The caption with a photo accompanying an article last Wednesday on Underwriters Laboratories incorrectly identified the subject and task depicted. The photo showed UL engineer Bob Garcia testing the wiring of a television monitor. (Published 12/01/1999)

It's on the alarm clock that rousts you out of bed in the morning, the reading lamp you turn off at night. It can be spotted on your coffee maker and toaster, your refrigerator, stove and gas grill--and your TV, CD player, telephone and computer monitor.

It's the UL mark, a small circle surrounding the letters UL. It certifies that the appliance, no matter what size or purpose, has been approved by the world's largest independent testing service, Underwriters Laboratories. Stamped on nearly 15 billion products a year, it is, in the words of a top UL official, "the American mark of safety."

But over the past several years, a number of UL-approved products have not been safe. Space heaters, halogen lamps, baby monitors and toasters have caused fires. A popular fire sprinkler system failed 30 percent of the time. And some smoke detectors and carbon monoxide alarms designed to pass UL laboratory tests didn't work in real-world situations.

There's no question that UL-approved products, such as Christmas lights, are safer than ones that don't have the mark, and even UL's sharpest critics say the not-for-profit company provides a valuable service. But many federal safety officials, local firefighting officials, building-code administrators and international fire investigators say they are increasingly troubled that the gold standard of safety signified by the UL mark is falling short in an era when Americans increasingly put their faith in devices purported to save time and lives.

"We're having more problems than we had before," said David Smith, president of Associated Fire Consultants, an Arizona fire-investigation firm. "A lot of products seem to be hitting the market that are not fire safe but have been through UL."

Debra Rade, UL's chief legal officer and senior vice president of administrative operations, said that among the 17,000 different products tested by UL, "there are very few [approved] that present a substantial hazard." Problems that do occur, she said, are caused by new technology--or old technology put to a new use. Through those problems, what "we've learned is that the system works," Rade said. "As soon as problems are uncovered, the wheels are set into motion to analyze the issue and respond."

She added: "Our 4,000 employees live, eat and breathe safety for breakfast, lunch and dinner. If we see there's a problem, we investigate it thoroughly and responsibly."

High-profile recalls, contentious lawsuits and private industry spats have put UL under increased scrutiny. Firefighting officials, who for many years championed UL as a world leader in safety testing and standards, have begun to openly express doubts in the wake of the recall of more than 8 million sprinklers. In fact, some fire marshals are considering challenging UL's tax-exempt status, granted by Congress 45 years ago to organizations that test "for the public safety."

Interviews with more than 50 fire experts, safety officials, building-code authorities, engineers and lawyers around the country and a review of thousands of pages of documents obtained from court suits and the U.S. Consumer Product Safety Commission under the Freedom of Information Act highlight a number of concerns about UL:

* UL's safety tests may not reflect what happens in the real world. More than 350 smoke detectors have failed to sound an alarm in residential fires; about one-third of those same detectors were sent back to the manufacturer for retesting and were found to have passed a UL smoke test. And in testing pop-up toasters, UL has never included any kind of food in its fire tests--even though there have been an increasing number of reports of bagels, pop-up pastries and thick slices of bread getting stuck in toasters and starting serious fires. Between 1993 and 1996, the latest figures available, there have been at least 30 incidents of fires caused by toasters that failed to shut off.

* UL doesn't always consider factors that could affect the long-term integrity of the product, and it rarely tests products once they leave the factory. UL, for instance, didn't consider how the key components of the Omega fire sprinkler would react over time with some of the additives and chemicals commonly found in the sprinkler's water supply. The defects came to light only when the sprinklers failed to operate in two fires. More than 15 years after the product was first approved, it was found to have a 30 percent failure rate, forcing thousands of buildings to be retrofitted with new sprinkler systems.

* When a problem develops, there is evidence that UL is slow to react--and when it does, UL has first faulted either consumers for not using the product properly or electricians, plumbers and other workers for not installing it correctly. To correct problems, UL may require new warnings or consumer instructions before requiring modifications to the product itself. And when fixes have been proposed, some were not thoroughly tested or adequate.

For example, after a number of fires were started by halogen lamps, UL first told consumers to reduce the wattage of the halogen bulb from 500 to 300 watts and then, a few months later, directed manufacturers to place "Hot Surface!" warning labels on the lamps. But 300-watt halogen bulbs were also found by the Consumer Product Safety Commission to start fires--even though they passed UL's fire tests. It took two more years before UL adopted a tougher fire test.

Similarly, UL blamed the faulty Omega sprinkler on bad installation and anomalies of local water systems for more than two years, repeatedly saying it was a "site specific" problem. UL maintained that position even after millions of sprinklers were recalled for what the CPSC labeled a design defect. UL stopped calling it a site-specific problem after the CPSC accused UL of misleading the public.

