2003National Reporting

A Staggering Abundance of Crash-Causing Glitches

What ails the Harrier? The list goes on and on. And even when everything is working properly, the plane is a bear to fly.
Kevin Sack and Alan C. Miller
Times Staff Writers
December 16, 2002,
Part 2

previous | index | next


CHERRY POINT, N.C. -- The pigeons in the hangar had worn out their welcome.

So late one night, three Marine maintenance workers launched an artillery assault on the squatters in the rafters. Armed with slingshots and ball bearings, they fired round after round, unfazed by the $28-million Harrier attack jet parked nearby.

They bagged one bird before being told to knock it off. The Marines then searched the hangar, inspected the plane and thought they had found all the bearings.

The next day, having flown the Harrier to Ohio, Capt. Stephen E. Brooks had just taken off for home when the engine shuddered with a thunderous crack. Finding himself powerless 150 feet above a busy interstate, he pointed the nose toward a cornfield and pulled the ejection handle. He walked away with scrapes, bruises and a medal for steering his plane away from a strip mall.

walking the tarmac

Marines walk the tarmac in Cherry Point, N.C. looking for small objects that could be sucked in and cause damage to the Harrier's single Rolls-Royce engine. (photo: Tyrone Turner for the Times)

An investigation into the October 1997 crash determined that a foreign object had shaken loose during takeoff and ricocheted through the whirring engine, reducing compressor blades to shrapnel. A metallurgical analysis confirmed what was widely assumed: A dent in one of the damaged blades matched the ball bearings like a fingerprint.

Harrier pilots have encountered a mind-bending array of calamities, some unfathomable in their recklessness, others dispiriting in their predictability. In the 31 years the Marines have flown the plane, just about anything that could go wrong has gone wrong.

Time and again, engines have flamed out, compressor blades have cracked, control sticks have jammed, wing flaps have frozen, nose wheels have veered off runways. A widely respected pilot and a flight surgeon lost their lives when the acrylic canopy over their cockpit imploded.

"If you can think of a situation, it's happened," said Durward Savage, a retired Harrier pilot.

Military officials knew about defects in the flaps and ejection system for years before fixing them, while planes crashed and pilots died. Problems with the plane's engine persisted as the Marines and the manufacturer, Rolls-Royce, engaged in finger-pointing over who was responsible.

The Harrier's mechanical breakdowns have been compounded by maintenance lapses that have helped make it the most accident-prone airplane in the U.S. arsenal.

The corps has been devoted for decades to creating an aviation fleet that can take off and land vertically, with the Harrier as its prototype. The plane can rise like a helicopter, hover in precarious balance and then roar off toward the horizon.

But the Marines' Icarus-like pursuit of vertical flight has strained the limits of the plane, its pilots, maintenance crews and budget, trumping the corps' renowned can-do resourcefulness.

They have cut corners to save time and money. Rather than spending millions on engine testing, for example, they have let pilots discover the plane's failings at 15,000 feet.

"We took a revolutionary airplane with an underdeveloped engine program and we put it in the fleet," said Maj. Gen. Charles F. Bolden Jr., recently retired commander of the 3rd Marine Aircraft Wing. "We let the fleet figure out what the problems were and we killed people."

In all, 45 Marines have died in 143 noncombat accidents since the corps first bought the Harrier from the British in 1971.

Marine Corps officials say they have never knowingly allowed an unsafe aircraft to fly. The current chief of Marine aviation, Lt. Gen. Michael A. Hough, said safety improvements on the Harrier have always been a top priority. But he said it often took years to design, test and implement such changes.

"And you can't down your fleet and say, 'Well, we'll take a little pause here for about three or four years, boys, and get back to you later,' " he said.

He said the Marines have never deferred safety-related fixes because they lacked funding. But he is contradicted by predecessors, other Marine officers, accident records and official reports.

When asked which problems he would have addressed with more money, retired Lt. Gen. Harold W. Blot, a former head of Marine aviation, hardly knew where to begin.

harrier in hangar

Harrier are serviced in a maintenance hangar at the Marine Corps Air Station in Yuma, Ariz. The plane's complex nature presents a special challenge to the young enlisted Marines who maintain it. In one survey, nearly half of them said they felt they were under-trained. (photo: Don Bartletti, Los Angeles Times)

"Well," he said, "there was a nose wheel steering issue. There was an inlet guide vane issue. There was the flap impingement issue."

