Primates

, Volume 59, Issue 3, pp 215–225 | Cite as

Beneficial effect of hot spring bathing on stress levels in Japanese macaques

  • Rafaela S. C. Takeshita
  • Fred B. Bercovitch
  • Kodzue Kinoshita
  • Michael A. Huffman
Original Article

Abstract

The ability of animals to survive dramatic climates depends on their physiology, morphology and behaviour, but is often influenced by the configuration of their habitat. Along with autonomic responses, thermoregulatory behaviours, including postural adjustments, social aggregation, and use of trees for shelter, help individuals maintain homeostasis across climate variations. Japanese macaques (Macaca fuscata) are the world’s most northerly species of nonhuman primates and have adapted to extremely cold environments. Given that thermoregulatory stress can increase glucocorticoid concentrations in primates, we hypothesized that by using an available hot spring, Japanese macaques could gain protection against weather-induced cold stress during winter. We studied 12 adult female Japanese macaques living in Jigokudani Monkey Park, Japan, during the spring birth season (April to June) and winter mating season (October to December). We collected faecal samples for determination of faecal glucocorticoid (fGC) metabolite concentrations by enzyme immunoassay, as well as behavioural data to determine time spent in the hot springs, dominance rank, aggression rates, and affiliative behaviours. We used nonparametric statistics to examine seasonal changes in hot spring bathing, and the relationship between rank and air temperature on hot spring bathing. We used general linear mixed-effect models to examine factors impacting hormone concentrations. We found that Japanese macaques use hot spring bathing for thermoregulation during the winter. In the studied troop, the single hot spring is a restricted resource favoured by dominant females. High social rank had both costs and benefits: dominant females sustained high fGC levels, which were associated with high aggression rates in winter, but benefited by priority of access to the hot spring, which was associated with low fGC concentrations and therefore might help reduce energy expenditure and subsequent body heat loss. This unique habit of hot spring bathing by Japanese macaques illustrates how behavioural flexibility can help counter cold climate stress, with likely implications for reproduction and survival.

Keywords

Faecal glucocorticoids Thermoregulation Dominance rank Macaca fuscata 

Introduction

Animals rely on a variety of thermoregulatory mechanisms, including morphological, physiological and behavioural adjustments, to maintain homeostatic internal body temperatures in response to severe climatic conditions. During hot weather, koalas (Phascolarctos cinereus) hug tree trunks to cool down (Briscoe et al. 2014), and chacma baboons (Papio ursinus) increase their resting time to minimize heat gain (Hill 2006). In cold weather, some nonhuman primates living in temperate zones migrate to warmer altitudes (Rhinopithecus spp.; Cui et al. 2006; Li et al. 2000; Liu and Zhao 2004) and behave in ways that can reduce heat loss, such as sun bathing (Bishop 1979; Cui et al. 2006; Liu and Zhao 2004) and huddling (Kelley et al. 2016; McFarland and Majolo 2013; Nowack et al. 2013). Autonomic thermoregulatory responses include shivering, involuntary muscle contraction, and non-shivering thermogenesis through metabolism and heat production in brown adipose tissue (Chaffee and Allen 1973; Terrien et al. 2008). Besides low temperatures, winter is often characterized by lower food abundance and quality, so maintaining a sufficiently high energy intake for keeping a stable body temperature is challenging (Li et al. 2000; Persson 2005; Tsuji 2010). By reducing heat loss, animals can conserve energy to increase their survival chances in cold environments (Grueter et al. 2013).

Japanese macaques (Macaca fuscata) are the most northerly species of nonhuman primates in the world; they have adapted to live in seasonally cold environments (Yamagiwa 2010). Hori et al. (1977) reported that during winter, Japanese macaques living in Jigokudani Monkey Park maintain their normal body temperature, possibly due to having thicker and longer fur during this season. Furthermore, these Japanese macaques have the unusual habit of bathing in a hot spring pool. This behaviour was first observed in 1963, when a juvenile female was observed bathing in an outdoor hot spring belonging to a nearby hotel on a snowy day. Other juveniles and adults did likewise, until the end of March, when they ceased entering the hot spring. They were next observed bathing the following winter (Suzuki 1965). For hygienic purposes, the park management decided to build a hot spring only for the monkeys, and by 2003, 31% of the females in the group regularly bathed in winter (Zhang et al. 2007). The fact that this behaviour mostly occurs during winter suggests that the monkeys use the hot spring to stay warm (Suzuki 1965; Zhang et al. 2007), but to date there are no physiological data in support of this hypothesis.

In humans, taking hot baths is reported to lower stress levels and can aid in thermoregulation (Toda et al. 2006). Humans respond to cold stress with a cortisol response, increasing metabolic heat production (Izawa et al. 2009; Leppäluoto et al. 2008). In nonhuman primates, faecal glucocorticoids (fGC) have been widely used as a noninvasive way to measure stress in wild and captive populations (Behringer and Deschner 2017; Heistermann et al. 2006; Monfort 2003), and previous studies have reported an increase in fGC levels during cold stress in Japanese macaques (Takeshita et al. 2014), chacma baboons (Weingrill et al. 2004) and geladas (Theropithecus gelada; Beehner and McCann 2008).

We hypothesized that Japanese macaques in Jigokudani Monkey Park adopt hot spring bathing as a thermoregulatory mechanism to cope with cold stress. If the hot spring is a restricted but high-value resource, in a large group, high-ranking individuals will have priority of access compared to lower-ranking individuals. We predicted that: (1) Japanese macaques will spend more time in the hot spring during winter than during summer, and during colder ambient temperatures, (2) high-ranking females will spend more time in the hot spring than low-ranking females, (3) fGC concentrations should be lower when individuals bathe in the hot spring than when they do not bathe, and (4) the time spent in the hot spring will be negatively associated with fGC concentrations.

