Arthropod Resource Partitioning Among Omnivorous Tanagers (Tangara Spp.) In Western Ecuador

By Naoki, Kazuya

ABSTRACT.-

The distribution and abundance of food are primary factors affecting resource-use patterns in birds. Many bird species eat several food types, which may differ in their distribution and overall abundance. I studied foraging ecology of seven sympatric species of Tangara at Mindo, Ecuador, to determine whether the patterns of resource use differed between two food types: arthropods and fruits. Interspecific differences in arthropod-foraging were manifested in the fine segregation of microhabitat preference combined with different habitat use. By contrast, interspecific differences in fruit-foraging were manifested in preferences for different plant genera, often associated with different habitats. No evidence was found for spatial partitioning of the same fruit species. Interspecific overlap in fruit-for-aging was 3 higher than that in arthropod-foraging, and species of Tangara that frequently joined the same mixed-species flocks differed largely in arthropod- foraging but overlapped greatly in fruit-foraging. The differences in patterns between arthropod and fruit-foraging may be explained by the different characteristics of arthropods and fruits as food resources. High sympatry of species of Tangara and other omnivorous tanagers, in general, appears to be maintained not because fruits are abundant, resulting in little competition for them, but because these tanagers specialize on different microhabitats for foraging arthropods.

Received 22 July 2003, accepted 7 February 2006.

Key words: arthropods, Ecuador, foraging ecology, omnivorous tanagers, Tangara.

Segregacin de Artrpodos como Recurso Alimenticio de Tangaras Omnvoras (Tangara spp.) en el Oeste de Ecuador

RESUMEN.-La distribucin y abundancia de los alimentos son factores principales que afectan los patrones del uso de recursos de las aves. Muchas especies de aves comen varios tipos de alimentos que podran diferir en su distribucin y abundancia. Estudi la ecologa de forrajeo de siete especies simptricas del gnero Tangara en Mindo, Ecuador, para determiner si los patrones de uso del recurso difirieren entre dos tipos de alimentos: artrpodos y frutos. La diferencia interespecfica en el forrajeo de artrpodos se manifest por la fina segregacin en la preferencia del microhbitat combinada con el uso del diferentes hbitats. En el contraste, la diferencia interespecfica en el forrajeo de frutos se manifest por las preferencias por diferentes gneros de plantas, a menudo asociado con los diferentes habitats. No se encontr ninguna evidencia en la segregacin espacial de las mismas especies de frutos. El traslape interespecfico en el forrajeo de frutos fue tres veces mayor que el traslape en el forrajeo de artrpodos, y las especies de Tangara que frecuentemente se encontraban en las mismas bandadas mixtas difirieron mayormente en el forrajeo de artrpodos pero se solaparon en el forrajeo de frutos. Las diferencias entre el forrajeo de artrpodos y el de frutos podran explicarse por las diferentes caractersticas de artrpodos y frutos como recursos alimenticios. La alta simpatra de Tangara y otras Tangaras omnvoras, en general, parece ser mantenida no porque los frutos sean abundantes y causen poca competencia, sino porque estas tangaras se especializan en diferentes microhbitats para forrajear artrpodos.

(ProQuest Information and Learning: … denotes formulae omitted.)

PATTERNS OF RESOURCE partitioning among sympatric species have been one of the central issues in community ecology (e.g., Schoener 1974, Diamond 1978, Tilman 1982, Connell 1983, Grace and Tilman 1990). Since MacArthur’s (1958) study of sympatric warblers (Dendroica spp.), many avian community ecologists have focused on differences in foraging ecology among closely related species, particularly congeners (e.g., Parus [Lack 1971], Nectarinia [Gill and Wolf 1978], Geospiza [Grant 1986], Phylloscopus [Price 1991], Aerodramus [Lourie and Tompkins 2000]). Because congeners share a large part of their evolutionary histories and presumably have similar morphological, behavioral, and physiological characters, sympatric congeners tend to have more intense interspecific interaction than sympatric noncongeners. Therefore, the elucidation of resource-use patterns among sympatric congeners is more likely to reveal important aspects of species coexistence (Tokeshi 1999).

Most previous studies of avian communities have focused on the partitioning of one type of food resource, such as microhabitat preference in arthropod-foraging, the time spent visiting different flower species, the composition of fruit species found in feces, and the arthropod taxa in stomach contents (e.g., Gill and Wolf 1978, Holmes et al. 1979, Sherry 1984, Loiselle and Blake 1990, Remsen 1990). These studies showed that the distribution and abundance of food generally affected the distribution and abundance of the birds that fed on them, and that the sympatric species partitioned the resource at several different levels: food size, food type, microhabitat, habitat, tree species, or a combination. Many bird species, however, eat more than one food type to various degrees. This is especially true for so-called “frugivorous” birds, most of which supplement their fruit diet with protein-rich foods, such as seeds, insects, and vertebrates (Moermond and Denslow 1985, Remsen et al. 1993). Many are also called “omnivorous” or “frugivorousinsectivorous” (Buskirk 1976, Blake and Loiselle 2000). If the distribution and abundance of a resource influences the way birds partition it, how do omnivorous species feeding on more than one food type partition them? Do they use two food types differently, according to the distribution and abundance of each food type? Does one food type have stronger influence than the other on the structuring of the community? If so, what are the main characteristics of a resource that gives it a stronger influence on community structure?

