# News this Week

Science  14 Dec 2007:
Vol. 318, Issue 5857, pp. 1704
1. BIOMEDICAL RESEARCH FACILITIES

# London's Super-Lab Faces Hurdles

1. John Travis

Imagine that the National Institutes of Health and the Howard Hughes Medical Institute teamed up with a megacharity and a university to buy land in the heart of New York City next to Grand Central Station and agreed to spend more than $1 billion to build a biomedical research facility there. And as part of the project, the famed Cold Spring Harbor Laboratory on Long Island would close, with only some of its labs relocated to the new facility. That suggests the scope, and controversy, of the ambitious project U.K. Prime Minister Gordon Brown unveiled last week when he announced the sale of a key central London site to a coalition composed of the government's Medical Research Council (MRC), two medical charities, the Wellcome Trust and Cancer Research UK, and University College London (UCL). Next to the reopened St. Pancras station, which has high-speed rail links to the rest of Europe, the groups plan to build a £500 millionplus facility—£85 million for the land, about £350 million for the building, and the rest for equipment—that will house some 1500 scientists, many of them from MRC's celebrated National Institute for Medical Research (NIMR). “This is going to be great for British medical science, European medical science, and world medical science,” says UCL Vice-Provost Edward Byrne. Although there's no guarantee the project will live up to that promise, Richard Lerner, president of the Scripps Research Institute in San Diego, California, says the effort does signify that the prime minister believes “biological science is important to the future of the country.” Still, to achieve Brown's goal of improving the U.K. economy and the health of its citizens, the project will have to overcome major hurdles. London officials, who had designated the land for affordable housing, may try to stop it, as may those fearful of research on dangerous infectious pathogens being performed in central London near rail links to Europe. And it's not clear how the new center will blend labs from UCL, Cancer Research UK, and NIMR, which is being forced to close, to the dismay of many of its scientists. “Making it a single entity rather than a collage of contrasting colors is a strong challenge,” says Frank Gannon, Science Foundation Ireland's director general and former head of the European Molecular Biology Organization. “It's not simply having great scientists in the same location.” The proposed facility has emerged out of the struggle over NIMR, whose 19-hectare campus in north London is home to more than 500 scientists and support staff. MRC has long wanted to strengthen NIMR's efforts in translational medicine by marrying it to a research university and hospitals. In 2005, it announced that NIMR would relocate to a 0.3-hectare central London site in a project with UCL. NIMR scientists protested the move, arguing that it was an attempt to downsize their institute (Science, 4 February 2005, p. 652). When the hectare-sized lot near St. Pancras became available, however, MRC and UCL joined forces with Cancer Research UK, which was looking to relocate some 550 scientists specializing in cell growth, signal transduction, and genome maintenance from its aging London Research Institute (LRI). “Pooling our resources helps us invest in technologies we might not on our own,” says Harpal Kumar, Cancer Research UK's chief executive. The charity and MRC will also shift technology-transfer units to the new center, a move they hope will speed research findings to the clinic. With Cancer Research UK, the project “became much more exciting,” says MRC Executive Director Nick Winterton. “This is a new vision.” UCL will enable access to its teaching and specialist hospitals and has committed about £50 million for construction. UCL will also embed some 150 scientists at the new center. “The relatively small number of UCL scientists will be a bridge to the whole university,” says Byrne. Researchers will be able to “interact with almost every imaginable discipline.” The Wellcome Trust, the U.K.'s largest nongovernmental funder of biomedical research, has also committed at least £100 million. Wellcome Director Mark Walport notes that the St. Pancras location will encourage international collaboration and educational outreach through the British Library. “We immediately saw it as an important opportunity,” he says. Now comes the hard part. The project has a site, partners, estimated completion date—2013—but little more. MRC, for instance, won't detail its financial contribution yet, although the previous plan with UCL would have required it to contribute more than £200 million. And will one scientist run a unified institute? “Exactly how the governance will work has to be dealt with quite soon,” says LRI Director Richard Treisman, noting that Cancer Research UK needs to have a visible presence, as it depends on public donations. British Nobel laureate Paul Nurse, president of Rockefeller University in New York City, heads a committee that will develop science plans for the project by next year. “The aim of this group is to be as ambitious as possible,” says Byrne. What to incorporate from NIMR will likely be the stickiest topic Nurse faces. MRC has said that the new site will have animal research facilities comparable to those at NIMR now and that NIMR's World Influenza Centre will be part of the project, but it has made no assurances to other labs. “Not everyone will relocate. That's clear,” says Winterton. Although not agreeing that relocation is needed, several NIMR scientists who talked to Science acknowledge that the new proposal has much more appeal than the smaller union with UCL. Yet MRC, they note, hasn't said how many NIMR labs will be eliminated. Given the potential uncertainty during the next 7 years, they worry that colleagues will simply leave. Some even wonder whether the institute's name will live on at the new facility. “I don't know if this will mean the real end of NIMR,” says Jonathan Stoye, a virologist at the institute. 2. SPACE SCIENCE # China's Crystal-Sharp Moon Map Sets the Internet Abuzz 1. Hao Xin, 2. Richard Stone* 1. With reporting by Andrew Lawler. BEIJING—Last month, China feted its space scientists for sending the Chang'e-1 spacecraft to the moon, the nation's first mission beyond Earth orbit. Last week, however, the Chinese space establishment found itself on the defensive, after anonymous individuals in Internet forums attacked the authenticity of Chang'e-1′s first mosaic view of the lunar surface. The critiques have touched a sore spot. Chinese Premier Wen Jiabao, who unveiled the picture at a celebratory event on 26 November, hailed Chang'e-1 as “the third milestone in China's space exploration,” after placing its first satellite in orbit in 1970 and its first astronaut in orbit in 2003. Some Chinese scientists have equated questioning the veracity of the picture—a mosaic of 19 scan strips of the lunar surface taken over 3 days and processed and stitched together by an army of imaging specialists—with an attack on China itself. “Doubting the authenticity of the Chang'e moon photo is insulting the country,” the mission's chief scientist, Ouyang Ziyuan, told the newspaper Guangzhou Daily last week. Chang'e-1, named after a legendary fairy who flew to the moon, is the first stage of a China National Space Administration (CNSA) program to orbit the moon, land a probe, and return a sample to Earth. It is also the second of four planned missions—others are from Japan, India, and the United States—aiming to learn about the moon's origins and composition (Science, 31 August, p. 1163). Launched on 24 October, Chang'e-1 maneuvered into a polar orbit, 200 kilometers above the surface, on 7 November, according to the Beijing Aerospace Command and Control Center. It travels above Japan's Kaguya spacecraft, launched in September in a polar orbit 100 kilometers above the surface. Chang'e-1′s payload includes a microwave radiometer to measure soil depth, gamma ray and x-ray spectrometers to determine soil composition, a solar plasma detector to chart the distribution of solar particles impinging on the moon, and a charge-coupled device (CCD) stereo camera for mapping. “All the sensors are working well,” says Wu Ji, director of the Center for Space Science and Applied Research of the Chinese Academy of Sciences, which managed the payload and developed the radiometer and plasma detector. Wu is particularly proud of the radiometer, the first of its kind on a lunar mission. The moon image, too, is a source of pride. The sharp resolution of the mosaic, representing a swath of lunar surface between 57° and 83° longitude east and 54° and 70° latitude south, surprised even some project scientists. After seeing the first photo sent back by Chang'e-1, the camera's chief designer, Zhao Baochang, told the Chinese newspaper Science Times, “everybody at the unveiling scene was dumbstruck. … We absolutely did not expect the photo to be this clear!” Space administrators and scientists reportedly received bonuses totaling more than$1 million for “meritorious service.” Western experts are also impressed. “To be able to match the 19 images and remove the seams, they produced a very aesthetically pleasing mosaic,” says Mark Rosiek of the U.S. Geological Survey's Planetary Geomatics Group in Flagstaff, Arizona.

