Isis the lioness arrived at Chicago's Brookfield Zoo six years ago with a job: to give birth. Brookfield hadn't seen lion cubs in decades.
The zoo carefully planned her courtship with its new male, Zenda. Keepers took Isis off birth control. The community was excited.
But nothing happened.
And it wasn't just at Brookfield. The lioness Asali, at Ohio's Columbus Zoo, wasn't getting pregnant. Nor was Kiki, in Atlanta, or Neka, in Oregon.
At the same time, scientists at the St Louis Zoo were coming to a startling conclusion: dozens of lionesses, including Isis, had been on birth control - but the drug wasn't wearing off.
The implants, meant to last six months to a year, were still in effect three, four, even five years later.
Now researchers across the United States are trying to figure out the extent of the problem, the impact on zoo breeding programs - and whether some animals will ever bear offspring again.
The Wildlife Contraception Centre, an arm of the Association of Zoos & Aquariums housed at the St Louis Zoo, is the distributor of Suprelorin, the drug used on the lionesses, to accredited zoos in North America. The centre has recorded more than 3300 treatments for about 1600 exotic animals, from polar bears to monkeys to meerkats.
But it has documented only 88 animals that have "reversed" - in other words, fallen pregnant or produced sperm - in nearly a decade of treatments.
Of the 200-plus species treated, the centre has seen reversals in just 50. Not one of the six or seven polar bears given Suprelorin, species experts say, has had cubs.
It's too early to draw conclusions, researchers say. Still, the zoo community is unsettled.
"I think we all should be worried," says Budhan Pukazhenthi, a reproductive physiologist at the Smithsonian Conservation Biology Institute at the National Zoo in Washington who works with lions.
"I think we also should use a lot more caution when we make the decision to place an animal on contraception."
Spokesmen for Virbac, the US subsidiary of a French pharmaceutical group that makes Suprelorin, have not returned multiple emails and phone calls on this subject.
The issue is key to the future of North American zoos. Zoos today often say they exist for two primary reasons: to educate the public on wildlife conservation, and to provide a back-up plan should species die out.
To accomplish the first goal, leaders prefer to exhibit animals in wild settings - herds of elephants, bands of gorillas, prides of lions. But to achieve the second, and for the health of their animals, they need to monitor those family groups, and only allow breeding between individuals of the right genetics, at the right time.
The alternative, zoo leaders note, is to kill unwanted animals, as do some European zoos.
Denmark's Copenhagen Zoo recently shot a two-year-old male giraffe, dismembered it in public and fed the remains to its lions - despite an online protest petition with thousands of signatures. The zoo said the giraffe was killed to ensure the breeding program remained healthy.
Contraception, Copenhagen Zoo officials noted, has health repercussions.
At the Association of Zoos & Aquariums 2013 conference, held in Kansas City, contraception centre leaders called a special session on the topic to plead for data from the continent's zookeepers.
To fix the problem, they explained, they first need to understand how serious it is.
Lions were the poster children at the meeting. Keepers had already begun sharing data with the St Louis contraception centre on mating, health and reproduction.
And they had some numbers: 118 lionesses had been treated with Suprelorin, said Hollie Colahan, large mammal curator at the Denver Zoo and co-ordinator of the association's lion species survival plan.
Only nine had documented reversals.
Like most US zoos, Brookfield has long had lions, but they also weren't really a priority 30 years ago, according to Bill Zeigler, Brookfield's senior vice president of collections and animal care. That was because they were not considered endanged.
But then lions in the wild started dying in the thousands - killed by disease or by people, says Luke Hunter, president of the wild cat conservation group Panthera.
Suddenly zoos became more interested in telling the story of lions, Zeigler says.
Brookfield, however, didn't have a female. Only in 2008 was the zoo sent Isis, then aged two, and Zenda, a male the same age, through a species survival plan overseen by the Association of Zoos & Aquariums.
Brookfield introduced the cats slowly. Sex between wild animals isn't a sterile event.
"Lion mating involves teeth and claws and a little bit of blood," says Denver's Colahan. When things don't work out, animals can die.
Brookfield initially put Isis and Zenda in nearby pens, only lifting the gate after four months.
Lions are assigned a number that designates their genetic worth. Isis and Zenda weren't the most genetically valuable at that time, so the association team didn't move them to the front of the line. Besides, Isis was still a little young.
So, in 2009, Brookfield put Isis on one dose of Suprelorin.
Suprelorin was a godsend at first because an earlier contraceptive, Melengestrol acetate (or MGA), had been linked to uterine overgrowth, lesions, sterility and, sometimes, cancer, in lionesses.
Then Cheryl Asa, director of the contraception centre and of research at the St Louis Zoo, stumbled onto Suprelorin, developed in Australia for domestic dogs by a company called PepTech Animal Health, later purchased by Virbac. It was - and still is - unapproved for use in the US.
But Asa helped swing a deal with the US Food and Drug Administration: the contraception centre would import the drug for investigational use and carefully track all treatments. (An FDA spokeswoman declined to comment, saying federal regulations bar the department from even acknowledging the existence of such a study.)
The zoo community embraced the contraceptive. Early studies - some by PepTech itself - indicated Suprelorin was safe, effective and promised to be reversible. For several years, all went well.
"We were focused on health," says Asa. "And it hadn't occurred to me that it wouldn't be reversible."
Keepers were perhaps the first to realise something was wrong. Zoo Atlanta, which had given a female lion called Kiki one dose of Suprelorin, wondered if she would ever get pregnant again when she failed to come back on heat a year later.
In 2012, the zoo association's Population Management Centre analysed the status of zoo lions. In general, the study found, things were fine. Still, it warned, "If the females on contraceptive do not breed again, the population will experience a much more severe decline."
No matter how this ends, some zoo staffers say they'll be more cautious in future.
"I think it was sort of recommended across the board without really knowing what the long-term consequences were going to be," says Rebecca Snyder, Zoo Atlanta's curator of mammals. "I think we all learnt a lesson from that."
