DATELINE: OAK RIDGE, Tenn.

Shortly before noon, after three hours of false starts and frustrations, Jack LaForge hit the big green "on" button and the experiment that had drawn these 11 men into the crowded laboratory began in earnest.

Project leader Glenn Allgood - weapons builder, mathematician and engineer - peered intently at one of the half-dozen computer screens around him.

At that moment, he was on the verge of solving a problem that had plagued his team for almost a year.

Defective towels.

Not just defective towels; towels so defective that they have to be sold by WalMart for a discount.

This, it turns out, is the high-tech battlefield of the 1990s. It's a battlefield onto which the U.S. government is throwing billions of tax dollars in its technology effort.

The 11 men had gathered in this historic place, where a half-century ago scientists produced fuel for the atomic bomb, to do battle with the new foe: those who would steal textile jobs from U.S. workers.

In this case, the enemies are the low-wage nations of Asia that have decimated America's apparel and textile business.

And in the end, Allgood's mission will be judged a success not if he puts a man on the moon but if he can reduce the number of those defective towels that WalMart has to unload at a cut rate.

The goal of the team assembled at the Oak Ridge National Laboratory is to save this $50 billion industry - and its 1.6 million jobs - by developing revolutionary technologies to speed textile production and cut waste.

In many ways, the effort to defend the American economy has been portrayed as a patriotic mission, fundamental to the well-being of all Americans. To that end, Congress has created more than 50 programs aimed at boosting U.S. companies - at an estimated cost to taxpayers of $6 billion a year.

But there's a catch.

Although the technology programs have been cast as a way to create high- paying jobs for U.S. workers, there are no guarantees that these technologies will remain in the United States.

And once they go offshore, U.S. workers may well find their jobs at risk.

There's another irony, as well: At the same time that the U.S. government is spending millions of dollars to help modernize textile manufacturing and create U.S. jobs, it has, through trade policies such as NAFTA and GATT, been throwing open U.S. borders to cheap imports and sending textile jobs overseas.

Thus, what Washington gives with one hand, it takes away with the other.

The government estimates that GATT, the General Agreement on Tariffs and Trade, could cost American apparel workers upward of 120,000 jobs - the same number of jobs that it expects to create through its technology initiative.

In other words: a wash.

To blunt criticisms that it is playing favorites by selecting individual companies for high-tech handouts, the Department of Energy increasingly is committing its $8.2 billion research resources to rescuing whole industries. Other agencies are following suit.

Instead of awarding grants and subsidies to individual companies, the agencies are directing money to industry trade groups and nonprofit institutions, which are supposed to share the technological advances with all industry members for little or no cost.

Under terms of the government-industry partnerships, the industries themselves target areas in need of technology triage. Scientists at former weapons facilities, such as the Energy's Department's lab here, perform generic research, available to all industry partners.

Cost of the research usually is shared by industry and government. In some cases, partners pay a fee to join the consortium. In return, they get exclusive access to the technologies for a fixed number of years.

The automobile, semiconductor, battery and financial industries are among more than a dozen groups that have formed alliances with the federal government. Since 1988, these industry partnerships have benefited from more than $1 billion worth of taxpayer-financed research and subsidies.

The oldest and largest industry partnership - Sematech - has received nearly $800 million in subsidies to help U.S. computer-chip manufacturers in their battle with foreign competitors.

Among the expenses taxpayers have helped underwrite: the $599,151 pay package Sematech president William J. Spencer received in 1993.

U.S. chipmakers once again dominate the world market. But even a spokesman for Sematech plays down the organization's role.

"That has basically been American firms fighting the global battle," said spokesman Miller Bonner. "We would not argue that point at all."

With about 1.6 million workers, the industry accounts for nearly 12 percent of U.S. manufacturing jobs. About 800,000 of the jobs involve production of clothing and accessories, a field where employment has plummeted since the 1970s, mostly due to foreign competition.

Without new technologies to make these industries more efficient and competitive, warned government and industry officials, the exodus of jobs would increase and thousands more would be thrown out of work.

And so in March 1993, amid the worst blizzard to hit the Eastern seaboard in 100 years, officials gathered at North Carolina State University to
announce the formation of the American Textile Partnership - Amtex for short.

Grand things have been promised from Amtex, which represents 82 companies, a tiny fraction of the 26,000 firms that make up the textile and apparel industry.

The partnership is dominated by giants such as Fieldcrest Cannon, with $1 billion in annual sales, and Milliken & Co., a private company with estimated sales of $2.5 billion. It includes smaller firms as well.

Energy Secretary Hazel R. O'Leary, who traveled to Raleigh, N.C., to attend the signing ceremony only days after taking office, offered glowing predictions:

"We could save 350,000 jobs over the next five years, and projections are that we could create 200,000 new and better-paying jobs in the American textile and apparel industry in the fifth through 10th years. Our work with Amtex can be a model for our nation."

