Nearctic and Neotropical species.—

Flight calls have been studied most intensively in North America, and Evans and O'Brien (2002) compiled a guide to the flight calls of migratory birds that occur in the eastern part of the continent, mostly east of the 100th meridian. This unique resource provides detailed information on flight calls, including descriptions of the vocalizations and calling behavior, spectrographic representations, examples, and identification tips. Not all of the species contained in the guide regularly give flight calls, and not all of those that regularly give flight calls utter them at night. For example, cuckoos, woodpeckers, corvids, larks, swallows, thrushes, wood-warblers, tanagers, grosbeaks, emberizid sparrows, blackbirds, and finches (among other groups) give flight calls regularly, but most woodpeckers, corvids, larks, swallows, and finches rarely call at night (these are primarily diurnal migrants). Groups of species that do not regularly give flight calls when moving during day or night include New World flycatchers (Tyrannidae), vireos (Vireonidae), and mimids (Mimidae).

Palearctic and Paleotropical species.—

Although less intensively studied than Nearctic species, numerous Palearctic and Paleotropical species also utter flight calls (Chappuis 1989, van den Berg et al. 2003). Some of these are closely related to vocal New World species, such as Turdus thrushes (e.g. Fieldfare [Turdus pilaris]; Redwing [T. iliacus]), cardueline finches, pipits, and Regulus crests; other species are more typically Old World, such as bee-eaters (Meropidae), Emberiza buntings (e.g. Rustic Bunting [Emberiza rustica], Ortolan Bunting [E. hortluna]), many wagtails (Motacillidae), larks (Alaudidae), and fringillids. Like some Nearctic birds, not all those species regularly vocalize at night. In fact, it is primarily European Turdus thrushes (Siivonen 1936; Browne 1953; Vleugel 1954, 1960; Chappuis 1989; van den Berg 2003) and some Emberiza buntings and Regulus crests (M. Robb pers. comm.) that regularly give flight calls during nocturnal migration. Other Palearctic-Paleotropical species also give flight calls at night, including several species of pitta (Fairy Pitta [Pitta nympha], S. Lin pers. comm.; Blue-winged Pitta [P. moluccensis], P. Round pers. comm.), some Asian and Australo-Papuan cuckoos such as Long-tailed Koel (Eudynamys taitensis) and Pied Cuckoo (Clamator jacobinus) (N. Olliver pers. comm.), and Woodland Kingfisher (Halcyon senegalensis) (D. Mostert pers. comm.). Research in Africa, Asia, and Australia will probably identify numerous additional species that utter flight calls at night.

Like New World exceptions that rarely give flight calls, there are also Palearctic species that rarely give flight calls: Old World flycatchers (Muscicapidae) and Old World warblers (Sylviidae) are generally silent during migration. However, Pied Flycatchers (Ficedula hypoleuca) and Spotted Flycatchers (Muscicapa striata), which are not normally heard during nocturnal movements, apparently vocalize when visibility is poor (B. Bruderer pers. comm.; see Herremans 1993), and some sylviid warblers infrequently vocalize (similar to fledgling calls) during daytime movements (Blackcap [Sylvia atricapilla] and Chiffchaff [Phylloscopus collybita], M. Herremans pers. comm.; River Warbler [Locustella fluviatilis], J. Kriek pers. comm.).

