Northern Cardinal Cardinalis cardinalis
Version: 2.0 — Published February 12, 2021
Account navigation Account navigation
Sounds and Vocal Behavior
Welcome to Birds of the World!
You are currently viewing one of the free accounts available in our complimentary tour of Birds of the World. In this courtesy review, you can access all the life history articles and the multimedia galleries associated with this account.
For complete access to all accounts, a subscription is required.
Already a subscriber? Sign in
Songs are typically pure-toned, whistle-like series of repeated syllables whose fundamental frequencies range from approximately 1–8 kHz; consecutive syllables of songs are separated by silent intervals of < 1 s, and different syllables are often sung in a set order (Figure 2; additional sound spectrograms in 241, 2, and 12). Many songs are about 2–5 s long, but cardinals occasionally sing songs lasting 15 s or more (SLH). The same songs are sung by both males and females, following structural rules. Three widespread categories of songs are commonly verbally described as what cheer, cheer, cheer (Figure 2C, first 4 syllables); cheer, cheer, what what what what; and birdie, birdie, birdie (Figure 2B). While these verbal descriptions evoke relative frequencies and timing of component syllables, they do not precisely specify frequency, timing, or fine details of syllable structure, and an individual’s repertoire may include multiple songs that fit the same verbal description. For example, cardinals in south-central Wisconsin sang two types of what cheer, cheer, cheer songs: one had ~0.1 s long what syllables that ascended from ~2 to 4 kHz, followed by ~0.6 s long cheer syllables that descended from ~6 to ~2 kHz; the other had ~0.5 s long what syllables that ascended from ~2 to 4 kHz, followed by ~0.4 s long cheer syllables that descended from ~2.5 to 1 kHz. The least pure-toned song component is the chirr syllable ( ), compared by Lemon (241, "sound Z") to the chatter of a red squirrel (Tamiasciurus hudsonicus); the lower amplitude of chirr, relative to other song components, is shown in Figure 2C. See further information on chirr syllables in Pattern of Presenting Repertoire, Social Context and Presumed Functions of Vocalizations, and Mechanisms of Sound Production. Evidence from field and laboratory studies indicates that shared repertoires of neighbors are produced by generations of song learning with occasional improvising and copying errors (7).
Calls are short sounds with broad-frequency bandwidths; time elapsed between repeated calls, number of repetitions, and sequencing of calls are far more variable than for songs. Both sexes use many of the same calls. At least 16 different calls are described for Northern Cardinals; sound spectrograms and contexts are given for many by Lemon (4). Calls vary from short sounds (< 0.05 s) with broad frequency ranges, to longer (around 0.6 s) sounds with narrow frequency ranges or harmonic series (4). Different calls may be given in rapid succession during agonistic interactions.
The most commonly heard calls of Northern Cardinals are chip calls (Figure 3A and 3B, ). Chip calls have most of their energy between 4 and 8 kHz, though they may extend into lower frequencies at lower amplitudes (4). Chip calls are used in many “contact” situations: in response to territorial invasions and in conflicts at feeding stations; when predators approach nests; by territorial males to females trespassing in their territories; and “when excited, as when bathing” (all from 4); by females approaching their nests to incubate eggs or brood young; by both males and females bringing food to the nest (242, 12); between mates foraging out of one another's view (SLH); and by parents apparently trying to lure young to fledge (100, SUL). Chip calls are often given in groups; amplitude and rate of utterance vary, and further study may show that this category includes multiple different calls, or that chip calls intergrade into other call types, particularly chuck and chitter (4, SLH).
Chuck calls differ from chip calls primarily in their broader frequency range, with peak energy extending from below 4 kHz through ~8 kHz, with more time elapsed between sequential calls than is typical for repeated series of chip calls. Chuck calls are used in agonistic contexts (4, ).
Chitter calls are given in closely-spaced series, with most energy between 7 and 8 kHz, when hawks flew overhead, when tape recordings (presumably of conspecific song) were played in their territories, and by captive cardinals presented with “strange”/novel objects (4).
Took calls (Figure 3C, ) are used mainly in the context of one cardinal bringing food to another. Took calls are commonly given by one or both mates when males approach the nest with food for the female and/or nestlings, by the female when she approaches the nest with food (242, 12), and by either or both parents in the course of feeding nestlings. They are also given by females “both to the eggs when incubating and to the young nestlings,” by the female apparently soliciting courtship-feeding, and by the male and female to one another when a second male is chasing the first and also courting the female (243: 118–120).