Methods

Study site and subjects

This study was conducted over 2 months of the spring birth season (April to June) and 2 months of the winter mating season (October to December) of 2014 on a troop of Japanese macaques at Jigokudani Monkey Park (JMP), Japan (36º43′58″N, 138º27′46″E). Individuals were identified by physical features such as scars, nose shape and birth marks. The group contained about 160 individuals, including adult females (> 5 years old), adult males (estimated to be > 7 years old), sub-adult males (estimated to be 5–7 years old), juveniles (1–4 years old) and infants (< 1 year old). Age classifications are based upon Hamada and Yamamoto (2010). The hot spring available to the monkeys is ca. 40 m2 and 40 cm deep, with a water temperature of ca. 40 °C. Tourists are not allowed to feed the monkeys or to enter the hot spring. Food is provided daily to the group by JMP staff at 09:00, 12:00 and 15:00, consisting of ca. 20 kg of wheat grains and a small amount of soybeans. Apples are provided to the monkeys at the end of the day in spring but not winter. The monkeys also consume natural foods and spend the nights in the mountain forest outside of the park.

Data collection

From approximately 50 adult females in the group, we selected 12 individuals (mean age ± SD = 11.58 ± 4.48 years) to obtain a representative sample of socially high-, middle- and low-ranking individuals, based on information provided by park staff. Table 1 provides key information about our study subjects. Eight females gave birth during spring; they were classified as pregnant until parturition, and then as early lactating until the end of spring, and as mid-lactating during winter. One female’s infant died 2 days after birth. Six faecal samples from this female after the infant died were removed from the analysis to avoid potential consequences for fGC levels. Because all four females who did not give birth had a 1-year-old infant, around the age of weaning (Hamada and Yamamoto 2010), they were classified as late lactating in spring (birth season) and as cycling the following winter (mating season). To determine which females became pregnant by the end of the study (winter), we used birth records provided by JMP staff the following year. We estimated the conception date by subtracting the mean gestation period of the species (173 days; Nigi 1976) from the birth date.
Table 1

Classification of subjects

ID

Rank position

Reproductive state (spring/birth season)

Reproductive state (winter/mating season)

Number of bathing weeks (Total = 17)

Absolute

Relative

Tokkuri00

1

1

Pregnant/early lactating

Mid-lactating/pregnant

7

Tokkuri92

2

2

Pregnant/early lactating

Mid-lactating

8

Towano99

6

3

Pregnant/early lactating

Mid-lactating

3

Wakako05

7

4

Pregnant/early lactating

Mid-lactating

4

Toeiti9807

9

5

Late lactating

Cycling

5

Tomano9404

10

6

Pregnant/early lactating

Mid-lactating/pregnant

9

Toraiti03

14

7

Pregnant/early lactating

Mid-lactating

1

Towashiti03

23

8

Late lactating

Cycling/pregnant

0

Togura9398

24

9

Pregnant/early lactating

Mid-lactating

0

Toeko9505

30

10

Late lactating

Cycling

1

Keroiti05

40

11

Late lactating

Cycling/pregnant

8

Togura9308

42

12

Pregnant/infant lossa

Cycling

0

aSamples collected from infant loss until the end of spring were excluded from the analyses

Faecal samples were collected opportunistically between 9:00 and 16:00 from all subjects, resulting in a total of 212 and 166 faecal samples during spring and winter, respectively (weekly mean ± SD = 1.77 ± 1.32 samples/female). After collection, samples were stored in a cooler bag until transfer to a freezer (−20 °C) within 1 h after collection. We discarded any sample contaminated with water or urine.

We used focal instantaneous sampling (Martin and Bateson 1993) to determine rates of aggression given (e.g., bites, chases, and attacks) and received (e.g. grimaces, screams, flights), time spent (continuously) in affiliative interactions (grooming, huddling), and hot spring bathing time of focal individuals, recording every minute in a 30-min session (3 sessions/week/subject). We collected 115.5 h of data during spring and 160 h during winter (mean ± SD = 22.95 ± 0.58 h/female). The three sessions for each individual each week were scheduled at different times of the day (morning, early afternoon, late afternoon), to get a representative pattern of individual activities throughout the day. We standardized data on hot spring bathing in two ways. First, we categorized each week as a ‘bathing week’ if the female was observed in the hot spring, and as a ‘non-bathing week’ if the female was not observed in the hot spring. Second, we quantified the data by calculating the weekly average amount of time each individual spent in the hot spring. In addition, we recorded all adult females’ agonistic and approach–avoidance interactions between dyads on an ad libitum basis (772 h) to assist in establishing the social rank position of our focal subjects, using David’s score (Gammell et al. 2003). Table 1 indicates the absolute rank positions of our subjects in the group and the number of bathing weeks per individual.

We recorded daily minimum temperature using a mercury thermometer (Ishihara Ondokei Seisakusho Y.K., Tokyo, Japan) placed in the shade, about 5 m above the ground. We also recorded the daily number of visitors to the park.

Hormone extraction

Hormones were extracted from faecal samples and analysed by adding 5 ml of 80% methanol to 0.10 g of freeze-dried, pulverized faeces. After centrifugation (3000 rpm × 5 min), the supernatant was subjected to hormonal analyses by enzyme immunoassay for glucocorticoid determination, following a method previously used for other species (Kinoshita et al. 2011; Mendonça et al. 2016).

