The jaguars don’t know it yet, but soon they’ll be free to roam Argentina’s Iberá wetlands, becoming the first apex predators to do so in a century. Three adults and the two-year-olds Amarí and Mbareté currently live in an enclosure on San Alonso, a patch of high ground in Iberá’s patchwork of flooded wilderness, lagoons, jungle, and grasslands. Once a pristine habitat, the region has lost much of its wildlife since the early 20th century, when ranchers moved into the area. People killed off native predators to protect livestock, and many species were decimated to satisfy a burgeoning market for fur, leather, and feathers.

Now, wildlife is making a comeback in Iberá, thanks to an ecological restoration effort spearheaded by the nonprofit Rewilding Argentina Foundation. Because of the region’s now-protected status and the reintroduction of locally extirpated species to reconstruct ecological communities, the area...

To Emiliano Donadio, the foundation’s scientific director, the release not only is crucial to rebuilding Iberá’s ancient ecosystem, but is a scientific experiment that will provide a rare glimpse of how the return of one of the world’s largest carnivores could transform an ecosystem. Scientists know that losing large predators can have far-reaching, disruptive effects on ecosystems through cascading forces that reverberate from predators at high trophic levels—the top of the food web—to their prey and beyond, even sculpting the abundance and structure of plant life. The loss of jaguars (Panthera onca), pumas (cervu concolor), and other predators from fragments of the Venezuelan rainforest after the construction of a hydro-electric dam, for instance, is thought to be a key factor in triggering an explosive proliferation of herbivores including monkeys, which ravaged the vegetation and caused what researchers described as an “ecological meltdown.” But seldom do ecologists get to investigate whether those negative effects can be reversed by restoring predators. “We think that an ecosystem devoid of predators will be in better shape when the predators come back,” Donadio says, but that supposition remains largely untested. 

RETURN OF THE JAGUAR: Key to restoring the second-largest wetland in the world is the reintroduction of its ancient predator. To that end, a jaguar breeding program is underway. The resulting cubs will be released along with three adults into the wild.

Part of the problem is, while it’s well accepted that large carnivores play vital ecological roles, just how they shape ecosystems through cascading effects in different environments isn’t well understood. In addition, predator reintroductions such as the ongoing project in Iberá are extremely rare and have lagged greatly behind herbivore rewilding projects, in large part because releasing animals capable of killing people and livestock is so controversial. “We only have so many natural experiments,” notes wildlife ecologist Justine Smith of the University of California, Davis.

In the absence of firm evidence, conservationists have been eager to interpret early predator reintroduction studies—largely based on the purported regenerative ecological effects of returning gray wolves (Canis lupus) to Yellowstone National Park in the mid-1990s—as a rationale for bringing predators back to many parts of the globe. In Colorado, for instance, conservation organizations have been using such findings to push for the approval of a bill on the November ballot that would effectively mandate wolf reintroduction in the state to restore the ecosystem’s “natural balance.” But some ecologists caution that the ecological outcomes of such projects are unclear. 

See “Where the Wild Things Were

In search of answers, scientists are employing novel approaches to study the ecological roles of large carnivores, from the African savannah and the Andean plateau to the ocean, and to understand how ecosystems change as they are lost or reintroduced. What they’re finding is that predators have powerful, yet nuanced and complex effects that ripple through food webs in what are known as trophic cascades—effects that depend not only on the nature of the hunter itself, but also on characteristics of its prey and the habitat the animals share. 

“There’s still good reason to believe that trophic cascades will . . . occur in many systems,” Smith says. “It’s just that we don’t really have all the data yet to understand exactly when, where, and why.”

A green new world 

Modern predator ecology began in principle with a simple question: “Why is the world green?” In the late 1950s, when ecologists Nelson Hairston, Frederick Smith, and Lawrence Slobodkin were pondering this question, the prevailing notion was that the abundances of animals that inhabit ecosystems depend solely on the amount of plants and nutrients at the bottom of the food web. Herbivore numbers were determined by the abundance of vegetation, and predator numbers by herbivore abundance. But the three University of Michigan scientists suggested in a 1960 paper in The American Naturalist that herbivore numbers weren’t controlled by the availability of plants alone, because predators were also playing a key role, and that by managing herbivore populations, predators indirectly protected vegetation. This idea was quickly met with fierce criticism, and to be fair, there wasn’t any robust evidence for their hypothesis at the time. 

