Special Collection on Bird Song

47 Open Access articles from The Auk: Ornithological Advances and The Condor: Ornithological Applications, plus 3 other articles

What Can We Learn from Bird Song? Recent Advances in Functional and Applied Avian Bioacoustic Research

Shelby Lawson and Mark E. Hauber
Department of Animal Biology, School of Integrative Biology, University of Illinois, Urbana-Champaign, IL USA

A major area of ongoing interest in ornithology is the development, production, and function of avian vocalizations and sonations. By investigating songs and calls—their acoustic structure, their production and perception by conspecifics, and the context in which they are produced—vocal communication research in birds increases our understanding of the behavioral ecology and evolution of communication systems. Recent research published in The Auk: Ornithological Advances and The Condor: Ornithological Applications on basic and applied aspects of the sounds produced by birds has advanced our knowledge of the subject and will influence the direction of future research in the field. Here, we overview the literature on avian vocalizations that has been published in The Auk and The Condor over the past 5 years—to better understand the focus, approach, and direction of current research trends in vocal communication, and to generate suggestions for areas of research that need more attention.

As of 2014, we have refocused our two journals: The Auk on expanding our knowledge of fundamental biological concepts and avian species, The Condor on the application of ornithological concepts to promote better ecological management and conservation of species. When these aims are applied to vocal communication, we find that research in The Auk has tended to explore ultimate, evolutionary questions about the structure, function, and behavioral ecology of vocalizations. By contrast, research in The Condor has probed proximate questions about detectability, conspecific attraction, and species-specific behaviors, all of which may be incorporated into improved conservation and population-management strategies for the focal species.

Among the many publications in these two journals over the past 5 to 6 years, only a small but consistent fraction involves study of vocal communication or sound production, yet the field is well represented here, with a good range of topics and functions covered. Moreover, the methodologies are diverse in both journals, and there is a good amount of technical overlap between them, which makes comparisons feasible between basic and applied questions and studies. We have formatted our overview to begin with papers on the structure and production of sound signals, then shift to sound in the environment, and end with a look at behaviors of receivers when they hear other birds’ vocalizations.

Structure of Acoustic Signals

The acoustic features that distinguish each species’ vocalizations are important not only for classifying species-specific calls and songs, but also for defining variations in these vocalizations within and across populations. Over the past few years, two specific studies in these journals sought to analyze song repertoires and displays given by species whose vocalizations had not been well studied. The first study recorded Canyon Wren (Catherpes mexicanus) vocalizations and found that both males and females sing, but that the frequency of time spent singing and the variability of song type were greater in males than in females. Additionally, song types were shared between males, which suggests a high degree of song learning between wrens that may be enhanced by certain ecological and life-history traits of the Canyon Wren (Benedict et al. 2013). The second study examined the courtship displays of 3 closely related South American “bee” hummingbird species by placing female conspecifics in a cage to solicit displays from different males. The 3 species—Chilean Woodstar (Eulidia yarrellii), Peruvian Sheartail (Thaumastura cora), and Oasis Hummingbird (Rhodopis vesper)—varied in the complexity of their songs and displays, as well as in the compexity of the mechanical sounds produced during dives, which, in turn, were related to feather morphology (Clark et al. 2013).

Figure 1. Examples of 2 Canyon Wren song types recorded among 6 locations within the range of the species (gray area). (figure 3 in Benedict et al. 2013) Canyon Wren

Song data, along with displays, morphology, and genetics, can also be used in some cases to justify the systematic separation of species that had formally been grouped together. For example, analyses of acoustic variation and morphological measurements of 8 subspecies of the House Wren (Troglodytes aedon) revealed that at least 4 of the subspecies show disparities in song and/or morphology that are greater than those in other pairs that are recognized as separate, full species (Sosa-López and Mennill 2014). Similarly, the songs and calls of 2 subspecies of a North American “bee” hummingbird, the Bahama Woodstar (Calliphlox evelynae), were found to be distinct and individually recognizable, possibly meriting species status for the subspecies (Feo et al. 2015). On the other hand, 2 other pairs of “bee” hummingbird species were found to be producing hybrids with songs and displays intermediate to those of the parent species, which could be a critical indicator of hybridization of the species pairs (Clark et al. 2012).

