(1) Infections can change social behaviour in multiple ways, with profound impacts on pathogen transmission. However, these impacts might depend on the type of behaviour, how sociality as a biological trait is defined (e.g. network degree versus mean edge strength), and the type of social relationship between the interacting individuals. (2) We used the highly social common vampire bat (Desmodus rotundus) to test how an immune challenge by lipopolysaccharide (LPS) injections affects two different social behaviors and three alternate measures of sociality, and whether the LPS effect differs by kinship relationship. (3) Effects of sickness should be lower for social behaviours that bestow greater benefits to inclusive fitness, such as food sharing. As predicted, immune-challenged bats experienced a greater reduction in allogrooming received than food sharing received. (4) Sickness effects might also depend on how a social interaction is defined (e.g. the number of grooming partners versus the duration of grooming events). We predicted that sickness would impact both the number and duration of social encounters, but we only detected a decrease in the number of grooming partners. (5) Finally, sickness effects might vary with social relationship type. We predicted that sickness effects should be smaller for interactions among close kin. As expected, the immune-challenge had smaller effects on mother-offspring interactions. (6) In conclusion, our results highlight the need to explicitly consider how the effects of sickness on social network structure can differ depending on the "who, what, and how" of social interactions, because these factors are likely to influence how sickness behaviour alters pathogen transmission.
