New information, closely held until now by Britain and Canada, casts serious doubt on President Reagan's repeated claim that toxin weapons have been used by the Soviets and their allies to kill thousands of people in Southeast Asia.

Two large-scale efforts by Britain and Canada to take samples in Southeast Asia have demonstrated that the poisons the United States has claimed to be chemical warfare agents used by the Soviets actually occur naturally in the area as a fungus that probably infests food.

Further, the British government has revealed that hundreds of samples taken from thirty-five separate alleged attack sites showed that there was none of the supposed "yellow rain" chemical warfare agent present.

The Canadians, in work done in 1984 and finally released yesterday, carried out the only large-scale sampling of the blood of a variety of people near the war zone in Laos and Kampuchea, and found that about 2 percent of the people tested had the toxins -- called tricothecenes -- in their blood. But they were not believed to be victims of any chemical attacks. Rather, they were probably poisoned by fungus infested food.

The United States began a high-level public campaign in September 1981, when then-Secretary of State Alexander M. Haig Jr. said it had found the "smoking gun" showing that the Soviets, through the Vietnamese, were conducting a lethal chemical and toxin warfare campaign in Laos and Kampuchea.

The smoking gun was a single leaf and stem, covered with mold, that contained tricothecene toxins -- poisons produced by a fungus that the State Department maintained did not exist in Southeast Asia.

The president has hammered the Soviets more than 15 times in documents and speeches for their use of "grisly chemical and biological weapons, including yellow rain" and for the "flagrant violation" by the Soviets of the Geneva Protocol (on chemical weapons) of 1925 and the 1972 Biological and Toxins Weapons Convention.

The new information was released by the Canadian government after inquiries by The Washington Post. The British released their data after a report in the British journal Nature and a parliamentary inquiry.

"I think the matter is now settled," said Matthew Meselson, a Harvard biologist who has spent several years pressing his challenge to the U.S. claims. It was Meselson who earlier reported that virtually all the samples of the alleged chemical warfare agent contain pollen, and reported that the "yellow drops" cited in the alleged attacks were probably pollen-laden bee feces.

Meselson added, "This is the last gasp of one of the least careful jobs of intelligence work done in recent years."

A State Department spokesman, asked about the Canadian and British results, said that the U.S. position is unchanged and that the toxins are not native to the region. He said that the government still believes that the Soviets made and sent to Vietnam the toxins, and that they were used to kill thousands, many of them Hmong people in the hills of Kampuchea and Laos.

He said that the State Department believes that a variety of lethal chemical warfare agents, possibly including nerve gas, were used in Southeast Asia and the tricothecene toxins were "only one of the things used."

Dr. Joseph Rosen of Rutgers University, who in the past has supported the idea that tricothecenes have been used as weapons in the Southeast Asia, and who found one sample positive in his own laboratory, said yesterday that the new information "says at the least that maybe the tricothecenes have been much more prevalent in Southeast Asia that we thought."

"This adds to the confusion on the whole matter. There are still arguments that [toxin] weapons were used and there is good evidence for the natural occurrence . . . this is obviously on the side of natural occurrence," said Rosen.

Donald B. Louria, a specialist in toxin poisoning, said, "People have been warning them [the U.S. government] that something like this would come along . . . . This really weakens any data they had. It should tell them it is not good to go off half-cocked without any real checks."

Louria is chairman of the department of preventive medicine of the University of Medicine and Dentistry at New Jersey Medical School. He visited Thailand when the U.S. was collecting samples of the alleged chemical warfare.

In addition to the new data, a 1981 Army analysis obtained by The Washington Post stated that symptoms reported by refugees from the war area were not consistent with a single agent, and that hemorrhaging, often cited as the telling symptom of yellow rain poisoning, was not common among the refugee reports. The analysis was done by Lt. Cmdr. William W. Edwards in November 1981.

The Canadian report covers the 1984 testing of 280 blood samples collected in five areas of Thailand.

Ron Cleminson, an official of the arms control and disarmament section of Canada's Department of External Affairs, said that five people among those tested showed evidence of tricothecene poisoning in their blood.