Hormonal analyses

To validate the glucocorticoid assay for Japanese macaques, we conducted an adrenocorticotropic hormone (ACTH) challenge at the Primate Research Institute of Kyoto University. We collected daily faecal samples from two adult male Japanese macaques starting from 2 to 3 days before administration of ACTH (1.8 UI/kg) to the experimental subject, or saline solution (1.274 ml) to the control subject to determine baseline fGC concentrations. We used male subjects to avoid possible influences of female reproductive state on circulating glucocorticoid levels. We then collected faecal samples, whenever possible, at 12, 24, 36, 48, 60 and 72 h post challenge (N = 17, 8.5 ± 2 per male). We observed a peak in fGC levels 1 day after the procedure in the ATCH-challenged individual, but not in the control (Fig. 1), which indicates that the assay detected fGC response. The polyclonal cortisol antiserum (FKA404E; Cosmo Bio Co., Ltd.) which was raised in rabbits against cortisol-3-CMO-BSA cross-reacted 100% with cortisol (compd. F), 11.5% with 11-deoxycortisol (compd. S), 4% with cortisone (compd. E), 2% with corticosterone (compd. B), 0.2% with 17alpha-hydroxy-11-deoxy-corticosterone (compd. A), 0.04% with 17alpha-hydroxypProgesterone and 0.0% with other steroids. To test accuracy, five pools of faecal extracts containing samples from three different monkeys were spiked with cortisol standards. The mean recovery was 10l.3%. To test for parallelism, four pools of faecal extracts containing samples from three different monkeys were serially diluted. An F test showed that the slope generated by the four pools did not differ from the standard curve (F = 0.74, P = 0.67). The sensitivity of the assay was 0.02 ng/ml. The mean intra-assay coefficient of variance was 6.6% (N = 378) and the inter-assay coefficient of variance was 13.7% (N = 21). This research adhered to the Primate Society of Japan (PSJ) principles for the ethical treatment of nonhuman primates. All manipulations of the subjects were approved (research clearance no. 2015-127) by the Primate Research Institute (PRI) Ethics Committee of Kyoto University and conformed to the PRI’s Guidelines for Care and Use of Nonhuman Primates (PRI 2010).
Fig. 1

Changes on fGC concentrations with time in two Japanese macaques after administration of ACTH (solid line) and saline solution (control, dashed line)

Statistical analyses

All statistical tests were conducted in R software (3.3.0). To examine the effects of season on hot spring bathing, we used non-parametric tests due to the highly skewed data with zero inflation. To control for pseudo-replication of individual data, we calculated individual averages of time spent in the hot spring each season. We used Wilcoxon signed-ranks tests with continuity correction to compare bathing time between winter and spring, and Spearman’s correlation to examine the relationship between social rank and bathing time. In addition, we tested the effect of ambient temperature by calculating the weekly means of minimum temperature and the weekly means of bathing time per female only during bathing weeks. We removed non-bathing weeks in this analysis to exclude possible effects of restricted access to the hot spring (e.g., presence of too many dominant females, or harassment by males).

For testing predictions about hormonal concentrations, the data were normalized by log transformation, and given that there were no zero-inflation issues, we used general linear mixed-effect models (GLMM) with the package “lme4”. Before each model, we tested for potential multicollinearity between fixed factors by calculating variance inflation factors (VIF) with the package “car”. All fixed factors with a VIF value > 2 were considered problematic and removed from the model. To assess the equality of variances of categorical fixed factors, we used Levene’s test (Levene 1960). Because season was not homogeneous (F1169 = 4.35, P = 0.03), we analysed each season separately. To compare hormonal concentrations between seasons, we calculated average fGC levels for each female per season and used the Wilcoxon test, controlling for reproductive state.

First, using individual samples, we tested the possible effect of sampling time on fGC concentrations. To control for season, we selected only spring samples. The initial model included collection time (morning or afternoon), age, reproductive state, daily minimum temperature, and rank position, with individual ID as random effect. Following Burnham and Anderson (2002), we sequentially removed the fixed factors and interactions to select the model with the lowest Akaike information criterion (AIC). We also considered other models that did not differ significantly from the best model (∆AIC < 2). The final model included only reproductive state, daily minimum temperature, and collection time. As in a previous study of Japanese macaques (Takeshita et al. 2017), we found no difference between morning and afternoon samples (Table 2), so we used the weekly mean of all our quantitative data (fGC, daily minimum temperature, number of tourists, time spent in the hot spring, rates of aggression [given, received, and both], and time spent in affiliative behaviours) per individual in the following analyses.
Table 2

Mixed-effect models parameter estimates showing the relationship between log-transformed fGC and fixed effects in Japanese macaques

Model

Variable

Estimate

Std. error

Z value

P value

1. Collection time (spring only, individual samples)

Intercept

5.111

0.159

32.084

< 0.001

Collection time (morning)

0.138

0.102

1.346

0.178

Reproductive state (pregnant)

0.466

0.153

3.051

0.002

Reproductive state (late lactating)

0.162

0.215

0.755

0.450

Daily minimum temperature

0.057

0.014

4.092

< 0.001

2. Spring birth season (weekly means of fgc)

Intercept

4.873

0.181

16.953

< 0.001

Weekly means of minimum temperature

0.071

0.015

4.731

< 0.001

Bathing week (yes)

0.050

0. 194

0.258

0.796

Rank

0.004

0.007

0.538

0.590

3. Winter mating season (weekly means of fGC)

Intercept

5.418

0.133

40.616

<0.001

Weekly means of aggression rates

0.049

0.030

1.654

0.098

Bathing week (yes)

0.282

0.124

2.286

0.023

Rank

0.015

0.005

3.059

0.002

All variables included in the final selected models are shown. Comparisons were made against the intercept of first levels of each factor (collection time = afternoon, reproductive state = early lactation, bathing week = no). Individual identity was entered as random effect. Data are based on 378 samples of 12 female Japanese macaques

We built one model per season, initially including all fixed factors (reproductive state, age, bathing/non-bathing week, weekly means of daily minimum temperature, relative rank, and weekly means of: tourists, time spent in the hot spring, rates of aggression, and time spent in affiliative behaviour), with individual as random factor. We selected the best model following the same methods described above. Measures of central tendency report the mean ± standard error. We set the alpha level at 0.05.

