Arthur “Bud” Kelder, Eakin’s cousin, had died while a POW during World War II, but his body had never been found. Bud’s parents had sent handwritten letters to the Army for years, asking to have their youngest son’s body returned to them in Illinois.
“It is our hope that his remains may be sent here, for burial at home,” pleaded one.
Six decades later, Eakin, a stubborn Texan who was himself a vet, resolved to find out exactly why Bud had never come home.
In the fall of 2009, Eakin reached out to family members and found that Bud’s older brother had kept a trove of historical documents laying out Bud’s saga: the telegram announcing he was a POW, newspaper clippings, letters sent between Bud and his parents.
Before the Japanese invaded the Philippines, Bud had worked as a dental assistant at the American hospital in Manila, a plum assignment in a tropical getaway. In late 1941, the Army private wrote to his parents about saving his paycheck to buy a custom-made sharkskin suit. It “really is a peach,” Bud wrote.
After war broke out in December 1941, Bud was among 12,000 American troops who were besieged by the Japanese for four months on the Bataan peninsula, just south of Manila.
“What I dream is there will be no more separations between us again and we’ll spend more time together,” he wrote to his parents in cursive two months later. “P.S. Mother! Don’t worry about me.”
The Americans surrendered on April 9, 1942. The Japanese made Bud a driver during the infamous five-day, 65-mile Bataan Death March. He eventually ended up at Cabanatuan, one of the largest POW camps for American troops.
When Eakin tried to figure out what happened next, he ran into an unexpected roadblock.
The Pentagon spends about $100 million a year to find men like Bud, following the ethos of “leave no man behind.” Yet it solves surprisingly few cases, hobbled by overlapping bureaucracy and a stubborn refusal to seize the full potential of modern forensic science. Last year, the military identified just 60 service members out of the about 83,000 Americans missing from World War II, Korea and Vietnam, around 45,000 of whom are considered recoverable.
At the center of the military’s effort is a little-known agency, the Joint Prisoners of War/Missing in Action Accounting Command, or J-PAC, and its longtime scientific director, Tom Holland. He alone assesses whether the evidence J-PAC has assembled is sufficient to identify a set of remains: A body goes home only if he signs off.
Over Holland’s 19-year tenure, J-PAC has stuck with an outdated approach that relies primarily on historical and medical records even as others in the field have turned to DNA to quickly and reliably make identifications.
Though finding missing service members can be difficult — some were lost deep in Europe’s forests, others in Southeast Asia’s jungles — Holland’s approach has stymied efforts to identify MIAs even when the military already knows where they are. More than 9,400 service members are buried as “unknowns” in American cemeteries around the world. Holland's lab has rejected roughly nine out of every 10 requests to exhume such graves.
Holland’s cautious approach is animated by a fear of mistakes.
“Our credibility is only as good as our last misidentification,” he said in an interview. “It doesn't matter that I've identified 500 people correctly. If I misidentify one, that’s what going to be the focus. That’s what's going to be on the news. That is what is going to erode the credibility. That’s what I go home with every night.”
The top military official at J-PAC, Gen. Kelly McKeague, said he believed the standards for laboratory work to identify a veteran should be higher than the FBI lab’s standard for a death penalty case. With what J-PAC does, he said, there’s “a lot more at stake.”
In recent years, J-PAC and the other agencies responsible for the MIA program have come under intensifying scrutiny. In 2010, when Congress added World War II to J-PAC’s mission, it mandated at least 200 identifications overall a year by 2015 — a benchmark the agency has already said it will not meet. The problems, including those of DNA, go beyond J-PAC. Last month, Secretary of Defense Chuck Hagel ordered a review of how the military manages the effort.
Time is running out. There are 35,000 missing who experts say are findable from WWII. But the MIA’s closest relatives are dying off — and often with them so are the chances for using DNA to finally identify their long-lost loved ones.
Working from a home he built by hand in Helotes, just north of San Antonio, the gray-haired, lanky Eakin began his search by requesting Bud’s records from the Army.
