News this Week

Science  11 Aug 2000:
Vol. 289, Issue 5481, pp. 842

You are currently viewing the .

View Full Text

Log in to view the full text

Log in through your institution

Log in through your institution


    New Arenavirus Blamed for Recent Deaths in California

    1. Martin Enserink

    Hemorrhagic fever. The words conjure up images of people in far-off places dying horrific deaths from diseases like Ebola and Lassa fever. Now, researchers say a similar viral disease is endemic—but at very low levels—in the western United States. Last week, the California Department of Health Services announced that a recently discovered virus carried by wood rats and pack rats killed a 14-year-old girl in April; moreover, the department says, there's strong evidence that the virus has caused at least two other deaths within the last 14 months.

    The agent, called the Whitewater Arroyo virus, probably infects humans when they inhale aerosolized rat urine. Hantaviruses, another family of rare viruses carried by rodents, infect people by a similar route. Although the evidence isn't airtight yet, researchers say two of the three patients' symptoms—high fever, internal bleeding, and liver problems—suggest that the new disease is indeed a hemorrhagic fever. If the Whitewater Arroyo virus is the culprit, that would be true to form: The virus is an arenavirus, a family that includes Lassa as well as several South American viruses known to cause hemorrhagic fever.

    “People don't need to be alarmed,” says Carol Glaser, an infectious-disease physician at the state's Viral and Rickettsial Disease Laboratory in Berkeley. “We think this is a rare event.” At this stage, researchers don't know how widespread the virus is: So far, it has been found in rodents in six western states. Nor do they have any idea how many people infected with the virus go on to develop disease. Even so, Glaser and others suspect that more people may have succumbed to the disease in recent years—which is why they plan to go back and look at blood and tissue samples from patients who died of unexplained illnesses that bear some hallmarks of hemorrhagic fever.

    Some arenaviruses are treatable with a drug called ribavirin, raising hopes that future patients may be saved if the disease is diagnosed in time. “It's not like with hantavirus, where you can only sit by the bedside and hope to God they get through it,” says Charles Fulhorst, a virologist at the University of Texas Medical Branch (UTMB) in Galveston, who worked with Glaser to nail down the new virus as the culprit. “We may actually be able to do something.” In addition, other arenaviruses have been known to cause secondary infections when hospital workers come into contact with blood from infected patients, which can occur easily because they're often bleeding extensively, says Robert Tesh, also of UTMB. Patients with hemorrhagic fever- like symptoms should probably be handled with extra care, says Tesh, who has studied South American arenaviruses such as Guanarito and Junin, which cause Venezuelan hemorrhagic fever and Argentine hemorrhagic fever, respectively.

    North America had seemed blissfully free of dangerous arenaviruses until now. Although there are at least 14 arenaviruses known in South America, four of them deadly, the only previously discovered North American member of the family—the Tamiami virus, found in cotton rats in South Florida in 1970—doesn't seem to make anyone sick. But in 1996, Fulhorst, then at the Centers for Disease Control and Prevention (CDC) in Atlanta, discovered the Whitewater Arroyo virus—named after the site in New Mexico where it was first found—in several rat species of the genus Neotoma. Over the past 3 years, Fulhorst, working with rodentcontrol officials in several states, discovered the virus in western Oklahoma, Colorado, Utah, Texas, and California. Explains Fulhorst: “The next question then became, Does it also make people sick?”

    The first clue came last year, when Glaser's lab was asked to test a blood sample from a middle-aged woman from Riverside County, east of Los Angeles, who had died in June 1999 of a disease resembling hantavirus pulmonary syndrome. She tested negative for hantavirus, as well as for dozens of other infectious agents. The case might have gone into the record books as another undiagnosed fatality if it hadn't been for Michelle Jay, a veterinarian at the lab who happened to be familiar with Fulhorst's work. She suggested that he test the sample for the arenavirus. Using the polymerase chain reaction, Fulhorst soon found traces of the virus's genetic material in the woman's blood, suggesting she died from an infection. Fearing the test might be a false positive, Fulhorst and Glaser decided not to go public yet.

    Stronger evidence came when a 14-year-old girl in Alameda County, in Northern California, died with similar symptoms in April. In her blood samples, Fulhorst not only found traces of the virus's RNA that matched that of the Whitewater Arroyo virus found in Californian rats; he also isolated the virus itself—sufficient proof, Fulhorst says, that it was indeed the culprit. And just when the team was preparing to announce those findings in June, a 30-year-old woman died of similar symptoms in Orange County. Again, Fulhorst found genetic traces of the virus; he is now trying to isolate the agent itself from the samples—a task that could take a couple of weeks. But with one case nailed down and strong evidence in two others, the researchers did not want to wait to go public. “A new patient may be walking into the hospital next week,” says Fulhorst.

    “From what I've seen, the data look pretty compelling,” says Thomas Ksiazek, acting chief of CDC's Special Pathogens Branch. Although the disease doesn't seem to pose a big public health threat, he says, CDC plans to step in by developing new diagnostic tests and disseminating them to state labs, and by putting together a case definition so that doctors will know what to look for when they see possible new patients. The agency will also assist in finding out how widespread Whitewater Arroyo infections may be by looking at old blood and tissue samples from California and other states, says Ksiazek. Meanwhile, the one confirmed case emphasizes the importance of avoiding rodent-infested places or taking precautions to prevent exposure to rat urine—a CDC warning already in place since the Sin Nombre virus, a hantavirus, was discovered after an outbreak in 1993. Since then, more than 250 people have come down with hantaviruses.

    The new finding also drives home a point infectious-disease experts have been making for years: If they have the resources to monitor for new viruses and track their distribution before they cause disease, researchers are much better prepared when an outbreak does occur. During the 1993 Sin Nombre outbreak in the Four Corners area, says Fulhorst, no one had a clue what the causative agent was, and a small army of epidemiologists, microbiologists, and toxicologists was sent in to find out in a rush. Only afterward did researchers realize that the Sin Nombre virus had been around, occasionally killing people, for years. “With this one, it's exactly backward,” says Fulhorst. “We said we thought there may be arenaviruses out there; we showed they were in the rats; and now we're saying they're causing disease.”

    There may be much more to discover this way. During his rodent surveys, Fulhorst says he found a third North American arenavirus, this one in deer mice in California. The agent, whose discovery hasn't been published yet, has been called the Bear Canyon virus; whether it makes people sick is, at this point, anybody's guess.


    Monsanto Donates Its Share of Golden Rice

    1. Dennis Normile*
    1. With reporting by Elizabeth Pennisi.

    Monsanto Co. has agreed to provide royalty-free licenses to speed up work on a genetically modified rice that could alleviate vitamin A deficiency around the world. Researchers welcomed last week's announcement, but warn that a thicket of intellectual property claims surrounds the technology and that significant legal hurdles remain before the rice can become widely available to farmers in developing countries.

    “Monsanto is the first company with the good will to offer this technology free for humanitarian purposes,” says Ingo Potrykus, a plant molecular biologist at the Institute for Plant Sciences of the Swiss Federal Institute of Technology in Zurich. “I hope I can use this to convince other companies to give up their intellectual property rights,” adds Potrykus, who developed the variant in collaboration with Peter Beyer of the Center for Applied Biosciences at the University of Freiburg in Germany (Science, 13 August 1999, p. 994; and 14 January, p. 303). Monsanto CEO Hendrik Verfaillie says his company, a subsidiary of Pharmacia Corp., is taking this step “to minimize the time and expenditure associated with obtaining licenses needed to bring ‘golden rice’ to farmers and the people in dire need of this vitamin.”

