Behavioral Ecology and Sociobiology

, Volume 67, Issue 1, pp 101–111

Sneak copulations in long-tailed macaques (Macaca fascicularis): no evidence for tactical deception

Authors

    • Ethology Research, Animal Science DepartmentBiomedical Primate Research Centre
    • Behavioral BiologyUtrecht University
  • C. U. Olesen
    • Ethology Research, Animal Science DepartmentBiomedical Primate Research Centre
    • Behavioral BiologyUtrecht University
  • H. de Vries
    • Behavioral BiologyUtrecht University
  • B. M. Spruijt
    • Behavioral BiologyUtrecht University
  • E. H. M. Sterck
    • Ethology Research, Animal Science DepartmentBiomedical Primate Research Centre
    • Behavioral BiologyUtrecht University
Original Paper

DOI: 10.1007/s00265-012-1430-4

Cite this article as:
Overduin-de Vries, A.M., Olesen, C.U., de Vries, H. et al. Behav Ecol Sociobiol (2013) 67: 101. doi:10.1007/s00265-012-1430-4

Abstract

Sexual competition is highly prevalent within multi-male multi-female primate groups and may lead to copulations in absence of potentially interfering bystanders. Such avoidance of bystanders may result from tactical deception or from simpler mechanisms such as taking advantage of encountered situations without bystanders, operant conditioning or a peripheral positioning of non-alpha males. We investigated which individuals are avoided as bystanders, how individuals react to the presence of bystanders and whether copulation partners separate themselves from the group in a tactical way. Our observations of a group of 15 female and seven male long-tailed macaques housed in three interconnected, but visually separated compartments revealed that both males and females can interrupt sexual behaviour and that bystanders of both sexes were avoided during copulations (n = 256). The strength of the effect of bystanders tended to decrease with the dominance rank of male bystanders, but did not depend on the dominance rank of female bystanders. The audience effects of non-alpha individuals did not depend on the strong audience effect of the alpha male in combination with proximity with the alpha male. The effects that we found for separate bystanders suggest that sexual competition concerns rank dependent male–male competition and rank independent female–female competition. Additionally, both male and female copulation partners paid attention to the presence of bystanders and conducted fewer copulation solicitations in their presence. The timing of a male and female’s separation from the group suggests that exploitation of the peripheral position of non-alpha males, and not tactical deception, may cause these audience effects.

Keywords

Tactical deceptionAudience effectMating tacticsSexual competitionPrimatesMonopolizationMacaques

Introduction

Many animal species live in multi-male multi-female groups. Within these groups, high-ranking males typically have priority of access to fertile females (Altmann 1962), which is reflected by a correlation between mating success and male rank in many animal groups (for review: Cowlishaw and Dunbar 1991; for different species, e.g. Macaca species: de Ruiter and van Hooff 1993: Paul et al. 1993; Garcia et al. 2009; Massen et al. 2012; Marmota flaviventris: Huang et al. 2011; Capra ibex: Willisch et al. 2012), but not all (e.g. Macaca fascicularis: de Ruiter et al. 1992; Porphyrio porphyrio melanotus: Lambert et al. 1994; Crocuta crocuta: Engh et al. 2002). Subordinate males and females do not necessarily benefit from increased mating success of dominant males and use several alternative reproductive tactics (Berard et al. 1994; Soltis et al. 2001; Alfaro 2005) to increase the reproductive success of subordinate males and enhance female promiscuity. One of these tactics resulting from sexual competition is sneak copulation, involving sexual behaviour in the absence of competitors. Sneak copulations have been reported in multiple animal species, including birds (Davies 2000), seals (de Bruyn et al. 2011) and ungulates (Hogg 1987; Willisch and Neuhaus 2009). Studies in primates that report sneak copulations range from observations at a single field site (Kummer 1968; Soltis et al. 2001) to a large compilation of these anecdotal observations (Byrne and Whiten 1990) and experimental data (Ruiz de Elvira and Herndon 1986; Gygax 1995; Kummer et al. 1996; Overduin-de Vries et al. 2012).

Non-human primates are our closest animal relatives and are proposed to share some complex cognitive capacities with humans, including tactical deception (Byrne and Whiten 1992). Sneak copulations in primates may be an example of tactical deception (Byrne and Whiten 1992), yet it is not known which cognitive mechanisms actually lead to the hiding of sexual behaviour. The reported hiding behaviours (e.g. Ruiz de Elvira and Herndon 1986; Byrne and Whiten 1990; Gygax 1995; Overduin-de Vries et al. 2012) may have resulted from four cognitively distinct mechanisms. Tactical deception is the cognitively most complex mechanism and comprises active avoidance of a particular competitor (Byrne and Whiten 1992). Second, sneak copulations may result from passively taking advantage of an encountered situation where the competitor is absent (Byrne and Whiten 1992). Third, operant conditioning, an even less advanced cognitive process, may result in inhibition of sexual behaviour near former punishing individuals. Last, subordinate males often reside in the periphery of the group (Macaca fuscata: Hayakawa 2007; Cebus apella: Janson 1990), and when females approach these males, copulations automatically take place at locations distant from the often centrally positioned alpha male. This mechanism is simpler than the former strategy as it does not require recognition of the absence of potentially punishing individuals. Above all, in order to label sneak copulation an example of tactical deception, evidence is needed that “the hiding is tactical and not simply an opportunistic response to being left alone by chance” (Byrne and Whiten 1992, p. 613) or the consequence of simple conditioning or general positioning of non-alpha males.

