TY - JOUR
AU - Ornelas, Juan, Francisco
AB - Abstract We studied the songs of Wedge-tailed Sabrewings (Campylopterus curvipennis) in six localities from central Veracruz, Mexico, to document structure and variation within and between singing groups in the same geographic region. Wedgetailed Sabrewing songs were acoustically, structurally, and behaviorally complex, rivaling those of other taxa with complex signals. Songs of individual birds were composed of >45 well-differentiated and structurally complex syllables. We found 239 different syllable types across eight recorded singing groups of Wedge-tailed Sabrewings (∼20 syllable types per singing group), with the greatest versatility recorded in hummingbirds to date. The acoustic variation (15 variables) was summarized in three principal components (58% of acoustic variation), in which intragroup variability accounted for most of the observed variation. We found significant differences between and within groups in terms of syllable sharing (Jaccard’s similarity coefficient). Individuals generally shared >50% of syllable types within groups, whereas syllable sharing was <10% between individuals from different groups. The same microgeographic pattern was supported in a UPGMA (unweighted pair-group method with arithmetic mean) analysis where individual songs from each singing group clustered separately. However, songs recorded at the same location differed between seasons, which suggests that this species does not exhibit geographically distinct dialects that are consistent across time. The interplay among this species’ social system, distribution of its floral resources, and microgeographic and temporal variation of its song requires further research. Resumen Estructura y Variación Microgeográfica del Canto de Campylopterus curvipennis en Veracruz, México Estudiamos los cantos de Campylopterus curvipennis en seis localidades del centro de Veracruz, México para documentar la estructura y variación dentro y entre asambleas de canto de la misma región geográfica. Los cantos de C. curvipennis fueron acústica, estructural y conductualmente complejos, rivalizando a otros taxa con cantos complejos. Los sonidos de estos individuos están compuestos por >45 sílabas complejas y bien diferenciadas. Encontramos 239 tipos de sílabas diferentes a través de ocho grupos grabados de C. curvipennis (∼20 tipos de sílabas en cada grupo), con la mayor versatilidad registrada a la fecha para colibríes. La variación acústica (15 variables) fue resumida en tres componentes principales (58% de la variación), en donde la mayor parte de la variación observada se atribuye a variación intragrupo. Encontramos diferencias significativas dentro y entre grupos en los tipos de sílaba compartidos (coeficiente de similitud de Jaccard). Los individuos generalmente compartieron >50% de tipos de sílaba dentro de cada grupo, pero <10% entre aquellos de grupos diferentes. El mismo patrón microgeográfico fue apoyado por un análisis de UPGMA (“unweighted pair-group method with arithmetic mean“), donde los individuos se agruparon por sus cantos de acuerdo a su procedencia. Sin embargo, cantos grabados en el mismo sitio fueron diferentes entre estaciones, lo cual sugiere que esta especie no presenta dialectos geográficamente distintos que son constantes a través del tiempo. La interacción entre el sistema social de esta especie, la distribución de sus recursos florales y la variación microgeográfica y temporal de su canto requieren investigación futura. Hummingbird vocalizations, in their variety, offer many opportunities for studying the function and variability of song advertisement. Males and (in some species) females vocalize in several social situations (Wiley 1971, Gaunt et al. 1994, Ficken et al. 2000, Jarvis et al. 2000); but the acoustic characteristics, structure and organization, function, and evolution of vocal display repertoires in hummingbirds are largely unknown (Kroodsma et al. 1996). The range of variation in hummingbird song among individuals (Wiley 1971, Snow 1977, Gaunt et al. 1994, Ficken et al. 2002), among populations of the same species (Wiley 1971, Atwood et al. 1991), and among species (Gaunt et al. 1994, Ornelas et al. 2002) is poorly known. Careful study of hummingbird vocalizations should yield insights into patterns of vocal development, geographic variation, sexual selection, and a variety of other biological phenomena, such as territoriality (Kroodsma et al. 1996). For a long time, hummingbirds were believed to have relatively simple calls or songs (Johnsgard 1997); however, it is now known that at least some species are song learners (Baptista and Schuchmann 1990, Gaunt et al. 1994, Gahr 2000, Jarvis et al. 2000), and some have complex songs (Ficken et al. 2000, Ornelas et al. 2000). A wide range of complexity in song type occurs within certain genera. For instance, some species of Amazilia have simple songs, whereas other species have songs that are either complex or intermediate in complexity (Kroodsma et al. 1996). Authors have described the songs of the Long-tailed Sabrewings (Campylopterus excellens) from Veracruz and Oaxaca, Mexico (Winkler et al. 1992), and of the Juan Fernández Firecrown (Sephanoides