Widespread Behavioral Responses by Mammals and Fish to Zoo Visitors Highlight Differences between Individual Animals

Simple Summary

It is important to understand the impacts that humans have on zoo animals to ensure that zoo animal welfare is not compromised. We conducted multiple short-term studies of the impact of zoo visitors on 16 animal species and found that 90.9% of the mammal species and 60.0% of the fish species studied exhibited some change in behavior related to zoo visitors. Animals with behavioral changes were housed in exhibits with no direct contact with humans and exhibits with direct contact. These changes in behaviors were not always consistent across species, and often individual animals of the same species and living within the same exhibit had varied behavioral responses. We recommend (1) using short-term assessments to identify behavioral responses that may be of concern; (2) monitoring individual responses of zoo animals to humans; and (3) creating refuges where animals may choose to retreat.

Abstract

The impact that humans have on zoo animals can vary based on the species of animal, exhibit design, and individual differences in behavioral responses. We independently analyzed data from 10 never-published studies that examined the impact of zoo visitors on zoo animal behavior. Of the 16 species studied, 90.9% of the mammal species and 60.0% of the fish species demonstrated a change in at least one behavior based on zoo visitor abundance or visitor behavior (e.g., noise, solicitation of interactions from zoo animals). In addition, behavioral changes associated with zoo visitors were present in animals housed in exhibits where there was direct contact with zoo visitors, as well as in exhibits where there was indirect contact and no direct contact. Individuals often varied in their behavioral responses, and some individuals appeared to seek out interactions with visitors. Our findings demonstrate that short-term research projects can provide valuable insight into individual animal-level and species-level responses to visitor abundance and visitor behavior in the zoo setting. We recommend that behavioral assessments focus on the analysis of behaviors of individual animals whenever possible, and we recommend that exhibits provide areas that allow for animals to retreat from the public view.

1. Introduction

Zoos can play important roles in conservation education, and the experiences that visitors have at a zoo can influence visitors’ perceived connections with the animals, as well as visitors’ responses to conservation messages [1,2]. Zoo visitors often rate their visits as positive when they can see the animals at close proximity [3,4]. Furthermore, visitors often show more interest in animals that are active [5,6], and some visitors report having positive emotional responses after having a direct interaction with a zoo animal [7]. However, popular zoo species can experience greater levels of noise from visitors, as larger crowds can lead to increased ambient sound [8]. As a result of factors related to zoo visitors (e.g., increased visitor abundance, increased noise), some animals have been shown to exhibit behavioral and physiological changes that can be indicative of stress [9,10].

Although studies on the impacts of human visitors on zoo animal behavior have existed for decades [11,12], documented responses of zoo animals to visitors have been inconsistent, and such variation is likely due to a range of factors (e.g., species studied, the animals’ individual personalities, characteristics of the zoo visitors, variability in exhibit design [13,14]). Furthermore, there is not full agreement on whether certain interactions between the visitor and zoo animal are positive, neutral, or negative [15]. Studies have examined a range of variables, from visitor abundance, density, and proximity [16,17], to visitor noise and visitor activity levels [8,18]. Additional studies are needed to understand the extent visitors impact the welfare of the zoo animals [19], as behavioral changes in an animal do not necessarily indicate a negative effect on animal welfare [13,20].

To date, most zoo visitor studies have focused on primates and felids [21]. Primates sometimes respond to zoo visitors [12,17], and at times the primates’ behavioral responses appear to signal that the animals are stressed by the zoo visitors [22,23]. Markers of behavioral stress include behavioral changes [11,24,25], differences in visibility while on exhibit [26], and elevated levels of glucocorticoids [27,28]. Nevertheless, it is difficult to make generalizable statements about the impacts of zoo visitors on primates. For example, Kuhar [26] found that overall western lowland gorillas (Gorilla gorilla gorilla) did not exhibit much behavioral change as visitor abundance increased; however, when crowd size was large, aggression increased in a bachelor male group, but not in a family group. Carder and Semple [29] noted a visitor effect in gorillas at one zoo, but not at another. Stoinski et al. [30] documented that overall gorilla behavior did not appear to be impacted by visitor abundance, but differences existed in the responses by individual gorillas. However, a study of chimpanzees (Pan troglogytes) and gorillas found that overall the animals’ behaviors and exhibit use did not vary with the number of visitors [31].

Similar variation also exists in felids. Margulis et al. [5] did not find evidence that visitor presence impacted six cat species: lion (Panthera leo), Amur leopard (P. pardus orientalis), Amur tiger (P. tigris altaica), snow leopard (P. uncia), clouded leopard (Neofelis nebulosi), and fishing cat (Felis viverrinus). However, Suárez et al. [32] noted behavioral differences based on the presence or absence of zoo visitors in five felid species: Eurasian lynx (Lynx lynx), jaguar (P. onca), bobcat (L. rufus), ocelot (Leopardus pardalis), and Asiatic lion (P. l. persica). A separate study of jaguars found that visitor density and noise level impacted the amount of time the cats spent visible and the male’s level of aggression, while visitor noise levels were associated with increased pacing by the female [33]. Therefore, even with the most-studied taxa (primates and felids), behavioral responses vary.

Other taxa also show a range of responses. Zoo visitors had little to no impact in studies of flamingos (Phoenicopterus roseus, P. ruber, P. chilensis, P. andinus, and P. minor [34]), greater rheas (Rhea americana [35]), slender-tailed meerkats (Suricata suricatta [36]), and African penguins (Spheniscus demersus [37]). However, zoo visitors impacted aggressive, huddling, and avoidance behaviors [38], as well as location within the exhibit area [39], in little penguins (Eudyptula minor), aggression and activity levels in female Galápagos giant tortoise (Chelonoidis nigra [40]), and vigilance in koalas (Phascolarctos cinereus [41]). Furthermore, studies have found variable patterns related to inactive behavior: from resting more when visitor abundance was low [25] to resting more when visitor abundance was high [42], to no change in resting behavior [39].

