p.p1 {margin: 0.
0px 0.0px 12.0px 0.0px; text-align: center; font: 16.0px Times; color: #000000; -webkit-text-stroke: #000000}p.p2 {margin: 0.0px 0.0px 12.
0px 0.0px; font: 12.0px Times; color: #000000; -webkit-text-stroke: #000000}p.p3 {margin: 0.0px 0.0px 12.0px 0.0px; font: 12.
0px Times; color: #000000; -webkit-text-stroke: #000000; min-height: 14.0px}p.p4 {margin: 0.0px 0.0px 12.
0px 0.0px; text-align: justify; font: 12.0px Times; color: #000000; -webkit-text-stroke: #000000}p.p5 {margin: 0.0px 0.
0px 12.0px 0.0px; text-align: justify; font: 12.
0px Times; color: #000000; -webkit-text-stroke: #000000; min-height: 14.0px}p.p6 {margin: 0.0px 0.0px 12.
0px 0.0px; text-align: justify; font: 12.0px Times; color: #9d45b8; -webkit-text-stroke: #9d45b8}li.li4 {margin: 0.0px 0.
0px 12.0px 0.0px; text-align: justify; font: 12.0px Times; color: #000000; -webkit-text-stroke: #000000}span.s1 {font-kerning: none}span.s2 {font-kerning: none; color: #ff5f5e; -webkit-text-stroke: 0px #ff5f5e}span.s3 {font: 14.
4px Times}ol.ol1 {list-style-type: upper-roman}ul.ul1 {list-style-type: disc}Does tolerance in a shared task increase short term grooming and proximity? AbstractIntroductionIn primate groups, cooperative behaviours are often observed among group members. Cooperative behaviours are defined as one individual « helping » another one, where the cooperative act benefits both the actors by enhancing their fitness. Known cooperative acts in primates groups includes for example group hunting in chimpanzees, cooperative breeding in marmosets, or grooming, coalitionary support and alarm calling in vervet monkeys. The broad definition of cooperative behaviours also include altruistic behaviours, where only the « helped » individual receive benefits, while the « helper » provides costly help, with no benefits. Altruism, when directed toward a family member, is called nepotism.
In this paper, we focus on the cooperative behaviours where both actors benefit from the interaction.As humans, we easily see how cooperation can be beneficial, however, those benefits are not so obvious from an evolutionary point of view. Indeed, cooperation leads to the temptation of defection and free riding when it comes to cooperating back. In that regard, the evolutionary stability of cooperative behaviour poses some challenges.Over the years, three big types of cooperating situations were identified, and for each type, an explanation as to how it could be evolutionary stable was found.
These three « types » of cooperation basically answer the question « Why should an individual cooperate? », and form the three pillars of cooperation », with are: Nepotism, Reciprocity, and Mutualism as mentioned in van schaik and kappeller. Despite the necessity of classifying cooperation into different types to explain how the evolutionary mechanisms are working, it is very important to remember that observed cooperative behaviours are likely to be the evolutionary product of nepotism, reciprocity, and mutualism altogether.We will now describe shortly mutualism and nepotism, and then focus on reciprocity.
MutualismMutualism is probably the most straightforward kind of cooperation, as both actors obtain the benefits at the same time. One the simplest examples of mutualistic interaction is found in group living animals, where the simple fact of living together helps reducing the predation risk. In this situation, all actors receive the benefits at the same time. NepotismHamilton proposed that cooperating with someone related, actually helps the cooperator. Helping a family member benefits the copies of one’s own genes contained in that family member. An individual who would initiate cooperative behaviour with its family would be advantaged and selected by natural selection.
This theory is now known as the kin selection theory, and is a powerful tool to explain nepotistic behaviours observed in nature.Kin selection has been ever since well documented among primates 1, especially among female lineage.The inclusive fitness theory of Hamilton 2 predicts that one should cooperate with another according to : the degree of relatedness they bear (R), the benefits provided to the cooperator (B), and the cost (C) for the one cooperating, resulting in the most famous inequality Br > C.
This theoretical advancement lead to a better understanding of mechanisms from which cooperation among kin is likely to have evolved. However, when it comes to unrelated individuals, such evolutionary mechanisms are much more contentious. ReciprocityA possible answer was found in the vast field of reciprocity, which has been argued to be a serious candidate by Axelrod 2. The first variant described was direct reciprocity, where individuals are thought to base the decision of cooperating with someone by considering whether that particular individual did help previously. A way to model direct reciprocity is the repeated prisoners’s dilemma, and from this classic test in game theory, one strong argument for the evolutionary stability of direct reciprocity was found in the Tit for Tat strategy, where the individual starts by cooperating, and then just copies what his partner does, whether he cooperates or defects. This as been shown experimentally, with some studies showing that cooperation between two individuals is enhanced in a reciprocal context 3 4.Then came the theory of indirect reciprocity through, among others, image scoring.
