Originally Kim insisted the whole family eat only the foods Tessa could, but then she realized it wasn't fair for her younger daughters not to be able to eat like other kids at school and birthday parties. Suppose you couldn't walk, Kim said, explaining her thinking to Tessa; should I make your sisters sit in wheelchairs too? Kim herself kept to Tessa's diet, however — Tessa never saw her mother so much as add milk to her tea.

Yet this carefully constructed world was in constant danger of collapse. In 2011, Tessa almost died twice. First, when she was 7, a piece of rye toast turned out to contain traces of wheat. Then, 10 months later, Kim took a small uncharacteristic break from their rigid food routine and bought some Vietnamese summer rolls from a restaurant after quizzing the staff about each of the ingredients. But the clear noodles that she was told were rice turned out to be made of wheat, and soon Tessa was losing consciousness. She didn't have hives or other external signs parents often rely upon, but internally her body shut down. At her doctor's office, the medical team had to use two EpiPens, adrenaline-loaded syringes, along with steroids and an array of drugs to bring her back. (Injected adrenaline is the only known antidote with the power to arrest anaphylactic shock, the allergic-immune response that causes tissues throughout the body to swell until the windpipe closes, the lungs collapse and the heart fails.)

A week later Tessa began having panic attacks. She no longer wanted to leave the house without her mother — even to go to school or diving practice. She was afraid to eat. "Her belief was, If I don't eat, I can't die," Kim recalled in one of our many conversations over the last year. When Kim went away for the weekend, she returned to discover her daughter had eaten only one bowl of plain white rice in 48 hours. At school Tessa didn't want to touch anybody or anything. What if at recess the kickball had rolled through a splash of milk or some bread crumbs in the courtyard where kids ate their lunches? Costly sessions with a child psychiatrist were of limited value. Treatment for obsessive-compulsive disorders like germ phobia teaches the sufferer to distinguish anxiety from reality, but the reality for a severely allergic child is that invisible trace contaminants can kill.

Kim, a pretty 42-year-old with wholesome girl-next-door looks, has a particular optimistic energy indigenous to Silicon Valley: an entrepreneurial drive, buoyed by a sunny, hopeful Northern California temperament, embedded in a community of people who make dreams — at least of a certain sort — come true. Although she and her husband, Andy, an executive recruiter, were not in a position to write big checks, Kim had been involved in enough philanthropic efforts to know she could mobilize tremendous resources. She grew up in the area and was the kind of person who couldn't walk into a coffee shop without people coming over to thank her for her help with some enterprise or to try to enlist her for another.

Kim realized that her daughter was born on the cusp of an epidemic. When milk first splashed on Tessa's skin, Kim was thrust into a strange, lonely world, where few parents shared or understood her anxieties. In the following years, however, she watched food allergies explode into a major public-health problem. The rate of food allergies has more than doubled over the past decade, and there are now an estimated 5.9 million children in the United States with food allergies (along with another 2.3 million adults); to put it another way, in every classroom, roughly 1 in 13 children will have a food allergy. That number may rise in the future, as food allergies are now even more prevalent in children between 3 and 5 years old — nearly 1 in 10 preschoolers — and it seems as if children aren't growing out of their allergies at the rate they did before. In our children's (nut-free) nursery school, their EpiPen packs are on a shelf with five others — the first time their teacher has seen that many in her long career.

In April 2009, Kim heard that Kari Nadeau, an M.D./Ph.D. and an associate professor of allergies and immunology at Stanford University School of Medicine and Lucile Packard Children's Hospital, was giving a lecture at Stanford. Kim had recently flown across the country to meet an expert in New York to see if anything could be done for her daughter. He wasted no time in telling her, "We're not hiding the answer under a rock here," and advised her simply to continue their regimen of strict avoidance. While young children frequently outgrow allergies, Tessa's case offered little hope, because her blood work showed high quantities of an immune protein called IgE, which is used as a marker of the severity of the allergic response.

Nadeau and others, however, were having success with a trial of a treatment known as oral immunotherapy that could desensitize children with severe peanut allergies. The treatment re-educated the hyperactive immune systems of allergy patients by giving them minute doses of peanut every day, gradually escalating the amount over the course of several years. Eventually patients build up their tolerance for the food, and it is no longer dangerous.

