Volume 329, Issue 5861 COVER DEPARTMENTS Laser-surveyed topography ofthe elevation in a 255 Science Online Pennsylvania landscape, showing evidence of 257. ThieWaekin Science sediment filed ealy American millgonds along 259 Editos” Choice astream valley Tertaces around the steam 260 channel descend stepwise from upstream to 263, donsieam (dark purple, ight purple, pink, 265 ‘each associated wth a dam The deeper stream 349 incision int the to donnsrear terraces is ‘due to more complete dam breaching, See page 289, ae EDITORIAL Image: U5. Geological Suey (UDAR deta and preliminary processing); Michael Rahnis 258 Esozping is oa ‘analysand generation by lean sce crore a " Norman Neureiter NEWS OF THE WEEK LETTERS Uncle Sam’ Biomedical Archive Wants Your Papers 266 Conservation nth Sense Ltr etal 281 Setelit Company Offers Eat Observing Researches 267 Scientific Meetings: Worth Attending A Nutt ae Scientific Meetings: Cel In Instead Ror Putting 3 Human Face en Energy Usage F Burs or Neurons ds Stayin ane 9 : Pe nay el Song Seyi s Fair Game for Chimpanzees £. Visalberghi and J. Anderson SRE: badd Response K, Jensen, J. Coll, IA Tomasello New Dark Wetter Map Reveals Where Galaxies Gambel 270 Polynesians Took the Expres Tain ThroughhMelaresia 270 BOOKS E7 AL. to the Pacific Cool It The Skeptical Erwironmentalis’s Guide to 285 Most Mave Black Hoe Conims Reatvty ules 271 Global Warring the Unverse 8. Lobo, reviewed by WF. Ruddnn Isolated Tribe Gives Clues to the Origins of Syphilis 272 (US, and U.K. editions) ° Caledatng e's Death TOIL WHO Study BaGs 273, POLICY FORUM he RS lone Estimate Aging nkastvcre and Ecosptemestoration 286 NEWS FOCUS MUL Dope etal. Gone Tess for Pachitc Rik Pare Researchers 274 { ‘Hoping for a Glimpse of What's Abead PERSPECTIVES Seeking the Roots of Ritual 278 Organizing the Source of Memory 288 = Just Deit Cal the Garden of en EA Grove Vv ion Continues to Suprise 269 CR ODel ond. ownsey pr. 309 ‘Managing Coastal Wetlands: 290 ‘Wolo and. Fx Dreams of Natal Steam a D.R. Montgomery Probing Quinum Magnetism vith Cold Atoms 292 Metevensten and A Samper CONTENTS continued >> wawescioncemag.org SCIENCE VOL319 18 JANUARY 2008 249Science SCIENCE EXPRESS MOLECULAR BIOLOGY | Shared Docking Mott ia TRFL and TRF2 Used for Differential, Recruitment of Telomere Proteins chen eal. ‘wo sitar members the protein complex that protec the free ends of oosones hae sin binding sites for ober complex mame and accessory pots. Jo 6icienee. 1151804 CELLBIOLOGY Differential Regulation of Dynein and Kinesin Motor Pro BR. Din, .L. Ross, YE. Goldman, EL. F Holebaur The moter protees een ane kesh th encunter te protein tas they mowe slong the microtubules the lormer eras ection, wberesthe later etches, TOALI26icience. 1152993 is by Tau BREVIA PLANT SCIENCE Arabidopsis CLV3 Peptide Directly Binds UV: Ectodomain M. Ogowa, i. Shinchora ¥ Sakagani, ¥ Matsubayashi Petes tht maintain te stm cls he shea apa maritom ‘of Arabidopsis at by Ding 10 Ue exraclila portion ofa receptor-siexnae 294 wwescioncemag.ora, SCIENCE VOL 319 CONTENTS L MEDICINE Clonal Integration ofa Poiyomavirus in Human Merkel Cell Carcinoma H.Feng, M.Shuda ¥. Chang, FS. Moore {tare but highly aggressive crm of human skin cancer may be cause ya previously urchraceized human plyemavis IO26isience 1152586 GeoPHysics Rogue Mantle Helium and Neon EAbaréde Anomalous high als of He ta ie inte recjcled basals under ecean anes may rest rem helm diusng in for more pristine, prinive marti OL 126%science 1150060 RESEARCH ARTICLES Pavsics Time Resolved Observation and Control of Superexchange Interactions with Ultracold Atoms {in Optical Lattices 5. otky et al trod atoms vapoed at sites of optical nies ae wed Investigate th superexchange ineraction between neg noorng sins. >> Pose 295 Naural Streams and the Legacy of Water-Powered til R.C.Weler end D. J Merits ooeplans and sreams in the eae Unie States were altered ‘tensive by mild nthe 1700s nd 1000s, challenging recent hytoloicinerpcetatrs an restaton approaches. 299 NEUROSCIENCE U2 Selector Activity Specifies Cortical Mentity and Suppresses Hippocampal Organizer Fate V.S.Mangale eal. ‘Te train's cortex begin 25. one-click sheet ose cel, hes timate identity is spected by agone that suppresses oneal ets. REPORTS ASTRONOMY Million-Degree Plasma Pervading the Extended Orion Nebula Gil eta. ‘tion degie as isthe Orion Nebula, implying that sho g2strom sella oslo i camimon inoue Gay 304 309 ed CONTENTS cont 18 JANUARY 2008 251Science REPORTS CONTINUED. cHEMsTRY Elementary Structural Motifs in a Random Network 312 of Cytosine Adsorbed on a Gold(111) Surface R. Otero etal ‘Upon cooing, cytosine mlacles on a gol surac forma ciordced netnrk based cn the ester other elementary stra unis, wien may have ansocies mith gases ceoriysics The Subcvation Zone Flon Feld from Selemic 315 Anisotropy: A Global View ‘MD. Long and PG. Siver entiation of te fstet seismic nave propagation peed \nsunxtuction nes reves tat enh migration induces tow Inthe nati above and beneath the subducting sa. EvoLUTION Localized Negative Genetic Correlation Constrains 218, ‘Microevolution of Coat Color in Wild Sheep L.Gratten eal ‘Auhough the nes wie shee increases wih sue. age, dark Sheep are bering raver because colors genetical inke to genes tha decease ines. EcoLoGY Coastal Ecosystem-Based Management with 321 Nonlinear Ecological Functions and Values E.B. Barbier eta. Target accoun! he retina eaten between presaredhabtat ea and wove alteration focitatesintegated management of (cast conantion ad development spetvep, 299 DEVELOPMENTAL BIOLOGY {Catenin Defines Head Versus Tat Identity 323 During Planarian Regeneration and Homecstasis KA. Gurley). G Rink AS. Alvorado Smeteateni-1 Is Required for Anteroposterior 327 Blastoma Polarity in Planarian Regeneration GR Petersen and P W. Reddien [er the head or tao! 3 lanai i severed, the signal intent of a prcrinent developmental signaling patnay contol whetes {new head ort egpneates NETIC Natural Genetic Variation in iyeapene Epson Cyclase 330 Tapped lor Maize Bioforitication CE Hanes etal detain efthe gone that conolsvituin Aves in maize wil alonprcucon of varieties that can imxoveclobal heath vwthout sing tanegenic method CONTENTS L 290 & 321 Dual Positive and Negative Regulation of 333 Wingless Signaling by Adenomatous Poy posis Col CM Takacs etal An inoortant developmental signaling molecule knownto be a tuner suppressor can abo achate goth, possibly explaining the responses same canes MEDICINE Initiating and Cancer-ropagating Cells in 336 TeL-ANIL1-Associated Childhood Leukemia D. Hong etal. leona was eachcaty pretoutemi alls containing the Characteristic chromosomal ransocation butery one wnderooes futher gene changes an deans cen. BIOCHEMISTRY Effects of Molecular Memory and Bursting on 339 Fluctuations in Gene Expression Je. Pedraza and). Paulsson -Atheory of stochastic gone expression sugges that noe canbe modulate mthoutfedback loops, complicating interpretation of single cl experirens PVAAAS 258 eric corguntegeestpunitrsncd ae cone pure top ore es eontpeyene pra CONTENTS continued >> wawescioncemag.org SCIENCE VOL319 18 JANUARY 2008 253‘Achange of heat. SCIENCENOW svusclencencut.org Building a Now Heart From Old Tissue ‘When nenbom cls are transolarted onto a “skeleton of {an adult heart, becomes 3 pumping organi heb, How to Make a Submarine Disappear Making gracetl ext Researches think objects can be made acoustical “sible” High Prices Just Fee! Good SCIENCE CAREERS “Anlcipation of quay activates ou bran’ pleasure cents. vv sclencecareer.org CARER A IMisciNet: Following the Image A.Sasio ‘Ara Skopuita carer fom tings shers pass oer, om cellar dts to shy high schoo stents ‘Mind Matters: Leaving a Lb Graceully 1.5. Levine When i'tie to eave the atthe key isto make ste heend ist biter Tooling Up: Dealing With Men Who Have a icblem With Women DJensen Youre ikl to encunter mene have never leaned how tlt tostrong, competent wore. From the Archives: You've Worked Hard to Get This Far LAustio ‘Don't tow your acséemtc jo interview Proteins escape from the ER surface, SCIENCE SIGNALING twvshecorg THE SIGNALTRISOU PERSPECTIVE: he Enoplasnc Reticulum Takes Center RGIENCE POD CAS] ar Stage in Cell Cycie Regulation Fearon and 0. Cohen fix The EReppeacs to pay ake olin controling the spatial localization of proteins inated inthecl ce ST NETWATCH: UCSC Genome Bioinformatics skin cancer, the legacy of ‘Analy you gene a incerest na range of organisms using the tools r Fae water-ponered mills, and more avaliable at the UCSC Genome Bronee,n Binlormatis Resouces, onscercemag pound Listen tothe 18 January Science podcast ta heat about genetic testing for psychiatric disorders, possible viral culprit ina vere ‘Separate individual or nstttional subscriptions to these products mey be required for ful text access, wawescioncemag.org SCIENCE VOL319 18 JANUARY 2008 255Evolution in Soay Sheep (uenttative trait loci OTL) ae genetic loc cre Late a typialyadatve phenotypic ai sueh as height or cor. In Soay sheep dark coat cole is asaoated th Large boy sa, whichis hetable and postvly correlated wit ress. Through ‘mapping ofthe resin associated withthe lc determining coat colt, Gratten et a (p. 318) etl the linkage of several causal mutations win a smal genomic region thin ree-tving sheep population. Astong cmelation was und beeen coat clr ané esa, but these tations showed antagonistic eects between body sine and Lifetime fines. This study provides emgircal sup ptfor the ol ofa negative genet correlation a {he evoionary nari of «natural populin, Organizing Brain Development In early development, the cortex ofthe mam ‘malian bain is built layer by lye, Irom the inside ut. But before that ocur the neural epithelium consist only a single ayer of stom cells. Man= gale etal. (p. 304; see the Perspective by Grove) have now determined that inmnice the gene Uhi2 determines cota identity for cells at these ear cat stages. Un2 etablshes whether cells averse te the cortial hem, which i nw seen to function «an organize of hippocampal developmen. Describing Disorder Insights int the structure of amorphous mater alsuvally come fom theoretical modeling or from microscopy studies of larger colloidal part cles, Although canning tunneing microscopy (SIM) methods can achieve atomic resolution, molecular overayers on mata substrates tend wwwsciencemag.org SCIENCE VOL 319 Trapezium, pri et al. (p. the galactic plane, ether to form bulkctke stands or crystallize into ordered films. Otero et a. (p. 312, publishes online 13 December report SIM images of eto sine molecules adsorbed on gold surfaces which show thatthe molecules are highly mobile at ‘oom temperature. Upon cooking to below 150 vin, disordered hydrogen-bonded netaorks {orm that can be characterized as being bull up from three types of subunits—zgzag filaments And five: and si-merbate rings. The Not-So-Old Mill Streams (Our understanding of natural stream devel ‘opment has been greatly influences by ‘many studies inthe easein United States. Walter and Merritts (p. 299, cave; see the Pe spective by Montgomery) now show that mill dams ware parva sivethere nere more than 65,000 mils in the 19th century-—and that these structures thor ‘oughly changed the natural characte ofthe floodplains and streams. Mapaing of about 20 streams in detall shoned that these dams, co: {ectvely spanning the fll nth of mary streams, alongwith widespread deforestation could cause up to 5 metersof sediment to accu mulat in floodplains. Steams were changed {om branching channels across broad wetlands to confined meandering channels Vitamin A in Corn amin a detceny affects the vison of i tions of eilten each yea, a problem that could be addressed wth dietary adjustments. Hares et ap. 330) show that naira maze vatin's offer a wide range of the precusors to vitamin A The clase gene (critical in assigning EDITED BY STELLA HURTLEY AND PHI << Shockingly Hot Four extremely bright and massive stars, called the ary illuminate the Orion Nebula. Gidel 9, published online 29 November see the Perspective by O'Dell and Townsley} show that the Orion Nebula glows in x-rays because it is flooded with very hot gas in excess of L milion kelvin. This heating is ely the result of shocks from powerful outflows from one bright SHar in the Trapezium, The majority ofthe stats in our Galany are found in regions similar to the Orion Nebula, s0 this phenomenon should be widespread throughout ‘carotenoid varians to different metabolic path ‘ways. Aselectve breeding program, rather than ‘complicateé transgenic methods, may be able te deliver more nutitious maize, Heads or Tails Planacans—tubulaylatworms—can regener: ate all oftheir body part and entire organ sys ters aftr amputation, However, the mecha tn by whic the arial “kaon” ha to ge erate head after head removal or tal ater ‘ail removal, a propery called regenera tion polarity, is unknown. Gurley et ap. 323, published ontine & ‘ecembet and Petersen and Redden (p. 327, pubished enine 6 Deceme) now ind that acon served factor within the Wnt sig talingpatnayisused to distin ish ead fom til. Deceased Wat signaling through tember 2004, surveyed 988 houscholds in 3 clusters across the county. They rived ata figure of 98,000 “extra deat since the invasion, about half due to violence, Soon § after this, a team Led by Pedersen and the wyww.sclencomag.org United Nations Development Programme, which had used much larger sample of 21,668 households in 2200 clusters, pro: duced an estimate for toughly the same period of about 25.000 violent deaths. As the invasion gave way to occupation and insurgency. Roberts and Burnham Human cost. The WHO-Ied aq Family Health Survey estimates as many a5 223,000 deaths since the rq ivasion, far fener than @ survey by Burnham and Roberts estimated ‘mounved another study. This time they lef the surveying entirely to the Iraqi team, ‘communicating from abroad. Published in Oetober 2006 in The Lancer, the second survey-based on 1849 households in 47 clusters—estimaved that 601,000 Iraqis died violent deaths between the 2003 iava- nd July 2006, To many, the number realistically high. Some also faulted the authors for not fully answering ns about the survey's methods 10 Oetobes 2006, p. 396). ‘Now comes the WHO survey, Conducted wwiththehelp ofthe Iraqi government, itisby far the most comprehensive mortality assessment to date. Interviewers visited 9345 homes in more than 1000 clusters. But its estimate of 151,000 violent deaths has ‘come in for some criticism, oo. Unlike other Iraq casualty surveys, this one includes an upward adjustmeni ef 35% wo account for NEWS OF THE WEEK L really an arbitrary fudge factor,” says ‘Debarati Guha-Sepir, an epidemiologi the WHO Collaborating Centre for Research oon the Epidemiology of Disasters in Bru sels, Belgium. But the number falls quately within the ranze produced by & meta-anal of all available mortality studies by Guha-Sapir and fellow centre epidemic gist Olivier Degomme. The Johns Hopkins Figure is an outlier, she says. Why the Hopkins study came up with sucha high figure is not clear. Criticism of the study hasin fact intensified sinee Burnham and Roberts released a data set to selected peers last year. “It did not include the standard kinds of ™ says Seppo Laaksone! statistician at the University of Helsinki in Finland and a sp. cialist in survey methodology. ple, he says, it was le “to check the objec- tivity and randomness of cluster selection.” Scheuren, who also received the data, wanted 10 ‘compare results obtained by dif ferent incerviewers w "wet han die on noise” and check for fabri= cation by surveyors, Roberts declined to provide all the details, according to Seheuren, saying that he was concerned that this would risk the safety of the interviewers Burnham told Sefence, how= ever, that the Johns Hopkins team doesnot have such dewiled information. “Our goal was to reduce any type of risk to the community and the partie says Burnham, “While we have much of the Faw data, we requested that anything desi interviewers or the location ofthe neighbor. hhonds visited not be sent ro us Laaksonen responds that he would not have published “any figures for the country” if he didn’t hhave direct aceess to such raw information from surveyor Burnham is not retreating. Because the WHO survey was conducted by Iraqi gov: ernment personnel, ‘people may have been hesitant to answer honestly.” he says, He claims that unlike those in the WHO stud, nearly all of the deaths tallied by the 2006 Lancet study were verified with death eer. tificates, Even iPthe debate may be drawing toa close about whether the number of vio- lent deaths in postinvasion Iraq could be as high as 600,000, the argument about meth= ‘underreporting” of deaths duetomigration, ods is clearly far ftom settled. memory lupse. and dishonesty. “That is JOHN soMANNOR SCIENCE VOL319. 