreciable loss. In contrast, Sc-5C and Se-7C exhibit exceptional values of IRe{i)! Im(y)I» .12 at 1300 run. Figure 3A illustrates !he dispersion ofIRe("(Ylm(,y)1 fl·Q111 1300 to 1550 nm for all the compounds shown In Fig. 2. Sc-5C exhibits a ratio at 1300 nm of nearly 100, whereas Se-1C possesses similarly large values but maintains this over the entire spectral region of interest (2.:>;. These ratios dwarf those of DOn-9C, a system previously reported to show promise ill terms of.good two-photon rOMs (10).

The two-photon FOMs indicate that detrimental 2PA can be avoided in selenupyrylium polymethines, presumably BS a result of energetics as shown in Fig. lB. To determine the positions of the two-photon resonance.'! in these polymethines, we measured nondegenerate 2PA spectra for all the dyes in solution (Fig. 4). A salient feature is that all spectra ex hibit. a favorable region over v"hich 2PA is relatively weak; this lowtoss windowappears to be-tunable simply by modi:Fyi I1g the position of lJQI {or; equiv.alendy. mol ecular

, f


, I.j





'11 ~"J ~'!I fUQ

TmJ'I5iliOll Wndan;th loml

- jff(;_

j ~Jl

rmJlJiliom ~·I\ ... h.!lWlh rrml


_' '_ . ~ I.. '"

J 1_ld

_, It ,,0

, ~

I >

r i


I r

_.400 g


'! .'"' ~

I ~ "_ r T" r If ..



~ I)O(I'! 100

1 fIIn~lliCl'n Wl:l\('lo:l'\(l~h Inm)

Length). These observations result from (I COI"i:J.bination of'cbaracteristics that is unique to cyanmelike dyes. As is typical for cyanines (26), the 2PA transitions lie at lJO! """ 1.1 x Eo)l and e03 ::::; 1 . .7 x EOb leading to a large energy gap between these resonances, The bauds are also reasonably narrow, likely because of (nearly) identical ground, and exeited-stare equilibrium geometries (14). 10 comparison, 2PA spectra can be substantially broader fur dipolar-type compounds that have lather different ground- and excited-state equilibrium geometries (27)- The streogtbs of the 2PA transitions in the polymethine systems are also modest. TheIower-lying transition (E02) is typieally relatively weak, as it is ascribed to vibronically assisted 2PA (16, 28) thataccesses a. Slate which would otherwise be 2PA-forbidden. The moderate activity of the higher-lying two-photon state (Eoo) is a consequence of relatively small values of M«: (<4 D; see fig. S8). The importance of this observation is apparent when contrasted to squaraine molecules, which bear a number a sirnilarities to cyanines but differ in possessing large values ofM/re· (>16 D) and therefore exhibit peak two-pho Ion absorptiviti es mare than 30 times th ose ofcyanines (29). USing these 2PA spectra-and the linear optical properties given in Table 1, we detennmed the dispersion of IRe(y)f[m(y)1 according to the simplified perturbation theory expression (12) (Fig. 3B). 'The theory accurately reproduces the experimentally determined values shown in Pig. 3Aand clearly illustrates both the broadness. (..,300 run) and tunability of the favorable twophoton FOM window.

Our strategy provides 11 route toappreciable enhancement of the third-order polarizability and

• I , ,

-:'00 .!~l <III)

·{'<l.'\..lll0'1 WI!~e'I!lfllllh 1111111

-~ l
_. ,I> 0
. "
.. -., two-photon FOMs of organic ehromophores fOI· AOS Materials with high ehrornophore density )hat are .a180 photochemically robust are required if such chromopbores are to be incorporared into devices for photonie applications. However. the molecular design flexibility afforded by cyaninelike chromophores should provide pathways to address this and other issues. Consequently, our results bave important implications for Cb8 design of a new generation of materials that could enable AOS with low power and high contrast.

References and Notes

1. 5. R. M<!rd~r 1'/ 01., Stience261., 186 (1993).

2. F. MeYff5, 5. R. Marder, B. M. Pierre, J. t, Bieda" J. Am.

Cltem; Soc 1~6, 10703 (;1.994).

3. T.-D. Kim et 01., j .. Am. Chem. Soc_ 1:29, 4Blt (ZOo/).

4. M. Rumi 1'/ al., J. Am. Chem. Sac. 122, 9500 (2000).

S. J.·l. BrEda" C. Adam, P. Iackx, A. rerSOOI1I, B. M .. PierCE.

Chern Rev. '.14, 243. (1994).

6. M. Horhberq e/ ai., Nat. .Ma/er. 5, 703 (2006).

7. c. Koo!i et 01., Nat. P/tQtol1ic, 3. 216 (2009).

8,. J. M. Hales., J. W. Perry, in Introduction 10 Orgollir ff.ecfroni, and OploeiectrOl/i( Material:; ond Devices, 5.·5 Sun,

L. Dalton, Ed,. (OK Pre.lS, Orlando, fl, 2003), pp, 521-579.

9. V. Mi,rilhi. K. W. Oelong, G. I. St~g"miln, M. A S~jfl, M. J. Andrejm, Opt. Lett. H, 1140 (1989).

10. 1. M, Hille" 5 Zhanq, 'i Barlow, S_ R. M;Hd~f. ] W, pe.tty, }. Am. Chern. Soc 12:8, 1136l (2006).

11. 5. PQly~kov, F. YOI~ino, M. Uu, G. Stegeman. Phys. Rev. ·8

69., 115421 (2004).

U.s,," supporting t(lil!l;naL on SCience Ohlin",.

13. T. Kogej el ol., C/lem. Phys. Le/t. 298, 1 (1998).

14. D. Q. Lu.G. H. (hen, J. W. Perry, W. A Goddilrd, j 11m.

them. 50(, 116, 10679 (1994).

U H. KUhn.)_ (hem. Phy,. 17. 11.'78 (949)-

16. J. P. Hermann, Opt. Commun. 12,,102 (1974).

17. I. c. Mily et al., Opt. lett. 10,3057 (2005). ta. M. G. Kuzy~, Op/. Lett. 25, 1183 (2000).

19. ]. S. (raw, J. R. Reim"'~, G. B. Ba(,~ay, 1\. T. Wong,. 11(. S. H us h, (/Ielll< Pilys, 167, 77 (1992). SCIENCE VOL 327 19' MARCH 2010



,2 ~ _ t. M, Tolbert, )j__ Zhao, 1, Am, Chem, ".,,-, 119, :3 ,2 53 (1997). 21_ (, S, Gorman., S. R, Marder, [h~m_ Mater 7, ,215 U 995,). 22. A_ CK Kafhk{)vll:1l't aI., Dyes Pigmtl,1t'i 5, 295 (1984). n _ 5_ R. M"rd~r et al, Science 276, 11'33 (1997l

24. K.]. Thorley, I. M. Hales, H. L. Anderson, 1. W. Pwy, AngeIV_ [hem. lilt Ed. 47, 7095 (2008).

2. 5 _ It l'Ial a 1.;0 Inuod tha r a 7,0:. f bon puLyme-rh in e wi tn thiQPYryI tum te rrmua I 9roup~ (S -7() exhi bited slightly lower nonlin ea fin"$ but co m para Me two-ph 0100 !'OMs to se- 7 ( fTa ble 53),

211. ]. Fu e/ QI., j. Opl. Soc. Am. B 24, 56 (2007).

.27_ L. Beverjn~ el at, 1_ Am, Chem_ .10(, 127, 7282 i200St

28, D_ 'it~efet et ol, Chem, Phy;_ 279, 179 (2DO~)_

29. S. I Chung EI at; }. Am, Chem, 50,_ 128, 14444 (lOOo).

3"0_ This m~tErial ts based on work ,upported in part by the Science and Technology centers Program oT NSf under agre~!lieo! DM~"012096 7, the DeJe~)e A(IV'm~",d

ReI earth Projects Ag en cy( DARPA) MO RP H Pro gl1l m and Office of N~val R")f!<l[.ch (N00014-04-009S and r.l00014"06-1·0897), and tM DARPA Z0t ~rog'a", (W31P4Q·09-1-001~)_ We thank W_ Hermann fur helpfUl dis(u!>ioos and corrcbcranve nonlinear optical

chara ct€ ri za n on mea sursrnents,

Supportillg Online Material

\WIIW _ <cien,emag . Drg/[gilmnt en tlfu iVsc:ienc<,_118 Sl17/D[1 Ma tl!rials ;3 n d Me tn 0 d5

SOM T€xt

Fig,. 51 to 58", S1 til S4


24 ill o\l~mbet 2 009; ~C(€ Pled 2 febf~ a ry 2010 Pu bli Ihed unl i ne 1 a Feb W' ry 201 0;

IO. 112Mcien(€.1 IB5117

indude this in form ation wh~n d t1 nq this. pa per.

Ferrous Polycrystalline Shape-Memory Alloy Showing HugeSuperelasticity

Shape-memory alloys, such as Ni-Ti and Cu-Zn"A[, show a reversible strain of more than several percent due to superelasticity, In partie ular, the Ni-Ti-based alloy, which' exhj bits some ducti t.ity and exce lie n t S upere lastic strai n, 'is th eon ly sepere la stic rna ten a I ava i La b le for pra cti ca l a ppli ca ti a n 5 at present. WI' herein describe a ferrous polycrystalline, high-strength, shape-memory alloy exhibiting a superelastic strain of rna [I' tha n BOlo, with a tensile strength above 1 gigapasca~ which is almost twice the maximum superelastic strain obtained in the Ni-Ti alloys. Furthermore, this ferrous alloy has a very large damping ~.;lpa('ity i1nd exhibil5 a l~rge reversible change in magnetization durin.9 loading and unloading. This ferrous shape-memory alloy has great potential as a high-damping and sensor material,

Elastic strain in metallic materials is USl)a.L1y limited to about 0.2"10" and application of stress over a cri ti cal value causes plastic deformation due to slip or twin boundary III Qti on , Many shape-memory alloys (SMAs.), such as Ni-Ti and Cu-2n-A1, show a large reversible strain of great~- than several percent due to "pseudoelasticity" or "superelastieity" (1). TIlls 5tr3it~ relies 011 the occurrence of a thermeelastic martensitic transformanon and arises from a combination of stress-induced rransformarion upon loading and its reverse rranstbrmation upon unloading, However, most polycrystalliue 'SlV1As that have all atomically ordered structure are extremely brittle. Exhibiting some duenl ity and a superelastic strain of 1110re than 7'Y"t the Ni- Ti based polyerystalline SMAs are used fur many products, such as cellular phone antennae, spectacle frames .. medical guidewires. and stems. However, Ni-Ti-based polycrystalline SMA specimens are easily fractured, for (0ample, by large defamations of over 300/" that occur aftercold rollmg without intennediete annealing, Furthermore, their ductility is [Jot suffieient fur most applications. Reduced productivity and the high cost of'fhbrication Caused by the low cold-workability impede further application ofthese materials to other 'fields.

Polyerystalllne B lloys, such as Fe-M n-S i (1, 3)_, Fe-Ni-C (4); and Fe-Ni-Co-Ti (5-8), have been

IDepartme!l1 01 Materials Sci@nce, Graduate School of Engi" l'leering, T ohoku Un iVel'5ity; Sendai 98 O-SS 7 9, ]apll n. 7.1 nstitute of Multidisciplinary Re,earel1 for Adv~n(~d Materials" Tohoku U niversi ty, S"endai 9 So-a 5 7 9, -I" pan.

"To Whom correspondence should be addressed. E·mail: kainuma@tageo_!


developed as "ferrous SMAs," which, because of their better workability and lower cost, are commercially more attractive than the Ni-Ti-based SMAs. The Pe-Mn-Si-based a lloys are now used fur pipe couplings and fishplates, One drawback of existing ru11l11S SMAs is that superelasticity can barely be obtained at room temperature, because their martensitie uansformations "{ [facecentered cubic (fce))/E [hexagonal close packed OlCp)] and ,//0.' (body-centered cubic (bee) or body-centered tetragonal (betj], are basically nornhennoelastic, Mak:i and others have SilOcceded in obtaining a thcrrnoelastic 't /0,' (bet) transformation in a pol yc ry sta lline F e- Ni -Co- Ti alloy by the precipitation of a :y'{Ni,Fc,CO)3 Ti phase. with a L12 structure (6). However, the superelastic strain obtained by a bending test at 240 K is only 0.7% (7), which is insufficient fat practical usee Itis known that Fe-Pd (9) and Fe-Pt (10) alloys exhibit a thermoelastic tcc/tace-cernered tetragonal (fet) transfonnati oil, but in spite 0 f 6Xtensive studies, no sopcrclasticity at room temperature bas been reported since the discovery of FecNi-Co~Tj alloy in 1984.

We present here a ferrous polycrystallme SMA ahowing a large superelastic strain of over 13% at mom temperature due to a thermoe lastic "110.' transfbrmation. This ferrous superelastlc alloy with a composition of Fe-28Ni-17Co~11.5A1"2.5Ta"O.05B [atomic percent (at%); Iiereaiier cotated as NCATB] is mechanically strengtheued by fine and coherent precipitates of'a r'~(Ni)./(l,Coh(A1, Ta) phase 'with a Ll2 structure due to agi n g (ti g. S I) and a strong recrystallization {035)<lOO> texture obtained by thermomechaniea) trearment, Here, the texture, intensity of the <JOO>r component in the rolling

direction is 25.4 (fig. S2), It is known that 'in the Fe-Ni-Al ternary alloy, the y' -phase with the L 12 srrueture coheren.tly precipitates in the y matrix by aging (It, ]2), as wellas in the Fe-Ni-Co-Ti alloy, although the yIn' marteositic transform B.tion in this system is not thcnnoelastic, Before the present study, we examined the effects of alloying elements and ofmierostructures conrrolled by various thermomecbanical treatments on the ductility and (he superelasticity for the Fe-Ni.-AI alloys. We have now found a combination of composition and microstructure suitable 10 ObtA,in the high level of supcrelasticity [supporting online material (SaM) text].

The lye li c tensile stress-strain (S S) curves obtained at room temperature for NCt\TB superelastic alloy are shown in Fig, 1, A to C (13). An SS curve. obtained from a commercial ).fi- Ti superelastie wire (Ni-49A at,% Ti) is also presemed in Fig. l C, Figure IE shows a series of optical micrographs taken from the NeATB alloy during 11 % tensile strain cycle (movie S I). The microsnucmral evolution indicates that stress-induced (/.' martensites appear and varus h during the strain cycle and that this pseudoelastic behavior is caused by stress-indu (led thermae lasti c Iran sformation For applied strains up to spproxnnarel;y 15%, the NCAl'B exhibits ahn.oSt complete superelasticity (Pig. IB), The maximum superelastic strain obtained is about 13.5%, which is approximately 20 times larger than that obtained in the I./e--Ni-Co-Ti alloy and almost twice that of the commercial Ni-l'l alloy. The NCAT13 alLoy shows a very high tensile strength of 1200 MPa.

Figure 2A shows an electrical resistivity curve for the NCATB superelestic alloy in the temperature region of the manenshic uansformatioo. The martensitic transformation strut temperature Ms and the reverse uansfomraticn finish temperature Af are 187 and 211 K, respectively. The thermal hysteresis of the transformation, defined as Til '" Ar- !y!gj is only 24 K .. which is Sib'll!fieantly smaller than that of nonthermoelastic transformations in steels and general ferrous alloys (- 400 K) but is comparable to that repeated in the Fe-Ni-Co-Ti alloy (8), which -bas a ihermoclastie transfermarion. 'The origin of the ihermcelestic transformatirm observed in the NCATB alloy can be understood on the basis of d;le arguments put forward by Maki and others for the Fe-Ni-Co-Ti alloy (1, 6; 14), namely, higb hardness (Vickers hardness = 402) large tetrag-

19 MARCH 201'0 VOL 327 SCIENCE

anality of the bet martensite phase (da = 1.11, where c and a are the lattice parameters for the c and a axes, respectively), and the partial atomic ordering of the parent phase because of coherent precipitates ofthcy'e(N~Fe,Coh(Al, Ta) ordered fcc phase (table SI).

[0 the present alloy, additions of Ta and Bare necessary fer obtaining excellent superelastic properties, Figure 2, Band C, shows the electrical resistivity curves for Pe-28Ni-17Co" II.5Al (NCA) and Fe-2 aNi " 17Co-l] .SAl"2.STa. (NeAT) alloys, respectively In the NCA quaternary alloy shown in Fig. 2J3, the "fir;!.' transformation is nouthermoelastic exhibiting a large thermal bysteresis greater than 5QO K. Therefore, superelasticity cannot be obtained, The appearance of the 000- thermoelastic transformation is a result of its low hardness and small tetragonality of martensite, which may be due to a small amount of the 1" precipitates because of its low y' solvus temper" sture (table SI). TIle addition of Ta, a strong y'forrning element, increases the volume fractioo of the y' phase and heightens the hardness and tetragunality of martensite (table Sl), which drastically changes the transforrnation behavior from nonthermoelastic to thermoelastic (Fig. 2C). However, in the Bvfree NeAT alloy, ~-NiAl ordered bcc phase precipitates along the grain

boundaries in the matrix phase (Fig. 2C, inset), and the ductility of the alloy rapidly decreases because of'intergranular fracture. The addition of a small amount of boron results in a drastic suppression of the undesirable 1'5 phase, although the precipitation of the ~. phase cannot be completely suppressed at high-angle boundaries (Fig. 2.A, inset). The ·NCATBalloy "vith a random texture, where most anile groin boundaries (about 90%) are composed of high-angie boundaries covered by the ~ phase, is still britt! e and fla ctures before showing superelasticity (Fig. lD)

It is known that sheet specimens with a strong texture have many low-energy grain boundaries, such as small-angle and coincidence grain boundaries U 5)_ In the NeATB alloy, the belioo of high-angle boundaries (about 40%) could be lowered by a strong ~ 035}< toe- recrystalization texture developed because of thennomechanical treatment (fig. 52). Consequently, improvement in the mechanical properties' was obtained by the suppression of the grain boundary precipitatiou of the ~ phase (Fig. lA, inset), Although the strong {035)<lOO> texture develops due to a similar thermornechanical treatment in the B- free NeAT alloy, the grain boundary precipitation of the .11 phase cannot be suppressed without the doping of boron.


~BOQ II, e 600 ~,



O~----~--~~~~~=====r~---~-~~'--~~Stnlln (%)

11400 1200,



Fig._ 1_ Tensile 5S curves at room temperature obtained in the NCATS alloy (A to C) with stronq {035}<100::e texture and (0) with random texture and [inset of (C)] commercial Ni-Ti superelast1c wire (Ni-49.4 at % Ti), The tensile directiWl is parallel to the rolling direction, that is, tilt' <100>.y dfrettion of the {035j<100> textured sheet (E) A series of optical micrographs taken from the NCATB superelastic alloy during the 11% tensile strain cycll:!, showing stress-induced martensite and its reverse transformation.



It is known "[hat both the shape memory and superelastic properties are also influen ced by cootrol of the texture, because the transformation strain is strongly dependent on the deformation direction in crystals U 6 J 9). The lattice pararneters of the y and a' phases of NeATE, as determined by x-ray diffraction. are ao '" 0.3604 om and a = 0.2771 nm, C = 0.3069 TI1n, respectively, producing a volume change due ID the transformation of +0.68%, Hence, the tensile superelastic strain expected in the <100>,( direction is simply evaluated from the lattice correspondence by


In the case of single crystal NCATE, the e~6if is 8.7%. According to theoretical calculation based on a phenomenological theory, the superelastic. strains expected in directions <liD>,. and <l l l>, are about 4.1 and 2.1 %, respectively.and the strain in the direction <I DO>,. bas. a maximum value (fig 83). The result for the <lOO>y direction, however, does not agree with the superelastic strain of 13.5% experimentally obtained in the rolling direction <100>'1' for the (03 5) <100> textured sheet (Fig. IC). One possible explanation for this discrepancy is the occurrence of a two-step transfbrmation, fcc. -7 bet (da '" I.11)'__" bee. If the bee martensite appearing in the final stage has the same molar ve lume as that of the bet marteosi te, th e I arti ce param eter of th e b co. p base IS a = 0.2867 nm and the Ef&f becomes 12.5%. whieh 1S almost comparable to the experimental

'I~--------~=-----~~~~ 1'0 Uil 01 0'




Fig. 2 .. Electrical resistance curves obtained from (A) NCATB sheet with strong {035}<100> texture, (B) NCA, and (C) NCAT sheets with random texture. The electrical resistivity ratio at .298 K is scaled as 1.0. The optical micrographs taken from the random and {035}<100> textured NCATH alloys are shawn in the inset ol IN, and that from the random textu red N CAT a lloy is in the inset of (C). SCIENCE VOL 327 19 MARCH 2010



value. Furthermore, the. gradient of the super-elastic curve ill the plateau region drastically c hanges a I a stra in of' tel {Fig. IB), and the strain region with the low gradient given by;,\em = 1001 !;ell! where 8<ijis the critical strain ofstress-induced transfermation, is about 8.5%. 111i8 value is roughly the same as the theoretical value fur the <1 00> y direction in the bet martensite. These results support the conjecture thar (he origin of the large superelastic strain is due to the two-~'tep transfbrmation. Th6 two-step transformati 00 caused by the appearance ofa secondmartensite phase, however, has not been observed, as shown in comparison between the micrographs 4 and 5 with 10.3 and 1.1,0% strains. respectively (Fig, !E and movie S 1). There is a possibility [hat 10 the second stage over the Eo,.] of about J 0%, the cia ratio of the bot rx' rnartensi te continuously decreases with applying stress, which yields the recoverable strain, but further work is required to con finn this.

In contrast to the Ni-Ti-based SMAs, the NCATB alloy has many physical properties besides the large superelastic strain that are useful for practical applications. One of them is an excellent mechanical damping property due to the large energy-absorption in the sup erelastic .cy de. 11:10 superelastic Ni-Ti allays have drawn attention as hlgh-damping materials with a recentering capacity due to theirsuperelastlcity, Because they absorb a large amount of energy during a superelastic cycle, where the absorbed energy corresponds to the area enclosed by hysteresis loop io the SS curve, they have been considered for seismic a ppli cati ens su ch as dam pers in buildings and bridges (20 22). The hysteresis in the SS curves of the NCAIB superelastic alloy is considerably larger than that of the Ni-Ti superelastic alloys (Fig. 1 AJ. Plotted in Fig. 3 is the energy absorbed by a superelastic cycle per unit volume a s a fun cti on of a ppli ed teosil e strain for the NCATS superelastic alloy, compared with those fbr conventional superelastic nonferrous polycrystalline SMAs V 7, 23). TIw, maximum energy


-- &~
.~ 80 ,_
~ m
~ -;.
~ GO E
• &Nor!
e c
• :8
'0 40
G ~
._ -
o· a
lit 20 . .:
.a Ill!
c( ftI
:I o~·.sC*==~~~~~ o 4 8 U $


Fig, 3. Energy absorbed by one superelastic cycle at room temperature asa function of applied tensile strain for NCATE, Ni-49.4 at.% Ti, Ni-46.4 at.% Tic61\lb (23), and (u-16.5 at.% Mn-9.2AHNi (17) superelastic polycrystelline SMAs.


absorbed by a complete superelastic cycle for the present furrous alloy with applied strains up to 15% is 81 MJ/m", which is more t}1110 twice as large as that in the Ni-Ti-Nb alloy, Which shows the largest energy absorption of all noefcrrous SMAs (38 MJ/m3 in 8% superelastic cycle) (23) Bod is also almost 5 times larger than that in the Ni-Ti alloy (1.6 Ml/m3 in 8% superelastic oycle),

The NCATa superelastic alloy also undergoe~ a large cban,ge in spontaneous magnetizatierr in, dueed by the martensltic transfhrmation. No ape parent change ill magnetization due to th erma J martensitic transformation was detected in the thermomagnetizatioucurves (ng. S4), because the transfonuation hardly proceeds duringcooling (Fig. 2A} Figure 4 shows the magnetizationversus-magnetic field curves at room temperature obtained under some. fixed tensile strains for an NeATH superelastic sheet specimen, whew the magnetic field W::lS applied in the direction perpendicular to fhe tensile direction. It is seen thatthe spontaneous magnetizarion drastically il1- creases with increasing applied srwin and reaches ! 40 electromagnetic units (emu)/g at a strain of 1,2%, which is about 3.5 times. larger than that before loading. TIle spontaneous magnetization decreases during unloading, The magnetization after unloading is almost the same as that befbre loading, although there is some residual strain perhaps que [Q slip defbrmation. Thus, llle stressinduced a' martensite has a spontsneous magneti. zation larger than 140 emu/g at room temperature, and the change in magnetization of the alloy in the superelastic cycle is reversible, whereas some bysteresis occurs between the loading and unloading ourves. This physical property can be used praoticallyas a noncontact sea In sensor covering a l:;Irge repeatable Strain of over 10% and is detectable ill the magnetizafion, If this alloy 1S used as a mechanical damper in constructions, one can nondestructivelyand noncoutactually monitor the strain in the damper through magnetization measurements,

This alloy also has excellent cold -workability, The polycrystalline specimens before aging tin:

110t tl'actu:red, even by a 99% reduction in cold rolling william intermediate anneafing, Therefore, various shapes, such as wires, rubes, and thin foils. can be obtained more easily eornpared with the Ni-Ti alloys fiaetured by deformation of over' 30%, and the costs for fabrication call be lowered because of the advantages otthe processing cost,

The NCATE -alloy-sheetwith a strong texture exhibits an excellent superelastic effect of more ihan 13%. This alloy shows several unique physical properties, such as a large superelastic bysreresis, a large .ehange in magnetization and electric resistance during loading, and an excellent ductilizy. Given these properties, the present ferrous superelastic alloy is" expected to be used for 1U8JJY practical applications such as superelastic materials, damping materials, and sensor materials in various fields>

References and Notes

1. K. 015 uka, C. M. Wa)'ma n, Ed! .. Shap~ Memory Mo f cr ia/J; (Combridg~ Vniv. Pre51, Cambridge, 1993).

2. A. 5010, ~. (hi,~[mil., K. Soma, r, Mor;, Acta Metal/. 30, 1177 (19821.

3. s Kaji\'l~(il, A. BilTuj, T. ~ikudli, 1'1. Shinya; Piar;. SP/~ 5053. 250 (2003),

4. S. Kajiwar.a, T. Kiku,~i, Acta Metall. 38, 847 (1990).

5. N .. N. iSanluLl. Y. N. KO!lal, V. V. Kokorin, I;'hysc Mel.

Mttnll. 47, 186 (19'19;1.

