The chatty green parrots had a front-row seat to a spine-chilling show. Tethered to a tree branch not far from their cage, another parrot, similar in appearance but of a different species, was armored in a small leather vest. As the green parrots looked on, a man approached the lone parrot with yet another bird leashed to his arm: a red-tailed hawk. The hawk lunged at the parrot in the vest, wrapping its talons around it. The parrot screamed, a sound only made when death is imminent. Satisfied, the man pulled the hawk off.
This simulated attack—don’t worry, the parrot was unscathed thanks to the vest—was a ruse, aimed at Puerto Rican parrots about to be released into the wild for the first time. A critically endangered species found nowhere else in the world, those chatty green birds had never known life outside of captivity. So they were being taught what to look out for when they headed into hawk territory.
“We wanted an experience that would instill in those birds a very real fear and recognition of a red-tailed hawk as a deadly predator,” says Tom White, a wildlife biologist with the U.S. Fish and Wildlife Service’s Puerto Rican Parrot Recovery Program. In 2001, White helped spearhead this fear-based training regimen for parrots before reintroducing them into the island’s forests. The training involves several other phases before the simulated attack, including “flying” a hawk-shaped cutout over the parrots’ aviary, playing recorded hawk calls, and having a live hawk attack their cage.
For endangered species, even a small boost in numbers can help shift the needle away from extinction. But isolating animals for safe breeding comes with a cost. Captive “insurance populations” of endangered species, like the protected group of Puerto Rican Parrots, can lose their survival edge, both behaviorally and genetically, in just a few generations. Once they become easy prey, they’re less likely to survive if released.
This situation has left some conservationists in an ethical bind. So far, wildlife-reintroduction programs have been hesitant to embrace an approach where live predators possibly injure or kill one animal for the sake of teaching many—even if, as in the safety-vested parrot’s case, the animal that’s attacked isn’t endangered. But White’s results seem promising. And researchers behind these programs know that pampering wildlife in captivity won’t help if the species can no longer make it on its home turf.
Earlier this year, on an island thousands of miles from Puerto Rico, another group trying to bring back a critically endangered bird called White for advice. The group, based primarily on Hawaii’s Big Island, had released five Hawaiian crows—raven-like birds with bristly feathers framing dark beaks—from a captive flock into their native forest, a place the species had not been in nearly 15 years. But within a week, three of the crows had died. Two were killed by their native predator, the Hawaiian Hawk, or ‘io.
Having some deaths in a reintroduction program, especially early on, is not unusual, says Alison Greggor, a researcher with San Diego Zoo Global, one of a few institutions collaborating to reintroduce the crows. Nevertheless, this initiative—known as the the ‘Alalā Project, after the crows’ native name—collected the surviving birds back into captivity, and vowed next time would be different. While the group had attempted antipredator training before releasing the crows, they realized it wasn’t rigorous enough. For these ‘alalā, approximately five generations removed from the wild, to have a chance at survival, the species needed to relearn to outwit its old adversary.
White advised his colleagues in Hawaii to show the ‘alalā that hawks mean business, and recommended a similar approach to the training he employs with the Puerto Rican Parrots. If the ‘alalā were going to watch a non-native bird get attacked, “I told them that you don’t need to worry about a vest,” White says. “Just go ahead and have the real experience.”
But the “real experience”—allowing a hawk to actually kill a bird in front of the ‘alalā, as opposed to merely simulating an attack—might not go over well with everyone. Animal-welfare permits and public perception are concerns, and the ‘alalā is revered as a family guardian in Native Hawaiian culture. Besides, the ‘Alalā Project wanted to take a “data-driven approach” to the training and adapt it to the learning style of crows, which are notoriously perceptive, social birds. To narrow down the best techniques, the group performed pilot trials, observing how the crows reacted to multiple scenarios, including one in which a hawk attacked live prey.
When they settled on what would go into the actual training regimen, the ‘Alalā Project tried to incorporate as many of the elements of a real attack as they thought would be effective. In line with many of White’s techniques, a taxidermied ‘io with wings spread wide was rigged to “fly” across the crows’ aviary while they heard its cries, along with ‘alalā alarm calls. A live ‘io appeared just outside their cage, flapping its wings as if swooping toward them. Finally, they saw and heard the consequences of ignoring those signals: the sound of ‘alalā shrieking in distress and the body of a dead American Crow.
Throughout these experiences, the researchers watched for what they considered appropriate reactions from the ‘alalā: calling in alarm, acting vigilant, flying, or making themselves bigger and more threatening, a behavior called “mobbing.” The approached seemed to work. Even though the crows didn’t observe the gruesomeness of an actual kill, Greggor, who oversaw the training, thinks the birds still got a frightening enough picture of a predator. In fact, during the training, one ‘alalā that had been taken back into captivity after the last release started mobbing the hawk outside their cage, which prompted a second round of training to see if other birds picked up this behavior. That time, three other crows joined in.
In the end, the group didn’t feel that a real kill would be “more effective or more reliable” than their staged attack, Greggor says. Their approach, she notes, “doesn’t mean that that’s now the perfect formula forever moving forward, but this was our attempt at taking an evidence-based approach to designing a training scenario that would be most relevant to the birds.”
Katherine Moseby, a wildlife-reintroduction biologist at Australia’s University of New South Wales who isn’t involved with the ‘Alalā Project, was encouraged to hear that Greggor and her colleagues used a live ‘io in the training. But Moseby remains somewhat skeptical of just how much of an effect simulated encounters without actual attacks can have. “Here is a question, what was the first animal that bit you?” she asks. “I bet you still remember it, even though it likely happened when you were young. Real experiences with real animals that truly scare you are likely to have stronger and longer lasting effects.”
White acknowledges that some people might think his recommendation—in which a live bird is attacked or even killed by a live predator—is cruel or drastic. “But in my opinion, it’s more cruel and more irresponsible to release captive-reared animals that have not had the best training possible,” he says. He stands by this view even if not training some of the released animals at the same time would be a more scientific way of discerning what about that training does and doesn’t work. “If we are then going to expose them to the real world, then we need to prepare them for the real world,” he says.
The ‘Alalā Project may soon have a better idea of how much its training has sunk in. Over two occasions this fall, they released 11 total ‘alalā into Pu‘u Maka‘ala Natural Area Reserve on Hawaii’s Big Island. Still, Moseby points out, “most practitioners train all of the released animals, so it is difficult to isolate the effect of training.” Not training a few of them would mean allowing some crows to encounter their first hungry hawk after flying blithely into the forest for the first time. But with less than 130 ‘alalā left on the planet, the species may not be able to afford that.