fernandensis) from the Juan Fernández Islands, Chile (Kroodsma et al. 1996), as long, loud, and elaborate. Ficken et al. (2000) documented the complexity, organization, and function of the Blue-throated Hummingbird’s (Lampornis clemenciae) song and suggested that hummingbirds with complex vocalizations tend to inhabit dense vegetation. Ornelas et al. (2002) documented song complexity in its congener, the Amethyst-throated Hummingbird (L. amethystinus), and hypothesized that the evolution of diverse acoustic signals may be favored, independently of habitat characteristics, by the complex foraging interactions that humming-birds encounter daily, which require them to communicate intra- and interspecifically. Snow (1968) documented microgeographic song variation in Little Hermits (Phaethornis longuemareus), in which neighboring males had similar songs that differed from other groups in the same singing assembly. Wiley (1971) confirmed the same pattern and hypothesized that their elaborate songs, displayed in relatively dark locations, may function—in place of conspicuous plumages—as a means of intraspecific communication at long distances and he related them to the differentiation of song groups. Subsequent studies have described similar patterns of microgeographic song variation in other hummingbird species as evidence of song learning (e.g. Snow 1977, Stiles and Wolf 1979, Vielliard 1983, Baptista and Schuchmann 1990, Atwood et al. 1991, Gaunt et al. 1994); but species studied to date have very simple vocalizations, and differences between song groups are slight. The purpose here is to examine acoustic variation in Wedge-tailed Sabrewing (Campylopterus curvipennis) songs in Veracruz, Mexico. We first describe the acoustic characteristics, structure, and contextual use of vocalizations. We then survey the vocal variation at multiple sites in the same geographic region to assess whether microgeographic song variation exists in this species. Methods Study species The Wedge-tailed Sabrewing is a resident, plumage-monomorphic species distributed from San Luis Potosí and Tamaulipas, south through eastern Mexico, and from northeastern Chiapas to the Yucatan Peninsula, Guatemala, and Belize. The species occurs in open woodlands and humid, densely shaded forests, forest edges, and second-growth vegetation; it sometimes also uses more open habitat (Howell and Webb 1995). Usually, the species can be found at low elevations; its range rarely extends above 1,200 m above sea level. Wedge-tailed Sabrewings are typically territorial and aggregate around food resources (Howell and Webb 1995). Birds can sing as solitary individuals or in groups of up to three individuals (Johnsgard 1997). Males sing among dense tangles of vegetation, at heights that vary from near ground-level to high in the canopy (Winker et al. 1992). Songs are sometimes emitted in flight, as the bird goes from perch to perch, and are often preceded by a short trilled call. The song itself is an extended mixture of whistles and gurgles, described as one of the most melodious and versatile songs of any North American hummingbird (Johnsgard 1997, Ornelas et al. 2002). However, no one has conducted a detailed spectrographic study of its song structure. Study areas and recordings Eight groups of birds (hereafter “singing groups“) were recorded at six different locations in central Veracruz, Mexico, between November 2001 and March 2003. Locations of singing groups and dates of recordings are: La Orduña (ORDU: 19°26′N, 96°57′W; 1,150 m above sea level; February to April 2002); Briones (BRIO: 19°30′N, 96°56′W; 1,440 m above sea level; October 2002); Coatepec (COAT: 19°26′N, 96°53′W; 1,180 m above sea level; November 2002); Coapexpan (COAP: 19°31′N, 96°56′W; 1,440 m above sea level; December 2002); Clavijero (19°30′39′′N, 96°56′34′′W; 1,440 m above sea level; November 2001 [CLA1] and November 2002 [CLA2]); Macuiltépetl (19°32′N, 96°55′34′′W; 1,440 m above sea level; November 2002 [MAC1] and February 2003 [MAC2]). Most locations were separated from one another by more than 2 km (see below). Some observations of behavioral contexts associated with vocalizations and frequency of singing activity were taken during the study. Various types of vocalizations and behavioral contexts were detected during our preliminary observations. Here, we focus on full songs, defined as continuous series of vocalizations composed of different syllable types, emitted by individual birds attending singing assemblies. Vocalizations were recorded with a digital Tascam DA-P1 tape recorder (Teac Corporation, Tokyo, Japan) and a Sennheiser MKH-70 shotgun microphone. Songs were digitized and spectrograms generated in CANARY, version 1.2.4 (Charif et al. 1995). Sampling rate was set at 44,100 Hz, and spectrograms were made with a 349.7-Hz filter bandwidth and frame length of 512 points (temporal resolution = 1.45 ms). On any given day, we were able to simultaneously identify at least five individuals based on short-term perch maintenance. For long-term monitoring and individual vocal identification, use of marked individuals is desirable, but marking hummingbirds is difficult and could alter their singing behavior in the assemblies. Even