This past October, drawn by a mystery that has long since faded from public memory, I drove 20 miles on a rugged jeep road to a trailhead deep in the Holy Cross Wilderness southwest of Vail. From there, I trekked solo for several miles toward the face of 13,371-foot Gold Dust Peak. As I ascended the final pitch, I scrambled over massive rocks, some the size of beer kegs, others as large as Toyota Camrys.

Two days before I ventured into the wild, I met with Lieutenant Colonel Tony Somogyi at the Eagle County Regional Airport. Somogyi runs the Colorado National Guard’s High-Altitude Army Aviation Training Site (HAATS), a helicopter unit that flies more mountain rescues than any other group in the United States. I stopped by to borrow a GPS locator beacon, to study the team’s aviation charts and topographic maps of the nearby wilderness, and to assure Somogyi that he would not have to send a team to search for me in the coming days.

Eighteen years earlier, in July 1997, HAATS helicopters had transported a team of Air Force pararescue jumpers, known as PJs, and me to Gold Dust Peak. The PJs scoured the mountain that day looking for debris from a plane that had crashed into the hillside. At the time, I was an inactive member of the Army National Guard training in Aspen; I had heard news of a military plane veering off course and careening into the mountains near Vail and was curious about what had happened. Thanks to my military background—and the fact that I’d attended an Army Special Forces combat diver course with some of the PJs—the Air Force team allowed me to join them on their mission.

In the years since, I’ve often thought about the crash. I’ve thought about how, even after the military searched for months across thousands of acres in multiple states, one important question remained—a question that has gone unanswered for nearly two decades. When I retired from the military last September, I decided to return to the location searched by the PJs. On my way to the talus-covered summit in October, I discovered wreckage strewn about the landscape: fragments of a cockpit and mangled sections of rubber tires, tubing, and wires. Finally atop the mountain, I paused to take in the view of jagged peaks and high alpine lakes and considered what must have been running through the minds of the PJs as they meticulously searched the area: What happened to the bombs?

On the morning of April 2, 1997, Air Force pilot Craig Button left the Davis-Monthan Air Force Base in Tucson, Arizona, in the cockpit of a single-seat A-10 Thunderbolt jet. Button piloted one of three A-10s on a routine training mission to the Barry M. Goldwater bombing range in Arizona between Tucson and Yuma. As part of the mission, Button’s jet was equipped with live bombs, which he’d been instructed to drop on simulated enemy targets.

Roughly an hour after takeoff, 32-year-old Button unexpectedly broke formation, flew past the Goldwater range, and continued several hundred miles off course. Air Force radar traced Button’s flight path diagonally across Arizona, toward the Four Corners region. Button entered Colorado and likely flew over Telluride and Aspen. He eventually entered Eagle County airspace, where he crashed into the northwest face of Gold Dust Peak.

Locals reported hearing an explosion and seeing a smoke plume somewhere in the Holy Cross Wilderness, and satellites picked up an infrared event near the town of Eagle. The Department of Defense deployed an aerial search team that included helicopters and low-flying Cessna planes from the Civil

Air Patrol. Air Force jets scanned the area from miles above. On April 20, 18 days after the crash, a Colorado National Guard helicopter pilot spotted scraps of paper from the pilot’s flight plan and several pieces of metal in the snowfields on the north slope of Gold Dust Peak. Severe weather combined with the rugged terrain made rescue efforts challenging. It wasn’t until later in the week that the PJs recovered portions of the wreckage and Button’s remains.

Because of the harsh conditions, Air Force officials decided not to return to the crash site until early July, when dozens of military personnel used ropes and climbing harnesses to access the debris. At the time, I spoke with the officer in charge of the mission, Brigadier General Donald Streater, who’d examined what had been recovered from the mountain. The evidence, Streater said, suggested that the four 500-pound bombs aboard the A-10 when Button took off were not on the plane when it crashed.