To investigate the cognitive mechanisms underlying sneak copulations, it is necessary to determine which individuals are avoided as bystanders during copulations and on which grounds. Regardless of the cognitive mechanism underlying sneak copulation, an important benefit of hiding sexual behaviour is avoidance of harassment, raising the focus on harassers. Harassment of copulations is a tactic that is often used by high-ranking males to reduce mating success of their group members for the benefit of their own mating success (Dixson 1998, p. 72, Table 4.5). Whether male harassment results in interference of the copulation before an ejaculation can occur depends on the relative ranks of the males involved (Niemeyer and Chamove 1983) and is more often used by high-ranking than by low-ranking males (Chapais 1983). Moreover, although less frequently reported than harassment by males, females may also harass copulations of group members (Niemeyer and Anderson 1983). Yet, harassment by females only rarely results in actual termination of copulation, and ejaculation generally still occurs (Niemeyer and Anderson 1983). The function of female harassment is more ambiguous than male harassment and less investigated. Harassment by females may function to test male mating quality (Niemeyer and Chamove 1983), to limit future competition over resources by reducing the number of births in the group (Niemeyer and Anderson 1983) and to reduce competition for access to benefits given by male copulation partners, such as agonistic support (Buchan et al. 2003). Female harassment may be rank dependent since the alpha female is often the most frequent harasser (Macaca arctoides: Gouzoules 1974; Macaca mulatta: Wilson 1981). To sum up, based on the reported observations of interfering behaviours, it is expected that high-ranking males and females are avoided as bystanders during sexual behaviour.

Although some empirical studies confirm the avoidance of potential harassers (Ruiz de Elvira and Herndon 1986; Soltis et al. 2001), the effect of specific individuals is usually not determined since most studies examined the combined effect of multiple individuals on hiding copulations and did not discriminate which individuals did and which did not influence the occurrence of copulations. Although in a previous study we found that in rhesus macaques the presence of the alpha male and some other high-ranking males individually inhibited copulations of their group members (Overduin-de Vries et al. 2012), it is not clear whether females have the same effect. Moreover, while we found that females reduce their sexual behaviour in the presence of a potential harasser (Overduin-de Vries et al. 2012), it is not clear whether males do the same.

Another central question when studying the cognitive mechanisms underlying sneak copulation is whether the male, the female or both partners involved in the copulation are paying attention to bystanders. Because of the costs associated with interferences, both males and females may be motivated to hide their copulations from potentially interfering bystanders. Disruption of copulations may have consequences for the reproductive success of lower ranking males. In M. arctoides, interruption of copulations results in a significant shortening of the pair-sit behaviour (Bruce and Estep 1992), which impedes positioning of the copulatory plug, and increases susceptibility to sperm competition (Dixson and Anderson 2002). Additionally, males may experience injury or energy loss from female harassment since female harassment is mostly directed at the male (M. mulatta: Wilson 1981). Likewise, females risk injury from interferences since they are often the target of aggression after interferences by males (e.g. M. mulatta: Ruiz de Elvira and Herndon 1986; Manson 1996). Thus, both male and female copulation partners are expected to pay attention to the presence of these bystanders and to act accordingly.

This study was designed to investigate (1) which individuals in a social long-tailed macaque group inhibit sexual behaviour of their group members and whether this inhibition is linked with the bystander’s interference behaviour, sex, rank or proximity to the alpha male; (2) whether both males and females separately adjust their rate of copulation solicitations to the presence of potentially harassing bystanders; (3) whether sneak copulations result from peripheral positions of subordinate males; and (4) whether the individuals involved in a sneak copulation separate themselves from the rest of the group in a tactical way: i.e. whether the first individual’s separation from the group is shortly followed by the second individual.

Methods

Subjects

Observations were conducted on one group of 27 socially housed long-tailed macaques (the ‘lixa group’) at the Biomedical Primate Research Centre in Rijswijk, the Netherlands, formerly housed at the ‘Ethology station’, Utrecht University. Long-tailed macaques of this colony are non-seasonal breeders, consistent with other captive long-tailed macaque colonies (Honjo et al. 1978), but different from most wild populations (Kavanagh and Laursen 1984). Long-tailed macaques often live in multi-male multi-female groups of 6 to 58 individuals (van Schaik and van Noordwijk 1985) where alpha males try to monopolize fertile females and father the majority of infants (de Ruiter and van Hooff 1993; Engelhardt et al. 2004), but females are promiscuous and actively approach non-alpha males to copulate (Nikitopoulos et al. 2005). Females copulate with more than one male during their fertile phase (Engelhardt et al. 2004) and copulate outside ovarian cycles in order to confuse paternity (Engelhardt et al. 2007).

