# News this Week

Science  17 Aug 2001:
Vol. 293, Issue 5533, pp. 1234
1. BIOMEDICAL POLICY

# NIH Wins an Exemption From HHS Peer-Review Overhaul

1. Eliot Marshall

The peer-review system at the National Institutes of Health (NIH) rests on a simple notion: Fly in a panel of experts and let them decide, face to face, on the best science. It's a trusted system and so widely admired by most scientists that any proposal for improving it can easily be perceived as a threat.

That's what happened this month, after NIH officials received a leaked memo on 1 August from their political bosses at the Department of Health and Human Services (HHS). The memo appeared to call for cheaper ways to obtain expert advice. NIH officials reacted immediately, and within days, an HHS manager was trying to calm the waters, explaining that NIH's scientific programs were not a candidate for such reforms. The memo might have gone unnoticed had it not arrived on the heels of two other recent HHS directives, one clamping down on staff travel and the other cutting back on scheduled salary increases at NIH. The new policy thus seemed to fit a pattern of HHS asserting its management authority over one of its administrative units.

The memo, written by HHS deputy assistant secretary for grants management Terrence Tychan, said that all HHS divisions should try to process grants more efficiently by cutting out face-to-face peer-review meetings. To reduce travel costs and other hassles, it proposed relying on “field readers.” Opinions would be gathered by mail, the memo explained, and program managers would consult submitted comments in deciding which projects to fund. Under the plan, HHS offices would implement “standard application review processes,” “consolidate and accelerate annual grant planning by linking it to the president's budget,” and bring about “standard scoring of competing applications.” According to Tychan's memo, the proposals were “favorably received” on 27 July by HHS Secretary Tommy Thompson and were headed for “a realistic implementation strategy.” Because the plan has not been finalized, the memo notes, it “is not to be shared with the world.”

But some NIH officials, worried that the plan might take effect without wider discussion, did not keep quiet. The matter reached Wendy Baldwin, NIH assistant director for extramural research. After talking with Tychan, Baldwin says she's confident that NIH's peer-review system is not going to be changed. “We have already been engaged in streamlining,” Baldwin says. “I don't think [this memo] was talking about things that would really affect us. I have no indication that this is going to affect the way we do scientific peer review.”

One NIH official, who asked not to be identified, remains concerned that the proposed system would give more responsibility to office managers and less to scientists directly involved in research. Another NIH official asserts that NIH's face-to-face reviews are of higher quality than mail reviews, because participants perform better when they must present their views to a live audience of peers. If the memo's policies were adopted, said one institute director, “there would be a revolution” at NIH.

Tychan acknowledges that NIH scientific review is “different and unique.” The memo, he said in a telephone interview, was not intended for NIH but rather for a policy group that's “trying to streamline the way we do grant review at HHS.” Their goal is to “make some improvements in the operating divisions” that process HHS awards for research on topics that range from child welfare to Medicare benefits. After talking to Baldwin, he says that “I don't want anybody at NIH to get worried; this is not designed to make vast changes to peer review” at NIH.

View this table:

NIH runs the largest peer-review operation within HHS, tapping more than 10,000 scientists each year, according to an official in NIH's Center for Scientific Review (CSR). The reviewers—who are reimbursed for travel and lodging and receive $200 a day for expenses—are grouped by discipline into 150 study sections. They meet roughly three times a year to help sift through the 44,000 applications NIH receives annually. CSR deputy director Brent Stanfield estimates that each review costs NIH just over$1700—“very efficient,” he claims.

Tychan's interpretation of the memo may put the concerns at NIH to rest. Even if it doesn't, however, the recent dialogue suggests that NIH wields sufficient clout to carve an exemption to a policy edict even before the edict is issued.

2. ASTRONOMY

# Pull of Gravity Reveals Unseen Galaxy Cluster

1. Andrew Watson*
1. Andrew Watson writes from Norwich, U.K.

Just as the brightest headlights on the nighttime freeway don't necessarily adorn the heaviest trucks, the brightest objects in the nighttime sky may not be the weightiest. Yet even though it is mass that ultimately determines the structure of our universe, astronomers traditionally flock like moths to bright sources, mainly because those are the ones they can see.

Now astronomers are taking long strides into the ponderous realm of dark matter. Tony Tyson of Lucent Technologies' Bell Labs in Murray Hill, New Jersey, and colleagues have discovered a whole new cluster of galaxies and calculated its distance, without relying on its emitted light. Instead they inferred the unseen cluster's existence from the way its gravity rerouted light from more-distant galaxies beyond. Tyson's team is one of several racing to show that the technique, known as gravitational lensing, can be used to map matter in deep space (Science, 17 March 2000, p. 1899). But the new work, to be published in the 20 August issue of Astrophysical Journal Letters, “is the first convincing demonstration this goal will actually be realized by gravitational lensing,” says University of Chicago astrophysicist Wayne Hu.

Astronomers believe that about 90% of mass in the universe is dark. Telescopes can't see it, but its gravitational pull blows its cover. “Gravity doesn't care whether matter is dark or luminous,” says Tyson's Bell Labs colleague David Wittman, lead author on the paper describing the work. “All you need are background sources of light, which are all over the sky, and in principle you can find all the matter between us and the background sources.”

The team, which also includes Vera Margoniner of Bell Labs, Judith Cohen of the California Institute of Technology in Pasadena, and Ian Dell'Antonio of Brown University in Providence, Rhode Island, used a special wide-field camera attached to the 4-meter Blanco telescope near La Serena, Chile. The astronomers studied a square of sky about twice as big as the full moon, containing tens of thousands of galaxies but no previously known galaxy clusters. They analyzed each individual galactic speck for telltale distortions that might be caused by massive but invisible objects closer to the telescope. Typically, images of distant galaxies appear to wrap around the lens core like the rim of a wheel (see figure). By scrutinizing the whole image a piece at a time and measuring the distortion in each region, the team created a “mass map” of the intervening space. A dense patch in one corner of the map revealed a cluster of galaxies, which the team subsequently confirmed using a conventional telescope.

To gauge the distance to the lens, the astronomers exploited the fact that the more distant a light source is beyond a gravitational lens, the more the lens distorts the light in transit. So determining both how far the sources are from Earth and the amount that the images are distorted reveals the lens's location.

To measure very large cosmic distances, astronomers rely on redshift, the reddening of light that takes place as expanding space stretches the light's wavelength. The more distant the object, the redder the light. Tyson's team estimated the remoteness of the source galaxies by comparing their colors to those of other galaxies at known distances. Then, after studying how much the lens cluster distorted the light of thousands of sources, they calculated the distance to the cluster, again in terms of redshift. The team pegged the lens cluster's redshift at 0.3, corresponding to a distance of about 3.5 billion light-years. Double checking the spectra from some galaxies in the lens, the team confirmed the redshift value to within 10%. “I am still surprised at how well it works,” Wittman says.

Mapping clusters by mass should help to close the gap between predicted and observed cluster abundances, says Peter Schneider of the Max Planck Institute for Astrophysics in Garching, Germany. “In that respect, this sort of work is very valuable.” Hu goes further. “A catalog of clusters selected by their mass will be invaluable for the study of dark matter and dark energy in the universe,” he says. “I expect great things.”

# NSF Launches TeraGrid for Academic Research

1. Jeffrey Mervis

Promising benefits to researchers working on everything from drug discovery to climate forecasting, the National Science Foundation (NSF) last week launched what will be the nation's most powerful network for scientific computing. NSF has pledged $53 million to four U.S. research institutions and their commercial partners to build and operate a system expected to be up and running by 2003. Its official name is the Distributed Terascale Facility, taken from its targeted capacity to perform trillions of floating-point operations per second (teraflops) and store hundreds of terabytes of data. But if it's a success, it may go down in history as Internet 3. The institutions—the University of California, San Diego; the University of Illinois, Urbana-Champaign; the California Institute of Technology in Pasadena; and Argonne National Laboratory outside Chicago—are no strangers to supercomputing. San Diego and Illinois, for example, are home base for NSF's Partnership for Advanced Computational Infrastructure program. Last year NSF gave$45 million to the Pittsburgh Supercomputing Center for a 6-teraflops machine. But the TeraGrid, as it's been dubbed, is touted as a new breed of supercomputer, with software that will allow high-speed, high-bandwidth connections previously not possible.

“It's not just size or speed,” says Fran Berman, head of the San Diego Supercomputer Center. “This will change how people use data and how they compute.” Her counterpart at Illinois's National Center for Supercomputing Applications, Dan Reed, says the TeraGrid will “eliminate the tyranny of time and distance.”

It's already changed the sociology of supercomputing, with its cutthroat competition to have the biggest and fastest machine on the block. The winning institutions were the only entrants in what was scheduled to be a competition. “We were under a lot of political pressure to get this out by September,” says an NSF official, “and we only gave [applicants] 3 months to put together their bid. We knew that would be a tough deadline for people to meet.” Despite being the only applicant, the winners put together a proposal “that passed [peer review] with flying colors,” says Bob Borchers, NSF's head of advanced computing.

The TeraGrid will build on an existing 40-billion-bits-per-second fiber-optic network, the so-called Internet-2 created by Qwest, one of three key industrial partners in the facility. It will rely on clustered Linux servers from IBM powered by thousands of Itanium-family processors from Intel. Each of the four institutions will contribute elements to the TeraGrid; by April 2003, it is expected to deliver 13.6 teraflops of computing power and more than 450 terabytes of storage.

NSF officials are hoping that this fall Congress will give the agency enough money to connect the Pittsburgh center to the grid in a few years' time. That will be followed, says Borchers, by a “deepening” of the network to connect a steadily rising number of regional and local sites. That's the path NSF followed to help create its previous research backbone that became the Internet.