Figure 2. Recording locations of subspecies of House Wren and other wren species. Open circles represent recordings from the 8 subspecies of House Wren whose song was analyzed. Examples of song from each of these are shown in spectrograms on the right. Recordings from other species of wren are shown in shaded symbols and were used as a comparison. Spectrograms of song from these species are shown on the left. Dark gray areas show the distribution of House Wrens during the breeding season, and dashed lines mark the geographic boundaries of the different subspecies. (figure 1 in Sosa-López and Mennill 2014) House Wren

Notably, variation in acoustic song and call structure within a species does not necessarily indicate that its species status needs to be reevaluated. Populations of birds can exhibit diversity within their vocalizations across geographically distinct ranges, which are referred to as “dialects.” The cause of dialects can range from differences in song-learning ontogeny and ecological pressures within populations to isolation of groups that prevents sharing of song types. For instance, Golden-crowned Sparrows (Zonotrichia atricapilla) occupy and breed in territories in montane western North America, across which 13 distinct song types were found to be used. However, most birds use only 1 of the 5 song types that are most common species-wide, which indicates that song sharing is negatively correlated with migration distances in this species (Shizuka et al. 2016). As seen above, House Wren subspecies also display dialects across geographic ranges, to the extent that species status may be warranted (Rendall and Kaluthota 2013).

Populations that become more isolated from each other, geographically and/or genetically, also may undergo song divergence. An example is the Common Yellowthroat (Geothlypis trichas), in which the acoustic features of song notes vary between its eastern, western, and southwestern ranges. The variation in habitats among those ranges likely create a drive for the evolution of songs that transmit best in a particular habitat (Bolus 2014). In turn, flight calls of Red Crossbills (Loxia curvirostra) on Newfoundland, Canada, were found to be distinguishable from U.S. conspecifics, with alarm and chitter calls differing as well (Hynes and Miller 2014).

Finally, dialects are usually seen in species that learn song, because this allows for a cultural transmission of the learned song. However, acoustic variation across a large geographic range was detected in the calls of subspecies of Barred Owls (Strix varia), which do not learn their vocalizations. This variability was uncorrelated with geographic distance or genetic introgression between subspecies, though, indicating that further study is needed to better understand the evolutionary and developmental origins of this phenomenon (Odom and Mennill 2012).

Song divergence within and between populations can ultimately lead to speciation through behavioral isolation, and song divergence can appear more quickly and initiate speciation nearly instantaneously (Lamichhaney et al. 2018). Nevertheless, longitudinal studies are essential for measuring discrepancies in vocalizations between and within populations over time to better understand this process. No study in either journal in the last several years has used song data collected continuously for a length of time, but a handful have compared data from time-points years apart, as a way to analyze the stability of the population’s song after many years. For instance, the geographic variation in dialects of Golden-crowned Sparrows was recorded 15 years apart, and interestingly, the most commonly used song types remained stable over time, although their geographic distribution has changed (Shizuka et al. 2016). Similarly, the structural components of various dialects in a population of Brown-headed Cowbirds (Molothrus ater) have also remained stable over a 30-year period, only showing an increase in one parameter, the whistle duration (O’Loghlen et al. 2013). In contrast to the stability seen in these 2 species, the Hermit Warbler’s (Setophaga occidentalis) type 1 dialect, measured at 4 time-points during a 14-year span, underwent large changes, particularly in the introductory notes, and exhibited a loss of the terminal syllable (Janes and Ryker 2013). The varying ecological traits and life histories of these species are likely responsible for differences in song stability. For instance, juvenile Hermit Warblers may have experienced copying errors when learning song from adults, and high rates of fidelity and survivorship would have contributed to the spread of this new song throughout the population.

