Back

Vocal Similarity And Differences In Four Species Of The Genus Anthus: Implication On Communication And Taxonomic Relatedness

Dragonetti, M.

2022-09-26 animal behavior and cognition
10.1101/2022.09.25.509367 bioRxiv
Show abstract

Birds of the same genus often share similar call repertoires, the aim of this paper is to find species-specific and common acoustic features across species. This can be a useful tool for identification purposes and for studying intra-interspecific communication. Similar flight call (tsip) in two closely related species (Anthus pratensis, Anthus spinoletta) were studied to find characteristics, that allows to discriminate the two species with acoustic means. Three different call types (tsip, soft and alarm) of four species of the genus Anthus (A. pratensis, A. spinoletta, A. petrosus, A. cervinus) were also studied to find whether these common call types show different degrees of similarity. Discriminant function analysis correctly classified 98.4% of A. pratensis and A. spinoletta flight calls. Three acoustic parameters showed the highest discrimination power: the frequency modulations, the maximum frequency value and the minimum frequency value of the peak frequency contour. Using these three values I proposed a simpler procedure for recognizing these two species, that allowed a correct classification of 96% of calls.The three call types of the four Anthus species were studied using cross correlation among spectrogram contours. Alarm calls of the four species showed stronger similarity, while the other call types were more distinctive, with soft call seeming to have a lower similarity between species and hence a higher distinctive power. These results suggest the hypothesis that alarm call is similar, because it retains features of a common ancestor easing heterospecific communication, while the other calls showed decreasing similarity and more species-specific features.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.