Desensitization is a straightforward idea; it's the same principle, according to legend, that King Mithridates VI used to cultivate immunity to poison (an occupational hazard of royalty in the ancient world). The therapy has been used successfully for environmental allergies for decades, by giving patients small injections of pollen or cat dander or other allergens, but it was considered too dangerous to try with food allergies until recently. A study testing various allergens was pioneered in Europe in the 1980s, and in the past five years, ongoing studies at Mount Sinai School of Medicine in New York and other centers have shown that children can be safely desensitized to a single allergen, peanut, and, in separate trials, to milk and egg.

But Tessa was allergic to more than a dozen foods. If she were to enroll in single-allergen trials, simply being desensitized of three of her allergens would take the better part of a decade.

People hearing Nadeau talk for the first time often describe falling under her spell. As Nadeau lectured, Kim became filled with conviction: This is the person who will cure my daughter. Afterward she approached Nadeau and asked what she could do for her 6-year-old, who was fatally allergic to most major food groups.

"I am not sure," Nadeau told her, "but I promise, we will figure it out."

Kim's question intrigued Nadeau. Could patients be desensitized to more than one allergen at a time? No one had ever tried it, but more than a third of children with food allergies are allergic to more than one food. If it was safe to give patients x milligrams of one allergen, would it be safe to give them one-fifth of x milligrams of five different allergens, as long as the total dose remained the same? That would assume that allergens function in a linear, additive fashion — rather than a multiplicative one; it was also possible that they could interact with one another to produce a more severe reaction.

Nadeau experimented with blood samples of allergic patients and was encouraged to see that the allergens seemed not to interact with one another. She consulted with senior colleagues in the field to see if anyone would collaborate on a multiallergen study, but no one was interested. Scientifically the results would be harder to interpret than single-allergen trials. Moreover, each allergen would require getting separate F.D.A. approval, and it was difficult to get even one application approved. When she found herself home sick in bed with a virus for a few days in 2011, she decided she would "knock them all out" and wrote 13 Investigational New Drug Applications, each 90 or so pages long, and soon received F.D.A. approval for each one.

Even more daunting was the question of how to finance the study. Each child would cost between $20,000 and $30,000 to treat annually, and treatment could take several years. Flour would have to be manufactured from the proteins of each allergen to prepare precisely measured, minute doses, and it would have to meet the high purity standard for drugs. While there were more than enough parents in the area who could afford this treatment, ethical rules governing trials prohibit individuals from "buying" places by paying for themselves.

Nadeau applied for research grants, but none materialized. While food allergies cost an estimated $500 million a year, Congress recently appropriated only $28 million a year for research (compared with, say, $1 billion for diabetes and $5 billion for cancer). The Food Allergy Research and Education group (FARE), the only major food-allergy organization, spent an additional $3.4 million for research last year. Nadeau believed she had an important idea — one that could revolutionize her patients' lives — but in a situation that young researchers often face, she couldn't compete for grants without proving her idea, and she couldn't do that without financing.

The only alternative seemed to be raising money directly from private individuals. No one had heard of this kind of major trial being financed through grass-roots philanthropy, but Nadeau often reminded herself of Margaret Mead's words: "Never doubt that a small group of thoughtful, committed citizens can change the world." She and Kim began to talk about how to do it, and Kim organized a group of mothers of children with severe allergies to work with Nadeau's Stanford Alliance for Food Allergy Research. The women, who call their group the Safar Community Council, immediately bonded: no one understood the lives they all were living better than they did. The group had an abundance of passion, skill and advanced degrees to draw upon: many of the women had forsaken careers to care for their allergic children.

The council members reached into their own pockets again and again and implored family and friends. When the financing fell short, two sets of Safar parents who had already made initial donations, Melissa and Scott Kepner and Shirley Chu Orsak and Mike Orsak, agreed to split the balance. To keep giving, the Kepners canceled several family vacations: when they thought about whether they really wanted to go skiing for Thanksgiving or give more to the multiallergen study, they voted for the study. (Traveling with their young daughter Kate, who was severely allergic to eggs, milk and nuts, was no small feat. The first thing Melissa did at hotels was ask to use a vacuum, to ensure their toddler wouldn't find a peanut under the sofa — perhaps a first for the Four Seasons. Then they set up camp with their own toaster, hot plate and food rather than risk a restaurant.) Steve Carell volunteered to host a fund-raising auction, because Nadeau treated his daughter, Annie, for a dairy allergy in an earlier study.

Nadeau herself decided to forgo a salary for three years. She was left with only a small salary from clinical care, which — considering the high cost of child care — left her with negative cash flow. "My husband was a bit cross," she told me, laughing.