Production of Acoustic Signals

For consistent regional variations in acoustic structure to be possible, birds must have mechanisms in place to produce their signals correctly and consistently in the local environment. This capability can be maintained or modulated by the physical structures for producing sound, the neural mechanisms that act on these physical structures, and the internal state of the animal. The syrinx, the vocal organ in birds, is an important and necessary structure for song production. Unsurprisingly, there is a great amount of focus in research on the morphology of the syrinx in different species and the mechanisms that regulate song output. For instance, Christensen et al. (2017) examined muscle-fiber composition and size of syringial muscles across 10 avian species and found that both male and female syrinxes contained “fast” and “superfast” fibers used for quick contractions, even in species in which females did not sing. A recent study found that singing behavior is likely ancestral in birds, for both sexes (Odom et al. 2014), so weak but continued selection pressure on female fiber composition may still explain the lack of dimorphism in vocally dimorphic species. Accordingly, females in some of the species tested may also utilize fast and superfast fibers in complex calls, even though they do not sing, but this needs further research.

In another study, the functional morphology of the syrinx in Pectoral Sandpipers (Calidris melanotos) and the synchronization of their wing movements and hoots were explored to determine how their courtship sounds are produced. The hoot sound, in particular, was produced through a closed beak during flight, demonstrating the beak’s role in sound output (Riede et al. 2015). In turn, Common Yellowthroats with smaller beaks were shown to sing at lower frequencies, thereby affecting transmission of song through the habitat and possibly driving song divergence (Bolus 2014). Variation in other physical structures can also affect outputs of nonvocal acoustic signals, like that of the drumming display in the Ruffed Grouse (Bonasa umbellus), which uses its wings to create a directional sound that is loudest along the longitudinal axis of the bird (Garcia et al. 2012).

Physiologically, circulating hormone levels are known to affect vocal output as well, by modulating the neural song circuits or acting directly on musculoskeletal targets for producing song. For example, inhibition of androgens in Golden-collared Manakins (Manacus vitellinus) altered the duration of the “chee” note and the frequency of the “poo” note in their songs (Fuxjager et al. 2014). In another study, Reichard et al. (2017) recorded courtship behavior and levels of testosterone and cortisol in paired and unpaired male Dark-eyed Juncos (Junco hyemalis) in response to a caged female. Although testosterone levels didn’t vary between paired and unpaired males, cortisol was found to be higher in unpaired males. This may explain the difference in types of song produced and the components of the courtship displays delivered between paired and unpaired males. In a study of Brown-headed Cowbirds, cortisol was correlated positively with song repertoire size, but negatively with immune function. Male cowbirds with larger repertoires generally engaged in more aggressive interactions with other males, and the increased energy expenditure of this activity may have hindered their immune system functions, although further research is needed on this (Merrill et al. 2013).

Figure 3. Song repertoire size of male Brown-headed Cowbirds in relation to their bactericidal capacity against E. coli. (figure 1 in Merrill et al. 2013) Brown-Headed Cowbird

Figure 4. Song repertoire size of male Brown-headed Cowbirds in relation to baseline corticosterone (CORT). (figure 2 in Merrill et al. 2013) Brown-Headed Cowbird

External factors can also influence song, particularly the type of song or the frequency of production. Canada Warblers were found to increase their use of particular song types, depending not only on their pairing status but also on the seasonal phase of the nesting cycle they were in (Demko et al. 2013). Finally, singing frequency by male Song Sparrows (Melospiza melodia) was shown to decrease in high temperatures, but males who had larger bills, and likely had better thermoregulation, were able to continue to sing at higher rates than males with small bills (Luther and Danner 2016).

Sound in the Environment

After a signal has been produced, it must travel through the environment to reach a receiver. Degradation of the mechanical signal as it travels is inevitable, but signalers can minimize this by adjusting their vocalizations or by producing them so as to avoid interference or noise (Laidre 2017). Birds may also use degradation of incoming song to perceive the distance of the singer. This is possible through mirror neurons, which produce memories of a learned song type when birds are singing and when they hear song. To determine its distance from the singer, the receiver can compare its memory of a song type with the incoming song that has been degraded by the environment; further degradation would mean the song had traveled a longer distance (Morton 2012). Understanding how sounds travel through an environment, and the measures birds take to best transmit their vocal signals, is pertinent to researchers for detecting the most vocalizations for population surveys and studies of song. For example, sounds played in a forest attenuate more quickly than sounds played near a road, which has implications for avian point-count surveys. Roadside surveys may be overestimating numbers of birds present, because vocalizations at a distance can travel farther in open areas like roads (Yip et al. 2017). In addition to location, detection probability is affected by timing: Birds are often more vocal during the dawn and dusk. Passive detection probability of locally threatened Golden-winged Warblers (Vermivora chrysoptera) decreased with time elapsed since sunset and with distance from territories. The use of playbacks promoted responses and maintained detection probability throughout the day, but the song types that were played influenced the song-type responses (Aldinger and Wood 2015).

