Only a few ounces of phosgene, one of the deadliest chemicals made, splashed Dan Leitten when his elbow bumped a valve handle at the PPG Industries plant. He held his breath, but the tiny amount he inhaled was still enough to kill him.

Leitten may well have saved other lives inside and outside the plant by quickly turning the handle to stop the flow of phosgene. The cold liquid - which was under 100 pounds of pressure - could have formed a huge gas cloud in seconds and "cleared the city of La Porte," one worker said.

No one nearby was wearing a respirator, he said. "We couldn't have shut the valve off. No one could have gotten there."

Larger releases of the chemical that terrified soldiers as a gas in World War I have been reported at PPG at 1901 Avenue H in La Porte, and three other Houston-area chemical plants on at least seven occasions in the last two years.

In one accident on Nov. 12, 1984, an estimated 5,000 pounds of phosgene gushed from a pipe when a bolt broke at a plant then owned by The Upjohn Co., about 1 1/2 miles northeast of Deer Park on Battleground Road. No one was hurt, but industry and government officials say the results could have been catastrophic.

As that gas cloud floated across the plant boundary, the phosgene concentration was more than five times the level the Environmental Protection Agency says can be "rapidly fatal after even short exposure," state officials calculated.

But a northeast wind blew the phosgene through water vapor rising from a cooling tower - which apparently neutralized some of it - and then over an open field in the direction of Texas 225 and Deer Park.

"They were lucky," said Bill Smalling of the Texas Air Control Board. "This release was enough to kill somebody off the plant property."

A quick 5,000-pound phosgene leak can kill or cause serious injury more than 12 miles away under "worst-case" weather conditions, according to an EPA chart for estimating the dangerous range of toxic chemicals in the air. Twelve miles is the approximate distance a 600-pound release of phosgene may be dangerous, but the chart does not allow projections for larger quantities. A leak as small as a half-pound can be dangerous up to 200 feet away.

To help local officials prepare for emergencies, EPA recently issued a list of 402 chemicals that can kill or injure if a large enough amount is accidentally released to the air. Some may no longer be made in the United States or imported. But every hazardous chemical in this country is probably present in the Houston area at one time or another, according to District Chief Max McRae, who heads the Houston Fire Department's hazardous materials response team. Phosgene typifies the most dangerous substances on that list and illustrates industry's record in handling them.

It is no secret why a company official calls phosgene "a classical poison" and a toxicologist says its assault on the body is "insidious." As troops in the trenches learned during World War I, a person can inhale a fatal amount without knowing it. And the damage is usually delayed - with the onset up to 24 hours later - so hospital observation is strongly advised.

The effects can be devastating. Phosgene - a liquid below 47 degrees and a colorless gas above that temperature - breaks down into carbon monoxide and hydrochloric acid when it meets water. In the lungs, that can produce excessive fluid in a condition called pulmonary edema, pneumonia and sometimes abscesses. In fatal cases, the victim essentially drowns.

One local plant worker, who asked not to be identified, went to an emergency room after breathing phosgene, but the doctor told him only that he probably had not received a harmful dose.

"For the next four to six hours, I didn't know whether I was going to live or die," said the worker. The next day, co-workers asked how he felt. "I don't know," he said. "I may be dead 12 hours from now."

Dan Leitten was not as lucky while attaching a section of pipe to a phosgene line at the PPG plant on June 28, 1984. The 25-year-old maintenance foreman stepped up a ladder and knocked a "quick-opening" valve handle, according to a report by an investigator for the Occupational Safety and Health Administration. Goggles were his only safety equipment.

Moments before, co-worker James Walden recalled in a court deposition, he had warned Leitten: "Be careful. That line is hot. I'll be putting plastic flowers on your grave."

Judy Haechten, Leitten's widow, was working as an engineer at the PPG plant at the time. She saw him 30 minutes after the accident, and he told her some phosgene had gotten on his face. But he did not think any got into his lungs, because he exhaled immediately and held his breath on the way to a safety shower.

Leitten received oxygen but decided not to go to the hospital, because another worker had been hospitalized that day and he did not want to add a second blemish to the plant's safety record for a mere "discomfort."

At home that night, however, he began having trouble breathing, so Haechten drove him to Pasadena Bayshore Hospital. He was given oxygen and drugs to rebuild the tissues the chemical was destroying, but he died at 7:45 the next morning.

OSHA levied a $1,800 fine against Payne & Keller, the maintenance contractor at the PPG plant for whom Leitten worked, because a phosgene line with a quick-opening valve had not been blocked and tagged with a warning, and he had not been made to wear a respirator.

Haechten and Leitten's parents are suing PPG, claiming that company was negligent for basically the same reasons, said attorney Ken McConnico.

A PPG lawyer declined to comment on the charges, but E.G. (Henry) Ramirez, manager of PPG's La Porte plant, said safety procedures were reviewed after Leitten's death. Some were changed, and the company began an aggressive program to educate workers about phosgene's dangers. "Our objective was to learn from this unfortunate accident."

Although company officials say the potential is very small for a phosgene leak to get beyond plant boundaries, the 1984 incident at the Upjohn plant shows it can happen.

And it happened again on Sept. 27, 1985, at PPG, when part of a phosgene release drifted beyond the plant's boundary and over Fairmont Parkway and Sixteenth Street, just outside the plant. The company told environmental officials a power failure disabled a chemical incinerator, and 8 to 10 pounds of phosgene escaped - an amount the EPA formula says can be dangerous up to 1,600 feet away.

Ramirez confirmed an account by workers that one employee smelled the chemical at the intersection while driving to work. The man went to a hospital for observation, as did another employee who was exposed on the plant property, but neither was injured, he said.

