# Recolonizing wolves influence the realized niche of resident cougars

> When wolves returned to the Southern Yellowstone Ecosystem, the cougars already living there had to adjust. Following collared cougars to more than 650 kills over a decade, researchers found that cougars took more mule deer and fewer elk in summer, hunted more in forest, and lost kittens to wolves. The authors argue that wolves are reshaping the cougar’s place in the ecosystem.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2015-elbroch-lendrum-recolonizing-wolves-influence-realized/

## Citation and access

- **Authors:** L. Mark Elbroch, Patrick E. Lendrum, Jesse Newby, Howard Quigley and Daniel J. Thompson
- **Year published:** 2015
- **Journal or source:** Zoological Studies
- **Full citation:** Elbroch, L. M., Lendrum, P. E., Newby, J., Quigley, H. and Thompson, D. J. 2015. Recolonizing wolves influence the realized niche of resident cougars. Zoological Studies 54: 41.
- **DOI:** 10.1186/s40555-015-0122-y
- **Access status:** Open access (CC BY 4.0)
- **PDF:** https://www.originalwisdom.com/wp-content/uploads/2026/11/2015_Elbroch_and_Lendrum_etal_recolonizing-wolves-influence-the-realized-niche-of-resident-cougars.pdf

## Tags

- **Species:** Mammals > Grey wolf; Mammals > Mountain lion
- **Region:** North America > United States
- **Tracking:** Bones; Feeding signs; Hair; Kill sites; Tracking; Tracks and signs identification
- **Behaviour and ecology:** Behaviour; Carnivore coexistence; Competition; Feeding; Habitat selection; Hunting; Interspecific competition; Intraguild interactions; Predation; Range expansion
- **Methods and technology:** Akaike Information Criterion (AIC); Analysis of variance (ANOVA); Global Positioning System (GPS) collars; GPS cluster analysis; Linear regression; Logistic regression; Radio telemetry; Very high frequency (VHF) radio collars
- **Data type:** Existing data; New field data
- **Observer qualification:** Not reported
- **Discipline:** Behaviour and ecology; Wildlife tracking
- **Year published:** 2015

## Detailed keywords

- **Species:** Studied: cougar (Puma concolor), and gray wolf (Canis lupus) as the competitor whose numbers rose. Prey found at cougar kill sites: elk (Cervus elaphus), including calves, and mule deer (Odocoileus hemionus); “other ungulates” were bighorn sheep (Ovis canadensis), moose (Alces alces) and pronghorn (Antilocapra americana); 29 prey items were not ungulates (not named). In the study area but not studied: black bear (Ursus americanus), grizzly bear (Ursus arctos), coyote (Canis latrans), red fox (Vulpes vulpes), bison (Bison bison) and a small number of white-tailed deer (Odocoileus virginianus).
- **Location and habitat:** Southern Yellowstone Ecosystem (SYE), about 2300 km² north of the town of Jackson, northwest Wyoming, United States, including parts of Grand Teton National Park, the National Elk Refuge and the Bridger-Teton National Forest. Elevations from 1800 m in the valleys to over 3600 m. Cottonwood riparian zones and sagebrush uplands at lower elevations; aspen, Douglas fir and lodgepole pine at middle elevations; Engelmann spruce and subalpine fir higher up. Short cool summers, long snowy winters, snow up to more than 245 cm deep. Elk and deer migrate seasonally; elk receive supplemental winter feeding.
- **Topics:** wolf recolonization after reintroduction; competition between large carnivores (exploitation and interference competition); realized niche; prey switching from elk to deer; prey selection by season; hunting habitat (forest versus open); predation on elk calves; kill rates; cause-specific mortality of cougars, including kittens killed by wolves and starvation; investigating kill sites and death sites.
- **Statistics or models:** generalized linear models (Poisson) of changes over time in diet proportions, calf proportions and kills in forest; multiple regression with Akaike’s information criterion adjusted for small samples (AICc), Akaike weights and model averaging; pairwise correlations (cutoff 0.6) to remove correlated predictors; logistic regression of GPS cluster attributes to predict kill sites, with receiver operating characteristic (ROC) curves (area under the curve 0.827) and sensitivity and specificity curves; an alternative “night > 1” rule (clusters spanning more than one night counted as kills); two-way analysis of variance and linear regression with individual cougar as a random effect for kill rates. Software: JMP 11; a Python script for cluster identification. Inconsistencies: (1) the text says wolves rose from 15 to 91, “a 600 % increase” (15 to 91 is about a 500% increase); (2) the Results report 33 kittens of known cause of death, but the listed causes add up to 26, and the percentages match 26, not 33; (3) the Discussion says 5 kittens killed by wolves were 23% of known-cause deaths and 9 adult starvations were 20%, figures that do not follow simply from the Results counts; (4) the kill-site model section cites Table 1 for cluster attributes, but Table 1 holds the yearly wolf, prey and diet figures; (5) the Introduction calls one wolf killing of a cougar “intraspecific” where interspecific is meant.
- **Data type:** New field data: cougar captures from 2001, kill sites and prey from January 2002 to 1 October 2012, cause-specific mortalities 2002 to 2012, and GPS clusters and kill rates 2005 to 2012. Existing data: yearly elk counts and cow-calf ratios and mule deer counts from the Wyoming Game and Fish Department, wolf numbers from annual wolf recovery reports, and snow depth from a SNOTEL station (Gros Ventre Summit).

