«Item type Thesis or dissertation Authors Davis, Nicolas Citation Davis, N., Schaffner, C. M., & Smith, T. E. (2005). Evidence that zoo visitors ...»
Line, K. Morgan, H. Markowitz, & S. Strong, 1989; Line, Morgan, Markowitz, & Strong, 1990). However, this requires the implantation of telemetry devices which may confound and compound measures of stress (Honess & Marin, 2006a). Another method which has been successfully used involves the monitoring of leukocyte activity, which is known to be affected by psychological stress in humans (Ellard, et al., 2001). This technique has also been adopted in badgers in the wild (Montes, McLaren, Macdonald, & Mian, 2004) and more recently in non human primates (Honess, et al., 2005). However, the procedure requires blood samples which involve separation, capture and restraint, all of which are known to cause a stress response (Mormede, et al., 2007).
Finally, the monitoring of the neuroendocrine system has also been successful in measuring the stress response, and the most widely studied and dependable index uses adrenocortical hormones (Kikusui, et al., 2006). There are a number of hormones involved in the HPA axis including CRH, AVP, ACTH and GCs (see section 1.4.2). These cholesterol-derived steroids are produced as cortisol in most mammals and fish, and as corticosterone in rats and birds (Mormede, et al., 2007).
Although increased levels of GCs have been recorded in many species and in every vertebrate genus (S. L. Klein, 2000), there are large inter-specific differences in basal levels of cortisol. Such differences are however comparable for the majority of physiological and behavioural parameters (Lane, 2006). The measurement of these GCs following activation of the HPA axis is a well established means of assessing stress levels in animals (Buchanan & Goldsmith, 2004; Mendoza, Capitanio, & Mason, 2000) and can provide important information which can help in assessing the welfare status of an individual or group of animals.
1.4.9 Factors effecting glucocorticoid levels There are a number of factors that need to be considered when using GCs as a measure of stress (Honess & Marin, 2006; Lane, 2006; Millspaugh & Washburn,
2004) and these should ideally be controlled for in any research study. For example age, sex and reproductive status may influence the activity of the HPA axis (Millspaugh & Washburn, 2004), although the means by which it is influenced is not clearly understood (Tilbrook, Turner, & Clarke, 2000).
Although there is only limited evidence of any sex differences in basal GC levels (Lane, 2006; Tilbrook, et al., 2000), there is considerable evidence that there can be significant sex differences in responses to different types of stressors (Lane, 2006; Silva, Ines, Nour, Straub, & Da Silva, 2002). Cortisol levels in females are affected by physiological changes during ovulation, pregnancy and lactation. For example, cortisol levels in humans are approximately three times higher during pregnancy compared to non pregnant levels, rising to five times in late gestation (Keller-Wood & Wood, 2001). There are also well established links between reproductive status and basal cortisol levels with elevated cortisol levels during late gestation demonstrated in a variety of other primates (Cavigelli, 1999; T. E. Smith & French, 1997a; Weingrill, Gray, Barrett, & Henzi, 2004).
It has been suggested that exercise should be considered as a variable when considering GCs as a measure of stress (Coleman, et al., 1998). Cortisol has a role in metabolic homeostasis, in particular in the regulation of energy, thus any change in the HPA axis is not always necessarily a consequence of a stressful stimulus (Mormede, et al., 2007). For example, an increase in urinary cortisol levels has been found to be positively correlated with locomotion in some primates, and may be part of a generalised stress response in some species (T. E. Smith, et al., 1998). However, in other studies where this has been assessed, significant increases have only been found for extreme levels of exercise (Lane, 2006). These studies indicate that providing energy requirements can be met by existing fat and carbohydrate stores, for example during moderate or short intense periods of exercise, levels of GCs would not be affected.
Other factors that need to be considered when using GC levels include the natural circadian cycle of circulating cortisol, with GC levels known to be much higher in the morning than later in the day for diurnal animals (Mendoza, et al., 2000). Seasonality has also been reported to have an effect on GCs in a variety of animals, whereby levels vary at a predictable level throughout the annual cycle (Millspaugh & Washburn, 2004). This can be linked to seasonal changes in environmental conditions that can impact on the individual’s metabolic demand (Lane, 2006), such as extremes of temperature and humidity (Weingrill, et al., 2004), reproductive status (Honess & Marin, 2006a) or food availability (Cavigelli, 1999).
