«Item type Thesis or dissertation Authors Davis, Nicolas Citation Davis, N., Schaffner, C. M., & Smith, T. E. (2005). Evidence that zoo visitors ...»
3.1.2 Zoo visitor studies There has been conflicting evidence regarding the effect that visitors have on the lives of captive animals (Hosey, 2000; Hosey, Melfi, & Pankhurst, 2009b). For example, while no significant effect was found on the behaviour of six species of felids (Margulis, Hoyos, & Anderson, 2003), or captive cheetahs (Acinonyx jubatus) (O'Donovan, Hindle, McKeown, & O'Donovan, 1993), visitors did have an effect on captive Indian leopards (Panthera pardus) (Mallapur & Chellam, 2002) and jaguars (Panthera onca) (Sellinger & Ha, 2005). However, the vast majority of visitor studies have been carried out on primates with studies also showing a variety of effects (Davey, 2007). Some researchers have reported no effect of zoo visitors on primates (Adams & Babladelis, 1977; Synder, 1975), whereas other researchers report an enriching effect. Cook and Hosey (1995) reported how chimpanzees (Pan troglodytes) voluntarily interacted with visitors in order to try and obtain food, while Fa (1989) reported positive effects when visitors threw food towards green monkeys (Cercopithecus aethiops), although such behaviour is likely to have negative consequences towards their general health. The majority of studies however appear to demonstrate negative effects to various degrees (Chamove, et al., 1988; Glatston, Geilvoet-Soeteman, Hora-Pecek, & Van Hooff, 1984; Hediger, 1969; Mallapur, Sinha, & Waran, 2005; Mitchell, Obradovich, Herring, Dowd, & Tromborg, 1991;
Skyner, Amory, & Hosey, 2004; Wells, 2005) ranging from an increase in locomotion in a variety of species (Hosey & Druck, 1987), an increase in aggression in mangabeys (Cercocebus galeritus chrysogaster) (Mitchell, Herring, et al., 1991), increasing aggression and stereotypic behaviour in gorillas (Gorilla gorilla) (Wells, 2005), pied tamarins (Saguinus bicolor bicolor) (Wormell, Brayshaw, Price, & Herron, 1996), and mandrills (Mandrillus sphinx) (Chamove, et al., 1988), increases in abnormal behaviours in lion tailed macaques (Macaca silenus) (Mallapur, et al.,
2005) and self-harming behaviour in gibbons (Hylobates pileatus) (Skyner, et al., 2004). A prevalence of aversive consequences may be linked to a closer taxonomic relationship between visitors and other primates and possibly linked with more familiarity in communicative signals (Hosey, Melfi, & Pankhurst, 2009a).
3.1.3 Factors influencing the impact of visitors on zoo primates There are many factors that could be influencing the impact of visitors on zoo animals’ wellbeing which have been outlined by Hosey (2000). Firstly, there appears to be a considerable amount of inter-species variation (Chamove, et al., 1988; A. S.
Clarke & Mason, 1988; Mitchell, Herring, et al., 1992; Wormell, et al., 1996) that may be explained by the degree to which an animal may see the human as a threat.
This perception can be related to the animals body size, its social organisation, species typical responses to environmental events and the extent of habituation to humans (Hosey, 2005). Secondly, the design of the enclosure can also be significant (Carlstead & Shepherdson, 2000; Chamove, et al., 1988; Glatston, et al., 1984;
Hosey, 2000; Mitchell, et al., 1990; Wormell, et al., 1996). The size and complexity of the enclosure space will affect how the animal will respond to visitors, with larger more complex exhibits allowing more retreat space and opportunities for individuals to remove themselves from the view of the public (Hosey, 2005). Thirdly, the way in which visitors view the animals can also be an important factor with the size of viewing windows (Blaney & Wells, 2004), height of viewing (Chamove, et al., 1988) and position of exhibits (Margulis, et al., 2003; Mitchell, et al., 1990) all potentially having an impact. Finally, the behaviour of visitors can also make a difference, with noise, size of crowd and activity level all impacting on the organisms (Birke, 2002;
Hosey, 2000; Hosey & Druck, 1987; Mitchell, Obradovich, et al., 1991).