Results

Daily temperature varied from −4 to 13 °C (mean = 4.5 °C) in spring and from −7 to 7 °C (mean = 0.8 °C) in winter. Females spent significantly more time in the hot spring during winter than during spring (Fig. 2; Wilcoxon signed rank test: V = 44, N = 12, P = 0.01), and there was a significant negative correlation between weekly means of daily minimum temperature and time spent in the hot spring during bathing weeks (Fig. 3; Spearman rank correlation: rs = −0.80, N = 12, P = 0.0001).
Fig. 2

Seasonal differences in hot spring bathing time in 12 female Japanese macaques. Error bars indicate standard error. *P = 0.01

Fig. 3

Correlation between hot spring bathing time and ambient temperature. Each dot represents the mean values of one week of the study period. Error bars indicate the standard error

Dominance rank was positively correlated with bathing time during the winter (Spearman rank correlation: rs = −0.76, N = 12, P = 0.004), with high-ranking females spending significantly more time in the hot spring than low-ranking females (Fig. 4), but there was no significant correlation in spring (Spearman rank correlation: rs = −0.37, N = 12, P = 0.32).
Fig. 4

Correlation between means of time spent in the hot spring and relative rank order of 12 female Japanese macaques during winter. Error bars indicate standard error. The lower the number, the higher the relative rank position

The first model to test our hormonal predictions included individual samples from the spring season, and showed no effect of sample collection time or of subject age, but early-lactating females had higher fGC levels than pregnant females, and daily minimum temperature was positively related to fGC levels. A Wilcoxon test showed no significant difference in fGC concentrations between spring (323.24 ± 41.70 ng/g) and winter (199.39 ± 23.99 ng/g), using only individuals that were not pregnant during spring (Wilcoxon paired test: V = 3, N = 4, P = 0.625). The following two models were analysed using weekly average fGC levels. The spring model revealed that weekly mean minimum temperature was positively associated with fGC, consistent with the first model using individual samples (Fig. 5). There was no effect of rank, bathing week, age, reproductive state, number of tourists, time spent in the hot spring, or aggressive or affiliative behaviours on fGC.
Fig. 5

Correlation between weekly means of fGC and weekly means of daily minimum temperature during the spring and winter in 12 female Japanese macaques. Each dot represents the weekly mean of one female

For the winter season samples, the best model included bathing behaviour, dominance rank, and aggression rates (Table 2). Dominance rank significantly influenced hormonal levels, with higher fGC in high-ranking than low-ranking females. There was also a significant effect of bathing behaviour, with lower fGC during bathing weeks than non-bathing weeks (Fig. 6), but no effect of weekly means of time spent in the hot spring. Aggression rates showed a positive association with fGC levels. One model that did not differ significantly from the best model (delta AIC < 1.06) included grooming time in addition to bathing behaviour, dominance rank, and aggression rates. Although grooming was not significantly correlated with fGC levels in this model (GLMM: 0.008837 ± 0.009044, Z = 0.977, P = 0.34), aggression rates had a stronger positive correlation with fGC levels (GLMM: 0.056187 ± 0.030469, Z = 1.844, P = 0.06). Bathing behaviour (GLMM: −0.315148 ± 0.128326, Z = −2.456, P = 0.01) and dominance rank (GLMM: −0.015446 ± 0.004879, Z = −3.166, P = 0.002) also had a significant relationship with fGC levels, in agreement with the best model. Weekly means of minimum temperature, number of tourists, reproductive state, and age were not included in any of the winter models, and thus did not significantly influence fGC concentrations.
Fig. 6

Comparison of weekly means of fGC of 12 female Japanese macaques during hot spring bathing weeks and non-bathing weeks in winter (mating season) and spring (birth season). Error bars indicate standard error. *P < 0.05; NS not significant

Discussion

Our results confirmed that female Japanese macaques use the hot spring for a significantly longer time during winter than spring, and during colder weeks. Higher-ranking females spent significantly more time bathing in the hot spring than lower-ranking females. Daily minimum temperature was positively correlated with fGC during spring, but not winter. In contrast, during winter, females had significantly lower fGC levels in bathing weeks than non-bathing weeks, and higher-ranking females had higher fGC than low-ranking females.

The association of hot spring bathing time with winter and low ambient temperatures supports the premise that the macaques use the hot spring for thermoregulation. One study using human serum samples has shown that cortisol levels initially increase during a hot bath due to heat stress (Moller et al. 1989), but they should progressively decrease with time, especially when the air temperature is low (Agishi and Ohtsuka 1998). In addition, daily hot bathing has been shown to reduce stress levels and improve women’s sleep during winter (Sung and Tochihara 2000). Our study showed no correlation between bathing time and fGC levels, but we found that the weekly mean levels of fGC were lower during bathing weeks, indicating that, as in humans, the hot spring has a stress-reducing effect in Japanese macaques. The lack of association between fGC and bathing time might be due to the time lag between hormonal secretion in the serum and excretion in faeces, preventing detection of the immediate adrenal response. Further investigation using serum or saliva samples might be useful to detect short-term changes in fGC levels.

The beneficial effects of the hot spring can explain the influence of rank on bathing time. Zhang et al. (2007) reported that dominant female Japanese macaques spent more time in the hot spring than subordinate females during winter, a pattern duplicated during this study. The two reports strongly suggest that the hot spring is a limited resource, with dominant females restricting access for lower-ranking females. Despite their priority of access, higher-ranking females had higher fGC than low-ranking females, similar to male Japanese macaques (Barrett et al. 2002) and female lemurs (Lemur catta; Cavigelli 1999). By contrast, studies have reported the opposite relationship in male baboons (Papio anubis; Sapolsky 1990), or no significant association, as in male long-tailed macaques (Macaca fascilularis; Girard-Buttoz et al. 2009), female Japanese macaques (Takeshita et al. 2014), female baboons (Papio spp.; Weingrill et al. 2004), and female mandrills (Mandrillus sphinx; Setchell et al. 2008). As reviewed by Creel (2001), this variation is probably due to the fact that several factors can influence fGC levels and its relationship to dominance rank, including environmental setting (captive/wild), season, food availability, genes, and whether the individuals receive social support or not (Abbott et al. 2003). In this study, the rates of aggression had a weak, but positive association with higher rank, supporting the hypothesis of the costliness of maintaining a high position in the hierarchy (Barnard et al. 1993; Creel 2005). The relationship between aggression and fGC might have been weakened due to other factors, including hot spring bathing, which was shown to have a negative relationship with fGC levels. Possibly, changes in fGC levels caused by aggression were weakened by hot spring bathing in a 1-week window. However, dominance rank was a good predictor of fGC levels, and indicates the overall pattern of stress levels associated with social dynamics.