    Potrykus's team transferred into rice the multiple genes needed to create a synthesis pathway for β-carotene, which the body converts into vitamin A. It marked the first time a trait requiring multiple genes had been transferred into a plant. The enriched rice, which has a golden hue, promises to help alleviate a widespread public health problem: vitamin A deficiency, which afflicts 400 million people and can lead to vision impairment and increased susceptibility to disease.

    Before the promise can be realized, however, the β-carotene pathway must be transferred to the different varieties of rice grown in each region. Potrykus's group has tried to hasten the process by offering to make its germ plasm and related technology freely available. The offer drew interest from several national rice research labs, as well as the International Rice Research Institute (IRRI) in the Philippines.

    But claims of intellectual property rights have kept the technology bottled up in Potrykus's lab. Two outside studies have found that between 25 and 70 proprietary techniques and materials are involved in the gene transfer. Potrykus says Monsanto's most important claim is on the 35S promoter, which boosts the expression of the genes introduced into the rice. That change makes possible a high level of β-carotene synthesis.

    But other important pieces of the puzzle lie outside Monsanto's control. Incorporated into the golden rice, for example, are two “bits of DNA” acquired, says Potrykus, under separate agreements that require the donors' consent to be passed on to third parties. The agreements have prevented him from distributing the germ plasm to any of the rice breeders interested in working with it.

    Potrykus has spent a year negotiating with the two companies that hold rights to the DNA to enable the technology to be used by public-sector research groups. He declined to identify the companies to avoid jeopardizing what he calls a “delicate stage” of the negotiations, which he hopes will be concluded “within 2 months.” Potrykus and others hope that Monsanto's announcement—made at an agricultural biotechnology symposium in Chennai, India—will set a precedent for companies with a stake in the necessary technologies. “It may encourage others to work to cut through some of the red tape,” says Ronald Cantrell, IRRI's director general.

    Last week Monsanto also announced progress on its earlier promise to make public a draft of the sequence of the rice genome unveiled this spring in a collaboration with Leroy Hood, then at the University of Washington, Seattle, and now president of the Institute for Systems Biology in Seattle (Science, 14 April, p. 239).* The data have already been transferred to Japan's Rice Genome Research Program (RGP), which is the lead agency for the International Rice Genome Sequencing Project (IRGSP). The Japanese group will pass the Monsanto data to other IRGSP members once legal issues are resolved. Takuji Sasaki, RGP director, says that the Monsanto data, although “rough,” should hasten completion of the sequencing project, whose status will be discussed next month at a meeting in South Carolina.


    Gates Foundation on Big Funding Spree

    1. Michael Hagmann

    For Eleanor Riley, an immunologist at the London School of Hygiene & Tropical Medicine, it must have felt like Christmas in July. The source of her midsummer cheer: $40 million from the Seattle-based Bill and Melinda Gates Foundation. “I'm absolutely delighted. This is at least 10 times [the amount of grant money] I would have expected for my entire career,” says Riley, who last Monday received the funding for a 5-year project to develop and test new ways of fighting malaria.

    But the grant for Riley and her colleagues was only one slice of the high-calorie funding cake—worth almost $200 million in all—that the cash-brimming Gates Foundation dished up for scientists in various fields late last month. Other beneficiaries are tuberculosis specialist Jim Yong Kim of Harvard Medical School in Boston, who received almost $45 million to develop a program to control multidrug-resistant tuberculosis (MDR-TB), and Alfred Sommer of the Johns Hopkins University School of Hygiene & Public Health in Baltimore. His $20 million grant “came at a very critical time,” Sommer says. “We are just in the process of starting—and can now scale up—four large field projects in Nepal, Bangladesh, India, and Zanzibar” to study how cheap vitamin and mineral supplements can reduce maternal and child mortality in developing countries.

    The grants for research into malaria and TB are the third big chunks of money the Gates Foundation has lobbed into the fight against these major killers within the past year. The foundation—one of the world's largest science-funding philanthropies, with assets of more than $20 billion—kicked off its spending spree in mid-July at the XIII International AIDS Conference in Durban, South Africa (Science, 14 July, p. 222), by announcing several AIDS/HIV-related grants totaling $90 million.

    With the new money, Riley and her colleagues intend to expand research into new drugs and insecticides and to set up centers of excellence in malaria-endemic areas of Africa. “We hope the Gates money is acting as some sort of catalyst to bring other partners on board,” Riley says. Meanwhile, Kim and his team—in collaboration with the World Health Organization, local health authorities, and other partners—are gearing up to develop a multidrug treatment program for MDR-TB patients in Peru. “This will have an enormous impact,” Kim says, noting that the treatment protocol will be adaptable to other developing countries.

    View this table:

    Even greater largesse may be in store. At the July G-8 meeting in Japan, leaders of the world's economic heavyweights resolved to halve the death toll of malaria, TB, and HIV by 2010. “There are rumors that the European Union will announce a major new investment this fall,” Kim says. Until the noble words are backed by cash, the Gates money is paving the way.


    Rescue Droids Stumble in an Urban Jungle

    1. Mark Sincell*
    1. Mark Sincell is a science writer in Houston.

    AUSTIN, TEXASBroken bodies were strewn on the floor like discarded rag dolls. Severed legs protruded from under shattered furniture. Twitching fingers pushed hopelessly at a pile of bricks. Somewhere, a baby cried. Agonized moans drifted up from the basement, while unseen fingers tapped feebly on metal pipes. Then, clambering through the wreckage, came the robots.

    It wasn't the set of Terminator 3, but the floor of the Austin Convention Center. There, in the midst of the 17th Annual National Conference on Artificial Intelligence,* four teams of engineers fielded mechanical contestants in the first annual urban ruin search-and-rescue competition—a simulated catastrophe created to test intelligent lifesaving robots that may one day lead rescuers to people trapped in the precarious rubble of collapsed buildings. The competition indicated that the technology still has a way to go.

    When a building tumbles down, rescuers have about 30 minutes to start searching for survivors, says Robin Murphy, a competitor and robot expert from the University of South Florida, Tampa. After that, the chance of survival drops precipitously. Flesh-and-blood rescue teams must wait 3 hours for the unstable building to settle before going in. Expendable robots, however, could get to work immediately, sniff out the survivors, give them food and first-aid supplies, and lead the rescue team straight to them.

    Such robotic St. Bernards would have to be both agile and clever—mobile enough to climb over wreckage, smart enough to do it without human help. Radio control is too unreliable, and the clutter of a collapsed building makes tethers impractical. So rescue robots must solve one of the hardest problems in artificial intelligence: traversing a changing three-dimensional maze. To do that, an array of light, heat, and ultrasonic sensors feeds environmental information to the robot's computer brain. The robot then combines this with internal data about its moving parts to build up an internal map to guide it through the maze.

    That's the theory, at least; the test course showed how hard it is to put it into practice. Designed by engineers from the National Institute of Standards and Technology (NIST), it resembled an Ikea showroom in hell. Meter-high walls of blue particle board and Plexiglas delineated several rooms in three increasingly realistic devastated buildings. Overturned chairs, tables, and a bed were spread through the spacious rooms of the easily navigable “yellow” course. The more difficult “orange” course was clogged with rubble piles and featured a loft reachable by a ramp, stairs, and a rope ladder. The extremely difficult “red” course was just an unstable pile of boards, aluminum siding, cinder blocks, chicken wire, and plastic tubing. Realistic plastic victims with internal heaters were hidden throughout the courses like gruesome Easter eggs.

