# News this Week

Science  18 Jan 2008:
Vol. 319, Issue 5861, pp. 266
1. SCIENTIFIC PUBLISHING

# Uncle Sam's Biomedical Archive Wants Your Papers

1. Jocelyn Kaiser

If you have a grant from the U.S. National Institutes of Health (NIH), you will soon be required to take some steps to make the results public. Last week, NIH informed its grantees that, to comply with a new law, they must begin sending copies of their accepted, peer-reviewed manuscripts to NIH for posting in a free online archive. Failure to do so could delay a grant or jeopardize current research funding, NIH warns.

The public-access law, touted as a way to enable taxpayers to learn about the research they fund, was part of a spending bill Congress passed in mid- December. It makes mandatory a policy in effect since May 2005 that requests that NIH-funded investigators submit manuscripts accepted by a journal to NIH; the agency posts the full text in the archive, called PubMed Central (PMC), no more than 12 months after the article is published. Only about 12% of authors are complying with the voluntary policy, however, and of 80,000 eligible articles per year, only 20% to 25% are being submitted either by authors or directly by journals, says David Lipman, who oversees PMC.

NIH says it is ready for the glut of manuscripts it will soon receive, but there are signs that some scientists may be confused about what to submit. For example, NIH is already removing old papers that authors mistakenly posted in PMC. Lipman acknowledges that there will be “a learning process” but notes that traffic on the site is already “huge,” with 12 million article views each month.

NIH's brief notice simply states that the policy is mandatory for all articles accepted on or after 7 April. Initially, NIH requested that only original research be archived at PMC, but now the agency says the policy has been expanded to include review articles if they were peer-reviewed. Many journals retain copyright of the manuscripts they publish, so authors must obtain permission to post a copy on the NIH site. It is up to investigators and their institutions to figure out whether their submissions comply with the journals' policy.

To give scientists a nudge, NIH will require them to include the PMC number when they cite their own papers in grant applications and progress reports. Other possible ways of forcing scofflaws to comply range from having a program director call with a reminder to “the most extreme: suspending funds,” says NIH Deputy Director for Extramural Research Norka Ruiz Bravo. “We hope we're not going to get there,” she says.

The new law puts NIH in line with some other funding agencies that require grantees to send their papers to PMC or a U.K. version of the archive; these include the U.K.'s Medical Research Council and Wellcome Trust, which adopted such policies in 2006, and the Howard Hughes Medical Institute (HHMI) in Bethesda, Maryland, whose rule goes into effect this month (see table). All three institutions require that papers be posted within 6 months of publication in a journal. NIH differs in one way: Whereas other funders help pay author fees that some journals charge to make the full text immediately available, NIH is not offering any extra money for “open access,” Ruiz Bravo says.

Critics have long warned of potential snags. For example, they note that it will often result in two versions of an article on PMC: the accepted manuscript, which hasn't been copyedited, and the published paper. Publishers also have worried that making articles available for free will cut into the subscription income needed to run journals and other society activities. The Association of American Publishers had no immediate comment on the NIH plan, but AAP has warned that a mandatory policy “undermines” publishers' copyright and is “inconsistent” with U.S. laws (Science, 11 January, p. 145).

Scientists who have been sending their papers to PMC say the process is relatively easy, but keeping track of each journal's copyright policy i s not. For example, Karen Allendoerfer, lab manager for HHMI investigator Susan Lindquist of the Whitehead Institute in Cambridge, Massachusetts, says figuring out whether the author must submit articles directly or can rely on the journal to do it is “a pain in the neck.” (NIH has tried to help by posting a list of some 300 journals that automatically upload articles to PMC.)

As for journals, although most major biomedical publications (including Science) already allow authors to submit manuscripts to PMC, some publishers say they will need to police the site for articles mistakenly posted, such as those not yet released from the journal's embargo or those published before 2005. Martin Frank, executive director of the American Physiological Society, says APS asked NIH to remove 78 papers last year, and he expects “hundreds” of similar errors when the mandatory policy kicks in. Lipman acknowledges that NIH had to remove some papers. But complying with copyright, he says, is not NIH's responsibility; it's “between the author and the publisher.”

2. REMOTE SENSING

# Satellite Company Offers Earth-Observing Researchers a Ride

1. Daniel Clery

Looking for a way to get multiple daily observations of Earth from space? A satellite communications company called Iridium may have a deal for you.

Iridium is offering space on its next-generation fleet of 66 satellites for researchers to do earth observation. Built and paid for by government agencies, the instruments would be launched by Iridium, which would manage the data transmissions for a fee. Next week, it will pitch the idea to scientists, funding agencies, and instrument manufacturers at a meeting at the Royal Society in London. But the $1.6 billion price tag could prove a bit too steep. “The concept can't be argued with,” says Mick Johnson of the European aerospace company EADS Astrium, who also heads the U.K.'s Centre for Earth Observation Instrumentation. “The question is: What's the performance, what's the cost, and what will we then not be doing?” Motorola set up the Iridium system during the 1990s as a worldwide mobile phone service. But it proved a financial disaster and was bought by a group of private investors in 2000. Iridium, which can handle both voice and data, now has about 225,000 customers, mostly in the military and in the shipping, aviation, and petroleum industries. Remote research outposts such as the Amundsen-Scott South Pole Station rely heavily on it. The satellites, however, provide limited bandwidth—at 10 kilobits/ second, it's slower than a dial-up modem—and are expected to begin failing during the next decade. So in late 2006, the company began planning a replacement constellation, dubbed Iridium NEXT, and came up with the idea of adding remote-sensing capabilities as a moneymaking venture. It's working with Trident Sensors, a British company that makes Iridium receivers for scientific uses such as ocean-monitoring buoys and autonomous submersibles. Most earth observation these days is done by single satellites, often huge behemoths bristling with sensors. Although this approach allows researchers to compare different types of data, it is less useful for studying phenomena that evolve rapidly, because a single craft can take days to return to a particular spot. Iridium's idea is entirely different: The satellites' main function will remain communications, but the company will make available on each one enough room for a suitcase-sized instrument, weighing no more than 50 kilograms and using less than 50 watts of power. Government agencies would choose what instruments to install, build them, and control them in orbit; Iridium would charge to download the data. Because there are 66 satellites with many carrying the same instrument, any location can be viewed many times a day, capturing fast-changing phenomena such as plant growth or how air pollution levels vary with the time of day. “Science needs this. It's unparalleled to have so many platforms,” says atmospheric chemist Paul Monks of the University of Leicester, U.K. Iridium and Trident have worked out one possible scenario, in which different groups of the 66 satellites would carry out four separate missions (see table): altimetry, radiometry, imaging, and radio occultation, which gives temperature and moisture profiles (Science, 7 April 2006, p. 48). But the space constraints onboard the satellites mean that not all types of earth observation would work. “Only some instruments would be feasible,” says Volker Liebig, head of earth observation at the European Space Agency, adding that similar instruments are two to four times that mass. The bigger obstacles are timetable and cost. Executive Vice President Don Thoma says Iridium hopes to sign a contract to build the satellites within a year and so needs to know by then whether to make space for instruments. The satellites are scheduled to be launched between 2013 and 2016 and will operate until at least 2030. If government agencies take up the offer and provide instruments, they would have to pay Iridium a “hosting fee” in the form of a prebuy of data services. Bill Simpson of Trident says the$1.6 billion for instrument construction and hosting is a fraction of what it would cost funders to build and launch their own constellation.