The lack of influence of rank and aggression on stress levels in spring suggests that females experience higher social stress during winter. Besides the cold and reduced food availability, winter corresponds to the mating season and is characterized by increased agonistic interactions and copulations (Huffman 1987). This is consistent with previous studies showing elevated fGC levels during the mating season in female and male Japanese macaques (Takeshita et al. 2014, 2017). However, we did not find a significant difference in fGC levels between seasons. In addition, the effect of daily minimum temperature on fGC only during the non-mating season was in contrast to previous studies that showed that fGC levels are higher in low temperatures, as a response to cold stress (Beehner and McCann 2008; Takeshita et al. 2014; Weingrill et al. 2004). In the current study, the lack of a seasonal difference in fGC levels or temperature associations with stress levels in winter suggest that hot spring bathing does aid in thermoregulation of Japanese macaques. However, our results suggest that increasing spring temperature raises fGC levels during acclimatization to the changing weather (minimum temperature varied between −4 and 13 °C during spring). Given that Japanese macaques do not sweat to dissipate heat (Tokura and Sugiyama 1975), the rise in ambient temperature could increase their body temperature, resulting in elevated fGC as a way to provide energy for keeping a stable body temperature. Furthermore, we observed the highest peaks of fGC towards the end of spring, which coincides with the period of moulting in this species (Inagaki and Nigi 1988). In northern elephant seals (Mirounga angustirostris), moulting has been associated with high fGC concentrations due to metabolic demands for new pelage synthesis and the need to spare mobilized protein (Champagne et al. 2015). Our data suggests that Japanese macaques undergo the same pattern of elevated fGC levels during springtime moulting.

Reproductive state did not interact with bathing behaviour, indicating no influence of pregnancy or lactation in fGC response to hot spring bathing. However, early-lactating females had higher fGC levels than pregnant females. In mammals, fGC levels vary throughout pregnancy, but usually increase at late gestation (Fanson and Parrott 2015). At least three explanations might apply to our results: (1) we did not obtain samples from all females shortly before parturition, when fGC levels are highest; (2) the latency for cortisol to return to pre-pregnancy levels after parturition means that high fGC levels of early-lactating females might be remaining from pregnancy levels; (3) early-lactating females are experiencing nutritional or social stress. The first explanation is associated with the high individual variability in fGC levels, as reported by Bardi et al. (2003) in peripartum Japanese macaques, and we acknowledge that our sample size was small. The second explanation is corroborated by the lack of significant differences among reproductive states when using weekly means of fGC levels. The third explanation is supported by previous studies showing that lactating primates have high fGC levels due to the energetic demands of milk production (Maestripieri et al. 2008; Setchell et al. 2008; Weingrill et al. 2004). In baboons, the costs of lactation are higher towards weaning (Altmann 2001), which in Japanese macaques is completed at 1 year of age (Nakayama et al. 1971). These explanations, along with the fact that weekly average fGC levels did not differ among the three reproductive states, suggest that pregnant, early-, and late-lactating females undergo similar levels of physiological stress.

Tourism did not affect stress hormones in this population of Japanese macaques, a finding opposite to that reported in spider monkeys (Ateles hybridus; Rimbach et al. 2013), howler monkeys (Alouatta pigra; Behie et al. 2010), Barbary macaques (Marechal et al. 2011), and European pine martens (Martes martes; Barja et al. 2007). The Jigokudani monkeys have been provisioned and habituated for over 50 years, with an average of 480 visitors per day. Studies of some populations of orangutans (Pongo pygmaeus morio; Muehlenbein et al. 2012), howler monkeys (Alouatta seniculus; Rimbach et al. 2013), and spotted hyenas (Crocuta crocuta; Van Meter et al. 2009) have also shown that habituated mammals are not stressed by the presence of tourists. Hence, no universal rule can be applied to animals with regard to whether or not ecotourism is stressful.

In conclusion, our data strongly indicate that Japanese macaques bathe in the hot spring as a thermoregulatory strategy to cope with cold during the winter months. High-ranking females can take advantage of their position to spend more time bathing, but high social rank comes with costs, reflected in high fGC, which is in turn associated with aggression rates. Because hot spring bathing is reported in only one population of Japanese macaques, we presume that other populations use alternative ways to cope with the bitter cold winter temperatures, such as huddling (Hori et al. 1977) and forming large clusters of individuals (Zhang and Watanabe 2007). We suggest that hot spring bathing by this group of Japanese macaques is an opportunistic tradition that provides physiological benefits to the monkeys. Despite the uniqueness of this behaviour and its limited distribution, it demonstrates behavioural flexibility as an adaptive mechanism to cope with cold stress.

Notes

Acknowledgements

We would like to thank Mr. Haruo Takehushi, Mr. Toshio Hagiwara, Mr. Atsushi Takizawa, Mr. Yukihiro Sato, and Ms. Kayo Miyata from Jigokudani Monkey Park and Ms. Yukari Murano for their assistance during the study and for providing us support and useful information. Our gratitude to Mr. Akihisa Kaneko, Ms. Mayumi Morimoto, and all staff from Center for Human Evolution Modeling Research at the Primate Research Institute for their assistance during the biological validation of the hormonal assay. We also thank the associate editor and two anonymous reviewers for their constructive comments, which helped us to improve the manuscript. The study was funded by the Primate Research Institute; the Leading Program in Primatology and Wildlife Science (PWS); a grant-in-aid from the Japan Society for the Promotion of Science (JSPS) no. 16J00399, and a scholarship to RSCT by the Nippon Foundation.

