A Lesson for Ravens: Don’t Eat the Tortoises

Can fake tortoise shells teach predators to stop devouring soft-shelled juveniles?

The remains of a desert tortoise near Joshua Tree National Park, in California's southern desert.
Shelie Puffer / AP

Tim Shields didn’t see any young tortoises himself. For the most part, the only sign of them was their shells, desiccated and punctured, scattered around the landscape and piled under the occasional Joshua tree. He was working on a long-term monitoring project in the Mojave and Sonoran Deserts, where over the past few decades juvenile tortoises had all but disappeared from the study areas. This particular year, the scientists and volunteers spotted a single live juvenile tortoise—in the beak of a raven, flailing its legs as it was carried away.

That was eight years ago, and it marked a turning point for Shields, he told me, sitting in his kitchen in Joshua Tree, California. While desert tortoises face a range of threats, for juveniles whose shells are still soft, raven predation is perhaps the biggest concern. Shields knew that past efforts to control the canny birds by shooting or poisoning them had faced legal challenges from animal activists; many scientists think, as well, that killing enough ravens to protect tortoises simply isn’t possible. So Shields went in search of an alternative.

After talking with other biologists, engineers, and even some rocket scientists—with anyone, really—he concluded that he needed to teach ravens a new way of life. Working with a diverse cadre of collaborators, including the noted raven-biologist William Boarman, the design software company Autodesk, and a science teacher from Shields’s hometown, his company, Hardshell Labs, created the “techno-tort”—an educational tool for ravens designed to get across one lesson: They ought not to eat tortoises.

Shields introduced me to the techno-tort in the front yard of his house. Originally inspired by basic styrofoam tortoises that Boarman made in the 1990s, the techno-tort is now uncannily lifelike. Lying in the sand, the 3-D-printed shell had a similar solidity and texture to the real thing but was considerably lighter. This one had been color-printed, but earlier prototypes were hand-painted—and subsequent shells have been too, while the company searches for a printing pigment that will last under the desert sun.

In the field, techno-torts will be fitted with accelerometers and methyl anthranilate, a nontoxic bird deterrent. When a raven disturbs the shell, it will let out an explosive spray with a noxious taste and odor. A more low-tech version will involve packing shells full of meat treated with another nontoxic substance that will briefly nauseate the raven. The aim is to effectively communicate, through the experience of fright or queasiness, a simple message: Stay away.

Until recently, the main approaches available to conservationists for dealing with difficult wildlife have been limited—basically fencing them out or killing them. But now pedagogic alternatives are beginning to emerge, raising the question of whether fractious relationships might be modified by a little interspecies education.

Conservationists have tried to reeducate wildlife in the past. A range of animals, from stout wallabies to the giant lizards of the Canary Islands, has undergone training on avoiding predators, with mixed results. There have been efforts to teach golden lion tamarins to eschew dangerous foods, and whooping cranes to migrate. But these projects had only small, usually captive classes of students. In recent years, behavioral interventions have been significantly scaled up.

In the Kimberley region of Western Australia, for instance, researchers have launched the world’s largest cane-toad-mitigation effort, which depends on teaching wild animals about the dangers of these toxic toads. Introduced in 1935, cane toads have caused severe population declines among many species of larger predators, from freshwater crocodiles and giant monitor lizards to the critically endangered northern quoll, a small but fierce nocturnal marsupial. To these animals, accustomed to eating local amphibians without consequence, cane toads look like a tempting meal.

Many of these predators might survive an encounter with a smaller cane toad, but the toads at the front of this advancing wave are large, highly toxic animals. So Rick Shine of Macquarie University in Sydney and his collaborators have sought to create their own advance guard. In a study led by Samantha J. Price-Rees, they introduced bluetongue lizards to nauseating sausages made of minced cane toad; in another, led by Georgia Ward-Fear, they exposed monitor lizards to juvenile, less toxic “teacher toads.” Both of these animals proved to be quick studies and began to avoid adult toads. Shine, along with a group of research and conservation partners known as the Cane Toad Coalition, is currently rolling out a multiyear project to drop teacher toads and sausages ahead of the main toad wave.

