When Perera entered graduate school in the environmental health sciences program at Columbia in the mid-1970s, the field had begun to move away from worrying about microbes. In Western societies antibiotics had controlled menaces like tuberculosis and typhus, and the mechanisms of bacteria and viruses were well understood. Environmental health, the study of disease transmitted through the environment, included another focus: industrial toxins.

During the 1970s health officials commonly stated that as much as 90 percent of cancer was environmental in origin, with pesticides and industrial chemicals responsible for half the incidence. Even as regulatory measures were put in place—the Occupational Safety and Health Act, the Clean Water Act, the Clean Air Act, the Toxic Substances Control Act—epidemiologists hurried to analyze the risks. "The National Institute of Environmental Health Sciences was born out of the notion that we were living in a sea of chemicals," says Gwen Collman, a program manager at the institute who works with Perera today. "There was a lot of pressure to understand what was going on."

By recording who got sick after years in the workplace and who did not, and then comparing their exposures to chemicals, investigators were able to put hard numbers on the carcinogenic potencies of many substances. The most pervasive carcinogen, cigarette smoke, was identified outside the workplace. Exposure to it increased the risk of lung cancer 9 or 10 times over that of nonsmokers. None of this research required any laboratory biology. When researchers compared retrospective studies of smoke exposure and death rates, the impact was unmistakable.

Smoking is voluntary, of course, and Perera began to wonder about involuntary low-dose exposures to other common chemicals. "I'm seeing this research happen, and it's great," she says. "But at the same time hundreds of thousands of others are being exposed. The biggest challenge was air pollution. It's so hard to measure compared with waterborne contaminants. I thought, That would be the thing to tackle, the air pollution story."

In 1976 an environmental advocacy group, the Natural Resources Defense Council (NRDC), hired Perera as a health scientist. Then as now, the council took a tough stance on toxic threats to the natural world and human health. She and a colleague at the NRDC wrote a book, Respirable Particles, in 1979, which warned that the government's air pollution standards for American cities and industries were failing to curb the very smallest particles. "The method [of detection] was to weigh the air filters," Perera says, "and these captured only the bigger particles, not the ones down to one or two micrometers [about one-fiftieth the width of a human hair]. We said that the Environmental Protection Agency should regulate these fine particles and undertake a research program."

To read Respirable Particles today is to appreciate how far ahead of her time Perera was, or at least ahead of the regulators. Studies associating urban death rates with air pollution have since drawn a tighter and tighter circle around fine particulate matter. As a result, the EPA began to monitor and regulate particles as small as 10 micrometers, also called microns, in 1987, but did not get down to the level of 2.5 microns, as Perera had urged, until 1997, when standards were tightened. Fine particles are considered especially dangerous because they lodge deeply in the lungs and aren't cleared by coughing. Trapped there, the toxic fractions, including hydrocarbons, can pass into the lung tissue.

"Evidence came out that fine particles could be more dangerous than believed," says Raymond Werner, chief of air programs for the New York and New Jersey region of the EPA. "We're able to home in more and more, based on what we've learned from her studies and others' studies. We've upped the ante on the kind of information we're seeking."

While at the NRDC, Perera became interested in a hydrocarbon compound called benzo[a]pyrene, or B[a]P for short. It had a long rap sheet in toxicology. B[a]P is found in coal tar, and coal tar was linked to cancer in English chimney sweeps as long ago as the 1700s. B[a]P is a constituent of the sooty yellow emissions of coke ovens (coke, derived from superheated coal, is used in steelmaking). Workers tending coke ovens were at high risk of developing lung cancer, according to the occupational health research, and the workers who smoked were at even more risk. Cigarette smoke also contains B[a]P. In 1981 the National Institute of Environmental Health Sciences declared that benzo[a]pyrene was "reasonably anticipated" to cause cancer in humans.

When Perera returned to Columbia in 1979 for her doctorate in public health, she was eager to close the case on B[a]P. She wrote a long paper, her first work in a scientific journal, about the compound's harmful properties. She called for the EPA to rein in B[a]P pollution under the Clean Air Act. But by then the general cleanup of America's air had resolved the worst of the problem. The EPA never did regulate the substance.