And there were earlier occasions when PPG employees feared a release might drift into residential areas, said Rindy Gremillion, an office worker at the plant from 1978 to 1982.

"People at the plant would call and say, `Are you sleeping with your windows open?' You'd say no, and they'd say good," she said, adding that she believes safety measures have improved at the plant since then.

Some workers are not reassured by federal regulations limiting the amount of phosgene they may breathe. "You know you're breathing this stuff," said a man who has worked around phosgene. "But everyone gets an attitude you're selling your health to sustain your home life. You know you're working with this stuff and it's killing you."

Later this year, state officials will require all plants with air pollution permits to start regular programs to discover and repair sources of tiny leaks of chemicals, including phosgene. Some have already done so, but the practice is not yet uniform at plants that handle phosgene, industry officials say.

Using a portable detection device, Dow has started searching for very small leaks at one of its two Freeport plants that make phosgene and at the La Porte plant formerly owned by Upjohn. PPG has that program in place at a unit that uses phosgene, but not yet at the one that makes it. The SDS Biotech plant, south of Interstate 10 near Greens Bayou, has not started the program. The Mobay Chemical Corp. plant at Baytown has been doing some of this kind of leak monitoring and plans to expand the effort.

Those programs illustrate the differences at the plants in other areas of safety and environmental monitoring.

Officials of the four companies use a variety of safeguards that they said provide extensive protection for workers and people around their plants. They include emergency generators, systems to detect phosgene releases, computers to predict whether a release will leave a plant, vacuum systems to collect hazardous chemicals, and equipment to neutralize or incinerate them.

Martin Heifele, Mobay plant manager, said safeguards vary from plant to plant partly because of differences in the design of units that produce or use phosgene. Another factor, he said, is industry groups are more likely to develop voluntary safety standards for chemicals in wider use, such as chlorine.

Phosgene is only produced to make other chemicals. Some plants that produce phosgene also store it. Some consume it in a chemical process as it is made. Since the chemical disaster in Bhopal , India, the chemical industry has given increased attention to reducing stockpiles of highly toxic substances, such as phosgene.

At Dow's Freeport plants, phosgene is consumed as it is produced, in the same unit at the same rate, said B.A. Allen, environmental manager for the company's Texas operations.

Heifele said Mobay makes and uses phosgene internally. It is basically used immediately in the same system or soon afterward. The only amounts stored are for delivery to other plant units.

SDS Biotech produces phosgene only as an unintended byproduct of a reaction process, Schwendeman said. It is normally consumed in a chemical reaction or neutralized completely in a safety system.

At the former Upjohn plant, which Dow began operating last August, and at the PPG plant, phosgene is produced for use on site and stored. Both plants have outside customers for the chemical.

The maximum amount stored at the Dow plant has been reduced to 700 gallons - or about 3 .5 tons - in one tank, Allen said.

Ramirez said the PPG plant stores phosgene in one-ton tanks and has reduced the maximum on hand to 45 tons. About half is used at the plant and the rest is sold to the adjacent Matheson Gas Products plant for repackaging in smaller containers, he said.

Besides reductions in the amounts stored, company officials described other improvements they are planning or have recently made. Among examples:

Dow will install an automated emergency response system and a "water curtain" to blanket chemical leaks at the former Upjohn plant to match systems already in place in Freeport. SDS Biotech plans to add an incinerator as a backup to its neutralization system. PPG recently bought a computer to predict the hazard of a release. Mobay is working with a British company to develop a laser to continuously scan a plant for phosgene leaks.

Public records reveal several incidents in recent years, however, in which some plants handling phosgene took certain safety actions only after an accidental release or citation from government officials:

In 1979, OSHA inspected the PPG plant after a phosgene release exposed several workers. It cited the company for the "serious violation" of not providing emergency alarms and appropriate respirators.

SDS Biotech installed an emergency generator last year after a power outage unleashed a cloud of toxic chemicals - not including phosgene - that led to the evacuation of 500 residents.

The 5,000-pound release occurred at the Upjohn plant in 1984 when a worker tried to stop a small leak and one bolt on a two-bolt pipe-fitting broke. Later, the company told the Texas Air Control Board it was installing four-bolt fittings to prevent a recurrence.

Some company officials concede privately the 5,000-pound release demonstrates how easily hazardous chemicals can escape, despite safeguards. Other instances detailed in government documents illustrate the same point:

According to OSHA files, four workers at the Upjohn plant went to the hospital in 1980 after a pipe that had been eaten away by phosgene failed and the wind shifted unexpectedly. At the same plant in 1982, 13 workers went to the hospital after a worker opened a valve on a tank that had not been double-checked to see if it contained phosgene. And a worker at the Upjohn plant died in 1981 in an accident similar to the one that killed Leitten - he opened the wrong valve on a phosgene line that OSHA later said should have been blocked.

Several industry officials said the Bhopal catastrophe made them redouble efforts to assure that their plants handle phosgene and other highly toxic chemicals as safely as possible. "Like everybody else in the chemical industry, we've tried to learn from that incident," Ramirez said.

Judy Haechten, Leitten's widow, said people in the Houston area should constantly remind themselves they live near industries with an enormous potential for danger, and not let precautions lag. The row of plants on Texas 225 is a perfect example, she said.

"I can imagine driving down 225 and if one unit happens to go, the entire road is gone. That means people on the roadway would be affected, as well as people in the plant. You can't draw a line around the plant and say everything outside is safe."

Next: Scientists are still trying to figure out how tiny amounts of toxic chemicals affect you over long periods.