Milliken's president, Tom Malone, was equally upbeat: "I believe this can be a role-model example of total manufacturing industry-government- university collaboration dramatically increasing the competitiveness of the industry to preserve and create tens of thousands of new jobs in the U.S.A."

Amtex officials projected savings of $12 billion a year, for an industry that has annual sales of $220 billion.

The textile industry's record on technology investment is poor, according to a 1992 National Science Foundation study. The industry spends only four- tenths of a percent of total sales on research and development.

The fact that the industry now wants taxpayers to help make up for this bothers some, including C. Paul Robinson, vice president at Sandia National Laboratories, which has several Amtex projects.

"I do feel that people think there is a sense of entitlement," said Robinson of textile officials. "We're a big industry; we're an American industry, and we're losing jobs like crazy, so we should be supported."

Since its creation, the Amtex partnership has spent $102.6 million, 45 percent of which has been in taxpayer-provided government services - scientists' time and use of equipment. Its 1995 budget is $53.6 million.

The consortium has picked eight areas of research to pursue, including development of software to speed production and distribution, new sensors to detect manufacturing flaws and new processes to reduce fabric and processing wastes.

It was the sensor project - called Computer Aided Fabric Evaluation, or CAFE - that brought the 11 men to this makeshift lab at Oak Ridge in March of this year.

To Allgood's team had fallen the task of putting "eyes" on America's textile looms.

They are among 110 scientists from 12 national labs assigned to this project, working with engineers from the textile firms. The budget for the CAFE program this year is $10.1 million, about half from taxpayers.

The team was here to test the initial research on the car-size green-and- silver loom sitting at the end of the crowded room. At that moment, so many wires and sensors were snaking from it that the machine resembled a patient in intensive care, its every movement detailed on nearby computers.

For large textile manufacturers such as Fieldcrest Cannon, flawed products, or "seconds," represent a significant loss.

About 8 percent of the company's bedsheets and 8 percent to 10 percent of its towels are seconds. On the company's lower-end brands, these seconds must be significantly discounted. They sell at 30 cents per dollar of the company's production costs.

Fieldcrest Cannon is in search of technology that can reduce the number of seconds rolling off its looms and printing equipment.

"We want to be able to stop the machine and correct the mistake immediately," explained Chris Kametches, senior vice president for manufacturing and engineering at company headquarters in Kannapolis. "We don't have that technology now. But it is technology the government already has."

Soon, Fieldcrest and other textile mills will have it, too. Allgood and his colleagues hope to test a sensor system in a mill by year's end and have a system working in five years.

Until then, though, all that stands between a first-quality roll of fabric and one that must be sold as a second are people such as John Trexler, who has worked at Fieldcrest Cannon's Kannapolis plant for 26 years.

Trexler, a quiet, stocky man in a T-shirt and baseball cap, spends his days watching as vivid rolls of printed fabric stream past him, often at 60 yards a minute. During an average shift, he will inspect eight miles of fabric.

Although Trexler is legendary at the mill for his ability to detect imperfections no larger than a speck of dust, Fieldcrest Cannon still loses about 4 percent of its weekly output of 1 million yards to printing flaws. That's about $140,000 worth of fabric that must be discounted each week.

When members of Glenn Allgood's team wanted to learn about fabric inspection, they sought out Trexler and Jeff Mills, who supervises the printing operation.

The challenge was to take components of a sensor system that was used to identify military targets during the Persian Gulf War, and reprogram the system to detect tiny flaws in a bedsheet fabric full of random speckles and faded colors.

As Allgood has discovered, getting a sensor system to function in a combat zone is one thing; having it perform on a bedsheet in a dusty and noisy textile plant is quite another.

The purpose of this day's experiment was to isolate routine "background noise" from the start-up of the loom so the sensors could be programmed to ignore it.

The industry has identified 60 to 80 common problems that plague production lines and represent "big money costs" for manufacturers, according to Allgood. These are what the labs are attempting to eliminate.

The scientists and engineers re-created the defects in the lab and are experimenting to determine if the sensors can detect the flaws.

Next comes the complex task of writing the algorithms - or programs - that will enable a computer attached to the sensor to identify the problem and alert inspectors such as Trexler.

Trexler won't be out of a job, officials say. His expertise still will be needed.

"The brain is the best computer around - still," said Oak Ridge's Van Hoy, who spent time with Trexler on the printing line.

There are no guarantees that new technologies will not eliminate some jobs in the industry, especially in the short run.

"But that's the way technology has always been," said Amtex executive director Richard K. Quisenberry. "In the macro sense, it enhances the competitiveness of that industry, and we eventually rebuild jobs - not always the same people, unfortunately."