Within hours after the picture's release, however, critics were chipping away at it. In the “Beautiful Science” forum of People Net, one anonymous individual drew attention to a tiny rectangular shadow and asked, facetiously, whether it is the U.S. lunar rover that was abandoned on the moon. On other Internet sites, critics pointed out that the mosaic is similar to images and maps from earlier lunar missions, including a 1994 image from Clementine, a U.S. probe. Some even alleged that the mosaic is a fake.

Chinese space officials came out fighting. In a forum held live on SpaceChina Net on 29 November, Ye Peijian, designer-in-chief of the Chang'e-1 mission, said that Chang'e-1 obtained the map “using our own equipment—it is absolutely true.” He lambasted critics for doubting the image's authenticity. “This kind of speech is either irresponsible, or with ulterior motives.”

Western analysts view the Chang'e-1 mosaic as bona fide. The Chang'e-1 and Clementine mosaics are “evidently not the same,” says Emily Lakdawalla, a blogger for the Planetary Society. She points out that the lighting angle is different: “The Clementine image is lit from the top [north], while the Chang'e image is lit from the northwest,” she wrote on her blog, referring to her high-resolution image (planetary.org/blog/article/00001248).

However, it's not clear how this Chang'e-1 mosaic is oriented. Mapmaking involves projecting craters and other three-dimensional features onto a flat surface. In a Mercator projection, the north and south poles are spread out, resulting in a map with equally spaced longitudes and latitudes, and a constant compass bearing. Google Moon uses a Mercator projection.

The Chang'e-1 image Lakdawalla analyzed includes no frame and no indication of whether it is a Mercator projection. Matthew Hancher, a researcher at NASA's Ames Research Center in Mountain View, California, on the other hand, looked at a framed Chang'e-1 moon image—the one unveiled by the government—and concluded that “the map is perfectly clear” in representing a sinusoidal projection. In a sinusoidal projection, longitudes converge at the poles and latitudes are parallel. A low-resolution image is available from the CNSA Web site (www.cnsa.gov.cn/n615708/n620172/n677078/n751578/images/1798046.jpg). Rosiek came to a similar conclusion after finding that a sinusoidal projection of the Chang'e-1 mosaic almost matched a sinusoidal Clementine base map, but with a small slant, which Hancher estimated to be 5° from north.

The peculiar presentation of the Chang'e-1 mosaic and its similarities to the Clementine base map do not mean that it is a fake. “The forgery idea just doesn't make very much sense,” Hancher says. “The Chinese would be backing themselves into a corner from which it is unclear how they could possibly hope to escape.” He and others point out that Chang'e-1′s mosaic has a finer resolution than Clementine's. And Ouyang, who told Science he's still “angry” at the doubters, says new Chang'e-1 images, including a view of the moon's dark side, have been posted to the lunar program's Web site (www.clep.org.cn/index.asp?modelname=index%5Fzt%5Fkxtc&tit leno=cggxiang).

Wu and his colleagues hope to put the controversy behind them as they plan an encore: launch of the backup Chang'e spacecraft in late 2009 or early 2010. The duplicate probe would be equipped with a higher resolution CCD camera to help determine a landing site for the planned 2012 lander mission. A decision on the payload and launch date of the second probe is expected in early 2008, Wu says.

3. CLIMATE CHANGE

# Senate Bill Would Provide Billions for Deploying Cleaner Technologies

1. Eli Kintisch

A Senate panel has approved a sweeping climate change measure that would provide billions of dollars annually to commercialize energy sources that emit little or no carbon. Utility companies have long regarded the type of cap-and-trade system contained in the legislation (S. 2191) as a stick that punishes them for generating energy that the country needs. But lawmakers are also hoping that vast numbers of technological carrots will help curb the global rise in greenhouse gas emissions. “Many have said that we need a Manhattan Project for energy,” said Senator Max Baucus (D-MT) during the panel's 5 December markup of the bill. “This is it.”

So far, the debate about climate legislation has focused on who should pay as the United States overhauls its energy system. The question of how to allocate the vast sums that a federally managed system might produce, however, has received little attention. Neither has promoting synergy between the public sector and industry, on whose shoulders rests most of the burden for reducing emissions.

The Climate Security Act of 2007, approved on a vote of 11 to 8 by the Senate Environment and Public Works Committee, would require greenhouse polluters—refineries, factories, and fuel importers, as well as power companies—to procure emission permits in order to operate. Some of the permits would be provided for free, and others would be sold through a yearly auction. (Over time, the total number of permits issued would decrease, as an incentive to reduce the level of emissions, and a larger fraction would be sold.) The bill, which the Senate is expected to take up next year, would lower U.S. greenhouse emissions by an estimated 70% by 2050. Economists have calculated that an auction could generate as much as $3 trillion between 2012 and 2050. Some lawmakers, such as Senator Bernie Sanders (I-VT), see those revenues as a golden opportunity to stimulate the development and use of green technologies such as solar and wind power, along with more energy-efficient products. “Sustainable energy is and will be the future of this country. We need to give it a fair shake,” he told Science. Just what is fair, however, is a key unresolved question. Sanders had criticized an earlier version that would have forced renewables to compete for deployment funds with nuclear plants, which have been subsidized by the government for decades. The bill would now create a separate pool specifically for renewables (see The Carbon Payoff chart). Elizabeth Salerno of the American Wind Energy Association in Washington, D.C., applauds that change, although she says coal and nuclear plants will still retain “a competitive advantage.” Solar-energy researcher Nathan Lewis of the California Institute of Technology in Pasadena likes the fact that hundreds of millions of dollars in revenue each year would be allocated for basic research at the Department of Energy's new Advanced Research Projects Agency-Energy. “That's a step in the right direction but not nearly enough to make up for the deficiencies in basic research over the last 3 decades,” he says. Lewis thinks the government should spend as much on basic and applied energy research as it provides the National Institutes of Health, whose annual budget is$30 billion.