He got back a remarkable document. Despite horrid conditions, the Americans prisoners had left a road map to find Bud and others like him: the “Cabanatuan POW Camp Death Report.”
Barely surviving on a little more than two cups of rice every day, the POWs had still managed to document the deaths of their comrades. It was a monumental task. Of the several thousand prisoners housed at the camp, only about 500 made it out.
Each day, the survivors dug a single unmarked grave to bury that day’s dead. And each day, the officers kept a meticulous ledger of the men who had died. However little that death roster meant in the prison camp, later, the POWs knew, it would be crucial to getting the fallen home.
The entry for Nov. 19, 1942, lists 14 men. Among them: Private Arthur “Bud” H. Kelder, dead at 4:35 p.m. The 26-year-old had succumbed to pellagra, a vitamin deficiency common among the starved prisoners.
Bud’s parents, unaware of his death for eight more months, kept writing him. On May 20, 1943, Bud’s father wrote, “Dear son, another week has gone and still no word from you. We hope you have a few letters on the way to us.”
Bud’s 34-page file had sat untouched by the government for nearly 60 years, a forgotten folder in a vast repository at the National Archives. Had anyone looked it over, they would have found what Eakin did: The military actually had a pretty good sense of where Bud was.
The files included not only a date for Bud's death, but also a specific grave for him. After the war, the Army had gone to the POW camp to dig up and identify the bodies of more than 2,700 men who died there. Using the Cabanatuan camp report and recollections of the survivors, the Army numbered the communal graves and matched them with the death list. Bud and 13 others had been in Grave 717.
Limited by the science of the time, the Army was able to identify and send home four men from that grave but not Bud. In the early 1950s, the remaining 10, along with more than 900 others who died at Cabanatuan, were reburied individually as unknowns in an American cemetery in Manila. White crosses marked their graves, bearing the words “Here Rests in Honored Glory A Comrade in Arms. Known but to God.”
To Eakin, it seemed obvious that Bud was among the 10 unidentified sets of remains from Grave 717 — a relatively small group of bones that the military could dig up and test against the DNA of family members. He tried to make it easier for the Pentagon by tracking down relatives of the men in the grave, so they could provide DNA for comparison. He even turned over an envelope Bud had licked.
Eakin said he thought identifying Bud would be "such a no-brainer.”
Reviewing details of the case later, Joshua Hyman, the head of the University of Wisconsin-Madison’s DNA Sequencing Facility, described it as a “piece of cake.”
But no one in the military seemed to show much interest. Every month, Eakin called the Army, asking, “When are you going to go dig up these remains?”
Bud’s fate rested with Holland and J-PAC’s lab on a joint Navy-Air Force base in Honolulu.
With his beard, slightly ruffled appearance, and a penchant for storytelling, Holland has the air of a liberal arts professor. Walking through the hallway outside his lab, he pointed proudly to displays of artifacts found with recovered remains: Tattered uniforms, faded, crinkled pictures and pineapple-shaped grenades.
Holland joined J-PAC in 1992 and has spent his entire scientific career there, becoming the lab's director in 1995. “The lab has been his life,” said Mark Leney, a former J-PAC anthropologist who worked with Holland for six years. “He has an enormous sense of ownership of it.”
Though J-PAC is a military command, the generals who rotate in and out for temporary assignments take no role in the scientific decisions, leaving them to Holland. Almost a dozen current and former staffers described Holland as someone who bridles at being challenged and fiercely protects his fiefdom.
“Expressing dissent was clearly not appreciated and frowned on,” said Leney, who now teaches at the University of Massachusetts Medical School.
Leney said he and another anthropologist wrote Holland a memo in 2002 about problems with procedures and standards at the lab. They asked for guidance and clarification. “Here is my guidance:Don’t ever write a memo like this again unless it is stapled to your resignation,” Holland wrote back.
