New witnesses have disputed the U.S. Department of Agriculture's official account that the only American animal found to have mad cow disease could not stand up or walk when it was slaughtered, challenging anew the underpinning of the agency's approach to detecting the disease.

The new information brings to five the number of workers in Washington state who say the infected animal was not a "downer," according to an investigation by the agency's inspector general.

If the animal could and did walk, it would not have been a high priority for testing -- making its discovery more a matter of luck than effective surveillance.

A second report by the inspector general yesterday found major flaws in the detection program for mad cow disease. The draft report, released by Rep. Henry A. Waxman (D-Calif.) in advance of a congressional hearing today, cited deficiencies in how the agency collects samples for testing and how it calculates infection rates, and questioned why the agency does not test some slaughtered animals that are most likely to be infected.

Inspector general Phyllis K. Fong and USDA Secretary Ann M. Veneman are scheduled to testify today before the House Government Reform Committee.

Top USDA officials yesterday strongly defended the determination that the Washington state animal was a downer and the quality of the agency's expanded surveillance program.

Ronald DeHaven, administrator of the agency's Animal and Plant Health Inspection Service, said in a teleconference that the Washington state cow was the kind of questionable animal that should be tested, regardless of whether it could sometimes stand and walk. And he said that while the disease surveillance program was not working adequately in the spring -- when the inspector general's office evaluated it -- it has been greatly expanded and improved since then.

"We've recognized the shortcomings of the previous plan," DeHaven said. "Our new plan reflects changes in the program not in place previously."

The issue of how the Bush administration has handled the mad cow issue has become increasingly politicized. Democrats have accused the USDA of sloppy procedures and making inaccurate statements about the December discovery of an infected animal, while Republicans have said the administration limited the damage and protected public health. Mad cow disease, which is caused by deformed proteins called prions, destroys the brain and is incurable.

Waxman's release of the draft inspector general's report -- which did not include the usual responses from the agency being inspected -- was itself unusual.

The USDA's assertion that the animal was a downer helped reassure the public when the animal tested positive for mad cow disease, indicating that the government surveillance system was able to detect animals with the disease.

But the new information, released in a letter from Waxman, paints a different picture. Two workers at the slaughterhouse where the animal was killed said early this year that they did not think the infected animal was a downer, and now three men at the dairy farm where the animal lived just before it was taken to slaughter have said the same.

In addition, the USDA veterinarian who labeled the animal a downer told the House Government Reform Committee there was a "distinct possibility" that the cow stood up again soon after he examined it. According to Waxman's account, the inspector general also found that three USDA officials in the western region were aware that the slaughterhouse that killed the infected cow routinely tested ambulatory cows and had a general policy of not accepting downers.

"Your claim that the infected cow was a downer reassured the public that USDA's testing program was working," Waxman wrote. "But it now appears that these assurances lacked foundation. Even a cursory investigation would have found that the infected cow stood and walked on the day of slaughter."

Among the other conclusions reached by the inspector general were:

* Cattle rejected by slaughterhouses because they appeared to have disorders of the central nervous system -- believed to be at highest risk of carrying the infection -- were not always tested. The inspectors said that 162 of 680 cattle rejected between 2002 and 2004 for those disorders had been tested.

* The national sampling program for mad cow disease is not random because it is voluntary. The federal government could require brain samples to test for the disease but has declined to do so except at certain slaughterhouses.

* The USDA surveillance program assumes mad cow disease is confined to high-risk cattle -- those that cannot walk, have disorders of the central nervous system or have died of unknown causes -- but studies have shown that healthy-looking animals can also be infected.

DeHaven said the USDA's new surveillance program, begun June 1, has already tested more than 11,000 animals. He said that samples are coming from a broad range of sites -- farms, slaughterhouses and rendering facilities -- and that many are from animals that died of unknown causes.

"We have had outstanding cooperation with the industries we're working with," he said, adding that the agency is on target to collect 260,000 samples within 18 months. The goal of the expanded surveillance program, he said, is to determine whether there is mad cow disease in the American herd. The infected animal found in Washington was born and raised in Canada.

"Nothing in the report suggests any compromise of the public health," DeHaven said.

The USDA assertion that the sick cow was a "downer" helped reassure the public.