* Product-safety decisions are typically made in private, with manufacturers having greater opportunity to comment than the public or other interested parties, including competitors. And when questioned, UL often cites client confidentiality, making it hard to research how decisions were made and thus difficult to get standards or listing decisions changed, revised or reconsidered. In 1995, UL approved a special electrical connector to hook copper wiring with aluminum wiring--even though the CPSC had for years declared these connectors unsafe. The CPSC was never consulted and has repeatedly urged UL to reverse its decision, maintaining that the connectors are fire hazards. UL did give the CPSC its findings in this case, but after more testing continues to list the product as safe.

In many other countries, standards are set or approved by a government entity with industry involvement. U.S. safety standards, on the other hand, are set primarily by private industry--either in independent labs such as UL or by industry associations or organizations. The CPSC, an independent regulatory agency charged with protecting consumers from hazardous products, imposes federal regulations only when it believes industry's voluntary efforts are insufficient.

Because U.S. companies rely on self-regulation, the issue of how well UL is doing its job becomes all the more critical. UL, which faces only a handful of competitors, is the dominant standards writer for electrical and fire-safety testing.

Many experts interviewed contend that UL's recent problems can be traced to the way the company is organized and funded--with more than nine-tenths of its revenue coming from companies for testing products. UL also sets industry safety standards--which it then measures products against--but does not charge for that.

In 1998, UL's tax returns show revenue totaling $407 million; $376 million came from testing. Excluding the $26 million in investment income and government contributions, testing accounted for 98 percent of UL's revenue. The actual testing fees are relatively small--say, $7,000 for a toaster per round of testing--but UL also receives annual payments from companies that wish to display the UL mark on its products.

Because manufacturers provide almost all of UL's revenue, many fire-safety officials say UL doesn't always set the most rigorous safety standards or follow through to make sure the products remain safe once they have been sold and are in use.

Few manufacturers contacted would comment on UL, but UL officials reject any suggestion that they are too close to manufacturers. "We are a completely independent organization, dedicated to safety," Rade said.

Not all fire and safety officials find fault with UL. Patrick Coughlin, executive director of the Residential Safety Institute, a public interest group that promotes fire protection, said he thinks UL's critics are wrong. "It's easy to be critical of them based on anecdotal evidence, but I think UL is very open and responsive," he said.

But among other safety and firefighting officials, concerns have mounted to the point that the National Association of State Fire Marshals is considering challenging UL's tax-exempt status because it believes UL is not fulfilling its public-service duty.

"In the last couple of years, we've had cause to reconsider and reevaluate that maybe things can be done better," said Donald Bliss, New Hampshire state fire marshal and chairman of the fire marshals association's task force on consumer product safety, referring to the recall of millions of fire sprinklers and the growing number of lawsuits involving UL-listed smoke detectors.

"We're concerned that UL relies heavily upon revenues from manufacturers and developers of products," Bliss said. "If they have such an intimate relationship with the manufacturer and are designing safety standards at the request of the manufacturer, is that in the public interest?"

If UL sets too tough a standard, some believe, it may not have many products to test. As a result, fire-safety and other experts suggest, UL chooses the lowest common denominator for its safety standards to gain as many testing clients as possible.

"They've got to make money off these folks to stay in business," said Mark Chubb, a private fire-safety consultant and formerly executive director of the Southeastern Association of Fire Chiefs. "Don't they have to please the folks who pay them for tests? If so, are they playing to that audience instead of public safety?"

Those concerns have become an issue in recent litigation. A former employee of a smoke-detector manufacturer testified in a lawsuit concerning a faulty smoke alarm that UL's approval process has a significant loophole that allows manufacturers to fraudulently modify sample units to pass UL tests. "They [UL] allow the fox to watch the henhouse," said David A. Minnis, formerly an electronics engineer with BRK Brands, which makes the First Alert and Family Guard brands.

According to Minnis, when a product fails any UL tests, engineers simply return it to the manufacturer with a note saying it didn't pass, without any analysis of the flaw. Minnis, who was BRK's "point person" with UL, testified that when the horn of one of BRK's alarms failed to sound after a UL corrosion test, he oiled the horn contact to make it sound and pass the UL test--even though oiling the contacts was not going to be part of the assembly-line process.

"By not analyzing those failures, they [the UL engineers] were relying on somebody to tell them" what was wrong, Minnis testified.

BRK and UL officials, who dismiss Minnis as a disgruntled employee, said they have investigated his allegations and determined the alleged alterations never took place. Rade said when a company resubmits a failed product, it must describe what changes it made and a UL engineer will determine whether the fix is legitimate. She added it is not UL's job to tell companies what's wrong with products when they fail.

Rade acknowledged that sometimes UL's decision-making process may seem slow and mysterious, but that's only because UL "is an engineering organization," she said. "We pay very careful attention to detail" to make sure everything is in order before issuing any announcement, decision or revision, she added.