Other calls are described by Lemon (4), with sound spectrograms and contexts for many, including alarm shriek, female precopulatory see, agonistic pee-too, piping of male to initiate courtship-feeding, twitter (possibly submissive), and food-begging calls of fledglings. “Chattering” begging calls of nestlings and fledglings, as well as pfitt, pfitt calls of male driving away young of the previous brood once the next brood hatched, are described by Laskey (100). Nice (244) added ya, hiss, and snarl given by male driving smaller species from feeder.
Age-Related Changes in Maturation of Vocalizations
In the laboratory, song develops gradually during the first year (1). Soft warbles at about 3 weeks of age typically become more songlike during autumn. Singing then ceases until January, when subsong resumes as long sequences of irregularly spaced whistles that fluctuate rapidly in frequency. Subsong also includes several recognizable adult vocalizations: chirr song element (see Social Context and Presumed Functions), as well as pee-too, chuck, and chitter calls. By April, young birds sing adultlike songs.
Song development in nature is similar. “Warbling soft whistle” reported from young 13 days after fledging (245). During August, apparently both males and females warble softly, and males include some adult syllables (100); some juvenile birds (sex not determined) sing “fairly well” in August (7). During late February and early March, two males thought to be in their first year used local syllables but in atypical combinations, in songs with a greater variety of syllable types than normal; their songs also included call notes; syllables were often quavering and sung at relatively low amplitude (1). Normal adult songs followed within a week. Females have a similar developmental sequence (1). Repertoires of syllables and song types do not appear to change from year to year in adults of either sex (241, SLH).
Evidence for or Against Vocal Learning
Songs are learned (1, 246, 247, 7, 21). Cardinals isolated from adult song as nestlings produce atypical repertoires of syllables and song types; some structural elements seem normal, however, and syntactic structure of isolate songs is the same (males) or similar (females) to that of wild-type songs. Male isolates reared together learned from one another to develop a largely shared repertoire of syllable types. Also, deafened juveniles develop abnormal song elements, and birds deafened as adults cannot maintain proper syllable structure.
Hand-reared males and females learned songs from tutor tapes of either male or female cardinals, but not of Blue Grosbeak (Passerina caerulea) (246, 21). Lab-reared cardinals also improvised more than would be expected from the large portion of learned song repertoires that neighbors share in wild populations (2, 12).
In contrast to songs, calls appear not to be learned; calls that were given by individuals reared in isolation, and by birds deafened before they had given these calls in their own vocalizations, did not differ from calls developed by wild individuals with intact hearing (247).
Sensitive Periods of Learning
Hand-reared males and females first learned songs from tutor tapes heard between day 11 and day 30 (21), when young cardinals in nature are still on the territories where they hatched. One male taken into captivity 3 weeks after fledging later sang songs with some features unique to the area where he hatched (7), and males captured in their first December also sang some syllable combinations typical of their capture localities (1). In the laboratory, some males continued to learn new songs from tutor tapes through at least day 295, but no female learned new songs after day 100 (21; but see Beecher  for relative ineffectiveness of tape tutors). Wild yearling males apparently can learn songs from their neighbors in their first spring: 2 yearling males sang some of the atypical songs of an older male in an adjacent territory (1).
Cardinals share a large proportion of syllable and song types with their mates and neighbors. Songs become less similar with increasing distance from a particular locality (2, 249), but Lemon (7: 395) estimated that a cardinal “would have to travel perhaps hundreds of miles to get to an area where most of the repertoire is unfamiliar,” and stated that some syllable and song types occur throughout North America 2).
Songs can differ geographically in other ways, too. (1) Syllable types may be combined into songs in different ways (2). (2) Number of syllable types per song varies. For example, in southern Ontario (241), Texas (136), and south-central Wisconsin (242, 12) almost all songs are composed of 1–2 syllable types, but in southern Minnesota, songs with 3 syllable types are fairly common (SLH). (3) Anderson and Conner (249) suggested that songs might differ by habitat in ways that favor sound transmission, but small sample sizes and varying cardinal densities and rates of song interaction at the three sites sampled indicate that more data are needed to determine whether this is the case. Also see Singing in Noisy Environments.