18 JANUARY 2008, 273274 NEWSFOCUS ‘SAN DIEGO, CALIFORNIA—Starting nology company was rex part of John Kelsoe’ life plan. A respected psychiatric zenetiist here at the University of California, San Diego (UCSD), he has spent 20 years patiently arching for the genes behind bipolar dis. coder, tracing patterns of inheritance in ilies from Amish Pennsylvania to Teeland, Kelsoe has now laid this solid reputation on the line: He has founded a company that last year quietly b ne test 0 help diagnose people with bipolar disorder. \shich affects about [*% ofthe population, Kelsoe knows well that he is opening himself up cise, And it has b Collins, director of the National ‘Genome Research Inst IGRI in Beewls, M land, was taken aback to learn that a gene test far bipolar disorder was fon the market, Being diagnosed with a mental disorder carries nificance” says Collins who, like many’inthe fel, thinks atest based cn current kaowled is likely to be misleading. In common diseases, “there are no nities” yet for bipolar disorder, 18 JANUARY 2008 Kelsoe, tall and tri, with the sig stoop ofanacademic who has spent yearsin the la netics of bipolar are own frustrated with oratory, agrees that the g still cloudy. Bur he ha nosis and spoty treatment ofthe disease. On average, 7 years elapse from the ‘onset of symmpioms to diagnosis an! patients nosed three times. “There elu says. “AS opposed to playing it safe and conservative scientifically, [think we also have to think about our obligation w provide sornethit fr physiciansand for patents, Peet VOL 319. SCIENCE His company, to market a test in psychiatric genetics, but it ‘won't be unique for long. Another compan in Louisville, Kentucky, plans next year 10 star offering a similar tes for schizopheenia, A third in Boulder, Colorado, will launch a senctic screen w help identify those at risk of Suidal thinking from antidepressants, Psychic of the fas-expandi but it arouses Fierce emotions, There tests are very lifferent from trditienal for rare and often fatal diseases like eystie rosis, Tay-Sachs, and sickle cellancrnia, With these, the cause is not in question: An individ cursed with faulty genes from both parents has a 100% chance of contracting the disease Bipolar disorder is another animal entirely: Dozens, perhaps even hundreds of genes may con- lanyely unknown environmental heart disease, adult cancer, and type 2 diabetes. Each gene vari uly isk only s say. from 1% to that the vast majority of people swith these variants never fall ilMoving target. Nany researchers sey that early studies linking genes and mental less need to be replicated from that disease, Furthermore, in bipolar disorder, suspect genesare still Being uncov- fered, There'smuich debate about which finde ings will endure and which will bepicked up ‘once or wice and never seen again. use ofthis uncertainty, not ie mene tion concern about how test results will be imerpreted, researchers and ethiciss are ask- ing difficult questions. Why test for genes ‘whose science is shaky orthat boost risk only slightly? Who will buy these tess? And how will consumersunderstand andcope with the ‘genetic information they re given? The genes ics, including “offices, se ther selves ging this nascent Field and carving ot an early niche. They also believe, in the swords of UCSD business school assistant dean and Psynomis CEO Kurt May. that "don't have to win the Nobel Prize on we don’t have to solve the problem forall time, All ve have todo is resent a husiness movel tha is one beter ‘than psychiatry today.” ‘At the heart ofthat business model is GRK3. 1 yenc on chromosome 22 tht Kelso identified in the late 1990s. In 2003, he reported in Molecnlar Psychiatr that a change in an individual DNA base in GRK3, called single-nucleotide polymorphism (SNP), was associated with about a three ‘ims increase inthe risk of bipolar disorder in 28 families studied. Bu the varie was rar, found in only 390 those with bipolar an 1% ofthe unaffected Since then, Keo has hiton a combina tion of four SNPs in GRAS that together scemto heassocited witha doubling of di cae risk and are more common. They were found in 13% of bipolar patients (and 5% of those without the disorder) in 181 Caucasian families, That work was published last nionth. UCSD has filed For global patent protection on both sets of variants and Ticensed them to Psynomics: Kelsoe will shave in licensing fees and any profit. The Paynomics test looks For bath GRAS var ans, Italo checks fora pubtely available ‘tition in the serotonin transporter see, ‘shich is believed to influence a person's response to certain antidepressanis, often a 2 component of treatment for bipolar disorder. & “Does GRK play compelling role in bipolar disorder? It mey oF R may not. "I'S & alays possible” tha i volved. but “20 wowwsclencomag.org ime it seems unlikely.” says David Collier, a neuropsychiatric genetiist athe Institute of Psychiatry in London, [n December 2006, he published a paper arguing that bipolar | disorder, the more severe of oxo known, forms of bipolar disorder, wats not asso ated with GRA in a suimple of 410 patients and controls from Scotland, But Collier adds that because of its relatively small size, his town siudy “can’t be regarded as definitive.” Collier is keeping an eye on genome- wide association Studies. which are revolu tionizing the genetics of complex disorders such as type 2 diabetes. One of the first whole-genome studies that included bipo- lar disorder, published last year by the ULK-S Wellcome Trust Case Control Con sortium, did not find GRK3:italso failed to find many other genes previously linked to 0 amass me 2 7 ’ on od exe: oo 2 ” 2 we, tink mth Expanding. Psynomics expects its test to row from mo genes to seven Cris say few. E any, genes for bipolar eisorder have boon validate. the disease, says consortium member Col- lier. Although such seans are not designed to pick up every disease gene, some propo: nents of the whole-genome approach say their confidence in genes identified the " e sets of affected Families, is declining. Psynomics “is using data from the other side of the bridge, the bad side," says NHGRI’ Collins, referring to family linkage studies that have been hard to replicate, Pablo Gejman, a psychiatric geneticist .with Evanston Northwestora Healthcare and Nontwvestern University in Evanston, I SCIENCE VOL3I9 NEWSFOCUS i nois, is leading a project that he hopes can cconfitm or reject initial schizophrenia gene discover -wide association study of 4500 people with this disease and. 4500 without, Results are due out later this year. (Kelsoe is leading a parallel stady in bipolar disorder.) Gejman wonders if even larger trials —of 15,000 people. perkaps— will be novessary. This week, he published a paper in the American Jounal of Povehia- ry, reporting that none of [4 previously described schizophrenia genes played a big role in disease risk in neatly 3900 people. “My message.” says the native Argentinean, ‘woul be that we need to be patient, [con= duct] systematic studies, collaborate .. and tty not to jam woorapidly the gun" by selling ci tests befove the data are solid, The doctors Kelsoe acknowledges that GRA3 may turn ‘ou play a minorrole in bipolar disorder, or ‘even no role a all although he considers that unlikely. No independent group has pub- lished a second GRK3-bipolar disorder asso ciation, but Kelsoe says he knows of two ‘groups. in Israel and in Canada, that have foundalink. Meansshile, inthe next 3 months Peymomies hopes to add Five more genes to its test. including four that Kelsoe discovered, and then ramp up marketing to physicians and patients, Of the five, three are thought to modulate response to lithium, a mood stabilizing drug used 0 treat bipolar disorder; ‘one, CACNG2, is considered indicative of bipolar risk and lithium response; and another, SP4, has been linked (o both bipolar disorder and schizophrenia. Tn determining which genes. ‘commercialization, “we've cut threshold, and that is replication in ‘one inxkeperdlent stay.” says Kelsoe by a different group or his own group exam ining new samples. “Docs make decisions. ‘every day on allot less information” than thi Gejman and Collins may shake thei heads. bat psychiatric gene testsare coming. Mark Brennan, founder of SureGene, a sehizophrenia-based test company in Louisville, says such tests are “going 10 happen with or without” his own efforts, Brennan is SureGene’s chief scientific offi ‘eer and a human genoticist atthe University ‘of Louisville in Kentucky. His company iy designing 4 risk predictor and diagnostic Brennan discovered. It's good for scientists to get into marketing, he argues, because “we have to step up and make sure i’s done ina way that really helps people: Brennan has so far published informa: 18 JANUARY 2008 275a NEWSFOCUS 276 Hoping for a Glimpse of What's Ahead Its one ofthe fist questions people ask about testing fora gene associated wth mental illness: Who would want to knowthe result? One ready volunteer {or testing is Holly Finn, 13, nfo lies in Grand Rapids, Michigan, with her parents and her 16-year-old sister, Katherine. Holy’s mother, Ki was diagnosed wit bipolar disorder 3 months before her 17tnb says that Katherine began showing signs of depression and difficlty concen: trating inthe fourth grade. Italy pegge as having attention efit hyper axctivty disorder (ADHD), Katherine received a bipolar diagnosis in July of 2008, nhen she was 12. t's been a rocky road: Katherine has smitched schools three times and suffered taunting from her pees, says her mother, an active ‘member of the Depression and Bipolar Suppor Alliance in Chicago, Ini. Nowthe younger Finn daughter, Holy s being treated for ADHD and anx- ‘ety Her symptoms are different rom ne sister's, and she isnot though to have bipolar disorder. But, Holly says, having witnessed her sister struggle for yeas, she would welcome testing fora bipolar gene. “I thine might have a hance of having bipolr disorder," ays Holly, who writes songs she plays on her guitar to cope wit her sister's ups and downs. Atest could help her “get tothe bottom of” her symptom, Holl say, or at eas prod her and her psy- ‘hiatrit to keep an eye out for bipolar disorder down the road, “It certainly wouldn't hart.” But hurt is exactly what many are afraid of, “Some people would say \neividuals havea right to this information and shoutda tbe barred by the ‘medical community frm receiving it,” says psychiatrist Jinger Hoop of the tion on two of the genes, which are on chromosome 22; SureGene is working with outside scientists to replicate the results. The company declines to name erbate their disease. “Doctors are operat ‘ain Fied is Herbert Meltzer, phrenia specialist at Vander in Nashville, Tennessee. Meltzer, who is fon SureGiene's advisory board and says he hhas received company stock, is enthusias= tie. “If we proceed and it doesn't look as promising to me your ivory tower and s Ssitloaks ghee Company Test available Disease then P11 do my best tokesp itoft the Newolark mid-2008 jr depression rurket he promises Before launching — Psynomics. now Bipolar disorder the SureGene tes slated to happen —sueGene —mid-2007 —_Sciophrenta about 18 months fiom now—Breman PIPASEsH Hema Rosey ‘sams the goes repli ind err * he says, strugeling to find the right ddiggnosis, the proper recipe of medications. “Noto ge! snooty about it—it'sfineto sitin Offer this until irs perfect’ Meanwhile, that Players in the Psychiatric Gene-Testing Business Medical College of Wisconsin in Milwaukee, who is studying how people with schizophrenia or a rk frit respond to genetic data, Her work has «confirmed what others have seen: Consumers often reat favorably to hypo" thetical genetic hnouledge. “But we don't know alot about what the actual risks and benefits are” of receiving genetic results, Hoop notes. “We really need empirical data to let us know hom people nould interpret very subtle changes in sk” One ofthe fen projects tackling ths question i the Rsk Evaluation and Education for Alzheimer's Disease (REVEAL) stucy headquartered at Boson University School of Medicine. Neurologist and epidemiologist Robert Green and his colleagues here and at other centers ae examining how neatly 700 volunteers witha family history of Alaheimer’s disease respond to learning whether they cary 2 deleterious version ofthe APOE gene. A single copy raises Alzheimer's risk roughly three times, and two copies raise risk 12 t0 15 times. "The bottom line is that people do want ths informa tion, and they seem to sel-select for peop who can handle it,” says Green. ‘Those “who felt that they couldn't handle such information generally didn't fallow throuch and obtain’ their test results, REVEAL has ecamined the role of genetic counseling; whether individuals vith the hazardous form, called APOE e4, ate mor likely to buy long-tenm «are insurance—they are—and more broadly, what they take aay from knowing ther APOE ed status, especialy given that theres no ay total the onsetof Alzheimer’s, One concern, notes Hoop, is thatthe REVEAL study has found that con sumers tend to recall not their numerical risk of disease but whether they paying up to $750 and providing aphysivien’s name, Soon after, a small Blue and white pls tic container arrives in the mail. The capsule preserves a saliva sample, which the eon sumer ships to a certified UCSD lab for genetic testing, with DNA results going back to the company: Results also go to the desige nated physician but not tothe consumer. The process is meant to control who takes the test: Kelsoe has rejected risk test