Fifteen (>3.5 years) females and seven (>6.5 years) males were included in this study (Table 1). The remaining five monkeys in the group were under the age of 1 year and therefore sexually immature. In total, eight females were pregnant or lactating during our observations (Table 1). Eight of the females had been given contraceptives (implanon implants; generic name: Etonogestrel). Since four of the males were non-natal and the group consisted of three unrelated matrilines, all subjects had the option to show sexual behaviour with multiple unrelated sexual partners [mean ± SD = 6.5 ± 0.5 (for female subjects) and 13.9 ± 1.6 (for male subjects)].
Table 1

Study subjects

 

Rank

Age (years)

Reproductive status

Hormonal implant

Females

    

Voila

1

11.5

Pregnant

No

Silva

2

7.3

Lactating

No

Fossa

3

3.5

Pregnant

No

Sepia

4

12.0

 

Yes

Riva

5

8.0

 

No

Virginia

6

15.9

Lactating

Yes

Tremaa

7

11.7

Lactating

Yes

Foetsie-ba

8

11.0

 

Yes

Mokka

9

8.2

Lactating

No

Baklava

10

10.1

Lactating

Yes

Goa

11

15.1

 

Yes

Taiga

12

10.1

 

Yes

Yukka

13

15.9

 

No

Kaa

14

24.1

 

Yes

Karamba

15

12.3

Pregnant

No

Males

Natal?

Nacho

1

13.5

No

 

Bolero

2

13.6

No

 

Potato

3

26.6

Yes

 

Pesto

4

6.8

Yes

 

Rokoko

5

6.5

Yes

 

Voodoo

6

7.11

No

 

Bamboo

7

15.1

No

 

Information of subject’s name, rank and age (in years). Additionally, for females, their reproductive status and presence of hormonal implant are provided, while for males, their origin (natal or non-natal) is provided

Housing

Three visually separated compartments were used during the observations. Two inside enclosures measured together 72 m2, 3 m high, while an outside enclosure measured 209 m2, 3 m high. Some adjustments (cf. Overduin-de Vries et al. 2012) ensured the visual separation of the inside and the outside enclosure (Fig. 1). Two inside compartments were visually separated from each other by a concrete wall, the third inside compartment was locked and not available to the monkeys during the observations. Monkeys were able to walk freely between these compartments through tunnels and openings in the walls.
https://static-content.springer.com/image/art%3A10.1007%2Fs00265-012-1430-4/MediaObjects/265_2012_1430_Fig1_HTML.gif
Fig. 1

Schematic layout of the cage. Enclosures were surrounded by a concrete wall, indicated by bold lines, and wire mesh, indicated by thin lines. Openings in the wall are indicated with an asterisk; these permitted monkeys to walk through, but not see through as they were covered by opaque plastic flaps. This way, three visually separated compartments were created indicated by different shadings. Blank areas were not used by the monkeys during observation slots

This study was performed making use of the normal housing conditions without hindering the animals’ normal behaviour. The monkeys were fed monkey chow, complemented by fresh fruit, vegetables or bread on a daily basis. Water was provided ad libitum. The inside enclosure had a sawdust bedding and environmental enrichment, consisting of fire hoses, tires, ladders and an outside swimming pool was permanently available. Extra enrichment containing food was provided every other week (Vernes and Louwerse 2010).

Observations

Observations were performed from 9 November 2009 until 19 March 2010 in two-hour slots (total observation time = 106 h). Observations were balanced between morning and afternoon sessions. Two observers simultaneously recorded the presence of the focal animals in the two inside and one outside enclosure using The Observer 5.0 (Noldus 1991). The two observers communicated between inside and outside enclosures through walkie-talkies to validate the identity of individuals moving between compartments. A bystander was designated ‘present’ if the bystander was in the same enclosure as the dyad involved in a sexual interaction. All occurrences of sexual and dominance interactions were recorded with the main focus on sexual behaviour (cf. Overduin-de Vries et al. 2012). The recorded sexual behaviours included: copulation solicitations: sexual presentation (females), lift female and grasp waist (males); copulations: mount with thrusting; and other sexual behaviours: inspect, touching genitals (Angst 1974). In addition, interferences in sexual interactions were recorded. Interference was defined as an aggressive behaviour towards one or both partners involved in a sexual interaction, but it did not necessarily lead to interruption of the sexual interaction. A female was considered proceptive on a given day if she sexually presented at least once. The dominance hierarchy of the group was determined on the basis of submissive behaviours (teeth baring and unprovoked avoidance). From the resulting matrix, the dominance hierarchy including all adult individuals was obtained (de Vries 1998) using Matman 1.1 (de Vries et al. 1993), and the dominance hierarchy was significantly linear (h′ = 0.55, p < 0.0001) (de Vries 1995). Subsequently, the males and females were ranked within each sex with the most dominant animal assigned rank one.