6. T. Maki, K, Kob;lya,hi, M, Minaro, I. tamura .. SCI: Me/a./[ ~8, 1105 (19S4).

7 IV. V. Kokorin, Yu, I. 5amw~ov, V. A .. Oiernenko,

0" M. 5hl!\'(henKa, Phyr" Mct, MetaU. 67, 202 (1.9a9). '8. T. Ka~"5hita ~t ol., Sa. Me/all. 19, 973 (19B5) .

9. T. 50hmuril, R Oshima, F. E. Fujita, Ser. Metoil. 14, ass


10. D. P. DUrne, ~. ~ W~)'m.n, MelllJL Trans; 4, l,37 (1973). 11.. o. HornlJogefl, W. M. Ma)'l!f, Ada Me/oil. is, 584 (196 n. 12. E. Homboqsn. W. M. Mayer, Z. Melallk. 58, 3l'Z (196 n. 13 Material<; and methods are available as ~upporting

mil te ria I on Science o nli ne,

14. S. KajilYaro, MalE(. Sci. lillg .. A 2H~275, 67 U 999).

15. T. W;1jilnabe, 1:1. Fujii, 11. Of~aw", I(; I. Mai, Acta Metoll, 31,941 (l98\i).

16. Y. C. Shu, K. B~at~(hary;i, Acta Me/all, 46, S4S 7 (~99B1.

17. Y. Slitau, T. Omori, It Ka in urn a, N. Ono, K. lshida,

Mllter. Itans: 33A, 2817 (;1002).

18 H. Sehltoglu ef ot., Acfl1 Me/oiL 411, 3311 (2000).

19. I. D~dd. ~t 0/, Mofer, Trans. l',lll" 2026 (200S).

20. I. M[Cormik, J. Tyber, R, D€sRo(he5, Ie. Gall, H.J. Mai€f, ). I!M. Mec/!. 133, 1019 (2Q07).

21. R. IJE'SRo mE'S, M. D!'lemM t, 6ng. :'ilrud!lres 24, 325 (2002).

22. L. Di >jlln\/, A. 5. Elnash,j, Ptoq. srrutt ling, M(lt'~r: 5, 60 (2003).

23 T. Takogl, Y Sutou, R. Kainwmil, K. Yarnauch i, IC. I>~[da, Ilppt Bromater. 768, 17 9 (2006).

24. The authors are grateful to K, R. liebeck a t Cambndge u~ lyersi ty lor his help i 0 crttica I ,eo din g. This work Wil s supported by G ra n t- in· A i dl from the J a pa nese S odety (Of the Prornntton "I ,oen~" G SPS), ill ew Energy and

10 dus trio I T €ch nolagy D~v€lop.lllen t Organ iza fio n (N WO), (~r~ Res€~rch For EvoIOt[o~"1 Scien(~ and T~,hnology

Ie RE sn, Japan 56 eacea nd T ~chn Dlogy Agency 11 ST),an d GIG ba I (OE Progf a m, T G ha~ u Uoivers·i IY, M EXT, I:'p~ n,

Sl.lpportillg Onlille Materi,,1 WWW.>Qen!{gileonrenllmllln7/5 9 n/14fr8[D( 1 Mater[~(s and Meth~d!

50M Text

Fig,. 51 to ,4

Tabl~ 51


Movie 51

12 Odab~r 2009; acceptad 25 lanua!y 2010 1 O.l1l6/sdene€ .1181 16'9

19 MARCH 201'0 VOL 327 SCIENCE

M~gnetlc FIeld (kat)

fig. 4. Magnetization turves examined at room temperature under some fixed tensile strains for NCATB superelastic alloy. Solid lines, loading process; dotted lines, unloading process.

Eliminating Turbulence in Spatially Intermittent Flows,

Bjorn Hof,l* Alberto de LOiar,l Marc Allil;),l Xiiloyun TII,lt Tobias M. Schneider2

Flows through pipes and 'Channels are the most common means to transport flt.Jids in practical

a p plica tiOflS. and eql,la Uy occur i n numerous n (lWril! sy<;tem~. In gene ra 1, the transfer of flui ds i~ energetically far more efficient if the motion is smooth and laminar because the friction losses are lower. However, even at moderate velocities pipe and rhanne] flows are sensitive to minute disturba nces, and in practice most flows a re eu rbulent, I nile ,ligating the motio nand spatial distribution of vortices, we uncovered an amplification mechanism that constantly feeds energy from the mean shear into turbulent eddies. At intermediate flow rates, a simple control mechanism suffices to i nlercept this energy tr anster by reduci ng inflection points in the velocity profile. Wh en activated, an imrnedi'ate collapse of turbulence t~ observed, <Jnd the flow relarninarizes.

At lOW. v. clocities, the. .. motion 01'11. uids is wellordered and takes the form of'parallel layers that progress downstream without mixing, Once flow rates are sufficiently large, 'however, this laminar flow gives way to highly disordered turbulent dynamics. The likelihood of a tlow to be turbulent can be estimated from Ole Reynolds (1) number Re = UD/v, which is a mea-sure ofthe ratio of'inertial to viscous forces [0 and D (the pipe diameter) are a typical velocity and length scale oftbe flow, respectively, and v is the kinematicviscosity of the fluid (2)}. Consequent. ly, large-scale flows with high velocities (such as wind gusts or ecean currents) are strongly turbulent, whereas flows at low velocities or on small scales are laminar. For turbulence 10 arise, an instability rnec ballism is required' that transfers kinetic energy from the mean flow into eddies and Cl'OSS flows. Forecnvective or rotating systems, such instabilities can often be directly deduced from the llneenzedgcvemrog equations. However, fur many shear flows (such as pipe, duct. or channel flows) [he sarne.appreaeh has failed to J?rcg,icI tbe experimentally observed nansinon to turbulence (3 5). It bas been shown that the transition process strongly depends on the magnitude of perturbations (6 9), but a detailed understanding of the underlying energy transfer mechanism is still 1 ackin g.

In natural and industrial processes, turbulence can have unfavorable and even detrimental effects, so means of suppressing the aocompanying erratic velocity and pressure fluenarions have importent practical implications. The eaergy dissipation mill skin friction of turbulent flows are much Ierger than those oflaminar ones, and consequently th e fluid transport is fa r ] ess efficieruif fl ows are turbulent. Particularly large pressure and shear stress e ucruations occur in the transitional regime ..

, M<1X Plan(K lnstituta for Dynamics and Self-Organisation, Bunsenstrasse 10. H073 G5tlingell. Germany. 'School of Engineering and Applied Sciences, Harvard ,Univ~r';j:y. 29 OdQfd Street, CiJmbridge. MA 0213 A, USA.

=Io whom rorr~~pol1d~nre should be addressed. ~-maH;

tPresent oddtess: lilVision GmbH. 31081 Gi:it~!1gen. Gennany.

in which the flow switches intennittently between turbulence and laminar. This erratic dynamical bee havior can lead to increased' structural vibrations (10) and the damaging of equipment; in thecase of arterial blood flows, this behavior is believed to be one cause for the growth and ultimately the rupture of'aneurysms (JJ, f 2)_ For shear flows, a number of active and passive control strategies (/3 18) have been suggested in the past in an effort to r-educe the turbulent drag toward its laminar value, Although many of the Concepts are Vel), promising and locally lead to a drug reducti011, the control costs are typically much higher than the energy gain, and to our knowledge none have led to a complete relarnmarization of turbulence in practice. Ei-videnve that such a transition from turbulent to laminar flow may even be achieved at minimal energy cast was found in recent studies V9, 20) that surprisingly show turbuleocecan decay after extremely long times. These observations suggest thatihe turbulent and laminar states may remain dynamically connected instead of being separated by a larg(J. poternial well as previously believed. The aim of this study is to extra (,1. the instability mechanism regenerating UlJ'bulent eddies in spatially intermittent flows and to force turbulence to decay by intercepting this 1.11 sobilJlis!TI_

In-smooth pipes ofcircular cross-section, turbulent flow strucrures witb an appreciable lifetime can first be observed at Reynolds numbers of -1700 (19, 2UJ. In this transitional regime. turbulence occurs in localized patches, or "puffs" (1l)', which have a leogtb of about 30 D and are adveeted downstreamat a fixed velocity (close to the mean flow velocity 0). In laboratory experiments, the velocity field of such puff$ call be recorded by using 'a stereoscopic particle linage velocirnetry (PIV) system (La Vision GmbH, G6ttingcn, Germany). A cross-sectional plane of the pipe (IiOITnaJ to the axial direction) is illumiuated by. a laser light sheet, and the fluid is seeded with spherical, oeti1TIllly buoyant particles (-'13 um in diameter), The measurement technique is identical to that described in (22, 23), and in the present setup all-three velocity compeneots could be obtained instantaneously at approximately 2000



points equally spaced over the cross section. The pipe has a total length of 12 m arid is made of I -m-Ioug segments of precision glass tubing with an inner: diameter of D = 30 (± 0.01) mm,

The specially designed pipe inlet uno the smooth connection between pipe pieces ensured that here, unlike in ordinary pipes, flows remained laminar for Re « 5000. This allows the creation ofturbulent puffs in a controlled manner by injecting ajet of water through a l-mm hole in the pipe wall into the fully developed laminar flow. ill order to ell sure that any initial transien rsh ave decayed, the developing puff was left to evolve fur ZOO tirneunits (P10) before PrY measurements were carried out at the corresponding downstream position, The axial velocity component at the pipe centerline (U,,) during the passage of a turbulent puff isshown in Fig. IA, Upstream of the turbulent puff [LID <; -5, where L is the axial distance (Fig. 1 A)], the flow is laminar and :fully developed with U" =20: At the upstream (rear) laminar turbulent interface (IJ D '" 0), the centerline velocity drops Very sharply Bnd. I,hen flucmares over a lengtb of 5 to )0 liD. Further downsneam (Fig. l A, left), as (be ilucb.Jations cease Uc gradually reams to, its laminar value «u: As shown in Fig. 1 A (inset), across this interface the velocity profile adju~1:.S from the upstream parabolic shape (shown in green) to the average turbulent profile (showa in red). The mismatch of the azimuthally averaged velocity pmfiles ill the laminar (green) and turbulent (red) region strongly distorts the profile at the interroce: The upstream laminar fluid has. to decelerate in the centra I part of the pipe and accelerate close to tile walt which gives rise to iafleetion points in the velecity profile (fig. lA, inset, blue profile). The latter tUltUIs Rayleigh's inflection point criterion (24), thus suggesting a hydrodynamic instability (25) which drives the turbulent dynamics.

At thelocation of the .::.1:l.'ongest inflection point, the centerline velocity shows a sedden drop (Fig. lA.}, whereas the turllll1e:n.tkinetic energy shows a sharp increase (Fig, I C), supporting the proposed hydrodynamic i11Stability. Close inspection reveals that stream wise vortices are generated at the same location (movie S I), In order to quantitY the 1110- tionand production of such streamwise rolls wnhin tile tu.rbulent puff; we considered the vorticity transport given by the cross-sectiona I average- of tile product oflhe magnitude of the axial vorticity and the motion in the axial direction relative to the! meal} velocity: (lol,:!(u~ - u) (Fig. IE). TIle zero crossing of this quantity [at UD = 0 (Fig. 1)] pinpoints the location of'vorticiry production: To the left ef'this point, vorncrty travels downstream relative to the mean tlow [(I@~l(t~L- U) > OJ, whereas to the right eddies in the near- wall region move upstream [( I@,I (ld-, - UJ> < 0] and penetrate and distort the ensuing laminar flow. In doing so, they sustain the infIettiOD points. which in tum cause the instability (located at UD = 0) that regenerates vouicity, Because the time evolution of the turbulent stru cture is not 1)CCt!Ss[b] e in the experiment. highly resolved dire{:.t numerical slmu- SCIENCE VOL 327 19' MARCH 2010



JatiOI1S 'were carried out in a long periodic domain (50 LID). This allows a detailed investigation of the instability mechanism and in particular enables time-averaging of the vorticity transport as well as the kinetic energy (Fig. lC). 111c strength of {he inflection point was determined from the curvature change of the profile (16). As shown in Fig. lC, the strongest inflection point indeed coincides with the location of'jhe vorticity production and a very sharp increase in the tUl'bulent kinetic <lnergy, which reaches a pea k ~2 D downstream. This confirms out proposition that turbuIeooe is indeed sustained by an inflection point instability, and the localized nature of thls process is su bstantiated by th e el ea r ex pon en tial decrease of the kine1;ic energy In the up- and downstream direction (Fig. l C, green cLU"Ve). Unlike the elongated turbulent eddies and streaks. the inflection point is spatially distinct, enabling a. local manipulation of turbulence that potentially can be carried out at mud] lower energy costs.

The currently believed mechanism (27-29) driving turbulent spots and P1JfI'S is based on socalled -parent-offSpring scenarios (30) for the vorticity production: Existing- vortices seed new ones by means Ofl'OUCUp or hairpin recreation processes. SOIIl6 vortices indeed roll up at the upstream imerface (UD < 0) and progress downstream, as suggested ill (27). The vast majority of vortices in this region (0 <])D < -5), however, are traveling upstream With respect to the mean flow (Fig, I G and 111,ov1e S I) and are therefore not created through roll-up but instead emerge directly from the inflection point. Inflectional instabilities have been suggested to play an important role in seven. II proposed turbulence regeneration mechanisms in higher Reynolds number flow'S (30 32). Although they differ in various details (mainly the vortex regeneration), they also share certain features with the mechanism described here. In all cases, streamwise vortices distort the velocity profile by moving slow fluid from the walt to the center (lift-up mechani Sill). gi viog cis e to inflecti onal1y un stable streaks. The susceptibility of laminar pipe flow to inflection points and the resulting transition to turbulence via azimuthal modes {the direct creation cfstreamwisevortices at an inflectional instability} bas been demonstrated "in ($3) and is in excellent agreement with the vorticity production observed in the present study.

To test the Pl'OP05Cd manipulation of the 10. calized instability, fully resolved computer simulations of a turbulent puff at Re "" 1900 have been carried out. The simulations are based on a spectral code (34) with more than 4 x IOfi degrees of'freedom tor a 50-D-long pipe. This resolution ensures that all scales are fully resolved at the Reynolds numbers .eonsidered and, as demonstrated with a similar numerical scheme, that excellent agreement is found with experiments (35). A localized fOl"cillg(26) co-moving with the puff was employed at the.rear turbuleot-laminar interface, 111e forcing locally reduces the downstream velocity in the central part ofthe pipe and increases that in the near-wall rcgio)) (keeping the


mean velocity and thus Re constant). Hence, the parabolic profile is distorted to a plug-like profile, reducing fhe sudden change of the axial Velocity across the roar oftbe puff. The forcing is active only over a section of'less than 7 D in length, and the maximum change in the center line velocity in this region is 15%, The vorticity of the puff before applying the forcing is shown in Fig, 2 (top). The magnitude of the vortices decreased significantly 30 time units later [time (I) = 80 in Fig. 2], and. at I = HO they almost completely disappeared, Equally, the turbulent kinetic cneJ;gy (Fig. 2, bottom) continuously decreases subsequent to the localized distortion of the velocity profile at. the laminar turbulent interface .. A fter 125 time units, the turbulence bas diminished beyond recovcry, and even when the forcing is switched off the flow continues to relaminarize.

Because in experiments a distortion of the protile at the turbulent laminar iuterfaee cannot be



as readily implemented as in the simulations, we chose a different method The velocity profile at tile rear end of the puff was distorted by inducing a second turbulent puff several D upstream of the original one (Fig. 3). Although the fluid in between the 1:\VO puffs was not turbulent, the short distance between them wasinsufficieot to allow a parabolic profile 10 fully develop. Directly upstream of thy leading puff [UD:::;: 6 (Fig. 3)J) the profile has a lower centerline velocity (U" < 20) thad does a fully developed laminar flow, Hence, the sudden change in the axi al velocity at the rear of the leading puff is reduced. As indicated by the dotted lines, the velocity changes. are equivalent to those found when the forcing is applied in the numericalsimulations. Comparing the azimuthally averaged velocity profile at the rear of the leading puff (UD:::: 0) with that of the upstream puff (UD;::;: 1 i) (Fig. :lB, left). it isevident that the inflection point at the rear of the downstream

~ .

z O'l...!II"""·-.,;-; .. - -t~··~-P~-

( I








fig .• 1. Instability mechanism. (A and !I) Turbulent puff Tll the experiment at Re = 2000. The flow direction is from right to left, (A) Centerline velocity during the passage of the puff. (Inset) Azimutha lIy averaged vel.oQty profiles at three positions .. Allhough the profile at LID > 0 has the typica I shape. of an averaged turbulent velocity field, close to LID = -1 a' stron 9 inflection point is observed, and further upstream (LID < 0) the laminar parabolidLow is quickly recovered .. (B) The vOrticity transport shows that for LID positive, vorticity moves downstream, whereas for LID negative vorticity moves upstream. At LID = o vorticity is created. (C') Time-averaged qua ntities for a n umerically simulated puff (Re = 1900). Positions a re given in the co-m ovi ng fra m e 0 f r efe re n ~e so tha t th e 10 cati 0 n 0 f the vorticity source (d ashed li n e) is fixed. The magnitude of the inflection point (26) is shown in: red, the vorticity transport in black, and the turbulent kinetic energy in green. The latter is plotted on a log-linear stale.

19 MARCH 201'0 VOL 327 SCIENCE

puff' has diminished. As in the numerical simulations, the instability ceases, and the fionr puff eventually decays, whereas the rear OM survives, The apparent dependence of spots of turbulence on adjacent laminar regions allows us to develop a control method fur the spatially intermittent flow regime: By applying a sufficiently strong pertur-

bation at a fixed location in the pipe, the laminar segments separating the turbulent spots are disturbed As a result, tbe fluid at this "control point' is uniformly turbulent. AI; the turbulent eddies travel downstream, they are followed by equally turbulent fluid and are Lacking any laminar turbulent interfaces in the vicinity where sufficiently



Ii: 0..01

Fig. Z. Control in the numerical simulations. An axially localized forcing is applied to the rear interface of a turbulent pulf so as to distort the laminar profile, (Top) Snapshots of streamwise Vorticity isocontours of th:e puff at time instants separated by 30 time units, (BaHam) Time-series of the turbulent energy of the puff. The red line corresponds to the puff when no eonfrol strategy 'is applied, whereas the black line shows the effect of applying the forcing (the circles correspond to the time instants when the, snapshots of the top panel have been taken), The forcing 15 introduced at t = 50, and its intensity is ex.ponentIaUy increased until it satura tes at about t = 100 (blue dashed line).

Fig. 3. Interaction of two turbulent puffs in the experiment (A,) The solid line indicates the centerline velocity of a pipe flow at Re = 2000. The two puffs are separated by LID = 6. The centerline velocity in between the puffs does not recover the value of the La rninar Hagen-Poiseuitle flow, The dotted lines indicate the centerline velocity at corresponding times in the nurnerira] !>imulations for a putt before (blue) and after (red) the localized forci ng has been applied. (8) (Left) The (azimuthaUyaveraged) velocity prof le at th e rea r of the up strsa m (sec 0 nd) puff (LID = 12) shows a stro n 9 i nflectio n po i n t, wh erea s the inflection point at the rear of the downstream (first) puff (UD = 6) is much weaker and strongly influenced by the presence of the upstream puff following it. Ev.entuaUy, the front puff disappea rs, whereas the rear one remains unchanged. (Right) Qualitatively, the same reduction of the inflection point is observed in the numerical simulation when the forcing has been applied, again resulting in the decay of turbulence.



strong inflection points could develop. Consequently, the energy transfer mechanism is subdued, which in tum leads to a steady decay of the turbulent intensity downstream of tbe control point.

In order to implement this control in the experiment, an intenruttent flow was created at Re = 2000 by the addition of a small perturbation to the pipe inlet that periodically generates turbulent puffs. In ordinary pipes Jacking the smooth inlet conditions of the laboratory setup, laminar -nabuleot intenninency occurs naturally at 1700 < Re < 2500, These experiments were carried out in a D = 10 (:-:0,0]) mm pipe of lengtb LID'" 600, which otherwise was identical to the setup described above. To visualize the fluid motion, tracer particles were added to tbe water- Turbulent regions, can be clearly distinguished from the. Laminar ones (Fig .. 4A). The time evolution of the intermittent flow was captured by sampling images 150 D downstream oftbe pipe entrance at 25 Hz. and the gray values of each image were averaged across the pipe radius RI1d plotted in a space-lime diagram (Fig, 4B} The (almost) vertical stripes in Fig. 48 constitute the periodic passage of turbulent puffs at this location (the sl ight tilt of the stripes results from the puff's advection velocity), A fulther' 100 D downstream, the flow could bel manipulated by the continuous injection and simultaneous withdrawal of water through two small holes in the ",;;111 (Fig, 4). In tbe initial phase of the experiment, before the control bad been actuated, turbulent puffs appeared periodically upstream (Fig. 4B) as well as downstream (Fig. 4C) of the injection point The control was triggered after 40 s and kept on for tlre remainder of the experiment. The space-time diagram constructed from images captured 150 D downstream of the injection point (Fig. 4C) reveals that initially turbulent puffs occur at the same rate as upstream of the control point. Shortly after the control is turned on. the fluid emerging, from the manipulated zone is uniformly laminar, as revealed by the disappearance of the stripes in Fig. 4C. The pressure difference required to pump fluid through the remainder of the pipe is reduced to the laminar value (Fig, 40), and the pressure flucnrarions fully disappear. A similar drag reduction is achieved when applying the same method to an intermittent flow in a channel (of height 11 = 4 mm and width W' = 120 rmn) and a square duct (H = 8 mm) in the intermittent regime (Fig, 40).. Because of its localized nature, the cost of this flow manipulation is. low: The energy gained limn. eliminating nubulencc is approximately five times larger than the cost. depending on Re and the details of the applied perturbation.

