It was in August 1979, while searching for arrowheads inside Idaho’s Civil Defense Caves – ancient lava tubes running from Yellowstone National Park, partly repurposed as a cavernous fallout shelter during the Cold War – that one unfortunate family found it: a human torso in a hessian sack, buried 20cm beneath the sediment.
The identity of the deceased – whose other body parts were missing – and how they ended up in a cave just east of Dubois was debated from the beginning. Clark County Sheriff Earl Holden guessed that the clothing on the torso – a pin-striped shirt, white cotton T-shirt, knitted maroon sweater and black woollen trousers with braces – dated back 60 years or so, and jotted in his notes that they were the kind of thing a gambler of that era might wear. The coroner, Holden’s best friend Ernie Sill, disagreed, putting the death at a maximum of ten years prior due to the preservation of the remains. Doug Ubelaker, the pre-eminent Smithsonian anthropologist assisting the FBI on the case, guessed that it was between six months and five years.
Clark County is large, sparse and rural; Dubois, with a population of a little more than 400 then (about 600 now) is its county seat. The body, which caused a stir among locals, was a dark blot on the landscape, but absent of context. Pre-DNA testing and with scant information, the case was relegated to legend: told and retold in the local area, but otherwise left to go cold.
Then, in March 1991, while treasure hunting in the cave with her family, 11-year-old Anna Rogers padded into the dark along the cave’s right edge and turned her torch toward a pale white human arm. Authorities promptly uncovered more parts in hessian – everything except the head. The Idaho Bureau of Investigations and the FBI were called in, the bones were sent to an FBI lab, and plans for excavation by the Idaho Museum of Natural History and Idaho State University were put in place. “I’d like to get that skull,” Craig King, then sheriff, declared to AP. “That’s where the cause of death would show.” A systematic dig yielded nothing.
Ubelaker, who uses the story as an exemplar of difficult cases in his lectures, speculates in his book on the effect the experience may have had on young Anna Rogers, settling on the hope that she saw it through the childlike prism of adventure. Nearly 30 years after her discovery, Rogers tells me via email: “It changed the way I saw Dubois, it didn’t seem as safe as I thought it was. It was a small town compared to Seattle [where she lived at the time].”
“I felt sorry for him, thought maybe his family must have missed him. I was also worried there might be a murderer still on the loose.”
Anthony Redgrave didn’t know his father, who left his family when he was three months old. Growing up in Baltimore, Maryland, all he had to remember his dad were a few artefacts and stories. “Two photographs: one of him holding me, one of him just sitting on our sofa; a stuffed dragon my mom had gotten from him as a present; and a tape copy of Blue Öyster Cult’s Some Enchanted Evening,” he says over Skype, sitting alongside his wife Lee Bingham Redgrave in their Massachusetts home. “I didn’t really know anything about him, but one of the photos had his birthdate on the back.”
Anthony, who is trans and recently found out he was born intersex, says he’s always been a computer guy. He enjoyed science growing up, but also had a creative streak, and was the kind of kid who was into anime culture and Bowie. He went to school for visual arts, all the while teaching himself the fundamentals of genealogy – the study of family history – as he sought to find any evidence of his dad online.
He met Lee on OkCupid in 2006, and they bonded over their family searches; Lee, an adoptee, found out the identity of their birth mother (Lee uses they/them pronouns) at 18 and was later able to track down their father, confirming his identity using a home DNA kit from AncestryDNA. It was the first project the couple worked on together.
They operated well as a research team, going on to source contacts for Anthony’s father’s siblings in late 2010. “It was the day after Thanksgiving that I called one of the people who I thought was probably my aunt,” Anthony says. Before he knew it, the phone was getting passed around the whole family. “Now I have more than 100 cousins and a dozen aunts and uncles just from that side of my relatives.”
Lee previously worked as a doula, and sees some parallels with genealogy. “Maternity science is kind of a combination of art and science, it’s a fuzzy science,” Lee explains. “So is genealogy. The DNA part is science, sure, but a lot of genealogical research is kind of an art that you learn. It’s a weird hybrid.” The Redgraves honed their art over years as a side hustle, undertaking formal and informal training and helping adoptees find their parents for free doing “search angel” work. Later, they would go on to offer services such as full family tree reports to paying clients.
Then, in January 2018, something shifted. Their friend Christa Steele-Knudslien, a prominent trans rights activist in western Massachusetts who founded Miss Trans America and Miss Trans New England, was murdered by her husband.