Under the best circumstances, flying the Harrier is a formidable task. In an era of advanced avionics, with computers doing much of the flying, it still requires considerable manual dexterity (known among pilots as "monkey skills") and mental focus ("headwork").

Charles E. Myers Jr., a former director for air warfare in the Pentagon, likens the allure to riding "the nastiest horse in the rodeo."

While a conventional jet has one way to take off and one way to land, the Harrier pilot must master four basic methods for each. In addition to handling the throttle, control stick and rudder pedals, he also must maneuver the four rotating nozzles that allow the plane to ascend and descend on powerful blasts of hot air.

While flying vertically, he must pay close attention to wind direction. Add the challenges of operating at night or from a moving ship and the task becomes daunting.

Retired pilot Brooks, whose plane was damaged in the pigeon hunt, compares it to "speeding your car 90 mph through a crowded shopping mall parking lot while playing the hardest X-Box video game imaginable and talking on your cell phone."


Investigators comb the wreckage of Capt. Stephen E. Brooks' Harrier in 1997. Brooks had just taken off when the engine exploded. He pointed the plane toward a cornfield and ejected. The crash was blamed on a stray ball bearing fired by a maintenance work at pigeons in a hangar. (photo: Dayton Daily News)

Proficiency in the Harrier cockpit requires, at minimum, 15 to 20 hours in the air each month, according to the Marines. But the plane's frequent groundings -- there have been 31 in the last 12 years alone -- mean pilots often have to make do with simulators. As recently as 2000, they averaged 8.2 flight hours a month; that has since increased to 13 hours.

The downtime has contributed to numerous accidents caused by pilot error. But in some instances, pilots have been assigned posthumous blame even when inadequate training or mechanical problems were significant factors.

The aviators risked their lives every day but did not always get the support they deserved, said retired Maj. Gen. Michael D. Ryan, a former Marine wing commander who is now executive vice president of government business for Rolls-Royce North America Inc.

"These pilots were young guys, hard chargers, superheroes," he said, "who were working for leadership that sometimes didn't provide them with what they needed."

Flap Problems

From the time the Marines first discovered a dangerous flaw in the Harrier's wing flaps, nearly eight years passed, three planes crashed and two pilots died. Only then did the Marines and their Navy counterparts decide to completely redesign the problem part. It took another eight years to find the money and do the work.

In 1986, as 1st Lt. Edward C. Jasiewicz was trying to land his Harrier at the Marine Corps Air Station at Cherry Point, the flaps suddenly froze, pitching the nose forward and prompting the pilot to eject to safety.

An investigation focused on the electronic device that controls movement of the flaps, which help provide the plane with lift. The report noted that the failure rate for the device, known as a flap electronic controller, had been nearly three times higher than predicted.

Two years later, another flap failure forced 1st Lt. Kerry D. Dale to eject only 33 feet above the ground as his plane plunged into a stand of Carolina pines and exploded into bits.

"It was a light body bag," said his father, Jim P. Dale.

An investigation found that moisture had seeped into the flap controller and shorted it out.

The Marines took steps to reduce the problem, like coating the circuitry with waterproofing, while the airframe manufacturer, McDonnell Douglas Corp., explored longer-term solutions.

In the meantime, pilots were expected to fly their way out of trouble by following directions in their flight manuals, according to officials from the Naval Air Systems Command, which oversees Marine aircraft safety.

In 1989, Maj. Woody F. Gilliland managed to land his plane after the flaps seized up at more than 20,000 feet. Again, the investigation identified moisture as the culprit, according to Gilliland.

Four years later, Capt. William P. Delaney had just taken off from Cherry Point in a light rain when the flaps on his Harrier jammed. Delaney ejected just before his plane smacked the runway, broke apart and burst into flames.

The parachuting pilot descended into the fireball and was killed.

The crash investigation found that two wires to the flap controller had broken, cited moisture as a "contributing factor" and again recommended a complete redesign of the controller box.