ing for individuals who have no syrmp= toms but are curious, ‘We don’t want 10 Ta akc he Tee, pata a Danes alice ze tn oo pally of pets ana rz tyme tmate peo ‘response to antipsychotics Divine bling, een so, Kelsoe a thar there are ways to ceaed “in at least several hundred samples fiom different centers.” That far fewer than the thousands Gejman argues are necessary Even if the genetic associations he and Brennan have found are small clues, Kelsoe believes that they can still help. One reason for pushing forward, he says, is that patients need information now. Among the veterans he sees at the local VA hospital, a short walk from his UCSD. office down paths lined by eucalyptus ‘1B JANUARY 2008 VOL319_ SCIENCE doctor's down there trying to work with whatever he wou” ‘The market Paynomies is not relying much on doctors, however, to help it reach its goal of selling 1800 tests in 2008 and 30,000 in the next 5 years, “This s going to initially be driven by the patients" says Kelsoe, who believes that Aociors are wary of a psychiatric gene test Psynomies allots anyone to order its test by _get around Psynomics’s procedures, A bipo= lar patient, for example, could send in her ‘childs saliva sample and claim its her own, In the end, predicts May, the market will demand risk testing, and "the market is going 0 W Business has already won out in another facet ofthe Psynomics test: how the consumer will be advised of the results, Kelsoe says he initially wanted the company to provide & individual counseling for clients. After about & nw scloncemag.orgharbor the “bad” gene. This is a ‘concern in testing for bipolar disorder or any psychiatric ill. ness: People may think that car tying the gene means that they have or will develop the disease Making this mistake could, ‘among other things, induce stress, which can exacerbate pe iat eisorders, How much of @ boost in risks meaningful to consumers? The bipolar test from the company Paynomics in San Diego, Califor: nia, currenily includes the gene GRK3, which the company says Increases risk two to three times. v In someone of averace risk, that’s a shift from 1% to 2% or 3%. For an ingivigval with a bipolar parent or sibiing, whose baseline isk is about 10%, the incase may be more meaningful—but only ifthe gene com- Pounds family precisposition. ‘Another concer is how having such inlermation willalter a person's Be “Tome, the question snot really the isk elatonship" | ase, s3ys Wye Burke, a genetcistand bioethics atthe University of Washing ton, Seale, "Is nhat ace you going to do¢ilferenty with that information?” “ao longer ial” Interpreting and coun selingon tes results willbe up t doctors, who rece summary noting that™apos- itive Gi makes a diagnosis of Bip larDisorder2 3 tines morelikely"forindid- uals with symptoms anda family history: This is not much help, some sey. “I'm in the bipolar disorder geneties field, and | ‘woulldn't know how to interpret the results of a bipolar genetic test,” says James Pota director of moed disorder Hopkins University in Baltimore, Mary Potash, who knows Kelsoe and calls him a “very capable, very solid guy.” says that offering a test now is premature and that doctors ed onto decode i sting the results of y to grow knottier as many psychiatric neluded. Inter poorly understood 2 disease genes may predispose a person to more than one psychiatric disease, Indeed, even though SureGene has positioned itself in schizophrenia, Brennan notes that its test may be more instructive about psychosis, risk generally. Psynomies’s CEO May frames this as in opportunity: “We would g lke to establish that as a brand. this is the E destination tor information, tor diagnost B for therapeutics” in psychiatric iliness. g Another company, NeuroMark in Boul- der, Colorado, has similar ideas, NeuwMark wunwsclencomag.org Cay ‘isorder, Younger daughter Holy cee Pee td cas NEWSFOCUS L Hollis already contemplating this question Assegrons up and considers having a child of her nn, she say, she could imagine testing her son or daughter at about age 2 for genes associated withthe disorder so as tobe teady to spot iit surfaces. Currently, testing of chitden without symptoms is not avail able, nor is prenatal testing, Physicians focus instead on sta biting a woman's bipolar disor der vile pregnant—something that Kristin inn discussed in her 2007 book, Bipoler and Preg ant. But even the skeptics agree that validated genes for schizophrenia, bipolar disorder, and othe psychi atric diseases ae likely to emerge soon. “Biotech companies are going to do what they're going to do within the limits ofthe lav,” says Kay lmison, 2 psychologist at Johns Hopkins Univesity School of Medicine in Balt betncen genes ae dis is desizning a pharmacogenetic test For daliy risk from antidepressants, side effect for 2% t0 6% of those taking the drugs. The ‘company hopes to launch its test, focused on fone antidepressant, Celexa, later this year, before extending itto related drugs and devel ‘oping bipolar andl schizophrenia risk tests as well. Then there are broader gene testi nies such as 23andMe in Mountain View, California, which offer information on. SNPs behind a host of diseases, Although 2SandMe could test far SNPS in depression and other psychiatric illnesses it would ne permission from companies like Psynomics and SureGene ‘@ sell information on the SNPs those companies have licensed Despite the psychiatric space, Paynomies is very much a work in progress ‘Only about a dozen people have taken the test so far. The company hasn't determined how it Will follow its clients, although May says tis 2 them a satisfaction questio Means, Psynomics is hunting for venture capital tide it over until west sies ramp up, One critical but unsettled factor is whether Congress will pass the G Information Nondiserimination Act, which ‘would protect individuals against diseri nation by insurance companies and employ ers based on genetic information. Biotech ‘companies and researchers have lobbied for it for years. but it hasn't been approved. Another uncertainty is how gene tests will yr many genetic test. inclu SCIENCE VOL3I9 ‘more, Maryland, who bas wl with bipolar disorder and who generally supports making gene tests avait able, may take years before it's clear what that means for people with psychiatric illnesses and ther families extensively about her personal batle ic. ing that offered by Psynomics, regulation consists of lab certification but no 28s iment ofthe science bchind the test. “This is different kind of infornsation”™ than has been provided in the past, says Boston University neurologist and epidemiol logis Robert Green, who studies Alzheimer's Sear whatts ling ‘Gene testing could be dane pr natally for example, to cheek for bipolar risk in the fetus. Kelsoe shudders atthe tha expe cially because bipolar genes are unlikely to guarantee disease: "T think this is very dan- gerous. scary. eugenies-type stuff. and [ don’t want to get anywhere near it,” he says. Kelsoe says his company would refuse to hhonor such a request. But if the question ‘comes up in the Future, the decision may nat be his to make In Kelsoe's UCSD office, decorated with his children’ art, a small scrap of paper is taped to his computer. In black ink is serawled a quote from Virgil: “Fortune Favors the BOLD!” Despite that encourage e scary” he says over lunch, “Tvespent my whole career tying... 10 people's respect that I'm responsible and cautious, So the concern is, are they’ going to think this is irresponsible. am I going to lose credibility because of it?" But then Kelsoe leans in, a wieked grin on his tee. “Sor times," it's fun to stir things up.” JENNIFER COUZIN 1B JANUARY 2008 277Sacred circle. Carved central slabs are surrounded by massive monoliths neatly S meters al. tiated wheats are from this region, “Yo cean make @ good case thi the real origin of complex Neolithic societies.” IfGobekli Tepe istruly 11,000 yearsokl it also challenges the idea that symbolism and agriculture were first developed in the ‘Levant the area that ineludes modera Jor- ddan, Israel, and Syria—and spread north, “The idea that the origins of monumental architecture were in the south has been turned on its head,” says archaeologist Gary Rollefson of Whitman College in Walla Walla, Washington, and an editor of the journal Neolithics. “What Klaus is picking lout there now is far earlier than anyone expected. That’s all back on the drawing beard now. A hunter-gatherer's paradise Gobekli Tepe means “navel hill” in Turkish, Seeking the Roots of Ritual and at 780 meters high. it seems a natural eeking the hoots of hitua thering spot, the highest point for kilo ers. The site was first examined—and {In the hills of Turkey, researchers are slowly uncovering the world’s oldest monumental dismissed in the 1960s by University of structures, strange monoliths built by hunter-gatherers perhaps 11,000 years ago Chicago anthropologist Peter Benedict who assumed that the flints and broken GOBEKLI TEPE, TURKEY—Dawn was still Tepe, in southeastern Turkey, “is the first slabs of limestone littering the area were halfan hour away when Klaus Schmid manmade holy place.” says Schmidt, He the remnants of an abandoned medieval or arrived at his hilltop dig site here, but a and others say that the site upends tradi-_ Byzantine cemetery half-dozen workers from nearby village tional notions about the development of Schmidt first visited the hilltop in 1994, were already waiting for him. It was the symbolism. Archacologists once hypothe- affera local farmer had run hs plow into a end of the dig season, and the enthusias- sized that agriculture gave early people the reciangular piece of limestone. Schmidt tic, white-haired German was there for cimeand food surpluses that they needed (@ found noc a gravestone but the top of a inne last look, He picked his way down build monuments and develop a rich sym- large, buried pillar, and he quickly recox- Steep excavation layers toward a massive, fed thatthe scattered Mlint T-shaped slab of rock rising 3 meters out tools on the surface resembled ofthe ground. those from nearby sites tha pre= The went i awn illuminated date pottery. As he began i= moze Fshaped limestone megaliths and ging, the mystery grew. He then still more, arranged in circles and ancovered one huge, elaborately ovals down the hillside, Some were carved carved ritual structure after with sivlized arms: on others, a carved eek nother, bt no houses or signs menagerie of shes, spiders, boars, faxes, of settlement. The people who birds, and other beasts emwled and soared built the monaments apparently Schmidt of the German Archeological | “€¥P aus SYRIA backfilled them thousands of Institute (DAL) in Berlin, claims that these 2 years ago, and the bones that are the world’s oldest monumental struc LesangN iumed up inthe loose fil were tures, which he has painstakingly uncov- ae all from either humans or wild ered since 1995 (see sidebar, p. 280). bolle vocabulary. But Gilbekli Tepe raises animals, net fom domesticated animals Radiocarbon dating and stone tool compar- the aiternative possibility that the need 0 So just how old are the monuments? isons indicate that they are 11,000 years feed large groups who gathered to build or Because most of the bones found are from oli, So these great stone circles were worship at the huge str backfill, dating them yields muddled erected before metalworking, before pot first steps toward agri results, and Schmidt has fewer than two sociocultural changes come first dozen direct radiocarbon dates. Instead, he ; chiefly relies on comparing the sione tools Hodder of Stanford University in PaloAlto, and other arificts with those from sites California, who notes that the first domes- with more radiocarbon dates, This relative ven before most signs of Although people beg ces thousanils oF years. 278 1B JANUARY 2008 VOL319 SCIENCE wz scioncemag.orging puts the early layers at Gobekli Tepe n the beginning of a distinctive period called the Pre-Potiery Neolithic B, dated elsewhere 10 about 11,000 years ago. Although the dates aren't preeise, the a fact comparisons are convincing, and a few hundred years of fudge time “don’t really says Harvard University arehaeolo- sist Ofer Bar-Yooet Although the site may predate system- wsriculture, the people who carved these pillars faced no shortage of resources. Animal and plant remains suggest that 11,000 years ago this place teemed with gazelle, aurochs, and deer. Groves of fruit ‘nul mul trees lined the rivers, and flocks of migrating bitds paused here regularly. “It must have looked like a paradise, ideal for hunter-gatherers,” says Angela von den Driesch, an emeritus archaeozoology pro- fessor at Ludwig Maimiian University in Munich, Germany, who has classified ani ml remains atthe site, The region was so rich that people could have settled down ‘while sill supporting themselves with hunt- is have oft ste "Yon den Diss ne ofa nib of archavolosists, plus a rovelving cast of cr German skis and about 60 focal srrlera ho ercevt tha sts for 4 ont cach youu They dre othe se before to dig teumibg fa tcalerocn wo 9 ba tional high-walled compound Schmidt towns ia the mide of Uta ol ety One day ls fal. a dozen students quell sored cui weighed bone agnensanda the compound’ sourard a the slerncon cooled late eveniag. I's Farpraiagly feprecalins peiaion: gen he pose § ing nature of the site and the roughly # Ssho.onn in funding Schmit receives each 4 year fromm DAL andthe Germen Research i rout ion. Clearing a space a along table, Scher § pulled oat amet purple scheraoe of he {Mite a sor ot bind eye uy put eget 4 after ground-pentrating radar sean tn § 2003. The excavated portion 3300 square metes,ust Yeo! te hectare ste ol ; a : 6 E § was represented by a white square. AIL around it, the sean revealed oval groups of buried monoliths dotting the hillside like dark-purple bomb craters. ‘Schmidt says there are at least 20 cere= monial structures still underneath the soil, and perhaps more. "I's much more complex and advanced than we thought,” he says. ‘Clearly. they cou! communicate with each wunwsclencomag.org other andor ly compli- cated. [didn’texpect hunter-gatherers could ‘or would build something like this.” “The most spectacular ritual space, whieh ‘encompasses features seen at the others. is nestled in the dig’s western corner, Two -meteriall, T-shaped pillars stand in the center of a cirele of slightly smaller slabs, ‘weighing an estimated 5 107 tons. Pat= alle! lines earved into the slabs’ sides slant ‘back ina V and meet at the narrow front edge like clasped hards. suggesting a styl- ized person. The two central stones face the valley, and the surrounding pillars face anize something re Peering into the past. Klaus Schmidt (lop) uncovered pillars Carved ith foghtening bests te th an. SCIENCE VOL3I9 NEWSFOCUS i inward toward the center of the cirele. In shatp contrast to the stylized human shapes, the sies of the sal ved with images of animals: complex arrangements of spi- dors and snakes, foxes and wild boars, vul tures and cranes There's no way to know for sure what these figures meant, but there are some clues. Schmidt says the lack of female sym- bology largely rules out Fertility