Seasonal changes in the nocturnal singing routines of Common Nightingales Luscinia megarhynchos

Male Common Nightingales Luscinia megarhynchos famously sing at night. There are two distinct types of nocturnal singing routine (the dusk-to-dawn pattern of variation in song rate): (1) dusk and dawn choruses, with little or no song during the middle of the night; (2) a rapid increase in song rate after dusk, reaching a broad peak in the middle of the night, declining towards dawn, and followed by a dawn chorus. Males sing different nocturnal singing routines at different times in the breeding season. Earlier in the breeding season, most males sing Type 2 routines. Later in the breeding season, most males sing Type 1 routines, as do birds on the wintering grounds. At least some individuals may also sing Type 1 routines during the first few days after their arrival on their breeding territories, before the arrival of females. The main function of nocturnal song appears to be mate attraction. The patterns of variation in song rate over the course of the night are qualitatively similar to those predicted by stochastic dynamic programming (SDP) models of daily singing and foraging routines, for birds that do not forage at night, in circumstances when birds can pair at night (Type 2 routines), and when they cannot (Type 1 routines). The observed seasonal changes in the types of routine sung are also consistent with the predictions of these models.     

The effect of variability in the food supply on the daily singing routines of European robins: a test of a stochastic dynamic programming model

A stochastic dynamic programming (SDP) model offers a general explanation of daily singing routines in birds, but remains almost untested empirically. I examined a central prediction of the SDP model, that a more variable food supply decreases the bird's song output at dawn, relative to its song output at dusk. I provided supplementary food to make the food supply more or less variable over 2-week periods in the territories of free-living European robins Erithacus rubecula. Robins sang relatively less at dawn than at dusk after weeks in which their supplementary food supply was variable, and more at dawn than at dusk after weeks in which their food supplementation was constant. These results provide strong support for the prediction of the SDP model. Copyright 1999 The Association for the Study of Animal Behaviour.     

Two explanations of the dawn chorus compared: how monotonically changing light levels favour a short break from singing

I used optimality modelling to compare two of the most plausible and general explanations for the dawn and dusk peaks in bird song output. Kacelnik's explanation is that foraging is inefficient in poor light, but that social interactions are less affected, making singing more worthwhile than foraging. McNamara et al.'s explanation is based on stochasticity in foraging success and overnight energy requirements; it has been extensively analysed with stochastic dynamic programming models. Both explanations are now incorporated into this sort of model. I used various functions to link success of foraging and singing to time of day, but assumed that above some light level there is no further effect. Kacelnik's explanation has as strong an effect as stochasticity in generating dawn and dusk choruses. It also predicts short pauses in the singing output just after the dawn chorus and before the dusk chorus. The former arises because birds delay foraging when it will become more profitable later, until foraging success reaches a plateau, when the energetic debt accumulated makes them forage. The principle of this see-saw double switch in behaviours may apply to other explanations for the dawn chorus, and to other shifts in behaviour when conditions change gradually. The model predicts that from day to day cloud cover determines when a dawn chorus starts, but that overnight temperature and wind strength have more effect on chorus intensity and duration. I discuss what sort of observational and experimental data on singing routines would better test this model. Copyright 2002 The Association for the Study of Animal Behaviour. Published by Elsevier Science Ltd. All rights reserved.     

The relationship between daily routines of singing and foraging: an experiment on captive Great Tits Parus major

Unpaired, singing, male Great Tits Parus major kept in aviaries were subjected to two different daily regimes of foraging profitability. Birds learnt to adjust their foraging patterns to avoid feeding at periods of low foraging profitability. Once learnt, this had no effect on their singing regimes, indicating no direct causal link between the organization of daily regimes of singing and foraging. When first faced with low foraging profitability early in the morning, birds reduced morning song levels but these returned to normal, although no higher than normal, as they learnt to anticipate the daily routine. This suggests that unpredictable periods of low foraging profitability cause reduced singing activity but predictable time of day effects do not. Birds did not spend all their non-foraging time singing and appeared to have 'free time'.     

