It took a lot of digging to bring back to life the Spanish influenza virus of 1918. Some was done with invisible molecular primers in a PCR machine in Rockville. Some was done with pick and shovel in the frozen ground of Alaska.

Either way, it was a huge amount of work on a project whose chance of success at the start seemed very, very slim. Now, it will go down as one of the most astonishing technical feats in the history of science -- the viral equivalent of bringing dinosaurs back in the fictional "Jurassic Park."

It may also prove to be unusually useful -- not an elaborate biological parlor trick, but a vital service to global public health.

The Spanish flu killed at least 50 million people around the world in slightly more than a year -- late winter 1918 into the spring of 1919. Researchers have never figured out what made the virus so lethal, in part because there were no samples to study. Although viruses had been discovered by 1918, the flu virus was not isolated until 1933.

With the genome of 13,600 nucleotides known and published in the journals Science and Nature, the 1918 virus is already shedding light on its own history. It was a bird virus that appears to have become a human virus through the slow accumulation of mutations, not through the sudden trading of genes with another flu strain.

It is also illuminating the possible future of viruses that are worrying flu experts now. Some of the H5N1 "bird flu" strains seen recently in 10 Asian countries carry a few of the mutations seen in the 1918 virus, suggesting that they, too, may be slowly adapting to human hosts.

With more work, scientists will probably be able to figure out why the 1918 strain was so dangerous. Experiments with the reborn virus began in August at the Centers for Disease Control and Prevention in Atlanta and have already answered some questions, which may lead to better vaccines and drugs.

The story of how this feat came about has several beginnings. In hindsight, it is clear that perhaps the crucial one occurred 55 years ago with Johan Hultin.

Searching in Permafrost

Hultin had taken a break from medical studies in his native Sweden to study for a doctorate in microbiology at the University of Iowa. At a departmental lunch in 1950, he heard a professor make a passing reference to the idea that intact samples of the infamous 1918 strain might still exist in bodies frozen in the Arctic. Hultin was looking for a dissertation project. He proposed to his adviser that he try to recover the virus for use in a vaccine. The idea was approved.

While the percentage of people who became ill and died of the 1918 flu -- the "case-fatality rate" -- was 2 percent to 5 percent in the United States and Europe, it was more than 50 percent in some isolated native groups. In Alaska, some villages were virtually wiped out.

Hultin had spent the summer of 1949 in Alaska, helping a paleontologist named Otto Geist perform excavations. He had driven up on the newly opened Alaska Highway, which he said "was itself a great adventure." He figured there were mass graves from the 1918 pandemic there. He wrote Geist and asked him to contact missionaries working in Inuit villages. Specifically, he wanted to know whether there were records of epidemic deaths in 1918 or 1919, and if so, what the symptoms were.

Hultin heard from seven or eight missionaries. They sent him notes copied from mission record books, often in Norwegian, which he could read. He got a map that showed the extent of permafrost -- land where the ground never thaws. He chose three villages in the permafrost zone that had mass graves containing corpses from an epidemic that sounded like influenza.

The young graduate student surveyed the sites, all on the Seward Peninsula, which stretches westward into the Bering Sea. In one, a river had changed course, disturbing the permafrost. In another, a beach had eroded, exposing the grave. But the third, a place called Teller Mission, looked good.

Seventy-two of 80 residents of Teller Mission died between Nov. 15 and 20, 1918. The Army buried the victims with a steam-powered excavator used by miners.

Hultin went to the village, whose name has since been changed to Brevig Mission, and requested permission to excavate the grave. Through a translator, he emphasized the benefit of making a vaccine. The villagers had been vaccinated against smallpox, so they knew what he was talking about. And at the meeting were three of the eight survivors from 1918.

"They told us their terrible story about all the other people in the village dying. That convinced the rest of them to let me help," Hultin recalled recently.

On June 25, 1951, he, two Iowa professors and the paleontologist went to work. They dug through three feet of tundra and gravel, and then three feet of permafrost. They wore masks. There were no observers or reporters. They sampled four bodies; all had evidence of pulmonary hemorrhage, the hallmark of rapid death from influenza alone. They took blocks of tissue from various organs and quickly put them into steel containers that were then sealed in steel boxes.

"Preserving the specimens and getting them safely and quickly to their medical laboratories in Iowa City was now the problem," wrote a Washington Post reporter three months later in a brief account.

"A wild storm whipped the bay to waves of almost impassable heights. Dry ice, brought from the States to refrigerate the specimens, had evaporated . . . In the emergency, the scientists used a fire extinguisher whose foamy carbon dioxide contents, spurting from its nozzle, formed dry ice. With native help, the expedition members detoured the hazardous bay crossing, made their way overland to a narrow strip of the bay, and got back to the town of Teller," wrote the reporter, N.S. Haseltine.

