# News this Week

Science  19 Feb 1999:
Vol. 283, Issue 5405, pp. 1090
1. AIDS RESEARCH

# Researchers Urged Not to Inject Virulent HIV Strain Into Chimps

1. Jon Cohen

An unusual coalition of prominent AIDS researchers, primatologists, and animal conservationists is urging vaccine developers not to inject chimpanzees with recently isolated strains of HIV that can cause AIDS-like disease in the animals. In a letter published on page 1117, virologist Alfred Prince of the New York Blood Center, chimpanzee advocate Jane Goodall, and nine others raise scientific and ethical objections to such experiments, but stop short of calling for a ban on them. “Before we jump off the diving board and use a virulent strain, we should stop and reassess this,” says co-signer Jonathan Allan, an AIDS researcher at the Southwest Foundation for Biomedical Research in San Antonio, Texas. “We're redefining what the ethical limits are for chimp studies.”

The letter is the latest twist in a long-running debate about the value of various animal “models” for studying AIDS. Behind that debate are vehement disagreements among leading researchers about the most fundamental aspects of what, exactly, a vaccine designed to thwart HIV should do. Now, the prospect of conducting potentially lethal experiments on chimps has sharpened those disagreements.

The spark that rekindled this debate came from one paragraph in a review article about AIDS vaccine progress that Norman Letvin of Harvard's Beth Israel Deaconess Medical Center published in the 19 June 1998 Science (p. 1875). The great apes, Letvin noted, are the only animals other than humans that can be infected by HIV-1. But chimps have been poor models for testing vaccines, he argued, because HIV doesn't replicate well in the animals or appear to make them sick. Inoculating a chimp with a candidate vaccine and “challenging” it with HIV has not provided a rigorous test of whether the vaccine is likely to help humans, Letvin argued. But, he suggested, that may be about to change: A year earlier, researchers at Emory University's Yerkes Regional Primate Research Center in Atlanta, Georgia, reported that an HIV-infected chimpanzee named Jerome had developed an AIDS-like illness.

Unlike nearly 200 other chimps that researchers had infected with HIV, Jerome had a steep drop in CD4 cells—the main immune system warriors that HIV targets and destroys—and a coincident increase of virus in his blood. When Emory's Frank Novembre transfused another chimp, Nathan, with blood from Jerome, the virus again replicated well and in 6 months depleted his CD4s. An HIV isolate subsequently isolated from Nathan, whose health Novembre says is now “going downhill,” also decimated the immune systems of two other chimps. Patricia Fultz of the University of Alabama at Birmingham has seen similar results in three chimps she infected with HIV derived from Jerome.

In Letvin's Science review, he wrote that a stock of this strain “would provide an important new tool for testing vaccine approaches.” Letvin pointed out that scarcity of chimps and the steep fees primate centers charge researchers to use them in HIV experiments—at least $50,000 per animal—would limit their use compared with monkeys, which develop AIDS when infected with either SIV, HIV's simian relative, or a lab-made hybrid of the two viruses called SHIV. Still, a pathogenic strain of HIV adapted to chimps might allow researchers to conduct critical tests to determine whether candidate vaccines could foil infection by an aggressive virus or, failing that, prevent or delay disease. Many AIDS researchers emphatically agree with Letvin's point of view. “In order to really test the efficacy of an HIV vaccine, we really need a disease endpoint in an animal as close as possible to man,” says Malcolm Martin of the National Institute of Allergy and Infectious Diseases (NIAID), who himself has attempted to find an HIV that would cause disease in chimps. Although no vaccine tests are currently planned with the virus, Letvin's suggestion drew a sharp response from Prince, who runs a chimp colony in Liberia, and his co-worker Linda Andrus. In an initial letter, published in the 18 December 1998 Science (p. 2195), they said “the prospect of causing a rapidly progressive and fatal disease in this near-human species is abhorrent.” A more acceptable test of a vaccine, Prince and Andrus wrote, is whether it can prevent a virus from establishing a chronic infection. They pointed out that if a vaccine can block chronic infection, then disease would not occur. “Prevention of disease is not relevant,” they wrote. And, they argued, several nonvirulent strains appear to replicate well enough in chimps to provide a realistic challenge. Now, Prince, Andrus, and their nine new co-authors have taken the argument a step further. In the letter published today, they contend that the virulent strain may be too “hot”: It destroys a chimp's immune system in a few weeks, while in humans the same process typically takes years. This could “seriously jeopardize the HIV vaccine effort” by ruling out vaccines that fail to protect against this strain, but which might be effective against wild-type HIV, they write. And they suggest yet another alternative “challenge” virus: a strain called Han-2, recently described by European researchers, that replicates well in chimps but does not appear to cause disease. This virus should provide a good test of a vaccine's ability to stave off chronic infection, Prince says. Fultz argues, however, that if Han-2 does replicate to high levels in chimps, it will cause disease. She also does not believe that the Jerome-derived strains of HIV are too hot, stressing that infected animals have lived up to 4 years even though their immune systems are damaged. “They're not as lethal as Prince is implying,” says Fultz of the strains. And if Jerome-derived strains do cause disease somewhat faster in chimps than HIV normally does in humans, says Novembre, that has an advantage: “This virus will tell you a lot quicker” whether a vaccine is working. “You don't want to wait 10 years.” Fultz also takes strong exception to Prince and Andrus's statement that preventing disease in the chimp model is not a relevant criterion for judging vaccines. Says Fultz, “That's one of the stupidest statements I've ever heard.” Letvin, too, dismisses the contention. Focusing only on chronic infection might lead researchers to overlook a useful vaccine, he argues. “If we have a vaccine that can make people live decades longer, we need to know that,” says Letvin. The next step in this debate could be a meeting to hash out these issues. Alan Schultz, who oversees AIDS vaccine research at NIAID, says he will do his “public servant best” to organize one. 2. SCIENTIFIC MISCONDUCT # Baylor Saga Comes to an End 1. Jocelyn Kaiser Molecular physiologist Kimon Angelides last week ended a long, costly battle against his former employer, Baylor College of Medicine, which had found him guilty of fabricating data, stripped him of tenure, and evicted him from his lab. On 10 February, Angelides settled a civil suit filed against Baylor and 14 individuals at the university, just hours after a federal appeals board had released a report backing Baylor's findings that Angelides had “committed scientific misconduct.” Angelides has agreed to accept the appeals board's decision and will receive no payment, although Baylor will pay his attorneys$500,000.

“We're quite pleased with the result,” says Baylor lead trial counsel Gerard G. Pecht of the Houston-based firm Fulbright & Jaworski LLP, who says those sued “have been totally vindicated.” The settlement also may bring a measure of relief to officials at other universities, who have worried about being sued simply for following the federal government's requirements to investigate misconduct allegations (Science, 12 February, p. 913). “This kind of suit shouldn't have gotten to this point at all, in our view,” says Allan Shipp of the Association of American Medical Colleges (AAMC).

The saga began in 1992, when a Baylor department chief questioned data in Angelides's grant applications for research on the transmission of nerve impulses through sodium channels. After a 2-year investigation, a Baylor panel found that Angelides had falsified and fabricated figures in five journal articles and five grant applications. In 1997, the Office of Research Integrity (ORI) of the Department of Health and Human Services (HHS) concurred with Baylor's findings and barred Angelides from receiving federal grants for 5 years.

Angelides, who claimed that other scientists in his lab were the ones who had falsified data, appealed the ORI ruling. He also sued Baylor, its president, the seven panelists who examined his case, two former members of his lab, and four others for slander and denial of due process.

A jury had listened to more than a week's worth of plaintiff's testimony when the HHS appeals board released its 171-page report on 10 February. The board, which conducted its own investigation, found that Angelides's “accusations against other researchers were unsubstantiated.” The evidence, the board concluded, showed “not honest error, not disputes in interpretation of data, not preliminary results that later proved overly optimistic, not even carelessness, but rather intentional and conscious fraud.”

According to Angelides's attorney, James Pianelli of McGehee and Pianelli LLP in Houston, “the timing of the [appeals board report] influenced our decision to settle the litigation.” Under the 10 February agreement, Angelides accepts the appeals decision and ORI debarment and will neither appeal nor criticize the decision publicly, will not claim “he has been exonerated or vindicated,” and dismisses all claims against the defendants.

The appeals board's validation of Baylor's findings “certainly says that the system is working properly,” says Barbara Mishkin, a Washington, D.C.-based attorney who specializes in scientific misconduct. But while Baylor came out ahead, the Angelides case may still discourage universities from pursuing misconduct cases—and scientists from serving on review panels, experts say. “It's not reasonable for people to make this very difficult, painful decision and expose their personal assets to risk,” says C. K. Gunsalus, associate vice chancellor for academic affairs at the University of Illinois.

The case may yet leave a positive legacy for researchers. In response to the Angelides affair, the AAMC and ORI have argued that universities and faculty who conduct proper scientific misconduct investigations should be shielded legally from lawsuits. At least, argues Pecht, any civil action should be delayed until the case has been through appeal at HHS. ORI acting director Chris Pascal says HHS “is considering whether additional legal protections are needed in this area,” via legislation or regulation. “Otherwise,” says Pecht, “the inclination may be for some institutions to sweep the problem under the rug.”

3. HUMAN GENETICS

# A Gene That Scrambles Your Heart

1. Michael Hagmann

Building the perfect heart is hard. Each year about 30,000 babies are born with one of the more than 30 different types of congenital heart defects (CHDs), making these the most common of all human birth defects. Despite much searching, until now the genes behind only three rare disorders had been found. But on page 1158, researchers identify a gene that appears to be key to a widespread form of CHD associated with DiGeorge syndrome, which is second only to Down syndrome in causing malformations of the heart.

The findings may finally end a frantic hunt for the DiGeorge gene, which when damaged or missing prevents a proper connection between the outflow of the heart and the main blood vessels and also causes malformations in the facial bones and thymus gland. Surprisingly, the gene encodes a component of the cell's protein degradation machinery, supporting the notion that these “garbage disposal” pathways may be important in organ formation, says the study's lead author, developmental biologist and pediatric cardiologist Deepak Srivastava of the University of Texas Southwestern Medical Center at Dallas.

“This is a major breakthrough,” says developmental biologist Paul Krieg of the University of Texas, Austin. “It opens up a whole new area of research in heart development, because it links a clinical syndrome to a new and exciting pathway in cell biology.” Others are more cautious, arguing that it's still possible that other genes also contribute. “This is beautiful work,” says Christine Seidman, a cardiologist and geneticist at Harvard Medical School. “But I think it's not yet possible to attach the DiGeorge syndrome to a single gene.”