3.1.4 Using HPA activity to assess visitor impact Although there is a general consensus that visitors can have a negative impact on zoo animals, there is sufficient inconsistency to warrant further study (Hosey, 2000). Previous studies exclusively relied on behavioural indices to assess the impact of visitors on animals, in particular monitoring changes in affiliative and abnormal behaviours (Davey, 2007). While this can be an effective method, it can be difficult to interpret how behavioural changes can affects an animals’ welfare, particularly regarding the presence of abnormal behaviours (G. Mason & Latham, 2004) (see Chapter 1, section 1.3.4). The use of physiological measures as a means of assessing the stress response provides additional evidence and insight into the effect of visitor numbers on individuals. To my knowledge, only one previous study has used physiological indices to assess visitor impact (Kalthoff, Schmidt, & Sachser, 2001).
This used salivary cortisol and behaviour in several mammal species including rhinos, although no significant relationship was found between GC levels and visitor numbers. My study attempted to increase the understanding of visitor effect by assessing the relationship between visitor numbers and one aspect of an animal’s physiology; activity in the HPA axis.
Previous studies have demonstrated that the analysis of urine and faeces can provide an effective method of measuring reproductive steroid metabolites in Neotropical primates, such as marmosets and tamarins (French, et al., 1996; T. E.
Smith, Schaffner, & French, 1997; Ziegler & Snowdon, 2000; Ziegler, Wegner, Carlson, Lazaro-Perea, & Snowdon, 2000). Campbell et al. (2001) have used urine and faecal analysis to investigate levels of pregnane-diol 3α glucuronide and estrone conjugates in the ovarian cycles in female spider monkeys (Ateles geoffroyi).
Cortisol, another steroid hormone, is the end product of HPA activity and is an effective marker for assessing physiological stress in captive animals (Boinski, et al., 1999; Crockett, et al., 2000; T. E. Smith & French, 1997a; Whitten, et al., 1998;
Ziegler, et al., 1995; see Chapter 1, 1.4.10). Therefore measurement of cortisol in the urine can potentially provide information about the physiological response of a nonhuman primate to a potential stressor, such as visitors.
The aim of my study was to investigate the physiological impact of visitors on the spider monkey HPA axis by quantifying levels of urinary cortisol in samples collected during days of varying visitor numbers (i.e. 0 to 16,500 visitors). The relationship between concentrations of urinary cortisol and actual visitor numbers was then investigated. I predicted that if visitors adversely impacted the animals then a positive relationship would be identified between urinary cortisol and visitor numbers.
In February 2001 foot and mouth disease (FMD) appeared in the UK with a devastating impact on all livestock industries (Mepham, 2004). Consequently, strict restrictions were imposed on the movement and handling of animals throughout the country. Farms, zoos and safari parks were closed to all but essential staff for the duration of the outbreak. Chester Zoo was closed for a total of six weeks from February 25th, 2001 to April 6th, 2001. During this time only essential keeping staff members were allowed into the zoological park. The closure of Chester Zoo provided a unique opportunity to collect data for a period when there were no zoo visitors.
3.3.1 Subject and housing The study involved five adult females, one adult male and three juvenile males from a breeding group of Colombian spider monkeys (Ateles geoffroyii rufiventris) housed at Chester Zoo (Table 3.1). The animals had access to both the indoor enclosure throughout the study, apart from the adult male who due to management reasons was separated on his own in the separation pen inside (for details of the enclosure and husbandry see Chapter 2, section 2.16 and Figures 2.4
184.108.40.206 Urine collection Urine collection was conducted three to four times a week between 0700 and 0800 hrs between 23.01.01 and 25.05.01 (See Chapter 2, 2.2 for further details).
220.127.116.11 Quantification of levels of cortisol Levels of cortisol were measured in all urine samples using the EIA as described in Chapter 2 and corrected for urine dilution using the modified Jaffe endpoint assay (Burtis & Ashwood, 2001; Chapter 2, section 2.3.2). Samples were diluted 1:256 to 1:512 as necessary and run in duplicate (see Chapter 2, 2.3) and corrected for creatinine concentrations following the procedure outlined (see Chapter 2, 2.3).