Strategic diel regulation of body mass in European robins

Stochastic dynamic programming (SDP) is a computational technique that has been used to model daily routines of foraging in small birds. A diurnal bird must build up its fat reserves towards dusk in order to avoid starvation during the night, when it cannot feed. However, as well as the benefits of avoiding starvation, storing fat imposes costs such as an increased predation risk and higher flight and metabolic costs. There is therefore an optimal level of fat reserves for a bird to reach at dusk in order to survive overnight without being left with excessive fat reserves at dawn. I tested a prediction common to all SDP models of daily foraging routines, that a bird will attempt to reach this level at dusk, regardless of its fat reserves the previous dawn. I provided supplementary food to manipulate the fat reserves at dawn of free-living European robins, Erithacus rubecula. Diurnal changes in body mass (a reliable estimate of fat reserves) were then monitored remotely. Robins provided with an ad libitum food supply reached almost exactly the same body mass at dusk, regardless of their body mass at dawn, supporting the prediction that birds attempt to reach a target level of reserves at dusk. Copyright 2000 The Association for the Study of Animal Behaviour.     

Long-term influence of simulated territorial intrusions on dawn and dusk singing in the Winter Wren: spring versus autumn

Males of many songbird species have peaks of singing activity at dawn and dusk. Singing during those twilight periods can function in territory proclamation, and males are suggested to adjust song output to the level of intruder pressure. We used song playback during the breeding season to simulate intrusions into territories of male Winter Wrens (Troglodytes troglodytes) shortly after dawn. We then compared male singing behaviour during the dawn and dusk chorus before and 1 day after the simulated intrusion. One day after the playback, male Wrens increased their song output before sunrise, which confirms our results from a previous study on dawn singing in autumn territories. At dusk, on the evening following the playback, males slightly increased song output after sunset, but singing activity at dusk was generally very low. We found no significant changes of song output after sunrise, before sunset, and between 2 days of control without playback. These results are consistent with the hypothesis that dawn and dusk singing is important for territory defence in spring. Unlike in autumn, however, increased singing in spring at dawn and dusk could also serve to defend other resources such as fertile mates or to strengthen the pair bond after a territorial challenge. In comparison with the results on autumnal singing, male Wrens started singing earlier at dawn during the breeding season, and they generally sang more songs at dawn and immediately after playback. The increase in absolute numbers of songs sung in the morning after playback seemed greater in spring than in autumn; however, the proportional increase relative to overall song output was similar in both seasons.     

Raised thermoregulatory costs at exposed song posts increase the energetic cost of singing for willow warblers Phylloscopus trochilus

Sexually selected displays, such as bird song, are expected to be costly. We examined a novel potential cost to bird song: whether a less favourable microclimate at exposed song posts would be predicted to raise metabolic rate. We measured the microclimate and height at which willow warblers Phylloscopus trochilus sang and foraged. Song posts were higher than foraging sites. The wind speed was 0.6±0.3 ms⁻¹ greater at song posts (mean±SD, N=12 birds). Song rate and song post selection were not influenced consistently by temperature or wind speed, but the birds sang from lower positions on one particularly windy day. This may have resulted from difficulty in holding on to exposed branches in windy conditions rather than a thermoregulatory constraint. The results suggest that the extra thermoregulatory costs at song posts would increase metabolic rate by an average of 10±4% and a maximum of 25±8% (N=12 birds) relative to birds singing at foraging sites. We estimated that metabolic rate would be 3–8% greater during singing than during quiet respiration because of heat and evaporative water loss in exhaled gases. The combined energy requirements for sound production, thermoregulation at exposed song posts and additional heat loss in exhaled air could increase the metabolic rate of willow warblers by an average of 14–23%, and a maximum of 42–63%, during singing. The energetic cost of singing may thus be much greater for birds in a cold, windy environment than for birds singing in laboratory conditions.     