4. PLANT SCIENCE

# How Seedlings See the Light

1. Josh Gewolb

Seedlings start to turn green the instant they pop out of the earth and receive sunlight. Exactly how light touches off the chain of events that converts a ghostly pale seedling into a green, photosynthesizing plant has long been a mystery. Now, a team of scientists has filled in one of the major gaps in understanding this photomorphogenesis, as it's called, by uncovering a surprisingly simple three-step pathway involving blue light.

Plant scientists have known for roughly a decade that a plant protein called COP1 is a master regulator of photomorphogenesis. When seedlings germinate in the dark, COP1, which was discovered by Xing-Wang Deng's team at Yale University, keeps the genes that bring about the process in the “off” state by fostering the degradation of transcription factors needed for the genes' activity, including one well-characterized factor known as HY5. Then, when seedlings encounter light, COP1 levels in the nucleus fall, allowing the transcription factor levels to rise and switch on the photomorphogenesis genes.

Since COP1 was discovered, a number of laboratories have tried to identify what they assumed was a cascade of proteins connecting it to the photoreceptors that detect the light. But they could find no such proteins. Now Deng and colleagues have explained why this search has been futile. In a report published online this week by Science (http://www.sciencexpress.org/), they show that in the plant Arabidopsis thaliana, blue photoreceptors known as cryptochromes interact with COP1 directly. Thus, they suggest that the light signal may be transmitted to COP1 without the intercession of other proteins.

Plant physiologist Roger Hangarter of Indiana University, Bloomington, says the study is “profound, because people have been struggling for a long time to see how the photoreceptor gets information into the nucleus.” But plant geneticist Albrecht von Arnim of the University of Tennessee, Knoxville, cautions that the researchers have yet to show that a change in light conditions actually changes the interaction between cryptochromes and COP1.

The Deng team's current work was inspired by results published last year by Anthony Cashmore at the University of Pennsylvania in Philadelphia and his colleagues. They created Arabidopsis mutants in which cryptochrome structure was altered and that, consequently, experience photomorphogenesis even when reared in darkness. The observation that a change in cryptochrome shape blocks inhibition of the light response by COP1 suggested to Deng and his colleagues that there might be a direct interaction between COP1 and the cryptochromes.

A battery of tests performed by the team repeatedly caught the two sets of proteins in flagrante delicto: For example, antibodies against COP1 fished one of the two common classes of cryptochromes out of Arabidopsis extracts along with COP1, an indication that the two proteins do in fact associate in plant cells. The researchers also studied the overall gene expression profiles of plants with deficiencies in photoreceptors and COP1, and the similarities they found further the circumstantial case for a link between the two proteins.

Based on the data, the team members propose that blue light alters cryptochromes so that they can attach to COP1. This stops the protein from its normal business of tagging HY5 and other transcription factors for degradation. As a result, the gene-transcribing machinery, which is ready and waiting in the nucleus, can respond almost instantaneously to light-signal changes.

Plant researchers say that this three-step COP1 pathway is unexpectedly simple: Experience with other signaling pathways shows that there could easily have been a dozen regulators between cryptochromes at the cell surface and COP1. But the short pathway is not the end of the story.

For one, cryptochromes are not the only photoreceptors involved in photomorphogenesis. Red light receptors called phytochromes control another short pathway that also regulates the light response on time scales as short as those of the new cryptochrome pathway. Indeed, the two pathways may be connected, because Cashmore showed a few years ago that phytochromes can directly interact with cryptochromes. In addition, plant geneticist Chentao Lin of the University of California, Los Angeles, found a protein, SUB1, that now looks like it might modulate the COP1 cascade.

5. ECOLOGY

# Field Test Backs Model for Invader

1. Christine Mlot*
1. Christine Mlot is a writer in Madison, Wisconsin.

MADISON, WISCONSIN—The growing threat from invasive species has spurred researchers to try to forecast whether a newcomer species will fade into oblivion or become the next kudzu. To do so, scientists have devised computer models for target species. Here at the annual meeting of the Ecological Society of America last week, a researcher at Washington State University in Pullman presented results that meld such modeling with field tests of a weed currently plaguing the southeastern United States. And although the results are grim, they demonstrate the importance of attacking weeds early in their invasion, before they take hold.

The work involves the Chinese tallow (Sapium sebiferum), introduced in the South as an ornamental tree in the 1770s and outlawed in recent years by the state of Florida because of its invasiveness. Robert R. Pattison, who works in plant ecologist Richard N. Mack's group in Pullman, began by plugging the temperature and moisture conditions of the plant's native range in China into a commercially available computer model called CLIMEX. When he applied the same parameters to other parts of Asia and Australia, the model—which has been widely used to predict the distribution of introduced biocontrol agents—generated a map that corresponds to areas where the tree has indeed invaded. When Pattison applied the model in the United States, it showed that Chinese tallow has a potential that is “well beyond its current range” in the South, he told those attending the meeting. The model predicts that the weedy tree could live as far north as Illinois and New Jersey and in scattered spots along the West Coast.

Pattison then took a novel step: putting the forecast to a test in the field. “That's what is nice about it,” says ecologist Erika Zavaleta of Stanford University, who studies invasive species and global change. Pattison planted Chinese tallow seedlings at seven eastern research sites, all within the U.S. range predicted by the model. The seedlings, now leafy 2-meter saplings, have thrived in the 2 years since planting at sites as far away from the current range as Maryland. The plants did well in both dry and moist areas, and did especially well when planted at sites with an open canopy. “They grew like crazy,” says Pattison.

The field results suggest that the model lacks other variables that might account for why the plant has not yet spread to Maryland and other northern regions in the 200 years it has been in the United States, says Zavaleta. Still, she says, Pattison's work in coupling field testing with modeling “is really valuable at giving a rough cut” of a plant's potential distribution.

That forecast is also heavily influenced “by what people plant,” says Pattison. Chinese tallow's recent appearance in the wild in California, he says, is a likely result of its popularity among backyard gardeners.

6. HUMAN GENOME

# Less Can Be More, U.K. Study Finds

1. Ben Shouse*
1. Ben Shouse is an intern in the Cambridge, U.K., office.

HINXTON, U.K.—In addition to setting a scientific milestone, the publication of competing drafts of the human genome sequence last February marked a struggle for priority in which the rival groups tried to undermine each other's claim by challenging the quality of the results. Now, the first formal comparison of the public and private genome maps confirms that there are indeed major differences between the data sets. And it suggests that the version produced by Celera Genomics of Rockville, Maryland, is more accurate.

At the Genome Informatics Conference here on 9 August, Colin Semple, head of bioinformatics at the Medical Research Council's Human Genetics Unit in Edinburgh, described an analysis of a 6.9-megabase stretch of chromosome 4 (4p15.3 to p16.1), a region implicated in bipolar disorder. At least 96% of this challenging region had been sequenced independently by Kathy Evans and her group at the University of Edinburgh using a map-based method.

Semple's team compared the sequences published by Celera, the publicly funded Human Genome Project (HGP), and the Evans team. Contrary to speculation, Celera's approach of breaking the whole genome into random fragments for sequencing yielded better data than the map-directed approach used by HGP. For this swath of DNA, Celera made half as many “misassemblies”—putting a fragment in the wrong order, or flipping it—as the public effort did, logging 2.08 misassemblies per megabase. However, Semple's team found that the Celera stretch is still full of holes: Celera had sequenced only 23% of the region, while HGP had managed 59% of it. Celera has “the best quality data, possibly as a result of having so little sequence in it,” says Semple. He notes that his group analyzed data that were publicly available as of 1 September 2000, so both sequences undoubtedly have been polished since then. And it's unknown whether the accuracy rates in this chromosome 4 region can be extrapolated to other regions.