* wild caught ¥ estimate 18.104.22.168 Visitor study The physiological impact of visitors on the monkeys was investigated by assessing the levels of urinary cortisol. Samples were collected during the FMD outbreak when no visitors were in the zoological park and throughout the year when visitor density fluctuated widely. This study used the total number of visitors in the zoo and related this to the concentration of urinary cortisol in the sample collected the next morning to account for a lag in the excretion of urinary cortisol (Bahr, et al., 2000; Whitten, et al., 1998). I selected samples using the Chester Zoo records and diary notes from the urine file. Samples were only included when I was confident that no other physical or social stressful events were occurring. I avoided any samples collected in the three days following a social or physical stressor and samples that preceded a known social conflict between animals as such events have been shown to impact cortisol levels in this group of spider monkeys (Chapter 5). In addition, a study of marmosets revealed that cortisol levels increase prior to the outbreak of serious conflict between animals (T. E. Smith & French, 1997a). Cortisol values were only used on days when there were data points from two or more animals. The values from all animals were averaged on each day to provide one data point per time sampling interval.
It must be noted that for the period of seven months before the FMD outbreak and throughout the zoo closure, the breeding male had been separated from the rest of the group for animal management reasons. Although isolated from the rest of the group he still had full visual and also limited physical contact with the females and juveniles. He was reintroduced back into the group three weeks after the zoological park reopened. To ensure that this separation was not itself a source of stress we used a matched paired t-test to compare the mean urinary cortisol values for all members of the group during two periods when the breeding male was separated and after he had been reintroduced [t (5) = -.017, p = 0.987, Ric in: M = 2.13 ug cortisol/ ml per mg creatinine, SE = 0.40; Ric out: M = 2.14 ug cortisol/ ml per mg creatinine, SE = 0.29]. This indicated that the separation event did not confound the study.
3.3.3 Data analysis The data presented here are derived from 179 urine samples (which resulted in 77 data points) collected across 77 days of urine collection. For details regarding contribution to the study see Chapter 2 (Table 2.4). On average we collected 35 urine samples from each monkey (range of 13 to 51 samples per monkey). I required at least two samples from two different monkeys on a given day to assess whether absolute visitor numbers were associated with urinary cortisol.
To determine whether there was a relationship between visitor number and cortisol we performed a Spearman’s rank correlation, because the presence of 10 data points with a value of 0 led to a skewed distribution as revealed by a significant Kolmogorov-Smirnov test (Z = 1.62, N = 77, p 0.01). To explore the impact of visitors on HPA function we used a repeated measures one-factor analysis of variance (ANOVA) to compare cortisol levels across four categories of visitor numbers. These were no visitors (0), low (1-999), medium (1000-6999) and high (7000) and were based on archived visitor data on the total number of visitors through the gate for a given day. When relevant, I assumed a two-tailed distribution and adopted an alpha level of 0.05 for all statistical tests.
Mean values for all samples illustrated a trend for increasing cortisol values with higher numbers of visitors (Figure 3.1) and a positive correlation between urinary cortisol and number of visitors (rs = 0.43, p 0.001, N = 77, Figure 3.2) was identified. A repeated measures ANOVA was then used to compare cortisol levels from the five subjects across four visitor categories. We corrected for sphericity problems using Huynh-Feldt correction as recommended by Keppel (1993). We identified a non-significant result [F (3, 12) = 2.57, p =.156). However, further investigation of the data revealed that the cortisol levels in one subject (Fay) showed a conflicting trend with levels of cortisol decreasing with increasing visitor number categories. When the data from this individual subject were excluded from the ANOVA, a significant difference in cortisol levels was observed across the four visitor categories [(F (3, 9) = 10.82, p = 0.002].
The aim of this study was to increase the understanding of zoo visitor impact on spider monkeys by incorporating a physiological measure. My study supports previous behavioural research that visitors can have a meaningful impact on primates in zoos (Chamove, et al., 1988; S. Cook & Hosey, 1995; Fa, 1989; Hediger, 1969;
Hosey, 2000). I found that as absolute visitor numbers increased, urinary cortisol increased, which suggests that visitors had an impact on spider monkey physiology.