Researchers already knew that in 90% of DiGeorge patients, chromosome 22 is missing a large chunk of DNA—about 3 megabases. This presumably causes the syndrome by eliminating one or more crucial genes, and human geneticists have been trying to pin them down. Srivastava, however, plucked out the key gene not through clinical studies but through basic research—in mice. He and colleagues were studying a transcription factor called dHAND, which turns on an array of genes crucial to the development of the mouse heart. Notes Krieg, “This is a nice example of how basic research can yield clinical answers.”

As the heart takes shape, so-called cardiac neural crest cells migrate from the neural fold (the spinal cord precursor) into specific niches in various tissues. These neural crest cells form the connection between the heart chambers and nearby vessels (see figure)—which are precisely the regions affected in DiGeorge syndrome.

In mice lacking the gene for dHAND, these cells did not develop properly. The researchers picked out a dozen genes normally activated by dHAND by looking for messenger RNAs found in normal mice but absent when dHAND was shut down. One corresponded to a gene called Ufd1 (for ubiquitin fusion degradation), which was infamous for being one of 25 or so genes known to lie within the DiGeorge deletion site.

The link between Ufd1 and the syndrome tightened when they studied the distribution of its protein product in mouse tissues. “Ufd1 showed up in virtually all tissues that were affected by the DiGeorge syndrome,” says Srivastava, including structures that give rise to the thymus and facial bones. Srivastava then turned to humans and found that of 182 DiGeorge patients, all were missing the gene for UFD1. The team also came across one patient who had all the classic symptoms, yet, like 10% of all DiGeorge patients, had no apparent genomic deletion. But after more detailed analysis, the team found a minideletion affecting only two genes, UFD1 and a cell cycle control gene called CDC45. Although he admits that CDC45 cannot be formally ruled out, Srivastava says that “together this indicates that UFD1 is the cause for the 22q11 deletion phenotype.”

Not quite, cautions cardiologist Seigo Izumo of the Beth Israel Deaconess Medical Center in Boston. “UFD1 is the most attractive candidate,” he says, but “it could still be a combined effect of UFD1 and CDC45.” Indeed, the DiGeorge syndrome is probably a game of several players, comments Beverly Emanuel, a human geneticist at the University of Pennsylvania in Philadelphia. “It's clear that UFD1 contributes, but this is not the complete answer,” she says. She notes that there are patients who have genetic disruptions at the suspect region, but seem to have an intact UFD1 gene. “They need to be explained. Clearly there are other things going on at this locus,” she says.

Srivastava, however, is already seeking the proteins that Ufd1 normally helps degrade. Their untimely accumulation when one copy of the gene is missing might somehow cause the developmental problems, he suggests. And Izumo thinks the discovery may even eventually brighten the outlook for afflicted infants, many of whom must currently undergo open heart surgery. “New studies may eventually lead to a better treatment and perhaps even preventive interventions” for those whose hearts need a little help to be made whole.

4. VIROLOGY

# Virus Suspect Identified in Elephant Deaths

1. Dan Ferber*
1. Dan Ferber is a writer in Urbana, Illinois.

When Kumari, the first elephant ever born at the Smithsonian Institution's National Zoological Park in Washington, D.C. was just months old, the youthful pachyderm would frolic for adoring crowds, splashing in the pool or playing with the pumpkins she got on Halloween. But the good times didn't last for Kumari: On a sunny spring day in 1995, after a 5-day bout with a mysterious illness, the 16-month-old Asian elephant lay down and died. At the time, zoo scientists had no idea what had killed the 1000-pound youngster.

But now, on page 1171, a team led by Laura Richman and Gary Hayward of Johns Hopkins School of Medicine in Baltimore and the National Zoo's Richard Montali reports that it has found the killer—a novel herpesvirus distantly related to the virus that causes cold sores in humans. The new virus has killed at least seven other juvenile Asian elephants at zoos.

Exactly how Kumari became infected with the virus is unclear, but it may have been transmitted to her or her mother by an African elephant also kept at the National Zoo. Richman, Hayward, Montali, and their colleagues have found that the same virus that killed the Asian elephants is carried by African elephants. In these animals, however, the virus apparently causes only relatively innocuous skin or genital warts. The researchers also suspect that Asian elephants harbor a virus that is fatal to their African cousins.

By raising the possibility of fatal crossover infections, the work could influence how the world's zoos take care of their elephants, which can no longer be imported from the wild because most populations are dwindling there. “This has tremendous implications for whether or not [zoos] mix these two species,” says Michael Hutchins, director of conservation and science for the American Zoo and Aquarium Association, which oversees a species survival plan for the endangered animals.

The first clues to the cause of Kumari's death came in a postmortem conducted by veterinary pathologists Montali and Richman, who was then at the National Zoo. Their initial exam revealed that Kumari had suffered extensive internal bleeding—a finding that, along with other necropsy findings, “didn't add up to anything we were aware of in elephants,” Montali says. The next day, however, while examining slides of Kumari's tissues under the microscope, they spotted a telltale sign of a virus infection: amorphous inclusion bodies in the nuclei of cells from her blood vessel linings. Further examination under an electron microscope revealed dark, round particles with the expected diameter, about 90 nanometers, of a herpesvirus.

The group followed this lead by hunting for herpesvirus genes in Kumari's infected cells with the polymerase chain reaction (PCR), a sensitive DNA amplification technique. The PCR allowed them to pull out the gene encoding an enzyme, called terminase, that helps to assemble the herpesvirus particle, thus confirming that Kumari had suffered a herpesvirus infection. “Of course, we're thinking [Kumari's] can't be the only case that ever occurred,” Richman says. And indeed, a review of a century's worth of elephant studbook records uncovered 26 suspicious deaths at zoos throughout North America. After collecting preserved tissue from more than 20 long-dead elephants, Richman found that the damaged tissues of seven Asian elephants carried the same viral terminase gene as Kumari, indicating that they had succumbed to the same infection.

To find out where the virus came from, the researchers scrutinized herpesviruses obtained from skin and genital warts of several otherwise healthy African elephants. The viral terminase sequence turned out to match exactly that from the dead Asian elephants—strong evidence that both species were infected by the same virus, Richman says. She suspects that the virus causes only skin and genital sores in African elephants, but becomes lethal when it infects Asian elephants who were not previously exposed.

Other viruses may have made the opposite crossover. In 1996, an 11-month-old male African elephant died at the Oakland, California, zoo with symptoms much like Kumari's. A combination of PCR and DNA sequencing by the researchers showed that the virus that killed him, and one other African elephant, was closely related—but distinct—from the one that killed the Asian elephants. The team suspects that the virus that killed the African elephants originated as a mild strain in Asian elephants, although they haven't shown that directly.

If cross-transmission of the herpesviruses does turn out to be causing the fatal infections, developing vaccines could be one solution. Until then, however, zoo keepers may have to consider keeping Asian and African elephants separated to prevent the lethal disease—a difficult task, Richman says, because some zoos don't have the space or the facilities and would have to build new barns.

But at least there's hope of treating the new disease, now that it has been recognized. Montali, Dennis Schmitt of Southwest Missouri State University, and other team members managed to save an infected female Asian calf in 1997 at the Springfield, Missouri, zoo, by giving her the antiherpes drug famciclovir, in elephant-sized doses.

5. BIOENGINEERING

# Preliminary Data Touch Off Genetic Food Fight

1. Martin Enserink

The controversy in Britain over genetically modified food reached a new high on 12 February, when preliminary data from experiments on potatoes made headlines for the second time in 6 months. The latest media frenzy was touched off when 21 European and American scientists released a memorandum in support of Arpad Pusztai, a protein biochemist who was suspended last year by the Rowett Research Institute in Aberdeen, Scotland, after he appeared on a TV show and sounded an alarm about potatoes altered to resist pests (Science, 21 August 1998, p. 1124). After reviewing the case, the scientists said Pusztai's statements were correct and demanded that the Rowett Institute exonerate him.

Their action immediately prompted members of the British House of Commons to urge a moratorium on genetically modified food and triggered allegations that the government or the biotech industry had a hand in suppressing the data. “This raises questions about the extent to which the biotech industry seeks to permeate every level of government,” says Labour MP (Member of Parliament) Alan Simpson.

The Rowett affair erupted on 10 August 1998, when Pusztai appeared on Granada's TV show World in Action and declared that transgenic potatoes had stunted growth and suppressed immunity in rats that had eaten them for 110 days. The potatoes contained a gene encoding a lectin, a plant protein that can deter insect pests. The world press immediately besieged Pusztai's institute, which initially supported the claim; Rowett chairman and European Parliament member James Provan urged European Union (EU) President Jacques Santer and British Health Secretary Frank Dobson to require more rigorous testing of transgenic food. Just 2 days later, however, the institute's director, Philip James, said Pusztai's data turned out to be “a total muddle”; the disconcerting conclusions, James said, were based on experiments with nontransgenic potatoes spiked with a lectin. The institute apologized for spreading “misleading information,” suspended Pusztai, and turned over his data to a four-member audit committee for investigation.

That committee's report, released on 28 October 1998, didn't mention the alleged mix-up; instead, it acknowledged that experiments with lectin-transgenic potatoes had been carried out, but concluded they did not support the suggestion that the potatoes affected growth, organ development, or immune function in rats. Pusztai, who was forbidden by Rowett to talk to the press, sent copies of the audit report, his own rebuttal to it, and a transcript from the World in Action show to dozens of scientists who had asked for them, asking them to review the material.

The responses, collected by protein chemist Edilbert Van Driessche of the Vrije Universiteit in Brussels, were presented along with a statement last week at a press conference in the House of Commons. The statement contends that Pusztai's data do suggest that the transgenic potato affected the rats' immune systems, affected their organs, and slowed their growth. The data in the audit report, it says, “appeared to be arbitrarily selected and biased towards brushing aside the conclusions of the experimental findings.”