Semple's presentation provoked surprisingly little rancor at the meeting, sponsored by the Wellcome Trust and Cold Spring Harbor Laboratory, which was a mostly cordial affair punctuated by hearty laughs in response to inside jokes and kooky names for genome annotation algorithms such as FANTOM (for functional annotation of mouse). Indeed, the comparative study's implications depend on one's point of view. “It's a question of whether you want Havarti or Swiss cheese. The public assembly doesn't have that many holes, but the holes it does have are much bigger,” says Jim Kent of the University of California, Santa Cruz, author of the computer program used for the initial assembly of the HGP genome sequence. Kent advises scientists considering a subscription to Celera's database to take a utilitarian approach. First they should examine the region of interest in the public database, he says: “If it's in good shape, then praise the lord they've just saved themselves $20,000.” 7. EVOLUTIONARY BIOLOGY # Queens, Not Workers, Rule the Ant Nest 1. Elizabeth Pennisi Queen fire ants not only populate their colonies, but they can also influence its sex ratio by limiting the number of female eggs they produce. This finding, reported on page 1308 by evolutionary biologist Laurent Keller of the University of Lausanne, Switzerland, and his colleagues, is contrary to expectation. Most entomologists thought that the workers, the queen's female daughters who actually raise the colony's young, determine the ratio of males to females, as they have the ability to kill or starve unwanted eggs. But, says Lotta Sundström, an evolutionary biologist at the University of Helsinki, Finland, “this is really the first experimental test of who determines [sex] ratios.” The results are “compelling,” adds Jon Seger, an evolutionary biologist at the University of Utah in Salt Lake City. The idea that the female workers should shape the relative proportions of females and males in Solenopsis invicta ant colonies dates back to work done 25 years ago by Harvard's Robert Trivers and Hope Hare. They predicted that worker females would favor females over males for evolutionary reasons: Insect males develop from unfertilized eggs and carry just one copy of every chromosome, while the females arise from fertilized eggs and have two copies, one from each parent. This so-called haploid-diploid system of sex determination skews the relatedness of the offspring such that sisters have more genes in common and, therefore, are more closely related to one another than to their brothers. Thus, while queens benefit equally by producing sons or daughters, worker females should prefer to raise sisters so they can pass on more of their genes. Trivers and Hare's subsequent analysis of sex ratios—using the relative dry weights of males and females in preserved ant, bee, wasp, and termite colonies in museum collections— supported an apparent bias in favor of females among these social insects. In contrast, solitary species tended to have the usual one-to-one proportion of females to males. To reflect the degree of genetic relatedness between the two sexes, ant colonies should have three females to every male. And although several studies bore out this prediction, “there have been these nagging inconsistencies,” notes Kenneth Ross, an entomologist at the University of Georgia, Athens. Some ant colonies had more males than predicted, and fire ants sometimes even had colonies with many more males than females. Keller and his colleagues decided to try to find out why. In 1999, they collected 24 fire ant colonies, each of which had a queen. The researchers reared those colonies in the lab for at least a week, then counted the number of males and females among 100 individuals selected from each colony. Eleven turned out to have mostly males, and 13 consisted almost entirely of females. The researchers then switched queens from male-dominated colonies with queens from female-dominated ones. Five weeks later, the researchers found that the sex ratios had reversed. The formerly male-dominated colonies that were now headed by queens from female-dominated colonies were biased toward females, and the previously female-dominated colonies with their new queens now had excess males. Furthermore, an examination of the DNA in the eggs revealed the reason for this change. More than half the eggs laid by queens from male-dominated colonies were haploid—or male. The queens—not their worker daughters—were influencing the sex ratio, by laying either more male or more female eggs. Most likely, they controlled the number of eggs fertilized by their internal stores of sperm. The researchers still don't know why some colonies are nonetheless biased toward females. Older females may simply be less efficient at fertilizing eggs and thus produce male-biased colonies, or they have a genetic disposition to produce mostly one sex over another. “We need to look at a colony to see if a queen's rate of fertilization is stable through time,” says Ross. The Keller team hopes to do those experiments and look at other species, too. When they do, entomologist Madeleine Beekman of the University of Sydney, Australia, predicts that they will find a variety of scenarios based on “which of the conflicting parties can exert power over the other parties.” The outcome, she adds, “is likely to depend on the idiosyncratic details of [each species'] biology.” 8. ECOLOGY # Parasitic Wasps Invade Hawaiian Ecosystem 1. Erik Stokstad Like specialized shock troops, insects have frequently been drafted to fight invasive weeds and pests. But these so-called biological control agents also have the potential to run amok, attacking innocent native species. Now comes the most detailed example yet of how deeply a control agent can infiltrate an ecosystem. On page 1314, ecologists M. Laurie Henneman and Jane Memmott of the University of Bristol, U.K., report that parasitic wasps from Texas and China that were introduced into Hawaii more than 50 years ago to prey on sugarcane pests are now dominant players in the food web of a remote native forest. Use of parasitoid wasps, which kill other insects by laying eggs in them, is a popular strategy for trying to control pests. And although there have been hints that such wasps—for instance, those introduced to combat the gypsy moth—can harm native species, “we had no idea that the nontarget impacts that exotic parasitoids had on native invertebrates could be so great,” says biocontrol expert Pauline Syrett of Landcare Research, a government-funded institute in Lincoln, New Zealand. “The results will shock many ecologists and managers of natural areas.” Syrett and others suggest that the finding justifies more stringent prerelease evaluations of biocontrol agents, including mandatory tests to assess the number of species they attack. “It's a call for safer practices” and for narrower specialists that will target only pests, says Robert Pemberton, an entomologist and botanist at the U.S. Department of Agriculture's (USDA's) Invasive Plant Research Laboratory in Fort Lauderdale, Florida. To study how the biocontrol insects interact with native species in the Hawaiian food web, Henneman and Memmott went to Kauai's Alakai Swamp —a boggy forest much higher, cooler, and wetter than the lowland fields where more than 122 parasitoids have been released in the last 100 years. Because the swamp is an extreme environment for Hawaii, the researchers reasoned, any effects from the parasitoids would be a minimum estimate of effects elsewhere. In the swamp, Henneman collected moth caterpillars from each of two 200-meter-by-25-meter plots, while keeping track of which of 52 kinds of plants the caterpillars were feeding on. Back in the lab, she reared 2112 caterpillars, carefully feeding them leaves she brought to the Kauai Agricultural Research Center and watching to see if parasitoid wasps would burst from the caterpillars. The vigil paid off; parasitoids emerged from 216 caterpillars, and an equal number of dead caterpillars contained wasp larvae. All told, Memmott and Henneman estimate that about 20% of the caterpillars had been parasitized. Of the wasps, 3% were native and 14% had been accidentally introduced to Hawaii; a whopping 83% were biocontrol agents. “The parasitoids have an amazing ability to move into a pretty difficult environment,” says Pemberton. The biocontrol wasps belonged to three species, and when Henneman and Memmott looked up their release dates, they learned that all three had been set loose more than 50 years ago. This makes it difficult to assess whether the wasps have damaged the swamp ecosystem, because little is known about the original community. Memmott says that the native moths have most likely been attacked for decades and can sustain the rate of parasitism. However, she notes, the most vulnerable moths could have been driven extinct soon after the biocontrol parasitoids arrived in the 1940s and '50s, and there would be no way to detect that. There is some good news, however. A large population of biocontrol caterpillars, released to control exotic blackberry species, has also reached the swamp, and had it been a target of the parasitoids, it could have increased the number of wasps, thereby exacerbating their effect on native caterpillars. But Henneman and Memmott found that the wasps don't parasitize the biocontrol caterpillars. Also encouraging is the absence of more recently released parasitoid wasps. “It means that biological control is much safer today,” Memmott says. Peter Follett, an entomologist at the USDA's Pacific Basin Agricultural Research Center in Hilo, Hawaii, agrees. “We're definitely more cautious [about releases] now than we were in the '40s, and even in the '80s,” he says. But others aren't sure that the absence of more recent releases is significant. Frank Howarth, an entomologist at the Bishop Museum in Honolulu, points out that these insects simply may not have arrived yet at Alakai Swamp. If they do, the swamp moths will have to face even more renegade mercenaries. 9. BIOMEDICAL POLICY # Bush Squeezes Between the Lines on Stem Cells 1. Gretchen Vogel* 1. With reporting by Constance Holden, and Pallava Bagla in New Delhi. President Bush's decision to support the use of existing embryonic stem cell lines surprised many and angered some. It also triggered a hunt to find them Ending months of speculation, President George W. Bush told a national TV audience on 9 August that he would allow the federal government to fund work on embryonic stem (ES) cells. But rather than settling the debate, Bush's compromise has sparked a new round of scientific questions, including exactly how much research the new policy will permit. Although the president's decision may temporarily halt a political push for more research opportunities, his assertion that some 60 ES cell lines would be available for research is more than twice the number most scientists would cite. In his first televised speech since January, Bush announced that the National Institutes of Health (NIH) can fund work with human ES cells—but only those cell lines derived before his announcement. The spin has been generally favorable: Both supporters and opponents of the research, who had condemned such a compromise last month (Science, 13 July, p. 186), praised the decision. “The compromise he's tried to achieve is a laudable one,” says cell biologist Douglas Melton of Harvard University, who works with both human and mouse ES cells. And although some pro-life groups, including the U.S. Conference of Catholic Bishops, condemned the decision, others said the president had kept his campaign pledge to oppose research that destroys embryos. Embryonic stem cells can in theory develop into any cell type in the body, and many scientists think they could eventually be used to treat chronic diseases such as diabetes or Parkinson's. Any clinical applications, however, are many years away. But because the cells are derived from week-old human embryos, which are destroyed in the process, their use has been controversial. Last summer, NIH issued guidelines to govern work on the cells, but the new Administration halted their implementation this spring while it reviewed the issue. When the White House floated this compromise several weeks ago, scientists in favor of ES cell work said that too few cell lines would be available to fully determine the potential of this new field. They were also concerned that many of the existing cell lines have commercial strings attached that could limit research. In response, Secretary of Health and Human Services Tommy Thompson asked Lana Skirboll, head of science policy at NIH, to tally the number of cell lines worldwide. To her surprise, Skirboll said at a 10 August press conference, her staff tracked down at least 60 from five countries: the United States, Australia, Sweden, Israel, and India. All 60 lines are propagating, she said, have cell surface markers characteristic of ES cells, and should be available to researchers within months. The unexpectedly high number of cell lines arose, Skirboll said, because many researchers have been keeping quiet and agreed to talk to NIH only under conditions of confidentiality. “If there are 60 cell lines, that's news to me—and good news,” Melton says. He cautions, however, that the properties of the purported cell lines are unclear, as are intellectual property rights. Cell surface markers, for example, are a “necessary but not sufficient” indication of an ES cell line's characteristics, he says. An informal search by Science turned up 34 distinct cell lines in four countries. At least seven cell lines have been described in peer-reviewed publications: five by James Thomson of the University of Wisconsin (UW), Madison, and two by Martin Pera of Monash University in Melbourne, Australia, and his colleagues. (Scientists from the Jones Institute in Virginia published a description of two ES cell lines, but they do not qualify for NIH funding because they were derived from embryos specifically created for research.) The remaining new lines have been described at scientific meetings or in conversations with Science. Pera's team has derived four more lines, he says. Joseph Itskovitz-Eldor of Technion University in Haifa, Israel, who collaborated with Thomson to derive the first ES cell lines, told Science he has derived three more of his own. Bresagen, a company based in Atlanta, announced having derived four. Peter Eriksson told Science that he and his colleagues at Göteborg University in Sweden have derived five lines and have so far characterized three. Michael Andäng of Huddinge University Hospital outside Stockholm told Science that he and his colleagues are characterizing “five to 10 human ES cell lines” they have derived. Roger Pedersen of the University of California, San Francisco (UCSF), has derived at least one. Science could not confirm reports of work with human ES cells in India. Reliance Life Science of Bombay markets a product called ReliCord, derived from umbilical cord blood. The company was rumored to have ES cell lines, but Anand Rao, research director for cell biology, told Science it has none. Manju Sharma, India's biotechnology secretary, says she knows of no ES cell lines in the country. Pera echoed the general surprise at NIH's total, suggesting that they may still be in early, uncertain stages of derivation. “It is no small chore to derive, cryopreserve, and properly characterize three or four lines,” he says. “Someone must have a factory somewhere, or we are talking about potential cell lines.” But at the briefing Skirboll stood by her number and predicted that even more cell lines would come to light. ## Relaxed requirements Even as the president sought to limit criticism by opponents of ES cell research, the ethical requirements that he laid down last week are much simpler than those issued last summer by NIH. The new criteria require that cell lines have been