Data analysis

We determined the inhibiting effect of bystanders on copulations and on copulation solicitations. Each adult individual (male and female) was treated as bystander. First, the effect of each bystander’s presence on copulation rate was determined. Second, the effect of each bystander on the rates of female or male copulation solicitations was determined.

We calculated the rate of sexual behaviour (cf. Overduin-de Vries et al. 2012), i.e. for copulations or for solicitations, in the presence and absence of a particular bystander using the Observer XT 7.0 (Noldus 1991). Since the monkeys were able to move around freely, their time spent with a particular bystander differed from their time without this particular bystander. A simple comparison between the number of sexual behaviours with and without a high-ranking bystander male can be biased by the fact that subordinate males avoid the presence of high-ranking males also during non-sexual behaviour. Therefore, we calculated the rate of sexual behaviours per hour for each female–male dyad across the summed time in which they were together in the same enclosure with the bystander present S(pr), and similarly across the summed time that they were together in the same enclosure, but in absence of the bystander S(ab). In order to correct for the variation between dyads, for each dyad these rates were centred around zero by subtracting for each dyad the mean of S(ab) and S(pr) from S(ab) and S(pr).

Finally, we analysed whether males and females coordinate their movements preceding a sexual event to avoid the alpha male, the individual that is expected to have the strongest audience effect. For each dyad that was observed to engage in sexual behaviour in the outside compartment, the intervals (in seconds) between the male and female going outside, from now on called ‘latencies’, were compared between episodes with and without a sexual event following their movement from the inside to the outside compartment when the alpha male was inside, further termed ‘movement outside’. Only movements outside were taken into account because we had the impression that the majority of the group was most of the time inside and that individuals went outside to separate themselves from the group. Latencies were calculated in two conditions: with and without sexual behaviour in the outside compartment by the male–female dyad. For each dyad, these latencies were centred around zero by subtracting the mean latency of dyad d from dyad d’s latencies in both conditions.

Permutation tests were conducted (1) to investigate whether sexual behaviour was more often observed in the absence than in the presence of a bystander and (2) to investigate whether latencies were shorter before a sexual event than before a non-sexual event. To control for variation between dyads, permutations were conducted within dyads; that is, for each dyad the observations were put in a random order, thus, assigning each observation randomly to one of the two conditions. Permutation tests included 10,000 sets of random permutations, under the null hypothesis that all observed copulation/solicitation rates or latencies were randomly distributed over both conditions; with or without audience (for sexual behaviour rates) or with or without sexual behaviour (for latencies). This created a null distribution of the mean difference between two conditions (from now on called “mean randomly obtained difference”, MRD), The proportion of times that the MRD was greater than or equal to the absolute mean observed difference (MOD) yields the right-tailed p value Pr. This value was multiplied by two to obtain the two-tailed p value. The MOD, representing the observed difference between copulation rates or solicitation rates in and out of view of the bystander, is an indicator of the strength of the audience effect of a particular bystander.

In order to disentangle the effects of certain characteristics of bystander individuals on the strength of the audience effect, we ran linear models, with as dependent factor the MOD and as independent factors the bystander’s dominance rank, proximity to the alpha male and proceptivity. Proximity to the alpha male was calculated by summing the number of hours spent in the same compartment as the alpha male. To diagnose the presence of multicollinearity, we calculated the variance inflation factor for each of the predictors. These VIFs varied between 1.17 and 1.35, which are well below 5, and thus do not indicate serious multicollinearity. Inspection of the residuals showed that these were homogeneously distributed across the different predictor values, and they did not deviate significantly from normality (Kolmogorov–Smirnov test, p between 0.58 and 0.99)

All statistical tests were two-tailed with α set at 0.05. All tests were done using the software package “R” (R Development Core Team 2009) version 2.10. The “lm” function for linear models was used from the “stats” package, while the “sample” function for permutations was used from the “base” package. The sharper Bonferroni procedure for multiple testing (Hochberg 1988) was applied to test the audience effects of multiple males and females, with the family-wise error rate set at 0.05. Additionally, for single tests, we report trends with p between 0.05 and 0.10.

