Humpback whale song is arranged in a nested hierarchy with multiple levels [30,39]. At the lowest level, each vocalization (defined as the shortest continuous sound to our ears) is called a ‘unit’ (figure 1). A few units are combined to form a string of units called a ‘phrase’. Multiple repeating phrases create a ‘theme’, and several themes sung in a specific order form a ‘song’ that typically lasts between five and 30 min [30,40,41]. Repeating songs are called ‘song sessions’ that can extend over several hours [30]. At the highest level, songs are classified into ‘song types’ based on their theme structure and order [18]. Unlike social sounds that are produced by both females and males [42,43], only adult male humpback whales engage in singing [30,39,44]. This highly stereotyped, repetitive and progressively evolving song occurs primarily on winter breeding grounds and during migration [39,44–46]. Humpback whales spend their summers in high latitude feeding grounds, where they forage [47–49]. After summer, humpbacks undertake long migrations to low latitude winter breeding grounds to rest, mate and give birth [39,48,50]. The function of song remains contested, although it is thought to play an important role in male breeding success [51]. Whether it serves as a courtship behaviour targeted toward females [46], to mediate male–male interaction [51–53], or as a multi-message signal, remains unclear [51,54].
Figure 1.
In the South Pacific region, changes to humpback whale song may occur in two ways: progressive cultural evolution, and cultural ‘revolution’ [41]. Progressive evolution may occur when males embellish their own song through addition, substitution or deletion at any hierarchical level and through copying changes from other males [12,44,55]. Since song is learnt from conspecifics, this individual variation results in the same song slowly evolving over several breeding seasons [39]. By contrast, cultural ‘revolutions’ describe rapid replacement of one song type with an entirely novel one introduced from a neighbouring population, although the underlying mechanism(s) driving this phenomenon remain elusive [41,56]. Noad et al. [56] first observed a cultural revolution of humpback whale song spreading from one population to another in Australia. The eastern Australian population of whales adopted the song previously sung in western Australia in a time span of merely 2 years—a time frame far too brief to be considered progressive evolution [56]. Subsequent work has shown that humpback whale song revolutions continue to spread in a unidirectional pattern eastward from eastern Australia across to French Polynesia [41,57,58]. One song type described in eastern Australia in 2003 was tracked across multiple genetically distinct breeding populations and arrived in French Polynesia in 2005, with this pattern of transmission consistent for several different song types [41,56,59]. In some rare occasions, hybrid songs containing themes from both old and new song types were recorded during seasons of song revolutions [60]. Such instances of hybrid songs were suggested to be active song learning by a whale that is switching from singing the old song type to the new one [60]. Revolutionary songs appear simpler (less complex in terms of number of unit types, length and number of themes) than the songs they replace [61]. The underlying mechanism of transmission remains unclear, particularly given the genetic distinction between populations and strong site fidelity [59,62,63]. However, the following mechanisms may allow song transmission [55]: individuals moving from one breeding population to another (1) within or (2) between seasons may introduce new songs, or (3) song may be exchanged on shared or partially shared migration routes as well as summer feeding grounds (e.g. around Antarctica) [55,57,60,64]. Recently, a location in the western South Pacific—the Kermadec Islands—was suggested to have served as a stopover for humpback whales from multiple populations during their southward migration [58]. Similarities in song themes linked the Kermadec Islands to several wintering grounds, and provided the first indications of where cultural transmission between otherwise acoustically isolated populations on the winter breeding grounds may take place [58] (as song transmits for tens of kilometres [65,66]).
The vocal connectivity of western and central South Pacific populations, and the consistent, unidirectional spread of humpback whale songs from western Australia to French Polynesia raise the question of how far a song may be transmitted eastwards [41,58,67,68]. The distance from western Australia to French Polynesia spans just under 10 000 km, with multiple ‘populations’ of whales (specifically, eastern Australia, New Caledonia, Tonga, American Samoa and the Cook Islands) located in between [41,69]. However, the next major population of humpbacks eastward of French Polynesia migrates along the west coast of South America to breeding grounds located primarily off Ecuador and Colombia, about 8000 km away [69] (figure 2a). Here, we investigated whether song revolutions continue to spread in a unidirectional pattern eastward from the central South Pacific (French Polynesia) into the eastern South Pacific (Ecuador) region. Using similarity analyses (Levenshtein distance similarity index (LSI) and Dice's similarity index (DSI)) songs from three consecutive and concurrent years (2016–2018) were quantitatively compared to investigate the direction and strength of connection. The geographical bounds of song transmission are unknown but are hypothesized to be circumpolar, based on recent cultural evolution models [70]. This current paper employing empirical data provides the first understanding of connectivity across the entire South Pacific Ocean basin.
Figure 2.
