# The effects of puma prey selection and specialization on less abundant prey in Patagonia

> Scat studies said Patagonian pumas lived on hares. When certified trackers checked 694 GPS clusters left by 9 collared pumas in Chile, they found mostly guanacos, and hares made up less than 2% of the meat eaten. The real surprise was individuals: just 2 pumas drove most sheep kills, and 2 others selected the endangered huemul deer.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2013-elbroch-wittmer-effects-puma-prey-selection/

## Citation and access

- **Authors:** L. Mark Elbroch and Heiko U. Wittmer
- **Year published:** 2013
- **Journal or source:** Journal of Mammalogy
- **Full citation:** Elbroch, L. M. and Wittmer, H. U. 2013. The effects of puma prey selection and specialization on less abundant prey in Patagonia. Journal of Mammalogy 94(2): 259–268.
- **DOI:** 10.1644/12-MAMM-A-041.1
- **Access status:** Hosted (freely available elsewhere)
- **PDF:** https://www.originalwisdom.com/wp-content/uploads/2026/11/2013_Elbroch_and_Wittmer_the-effects-of-puma-prey-selection-and-specialization-on-less-abundant-prey.pdf

## Tags

- **Species:** Mammals > Mountain lion
- **Region:** South America > Chile
- **Tracking:** Bones; Feeding signs; Hair; Kill sites; Tracking; Tracks; Tracks and signs identification; Trailing
- **Behaviour and ecology:** Behaviour; Feeding; Hunting; Kleptoparasitism; Predation; Scavenging
- **Methods and technology:** Analysis of variance (ANOVA); Chi-square test; CyberTracker; CyberTracker Tracker Certification; Distance sampling; Geographic Information System (GIS); Global Positioning System (GPS); Global Positioning System (GPS) collars; GPS cluster analysis; Logistic regression; Very high frequency (VHF) radio collars
- **Data type:** New field data
- **Observer qualification:** Certified trackers
- **Discipline:** Behaviour and ecology; Human–wildlife conflict; Wildlife tracking
- **Year published:** 2013

## Detailed keywords

- **Species:** Studied: puma (Puma concolor), 9 animals fitted with GPS collars (4 males and 5 females). Prey found at kill sites: guanaco (Lama guanicoe), including first-year young called chulengos; domestic sheep (Ovis aries); European hare (Lepus europaeus); huemul (Hippocamelus bisulcus), an endangered deer; Patagonian haired armadillo (Chaetophractus villosus); upland goose (Chloephaga picta); culpeo fox (Lycalopex culpaeus). In the study area but not studied: about 150 cows, Andean condor (Vultur gryphus), southern caracara (Caracara plancus), Chimango caracara (Milvago chimango) and black-chested buzzard eagle (Geranoaetus melanoleucus). Mentioned from other studies: orca (Orcinus orca), sea otter (Enhydra lutris), bighorn sheep (Ovis canadensis), snow leopard (Uncia uncia), red fox (Vulpes vulpes), jaguar (Panthera onca), mule deer (Odocoileus hemionus) and white-tailed deer (O. virginianus).
- **Location and habitat:** About 1,100 km² north of Lago Cochrane, southern Aysén District, central Chilean Patagonia, Chile: Lago Cochrane National Reserve (69 km²), the private Estancia Valle Chacabuco (690 km²) and about 440 km² of Jeinimeni National Reserve, all to be combined into the future Patagonia National Park. Rugged mountains with open Patagonian steppe, high-elevation deciduous lenga (Nothofagus pumilio) forest, and lower shrublands of ñirre (N. antarctica) with chaura (Pernettya mucronata) and calafate (Berberis microphylla). About 2,500 sheep were kept in the eastern part of the estancia; sheep farming was the main land use beyond the park borders.
- **Topics:** puma diet; prey selection versus prey specialization; individual versus population-level foraging; rare and less abundant prey; endangered huemul; recovering guanaco populations; livestock predation and human–puma conflict; apparent competition; GPS cluster investigation versus scat analysis; misidentified scats; kill site search area; time until investigation; cluster length and prey size; time of day of kills; bone marrow health index; puma tooth marks on guanaco skulls; kleptoparasitism (a male taking over kills made by females); removal of individual problem pumas versus wide-scale culling.
- **Statistics or models:** GPS clusters defined as 2 or more locations within 150 m; chi-square goodness-of-fit tests for second-order and third-order resource selection (after Johnson 1980) with Bonferroni Z-tests; logistic regression (time until investigation and finding remains); analysis of variance (ANOVA) on marrow scores; Spearman rank correlation (rs); prey densities from program Distance 6.0 (published in Elbroch and Wittmer 2012); biomass estimates from growth and weight data; fixed-kernel home ranges (from Elbroch and Wittmer 2012). Inconsistencies in the paper: (1) the text reports 433 kill sites, Table 1 gives 423 kills for all pumas (the individual puma counts add up to 423), and Figure 2 lists 463 prey (it appears to include the 30 guanacos killed by unmarked pumas); (2) huemul are said to make up 0.9% of the biomass killed overall, yet 1.3% for females and 1.6% for males, both higher than the overall figure; (3) Table 1 marks no individual result as significant, although three values (M2 huemul, M3 sheep, M4 sheep) pass the cutoff and the text counts them as selection; (4) the inverse correlation between time until investigation and finding remains is given as rs = 0.01 with no minus sign; (5) female F4 is said to have killed 5 huemul but only 4 are described; (6) the results mention a Telonics collar, a make not listed in the methods.
- **Data type:** New field data: pumas captured March to September in 2008 and 2009; GPS clusters and kills from 2008 to 2009; collars worn for a mean of 9.33 months.

