# News this Week

Science  07 May 2010:
Vol. 328, Issue 5979, pp. 674
1. Gulf Oil Spill

# Will Deepwater Horizon Set a New Standard for Catastrophe?

1. Richard Kerr,
2. Eli Kintisch and

The fiery destruction of an oil drilling platform in the Gulf of Mexico on 20 April may have triggered one of the worst environmental disasters in U.S. history. The impact of the crisis, which began with the deaths of 11 workers and then simmered for several days before an expanding oil slick grabbed worldwide attention, promises to test the federal government's ability to protect habitat, wildlife, and the economic well-being of a four-state region on a scale never before imagined.

Secretary Ken Salazar of the Department of the Interior declared on 2 May that the government's role is to “keep its boot on the neck of BP,” the British-based oil giant that had contracted to use the drilling rig, Deepwater Horizon. And BP CEO Tony Hayward said his company “will absolutely be paying for the cleanup operation. There is no doubt about that. It's our responsibility—we accept it fully.”

As Science went to press, severe weather was slowing the spread of the oil to the shore. BP officials were hoping to deploy an oil-topping containment vessel in the coming days. Biologists have their fingers crossed, fretting over a blow at the worst possible time to an already fragile ecosystem. “I don't want to be alarmist, but I'm alarmed,” says ornithologist James Remsen of Louisiana State University (LSU), Baton Rouge. As a precautionary step to ensure the safety of the food supply, the National Oceanic and Atmospheric Administration (NOAA) has imposed a 10-day ban on commercial fishing in the region.

## What went wrong

The reasons the drilling vessel failed to contain the 100-million-pascal pressures in the well it had drilled 5500 meters into the sea floor, at a water depth of 1500 meters, are still unknown. But it probably wasn't because its crew pushed too hard on the envelope of modern technology. Just 8 months earlier, in the Gulf of Mexico, Deepwater Horizon safely drilled the world's deepest well through 9424 meters of rock in 1259 meters of water.

The latest drilling, 65 kilometers off the Louisiana coast, targeted a shallower, well-known oil-bearing stratum. That made for “rather routine” drilling, says petroleum engineer Kenneth Gray of the University of Texas, Austin. By 20 April, workers had finished drilling the hole using the weight of heavy drilling “mud” filling the hole to contain the pressure. A steel pipe or casing was inserted partway down the hole to reinforce it against collapse before cement was pumped between the casing and the wall of the hole.

This “cement job” is often the culprit in oil spills. According to a 2007 report from the U.S. Minerals Management Service, the agency that supervises offshore drilling, cement-related problems caused 18 of 39 blowouts in the Gulf of Mexico over a 14-year period. Something about the design or execution of the cement job could have prevented the cement from setting properly. That flaw could have allowed oil and gas to work its way upward until containment failed entirely.

Estimates of the flow started at zero but quickly rose to 1000 barrels and then 5000 barrels per day. Before drilling began, BP had projected a worst-case scenario of 162,000 barrels escaping per day. Why that hasn't happened is anyone's guess. The answer could lie in constriction in the well, in the blowout preventer sitting at the top of the well, or in the pipe that once connected the well and the drilling vessel now lying crumpled on the sea floor.

The duration of the spill is another open question. The blowout preventer, which failed to cut off the initial surge, has so far refused to respond to remotely operated submersible vehicles. In a week or two, workers will attempt to capture much of what is still leaking out by lowering a 70-ton concrete-and-metal box over the wellhead leak site. The captured oil would then be pumped out the top to waiting barges. No one has ever attempted such a maneuver at 1500-meter depths.

The ultimate solution would involve drilling back into the pressurized rock that's causing all the trouble. The approach requires drilling a new hole from a floating platform, intersecting the original 20-centimeter-wide well kilometers beneath the sea floor, and then pumping in drilling mud to kill the flow. It's an ambitious project that would take at least 3 months.

## Dealing with the oil

The typical southern Louisiana crude rising from the well emulsifies into a foamy “mousse” by the time it reaches the surface, and the more volatile, more toxic components quickly evaporate. In the current spill, “Mother Nature is helping us,” says marine chemist Edward Overton, professor emeritus at LSU Baton Rouge and head of a chemical hazards assessment team under contract to NOAA. “The possibilities are horrible, but the probabilities are less so.”

The one sample Overton and colleagues have been able to analyze “looked like roof tar,” he says, and such goo is expected to eventually form tarballs and tar mats. “We won't see a ‘black tide’ in this spill as we did after Exxon Valdez,” he predicts. The tarballs “are going to be very sticky but not very toxic. … We're looking at an all-summer event.”

Rough seas have stymied the government's ability to skim and collect the oil on the surface, and hindered some floating protective booms from being put into place properly. Several small fires were used to reduce the total amount of oil on the surface. Once the oil slick began moving toward the coast, BP started the aerial spraying of dispersants—detergent-like chemicals designed to break up the oil. Then engineers added dispersants more directly to where the oil came out of the well. The aim is to reduce the risk to seabirds and mammals, and to protect sensitive habitats from being soaked in oil. But the technology has a side effect: By spreading the oil throughout the water column, fish such as spawning bluefin tuna that would otherwise not be exposed are put at risk of toxic effects.

One possible long-term weapon is to use microbes already present in the water. Another technique, adding fertilizer to coastal waters to enhance the ability of natural microbes to metabolize the oil, has shown promise in small field tests, said Ken Lee of the Bedford Institute of Oceanography in Nova Scotia, Canada. A third approach would involve burning oil-covered wetlands in a way that preserves the roots of the plants.

A pressing economic issue for the region is ensuring the safety of shrimp, mussels, and clams that grow in gulf waters. Generally, local oil spills lead the government to shut down nearby fishing or seafood harvesting areas, and they are reopened after the catch is tested and found to be safe. But fishers lack easy screening tools to help them estimate contaminant levels in their catch, and there are no industry-wide techniques to do the tests accurately and quickly. “We don't know what [constitutes] a toxic dose,” says former Food and Drug Administration regulator David Acheson. The challenge, he adds, is measuring long-term rather than acute exposure to oil-tainted seafood.

## The potential impact

The biggest biological impacts of the spill are expected to be on islands that host large colonies of breeding birds and in the rich coastal wetlands, which nourish young fish, shrimp, and shellfish. “The area of the blowout is one of the most productive areas of the Gulf of Mexico,” says Gilbert Rowe, a marine biologist at Texas A&M University, Galveston.

Scientists are especially concerned about the Breton National Wildlife Refuge. The many breeding seabirds on the barrier islands include some 2000 brown pelicans, which only last year had recovered enough to be removed from the federal list of endangered species. If the oil can't be kept off the beaches, an entire generation of pelicans and other species could be wiped out, says LSU's Remsen.