## Synopsis

### 1. What they did

Wolves were reintroduced to Yellowstone National Park and central Idaho from 1995, after being absent since 1926. In the Southern Yellowstone Ecosystem, wolf numbers then rose from 15 to as many as 91 between 2002 and 2012, while a resident cougar population was already there. The researchers used this natural experiment to ask whether wolves, the dominant competitor, were changing how cougars live: what they hunt, where they hunt, how often they kill, and how they die.

- Capturing and collaring cougars. From 2001, cougars were caught in winter, when snow made it easier to find them, using hounds to tree or corner them. They were fitted with Very High Frequency (VHF) radio collars or Global Positioning System (GPS) collars with mortality sensors.

- Finding kills. The team searched sites where VHF-collared cougars had stayed put for 24 hours or more, and GPS clusters (groups of locations) where a cougar had stayed for 8 hours or more. They identified prey from hair, skin, rumen and bone fragments, used bite marks and feeding patterns to tell a kill from scavenging, aged prey by their teeth, and recorded the habitat.

- Investigating deaths. When a collar stopped moving, they visited the site as soon as possible and used signs of struggle and of other animals to work out the cause of death. Necropsies, usually led by a veterinarian, and laboratory disease tests added to this.

- Testing for change. They tested whether the share of elk, deer and elk calves in cougar diets, and the share of kills in forest, changed from 2002 to 2012, in summer and in winter. They then compared models explaining the changes with elk availability (elk per wolf), deer availability (deer per elk), snow depth and, for calves, elk cow-calf ratios.

- Kill rates. For GPS-collared cougars (2005 to 2012), they built models to predict which GPS clusters held kills, checked them against clusters visited in the field, and used them to estimate kills per week over time.

### 2. What they found

Prey and diet

- Between January 2002 and October 2012 they recorded 411 winter and 239 summer prey killed by 28 female and 10 male cougars, plus 37 prey of unmarked cougars. Only 29 prey were not ungulates.

- In summer, deer made up 42.4% of the diet and elk 38.3%. In winter, when deer migrate almost entirely out of the area, deer were only 7.2% and elk 74.4%.

- In summer, the share of deer in cougar diets rose and the share of elk fell over the study. In winter, there was no change.

- The best model for both changes was elk availability (elk per wolf). Both elk and deer were declining, deer faster than elk, while wolves increased steeply. By 2012, elk per wolf was about 14% of the 2002 value.

Elk calves and habitat

- The share of calves among elk killed rose in summer but did not change in winter. Cow-calf ratio was the strongest predictor, followed by snow.

- The share of summer kills found in forest rose over the study; in winter it did not change. On average 76% of summer kills and 58% of winter kills were in forest.

Kill rates

- From 2005 to 2012, 17 GPS-collared cougars formed 2729 clusters. The team visited 309 clusters from 14 cougars, an average of 8 days after the cluster formed, and found kills at 269.

- The best statistical model classified only 49.7% of known kill and non-kill clusters correctly. The simple rule “the cougar stayed more than one night” found kills at 86% of kill sites but also assigned kills to 40% of sites where none were found. The authors used this rule, as visits were often late and some kills were likely missed.

- Kill rates did not change over time in either season. Mean rates were 0.90 animals per week in summer and 0.99 in winter.

Cougar deaths (2002 to 2012)

- Kittens: 10 starved, 5 were killed by wolves, 5 died of unknown causes, 2 were killed by an unidentified predator, 2 by other cougars, 1 died of disease and 1 was killed by wildlife managers.

