Puma responses to close approaches by researchers

Transcription

1 PUMA RESPONSE TO RESEARCHER APPROACH 1 Puma responses to close approaches by researchers Linda L. Sweanor, Kenneth A. Logan, and Maurice G. Hornocker Abstract We documented behaviors that wild pumas (Puma concolor) exhibited when approached by researchers during a 10-year study of a puma population in New Mexico. We approached and visually observed responses of 75 radiocollared pumas on 251 occasions and 7 noncollared pumas on 5 occasions. These pumas seldom had contact with people, except in the context of puma research activities. Pumas we approached included adult females with nursing (n = 71) and weaned (n = 46) cubs, solitary independent females (n=77), solitary independent males (n=46), breeding pairs (n=5), and weaned cubs in the absence of their mother (n=11). Approach distance ranged from m (median=30.8 m), number of observers ranged from 1 8 people (median=2), and duration of observation ranged from <1 to >120 minutes. Pumas exhibited threat behavior (e.g., hissed, growled, approached, charged) in 16 instances (6% of observations). Females accompanied by cubs exhibited 14 of the 16 threats. Although we observed all threats when we were 2 50 m away (median=18.5 m), in another 156 observations when we approached to within similar distances (range = 3 50 m, median = 20.0 m), pumas showed no threat response. In these cases, the pumas left the area (n=114), watched us without leaving (n = 37), or showed no detectable response (n = 5). These responses by pumas to human approaches provide insight into the range in behaviors exhibited by pumas in environments where they have not become habituated to humans. This may aid our understanding of puma behavior, including interpretations of behaviors that might be exhibited by pumas in more human-dominated environments. Key words behavior, humans, New Mexico, people, puma, Puma concolor Puma attacks on humans are rare, numbering 108 verified accounts from in the United States and Canada (Beier 1991, Fitzhugh et al. 2003). However, the frequency of attacks is increasing, with more fatal attacks occurring from than during the previous 80 years (Beier 1991). Puma attacks on humans have generated recent popular books that describe details of attacks and naturally focus on people s behavior, reactions, and perspectives (Deurbrouck and Miller 2001, Etling 2001, Baron 2004). Yet, relatively little detailed information is available on behaviors exhibited by the puma that result in an attack. The focus on rare puma attacks also leaves out information on puma behavior during other, likely more frequent occasions when pumas encounter people but do not attack. There is popular speculation that pumas in more frequent proximity to humans behave differently than those that rarely come into contact with people (Baron 2004). One possibility is that pumas subjected to repeated, neutral encounters with people might stop responding to them and thus become habituated to humans (Whittaker and Knight 1998). In this paper we provide original, quantitative, observational data on responses of wild pumas to close encounters with people that did not result in attacks. These data are intended to inform managers and citizens in puma habitat about the variation of puma behavior during close encounters with people so that they may Authors address: Hornocker Wildlife Institute, P.O. Box 3246, Moscow, ID 83843, USA. Present address for Sweanor and Logan: Colorado Division of Wildlife, 2300 South Townsend Avenue, Montrose, CO 81401, USA; for Sweanor: lsweanor@mindspring.com. Present address for Hornocker: Wildlife Conservation Society, 2023 Stadium Drive, Suite 1A, Bozeman, MT 59719, USA. Wildlife Society Bulletin 2005, 33(3):1 xxx Peer refereed

2 2 Wildlife Society Bulletin 2005, 33(3):XX XX better diagnose puma behavior and judge risk of attack. During a 10-year puma study in New Mexico, we approached and visually observed pumas to gather data on puma population biology and predation and to address a series of scientific hypotheses (Logan and Sweanor 2001). Analyzing puma responses to close encounters with people was not a preconceived objective of our study design; however, we later recognized that our detailed field notes on puma responses to our approaches provided unique, quantitative data on the range of behaviors wild pumas exhibit during encounters with humans. Study area The San Andres Mountains study area of southcentral New Mexico encompassed 2,059 km 2. The mountains were located at the northern edge of the Chihuahua Desert, had a semi-arid climate, and rose from 1,280 2,730 m. Logan and Sweanor (2001) described the area in detail. The area was remote, and because it was almost completely within White Sands Missile Range, human activity during our research was minimal. The human density was about 0.65 persons per 100 km 2 and