It states that if one helps someone, he is more prone to be helped by someone else, regardless of their identity 5 6 7.This theory has the advantage of getting free of the constraint of having two individuals who must stay together for cooperation to evolve.The next step in reciprocity theory was introduced with the generalised reciprocity , and the idea was that individuals cooperate and do not pay attention to who they are cooperating with, as long as they have been helped previously. This generalised reciprocity theory has been demonstrated 8 9, in an experimental design where females rats were tested for their propensity to cooperate after experienced either cooperation or defection from another individual, regardless of its identity.
Females were first trained to pull a stick in order to provide a food reward for another rat, but not for herself. For three trials, they were then either helped every time, or not at all by another rat, and were then presented the opportunity to help or not another rat in response to their experience. Results were quite clear, showing that cooperation rate was overall higher when females did experience previous help. This study provided a serious empirical support for the generalised reciprocity theory. Despite several empirical examples arguing toward the validity of reciprocity, its relative role was thought to be minor when compared to the importance of kinship 10 and the general opinion about reciprocal behaviours was that they were quite rare.
In response to this common view, a meta-analysis 11 was conducted on the allogrooming behaviour in 9 primates species, including 25 social groups. The statistical analysis aimed to assess the relative roles of both kinship and reciprocity in grooming behaviour variance through the use of semi- partial correlations. The results showed that when controlling for kinship, reciprocity explained 20 percents of the variance, whereas when controlling for reciprocity, kinship only explained 3 percents of the variance.
Even though it could not validate the role of generalised reciprocity in primates, this study was of the most importance to clearly establish the role of direct reciprocity, and emphasise its weight against kinship in how primates make their grooming decisions. Another interesting work has been conducted by chapais , where he considers the importance of individual competence when it comes to choose a partner to cooperate with, opposed to what he describes as the overestimated role of kinship. In cases of competence-dependant cooperation, the decision of choosing a kin-partner over a non-kin partner should only happen when the situation does not require a high skill level. An example of such a situation can be found in thermoregulation with collective sleeping, where the only requirement for an individual to be a good cooperator is its ability to generate body heat. As all individuals are able to produce heat, they all qualify to be good cooperators so choosing anyone will provide the same result, but choosing a kin-partner will also provide the Br part of the benefits we find in Hamilton’s equation. If we consider B to be the thermoregulation benefits, and Br the interaction between the degree of relatedness r with B, we obtain B+Br > B. We see that in the case of low-competence condition, cooperation is expected to be kin-biased, but it is different with high-competence situations.
Despite being enlightening this study was restricted to grooming in response to grooming, and did not take into account key features of primates social life, such as agonistic support or tolerance at foraging sites. Emphasising such features would have been very helpful in order to fully comprehend the extent to which all types of reciprocation drives primates decision rules. Hopefully, those features were assessed in an experimental study by Borgeaud and Bshary (2015). They investigated the decision rules of females vervet monkey grooming reciprocation, and especially focused on tolerance toward subordinates and coalitionary support formation. The experimental design consisted in a shared box experiment meant to be performed by two females among the same group.
Because of the strong matrilineal hierarchy in vervet monkeys, each pair of females consisted of a dominant and a subordinate individual. Females were first trained to recognise a unique colour pattern signalling their own box, to avoid confusion as to who was meant to perform the experiment. Once the training was over and successful, the experimenter was looking for a grooming between two trained females, and once the grooming was over, set up the box experiment. What they looked at was whether the dominant female of the pair was more likely to tolerate the subordinate after being groomed by this specific individual, which meant no aggression would happen during the experiment. They also looked at (i) whether a female joining a conflict was more likely to support a recent grooming partner over a third party; and (ii) if a female was more likely to join a conflict if a recent grooming partner was involved.
They found that females did show increased tolerance toward a recent grooming partner in the box experiment than to someone who did not groom them in the last 60 minutes. Regarding coalitionary support, they found that female were only supporting higher ranking individuals in a conflict. However, the decision of actually joining a conflict was well explained by short term reciprocity with grooming partner : the females were more likely to join a conflict involving a recent grooming partner if the other individual involved in the conflict was lower ranking.
