Effects of scent enrichment on behavioral and physiological indicators of stress in zoo primates

Captive breeding is vital for primate conservation, with modern zoos serving a crucial role in breeding populations of threatened species and educating the general public. However, captive populations can experience welfare issues that may also undermine their reproductive success. To enhance the wellbeing of endangered zoo primates, we conducted a study to assess the effects of a new scent enrichment program on captive red‐ruffed lemurs (Varecia rubra), black howler monkeys (Alouatta caraya), siamangs (Symphalangus syndactylus), lar gibbons (Hylobates lar) and orangutans (Pongo pygmaeus pygmaeus). We combined behavioral observations and fecal endocrinology analyses to evaluate the effects of a series of essential oils (benzoin, lavender, lemongrass) on five captive troops (N = 19) housed at Dudley Zoo & Castle and Twycross Zoo (UK). We recorded observations of natural species‐specific and abnormal stress‐related behaviors for 480 h using instantaneous scan sampling. We collected 189 fecal samples and measured the fecal cortisol concentrations using radioimmunoassay. We found a significant effect of the scent enrichment on behaviors, with red‐ruffed lemurs and black howler monkeys reducing their social interactions, as well as red‐ruffed lemurs and lar gibbons decreasing their stress‐related behaviors after they were exposed to the series of essential oils. We also found that red‐ruffed lemurs displayed a significant increase in fecal glucocorticoids following exposure to essential oils. Our contradictory findings suggest that the effects of this series of essential oils may change depending on the species‐specific social lives and olfactory repertoires of primates. In conclusion, we cannot recommend using these essential oils widely with zoo primates without additional evaluation.