consisted of about 30 year-round residents, including 4 5 puma researchers. Consequently, pumas living there had few opportunities for human encounters, except in the context of puma research activities. Methods We followed capture and handling guidelines adopted by the American Society of Mammalogists (Committee on Acceptable Field Methods 1987). We captured independent pumas and some large cubs with foothold snares (Logan et al. 1999) and fitted them with radiocollars. We captured most cubs and marked them with ear tattoos and ear tags at nursery sites when they were 4 8 weeks old,and radiocollared several. We captured and tagged 241 pumas, 126 of which wore radiocollars. Thus, pumas involved in our visual observations were subjects of capture efforts by members of our research team. We caught individual pumas 1 6 times, although most only once or twice (Logan and Sweanor 2001). Even though we attempted to minimize trauma and discomfort during the capture and handling process, we probably subjected captured pumas to relatively traumatic, or at least negative, experiences with people. We approached radiocollared pumas in attempts to document health, association with a kill cache, associations with unmarked pumas, breeding behavior, presence of cubs, or cub survival. When females had nursing cubs, we typically returned to the nursery site at a later date (preferably when the mother was away) to capture and mark cubs. Approaches more frequently involved female pumas because this was our main method of obtaining data on reproduction and cub survival (Logan and Sweanor 2001). Therefore, our approaches were not a random or systematic sample of the puma population. Prior to approaching a puma, we triangulated its position using radiotelemetry by taking bearings on its collar s peak aural signals from at least 2 vantage points (Logan and Sweanor 2001). The puma s location was then plotted on a 7.5-minute United States Geological Survey topographic map. We planned an approach route that would best conceal our movements and keep us at a higher elevation than the puma. We continuously monitored the puma s radiotransmitter during the approach. When >1 researcher was present, the observer with the telemetry equipment typically would hike a set distance toward the peak signal while the remainder of the research team watched the location of the puma with binoculars. After reaching a new vantage point, the observer would stop and watch the puma location with binoculars while the remainder of the team hiked to join him or her. We continued this relay approach, cutting the distance between the puma and us until we were able to visually observe the puma. We mapped the puma s location, as well as our own, on a 7.5-minute topographic map. If the puma left the area, we paced the distance between our location and the puma s closest observed location. If the puma did not leave, we visually estimated the distance and crossreferenced that estimate with the mapped locations. On a few occasions, we were able to return to the site at a later date and estimate the distance by pacing. We defined the approach distance as the closest distance between the observer(s) and the puma during the visual observation. Pumas we approached included families (i.e., females with nursing or weaned cubs), breeding pairs, solitary independent males and females, and weaned cubs in the absence of their mothers. Independent pumas were either subadults (not yet breeding) or breeding adults. We considered cubs >9 weeks old weaned (Logan and Sweanor 2001).

3 We recorded all observable puma behaviors in our field journals. Because our data were descriptive and gathered opportunistically,we only address factors that we recorded consistently: the puma s status, our approach distance, the number of people involved, and puma behavior. We report ranges for observations, and where the distributions of observations were strongly tailed, we report the medians (instead of means) for the central tendencies. Modes also are reported when there are a large number of identical measurements (Zar 1984). Based on our observations of puma behaviors, we developed 4 response categories: 1) puma detected the researchers, as expressed by apparent visual contact, and moved away; 2) puma stayed at first observed location, visually detected the researchers, and watched; 3) puma appeared to ignore or did not appear to detect the researchers; and 4) puma displayed a threat response toward the researchers. Pumas that moved away did not always leave immediately,but they moved distances >100 m or beyond a significant visual barrier such as a ridgeline. Pumas that stayed showed obvious signs they had detected the observer(s), such as staring, shifting body position, or seeking denser cover at the site. Staying pumas might move a few meters during the course of the observation but not toward the observer. During