Bird vocalizations are critical cues in social interactions as they convey temporary information varying with the social context, e.g. the signaler motivation when facing a rival or a potential mate. To date, literature mainly focuses on learning birds. Burrowing petrels (Procellariidae) are non-learning birds with a limited vocal repertoire. Bachelor males communicate with conspecifics with a single call emitted in three situations: in absence of a certain auditory (spontaneous calls), towards females (female-directed calls) and toward males (male-directed calls). We first hypothesized that, although the call structure is preserved, temporal and spectral parameters vary between the three call types of bachelor males, translating different motivations (Motivation Hypothesis). To go further, we hypothesized that acoustic variations in male-directed calls indicate the signaler’s aggressive motivation and therefore the variations are similar whether calls are produced by breeder or bachelor males (Breeding-Status Hypothesis). We tested the two hypotheses performing field playback experiments on two petrel species: the blue petrel (Halobaena caerulea) and the Antarctic prion (Pachyptila desolata). Despite the obvious call stereotypy, we observed temporal variations and frequency shifts when males react to a female or a male, which may translate the sexual or aggressive motivation of the signaler. Furthermore, the similarity of variations in male-directed calls of both breeder and bachelor males suggests the aggressive motivation. So far, vocal plasticity in non-learning birds have been greatly underestimated. Here, we highlighted the expression of different motivations through vocal variations and the ability to produce frequency variations in species with genetically coded vocalizations.

To compare the acoustic structure of male calls according to the social situation and to their breeding status, we recorded the three call types (spontaneous, female-directed, and male-directed) from males of two burrowing petrel species, the blue petrel Halobaena caerulea and the Antarctic prion Pachyptila desolata. Records were made using field playback experiments on a small sub-Antarctic island within the Kerguelen Archipelaguo (Ile Verte, 49°51′ S, 70°05′ E) during the birds' breeding season in 2013, 2017, 2018. In total, we obtained N = 678 phrases of blue petrels (117 on breeders and 561 on bachelors) and 428 phrases of Antarctic prions (71 on breeders and 357 on bachelors). 

To describe and compare the call structure, we measured 34 acoustic parameters in both temporal and spectral domains at two syntactic levels (syllables and phrases) on recorded calls using the software Avisoft-SASLab Pro v5.2.11. To increase the accuracy of frequency measurements, we preliminarily downsampled recorded calls from 44.1 to 11.025 kHz using the automatic “Sampling Frequency Conversion” functionality of Avisoft. We also high-passed filtered (cutting frequency: 0.10 kHz, FFT filter) recorded calls to remove low-frequency background noise, which does not affect recordings. We counted the number of syllables and phrases on sonograms. We did not analyze syllables C2 to Cn as they are not present in all individuals. We automatically extracted temporal variables on the amplitude envelopes using the software functionality “Pulse Train Analysis”, with a resolution of 11.6 ms. This functionality automatically measures the temporal structure of waveforms using pulse recognition algorithms. We automatically extracted fundamental frequencies and variables describing energy spectral distribution on linear amplitude spectrum with a resolution of 22 Hz. Automatic extractions were based on similar pre-sets for all birds to ensure replicable measurements.