Although this vCIY simple strategy works well for suffici ently small Reynold s num b ers (R e < 2000 in pipes. Re < 1400 in channels, and Re < 1800 in ducts), it becomes less efficient as Re increases, and it fails once the regime of spatially expanding turbulence is reached (Re 2:: 2500 in pipes), In numerical simelations of pipe f1Qw, however, it is possible to apply a volume. forcing of larger amplitude along tbe entire 'pipe and turbulence still decays (fig, S3), The drag can be SCIENCE VOL 327 19 MARCH 2010



reduced by more than a factor of'two, suggesting that even at these larger Reynoldsnumbers turb ul ence is still driven by inflectional instab iliti es, In laboratory experiments, such a spatially extended volume forcing is unfortunately not readily available" and hence the necessary distortion of the flow profile cannot be easily unplememed in fully mrbulent flows. An exception .are electrically coo-

ducring fluids, such as liquid metals or plasmas, in which velocity profiles can be manipulated by magnetic fields,

We have identified a mechanism that ami" stantly converts energy from the mean shear into turbulent eddies, and these insights into the energy transfer have been exploited to eliminate turbulence in intermittent shear flows.


---~- ~~'

-- - -

- ---- ---



axral dfstance UP



0: 'Iio W :r: o


. I





o :: 80


40 I s



timels Channel fI,ow


Duct flow

Pipe flow


"11:11 " 1.2


Ll ~"'INI' I· .. ff


10 20 30 time /s

·1000 SO tlme/s

Fig. 4. Control in the experiment. (A) Flow visuallzationimage of an axial cross section ofa segment of the pipe in the intermittent regime at Re = 2000. The flow (ftom right to t~ffi is laminar on the left a nd turbulent on the right For the visualization, the water was seeded with 50~~tm particles (Mearlrnaid AA, ludwigshafen, Germilny). The particles are un-isotropic flat platelets that tend to align with the shear, resl)lting in a uniform light reflection when the flow is laminar and a patchy nonuniform one iothe presence of turbulence. Images were recorded (at a frame rate of .25 Hz) at two locations: one 100 D upstream and .the other 100 D downstream of the (ontrol point. (B and C) The laminar-turbulent intermittency in pipe flow is shown in space-time plots (B and Q, whereas the gray values of the flow visualization images (A) are averaged along the pipe radius. The ave raged gray values obtained for each flow~visualization image are then plotted as vertical lines, hence displaying the axial variation of !he flow in the vertical direction and the time variation horizontally. The space-rime plot resulting from the data sampled 100 D upstream of the control point is shown in (B) and that from data sampled 100 D downstream is shown in (C). Each occurrence of a turbulent puff in the observation section re5tjlts in a dark, almost vertical band in the figures. Upstream (B) of the control point, turbulent puffs appear at regular intervals (every-S s), Downstream (C) of the control point, the vertical bands and hence the turbulence completely disappear shortly after the control has been actuated at t = 40 s, The broad darker and lighter horizontal stripes in (B) and (el are due to intensity variations, (0) Pressure drop measured over -250 D in pipe flow (Re "" 2030) and ~ 250 H in plane channel flow (Re = 1400), and in a squarE' duct (Re ~ 1740) relative to the presSl]re drop in laminar flow.


References and Notes

1. 0, Reynold~i Pro«. R. Soc lond, '35, 84 (1811~) .

2. In our {ale, a Wa> chosen as the mean flow speed. and D WilS chosen depending DO th~ geometJ)' of

e 1 ther the diameter of th e pi pe O( the {ho nn e I (or duct) heig ht, I n ~ It ex peri men 1.>, Wa ter waS' us sd a, til e w'orkiOIl fluid and has a kinematic v1s{osizy 'of v .. 1 mm'ls at 20' C. Mea 0 velD d Ii es a t Re '" 2 000 a re therdor e

6.7 {mil 1 n the 30·mm pi pe, 25 (ml, in th e duct, and 'i 0 <mI, in the. rha nn~1 flow.

3. S. Grossmann, Rev. Mort phys. rz, 603 (ZOOO).

4. B_ Eckhardt, T, 'M, Schneider, S, HoI, ), West~rweeL Annu. Rev. Fluid MIlCh, 39, 447 j2007).

3. P. G. Dr"lin, w. H. Reid, /fydrixlynamic 'ilahi/ity (Glmilrldg< Univ. Press. Cambridge, UK, 1981).

6. A. G. Darbyshire, T. Mullin, j. Fluid Man. 269, 83 (1995).

7. A. A. Draad, G, D. C Kuiken, F. T. M. NieuW5tadt, J. Fluid Mech 377. 267 (1998).

'8. 6. H of. A. juel, T. Mullin, phy;. Rev. lett. 91, Z4 4 5 02 (ZOO 3). 9, F, Mellibovsky, A. Me,e9uer. Phy,. Fhlids 18, 074102 (2006).

10. S. Park. G. r. taU[J1le.1- Sound Vib{lI~ 319, 1067 (2009)

11. ~. M. Khanaln I. t. Bull. G. R. Upchurch jr., R. Berguer;

Anl1. VaK 5urg. 21, 67 (2007).

12. F. P. P. Ton et at., (omput stiuc. 87. 680 (2009). U. l. Simvi<;h, S. Karlsson, Nature 388,.1S3 (1997).

14. Y. pu G. E. Karni~d"ki" Science 28a, 1230 (2000).

15. R. Rathoasln9haITi~ K. S, Breuer, j, Fluid Mech 495. Z09 (2003).

16. J. H, M. Franllon, A. Talarnelli, L. Brandt, C. COISU, Phys Rev_ 1 ell. 96. 064 SOl (2006),

17. J. Kim, T. R. Bel'd.ey. AIln£I. Rev. Fluid Mrxh. 19, 3 83 (2007). lao M. H"9b~rg, T R. B",,;ley. D. S Henninqson, 1- Fluir.! Mech_ 4B1. 149 (Z003).

19. B. H of, J. W~st€rwee~ T. M. 5ch neider. B. o<:kh ardt, Nature 4 43. 59 (2006).

20. 6 .Hof, 1\, de low, D, J- Kl.!il\, I, Weslerw~e\. Phy,. /l~v. lett. 101. 214501 (ZOOS).

21, I. J- W.YgnaO.lki. F. H. Champagne, j, Fl!lid Moen. 59, 281 119]3.).

22. a. Hoi e; ol., 'i,jem;e 305, ~594 (20M),

23. B. Hoi, C. W. H. van Docrne, J. W€5terwee~

F. 1. M_ Ni~uwstadt, pny.,. Rev. lett. 95, 214 502 (2005).

24. L. R.ayltigh. rroc. tona. Math. Soc. 11, 57 (1880). 2..'i. Th;; in fl ;,tti on P Din I closest to th. w~lI a IsQ fulfiUI the sufficient imtobility rriterion for invisdd flows (Fjortoft's theorem),

26. MateMals and rnethcds are avaUable as 5upporting material an Si:i'ilir~ Onlin •.

27. P. R, liandyop~dhyay. j. Fluid Me[h. 163. ,439 (1986). ZlL (_ W, H. van Doome, L W~lteJWeel pllil. Trans. R 50(, A 367, 4.89 (2009).

29. M. Sfiimizu. 5, Kid:!., Fluid Dyn. RI;5 41. 0455G 1 (2009).

30. W. Schopp<!, F Hos>ain, J. Fluid Mech. 453. 57 (;?OOZl.

31. J. M, Hamilton, l- kim. F. Wale/fe, J_ FlUid Mech. 28], 317 (1995).

32. J. Jimenez, A, Pinelli,}. Fl!lid MI!~h. 389, 335 (1999).

33. M. I. Gavarini, A. Bottaro, F. T. M. Nieuwltadt,}. Fluid Mer:h. 517, Ul (2004).

34. A. Meseguer, r. Mellibov,ky, Appl. Numer. Moth. 57. ~ZO {20011,

35. M. A1/1la, A. Willi>, B. Hoi, J. Fluid Mech. 646, 127 (2010).

36,. We thank F. M.elUbovsky and A. Me,€guer for Iharing their spectral 'fld~. This re,earch was supported by the Max Planck 5oc:iely (B.H" A.dL •. M.A., and .x.T.)., Th~ Delltsch e fors(h(mgsg~mel nschatt und er gf anI S,h n 116711-1, Kavli lnstitut lor BiommQ S[i€n~e and Te,hnQlogy ~t Harvard, "Old Harvard M~t€ri;)(s Research and Engineering Center (r,M,S.).

Supportillg Onlille Material

I'/WIHc:iencemag . qrgkg iltDnt ell tlfu\V3 27/5 9., 2114 9V DC 1 Materials and Meti1od,'

Fig,. 51 to 53


Mol'i~ Sl

17 De, emb. r 2009 i acrep led 4 Februa ry 2010 10.1126/1 ci ence .1186 091


Microcavity Laser Oscillating Circuit-Based Resonator

In a

Christeph WaLther," Giacomo Scalari, Maria lnes Amann, Mattias Beck, Jerome Faist"

lasers bas§'d on micro(avities are extremeLy attractive for their compactness, Low power dissipation, and potential for ultrafast modulation speed. VIle describe an ultrasmall laser based on a

su bwavelength electronic in ductor-capacitor (lC) reso nant ci (cui! that allows for extre me confinement of the electric fie ld. This electricalLy injected lase r operates at a frequency of

1.5 terehertz, and the mode volume is strongly subwaveLength. The design concept of the

L( resonate r can be exten ded from the terahertz range to higherlrequen des and also

applied to detectors and modulators.

In mierocavines, the large electric field gencrated b~ vacuum flw: .. tua~ons enables the observation of'quennnn optics effects such as the Purcell effect (1); which is, in that case, the enhancement of'the spontaneous emission by the presence of a strongly confined optical mode (2, 3) or the strong light-matter coupling (4). A measure of the confinement is tbe effective mode volume Vefftaken by the energy density of the oscillating electric field. Lasers based on small cavities confined by dielectrics (5), such as vertical-cavity surface-emitting lasers (6), are limited to V"'I values larger than ...:2 (AJ2J1..n::l-:J, where Ie is tbewavelengtb, arid n"tt is the l.ffh,tive refractive index (7-11). Further reduction of the mode volume beyond these values can only be achieved with the use of metallic cavities. Progress has been made in the use of plasmonic cavities in a waveguide configuration (/2), witb plasmon resonances of nanowires (I3), OJ with resonances of nanoperticles (14, .15),.

We demonstrate that in the infrared range, th e large n egati ve va I tie of the di electric constant of metals enables the fabrication of lasers using cavities that borrow their concepts from C011VOOtional electronics and microwave technology (J 6). These cavities can be seen (Fig. lA) as consisting of electronic lumped circuit elements: itt our case, an inductor-capacitor (LC) resonant circuit (17. 18). Recently, rnetarnateria Is based on planar subwavelength lumped circuit resonators were used to design artificial magnetic and electric response in the terahertz range (l9), and even up to the near-infrared range (20). 111e gain medium [in our implementation, iii. terahertz quantum cascade (21, 22) active region] is inserted. between the plates of the capacitor. Amplification of the electric field in the capacitor by the optical gain medium allows the resonator losses to be overcome, thus leading to self-sustained laser oscillations. This approach is especially well suited for an electrically pumped active region biased by a voltage applied between the metallic capacitor plates, The LC laser enables an effective mode volume much smaller than the cube of the half-

I nstituts Io.r Quantum Haetrcnlcs, ETH Zurieh, WolfgangPauli·Strasse 16, 8Q93 Zurich, Switlerl~l1d.

"To. whom corresp on den (1" sh ou I d be ad dressed. E· rnai I: waUher@phys.elhuh (CW.); jerome.rai~@phy,.elh~.('h OJ.)

"I..tI ,. 5

F~uency (lHzl

Fig. 1. (A) Schematic of Ih~ LC laser. j·is the alternating current in the resonator,S is the induced 'magnetic field, and E is the electric field. The active gain medium is biased by the voltage source Voe. (8) Scanning electron micrograph picture of the LC laser device. (() Schematic ('fOSS ~ction throuqh the dewce along the sym metry axis. The red laye r is undo p ed A 10. 5Gao.sAs a nd prevents current i nJectio n into th e active regi 0 n below the bonding pad. (D and E) Finite-element simuLations of the electromaqnetic field in the resonator showing the dominating electric field component £, and the norm 01 the magnetic field Iii. (F) Measured reflectivity at 10 K of an array of 400 identical lC resonators, shown in the inset and rlesigned for a freqLlency of 1,45 1Hz, withoutgain medium and without electrical connection.

wave I en gth in the IT:I ateri al Vc>ff = 0.12(JJ2ljeffl and mila smaller than has previously been reported for other electrically pumped. cavities (9, 12,23). Moreover, a further reduction in size is possible by a mere resealing of tbe elements. The quality factor Q of the cavities investigated here is limited by ohmic losses. However, it is noteworthy that the radiative losses of resonant LC cavities will decrease as the volume of the cavity de-





creases (24), offering the possibility of a combination. of ultrasmall size and vel')' high Q for cavities based on superconductors (25).

A scanning electron micrograph image of our device (Pig. I B) shows that the resonator consists of two half-circular-shaped capacitors connected by a short line acting as an inductor, The geometry that we chose is a compromise between size. losses, and ease offabrtcation. To allow electrical pumping of the active medium, we selected 1'1 symmetric LC resonator that naturally defines a virtual ground for the resonance frequency in the middle of the inductor that is used for theconnectinn to the bonding pad A 8-pm-thick active region of a quantum cascade laser having the gain peaked at 1.5 THz (26) is inserted between the capacitor plates (Fig. I C). The fabrication of the LC laser IS similar to the fabrication tecbnique used for terahertz double-metal waveguides (27). The active material below tbe bonding pad is electrically isolated fiom the metal to avoid pumping of the latter. Blectromagnetic mode simulations are performed with the use of a commercial software package (2R). The electric field is found













12 SCI ENCE VOL 327 19 MARC H 201 0



-3: 10
~ 10G B
~ :t::::::
10 .£ '"
"' ~
~ u
1,5 0.5

1.0 Currant [mAl

Fig. 2. (A) Electrical (red CUlVe) and optical (bLue curve) characterization at 10 K of the l( laser. The output power is measured with an He-cooled bolometer. Black dots. represent the full width at haLt maximum of the emitted radiation, The emission spedra at the device is measured with an FTIR spectrometer, and the linewidth is obtained by a torentzian fit of the spectra, The green area indicates the resolution limit of 3 GHz of our FTlR spectrometer. (B) NormaLized measured spectra at various injection currents corresponding to points in (A). (el ElectricaL (red) and optical (blue) mei!surements at 4,2 K for an applied perpendicular magnetic field of 2,3 T, The differential reslstanl'.e (black curve) shows a la'rge discontinuity at the onset of the lasi ng threshold at 1.1 mAo

10 be mainly polarized in the z direction, perpendicular to the capacitors plates, as requi red by the selection rules of'the active .gain material. A pLot oftb:e electric fi~ld component Eo; in tbe resonaot mode on 0- section through the LC resonator <It belf-gain medium height (Fig. ID) 'shows that th e electric f e I do is antisymmetric with respect to the axis of .sYlnme!tY of the structure and is mostly concentrated and uniform under the capacitor plates. The magnetic field (Fig. 1 B),. on the other hand, is mainly concentrated around the inductor. The laser resonator shown ill Fi_g..lB is designed fur 1.45 THz (Ie "" 207 jim). We used room-temperature values fur the complex refractive index of gold (29) to compute a value ofthe conf n em en t factor of the el como fie ld en ergy in the active region r = 0.85 and a cold resonator Q01,,,, = 53, mainly caused by ohmlc losses in the gold layer. The simulated radiative losses lead to a relatively large radiative Q,"'d ""' l 89, which is ill good agreement with the upper limit of220 _given by Wheeler's formula (241'. The resonator is expected to exhibit a final Q '" 41, limited by the ohmic losses [supporting online material (SOM) text section S 1 J A Purcell factor of ! 7 is computed, quantifymg the enhancement of the spontaneous emission, and mainly limited by the noruadiative transition broadening of the sponta·060US emission linewidth. For the validation of the LC resonator concept, we measured the frcqcency-dependeot reflectivity of an array of identical LC resonators without gain medium or electrical connection. The observed absorption resonance ofrhe reflectivityat 1.45 Q = 20 at 10 I\. (Fig, IF) is in good agreement with


1.45 1.46

1.47 1.48

Frequency [THz]



., 2 .. 0 B ...
(J) Q)
!l "
-0 0
> 1.6 4 [J.. fig. 3. Lithographic tuning of the Laser frequency by changing the inductor le ngth of th e rese n ator,

0.0 0.2 0.4 0:6 0.8 1.0 U! 1.4

CUrrenl [rnA]

Wavenumbel (cITI"1)









1.5 frequency (THz)





E 2.5 g, 20


;g 1.5

> 1:0


0.0 '-"'--'LL.. __ _.__..ILl_.___.______..'----- ---'--_ ........



the mode simulation and fits well with a simple calculation based on microwave formulas, An inhomogeneous broadening due to. lithographic issues and coupling between resonators explain the lower measured Q of the LC array, compared with the single-resonator Q (SOM text S2).

TIle light and voltage versus current cbsracterisncs of a LC laser operated at 10K show a strong increase in the optical power observed at a current of -·1.20 mil., as the voltage reaches the value corresponding to the alignment of the structure, Wit11 a maximum detected peak power of -80 pW reached at 1.55 rnA (Fig. 2Aj. To charaeterizc the nature of the emission, we perfoniled a measurement of the device spectnrm as a function of increasing current. The spectra in Fig. 28 show a single emission line peaking at a frequency 'Of 1.477 Tl'lz, close to tbe value expeered by simulation of 1.45 THz. At currents

between 1 .01 and 1.10 mA, corresponding to the lowest detectable optical signal, the linewidth corresponds to Q = 1 J to 21, which is lower than the value computed for the cold cavity. This discrepancy is attributed to two. different factors:

First, during the alignment regime of the quannun structure Mar 'l ma, the active-region losses may be higher than at the threshold (30). Second. as (he transition occurs between two states with a clear spatial separation, the gain curve exhibits a strong Stark s.hift with the applied bias. At low biases, thegain peaks at 131Hz arid is therefore dearly detuned from the cavity resonance, which acts as a filter to further broaden the emission spectrum The observation of elecrroluminesceoce at such low current and frequency points toward a large value of the Purcell factor. As the current is increased, tbe.linewidth decreases steeply and falls below the resolution limit of our Fourier

19 MARCH 201'0 VOL 327 SCIENCE

transform infrared (FTIR) spectrometer, corresponding to a linewidth of-3 QHz at a eurrent of 1.4 1M and' a detected power of -7 pW_ As the current is increased beyond this 'Value, any further narrowing of the transition is hidden behind the limitation of ow- measurement system. The plot of the linewidrh versus I njected current comb. ned with the super-exponential increase of the optical power in Fig, 2,A strOUg!y suggests that the device is reaching the laser threshold regime near to 1.S mAo However, the gain in oursttuoture is not sufficient to operate the device dearly in the regime above the lasing threshold,

To access the operation in this regime, we have enhanced the gain of the active medium by immersing the device In a st;:roog magnetle field perpendicular to the quantum wells plane. The magnetic field breaks the ill-plane parabelie energy dispersion of the electronic subbands, and leads to the. formation of discrete Landau levels, allowing tlre crearien of quasi zero-dimensional states. The spacing of the Landau levels is equal to the cyclotron energy Eo = tteB/III"", where Ii is Planck's constant divided by 2r;:, B is the rna gnetlc tie ld, G' is th e e lectron charge, and m" is the effective mass of the electron. By varying the relative. ratio between the cyclotron energy and sub-band energy spacing, the electron dylJarnic,_~ and the subsequent lasing properties of a terabel17 quantum cascade laser are strongly affected (31). The optical characteristics of tile 1£ laser In a magnetic field of 2.3 T (Fig. 2C) show that, above the threshold, the power increases linearfy with thecurrent up to I6 oW until the electrical roll-over is reached. TIle onset of lasiog is aecouipauied by an abrupt decrease of'tbe differentia L resistance and is a clear demonstrationof'the lasing threshold (31). By measuring the threshol d current of a control laser with a cenvemional metal-metal waveguide, we estimate an enhancemeot of the gain due to the magnetic field by 5 to 10% at 2.3 T.

A number of resonant ea vities have been fabricated, continuously modifying the value of t}le inductor. As expected, the emission spectra (Fig. 3) shift correspondingly between 1.4 3 and 137 THz:. The spectra of the devices operating ill the edge of thega [0 curve are barely reBobing the threshold regime because they are strongly detuned from the maximum of the gain curve,

Our ultrasmall, electrically injected laser based on a resonator combines elements from microwave electronics with 11 gain medium. The strong confinement achievable with resonant LC cavities is of great interest for both applications and quantum optics studies. Our results can be seen as implementing gain in one cell of a LC-based metamatcrial (33). Because of its monomode out. put, low dissipation, and tunability, such a source could be integrated with hot-electron bolometers 10 create arrays ofhighly sensitive heterodyne receptors. for demanding applicaticns such as radio. astronomy (34), The strong coupling that isachieved in such resonators could also lead to, interesting applications, not only fur optical sources, but also

fbi" detectors and modulators, as the latter devices do not require very high Q factors and would benefit greatly -6:0111 the reduction of the mode volume brought by this resonator; VeIT can be roc dueed almost at long as t11<: reduction in the capacitance can be compensated by an increase of'the induotaace to keep the product constant, An attractive feature of the LC resonator is the spatial separariou ofrhe.elecrric and magoenc fields, simultaneously allowing extremely high eonfinement of the electric field and an efficient out-coupling to the enviromnent by the magnetic field. An LC resonator with a bent inductor could act as a magnetic 'loop antenna, which is effieiently coupled to both the vacuum and the matter, such as a quantum dot (SOM text S3). The design concept of the LC resonator can be extended from the terabertz r11l1ge to near-infrared frequencies (SOM text S4),

References and Notes

1. E. M, Purcell Pilys. R'ev; 69, 681 (1946).

2. }. Ger~rd at 0/., Plly,. R~v. Lett. 81, 1110 U99S).

3. S. Ogawa, M. lrnada, S. Yo~WnDt\l'. M O~~n<i. S. NQdil.

S(ience 305, 227 (20Q4)i pu~li5h~d online 3 IlJn~ 2004 ilO.11261>Cient€.109799S).

4. C. W B1sb~ d1., M. fIIi,hioka, A Ishikawa., Y. A r a kiIWiI,

.Phy', Re~. Leti. 69, 3314 (1992).

5. K. 1 Vahala NM('r~424, 839 (2003").

6. F. I(oyama, Lightwave Iedmol. 24,4502 (2006).

7. R. Cord 011, AA. B'oroditsky, K. W. Kim, Y. R~hmal-Sa'nii, E. Yablonovitd1, IEf Ptoc: Oploc-iedron. 145, 391 (1998).

8. O. Pointersl (1/., soence 284, L'l19 (1999).

9. H .. G. P·ark er ot; Sdence 305, 1444 (2004).

10. 1 Scheuer, w. Gleer\, G, DeRose, A. Yjlrtv, Appl. Phys,. len

86, 251101 (WOS).

11. K. IIlOlakj, T. Babi!. App/. I'h15. Lett. ss. 21H01 (2006),

12. M. Hill Bt at. Not Phatonlcs< 1 .• 589 (2007).

13. R. F. Oulton et ai., /!Jafll(e46l, 629 (2009),



1~ s, Kuh~. u. H;i~"ns:on, t, RagobEt~, V. Sondngndar. pliY\.

~fV_ ter: '97, 017402 (20~6).

15. M. A. 1Il0'liMV et (1/" Nature 460, 1110 (2.lJ09)

16. J. Slater, kev. Mod. Phys_ 18 .. 441 (J 946).

17. Th e lumped el em en! 9 r,ui t ~p pfaach ilua, p reposed a!

QP tico I Ire quen des \181.