Oil harms seabirds by allowing water to soak through their feathers, chilling the birds and making it harder to gather food. Dolphins are also liable to suffer organ damage from breathing toxic vapors or ingesting too much oil. Sea turtles are at particularly high risk because they feed at the oil-slicked surface.

The wetlands themselves are fragile, too. Having been damaged for decades by dredging, drilling, and other human impacts, Louisiana's wetlands are disappearing at a rate of 4400 hectares per year (Science, 25 November 2005, p. 1264). Oil can harm the vital grasses, which keep the sediment from eroding, by preventing them from photosynthesizing or exchanging gases. Plants cope by sending out new leaves, up to a point. So researchers are holding their breath to see the impact, especially for the most sensitive wetlands farther inland. If the leak isn't plugged, they could be hard-hit by tropical storms blowing oil inland this summer. “We may be talking catastrophic [effects],” says plant ecologist David White of Loyola University in New Orleans.

One obvious need is for additional research to track the impact of the oil on existing study sites. “We're poised to do lots of follow-up studies,” says ecologist Denise Reed of the University of New Orleans. “It's pretty clear that quite a lot of oil is going to get into the wetlands.”

2. Gulf Oil Spill

# Three Historic Blowouts

1. Lauren Schenkman

The decade from 1969 to 1979 witnessed three massive spills from offshore oil wells around the world. Here is how they compare in size and impact.

IXTOC 1 The biggest well-related spill was triggered on 3 June 1979, when a lack of drilling mud allowed oil and gas to shoot up through the 3.6-km-deep IXTOC 1 exploratory well, about 80 km offshore in the southern Gulf of Mexico. The initial daily outflow of 30,000 barrels of oil was eventually reduced to 10,000 barrels. The well was finally capped more than 9 months later. Mexico's state-owned oil company, PEMEX, treated the approximately 3.5-million-barrel spill with dispersants. U.S. officials had a 2-month head start to reduce impacts to the Texas coastline.

Ekofisk The first major spill in the North Sea resulted in the release of 202,000 barrels of oil about 250 km off the coast of Norway. The 22 April 1977 blowout caused oil to gush from an open pipe 20 m above the sea surface. The well was capped after a week. Between 30% and 40% of the spill evaporated almost immediately. Rough waters broke up the slick before it reached shore.

Santa Barbara A blown well 1 km below the sea floor and 9 km off the coast of Santa Barbara, California, spewed out a total of 100,000 barrels of oil. The initial eruption occurred on 28 January 1969, and the well was capped by mud and cement on 7 February, but the pressure forced oil through sea floor fissures until December. The oil contaminated 65 km of coastline. At least 3700 birds are known to have died, and commercial fishing in the area was closed until April.

3. Biomedical Research

# Peering Over a Cliff at the Poststimulus World

1. Jocelyn Kaiser

A year ago, the mood in the U.S. biomedical community was euphoric as researchers scrambled for a piece of the National Institutes of Health's (NIH's) stimulus windfall. More than $8 billion went to extramural research (and$1.8 billion for construction and equipment); thousands of scientists are now toiling away on the studies it funded. But reality is setting in. The 2-year grants will run out in 2011, and when that happens it could cause a nasty shock. Barring a new windfall—and none is in sight—NIH's budget will drop sharply next year. Much of the work recently begun will be left short of cash. The result could be the lowest grant funding rates in NIH history, and the academic job market will suddenly dry up—especially for young researchers.

“There will be a reduction in capacity. How abruptly and exactly when it hits we're still trying to piece together,” says Howard Garrison, public affairs director for the Federation of American Societies for Experimental Biology (FASEB) in Bethesda, Maryland. Garrison issued a gloomy overview of the figures last month.

Last week, in his first budget defense as NIH director before a House of Representatives Appropriations subcommittee, Francis Collins was asked about the poststimulus problem. He showed a graph of NIH's budget over the years with a label at the point where the stimulus money runs out and the total will drop $4 billion in 2011, assuming NIH gets the$32.2 billion President Barack Obama has requested (see graph). (NIH's baseline budget, not counting stimulus money, would increase by 3.2%.)

“This is the cliff that people are talking about, ” Collins said. “We are going to face a crunch” in 2011, he said. The success rate, or portion of reviewed applications that receive funding, which hovered around 30% a few years ago and 20% this year, “will be more like 15%.” The committee chair, Representative David Obey (D–WI), asked whether the NIH windfall was a mistake. No, Collins replied, it has been a “wonderful investment.”

The looming cliff may not look scary to some NIHers because they have weathered booms and busts before. NIH's budget doubled between 1999 and 2003, encouraging academic research institutions to go on a building spree, bringing new researchers into the system, and spurring others to expand their labs with multiple grants. After that, NIH's budget stayed essentially flat, or fell if adjusted for inflation, until 2009–10. Then the stimulus money in the American Recovery and Reinvestment Act (ARRA) threw a lifeline to labs that were in jeopardy; many NIH panels reached below the “payline” and funded grants that had just missed the cutoff for funding from NIH's regular budget.