Additionally, some zoo exhibits encourage direct physical interactions between the visitor and animals [43,44]. In visitor-feeding experiences with crowned lemurs (Eulemur coronatus [44]) and giraffe (Giraffa camelopardalis [45,46]), zoo animals did not appear to be negatively impacted by the increased visitor contact. Although petting-zoo goats (Capra hircus), llamas (Llama glama), and Vietnamese pot-bellied pigs (Sus scrofa) did not respond to zoo visitors in the same manner, Farrand et al. [47] concluded that the animals’ welfare was not negatively impacted by visitor interactions. Petting-zoo African pygmy goats (Capra hircus) and Romanov sheep (Ovis aries) showed a reduction in undesirable behaviors when a retreat area was available, underlying the importance of exhibit design in animal welfare [48]. When multiple exhibit types were studied, Bennett’s wallabies (Macropus rufogriseus) fed more and exhibited more interactive behaviors in the no-interaction exhibits; however, resting, locomotion, and vigilance did not differ between exhibit types [49]. In addition, quokka (Setonix brachyurus) were less likely to be visible when visitors were present in walk-though exhibits, but the effect of visitor presence was not considered to be great [50].

Individual animals can vary in their responses to zoo visitors [21,51,52]. A study of three polar bears (Ursus maritimus) found one animal increased and two animals decreased stereotypical behavior during periods of higher visitor density [53]. A recent review [21] highlighted that individuals may not perceive their environment in the same manner, even when these individuals are in the same exhibit; thus, individual traits and temperament may greatly impact behavioral responses. For example, the social rank of a Japanese macaque (Macaca fuscata) predicted probability of aggression towards zoo visitors [54]. Therefore, it is important to also consider individual differences when assessing zoo animal behavior [26] and welfare [55].

Although the behavior of some zoo animals may change when visitors are present, such behavioral changes are not necessarily negative: in a study of 12 primate species, the primates spent more time in the front of the exhibit and attempted to interact with visitors when large, active visitor groups were present [12]. When orangutans were given a preference test, they did not show aversion to the public viewing area [56]. When treatments were applied that varied in the extent of interactions that servals (Leptailurus serval) had with zoo visitors, the overall decrease in stereotypic pacing led to the conclusion that some visitor-encounter programs may have a short-term positive benefit for the animals [57]. Furthermore, the relationship between variables (e.g., visitor abundance, zoo animal behavior) is not necessarily causal: an increase in certain behaviors may attract zoo visitors [23]. Therefore, visitors and zoo animals may mutually impact the other’s behavior [5].

Here we present findings from 10 independent studies on 16 species that all focused on the impact of human visitors on zoo animals. Each of these 10 studies examined the extent to which one or more independent variable (e.g., abundance, noise, proximity, solicitation of interactions with zoo animals) impacted zoo animal behavior. Because individual animals may have different behavioral responses, whenever possible, we focused our analyses on individual animals, so that animals that demonstrated a change in behavior could be immediately identified, which is important for managing individual animals in a zoo [58,59]. Based on the findings from previous studies (summarized above), we predicted that zoo visitors would be associated with changes in interactive, vigilance, social, aggression, stereotypic, rest, and visibility behaviors of the zoo animals.

Our work contributes new species to the body of literature on the zoo visitor effect: 10 of the 16 species studied were not represented in a recent review of the literature on zoo visitor impacts on animal behavior [21]. Four of these newly studied species were fish, on which there have never been published studies documenting their responses to zoo visitors [21], even though fish (e.g., Chondrichthyes) are often part of animal–visitor interactive exhibits [60]. Findings from research on how visitors affect zoo animals can be integrated into plans for exhibit designs, and hopefully create experiences that are positive for both the visitor and the zoo animal [61]. However, without published data for a wider representation of taxa at zoological parks, analyzing taxonomic-wide patterns of the impacts of zoo visitors on zoo animals is not fully possible. Based on findings from our studies, we make recommendations for zoo animal management, specifically in the context of the responses of zoo animals to humans.

2. Materials and Methods

2.1. Study Subjects and Exhibit Design

Behavioral data were collected on 128 individual zoo animals representing 16 species of mammals and fish (

Table 1. Animals studied and their exhibits. An asterisk (*) represents species for which we found no previous peer-reviewed scientific literature related to the impact of zoo visitors on the species.

Common Name	Species	Exhibit 1	Study ID 2
Grey wolf	Canis lupus	O; 1	1
Cheetah	Acinonyx jubatus	O; 1	2
Lion	Panthera leo	O; 1	3
Southern tamandua *	Tamandua tetradactyla	I; 2 3	4
Garnett’s greater galago *	Otolemur garnettii	I; 2 3	4
Mona monkey *	Cercopithecus mona	O; 1	5
Northern night monkey *	Aotus trivirgatus	I; 2 3	5
Northern white-cheeked gibbon	Nomascus leucogenys	O; 1, 2 4	5, 6, 7
Bonobo *	Pan paniscus	I/O; 2	5
Western lowland gorilla	Gorilla gorilla gorilla	O; 2	5
Sumatran orangutan	Pongo abelii	O; 1	8
Cownose ray *	Rhinoptera bonasus	SE; 3	9, 10
Southern stingray *	Hypanus americanus	SE; 3	10
Bonnethead shark *	Sphyrna tiburo	SE; 3	10
Brownbanded bamboo shark *	Chiloscyllium punctatum	SE; 3	10
White-spotted bamboo shark *	Chiloscyllium plagiosum	SE; 3	10

¹ Exhibit descriptions include their exposure to the local environment (I = indoor; O = outdoor; I/O = indoor and outdoor sections; SE = semi-enclosed tented area), and the extent to which the exhibit was designed to encourage interactions with humans (1 = typical zoo exhibit design, i.e., visitors are separated from zoo animals by a moat, water body, landscaping, or fences; 2 = exhibit allows non-direct interactions between zoo animals and zoo visitors, i.e., indirect contact can be made via glass partition; 3 = exhibit allows direct interactions between zoo animals and zoo visitors, i.e., contact occurs). ² The study identification number is used throughout the manuscript so that the studies may be quickly referenced. ³ Exhibit was in the nocturnal house, so exhibit was dark during data collection hours. ⁴ One exhibit was a typical zoo exhibit, while the other exhibit allowed for interactions through glass windows.