But that doesn't mean funders have budget available for such a project in the next 12 months. The British National Space Centre, which coordinates U.K. spending, calls the Iridium proposal “an interesting opportunity.” ESA's Liebig says the total is “an enormously high sum” and says the agency has “no activity in this direction.” The U.S. National Oceanic and Atmospheric Administration is having a meeting this month to discuss doing earth observation with commercial satellite systems.

Researchers are hoping that some common ground will be found at next week's London meeting. “It's very easy to say no,” says Monks. “But you can't let every opportunity go by just because it's difficult.”

3. NEUROSCIENCE

# Mirror Neurons May Help Songbirds Stay in Tune

1. Greg Miller

The discovery of so-called mirror neurons a decade ago has fueled much discussion in neuroscience ever since. The original experiments identified neurons in monkeys that fire when an animal reaches out an arm or sees another animal make a similar movement. Brain-imaging studies have found analogous neural activity in the human brain in a variety of other contexts, prompting speculation that mirror neurons could have roles in perception, learning, and empathy (Science, 13 May 2005, p. 945).

In this week's issue of Nature, researchers describe mirror neurons in songbirds that fire when a bird sings or hears another bird sing a song similar to its own. The finding may pave the way to insights into how songbirds learn and maintain their complex songs—one of the few instances of learned communication aside from human language. Some researchers see broader implications as well.

“This is the first description of a mirror system in a species other than primates,” says Pier Ferrari, a neuroscientist at the University of Parma in Italy, a current member of the team that made the original mirror neuron discovery in monkeys. Ferrari thinks mirror neurons may turn out to be “a basic feature of the vertebrate brain” that initially evolved to help animals refine their movements and eventually became co-opted, at least in some species, for more complex functions, such as understanding the behavior of others.

In the new study, neuroscientists Jonathan Prather, Richard Mooney, and colleagues at Duke University mounted an ultralightweight device on the heads of swamp sparrows that enabled them to record the activity of individual neurons as the birds sang and listened to samples of recorded songs. In a brain region called HVC, part of a forebrain circuit that mediates singing, the researchers identified neurons that fired in a consistent pattern when the bird sang a particular song. (Swamp sparrows typically have several songs in their repertoires.) During singing, for example, a neuron might fire at the onset of a particular sequence of notes. When the researchers played the bird's own song through a speaker, the same neuron would again fire whenever that sequence appeared. Sometimes the same neuron would also fire in response to another swamp sparrow's song—but only when the other bird's song contained a similar sequence of notes. Mooney's team also found HVC neurons with similar responses in Bengalese finches, another songbird.

“These were damned difficult experiments,” says Daniel Margoliash, a birdsong researcher at the University of Chicago in Illinois. Margoliash says the findings are exciting because the HVC neurons identified by the Duke team may help explain how individual songbirds are able to maintain their complex and distinctive songs for life. The activity of these neurons during singing may be a neural representation of what the bird intended to sing that it compares against auditory feedback of what it actually sang, Margoliash says. Such a comparison is probably necessary for adult birds to maintain their songs, he says, and it almost certainly plays a role as young birds first learn their songs as well.

The HVC neurons may also help songbirds decode each other's songs, Mooney says. Male swamp sparrows sing back and forth to defend their territories, and hearing the song of a rival would activate some of the HVC neurons that fire when a bird sings a similar song itself, Mooney says. That might enable a bird to compare his rival's song to his own repertoire—and select an appropriate retort.

4. ASTRONOMY

# New Dark-Matter Map Reveals Where Galaxies Gambol

1. Govert Schilling*
1. Govert Schilling is an astronomy writer in Amersfoort, the Netherlands.

AUSTIN, TEXAS—The most detailed map of dark matter ever made confirms that galaxies cluster together where the density of the mysterious dark stuff is highest. The new survey, which spanned a patch of space 16 million light-years across and required 5 days of observing time on the Hubble Space Telescope, also shows that galaxies affect one another most strongly on the outskirts of these high-density regions. “It's as if people move in from rural areas to city centers, and most interactions occur in the suburbs,” says astronomer Meghan Gray of the University of Nottingham in the United Kingdom, a member of the team that made the map.

Dark matter can't be seen, and no one knows what it is. But the gravity of a dark-matter clump slightly bends the light from faint galaxies in the background in a process known as weak lensing. Thus, a statistical analysis of the shapes of tens of thousands of remote background galaxies reveals the distribution of dark matter in the foreground. “It's a very, very weak effect,” says team member Catherine Heymans of the University of British Columbia in Vancouver, Canada.

The new survey, presented here last week at the 211th meeting of the American Astronomical Society, was aimed at the relatively nearby supercluster Abell 901/902, 2.6 billion light-years from Earth. As predicted by current theories on cosmic evolution, dark-matter condensations within the supercluster appear to be the gravitational pits that galaxies fall into. Four particularly dense clumps (shown as pink blobs on the map) coincide with strong concentrations of galaxies. The map also confirms earlier indications that most galaxy interactions (such as tidal deformations and mergers) occur in regions with a moderate dark-matter density, on the outskirts of the densest blobs. Gray says galaxies in the dense cores may be shooting past one another too quickly for much interaction to take place there.

Physicist J. Anthony Tyson of the University of California, Davis, warns that many more clusters need to be studied before researchers can draw firm conclusions about how galaxies evolve and clusters grow. “I think we are witnessing only the beginning of an exciting exploration of the history of assembly of structures in the universe,” says Tyson, who is also director of the planned Large Synoptic Survey Telescope (LSST) (Science, 27 August 2004, p. 1232). “LSST will survey an unprecedented volume of the universe, charting billions of galaxies back to when the universe was a quarter of its current age.” Tyson expects the telescope, to be built in northern Chile, to be ready in 2014.

5. EVOLUTIONARY GENETICS

# Polynesians Took the Express Train Through Melanesia to the Pacific

1. Constance Holden

The Polynesians who settled the far-flung islands of Remote Oceania several thousand years ago accomplished one of humanity's most rapid feats of colonization. But who were these early seafarers? Researchers have long debated various origins for them, from Taiwan to the islands of Melanesia.

Now, modern genetics has offered the most definitive answer yet. A study led by anthropologist Jonathan S. Friedlaender of Temple University in Philadelphia, Pennsylvania, indicates that Polynesians bear a much closer relationship to aboriginal inhabitants of Taiwan than to the Melanesian groups who occupied New Guinea and surrounding islands from 50,000 to 30,000 years ago.

The new study “makes excellent sense,” says archaeologist Peter Bellwood of Australian National University in Canberra. It supports the “express train” theory, which posits that people from Taiwan moved rapidly through Melanesia, leaving little genetic footprint. Other ideas have included the “slow boat” hypothesis, in which the migrating Polynesians mixed with Melanesians on their way east. Then there is what has been called the “entangled bank” scenario, which holds that patterns of ancient migrations are too enmeshed to be accurately reconstructed (Science, 2 March 2001, p. 1735).

