The genetic variant is so common that some 50 percent of people in European populations carry one copy of it, and about 20 percent of people have inherited two copies, one from each parent. It is much less prevalent in people of African descent, the scientists said.

How it works is not yet known, but carriers of a single copy have a 15 percent to 20 percent greater risk of heart disease, while those with two copies are up to 60 percent more likely to develop heart disease than people who have none. The risk is even higher for people who suffer a heart attack at an early age, defined as men under 50 and women under 60.

The new finding, published online yesterday in the journal Science, is one of a spate of discoveries about the genetic bases of common diseases. Last week seven new genetic variants involved in the most common form of diabetes were identified, and a batch of new genes from other common diseases is expected to be reported in the next few weeks.

These discoveries are a long-promised fruit of the $3 billion Human Genome Project, which was essentially completed in 2003. There have been two principal approaches to scanning the genome for disease genes, which are culminating in photo-finish results by the proponents of each method.

One competitor is DeCode Genetics, a private company based in Reykjavik, Iceland, that has used the comprehensive health care records and known genealogy of the Icelandic population to track disease. DeCode has dominated the gene-finding field for the last several years.

DeCode’s rivals are medical researchers based at universities in the United States and Europe. They have made a slower start because, without an Icelandic-type data set, they have had to wait for construction of the HapMap, a survey of common genetic variations on the human genome in Africans, Asians and Europeans. These common variations, known as SNPs or “snips,” are thought to be the genetic bases of the common diseases.

Both sides have been helped by a technical development, the construction by companies like Affymetrix and Illumina of instruments known as microarrays or chips that can now detect up to 500,000 snips. With the chips, the genomes of patients with a disease can be compared with those of healthy people, allowing snips that seem associated with the disease to be identified.

Last week both DeCode and three academic consortiums reported new diabetes genes. This week’s reports on heart disease come from DeCode and another academic consortium, led by Dr. Ruth McPherson of the University of Ottawa Heart Institute and Jonathan Cohen of the Southwestern Medical Center at the University of Texas in Dallas.

Dr. McPherson said that it came as a “complete surprise” that DeCode had submitted its report at the same time to the same journal, but that “at the end of the day we are very happy they came out with the same result.”

Both groups identified snips in a small region of Chromosome 9 (the human genome is packaged into 23 pairs) as being associated with higher risk of heart disease.

But as if to prove how much remains to be understood about human biology, the snips lie in a stretch of DNA that contains no gene or genetic element with known functional purpose. It is only on the basis of rigorous statistics that the two groups believe their snips must be causally associated with heart disease.

This means the researchers still have no idea of the mechanism by which the snips raise the risk of heart disease. When the mechanism comes to light, it may be possible to design drugs that interfere with it so as to avert the risk. For the moment, however, the practical use of the new finding is in diagnostic testing.

DeCode says it will use this and other heart disease-related variants it has found as the basis for a test that gauges the inherited risk.

But Dr. McPherson said, “It would be unfortunate if snips using these tests were patented and physicians couldn’t use them without paying a royalty to DeCode.”

A test would help address the big question in heart disease of for whom the many available interventions should be recommended, Dr. McPherson said. Risk is now assessed on conventional factors like smoking.

But Dr. Christopher O’Donnell, an expert on heart disease genetics at the National Heart, Lung and Blood Institute, said it would be premature to recommend testing until the mechanism of the new variant was known. Figuring out the mechanism would be a major fruit of the new studies, which have produced “one of the few strongly replicated findings that has emerged from the field of heart attack and coronary heart disease,” Dr. O’Donnell said.

The new heart disease risk snip is much less common in Africans and African-Americans and seems not to be correlated with heart disease in these populations, the Ottawa-Dallas team reported.

The idea behind the HapMap, that common genetic variants are associated with common diseases, assumes that natural selection is unable to get rid of disease genes that act only late in life, well after the age of reproduction. But the heart disease risk variant is so extraordinarily common that it may confer some unknown benefit.

“It is clear that this variant must have some advantage or it couldn’t be in 50 percent of the population, so I think it’s likely there has been some selection for it,” said Dr. Kari Stefansson, DeCode’s chief executive.

The variant lies at a site on Chromosome 9 very close to one of the diabetes variants reported last week. At a news briefing last week, Dr. Francis S. Collins, the director of the National Human Genome Research Institute, commented on the coincidence that a single region of the genome could contain factors for both diabetes and heart disease.

“I think this is a stunner,” Dr. Collins said. “This is like the seat of the soul of the genome.”

Other biologists suggested the link hinted at an unsuspected common biology between the two diseases. But Dr. Stefansson dismissed this idea emphatically, saying he had tested it in thousands of people with both diseases. “I can say with substantial authority that these markers are uncorrelated,” he said, meaning that the two variants lie close to each other only by coincidence.