; Appendix A) in 15 exhibits at the Memphis Zoo in Memphis, TN, USA. Data were collected during the months September and October 2010–2018, with each species studied once during a period of 4–5 weeks, with the exception of cownose ray and two groups of northern white-cheeked gibbons, all of which were studied twice in two different years. All studies were observational studies, and therefore did not require formal Institutional Animal Care and Use Committee (IACUC) approval. None of the data presented in this manuscript have been previously published.

Mammals were studied from 14 different exhibits, and fish from one multi-species exhibit. Eight mammal exhibits were outdoor, five were indoor, and one exhibit had both indoor and outdoor sections that were visible to the public (Table 1). There were no species which had one group housed in an outdoor public exhibit and another group housed in an indoor public exhibit. All exhibits contained enrichment items for the animals. Data collection did not occur during daily “keeper chats” where animals were often provided additional enrichment. The only studies that involved enrichment items directly in the research design were the studies of the fish, which were part of a touch tank and where zoo visitors could provide food to the animals.

The extent to which each exhibit encouraged interactions between humans and zoo animals varied. Each exhibit was categorized as: (1) traditional zoo exhibit design (n = 6): visitors were separated from zoo animals by a moat, water body, landscaping, or fences; (2) indirect contact design (n = 8): exhibit allowed non-direct interactions between zoo animals and visitors (e.g., glass window was a physical barrier between the zoo animals and visitors); or (3) direct contact design (n = 1): exhibit allowed direct interactions between zoo animals and visitors, allowing for physical contact (e.g., touching or feeding animals; Table 1). The only exhibit that allowed for direct contact was the interactive touch pool (Stingray Bay) where all the fish studies were conducted.

Most studies research one species at a time, and when meta-analyses are conducted, variation can be large [14]. Furthermore, there can be difficulties with small sample sizes, or confounds related to seasonal variation [26,62]. Our study examined a range of species during the same time of year at the same geographic location, thereby minimizing seasonal changes in behavior.

2.2. Data Collection and Analysis

Each of the 10 studies represented an independent study to test to what extent visitors impact the behavior of zoo animals. All 15 researchers were trained and mentored by the same researcher from 2010 to 2018, and all studies used instantaneous scan-sampling methods [63]. Because each study was independent from all other studies, each study involved behavioral data collection by only one or two researchers; when two researchers worked together, one person collected scan sampling data for a species at a particular time. Scan sampling occurred every minute or every two minutes, depending on the specific characteristics of each study (e.g., number of individuals sampled, size of exhibit, species-specific characteristics such as quickness of movement across exhibit; details available in Appendix A). Such time intervals typically result in the same estimates for behaviors [64,65]; therefore, it was not a concern that our 10 studies used a mixture of one-minute or two-minute scan intervals. A typical observation period lasted 120 min (Appendix A). During each scan, the variables associated with human presence (

Table 2. Human-focused independent variables addressed in 16 species across 10 studies ¹.

Variable	Definition	Species of Study 2
Abundance	Number of zoo visitors located in the public viewing area. Four categories: 0, 1–4, 5–9, or ≥10 visitors	Wolf, cheetah, lion, mona monkey, owl monkey, gibbon (ID: 5, 6), bonobo, gorilla
Abundance (aquatic) 3	Number of zoo visitors peering or reaching into the pool. Four categories: 0, 1–10, 11–20, and ≥ 21 visitors	Cownose ray (ID: 9)
Presence	Two categories: (1) 0 visitors and (2) ≥1 visitors present	Gibbon (ID: 7)
Noise	Three categories: (1) low (visitors walking, talking softly, or pushing a stroller); (2) medium (“low” plus running or talking loudly); (3) high (“medium” plus playing loud music, yelling, and/or howling)	Wolf
Solicit interaction	Two categories: (1) visitor present at the exhibit window, but is not initiating contact (not touching the window); (2) visitor present at the exhibit window, and is touching, tapping, or banging on the window	Gibbon (ID: 7)
Proximity (1)	Three categories: (1) close (at the exhibit’s glass window); (2) medium (<1 m from the window); and (3) far (1–2 m from window)	Tamandua, galago
Proximity (2)	Two categories: (1) human observer and study animal within closest proximity; (2) human observer and study animal at greater distance than first category	Orangutan
Food provisioned (aquatic)	Two categories: (1) visitor(s) put fish into pool; (2) no fish was added to the pool	Cownose ray, southern stingray, bonnethead shark, brownbanded bamboo, white-spotted bamboo shark (ID: 10)

¹ Each study measured one or more of these human-focused variables. Most studies addressed only one human-focused variable. The human variable was recorded at each scan sample. Details are available in Appendix A. ² For species that were studied multiple times, the study identification number (ID) is listed for clarification. ³ The classifications for aquatic visitor abundance were larger than for non-aquatic visitor abundance because the interactive pool allowed for visitors to position themselves around the entire perimeter of the pool. None of the non-aquatic exhibits allowed for visitors to be around the entire perimeter.

) and the behavior of each animal in sight (

Table 3. Ethogram for the behaviors studied ¹.