Determining that their previous "Band-Aid" fixes had not been sufficient, the Marines and the company then decided to redesign the box, said Vince L. Higbee, Harrier program manager for Boeing Co., which bought McDonnell Douglas in 1997.

It took another two years for the Marines to find the money, Higbee said, adding that such a period was "not atypical" for military procurement.

It was only last year, 15 years after the Jasiewicz crash, that the company completed delivery of the newly designed parts to the Marine Corps.

The design changes rerouted the flow of moist air around the controller box and programmed the flaps so that they automatically override mistakes made by pilots.

The cost was $21 million. There have been no reported flap-related accidents since.

"It took Billy's accident for the Pentagon to get the money to fix that system," said William Delaney, the pilot's father. "Had they spent the money, they would have saved his life. He would still be here."

In a written statement, the Naval Air Systems Command said the true nature of the flaps problem wasn't understood until the Delaney investigation.

Higbee said each decision was based on the information available at the time. "In hindsight," he said, "you would have done something different."

The slowness in dealing with the flaps reflects a pattern of deferred repairs to critical Harrier components.

Ejection System

The ejection system killed three pilots between 1990 and 1998 and seriously injured several others before the Marines made modifications, records show.

A 1998 report of the Harrier Review Panel, a Marine Corps commission, said that while the ejection seat had worked at least 94% of the time since 1985, it also suffered from "known deficiencies."

The report cited concerns with the seat's computerized selection of ejection speeds and with an apparent design flaw that sometimes caused the harness straps to slap forcefully against the pilot's helmet. Because the Harrier operates so close to the ground when flying vertically, its ejection seat is especially fast-opening and therefore especially violent.

In addition, the Navy declined for at least 10 years to pay for a recommended steering mechanism that would have given parachuting pilots greater maneuverability. All other Navy and Marine aircraft have such a system.

The ejection seat manufacturer had proposed testing a four-line release system in the Harrier parachute as far back as 1988. But the Naval Air Systems Command responded that there were "no funds available," according to an internal document.

The systems command found the money after a prominent Marine pilot was killed in 1998 while ejecting from a Harrier.


Marine Corps Harriers on the flight deck of the Navy assault ship Peleliu. Three AV-8Bs have crashed this year. The causes are being investigated. (photo: Don Tormey, Los Angeles Times)

But the parachutes were not equipped with the new system in time to save Col. Kevin E. Leffler, who ejected into high winds over Death Valley National Park in 1999. With little control over his descent, Leffler dropped into a canyon and died when his head struck a boulder.

The Marine Corps investigation concluded that a four-line release steering system "might have saved the pilot's life." It has since been provided as part of a $7-million package of parachute improvements.

Engine Trouble

The Harrier's single Rolls-Royce engine has been a persistent source of trouble, playing a role in more than half of all Harrier accidents between 1980 and 2001, according to a Times analysis of the Naval Safety Center's aviation database. Even after the Marines swapped out the Harrier's original AV-8A model for the improved AV-8B, the new engine proved to be five times less reliable than the engine in the F/A-18 Hornet, according to a 1992 audit by the Defense Department's inspector general.

"The whole 20 years I was there, they were always doing engine modifications but they could just never fix it," said Clinton M. Higginbotham, a retired Marine Corps major who spent much of his career maintaining the Harrier. "It was just a bad engine from Rolls-Royce."

The company has paid to correct two design defects. But Rolls-Royce officials said other engine problems stem from the failure by the Marines and the Naval Air Systems Command to make engineering changes recommended by the company or to invest enough in testing and maintenance.

When Capt. Daniel J. Gilkey had to bail out of his Harrier at 17,000 feet over the California desert in June 2000, the engine part that failed had been marked for modification years earlier by Rolls. The same part had malfunctioned in at least four other Harriers.

"The level of funding for the AV-8 has always been substantially less than other combat aircraft," said Colin H. Green, president of Rolls-Royce's Defence Aerospace division.

Since a meeting two years ago between the Marine Corps commandant and the Rolls-Royce chairman, cooperation has improved and both have dedicated additional resources to the engine, officials on both sides say. After Gilkey's crash, the Naval Air Systems Command grounded 101 Harriers while working with Rolls-Royce to make modifications the company had recommended previously, said Ryan, the Rolls executive. The planes remained out of service for at least six weeks and some did not return to the air for close to a year.

