Louisiana State University LSU Digital Commons
LSU Doctoral Dissertations Graduate School
2003 Evolution of ecological diversity in the neotropical tanagers of the genus Tangara (Aves: Thraupidae) Kazuya Naoki Louisiana State University and Agricultural and Mechanical College, [email protected]
Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_dissertations
Recommended Citation Naoki, Kazuya, "Evolution of ecological diversity in the neotropical tanagers of the genus Tangara (Aves: Thraupidae)" (2003). LSU Doctoral Dissertations. 2021. https://digitalcommons.lsu.edu/gradschool_dissertations/2021
This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Doctoral Dissertations by an authorized graduate school editor of LSU Digital Commons. For more information, please [email protected].
EVOLUTION OF ECOLOGICAL DIVERSITY IN THE NEOTROPICAL TANAGERS OF THE GENUS TANGARA (AVES: THRAUPIDAE)
A Dissertation
Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of Doctor of Philosophy
in
The Department of Biological Sciences
by Kazuya Naoki B.S., Universidad de Costa Rica, 1996 August 2003
ACKNOWLEDGMENTS
Many people and institutions in many countries helped make this dissertation possible.
First I thank my parents and my sister in Japan to understand and support my interests in tropical biology for last 14 years. Even though I have failed to get in touch with them frequently, they always showed great comprehension on my passion and love to nature and offered to support me in all possible ways. I thank my advisor, J. V. Remsen Jr. (Van), for many, many advices at critical moments of my Ph.D. for last six years. Without his help, I would not have finished my
Ph.D. I also thank my graduate committee: Frederick H. Sheldon, J. Michael Fitzsimons, Kyle
E. Harms, James P. Geaghan, and L. Lee Southern for their time and support. At LSU I benefited greatly from fellow graduate students.
Funding for fieldwork was provided by the following organizations: the National
Geographic Society, the Cooper Ornithological Society, the Wilson Ornithological Society, the
American Ornithologists’ Union, the American Museum of Natural History, the Louisiana State
University Museum of Natural Science, SigmaXi, LSU Dept. Biological Sciences.
Many people made my fieldwork possible. In Ecuador, 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. I thank L. Chaves, M. Jácome (Fundación
Ornitológica del Ecuador), Francisco and Fernando Sornoza for providing advice and field assistance; E. Freire, H. Vargas, E. Narváez, Elsa Toapanta, and J. C. Ronquillo (Herbario
Nacional del Ecuador) for identification of plant samples; and E. Bastídas 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 Hugolino and Alicia Oñate, Efraín Toapanta,
ii Fundación Pacaso y Pacaso, and Amigo de la Naturaleza. N. Krabbe, P. Greenfield, and J. C.
Matthew shared observation on natural history of tanagers.
In Costa Rica, I thank Julio E. Sánchez, Hernan Araya, Leonardo Chavez, and Andres
Vega for helping us to find a study site and also provided logistic support. Don Beche, Doña
Lucia, Don Wiliam, and Doña Estrella in Tausito offered us a shelter, meals, and great friendship
during the fieldwork. Don Beto Chavez Mora and Don Leonel of Reserva Biológica El Copal let
us use their reserve for the study. Don Beto and his family also helped our fieldwork and offered
their companionship. Armando Estrada of Museo Nacional de Costa Rica kindly identified all
the plant samples. Marco Tulio kindly made slides for Naoki’s talk presented for Asociación
Ornitológica de Costa Rica. I also thank Ministerio del Ambiente y Energia for permitting to
conduct this research in Costa Rica. I thank Maria Isabel Gómez and Ernesto Carman for
collaboration in fieldwork.
In Bolivia, I thank Carmen Quiroga, Omar Rocha, James Aparicio, Alvaro Garritanos
(Colección Boliviana de Fauna), Sebastian Herzog, Bennett Hennesy (Armonia) for giving me
indispensable advise in conducting fieldworks. Omar Martinez and M. Isabel Gómez kindly accompanied me into the fields. Finally, I want to thank M. Isabel Gómez and her family for their kindness and hospitality.
iii TABLE OF CONTENTS
ACKNOWLEDGMENTS ...... ii
ABSTRACT...... ……...…v
CHAPTER 1 INTRODUCTION ...... ……...... 1
2 DICHOTOMOUS DIFFERENCES IN RESOURCE USE AMONG OMNIVOROUS TANGARA TANAGERS …...... ……...... ……..5
3 THE RELATIVE IMPORTANCE OF ARTHROPODS AND FRUITS IN FORAGING BEHAVIOR OF TANGARA TANAGERS ….....…...... ……..73
4. SEASONAL CHANGES IN FORAGING ECOLOGY OF TANGARA TANAGERS IN COSTA RICA: FOOD-TYPE DEPENDENT RESOURCE PARTITIONING AND TEMPORAL VARIATION...... …...……..83
5 EVOLUTION OF ECOLOGICAL DIFFERENCES IN TANGARA TANAGERS ….....…...... …...... …...... …...... …...... …...... 109
6 SUMMARY AND CONCLUSION .…...... …...... …...... …...... ………...... …….142
REFERENCES ………………………………………………………………………………...154
APPENDIX 1 DATA FOR ARTHROPOD FORAGING …………………………………………….163
2 DATA FOR FRUIT FORAGING …………………….……………………………….167
3 DATA FOR HABITAT USE ……………………………………..…..………….….…175
4 DATA FOR ELEVATIONAL DISTRIBUTION (M) .…….…………………….……177
VITA ………………………………………………………………………………………...…179
iv ABSTRACT
The origin and maintenance of biological diversity has been one of the fundamental issues in biology. However, the evolution of ecological traits that affect species coexistence and species diversity is poorly known. My research aimed to investigate the evolution of species- specific ecological and morphological traits and to understand the process of ecological diversification and species coexistence in Tangara tanagers (Thraupidae) by using phylogenetic comparative methods. Tangara is the largest avian genus in the New World with 50 recognized species. As many as ten species of Tangara are found sympatrically in the same Andean cloud forest, and many syntopic species travel together in mixed-species flocks. The distribution of
Tangara covers all of subtropical and tropical America from sea level to tree line; thus, Tangara species show a wide range of habitat preferences as well as strong variation in number of coexisting species and species combinations. Like many other species of tanagers, Tangara species are omnivorous; their diet consists of both insects and fruit. I collected extensive ecological and behavioral data at six study sites to quantify ecological differences among sympatric species. I measured museum skins and skeletons to define morphospace of each taxon. DNA sequences were used to build a molecular phylogeny, which reveals the speciation pattern. I combined ecological data, morphological data, distributional data from literature, and a molecular phylogeny by two phylogenetic analytical methods to elucidate evolution of ecological diversity among 25 Tangara taxa. Permutational phylogenetic regression analyses showed significant phylogenetic effects for arthropod foraging, but not for fruit foraging, habitat use, and elevational distribution. A disparity-through-time plot showed that the relative disparity of arthropod foraging decreased more rapidly than the other niche axes. These analyses revealed diverse evolutionary patterns unique to each niche axis. The relative strength of phylogenetic
v effects, frequency of homoplasy, mode of evolution, and association with morphology differed substantially among the four niche axes. Fruit foraging and habitat specialization showed the greatest ecological plasticity in relation to phylogeny, and the variation in microhabitat preference in arthropod foraging associated with species-specific attack maneuver was the most conservative and consistent with the phylogeny.
vi CHAPTER 1 INTRODUCTION
One of the fundamental goals in biology is to explain the pattern and origin of biological diversity (Hutchinson 1959; Magurran and May 1999; Ricklefs and Schluter 1993; Rosenzweig
1995). Most attempts to do so have been conducted at small temporal scales. Such studies have examined how local ecological processes, such as competition, predation, mutualism, and resource availability, influence local community structure (McPeek and Miller 1996; Schluter and Ricklefs 1993). Community diversity, however, is determined not only by local but also by regional processes, such as species formation, dispersal, and extinction (Ricklefs 1987; Ricklefs and Schluter 1993). Hence, the incorporation of evolutionary thinking is one of the most urgent tasks in the study of biological diversity (McPeek and Miller 1996; Ricklefs and Schluter 1993;
Rosenzweig 1995; Tokeshi 1999). Conventional ecological approaches through direct observations or through field experiments can reveal short-term changes in ecological traits, such as habitat choice, diet selection, and geographical distribution, which are crucial to understanding the patterns of community diversity and species coexistence. These approaches, however, do not allow for direct examination of evolutionary changes in ecological traits at a geologic time scale.