Behavior	General Definition
Mammals
Interactive	Approach human with eyes on the human; approach the glass window or place appendages on window; make direct eye contact with or gesture towards human (Used only in primate and tamandua studies).
Vigilance	Prolonged stare at a specific location, following a human with eyes or head, or scanning the exhibit with eyes and head (Used only in felid studies: cheetah and lion).
Alert	Look in one direction while resting with head up, standing, walking, running, stalking, sitting upright, or vocalizing (Used only in wolf study).
Social	Interact with another conspecific individual through touch (e.g., grooming, huddling bodies).
Aggression	Agonistic interactions between individuals through physical contact.
Stereotypic	Repetitive movements (e.g., pacing).
Rest	Remain in one location without movement. Sitting or prone position.
Out of Sight	Animal confirmed to be on exhibit but hidden from sight of the human observer.
Fish
Swim	Forward movement in the water.
Rest	No movement in the water.
Enrichment	Swim through or on top of, or rest within or on top of, enrichment item.
Social	Within two body lengths of one of more conspecifics; if swimming, swimming in the same direction.
Solitary	More than two body lengths from a conspecific; if swimming, more than two body lengths or swimming in the opposite direction.

¹ Specific ethograms for each of the 10 independent studies of 16 species are provided in Appendix A. The behaviors listed here were all behaviors analyzed. These behaviors were part of the ethograms for multiple species, but these behaviors were not analyzed when the species exhibited them infrequently (<5% of behavioral scans). All ethogram behaviors were analyzed for at least one species, except for aggression. Researchers recorded when instances of aggression occurred, but these occurrences were so infrequent (or did not occur at all) that they did not comprise a sample size large enough for analysis for any of the species (or individuals).

) were recorded.

For most of the 16 species, the ethogram (Table 3) was consistent across the studies. Because there were some species-specific behaviors (e.g., swimming was included only for the fish), full details for each of the 10 studies are provided (Appendix A Methods). We focused on the behaviors listed in the ethogram (Table 3) because these behaviors have been often analyzed in previous studies on the effect of zoo visitors on zoo animals (see [21] and the introduction of the current manuscript for a review).

For three studies (Study IDs 7, 8, and 10) representing seven species (white-cheeked gibbon, Sumatran orangutan, cownose ray, southern stingray, bonnethead shark, white-spotted bamboo shark, and brownbanded bamboo shark), the physical locations of the zoo animals were also noted at each scan (details provided Appendix A Methods).

When possible, behavioral scans were conducted on individual animals, based on an individual’s recognizable physical characteristics. Such single-subject analyses can be important because individuals can have varied responses to variables [59]. When individual identities could not be confirmed throughout a study’s duration, group scan sampling was used to record overall group behavior on an interval. Although group sampling does not provide the detailed level of individual behavior that individual scan sampling provides, this methodology is reliable in studies of free-ranging animals in their natural habitat [66].

Each of the species in the 10 studies were analyzed separately because these 10 studies were fully independent of each other. We tested for the presence of a relationship between a human variable (e.g., visitor presence, visitor abundance, visitor proximity, or noise level; Table 2) and zoo animal behavior (Table 3) for each study independently (full explanation of statistical analyses are provided in Appendix A). Most studies addressed one independent variable (Table 2), unless stated otherwise.

Changes in such behaviors do not necessarily indicate a positive or negative welfare status [21]; therefore, our analyses focused on whether or not the human-associated variable was associated with changes in the zoo animals’ behaviors, and to what extent the behavior changed. We defined a change in behavior occurred when p ≤ 0.05 or p ≤ 0.025, based on the number of comparisons made, and we ran Bonferroni-corrected post-hoc tests to determine patterns in behavioral changes (please see the details in Appendix A). Whenever possible, the analyses focused on each individual animal, instead of overall patterns by the group of individuals. We chose this individual-based approach because zoo management plans that focus on individuals, instead of generalized responses by a population, can help focus on the well-being of each individual zoo animal [58].

No statistical analyses focused on identifying predictive variables indicating the likelihood that certain species would respond to humans. Such predictive variables could not be identified because (1) behavioral changes were noted in a large majority of species; and (2) confounding variables existed: for example, the fish were the only species in an exhibit where direct contact with zoo visitors occurred.

3. Results

3.1. Across All Taxa

Of the 16 species studied, 81.3% (n = 13) exhibited a change in at least one behavior related to the presence of humans. These behavioral changes occurred across taxa: 90.9% of the mammal species (grey wolf, cheetah, lion, Garnett’s greater galago, northern night monkey, Mona monkey, northern white-cheeked gibbon, bonobo, western lowland gorilla, and Sumatran orangutan) and 60% of the fish species (cownose ray, southern stingray, and bonnethead shark) exhibited some changes (Figure 1A). Behavioral changes were not documented in three species: southern tamandua and the two species of nocturnal bamboo sharks. Although variables were kept as constant as possible across the 10 independent studies, species differed in their expression of behaviors of focus (Table 3; detailed results of statistical analyses for behaviors exhibited for ≥5% of the scans are noted in Table S1 and Figures S1–S6). Aggressive behavior rarely occurred (<5% of scans).

Behavioral changes occurred in species living in exhibits that differed in the level of contact with zoo visitors (Figure 1B). These descriptive results are provided to demonstrate the range of species that showed a behavioral change in the studies; however, these species and exhibit characteristics are not necessarily independent of each other (e.g., all fish species were in a direct-contact exhibit, while none of the mammals were).

While individuals within a species (e.g., lion, cheetah, orangutan) had similar behavioral responses to zoo visitors, differences in individual responses occurred in other species (e.g., wolf, galago;

Table 4. Summary of behavioral changes associated with zoo visitors ¹.