But experts believe what's on the shelf right now could have a significant impact. A recent study estimated that current technologies, if deployed widely, could reduce U.S. carbon emissions to 80% of 2005 levels by mid-century. “We need to take the solutions we have today and apply them,” says mechanical engineer Charles Kutscher of the National Renewable Energy Laboratory in Golden, Colorado, who led the study.

Others, including some advocates of renewables, believe the government shouldn't be charging companies for permits at all. George Sterzinger of the Renewable Energy Policy Project in Washington, D.C., complains that an auction would favor utilities in low-carbon-emitting regions, such as hydropower operators in the Pacific Northwest, whereas consumers in coal-dependent Ohio would pay 15% more, he estimates, as companies pass along the cost of their permits. Oil giant Shell complained earlier this year in a letter to lawmakers that forcing companies to pay for emission permits in addition to the cost of cleaning up their act could “withdraw capital from the industries and firms” involved and harm consumers.

Sterzinger prefers incentives, including tax breaks, as a way to spur renewables. That approach is part of an energy bill that the House of Representatives passed last week. (Senate action was pending as Science went to press.) An aide to Senator Joseph Lieberman (I-CT), who co-sponsored the Senate bill, agrees that companies, spurred by emissions caps, will have to do most of the heavy lifting if the country hopes to lower its carbon pollution. “The private investment will actually dwarf the public funds available,” he says. “That's the big way this bill is a Manhattan Project.”

4. RESEARCH POLICY

# NIH Weighs Big Changes in Peer Review

1. Jennifer Couzin

Peer review, a cornerstone of biomedical science, appears headed for an overhaul, to judge by a sweeping examination unveiled at the National Institutes of Health (NIH) last week. Since July, scientists have flooded two working groups established by NIH Director Elias Zerhouni with several thousand comments and ideas. This outpouring indicates that the community is frustrated by the system's administrative burden and deeply concerned about the fate of talented new investigators. Zerhouni has promised quick action.

At a meeting of the advisory committee to the NIH director on NIH's Bethesda, Maryland, campus last week, leaders of this review highlighted the recommendations they may deliver to Zerhouni in February. No final decisions have been made, however, and the committees are weighing everything, including shortening grant applications to seven pages from the current 25 pages and recommending an “editorial board” model that would refer some grant proposals to outside experts.

Molecular biologist Keith Yamamoto of the University of California, San Francisco, who also serves as co-chair of the external working group that solicited comments from outside NIH, suggested ways to ease a reviewing backlog. (Lawrence Tabak, director of the National Institute of Dental and Craniofacial Research, co-chaired the internal group.) Currently, Yamamoto noted, most applicants are permitted to resubmit their proposal twice if it's rejected the first time around, which happens most of the time. But the appeals, Zerhouni said at the meeting, have created a “traffic jam” and a system that “penalizes the new entrant to a very extreme degree.”

Yamamoto thinks reviewers ought to assess applications first for their scientific impact and, in cases that seem hopeless, communicate that unequivocally to the applicant without allowing resubmissions. “Right now, if an application is triaged”—left unscored—“many times it's unclear what the reason is,” said Yamamoto in a conversation after the meeting. “Here, the goal is to say, ‘Let's stop all that.’”

Streamlining applications—perhaps by vastly reducing the amount of preliminary data that's included—is also a possibility, as is eliminating the current scoring system and having reviewers rank only the top 10 grant proposals that they consider. Some study sections, the working groups believe, have too many members, having ballooned from the usual 15 or 20 members to as many as 80, to accommodate the increasingly specialized science being proposed. Sending applications containing certain technical details to outside experts, who would consider those elements alone and report back to the study section, is one way to slim study sections down. Shorter grant proposals, meanwhile, could allow each one to be evaluated by four people instead of the usual two.

Zerhouni and his advisory committee seemed enthusiastic, but several members wondered if the proposed changes went far enough. “The biggest [issue] on the minds of the people I talk to is getting the best people to serve on study sections,” said advisory committee member Thomas Kelly, director of the Sloan-Kettering Institute in New York City. And, he added, “I'm skeptical” whether the incentives proposed will be sufficient to coax these people to serve. Bioengineer Annelise Barron of Stanford University in Palo Alto, California, worried that slimming down the applications might mean those from less prestigious universities would not fare as well, because with shorter applications, name recognition could carry more weight. The NIH system cannot allow such a bias or must find a way to manage it, agreed David Botstein of Princeton University, who's working with Yamamoto on reviewing peer review. Some wondered whether blinding the names and affiliations of grantees would be possible.

No matter what the working groups decide, it's critical that NIH retains the scientists it's helped train and gives investigators “a sense of commitment that's real,” says pharmacologist and cardiologist Garret FitzGerald of the University of Pennsylvania, an NIH adviser not directly involved in this review. “Otherwise,” as the average age of first-time grantees continues to rise, says FitzGerald, “what rational person would choose to go into a career where you begin to be independent when you're 45?”

5. NEUROBIOLOGY

# Immune Molecules Prune Synapses in Developing Brain

1. Greg Miller

The complement cascade is part of the body's innate immune defense: a protein work crew whose duties include tagging bacteria and other bad guys for elimination. A new study suggests that complement proteins may have a surprising yet analogous function in the developing brain, tagging unwanted synapses for removal. The work also hints that these proteins may promote synapse loss in early stages of neurodegenerative disease.

“It's a pretty provocative finding,” says Greg Lemke, a neurobiologist at the Salk Institute for Biological Studies in San Diego, California. “This is part of a growing body of evidence that many molecules of the immune system have a second set of jobs in the brain,” says Lisa Boulanger, a neurobiologist at the University of California, San Diego.

The new study, which appears in the 14 December issue of Cell, began as an attempt to determine whether neural support cells called astrocytes have a role in refining synaptic connections between neurons during development, says senior author Ben Barres of Stanford University in Palo Alto, California. Postdoc Beth Stevens and colleagues used gene chips to look for changes in gene expression in neurons from the developing retinas of rats when the neurons were cultured with astrocytes.

To their surprise, astrocytes spurred the neurons to crank out a complement protein called C1q, which elsewhere in the body kicks off a cascade of chemical events that culminates in the destruction of an intruding cell. In experiments with mice, the researchers found that C1q concentrations in the retina and brain peaked a week or so after birth and dropped dramatically as mice matured. The peak coincided with the period when unwanted synapses are pruned. More intriguing, C1q seemed to concentrate at puny, immature-looking synapses in the developing nervous system.