## Synopsis

### 1. What they did

Earlier studies of puma diets in Patagonia were mostly based on scats and reported that pumas lived largely on European hares. The authors point out that puma and fox scats are easily confused, and that scats cannot show what individual pumas are doing. They wanted to know whether pumas as a population, and as individuals, choose some prey more than their numbers on the landscape would suggest, and what this means for less common prey such as the endangered huemul deer and domestic sheep.

- Capturing and collaring. In the future Patagonia National Park in Chile, the team rode on horseback until they found fresh puma tracks, then used hounds to bring the pumas to bay in a tree or on a rocky outcrop. They fitted 9 pumas (4 males and 5 females) with Global Positioning System (GPS) collars that took a location every 2 hours and sent the data by satellite (Argos) every few days.

- Visiting GPS clusters. A cluster was 2 or more locations within 150 m. CyberTracker-certified observers visited every cluster with at least one location at night or in twilight. Daytime-only clusters were checked for the first 2 months, never showed a kill, and were then treated as day beds.

- Reading the remains. Hair, skin, stomach contents and bone fragments identified the prey. Bite marks and what had been eaten showed whether the puma had killed or scavenged. They also found kills by uncollared pumas by following tracks and scavenging birds, counting them as puma kills only when tracks confirmed a puma and wounds and signs of struggle showed it was a kill.

- Ageing, health and tooth marks. They aged guanacos and huemul by their teeth, scored prey health from femur bone marrow, and checked guanaco skulls for puma tooth marks.

- Comparing diet with what was available. They estimated prey weights and used prey counts from transects to work out the biomass of guanacos, huemul, hares and sheep on the landscape. They compared what each puma killed with prey available in their own home range, and with what the puma population as a whole killed.

### 2. What they found

GPS clusters and kill sites

- They investigated 694 GPS clusters and found prey remains at 63.3%, including 433 kill sites and 6 scavenging sites.

- Kill sites were on average 8 m from the nearest puma GPS location.

- Clusters were visited on average 11 days after the puma left (range 0 to 78 days). The longer the wait, the less likely they were to find remains. One collar failed to send data, so that puma’s clusters were visited about 2 years later.

- Clusters shorter than 14 hours yielded kills 37% to 59% of the time; at 16 hours, 91%; at 18 hours, 98%; and longer clusters always held a kill, except two long stays by females at den sites.

- 64% of kills were made at night, 26% in twilight and 10% by day.

- Pumas stayed longer at bigger prey, but the pattern was loose: 120 kg prey were found at clusters as short as 2 hours.

- All 6 scavenging events were one male (M2) taking over guanacos killed by collared females.

What the pumas ate

- Guanacos were by far the main prey: 332 killed, about 23,248 kg, and 88.5% of the biomass. Sheep made up 8.7%, hares 1.9% and huemul 0.9%.

- Males and females killed different prey. Females killed no sheep, and only females killed prey under 8 kg, including all 73 hares. One female killed 25 hares and 2 armadillos in her first 95 days of monitoring, before her first litter, before killing her first guanaco.

- Males and females killed guanacos in similar health.

- Puma tooth marks were found on the head or jaw of 87.5% of young guanacos (chulengos), 60% of second-year guanacos, but only 25% of adults.

- Two collared pumas killed 7 huemul. A male killed 2 old females in poor health; a female killed 5, all healthy.