    In the last round of the competition, the robot finalists had 20 minutes each to conquer the course. Designs varied widely. Two contestants—from the University of Arkansas and the University of Kansas—mimicked the rolling-trash-can look of the “droid” R2D2 from Star Wars. The eventual winner, from Swarthmore College in Pennsylvania, resembled a Shop-Vac vacuum cleaner ringed with ultrasonic and infrared sensors and topped by a rotating camera. The brainy appliance mapped one room of the yellow course—and found one victim by crashing into him—without human help. But its small wheels faltered in the obstacle-strewn orange course.

    The fourth team, Murphy's Floridians, fielded a pair of rugged radio-controlled robots, one resembling the tracks of a 61-centimeter-long bulldozer and the other an oversized Tonka truck. Watching for obstacles through the robots' camera “eyes” (a process Murphy compares to “driving while looking through a straw at mouse-eye level”), two graduate students guided the pair all the way to the end of the red course. Along the way they found several victims and a local TV news camera operator.

    In the end, none of the robots came close to locating all the victims, and some wandered out of the course to “rescue” unsuspecting conference participants. “The teams aren't doing all that well,” admitted contest organizer Alan Schultz after the preliminary rounds. Schultz emphasizes, however, that robots historically perform poorly in the first year of contests like the urban ruin course, but that engineers rapidly rise to the challenge. “This contest will be an inspiration to the community,” he says. “I think in 5 years there is a reasonable chance that these robots will be able to work in real structures.”

    Lisa Meeden, a computer scientist at Swarthmore College who helped develop the winning robot, agrees that tough challenges are best. If the course is too easy, “you can fool yourself into thinking you have solved the problem when you haven't,” she says. “We wanted to push the envelope.”

    • *30 July to 3 August 2000.


    Political Scientist Becomes Cause Célèbre

    1. Robert Koenig

    BERN A right-wing political leader's successful defamation suit against one of Austria's most prominent political scientists, Anton Pelinka, has become a lightning rod for an international debate about academic freedom of speech. This suit and a second filed against Pelinka represent “a cynical effort to punish and suppress Professor Pelinka for his professional expression,” contends Irving Lerch, director of international affairs for the American Physical Society, who sent a letter last month to Austrian President Thomas Klestil backing Pelinka's right to free speech.

    On 11 May, a criminal court in Vienna found Pelinka guilty of defaming former Freedom Party leader Jörg Haider and fined him roughly $4500, although the court will not impose the fine unless Pelinka is convicted of making similar statements in the future. Haider, an outspoken populist who has made questionable comments about Nazism, accused Pelinka of defaming him for asserting in an interview on Italian television in May 1999 that some Haider statements had “trivialized” Nazism.

    Haider is pursuing a second defamation suit against Pelinka for comments he made to CNN last year. (A decision is expected in October.) He has also sued two other Austrian intellectuals, including Wolfgang Neugebauer. Neugebauer is a history professor who directs the Archives of Austrian Resistance, which documents the country's resistance movement during the Nazi era.

    Pelinka, a tenured University of Innsbruck professor who heads Vienna's Institute for Conflict Research, told Science that he is appealing the ruling to a higher court and, if necessary, will take his case to the European Court of Human Rights. At least two human rights conventions signed by Austria guarantee freedom of expression without undue interference by public authorities.

    In the 3 months since the judgment, an array of academic and human rights groups have rushed to Pelinka's defense. “Everything Pelinka said was consistent with normal public discussion about political figures in a democracy,” asserts Aaron Rhodes, executive director of the International Helsinki Federation for Human Rights, which is supporting Pelinka's case. Lerch's letter to President Klestil—sent on behalf of the scientific freedom and responsibility committee of the American Association for the Advancement of Science (AAAS, the publisher of Science)—states: “We are worried that the judicial system may be exploited for purely political purposes to intimidate scholarship and restrict freedom of expression.”

    Lerch, who chairs the AAAS panel, says his committee is concerned about possible political influence in the court case, because the man who is now Austria's minister of justice, Dieter Böhmdorfer, had represented Haider when the defamation lawsuit was filed last September. However, an Austrian government spokesperson says that Böhmdorfer withdrew his name from that Vienna law firm when he became justice minister in February—after Haider's Freedom Party became a partner in Austria's new center-right coalition government—and says there is no evidence that he sought to influence the case.

    Pelinka—a prolific author on comparative political science and the winner of a 1998 award for his leadership in criticizing neo-Nazism—says he worries about the impact of such defamation judgments on the willingness of untenured professors to speak candidly on controversial issues. Fearing government reprisals, some academics might toe a less risky line to avoid losing out on tenure, as most Austrian universities are government institutions. “I'm not afraid to say what I think,” Pelinka says. “But such rulings could have a chilling effect on the willingness of my younger colleagues to speak out.”


    Long-Sought Protein Packages Glutamate

    1. Laura Helmuth

    Among neurotransmitters, two stand out as stars, communicating most of the brain's urgent messages. These fast-acting, ubiquitous chemicals—GABA and glutamate—send the basic “stop” and “go” signals that most other neurotransmitters merely modulate. Glutamate is called into action wherever rapid-fire excitatory signals are needed—say, for vision or learning. For decades, researchers have been looking for the protein that packages glutamate for express delivery to other neurons. On page 957, Robert Edwards of the University of California, San Francisco (UCSF), and colleagues report that they've found this elusive transporter. “This has been a long time coming,” says neurobiologist Marc Caron of Duke University, who points out that many labs have been on the protein's trail.

    To prepare for launching from one neuron to another, neurotransmitters have to be stuffed at high concentrations into bubbles called synaptic vesicles. When it's time to send a signal, these vesicles fuse with the axon wall, and the neurotransmitter within bursts into the space between the cells. Vesicular transporters do the stuffing—they are proteins embedded in the vesicle wall that pump neurotransmitters, which are built in the cytoplasm, into the bubble.

    Two other transporter families have been identified to date: the one that escorts GABA and another that waltzes monoamines such as dopamine and serotonin into their respective vesicles. But the glutamate vesicular transporter had defied such efforts, possibly because there's so much glutamate floating around the cell that it's tricky to pinpoint proteins that act on it, says Caron.

    Edwards's team actually worked with this long-sought protein, now dubbed VGLUT1, for a few years before realizing its true identity as a glutamate transporter. Other researchers had pegged it as a phosphate transporter. But there were hints that it had greater powers: A mutation of a closely related protein in the worm Caenorhabditis elegans appeared to disrupt glutamate transmission. And mutations of another related protein in humans cause acid to build up in cellular bubbles called lysosomes. The UCSF team originally suspected that the protein might, through its ability to transfer phosphate, modulate some other protein's ability to pack glutamate into vesicles. They tracked the protein down to vesicle walls, suggesting that it was a key player in hustling chemicals in and out of these packages. As evidence mounted that the protein was more than a phosphate transporter, “we had to change our thinking,” recalls Edwards.

    To test whether VGLUT1 could transport glutamate into vesicles itself, Edwards's team, which includes Elizabeth Bellocchio, Richard Reimer, and Robert Fremeau Jr., inserted the gene that codes for the protein into some cells of a standard cell line. They then caused the cells to build mock vesicles and tested how much glutamate got into these chambers. Cell lines that contained VGLUT1 packed two to four times as much glutamate into their model vesicles as did those without. “I think their characterization is very convincing,” says Caron.