J-PAC currently has about 500 employees, including historians and military logisticians to coordinate overseas digs, but it’s work in the lab that people are most likely to conjure. The lab has even been featured on the new “Hawaii Five-0.” (Holland had a cameo.) Holland’s longtime work at the lab recently earned him a lifetime achievement award from the American Academy of Forensic Sciences.
Swiping his badge and entering the bright, sterile lab, which is not much bigger than a classroom, Holland walked by 16 long metal tables with bones laid out, some with complete skeletons, and others with pieces of bone that look like piles of jagged stones. Some bones were grayed out like seashells, and others were dark brown, almost burnt looking – the different colors betraying the countries in which they were found. Vietnam’s acidic ground bites at bones, leaving them pitted. One skeleton had a green-tinted sternum and ribs, a patina from oxidation of a copper belt with bullets that was strapped to the vet’s chest.
“There’s no formula that applies to every case,” Holland said.
Under Holland’s direction, J-PAC’s lab hasn’t prioritized DNA analysis, despite it being an advancement that has revolutionized forensic science. J-PAC’s method begins with historians sifting through archival material to start to narrow down who someone might be.
“It may be 100 people, 200 people, 500 people, that’s fine,” Holland said.
Then they compare bones to dental charts and other medical records. Combining that with archaeological analysis and artifacts, they try to winnow down the list to one person. DNA comes in last, only as a confirmation tool.
Scientists engaged in similar work elsewhere do the opposite. They start with DNA and let it drive the process, taking samples from bones they dig up and cross-referencing them against databases of DNA from the families of the missing to find a match.
“It’s how you get people identified these days,” said Clyde Snow, one of the world’s foremost forensic anthropologists. “It opened up a whole new world for us.”
Since the earlier 2000s, the DNA-led approach has been used in more than 30 countries to efficiently identify casualties of mass tragedies, including the United States after Sept. 11.
In post-conflict Bosnia, scientists initially used the traditional anthropological techniques that J-PAC relies on now and identified only seven out of the more than 4,000 bodies from the Srebrenica massacre, according to Ed Huffine, a forensic scientist who later designed a DNA-led process there.
Once they turned to DNA, Huffine said, they were able to make 400 identifications per month at the peak of their efforts.
J-PAC does occasionally start from DNA when bones from many people have been mixed together – as they have done with a complicated case involving 500 co-mingled remains from the Korean War.
But Holland’s deputy at J-PAC, John Byrd, said the lab rarely needs to resort to that. While using DNA first makes sense in places like Bosnia, where authorities lacked medical records for the missing, the U.S. military keeps copious records. And even advocates of DNA agree that relying on records can make sense in some cases.
J-PAC has also faced cases in which DNA wasn’t an option. Soldiers buried as unknowns from the Korean War were embalmed, making DNA extraction impossible. Holland’s team developed a much-admired innovation to get around that limitation, matching clavicle bones to chest radiographs taken to screen for tuberculosis.
But those cases are an exception. Typically, Holland’s lab has been able to extract DNA, including on all WWII cases it has worked on.
“If we worked together, concentrating on DNA, we could decrease greatly the time it takes to make identifications,” said a current J-PAC anthropologist.
Holland insists his process works. Making an identification “is an awesome burden,” Holland said, sitting in his paper-strewn office. Sticky notes act as a Rolodex, one wall displays dozens of photocopies of his hand with notes he wrote on his palm during meetings, and on a shelf in his bookcase are copies of two novels he penned – starring a fictionalized version of himself. “At the end of the day, I carry that burden.”
“If there is a better way to do it, I’m willing to take a look at it, but at some point the government pays me to do my job,” he said in a measured cadence that conveyed both annoyance and self-restraint. “And clearly I’m biased here, but I think I do a fairly good job.”
Holland’s job has been made harder by the overall military’s failure to systematically collect and sort comparison DNA samples from family members of the missing.
Scientists in Argentina, where around 9,000 disappeared in the country’s “dirty war,” started assembling a database of such samples even before they had the technology to analyze them.