| INTRODUCTION
Almost half of the total primate species recognized today worldwide are classified as endangered or critically endangered in the wild, primarily due to human activities (Estrada et al., 2017). Therefore, raising global scientific and public awareness of the plight of the world's primates is now vital (Estrada et al., 2017). Zoos may play a major role (Mellor et al., 2015) as zoo animal populations are usually managed to educate the public regarding wildlife and their habitats and to preserve endangered species through captive breeding and reintroduction programs (Schulte-Hostedde & Mastromonaco, 2015).
However, captive populations, potentially serving as buffers against extinction, experience problems that impair them from being viable for reintroduction into the wild. More specifically, zoo animal populations face reproductive challenges which have so far inhibited them from serving as viable "reserve populations" (Meier, 2016).
Additionally, managing zoo populations is challenging because of the mismatch between natural and captive environments and the knockon effects this has on the repertoire of behaviors exhibited (Carroll et al., 2014). Primates have evolved distinct behavioral patterns and difficulty in engaging in these behaviors can cause frustration or boredom, which, in turn, can lead to stress and development of abnormal behaviors (G. R. Hosey, 2005) that may undermine their individual welfare and ultimately their breeding success.
To maintain captive healthy populations modern zoos take part in conservation breeding programs. As reproductive success is linked to how closely captive environmental conditions mirror those that primates would be experiencing in the wild (Meier, 2016), zoos also use environmental enrichments to manage captive populations.
Environmental enrichment and conservation breeding programs are directly related, as enrichment is a dynamic iterative process that changes an animal's environment, increasing its behavioral choices and prompting a wider range of natural and species-specific behaviors and abilities (Ben-Ari, 2001). Furthermore, enrichment can contribute to promoting resiliency to stress, which helps animals recovering from adverse stimuli (Quirke & O'Riordan, 2011), improving both the exhibit from the visitor perspective and the reproductive performance of the hosted animals (Carlstead & Shepherdson, 1994). Enrichment can also foster the essential skills that animals need for their survival if reintroduced into their habitat (Danial Rioldi, 2013).
Scent-based enrichments can be effective at increasing active behaviors in zoo animals and improve their welfare (Fay & Miller, 2015;Quirke & O'Riordan, 2011;Samuelson et al., 2017). However, this is not always the case and some authors reported findings that are less clear or indicate that scent enrichment has little effect (Myles & Montrose, 2015;Wells et al., 2007). The delivery mechanism of the scent and the type of scents used are crucial for the implementation of novel olfactory enrichment programs (Baker et al., 2018). The majority of studies have used spices or essential oils rather than focusing on natural or biological scents, but this may not necessarily be appropriate for all species (Wells et al., 2007). The main goal of olfactory enrichment is to improve the welfare of animals in captive environments, but there is also the possibility that the use of scents can have additional positive impacts. For example, scents may elicit both behavioral and physiological responses and therefore the use of olfactory enrichment can be potentially used to promote beneficial impacts on reproductive success (Rafacz & Santymire, 2014).
Primates are traditionally considered "microsmatic" (i.e., with a reduced olfactory sense; Negus, 1958) and, as many uses of enrichment are ad hoc and unrecorded, only a small proportion of formal studies on olfactory enrichment has been undertaken on primate species (Clark & King, 2008). However, various lines of evidence suggest that chemical communication may be important in primates (Setchell et al., 2010). In particular, it has become increasingly clear that the sense of smell plays a crucial role in primate sociosexual communication, with semiochemicals (i.e., behavior-and physiology-modifying chemicals; Norland & Lewis, 1976) being important for kin recognition, mate choice and the regulation of sociosexual behaviors (Vaglio et al., 2016). However, little is known about the overall effects of olfactory enrichment on primate species.
The overarching aim of our work is to design and test a new scent enrichment program to enhance the well-being of critically endangered zoo primates. In this context, we carried out a preliminary study which aimed to assess the effects of a series of essential oils (namely, benzoin, lavender, and lemongrass) on behavioral and physiological indicators of stress in five captive primate species: Red-ruffed lemurs (Varecia rubra), black howler monkeys (Alouatta caraya), siamangs (Symphalangus syndactylus), lar gibbons (Hylobates lar), and orangutans (Pongo pygmaeus pygmaeus).
As the majority of studies of scent enrichment on zoo primates focus on essential oils, spices or herbs (Wells et al., 2007), we chose three essential oils due to their ecological relevance to non-human primates (benzoin; e.g., Horvath et al., 2007), effectiveness in domestic animals and humans (lavender; reviewed in Wells, 2009), and efficacy in sheltered cats and dogs as well as in zoo-housed exotic animals (lemongrass; e.g., Ellis & Wells, 2010;Holland, 2018;Wells, 2004).
The primate species investigated in this study are currently classified as critically endangered (red-ruffed lemurs, orangutans), endangered (lar gibbons, siamangs), or threatened (black howler monkeys) largely due to the deforestation, logging, and hunting activities that threaten the habitat and survival of these species across their ranges (IUCN, 2020). Therefore, designing and implementing strategies that improve the welfare and breeding success of these species in captivity is particularly crucial.
In this study, we predicted that the scent enrichment would reduce the stress levels of zoo primates, which would be reflected in significant changes in behavioral (i.e., increase of the frequency of social behaviors, and decrease of the frequency of stress-related behaviors) and physiological (i.e., decrease of fecal glucocorticoid concentrations or FGCs) indicators of well-being when comparing before (i.e., baseline period) and after (i.e., post enrichment period) the scent enrichment program.
Particularly, this should occur in relatively "macrosmatic" primates (i.e., primate species with greater levels of olfactory function; Smith & Bhatnagar, 2004) such as lemurs.

| Study subjects and housing
We studied five captive troops of red-ruffed lemurs, black howler monkeys, siamangs, lar gibbons, and orangutans housed at Dudley Zoo & Castle (red-ruffed lemurs, lar gibbons, orangutans) and Twycross Zoo (black howler monkeys, siamangs) in the United Kingdom. We carried out behavioral observations and fecal sampling from July to September in 2016-2019 (Table 1). All troops lived in indoor enclosures (heated to 28°C) with access to outdoor enclosures (i.e., "visitor walkthrough" enclosure in the case of red-ruffed lemurs).

| Study protocol
We divided the overall study period into three periods: Baseline, scent enrichment, post enrichment. We collected behavioral data and fecal samples for 2-6 days per study period (10 days in total), 2 days per week over a 3-month period (1-week baseline; 3-week scent enrichment, i.e., benzoin, lavender, and lemongrass; 1-week post enrichment), for each species (Table 1) to use a combination of both behavioral (e.g., naturalistic species-specific behaviors, stereotypic behaviors) and physiological (e.g., corticosteroid levels) methods to assess the effects of scent enrichment (See Sections 2.2.2 and 2.2.3).