Until now, genetic studies have yielded conflicting answers. Data from mitochondrial DNA (solely from women) told of little interbreeding between Polynesians and Melanesians, but a Y chromosome study suggested extensive blending between the two.

The new study, reported online on 17 January in PLoS Genetics, is based on 890 genomic markers from the nuclear DNA of 952 people in 41 populations. The conclusion? “There was remarkably little genetic intermixture,” says Friedlaender. Melanesians speaking Papuan languages—their own ancestral tongues—showed no sign of genetic contributions from Polynesians. Even among Melanesians sharing languages with Polynesians, there was minimal genetic mixing, never more than 20%. “When different peoples interact, … genes tend to be exchanged far less frequently than ideas and languages,” concludes Friedlaender.

Geneticist and skeptic Martin Richards of the University of Leeds, U.K., says, however, that the researchers could not distinguish whether the Polynesians originated in Indonesia or Taiwan; he argues that the paper “does not really move the debate forward very much.” But archaeologist Patrick Kirch of the University of California, Berkeley, calls the work “truly a major accomplishment.”

6. ASTRONOMY

# Most-Massive Black Hole Confirms Relativity Rules the Universe

1. Govert Schilling*
1. Govert Schilling is an astronomy writer in Amersfoort, the Netherlands.

AUSTIN, TEXAS—You can run from Einstein, but you can't hide. Astronomers observing a supermassive black hole 3.5 billion light-years from Earth have found that the object brightened dramatically at just the time the equations of general relativity predicted. As a bonus, scientists pegged the black hole's mass at a staggering 18 billion times the mass of the sun, six times as big as the previous record holder.

“This is a pretty spectacular extension of the regime in which the effects of relativity have been measured,” says theoretical physicist Robbert Dijkgraaf of the University of Amsterdam, the Netherlands, who was not involved in the study. It's also bad news for proposed alternative theories of gravity that give slightly different predictions, Dijkgraaf says.

The supermassive black hole, known as OJ287, resides in the core of a distant galaxy, where its gravity sucks in matter from an enormous disk of gas and dust rotating around it. As the matter plunges toward the black hole, the galaxy's core turns into a glowing energy furnace. Twenty years ago, astronomers sifting through photographs dating back to 1891 discovered that every 12 years or so, the radiation from this accretion disk flares up to about 100,000 times normal, in two bursts just over a year apart (see figure).

To astronomers, such periodic behavior smacked of orbital motion. Mauri J. Valtonen of the University of Turku in Finland proposed that a smaller companion goes around the supermassive black hole, following a very elongated orbit strongly tilted with respect to the primary's accretion disk. Every 12 years, at its closest approach, the smaller black hole passes through the accretion disk twice, stirring it up and heating nearby matter enough to produce a brightness peak. After new flares appeared in 1994, 1995, and November 2005, Valtonen and colleagues modeled the binary's orbit and predicted that the next peak would occur on 13 September 2007.

About 30 professional and amateur observatories, including the German 3.5-meter telescope at Calar Alto Observatory in Spain and the 2.5-meter Nordic Optical Telescope on the Canary Island of La Palma, joined forces to check the forecast. At the 211th meeting of the American Astronomical Society held here last week, Valtonen proudly presented the results: The flare showed up as predicted, beautifully confirming the binary model. From the orbital motion, the researchers calculated the masses of the two black holes to be 18 billion solar masses for the primary and a more modest 100 million suns for the companion.

The observations also marked yet another triumph for Einstein's theory of general relativity, which states that the extremely strong gravitational field near the supermassive black hole should distort spacetime in ways that rapidly change the orientation of the companion's orbit. It also predicts that the loss of energy from the emission of gravitational waves should cause the orbit to shrink—an effect that won a 1993 Nobel Prize for two physicists who first detected it in a binary neutron star. Without spacetime curvature, the flare would have occurred 10 days earlier than predicted, Valtonen says; without gravitational waves, 20 days later.

“The result is of fundamental importance, since precision measurements that really test general relativity only became available quite recently,” Dijkgraaf says. On a cosmic time scale, this particular test case will be gone soon as well: The two black holes are expected to collide and merge within some 10,000 years. When that happens, astronomers will really have something to stare at.

7. INFECTIOUS DISEASES

# Isolated Tribe Gives Clues to the Origins of Syphilis

1. Carl Zimmer*
1. Carl Zimmer is the author of Microcosm: E. coli and the New Science of Life, to be published in May.

In 1494, King Charles VIII of France invaded Italy. Within months, his army collapsed and fled. It was routed not by the Italian army but by a microbe. A mysterious new disease spread through sex killed many of Charles's soldiers and left survivors weak and disfigured. French soldiers spread the disease across much of Europe, and then it moved into Africa and Asia. Many called it the French disease. The French called it the Italian disease. Arabs called it the Christian disease. Today, it is called syphilis.

The sensational debut of syphilis inspired centuries of debate. Some have argued that Columbus's crew brought the disease from the New World to Europe; others say the disease existed unrecognized for centuries in the Old World before turning virulent. This week in the journal PLoS Neglected Tropical Diseases, a team of researchers argues that neither hypothesis is correct. Syphilis originated as a milder, nonsexual disease in the New World, they say, and it evolved into its current form after Europeans arrived. Among the evidence they offer is a mysterious disease restricted to an isolated tribe in a South American jungle. Its DNA, they argue, reveals that it is a kind of protosyphilis.

The disease was discovered in 1999 by a team of Canadian doctors who travel each summer into remote jungles in Guyana to provide medical care to the Akwio tribe. Michael Silverman, a clinician at the University of Toronto, noticed that some of the Akwio children had open sores on their arms and legs.

To Silverman, it seemed like a combination of two diseases, syphilis and yaws. The former, caused by bacteria known as Treponema pallidum pallidum, produces open sores. But it is spread through sexual contact and forms sores around the genitals. Another strain, T. p. pertenue, causes yaws, which is spread by skin contact rather than sex and produces sores on the limbs. But the Guyana disease was not quite like yaws either. That disease causes raspberrylike eruptions, not the open sores the doctors saw. “I thought, ‘This is bizarre,’” says Silverman.

The following year, Silverman and his colleagues proved that the disease was caused by a form of Treponema, which infected about 5% of the children in the tribe. They began treating them with penicillin. Just before boarding a flight for his 2005 mission, Silverman got a call from Kristin Harper, a graduate student at Emory University in Atlanta, Georgia. She was sequencing DNA from every known strain of Treponema and asked him to get her some from Guyana.

Harper thought genetic analysis could shed new light on the origins of syphilis. Until recently, the debate revolved around bones, which can be scarred by syphilis. Some researchers claimed to have found lesioned bones of Europeans who lived centuries before Columbus's voyages; others dismissed those findings as inconclusive.

Harper and her colleagues hoped to reconstruct the evolution of different strains of Treponema by comparing their DNA. But gathering the DNA was difficult. Although syphilis is common, the other strains are rare, and none can be cultured easily outside its human host. To get the DNA samples, Silverman dipped swabs into sores, preserved them in alcohol, and sent them to Harper, who extracted fragments of Treponema DNA. Months later, Silverman got a call from Harper. “She said, ‘You found the origin of syphilis!’” he recalls.

