Abstract
A fundamental attribute of social intelligence is the ability to monitor third-party relationships, which has been repeatedly demonstrated in primates, and recently also in captive ravens. It is yet unknown how ravens make use of this ability when dealing with different types of social relationships simultaneously during complex real-life situations. Free-ranging non-breeder ravens live in societies characterized by high fission–fusion dynamics and structured by age, pair-bond status and kinship. Here, we show that free-ranging ravens modify communication during conflicts according to audience composition. When being attacked by dominant conspecifics, victims of aggression signal their distress via defensive calls. Victims increased call rates when their kin were in the bystander audience, but reduced call rates when the bystanders were bonding partners of their aggressors. Hence, ravens use social knowledge flexibly and probably based on their own need (i.e. alert nearby allies and avoid alerting nearby rivals).
Highlights
Sociality is thought to have driven the evolution of cognitive abilities in primates [1,2,3,4] and other mammals, and possibly in birds like parrots and corvids [7]
Our results show that victims modulate their call rates according to the presence of particular individuals in the audience
Compared with audience effects found in avian alarm calls [19] or food calls [20], signalling ravens take into account bystanders with whom they have a valuable relationship, and bystanders that have a valuable relationship to their aggressors
Summary
Sociality is thought to have driven the evolution of cognitive abilities in primates [1,2,3,4] and other mammals (e.g. cetaceans [5], carnivores and ungulates [6]), and possibly in birds like parrots and corvids [7] (but see [8,9] for an argument against this notion in hyenas). Aside from age, sex, rank and bonding status, victims’ signalling behaviour could be altered in response to the size of the audience, or the behaviour of valuable social partners in close proximity. We first examined the factors sex, age class, bonding status, rank difference of victims and aggressors and the total number of birds present for possible multi-collinearity by calculating variance inflation factors using the HIGHSTATLIB v. The model investigating the effects of audience composition on call rates included the four binomial factors absence/presence of kin of victims, absence/presence of kin of aggressors, absence/ presence of bonding partners of victims, absence/presence of bonding partners of the aggressors, rank difference of opponents and the total number of birds present. Post hoc Mann – Whitney U-tests were calculated on estimated mean values derived from the final model
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
