Prosociality is the intent to improve others’ well-being. Existing hypotheses postulate that enhanced social tolerance and inter-individual dependence may facilitate prosocial preferences, which may favor the evolution of altruism. While most studies are restricted to ‘tolerant’ (cooperatively breeding and self-domesticated) species, despotic societies provide an alternative opportunity to investigate prosociality due to nepotism and ample inter-individual dependencies. Japanese macaques live in hierarchical matrilineal societies, with strong kin bonds. Besides, tolerance among non-kin may persist through reciprocity. Using a group service food-provision paradigm, we found prosocial preferences in a semi free-ranging group of Japanese macaques. The extent of provisioning was at levels comparable to tolerant species. Dyadic tolerance predicted the likelihood and magnitude of provisioning, while kinship predicted the magnitude. We emphasize the role of a complex socio-ecology fostering individual prosocial tendencies through kinship and tolerance. These findings necessitate a framework including different forms of interdependence beyond the generally tolerant species.

(a) Study site and subjects: The study was conducted with a semi-free-ranging group (~170 individuals) of Japanese macaques at Affenberg Landskron in Austria from February to May 2022. The macaques live in natural socio-environmental conditions with limited human intervention, except for feeding and emergency medical purposes (such as veterinary care). However, the group is habituated to human presence as visitors can enter the enclosure within the scope of guided tours offered by the park from April to September. Our study gained high ecological validity due to the population's living conditions and naturalistic group setting.

A total of 25 self-trained individuals voluntarily participated in the study while remaining in their social group. Direct interaction between the subjects and the experimenter was avoided by placing the apparatuses needed for our experiments in small research huts (with wire mesh separations), which the macaques could operate from outside. Affenberg Landskron provided age, sex, and rank data (see raw data sheet “large”) of the individuals as long-term behavioural observations were being conducted for other studies with the population.

(b) Apparatuses: We used a seesaw mechanism for the group service paradigm (Burkart and van Schaik, 2013; Horn et al., 2020), whereas a string-pulling task was conducted to measure dyadic social tolerance (Massen et al., 2015). The seesaw tool used a wooden board (length ~ 1.5 m) attached to the grids of the wire mesh separation of a research hut. The wooden board had two transparent plastic pipes (Ø ~ 3 inches) attached to two extremes (pos 0 and pos 1) through which food rewards (i.e., peanuts) could move. By default, the board was tilted towards the experimenter. A metal handle was connected to one of the ends of the board (pos 0) and projected outside, which upon pressing, could tilt the board towards the macaques, resulting in food rolling down through the pipe(s) (if present). In an attempt to press or when an individual released the handle halfway, the seesaw moved back to its original position, and the food (if present) would role out of reach again. Therefore, it was essential to press fully and hold the handle to get the food rewards.

For the string-pulling (dyadic tolerance) task, we used a wooden platform (length ~ 1.5 m) to which two strings were attached at the two extremes. Food rewards were placed out of reach of the macaques on two ends of the platform, which could only be moved by pulling one of the strings. The task required no joint action. The length of the apparatuses was kept constant.

(c) Experimental Procedure:

(i) Group service paradigm – The protocol consisted of several phases: habituation to the apparatus, social tolerance phase, apparatus training, group service (test and empty control), and blocked controls in a pre-determined order (see below). Each phase served a purpose and came with a criterion that needed to be fulfilled to begin the next phase. The rewards used were peanuts.

Phase 0 (Habituation phase)

This phase was carried out to reduce and/or eliminate any potential neophobia towards the seesaw apparatus and to habituate individuals to the setup. Food items were spread over the wooden board at regular intervals. The experimenter tried to catch the attention of the individuals by calling “Monkeys”. Three sessions were conducted (each on one day for ~1 hour). A total of 25 individuals participated in this phase voluntarily and obtained food rewards successfully.

