Contrary to what many pilots believe, it's not only intermittent high loading that fatigues metal; it's any flexing due to changes in loading, even the small changes that turbulence causes during routine cruising flight. Given enough time in service, all aluminum wings will eventually fail from fatigue, but airplane structures are designed to support many tens of thousands of hours of flexing.

"The amount of fatigue that occurs," Braly says, "depends both on the [magnitude of] stresses the structure experiences and on the number of times they occur." In other words, thousands of hours of cross-country cruising will fatigue a structure as much as repeated high-G loadings occurring a few times a day. Unfortunately, fatigued material looks the same as new material, until it's far enough gone for cracks to appear.

Evidence of fatigue cracking in even a single airplane raises a red flag with the FAA. If one airplane has cracks, it's likely that others of that type do as well. Within a month of the Sky Warriors accident, the FAA issued an emergency Airworthiness Directive, or AD, temporarily limiting all civil T-34s to 2.5 Gs positive and prohibiting them from exceeding 175 mph. The emergency action was not so drastic as some the FAA had taken, completely grounding Learjets in one case and Cessna 441 Conquests in another, but it was still a burden for airplanes that are widely used for aerobatics.

The FAA enlisted Raytheon Corporation, the parent of Beech Aircraft, to determine how best to ensure the future safety of the T-34 fleet. Raytheon spent almost two years on the problem while T-34 owners dangled in suspense. From the first, some owners darkly suspected that it was probably in Raytheon's interest, from the standpoint of limiting its liability exposure, to wipe out the whole fleet. Others, more charitable, thought that Raytheon's Beech engineers were as eager as anyone to keep the fine old airplanes flying, and that the length of time they spent coming up with a prescription was really intended to give beleaguered T-34 owners, who were faced with the possible reduction to junk-bond status of their $200,000 investments, a little breathing room.

Whatever Raytheon's motives may have been, the company's eventual response was draconian. According to an AD issued in August 2001, each T-34 front spar and rear spar attachment would have to be subjected to a magnetic eddy-current inspection for cracks every 80 flight hours. The inspections, which are relatively difficult to perform (and also prone to yield occasional false positive findings), would cost thousands of dollars, even after initial modifications had been performed to make the suspect areas more accessible. For a heavily used airplane, the 80-hour interval could mean two or three inspections a year. Nobody would want an airplane saddled with such demanding inspection requirements; in five years, says Lima Lima's Bill Cherwin, the requirement would "turn the whole fleet into beer cans."

Even before the FAA published the Raytheon inspection procedure, however, T-34 operators and their support organizations and businesses had begun to think about what the FAA calls alternative means of compliance. The FAA allows AMOCs as an avenue for independent solutions to engineering problems. Raytheon had not come up with a solution that T-34 owners could afford. If there was to be an affordable solution, owners would have to come up with it themselves.

In principle, two paths lay open. One would completely eliminate the suspect spar components, and therefore inspections for fatigue cracking. The other would perform one eddy-current inspection of the existing spar to ensure that it was free of cracks, and then strengthen it so as to preclude future fatigue failures.

The simplest repair, the Saunders Strap, had been around for decades-the T-34 was not the first Beech airplane to have spar problems. The first was the pre-World War II Beech 18. When Model 18 spars got into trouble, Dave Saunders, a freelance engineer, stretched a stainless steel strap under the belly of the airplane from one outer wing panel to the other. The strap took over a share of the load being carried by the lower elements in the spar, which were the only ones subject to significant fatigue.

Other Beech aircraft-Queen Airs, King Airs, and Beech 99 Airliners-experienced fatigue cracking, and Saunders adapted his straps to all of them. When it became apparent that T-34s were going to need spar work, Saunders lost no time designing a strap for them; it is also suitable for Barons and Bonanzas.

The Saunders Strap costs only $12,000 installed, and the down time, assuming that the required eddy-current inspection doesn't turn up cracks in the spar, is only two to three weeks. While no one questions the structural effectiveness of the strap-no airplane equipped with one has ever suffered a wing failure-some T-34 owners object to the ridge it produces along the underside of the wing; purists don't like the idea of tacking a conspicuous structural Band-Aid to the outside. And, though Saunders denies it, some say that the strap saps performance.

Earle Parks' Amarillo, Texas shop is equipped to rebuild T-34s from any condition. In addition to a huge inventory of spare parts, Parks has enough tooling to build an entire airframe from scratch-if it were legal to do so.

Parks had his own ideas about the T-34 spar. He had seen enough disassembled T-34 wings to know that there was some random variation in the size and shape of the many small parts, some of them shims and spacers to bring larger members into alignment with one another. Beech had eliminated the buildup of small parts when it designed a new spar that it has installed in new Bonanzas and Barons since 1973, and Parks decided that he could do the same. He replaced the inboard section of the shear web-the thin vertical element of the I-shaped main spar-with a sheet of heavier stock, eliminating the joggled lap joint that coincided with the location of the fatigue failure in the Sky Warriors airplane. He replaced the complicated buildup of small parts in the lower spar cap-only the lower cap is subject to significant fatigue-with a single long part machined from a solid piece of aluminum.

Replacing a spar sounds like a huge job. In most wings, skins and ribs are riveted directly to the spar, so removing the spar entails drilling out nearly every rivet in the wing. The Beech wing, however, is an unusual design. It consists of three separate assemblies: the D-shape leading edge, the main spar, and the main torque box, a sheet metal structure between the main and rear spars. The three assemblies are neither riveted nor bolted together; instead, they are joined by stainless steel wires, about the thickness of a wire hanger, that run the full length of the wing through interlocking piano-style hinges. To separate the spar from the rest of the wing, all you do is pull out the wires. In building the first post-World War II high-performance personal airplane, Beech seems to have anticipated that periodic spar inspections might be needed someday.

