Trailed: One Woman's Quest to Solve the Shenandoah Murders

by Kathryn Miles

A recent study found that 80 percent of all wrongful convictions occurred because of demonstrable confirmation bias on the part of both investigators and prosecutors.

I asked him what he would recommend someone like me do. “Start at the beginning and investigate the case all over again. Assume they got it wrong the first time,” he said, finishing his coffee. “Doubt leads to inquiry. And inquiry leads to the truth.”

“Women and girls being killed, women and girls going missing—it’s a major problem we still aren’t addressing,” she said. “In our culture, the lives of women and girls seem to be not worth so much—particularly to law enforcement.”

It’s what theorist Kate Manne describes in her book Down Girl: The Logic of Misogyny as an extreme example of the policing and enforcing of women’s subordination by punishing those who flout “patriarchal law and order.” Misogyny, she contends, works by creating threatening consequences for women who violate or challenge those perceived

patriarchal norms. Today, there is still a subset of white men who view the wilderness as exclusively their domain and actively employ misogynistic or racist techniques in a misguided attempt to maintain that.

They included Gerald (Gerry) Zerkin, who defended Earl Washington, an innocent man who had been on death row for nearly twenty years after being convicted of raping and killing a woman in Culpeper, Virginia, in 1982. Washington, who has an IQ of 69, was coerced into making a false confession by police. Initial forensic testing revealed that the assailant’s blood was marked by a rare plasma protein. After it was determined Washington didn’t carry that protein, the state altered its forensic report to make it appear as if the tests to identify the protein were “inconclusive.”

Prior to an indictment that includes the death penalty in any federal case, the Office of the Attorney General is required to submit that case to a Department of Justice committee on capital cases for review. That same committee reviews the attorney general’s case as well as any counterarguments presented by the defense. Ashcroft’s office skirted that entire process,

The landmark 1963 Supreme Court case Brady v. Maryland requires federal prosecutors to disclose all exculpatory evidence prior to a trial. In 2002, after the government began its case against Rice, it provided the defense with about seventy-five hundred pages of documents, along with audiotapes and videotapes. Deirdre Enright began to piece together that those files were only a fraction of the key documents involved: one document would reference another not included in the files, or a twenty-seven-page file would be missing half its pages. Enright began keeping a list of what was missing. In court, she’d request the missing pages.

As far as Zerkin is concerned, there is a direct correlation between the notoriety of a case, the stakes for the prosecution, and the pressure they feel to get a conviction—sometimes at any cost. The Shenandoah case, he theorized, was particularly problematic because it was so high profile: the violent double murder had put the entire country on edge and called into question the safety of our national parks.

As far as I could tell, the medical examiner’s determination that the women had died on May 28 has never been called into question. Certainly, in all documents generated during the first year of the investigation, the FBI and NPS made regular mention of the twenty-eighth as the probable date of the crime (it was even written on that board in the storage unit). During the first year of the investigation, several eyewitnesses came forward to say they had encountered

Julie and Lollie after May 24. They included three cyclists who had stopped at the Panorama lodge to get out of the rain on the morning of Sunday, May 26, 1996.

The manager of the Panorama resort was also interviewed that month by the FBI and park service police. He told agents that he remembered talking to Julie in the lodge’s gift shop on the afternoon of either Saturday, May 25, or Sunday, the twenty-sixth,

Although Fleming continued to state she couldn’t be certain it was Lollie and Julie she had picked up at Hawksbill, Yee concluded in his subsequent report that her account was credible. In her 2002 interview with Yee, the Panorama waitress recounted a nearly verbatim version of the account she had given to agents and rangers in 1996. Nevertheless, Yee wrote in his 2002 report that he questioned the legitimacy of the waitress’s story.

Yee remained dismissive. In his 2002 report, he also discredited her account by noting that the two women were vegetarians and thus would never order breakfast meat, which ignored the waitress’s repeated insistence that she wrapped up the meat for their dog.

Yee asked him to recall the particular shade of her rain jacket. The manager said dark blue. Wrong, said Yee. It was royal blue. Next, Yee showed the manager a photo of Julie in her shorts and sports bra. He asked the manager if he was sure that that was the woman he’d met six years earlier. The manager said yes. Yee asked the manager if he was certain there weren’t any physical differences between the woman he remembered and the person in the photo: like maybe their weight, for instance? The manager conceded that the woman in the sports bra could have been ten pounds thinner than the woman in the bulky rain jacket, but that it was hard to tell. Yee ruled the manager’s eyewitness account uncredible, saying that Julie was fit and trim to the point of being skinny and that a difference of ten pounds was surely indication that the manager had met someone else.

That same week, Yee also visited Emma Garthoff and Zella Dingus, the two Potomac Appalachian Trail Club volunteers. Both distinctly remembered meeting Julie and Lollie. One produced her daily nature center journal, which cited Lollie by name on May 27. Yee discredited the volunteers’ 2002 eyewitness accounts as well.

It is also true that Clyde Yee’s park service career ended with significant credibility problems. From Shenandoah, he was transferred to the Grand Canyon area. While there, he received at least one letter of reprimand for insubordination. His supervisor also reported that he was dealing with multiple “serious problems” Yee had created in other cases. In

Berkowitz believes that noble cause corruption remains rampant in the park service. In his book Legacy of the Yosemite Mafia, Berkowitz points to systemic issues, like an agency-wide cover-up of the murders of multiple rangers that went on for decades, along with individual cases in which specific rangers suppressed evidence or provided false alibis for one another.