| Scent enrichment
We cut white cotton sheets into 75-cm-long and 5-cm-wide strips, which were soaked with 20 drops Naissance 100% pure essential oil diluted with 12 ml of cold boiled water. We prepared the scent cotton strips during the early morning of each sampling day over the scent enrichment period. We positioned these strips around both indoor and outdoor enclosures; focusing on the outdoor enclosure, we tied them approximately 1 m from the ground around the climbing frames as these were the most used areas of the enclosures.
We utilized one essential oil (benzoin, lavender, lemongrassrespectively) per week during the scent enrichment period of the study.

| Behavioral data collection
We collected behavioral data by instantaneous scan sampling (Altmann, 1974) of some behaviors (Table 2), as a comparable straightforward assessment of major behavioral states which may indicate the expression of significant stress-related (i.e., selfscratching, pacing) and non-stress-related (i.e., resting, sleeping, grooming, playing) behaviors, with behaviors recorded at 5-min intervals over the duration of 6 h from 9 a.m. to 3 p.m., 10 days over a 3-month period. We recorded a total of 480 h of observations over the study period, with 50 scan samples each sampling day on each group.

| Hormone sampling and measurements
We collected fecal samples every morning before behavioral observations, whenever defecation was observed and the identity of the animal was known. In total, we collected 189 samples (red-ruffed lemurs = 25; black howler monkeys = 56; lar gibbons = 53; siamangs = 16; orangutans = 39). The samples were stored in a freezer at 20°C right after collection. At the end of the study period, the collated samples were fully prepared by adding biological hazard labels onto each pot before being delivered using dry ice to the Department of Veterinary Medical Sciences and Animal Production Science of Bologna University for radioimmunoassay (RIA).
Cortisol concentrations were determined by RIA. All concentrations were expressed in pg/mg of fecal matter. The extraction methodology followed the methods of Fontani et al. (2014). In brief, 5 ml of a methanol:water (4:1 v/v) solution were added to 60 mg (wet weight) of feces in capped-glass tube vials. The vials were then vortexed for 30 min using a multitube pulsing vortexer. After centrifugation at 1500 g for 15 min, 5 ml ethyl ether (BDH Italia) and 0.2 ml NaHCO 3 (5%; Sigma Chemical Co.) were added to 1 ml of supernatant. This preparation was vortexed for 1 min and centrifuged for 5 min at 1500g. The ether portion was aspirated with a pipette and evaporated under an airstream suction hood at 37°C.

| Statistical analyses
To assess the effect of scent enrichment on primate behavior and FGCs, we first generated three behavioral categories from the individual behavioral measures that we collected. More specifically, we generated (1) a resting category by adding up our data on resting and sleeping behaviors, (2) a social category by combining our data on grooming and play, and (3) a stress category by combining our data on pacing and self-scratching behaviors (we included scratching in this category as this is commonly considered an indicator of anxiety; Maestripieri et al., 1992 between the two models to find the model with the best fit (i.e., with the lowest AIC value). Finally, to estimate the effect size for the LMM models, we use the "r2" function implemented in the "performance" package in R (Nakagawa & Schielzeth, 2013). All models met the assumptions of homogeneity of variance and normality of residuals.

| Ethics statement
This study followed the guidelines for the care and use of captive

| RESULTS
Our analyses showed that enrichment condition did not have a significant effect on resting rates for any of the species examined (Table 3). Conversely, the LMM analysis examining the effect of scent enrichment on social behavior among howler monkeys, orangutans and siamangs revealed a significant effect of enrichment condition on rates of social behavior among these species (Table 4), with eight out of the 11 subjects studied exhibiting a decrease in social behavior after the introduction of scent enrichment (Figure 1).

Behavior Description
Resting Lying or sitting while awake, with eyes open and arms down by side of the body.

Sleeping
Lying on back, front or side, eyes closed and the whole body is relaxed.

Grooming
Using fingers or mouth to pick through the coat, removing any foreign bodies from a conspecific.

Playing
Animal is engaging in activities such as chasing others, leaping around the enclosure, and so forth, in a playful context.
Self-scratching An animal rubs their own body at a fast pace.