By November 2011, the Safar Council had raised enough money for two multiallergy trials with 85 patients, each of whom could be desensitized to up to five allergens. (To date, the group has raised $6.2 million, including recent grants from FARE and the National Institutes of Health — $2.2 million to finance the multiallergen trial and $4 million to establish an endowment for food-allergy research.)

Food allergies are a peculiar disease, because most of the time the child is not sick — indeed, she may be bursting with health — but is in omnipresent danger. Statistically the chance of dying is slight. Although the number of emergency-room visits for anaphylaxis caused by food has gone up significantly in the past decade — to as many as 90,000 in a year — only 100 to 200 people die (although statistics are difficult to collect because such deaths are often coded as cardiac arrest). Even for a severely allergic child like Tessa, the mortality rate is estimated at roughly 1 in 1,000, because parents of such children tend to be extremely careful. But food allergies amplify a kind of fear every parent experiences — of a child dashing suddenly into the street and, just like that, being gone. Your child is always playing near a precipice that is visible only to you: you may be able to keep her from falling off, but you can never move her away from the edge.

At Safar Council meetings (which I joined last fall), parents would look around the room and ask themselves if their story would be like Shirley Chu Orsak's, whose daughter, Nicole, felt her throat close up one day in fifth grade. Her mother always told her, "No one knows your body better than you," so she ignored the adults urging her to wait for paramedics, got the EpiPen from her backpack, injected herself and saved her own life. Or would their story be like that of Brian Hom's, whose youngest son, Steven, is in the trial now in the wake of the death of his older brother, B.J., who ate a dessert that turned out to contain peanut while on a family trip to Mexico to celebrate his high-school graduation? The family didn't have an EpiPen, because they thought the worst that could happen was that he would get some itchy hives, as he had before. They still thought he would be O.K. when the medical team arrived, but then the doctor said, "I'm sorry," and pulled a blanket over his head and everything was over.

Food allergies are nonsensical. "Why would evolution have us be allergic to the things that sustain us?" Nadeau asks. "What would Darwin say?"

In an allergic reaction, an ordinary food protein like peanut is misidentified as an invader, like a virus or bacterium, to which the body needs to mount a defense. In the body of a person with a peanut allergy, for example, the presence of peanut protein causes the release of peanut-specific IgE antibodies, which sets off the release of a flood of chemicals, including histamines, into the bloodstream. Those chemicals lead to the release of still more chemicals, and the chemical cascade escalates in a self-sustaining feedback loop until the response causes the heart or the lungs to fail, and the person collapses — a victim of friendly fire in a battle with an imaginary enemy. An EpiPen will prevent death if it's used within 20 minutes of exposure, but after that window, the reaction cannot always be halted.

Until recently, one common explanation for the rise of food allergies was the so-called hygiene hypothesis — the idea that the modern obsession with sanitation prevents children from being exposed to the bacteria and the parasites that enable the immune system to develop. With the immune system underemployed, so to speak, it begins to attack harmless targets. This hypothesis, however, has fallen out of favor; among other things, it turns out that allergies are rising worldwide, from Rio de Janeiro to Shanghai, in Europe and Britain, across environments that have a range of sanitary conditions. Moreover, children who grow up on farms with exposure to dirt and parasites and animals do not appear to have lower rates of food allergies (although they do have lower rates of environmental allergies, like hay fever or reactions to animals).

Having a parent or a sibling with allergies (food or environmental) increases the risk that a child will have food allergies. One study of identical twins has led researchers to believe that food allergies are about 70 percent genetic in origin and 30 percent environmental.

If allergies are primarily genetic, though, how could their incidence have risen so quickly? The traditional model of genetic change involves natural selection operating through a slow process of mutation, generation by generation, that, furthermore, results in traits that increase survival — not cause sudden death.

Epidemiologically, food allergies parallel the steep rise of other contemporary epidemics like asthma, diabetes and autoimmune diseases — a phenomenon for which there has been no convincing explanation. Emerging evidence suggests that food allergies, however, fall in the province of the new field of epigenetics: the science of how the environment can alter the genetic inheritance one generation passes on to the next. One focus of Nadeau's lab is studying whether the toxins found in pollution, pesticides or tobacco smoke damage the genes in ways that make children more likely to have allergies and the intimately related disease of asthma. There is evidence that having a parent or a grandparent who smoked — even if the child was never exposed to smoke — is a risk factor for food allergies, as is living in an urban area with elevated pollution.