References

  1. Abbott DH, Keverne EB, Bercovitch FB, Shively CA, Mendoza SP et al (2003) Are subordinates always stressed? A comparative analysis of rank differences in cortisol levels among primates. Horm Behav 43:67–82CrossRefPubMedGoogle Scholar
  2. Agishi Y, Ohtsuka Y (1998) Present features of balneotherapy in Japan. Glob Environ Res 2:177–186Google Scholar
  3. Altmann J (2001) Baboon mothers and infants. University of Chicago Press, ChicagoGoogle Scholar
  4. Bardi M, Shimizu K, Barrett GM, Borgognini-Tarli SM, Huffman MA (2003) Peripartum cortisol levels and mother-infant interactions in Japanese macaques. Am J Phys Anthropol 120:298–304.  https://doi.org/10.1002/ajpa.10150 CrossRefPubMedGoogle Scholar
  5. Barja I, Silvan G, Rosellini S, Pineiro A, Gonzalez-Gil A et al (2007) Stress physiological responses to tourist pressure in a wild population of European pine marten. J Steroid Biochem Mol Biol 104:136–142.  https://doi.org/10.1016/j.jsbmb.2007.03.008 CrossRefPubMedGoogle Scholar
  6. Barnard CJ, Behnke JM, Sewell J (1993) Social behaviour, stress and susceptibility to infection in house mice (Mus musculus): effects of duration of grouping and aggressive behaviour prior to infection on susceptibility to Babesia microti. Parasitology 107:183–192CrossRefPubMedGoogle Scholar
  7. Barrett GM, Shimizu K, Bardi M, Asaba S, Mori A (2002) Endocrine correlates of rank, reproduction, and female-directed aggression in male Japanese macaques (Macaca fuscata). Horm Behav 42:85–96CrossRefPubMedGoogle Scholar
  8. Beehner JC, McCann C (2008) Seasonal and altitudinal effects on glucocorticoid metabolites in a wild primate (Theropithecus gelada). Physiol Behav 95:508–514CrossRefPubMedGoogle Scholar
  9. Behie AM, Pavelka MS, Chapman CA (2010) Sources of variation in fecal cortisol levels in howler monkeys in Belize. Am J Primatol 72:600–606.  https://doi.org/10.1002/ajp.20813 PubMedGoogle Scholar
  10. Behringer V, Deschner T (2017) Non-invasive monitoring of physiological markers in primates. Horm Behav 91:3–18.  https://doi.org/10.1016/j.yhbeh.2017.02.001 CrossRefPubMedGoogle Scholar
  11. Bishop NH (1979) Himalayan langurs: temperate colobines. J Hum Evol 8:251–281CrossRefGoogle Scholar
  12. Briscoe NJ, Handasyde KA, Griffiths SR, Porter WP, Krockenberger A et al (2014) Tree-hugging koalas demonstrate a novel thermoregulatory mechanism for arboreal mammals. Biol Lett.  https://doi.org/10.1098/Rsbl.2014.0235 PubMedPubMedCentralGoogle Scholar
  13. Burnham KP, Anderson DR (2002) Model selection and multi-model inference: a practical information-theoretic approach. Springer, New YorkGoogle Scholar
  14. Cavigelli SA (1999) Behavioural patterns associated with faecal cortisol levels in free-ranging female ring-tailed lemurs, Lemur catta. Anim Behav 57:935–944.  https://doi.org/10.1006/anbe.1998.1054 CrossRefPubMedGoogle Scholar
  15. Chaffee RR, Allen JR (1973) Effects of ambient temperature on the resting metabolic rate of cold- and heat-acclimated Macaca mulatta. Comp Biochem Physiol A 44:1215–1225CrossRefPubMedGoogle Scholar
  16. Champagne C, Tift M, Houser D, Crocker D (2015) Adrenal sensitivity to stress is maintained despite variation in baseline glucocorticoids in moulting seals. Conservation physiology 3:cov004CrossRefPubMedPubMedCentralGoogle Scholar
  17. Creel S (2001) Social dominance and stress hormones. Trends Ecol Evol 16:491–497CrossRefGoogle Scholar
  18. Creel S (2005) Dominance, aggression, and glucocorticoid levels in social carnivores. J Mammal 86:255–264CrossRefGoogle Scholar
  19. Cui LW, Quan RC, Xiao W (2006) Sleeping sites of black-and-white snub-nosed monkeys (Rhinopithecus bieti) at Baima Snow Mountain, China. J Zool 270:192–198Google Scholar
  20. Fanson KV, Parrott ML (2015) The value of eutherian-marsupial comparisons for understanding the function of glucocorticoids in female mammal reproduction. Horm Behav 76:41–47.  https://doi.org/10.1016/j.yhbeh.2015.05.012 CrossRefPubMedGoogle Scholar
  21. Gammell MP, de Vries H, Jennings DJ, CoM Carlin, Hayden TJ (2003) David’s score: a more appropriate dominance ranking method than Clutton-Brock et al’s index. Anim Behav 66:601–605.  https://doi.org/10.1006/anbe.2003.2226 CrossRefGoogle Scholar
  22. Girard-Buttoz C, Heistermann M, Krummel S, Engelhardt A (2009) Seasonal and social influences on fecal androgen and glucocorticoid excretion in wild male long-tailed macaques (Macaca fascicularis). Physiol Behav 98:168–175.  https://doi.org/10.1016/j.physbeh.2009.05.005 CrossRefPubMedGoogle Scholar
  23. Grueter CC, Li D, Ren B, Li M (2013) Overwintering strategy of Yunnan snub-nosed monkeys: adjustments in activity scheduling and foraging patterns. Primates 54:125–135CrossRefPubMedGoogle Scholar