18. I\l. Engheta, Science 3'17, 1698 (2007).

19. T.]. Yen ef ol., scienee 3()3, 1494 (2004).

20. s, I inden et ol.; Science 306, 13"\ (2004). 21, J. Fai.ll er at., Science 2M, 553 (1994).

22. R. KBhler et ol., N dl/ll~ 417, 156 (2002).

23. Y. Chol51~gnellx el 0/., AppL Phy,. tett: 90, 091113


24. H. Wheeler, IRE Trails. AnfennQ5 f'rapQg. 35, 1419 (1947),

25. A. Wollrafl e/ QI., Nature 4l:L. 162 (2PP.4).

26. C. Wa lther et ot., Appl. Ph~ Left. 91, 131122 (2007 J.

27. B. William" $. Kumar, G. Hu, j. R~nQ, Opt, f.xpre:;s B, n31 (;'Po,).

2.8-. II. ulti pnY,1c1 Mo del in 9 a nd 51 mu lati on Soflvm11 (COMSOl) W~'I used lur nur simulillian, (\'IP'/'oLmm).

29. M. Oldill, R. llelL R. Ale.>:andfr lr., L. long, M. Quell)',

API'/. 01'1. 26, 744 (1987).

30. G. SGl.l~1i e/ al., Laser and PhO/'OilifS Rev. 3, 45 (2009). 3t, G; Scalari eI (1/., I'/!~ RRY. lett 91, 237403 (2004) }2. C. Sirlori cr al; IEEE J. OUQlll!im Electroll. 34, 1722 (1998). ~a. M. Weg€ne.r el QI .• opr. Expre>516, 19785 (20PS!,

34. I. Gao st Q/., Appl. Phjls. IMt. 86, 24;1.104 ,(ZOOS).

3-'. We thank !he 5,'li5s [IIil.tian"( $cien<;e Fau~diltia~ and

the Nation']l (enter of C:ompften[e in Rese,Jr!:h

• Quo ntum PhQtoni cs' for S UPPOft an d L, ~evo u I o r RfY rcntri buti om to th ere fledivity mea sure men t. This \\,ork 1rIr.l.~ palt! y. GI tried out in the [erlter lor M iUG' and fIIa~ oteenn ology (F I RST) a r ElH Zu ri ch, J. F, wou Id I ike to atknowled g"" frl.!itlyl di scussi ens with r. YahlomlVitd1~nd A. Imamllglu.

Supporting Online MiI.terial

·l{w'II.')Lieocenlag . orgltgilro n lenUfu U/3 '2J 69 721149 5/ DCl Motena!!; and Methods

SOM Text

Fig~. sie 54


12 October 2.009; .tRpt~d 11 J.anuary 2010 10.1126rldence.~ 1831'67

The ClimaticSignatur'e of Incised River Meanders

Colin P. Stark/* Jonathan R. Barool!r/ Yukl'li S_ l"Ii:lyakawa,2 TSLlyoshi HaUanji,3 Niels Hoviu5,4 Hongey (hen,S Ching.-Weei Lin,6 Ming-Ji;line HOl'ng,1 Xu,"'~ Yukitoshi Fukahata10

climate controls la ndscape evolution, but quantitative signatures of climatic drivers have yet to be' found in topography on a broad scale. Here we describe how a topographic signature of typhoon rainfdll is recorded in the rneanderinq of incising mountain rivers in the western

North Pacific Spa tiall Y i;lvefi;lged river Sinuosity generated from digita I elevi1 tion data pea ks in the typhoon-dominated subtropics, where extreme rainfall and flood events are common, and decreases toward the equatorial tropics a nd mid-latitudes, where such extre rnes are ra re,

Once eli matte trends .are removed, the primary control on sinuosity is rock weakness. Our results indicate that the weakness of bedrock channel walls and their weakening by he·avy rainfall together modulate rates of meander propagation and sinuosity development in incising rivers.

generate sinuous. cbanaels (5-10, 13 20), Very broad channels (8, 9), bedrock-floored cut (strathi terraces (2,5, 6, 13,21), asymmetric valleys (7, 17 i 9), and even meander cut-off loops (8-10, 13, J 5-20). 111 particular, field observarious (8- i 0) and theoretics I oonsiderati ODS (21, 23) indicate tha~ th e maj ori ty of th e meanders and high sinuosities 'Seen in such bedrock rivers are actively generated by erosive lateral chan- SCIENCE VOL 327 19 MARCH 2010


III humid. rapidly eroding (l) montane envirotuncn:s sUdh as the Cascades. (2), Japan l3), and Taiwan (4), lateral erosion along the banks. and walls of bedrock-floored river dUU1~ nels (5-13) can be as important a process as vertical incision, Rates of lateral erosion can exceed down-cutting rates by over 10 times as a bedrock channel widens {l t, } 1) or meanders {5-1 0; ! 3, 14). With time, such horizontal cutting isknown to


nel motion (5-9), although there are certainly cases in which meandering forms arc inherited U 5) as channels maintain their sinuous shape during down-cutting (19,20).

Three key factors determine the rule of lateral channel erosion: (i) flood shear (5, 6., JO-J 2) and particle momentum flux (24) on the channel walls, which induce bedrock wear (12. 14, 25,26) and remove debris dumped from adiacent hill slopes; (ii) bedrock weakness (5,6, 9,25,26), which sets both the basic rate of flow-driven wear (2.)) and the instability of channel bedrock walls and hill slopes (27); and (iii) rainfall intensity and duration. which modulare bedrock instability and adjust the rate of wall erosion by rock falls and landslides (27). Transient alluvial cover of the

l lamont-Doherty Earth O~seIVatory of Columbia University, Palisades, New York 10964, USA. 'Center lor Spatial Information Science, University 01 Tokyo, 5+5 Kashiw~noh", Kashiwa, Giiba 277·8568, lapan. 'Gradtl.lte School of Life and Environmentill Sciences, University [If Isukuba, lbaraki 305· SS 72, Japan. "Department of Earth Sciences, Unlversity of Carnbridqe, Downing Street, Cambridge C82 3 EO, UK_ sDepartment of Geosciences, National Taiwan Unlv€fsity, No. 1, Sec. 4, Roosevelt ROod, Taipei 10617, Iaiwan. "Department 01 Earth So en (13, National Ch eng Ku n (j Un iversily. T aina n, Tai\lldn. 'Wat~r Resources Agency, Ministry (li Economic AIf~irs, Hsin-ri Road, -Taipei, T~iwan. 8Nation.1 lnstitute for Environmenta I Slu dies, 16-2 0 noqa wa, 1 s ukub.l 305-8506, lapa n, 9S ta te K€y La bo r atory of Water Resources and Hydropower Eng ineeri n 9 Scient e, Wuha n Universi II', Wu ha n 430072, China. IODi",>ler Prevention Research Ih>liM!!, Kyoto Uni· v€rsity, Gokasyo, Uji, Kyoto 611--0011, Japan.

'To whom correspondence should be addressed. E-mail: (sl~rk@ldeo,(o!(lmbia.~du

Fig. 1. (A) Dra~nage deli neation and si nuosity cal(ulation. (8) A znorn shows how each reach sinuosity s = ilL is the ratio of a Iong-strea m Ito the disfa nee the (rOW flie~ L between channel junctio ns, Regionaliz.ed sinuosity X is the weighted sum (28) of th e Ii nk si nuosities within the 100-km-diameter drde in (A) and {Q; X is g ridded at a resolution of 20 km. (D) Simplified bedrock geology of Japan based 0(1 a 1:200,000 synthesis carried out by the Geological Swvey of Japan. Large contiguous areas of QliiItemary sedi· rnents are i nd icated in gray, as are metamorphic rocks, which represent a very small fraction of tolal bedrock exposure; both a Ie excluded from sinuo.sity analysis. Regionalized si nuosity 'x. was mapped by deUn eating reach sinuosities on each lithology mask separately.


bedrock bed is also important, but only in the relative sense (14, 21} of reducing down-cutting rates willie leaving wall erosion rates unaffected.

These observations led us to pose two hypotheses: (i) The degree of sinuosity development in a network ofbedroek rivers may provide a quantifiable measure of the rate of latera I channel erosion; and (ii) catchment- scale averaged sinuosity in humid mountain rivers should correlate on a regional scale both with rock type and with rainfall and flood frequency and magnitude. Underpinning these hypotheses is the observation (J 9) (fig. 82) that a range of pbenomena-efluxes of debris from adjoining bill slopes, djfferential erosion into heterogeneous bedrock, big shifts in formative discharge through stream capture, catchment growth or shrinkage, and climate change, not to mention necking and loop eut-off-> act to disrupt meander growth and ultimately limit the development of sinuosity. Over time, competition between lateral channel erosion and meanderlimiting processes yields a spread of reach sinuosities acros-s a catchment whose average~ is expected to depend on bedrock erodibility and climatic conditions.

Previous support for the two hypotheses comes from studies of laterally mobile bedrock rivers in Japan (5 7), where flood shear and rock strength Were found to correlate well with horizontal erosion and lateral channel migration duringJO,OOO years of vertical incision. However, no studies of sinuosity have been attempted at a scale large enough to identify climatic trends,

chiefly because the necessary data have not been available in consistent digital form at a regional scale. With the recent availability of 3 -arc sec Shuttle Radar Topography Mission (SRTf\.1) dig. ital elevation model data for much of the globe, as well as digital geology maps and archives of meteorological and hydrologic data, this impediment no longer exists,

To test the two hypotheses, we quantified upland river sinuosity (Fig. I, A and B) in the western North Pacific using SRTM data. We initially focused on Japan because of its broad dynamic range of storm rainfall, We processed a seamless, high-resolution, vector geology map of Japan (3) (Fig, 1, C and D) to identify incising bedrock channels, segment them into reaches between stream junctions, and group them into five broad lithological classes (28) tbat span areas sufficient fur statistical analysis. Channel-reach sinuosities s for each of the five classes were then aggregated within IOO·k1n moving windows and converted into gridded maps of regionalized average sinuosity X- (28).

Next, we correlated X with daily rainfall statistics, generated from 20- to 3D-year records from a dense network of 1666 stations (Fig. 2A) across the archipelago (28), to assess the dual influences of lithology and climate on sinuosity on a subregional scale. Scaling up to the western North Pacific, we generated a single grid of sinuosity X. tor a group of lithologies chosen 6'0111 the Japan analysis for their similar apparent erodibility (Fig. 2B). Finally, we correlated the sinuosity







ftbic iUI!fJ Al~,ytl tilar~ VQI(;ipJl~

"'Iui-Q,nres SEuj'IT'IfI~I~

e, \
... \
'" 130'E

_. - ,





grid with typhoon-track (Fig. 3), stream-gauge, and rain-gauge data (Fig, 4) around the western Pacific Rim (2b~,

In the Japan analysis, spatial correlation oftbe sinuosity 'X, grid with station rainfall statistics [such as frequency of rainfall exceedance P rx of

Fig. Z. (A) Sinuosity X versus storminess expressed as the frequency Frx at which rainfall accumulation exceeds 50 mm in 24 hours; (8) PDFs of detrended sinuosity p(XlFrx = 10), generated by projecting each value of 'X(f!;Xl onto %(10) along lines with a common origin at 'X" = 1.16 (the estimated average for a:ll the data combined). Blue diamond" accretionary co mplex: green squares, nonaccretiona ry complex sedimentary rock; yellow triangles, plutonics; orange triangles, felsic an d alkali mafic volcanks; purple circles, norralkali mafic volcanics; dashed lines, li near regressions. Error bars are lcrSDs.

Fig. 3. (A) Regionalized si n uosity X (green/yellowl fed for range 1 to 15) a nd typhoon strike frequency (blue contour intervais of 0.25 per year) a (ross the western North Pacific. "The studied islands are shown as follows: (8) the Philippines (green), (0 Japan (dark blue), (D) Borneo (brown), (El Taiwan (purple), and (F) NEW Guinea (cyan). Analysis was restricted to highrelief areas on subduction complex lithologies chosen from the ]a pan a na lysis for si mila rity in their apparent erodibility; maskedout regions are shown in gray.

. c»


. o


50 mm/day (28)] revealed approximately linear trends (table SI) for almost all rock types (fig. 2A), with sinuosity X rising with increasing "storminess' from north to south. To isolate the dependence all rock type (Fig. 2), values of 'X were projected troin a common origin onto areference






I 1.35

1.15 0,0 uz o.~ 0.6 0,8 1.0 12 1 A

BI~e€danca Ireq dallV rain, » 50mm ~;. II09,J1IY)1



P,." =I 0 year'", and a probability density function [PDF) p{xJ was generated for each lithology class.

The PDPs show (Fig. 28) that the most sinuous incising rivers are found in accretionary complex lithologies (8,9). These rocks are predominantly low-strength, highly erodible (6, 8,9), silt and fine-sand turbidites with intermittent coarser, stronger units; channels cut into these rocks arc known for their bigh degree of meandering (8, 9). The least sinuous incising rivers are found in generally erosion-resistant (5), 1.10nalkali mafic volcanics, which include rocks ranging in strength from very strong, fresh basaltic lavas to weathered, unwelded, much weaker tuffs (5). The greatest regional variation in sinuosity X and its strongest correlation with F,"", are seen in the weak accretionary complex rocks, whereas the strong, nonalkali mafic volcanics show the least variation ill X and its weakest correlation w ith F rx (Fig. 2A and table S 1),

When combined with past studies of erosion rates and fractured rock strength in Japan (5, 6), ow' results (Fig. 2 and table S 1) indicate that rock erodibility (5, 6,9,25) is the primary cootrol on sinuosity development and incised meandering (8, 9). !::Ieavy rainfall fiequency or storminess exerts a dear secondary control on sinuosity development, particularly in weaker rock types.

Next we examined sinuosity correlations on a subhemisphenc scale. By combining the highresolution geology data in Japan with lowresolution digital geology for the western North

~ 400N

. o


:::, SCIENCE VOL 327 19 MARCH 2010



Pacific, we were able 10 restrict % mapping to a subset of Iifhologies (mafic and nonmafic volcanics; plutonics, and sedimentary rocks) with broadly similar sensitivity of sinuosity tostorm frequency Fo. (Fig. 2) •. We excluded geological formations dissimilar to those found in Japan (suchas those til southern Borneo), whose sinuosuy sensitivity was ntlt calibrated by the Japan analysis. The resulting clipping mask for this geological assemblage was used to prune the stream networks arid generate another map of sinuosity X spanning an area from the equatorial tropics to the mid-latitudes (Fig. 3).

'vVe correlated {Fig. 4) the regional sinuosity X grid with records of monthly mesn rainfall, daily river discharge data, and a spatia) PDF of typhoon strike frequency (Fig. 3) generated from tropical cyclone track records (28) , We found roughly linear trends in sinuosity with both rainfall variability, expressed as the coefflcieot of variation Rev [Fig. 4J\; Kendall's rank eorreiation coefficient (tk.) = 0.44, Pearson's correlation coefficient (r2) == GAOJ. and with relative flood intensity, expressed as the 99th percentile discharge non:nalized by catchment area Q99 fA, (Fig, 4B; 1k= 0.48, rJ"" 0.42). Sinuosity is highest in the monsconal, typhoou-prone northern Philippines, southern Japan, and Taiwan, where heavy rainfall frequency and floodintansity are highest

1:1 1.25
Q) -
"iii ..
0::: •
0 1.2
'" .-
IT: ~e,;' '".
1.15 0.5

1 RcvH


T ';;;.1.4


·iii 1.35

00 :>

.~ 1.3


~ 1r25

II! c

.~ 1.2


II 1 15


(1.3,4,29). Low-sinuosity livers are more COlTImonly found in mid-latitude northern Japan and in tbc equatorial tropics, where cstremely heavy storms are far less frequen t and flood van ability is much lower. This regional dependence on stonn climatology is confirmed by the correlation seen between sinuosity and typhoon strike frequency (Fig .. 4C; tl;. =" 0.50,1'2 = 0.45) <U1d is corroborated by the peaked relationship seen between sinuosity and Iaiinrde (Fig .. 4D).

Our combined results sbow that once account is taken of rock type, sinuosity records a measurable signature of storminess a key factor in climate-landscape dynam.ics (30). Conversely, our observations show that if the modulating effect of storminess is removed, sinuosity provides a proxy measure ror bedrock erodibility

111is dependence of sinuosity X on bedrock erodibility (Fig, 2), as well as the ubiquitous association of'incised meanders with weak bedrock. (5, 6, 8 10, 13, 10), indicates tbat horizontal channel erosion rates are much more sensitive 10 greater rock erodibility than are vertical channel erosion rates. This asymmetry 1 i es in th e dua I coutrol of lateral channel motion by both flowdriven wear of the. channel wall (12,24,26) and mass-wasting erosion of'the adjoining hill slope, versus the single control on vertical incision by flow-driven wear of the bed alone (/4. 25). Bed

sediment cover augments the asymmetry (J 4) but is not its cause.

The climatic signature expressed in sinuosity comes from an amplific-ation of this. asymmetric sensitivity. More frequent heavy rainfall means thai channel walls and adjoining hill slopes are destabilized more often and are more prone to lateral erosion (27); if· also means an increase in flood frequency, flow shear, and sediment rnomentum flux On both channel walls and boos. The combined effect is to raise wall erosion rates mote than bed erosion rates and to accelerate sinuosity development (fig. S2). In the western North Pacinc, this phenomenon ultimately leads to a con-elation between typhoon' strike frequency and in cised chan ne I sin lJOS i ty,

The strength of the observed climatic signature underlines the strong coupling between climate and erosion in these island mountain landscapes {l, 4, 29, 31f} and tbe extent to which their morphology reflects recent [late Pteistoccne to present (5, 6. 13)] climatic conditions. This coupling is much weaker where erosion rates are much lower, and therefore sinuosity development in suoh errvirorunents [suoh as the Colorado Plateau (20)] will reflecta much longer-rerm integration of'weather patterns.

Even for the western NOl1:h Pacific there is significant disparity between the. brief time span of available rneteorologiral and hydrologic data [-20 to 60 years (28)] versus the time scale over which bedrock meanders have evolved I-I 000 to -100,000 years (5, 6, J 3)]. We nevertheless have confidence. !hat our corre lations with storminess have meaning, because the regional pattern of storminess is so strongly tied to the average distribution of typhoon tracks (FIgs. 3 and 4). The shape ofthis distribution is determined by Ihe location of tropic a] cyclogenesis (currently between 8"N and 15~N in the western North Pacific) and by the steering effect of the largescale circulation of the atmosphere (3/). Both undoubtedly change on interannual to multimilIenial time scales, but it is reasonable to infer that the broad latitudinal trend in typhoon strikes, with its peak in the sabtropics (Fig. 40), has remained roughly constant since the current platetectonic configuration of the western North Pacific was established several million years ago.

The studied landscapes and their high sediment fluxes also exhibit a strong coupling with tectonics (1, 4,30), but our results suggest that rock uplift rare is a much less plausible explarraiQ1Y variable for sinuosity X than climate. Sinuosities are highest io lhe· north em Philippines (Luzon; Figs. 3 and 4) and second highest in Taiwan, which is. consistent with our observation that typhoon strike frequ~mcies and storminess measures are highest in Luzon and.second highest in Taiwan, bat is inconsisteot w ith the relati vely low late Pleistocene-Holocene uplift rates seen in Luzon (-1 mrn/ (32) as compared with those in Taiwan [-·6 inm/year (4, 30, 33)]. Similarly the north-soutbtreod in increasing sinuosity seen itt Japan (Fig. 2A) is not expressed in the regional

19 MARCH 201'0 VOL 327 SCIENCE

£ 1.3
"U 1.25
a 1.2
II •

:::c 1.35 B


• ..

• • • •

w..- .

- .



46 8

DejA [mJs) X 10-7

~ 1.4
z- 1,35
<: 1.3
Ql 1.25
0 1.2
II 1.15
iOS 0 i


pattern Qf uplift rates inferred from marine isotope stage 5.5 marine terrace elevations (33). If uplift rate does playa role, it may be in modulating channel steepness as a fimctiou of sinuosity (fig. S3).

Ri?ferences and Noles

1. 1. D. Mmimao,]. p_ M. S','Iillki, J. Grot • .100, 52'S (19921.

2. F.). Pauilglia, M. T. Bral1don, 11m.}. Sd. ]01,385 (201)!).

3. Y_ S_ l1il)fahIVa, T_ OgUchi, (ieoff/Qrpho/(Jg!l 1,11, U


4. 5. 1. Dadson es vi,., Nature 426, 648 (2003).

5. T. 5~,uk!, Trans. lpn. Ge~morphQL Unio!! 3-,1, 1 (1982). 6 _ T _ Suzuki, rran~_ jpn, GeIJmorphal_ Union 4- t 33 (19 B 3).

7. M. H iran 0, frans, ]pl1.Goomorphol, Union 1-2, 11'1 (1931).

8. R. Yagi, H, lked;" Bull, Envirol1. Re". CentEr ,univ. Ttfltk!lbo 22, 1 (1997).

9 A, lII"k~nil, H. I~"da, BulL EnvlrQfL ReS, Center UIi/v, TJ;!lktloa 2.4, 1 (1999).

w. 1. R, B~rbour Sf 0/., Geoph}'5. Res. u« 36, L04401 (20()9). u. c. Po Stark, Geop_hys. Reo:. Lett. 33, 1:04402 (2005)_

J2_ ], M- Turowski, D_ Logue, N, Hovius J_Gf!fJphY!i. Re:;, ~14 ~F3), fD3016 '(2009).

n. ]. B. H_ Shyu, K, Sieh, ].-P_ MOUilC, W.-S_ Chen. 'Yo-G. Chen, J Geop"~_ R5. n~ (BB), B08403 (2006).

14. ], M. TUtol'l,kj" [), t~9ue, III. Ho\illJs, J- Geaphyo_ Res: 1.12,

(F4l, f04006 (;:1007),

15. W, M. Davi" Science 2:1" 225 (1B93).

16. A. Winslow, Science 22, n (1891).

17 _ R. H_ M,h~r~, J GEomorp~al. 5, 12 (1942)

IS. H. Bia"k, GeQI, Soc: Am. BuILS1, 3135 (1970L 19_ G, H. Oury, Am 1- S<'f_ 25~ J93 (195~),

20. 0, Ii, Holden, (j_~o~ Soc Am. BulL 102, 233 (1990).

n. G, S. Hampel:, R. 5. An derson, Gfol. Soc, A ff/. Bull, 1t4, 1131 (2002)_

22_ 5, Iked"_ G. Porker, K, Silwai,]. FW/d Meeh, 112,363


21. G_ Semina'". J, Fluid Mech. 554, 271 (2006),

24. ), G. A Bitter, Wror 6, 169 {19631,

25. L S. Sklar, W_ E, Dietrich, Geology 29, ~D87 (2001.)_

26. K. Hartshorn, ~. Hovius, W. B. D;Jde, R, L_ SlingN\~nd, Science 297, 2036 (2002),



27_ F_ G~;:zetti, S_ P~fllcracri, M_ Rossi, r P_ Stark, Mel~orQL

At1)105. P~'98" 23-9 12001).

28. Sl;~ supportinq material on 5tif:l1ce Online.

29. I. Gale\\-~lty el ot., j. Geoph% lie;. 111 (8), F0301~ (2006).

30. S. D. Willett e/ al., fd,'" GSA Spec Pop_ 398, Penrose

(Ollj, Ser. (2006)

31. ). c. L. Ch~n, Aonu. Rev_ FlUid Mffh_ 37, 99 (2005). 3i. v. Ma~da et oi. QlIal. 11)1. 115-116.1.5 (2004). rs, Y. Ota, M, Yamaguchi, Qua( Int. 120, IDS (2004),

34. This ,Iudy l'Ia, supported by IIII\SA , NSf (Earth soences Division, GeomorpnDlogy and tand-use Dynamkl Progriln1),and Ille TaiWan National Science Council. W"- tha~'" '(II E:, o ie:lli[n , G_ Pa(~er, and

E;, E. Wahl tor Imljghlfning diml,-sion.! and C.·H. len,

(_ Huanq, T,-C, Yi, dod M. T~jik"r~ lor help io the field.