Human Variable 2	Zoo Animal Behavior 2	Common Name (# of Individuals with Change)
Abundance↑	Interactive↑	Mona monkey 3; gibbon (3 of 4—ID: 5); gorilla (females) 3
	Vigilance↑	Cheetah (2 of 2); lion (3 of 3)
	Alert↑	Wolf (2 of 4)
	Social↓	Gibbon (3 of 4—study ID: 5) bonobo 3; cownose ray 3 (ID: 9)
	Social↑	Night monkey 3
	Rest↓	Lion (3 of 3); Mona monkey 3; night monkey 3; gibbon (4 of 4—ID: 5); gorilla (females) 3
	Rest↑	Bonobo 3
	Out of sight↓	Mona monkey 3
Abundance↓	Interactive↑	Gorilla (1 of 1 male) 3
Presence: Yes	Interactive↑	Gibbon (2 of 2—ID: 7)
	Rest↓	Gibbon (1 of 2—ID: 7)
	Social↓	Gibbon (2 of 2—ID: 7)
	Located at window↑	Gibbon (2 of 2—ID: 7)
Noise level↑	Alert↑	Wolf (2 of 4)
Proximity: Close	Interactive↑	Galago (1 of 4); orangutan (2 of 2)
	Rest↓	Galago (2 of 4); orangutan (2 of 2)
	Social↑	Orangutan (1 of 2)
	Out of sight↑	Galago (1 of 4)
	Out of sight↓	Galago (2 of 4)
Food provisioned	Rest↓/Swim↑	Southern stingray 3 (ID: 10)
	Periphery of exhibit↑	Bonnethead shark 3 (ID: 10)

¹ Listing is only when behavioral differences occurred, and only for results where a clear pattern emerged (e.g., an increase in alert behavior when visitor abundance increased). A behavioral change was determined when p ≤ 0.05 or p ≤ 0.025, depending on the number of variables analyzed (Appendix A; Table S1). Patterns of the responses were determined based on Bonferroni-adjusted post-hoc tests (Figures S1–S6). For species where there were multiple studies, the study identification number (ID) is noted. A behavior was analyzed for an individual or species only if the behavior was represented in ≥5% of the behavioral scans. Details on the behaviors tested are in Appendix A. ² The human variables are defined in Table 2, while the zoo animal behavior is defined in Table 3. ³ Results represent patterns from group scan sampling, not individual scan sampling.

). These differences were minor (gibbons, study ID 5: one behavior for one individual did not follow the pattern) to relatively complex (galagos: individual differences with four behaviors). In the gorillas, females showed more interactive behavior with humans when visitor abundance was high. Such behavior was greatest in the male when visitor abundance was low; however, the male engaged in interactive behavior twice as often as the females.

3.2. Mammals: Visitor Abundance, Presence, Noise, and Solicitation of Interactions

Behavioral changes occurred in 90.9% of the mammal species studied (Table S1). The response of zoo animals to visitor abundance was studied in eight species, and all eight species (wolf, cheetah, lion, night monkey, mona monkey, white-cheeked gibbon, bonobo, and gorilla) demonstrated behavioral changes to some extent. For most species, as zoo visitor abundance increased, interactive, vigilance, and alert behaviors increased, and social, rest, and out-of-sight behaviors decreased (Table 4). However, these patterns did not hold for all species (e.g., social increased in night monkeys with greater visitor abundance; rest decreased in bonobos when visitor abundance low). Although individual responses were consistent in the cheetahs, lions, and most of the gibbons, for some behaviors there were differences in the responses by individual wolves (alert and rest; Figure S1), and between the male gorilla and three female gorillas (interactive and rest; Figure S3).

There were two groups of gibbons, and both groups were studied twice (two different years). The group of gibbons housed in the indirect-contact exhibit, where visitors and gibbons could interact through a set of long windows, consistently demonstrated behavioral changes based on the abundance (ID: 5; Figure S3) and presence (ID: 7; Figure S4) of visitors. In both studies, as visitor abundance increased or as visitors were present, the gibbons increased their interactive behaviors (e.g., placing hands on window, making eye contact), but decreased their conspecific social and rest behaviors. These patterns held overall for the other group of gibbons housed in a traditional exhibit during one study (ID: 5; Figure S3), but not for the second study (ID: 6; Figure S5). Such findings highlight the importance of studying animals at different time points, as variables may change over time (e.g., Study IDs 5 and 6 differed because the son of the mated pair was transferred to another zoo). However, Study ID 6 did not examine behavior for each individual gibbon. Therefore, it is likely that data collection methods may greatly impact one’s conclusions: individual scan sampling occurred for all of the gibbon studies, with the exception of ID 6, which employed group scan sampling methods. As seen in several species (Table 4), often there were individual differences in behavior. Such differences between individuals can be muted when data are collected via group scan sampling. We acknowledge that these differences in results may be due to year-to-year changes in group composition, or due to differences in sampling methods. That said, we think it is important to highlight these factors because often studies address one social group for a short period of time, or studies may approach analyses from the level of only the group and not of individual animals. Our findings suggest that whenever possible, it is best to collect and analyze data on the level of the individual animal.

Lastly, two studies examined the variable of visitor abundance or presence one step further, by categorizing the human behaviors. The wolf study categorized visitor noise level and found that three of the four wolves exhibited a behavioral change, with two wolves having increased alert behavior when visitor noise was at the greatest level (Figure S1). These findings suggest that crowd size and/or noise may be important factors in behavioral responses by some animals. In one of the gibbon studies (ID: 7), visitor presence increased gibbon interactive behavior, increased the percent of time that gibbons were located at the exhibit window, and decreased conspecific social and rest behaviors; however, no behavioral differences were detected when comparing situations when visitors were present and soliciting interactions from the gibbons (e.g., touching, tapping, or banging on the windows) versus visitors were present and not attempting to engage with the gibbons (Figure S4). These findings suggest that the gibbons did not avoid the visitors when the humans were soliciting engagement.