And the way to do that is to create a system so technologically sophisticated and responsive to the market that it eliminates the advantage now held by low-wage competitors such as China, Hong Kong and Mexico.

Put another way, the hope is to give U.S. apparel and textile firms a powerful incentive to keep their production and assembly operations on U.S. soil.

That's no small challenge.

Since 1970, the high-water mark for employment in clothing and textile manufacturing, nearly 700,000 jobs have been eliminated - one-third of all such jobs.

In 1970, more than 2.3 million Americans, many of them women and minorities, worked in garment and textile factories. By 1993, the figure had shrunk to about 1.6 million.

Most of those lost jobs were exported to low-cost foreign shops. In some cases, American companies set up operations offshore. In others, they contracted out work. The reason in both cases: wages.

American workers in this industry earn an average of $8 to $12 an hour - more than 10 times average wages in foreign plants.

"The driving force, above all others, is labor costs," Quisenberry said. ''U.S. workers earn far more than workers in the Pacific Rim countries. Depending on the country, some workers might only be paid 30 cents an hour. That's what drives it."

The impact of foreign competition on the U.S. market has been dramatic. As recently as 1980, nearly three-fourths of the clothing sold in the United States was made here. In 1993, only 40 percent was American-made.

Amtex officials hope to check imports at their present level - 60 percent of the retail market. Any future growth would then accrue to U.S. firms.

"I think what we'd like to do - and realistically can do - is put in place technology that will allow us to hold the share of where we now are in the domestic market," said Quisenberry.

The domestic market has been growing about 3 percent a year, he said. "If we can just hold our share and let the market grow, we can create 120,000 jobs for people over the next 10 years."

A key to doing that is finding ways to shorten the time it takes to produce and sell a garment. Now, the cycle from farmer's field to retail rack is about 66 weeks, Amtex managers say. Much of that is idle time in storage and inventory - a clog in the apparel pipeline. The cost of that inventory backlog: about $25 billion, according to Quisenberry.

Amtex officials hope to reduce inventory gridlock by 28 weeks and save $12 billion a year by creating a computer model of the textile chain, from start to finish, and developing computer software to link the players.

"If you are a retailer, whether you buy from a Sri Lankan or whoever, you get the best price you can," said Jerry Work, a manager at the federal Pacific Northwest Laboratory in Richland, Wash.

"But if you can shorten that 66-week pipeline by increasing the level of technology, you can lower that wage-rate gap . . . and give U.S. companies a competitive advantage."

Said Quisenberry: "You have to give U.S. manufacturers an economic incentive to source domestically, rather than offshore. They're not going to do it for patriotic reasons."

Amtex members are prohibited from selling or using new technologies in foreign operations for five years. That should help U.S. companies compete more effectively in the United States, Amtex executives said.

What happens after five years?

Critics say the technologies will be copied by competitors or exported to low-wage countries - in either case, eliminating any competitive advantage.

Amtex hopes the technologies will help U.S. firms gain a significant share of emerging world markets.

But the days of free-ranging research at the weapons labs are on the wane as their budgets are cut and major new projects are infrequent. Oak Ridge is a good example.

The lab Allgood has commandeered for his loom until recently served as the control room for a fusion research project. It was scrapped a few years ago - "a victim of the balanced-budget effort," said Jack LaForge.

Sitting around the lunch table with their colleagues at the Sagebrush Steakhouse & Saloon, a few minutes from the unmanned security gates of Oak Ridge's Y-12 facility, LaForge and Van Hoy expressed uncertainty about the lab's future - and their own.

Both men, who are employees of Martin Marietta, which operates Oak Ridge for the Department of Energy, hope for a positive side to linking up with industry: a better-managed laboratory.

It's all new to Allgood and his team. "There is a cost structure" established by the industry partners, Allgood explained. "There are (financial) constraints that have been applied to this technology, which requires creativity on the lab's part."

And there's a deadline.

"We have to be very, very aggressive in moving this technology along," Allgood said. "It couldn't be one of these programs that went on for years and years."

Chris Kametches at Fieldcrest is one of Amtex's biggest boosters. He regards Allgood as a "genius" and the research being done in the labs as his industry's salvation.

The textile industry, said Kametches, does "not have scientists. A scientist is an individual who comes up with new thoughts, new ideas. We don't have the money to fund that kind of operation. We operate on a very thin profit margin."

Increasingly, large-volume customers, such as WalMart, demand that Fieldcrest fill huge orders within days. That requires the manufacturer to produce and maintain large inventories, hoping it will have on hand the color and style of towel or sheet WalMart orders.

"Over a 30-day period, we only guess right about half the time," said Kametches. "The faster we can respond (to an order), the later we can wait to do something. And that reduces our chances of making a mistake."

For that reason, the company is participating in another Amtex initiative, funded this year at $23.4 million. Taxpayers are picking up $12 million of that total.