When Harper and colleagues used the variations in the Treponema DNA to draw an evolutionary tree, the deepest branches belonged to samples of yaws from the Old World. Newer forms of yawslike diseases emerged from the ancestral bacteria. The strange Treponema strain in Guyana sat on a relatively young branch, suggesting that yaws had been carried into the New World by the first immigrants some 12,000 years ago, and that the Guyana strain evolved there. It shares a close ancestry with all strains of syphilis. The relationship also suggests that the Guyana strain is a transitional form that had already acquired some of syphilis's traits, such as the open sores.

In this form, the bacteria could move easily through skin contact, because its hosts wore little clothing. But once the bacteria infected the more heavily clad European explorers, the researchers speculate, it may have had to find a new route of transmission. “You bring in the Europeans who only touch skin when they have sex, and it takes off as a venereal disease,” Silverman says. For reasons scientists don't yet understand, the pathogen also evolved into a far more dangerous form that could trigger an epidemic in Europe.

John Logsdon, an evolutionary biologist at the University of Iowa in Iowa City, praises the research but doesn't think it offers definitive proof of where syphilis came from. “There's very little data to distinguish between the hypotheses,” he says. The new evolutionary tree is based on variations at only four sites in the Treponema DNA, which Harper agrees is a small number. One reason the data are so sparse is the challenging work conditions in the jungles of Guyana, where Silverman preserved samples in unrefrigerated alcohol—far from the ideal way to keep DNA from degrading.

Logsdon and Harper agree that the best way to test the new hypothesis would be to draw another evolutionary tree based on the entire genomes of the strains. But that may not be possible, as Silverman and his colleagues have not been able to get any more DNA from the Guyanese strain. On subsequent missions, they failed to find anyone infected with it. It appears that they have eradicated the disease. “We're still looking for one more case,” says Silverman.

8. EPIDEMIOLOGY

# Calculating Iraq's Death Toll: WHO Study Backs Lower Estimate

1. John Bohannon

A team led by the World Health Organization (WHO) has produced a new estimate of the number of Iraqis who died violently in the first 40 months following the U.S.-led invasion: between 104,000 and 223,000. This figure, published online last week by the New England Journal of Medicine, hews close to some other attempts to quantify the toll but comes in far below a controversial 2006 study led by researchers at Johns Hopkins University in Baltimore, Maryland. That group estimated approximately 600,000 violent deaths during the same period. The discrepancy has prompted critics to renew their charge that the Johns Hopkins results are not credible.

Data from a war zone are never fully reliable; the best researchers can hope for is “getting the numbers roughly right,” says Fritz Scheuren, a statistician at the University of Chicago in Illinois and past president of the American Statistical Association. Escalating violence in Iraq after 2003 put a limit on quality control, but researchers do have a quantitative starting point: the casualty tally made by Iraq Body Count, a nonprofit advocacy group based in London. By controlling for multiple accounts of the same car bombs and shootings, the group estimates from media reports that between 81,000 and 88,000 violent deaths have occurred in Iraq since the invasion. The figure is useful as “a lower bound on the true number,” says Jon Pedersen, a statistician at the Fafo Institute for Applied International Studies in Oslo, Norway.

To get the upper bound, says Pedersen, you have to knock on doors in what is known as a two-stage cluster survey. That's the method used by the WHO and Johns Hopkins teams, among others. Researchers divide the country into regions and then sample clusters of households within each. Finally, they extrapolate mortality rates from those clusters to the total population.

Epidemiologists Les Roberts and Gilbert Burnham of Johns Hopkins published the first Iraq cluster study in November 2004 in The Lancet. They used data collected by Roberts and an Iraqi team, which, in September 2004, surveyed 988 households in 33 clusters across the country. They arrived at a figure of 98,000 “extra” deaths since the invasion, about half due to violence. Soon after this, a team led by Pedersen and the United Nations Development Programme, which had used a much larger sample of 21,668 households in 2200 clusters, produced an estimate for roughly the same period of about 25,000 violent deaths.

As the invasion gave way to occupation and insurgency, Roberts and Burnham mounted another study. This time they left the surveying entirely to the Iraqi team, communicating from abroad. Published in October 2006 in The Lancet, the second survey—based on 1849 households in 47 clusters—estimated that 601,000 Iraqis died violent deaths between the 2003 invasion and July 2006. To many, the number seemed unrealistically high. Some also faulted the authors for not fully answering questions about the survey's methods (Science, 20 October 2006, p. 396).

Now comes the WHO survey. Conducted with the help of the Iraqi government, it is by 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, too. Unlike other Iraq casualty surveys, this one includes an upward adjustment of 35% to account for “underreporting” of deaths due to migration, memory lapse, and dishonesty. “That is really an arbitrary fudge factor,” says Debarati Guha-Sapir, an epidemiologist at the WHO Collaborating Centre for Research on the Epidemiology of Disasters in Brussels, Belgium. But the number falls squarely within the range produced by a meta-analysis of al l avai lable mor tality studies by Guha-Sapir and fellow centre epidemiologist Olivier Degomme. The Johns Hopkins figure is an outlier, she says.

Why the Hopkins study came up with such a high figure is not clear. Criticism of the study has in fact intensified since Burnham and Roberts released a data set to selected peers last year. “It did not include the standard kinds of data,” says Seppo Laaksonen, a statistician at the University of Helsinki in Finland and a specialist in survey methodology. For example, he says, it was impossible “to check the objectivity and randomness of cluster selection.” Scheuren, who also received the data, wanted to compare results obtained by different interviewers to “get a handle on noise” and check for fabrication by surveyors. Roberts declined to provide all the details, according to Scheuren, saying that he was concerned that this would risk the safety of the interviewers.

Burnham told Science, however, that the Johns Hopkins team does not have such detailed information. “Our goal was to reduce any type of risk to the community and the participants,” says Burnham. “While we have much of the raw data, we requested that anything designating the interviewers or the location of the neighborhoods visited not be sent to us.” Laaksonen responds that he would not have published “any figures for the country” if he didn't have direct access to such raw information from surveyors.

Burnham is not retreating. Because the WHO survey was conducted by Iraqi government personnel, “people may have been hesitant to answer honestly,” he says. He claims that unlike those in the WHO study, nearly all of the deaths tallied by the 2006 Lancet study were verified with death certificates. Even if the debate may be drawing to a close about whether the number of violent deaths in postinvasion Iraq could be as high as 600,000, the argument about methods is clearly far from settled.

9. SCIENCE AND COMMERCE

# Gene Tests for Psychiatric Risk Polarize Researchers

1. Jennifer Couzin

A small California company is the first to venture into psychiatric gene testing. But is the science ready?

A small California company is the first to venture into psychiatric gene testing. But is the science ready?

SAN DIEGO, CALIFORNIA—Starting a biotechnology company was not part of John Kelsoe's life plan. A respected psychiatric geneticist here at the University of California, San Diego (UCSD), he has spent 20 years patiently searching for the genes behind bipolar disorder, tracing patterns of inheritance in families from Amish Pennsylvania to Iceland. Kelsoe has now laid this solid reputation on the line: He has founded a company that last year quietly began selling the first gene test to help diagnose people with bipolar disorder, which affects about 1% of the population.