derived from embryos that were created for fertility treatments but are no longer needed. In addition, the couples donating the embryos must have given their informed consent, without any financial inducements. By contrast, the NIH guidelines issued last summer stipulated that NIH-approved cell lines must have been derived from frozen embryos, and those doctors procuring the embryos could not also derive the cell lines. In addition, the consent form had to meet certain strict criteria, explicitly stating, for example, that the cell lines “may be kept for many years” (Science, 1 September 2000, p. 1442). The NIH guidelines, which were never implemented, would have forced researchers to derive new cell lines, while Skirboll says that all 60 lines meet Bush's looser ethical criteria. The new policy may also ease some administrative burdens. Skirboll said that NIH was reviewing a prohibition against commingling any federal research funds—even support for building maintenance and upkeep—with private money used for embryo research that NIH is not allowed to fund. That policy has forced some researchers, including UW's Thomson and Pedersen at UCSF, to set up separate laboratory space for their human ES cell work. Bush's announcement does not affect research in the private sector. Companies or university researchers with non-NIH funding can derive new cell lines, but the NIH will not fund work with such lines. NIH is also scrapping its Human Pluripotent Stem Cell Review Group (HPSCRG), which was to have met in April to determine whether several newly derived cell lines would qualify. “There will be no new derivations for the NIH to approve,” says Skirboll, “so there is no need for the HPSCRG.” NIH has already begun setting up a registry of qualifying cell lines, said Skirboll, with cell lines that pass muster going into a database. Scientists will have to specify which cell line or lines they plan to use and make the necessary material transfer agreements (MTAs) with companies that own the cells. At the 10 August press conference, Thompson said that a scientist has been chosen to lead the database compilation but declined to disclose the name. He also said NIH could begin funding new grants to work with ES cells by early next year, and scientists could apply for supplemental grants even sooner—perhaps within weeks. Which cell lines come with strings attached—and how tight those strings are— remains unclear, as much of the derivation research to date has been funded by companies. “There are strong property and patent issues to work through,” Thompson acknowledged at the press conference. Several scientists, for instance, chafed at the requirements of an earlier MTA from WiCell, the not-for-profit company UW set up to distribute the human and rhesus monkey ES cells derived by UW researchers. Skirboll said that the owners of several cell lines have promised to work with scientists. The compromise seems to have slowed momentum in Congress for passage of a bill, sponsored by Senators Arlen Specter (R-PA) and Tom Harkin (D-IA), that would allow NIH also to fund work on the derivation of new cell lines. Robert Rich, president of the Federation of American Societies for Experimental Biology, which had lobbied hard for the Specter-Harkin bill, predicted that advocacy groups would ease up on their efforts, at least in the short term. Says Rich: “Most of us will take a deep breath for now and will wait for the 60-plus stem cell lines that the president says are available for research to appear.” Even if the research community gains access to all 60 cell lines, Philip Noguchi of the Food and Drug Administration says that it is very unlikely that any of the existing cell lines would be approved for clinical trials. That drives home one point in President Bush's speech: Despite the promise of ES cells, real treatments are still many years away. 10. BIOMEDICAL POLICY # New Chair of Bioethics Panel Wants National Debate on Issues 1. Eliot Marshall President George W. Bush's decision on embryonic stem cells may have dimmed the hopes of some federal researchers, but it has given ethicists a new lease on life. The president intends to create a new White House Council on Bioethics, to be chaired by University of Chicago moral philosopher Leon Kass. The new panel will succeed one created by President Bill Clinton—the National Bioethics Advisory Commission (NBAC)—that will disappear on 3 October. The White House has provided little information on the new council. In a telephone interview, Kass said that he was invited to become its chair only 2 weeks ago. According to the White House, the panel will monitor studies of embryonic stem cells, recommend regulations, and “consider all of the medical and ethical ramifications of biomedical innovation.” Kass hopes that its members—numbering “in the teens”—will be named soon. The panel will include people with a wide spectrum of views and have a “broad mandate,” says Kass, adding that it will do what “the president himself did in reaching his decision: namely, consult widely and make sure that all responsible points of view are heard.” Regulatory issues will be handled by the National Institutes of Health (NIH), according to a government spokesperson. With an M.D. from the University of Chicago and a Ph.D. in biochemistry from Harvard University, Kass worked at NIH in the late 1960s and later contributed to the founding of a bioethics research group, the Hastings Center in Garrison, New York. The center's founder, Daniel Callahan, thinks that Kass “will be very careful to get a fair range of people on the council.” NBAC's chair, former Princeton University president Harold Shapiro, says that “while I disagree with a lot of things Leon Kass says, I have a lot of respect for him.” Shapiro adds that he's glad the president “feels the need for continuing advice” on biomedical ethics. Kass worries that his opposition to human cloning and advocacy of restraints on medical technology will cause some researchers to regard him as “a Luddite.” But he expects to prove them wrong by making the council a forum for debate. “People who are nervous about me should wait and see. … My vision is for this council to become a kind of teacher to anybody who is interested” in bioethics. 11. BIOMEDICAL POLICY # "I Have Given This Issue a Great Deal of Thought ... and I Have Found Widespread Disagreement" The White House released this transcript of President Bush's nationally televised 9 August speech on stem cell research. Good evening. I appreciate you giving me a few minutes of your time tonight so I can discuss with you a complex and difficult issue, an issue that is one of the most profound of our time. The issue of research involving stem cells derived from human embryos is increasingly the subject of a national debate and dinner table discussions. The issue is confronted every day in laboratories as scientists ponder the ethical ramifications of their work. It is agonized over by parents and many couples as they try to have children, or to save children already born. The issue is debated within the church, with people of different faiths, even many of the same faith coming to different conclusions. Many people are finding that the more they know about stem cell research, the less certain they are about the right ethical and moral conclusions. My administration must decide whether to allow federal funds, your tax dollars, to be used for scientific research on stem cells derived from human embryos. A large number of these embryos already exist. They are the product of a process called in vitro fertilization, which helps so many couples conceive children. When doctors match sperm and egg to create life outside the womb, they usually produce more embryos than are planted in the mother. Once a couple successfully has children, or if they are unsuccessful, the additional embryos remain frozen in laboratories. Some will not survive during long storage; others are destroyed. A number have been donated to science and used to create privately funded stem cell lines. And a few have been implanted in an adoptive mother and born, and are today healthy children. Based on preliminary work that has been privately funded, scientists believe further research using stem cells offers great promise that could help improve the lives of those who suffer from many terrible diseases—from juvenile diabetes to Alzheimer's, from Parkinson's to spinal cord injuries. And while scientists admit they are not yet certain, they believe stem cells derived from embryos have unique potential. You should also know that stem cells can be derived from sources other than embryos—from adult cells, from umbilical cords that are discarded after babies are born, from human placenta. And many scientists feel research on these type of stem cells is also promising. Many patients suffering from a range of diseases are already being helped with treatments developed from adult stem cells. However, most scientists, at least today, believe that research on embryonic stem cells offer the most promise because these cells have the potential to develop in all of the tissues in the body. Scientists further believe that rapid progress in this research will come only with federal funds. Federal dollars help attract the best and brightest scientists. They ensure new discoveries are widely shared at the largest number of research facilities and that the research is directed toward the greatest public good. The United States has a long and proud record of leading the world toward advances in science and medicine that improve human life. And the United States has a long and proud record of upholding the highest standards of ethics as we expand the limits of science and knowledge. Research on embryonic stem cells raises profound ethical questions, because extracting the stem cell destroys the embryo, and thus destroys its potential for life. Like a snowflake, each of these embryos is unique, with the unique genetic potential of an individual human being. As I thought through this issue, I kept returning to two fundamental questions: First, are these frozen embryos human life, and therefore, something precious to be protected? And second, if they're going to be destroyed anyway, shouldn't they be used for a greater good, for research that has the potential to save and improve other lives? I've asked those questions and others of scientists, scholars, bioethicists, religious leaders, doctors, researchers, members of Congress, my Cabinet, and my friends. I have read heartfelt letters from many Americans. I have given this issue a great deal of thought, prayer and considerable reflection. And I have found widespread disagreement. On the first issue, are these embryos human life—well, one researcher told me he believes this five-day-old cluster of cells is not an embryo, not yet an individual, but a pre-embryo. He argued that it has the potential for life, but it is not a life because it cannot develop on its own. An ethicist dismissed that as a callous attempt at rationalization. Make no mistake, he told me, that cluster of cells is the same way you and I, and all the rest of us, started our lives. One goes with a heavy heart if we use these, he said, because we are dealing with the seeds of the next generation. And to the other crucial question, if these are going to be destroyed anyway, why not use them for good purpose—I also found different answers. Many argue these embryos are byproducts of a process that helps create life, and we should allow couples to donate them to science so they can be used for good purpose instead of wasting their potential. Others will argue there's no such thing as excess life, and the fact that a living being is going to die does not justify experimenting on it or exploiting it as a natural resource. At its core, this issue forces us to confront fundamental questions about the beginnings of life and the ends of science. It lies at a difficult moral intersection, juxtaposing the need to protect life in all its phases with the prospect of saving and improving life in all its stages. As the discoveries of modern science create tremendous hope, they also lay vast ethical mine fields. As the genius of science extends the horizons of what we can do, we increasingly confront complex questions about what we should do. We have arrived at that brave new world that seemed so distant in 1932, when Aldous Huxley wrote about human beings created in test tubes in what he called a “hatchery.” In recent weeks, we learned that scientists have created human embryos in test tubes solely to experiment on them. This is deeply troubling, and a warning sign that should prompt all of us to think through these issues very carefully. Embryonic stem cell research is at the leading edge of a series of moral hazards. The initial stem cell researcher was at first reluctant to begin his research, fearing it might be used for human cloning. Scientists have already cloned a sheep. Researchers are telling us the next step could be to clone human beings to create individual designer stem cells, essentially to grow another you, to be available in case you need another heart or lung or liver. I strongly oppose human cloning, as do most Americans. We recoil at the idea of growing human beings for spare body parts, or creating life for our convenience. And while we must devote enormous energy to conquering disease, it is equally important that we pay attention to the moral concerns raised by the new frontier of human embryo stem cell research. Even the most noble ends do not justify any means. My position on these issues is shaped by deeply held beliefs. I'm a strong supporter of science and technology, and believe they have the potential for incredible good—to improve lives, to save life, to conquer disease. Research offers hope that millions of our loved ones may be cured of a disease and rid of their suffering. I have friends whose children suffer from juvenile diabetes. Nancy Reagan has written me about President Reagan's struggle with Alzheimer's. My own family has confronted the tragedy of childhood leukemia. And, like all Americans, I have great hope for cures. I also believe human life is a sacred gift from our Creator. I worry about a culture that devalues life, and believe as your President I have an important obligation to foster and encourage respect for life in America and throughout the world. And while we're all hopeful about the potential of this research, no one can be certain that the science will live up to the hope it has generated. Eight years ago, scientists believed fetal tissue research offered great hope for cures and treatments—yet, the progress to date has not lived up to its initial expectations. Embryonic stem cell research offers both great promise and great peril. So I have decided we must proceed with great care. As a result of private research, more than 60 genetically diverse stem cell lines already exist. They were created from embryos that have already been destroyed, and they have the ability to regenerate themselves indefinitely, creating ongoing opportunities for research. I have concluded that we should allow federal funds to be used for research on these existing stem cell lines, where the life and death decision has already been made. Leading scientists tell me research on these 60 lines has great promise that could lead to breakthrough therapies and cures. This allows us to explore the promise and potential of stem cell research without crossing a fundamental moral line, by providing taxpayer funding that would sanction or encourage further destruction of human embryos that have at least the potential for life. I also believe that great scientific progress can be made through aggressive federal funding of research on umbilical cord, placenta, adult and animal stem cells which do not involve the same moral dilemma. This year, your government will spend$250 million on this important research.