Another focus of Nadeau's research is diet. "I see immigrants who come here from a country with a traditional diet, and then they adopt a Western diet, and their children have food allergies," she says. "We have to wonder: Is it the change in their diet or something else?"

Important evidence that diet during pregnancy and early childhood affects food allergies has emerged from Europe. A recent study found that eating hydrogenated oils found in fast food during pregnancy is associated with an increase in allergies and asthma. Another study suggested that eating a Mediterranean diet during pregnancy (fruits, vegetables, olive oil, yogurt, fish and so forth), high in vitamin D and calcium, is associated with a decrease. And according to a newly published study, introducing rice and corn at 6 months, like Gerber's rice cereal — "Baby's first food" — increases the risk of food allergies, while introducing wheat before 6 months and eggs and fish before 9 months results in a decreased risk for asthma and allergies.

Nadeau herself is aligned with the school of thought that speculates that the cause of food allergies could be "some element of interaction between genes and the environment — air pollution, tobacco smoke, chemicals in water or the food you eat. We are ingesting the proteins that are causing the allergies in a very different form and immune environment than when primitive man did."

As a child living on a houseboat off the Jersey Shore, Nadeau suffered from severe asthma. She recalls the horrible feeling of lying awake at night laboring to breathe, without understanding how the damp moldy environment contributed. Her mother talked fondly about having been a cheerleader, but Nadeau's interests ran to science, and her desires focused on service. In college and medical school, she worked to bring medical care to disadvantaged people, from treating torture victims in Haiti to documenting tuberculosis in children in immigration detention centers in Texas.

After receiving an M.D./Ph.D. from Harvard, she did a pediatrics residency, specializing in oncology-hematology. Most oncologists cope with their patients' suffering by detaching, but she couldn't, and she found the destructiveness of chemotherapy — the stark pain brought on by the treatment itself — too hard to bear. Because her dissertation focused on the immune system, she decided to switch to allergies and immunology. In between fellowships, she went to work in biotechnology and led the successful development of a new drug for non-Hodgkin's lymphoma. But she found working in the corporate world alienating and decided to return to research and caring for patients.

She and her husband, Paul, a neurosurgeon with a Ph.D., decided that, although they loved children, given the demands of their research, they should content themselves with a small family and a lot of pets. They had a son, Christopher, and a menagerie, but then Nadeau became pregnant with twin girls, Jennifer and Stephanie. She and Paul both took positions at Stanford, and they bought a fix-up ranch house in Los Altos Hills and spent weekends at Home Depot.

In 2004, when she was 38, Nadeau found out she had a rare cancer — cancer of the T-cell — for which there were no effective therapies. (T-cells play a critical role in the immune system and, coincidentally, the T-cell dysfunction that leads to food allergies was already a focus of her research.) Life expectancy was only a couple of years. "I went home and cried, and then I watched 'The Sound of Music' with my kids," she told me, "and that was the last time I cried." Although she said she doesn't believe you can pray illness away, she drew strength from her faith and the community of her church.

In a small subset of patients, however, the cancer cells are actually caused by a virus, and over time it became clear she was in that lucky group — instead of multiplying, her cancer cells disappeared.

"I carry that experience with me," she said of having cancer. "Now I know how patients feel when they get these diagnoses." Soon thereafter, she found herself unexpectedly pregnant with her second set of twins, Katherine and Elizabeth. She became even more focused. "Every moment is managed," she said, from the half-hour she spends on the elliptical machine (while reading e-mail on her iPhone) to the spreadsheet on which all five children check off that they have completed their daily piano practice. She takes the children to scientific conferences when she travels, and she and her eldest three spent their last summer vacation on a trip to rural Mexico building houses.

I met Nadeau a year and a half ago, after moving to Palo Alto, when we consulted with her about our son, Kieran, who was nearly 2. When Kieran was 5 months old, my husband and I noticed that 20 minutes after he drank breast milk, a red rash appeared around his mouth. The same thing happened with cow- and soy-milk formula, so we took him to an allergy doctor, who recommended hemp milk, a watery substance extracted from the hemp plant. When he started solid food (beginning with rice cereal), she suggested we try cheese, as the milk proteins in cheese are different from those in raw milk.

Early one evening, I gave him a nibble of Havarti — which I chose because it somehow struck me as particularly innocent. Immediately he began to fuss and rub his eyes. I stepped into the bathroom to get him some eye drops; when I returned a few seconds later, I almost didn't recognize him. It was like a moment in a fairy tale when a child is transformed into a beast: his face was swollen and covered with scaly red splotches, and his eyes looked small and sunken. We ripped off his clothes and saw that horrible scales were spreading down his torso. He began to wheeze and gasp for breath. We threw him in the car and drove to the hospital, a few minutes away.