rituals, And the contrast between the designs carved on the pillars and bones tha litter the backfill seems meaningful. The backfill yields savelle,aurochs, rel deer, boat, goats sheep, oxen all wild plus a dozen different bird species, including vultures, ducks, and geese. Yet the pillar carvings are dominated no by prey but by more danger: ous creatures: leopards, lions, foxes, and vultures, plus spiders, snakes, and scerpions. “The fed by animals.” says Stanford's Hodier. “I's a scary, fantastic ‘world of nasty-looking beasts.” The cathedral on the hill So what was this place? Schmidt ts adamant about what i's not, although not everyone agrees ‘with him. Despite the site's size and the contemporary villages ichmidt insists this was no settlement, He's con- vinced that the circles were designed to be open to the sky, like Stonehenge. Telliale signs of settlement such as hearths, twash pits, and stall fertility fige uurimes—are conspicuously absent, And the hilltop is a long hike from any water sources “We know what settlements from these ti Schmidt says, one of them.” Instead, Sehmidt argues that huncer-gatherers. from across the region gathered here periodically, pooled their resources temporarily to build the monuments for some ritual purpose, andl then let Rollefson and others see that asunlikely arguing thatthe scale ofthe site would have required at least small group of year-round residents. “To have this kind of magnet out in the middle of nowhere would he unpree 1B JANUARY 2008 279a NEWSFOCUS 280 Just Don't Call It the Garden of Eden UREA, TURKEY—Fromfis one-man bocth atthe Urfa bus sation tours of: al Serdar Avd spends his days handing out maps tothe sights arcund this Turkish city. To his clsmay, one destination is conspicuously missing from the tourist quides: Gabelli Tepe. "People want to come here to see these temples, but Klaus Sdymidt is digging so slow!” Au says. “How mary years do wehave to waitto go inside? Thsis a big problem, think.” Everyone in this ity of almost a milion people seems to know Schmidt's ame and lave an opinion about his wok excavating the wold’ ist mon- umenta architecture (see main tea), Schmidt's progres is charted by reg ular teports in the local papers, ane a 2006 caver story in the German weekly Der Spiegel speculated thatthe site might be linked to the biblical Garden of Eden. Much to Schmidt's cismay, Turkish papers picked up the story andthe ensuing debate about whether the site nas the birthplace of ‘Adam—79,000 dams inthe United States, 3500 have boen rated as unsafe, collec tively in need of $30 billion for rehabilitation, repair, orremoval (1). Levee inventories are Jess cleat, but estimates exceed 25,000 km (9), with unknown structural integrity 0) Todte, “600 damshine been renkned prima rily for safety and economic reasons, Dam removals flloweal by rapid recovery of inver- tebe. fish, and rparian vegetation communi- ties (/7), Levee decornmissioning often aban ddoning breached levees, reduces economic demands of kvee reconstruction while improv Ing flocaplain habitat and water quality (12) OF the>6 million km of rads inthe United States, $85,000 kim aze on pte lands mine tained by federal land agencies, a portion of which are rarely used (13). The US. Forest Service (USES), with 250,000 km of roads cover 50 years old. estimated its maintenance backlog at ~$10 billion for 26% of system rads (13). The USFS decommissioned 7900 km of reads between 2002 and 2005 and has ‘demtified almost 300,000 km for possible Alccommissioning over the next 40 years (1), Decommissioning decreases economic Kabili- ties but i also an important tool for restoring forest ecosystems (13). Tn US. federal waters, there are =3900 off shore ofl and as platforms, primarily in the Gulf of Mexico (GOM), about one-third oF Targeted decommissioning of deteriorated and obsolete infrastructure can provide opportuntis fr restoring deqraded ecosystems, U.S, POPULATION Dans (thousand) NK BRIDGES 707) SEWER PIPE 600 200 (Ghousands of nd RX 5 SURFACED ROADS (rion ok) OFFSHORE O1L/ GAS PLATFORMS (thousands) 1900 1920 1940 3960 3990 2000 USS. Antrastructue inventory. For sources, se (2. yw sciencomag. org,which are idle (15). Fed eral policies regi that pkatiornsberemval with in L year of lease expira- tion nd 2700 plaforms have been removed. Re moved platforms have bbeen primarily shallow water platforms and mostly dlsposal of orshore (15), ‘Cosssandenvieoamentian pacts of remening deep vwater platforms are sub- stantial ull platforra removal has drasboeks, inludin vironmental impacts and loss of the potential ecological value of the structure as an artificial reef (/6), Research shows that plaiforms. facilitate the expansion of com populations inthe GOM. (U7) and act as refuges for juvenile fish, rereasing fish production off the coast of (California (18), Rigs-1o-Reets Programs allow reuse of decommissioned siucturesasatficial reefs, Through 2004, 190 retired platforms were dedicated for fisheries enhan: ‘which reduced decommissioning costs and led to >S20 millon in industry donations to state ‘environmental management trust fax (13) Department of Defense (DOD) facilities pose an unusual challenge ane opportunity. OF he 257 million of federal andsinthe United States, >10 million ha belong w the DOD (19). Access restrictions have made military bases some ofthe richest ecological reserves of any ‘oF the nation’ public lands 79). Through the Base Realignment ard! Closure (BRAC) pro _gram, 400 military sites were closed or roclas- sified between 1988 and 2002, To date, man- agement of 21 bases on 445,004 ha as been transferred from the DOD to the US. Fish and Wildlife Service to become National Wildlife Refuges (e.., Jefferson Proving Ground be- ‘came Big Oaks National Wildlife Refige). Perhaps the largest combination of in- fastructure management and ecosystem restoration is the Comprehensive Everulades, Resioration Project(CERP), aneffortto restore ‘he2.3 million-hawatershedandits ecosystem (3), Hydrology in the Ev ne Fated dhrough hundreds of control gates, thoat- sands of kilometers of levees, 2900 km of ceanals, and dozens of pump stations, which largely continue w function well. Tae CERP is, based en infrastructure modification and par- al remeal to move the ecosystem 1 a more natural and sustainable configuration throug 40-year. 320 billion project. wunwsclencomag.org Infrastructure decommissioning and removal. Removal of Carbonton ‘Dam ceslted in rapid ecovery of tho federally listed ondangerd Cape Fear shines (owopis mekistochol), found ine former impoundment ater <2 {ees of removal. The dam was removed to generate emironmental restora tion credits, which nee then sold to ost stream impacts ebehere. Policy Directions and Exit Strateaies Infrastructure policy should do more than fund projects; it should set national priosties and initiauves, The National Commission on Infrastructure would set such strategie prioti= ties and consider infrastructure financing rha- Dilitation, and maintenance, Rehabilitation under the NIIA includes considering remeval of infrastructure that is deteriorxed oF no Jonger usefil. When infrastructure has been decommissioned, ecological restoration has ‘been aside benefit Priostizing decommission- ing sites based on a combination of ecological, ‘economic, and safety concerns can benefit multiple Sakehokers, possibly ralucing over all decommissioning costs(/5) Infrastructure decommnissioning to oocur during discrete windows of opportu nity These may be poliey-relaced (epi tion ofa dam license). natural eg. Hooding). or through deliberate legislation ( However, politcal wil for sich expenditures. is difficult © maintain, particularly during transitions (20). A less. broadly mechanism is th ‘market lke principles in which infasructure decommissioning is use 1 renerate credits to offset environmental impacts elsewhere reaove) igreaiest esson from current aging Infrastructure is the need for exit strategies, ‘which vary greatly among infrastructure types. 3s for decommissioning dams are sur- and vague (17), whereas policies for decommissioning offshore platorms are ‘unambiguous (15). Because the costs of decommissioning and cleanup for infrastruc= should equire provisions for decommissioning SCIENCE VOL3I9 POLICYFORUM t as part ofinfasucture license or lease terms, perhaps similar wo that for offshore facilities, ‘where bonding requirements are specified and ‘based on the estimated cos of fall removal Any infrastructure policy approach must confront the national conundrum of pressing infrastructure problems and contining envie ronmental degradation. Specifically. National ‘Commission on Infrastructure should squarely face decommissioning as a viable option, and the environmental denctits gained through such decommissioning should be assessed as definable benefits and leveraged when possible and practicable References and Notes Arian Scie Crt grees ASCE eprt Crd Jovazeries native OSCE Nr, 2099) 2 ene cere Se oe on’ Rossen (Comtesse ini rs, mde 2002, aera earl Cec Neg onoeig fe Eel Umar oe Keats (apenas Pes. sigan, C209. Cason ear, sence 308 808 205) 5 atm tneanmentl gece OfsrOiend Govan ewte Patinn, s Ne 19) {6 NC. Conpnsig Ween Lsestinter Cer rr Act ara ey Pres, Wash OC 2m, 1. Eten a, Scene 308 638 2008 1 NC, rages Taner ecerng te vera Osea ‘dames pce nation, 208 9 Gl noter ein Bol 31.339 0909 10, tein eve Ply HevenComte. Pe anor! ee Calege Cede tnerercy anager Acton, Vosngon 200. 11, MM Dales romerpaegy 227 200) 12, O-Gabtero, scene ae 721 0978, 1. 8 Mate, No Pecillo, ng, 14, Dae a. Remesiog infos Communes na Fore etree ence Deveepment Ata C203) 1, Mtr Paseo, Dero ee 36, 125 16, 0. Seoete M.S. te, cen Cos anny. 4.21 1. PS ea ar lrg Se 20,129 GO 1, MLS reat fh Bo 108,388 000. 19, AL Cete et ct Center Batya ny es (US. eprtnents Defense, wsiraien, 9.1998, 2a, Naat ee or Pe: ers npr PND, Fag Feuer eget Peace oa Congressman ves OC 1989), Soe: eps (U5 Cenas Bret, 2007 ripe Fer Higa Aniston GHA, DONT sce es aor High Sat Sarco oa eta stance oer Sone siptatt or cnens ene USC, Herons Ina ame, ter, 207 et Pipes tmameett reece age, he eae Iter ing noe estar Copan. 202: and tae poms Wer Magee! Seve US Deparment the nei, 207) rst [stan mambo ie ey for he US oer certe ta ‘a, 21, face US. npatan tee 09061502 147s ane WS DOC O82177) Pcp be anotengrant ne vow any ipa aa fon Wena Seen sing ML) Sa nd sa n20eeee 19882 18 JANUARY 2008 287288 NEUROSCIENCE Organizing the Source of Memory Elizabeth A. Grove he hippocampus is essential for a I icaningfal human life. Without it, we cannot form new memories, Engaging inva normal social life would not be possible because new experiences would fade from memory within minutes (1), In keeping with its complex function, the hippocampus, putt of the brain'scerebral cortex, hasadisinctive, shaped called CA, CA2, and CA3,is dentate gyrus, packed sheet of granule neu rons, curved around the pyramidal cell Layer. Dissected free of the brain, the entre rolled. snuctare bears an uncanny resemblance 1 a seahorse of, in Greek, “hippocampus” (2),On page 304 of this issue, Mangale er al. (3) reveal how this intricate stracture develops in an embryo. ‘One hippocampus develops on e of the brain, close to the brain’s midi Within this midline lies the “cortical her,” a source of secreted proteins that regulatea vari- ety of developmental processes (4). Among these signaling molecules are bone morpho= genetic proteins (BMPs) and Wingless-Int (Wnt) proteins, whose functions include directing fields of cells to develop as a spe~ cifiecell type or part ofthe body: Although the cortical hem has been a likely candidate to specify the hippocampus, evidence huts been surprisingly Investigators have focused on manipulating Wat or BMP signaling in mice, the favored species for stiles of mammalian develop= ment, When Wnt signaling is disrupted, hip- pocampal fies are reduced or lost, but this «ean be explained by decreased cell division or imerease cell death (3-7). When Wat signal ing ivaciivated throughout the embrycnic co . small patches of cells resembling hip- pocampal neurons develop out of place: how= vero particular source of specifying signals ished (8). Loss of BMP recep- ty causes lossof the hem itself (9). Mangale et af. identified an elegant solu sionto the technical problems of transplanin aheminio amanmalian brain, to test its effect, ‘on hippocampus development. They used genetic methods to implant several hems into he autor 5 in the Oeparment of Newobitoy University of Cheap, chicage, 1 60837, USA. E-mal aromegbis van es T8JANUARY 2008 VOL319 SCIENCE suc corgrieer corganieer Organizers. (op) Transplant ofa second Spemann’s ‘organizer in te erly frog emibiya generates a sec- ‘ond body ans, resting in joined, twin tadpoles. ‘ottom) Insertion of entra conta hem organizers ‘into the mouse forebrain induces etr3 hippo «cempusesto frm (ight nal brain let). the embryonic forebrain, next to tissue co petent to become the hippocarpas, In the chimeric mice that they generated. normal cortical cells intermixed with cells that were labeled with green fluorescent protein and genetically engineered to lack the transcrip tion factor Lhs2. In the absence of functional Lx2, cells are predicted toexpressicomple- mentof other genes typical of thehern, includ ng Wins and Bp (70.11). This method of introduc a chimeric forebrain would not have worked ‘without an imeresting cell property demon strated by Mangale et al. Lhx2-deficient cells prefer to stick t0 each other, 2s do cells Signals from the midline ofthe embryonic brain direct formation ofthe hippocampus expressing Lhx2. Thus. in adlnesion created patches of e normal development, M. that such cell affinities help co form a sharp boundary between the H/2-negativehemand Linx2-expressing corte. Next to patehes of hem tissue, new hip- pocampal fields emerged in regions wit Lh? function. The dentate gyrus, CA3, and. CAL were oriented in the appropriate order beside hom tissue, The most straightforward imterpretation of these results és that Lhx2 imukes a region of the forebrain competent (0 ‘generate cerebral cortex, and the hem organ- izes hippocampus inthe compesent tissue entification of any embryonic \vith “organizer function” recalls the historic discovery ofa primary embryonic organizer by Spemann and Mangold (2). Ata stage of development when the salamander embryo Jooks ikea hollow, collapsing bal of dividing cells, transplanting @ portion of the bal “Spemann’s organizer.” to another embryo duplicates the body axis, resulting in con- Joined twin animals (see the figure). The hema ‘snot a primary, but “secondary” organizer It regulates a particular component of the bods. the hippocampus. Nonetheless, the position of this secondary organizer deter- mines a major develope to its function, because information travels largely one Way. from the dentate gyr through the CA Fields. Disruption ofthe axis after birth eauses profound memory loss (1) Remaining questions concern the mech nisms of axis formation and the molecular identity of hippocampal organizing signals, including, but probably not limited to BMPS ‘and Wnts, Answers may shape investigation ino the genetic hasisof human disorders that implicate the hippocampus, such as epilepsy and schizophrenia. As for experimental mice ‘engineered to develop more than one hip- ocampus per hemisphere, are more hip- Pocampuses beter or worse for their memory after birth? Orher pa fied in and around the embryonic corte, based on their expression of genes encoding secreted signaling proteins (4). model has en proposed in which organizers regulate the cell types, subdivisions, and finctionally izershave been iden wunmsciencemag.orgspecialized areas of the cerebral cortex. Organizer function has new been demon- strated for two such signaling sourees: thecor tical hem, and a region ofthe forebrain that secretes fibroblas! growth factors to direct, neocortical organization (/3, 14). A broad implication of the work by Mangale er af is that it tips the balance in favor of this model, transforming it from a provisional toa work ing model and a platform for future studies of how the coetes is frst organized (4,15). References 1. LILSqire Memay/ondsan (ver Un Fes New Yea 587, 2 Hs, Ferg. andometelNevmoneny (reeman Mew Yr 1980. 