Meredith J. Bashaw a, Angela S. Kelling b, Mollie A. Bloomsmith b & Terry L. Maple b a TECHlab, Zoo Atlanta, and Center for

Transcription

1 This article was downloaded by: [Dr Kenneth Shapiro] On: 09 June 2015, At: 07:05 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: Registered office: Mortimer House, Mortimer Street, London W1T 3JH, UK Journal of Applied Animal Welfare Science Publication details, including instructions for authors and subscription information: Environmental Effects on the Behavior of Zoo-housed Lions and Tigers, with a Case Study o the Effects of a Visual Barrier on Pacing Meredith J. Bashaw a, Angela S. Kelling b, Mollie A. Bloomsmith b & Terry L. Maple b a TECHlab, Zoo Atlanta, and Center for Reproduction of Endangered Species, Georgia Institute of Technology b TECHlab, Zoo Atlanta, and Center for Conservation and Behavior, Georgia Institute of Technology Published online: 05 Dec To cite this article: Meredith J. Bashaw, Angela S. Kelling, Mollie A. Bloomsmith & Terry L. Maple (2007) Environmental Effects on the Behavior of Zoo-housed Lions and Tigers, with a Case Study o the Effects of a Visual Barrier on Pacing, Journal of Applied Animal Welfare Science, 10:2, , DOI: / To link to this article: PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content ) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views

2 expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at

3 JOURNAL OF APPLIED ANIMAL WELFARE SCIENCE, 10(2), Copyright 2007, Lawrence Erlbaum Associates, Inc. ARTICLES Environmental Effects on the Behavior of Zoo-housed Lions and Tigers, with a Case Study of the Effects of a Visual Barrier on Pacing Meredith J. Bashaw TECHlab, Zoo Atlanta, and Center for Reproduction of Endangered Species Georgia Institute of Technology Angela S. Kelling, Mollie A. Bloomsmith, and Terry L. Maple TECHlab, Zoo Atlanta, and Center for Conservation and Behavior Georgia Institute of Technology Tigers and lions in the wild are nocturnal nonhuman animals who may hunt and mate opportunistically during daylight hours. In captivity, they spend most time on exhibit sleeping or pacing. To better understand their activity budget, this study examined the daily behavior patterns of 2 Sumatran tigers and 3 African lions in different housings. The proportion of scans the large felids spent engaged in stereotypic pacing varied by time of day and environment. The tigers spent different amounts of time pacing when housed in different exhibits; the lions paced more in off-exhibit housing than when on exhibit. These differences suggest changes to the cats immediate housing environment may decrease pacing but provide little insight into altering specifics. Carnivores pacing relates to their inability to control sensory access to social partners. Both environments with increased pacing contained chain-link fencing. allowing Correspondence should be sent to Meredith J. Bashaw, Department of Psychology, Franklin & Marshall College, P.O. Box 3003, Lancaster, PA mbashaw@alumni.duke.edu