In the past twenty-five years, dozens of incidents have come to light in which NPS rangers conducted searches without warrants, continued to interrogate suspects after they requested attorneys, and used aggressive apprehension techniques. In one case, an individual died after being tased by a ranger when the man failed to leash his dog.

Meanwhile, a recent study by the National Registry of Exonerations found that at least 2,663 innocent people have been exonerated of criminal convictions since 1980. In more than 50 percent of these cases, official misconduct (including witness tampering, concealing or fabricating evidence, and inappropriate interrogation techniques), directly contributed to the false convictions.

Mike German. A graduate of Northwestern University School of Law, German had spent several years investigating white-collar crime for the FBI before becoming an undercover agent in 1992. A self-described “popular and successful agent” and the author of two books, German characterizes himself in his most recent memoir as “known and respected throughout the bureau.”

After that interaction, the FBI took more extreme measures. In June 2000, the agency concocted a new scenario. Mike McCarthy,

they determined, would commit a similar crime to the murder of Julie and Lollie and also in Shenandoah National Park. To make the story believable, they fabricated issues of the Washington Post with what appeared to be legitimate news stories detailing the murder of two unidentified women and a forest fire believed to have been perpetrated by their killer in an attempt to destroy evidence of the crime. Mike McCarthy sent Rice the clippings, along with an ominous assertion that he had to flee the country. The Richmond FBI then enlisted the agency’s Sensitive Operations and Support Group with the task of obtaining a postcard from Copenhagen, Denmark, and creating a postage cancellation stamp that would make it appear as if the card had been addressed and sent from there. On the postcard, McCarthy alludes again to the crimes and his responsibility for them.

On October 1, 2003, the lab sent back its findings: Rice’s DNA again did not match that of the hairs. Evonitz, on the other hand, could not be excluded. On October 21, 2003, the Richmond office again requested a DNA analysis of the crime scene hairs. This time, they requested that it only be compared against Rice, and not against Evonitz.

After reading the lab results from Evonitz’s DNA test, I contacted at least twenty different leading international geneticists. All said the same thing: that not only could you not exclude Evonitz based on that one base pair difference, but also the location of that difference made it all the more likely he had contributed the sample because of a condition known as “heteroplasmy.”

Deirdre Enright had first uncovered this phenomenon during her research in 2003. At the time, what was known was that a body sometimes misfires a protein in a particular place, stamping a C protein when it would normally stamp a T, for instance. As a result, two hairs taken from a person with heteroplasmy may not be perfect matches to each other, even though they share identical DNA.

I thought back to my visit to the FBI lab at Quantico almost exactly three years earlier: Even if we wanted to, we’re not the ones who get to open that vault, the scientists had told me. The lab exists in support of the field. All we do is examine what they tell us to. So why had no one told them to test Evonitz again?

She told them about that Hail Mary attempt to find Darrell Rice’s DNA on the women’s gags. The FBI knew that there was male DNA on at least one of them, and it was sent to the independent lab for the more sophisticated Y-STR testing. It, of course, excluded Rice as well. One of the students asked if that test had also excluded Evonitz.

According to his sisters, Marc began making obscene phone calls at the age of thirteen. Two years later, he began breaking and entering. He had rage issues and was prone to physical violence when angry. At a young age, he began watching sadistic pornography with his father and had a particular fondness for snuff films. He sexually molested his sisters and tried to persuade them to have sex with each other. He claimed to have killed a prostitute in Florida. His first wife told similar stories, that he was particularly interested in rape fantasies and had an explosive temper. He would punch holes in walls and break tables when he was angry; he’d go after drivers who cut him off in traffic. His violence became worse with his impotence in early 1996, and his behavior became more erratic then as well. In September of that year, he abruptly announced to his wife that they had to sell their Volkswagen Beetle, which had been his beloved car. He insisted they trade it in for a Taurus.

According to a 2018 internal agency memo, an FBI supervisor issued a moratorium on any further testing of evidence related to Marc Evonitz.

In the letter explaining why my request had been denied, an FBI FOIA officer stated that that information “is not in the public’s best interest and thus does not meet the Freedom of Information standard.”

But that sophistication has also come with a problematic legal cost. A few years ago, a father was convicted of raping his young daughter after his semen was discovered in her underwear. Subsequent forensic work found his semen in the clothing of his other relatives as well. It wasn’t until experts determined that something as simple as washing a load of laundry was enough to deposit those cells on other people’s clothing that he was exonerated. A few years ago, European officials were convinced they had a continent-wide serial killer after DNA taken from more than forty evidence kits pointed to the same individual. It took months before one enterprising lab scientist determined that the DNA actually came from a factory worker who had packaged the swabs used to collect evidence.

When I called him about the Shenandoah case, he asked to see the lab reports generated by the earlier testing. I sent them to him. We spoke on the phone a few days later. “This is exactly the kind of case we like to take on,” he said. “We have availability for a new one. Let’s see if the FBI will give us this one.” As of this writing, neither he nor I have heard back on that offer.

I can’t say for certain that Darrell Rice did not kill Julie Williams and Lollie Winans. But of the hundreds of experts I’ve spoken with, only the three agents involved in the initial investigation say they believe he is the perpetrator. Every other profiler, investigator, and scientist says that the preponderance of the evidence makes a strong case that Rice is innocent and that Marc Evonitz is a far stronger suspect.

That summer and fall, Deirdre Enright and I continued sorting through boxes and burrowing down one rabbit hole and then another. In our own Hail Mary attempt, we summarized everything we had found, and Enright sent it in a letter to a former federal attorney she knew was sympathetic to the case. We never heard back. But we kept at it.