Supporting Online Material

IWJW .>{gil (0 nten !flu un 2 715 9 72114971 DC 1 Me!lind,

fig,. 51 to 53 Table 51

'9 III ove m her 2009; ~ ccepted S Fe brua ry 2010 10.112615 d ence .11M 4 06

Transition-State Spectroscopy of Partial Wave Resonances in the F + HD Reaction

W~!ltui Dong,* Chuillei )(jao;* Tao Wang, Dongxu Dai, Xueming Yang, t DO)1g H, l.hangt

Partial wave resonances, quasi-bound resonance states with well-defined rotation in the

transitie n state region ofa chemical reaction, playa govl'rni n 9 role in reactio n dyrra rnics but have eluded direct experimental characterizatio[J. Here, we report the observation of fndividual

partial wave resolved resonances in the F + HD - HF + D reaction by measuring the collision energy-dependent, angle- and state-resolved differentiaL cross section with extremely high resolution, providing a spectroscopic probe to the transition state of F + HD - HF- + D. The agreement of the data with the high-level theoretical calculations confirms the sensitivity of this probe to the subtle quantum mechanical factors guiding this benchmark reaction.

D irect characterization of transition states, the fleeting structures at lhe. boundary between, the reactants 'lind products in molecular transformations, is one of the grand challenges in physical chemistry (1 3). Substantial progress toward this goal has been made in the past two decades by using negative-ion photodetachment spectroscopy, which probes a reaction transition state by directly pboiodetachmg an electron fhnn an appropriate negative-ion precursor that bas a similar geometric structure to the transition state (.f)_However, it is much more difficult to measure the transition-state structures in crossed molecular-beam experiments, even with full quantum mechanical Slate resolution of the initial and final states, because in most chemica! reactions experimental observables come fi:"om the eoetributions of collisions with mall), partial WilV6S, or in classical. terms, of collisions with a range of impact parameters or total angular lUO. mentum. A partial wave is a quantum-scattering

Stale Key Laboratory of ,Molecular lI>eacftof] Dynamic~, Dalian Institute (Ii Chemio;al Physic;, Olh,ese Alad~my (Ii Srienrss, Dalion 116023, Li<loninl), People', Rep(lbli(()f China,

'Til es~ ~ uth or s ron tri b u ted e q ua! Iy 1(1. th i5. work.

tTo Whom cerrespondence should be addressed, E-mail: ~my<ll',(n (X.Y.);· zhal'lgdh@dfq},.aun (D.H ,Z.)

Fig. 1. Experimental and theoretical DeS of the HF(V = 2, ! = 6) product of the Fep3/Z) +

HD(j ca 0) re'ilctlon in the ~_

ba ckw'a rd sea tte ri ng di - e!

rection, The solid circles g II o.l

ijre experimental data; 'R_I~

the re d curve, the re sult of full quantum dynamics calculations convoluted with the exp en mental resolution and shifted 0.03 kcaVmollower in energy. The error 'bars in the experimental data are the estimated measurement errors (Llo) for !h" HF(v' =

2,! = 6) product peak

intensity (fig, 51) in the: col.lisionenergy scan. The three peaks are assigned to the partial wave Feshbach resonances of J = 12, 13, and 14 in the F + HD .... HF + D reaction, as explained in the text The threedimensional Des shown was measured at 1.285 kcaVmo~ with F and' Bindicating the forward- and backwardoscattering, respectivelY,directions for HF with respect to the F-atom beam direction.

state With a well-defined total angular momentum quantum number (1), which specifies the overall rotation of the reaction system and is conserved in the reaction precess, A full reaction

process can be considered as a combination of reactions of many partial waves (J = 0 to .Jrm ... ) that contain the most detailed information on the reaction transition state. However, contribunon of a single partial wave to the experimental observables isextremely hard to resolve because different partial wave contributions are severely overlapped under most circumstances.

Reaction resonances are transiently stabilized quantum states in the transition-state region with well-defined energies, and energy widths, which change with individual partial waves, When the energy changes associated with certain partial Waves ate larger tban the energy width of the resonance stares, the experiment-al observables may be largely determined by a few or even one partial wave 10 that energy region. As 11 result, detection of partial wave reaction resonances would offer th e m OSl detail ed pi oture of the quantum mechanical factors guiding chemical reactivity and so bas attracted great attention in the field. The




12 'Collis)i,n :Energy ('kcalli1l<lI)

www.sciencernaq.orq SCIENCE VOL 327 19 MARCH 2010



F + Hl reaction system has been a textbook case io the- study of resonances (5 8). 10 this work, we report the experimental observation of partial wave resolved resouaoces .il:i the F + HD -> HF + 0 reaction, providing an extremely accurate probe orits reactive resonance potentia] energy surface (PES),

We performed a fully quantum state-resolved, crossed-beam reactive. scartering experiment on the FeP1r!) + HD(j = 0) ~ HP()!',f) + D reaction by using tbe D-atom Hydberg tagging technique (9). A unique daub la-stage discharge P,.atom beam was used in this experiment So that a high- intensity F·atom beam with high-speed ratio (v/ov) could be obtained (10). Both beam sources (F and HD) were cooled cryogenically to maximize resolution

Time-of-flight (TOF) spectra of the D atom products from the F + HD reaction in the backward scattering direction. were measured at vannus collision energies from 0.9 to 1.5 kcal/mcl and are shown in fig. SI. Full product (HF) rotational state resolution was achieved, and differential cross sections' (DeS) fo!" Hf{)I' == 2) products in individual rotational states were obtained. A. clear osci Ilatory structure is evident (Fig, I) in the collision energy dependence of the nes for the HF(I/ = 2,j' '" 6) product in. the backward scmtering direction, Theoretically, we carried Gut full quantum scattering calculations on a recently repented PES (1J). The rheoretloal DCSfn the backward direction for the HF(II "" 2.)' = 6) product were computed at various collision energies. For better comparison, the experimental resolution factor W.8S included in the theor-etical snnwa.timl. TIle original theosetical IcsuIts are shifted

lower: in energy by 0.03 keel/mol in Pig. i to compare ,vith the experimental results. With this small shift, the agreement between the theory and the experiment is remarka ble,

in order to trace. the dynamical origin of the oscillatory structures. we calculated the DeS in the backward direction for the HF(j!' ,.,. 2,)' = 6) product with different .1",.",. It turns out that the three rna in peaks exhi bited 10 Fig. I emerge subsequently by taking Jinm/ = l2, 13, and 14 (fig. S2). In other words, the assigned J = 12 peak in fig. I emerges only after including the J = 12 partial wave in the cal culation s. The sam e is true for the peaks 'labeled J = 13 and 14. This clearly confirms that the detected oscillations are eontributed by the J = 12, 13, and 14 partial wave Fesbbaeh resonances in the F + HD -I' HF + D reaction. It is interesting to point out that the final heights of'tbe specific peaks of these oscillations can be obtained onJy by including larger.J cootri buti 0l1S in the calcula tions, implying tImt qu antum interference among the partial wave resonances is qui\e signi:fican,t in the observed DC$, In addition to the HF(v' = 2, f' = 6) product, partial wave resonances are also observed 1:0 the Des of other well-resolved rotarionally excited HF(v' = 2} products in the backward direction.

The partial wave resolved reSOnal1DCS can only be observed in both angle- and state-resolved DCS measurement with extr~mely translational energy resolution in a scattering experiment. It is necessary 10 point out that the reaction resonance that we are dealing with here is the same rescnance studied in (7) and {In. Des in different scattering directions WdS also measured in this

work at the collision energy of 1.285 keal/mol (Fig. 1). The three-dimensioual DeS plot shows a narrow forward-scatteriug peak with a broader backward-scaneringdisnilnnion, indicatiug a strong resonanoe effect in the reaction. The present experiment provides a spectroscopic probe of the resenanoe potential far beyond the generally accepted "chemical. accuracy" of'about 1 keel/mol. The observation of the partial wave resolved resonances provides opportunities to study their effect on chemical reactivity at the truly state-tostate- to-state I evel,

Refer~n{)e s and Note~

1. G, C, SIhat1, Science ·28"8, 1599 (2IJoi)).

2. F. Fern;! n dei" Alor)s 0, R, N. Za,~~ Ann!!. Rev. Pbys. r; bern.

53, 67 (2002).

3. 1- c. pOliln.yi, A_ H. le\'lilill1c,. (l1em_ Res. 28, 119 (;1.995)

4. D. 1'/1. Neuma,-k, Science UZ, 1446 (1996).

5. D. E. Maoolopowlos et 0/" Sdm(e 262, 1852: (;1993).

6. R. T. Skodje et ol., Phys, Rev. Lelt. 85, 12M (2000).

7. R. 1 S~odje et oi, 1. Ch~m, I'hy.>. 1.12, 4536 (2000). B. M. Qiu el 01 .. Science 311, 1440 (2006).

·9. Th~ exp erlmen ta I techn lq ue used in thiS work is d estri bed in delaiL in the supporting material> <l\Iaililbl~ on 5LielJ(e Oolin s.

10. 2. Ren e:t at; Rev. Sd. Inslmm. 77, 016102 (2006).

11. Z. Rene/ (1/., rroc. Natl kad Sci_ U:S.I1. ros, 12662 (2008).

12. We ~c~nQWleil£w tile support ill \his work 'by!lie C~ine,e.

Aeademy of Sdences, tIff National Natural Sden(~ Foundation of C!1in., end the Mil11>UY of SdenGt' and T~d1nrnagy 01 Chino.

Supporting Online Mat~rial

~IWW .• cien(:emagorg/cg ,Imn I en tlfu U13" 7/5 9 nn 5 0 VD[ 1 Ma lerial! an d Meth 0 ds

I'ig,. 51 and 52


B D~i'~mber2009; ~'(~pt~d 2B J~;IU"ry 2010 lG.l1l6/5dem:€.1185694

MechanosensitiveSelf-Replication Driven by Self-Organization

J,H:qui M. A. (arnall,' Christopher A. Waudby,1,2 Ana M. Beleng,u!!r/ Marc c. A. Stuart,]'" Jerome J.-P. Peyralans,4 Sijbren Otto4*

Self-replicating molecules· are likely to have played an important role in the origin of life, and a srnal! number of fully ,ynthetit self-rephcators have already been described. Yet it remains an open question which factors 'most effectively bias the replication toward the la!'-from-equilibrium distributions characterizinq even simple organisms. We report here two self-replicating peptide-derived macrocydes that emerge from a small dynamic combinatorial library am! compete for a common feedstock. Repl.ication is driven by nanostructure formation, resulting

f rom the a ssernh ly of the peptides into fi bers held together by !1 sheets. Which of the two replkators becomes dominant is influenced by whether the sample is shaken or stirred. These resu Its establish that mech~ nico;Ji forces can <Jet as a sl;'l~cti on pressure in to e competition between rE!plka!or~and can determine the outcome of a covalent synthesis.

The ab. ility to ~ an essential component of evolvable life, yet how replication emerged during the origin of life remains an unanswered question U 3). TIle experimeotal approach to this subject bas focused lilrgely 00 kinetically controlled autoearalysis, whereby a molecule is able to catalyze its own formation from a set of precursors, This outcome has been achieved using both biological molecules

such as DNA (4,5), RNA (6) and a-hellcfll peptides (7, 8) -and nonbiological molccules (9 11). Also. cross-caralytic systems have been reported wherein two or more sets of compounds induce one another's synthesis (13. Jif). These relatively simple systems are still far from the complexiry exhibited by contemporary organisms, which can undergo Darwiman evolution and exhibit a complex internal organization.

We previously proposed that it should be possible to use tlynal11ic combinatorial libraries to develop molecules capable of promoting their own fonnation, while forming extended. assemblies at the same time (15), Dynamic cotnbinatonal libraries am created by mixing building blocks that C8111"Cact with each other tlu"Ough the formation of reversible covalent bonds, leading to a mixture of products that are all in rapid equilibrium. We reasoned that if two or more molecules of a particular product could stabilize one another through noncovalent binding, the equilibrium would shift toward formation of this product at the expense of the other library members. A number of examples have very recently appeared that exploit this principle. Giuseppone described a dynamic equivalent of the Rebek replicator (1 6) ~s well es an a'umpoietlc system

I University of Cilmbridge, Department of chemistry, Lenlljeld Road, Carnbridqs (B2. lEW, UK, ~Departmenl of S tru ctu rat a no Mole, u tJ r Bi ol.ogy, University C{)!I~ge lcnden london W(lE 6ST, UK. 3Groningell BiomQI~~ular Sd~IlC~5 ~ n d Bi 011>[ hno\ogy IOSD tut~, U II iver~i ly 0 f G roning en, Nij E'<nbgrgh 4, 97 47 AG Grol1i 119 en, N ~ther\d 11 ds, ~ (eAtr~ lor Systems Chern! Illy, Slfa~llgh Institute, Ulliv@r>ity 0.1 Grn f11 ngen, N ijenborg h 4, 97 '+7 AGGron i ng en, N eUlerliln ds.

"To l'I"hom rcrrespundertre should be addressed. E-mail:


19 MARCH 201'0 VOL 327 SCIENCE WI/

where reversible formation of an amphiphilic molecule was able to promote the formation of more umpbipbile by accelerating dlssolution of the start-

ing material (J 7). Sadownik and Philp coupled an established replicating system to asmall dynamic combinaronal library of starting materials and eom-




, - GI)'-I.e1J -l V -Leu.-t. vs-OH

2. ,=OOOH

S F tOOl htwu
~ !If] lnI.
<II 4 .... ,
160 M .. ,
C '0
Ii ill -
.... IJ
0 10 I! 2(1
u " G TIm {ow,.
aQ ,.
!!it! h'III'
S Jnur.
r. :a mIT 4(1 J""',
~ , r
5 10 IS
H ..... IOO 100
t- .. .~ .
II! 80 E 60
to< <:I' f/
-'" .' i /
! eo 6(1
i 40 '§ 40 I
2 I
Q ':;
S ao a dt" ~ :w ""'II~'I' • ~.
iii ij ,01
!e l'
.... o. 1.6. 0
0 2: 4 B 10 12 0 2 " I) a IIJ
1'Ime1·~) TIm{11 1JY'i1 Fig. 1. (A) Schematic illustration of a small dynamic combinatorial liorary made from dithiol building bloch. High-periormance liquid chromatography analyses (monitored at 254 nm) of product mixtures obtained by oxidation of (8) building block 2: (3.8 mM) after stirring at 1200 rpm for 15 day.s; (() building block 1 (3.8 mM) after stirring at 1200 rpm for 15 days; (0) building block 1 (3.8 mM) after 16 days with no agitation; and (E) building block 1 (3.8 mM) after shaking at 500 rpm for 20 days. Evolution of the product distribution with time upon agitating a solution of 1 (3.8 mM) by (1'-) shaking (500 rpm) or (6) stirring (1200 rpm) . .Mass spectra of the material c.orresponding to the peaks in the chromatograms are shown in fig. S1. Growth of cydic hexarner (rirdes) and heptarner (squares) upon seeding an irnrnaMe small dynamic combinatorial library made from building block 1 (3.8 mM) with (H) S mole percent (mol %) of 16 followed by shaking at i200 rpm or m 5 mol % of 17 followed by stirring al 1200 rpm.



petmg reagents (lm. Ulijn and colleagues have reported an enzyme-mediated peptide synthesis ill which gelation drives formation of a single product from a dynamic mixture (19).

We now report the results of our OW11 efforts using small dynamic combinatorial libraries made from self-binding building block 1 (Fig. lA), which features a peptide sequence with altern ati~g hydrophobic (leucine) and hydrophilic (lysine) c-amino acids. Peptides with such eharacteristics have a bigb propensity to assemble uoncovaleutly into a ~-sbeet structure (20, 2J). Molecules oft can also bond covalently to one another through oxidative disulfide fQl1uation from !heir pendant thiol groups, producing a mixture of different macrocycles in the presence of oxygen from the air (22).

AI approximately neutral rH the resulting produce mixture would normally be expected to be under thermodynamic control with a ring size distribution reflecting the extent to which inter" molecular peptide-peptide interactions stabilize each macrocycle, However, kinetically controlled processes can also determine which products are formed, as we show herein, We first performed a control experiment in which we observed by liquid chromatography mass spectrometry (LC~MS) the product distribution resulting from stirring dithiol 2, which does not feature a peptide chain, in borate buffer at pH 8.0 {23}. The final composition, a mixture of'cyclic trimerand tetrarner (Fig. 1 B), remained unchanged over a period of 15 days .. The S11me experiment with peptidefimctionalized building block 'I 'was also monitored by LC-MS and gave very similar behavior during the first 4 days. However, after that period a sudden change in composition occurred and the cyclic heptarner rapidly became the dominant product, consuming most of the trimer and tetramer in the process (Fig, J C). This behavior depeods critically on the mode of agitation. Repeating the experiment with peptide 1 in tbe absence of mechanical agitation resulted in a mixture COI1- tainin g mostly trimer a nd tetrarri er that rem ained unchanged over a period of?5 months, similar to the behavior of control compound 2 (Fig. {DJ. Rernarkablj; repeating the experiment with peptide 1 using shaking rather than stirring as the mode of'agitation induced the preferential fonn~tion of cyclic. hexarner (Fig. I E).

The sigmoidal nature of the growth of the hexamer and hepramer in the above experiments (Fig. 1, F and G) suggests that these macrocycles ate able to promote their own formation, which we continued in two separate experiments: We added a small amount of hexamer to a sample of 1 that was then shaken fur 9 days (Fig. lH) and also a small amount of hepramer toa sample of 1 that was then stirred for 6 days (Fig. 11). In both cases, the samples consisted mostly .of trimer and tetramer at the point of addition. TIle addition of a small amount ofhexamer or hepramer clearly induced the formation of more of the respective maerocycles,

We have characterized the solutions. of hex amer and heptamer using cryogenic transmission electron microscopy (cryo" TEM) and circular SCIENCE VOL 327 19' MARCH 2010


f 1O
i °
~ .'0
0 '., ,_



'. ",

Fig, 2. (ryo-TEM images of (A) 16; (8) 1] (inset: magnified view of representative fibNS); (C) rircutar dirhroinn spectra of 16 (solid line), 1] (dashed line), and a mixture of 13 and 14 (dotted \ine); (D) schematic representation of the proposed formation of fibers of 16, The benzenedithio! core of building block 1 is shown in ye!!.ow and the peptide chain in blue. Stacks of hexamer are held together by the assembly of the pe pride chai ns into eta nga ted cross -a s h eels. The stac ksg row fro m thei r ends, which increase in number through stack fragmentation, enablil'lg exponer]tial growth.

dichroism (CD), Fourier transform infrared (FTJR), ultraviolet (UV). and fluorescence spectroscopy. In both so luti ens, long '11 in fibers are clearly visible by cryo-TEM (Fig. 2, A and B). These fibers are up to 1 to 2 um ill lengih, The hexamer and heptamer fibers have comparable diameters in the range of 4.7 to 4.9 nm. These dimensions are in agreement with the diameter of a single macrocycle with peptide chains extending radially frem the core ina f~-&heet conformation .. as determined from inspection of Corey Pauling Koltuu (CPK) models.

Having established lhe existence- of fibers, we next investigated the nature of the interactions holding them together. 111e CD spectra of solutions containing mainly bexamer (solid line in Fig. 2C) or heptamer (dashed line) show most of ale feamres typical for p-slroet fonnation (i.e., maxirnum at 200 nm, minirmnn at 220 nm), whereas the spectrum ofa solution containing mostly trimer and tetramer (dotted Line) shows features typical lor random-coil p eptid es (i, e., minimum at 200 11m. low ellipticity above 215 nrn) (24).

Furfher evidence for the formation of ~-shee1- type structures in fhe samples dominated by bexamer and heptamer was obtained by fluorescence spectroscopy using thioflavin 1. TIlls dye exhibits an increased fluorescence wh <)11 bo und to peptides ihat fbrm extended B sheets (i.e., amyloid fibers) (25), Figure 82 shows fluorescence for hexamer and heptamer samples but not for a sample con" taining trimer and terramer. We also used Congo red, which shows characteristic apple-green blrefringcnce and a red shill in its UY/visible spectrumwhen bound to amyloid libel'S (26,2 nann observed both effects fur samples ofhexamer and heptamer (figs. S3 and 84). Additionally, samples of both hexamer and heptarnerwere examined by F11R spectrescopy (fig. S5). and C""O absorbance bands were observed at 1634 em I and 1627 em-I. respectively; These values are with in the range expected for ~-sheet pepudes (28).

Indirect evidence for the importance' of ~sheet fonnarion in the production of hexamerand


A O.3~------- ..... ~

c: 0.2 ID ~




a 0.3 ,.-------- .....

~ 0.2

co ;) tT

[L! Q I


250 cOO

0- ....... _ ............ _--

150 1000 0 250 500 750 1000

Eo 3,---------


o 2SO 500 750 1000 500 750 tOOO

Joer I e<ngll\ I n ITl Fiber length I nm

Fig. 3. Distributions of the fiber lengths irom eryo-TEM analysis of samples of hexarner (left) and heptamer (right) after (A and D) 48 hours without agitation, (8 and El stirring for 48 hours, and (C and F) sha king for 48. hou rs,

250 500

C 03~-------~ g 0·2 §I



heptamer was obtained from a study of control peptide 3 (Fig. 1 A), which feamres the same aamino-acid residues as 1 but ill a dilIerent sequence, Building block 3 produced a mixture dominated by cyclic trimer and teiramer independent of the mode of agitation (fig. 86). The CD spectrum 0 r this mixture resembles lh at 0 f 11l!1dOli1- coil pepti de (fig. S7). 111 e absence of su bstantial amounts of'hexamer or heptamer for fhis peptide indicates that formauon of these large macrocycles requiresa specific peptide sequence.

Based on the evidence for a-sheet formation and the cryo-TEM images, we propose {hat the hexamer and heptarner self-organize as S1101N11 schematically for the hexamer lit Fig, 2D.

250 500

150 \000

The forrnation of libel'S by the hexamer and heptamer may explain the mechanosensitivity (29-32) observed daring their Iormation. Indeed, ills known that the assembly and morphology of amyloid fibers can be influenced by mechanical energy (33). Three phases can be distinguished during the growth ofthe hexamer or heptamer; all initial nucleation phase" a growth phase in which theconcentration of these macrocycles increases exponentially, and a third phase in which die raie of growth levels off until most of the dithiol building block is consmuedff-ig. L P and G) Tt is likely that the fibers grow from their two ends .. IV; long as sufficient monomer is available, ~l:us process would result in a constant rate of formation


A:. • eo


• '311 •

... . ~ ~ ..
~ 7 0 I 1 • ,~ G.
n ~1IoJIj"' :: J I

~, •

Q --

I) j ~ 3 4 5! 7


Fig. 4. Growth curves for 16 (circles) and 17 (squares) from thr('>e solutions seeded at t = 0 wIth .<1 mixture of hexarner and heptamer (S rnol % of each) and subjected to different agitation conditions: (A) no agitation, (B) stirring at 1200 rpm, and (C) shaking at 1200 rpm.

of'hexamer or heptamer (linear growth). For hexamer and heptamer concemrations to increase exponentially, as observed, such simple growth 'from the fiber ends \5 not sufficient, More growing ends need to be generated as more hexamer or heptamer is being produced. This may be achieved by the breaking of fibers through mechanical agitation. Indeed, clear evidence was obtained that mechauical agitation induces fiber breakage. Figure 3 shows the length disributiens determined by cryoTEM of samples of hexamer and heptamer fibers, respectively, that were not agitated (A and D), stirred {B and E] or shaken (C and F). Curoulative length distributions are provided infig, S8. These distributions indicate au increased degree of fiagmentation upon l;tgjtati.O}l~ The hexamer and heptamer fibers were fragmented to a comparable extent by stirring, whereas shaking appeared more .eftlcient in fragmeuting hexamer

fibers than heptamer fibers. -

Control experiments confirmed that the rate of seeded fiber growth is influenced by the number of fiber ends contained in the seed, We divided a seed solution dominated by hexamer fibers in three portions: One was shaken for 48. hours to induce fiber fragmentation, another was shaken for 48 'hours and then sonicated for 5 min, and tile third was not agitated. We then divided a t1Citl<!girattd solutiomnade from 1 eonta ining, besides remaining thiols, mostly trimerand tetramer macrocyeles imo three portloas. We seeded one portion witb the shaken hexamer sample, the second with the sou" lcated sample, and the third with the nonagitated sample. Subsequent hexamer growtb was faster in the samples containing the agitated and hence fragmented fibers, as compared with that containing the nonagitated seed (fig. S9).