We see in Borgeaud & Bshary 2015 that : grooming a dominant can result in being tolerated by her at a foraging site, here the box experiment.females joining a conflict supported every time the highest ranked individual involved in the conflict.support is given significantly more when grooming happened before (last 60 minutes)In other words, this study shows that grooming can be traded for tolerance, and that it can also be traded against coalitionary support if both grooming partners outrank the individual engaged in the conflict.This study put the emphasis on what the dominant was going to do after being groomed, when they investigated whether the subordinate would be toleratedIn this study, we aim to focus on the subordinates, and investigate their responses to tolerance from a dominant. By looking at how they react, we hope to provide a better understanding of trading dynamics. Following the results of Borgeaud & Bshary 2015, we investigate here several possible reactions to tolerance from a dominant:(i) Given that grooming can be traded for tolerance (Borgeaud & Bshary 2015), we will test whether tolerance can be traded for grooming. We will investigate whether the subordinate is likely to groom the dominant after being tolerated.(ii) We will assess whether tolerance can be exchanged for coalitionary support.
To do so, we will look at how long the subordinate stays within 5 meters of the dominant after being tolerated. We saw in Borgeaud & Bshary 2015 that individuals only supported higher-ranked individuals. Since we here look at the actions of subordinates toward dominants, and that joining a conflict requires to be near the individual one could potentially support, we propose that staying within 5 meters of a dominant could be a sign of willingness to potentially join a conflict.MethodsSpeciesVervet monkeys (Chlorocebus aethiops) are old world monkeys found in southern Africa. The groups they live in are composed of several matrilines, and the overall group size can vary from 5 to 80. Males are the dispersing sex.
They are, like many other monkey species, known for cooperating through, for example, grooming. (Seyfarth & Cheney 1984)Group and study siteThe data collection took place in the Mawana game reserve, in South Africa, in the province of Kwazulu-Natal. This reserve is where the Inkawu Vervet Project is based. It is a large reserve (12000h) where the monkeys can move freely, as much as they want, in natural conditions. There are 5 groups of vervets, among which 3 are frequently studied (Noha, Ankhase, Baie Dankie).In this study, we followed the group of Noha (NH), which is the most habituated group, as they live the nearest from the research station, and are very often followed and experimented on. the group was composed, at the time of the study, of 18 adults, with 6 males and 12 females (see females hierarchy on table 1), and 33 juveniles, for a total of 51 individuals. One of we female we used to follow unfortunately died (Xaix).
Because we could not gather enough data with her, we had to exclude her from the analysis.We also excluded Gene from the analysis, as she was the highest ranked female, and we only followed the subordinate in each experimenting pairs. In total we used 4 females, which were the ones already trained in Borgeaud & Bshary 2015, for the statistical analyses.Experimental protocolThe general idea consisted of performing the same box task they used in Borgeaud & Bshary 2015 , with two adult females.
The goal was to create a situation where the dominant would tolerate the subordinate in an artificial foraging condition.We started following the group at sunrise, when they usually wake up. We looked for 2 females not too distant from each other, with whom we could start an experiment.
We then tried to have their attention by showing to them their own box covers, and the apple which served as a reward. When they were interested, we set up the two boxes, typically distant from less than 5 meters. Once everything was ready, we withdrew so they could access to the boxes, and when they were both in front of their boxes, displaying no aggression, we opened the boxes simultaneously, so they could take the slice of apple.
When the two females took their reward, we counted one successful trial.To make sure that the subordinate was truly being tolerated, we required 5 trials without aggression in a row, through repetitions of the process described above. We then monitored how the subordinate reciprocated for being tolerated by following them after the experiment for a 3 hours long focal.
We recorded every bout of grooming and every bout of time spent within 5 meters of any monkey. We recorded the interactions with every monkeys to see if, in addition to the interactions with the dominant they shared the experiment with, their behaviour toward other monkeys would be affected.Experimental boxes and coversThe boxes and the covers we used were the same as used in Borgeaud & Bshary 2015. The boxes were made of wood, with one door which could be opened, made of plexiglas (Fig??) A mechanism inside the box allowed it to be opened with a remote, controlled by the experimenter.
The remote was made so that two boxes could be opened at the same time. The boxes were made so that they could be nailed onto the ground, preventing the monkeys to take the boxes. During the experiments, a slice of apple was placed inside, serving as the reward. On top of the boxes could be placed covers painted with unique patterns of colors and shapes (see fig??).
Each unique cover was associated to one female, so that we could signal to the monkeys who was supposed to perform the experiment. In that way, we were able to get the pairs of females we wanted to test to come to their boxes, and complete the task.Focal data were voice recorded in order to allow more precision in data collection.
