Unique rhythmic signatures in arrhythmic birdsong

Abstract Animal rhythms are gaining increasing attention in the studies of behaviour and musicology. Recently, it has been shown that rhythm itself can be used as an information coding channel. Now we ask: does this hold true for arrhythmic sequences? To answer that, we analysed songs of the tawny pipits ( Anthus campestris ), migratory songbirds known for their simple songs. Using focal recordings of 384 individuals from six populations collected across their European breeding range, we calculated an extensive set of rhythmic and temporal indices to describe each song. First, the pattern of these songs was inspected and shown to be arrhythmic. Second, songs belonging to specific individuals and populations were compared using permuted discriminant function analysis and supervised uniform manifold approximation and projection. To describe the level of individuality of the rhythmic structure alone, we calculated Beecher’s statistic for all songs, as well as the potential of identity coding for each parameter separately. We show that tawny pipit males sing using individual rhythmic patterns in their arrhythmic songs, and that modern rhythmic indices, such as beat precision and integer ratios, are among some of the most individually distinct parameters of their songs. Furthermore, contrary to previous investigations based on the spectral shape and basic frequency and temporal characteristics of these songs, we show that the species displays population-specific temporal patterns, with significant differences throughout its European range. This study is the first to demonstrate geographic scale differences in birdsong rhythm, and to show that rhythmic analysis can provide useful descriptions of arrhythmic sequences.