Shaking and stirring differ in the magnitude and localization of the induced shear stress, We estimate (23) that the maximum shear stress proehl(; ed by the rotating sti r barf n out experim ems is on the order 00 x J 02Nm 2, assuming that the stir bar is separated from the bottom of'the vial by a layer of fluid with a thickness on the order of I urn (34). In contrast, the maximum shear stress experienced by a fluid on the bottom of's cylinder undergoing orbital shaking under the conditions of our experiments is estimated to be on the order of 2 Nm-2 (23). Although the maximum shear stress induced by sbaking is substantially less than that induced by stirring. shaking can still be

more effie ient tban stirring in fragmenring the hexamer fibers, because it acts on a larger fraction of the sample volume. The shearing upon stirring is localized in the small part of the sO]-Ution between stir bar and vial bottom, rather than caused by bulk fluid dynamics, as evident from the following experiment: Three sample vials COD' tainmg identical solutions of J were stacked on top of each other, each equipped with 11 stir bar and stirred on a magnetic stirrer, All three samp] es produced heptam er, but th e rate of growth diminished with increasing distance from the stir" rill' plate (fig. S10). Tberare of stirring in each sample was examined using a. high-speed eamera and was coofirmed to b 1:\ identical in the three samples, which must therefore exhibit comparable bulk fluid dynamics, The difference between the three samples is the force that pulls the magnet toward the bottorn ofrhe sample vial, reducing the average di stance between stir bar and vial bottom, boo ce .increasing local shear stress, Indeed, similar trends were observed when we compared how the force pulling the magnet to the bottom of the 'Vial and the apparent rate constant for the expoueotial phase ofheptamer growth depend on the distance between stir bar and stLt plate (fig, S 11).

To further investigate the link between agitation and product selectivity (i.e., hexamer formacion upon shaking and heptamer formation upon stirring], we performed a series of competition experiments iu. which the hexamerand heptatner competed for building block. 1 under a variety of agitation conditions, In these experiments, the nucleation phase was bypassed by adding equal amounts of hexamer and heptamer to a solution that contained mostly trimer and tenamer; These experiments were set up using three different agitation modes shaking, stirring, and 110 agitation- -and' the compositions of the samples were monitored over time using LC-MS. The results (Fig. 4) sbow that once present, the heptamer is ab le topropagate even in the absence of agitation. thllt is, wheo fragrneutation plays a reduced Tole in fiber growth (Fig. 4A). The absence of substautialhexamer growth in the same sample indicates that tile growth of pre-existing fibers is faster for !he heptam er than for the hexamer under these conditious, Heptamer growth is even more efficient when the sample is stirred (Fig, 4B). The heptarnerout-competesihe hexamer under both of these conditions, and only



shaking allows the hexamer to grow at a COinpethive rate (Fig. 4C). This final observation is consistent with shaking being more efficient than the other agitation conditions ill causing the breakage of fibers of the hexamer (Fig. ]). The data in Fig. 4B show that fibers. of'hexamer, when subject to stirring, are not C>Ol1S1lme{! and converted to heptamen The same conclusion was reached in a separate experiment in which a solution made from 1 was seeded with only hexamer an d su bjeeted to d ifferern agitation con diti on s (fig. S 12). Similarly, Fig. 4C shows that fibers of heptamer persist Wh611 shaken. These observations suggest that the selectivity in forming tbe larger macrocycles is. kinetically determined during the fiber growth process. Depending on the history ofthe sample (shaktng. stirling, or no agitation), We obtained hexarner, heptamer, or a mixture of biller and tetramer, respectively, No subsequent change of any of these samples to produce some common thermodyoamically controlled distribution could be observed even after J months without agitation,

The above data support a model for mechsnnsensitive tiber growth that involves two processes; elongation of fibers (i.e., linear growth) and breakage of fibers to produce more growing ends (which enables exponential _grmvtlj). It is clear that fiber breakage through mechanical force IS critical to tbe fonnation of substantial guaotities of fibers, because in the absence of mecbanical force no fibers oi'be.;;:amer or heptamer could be observed. The outcome of the competition between hex amer and beptamer is primarily determined by differences in their exponential growth rates, that is, by differences in the efficiency of fiber fiagmeotation, However, when the two competing fiber populations possess comparable rates of'fiugrneoration, the outcome can instead be determined by the relative efficiencies of'their lineargrowth processes. Under stirring con di lion s, the bexamer and tlre heptamer fibers arc fragmented to a comparable degree (Fig. 3). The heptamer dominates, however, because this .eompotmd is more efficient at elongating its fibers through linear growth (!IS evident from Fig, 4AJ Shaking is more effective at fragmenring hexamer fibers ihml heptamer fibers (Fjg. 3); hence, we obtain hexsmer under these conditions. The greater susceptibility of the hexamer to fragmentation upon shakii!g can be rationalized, as the macrocycles in the bexamer fibers are able to form fewer 13' sheets than those in the hcptamer fibers.

The observation thet hexamer (or heptamer] accelerates [he formation of more of the likesized macrocycle may be explained by the fibers acting ill> kinetic traps: Through the incorpozationof hexamer or heptamer into-fibers, their reequilibration into different macrocycle sizes is inhibited and the overall distribution is shifted increasingly to the macrocycle size residing in the fibers" We cannot exclude tile possibiliry that the fiber ends also actively accelerate theformation of like-sized rnaereeycles through an autocatalytic templating effect. SCIENCE VOL 327 19 MARCH 2010



The implic.atioh of our results for dynamic combinatorial chemistry is that this technique need not be limited to exploring the thermodynamic minima of molecular networks: kinetic control can dominate, provided that the noncovalent interactions aJC sufficiently strong and/or numerous, Our results further show that it is possible to obtllin kinetic products trom an assembly process where all tbe individual steps (covalent disulfide exchange 1Il1d noneovalent peptide-peptide interactions) are reversib 16. Sue b a transition from thermodynamicelly controlled self-assembly to kinetic controlrmrst have been lin important step in the orlgio of life, as life is far from equilibrium (35). Finally. our approach represents. a promising method for the discovery of self-synrhesizing materia Is in general and noncovalent polymers in particular (36).

References and

1. L f. Orgel, Nalure ~58, 2Q~ tl992l-

2 _ L. E, Oigel. I1c~_ (hem. fles. 28. 10-9 (1995)-

3. E_ S,athm~ry, J. Mayn~Td Smith. J tbeat: Bfal_ 187,. 555 (1997).

4. G. von Kiedrowski. Anoow. Chem. tnt. fd. [lll;!- 25, H2 (1986).

5 G. Von Ki~dm\'ll~i. B_ WW>:Iia. J" Helbing, M. M"\z~n.

S. Jordan. Arw~w. (hem, In I. Ed. Engl. 30, 423 1199lJ.

6. N. ~aul, G_ F_ loyce, srec Nat/. AClJd_ Sci. U.s.A 99. :12733 (2002).

7 _ D, H. Lee, J _ R. Gr anja, J. A Mil r~h"'. K. S~"eti n, M., R. Ghadri. Nature 382. 525' U 996).

8_ S_ Y.;10, L Ghosh, R. Zurshi, 1. (hmie~~w,~i, j. Am, C~eilJ. Soc. 119, 10559 {L99n

9- A_ Vldoline, D. Philp. 1M. l- Org. Ch~m. 200~. 593 (200'l).

1(1. T. Tjiyikua, P_ llallester, J. Rebek, J. Am_ Chern, So~, n2. 1249 (1990).

11. V Rot~Uo, 1- I. Ho~g, J R~b;.ki 1. Am. Ch~m, Soc- 1~,~, 94ll (1991).

12. B Wang. I, O. Sutherland, CiJem. [ommun. (Camb.) 16, 1495 (1997).

13. R, l" Pi ete", t Hu r, J. R'.b. k. Angew_ (hem, /nt EeL lng/. .33, 1579 (1994),

14. D. H. tee, K. SeveJio, y, 'Yokobi3ya,hi, M R, Gh~dfri,

Nature' 39(). 591 (1991').

15. P_ T Corbett et (11., Chern !UN_ 106. 365.2 (2006)_

16. S. ·Xu. N. Giuseppone, i Am, (/:rem. Soc. 130, 1326 (ZOO!)).

17. R_ Nguyen, L. Allouch2, 0_ BUhler, N_ Giul€ppone.

Allgew. C/1el11. lilt Ed. 48., 1093 (2009 J.

is. 1. W_ Sildownik. D"Philp, Angel!'- (h~f"'- !n~ [;1 47,9965 (200a).

19. R, l. Wllli<lml et G/" Naf. Nanotec/mo/. 4., 19 (c009). ZO. W_ F. DeGriido. J'. D. !.ear,]. 11m, ('hem. Sac" 101,.7684 (1985).

21_ Y. Kri~hn"n·Gho~h. 5 B"lasubr~manian. Angew. Ch~m. lilt. Ed. 42, 2171 (2003)_

2 2 _ 5. Otto, R. L E _ Furia n, i. K. M~ 5" nders, J ilro.

Chern. 50(. 12.2. 12063 (2000).

23_ Materi,,[s and method, Me aVilil~ble as SUPPQtli"rJ

material 011 Science Online.

24. ~. J. G,e.nfi~ld, Nat_ Proti»; 1, 2816 {200n

25. H .. levine Ill, PrQre/1l 5tJ. 2, 404 (I 9~ 3).

26_ H Puchtler, F_ Sweat, M_ le~iM.1_ I-jistoehem. (yj'(J~hem'10, 3SS U \162).

27 _ W _ ~. Klunl\, l W. P,ttegr<lW. G_ ] _ ilbr'ilh~rfl, 1. Hislocnem. Cytochem, 37, 1273 (1989).

28. I. KOi1g. S_ vu Acta CliQ{him_ Bfophys. Sin. (Shanghai) 39, 549 (2007).

29. R',port, of iTlolB,LJlar \ystenis who)". !Ii" ha~o'~nsi~vI\y atlerts defined covalent trarsformanons are rare. ~or exampl~s of such systems, .ee 011-32).

30. C. R. Hk~nboth ~r ot., Natum 446. 423 (2007).

31. 11. A. D.wis et' al_. Naturf'! 459, 68 (2009)_

32. A. Piffmaltei. 5. Karthikeyan. R. P. SijOesma. Not: Chem. 1,133(2009)_

33. A. T. Pet~ovil et al., Sciel1ce 307 I 262 (2005).

34_ Tni" estimate is b,;,,,,d on the I:ypiCilI tlJidme5S 01 a I")ler 01 (ubriomt be !Wee n 51i di ng non mnform a I tonto ets, 5 ee '(37). ~5_ A. Pross. /. Iheor. Bioi. 2~0. 393 (20Q3J_

36. L. Brun5V€ld, B.l. B. Folrnel, E. W. Meijer, R P. 5ijbelmil, Chem RfN- 101, 407.1. (~001)_

n. R. 5. Dl\ryer· Joyce, B. W. Drin~watfr. C]. lJonohoe, rtoc. R, Soc. tond. A 459,957 (2Q03J_

33. We thank J. I<. M. Sanders and 1. B. F. N. El1gberl> F{)r ~ri~[a( re"drng of the manusrript ,,~d

C. M. Dobson 10) usefu I discu >Ii on" This reseann has been sup~or~d by the Engineering and

P hYlkal Sden (5 f<e~ea rch Ccim ell, lite Roi'~ I SOciEty, the 0)'00 m ic (om bi no to~~ l (hell'li5tiY

M a rie-( uris R-e sear(h Tra ioio g Network, and Coopera H u n ln 5cie"o~ and Technology iCOSr1 CMQ7D3_

Supporting Onlille Mat~fial

Wi'IW .>,jen(e~'~g"orglrgilrQ n telllffu lin 2715 9-72/150 2J DC 1 Milli>rials and Me.tliod;;

figl. 51 to 512-

Re r'~rences

1 Odobsr 2 0!)9; at c~p. ted 10 r;e!lmil ry 2 010 10.11261\ d enoe .1182 767

Seminal Fluid Mediates Ejac'ulate Competition in Social Insects

Susanne P. A. den BOE'rr1 Boris Baer,2,J Jacobus]. BOQmsmal~

Queens of <lots and bees normally obtain a lif~time supply of sperm on a single day of sexual activity, and sperm competition is expected to' occur in lineages where queens receive sperm from multiple males. We compared singly mated (monandrous.) and multiply mated (polyandrous) sister groups

of ants and bees and show. that seminal flutd of polyandrous species has a more positive eftett on the survival of a male's own sperm than 0[1 other males' sperm. This difference was not observed in the monendrnus species, sugg.esting that incapacitation of competing sperm may tUJVe independently e\folv\ld in both bees and ants. In Attu leafcufter ants, the negative effect of the' semi ria I fluid of oth~r males was negated by secretion from the queen sperm-storage organ, sli9gesting that qu~ens may control ejaculate competition after sperm storage.

'MUCh sexual selection ill polyandrous

. . species oceu. TS. a.fief matin .. g in the

_ form of sperm competition and cryptic

female choice (1-3). In most animals, males seek addjtiOblll mates-to increase the number of their offspring, and females may rernare to gain direct or indirect benefits to promote offspring

'G:!ntrg lor Sodal Evolu!lon. Department 01 Biology, Uoiver'iity 0·1' Copenhagen. Universitetspari<en 1;;. 2100 Gopel1hogen, D enmil rk. 'A RCCel1tr.e of Ex~€Ilen(e in PIa nt Ener gy Bi ology, M($ B.uilding M310, lh~ University of Western Australi~. 6009 (filwl~y, Au ill d!ia. 'Centre for EVoltl6o~ilry Biology, Scll 001 of Animal Bi 01 OgY (MOn), The Un iveBity of WlC<;tern Austr<l\ia, 6009 Cr<lwley, A ustr alia,

'To whom certespondence ,hould be addressed. E-mail: jjboom,


quality [e.g., (1 Sl]' The eusccial ants, bees, wasps, and termites, in which only relatively few individuals have the opportunity to mate, are excepnons, because they evolved from strictly monogamous- ancestors (6. 7). Newly eclosed queens of B,n\R, bees, and wasps are receptive for mating during a very brief period of time (<'I few hours to a few days) and never rematc (7, 8). Whereas queens from basal lineages store only a single ejaculate, obligate multiple queen-muting bas evolved secondarily in honeybees, vespine wasps, leafcuner ants, army ants, harvester ants, and a few minor taxa (6. 7. 9). Thus, opporrunities for postcepulatory sexua! conflict have repeatedly emerged and may have induced converge-lit adaptive responses.

The- absence of remaring implies that ejacuboles from multiple males coexist 'within a queen's sperm srorag.e organ (sperrnatheca) throughout her life (7, j(f). This situation is likely to have undergone selection fbr prudent mutual exploitation among partners, similar to rmrtualisms characterized by lifetime oommirment (iJ. 11)" Ejaculate competition might occur shortly after multiple insemination, if it has no· 11'1BjOt negative effects on queen health and longevity (10, 13) and leaves sufficient higb-quality sperm fu," her to realize- her lifetime reproductive potential. However, selection is expected to act against antagonistic interactions betweenejaculates after sperm storage, be cause a female's reproductive life span ultimately depends on h er ability (0 fertil ize eggs from this nonrenewable stock of sperm (14).

Queens of Atla leafcutter ants use few sperm to fertilize each egg, consistent with a correlation between lifetime reproductive success and sperm-storage lhnitations (15),. This implies (hat selection- Ot1 sperm viability should be strong and that both [be male seminal fluid and queen spermarhecal fluid are necessary fur reproductive success Male accessory gland (AG) secretion confers a positive effect. oft sperm survival, even in very small quantities, in leafcuner ants {l6) and honeybees (/7), and the spennathccal fluid of honeybee queens also positively influences the viability of stored sperm (17),

We used an in vitro sperm survival assay (J 6-! /Y) to test the effeots of own A G -secretion. AG secretion of other males, and queen

19 MARCH 201'0 VOL 327 SCIENCE

sperrnathecal fluid on sperm survival in two species of corbiculate bees and three species of fungus-grcwing ants. This assay (i 8) provides relative sperm survival data that can be directly compared within experiments, but does not allow accurate comparisons of absolute sperm survival percentages across species, because sperm and seminal. fluid eontributions may vary across natural ejaculates (HI).

Our choice of species exploited the sharp evolutionary transition shown by both the COl"bicularc bees and the fimgus-growing ants from exclusive single mating of queens, i.e., storing sperm tum a single ejaculate tBombus terrestris bumblebees and Tradiymyrmex cf ;;el&ki ants), [0' mill tiple III ari ",g. i, e., queens storing sperm from lTIuJ tip I e mal es with whi ch th ey mate in qUick succession tApis mellifem honeybees and Atta colombica and !1cromymleJ: echinatior Ieafcuner ants) (6, 2U). We found that providing sperm with AG secretion from tbc same male significantly increased sperm survival in all five. species (Fig. 1), indicating that sperm is consistemiy protected by the quantities of AG secretion that we used in our assays, The absolute sperm survival values arc almost certainly underestimates for sperm viability in natural ejaculates, because the Hayes saline buffer used dOesoot nourish sperm, so that bigher AO concentrations may be-needed to reach tbe close to 100% in vitro Sperm survival that can be achieved with this 'assay (16)

We next investigated whether the AG secretion of other conspecifie males would be equally effective in enhancing sperm survival (J S). We expected that AO secretions might be at least partially hostile to other males' sperm in the polyandrous species, but that no such effects should occur in their monandrous sister groups 'where «jacul!l1es never interact in vivo. We used generalized estimating equations, nesting species within mating systems and using individuals as. repeated measures (18), to compare the effects of AG secretions fiom a focal male, brothers of the focal male, and unrelated males on sperm survival We found significant effects of mating system (polyandrous versus menandrous, i = 22.33, elf = I, P <: O.OOl) and treatment (AG origin, '/ = 19. t 7, df = 2, P <: 0.00 I). The overall mating eystem difference may not be biologically meaningful because the absolute sperm survival percentages across species (nested within mating system) are 110t directly comparable. However, a significant interaction between mating system and treatment (X2 = 18.62, df"" 1, P < 0.001) indicated tbat sperm exposed to alien seminal fluid showed reduced levels of spenn survival in polyandrous species, but not in monandrous species (Fig. 2). TIlls was confirmed by a rerun of the analysis for the monandrous species seperarely, where no' &lgni6c~m effect of treatment could be detected (r! "'" 2 .13, df = 2, P = 0:346 for Bombus and i- = 0.0'2, df = 1, P = 0.883 for Trachymymwx), The origin of'nonsclf


Fi g. 1. Positive effem of male AG secretion on sperm su rviva I E so

<JCW5S three species of fungus- III

g. eo growing ants (Trachymyrmex d. .,


zeteki, Acromyrm61: echinatior, Atta .2 40

colombica), .<1 bumblebee (Bom~ ~20 bus terresfris), and' a honeybee lApis melliferal. The upp_er five charts 5 how the a bsolllte spe rm viabitity values in the control treat-

ments with HilYes Siltine (left) and !!l

the AG supptementations (right) ffi

that produced the AGlcontrol sur- ~

vival ratios in the lower pa nel, The ~ 1 0

datil for A. (olombico and A. ~

mEilifera are from previous studies .s (is, 16), w'hereas those lor the other three species were collected in the present study. The slope of the regression in the lower panel (y = 0.824x + 0.7B4; II = 0.876; P = 0.019) is not significantly different from 1 (i.e., approximately parallel to the diagonal: 95% confidence interval slope: 0.2S4-1.3(5). The intercept of the regression is significantly dif-

ferent from 1 (zero after log-transformation: t'J. = 7.196; P =,0.0(6), indicating that the.AG treatment had a consistently positive effect relative to control exposure to Hayes saline, Tl'ie in~et shows the male AGs and eceessory testes (A Ts) of a representative species (ll terrestris) as they were dissected, All bars are SEM.




E i

a. III



Live/dead sperm in control treatments

Fig. Z. The effect of own ~ re lated (broth er), and unrelated' male AG secretion on sperm survival (mean ± 5EM) in the monandrous bumble bee B. tenestn: and fungl) :ant Ti"t1~ chymyrmex d. zelekJ, and the polyandrous honeybee A. melli/era and fungus~growing a nts A. r;chinatior a od A, (%mbica. Photos of males are given in the panels. The brother !:realmer'!! {Ould not be perfor'lTEdfor Ttut'hy.rrrjtmf'X ct zeteki (1B). Thosebars labeled with differen t letters (a or b) .differ sign.ificanUy (all P < 0.001), whereas those with the same label do not (all P > 0.1).





6 90 ~

~ so

011 >

:J 70


Api< m"i1iler~

E 90.


~ BO


~ 70


&ambUs leffEJ"lf/S




TrMhymytmei ~r. ;ls/eM

90 E

[ 80

'" 011

I:: 70 ...J

~ 60

E 60 t CoSO "l


.2: 40 ...J

# 30

Ii a

AtJiumylffl<Pl echinaHor


E 90 t

CollO "l


.2: 70 ...J


AG seoretion source: c::I Same male

o Brother

Ll Unrelated male SCIENCE VOL 327 19 MARCH 2010



AG secretion. a brother OJ an unrelated male) did uot differentially affect sperm survival (X'2 = 2.44, df= I, P"" 0.118 for ipis; '"I: = 1.16, df= I, P = 0.282 for Acrom.vrme.x; and X? = 0.25, df > I, P = D.616 for 11rta_). This suggests that the negative effects on sperm survival are triggered by a. self versus nonself recognition process.

These results are consistent with some f0I111 of ejaculate competition in polyandrous honeybees and leafcuner ants, and suggest that AG secretions contain both compounds that are favorable fat sperm survival (Fig. 1) and substances that may incapacitate sperm of competing males (Fig. 2). To assess the combined effect of these putative forces on sperm survival. we repeated the. seoond experimeut for the three polyandrous species (but without the brother treatment), whileadding an equal mix of own and alien AG secretion as an additional treatment (18) {Fig. 3). Overall treatment affects were signifieaat inall oases (X2 = 8.3 I, df = 2, p = 0.016 for Apis; l = 7.:U.1, df > 2. P = 0.027 for Acromyrmex; and i'" = J 0.82, df= 2, P = 0.005 for Alia), However, own AG secretion does. not appear to offset the negative effects of alien AG secretion, as pure alien and mixed AD treatmentsalways resulted ill similar but reduced sperm survival compared to tile pure own AG treatment (detailed statistics are in the legend 10 Fig. n

After )0 min of exposure to alien AG seeretion, sperm survival was reduced by 6.6% in Apis, 14.7% in Alta. and 18.QilIo in Acromyrmex relarive to treatment with own AG secretion (Fig. 2). However, the residence time of inseminated honeybee spenn in the bursa copulatrix and lateral oviducts before ooillg stored or discarded is 40 to 90 hours (21), whereas ejaculates of A. colombica are probably transferred to the spermatheca almost immediately

t22} Because rival ejaculates of Alia have less time to compete before being stored, there may be more damage pel" minute of exposure, This is consistent witb A. melltfora queens expelling some 95% of the sperm provided by 10 to 20 matings 0])' Whereas A. colombica. queens store essentially an sperm provided by 2 10 5 matmgs (22. 23). Typical sperm presto rage time in A. echinutior is unknown, butis up to 5 hours in i/. versioalor, where queens appear to store around 10% of inseminated sperm (24).

Because every sperm cell provides a potential fitness benefit after the sperm storage process is complete (14~ 15), we expected queens to either prevent hostile AG secretion from being stored, or to neutralize antagonistic interactions between ejaculates shortly after they enter the spermatheca, We tested this by collecting virgin A. eolombica queens and artiflcially inseminating them with Hayes solution to obtain spennatlreeal fluid suspensions (I8). Exposure of freshly dissected sperm from male BO~SSOry testes to mixtures of spermarhecal fluid and seminal fluid of another male showed that spermatheeal fluid maintains sperm survival at normal levels (Fig. 4). Thus. we eonclude that spermath ecal fluid has til e potential to pre vent the negative effects of AG secretion on alien sperm.