Re-expression of songs deleted during vocal development in white-crowned sparrows, Zonotrichia leucophrys

White-crowned sparrows learn and produce multiple song types as juveniles, but most individuals stop singing all except one by the end of the first singing season. This single song type is generally maintained throughout adulthood. We demonstrate that, at the start of the second and subsequent singing seasons, this species can recall songs that had been deleted during the first singing season. The re-expression of song occurred in both the oriantha and the gambelii subspecies. Although all our males recrystallized the original song in the second year, our results indicate a mechanism for seasonal song change without new song memorization. The traditional dichotomy of closed-ended versus open-ended learning is inadequate for birds that learn early in life but can change their song output seasonally. We suggest that species can exhibit a closed sensitive period for song memorization and first production, with the ability to recall deleted songs later in life. This type of learning, selective attrition followed by subsequent re-expression, may be used by some species currently considered open-ended learners. Copyright 2000 The Association for the Study of Animal Behaviour.     

Phylogenetic and ecological determinants of the neotropical dawn chorus

The concentration of avian song at first light (i.e. the dawn chorus) is widely appreciated, but has an enigmatic functional significance. One widely accepted explanation is that birds are active at dawn, but light levels are not yet adequate for foraging. In forest communities, the onset to singing should thus be predictable from the species' foraging strata, which is ultimately related to ambient light level. To test this, we collected data from a tropical forest of Ecuador involving 57 species from 27 families of birds. Time of first song was a repeatable, species-specific trait, and the majority of resident birds, including non-passerines, sang in the dawn chorus. For passerine birds, foraging height was the best predictor of time of first song, with canopy birds singing earlier than birds foraging closer to the forest floor. A weak and opposite result was observed for non-passerines. For passerine birds, eye size also predicted time of first song, with larger eyed birds singing earlier, after controlling for body mass, taxonomic group and foraging height. This is the first comparative study of the dawn chorus in the Neotropics, and it provides the first evidence for foraging strata as the primary determinant of scheduling participation in the dawn chorus of birds.     

Social facilitation of male song by male and female conspecifics in the zebra finch, Taeniopygia guttata

Zebra finches are a ubiquitous model system for the study of vocal learning in animal communication. Their song has been well described, but its possible function(s) in social communication are only partly understood. The so-called ‘directed song’ is a high-intensity, high-performance song given during courtship in close proximity to the female, which is known to mediate mate choice and mating. However, this singing mode constitutes only a fraction of zebra finch males’ prolific song output. Potential communicative functions of their second, ‘undirected’ singing mode remain unresolved in the face of contradicting reports of both facilitating and inhibiting effects of social company on singing. We addressed this issue by experimentally manipulating social contexts in a within-subject design, comparing a solo versus male or female only company condition, each lasting for 24 h. Males’ total song output was significantly higher when a conspecific was in audible and visible distance than when they were alone. Male and female company had an equally facilitating effect on song output. Our findings thus indicate that singing motivation is facilitated rather than inhibited by social company, suggesting that singing in zebra finches might function both in inter- and intrasexual communication.     

Ways to test stochastic dynamic programming models empirically

Stochastic dynamic programming (SDP) models are widely used to predict optimal behavioural and life history strategies. We discuss a diversity of ways to test SDP models empirically, taking as our main illustration a model of the daily singing routine of birds. One approach to verification is to quantify model parameters, but most SDP models are schematic. Because predictions are therefore qualitative, testing several predictions is desirable. How state determines behaviour (the policy) is a central prediction that should be examined directly if both state and behaviour are measurable. Complementary predictions concern how behaviour and state change through time, but information is discarded by considering behaviour rather than state, by looking only at average state rather than its distribution, and by not following individuals. We identify the various circumstances in which an individual's state/behaviour at one time is correlated with its state/behaviour at a later time. When there are several state variables the relationships between them may be informative. Often model parameters represent environmental conditions that can also be viewed as state variables. Experimental manipulation of the environment has several advantages as a test, but a problem is uncertainty over how much the organism's policy will adjust. As an example we allow birds to use different assumptions about how well past weather predicts future weather. We advocate mirroring planned empirical investigations on the computer to investigate which manipulations and predictions will best test a model. Copyright 2000 The Association for the Study of Animal Behaviour.     