Pacing
Animal walks back and forth in a distinct, unchanging pattern within the enclosure.
Only three individuals (two Siamangs and one orangutan) showed an increase in social behavior after exposure to scent enrichment ( Figure 1). Similarly, the regression analysis conducted on lar gibbons and red-ruffed lemurs showed that the interaction between species and enrichment condition had a significant impact on rates of social behavior (Table 4). This analysis revealed that though rates of social interactions among lar gibbons were comparable between before and after exposure to the scent enrichment, among red-ruffed  We did not find any significant effect of enrichment condition on rates of stress-related behavior on howler monkeys, orangutans and siamangs via the LMM analysis (Table 5). By contrast, we found a significant effect of the enrichment condition on rates of stressrelated behavior among lar gibbons and red-ruffed lemurs in the regression model. This analysis showed that both species exhibited a significant reduction in rates of stress-related behavior following the exposure to scent enrichment ( Figure 3).
The LMM model that investigated the effect of scent enrichment on FGCs among howler monkeys, orangutans, siamangs and red-ruffed lemurs revealed a significant interaction between species and enrichment condition (Table 6). More specifically, the analysis showed that enrichment condition affected FGCs in redruffed lemurs but not in other study species. Contrary to our expectations, however, we found that FGCs increased after exposure to scent enrichment, compared to before the introduction of the scent ( Figure 4). Interestingly, Figure 4 shows that orangutans seemed to decrease their FGC levels following exposure to scent enrichment, although the effect failed to reach statistical significance. Finally, among lar gibbons, although mean FGC con-

| DISCUSSION
The effects of scent enrichment have previously been tested on several domestic, farm, laboratory, and zoo-housed animals (Blackie & de Sousa, 2019;Heitman et al., 2018). However, olfactory stimulation is still one of the least studied forms of enrichment (reviewed in Campbell-Palmer & Rosell, 2011). In addition, there are mixed and conflicting assumptions regarding the benefits of olfactory F I G U R E 2 Mean rates ± standard error of the mean of social behavior among lar gibbon and red-ruffed lemur. The regression analysis showed that the introduction of the scent enrichment induced a significant reduction in social behaviors among red-ruffed lemurs Unexpectedly, we found a significant reduction in rates of social interactions after being exposed to the series of essential oils in both Regarding primates, Gronqvist et al. (2013) showed that olfactory enrichment significantly increased the frequency of natural speciesspecific behaviors in captive Javan gibbons (Hylobates moloch) although the interest in the new scent decreased rapidly after the first day, though no significant effects on individual behaviors were found in ring-tailed lemurs (Lemur catta) (Baker et al., 2018) and gorillas (Gorilla gorilla gorilla) (Wells et al., 2007). The effect that our scent enrichment exerted on social behaviors, with decreased rates of social interactions in red-ruffed lemurs and black howler monkeys, but no significant effects on siamangs, lar gibbons, and orangutans, might be related to differences in social organizations and structures among these species. Specifically, red-ruffed lemurs and black howler monkeys are social species living in small groups including both adult males and females, whereas siamangs and lar gibbons are monogamous and orangutans are solitary. Red-ruffed lemurs and black howler monkeys, thus, display more social affiliative behaviors which have a stress-reducing effect. We, therefore, speculate that redruffed lemurs and black howler monkeys could have reduced their rates of social behaviors because our scent enrichment might have decreased the need for reassurance-derived social interactions.
F I G U R E 3 Mean rates ± standard error of the mean of stress-related behaviors among lar gibbon and red-ruffed lemur. The regression model showed that both lar gibbon and red-ruffed lemur exhibited a significant reduction in stress-related behaviors following the exposure to the scent enrichment However, we recognize that further factors may have induced such differences between the effects of our scent enrichment on individual species; for instance, it is possible that the new unfamiliar scents increased the stress levels in red-ruffed lemurs because they perceived them more intensely than the other study species, or that decreased rates in social behaviors in red-ruffed lemurs and black howler monkeys are due to increased rates in other behaviors (such as inspections, locomotion, etc.) which were not measured during our study.
We also found a significant reduction in rates of stress-related behaviors after red-ruffed lemurs and lar gibbons were exposed to the series of essential oils, which is the most promising outcome of this preliminary study about the potential positive effect of such essential oils. Similar findings have been reported in non-primate species. For example, Uccheddu et al. (2018)  behaviors reflecting low-level acute stress or anxiety (Maestripieri et al., 1992) but FGCs reflecting high-level chronic stress (Sapolsky, 2002). Accordingly, self-directed behaviors have been found to increase in anxiety-inducing contexts, such as when animals are given anxiogenic drugs (Schino et al., 1996) or after aggression (Schino, 1998); FGC concentrations have been shown to increase when animals are exposed to high levels of stress, such as in the presence of tourists (Barja et al., 2007) or when exposed to the odor of a predator (Monclús et al., 2006). Additionally, although glucocorticoids are commonly associated with the negative aspects of stress, these steroid hormones play many important roles both in mediating the response to stress and in the circadian rhythm (McEwen, 2019).
Thus, another potential explanation for elevated FGCs is increased energy expenditure. For instance, it is possible that increased FGCs in red-ruffed lemurs may be due to enhanced positive arousal related to increased rates in other behaviors (such as investigatory behaviors and locomotion) which we did not measure during our study.
Hence, as suggested by other authors (reviewed by G. Hosey et al., 2013), we emphasize that both behavioral and physiological indicators should be used to investigate the stress levels of individual animals, whereas behavioral indicators of anxiety alone should not be interpreted as definite indicators of glucocorticoid production.
Interestingly, we found that our scent enrichment exerted both behavioral and physiological effects only on red-ruffed lemurs.
Although primates have traditionally been considered to be "microsmatic" with a simultaneous amplified emphasis on vision (Dominy & Lucas, 2001;Fornalé et al., 2012;Gerald, 2003), several studies suggest that chemical communication is important also for F I G U R E 4 The linear mixed model showed that the scent enrichment elicited an increase in fecal glucocorticoid concentration levels in red-ruffed lemurs, but not in any other study species.
primate species (reviewed by Drea, 2020). Particularly, it is established that some species rely heavily on olfaction in addition to vision and auditory senses; for instance, this is the case of several lemurs (Gould & Overdorff, 2002;Janda et al., 2019;Scordato & Drea, 2007) and squirrel monkeys (Laska et al., 2000). This would explain the significant impact of our scent enrichment on red-ruffed lemurs, rather than the other study species for which no such response was observed, as lemurs have retained a greater olfactory complexity than other lineages such as monkeys and apes. However, we recognize that other factors may have contributed to such effects of our scent enrichment on individual species; for instance, it is possible that the effect on red-ruffed lemurs could be related to their different enclosure design (i.e., "visitor walkthrough"-including a section in which the public could be very close) which ultimately could have led the lemurs being exposed to a different olfactory environment (i.e., anthropogenic) than the other study species.
Finally, we have to acknowledge some major limitations of this preliminary study. First of all, although our study is ambitious in many respects (i.e., we worked on several species, over several years, across three conditions and with multiple measures intended to assess welfare), we focused on limited data pools which included a relatively small sample size and unit of analysis. Additionally, we did not record behaviors, such as normal locomotion, foraging, inspections and investigatory actions (e.g., exploring around the scented cloths), but changes in these behaviors could also be very informative.

