First insights into the vocal repertoire of Etruscan shrews

Affiliation of presenting author: University of Veterinary Medicine Hannover Vocal communication in animals serves the coordination of social interactions, with different call types emitted in different contexts. An acoustical description of these call types is catalogued in a vocal repertoire. One method to establish a vocal repertoire is the unsupervised cluster analysis, whereby calls are grouped together based on their acoustical properties. Aim of this study is to utilise unsupervised cluster analysis to describe the vocal repertoire of adult Etruscan shrews (Suncus etruscus), the smallest extant mammal suggested to be an interesting animal model for hearing research. We conducted 47 social confrontation experiments with 33 Etruscan shrews, whereby two animals at a time were confronting each other. The composition of the dyads differed in sex, familiarity and housing condition. The experiments were video- and audiorecorded. Temporal and spectral acoustic parameters of the vocalisations were measured. Based on these acoustic measurements, a two-level unsupervised cluster analysis was conducted, using both hard and soft clustering approaches, while dimensionality reduction techniques were also utilised. For the behavioural context analysis, a video analysis was carried out. Finally, influencing factors on the call rates were tested with the help of linear mixed models. Four call types (“screech”, “screams”, “chirp” and “very short tonal”) and a transition between the “screech” and the “scream” were described based on the results of the unsupervised cluster analysis. The “chirps” and the “very short tonal” were mainly observed during socio-positive behaviour, while the “screeches”, the “screams” and the “screech – scream” transitions were emitted during agonistic interactions. Socio-positive calls occurred mainly in male – female dyads. In addition, an effect of the housing condition on the call rate of socio-negative calls was observed, with the animals housed in same sex groups emitting significantly more socio-negative calls. To conclude, four call types and a transition as well as their behavioural context were described in this study. Nevertheless, additional research is required, since confrontation experiments between unfamiliar animals favoured the occurrence of agonistic interactions. Therefore, recordings in the housing cages are advisable, as they would allow the documentation of a more natural behaviour and possibly additional call types.