What began as a search for a pilot and his jet shifted to a search for the missing bombs. As a safety precaution, the Forest Service closed the area to hikers while military officials conducted a detailed hunt within a mile-and-a-half radius of the crash site. They found nothing, and the wilderness area was deemed safe and reopened.

That summer, divers and explosive specialists combed every body of water under Button’s presumed flight path. Again, they found nothing. Officials concluded the search later that year having never recovered the bombs that were aboard Button’s jet. The official summary of the operations released by the Air Force states: “Still missing are four MK-82 bombs. No evidence has been recovered to indicate that the bombs were on the A-10 when it crashed.”

Forensic analysis of Button’s remains ruled out drugs, alcohol, disease, and hypoxia (lack of oxygen) as causes of the accident; the military declared his death suicide. As for the bombs, military officials thought Button must have dropped them unarmed—a pilot must manually arm bombs from inside the cockpit—somewhere along his flight path. Although there was no black box, Button’s A-10 was equipped with computer technology that would have recorded the moment he dropped the bombs, per Streater’s notes from the recovery operation. However, the computer pieces recovered in the wreckage were too damaged to provide a readout. Detailed flyovers of Button’s route never detected images of bombs or produced photographs of bomb craters. According to the military’s report, it’s possible the ammunition could have burrowed below the Earth’s surface. Still, the question remained: Where?

Theories as to the whereabouts of Button’s bombs have existed since the crash. One such explanation is that the pilot dropped them somewhere over Hopi Nation or Navajo Nation, which each stretch for miles across the Southwestern United States. That idea persists because the federal government reportedly had difficulty getting access to thoroughly search the area back in 1997.

Recently, though, I spoke separately with two men who worked on the crash recovery who each shared the same alternate hypothesis—one that I’ve not seen reported. One of the men I spoke with was Joel Best, a retired brigadier general and the commander of HAATS at the time of the crash. Best said he thinks Button could have dropped the bombs on the Goldwater bombing range—the destination of the original training mission—before peeling off toward Eagle County. “Button,” Best says, “would not have had the fuel to fly to Gold Dust Peak unless he’d dropped the bombs very early in his flight.” Best also says the theory explains why the ammunition was never found: Button’s bombs would be difficult to locate and identify amid a bombing range that contains thousands of craters and various metal fragments.

Dirk Applegate, a civilian project manager contracted by the Air Force in 1997 to assist in the recovery efforts, told me he’d also heard the supposition that Button dropped the bombs over the Goldwater range. Just as Best had, Applegate posited that Button would not have had enough fuel for the lengthy flight to Eagle County with the heavy bombs aboard.

Earlier this year I also contacted Streater, who is now 68 years old. He offered a far less specific theory. “He jettisoned them somewhere along the 495-mile flight path,” Streater says, “but not necessarily on the Goldwater bombing range.” Streater did say he was certain the bombs were not within a mile and a half of the crash site “based on everything we used to look for them and the complete absence of any evidence of the bombs in that area.” According to Streater, the final Air Force report was more concerned with the cause of the crash than the missing bombs. And Streater told me he was not aware of any follow-up investigation.

The Air Force was also unable to offer an explanation. Captain William Russell, who works as a public affairs officer at the Peterson Air Force Base in Colorado Springs—which handled the bulk of the recovery and cleanup operation in ’97—had not heard of Craig Button or his missing bombs. Russell’s predecessor, Colonel Robyn Chumley, who is now retired but served as an Air Force media liaison in 1997, did remember the crash. She said it was sobering to have to return a fellow airman’s body to his parents under those circumstances. But Chumley wasn’t aware of the Goldwater theory and said she had no idea where the four MK-82s were located.

Before I began my descent from Gold Dust Peak last fall, I paused one last time to survey the crash site—and attempt to let go of my own obsession with this unsolved mystery. After almost 20 years, getting a definitive answer on the location of the bombs seems unlikely. As Streater puts it: “Considering you’d need to conduct a ground search three miles wide and 495 miles long, a really comprehensive search of the flight path is mission impossible.”