To answer these questions, I studied the foraging ecology of seven sympatric species of Tangara at Mindo, Ecuador. These tanagers are small-bodied, canopy-dwelling passerines endemic to the Neotropics (Isler and Isler 1999). The genus Tangara contains 50 species, more than any other avian genus in the New World (Stotz et al. 1996). Up to 10 species are found sympatric in the same Andean cloud forest, and as many as 9 species travel together in the same mixed-species flocks (Hilty et al. 1986, Isler and Isler 1999). Although they feed on a wide variety of food items, fruits and arthropods constitute most of their diet (Snow and Snow 1971, Hilty 1977). Many species of Tangara are “colorful” and “conspicuous”; however, only a few species have been studied intensively, and the biology of many species is poorly known (Isler and Isler 1999). At a community level, only two studies have adequately sampled the foraging ecology and resource-use patterns of this diverse group (Snow and Snow 1971, Rodrigues 1995). These studies surveyed all sympatric tanagers of the family Thraupidae but were conducted in relatively species-poor Trinidad and southeastern Brazil, where only three species of Tangara were found.

The first objective here was to examine foraging patterns among sympatric species of Tangara in two food types: arthropods and fruits. I applied multivariate techniques to analyze various foraging parameters simultaneously and to elucidate the principal axes in foraging ecology of each food type. The second objective was to test whether interspecific foraging differences result in resource partitioning among these species. For this, I used a Monte Carlo method to compare the observed interspecific foraging overlap with one expected by chance in the absence of species interaction.

METHODS

Study site.-The study was conducted in the vicinity of Mindo, province of Pichincha, Ecuador (002′S, 7846′W). Mindo is a small village on the western slope of the Andes at 1,250 m elevation. The area corresponds to the transitional zone from foothill forest to subtropical montane forest (Ridgely and Greenfield 2001; from Upper Tropical to Middle Montane elevational zones in Stotz et al. 1996). As a result, Mindo possesses an extremely rich avifauna that ranges from lowland humid forest to Andean cloud forest; >360 bird species have been recorded in this area (Ridgely and Greenfield 2001). The vegetation around Mindo consists of a mosaic of secondary forest and patches of pasture, though some large trees >25 m high and remnants of primary forest are also present. The area is used as a buffer zone for the Mindo-Nambillo protected forest, which preserves 19,200 ha of primary forest. Rainfall averages 2,688 562 mm year1 (n = 13 years), and annual mean temperature is 20.3 0.2C (n = 11 years; unpublished data from Instituto Nacional de Meteorologia e Hidrologia). The dry season lasts from mid-May to mid-December.

Foraging behavior was quantified along a 10-km trail southeast of the village of Mindo (1,300-1,600 m) during the dry season, between 1 June and 15 December 1999. I used an additional 4-km trail for observations after 14 August. This second trail was at a private farm 1 km west of Mindo. Although the elevation of the farm was similar to that of the first trail (1,300-1,50\0 m), the vegetation was more disturbed and had more plant species from lower elevations than the first trail (K. Naoki pers. obs.). At this second site, I found higher densities of species of Tangara typical of foothill forests (Tangara ruflgula, T. gyrola, and T. icterocephala).

Foraging data.-Oata were collected all day between 0600 and 1800 hours. I located birds by sight and sound while slowly walking along a trail. I observed individual birds through 10-40x binoculars and recorded foraging behavior by using a microcassette recorder. At every encounter, I recorded only the first foraging attempt per individual bird to avoid sequential observations and serial correlation problems in data analyses (Hejl et al. 1990, Martin and Bateson 1993). For each foraging observation, I recorded the following foraging parameters: “food item,”"attack maneuver,”"substrate type,”"substrate size,”"perch diameter,”"perch angle,”"foliage density,”"height above ground,”"distance to canopy,”"horizontal position,” and “habitat.” Each food item was classified as a fruit, arthropod, nectar, flower bud, or Mullerian body. But in the multivariate analyses, I used only arthropod and fruit observations because of small sample sizes of the other food items. Substrate categories for arthropod-foraging were (1) moss or thickly moss-covered branch, (2) partially moss-covered branch, (3) bare branch, (4) dead branch, (5) live leaf, (6) dead leaf, (7) flower bud, and (8) air. For fruit-foraging, I recorded plant species instead of substrate type, though in analyses I used genera instead of species. Substrate size and perch diameters were estimated in relation to a bird’s body size and later calculated by using measurements taken from live birds. Foliage density was measured to the nearest 10% in a 1-m-diameter sphere around the bird, following the method of Remsen and Robinson (1990). For horizontal position, I used four categories: three parts of a tree (inner, middle, and foliage) and air (outer). I calculated vertical position as height above ground/(height above ground + distance to canopy). Classification and nomenclature of attack maneuver and perch angle followed Remsen and Robinson (1990). Habitat was categorized as (1) primary forest, (2) secondary forest, (3) semiopen, (4) scrub, and (5) orchard-garden. Primary and secondary forest were further divided into forest interior, forest canopy, and forest edge. “Semiopen” was recently cleared land with some tall trees left to provide shade for cattle. “Scrub” is young secondary growth composed of grasses and trees <5 m.

