In an experiment that could lead to mass production of strong, lightweight silk, scientists at a Canadian biotechnology company and a United States Army research center have spliced spider genes into cells from cows and hamsters and induced the cells to churn out silk. The silk, grown in tissue cultures, has been spun into threads that are comparable to those produced by spiders.

The scientists are seeking to produce dragline silk, what spiders use for the radiating spokes in their webs, a fiber that, pound for pound, is stronger than steel but also light, elastic and easily recyclable.

"No one has been able to make fibers like this before," said Dr. Jeffrey D. Turner, president and chief executive of the company, Nexia Biotechnologies. "That's a huge advance in this field."

Nexia, of Vaudreuil-Dorian, Quebec, is planning to expand its silk production to a commercial scale with goats that have been genetically altered to secrete silk proteins in their milk. The company has bred the goats, but they have not yet begun producing milk.

The researchers, from Nexia and the United States Army Soldier and Biological Chemical Command in Natick, Mass., describe their work in today's issue of the journal Science. They said they had inserted genes derived from two species of spiders into kidney cells from baby hamsters and udder cells from cows. The genes caused the cells to produce silk proteins, which were collected and squeezed out a syringe into fibers.

The military is interested in silk for medical sutures and bulletproof vests.

Dr. Cheryl Y. Hayashi, a professor of biology at the University of California at Riverside who studies silk genes, called the research "a big breakthrough." After decades of work, "finally, we have a way to turn these recombinant silk proteins into something that resembles a spider silk fiber," Dr. Hayashi said.

Scientists have long sought an artificial way to produce silk, because spider farms are not practical; as predators, spiders do not like to live near one other. Silk worms, which produce a fiber that is weaker than spider silk, can be cultivated.

Researchers' efforts to produce silk until now have had limited success. Silk protein molecules are long, and it is difficult to deduce the DNA that produces them. It is also difficult to insert long pieces of spider DNA into another organism's genetic code. In earlier experiments with bacteria and yeast, scientists were able to produce only a truncated, weaker silk protein.

The silk proteins produced by the mammal cells are longer but still considerably shorter than proteins found in natural spider silk. In addition, the cells produce only one of the two types of proteins commonly found in spider silk.

Scientists do not fully understand how spiders spin their syrupy solution of silk proteins into threads. The Nexia and Army scientists found that giving the fibers a strong tug after they had been squirted out vastly improved them, probably by lining up the proteins and allowing them to interlock more tightly, like Velcro.

The synthetic silk fibers are as tough as natural fibers — a property useful for stopping projectiles like bullets — but only about a third as strong, meaning they can support one-third the weight before breaking.

"There's still something wrong, several things wrong," said Dr. David P. Knight, a researcher at the University of Oxford in England.

The second silk protein was needed for strength, Dr. Knight said, and a closer copy of the spider's spinning apparatus would produce better results. Nonetheless, he said he was "very excited" by the advances.

Dr. Costas N. Karatzas, Nexia's vice president for research and development and senior author of the Science paper, said the researchers would now try to produce the second protein and try varying mixtures of the two proteins to improve the silk.

The company's tissue cultures produce only small amounts of silk — the experiments reported in Science generated only half an ounce.

For mass production, Nexia will turn to a herd of goats. In 2000, it announced the birth of two genetically engineered male goats, Peter and Webster. The researchers added to the goats' genome a silk gene designed to switch on in the mammary glands of lactating females. They then bred the two with females in the herd in hope that their daughters would make silk in their milk.

The researchers verified last year that the genes were present and working in Webster's and Peter's descendants by injecting hormones into some of the young female offspring to induce them to lactate. The goats are now pregnant and are expected to start producing milk naturally after they give birth in a few months.