But while lawmakers are likely to applaud the call for reforming graduate education, the idea of a massive boost in federal support flies in the face of current fiscal realities. The prospect of a $1.3 trillion budget deficit and a freeze next year on most domestic programs may cause policymakers to put this report on their bookshelves rather than in their legislative agendas. “There is a lot of interest in this topic within the Obama Administration,” says Under Secretary Martha Kanter, who oversees postsecondary education at the Department of Education and who spoke at the report's unveiling last week. “But there are always a lot of competing ideas. And most of them don't get to the finish line. That's especially true in a flat budget year, which it looks like 2011 will be. What we don't do as much of is ask, ‘What can we stop doing?’” By coincidence, The Path Forward was released the day after the House of Representatives Science and Technology Committee approved a reauthorization of the 2007 law, the America COMPETES Act. That legislation drew heavily from the academies' Gathering Storm report and provided the justification for Congress to boost funding for research. But the committee's actions on 28 April indicate how even the staunchest supporters of academic research and training have had to curb their enthusiasm. During an 8-hour markup of the bill (H.R. 5116), the committee remained solidly in favor of the Administration's promise to double over 10 years the budgets of the National Science Foundation (NSF), the Department of Energy's (DOE's) research programs, and the National Institute of Standards and Technology. In addition to spelling out operating procedures for DOE's new energy hubs and its Advanced Research Projects Agency—Energy, creating an Office of Innovation within the Department of Commerce, and giving NSF the authority to tackle knotty research challenges with prizes rather than grants, the bill also embraces several new federal initiatives to improve science and math education. At the same time, however, the committee bowed to growing congressional concern about the federal deficit by lopping 10% off authorized agency spending levels spelled out in a version of the bill introduced only a few days earlier by the committee's chair, Representative Bart Gordon (D–TN). The lower levels were proposed by Gordon himself, who had anticipated the push for a lower number by crafting an amendment to his own bill that pared its original cost from$93 billion to $82 billion over 5 years. The amendment passed with bipartisan support. “These [new] levels are lower than I'd like them to be,” Gordon acknowledged at the start of the session. “But I consider them to be practical. … We will maintain a doubling path but on a slightly less steep trajectory.” Authorizing a new doctoral training program would require the science committee to find new sources of revenue or cut into existing programs. Neither is a “winnable argument” under the current fiscal restraints, says one committee aide. Gordon, who is not running for reelection in November, hopes the full House will pass the COMPETES bill before the end of May. Its reauthorization would be a crowning achievement of a 26-year congressional career. But will the Senate go along? Although the bill has many supporters, the Senate's legislative calendar is already jammed. 7. Lab Safety # Radiation Accident a ‘Wake-Up Call’ For India's Scientific Community 1. Pallava Bagla NEW DELHI—The improper disposal of a derelict gamma-ray research device at the University of Delhi has resulted in the death of a scrap-metal worker—and drawn scrutiny of how India's academic institutions handle radioactive materials. India's nuclear watchdog, the Atomic Energy Regulatory Board (AERB) in Mumbai, has labeled the incident as a “serious violation” and slapped India's premier university with an indefinite moratorium on radiation-related activities. That “such an outstanding university was so callous is mind-blowing,” declares former AERB Secretary K. S. Parthasarathy, who says the tragedy “should be a wake-up call.” On 30 April, India education minister Kapil Sibal ordered universities to beef up guidelines for the handling and use of hazardous materials.” India's accident is the latest in a series of radioactive mishaps worldwide. Since 1993, the International Atomic Energy Agency in Vienna has logged several dozen incidents involving the loss or theft of “dangerous” radioactive sources, including cobalt-60. One of the worst cases of cobalt-60 exposure occurred at a scrap yard in Samut Prakarn, Thailand, in February 2000, when the dismantlement of a radiation-therapy unit killed three people and injured 10. The accident in India occurred after plant biotechnologist Deepak Pental, the University of Delhi's vice chancellor, ordered a campuswide spring cleaning to create space for newly recruited staff. On 26 February, the university auctioned to a scrap dealer a Gammacell 220 research irradiator, which university chemist B. K. Sharma had imported from Atomic Energy of Canada Ltd. in 1968 but which had lain unused since Sharma's retirement in 1985. At the scrap yard, unaware that the machine they were dismantling contained cobalt-60, seven workers fell ill; two are in critical condition. One died—India's first known fatality from a radiation accident. Some of the cobalt-60, a gray-blue metal resembling nickel, is unaccounted for. The incident has raised questions about the University of Delhi's competence in radiation safety. An auction committee comprised of top university chemists had determined that radiation levels from the Gammacell's cobalt-60 would be “manageable,” Pental says, as the isotope has a half-life of 5.27 years. (In fact, the material in such a device can release lethal radiation for decades.) AERB's chair, nuclear scientist S. S. Bajaj, told Science that he was “shocked” by the Delhi researchers' “ignorance.” Bajaj acknowledges that the fate of the Gammacell device was not being tracked by AERB and should never have “disappeared from [AERB's] radar,” but he says the university was obliged to inform AERB before disposing of the device. An agency team would have helped ensure that this was done safely. AERB has launched an audit of the roughly 140 Gammacells in India's academic institutions. Pental says he takes “moral responsibility” for any negligence on the part of the University of Delhi. AERB and Delhi police are now conducting an investigation; anyone found guilty of unauthorized disposal of a radioactive source can be jailed for up to 5 years. “Zero tolerance,” says Parthasarathy, “is the only way forward.” 8. ScienceInsider # From the Science Policy Blog The nation's first cancer vaccine will soon hit the market. The U.S. Food and Drug Administration has approved for sale Provenge, meant to treat prostate cancer that has spread throughout the body. Patients given the vaccine lived about months longer, nearly 26 months total, than controls. A course of treatment costs$93,000.

The National Institutes of Health has approved four stem cell lines submitted by WiCell, the nonprofit associated with the University of Wisconsin, Madison. The lines were approved under the Bush-era stem cell policy, but they had to undergo new scrutiny to make sure they meet stiff ethics rules.

The plight of postdocs, and the glacial pace of negotiations between a union that represents them in California and the University of California (UC), was the subject fairly acrimonious congressional hearing. About 6000 postdocs—10% of all U.S. postdocs—work at UC. The union is seeking pay structure modeled on federal guidelines.

A new handbook offers hints for universities trying to increase diversity while staying within the law. The joint effort of AAAS (which publishes Science) and the Association of American Universities includes tips, examples of what works, and legal analyses.

Drug company scientists released news of high-profile clinical trial of a drug to treat fragile X syndrome by divulging the results to a reporter at The New York Times. Researchers feel “pretty good about the unpublished] data,” said a scientist with Novartis.

Are genomewide association studies useful? A provocative paper in Cell says no, arguing that by linking common gene variants increased disease risks, they fail to uncover helpful connections. A better approach, says the much-discussed paper, may be to look rare variants.

See the full postings and more at news.sciencemag.org/scienceinsider.

9. Paleogenetics

# Close Encounters of the Prehistoric Kind

1. Ann Gibbons

The long-awaited sequence of the Neandertal genome suggests that modern humans and Neandertals interbred tens of thousands of years ago, perhaps in the Middle East.

It's the mystery of Mount Carmel. On this limestone ridge overlooking the coast of Israel, modern humans lived in caves off and on for tens of thousands of years, starting more than 100,000 years ago. Then, perhaps as early as 80,000 years ago, members of another species reached and occupied the caves: heavy-bodied Neandertals, who were escaping a cold spell in Europe and moving south into the Middle East. Did the two species meet here? Did they mate?

The archaeological record in the caves is ambiguous on that question, and anthropologists have fought bitterly over it. Some claim that the anatomy of fossils shows that Neandertals, our closest cousins, did mate with modern humans, either in the Middle East or in Europe. But others thought modern humans coming out of Africa completely replaced Neandertals with little or no interbreeding. And the genetic evidence from ancient bones showed no sign that Neandertals had swapped genes with our ancestors—until now.