Data analyses.-For multivariate analyses, I used nine foraging parameters: (1) attack maneuver, (2) substrate type, (3) perch diameter, (4) perch angle, (5) foliage density, (6) height above ground, (7) vertical position, (8) horizontal position, and (9) habitat. These parameters included both categorical and continuous variables. The continuous variables-such as perch diameter, height above ground, and distance to canopywere grouped into four to six categories. This allowed analysis of all the foraging parameters simultaneously with multivariate techniques. Each foraging category was expressed as a proportion of total foraging observations for that species. This standardization eliminated problems arising from unequal sample size when applying ordination techniques (Loiselle and Blake 1990). I used correspondence analysis (CA) to find the principal foraging parameters that explained most of the foraging variation. Correspondence analysis has been shown to be the preferred method for analyzing categorical foraging data, because it recovers more variation from the original data sets and is more consistent in magnitude and sign of the coefficients from eigenvectors than other multivariate methods, such as principal component analysis and factor analysis (Miles 1990). When an interspecific difference was unclear in the first three CA axes, I analyzed the other axes to see whether any interspecific difference was explained by other foraging parameters.

After conducting multivariate analyses, I tested whether sympatric species of Tangara used different parts of the same fruiting trees to partition fruit resources, as found in other frugivorous bird communities (e.g., Terborgh and Diamond 1970, Rodrigues 1995). For this purpose, I used the subset of fruit- foraging data from Ecuador: foraging observations on the two most commonly eaten fruit species, Miconia brevitheca and Trema micrantha, which together accounted for two-thirds of all the fruit- foraging observations. I conducted a multivariate analysis of variance on foraging height and vertical position of each of the two fruit genera by using species of Tangara as independent variables.

To test whether interspecific foraging differences result in resource partitioning, I constructed three-way tables composed of species and two principal foraging parameters, habitat and substrate type in arthropod-foraging, and habitat and plant genus in fruit- foraging, and calculated Pianka’s overlap index (Krebs 1999):

…

where P^sub ij^. is the proportion that resource i is of the total resources used by species j, P^sub ik^ is the proportion that resource i is of the total resources used by species k, and n is the total number of resource states. Pianka’s index varies between 0 and 1, in which O indicates no overlap in resource use and 1 indicates complete overlap.

I used a Monte Carlo method (Gotelli and Entsminger 2006) to test the null hypothesis that species of Tangara used arthropod and fruit resources randomly without influence of interspecific interaction. Rejection of the null hypothesis would indicate that sympatric species of Tangara avoid the resources used by the other species of Tangara (i.e., the observed foraging overlap was smaller than one expected by chance), or that they prefer the resources used by the other species (i.e., the observed foraging overlap was larger than one expected by chance). A null distribution was generated by 1,000 permutations using ECOSIM (Gotelli and Entsminger 2006) with randomization algorithm 2 (RA2). Randomization algorithm 2 assumes that the resource states unused by a species are unavailable for the species even in the absence of species interactions. Therefore, in each randomization, RA2 substitutes a random number for the resources, whose values are not zero (Gotelli and Entsminger 2006).

RESULTS

During six months, I observed 11 species of Tangara. Two of them, T. vassorii and T. heinei, were recorded only a few times. Tangara rufigula, T. gyrola, and T. icterocephala were uncommon in Mindo and were mostly limited to a highly disturbed, drier, lower part of the observation road <1,400 m above sea level. Tangara cyanicollis was essentially a solitary nonforest species and was mostly found in pairs in a semi-open area. The other five species-T. arthus, T. labradorides, T. nigroviridis, T. parzudakii, and T. ruficervix- wete common in tall wet forests at higher elevations and were often found in the same mixed-species flock (K. Naoki unpubl. data). These five species, with . cyanicollis and T. rufigula, yielded the bulk of my 1,340 foraging observations (Table 1).