if we could mark birds, they move too fast to be spotted in dense vegetation. Because males are quite similar to females in plumage, we did not distinguish sexes. One hundred and seventy-three vocalizations were recorded and analyzed. Because individual identification of recorded birds was not possible, we used correspondence analysis (CA; Greenacre 1984) on all song recordings from each group separately to provide an estimate of how many different birds contributed to each group’s set of songs. Correspondence analysis is a descriptive technique that produces a graphic representation of relationships among objects (songs, in our case). In a manner somewhat analogous to principal component analysis (PCA), CA produces an economical description of the variation among objects measured with a large number of variables. Instead of using measurement variables, as in PCA, CA operates with a matrix of presence-or-absence variables—in our case, presence or absence of 239 syllable types in recordings (173) made at each site. The usefulness of CA in identifying individual birds was tested using the presence or absence of syllable types as variables on the songs of three Wedge-tailed Sabrewings recorded in April 2004. Those birds, from a different group and part of an ongoing study, were recorded on their individual territories, using the locations of singing birds as clues to individual identity (long-term perch fidelity). The CA successfully identified three aggregations of individual recordings, corresponding to each of the three known individuals, and explained 40.3% of total variance (Fig. 1A). Therefore, our adoption of CA appears to be justified for the analysis of our data. The procedure was then implemented for data from each of the singing groups (Fig. 1B). From the axis 1 and 2 joint plots, we drew conclusions regarding individual identification in each site, on the basis of the clear aggregations of points (recordings) in the dimensional space. Outlier recordings in ordination diagrams were removed from further analyses, and retained if clear aggregations occurred. After excluding individual recordings with uncertain classification (23), we ended up with 150 song recordings corresponding to 32 individuals (Table 1). We used STATISTICA, version 5.5 (Statsoft, Tulsa, Oklahoma), in all CA applications. Table 1. Variables measured in Wedge-tailed Sabrewings’ songs. ‘Syllable versatility index’ is number of syllable types multiplied by number of transitions from one syllable type to another divided by total number of syllables (Kroodsma and Verner 1978, Conner et al. 1986). ‘Syllable transitions’ is number of transitions from one syllable type to another. Data are means ±SE across individuals within each group. Numbers in parentheses are individual birds estimated per singing group Open in new tab Table 1. Variables measured in Wedge-tailed Sabrewings’ songs. ‘Syllable versatility index’ is number of syllable types multiplied by number of transitions from one syllable type to another divided by total number of syllables (Kroodsma and Verner 1978, Conner et al. 1986). ‘Syllable transitions’ is number of transitions from one syllable type to another. Data are means ±SE across individuals within each group. Numbers in parentheses are individual birds estimated per singing group Open in new tab Fig. 1. Open in new tabDownload slide Biplots based on correspondence analysis of (A) individuals of known identity recorded in April 2004 and (B) unknown individuals from Macuiltépetl (MAC1) recorded in November 2002, showing 40% and 58% of the total variance in the data, respectively. We show the plane of axes 1 and 2. Variation explained by each axis is given in parentheses. Symbols stand for recordings from three known individuals and numbers stand for 14 individual recordings from unknown individuals of MAC1 Fig. 1. Open in new tabDownload slide Biplots based on correspondence analysis of (A) individuals of known identity recorded in April 2004 and (B) unknown individuals from Macuiltépetl (MAC1) recorded in November 2002, showing 40% and 58% of the total variance in the data, respectively. We show the plane of axes 1 and 2. Variation explained by each axis is given in parentheses. Symbols stand for recordings from three known individuals and numbers stand for 14 individual recordings from unknown individuals of MAC1 Song structure We examined spectrograms of all song groups (150 spectrograms) and identified syllable types as units by eye, on the basis of distinct tracings on the spectrogram. Syllables were letter-coded and categorized by C.G. for further analysis. To explore the sampling effort, we generated plots of the cumulative number of syllable types detected in songs (150 recordings) of each singing group (Fig. 2). With the aid of ESTIMATES, version 6.0b1 (Colwell 2000), we randomized 1,000× the number of syllable types detected to remove the effect of sampling order. Most of the groups reached the asymptote within the acquired sample size (∼30), but new syllables were still detected at the end of sampling in the CLA2, MAC1, and MAC2 groups. Fig. 2. Open in new tabDownload slide Cumulative number of syllable types found in each singing group against