Previously, hypotheses that involved long-term evolutionary changes were testable only by examination of fossil records. Because most ecological and behavioral traits are not preserved as fossils, evolutionary changes in these traits could only be inferred indirectly.
Integration of systematics and biogeography with ecology in the late 1980s opened a new field, historical ecology (Brooks and McLennan 1991; Harvey and Pagel 1991). This approach allows for the reconstruction of ancestral character states of ecological traits, the direct examination of evolutionary pathways, and the testing of hypotheses involving long-term evolutionary changes
1 (Brooks and McLennan 1991; Harvey and Pagel 1991; Martins 1996). Although phylogenetic comparative methods are now widely used, few studies have applied phylogenetic approaches to elucidate the evolutionary changes of ecological and morphological traits that determine species coexistence and formation of biological diversity (Losos 1996b).
The objective of my research was to investigate the evolution of species-specific ecological traits in Tangara tanagers (Thraupidae) and to understand the pattern of ecological diversification and species coexistence of this large genus. The objective was approached as follows: (I) Each species’ ecological niche was investigated through examination of resource partitioning among sympatric species at several study sites in the Neotropics as well as differences in elevational distributions in response to habitat and environmental conditions. (II)
Patterns of speciation were investigated by reconstructing a molecular phylogeny and by studying the geographic distributional patterns among the various species and lineages. (III)
Evolutionary change of various ecological and morphological traits, such as microhabitat use, habitat preference, body size, and body shape, were studied by mapping these characters on a molecular phylogeny.
Only a handful of studies have used phylogenetic methods to investigate the evolution of ecological differences in bird communities. For example, Richman and Price (1992) examined community structure in the Old World warbler genus Phylloscopus. By comparing three communities, Richman (1996) concluded that two peripheral sites had different evolutionary history; however, the lack of ecological data from these two sites prevented further exploration of the factors that caused the difference. Joseph and Moritz (1993) studied historical patterns of the ecological diversification among Australian Sericornis scrubwrens, and Cicero and Johnson
(1998; 2002) investigated the evolution of habitat preference in New World Vireo and
2 Empidonax Flycatchers and concluded that habitat changes played an important role in speciation of these genera. Studies of Anolis lizards in the Caribbean (Losos 1992; Losos 1996a;
Losos 1996b) have provided a detailed picture of the assembly of these lizards’ communities. A comparison among islands in the Greater Antilles revealed that these communities have converged to a strikingly similar structure, indicating high ecological determinism (Losos et al.
1998). That study, however, focused on depauperate fauna of small islands; thus, its applicability to rich continental faunas is unknown. A study of cichlid fishes in Lake Malawi provided an example of how ecological diversification may have promoted rapid radiation of lineages (Danley and Kocher 2001). In contrast, a study of tiger beetles found no association between ecological disparity with species coexistence or radiation of lineages in North America
(Barraclough et al. 1999). Of the above studies, only the latter examined statistically whether ecological diversification promoted cladogenesis and coexistence of congeners.
My research is the first to study evolution of ecological niches in the continental
Neotropics, where biological diversity peaks for most groups of terrestrial organisms. My research also is the first to combine phylogenetic information with field data on ecological and behavioral traits of numerous species from several study sites.
BACKGROUND BIOLOGY OF TANGARA
The Tangara tanagers (Thraupidae: Aves) are small-bodied, canopy and forest-edge dwelling birds in the order Passeriformes. Several factors make Tangara an excellent choice for studying the evolution of ecological diversity and species coexistence. The genus Tangara contains 50 species, more than any other avian genus in the New World (Isler and Isler 1999;
Stotz et al. 1996). The Tangara tanagers are dominant components of forest communities in the mountain regions, especially in the Andes where the genus has the highest number of species
3 among avian genera in Venezuela, Colombia, Ecuador, Peru, and Bolivia (Stotz et al. 1996).
The distribution of Tangara covers all of subtropical and tropical America from sea level to tree
line; thus, Tangara species show a wide range of habitat preferences as well as strong variation
in number of coexisting species and species combinations (Isler and Isler 1999; Ridgely and
Tudor 1989). The geographic ranges and elevational distributions of most Tangara species are well known (Isler and Isler 1999; Ridgely and Tudor 1989; Stotz et al. 1996), which allows for the evaluation of the degree of sympatry between species and lineages.
As many as 10 species of Tangara are found sympatrically in the same Andean cloud forest (Isler and Isler 1999); many syntopic species travel together in mixed-species flocks. Thus, it is unclear how so many apparently similar species can coexist. Like many other species of tanagers, Tangara species are omnivorous; their diet consists of fruits, arthropods, nectar, flower buds, and Müllerian bodies. Some authors suggested syntopic Tangara species show ecological segregation in the way they forage on arthropods more than on fruits (Isler and Isler 1999;
Ridgely and Tudor 1989; Snow and Snow 1971). However, the quantitative behavioral data is available only in species-poor Trinidad and Southeastern Brazil (Rodrigues 1995; Snow and
Snow 1971).
No previous phylogenetic work has been done with Tangara tanagers. Isler and Isler
(1999), however, classified species into 13 different groups based on geographic distributions, plumage, behavior, vocalizations, and nest sites.
4 CHAPTER 2 DICHOTOMOUS DIFFERENCES IN RESOURCE USE AMONG OMNIVOROUS TANGARA TANAGERS
SUMMARY
The distribution and abundance of foods is a primary factor affecting the resource-use patterns of birds. Many bird species eat several food types, which may differ in their distribution and overall abundance. Foraging ecology of sympatric Tangara tanagers was studied at three sites: Mindo, Ecuador; El Copal, Costa Rica; and Serranía Bella Vista, Bolivia. The goal was to determine whether the patterns of resource partitioning differed between two food types: arthropods and fruits, and whether the patterns differed among the three study sites. Interspecific differences in arthropod foraging were manifested by the fine segregation of microhabitat preference combined with different habitat use. In contrast, interspecific differences in fruit foraging were manifested by preferences for different plant genera, often associated with different habitats. No evidence was found for spatial partitioning of the same fruiting tree.
Interspecific overlap in fruit foraging was four times higher than in arthropod foraging, and
Tangara species that frequently joined the same mixed-species flocks differed largely in arthropod foraging but overlapped greatly in fruit foraging. These patterns were observed at all three study sites, despite differences in the number of sympatric species and species composition.
The differences in patterns between arthropod foraging and fruit foraging may be explained by the different characteristics of arthropods and fruits as food resources. High sympatry 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.
5 INTRODUCTION
Patterns of resource partitioning among sympatric species have been one of the central issues in community ecology (e.g., Connell 1983; Diamond 1978; Grace and Tilman 1990;
Schoener 1973; Tilman 1982). Since MacArthur’s study of sympatric Dendroica warblers
(1958), 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)]. 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 non-congeners; 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; Loiselle and Blake 1990;
Remsen 1990; Sherry 1984). 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 their combinations. 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), and many are also called
6 “omnivorous” or “frugivorous-insectivorous” (Blake and Loiselle 2000; Buskirk 1976). If the
distribution and abundance of a resource influences the way birds partition it, then how do
omnivorous species feeding on more than one food type partition them? Do they partition two
food types differently according to the distribution and abundance of each food type? Does one
food type have stronger influence than the other 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 patterns and overlap of resource use in two food
types of Tangara tanagers at three sites where the number of sympatric species and species
composition largely differ. Tangara 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 sympatric in the same Andean cloud forest, and as many as nine 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 (Hilty 1977; Snow and Snow
1971). Many Tangara tanagers 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 (Rodrigues 1995; Snow and Snow 1971). These two studies surveyed all sympatric tanagers of the family Thraupidae but were conducted in relatively species-poor Trinidad and southeastern Brazil, where only three Tangara species were found.