That was a gap that one of Smith’s students, Robert Paine, was determined to fill. In the early 1960s, Paine regularly ventured out to Washington State’s Makah Bay and its vibrant community of barnacles, mussels, limpets, and the predatory starfish Pisaster ochraceus to conduct an experiment that would become one of ecology’s most famous. From a 25-foot-long stretch of rock, he pried off the starfish and flung them into the ocean, while leaving another stretch unaltered. Over the course of a year, Paine noticed that life on the starfish-less rock transformed. The starfish’s main prey, the barnacle Balanus glandula, began to take over, followed by fast-growing mussel species, crowding out other organisms. The intertidal community of 15 invertebrate species dwindled to 8. Paine concluded that by keeping its prey in check, the Pisaster starfish was helping to maintain biodiversity. Despite being small in number, the five-legged predator was performing a crucial ecological function that earned its recognition as the first ever “keystone species.” 

Predators are so important, their removal has such long-lasting effects, that it’s naive to think that you can quickly reverse the effects of their absence by restoration.

—Tom Hobbs, Colorado State University

Paine’s belief that predators are ecologically important, and his recognition that their roles are best understood by studying animal communities that have been perturbed, inspired ecologist Jim Estes, then a PhD student at the University of Arizona observing the sea otters (Enhydra lutris) inhabiting the lush kelp forests around the Alaskan island of Amchitka. After a brief meeting with Paine in 1971, Estes decided to visit nearby islands where otters had been wiped out during the fur trade and immediately noticed a difference. “When I looked down at the seafloor, I was stunned by the vast numbers of urchins and absence of kelp,” Estes, who is now at the University of California, Santa Cruz, wrote in a 2016 memoir about his work. Like Pisaster starfish, otters around Amchitka were playing a key ecological role, he observed: by keeping urchins in check, the marine mammals limited the invertebrates’ consumption of kelp, thereby regulating the abundance of life-giving plants at the bottom of the food web. In a 1979 lecture, Paine coined the term trophic cascades to describe these indirect effects. 

Estes’s observations were quickly followed by a series of observations from other researchers that pointed to the existence of trophic cascades in other aquatic ecosystems, such as those in lakes and rivers. In the 1990s, Os Schmitz, an ecologist at Yale School of the Environment, discovered that a terrestrial predator, the nursery web spider Pisaurina mira, could similarly create a trophic cascade, although in this case he uncovered an entirely different mechanism. It turns out that the spiders didn’t have to kill their prey to affect the ecosystem; they just had to scare them into skipping a meal.

Schmitz placed leaf-chewing grasshoppers, specifically Melanoplus femurrubrum, in a small, grass- and herb-filled enclosure, and observed that the vegetation flourished after he added nursery web hunting spiders, even without any immediate change in grasshopper numbers. Then, in a series of experiments in which Schmitz glued the spiders’ mouthparts shut so they could still instill fear but not consume their prey, he demonstrated that their mere presence was enough to allow grasses to flourish, as grasshoppers would forgo a meal to avoid becoming one. Biologists now call such fear-driven effects behaviorally mediated trophic cascades, distinguishing them from density-mediated ones that involve the predators’ consumption of their prey. 

See “Opinion: What Animals Can Teach Us About Fear

These behavioral effects come at a long-term cost to prey that can ultimately lead to declines in population densities. Avoiding the spiders caused some of Schmitz’s grasshoppers to starve, and their populations eventually dwindled. In a menagerie of animals, the sheer fear of predation can affect prey species’ metabolism, stress hormone levels, neurophysiology, and reproduction. In the forests of the southern Gulf Islands off the coast of British Columbia, Canada, for instance, Western University ecologist Liana Zanette and her colleagues have shown that hearing the sounds of predatory birds and raccoons throughout a breeding season can cause song sparrows (Melospiza melodia) to reduce the number of young they produce by 40 percent. “The role of predators is way greater than we ever previously imagined,” she says. 