3.3. Mammals: Human Proximity

Proximity between humans and zoo animals was studied in two studies (Study IDs: 4 and 8). The first study analyzed behaviors based on the proximity of the human to the exhibit window of the galagos (3 exhibits) and tamanduas (1 exhibit). All four galagos demonstrated some change in behavior across three levels of human proximity (close, medium, far), but the behaviors that changed (e.g., interactive, rest, stereotypic, and out of sight) varied for the four individuals (Table 4; Figure S2). For example, two individuals exhibited stereotypic behavior for ≥5% of the scans (5–6% of scans for both individuals), but one galago increased stereotypic behavior when humans were at the exhibit window, and the other increased stereotypic behavior when humans were within 1 m of the window. These two individuals were housed separately: one with another galago that exhibited stereotypic behavior for < 1% of the scans; the other galago was not housed with a conspecific.

Neither tamandua differed in their behaviors based on visitor proximity. The male exhibited stereotypic behavior for 61.8% of the behavioral scans (Figure S2), and the stereotypic behavior did not change based on the proximity of the human. The female, however, exhibited minimal stereotypic behavior (0.32% of scans). Both the female and male were out of sight for a relatively large amount of time (38.8% and 20.8%, respectively), in comparison with the other species in this study.

While the galago and tamandua proximity study examined behavioral responses based on human proximity to the exhibit window, the orangutan proximity study focused on determining if the proximity of the human impacted the physical location of the orangutan, and if orangutans showed behavioral differences when the human and orangutan were in closest proximity. One of the two female orangutans was in closest proximity with the human observer more often than by chance; the second female did not show such a pattern. In fact, this female’s behavior closely approached statistical significance for avoiding the human observer (Table S1). When the females were in closest proximity to the human, both engaged in interactive behavior more and rested less (Figure S5). There were individual differences in social behavior; the female who spent little time in closest proximity with the human engaged in more social behavior when in closest proximity with the human. This finding is likely because the other female, with whom she was social, was in closest proximity with the human more often than by chance.

3.4. Fish: Visitor Abundance, Presence, and Food Provisioning

Cownose ray, southern stingray, and bonnethead shark all exhibited some extent of behavioral differences based on human behavior (Table 4). The variable visitor abundance was studied for only one species, the cownose ray (Study ID 9). The cownose ray decreased social behavior and increased solitary behavior as the abundance of visitors increased along the perimeter of the direct-contact touch pool (Figure S6A).

In the analysis of the impacts of food provisioning on behavior and location within the exhibit (Study ID 10), southern stingray increased their swim and decreased their rest behaviors when visitors added food into the exhibit pool, and bonnethead sharks increased the percentage of scans they were located on the periphery of the exhibit when food was provisioned (Figure S6B); however, the southern stingray did not exhibit differences in their use of the periphery or inside of the pool, and bonnethead sharks did not exhibit differences in swim and rest behavior based on food provisioning (as swim behavior occurred more than 80% of the scans when food was provisioned and when no food was provided). The cownose ray did not exhibit differences in swim behavior (it remained more than 97% of scans under the conditions of food provisioned and food not provisioned), nor in location. The two species of nocturnal sharks (brownbanded and white-spotted bamboo sharks) did not exhibit behavioral changes; they spent more than 85% of all scans resting in the inside section of the pool. Such findings were not surprising given that these sharks are nocturnal, but the exhibit was a diurnal-focused exhibit, located in a partially open-air tent.

4. Discussion

In 13 of the 16 species studied, we found evidence that visitors impact the behavior of zoo animals. Behavioral changes were noted across taxonomic groups, with all but one mammal species demonstrating a behavioral change associated with zoo visitors. Furthermore, we noted behavioral changes in animals housed in exhibits with no contact allowed between zoo visitors and zoo animals, exhibits allowing for indirect contact (e.g., glass window), and exhibits allowing for direct contact between zoo visitors and animals. We acknowledge that our 10 studies were independent studies that did not focus on the same independent variables for all studies, and that direct comparisons cannot be made across all 16 species as to what primarily influences behavioral changes in zoo animals. However, such comparisons were not the intent of our analysis. Instead, our goal was to document the extent to which a range of variables associated with zoo visitors were related to behavioral changes in zoo animals.

Overall, we found that as visitor intensity increased in abundance, noise, and/or proximity, a majority of the animals studied demonstrated an increase in alert, vigilance, or visitor-interactive behaviors, and a decrease in social and rest behaviors. However, patterns were not consistent across all species, and behavioral differences sometimes existed between individuals living within the same social group. Furthermore, zoo visitor abundance was not consistent across exhibits: both exhibits of white-cheeked gibbons attracted relatively large numbers of visitors (Appendix A), while the northern night monkeys never had more than seven visitors at one time, suggesting that that the visitor effect can potentially impact individuals in particular exhibits more frequently and for longer durations than individuals in exhibits that are not frequented by visitors as often and to the same extent.

The results from our studies support previous findings that the responses to zoo visitors can be variable [13,14]. Zoo visitors had an impact on the behaviors for 81.3% of the species we studied, but the lack of a behavioral change noted for three (18.7%) of the species we studied does not lead us to conclude that these three species were not impacted by zoo visitors. First, in some of the studies, the researchers did not identify individual animals; therefore, potential differences in individual animals’ behavioral responses were not detectable in those studies. Second, the three species for which we did not detect behavioral changes (tamandua, brownbanded bamboo shark, and white-spotted bamboo shark) are primarily nocturnal. While one species, tamandua, was housed in an exhibit set for nocturnal conditions, the two species of bamboo sharks were in a diurnal exhibit. These bamboo sharks spent most of the study resting in one location that was not in close proximity to where human visitors stood. Third, our 10 studies rarely measured more than two zoo visitor variables at a time. It is possible that for studies where no behavioral change was detected, had the study examined additional variables, the results may have differed. For example, Choo [16] found that overall orangutans were impacted by the proximity of zoo visitors, but not by visitor abundance or activity. In our study, the wolves’ behaviors often changed as visitor abundance and visitor noise changed, but there were no differences based on visitor presence versus absence. Therefore, future studies that record both visitor abundance as well as visitor noise and proximity could help tease apart the factors that most prominently impact the zoo animals. Such information can then be used directly by the zoo in their communication with visitors. For example, Sherwen et al. [36] found that the use of signs and the positioning of zoo employees reduced visitor noise around a meerkat exhibit.

