ALO ALTO, Calif.

ANY absent-minded scientist can let a saucepan of milk boil over. But it takes a true genius to do it twice in 20 minutes.

Harold McGee, who first bonded rigorous science to popular cookery in his 1984 book, "On Food and Cooking," was doing a run-through for Thanksgiving dinner. His culinary investigations did not end with the book's publication. He has spent the last 10 years working on a revision, just published by Scribner. Illustrating some of his new findings in his kitchen here, sliced apples for pie were dehydrating in a bowl, green beans were boiling in two pots of water (one salted, one not, to learn the difference), and a fat turkey sat on the counter with ice packs resting on its breasts.

Chilling the white meat, Mr. McGee explained, makes it cook slower than the rest of the turkey, preventing the breasts from drying out before the dark meat is done. "I used to strap the ice packs on with an Ace bandage," he said. "But my family found it too unappetizing."

Just then a pan of milk that was supposed to be caramelizing foamed over and spilled onto the stovetop, illustrating the peculiar behavior of milk proteins exposed to high heat. "They migrate to the inside of the steam bubbles and bond there," Mr. McGee said ruefully, looking for a kitchen towel. "Then they stretch and stretch, and when they can't get any bigger, they collapse."

The challenges of Thanksgiving have caused many an accomplished cook to crack like a pecan. Too many recipes to coordinate, never enough oven space and too many people underfoot seem the culprits. But Mr. McGee knows better. "What makes Thanksgiving food hard," he said, "is that you're trying to cook these bundles of different matter — white meat and dark meat, apples and pie crust — while they're all stuck together. You want some parts to be soft and moist and some to be crisp and dry. No wonder it's so hard to get them just right."

Insights like these, giving cooks new ways to think about what they do in the kitchen, have helped Mr. McGee gain renown and a cult following over the last 20 years, although his formal scientific training is minimal, and he has no university affiliation or laboratory access. When Mr. McGee began his research, he had a Bachelor of Science degree from the California Institute of Technology and a doctorate in English from Yale, where he was teaching literature.

"On Food and Cooking" has sold more than 100,000 copies, and the revised edition includes mostly new material, which reflects both Mr. McGee's further research and American food fashions. "I could never have anticipated that people would have this bottomless appetite for information about chocolate, coffee and tea," he said. "Not to mention foams, caramelization and trans-fats."

Gathering strands of botany, history, animal husbandry, genetics, chemistry, thermodynamics, physiology and physics, "On Food and Cooking" answered eternal kitchen questions: what makes hollandaise curdle, how onions make us cry, what causes cakes to fall. Some of Mr. McGee's findings challenged the strict rules that still governed cookery at that time, most based on traditional French methods. His assertion that searing meat does not actually "seal in" the juices rocked the culinary world.

"As a cook I wanted to believe that chefs were right, that their experience of doing these things over and over must prove something," he said. "But as a scientist I could see that the evidence didn't hold up." The point of searing, he said, is to create flavor through the browning of surface chemicals. Its only effect on the juices is to pull them out of the meat, making it drier.

The curiosity that fueled the experiments in Mr. McGee's first book is undiminished after 20 years, and his approach to cooking is still skeptical. He tries to take as little as possible for granted, asking at each step: Why am I doing this? Is there a better way? All this questioning has yielded conclusions, some more useful than others, and many of them heretical in culinary circles.

For example, although brining the turkey is now part of the Thanksgiving ritual for many cooks, Mr. McGee does not do it. "The bird does become juicier, but it's just absorbing tap water, not the true juices that make a bird flavorful," he said. "And the drippings become so salty that you can't use them." He says that his own experiments with turkey, though far from complete, show that drying the bird out, rather than infusing it with water, is more likely to make it flavorful and juicy with crisp skin. He unwraps his turkey a day or two before cooking, letting it air-dry in the refrigerator, and then cooks it at high temperature.