Prey selection and specialization

- As a population, pumas specialized on guanacos (killed more than any other prey). They selected sheep, killing 3.8 times more than expected from their numbers, and killed fewer hares than expected. They did not select huemul.

- Seven of the 9 pumas specialized on guanacos. One male (M4) specialized on sheep. One female (F5) specialized on hares by number, but on guanacos by weight.

- Compared with prey in their own home ranges, 3 of 9 pumas showed selection: 1 for huemul and 2 for sheep.

- Compared with the population as a whole, 5 of 9 pumas differed: 2 selected sheep, 1 guanacos, 1 huemul and 1 hares.

- Just 2 sheep-selecting males were enough to create the population-level pattern for sheep.

### 3. What they concluded

- GPS cluster investigation gave a very different picture from earlier scat studies in Patagonia: hares made up only 1.9% of what pumas ate here, not 40% to 52%. The authors suspect earlier studies counted some fox scats as puma scats.

- Pumas may or may not specialize on or select rare or less abundant prey, but when a few individuals do, their impact can be large.

- Only by looking at individuals did it become clear that 2 of the 6 pumas sharing their range with huemul selected them more often than expected.

- The authors conclude that removing individual pumas proven to select huemul or livestock is the best strategy for pumas, huemul and livestock owners. Wide-scale culling would be ineffective unless it happened to remove those individuals, and non-targeted removal over small areas has failed elsewhere because new pumas move in.

- The two ways of measuring selection (against home-range prey and against the population) both picked out pumas selecting rare prey, and should be compared in future work.

- Further research is needed on how pumas affect huemul and guanaco populations, how hares and sheep subsidize pumas, and whether individual diet variation occurs across the puma’s range.

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

- Capture: captures took place from March to September, when snow made pumas easier to find. The team rode on horseback until they found fresh puma tracks, then used hounds to bring the pumas to bay in a tree or on a rocky outcrop.

- Kill site investigation: at each GPS cluster, observers used tracks and signs to identify prey and decide whether the puma had killed or scavenged: hair, skin, stomach or rumen contents, bone fragments, where the bite marks were, and what had been eaten. They marked the kill site at the rumen for ungulates, or at the stomach, intestines or largest pile of remains for small prey.

- Uncollared pumas: additional kills were found by following tracks and scavenging birds, and counted only when tracks confirmed a puma and wounds and signs of struggle showed a kill.

- Ageing and health: teeth were used to age prey, and the colour and texture of femur marrow to score health.

- The authors note that felids often pluck and drop fur, even from prey as small as hares, which makes cluster work well suited to cats.

### 5. Tracker or observer certification

Field investigations were carried out by CyberTracker-certified observers, citing Elbroch et al. (2011) and Evans et al. (2009). The certification level is not given. N. Wight, C. Rivera, C. McFarland and R. Millacúra are thanked for help with kill site investigations.

### 6. Other technologies, skills or innovations

- Argos-GPS collars (SirTrack, Televilt and Lotek models; a Telonics collar is also mentioned in the results) and very high frequency (VHF) collars.

- Handheld Garmin GPS units (accuracy 5 to 10 m) to find clusters and mark kill sites; ArcGIS to map locations.

- A two-part bone marrow index (colour and texture) to score prey health in the field.

- Distance sampling (program Distance 6.0) for prey densities.

- The distance between kill sites and the nearest GPS location (8 m on average) was measured to help future researchers know how large an area to search.

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

- Earlier, scat-based studies may have given a distorted picture of puma diets in southern South America.

- The huemul population, about 120 of roughly 1,000 left in Chile, could not absorb losses to a few pumas; the effects of current puma predation threaten its viability.

- Where only a few individuals select rare prey, management aimed at those individuals is more effective than population-level control.

- Understanding individual pumas will help in managing livestock predation where livestock share the land with native prey.

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

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

- The study pairs GPS with ground skills: the collars showed where pumas spent time, but certified observers reading hair, bones, bite marks and feeding signs determined what was killed, whether it was killed or scavenged, and its age and health.

- The authors’ concern that earlier studies mistook fox scats for puma scats shows how much diet research depends on correct scat identification.

- The search-area figure (kill sites on average 8 m from the nearest GPS point) and the cluster-length figures are practical guides for anyone checking GPS clusters in the field.

## Cite this summary

Lawrence, K. 2026. Summary of: The effects of puma prey selection and specialization on less abundant prey in Patagonia (Elbroch and Wittmer, 2013). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2013-elbroch-wittmer-effects-puma-prey-selection/