Kelsoe knows well that he is opening himself up to harsh criticism. And it has begun: Francis Collins, director of the National Human Genome Research Institute (NHGRI) in Bethesda, Maryland, was taken aback to learn that a gene test for bipolar disorder was on the market. Being diagnosed with a mental disorder carries “great significance,” says Collins, who, like many in the field, thinks a test based on current knowledge is likely to be misleading. In Collins's list of accepted genes for common diseases, “there are no entries” yet for bipolar disorder, he says.

Kelsoe, tall and trim, with the slight stoop of an academic who has spent years in the laboratory, agrees that the genetics of bipolar are still cloudy. But he has grown frustrated with the slow diagnosis and spotty treatment of the disease. On average, 7 years elapse from the onset of symptoms to diagnosis, and patients are misdiagnosed three times. “There is a huge need” for information, Kelsoe says. “As opposed to playing it safe and conservative scientifically, I think we also have to think about our obligation to provide something” for physicians and for patients.

His company, Psynomics, is the first to market a test in psychiatric genetics, but it won't be unique for long. Another company, in Louisville, Kentucky, plans next year to start offering a similar test for schizophrenia. A third, in Boulder, Colorado, will launch a genetic screen to help identify those at risk of suicidal thinking from antidepressants.

Psychiatric genetic testing is only a slice of the fast-expanding genetic testing market, but it arouses fierce emotions. The new gene tests are very different from traditional ones for rare and often fatal diseases like cystic fibrosis, Tay-Sachs, and sickle cell anemia. With these, the cause is not in question: An individual 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 contribute to its onset, along with largely unknown environmental triggers. This makes it similar to heart disease, adult cancer, and type 2 diabetes. Each gene variant may raise risk only slightly—say, from 1% to 2%—meaning that the vast majority of people with these variants never fall ill from that disease. Furthermore, in bipolar disorder, suspect genes are still being uncovered. There's much debate about which findings will endure and which will be picked up once or twice and never seen again.

Because of this uncertainty, not to mention concern about how test results will be interpreted, researchers and ethicists are asking difficult questions. Why test for genes whose science is shaky or that boost risk only slightly? Who will buy these tests? And how will consumers understand and cope with the genetic information they're given?

## The genes

The men running Psynomics, including Kelsoe, its chief scientific officer, see themselves guiding this nascent field and carving out an early niche. They also believe, in the words of UCSD business school assistant dean and Psynomics CEO Kurt May, that “we … don't have to win the Nobel Prize on this one, and we don't have to solve the problem for all time. All we have to do is present a business model that is one better than psychiatry today.”

At the heart of that business model is GRK3, a gene on chromosome 22 that Kelsoe identified in the late 1990s. In 2003, he reported in Molecular Psychiatry that a change in an individual DNA base in GRK3, called a single-nucleotide polymorphism (SNP), was associated with about a three times increase in the risk of bipolar disorder in 428 families studied. But the variant was rare, found in only 3% of those with bipolar and 1% of the unaffected.

Since then, Kelsoe has hit on a combination of four SNPs in GRK3 that together seem to be associated with a doubling of disease risk and are more common. They were found in 15% of bipolar patients (and 5% of those without the disorder) in 181 Caucasian families. That work was published last month. UCSD has filed for global patent protection on both sets of variants and licensed them to Psynomics; Kelsoe will share in licensing fees and any profit. The Psynomics test looks for both GRK3 variants. It also checks for a publicly available mutation in the serotonin transporter gene, which is believed to influence a person's response to certain antidepressants, often a component of treatment for bipolar disorder.

Does GRK3 play a compelling role in bipolar disorder? It may or it may not. “It's always possible” that it's involved, but “to me it seems unlikely,” says David Collier, a neuropsychiatric geneticist at the Institute of Psychiatry in London. In December 2006, he published a paper arguing that bipolar 1 disorder, the more severe of two known forms of bipolar disorder, was not associated with GRK3 in a sample of 410 patients and controls from Scotland. But Collier adds that because of its relatively small size, his own study “can't be regarded as definitive.”

Collier is keeping an eye on genome-wide association studies, which are revolutionizing the genetics of complex disorders such as type 2 diabetes. One of the first whole-genome studies that included bipolar disorder, published last year by the U.K.'s Wellcome Trust Case Control Consortium, did not find GRK3; it also failed to find many other genes previously linked to the disease, says consortium member Collier. Although such scans are not designed to pick up every disease gene, some proponents of the whole-genome approach say their confidence in genes identified the old-fashioned way, in large sets of affected families, is declining. Psynomics “is using data from the other side of the bridge, the bad side,” says NHGRI's Collins, referring to family linkage studies that have been hard to replicate.

Pablo Gejman, a psychiatric geneticist with Evanston Northwestern Healthcare and Northwestern University in Evanston, Illinois, is leading a project that he hopes can confirm or reject initial schizophrenia gene discoveries: a genome-wide association study of 4500 people with this disease and 4500 without. Results are due out later this year. (Kelsoe is leading a parallel study in bipolar disorder.) Gejman wonders if even larger trials—of 15,000 people, perhaps—will be necessary. This week, he published a paper in the American Journal of Psychiatry, reporting that none of 14 previously described schizophrenia genes played a big role in disease risk in nearly 3900 people. “My message,” says the native Argentinean, “would be that we need to be patient, [conduct] systematic studies, collaborate, … and try not to jam too rapidly the gun” by selling genetic tests before the data are solid.

## The doctors

Kelsoe acknowledges that GRK3 may turn out to play a minor role in bipolar disorder, or even no role at all, although he considers that unlikely. No independent group has published a second GRK3-bipolar disorder association, but Kelsoe says he knows of two groups, in Israel and in Canada, that have found a link. Meanwhile, in the next 3 months Psynomics 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 to treat bipolar disorder; one, CACNG2, is considered indicative of bipolar risk and lithium response; and another, SP4, has been linked to both bipolar disorder and schizophrenia.

In determining which genes are ready for commercialization, “we've cut kind of a low threshold, and that is replication in at least one independent study,” says Kelsoe—either by a different group or his own group examining new samples. “Docs make decisions every day on a lot less information” than this.

Gejman and Collins may shake their heads, but psychiatric gene tests are coming. Mark Brennan, founder of SureGene, a schizophrenia-based test company in Louisville, says such tests are “going to happen with or without” his own efforts. Brennan is SureGene's chief scientific officer and a human geneticist at the University of Louisville in Kentucky. His company is designing a risk predictor and diagnostic test around six schizophrenia genes that Brennan discovered. It's good for scientists to get into marketing, he argues, because “we have to step up and make sure it's done in a way that really helps people.”

Brennan has so far published information 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 them all. But one who agreed to be identified is Herbert Meltzer, a longtime schizophrenia specialist at Vanderbilt University in Nashville, Tennessee. Meltzer, who is on SureGene's advisory board and says he has received company stock, is enthusiastic. “If we proceed and it doesn't look as promising to me as it looks right now, then I'll do my best to keep it off” the market, he promises.

Before launching the SureGene test—slated to happen about 18 months from now—Brennan wants the genes replicated “in at least several hundred samples from different centers.” That's 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 trees, are many bipolar patients who have been misdiagnosed and erroneously put on antidepressants alone, which can exacerbate their disease.