Identification.—

How are calling birds identified when they are migrating at night and not visible? Identification of some calls is simple, because the nocturnal vocalizations are the same as the diurnal ones (Catharus spp.; Howes 1912, Evans 1994). However, identification of many species is often more complicated and requires additional information, which generally comes from two distinct sources (Evans and Mellinger 1999, Evans and Rosenberg 2000): (1) Comparisons of spectrograms of diurnal flight calls of known species and unknown nocturnal flight calls—many birds observed in visible morning flights often give flight calls (Evans and Rosenberg 2000, Evans and O'Brien 2002; see Gauthreaux 1978, Hall and Bell 1981, Wiedner et al. 1992 for descriptions of the morning flight phenomenon). Also, direct comparison is possible of unknown nocturnal vocalizations and flight calls recorded from birds in captivity or from birds carrying miniature microphones (Hamilton 1962, A. Cochran and M. Lanzone unpubl. data, W. Cochran unpubl. data). Figure 2 shows examples of these types of comparisons. (2) Correlating the seasonal timing and geographic range of nocturnal calls with known timing and migration ranges for each species—species-specific migration calendars are available for many species and locations in North America, often generated from accounts of species killed at night in collision with television towers, lighthouses or buildings, and historical arrival and departure dates (see Evans 1994, Evans and Rosenberg 2000; see also Hedges 2001).

Effects of atmospheric conditions.—

Several relationships between call counts and atmospheric conditions are apparent from the literature and to observers in the field during migrations. Counts of calls increase with increasing cloud cover and decreasing cloud ceiling, especially under artificial lighting (Cochran and Graber 1958, Graber and Cochran 1960, Ogden 1960, Dorka 1966, Graber 1968, Clemens 1978, Thake 1983, Evans and Mellinger 1999). Call counts also increase as birds approach boundaries between air masses of different density (Peterssen 1956), where conditions unfavorable for migration—such as precipitation, high winds, and poor visibility—force birds to pile up or descend (Graber and Cochran 1960). Calling usually occurs during periods of seasonally appropriate wind directions (Graber and Cochran 1960). Vleugel (1960) found that call counts of Turdus in Holland during autumn increased with the passage of cold fronts and decreasing temperature. Call counts are also positively correlated with 24-h trends of falling temperatures in autumn, whereas the inverse is true in the spring (Graber and Cochran 1960). Two potential caveats exist when extrapolating from the results of these studies: the studies represent site-specific results and there is no information about rates of calling and their relationship to independent measure of bird numbers.

Effects of altitude and topography.—

The effects of altitudes on calling rates are poorly known. Evans (2000) recorded many vocalizations within 500 m of the ground during autumn migration (see also Black 1997). Evans and Rosenberg (2000) and Evans (2000) indicated that flight altitudes of calling wood-warblers were less than 200–300 m those of calling thrushes (up to 450–500 m). There is also temporal variation in the altitudes of calling birds, and average altitudes may be substantially lower or higher on different nights (Black 1997, W. Evans pers. comm.). It is not known whether calling is primarily a boundary-layer phenomenon, occurring only in the atmospheric strata close to the ground.

Although migration occurs across broad spatial scales (Lowery and Newman 1955, Parslow 1969, Gauthreaux et al. 2003), evidence suggests that topographic features such as mountains or hills and coastlines concentrate birds (Eastwood 1967; Bruderer 1978; Richardson 1978, 1990; Åkesson 1993; Williams et al. 2001); these features also appear to concentrate flight calls. Evans and Mellinger (1999) found that changes in wind conditions resulted in larger numbers of calls counted on the coast of Texas; southwesterly winds forced birds migrating inland toward the coast, and to avoid drifting over the Gulf of Mexico, these birds piled up on the coast and then moved north along it. Evans (2005) reported that when the cloud ceiling is low, altitudinal variations of terrain disrupt the flight of calling migrant passerines and effectively concentrate calling birds in areas of lower altitudes.

Temporal patterns.—

Despite the variability in all these relationships, patterns of call counts across seasons and years are often consistent and probably represent some true behavioral and biological patterns (e.g. the migration timing of different species). Conversely, nightly temporal patterns of calling are much more variable. These patterns could represent site-specific differences and additional unknown behavioral and biological patterns. Ball (1952) recorded ≈90% of thrush vocalizations in the hours just before dawn, with a ratio of 27 calls after midnight to 1 call before midnight (from 33,921 calls). Graber and Cochran (1960) supported this conditionally, though they detected migration consistently at any hour of the night, but a marked peak in calling occurred in the hours before dawn if migration occurred all night. Farnsworth and Russell (2005) reported a similar pattern in an acoustic study of migration over the Gulf of Mexico, finding that the nightly peak of call counts occurred in the two hours just before dawn. By contrast, call counts of Turdus in Europe usually peaked in the hours closest to local midnight, with deviations from this pattern usually associated with the passage of a front (Siivonen 1936; Browne 1953; Vleugel 1954, 1960). Furthermore, flight-call counts have varied extensively throughout the night, though on many nights, peaks occurred in the hours close to local midnight (Ross et al. 1995, Farnsworth et al. 2004).