Playbacks may be more useful in some cases for collecting responses than passive listening, so researchers should consider this for further surveys. Automated detection of vocalizations and modeling of the singing activity are also becoming more prominent as technology advances. For example, a Bayesian change-point analysis of male Common Nightingale (Luscinia megarhynchos) vocal activity was used to determine whether the model could predict mating status of the males on the basis of singing behavior. Afterward, birds were captured and pairing status was found to be consistent with the model (Roth et al. 2012).

Recently, more research has been focused on how anthropogenic noise is affecting birds’ ability to send signals effectively. As human presence in and near avian habitats grows, so does the level of noise, which can mask vocal communication signals. For instance, conspecific approaches to playbacks of Tufted Titmouse (Baeolophus bicolor) mobbing calls were severely reduced when masked by anthropogenic noise. This is critical, because mobbing calls bring birds together to cooperatively harass a predator, but if there is a lack of response due to calls being masked by noise, then mobbing a predator can become more life-threatening for individual birds (Damsky and Gall 2017). On the other hand, some avian communities, such as grassland communities, may be less affected by anthropogenic noise. Despite increased noise from a new wind energy facility on the Nebraska Sandhills, the acoustic diversity and species richness seemed unaffected by wind turbines (Raynor et al. 2017). Human-created noise varies in its effect on avian communities as a result of many factors, including the acoustic attributes of the noise and vocalizations being produced, the hearing ability of the species in the community, and the proximity of the community to the source of the noise. Further research is needed on how different communities will be affected by anthropogenic noise. Understanding the many sources of noise in a habitat—from anthropogenic to biological to geophysical—is essential for researchers in defining the soundscape for that habitat. The soundscape contains useful information about the environment’s effect on sounds, the biodiversity within the habitat, and the extent of human disturbance (Gasc et al. 2017).

Figure 5. Spectrogram of a Tufted Titmouse mobbing call (A) and anthropogenic vehicular and air traffic noise (B), both of which were recorded on Vassar Farm and on the Vassar College campus in Poughkeepsie, New York. (C) Spectral density (dB/kHz) of titmouse calls and anthropogenic noise. Calls below 2.2 kHz are masked by noise. At 3 kHz and above, 5 kHz calls have the most energy. (figure 1 in Damsky and Gall 2017) Tufted Titmouse

Functionality of Vocalizations and Other Sound Signals

The result of a successfully transmitted signal is that the sound reaches the intended receiver. The functions of vocalizations and their effects on the receiver are areas of interest in behavioral ecology. Low-amplitude vocalizations and nocturnal vocalizations, though often thought of as uncommon, occur more often than previously believed, and their potential functions are discussed in reviews by Reichard and Welklin (2015) and Van (2012). Information for a diverse array of functions can be encoded in the acoustic features of songs and calls, even in species with small repertoires. Spotted Crakes (Porzana porzana), for example, have a very small repertoire of calls but were found to modulate the temporal distribution of those calls in response to a simulated intrusion. Changes in the calls’ acoustic structure and frequency varied, depending on the level of aggression and other behavior the birds employed toward the playback (Ręk 2015). Although studies on the acoustic structure and composition of songs were common in our sample, these studies did not explore the function of the song, or its effect on conspecifics that were exposed to it. Song is often used for mate attraction or courtship, but the features that the receiver finds salient in determining whether to mate, and whether these features demonstrate fitness, are questions often left unanswered. However, one study explored how male White-crowned Sparrows (Zonotrichia leucophrys) assess competing males and found that territorial responses varied according to the song performance of the intruder, particularly regarding the terminal trill (Phillips and Derryberry 2017).