“We were collecting samples even though there was no possibility to process them, but the relatives were dying,” said Mercedes Doretti, co-founder of the Argentine Forensic Anthropology Team. Later, Argentina launched a national campaign — funded in part by the U.S. government — to collect more DNA, using famous actors and soccer teams to promote it.
There has been no such campaign in the U.S. In an era of Facebook and Twitter, J-PAC officials heralded a 2001 Pentagon letter that ran in the widely syndicated “Ask Ann Landers” column. It didn’t work. Fewer samples came in after the letter than before, an Army official said.
Huffine and other experts say using DNA effectively requires one central database and single authority overseeing it. But the Pentagon has six different agencies handling aspects of DNA testing and collection, spread out from Hawaii to Delaware. Each military branch is tasked with collecting samples from relatives of the missing from their service.
The Pentagon has relatively complete records for Vietnam and Korea but only a fraction of the needed samples from WWII. Nothing was on record for Bud or the others from Grave 717 until Eakin got involved.
The Army organizes its samples broadly by war and not by theater, major battle or event. Officials said they had no way to discern how many samples they had in hand related to the Cabanatuan unknowns, for example.
Once the military does get DNA samples, there are further delays. The Armed Forces DNA Identification Laboratory, which processes samples for J-PAC, takes 110 days on average to sequence DNA, much longer than commercial labs.
“Just waiting for DNA and trying to find a donor for DNA” from a family member — “a lot of the pauses in progress occur there,” a former J-PAC anthropologist said.
Eakin first heard of Bud as a teenager, when, looking at framed photographs on the wall, he asked about a black-and-white photo of a young man he didn’t recognize.
“That’s Bud,” his grandfather told him in a quiet voice.
“That was the only time I saw my grandpa cry,” Eakin recalled.
The correspondence Bud’s parents had from the military ended in 1950 with a letter telling them that Bud was “not recoverable” and “should any additional evidence come to our attention indicating that his remains are in our possession, you will be informed immediately.”
His parents died in the 1960s, without any resolution about their son.
Herman Kelder was Bud’s only sibling. He left his son, Doug, a file box full of documents. Doug and Ron Kelder, Bud’s cousin, have looked to the dogged Eakin to solve the family mystery.
Bud’s story wasn’t meaningful to Eakin just because he was family. As a rowdy teenager in rural Indiana, Eakin joined the Army after a night of drinking with his buddies and did two tours in Vietnam. He wouldn’t easily give up on the cousin who hadn’t been as lucky as he was to make it home.
In the spring of 2010, Eakin had two breakthroughs.
He tracked down an audio tape recorded by Bud’s older brother in 1994, shortly before he died.
“I got out of dental school in 1935,” Herman Kelder said on the tape. As he started to build his practice, he had worked on Bud and “put some gold inlays in his mouth where he had some silver fillings.”
Distinctive dental work like that could help identify Bud among the 10 unknowns from Grave 717, Eakin realized. Perhaps the military had noted in its files if any of the bodies had gold inlays.
A week after making the discovery, Eakin attended a meeting hosted by the Defense Prisoners of War and Missing in Action Office, or DPMO, another Pentagon agency whose mission overlaps with J-PAC’s. The meeting’s purpose was to update family members on the search for their loved ones, but officials told Eakin that even aided by his new information, Bud was unlikely ever to be identified.
Not long after that, Eakin received a package in the mail from a sympathetic source inside the Pentagon: It contained what the government calls X-files for the 10 unknowns buried in Grave 717.
When the military couldn’t identify a set of remains in the years after WWII, it put all the documentation it had assembled — where the body was found, skeletal details and dental charts — into a file assigned an X-number to stand in for a name. There are about 8,500 X-files from World War II.
Eakin went through the X-files for the bodies in Grave 717. Only one — X-816 — had gold inlays.
Eakin was ecstatic that the clues led so directly to his cousin. Unknown soldier X-816 — seemingly, Bud — was buried in grave A-12-195 in the Manila American Cemetery. Eakin went back to the Army that April, thinking, “We’ll have him home in a week.”