The FAA required that Parks perform a stress test to the 9 G ultimate load. Parks built a heavy steel fixture, put a wing into it, and pushed on the wing with a hydraulic ram. The steel fixture deformed, but the wing did not. Parks got his Supplemental Type Certificate.

The most technically economical response to the spar situation emerged from GAMI, George Braly and Tim Roehl's company. GAMI first did a computer survey of the T-34 spar, using the now-universal method called finite element analysis, and found a hot spot of concentrated stress at the exact point where the Sky Warriors wing had failed. Then, using electronic strain gauges affixed to wing and spar surfaces, the engineers recorded the structure's reactions to G-loads applied in flight. The results confirmed the computer's diagnosis.

FAA Designated Engineering Representative Victor Juarez then designed a small, artfully tapered gusset that bridges the critical area, eliminating the stress concentration. The GAMI team hardened the perimeters of rivet and bolt holes in the affected area, increasing their lives several-fold. The FAA approved the modification without testing a wing, solely on the basis of extensive analytical documentation the company supplied.

GAMI intended to turn over the rights to the AMOC to the T-34 Association, but to Braly and Roehl's surprise, the association's board chose not to involve itself in the airframe repair business or to endorse any particular AMOC. So Braly and Roehl formed the T-34 Spar Corporation, which provides the required inspections and modifications at a number of sites for $14,000.

The costliest repair is offered by Nogle & Black Aviation of Tuscola, Illinois. Charlie Nogle and his son Jud are, like Earle Parks, longtime eminences in the T-34 community. The Nogles scrap the existing spars and replace them with one more massive and better made than the original. The new spar also provides a shear web that runs all the way out to the wingtip (the original T-34 web stops a little outboard of the landing gear). The full-length web looks stronger, but its real function is to provide support for supplemental rubber fuel bladders in the leading edges outboard of the standard tanks, whose 50-gallon capacity may not be enough for the bigger-engine airplanes.

New spars-two are required-cost around $12,000 each, and the additional cost of dismantling the airplane and wing and refurbishing the wings as needed (it would be foolish to take the wings apart and not bring them up to near-new standard) runs another $12,000 to $20,000. Nevertheless, says Jud Nogle, the mod sells itself. "Pilots take one look at the new spar next to an old one, and they want to have it."

T-34 owners grumbled over the cost of modifying their airplanes. Many felt that the FAA had overreacted to the cracks. After all, only one airplane had had the problem, and that airplane, they felt, had been systematically abused. The same might be true of all airplanes used in air combat schools, which also provides "upset training" to pilots wanting to learn to recover from unusual attitudes, such as those produced by an encounter with the wake of a larger airplane. Surely it was unfair to lump them together with the T-34 operators who used their airplanes primarily for weekend outings and formation flights, and who rarely, if ever, indulged in acrobatics. It was rumored that in addition to the fatal Sky Warriors accident, a couple of other T-34s had experienced partial wing failures and had flown home to be repaired, and that these too had been air combat school airplanes.

As more and more airplanes underwent eddy-current testing and emerged with clean bills of health, and as more and more of the spars removed and inspected by the Nogle and Parks organizations proved free of cracks, owners and maintainers increasingly complained that the whole business was, as Earle Parks put it, a "fool's errand."

Then, on November 19, 2003, a second wing failed, in nearly a carbon copy of the first: The airplane, operated by Texas Air Aces of Houston, shed its right wing while maneuvering. The failure occurred at the same point as the failure on the first airplane. Both pilots died. The airplane's spars had not been modified, and it seemed probable that the airplane had been operating beyond the 2.5-G, 175 mph limits imposed by the AD.

Now the FAA raised new concerns. It had found cracks in the rear spar of the accident airplane. Owners became frantic, worrying over rumors of a new AD and possible cancellations of the existing AMOC authorizations. In the meantime, stung by suggestions that the local office had looked the other way while Texas Air Aces continued to operate its non-AMOC'ed airplane in violation of the 1999 flight restrictions, the FAA ordered a fine-tooth-comb inspection of all T-34 logbooks, paperwork, and service and maintenance histories.

The resentment many owners felt toward the air combat schools rose sharply. Was there not some way to differentiate between ordinary users and those who routinely applied high stresses to their airplanes?

The difference was not as great as all that, countered Robert Gold, owner of Sky Fighters, a Denver T-34 operator specializing in mock dogfighting and upset training. His company, Gold said, specifically discourages pulling lots of Gs during air combat maneuvering; in fact, it has a "strict and absolute limit of four Gs" and usually pulls no more than 3.5 on a simulated combat mission. Gold summed up the bottom line with an earthily persuasive argument: "It's not that much fun cleaning vomit out of the airplanes." In fact, said Gold, upset training occasionally involves higher G loadings than the "very choreographed" dogfighting does. Other operators of air combat schools flying different aircraft types, such as Siai-Marchetti 260s and Extra 300s, back Gold up. They don't pull lots of Gs because most customers don't like it.

The FAA had already turned a cold shoulder to the proposal that combat school airplanes be treated differently, arguing that because parts of airplanes often get exchanged without detailed record-keeping, it is impossible to know the service history of, say, a given set of wings. Nevertheless, it began to categorize T-34s as Type 1, those used in aggressive air combat or upset training, and Type 2-all others.

Early in March, the FAA published the latest revision of the T-34 AD. Citing new cracks found in the vicinity of the landing gear pivot fitting on the rear spars of several aircraft, it cancelled all the existing AMOCs and grounded, as of March 15, all T-34s that were not in compliance with the original 80-hour-interval Raytheon inspection requirement.