2. WS. Mangala etl Soe 39,30 2000) © CWitagatle 6 Gove, Cr Op. erebo , soon. 5: Sia tel One. 1 [ere ewepen 127, 469 ooo. 7. AMigjuepie ot, ee Me, ed 6. USA 40412919 2007. PERSPECTIVES t 8, Mahan et oe 32,23 2007. 9, Meranda, Delon 13,3789 2007 10, Sterna oa tch De 108, 16 (00. BLE Sade ED Ferby.C A Math Men 32,591 (coon. 22, Speman H Mangal, Rach. Ee Mech, 20, smn, 13, fc timo A Ge, See 294 1071 Goon. 14, 5 Gore ot, veep 130.1903 2008. 15. 0.0 O\eay etal. teran $6,252 (207) ASTRONOMY Orion Continues to Surprise ©.R. Doll nd Loi Townsley he closest example ofa region of mas: sivestar formation is the famous Orion Nebula, which is easily visible even ‘vith small binoculars. The accompsnying ure is a mostic of Hubble Space Telescope optical images and shows the rich detail in this, beautiful object. The brightest tars in the mid die are the most massive of the several thou- sand stars that form the eponymous cluster, Star formation within Orion's Giant Molecular hascontinued through recent times. The Orion Nebula has frequently been one othe firs sar gets viewed with new astronomical techniques, (2), bezinning with the optical telescope, the Introduction of astronomical spectroscopy, and mos. 1Nelength observa- tions, Now on page 309 in this issue, Giidel ral. 2) present x-ray measurernents with the X-Ray Multi-Mirror (XMM)-Newton satel- thatreveal something new: hoe as that can ‘only be powvered by hypersonic winds from the massive stars. Understanding these new observations requires undersianding the threecimensional structure ofthe Orion Nebula (3). The location ‘othe center of star formation is somewhat like Goldilocks porridge, neither too deep. within the cloud (or we wouldn't se the ars) or too close to its surface (or there would not have been enough material to form the cluster stars), This selection effect means that the Orion Nebula is an irregular concave thin ng ionized by ultraviolet pho- fo the hotest sta. The ionized .2.0'00 ie he Depatnent fyi and Asuenony Vancebit Unvetsty nastle,1H'37235, SAL Towsley Is in the Departnent of astonomy ane Atrophyice, Penojtania Sate Univrty, Unter) atk PR eid? USK, emai conde vance ey, tee pecs voya:sciencomag.org thin layer is near the surface ofthe molecular cloud. Themassive stars liest the center of the cluster and are on the observer's side of the ionized layer. Finally, there isa thin layer of asad dust that les inthe foreground and forms a semitransparent veil throug Which we see the nebuls. The thickest postion region of the nebula, superimposi thumb outline in the figure ‘The cavity between the ion- ized yerand ihe veil opens to the lower right in the igure, and there is a slow flow of low-density gas in that direction, The ionized yas of the visible nebula hase temper- sure of about 9000 K. This is the equilibrium estab lished benween the heating that occurs fiom the excess ‘energy of the photons that ionize the nebula’s hydro- gen and the cooling that ‘occurs by collisional exett tion of heavy elements such nitrogen, and oxygen. The hot gas ob- served by Giidel etal. isata temperature of 2 million K and so must be formed in very different manner. Higher-temperauure ‘been seen in earlier studies fof oxher similar abject, bu the diffuse light caused by scattering in the Chandr Observatory’s eamera pre cluded detecting any diffuse souree of x-rays in Orion as carbon, Trapezium resion tot SCIENCE VOL 319 Observations ofa nearby nebula reveal an unusual region of extremely hot gas whose high temperature is te result of wind from massive stars. ay also be that al obscures emission (4) near the brightest star. It the thickest veil mae from hot gas inthe central region, “Thenew study shows twoextenda regions 2 million K. one in the 3:30 o"clock direction from the compact group of four bright ars in the cemer ofthe nebula and the ther in the 5:000elockdirestion, both Ii within the outer rim of the extended nebula, Cradle of star formation, This mossiimage depicts 12.8 equate ight year portion ofthe Orion Nebula, the nearest region of mess mation. The bright area in the upper left canter s the well-studied 3 of 9O0C K gas, and the muck larger nearly lower righ isthe Ete fuse ray emision fram 2 mien K gaewas ebsered by Gide! ta rulor fed Orion Nebula within which di 'B JANUARY 2008 2897 PERSPECTIVES 290 Staliht cannot heat 2st this high tempera ‘ure andthe power must come from thelyper- sonic wind eitedby massivesars(5). This wind carries about 10,000 times as much power as the observed x-rays. which vans that tis the prime candida to be the tengine driving the oray emission, The wind willovertake the eseaping low-density gas in the outer nebula, peoducing shock fronts that immediately reach remperatures of of kelvin, This is the same mech invoked.o explain be hoter xray emission in the similar but more populous Omega and Rosette nebulae (4), but many questions remain in Orion. Why i the x-ray emission confined to these two areas? Is there a chan- neling elect ofthe stellar wind, a rapid coo ing ofany other shocked gas close tothe sa, or does extinction in the veil simply prelude observation of hot sas inthe opal bright est part ofthe nebul? Researchers can answer these quesions hy reanalyzing the earlier Chandra Observatory EcoLoay ‘observations and by employing careful sub ‘waction ofthe instrumental scattered light and ‘mapping the extinction of optical light from the Extended Orion Nebula, Once these obser- ‘ational questionshave been resolved, the ball ‘vill be in the theoreticians’ court: itis they ‘sho must then confiont the problems of why the Orion gas is at 2 million K and why itis located where itis Rterencesané Notes 1 CR Oa, he Orn ebuo hard Uni res, Cambie MA 2008) 2 MGitl eel enc 289, 309 200; ube ‘tee 29 Nevember 209 (101126 119826 3. ERO, ara Re diton Asp 39.99 Goon. 4A. Li Tons Ameo £593, 674200) 5. C tether eto, eopy. 326,356 (908 E. The popu of rear wes apperiedin pa bythe space Teese Scene nue grat CO 1096) nde ire and Chonda iy Observe (rant GOE-70008 3 Pevwsyoi Sate Uses ao.a265eee. 1259476 Managing Coastal Wetlands Ivan Valiela and Sophia E. Fox ‘Wetland management may be improved by evaluating nonlinear relationships of economic value and ecological services. bout 30 to 50% of the area of Earth's degraded during past decades (/). The Fates ofthese losses exceed those of the much better publicized losses of topical for- ext. Loss of coastal habitat areas reduces key ccological services; for example, fish and shellfish stocks may decline and shorelines ‘may be destabilized (2), On page 32] of this issue, Barbier ete (3) find that habitat area ‘might not relate linearly to economic value of certain ecological services furnished by coastal wetlands and argue that ecosystem ‘management sitategies might benefit from consideration of such nonlinear Functions. a basis for coastal preservation, Previous work has suggested that the more salt marsh area present, the greater the likely harvest of Shrimp in adjacent coastal waters (4), Sach linear results suggest that the best course isto proteet as much of these environ ‘he authors ae at the fcessters Gener and Boston LUnvesty Marne Pegran, Marne Belial Labrto, nds Hole, WA2543, USA E-mail Nabe et ‘eto T8JANUARY 2008 VOL319 SCIENCE _ments as possible: this has largely been the basis formanagement strategies, Barbier era show that for atleast one ecological service provided by weilands such as marshes and ‘mangrovesthe wave attenuation servi Costsand benefits of conversion. Ponds ae constuced to ise shim in areas previously covered by mangrove forests in northe vest 21, Indonesia, that protects coastal areas from storms and. isunamis—the relation between service and wetland area is not linear, but is decid edly curved. Such nonlinear links have consequences for management, Barbier eral. focus en the ey management problem in mangroves: con version of forests to shrimp ponds. Is thece some best solution tothe dilemma of weigh- ing the value of ecosystem services against the economie value of the shrimp crop? The authors calculated the “economic value” of intact mangrove forests in Thailand, incorpo- rating the nonlinear behavior of wave atentsa- siomas well as the value of shrimp production after mangrove conversion. Economie value peaked tan intermediate level of conversio fiom mangrove forest to shrimp ponds, Thus, itmay be possible to define an economic opti- mum for comersion of mangrove forest area ‘oshrimp ponds. “The work is based on the premise hat an ecological function can be converted into a ccurrency ditectly equivalent to money (5, 6. We have at least wo qualms with regard to thisinceresing idea, First available information pr ‘em valuation are often forced to make dering {and not always compelling) leaps of faith 10 comertecological service to monetary vals Seconel, by putting a price on 2 natural environment. we put it up for sale—and the ‘monetary value of natural services will scl dom reach the value aecruedby conversion to industrial o¢ commercial use, For instanc some years ago there was a proposal to build an industrial plant on a weidand parcel in the Hockensack Meadowlands of New Jersey. AS counterpoint to the proposal for const ‘ion, decision-makers asked fora valuation of the wetland, which came to about '$9000 per ere, per year. in per petuity. The builders offered $200,000 per acre and showed that the add jobs, taxes, ete. ‘would by far exceed wetland “proceeds” in the short andl long term, Today a refinery stands That said, there are several valuable features of the results reported by Barbier ei ul. They show that considerationof nonlin- ear aspects of coastal variables improved the ability to incorpo: rate conilicting demands in mane agementstatevies. Use of nonlin- ‘ear relationships could elucidate diminishing ecological returns sssociated with different amounts ofare wunmsciencemag.orgvide an objective way to decide on the degree ‘of preservation of different environments of the world Scoasts, ‘The results also suggest several fertile areas for further esearch, Firs it isimportant to determine how general nonlinearities be; in the Thai mangrove example, the ‘of cuasial protection by wave attenuation was nonlinearly related to the area of mangrove loss, but yields of fishery and wood products were linearly related to habitat area. Second, scientists must establish whether the devia. tions from linearity are significant in both ‘ecological and management terms to make sure that implementation of policies resulting from use of curved functions leads to ewectable differences. Thied. mote solid evi sdenceis ncedd that a serviee is actually being provided: for instance, there has been consid ‘erable debare about how much wave protes tion is provided by mangroves 7-17). Fourth, uwure stuies should more comprehensively cover the suite of locally relevant factorsinflu- fencing economic value, Barbier 2) made 2 ‘good star toward addressing local issues by ‘quantifying the net effects of economic costs and benefits resulting from conversion of mangrove forests to shrimp farming, but the valuation might benefit by addition of a few GEOLOGY kkey terms, In Ecuador, for example, the shrimp industry came to a standstill not be- ‘cause of deficit income, but because the sup- ply of juvenile shrimp in nesaby waters (used as the "see for shrimp ponds) was depleted by overfishing. In Eouador and elsewhere, the ‘mangrove conversion rate may be azcelersted by the need to abandon ponds that, asa result ofhigh-yield culture methods, have become to chemically altered 1@ be suitable for shrimp growth, Finally. Barbier e7 af. point to the need to determine ecosystem-based manage~ iment for coastal areas by weighit incorporating the interests of several pl crs: shrimp farmers, who will think about the price of shrimp, but are unlikely 1 con. sider long-term, regional benefits of coastal protection when deciding whether 10 dig another pond: ouiside investors, who ‘might not know or eare about mangrove services: and officials, who might be responsible for implementing regional ‘environmental strategies that foster eco- logical services. The report by Barbier et af. highlights the ‘compleaities involved in making the com= promises needed for future coastal manage- ‘ment, Research from the study areas poined Dreams of Natural Streams ural river look like? The autor isi the Deparimert of farih and space Scenes. Quatemary Reearh Contr, Uiersty of Washington, seattl, WA 98195, USA. Emel: aveg scacinion vwyn:scioncomag.org crucial for river restoration, an seademically young discipline that is rapidly maturing into a billion-dollara-year industry. The classic sinuous form of meandering channels (2) as ‘come to represent a natural ideal in channel restoration design—even for rivers for which such an ideal is historical fiction. Modern fluvial geomorphology—the study of rivers—evolvel out of the stadies of Luna Leopold, M. Gordon Wolman, and their colleagues inthe 1950s. As the field devel- oped. pioneering studies of streams in the eastern United States contrived toa stan ddard model for how fuvial processes shape rivers and floodplain environments. This ‘model has been elaborated upon and exported around the world, Indeed. these now classic studies provided the basis for the so-called natural channel design central to many tiver- resorationetfortsacross the United Stes (3), ‘Walterand Mecrits now show that some of the rivers studied by Leopold, Wolman, and HHeagues were not so natural afterall. The SCIENCE VOL 319 PERSPECTIVES I out above will show whether “bent” rela- tionships make for compromises that are not only evologically desirable, but also enable compromises for planned management of ‘coastal wetlands that are acceptable to the diverse stakeholders, References ane Notes 2, Care te loa Leo Cont Haat Uunsioen 1A Maer 10 Date 7007) Aisa ‘Meconeence aie tyme! sta 2, Cian leat ase Change cel coe 4 ACE Tamer insti the ef oes Wont {os Stroud £4 era Caen Maer CGnszratin,Snanah Gk, 1921 9997-104 5S. Casares to ete 397,253 (0997. 6. ER Batt fen Foi 22, 1712000. 