4 96 BASHAW, KELLING, BLOOMSMITH, MAPLE uncontrolled sensory contact. Where the tigers paced, the study placed a visual barrier between one female and keepers or conspecifics cues. This did not significantly decrease pacing. However, the study suggests considering sensory access and environmental variables when designing environments for captive carnivores. Although at least one species of large felid is displayed in almost every zoo, big cat exhibits can be frustrating and disappointing for zoo visitors (Carlstead, 1998). Most modern zoos use naturalistic exhibits in an attempt to improve animal well being and visitors perceptions of, and respect for, animals (Finlay, James, & Maple, 1988). However, these exhibits contain visual barriers that, combined with the natural camouflage of these animals, may make the animals difficult for visitors to find within the exhibit (Bashaw & Maple, 2001). This problem is exacerbated by the animals inactivity during peak visitor hours (10:00 to 15:00; Bitgood, Patterson, & Benefield, 1988) and their tendency to develop pacing stereotypies, which visitors interpret as indicative of stress or boredom (Carlstead, 1998). Indeed, stereotypic behaviors, including pacing, are often used as indicators of compromised well being (Broom, 1991). Numerous authors have documented within-subjects differences in behavior when animals are housed in different environments. Among the housing variables that have been demonstrated to influence animal behavior are the following: 1. Whether the enclosure is indoors or outdoors (Hoff, Forthman, & Maple, 1994), 2. The social density of animals (Tenneson, 1989), 3. The presence of objects in addition to substrate (Ogden, Lindburg, & Maple, 1993; Wilson, 1982), 4. The amount of usable surface area (Neveu & Deputte, 1996; Perkins, 1992), 5. The number of visual barriers, retreat spaces, or hiding places (Anderson, Benne, Bloomsmith, & Maple, 2002; Carlstead, 1991; Mellen, Hayes, & Shepherdson, 1998), and 6. The presence of visitors (Mitchell et al., 1991). ErwinandDeni(1979),Stevenson(1983),andCarlstead(1998)areonlyafewofthe authors to suggest that stereotypic behaviors, including pacing, are a result of an abnormal organism environment interaction (Carlstead, 1998, p. 172). A comparison of behavior of the same animals in different environments may allow identification of housing variables that affect behavior. These variables may then be experimentally manipulated to determine whether they cause observed behavior differences. Wild tigers (Panthera tigris) and lions (Panthera leo) are nocturnal animals. Tigers are truly nocturnal, mainly active between 19:00 and 05:00 each day (Sankhala, 1977; Schaller, 1972); lions activity peaks are after 17:00 and before

5 EFFECTS OF A VISUAL BARRIER ON PACING 97 08:00 (Schaller, 1972). However, both lions and tigers will kill and mate opportunistically and have been observed engaged in both these activities during daylight hours (Sankhala, 1977; Schaller, 1972). The flexibility of lions and tigers to become active during the day likely accounts for daytime activity under captive conditions when animals are shifted between environments and when feedings occur during the day. Increases in the frequency of feeding small felids have increased behavioral diversity and exploratory behavior, while reducing duration and frequency of stereotypic pacing (Shepherdson, Carlstead, Mellen, & Seidensticker, 1993). At Zoo Atlanta, feeding and shifting the lions and tigers occur at the beginning and end of the zoo s visitor hours, around 09:00 and 17:00. We first examined patterns of behavior for Zoo Atlanta s two Sumatran tigers and three African lions by time of day, as well as in different housing situations to identify potential variables that may influence activity and pacing. We hypothesized that there would be differences in behavior of the same animals in different housing conditions and that activity budgets of all large felids would vary by time of day, with the large felids most active in the morning and late afternoon. Method OBSERVATIONAL STUDY Subjects. Subjects in the observational study were one male and two female African lions and one male and one female Sumatran tigers. All animals were reproductively viable except one immature (2-year-old) female lion. For more information on these animals, see Table 1. All large felids were fed Nebraska feline diet Species Sex Name TABLE 1 Detailed Information About the Study Subjects Age at Study Start Birth and Rearing Reproduction African lion Female Kalihari 6 years Wild-born, hand-reared African lion Female Kariba 2 years Captive-born, hand-reared African lion Male Zulu 9 years Captive-born, not recorded Sumatran tiger Female Sekayu 12 years Captive-born, parent-reared Sumatran tiger Male Jalal 6 years Captive-born, parent-reared None None Sired cubs Dam of cub Sire of cub

