Behavioural and physiological responses of domestic dogs (Canis familiaris) to agonistic growls from conspecifics

•We tested dogs’ behavioural and physiological responses to aggressive growl stimuli.•Dogs responded to growls with Proactive, Passive and Neutral behaviours.•Post-test cortisol levels differed between behavioural response categories.•Behaviours reliably predicted physiological stress as measured by salivary cortisol.•This test may provide another tool to aid in shelter dog adoption assessments.

Motivation–structural rule theory predicts that a sender producing harsh, low frequency sounds directed at a conspecific modifies the receiver's behaviour, in part, by communicating its willingness to escalate to an attack. Motivation–structural (MS) rules generally assume that receivers respond to this signal by retreating because of the threat encoded in the acoustic characteristics of the vocalisation. This assumption does not consider if alternative behavioural responses exist or how internal and environmental contexts affect receivers. This becomes an area for potential development of the MS theory when acknowledging that physiological and behavioural reactions may be related to distinct antithetical responses, such as Passive and Proactive coping strategies. To test if aggressive sound stimuli elicit consistent retreat responses by receivers, 42 dogs from shelters and private dog breeders were graded on behaviour and measures of salivary cortisol (a stress related steroid hormone) in response to an agonistic growl from an unknown, similarly sized conspecific. Results revealed that 52.4% (22/42) of dogs displayed retreat behaviours (Passive response), 33.3% (14/42) actively approached the growl source (Proactive response) and 14.3% (6/42) neither approached nor retreated (Neutral response). Post-test cortisol levels differed significantly between dogs in the Proactive, Passive and Neutral categories. Passively responding dogs averaged an 80% change in cortisol levels, Proactive dogs exhibited a 16.5% average cortisol change and Neutral responders displayed only a 6.4% change in cortisol. Although shelter dogs exhibited greater changes in cortisol, they did not differ significantly from the privately-owned dogs. This study verified that overt, observable behavioural responses reflect the physiological stress responses as measured by cortisol changes exhibited by domestic dogs. Results also suggested that experiences may influence how dogs to threatening stimuli from conspecifics respond—behaviourally and physiologically. A protocol providing a simulated threatening interaction with a dog, which does not bring the animals into direct contact, would have important implications for shelter staff in terms of how dog–dog introductions are managed, housing arrangements, potential training, matching dogs to owners and what advice shelter staff could offer to new owners.