Differences in the level of male responses to songs from different geographic locations has been reported based on playback experiments, but it is often unclear whether the choices of songs or natural or constructed song sequences used in these experiments have adequately controlled for song characteristics that vary with context, and might be expected to elicit different levels of response (see Social Context and Presumed Functions). Southern Ontario males responded more strongly (measured by approach to speaker, number of songs sung, and song-matching) to local songs than to similar song types recorded 13 km away (250), or to song recorded 34 km away that did not closely resemble any local song types (3). Song from a Texas cardinal, dissimilar to any local song types, was apparently ignored, and may not have been recognized as conspecific (3, 250). Provost et al. (107) found that male Northern Cardinals from the igneus and cardinalis subspecies groups on opposite sides of the Cochise Filter Barrier did not respond to one another’s songs in a way statistically distinguishable from their responses to Cactus Wren (Campylorhynchus brunneicapillus) song. Males from one population responded equally to local and distant songs from their side of the filter barrier, but males from the other group responded strongly only to local songs. Confounding factors render these results somewhat tentative, including lack of controls for whether the playback songs were recorded during song exchanges with conspecifics, non-reciprocity of playback testing, qualitatively different responses from birds in the two groups, and whether songs were played from within territories (see Priorities for Future Research). Additional research is also needed to determine whether male responses to playbacks can be used to predict female mate choice decisions that could produce reproductive isolation. Given the role of song type matching in Northern Cardinal communication, choosing a mate with shared song types might enhance communication, although at least males can learn new songs into their first spring, and a great deal of geographic variation exists in song type repertoires of subspecies cardinalis over its large and presumably reproductively continuous geographic range (see Places of Vocalizing: Singing from Nest, and Social Context and Presumed Functions of Vocalizations, and Sensitive Periods of Learning).
Laboratory-raised Arizona females (presumably of subspecies superbus) did not discriminate among potential mates on the basis of song dialect (21). Number of Precopulatory Displays (see Behavior: Sexual Behavior) given by estradiol-injected females when a videotaped image of a courting male was accompanied by either familiar or unfamiliar songs did not differ. Sample sizes were small, however, and the conditions under which these females were raised and tested was highly artificial, so their relevance to the behavior of wild cardinals is unknown.
Considerable individual variation has been noted in renditions of the local repertoire of syllable types (251). In a Mississippi population, Gentry (212) found that focal males responded more strongly (measured by approach to the speaker, number of songs sung, and persistence in the area of the simulated intrusion) to songs recorded from known neighbors with which they shared a territory boundary, compared to song recorded 10 or more years before from cardinals that lived in different but unspecified areas within 80 km of the focal males, when both songs were played on territory centers. Gentry (212) also found seasonal variation in the strength of the response to songs of known neighbors, compared to these older recordings, with stronger responses to known neighbor song in the spring and equal responses to both types of stimuli in the winter. Similarity of the older recordings to current song types of the focal males was not assessed.
Singing in Noisy Environments
Minimum song frequency increased in Ohio forest stands with higher levels of background noise from surrounding urbanized areas; while this frequency shift may help avoid overlap with low-frequency noise, it may also be an artifact of songs being sung at higher amplitude in urban areas (18). Despite the trend for urban cardinals in Ohio to sing songs at higher frequencies than do rural cardinals, males in both rural and urban areas that sang songs of lower frequency had nestlings in better condition, and occupied territories with sparser understory vegetation. Song frequency was not correlated with the number of young fledged over the season (20). Thus, the higher-frequency songs of urban cardinals did not correlate with net reproductive advantages in urban habitats, perhaps due to a mismatch between cardinal preferences for dense understory vegetation and the low quality of the habitat provided by invasive dense understory shrubs in urban areas, and with stochastic predation accounting for much of the variation in reproductive success in urban habitats (252, 253, 54, 18, 20).
Males sing throughout the year, with song most common February–September in Tennessee (100); February–July in Norman, Oklahoma (sex not specified, presumably males; 244); late February–July in Wisconsin (sex not specified, presumably males; 254), February–April in central Kentucky (10); and December–October in southeastern Mississippi (255). Males sing least in November and December in Tennessee (100) and North Carolina (17). The female song season is shorter than that of males: from mid-February or March to August in Tennessee and Kentucky, January-September in North Carolina (243, 100, 9, 17).