7 The first objective of this study was to determine major foraging differences among
sympatric Tangara tanagers in two different food types: arthropods and fruits. I applied
multivariate techniques to analyze various foraging parameters simultaneously and to elucidate
the principal dimensions in foraging ecology. The second objective was to compare patterns of
resource partitioning between the two food types used by the same set of congeners during the
same time period. The third objective was to compare the overall interspecific overlap in
resource use between the two food types. The fourth objective was to relate the differences in
resource-use patterns to the presumed differences in distribution and abundance between the
food types. Finally, I compared the resource-partitioning patterns among three study sites to find
whether any trend is discernible.
METHODS
FORAGING DATA
All data were obtained between 6:00 and 12:00, and 13:00 and 18:00, when birds were most active. I located birds by sight and sound while slowly walking along a road. I observed individual birds through 10X40 binoculars and recorded observations 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.
“Food item” was classified as a fruit, arthropod, nectar, flower bud, or Müllerian body, but in the multivariate analyses I used only arthropod and fruit observations. Substrate categories for arthropod-foraging were: (a) moss or thickly moss-covered branch, (b) partially moss-covered
8 branch, (c) bare branch, (d) dead branch, (e) live leaf, (f) dead leaf, (g) flower bud, and (h) air.
For fruit foraging, I recorded plant species instead of substrate type although in analyses I used genera instead of species. Substrate size and perch diameters were estimated relative 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. 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”) and used this parameter instead of “distance to canopy” for the analyses. Classification and nomenclature of “attack maneuver” and “perch angle” followed Remsen and Robinson (1990).
DATA ANALYSIS
Foraging parameters in this study included both categorical and continuous variables.
The continuous variables, such as perch diameter, height above ground, and distance to canopy, were grouped into four to six categories. This allowed the 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 and maximize interspecific differences. Correspondence analysis has been shown as the preferred method for analyzing categorical foraging data because it recovered more variation from the original data sets and was more consistent in magnitude and sign of the coefficients from eigenvectors than other multivariate methods such as PCA and factor analysis (Miles 1990). When an interspecific difference was unclear in the first three CA
9 dimensions, I analyzed the other dimensions to see whether any interspecific difference was
explained by other foraging parameters.
After conducting multivariate analyses, I tested whether sympatric Tangara tanagers 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). For this purpose, I used the subset of fruit
foraging data from Ecuador: foraging observations on two most commonly eaten fruit genera:
Miconia and Trema, which together accounted for two thirds of all the fruit foraging observations. I conducted ANOVAs on foraging height and vertical position of each of the two
fruit genera by using Tangara species as independent variables.
To measure degree of overlap in resource use, I calculated Pianka’s measure of niche
overlap for arthropod and fruit foraging (Krebs 1999, p. 471).
n ∑ pij pik O = i jk n n 2 2 ∑ pij ∑ pik i i
where Pij = Proportion that resource i is of the total resources used by species j
Pik = Proportion that resource i is of the total resources used by species k
n = Total number of resource states
I constructed 3-way tables comprised of species and two principal foraging parameters: “habitat” and “substrate type” in arthropod foraging, and “habitat” and “plant genus” in fruit foraging.
The 3-way tables generated some foraging categories with few or no foraging observations. I lumped the categories with fewer than three observations to avoid having numerous cells with zero observations. Because the categories used for arthropod foraging and fruit foraging to calculate niche overlaps were not identical, additional care was taken when comparing these two
10 foraging categories. For example, several species search the same substrate in the same habitat
for arthropods, but they may take different arthropods based on arthropod body size or
taxonomic groups. In this case, the measure of niche overlap is biased towards detecting more
overlap than really exists. To lessen this kind of bias and to compare objectively niche overlap,
observed degrees of niche overlap were tested against expected from chance by using a Monte
Carlo approach. I used the program EcoSim (Gotelli and Entsminger 2003) to generate a null
distribution by 1,000 randomizations with randomization algorithm 2 (RA2). RA2 assumes that
certain resource states are unavailable for each species even in the absence of species interactions
but the amount of specialization may change.
To avoid the redundancy of presenting the numerous data tables that do not show clear
interspecific differences, I present complete data and most through analyses only for the data
from Mindo, Ecuador. For the analyses of Costa Rican and Bolivian data, I present only the data
of the foraging parameters recognized as important by correspondence analyses and others
relevant to the discussion. The complete data sets for all the studied Tangara species are in
Appendices 1-3.
STUDY AREAS
Ecuador
The study was conducted in the vicinity of Mindo, prov. Pichincha, Ecuador (0°02’S,
78°46’W; Fig. 2.1). 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 from both lowland humid forest and Andean cloud forest; over 360 bird species have been
11 12 recorded in this area (Ridgely and Greenfield 2001). The vegetation around Mindo consists of a
mosaic of secondary forest and patches of pastures, although some large trees over 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/year (n = 13 years), and annual mean temperature is 20.3 ± 0.2°C (n = 11 years;
unpubl. data from Instituto Nacional de Meteorología e Hidrología). 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 the first trail (1,300-1,500
m), the vegetation was more disturbed and had more plant species from lower elevations than did
the first trail (pers. obs.). At this second site, I found higher densities of Tangara species typical
of foothill forests (Tangara rufigula, T. gyrola, and T. icterocephala).
Costa Rica
The study was conducted at Reserva Biológica El Copal, prov. Cartago, Costa Rica
(9°47’N, 83°45’W, 970m; Fig. 2.1) and its vicinity from December 2000 to June 2001. El Copal is a newly established private biological reserve that protects 200 ha of little-disturbed subtropical wet forest. El Copal is surrounded on three sides by Tapantí National Park and by La
Amistad International Park, whose 255,000 ha forest extends from central Costa Rica to the
Panama border.
Rainfall averages 4,699 ± 418 mm/year (n = 10 years), and annual mean temperature is
22.6°C (n = 36 years; unpubl. data from Instituto Costarricense de Electricidad). The dry season
13 lasts from mid-December to mid-May. The more detailed climatic information is described in
Chapter 4, where I analyzed the effect of seasonality.
Foraging ecology was quantified along a 3-km trail inside El Copal and a 4-km road from
the entrance of El Copal toward Cartago. The vegetation along this 4-km road was more
disturbed with several small sugar cane fields. Both observation trails are between 960 and
1,200 m in elevation and were marked with numbered color tapes at every 50 m. Although the
data were collected during both dry and wet seasons, which correspond to non-breeding and
breeding seasons respectively, I used only the data from dry season in this chapter to make
comparisons among three study sites. In Chapter 4, I present the complete data set from Costa
Rica.
Bolivia
The study was conducted at Serranía Bella Vista, depto. La Paz (15°39’S-67°28’W,
1,250-1,550m; Fig. 2.1), 20 June-5 August 2000, 24 September-11 October, and 27 June 31 July
2002. Serranía Bella Vista is in the Upper Tropical Zone of Meyer de Schauensee (1970), and its vegetation is classified as subhumid pluvistational forests (Navarro and Maldonado 2002).