"Next time," the E.R. doctor told us, "just call 9-1-1."

Next time? I thought, staring at him with horror. What kind of parents do you think we are? There will never be a next time. But there was: a next time and a time after that and a time after that. By the time he was 3, he had to be revived with an EpiPen three times. He turned out to be allergic to legumes as well as to dairy and soy, and to the egg in the dairy-free apple cake at his first birthday, so that the party ended with a hospital stay. Six months later, he got hold of a cookie that contained eggs and nuts. I saw him pop it in his mouth, and I grabbed it before he bit down and dragged him to the sink to wash his mouth out. I thought I caught it in time, but we decided to drive to the hospital and sit in the parking lot to be safe. In the car, red hives began to speckle his face, and his lips turned blue. I didn't know then that blue lips are a sign of oxygen deprivation, but I understood something terrible was happening. I took the EpiPen and plunged it into his thigh.

By the time he was 2, he outgrew his allergies to dairy, soy and legumes but not to eggs or nuts or sunflower seeds, and like Persephone's pomegranate seeds, they have brought him to the mouth of Hades, again and again. Each night, I would wake to check him in his crib. I would touch his face until, bothered, he would sigh or flop over, and I would know he was only sleeping. Sometimes I'd wake from a nightmare in which all of our EpiPens had disappeared and he was gasping for breath again, his eyes round with terror. I would open the drawer to my night table, where I keep an extra EpiPen for reassurance, and close my fingers around it. One in a thousand. During the day, the mortality rate seemed slight, but at night it loomed large.

Kieran began treatment last March, the way all the children in the trials begin, by verifying his allergies through a "food challenge," in which just enough of the food is given to provoke a small reaction. His first challenge was to cashew — a nut we had been told he was allergic to based on blood work but which he had never actually eaten. I knew that Nadeau had done more than a thousand food challenges and rarely had to use an EpiPen, but I still felt sick with dread. He began at one milligram (1/360 of a cashew) and then, an hour later, a larger dose. After a few more doses, I was beginning to wonder if he actually had a cashew allergy. Then he began to cry and flail his arms, as if he had been dropped overboard and was trying to keep from drowning. One by one over the next 20 minutes, the maximum doses of different medications failed to halt the hideous metamorphosis, and they had to use the EpiPen.

Forget this, I thought. My son is not going to be the canary in the coal mine.

Kieran fell asleep in his car seat on the way home, and he woke up happy as always, chattering about a squishy purple rubber toy that Tina, the physician assistant at the clinic, gave him. By the time we were home, my own feelings had changed; I realized that the severity of his reaction only underscored the need for the treatment.

For some children, food allergies are part of what is known in medical literature as the atopic march: a classic progression of immunological problems, from having eczema as a baby to developing food allergies between 1 and 3 years and then asthma between 4 and 6. Asthma and allergies are a dangerous combination, because allergic attacks initiate respiratory distress that sets off asthma; having both conditions greatly increases the risk of dying during anaphylaxis. Because both are believed to be part of a common pathology, one important research question Nadeau is trying to answer is whether treating food allergies early in a child's life could prevent the child from developing asthma. Food allergies also impair growth: children with food allergies have a lower body mass index than children without, and those with more than two allergies are shorter as well.

Justine (who asked that I not use her last name in order to protect her family's privacy) has a son who experienced just this sort of frightening trajectory. When Jack was 8 weeks old, Justine decided to stop breast-feeding so she could go back to work at the nonprofit she founded, and "all hell broke lose," she recalled. Jack developed severe eczema, consumed very little formula, cried inconsolably when they tried to feed him and rapidly lost weight.

She took him to countless specialists from the time he was an infant, but it was not until 2009, when Jack was 4, that Justine found her way to Nadeau, who diagnosed severe allergies to nuts, citrus, watermelon, garlic, onion, cucumbers and many other foods. By the time he was 6, he outgrew all his food allergies except cashews and pistachios. But his asthma was very severe; for unknown reasons (perhaps because the asthma made the act of breathing itself burn too many calories), he stopped growing when he was 5, and by the time he was 7, he had dropped to the 20th percentile in height from the 70th, while his weight fell to below the first percentile. He also suffered from severe reflux and, perhaps because of that, was often nauseated and reluctant to eat.