    Still, there are a few loose ends. Glutamate buzzes throughout the brain, but VGLUT1 is much more localized. Even so, says Edwards, this doesn't disqualify it as the long-sought glutamate transporter: Another, closely related protein shows up in the areas of the brain where VGLUT1 is absent. He thinks the two carry out similar functions but are distributed in different parts of the brain.

    Now that they've identified the glutamate vesicular transporter, the UCSF researchers hope to figure out how it works. True transporters actively escort neurotransmitters into a vesicle, pulling them uphill against the gradient between the tightly packed neurotransmitters inside and the low concentration outside the vesicle. It's a slow process, but it can pack in more chemicals than the alternative, a channel. Channels essentially open up part of the vesicle wall, enabling chemicals to surge in, attracted by a charge or pH gradient. Strangely, the glutamate transporter appears to have properties of both: It packs glutamate into vesicles both extremely tightly and extremely quickly. “We're thinking about all sorts of models right now,” says Edwards.


    Hunt for Mad Cow in Sheep Reassuring

    1. Michael Balter

    Experts on brain-riddling spongiform diseases have grown steadily more uneasy over signs that so-called mad cow disease, linked to a lethal human illness, may be lurking in sheep. Not only would that open up a new front in the battle to purge bovine spongiform encephalopathy (BSE) from livestock, but it would also suggest that far more people—at least those who eat lamb, anyway—are at risk of contracting the human killer, variant Creutzfeldt-Jakob disease (vCJD). A study in the 10 August issue of Nature now offers evidence that BSE is not rampant in sheep after all, although scientists are far from ready to let their guard down.

    Spongiform diseases rob their victims—be they sheep, cattle, or people—of their balance, their minds, and inexorably their lives. As of 30 June, 75 confirmed or probable vCJD cases had been diagnosed in the United Kingdom, according to a study led by the U.K.'s National CJD Surveillance Unit in Edinburgh that appeared in the 5 August issue of The Lancet. vCJD cases have increased an average of 23% each year since 1994, and projections based on average incubation times of the disease have raised the specter of tens or even hundreds of thousands of deaths in the coming years from exposure to BSE.

    With BSE in British cattle on the wane, some experts now worry that BSE might be circulating in sheep, where it could be masked by scrapie, a related disease with very similar symptoms (Science, 17 March, p. 1906). Although there is no evidence to date that sheep can become infected naturally with BSE—nor that scrapie itself is transmitted to humans—sheep experimentally infected with BSE develop scrapielike symptoms. This raises fears that if a BSE epidemic occurred in sheep, it might be confused with scrapie. Indeed, there has been speculation that BSE may be to blame for a scrapielike illness recently reported in some sheep in Vermont.

    Like BSE, scrapie, which has flourished in British flocks for more than 250 years, is thought to be caused by infectious prion proteins. If BSE jumped into sheep from cattle, there might be a spike in the incidence of reported “scrapielike” cases during the height of the BSE epidemic in cattle between 1990 and 1995. To investigate this possibility, a U.K. team from the Institute for Animal Health in Compton, the Veterinary Laboratories Agency in Surrey, and the University of Oxford analyzed scrapie incidence between 1962 and 1998, using data from an anonymous mail survey of 11,554 sheep farms. Their numbers show that scrapie incidence has risen slowly, but there is no noticeable jump in scrapie cases before, during, or after the height of the cattle BSE epidemic in the early 1990s, when up to 36,000 cases of mad cow disease were reported each year. The authors interpret this to mean that the BSE epidemic in cattle did not lead to a comparable BSE epidemic in sheep. The data also show that farms raising both sheep and cattle did not run a higher scrapie risk, which might have been expected if many BSE cases were masquerading as scrapie.

    Although the survey doesn't eliminate the nightmarish possibility of BSE in sheep, “one can at least be confident that there has not been an epidemic in sheep equivalent to that in cattle,” says epidemiologist Peter Smith of the London School of Hygiene & Tropical Medicine, acting director of the U.K.'s Spongiform Encephalopathy Advisory Committee. The authors note that although their analysis was fine enough to detect a doubling of scrapie cases over the baseline incidence, a smaller blip, representing a smaller scale epidemic, may have gone unnoticed. Direct biochemical tests for BSE infection in cattle are now in use in Europe, but tests to distinguish BSE from scrapie are not yet available—leaving indirect studies as the only way to gauge infection rates. Thus, says Smith, “we cannot exclude the possibility that BSE was in sheep at a low level.”

    Although the new findings are comforting, researchers plan to keep a wary eye on sheep. As they know all too well, the history of the BSE epidemic and the emergence of vCJD illustrate the high price of complacency.


    A New Breed of High-Tech Detectives

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

    Forensic scientists are equipping police investigators with powerful tools for collecting and analyzing evidence. Whether some techniques infringe on civil liberties is a matter for public debate

    On an early spring day in April 1999, police in Albuquerque, New Mexico, found Stephanie Murphy's body wrapped in a sheet on the back seat of her car. Catherine Dickey, a forensic scientist on the force, suspected that the murdered 37-year-old had been sexually assaulted. Finding evidence for such a crime would be difficult, however. Murphy's body had been shut in the car in the hot desert sun for a few days, so any semen left on or around the body would have started to break down. “They knew that in that advanced state of decomposition they'd have a hard time finding any kind of evidence of a sexual assault,” says Colin Smithpeter.

    Smithpeter thought he might just be able to help. He and his colleagues at Albuquerque's Sandia National Laboratories are developing a unique camera designed to uncover just the sort of hidden evidence that the police were hunting for. Smithpeter met Dickey at the morgue, where she went to work examining the body with a blue light and tinted goggles, searching for traces of semen that would fluoresce in response to the lamp. She found nothing. Then Smithpeter unveiled his camera, which detects and processes faint glows that are elicited by an intense beam of blue light. The camera picked up three tiny stains on Murphy's skin. One turned out to be dried semen, a later lab test showed. The residue was shipped to the New Mexico state crime lab for DNA analysis, in anticipation that the killer had left behind a biological calling card.

    Murphy's killer remains at large, but at least investigators now have something to go on. “No criminal is so clever that he or she never leaves a trace,” says Edward Crew, chief of police in Birmingham, U.K. That's the guiding principle of forensic scientists, who are developing a cornucopia of techniques for detecting and analyzing elusive shards of crimes. Most of the new tools have grown straight out of basic research. Confocal microscopes, for example, can reveal idiosyncrasies in handwriting, while chemistry labs can sniff out molecules of naphthalene and other combustion products in a spent shotgun cartridge, offering a means of determining when a gun was fired even days later. Computer software can now calculate the exact position of a victim's body by reconstructing the trajectory of a spray of blood on a wall. And sophisticated databases can store vast amounts of information and search for unforeseen clues and connections between cases.

    No area of forensics, however, can boast more progress than DNA profiling, itself a product of the revolution in genetics research over the past couple of decades. Single hairs and specks of bodily fluids can provide enough DNA to help pin a crime on—or clear—a suspect. And handheld DNA devices, laser scanners, and digital cameras could soon bring the forensics lab to the crime scene, beaming back evidence to databases that could become Orwellian in scope, containing information on every one of us.

    Leading the way to a technology profusion that would surely astound even Dick Tracy are a handful of outfits throughout the world, including the FBI, the U.S. National Institute of Justice (NIJ), the U.K.'s Forensic Science Service (FSS), and the U.S. national laboratories. Their influence on crime solving will only grow, insiders say. “There is still a huge amount of untapped potential in the use of forensic science to fight crime,” argues FSS research director Trevor Howitt.

