Discussion
The singing behaviour of territorial male Bermuda Vireos was related to their breeding activities. Although the Bermuda Vireo can be heard year-round, our study suggests that males become more conspicuous during the breeding season by ascending to higher song perches, before returning to their usual haunts in the understory vegetation for the remainder of the year. Males used discrete songs more extensively than rambling songs year-round, and discrete song rate did not differ between the breeding and non-breeding seasons. During the breeding season, however, males with nesting duties sang fewer songs than males without nesting duties, and males caring for nestlings produced the fewest songs of all. Song perch height was higher during the breeding season than during the non-breeding season, for breeding males without nesting duties than for breeding males with nesting duties, and when males sang discrete songs rather than rambling songs.
White-eyed Vireos, like several other species in the genus Vireo , have a species-typical song and a longer, faster, run-on song (Lawrence 1953; Graber 1961; Nolan 1960; Nolan 1962; James 1978; Bradley 1980; Robinson 1981; Gomez-Montes and Moreno 2008; Hedley 2016); for V. griesus , the former and latter song types are known as “discrete songs” and “rambling songs,” respectively. That the production of discrete song did not differ between the breeding and non-breeding seasons suggests that discrete song functions, at least partially, in year-round territory defence. Similar patterns of song production have been described for non-migratory tropical birds that also defend year-round territories (Tobias et al. 2016). The rambling song was generally rare, as it is in continental White-eyed Vireos (Bradley 1981), and we were unable to compare its rate of production between seasons. Nevertheless, several observations support a territorial defence function for rambling song. It is produced in the non-breeding season and, compared to discrete songs, is produced lower in the canopy, where male-male interactions typically occur (Liu 2004). Although anecdotal, Bermuda Vireos in our study sang rambling songs during several close-quarter countersigning exchanges with neighbouring males. We note, however, that one male also directed rambling song towards a female moments before copulating with her, suggesting that rambling songs might also function in a breeding context.
Our findings also provide evidence that Bermuda Vireo song functions in mate attraction. During the breeding season, males without nesting duties usually were unaccompanied by a female and spent this time singing discrete songs at a high rate, whereas males with nesting duties were most often accompanied by a female and sang significantly fewer discrete songs. Similar singing patterns have been described in Bell’s Vireos (Vireo bellii ; Nolan 1960), Yellow-throated Vireos (James 1984), and Warbling Vireos (V. gilvus ; Howes-Jones 1985), and for several avian taxa beyond the Vireonidae (Powlesland 1983, Hayes et al. 1986, Staicer et al. 2006, Foote et al. 2017). Our findings that song perch height was higher during the breeding season than during the non-breeding season, and higher among breeding males without nesting duties than among breeding males with nesting duties, provides further support that Bermuda Vireo discrete song functions to attract mates. Males with no nesting duties performed lengthy song bouts of discrete song whilst remaining stationary on higher branches in the tree crown, before repeating this behaviour at another elevated and frequently visited perch in the territory. These behaviours have also been described for unmated males in Blue-headed Vireos, Yellow-throated Vireos, South Island Robins (Petroica australis ), and Chipping Sparrows (James 1978; Powlesland 1983, Liu and Kroodsma 2007); these researchers suggest that singing from elevated perches increases an unmated male’s visual and acoustic conspicuousness to prospecting females. Three anecdotal observations provide further support that singing from elevated perches is a mechanism for attracting prospective females: (1) breeding pairs travelled primarily in the understory, (2) nests were never built in the canopy, but, rather, from forked branches, usually 2–3 meters above the ground, and (3) males often returned to canopy perches after their mate disappeared, typically following nest failure. The tendency of males to select higher perches when singing discrete song versus rambling song might be because discrete song is louder than rambling song (Bradley 1980); the combination of being louder and being sung from higher perches may reflect a history of selection for maximizing signal transmission distance (Sprau et al. 2012, Podos and Sung 2020).
Most of the nest predation in our study occurred during the nestling care stage, which is consistent with the predictions of the nesting stage hypothesis (Morton et al. 1993, Burhans et al. 2002). Given the increased risk of nest predation during nestling care, and the fact that songs can alert predators to nearby nests (Ellison and Ydenberg 2019), it is perhaps not surprising that Bermuda Vireos produced the fewest discrete songs during this time. Similar declines in vocal activity during nestling care, compared to other stages of the nesting cycle, have been documented in species spanning multiple avian familes, including House Wren (Troglodytes aedon , Wilson and Bart 1985), Yellow-throated and Blue-headed Vireo (James 1999), Golden-cheeked Warbler (Bolsinger 2000), Chipping Sparrow (Liu and Kroodsma 2007), and Common Reed Bunting (Emberiza schoeniclus , Brunner and Pasinelli, 2010). Our observation of song resurgence during the fledgling stage (Figure 4) appeared to occur because males often used discrete song to guide fledglings throughout the natal territories and to bring them close after securing a food item. There have been no observations of predation of adult or fledgling Bermuda Vireos, which spend most of their time in thicket understories. Thus, in addition to its established role in passerine song learning (Nowicki et al. 1998), the increased singing rate of males during fledgling care possibly reflects this low predation risk, further supporting the nesting stage hypothesis.
Future research should attempt to further distinguish the functions of the two main song types used by Vireo species, as has been done in the two-category singing system of North American parulids (Spector 1992). The many observational studies that preceded our work not only brought to light the ubiquitous nature of the two-category vireonid song system, but also provide a list of vireonid species that can serve as candidates for hypothesis testing. Future research could also test the effects of feather moult on singing rate. In Bermuda, August marks the peak of feather moult in vireos (M. Mejías pers. obvs.), when most adults were seen hastily feeding while missing some or all their tail feathers. Feather moult is one of the most energetically expensive and time-consuming life stages in birds (Rohwer et al. 2009, Kulaszewicz and Jakubas, 2015), and could also explain the decline of song in August, with males prioritizing intensive foraging over vocalizing, as observed in moulting Blue-headed and Yellow-throated Vireos (James 1999).
In conclusion, we found that male Bermuda Vireos are year-round singers that alter their singing behaviour in relation to breeding stage. Our results provide support for the territory defence and mate attraction hypotheses of passerine song, and suggest that nesting birds reduce their production of conspicuous songs during nestling care when their nests are at the greatest risk of predation.