Back

Vocal Interactions Between Singing Humpback Whales (Megaptera novaeangliae)

Bruno, J. H.; Mercado, E.; Tietjen, N.; Levin, I.; Ryan, M.

2025-06-12 animal behavior and cognition
10.1101/2025.06.12.658025 bioRxiv
Show abstract

Humpback whales construct predictably patterned sequences within multi-hour sessions of sound production. These sequences (called "songs") are some of the most structurally and acoustically complex vocal patterns produced by any mammal. Humpback whale song production is highly dynamic in that individuals constantly modify the properties of songs across multiple time scales throughout their adult lives. Past analyses of co-vocalizing humpback whales suggest that singers within earshot of one another may dynamically and interactively adjust their vocalizations in reaction to what they hear. The current study investigated whether singing humpback whales coordinate their production of overlapping sound patterns either by modulating how long they produce similar patterns or by dynamically adjusting the spectral properties of individual sounds and/or sound sequences. Our results show that the vocal adjustments singing humpback whales make when singing in dyads are diverse and appear to depend on the acoustic context. These findings confirm that individual singers flexibly adjust song features in real-time. When pairs of singers adjust song content, they do so in ways that may increase or decrease the acoustic similarity of concurrent sound patterns. No evidence was found of singers adjusting the number or timing of sound pattern repetitions to alter pattern overlap. However, most singers modified how they produced units after a second singer began singing. The current results suggest that co-singing humpback whales attend to the vocal actions of other audible singers and may sometimes modulate song content to initiate social interactions from long distances. Understanding how singing whales vocally interact can clarify the relevance of singing in coordinating vertebrate behavior.

Matching journals

The top 7 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.