Foraging ecology of Tangara species was quantified along the 8 km dirt road between 5 km northwest of village Km 52 and the summit of Serranía between 1,250 and 1,550 m. This dirt road is a main highway connecting Caranavi and Yucumo. Although the area between village Km 52 and Serranía was mostly deforested for agriculture and cattle raising, the forest at
Serranía was disturbed only slightly because of its steep terrain. Cracids and large frugivorous birds, which are most sensitive to deforestation, were still common in this area (pers. obs.).
There is no weather station in this area; however, the wet season begins in approximately
14 December and lasts until April as most of the wet eastern Bolivian Andes (pers. comm. from
local people).
RESULTS
ECUADOR
At Mindo, I observed 11 species of Tangara tanagers. Two of them, Tangara 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 below 1,400 m above sea level. Tangara cyanicollis was essentially a solitary, non-forest species and was mostly found in pairs in a semi-open area. The other five Tangara species, T. arthus, T. labradorides, T. nigroviridis, T. parzudakii, and T. ruficervix, were common in tall wet forests at higher elevations and were often found in the same mixed-species flock (unpubl. data). These five species, with T. cyanicollis and T. rufigula, yielded the bulk of my 1,340 foraging observations (Table 2.1).
Arthropod Foraging
Correspondence analysis of arthropod foraging showed that the first two dimensions accounted for 84% of the total variance (Fig. 2.2, Table 2.2). These two dimensions were heavily weighted by three foraging parameters: “attack maneuver,” “substrate type,” and
“horizontal position,” which represented microhabitat preference and together contributed to
0.70 of the weighted partial contribution (see WPC per parameter in Table 2.2). The remaining dimensions were weighted by “substrate type” and “attack maneuver,” but the fourth dimension was weighted by “habitat.” Dimensions 1, 2, and 3 separated seven Tangara species to four species-groups based on microhabitat preference (Fig. 2.2): (1) Tangara arthus and T. parzudakii used thick branches at interior and middle position by reach-down and hang-down maneuver;
15 Table 2.1. Body mass and food types used by seven Tangara species in Mindo, Ecuador.
Percent in each food category Body mass (g)a Flower Müllerian Species Code mean ± SD (n) n Arthropod Fruit bud body Nectar Tangara arthus A 22.1 ± 1.6 (11) 312 33 66 0 0 0 T. cyanicollis C 18.2 ± 1.3 (10) 127 24 73 2 0 0 T. labradorides L 14.6 ± 1.6 (7) 106 40 60 0 0 0 T. nigroviridis N 17.9 ± 1.0 (12) 115 38 62 0 0 0 T. parzudakii P 26.6 ± 1.4 (9) 197 40 59 1 0 1 T. ruficervix X 19.7 ± 1.6 (16) 231 21 67 0 7 4 T. rufigula R 19.1 ± 0.8 (11) 136 40 52 3 3 1
a Taken from the specimen labels at Louisiana State University Museum of Natural Science.
16
17 Table 2.2. Partial contributions of foraging variables to inertia of arthropod foraging in Mindo, Ecuador.
WPC per Foraging parameter Categories CA dimensions WPC** parameter Dim1 Dim2 Dim3 Dim4 Dim5 Dim6 (54%)* (30%) (8%) (4%) (3%) (2%) Attack maneuver glean 0.00 0.04 0.01 0.00 0.00 0.06 0.02 0.19 reach-up 0.01 0.01 0.05 0.00 0.01 0.04 0.01 reach-out 0.00 0.00 0.00 0.01 0.00 0.09 0.00 reach-down 0.03 0.00 0.00 0.00 0.05 0.00 0.02 hang-down 0.06 0.00 0.01 0.00 0.02 0.01 0.03 hang-side 0.00 0.00 0.01 0.00 0.00 0.01 0.00 hang- upsidedown 0.00 0.01 0.03 0.00 0.00 0.17 0.01 probe 0.01 0.00 0.01 0.02 0.08 0.02 0.01 pull/bite 0.00 0.01 0.01 0.00 0.04 0.03 0.01 18 sally 0.05 0.16 0.00 0.02 0.01 0.00 0.08 Substrate moss 0.14 0.01 0.05 0.00 0.14 0.01 0.08 0.28 partially-moss- covered branch 0.03 0.00 0.01 0.00 0.04 0.00 0.02 bare branch 0.01 0.00 0.19 0.01 0.16 0.01 0.03 dead leaf 0.00 0.03 0.08 0.05 0.01 0.00 0.02 leaf 0.05 0.08 0.09 0.02 0.00 0.00 0.06 flower bud 0.00 0.00 0.01 0.03 0.01 0.02 0.00 air 0.04 0.17 0.00 0.00 0.00 0.00 0.08 Perch diameter < 5 mm 0.02 0.04 0.03 0.03 0.01 0.09 0.03 0.11 5-10 mm 0.01 0.01 0.00 0.00 0.00 0.03 0.01 10-20 mm 0.05 0.00 0.00 0.03 0.00 0.01 0.03 20-30 mm 0.04 0.01 0.00 0.07 0.02 0.00 0.03 30-60 mm 0.01 0.00 0.02 0.00 0.02 0.01 0.01 60 < mm 0.01 0.00 0.00 0.00 0.00 0.00 0.00
Table 2.2. (cont.)
WPC per Foraging parameter Categories CA dimensions WPC** parameter Dim1 Dim2 Dim3 Dim4 Dim5 Dim6 (54%)* (30%) (8%) (4%) (3%) (2%) Perch angle horizontal 0.00 0.00 0.00 0.01 0.00 0.03 0.00 0.01 diagonal 0.00 0.00 0.01 0.00 0.00 0.03 0.00 vertical 0.00 0.01 0.01 0.01 0.00 0.02 0.00 Foliage density 0 0.05 0.00 0.00 0.04 0.02 0.00 0.03 0.05 0-5% 0.00 0.00 0.01 0.01 0.00 0.00 0.00 5-25% 0.01 0.00 0.00 0.02 0.00 0.00 0.00 25-75% 0.01 0.01 0.02 0.00 0.03 0.02 0.01 Foraging height < 5 m 0.00 0.00 0.02 0.03 0.02 0.00 0.01 0.02 5-10 m 0.00 0.00 0.02 0.00 0.00 0.00 0.00 10-15 m 0.00 0.01 0.04 0.00 0.01 0.00 0.01 19 15-20 m 0.00 0.00 0.00 0.00 0.00 0.00 0.00 20-30 mm 0.00 0.01 0.00 0.05 0.03 0.00 0.01 Vertical position < 5 0.02 0.00 0.00 0.01 0.00 0.01 0.01 0.05 5-6 0.02 0.00 0.00 0.00 0.05 0.00 0.01 6-7 0.00 0.00 0.01 0.02 0.03 0.00 0.00 7-8 0.00 0.00 0.01 0.02 0.08 0.04 0.01 8-9 0.00 0.00 0.02 0.00 0.00 0.01 0.00 9-10 0.03 0.00 0.01 0.05 0.00 0.00 0.02 Horizontal position inner 0.06 0.00 0.01 0.01 0.00 0.00 0.03 0.23 middle 0.12 0.01 0.01 0.01 0.01 0.00 0.07 foliage 0.03 0.11 0.00 0.00 0.01 0.01 0.05 outer 0.04 0.18 0.01 0.00 0.00 0.01 0.08
Table 2.2. (cont.)
WPC per Foraging parameter Categories CA dimensions WPC** parameter Dim1 Dim2 Dim3 Dim4 Dim5 Dim6 (54%)* (30%) (8%) (4%) (3%) (2%) Habitat primary forest 0.00 0.00 0.00 0.00 0.00 0.17 0.00 0.05 secondary forest 0.00 0.02 0.03 0.15 0.01 0.00 0.01 secondary growth 0.00 0.00 0.04 0.01 0.02 0.00 0.01 semiopen 0.00 0.04 0.08 0.23 0.03 0.01 0.03 orchard /garden 0.00 0.00 0.00 0.01 0.00 0.00 0.00
* Percent explained by each dimension was shown in parentheses. 20 ** WPC: the weighted partial contribution of each foraging variable was calculated as the sum of the multiplications of partial contribution to each dimension by the percent explained by each dimension.