When the researchers examined the brains of mice lacking a functional C1q gene, they found that development had gone awry in the lateral geniculate nucleus, a relay station in the brain that receives synaptic inputs directly from retinal neurons. In normal mice, geniculate neurons initially receive inputs from both eyes and then prune them so that they only receive input from one eye or the other. In the mutant mice, geniculate neurons maintained extraneous inputs from both eyes into adulthood.

That's a striking finding, Boulanger says: “When you get rid of these proteins that we thought just functioned in the immune system, it disrupts a very specific event that we think is involved in making the precise, final connections in the developing visual system.” Many questions remain, however. Barres suspects that complement proteins mark unwanted synapses for removal by microglia, immune cells in the brain. More work is needed to demonstrate that, Boulanger says, and to figure out why only certain synapses are flagged for removal.

Finally, Barres and colleagues collaborated with Simon John's group at the Jackson Laboratory in Bar Harbor, Maine, to investigate whether C1q might have a role in synapse loss in a mouse model of glaucoma. Compared to normal adult mice, adult glaucoma mice exhibited elevated C1q levels: The protein accumulates at retinal synapses early in the disease, even before synapses disappear and neurons die off.

Synapse loss precedes cell death in Alzheimer's and other neurodegenerative diseases, Barres notes. He speculates that drugs that block the complement cascade may forestall neurodegeneration in a number of disorders. It's an exciting idea, says Monica Vetter, a neurobiologist at the University of Utah in Salt Lake City: “There's good evidence that these complement components are upregulated in other diseases.”

6. PHYSICS

# Simple Scheme Stores Light by Converting It Into Vibration and Back

A few years ago, physicists slowed light to a crawl and then stopped it entirely (Science, 26 January 2001, p. 566). To do that, they exploited strange quantum-mechanical interactions between light and atoms in a gas, converting a pulse of light into a subtle arrangement of spinning atoms. On page 1748, three physicists report a simpler way to hit the brakes: They convert light in an optical fiber into a slow-moving vibration and then back into light.

“This has the enormous advantage of simplicity,” says Stephen Harris, an applied physicist at Stanford University in Palo Alto, California, and a pioneer of the atomic techniques. “Conversely, it can't do some things that the other techniques can.”

To store a pulse of laser light in a cloud of atoms, researchers shine a second laser into the cloud at the same time. The overlapping light fields interact with the atoms in a way that greatly decreases the light's speed. The light also nudges each atom into a strange quantum-mechanical condition in which it spins in two different directions at once. The precise spin mixture varies from point to point in the cloud, effectively freezing the light pulse into the atoms when the reference laser is turned off and holding it until the laser comes back on. Others have managed to store light by shunting it into tiny optical “resonators” for a fraction of a nanosecond.

To find another way, Zhaoming Zhu and Daniel Gauthier of Duke University in Durham, North Carolina, and Robert Boyd of the University of Rochester, New York, opted for an optical fiber. They fed a “data” pulse in one end and a short, intense “write” pulse in the other. When the two collided, the data pulse disappeared and was replaced by a vibration crawling along at just 1/40,000 the speed of light in a fiber. To convert the vibration back to light, the researchers hit it with a “read” pulse identical to the write pulse.

The fiber vibrates because the light makes it contract in the spots where the light is most intense. To make the conversion efficient, the team tuned the frequency of light in the read pulse slightly lower than that in the data pulse. The two had to differ by the frequency of the vibration, which was fixed by the properties of the fiber. The researchers showed they could store a train of three 2-nanosecond pulses and retrieve it as much as 12 nanoseconds later.

The new technique works for any frequency of light that will pass through the fiber, Gauthier says. The atomic and resonator techniques generally work at one frequency.

The conversion doesn't depend on quantum mechanics, notes Lene Hau, a physicist at Harvard University and one of the first to stop light. That should make the effect more robust but rules out truly bizarre embellishments. For example, Hau and colleagues have encoded a light pulse in one cloud of atoms and revived it in another cloud by letting a few atoms drift between the two, as they reported 8 February in Nature. Such a feat would be impossible with the fiber technique. Still, Hau says, “it's very important to try different systems.”

The atomic systems might someday provide the memory for quantum computers, Harris says. Gauthier sees more immediate uses for the fiber-optic approach. For example, it might be used to measure the correlations between signals in optical networks. But first researchers must increase the storage time and reduce the power in the read and write pulse from a walloping 100 watts. That's enough to shake up anybody.