Song production by laboratory-reared males in their first spring was associated with a significant increase in testosterone and luteinizing hormone; females singing in their first spring had a much smaller associated increase in these hormones; wild male cardinals also had far higher breeding-season levels of testosterone and luteinizing hormone than did wild females (21, but see Other: Physiology: Hormones).
Daily Pattern of Vocalizing
Varies seasonally for males: In south-central Wisconsin, first daybreak song 10–15 min after civil twilight from late December through March and mid-August to September, at civil twilight from April to mid-May and July to mid-August, and 10–15 min before civil twilight from mid-May through June. In the evening, cardinals last sang at relatively higher light levels than before daybreak (mean 0.83 foot-candles greater for evening than daybreak song April–July; 256). In eastern Tennessee, in early spring, males sang largely in the early morning (from before sunrise to 0900 Eastern Standard Time); later in the season, song was also reported in the late afternoon, but was particularly uncommon in the early afternoon; singing was also rare on overcast days and during precipitation (211). During the dawn chorus, males switch between song types more frequently (i.e., after singing fewer songs of a given type) than in “broadcasting” singing later in the day (SLH, differences not quantified). Daily patterns of singing have not been described for females.
Nocturnal singing of short songs common from April–July in Tennessee (reported times between midnight and “about a half hour before dawn”); sex of these singers was generally not known, but in one case, song came from where a female had been roosting (257: 2, 100). Nocturnal singing reported on both moonlit and dark nights in “spring and summer” in the Houston, Texas, area (139).
Places of Vocalizing
Sites; Differences in Vocalization According to Site
Song perches of males are usually > 10 m above ground, with 4–7 per territory (67, SUL); males sometimes sing from low perches in trees or shrubs, and occasionally from the ground, generally at lower amplitude than when singing from higher song perches (SLH). Females usually sing from shrubs or lower perches in trees, and usually sing at lower amplitude than males, but occasionally sing from high, exposed perches, and occasionally sing as loudly as is typical for males (SLH, JMJ). Male cardinals singing within 100 m of Cooper's Hawk (Accipiter cooperii) nests, sing from less-exposed perches than do male cardinals singing from matched sites without Cooper's Hawk nests, presumably to reduce the risk of being preyed upon, and also implying a communicative advantage to singing from exposed perches when the risk of predation is lower (258).
Singing from Nest
Females sing from the nest while incubating and brooding, generally in response to their nearby mate's chip calls or songs. These songs and vocal exchanges range from brief and simple to elaborate. Female song from the nest appears to provide information to the mate about when to come to the nest with food (and may also convey information specifically about when to stay away) (242, 12, JMJ). This information may allow the visually conspicuous male to restrict his visits to the nest to times when food is needed, thus reducing flights that may alert predators to the nest's location. Correlations between female song and subsequent male behavior indicate that the male's baseline rate of bringing food to the nest can be either raised or lowered by song messages from the female. During the nestling period of 9 pairs in south-central Wisconsin, when males gave chip calls without song within 12 m of the nest, they came to the nest on average 80% of the times that their mate responded with song, compared to only 45% of the times that she did not sing. If males gave song vocalizations within 12 m of the nest, they came to the nest 64% of the times that their mate responded with a non-matching song type, 36% of the times that she did not sing, and only 9% of the times that she sang a song that matched the male's (i.e., shared at least 1 of the syllable types in the male’s song). The linked sound file (ML541393 ) provides an atypically long and elaborate song exchange between a female brooding nestlings, and her mate singing from a perch within 5 m of the nest. The male sings the first song: the pair were recorded in stereo, and his songs are louder on one channel, and the female’s songs are louder on the other channel. This song exchange includes especially long songs from the female that incorporate chirr syllables; in some parts of the exchange, the male and female are simultaneously singing series of the same syllable types, with the start times slightly offset. After this song exchange, the male left the area without visiting the nest, and the female stayed on the nest. This song exchange is much more prolonged and complex than is typical of nest-vicinity song exchanges, and demonstrates the considerable singing abilities of female cardinals. The use of songs with matching syllable types was typical in the south-central Wisconsin population when the male did not visit the nest after a song exchange. The apparent visit-lowering message of song matching in this population may not be universal; limited data from Minnesota and Connecticut populations indicate that any singing by the female from the nest increases the probability that the male will visit the nest, whether the female's songs match her mate's, or not (SLH). There is also a report of a Virginia male that approached the nest and briefly sat on eggs on a cold day, 15–20 seconds after a female sang a song that ended with a chirr from the nest or nearby, and then departed (259).