6 98 BASHAW, KELLING, BLOOMSMITH, MAPLE 6 days each week and, on the 7th day, were given bones to simulate their irregular eating pattern in the wild and maintain their dentition. During this study, the two female lions were housed with full access to each another, whereas the male lion and both tigers had visual and olfactory but not tactile access to conspecifics. All large felids were shifted into their daytime environment at about 09:00 and remained there until shifted again at 17:00. For the rest of the day, all animals were housed in concrete-floored indoor outdoor areas with no public access. Enrichment items, including hanging logs, scents, boomer balls, empty plastic drums, and hay piles, were added to both indoor and outdoor enclosures on a rotating schedule. During the day, tigers were housed in one of two outdoor areas where they were on display to the public (hereafter referred to as on exhibit ). One of the tiger exhibits (hereafter the large exhibit ) had an area of 957 m 2 and consisted of a gently sloping hill with a stream running through the middle, a pool near the glass separating the exhibit from the public, and a bamboo thicket and two large bushes that provided visual barriers for the tigers from both conspecifics and humans. The second exhibit (hereafter the small exhibit ) had an area of 149 m 2 and had a similar pool, a raised ledge along the back, steep downhill slopes, and no visual barriers. The lions were are also housed in two areas: (a) an outdoor area where they are on display to the public (on exhibit) which was 1,073 m 2 and consisted of a grassy hill, a large rock structure that provided shade, a visual barrier, and opportunities to lie in the sun; and a pool near the glass that separated the exhibit from the public. On alternate days, lions were housed either on exhibit or (b) in a smaller, concrete-floored indoor outdoor area surrounded by mesh fencing and with no public access (in holding). The holding area was divided into six interconnected areas: four indoor rooms (12 m 2,24m 2,17m 2, and 17 m 2, respectively) and two outdoor patios (31 m 2 and 22 m 2, respectively). During the observational study, lions in the holding area had access to all six rooms unless the animal-care staff was cleaning, in which case lions were excluded from the room in which the staff member was working. Data collection. Each cat was observed for 1-hr sessions with instantaneous scans of the behavior of all individuals in a housing condition at 1-min intervals (Martin & Bateson, 1986). Behavioral categories analyzed included the following: 1. Resting (lying down or sitting with eyes closed and no movement), 2. Resting but awake (lying down or sitting with eyes open and remaining alert), 3. Pacing (moving around the enclosure on a set path, with at least two repetitions), and

7 EFFECTS OF A VISUAL BARRIER ON PACING Nonstereotypic activity (any behavior other than rest, rest/awake, stand, or pace). Data recording sessions occurred between 10:00 and 12:00, between 12:00 and 14:00, and between 14:00 and 16:00. All lions were observed for equal amounts of time on exhibit and in holding, and both tigers were observed for equal amounts of time in each of the two exhibits. Data were collected from the public viewing areas when the animals were on exhibit and from keeper access areas when they were in holding. Observations were made by 11 data collectors with an index of concordance of at least 88.5% in simultaneous data collection with the primary experimenter. A total of 540 hr of data were collected over all subjects and locations. Data analysis. Because data were not normally distributed, the number of subjects was small, a repeated-measures design was used, and nonparametric statistics were employed for evaluation of the data (Runyan & Haber, 1980). The Friedman s test, the nonparametric equivalent of a repeated-measures analysis of variance, was used to evaluate differences in behavior by time of day. Mann Whitney U tests, nonparametric t-test equivalents that do not assume normally distributed data, were used to establish differences in behavior between the different environments. In all tests, the p <.05 criteria was used to establish statistical significance. Results For all large felids, there was a significant (F = 6.14, p =.046) difference in resting by time of day, with more time spent resting in the midday (12:00 14:00) and afternoon (14:00 16:00) sessions than in the morning (10:00 12:00) session (see Figure 1). For nonstereotypic activity, the opposite trend was evident (F = 5.7, p =.058), as more time was spent active in the morning session than in the midday or afternoon. These changes in resting and activity were consistent for each individual cat and for both species. Large felids also showed consistent differences in behavior as a function of their daytime housing. Lions spent a significantly greater proportion of their time (a) resting on exhibit than off exhibit (U = 9.0, p =.05), and (b) a significantly smaller proportion of their time resting but awake on exhibit than off exhibit (U = 9.0, p =.05; see Figure 2). Lions also paced for a significantly greater proportion of time off exhibit than on exhibit (U = 9.0, p =.037). In fact, pacing was never observed on exhibit, whereas it occurred in 7% of scans off exhibit. For tigers, although less nonstereotypic activity and more pacing were exhibited in the small exhibit than at the large exhibit,

8 100 BASHAW, KELLING, BLOOMSMITH, MAPLE FIGURE 1 Mean behavior for all lions and tigers by time of day for the observational study. The filled star indicates a significant difference in the proportion of scans spent in that behavioral category, whereas the open star indicates a trend toward a significant difference. FIGURE 2 Mean behavior of lions by enclosure for the observational study. These data are averaged for three lions and all times of day. Stars indicate significant differences in the proportion of scans spent in that behavioral category. the results were not significant (for activity, U = 1.34, p =.180; for pacing, U = 4.0, p =.121; see Figure 3). Discussion As predicted, the large felids activity budgets varied by time of day. The large felids were most active in the morning and late afternoon and spent the most time resting at midday. These activity patterns have implications for the