“Doctors are operating in constant uncertainty,” he says, struggling to find the right diagnosis, the proper recipe of medications. “Not to get snooty about it—it's fine to sit in your ivory tower and say, ‘We don't want to offer this until it's perfect.'Meanwhile, that doctor's down there trying to work with whatever he's got.”

View this table:

Psynomics 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 is going to initially be driven by the patients,” says Kelsoe, who believes that doctors are wary of a psychiatric gene test. Psynomics allows anyone to order its test by paying up to $750 and providing a physician's name. Soon after, a small blue and white plastic container arrives in the mail. The capsule preserves a saliva sample, which the consumer ships to a certified UCSD lab for genetic testing, with DNA results going back to the company. Results also go to the designated physician but not to the consumer. The process is meant to control who takes the test: Kelsoe has rejected risk testing for individuals who have no symptoms but are curious. For a healthy person, he says, receiving a positive test result could do more harm than good, especially because data from asymptomatic people are lacking. Even so, Kelsoe admits that there are ways to get around Psynomics's procedures. A bipolar patient, for example, could send in her child's saliva sample and claim it's her own. In the end, predicts May, the market will demand risk testing, and “the market is going to win.” Business has already won out in another facet of the 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 2 years, he came to see that this was “no longer necessary or practical.” Interpreting and counseling on test results will be up to doctors, who receive a six-page summary noting that “a positive GRK3 test … makes a diagnosis of Bipolar Disorder 2–3 times more likely” for individuals with symptoms and a family history. This is not much help, some say. “I'm in the bipolar disorder genetics field, and I wouldn't know how to interpret the results of a bipolar genetic test,” says James Potash, director of mood disorders research at Johns Hopkins University in Baltimore, Maryland. Potash, who knows Kelsoe and calls him a “very capable, very solid guy,” says that offering a test now is premature and that doctors cannot be counted on to decode it. Paradoxically, interpreting the results of Psynomics's test is likely to grow knottier as the test is refined and more genes are included. Interactions among genes are poorly understood, and many psychiatric 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. Psynomics's CEO May frames this as an opportunity: “We would like to establish that as a brand, this is the destination for information, for diagnostics, for therapeutics” in psychiatric illness. Another company, NeuroMark in Boulder, Colorado, has similar ideas. NeuroMark is designing a pharmacogenetic test for suicidality risk from antidepressants, a side effect for 2% to 6% of those taking the drugs. The company hopes to launch its test, focused on one antidepressant, Celexa, later this year, before extending it to related drugs and developing bipolar and schizophrenia risk tests as well. Then there are broader gene testing companies such as 23andMe in Mountain View, California, which offer information on SNPs behind a host of diseases. Although 23andMe could test for SNPs in depression and other psychiatric illnesses, it would need permission from companies like Psynomics and SureGene to sell information on the SNPs those companies have licensed. Despite being first in the psychiatric space, Psynomics 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 it is sending them a satisfaction questionnaire. Meanwhile, Psynomics is hunting for venture capital to tide it over until test sales ramp up. One critical but unsettled factor is whether Congress will pass the Genetic Information Nondiscrimination Act, which would protect individuals against discrimination by insurance companies and employers 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 be regulated. For many genetic tests, including that offered by Psynomics, regulation consists of lab certification but no assessment of the science behind the test. “This is a different kind of information” than has been provided in the past, says Boston University neurologist and epidemiologist Robert Green, who studies Alzheimer's genetics and early detection. “It's not always clear what it's telling us.” Gene testing could be done prenatally, for example, to check for bipolar risk in the fetus. Kelsoe shudders at the thought, especially because bipolar genes are unlikely to guarantee disease: “I think this is very dangerous, scary, eugenics-type stuff, and I don't want to get anywhere near it,” he says. Kelsoe says his company would refuse to honor such a request. But if the question comes up in the future, the decision may not be his to make. In Kelsoe's UCSD office, decorated with his children's art, a small scrap of paper is taped to his computer. In black ink is scrawled a quote from Virgil: “Fortune Favors the BOLD!” Despite that encouragement, “it's a little scary,” he says over lunch. “I've spent my whole career trying … to gain 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 wicked grin on his face. “Sometimes,” he says in a low voice, “it's fun to stir things up.” 10. SCIENCE AND COMMERCE # Hoping for a Glimpse of What's Ahead 1. Jennifer Couzin Researchers are beginning to tackle the question of how people who undergo genetic testing for psychiatric disorders interpret the results and how such information alters their lives. It's one of the first questions people ask about testing for a gene associated with mental illness: Who would want to know the result? One ready volunteer for testing is Holly Finn, 13, who lives in Grand Rapids, Michigan, with her parents and her 16-year-old sister, Katherine. Holly's mother, Kristin Finn, was diagnosed with bipolar disorder 3 months before her 17th birthday. She says that Katherine began showing signs of depression and difficulty concentrating in the fourth grade. Initially pegged as having attention deficit hyperactivity disorder (ADHD), Katherine received a bipolar diagnosis in July of 2004, when she was 12. It's been a rocky road: Katherine has switched schools three times and suffered taunting from her peers, says her mother, an active member of the Depression and Bipolar Support Alliance in Chicago, Illinois. Now the younger Finn daughter, Holly, is being treated for ADHD and anxiety. Her symptoms are different from her sister's, and she is not thought to have bipolar disorder. But, Holly says, having witnessed her sister struggle for years, she would welcome testing for a bipolar gene. “I think I might have a chance of having bipolar disorder,” says Holly, who writes songs she plays on her guitar to cope with her sister's ups and downs. A test could help her “get to the bottom of” her symptoms, Holly says, or at least prod her and her psychiatrist to keep an eye out for bipolar disorder down the road. “It certainly wouldn't hurt.” But hurt is exactly what many are afraid of. “Some people would say individuals have a right to this information and shouldn't be barred by the medical community from receiving it,” says psychiatrist Jinger Hoop of the Medical College of Wisconsin in Milwaukee, who is studying how people with schizophrenia or at risk for it respond to genetic data. Her work has confirmed what others have seen: Consumers often react favorably to hypothetical genetic knowledge. “But we don't know a lot about what the actual risks and benefits are” of receiving genetic results, Hoop notes. “We really need empirical data to let us know how people would interpret very subtle changes in risk.” One of the few projects tackling this question is the Risk Evaluation and Education for Alzheimer's Disease (REVEAL) study headquartered at Boston University School of Medicine. Neurologist and epidemiologist Robert Green and his colleagues here and at other centers are examining how nearly 700 volunteers with a family history of Alzheimer's disease respond to learning whether they carry a deleterious version of the APOE gene. A single copy raises Alzheimer's risk roughly three times, and two copies raise risk 12 to 15 times. “The bottom line is that people do want this information, and they seem to self-select for people who can handle it,” says Green. Those “who felt that they couldn't handle such information generally didn't follow