the number of songs or recordings sampled. Samples were randomized 1,000× to remove the effect of sampling order. The ORDU, COAT, BRIO, CLA1, and COAP song groups reached the asymptote with the acquired sample size, but new syllables were still detected at the end of sampling in CLA2, MAC1, and MAC2 groups Fig. 2. Open in new tabDownload slide Cumulative number of syllable types found in each singing group against the number of songs or recordings sampled. Samples were randomized 1,000× to remove the effect of sampling order. The ORDU, COAT, BRIO, CLA1, and COAP song groups reached the asymptote with the acquired sample size, but new syllables were still detected at the end of sampling in CLA2, MAC1, and MAC2 groups To document song structure and acoustic variation within and between singing groups, we measured and estimated 17 variables on each spectrogram and averaged values for each bird (Fig. 3; Table 1). Out of the 17 variables, the “I“ syllable was the most common in the repertoire and appeared in all individuals of every singing group; accordingly detailed measures were taken of it. Acoustic measurements were made using measurement cursors in CANARY (Charif et al. 1995). Fig. 3. Open in new tabDownload slide Fragment of one of the Wedge-tailed Sabrewing songs showing components and descriptive terminology as well as some of the measurements taken of each song Fig. 3. Open in new tabDownload slide Fragment of one of the Wedge-tailed Sabrewing songs showing components and descriptive terminology as well as some of the measurements taken of each song Microgeographic song variation Acoustic variability in Wedge-tailed Sabrewing songs were assessed among and within singing groups in the same geographic region. To eliminate multiple-comparison problems and reduce the dimensionality of descriptions to a smaller system of uncorrelated responses, we conducted PCA with averaged values for each bird (n = 32) of all acoustic measures using STATISTICA. Syllable versatility index and syllables per second were excluded from the model, because the variables used to calculate those indexes are included in the analysis. To improve normality, mensural data were log (x + 1) transformed and counts were square-root (+0.5) transformed before analysis, but untransformed data are reported in tables and figures. After factor scores were calculated, we used a series of nonparametric one-way analysis of variance tests (Kruskal-Wallis ANOVA) to evaluate group variation among acoustic variables using STATVIEW (Abacus Concepts, Berkeley, California). Besides comparing acoustic variables among singing groups, we also compared syllable-sharing among and within groups. With the aid of ESTIMATES (Colwell 2000), we used Jaccard’s similarity coefficient (Sj) to quantify pairwise similarity (496 comparisons) among songs of individual birds (n = 32) from all groups on the basis of presence or absence of syllables in their respective repertoires (presence or absence of syllables registered in any recording of individual birds): Sj = a / (a + b + c) where a is the number of syllables shared by two singing groups (A and B), b is the number of syllables present in the repertoire of the A group but absent in the repertoire of the B group, and c is the number of syllables present in the repertoire of the B group but absent in the repertoire of the A group. In addition, we used the UPGMA (unweighted pair-group method with arithmetic mean) clustering algorithm, with Jaccard’s coefficient as the distance measure, to group individuals from different singing groups on the basis of presence or absence of syllable types. Cluster analysis was performed with the aid of STATISTICA. Finally, the Mantel correlation statistic (Mantel 1967) was computed using R PACKAGE, version 4.0 (Casgrain and Legendre 2001) to test for significant correlation between song similarity among groups and geographic distance. Results Singing behavior The ORDU singing group was spaced out over small adjacent territories in second-growth tangled vegetation near coffee and sugarcane plantations. Each territory (∼20 m2) was defended by one individual. Males usually sang just the introductory song syllable while perched, but they sang the entire song in flight either to or from the territory. Three of those territories were observed for three days by five observers from 0630 to 1930 hours, a total of 52 sessions of 10-min observations with 5-min pauses. Singing frequency was constant throughout the day, with a peak of song activity before noon. The COAT, COAP, and BRIO groups occurred within gardens in towns. At the first two sites (COAT and COAP), Wedge-tailed Sabrewings were mostly situated in areas with large patches of flowering Malvaviscus sp. (Malvaceae); the BRIO group was situated with a Spathodea campanulata (Bignoniaceae) tree. Birds were constantly foraging and, between foraging bouts, they generally sang from hidden territorial perches. We did not observe individuals vocalizing while flying; we did observe some territorial chases. Birds from CLA1 and CLA2 congregated in a botanical garden, surrounded by a cloud forest remnant, specifically in an area containing trees mainly of Dombeya wallichii (Sterculiaceae), an