On page 710, an international team of researchers presents their first detailed analysis of the draft sequence of the Neandertal genome, which now includes more than 3 billion nucleotides collected from the bones of three female Neandertals who lived in Croatia more than 38,000 years ago. By comparing this composite Neandertal genome with the complete genomes of five living humans from different parts of the world, the researchers found that both Europeans and Asians share 1% to 4% of their nuclear DNA with Neandertals. But Africans do not. This suggests that early modern humans interbred with Neandertals after moderns left Africa, but before they spread into Asia and Europe. The evidence showing interbreeding is “incontrovertible,” says paleoanthropologist John Hawks of the University of Wisconsin, Madison, who was not involved in the work. “There's no other way you can explain this.”

As a result, many people living outside Africa have inherited a small but significant amount of DNA from these extinct humans. “In a sense, the Neandertals are then not altogether extinct,” says lead author Svante Pääbo, a paleogeneticist at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, who was surprised to find he was part Neandertal. “They live on in some of us.”

The team also used the Neandertal DNA like a probe to find the genes that make us modern. Even though the genomes of humans and Neandertals are 99.84% identical, the researchers identified regions that have changed or evolved since our ancestors and Neandertals diverged sometime between 270,000 and 440,000 years ago—their new, slightly younger estimate of the split. So far, the team has detected tantalizing differences in genes involved in metabolism, skin, the skeleton, and the development of cognition, although no one knows yet how these genetic changes affect physiology. “This is a groundbreaking study!” enthuses evolutionary geneticist Hendrik Poinar of McMaster University in Hamilton, Canada. “We can actually discuss an extinct human species—Neandertals—on a genetic level rather than strictly on morphological grounds.”

## Mixed marriage

The discovery of interbreeding in the nuclear genome surprised the team members. Neandertals did coexist with modern humans in Europe from 30,000 to 45,000 years ago, and perhaps in the Middle East as early as 80,000 years ago (see map, p. 681). But there was no sign of admixture in the complete Neandertal mitochondrial (mtDNA) genome or in earlier studies of other gene lineages (Science, 13 February 2009, p. 866). And many researchers had decided that there was no interbreeding that led to viable offspring. “We started with a very strong bias against mixture,” says co-author David Reich of Harvard Medical School in Boston. Indeed, when Pääbo first learned that the Neandertal DNA tended to be more similar to European DNA than to African DNA, he thought, “Ah, it's probably just a statistical fluke.” When the link persisted, he thought it was a bias in the data. So the researchers used different methods in different labs to confirm the result. “I feel confident now because three different ways of analyzing the data all come to this conclusion of admixture,” says Pääbo.

The finding of interbreeding refutes the narrowest form of a long-standing model that predicts that all living humans can trace their ancestry back to a small African population that expanded and completely replaced archaic human species without any interbreeding. “It's not a pure Out-of-Africa replacement model—2% interbreeding is not trivial,” says paleoanthropologist Chris Stringer of the Natural History Museum in London, one of the chief architects of a similar model. But it's not wholesale mixing, either: “This isn't like trading wives from cave to cave; the amount of admixture is tiny,” says molecular anthropologist Todd Disotell of New York University in New York City. “It's replacement with leakage.”

Although the 1.3-fold coverage of the Neandertal genome is a remarkable technical feat, one-third of the genome is still murky. In a separate paper (p. 723), the team describes and successfully tests a new method for filling in gaps in the rough draft of the genome.

The team also used three methods to nail down the interbreeding result. First, they compiled the Neandertal genome using DNA from the limb bones of three female Neandertals who lived in Vindija Cave in Croatia from 38,000 to 44,000 years ago; they confirmed parts of the genome with much smaller amounts of DNA from Neandertals who lived in Spain, Germany, and Russia.

Once they were satisfied that the composite genome was a fair representation of Neandertals from across a great part of their geographical range, researchers compared the Neandertal genome to a chimpanzee's to determine which genetic variants were primitive, ancestral forms. Then they compared the new, derived genetic variants in Neandertals to those in the complete genomes of five living humans, including a San from Southern Africa, a Yoruba from West Africa, a Papua New Guinean, one Han Chinese, and one French European.

The team measured the genetic proximity of Neandertals to pairs of modern humans from different continents, first using single-nucleotide polymorphisms (SNPs), or sites in the genome where a single nucleotide differs between individuals. When they compared a Neandertal with a European and an Asian, they found that the Neandertal always shared the same amount of derived (or more recently evolved) SNPs with each of them. But when they compared a Neandertal with an African and a European, or with an African and an Asian, the Neandertal always shared more SNPs with the European or Asian than with the African. “We've shown that Neandertals are significantly more closely related to non-Africans than Africans on average,” says Reich.

Even though they looked at just two Africans for this part of the study, those two have a particularly ancient, diverse heritage, so they are a good proxy for much of the genetic diversity in Africa. But sequencing additional Africans would be a good idea, says Reich.

For now, it seems Neandertals interbred with the ancestors of Europeans and Asians, but not with the ancestors of Africans. At first, “we were baffled that this affinity with Neandertals was not only in Europe and West Asia [where it was most expected], but also in Papua New Guinea” where Neandertals never set foot, says Pääbo.

To be certain, they used two other methods to detect gene flow between Neandertals and Eurasians. Using the published genome of an African American from the Human Genome Project, they compared large regions of African and European ancestry in this single genome to Neandertal regions. In this person's genome, the European and Neandertal segments were more similar to each other than either was to the African segments.

Finally, population geneticist Rasmus Nielsen of the University of California (UC), Berkeley, scanned the human genome for “ancient” genomic segments—those that might predate the time when modern humans arose, about 200,000 years ago. Before receiving the Neandertal DNA sequences, he identified 13 genomic regions that were unusually variable, and therefore likely to be evolutionarily ancient, in 48 people outside of Africa. He identified 13 “old” variants as possibly coming from Neandertals or other archaic ancestors, because they were missing from the genomes of 23 African Americans (used as proxies for Africans). Then the team looked in the Neandertal genome—and found 10 of the 13 ancient variants. “There are places in the genome where we can say this section is really, really likely to be from a Neandertal,” says Reich.

When and where did modern humans pick up those Neandertal genes? The most likely scenario “was the movement of a few Neandertals into a group of moderns,” says co-author and population geneticist Montgomery Slatkin of UC Berkeley. If a few Neandertals interbred with members of a small population of modern humans, Neandertal gene variants might persist in subsequent generations of modern humans if the interbred population expanded rapidly, thereby spreading Neandertal DNA widely.

