Data on allogrooming in mandrills and Japanese macaques and food sharing in vampire bats

Cooperation events from three studies. The first two studies were conducted on mandrills (Mandrillus sphinx) and Japanese macaques (Macaca fuscata) housed at the Rome Zoo (Bioparco) in Italy. An observer recorded the duration of all female-female grooming episodes involving a focal subject, either as actor or recipient. All female subjects were available as potential grooming partners throughout the study. Kinship was based on maternal pedigrees. The first dataset contained 1703 observations of mandrill allogrooming collected between July 2014 and June 2015 from 10 sexually mature female mandrills in a group that also included two mature males. A past study of six female mandrills from the same captive population found that allogrooming A to B predicted allogrooming B to A, when controlling for kinship (Schino and Aureli 2010c), or when controlling for kinship and rank and excluding recent reciprocal grooming (Schino and Pellegrini 2009).

The second dataset contained 737 observations of macaque allogrooming collected between April and November 1996 from 22 sexually mature female Japanese macaques (Macaca fuscata) in a group of 71 that also included mature males and juveniles. Similarly, analyses of grooming in the same captive population of Japanese macaques also found symmetry in females allogrooming, and also found that allogrooming predicted support in social conflicts when controlling for kinship, rank, or time spent in proximity (Schino et al. 2007); however, grooming was better predicted by kinship than by grooming received (Schino and Aureli 2010c).

The third dataset included 408 regurgitated food-sharing donations among 15 female common vampire bats (Desmodus rotundus) from previous studies, where food sharing was induced by fasting a subject (for details, see Carter and Wilkinson 2013; Carter and Wilkinson 2015). Each donation size was estimated by the total seconds that the unfed subject licked the mouth of a fed groupmate. The potential donors varied across some trials, and we accounted for this in our permutation procedure by providing a list of potential fed donors that were present for each observation. Kinship was estimated using a maternal pedigree and maximum likelihood estimates applied to genotypes of 19 polymorphic microsatellite markers (for details see Carter and Wilkinson 2015). Past analyses of the same data found that food sharing was better predicted by reciprocal sharing than by kinship, when controlling for grooming and donor sex (Carter and Wilkinson 2013a; Carter and Wilkinson 2013b). This conclusion was supported by later experiments showing that the bats were attracted to the calls of unrelated past helpers more than related non-helpers (Carter and Wilkinson 2016) and that bats that previously fed more nonkin had food-sharing networks that were more robust to removal of donors (Carter et al. 2017).