Results

Copulation, copulation solicitation and interference behaviour

Out of the 256 observed copulations, 126 involved non-alpha males. Male copulation solicitations (N = 488) involved non-alpha males 367 times. Female solicitations (N = 222) were directed towards non-alpha males 92 times. In general, males and females that copulated more than once during the mating season had more than one copulation partner (Fig. 2). There were two females that only copulated once. The mean number of proceptive females per day was 2.57 ± SD 1.54 (range, 1–6). There were 22 observed instances of a female copulating with more than one male on the same day. These instances included seven different females, and 15 of these instances included copulations with the alpha male. On a day a female was proceptive, she occasionally behaved proceptively toward more than one male, by sexually presenting toward, on average, 1.13 ± SD 0.36 (range, 1–3) males.
https://static-content.springer.com/image/art%3A10.1007%2Fs00265-012-1430-4/MediaObjects/265_2012_1430_Fig2_HTML.gif
Fig. 2

Distribution of copulations in the group, males on the bottom line and females in the arch are ordered according to dominance rank from high to low ranking from left to right. The width of the lines between individuals indicates the number of copulations observed for each dyad

The number of copulations was significantly correlated with male dominance rank (Spearman rank correlation, ρ = −0.857, N = 7, p = 0.024). Female dominance rank was not related with the number of copulations (Spearman rank correlation, ρ = 0.102, N = 15, p = 0.72). There was no significant difference in the number of copulations between pregnant or lactating females and other females (Mann–Whitney test, U = 18.5, N = 7, 8, p = 0.29). Likewise, there was no significant difference in the number of days a female was proceptive between the group of lactating and pregnant females and other females (Mann–Whitney test, U = 35, N = 7, 8, p = 0.45). Females with hormonal implants copulated significantly less often (Mann–Whitney test, U = 45.5, N = 7, 8, p = 0.049) and were proceptive on fewer observation days (Mann–Whitney test, U = 49.5, N = 7, 8, p = 0.014) than females without implants. Because in subsequent pairwise comparisons investigating bystander effects females were their own control, existing differences in sexual behaviours between the different reproductive stages are automatically controlled for.

In total, eight sexual interactions (N > 710) of which four involved copulation (N = 256) were interfered. The alpha male showed four cases of aggressive interference in sexual interactions, while eight out of ten copulations in his view were not interfered. Non-alpha males, namely both the second ranking and third ranking male, also interfered sexual interactions of group members. Also, females interfered in sexual interactions; the 9th and 15th ranking female both once interfered in a sexual interaction. None of the other individuals was observed to interfere sexual interactions of group members.

The alpha male directed his aggression during interferences towards either the male (two times) or both partners (two times) involved in the sexual interaction. The two interferences by non-alpha males concerned aggression to both copulating partners, whilst interferences by females both times were directed towards the female. Aggression involved in interferences was almost exclusively directed at an individual that was lower in rank than the aggressor, with one exception: the 15th ranking female once directed her aggression toward the 9th ranking female that was copulating with the 2nd ranking male. On four occasions, the harasser (one female, three times the alpha male) copulated with the interfered opposite-sexed individual afterwards.

Bystander effects on copulations

The presence of a particular bystander male may affect sexual behaviour of group members. The key result is that dyads of females and non-alpha males copulated significantly less often in the presence than in the absence of the alpha male (Table 2). The majority of the copulations (116/126) with non-alpha males occurred in absence of the alpha male. Each non-alpha male copulated on average 19.33 ± SD 28.8 times in absence and 1.67 ± SD 3.2 times in presence of the alpha male.
Table 2

Statistical results of the permutation tests for the audience effects of all possible bystanders (column 1) on female solicitations (columns 2–3) and male solicitations (columns 4–5)

Bystanders

Rank

Female solicitations

Male solicitations

 

MOD

p

MOD

p

Males

     

Nacho

1

0.068

0.0008*

0.149

0.0002*

Bolero

2

0.048

0.0744

0.113

0.0002*

Potato

3

0.013

0.7598

−0.015

0.3752

Pesto

4

0.025

0.3342

−0.004

0.8360

Rokoko

5

0.029

0.3310

−0.039

0.0304

Voodoo

6

0.029

0.2498

0.082

0.0088

Bamboo

7

0.022

0.1646

0.062

0.0016*

Females

     

Voila

1

0.052

0.0176

0.104

0.0002*

Silva

2

0.061

0.0022*

0.102

0.0002*

Fossa

3

0.077

0.0002*

0.097

0.0002*

Sepia

4

0.070

0.0006*

0.071

0.0002*

Riva

5

0.040

0.2392

0.106

0.0004*

Virginia

6

0.083

0.0004*

0.113

0.0014*

Tremaa

7

0.057

0.0062

0.095

0.0006*

Foetsie-ba

8

0.068

0.0008*

0.053

0.0196

Mokka

9

0.028

0.2316

0.070

0.0018*

Baklava

10

0.075

0.0006*

0.102

0.0004*

Goa

11

0.231

0.0002*

0.090

0.0004*

Taiga

12

0.062

0.0004*

0.113

0.0002*

Yukka

13

0.080

0.0054

0.082

0.0092

Kaa

14

0.097

0.0004*

0.084

0.0002*

Karamba

15

0.044

0.0282

0.025

0.2342

The audience effect strength is measured with MOD = mean observed difference in copulation (or solicitation) rate per hour with or without bystander

