Song function in humpback whales (Megaptera novaeangliae): the use of song in the social interactions of singers on migration

Abstract

Male song and other acoustic signals are often sexually selected traits that are common in many taxa, such as birds, anurans and insects. Song often serves a dual function in female attraction and male-male competition and song characteristics such as duration, rate, repertoire size and amplitude are important for both functions and to be correlated with male reproductive success. Male humpback whales are well known for singing a complex and highly stereotyped song during the breeding season and while the song appears important in the social interactions of humpback whales during the breeding season, the function of song is still unclear. Current debate surrounds whether song is primarily directed towards females as an elaborate acoustic display or towards other males to facilitate male social ordering. The aim of this thesis was to investigate the nature of interactions between singers and conspecifics and document the context in which singing occurs. This was achieved by: (1) investigating the movement patterns and interactions between singing and non-singing whales and (2) examining differences in the song structure and source levels of song within a social context. To investigate song function, simultaneous observations of the positions and movements of singing and non-singing whales were obtained in real time during their southward migration off the east coast of Australia in September and October of 2002 - 2004. Data on the interactions of whales were collected by acoustic tracking using a hydrophone array, visual tracking using land-based hilltop observations and observations collected from a small boat. Of a total 114 singers, 48 did not associate with other whales whereas 66 were involved in 63 associations. Singers were significantly more likely to join groups containing a female and calf than any other group type and the only groups with which singers started to sing after joining were unescorted mother-calf pairs. Singers also associated for longer and sang for a significantly greater proportion of time in the presence of a female-calf pair than any other group type. Previous studies demonstrate that associating with females with a calf can be a reproductively successful strategy for males. In contrast, whales that joined singers were usually lone males, these associations were brief and singers typically stopped singing in the presence of other males. These findings provide the highest reported incidence in humpback whales of males singing when escorting females and support an intersexual function of song in humpback whales. A suggested explanation for observations of males joining singers is that these males are prospecting for females rather than engaging in male social ordering and that singing may incur the cost of attracting competing males. To investigate the importance of surface active behaviours in interactions involving singers and whether singers utilize the acoustic cues from these behaviours to locate other groups, the frequency and type of surface active behaviours were quantified in singing and non-singing groups. Compared to singers that were alone and did not interact with other whales, surrounding groups were significantly closer to a singer (on average 2.8 km distance) when singers joined other groups, and singers joined other groups that were surface active in 54 % of cases. While source levels of different surface active behaviours vary and certain behaviours might be better suited for inter-group communication, it is unlikely that singers rely primarily on acoustic cues from surface active behaviours to locate other groups. Surface activity was not more common in groups just prior to singers joining them, and while not significantly so, surface activity increased when singers affiliated with other groups suggesting an importance in intra-group dynamics. Whereas lone singers were the least surface active, groups containing a calf were found to exhibit a significantly higher rate of surface activity than groups without a calf, for which the calf displayed the majority of behaviours. Considering singers predominantly joined mother-calf pairs, it could be speculated that a high rate of surface activity by a calf might inadvertently attract singers and other surrounding males to these groups and could increase the probability of singers and other males locating mother-calf groups. To further explore a context of singing, the structure and amplitude of songs were investigated over two years for singers in two distinct social contexts; lone singers and singing escorts. A total of 274 songs from 27 singers in 2003 and 2004 were analysed, with a subset of data used for source level comparisons. Lone singers consistently sang longer songs compared to singing escorts due to a greater repetition of phrases most noticeably within one theme (theme C), the theme in which all singers sang the most common song unit (the moan) at higher amplitude compared to the other three themes in which it occurred. These findings suggest that lone singers might increase the repetition of phrases sung at higher amplitude to maximise the broadcast of their song and maximise their probability of detection. In contrast, singing escorts sang shorter songs due to a reduced repetition of phrases in most themes, particularly the theme containing the loudest song units (theme C). Consequently, singing escorts sang the song more quickly, resulting in a higher song rate compared to lone singers. Comparisons of the average broadband source levels of nine song units in the 2004 song showed significant variation among the nine different units and among the 22 individual singers for each song unit. Source levels of one of the nine song units also varied significantly between lone singers and singing escorts. Singing escorts produced higher source levels than lone singers for the ‘cry’ which showed a difference of 4 dB higher than lone singers. Lone singers typically produced higher source levels for the ‘moans’, which were low frequency sounds that propagate well, although this was not significant. We suggest that amplitude of the song might be important in female assessment of singers and the particular song units that singing escorts sing more loudly are the basis for this assessment. These findings further support an inter-sexual function of song in humpback whales. This study has described in detail the contexts in which males are observed to sing and has been able to provide new interpretations on the social interactions of singers. The results within this thesis provide the highest reported occurrence in humpback whales of males singing in the presence of females and strongly suggest that song has an inter-sexual function. Further work is needed that investigates song structure within a social context and the potential for song functioning as a long distance advertisement display and/ or courtship display.