Pusztai's supporters also point to a follow-up study performed last fall by Stanley Ewen—a pathologist at Aberdeen Royal Hospitals who has worked with Pusztai for 10 years—who examined the guts of the rats from Pusztai's experiments under a microscope. Ewen, who presented the results at an EU-sponsored lectin meeting in Lund, Sweden, in November, found that the animals fed a transgenic diet had symptoms of infection, with white blood cells accumulating in their gut lining. The same reaction didn't occur in rats that had been fed a nontransgenic potato diet spiked with the same lectin. Although it's unclear how the diets could have had different effects, “they are profound changes,” says Thorkild Bøg-Hansen, a lectin expert at the University of Copenhagen, “that require further investigation.”

The audit committee's chairman, Rowett senior scientist Andrew Chesson, says he stands by his report but doesn't want to discuss the reviewers' findings, to avoid a debate about raw data in the press. Pusztai should publish his results in a scientific journal, Chesson says: “If the data are sound, I don't think he'd have any problem publishing them.”

The new analyses of Pusztai's data immediately led Simpson to demand a “complete moratorium” on genetically modified food—a measure British Prime Minister Tony Blair said he wasn't ready to take, as he strongly believed the new food was safe. Simpson also says, “If the data are now being corroborated, someone has to explain the basis upon which his research was suppressed.” Several MPs expressed suspicions about the government's role in the affair, which were stirred up even further when a newspaper revealed on 16 February that science minister Lord Sainsbury once had a financial interest in a company that owns a patent on the cauliflower mosaic promoter, a gene often used in plant genetic modification. Conservative MPs said Sainsbury was a biotech “advocate” and demanded his resignation. But Chesson says his institute was not influenced by the government or the industry and has “never ever” attempted to suppress any results. “The sooner the data get into the scientific journals, the happier we'll be,” he says.

Whatever the fate of the findings, most parties agree on at least one thing: The affair has been an outstanding example of how not to communicate scientific findings to an already confused and worried public.

6. COSMOLOGY

# Superheavy Particles From the Big Bang?

1. James Glanz

CHICAGO—Pity the poor Wimps. Although theorists have proposed that these Weakly Interacting Massive Particles—hypothetical slow-moving, exotic relics of the big bang—could account for much of the mass of the universe, no one has conclusively observed a Wimp. Worse, in their bid for the title as the unseen “dark matter” that astronomers believe our galaxy must contain in large amounts, they have to compete with big, brawny lumps of common stuff—stones or gas—that go by the acronym Machos. As if all that were not bad enough for the effete Wimps, a formidable new rival has just emerged: a Godzilla of a particle called the Wimpzilla.

Described here last month at a gathering of the world's leading cosmologists,* Wimpzillas could be millions to trillions of times more massive than Wimps and would have been created even earlier in the mayhem of the big bang. Their large mass means that relatively few of them could account for most of the weight of the universe. “Size does matter,” growls Rocky Kolb of the Fermi National Accelerator Laboratory and the University of Chicago, who presented the work for a team of theorists.

Wimpzillas are as much a figment of theory as their lighter cousins, but for some theorists they're an especially welcome one. They could turn out to be the very same particles that are the linchpins of an effort to explain all the forces of nature in a single framework—a so-called grand unified theory (GUT) –put forth in 1990 by John Ellis of CERN in Geneva, Switzerland, Dimitri Nanopoulos of Texas A&M University, and others. And debris thrown off by Wimpzillas when they decay, as the GUT predicts, might explain the rare, mysterious cosmic rays that slam into Earth's atmosphere at astonishingly high energies (Science, 1 September 1995, p. 1221 and 22 December 1995, p. 1923). Nanopoulos says Wimpzillas have him so excited, “I am almost getting white hair.”

The monster particles emerge naturally in cosmologists' standard creation story, says Kolb. The story begins when the tiniest mote of the primordial emptiness happens to pop into a state called a “false vacuum,” setting loose a tremendous, exponential expansion. The false vacuum has more energy in it than ordinary emptiness, and according to Einstein's equations of relativity, this energy acts like gravity thrown into reverse, driving the expansion—a runaway process called inflation.

Inflation goes on for 10−35 second or so, creating more and more space filled with false vacuum—and nothing else. “There's no radiation. No matter. No House managers. It's a good universe,” says Kolb, in one of the symposium's many tilts at the impeachment proceedings then playing out in Washington. The chilly symmetry of the false vacuum somehow shatters at about 10−35 second, ending the era of exponential expansion. Its energy is converted into an outrushing fireball of particles and radiation—the start of the big bang.

The heat of that fireball could have gone into creating ordinary Wimps, with masses as high as a million times the mass of the proton (106 giga-electron volts, or GeV). They would have been spawned as particles of both matter and antimatter, which would annihilate each other when they meet. But the weak attractions between Wimps and the continued expansion of the universe, which would have swept some Wimps out of harm's way, could have ensured enough survivors to account for the large fraction of cosmic mass—up to 90%–thought to be dark matter.

But now Kolb and his collaborators Antonio Riotto of CERN and Daniel J. H. Chung of the University of Michigan, Ann Arbor, have come up with the heavyweight challenger. Following an example set by Andrei Linde of Stanford University, Lev Kofman of the Canadian Institute for Theoretical Astrophysics at the University of Toronto, Alexei Starobinsky of the Landau Institute in Moscow, and others, they began eyeing the instants just after inflation and before the main fireball, when higher energies—and, hence, higher masses—might be available.

“All of a sudden we have found that this is a pretty rich physics regime,” says Linde—and a rich source of particles. Riotto says that he, Kolb, and Chung soon found several ways to produce superheavy particles. The trio's favorite relies on the pairs of virtual particles that pop in and out of existence in any vacuum, according to quantum mechanics. The “reverse gravity” still in effect at the end of inflation rips any such pair apart, so that instead of meeting up, annihilating, and disappearing, the particles take on a real existence. The Wimpzillas would have been scarce enough to avoid meeting each other and annihilating when expansion slows in the later fireball.

Such particles could be as heavy as 1013 GeV—femtograms, a mass normally in the domain of high-resolution chemistry, not particle physics—so just a smattering of them could account for dark matter. And since annihilation is never a threat, intrinsically weak interactions are not required. “Wimpzillas might be charged,” Riotto says. “They might also have strong interactions.”

But it's a weakly interacting Wimpzilla around 1012 GeV that interests Nanopoulos, because his and Ellis's so-called “flipped SU(5)” GUT long ago predicted a heavy analogue to the proton at about that mass. Their theory predicts that such a particle, which they called a crypton, should decay after a long but finite lifetime, flinging off particles that could slam into Earth's atmosphere as ultrahigh-energy cosmic rays.

While the new calculations have delighted some researchers, they have hit others like a punch in the stomach. Wimps fit naturally into a less ambitious particle theory called supersymmetry, which many physicists favor. Observers might also be feeling queasy, since if rare, lumbering Wimpzillas make up the dark matter, then current Wimp searches (Science, 1 January, p. 13) would have no hope of turning up a signal unless the universe is populated by an even more bizarre mixture of the two particles. And it's only getting worse. Linde and colleagues, for example, say they might have found a way to make relics as heavy as 1018 GeV. Says Linde, “We call our particles fat Wimpzillas.”

• * The Pritzker Symposium on the Status of Inflationary Cosmology, University of Chicago, 29 to 31 January, with a closely related workshop from 1 to 3 February.

7. FEDERAL RESEARCH

# Efforts to Evaluate R&D Found Wanting

1. Jeffrey Mervis

Follow the rules, work together, use outside experts—and don't neglect the young ones. That's the message from a National Academy of Sciences (NAS) panel asked to help federal agencies evaluate their R&D efforts as part of a 1993 law that many research officials have sought help in implementing.

The Government Performance and Results Act (GPRA) requires every federal agency, starting this year, to link its budget to its program goals and explain how it plans to measure progress toward those goals. The exercise has challenged officials at agencies, such as the National Science Foundation (NSF) and the National Institutes of Health (NIH), which support basic research that often may take decades to generate any social or economic payoff. Some officials and scientists have argued that any evaluation is doomed to fail or—worse—that it will force agencies to emphasize trivial results that can be easily quantified.