These results reveal consistent posunsring sexual selection and sexual conflict In eusocial insects. They complement work elucidating why multiple queen-menngevolved (9, 15~J7) and show that ants and bees are suitable models for testing sperm competition theory. This is because males have no influence on the fate- of ej aGW ates after ma ling, and sperm storage has two distinct phases: a provisional one, varying across species from seconds to a few days, during which sperm competition may serve


Fig. 3. Own AG secretion does

not counteract the negative effects 0 n spe rm s u rviva I (mean J.. 5EM) of the AG secretion of other males in polyandrous ants and bees: A. melli/era, A. ech;natior,and A.

coiombica. Overall treatment E

effecl:5 were significant in <1U cu

three c"se5 (see re xt). 5 p e d fic .g. 70 111

con tra s ts were not significant>

for pure alien AG secretion :::i 60

versus a mixture of own and "

alien AG secretion (l ~ 0.00, df ~ 1, p:= 0.964 for Api$; X· = 0.05, df = 1, p = 0,823 for Atromyrmex; and X2 = 0.40, df = 1, P = 0.529 for Alta).

However, "Contrasts were. signif- 30 .J.._-"'==J'-----'-"""'""'=-L..-'=~-'-_""'"'__'__j!.=="'-_ ...... =!L......J

icant for pure own versus pure Apis Acromyrmex Alta

alien AG secretion (%~ = 4.86, melllfera eci?inatior cotomaice

df = 1, p = 0.028 for Apis; X2

= 6.75, df = 1, P = 0.009 for Acromyrmex; and ,/-.= 9.25, df = 1, P = 0.'(}02 for Alta), .and for pure own versus a mixture of own and alierr AG secretion (X; = 6.58, df"", 1, P = 0.010 for Apis; X~ = 4.62, df = 1, P= 0..032 for Acromyrmex; and x,~ = 9.89, df = 1. P = 0.002 for AttaJ.






A G s ecretio IT sou rce: [::!I Same male

o Unrelated male

~ Same + unrelated male,

female interests; and a much 1011g"er permanent phase during which queens are expected to maximize the survival of all sperm, It therefore seems reasonable to expect that selection on male traits that enhance ejaculate competition is proportional to the duration of prestorage ejaculate interactinn and to the ratio of inseminated versus loag-terrn-stered sperm:

Antagonistic effects similar to those reported here may have arisen in ether eusocial Hyrnenoptera that have evolved from single to tnllitipic queen-mating l6.25 27), and spennathecal fluids thai neutralize aggressive male AG compounds (compare to Fig. 4) may also be found in dwarf honeybees where sperm is directly trallsretTed to the spennathera (28), as in Atta (22). However, although phenotypic expressions of these sexual conflicts might be simil ar; tb eir indepcnden t ev 0 I u tionary origins suggest that the genes and patlrways involved are different (7). Clarifying such molecular mechanisms is. now feasible. as recent studies have shown that honeybee spennatbecal proteomes differ appreciably :OUIn male AG proreomes (17, 29 30), consistent With functional differences connected with operating when spenn is in a state of long-term dormancy Of metabolically active.

E ~

g. 80






Own AG Alien AG Alien AG

secretion secretion sscrettcn + spermathecal fluid

Fig. 4. Sperrnathecal fluid eliminates the I]eg.ative. effects of other males' AG secretions on sperm survival in A. colombica (mean ± SEM). The inset shows a dissected spermatheca 1 hour after having been artificially inseminated with Hayes saline. The first two bars differ significantly (·l = 15.04, df = 1, P 0:: 0.001), as do the second and third bar (X2 == 13.59, df = 1, P .c;; 0.001), but the bar on the does not differ significa ntly from the bar toward the left (X2 = .2.38, df = 1. P = 0 .. 123).

19 MARCH 201'0 VOL 327 SCIENCE

R efere n ce s and Noles

L G_ A. P~k~r, Bioi Re~_ (amb. Philo", So<:. 4~, 525 (1970). Z'. W. G. Eb€rhaJd, reroale f(Jl)lroi' SexU11/ Selection by Crypfi( Femole Choice (Princeton Voiv, Press, Princeton, NI, 1996).

3. l. W. Simmonl, Sperm Competitioll and tts fV/Jluffonmy Consequences In the Ins~,ts (Rr;ncetfH\ UI\1V_ Pre", Prln[eto~, NJ, 2001).

4. T. el1"pm~n, Go Arnqvist, J- Bamjham, L Row~, rrmds [(O/. [voL U, 41 (jOQ'.).

S_ M- D, Je~njolJs, M. Petri~, Bioi, Rev_ Camb. Pili/o<, SOL 75, 21 (2000).

6. w, 0, H, Hugj1e\, B. P_ Old,oyd, M. Beekman, F. 6. W, R.otniel;5, Science :320, 1213 (2008),

7- J.l [Iom(r;'mc, PII;w" rlan~ R: ~-I}C fprjd, Sfj: B 3~4, 31 ~1 (2009).

8. B. 1-i61ldobler, S. Bart;, in Experimenwl Behavioural cwloJ1Y anil SociDbiology, B, H 1i lldob ler, M. Un do uer, fd~, (GUstav Fischer Vertag, 5tutlgan and New vcrk, 1985), pp. B7~2'i7_

9. J.). Boomlma, D. J. (, Kr()na uer, i .. 5, Pfd ersen, in argollizmfoll of msea SocietiC!s: From @nome (-0 SocioCilil]pl1li ily, J. Go cia u l. F~well. E-rls, (HaNa rd Univ. PI""" C~lTIbri(lge, Mil, 20(9). pl). 3-25_

10. 1. ). Boomsma, B. liaer, I. Heinze, AnmJ, Rev. Enlomo/. 50, 395 (2005).

11. D_ K, AanM el al., Sclel1ce 326, U03, ,(2009)_

12. f. G_ Leigh Jr-, 1- Evol, BioL 23, 6 (2-010).

13, T_ Ch~pl"nan, I,. F. Liddl~, J- M- Kalb, M, F. Wolloer,

L. Partlidg~, nanue 313, 2111 (1995).

14. B, (oil', Behav. u« Sodobtoi: 12, 191 '(1983)_

15. 5. P. A de~ Boer el oi; Prn(. R. Soc. B 276, 3945 (20(9)_ 16_ 5_ P _ A_ de~ a 0 er, 1 1- Boomsma, B_ Bo er, BehoV_ Lcol,

SQtiobiol. 62; 1843 (2M8).

17- S. P. A. d~11 BO~f, J. I. Boomsma, B_ Baer, l. tnsea PhY'iioL '55, 538 (2009).

18, M~!eri.15 and me!hDd, are available a, ,uppmtlog material on Soen(e OnliQe,

19. L, Holman, Behov . .fcol Soc/obiol_ 63, 1579 (11)09)_

20. P. Villf>en, T. MUfilkaml, 1. Ii. Schultz, J.]. 6aom.lmi!,

Pt(Jc_ R_ S"m;, Be 269, 1541 (2002)_

21. ~. enr, l1pidologie 36, 187 (ZOOS).

22_ B, Baer, J, I. Bool11sm~, ]. Morpl104 267, n65 (2006).

23. S. B~ er, S. A _ O. Armitage, J. J. S()omsm a, Nafure 441.

M2 (2006).

24. A_ Rei(h~rd~ D_ Wheeler, Bei1Qv_ feo/. Sor:fobio/_ 38, l!l9 11996)_

2 5. ]. J- &J omsma, r. l- W _ Ra !nie~, Pilil as. rr~n~ R. S D(. [J 3,s~, 94'1 (1996).

26. ]. E. 5tr~>smann, IIl5eC/5 50c. 48, 1 (2001).

n _ R_ H. QDli", E. J. Fi~rdj~gstild, Ann. Zoal Fenr: 38. fb7 (lOOl!.

28. N, Koeniger, G. KoenigeT. S. Wung5iri, Apfdologie 20, 4~3 (19391_



2'1_ B. Baer, H. Euhel, N. L T.ylor, N. O'r{){)le, II. H_ Milt;.r, Genome iliol. IG, R67 (2009).

30_ S_ Boer, I. l. HfW(~wood, N_ L. Taylor, H. EubeL A. H. Mi lla r, Pro leom r.s 9, 2085 (2009).

3 ~ We th~n" G. N~eh rna iI a nd D_ R Na sh for, .,tj5ti<;al h~lp. l. Holman and D. R. Na,lh tor disOI»ion and comments, Wi_ StCirup tor collecting ant, in Ihe field, Wi_ King for wllEcri ng part of the. s pe Il1l vi a bilily do la lor non ~yb ees, the honeybee keeper> of Western Au>t(alia. (aeller Be5 of WA) lor pfovidirl9 the /l€Ce5lary bee material, the Smi[nsonian Tropical Resear(h InstitutE in Pan"m~ for fa{iUti~5 and IUllil!k 54PPO(t during fieldw()rk, and the 1\01 mi do d !-be; ona( de Anl bien!~ (1\ NAM) for i)sLl.i ng cotletri ng a nd expert p errni I> for lung us ·growi ng a nts, Supported by ~ grool from the Oimisn Nation~l R~S€afCh F()undatlon (to H B.) a nd a n AU, tr illia n R~ sea rch Cou neil Di,mvery Proj"D (DP QS] BIG7) ~nd Queen Elililbetn II Fellowship (Op 0770050) (to B.B,).

Supporting Online Material

\WI\Y .>Lience",~g.otgkgil eM tent/fu 11/3 2 715 972115 0 61 DC 1 Ma IE rials and Melli ods


16 November 2009~ accepted 17 february 2010 10. 1126f,(ji'nce_ '~84709

Patterns of' Diversity in Marine Phytoplankton

ArIdre'W D. Bartoll,l, Stephanie Dutkiewitl} Glenn Flierl,l Jason Bragg/t Mkhael ). Follows1

Spatial diversity grild,]enl:5' are a pervasive featlJre of life on Eatth. We examined il global

ocean circulation, bioqeoehernistry, and ecosystem model that indicated a decrease in phytoplankton diversity with increasing latitude, consistent with observations of many marine and terrestrial taxa, In the modeled subpolar oceans, seasonal variability of the environment

led to competitive exclusion of phytoplankton with slower growth rates and lower diversity. The relatively weak seilsonatity of the stable subtropica land tropical oceans in the gLo bal modeL enabled Long exclusion time scales and prolonged coexistence of muLtiple phytoplankton with comparable fitness, Superimposed on the decline in div<,rsity' seen from equator to pole were "hot spots" of en ha m;ed div!:!fsity in 50m e regions of enE:fgetic ocean circLllation, which refle~ted latera L dispersal.

10 both marine and terrestrial environments, many taxa exhibit a decline itt species di~ .. versity witb increasing latitude (1. 2). and this pattern 'bas important implioatioas for eCOSYSlem structure and function (3). 'The extent to which and why rnariue phytoplarjktQo may follow such patterns is not yet dear, although it has been argued that the biogeography of'microbes is governed by a similar sci of processes as for macro. organisms (1'1). There is some evidence of I atitudinal diversity gradients among certain taxa of marine microbes, including bacterioplankron (5, 6) and coceolithophorids (7, 8), althougb thegenerality

loeparu"enl of t~rth, Atmospheric" ~ nd Planetary Sciences, Massa~hu>elt<; Inllitute ofTedl[1o\ogy, 77 Ma<&lchulem- I\venllf', Combridg~, MA 02139, USA. 2Depilrtrrw~! of CiVil and Enviroml1 ental E ngi 11e€rl nq, Ma,sa dlUS!'tl5 r nsti Me ef Techno!· Qgy, 77 Massachu>eds Avenwe, Cal.nbridge, /WI 02139, USA

'To whom correspondence should be addressed. E·mail: adbar1on@mil,edll

+Present address: Commonwealth Stisntifir, and Industrial R e,learch 0 rga ni5ati 0 n (CSI RO) Pia n tin d ustry. (a n ben 0, ACT 2601. Austfilli~.

of ihese patterns, particularly in the open ocean, is, as yet, equivocal (9, JU).

In a recent study, a ihree-dimensional and timevarying global ocean circu1atiol}, biogeochemistry, and ecosystem model was initialized with a relatively large number (78) ofvirmal phytoplaokton types whose traits were assigned stochastically fi'Om plausible ranges of possibilities uo 12). The modeled phytoplankton communities "self assembled" according to the relative fitness of the phytoplankton types in the regionally and seasonally varying resource and predatory environment, The emergent pbytoplankmn populations captured the observed large-scale oceanie patterns in th 6 distri ibu tioo of pb ytop lankton biomass andcommunity structure, including the observed niche differentiation among ecotypes of the cyanobacterium Prochloroeoccus in the Atlantic Ocean (J 1).

Here, we have studied an ensemble of J 0 integrations of the global model, each member having a diffcrcnr, stochastically seeded selection of phytoplankton types, to examine and interpret

the emergent patterns of'phytoplankron diversity. In each of the solntious, after a decade of inc regration, a dozen or SO phytoplankton types aecount for more man 99% of the total global phytoplankton biomass. Others persist at low abundance or with limited geographic distribution, and some decline toward virtual eXMctiOfL Fast-growing "oppommisr" ~hytoplal1kton tend to dominate the biomass of the variable h(g'lJ Latitudes, whereas "gleaners' (those best able to survive OIl minimal resources] dominate the stable, low-latitude seas (12, 13). There is also a degree of local coexistence among phytoplanloon types, On an annual, vertically averaged basis, the phytop lankton diversity in the euphotic zone- (here assumed to be Q. to 260-111 depth) is lower in tbe polar and subpolar oceans and higher in tropical and SUbtropical latitudes. (Fig. lA). This meridional gradient is dearly seen in the zonally averaged view (Fig. 18) and is consistent with numerousobservations of marine and terrestrial ecosystems (1, 2), including the sparse observations of marine microbial diversity (5-,s)_ Supelimposed on the model's meridional gradient are "hot spots" of'bighest diversity, which are generally associated with regions of energetic circulation such as the western boundary currents, The Atlantic Ocean hot spots appear to be consistent with observations of increased diatom diversity near the North African and South American coasts (8).

The mechanisms for maintaining the diversity of life on Earth have Long imerestedecologists (14, j 5)" and the explanations for the meridional diversity gradient have been classified as historical, evolutionary, or ecologies I in nature (6, 16). Historical explanations invoke events and changes in Earth history, such as Milankovitoh cycles, in setting current species diversity. Evolutionsry explanations examine the rates of'.speciation and extinction and their balance through time (17, 18). These pmccsses arc not resolved in SCIENCE VOL 327 19 MARCH 2010



this model, yet diversity gradients are still apparent, Thus, we seek ecological explanations for the model diversity gradients, acknowledging that some real-world processes are not being considered" Niche differentiation, including seasonal succession, plays a role in determining the regional and seasonal habitats of phytoplankton types, adding to, but not fully explaining, the spatial diversity patterns (10), We find dispersal and temporal variability of the environment to be the most important ecological controls on phytoplankton diversity gradients in this model, whereas other factors are of lesser importance or not resolved (1 (f).

Resource competirion theory (12,13,19-21) provides a useful framework for illustrating the role of tempcral variability in the global model. Consider an idealized system witb a single, limiting nutrient, where the rate of change of biomass is determined by the balance bstweeo growth and mortality and the rate of change of the nutrient is determined by consumption by phytoplankton and its environmental resupply (10), At equilibrium in this system, the phytoplankton type with the lowest environmental nutrient concentrarion at which the growth and mortality are ill balance (designated as R*) (10) is. expected to outcompete other phytoplankton types over time (12, 2{}). This limit is relevant to the subtropical oceans, whichare characterized by a relatively weak seasonal cycle, and a strongly stratified, oligotrophic surface ocean, An emergent feature of the global model solutions was the coexistence of multiple physiologically distinct phytoplankton types with similarly low Rt- in the tropical and subtropical regions (Fig. I C) (J 2), at least for the time scale of the model irrtegratiens, Because the R* fur ead, phytoplankton type depends on imposed pbysiolcgical characteristics and mortal It;)!, there are, in theory, many possible combinations that can achieve the same maximal fitness (lowest R"). Moreover, the emergent, coexisting community of physiologically distinct but R"equivalent organisms is consistent with studies of laboratory populations: of manipulated bacteria (22) and the hypothesis that such a mechanism may be important in maintaining the diversity of marine phytoplankton (2n This model outcome itself points to a possible explanation for enhanced phytoplankton diversity at lower latitudes and echoes the neutral theory of ecology and the hypothesis of ecological equivalence (l3).

We analyzed the diversity dynamics within me idea lized resource competition framework for the special case where all phytoplankton types have identical R*. In support- of the emergent pattern in the global model, the idealized simulations indicate that the relatively steady environmental conditions in the tropical and subtropical oceans enable the pro longed coexistence ofrnsny phytoplankton with equivalent fitness (equal R*) and enhanced diversiry (Fig. 2M (/0). However, the oceans 'are constantly perturbed by atmospheric forcing and internal physical phenomena



80r--° 0_.5~_=i:=__-1----1___,.5.___::_----------___,








Threshold Inde;x





10-2 10,-1

log R" (mrn.ol N m-3 )

fig. 1. (A) Diversity of modeled phytoplankton types in the uppermost 260 m, averaged annually across 10 ensemble members. Diversity is defined as the number of phytoplankton types comprising greater than 0.1% of the total biomass. (8) Zonal mean diVersity, as well as the Shannon Index (10), lor the map .shown in (A). (el Annual mean R'" (small black dots) of all phytoplankton types with a concentration above 10-lZ mmol N m--3 along a meridional transect through the.Atlantic Ocean at .20oW in an idealized global model with a single limiting nutrient (12), The large red dots show the R* for only the most abundant typE' in each latitude,

across 11 vast range of spatial and temporal scales, fntroducing a time-varying, periodic nutrient source to th e idealized simul atiou S eventually 1 eads to competitive exclusion of all but the single phytoplankton type that grows. fastest under optimal conditions (Pig. 2B), even if the equivalence of R*' isimposed, The slower-growing phytoplankton types need a higher time-averaged nutrient concentration to compete with the faster growers and are excluded over time (10), Environmental variability creates a competitive structure such that the number of extant phytoplankton types can be reduced through competitive exclusion. indeed, in the higher latitude, strongly seasonal marine eovirorunents where the global model solution exhibits lower diversity (Fig. 1), high growth rate, and not low R*, is the, most appropriate measure of organisma I fitness (12).

Using tbe idealized, experimental system, we investigated a range of natural frequencies and amplitudes of variability in nunient supply and defined the time taken until Due phytoplankton type accounts for more than 90% of lhe total biolU11SS as the time scale of competitive exelusian, or't("I;t. This time scale can exceed a thousand years when the envircorrrental -variability has either short (hours to days) or tong (annual

and longer) periodicity (Fig. 3), In contrast, when the environment varies with a period of months. competitive exclusion occurs within a few years or less. Large amplitude variations promote rapid exclusion" whereas small amplitude variations allow fur extended coexistence. Therefore, in the subtropical and tropical oceans, where. seasonality is relatively weak, we expect the time scale of competitive exclusion to be long (centuries OJ" more, which is long relative to the length of global model integrations) for pbytoplaokton types with equivalcntR*' (Fig. 2A}. In contrast, the subpolar and polar oceans are sub j ect to strong season al variations, including changes in the mixed layer depth that regul ale lightand nutrient availability. Here, opportunism is' favored and the exclusionary pressure by the fastest growing phytoplankton 011 those with lesser growth rates is strong (Fig. 3). The exclusion time seale here may 'be as short as several years, and the long-term coexistence of many pbytoplankton types is not sustained. 'Vtuiabili ty in growfl 1 rate, 111' hieh is sensitive to changes in temperature and light, led to similar results (10).

The time. scale of competitive exclusion is set by the character of environ mental variability, but local diver-sity in the global model is ultimately a

19 MARCH 201'0 VOL 327 SCIENCE

z 0.031
E 002 Y·e·ars



o 200 400 600 800 1000



Growth Rate for Phytoplankton Type j ( day -1 )

-0.6 -1.0 -1.5-2.0 -2.5

Fig. Z. lI.bundances for five hypothetical phytoplankton types with a constant (A) and periodically varYing (B) nutrient source ((j)~1 = 365 'days, and the amplitut!e is 0.5). Colors r~present phytoplankton with different maximum growth rates (~l). Data in (B) were annually averaged.

Fig. 3. The relationship between exclusion time scale heE) and Ihe period (00-\ solid line) and amplitude (A, dashed line) of the external nutrient source in th .. idealized model (lO). We considered a system 10 be in a state of competitive exclusion wilen one species comprises greater than 90% 01 the total biomass.




W o

P 500

balance between the removal of species by exclusion and the replenishrnent of phytoplankton types through physically mediated dispersal (17). The rate of'lcng-range dispersal of phytoplankton types in the ocean can he fil st in swill curren Is (weeks to months) but is generally slower (dec. ades to cenruries) within and between ocean gyres (24) .. In the high latitudes, exclusion is generally rapid relative to dispersal and mtergyre exchange, and diversity is consequently lower, In the tropical and subtropical oceans, the exclusion time scale is typically long relative 10 the rcdistribution of phytoplankton by dispersal. Here, a higher diversity of similar R* types. can be maintained,

In the "hot spots" of highest phytoplankton diversity, ocean dynamics, such as lateral advection and stirring due 10 planetary waves, mix organisms 110m different habitats. For example, the elevated diversity in the region of the Gulf Stream reflects the rapid poleward and eastward advection of organisms adapted to tropical and SUbtropical enviroumeurs, as. hac; been observed (25). As the boundary current transports away rhe subtropical communities and their environments, the transported waters are, mixed and (heir-phyroplankton intermingled with locally adaptodorga-

Amplitude of Nutrient Perturbation (A)

l500ITO~O_._1 __ U_.2 __ U,.~3 __ O_.4_· __ O_,5 __ 0_,_6 __ O_,7 ~~g __ O~.9 __ ,

-- Rel"tiQo ot 1:c" to Period (r~.1)

- - -. - - Relation ohce: to Amplitude (A)




ciationand climate change, should also be explored, Further laboratory or mesocosm experiments migbt be designed to address the potential for coexistence of'microbes with equal fitness. New, molecular approaches (5, 6) will enable efficient, systematic surveys. in the near future, and we suggest that a targeted survey of phytoplankton diversity (prokaryotes-end eukeryotes), crossing fi"om a subpolar regime, across a boundary current, dispersal-dominated region (the model hot spots), and into the interior of'a subtropical gyre, could provide a valuable test of the hypothesized patterns and mechanisms that emerge. from tills study,

Rilferen(e~and Notes

1. D. J. Cutrk Am. Nat. 131, 27 U991},

2. H. Hillebtll~~, 11m, /'ItJt lU, 192.(2004)

3. R. Ptacnik el al" PfOc"Nati. Mad, Sci. USA. 105, 5134 (20.0.8).

4. I. 8. Ma rtiny ~f oi; Nat f/Gv.Microbiol. 4, 1.0.2 12006l.

s, T. pommi€ r et at. Mol, Ecol: 16, 867 (2 007),

6. J A Fuhrman e-t al. Pf(Y(. Natl, Ikad. sti. U,S.A. 105, 7774 (200.8).

7. s H\i~jo, H Ol\ada, Mkropal~olltol(j9Y 20, 20.9 (1974).

8. P. C~'mei\o et ot. P.M. Nail. Acad. Sci. JHA. t05, 20344 (2008).

9, P. Ctrmeiio e/ at; timno: OfeQllo[jf_ 53, 112 (zoas). 10. Supporting material, is available on Science. On!iQe.

1 L M.]. Poll oW';, S. ou tijewiCl, S. Gran t. S. W. Ch isholm, Science U5, 1841 (2o.o.J).

12. S, ou t kiewicz, M. J, fallow!, J. IS. Bragg, Global

8iaf}fOchem. cycles 23'. 15640. 17 (2o.o.9l.

13. ]. P. Grover, Am. Natl::! 6, 771 (J.99Ql.

