Wildlife managers in the greater Yellowstone ecosystem have documented a previously underestimated pattern of natural mortality among adult grizzly bears. Research indicates that wolves, mountain lions, and other grizzlies are killing adult bears more often than the models used to manage the population had accounted for, and a small number of deaths remain unexplained.
The grizzly population in Yellowstone is one of the most closely monitored large mammal populations in the world. After the species was listed under the Endangered Species Act in 1975, when the population was estimated at roughly 136 animals, intensive management and habitat conservation brought the number back to an estimated 700 to 1,000 bears today. Individual bears are tracked through ear tags, GPS collars, den cameras, and mark-recapture surveys.
For years, the documented causes of grizzly mortality fell into familiar categories. Vehicle and train strikes accounted for a steady share of annual deaths, particularly along transportation corridors. Management removals occurred when bears came into conflict with livestock or threatened human safety. Adult males killing cubs or younger bears was a known but secondary factor. Human-caused mortality was the primary driver in the models that informed recovery decisions.
The anomaly that drew researchers’ attention was a cluster of adult deaths in specific areas that did not match those categories. The events were initially attributed to statistical noise, but they persisted long enough to warrant investigation. The signal appeared and intensified during the 2000s and 2010s, when the population reached the higher end of its trajectory and densities increased.
A grizzly bear’s position in the ecosystem is more complex than the phrase “apex predator” suggests. Yellowstone grizzlies are omnivores with a diet that shifts across seasons. They feed on whitebark pine seeds, army cutworm moths, cutthroat trout, and ungulates such as elk, bison, and deer through a combination of hunting and scavenging. They regularly take elk calves in spring and can kill adult ungulates in winter when those animals are compromised by weather or injury. They also steal kills from wolf packs, a behavior that reinforced the perception that grizzlies were dominant over wolves in direct encounters.
The prevailing assumption was that a healthy adult grizzly, particularly a large male, faced no natural predators. That assumption did not account for the bear’s specific physical vulnerabilities. Grizzlies are formidable when alert but less so in terrain that limits their ability to confront an attacker directly. They are also physically compromised at two predictable points in the year: when they emerge from hibernation in reduced condition, and in late fall when the demands of building winter fat reserves limit time for vigilance.
The first clear evidence that something was wrong came from subtle patterns in the data. GPS collars on adult bears, especially females with cubs, transmitted mortality signals in areas where terrain made carcass investigation difficult. When investigators could reach the sites, the wound evidence did not always match the standard signatures for vehicle strikes, human-caused deaths, or fights with other grizzlies.
Trail cameras in the affected zones captured behavioral changes as well. Adult bears, including males, showed increased vigilance in specific areas, avoided terrain features they previously used routinely, and shortened their foraging bouts. Movement data showed that surviving bears were systematically avoiding ridgelines, forested saddles, and areas with limited sightlines within the mortality clusters. The behavior was consistent with animals responding to a perceived predation risk, even though the standard ecological model said they had no reason to do so.
Scat analysis added another dimension. The pattern of scent marking and rubbing behavior in the anomalous zones changed during the years when deaths were clustered there. Researchers described the change as consistent with the pattern seen when a significant predator or competitor enters an area and resident animals recalibrate their assessment of risk.
The first confirmed cases of wolf predation on adult grizzlies were documented through a combination of GPS collar data, trail camera evidence, and carcass investigation. A single wolf or a small group cannot successfully prey on a healthy adult grizzly. The size difference is too large and the bear’s defensive capability is too formidable. What changes the calculus is pack size and coordinated attack behavior.
Wolf packs of eight or more animals can, under the right circumstances, coordinate attacks that exploit the bear’s limitations rather than confronting its strengths. The bear’s defensive strengths are its size, forelimb power, and thick hide. Its limitations include speed over distance, reduced maneuverability compared to canids, and vulnerability to simultaneous attacks from multiple directions. A coordinated attack engages the bear from several directions at once, dividing its defensive response.
The bears killed by wolves in documented cases were almost universally taken in terrain where escape was limited, such as areas that channeled movement or restricted refuge. Early spring, when bears emerge from hibernation in reduced condition, was a period of particular vulnerability. The packs involved in confirmed kills tended to be large and to hold territories where grizzly denning habitat overlapped wolf core territory.
The confirmed cases are likely a fraction of the actual events. Mortality investigations require finding a carcass, accessing it before scavengers disperse the evidence, and attributing cause of death from what remains. Wolf kills of large animals are often partially or fully consumed and dispersed quickly in an ecosystem with a substantial scavenger community. Researchers estimate that documented wolf-caused grizzly deaths represent between 20 and 40 percent of actual events, meaning the true rate is roughly two to five times higher than confirmed documentation suggests.
Mountain lions are also implicated in grizzly mortality, though the evidence is more limited. Cougars are ambush predators that rely on striking velocity, positional advantage, and a killing technique directed at the back of the skull and upper spine. Against an adult grizzly, this technique works only if the positional advantage is absolute and the bear never has the opportunity to turn around.
The terrain features in confirmed cougar-on-grizzly kills cluster around specific locations: overhead ambush positions, cliff edges, fallen timber at elevation, and boulder fields that allow a cougar to position above a moving bear while limiting escape routes. These are locations that experienced cougars with home ranges overlapping grizzly territory have apparently learned to exploit.
