tment of Agriculture's chief veterinarian, who added that federal investigators were searching in many states, as well as in Canada.

Locating the source of the fatal brain-wasting disease now tops the USDA's agenda, but the agency announced yesterday actions designed to stanch the erosion of public confidence in the food supply and assist beef industry officials facing financial losses from this week's discovery. Government officials said the number of cattle tested for mad cow during the next fiscal year would nearly double, to 38,000 from 20,600 -- still a fraction of the roughly 50 million US cows slaughtered annually for food.

Yesterday, the National Cattlemen's Beef Association, representing the $175 billion industry, endorsed a "test and hold" system under which any cows tested would be withheld from the food supply until the results come back.

While the cuts of beef used for human consumption do not carry the infectious agent, "this precautionary measure would give consumers additional confidence in the safety of the food supply," Terry Stokes, the group's CEO, said of the industry's new position.

The Washington cow's muscle meat had entered the human food supply during the two weeks it took to get the test results showing it was infected with the disease, and despite Wednesday's recall of meat from the slaughterhouse that processed that cow, none has been recovered.

Federal officials said that the holiday had prevented recall notices from getting to stores and that the muscle meat would not have been infected anyway. The parts that were -- the brain and spinal cord -- were removed either for testing or for sale to a plant that makes pet food and feed for chickens and pigs.

Critics asserted this week that the government tests too few animals and uses antiquated, imprecise tests. Under the current system, all cows that are too sick to walk are supposed to be tested because they are susceptible to disease and the likeliest to have mad cow. Critics, however, say many such cows escape testing at the slaughterhouse. They contend that in Japan, all slaughtered cows are tested, as are all diseased cows and most older ones in Western Europe.

US officials respond that their program is not intended to detect every cow with the disease. Instead, it is a surveillance system meant to alert officials only if the infection rate exceeds 1 cow out of every 1 million, The New York Times reported yesterday.

USDA officials said yesterday that they are considering a number of measures to tighten the monitoring system but have not committed to any specifics.

Also yesterday, federal officials quarantined a second herd of cattle in Washington state, after learning it was the home of an offspring from the infected cow. Cows are not believed to be contagious, but the quarantine was a precaution, said DeHaven. Another one of its offspring was still at the Mabton farm where the infected cow lived before being slaughtered Dec. 9. The 4,000 cows on that farm were quarantined earlier this week on the off chance that those cows ate contaminated feed.

"The reason for concern with these calves is that even though it is an unlikely means of spreading the disease, there is the potential that the infected cow could pass the disease onto its calves," said DeHaven. With the key beef markets of Japan and South Korea now banning the importation of US beef, the Bush administration said yesterday it would send a trade delegation to Asia next week to argue against the ban. Japan alone bought $1.03 billion in US beef last year.

Asked for Bush's reaction to the mad cow discovery, White House spokesman Scott McClellan said the president would continue to eat beef. The administration, he said, will be "looking at whether or not there are any additional safeguards that there need to be in addition to the strong safeguards we have in place."

Scientists think mad cow generally takes four to six years to incubate, so USDA officials contend the dairy cow at the heart of the investigation may have been infected by eating contaminated feed shortly after its birth 4 1/2 years ago, a time period during which the documentation on the birth, trade, and slaughter of cows may be scarce.

Other cows in that herd could have been infected, too. Federal officials want to test, quarantine, and possibly destroy all remaining cows from that herd, but finding those cows will be a challenge. "We are only in Day Three of what is, no doubt, going to be a complex and complicated investigation," DeHaven said.

Mad cow typically spreads when cows eat feed made from mashed brain and spinal tissue from other infected cows. All brain-based feed for cattle was banned in 1997. But compliance initially was incomplete, according to a federal audit, and only recently have most processing facilities fallen in line.

Federal investigators know from records that the sick cow was purchased in October 2001 by the Mabton farm. But the trail stops there: It is not clear whether the cow came from a dairy finishing farm, where calves are reared to produce milk, or a livestock market, where they are readied to become meat.

It took Canadian officials several months using records and DNA testing to get to the root of their mad cow case this past spring. "Potentially, many states could be involved in the investigation," DeHaven said.

A few nations, including New Zealand and Canada, implant most calves with tiny microchips and insist on tight record keeping, which enables them to quickly retrace the movement of cows. USDA officials are considering such a system, but DeHaven warned that "this is a huge undertaking, . . . not something that's put in place overnight."

Even if federal officials track down the birth herd, they may not be able to establish which batch of feed was contaminated.

"Think yourself about what you might have eaten four or five years ago," said Stephen Sundlof of the Food and Drug Administration.

Raja Mishra can be reached at rmishra@globe.com.

--------------------------------------------------------------------------------