This scenario apparently fits with fossils and stone tool data from the Israeli caves such as Skhul, Qafzeh, and Tabun, where Neandertals show up in the region as early as 80,000 years ago, when moderns were already there. Although each group may have occupied the caves intermittently, some say they may have overlapped for up to 10,000 years. Neandertals and moderns apparently even occupied the same cave, Tabun, at different times. The two species had much in common: Both lived in caves, used similar toolkits (although Neandertals may have made better spear points), and hunted the same fallow deer and gazelles. “It doesn't surprise me,” says archaeologist Ofer Bar-Yosef of Harvard University about the ancient DNA finding. “We always predicted low-level mixing,” because some Neandertals in the Middle East, such as a female skeleton at Tabun, look less robust than Neandertals in Asia and Europe. Mixing in this region could also have happened later, when another group of modern humans came out of Africa about 60,000 years ago and perhaps met Neandertals, who were still occupying caves in the Middle East until 50,000 years ago, says Stringer.

Finally, the researchers cannot rule out the possibility that what they see as “Neandertal” motifs are really ancient genetic variants that Neandertals and some modern humans inherited from a common ancestor they shared before Neandertals split off. Although all early modern populations, including in Africa, interbred, that gene flow was not complete enough to pass these Neandertal motifs to all Africans. Human populations that were more closely related to the ancestors of Neandertals carry those motifs while Africans do not, says Reich.

To date, the genomic data don't support interbreeding in the time and place when everyone most expected it: between 45,000 and about 30,000 years ago in Europe. Neandertals and moderns lived in such proximity in France, for example, that some researchers think Neandertals imitated modern stone-tool and beadmaking technologies. But such late European mixing cannot explain the current findings, in which Asians and Europeans are equally similar to Neandertals. It's still possible that Neandertals and modern humans in Europe inter-bred rarely and that the Neandertal genes were swamped out in a large population of modern humans, says Slatkin.

In some ways, it is surprising that there isn't more evidence of interbreeding, now that researchers know it was biologically possible. “For some reason, they didn't interbreed a lot—something was preventing them,” says evolutionary geneticist Sarah Tishkoff of the University of Pennsylvania. “Was it a cultural barrier?”

## Modern motifs

The Neandertal genome also gives researchers a powerful new tool to fish for genes that have evolved recently in our lineage, after we split from Neandertals. The team compared the Neandertal genome with the genomes of five diverse modern humans. They found 78 new nucleotide substitutions that change the protein-coding capacity of genes and that are present in most humans today; just five genes had more than one such substitution. That's a tiny fraction of the 3 billion bases in each genome. “Only 78 substitutions in the last 300,000 years!” says Poinar. “The fact that so few changes have become fixed on the human lineage is amazing.”

But the mutations they've found so far “are all very interesting, precisely because there are so few,” says Pääbo, whose team is trying to identify their function. The catalog includes changes in genes that encode proteins important for wound healing, the beating of sperm flagellum, and gene transcription (see table, above). Several of these newly evolved modern human genes encode proteins expressed in the skin, sweat glands, and inner sheaths of hair roots, as well as skin pigmentation. “The fact that three of six genes carrying multiple substitutions are in skin is fascinating,” says Poinar. Pääbo speculates that these changes “reflect that skin physiology has changed but how, of course, we don't know yet.”

Some of those changes are likely to be neutral changes that accumulated through genetic drift, but the team also used the Neandertal data to find other evolutionary changes that were beneficial to modern humans and so rose to high frequencies in some populations. Specifically, they have identified 15 regions containing between one and 12 genes. The widest region is located on chromosome 2 and contains the gene THADA, a region that varies in modern humans and that has been associated with type 2 diabetes. Changes in this gene may have affected energy metabolism in modern humans.

Other mutations appear to be in genes important in cognitive development and that, when mutated in living people, contribute to diseases such as Down syndrome, schizophrenia, and autism. One gene, RUNX2, is associated with a disease that leads a spectrum of developmental abnormalities, including misshapen clavicles and a bell-shaped rib cage. Suggestively, Neandertals had bell-shaped rib cages and possibly peculiar clavicles. But precisely how all these genetic differences are expressed physiologically is the next frontier. “We need to follow up. Are there regions that are functionally significant?” says Tishkoff. By 7 May, the Neandertal data should be posted on Ensembl and the UC Santa Cruz browser, so other teams can do just that, says Pääbo.

His own group is already working on such functional studies. Postdoctoral researcher Matthias Gralle is analyzing the way these recently evolved genetic differences change the way proteins are expressed. Such studies may eventually offer clues about why Neandertals went extinct—and our ancestors didn't. “The mystery isn't just why they disappeared,” says paleoanthropologist Jean-Jacques Hublin of the Max Planck Institute for Evolutionary Anthropology. “It is why we were so successful that we replaced all the others.” For now, researchers are delighted that this “groundbreaking” genomic work has made it possible to ask such interesting questions, says Poinar. “This is the real appeal of this project: What will the genome of the Neandertal tell us about functional differences between the two [species],” says Poinar.

10. Paleogenetics

# Cloned Neandertals Still in the Realm of Sci-Fi

1. Elizabeth Pennisi

Since the idea of sequencing the Neandertal genome became more than a glimmer in a paleogeneticist's eye, some have asked, "Could we, should we, would we, bring this extinct human species back to life?" But for both technical and ethical reasons, experts say, bringing back a Neandertal is a pipe dream.

Science-fiction writers have been resurrecting Neandertals in novels for decades, imagining what it would be like to see and communicate (not to mention mate) with another species of human. So once the idea of sequencing the Neandertal genome became more than a glimmer in a paleogeneticist's eye, some have asked, “Could we, should we, would we, bring this extinct human species back to life?” After all, biologists are trying to bring back the woolly mammoth by cloning. But for both technical and ethical reasons, experts say, bringing back a Neandertal is a pipe dream.

Could we do it? Robert Lanza laughed at the thought. Chief scientific officer for Advanced Cell Technology in Worcester, Massachusetts, he and his colleagues have cloned species from cows to goats to mice and extended their efforts to include endangered species and human embryos. But cloning Neandertals is fantasy, says Lanza. “You can't clone from stone, and you can't clone from DNA that has been destroyed from weather and the elements,” he points out.

The Neandertal genome sequence reported on page 710 (and see main text, p. 680) reflects the battered state of the starting DNA, which came from bones that are 38,000 to 44,000 years old. Because the isolated DNA was in pieces typically about 50 bases long, there are many missing stretches, particularly repetitive regions. “We will never have a finished sequence for the Neandertal the way we have for a mouse,” says Svante Pääbo, who led the Neandertal sequencing project at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany. Jurassic Park aside, reconstructing an organism with a partial genome would be like constructing a building with a partial blueprint.