Arthropod-foraging.-Correspondence analysis of arthropod- foraging showed that the first two axes accounted for 84% of the total variance (Fig. 1). These two axes were heavily weighted by three foraging parameters: attack maneuver, substrate type, and horizontal position, which represented microhabitat preference and together contributed to 70% of CA’s partial contribution. The remaining axes were weighted by substrate type and attack maneuver, but the fourth axis was weighted by habitat. Axes 1, 2, and 3 separated seven species of Tangara to four species groups based on microhabitat preference (only axes 1 and 2 shown in Fig. 1, because axis 3 did not make further separation): (1) Tangara arthus and . parzudakii used thick branches at interior and middle position by reach-down and hangdown maneuvers; (2) T. nigroviridis foraged on thin branches in foliage by using glean and hangdown; (3) T. labradorides and T. rufigula searched leaf surfaces by using glean and reach-up; and (4) . cyanicollis and T. ruficervix mostly sallied into air (Fig. 2). Axis 4 separated these pairs of species of Tangara along a forest-nonforest habitat gradient and assigned T. cyanicollis as a nonforest habitat user (Figs. 1 and 2).

Species found close together in the first three axes, T. parzudakii-T. arthus and T. rufigula-T. labradorides, were separated in other axes. Tangara parzudakii and T. arthus reached the maximum distance in axis 5, which was weighted by moss and bare-branch substrate. Tangara parzudakii searched moss and partially moss- covered branch in 94% of the foraging observations in contrast to 63% in T. arthus (difference highly significant in a G-test of independence with William’s correction with 2×2 contingency table; Gadj = 32.0, P < 0.0001; Fig. 2). Tangara parzudakii often probed into thick moss or pulled away pieces of moss (30%; Fig. 2). These subsurface maneuvers and substrate manipulations were rare for T. arthus, which usually searched the surface of mossy or bare branches (Fig. 2). Tangara rufigula and T. labradorides reached the maximum distance in axis 6, which was weighted by the hang-upside-down attack maneuver. Tangara rufigula used significantly more acrobatic attack maneuvers, such as hang-upside-down and sally, than T. labradorides (22% vs. 7%, G^sub adj^ = 5.8, P < 0.05; Fig. 2). This difference in attack maneuver possibly reflected a finer difference in substrate use, which was not included in the correspondence analysis. Tangara rufigula used leaf undersurfaces significantly more often than T. labradorides (76% vs. 53%, G^sub adj^ = 3.9, P < 0.05). In addition, T. rufigula caught arthropods from the large leaves of Cecropia gabrielis significantly more often than T.labradorides (24% vs. 3%, Gadi = 7.4, P < 0.01). In Mindo, T. labradorides was found mainly in tall, wet forests at higher elevations, whereas T. rufigula was found in disturbed, drier forests at lower elevations; thus, they were rarely found in the same mixed-species flocks. Furthermore, these two species differ largely in elevational distribution: T. labradorides is found in the subtropical zone between 1,300 and 2,000 m, whereas T. rufigula is found in the foothills between 500 and 1,400 m (Ridgely and Greenfield 2001).

Differences in other foraging parameters reflected the differences in microhabitat preference. For example, T. arthus and T. parzudakii, which fed at the inner and middle horizontal positions, used, on average, larger perch diameter and lower foliage density than the other species of Tangara that fed at foliage and the outer horizontal position (Table 2). Tangara labradorides and T. rufigula, which mostly searched leaves for arthropods, foraged at the highest average foliage densities (Table 2). The most acrobatic T. rufigula used horizontal perches less often than the others. Tangara ruficervix, which sallied into the air in the forest canopy, more often used higher branches for foraging.

Fruit-foraging.-The analysis of fruit-foraging was conducted in the same manner as that for arthropod-foraging. The first three axes explained 83% of the total variance (Fig. 3). All axes except axis 5 were weighted by different plant genera. This foraging parameter alone explained 43% of the partial contribution, followed by attack maneuver (16%) and habitat (13%). Besides fruit genera, axes 1, 2, and 5 were also weighted by attack maneuver, and axes 2, 3, and 4 by habitat. Axis 1 separated T. arthus, T. parzudakii, and T. ruficervix, with a high proportion of Cecropia fruit in their diet and more frequent use of the hang-side attack maneuver, from T. labradorides and T. nigroviridis, which had a high proportion of Miconia fruits and glean-attack maneuver (Fig. 4). Axis 2 separated nonforest T. cyanicollis from the other species (Figs. 3 and 4). Axis 3 separated T. rufigula, with a higher proportion of Palicourea and Bocconia fruit, from the other species (Figs. 3 and 4). Tangara labradorides and T. nigroviridis were separated to the greatest degree in axis 5, which was weighted by reach-down attack maneuver. Tangara labradorides used reach-out and reach-down more often than T. nigroviridis, though the difference was not significant (G-test of independence with William’s correction; Gad. = 2.5, P = 0.11). Tangara arthus, T. parzudakii, and T. ruficervix achieved the most segregation in axes 4 and 6.