It was not the only time engine problems compromised Marine readiness. In 1991, the new night-attack version of the Harrier was held out of the Persian Gulf War after a nonfatal crash revealed that the engine casing flexed at high speeds, creating friction that caused fires. The corps and the Naval Air Systems Command had chosen not to subject the new engine to expensive preflight testing.

"Oh, it just drove us crazy," said retired Col. Ronald V. Deloney, who led the squadron that had to stay home. "We watched the war on CNN."

Money Problems

Many Marine leaders maintain that money for the Harrier has always been scarce.

Some blame the Navy, which controls the corps' aviation budget. They say the Navy has been tightfisted because the Harrier is flown only by the Marines. And several retired generals said the Navy, which is largely defined by its aircraft carriers, has felt threatened by the Harrier's ability to launch from assault ships with shorter decks.

But former Navy secretaries said they never intentionally shortchanged the plane and that the Marines set their own budgetary priorities. The Harrier, said Sean O'Keefe, who was Navy secretary during the first Bush administration, simply "proved to be more expensive, harder to maintain, harder to operate and required a lot more care and feeding than other tactical aircraft."

Retired Rear Adm. Robert H. Gormley said Navy leaders were "very resentful" that the corps poured money into a plane that never seemed to outpace its problems. "They just didn't see what the return was for these dollars," he said.

In 1998, the review panel concluded that the AV-8B's risk was "uniquely high" and recommended more than 50 fixes and upgrades. The Marine Corps and Navy committed $133 million over six years to make the improvements. About a third of the money had been spent as of late last year and, as of Oct. 2, 29 of the panel's recommended improvements had been completed, the Marines said. Another 19 are underway.

Now, said Lt. Col. R.E. Claypool, who commands a Harrier squadron in Yuma, Ariz., "we've got our arms around the engine."

Retired Marine Commandant Charles C. Krulak, who convened the review panel, called the new and improved Harrier "an ass-kicking machine."

That same machine crashed three times this year.

Deadly Fixes

On his first day as a Marine mechanic, Larry Stoneroad strolled out on the tarmac at Cherry Point to marvel at a hovering Harrier. His awe quickly turned to horror as the plane rolled over and slammed to the ground, fatally injuring the pilot. "I thought, 'Oh, my gosh! How many of these am I going to see?' " he said.

Stoneroad saw enough that he retired in 1999, after 25 years, without ever accepting a pilot's offer to take a spin in one of the very planes he maintained.

"I'd seen too many of them go down," he said.

Marine mechanics have long faced the unenviable task of ensuring the Harrier's safety. The plane's inaccessible engine and myriad breakdowns are challenge enough. But many also have tackled their jobs with inadequate training and support, under tremendous pressure to keep the planes flying.

Stoneroad's longevity made him something of an oddity among Harrier mechanics.

Gunnery Sgt. John Higginbotham, a senior Harrier mechanic at Cherry Point, said it was not so long ago that, with just three years under his belt, he was the most experienced mechanic in his squadron.

In Britain, where maintenance-related mistakes are relatively rare, some Harrier mechanics have worked on the plane for more years than their American counterparts have been alive.

Some Marine leaders acknowledge that the Harrier, quite simply, is often too complex for the recent high school graduates who typically maintain it.

"We had regular guys fixing them, not engineers," said retired Lt. Gen. Charles H. Pitman, a former chief of Marine aviation," and so we found that some of the problems were caused by us doing something we shouldn't have done."

For instance:

Maintenance workers installed the wrong size washer in Col. Leffler's plane, which led to an engine fire that forced his ejection over Death Valley in 1999, investigation reports show.

The Marines knew the wrong washers had been installed in some planes. Three years before Leffler's accident, the same mistake had caused two nonfatal crashes. The Naval Air Systems Command ordered inspections.

Harrier engine

The Harrier's engine has been an ongoing source of trouble, figuring in more than half of all accidents between 1980 and 2001 (photo: Tyrone Turner for the Times)

But military officials did not believe all Harrier engines contained the bad part. And with some Harriers deployed overseas, they did not want to ground the entire fleet, according to an internal maintenance bulletin. Leffler's plane was among those exempted.