(2) T. nigroviridis foraged on thin branches in foliage by using glean and hang-down; (3) T.
labradorides and T. rufigula searched leaf surfaces by using glean and reach-up; and (4) T.
cyanicollis and T. ruficervix mostly sallied into air (Figs. 2.3-2.5). Dimension 4 separated these
pairs of Tangara species along a forest/non-forest habitat gradient, and assigned T. cyanicollis as
a non-forest habitat-user (Fig. 2.2 and 2.6).
Species found close together in figure 2.2, T. parzudakii - T. arthus and T. rufigula - T.
labradorides, were separated in other dimensions. Tangara parzudakii and T. arthus reached the maximum distance in dimension 5, which was weighted by moss and bare branch “substrate”
(Table 2.2). Tangara parzudakii searched moss and partially moss-covered branch in 94% of the foraging observations in contrast to 63% in T. arthus (Fig. 2.4; difference highly significant in a
G-test of independence with William’s correction; G1 = 32.0, P < 0.0001). Tangara parzudakii often probed into thick moss or pulled away pieces of moss (Fig. 2.3; 30%). These subsurface maneuvers and substrate manipulations were rare for T. arthus, which usually searched the surface of mossy or bare branches (Fig. 2.4). Tangara rufigula and T. labradorides reached the maximum distance in dimension 6, which was weighted by the hang-upsidedown “attack maneuver.” Tangara rufigula used significantly more acrobatic attack maneuvers, such as hang- upsidedown and sally, than did T. labradorides (Fig. 2.3; 22% vs. 7%, G1 = 5.8, P < 0.05). 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 did T. labradorides (76% vs. 53%, G1 = 3.9, P < 0.05, G-test of independence with William’s correction, with 2 X 2 contingency table). In addition, T. rufigula
caught arthropods from the large leaves of Cecropia gabrielis significantly more often than did
T. labradorides (24% vs. 3%, G1 = 7.4, P < 0.01). In Mindo, T. labradorides was found mainly
21 100
sally 80 pull probe 60 hang- upsidedown
hang-side 40 hang-down reach-down reach-out
Percent Attack Maneuvers Attack Percent 20 reach-up glean
0 A C L N P X R (104) (31) (42) (44) (78) (49) (55) Species
Figure 2.3. Attack maneuvers used by seven Tangara species for arthropod foraging in Mindo, Ecuador. A = Tangara arthus, C = T. cyanicollis, L = T. labradorides, N = T. nigroviridis, P = T. parzudakii, X = T. ruficervix, and R = T. rufigula. Numbers in parentheses are the sample sizes of independent foraging observations.
22 100
80 air flower bud 60 dead leaf leaf branch 40 partially -moss
Percent Substrate Use covered 20 -branch moss
0 A C L N P X R (104) (31) (42) (44) (78) (49) (55) Species
Figure 2.4. Substrate used by seven Tangara species for arthropod foraging in Mindo, Ecuador. A = Tangara arthus, C = T. cyanicollis, L = T. labradorides, N = T. nigroviridis, P = T. parzudakii, X = T. ruficervix, and R = T. rufigula. Numbers in parentheses are the sample sizes of independent foraging observations.
23 100
80
60 outer foliage middle 40 inner
20 Percent Horizontal Foraging Position
0 A C L N P X R (100) (31) (41) (44) (76) (49) (54) Species
Figure 2.5. Horizontal position used by seven Tangara species for arthropod foraging in Mindo, Ecuador. A = Tangara arthus, C = T. cyanicollis, L = T. labradorides, N = T. nigroviridis, P = T. parzudakii, X = T. ruficervix, and R = T. rufigula. Numbers in parentheses are the sample sizes of independent foraging observations.
24 100
80
60 scrub semiopen primary 40 forest secondary forest Percent Habitat Use Habitat Percent
20
0 ACLNPXR (101) (31) (41) (40) (76) (48) (54) Species
Figure 2.6. Habitat use by seven Tangara species for arthropod foraging in Mindo, Ecuador. A = Tangara arthus, C = T. cyanicollis, L = T. labradorides, N = T. nigroviridis, P = T. parzudakii, X = T. ruficervix, and R = T. rufigula. Numbers in parentheses are the sample sizes of independent foraging observations.
25 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 Tangara species that fed at foliage and the outer horizontal position (Table 2.3). Tangara arthus and T. parzudakii also foraged in the canopy less often than the other species. Tangara labradorides and T. rufigula, which mostly searched leaves for arthropods, foraged at the highest average foliage densities (Table 2.3). The most acrobatic T. rufigula used horizontal perches less often than the others. Tangara ruficervix, which sallied into the air in the forest canopy, used more often 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 dimensions explained 83% of the total variance (Fig. 2.7). All dimensions except dimension 5 were weighted by different “plant genera”(Table 2.4). This foraging parameter alone explained 0.43 of the weighted partial contribution, followed by “attack maneuver” (0.16) and “habitat” (0.13). Besides fruit genera, dimensions 1, 2, and 5 were also weighted by “attack maneuver,” and dimensions 2, 3, and 4 by “habitat.” Dimension 1 separated
Tangara arthus, T. parzudakii, and T. ruficervix, with a high proportion of Cecropia fruit in their diet and used more often hang-side “attack maneuver,” from T. labradorides and T. nigroviridis,
26 Table 2.3. Foraging site characteristics of seven Tangara species in Mindo, Ecuador. All data are mean ± SD.
Arthropod foraging Fruit foraging Species Perch Foraging Foliage Perch Foraging Foliage diameter height density diameter height density (mm)a (m)b (%)c (mm)d (m)e (%)f Tangara arthus 20.6 ± 16.0 10.7 ± 4.0 9 ± 10 8.2 ± 7.6 11.2 ± 4.4 21 ± 12 T. cyanicollis 6.0 ± 2.8 10.7 ± 4.7 15 ± 11 6.3 ± 4.5 12.1 ± 5.8 21 ± 10 T. labradorides 5.2 ± 1.9 12.5 ± 4.8 22 ± 14 4.7 ± 1.4 12.5 ± 4.3 23 ± 11 T. nigroviridis 5.2 ± 2.7 11.4 ± 3.8 13 ± 13 4.7 ± 1.4 10.1 ± 4.2 24 ± 11 T. parzudakii 22.3 ± 18.7 11.2 ± 4.6 6 ± 7 8.5 ± 9.6 11.1 ± 5.0 20 ± 13 T. ruficervix 7.8 ± 5.4 12.0 ± 4.0 16 ± 13 10.3 ± 11.2 14.6 ± 7.1 17 ± 11 T. rufigula 7.7 ± 6.7 11.3 ± 5.0 20 ± 11 6.1 ± 4.2 12.9 ± 3.9 19 ± 10
a b c ANOVA: F6,337 = 19.9, P < 0.0001. F6,392 = 4.6, P < 0.0001. F6,365 = 15.5, P < 0.0001. d e f F6,751 = 7.7, P < 0.0001. F6,798 = 2.7, P < 0.05. F6,786 = 4.6, P < 0.0001. 27
