Chimpanzees had been dying mysteriously at the Tacugama sanctuary in Sierra Leone for a decade or so by 2016 when Tony Goldberg set to work on figuring out why.

The disease was not contagious, did not infect humans and did not appear at other sanctuaries, but it killed chimps at Tacugama in an unmistakable, alarming pattern.

"It was always in the same season and always the same symptoms," said Andrea Pizarro, conservation manager at the sanctuary.

The chimps would show what seemed to be neurological symptoms: lack of coordination, difficulty walking and seizures. They would also exhibit signs of gastrointestinal distress, like distended abdomens and vomiting. If the syndrome appeared, not one afflicted chimp survived. Sometimes, chimpanzees that seemed fine one day were found dead the next.

The sanctuary, the only place for orphaned chimps in Sierra Leone, houses an average of a little more than 90 chimps. These are Western chimpanzees, a critically endangered subspecies. Fifty-six chimps at Tacugama have died from this mysterious disease.

What made the mystery even more puzzling was that the pattern of disease occurred only at Tacugama. Several investigations that focused on viruses or toxic plants produced no clear answers.

In 2016, Goldberg, a public health researcher and veterinarian at the University of Wisconsin, Madison, and head of the Kibale EcoHealth Project, was approached by the Pan African Sanctuary Alliance to try to solve the mystery. He and his colleagues joined forces with other veterinarians and biologists to undertake a comprehensive analysis of blood and tissue from the dead chimps that had been frozen at a nearby hospital.

"It took me five years," he said.

He and other researchers reached a landmark in their veterinary detective work with a report published in Nature Communications that identified a new species of bacteria clearly connected to the syndrome.

So far, the research has not found the bacteria to be the sole cause of the disease, but it has opened a new window on the bacterial genus Sarcina, which may include more unidentified species that threaten the health of humans and animals.

Goldberg emphasized that this was not a pandemic in the making. The bacterium is not contagious.

Researchers started with an extensive screening of blood and tissue from healthy and sick chimps for viruses, bacteria and parasites, using genomic studies, visual examination of the tissue and other techniques.

Leah Owens, a candidate for a doctorate and a veterinary degree working in Goldberg's lab, started concentrating on bacteria after initial DNA surveys showed only one likely culprit: a bacterium that was in 68% of the samples from sick chimps but none from healthy chimps.

Owens tried growing the bacterium in culture, sending it to other labs for sequencing, looking for it in tissue samples. Almost impossible to grow in the lab, the bacterium finally proliferated in a smear of brain tissue. Under a microscope, the tissue revealed the common shapes of bacteria: spheres and cylinders.

"I get to this one that just is nutso looking," she said. "When you look at it straight on, it looks like a four-leaf clover," but it is actually a cube of four spheres.

That indicated that it belonged to the genus Sarcina, which had included only two known species. One lives in the soil, and the other, first identified in 1844, is called Sarcina ventriculi and was known to cause gastrointestinal symptoms in humans and animals like those the Tacugama chimps suffered.

In humans, Sarcina ventriculi can thrive after surgery and produce gas that fills the walls of the intestine. Once the infection reaches that stage, people almost always die.

The technical term, Goldberg said, is emphysematous gastroenteritis, and "that's what the chimps had."

As Owens investigated further, it became clear that the bacterium in the chimp samples, including in the brain tissue, where an intestinal bacterium certainly did not belong, was not the same as the species reported in humans and animals over many years. It was bigger, and its genome had significant differences.