Cooperation in Capuchin Monkeys
...d show a higher level of cooperation than that shown in the study by Chalmeau et al. (1997) by presenting a more intuitive task under more controlled conditions. They also tested several predictions to determine whether the monkeys were taking into account their partner’s actions. The researchers hypothesized that by eliminating visual contact between partners, the monkeys’ success would decrease over the course of trials. They also predicted that the monkeys would adjust their actions to coincide with those of their partners. The researchers in this study designed a task in which two subjects needed to work together to obtain a food reward. The subjects were adult capuchins from two social groups and were housed in indoor-outdoor pens that were visually separated by an opaque screen, although they could still have auditory contact. After each monkey had learned individually to pull a tray into a locked position in order to obtain a food reward, the monkeys were tested in pairs and the tray was counter-weighted according to the particular test type. Instead of allowing all individuals access the apparatus at the same time, the researchers designated pairs of monkeys and tested them separately from the rest of the group. In the test chamber the subjects were separated by a mesh partition and were presented food that could only be reached if both individuals simultaneously pulled on the bars. Each monkey in a pair was able to serve as the recipient of the food reward and an unrewarded helper on alternating trials. This set-up, along with the mesh partition, prevented any monopolization by dominant individuals. One of the four test types that the researchers used in this study was a cooperation test (COP), in which two monkeys were needed to pull the bars in order to obtain the food. The second test was a solitary effort test (SOL), in which the tray was weighted so that one individual could pull the food within reach by itself. The third test was an obstructed view test (OBS), which was similar to the COP test except that an opaque panel was placed between the monkeys instead of a mesh partition, in order to eliminate visual contact. The fourth test type was an unrestricted cooperation test (UCP), in which the researchers removed the plastic panel that normally kept the monkeys from leaving the test chamber and entering the group cage. This allowed the second monkey to move between the test chamber and the adjoining group cage whenever he or she preferred. As the researchers predicted, results showed that individual 1 in each test group glanced more often at its partner in cooperative tests, when assistance was needed, than under the solitary effort condition. The success rate was also found to be higher in cooperative tests than in obstructed view tests, and success decreased over the course of trials, as the researchers hypothesized. Furthermore, individual 1 pulled at a significantly higher rate when individual 2 was in the chamber than when he or she was outside. In a study by Visalberghi et al. (2000) the researchers tried to increase task predictability and to improve the motivational and emotional aspects of the test environment for capuchin monkeys, without altering the cognitive aspects of the task used in Chalmeau’s (1997) study. They tested the monkeys in pairs so that the partners were the same across trials in order to make their behavior more predictable for one another. Furthermore, both partners were rewarded when pulling a handle simultaneously, and the elements of their task, as well as the consequences of their pulling actions, were visible to the monkeys. Visalberghi et al. (2000) wanted to test the extent to which the monkeys take into account their partner’s behavior and spatial position and whether they understand what cooperation entails. The researchers hypothesized that their task and procedure were more favorable for cooperation than the study performed by Chalmeau et al. (1997). Eight capuchin monkeys were used and were first trained to pull a handle when a reward was on a tray so that the researchers could better evaluate each monkey’s appreciation of the partner’s role (Visalberghi et al. 2000). The test apparatus was made of Plexiglas and had two metal handles, a trough below each handle, and a tray above both of them. A peanut reward could only be obtained when the two handles were pulled simultaneously, tilting the trays. Although the task was solved by all of the pairs of monkeys, the number of instances of cooperation was different among them. In addition, three of the monkeys used in this experiment had also been used in the study conducted by Chalmeau et al. (1997). Results showed, however, that even though these monkeys were more experienced with the task and had a greater chance of mastering it, they did not cooperate more than the other, less-expert subjects. The number of occasions of cooperation accomplished by each monkey did not correlate with the number of pulls it performed. The calculated number of pulls per session showed that pulling was significantly higher when the reward was on the tray than when it was not. Furthermore, the position of the partner affected the number of pulls performed by the monkeys. In particular, the monkeys pulled with the same frequency whether the partner was on or off the platform, but they pulled less often when the partner was on the platform and not within reach of the other handle than when the partner was within reach. The monkeys also pulled the same amount when the partner was on the platform and within reach of the other handle as when the partner was actually pulling the other handle. These results suggest that whereas the presence of the partner near the handle increased the rate of pulling, the action of pulling performed by the partner did not (Visalberghi et al. 2000). The finding that capuchins pull and keep pulling in impossible situations indicates that they have no understanding of the role of the partner in making their pulling action effective (Visalberghi et al. 2000). The last study I am going to discuss was done by de Waal and Davis (2003). The researchers wanted to imitate a group hunting situation, in which the individuals collect around their captured prey, by allowing the monkeys in this experiment to move around freely during the pulling task instead of being confined to separate areas, as done previously. This way, the researchers assumed that cooperation partners could compete over the acquired food (de Waal & Davis 2003). They further manipulated opportunities for competition by presenting the food in clumped versus dispersed distribution, and they manipulated the tendency for competition by comparing related and unrelated pairs of monkeys. One goal of the study was to analyze decision-making regarding cooperation by having the subjects face a cooperation task with variable partners and variable food distributions. Second, by investigating how cooperative tendency varied with the potential for competition, the researchers wanted to determine the speed of the decision-making process. They wanted to know whether monkeys learn incrementally or if they make instantaneous adaptive decisions. Adult capuchin monkeys were used from two different social groups. Of the female pairs, half were related, which included all mother-daughter pairs. None of the monkeys had been in a study like this one before, but initial training was very easy. For testing, each monkey pair was separated from their group in a large indoor mesh cage with a tray on a track outside the cage. Three distinct reward distributions were created including C, clumped (where both cups were together in the middle of the tray, each holding food), S, standard (where cups lined up with the pull bars) and D, dispersed (where cups were at the extreme ends of the long tray). The researchers hypothesized that competition would occur the most under the clumped condition. As expected, the monkeys performed best on standard tests and there was a gradual drop in performance over the duration of all three test conditions. Performance in corresponding trials was the lowest under the clumped condition and this was visible in the very first trial and maintained throughout the test. A rotating test, in which C-trials were interspersed with the other two types of trials, showed basically the same results as tests of a single condition, with lower success in C than D trials. The negative impact of clumping on cooperative performance was therefore instantaneous and was not disturbed by alternating it with other conditions (de Waal & Davis 2003). Further results showed that cooperative success dropped under the clumped condition, especially in non-kin pairs, which was explained by the tendency of dominant non-kin to claim more than half of the rewards under the clumped condition. All the effects including food distribution, trial number, and kinship were found to be significant. According to de Waal and Davis (2003), this study demonstrates that in an experimental task, capuchin monkeys decide to cooperate based on their chances at obtaining rewards. Furthermore, the monkeys based decisions about cooperation on anticipated pay-offs, taking into account both the possibility and the likelihood of competition with their partner (de Waal & Davis 2003). Discussion Contrary to the findings by Chalmeau et al. (1997) and by Visalberghi et al. (2000), that cooperation among capuchins may be due do the chance co-occurrence of individual actions and is not cognitively grounded, the results obtained by Mendres and de Waal (2000) showed that capuchins are able to learn certain aspects of their partners’ roles in a cooperative task. The main evidence is that the capuchins were significantly less successful in completing the task when visual contact between them was blocked. The monkeys did not pull the bar less under this condition, but they were less successful, suggesting that coordination, not motivation, was the issue (Mendres & de Waal 2000). The ...