Although Justine was a founder of the Safar Council, she was scared to allow her own little boy to participate in the multiallergen trial, because she felt that the science was too new and feared destabilizing his precarious health. Then one day in first grade, Jack was eating a sandwich Justine prepared with whole-wheat bread, and he suffered a severe allergic attack. When Nadeau analyzed the bread in the lab, it turned out to contain traces of cashew. (The company subsequently added a label stating that the bread was made on shared equipment with cashew. Labeling in the United States permits companies not to list so-called trace contaminants, which are defined as up to roughly 200 milligrams, or about a whole peanut — more than enough to cause anaphylaxis in some people.) It was clear to Justine that there was no way to protect Jack — any time he ate a bite of anything, his life was at risk.

"I suddenly realized the trial was the opportunity of a lifetime," she said. "We couldn't afford to wait any longer."

For Tessa, not eating dairy, wheat, egg, peanut and almond had been the inviolable commandment of her life. Now she was being asked to forget all that. Her starting dose, in January 2012, when she was 9, was one milligram each of flour made from the proteins of her allergens — a dose so small it looked like a few shavings of nutmeg. Over the course of treatment, she would increase the amount to 4,000 to 6,000 milligrams of each food — the equivalent of about 2 pieces of toast, 1 egg, 1 cup of milk, 16 peanuts and 16 almonds.

She and her parents had to commit to a rigid regimen of taking the dose every day at roughly the same time. Each afternoon, Tessa would have to hurry home from diving or soccer practice to eat dinner so she could take her dose on a full stomach two hours before bed. Her parents would then anxiously monitor her for two hours, while she did homework, avoiding exertion, hot showers or anything else that would make her body temperature rise, which can facilitate anaphylaxis. If she reacted to a dose at home the night before (or if she was sick), she could take a half-dose the next day, but she could never skip a day.

Every two weeks, she would visit the hospital for what is known as an updose.On those days, Tessa would dawdle while her family filled the small hospital room, eyeing her anxiously as she finally swallowed her dose mixed with applesauce, then waiting to see if her body would accept the increased amount. Spirited and sassy, Tessa could go from dancing around the room before the dose to sullenly working on a sticker book of cupcakes. Her favorite TV show was "Cake Boss" — she loved looking at all the magnificent varieties of cakes made with ingredients she had never tasted. Although she had perfected a theatrical glower when her mother displeased her — along with a reproachful "Mom!" — the intimacy between them was always visible.

Many parents try to keep their children from the knowledge that their allergies could kill them. (When Maya Bodnick, who is 9, told me she could get sick from nuts but knew she could not die, her mother's eyes met mine over her head.) But Tessa was all too aware of what was at stake. When I asked her to describe what would happen if she accidentally ate an allergen, she replied sardonically, "I might take a long nap underground."

Tessa was 1 of 35 children in a multiallergy clinical trial who were injected with a powerful drug, Xolair, that suppresses allergic reactions. Xolair, which has been approved for use in patients with asthma (which Tessa also has), blocks the allergic response by inhibiting the IgE antibody that creates anaphylaxis. Using Xolair, Tessa was able to rapidly escalate her doses and complete her treatment in just four months — a fraction of the time it would ordinarily take.

My son, Kieran, was too young to take Xolair, so his treatment progressed slowly, and he sometimes had reactions, which meant his updose would have to be postponed. We were mystified to find that after days of giving him a particular dose without incident, a cluster of hives would appear like a swarm of insects on his cheeks, or his face and hands would swell, and he would cry and ask for a Band-Aid or vomit or say, "My food hurt me," and show me his swollen tongue.

Each time he had a reaction at home, I was able to reach Nadeau immediately, and she talked me through using various medications to control the reaction. She then stayed on the phone until the reaction subsided, so that we didn't have to spend the night in the E.R. Once, I reached her on her cellphone while she was on vacation with her family in Hawaii. She talked me through the reaction for the better part of an hour and, when the crisis passed, she told me I could call back in the middle of the night if Kieran had any problems later — her cellphone would be by her bed.

"My family understands I need to be available to my patients 24/7," she said.

I rued having interrupted her vacation, but I found out that a half-dozen other parents in the trial had called her with their own questions and crises as well, and that she considered her trip, and indeed all her vacations, to be holidays for her family and "working vacations" for her. She has a missionary temperament, which is fortunate, because the job requires it. One problem with trials involving children is the perennial shortage of saints: of physicians who are not only willing to be constantly available to anxious parents but who have the knack for forging strong alliances with the children themselves to motivate them.