The killing numbed them. They had met Steele-Knudslien after the suicide of another friend, Lars, in 2008. His passing was a catalyst for their personal journeys into genealogy, they explain, and Steele-Knudslien’s death pushed them to find some deeper meaning within it. “Her husband killed her and there’s nothing we can do about that. So what do you do with that energy? You throw yourselves into work,” Lee says.
Aware they were mainlining casework as an escape, a true-crime-enthusiast friend directed the couple to a Reddit post by a new non-profit called the DNA Doe Project. Founded by Margaret Press and Colleen Fitzpatrick, the volunteer organisation aimed to identify John and Jane Does – unidentified corpses, which were often the victims of crimes.
Initially, the Redgraves demurred, “eyeballs deep” in an adoption case. But their friend insisted they reach out, saying that their skills fitted perfectly and that they needed something to channel their grief towards. “We ended up on the phone with Colleen that night, for like an hour and a half,” Lee says. They soon became team leaders on some of the DNA Doe Project’s first cases.
uzzle solvers by nature, genealogists tie the loose ends of a person’s familial tapestry for answers. Under the tutelage of Press and Fitzpatrick, the Redgraves entered the world of forensic genealogy. It’s an emerging practice that pairs genealogical research with genetic data – typically gleaned from direct-to-consumer DNA-testing companies – to help identify victims (and sometimes suspects) in criminal cases. The DNA Doe Project is focused on identifying deceased individuals: there are around 40,000 “Does” in the US, possibly more.
Fitzpatrick had contacts in law enforcement agencies across the country through her company IdentiFinders, which helps to trace identities using Y-DNA – the genetic marker passed down from father to son – and public databases. The agencies were interested in the DNA Doe Project’s new technique, but neither Ancestry nor 23andMe, the two largest consumer DNA testing companies, were receptive. Both businesses’ terms of service require a living person to spit in a tube for a genealogy report, and they will only release a user’s data to law enforcement when legally required to do so. (23andMe’s guidelines state it “chooses to use all practical legal and administrative resources to resist requests from law enforcement”.)
The workaround was whole-genome sequencing, which would have cost millions once, but is now around $1,500 (£1,140), and is much more accurate than the Y-DNA method. A lab could take a Doe’s DNA and create a digital file of the whole genome. Crucially, a bioinformaticist could then reduce the file from three billion to roughly 600,000 letters, to be compatible with a service called GEDmatch.
GEDmatch is an open data personal genomics database which allows anyone to upload their DNA data files from different direct-to-consumer testing companies, and find the contacts of potential relatives who have done the same. The recent popularity of home DNA testing means a greater number of people are uploading their DNA to GEDmatch, increasing the breadth of its dataset (though it’s still small enough that forensic genealogists are only ever likely to see 3rd or 4th cousin matches).
“It was such a Hail Mary,” Press says over Skype. “We didn’t know what a dead person’s degraded DNA would look like in GEDmatch: whether there would be any hope of any matches at all… whether enough DNA could be extracted from the samples. But step by step, it just worked.” Behind Press’s shoulders, I make out cases represented by sticky notes, which are allotted into groups: solved on the left (30), active on the right (35)
Law enforcement agencies and medical examiners submit their cold Doe cases directly to the DNA Doe Project, and donations sometimes help cover the cost. The process works as follows: DNA from the Doe is extracted by a lab and then sequenced, often at another location. This may take several attempts due to the quantity, contamination, degradation or previous mistreatment of the DNA. A bioinformaticist then bridges the gap between the lab results and the GEDmatch upload, running sequencing data that can reach hundreds of gigabytes through an algorithm to produce roughly 13 megabytes of data suitable for upload. Once uploaded to GEDmatch, volunteers look for matches and begin populating an extensive family tree, possibly numbering thousands of distant relatives, using existing trees and documents.
The DNA Doe Project started with 12 or so remote volunteers but has since grown to between 60 to 70 dotted around the US and internationally, with hundreds on the waiting list. Volunteers have to be in the genealogy community and have experience with multiple unknown parentage searches. “You have to know what you’re doing, and how tricky those cases can be,” Fitzpatrick says.
The project’s breakthrough case was known as “Buckskin Girl”, a woman found murdered in Miami County, Ohio in 1981. The team uploaded her DNA to GEDmatch in March 2018 and within four hours, they knew her identity. At a press conference, she was confirmed as Marcia King, a 21-year-old born in Little Rock, Arkansas.
This was two weeks before the Golden State Killer was identified by genetic genealogist Barbara Rae-Venter and her team using similar methods, in a case that thrust forensic genealogy into the headlines. “So this was the first time in the world anyone had ever heard of genetic genealogy being used in this way,” Press says. “Then we started getting tons of calls. That was when we knew we’d be around for a while.”