Cooperative dyads We first trained the monkeys to form cooperative dyads in order to get the reward from the boxes. As a follow-up of the study by Borgeaud and Bshary 2015, our experiment was built on the monkeys’ knowledge of their own box covers 12. To ensure the monkeys remembered how the boxes worked, we made preliminary experiments with higher distances between the boxes, such as 7 to 10 meters, in order to prevent high levels of aggression. Once the monkeys were reminded of how the task worked, we reduced the distances to less than 5 meters, in order to create a true cooperative task where tolerance was needed. Our intention being full cooperation between monkeys, the boxes were not opened until the two requested participants were in front of their boxes, as they were trained to be. Should the dominant individual prevent the subordinate to access their box, no reward was delivered for both of them until the subordinate was tolerated in further trials. We considered tolerance to be achieved when 5 trials could be performed in a row. Focal data The focal begun 10 minutes after the end of the experiment, so the monkey could return to a natural behaviour, with less human interference.
In order to collect enough social interactions, each focal lasted for ideally 3 hours. During the focal, an appropriate distance was kept in order to let the monkey act as naturally as possible. This distance varied from 10 meters for the most habituated subjects, to more than 30 for the shyest ones.We obtained collected and analyzed approximately 80 hours of focal data.We also followed females which were not involved in any experiments on that day, in order to collect baseline focals, to which we compared what was happening in the experimental focals. Statistical analyses The first thing we did with the data was to assess, for each female focaled in both baseline and experimental conditions, whether their proximity behaviour was altered by the box experiments. To do so we took, for each female all the datapoints in baseline conditions, and all the datapoint in every experimental condition they had been involved in. We compared the proportion of time they spent with every individuals in the group in both conditions (see table 1) with a paired wilcoxon signed rank test.
We then aimed to assess, for one focal female, whether she would spend more time within 5 meters of the dominant female she experimented with, than in baseline condition. To do so, we used a paired wilcoxon signed rank test, comparing every bouts of time they spent together in Baseline against Experimental condition (See table 4).For the next analysis, we calculated the proportion of total proximity time each focal female spent with the other females (See all the pairs in Table 4) she performed an experiment with in experimental condition, and then in baseline condition. We put together the data we calculated (see table???).
We then tested with a paired wilcoxon signed rank test whether the focals would spend a different proportion of time with an individual they performed the experiment with, in experimental condition than in baseline condition.We followed the exact same steps for the grooming behaviours.All statistical analyses were performed with R version 3.3.3 (2017-03-06) — « Another Canoe », on max OS X Yosemite (10.10 ).ResultsProximityThe comparison of proximity data with every monkeys for each females in baseline against experimental condition yielded no significant results with the wilcoxon signed rank test.
Suggesting the boxes experiments did not make a significant change in their global proximity behaviour. (See p-values in table 2) The test of the proportion of time spent within five meters in experimental condition against baseline condition, for each pair of focal individuals, and experiment partners (See Table 4). The test did not work when no data was available for one condition.
We will get over this problem in the next analysis. We learn here that, when the test worked, no result was significant. II Bis) Can we use a big chi-square to assess with who she spent different times in exp than base?III) GroomingThe comparison of grooming data with every monkeys for each females in baseline against experimental condition yielded no significant results with the wilcoxon signed rank test. Suggesting the boxes experiments did not make a significant change in their global proximity behaviour. (See p-values in table 3). It seems however, that the females were more disturbed in their grooming behaviours than in their proximity behaviour.
II)DiscussionThe tests performed in the first (P.I) proximity analyze, with every tests non significants, suggest that the stress induced by the boxes experimentations was not enough the significantly alter the females’ proximity behaviour. Three out of four p-values are very strong, showing that the individuals were almost not affected by the experimentation’s interference in their proximity behaviours. It appears, however, that the individual Pret was more disturbed than the three other females. As we explained earlier, the individual Xaix, who was one rank above Pret (See Table 1), died in the middle of our data collection.
As this change in the hierarchy was directly, above her, she took Xaix’s place in the hierarchy. We suggest that it induced a stress on Pret, and that, possibly coupled with the stress of the experimentations, her proximity behaviour was altered, although not significantly.The tests performed in the first (G.
I) proximity analyze, with every tests non significants, suggest that the stress induced by the boxes experimentations was not enough the significantly alter the females’ proximity behaviour. Interestingly, the individual Bogo almost had increased grooming habits in experimental conditions.LimitationsLike every study, our design has flows, which we tried to limit as much as we could.
Our results concerning the proximity with other monkeys are not the same as they would have been without the interference of the observer. Indeed, the presence of the observer may have prevented the shyest individuals to approach the female being followed, as they might have done in perfectly natural conditions. Despite the disadvantages of following the monkeys, it was the only way we could collect the data.The increased presence within 5 meters of the female the focal individual experimented with may have nothing to do with our experiments. As we said in the methods, we started the box experiment whenever we found two females not too far from each other.
The observed increased presence within 5 meters might be due to the fact that the two females were just close to each other, and followed the same foraging path along the focal. However, we often observed monkeys to switch from different subgroups along the day, which makes our results more likely to be significant.ConclusionsAnnexe