Cougar kills of large prey are substantially underreported. Cougars are solitary animals that move their kills to concealed locations for extended feeding. A grizzly killed by a cougar and cached in a rocky crevice or under dense timber may simply never be found. The confirmed events are those where a GPS collar provided a precise location that allowed investigators to reach the site quickly.
The reverse interaction is also documented. Grizzlies will take cougar kills when they locate them and will kill cougars if they encounter them at an advantage. Telemetry studies tracking both species simultaneously show mutual wariness rather than simple dominance. Experienced cougars adjust their habitat use to minimize encounters while still exploiting terrain where ambush opportunities occasionally present themselves.
Intraspecific mortality, grizzlies killing other grizzlies, has also increased in ways the early models did not predict. Adult male infanticide has been documented for decades as a reproductive strategy, and fights between adult males over resources and mating access always carried mortality risk. What the population recovery appears to have done is intensify that pressure.
As the population has grown and available habitat has filled, overlap between individual home ranges has increased, and so has the frequency of aggressive interactions between adult males. High-density zones, particularly where late-season food resources concentrate large numbers of bears, produce encounter rates that translate into aggressive interactions at frequencies not represented in the data that built the original population models.
The bears most affected by elevated intraspecific mortality are not the dominant adult males doing the killing. They are subadult males and adult females, particularly females with cubs, whose survival is suppressed by encounter rates with dominant males. The cascade effects run through the population structure, changing age and sex distribution, increasing stress, reducing reproductive output, and displacing subordinate bears into suboptimal habitat.
Competition intensifies in years when critical food resources fail. When the whitebark pine crop is poor, elk calves are less abundant, or cutthroat trout are unavailable, bears compete more intensely for what remains. The population appears to be encountering the carrying capacity of its habitat not through starvation, which the models predicted as the primary density-dependent regulator, but through increased intraspecific violence, which the models treated as a secondary and roughly constant factor.
Even after accounting for wolf predation, cougar kills, intraspecific mortality, and the standard human-caused and vehicle-related deaths, a category of grizzly mortality events remains unexplained. These are adult deaths, including adult males, where the wound evidence does not match wolves, cougars, intraspecific fights, or human causes.
The characteristics of these events form a pattern. The location of the injuries, the bite radius, and the damage to skeletal structure suggest an attacker with different physical dimensions and attack behavior than any of the documented predators. The deaths cluster in a specific geographic zone of the greater Yellowstone ecosystem, in a consistent terrain type, and within a relatively narrow seasonal window.
Researchers have considered and tentatively ruled out disease, poisoning, equipment malfunctions producing false mortality signals, and misidentification of cause of death. The internal analysis, not yet published in peer-reviewed form, has informally pointed toward the possibility of a large predator operating in a specific portion of the ecosystem whose presence has not been formally confirmed.
The context is relevant. The gray wolf was believed absent from Yellowstone before the 1995 reintroduction, but credible reports of wolf presence existed in the years before the official reintroduction. The monitoring system is extensive but not comprehensive. Animals present at low density in specific terrain can go undetected for extended periods.
The hypothesis discussed informally is not a dramatic one. It does not require an unknown species. It requires the presence of an individual or small group of a known large predator species, most likely a large male grizzly with atypical predatory behavior or an extended-range mountain lion of exceptional size, operating in a way that current monitoring infrastructure has not directly documented. Resolving the question would require targeted deployment of additional monitoring in the specific geographic zone where the deaths are clustered.
The implications extend beyond Yellowstone. It is the most intensively monitored large carnivore ecosystem in the world. If its monitoring system is producing mortality data with a significant unexplained component, the gaps in less monitored systems are likely larger. Mortality gaps in grizzly populations in the northern Rockies or in wolf and cougar populations elsewhere may reflect undocumented causes rather than an absence of causes.
The management implications are also specific. Population models built on the assumption that adult grizzly mortality is driven primarily by human-caused events need updating to incorporate natural mortality sources that earlier data did not capture. Decisions about whether to delist the greater Yellowstone grizzly population from Endangered Species Act protection have included population viability analyses that incorporate mortality data. If natural mortality is larger than previous models assumed, those calculations need recalibration.
The evidence also illuminates the interactions among the three large predators in the ecosystem. Wolves suppress cougar populations through direct killing and behavioral displacement. Grizzlies displace wolves from kills and sometimes kill wolves. Cougars and wolves compete for ungulate prey. Each of these interactions is embedded in a web that the current data is only beginning to characterize at the system level.
New monitoring infrastructure includes increased collar deployment on adult animals in the zones with anomalous mortality, additional camera trap grids in the specific terrain types associated with the unexplained deaths, and genetic sampling protocols designed to detect large predators through environmental DNA in areas where direct observation has been insufficient.
The full picture is that Yellowstone’s grizzlies are being hunted, though not in a way that threatens the population’s overall recovery. Wolves in large packs under specific conditions and seasons are taking adult grizzlies. Cougars using terrain advantages are taking adults that the standard ecological framing said should be unchallengeable. Bears at increasing density are killing each other at rates the early recovery data did not predict. And something in a specific zone of the ecosystem is killing adult grizzlies in ways that none of these explanations fully account for.
The population’s place in the Yellowstone food web remains an open question. The animal at the top of the chain is still teaching those who study it what it means to share a landscape with other predators.