Perera's thesis adviser at Columbia was a molecular biologist named Bernard Weinstein. "My lab was interested in DNA adducts, how B[a]P bound to DNA," Weinstein recalls. "My group in 1979 showed the molecular structure of the B[a]P adduct. We thought, Wow, that tells you the DNA has been clobbered by a carcinogen. It was a carcinogen on a critical molecule. We could use that as a marker."

Perera began working with Weinstein on a small study comparing two groups of hospital patients: One group was suffering from lung cancer; the other was a group of orthopedic patients as a control sample. Perera and Weinstein detected the benzo[a]pyrene adduct in some of the cancer patients but not in the control group. Although the patients' exposure to cigarettes and other hydrocarbon sources was not integrated with the results, the idea of a marker for lung cancer held up.

"We showed it could be done," Perera said. "There was a suggestion of risk in humans. Epidemiology can be done without markers, and it can be very elegant, but to me it was important to know where and when to intervene, before the tumor is locked in and inevitable."

Next, she and Weinstein wrote a conceptual paper that would become the founding document of molecular cancer epidemiology. The paper laid out four categories of markers according to what Weinstein called "a continuum of causation."

The first type of marker simply measures a dose. A foreign chemical, like lead, is detected in fat or blood or urine, but the body hasn't done anything special to it except to retain it or eliminate it. The second type of marker has been transformed through a "biologically effective" reaction. A DNA adduct is an example, because the toxic agent has bonded, at least temporarily, with genetic material. Individuals differ greatly in the amounts of adducts that their blood and tissues will express in response to an exposure: The range can be 100-fold or more.

The third kind of marker Perera calls "preclinical," meaning that a medical development may be in the offing. Examples are chromosomal aberrations or a mutation in a gene that has been implicated in the formation of a tumor. Finally, there are innate markers of susceptibility. Certain genetic variations make certain people more likely to succumb to an illness. Over the last 20 years, genetic research has provided many markers of susceptibility, just as Perera and Weinstein predicted.

Overall, the environment's suspected role in provoking cancer is becoming less prominent as genetic and other predispositions are uncovered. Lifestyle factors, like diet and smoking, still weigh heavily, but professionals say that pollution and industrial carcinogens account for 5 percent or less of cancer incidence. Of Perera's peers in molecular epidemiology, perhaps the best known is John Groopman of Johns Hopkins University. Groopman doesn't study industrial toxins at all. He examines how aflatoxin, produced by mold on grains, peanuts, and other crops, leads to liver cancer in Asia and Africa.

The secret of Perera's success, I concluded, was adaptability, both personal and professional. When air pollution declined in American cities in the 1980s, Perera found new urban sites abroad, first in Finland and then in Poland, where she could pursue hydrocarbons and their biological signatures. The opportunity to work in Poland came after the collapse of Soviet Communism. U.S. health officials offered to help their counterparts who were struggling with rampant, unrestricted coal burning by factories and homes.

The pollution in Poland jolted Perera into thinking about children as study subjects. "The air stung my eyes," she said. "I was alarmed—well, not alarmed, but I was concerned. I thought about the children. Let's go into the womb, I thought."

In the 1990s health officials and researchers converged on the view that children were more sensitive than adults to pollutants and so merited additional protection. Exhibit A was lead, which set back children's mental development without appearing to harm adults. Pesticides in foods and contaminants in air and water posed extra hazards, if only because children absorb relatively more of these substances. Perera and her associate Robin Whyatt pointed out in a 1995 paper that children have "higher breathing rates, ingest more drinking water, and consume more calories of food per unit body weight than do adults." So the researcher adjusted her sights for cancer risk from adults to children and added markers for cognitive development and asthma to her molecular toolbox.

"The value of a susceptible subpopulation like children," says Gwen Collman of the National Institute of Environmental Health Sciences, "is that it helps you get away from the mean, which may cloud the truth. Why waste your efforts on protecting everybody when not everybody is affected?" When the institute and the EPA awarded funding in 1998 to create children's health research centers around the country, Perera set up shop as director of the Columbia Center for Children's Environmental Health.