Nonsense, says the Committee on Science, Engineering, and Public Policy (COSEPUP), a joint panel of NAS, the National Academy of Engineering, and the Institute of Medicine. “Both applied and basic research programs can be evaluated meaningfully, and on a regular basis,” concludes the panel in a report issued on 17 February, for which six federal agencies anted up $300,000. The tricky part is devising the right yardsticks, says the panel, chaired by Phillip Griffiths, director of the Institute for Advanced Study in Princeton. The panel, which met with agency officials after they had prepared their first performance plans to accompany last year's budget request, endorses the use of outside experts. It would be an expanded form of the peer review process that judges individual proposals, with reviewers looking at the quality and relevance of an agency's entire research portfolio. “The research needs to be done well, worth doing, and able to stand up to international comparisons,” says panelist Morris Tanenbaum, a former chair of AT&T Communications. The report suggests that one agency serve as a focus for research supported by many agencies, such as global change or information technology, to make sure that national goals are also being addressed. In particular, the report notes that most R&D agencies played down their training roles when writing up their plans. “The defense and energy departments train the majority of engineers and physical scientists in this country, but those agencies are downsizing and nobody's picking up the slack even though there is heavy demand by industry in some sectors,” complains panelist Mildred Dresselhaus of the Massachusetts Institute of Technology. Academy officials would like to follow up this week's report with a longer study that Congress proposed in legislation passed last fall. It invites the Office of Science and Technology Policy (OSTP) to contract with the Academy “to develop methods for evaluating research programs,” including knowing when to pull the plug. The exercise is also contained in a Senate bill (S. 296) that would double federal R&D spending by 2010. The study has been blocked by White House objections, however. In a 7 October letter to the Senate Commerce Committee, OSTP Director Neal Lane explained that the Results Act already “provides the correct framework for developing performance goals for federal [R&D] activities” and that the new study “would depart from the GPRA approach by mandating alternative forms of evaluation.” Agencies can now request permission to use nonquantitative measures, but some Administration officials see the proposal as an attempt to skirt the current law by substituting Academy criteria for those approved by the White House. Not so, says Michael Lubell of the American Physical Society, one of many professional organizations lobbying hard for the measure. “The Academy study would provide additional options for evaluating research in a way that Congress could use to judge the success or failure of programs,” he says. “Nobody wants to scrap GPRA, but scientists tend to worry when somebody other than an expert in the field tries to judge the quality of their work.” 8. COMPUTER SCIENCE # New Interface Makes Virtual World Tangible 1. Joseph Alper* 1. *Joseph Alper is a writer in Louisville, Colorado. When you explore or manipulate an object in the real world, it helps to use your hands as well as your eyes. Handling a flexible plastic film, for example, requires sensing small pressure variations across the finger tip. Inserting a pin into a small hole can require real-time information about friction and vibration. In the virtual world of computer models and remote-control robots, users generally lack such tactile, or haptic, feedback, which makes delicate manipulative tasks even more difficult. Now physicist Ralph Hollis and graduate student Peter Berkelman, of Carnegie Mellon University's Robotics Institute in Pittsburgh, have developed a new way to bring the sense of touch to computer interfaces: a magnetically suspended joystick that allows a user to manipulate—and feel—objects in the virtual world. It's not the first haptic interface ever constructed, but it is the first to rely on magnetic levitation instead of conventional bearings, which eliminates friction and allows the device to reproduce more subtle tactile feedbacks. “It's quite an impressive interface,” says haptics researcher Edward Colgate of Northwestern University, who tested the device when it was unveiled last November at a haptics symposium.** “You get the complete sensation of both motion and physical interaction in a very real way.” Next May, the device will be shown in action at a robotics conference in Detroit. It works by harnessing the Lorentz force, an effect discovered in the 19th century in which a wire carrying an electric current and immersed in a magnetic field experiences a force at right angles to both the current and the magnetic field. The force can be used to suspend a current-carrying object between two magnets—what Hollis calls Lorentz levitation. It can also cause the suspended object to move when the current flowing through it is changed. Hollis reasoned that the right arrangement of magnets and current-carrying coils could generate force feedbacks that a user grasping the suspended object would sense, and he thought it could lead to a more realistic haptic interface than ones based on motors, wires, and pulleys. “In general, these types of devices haven't been able to achieve the degrees of freedom needed to manipulate an object's motion in space and they have too much friction, which doesn't allow the user to feel anything but the biggest haptic sensations,” says Hollis. To put theory into practice, the Carnegie Mellon group built a device consisting of a joystick handle attached to a bowl-shaped nonmagnetic surface that contains six wire coils. This assembly, called the flotor, is suspended in the air gap between six pairs of permanent magnets mounted inside and outside the flotor on bowl-shaped stationary surfaces. An arrangement of three light-emitting diodes (LEDs) and matching optical position sensors allows the device to sense motion in the flotor. When an operator moves the joystick, each LED-sensor pair records movement in both the ×- and y-axes around that sensor. Combining the output of the three pairs gives six independent measures of movement, corresponding to the six degrees of freedom—×, y, and z, plus pitch, roll, and yaw—needed to describe the motion of any object manipulated in space. Software, analogous to but far more complicated than the device drivers used to translate the motion of a trackball to movement of the cursor on a computer screen, translates the sensors' electrical output into movement of an object in a three-dimensional computer simulation of, for example, a bolt being fitted into the correct hole on a jet engine or a drug docking with its receptor. The interface transmits haptic feedback to the user when output from the computer simulation alters the currents flowing to the six coils, generating forces and torques on the joystick. Crash a virtual object into a virtual wall, for example, and you would feel the impact and recoil through the joystick. The technology allows only a limited range of motion: 15 to 20 degrees—about 2.5 centimeters—in all directions for the current device. “To get a larger range of motion you need a bigger flotor, and that would require much bigger magnets, so the device would get cumbersome,” says electrical engineer Blake Hannaford of the University of Washington, Seattle. “But no device is going to be perfect, and Ralph's is very good.” The next challenge is to improve the software so that the device can interact with a wide range of computer simulation programs. “What we want to create,” says Hollis, “is software that is application independent,” much as it is for a mouse today—software that would allow a user to simply plug in the haptic interface and touch the virtual world. • ** 17th annual Symposium on Haptic Interfaces for Virtual Environments and Teleoperator Systems, sponsored by the American Society of Mechanical Engineers, Anaheim, California, 19 to 20 November 1998. 9. CANADA # Planned Chiropractic Merger Riles Faculty 1. Wayne Kondro* 1. Wayne Kondro writes from Ottawa, Canada. OTTAWA, CANADA—The tape-recorded message on the phone at the Canadian Memorial Chiropractic College (CMCC) explains that chiropractics is a “science and an art and a philosophy. It's about hands-on healing.” But several scientists at Toronto-based York University say it's none of those things. And they have enlisted prominent medical researchers from around the world to help them fight a planned affiliation between CMCC and York, Canada's third-largest university. Last May, York's governing senate approved in principle a plan to offer an undergraduate degree in a field that relies on spinal manipulation, rather than drugs or surgery, to treat disease and improve health. A major component of the arrangement would be a$16 million classroom facility built by CMCC on land leased from the university. Officials view the deal as an opportunity for the university, which lacks a medical school, to expand its presence in the health sciences. “More and more, chiropractics are part of an overall approach to health,” says York's associate vice president for research, Brock Fenton.

But opponents say that keeping company with chiropractors will have the opposite effect. “Surely this kind of affiliation will badly influence our academic reputation,” says chemistry professor Deithard Bohme. Last fall, the university's Faculty of Pure and Applied Science voted against affiliation by a margin of more than 2-to-1, and in recent weeks faculty members have gathered support from such scientific luminaries as Stanford's Arthur Kornberg and Dana-Farber Cancer Institute's Beruj Benacerraf, who have signed onto a statement that “the affiliation would serve to legitimize the unscientific and even antiscientific philosophies and practices associated with chiropractic.”

That hostility already has forced York officials to abandon a plan to make chiropractors members of the pure and applied science faculty. They are now considering creating either a stand-alone entity for chiropractors or one that includes all health-related disciplines. But one leading critic, physics and astronomy professor Michael De Robertis, believes that idea is also flawed because the university doesn't have the expertise to exercise proper oversight of chiropractic training. “They should affiliate with a university with a medical faculty to learn the techniques that will help them become more scientific,” he says, adding that other disciplines “started out being equally voodoo” before becoming respectable by adopting scientific methods.

CMCC President Jean Moss dismisses the controversy as a “hangover from 20 or 30 years ago” and says that recent articles in reputable journals make a strong case for the contribution of chiropractors to a patient's overall well-being. “As an institution, we believe that the body does have the inherent ability to heal itself and that we do assist the body by manipulation of the spine.”

The heated debate could come to a boil next month at a meeting of the faculty senate, where four panels are ironing out details of the planned affiliation on issues ranging from curriculum to research. Opponents want the university's president, Lorna Marsden, to convene an independent group to examine chiropractic's scientific merits. A spokesperson says Marsden has no plans to do so, saying it would circumvent last year's vote. Bohme says he will introduce a motion for such an external review if Marsden declines to act.

Chiropractors are generally trained in independent schools offering a 4-year program, and half the provinces in Canada and most U.S. states recognize chiropractic care in patient health plans. Yet there remains considerable controversy about the efficacy of treatment. A recent editorial in the New England Journal of Medicine by Paul Shekelle of the Los Angeles Veterans Administration Medical Center (8 October 1998, p. 1074) says chiropractics has been shown to provide “somewhat effective symptomatic therapy for some patients with acute low back pain” but that it is inappropriate “as a broad-based alternative to traditional medical care.”

CMCC failed in attempts earlier this decade to link up with three Canadian universities. Moss says the stumbling blocks were unrelated to the quality of care provided by chiropractors and included the logistical problem of absorbing 600-plus students. But provost Penelope Codding of the University of Victoria, which declined a CMCC offer to affiliate, notes that senate records of the decision include a belief that “research in chiropractic at this time is not consistent with contemporary standards of university research.” A spokesperson for Calgary University, which also weighed affiliation, says president Terry White “cannot comment” on the school's experience with CMCC.

York's Fenton contends that the union will strengthen the university through greater diversity. “You have to be broad-minded about what goes on in universities,” he says. But critics don't buy that argument. “It's irresponsible of York to do this,” says Montreal-based pediatric practitioner and McGill instructor Murray Katz, who has criticized chiropractics as an expert witness in coroner's inquests and civil lawsuits. “It's a treatment in search of a disease.”

10. # From Junk Bond King to Cancer Crusader

How Michael Milken helped transform the research and funding of prostate cancer by applying the lessons of Wall Street to the support of science

Paul Lange and Leroy Hood ran up against a familiar problem in 1995, when they were trying to identify genetic changes that occur in cancerous prostate cells: They desperately needed to locate and study families with several living prostate cancer victims—a tall order, since this cancer is typically diagnosed late in life. Frustrated by the glacially slow process of tracking down families through men who came in for treatment, the University of Washington, Seattle, scientists turned to an unusual ally: financial wizard and prostate cancer survivor Michael Milken.

Milken—a crusader for prostate cancer research since he was diagnosed in 1993 with an advanced case of the disease—came up with what Hood thought was “a hare-brained idea.” Lange recalls that Milken said, “No problem, we'll just go on Larry King Live” to ask for volunteers. When King's producers at CNN were leery of such a depressing subject, Milken says he phoned some high-level contacts at the network. “I've known Larry for a long time and have been involved in the financing of CNN,” he explains. And so, on 13 November 1995, King interviewed Milken, Hood, and prostate cancer survivor General Norman Schwarzkopf. More than 3000 people called in. In just 3 weeks, Hood and Lange signed up nearly 300 families—as many as most institutions typically enroll in several years.

Once called the most powerful man in American finance, before he was jailed for securities-law violations in 1991, Milken is a man with a mission. He has put the energy, contacts, and ambition that once earned him billions on Wall Street into finding a cure for prostate cancer, the second leading cancer killer among U.S. men. When he's not pulling strings to get the message out (see sidebar), his foundation has been bankrolling research, cajoling companies, and trying to break down barriers between institutions. Milken has brought to the field a survivor's sense of urgency, vast amounts of cash, and a knack for remaking organizations. And in the process, he is shaking up the culture of the prostate cancer research community.

Within 1 month of his diagnosis at age 46—when he was given less than 18 months to live—Milken set up a foundation formally called the Association for the Cure of Cancer of the Prostate, but widely known as CaP CURE. Observers say it has raised both the public and scientific profile of prostate cancer research. Indeed, Milken has achieved for prostate cancer what a well-organized coalition of breast cancer survivors has done for that disease: “CaP CURE has put prostate cancer on the map,” says cell biologist Joy Ware of the Medical College of Virginia Campus of the Virginia Commonwealth University, in Richmond.