9 EFFECTS OF A VISUAL BARRIER ON PACING 101 FIGURE 3 Mean activity level, stereotypic and nonstereotypic, of tigers by enclosure for the observational study. These data are averaged for two tigers and all times of day. experience of zoo visitors. Visitors tend to find active animals more interesting, as indicated by greater attention to an active animal s behavior (Bitgood, Patterson, & Benefield, 1986, 1988) and a longer time spent at exhibits when animals are active (Jackson, 1994). Visitors tend to interpret high levels of inactivity and stereotypic pacing as indicative of stress or boredom (Carlstead, 1998). The midday peak in inactivity supports Bitgood et al. s (1988) cross-species study, which suggested large felids were inactive much of peak zoo-visitor hours (between 10:00 and 15:00), and emphasizes the importance of providing information about their activity patterns to zoo visitors. Added interpreters or interactive graphics could engage and educate visitors about why tigers are inactive at midday and suggest better visiting times. In addition, the behavior of the same animals was different in different housing conditions. These differences in behavior, and particularly stereotypic pacing, between exhibits support Brooms s (1991) conclusion that changes to the environment had the potential to improve well being and create a more enjoyable and educational experience for the zoo visitor (Finlay et al., 1988). The presence and number of visual barriers in exhibits have been identified as the most significant predictor of proportion of time spent pacing in small cats, with an increased number of visual barriers predicting a decrease in time spent pacing (Carlstead, Brown, & Seidensticker, 1993; Mellen et al., 1998). Mellen et al. found that amount of time spent pacing in small felids varied more with the number of visual barriers than with any other variables that they measured. It is not surprising that animals who are solitary in the wild may exhibit behaviors indicative of stress when unable to distance themselves from each other in captivity (Carlstead, 1998). Visual barriers may allow animals to better adjust their perceived distance from conspecifics or other sources of social stimulation. For example, the placement of a visual barrier between a red panda (Ailurus fulgens) and the panda s own reflection

10 102 BASHAW, KELLING, BLOOMSMITH, MAPLE greatly reduced pacing in this solitary animal (Bouwens, 1999). Anecdotal evidence suggests that both tigers and lions at Zoo Atlanta tend to pace in areas where another cat or a human could be seen through a mesh or chain-link barrier (Bashaw, 2000). One of the differences between the environments in the observational study is the presenceofthesetypesofbarriersbetweenthelionsandtigersandtwosourcesofsocial stimulation: keepers (Mellen, 1991) and conspecifics. It is possible that altering thesemeshbarrierscouldbeaneasyandinexpensivemeanstoreducepacing.acase study was conducted to evaluate whether the addition of a visual barrier between a tiger and social stimuli would reduce pacing. Method CASE STUDY Subjects. As a result of animal management between experiments, only the one female tiger (Sekayu) was available for study in this experiment. She was housed in the larger exhibit (described previously) for the duration of the study. In all other respects, her management was as described in the observational study. Manipulation. Behavioral data were collected before and after the attachment of an opaque, corrugated plastic barrier outside the mesh fencing at the back of the tiger exhibit. This barrier eliminated the subject s visual access to conspecifics or keepers present in the tiger area but was not itself accessible to the tiger. The phases of the study are referred to as (a) the baseline 1 condition (25 days prior to placement of the barrier), (b) the experimental condition (24 days while the barrier was present), and (c) the baseline 2 condition (38 days following the removal of the barrier). Data collection. Previous studies have demonstrated that the duration and frequency of pacing may vary independently (Shepherdson et al., 1993). The observation of only a single individual made it possible to keep more detailed records of behavior that would include both duration and frequency information, so continuous focal data were collected using the Observer software package. The same behavioral categories were analyzed as in the observational study, except that behavior directed toward the fence to which the barrier was attached was recorded as its own category: barrier-directed behavior. Based on the activity patterns identified in the observational study, we elected to narrow the observation periods to 90-min windows and disperse them more broadly across the day. We observed between 09:00 and 10:30 (morning), 12:00