Back in Iowa, Hultin thawed the tissue and tried to recover the virus. He exposed ferrets -- the species whose response to influenza is most like people's -- to tissue extracts. The animals did not get sick. None of his experiments succeeded. He concluded there was no live virus in the Inuit corpses.

Hultin believes he could have gotten a doctoral dissertation out of this meticulous but failed effort. But he never got around to writing it. Soon after his many months of experiments had proved fruitless, he was invited to enter medical school at the University of Iowa. He accepted the offer, became a pathologist and spent much of his career at a hospital in California. Now retired, he turned 81 on Friday.

No scientific publications came out of Hultin's project. But it was not entirely lost to history. A historian named Alfred Crosby mentioned it briefly in his 1989 book, "America's Forgotten Pandemic: The Influenza of 1918."

As it happened, Hultin was not the only person who attempted to get the Spanish flu virus out of the ice. The same year he tried, U.S. Army researchers did also. They excavated a mass grave near Nome, Alaska, finding only skeletons. Hultin had been there three weeks earlier and had rejected the site.

Four decades later, however, the Army returned to the story.

A Key Institution

One of Washington's more obscure but important institutions is the Armed Forces Institute of Pathology in Rockville. It provides pathology services for the military, including autopsies of war dead. It also functions as a kind of Supreme Court for difficult cases. Pathologists unsure of a diagnosis, for a small fee, can consult its experts and send them microscopic slides or other samples for review. Part of the institute's value lies in its pathological specimens dating to 1862 -- 3 million pieces of preserved human tissue.

Jeffery K. Taubenberger is a civilian pathologist who heads the institute's division of molecular pathology. His laboratory is one of the few in the country with expertise in rescuing and restoring genetic material from damaged or decayed tissue. In 1995, Taubenberger wondered whether it might be possible to get the 1918 virus out of dried and fixed tissue from the Spanish flu pandemic. "I really wanted to see if there was some way we could make use of this vast, wonderful collection for this," he recalled.

He and his colleagues reviewed slides of lung tissue from 78 soldiers who had died in the pandemic. They narrowed the search to 10 slides in which the microscopic appearance showed that the men died only of viral pneumonia, not of a secondary bacterial infection that was more often the cause of death.

They tested preserved, leftover pieces of lung tissue from all 10. Two came up positive for influenza A, the broad family that includes Spanish flu. One was from a 21-year-old private who died in South Carolina on Sept. 26, 1918. The other was from a 30-year-old private who died in Upstate New York on the same day.

Using polymerase chain reaction (PCR) technology to amplify the genetic material, and primers -- short, important stretches of genetic material -- from human, animal and bird viruses, Taubenberger, Ann H. Reid and Thomas G. Fanning fished out fragments of the 1918 microbe. There were multiple copies of the virus in the sample, but they had broken into small pieces. Matching the overlapping ends of the fragments, the researchers reassembled the fragments in the right order.

The first gene they recovered, called NS, was virtually identical in the two cases.

Influenza A has eight gene segments. When Taubenberger published the report on the first one, Hultin read it. He realized, at long last, that there might be value in dead Spanish flu virus -- and he thought he might still have a source. He contacted Taubenberger and asked if he would be interested in frozen organs of 1918 victims, should any still exist. Taubenberger said yes. Hultin set off two weeks later.

He returned to Brevig Mission and again sought permission from the village council to dig. "I said that the virus was dead in 1951 and was even deader now," he recalled.

The village leaders talked a long time in Inupiat, the local language. They were worried about the release of evil spirits, not contagion, Hultin said. Then someone recalled that the victims had received Christian burials, which were supposed to have chased away the evil spirits. Permission was granted.

On Aug. 20, 1997, Hultin and a local crew opened the grave. The four bodies he had sampled in 1951 were decomposed. But he found one that had been missed the first time. It was of a woman in her thirties who was very fat. All that was left of her clothes was a row of bone buttons lying on her chest. But her body was intact and frozen, apparently insulated by the fat from the occasional brief thaws. "I sat on an upside-down pail and I looked at this, and I got the flash in my mind," Hultin said. "Maybe this is where I can find it."

With only gloves and a face shield for protection, Hultin removed her lungs and sampled her spleen, liver and heart. He cut the tissue into one-inch cubes and put them in a preservative solution. The grave was closed for a final time.

Hultin and Taubenberger hoped the Alaska material would contain virus material that was more nearly intact than the material from the soldiers. It did not. In fact, it was a bit more fragmented. The longest strands of RNA -- flu's genetic material -- in the institute's slides were about 130 nucleotides, or letters, long. In Hultin's material, the longest was 110.

Nevertheless, Hultin had provided Taubenberger with all the material he would need to reconstruct the 1918 virus. Eight years later, it was done.