Group-living animals often maintain a few very close affiliative relationships – social bonds – that can buffer them against many of the inevitable costs of gregariousness. Kinship plays a central role in the development of such social bonds. The bulk of research on kin biases in sociality has focused on philopatric females, who typically live in deeply kin-structured systems, with matrilineal dominance rank inheritance and life-long familiarity between kin. Closely related males, in contrast, are usually not close in rank or familiar, which offers the opportunity to test the importance of kinship per se in the formation of social bonds. So far, however, kin biases in male social bonding have only been tested in philopatric males, where familiarity remains a confounding factor. Here, we studied bonds between male Assamese macaques, a species in which males disperse from their natal groups, and in which male bonds are known to affect fitness. Combining extensive behavioural data on 43 adult males over a 10-year period with DNA microsatellite relatedness analyses, we find that post-dispersal males form stronger relationships with the few close kin available in the group than with the average non-kin. However, males form the majority of their bonds with non-kin, and may choose non-kin over available close kin to bond with. Our results show that kinship facilitates bond formation, but is not a prerequisite for it, which suggests that strong bonds are not restricted to kin in male mammals and that animals cooperate for both direct and indirect fitness benefits.

This data is part of a long-term research project on a population of fully habituated Assamese macaques ranging in their natural habitat in the Phu Khieo Wildlife Sanctuary (PKWS) in north-eastern Thailand. We studied one group from 2006 to 2012, 3 groups from 2012 to 2014 and 4 groups from 2014 to 2017 on a total of 43 adult males. The study period was divided into stable periods (6-8 per group), defined as periods during which adult male group composition did not change. Observational data were collected almost daily, from dawn to dusk, using 20 to 40 minutes continuous focal animal sampling. During these protocols, frequencies and durations of affiliative behaviours (i.e. grooming, body contact and proximity within a 1.5 m radius) were recorded, as well as agonistic interactions. To measure social relationship strength, we used the dynamic dyadic sociality index (DDSI); dominance ranks were calculated based on the Elo-rating method. To assess relatedness we genotyped all 136 adult individuals of the population at 17 microsatellite loci and one mtDNA locus (details see manuscript). We estimated relatedness using the triadic maximum likelihood estimator, and assigned dyads with r > 0.25 to close kin and dyads with r < 0.125 to non-kin. Our research was completely non-invasive and followed the ASAB/ABS Guidelines for the Use of Animals in Research. Data collection and export of samples were authorized by the Department of National Parks, Wildlife and Plant Conservation (DNP) and the National Research Council of Thailand (NRCT) with a benefit sharing agreement (permit numbers: 0004.3/3618, 0002.3/2647, 0002/17, 0002/2424, 0002/470).