introduced African species with big pink inflorescences with apparently high nectar volumes. We observed those groups during November for two consecutive years. The individuals had the same singing behavior as the three aforementioned groups (COAT, COAP, and BRIO). Birds from MAC1 and MAC2 congregated in a second-growth area composed of big patches of Abutilon hybridum (Malvaceae), surrounded by cloud-forest remnants. Birds were recorded in two seasons but probably constituted the same singing assembly. Singing behavior in that group was more variable than in the rest of the groups. We observed some individuals singing while flying (similar to ORDU song group), whereas others sang while perched; but well-defined territories (as in the ORDU song group) were not detected at Macuiltépetl. Song structure Wedge-tailed Sabrewing songs are long (8–10 s), loud, and high-pitched (>7 kHz), composed of >45 well-differentiated (depending on the singing group), structurally complex syllables sung at a high rate (5 syllables per second) (Table 1). We found 239 different syllable types across the eight singing groups (Fig. 4) and commonly >20 syllable types per individual (Table 1). Acoustic structure of syllables was very complex and variable. Syllables lasted <0.2 s, except for the introductory syllable, which was more variable and of longer duration (0.4–0.7 s) (Table 1). The introductory syllable’s first note differed from the rest of the notes in every singing group, and total number of notes was also variable. The full song covered a wide range of frequencies (7.8 ± 0.2). Songs were very versatile (see Table 1), and syllables were rarely repeated more than once in succession. Songs began with an introductory syllable, which differed among groups (Fig. 5). On average, songs proceeded as a series of 45–52 syllables. Some syllables appeared only once in a sequence; others appeared multiple times. Fig. 4. Open in new tabDownload slide Open in new tabDownload slide Spectrograms (first 2.5 s) of individuals from each singing group, showing the great variation and structural complexity of syllable types, and some of their respective alphabetical codes. Syllables can be separated into one or several notes with diverse structure (including trills, short bursts, harmonics), with rapid or slow frequency modulations, and lacking pure tones. (Continued on next page) Fig. 4. Open in new tabDownload slide Open in new tabDownload slide Spectrograms (first 2.5 s) of individuals from each singing group, showing the great variation and structural complexity of syllable types, and some of their respective alphabetical codes. Syllables can be separated into one or several notes with diverse structure (including trills, short bursts, harmonics), with rapid or slow frequency modulations, and lacking pure tones. (Continued on next page) Fig. 5. Open in new tabDownload slide Complete repertoire of introductory syllables recorded from each singing group. Note that two introductory syllables were recorded from two of the groups (ORDU and MAC1). In all groups, the first note of the introductory syllable is different from the rest of the notes, and the total number of notes is variable Fig. 5. Open in new tabDownload slide Complete repertoire of introductory syllables recorded from each singing group. Note that two introductory syllables were recorded from two of the groups (ORDU and MAC1). In all groups, the first note of the introductory syllable is different from the rest of the notes, and the total number of notes is variable Microgeographic song variation A PCA of 15 acoustic variables yielded three components (eigenvalues > 1.5), which accounted for about one-half (57.7%) the variation recorded in Wedge-tailed Sabrewing songs (Table 2). Principal component 1 (29% of the variance) exhibited positive loadings for total song duration, number of syllables, number of different syllables, and number of transitions from one syllable type to another. Principal component 2 (18.5% of the variance) was mainly explained by duration and number of notes of the introductory syllable. Principal component 3 (10.2% of the variance) represented a pitch element explained by maximum and minimum frequencies (Table 2). Although factor scores differed significantly among groups for PC 3 (Kruskal-Wallis ANOVA; PC 1: χ2 = 3.14, df = 7, P = 0.87; PC 2: χ2 = 10.87, df = 7, P = 0.14; PC 3: χ2 = 18.77, df = 7, P = 0.008), the lack of group discrimination in the PC 1 versus PC 2 scatterplot indicates that intragroup variability accounted for much of the observed acoustic variation (Fig. 6). Table 2. Principal component analysis of the correlation matrix among 15 acoustic variables in eight singing groups of Wedge-tailed Sabrewings in Veracruz, Mexico. Correlation loadings above 0.50 are in boldface Open in new tab Table 2. Principal component analysis of the correlation matrix among 15 acoustic variables in eight singing groups of Wedge-tailed Sabrewings in Veracruz, Mexico. Correlation loadings above 0.50 are in boldface Open in new tab Fig. 6. Open in new tabDownload slide Bivariate plot of factor scores produced by a principal component analysis of 15 acoustic variables of Wedge-tailed Sabrewing song