through and obtain” their test results. REVEAL has examined the role of genetic counseling; whether individuals with the hazardous form, called APOE e4, are more likely to buy long-term care insurance—they are—and more broadly, what they take away from knowing their APOE e4 status, especially given that there is no way to stall the onset of Alzheimer's. One concern, notes Hoop, is that the REVEAL study has found that consumers tend to recall not their numerical risk of disease but whether they harbor the “bad” gene. This is a concern in testing for bipolar disorder or any psychiatric illness: People may think that carrying the gene means that they have or will develop the disease. Making this mistake could, among other things, induce stress, which can exacerbate psychiatric disorders. How much of a boost in risk is meaningful to consumers? The bipolar test from the company Psynomics in San Diego, California, currently includes the gene GRK3, which the company says increases risk two to three times. In someone of average risk, that's a shift from 1% to 2% or 3%. For an individual with a bipolar parent or sibling, whose baseline risk is about 10%, the increase may be more meaningful—but only if the gene compounds family predisposition. Another concern is how having such information will alter a person's life. “To me, the question is not really the risk relationship” between genes and disease, says Wylie Burke, a geneticist and bioethicist at the University of Washington, Seattle. “It's what are you going to do differently with that information?” Holly is already contemplating this question. As she grows up and considers having a child of her own, she says, she could imagine testing her son or daughter at about age 2 for genes associated with the disorder so as to be ready to spot it if it surfaces. Currently, testing of children without symptoms is not available, nor is prenatal testing. Physicians focus instead on stabilizing a woman's bipolar disorder while pregnant—something that Kristin Finn discussed in her 2007 book, Bipolar and Pregnant. But even the skeptics agree that validated genes for schizophrenia, bipolar disorder, and other psychiatric diseases are likely to emerge soon. “Biotech companies are going to do what they're going to do within the limits of the law,” says Kay Jamison, a psychologist at Johns Hopkins University School of Medicine in Baltimore, Maryland, who has written extensively about her personal battle with bipolar disorder and who generally supports making gene tests available. It may take years before it's clear what that means for people with psychiatric illnesses and their families. 11. ARCHAEOLOGY # Seeking the Roots of Ritual 1. Andrew Curry* 1. Andrew Curry is a writer based in Berlin. In the hills of Turkey, researchers are slowly uncovering the world's oldest monumental structures, strange monoliths built by hunter-gatherers perhaps 11,000 years ago. In the hills of Turkey, researchers are slowly uncovering the world's oldest monumental structures, strange monoliths built by hunter-gatherers perhaps 11,000 years ago GÖBEKLI TEPE, TURKEY—Dawn was still half an hour away when Klaus Schmidt arrived at his hilltop dig site here, but a half-dozen workers from a nearby village were already waiting for him. It was the end of the dig season, and the enthusiastic, white-haired German was there for one last look. He picked his way down steep excavation layers toward a massive, T-shaped slab of rock rising 3 meters out of the ground. The gentle light of dawn illuminated more T-shaped limestone megaliths and then still more, arranged in circles and ovals down the hillside. Some were carved with stylized arms; on others, a carved menagerie of snakes, spiders, boars, foxes, birds, and other beasts crawled and soared. Schmidt, of the German Archaeological Institute (DAI) in Berlin, claims that these are the world's oldest monumental structures, which he has painstakingly uncovered since 1995 (see sidebar, p. 280). Radiocarbon dating and stone tool comparisons indicate that they are 11,000 years old. So these great stone circles were erected before metalworking, before pottery, before the domestication of animals, even before most signs of agriculture. Although people began painting in caves thousands of years earlier, Göbekli Tepe, in southeastern Turkey, “is the first manmade holy place,” says Schmidt. He and others say that the site upends traditional notions about the development of symbolism. Archaeologists once hypothesized that agriculture gave early people the time and food surpluses that they needed to build monuments and develop a rich symbolic vocabulary. But Göbekli Tepe raises the alternative possibility that the need to feed large groups who gathered to build or worship at the huge structures spurred the first steps toward agriculture. “This shows sociocultural changes come first; agriculture comes later,” says archaeologist Ian Hodder of Stanford University in Palo Alto, California, who notes that the first domesticated wheats are from this region. “You can make a good case this area is the real origin of complex Neolithic societies.” If Göbekli Tepe is truly 11,000 years old, it also challenges the idea that symbolism and agriculture were first developed in the Levant—the area that includes modern Jordan, 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 out there now is far earlier than anyone expected. That's all back on the drawing board now.” ## A hunter-gatherer's paradise Göbekli Tepe means “navel hill” in Turkish, and at 780 meters high, it seems a natural gathering spot, the highest point for kilometers. The site was first examined—and dismissed—in the 1960s by University of Chicago anthropologist Peter Benedict, who assumed that the flints and broken slabs of limestone littering the area were the remnants of an abandoned medieval or Byzantine cemetery. Schmidt first visited the hilltop in 1994, after a local farmer had run his plow into a rectangular piece of limestone. Schmidt found not a gravestone but the top of a large, buried pillar, and he quickly recognized that the scattered flint tools on the surface resembled those from nearby sites that predate pottery. As he began digging, the mystery grew. He uncovered one huge, elaborately carved ritual structure after another, but no houses or signs of settlement. The people who built the monuments apparently backfilled them thousands of years ago, and the bones that turned up in the loose fill were all from either humans or wild animals, not from domesticated animals. So just how old are the monuments? Because most of the bones found are from backfill, dating them yields muddled results, and Schmidt has fewer than two dozen direct radiocarbon dates. Instead, he chiefly relies on comparing the stone tools and other artifacts with those from sites with more radiocarbon dates. This relative dating puts the early layers at Göbekli Tepe in the beginning of a distinctive period called the Pre-Pottery Neolithic B, dated elsewhere to about 11,000 years ago. Although the dates aren't precise, the artifact comparisons are convincing, and a few hundred years of fudge time “don't really matter,” says Harvard University archaeologist Ofer Bar-Yosef. Although the site may predate systematic agriculture, 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 and nut trees lined the rivers, and flocks of migrating birds paused here regularly. “It must have looked like a paradise, ideal for hunter-gatherers,” says Angela von den Driesch, an emeritus archaeozoology professor at Ludwig Maximilian University in Munich, Germany, who has classified animal remains at the site. The region was so rich that people could have settled down while still supporting themselves with hunting and gathering; archaeologists have found such villages within 160 kilometers of the site. Von den Driesch is one of a number of archaeologists, plus a revolving cast of eager German students and about 60 local workers, who excavate the site for 4 months each year. They drive to the site before dawn to dig, returning in the afternoon to a traditional, high-walled compound Schmidt owns in the middle of Urfa's old city. One day last fall, a dozen students quietly sorted and weighed bone fragments and artifacts in the compound's courtyard as the afternoon cooled into evening. It's a surprisingly unpretentious operation, given the imposing nature of the site and the roughly$300,000 in funding Schmidt receives each year from DAI and the German Research Foundation.