By September, the only progress the government had made was to produce a research memo on Bud’s case. Heather Harris, a DPMO historian, concluded that the archival evidence — mainly the death report ledger that the POWs had kept at Cabanatuan — was unreliable and therefore insufficient to warrant disinterment. She also noted another complication: Some of the remains appeared to have been commingled when they were first disinterred in the 1940s. It was possible that identifications made then had been wrong.
On Sept. 21, a J-PAC dentist compared the dental records in the Grave 717 X-files with other dental records for the 10 men buried in that grave. He couldn’t conclusively match any of them. There’s no indication in the report that Herman Kelder’s taped statement about Bud’s gold inlays was part of the analysis.
Eakin had hit another wall.
By this time, proving X-816 was really Bud had become an obsession. Eakin was prone to such quests. When serious injuries from a helicopter crash ended his career as a pilot in the mid-’80s, Eakin combed through FAA data to understand the cause of the accident. He couldn’t simply file it away as an unfortunate accident and move on.
“I had to know why it fell out the sky,” he said.
Teaching himself to write computer software, he turned the agency’s 4-foot stack of 9-track tapes into a searchable database, then sold the agency his improved version of its own data. Consulting on aviation accidents became Eakin’s full-time job.
“If J-PAC thinks he’s going to give up, they’re wrong,” said Joan Eakin, his wife of 38 years.
Increasingly frustrated by the military’s inaction, Eakin broadened his efforts beyond Bud. After a drawn-out battle with the Department of Defense, Eakin obtained access to all World War II and Korea X-files, more than 9,400 in all. Then he built a database with material from 3,000 files for unknowns buried along with Bud in the American cemetery in Manila, methodically entering information for weeks in the evenings.
He discovered that all the way back in the 1940s, the military had made tentative identifications for more than half of those X-files, tying each set of remains to a few service members or, in some cases, just one. Many of the cases seemed as solvable as Bud’s. In one instance, Eakin concluded, a Silver Star recipient had never been identified because his name was spelled wrong.
“It’s not rocket science,” he told anyone who would listen.
Eakin didn’t understand why the military didn’t just exhume all the men who had been buried in Grave 717.
While some oppose focusing on disinterments, arguing those men have already been found, the Pentagon began encouraging J-PAC to exhume more unknowns 15 years ago based on the emerging promise of DNA. J-PAC has done just 111 disinterments since then. Half of those were exhumed in the past two years, in part a response to the mounting pressure to make more IDs.
“The laboratory is pursuing disinterments very aggressively,” Holland said.
J-PAC recently set a goal to dig up 150 sets of remains per year by 2018. Even at that pace, the agency would need until about 2081 just to get the 9,400 unknown World War II and Korea service members out of the ground.
Holland approves just a handful of the disinterment requests that come across his desk. Cases go first to the head of the disinterment unit, who dismisses about 80 percent of them, Holland said. The rest go to Holland, who said he rejects another 80 percent. That means only 4 percent of cases considered for disinterment move forward.
Holland said he is “handcuffed” by a 1999 Pentagon policy that requires a “high probability of identification” before exhumation.
There is disagreement within the military about whether the policy is still in effect. Officials at DPMO, which frequently squabbles with J-PAC, say it isn’t.
“It’s J-PAC’s choice,” said Navy Capt. Doug Carpenter, chief of accounting policy for DPMO. “How they choose to hold themselves accountable for disinterment is up to J-PAC, and they do a fine job.”
Regardless, Holland has the power to define the exact standard for disinterment. He has imposed strict parameters for how many people a set of remains could possibly be before moving forward to dig them up: typically, no more than five. He will not lower the threshold even in instances where comparison DNA could be used to identify disinterred remains.
Holland might still be smarting from a 2003 blunder in which J-PAC dug up remains of unidentified sailors who had been on the USS Oklahoma when it sunk in Pearl Harbor. Agency officials had failed to investigate the historical archives adequately and exhumed a container with bones from hundreds of people instead of the five they were expecting.