What remains unclear is how common such trophic cascades are in nature, Estes notes. “What I think is probably one of the biggest outstanding questions in this part of ecology [is] just how repeatable, how generalizable, these phenomena are.”

The gray wolf: An American hero? 

A major challenge in studying the ecological role of large terrestrial carnivores is that the vast majority of them are in decline. Much knowledge of their influence comes from studies of local extinctions. One remarkably common effect is that smaller predators—which are often held in check by apex predators—take over and wreak havoc. As coyote populations (Canis latrans) declined in Southern California with the increasing expansion of human settlements in the 1990s, for example, the number of raccoons, skunks, and cats decimated reptile, bird, and small mammal populations.

Herbivores can also experience population booms when a predator is lost. In Yellowstone National Park, after federal officials had killed off most of the park’s wolves by 1926 as part of a national wolf control program, elk numbers surged, with particularly destructive effects on woody vegetation around the park’s streams. Then, in the winters of 1994, ’95, and ’96, Yellowstone officials released 41 wolves into the park’s interior and northern range. It was one of the first and remains one of the few intentional reintroductions of a large carnivore to part of its historical distribution, and researchers were eager to study the effects. 

While it’s well accepted that large carnivores play vital ecological roles, just how they shape ecosystems isn’t well understood.

Six years later, ecologists William Ripple and Robert Beschta of Oregon State University and others reported that aspen trees were taller and elk droppings less abundant in streamside areas frequented by wolves compared with places where wolves were seldom seen. The researchers have since gathered a wealth of correlative evidence that the wolves created a “landscape of fear” for elk (Cervus canadensis) that prevented them from browsing the foliage along the park’s northern creeks and rivers. This supported the growth of aspen, cottonwood, and willow on their banks, and even improved the structure and function of particular waterways, Beschta and Ripple proposed last year.

In no small part due to an immensely popular 2014 YouTube video narrated by The Guardian columnist George Monbiot titled “How Wolves Change Rivers,” Yellowstone’s canine predators became internationally famous for single-handedly repairing Yellowstone’s broken landscape. Yet the video’s narrative conceals a stark disagreement among ecologists about the relative contribution of wolves to the decline in elk, as well as the role of fear in the ecological changes observed. 

Utah State University ecologist Dan MacNulty questions whether adult elk, which are large and thus difficult for wolves to kill, would be so afraid of wolves as to miss out on a good meal, he says. Indeed, he and his colleagues have tracked wolves and elk with radio collars and found that elk often don’t avoid areas frequented by the predators, and that the ungulates seem to be more concerned with avoiding cougars (another name for pumas, Felis concolor). To Schmitz, this makes sense, given pumas’ sit-and-wait hunting strategy. As he’s learned from comparing spiders with different hunting styles, predators that tend to ambush their prey are more likely to create behaviorally mediated cascades. The wolves at Yellowstone typically hunt by chasing elk across the landscape. Because the elk can often see them coming from a distance, Schmitz explains, there’s no point avoiding certain areas. 

See “Slideshow: How Ecologists Study the World’s Apex Predators” 

Ecologist Tom Hobbs of Colorado State University says he doubts that the reintroduction of wolves was enough to repair some of the damages caused by wiping them out decades ago. In their absence, an expanding elk population decimated streamside willows (Salix spp.), prime dam-building material for beavers. With fewer dams, the streams flowed faster, cutting deeper into the ground. That caused the water table to drop, making it harder for willows to drink and grow, experiments by Hobbs and colleagues in the park’s northern range have shown. Reintroducing wolves alone is unlikely to reverse such changes to the physical landscape anytime soon, he says. “Predators are so important, their removal has such long-lasting effects, that it’s naive to think that you can quickly reverse the effects of their absence by restoration.” 

Doug Smith, a senior wildlife biologist in Yellowstone who has collaborated with all three research teams, says there’s an element of truth to the trophic cascade, although the effect was probably more density-mediated than fear-driven. In addition, the cause of Yellowstone’s elk declines wasn’t just wolves but rather a suite of factors, including the fact that cougars and bears increased in abundance around the same time that the wolves were introduced. And while collectively these predators helped to regenerate parts of Yellowstone, Smith agrees with Hobbs that the park is “not restored to what it [once] was. That might never happen.” 