The differences in some other foraging parameters reflected preferences for different fruit genera. Tangara arthus, T. parzudakii, and T. ruficervix, which often fed on Cecropia fruits by perching on the thick fruit, used thicker perches (Table 2) and vertical perching position more often. Tangara ruficervix, with a high percentage of Trema micrantha in its diet, showed, on average, higher foraging height, which probably reflected the fact that most Trema micrantha were >15 m in this area. Otherwise, few interspecific differences were found in foliage density, vertical position, and horizontal position.

Although interspecific differences were observed in fruit- foraging, seven species of Tangara did not differ significantly in foraging height or vertical position when they fed on the same fruit species: Miconia brevitheca and Trema micrantha (Table 3). Miconia brevitheca alone accounted for nearly half or more of all fruits consumed (48 17%; range from 33% in T. ruficervix to 77% in T. nigroviridis; Fig. 4), and M. brevitheca and T. micrantha combined accounted for nearly two-thirds or more of all fruit-foraging observations (63 19%, range: 44-94%; Fig. 4).

In short, CA separated seven species of Tangara to four species groups on the basis of fruit preference: (1) two forest species (T. labradorides and T. nigroviridis) that fed heavily on small fruits of Miconia and Trema; (2) three forest species (T. arthus, T. parzudakii, and T. ruficervix) that fed more frequently on larger fruits; (3) one nonforest species (T. cyanicollis); and (4) T. ruflgula, with a higher proportion of Palicourea and Bocconia fruit.

Foraging overlaps.-The foraging overlap between 21 species-pairs was 7 SD = 0.28 0.28 (range: 0.00-0.93; Table 4) in arthropodforaging and … SD = 0.81 0.13 (range: 0.48-0.97; Table 4) in fruit-foraging. The mean overlap in fruit-foraging was 2.9x higher than in arthropodforaging. In fruit-foraging, most species- pairs showed resource overlap higher than 0.70, with the exception of T. cyanicollis, a nonforest species, which showed lower overlap in fruit-foraging with the other forest species of Tangara (Table 4). The average observed overlap in fruit-foraging was significantly larger than that expected by chance (Table 5). By contrast, in arthropodforaging, only two species-pairs showed resource overlap >0.70 (Table 4), and the average observed foraging overlap was significantly smaller than that expected by chance (Table 5). The differences in foraging ecology of these two species-pairs, T. arthus-T. parzudakii and T. labradorides-T. rufigula, were described in detail in the section on arthropod-foraging above.

DISCUSSION

Most variation in foraging ecology of Tangara is explained by relatively few foraging parameters. Differences in microhabitat preference characterized by the combination of substrate type, horizontal position, and attack maneuver explain the major differences in arthropod-foraging among sympatric species of Tangara. Differences in habitat choice further explained part of the remaining variation. The observed average overlap in arthropod- foraging is smaller than that expected by chance, and most species- pairs share <60% of arthropod resources. High resource overlap of two species-pairs is partially attributable to the coarse resolution of some foraging variables, such as leaf and moss types, species of Tangara that inhabit the same forested area and often join the same mixed-species flocks exploit different microhabitats for arthropod- foraging.

The differences in fruit-foraging among species of Tangara are the consequence of differences in habitat use, rather than differences in microhabitat such as different height or horizontal position of the same trees. This apparent lack of microhabitat partitioning in fruiting trees was also observed among three species of Tangara and other omnivorous birds in Costa Rica (Daily and Ehrlich 1994). Although the main differences in fruit-foraging appear to be differences in habitat use, the differences in fruit- foraging among the forest species cannot be explained solely by habitat, because these species prefer different fruit species even when in the same mixed-species flocks (K. Naoki pers. obs.). For example, two small species, T. nigroviridis and T, labradorides, preferably feed on small fruit species, M. brevitheca and Tretna micrantha, which constitute 86% and 93% of their fruit-foraging observations, respectively. Their small body size appears to limit access to some large fruits and also to C. gabrielis, which requires the hangside attack maneuver at a thick Cecropia catkin. Despite the differences in various foraging parameters, the average resource overlap among species of Tangara is larger than expected by chance in fruit-foraging.