The goal of our study was to determine to what extent the zoo animals demonstrated behavioral changes associated with changes in zoo visitor abundance, presence, noise, proximity, or food provisioning. We acknowledge that individual animals may vary drastically in their previous experiences with zoo visitors (e.g., frequency, intensity, and overall nature of interactions with visitors). Further, individual animals can vary in how they cope with changing environments and stressful situations [21,54]. That said, given the dearth of information published on some of the species we studied, documenting the behavioral changes that occur is a first step in gaining a better understanding of the visitor effect, and understanding how individual animals respond. Our studies examined behavioral changes at the individual level, whenever possible, because of the importance of understanding individual animals’ responses [58,59].

Although some behaviors (e.g., stereotypic behavior, aggression) are often associated with negative animal welfare, and other behaviors (e.g., play, affiliative social interactions) are sometimes associated with positive welfare [40,67], Sherwen and Hemsworth [21] stressed that many behaviors may not have clear associations with welfare status, and behavioral changes do not always indicate that the change is negative [68]. We acknowledge that an animal’s behavioral response (for example, decreasing rest behavior) does not necessarily indicate a particular emotional response. Therefore, we did not categorize the behaviors based on what the welfare implications were. We present our findings to add to the body of scientific literature on the responses of animals to zoo visitors, to help form a better understanding of zoo animal responses.

Based on our findings from studies of 16 species, we discuss below our recommendations for the management of zoo animals. These recommendations address (1) the importance of short-term studies that allow for the assessment of behavioral responses by individual animals, (2) the extent to which exhibit design may impact individual animals, and (3) considerations for future research studies.

4.1. Study Design: Short-Term Assessments on Individual Animals

We recommend that when it is not feasible to conduct long-term projects on many species in multiple exhibits at a zoo, short-term monitoring programs that are based on well-established, species-appropriate ethograms can provide a great deal of information on a range of species at a particular zoo, and across multiple institutions. The benefits of individual-based monitoring are known [58,59]. Such monitoring programs could involve animal keepers and zoo educational staff, as well as members of the public (e.g., trained zoo volunteers, students taking a behavior course). These short-term assessments can quickly highlight if there are species or individuals that may be of concern.

Previous studies have demonstrated that individuals can vary in their behavioral responses [21,51,52]. We found further evidence that individual animals of the same species in the same exhibit do not always respond in the same manner to humans. Therefore, we also recommend that studies are designed so that behavioral responses by individual animals may be detected. Although it may be necessary at times to collect data on the entire group, such as when it is not possible to accurately identify individuals, we suggest that data on individual animals are collected whenever possible. Single-subject experimental designs (SSDs) can also be important [59], especially if there is an individual animal of concern. While it is important to gain a general understanding of the impact of visitors on all species at a zoo, individuals within a species may react in different ways to visitor presence and other variables associated with visitors [8,30]. Therefore, it may be worth doing these initial assessments, and then consider SSDs or more involved, long-term studies to examine if changes in exhibit design, visitor behavior, or management practices lead to changes in the impact on the zoo animals (in the case that the impact was originally deemed to be negative).

4.2. Exhibit Design

Zoo visitors are often interested in seeing zoo animals in close proximity [4], but proximity to humans can be a source of stress for zoo animals [69]. Sometimes modifications to enclosures can minimize the visitor effect: when netting in front of an enclosure modified the visibility between gorillas and zoo visitors, the gorillas’ behaviors changed (reduced aggression and stereotypical behavior), but the visitors’ perceptions also changed [70]. In a separate study, when visual contact between zoo visitors and capuchins (Sapajus apella) was reduced, the capuchins exhibited a decrease in aggression and in fecal glucocorticoid metabolite concentration, but the number of visitors also decreased [28]. Therefore, having zoos monitor the visibility of the animals [71] can address the zoo visitors’ experiences as well as identify potential issues relating to animal stress.

We found widespread behavioral responses of animals to zoo visitors, and these responses occurred in animals in traditional exhibits to animals in exhibits that were designed to encourage interactions between zoo animals and visitors. All species (100%) that had more traditional exhibits (e.g., zoo animals and visitors were physically distanced from each other due to a water body, elevation difference, and/or vegetation) demonstrated some level of behavioral change associated with zoo visitors. Behavioral changes also occurred for the majority of species in indirect-contact and direct-contact exhibits (87.5% and 60% of species, respectively). These findings illustrate the widespread extent to which zoo animals respond to visitors.

An effective exhibit design can help protect the animals from potential negative consequences of large numbers of visitors [16,31]. In some exhibits, animals rotate on and off exhibit; it is possible that having restricted access to on-exhibit areas (which are often outdoors) could impact the animal’s behavior, as could being off exhibit where zoo visitors are not present. Based on our findings, we recommend that attention is paid to how individuals use their exhibit, and we recommend that all exhibits offer areas of refuge that are adequate in size for all individuals to enter at one time.