Acoustic features, if they are an honest signal of fitness, can also be informative for females in choosing a mate. For example, it has been hypothesized that female mate choice acts as a stronger selective pressure in migratory species, and this may have led to the evolution of more complex song compared with sedentary species. However, a review of songs from the genus Geothlypis found that distance of migration was not correlated with song elaborateness (Byers 2015). No studies in the recent years of The Auk and The Condor have explored relationships between song traits and mating success, but one work investigated whether acoustic features of the mechanical swish in Greater Sage-Grouse (Centrocercus urophasianus) were related to mating success. Although no features of the swish were significantly correlated with mating success, more successful males did display larger ranges in swish frequency and adjusted their behavior more when females approached (Koch et al. 2015).

Vocalizations are often intended for conspecific receivers, whereas signals can be eavesdropped upon by heterospecific bird species—which may modify their behaviors in response to these “unintended” signals as well. For example, one study used playbacks of calls of obligate ant-following birds, which specialize in preying on invertebrates that are flushed out by ant armies, to see whether other bird species eavesdrop on this signal to locate food sources. The research found that ant-following birds do take advantage of these calls, but generally only if the playback is from a smaller, subordinate ant-follower species, from which it is likely easier to steal food sources (Batcheller 2017).

Birds may also adjust their behaviors after eavesdropping on predators to minimize the risk of predation for themselves or their offspring. Hooded Warblers were shown to respond differently to 2 different predator calls, with adults quickly returning to nests in response to eastern chipmunk (Tamias striatus) calls but delaying their return after hearing Cooper’s Hawk (Accipiter cooperii) calls. Adults likely hide after hearing hawk calls because hawks prey on the adults, whereas chipmunks prey only on nestlings, so chipmunk calls would prompt adults to return to protect their offspring (Schaef and Mumme 2012). Another threat to nests are brood-parasite birds, which lay their eggs in the nests of heterospecifics, thereby forcing the latter to care for the parasite’s offspring. There is evidence that calling behavior in Bell’s Vireo (Vireo bellii) is negatively correlated with parasitism by the Brown-headed Cowbird, in that birds with nonparasitized nests sang more than birds that had been parasitized. However, further research is needed; it is unclear whether the calls are a signal that decrease parasitism, or whether parasitism exhausts the host, which decreases the call rate (Steckler and Conway 2012).

Vocalizations in some species have features whose function is to provide identifying information on the species or sex of the caller. Species-recognition cues are particularly relevant for parasitic species, which are raised in the nest of a different species. In parasitic African indigobirds (Vidua spp.), mimicry of vocalizations learned from their heterospecific hosts was found to be important for species recognition, both in territorial defense and in directing juveniles to conspecifics after dispersal (DaCosta and Sorenson 2014). Sex-specific identifying features in vocalizations are often overlooked, because in many of the studied species only males produce song, and the acoustic structure of calls that both males and females produce are understudied. However, one study sought to determine whether short- and long-distance calls of White-naped Cranes (Antigone vipio) contained features that could be used to identify the sex of the caller, as well as individual identity. Only one call type contained any sex-specific information, and the lack of individual identifying features may have been because individuals almost always have visual contact with each other, such that there isn’t a strong drive for these identifying features to be encoded in their vocalizations (Bragina and Beme 2013).

Some vocalizations are functionally important for communication between parents and their offspring. Nestlings, for example, use begging vocalizations and other cues to relay their level of need to the parents. Accordingly, food-deprived nestling Bank Swallows (Riparia riparia) became more vocal with their begging and moved toward the tunnel entrance, where parents return. Both of these activities are energy costly and seem to be used only by the hungriest nestlings (Brzęk and Konarzewski 2014). In the Red-backed Fairywren (Malurus melanocephalus), nestlings imitated acoustic features of their mother’s call to secure more parental provisions. Additionally, nestlings that had more exposure to the call, both in the egg and as hatchlings, were better imitators and received more food from the mother (Colombelli-Négrel et al. 2016). Besides calling, parents in some species, such as Tree Swallows (Tachycineta bicolor), use alarm calls to signal their nestlings to be quiet to minimize the risk of predation. This quieting response is mainly seen in older nestlings, which beg even when their parent is absent, unlike younger nestlings, which beg only in response to the parental food call (McIntyre et al. 2014).