Arthropod-foraging in species of Tangara is characterized by (1) interspecific differences in microhabitat preferences and subsequently differences in habitat use, (2) low average resource overlap, and (3) particularly low resource overlap among the species that share the same habitat and participate in the same mixed- species flocks. By contrast, fruit-foraging is characterized by (1) interspecific differences in plant-genus use, often associated with differences in habitat use; (2) high average resource overlap; and (3) particularly high resource overlap among species that share the same habitat. In other words, species of Tangara that use the same space for foraging and form mixed-species flocks in the same habitat differ largely in arthropod-foraging but overlap greatly in fruit- foraging. These conclusions are similar to previous foraging observations of omnivorous tanagers (Snow and Snow 1971, Ridgely and Tudor 1989, Isler and Isler 1999; but see Rodrigues 1995).

Many plant species in the tropics are distributed patchily (Loiselle and Blake 1993), and fruiting of individual plant species is seasonal and poorly predictable in time (Hilty 1980, Levey 1988, Loiselle and Blake 1991). Therefore, specializing on individual fruit species or even a genus would be difficult. Most bird species consume a variety of fruits in various families (Snow 1981, Wheelwright et al. 1984). Although fruiting of individual species is limited in time and space, fruits are easy to find and tend to be superabundant when and where available, thus allowing many species and individuals to feed on the same fruiting trees without severe competition (Willis 1966, Leek 1969). In addition, fruits can be plucked by birds in only a few different ways (Snow and Snow 1971), and morphologically similar congeners mostly use the same foraging techniques (see fig. 7 in Moermond and Denslow 1983). Thus, syntopic congeners do not show clear partitioning of fruits and often are grouped in the same foraging guild when analyzed together with other frugivorous genera (Loiselle and Blake 1990, K. Naoki unpubl. data).

The substrate types for arthropod-foraging used here are more homogeneously distributed in both space and time, even though arthropod taxa found in those substrates may vary seasonally. Arthropods are often limited in quantity and, most importantly, many are cryptic and have developed ways to avoid depredation. These features favor the development of specialization in predators to exploit group\s of arthropods that adopt similar predator-avoidance mechanisms. With few exceptions (Sherry 1985), most insectivorous birds are thus highly specialized on both the substrate searched and the foraging maneuver used to capture arthropods (e.g., Robinson and Holmes 1982, Sherry 1984), and this specialization in a searching method and substrate type enhances foraging efficiency (Robinson and Holmes 1982).

Data collection for the present study was limited to the dry season, presumably the season of minimal breeding in the tropics. The relatively short duration of the study may have influenced the conclusion concerning the relative importance of arthropods. However, arthropod consumption is expected to increase during the breeding season, which is usually the beginning of the wet season, and to match the higher protein demand for egg production and feeding of nestlings in most fruit-eating birds (Moermond and Denslow 1985, Poulin et al. 1992). Even during the nonbreeding season, species of Tangara spent 60% of their foraging time searching for arthropods, which constituted one-third of the foraging observations (Naoki 2003). Thus, it is unlikely that the relative importance of arthropods would diminish if the study had been conducted over a longer period. Interestingly, many frugivorous birds breed at the same time as insectivorous birds when fruit availability is often low (Hilty 1977, Levey 1988, Loiselle and Blake 1991). This suggests that some of the key biological aspects of these fruit-eating birds, such as population density and community organization, are governed by arthropod availability, rather than fruit availability as often assumed.

In summary, the difference of fruits and arthropods as food resources appear to explain (1) apparent lack of segregation in fruit-foraging among species of Tangara, in contrast to the fine segregation in arthropod-foraging; and (2) little specialization in fruit-foraging with wide variety of fruits consumed by each species of Tangara, in contrast to highly specialized arthropodforaging in both substrate type and attack maneuver. High degrees of sympatry of Tangara and other omnivorous tanagers, in general, appear to be maintained not because fruits are abundant and cause little competition, but because these species specialize on different arthropods, which are presumably more limited but of higher quality than fruits.

ACKNOWLEDGMENTS

I am grateful to M. Moreno Espinosa (Museo Ecuatoriano de Ciencias Naturales) and S. Lasso (Ministerio de Medio Ambiente) for help in obtaining the research permit in Ecuador. I thank L. Chaves, M. Jcome, and Francisco and Fernando Sornoza for providing advice and field assistance; E. Freire, H. Vargas, E. Narvez, E. Toapanta, and J. C. Ronquillo (Herbario Nacional del Ecuador) for identification of plant samples; and E. Bastidas and V. Zak for kindly allowing me to work in their reserve and private property. Many people in Mindo helped me in various aspects of the research. I especially thank H. Onate and A. Onate, E. Toapanta, Fondation Pacaso y Pacaso, and Amigo de la Naturaleza. N. Krabbe, P. J. Greenfield, and J. C. Matthew shared observations on natural history of tanagers. The manuscript benefited from comments by J. V. Remsen, Jr., B. O’Shea, P. C. Stouffer, A. R. Styring, T. Valqui, Ramiro P. Lpez, M. Isabel Gomez, and two anonymous reviewers. The study was supported by the American Ornithologists’ Union, American Museum of Natural History, Wilson Ornithological Society, and Louisiana State University Museum of Natural Science.