Perceived wintering latitude determines timing of song output in a migratory bird

Migratory bird populations frequently consist of individuals that overwinter variable distances from the breeding site. Seasonal changes in photoperiod, which varies with latitude, underlie seasonal changes in singing frequency in birds. Therefore, migratory populations that consist of individuals that overwinter at different latitudes with large overwintering ranges could experience within‐population variation in seasonal production of song. To test the influence of overwintering latitude on intrapopulation variance in song production in the spring, we subjected two groups of Eastern Song Sparrows (Melospiza melodia melodia) from the same partially migratory breeding population to different photoperiodic schedules associated with a 1,300‐km difference in overwintering location. One group remained on the natural photoperiodic schedule of the breeding site (resident group) while the other group experienced a nonbreeding photoperiod that mimicked a southern migration in the fall followed by a northern migration back to the breeding site in the spring (migratory group). We compared song output between the two groups in three different stages (nonbreeding, prebreeding, and breeding). Little singing occurred during nonbreeding stage sample dates (20 November, 6 December) for the resident group, and no singing occurred for the migrant group. During the prebreeding stage (27 January, 7 February), significantly more singing occurred in the resident group than in the migrant group. During the breeding stage (21 March, 4 April), after a simulated migration for the migrants, song output was similar in both groups. These results suggest that within‐population variation in wintering latitude may contribute to variation in seasonal changes in singing behavior, which may covary with readiness to breed. Studies utilizing confirmed migrants and residents, rather than merely simulated migrants and residents, are also needed to better understand these processes.     

Sleep, learning, and birdsong

Neuroethological research combines approaches derived from animal behavior and neurobiology to examine the neuronal mechanisms of behavior, often in the context of laboratory experiments on species chosen for particular adaptations. Typically, these species are not traditional laboratory animals yet they contribute greatly to a broad, evolutionarily diverse view of nervous system function. The surprising role of sleep in the vocal learning process of songbirds is one such example, described here. Juvenile zebra finches show sleep-dependent daytime fluctuations in their patterns of singing starting after their first exposure to tutor songs. Nighttime bursting activity in the vocal control song system also changes after the onset of tutoring, with the neuronal changes preceding the changes in objective behavior (daytime singing). After tutoring, the nighttime bursting increases and exhibits structure that depends on the particular tutor song, and the nighttime expression of these changes requires normal auditory feedback during daytime singing. These observations shed light on the information carried in neuronal activity during sleep and on the adaptive plastic mechanisms engaged during sleep. They suggest a new hypothesis of sensorimotor learning, whereby sensory memories act indirectly on sensorimotor feedback by modifying networks through plastic changes at night. Sleep may also contribute to adult song maintenance, with nighttime neuronal replay conveying information about songs produced during the day and possibly mediating daily changes in the structure of premotor bursts. Collectively, these insights contribute a comparative perspective to theories of sleep and memory, which also help to inform a developing understanding of how humans acquire and retain memories.     

Detecting diel patterns in the songs of Chipping Sparrows using citizen-science data

Previous studies have revealed diel patterns in the songs of Chipping Sparrows (Spizella passerina), with songs shorter in duration before dawn than after. However, the extent to which this phenomenon generalizes to the full geographic range of these sparrows is unclear, as is the question of whether citizen-science data can be used to detect diel patterns in song. We analyzed all available songs of Chipping Sparrows from the Macaulay Library and xeno-canto databases and compared the distributions of song features of recordings made at different times of day. We show that, across their entire geographic range in North and Central America, Chipping Sparrows sing shorter songs before sunrise (dawn song) than after sunrise (day song). Furthermore, we show that Chipping Sparrows shorten their songs by singing fewer syllables, not by singing faster: the number of syllables per song accounts for the observed difference in duration, not the syllable nor the intersyllable duration. Our results demonstrate that recordings from public repositories can be used to determine whether daily song patterns exist in species even in the absence of prior fieldwork, and we further propose that citizen-science recordings can be used to inform cross-species hypotheses and facilitate future studies to determine whether diel patterns in song are associated with differences in social behavior.     

