Even more damning, a lightly soiled pajama pocket flap found near Colette’s body shows that she bled directly on it before it was ripped off. The area covered by the flap was soaked in blood, so we know the flap came off first. One sleeve was bloodied before it was ripped down the seam. The defense didn’t even try to dispute this evidence at trial.

Morris breezes past the inconvenient pajama threads and blood spatter. As far as he’s concerned, the whole crime scene was hopelessly corrupted. Everyone agrees that the military police lost control of the scene but that doesn’t explain how Colette bled on the top before it was torn, or how those threads ended up under her body.

Morris spends a lot of time trying to discredit a model the prosecution used to show that MacDonald stabbed his wife’s corpse 21 times through his own pajama top, creating 48 holes in the garment. Morris discounts the simulation because it relies on unsupported assumptions about how the top was draped over Colette’s body. If Morris is right, the prosecution’s model should be downgraded from incontrovertible proof of MacDonald’s guilt to highly plausible conjecture. But even if we throw out the prosecution’s pajama model, the rest of the fiber and blood evidence still points squarely at MacDonald.

MacDonald claimed that he attempted CPR on his daughters, even though they were both obviously dead when he found them. Yet his daughters, 5-year-old Kim and 2-year-old Kristy, were found on their sides with their mouths closed, posed as if they were sleeping. If he’d actually performed CPR on the dead children, they’d have been on their backs with their mouths open. We know Kim’s body was moved after she was mortally wounded because her blood and cerebral spinal fluid were found in the master bedroom. The fact that blue pajama fibers were found under her bedclothes suggests she was moved by someone wearing her father’s pajama top.

There are many, many more instances where MacDonald’s story conflicts with the physical evidence. And yet the jury might still have been inclined to believe the doctor—if they hadn’t also heard about the outlandish lies he told about the case. (He told his father-in-law and a family friend that he located, tortured, and murdered one of the intruders, for example.)

So why is Morris willing to believe that MacDonald might be innocent? Basically, because MacDonald seemed like such a good guy. Morris simply can’t believe that a Princeton-educated Green Beret doctor who was widely regarded as a loving husband and father would do such a thing. He thinks it’s perfectly plausible that a bunch of hippies would decide to attack a Green Beret for no particular reason, but he can’t imagine that MacDonald might have a dark side.

Morris also believes MacDonald might be innocent because of the confessions of local drug addict named Helena Stoeckley. Stoeckley was one of scores of Fayetteville-area hippies who were questioned in connection with the MacDonald case. At the time, she was a 17-year-old runaway who supported her omnivorous drug habit by working as a police informant. When her handler asked her if she knew anything about the murders, Stoeckley, still tripping, said she “felt in her mind” like she was there.

Stoeckley’s handler asked her about the murders because she sometimes wore a blonde wig and a floppy hat, like the one MacDonald described.

A military policeman saw a young woman with long hair and a wide-brimmed hat standing on a corner near the MacDonalds’ home as he rushed to the scene. Neither the officer nor MacDonald could identify Stoeckley from a photograph, however.

For years, Stoeckley alternated between confessing to anyone who would listen and swearing she didn’t remember anything. When she was finally called to testify at MacDonald’s trial, she said she was so high on mescaline that she had no idea where she’d been.

Morris suggests that Stoeckley knew things she couldn’t have known unless she was in the house that night, but these claims are unconvincing. Stoeckley was interrogated so many times over the years, and the case received so much publicity, that it’s impossible to know which details she picked up from her interrogators. Like many who trust Stoeckley, Morris tends to seize on what she got right—when she got it right—and forget about all the errors and wildly implausible claims she made over the years. Morris omits Stoeckley’s most outlandish “confessions,” such as her claims that she applied to babysit for the MacDonald children, burglarized their house three weeks before the murders, and got high and had sex with Jeffrey MacDonald.

Lindsay Beyerstein is a staff writer at In These Times and the lead writer at the Sidney Hillman Foundation. She blogs at Duly Noted and Clear It With Sidney.