LITERATURE CITED

BLAKE, J. G., AND B. A. LOISELLE. 2000. Diversity of birds along an elevational gradient in the Cordillera Central, Costa Rica. Auk 117: 663-686.

BUSKIRK, W. H. 1976. Social systems in a tropical forest avifauna. American Naturalist 110: 293-310.

CONNELL, J. H. 1983. On the prevalence and relative importance of interspecific competition: Evidence from field experiments. American Naturalist 122:661-696.

DAILY, G. C., AND P. R. EHRLICH. 1994. Influence of social status on individual foraging and community structure in a bird guild. Oecologia 100:153-165.

DIAMOND, J. M. 1978. Niche shifts and the rediscovery of interspecific competition. American Scientist 66:322-331.

GILL, F. B., AND L. L. WOLF. 1978. Comparative foraging efficiencies of some montane sunbirds in Kenya. Condor 80:391-400.

GOTELLI, N. J., AND G. L. ENTSMINGER. 2006. ECOSIM: Null models software for ecology, version 7. Acquired Intelligence & KeseyBear, Jericho, Vermont. [Online.] Available at garyentsminger.com/ ecosim.htm.

GRACE, J. B., AND D. TILMAN, EDS. 1990. Perspectives on Plant Competition. Academic Press, San Diego, California.

GRANT, P. R. 1986. Ecology and Evolution of Darwin’s Finches. Princeton University Press, Princeton, New Jersey.

HEJL, S. J., J. VERNER, AND G. W. BELL. 1990. Sequential versus initial observations in studies of avian foraging. Pages 166-173 in Avian Foraging: Theory, Methodology, and Applications (M. L. Morrison, C. J. Ralph, J. Verrier, and J. R. Jehl, Jr., Eds.). Studies in Avian Biology, no. 13.

HILTY, S. L. 1977. Food supply in a tropical frugivorous bird community. Ph.D. dissertation, University of Arizona, Tucson.

HILTY, S. L. 1980. Flowering and fruiting periodicity in a premontane rain forest in pacific Colombia. Biotropica 12:292-306.

HILTY, S. L., AND W. L. BROWN. 1986. A Guide to the Birds of Colombia. Princeton University Press, Princeton, New Jersey.

HOLMES, R. T., R. E. BONNEY, JR., AND S. W. PACALA. 1979. Guild structure of the Hubbard Brook bird community: A multivariate approach. Ecology 60:512-520.

ISLER, M. L., AND P. R. ISLER. 1999. The Tanagers: Natural History, Distribution, and Identification. Smithsonian Institution Press, Washington, D.C.

KREBS, C. J. 1999. Ecological Methodology, 2nd ed. Benjamin/ Cummings, Menlo Park, California.

LACK, D. L. 1971. Ecological Isolation in Birds. Harvard University Press, Cambridge, Massachusetts.

LECK, C. F. 1969. Observations of birds exploiting a Central American fruit tree. Wilson Bulletin 81:264-269.

LEVEY, D. J. 1988. Spatial and temporal variation in Costa Rican fruit and fruit-eating bird abundance. Ecological Monographs 58:251- 269.

LOISELLE, B. A., AND J. G. BLAKE. 1990. Diets of understory fruit- eating birds in Costa Rica: Seasonality and resource abundance. Pages 91-103 in Avian Foraging: Theory, Methodology, and Applications (M. L. Morrison, C. J. Ralph, J. Verner, and J. R. Jehl, Jr., Eds.). Studies in Avian Biology, no. 13.

LOISELLE, B. A., AND J. G. BLAKE. 1991. Temporal variation in birds and fruits along an elevational gradient in Costa Rica. Ecology 72: 180-193.

LOISELLE, B. A., AND J. G. BLAKE. 1993. Spatialdistribution of understory fruit-eating birds and fruiting plants in a Neotropical lowland wet forest. Vegetatio 108:177-189.

LOURIE, S. A., AND D. M. TOMPKINS. 2000. The diets of Malaysian swiftlets. Ibis 142:596-602.

MACARTHUR, R. H. 1958. Population ecology of some warblers in northeastern coniferous forests. Ecology 39:599-619.

MARTIN, P., AND P. P. G. BATESON. 1993. Measuring Behaviour: An Introductory Guide. Cambridge University Press, Cambridge, United Kingdom.

MILES, D. B. 1990. A comparison of three multivariate statistical techniques for the analysis of avian foraging data. Pages 295-308 in Avian Foraging: Theory, Methodology, and Applications (M. L. Morrison, C. J. Ralph, J. Verner, and J. R. Jehl, Jr., Eds.). Studies in Avian Biology, no. 13.