"All UL standards are developed to anticipate real world events and what's reasonably foreseeable," Rade said. "If we didn't anticipate for everything, if there's a misuse of a product that we never thought of, then we change our standard." And increasingly, she said, UL is trying to open up the standards-setting process to allow "anyone who has a complaint" an opportunity to comment.

UL officials are "very proud of what we've accomplished," Rade said. "The U.S. enjoys the highest level of safety in the world. That's indisputable. And one of the reasons the U.S. enjoys that is because UL set the entire foundation for product-safety certification."

There's no question that there is a "high level of expectation and high level of scrutiny on us," she said. ". . . Our critics have an interest in perfection and we share that interest."

UL's origins can be traced to 1893, when the Chicago Board of Fire Underwriters sent electrical investigator William Henry Merrill to discover the cause of fires at the Columbian Exposition. Seeing a need for a safety-testing organization, Merrill then launched UL, in a spare room at a Chicago fire station, with the backing of the insurance industry.

Initially, UL operated as a not-for-profit organization, but the Internal Revenue Service revoked that tax-exempt status in 1935, on the grounds that it was more of a commercial business than a scientific endeavor, according to Joseph O'Neil, former executive director of the American Council for Independent Laboratories.

UL appealed the IRS decision but was turned down in an opinion issued by the U.S. 7th Circuit Court of Appeals, which said UL "did not operate on the basis of science for the sake of science" but was "science for the sake of business."

In 1954, however, UL again obtained tax-exempt status after Congress amended the tax laws to include organizations "testing for the public safety" in the same nonprofit status as religious groups, educational institutions and charities. "There was very little explanation for the modification--no hearings, no legislative history," O'Neil said. The only clue is a UL letter in the hearing record requesting the exemption so it can "expand its testing facilities out of earnings."

What started in 1893 as a two-person operation, with $350 worth of equipment, has now grown into an international corporation with $512 million in assets, $407 million in annual revenue, more than 5,200 employees and 13 laboratories worldwide. It oversees more than 700 safety standards and runs more than 89,000 product investigations a year.

A walk through UL's Northbrook, Ill., headquarters and testing facilities shows how varied are the products on which the UL mark is applied: window glass, roof shingles and wallboard; bulletproof vests, safes and locks; televisions, CD players and pinball machines; vacuum cleaners, toasters and pizza cookers; hair dryers, disposals and flashlights; medical beds, garage doors and even pet-bed warmers.

"We don't test for quality, but for any foreseeable hazard--fire, shock, sharp edges, radiation," said John Drengenberg, UL's manager of consumer affairs. "We look for the worst possible condition, simulate it and test it to make sure that if a product fails, it fails safely."

So fire-resistant safes are subjected to 2,000 degrees of heat and then dropped the equivalent of three stories onto broken cement blocks to make sure the safes won't pop open and the papers inside aren't charred beyond legibility. A hand-held hair dryer is dropped three times on hardwood floors to see if it breaks to the point that consumers can come into contact with any live wires or parts and be shocked or electrocuted. Its cord is flexed 3,000 times--10 per minute--to make sure the cord isn't likely to break during the dryer's normal use.

TVs are deliberately short-circuited to see if they start fires. A refrigerator door is opened and closed 300,000 times to see if the door can still be opened from the inside so children won't be trapped at the end of its normal life.

For pop-up toasters, temperature tests are run to see if cords, wires and plastic housing get too hot. But no tests are run to see what happens when the food being toasted ignites, an increasingly current phenomenon as more consumers heat up large bread products such as bagels and pastries.

Penny Cassella learned that in March when her son Joshua plopped a toaster pastry into the pop-up toaster. After a quick trip to the bathroom, the 9-year-old returned to the kitchen, only to find the toaster, cabinet and ceiling all on fire. Everyone managed to get out of the house safely, but by the time the fire was extinguished 20 minutes later, the Cassella residence in central Florida had suffered about $150,000 in damage--all because the pastry got stuck in the toaster and jammed the heating element, keeping it from popping up and shutting off.

UL has declined to add a food test, saying foods such as bread varied so greatly that it would be scientifically impossible to create a test that could be repeated precisely in different labs around the country. It took 2 1/2 years for UL to propose another solution: an automatic shutoff switch. But that proposal, issued only last month, is not scheduled to take effect until 2002 at the earliest.

"The UL mark guarantees that the product is probably safer than if UL were not around," said Jesse Aronstein, a New York engineer who has persistently taken UL to task over the past 20 years, challenging many of its standards. But he added: "The question is how well it is doing its job, whether it is truly independent and whether it moves fast enough when problems occurs--or even moves at all. Those questions have to be answered on a case-by-case basis."