11 EFFECTS OF A VISUAL BARRIER ON PACING 103 and 13:30 (midday), and 15:30 and 17:00 (afternoon) to attempt to capture periods of higher activity while maintaining comparable times of day. One 60-min focal session was collected within each 90-min window 3 days a week, until 10 observations were conducted in each time period for each condition; however, only 9 observations occurred between 15:30 17:00 in the experimental condition. Observations were made by four data collectors with index of concordance of 85% or more in simultaneous data collection with the primary experimenter. Data analysis. To evaluate changes in behavior, confidence intervals (CIs) were computed from the data collected in both baseline conditions for each behavioral category (Bard & Nadler, 1983; Hays, 1994; Tarou, Bashaw, & Maple, 2000). To compensate for computing multiple intervals, 99% confidence intervals were used (Hays, 1994). The mean percentage of scans in which each behavior was observed in the experimental condition was compared to the confidence intervals; means that fell outside the intervals were considered statistically significant. The two baselines were also compared to evaluate the consistency of behavior among the study periods and combined where they did not differ. We report both the mean and confidence intervals where appropriate. To provide validation of consistency with the observational study, data were also visually examined for effects of time of day on behavior. Results Condition. The frequencies of events in each behavioral category number of times per hour behavioral category occurred in the two baseline conditions were identical, so confidence intervals encompassing both baselines were used to compare to the experimental phase. The experimental mean, direction of change, baseline mean, and confidence intervals for the frequency of occurrence of events in each behavioral category are presented in Table 2. In addition to these changes in the frequency of behavior, the tiger also moved into the portion of the exhibit containing the barrier more often in the experimental condition (M exp = 4.1, CI = 5.4 > x > 2.9). Duration of events in each behavioral category differed significantly between baselines, with time engaged in nonstereotypic activity in baseline 2 greater than in baseline 1 (M BL2 = 22.3, CI BL1 =19.3>x>9.5). The baselines were therefore not combined for comparison to the experimental condition. Duration of behavioral categories (measured as a percentage of time for the 60-min session) in the experimental condition was the same as baseline 1, differing significantly from baseline 2 only in that a smaller amount of time was spent in barrier-directed

12 104 BASHAW, KELLING, BLOOMSMITH, MAPLE TABLE 2 Comparisons of Frequency of Behaviors in the 60-Min Sessions for Case Study Between the Combined Baselines and the Experimental Conditions Behavior Mean of Experimental Condition Direction of Significant Change Mean of Combined Baselines 99% Confidence Interval Resting > x > 1.41 Rest but awake > x > 5.11 Pacing > x > Nonstereotypic activity > x > Inactive > x > Barrier directed > x > 3.64 Note. For the experimental condition, each mean representing a statistically significant increase is followed by a +, each statistically significant decrease is followed by a, and each mean indicating no change is followed by an 0. Confidence intervals were computed based on the combined data from both baseline conditions. behavior (M exp = 0.7, CI BL2 =3.2>x>0.8). There was no change in the amount of time the tiger spent in the area of the exhibit containing the barrier. Time of day. Although the data cannot be statistically compared because of the single-subject design and slight differences in the observation times, visual inspection of the pattern of activity levels among the three time periods (morning, midday, and afternoon) appeared similar to the observational study. Pacing was lowest in the midday sessions and approximately equal in the morning and afternoon sessions. This pattern of pacing was accompanied by the opposite trend in resting, with the midday sessions showing the highest percentage of resting behavior. Overall, the majority of the time in the midday session was spent resting or resting but awake, whereas in the other two sessions, the majority of the time was spent pacing. Discussion Although the addition of a visual barrier between a tiger and the potential social stimulation provided by keepers and conspecifics did result in decreases in frequency of resting but awake and both frequency and duration of barrier-directed behavior, it did not reduce duration of pacing and actually increased frequency of pacing. Patterns of behavior by time of day were similar to those reported in