I will also name a President's council to monitor stem cell research, to recommend appropriate guidelines and regulations, and to consider all of the medical and ethical ramifications of biomedical innovation. This council will consist of leading scientists, doctors, ethicists, lawyers, theologians and others, and will be chaired by Dr. Leon Kass, a leading biomedical ethicist from the University of Chicago.

This council will keep us apprised of new developments and give our nation a forum to continue to discuss and evaluate these important issues. As we go forward, I hope we will always be guided by both intellect and heart, by both our capabilities and our conscience.

I have made this decision with great care, and I pray it is the right one.

Thank you for listening. Good night, and God bless America.

12. HUMAN EVOLUTION

# Max Planck's Meeting of the Anthropological Minds

1. Michael Balter

Four years since it began, a novel German experiment in tackling anthropology's thorniest problems is beginning to bear fruit

LEIPZIG, GERMANY—On a summer day in 1997, four near-strangers climbed high into the German Alps to mull over an offer that promised to transform their lives. The Max Planck Society had just asked the researchers—primatologist Christophe Boesch, linguist Bernard Comrie, geneticist Svante Pääbo, and psychologist Michael Tomasello—to create an institute on human and primate origins, evolution, and culture. Max Planck was tempting them with scientific liberation: generous funding that would enable them to assemble a world-class team and free them from grant-writing for the rest of their careers. But the prospect was a gamble. Did they really want to relocate to the former East Germany, where the Max Planck insisted that the new institute be located? Also preying on their minds was a huge wildcard in any attempt to form a cohesive team: How would the four acquaintances get along?

By the time they came down from the mountains, the answer was clear. “The chemistry between us really worked,” says Pääbo. And when they visited Leipzig after their alpine encounter, they found that this historic university town 180 kilometers southwest of Berlin was cultured and livable, not a drab Soviet holdover. “I thought, ‘I can recruit people, I can sell this place,’” says Tomasello.

That's just what Tomasello and his colleagues have done. Dozens of scientists from all over the world have flocked here since the Max Planck Institute for Evolutionary Anthropology opened in early 1998. (Now renting space in a former printing company, the institute plans to move into a new $30 million facility by 2002.) Part of the allure, says Max Planck president Hubert Markl, is that the institute helps fill an important gap. “Modern evolutionary biology was not very well represented in Germany,” he says. The institute also fits with the Max Planck's goal of strengthening science in the former East Germany (Science, 22 December 2000, p. 2244). Since coming to Leipzig, the four scientific directors of the new institute have continued much of the research that made their names: Pääbo's work with ancient DNA, Boesch's studies of wild chimpanzees, Tomasello's research into human and ape cognitive abilities, and Comrie's study of language diversity. But Leipzig's generous funding has also enabled them to strike out on new paths, as well as sow the first seeds of interdisciplinarity. While Pääbo's group has begun exploring genetic differences between humans and apes, Boesch has created a genetics lab to probe family relationships among chimps. And as Tomasello shifts into new studies of language acquisition, Comrie has teamed up with Pääbo's group to examine the relation between languages and the genetics of the people who speak them. Although a handful of interdisciplinary anthropology programs exist elsewhere, they are “not as coordinated and focused” as the Leipzig center, says University of Pittsburgh paleoanthropologist Jeffrey Schwartz. As Alan Cooper, director of the Henry Wellcome Ancient Biomolecules Center in Oxford, U.K., puts it, the institute “promises to synergize” research areas that often have little to do with each other. But some researchers say it is too early to tell how much synergy may develop. “It's still very new,” says cognitive neuroscientist Marc Hauser of Harvard University. So far, he says, “the impact is more from the individual players.” ## DNA detective work The idea for the experimental venture gestated for several years. In 1993, a Max Planck representative first ran the concept past Pääbo, a Swedish researcher at the University of Munich with a formidable reputation in ancient DNA research. A couple of years later, an outside committee chaired by geneticist Walter Bodmer of Oxford University endorsed the project and recommended hiring Pääbo to get the ball rolling. “The thought was that [Pääbo] would be the leader who would bring in the others,” Bodmer says. Pääbo has used his share of the Leipzig largesse—the institute has a$7.8 million budget spread across four departments—to build a state-of-the-art ancient DNA lab. Teasing DNA from millennia-old specimens is challenging, requiring painstaking use of the polymerase chain reaction (PCR)—which amplifies tiny amounts of DNA so it can be sequenced—in a lab shielded from modern DNA. Pääbo has campaigned hard for rigorous techniques to avoid the contamination that has sent many lofty claims crashing to earth. “In a field plagued with poor operating practices,” says Cooper, Pääbo's lab “is one of the few with an unquestioned reputation.”

Pääbo also maintains a long-standing interest in human origins. In a paper in Nature last December, his group, working with Ulf Gyllensten's team at the University of Uppsala in Sweden, reported the sequencing of the entire mitochondrial DNA (mtDNA) sequence from several dozen people around the world. Since the late 1980s, a number of researchers—including Mark Stoneking, an mtDNA pioneer now in Leipzig—showed that human evolutionary trees could be created by measuring the variation in mtDNA sequences among living humans. The new mtDNA data, Pääbo and Gyllensten claim, provide some of the strongest evidence yet that all modern humans are the descendants of ancestors who lived in Africa nearly 200,000 years ago.

And the team has moved full tilt into one of the hottest areas in molecular anthropology: deciphering the genetic differences among humans, chimps, and other apes. This may provide clues to how and why the human and ape lineages diverged—as well as which genes might be implicated in the cognitive talents that distinguish humans from other species. Although human and chimp DNA differs by less than 2%, little is known about the few genes that appear to make all the difference. This year Pääbo's group described key differences among humans, chimps, and monkeys in how homologous genes are expressed and regulated, especially in the brain (Science, 6 April, p. 44).

As one institute researcher says, half-jokingly, “the Max Planck thinks that Svante is going to win the Nobel Prize, and the rest of us are here to be his playmates.” The truth, insists Pääbo, is that he and his group are thriving in the interdisciplinary environment.