In a sense, the foundation, which to date has given $65 million to prostate cancer research, is Milken's own grand experiment: an effort to bring a business approach to the task of curing prostate cancer. “Everything I learned in business, in financing companies, I've tried to bring to bear here,” he says. Quick decisions and a sharp focus on a cure are hallmarks of that approach. Last month, for example, dozens of prostate cancer researchers cashed CaP CURE checks for a total of$7.5 million in grants they had applied for only a few months previously. Half of these grants relate directly to treatments for advanced prostate cancer. The foundation is sponsoring more than 80 clinical trials, and next month top clinicians will gather in New York City to hash out new clinical trial protocols at a CaP CURE meeting.

## The bottom line

Perhaps because of Milken's frustration with his own limited treatment options, CaP CURE zeroes in on cures rather than basic research. The idea is both to create new treatments and speed them to patients, hastening the pace of clinical trials, says Howard Soule, the foundation's scientific director and acting executive director.

As a first step in this acceleration, Milken wanted to overcome a lack of coordination—even competition—between cancer centers, which often left researchers working only with colleagues in their own institution. So in 1996, he offered leading medical centers $300,000 each in annual funding—but only if they all participated in planning and conducting joint trials. Most foundation-sponsored trials take place within this consortium, which now has 10 centers* that receive a total of$2.5 million from CaP CURE each year. The hospitals' computer systems are linked, thanks to a donation Milken secured from software giant Oracle (he had helped CEO Larry Ellison finance the company), so researchers can check on patient accrual and protocols at any center. In a typical month, several phase II trials are being jointly run by two or three centers, says Soule.

Although big pharmaceutical companies tend to arrange trials themselves, executives of smaller biotech companies say they are eager to work with the group. “It would be like doing a trial with the who's who of prostate cancer in the United States,” says immunologist Frank Valone of Dendreon Corporation in Seattle. Dozens of companies have been and are collaborating with consortium members to test potential therapies, many of which come out of industry labs. The foundation doesn't fund companies directly, but hooks them up with academics who have CaP CURE grants and can therefore shoulder some of the cost of the trial, indirectly lowering costs for the biotech firms, explains Valone. “We're doing one [trial] this year that we otherwise wouldn't have,” he says.

And even though major pharmaceutical companies are less involved in the consortium's trials, collaborations extend to their researchers too. “CaP CURE has made it possible for us to interact, to set up the exchange of reagents, and they serve as a matchmaker for potential relationships,” says Blake Neubauer, an endocrine pharmacologist at Eli Lilly in Indianapolis.

To provide a venue for fostering more informal connections among academics and companies, Milken again harked back to his business glory days. He was famous for hosting an annual junk bond seminar known as the Predator's Ball, where corporate raiders, companies, and investors were entertained by the likes of singer-actress Diana Ross. Now, he sponsors an annual scientific retreat, where researchers from biotech and academia attend sessions that can last 12 hours.

The networking continues at a reception at Milken's Lake Tahoe mansion with such entertainers as singer-songwriter Paul Anka. Many researchers describe this invitation-only, all-expense-paid meeting as simply the best in prostate cancer—or any field. “It's like a Gordon Conference, only more intense,” says Coffey.

Observers say all this activity has indeed helped speed potential therapies for prostate cancer. According to an industry survey of pharmaceutical companies (see graph above), since 1993, the number of drugs in development has more than doubled, to 36. “CaP CURE has had a major role” in this proliferation, says Coffey. “They've funded almost every one of those developments.” And compared to the early 1990s, when most new drugs were versions of anti-androgens, today's candidates are much more sophisticated and diverse, he adds. “From gene therapy to monoclonal antibodies, there is really interesting stuff coming,” Coffey notes. “And there's more on the way.”

## No more business as usual

As activity in prostate cancer swells, other funders are picking up on some of Milken's business approaches. The Juvenile Diabetes Foundation, for example, will begin handing out next month the first of its new, 2-year, $100,000 “rapid response grants,” reviewed in just 3 months. And this year, NCI will offer$10 million through a program called Rapid Access to Intervention and Development—an attempt to speed academic discoveries into drug trials, just the sort of emphasis that CaP CURE prizes. NCI is also planning a new consortium for clinical trials, with prostate cancer as one of the pilot projects, which has aims similar to those of CaP CURE's and will share many of the same features. However, it will be more open than CaP CURE's therapy consortium. “It's national; anyone can come into it,” says NCI's director, Rick Klausner.

Indeed, some researchers are bothered by the exclusivity of CaP CURE's meetings and consortia. For example, its prostate tissue banks at the moment are open only to the four universities inside the foundation's Genes and Family Studies Consortium, which Hood runs. CaP CURE also relies on a hand-picked coterie of specialists as grant reviewers—and much of the money is awarded to their home institutions. “Some people have said that it's an old-boy network, and to some extent they're right,” admits Holden, who nevertheless says that the institutions involved with CaP CURE are “topnotch” by any measure.

And although everyone is happy to receive a check 3 months after submitting a grant, some researchers grumble about the limited 1-year duration of the awards. “A year makes people a lot more nervous,” says Coffey, and makes it difficult to hire staff.

Meanwhile, although biotech executives praise Milken's promotional and fundraising skills, some are leery of getting too close to him. “Mike Milken, being the head of CaP CURE—and a convicted felon for stock fraud and manipulation—is not someone you can entertain as a supporter of a small start-up company,” says virologist Daniel Henderson, CEO of Calydon in Sunnyvale, California. “You can't get quality investors to invest alongside him. So you have to stay away.”

Most substantively, some researchers fault the foundation's focus on late-stage disease. Neither the NCI nor the American Cancer Society (ACS) takes that approach. “We like basic research, because time has shown that's where everything has come from,” says Dawn Willis, ACS's director of research promotion and communication.

Holden and Milken are the first to admit that their efforts haven't paid off yet for patients in the most dire straits. “You can't say that a patient diagnosed with advanced prostate cancer in 1999 is going to live longer than he would have in 1993,” says Holden. Still, many researchers say that CaP CURE has hiked the chances that new therapies will emerge sooner rather than later. “They've seeded the area,” says Neubauer. “I think there are some signs that it's starting to bear fruit.”

Hood and Lange would agree. After analyzing DNA from families recruited through Larry King Live, they have identified a gene that seems to predispose 15% of those men to prostate cancer. They've also uncovered more than 200 molecules that are overexpressed in cancerous prostate cells, and they're talking to pharmaceutical companies about trying to use these to diagnose and even treat the disease. If they and others succeed, Milken's entrance into cancer research may turn out to have been his most important business venture.

• * Columbia University, New York City; Dana-Farber Cancer Institute-Partners Care, Boston; Johns Hopkins Oncology Center-Brady Urological Institute, Baltimore; Memorial Sloan-Kettering Cancer Center, New York City; University of California, Los Angeles, School of Medicine; University of California, San Francisco-Mount Zion Cancer Center; University of Michigan Cancer Center, Ann Arbor; University of Texas-M. D. Anderson Cancer Center, Houston; University of Virginia Cancer Center, Charlottesville; University of Wisconsin Cancer Center, Madison.

11. # Michael Milken's World

WASHINGTON, D.C.—One thing you quickly notice about Michael Milken is how easily he works large sums of money into the conversation. Take the following scenario: Sitting at a dinner table with several reporters at a posh hotel here, on the eve of a major cancer rally, he's holding forth on cancer research: “For just $20 billion, the government could test every known chemical compound against every known cancer cell line in 2 years,” he argues. Tanned and healthy-looking at 52, prostate cancer survivor Milken talks nonstop about big ideas, ranging from his prostate cancer foundation CaP CURE (see main text), to how Congress could conquer all cancer, to his plan for drastically altering the American diet. Thinking big is a Milken trademark. He rose to fame—and infamy—in the 1980s by pioneering the use of junk bonds (high-risk and potentially high-yield bonds) for corporate takeovers. This kind of paper money financed huge expansions at companies such as CNN and MCI, and it catapulted Milken into powerful circles, reaping him enormous riches–$550 million in 1 year alone. But in 1989, Milken pled guilty to six counts of securities-law violations, was fined $1.1 billion, and was sentenced to 10 years in a low-security federal prison in California. After serving only 2 years, Milken was released early. The next day, he was diagnosed with advanced prostate cancer. Since that day, Milken has capitalized on his contacts in business and the entertainment industry to bring attention and funds to prostate cancer. Celebrities such as Bill Cosby, a member of CaP CURE's honorary board of directors, liven up foundation events. Arnold Palmer and other major sports figures help out with charity golf tournaments and baseball games. Last year, singer Rod Stewart had dinner guests dancing onstage at New York City's Waldorf Astoria Hotel, raising$3.5 million for CaP CURE. Throwing good parties is Milken's “personality, his history,” says prostate cancer researcher Don Coffey of Johns Hopkins University, who advised Milken during the setup of his prostate cancer foundation but has no direct involvement in it. “I'm sure he was involved in financing everyone [at the Waldorf bash],” he says, “and he's called in all his chips.”

Though Milken's foundation focuses mostly on medical research aimed at finding a cure, he's also concerned with prevention. His latest crusade is diet: He thinks low-fat, high-soy meals could slash cancer rates, and he's promoting his self-published cookbook with everyone from Martha Stewart to Barbara Walters. At the elegant but low-fat Washington dinner, for example, he points out that the mushroom soup, prepared by his personal chef in the hotel's kitchen, was creamed with barley. As always, Milken eventually brings the argument back to dollars, noting that America spends $100 billion on cancer treatment each year. Anything that prevents that expenditure, of course, only makes sense. 12. SPACE SCIENCE # To Mars, En Masse 1. Oliver Morton 1. Oliver Morton is writer in Greenwich, U.K. A fusillade of probes heading to Mars in the next decade marks a new era in space exploration PARIS—William Boynton labored for 8 years on the gamma ray spectrometer for Mars Observer, the first instrument he had ever built for space. Then disaster struck: the ill-fated spacecraft was declared missing, presumed dead, as it neared its target in August 1993, taking with it not just Boynton's instrument but a host of others. The$845 million mission had been the first attempt that U.S. scientists had made for Mars in almost 2 decades. When it vanished, the blow was crushing. “It was almost like a death in the family,” recalls Boynton. “People in the building were uncomfortable. They didn't know what to say to us.”