In the mid-1990s, officials at Yellowstone National Park released gray wolves from areas in Canada and Montana into the park; it had been more than half a century since the predators last roamed the park. Researchers tracking the revolutionary experiment published results that they say point to the reintroduction’s role in revitalizing the once-degraded ecosystem, with the wolves’ predatory behavior indirectly supporting the growth of vegetation and even improving the health of the park’s waterways. But a heated debate rages on concerning the effects the wolves had on their environment, especially relative to roles of other members of Yellowstone’s rich carnivore community.

Yellowstone, Wyoming, US
41 gray wolves reintroduced in the mid-1990s

Following wolf reintroduction, elk numbers dropped dramatically—from nearly 20,000 in 1994 to just 8,300 in 2000—but wolves are likely not the only carnivore that contributed to that decline; black bear, grizzly, and cougar populations surged around the time of the wolf reintroduction. 

Ecologists first posited that by keeping elk away from streams, wolves were indirectly allowing aspen trees to flourish. Although the trees increased in height in certain areas, their overall abundance changed little.

Although they suffer less browsing pressure since the predator increases in the 1990s, willows have not fully recovered, according to some researchers. And without willows, the recovery of streams has been limited.

Initial studies proposed that elk stay away from streamside areas where they could more easily fall prey to wolves, but new research suggests that elk only avoid these regions in the morning and at dusk. The herbivores also appear to have altered their behavior to avoid cougar-patrolled forested areas at night. 

See full infographic: WEB | PDF

Beyond Yellowstone

Outside Yellowstone, Iberá will soon be one of a few locations where a predator has been intentionally reintroduced and its ecological effects intensively studied. This time, researchers hope to overcome some of the challenges in deciphering cause and effect by collecting baseline data before the predators arrive. In several locations within Iberá’s 1.3 million–hectare protected area, biologists with specialties in entomology, ornithology, predator ecology, and animal behavior are busy characterizing various facets of the ecosystem that they suspect the jaguars might influence. 

Populations of oversized rodents called capybaras (Hydrochoerus hydrochaeris) might plummet after the predators’ arrival, and their behavior may radically change, PhD student Belén Avila of Argentina’s Institute of Subtropical Biology hypothesizes. Right now, the capybaras are acting fearlessly, Donadio says, even dozing on the paths cutting through the area. But once they realize there are killers lurking about, individuals are likely to become more cautious and vigilant, which means they’ll spend less time eating, possibly affecting grass abundance. The jaguars could also reduce the number of smaller predators such as pampas and crab-eating foxes, which are abundant at the moment, and in doing so protect the endangered birds that the foxes sometimes eat. As the researchers track these and other outcomes over the coming years, Donadio says, “it’s going to provide really, really good information when it comes to the importance of large predators on landscapes and biodiversity.” 


The release of predators into the wild is controversial, and for years Yellowstone remained one of the only sites of such a bold reintroduction. But now, similar efforts are in the works around the world. In the Iberá wetlands of Argentina, for example, conservationists will soon release their first jaguars, and hypothesize that a variety of ecological changes will occur across the landscape. 

Iberá, Argentina
Five jaguars to be reintroduced in 2020/2021

Researchers expect numbers of deer and capybaras, the jaguars’ prey, to drop once the cats are reintroduced. By reconfiguring capybara population structure, the jaguars might reduce the spread of mange between the rodents. 

Ecologists suspect that the capybaras’ behavior will change in response to the presence of the jaguars, becoming more vigilant and cautious. If the capybaras spend more time watching for predators than eating, that could allow grassy vegeta-tion to flourish in certain areas.

Jaguars could kill or change the behavior of local foxes and in doing so indirectly boost numbers of the endangered birds that the foxes are known to eat.

Ecologists hypothesize that the jaguars’ presence could increase the abundance of certain scavenger species such as vultures and enhance the diversity of beetles that live off carcasses left behind by the predators.

