Random cell growth and division and random differences in which genes get turned on and how active they are during development can cause identical twins to have different numbers of cells in their kidneys and even different patterns of folds in their brains, Dr. Finch pointed out. And random differences in development early in life can set the stage for deterioration decades later.

But seemingly random events can still come as a shock. That’s how Annmarie Bald felt when her identical twin, Catherine Polk, died in her sleep of a heart attack. It happened seven years ago, when Ms. Polk was 43. To this day, Ms. Bald, of Forked River, N.J., lives in fear that the same thing will happen to her. She nervously sees her doctor every year for a checkup, and every year her doctor tells her the same thing: her heart is fine.

“The question in my mind every day is, ‘How did I end up still here and she’s gone?’ ” Ms. Bald said. “It’s not something you ever get over.”

Yet even diseases commonly thought to be strongly inherited, like many cancers, are not, researchers found. In a paper in The New England Journal of Medicine in 2000, Dr. Paul Lichtenstein of the Karolinska Institute in Stockholm and his colleagues analyzed cancer rates in 44,788 pairs of Nordic twins. They found that only a few cancers — breast, prostate and colorectal — had a noticeable genetic component. And it was not much. If one identical twin got one of those cancers, the chance that the other twin would get it was generally less than 15 percent, about five times the risk for the average person but not a very big risk over all.

Looked at one way, the data say that genes can determine cancer risk. But viewed another way, the data say that the risk for an identical twin of a cancer patient is not even close to 100 percent, as it would be if genes completely determined who would get the disease.

Dr. Robert Hoover of the National Cancer Institute wrote in an accompanying editorial: “There is a low absolute probability that a cancer will develop in a person whose identical twin — a person with an identical genome and many similar exposures — has the same type of cancer. This should also be instructive to some scientists and others interested in individual risk assessment who believe that with enough information, it will be possible to predict accurately who will contract a disease and who will not.”

Alzheimer’s disease also has a genetic component, but genes are far from the only factor in determining who gets the disease, said Margaret Gatz of the University of Southern California and Nancy Pedersen of the Karolinska Institute.

Dr. Gatz and Dr. Pedersen analyzed data from a study of identical and fraternal Swedish twins 65 and older. If one of a pair of identical twins developed Alzheimer’s disease, the other had a 60 percent chance of getting it. If one of a pair of fraternal twins, who are related like other brothers and sisters, got Alzheimer’s, the other had a 30 percent chance of getting it.

But, Dr. Pedersen noted, Alzheimer’s is so common in the elderly that it occurs in 35 percent of people age 80 and older. If genes determine who gets Alzheimer’s at older ages, Dr. Pedersen says, “those genes must be very common, have small effects and probably interact with the environment.”

As for other chronic diseases of the elderly, Parkinson’s has no detectable heritable component, studies repeatedly find. Heart disease appears to be indiscriminate, striking almost everyone eventually, says Dr. Anne Newman of the University of Pittsburgh, who has studied it systematically in a large group of elderly people.

But the general picture is consistent in study after study. A strong family history of even a genetically linked disease does not guarantee a person will get it, and having no family history does not mean a person is protected. Instead, chronic diseases strike almost at random among the elderly, making it perhaps not so surprising that life spans themselves have such a weak genetic link.

Matt McGue, a psychology professor at the University of Minnesota who studies twins, contrasts life spans with personality, which, he says, is about 50 percent heritable, or attention-deficit hyperactivity disorder, which is 70 to 80 percent heritable, or body weight, which is 70 percent heritable.

“I’ve been in this business for a long while, and life span is probably one of the most weakly heritable traits I’ve ever studied,” Dr. McGue said.

Seeking Rare Families

At the National Institute on Aging, the question still hovers: Is it possible to find genetic determinants of exceptional health and longevity?

“If you could identify factors for exceptionally good health, that might allow people to avoid disease,” said Evan Hadley, director of the institute’s geriatrics and clinical gerontology program.

There are two methods to do this, Dr. Hadley said. One is to look at how the genes of centenarians differ from those of the rest of the population. But, he said, that requires that if longevity genes exist, they are common among centenarians. And, so far, such studies have not yielded much that has held up — with one well accepted exception: a gene for a cholesterol-carrying protein that affects risk for heart disease as well as Alzheimer’s disease. Those who have that gene have double the chances of living to 100. But that chance is not much anyway. Only about 2 percent of people born in 1910 could expect to reach 100. The second approach is to look for rare genes in unusually long-lived families. “If there is something in a family, it may be in only one or a few families,” Dr. Hadley said. But it may have a big effect.

So the National Institute on Aging is starting a research project with investigators at three United States medical centers and at Dr. Christensen’s center in Denmark. The plan is to find exceptional families, those in which there is a cluster of very old, closely related members — two sisters in their 90’s, for example — whose children, who would typically be in their 70’s, and grandchildren, can be studied too.

Today, many families have a few members living to advanced ages, but very few families have many of them. And in large families, just by chance, someone may live past 90, but it is unlikely that most of the brothers and sisters will get there. For these families, there does not appear to be a genetic component to life spans.

For now, the study is in a pilot phase, testing a scoring system to define the families who seem to fit the criteria.

“If you are really, really old in a family, that gets you more points,” Dr. Hadley said. “You get more points for being 97 than for being 92. But we also are looking at the whole family structure. If there are just two siblings in a family and both live to 98, that’s very exceptional. But suppose there are eight kids and they all made it to 87. That’s pretty unusual, too.’’

If the researchers find genes in the oldest family members that seem to be associated with protection from a disease like heart disease and with a long life, they will follow the younger members of the family, children in their 60’s and 70’s, asking if the same genes seem to protect them as they age.

Some wonder if the project can succeed, said Dr. Newman, who is directing one study center, at the University of Pittsburgh. “The big debate is, is it possible for there to be a few genes that are protective or is it going to be so complicated that we won’t be able to figure out the genetic factors? Is it going to be that some people are just lucky?”

She is optimistic, reasoning that since some families tend to have early onset of certain diseases, others probably have a genetic predisposition to get diseases like heart disease, cancer and Alzheimer’s so late that most members do not get them at all and live very long and healthy lives.

“This would be the flip side of early onset,” she says.

Mrs. Tesauro is in the pilot study. She had always been healthy and active, a self-described tomboy growing up who played tennis until she was 85. “I just can’t sit still,” she said.

She was a woman who knew her mind, so eager to go to college that she defied her father, who thought it was a waste of money, and worked her way through. She ended up with a master’s degree in education and a career as a high school teacher.

Her twin was different. She was the frilly type, Mrs. Tesauro said, and was not much of a student. She failed a grade in high school and barely graduated. Both Mrs. Tesauro and her sister married and had children.

Mrs. Tesauro was born first, and it is a common belief even among scientists that the twin born first is stronger and lives longer. But when he looked at the Scandinavian data, Dr. Christensen said, he found that birth order made no difference in health or longevity.

The day before visiting Mrs. Tesauro for the first time, the Pittsburgh investigators tried to call her, just to be sure she was still alive and still healthy enough to be interviewed. When they could not reach her, they began to worry.

But all was well. Mrs. Tesauro answered the phone the next morning and explained why they had had such trouble. She was out running errands.

Live Long? Die Young? Answer Isn’t Just in Genes - New York Times