Clearing a space at a long table, Schmidt pulled out a mottled, purple schematic of the site—a sort of bird's-eye x-ray put together after a ground-penetrating radar scan in 2003. The excavated portion—3500 square meters, just 5% of the 9-hectare site's total—was represented by a white square. All around it, the scan revealed oval groups of buried monoliths dotting the hillside like dark-purple bomb craters.

Schmidt says there are at least 20 ceremonial structures still underneath the soil, and perhaps more. “It's much more complex and advanced than we thought,” he says. “Clearly, they could communicate with each other and organize something really complicated. I didn't expect hunter-gatherers could or would build something like this.”

The most spectacular ritual space, which encompasses features seen at the others, is nestled in the dig's western corner. Two 5-meter-tall, T-shaped pillars stand in the center of a circle of slightly smaller slabs, each weighing an estimated 5 to 7 tons. Parallel lines carved into the slabs' sides slant back in a V and meet at the narrow front edge like clasped hands, suggesting a stylized person. The two central stones face the valley, and the surrounding pillars face inward toward the center of the circle. In sharp contrast to the stylized human shapes, the sides of the slabs are carved with images of animals: complex arrangements of spiders and snakes, foxes and wild boars, vultures 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 symbology largely rules out fertility rituals. And the contrast between the designs carved on the pillars and bones that litter the backfill seems meaningful. The backfill yields gazelle, aurochs, red deer, boar, goats, sheep, oxen—all wild—plus a dozen different bird species, including vultures, ducks, and geese. Yet the pillar carvings are dominated not by prey but by more dangerous creatures: leopards, lions, foxes, and vultures, plus spiders, snakes, and scorpions. “The symbology is dominated by nasty animals,” says Stanford's Hodder. “It's a scary, fantastic world of nasty-looking beasts.”

## The cathedral on the hill

So what was this place? Schmidt is adamant about what it's not, although not everyone agrees with him. Despite the site's size, and the contemporary villages not far away, Schmidt insists this was no settlement. He's convinced that the circles were designed to be open to the sky, like Stonehenge. Telltale signs of settlement—such as hearths, trash pits, and small fertility figurines—are conspicuously absent. And the hilltop is a long hike from any water sources. “We know what settlements from these times look like,” Schmidt says. “This isn't one of them.” Instead, Schmidt argues that hunter-gatherers from across the region gathered here periodically, pooled their resources temporarily to build the monuments for some ritual purpose, and then left.

Rollefson and others see that as unlikely, arguing that the scale of the site would have required at least a small group of year-round residents. “To have this kind of magnet out in the middle of nowhere would be unprecedented. There would have to be so much work to carve these pillars,” he says. Indeed, Schmidt says molding and carving a single pillar would have taken months.

Some say it's just a matter of time before evidence of settlement shows up. “They haven't found much human habitation, but they will,” predicts Bar-Yosef. “It's impossible to have such a large site without people there to take care of it.” Schmidt acknowledges there must have been a few people—“personnel,” he says—but insists the site was exclusively a ritual destination rather than a settlement, which would make it unique for this period.

And researchers agree that Göbekli Tepe was a convergence point. “Certainly it was a major focus for regional celebrations or ritual activity,” says Rollefson. There are hints of a regional culture: The symbols found at Göbekli Tepe are echoed at similarly aged villages hundreds of kilometers to the south in Syria, such as one called Jerf el-Ahmar, and at slightly younger sites closer to hand in Anatolia, such as Nevali Çori. But Göbekli Tepe is clearly the grandest site and the oldest of such magnitude. “Göbekli Tepe's really the only one with that megatemple approach,” says Rollefson. “After it was built, quite a few sites with the same architecture and iconography and style appeared.” Schmidt agrees: “Here we have the religious center for settlements at least 50 kilometers away,” he says. “Those were village churches; this is the cathedral on a hill.”

Schmidt argues that the site's antiquity and the lack of domesticated animal and plant remains is strong circumstantial evidence that symbolism and religion led to agriculture and domestication, not the other way around. “Developing from hunter-gatherers to farmers happened here and spread south,” Schmidt says. “Not just architecture and monumental architecture, but turning wild animals into domestic livestock happened here. This is the starting point for a whole front of innovation.” Indeed, the region is home to a number of domestication “firsts,” including the first domesticated wheat, which emerged only 30 kilometers away at Nevali Çori about 500 years after the Göbekli monuments were built. “Its real importance is that it's early and large and symbolic a long time before agriculture,” says Hodder of the site.

But there are claims of domesticated rye in Syria as far back as 13,000 years ago. And many researchers think people planted wild crops for a long period before actual domesticated varieties appear (Science, 29 June 2007, p. 1830). Bar-Yosef notes that there's no evidence to rule out farming in this area at this time, complicating the vision of a hunter-gatherer paradise. “There's no reason to assume the builders were simple hunter-gatherers,” he says. “People can be planting wild foods.”

Bar-Yosef is also eager to have more context about the larger society that produced the site. How big did the tribe need to be to support such an intense process? “And in a simple American way,” he asks, “who paid for it all?”

Schmidt, mindful that future archaeologists may develop better techniques, refuses to be rushed. At 53, he's got 12 more years of digging until he retires, and he envisions the site's excavations lasting another half-century. Mindful of archaeology's destructive nature, he's carefully leaving something for future excavators. “The idea is not to excavate the entire settlement. The idea is to excavate as little as possible,” he says. “We just have to have enough to be sure we understand what was really going on at the site.”

12. ARCHAEOLOGY

# Just Don't Call It the Garden of Eden

1. Andrew Curry

Everyone in the city of Urfa, Turkey, seems to know Klaus Schmidt's name--and to have an opinion on his work excavating the world's oldest monumental architecture. As Göbekli Tepe's fame spreads, tourism may become Schmidt's biggest concern.

URFA, TURKEY—From his one-man booth at the Urfa bus station, tourist official Serdar Avci spends his days handing out maps to the sights around this Turkish city. To his dismay, one destination is conspicuously missing from the tourist guides: Göbekli Tepe. “People want to come here to see these temples, but Klaus Schmidt is digging so slow!” Avci says. “How many years do we have to wait to go inside? This is a big problem, I think.”

Everyone in this city of almost a million people seems to know Schmidt's name and have an opinion about his work excavating the world's first monumental architecture (see main text). Schmidt's progress is charted by regular reports in the local papers, and a 2006 cover story in the German weekly Der Spiegel speculated that the site might be linked to the biblical Garden of Eden. Much to Schmidt's dismay, Turkish papers picked up the story, and the ensuing debate about whether the site was the birthplace of Adam—considered a Muslim prophet—briefly threatened to derail further excavation on religious grounds.

Schmidt has been assigned a full-time monitor from the Turkish Ministry of Culture and Tourism in Ankara to make sure he sticks to a preapproved dig plan submitted months in advance. Colleagues say the notoriety has made it difficult for him to run the small, controlled dig he'd like. “He's always under pressure from the Turkish government,” says colleague Gary Rollefson of Whitman College in Walla Walla, Washington. “He's whipsawed back and forth. If he digs too fast, he's accused of not paying attention to detail. Too slow, and there's pressure for him to get something for tourists to see.”

As the site's fame spreads beyond Urfa, tourism may become Schmidt's biggest concern. Together with Turkish authorities, he's developing plans for a tourist center that would house replicas of the site's striking T-shaped pillars and spectacular carvings, to take pressure off the excavators and bring money to the impoverished village nearby. Construction may begin next year.

But with funding hard to come by, a finished facility is years away. In the meantime, tourists find their way to the dig even without Avci's helpful directions—about a busload of them every other day. One day this fall, 11 buses full of German retirees rolled up in a cloud of dust. Schmidt assigns a student to show visitors around or sometimes does guide duty himself. “As archaeologists, we're not used to attention like that,” says Julia Wagner, an excavator and student at the Free University of Berlin. “Working at the site every day, you forget a little how important it is.”