One such collaboration is using linguistics to reconstruct hidden strands of human prehistory, such as early migration patterns. Researchers in Pääbo's department, especially Stoneking, have teamed up with Comrie's linguists to probe how closely languages and genetic profiles are correlated among the peoples of the Caucasus Mountains. Geneticist Luigi Luca Cavalli-Sforza, now at Stanford University, pioneered similar methods, albeit with less sophisticated techniques, in the 1970s. The new work, claims Comrie, suggests that “the correlations between linguistic patterns and population genetics are not as close as Cavalli-Sforza thought they would be.”

Armenians, for example, are closer genetically to Azerbaijanis —their nearest neighbors, whose language has little in common with Armenian—than to other Armenian-speaking populations living farther away. This finding suggests that in the Caucasus, at least, replacement of one language by another—either Armenian by Azerbaijani or the other way around—has played a greater role in current linguistic patterns than has population migration. This and other studies under way in Leipzig are “already making major contributions,” says linguist Peter Cole of the University of Delaware, Newark.

Comrie, who still tackles purely linguistic issues such as language diversity, says he could never have found enough funding for these more classical studies if he hadn't come to Leipzig. For someone with a humanities background, he says, “if you get an outside grant, it's usually a small amount for a small project.”

## Family ties

Molecular approaches, as a complement to meticulous fieldwork, are also big in Boesch's primatology department. Since 1979, he and his colleagues have been studying wild chimps at Taï National Park in the Côte d'Ivoire, using observation techniques that minimize contact with the animals and thus reduce the degree of human influence on their behavior. Such studies have shown that chimps in different parts of Africa have different “cultures,” including variations in tool use, grooming practices, and courtship behaviors (Science, 25 June 1999, p. 2070).

Boesch, who worked at the University of Basel in Switzerland until coming to Leipzig in 1998, says the move has given him “a marvelous opportunity” to create a new primatological research group and assemble his own team of crackerjack geneticists rather than relying on outside scientists. This has enabled his group to help resolve controversies around chimp behavior that have dogged the field.

Boesch is at the center of one of these controversies. In 1997, for instance, he co-authored a study published in Nature that concluded from a genetic analysis of infant Taï chimps that females were surreptitiously mating with males outside their community. Half the Taï offspring, they found, were not related to their supposed fathers. Primatologist Jane Goodall and others challenged this finding in last May's issue of Molecular Ecology, arguing that they saw no evidence for such “extragroup paternity” in chimps in Tanzania's Gombe National Park. Goodall's group came to a similar conclusion after reanalyzing the Taï data as well, suggesting diplomatically that further genetic studies could help resolve the discrepancy.

At Leipzig, with the genetic tools needed to conduct a more sophisticated analysis of the Taï data, Boesch has taken up the gauntlet. New work, led by geneticist Linda Vigilant in Boesch's lab, analyzed DNA from hair and feces left by Taï chimps. Although such collection techniques are not new, Vigilant and her co-workers employed strict checks—such as ensuring that samples have enough DNA for analysis and repeating each PCR run many times—that enabled them to distinguish reliable DNA samples from those likely to give spurious results. After analyzing a larger subset of the Taï community, they concluded that extragroup paternity was, in fact, minimal, as Goodall had suggested.

Whereas Boesch must travel to Africa to observe chimps, a new $14 million primate facility at the Leipzig Zoo serves Tomasello's purposes just fine. He and his co-worker, psychologist Josep Call, are conducting numerous experiments at the facility to determine how human contact affects chimp behavior (see sidebar, p. 1247). Tomasello is also undertaking an intense study of human language acquisition. He's devoting a substantial chunk of his Leipzig funds to recording several hours of conversation each week between several mothers and their children. “We pay the mom to be a research assistant, and she turns on the tape recorder,” he says. The recordings, which begin at about age 2 when children start speaking more than one word at a time, require as much as 20 hours of transcription per hour of tape to ensure their accuracy. Aided by an army of transcribers, Tomasello's group has recorded two children for a full year. “We want to trace the language back all the way, where did it come from. If you go back day by day, it's like an archaeological record.” Although the Max Planck has brought together some of the brightest lights in their fields, observers say it is too early to judge whether the Leipzig recipe will succeed in spawning a new era of interdisciplinary work in anthropology. The institute plans to create two more departments—one in biological anthropology and one in social or cultural anthropology—and recruit topflight scientists in the coming months. These will be critical partners in the interdisciplinary marriage and position the institute to move beyond the compartmentalization that many researchers believe is slowing down the field. “The real question,” says Harvard's Hauser, is whether uniting these disparate fields will “actually influence how they do their work and think about [research] problems.” 13. HUMAN EVOLUTION # Zoo's New Primate Exhibit to Double as Research Lab 1. Elizabeth Pennisi LEIPZIG, GERMANY—Michael Tomasello and Josep Call want to know what it takes to make a mind human. Not only are these two comparative psychologists here at the Max Planck Institute for Evolutionary Anthropology tackling one of biology's fundamental questions, but they expect to conduct their work in full view of thousands of curious onlookers every day. The researchers are hoping to help bridge the gap between scientists and the public at a sleek new facility at the 120-year-old Leipzig Zoo. Four years in the making, the 13,552-square-meter Wolfgang Köhler Primate Research Center is one of the largest in the world, says Michael Seres, the research coordinator who helped set up the center and now oversees it. The design does double duty, satisfying educational and research needs. A boardwalk along the outdoor habitat leads into an artificial cave, where zoo patrons can get a closeup view through a glass partition of four primate species—including 15 chimps, a half-dozen gorillas, seven orangutans, and four bonobos—in surroundings resembling their natural habitats. Casual visitors get few hints to the remote platforms, cameras, and the dozen veterinarians and keepers who mind the animals and enable the science, including studies on the role of gestures as communication tools. “Doing cognitive research in a zoo setting is catching on,” notes Lisa Stevens, a senior curator at the National Zoo in Washington, D.C. Crucial to the research is housing the chimps on a 4000-square-meter island, separated from the other primates, such that they form a troop. This hierarchy of males and females will allow the researchers to monitor chimp cognitive development and the use of social skills such as grooming and facial expression. Very few zoos allow chimps to form a full troop, for fear of how the animals might behave in front of people. “Chimps are not as nice as gorillas or orangutans,” Seres explains. “They can kill one another and be quite aggressive.” Zoo managers are leery of how members of the public may react to chimp violence, or to cantankerous individuals spitting, defecating, or throwing objects at them. Most exciting to Seres is a now-empty boulder-lined enclosure connected to the island by a tiny doorway blocked with heavy mesh. The enclosure forms the front yard of a nursery, complete with incubators and a kitchen. That's where Seres and Tomasello hope to study how chimps come to understand the world around them, and in particular, how they learn to perceive the behavioral and psychological states of others. In doing so, the researchers will explore the influence of human interactions on a developing chimp mind. The experiment may help resolve a long-standing controversy over just how intelligent our closest cousins are and should dovetail nicely with their work on cognition in other apes. The controversy ignited in the early 1980s, when Sue Savage-Rumbaugh and her colleagues at Georgia State University in Atlanta reported that chimps demonstrate skills once thought to be the sole province of humans: language and an awareness of what others are thinking. The Georgia group had been trying to teach Matata, an adult female bonobo, to understand human language and to point to symbols to communicate. They didn't have much success with Matata, but to their surprise her adopted 6-month-old son Kanzi seemed to pick up language skills by observing her handlers. By age 2, Kanzi could use symbols to communicate, and he continued to expand his vocabulary by interacting with researchers and their symbol boards. Since then, the Georgia team and others have succeeded in getting other primates, including chimps, to demonstrate language, math skills, and self-awareness. But their mixed results—individuals often fail to master these skills—have left many experts uncertain of the degree to which primates are capable of humanlike brain power. Defending the work, psychologist Duane Rumbaugh, Sue's ex-husband and a collaborator at Georgia State, thinks that success depends largely on the degree of interaction between the primate and the researchers. “The animals won't respond,” he says, “unless they are brought up fundamentally as human children [are], with a rich, social, interactive life.” Given this background, he asserts, “they have the capability in many ways of becoming nonanimals just as much as we are.” Although such interpretations stir academic controversy, the work is combustible for another reason: Some scientists and animal activists contend that the experiments do not respect the animals for what they are and force them to do unnatural tasks. Now Seres and Tomasello are about to plunge in. Their research subjects are orphans: infant chimps abandoned by captive mothers, a common phenomenon at zoos. These infants “would die if humans didn't raise them,” says Tomasello. Once the Leipzig facility was nearly finished last year, he put the word out that the center would raise orphans. Earlier this year, a 3-week-old male arrived on the center's doorstep, even before the nursery was completed. “We had to run out and quickly buy the cribs,” Seres says. Tomasello and his colleagues say they have no intention of raising chimps like children or transforming them into “nonanimals.” Their goal is to use the orphanage to explore factors that influence chimp brain development with minimal disruption to the animals' normal upbringing. Although fed and cared for by human keepers, the young chimps will spend most of their time with their peers. Gradually they will be incorporated into the troop, although at first they will just see, smell, and touch the adult chimps and their young through a mesh. About half the youngsters will experience, with a person, intense parent-child interaction: playing games or being taught symbols. The idea is not to teach the animals how to take part in human activities but to give them the kind of intellectual stimulation generally showered on human infants. The rest of the animals will receive the care chimps typically get from their keepers. In this way, the researchers can test whether the degree of human-chimp interaction indeed makes a difference in chimp cognitive abilities. The work, Tomasello predicts, in conjunction with studies of the adult chimps and other great ape species, “should be extremely helpful to people trying to reconstruct the evolution of primate, including human, cognition.” 14. SCIENCE AND COMMERCE # Chemists for Hire: Have Flask, Will Travel 1. Joe Alper* 1. Joe Alper is a freelance writer living in Louisville, Colorado. A new breed of entrepreneurs—synthetic chemists—are selling their skills to drug companies through contract shops Two decades into the molecular biology revolution—which was supposed to herald the end of the age of chemistry—organic chemists are suddenly in hot demand. Potential drug targets are piling up faster than companies can adequately test them, for lack of those skilled in the art of organic synthesis. That leaves firms such as Bristol-Myers Squibb and Eli Lilly scrambling to add hundreds of chemists to their research staffs—and offering them$80,000 a head, signing bonuses of up to $40,000, and moving expenses. Even with those perks, however, there doesn't seem to be enough chemists to go around. Enter the chemist-entrepreneurs. A growing group of skilled synthetic chemists is seizing the opportunity to build chemistry-for-hire companies that custom-synthesize organic molecules. It's a strategy that many in the pharmaceutical and biotech industry admit is often cheaper and faster than doing it themselves. “There's been an explosive growth in companies set up to do contract synthesis at scales ranging from the milligram to the kilogram,” says Gifford Marzoni, a chemistry agent at Davos Chemical Corp. in Englewood Cliffs, New Jersey. Davos is a virtual synthesis shop that booked$100 million in revenues last year, up 10-fold over the past 5 years, by matching companies that need molecules with those that can make them.