As forensic genealogy enjoyed its first public successes, however, it was also met with increased scrutiny, in particular around concerns over how people’s genetic data could be accessed and used by law enforcement. On May 18, 2019, GEDmatch changed its privacy policy so that users had to opt in to allow law enforcement agencies to access their information, following an incident in which detectives were allowed to use the site to find a perpetrator of violent assault (the site’s own terms of service said it permitted access only for sexual assault and homicide cases). Overnight, the DNA Doe Project’s matches on a dozen or so cases were wiped. The cousin of the Buckskin Girl hadn’t opted to make their information accessible; the match wouldn’t have been made.
The DNA Doe Project also uses a similar database, FamilyTreeDNA, which had its own scandal when it emerged it was allowing the FBI to access users’ data without disclosing this to customers (it now also allows users to opt out).
As with any new field, standard practices are evolving, and there is contention around when law enforcement should be able to access people’s DNA data in an attempt to identify victims or suspects. In November 2019, Florida police obtained a warrant to peruse the entirety of GEDmatch – including those who hadn’t opted in – which academics warned set a dangerous precedent.
“It’s a brand new field,” says Fitzpatrick, who has since left the DNA Doe Project to focus on IdentiFinders. “And there are rogues. We’ve been trying to be as cautious as we can.”
professor at Idaho State University (ISU), met the duo and invited them to give a lecture to her students at the University of New Hampshire. There, she told them about the mystery even the FBI couldn’t solve.
Together with Samantha Blatt, who joined ISU as an assistant professor in 2018, Michael compiled a biological profile of the body. In it, they wrote that the body had likely been dismembered post-mortem, “using multiple methods and tools”, with no apparent cause of death. The person was probably of European ancestry, had brown/reddish hair, and was possibly between 25-45 years old. They sent the biological profile to Deputy Sheriff John Clements, the investigating officer on the case, and, with the sheriff’s office’s permission, submitted it to the DNA Doe Project in May 2019.
With the DNA extracted, sequenced and uploaded to GEDmatch, the DNA Doe Project volunteer team got to work in July, and was led by the Redgraves with various spreadsheets and an expanding family tree. It was clear this wouldn’t be a four-hour puzzle like Buckskin Girl. A common family name that came up was Loveless, or Lovelace – Mormon pioneers (and polygamists) from Utah, which created a “family tree spaghetti”, according to Lee, adding complexity to the process. Hundreds of first cousins would descend from a single grandparent; there was endogamy – marriage within a small community – and half-relationships, unpredictably affecting the way shared DNA presented between people in the tree. The broad estimate for the Doe’s age and post-mortem interval – the time elapsed since death – didn’t help.
But when bioinformaticist Greg Magoon found the Loveless name in the Y-DNA markers in October, and another volunteer traced the Doe’s DNA to a patrilineal ancestor, things began to crystallise. “It flopped it around, because [before] I was looking for somebody who descended matrilineally from Loveless instead of the other way around,” Anthony says. “And so that’s when things started really tightening up, and we went from 30 different people that it could be, to three.”
They researched these three, looking for proof of life and building out timelines. Two demonstrably lived and died elsewhere, working on the railroads. Records on FindAGrave.com showed that the third, Joseph ‘Henry’ Loveless, lay in his grave in Payson, Utah.
Except that couldn’t be true. The volunteers spotted that Henry’s wife Agnes Loveless’s FindAGrave profile came complete with a detailed death record dated to 1916: “homicide by axe”.There was a family story on one genealogical website about how Loveless had been murdered the same way. But Henry’s FindAGrave profile had no such death record, only a death date of 1915 – which, on closer inspection, was missing on the photo of his tombstone: “1870-“
On the phone, Payson City Cemetery confirmed there was in fact no internment associated with the plot. A family member had apparently confused Henry with his brother Jedediah, who shared the same first initial and who was killed by a “runaway team” of horses in 1915 – and mistakenly inputted his death date on Henry’s FindAGrave profile.
The volunteers looked at newspaper clippings from the time to corroborate tales of Joseph Henry Loveless’s life. His story remains undeniably wild for modern-day sensibilities and evokes the anarchic spirit of the frontier: born in Payson, Utah Territory in 1870, he was a bootlegger, thief and outlaw. His wife Agnes (also a bootlegger) was murdered by a suspected “Charles Smith” on May 6, 1916. On May 7, an Idaho newspaper reported, a “Walter Currans”, who also gave the name “Smith”, was arrested for the killing of Agnes, “his common law wife”. Several days later, according to another newspaper article, “Walter Cairns” escaped from his cell with a saw hidden in his boot, never to be caught.