Although zoo animals may seek refuge at times, visitors can potentially be a source of environmental enrichment for the zoo animals [20,56]. In our study, when visitors were present at a window, the gibbons spent more time interacting with the humans than they did with each other, and the gibbons spent more time at the window when visitors were present. However, when visitors were present at the window, there was no difference in gibbon behavior when the visitors solicited interactions with the gibbons compared to when the visitors did not initiate contact. These findings suggest that the visitors attracted the attention of the gibbons, but whether or not it was the visitor who initiated contact did not influence the behavior of the gibbons. Because the gibbons were able to access the entire exhibit, but the interactive window was only available on one side of the exhibit, the gibbons had to approach the window to interact with the humans. Our findings appeared to indicate that the gibbons sought out interactions with the humans. The gibbons did not rest or engage in social behavior as often as they did when no visitors were present, but additional study is needed to determine to what extent visitors could potentially impact short- and long-term social bonds between conspecifics in the same social group.

We found that some individuals appeared to initiate interactions with humans (e.g., the gibbons in the indirect-contact exhibit), not all animals did so. For example, the wolves’ behavioral changes (e.g., increased alert behavior and decreased rest behavior in some individuals) appeared to be a response to the visitor abundance and visitor noise. For both tamanduas we studied, no differences in behaviors were detected based on human proximity, but the sustained stereotypic pacing of the male tamandua and the frequency that both animals were on exhibit yet hidden from view are important findings to address overall in regards to the housing and husbandry of these individuals. These differences in responses could be due to species differences or each individual animal’s history with humans; a more extensive data set is needed to draw such conclusions.

In our study of the orangutans, the location of the human observer appeared to impact the physical location of one of the female orangutans. Furthermore, for both female orangutans, they rested less and interacted with humans (e.g., direct eye contact, reaching toward a human) more when they were in closest proximity with a human observer. Although this study did not focus on quantifying zoo visitor abundance, it did provide novel findings about the potential impacts of human observers on zoo animals. We found the orangutans interacted with humans through gazing and gesturing behaviors, both of which are ways to communicate information between two individuals [72,73,74]. Whereas Kaplan and Rogers [73] found that captive orangutans avoided direct stares and exhibited less direct gazes in comparison to wild orangutans, the two female orangutans at the Memphis Zoo did directly look at humans. However, the responses of the two female orangutans were not identical, as one of the females spent much less time in close proximity with humans, and the male, when on exhibit, spent all of the time in one spot, elevated and at a distance from humans. Such varied responses to exhibit use and humans highlight the importance of assessing individual animals’ behaviors, and determining to what extent the exhibit’s features allow each individual to choose the extent to which they are in view of (or proximity to) humans.

The only exhibit that allowed for direct contact between zoo visitors and zoo animals was the interactive fish exhibit. Behavioral changes were noted in the cownose ray, southern stingray, and bonnethead shark. Although cownose rays increased solitary swimming behavior (and decreased social swimming) when more visitors were present, food provisioning impacted neither the time spent swimming versus resting, nor the physical use of the pool (periphery versus inside). These findings suggest that there may be variability between the individual cownose rays (as the groups were not the same from year to year), or that the rays had different responses to visitor variables. For example, the abundance of visitors may impact the rays in a different manner than whether or not the visitors are engaged in feeding the rays at a particular time, as the rays may anticipate receiving food from the visitors.

4.3. Recommendations for Future Studies

Visitors tend to show a greater interest in mammals, as well as animals with larger bodies and higher activity levels [6]. Furthermore, much of the published literature also focuses on mammals, with emphases on primates and felids [21,75]. Our study includes findings on understudied (or never-studied) species, which are first steps in adding to the general understanding of how zoo animals respond to humans. However, many unknowns still exist regarding the impacts that zoo visitors have on captive species. Ideally, we would be able to identify the primary variables associated with species (or individuals) that demonstrate behavioral changes associated with the visitor effect. However, determining these predictive variables is a difficult task, as the responses by the animals may be based on a variety of factors. Multiple variables (e.g., enclosure design, interactions with zoo public, proximity to potential predators, interactions with animal keepers) can impact the behavioral and physiological responses of zoo animals [11,76,77,78]. As the number of studies increase, and the number of species expands, identifying predictive variables may be more possible. We recommend that studies of the impacts of zoo visitors expand to include species that are underrepresented in the literature, and, when possible, take note of behaviors associated with individual animals.

Recently there has been an increase in research addressing the welfare of zoo animals [79], specifically research that measures behavior and physiology [80]. There is no single strategy for assessing welfare that is most appropriate for all zoo animals [81], as the needs and responses of these animals vary by species [82], as well as on an individual basis [59,83]. Furthermore, it may be that factors other than (or in addition to) zoo visitors are primarily impacting an animal’s behavior. For example, in our study the male tamandua spent 61.8% of the scans exhibiting stereotypic pacing, and such pacing was consistent throughout the study, at different levels of visitor presence. Our finding suggests that additional factors may have been contributing to the stereotypic behavior of this individual animal, but we cannot rule out that zoo visitors did not contribute to the pacing behavior. In such situations, we recommend a holistic approach to animal management to examine multiple factors that may be impacting a particular individual animal.

Noise and disruptions may lead to stress in some zoo animals, resulting in physiological changes as well as behavioral changes [9,10,27,69]. In addition, minimizing such disruptions could be critical for targeted breeding programs of threatened or endangered species. Further research on the hormone profiles of zoo animals could provide a better understanding of both behavioral and physiological factors related to zoo visitor presence. Such information taken together would be helpful in then assessing whether any changes indicate an animal welfare concern [19].

5. Conclusions

We found that more than 80% of the species in our study indicated some degree of behavioral change related to the presence of zoo visitors. Furthermore, we documented behavioral changes in individuals housed in exhibits that vary in their exposure to zoo visitors (no contact to direct contact allowed). Our findings provide evidence that a range of individuals modified their behaviors; however, sometimes some individuals of a species within the same exhibit exhibited a behavioral change while other individuals did not. We recommend that the monitoring of zoo visitors’ impacts on zoo animals should be expanded to account for the variety of species housed in captivity, and to ensure that individual zoo animals are not responding in a manner that suggests that they are experiencing stress from the zoo visitors.