7. ECOLOGY

A new study suggests that fish farming could rapidly wipe out some populations of wild salmon in British Columbia. Although some researchers are calling for dramatic controls on the industry, others say the risk hasn't been established firmly enough. At stake is the $450 million aquaculture business. One of the top concerns about aquaculture is the spread of disease and parasites to wild species. On page 1772, the first population-level analysis suggests that sea lice from farmed salmon will cause several populations of one species of salmon in British Columbia to plummet by 99% within 8 years. “It's a shocking number,” says salmon conservation expert John Reynolds of Simon Fraser University in Burnaby, Canada, who was not involved in the research. But environmental physiologist Scott McKinley of the University of British Columbia in Vancouver worries about rushing to judgment. “You cannot conclude anything from a correlation,” he says. Sea lice are small crustaceans that latch onto salmon and other fish. They feed on tissue and create lesions that make it hard for fish to regulate their body fluids. The saltwater parasites naturally occur on adult salmon in the sea but not on juveniles, which hatch in fresh water and then swim to the sea. In 2001, however, researchers found significant numbers of sea lice on wild juveniles that had passed by fish farms in British Columbia. The situation was alarming because young pink salmon are more vulnerable to damage from lice than adult salmon are. Graduate student Martin Krkošek of the University of Alberta, Edmonton, started studying the problem in 2003. In previous papers, he and colleagues calculated that juvenile pink salmon are 73 times more likely to be infected with sea lice after they passed by salmon farms than are fish that didn't pass by and that lice can kill between 9% and 95% of juvenile pink salmon, depending on how many fish farms they must swim by. Some researchers are unconvinced, however, and point to other studies that suggest lower mortality from sea lice. In the new work, Krkošek and colleagues investigated the extent to which sea lice are affecting pink salmon populations throughout the Broughton Archipelago near Vancouver Island. They analyzed 35 years of records from the Canadian fisheries agency on the number of salmon in seven rivers that flow into marine channels with fish farms. They also looked at 64 rivers from which migrating salmon do not pass by fish farms. Using a standard model, they calculated that pink salmon not exposed to fish farms showed the same range of population size for all 35 years, varying from year to year. The pink salmon that swam past salmon farms showed the same pattern, until the lice infestations began in 2001. Then all seven populations shrank year after year. If these populations continue to decline at this rate, they will be 99% gone within four generations. “It's very fast,” says Krkošek, who says immediate conservation steps are necessary. “We can't sit around and do more research, because these fish will be gone.” Senior author Mark Lewis of the University of Alberta in Edmonton and another co-author were among 18 scientists who in September called for requiring salmon farms to be surrounded by barriers to prevent the spread of parasites or disease. As with previous papers, the reaction to the new finding is polarized. McKinley and others say that there are too many unknowns to conclude that sea lice from farms harm wild salmon. Many factors influence their abundance, including fluctuations in ocean nutrients. But fisheries biologist Ray Hilborn of the University of Washington, Seattle, says it is too risky to farm fish in open pens near wild relatives: “The bigger concern is that [sea lice] are just one of many pathogens. There could be other things out there that we don't know about.” 8. REEFS IN TROUBLE # Spawning for a Better Life 1. Elizabeth Pennisi With our planet's besieged corals the focus of the International Year of the Reef in 2008, scientists are racing to decipher the riddles of coral reproduction The numbers aren't working in Michael Henley's favor. Two months earlier, this invertebrates aquarist at the National Zoo in Washington, D.C., returned from Puerto Rico with three Nalgene bottles brimming with coral larvae—12,000 or so of the tiny creatures—to do his part in an international bid to grow an endangered coral species from scratch. He hopes to nurse the elkhorn coral (Acropora palmata) larvae through the fragile swimming stage of their life cycle and entice them to build miniature reefs in a saltwater tank. Henley has labored to make the baby corals feel right at home. Inside the 350-liter aquarium, two underwater jets pulse, simulating surges, and a pair of 400-watt lamps suspended over the tank are stand-ins for the blazing Caribbean sun. He's fine-tuned the filtering system to get the turbidity just right, and he nourishes the coral with newly hatched brine shrimp, oyster eggs, and rotifers. Henley spices up the cuisine with algae, which adult coral polyps host as symbionts that provide carbohydrates and other nutrients. “If they don't take [the algae] up, they won't live long,” says Henley. Despite Henley's tender care, many did not live long. One month after the larvae arrived at the zoo, only 158 found safe haven in tiny grooves etched into ceramic tile terraces in the tank. Another month later, Henley has found just two millimeter-long animals. “It doesn't look good,” he says. Like coral lovers around the world, Henley feels a sense of crisis. Climate change, disease, and human activities such as overfishing and coastal development have destroyed 20% of the world's 285,000 square kilometers of known reefs, threatening biodiversity hot spots that generate an estimated$30 billion a year in revenue, mostly from fisheries and tourism. In the Caribbean, populations of elkhorn and staghorn corals have dropped so low that in May 2006, the United States listed them as threatened species. Last September, for the first time, corals were inscribed on the Red List of Threatened Species. Two coral species from the Galápagos—Floreana coral (Tubastraea floreana) and Wellington's solitary coral (Rhizopsammia wellingtoni)—are considered critically endangered, and a third, Galápagos coral (Polycyathus isabela), is designated as vulnerable. According to the Status of Coral Reefs of the World: 2004, nearly half of remaining reefs are imperiled and could collapse as soon as 20 years from now.

Prospects for recovery are grim. With more frequent and severe spikes in ocean temperatures, corals seem to be having an ever-harder time reproducing and surviving. Although reefs are built in large part through asexual cloning, “sexual reproduction is possibly the most important process in the replenishment of degraded reefs,” says James Guest, a marine biologist at the University of Newcastle upon Tyne, U.K. Larvae can migrate to where conditions are more amenable, and the mixing of gene pools may be critical to survival in a rapidly changing world.

Henley, part of a consortium called SECORE, short for Sexual Coral Reproduction, hopes that by growing corals in aquaria, he and his colleagues can keep beleaguered species from disappearing. Marine biologists, too, want to jump-start new reefs. “Our vision is to do for coral reefs what [others] do for forests” by developing seedlings for reforestation, says Alina Szmant, a marine biologist at the University of North Carolina, Wilmington. Other teams are blending old-fashioned husbandry and 21st century science to work out how corals know when to spawn and what fate befalls larvae. “A lot of our state of knowledge is an accumulation of hunches,” laments Margaret Miller, a marine biologist with the National Marine Fisheries Service Southeast Fisheries Science Center in Miami, Florida.

Some hunches are paying off. By combining satellite climate data and spawning records, one team has begun to overturn the long-held view that surface water temperatures regulate the spawning clock. DNA studies have uncovered proteins that help detect light and fine-tune the spawning schedule. Studies are revealing the molecular pas de deux of coral-algal partnerships. And genetic tags are making it possible to trace reef genealogies to find out which are the fittest and likeliest to withstand environmental perturbations that are devastating more vulnerable reefs.

## Time to spawn

Until recently, catching corals in the act has been more art than science. Spawn-chasers must contend with bad weather, miscalculated spawning times, problems with permits to transport live corals, and a host of unknowns about conditions conducive to survival of young coral (see sidebar, p. 1715).

Although coral clones form the bulk of the pillars and billowing mounds of healthy reefs, after a few years' growth, individuals in the colonies become sexually active. Most species spew gametes into the water in a milky frenzy of fertilization en masse. For many corals, this is a once-a-year event. Thus, “spawning together is a biological imperative,” says Miller.

To ensure sufficient concentrations of egg and sperm for fertilization, corals need to know the right month, the right day, even the right hour to release gametes. For decades, researchers assumed that warming sea surface temperatures stimulate the production and maturation of gametes, which burst forth a few days after the full moon, at a precise time after sunset.

The role of temperature as the seasonal cue never felt quite right to Robert van Woesik, an ecologist at the Florida Institute of Technology in Melbourne. “The problem for decades has been that we have been studying coral spawning at mid-latitudes,” he explains. In the tropics, water temperatures vary by only about 3°C—not much of a cue, he contends. At higher latitudes, corals off western Australia spawn in late summer, whereas those off the east coast spawn in spring, even though water temperatures peak in both regions in summer.

From September 2002 until February 2003, van Woesik and research fellow Lolita Penland kept tabs on two tropical reefs in Palau. They discovered that multiple species spawned multiple times over the course of the year. Temperature didn't seem to matter. Rather, most of the time, spawning coincided with the spring and fall equinoxes, when the amount of insolation, or sunlight hitting any one spot, is highest. A literature review revealed that spawning schedules in the Great Barrier Reef and off Japan also track insolation.

Van Woesik and his colleagues next looked at spawning in the Caribbean. They combed the literature and got timing information for 12 species, including elkhorn, brain, and star corals, from Venezuela to Bermuda. Again, insolation seemed to set the spawning clocks. It could be that insolation fluctuations affect symbiont productivity, periodically fueling gonad growth, van Woesik says.