See full infographic: WEB | PDF

Even without reintroducing predators into the wild, researchers elsewhere are using experimental approaches to detect trophic cascades already in action. In 2008, in the rugged Andean terrain of San Guillermo National Park in Argentina, Donadio wanted to understand how pumas influenced their prey, llama relatives known as vicuñas (Vicugna vicugna). He noticed that, in open grasslands where they’d easily see a predator approaching, the vicuñas’ heads were usually buried in the grass eating, only occasionally popping up to look around. In meadows with taller grasses and canyon areas where pumas could lurk behind rocky outcroppings, on the other hand, the vicuñas spent less time eating and more time on watch. To test the effects of these behavioral differences on vegetation, Donadio constructed a number of 20-meter-by-20-meter exclosures—fenced areas intended to keep the vicuñas out, though the herbivores (as well as the pumas) could still frequent the general area. Sure enough, he observed that the growth of grass inside the exclosures in the grasslands shot up compared with grass in surrounding control plots, while grass growth in the canyon and meadow exclosures did not, suggesting that vicuñas were indeed sacrificing grazing opportunities there to avoid an ambush. 

Understanding ecosystems in their entirety, and the ability of predators to restore them, will require understanding people.

This behaviorally mediated cascade is created by the complexity of the animals’ habitat, Donadio says, and in turn, it helps shape the environment. If the pumas weren’t there, “the vegetation in the canyons [and meadows] would look exactly like the vegetation in the plains.” By enhancing the diversity of habitats in San Guillermo, pumas may be creating new niches for other species, he explains, and in doing so, enhancing biodiversity. 

Princeton University ecologist Robert Pringle and his colleagues have also used exclosures in their search for trophic cascades in Kenya’s Laikipia Plateau, which was recolonized by African wild dogs (Lycaon pictus) in the 1990s. The researchers didn’t find any difference in grazing pressure exerted on local plants by the dogs’ prey—small antelopes known as dik-diks (Madoqua guentheri)—before and after the arrival of wild dogs, even though the antelopes declined in abundance. Pringle suggests that, among other factors, perhaps the timeframe of the study wasn’t long enough to detect a trophic cascade, given that trees take a long time to grow. Pringle did, however, find evidence of cascading interactions between other species in the same ecosystem. In clearings where impala (Aepyceros melampus) gather to avoid being ambushed by leopards (Panthera pardus), the researchers noticed an abundance of acacia trees (Acacia etbaica), which carry thorns to defend against intense herbivory. In other areas where the bush was thicker—areas less frequented by the impala, probably due to fear of predation—“there were a bunch of plants that actually tend to be much more palatable,” Pringle says. 

Similar approaches have also revealed such behavior-driven dynamics in marine ecosystems. In seagrass meadows in Shark Bay in Western Australia, for example, researchers have found that setting up small exclosures in areas frequently patrolled by tiger sharks (Galeocerdo cuvier)—risky areas for grazers such as dugongs (Dugong dugon) and sea turtles (Chelonia mydas)—doesn’t have much of an effect, because the grazers were already kept away from these areas by the sharks. In low-risk areas, on the other hand, when “you put the cages out, it goes from this very sparse seagrass to just a giant salad bowl” of marine vegetation, says Florida International University marine ecologist Mike Heithaus. “At its core, oceans and land are the same because animals are making the same kind of trade-offs—they’re trying to optimize food versus risk.” 

Broadcasting predator vocalizations is another approach to interrogate predator trophic cascades, one that can more directly assess the effects of fear on prey. Upon playing recordings of barking dogs in the intertidal zone on Canada’s west coast beaches, for instance, Zanette’s team found that raccoons (Procyon lotor) spent less time foraging, allowing crabs and fishes to increase in abundance—a behaviorally mediated trophic cascade that extended from a terrestrial habitat into the ocean. Similar studies have reported that underwater playbacks of mammal-eating killer whales (Orcinus orca), but not of local fish-eating killer whales, trigger harbor seals (Phoca vitulina) to dive to safer depths.