28 Table 2.4. Partial contributions of foraging variables to inertia of fruit foraging in Mindo, Ecuador.
WPC per Foraging parameter Categories CA dimensions WPC parameter Dim1 Dim2 Dim3 Dim4 Dim5 Dim6 (33%) (26%) (24%) (9%) (4%) (3%) Attack maneuver glean 0.03 0.00 0.02 0.00 0.03 0.00 0.02 0.16 reach-up 0.00 0.01 0.00 0.00 0.06 0.00 0.01 reach-out 0.01 0.00 0.06 0.00 0.04 0.06 0.02 reach-down 0.03 0.00 0.01 0.00 0.08 0.00 0.02 hang-down 0.01 0.00 0.00 0.03 0.01 0.05 0.01 hang-side 0.10 0.00 0.00 0.01 0.03 0.00 0.04 hang- upsidedown 0.00 0.01 0.00 0.02 0.06 0.02 0.01 probe 0.01 0.11 0.03 0.01 0.00 0.01 0.04 sally 0.00 0.00 0.03 0.01 0.00 0.00 0.01 29 Fruit genera Acnistus 0.00 0.05 0.04 0.01 0.03 0.02 0.03 0.43 Banara 0.00 0.00 0.04 0.00 0.00 0.00 0.01 Bocconia 0.00 0.01 0.08 0.00 0.01 0.00 0.02 Brunellia 0.00 0.03 0.05 0.02 0.02 0.00 0.02 Cavendishia 0.01 0.00 0.00 0.00 0.00 0.00 0.00 Cecropia 0.17 0.00 0.00 0.01 0.01 0.00 0.06 Cestrum 0.00 0.00 0.00 0.03 0.00 0.01 0.00 Cordia 0.00 0.08 0.02 0.06 0.00 0.04 0.03 Coussapoa 0.00 0.01 0.01 0.00 0.01 0.02 0.01 Eugenia 0.00 0.09 0.00 0.03 0.01 0.05 0.03 Ficus 0.01 0.02 0.00 0.00 0.00 0.01 0.01 Freziera 0.00 0.00 0.00 0.03 0.00 0.01 0.00 Hedyosmum 0.02 0.00 0.00 0.05 0.00 0.01 0.01 Hyeronima 0.00 0.00 0.03 0.02 0.02 0.01 0.01 Miconia 0.08 0.04 0.01 0.06 0.00 0.01 0.05
Table 2.4. (cont.)
WPC per Foraging parameter Categories CA dimensions WPC parameter Dim1 Dim2 Dim3 Dim4 Dim5 Dim6 (33%) (26%) (24%) (9%) (4%) (3%) Myrsine 0.00 0.01 0.02 0.00 0.00 0.04 0.01 Oreopanax 0.00 0.01 0.06 0.00 0.00 0.02 0.02 Palicourea 0.01 0.02 0.15 0.00 0.02 0.00 0.04 Solanum 0.01 0.00 0.05 0.01 0.02 0.08 0.02 Trema 0.01 0.06 0.00 0.10 0.03 0.00 0.03 Turpinia 0.01 0.00 0.00 0.01 0.04 0.02 0.01 Perch diameter < 5 mm 0.07 0.00 0.00 0.00 0.06 0.01 0.03 0.10 5-10 mm 0.01 0.01 0.00 0.00 0.01 0.03 0.01 10-20 mm 0.06 0.00 0.00 0.00 0.05 0.02 0.02 20-30 mm 0.05 0.00 0.00 0.00 0.05 0.03 0.02 30 30 < mm 0.05 0.01 0.00 0.03 0.01 0.00 0.02 Perch angle horizontal 0.00 0.00 0.01 0.00 0.01 0.00 0.00 0.04 diagonal 0.01 0.00 0.01 0.01 0.00 0.00 0.01 vertical 0.07 0.00 0.01 0.02 0.03 0.01 0.03 Foliage density 0-5% 0.02 0.01 0.00 0.00 0.02 0.00 0.01 0.03 5-25% 0.01 0.00 0.00 0.02 0.00 0.02 0.01 25-75% 0.02 0.00 0.01 0.03 0.00 0.03 0.01 Foraging height < 5 m 0.00 0.09 0.01 0.02 0.01 0.00 0.03 0.06 5-10 m 0.00 0.00 0.00 0.01 0.03 0.02 0.00 10-15 m 0.01 0.02 0.00 0.01 0.00 0.02 0.01 15-20 m 0.02 0.00 0.01 0.00 0.00 0.00 0.01 20-30 m 0.00 0.00 0.00 0.05 0.04 0.01 0.01
Table 2.4. (cont.)
WPC per Foraging parameter Categories CA dimensions WPC parameter Dim1 Dim2 Dim3 Dim4 Dim5 Dim6 (33%) (26%) (24%) (9%) (4%) (3%) Vertical position < 5 0.01 0.01 0.00 0.06 0.00 0.04 0.01 0.05 5-6 0.01 0.00 0.01 0.04 0.01 0.02 0.01 6-7 0.00 0.00 0.00 0.04 0.00 0.00 0.00 7-8 0.01 0.01 0.00 0.00 0.02 0.10 0.01 8-9 0.00 0.00 0.01 0.00 0.00 0.05 0.01 9-10 0.00 0.00 0.02 0.01 0.03 0.00 0.01 Horizontal position foliage 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 middle 0.01 0.00 0.00 0.01 0.00 0.00 0.00 primary Habitat forest 0.00 0.00 0.03 0.09 0.01 0.02 0.02 0.13 31 secondary forest 0.00 0.07 0.02 0.00 0.00 0.00 0.02 secondary growth 0.02 0.01 0.04 0.00 0.00 0.05 0.02 semiopen 0.01 0.08 0.00 0.01 0.01 0.00 0.02 orchard/ garden 0.01 0.07 0.09 0.02 0.04 0.00 0.05
which had a high proportion of Miconia fruits and glean “attack maneuver” (Figs. 2.8 and 2.9).
Dimension 2 separated non-forest Tangara cyanicollis and T. rufigula from the other species
(Figs. 2.7 and 2.10). Dimension 3 separated T. rufigula, with a high proportion of Palicourea
and Bocconia fruit, from T. cyanicollis (Figs. 2.7 and 2.9). Tangara labradorides and T.
nigroviridis were separated to the greatest degree in dimension 5, which was weighted by reach- down “attack maneuver.” Tangara labradorides used reach-out and reach-down more often than did T. nigroviridis, although the difference was not significant (G-test of independence with
William’s correction; G1 = 2.5, P = 0.11). Tangara arthus, T. parzudakii, and T. ruficervix achieved the most segregation in dimensions 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 its thick fruit, thus used thicker perches (Table 2.3) 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 that most Trema micrantha were over 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 Tangara tanagers did not differ significantly in foraging height or vertical position when they fed on the same fruit species: Miconia spp. and Trema micrantha (Table 2.5). Miconia 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. 2.9), and Miconia and Trema micrantha combined accounted for nearly two- thirds or more of all fruit-foraging observations (63 ± 19%, range 44 - 94%; Fig. 2.9).
32 100
80 sally bite hang- 60 upsidedown hang-side hang-down 40 reach-down reach-out reach-up Percent Attack Maneuvers Attack Percent 20 glean
0 A C L N P X R (207) (93) (64) (71) (115) (146) (71) Species
Figure 2.8. Attack maneuvers used by seven Tangara species for fruit foraging in Mindo, Ecuador. A = Tangara arthus, C = T. cyanicollis, L = T. labradorides, N = T. nigroviridis, P = T. parzudakii, X = T. ruficervix, and R = T. rufigula. Numbers in parentheses are the sample sizes of independent foraging observations.
33 100
80 others Oreopanax Palicourea 60 Hyeronima Solanum flower bud Myrsine Hedyosmun 40 Ficus Acnistus Brunellia Eugenia
Percent Fruit Genera Eaten Fruit Genera Percent 20 Cordia Cecropia Trema Miconia 0 A C L N P X R (207) (95) (64) (71) (116) (172) (80) Species
Figure. 2.9. Fruits eaten by seven Tangara species in Mindo, Ecuador. A = Tangara arthus, C = T. cyanicollis, L = T. labradorides, N = T. nigroviridis, P = T. parzudakii, X = T. ruficervix, and R = T. rufigula. Numbers in parentheses are the sample sizes of independent foraging observations.