    But as the O. J. Simpson case and recent pretrial hearings on the admissibility of expert testimony warn, these fancy new techniques can be vulnerable to challenge in court. Some developments are also raising civil liberties concerns. And as detection techniques become increasingly sensitive, the problem of sifting the wheat from the chaff at a crime scene mushrooms, guaranteeing a role for the human investigator despite all the gadgetry.

    The DNA revolution

    When Ronald Keith Williamson found himself a free man in early 1999, he had his genes and molecular biology to thank. Eleven years earlier, Williamson had been convicted of the 1982 rape and murder of Debra Sue Carter in Ada, Oklahoma. Recent DNA testing of hair and semen found at the crime scene failed to link Williamson (or a second man charged) to the crime. But the DNA did implicate a former suspect in the case. Williamson is one of eight people to be freed from death row so far in the United States on the strength of DNA evidence.

    DNA profiling is a powerful technique for gauging the likelihood that a biological sample, such as blood or semen, came from a specific individual. In Williamson's case, the analysis belatedly ruled him out. The technique is the jewel in the crown of modern forensics. “That's the main area where we have really advanced in the past several years,” says Joe DiZinno, scientific analysis section chief of the FBI Laboratory in Washington, D.C. The use of profiling to identify suspects was pioneered by geneticist Alec Jeffreys of the University of Leicester, U.K. His approach, called multilocus profiling, used restriction enzymes to cleave the DNA at specific sites. The resulting fragments differ in size from person to person, as revealed by gel electrophoresis, in which fragments of different masses migrate at different speeds when subjected to the pull of an electric field. That was 1985, when a single, relatively large specimen took weeks to process and yielded a barcodelike output that was difficult to compare to other DNA samples.

    Modern profiling relies on short tandem repeat (STR) analysis, which debuted in the forensics world in 1994. This technique looks at specific areas of DNA molecules containing simple blocks of base pairs; the blocks are repeated end to end. The number of occurrences of each block, or repeat unit, varies by individual. Examining several DNA regions, or loci, and counting the number of repeated units in each area generates numbers that form a molecular label of the DNA's owner. STR analysis as it's used for DNA profiling relies on the polymerase chain reaction (PCR)—“the cornerstone of the whole process,” says Bruce Budowle of the FBI in Quantico, Virginia —to make millions of copies of the selected STR regions. The resulting DNA fragments are then sized up by gel electrophoresis, which yields the number of times each repeat unit appears in the fragment. This gives the DNA profile as a numerical tag for easy database comparison.

    The latest variation on the technique, introduced in Britain in mid-1999, targets 10 DNA loci—enough to guarantee that the odds of someone else sharing the same result are slimmer than one in a billion. Chance matches are even less likely in the United States, where the FBI routinely examines 13 STR sites. “The chance of two [unrelated] individuals on average having the same DNA profile is about one in a million billion,” claims Budowle.

    Thanks to PCR, today's routine work needs genetic material from a mere 600 cells containing about a nanogram of DNA; a speck of blood just 2 square millimeters in size will do nicely. Such sensitivity permits scientists to type DNA from, say, the mouth area of a mask tossed away after a bank heist, cigarette butts, or licked postage stamps on envelopes. The FSS “had a case where a guy picked up a brick and threw it through the window of a liquor store, and off the brick they picked up his DNA profile,” says geneticist Keith McKenney of George Mason University in Fairfax, Virginia. In another case, a discarded match supplied enough damning DNA. In these instances, saliva or dried blood didn't provide the smoking gun. Every 24 hours the cornea is replenished, and the dead cells are sloughed off into tears. By rubbing their eyes, the brick throwers and match users could have transferred these or other cells to their fingers, regularly reloading the DNA pen that left their telltale signatures.

    As powerful as DNA evidence is in court, forensic scientists are hoping to make it even more useful to crime investigators by providing them with the tools to infer suspects' physical characteristics. Geneticists can assess the likelihood that a person is a redhead simply by testing for mutations in the gene for the receptor for a hormone that spurs production of the pigment melanin. Ethnicity can be inferred from the frequencies of alternative forms, or alleles, of genes; allele patterns differ by racial origin. And the FSS is backing research at University College London's Galton Laboratory on the genetics of facial shape. A noble Romanesque profile or deeply cleft chin could be a villain's downfall. “All facial characteristics are on the agenda,” says FSS molecular biologist Gillian Tully, who acknowledges the huge challenge of deciphering the genetics of such traits, many of which are the products of several genes. Nevertheless, she says, “within 10 years we might be looking at genetic tests for the basis of the main facial characteristics like, for example, nose, chin, and forehead shape.”

    The FSS is also exploring the use of genetic markers called single-nucleotide polymorphisms, or SNPs, for DNA analysis. SNPs—individual base pairs of the DNA molecule that are known to vary between individuals—are hot items for tracking down disease genes and determining a patient's potential reaction toexperimental drugs (Science, 17 March, p. 1898). Although it would require about 50 SNP sites to achieve the same level of confidence as with STR analysis, SNP analysis holds a trump card: easy miniaturization, which means faster processing. Indeed, researchers and biotech companies are already developing rapid SNP identification techniques. “The potential exists to analyze several hundreds or even thousands of samples on a single microscope slide-sized device,” says Tully. Complementing such a SNP lab on a chip would be miniaturized setups for extracting and amplifying DNA.

    FSS chief executive Janet Thompson and others envisage technology, including handheld forensics devices, that can be brought to a crime scene and transmit data to a central facility for onsite analyses. Experts are debating how far down this track investigators should go. “I'm not so sure how desirable that is,” says DiZinno. He argues that evidence is better analyzed in a lab environment to avoid contamination of samples with extraneous DNA. Population geneticist Lisa Forman of the NIJ in Washington, D.C., agrees: “Although there may be a place for some onsite analysis at the crime scene, nothing will replace the laboratory bench,” she says, “because the kind of care and control that can occur at a laboratory bench is crucial for forensic evidence.” Budowle has another concern: “You don't want the police taking over that analytical process, because, remember, there's the whole interpretation of the results. Then when it goes to court, suppose there's a challenge on the evidence. Is the police officer going to go in and defend that challenge?”

    All in the family

    The DNA profiling that links dresses to presidents and Coke cans to cokeheads relies on DNA from the cell nucleus. A new star in the firmament for U.S. forensics efforts is the DNA present in mitochondria, the cell's powerhouse. At first glance mitochondrial DNA (whose forensic potential was pioneered in the United Kingdom) may seem like an odd tool. It offers less variety than nuclear DNA and is less discriminating, because the same sequences are passed from one generation to the next from mother to child, via the egg. Brothers and sisters, for example, share identical mitochondrial DNA. More than compensating for these shortcomings is that “there's probably 10,000 times as much mitochondrial DNA as there is nuclear DNA,” says McKenney. “In a sample that's aged or degraded, it's quite common that the nuclear DNA has been degraded beyond the point of recovery, and yet there is mitochondrial DNA that can be recovered.”

    For this reason mitochondrial DNA is associated with some of the glitzier results in DNA analysis. It has been extracted from 100,000-year-old Neandertal remains, has linked the 9000-year-old remains of “Cheddar Man” to a relative living today just down the road in Cheddar, England, and has identified the remains of Tsar Nicholas II based on similarities in the mitochondrial DNA of a close relative, Britain's Prince Philip.