"Given the critical consequences of this decision," the investigators wrote, "it would have made more sense to inspect all engines."

In 1983, a TAV-8A, the first-generation two-seat training model of the Harrier, crashed after the plane's control stick jammed during a vertical takeoff. Capt. Paul L. Spargo Jr. ejected too late, dying instantly as he slammed into the ground.

Investigators discovered that loose parts under the floorboard had caused the problem. Mechanics discarded the parts six years earlier, after using the cockpit to test experimental equipment while the plane was sidelined by an earlier crash. With the parts undetected, the plane was returned to service.

Just six days before Spargo's crash, another pilot reported that the control stick was binding. Mechanics could not replicate the problem on the ground and cleared the plane for flight.

"Everybody was shocked that something like this would happen," said Newton A. Collyar, the squadron operations officer at the time. "Somebody really dropped the ball."

compressor blades

Damaged compressor blades were recovered after Capt. Danial J. Gilkey had to bail out of his AV-8B Harrier at 17,000 feet in June 2000. Gilkey parachuted to safety. (photo: U.S. Navy)

Capt. Richard F. Davis died in 1975 when his AV-8A rolled on its side and crashed as he attempted a vertical takeoff.

Investigators discovered that a maintenance worker had left a flashlight in the engine bay, the equivalent of stitching up a patient with a scalpel still inside. The flashlight created "a severe loose article hazard condition" that could have caused the accident, the investigation report said.

Annie Davis Kennedy, the pilot's widow, did not know about the flashlight until The Times contacted her last summer. She had been told only that there was a fire.

"After all these years, to think it might have been human error," she said. "It rips me apart."


Maintenance workers were surprisingly frank about their shortcomings in a survey by the Harrier Review Panel for its 1998 report.

Forty-three percent said they were undertrained. Seven in 10 said they did not have enough equipment or spare parts. Sixty percent said they had seen a colleague do something wrong or dangerous and 35% said they had been forced to do so themselves.

Stoneroad recalled several instances when higher-ranking Marines overrode his objections and signed off on planes he considered unsafe.

"They'd just say, 'We need the airplane,' " he said. In those cases, Stoneroad said he would quietly warn the pilot during preflight checks: "Did you see that I didn't sign off on it? You just be prepared if anything happens." The pilots invariably would fly anyway, he said.

Lt. Col. Lee Schram, the Harrier coordinator for the Marine Corps and a former Harrier squadron maintenance officer, said he had never heard of a Marine mechanic being asked to do something improper. He said a series of checks ensures that "safety is stressed at all levels."

At times, Harrier mechanics have worked exhausting hours that only increased the chances they would make mistakes.

"You're not going home till it's done," said Harold Bunch, 25, a Harrier mechanic who left the corps earlier this year. "So if you see ways to cut corners without putting a pilot in danger, you do it. Of course, people have different judgments. There's a lot of gray areas when it comes to safety."

engine replacement time

Because of the Harrier's unique design, maintaining the aircraft is a time-consuming chore.

To preserve the plane's balance while hovering, its engine is located in the middle of the fuselage. It can be removed for repair only by taking off the wings, a gargantuan task requiring the disconnection of various controls, hydraulic lines and wires.

On average, it takes 550 man-hours to remove and replace the engine, the Marines say. It takes nine hours to perform the same work on the twin-engine F/A-18 flown by the Navy and Marines and 10 hours on the Air Force's single-engine F-16 Fighting Falcon.

The Harrier required about 25 hours of maintenance for every hour of flight in 2002, two to five times the hours needed for various models of the F/A-18. The Harrier's cost per flight hour, a figure that includes maintenance, was $5,351 in 2001. For the Marines' F/A-18C it was $3,871.

The demands on time and money can create a disincentive to perform heavy maintenance.

For example, the Marines do not regularly schedule the Harrier engine for major overhauls, as is done with most military aircraft. They order major repairs only when a problem is found or near the end of the engine's life expectancy. The savings are considerable. But the money has not always been available for the rigorous testing needed to forecast when various parts will wear out.

The Marines also scrimp on spare parts, causing mechanics to cannibalize components from one plane to keep others in the air.