34 100
80
60 scrub semiopen secondary 40 forest primary forest Percent Habitat Use Habitat Percent
20
0 ACLNPXR (197) (88) (62) (67) (113) (138) (67) Species
Figure. 2.10. Habitat use by seven Tangara species for fruit foraging in Mindo, Ecuador. A = Tangara arthus, C = T. cyanicollis, L = T. labradorides, N = T. nigroviridis, P = T. parzudakii, X = T. ruficervix, and R = T. rufigula. Numbers in parentheses are the sample sizes of independent foraging observations.
35
Table 2.5. Foraging site characteristics of seven Tangara species when feeding on Miconia and Trema micrantha.
Miconia Trema micrantha Foraging Foraging height Vertical height Vertical Species n (m)a positionb n (m)c positiond Tangara arthus 103 10.9 ± 3.6 8.4 ± 1.7 24 11.4 ± 3.8 8.0 ± 1.6 T. cyanicollis 35 11.4 ± 4.0 8.6 ± 1.3 6 11.3 ± 4.7 8.8 ± 1.2 T. labradorides 42 12.9 ± 4.7 8.6 ± 1.2 13 13.7 ± 5.0 8.4 ± 1.7 T. nigroviridis 54 11.4 ± 3.7 8.7 ± 1.4 12 12.1 ± 4.4 8.7 ± 1.2 T. parzudakii 39 11.5 ± 3.8 8.6 ± 1.3 18 14.2 ± 5.6 8.7 ± 1.4 T. ruficervix 48 12.2 ± 4.0 8.7 ± 1.5 47 11.6 ± 3.8 8.4 ± 1.6 T. rufigula 29 10.8 ± 4.3 9.3 ± 1.0 11 13.6 ± 5.1 8.8 ± 1.0 a b c ANOVA: F6,343 = 1.7, P = 0.13. F6,343 = 1.4, P = 0.21. F6,126 = 1.3, P = 0.25. d F6,126 = 0.7, P = 0.63.
36
In short, CA separated seven Tangara species to four species groups based on 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 non-forest species (T. cyanicollis); and (4) one
foothill species (T. rufigula) that reaches Mindo as its highest distribution limit.
Overlap in Resource Use
The resource overlap between 21 species-pairs was x ± SD = 0.28 ± 0.28 (range 0.00-
0.93; Table 2.6) in arthropod foraging and x = 0.81 ± 0.13 (range 0.48-0.97; Table 2.7) in fruit
foraging. The mean overlap in fruit foraging was 2.9 times higher than in arthropod foraging. In
fruit foraging, most species-pairs showed resource overlap higher than 0.70, with the exception
of T. cyanicollis, a non-forest species, which showed lower overlap in fruit foraging with the
other forest Tangara species (Table 2.7). The average observed niche overlap was significantly
larger than the expected by chance (Table 2.8). In contrast, in arthropod foraging, only two
species-pairs showed resource overlap larger than 0.70 (Table 2.6). 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 of arthropod foraging. The average observed niche overlap in arthropod foraging was significantly smaller than the expected (Table 2.8).
COSTA RICA
At my Costa Rica site, seven Tangara species were present. Tangara dowii is a common resident in mountains at 1,200-2,750 m (Stiles and Skutch 1989) and was found occasionally only at the highest part of the observation road . Tangara inornata had been recorded at lower than 400 m, although we saw a pair visiting my study area at 900 m high for a few days. The other five species were permanent residents in the study area, and I obtained 1,311 foraging
37
Table 2.6. Matrix of niche overlap values in arthropod foraging by using three-way tables constructed by species x substrate x habitat.
Species A C L N P X Tangara arthus (A) T. cyanicollis (C) 0.06 T. labradorides (L) 0.06 0.14 T. nigroviridis (N) 0.47 0.11 0.56 T. parzudakii (P) 0.78 0.00 0.00 0.09 T. ruficervix (X) 0.05 0.49 0.54 0.34 0.00 T. rufigula (R) 0.04 0.16 0.93 0.49 0.00 0.48
*The niche overlap values larger than 0.70 are underlined.
38
Table 2.7. Matrix of niche overlap values in fruit foraging by using three-way tables constructed by species x fruit genus x habitat.
Species A C L N P X Tangara arthus (A) T. cyanicollis (C) 0.72 T. labradorides (L) 0.92 0.68 T. nigroviridis (N) 0.92 0.69 0.97 T. parzudakii (P) 0.94 0.66 0.87 0.80 T. ruficervix (X) 0.84 0.48 0.82 0.73 0.93 T. rufigula (R) 0.89 0.66 0.94 0.91 0.89 0.85
*The niche overlap values smaller than 0.70 are underlined.
39
Table 2.8. Monte Carlo approach using EcoSim7 (Gotelli and Entsminger 2003).
Arthropod foraging Fruit foraging observed expected observed expected mean mean ± SE mean mean ± SE Ecuador 0.28 0.36 ± 0.00 0.81 0.45 ± 0.00 P(obs.
Costa Rica 0.22 0.42 ± 0.00 0.80 0.55 ± 0.00 P(obs.
Bolivia 0.26 0.34 ± 0.00 0.89 0.48 ± 0.00 P(obs.
40
observations from them. Fruits and arthropods accounted for 78% and 20%, respectively, of
these foraging observations (Table 2.9).
Arthropod Foraging
Because of the smaller number of Tangara species at the Costa Rican site,
correspondence analysis provided only four dimensions, and the first two accounted for 79% of
total variation (Fig. 2.11). “Substrate,” “horizontal position,” and “attack maneuver” explained
0.37, 0.17, and 0.15 of the weighted partial contributions, respectively.
CA1 was weighted by leaf and moss “substrate”, and middle “horizontal position,” and
separated 5 Tangara species to branch foragers, which used the middle part of “horizontal
position” more often, from leaf and aerial foragers, which used the tips of branches (Figs. 2.11-
2.13). CA2 was weighted by leaf and air “substrate,” outer “horizontal position,” and sally
“attack maneuver,” and separated a leaf gleaner and aerial forager (Figs. 2.11-2.14).
Three branch foragers differed in their use of moss (Fig. 2.12). Tangara florida foraged
on moss-covered branches in wet, dense forests and was often observed picking up a piece of
moss and searching inside moss for arthropods (Fig. 2.14). In contrast, Tangara gyrola searched
mostly the undersides of bare branches by using reach-down and hang-down attack maneuver
(Figs. 2.12 and 2.13). Tangara icterocephala was intermediate between the former two species.
It mostly searched the undersides of partially-moss covered branches in forested areas, but without manipulating moss (Figs. 2.12 and 2.14).
Fruit Foraging
The first two dimensions accounted for 79% of total variation (Fig. 2.15). “Fruit genera”
(0.45), “habitat” (0.16), and “vertical position” (0.13) were the most important foraging parameters in fruit foraging. CA1 was weighted by Conostegia “fruit genera,” semiopen
41 Table 2.9. Body mass and food types used by five Tangara species in El Copal, Costa Rica.
Percent in each food category Species Code Body mass (g)a n Arthropod Fruit Flower Müllerian Nectar mean ± SD (n) bud body Tangara florida F 18.6 ± 1.1 (13) 161 19.3 80.7 0.0 0.0 0.0 T. guttata G 21.0 ± 1.7 (11) 331 23.6 74.9 0.3 0.0 1.2 T. gyrola Y 25.0 ± 3.0 (10) 336 17.6 82.4 0.0 0.0 0.0 T. icterocephala I 22.3 ± 1.9 (10) 240 12.9 87.1 0.0 0.0 0.0 T. larvata B 18.4 ± 1.4 (24) 243 28.0 65.4 0.4 3.7 2.5
a Taken from the specimen labels at Louisiana State University Museum of Natural Science.