The cane toad project is meant to protect predators from their own appetites, but other new initiatives, such as the techno-torts, are trying to save vulnerable prey by reeducating the animals that eat them. On the South Island of New Zealand, a massive campaign is under way to encourage more recently arrived predators—hedgehogs, rats, cats, and ferrets—to ignore the eggs and chicks of banded dotterels, wrybills, and other birds. In two study areas totaling 1,800 hectares, plus control sites, a team of ecologists has been smearing Vaseline infused with bird scents around the landscape. This system of “chemical camouflage,” developed by Catherine Price and Peter Banks of the University of Sydney (where I work as well), starts with spreading the scent of birds for a month or so prior to the breeding season; predators learn to ignore this olfactory cue as an unrewarding distraction. The trial results are still being analyzed, but previous efforts with other species and on smaller scales produced remarkable results. In an earlier study, the nests in treated areas had a 62 percent greater survival rate than the controls.

What unites all of these projects is a curiosity about how animals learn. But teaching animals means that researchers have to do their own deep learning on the ways different animals think and behave. In Australia, researchers have found that while bluetongue lizards can be effectively trained with cane toad sausages or teacher toads, monitor lizards require the latter to retain their aversion. In New Zealand, researchers need to worry about how different predators generalize: What combination of bird scents will teach a predator to ignore all avian species? This kind of question is part of a growing appreciation of the complexity of animals’ cognitive abilities. As Shine told me, “We can attempt to use those abilities to change the outcomes of encounters between potential predators and prey in a way that probably we weren’t thinking about a decade ago.”

In Joshua Tree, Shields and Boarman showed me footage from cameras they placed out in the desert with prototype techno-torts. In one memorable clip, an adult raven confidently approaches the techno-tort and, without hesitation, flips it over and pounds down on the underside of the shell. Was she curious about a strange object in the environment? Was she wondering where this bad replica of a tortoise shell had come from? Or did she believe that she had come across the familiar form of a tortoise? From the clip, the last possibility seems likely.

Shields and Boarman are passionately interested in understanding how ravens make sense of techno-torts. They need to be. For this approach to have any chance of success, the ravens need to mistake these shells for the real thing. Otherwise they won’t be learning to avoid tortoises, just techno-torts. In other trials, the researchers are changing the shape and color of the shells to see how these tweaks influence raven interactions. Which of these cues really matter?

As I drove away from Shields’s house that afternoon, I saw a group of ravens eating roadkill. This readily available food source is just one of the many transformations to the desert that have increased raven numbers. In recent decades, human settlement has provided a steady supply of food, water, and nesting places. These changes have also destroyed tortoises’ habitat, which is now crisscrossed by roads, overtaken by human uses, and modified by introduced plants and cattle. While ravens are a worrying threat to young tortoises, they are far from the only danger.

The techno-tort project, then, might only sidestep a larger problem. This criticism is leveled at many of these “educational” projects. After all, even if they succeed, abundant numbers of ravens, cane toads, and rats will still be left roaming around. And Shields and Boarman are pretty sure that whatever they develop will work for only a limited time with crafty ravens. Alongside the techno-torts, they are working on a suite of other tools, including modified lasers and drones, to both haze and educate these birds. In the end, Shields explained to me, it’s going to be a bit like a chess match, an ongoing effort to adapt to ravens’ own prodigious capacities for learning and experimentation.

These conservationists don’t see educational approaches as a cure-all. Instead, their projects are often meant to buy time. Boarman was explicit about this: All that mitigating the impact of ravens can do, he told me, is produce a whole lot of tortoise teenagers, which will still need a desert that can support them. Shields agreed: The aim of the techno-tort, he said, is simply to “stop the hemorrhaging.” But, he went on to say, until we can modify human behavior in the desert, “we need to buy time, and we can buy time by altering raven behavior. That’s the hope.”