As a result, planes often fly with known ailments, or "gripes," that are not considered serious enough to warrant immediate repair.

Ryan, the retired general now with Rolls-Royce, said it was common in the mid-1990s for Harriers to fly with 40 or 50 outstanding gripes, and sometimes with more than 100. Shortly after becoming a wing commander, he prohibited planes from flying with more than 10, a limit soon applied to all Marine aircraft.

"I felt it wasn't a healthy airplane," he said. "The can-do message is great, but we shouldn't be stretching it to encourage people to do things like that."

The money generated by the Harrier Review Panel report has paid for some maintenance improvements. It has provided financial incentives for experienced mechanics to reenlist and has purchased better equipment. Among the measures not approved was money to allow mechanics to remove engine sections, or modules, for repair without removing the entire engine.

Meanwhile, maintenance problems have continued.

In April 2000, Capt. Michael R. Brunnschweiler ejected over California after his Harrier lost hydraulic power and began tumbling through the sky.

Among several errors, investigators found that mechanics had improperly connected a hose that feeds hot air from the engine to the Harrier's 25-millimeter Gatling gun.

"The hot air leaks out right on the ammunition and it heats up and goes boom," Brunnschweiler explained. As he ejected, his parachute deployed with too much slack in the harness and a metal fitting slapped him in the head.

He lost sight in his left eye, fractured his skull and dislocated his right shoulder. After six surgeries, he gave up any hope of returning to the cockpit and took a medical retirement this year at age 31.

"I'm bummed, definitely," Brunnschweiler said. "All I ever wanted to do was fly."

Pilot Blame

In their final conversation, 1st Lt. Earle J. Anderson told his brother, Richard, that if he ever crashed in a Harrier, the Marine Corps would first bury him and then blame him.

"We are trained to believe -- and we want to believe -- that the aircraft will not let us down," Anderson explained that night. "So whenever something happens, it's almost always attributed to pilot error, so the rest of us can continue to believe."

Several months later, on Jan. 26, 1990, Anderson's Harrier plummeted from 35,800 feet to 22,300 feet in 24 seconds, entered a deck of clouds and disappeared into the sea near Japan. There were no radio transmissions and the Marines never found the plane or the pilot.

Despite a dearth of physical evidence, Marine investigators decided the most likely cause was pilot error, a bitter pill for Anderson's grieving survivors.

Anderson was not the only Harrier pilot to predict before dying that he would be faulted for the plane's failures. They have had good reason:

Capt. Dale W. Mulkey died in 1996 after bombs aboard his AV-8B detonated, sending it into a meteoric freefall.

The Marines discovered that ordnance officers had committed a series of mistakes, the most serious of which was arming the bombs with expired fuses, making them six times more sensitive to electrostatic effects. Even so, investigators initially blamed Mulkey, claiming he was diving too steeply and slowly when he released his bombs during a practice run.

Higher-ranking officials later removed the pilot error finding and substituted harsh language blaming the ordnance handlers "for recklessness and lack of attention to detail."

In 1988, Marine investigators attributed Kerry Dale's fatal crash to pilot error because he failed to safely land his AV-8B after the flaps froze. The flaps malfunctioned moments after takeoff and, with only seconds to react, he inadvertently sent the plane into a dive. Investigators dismissed the flap failure as only a "contributing factor."

1st Lt. Charles E. Simpson was blamed for the 1982 crash that took his life because he mishandled an emergency landing in a plane with a faulty external fuel tank.

Fuel did not transfer properly from one tank to another, creating an imbalance of 1,600 pounds on one side of the plane that "forced the pilot to perform a high-workload landing," investigators wrote. The fuel tank problem had been reported after a previous flight but not fixed.

The practice of blaming pilots in such cases is misdirected, said retired Maj. Gen. Paul Fratarangelo, a former Marine wing commander.

"When an airplane demands more than a highly trained and skilled pilot can deliver in an emergency situation, you fix the problem by redesigning the airplane, not by assigning pilot error," Fratarangelo said.

Pilots also have been faulted when investigators could not come up with other explanations.