Males sing more than females: females sang at about 10% the rate of males in a Kentucky study (9), and at about 20% the rate of males in a North Carolina study (251). Males and females sing the same repertoires of syllable types, and combine them into songs in the same ways (12, and references therein), although females singing from their nests occasionally give longer songs with more syllable types (SLH). Songs of females in southern Arizona (presumably subspecies superbus) have louder second harmonics relative to the fundamental, and lower stereotypy in successive renditions of the same syllable, than do male songs (13). Vondrasek (251), using different analysis methods, did not find these differences in North Carolina cardinals (subspecies cardinalis). In southern Arizona, both males and females sang in response to songs played within their territories that had been recorded from non-neighboring local cardinals of both sexes; pairs considered as a unit sang more songs in response to playbacks of songs sung by males than by females; when singers could be visually identified to sex, males sang more than did females when responding to male songs, and the sexes did not differ significantly in their singing responses to female songs (14). In similar experiments in North Carolina, Vondrasek (251) found no significant difference in responses to playbacks of male and female song, by either males or females. In central Kentucky (subspecies cardinalis), males sang in response to playback of songs of both sexes, but females sang (rarely) only in response to male song (9); however, elsewhere females have been reported to sing in response to playbacks of song of both males and females (21, 251). In North Carolina, females in newly formed pairs on new territories sang at higher rates than did females in established pairs on territories they had occupied in previous years; this pattern was not found for males (251).
Song control nuclei in brains of males are 1.5–2 times larger in volume than those of females, as has been found in most other oscine songbird species studied in which both sexes sing regularly (25 and references therein). This is a much smaller difference than is found in oscines in which females do not sing or sing only rarely. Both sexes also show statistically significant seasonal variation in the size of song nuclei (larger in March than in January); the volume difference in male versus female song control nuclei was the same at both the midwinter and early breeding season sampling points (25). See information on possible mechanisms of hormonal control in Other: Physiology: Hormones.
Repertoire and Delivery of Songs
Number of Song Types per Individual
Males. 9–14 syllable types (n = 95 birds) and 8–10 commonly used song types (n = 25 birds) in southern Ontario (241, 2); 8–21 syllable types in east Texas (n = 16 birds, 249); up to 18 syllable types in south-central Wisconsin (12).
Females. Up to 17 syllable types in south-central Wisconsin (12).
Variations of syllable types that may intergrade sometimes make classification somewhat arbitrary (2). Also, it is unlikely that all rare combinations of syllable types into songs are catalogued (SLH). Anderson and Conner (249: 437) calculated that “90–110 songs were required to obtain an approximate measure of syllable repertoire” for individuals with repertoires of 8–21 syllable types in east Texas. Requisite sampling effort varies with the rate of switching between song types, which varies with time of day and behavioral context.
Pattern of Presenting Repertoire
Repertoire of syllable types is presented in songs that usually consist of strings of repeated syllables (see Figure 2). Within a song, the duration of the silent interval following a syllable is positively correlated with the length of the syllable (6), likely reflecting the volume of expelled air to be replaced with the subsequent minibreath (see Mechanisms of Sound Production). The syllable repetition rate tends to increase as a song progresses (6).
Most syllable types have a typical position within songs. Those used to begin songs are either sung alone or followed by other syllable types that in most cases are not used to begin songs. Pairs of syllable types often occur together; in the rare cases in which one initiating syllable type may be followed by either of two others, the particular type of second syllable chosen may be correlated with the number of repetitions of the first (8). In a southern Minnesota population that regularly sang songs with 3 syllable types, within-song ordering of associated syllables was more variable (SLH).
Certain terminal syllables can occur at the end of any song. Chirr is a geographically widespread example used at the end of many different song types (see Figure 2C, ). A southern Minnesota population also had short high-pitched syllables that were used in this way (SLH).
A series of songs that begin with the same syllable type, or in which the same syllable type alternates between the first and second position in sequential songs, is referred to as a “bout”; within a bout, shorter songs with just one syllable type often alternate with longer songs in which that syllable type is followed by others (6, SLH). When a male is singing by himself (i.e., with no song interactions with other cardinals), the sequence of song types in successive song bouts tends to follow a first-order Markovian relationship; i.e., the current song type is good predictor of what the next song type will be, particularly if a relatively short period of time elapses between successive song bouts (6). In contrast to associations between syllable types, which tend to be unidirectional (i.e., A is often followed by B, but not vice versa), associations between song types in successive bouts tend to be two-way (6).