Some reef researchers are dubious. “We need experimental work to understand how these environmental cues are translated into physiological change by the organisms,” says Guest. And although the Caribbean data are “pretty solid,” adds Andrew Baird, an ecologist at James Cook University in Townsville, Australia, he is not convinced that the same patterns prevail elsewhere. “It's possible that these cues differ among regions,” says Baird, who in the past 5 years has sampled gametes from 20 sites in the Indian and Pacific oceans. He has found two peaks of spawning activity, loosely tied to the start and finish of the monsoon season.

No matter what puts corals in the mood, the spark that ignites a reproductive melee is moonlight. Most corals spawn within a week of the full moon. Yet corals lack eyes of any sort. Researchers have recently discovered that at least one species “sees” moonlight thanks to cryptochromes, proteins that sense blue light (Science, 19 October, p. 467).

After locating cryptochrome genes in the fingerlike coral Acropora millepora, Oren Levy, a marine biologist at the University of Queensland in St. Lucia, Australia, and his colleagues monitored the activity of two of those genes. One revved up protein production at dawn, whereas the other stirred hours later. When kept in constant darkness, the genes were quiet. One gene's activity tracked the moon's phase; protein production soared during the full moon. Szmant thinks these data are inconclusive. Yet if they hold up, the findings “provide an important mechanistic link in how mass spawning in corals is cued,” says Miller.

## Settling down

Nailing the timing is making it easier to predict and witness mass spawnings and to begin to understand this stage of the coral life cycle. Typically, spawning corals disgorge eggs and sperm packaged in a mucous bundle that floats to the surface and bursts, freeing the gametes. A cnidarian orgy ensues. Within a few days, the free-swimming larval progeny disperse and colonize.

The larvae are a weak link in the coral life cycle—a frailty exacerbated by climate change. Two years ago, Szmant and Miller gathered the larvae of elkhorn and mountainous star coral in the Caribbean and allowed them to settle on limestone plates that had been left for weeks or months on the reef to approximate natural surfaces and then brought back to the lab. They noted where the larvae preferred to land, put the plates back on the reefs, and watched what happened.

In lab experiments, the larvae tended to take root where there was encrusting red algae, but often new settlers would be over-run by the expanding algal patch. Star coral preferred the undersides of plates—even if the plates were put back upside down. This species fared poorly: Three-quarters of individuals disappeared within a month, and none survived a year. Elkhorn coral populated the top of the plates, and 3% remained alive after 9 months.

Puzzled about why coral would tempt fate by settling near the aggressive algae, Andrew Negri of the Australian Institute of Marine Science in Townsville took a closer look at this odd dynamic. He and his colleagues now argue that bacteria associated with the algae are the true cues for larval settlement. Szmant has evidence supporting this contention. Knowing the right cues, she says, “could lead to the development of targeted restoration measures.”

Temperature, too, is a huge factor. Szmant and her colleagues have utilized a fluorescent protein found in larvae to follow them as they put down roots. In the summer, if seawater temperatures climb 1 or 2 degrees Celsius above normal, larval survival and settlement plummet, she and her colleagues are finding. Similarly, a team led by Paul Sammarco, a marine biologist at the Louisiana Universities Marine Consortium in Chauvin, has found that larvae of the species he studies—such as the Atlantic brain coral (Diploria strigosa) and reef-building and soft corals on the Great Barrier Reef—succumb to above-average water temperatures.

Not every species is under siege. Even where key reef-builders are in trouble, “weedy” corals, such as mustard hill coral (Porites astreoides) and lettuce coral (Agaricia agaricites), are doing okay. Most of these corals brood their young, and Szmant hopes to zero in on differences that gird brooders against environmental forces that hit spawners hard.

And there is plenty of geographic variability. In the Caribbean, for example, the prognosis is not bleak for every reef. Larvae seem to settle just fine near Bonaire and Curaçao, even where development has resulted in pollution and high sediment loads. Szmant and Miller are panning for clues to these reefs' secrets to success.

In the Florida Keys, Miller has teamed up with molecular ecologist Iliana Baums of Pennsylvania State University in State College to probe whether survival depends on having the right parents. In 2005, Baums developed a genetic test that distinguishes individual elkhorn corals. The test uses short bits of variable DNA called microsatellites. Last summer, Miller and Baums collected egg and sperm from elkhorn and mountainous star coral, looked at their microsatellite makeup, and cross-fertilized gametes from a number of individuals. They tallied how many larvae of each cross settled on plates coated with various communities of microbes and algae. “We saw incredible differences in the performance of different crosses,” says Baums. Some crosses yielded few offspring, whereas others were prolific. Reef restoration may depend on collecting gametes from the hardiest parent colonies, says Baums.

## Matchmaking

Larvae lucky enough to find turf to colonize can't go it alone: They must latch onto a symbiont. They tend to be picky about which microscopic zooxanthellae algae will get a cozy intracellular home in return for generating a stable food supply. Mary Alice Coffroth, a marine biologist at the University at Buffalo, New York, and her colleagues have been teasing out the molecular signals that underlie a successful match.

Researchers suspect that larvae recruit algal partners from the water column. Coffroth and others are showing that larvae are selective but not dead set on particular species. Coffroth evaluated the potential of free-living Symbiodinium to join forces with young corals called octocorals by suspending trays of newly settled larvae 20 meters above a reef. At the beginning of the experiment, none of the larvae had algae, but by the end, Coffroth and her colleagues found that just three of the many Symbiodinium varieties in the water had taken up residence in coral polyps. They found a similar pattern with algae from reef rubble and other bottom surfaces where octocoral colonies live, they reported in the 5 December 2006 issue of Current Biology.

It remains a mystery how corals and zooxanthellae strike up a relationship. To eavesdrop on the molecular signals of courtship, Mónica Medina, a geneticist at the University of California, Merced, and her colleagues have collected larvae, extracted DNA, and pulled out bits of genes called expressed sequence tags (ESTs), which allow researchers to study genes without knowing the full sequence. So far they have 10,000 ESTs for each of two coral species and about 900 zooxanthellae genes represented. They are putting the DNA to work on chips that can monitor changes in expression in thousands of genes at a time. In 2006, Medina and others collected eggs and sperm of elkhorn and mountainous star coral from a Florida reef, then used DNA chips to monitor gene expression after fertilization. When they expose larvae to zooxanthellae, “the corals go wild,” says Medina, switching on scads of genes.

Virginia Weis, a cell biologist at Oregon State University in Corvallis, has some ideas about which genes are important. In studies of the sea anemone, a coral cousin that also has symbionts, she discovered that a nascent relationship hinges on run-of-the-mill genes involved in programmed cell death and cell division.

## Planned parenthood

Understanding the molecular details of symbiosis should one day help guide efforts to save coral species. But the National Zoo's Henley and his three dozen SECORE partners aren't waiting for answers. Nor are Szmant and Miller. They are taking coral sexual reproduction into their own hands.

