n one of the most remote spots on earth, a desiccated island in the Aral Sea, lie the remains of the world's largest biological-warfare testing ground. Since Russia abandoned Vozrozhdeniye Island in 1992 to its new owners, Kazakhstan and Uzbekistan, it has remained uninhabited, its laboratory complex deserted by all but the Kazakh scavengers who pilfer its parts.

The island was briefly used as a bioweapons testing range in the 1930's and then abandoned until 1954. Stalin, having caught up with the United States in nuclear weapons, had decided he was ready to return to biological ones. The Soviets planned to load their germs onto missiles and bombs; on Vozrozhdeniye Island, they concentrated on making them ever more deadly.

The only access to the island now is in the company of the scavengers, who say they began stripping the island bare back in 1996. Using two small boats, the last to ply the waters of the shrunken southern Aral Sea, they have carried away everything from floorboards to wiring. These days they are working on galvanized-steel piping, sealed and towed at a snail's speed to the mainland shore. ''We took out 120 tons last year,'' one said. ''We should get the last 80 tons this year.''

The trip to the island starts at the scavengers' base in Aralsk, a former Aral Sea port in Kazakhstan. It takes a day's travel over a bumpy road to reach the coast, where we board the scavengers' boat for a 15-mile ride to their camp on the other side. The next day, we travel another 20 miles in their four-wheel-drive truck, the only working vehicle on the island, to reach the abandoned testing site.

Relics of life in what was the island's only town, Kantubek, lie in the dust. They tell parallel stories of a comfortable year-round home for some 1,500 people and a practice field for the most hideous kind of warfare.

Rusting street signs primly warn against parking, or herald a clinic or a pedestrian crossing. A red fire truck, a kindergarten, a soccer field and a klub speak of ordinary life. Inside the buildings, the military side of Kantubek begins to emerge. In the rubble lies a portrait of Capt. Aleksandr Oleksenko: a meritorious soldier, the Soviet equivalent of an employee of the month. A yellowing copy of a January 1989 issue of Sovietskaya Rossiya includes the headline ''The Soviet People Are With the Party.''

In a garage with room for more than 30 trucks, two T-52 tanks and two armored personnel-carriers stand side by side. They were used to test their resistance to germ warfare.

Near the entrance of the laboratory complex, two miles from Kantubek, stand a pair of two-story buildings where animals were kept and monitored for good health before their exposure to bioweapons. Piled up in corners are hundreds of cages designed to hold guinea pigs, hamsters and rabbits. Horses and donkeys were kept in separate stables.

A germproof full-body suit, complete with a glass face mask and an airhose attachment in the back, lies in a corner. An odd smell -- ether, chlorine and something indefinable -- lingers in the air. Poking out of the rubble are dusty issues of The British Medical Journal and The Journal of Infectious Diseases.

It is all surprisingly low tech: nails are everywhere, but no screws. There are books by Marx and Lenin and yellowed, handwritten ledgers that would not seem out of place in a museum devoted to a 19th-century Russian writer.

Weeks later, I go to see Gennadi Lepyoshkin in his small apartment in Stepnogorsk, 800 miles away in northern Kazakhstan. Lepyoshkin, 55, a heavyset physician, microbiologist and retired colonel in the Soviet Army, once ran a huge bioweapons production plant in Stepnogorsk. He spent 18 summers supervising teams of researchers on Vozrozhdeniye Island. Testing was performed only in the summer, when 120-degree temperatures made the spread of pathogens less likely.

''I first went in 1970, and it was a beautiful place,'' he recounts over a glass of Kazakh Cognac. ''The water was clear; it came right up to the town, and we used to swim and sunbathe after work.'' Life's pleasures included dancing at the club and duck hunting in the northern part of the island, the opposite end from where the testing took place.

''The atmosphere was friendly, people were earning good money and we were provided with everything, except there weren't a lot of vegetables,'' he continues.

Lepyoshkin lived and worked at the laboratory complex. ''About one-third of our work was on weapons, like anthrax, plague and other bacteria,'' he recalls, ''and two-thirds on matters like testing vaccines or clothing or how long micro-organisms would survive in the soil.''

I show Lepyoshkin photographs of the lab complex. He flips through them gloomily. He stops at a shot of a long three-story building.

''That's where I worked,'' he says. ''That's the hot zone, where we kept the pathogens, where the animals were brought after the tests and where they died and were autopsied. It was cleared out so no one could even guess what went on there.''

Was it dangerous?

''Yes, there was always danger, but we never had an accident.''