    “The forensic application of mitochondrial DNA comes from the hair and fiber folks,” says McKenney. Hairs are often found at crime scenes. Although the fleshy root of a plucked hair contains nuclear DNA suitable for STR analysis, the shaft does not. It's possible, however, to grind up the hair shaft, copy the mitochondrial DNA using PCR, and sequence it. The FSS used mitochondrial DNA analyses to identify remains recovered from the Paddington rail disaster in London last year, by matching samples with DNA profiles of the victims' relatives.

    Although mitochondrial DNA can't conclusively link an individual to a crime, it can point a finger at a family, explains McKenney, who has employed similar analyses to study the familial links of 500-year-old Inca mummies. And that, he says, beats the more traditional and “somewhat subjective” microscopic comparison of the physical characteristics of crime-scene hair.

    The value of establishing family ties is not lost on missing-persons investigators. “Many people who have disappeared, or died, or whatever, didn't bother to leave a nice, clean sample of their DNA,” says McKenney. “So what you can do is, you can go to any one of their maternal relatives and get the same mitochondrial DNA.” The new FBI missing children's initiative, planned to come online at the start of 2001, will focus on gathering relatives' mitochondrial DNA. Samples from missing children (or their remains) recovered later would be compared to the relatives' DNA. “That's an example of how mitochondrial DNA is now, I think, ready for prime time,” says McKenney.

    Also making mitochondrial DNA more attractive to forensics investigators are new methods for quicker and cheaper sequencing. Tully helped develop a technique called minisequencing. “When you're sequencing, you look at something like 780 bases of the mitochondrial DNA,” she says. “When you're minisequencing, you're looking at 12, but you're looking at 12 of the bases that are most likely to differ between individuals.” This approach can slash the time it takes to get results from a batch of mitochondrial DNA samples from 3 months to 3 weeks. That's still ages compared to an STR analysis, which requires about 24 hours, but it provides a quick and dirty means of eliminating scores of potential suspects from an inquiry involving, say, a single hair at a crime scene. Or following a rape, it saves time being able to identify which hair needs to be sent for a more thorough analysis. Minisequencing “opens up many more cases for mitochondrial analysis,” says McKenney.

    Chemical clues to killers

    For all its power, DNA profiling can't nail all criminals: Not every crime scene throws up biological evidence. But forensic scientists are tapping into many other scientific disciplines. “The operative words are faster, better, cheaper,” says DiZinno, borrowing a phrase from NASA.

    The subjects of all this high-tech identification range from pen ink to automotive paints and explosives residues. Even duct tape, which is often used to bind victims or tape explosive devices in place, can help link a suspect to a crime. Researchers have developed a way to compare evidential duct tape to samples from known sources. The tape's adhesive has a characteristic chemical signature that allows a comparison between the questioned tape and known samples, as revealed by x-ray scattering or by a scanning electron microscope.

    Non-DNA chemical analyses could be destined for the field as well. The Forensic Science Center at Lawrence Livermore National Laboratory in California is developing a portable combined gas chromatograph and mass spectrometer that can churn through 200 high-quality analyses in the field each day. A prototype weighing just 20 kilograms is already in the hands of the FBI. Livermore researchers are also pushing portable mass spectroscopy to new limits. In a system under development, a sample is placed on the tip of a probe and inserted into a chamber. A laser blasts the sample into ion shards that are penned in by the electrical fields of an ion trap. The ions are then kicked out of the trap by electric pulses into a time-of-flight mass spectrometer tube, which identifies them by mass. The device is well suited to detecting minute traces of airborne chemicals such as nerve gas, for example, or to analyzing the composition of a hair for the chemical residues of drug use.

    Even the lowest of low-tech crimes are getting high-tech attention. Forcing open doors and windows with a wrecking bar, for example, leaves gouges that can be matched with the tool that caused them. At the FSS's behest, Isomark, a company based in Nuneaton, England, developed a fine-grain silicon putty for taking casts of tool mark impressions. The impressions have a resolution of a tenth of a micrometer and can be scanned to create a digital image.

    Standing in the dirt while levering open a window leaves a footprint—and telltale wear marks can point to the shoe's owner. Over the past year, Simon Bramble and his FSS colleagues have developed and patented a laser scanner for footwear marks, essentially an upturned flatbed document scanner with modified optics, held a centimeter above the footprint on three stubby legs. The scanner, which is making its debut at real crime scenes this summer, offers a number of advantages over old wet-film photography. “We can immediately see the result, so we are confident we have captured the mark,” says Bramble. And besides sometimes being sharper, digital images can be transmitted immediately for comparison with a library of footwear marks, fingerprints, or tool marks.

    The analysis of rotting bodies is also yielding to modern science. A one-of-a-kind facility in Tennessee is developing biochemical profiles of human corpses as they decompose to help investigators determine how long a body has been lying around (see p. 855). Other teams are applying DNA analysis to a time-honored method for making such a determination: analyzing maggots from corpses. Because the life cycles of maggot species vary by season and time of day, the type of larvae gives clues to when, exactly, a victim died. But because the larvae of many maggot species look alike, forensic entomologists are forced to keep the larvae in jars until they metamorphose into adults and can be told apart (see NetWatch, p. 827). Now, DNA technology is poised to make maggot-filled jars redundant. “DNA profiles of the larval forms can be used to identify what those maggots are,” says the NIJ's Forman, who says this approach has yet to be offered up in court.

    Fingering the criminals

    Long before DNA profiling burst on the crime scene, police investigators nailed hundreds of thousands of suspects with a familiar forensic technique: fingerprinting. Now, this time-tested technique is itself coming under scrutiny. The renewed interest is coming in part from court challenges to fingerprinting evidence in the United States that, says Vivian Baylor of the national security program office at Oak Ridge National Laboratory in Tennessee, “have really got the fingerprint community in an uproar.”

    Renowned fingerprint expert David Stoney of the McCrone Research Institute in Chicago and others have argued that the technique for comparing fingerprints—homing in on variations in only a few of the whorls that make up a print—is arbitrary and therefore unreliable. In pretrial hearings held in separate cases in September 1999 and April 2000, defense lawyers used those arguments to try to prevent fingerprinting evidence from being introduced in trials. Both challenges failed, but attacks on fingerprinting evidence are expected to continue. “The bottom line with fingerprints is that they have been accepted forever,” says Forman, but that “there is rather more limited scientific validation.” A startled NIJ is currently inviting submissions for research efforts aimed at providing additional scientific support for fingerprint identification.

    As some experts try to shore up the identification of fingerprints, others are working to wring more information out of the oily deposits themselves. A 1993 case in Tennessee indicated just how much researchers had to learn. A 3-year-old girl had been abducted, raped, and murdered by a friend of the family. The killer confessed to the crime while under the influence of drugs and alcohol, but when he dried out he denied it all. That forced Knoxville, Tennessee, police to turn to the physical evidence. They were certain the child had been in the suspect's car, but dusting revealed no sign of her fingerprints. Puzzled, police forensic scientist Arthur Bohanan made a surprising discovery: Children's fingerprints evaporate quickly. To help figure out why, he approached colleagues at the nearby Oak Ridge National Laboratory. There, Michelle Buchanan and her team discovered that the sebum exuded from the skin of children younger than about eight is laden with more volatile components than adult sebum. Thus a child's fingerprint vanishes quickly, sometimes in hours.

    The Oak Ridge team's studies have also suggested that fingerprints could provide more information than simply patterns of whorls. Their gas chromatography and mass spectrometry analyses revealed traces of estrogen in prints left by women and testosterone in those left by men, suggesting the ability to distinguish fingerprints by sex. And prints left by a smoker revealed traces of nicotine metabolites, raising the prospect of fingerprint analysis for metabolites of drugs such as cocaine, Baylor says.