Strategic Regulation of Body Mass and Singing Behavior in New Zealand Robins

The ‘small bird in winter’ paradigm states that body mass is a balance between the conflicting demands of carrying enough energy to survive nightly fasts while minimizing the risk of predation associated with carrying additional fat reserves. We conducted a short‐term food‐supplementation experiment during which New Zealand robins (Petroica australis) were provided with food on the second day of a 3‐d trial. This allowed us to test two predictions from models of strategic mass regulation in small birds: (1) individual birds reach the same end‐of‐day mass despite differences in their initial morning mass while, (2) using surplus energy for increased singing. As expected, robins gained mass at a higher rate early in the morning on the fed day than they did on either of the two control days, but there was no significant difference in their evening masses across the 3 d of the experiment despite birds on day 3 starting at higher initial masses than birds on day 1. Robins displayed a significantly higher rate of singing when receiving food supplements on day 2, supporting a link between energetic reserves and behavior. Our results suggest that potentially energetically costly behaviors, such as song production, are sensitive to short‐term changes in energy reserves, and that both state and behavioral predictions can be successfully integrated to provide tests of state‐based models of behavior.     

Rufous‐capped Warblers Basileuterus rufifrons show seasonal,temporal and annual variation in song use

In the majority of songbird species, males have repertoires of multiple song types used for mate attraction and territory defence. The wood‐warblers (family Parulidae) are a diverse family of songbirds in which males of many migratory species use different song types or patterns of song delivery (known as ‘singing modes’) depending on context. The vocal behaviour of most tropical resident warblers remains undescribed, although these species differ ecologically and behaviourally from migratory species, and may therefore differ in their vocal behaviour. We test whether male Rufous‐capped Warblers Basileuterus rufifrons use distinct singing modes by examining song structure and context‐dependent variation in their songs. We recorded multiple song bouts from 50 male Warblers in a Costa Rican population over 3 years to describe seasonal, diel and annual variation in song structure and vocal behaviour. We found that Rufous‐capped Warbler songs are complex, with many syllable types shared both within and between males’ repertoires. Males varied their song output depending on context: they sang long songs at a high rate at dawn and during the breeding season, and shortened songs in the presence of a vocalizing female mate. Unlike many migratory species, Rufous‐capped Warblers do not appear to have different singing modes; they did not change the song variants used or the pattern of song delivery according to time of day, season or female vocal activity. Our research provides the first detailed vocal analysis of any Basileuterus warbler species, and enhances our understanding of the evolution of repertoire specialization in tropical resident songbirds.     

Avian daily foraging patterns: Effects of digestive constraints and variability

Summary Birds show a typical daily pattern of heavy morning and secondary afternoon feeding. We investigate the pattern of foraging by a bird that results in the lowest long-term rate of mortality. We assume the following: mortality is the sum of starvation and predation. The bird is characterized by two state variables, its energy reserves and the amount of food in its stomach. Starvation occurs during the day if the bird's reserves fall to zero. The bird starves during the night if the total energy stored in reserves and the stomach is less than a critical amount. The probability that the bird is killed by a predator is higher if the bird is foraging than if it is resting. Furthermore, the predation risk while foraging increases with the bird's mass. From these assumptions, we use dynamic programming techniques to find the daily foraging routine that minimizes mortality. The principal results are (1) Variability in food finding leads to routines with feeding concentrated early in the day, (2) digestive constraints cause feeding to be spread more evenly through the day, (3) even under fairly severe digestive constraints, the stomach is generally not full and (4) optimal fat reserve levels are higher in more variable environments and under digestive constraints. This model suggests that the characteristic daily feeding pattern of small birds is not due to digestive constraints but is greatly influenced by environmental variability.     