*p = 0.05, MOD is significant at this family-wise error rate after applying the sharper Bonferroni procedure for multiple testing (Hochberg 1988)

Although the alpha male had the strongest audience effect (with the highest MOD), he was not the only male that inhibited copulations since the presence of the second, fourth, fifth and sixth ranking male also significantly inhibited copulations of group members (Table 2). The MOD of copulation rates of non-alpha males was neither dependent on rank nor on the bystander’s proximity to the alpha male (Table 3). However, in this model, the alpha male was excluded because his own proximity to himself is not a meaningful measurement. When including the alpha male to the model and eliminating the non-significant factor “proximity to the alpha male” we increase our limited sample size and improve our model. In this model, rank does significantly affect the audience effect strength (Table 4, Fig. 3a)
Table 3

Results from the linear model with as dependent variable the audience effect strength of bystander males (MOD) and as independent factors the dominance rank of the bystander male and its proximity to the alpha male

 

Coefficient

SE

T

p

Intercept

0.08

0.07

1.27

0.29

Dominance rank

−0.008

0.005

−1.61

0.21

Proximity to alpha male

0.0003

0.002

0.18

0.87

Table 4

Results from the linear model with as dependent variable the audience effect strength of bystander males (MOD) and as independent factor the dominance rank of the bystander male

 

Coefficient

SE

T

p

Intercept

0.11

0.01

7.67

<0.001

Dominance rank

−0.01

0.003

−3.22

0.023

https://static-content.springer.com/image/art%3A10.1007%2Fs00265-012-1430-4/MediaObjects/265_2012_1430_Fig3_HTML.gif
Fig. 3

The audience effect strength (MOD) in copulation rates between absence and presence of a a particular bystander male or b a particular bystander female. Bystanders are represented on the x-axis by their dominance rank number from 1 (highest ranking) to 7 or 15 (lowest ranking). Filled dots represent data points from bystanders with significant audience effects, open dots represent bystanders without significant audience effects

Out of the 15 females in the group, 9 had a significant audience effect on the copulations of their group members (Table 2). The MOD was not dependent on the rank of the bystander female (Fig. 3b), or with proceptivity of bystander females or with the bystander’s proximity to the alpha male (Table 5).
Table 5

Results from the linear model with as dependent variable the audience effect strength of bystander females (MOD) and as independent factors proceptivity and dominance rank of the bystander female and its proximity to the alpha male

 

Coefficient

SE

T

p

Intercept

0.11

0.19

0.57

0.58

Dominance rank

2.8 × 10−5

1.9 × 10−3

0.02

0.99

Proceptivity

−1.6 × 10−3

1.1 × 10−3

−1.48

0.17

Proximity to alpha male

−4.4 × 10−4

3.8 × 10−3

−0.12

0.91

Bystander effects on copulation solicitations

Non-alpha males invited copulations significantly more often in the absence than in the presence of the alpha male (Table 2). Additionally, the second and seventh ranking male significantly inhibited male copulation solicitations within their view (Table 2). Moreover, all but three (rank 8, 13 and 15) females had a significant audience effect on male copulation solicitations (Table 2).

Females sexually presented significantly more often to non-alpha males when the alpha male was absent than when he was present (Table 2). None of the other males significantly influenced female solicitation behaviour (Table 2). Additionally, 9 out of 15 females significantly inhibited female copulation solicitations of other females (Table 2).

Peripheral positioning

Subjects spent more time in the two inside (90.9 ± SD 11.5 h) than outside compartments (17.6 ± SD 11.5 h; Wilcoxon signed rank test, V = 0, N = 22, p < 0.001). The monkeys spent an equal amount of time in either of the two inside compartments (Wilcoxon signed rank test, V = 170, N = 22, p = 0.17). Males tended to spend more time outside (26.5 ± SD 16.3 h) than females (12.8 ± SD 4.9 h) (Mann–Whitney test, U = 26, N = 7, 15, p = 0.066). Males spend fewer hours in proximity to the alpha male (mean = 39.7 ± SD 5.2) than females (mean = 49.1 ± SD 2.3; Mann–Whitney test, U = 1, N = 6, 15, p < 0.001). Whereas the alpha male had a higher copulation rate inside (1.6 copulations/hour) than outside (0.48 copulations/hour), non-alpha males tended to have higher copulation rates outside (0.39 ± SD 0.67 copulations/hour) than inside (0.15 ± SD 0.25 copulations/hour; Wilcoxon signed rank test, V = 19, N = 6, p = 0.094).