| CONCLUSION
This preliminary study provided contradictory findings and suggested that the application of our new scent enrichment program may affect the stress levels of zoo-housed primates; particularly in the case of primate species where odor plays a crucial role, such as red-ruffed lemurs. Following the exposure to the series of essential oils (benzoin, lavender, and lemongrass), both red-ruffed lemurs and lar gibbons exhibited significantly lower rates of stress-related behaviors, such as pacing and self-scratching. Conversely, red-ruffed lemurs also significantly increased their levels of FGCs, which however might be explained by an increase in positive arousal. However, given that the exposure to the series of essential oils entailed a significant reduction in social behaviors in red-ruffed lemurs and black howler monkeys as well as a significant increase in FGCs in redruffed lemurs, we cannot even exclude negative effects by our scent enrichment. Therefore, in conclusion, we cannot recommend using this series of essential oils widely without further evaluation.
Future work would need to expand the investigation of the effect of our scent enrichment on primate welfare by focusing on both a larger sample size and a wider range of species across the major lineages. Also, it would be crucial to test further types of scent enrichment by considering the ecological/biological relevance of the scent enrichment to the study species. Many scents, including essential oils, are chosen based on their effectiveness in humans or domestic animals, but this may not necessarily be appropriate for all animal species (Wells, 2009). In particular, as previous authors have suggested, important factors to consider for the implementation of novel olfactory enrichment programs are the mechanism of delivery of the scent and the type of scents used (Baker et al., 2018), whereas the effectiveness of any intervention should be continually monitored to inform best practices.