Social learning of arbitrary food preferences in bonobos

Do chimpanzee food calls bias listeners toward novel items?

Social learning is beneficial in almost every domain of a social animal's life, but it is particularly important in the context of predation and foraging. In both contexts, social animals tend to produce acoustically distinct vocalizations, alarms, and food calls, which have remained somewhat of an evolutionary conundrum as they appear to be costly for the signaller. Here, we investigated the hypothesis that food calls function to direct others toward novel food items, using a playback experiment on a group of chimpanzees. We showed chimpanzees novel (plausibly edible) items while simultaneously playing either conspecific food calls or acoustically similar greeting calls as a control. We found that individuals responded by staying longer near items previously associated with food calls even in the absence of these vocalizations, and peered more at these items compared with the control items, provided no conspecifics were nearby. We also found that once chimpanzees had access to both item types, they interacted more with the one previously associated with food calls than the control items. However, we found no evidence of social learning per se. Given these effects, we propose that food calls may gate and thus facilitate social learning by directing listeners' attention to new feeding opportunities, which if integrated with additional cues could ultimately lead to new food preferences within social groups.

Bonobos assign meaning to food calls based on caller food preferences

Human communication relies heavily on pragmatic competence. Speech utterances are often ambiguous requiring listeners to use interaction history, shared knowledge, presumed intention and other contextual variables to make inferences about a speaker’s meaning. To probe the evolutionary origins of pragmatic competence we tested whether bonobos (Pan paniscus) can make inferences about the type of food available from listening to other group members’ food calls. We trained two group members to either prefer blue or pink chow and demonstrated these preferences to observers. A third group member served as an untrained control. In playback experiments, we broadcast the food calls of a trained demonstrator and the untrained group member to investigate whether subjects were able to infer which coloured chow was most likely available, based on the callers’ trained food preferences or lack thereof. As predicted, when hearing the untrained group member’s calls, subjects did not exhibit a bias, whereas they responded with a significant foraging bias when hearing a trained group member’s calls. These findings suggest that bonobos may take into account the idiosyncratic food preferences of others, although subjects probably differed in what they remembered.

A Food for All Seasons: Stability of Food Preferences in Gorillas across Testing Methods and Seasons

Decisions about which foods to use during training and enrichment for captive animals may be based on invalid assumptions about individuals’ preferences. It is important to assess the stability of food preferences given that one-time preferences are often used to inform which items are offered over a longer period of time. Presenting preference assessments using images of food items allows control over factors such as size, scent, and inadvertent cueing but requires validation. We presented three male gorillas with choices between randomly selected pairs of actual food items from their morning meal using PVC feeders. We also presented the gorillas with two-alternative forced-choice tests between images of these foods on a touchscreen computer. Ranked preferences were correlated across method and seasons. Furthermore, gorillas selected images of preferred over less preferred foods in a validation task on the touchscreen. However, selections of some food items changed within sessions, suggesting that preference may be relative to other contextual factors. Researchers should assess how choices affect subsequent preferences to understand whether animals demonstrate absolute preferences for particular food items, or prefer variety.

Food preference and nutrient composition in captive bonobos (Pan paniscus)

Food preference has been studied in a range of Hominoidea in the wild and in captivity, allowing for interspecific comparisons. Chimpanzees (Pan troglodytes) and gorillas (Gorilla gorilla gorilla) prefer low-fibre, high-sugar foods, suggesting that frugivory and their dietary overlap are a result of their shared preference for the same nutrients. Comparable tests of the nutritional preference of bonobos do not exist. In this study we examined food preferences of five captive bonobos for 23 familiar and ten novel food items. We performed paired-choice food tests, resulting in a clear rank order in food preference, with minor individual differences. Fruits were more preferred than vegetables. We correlated nutritional composition of the food items with the bonobos' preference. We found that preferences for familiar food items were positively correlated with total energy and carbohydrate content and negatively correlated with water and micronutrient (sodium, calcium, phosphorus, iron, zinc, manganese, selenium) content. Food preference for the novel food items also showed a significant positive correlation with total energy and carbohydrate content. Our study supports the idea that food preference among bonobos follows the pattern of the other great apes and that the shared frugivorous diets may be the result of a common preference for the same nutrients. In the wild, these preferences may be less clear due to the interference of preferred nutrients with secondary compounds. Combining food preference data and nutritional information can help in providing a healthy diet with a balanced nutrient composition in captivity. Individual food preferences can help in optimizing food choice for positive reinforcement training and food-related tasks in future research.