Cougar Predation

[Acinonyx sp.]

South American Cougars:
Ecology of the Patagonia puma in southern Chile. Biol Conserv
"Guanacos Lama guanicoe, especially young animals, were the puma's most important prey item by biomass, but European hares Lepus capensis were preyed upon more than expected relative to available biomass. "
"Mammals and birds accounted for 92 and 8% of the

prey items counted in 405 puma scats. European hares represented 51%, guanacos 23%, sheep 5%, upland

goose 5%, rodents 3%, and lesser rhea 1%"

*This image is copyright of its original author

---

North American Cougars:
Cougar Kill Rate and Prey Composition in a Multiprey System

*This image is copyright of its original author

Kill Rate 
Most of our annual kill rate estimates fell within the range of values reported for cougars elsewhere (Table 1), which is not surprising given the large variation among preestimates. Adult females and family groups in west-central Alberta tended to kill closer to the high end of earlier estimates, whereas adult males killed at the lower end (in terms of frequency, not biomass). The ratio method we used to calculate kill rate was substantially more conservative (i.e., >25°/o; Hebblewhite et al. 2003) than the IKI estimator used in most other field studies of cougar kill rate. Thus, the adult cougar kill rates we found were among the highest recorded using field data. Subadults were less effective predators, and our kill rate estimates were lower than those given previously (Table 1). Two of 3 previous estimates, however, used cluster models without visitation (e.g., Anderson and Lindzey 2003, Mattson et al. 2007), which tend to overestimate subadult kill rate (Anderson and Lindzey 2003; K. Knopff, University of Alberta, unpublished data). Our kill rate estimates indicate that adult cougars are highly effective predators, killing at rates at the upper end of those recorded for wolves in both frequency and biomass (Peterson and Ciucci 2003, Sand et al. 2008, Webb 2009). Our estimates were inconsistent with lower proposed values based on energetics calculations and movement models for cougars during summer in Idaho, USA (Laundré 2005, 2008). The Idaho estimates differed from our summer estimates by as much as 365% in terms of frequency of killing and 538% in terms of prey biomass. Because kill rate fundamentally influences the effect predators have on their prey, the discrepancy between studies represents a substan- tial difference in the capacity for cougars to impact ungulates. For instance, Laundré et al. (2006) used a deterministic population model and kill rates derived from energetics calculations (Laundré 2005) to analyze cougar- mule deer dynamics in southern Idaho, concluding that cougars did not contribute to the decline or impede the recovery of mule deer between 1992 and 2004. Incorporat- ing higher kill rate values we identified would alter this conclusion considerably. Some of the discrepancies in kill rates in the literature (Table 1) are likely a result of inaccuracies associated with indirect methods or lack of precision due to small sample size in field-based studies. Energetics models often underestimate actual kill rates by large carnivores (Peterson and Ciucci 2003), and classification success of clustering models remains far from perfect, yielding reasonable estimates of kill rate only over long monitoring intervals and under circumstances where false positives and false negatives at individual location clusters cancel appropriately (Webb et al. 2008, Knopff et al. 2009). Some inconsistencies might be due to different ecological conditions among regions, but it is currently impossible to ascertain how much of the variation among studies can be ascribed to ecological vs. methodological differences. We believe that visiting GPS telemetry clusters in the field and using a ratio estimator to calculate kill rate represents the current gold standard in kill rate estimation for cougars (see reliability of results below) and future studies using similar techniques in other places will help better define the natural range of variation in cougar kill rate.

 Influence of Prey Vulnerability 
Cougars in west-central Alberta killed primarily female ungulates just before, during, and just after the birthing period (Apr-Jun), males just before and during the rut (Sep- Nov; Fig. 3), and focused prédation on juvenile ungulates in spring consistent with the reproductive and juvenile vulnerability hypotheses, reinforcing the notion that cougars select for vulnerable prey. Similarly, whereas cougars are capable of killing prey as large as adult moose and feral horses, prey of this size were rarely taken (<2% of prey weighed >400 kg). Most (74%) moose and feral horses killed by cougars were juveniles, and all cougar age-sex classes killed a higher proportion of large ungulate species (i.e., ad >200 kg) in summer when smaller juveniles were available (Table 3). This pattern is likely due to lower risks associated with attacking smaller prey (Sunquist and Sunquist 1989). Although our seasonal result is novel, that cougar prédation on large ungulate species tends to focus on animals < 1 year old has been well-documented (Hornocker 1970, Turner et al. 1992, Ross and Jalkotzy 1996, Murphy 1998, Husseman et al. 2003). Although we were unable to distinguish relative contri- butions of elevated encounter rate (due to increased prey abundance), greater average susceptibility of prey to attack, or greater searching effort by cougars to higher kill rate in summer, these mechanisms were collectively more than twice as important as reduced handling time (Fig. 4). Because close grouping between mothers and young (e.g., moose) or hiding behavior of neonates (e.g., white-tailed deer) are strategies that limit searching efficiency during the pulse of juvenile availability in spring (Fitzgibbon 1990¿, Fryxell et al. 2007), we suspect that high vulnerability of juveniles to attack when encountered is the dominant force driving elevated summer kill rates, which is further supported by selection for juveniles in summer (i.e., proportions in cougar diet higher than max. possible annual production), even though encounter rates with juveniles might be low relative to older ungulates. That vulnerability to attack might influence cougar prédation runs contrary to the assumption that that felids kill prey as encountered (e.g., Wilmers et al. 2007). Rather, our results support studies that indicate selection for vulnerable individuals may be widespread in felid-ungulate systems (e.g., Pierce et al. 2000, Molinari-Jobin et al. 2004, Owen- Smith 2008). These patterns can be interpreted as part of a broader optimal foraging strategy where felids attempt to minimize risks associated with prédation by targeting easier prey (e.g., Ross et al. 1995, Logan and Sweanor 2001, Owen-Smith 2008, Knopff et al. 2010). Reply