Nutritional status also affects GC levels, particularly for faecal cortisol measurements (Millspaugh & Washburn, 2004), although day-to-day changes in diet are not thought to have a significant impact (Lane, 2006). These external factors are particularly relevant to wild and free range studies whereby such conditions, unlike captive studies, are not under the researcher’s control.
1.4.10 Glucocorticoids as a measure of stress Glucocorticoid hormones can be measured in several biological samples, including plasma, saliva, urine and faeces. Plasma is the most widely used in animal welfare studies (Mormede, et al., 2007), and its benefits include providing an instantaneous value of GCs in the blood at the time of the sample. However, there are also a number of conflicting factors that need to be considered.
There is potentially a large variation in values depending on when the sample is taken in relation to when the stressor occurred (Lane, 2006). The HPA axis is highly sensitive and it only takes a few minutes following an event before an increase in GCs can be detected in the blood. The response is then prolonged for around one hour following the cessation of the event (Mormede, et al., 2007). The timing of the blood sample is therefore important with at least ten minutes required before the GCs can be picked up in the blood (Mormede, et al., 2007). This sensitivity also requires consideration for blood sampling procedures. Separation of individuals from their social partners, capture, handling, physical restraint and even anesthetisation are often required, and these events stimulate HPA activity (Schaffner & Smith, 2005; T. E. Smith & French, 1997a). Even animals that are trained to present a limb for samples have shown a stress response to the procedure (Honess & Marin, 2006a). Such invasive procedures are likely to confound the interpretation of results. Therefore, where a hands-off study is not possible, the use of appropriate control animals is essential (C. J. Cook, Mellor, Harris, Ingram, & Matthews, 2000).
While remote blood sampling through various devices has been offered as a solution (Ingram, Crockford, & Matthews, 1999), the development of non invasive techniques are favoured as they minimise the impact on the animal and allow the study of animals whilst in their ‘natural’ situation (Buchanan & Goldsmith, 2004).
The use of salivary GCs has been used to monitor stress response in a variety of non–human primate species including rhesus monkeys (Macaca. mulatta) (Boyce, Champoux, Suomi, & Gunnar, 1995), orang-utans (Pongo pygmaeus) (Elder & Menzel, 2001) and baboons (Papio hamadryas) (Pearson, Judge, & Reeder, 2008). A number of methods have been adopted to collect saliva including offering flavoured rope to be chewed (Lutz, Tiefenbacher, Jorgensen, Meyer, & Novak, 2000; Pearson, et al., 2008). It offers the advantage of being relatively non invasive, and offers a less stressful alternative to blood collection for measuring short-term stressors. However, there are some limiting factors that need consideration, with difficulties in its use in untrained and wild animals as there are problems with considerable inter-individual variation in the time lag from the bloodstream to saliva and the impact of circadian rhythms (Lane, 2006). There are also some sensitivity and specificity issues (Mormede, et al., 2007).
The analysis of faecal samples for measuring GCs has been successfully used to monitor stress responses in a variety of animals (Boinski, Swing, T.S., & Davis, 1999; Shepherdson, Carlstead, & Wielebnowski, 2004; Whitten, Stavisky, Aureli, & Russell, 1998). It also offers a non invasive means of collection, and is particularly appropriate in free range and field studies (Cavigelli, 1999; Engh, et al., 2006) making it a particularly useful tool in the field of conservation biology (Millspaugh & Washburn, 2004). Variability has been reported in faecal GC measurements, which may be due to potential dietary effects, water content, collection and storage techniques and assay protocols (Lane, 2006).
Urinary GCs provide another non invasive method, although the practical difficulties in collection make it more prominent in captive studies. It has been used to examine cortisol levels in marmosets (Callithrix kuhlii) (Schaffner & French, 2004; T. E. Smith & French, 1997a), brown capuchins (Cebus apella) (Boinski, Swing, Gross, & Davis, 1999), pig tailed macaques (Macaca nemestrina), (Crockett, et al., 2000), chimpanzees (Pan troglodytes) (Whitten, Stavisky, Aureli, & Russell,
1998) and cotton topped tamarins (Saguinus oedipus) (Ziegler, et al., 1995).. Urine is also the main elimination pathway for GCs, therefore its measurement accounts for the accumulation of cortisol over several hours (Mormede, et al., 2007), although this is subject to species variation. The peak excretion in urinary cortisol in three species of primates was found to be around six hours post stressor (Bahr, Palme, Möhle, Hodges, & Heistermann, 2000). Urinary GCs have the benefit of adjusting for the fluctuations present in plasma levels, therefore providing an integrative, sensitive measure of their production over a period of time (Mormede, et al., 2007). However, such consistency could be seen as a disadvantage if more temporal precision is required (C. J. Cook, et al., 2000). It is also necessary for urinary cortisol levels to be corrected for diuresis by correcting for creatinine levels (Burtis & Ashwood, 2001).