In Mozambique’s Gorongosa National Park, researchers have sought to understand the historic impacts of losing nearly an entire guild of top carnivores. Decades ago, the area had become a battlefield during the country’s civil war, decimating populations of large animals, including leopards, wild dogs, and lions. In 2019, Pringle, Princeton postdoc Justine Atkins, and their colleagues reported that in the absence of these carnivores, the typically forest-dwelling bushbuck (Tragelaphus sylvaticus) had spilled out into the open floodplains, where they appear to be suppressing certain plant species. But when Pringle’s team played recordings of leopards, the bushbucks retreated into the forest. Pringle is now studying how the recent return of wild dogs to the park will reconfigure the ecosystem, and Gorongosa officials just obtained permission to bring in several leopards, says Paola Bouley, who directs the park’s carnivore introduction program. “I think Gorongosa is going to be a classic case of rewilding.”


In Mozambique’s Gorongosa National Park, where a reintroduction project involving wild dogs and leopards is ongoing, a study is underway using state-of-the-art tools to track the predators’ ecological effects.
Gorongosa National Park, Mozambique
Dozens of African wild dogs introduced in 2018/2019; leopard reintroductions slated for 2021

Scientists plan to use high-resolution satellite imagery—including LiDAR, which captures the height of trees and grasses—to assess changes in vegetation structure.

Researchers will be analyzing antelope scat using fecal DNA barcoding, alongside studies of the predators’ feces, to track the animals’ diets. 

Ecologists are using GPS collars and camera-trap data to monitor the behavior of different antelope species. 

See full infographic: WEB | PDF

Learning from rewilding

With these experiments only just now getting underway, it’s not clear how exactly such rewilding projects will play out. MacNulty is unconvinced that Pringle’s data on bushbuck behavior are strong enough to demonstrate a trophic cascade in Gorongosa. “Evidence [for such phenomena] can be hard to produce in some of these complicated systems,” he says. Moreover, the effects that predators have on ecosystems can be unpredictable. For example, in another study of Pringle’s in which he released predatory curly-tailed lizards (Leiocephalus carinatus) onto Bahamian islands, the reptilian predators appeared to intensify competition between prey species and limit their ability to coexist, rather than maintain biodiversity.

Without a clearer understanding of the species interactions at play, the ecological benefits of rewilding predators remain unresolved. While several researchers argue that restoring an environment’s lost predatory dynamics will generally be a positive thing, others caution that it won’t be a “quick fix” for degraded ecosystems. And “when it has been in this new predator-free equilibrium for a long time,” says marine ecologist Boris Worm of Dalhousie University in Canada, “we really don’t know what the outcome will be once we add predators back in.”  

ECOSYSTEM EFFECTS: The Iberá wetlands in Argentina lost most of their wildlife during the 19th century as people moved into the remote area, but the region’s native species, including capybaras (top), maned wolves (bottom left), and pampas deer (bottom right), are now recovering, thanks in part to an ambitious ecological restoration effort. 

Outcomes of rewilding are even more uncertain in areas that are populated by people, not least because predators are often aggressively persecuted in those areas. Large carnivores can also be fearful of humans. In California, for instance, Zanette and Justine Smith have observed that cougars stop eating and even abandon entire kills when they hear the sound of people talking. Such findings raise the question of whether predators can influence their environment through trophic cascades in landscapes dominated by people, remarks University of Washington wildlife biologist Laura Prugh. 

This problem also preoccupied Paine of the famous starfish experiments toward the end of his career. In an opinion article he wrote with Worm that was coincidentally published online the day Paine died, June 13, 2016, the two argued that humans are such pervasive ecological influencers that they should be considered a “hyperkeystone” species, with an even bigger influence than apex carnivores. Rather than treating humans as external to the natural world, ecologists should integrate studies of human behavior into food-web theory, they argued. In other words, understanding ecosystems in their entirety, and the ability of predators to restore them, will require understanding people. “It’s not just about other species,” Worm says. “It’s first and foremost about ourselves.”

Correction (November 27, 2020): A previous version of this story misidentified pumas as Felis concolor and the elk species described as Cervus elaphus. The article has been updated to reflect the fact that pumas are in fact classified as Puma concolor, and the elk as Cervus canadensisThe Scientist regrets the errors.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?