In total, it took 14 volunteers more than 15 weeks of work to crack the case, with around 250 DNA cousins’ family trees being used to place 31,730 individuals in a combined tree. Once they zeroed in on Joseph Henry Loveless, with his outlaw status, minimal official records and lack of evidence of life beyond 1916, it became clear that he was likely the man who was arrested for Agnes’ murder and later escaped – regardless of how many aliases he may have used.
What really nailed it for the Redgraves that Loveless was the body in the cave was the “Walt Cairns” wanted poster for murdering Agnes and then escaping his jail cell in St. Anthony. A physical description for Cairns matched Loveless’s immediate family, including his distinctive lack of or very faint eyebrows. He was also listed as wearing similar clothing found on the body in the cave when he escaped prison: red sweater, black trousers…
Exactly how he was killed, or by whom, is still unknown. One theory is it could have been by Agnes’ family, or people who knew her. His remains were preserved due to the unique, stable environment of the caves. If you visit them today, rations from the 1960s appear as new, stacked neatly in boxes in the fallout shelter, while the cave where Loveless was found had calcium deposits on the ceiling and fine silt for soil.
On November 5, 2019, the DNA Doe Project provided the Clark County Sheriff’s Office with its preliminary report, along with contact information of Loveless’s descendants. One grandson, 87, was willing to assist, and Deputy Clements set off on a 16-hour road trip from Dubois to California, DNA test kit in tow. The grandson looked eerily like the composite Anthony had made of Loveless from photos of close relatives and his physical description. The results were consistent with a grandfather-grandson relationship.
Anthony Redgrave and Margaret Press travelled to Dubois for the sheriff’s official announcement on New Year’s Eve, along with Blatt and Michael. Blatt said “oohs” and “ahs” went up in a little store across the street from the press conference as local residents heard the details: the body in the cave had finally been identified.
The day after I spoke with the Redgraves for the first time, in January 2020, Christa Steele-Knudslien’s ex-husband was sentenced. He received life in prison with parole eligibility after 25 years. Steele-Knudslien’s father, Robert Steele, a humble midwestern man who accepted and grew close with Christa in her final years, had a lawyer read out an emotional statement.
In August 2018, shortly after joining the DNA Doe Project, the Redgraves had decided to create a special interest group for possible trans and gender-variant Does. The Trans Doe Task Force, which is legally distinct from the DNA Doe Project, initially funnelled trans and gender-variant cases into the DNA Doe Project pipeline, including cases known as Julie Doe and Pillar Point Doe (the latter of which has been tentatively solved).
Historically, trans Doe cases may have fallen through bureaucratic cracks or suffered from bias and prejudice. Forensic anthropologists, and some medical examiners, make an estimation on a Doe’s sex and gender based on various parts of their skeleton. When the police get the report, they’ll tick the male or female box, and this will inform their work on the case thereafter. Trans Does could be incorrectly listed as “John” or “Jane”, not reflecting their lived identity. There are also usually only two options; NamUS, the missing persons system, only added an “other” sex option in June 2019.
Trans Does were also often assumed to be sex workers, Lee says. ”That’s two strikes against you: you’re some kind of pervert prostitute, obviously you’re going to get murdered. That’s not fair. And that’s not acceptable for forensic science, either. You can’t have that kind of bias on a case.”
The Trans Doe Task Force casts a wide net to identify potential trans Does, taking into account things such as clothing and physical features. It is possible that any Doe could come under their umbrella; similarly, they may take on cases in which it turns out the person did not identify as trans or gender-variant.
Once the Trans Doe Task Force takes on a case, it aims to use archival research and outreach in the queer community to trace the person’s lived identity, making contact with their chosen family if they were estranged from their biological family. They also aim to drive better representations and understanding of trans and gender-variant Does.
The Redgraves stopped volunteering with the DNA Doe Project in April 2020 to focus more on the Trans Doe Task Force, which they plan to reorganise as a student club inside a broader online training program for would-be forensic genealogists, as well as their own company, Redgrave Research. Students working with the Redgraves will get to work on difficult cases under their supervision, with those who are “trans informed” able to help with Trans Doe Task Force cases, too. The project forms part of Anthony’s doctoral degree in education.
Lee describes it as a “teach a man to fish” way of thinking. “There’s a lot of work to do, and we don’t have enough people doing it,” Anthony says. “So this is bridging that gap.” They also hope to lay a foundation for standard practices and ethics in the field, and to create a public LGBTQ+ missing persons database that they can cross-reference with cases. Lee hopes the work will be somewhat healing, helping people in death potentially marginalised in life.