Separation from the group

In episodes with a sexual interaction outside, the latency (the intervals between the male and female of a dyad going outside) was on average 379 ± SD 550 s, compared to 559 ± SD 804 s when no sexual interaction occurred outside (Fig. 4). Although the average latency in episodes with sexual behaviour was shorter than in episodes without sexual behaviour, the difference was not significant (permutation test, p = 0.15). In episodes without sexual behaviour, males were significantly more often the first of the dyad to appear outside (exact binomial test, male first = 1,130/1,667, p < 0.0001). In the episodes where a sexual behaviour occurred outside, males tended to be the first of the dyad to be outside (exact binomial test, male first = 58/97, p = 0.07). There was no significant difference in the proportion of times that the male was outside first between episodes with and without sexual behaviour (Fisher’s exact test, N = 1,761, p = 0.12).
https://static-content.springer.com/image/art%3A10.1007%2Fs00265-012-1430-4/MediaObjects/265_2012_1430_Fig4_HTML.gif
Fig. 4

The latency between the movement of a male and female partner to the outside compartment for events with (black lines filled circles) and without (grey lines open circles) sexual interactions. Dyads on the x-axis are ranked according to the latency during events without sex. Error bars indicate 95 % confidence intervals

Discussion

In order to label sneak copulations an example of tactical deception, it has to be shown that hiding is tactical (Byrne and Whiten 1990) and, furthermore, to investigate cognitive mechanism underlying sneak copulations, it is necessary to determine which individuals are avoided as bystanders and on what basis. We investigated whether the presence of group members affected sexual behaviour of group living long-tailed macaques, how certain characteristics of bystanders relate to their inhibiting effect and how individuals reacted to the presence of particular group members. Moreover, we investigated whether movements to the outside compartment are synchronized between male and female copulation partners to sneak away from the remainder of the group.

Our study confirms that male long-tailed macaques, by their mere presence, significantly inhibit copulations of group members in their view. The alpha male has a strong audience effect and reduces the number of copulations within his view. Additionally, four other males inhibited the occurrence of copulations within their view. Their audience effects were not a side effect of the alpha male’s audience effect mediated by proximity of bystander males to the alpha male, indicating that it is indeed the bystander male itself that has an audience effect. The motivation to avoid males as bystanders during sexual behaviour may be harassment. Although the observed rate of interference was low, it may not be representative of the actual risk of interference (cf. difference between risk and rate of predation [Hill and Dunbar 1998) or infanticide (Janson and van Schaik 2000)] since effective avoidance of potential harassers will decrease the number of interferences in a group. Indeed, we found ample evidence that dyads avoid bystanders during copulations. This suggests that the potential risk of interference by a particular bystander causes an audience effect of this individual.

The alpha male in our study was the most frequent harasser, but non-alpha males also interfered in copulations. This is consistent with other primate species where interferences are more frequently observed by higher than by lower ranking males (M. arctoides: Niemeyer and Chamove 1983; M. fuscata: Stephenson 1975; M. mulatta: Wilson 1981). In line with this, the audience effect strength tended to be higher in high- than in low-ranking males. Yet this effect of male rank disappeared when the alpha male was left out of the analysis, rendering it unclear whether the effect of dominance rank is important for audience effects between non-alpha males. However, the difference in audience effect strength between alpha and non-alpha males is evident. The absence or lower audience effect strength of lower ranking males may result from their reduced influence on particular dyads. Possibly, copulating dyads involving high-ranking individuals did not pay attention to the presence of males that were relatively low in rank. This can be due to the direction of interference, which is commonly directed from a higher ranking towards a lower ranking individual (this study; Chapais 1983). In summary, our data provide evidence for an avoidance of several bystander males as potential harassers during copulations. This is consistent with rhesus macaques that also show audience effects of alpha and non-alpha males on sexual behaviour of group members (Overduin-de Vries et al. 2012). The variability of audience effect strengths among males of different dominance ranks suggests that male audience effects result from dominance-based male sexual strategies such as interference.

Females also significantly inhibited copulations of their group members. Similar to the effect of non-alpha males, the audience effect of females is not a by-product of the alpha male’s audience effect in combination with female proximity to the alpha male since females with different proximities to the alpha male did not differ in their audience effect strengths. Females also aggressively disturb copulations (Niemeyer and Chamove 1983; this study), and hiding copulations from females may also function to prevent interferences. The strength of a female audience effect was not related to the dominance rank of the female. This is in line with the dominance rank of the females that interfered copulations: one was middle and the other low ranking. Our results contrast with the hypothesis that mainly dominant females are avoided because of their potential to interfere copulations (M. arctoides: Gouzoules 1974; M. mulatta: Wilson 1981). Alternatively, the hypothesis that the function of interferences by females is to test male quality (Niemeyer and Chamove 1983) predicts that a female’s potential to interfere copulations may depend on a female’s willingness to copulate. However, there was no significant relation between female proceptivity and audience effect strength, refuting this hypothesis. Altogether, many females were avoided during copulations, and avoidance of a particular female was not based on her dominance rank, her proceptivity or proximity to the alpha male. Since no particular female characteristic was found to determine a female’s audience effect strength and many females had an effect, it is possible that all females are potential harassers and benefit from interfering copulations, supporting the hypotheses that females prevent competition for future resources (Niemeyer and Anderson 1983) or for non-procreative benefits given by male copulation partners, such as agonistic support (Buchan et al. 2003), or for parental investment (Soltis 2002). If this is true, it indicates that female audience effects result from general female–female competition.