Fig. 6. Open in new tabDownload slide Bivariate plot of factor scores produced by a principal component analysis of 15 acoustic variables of Wedge-tailed Sabrewing song Jaccard’s similarity coefficient averaged 0.12 ± 0.04 (mean ± SD) between-group comparisons (n = 446) and 0.45 ± 0.19 within-group comparisons (n = 50). The UPGMA analysis clearly clustered 32 individuals into seven groups based on syllable sharing (Fig. 7). Clustering at a higher level was less apparent in the analysis. One individual of the CLA2 singing group was separated from the cluster, and individuals of MAC1 and MAC2 (which may have included some of the same individuals) were found intermixed in one cluster. We found common syllables present in the song of all singing groups. The “I“ syllable was the most frequently emitted (8× per song on average, range = 1–24). The “C“, “D“, “P“, “I(a)“, and “I(e)“ syllables were also present in the songs of every group, but not as commonly emitted. The “I(a)“ and “I(e)“ syllables, along with other syllables, seem variants of the “I“ syllable, and their presence varied among groups (Fig. 8). No significant correlation was found between song similarity and geographic distances between sites (Mantel test: r = 0.215, P = 0.12). Fig. 7. Open in new tabDownload slide Average Euclidean distance between clusters. Dendrogram based on 32 individual songs distinguished seven principal clusters corresponding to eight singing groups recorded. Note that MAC1 and MAC2 are in the same cluster, so they probably belong to the same singing group. The UPGMA cluster analysis revealed clear patterns of vocal similarity within groups based on shared syllable types. Note also that ORDU is clustering outside of all the other groups, except for one bird from CLA2, probably because of differences in the contextual use of vocalizations; however, a better representation of the flight context is necessary for comparisons Fig. 7. Open in new tabDownload slide Average Euclidean distance between clusters. Dendrogram based on 32 individual songs distinguished seven principal clusters corresponding to eight singing groups recorded. Note that MAC1 and MAC2 are in the same cluster, so they probably belong to the same singing group. The UPGMA cluster analysis revealed clear patterns of vocal similarity within groups based on shared syllable types. Note also that ORDU is clustering outside of all the other groups, except for one bird from CLA2, probably because of differences in the contextual use of vocalizations; however, a better representation of the flight context is necessary for comparisons Fig. 8. Open in new tabDownload slide Variants of the “I“ syllable in the songs of Wedge-tailed Sabrewing singing groups that suggest song learning Fig. 8. Open in new tabDownload slide Variants of the “I“ syllable in the songs of Wedge-tailed Sabrewing singing groups that suggest song learning Discussion Song complexity and contextual use Wedge-tailed Sabrewing songs are structurally, acoustically, and behaviorally complex, rivaling those of passerine birds in the complexity of their signal. In contrast to the songs of other hummingbird species that produce complex acoustic signals (e.g. Blue-throated and Amethyst-throated hummingbirds; Ficken et al. 2000, Ornelas et al. 2002), Wedge-tailed Sabrewing songs are loud and long, composed of very discrete units (syllables), with a highly variable and elaborate acoustic structure. We observed Wedge-tailed Sabrewings in two ecological contexts during our study. First, several individuals were observed aggregated in areas containing flower patches, singing for long periods between foraging bouts. We believe that the observed territorial chases may be associated with defense of floral resources, of territorial singing perches, or of both. Second, individuals aggregated in areas of dense vegetation, formed by small adjacent territories. Territories were defended by one individual (presumably a male) that emitted the full song in flight when an intruder initially arrived and after it arrived on the territory. Although we observed no attempted copulations, this singing behavior appears to have an intersexual display function. Because the territories did not contain any resources required by females (except perhaps males), and because they were close to each other (∼7 m), allowing acoustic contact between neighboring territory-owners, we believe that the males’ singing and territorial behavior can be characterized as a typical lek (Payne 1984), in which the function of vocalizations remains to be investigated. The existence of elaborate acoustic signals in Wedge-tailed Sabrewings is intriguing, because most lekking species studied in detail have simple, persistent, loud songs, often audible over long distances (>100 m) (Wiley 1971, Snow 1977, Stiles and Wolf 1979, Atwood et al. 1991, Gaunt et al. 1994). Wedge-tailed Sabrewing songs have the highest versatility among hummingbirds recorded to date (Ornelas et al. 2002; Table 1). Songs of other species in assemblies or leks consist of a single phrase, repeated without change for long periods (Stiles and Wolf 1979). We believe that the observed variability may serve not only for individual or