7. tatirear Rae, Evert Cont Sh Se, 5,401 0200) 8, Rib are tse Cotl Se St 6,539 (ot 9 Je Yermao U Tampa, tine Cat Sha Se 3, 112000 20, J. Yen UhampaeysEtarine eat Sh So ed 1, Ti Reser Federation een A 412 BNoumber 20m) ase 0 dati Comes con ely 23, 59 2003, Supported by ST got OB 081600" ER mon26ene 1159471 Human influences have fundarwentally ‘changed river morphologies in temperate regions around the wore new study dies not challenge their fund mental insights into how the interplay of hydraulics and sediment transport shapes river and siream channels, but i light of the new findings, what constitutes natu nel form requires reexamination. The results parallel findings in Europe and the Pacific Northwest of how historic: clearing of large wood and logjams altered river morphology. Before European rivers were cleated (0 promote waterborne com= large trees and logjams obstructed y al blockages spit flow into multichannel networks of branching streams (4). Similarly, downed wood split channels into branching nerworks of small channels. flowing across slough-rich valley bottoms in the forested Noodplains of the Pacific Northwest(5. 0) Walter and Merrits now present com- pelling evidence fora similar chonnge that rad ically altered rivers in the eastern United States. Thus. a comparable transition from 18 JANUARY 2008 2917 PERSPECTIVES 292 multithread channels to meandering si thread channels occurred inal thee regions. This observation sransformation accompanied human develop= ment of iemperate forestrivers. ‘Waker and Merrits also help to resolve a ‘Natural or not? single-channel mean- > J. The evolution in the upper doublet of states () corresponds tothe correlated tunneling of atom pais 27, whereas the superexchange takes place in the loner one (i. Both doubles are coupled by fs-order tunneling processes i. SCIENCE VOL319 18 JANUARY 2008 295ESEARCH ARTICLES ators in if) amd |). Here, = dul) is the eflacive interaction strength with of! being the postive scattering Kent forthe spin sates used in the experiment, Ais the mass of single ator, andy 4(4) denote the wae funtion for 4 packs ealized on the lef or right side ofthe double well, The state of the system can be ase 85 eco cf the Ie and at ve siacs 1,0) and 0.1!) ae spin the following, we wal focus on the cvokion ofthe populaon inbulance = n,m, land the Net order pameteroF "span aaa N= Gig.) mye rym yi) seating with double Wells tally propared sty). Here, Inyuace ~ Ginya.n) donate the corresponding quanium mechanical expectaion alice and "ntti of doing intrasons (>>, wvhon stating inthe subspace of sing ocupiod ‘asp hy 7.) a the ctl highyng stucs 70) and 07)) can ony be reach 38a” rtamodine sues in seen ‘oder taming procs, Such proses ed 0.4 bonkeal Gur) sparcxcharge iteration hich couples the sates 7.) and J) (Fag TA) Mae coal, fran arity sin corfiguraion with al rt ees Uy = Cig = Uy (282 sonthondr hopping csrts ae deeb by an oxo Heikeboreavpe effete spin Hao. nim nthe int U>>J 2, 15, 2.30) Fae = eke = Jal + 585) - eSiS% where Six Tine Sin = Mtn and Sf = Cini. g)2 denote the coresponding sin ‘operators of te sstem with Sig = Si The efetve coupling stent Jo apes the se perexclange and can radiy be evaluatd by penwhtion theory up Ho quad: eed in the tuending oer, which yield, = 229 ‘When a posit bias 8 > 01 applic, the egeneracy of the fvo iterate Stes in the superxshange proses i ited (Fig. 1A). For J v= U this leads to a modifeaion of the ‘icctivesupenescangs coupling ith 209 J PAU ea) PUA) =2F UU? XV (1). By ning the bias to A ~ Ui its posible to change the sign of Ja nd therefore 10 Switch between feromeic and ontfenemagntic superxchange interactions. Far J <= U3, the picture of an ellactive coupling via wwe wal iencrmedine sats is again vali and the {all versal 1 y= -2PIU is ud 9 be hol ford For symmetic daiste wall (4 = 0) the Honvitoian Ea, 1 can be diagonals arayt- ically give 8 vali picture forall yaks of J and U wikia the singl-band BIL A come mks fas seven hy the spin pls and singlet se £5) ~ (1) © LV ad the states 3) = 296 18 JANUARY 2008 VOL 319 SCIENCE (110) (01/2. Two of the eigensats are linear combinations of |) and), where the one hnaving the lager overlap with) is the sound state, The i s) and the state aleady cigenstaes themselves with cncrey 0 and U, respectively (Fig. 1B). As a deci cone ) cannot be read from the iii 9 Ny V2. Therefore, the dy smical evolution of the spin imbalance con- tains only two fequencies metho) 0 ‘The exaction of these lequencies foe tine- resolved measurements allows fr the determi: nation of 27 = h oy ant U= Moy — 02) ‘within the BHM. Because these fequencies can be measured swith high accuracy, we are able o observe de- ‘ations fom the simple BEIM. We obtain a is correction by the inclusion of nearestneihlor interactions (3) in an extended two site Bose Hubtant model [FBHME: see Fg, SU in GP) ‘This modification introduces the inervll inte tion cengy Ui = 2 * hvtinwatayte and a correction 10 the tunneling matrix element, ‘which becomes = J A, where Ad JGoopng(oyx. The interwell interaction leads to-a dincct spimexchange term, which in the Timit U >> J reduces the comecied super: exchange coupling to fa ~ 272 ~ Cie. Although we find tha the Gomeetions wo the pare twosite BIM are not negligible in the expsri- smentlly wlevant paramser region, numerical Fig. 2. State preperation and detection. (A) Spin trip- {et pairs ae created on dou ‘ly occpiedlatce sites and subsequentiy spit under the influence of a magnetic field ‘gradient to obtain artfer- {0,0 ‘tromagnetic Néel order. (B) W-w - \w. cacultiors basa! on the multand Schrier ‘equation show thatthe diect exchange can nov cer overcome the supereacharge coupling ten (Hig. SI) and therefore change the nature of the srcund slate wo be aniferomnagnetic, This is ‘agrooment with the Lieh-Matts theorem (32), Wich states thatthe ground ste for bo bosons has to be spin dynamics baw neighbor inially prepare a sarge of taco neta ors wth two relevant tral ses) and Up ina thro-dimensional GD) aray of double walls ith Nécitype aneromagnatic onl TH along one spatial direction (Fig. 24). ‘State preparation was started by loading a "Rb Bons insin condenses of typi =4 = 10 atom in the F— me 1) Zeon sblevel sch no dacomive thom fection om imnetc tip with high oft Fld ito 3D ‘ical hte of doubksl poi (25). Ths Rrctatioe" potential is chained by superine poning on onc two sting Fight ls with Pesiodiiy 325m (shor ice) and 768m Alon lc) and adda sing waves wih petedity 420m on fete perpecuras (2, Contulingall depts ane he else pane ofthesho anions tice allowsoncto tncthe Aoublewell configuration in rm of the Hae ‘oan prarrs Cand. The depts te Intices a given in units of hsbc vel ney f= (QMF th 265 tutthaic. Th eaingrampswcr opti to xor an cccpetion oft storm er ca well and avd bosin wo Hishsr Nitin Isck (47) Afr mong the dahl wet hy leeletet) Evolution of thecigenener- ag fiOzi aT ies with respect 10 the o2ils) ~ ‘groure-state ene during the spliting with Ving = 10 Ey, Veang= 25 Fy a0 gradient of &° = 17 Geom. ‘he notation liv] denotes Ep (KH) the number of vibrational 5 ‘excitation forthe fist and second paride. The inset offoonin, shows the eigenenergies and (eigenstates forthe fe ral bare height of Van = 44 €,.(€) Detection ofthe population and spin imbs- lance. The population ofthe left vets tansfared to a higher vizatoral level of the urdeying. long atice »> wel 25,27. Subsequent band mapping anda Ster-Getch filter allow one to determine xt and N( from the ime offight images. nes scioncemag.orgramping down the shor latice, a microwave pid adiabatic passage was used taste ll coms into the = 1 mp = 0) state. Subsequently. the magnate op is site off and while mainain- ing a homogsneous offset field of “12 G, atm Pairs wore coherently tansfred from p= OF ‘me~ 0 into spin pe pais (1-1)*\-13: Dy by means of spin-changing collisions (33, 44), ‘The two magnetic sublevels np ~ #1) corm spond to the nv spin states |7) and |). The remaining atoms in the yy ~ 0) stale €2, 09 singly occupied sites, are transfered into /F~ 2 ine = 0) and removed a feng sequence befire the detection 37). Firally, the shor latice was ramped up slowly in 20 ins, thereby inking a coberent spliting of the atom pairs and leaving the double wells in a site with one atom on cach side (20), For the time of the ramping up. magnetic fekl gradicat of 8 = 17 Gfem in the dietion of the superlative is swithal on. Therstre cneracy ofthe states.) and 1.1) im the douite well lined by ~900 Hz, Which enables an adiakutc aging of the state tring the spi 2B). Nuv rca inepration of procedure yicids an expecta fidliy of 97 for eweating an antiferemagnotic onder along the axis of the superlatice. The mean pou lation imbalance a() and spin ambalanee (0) ‘of the ensemble of double wells was devected bby applying a mapping technique (24, 27) combined with a Stera-Gerlach fiker (Fig. 20), A maximum spin. imbalance of 60 to Ys was observed for our initial state corresponiing 10 2 probability of 80 10 SS% for having prepared the desiod state (tl). We Ielieve that this measured value is mainly reduced as a result of our detction method, Directs the mapping sequence (23) and mixing of the spin configuration and thus a reduetion of the measured Néel order paramcier 31), Fig. 3. Spin and population dy- os namisin smmetic double wells The time evoution of tne mean 00 spin) (ve ccs) and pop- ulation imbalance 2) (orown & Time-resolved observation of superex- change interaction. The oyun chanics ae iniiacd by wily mmping down the shot late and‘heeby the doublesrell bare in 200 ps thus substantially creasing the tan acting and wipereschange couplings. A Re Kein he aytem evolve fra hd time the spin configratin ws frozen out by raping tp the basen 200 ps, quenching both J and “Reagan. The mcasirement ofthe ensemble sNerages 0) and NA) caida de Seri above hase typ time tees blaine by his pmo ae show in Fig. 3. or ow farce Alps UU Djs observes prononsd tine vohain ofthe sin inblce NA) comsiing fF two hogan) components with compa nian an Fsqucces (i. 3). With oe Srensingimraction oeay Ube 90, te froqncy ro inersscs, Kans kw com ponent wih lt ll amie cd a Tima iheoqaeasy modal vith sal ample (Fig 3B) The fst compen cor fesponds 10 fistordr tuneing dh i the Coupling of ant), which becomes more and more off-wsonant asthe base height inccased an therefore is decreased. For JU <2 1, completly suppres ad the nly piotess Vib is the sopsrecchange ‘ciiton (Fig SC), For all baer beg he population imbalance x) sti Ht, eras ing that eventhough strong spin comes are seat inthe sytem, no nt mas hw ean BS bse or or iia wate, We ft he fee 4) with sum of two damp sine waves wih yetble heer /De ond npn fy For th damping We awume Gassian chanelersi wih Vecampirg consents Ya “The oats ofthe Fite diplyed in Fie 4 Toe Vinee = 18 F, (Fig. 4A, ime, be can ed nb’ single fsqusney component comspon ing to he upeenchang clin with APC i fllroathe ample (i. 4, Wesel toobsce this Fequency down 104.80) Heat res) are saoam fr thee ber ‘er depths wthin the doube-nell & ‘The measured taces forthe spin imbelance ace ited with the sum ‘of two damped sine waves (blue lined The population imbalance 2) stays fat for al traces. & Population imbalance x(t) ‘and spin imbalance N,(t) 8 nw scioncemag.org 700 t(ms) SCIENCE VOL 319 18 JANUARY 2008, ARTICLES i IU ~ 0.023 for san = 20 E,. The damping of the signal can be explained by the inhomoge neous distibutio of coupling parameter due to the Gaussian shape of the latice bears, which leads to a dephasin of the evclation within dhe ‘ensernble For Vggq 2 17E,, ational damping mechanisms liketunnelin to empty adjacent at- sites (Getets) or small residual inhomoge- rnsous magnetic fick gradients hocome relevant and limit the measurements (Fig. 40). ‘The comparison of the results with the theoretical pretitions by the simple BHM shows ‘atstically significant deviations at low barrier ci mot be explined by our unee tutes in the tee cepts. In this region, the EBHM can model the experimental dua much RESEARC! more acouratcly. This can be understood by the with decreasing anier and hegins to nicely inflosnce the dynamies (3). In ft, the EBUM description vies We; ~ 03) = U> 3U and therefore directly exphins the upward tet of this frequen ciference for small shar-latice eps (lg. 52). For large barrier cights, our ‘experimersal tare compatible with both mack cs within the ungerainies of tie lative depths. However, hee the predictions of the EBHM are always close o the measured values ‘Sign reversal of the effective coupling parameter. To domonstate the controllability ‘of superenchange interactions, we investiga the spin dynamics wih an applied bias on the double wells fora short katice depth of 15 E and the same depus forthe long and tansverse lntce as before. Staring with an inital ani- Somagncically ordered site, as above, we Sirs the system evolve in syimmetie double wall (9 = 0) for fo = 4:5 ms until the fst node {0} 0 ofthe spin imbalance is eached for the sae (f) #5) V2 Cg. $1, Mer fheeriny out the mative phase between |) and 1) by ramping up the powmial barter, defined potential bias Ais applied and a sssond evolution sequence with bold time = 1 — ay is intiaized by ramping down the shor fate again 10 13 E, The subsequent detection follows the scheme described ahove, Figure $A shows the evolution of the sin iemhalance V.