Much of the basic research on vocal functionality is published in The Auk. However, there is one function of bird song that is heavily covered in The Condor: conspecific attraction. Birds use social information such as number or density of birds vocalizing in an area to determine what habitats other birds have chosen fit to breed in. Given that The Condor is the journal for applied conservation ornithology, it is unsurprising that a major focus would be researchers utilizing conspecific attraction to draw birds, especially threatened taxa, to existing, new, or restored habitats. Broadcasting conspecific song in suitable areas of the Florida Everglades influenced male Cape Sable Seaside Sparrows (Ammodramus maritimus mirabilis; an endangered subspecies) to establish territories in those areas (Virzi et al. 2012). In another study, playbacks of conspecific calls, as well as odors, were tested for the ability to attract Leach’s Storm-Petrels (Oceanodroma leucorhoa) and Fork-tailed Storm-Petrels (O. furcata) to newly reclaimed land. Prebreeding individuals of both species were attracted to the conspecific calls, and, interestingly, conspecific odors (feather, nesting material, stomach) attracted Fork-tailed Storm-Petrels in particular (Buxton and Jones 2012). The saliency of these cues for different species may vary throughout the year, and specifically during the breeding cycle if they are used to choose breeding grounds. Thus, some studies have examined behaviors in response to conspecific cues throughout the breeding cycle. Playback of Grasshopper Sparrow (Ammodramus savannarum) song did not influence colonization during the early breeding season, but it doubled the number of sparrows present at the playback sites in the later season, which suggests that Grasshopper Sparrows may use number of birds in the late season as an indicator of reproductive success for that habitat (Andrews et al. 2015). In some instances, however, conspecific presence doesn’t result in increased colonization of an area, and more research is needed to tailor conservation approaches to such populations, for example in the Golden-winged Warbler (Vermivora chrysoptera; Albrecht-Mallinger and Bulluck 2016). One study even found that playback had a repelling effect on colonization of Chimney Swifts (Chaetura pelagica) later in the breeding season. Juveniles looking to breed likely used number of vocalizations as a cue for population density and avoided areas with playback that were seen as overpopulated (Finity and Nocera 2012).

Both The Auk and The Condor have covered an extensive array of research related to avian vocalizations, but there are a few areas that need more attention and present an opportunity for future exploration. At the basic level, the methodologies of most of the studies discussed here rely on naturalistic observations and recording of acoustic signals and don’t involve any manipulations of subjects or their surroundings. Song research conducted in a lab is virtually nonexistent in these journals, yet it would provide a great opportunity to test ideas within a controlled environment. Another area of research that could be given more attention in a lab setting is the neurobiology of song, including its production and perception. Despite growing interest in the subject and its importance in understanding avian vocalizations, studies using neurobiological techniques are hardly represented in The Auk and The Condor, even though technological advances now include nonlethal and minimally invasive methods of neuroethological investigations of functionally relevant vocalizations (e.g., Louder et al. 2016).

There is a wealth of studies on conspecific attraction that are incredibly useful for conservation work, but future studies on this topic could be improved by documenting certain measures that are less standardized in other studies. Quantifying densities of birds before treatment and comparing the habitat suitability and quality at different playback areas would be useful for better understanding how these variables may be related to response to playbacks. The timing of the breeding cycle also likely influences the effectiveness of conspecific attraction for many species, but often it is not explored throughout the full breeding season, but rather during the early stages only.

Song research continues its historical bias toward song production by males, even though both males and females of many species produce song. Only one study in our sample examined female song production. Finally, there is room for more work on the acoustic features of song (of either sex) that are particularly salient to either competitors or potential mates, because those features relay information about the fitness or biological state of the producer. Acoustic research is often focused on birds’ production of signals and its function, but understanding how a signal’s various acoustic features are perceived is crucial for determining how the signal is interpreted by its receivers.