Even if scientists had the complete genome, it wouldn't be enough. DNA itself doesn't tell the whole story. Chemical modifications to the genome, the way chromosomes arrange in the nucleus, and maternal components in the egg all play a role in translating a genetic blueprint into a viable individual. “It's not just the DNA; there's a lot else going on,” says Lanza. None of that information is even available for Neandertals.

Then, too, cloning doesn't typically start with a genome; it starts with two cells. One cell provides a nucleus (with DNA inside), and one is an egg cell, most often of the same species, whose DNA has been removed. The nucleus is then transferred to the egg, sometimes by fusing the two cells. “If you have just got DNA, you are asking an enormous amount of the oocyte that you are going to put the DNA into,” explains Ian Wilmut, who cloned Dolly the sheep and now works at the University of Edinburgh in the United Kingdom. “It has to reform the nucleus and reprogram [the DNA].”

That leads to the next problem: What species' egg would play host to this DNA? The obvious candidate would be a modern human egg, but they are notoriously fickle and don't take well to nuclear transfer, even of modern human DNA. “There's something different about primates that we haven't identified,” says Wilmut. “[Cloning] works very poorly.” And incompatibilities between Neandertal DNA and the human egg might further diminish the chances of a viable embryo.

Molecular geneticist George Church of Harvard University has proposed another approach: modify the DNA in a human cell line to resemble the Neandertal. “This is a daunting task, but with future technological developments and enough time and money, it may be possible,” says Adrian Briggs, who worked on the Neandertal genome sequence and is about to join Church's lab. In theory, one could convert a human or chimp genome to a Neandertal genome—base by base—while it is still nicely nestled in a stem cell, then clone it. But there's on the order of a million differences between the Neandertal and human genomes, and the more changes needed, the greater the risk of introducing errors.

If, somehow, a viable embryo were produced, this developing chimera would need a surrogate mother. What species would that mother belong to? Again, the obvious choice is a human, but no one knows whether a modern woman's biochemistry would be compatible with that of a Neandertal fetus. And is it ethical for a human surrogate mother to birth a Neandertal baby? Church thinks ethical views will evolve as technology improves. Once cloning works well in a variety of animals and stem cell–derived organs become commonplace, “I think the resistance to it will disappear,” he says.

But others disagree. “We do not—and should not—create human beings just to satisfy our scientific curiosity,” says Pääbo, pointing out that Neandertals are a species of human, so cloning them raises many of the same ethical issues as cloning a modern human.

Cloning Neandertals would involve several “ethically deplorable steps,” says Briggs, including using surrogate mothers and risking high failure rates, abnormal births, and, sometimes, early death of clones. With a Neandertal, “all of these safety issues would apply, only writ large,” says Wilmut. And how would a Neandertal fit into modern human society? “I see no palatable conditions,” says Pääbo. “Not even for medical purposes are we thinking about creating a [modern] human being. Why would we consider something like this, which is much less pressing?”

11. Paleogenetics

# Computer Kid Makes Good

1. Elizabeth Pennisi

Everyone involved in the Neandertal sequencing project at the Max Planck Institute for Evolutionary Anthropology stresses the teamwork involved in sequencing the Neandertal genome. And yet Richard "Ed" Green, the postdoctoral fellow in charge, still stands out.

Late 2007 was a real low point for Richard “Ed” Green and colleagues at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany. A year earlier, in Nature, they had predicted that they could sequence the Neandertal genome using 20 grams of bone and 6000 runs using “next generation” sequencing technologies. They knew going in that most of the DNA in fossil bone is bacterial, with only a small percentage of Neandertal DNA. But it turned out that the bones to be sequenced had far less Neandertal DNA than the sample on which they based their projections. “We were in kind of an awkward situation of having announced to the world we were going to do it, and we were left with no concrete plan of how to do it,” Green recalls. “That was very scary.” Their fears increased when they discovered that their first million bases of Neandertal sequence were contaminated with modern human DNA.

“But we worked it out along the way,” says Green, the postdoctoral fellow in charge of the project. He and colleagues developed methods to control contamination by putting bar codes on all DNA coming from the fossils (Science, 13 February 2009, p. 866). They cut down on the amount of DNA to be deciphered by cutting up much of the bacterial DNA so that the sequencing reactions ignored it. Everyone, especially Green, stresses the team effort involved. “Many people here have been able to say they ‘saved the Neandertal genome project,’” he notes. And yet Green, 37, still stands out.

“Ed brought the quantitative and algorithmic horsepower needed to interpret the Neandertal data,” says David Haussler of the University of California (UC), Santa Cruz, where Green now works as an assistant professor. “He invented new analysis methods that allowed the Neandertal project to happen.”

That computational horsepower is what landed Green the job of shepherding the Neandertal genome. After getting a degree in computational biology from UC Berkeley, he joined Svante Pääbo's lab at the Max Planck institute in 2005 to explore the evolution of genes that can code for more than one protein. Pääbo and the sequencing company 454 Life Sciences in Branford, Connecticut, had just sequenced cave bear and mammoth DNA and were puzzling over the results: There was so much microbial sequence, it was hard to detect mammalian DNA. Green knew what to do: He enlisted a cluster of computers to compare the DNA with that of known sequences, including dog and elephant, so he could discard the microbial sequence and focus on the tiny bit of mammalian DNA.

“This was really the first large-scale snapshot of what the universe of [ancient] DNA looked like when it came out of a bone,” Green recalls. “Then Svante said, ‘Let's try Neandertal.’ It was obvious that this was a once-in-a-lifetime opportunity.”

He took charge of the bioinformatics effort, writing software to better detect Neandertal DNA and to deal with degradation. “He is able to design ways to analyze a whole genome under circumstances that are nonstandard,” says Pääbo.

Green also coordinated the design and logistics of the rest of the project, which involved about 50 people. He was “very patient in terms of helping and training others,” says Pääbo. Former graduate student Adrian Briggs agrees: “Without Ed's enthusiasm and competence, the project would never have proceeded so fast.” The job required long hours at the lab, but Green says he didn't mind because the Max Planck facilities were “maximally comfortable,” complete with Ping-Pong table, sauna, barbecue grill, and even a resting room.

Switching gears was not new to Green, who had started off in developmental biology as an undergrad and studied cancer biology in grad school before moving to computational biology. Now that he's settled in at UC Santa Cruz, Green expects to switch gears again. He wants to look at gene expression in nonmodel organisms while continuing to work with Pääbo on Neandertal DNA. “Ed is an incredibly skilled bioinformatician,” says Pääbo. “It would be great if we could continue to work together.”