- Subadults and adults: 16 were legally hunted, 9 starved, 5 were killed by other cougars, 5 died of disease, 5 of unknown causes, 3 were killed by wildlife managers, 1 was killed illegally and 1 died of old age.

- All starvations, kittens and adults, happened after 2005. No cougars were hunted from 2007 to 2012.

### 3. What they concluded

- Cougars switched from elk toward deer in summer as wolves increased. Because both elk and deer were declining, changing prey numbers alone did not explain the switch; competition with wolves at least partly did.

- Elk numbers fell only from 14,956 to 11,982, far more than the low-density cougar population needs, so the authors did not see this as cougars running short of elk (exploitation competition). They favoured another explanation: cougars moved into more complex, forested habitat to avoid wolves, where they met more forest-dwelling mule deer. The rise in summer kills in forest supports this.

- Cougars took more elk calves in summer, not fewer as predicted. Cow-calf ratios stayed steady, probably helped by supplemental feeding of elk, so calves may not yet be scarce enough to limit cougars.

- No change in kill rates was found, but the authors think this reflects their methods (late site visits, the more-than-one-night rule missing small prey and kills lost to competitors, and mixed GPS fix rates) rather than what was really happening.

- Wolves killing cougar kittens is strong evidence of direct (interference) competition. The starvations may also partly reflect competition with wolves.

- Together with earlier evidence that cougars chose home ranges further from wolves, the results show new and ongoing competition. Resident cougars may have been “naïve intermediaries” relearning to live with a dominant competitor, and their realized niche may now resemble the one they had when wolves last shared Yellowstone.

### 4. How tracking was or was not used

The study rests on reading tracks and signs at sites found by telemetry. At kill sites, the team identified prey from remains (hair, skin, rumen and bone fragments) and used bite marks and feeding patterns to decide whether the cougar had killed the animal or was scavenging. They aged prey from tooth eruption and wear. At death sites, they used signs of struggle and signs of other animals to judge the cause of death and whether animals present had acted as predators or scavengers, citing a tracks and signs field guide (Elbroch 2003). Site searches where snowfall may have hidden evidence were left out of the kill-site model. Cougars were captured in winter, when snow made it easier to tell whether a cougar was present, and hounds were used to trail them; the paper does not describe people following tracks themselves.

### 5. Tracker or observer certification

Not reported. No qualification or skill test is given for the people who investigated kill and death sites, although the prey, kill-or-scavenge and cause-of-death findings all depend on their sign interpretation. The authors note that field crews often reached clusters many days late and likely missed evidence of some kills.

### 6. Other technologies, skills or innovations

- GPS collars (Telonics, Televilt, Vectronics) taking locations every 3 to 10 hours, mostly six to eight a day, and VHF collars (Telonics), with mortality sensors that signalled when a collar had not moved for 8 hours or more.

- A Python script (developed by Knopff and colleagues) to find GPS clusters: two or more locations within 100 m and 48 hours.

- Logistic regression models of cluster attributes (more than one night, hours at night, fidelity, distance to the next cluster) to predict kill sites, compared with the simple more-than-one-night rule.

- Necropsies and laboratory disease testing at the Wyoming State Veterinary Laboratory.

### 7. Implications for conservation (from the paper)

- Recolonizing wolves are reshaping how resident cougars hunt and survive, through both prey switching and direct killing of kittens.

- Species reintroductions offer rare chances to test how a dominant competitor affects a subordinate one, since carnivore numbers are hard to manipulate in the wild.

- More intensive fieldwork may be needed to detect changes in kill rates, since late site visits and the methods used can miss small prey and kills taken over by competitors.

### 8. Implications for tracking (Original Wisdom's observations)

_These are Original Wisdom's observations, not the authors'._

- The kill-rate results show how much depends on getting to the ground in time and reading it well: with visits made on average 8 days after a cluster formed, the authors believe crews missed some kills, and the models built on those field checks performed poorly.

- Telling a kill from scavenging, and a wolf killing from another death, relies on bite marks, feeding patterns and signs of struggle. Reporting who did this work and how their skill was checked would make such findings easier to weigh.

## Cite this summary

Lawrence, K. 2026. Summary of: Recolonizing wolves influence the realized niche of resident cougars (Elbroch and Lendrum et al., 2015). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2015-elbroch-lendrum-recolonizing-wolves-influence-realized/