MOERMOND, T. C., AND J. S. DENSLOW. 1983. Fruit choice in Neotropical birds: Effects of fruit type and accessibility on selectivity. Journal of Animal Ecology 52:407-420.

MOERMOND, T. C., AND J. S. DENSLOW. 1985. Neotropical avian frugivores: Patterns of behavior, morphology, and nutrition, with consequences for fruit selection. Pages 865-897 in Neotropical Ornithology (P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgely, and F. G. Buckley, Eds.). Ornithological Monographs, no. 36.

NAOKI, K. 2003. The relative importance of arthropods and fruits in foraging behavior of omnivorous tanagers (Thraupidae): The comparison of three methods. Condor 105: 135-139.

POULIN, B., G. LEFEBVRE, AND R. McNEIL. 1992. Tropical avian phenology in relation to abundance and exploitation of food resources. Ecology 73:2295-2309.

PRICE, T. 1991. Morphology and ecology of breeding warblers along an altitudinal gradient in Kashmir, India. Journal of Animal Ecology 60:643-664.

REMSEN, J. V., JR. 1990. Community ecology of Neotropical kingfishers. University of California Publications in Zoology 124:1- 116.

REMSEN, J. V., JR., M. A. HYDE, AND A. CHAPMAN. 1993. The diets of Neotropical trogons, motmots, barbets and toucans. Condor 95:178- 192.

REMSEN, J. V., JR., AND S. K. ROBINSON. 1990. A classification scheme for foraging behavior of birds in terrestrial habitats. Pages 144-160 in Avian Foraging: Theory, Methodology, and Applications (M. L. Morrison, C. J. Ralph, J. Verner, and J. R. Jehl, Jr., Eds.). Studies in Avian Biology, no. 13.

RIDGELY, R. S., AND P. J. GREENFIELD. 2001. The Birds of Ecuador. Cornell University Press, Ithaca, New York.

RIDGELY, R. S., AND G. TUDOR. 1989. The Birds of South America, vol. 1 : The Oscine Passerines. University of Texas Press, Austin.

ROBINSON, S. K., AND R. T. HOLMES. 1982. Foraging behavior of forest birds: The relationships among search tactics, diet, and habitat structure. Ecology 63:1918-1931.

RODRIGUES, M. 1995. Spatial distribution and food utilization among tanagers in southeastern Brazil (Passeriformes: Emberizidae). Ararajuba 3:27-32.

SCHOENER, T. W. 1974. Resource partitioning in ecological communities. Science 185:27-39.

SHERRY, T. W. 1984. Comparative dietary ecology of sympatric, insectivorous Neotropical flycatchers (Tyrannidae). Ecological Monographs 54:313-338.

SHERRY, T. W. 1985. Adaptation to a novel environment: Food, foraging, and morphology of the Cocos Island Flycatcher. Pages 908- 920 in N\eotropical Ornithology (P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgely, and F. G. Buckley, Eds.). Ornithological Monographs, no. 36.

SNOW, B. K., AND D. W. SNOW. 1971. The feeding ecology of tanagers and honeycreepers in Trinidad. Auk 88:291-322.

SNOW, D. W. 1981. Tropical frugivorous birds and their food plants: A world survey. Biotropica 13:1-14.

STOTZ, D. F., J. W. FITZPATRICK, T. A. PARKER III, AND D. K. MOSKOVITS. 1996. Neotropical Birds: Ecology and Conservation. University of Chicago Press, Chicago, Illinois.

TERBORGH, J., AND J. M. DIAMOND. 1970. Niche overlap in feeding assemblages of New Guinea birds. Wilson Bulletin 82:29-52.

TILMAN, D. 1982. Resource Competition and Community Structure. Princeton University Press, Princeton, New Jersey.

TOKESHI, M. 1999. Species Coexistence: Ecological and Evolutionary Perspectives. Blackwell Publishing, Oxford, United Kingdom.

WHEELWRIGHT, N. T., W. A. HABER, K. G. MURRAY, AND C. GUINDON. 1984. Tropical fruit-eating birds and their food plants: A survey of a Costa Rican lower montane forest. Biotropica 16:173-192.

WILLIS, E. O. 1966. Competitive exclusion and birds at fruiting trees in western Colombia. Auk 83:479-480.

Associate Editor: P. C. Stouffer

KAZUYA NAOKI1

Department of Biological Sciences and Museum of Natural Science, 119 Foster Hall, Louisiana State University,

Baton Rouge, Louisiana 70803, USA

1 Present address: Instituto de Ecologa, Universidad Mayor de San Andrs, Casilla 6394, Correo Central, La Paz, Bolivia. E-mail: knaoki@entelnet.bo

Copyright American Ornithologists’ Union Jan 2007

(c) 2007 Auk, The. Provided by ProQuest Information and Learning. All rights Reserved.