Social Context and Presumed Functions of Vocalizations
In general, calls are used for communication at shorter distances than are songs (4, 7). Calls function to maintain contact when cardinals cannot see one another, and they are used in aggressive interactions, when predators are seen, by females soliciting copulation, and by males feeding their mates. Parents call when they approach the nest and while feeding at the nest, and nestlings and fledglings call as they beg for food.
Both males and females sing in territory defense (e.g., 5, 9, 14, 33, 212). Males countersing with male neighbors, and sing in response to naturally occurring and simulated territorial intrusions by another male. When males countersing with neighbors or mates, number of songs/bout decreases, number of syllables/song increases, and number of syllable types/song increases (10). As neighboring males approach one another more closely during territorial interactions, their rate of switching between song types increases (i.e., bout length decreases), and the amount of song-matching increases (5). Male singing rate and number of songs per bout were higher, and individual song durations were longer, during pre-nesting and nest-building periods in Kentucky than during incubation, brooding, or fledgling periods (10). When the periods with eggs or young in the nest were excluded, singing rates of males in North Carolina did not change from late April through early July (17).
Females may sing during or immediately after chasing other females in their territories (211, 9), and may sing in response to playback of songs of both females and males from within their territories (21, 251). Song by females during nest defense has been noted (15; 19). Females in new pairs sing more frequently than do females in established pairs, and sing at higher rates during the early breeding season than later in the breeding season (17).
Males sing in courtship displays (4), and males and females sing together before nesting, when they typically match one another's song types more so than do countersinging males (5). Intrapair song-matching may help mates from different dialect areas learn each other's songs (242), but if females do learn songs from their mates, their sensitive period is longer than indicated by their learning from tutor tapes in the laboratory (21). Song exchanges between incubating or brooding females and their mates appear to coordinate male feeding visits to the nest (12; see Singing from Nest). Countersinging of mated pairs may also function in unspecified aspects of “pair-bond maintenance,” and in announcing to other cardinals that a mated pair is present on a territory (11). The female's singing rate does not change when her mate is temporarily removed before nesting (11); experiments have not been conducted to determine whether male singing rates increase upon mate removal.
Correlations have been reported between song characteristics and male body size, territory quality, and reproductive success, although differences in context may not have been adequately controlled, especially when different population densities in different locations may have affected the probability that song sampling occurred during song exchanges with territorial rivals. In rural eastern Texas, males with shorter songs and fewer syllables and syllable types per song, had territories of higher quality (higher density of understory vegetation), and territories of higher quality had more nestlings and fledglings/nest, though the direct correlation between song characteristics and reproductive success was not statistically significant (50). In central Ohio, while the increase in minimum song frequency among urban cardinals (compared to more rural conspecifics) was best explained by increased noise levels within urban cardinal territories, the longer songs of urban males, made up of syllables repeated at higher rates, were best explained by increased cardinal densities in urban areas, with correlated increased rates of territorial interactions with conspecifics (18). In rural areas, larger males sang shorter songs, with fewer syllables and syllable types per song, and slower syllable repetition rates, and held territories with denser understory vegetation; male song did not co-vary with provisioning behavior, or with annual reproductive success (20).
