Social network analysis uncovers hidden social complexity in giant pandas

Abstract

Abstract: Analyses of animal social networks have traditionally been conducted on species that exhibit social behaviors such as group living, whereas relatively less work has been done on species that are thought of as solitary, are cryptic, and that communicate through scent-marking cues. We employed noninvasive fecal genetic sampling, conducted from March 2015 to February 2016, to identify individuals of one such species, the giant panda (Ailuropoda melanoleuca), across a study population in southwestern China. We then used spatiotemporal proximity thresholds to infer scent mark–based association networks and conduct social network analyses of the population. Results show social clustering in which cluster members preferentially associated with each other. Genetic relatedness was a positive predictor of associations outside the mating season but a negative predictor of associations in the mating season (Mar–Jun), potentially indicating a behavioral change that would reduce risk of inbreeding and kin competition that we term “kin-space recognition.” Our findings suggest that associations between individuals in species generally thought of as solitary may be widespread and dependent on complex behavior.