Reasons for the variability in nightly peak call counts are not known, but they might include locally varying weather conditions (Graber and Cochran 1960, Graber 1968, Evans and Mellinger 1999, Evans and Rosenberg 2000, Evans 2000) and variation in flock sizes and species composition (Marler 1956; Hamilton 1962; Thake 1981, 1983; Farnsworth et al. 2004). Some variability might result from different species descending at different times of the night and calling at different rates during descent (Graber 1968). Furthermore, rates of calling may vary highly among individuals. The hourly mean call counts for Swainson's Thrushes bearing small microphones “ranged from 0 to 37, including one individual [briefly] vocalizing 16 times minute⁻¹ and one individual that did not call for 3 hours” (W. Cochran pers. comm.).

Hemispheric patterns.—

Another intriguing but unexplained pattern is that, by nearly all accounts, vocalization by nocturnally migrating birds in the Palearctic seems to exhibit a different pattern than vocalization in the Nearctic migration system. Calls are uttered more frequently, at greater magnitude, and by more species in the New World. Whether this is a function of small sample sizes in the European studies, fundamental behavioral differences between the migration systems, phylogenetic effects, or some combination thereof is unknown. Anecdotal accounts from numerous European researchers suggest that flight calling is limited even in the species that regularly vocalize at night (compare Vleugel 1960 and Ball 1952).

Are flight calls learned?—

Although flight calls are one of the earliest calls to appear in the repertoire of juvenile cardueline finches (Mundinger 1979), these species learn and change these calls by imitation throughout life (Mundinger 1970). Is this pattern found in other passerines? Hamilton (1962) suggested that the calls were innate. Comparing the flight calls of captive-bred birds with diurnal and nocturnal vocalizations of wild birds could provide some answers. No detailed seasonal usage pattern of flight calls is available, and developing a timetable for many species should be informative. Understanding the ontogeny of flight calls is crucial for attempts to classify them and has important implications for determining their functions. Furthermore, if flight calls are learned in a diverse array of passerines, the influence of vegetation structure and ambient noise spectra may play an important role in their development (Hansen 1979, Nottebohm 1985).

What factors constrain flight calls?—

Recent studies show that different species of birds have different detection thresholds for signals in ambient noise (Klump 1996, Langemann et al. 1998), as well as different hearing thresholds (Dooling and Saunders 1975, Dooling 1982, Okanoya and Dooling 1987). Differences in the perceptual abilities of species could play important roles in determining variation in call frequencies. Flight calls may also be subject to different selection pressures related to encoding information (relative to selective pressure on songs). Reverberation, amplitude modulation rate, consistency of transmission, and spectral distribution of ambient noise are important sonic properties that define vocalizations, and these properties vary with selection pressures among habitats (Marler 1955, Morton 1975, Ryan and Brenowitz 1985, Wiley 1991). How these properties relate to the use and function of flight calls is not known. Also, although Hamilton (1962) did not believe that predators play a major role in shaping flight calls, Gill and Sealy (2003, 2004) found evidence that high-frequency seet calls alert individuals to brood parasites. Flight calls, which have similarly high frequencies and short durations, may be used to communicate information above the frequency thresholds of predators (Langemann et al. 1998). If these calls are related in some way to fledgling vocalizations (see Tyler 1916), there may be a direct relationship between such anti-predatory behaviors and flight calls.