Kieran was young enough to love the treatment — going to the hospital with his parents and being petted and praised by Nadeau and the staff. (His sister, Violet, felt left out and was thrilled when we discovered that she had developed a peanut allergy and can be in a trial this spring.) But the older children's feelings about their treatment were more complicated — a problem that has plagued other trials. Single-food-allergy trials in the United States and Europe have reported dropout rates of about 20 percent. Nadeau is intent on avoiding that.

"If you feel sad or discouraged, you call me — you," she tells the children, leaning in, entwining her pinkie in theirs and asking them to make her "a pinkie promise" that they will take their dose. She tells patients to call her by her first name, and her light, musical voice and lack of a white coat in the clinic all contribute to the magical different-from-other-doctors place she occupies in her young patients' minds. She gives them presents on every possible occasion or lets them pick out books or puzzles or Play-Doh from a bucket in the office. She made a deal with Justine's son, Jack, that if he stuck to his dose of the cashews he loathed, she would arrange for a special tour for him at the zoo where her son, Christopher, volunteered. She conveys to all the kids that they are doing something wonderful — something important that they are the very first to try — but that she knows it is difficult.

"The main thing is to keep going," she tells patients again and again. "Failure is not an option. We believe that everyone who keeps going succeeds. Everyone."

Last May 4, friends and family crowded Tessa's hospital room for her final dose: a full serving of egg, wheat, milk, peanuts and almonds. She ate; an hour passed and then two and then . . . nothing. Everyone squealed and cheered and cried as Kim brought out a celebratory cake. It was the first real cake — an ordinary white bakery cake with fluffy frosting — that Tessa ever had. Three years had passed since Kim first heard Nadeau speak, and her vision that this woman would change their lives had been fulfilled. Tessa was the first person in the world to be desensitized to multiple allergens at the same time.

That night, family and friends all went out to a pizzeria (where Tessa asked me what pepperoni was), followed by her first taste of ice cream. Congratulatory texts poured into Kim's phone, and she sent her friends snapshots of something amazing: her family, eating pizza together, like any other family.

But profound change is profoundly unsettling. Tessa lost a defining aspect of her identity. Would she not be special anymore? Would she get less of her mom's attention? Tessa still wouldn't eat any of the food that was formerly unsafe at parties or at school. While Tessa told a few close friends about the treatment, she was reluctant to reveal it more widely. When I asked her why, she pursed her lips and studied her nails. Was she afraid that it would make people doubt the severity of her former allergies? "Yeah," she said without looking up.

But over the next eight months, Kim watched as her daughter's anxieties began to dissipate. Tessa interacted more with other children in social situations and had her first sleepover. She still has a few allergies (shellfish and some nuts), but they are less burdensome. She had been on the Stanford Junior Olympics diving team, but during the trial, she became too anxious to dive. For months, Kim took her to practice, but she would just sit by the pool, unable to bring herself to get in. This winter, six months after completing the treatment, she decided to try again. She is currently working on a reverse tuck, which involves flipping backward, taking care to avoid the diving board on the way down. "Anyone can do soccer," she told her mom, "but diving is special."

Nadeau stresses that oral immunotherapy is still experimental. Her patients are not cured; they are desensitized enough that they can tolerate their former allergens. The reason that she doesn't call it a cure is that the child must continue to eat a maintenance dose of the food every day to avoid regaining the allergy. She often explains to her patients, "If you get off it for three days, you may become sensitive again." An egg-allergy trial found that when patients were taken off the maintenance dose for a month, roughly 60 percent regained the allergy (and there was no way to predict who those patients would be).

Eating a full serving of a food they previously regarded as toxic was challenging for many patients. They had to overcome the instinctive revulsion every day. Tessa found milk and cheese weird and eggs so sickening that she once threw up when her mother coaxed her into eating them, so she stuck to egg-white powder. When the children's blood and skin-prick testsbecome negative tothe allergens, which happens somewhere between six months and three years on the maintenance dose, Nadeau believes that a small amount of the allergens (for example, one peanut a day or the amount of egg and milk in one pancake) will be enough to prevent the allergy from returning.

For some of the children, when they had their last updose, the desensitization was complete, while for others, ghosts in their immune systems would sometimes surface, and they would have reactions to their maintenance doses. As a precaution, everyone was advised to keep carrying their EpiPens.