some observations we could not determine whether the puma detected us or was simply ignoring our presence. In these cases, we made no eye contact and the puma continued to behave as it had when first observed. Cat facial and postural signals change with increasing fear and aggression (Leyhausen 1979). Because we often could not determine whether a puma was reacting more from fear or aggression, we used the term threat response to include Puma response to researcher approach Sweanor et al. 3 Table 1. Puma responses to human observers who approached to m, San Andres Mountains, New Mexico, Median (range) Median (range) Puma response No. approach number of Ignored or Puma class a approaches distance (m) observers Left Stayed not noticed Family (55 300) 2 (1 5) Solitary (60 400) 1 (1 4) Breeding pair ( ) 1 ( Cubs ( ) 1 (1 2) Total (55 400) 1 (1 5) a Family includes F with nursing or weaned cubs; solitary includes adults and subadults; cubs includes weaned cubs weeks old and in the absence of their mother. either type of behavior, as long as it was interpreted as a threat toward the human observers. These threat behaviors included vocalizations such as hissing, spitting, growling (with ears forward or down and back), deliberate approaches or charges toward the observers, or a confrontational stance. A confrontational stance consisted of the puma quickly moving a short distance out from cover and in the direction of the observer, where it would stand stiff-legged in clear view of the observer. Results On 256 daytime occasions, we approached to within m (median = 30.8 m) of pumas and visually observed their responses. We made approaches on adult females with nursing or weaned cubs (n = 117), solitary independent females (n = 77), solitary independent males (n = 46), breeding pairs (n=5), and weaned cubs in the absence of their mothers (n=11). Threat responses were exhibited during 16 approaches (6%). Because we observed no puma threat responses for approach distances of >50 m, we analyzed those approach distances separately from those <50 m. Approaches >50 m We approached pumas at distances of m on 84 occasions. In 60% of those cases, the puma left the area (Table 1). A similar number of observers were present when pumas stayed and watched or ignored the researchers (median = 1, mode=1, range=1 3 observers) as when pumas left (median = 2, mode = 1, range = 1 5 observers). Approach distances for pumas that left ranged from m (median = m), whereas approach distances for pumas that stayed and watched or ignored researchers ranged from m (median=120.0 m). We made only 4 observations at distances >300 m from the puma, and in 3 of those cases we could not determine whether the puma detected or reacted to the person s presence. Observers watched pumas for an average of 26 minutes (range = minutes) during each event. One observation may

4 4 Wildlife Society Bulletin 2005, 33(3):XX XX Table 2. Puma responses to human observers who approached to 2 50 m, San Andres Mountains, New Mexico, Median (range) Median (range) No. approach number of Puma response Puma class a approaches distance (m) observers Left Stayed b Threat Family (3 50) 2 (1 5) Solitary (2 50) 1 (1 8) Breeding pair 3 50 (10 50) 1 (1 2) Cubs 8 20 (6 50) 2 (1 3) Total (2 50) 2 (1 8) a Family includes F with nursing or weaned cubs; solitary includes adults and subadults; cubs includes weaned cubs weeks old and in the absence of their mother. b Includes 5 observations when puma either ignored or did not detect human observer(s). have been misclassified as staying when the puma actually may have approached the observer. A single observer approached from below and to within 80 m of a radiocollared puma mother;over the next 68 minutes, she obtained a visual of the mother and 2 cubs. Visual contact with the mother was then lost, and her radiosignal became active. Because the signal became louder, with the peak signal pointing 90 degrees from the cubs location (and just down a thickly vegetated draw from the observer), the observer moved to higher ground, away from the cubs and mother. Approaches <50 m We approached to within 2 50 m of pumas on 172 occasions (Table 2). In 38 of 114 approaches where the puma left without giving a threat response, it stayed for minutes after visual contact with observer(s) prior to leaving. In 42 approaches the puma stayed where it was but did not give a threat response. Sixteen approaches (9%) Table 3. resulted in a threat response. From 1 5 observers (median = 2, mode = 2) were present when threat responses were elicited, whereas 1 8 observers (median= 2, mode = 1) were present during occasions when pumas left or stayed without exhibiting threats (Table 3). Pumas threatened single observers during 3 of 70 approaches (4%), 2 observers during 9 of 73 approaches (12%), and 3 or more observers during 4 of 29 approaches (14%). Threats occurred at a