When Boynton first heard that copies of the Observer instruments were going to be sent back to Mars on a series of missions-on-the-cheap—with his slated to be on the last, due for launch in 2001—he thought it was a crazy idea. But a vibrant new Mars program has made Boynton, and many of his colleagues, converts. At a conference* here earlier this month hosted by the French space agency CNES, scientists described one new Mars mission after another, some major campaigns—the cluster of missions intended to bring samples of the planet back to Earth in the middle of the next decade—and others tiny. Current plans foresee 20 spacecraft making the voyage before 2010, of which five are en route, and a series of “micromissions,” which could add many more (see table). The probes will scrutinize everything from subsurface ice to the edge of the atmosphere. No single setback can stymie all the new missions; there are more eggs in more baskets than ever before.

This is not just a matter of more Mars missions; it's a different approach to exploration. The missions are small, the risks are high, and the pace of innovation is quick. For example, the plans for carrying out the sample-return missions, of which the first is just 4 years away, have been completely rewritten in the past 6 months. To some, this continuous evolution is thrilling and satisfying: “It makes the whole program a hell of a lot more robust,” says Cornell University's Steven Squyres, principal investigator on the 2003 Surveyor rover mission. Others are unsettled by the ever-evolving effort's breadth. “With the whole variety of different things which are being put forward on the smorgasbord table right now, I don't think things have been sufficiently focused to identify the highest priority goals,” says Gerald Wasserburg, a geochemist at the California Institute of Technology in Pasadena.

View this table:

Adding to the ferment is the increasingly international nature of the effort. While the lion's share belongs to the United States—Russia's program is so diminished that no Russians even made it to Paris—there are Japanese and European missions, and, above all, a critical role for France. CNES is now an integral part of NASA's efforts. Negotiated over the past year, the CNES-NASA deal is a huge boost for both sides: French science minister Claude Allègre satisfies his desire to steer his country's program from manned to unmanned spaceflight, while the United States gets someone else to pick up some of the tab for the sample-return program. “They bring a very high level of commitment and a great deal of technical capability,” says NASA's Carl Pilcher. “From a program with very narrow reserves they've allowed us to convert to something much more robust.”

According to the latest sample-return plan, the United States will gather the samples and the French will bring them back. The idea is that NASA rovers launched with companion landers every 2 years or so will pack about 40 samples of rock and soil into 15-centimeter-wide canisters mounted on simple three-stage solid-rocket boosters just big enough to put such a canister into a stable orbit round Mars. (The booster was originally designed for the U.S. Navy 40 years ago as a minimal response to Sputnik: Brian Wilcox, the son of that project's manager, happens to work at NASA's Jet Propulsion Laboratory and realized that it would do the sample-return trick quite nicely.)

The French will bring the samples back to Earth, with an orbiter and four “netlanders” to be launched in 2005, on an Ariane 5, which will also carry the second NASA rover. (The Ariane 5, which is considerably bigger than the U.S. launchers that NASA could afford within its Mars exploration budget, is a crucial part of France's contribution.) With the help of radio beacons and laser guidance, the orbiter is meant to pick up both sample canisters deposited in Mars' orbit by the landers; if you remember the spacecraft James Bond's nemesis, Blofeld, used to swallow up other satellites in You Only Live Twice, you have the general idea. The orbiter will then fire its rockets and head back to Earth, dropping off the sample return canisters over Utah.

Ariane 5 is also vital to another strand of Mars exploration. When it launches communications satellites, the booster often has spare capacity that can be used to lift small payloads—and with the help of a lunar swing by, these payloads can be slung on a Mars trajectory. The U.S.-French team is planning a series of micromissions to get to Mars this way. The first two, included in NASA's 2000 budget request, will be a communications relay satellite to help Mars missions talk to Earth and to each other, and an aircraft that will, if all goes according to plan, fly along some of Valles Marineris, Mars's Grand Canyon, on the centenary of the Wright brothers' first flight at Kitty Hawk. Yes, this is a stunt—but one with some scientific rationale, as the aircraft could identify the composition of the thinly layered deposits that make up the canyon's sheer walls.