After our 13-pound bird had roasted for only an hour and half, the readings from four temperature probes attached to a Fluke 50 lab thermometer showed that the meat was perfectly done both in the thigh and the breast. "White meat and dark meat are completely different kinds of muscles in both birds and animals," Mr. McGee said. "White muscles store energy for sudden, rapid movements and dark ones for slow, sustained ones. That's why chickens, who only fly for a few seconds when they're startled, have white breasts, and ducks, who fly for long periods, have red ones."

But the skin was only golden, not as brown and crisp as we wanted. "Crispness is a matter of heating and dehydrating the proteins," Mr. McGee said, as he fired up an industrial-size blowtorch and began methodically stroking its blue flame over the turkey. Dissatisfied with the slow results, he switched to a heat gun, whose red-hot coils seemed to give a more concentrated heat. The skin turned from gold to clear and then to bronze, the juices visibly running out of it and through the turkey. "Caramelized has become a popular word since I wrote the first book," he said. "But everything browned is not caramelized."

Caramelization, he explained, is what happens to sugar — simple sucrose molecules — exposed to high heat. But the browning that takes place in savory foods like onions, potatoes, celery and turkey skin is a "Maillard reaction," the explosive meeting of a carbohydrate molecule (which may or may not be a sugar) and an amino acid in a hot, dry environment.

Maillard reactions take place when coffee or cocoa beans are roasted or when a bread crust turns brown. Mr. McGee said: "Maillard reactions contribute even more to the pleasures of eating than caramelization does. But of course it doesn't sound as good on a menu."

The alluring scent of Maillard reactions filled the kitchen as Mr. McGee's pie crust began to brown. Although it is almost impossible to do anything truly new in the kitchen — as Mr. McGee notes, it often turns out that even the most complex flavor combinations were routinely used by Roman cooks — his pie crust method seems revolutionary. "The goal of pie crust is to create thin, even layers of fat and flour," he said. "That's what makes them flaky. But the usual method isn't really optimal for that."

Instead of using his fingers to rub globs of fat into flour, then dribbling in ice water, Mr. McGee starts with square chunks of cold butter and a pile of flour on a board. With a rolling pin he presses and rolls the butter into the flour, flattening it into thinner and thinner flakes. Occasionally he scrapes the mixture into a bowl and freezes it for five minutes, to keep the butter from melting. Since the gluten is not activated until the water is added, there is no worry about overworking the dough, even though the process can take some time.

Finally, to add the water Mr. McGee fetched a plant mister. "I always found it was hard to get the water evenly into the dough" he said. "So I measured how many sprays of the mister it takes to get half a cup of water — it's 150, by the way — and I use that to get uniform droplets." Now working quickly, he lightly squeezed the mister over the dough 50 times, then turned the dough and folded it. After two repetitions the dough just held together. He divided it into two round discs and returned it to the freezer.

Hours later his careful work paid off in a golden-brown crust of unspeakable flakiness and buttery flavor: Mr. McGee's method means that there is no need to add shortening to ensure a good texture.

Although Mr. McGee is an excellent cook, and his work has many practical applications, his research has taken him into more ethereal realms of aroma and flavor.

Tracing molecular compounds through the universe of food, he is preoccupied with understanding why chestnut honey and corn tortillas smell almost exactly the same, how compounds called phthalides make celery and walnuts taste so similar and which terpenes are common to carrots and rosemary. "By the way, you can't do these kinds of flavor experiments with baby vegetables," he said politely, averting his eyes from some organic baby carrots I had brought along to roast. "Baby vegetables are completely undeveloped in terms of flavor, although they look nice."

At the end of the day the lab report was mixed. A successful dulce de leche was never achieved, though we had proved that an unwatched saucepan of milk always boils. Mr. McGee had the scientific explanation even for that. "You set the heat to boil a certain amount of liquid," he said. "Then as it boils down, you have less and less liquid and too much heat."

And for the record, as every cook knows, green beans cooked in salted water definitely have more flavor. "Somehow the salt must inhibit the osmosis of chlorophyll into the cooking water," Mr. McGee said. "Someday I'll figure out exactly how."