() in symmewic double wells szcher with the time traces for two different bias nengies A > U, yielding an effective eo pling of Jed) = ef ~ 0) and vel S~ 002, respectively. The sign reversal of Je de 10 the introduction of the bis is directly visible by the change in lope of de spin iba 4-H shou be noted that the m0 ‘oda fx bosons does timp vclation © LLeb-Matts theorem, because the new ground lai in this regime isthe spin triplet state 1.0) and the superexckange couples the fist and scoond excita sats 5) and), which have reverse order for A> U (Fig, SB, let inset), ‘The intoustion of a norzero tit leads © an increase semsitity ofthe exchange frequency to Mactuaticns de to the iahomogensitics inthe 297i RESEARCH ARTICLES 298 wave to the time traces obiinal for various tts yiblds the fequensy curve shown in Fig. 3B fogetier with the amplitude of the oscillaion nay of double wells, which are most ellscive around 4 = U, Therefore, the damping of the signal due 10 dephasing is stonger as ap poate UA Gots sige dmpensd sie Seringatand $0 He he ocllion Sequery 0 T 3 Bt © ayn ofore)2n A g ote came oe B oof B ee. s ee : 31 FT Pm, & Zost et q E c ‘ 001 & of) 4 8 8 10 12 5 0 1% 2 Veron (Ex) Vohort (Ex) J. Frequency components ofthe spin oscilatons (A) The frequencies obtained by iting the sein imate data for various values of Ving (fd and Due cies), 25 wel as the frequency diference and the geometrical mean (green and blac cles), are compare to the theoretical predictions ofthe simple BM (colored recions). The width ofthe regions represents a 2% uncertiny in all atticecenths. The 20 ‘error bars on te it results and extracted values are inal cases smaller than the datapoints. The dashed lines are the predictions ofthe EBHN. Th inset shows the data for high barriers, where only a single freauency component can be identified (B) The plot of relative amplitudes versus short lattice depth ‘shows the suppression ofthe fst-orer tunneling process toward smal values ef JU. The dameing ‘oefients fr the two oxilatory components versus ate depth are displayed in (C). The ertor bars denote the 2a unceraiaties as obtained from the fit results Fig. 5. Sign reversal of the supererchange cou pling. (A) Alter a quarter ‘of a superexchange 05° cilation in-symmetric double nell (lack and ‘rey dots, JU = 0.08), a defined bias applied, leading to a reversal in the slope of the spin i= balance and ongoing o:- «lations. For AU =1.32 (blue cries), the magni= tude of the couplings almost the same as for A = 0, whereas it is ‘decreased to ~50% for AU =1.70 (orange cit- ‘les) B) The fit of the data for various tts toa damped sine yields ares- fonance in frequency at A= U, wheee the ampli- tude reverses sgn (right inge.The soles show the theoretically expected frequencies forthe given Parameters where the shacing reflects the am- plitude ofthe corresponding Fourier comportent. The left inset shows the three lowest eigenenergies of the BoseHubbard Hamiltonian versus the potential bias. The error bars denote the 2c uncertainties as ‘obtained from the fit resus. & 2 8 U0, JU 008 ust eo wuet70 6 & Spin imbalance N,(t) 03 02 oO Frequency (kHz) 00 15 40 Potential bias A/U 05 20 18 JANUARY 2008 VOL 319 SCIENCE reaches a rsonance for VU = 1, where aryplitade rovers sign, leading tothe observed time ever in the dyaamicn ‘Summary and outiogk. We have dernonstratad limowsolved msasurcraens of supersxchange spin interactions between ulacold akms on rsighboring lace sites and have shown how to ‘omro! such intortions with opvicl. spertat- ties Comparing the measurements to theoretical Predictions of these spin interactions fiom fist Principles, we fi exellent agreement of our dat tan extsaded wosite veion of the BUM, Athugh superesehange interactions. become ‘exponentially suppressed fr dep optical latices, the coupling stength 2Javh can be several hhundied este for latioe dep of ound 12 10 1S Band thus almest a for of 1000 larger than the direct magnetic dipole-ipol interaction of Rb atoms on neighboring ltice ses. Coupling srents one order of magnitude larger than the ‘ones shown here, however, could stllbeackioved with the use of electric dipole-dipole mediated spin inkrictions. between grount-sare polar rmolsculs (35). ‘The demansrated scheme 19 change the supereschange coupling srength and reverse the sign of the spin inteaetion can also be apphed to the full 1D chain, offering now pes sibilities fore ia-spin interactions in ‘optical latices. It is BoM, eg, conceivable to ‘gine a setup with ferromagnetic interactions lng one and antieromagnetic iniractions along another lative direction. Funhesnor, one ‘dynamically switch between ferro and ntiferomagnetc interactions along a given lt- ection and follow the subsequent dynam ical evolution ofthe quantum spin system. ‘When the present kaxlng scheme scarred ‘out without any magnetic gradient eld during the sling press, a vakenoe-bond solid (VBS) type spin state (3, 26) can be ellicently eng may of robust Boll pais (37, 38). In principe, the superexchange interaction can be changed to be of ing type, ea. by tuning the inierspeces scattering length (Z3), Theebs, it can be used to ‘rene large ensangled states out of the ial slsconnected pas, which have been shows to be powerflesouress for mezsuremert-base un ‘um computaten (/2, 39), Moreover, controlling the superexchange inrations along different lticediostions also fers novel posible for the generation of tpological many-body aks fee quantum information procesing (1, 5}. References and Notes hing 2 Ps A38 253 0505, 2 ti tesering 2 Py 38,12 0920. 5. A Auch, tneaeg ethos and Queers Moyen (pes, Bei, 208 4 PARM Ba Poe Soe Lad Se A222, 66 fase. 5. PAM Bix Poe RS ladon So: A323, 74 fasan, 6, Neteserbeg 2 Mya 4,649 (920 ALA tamer Physic 12 0938. 8. dean Py Re 79, 350 (950) nes scioncemag.org9 8 ee M. Mgan, KG Wer, Rey Mod. Phys. 7827 an 10, 6 Bar, tas DP, Over, My fe 5, arcu AL | Pt et ay Science 309, 2180 009) BAW Rey Vite Le Denler, MO Ldin, Pps. Be let 99, W001 (207. BI tye ene Py oat 8 0 14 A ioe, An hy 323, 2 2000. 5. CA. Duan € Demy MD ai Phys Ro Lt $4, voce G03, 6 Saves Py Re. et. 93, 030401 2000. YF. M Lewontin cal 4a. A 6,203 (207, 1. Gen, Lato Me] stir Sarena, 1 607. 29, C pn © aol Wi PZ}. Cea, Quanta ng Compe. 3, 26.2000. 2D. A Ses Kade Phys a Lt 89, 274 (99. 21D Joh Bega) | Gin, Gone, Zeer, Py Rew Let 82, DTS (999) 22. Nana of a, etre 25, 937200) 22 ML Anere tc, ate MB, 452 (2007 eves Hips 24 LN, dn Ply Se 8, 232 (526) 25 Seb Sub, Medel, Pe, Pat, Pp. Boe A 73033605 2000 26 Ley Slee Pys Ro et. 98, 20405 C607, 27 5. Fling tay Metre 448 1028 (OD. 2A Fr the pi Soles wel cu epee he inarctoneeres fer te feet ictombinoens ot prs ay yen fen pet 29, AB alo 2 Siro, Phys Re Let, 90, 10801 (con. Any Wot, E Dee, MO ain, Ay. Pgs 5,113 BOD. ster ond mettre ale gpg rst en Scene Dane 22. Ce, Ort, Myo 325, 168 0862 BBA. Nea ery Ps Bo Lt. 95, 190105 (as) 2A F Geir Ae, §. leg 0: Mand ch, Py Boe 473, 43028) (C0. 25. A Mehl, Benen Zell, At hy 2,381 co FALE. Rak LA ie Pye A269, 122 (08) on at RESEARCH ARTICLES I 237. Coot to Phy er, et 92, 20402 (aon, BR C anger at Py Ror ete 95, 060802 (2005) BY. FYerton, 1 Cac, Py Re A 79, 0030200 (on, 42. We aire hep dcusins ih 8 Paes ted lunding bythe Seth Fenceigence hat the trope Union (OUNU, SCAU, ST i oe Dee of Serie Rese (UR, A455 97 13000, td Paar Foundation 51. scaled ona ‘agp rom the oat ci ef xeon rates th sang coelson WACOR). ‘Supporting Ontne Materat wmecenog are ISDEAYOCL 5a tet figs St ana s2 Relea tes 21 Setter 2007; spe 6 ember 2097 Fished ene 20 December 2007, 1oabeee 150682 faded hs nln ven cig this pope Natural Streams and the Legacy of Water-Powered Mills Robert C. Walter*t and Dorothy J. Merrtts't Gravel-bedded streams are thought to have @ characteristic meandering form bordered by a self- formed, fine-grained flocéplain. This ideal guides a multibilion-dllar steam restoration industry ‘We have mapped and dated many of the deposits along mid-Atlantic streams that formed the basis, for this widely acepted model. Thse data, as well as historical maps and recores, show instead, that before European settlement, the steams were small anabranching channels within extensive vegetated wetlands that accamulated litle seciment b ‘sored substantial erganic carbon. Subsequently, L to 5 meters of slacknater sedimentation, behind tens of thousands of 17th- to 9th-century illdams, buried the presetlement wetlands with fine sediment. These findings show that most floodplains ong mid-Atlantic streams are actully fl teraces, and histricaly incised ‘channels are not natural archetypes for meandering steams. ‘dred by a selfonme, fine-grained flood plain cmerged as the ameter mer fom based on piencesng stadies in mid Athi and westom steams of the United Sits (I). Today this ideal of alternating pools and rifles along sinwous channels with gravel point hare and fine-grained overbank floodplain deposits T: meandering grove: budded stream bor tectoniclly inactive mid Aiani Pisdmont of the Unita States are deeply incised, with seep era ing hunks, and cary ancmaloasly high amounts ‘of suspended sediment (7). Fine rained depots boreting many casera sircams are thicker than would be expected from just their eecert flood Asposis (J, 3. These Holovere dp ly fom broad surfaces, refered 10 as the Departren of ath and Endre, Fain ad Manta ‘alee, Pst fe 8 505, tance, A 17604-3003 us “These authors conbuted equal this wor. ‘To whom coespondence should be added. Ext beter lene LCM); dot metsgtondn. aon nw scioncemag.org fay” that were interpreted as hopin fo by acombinatcn of migating, meandering stam channels and overianke deposition of sts and clays (2, 3. 8). The geomety of singkechannet ‘meandering steams has been viewed asthe west of selfadjuting hydraulic variables in response to changing dkcharwe and sediment lead, and agricukure and uibanization have been cited Widely a¢ the eases of recent aggratation an adaion (2,3, 4,8 10). This patie of steam development and momphology has bas cone sided as typical of stvams an rivers in sable landscapes, We observe that cress of breached, historic silldans mere with valleyeflat surtices and that mest modem streams are incised deeply below this surface. This observation led us t hypothesize that a pid. regional transformation ‘of steam valleys had oeeurred in castem: Noh America, fam widespread aggradation a6 r2- sul of damming (base-level rise) to subscopent incision and lank erosion due wy dam breaching Ataseevel fll, We propase tht valley sedimen- tion not only xslt from accelerate hillslope rosin caused hy defection and agricukal SCIENCE VOL 319 development (8, 11) but ako. as coupled with widesread valley-bonicm damming for water power, after European setlement, foen the late Tih century though the carly” 20% century Danming ws essential to the estensve taping ‘of sediment in bro valley fats tht comespond to reanoie saris, We tes this lypotsis hy examining the fol lowing lens of evidence: (i) historical accounts ‘of widespread, intensive water-powered nll that impacted most Tist- to thicLorder streams in the mid-Adantc region: (i) historical maps sowing mabiple dams snd ponds, and our ob servations in the field and from ligt detection nd ranging (LIDAR) data of aggradation in these ponds thet caused sedimentation upstcam, imo whutaries ad snakes: i) hisercal, go Iogical, and. goochomical data showing rapid suimentation in valley botioms daring the pe= fd of early land ekearing: (iv) fel observations and remome-sensing data including LIDAR, showing that dowrstrean thickoning wedges of sediment grade to milldan heights and, hence that_dams produced temporary, higher base levels and (¥) field observations and hboratory data showing that the momiologies and fane- tions of prsetlement steams were substanially dlferoat flor these oF modery steams. We r= visited the same scams and specific reaches teed in cy stacs that pions modem hi vid gzomomphology, including fndumental ideas regarding meander migration, Noodpiin fern tion, hydraulic geometry, and Nivel sponse to tnd clearing. These steams include the Brandy wine River (in Pernsysania and Delaware) and Seneca Crock, Was Branch, and Westrn Run ‘in Maryland) (J-4, 8, 9, 11) all of which lie within the Piedmont physiographic province fof the mid-Attanie region. In all, we stcied Piednomt scams in 20 watersheds throughext Pennsyhania and Marykand (drinage areas fiom I to 1230 kn Bg, SD), ‘illdam history. Dam buiking for water ovr inthe easter United States began i the 18 JANUARY 2008, 299ESEARCH ARTICLES late 14008 and persist uni de early 1900, Dams and races that delivered water from the pond upstiear of a dam were built te run iron forges, fumaces, ming operations, and mest conmonly-—rnils; we tert all ofthese war powered activi. milling Before the adopsion ‘of steam engines during the Inte 19th eertury, every mill reqired 3 illum reservoir to supply a relatively constant head and selible supply of technologics fom Furope, wher thousands of ‘waterporvored mulls ined stems a carly as 1100 CE (22) {see supporting ene material (SOM) text, table S4, and fig. SS). Milling in- tensifed with econkmic zronth in early Amer ie, and dozens of mill acts dating 9 the 1700s encouraged mill and dam building. Our aralysis fof histori reconts in Lancaser County, Pem sybvania, or esample, indices that peak ill ‘eater Fay Ameican ster brought rai. devekynent ws om 17800 1860, Bat | sae boundaries in 1010 )- ant aensty in rtssg-km (ourder of its) ce z Wl > omoce¢so09 & i > 0029008 14370) | Ml >00stoa10 2577 BP all -c10t0020,:2000) MM -020:0005 1217 5 z 5 z & 5 z a =w Fig ww aw 7 1. Density of nater-ponered mills along eastern U.S, streams by 1640 by county (872 county boundaries are shown for 1820). The highest densities ae in the Piedmont and the Ridge-and- Valley physiographic provinces of Maryland, Pennsylvania, New York, and central New England. ‘ourties, southeastern Pennsyvani 2. Historic 1th-century mildams (vianges) on Piedmont streams in York, Lancaster, and Chester located from 2100 large-scale tonrship maps dating to 1876 (York), 1875 (Lancasten, and 1847 (Chest. The total numberof dams shown is 1025. Main stems of Conestoga (Lancaster) and Brandywine (Chest) rivers are highlighted in dark Boe 18 JANUARY 2008 VOL 319 SCIENCE powszed milling exer from 1710 10 1940 ‘Ow county-by-county compilation of US. man. ubscturing census dais reveals 65,000 water- powered mils in $72 counts im the este United Stes by 1840 C3) (Fig, 1). Water- powered milling was especially intensive inthe mid-Athntic Piedmont exon along and west ‘of the fall ine, whore seam gradi are cane ducive to milkdam construction and shi pons of dhe Coastal Pain ar’ in close proxinity (eg. Wibhingon, Delaware; Philodsphia, Pom sphania and Bokimore. Mayland The Palmer provided the bulk of manuldcturing and agri ‘uur goods to port cits anda lage portion ‘of the wheat and our for mid-Atlantic shipping (4,15). store maps show locations of mil bik ings. millponds, and races. dating 10 carly ‘American histry, and historic phots of dams date to the late 19% century. Laneaser, York, and Chester counties, Pennsylvania, have large. scale mid- 19 late 1h-contury historic maps fiom which we located 1025 milkéams (Fig. 2), whichis similar the number of mills recorded in the 1840 US. manutiacuring consis. Fr ‘county historical societies and the Pensy ania commoowealth’s inventory of di acquired photographs and recon of hundreds of historic milldams, The Pennyylvania Departe ment of Enviroamental Protection (P\ DEP) has an inventory of about 8400 dams in the commonwealth, of which 4100 are breached, ad estimates that $000 to 10,600 more might ‘evs (70). These eximates result in an average density of O14 to 0.15 dams per kn for the commonwealth of Pennsy ania. The possibility ‘of 16,000 10 18,000 danss in Pennsylvania is consent with tbe ~10000 mils sted for Pennsylvania in the 1840 US, manufacturing ‘exnsis corsiering tha il and dam bulking ‘continued throughout he 19h century In Chester ‘County, Pens ana, fr example, which inches much of the desinage area of the Brandywine River the 140 US. manufacturing census ists 379 waler-powered operations, and our exami- nation of historic maps from the 14s 10 the x60 revealed at eas 377 dams. The Brandy- wine Valley contained “the most notable con- ‘entation of mill indusries in the coke with 60 paper mills alone in 1797 (U7), Using density of 0.19 milklams por km? in Caster ‘Coane. which is only sighily grater than the density of waterpowered mills in other Pik mont counties of Pennsyhsania, Maryland, and Vinginer calcula from the 1880 U.S. ma facturing eersus data ‘Our field observations snd LIDAR snalysis {61> 200 millpond reaches) for the Conevious, Brandywine, and Codnus Rivers (among others) in Pomsyhania, and for Westem Run, Seneca Creek, Wats Branch, and the Monocacy River in Maryland (figs. S1 to SA), demenstate a sees ‘of aggactinal wees of fine-grain seiner that thins and extends upstzar of das on all nes scioncemag.org‘ofthese steams for several hlometers. Agent ational impocts extended upsteam te many un dammed tributaries a5 well, as