Her assumption was that if air pollution is bad for mothers and infants, it's worse in communities where mothers and infants are poor. Perera's team analyzes housing conditions and psychosocial stressors as well as polycyclic aromatic hydrocarbons, cigarette smoke, insecticides, toxic metals, and indoor allergens. The Columbia center not only collects health data from its subjects but also offers health instruction.

Counting staff and scientists, about 50 people are employed full- or part-time on projects. Perera has teamed economists and psychologists with toxicologists and analytical chemists. "I'm an interdisciplinary figure," she said, smiling at the grand description. "I like bringing together juxtapositions. As one of my advisers said, to be 'interdisciplinary' is to have several skills in one skull."

Perera has people skills in spades. Christopher Dickey, one of her graduate students and later a research assistant, recalled going to scientific meetings with Perera in the early 1990s, when molecular epidemiology was new. Dickey would watch in admiration as his boss picked the brains of potential partners. It wasn't just advice on biological markers she sought, it was also help in paying for the analyses. "She is a master of those collaborations," he said. "She's phenomenal at it. She gets people with different agendas to work together."

Dickey gave two reasons for her persuasiveness: "Sensitivity—I haven't seen it honed in other people as finely as it is in Ricky. Also, she's physically striking and extremely polished. It's a little intimidating even. There aren't that many with that charisma."

When I visited, Perera was confident enough to open all her staff meetings to me. She ran the agenda and dealt with problems and interruptions with an unfailing calm, her voice relying on a softly rising tremolo for its effect rather than a boost in volume. She penned a note to herself on a fingernail—"Wed at 9"—for the next meeting.

No edges were evident, nor any abrasions from the harder climb a woman has to make in the largely male world of science. The closest she came to a feminist statement was when I referred to the DNA adduct, in use for more than 20 years, as the granddaddy of biological markers.

"It's the grandmother, don't you think?" she says, arching an eyebrow.

On the last day of my visit we went to see her latest collaborator, Benjamin Tycko, a cancer geneticist at the Columbia-Presbyterian Medical Center with much fancier digs than her own. Perera had sent him cord-blood samples for a trial run of their project.

Tycko studies a phenomenon called gene silencing, in which one of a person's two copies of a gene is shut off. Normally a person's gene-silencing patterns are inherited from one parent or the other. But chemicals, including those in cigarette smoke, can shut genes off, too, or turn inactivated genes back on. Sometimes, when genes regulating growth are switched on or off, a cell turns cancerous.

Tycko is examining gene silencing in PEG-1, a growth-factor gene that's active in the placenta and in the development of the fetus. Perera would like to know if the gene's activity can be switched on and off by environmental toxins. If so, Tycko's PEG-1 marker might be paired with one of her own. Ideally, the markers would lie on the same path in the maze connecting the mother's environment, the placenta, and the newborn.

Perera turned the full beam of her attention on Tycko as he explained how the cord-blood samples showed that the gene's activity varied and that it varied in ways that suggested the differences were not random but environmentally influenced. In other words, the project that she had in mind, to test PEG-1 as a marker, was feasible.

"What's this gene's relation to cancer?" Perera asked.

"It's not clear-cut," Tycko said cautiously. "It's related to growth."

They began to discuss a joint venture. "You're the one who knows how to measure the exposures," Tycko said. Perera left the meeting encouraged, saying, as she strode briskly back toward the Hudson River: "We're very poor compared with the 'hard' sciences. On the other hand, the hard sciences are saying they need the environmental side."

The sky was a high, hard blue, clean as a whistle. If there is a single cloud hanging in the way of molecular epidemiology, it's called "validation." A validated marker can be used precisely, accommodating whatever question is asked of it, so that scientists and policymakers can take it off the shelf, plug it into their risk calculations, and have confidence in the results. Blood levels of lead predict neurological deficits far better than any DNA adduct or aberrations can predict a cancer. The day for cancer markers will come, but it will take more spending, and it will take a much deeper understanding of carcinogenesis.

"It's slow," Perera said, "but I'm patient." She read from notes she had written, so as to be clear. "The picture has become more complex. It won't be solved in my lifetime. I won't solve these problems, but I hope to establish methods for others to follow."