    Finding fingerprints can be a problem, however: Dusting works well only for fresh prints on smooth surfaces. Thus one challenge is to extend the repertoire of surfaces from which a fingerprint can be extracted—getting a print off toilet tissue, for example, is very difficult. Another is to unravel old, overlapping prints and ascribe dates to each, something that remains a pipe dream. A common trick for making prints visible is to fix them with superglue vapor (cyanoacrylate) before using high-energy blue lighting to view them. Apart from the fact that the trick fails in the dry heat of places like New Mexico, there are other problems. “First of all, you have to superglue the whole world, which isn't all that helpful for the people doing it, but second of all you also have to black out all of the ambient light,” says Forman.

    This is where Sandia's Smithpeter comes into the frame. His “multispectral” camera creates images of light emissions stimulated by a powerful blue light used to sweep a crime scene for fingerprints and other stains. The blue light triggers molecules in semen, saliva, or urine stains to fluoresce and emit light at a lower frequency, typically in the green-yellow range. The camera is fitted with filters designed to block out reflected blue light, but pass the fluorescent light through to a charge-coupled device (CCD) detector. The trick is to make the illuminating lamp flash on and off at a fixed frequency and to record at a slightly different frequency. Because the blue light is filtered out, most of the time the camera simply sees the ambient light. But every so often the camera is “on” at exactly the right moment to catch a glowing stain, which shows as a bright flash against the string of ambient light signals before and after.

    An alternative to blue lighting would be to induce fluorescence with an ultraviolet or visible laser. A group led by John DiBenedetto at the Special Technologies Laboratory, a Department of Energy facility near Santa Barbara, California, is developing such lasers to pinpoint and image fingerprints and stains in ambient light. The team's approach is to exploit the spectral signature in the fluorescence coaxed from the sample by the laser light to identify it, then use a CCD camera to create a digital map of the stain.

    Database indictments

    On Good Friday of this year, a 23-year-old university student, Sara Cameron, was murdered in the Newcastle area of Britain. Her killer left behind just enough DNA at the crime scene for forensic scientists to build a profile of the unidentified male suspect. (Police are withholding other details on the nature of the evidence.) When their investigation stalled, they decided to do something drastic: take DNA samples from up to 10,000 men living in the area. British police cannot force the public at large to provide cell samples, but this has not inhibited people from cooperating during inquiries. The DNA testing has eliminated several potential suspects from the investigation, but so far has not led them to Cameron's killer.

    Mass DNA screenings raise an obvious point: DNA evidence by itself can't identify an unknown criminal. But the chances of finding the owner of DNA left at a crime scene will grow as more and more DNA profiles are included in searchable databases. For example, a police investigation into the rapes of two teenage British girls in 1989 sputtered until a fellow named Colin Jacklin gave a false name at a routine traffic stop in March 1998. Jacklin was arrested and enrolled in the U.K.'s DNA database; it turned out that his profile matched DNA samples collected at the rape scenes. Jacklin was found guilty of committing two rapes and an indecent assault; he's now serving two life sentences.

    In 1995, the United Kingdom became the first country to create a national DNA database. Based on the 10-loci STR approach, it now holds about 800,000 profiles of people suspected or convicted of an imprisonable offense. It is a major force in police intelligence, with about 600 “hits” every week from samples collected at crime scenes. British lawmakers have been generous to crime fighters, allowing them broad latitude in the acquisition of samples for DNA profiling. Other countries with DNA databases—including Austria, Germany, the Netherlands, and New Zealand—are typically more restrictive about the circumstances under which DNA samples can be drawn from individuals.

    In the United States, each state has passed laws that allow a DNA sample to be collected from individuals convicted of rape or other sex crimes, and many also permit sampling from people convicted of murder, burglary, or even certain misdemeanors. “It varies from state to state as to which crimes will qualify an individual after conviction for DNA analysis,” says the FBI's DiZinno. The U.S. national DNA database system, called CODIS, or combined DNA index system, started up in October 1998. It collates data from all the state databases. There are maybe 100,000 profiles in CODIS now, but according to DiZinno a whopping 750,000 blood samples from convicted felons have yet to be profiled and those profiles incorporated into CODIS.

    At some point, society will be faced with a difficult question: Should everybody be included in a DNA database as a matter of course? As costs plummet—currently a single STR-type DNA profile costs about $100—a populationwide database will soon be financially feasible. “I think it's going to come,” says George Mason's McKenney, who is building a mitochondrial database with Department of Justice backing. “Once we understand the value of DNA profiling, and we put in sufficient assurances that it won't be abused or misused, then we will pass laws that say everybody will be profiled on birth,” he says. “It will be just like a Social Security number.”

    That prospect unnerves civil liberties lawyers and some forensic scientists. “At least in the United States that would never happen,” predicts DiZinno. “There are too many privacy issues involved.” And Budowle would rather it didn't happen: “You have to have a balance between the privacy of the citizen and the needs of the state,” he says. As forensic science grows more sophisticated, that balancing act will be ever harder to maintain.


    Where Dead Men Really Do Tell Tales

    1. Robert F. Service

    A research plot where human corpses are studied as they decompose is yielding a bumper crop of insights into how the body decays

    KNOXVILLE, TENNESSEE Take one look at the half-naked middle-aged white man sprawled on a patch of gravel next to a wooded path, soaking up the summer sun, and you'd never know that a few weeks earlier he was wrestling with a weight problem. His leathery skin sags around his bones and is beginning to disintegrate. The muscles, organs, and other soft tissues are long gone, liquefied and consumed by bacteria and insects. Some of the flies buzzing around probably got their start as maggots that gnawed their way through the cadaver to the outside world.

    The stench is overpowering.

    The thought of being left out in the elements to rot may be a bit unsettling, but this anonymous man donated his body to science and presumably knew, before passing away, that he could end up at the “Body Farm.” After suffering a fatal heart attack in mid-June, the man became part of a hunt for chemical markers that reveal the time elapsed since death. Behind a chain link fence topped with razor wire to keep out vandals, some 20 bodies lie prostrate in various states of decomposition. Most individuals died weeks ago and now appear human in outline only. Once inside, a trio of researchers led by Arpad Vass, a chemist at Oak Ridge National Lab in Tennessee, makes the rounds, taking soil samples from around the remains.

    The site, officially known as the University of Tennessee, Knoxville's (UT's) Anthropological Research Facility, was started in 1981 by a pioneer in forensic science, UT's William Bass. His exploits identifying remains and fingering suspects inspired Patricia Cornwell's 1994 novel The Body Farm—a name that stuck, much to the scientists' chagrin. “We don't like to call it the Body Farm. It's not as respectful as it should be,” says Vass. But whatever you call it, he says, “it's an absolutely one-of-a-kind facility.”

    The corpses are teaching scientists much about how to reconstruct the manner and circumstances of unexplained deaths. Perched above one body is a mechanical nose of sorts, a textbook-sized box with 32 separate chemical sensors that sniff noxious fumes wafting into the air. Back in the lab, the aromatic compounds are washed from the nose into a vial, from which they are fed into a gas chromatograph to identify each component. A neural network compiles unique patterns of chemical signals that correlate with precise times since death—patterns that might be discernible above other corpses. Ultimately, Vass says, he and his UT collaborators hope they will be able to pluck out a single compound that can do the same job. A handheld device that homes in on such a chemical could then be built for crime-scene investigators.