Jack was one of Nadeau's most difficult cases. In the months following his final updose last June, he suffered severe asthma, unexplained rashes, stomach pains and difficulty eating. He developed an inflammatory condition in his stomach that may or may not have been related to the irritating presence of his daily maintenance dose of 16 cashews. In September, Justine sent me an e-mail: "He is such a good boy, but he is tired. It makes me sad thinking he does not know any different. He is sick more than he is not."

At one of Jack's many long appointments, I asked him what it was like to have allergies. He looked down, his long lashes brushing his cheek. "It's hard," he said. I watched his tawny arm, graceful as a ballet dancer's, as he folded the wings of an origami swan he was making. "I don't really like being EpiPenned."

A silence fell. His mother's face sank into shadows. He has an unusually sweet and patient disposition; he would sit in the hospital bed without complaint, hour after hour, doing origami, puzzles and workbooks. But I recalled the day I saw him have an anaphylactic reaction during an updose and how agonized Justine was. "It's hard, because you tell your child to trust you and that the hospital is the safest place to be, and then out of nowhere, when their guard is totally down, they have to be EpiPenned," she reflected. She also told me about the day Jack met Brian Hom's son Steven, and after hearing that Steven's brother died of a peanut allergy, Jack asked Justine if he had almost ever died. Justine had to leave the room so he wouldn't see her cry.

The thought that his treatment serves the larger purpose of the trial seems to help him to get through it. In the fall, he was scheduled for a procedure in which a camera is slid down his throat to photograph his stomach to see if medication had been effective. When Justine told him it would have to be postponed because he had a cold, he said, "But I have to do it to help the research for other kids, right?"

Justine was startled: she had never told him that. "I don't know where he comes from," she said. "He is very focused on helping other kids through his hardship."

Justine, meanwhile, never doubted her choice to be in the trial. In an e-mail, she wrote: "And Kari — as always — will always be there for him. She is our godsend. Our angel. I don't have words to describe how I feel about her. . . . This has been the hardest journey of our lives, and she has saved my boy. Many times."

Last month, Jack's skin tested negative, and he is no longer allergic to either cashew or pistachios. (Treating cashew also treats pistachio, because the same protein causes both allergies.) He has gained a little weight and grown a small but significant amount, perhaps because his asthma is now moderate rather than severe, which Nadeau says may be a result of his participation in the Xolair trial. But he has to be careful; once he had an anaphylactic episode at school, and if he exerts himself within two hours before or after his dose, he can have a bad reaction. He is currently working on a cookbook to give to kids in the trial of all the ways to disguise disgusting nuts: in sushi, inside an olive or, his favorite, covered with chocolate and mixed into ice cream.

Just before Christmas, our son, Kieran, reached his first — and, for us, his most important — milestone. Nine months after we began, he ate 300 milligrams of each of his allergens (the equivalent of more than a full peanut, almond, cashew and hazelnut, as well as a bit of egg) and is no longer allergic to trace contaminants. Although it will take at least another year before he can eat a whole serving, he had become 300 times less allergic than he was at the start of the trial. When Violet's peanut allergy is treated, our family can take down the sign on our front door forbidding anything made with eggs or nuts and welcome Girl Scout Cookies.

After Kieran's updose in the hospital that day, we went down to the cafe to celebrate. We sat near the hospital's glittering Christmas tree, listening to piano music, and I bought a turkey sandwich for us to share. For the first time in his life, I didn't have to worry whether the bread had been cut with a knife or made on a surface that had also been used for mayonnaise or peanut butter. Kieran liked making a game of sharing the sandwich — a bite for him and a bite for me — but then he said, "I wish it all mine," and when I told him he could have it, he slid off my lap and hopped up and down with joy.

The current studies at Stanford are in the analysis phase. The next and final phase of Nadeau's multiallergy trial, which is in development, will very likely involve more patients and take place at multiple allergy centers across the country. Then, if that is successful and the treatment is approved by the F.D.A., it may eventually become standard. Research in Nadeau's lab has found that treating allergies actually changes the genes by epigenetics (a chemical modification of the genes that does not affect the actual DNA sequence), so desensitizing children may reduce the likelihood that they will pass on the disease to their children. One day, perhaps, fatal anaphylaxis may become a sorrow associated with an earlier age, like dying of appendicitis or polio.

I sometimes have an image of Kari Nadeau taking Kieran and Jack and Tessa and the other children by the hand and guiding them step by step, dose by dose, away from the grave. Of course, none of us will ever know which, if any, of our children would have slipped over the edge, but we do know the treatment rescued us from a lifetime of peering into the abyss.