1.5 Animal welfare in zoological parks
In recent years it has been recognised that zoos and wildlife parks have an important educational role in raising awareness of the environmental issues that threaten the survival of animal species in the 21st century (WAZA, 2005). For visitors to be connected or inspired to care about these issues it is important that the animals are maintained at the highest welfare standards and behave as naturally as possible (Kidd, Kidd, & Zasloff, 1995; Robinson, 1998). This is achieved by housing the animals in an appropriate naturalistic context, to encourage their ‘natural’ behaviour within the appropriate surroundings. In addition, any animals that show abnormal behaviours or appear to be suffering will be counter productive to the conservation message (Carlstead, 1996).
Negative or abnormal behaviours have been documented in zoos and include lethargy, inappropriate self-directed behaviours such as self rocking or self mutilation (Hosey & Skyner, 2007), coprophagy, excessive aggression and the performance of natural behaviours, but performed out of context or to an excessive manner (e.g. over grooming) (Carlstead, 1996). Stereotypies are also prevalent in zoos, with some species being particularly likely to display them (e.g. 82% of zoo carnivores or 47% of zoo or circus elephants (G. Mason & Latham, 2004).
The importance of natural behaviours in zoos was first highlighted by Hediger (1950). However, they were not used as a benchmark of animal welfare until much more recently (Chamove & Anderson, 1989; A. F. Fraser & Broom, 1990a;
Lindburg, 1988; Thorpe, 1967). Furthermore, assessment and the presence of natural behaviours needs to be based on scientific evidence, and not on preconceptions, perceived behaviour or anthropomorphisms (Robinson, 1998). It must also be remembered that the wild environment can be a challenging place and does not necessarily provide a blueprint for optimal welfare (Veasey, et al., 1996b).
However, other behaviours can be stimulus driven, and in the absence of the stimulus there is no motivation (Veasey, et al., 1996b). One such example could be predatory avoidance behaviour, and in these cases providing there is no stimulus, the absence of such behaviour would not indicate a loss of welfare. Therefore, the performance of a full repertoire of natural behaviours is not necessarily essential for the welfare of an animal kept in captivity, i.e. an animal behaving differently to that of wild conspecifics is not necessarily suffering (Veasey, et al., 1996b). While some work has been done with domesticated animals regarding behavioural needs (Dawkins, 1990; Jensen & Toates, 1993), such as dust bathing in poultry (Gallus gallus) (Dawkins, 1983) and nest building in sows (Sus scrofa) (Jensen, 1993), surprisingly little work has been done with zoo-housed animals. Indeed for many of the species kept in zoological parks very little information is known about their behaviour in the wild (Robinson, 1998). This presents a challenge for zoo management and their staff to provide the appropriate environments for their animals.
Animals housed in artificial environments are exposed to a wide variety of potential stressors (Morgan & Tromborg, 2007). While zoological parks are not constrained to the same degree as laboratories and some farming paradigms, there are still unique challenges to housing animals within a zoo setting (Hosey, 2005;
Robinson, 1998). These include the effects of visitors (Birke, 2002; S. Cook & Hosey, 1995; Hosey, 2000; Mitchell, et al., 1992), keeper animal interactions (Hosey, 2008; Mellen, 1991; Wielebnowski, Fletchall, Carlstead, Busso, & Brown, 2002), unpredictable noise (Owen, Swaisgood, Czekala, Steinman, & Linburg, 2004;
Shepherdson, Carlstead, & Wielebnowski, 2004), construction work (Powell, Carlstead, Tarou, Brown, & Monfort, 2006), proximity to predators, prey or competing conspecifics (Buchanan-Smith, Anderson, & Ryan, 1993; A. Lee, 1992;