  24. Hamada Y, Yamamoto A (2010) Morphological characteristics, growth, and aging in Japanese macaques. In: Nakagawa N, Nakamichi M, Sugiura H (eds) The Japanese macaques. Springer, Kyoto, pp 27–52CrossRefGoogle Scholar
  25. Heistermann M, Palme R, Ganswindt A (2006) Comparison of different enzyme immunoassays for assessment of adrenocortical activity in primates based on fecal analysis. Am J Primatol 68:257–273.  https://doi.org/10.1002/ajp.20222 CrossRefPubMedGoogle Scholar
  26. Hill RA (2006) Thermal constraints on activity scheduling and habitat choice in baboons. Am J Phys Anthropol 129:242–249CrossRefPubMedGoogle Scholar
  27. Hori T, Nakayama T, Tokura H, Hara F, Suzuki M (1977) Thermoregulation of Japanese macaque living in a snowy mountain area. Jap J Physiol 27:305–319CrossRefGoogle Scholar
  28. Huffman MA (1987) Consort intrusion and female mate choice in Japanese macaques (Macaca fuscata). Ethology 75:221–234CrossRefGoogle Scholar
  29. Inagaki H, Nigi H (1988) Annual changes in hair length of the Japanese monkey (Macaca fuscata fuscata). Primates 29:81–89CrossRefGoogle Scholar
  30. Izawa S, Kim K, Akimoto T, Ahn N, Lee H et al (2009) Effects of cold environment exposure and cold acclimatization on exercise-induced salivary cortisol response. Wilderness Environ Med 20:239–243CrossRefPubMedGoogle Scholar
  31. Kelley EA, Jablonski NG, Chaplin G, Sussman RW, Kamilar JM (2016) Behavioral thermoregulation in Lemur catta: the significance of sunning and huddling behaviors. Am J Primatol 78:745–754.  https://doi.org/10.1002/ajp.22538 CrossRefPubMedGoogle Scholar
  32. Kinoshita K, Inada S, Seki K, Sasaki A, Hama N et al (2011) Long-term monitoring of fecal steroid hormones in female snow leopards (Panthera uncia) during pregnancy or pseudopregnancy. PLoS One 6:e19314.  https://doi.org/10.1371/journal.pone.0019314 CrossRefPubMedPubMedCentralGoogle Scholar
  33. Leppäluoto J, Westerlund T, Huttunen P, Oksa J, Smolander J et al (2008) Effects of long-term whole-body cold exposures on plasma concentrations of ACTH, beta-endorphin, cortisol, catecholamines and cytokines in healthy females. Scand J Clin Lab Invest 68:145–153CrossRefPubMedGoogle Scholar
  34. Levene H (1960) Robust tests for equality of variances. In: Olkin I (ed) Contributions to probability and statistics, vol 1. Stanford University Press, Stanford, pp 278–292Google Scholar
  35. Li B, Chen C, Ji W, Ren B (2000) Seasonal home range changes of the Sichuan snub-nosed monkey (Rhinopithecus roxellana) in the Qinling Mountains of China. Folia Primatol 71:375–386CrossRefPubMedGoogle Scholar
  36. Liu Z-H, Zhao Q-K (2004) Sleeping sites of Rhinopithecus bieti at Mt. Fuhe, Yunnan. Primates 45:241–248CrossRefPubMedGoogle Scholar
  37. Maestripieri D, Hoffman CL, Fulks R, Gerald MS (2008) Plasma cortisol responses to stress in lactating and nonlactating female rhesus macaques. Horm Behav 53:170–176.  https://doi.org/10.1016/j.yhbeh.2007.09.013 CrossRefPubMedGoogle Scholar
  38. Marechal L, Semple S, Majolo B, Qarro M, Heistermann M et al (2011) Impacts of tourism on anxiety and physiological stress levels in wild male Barbary macaques. Biol Conserv 144:2188–2193.  https://doi.org/10.1016/j.biocon.2011.05.010 CrossRefGoogle Scholar
  39. McFarland R, Majolo B (2013) Coping with the cold: predictors of survival in wild Barbary macaques Macaca sylvanus. Biol Lett 9:20130428.  https://doi.org/10.1098/rsbl.2013.0428 CrossRefPubMedPubMedCentralGoogle Scholar
  40. Mendonça RS, Takeshita RS, Kanamori T, Kuze N, Hayashi M et al (2016) Behavioral and physiological changes in a juvenile Bornean orangutan after a wildlife rescue. Glob Ecol Conserv 8:116–122CrossRefGoogle Scholar
  41. Moller N, Beckwith R, Butler PC, Christensen NJ, Orskov H et al (1989) Metabolic and hormonal responses to exogenous hyperthermia in man. Clin Endocrinol (Oxf) 30:651–660.  https://doi.org/10.1111/j.1365-2265.1989.tb00271.x CrossRefGoogle Scholar
  42. Monfort SL (2003) Non-invasive endocrine measures of reproduction and stress in wild populations. In: Holt W, Pickard A, Rodger J, Wildt D (eds) Reproductive science and integrated conservation biology, vol 8. Cambridge University Press, Cambridge, pp 147–165Google Scholar
  43. Muehlenbein MP, Ancrenaz M, Sakong R, Ambu L, Prall S et al (2012) Ape conservation physiology: fecal glucocorticoid responses in wild Pongo pygmaeus morio following human visitation. PLoS One.  https://doi.org/10.1371/journal.pone.0033357 PubMedPubMedCentralGoogle Scholar
  44. Nakayama T, Hori T, Nagasaka T, Tokura H, Tadaki E (1971) Thermal and metabolic responses in the Japanese monkey at temperatures of 5-38 degrees C. J Appl Physiol 31:332–337CrossRefPubMedGoogle Scholar