Phase 1 (Initial training and habituation to procedure)

The mechanism of the seesaw was completely locked, and the platform was tilted towards the individuals. So, when placed in either position, food rewards would roll down to the individuals. We conducted a total of five sessions, and for each session,  food was provided in pos. 0 and pos. 1, alternately. The number of trials was determined according to the number of individuals who successfully passed phase 0 (25 * 5 = 125 trials/session, see (Horn et al., 2016)). This number was used for the following phases as well. A trial began when a food reward was placed (either in Pos. 0 or Pos. 1) and ended with an individual retrieving it or after 2 min. In this phase, an individual was considered trained if they received at least a total of 10 rewards throughout the different sessions. To fulfil the criterion, at least half of the habituated individuals needed to get trained. We recorded which individual(s) obtained the rewards for each trial.

Phase 2 (Social tolerance test)

The procedure was the same as in phase 1, with the mechanism still locked and the platform tilted toward the individuals. However, food rewards were always placed in Pos. 1 in this phase. We conducted two sessions, with each having 125 trials.

Phase 3 (Apparatus training)

The seesaw mechanism was operational from this phase and onwards. Therefore, individuals needed to learn to press the handle to tilt the board and get the rewards. Food rewards were always placed in Pos. 0. Thus, individuals pressing and holding the handle could make food rewards available to themselves. The phase was completed when at least half of the individuals obtained at least ten rewards over five sessions.

Phase 4 (group service: test and empty control)

This phase was the core of the experiment. In test sessions, food was placed in Pos. 1 for each of the 125 trials of the test sessions. An individual needed to press the handle in Pos. 0 to make food accessible for another individual in Pos.1. The trial began when the experimenter placed a peanut in the pipe in Pos. 1 while drawing the attention of the macaques (by calling “Monkeys”). An actor could provision by pressing and holding the handle within a maximum duration of 2 min. If not, the experimenter removed the reward from the pipe and started a new trial. Trials were conducted at different times throughout the day, from 0930 to 1830 hours. We carried out empty control sessions alternately with the test sessions. Here, the experimenter pretended to put a reward in Pos. 1 but left the pipe empty to check for potential stimulus enhancement effects. The maximum duration of a trial was again 2 min.

Phase 5 (Blocked controls)

The access to the food pipe in Pos. 1 was blocked by attaching plexiglass to the enclosure. Food reward was thus visible but not accessible to the individuals, even if actors pressed the handle. We conducted five blocked test sessions during which food was placed in Pos. 1 (similarly to test sessions in phase 4), alternately with five empty (blocked) control sessions (no food placed in Pos. 1).

Phase 6 (Re-test and empty controls)

To check for any potential bias due to the experimental order (Horn et al., 2016; Thornton and McAuliffe, 2015), we repeated two tests and two empty control sessions. The procedure was exactly the same as the regular test and empty control sessions.

We recorded the identities of the individuals pressing, provisioning, and receiving for all the above phases (wherever applicable). Also, we measured the latencies for pressing the handle for all the trials. For phases 4, 5 and 6, every fifth trial was a motivation trial. Food was placed in Pos. 0, accessible to the actor (if pressed). These motivation trials kept the monkeys interested in participating in the experiment. Additionally, these trials helped check if low pressing rates (if any) resulted from a general lack of willingness to participate in the experiment.

(ii) String-pulling (dyadic tolerance) task – The string-pulling task was designed to assess dyadic social tolerance. This task was independent of the group service paradigm and was performed afterwards.

Habituation to apparatus

We first conducted a habituation phase for a whole day (between 0930 and 1830 hours). We put food rewards on the platform and moved it slowly toward the monkeys. This phase aimed to familiarize the monkeys with the apparatus. Individuals were considered habituated after obtaining at least five rewards during the habituation phase.

Dyadic social tolerance

The experiment lasted three days, and we conducted 18 sessions of 20 trials each. Every trial began by calling “Monkeys” while placing food rewards (peanuts) on either end of the platform yet out of reach of the monkeys. Once the monkeys were attentive, we presented both strings at the same time. To access the rewards, at least one individual needed to pull. The trial ended when the platform was moved and rewards obtained or after 2 min. When the platform was moved, two pieces of food rewards were accessible, placed ~1 m apart, meaning that a high social tolerance was necessary for two individuals to retrieve the rewards simultaneously. Thus, monopolization of both rewards by a single individual was also possible.

We recorded all the phases of the two experiments using a Canon Legria HF G86 video camera mounted on a tripod. Additionally, we noted the identities of the individuals participating using a pen and a notebook. 

Microsoft Excel 365

R-version 4.2.2