Chirr syllables may have particular significance as a display of the singer's fitness. Chirr elements are produced at a rate of ~30 per second by rapid pulsation of the left side of the syrinx between open and closed, while the right side remains closed; the rapid trill rate precludes minibreaths between trill elements to prolong chirr duration (22, 260). Thus, inclusion and duration of chirr syllables in songs could potentially be used as an indicator of the singer's lung capacity and general cardiovascular health (22). Difficulty of chirr production may be showcased by the normal placement of chirr elements at the end of already long songs, and by the occasional use of chirrs to construct unusually long and elaborate songs by bridging sets of syllable types not normally sung sequentially (SLH, ML541393 ). Chirr syllables are sung at significantly lower amplitude than other syllable types in the same song, and thus presumably function primarily in short-distance communication with cardinals close to the singer. Chirr syllables are given in both territorial and courtship interactions when cardinals are in a relatively high state of arousal, including escalating territorial conflicts, courtship, and nest-vicinity singing by both sexes (12, 259, JMJ; R. Suthers, personal communication). Preliminary analyses of chirr syllables recorded from 32 cardinals (most presumed to be male) from diverse locations in the United States and Mexico showed extensive variation in the duration of chirrs sung by the same individual, and extensive overlap in the durations of chirrs sung by different individuals (SLH and D. A. Enstrom, unpublished data). Two males recorded between 31 March and 13 April of 2013 in east-central Illinois, sang statistically significantly longer chirrs in song exchanges with their mates, and when singing just before or after participating in a territorial pursuit involving another male or pair, and when with their mate when a second female was also present in their territory, than when singing alone, or when singing with another cardinal(s) outside their territory, outside the context of pursuit (SLH and D. A. Enstrom, unpublished data). More quantitative data are needed to compare chirr use in different contexts. Also potentially useful as a fitness indicator is the precision of vocal transitions during continuous sounds that span the ~3.5 kHz boundary at which airflow is redistributed between the two sides of the syrinx (22, 260). Finally, it is not known whether cardinals build up an oxygen debt during long songs of syllables repeated at rates close to the 16 per second minibreath limit; if so, the duration of unusually long songs with rapid syllable repetition rates could also indicate vital capacity (22, 260).
Mechanisms of Sound Production
From Suthers and Goller (22), Suthers (261, 262), and Goller and Riede (26). Sound is produced by the vibration of the medial and lateral labia of each side of the syrinx when the labia are adducted into the bronchial lumen. While this mechanism is presumably widespread among birds, Northern Cardinal is one of the species in which it was established that labial vibration (rather than the vibration of other syringeal membranes) is the source of sound generation. Cardinals use the left side of the syrinx to produce frequencies up to about 3.5 kHz, and the right side to produce higher frequencies; morphologically, the left medial labium is thicker than that on the right side. The majority of cardinal song syllables span the frequency ranges of the two sides, with a greater contribution from the left side for most syllables; the mid-whistle transition of airflow between the two sides is usually undetectable by ear or sound spectrogram. Among syllables produced by only one side of the syrinx, more are produced by the left side than by the right side. At syllable repetition rates up to 16 per second, cardinals take shallow “minibreaths” between syllables, thus prolonging potential song length; during minibreaths, air often flows through the right side of the syrinx at a higher rate than the left side. From Suthers and Goller (22), Fletcher et al (23), Riede et al. (24), and Suthers et al. (27), the size and shape of the oropharyngeal-esophageal cavity, the beak gape, and the tongue position all change significantly during singing to track and selectively enhance the changing fundamental frequency of the song. As fundamental frequency increases between ~2 and 9 kHz, the size of the oropharyngeal-esophageal cavity decreases and the beak gape increases, while the tongue remains in the lower mandible. Below about 3 kHz, the reduced beak gape, likely coupled with the shape of the oropharyngeal-esophageal cavity and tongue position, suppresses the magnitude of the second and third harmonics relative to the fundamental frequency. In experiments in which a singing male cardinal could not reduce his beak gape, the amplitude of the second and third harmonics was increased relative to the fundamental. At fundamental frequencies below ~2 kHz, the oropharyngeal-esophageal cavity remains enlarged, the beak remains nearly closed, and the tongue is elevated to a maximally vertical position in which the tip of the tongue seems to touch the upper palate at the back of the beak; tongue elevation likely shifts the resonance peak of the vocal tract to a lower frequency, but it is not known whether the degree of tongue elevation precisely tracks changes in these lower frequencies. However, the function of the continued increase in beak gape as frequencies rise above 3 kHz is not yet understood. Exceptions to these general trends for some syllable types may be explained by constraints imposed by the rate of frequency modulation or syllable repetition rate; one cardinal consistently produced a syllable with fundamental frequencies between ~0.8 and 2.2 kHz, in which the second harmonic was produced at higher amplitude than the fundamental. Beak gape is positively correlated with overall song amplitude, at least to some extent; thus, lower-frequency sounds or portions of sounds may tend to be sung at lower amplitude (as are chirrs, the syllables cardinals produce whose components are too closely spaced for air to be replenished by minibreaths). In contrast to the aforementioned parts of the vocal tract, tracheal resonance selectively amplifies specific frequencies, but does not appear to change to track changes in song frequency (24).
“A loud snapping of the wings” was heard from males landing next to decoys (live, caged males) placed in their territories (4).