13 EFFECTS OF A VISUAL BARRIER ON PACING 105 the observational study. These results suggest that the mesh between the tiger exhibit and holding area was a less salient stimulus when the visual barrier was in place but that visual access to social stimuli was not primarily responsible for high levels of pacing observed in the tiger. The continued pacing in the area of the exhibit nearest the holding area suggests that the function of this area for the tiger did not change when the visual barrier was added. Lyons, Young, and Deag (1997) pointed out that edge areas of enclosures are important: The edges of a captive environment constitute an enforced territorial boundary [and are] a source of several forms of stimulation (p. 72). The area where most pacing occurred, where the barrier was placed, is the only location in the exhibit where the cat could engage in social interaction with keepers and conspecifics, gain access to food, and enter den areas. Although we effectively limited visual access to social stimuli, the addition of a visual barrier did not change the use of this area for other stimuli. Locomotor stereotypies in felids have been closely linked to feeding (Carlstead, 1998), so it is possible that pacing in this cat was more closely tied to feeding than to ability to distance herself from social stimuli. If this were the case, we would expect her pacing to decrease as a result of changes in feeding strategies, especially with feeding techniques in which food is available in other areas of the exhibit at more variable times. Indeed, previous experiments with feeding enrichment have been shown to reduce pacing in tigers at Zoo Atlanta (Bashaw, Bloomsmith, Marr, & Maple, 2003). Tigers rely on auditory and olfactory cues both in addition to and in the absence of visual cues for communication with conspecifics (Schaller, 1967). Lyons et al. (1997) found that the separation of conspecifics led to increased pacing along barriers for both cheetahs and snow leopards. They suggest that to reduce stress if conspecifics must be separated, it is best to keep them out of sight of, and unable to communicate with, each other. Because the barrier we chose to use restricted visual access but not auditory or olfactory access it is likely that the barrier did not effectively limit social stimulation from keepers and conspecifics. Future exhibit design could incorporate visual, auditory, and olfactory barriers between areas in which tigers are likely to be housed alone to reduce stress or excitement resulting from social stimuli transmitted between animals. GENERAL DISCUSSION Both physical differences in the environment and temporal patterns affected the behavior of zoo-housed lions and tigers. These differences provide a starting point for evaluating the causes of stereotypic behavior performed by these animals and have important implications for evaluating the welfare of captive animals. Carlstead (1991) found that exhibit cleaning, unexpected noises, and high numbers of zoo visitors led to increased stereotypic running in a fennec fox (Vulpes

14 106 BASHAW, KELLING, BLOOMSMITH, MAPLE zerda). The shared element of these factors was that each of them elicited an escape response that could not be fulfilled because the animal was enclosed. High visitor number has been found to alter behavior in primates (Mitchell et al., 1991), and both tigers were observed to pace along the back of the exhibits. It is possible, rather than tryingtoaccessfoodordenspaceontheothersideofthemesh,theyweretryingtoescape from visitor traffic at the front of the exhibit. However, this is unlikely for two reasons. If pacing were entirely a result of an attempt to escape from visitors, we also shouldhaveseenlesspacingbylionsintheholdingareathanonexhibit:thelionson exhibit are exposed to visitors, whereas the lions in holding are not. Instead, the opposite trend was apparent, suggesting visitor presence is not the only stimulus inducing pacing in these animals. In addition, the number of visitors at Zoo Atlanta is typically highest in the midday. Because pacing is higher in the morning and afternoon than at midday, it is more likely that pacing is related to events scheduled at this time including feeding, social contact from keepers, and access to den space. However, a multi-institutional examination in which environmental design and visitor numbers are independently manipulated would clarify whether the opportunity towithdrawfromvisitorsispredictiveofpacinginbigcatsasinsmallcats(carlstead et al., 1993; Mellen et al., 1998). Rather than the social stimuli being a source of stress, perhaps it is the lions or tigers ability to control their access to social stimuli that could be related to pacing. Control over environmental events has been demonstrated to have a significant effect on an organism s response to both appetitive and aversive events (Bashaw, 2003). In both the lion-holding area and the small tiger exhibit, large felids were almost entirely unable to control access to social stimuli; keepers and/or conspecifics were on the other side of a permeable mesh fence, and no retreat space was available in which these stimuli could be avoided. In placing the visual barrier, we eliminated visual access to the holding area, but again the tiger could not control the amount of social stimulation it received. The visual barriers evaluated by Mellen et al. (1998) were located within the exhibit, so the animals could use them as a retreat to control social access. Future studies should experimentally compare different types of barriers added to exhibits to determine whether barriers that allow large felids control over their sensory environment are more effective than total barriers like the one used in this study. CONCLUSION Stereotypic behavior has been used as a yardstick to measure the well being of animals in their current housing conditions. Although the relationship between well being and stereotypy is obviously not this simplistic (Mason, 1991), if the same animals show consistent differences in stereotypic behavior when housed