median approach distance of 18.5 m (range=2 50 m), whereas pumas that left or stayed without exhibiting threats were approached to a median distance of 20.0 m (range=3 50 m). Families. Mothers with cubs exhibited 14 threat responses during 79 approaches (Table 2). Since we often approached females with nursing cubs with the intention of capturing and marking cubs, we examined behaviors of mothers that were with nursing and weaned cubs separately. In 33 of 60 (55%) approaches to nurseries, mothers left without threat, often after watching observers for several minutes. We subsequently captured and marked cubs on 32 of these occasions. During 16 other approaches (27%), females remained at nurseries without giving a threat. We observed these families for an average of 73 minutes from a distance of m (median=17.0 m). Cubs nursed, played, or slept, and females sometimes closed their eyes for short periods. Eleven approaches of puma nurseries (18%) elicited a threat response from mothers (Table 3). We approached nursing females that gave a threat response similar distances to those that were not. Characteristics of 172 approaches to 2 50 m on different puma classes that resulted in a no-threat/threat response, San Andres Mountains, New Mexico, Median Median Mean No. Mean cub age approach No. duration Puma class approaches (weeks) distance (m) observers (min) Family: Nursing cubs 49/11 4.5/ /19.7 2/2 47.2/67.5 Weaned cubs 16/3 21.4/ /6.0 2/2 18.9/7.0 Solitary male a 32/1 15.8/2 1/1 5.8/<1 Solitary female a 49/0 20.2/NA b 1/NA 8.2/NA Breeding pair 2/1 30.0/50 1.5/1 4.0/3.0 Cubs c 8/0 33.0/NA 20.0/NA 2/NA 1.6/NA Total 156/ /18.5 2/2 22.1/49.0 a Includes adults and subadults. b NA = Not applicable. c Includes weaned cubs weeks old and in the absence of their mother.

5 Puma response to researcher approach Sweanor et al. 5 During 24% of 172 occasions when researchers approached to within 50 m, the puma remained in place without giving a threat response. During 66% of 172 occasions when researchers approached to within 50 m, the puma left the area without giving a threat response. Median numbers of observers during threat and nonthreat responses also were similar (Table 3). In all 11 instances when the puma mother gave a threat response, she hissed or growled; in 2 instances she hissed and growled. Three mothers immediately fled the nursery after vocalizing. On 5 other occasions, observers watched family groups for minutes (median=131 minutes),but the mothers never left the nurseries. In 2 cases, mothers charged to within 5 m of observers (one with ears up and forward, one with ears down and back) before being diverted by the observers actions: standing erect, yelling, and beating sticks, telemetry antennas, and a catch-pole against vegetation and rocks. Both females subsequently fled over 200 m. In the final case, the mother moved upslope at a brisk walk toward 2 observers while growling. The 2 observers moved adjacent to one another,and the mother stopped her advance. She turned and walked away, stopping twice to sit and gaze at the observers and toward her cubs. She eventually moved about m away while her cubs bedded about 70 m from observers. We observed no threats 16 of 19 times (84%) we approached females with weaned cubs (Table 3). Of 16 approaches, mothers left 13 times, 2 mothers stayed, and 1 exhibited no observable response. Three other approaches (16%) resulted in threats. Females with weaned cubs did not give a threat response until we were 5 16 m away (median=6 m). In all 3 threats, mothers of weaned cubs vocalized by growling or hissing. In the first case, the cubs scattered when they saw human observers, but the mother sat up and growled. Observers immediately walked away from the site. In the second case, 1 of 2 observers accidentally approached to within 6 m, apparently without detection. Noise from the first observer alerted the family, and the cubs immediately fled from the area. The mother puma charged out from behind the rock outcrop where the family had been bedded, stopped in a confrontational stance 5 m from the first observer (the second observer was about 5 m upslope from the first), and vocalized by spitting and hissing. Observer 1 stood facing the puma while observer 2 backpedaled upslope about 7 m. The puma quickly averted her gaze, turned, and disappeared behind the rock outcrop. Within 10 seconds the puma reappeared about 20 m upslope from the observers and was trotting away from them. When she was 40 m away, the trot turned to a walk and the puma disappeared over a ridge. In the third case, a single observer approached a family group at a mule deer (Odocoileus hemionus) cache site. The cubs remained at the deer carcass, but the mother puma came toward the observer from behind the carcass, growling and hissing. The observer started backing up, but the puma slowly kept coming in a walking gait with its body in a partial crouching posture. When the puma was only 3 m away, the observer yelled and waved his arms. The puma did not react. The observer then fired a.22 magnum pistol in the air; the puma still did not react, but kept advancing. The observer stepped up on a ledge about 0.6 m up from the canyon bottom and backed upslope. When the puma reached the ledge, she stopped and looked around while the observer continued backing up. When the observer could no longer see