group recognition, but to establish more complicated communication networks that are not immediately apparent to human observers. Microgeographic song variation The acoustic characteristics (15 variables) of full songs of individuals from six localities did not differ significantly among singing groups. Songs from the same locality did not cluster together in a scatterplot of the first and second components from the PCA analysis. However, the singing groups differed markedly in general song structure and structure of the introductory syllable. Also, syllable-sharing between singing groups was significantly lower than syllable-sharing within groups (UPGMA cluster analysis). That indicates that the songs of individuals, presumably males, vary considerably, and that birds from the same singing group tend to have similar songs. We found syllables that occurred in each singing group, and the “I(a)“ and “I(e)“ syllables, along with others, seem variants of the “I“ syllable. The variation among those syllabic elements as well as the observed variability in the structure and complexity of vocalizations within or across birds of Wedge-tailed Sabrewing singing groups can be considered evidence of song learning. Local dialects arise when shared songs differ between neighboring groups (Marler and Tamura 1962, Harbison et al. 1999), but their functional significance remains widely debated (Catchpole and Slater 1995 and references therein). In central Veracruz, singing groups gradually disintegrated as floral resources were diminished, and groups recorded in different years in Clavijero and Macuiltépetl shared few syllables across years and, therefore, were probably composed of different individuals. Little information exists on how hummingbirds disperse, alone or in groups, or how they track floral resources through space and over time (Hobson et al. 2003). It is impossible to know whether individuals retain the same songs from year to year; however, it is likely that different groups of individuals, with different songs, assemble in the same places each year, or that the same individuals develop different songs in each group they join, or both. That songs recorded at the same location were different over two different seasons suggests that this species does not exhibit geographically distinct dialects that are consistent across time. In polygynous Yellow-rumped Caciques (Cacicus cela), males in any nesting colony learn many vocalizations from each other and can change them from year to year (Trainer 1988, 1989). At the same time, groups can have “signature“ notes, shared by all members of a group in any one year, like the introductory syllables of the Wedge-tailed Sabrewings. Also, we did not find any relationship between song similarity and geographic distance among song groups, and the clustering of song types (assessed by UPGMA cluster analysis) did not reflect geographic distances among sites. Therefore, the existence of song dialects with consistency across time and well-established boundaries among singing groups in Wedge-tailed Sabrewings cannot be claimed until more data become available (greater geographic scale). Nevertheless, the interplay among this species’ social system, distribution of its floral resources, and microgeographic and temporal variation of its songs are subjects for future research. Acknowledgments Comments by M. S. Ficken, A. Guillén, M. E. Mermoz, D. A. Nelson, S. G. Sealy, K. G. Smith, and three anonymous reviewers greatly improved a previous version of this paper. We gratefully acknowledge L. Jiménez, A. Castillo, A. L. Castillo, M. Ordano, and C. Lara for field assistance, and A. Espinosa de los Monteros, A. Hernández, D. Hernández, and C. Pytte for assistance in preliminary stages of our study. Thanks to C. Rojas Nieto for helpful and enlightening discussion. This work was supported with research grants from the Consejo Nacional de Ciencia y Tecnología, México (CONACyT) (Ref. 25922-N) and Comisión para el Conocimiento y Uso de la Biodiversidad (CONABIO) (Ref. H028) to J.F.O., and the Instituto de Ecología, AC (Ref. 902-11-563). Literature Cited Atwood , J. L. , V. L. Fitz , and J. E. Bamesberger . 1991 . Temporal patterns of singing activity at leks of the White-bellied Emerald. Wilson Bulletin 103 : 373 – 386 . WorldCat Baptista , L. F. , and K. L. Schuchmann . 1990 . Song learning in the Anna Hummingbird (Calypte anna). Ethology 84 : 15 – 26 . Google Scholar Crossref Search ADS WorldCat Casgrain , P. , and P. Legendre . 2001 . The R PACKAGE for Multivariate and Spatial Analysis, version 4.0 d6. User’s Manual. [Online.] Département de Sciences Biologiques, Université de Montréal. Available at www.fas.umontreal.ca/BIOL/legendre&sol; . WorldCat Catchpole , C. K. , and P. J B. Slater . 1995 . Bird Song: Biological Themes and Variations. Cambridge University Press, London. WorldCat Charif , R. A. , S. Mitchell , and C. W. Clark . 1995 . CANARY 1.2.1 User’s Manual. Cornell Laboratory of Ornithology, Ithaca, New York. WorldCat Colwell , R. K. 2000 . ESTIMATES: Statistical Estimation of Species Richness and Shared Species from Samples, version 6.0b1. User’s Guide and application. [Online.] Available at viceroy.eeb.uconn.edu/estimates. WorldCat Conner , R. N. , M. E. Anderson , and J. G. Dickson . 