Analysis of the female initiatives to copulate revealed that they react to the presence of the alpha male and most other females, yet do not react to the presence of subordinate males. Also, male copulation solicitations were inhibited by the alpha male and all but three females, and in addition by two lower ranking males. Altogether, the audience effects on both male and female solicitations show that both sexes pay attention to the presence of bystanders during copulation solicitations and adjust their behaviour accordingly. This suggests a benefit for both partners to hide their sexual behaviour, which is in line with the proposed costs of being interfered for both males (Wilson 1981; Bruce and Estep 1992) and females (Ruiz de Elvira and Herndon 1986; Manson 1996), such as the costs of receiving aggression and the reduced reproductive success due to interfered copulations. Alternatively, only one individual of the copulating dyad obtains direct benefits, while the other obtains an indirect benefit when its partner is more willing to copulate in occluded conditions.

Because both partners adjust their sexual behaviour in the presence of bystanders, there is a possibility that hiding copulations is a coordinated behaviour between both partners involved in the copulation. Namely, if it is a coordinated movement, sneaky mating may be the result of tactical deception, i.e. both partners are “tactically getting left behind” (Byrne and Whiten 1992, p. 613). We determined whether there is evidence for a coordinated movement to the outside compartment to avoid the alpha male, which was most of the time inside. Although the latency in episodes with sexual behaviour was shorter than in episodes without sexual behaviour, this difference was not significant. Therefore, we have no evidence for tactical deception. However, we cannot firmly exclude tactical deception either since the data were in the direction expected under the tactical deception hypothesis, and additional data collection might reveal a significant effect.

Alternatively, bystander effects may result from conditional learning. Because of the low incidence of interferences in our study, it was impossible to statistically link bystander effects with interference behaviour by a particular bystander. Despite the low incidence of interferences, it is possible that a few incidences of punishment in the past have resulted in a conditioned inhibition of sexual behaviour in the presence of certain individuals. Aggression-accompanied interference has been frequently reported (Gouzoules 1974; Wilson 1981; Chapais 1983; Niemeyer and Chamove 1983; Ruiz de Elvira and Herndon 1986; Manson 1996; Dixson 1998) and thus may be the underlying mechanism to bystander effects.

Moreover, sneak copulations may result from the general peripheral positioning of non-alpha males. Indeed, in our study males occupied peripheral positions during non-sexual events since they tended to reside more often in the (less occupied) outside compartment than females, spend less time in proximity of the alpha male and were more often the first of a dyad moving outside. Moreover, before a sexual event males tended to be the first outside as well, and non-alpha males, in contrast with the alpha male, copulated more often outside than inside. This suggests that sneak copulations may simply result from the peripheral positioning of non-alpha males. Yet the rate of sexual interactions was significantly higher in the absence than presence of bystander males, in particular of the alpha male. This enhanced incidence of sexual interactions suggests that, in addition to the opportunity created by the peripheral position of non-alpha males, dyads also exploit such a situation.

Altogether, the simplest explanation of our results is that copulations in absence of the alpha male result from exploiting the peripheral positioning of non-alpha males. Since bystander effects based on the spatial composition of the group do not require high cognitive mechanisms, they may be found in a wide array of species. It would be interesting to investigate whether similar effects of bystanders exist in species or groups where individuals occupy different dominant ranks and were the Priority of Access model does not predict mating success, such as spotted hyenas (C. crocuta: Engh et al. 2002), harbour seals (Phoca vitulina: Coltman et al. 1999) and pukeko (P. porphyrio melanotus: Lambert et al. 1994).

In conclusion, males and females both can harass copulating pairs, both have an audience effect and both inhibit the occurrence of sexual behaviour of their group members of both sexes. The audience effects of non-alpha individuals were not just a side effect of sitting close to the alpha male, and, therefore, represent a true avoidance of potential harassers. Bystander effects express male–male competition and female–female competition, indicating that both are important factors in the sexual dynamics of long-tailed macaques. Reducing sexual competition does not seem to be a result of employing tactical deception. Rather, our results suggest that long-tailed macaque males and females copulate sneakily because they exploit the peripheral position of non-alpha males. Investigating these mechanisms in non-primate species would give more information on whether this effect depends on primate social cognitive capacities or whether it is a broader effect which exists in many species with comparable multi-male multi-female mating systems.

Acknowledgments

We would like to thank Nick Stolk for his contribution to the data collection, three anonymous reviewers for the thoughtful comments and the animal caretakers for their care for the monkeys. This research was funded by the Biomedical Primate Research Centre, Rijswijk, the Netherlands.

Ethical standards

This research complied with protocols approved by the Animal Ethical Committee of the Biomedical Primate Research Centre and with the legal requirements of the Netherlands.

Copyright information

© Springer-Verlag Berlin Heidelberg 2012