6 6 Wildlife Society Bulletin 2005, 33(3):XX XX the puma and thought he was at a safe distance, he turned around and cautiously walked away. Solitary pumas. We approached solitary pumas to within 50 m on 82 occasions (Table 2). They included independent males (n = 26 adults, 7 subadults) and independent females (n =33 adults, 16 subadults). We approached solitary males that did not give a threat response to 3 50 m (median= 15.8 m), whereas we approached solitary females to 4 50 m (median = 20.5 m). Only 1 approach resulted in a threat (Table 3). An adult male puma hissed, then fled after a single male observer approached to within 2 m of his bed site. The observer had positioned himself about 10 m above the puma s bed for 70 minutes without obtaining a visual; he then moved down and into the rock outcrop where the male puma was bedded, finally causing the puma to react. The puma ran about 250 m to another rock outcrop and stopped. Breeding pairs. In 2 of the 3 cases in which we approached to within 50 m of breeding pairs, no threats were exhibited (Table 2). In both cases, the females ran several meters to denser cover and stopped. One male remained stationary. But after making eye contact with 2 observers, the other male slowly rose, walked 15 m away from the observers, lay down with his back to the observers, and appeared to fall asleep. In 1 approach,the male of the breeding pair gave a threat response. In this case the single observer made eye contact with the female puma but could not see the male. The observer then whistled, imitating a call he had heard made by an orphaned cub. The male puma immediately charged out from behind a rock outcrop beside the female and onto a rock pillar about 1 m above and in front of the female. The male made a confrontational stance and subsequent eye contact with the observer from about 50 m. The observer took 1 step upslope, diagonally toward the male puma, so that he could easily be seen. The male puma immediately turned around and ran upslope away from the observer. The female continued to watch the observer, so he took a few steps toward her. She turned around, slowly crept up a rocky chute, gained the canyon rim, and ran in the same direction as the male. Cubs without mother. Of 8 weaned litters approached when their mother was away (Table 2), 6 left and 2 remained at their bed sites. None of the cubs gave threat responses. Cubs ranged in age from weeks (x - =33 weeks). In 2 instances the cubs (a single 46-week-old female,and a litter of 3 52-week-old cubs) slowly walked away from 2 observers when they were m distant; in 4 other instances the cubs ran from 2 3 observers who approached to 6 50 m. The 2 remaining litters were approached to 6 20 m by 1 observer. These cubs simply moved into denser cover at the site. Variation among and within individual pumas We approached and observed each of 25 independent, radiocollared male pumas (either solitary or in breeding association) 1 7 times (median = 2, mode = 1), each of 42 independent, radiocollared female pumas (solitary, breeding, or with cubs) 1 12 times (median=4, modes=1, 2), and each of 8 radiocollared cubs (in the absence of their mother but along with uncollared siblings) 1 2 times (median=1, mode=1). Each of the 67 independent pumas also had been captured and handled 1 6 times (median = 2, mode = 1). Fifty-three independent pumas (23 males, 30 females) never gave a threat response even though each had been captured 1 6 times (median=2, mode=1) and visually observed 1 12 times (median = 2, mode = 1). We captured 8 of these pumas only once and subsequently approached them on only 1 occasion. Fourteen independent pumas (2 males, 12 females) exhibited threats on 16 occasions. We captured each of these 14 pumas 1 6 times (median= 2, mode=2) and visually observed them 1 11 times (median=5,modes=1,2,8). Each puma gave its first threat after being caught 1 4 times (median = 2, mode =1) and after being visually observed without threat on 0 6 (median=1.5, mode=0) other occasions. We captured 2 pumas only 1 or 2 times. They then gave threat responses during the 1 subsequent occasion they were each visually observed. Two other females gave threat responses during 2 different approaches. Puma F6 gave threats during the third and eleventh of 11 approaches, and puma F91 gave threats during the fifth and seventh of 8 approaches. Five of the 14 pumas that gave threats did so the first time we approached, and 2 pumas gave threats the last time we approached (out of 7 and 11 approaches, respectively). However, we approached 9 of the 14 pumas that gave threats on subsequent occasions when they did not give a threat. Puma F6 s behavior during several approaches demonstrated variation in within-individual response. On 1 occasion we observed F6 from 100 m and she gave no detectable response. We also