In 2001, Dirk Petersen and Michael Laterveer of the Rotterdam Zoo in the Netherlands set up SECORE to encourage aquarists to bring coral reproduction in-house rather than depend on harvesting coral from the sea. Working with larvae of two species, Atlantic brain coral and boulder star coral (Montastraea annularis) from Curaçao, Petersen figured out how to gather and fertilize gametes and raise larvae.

Last year, SECORE researchers and aquarists netted 900,000 elkhorn larvae in Puerto Rico. Each night during the spawning season that August, divers collected gametes and quickly returned them to shore or boat, where others would mix gametes from different spots, gently rocking coolers to encourage fertilization. “The whole crew had to work in 2-hour shifts, 24 hours a day for 4 days,” Petersen recalls.

They wound up with far more larvae than they could handle. Convention on International Trade in Endangered Species of Wild Fauna and Flora permits did not come through in time for the European participants. And there were fresh hassles. Previously, Petersen had transported larvae in carry-on luggage. But last year, new security rules relegated corals to baggage, and 100,000 larvae destined for the Columbus Zoo and Aquarium in Ohio got lost in transit for 2 days and perished. Months later, only corals under the care of Mitch Carl of the Henry Doorly Zoo in Omaha, Nebraska, had survived. He distributed many to other SECORE members, and 821 colonies are now thriving.

The second time around, last August, Henley and other SECORE colleagues each took home about 10,000 larvae. Once again, only Carl had the magic touch: He is raising some 900 incipient colonies. “I'll be happy if I have one colony left after 1 year,” says Henley.

Likewise, Szmant and Miller have little to show for their efforts to collect and fertilize gametes during spawning expeditions. They release some larvae right away and put others on the reef after these have settled on ceramic plates in the lab. “We have been trying various things to increase survivorship, but no major breakthrough yet,” says Szmant. Baums's findings suggest that a breakthrough may come through “breeding” corals that produce hardy offspring.

The stakes are high to get it right. “If the larvae can survive and settle and grow into adults,” Szmant says, “then there's hope that the reefs will recover.”

9. REEFS IN TROUBLE

# Moonlight Sonata on the Reef

1. Cheryl Jones*
1. Cheryl Jones is a science writer based in Canberra, Australia.

MAGNETIC ISLAND, AUSTRALIA—On the night of the full moon last October, a couple of dozen scientists from around the world converged on this tropical island off Townsville in northeastern Australia to witness a marvel of nature: an upside-down snowstorm.

Decades of observations suggested that Acropora millepora, a common coral of Australia's Great Barrier Reef, would erupt in an hourlong frenzy of reproduction precisely 5 days after the full moon. But the corals ad-libbed, releasing a blizzard of male and female sex cells into the warm tropical waters 2 days earlier than expected. “Coral rarely reads the script; this is biology,” says David Miller, a molecular biologist with the Australian Research Council (ARC) Centre of Excellence for Coral Reef Studies at James Cook University (JCU) in Townsville, Australia.

Fortunately, Miller and his colleagues were tipped off. They had on hand a sample of A. tenuis, which spawns a few hours before A. millepora. When the researchers saw A. tenuis spawn, they sped out to the reef to bring coral samples ashore in tubs of water. On the beach under a gibbous moon and amid the wails of the bush curlew, A. millepora released egg-sperm bundles in delicate mucous sacs into the “baby baths.” Later that night, the researchers, working in a makeshift lab in a seaside guesthouse, divided the egg-sperm bundles between plastic tubs in which they would raise larvae over the next few days.

The researchers hope that work on A. millepora—the species that has been most extensively studied at the molecular level—will point to ways to minimize global warming-inflicted damage to the 2000-kilometer-long Great Barrier Reef, which generates $4 billion a year in revenue. Diseases, most of them poorly understood (see sidebar, p. 1716), and urban runoff are among the villains. A looming threat is acidification of the seawater from dissolved carbon dioxide (Science, 4 May, p. 678, and p. 1737 of this issue). To address these critical issues, four teams came to Magnetic Island to raise larvae. There's no magic formula. Some scientists filter and change water regularly, whereas others say this stunts larval growth. In an even simpler approach, a team led by ecologist Andrew Baird of JCU raises larvae in$1 buckets for studies about how temperature affects the larvae's ability to latch onto energy-giving algal symbionts.

A group led by Miller and Eldon Ball of the ARC Centre for the Molecular Genetics of Development has the big picture in focus. They are using high-throughput sequencers to compile a catalog of expressed genes as part of a “poor man's coral genome project.” They hope to get most of the coding sequence out next year. Already, their labs have identified about 10,000 genes in A. millepora. Some, including genes critical to the vertebrate immune system, were once thought to be vertebrate innovations because they were absent from the fruit fly and the nematode worm, the archetype of invertebrates. “It's a vertebrate-centric view of the world,” says William “Bill” Leggat of JCU.

Miller and Ball hope to find out which proteins among coral's repertoire of tens of thousands are expressed at various stages of development and in response to environmental shocks. In one fine-grained study, Miller is using genomic data to zoom in on the molecular basis of symbiont uptake and calcification. The research will include a comparison of coral genes with the full genome sequence of the sea anemone (Science, 6 July, p. 86). Another project, by Victor Hugo Beltrán Ramírez of JCU, is probing two proteins that turn on corals' vivid green and red fluorescence. The proteins may be involved in either photosynthetic enhancement or photoprotection in adult corals.

Leggat and François Seneca are investigating coral bleaching, a phenomenon in which coral polyps expel their algal symbionts if water exceeds the seasonal average temperature by 2°C for a couple of weeks. They want to find out which genes are switched on during the crises, which can wipe out entire reefs. The Magnetic Island reefs were hit hard by bleaching in 2002. One hypothesis is that free radicals generated by the heat-stressed symbionts disable their ability to photosynthesize. Coral somehow senses this and ousts its partners.

Two days after the spawning at Magnetic Island, Lubna Ukani, a molecular biologist in Miller's group, was bent over a microscope in the guesthouse-cum-laboratory. The ground floor was covered with baby baths, and tables strained under a mass of microscopes and chemicals. Ukani was chemically preserving 1-millimeter-diameter, pearshaped white larvae for the genetic research.

“Tomorrow, they'll start swimming,” Ukani says. After reaching this stage out on the reef, some larvae corkscrew down to the bottom, looking for a patch of sea floor that they will call home for the rest of their lives—as long as 100 years. Other larvae drift for months, eventually settling far away. That wanderlust, an evolutionary adaptation to cope with changing sea conditions, may be the key to survival for coral reefs as the planet warms.

