For a while now, thanks in part to the reporting of the Washington Post’s Spencer Hsu, it’s been known that something was not quite right with the FBI’s hair forensics unit in the past. But only but only recently has the FBI admitted that failings within the unit led to hundreds, maybe thousands of questionable convictions before 2000.

In one particularly shocking case from 1978, two FBI-trained hair analysts who helped in the prosecution of a murder case couldn’t even tell the difference between human hair and dog hair.

The case involved a murder in Washington D.C. that year. The victim, a cab driver, was robbed and killed in front of his home. Before long, police centered upon Santae Tribble, then a 17-year-old local from the neighborhood, as a suspect.

Tribble maintained his innocence. But no matter what he said and how much his friends vouched, two FBI forensics experts claimed that a single strand of hair recovered near the scene of the crime matched Tribble’s DNA. Thanks to that evidence, which was groundbreaking at the time, Tribble was found guilty and sentenced to 20 years to life in prison after 40 minutes of jury deliberation, reported the Washington Post.

He would go on to serve 28 years until the truth came out: an independent analysis found that the FBI testimony was flawed. Not a single hair that was found on the scene matched his DNA. After attorneys brought the evidence to the courts, Tribble was exonerated of the crime, though he’d already been released from prison. “The Court finds by clear and convincing evidence that he did not commit the crimes he was convicted of at trial,” a judge wrote in the certificate of innocence released at the time, in 2012.

It gets worse. Not only did none of the hairs presented as evidence in trial belonged to Tribble, the private lab found that one of the hairs actually came from a dog.

“Such is the true state of hair microscopy,” Sandra K. Levick, Tribble’s lawyer, wrote at the time, in 2012. “Two FBI-trained analysts… could not even distinguish human hairs from canine hairs.”

Tribble’s case in not unique. In a Washington Post story released over the weekend, officials from the FBI and the Justice Department acknowledged the extent of their flawed use of hair forensics prosecutions prior to 2000.

The numbers are staggering. Over 95 percent of the cases involving hair evidence that the FBI has reviewed so far contained flawed testimony—257 out of 268 cases.

Thirty-two of these flawed cases involve inmates who are currently on death row. In 14 cases, the inmates have already been executed, or died in prison, The Post’s Hsu reports. About 1,200 cases across the country still need to be reviewed.

Now, the Post says that “federal and state prosecutors in 46 states and the District are being notified to determine whether there are grounds for appeals.” Inside the capitol, the only jurisdiction where investigators have re-investigated all the FBI hair-related convictions prior to 2000, three of seven defendants have been exonerated since 2009. Inmates in another two cases had already been exonerated before the review.

Specifically, the “flaws” mean that FBI experts gave “statements exceeding the limits of science,” and many instances where the shaky scientific ground that the statements were dependent on were not shared with defense attorneys. To put it simply, the testimony was presented as fact, when it was not.

In a joint statement to Hsu, the FBI and the Justice Department said they “are committed to ensuring that affected defendants are notified of past errors and that justice is done in every instance. The Department and the FBI are also committed to ensuring the accuracy of future hair analysis testimony, as well as the application of all disciplines of forensic science.”