7 Puma response to researcher approach Sweanor et al. 7 observed her from <50 m on 10 occasions: 4 times with nursing cubs, 3 times with weaned cubs, and 3 times without cubs. When with nursing cubs, she responded to observers by leaving (n=2),staying in place (n=1), or giving a threat (n=1). With weaned cubs, she responded by leaving (n = 1), staying in place (n=1), or giving a threat (n=1). When she was solitary, she responded by leaving (n = 2) or staying in place (n=1). Although she gave a threat when we approached her and her weaned cubs in September 1987, she and her family responded to an approach 1 month later by quietly slinking away. In 2 attempts to capture her cubs from a subsequent litter in August 1990, she either moved away or remained by the nursery. We captured 1 cub in each attempt. In the second case, F6 remained within 15 m of the cub during the entire handling process, even though she observed the cub being captured. In an attempt to mark a new litter in December 1991, 2 observers tried to coax F6 to leave her nursery. We located the nursery in the bottom of a draw at the base of a rock wall and behind boulders. One observer dangled a long pole over the rim and jostled a bush in front of the nursery while the second observer watched from about 30 m away on the opposite side of the draw. The disturbance resulted in F6 charging the observer with the pole. F6 leaped up the rock wall and ran toward the person while making continual eye contact and with ears perked up and forward. The person faced the advancing puma, yelled, and thrashed low bushes in her direct path with the pole. When F6 was about 3 m away, she abruptly turned at a right angle and ran away upslope. Discussion Our data represent puma responses to humans in 1 puma population in a remote mountain range in southern New Mexico. Clearly, the majority of pumas we studied avoided direct human encounters:65% of pumas left when people approached to m,and 66% of pumas left in encounters <50 m. Furthermore, only 6% of approaches resulted in threats by pumas, and no puma initiated an encounter or attacked a person. In general, pumas avoided close encounters with people or they remained still, as if hidden. Clearly, the pumas did not consider us potential prey as we approached. In North America 16 fatal and 92 nonfatal puma attacks on humans were documented between (Beier 1991, Fitzhugh et al. 2003). In contrast to our research, pumas probably initiated most of the encounters that resulted in those attacks. Consequently, pumas we approached were probably placed in a more subordinate position, and the threats they gave likely were more defensive rather than aggressive in nature. The observers dominant behavior (we instigated the encounters) as well as careful approach tactics may have reduced the chance of a threat. The negative, and possibly traumatic, experience that pumas in our study had with humans during capture events also may have contributed to the general avoidance reaction of pumas toward people during approaches. However, the number of times a puma was captured (with the idea that capture might deliver some degree of aversive conditioning) did not appear to affect the probability of a threat response during an encounter. Pumas that gave threats were captured and handled a similar number of times as those that did not give threats. Still, it is possible that pumas that have never been captured and handled might be more likely to give threats than those that have;it is also possible that intentional aversive conditioning during captures will result in a reduction in subsequent threat responses. We could not test these hypotheses because we could not quantify responses of noncaptured pumas and we tried to minimize capture trauma. The threshold distance for a threat response was <50 m; however, approaching to within that distance did not guarantee that a threat would occur. Only 9% of approaches to within 50 m resulted in a puma threat response. In fact, on 39 of 46 occasions (85%) when observers approached pumas to within 10 m, there was no resulting threat. Additionally, of the pumas that gave threats, 2 probably did not initially know that they were responding to human intruders. An adult male puma made a confrontational stance when he heard the observer whistle. It was possible the puma thought there was a male competitor nearby and rushed out to defend his mate. When the observer stepped out and became more visible, the puma fled. A mother puma made a confrontational stance after she and her cubs detected noise above and behind them. After a brief moment (time for her to recognize the intruder as human), she retreated. Consequently, only 14 of 172 approaches to within 50 m (8%) resulted in a threat response purposefully directed at humans. Aggressive, dominant behavior by people may deter a puma attack. In cases where a puma is sur-