15 EFFECTS OF A VISUAL BARRIER ON PACING 107 in different environments, it is likely that features of those environments are related to the levels of stereotypy observed. In this case, the prevalence of less pacing and more naturalistic behavior indicates that the lions are housed in a more complex environment on exhibit than in holding and that the larger tiger exhibit may be more appropriate than the smaller one. Our attempt to manipulate one variable (visual social stimuli) that differed between environments by adding a visual barrier between a tiger and the holding area was unsuccessful at reducing pacing. Access to olfactory or auditory social stimuli or feeding areas may be more important than visual stimuli in influencing pacing. However, this study suggests that using a within-subjects design to compare environmental conditions may be an effective tool in evaluating the association between the physical environment in which an animal is housed and stereotypic behavior. ACKNOWLEDGMENTS Mollie Bloomsmith is now affiliated with Yerkes National Primate Center. Meredith Bashaw is now in the Department of Psychology at Franklin & Marshall College. A Charles T. Bailey Fellowship and a Presidential Fellowship provided support to the first author during the completion of this project. We thank the elephant/carnivore keepers and managers at Zoo Atlanta, without whose cooperation this research could not have been accomplished. We also thank Kate Alexander, Natasha Kennedy, Robin Keough, Sharon LeMeer, Melissa Loree, Jeanne Peters, Missy Snyder, Lorie Tarou, and Tiffany Willis, who each collected a portion of the data for the observational study. REFERENCES Anderson, U. S., Benne, M., Bloomsmith, M. A., & Maple, T. L. (2002). Retreat space and human density moderate undesirable behavior in petting zoo animals. Journal of Applied Animal Welfare Science, 5, Bard, K. A., & Nadler, R. D. (1983). The effect of peer separation in young chimpanzees (Pan troglodytes). American Journal of Primatology, 5, Bashaw, M. J. (2000). To hunt or not to hunt? A feeding enrichment experiment with captive wild felids. Unpublished master s thesis, Georgia Institute of Technology, Atlanta. Bashaw, M. J. (2003, March). Consistent effects of controllability of environmental events? Paper presented at the annual conference of the Southeastern Psychological Association, New Orleans, LA. Silver Spring, MD: Association of Zoological Park and Aquariums. Bashaw, M. J., Bloomsmith, M. A., Marr, M. J., & Maple, T. L. (2003). To hunt or not to hunt? A feeding enrichment experiment with captive lions and tigers. Zoo Biology, 22,