  45. Nigi H (1976) Some aspects related to conception of the Japanese monkey (Macaca fuscata). Primates 17:81–87CrossRefGoogle Scholar
  46. Nowack J, Wippich M, Mzilikazi N, Dausmann KH (2013) Surviving the cold, dry period in Africa: behavioral adjustments as an alternative to heterothermy in the African lesser bushbaby (Galago moholi). Int J Primatol 34:49–64CrossRefGoogle Scholar
  47. Persson J (2005) Female wolverine (Gulo gulo) reproduction: reproductive costs and winter food availability. Can J Zool 83:1453–1459CrossRefGoogle Scholar
  48. Primate Research Institute (2010) Guidelines for Care and Use of Nonhuman Primates Primate Research Institute. Kyoto University, KyotoGoogle Scholar
  49. Rimbach R, Link A, Heistermann M, Gomez-Posada C, Galvis N et al (2013) Effects of logging, hunting, and forest fragment size on physiological stress levels of two sympatric ateline primates in Colombia. Conserv Physiol 1:cot031.  https://doi.org/10.1093/conphys/cot031 CrossRefPubMedPubMedCentralGoogle Scholar
  50. Sapolsky RM (1990) Adrenocortical function, social rank, and personality among wild baboons. Biol Psychiatry 28:862–878CrossRefPubMedGoogle Scholar
  51. Setchell JM, Smith T, Wickings EJ, Knapp LA (2008) Factors affecting fecal glucocorticoid levels in semi-free-ranging female mandrills (Mandrillus sphinx). Am J Primatol 70:1023–1032.  https://doi.org/10.1002/ajp.20594 CrossRefPubMedGoogle Scholar
  52. Sung EJ, Tochihara Y (2000) Effects of bathing and hot footbath on sleep in winter. J Physiol Anthropol Appl Human Sci 19:21–27CrossRefPubMedGoogle Scholar
  53. Suzuki A (1965) An ecological study of wild Japanese monkeys in snowy areas. Primates 6:31–72CrossRefGoogle Scholar
  54. Takeshita RSC, Bercovitch FB, Huffman MA, Mouri K, Garcia C et al (2014) Environmental, biological, and social factors influencing fecal adrenal steroid concentrations in female Japanese macaques (Macaca fuscata). Am J Primatol 76:1084–1093.  https://doi.org/10.1002/ajp.22295 CrossRefPubMedGoogle Scholar
  55. Takeshita RS, Huffman MA, Kinoshita K, Bercovitch FB (2017) Effect of castration on social behavior and hormones in male Japanese macaques (Macaca fuscata). Physiol Behav 181:43–50CrossRefPubMedGoogle Scholar
  56. Terrien J, Zizzari P, Bluet-Pajot M-T, Henry P-Y, Perret M et al (2008) Effects of age on thermoregulatory responses during cold exposure in a nonhuman primate, Microcebus murinus. Am J Physiol Regul Integr Comp Physiol 295:R696–R703CrossRefPubMedGoogle Scholar
  57. Toda M, Makino H, Kobayashi H, Morimoto K (2006) Health effects of a long-term stay in a spa resort. Arch Environ Occup Health 61:131–137.  https://doi.org/10.3200/Aeoh.61.3.131-137 CrossRefPubMedGoogle Scholar
  58. Tokura H, Sugiyama K (1975) Sweating in the Japanese macaque (Macaca fuscata). Primates 16:95–98CrossRefGoogle Scholar
  59. Tsuji Y (2010) Regional, temporal, and interindividual variation in the feeding ecology of Japanese macaques. In: Nakagawa N, Nakamichi M, Sugiura H (eds) The Japanese macaques. Primatology Monographs, Springer, pp 99–127CrossRefGoogle Scholar
  60. Van Meter PE, French JA, Dloniak SM, Watts HE, Kolowski JM et al (2009) Fecal glucocorticoids reflect socio-ecological and anthropogenic stressors in the lives of wild spotted hyenas. Horm Behav 55:329–337.  https://doi.org/10.1016/j.yhbeh.2008.11.001 CrossRefPubMedGoogle Scholar
  61. Weingrill T, Gray DA, Barrett L, Henzi SP (2004) Fecal cortisol levels in free-ranging female chacma baboons: relationship to dominance, reproductive state and environmental factors. Horm Behav 45:259–269.  https://doi.org/10.1016/j.yhbeh.2003.12.004 CrossRefPubMedGoogle Scholar
  62. Yamagiwa J (2010) Research history of Japanese macaques in Japan. In: Nakagawa N, Nakamichi M, Sugiura H (eds) The Japanese macaques. Springer, Kyoto, pp 3–25CrossRefGoogle Scholar
  63. Zhang P, Watanabe K (2007) Extra-large cluster formation by Japanese macaques (Macaca fuscata) on Shodoshima Island, Central Japan, and related factors. Am J Primatol 69:1119–1130.  https://doi.org/10.1002/ajp.20419 CrossRefPubMedGoogle Scholar
  64. Zhang P, Watanabe K, Eishi T (2007) Habitual hot-spring bathing by a group of Japanese macaques (Macaca fuscata) in their natural habitat. Am J Primatol 69:1425–1430.  https://doi.org/10.1002/ajp.20454 CrossRefPubMedGoogle Scholar

Copyright information

© Japan Monkey Centre and Springer Japan KK, part of Springer Nature 2018

Authors and Affiliations

  • Rafaela S. C. Takeshita
    • 1
  • Fred B. Bercovitch
    • 2
  • Kodzue Kinoshita
    • 2
  • Michael A. Huffman
    • 1
  1. 1.Department of Ecology and Social Behavior, Primate Research InstituteKyoto UniversityInuyamaJapan
  2. 2.Wildlife Research CenterKyoto UniversityKyotoJapan

Personalised recommendations