8 8 Wildlife Society Bulletin 2005, 33(3):XX XX prised, subsequent identification of the intruder as human also may deter a possible attack. In this study, pumas confronted, advanced on, or charged the observer(s) on 6 occasions. In all 6 cases, the observers helped avert a possible attack by making themselves more identifiable, standing erect and beating sticks or other objects against bushes or rocks, making continuous noise, banding together, or backing away to a higher (possibly more dominant) position. Beier (1991) reported that shouting loudly seemed to avert or repel attacks; Fitzhugh et al. (2003) emphasized that noise was most effective if it was loud and continuous. Female pumas accompanied by cubs are probably the most dangerous class of puma to approach (18% of close approaches resulted in threats). The likelihood of a threat response was similar for females with nursing (18% of close approaches) and weaned (16% of close approaches) cubs. We interpreted their threats to be in defense of their cubs. Females with cubs also were more likely to stay (42% stayed) than solitary pumas and cubs in the absence of their mother (24% and 25% stayed, respectively). Although 33% of 3 close approaches (i.e., <50 m) of breeding pairs resulted in a threat response by 1 of the pair, sample size was extremely small, and the 1 puma that exhibited a threat apparently withdrew as soon as he identified the intruder as human. Cubs in the absence of their mother may be the least dangerous class to approach (0 of 8 close approaches resulted in threats). But again, sample size was very small. Threat responses were most often given when >2 observers were present. This may result partly from greater efforts to capture or visualize cubs when we had more personnel. Solitary observers tended to be more cautious and did not usually attempt to get as close to nursery or bed sites as multiple observers. This contrasts with cases in North America where people did not deliberately initiate the encounter; under those conditions, solitary people were 3 times more likely than pairs or larger groups to have a puma encounter or be attacked (Fitzhugh et al. 2003). Variation in response between and within individuals was evident. Some individuals never showed a threat response, even though they were approached multiple times; others gave threat responses during the first and only approach. Still other pumas were aggressive in some instances but not in others. We could not predict whether a puma that behaved one way during an approach would behave in a similar way during a subsequent approach. Regardless, most pumas showed great restraint, considering that human observers (often singly) initiated the encounters. The majority of pumas left the area when they were closely approached. Many other pumas remained in place without exhibiting threat behavior. Sometimes pumas appeared to relax after tense moments associated with initial human-to-puma eye contact. Others tended to ignore the observers, or appeared to react as though observers could not see them. These pumas were living in a wild environment with a relatively low chance of encountering humans (except for the puma researchers) compared to pumas living in more human-developed habitats. We do not believe that their staying behavior was a result of habituation toward humans and human activity. We suggest that staying behavior is one of a number of natural responses a puma might exhibit when approached by a human as a tactic to avoid direct contact or, in the case of mothers, to stay in defense of cubs. Also, the puma may perceive that they are hidden, or they have not been seen. Thus, staying behavior does not necessarily indicate a puma has become habituated or lost fear of humans. Using staying behavior as the sole criteria to determine whether a puma is a public safety threat may not be appropriate. Moreover, pumas we studied showed varying behavior, staying on some occasions, fleeing on others. Our observations provide further insights into puma behavior and increase our understanding of the variation in behaviors expected from wild pumas. We encourage other researchers involved in intensive studies of wild pumas in habitats with variation in human development to record and quantify puma behaviors in relation to human activities. Managers can gain greater insights into puma behaviors that may or may not constitute threats toward people. Better yet,we recommend research focused on puma human interactions that identify the relative strength of factors that contribute to dangerous encounters between pumas and people. Our data have identified some influential factors, including puma sex, reproductive status, and approach distance, that can be used in a conceptual model to be tested through experimental research. Acknowledgments. Cooperators on the study included the Hornocker Wildlife Institute, New Mexico Department of Game and Fish (NMGF),