a result of backwater effects, Low stream grins (0.001 0 0.404 inthe Piekmon siden to mexdrste cam heights con trolled mlldar sposing: typically one dm was placed every 2.4 10 $ km akg the Brandywine Conssioga, and Codonis river, Sensca Crock, and Wesker Run. Hise patos ant our fe ‘observations indicate that ely milldams spanned centie valley botims of what ae now fist sovond and thi order seams (ig. S3). Such streams comprise greater than 70% of stream Jongh in the region, ant damning ther wuld have a substantial impact on a large portion ‘of watersheds, including upstream tnbutares Milkian heights were generally 25 to 37 m Dict stop calculations (J8) imply that” Now velocity would have been reduced! by as much 48 60% a Teast ta 3 kr upsrcam of mildams (low reduction is impacted to greater extent by higher cams andor fowcr steam gratiers), ‘Our estimates of trap ellcincies 79) of 6% for Piecmont reservoirs Supyest that Se tion rates woul! have bess igh dasing periods ‘of accelerated sol crasion. Various arly Amer ‘ean mill ets that regulned the rang of dams and compensation for flooded lands, both of Cre ere land “ees 21008081 Elevation (m) 7% Tie ‘which were common occumrences, als incate that milléans ad well-known back ar effets fon valley bettors (see SOM tex). Because of these backwater elles, ater ill acts were passed 10 contol nll crowding (J7, 20) (soe ‘SOM tex), Historical accounts alo Show that tildams altered the original srearn ecology: (i) In 1731, setters tore dewn a milldam an the Conesioga River because it was raining the local fishing industry, and i in 1763, a petition cited complaints euandng the abunkinee of nilkdam on the Conestoga River “as desroying the fomner fishery of shad, salmon, arc rock fish, which were before in abundance, and the Anbutay’ steams had plenty of tout all now one” 2), ‘Stream sediments. Based on geocherical and sta alysis of saiments exposed in steam banks (~100 sites) and samples Hom 445 backhoe trenches and >110 Hydraulic cores In adjacent fill traces, we consruced a come posite stratigraphic profile for a representative Piedinont scam (rst to tind order) (Fig. 3) ‘Sanilac profiles have been deseribed throughout the Piedmon, fom South Carolina wy Pennsy vanin (8, 22-2). The bedrock (weathered) vally floor typxally is overlain by: @) a thin (0.5.2) bed of angular 1 subangular quare- sich rave Gi). a thin (05-10 lem), dak Distance (m) Fig. 3. Steams throughout the mid-Atlantic region (ee abo figs. $1. and $2) have similar charac teristics: vertal to near-verical banks consisting of 1 to 5-m of laminated to masive fine-grained seciment overying a Holocene hyéric soil and a basal gravel overlying bedrock. (A) Westen Run, ‘Maryland. (B) Big Spring Run, Pennsylvania. Scale bers in (A) and (B) ave marked in 05m increments: the banks in) and () are ~2.2 and —1.4 m high, respecivey. (© Conceptval model based on composite stratigraphy frm mule sites, inudingstieat-bank exposes, trenches, and cores wawiscioncemag.org SCIENCE VOL 319 RESEARCH ARTICLES I (bck, 1OVR 2), onsaniesieh sit loan: and ii) thick (1-10 S.m) pale to yelois brown sequence of fine sand, sil, and elay [commonly referred 10 a6 siK loom and clay loam; US. Department of Agsiculture (USDA) Natural Re- sources Conservation Senice casiication). Ex: ‘xptions to this typical sequence occur where ream chamels were diverted along valley walls (eto dian valley botioms for agricul: ture} oF where steam hee were minod for me: ‘ores and gravel We interpre the dak, organi sit oa {210 9% by weight total earbon in the 2.0-mnm faction) above the basal gravel as a buried hy si (vetiand) soil (ables SI and $2). Thrash ‘oat the region, tis hyde sil contains abundant, wwollpresened woody debris, sod, 08, 1915, alzal mats, ana poll, and sometines is a peat ‘with mosses, AC muliple sts, we observe do «ious tree stumps (which ofcn show evden ‘of having been logged) rooted inthis stratum ‘and corduroy (wooden lg or plank) oes upon 1s surice The overlying fine-grain pale brown sediments are lower in organic matter (-1 0 2 by weight ot carbon inthe <2.0-mrn faction), generally horizontally bedded, typically finely [inated and contain rae kre of suboaed gravels ard histris artifacts that include brick fiagments, cut logs and plas, and pieces of ‘oa Seventy radipcarbon daze foe ths bua wweland soil (samples of wond, aves, oF sexx) Pennsylania and’ Maryland yk fiom 11.240 t0 300 years bef ‘presen (able $3), indicating tha those soils sccumubted and sored omganic matter through fot the Holocene epoch. We estima that these valley-bottom wetlands were efficient carbon sinks storing from 2.5 = 10° a 135 > 10" ky ‘of carbon por hoviare that now is bard bensath hier sediment (25) ‘We inet the undsrng grovel os de Fived from lorg-erm (10° 10 107 years) hillstope ‘erosion ans denudation of the kabdscipe, Many may have formed from downslope movement associated with episndie perighcil processes and tundra conditions during the late Pkistocene ‘or calier The prxdominance of quart, a highly resistant minor, and the position of these col- lial gmvets bencath pervasive Holocene Bt tlcil hyde soils suppost this deduction. ‘iming of millon si mentation using Pb geochnonclozy (26) and hisorie documents. Analysis at 140 sies where historic sediment is | and § m thick incats tha secular ecu Pb is reaches at a depts oF 1S 10 20 cm below the surface. This finding indicates tat sedimenis near the surface wore deposited by I8S0 of ealer and tha those cater reservoirs wached sediment storage cpeity by at Jest that year (lig $3). At mul tiple sites in Maryluxl and Penrsylania, we ‘observe groves of lange tees estimated 10 be up te 15D years old on valley fill deposits. This is somsistnt with fossil sed anit pollen eviden showing that a riparian forest bec 18 JANUARY 2008,ESEARCH ARTICLES ‘established circa 1850 in an estuary fill deposit i the upper Chesapeake Bay, after poak sedimentaion from the watershed occurred between 1840 and 1880 (27) Historic ai photos dating to the 1930s and maps dain © the cally 1800 inca thar pow diminish in size rt the mid 10 late 1800s and bocame stable swamps A 10 Beaton (m) 8 0 15 0 Distance from coriuence (km) and meadows thewalles, unl dam breaching and “The characteris of the presctamnent sai- ‘ments and organi material sggest that valley Dotons were trad. forested wetlands (alder shrubscrub) with smal, shallow (02 myc Kecally and at least O08 nyyeat along men stream reaches up stam of 2hh-esntury dam breaches. In Mary land, steamers setons monitored ford years ina sul of rsandr mization and Boop Maryan 2,9) ae dtd 1 kn upstream o breacined aivcentiry dam, which was. not recognized by previous workers. This -centiry ‘enlugement ofthe channel to upstream wiz ton, we propose thet enlmgement was atu response io dam bashing and incision nes scioncemag.orgWhere dans are unbreachad or parially breached, deep inciion has pot yot occured, and sacs are shalhn and wide as they spill ‘over the historic dams. In some cases, mare We cent dams have boon constructed with incisal channels near the site of older breaches many ‘of hse sovond generation dams are abo bach Jn Pomeyhania, these inset dams typically were ‘onsucted in the late 180 wo earky 1X Increasing charnel width ata breached dam site on the Brandyvine River din the pst ~100 years is consist! with 671 repeat mes ten of ir wo exdfentqualty for se Can hankiull dichawe at the Chadds Ford US. ‘Goological Suvey (USGS) sage station, which js located betwee two closely spaced breached nilldams (9° SOM text Our analysis of these data indistes that the ratio of wid o discharge fora given dschame range increased from Ot to OS between 1908 anki 2007. This increase Yields an estimate of atleast a 10-m increase in ‘channel wid, likely esuking fier bank ero- sion, since 1808 and is consistent with) cur mapping of bank erosion along this wach, (i) ir photos fom 1937 to the present tat sbow channel widening, and (i) the marked rise in suspended sediment keads recoded at this gage stain from 1964 to 1978 (USGS deta see SOM tex) Sinilarly, we calcula a power fine- tion exponent of €.26 for the rotten beeween width and discharge at the Chadds Ford gage station, whereas analysis of hyde geom- ‘tty relations in 1955 yielded an exponent of 0.17 (3), A 80% increase inthis exponent dur ing the pas ~$0 years is consistent with chan- nel evolution by laieral erosion and longctemn widening fier ncison in reponse to dam Loch ing (0. ‘Similar to carlierstuiss (2, 3, 8}, eur mea- surements of sediment si fe that modern channel-bed sediments te much coarser gained fecape suns and gravels) than the fine sed ments (mest silt and cay and some sand) that are exposed along sicam banks. We atwibute this cilerence not wo a meandering steam tha deposits gravel in bars and sit and clay on Aondplains during overbank how but rather to inreased flow velovtics and shear siesses that accompany deep incision of steams im fine- ‘grained millpond sediment. After dam breach ing, ines channels have shear stesses >5 10 0 tires as igh as those of presettemsn! sans (shear Stes is proponional 10 water depth and slope) ang, consequently, are capable ft porting Pleiecene gravels that exile! fom the buried toes of hillslopes (Fig. 4). Further, bank ‘erosion removes not nly vast amourts of ‘postsetement alluvium but ako wide areas of Holocene weiland soils and Pleistocene basal gravels that covered stew valleys in this region since the end of the last ee ae “Tasos te imput of bark erosion on sei= ment load of the mndem stars, we collected FC. data on sediments fern several south fem Pennsylvania strams and sircam banks nw scioncemag.org Cesiu-137, produced by atwespiheic muckear ‘bomb testing, reached its peak protacicn in 1963 and ceased in the miid-19703 with the banning of atmospheric born tess. This noa- al isotope read ated surface sedi- ‘mens and is Srongly swe onto the surfaces of fine-rained soil patcles (31, 32). It thus ies a mesic for sediment Jats an surase arson with suspended load ations indicates tht 30 10 81% ddan from gu fof suspended sediment is derived from bank fresion (25) Historic suean-bank sedinen's, ‘whieh we measured at 15 sites, ako have high levels of tol carbon, nitrogen, and phosphorus, ‘aging from 0.310 3.1% (C), 400 t0 2100 pats per ili (S), ste 340 so 958 pom (Ph; thus, ‘bank erosion contrbuting to the high nutrient loads of many of these stcamns (23). These r= subs ace consistent with bank erosion rates and ‘with sediment and nutrient Toads measured in ‘other watersheds in he United Stes and around the world (3436). Condusions. We concn tha vial agere dation and degrodation in the easier United ‘Stes were caused by human-induced base- level changes front the following peooesses (i) ‘widespread milan construction that inundated presetleent valleys axl converted them into a series of linked sbickowater ponds, coupled with forestaion and agricultural proctios that in- sreased sediment supply (i) sedimentation in ‘ubiquitous millpods that gradually convened these ponds © sedimentfiled reservcins (id subsequent dam breaching dat resuled in chan- nel incision though pestsetlemert alluvium and acceler bank erosion by meandering sea and (iy) the formation of an abandoned valley- fla temace and a lower inset oodpkin, This ‘evolution explins why so many easter streams hhave tunkfull(ischarge) heights thar are much lower than actual bank heights, Assesment of bankfull discharge are cnxial to estins of Hood poteotial anc o design criteria fr stam restoration Early workers considaed westom and east cm strcams in the Unked States to have similar ‘origins and forms feompare wih plate 2 in However, westom steams are in a move teton ly active ragien, generally have high steam ‘gracionts, and carry substantial badoad fine- tions, winerecs eastem Piedmena steams are in a tcctcnically quiescent region, have Io seca aon’. an! largely dil ox cary gravel (or even much fine sediment) during the Holocene bore formation of modem incised cts Formation of gril bars inthe wake of eroding, meandring channels. followed by everbank postion of ine seciment, ws not the ata pprcess of fleedplainy formation in the casiem United States before mill damming and subse quent chanos! incision afer dam breaching. Valley ‘botoms along eacer steams were charictrizt by latomlly extensive, wetland-dominated 59s- tems of forested meadows with stile vegosted iNands and mule snall channels during the SCIENCE VOL 319 RESEARCH ARTIC! Holocene, These findings are consent with research on streams in the Pacific Nonhwest, wich revealed that ge woody debris fio forssted riverine anas was crea 1 goo phic processes and habitat before logeing, cia no clearing, an ditching doing the late 18005 to carly 1900s (3739) In particular, logiams blocked channels and led t0 the formation of side charnels and floodplain sloughs, preducing multiple anabranching channels and riverine wwotlnds that are in stark contrast to the large Single chanel tha exs in these scams oa ‘Our resals explain the unasualy thick depos is of fine sediment in stream bunks ratve 10 the limes amount of modern overkank depo- sition Z.3). They abo explain the ack of evees fon mid-Aanic streams (J); the great contest ‘tween fine stcarnbank an couse stcan+bed retrial; the obsercation that casera incisal) scams, in comparison with westem steams, ge litle in wid a discharge increases a & station (3, 10}; and the anomalously high sus pend sediment and nutrient loads measured in Piedmont steams (7. Posisetlement.milldam constuction and nillpond solimentaion were rapid and pervasive, This prowess inuncdued, buried, and sequssered prcsttlement wetlands and altered. regional stan functions probably within 090 generations ‘of sotement, Our fick! an strical rescarch shows that stream valleys were once lined with nillponds, which thn sit into form ba fat Dotto lands. The modem, incised, mcandeting sam isan anit of the ese and fll of mide Atlantic strams in response to busran manip tultion of scam valleys for water power. Ox resis indicate that substantial portion of mod 4m suspended sediment and nutrient fads could ‘be due to dam breaching and ewsion of reservoir sxiiment, which is consistent with Suis of = ‘contd removals osowherv inthe United States (0,4, “These conslsions change the interpretation of hydraulic geometry in eastem US. steams that is hase on the archoxype of an “eal me- andring river form” and imply the need to r= ‘consider cument procedures fo steam reseation tha rly on rience reach conditions and the assumption that etedingcharnel_ hanks are natural and rplenishable. The cument condition ‘of single raveLbalded channels with high, fne- gmined banks and reatively dy yally- ft sur aes dionnectad from groundwater iin stk contrast tothe presetlement contin of snaenpy nl shallow anu tons for interpectation ‘of alluvial scimertation and steam chara form ara evolution in Ewope as well asin United States Tens of thousands of mils existed throughout Europe by the 18th century, and tis possible that Europe streams have been sponding to antheopozenic bas-level rise and SI for up to a millennium (SOE wet and fi, 5). We propose that widespread mid- to lte 18 JANUARY 2008, sll 303