    Although several U.S. national labs are redefining their post-Cold War missions to include projects that aim to come up with new crime-fighting tools, Oak Ridge has enjoyed the most success at putting together a forensic science program. Ongoing projects are attempting to recover fingerprints too faint to see with normal dusting procedures; spot crimes on video surveillance tape by clearing out hiss, noise, and distortion from poor lenses; and build portable chip-based devices for gathering and analyzing chemical forensic evidence in the field. Oak Ridge's suite of projects “is pretty impressive,” says Kevin Lothridge, co-director of the National Forensic Science and Technology Center, a Largo, Florida-based center that offers technical support for crime labs around the country. Since most funding for forensics gets plowed into casework, he says, new sources of novel investigative techniques are sorely needed.

    Indeed, most forensics techniques leave ample room for improvement. Take, for instance, a case that Vass says helped convince him to get into this field. About 20 years ago, Bass—Vass's mentor—was called in to help solve the mystery of William Shy, a colonel in the Confederate army. Shy's grave on his Franklin, Tennessee, plantation had been disturbed accidentally by the home's current owners. The body they uncovered, however, was surprisingly intact, with flesh still clinging to the bones. Because of the seeming freshness of the remains, Bass presumed that the body was a recent murder victim disguised in Shy's clothing. But closer inspection revealed that the corpse had never fully decomposed. It was buried in a lead-lined coffin, and the metal had leached into the body and surrounding soil, rising to concentrations high enough to prevent bacterial growth and decomposition. Simply eyeballing the body had misled the seasoned expert into pegging it 110 years younger than it was. “I thought, ‘;There has got to be a better way to do this,’” says Vass.

    There is now. While in graduate school, Vass worked out two new methods for dating remains. The first tracks the ratio of five fatty acids—valeric, and the straight and branched forms of propionic and butyric—which, he found, vary systematically as long as there is soft tissue. The various compounds can be sampled from any material around a corpse. Their concentrations rise and fall, but each day that passes after death brings a unique profile of all five, says Vass. A similar technique looking at ratios of seven inorganic compounds (including calcium sulfate and magnesium) that leach from bones into soil turned out to be an accurate marker after the flesh disappears, usually a few weeks after death.

    These new techniques are already beginning to land criminals behind bars. In 1997, after police in Cheshire, England, found the bones of a murdered 11-year-old boy lying on property belonging to his father's family, they zeroed in on Dad as the chief suspect. But when investigators sent Vass samples of soil from beneath the body, he found no traces of the telltale inorganics—even though the boy had died nearly a year earlier. The evidence clearly demonstrated that the victim's remains had only recently been brought to the property, Vass says. The father apparently knew that his wife was having an affair with a neighbor. When confronted by police, the neighbor allegedly confessed to the slaying, admitting that he and the boy's mother had killed the child in a twisted effort to implicate her husband.

    A recent case in Florida took a similar turn. In this instance, Vass says, a prisoner bragged to his cellmate that he had abducted a woman leaving a convenience store before raping and killing her. The prisoner also revealed that he had an accomplice help him bury and move the body several times, each time fearing they had been spotted. The cellmate told the police, but the prisoner later denied the story. Hoping for corroborating evidence, local police tracked down the accomplice, got his cooperation, and sent Vass soil samples from several potential temporary grave sites identified by the accomplice, including one on property belonging to the suspect's family. Analyzing the minerals and fatty acids, Vass found clear signs of the presence of a dead body in samples from all but one of the sites. The amounts of the chemical markers confirmed that they came from a large mammal, and the lack of fur or animal bones convinced investigators that the mammal was a person. The woman's body was never found. Still, after listening to the evidence, the defense attorney persuaded his client to accept a plea bargain for life imprisonment rather than possibly face the death penalty.

    Vass and his colleagues have now turned to volunteers willing to lend a hand (and then some) to help them expand their forensic chemistry toolkit. Their aroma studies are already beginning to pay off. With the neural network, Vass says, “we see significant changes between early and late individuals.” Now they're trying to home in on a single marker. “We've tracked hundreds of compounds thus far,” he says. At this point, the best candidate markers are phenols, ring-shaped organic compounds produced by the breakdown of amino acids that can even be detected in the air above shallow graves. If phenols live up to their promise, a handheld detector could prove a boon to detectives searching for missing bodies.

    Beyond the Body Farm, other Oak Ridge projects could soon put new tools in the hands of investigators. The lab is among the leaders in developing new techniques for analyzing the age-old art of fingerprint analysis. And Oak Ridge researchers are nearing completion of a suite of new software tools to allow crime labs around the country to recover incriminating images on grainy video surveillance tapes. Here too, the work has already helped police investigators solve cases. But it's the Body Farm that has put the Tennessee team on the forensic science map.


    Forensic Science on a Shoestring

    1. Robert F. Service

    With billions of dollars spent each year on law enforcement in the United States, one might think that funds for R&D on advanced crime-solving methods would be flowing freely. But the experiences of the U.S. national laboratories in trying to carve out a niche in forensics research suggest otherwise. “Justice and law-enforcement research funding isn't a blip on anybody's screen,” says Kevin Lothridge, co-director of the National Forensic Science and Technology Center, which provides technical support to crime labs.

    The main agencies that fund this sort of work in the United States—the FBI, the National Institute of Justice, and the Department of Energy (DOE)—together spend less than $50 million a year, Lothridge calculates. Although this amount “is much improved” over what was available a few years ago, he says, vagaries in year-to-year funding levels make it difficult to sustain ongoing programs. The insecure environment has snuffed out some fledgling crime-fighting efforts at the national labs. In 1995, for example, DOE gave researchers at Oak Ridge National Laboratory in Tennessee seed money to launch the Center for Applied Science and Technology for Law Enforcement (CASTLE). The program backed Oak Ridge projects such as a computerized system for rendering facial images from skeletal remains, a potentially faster and cheaper approach than the clay reconstructions currently in use. Compared to other national labs, Oak Ridge has sustained a healthy forensics research program (see main text). But CASTLE was phased out 2 years later after failing to raise enough grant money to keep itself afloat. In the end, says Vivian Baylor of Oak Ridge's national security program office, the computer facial reconstruction effort “withered on the vine.”

    That same pattern could be playing out at Lawrence Livermore National Laboratory in California, where researchers launched a Forensic Science Center (FSC) in 1991 to support local law enforcement and other agencies concerned with crime. Center scientists drew heat in 1994 over their controversial (and impossible to prove) hypothesis that a novel chemical reaction in a dying woman's blood produced a toxic gas, which sickened workers in a California emergency room trying to save her life. The circumstances surrounding her death remain a mystery. Among the FSC's achievements, however, is a portable gas chromatograph for analyses in the field. But the fancy hardware has not helped entice continued funding, says FSC chief Brian Andresen.

    Besides struggling for research dollars, the FSC has failed to attract stable funding to support analytical services it has provided to law-enforcement agencies looking for help in cracking tough cases. Because the FSC must charge up to $1000 a day per investigator to cover costs, work on a single case runs as much as $30,000, says Andresen. “Local law enforcement just doesn't have that kind of money,” he says. As a result, Andresen is scaling back his program to focus on one of Livermore's well-endowed core areas, chemical and biological weapons nonproliferation and counterterrorism. Over the years, Andresen says, he's heard plenty of rhetoric about how advanced equipment and techniques honed at the national labs could transform forensics casework. But “when push came to shove,” he laments, “law enforcement just didn't come up with the money.”