Study sites and subjects
Fieldwork was conducted at Wamba, Luo Scientific Reserve, Democratic Republic of Congo (00°10′N, 22°30′E). We followed two neighbouring communities of wild bonobos: E1 group (n = 39) has been habituated since 1974, and P group (n = 30) has been habituated since 2010. In 2014, the total sample size was 63 individuals, with 28 adults (16 females, 12 males), 12 adolescents (7 females, 5 males), 9 juveniles (6 females, 3 males), and 14 infants (8 females, 6 males). In 2015, the total sample size was 64 individuals, with 30 adults (18 females, 12 males), 8 adolescents (3 females, 5 males), 10 juveniles (7 females, 3 males), and 16 infants (10 females, 6 males).
This study was approved by the School of Psychology and Neuroscience Ethics Committee at the University of St Andrews, and permission to conduct the study was granted by the Ministère de la Recherche Scientifique et Technologie in the Democratic Republic of the Congo. Data collection took place from February to June 2014 and January to June 2015. We conducted daily observations from approximately 05:50 to 12:00, with a rough schedule of 4-day working and 1-day off, observing bonobos on a total of 204 days, amounting to ~1159 h of observation time.
We used focal behaviour sampling to film social interactions. Filming began whenever two or more individuals came within 5 m range of each other, in order to catch the beginning of social interactions. We recorded video footage using a Panasonic HDC-SD90 video camera, which has a pre-record feature that continually records the previous 3 s. Each day after returning from daily follows, we imported footage and sorted it into a clip directory using FileMaker Pro.
Gestures were defined as discrete, mechanically ineffective physical movements of the body observed during periods of intentional communication, including movements of the whole body, limbs and head, but not facial expressions or static body postures. We created a separate coding sheet in Filemaker Pro for each gesture instance, recording the following information: signaller, recipient, signaller age/sex, recipient age/sex, gesture type, part of sequence, part of bout, audience checking, response waiting, persistence, and signaller apparently satisfied. Signaller is the gesturing individual, and recipient is the individual to whom the gesture is directed. Age groups are taken from Hashimoto’s bonobo age classification (Hashimoto 1997): infant (<4 years), juvenile (4–7 years), adolescent (8–14 years), and adult (15+ years). Gesture type is defined by the physical form of the gesture, where possible following definitions are used with the chimpanzee (Hobaiter and Byrne 2011), but adding new definitions for gesture types that have not been reported in the chimpanzee. A sequence is defined as a series of gesture instances given by one individual, separated by <1 s. A bout is defined as a series of gesture instances or sequences given by one individual, separated by pauses of >1 s. Audience checking is when the signaller turns to face the recipient before or during gesturing. Response waiting is when the signaller pauses for >1 s after gesturing while maintaining visual contact. Persistence is when the signaller continues to gesture at the same recipient. Each instance of a gesture was required to meet at least one criterion for intentionality before it was accepted for analysis: audience checking, response waiting, or persistence.
For the expressed repertoire, we included all gesture types that an individual deployed. The understood repertoire, however, was not simply the gesture types that an individual received, but the gesture types that they understood. We took it that the recipient understood a gesture instance if the recipient reacted with an apparently satisfactory outcome (ASO)—i.e. a reaction that satisfied the signaller, as shown by cessation of gesturing. The signaller should start to react during gesturing or immediately following cessation of gesturing. Note that an ASO must be a change in behaviour: if the recipient remains in the same state and the signaller stops gesturing, there was no change in behaviour from the recipient, and thus, we coded “No response”, not “ASO”. For gestures occurring in sequences, if the recipient responded to the sequence with an ASO, that ASO was assigned to all gestures in the sequence, not only the final gesture instance in the sequence.
To corroborate the accuracy of our video coding, a second experienced coder, Dr Catherine Hobaiter, coded 100 gesture instances for the following information: gesture type, persistence, and signaller apparently satisfied. We calculated inter-observer reliability using Cohen’s Kappa, revealing agreement for all variables (gesture type K = 0.87, persistence K = 0.70, and signaller apparently satisfied K = 0.63).