16 108 BASHAW, KELLING, BLOOMSMITH, MAPLE Bashaw, M. J., & Maple, T. L. (2001). A note on the failure of signs to increase zoo visitors ability to see tigers. Curator, 44, Bitgood, S., Patterson, D., & Benefield, A. (1986). Understanding your visitors: Ten factors that influence visitor behavior. In American Association of Zoological Parks and Aquariums 1986 Annual Conference Proceedings (pp ). Minneapolis, MN. Bitgood, S., Patterson, D., & Benefield, A. (1988). Exhibit design and visitor behavior: Empirical relationships. Environment and Behavior, 20, Bouwens, N. (1991). Unpublished raw data. Broom, D. M. (1991). Assessing welfare and suffering. Behavioral Processes, 25, Carlstead, K. (1991). Husbandry of the fennec fox (Fennecus zerda): Environmental conditions influencing stereotypic behaviour. International Zoo Yearbook, 30, Carlstead, K. (1998). Determining the causes of stereotypic behaviors in zoo carnivores: Toward appropriate enrichment strategies. In D. J. Shepherdson, J. D. Mellen, & M. Hutchins (Eds.), Second nature: Environmental enrichment for captive animals (pp ). Washington DC: Smithsonian Institution Press. Carlstead, K., Brown, J. L., & Seidensticker, J. (1993). Behavioral and adrenocortical responses to environmental changes in leopard cats (Felis bengalensis). Zoo Biology, 12, Erwin, J., & Deni, R. (1979). Strangers in a strange land: Abnormal behaviors or abnormal environments? In J. Erwin, T. L. Maple, & G. Mitchell (Eds.), Captivity and behavior: Primates in breeding colonies, laboratories, and zoos (pp. 1 28). New York: Van Nostrand Reinhold. Finlay, T., James, L. R., & Maple, T. L. (1988). People s perceptions of animals: The influence of zoo environment. Environment and Behavior, 20, Hays, W. L. (1994). Statistics (5th ed.). Orlando, FL: Harcourt Brace. Hoff, M. P., Forthman, D. L., & Maple, T. L. (1994). Dyadic interactions of infant lowland gorillas in an outdoor exhibit compared to an indoor holding area. Zoo Biology, 13, Jackson, D. M. (1994). Animal activity and presence of docent interaction: Visitor behavior at Zoo Atlanta. Visitor Behavior, 9(1), 16. Lyons, J., Young, R. J., & Deag, J. M. (1997). The effects of physical characteristics of the environment and feeding regime on the behavior of captive felids. Zoo Biology, 16, Martin, P., & Bateson, P. (1986). Measuring behavior: An introductory guide. Cambridge, UK: Cambridge University Press. Mason, G. J. (1991). Stereotypies: A critical review. Animal Behaviour, 41, Mellen, J. D. (1991). Factors influencing reproductive success in small captive exotic felids (Felis spp.): A multiple regression analysis. Zoo Biology, 10, Mellen, J. D., Hayes, M. P., & Shepherdson, D. J. (1998). Captive environments for small felids. In D. J. Shepherdson, J. D. Mellen, & M. Hutchins (Eds.), Second nature: Environmental enrichment for captive animals (pp ). Washington, DC: Smithsonian Institution Press. Mitchell, G., Herring, F., Obradovich, S., Tromborg, C., Dowd, B., Neville, L. E., et al. (1991). Effects of visitors and cage changes on the behavior of mangabeys. Zoo Biology, 10, Neveu, H., & Deputte, B. L. (1996). Influence of availability of perches on the behavioral well-being of captive, group-living mangabeys. American Journal of Primatology, 38, Ogden, J. J., Lindburg, D. L., & Maple, T. L. (1993). Preference for structural environmental features in captive lowland gorillas (Gorilla gorilla gorilla). Zoo Biology, 12, Perkins, L. A. (1992). Variables that influence the activity of captive orangutans. Zoo Biology, 11, Runyon, R. P., & Haber, A. (1980). Fundamentals of behavioral statistics (4th ed.). Reading, MA: Addison-Wesley. Sankhala, K. (1977). Tiger! The story of the Indian tiger. New York: Simon & Schuster. Schaller, G. B. (1967). The deer and the tiger. Chicago: University of Chicago Press.

17 EFFECTS OF A VISUAL BARRIER ON PACING 109 Schaller, G. B. (1972). The Serengeti lion: A study of predator prey relations. Chicago: University of Chicago Press. Shepherdson, D. J., Carlstead, K., Mellen, J. D., & Seidensticker, J. (1993). The influence of food presentation on the behavior of small cats in confined environments. Zoo Biology, 12, Stevenson, M. F. (1983). The captive environment: Its effect on exploratory and related behavioural responses in wild animals. In J. Archer & L. I. A. Birke (Eds.), Exploration in animals and humans (pp ). Berkshire, UK: Van Nostrand Reinhold. Tarou, L. R., Bashaw, M. J., & Maple, T. L. (2000). The empty nest: A case study of maternal response to separation from a juvenile offspring in a captive Sumatran orangutan. Journal of Applied Animal Welfare Science, 3, Tenneson, T. (1989). Coping with confinement: Features of the environment that influence animals ability to adapt. Applied Animal Behavior Science, 22, Wilson, S. F. (1982). Environmental influences on the activity of captive apes. Zoo Biology, 1,