9 Puma response to researcher approach Sweanor et al. 9 United States Army at White Sands Missile Range, and the San Andres National Wildlife Refuge (SANWR). We are grateful to other project personnel who helped capture, radiotrack and observe pumas, including J. F. Smith, B. R. Spreadbury, T. K. Ruth, J. L. Cashman, and J. R. Augustine. We thank M. Weisenberger and S. Berendzen of SANWR for their observations of pumas. We thank K. M. Murphy, D. J. Mattson, and an anonymous referee for their suggestions for improving the manuscript. The puma study was funded by the NMGF and the Federal Aid to Wildlife Restoration Act, under Pittman- Robertson Project W-128-R. Publication of this paper was supported in part by WildFutures. Literature cited BARON, D The beast in the garden. W. W. Norton and Company, New York, New York, USA. BEIER, P Cougar attacks on humans in the United States and Canada. Wildlife Society Bulletin 19: COMMITTEE ON ACCEPTABLE FIELD METHODS IN MAMMALOGY Acceptable field methods in mammalogy: preliminary guidelines approved by the American Society of Mammalogists. Journal of Mammalogy 68 (4, Supplement). DEURBROUCK, J., AND D. MILLER Cat attacks. Sasquatch Books, Seattle,Washington, USA. ETLING, K Cougar attacks. The Lyons Press, Guilford, Connecticut, USA. FITZHUGH, E. L., S. SCHMID-HOLMES, M.W. KENYON, AND K. ETLING Lessening the impact of a puma attack on a human. Pages in S. A. Becker, D. D. Bjornlie, F. G. Lindzey, and D. S. Moody, editors. Proceedings of the Seventh Mountain Lion Workshop, Lander,Wyoming, USA. LEYHAUSEN, P Cat behaviour: the predatory and social behaviour of domestic and wild cats. Translated by B. A. Tonkin. Garland Press, New York, New York, USA. LOGAN, K. A., AND L. L. SWEANOR Desert puma: evolutionary ecology and conservation of an enduring carnivore. Island Press,Washington, D.C., USA. LOGAN, K.A., L. L. SWEANOR, J. F. SMITH,AND M. G. HORNOCKER Capturing pumas with foot-hold snares. Wildlife Society Bulletin 27: WHITTAKER,D.,AND R. L. KNIGHT Understanding wildlife responses to humans. Wildlife Society Bulletin 26: ZAR, J. H Biostatistical analysis. Second edition. Prentice Hall, Englewood Cliffs, New Jersey, USA. Andres Mountains, New Mexico and received her M.S. degree in wildlife sciences from the University of Idaho. She and her husband, Ken Logan (right), co-authored a book on the New Mexico research, titled Desert Puma: Evolutionary Ecology and Conservation of an Enduring Carnivore, which was awarded Outstanding Publication in Wildlife Ecology and Management honors by The Wildlife Society in From she led field research on puma human interactions in southern California for the Wildlife Health Center at the University of California, Davis. Linda is presently employed by Colorado Division of Wildlife and continues to be interested in studying puma behavior. Ken Logan has been studying pumas in the wild since His first project was on puma population characteristics and habitat use in the Big Horn Mountains, Wyoming, for which he earned a Master s degree in zoology and physiology from the University of Wyoming. Then he led the 10-year project on puma ecology on the San Andres Mountains, New Mexico while a researcher for the Hornocker Wildlife Institute. For that work Ken earned a Doctorate in wildlife sciences from the University of Idaho. From he was field research leader for the Southern California Puma Project for the Wildlife Health Center at the University of California, Davis. Presently, Ken is the carnivore researcher for the Colorado Division of Wildlife studying puma population dynamics on the Uncompahgre Plateau. Maurice Hornocker began studying pumas in 1964 in the central Idaho wilderness. Since then he and his students and colleagues have studied puma populations in several other regions in the western United States. He has also studied other big cat species around the world. He served as leader of the Idaho Cooperative Wildlife Research Unit at the University of Idaho from In 1985 he founded the Hornocker Wildlife Institute and served as its director until 2000, when the Institute merged with the Wildlife Conservation Society. He currently is a senior conservationist with the Wildlife Conservation Society. Linda Sweanor (left) was a wildlife researcher for the Hornocker Wildlife Institute from While with the Institute, she spent 10 years studying puma behavior and ecology in the San Associate editor: Krausman