He recalls the case of a woman who dropped a petri dish full of anthrax on the floor, then tried to cover up her mistake. The accident was discovered, but the woman's only punishment was a reprimand and a smaller paycheck. Lepyoshkin shrugs. ''No one got sick,'' he says.

Does he ever have any qualms about being part of a program that was making enough germs to kill the earth's population several times over?

''No,'' he says and shakes his head vigorously. ''Absolutely not. Because I knew the weapons would never be used. When nuclear weapons were made, no one thought they would be used. You'd have to be mad to use them.''

He pauses and adds: ''But now that there's terrorism, it's more scary. You know biological weapons are cheap. We calculated that to achieve an effect on one square kilometer'' -- and by ''effect,'' he explains, he means killing about half of the population -- ''it costs $2,000 with conventional weapons, $800 with a nuclear weapon, $600 with a chemical weapon and $1 with a bioweapon. One dollar.

''But we never discussed these things among us,'' he continues. ''We were doing interesting work, and we were proud of it. We discovered new methods to improve the immune system. We developed an anthrax vaccine that was given to the whole army, and it's considered to be the best in the world. Same with our plague vaccine; it's been used more than 40 years.''

When they arrived for last year's toil, in July, the scavengers discovered that an official U.S.-Uzbek expedition had come earlier in the year and burned down a row of eight warehouses.

But much of the contents of the warehouses survived the blaze, including a vast array of test tubes, bottles and petri dishes, some still in their original wrapping. The fire left some half-melted, looking like figures in a Dali painting, but most are intact underneath a coat of dust.

The Americans' primary mission was to destroy tons of anthrax spores hastily buried here in 1988 by the Soviets to evade possible detection at the site in Sverdlovsk where they were originally stocked, a violation of a 1972 treaty banning all biological arms. A U.S. Defense Department official said that the spores had indeed been destroyed.

Looking at pictures of the destruction, Lepyoshkin wonders aloud: ''Why did they have to burn the warehouses? The anthrax wasn't dangerous. You'd have to dig it out and rub an open wound on the exact piece of earth where there were spores to get infected. It was buried near the lab, and no one ever got sick from it.''

He says he was not on the island when the anthrax was buried, and no one was told about it at the time.

Ten miles from the lab, on a plateau, is the testing range itself. The scrubby trees that dominate the region have leaned into the road, evidence that it hasn't been driven for years. The scavenger driving the truck crashes through them with abandon.

The first site consists of a telephone pole and a row of three-foot-high concrete posts at 300-foot intervals, oriented in the direction of the prevailing winds, which are fairly constant.

''We used monkeys, about 200 to 300 each year,'' Lepyoshkin later recounts. ''Our staff would take them out to the range'' -- 15 miles from town -- ''and they would put them in cages next to devices that measured the concentration of germs in the air. Then after they were exposed, they were taken to the labs, where we would test their blood and monitor the development of a disease in them. They would die within weeks, and we would perform autopsies.''

Further on, four poles have been set horizontally on pickets two feet from the ground. Rusty chains hang down, even a few feed troughs. This is where the horses and donkeys were tied up. You can imagine them standing patiently in a row at dusk, when the wind would ease and deadly aerosols would be released.

At the highest point on the island, a 40-foot observation tower stands near the foundations of a gutted building. A spindly radio antenna still soars. It was a weather station. From the top of the tower, six dirt roads can be seen stretching in various directions to other test sites.

It is all very spare and quiet. The scavengers are silent, too.

Yet it is on this piece of arid scrubland that the legacy of the Soviet germ-warfare program is most menacing.

Before each test, poison was sprayed over the area to kill all insects and animals and make sure they didn't catch whatever disease was being tested. But since many burrow against the fierce summer heat, some probably survived.

Lepyoshkin and others who worked on the Soviet bioweapons program say it is most likely that some of these surviving local rodents were exposed to the weapons-grade bubonic plague bacteria and survived that too. Fleas would transmit the plague from generation to generation. The disease is resistant to antibiotics and more contagious than the natural kind, which affects a handful of people each year in Central Asia.

The scavengers, who have little knowledge of and less interest in what went on in Kantubek (''I don't see any microbes,'' one scoffed) are risking their lives, Lepyoshkin says. They are risking the lives of others too: if a scavenger contracts the plague and makes it to a hospital, he could start an epidemic.

And more visitors are coming. In 2001, the Kazakh government announced with great fanfare that the Aral Sea region contains major oil deposits. Lepyoshkin says that two shallow wells have been drilled by the Uzbeks on Vozrozhdeniye. So far, no one has fallen ill.