12. Profile: Frédérique Darragon

# Unraveling a Riddle in Plain Sight

1. Richard Stone

Amateur archaeologist Frédérique Darragon has spent 12 years documenting hundreds of mysterious towers in Southwest China—and winning over skeptical academics.

CHENGZI, CHINA—The jeep grinds over a rise on a rutted dirt road in the foothills of the Himalayan Mountains. “Stop here!” exclaims Martine Francoise Darragon. The svelte socialite-turned-explorer leaps out and holds up an old photo showing snowcapped peaks towering over gentle valley slopes, with a rock-strewn river in the foreground. The black-and-white landscape of western Sichuan Province, captured in the 1930s by the intrepid botanist Joseph Rock, matches the view from where Darragon, who goes by the name Frédérique, is standing. “This is where he took the picture,” she says in her French–New York City accent. But something from the photo is missing from the land today: two eye-catching stone towers whose beveled walls, viewed from above, would form eight-pointed stars.

In the 70 years since Rock's travels through the Tribal Corridor of Tibet and western Sichuan, a small village inhabited by Minyag people has sprung up here on the banks of the Chengzi River, some 3750 meters above sea level. Defying the thin air, Darragon, 60, bounds across a stone bridge over the Chengzi and makes a beeline to where the nearest tower in the postcard once stood. All that's left now of a structure that had been at least 25 meters tall is a dilapidated first-floor section: interior stonework and fill, and a doorway framing blue sky. Over more than a decade, Darragon has identified nearly 1000 such ancient structures in Sichuan and neighboring Tibet, from total wrecks like this one to largely intact towers exceeding 50 meters in height. Many more have been lost through the ages.

Why these Himalayan towers were built is an enduring mystery. Other structures—squat, square towers erected against northern invaders—once were widespread and are of scant scientific interest. But the much taller star-shaped towers and other arresting buildings suggest that the medieval kingdoms of these lands were more ingenious and sophisticated than many scholars have presumed. Clusters of towers on mountain slopes may have been status symbols in a game of one-upmanship among wealthy merchants, Darragon says, while lone towers in river valleys likely served as lookouts or way stations on the southwestern Silk Road. Some towers may have held religious meaning. And a star-shaped design may help withstand shaking in a region prone to strong earthquakes.

Over the past decade, Darragon has had fragments of wood beams from several dozen towers radiocarbon-dated, yielding approximate ages ranging from 300 to 1700 years old. Most presumably were built during this period, although any single tower's age is hard to pin down: Some beams could have been replaced after a tower was built, yielding a more recent carbon-14 date, while others could have been built using beams from older trees that predated the towers. Darragon may be an amateur, but her sleuthing and derring-do have earned the respect of Chinese scientists. “Some experts did not know what to think of her at first,” says Zhong Xiao-Hou, director of the National Architecture Institute of China in Beijing. “But we have come to admire her spirit and enthusiasm for our heritage.”

Drawing on Darragon's work, the State Administration of Cultural Heritage of China (SACH) is expected to soon nominate dozens of the more imposing structures to UNESCO's World Heritage list as the Diaolou Buildings and Villages of Tibetan and Qiang Ethnic Groups Cultural Landscapes. The towers “represent an extraordinary heritage and tradition, and deserve to be fully preserved,” says Francesco Bandarin, assistant director-general for culture at UNESCO.

But a UNESCO listing may not come in time for a clutch of towers near Danba, in western Sichuan, that are imperiled by the construction of a hydropower dam. The rGyalrong towers have become Darragon's latest cause célèbre.

## Unlikely savior

Darragon's wanderings in the highlands of Sichuan and Tibet, a forbidding swath of land she has come to know better than most Westerners—and Chinese—are a far cry from the racy days of her youth. As a teenager from a wealthy Parisian family, Darragon spent summers riding horses in England and winter breaks skiing in the Swiss Alps. All the while, she nurtured a counterculture side: “I was a communist at heart,” says Darragon, who worked one summer on a kibbutz in Israel.

She was also a playgirl at heart. When Darragon was 18, she inherited a small fortune from her father, an inventor and machinemaker. In early 1971, she sailed across the Atlantic as a bikini-clad deckhand in the first Cape Town–to–Rio de Janeiro race, then from Brazil to the West Indies before returning to Paris just in time to take final exams and graduate from the Universite Paris X de Sciences Economiques.

Darragon spent a few years managing her real estate and modeling. Then in the summer of 1978, she took up polo and was the first woman to play at the Bagatelle Polo Club in Paris before moving to Buenos Aires and becoming a record-setting player in Argentina. If those pursuits were not enough, Darragon has also raced as a jockey on thoroughbred horses, won renown as a samba dancer in Rio, and has had a lifelong passion for oil painting.

Along the way, Darragon collected prominent boyfriends and near-death experiences. The former include cable TV mogul Ted Turner, whom she has known since 1969 and lived with from 2000 to 2003. The latter include wiping out during a high-speed motorcycle chase in which she smashed her Suzuki into a car, flew over the hood, and landed 10 meters away without a scratch, and taking a polo ball in the mouth that crushed her jaw and knocked out several teeth. (She wrapped a scarf around her head and finished the game.) “At times I do regret my happy-go-lucky life,” says Darragon, who has no children and never married. But then she found a higher purpose in China.

Darragon visited China for the first time in the early 1990s and afterward spent months each year backpacking across the country. In 1993, her interest in the endangered snow leopard brought her to Tibet, where she would have her most serious brush with death. On a solo trek in the Himalayas near the border with Bhutan in 1996, Darragon decided to shelter overnight in a tiny cave. It was freezing, so she lit a fire inside. “Terrible idea,” says Darragon, whose Chinese name, Bing Yan, means “ice flame.” The fire sucked the scant oxygen from the thin air and, Darragon says, “I felt a snap inside my head.” She had suffered a stroke and couldn't control her left side, but managed to drag herself outside the cave before blacking out. Three days after she recovered consciousness, Tibetan shepherds came across her and carried her to a village, where she caught a ride to Lhasa, Tibet's capital. Enfeebled, Darragon holed up in the Holiday Inn—at the time, the only hotel in Lhasa with room service. “I could only crawl and didn't want my mother to see me in such a state,” she jokes. Four months later, she felt well enough to fly back to France.

Darragon was soon back in Tibet and venturing where few foreigners had ever gone, such as the remote valleys of Gongbu Jiangda, which once belonged to the ancient Nyangpo Kingdom. It was in places like that, off the beaten track, where Darragon encountered the stone towers, including ones with the astonishing star-shaped walls. Locals did not know who had built the towers, how old they were, or why they were built.