Other micromissions discussed at the meeting include derivations of the basketball-sized Deep Space 2 penetrators, which will be hurled at the face of Mars by the Polar Lander spacecraft just before the main craft soft-lands on the planet this December. Four pairs of these probes, each sharing a carrier spacecraft, could fit into the hitchhiker slots on a single Ariane 5—enough to form a network of probes for monitoring Martian seismic rumblings. The standard designs and cheap piggyback launches could keep the cost of such missions to as little as 30 million each. Such small numbers give the heady feeling that almost anything is possible. But even the leanest missions have a way of growing. The first sample-return flight, for example, was scheduled for 2001, but it had to be delayed to 2003, because the '01 lander had become too complicated and expensive. The '03 lander is also larded with extra instruments to characterize the problems that could face manned missions—how bad is the radiation, for example, and how inconvenient or toxic is the dust. One engineer puts it this way: The program planners “give you an eight pound bag and tell you to get ten pounds into it. You explain, and they come back with a ten pound bag and ask you to put 13 pounds into it. When you complain they look shocked and say ‘but we gave you a bigger bag!’” Some old and wise heads worry that the effort lacks a guiding strategy. “The number of missions and the extent to which we can achieve substantial goals is of concern to me,” says Wasserburg. In particular, he sees the experiments aimed at paving the way for human exploration as “absurd.” In a decade, he fears, we could end up with a lot of small triumphs but a number of big scientific questions still unresolved. If that were to be the case, it would be disappointing—but far less disappointing than the silence from Mars Observer. 13. MEETING: MICROBIAL GENOMES III # DNA Sequences Provide Grist for Microbiologists 1. Elizabeth Pennisi CHANTILLY, VIRGINIA—AT THE MICROBIAL GENOMES III MEETING HELD HERE FROM 29 JANUARY TO 1 FEBRUARY, BIOLOGISTS SHIFTED THEIR FOCUS FROM SEQUENCING TECHNOLOGY TO THE EXCITING NEW BIOLOGICAL INSIGHTS MICROBIAL SEQUENCES ARE PROVIDING. ## Bacterial Partners for Filaria When Barton Slatko signed on with the Filarial Genome Project, he expected to be determining the genetic makeup of a major human parasite. Filaria worms infect some 120 million people worldwide, causing the tropical diseases African river blindness and elephantiasis. But the work of Slatko, a molecular biologist at New England BioLabs Inc. in Beverly, Massachusetts, took an unexpected turn after the project's team found that some filaria worms are infected by an odd bacterium called Wolbachia. The bacterium is best known for infecting insects, where they are transmitted from generation to generation through the germ cells. Depending on the species, it can make the host infertile or alter the sex ratio of the offspring, such that only one sex survives. In contrast, in filaria the Wolbachia seem key to fertility, thereby making the worms susceptible to antibiotics that attack the bacteria. The bacteria “seem to have a pretty interesting role in terms of therapy,” Slatko reported. So he and his colleagues have decided to sequence a Wolbachia genome, hoping it will lead to new ways to control filarial infections. Because of the importance of these worms as both human and animal parasites—they also cause widespread infection in livestock—the World Health Organization set up a consortium in 1994 to map the genome of the filarial worm Brugia malayi and look for genes that might provide targets for new antifilarial drugs and vaccines. But as consortium members began analyzing the worm's genetic material, “we kept finding bacterial sequences,” says David Guiliano, a collaborator from the University of Edinburgh, Scotland. After confirming that the microbial DNA was not contamination, the researchers found earlier reports suggesting that the nematodes might harbor some unknown kind of bacteria. Starting 25 years ago, a few researchers had noticed “dense bodies”—possibly bacteria—in the embryonic and adult tissues of several species. One report even suggested that antibiotics seem to have antiworm effects. But until recently, “that work was not followed up on,” says Guiliano. Then in 1995, Claudio Bandi from the University of Milan, Italy, and his team reported that they had detected ribosomal DNA sequences indicating that the worms are infected by Wolbachia bacteria. Last year, Bandi, Edinburgh's Mark Blaxter, and their colleagues confirmed these data: They surveyed 10 filarial worms, finding Wolbachia DNA in nine of them. Based on their analyses, they concluded that the bacteria had long ago parasitized the nematodes. Although Wolbachia take a toll on insect fertility, “there's no evidence that they are parasites” in nematodes, says Guiliano. Instead, the bacteria seem to contribute to the nematode's reproductive success. Evidence for that comes from Bernhard Fleischer and his colleagues at the Bernhard Nocht Institute for Tropical Medicine in Hamburg, Germany. They reported in the January Journal of Clinical Investigation that the antibiotic tetracycline kills the bacteria living in the reproductive tissue of nematodes found in mice. The treatment resulted in smaller, infertile nematodes, but it had no effect on another nematode species that lacked signs of bacterial infection. “If you kill the Wolbachia, you will [eventually] kill the [infected] nematodes” by preventing their reproduction, explains Scott O'Neill, a vector biologist at Yale University in New Haven, Connecticut. And that's good news for filarial researchers, as it may provide a new way to try to interrupt the worm's life cycle. “They had been banging away at the nematodes without turning up much new, and then Wolbachia came on the scene,” O'Neill points out. Now, “the worm [research] groups are all latching onto [it].” The sequence may lead to better anti-Wolbachia drugs, which in turn may be more effective in killing the worms. The New England BioLabs's effort is likely to come out with the first Wolbachia sequences, as Slatko expects to be finished within 6 months. Filaria researchers can hardly wait, says Alan Scott, a molecular parasitologist at Johns Hopkins University in Baltimore who is part of the Filarial Genome Consortium. Anticipation of the sequencing results is creating “quite a bit of excitement.” ## Supermicrobe Decoded If any bacterium can be considered a supermicrobe, it is the one called Deinococcus radiodurans. Not only does this soil bacterium live on Antarctic granite, but it can survive doses of radiation a thousand times the lethal level for humans. After each assault, it simply stitches together its shattered chromosomes. At the meeting, biologists announced that they have completed sequencing the Deinococcus genome and uncovered one more surprise. Despite its super powers, Deinococcus is really a microbial Clark Kent, genetically much like other bacteria. The 3-million-base genome, decoded by molecular biologist Owen White and his team at The Institute for Genomic Research (TIGR) in Rockville, Maryland, indicates that Deinococcus owes its extreme radiation resistance to the same repertoire of mechanisms for repairing DNA found in other organisms. It just has more of them than most other life-forms. The genome also suggests how the microbe evolved its radiation resistance, an ability that could help cope with humanity's own messes. Other results presented at the meeting suggest that genetic engineers may be able to equip this hardy organism with genes that could enable it to degrade toxins and clean up metals at radioactive waste sites. The TIGR researchers started work on the Deinococcus genome in 1995, but the going was slow because it has a high proportion of guanine and cytosine bases, as well as many repetitive regions. Despite that, “TIGR has done a good job of it,” says Richard Alm, a microbiologist at Astra Research Center in Boston, Massachusetts. Now, the team's analysis is revealing the surprisingly conventional nature of the microbe's radiation defenses. One element of these defenses is an enzyme called MutT. Radiation damages cells in part by generating reactive forms of oxygen that oxidize key cellular compounds, including some of the nucleotide building blocks of DNA. These oxidized nucleotides can cause faulty DNA replication, but MutT protects against such mutations by helping rid the cell of the oxidized nucleotides. Most organisms have a single MutT gene, but with 20 MutT-like genes, Deinococcus is capable of “removing a whole lot of oxidative products,” White said. The sequence also offers a clue to how Deinococcus repairs the numerous breaks that radiation induces in its DNA. This repair requires that the microbe put the fragments back together in the right order. In 1995, Michael Daly and Kenneth Minton, molecular microbiologists at the Uniformed Services University of the Health Sciences, Bethesda, Maryland, had suggested that the microbe, which usually carries multiple copies of each of its three chromosomes, rebuilds the copies in parallel. This would be possible, they proposed, if the chromosome copies were aligned. The nearby fragments of one copy could thus provide information missing from the other, and complete chromosomes could be assembled (Science, 24 November 1995, p. 1318). But the genome seems to support a different picture, which first began to emerge when Daly did experiments with artificial DNA inserted into Deinococcus chromosomes. He showed that the fragments form circles before they build reconstituted chromosomes. When White and his colleagues realized that the Deinococcus genome contains some several hundred repeating stretches of DNA, they wondered whether these repeats were involved in making circles during repair. One possibility is that early in repair, a repeat at the end of a fragment loops around and links to one at the opposite end, forming a circle and thereby protecting the broken ends of the DNA. Alternatively, these stretches of repetitive DNA may help keep DNA fragments from copies of the same chromosome close by, enabling them to link up more readily and eventually reconstitute a full chromosome. But, Daly cautions, “it's early days,” and the roles of both repeats and the circles are still unclear. Why the organism would have evolved this Phoenixlike ability to rebuild itself after radiation exposure has been a mystery, but its lifestyle and its genome offer some clues. Desiccation damages DNA, and an analysis of the genome by Daly, Minton, and Kira Makarova of the National Center for Biotechnology Information hints that Deinococcus gained some of its all-round toughness by adapting to harsh, dry environments. The researchers showed that Deinococcus has genes for at least three desiccation-resistant proteins that thus far have been found only in plants. It also has other genes usually seen in eukaryotes and not in bacteria. “It seems whole families [of genes] were transferred [from other species],” Makarova suggested, although more sequenced genomes will be needed to confirm this. While some researchers mine the Deinococcus sequence for clues to the microbe's capabilities, others have been looking for ways to exploit its hardiness. At the meeting, Minton described the success of Daly's group in providing Deinococcus with genes that allow it to detoxify a common environmental contaminant—mercury. Daly and Hassam Brim in his lab showed that neither the mercury-altering genes nor another set of newly added genes for breaking down the chemical toluene interfere with the microbe's resistance to radiation. The engineered Deinococcus might thus serve to detoxify sites doubly contaminated with chemicals and radioactive wastes. “If we can take mixed waste and reduce it to a [pure] radiation waste, then we can handle it,” explains Marvin Frazier, a microbiologist with the U.S. Department of Energy. White is eager to see his supermicrobe go to work. “It's able to withstand a lot of environmental insults,” he emphasizes. “This is going to really be an industrially relevant organism.” 14. PHYSICS # Could Charge Stripes Be a Key to Superconductivity? 1. Robert F. Service Seen in ceramic superconductors, stripes are confounding many theorists but exciting a few as a possible clue to how these materials work Takashi Imai wears his stripes reluctantly. Like hundreds of other physicists around the world, Imai has been struggling to understand how an odd family of ceramics manages to conduct electricity without resistance at unprecedentedly high temperatures. Along with most other physicists studying high-temperature superconductivity, he was inclined to discount an idea that has hovered at the fringes of the field for several years—that charges percolate back and forth through these materials in a series of stripes, a few atoms wide, creating a landscape that helps current flow effortlessly through the material in all directions. “I was very skeptical [of the stripe theory],” says Imai, a young assistant professor at the Massachusetts Institute of Technology. The evidence for the stripes was patchy, and the theory “looked too simple to be true.” But that was before last summer, when Imai and his students began running a set of experiments for months on end that showed clear hints that charges were indeed running in defined lanes. By November, the group knew they were onto something big and resolved to double check every detail. “We kept running experiments 24 hours a day, 7 days a week. I skipped Thanksgiving and Christmas to keep taking data,” he says, adding casually, “This is a competitive field.” Competitive is an understatement. Deciphering the mystery of high-temperature superconductivity has been the prime obsession among condensed matter physicists since 1986, when the first superconducting ceramics were discovered. A definitive answer remains elusive. But Imai's discovery along with a couple of other recent reports is giving stripe proponents a big boost. At meetings around the world, “one of the main themes we're seeing is stripes,” says John Kirtley, a superconductivity researcher at IBM's T. J. Watson Research Center in Yorktown Heights, New York. “More and more people are starting to believe it.” But believing in the reality of the charge stripes is one thing; conceding that they have anything to do with superconductivity is something else entirely. The new evidence has by no means quenched the debate about the stripes' significance; indeed, it seems to show the stripes remaining fixed in the material like the stripes on a flag—a behavior that nearly everyone agrees should kill superconductivity, not promote it. As David Pines, a superconductivity theorist at the University of Illinois, Urbana-Champaign, and the Los Alamos National Laboratory in New Mexico, puts it, “There may be some evidence for stripes. But that doesn't say in any way, shape, or form that this helps produce superconductivity.” Seeing stripes. Early support for stripes came primarily from theorists, who were struggling to understand how the electrical and magnetic behaviors of the ceramic superconductors interact. The superconducting ceramics all share a layered structure, with sheets of copper and oxygen atoms sandwiched between layers of other atoms, such as yttrium, barium, strontium, and lanthanum. Each of the oxygen atoms in the copper-oxide sheets harbors an even number of electrons, which have “spins”—magnetic orientations—pointing in alternate directions. The alternating spins cancel each other out, so the atoms carry no net spin. The copper atoms, by contrast, have an odd number of electrons, and thus have an excess electron spin, giving these atoms a net magnetism. In these materials, opposite spins attract: When one electron's spin points up, its neighbor prefers to point down. Adjacent copper atoms do their best to align in opposite orientations, creating an alternating up-and-down pattern. This large-scale pattern tends to hamper the movement of free charges, which are introduced when researchers dope these “cuprate” materials with other atoms. Added to the lanthanum-based cuprate as a dopant, strontium atoms replace lanthanums, which normally donate electrons to the electron-hungry atoms in the adjacent copper-oxide planes. But strontium has one less electron than lanthanum to donate. The upshot is that copper atoms wind up with electron vacancies, or “holes.” These holes, it turns out, can percolate through the material, carrying a positive charge but no spin. In the early 1990s, theorists such as Vic Emery at Brookhaven National Laboratory (BNL) in Upton, New York, and Steven Kivelson at the University of California, Los Angeles, realized that as holes moved around, they would change the lattice of spins in the material. As a hole moves to a new copper atom, that atom's spin moves to the hole's previous location, disrupting the orderly up-down arrangement of spins. “That costs energy and tends to frustrate the movement of the holes,” says BNL experimental physicist John Tranquada. Stripes could form as the holes' way of getting around this restriction, he explains. Add enough holes, he says, and “if they clump together in a stripe then they can move along that stripe more easily without moving any spins.” The regions in between, meanwhile, hang onto their favored “antiferromagnetic” order of up, down, up, down spins. Though hints of stripes in the cuprates went back to the early 1990s, researchers had been unable to get a good look at them. Tranquada and his BNL colleagues suspected that the stripes were elusive because they were moving around, blurring their signature in the data. So the BNL group looked for a way to fix them in place. They added neodymium to a superconducting ceramic made with lanthanum, strontium, copper, and oxygen, in hopes that this dopant would alter the shape of the lattice just enough to keep the stripes from moving. The researchers then hit their sample with a beam of neutrons, which can probe the fine-scale magnetic structure of a material: Neutrons have an intrinsic spin that interacts with—and can reveal—the spins of electrons. The neutron scattering data suggested that these ceramics do indeed have stripes of magnetic order alternating with stripes rich with electrical charges. Pairing up. Encouraged by the result, Emery and Kivelson went further. They proposed that fluctuating stripes not only help charges flow through these ceramics but are crucial to high-temperature superconductivity itself. One hallmark of superconductivity is that electrical charges—either electrons or holes—skate through the material in pairs rather than singly, which keeps them from scattering off atoms in the lattice. And a major challenge in high-temperature superconductivity theory is explaining what promotes this pairing. Emery and Kivelson think it's the stripes. By confining charges to narrow regions, the magnetic barriers in effect raise the energy of the holes. To lower this energy, the holes work to spread out. Single holes can't readily manage the task. But by pairing up, the holes can work together to modify the arrangement of spins in the barriers and tunnel through, lowering their overall energy. Once formed, the pairs can move effortlessly through the material as a supercurrent. Theorists, who tend to favor their own pairing theories, have been slow to embrace this picture. And nearly everyone, including Emery and Kivelson, agrees that stripes should kill rather than promote superconductivity unless they can meander through the material like a winding river that often changes course. Due to a quantum mechanical effect, a fixed, parallel arrangement of stripes would conspire to pin down all the charges in the material, creating an insulator. “Just about everyone would agree that if you really localize the charge in stripes, you will not get superconductivity,” says Tranquada. Fortunately for stripe proponents, several early reports did find evidence of fluctuating stripes in the lanthanum-based cuprates. More support came from work reported last fall in Nature (8 October 1998) by a team of U.S. and British researchers led by Herb Mook, a neutron scattering expert at Oak Ridge National Laboratory in Tennessee, which also suggested fluctuating magnetic stripes in a different superconductor, made from yttrium, barium, copper, and oxygen, or YBCO. And even Tranquada's initial experiment could be taken as support for the idea that fluctuating stripes favor superconductivity. When the group added neodymium to pin down the stripes, the temperature at which the lanthanum-based material superconducted dropped sharply, from around 38 K to just a few degrees above absolute zero. But other experiments have begun to muddy this pretty picture and suggest that fixed stripes may not always be superconductivity killers. In work recently submitted to Physical Review B, a team of Japanese and U.S. researchers looked at a lanthanum-copper-oxide crystal with excess oxygen added as the dopant, a material that has the highest superconducting temperature of any lanthanum-based compound. The team's neutron scattering results clearly showed the presence of static stripes. The inescapable conclusion, says team member and MIT physicist Robert Birgeneau, is that “superconductivity and static magnetic order [fixed stripes] can coexist.” Now comes Imai's latest result, which underscores that puzzle. Because neutrons are strongly sensitive to the magnetic spins on the copper atoms, most neutron scattering work had revealed just the arrangement of magnetic spins in superconductors, leaving investigators to infer the charge stripes. Imai, however, has traced the charges themselves in the high-temperature superconductor lanthanum-strontium-copper-oxide. In a technique called nuclear quadrupole resonance, the team pulsed radiofrequency waves at the material and tracked the magnetic response, which indicated the spins of the copper nuclei. But nearby electrical charges also affect the magnetic signature, which allowed the researchers to piece together the location of the charges as well. Imai says the results, which have been submitted to Physical Review Letters, support the presence of “quasi static” stripes, largely fixed in place. “I didn't expect we'd see this phenomenon,” says Imai. “So I was very surprised.” So how can fixed stripes and superconductivity be present in the same hunks of ceramic? One possibility, says Princeton theorist Philip Anderson, is that any seemingly fixed stripes and superconductivity may be confined to separate regions of the material. Emery suggests another: The stripes aren't completely fixed after all, but meander about an average position. His reading of the recent data suggests that “there is some movement there,” he says. Whatever the answer, it appears that stripes are here to stay. They may be either a key to superconductivity or a false lead devised by nature to throw theorists off the track. To find out, says Pines, researchers now need to show that stripes not only coexist with superconductivity in the same region of material, but that they somehow improve its superconducting behavior. Until someone figures out how to pull off that experiment, the charge stripes will remain a disquieting mystery. 15. CLIMATE CHANGE # Big El Niños Ride the Back of Slower Climate Change 1. Richard A. Kerr After two “El Niños of the Century” in 15 years, climate researchers are finding explanations in long-term climate change Climate modelers were patting themselves on the back last year after successfully anticipating the arrival of El Niño in 1997. In the spring of that year, dutifully following predictions of a modest event, the tropical Pacific warmed sharply (Science, 24 April 1998, p. 522). But then the Pacific asserted its independence, confounding the models by soaring to record-breaking warmth in one of the most severe events of this kind on record—worse than the devastating “El Niño of the century” that struck in 1982–83. By the time it was over in 1998, El Niño-related weather extremes had caused 23,000 deaths and33 billion in damages around the world. Now, by deconstructing the symphony of longer-term climatic cycles that play out in the Pacific Ocean, researchers have found clues to why these events were so severe.