Over what distances are flight calls used?—

No one has studied the range of distances over which birds communicate through flight calls. Because the frequencies of signals indicative of long-range communication are lower (Marten and Marler 1977, Larom et al. 1997, Larom 2002), the characteristically high frequencies of these calls suggest that they may be used primarily for communication over short distances. However, the constraints associated with sound production during flight are not known, though they have implications for nocturnal communication and the architecture of nocturnal groupings. Furthermore, optimal in-flight communication using certain frequencies could shape flight vocalizations that either travel best in specific atmospheric and microclimatic strata or avoid specific ambient noise spectra (Larom 2002, Rundus and Hart 2002, Slabbekoorn et al. 2002, Slabbekoorn and Peet 2003).

How variable are flight-call characters?—

Intraspecific variation in flight calls is not a recent discovery (Ball 1952), but the extent of this variation in call characters, such as frequency, has only recently received greater attention (Hahn et al. 2001, Evans and O'Brien 2002, Sewall et al. 2004). Evidence suggests that the characteristics of certain species' flight calls, notably those of thrushes, larks, pipits, and finches, vary substantially (Evans and O'Brien 2002, M. Robb pers. comm., W. Cochran unpubl. data). The implications of such variation are also unknown, though Mundinger (1970, 1979) suggested that call-matching and imitation could be important. A few studies have explored the extent of phylogenetic signals in songs (e.g. kinglets; Päckert et al. 2003) and calls (e.g. herons; McCracken and Sheldon 1997), but few species are represented, and none of these studies explicitly addressed flight calls. Similarly, if character release exists in flight calls, the relationship between it and the diversity of related species is not known.

Are there potential applications for flight calls?—

Flight calls may be useful characters for comparative analyses among taxa (Mundinger 1979, Farnsworth and Lovette 2005), aiding in resolving cryptic species (Groth 1988) and delineating taxa (Mundinger 1979; Adkisson 1981; Groth 1988, 1993a; Hahn et al. 2001; Sewall et al. 2004). However, categorizing natural variation in flight calls and expanding the sampling of spatial and temporal distributions of flight-call data sets are critical precursors to pursuing such directions (Sewall et al. 2004). Flight calls may also shed light on habitat preference and morphology (Groth 1993a, b). For example, in phylogenetically controlled and uncontrolled analyses, Farnsworth and Lovette (2005) found little support for morphological constraints on flight-call frequencies in parulids. This pattern differs from the widely reported pattern in larger birds that vocalize at lower frequencies (Greenewalt 1968, Wallschläger 1980). There are also diverse and potentially powerful applications for monitoring flight calls to document broad front patterns of species-specific nocturnal movements and altitudinal distribution of calling migrants (Tyler 1916; Ball 1952; Graber and Cochran 1960; Graber 1968; Evans 1994, 2000; Evans and Mellinger 1999; Evans and Rosenberg 2000) and to identify the points of origin of calling migrants (W. Evans pers. comm.).

Patterns of nocturnal bird migration as detected by radar and acoustic methods do not always differ, though these methods illuminate what are likely fundamental differences in patterns of behaviors (Ross et al. 1995, Larkin et al. 2002, Farnsworth et al. 2004). Nocturnal call counts of migrating birds can be useful as indices of nocturnal bird density aloft (Larkin et al. 2002, Farnsworth et al. 2004), though extensive variation in calling rates (within and among species) poses a major challenge for measuring bird density from flight calls alone (Libby 1899, Graber 1968, Dierschke 1989, Evans and Mellinger 1999, Farnsworth 2001, Farnsworth et al. 2004). Moreover, there can be substantial differences between a high volume of migration and high incidences of calling (Graber and Cochran 1960, Ross et al. 1995, Farnsworth et al. 2004). More detailed studies of nocturnal migration using radar and acoustic methods simultaneously, and studies that span even larger spatial and temporal scales, will be necessary to resolve these issues. More studies using stand-alone methods based on flight-call counts will generate a base of departure and arrival data as well as relative proportions of species on different nights.