14. G.~. Hu((ili",m'l, 4m. NM 93. 145 (H~9).

15. G. E. Hutchinson, Am. Nat. 95, lH (1961).

16. G. G. Mfttelb.Jch ~I ai., [r.oL lett. ;to, ns ('200. n.

17. R, MacArthur., E. Wijson, The l1Jeory of f5iolld Bio!Joo[jmphy (Prin(eton Univ'ersily Presj, Princeton, N], 1967).

ie. A. p, AlI.eo, J. f. Gillooly, V:J,II. S;wage,]. H. BroW." no:

Natt Acad. fie i. us. A 103, 913 o (2006).

19. F. M. Stewarl, B. R levin. Am. NtJ~ ~07, 171 (1973).

20. D, Tilman, Ecology 62:" a02 (1981),

21- P. G. F.lkow,l:.i, M. ]. O!jver, Nat Rev. Mirmbfo/. 5, au (2007).

22. S. it Hansen, 5, p, Hubbell, Scie[1(e2iJ7, 1491 (1980).

23. S. HUbbell, The Uilifillli NeUlrfJi Theory Of Biodiversity o/lr,J Bjogoogf~phy (Pri~(eron Un,ver5it)' Pres" Prin,e1an, Nl, ZOOI).

24. A. C. M~ftiny, A P. K. Ta;, D. veneziano, F. Prime~u, S. W, (hl,holm, EnViron. MicrobiaL 11., 82] (2009]. 2.5. K. K. (ilv~nd"r'B."es, O. NI. KatL s. W. Chisholm, tseep-seo Res. 48, 23M (Zo.OI).

26. M. ii, L~ibold., M. II, McPeek, fcolagy 87, 13<}9 (2006).

27. A.D.~ .• S.D., ).B., and M.J. F. are grateM for SUPPDrt from l~e Gordon and Betty Moore Foundiltio~', MMfne Mjmlbiology hliiiaJlvB, MJF. and '5. O. are ako grateful fo~ ~upport. fro m NASA a nd NOM.

o 50 100 150. 200 250. WO 350 Period of Nutrient Perturbation (days)

nisms and evenrualjycuteempeted. The exclusion time scale here is long relative to the advective time scale, and the transported population contri butes to the local total b i OUID.'lS and di versi ty (Fig. IA). Similar processes ate responsible for the enhanced diversity in the tropical Eastern Pacific. In contrsst, the energetic Antarctic Cireumpolar Current region has low diversity because the neat-zonal circumpolar flow acts as a barrier to, 81~d not an agent of, C01TIm1JI1.icatlon between marine provinces.

Altbougb tbe global model presented here is a simplified system, the emergent patterns of diversity show features generally consistent with the sparse observations of marine microbial diversity. The model's diversity patterns primarily reflect a balance between dispersal and CQmpetitrve exclusion, with the latter modulated by en virOIl mental variability. Both n eurral co ex istence and niche differentiation play important roles ill regulating the diversity and biogeographies of model phytoplankton (26). Such a modeling approach might be extended 10 explicitly reflect a broader spectrum of marine organiSt1$, sucbas heterotrophic microbes: and zooplankton, and enable comparison 'with more observational data sets. The roles of' other processes, including spe-

Supporting Online Material

I'/WI'I. sdencernaq Drgll;gllmnt entlfu 1Vsc: len,e.1lS 496110(1 SOM text

Fig'. 51 !O S4


19 ~ ovemb er 200.9; ~CC€ pled 12 Februa ry 2010 Publi'hed online 2S FebilJary 2010;

lO' 11~'\Men(e.ll84961

Includ~ itris iolormation when dting this pa per .. SCIENCE VOL 327 19' MARCH 2010



Unicellular Cyanobacterial Distributions Broaden the Oceanic N2 Fixation Domain

Pia H. Moisander,l" Roxanne A. Beinart,lt Ian Hewson/* Angeli€que E. White,2 Kenneth S. lohnson/ Craig A. CarLs.on/ Joseph P. Montoya,s JXlnathan P. Zehr~

Nitrogen (N~)'-fixing microorganisms (diazotrophsl are an important source of biologically available fixed N in terrestrial and aquatic ecosystems and control the productivity of oligotrophic ocean ecosystems, We found that two major groups of urricellular N2-fixing cyanobacteria (UCYN) have distinct spatial distributions that differ from those of Trichodesmil1m, the N2-fixing cyanobacterium previously considered to be the most important contributor to open-ocean N2 fixation. The distributions and activity of the two UCYN groups were separated as a function of depth, temperature, and water column density structure along an BO~O-kilometer transect if) the South Pacific Ocean. UCYN group A can be found at high abundances at substantially higher latitudes and deeper in subsurface ocean waters than Itichodesmium. These findings have implications for the geographic extent and magnitude of basi n -sca I e ocean ic !II 2 fixation rates.

Nitr.Qgen (Nz}-fixin. g ,111:icroorganlsro.s (diazntrophs) are an important SOLU1le offixed N in oligotrophic ocean ecosystems (I, 2). Biological nitrogen (N0 fixation is catalyzed by the enzyme nitrogenase and provides about 100 to 150 Tg N per year to the. open ocean (3), about half of global biological N2 fixation (4). Large uncertainties remain iu the estimate, and N:;>. fixation Tates derived from gecchemics! evidence have not been entirely accounted for by existing biological data (4). TI1e organisms that u;<: N2 play a central role in providing N to SUpp011 primary productivity and the vertical downward flux of organic matter ("export") to the deep ocean that sequesters carbon from the atmosphere (5). The filamemcus cyanobacterium Trichodesmium W;;IS believed to be the most abundant and active oceanic Nr fixing micronrganism (6) until the discovery of two unicellular diazotrophic cyanobacteria [UCYN group A (1JCYN-A) and Crocosphaera watsonii {group B)], wbose abundancesand N2 fixation rates can be equal to or greater than those of Trichodesmiun: (7 10). UCYN-AandC \'Va/sonii are abundant and widespread in tropical and subtropical oceans (7) but are more difficult to visualize and quantify than Trichodesmium, and much less it; known about their distributions and growth requirements. UCYN-A is less t}l!;lll I I.U1] in diameter in size, has dim all tofl uo-

] D~partfr1~llt 01 OCe.3n Sdences, University 01 California Santa Cruz, 1156 High Street., Santa Cruz, CA 95 064, USA. '(olleg~ of ocea nic and A Ul1 ospheri c Sciences, 0 reg on 5 tate Uni~ VEfsity. 1.04 COAS Adll1ini~lrJti<ln Building, (orv~l[is. OR 97331" USA. lMonterey Bay Aquarium ReSl'ilf(h Institute. 7 700 Sand hoi dt Road, Mo,s land ing, (A 95039, USA. , Depa rtment of Ecology, Evolutiun and Mil tine Biology, Univ~f~ity of Calilonlii3 Santa Barbara, Santa Barb<i,f.3, (A 931O~, USA.'Sdl001 of Biology. Gt-orgia Institule of l~(hno\Qgy, 310 Ferst DriVE, Atlanta, GA 3 O~ 32, USA.

'To whom co rrespon d en r;e ,hou I d be add ressed, [0. mai I: pmoisafid@ucsc;.€du

tPre>ellt address: Deportment of Organismic and Evol!ltion~ry Blol.ogy, Hilf\l~rd UniversitY, c"mbridge, 1M 01~3a. USA; tPre5eni address: D~partn1el\t of Mkrobiology, 403 Wing Hall. Cornell Ul1iv~r5ity, I thaca, NY 14353. USA.


rescenee (f f), and has not been cultivated, whereas C. watsoni: is a cultivated phycoerythrinrich eyauobacterium 3 to S !Ull in size that can form small colonies (below we use Crocosphaera to Include. all C walsol1ii-like cells) (11). UCYNA is unusual in that it does not have genes for theoxygeo-evolviug photosystem Il 01' the carbon fixation enzyme rubisco and thus appears to btl a pbotoheterotropb dependent on organic carbon (J 3, 14), whereas O'tl('osphaerq has the phctosyntbeti c machinery typica I of cyanobacteria (1:11.

The salient differences iil the size and pbys'iology of the diazotroph groups (UCYN-A, Crocosphaera, and Dichode.smJ'illll) suggest that they occupy distinct niches and may differentially affect primary productivity snd the export of C and N. Yet in basin-scale models of oceanic N2. fixation, it is generallyassumed that a rather uniform set of environmental conditions equally eontrols the abundance and distribution of all diazotrophs (16, 17).. N2 fixation is typically formulated .as a fimction I'If SOlTIe combination of temperature, wind speed, water column stratification, photon fluxes, and the ratio ofN :P; reflecting the wann (~25"C), o~g.otrophio conditions generally held to be requited tor Tnchodesmium blooms (6, 18). Under Nfimhiag open-ocean

conditions, diazonophs compete with each other and other plankton for macro- and micronutrients, vitamins, and ather growth factors, especially iron, needed fur nitrogenase (J 9)., Little is known ahout how the physiological differences of oceanic diazotrophs are manifested in their global distributions and activities. This information is essential to fill the large gaps in global N2 fixation estimates (4)c We investigated distributions and activities of the major oceanic diazotrophs in the oligotrophic western Sauth Pacific Ocean, in parallel "villi measurement of an extensive set of environmental parameters (1 0).

Water samples were collected between 15" to 30°5 and 155"E to 1700W {Fig. l ), where slJli!l ce water temperatures decreased from th e. northernmosr to the southemmost stations, consistent with latitudinal differences in radiarive forcing. Diazotroph abundances were quantified by quantitative polymerase chain reaction (qPCR) of the nijll gene (euecding the iron protein in the nitrogenase enzyme) (.10). which cotrespends to cell abundance because there is 011e copy of nill-[ per genome in all three groups (VCYN-A, Crocasphaera, and Trichedesmium) (13). We expected to sec similar geographic distributions in abundance of all diazotrophs, reflecting the N-limited characteristics of the oligotraphic ocean. Unexpectedly, there was a transition from Crocosphaeradominated communities in w3J1n surface waters in the north to UCYN-A-dolninated communities in cooler waters in the south (Figs. 1 and 2) .. and temperature was the most important factor correlating with these distributions (table S I). This evidence shows that there is vertical and horizontal partitioning between populations of UCYN-A and Crocosphaera, UCYN-A had a maximum abundance of 2.2 ., 106 mIN gene copies per liter at the souibernmost stations. UCYN-A was. also detected at most other stations but deeper in the water column (Fig. 2), Crocosphaera was found at very higb abundances in the northeastern part of the study area, with a surface maximum at station 21 and subsurface maxima at stations 25 and 26, Peak abundance of Ctocosphaera (8 x 106 nijH copies per liter) at a depth of 37 In at station 25


26~ e



248. ~

e F


fig, 1. Sampling locations superimposed GIl a composite sea surface. te.mperature plot (March to April 20Q7) for the study area.


'is. possibly the highest Crocosphaa» cell density published to date. The UCYN populations were active because the 11i/1-1 gene W'dS expressed (table S2) and N2 fixation was detected by using 15N 2: uptake method. Maximum N 2 fixation rates were 0,026 mnol liter-I hour"! at station 4" dominated by Cnscosphaera and with little or no UCYN-A, and 4.5 rnnol liter-I hour-I at station 10, which was dominated by UCYN-A with little Crocosphaera or Trichodesmium observed, The distributions of fhe unicellular N2- fixing cyanobacteria differed from those of Trichodesmium. The latter had. surface blooms of abundances greater than 106 ni}li copies per liter at several stations in the northeastern part of the study area (Fig. I and table Sl), The distinct diazotropb distributions indicate that UeYN-A, Crocosphaera, and Trichodesmium biomasses are differentially affected by enviroomental conditions and controls.

The abundances of UCYN·Aand Crocosphaera in the. surface layers con-elated with water temperature (pigs. I and J and table Sl), but the diazotroph abundances were inversely correlated with e~ch other [r2= 0.397,P= 0,000, number of samples (11) = 96, cubic regression]. Abundances were weakly 01" not correlated with salinity; oxygen saturation: dissolved organic carbon (DOC); soluble reactive phosphorus (SRP); nitrate; nitrite; total dissolved nitrogen (TDN); chlorophyll a; fluorescence; and abun-

dances of'picocyenobacteria, picoeukaryotes, and other diazotroph phylotypes (table SI), Maximum abundances were observed at 24"C and 29"C for UCYN"A arid Crocosphaera, respectively (Fig. 3). Collectively. these data. suggest that UCYN-A has a lower temperature optimum than Croco..~phaertl, This is consistent wi1h datil "from the North Pacific, where highest UCYN-A and Crocospbaera abundances were detected at about 23.5°C and 26"C, respectively (21), and in the eastern Atlantic, where high UCYN-A abundances were observed at water temperatures ranging from 19° to 24"C (9). In addition, observations of nij/i transcripts of UCYN-A have been made at 12" to 19CC (22. 23). whereas Trichodesmium is thought not to be active below 20GC (6). Global distributions of UCYN-A and Crocosphaera from all known data indicate that UCYN frequently grow in waters with lower temperamres than, are considered necessary for Ttichadesmium blooms (Fig, 4). ill particular, this report shows that the geographic range of UCYN-A extends beyond the 25°C isotherm, indicating that previous regional or global N2 fixation estimates based on distributions of Trichodesmium have been underestimated,

The different depth distributions of UCYN-A and Crocosphaera indicate they might be adapted to differeet light intensities, similar to the high-Iight and low-light ecotypes of the nondiazotrophic

D r--------------, , - I


Fig. Z. Horizontal and vertical distributions of diazotrophs in the study area. (Al UCYN-A, (6) Croco5phaera, and (e) Trichodesmium spp. abundances [l0910 (niJH copies per liter/] with station numbers and sampling depths indicated. M!1ximum abundanre (p) and depth of maximum abundance (El of UCYN-A and Crocospnaera and (F) mean depth (±5D)of maximum abundance of UCYN'-A, Crocosphoera, and Trichodesmium spp, {one-way ANOVA}.



cyanobacterium Prochlorococcus. The peak abundances of UCYN-Ao_ccU!:ted significantly deeper than those of Cmcosphaera or Trichodesmium [P = 0_000, .11: = 13, one-way analysis of variance (ANOVA)] (Fig. 2). The. abundance maxirmnn ofUCYN-A, observed from the surface to 50- to 75~m depths at the southernmost stations, was centered on a weak and broad density fron: that characterized the entire study area and was dri ven by a latitudinal gradient in surface 'Iva ter temperannes, Moving north, peak abundances of UCYN-A at each station followcd approximately the 23 and 23.5 isopycnals (kg m -3) (Fig. I and fig. Sl); thus, peak. abundances were tound deeper in the 'water column in the northernmost stations. The depth of maximum abundance of UCYN-A was therefore positively correlated with sea surface temperature ('? = 0.390, P = 0.000, n = 23, linear regression), suggesting UCYN-A may grow deeper ill areas with warm surface waters. The different depth preferences of the UCYN groups may reflect their I1gbt, temperature, and nutriti onal requirements.

Intriguingly,. the peak abundances ofUCYN-A coincided with elevated nitrate concentrations 1.11 surface waters (fig. SI) that appeared to be related to shoaling isopyenal surfaces, indicating that there had been a vertical input of nutrionts and trace elements from deep watec, Although UCYN-A does not have genes fur assimilatory nitrate or nitrite reduction (14),. our data show fhat UGYN-A is abundant in waters where nunients were recently entrained. an observation consistent with recent reports of high abundances ofUCYN-A in uutricot-enriched estuarine and coastal Waters (22-24) and eddies (25), unlike what is known about Ttichodesmium (6). These 'UCYN-A-dominatedwaters may have remnants of noo-Nj-fixing phytoplankton blooms, stimulated by vertical mixing, A large area with elevated surface water chlorophyll a concentrations, indicative of increased phytoplankton biomass, was observed in ocean color satellite data south of the study area, associated with the general position of the Tasman Front (26). These observations support a link between high phytoplankton biomass and high UCYN-A abundance (fig. S2), which is expected on the basis of recent UCYN-A genome information suggesting that it requires a dose association with other organisms (13, 14), DOC concentrations followed meridional. trends previously described for the South Pacific (27) and' were lower in concentration in the surface where UCYN-A abundance was the greatest (Fig. 1, fig. S l , and table S 1). Although the reduced DOC concentrations at the location ofUCYN-A maxima are eonsisteot with recent vertical mixing in the area, heterotrophic microbial activity may contribute to tins trend and requires further study. The differential distributions. ofUCYN·A and Crocosphaera observed in the water masses, mixing in the area, and traced by temperature, may have beenaffected by difference,'; in the chemical composition of these water masses, SCIENCE VOL 327 19' MARCH 2010



The magnitude and geographic extent or oceanic N2 fixation determine the rate of primary productivity and vertical export of carbon in (he oligotrophic ocean. Most previous estimates of global N] fixation are dependent 011 distriburions or faorors that comrol the growth of Trichodesmium () 6, 17), TIle results of'this study show that actively N:!-uxing populations of the two major oceanic UCYN groups have broader latitudinal distribution than Trichodesmium, analogous to the global distributions of strains and ecotypes of

.... :1

ihe nou-Ns-fixing oceanic cyanobacteria (28), Observations of distributions ofOCYN-A at relatively high latitudes ;md in coastal waters geographically extends the oceanic regions where N~ fixation can occur. which contrasts "virh previous paradigms based on the temperature range [Oi' Trichodesmtum. 111e data show that temperature is a rna j O.t uri ver 0 f the diflerenti al di stributions for different taxa, although additional factors, such as non-Nj-fixing phytoplankton blooms, may be Important fur the growth of UCYN·A

B 8




H ~ ~ ~ U ~ n M U ~ W n ~ ~

T"II\p"' .. lllu.l"C~ Teo:nperah~ I"C)

Fig, 3_ Relationship of unicellular dazotroph abundances [10910 (nifH copies per liter)] and tempNatur.e. 1ft) U(YN-A, (18) Crocosphaera./ = 5.13 - O.1754x + O.063Bx" - O.0124K describes the nonlinear relationship betwf:!en UCYN-A. and temperature and f = --5.16 + 0.36lx the linear relationship between Crocosphaera and temperature.


Fig, 4. Global distribution and abundances of (A) U(YN-A and (8) Crocasphaem unicellular cyanobacteria compiled from all known published literature. The maximum value is shown for sites where multiple depths were an;;llyzed. LatHudinal1BO( and 25°( isotherms [average values of bimonthly means from July 2002 to J lily 2009 from NflSA Moderate Resolution Imaging Spectroradiometer (MODIS)] are indicated with dotted and dashed lines, respectively. Literature sources lor cyanobacterial distributions are listed in the supporting online material, EQ indicates equator.


Unicellular Ns-fixing microorganisms are much more. difficult to detect than larger diazotrophs such as Trichodesmium and may thrive in different geographic regions. In order to better consnain estimates ofthe N inputs to the global ocean via Nz fixation, it will be necessary 10 adequately represent the. activity of unicellular Nz-ftxing microorganisms ill ecosystem models"

Referen(es and Notes

1_ P_ Falkowski, Nature 387, 212 ~199n. 2. 1. iyrrdl, Nature 400, 525 ~1999).

3_ N, Gr~ber, 'i" Nitrogen in tlte Manirt Environment, D. G. Capone, E. J. Carpenter, D. A. Bronk,

M. R, MulhoUa nd, Ed I. !Aca demic Pr QII a nd Hsevi er; Burlington, MA, 1008), pp .. 1-49 .

4. J. N. Galloway et at. Biogeochemistry 70, 153 (2004).

5. D_ Karl et ol, NafUfI! 388, 533 ~1997).

6. D·. G. Capone, J. P. Zehr, H. W. poerl, B. Bergmal~,

E. J. Carpenter, Science 2.76, 1221 (1997).

7. 1- P. Zehr et al., NoM!? 412, 635 (ZOO)}_

a, J,~. ~o~tlWa et al., Na/ure BO, 1027 (2004).

9, R_ J. l,;mgIGi), D_ Hummer, J L~RG(he, Apt)! Pnv/rort Mi[fobiol. ts, 192 2 (2008).

10_ M.]_ (hur(h, S_ D, Je~~in~, 0 M, l<ilr~ J P_ Z~hr, Aquat Mi~r9b. Em/. ~ 8, 3 (2005)

U. N. l. Go~bel, C. A. £dward~, B. ]. Carter. K. M. Achme~, J. P. Z~hr, J. PllywL 44, 1212 (2008).

12 _ E. A. Webb, I_ M_ E hrenrei rh, 5_ l. Brown, F _ W _ Vatu)s,

J. B. Waterbury, Pnviroll. MicrobioL 11, 338 (2009). 13_ 1- P_ Zehr et ai., Sctence ~2Z, 1110 (2003).

1..4. H.J. Tripp et al. Nature 464, 90 (2010).

15 - J. Pc Z ehr, 5, 11" Ber\(h, E _ A, MGnG raqon, l. MoC;" rren, ~_ F. Delong, Pror:, Natl. Aw,d. 5(' U.S.A, 104, 17807

(2007). .

l~. R_ R, Hood, V. ], (ol~l, D, G_ CaP0Ile, J- Geaphys. Re5. 109, (06006 (2004),

17. C. Deutsch, J. L. 5ofmierlto, D. M. Sigman, N_ Gruber, J. P. Dunne, Nature 445, 163 (2001),

18_ A, E. White, Y_ H. Spill, R, M_ letelter, J-GcQphy,. R~. 112, (120(16 (2007)

19. $. A. 5aijudo.'Wilhelmy e/ ot., Nature 411, 66 (ZOOJ.).

20. Moterial5 and method! are av~ilabll> as supporting materia I G n 5c ience OJilin eo

21. M. J. Church, K. M. Bjofkm~n, D. M .. Kar~ M. A. Saito, I. P. Zehr, LimnoL OGeanogr. 53, 63 (200S).

22. 5_ M_ Short., 1- P_ Zehr, EnViron MicrobiaL 9, 1591 (2007)_

23. J. A. ~e~doba, R. A. Fcs M, C. Sakamoto, J. P. Z ehr, K. 5_ J 0 hnson, Limno/. OceanDgr, 5 Z, 1317 (200 7)_

24. A P. Ro:el", J. A. Gilbert, B .. A. KQUy·Gerreyn, Mar. Ecol.

P,-Oll- ser. 374,. 7'(2009).

25 M- 1 ChoJrth et 01_, Global Biagel'l(l)em. [jlcleS 23, GB2020 (2009)_

26. M. f. Baird et ai., Deep-Sea «es. I 55, 1438 (2008).

27. u, A. H a melt, C. A. Ca rlson, o. 1. RefWta, R. Sch Litzer, Oceanography ll, 202 (2009).

28_ Z_ I. John,Q~ et ai., Science 311, 1737 (2006).

29_ We thank R. P~M, (_ S~kamoto. 1. Cute, N_ Pereir~,

M_ turna I, B _ (a rter, M _ Ochia], K_ tendon, "- Pre> ton, M. Hogan, a nd personnel 0 n boa rd RN Kilo MOd na 10 r tedmka l assistance and C. Edwa rds for he Iplul discussions, lhis ,tudy was su pp orted by NSF-Oi"'; lion of Oma n S ciences (OCE! ~gralit 0425363), NSF Emerging Frontiers Program (Center for Mktobial Ocwnogr..phy: Resea''ch arid Edu,,""or\) ( 0424599), arrd Ih~ Gordon ,md Betty Moore FouMation ior 1.P.Z., and N'iHiu (grant 0425583) for J.P.M, A,E.W. was supported by NS,-OCE

(gra nt 0623 S 96, R. teteli er, principal i nV~11i gator).

Supporting Online MaleriaL

I'IfWIV. sdcncema g. o'g'cgi'wi1t~n IIlu W,oenc e.1Ia 5 4 681D( 1 Me ten a ls and Metho ds

Fig~_ st and 52

TaoLe, 51 to S3


3 DeceMl) 2009~ accepted 17 february 2010 Ptlblished online 2S Febru;)!}' 2010; 10.11.26/"Ie,n<:e.1185468

lnrlude this lnfcrmation when r:i~'lg tllil paper.


Sign up to vote on this title
UsefulNot useful