In Anderson's case, unable to examine either his body or the wreckage, they developed an elaborate scenario based largely on supposition: The pilot lost consciousness after becoming so "task saturated" that he failed to notice his cockpit was losing pressure.

That sort of speculative conclusion would never be acceptable to civilian agencies like the National Transportation Safety Board, said Edward L. Monhollen, a former Army pilot and veteran aviation accident investigator.

"I find it borderline irresponsible to state a cause when so many other possibilities exist," Monhollen said.

One reason investigators resort to guesswork is that the Marines have chosen, for financial reasons, not to equip the Harrier with a flight data recorder strong enough to withstand all crashes. Even when the plane's data storage unit does survive, it can be difficult and time-consuming to download, and sometimes provides virtually no clues.

The Harrier Review Panel concluded that pilots had caused a quarter of the major accidents involving the AV-8B. But a mind-set persists among Marine Corps leaders that pilots are responsible for the vast majority of accidents. Retired Lt. Gen. Thomas H. Miller, a founding father of the Harrier program, put the figure as high as "80% to 90%."

Some familiar with the program said there seemed to be a particular eagerness to blame the pilot when he was no longer available to explain or defend himself.

"The big joke was that if the pilot lived, it was mechanical error," said Clinton Higginbotham, the retired maintenance officer. "If the pilot died, it was pilot error."

The official cause of a major accident is determined after an internal investigation's findings are reviewed and approved up the line.

Marine aviation chief Hough reacted indignantly to the notion that pilots have been blamed unjustifiably.

"I've been in the service 40 years," he said. "I've never heard that in my life."

Short on Hours

Investigators looking into the 1981 crash of a Harrier flown by 1st Lt. David S. Noble made a disturbing discovery. He had flown just 7.5 hours in the previous 30 days -- half what the Marines say is needed to fly the Harrier safely.

The investigators suggested the lack of flying time may have "reduced his awareness" of proper procedures. But they still held Noble responsible for the accident.

"If you're flying every day, you're not afraid of anything," said Noble, who survived his ejection but saw his career shattered. "But if you're down for a few weeks, you forget stuff.... You're stressed about it."

Studies have found that inexperienced Harrier pilots are more likely to be involved in pilot error accidents than inexperienced pilots of other combat planes. Yet flight logs and crash investigation reports show that Harrier pilots struggle to accumulate anywhere near the flight time they need.

One Marine who died in a crash blamed on pilot error flew as little as two hours some months, according to his logbook. His widow, who declined to be identified, once said he told her he "scared himself" because of his lack of flying time.

Retired Lt. Col. John W. Capito, a former Harrier squadron commander, interviewed young pilots for the Harrier Review Panel and learned that many were flying just four to five hours a month right out of flight school.

"It's not enough to fly a Cessna, much less a Harrier," Capito said. "These guys were getting a third of the flight time they needed and then people were wondering why they get in accidents."

Within the tightknit fraternity of Harrier pilots -- there are about 350 today -- the procession of flag-draped coffins and 21-gun salutes has taken a toll. Wives have learned to live with the dread that strangers in uniform may come knocking on their door at any time.

Family members spend the rest of their lives wondering whether it was the plane or the pilot. They ask themselves how the Harrier could have vanquished so many men who seemed so indestructible.

Col. John H. Ditto, the highest-ranking pilot ever to die in a Harrier, had 4,900 hours of flight time in his 24 years as a Marine, including two tours in Vietnam. But Ditto's planes were the A-4 Skyhawk and the F-8 Crusader, and nothing he experienced in those cockpits prepared him for the Harrier.

In 1981, having been tapped to become a group commander, he chose to learn to fly the jump jet because several AV-8A squadrons would fall under him. On Jan. 19, with only 13.7 hours of flight time in the Harrier, he lost control during a vertical takeoff and ejected straight into the ground.

A Marine Corps investigation concluded that Ditto stayed with his plane too long and noted his "limited experience."

His widow, Susan Page, said Ditto described the Harrier as "a bear of an airplane." Even now, she feels heartbroken for his reputation.

"If you mention his name to anybody, they will say he was one of the best sticks in the Marine Corps," she said.

"I would love to think that there was something wrong with that airplane, love to think that in that heap of metal, maybe they missed something wrong with it."