In two papers soon to appear in the Journal of Climate, researchers show that these other, slower cycles of ocean warming and cooling have tended to be at or near their peaks—in some cases unusually high peaks—since the 1970s. By preheating the Pacific, they boosted the intensity of the El Niños.

No one knows what drives these cycles, although some researchers suspect that the past century's global warming may have helped push some of them to the warm side simultaneously. But identifying the cycles at work in a given year could be a first step toward forecasting how powerful a predicted El Niño will be—and ultimately how it will affect weather patterns around the world (see sidebar). “If we are to extract every ounce of predictability from the [climate] system,” says oceanographer David Enfield of the National Oceanic and Atmospheric Administration (NOAA), “we must try to understand how the other components work and how they interact with [El Niño] in modifying our climate.”

Untangling climate to understand its workings and future behavior takes a record long enough to include repeat performances of a given climate oscillation. Enfield, of NOAA's Atlantic Oceanographic and Meteorological Laboratory in Miami, Florida, and oceanographer Alberto Mestas-Nuñez of the University of Miami found theirs in a 136-year record of sea-surface temperature measured by ships around the globe, just compiled by oceanographer Alexey Kaplan and his colleagues at Lamont-Doherty Earth Observatory in Palisades, New York.

The Miami researchers parsed the complex climatic symphony of the Kaplan temperature record into its component parts. First, they removed the so-called El Niño-Southern Oscillation or ENSO variations themselves, the high-pitched drone of tropical warmings and coolings that return every 2 to 7 years. Next, they removed the bass crescendo of global warming—a trend of nearly half a degree Celsius per century, which may be driven by human activity. Finally, they used a sophisticated statistical technique to separate the remaining melody of the temperature record into several oscillations—a decade or two long and a half degree or more in amplitude—that were being played out in different parts of the ocean.

What they found was more of a cacophony of discordant cycles than a harmonious melody. In the eastern tropical Pacific, the researchers found that temperature swings slowly from warm to cold and back over 10 to 20 years. This oscillation, called an “interdecadal ENSO” because it is also located in the eastern tropical Pacific, began a sharp swing to the warm side in the late 1970s and reached a peak in the early 1980s, helping to drive the 1982–83 El Niño into record territory. When this shift was first recognized, some researchers suggested it was driven by greenhouse warming (Science, 28 October 1994, p. 544). But the Miami analysis shows that the interdecadal ENSO has been waxing and waning through the whole record—although its last comparable peak was in the 1860s.

Another cycle, unnamed, throbs to its own decadal beat in a large westward-pointing wedge of the central tropical Pacific. This cycle reached its greatest warmth in the early part of this century, but it, too, took a sharp turn toward the warm side in the late 1970s. And in the North Pacific, two distinct climate seesaws, each following its own rhythm, raise the temperature of the central part of the basin while lowering that of the rest of the ocean and vice versa. The more rapid of the two, which Enfield and Mestas-Nuñez equate with a well-known climate variation called the Pacific Decadal Oscillation, also showed tropical warming in the late 1970s. When El Niño struck in the early 1980s, all these cycles added up to trouble. “What's really spiking the '82-'83 event is the decadal-multidecadal variability,” says Enfield.

That seems to be what happened in 1997, too. In an independent analysis of a shorter sea-surface temperature record from 1955 through 1997, meteorologists William Lau of NASA's Goddard Space Flight Center in Greenbelt, Maryland, and Hengyi Weng of SAIC/General Sciences Corporation in Laurel, Maryland, found a similar combination of long-term global warming and decadal variations that “adds up to a huge [temperature] increase in '97,” says Lau. Their work is also in press in the Journal of Climate. And in their most recent analysis, Enfield and Mestas-Nuñez also see a convergence of decadal variations pushing the latest El Niño to its peak. “We can confirm what Lau is saying,” says Enfield.

All this is good news for those who'd like a bit of warning before a giant El Niño strikes. Because the oceans change slowly and these are longer-term cycles, forecasters can assume that the state of the oceans won't change too much in a year, so they can hope to warn of big El Niños 6 months to a year in advance, says Enfield. For even longer-term predictions, however, researchers will need to better understand what's driving the decadal fluctuations—and those answers may be a long time coming.

Although researchers have a variety of theories, no one really knows for sure what pushes the Pacific into these oscillations. It may be a feedback loop among winds, currents, and temperature. Occasionally, such loops could produce simultaneous warming in several cycles, according to recent work by modeler Gerald Meehl and his colleagues at the National Center for Atmospheric Research in Boulder, Colorado. But Lau and others say that global warming could be enhancing all the natural decadal cycles at once. “We've been looking at these phenomena as being separate,” says oceanographer Michael McPhaden of NOAA's Pacific Marine Environmental Laboratory in Seattle, Washington, “but we're beginning to realize there are these connections.”

These uncertainties mean that forecasting the likelihood of big El Niños 10 years out probably isn't in the cards, says Enfield. But by analyzing decadal climate variations, he and others are at least beginning to learn to follow the Pacific's tune.

16. # In North American Climate, a More Local Control

1. Richard A. Kerr

El Niño doesn't reign alone. Other cycles of ocean temperature sometimes intensify this warming of the tropical Pacific Ocean (see main text). And it now seems that at least one other ocean cycle wields influence over some of El Niño's worldwide effects: a decades-long temperature seesaw in the North Pacific.

El Niño, for example, often gives Minnesotans a break from their normally brutal winter cold. But not always, climate researchers Alexander Gershunov and Tim Barnett of Scripps Institution of Oceanography in La Jolla, California, have found. If the climatic mood of the North Pacific isn't right, Minnesotans may not get their usual respite, although other areas may be spared El Niño-related weather extremes. “They're onto something interesting and useful,” says meteorologist Nathan Mantua of the University of Washington, Seattle. “It should improve the skill” of long-range winter weather predictions.

In the classic El Niño winter pattern, tropical warmth displaces storm tracks in midlatitudes, leaving the northern contiguous United States warmer and drier and the south wetter and colder. But Gershunov and Barnett found that it's not that simple when they sorted a 61-year record of U.S. winters by the condition of the North Pacific. Every few decades or so, this basin swings from being unusually cold in its western and central parts and warm in the east to the opposite pattern, a cycle called the Pacific Decadal Oscillation, or PDO (Science, 10 July 1998, p. 157).

The researchers report in the 19 January issue of Eos that when the central North Pacific was cold, the El Niño pattern was stronger and more consistent. When the PDO was in its opposite phase, the El Niño winter pattern weakened. In the winter of 1997–98, for example, a powerful El Niño was felt even more keenly in North America because the PDO was in its cold phase, suggest Gershunov and Barnett.

They also found that the PDO interacted even more strongly with El Niño's opposite number, La Niña, a cooling of the tropical Pacific that currently holds sway. A cold North Pacific disrupted the La Niña pattern of a dry Southwest and wet northern states, and a warm North Pacific strengthened it. Because the PDO switched phases from cold to warm late last summer, Gershunov says, it should now be reinforcing the current La Niña, bringing chill and snow back to Minnesota.

That seems to be happening, with the Southwest being especially dry, although it's too early in the winter to be sure. Indeed, meteorologist Martin Hoerling of the University of Colorado, Boulder, cautions that Gershunov and Barnett's record is relatively short given the PDO's decades-long swings, so the correlation may only be chance. Minnesotans should still check the thermometer, rather than the PDO, before venturing out.