After this, the Pentagon decided to halt any further Oklahoma cases and required additional layers of bureaucratic approval for disinterments, but as a practical matter it is still Holland who makes the calls.
McKeague, the military commander of J-PAC, now has to endorse disinterment requests approved by Holland. Asked whether he ever questions Holland’s judgment, McKeague held up his hand like a stop sign and said that lab decisions were squarely Holland’s domain.
“His credentials are impeccable,” McKeague said.
The request next goes to DPMO, which has its own historians check the work, then to the assistant secretary of the Army for final approval, but the Army has never turned J-PAC down.
Holland acknowledged that what he called the Pentagon’s “knee-jerk” reaction to the USS Oklahoma mess left him reluctant to push for disinterments for fear of losing more autonomy. He expressed concern about having bodies that J-PAC can’t identify stack up in his lab.
“I might have a certain amount of discretion in how I interpret ‘high probability,’ but that discretion will be taken away from me” should the ratio of exhumed graves to identifications get out of balance, he said. “I guess in that sense, I am a little risk averse.”
In January 2011, Eakin’s persistence finally seemed to spur some progress.
J-PAC anthropologist Paul Emanovsky examined the cases of the 14 men who had been buried in Grave 717 in Cabanatuan, including Bud’s, and concluded that identifications were possible.
“I think it’s worth pursuing these cases, there are some pretty strong correlations for a couple of causalities, and others are reasonable,” Emanovsky wrote in a previously undisclosed email to Holland and the lab manager, John Byrd.
Regardless of the likely poor condition of the bones, “I think that all hope is not lost,” Emanovsky’s note said. Since the bones of the soldiers might have been commingled, he advised exhuming them all and comparing them with DNA from family members. “We could potentially identify several of these individuals,” he said.
Ten months later, the head of J-PAC’s office of disinterment reviewed the case and echoed Emanovsky’s findings in a brief memo, noting that while the bones were “eroded,” DNA “may help in identification of remains from Common Grave 717."
The memo was sent to Holland on Oct. 19, 2011. Even then, he did not act to disinter Bud and the others and carry out DNA testing.
At another family update meeting in Texas on Feb. 26, 2012, Eakin met with Johnnie Webb, J-PAC’s external relations chief. Despite the reports from members of the agency’s staff supporting disinterment, Webb told Eakin there was no evidence to support further investigation into Bud’s case.
Eakin had had enough. “We waited and waited and finally filed a lawsuit in federal court,” he said.
On Oct. 18, 2012, Eakin sued the Department of Defense, naming the secretary of defense, Webb and the head of DPMO, in United States District Court in Texas to force them to disinter Bud and the other World War II unknowns.
The lawsuit has survived several government motions to dismiss.
In a January 2013 memo prompted by the suit, Holland cited the policy that, again, he himself interprets.
“No definitive individual associations could be established based on the available documentation,” he wrote. “While it is possible that one or more individuals could be realized if all unknown remains from this incident were disinterred for analysis, the existing and available data do not meet the level of scientific certainty required by current DoD disinterment guidance.”
Citing the suit, J-PAC and DPMO declined to comment on any specifics of Bud’s case.
After years of fruitless struggle, Eakin has become convinced that J-PAC’s “job is justifying doing nothing — and they do their job well.”
It’s unclear whether the law will come down on Eakin’s side, but there’s reason to believe the judge is sympathetic to his claim.
“Notwithstanding giving his last full measure of devotion to this country,” U.S. District Judge Fred Biery wrote, “Private Kelder’s government declines, on technical legal reasons as opposed to spirit of the law, to give him a decent burial in a marked grave alongside others who died in service to the United States.”
The Kelders want to put Bud in the family crypt in the Norwood Park neighborhood of Chicago where Bud grew up. That’s where Bud’s parents, who so badly wanted to bring their boy home, are interred.
“All this isn’t about just digging up a bunch of old bones. This is about giving a family closure,” Eakin said. “These men gave their all, and we can at least give them their names on their headstone.”