The project is designed to produce a vast software system that will link every segment of the apparel industry - from cotton growers to knitting mills to T-shirt manufacturers to the cash register in the sportswear department at the local J.C. Penney's in the mall.

When sales data show a strong demand for purple T-shirts, the message is relayed back down the system and turns on the production pipeline. Now, once the purple T-shirts disappear from the rack, it often takes months to replenish supplies.

This is the system Amtex officials are betting on to regain lost business - and lost jobs.

Macaluso says he's never heard of Amtex.

Nor does he believe government scientists can improve his assembly operations or help him compete with Third World companies that pay workers 10, 20 or 30 cents an hour.

In fact, Macaluso thinks the federal government has signed the death certificate for America's beleaguered apparel business.

It's called GATT and it will open the floodgates to cheap garments, he says.

"All of this other talk is just a smokescreen. The politicians are trying to make us think they're still trying to save the industry. If they are, I must be missing the point."

Beginning this year, GATT phases out quotas on clothing imports over a 10- year period; the quotas had helped to blunt somewhat the impact of foreign competition. The flip side of the quotas is that they have inflated the prices paid by American consumers.

Researchers have estimated GATT could result in the loss of 50,000 to 250,000 American jobs - mostly in the apparel industry.

Macaluso puts it bluntly: "You're sending all of this work offshore and you're talking about creating jobs? Give me a break."

A 1994 study by the U.S. International Trade Commission on the potential impact of GATT said it would "likely stimulate further investment in apparel production in low-wage countries, adding to the competitive pressures facing the U.S. apparel industry."

The study estimated that foreign imports would increase by more than 15 percent, to more than 75 percent of the U.S. market. That would add to the sector's trade deficit and result in the loss of "over 5 percent to 15 percent" of the remaining apparel jobs.

Of the estimated 800,000 production jobs, GATT could result in a loss of 40,000 (5 percent) to 120,000 (15 percent) jobs - primarily affecting small companies such as Macaluso's.

Federal officials hope the industy will have time to adjust while GATT is phased in over 10 years. But they're concerned.

"I'm very worried about my textile (and apparel) workers. This administration is certainly looking at every possible way at retraining," said Rita Hayes, deputy assistant secretary for textile, apparel and consumer goods at the U.S. Department of Commerce.

Herman Starobin, research director for the International Ladies' Garment Workers Union, is also worried. After GATT is implemented, foreign firms could control up to 90 percent of the U.S. apparel market, he said.

As for Amtex and the government's job estimates: "Wet dreams. Worse than pipe dreams," Starobin said.

It's unclear if federal officials include the job-destroying impact of GATT when they talk about jobs they hope to create.

"I've not seen that number. I don't know what that number is," Amtex's Quisenberry said. "I don't know that the industry has ever come up with a number."

O'Leary first said Amtex would create 200,000 high-wage jobs. Later, Amtex officials lowered the estimate to 120,000 - a 40 percent drop. At other times, they have used estimates of 140,000 and "more than 100,000."

Amtex officials also have been imprecise about whether these will be new jobs or jobs recaptured by the industry. They have used both explanations.

In either case, the estimate seems to assume that U.S. companies will use the gains from any technological advances to hire new workers - not to increase profits.

That may or may not be the case, says Irving F. Stowers, who heads Amtex efforts at Lawrence Livermore National Laboratory in Livermore, Calif.

"What industry wants and government wants is often two different things. Voters and government would like to see more jobs created. Industry wants to see more profit," Stowers said. "Basically, if you raise the productivity of a business, it will profit, gain or turn around and invest and get back market share."

Assuming any jobs are created, Quisenberry said, the bulk of them "will primarily be entry-level manufacturing jobs in the apparel sector. But they're better jobs than I think you can find in the service sector."

Federal data contradict that claim.

The average hourly wage of apparel workers in 1993 ranged from $6.39 to $7.78. In the services sector, it ranged from $7.28 in laundry services to $15.28 in legal services.

"I think he's wrong. The service industry has gotten a bad rap. There are a lot of different service jobs and some of them pay fairly well," said Maureen Greene, a regional economist for the Bureau of Labor Statistics.

At Macaluso's three South Jersey facilities, full-time workers earn above- average wages. He said GATT would devastate the business.

"I have 300 people at three plants. My full-time people make between $9 and $12 an hour and get full health benefits. Where are they going to go? They (government officials) always evade the questions. They blow so much smoke."

His company already uses computers at its three locations, Macaluso said. It moves 47,000 garments a week and can turn around a product in two to three days, he said.

Sure, it's possible the system could be made more efficient, he said. "The kinds of things they are talking about can work, to a degree."

But "to me, it seems kind of late. With GATT, they're already putting the nails in the coffin anyway."