Evolutionary Chemistry, a biotech start-up in Boulder, Colorado, turned to Davos recently when it needed some modified nucleotides for one of its drug-discovery efforts. Marzoni, in turn, is putting together a deal between Evolutionary Chemistry and a small company that can handle the assignment. “We couldn't afford to hire the chemists that we need, let alone find them, so we have to go outside the company to keep our projects going at full speed,” says Ted Tarasow, Evolutionary Chemistry's director of chemistry.

The dearth of skilled organic chemists goes back at least a generation. In the late 1970s, say longtime academic chemists such as Robert Coates of the University of Illinois, Urbana-Champaign, the best and the brightest students went into molecular biology, not stodgy old chemistry. Indeed, with characteristic hubris, some molecular biologists claimed that biotechnology was going to relegate synthetic chemistry to the back bench of drug development efforts. “Protein therapeutics were the story of the day, then,” recalls Bruce Diel, a synthetic chemist and founder of ChemFinet, an online marketplace for synthetic chemists based in Overland Park, Kansas. “Synthetic chemistry was passé.”

Around the same time, the negative publicity surrounding environmental disasters such as Love Canal further depleted the talent pool. “The only time chemistry made the news was when something negative happened, and that drove away many good students,” says Thomas D'Ambra, who 10 years ago founded Albany Molecular Research, the grandfather of the chemist-for-hire industry, headquartered in Albany, New York. By the early 1990s, molecular biologists were generating a torrent of drug targets—and both pharmaceutical and biotech companies were realizing that protein-based drugs, with their large molecules, were not going to replace small organic molecules as drugs.

Today, with the completion of the human genome project, drug companies big and small concede that they shouldn't have built up their molecular biology capabilities at the expense of their chemistry groups. Bristol-Myers Squibb, for one, has stated that it intends to hire enough chemists over the next few years to shift the biologist-to-chemist ratio in its drug-discovery labs from 3:1 to 1:1. And many biotech companies admit in private that their drug development efforts are hamstrung by the inability of in-house chemists—if they even have any—to meet the demands of their biologists. “We have vice presidents asking us why we haven't done any work on a good target,” says one researcher at a midsized California-based biotech company.

That's why rent-a-chemist companies such as Albany Molecular Research and Array Biopharma in Boulder, Colorado, have seen revenues as much as double annually over the past few years, although admittedly from a small base. Each company has hired over 100 chemists, with more on the way. Although the two firms are engaged in their own drug-discovery efforts, their assembled chemical talents are largely directed at custom syntheses for partners. Albany Molecular recently placed seventh on Business Week's 2001 list of hot growth companies, and Array was able to complete its initial public offering this past December when other companies refused even to try. “We have been able to capitalize on the increasing demand for high-quality chemical synthesis,” says Robert Conway, Array's chief executive officer, whose company booked \$5.7 million in revenues during the quarter that ended 30 June, up 152% over the same period a year earlier.

Albany Molecular and Array Biopharma are full-service companies offering a host of chemical services. Clients pay on a sliding scale based on the number of steps in the milligram-to-kilogram journey. In a recent contract, for example, Albany Molecular synthesized more than 400 compounds for use in a cardiovascular drug screen. After the contracting firm identified a lead compound from the initial 400, it came back to Albany—for an additional, larger fee—to further develop the drug candidate. Albany's partners include Eli Lilly, Aventis, and DuPont, whose spokespersons declined to comment publicly about their use of such contractors.

Other companies are smaller and specialize in a certain type of chemistry. Synthon Chiragenics in Monmouth Junction, New Jersey, makes carbohydrate-based drugs using chemistries developed by the company's founder and chief scientific officer, Rawle Hollingsworth, for example. Fluorous Technologies of Pittsburgh, Pennsylvania, is an expert in organic syntheses involving fluorine.

If chemists are in such short supply, where do these entrepreneurs find them? “We've been remarkably successful at taking people away from the major pharmaceutical companies,” brags Array's Conway. Adds D'Ambra, “The market is tight, no doubt about it, but we work hard at creating a place that's attractive to good chemists.” That formula includes competitive pay, plenty of stock options, and what one chemist called “the opportunity to work with other chemical heads.” Says another, “It's like being in academia, only I have the chance to be rich. Why should molecular biologists have all the fun?”

And although pharmaceutical companies grumble that these chemist-entrepreneurs aren't helping to solve the problem but are merely cherry-picking from their labs, their presence hasn't set off the type of bidding war that recently swept the high-tech industry. So until the inevitable response to supply and demand increases the crop of chemists, Marzoni says that companies like his are a practical alternative. “For those companies that can't hire enough chemists, which is most of them these days, the only solution is to seek help outside the company.”

15. NATURAL DISASTERS

1. Robert Koenig*
1. *International Tsunami Symposium 2001, Seattle, Washington, 7–10 August.

Computer models, improved maps of the ocean floor, and new sensory equipment are giving scientists a handle on the causes of giant waves

ISTANBUL—When a colossal wave smashed into a spit of land along Papua New Guinea's coast on 17 July 1998, it destroyed three villages and killed more than 2100 people. That's when Costas Synolakis swung into action. The University of Southern California (USC) coastal engineer rushed to the site with an international team of tsunami scientists, including geologists, a seismologist, hydraulic engineers, and computer modelers, to find out every detail they could about the decade's most deadly tsunami. They measured marks left by the waves, surveyed damage, took statements from shaken survivors, and scrutinized seismologic and hydroacoustic data. But the evidence left them with a persistent puzzle: How could a moderate earthquake off the coast generate such a devastating tsunami, with 20-meter-high waves that impaled bodies on tree branches and smashed every structure in the 25-kilometer-long sand spit between the Pacific Ocean and the Sissano lagoon?

Three years of data collection, debate, and computer modeling may have turned up an answer. At recent tsunami conferences in Istanbul and Seattle,* and in an article being prepared for publication this fall in the Proceedings of the Royal Society, Synolakis, seismologist Emile Okal of Northwestern University in Evanston, Illinois, and several colleagues propose what they believe to be the culprit: an underwater landslide. Evidence of such a “slump” turned up in a detailed bathymetry survey, or map of the ocean floor, co-sponsored by the Japan Marine Science and Technology Center and the South Pacific Applied Geoscience Commission. The survey showed sea-bottom scars of a major slump about 25 kilometers offshore of the Sissano lagoon. Studying seismologic and hydroacoustic data from the day of the tsunami, Okal concluded that the quake caused the slump, which in turn unleashed the deadly tsunami. And Synolakis and others on the survey team developed a computer model of the area that, they contend, confirms that such a “massive slump” offshore could have caused the Sissano tsunami. “This is the strongest evidence yet that ‘local’ tsunamis” —killer waves that originate just a few miles offshore of the site—“can be generated by massive underwater landslides,” Synolakis says.

Some researchers dispute Synolakis's interpretation, arguing that the Sissano wave was caused mainly by the quake. But even more important than finding the cause of the Papua New Guinea tsunami, everyone agrees, are the tools developed in the search—including more sophisticated seabed imaging and computer models for tsunami generation and inundation. Marine geologist David Tappin of the British Geological Survey (BGS) calls the 3 years of ocean expeditions and scientific debates that followed the Papua New Guinea event “a watershed for tsunami research.” According to Tappin, the “unprecedented depth of research” into the Sissano tsunami may allow researchers to “better understand why some areas are tsunami-prone and even to consider estimating the magnitude of the risk.”

Despite those advances in bathymetry and computer modeling, other researchers argue that the most crucial challenge to tsunami research lies in gathering real-time data about killer waves before they near the shore. By spotting a tsunami early and taking its measure, proponents of such research say, scientists can help head off disaster before it strikes. Such data are now being provided by the Deep-ocean Assessment and Reporting of Tsunamis (DART) warning system, scheduled to deploy its sixth instrument this month to help predict and analyze the giant waves on the open ocean.

