It took just three weeks for two-thirds of all the world’s saiga to die. It took much longer to work out why.

The saiga is an endearing antelope, whose bulbous nose gives it the comedic air of a Dr. Seuss character. It typically wanders over large tracts of Central Asian grassland, but every spring, tens of thousands of them gather in the same place to give birth. These calving aggregations should be joyous events, but the gathering in May 2015 became something far more sinister when 200,000 saiga just dropped dead. They did so without warning, over a matter of days, in gathering sites spread across 65,000 square miles—an area the size of Florida. Whatever killed them was thorough and merciless: Across a vast area, every last saiga perished.

Richard Kock, a veterinarian and conservationist from the Royal Veterinary College, saw it all. He and his team were there on a routine monitoring trip to check the health of the population. “Mass mortality events are never nice things and I’ve experienced quite a few,” he says. “But the experience of the saiga was unprecedented, and unworldly. Even after 40 years of work, I just said: I don’t understand.”

The mega-death was all the more tragic because it struck at what should have been a time of celebration. After the collapse of the Soviet Union, poachers had brought the saiga to the edge of extinction, but governments and conservationists rallied to protect the species, and it rebounded. Saigas are good at that. Females can produce their first calf before their first birthday, and in every subsequent breeding season, most produce twins. So they recover quickly from disasters. By 2015, their population had quadrupled since the early 2000s, and it was predicted to do so again in a few years. “Everyone was saying: Oh great, we’ve really got them over,” says Kock. “They were beginning to talk about downgrading them off the endangered list. And then—bang—this happened.”

At first, the team suspected that a new infectious disease had spread through the population, but the pattern of deaths just didn’t fit. The saiga were dying too synchronously and too quickly. Also, all of them had died. “In biology, there’s certain rules, you know?” says Kock. “We accept that sometimes microbes can cause us harm, but not like this. Even very severe viral diseases or anthrax don’t do this. A good proportion of the animals would be fine.”

News of the die-off sparked outlandish explanations about Russian rocket fuel, radiation, and even aliens. But while conspiracy theories raged, a huge international team of scientists, led by Kock, got to work. Vets autopsied as many saigas as they could. Ecologists sampled the soil. Botanists checked the local plants. They couldn’t find any signs of toxins that might have killed the saiga. Instead, the actual culprit turned out to be a bacterium, one that’s usually harmless.

Pasteurella multocida normally lives in the saiga’s respiratory tract, but Kock’s team found that the microbe had found its way into the animals’ blood, and invaded their livers, kidneys, and spleens. Wherever it went, it produced toxins that destroyed the local cells, causing massive internal bleeding. Blood pooled around their organs, beneath their skin, and around their lungs. The saigas drowned in their own bodily fluids.

But that answer just led to more questions. Pasteurella is common and typically harmless part of the saiga’s microbiome. In livestock, it can cause disease when animals are stressed, as sometimes happens when they’re shipped over long distances in bad conditions. But it has never been linked to a mass die-off of the type that afflicted the saigas. What could have possibly turned this docile Jekyll into such a murderous Hyde?

The team considered a list of possible explanations that runs to 13 pages. They wondered if some environmental chemical or dietary change had set the microbe off. They checked if biting insects had transmitted a new infection that interacted with Pasteurella. They considered that Pasteurella might have gone rogue because of an accompanying viral infection, in the same way that Streptococcus bacteria can bloom during a cold, leading to strep throat. “We tested for everything and we couldn’t find anything,” says Eleanor Milner-Gulland from the University of Oxford.

Only one factor fit the bill: climate. The places where the saigas died in May 2015 were extremely warm and humid. In fact, humidity levels were the highest ever seen the region since records began in 1948. The same pattern held for two earlier, and much smaller, die-offs from 1981 and 1988. When the temperature gets really hot, and the air gets really wet, saiga die. Climate is the trigger, Pasteurella is the bullet.

It’s still unclear how heat and humidity turn Pasteurella into a killer, and the team is planning to sequence the bacterium’s genome to find out more. But for now, Kock says the connection makes sense. In laboratory studies, rats that are exposed to Pasteurella are more likely to get infected if you crank the humidity up. The idea of humidity as an environmental trigger also explains how the die-off happened so broadly, suddenly, and totally. It’s something pervasive, which would hit all the saiga at once, and influence the bacteria that they all already harbor.

The saiga might also be uniquely vulnerable to infections of this kind. They live in a world of variable climate and punishing droughts, which is why they’ve evolved their high-octane reproductive cycle. “They have this high-pressure lifestyle and they seem to be particularly prone to disease outbreaks,” says Milner-Gulland.

Then again, mass die-offs are new for them. There are no records of saiga perishing en masse before 1981. “We did a massive historical review, and talked to experts in Kazakhstani folklore and history,” says Milner-Gulland. “If it was normal for huge numbers of animals to die, you’d have thought there would be epic poems about it. The folklore tales mention deaths from harsh winters, but never animals just dropping dead on the steppes.”

As humans continue to change the climate, “this is evidence that some very strange things will happen” says Kock. People normally envision climate change in terms of droughts, floods, and melting ice caps, but disasters can take different forms. The changing climate could, for example, disrupt the normal relationship between an animal and its microbiome—with catastrophic consequences. “This could be the first case of what will be a bigger problem for biodiversity,” says Kock.

Indeed, mass animal die-offs are becoming increasingly common, with the number of such events increasing by around one a year for the last 70 years. Starfish. Swallows. Gazelles. Bats. All around the world, more and more animals are dying in record numbers.

As for the saiga, “what really frustrates me is that there is no magic-bullet solution to avoid these mass die-offs,” says Aline Kühl-Stenzel from the UN Environment Program. “It is out of our control to influence the climatic triggers, and the Pasteurella bacterium is widely spread. So, our conservation strategy is clear: We want large, resilient populations.”

She is hopeful. All the countries across which the saiga roams recently signed a mini-treaty to protect the antelope, and governments and NGOs are working hard to put an end to poaching. “The saiga really is a success story,” says Kühl-Stenzel. “Geographical, political, and cultural barriers have been overcome to bring the best people together to save this unique animal.”

The saiga is the ultimate rebounder—a species that’s adapted to bouncing back from catastrophe. If its population slips too low, a mass die-off might finish it off for good. But as long as a decent baseline persists, so too will the saiga. “All it needs is for us to butt out, to not poach it, to not build railway lines across its territory,” says Milner-Gulland. “Then it does just fine.”