Song environment affects singing effort and vasotocin immunoreactivity in the forebrain of male Lincoln's sparrows

Male songbirds often establish territories and attract mates by singing, and some song features can reflect the singer's condition or quality. The quality of the song environment can change, so male songbirds should benefit from assessing the competitiveness of the song environment and appropriately adjusting their own singing behavior and the neural substrates by which song is controlled. In a wide range of taxa, social modulation of behavior is partly mediated by the arginine vasopressin or vasotocin (AVP/AVT) systems. To examine the modulation of singing behavior in response to the quality of the song environment, we compared the song output of laboratory-housed male Lincoln's sparrows (Melospiza lincolnii) exposed to 1 week of chronic playback of songs categorized as either high or low quality, based on song length, complexity, and trill performance. To explore the neural basis of any facultative shifts in behavior, we also quantified the subjects' AVT immunoreactivity (AVT-IR) in three forebrain regions that regulate sociosexual behavior: the medial bed nucleus of the stria terminalis (BSTm), the lateral septum (LS), and the preoptic area. We found that high-quality songs increased singing effort and reduced AVT-IR in the BSTm and LS, relative to low-quality songs. The effect of the quality of the song environment on both singing effort and forebrain AVT-IR raises the hypothesis that AVT within these brain regions plays a role in the modulation of behavior in response to competition that individual males may assess from the prevailing song environment.     

Variation in singing style use in the reed bunting Emberiza schoeniclus: influencing factors and possible functions

The two main functions of bird song are territory defence and mate attraction. Considerable progress has been made in understanding how species adjust the use of songs to serve these and other (presumed) functions of bird song, but the striking variety of singing behavior observable in wild birds remains enigmatic. Some species make do with simple songs and small repertoires, while others show large, complex repertoires and still others have evolved several distinct singing styles. In most species with distinct singing styles, however, the functions of singing styles are poorly understood. Two distinct singing styles (type I and II, respectively) have long been known in the reed bunting Emberiza schoeniclus, while a new third one has recently been reported to exist. We first quantitatively investigated the evidence for the existence of three singing styles. Then, we tested predictions of the mate attraction hypothesis, the mate guarding hypothesis and the territory defence hypothesis by examining the relations between singing style use with social and temporal factors. Cluster and discriminant analyses supported the existence of three (instead of two) singing styles, which could be differentiated based on four variables referring to song structure and complexity. Use of singing styles was related to male mating status (consistent with the mate attraction hypothesis), but not to female breeding stage (no support for the mate guarding hypothesis). Finally, use of singing styles differed in relation to time of day, with the dawn chorus of paired reed buntings consisting almost exclusively of songs of the recently discovered type III singing style and daytime singing primarily consisting of songs of long‐known type I (in unpaired males) or II singing styles (in paired males). Our findings suggest that one singing style (type I) primarily serves to attract a social mate, although an additional territorial function of this singing style cannot be dismissed. The function(s) of the other two singing styles, both only sung by paired males, are not related to attraction of a social mate or to the own female's fertility, but appear to be important in the context of territory defence and extra‐pair matings.     

Effect of a vasotocin analog on singing behavior in the canary.

Groups of juvenile and 1-year-old male canaries were treated briefly with the vasotocin (VT) analog desGly(NH2)9d(CH2)5-[Tyr(Me)2,Thr4, Orn8]VT (dGVTA) during four time intervals between September and February. The canaries received subcutaneously testosterone-containing silastic implants at the start of the VT analog treatment to assure that despite age and season differences the birds would all have comparable plasma levels of testosterone. The VT analog was administered subcutaneously (0.7 micrograms/100 microliters) during the first 3 days (3 injections daily) of chronic testosterone treatment. Observations on the singing behavior were carried out between Day 8 and Day 30 after implantation of the testosterone-filled silastic tubing. The short-term administration of the VT analog influenced the amount of singing behavior during a 30-min observation interval measured 1 to 4 weeks later. Despite age differences the effect of dGVTA held and seemed more related to season than to age. The song duration (seconds of song/30 min) was affected in a dual mode. In early autumn the VT analog enhanced song duration of testosterone-primed canaries, but the same VT analog decreased song duration in the period November/January. These results suggest that the neuropeptide VT is implicated in control of seasonal changes in singing behavior.