Following are three case studies--involving fire sprinklers, carbon monoxide alarms and smoke alarms--that experts believe show how UL's testing and standard-setting falls short of optimum safety.

Safety Reconsidered

Here are five products approved by Underwriters Laboratories whose standards have since been challenged:

HALOGEN TORCHIERE LAMPS

PROBLEM: Fires; 232 incidents, causing 12 deaths from 1992 to 1998.

WHY: A bulb for a halogen lamp is much hotter than an incandescent bulb. Halogen lamps were first approved to illuminate spaces under construction; no new standard was written when lamps became so popular and placed in frequent contact with drapes, blankets and other flammable fabrics.

UL'S INITIAL RESPONSE: In April 1996, UL directed consumers to switch from 500-watt bulbs to 300 watt. However, 300-watt bulbs still caused fires despite passing UL's fire test -- a double layer of cheesecloth placed over a lamp for several hours. If the cheesecloth didn't ignite, the lamp passed the UL test, on the theory that if anything would catch on fire, cheesecloth would because it's such a light material.

PROPOSAL: The Consumer Product Safety Commission called for a new test in July 1996, using many layers of cheesecloth, to take account of new, heavier fabrics that could trap heat and start fires if in contact with the lamp. UL, however, maintained the double-layer test was "a reliable indicator of the risk of fire" and warned consumers to keep all combustible materials away from all halogen lamps, even those with 300-watt bulbs.

RESOLUTION: In May 1998, UL adopted a 20-layer cheesecloth test, effective for all lamps made after June 1, 1999.

POP-UP TOASTERS

PROBLEM: Fires; at least 30 incidents from 1993 to 1996.

WHY: Food such as bagels and pop-up pastries get stuck in the heating cavity and jam the mechanical lever that pops food back to the top. When the lever is left in the down position, the heating element remains on and fire can start. UL doesn't test for food fires in pop-up toasters.

PROPOSAL: In March 1997, the CPSC urged UL to conduct tests with toast.

UL Response: Said it was impossible to conduct a food test that can be repeated reliably in different labs. In November 1998, UL said it would revise its standard to address food fires.

RESOLUTION: In October 1999, UL issued a proposal requiring heating elements to shut off automatically, no matter what position the mechanical lever is in. The proposed rule is not scheduled to be effective until 2002 at the earliest.

OMEGA FIRE PREVENTION SPRINKLER

PROBLEM: High rate of failure.

WHY: Rubber O-ring would react to some chemicals and additives found in the water used in sprinkler systems.

UL REACTION: Called on fire officials and building owners in April 1996 to test sprinklers. In May 1997, after testing 800 sprinklers and finding a 30 percent failure rate, UL called the problem "site -- specific" and proposed a fix for existing sprinklers.

COURT ACTION: CPSC sued the manufacturer in March 1998; the lawsuit was settled in October 1998 with the recall of 8.4 million Omega sprinklers. As part of the settlement, manufacturers asked UL to withdraw its approval of Omega sprinklers. UL did but still called the problem "site specific." In January 1999, the CPSC accused UL of misleading the public by continuing to say the problem was "site specific."

RESOLUTION: UL immediately dropped "site specific" from its statements about Omega.

IONIZATION SMOKE DETECTORS

PROBLEM: Often fails to sound alarm in smoldering fires.

WHY: Manufacturers say it's due to "cold smoke" -- what happens in a smoldering fire that's not hot enough to drive the smoke up to the ceiling and detector. Several fire experts say the problem is with UL's standard smoldering fire test.

PROPOSAL: For years, fire experts have called on UL to reconsider its smoldering fire test. Some say it's based on 20-year-old tests that used different furniture than what's now found in homes. They note that UL conducts four flaming fire tests for detectors, but only one for smoldering fires.

UL RESPONSE: UL says that "a fire is a fire" -- it would be too difficult to develop a smoldering fire test to consider all the variable items found in a home.

RESOLUTION: UL is now in talks with the CPSC to discuss the possibility of new fire research to see if a new smoldering test is warranted.

CARBON MONOXIDE ALARMS

PROBLEM: False alarms.

WHY: The initial standard set too low a threshold at which an alarm must sound.

UL RESPONSE: Revised the standard.

NEW PROBLEM: Failure to sound and inability of consumer to know if sensor detecting gas still works after alarm leaves factory.

WHY: UL standard does not require long-term reliability testing; the "test" button on the detector only determines if the circuitry is working but not the actual sensor response to carbon monoxide.

UL RESPONSE: No simple way for consumers to test sensors, but UL recently proposed requiring manufacturers to pull hundreds of detectors off assembly lines every three months, and test these every quarter, to see how long sensors remain effective.

SOURCE: Underwriters Laboratories, Consumer Product Safety Commission and industry documents

CAPTION: UL consumer affairs chief John Drengenberg with test mechanism for wire connectors, nuts.