## Harbor-wave history

The relation between earthquakes and tsunamis has been known for more than 2000 years—ever since the Greek historian Thucydides connected an Aegean tsunami in 426 B.C. to the quake that preceded it. Nevertheless, modern tsunami science is in its infancy. Only during the past decade have hydraulic engineers and other scientists begun using computers to model the three-dimensional evolution of tsunamis and devising inundation maps and early-warning systems for them. “Tsunamis killed more than 4000 people during the 1990s, but we have surprisingly little data to help us analyze them,” says Eddie Bernard, an oceanographer who heads the U.S. National Oceanic and Atmospheric Administration's (NOAA's) Pacific Marine Environmental Laboratory (PMEL), which developed the DART system.

Tsunami generation involves intricate interactions among earthquakes, landslides, and “sympathetic” vibrations between the quake and the ocean above it. The Japanese word tsunami means “harbor wave”—a reference to the giant waves' ability to penetrate the protected harbors along Japan's coast. Although sometimes inaccurately called “tidal waves,” tsunamis are produced by sudden underwater disruptions—usually undersea earthquakes but also submarine landslides and, far less often, volcanic eruptions or meteorites that hit the ocean.

The typical tsunami begins as a series of waves in the deep ocean, where they are not particularly dangerous. Although the wave pulses can race through the deep sea at speeds exceeding 700 kilometers an hour, their energy is dispersed along a wavelength as much as 750 kilometers wide. So a tsunami wave on the open sea may be just a few meters high, with a slope that's sometimes too gentle for big ships to notice. It's not until tsunamis enter shallow coastal waters that they get higher and more dangerous—often “shoaling,” or squeezing together into narrow monster waves that can be as high as 10-story buildings.

Since 1990, 11 major tsunami events have struck coasts from Java to Chile, killing more than 4000 people and causing hundreds of millions of dollars in damage. In all, there were about 80 tsunamis during that decade, many of which caused little damage because they were small or struck undeveloped coastlines. The National Geophysical Data Center in Boulder, Colorado, estimated that more than 80% of the world's tsunamis appear to be generated by undersea earthquakes around the Pacific Rim, where colliding tectonic plates lead to an unusually high level of seismic activity. But Synolakis and others believe that some of those tsunamis—especially local tsunamis, which are generated just a few kilometers off the coasts they strike—may be caused mainly by offshore landslides that are shaken loose by local or more-distant quakes.

## Predicting the waves of the future

Whatever causes tsunamis, the more important challenge, Bernard believes, is to devise systems that can detect newly generated tsunamis in the deep ocean, collect information about them, and transmit it in real time to warning stations that could then evacuate threatened coastal areas. To provide a coastal warning system and to collect data about deep-sea tsunamis, NOAA has been deploying the DART system—starting in the North Pacific, along the Alaska Subduction Zone, which is the most dangerous generator of tsunamis that tend to strike Hawaii and the U.S. West Coast. Each DART assembly consists of a tsunami-detecting “bottom-pressure recorder” device on the ocean floor, which sends acoustic signals through the water to a car-sized surface buoy. The buoy transmits the data via satellite to ground stations, which relay them to NOAA'S PMEL lab and several tsunami warning stations—including stations in Alaska and Hawaii. The warning stations, as well as efforts to map the likely local impact of tsunamis at specific sites along the U.S. West Coast, are part of the U.S. National Tsunami Hazard Mitigation Program, which began in 1996.

So far, three DART instruments have been set up along the Alaska zone (near the Bering Strait), and two have been deployed closer to the Oregon coast, near the Cascadia Subduction Zone, which is thought to generate large tsunamis every few centuries. This month, a sixth DART station is being deployed in the deep ocean off South America's coast—not far from the site of the 23 June quake that generated a major tsunami along Peru's south coast. Oceanographer Frank L. Gonzalez, who directs PMEL's tsunami mapping center, says the new DART station “will intercept tsunami waves traveling from generation zones in the South American Subduction Zone to Hawaii, Japan, and other Pacific Rim countries.” Gonzalez says the new equatorial DART station “will certainly help the warning centers issue faster, more reliable alerts for tsunamis generated off South America.”

In the future, Bernard says, he plans to harness new science and technology to hone DART's ability to detect earthquake-generated tsunamis and spread warnings in advance. Better coverage will also be needed, he says. Bernard and his PMEL colleagues have proposed a worldwide program, tentatively called TROIKA, to deploy similar instruments in other tsunami-vulnerable regions. Such regions include the South Pacific, the Atlantic Ocean off the coast of Portugal (a 1755 tsunami destroyed much of Lisbon, killing 60,000 people), the Aegean Sea, and perhaps the Sea of Marmara and the Black Sea.

## The next wave of killer waves

Despite the promise of DART and other systems for analyzing and predicting tsunamis, some experts worry that future dangers may overwhelm any defenses scientists are likely to devise. Judging from new evidence about “megatsunamis” in the distant past, such as the wave that battered the islands of the ancient Minoan civilization about 3500 years ago (see sidebar), they warn that more-destructive waves eventually will strike heavily populated coastlines, with potentially devastating impact. Ground zero is the Pacific Rim, where a seismically active “ring of fire” extending from the Bering Strait to the South Pacific unleashes earthquakes that trigger tsunamis.

Aside from the traditionally tsunami-battered Pacific islands of Japan and Hawaii, one of the most vulnerable regions is the Southern California coast. Jose Borrero, a postdoctoral researcher at USC who has worked with Synolakis in analyzing the tsunami threat along the California coast, says that “Southern California's offshore geology makes it ripe for producing tsunamis. Even a small tsunami along that coast would have a large potential for damage.” Another West Coast threat comes from the Cascadia Subduction Zone, off the coasts of Washington, Oregon, and Northern California. Recent analysis of sand layers deposited in the region by ancient tsunamis suggests that one part of the Cascadia zone may be nearing a tsunami-generating earthquake, perhaps during this century.

Researchers also see potential danger in smaller seas such as the Mediterranean, the Black Sea, and even the tiny Sea of Marmara south of Istanbul. For example, a tsunami hit the French coastal city of Nice in 1979, and a small tsunami struck Izmit, Turkey, after an earthquake there in 1999. Istanbul, a metropolis of 13 million that rises along the Sea of Marmara and the Bosporus strait, could well be affected by a Marmara tsunami. “There will be another earthquake in this region, and it is likely to occur offshore, in the Sea of Marmara—making a tsunami likely,” says BGS's Tappin. Ahmet C. Yalciner, an ocean engineer at Turkey's Middle East Technical University and co-director of the recent NATO tsunami workshop, fears that “underwater landslides could be a very important factor” in worsening a potential tsunami if an earthquake shakes under the Sea of Marmara.

But Bernard and others hope that advances in tsunami observation and warning systems, although powerless to influence the course of the giant waves, will help reduce the danger to vulnerable coasts. “We now have the equivalent of seismometers in the tsunami world,” says Bernard. “Once we have collected data from 100 tsunamis in the deep ocean, we may have the potential to understand these complex events.”

• *NATO Advanced Research Workshop, “Underwater Ground Failures on Tsunami Generation, Modeling, Risk and Mitigation,” Istanbul, 23–26 May.

16. NATURAL DISASTERS

# Modeling a 3600-Year-Old Tsunami Sheds Light on the Minoan Past

1. Robert Koenig

The collapse of the Stronghyle volcano on the Greek island of Thera (now Santorini) in about 1600 B.C. generated a tsunami that smashed into the strongholds of the ancient Minoan civilization, Crete, and other islands in the Aegean Sea. In recent years, some archaeologists have speculated that the tsunami might have caused enough damage to doom the seafaring Minoans, whose mysterious disappearance has fueled speculation for centuries.

But recent research by tsunami modeling experts—based on data collected by geologists on coasts around the Aegean Sea—now indicates that the tsunami generated by that collapsed volcano caldera probably didn't do in the Minoans on its own. They say it could not have damaged agriculture, fishing, and society critically enough to condemn the Bronze Age Minoans, who were scattered across islands such as Crete, Rhodes, and Kos.

Sediment samples taken from the coasts of Crete, western Turkey, and other Aegean sites confirm that the volcano did indeed generate a tsunami, sedimentologist Koji Minoura of Japan's Tohoku University reported at a recent tsunami conference in Istanbul. But computer modeling of the event indicates that the waves were only between 5 and 8 meters high for most of the Minoan islands, says Greek-born tsunami modeler Costas Synolakis of the University of Southern California in Los Angeles. “At generation, the waves may have been as big as 120 meters, but they dispersed rapidly as they propagated toward Crete,” Synolakis says. “The size of the wave that actually reached Crete would have been disruptive, but it would not have devastated the Minoans to the point that they abandoned their palaces.”

Synolakis says he doesn't know what delivered the coup de grâce to the Minoans. “There is a vexing mystery as to what eventually destroyed their world,” he says. But geologist Floyd McCoy of the University of Hawaii's Windward Community College points out that the tsunami was only one of several scourges the Stronghyle eruption must have unleashed. “Add up all the regional effects of this massive eruption—ash fall, pumice drifts and fall, tsunami, earthquakes, destruction of the home base for the Cycladic culture—and no civilization could rebound, much less a Bronze Age one.”

To complicate things further, experts wonder whether the tsunami was caused by the volcano's actual collapse or by the “pyroclastic flow” of scalding gases, ash, and other particles that the eruption swept into the sea. Europe's top expert on volcano-generated tsunamis, geophysicist Stefano Tinti of the University of Bologna in Italy, says that “modeling tsunamis generated by such pyroclastic flow is extremely difficult,” because so many factors are involved, including the extent to which the hot gases vaporize upper layers of the sea. “We know far less about how to model the pyroclastic generation of tsunamis than we do about the more common sources, such as earthquakes and submarine landslides,” says Tinti. “This is a ripe field for future research.”