1986 . Relationships among territory size, habitat, song, and nesting success of Northern Cardinals. Auk 103 : 23 – 31 . WorldCat Ficken , M. S. , K. M. Rusch , S. J. Taylor , and D. R. Powers . 2000 . Blue-throated Hummingbird song: A pinnacle of nonoscine vocalizations. Auk 117 : 120 – 128 . Google Scholar Crossref Search ADS WorldCat Ficken , M. S. , K. M. Rusch , S. J. Taylor , and D. R. Powers . 2002 . Reproductive behavior and communication in Blue-throated Hummingbirds. Wilson Bulletin 114 : 197 – 209 . Google Scholar Crossref Search ADS WorldCat Gahr , M. 2000 . Neural song control system of hummingbirds: Comparison to swifts, vocal learning (songbirds) and nonlearning (suboscines) passerines, and vocal learning (budgerigars) and nonlearning (dove, owl, gull, quail, chicken) nonpasserines. Journal of Comparative Neurology 426 : 182 – 196 . Google Scholar Crossref Search ADS PubMed WorldCat Gaunt , S. L L. , L. F. Baptista , J. E. Sánchez , and D. Hernandez . 1994 . Song learning as evidenced from song sharing in two hummingbird species (Colibri coruscans and C. thalassinus). Auk 111 : 87 – 103 . Google Scholar Crossref Search ADS WorldCat Greenacre , M. J. 1984 . Theory and Applications of Correspondence Analysis. Academic Press, London. WorldCat Harbison , H. , D. A. Nelson , and T. P. Hahn . 1999 . Long-term persistence of song dialects in the Mountain White-crowned Sparrow. Condor 101 : 133 – 148 . Google Scholar Crossref Search ADS WorldCat Hobson , K. A. , L. I. Wassenaar , B. Mil&aacute , I. Lovette , C. Dingle , and T. B. Smith . 2003 . Stable isotopes as indicators of altitudinal distributions and movements in an Ecuadorean hummingbird community. Oecologia 136 : 302 – 308 . Google Scholar Crossref Search ADS PubMed WorldCat Howell , S. N G. , and S. Webb . 1995 . A Guide to the Birds of Mexico and Northern Central America. Oxford University Press, New York. WorldCat Jarvis , E. D. , S. Ribeiro , M. L. Da Silva , D. Ventura , J. Vielliard , and C. V. Mello . 2000 . Behaviourally driven gene expression reveals song nuclei in hummingbird brain. Nature 406 : 628 – 632 . Google Scholar Crossref Search ADS PubMed WorldCat Johnsgard , P. A. 1997 . The Hummingbirds of North America, 2nd ed. Smithsonian Institution Press, Washington, D.C. WorldCat Kroodsma , D. E. , and J. Verner . 1978 . Complex singing behaviors among Cistothorus wrens. Auk 95 : 703 – 716 . WorldCat Kroodsma , D. E. , J. M E. Vielliard , and F. G. Stiles . 1996 . Study of bird sounds in the Neotropics: Urgency and opportunity. Pages 269–281 in Ecology and Evolution of Acoustic Communication in Birds (D. E. Kroodsma and E. H. Miller, Eds.). Cornell University Press, Ithaca, New York. WorldCat Mantel , N. A. 1967 . The detection of disease clustering and a generalized regression approach. Cancer Research 27 : 209 – 220 . Google Scholar PubMed WorldCat Marler , P. , and M. Tamura . 1962 . Song “dialects“ in three populations of White-crowned Sparrows. Condor 64 : 368 – 377 . Google Scholar Crossref Search ADS WorldCat Ornelas , J. F. , C. González , and J. Uribe . 2002 . Complex vocalizations and aerial displays of the Amethyst-throated Hummingbird (Lampornis amethystinus). Auk 119 : 1141 – 1149 . Google Scholar Crossref Search ADS WorldCat Payne , R. B. 1984 . Sexual Selection, Lek and Arena Behavior, and Sexual Size Dimorphism in Birds. Ornithological Monographs, no. 33. WorldCat Snow , B. K. 1977 . Comparison of the leks of Guy’s Hermit Hummingbird Phaethornis guy in Costa Rica and Trinidad. Ibis 119 : 211 – 214 . Google Scholar Crossref Search ADS WorldCat Snow , D. W. 1968 . The singing assemblies of Little Hermits. Living Bird 7 : 47 – 55 . WorldCat Stiles , F. G. , and L. L. Wolf . 1979 . Ecology and Evolution of Lek Mating Behavior in the Long-tailed Hermit Hummingbird. Ornithological Monographs, no. 27. WorldCat Trainer , J. M. 1988 . Singing organization during aggressive interactions among male Yellow-rumped Caciques. Condor 90 : 681 – 688 . Google Scholar Crossref Search ADS WorldCat Trainer , J. M. 1989 . Cultural evolution in song dialects of Yellow-rumped Caciques in Panama. Ethology 80 : 190 – 204 . Google Scholar Crossref Search ADS WorldCat Vielliard , J. 1983 . Catálogo sonográfico dos cantos e piados dos beija-flores do Brasil, 1. Boletim do Museu de Biologia “Mello Leitão,“ Serie Biologia 58 : 1 – 20 . WorldCat Wiley , R. H. 1971 . Song groups in a singing assembly of Little Hermits. Condor 73 : 28 – 35 . Google Scholar Crossref Search ADS WorldCat Winker , K. , M. A. Ramos , J. H. Rappole , and D. W. Warner . 1992 . A note on Campylopterus excellens in southern Veracruz, with a guide to sexing captured individuals. Journal of Field Ornithology 63 : 339 – 343 . WorldCat © The American Ornithologists' Union, 2005
TI - Song Structure and Microgeographic Song Variation in Wedge-Tailed Sabrewings (Campylopterus Curvipennis) in Veracruz, Mexico
JF - Auk: Ornithological Advances
DO - 10.1093/auk/122.2.593
DA - 2005-04-01
UR - https://www.deepdyve.com/lp/oxford-university-press/song-structure-and-microgeographic-song-variation-in-wedge-tailed-t3ghcuxqlY
SP - 593
VL - 122
IS - 2
DP - DeepDyve
ER -