Archaeologists knew about some of the more accessible towers, and Taoping, a Qiang village with several towers a few hours from Chengdu, Sichuan's capital, has long been a tourist draw. Darragon has spent a total of 5 years roaming the hinterlands of Sichuan and Tibet, analyzing more than 250 standing towers, including a few dozen star-shaped ones and 750 or so other ruins. “The sheer amount of data she has collected about these architectural curiosities will be welcomed by a variety of specialists,” says John Vincent Bellezza, senior research fellow at the Tibet Center of the University of Virginia in Charlottesville.

Under the auspices of the Unicorn Foundation, a U.S. nonprofit that Darragon founded in 2001 with seed money from Turner, wood samples from 77 towers—54 in Sichuan and 23 in Tibet—have been radiocarbon-dated by Beta Analytic in Miami, Florida. With Darragon's help, over the past 6 years Achim Bräuning, a dendrochronologist at the University of Erlangen-Nürnberg in Germany, has been building a tree-ring database from the region. He has radiocarbon-dated wood from 16 towers; preliminary results, he says, corroborate Darragon's data. The earliest date—318 C.E., with a margin of error of 40 years—is from a star-shaped tower in Nyangpo. “It's probably the oldest one still standing in the world,” Darragon says.

## Search for meaning

Solving the riddle of the towers is a daunting challenge. When Darragon queried locals, she often just got shrugs. Sometimes it was lack of knowledge; sometimes it was a communication barrier. The region is a Tower of Babel of mutually unintelligible languages, and Mandarin Chinese won't get you far. Chinese annals from the Han Dynasty, which lasted from 206 B.C.E. to 220 C.E., refer to tall towers, according to Chen Zongxiang, a retired historian in Chengdu. Some old towers along the Min River in Sichuan are spaced several kilometers apart on a clear line of sight and must have served as watchtowers or beacons, says Yasuhiko Nagano, an expert on the Tibetan rGyalrong languages at the National Museum of Ethnology in Osaka, Japan. Untold numbers of smaller towers were built during the Jinchuan Wars of the 18th century, when the Manchurian emperor Qian Long sought to pacify the region. But many of the star-shaped towers lack classic features—arrow slits, for example—of defensive fortifications.

Scholars believe that the towers served various purposes. A majestic, hill-hugging assemblage in Danba may have arisen as merchants vied to outdo each other with taller and taller structures, similar to the origin of the San Gimignano towers in Italy, says Darragon. Other towers may have symbolized the dmu cord that in Tibetan lore connects heaven and Earth. In rGyalrong lands, now part of western Sichuan, “it appears that some towers were built to propitiate the deities” by ceremonially projecting ritual participants into a sky realm, says Bellezza.

One idea of Darragon's gaining support is that towers were way stations along the southern Silk Road, which passed through Tibet and Sichuan. “The more fabulous towers all lie along the trade routes,” she says. Many are located in strategic spots in river valleys along routes traveled by medieval caravans. According to Zhong, these towers “may have embodied the economic strength of each village.” Major commodities in the region then were silk, salt, tea, and musk of the forest musk deer. “Beyond the fact that we know that this area was exporting valuable musk in medieval times, we don't know very much about the mechanics of the musk trade,” says Anya King, a historian at the University of Southern Indiana in Evansville. Chinese scholars have postulated a “Musk Road” linking with the Silk Road in the Ngari region of western Tibet. “It is entirely possible that the towers served as regional depots of the trade routes, but at this point it won't be easy to prove,” King says. To test the idea, Darragon is attempting to organize a dig to sample soil for silk, tea, and musk remains at one undisturbed medieval tower.

The star-shaped construction, with its distinctive five to 13 points, is found almost nowhere else. Only a handful of such towers exist outside the region, in India, Iran, Tajikistan, and Afghanistan—including the Bahram Shah Minaret in Ghazni. The shape may help resist earthquakes. “That's what local people tell us,” says Li Chunxia, an anthropologist at Sichuan University in Chengdu who has collaborated with Darragon.

Earthquakes are not the main threat, however. Impoverished villagers have blown up or dismantled towers for building materials, Li says. An especially grievous loss was two ancient Nyangpo towers destroyed in 2006 and 2008, Darragon says. She has come across three villages named Bajiaodiao, or “eight-angle fortress,” only to find nothing but scattered stones. “The remaining towers desperately need protection,” she says. Tourism could help. The Unicorn Foundation has assisted some communities to restore towers and convert farmhouses into guesthouses. “If rightly harnessed, some of the proceeds could go to scientific pursuits in the region,” says Bellezza.

Another big boost for the preservation effort, Darragon says, would be to include the towers on UNESCO's World Heritage List. Over the years, Darragon and Unicorn have submitted maps, radiocarbon data, and oral history to SACH in support of a nomination. “She is so diligent. Even though she is not a professional, her work is hard to refute,” Li says. “Frédérique is a force,” adds Bandarin. “She has identified the value of this heritage, she has promoted the necessary research and scientific investigation, and she has attracted the interest of the local and national governments and of UNESCO,” he says. SACH is expected to decide soon on whether to nominate several groups of towers as Cultural Landscape sites. “The cultural self-esteem of about 50,000 minority people living in these regions is riding on it,” Darragon says.

If the nomination comes through, Darragon says she can rest easy—and move on. “I've never spent such a long time on any one thing in my life,” she says. But Darragon has set herself one last tower-related task: saving several rGyalrong towers near Danba, which have special significance to her. She glimpsed towers in China for the first time on a trip to Danba, when in 1997, she recalls, “in the pouring rain I caught sight of tall structures clinging to faraway mountain slopes.”

Several of these towers, including the tallest in Sichuan that is about 700 years old, are in jeopardy. “They will be under water,” Darragon says, if a second dam planned for the area is built. Chinese colleagues and other experts have joined with her to lobby authorities to revise the hydropower plans.

A recent trip to Danba allowed Darragon to reinforce that message. In Badi village high in the Danba hills, three young girls dash out of their home, smiling and giggling, as Darragon's jeep pulls up. She and her entourage are invited into the rGyalrong home for bowls of homemade cottage cheese and butter tea. They've known Darragon for years and treat her like family.

After a chat, Darragon takes her leave and gets back to work. As the afternoon shadows lengthen, she climbs into a ruined section of a fortress with an attached tower and uses a penknife to gouge a chunk from a wooden beam that she'll send for radiocarbon analysis. “That's it for today,” she says. Tomorrow she has more survey work and a meeting with Danba officials. “I will do my best to persuade them that a World Heritage nomination would benefit Danba more than a second dam would,” she says. Given her tenacity, it would be a wonder if her latest quest were to fail.