10. REEFS IN TROUBLE

1. Cheryl Jones

Disease is an unsung villain in the global degradation of coral reefs. Here are a few of the nastiest plagues that experts are keeping an eye on. All are expected to worsen as global warming nudges up average ocean temperatures.

Caribbean yellow-band disease is hitting reefs hard. “In Puerto Rico, some of the reefs that I have been studying for 8 years have lost 60% of their live coral tissue,” says Ernesto Weil of the University of Puerto Rico. The disease has struck large, old colonies of Montastraea, or star coral, the main reef builder in the region. Scientists have not pinned down the pathogen that leaves yellow rings as it chews through a colony, nor do they know if it targets the coral or its algal symbionts. Outbreaks, once confined to summer, are now “permanent,” says Weil, and assaults have been quicker and deadlier than in the late 1990s, when the disease first appeared. The increased ferocity seems to be correlated with a rise in average minimum water temperatures, Weil says.

The Caribbean white diseases are a group of bacterial diseases that show up as bands or patches of bare white skeleton. In the early 1980s, white-band disease wiped out as much as 95% of the acroporid corals throughout the Caribbean. White plague first struck in the late 1970s in Florida. Since then, it has appeared regularly across the Caribbean, afflicting 42 of the region's 60 coral species. The disease advances and retreats in sync with seasonal changes in water temperature. Although white plague remains “a major concern,” Weil says, yellow band is emerging as a bigger threat.

Aspergillosis, caused by the soil fungus Aspergillus sydowii, exacted a heavy toll on Caribbean sea-fan corals in the mid-1990s. The disease is now entrenched but at low levels, says C. Drew Harvell of Cornell University. Studies suggest that the pathogen will thrive in warming oceans, she says. One glimmer of hope is that the sea fans are fighting back. They are “resilient,” says Harvell, “and may have evolved increased resistance.”

White syndrome is a single disease—or a suite of diseases—that was first spotted about 10 years ago on the Great Barrier Reef and later detected in the Marshall and Hawaiian islands and in Palau. It can wipe out entire colonies in weeks or months. Although the causes are unknown, there is evidence for both bacterial infections and runaway cell death. A 6-year study of 48 reefs spanning 1500 kilometers of the Great Barrier Reef linked outbreaks to rises in sea surface temperatures and to coral density, says Bette Willis of James Cook University in Townsville, Australia. Global warming is likely to trigger more outbreaks, she says.

Scientists are warily eyeing Montipora white syndrome, which is attacking one of the three main coral genera in the Hawaiian Islands. “Changes in disease levels are starting to concern us,” says Greta Aeby of the Hawaii Institute of Marine Biology in Kaneohe. She and others suspect that urban runoff, especially in Oahu's south Kaneohe Bay, is fueling a slow but sure advance of the disease, the cause of which is eluding scientists. Says Aeby, “This is one disease we expect will get worse.”

11. REEFS IN TROUBLE

# Life on the Mean Reefs

1. Christopher Pala*
1. Christopher Pala is a writer based in Honolulu.

The short, nasty existence for reef-dwelling fish at two primeval atolls suggests that intensive fishing elsewhere has skewed predator-prey dynamics

Imagine an atoll in the time of Eden. It would be teeming with fish, a few big ones and a lot of little ones swarming among the coral reefs.

Think again, says a group of marine biologists who have been studying the Line Islands south of Hawaii. Led by researchers at the Scripps Institution of Oceanography in San Diego, California, they are comparing Kingman and neighboring Palmyra—U.S. possessions that are among a handful of Pacific atolls virtually untouched by humans—with Fanning and Christmas, which belong to Kiribati and respectively have some 3000 and 6000 residents.

To their surprise, the scientists found that life at Kingman is anything but idyllic. The ecosystem is dominated by large predators to an extraordinary degree: About 85% of the estimated mass of all fish is made up of apex predators such as sharks, large jacks, and snappers. For the prey fish, life in the real Edens of Kingman and Palmyra is just as the English philosopher Thomas Hobbes had described it for humankind without society: “Nasty, brutish, and short.”

“Inverted pyramids have been documented in plankton but never within a community of large animals,” says Scripps's Stuart Sandin, the project's coordinator. “The intensity of predation is new.” The more familiar pyramid occurs at Christmas Island, where sharks have been fished for their fins and jacks and snappers for food. There, apex predators make up only 15% of the biomass. Their findings are in review at two publications.

To begin to understand the future of reefs in a warming world, it's not only important to unravel the mysteries of coral but also essential to work out the dynamics of reef communities. The more researchers learn, the more acutely they feel the need to restore reefs to a state resembling primeval Kingman and Palmyra. “If we don't protect these places, it will be the end of true natural selection in the oceans,” says Alan Friedlander of the National Oceanic and Atmospheric Administration's biogeography branch in Hawaii.

Seen underwater, Kingman, a mostly sunken atoll 15 kilometers long, “looks like a small town at the start of a classic Western movie,” Friedlander says. “There's nobody out on the street except for a few big bullies, and the basements are crowded with petrified locals.”

“It looks weirdly empty,” agrees Sandin. “In Fiji, you'll see hundreds of colorful fish milling around, even though there's 20 times more fish [by weight] at Kingman.” Crevices are so precious that when one occupant ventures out to feed, another takes its place. “It's hot-bunking,” says Friedlander, referring to the Navy custom of having three men occupy one bunk in 8-hour shifts. As a result of this lopsided structure, a Kingman denizen from a prey species has little chance of reaching puberty. “That's why these fish are evolutionarily so important,” says Sandin. “Only the very fittest survive.”

How can so many predators live off so few prey? Because prey reproduce and grow much faster than predators do. “It's like your lawn,” says Friedlander. “The more you mow it, the faster and thicker it grows.” At Kingman, half-pints like surgeonfish, wrasses, and damselfish live about half as long as counterparts at Christmas, where the average prey is 20% bigger.

Still, life is literally no picnic at the top of the food chain. Examining predators' stomachs, the scientists found them mostly empty. “Kingman has shown us that unlike mammals, fish can survive with very little food. They simply grow slower,” Sandin says. And they don't turn on each other, he says, because they don't want to risk injury. Nowhere else has Sandin observed the voracious scrutiny that spearing elicits at Kingman and Palmyra. “Your buddy shoots his spear against a rock—ping!—and you notice that all the predators in the neighborhood turn around. He shoots a fish and they all come closer. If the fish gets away with an injury, it will be eaten within a minute,” he says. Desperation has made the predators fearless. “They nip at anything that moves: ears, ponytails, even pencils,” Sandin says.

Such primeval reefs need not be kept in pristine isolation: They could be a test bed for sustainable approaches to fishing, researchers say. “We need to study more pristine reefs to see how much you can fish without reducing the fish stocks,” Sandin says. “That's the key question we need to answer.”