42
43 100
80
air flower bud 60 leaf dead leaf bare branch 40 partially-moss -covered branch moss Percent Substrate Use 20
0 FGY I B (31) (78) (59) (31) (68) Species
Figure 2.12. Different substrate used by five Tangara species for arthropod foraging in El Copal, Costa Rica. F = Tangara florida, G = T. guttata, Y = T. gyrola, I = T. icterocephala, and B = T. larvata. Numbers in parentheses are the sample sizes of independent foraging observations.
44 100
80
60
outer foliage 40 middle inner
20 Percent Horizotal Foraging Position
0 FGY I B (31) (78) (56) (31) (67) Species
Figure 2.13. Horizontal position use by five Tangara species for arthropod foraging in El Copal, Costa Rica. F = Tangara florida, G = T. guttata, Y = T. gyrola, I = T. icterocephala, B = T. larvata. Numbers in parentheses are the sample sizes of independent foraging observations.
45 100
80 sally pull/bite probe 60 hang- upsidedown hang-side 40 hang-down reach-down reach-out reach-up Percent Attack Maneuvers 20 glean
0 FGY I B (31) (78) (59) (31) (68) Species
Figure 2.14. Attack maneuvers used by five Tangara species for arthropod foraging in El Copal, Costa Rica. F = Tangara florida, G = T. guttata, Y = T. gyrola, I = T. icterocephala, and B = T. larvata. Numbers in parentheses are the sample sizes of independent foraging observations.
46
47 “habitat,” and < 5 “vertical position,” and separated T. larvata and T. gyrola, which used non- forest habitat more often, from the other three species, which are more forest-dependent and for which Miconia accounted for more than 50 % of fruits eaten (Figs. 2.16 and 2.17). CA2 was weighted by Ficus and separated T. gyrola from the other species (Fig. 2.15 and 2.17).
The three forest species were separated in CA3 and CA4, which were weighted by
Cecropia and Viburnum respectively. Tangara guttata ate more fruits of Viburnum and
Cecropia than did the other two species (Fig. 2.17).
Overlap in Resource Use
The resource overlap between 10 species-pairs was x ± SD = 0.22 ± 0.28 (range: 0.01-
0.93; Table 2.10) in arthropod foraging and 0.80 ± 0.11 (0.68-0.99; Table 2.11) in fruit foraging.
The mean overlap in fruit foraging was 3.6 times higher than in arthropod foraging. In fruit foraging, all species-pairs showed resource overlap higher than 0.60 whereas only T. icterocephala and T. florida showed resource overlap larger than 0.60 in arthropod foraging. In fruit foraging the average observed niche overlap was significantly larger than the expected by chance whereas in arthropod foraging the average observed niche overlap was significantly smaller than the expected (Table 2.8).
BOLIVIA
At Serranía Bella Vista, I recorded 14 Tangara species and collected 1,591 foraging observations. Fruits and arthropods accounted for 69% and 31%, respectively, of all foraging observations (Table 2.12). Tangara argyrofenges and T. mexicana were occasional visitors: a small group of three individuals of T. argyrofenges was observed for three days in September
2000, and a pair of T. mexicana was observed during two consecutive days in July 2002. Two lowland species: T. chilensis and T. schrankii, were found mostly in disturbed areas and rarely
48 100
80
orchard /garden 60 semiopen secondary growth secondary 40 forest primary forest Percent Habitat Use Habitat Percent
20
0 FGY I B (126) (241) (259) (195) (153) Species
Figure 2.16. Habitat use by five Tangara species for fruit foraging in El Copal, Costa Rica. F = Tangara florida, G = T. guttata, Y = T. gyrola, I = T. icterocephala, and B = T. larvata. Numbers in parentheses are the sample sizes of independent foraging observations.
49 100
Others Viburnum 80 Tovomita Tetrochidium Saurauia Sabicea 60 Rubus Psychotria Phytolacca Phoradendron 40 Myrcia Miconia Marcgravia Hedyosmum
Percent Fruit Genera Eaten 20 Gonzalagunia Ficus Dendropanax Conostegia 0 Cecropia FGY I B Cavendishia (130) (249) (279) (209) (160) Acnistus Species
Figure 2.17. Fruits eaten by five Tangara species in El Copal, Costa Rica. F = Tangara florida, G = T. guttata, Y = T. gyrola, I = T. icterocephala, and B = T. larvata. Numbers in parentheses are the sample sizes of independent foraging observations.
50
Table 2.10. Matrix of niche overlap values in arthropod foraging in Costa Rica by using three-way tables constructed by species x substrate x habitat.
F G Y I Tangara florida (F) T. guttata (G) 0.01 T. gyrola (Y) 0.25 0.05 T. icterocephala (I) 0.93 0.07 0.34 T. larvata (B) 0.02 0.37 0.14 0.04
*The niche overlap values larger than 0.70 are underlined.
51
Table 2.11. Matrix of niche overlap values in fruit foraging in Costa Rica by using three-way tables constructed by species x attack maneuver x fruit genus.
F G Y I Tangara florida (F) T. guttata (G) 0.99 T. gyrola (Y) 0.73 0.72 T. icterocephala (I) 0.95 0.93 0.80 T. larvata (B) 0.70 0.74 0.68 0.73
52
Table 2.12. Body mass and food types of ten Tangara species in Serranía Bella Vista, Bolivia.
Percent in each food category Species Code Body mass (g)a n Arthropod Fruit Müllerian Nectar mean ± SD (n) body T. arthus A 21.4 ± 1.2 (13) 200 27.5 72.5 0.0 0.0 T. chilensis H 21.8 ± 1.8 (24) 47 14.9 85.1 0.0 0.0 T. chrysotis D 25.2 ± 1.7 (10) 93 38.7 61.3 0.0 0.0 T. cyanicollis C 17.3 ± 1.2 (20) 271 27.7 71.6 0.4 0.4 T. cyanotis T 15.7 ± 0.9 (15) 77 54.5 45.5 0.0 0.0 T. gyrola Y 20.6 ± 1.6 (21) 83 27.7 72.3 0.0 0.0 T. nigroviridis N 15.7 ± 0.9 (11) 235 20.4 79.6 0.0 0.0 T. punctata Q 15.0 ± 1.0 (16) 232 30.6 69.0 0.4 0.0 T. xanthocephala W 20.3 ± 1.8 (16) 253 34.0 65.2 0.0 0.8 T. xanthogastra Z 13.1 ± 1.5 (20) 83 50.6 49.4 0.0 0.0
53 a Taken from the specimen labels at Louisiana State University Museum of Natural Science.
inside the forest. Tangara ruficervix was uncommon. I obtained more than 70 foraging
observations for each of the other nine species.
Arthropod Foraging
The first three dimensions of correspondence analysis accounted 86% of the total
variation (Fig. 2.18). “Substrate”, “attack maneuver”, and “horizontal position” contributed
0.31, 0.21, and 0.19 of weighted partial contribution respectively. CA1 was weighted by
partially-moss-covered branch and leaf “substrate” and middle “horizontal position”, and ordered
nine Tangara species according to substrate use from thick branch, moss forager (T. chrysotis) to
aerial forager (T. cyanicollis) (Figs. 2.18 - 2.20). CA2 was heavily weighted by sally “attack
maneuver”, air “substrate”, and outer “horizontal position” (Figs. 2.19 - 2.21), and separated T.
cyanicollis from the other species (Fig. 2.18). CA3 was weighted by bare-branch “substrate”,
foliage and inner “horizontal position”, and pull/bite “attack maneuver”, and ordered five species
of branch foragers according to their dependence on moss (Fig. 2.18).
Among these five branch-foragers, T. cyanotis used mostly twigs (thinner branch in Table
2.13 and foliage in Fig. 2.20). Tangara gyrola mostly searched the undersides of bare branches
or bare parts of partially-moss covered branches (Fig. 2.19) by using reach-down and hang-down
attack maneuver (Fig. 2.21). Tangara chrysotis, T. arthus, and T. xanthocephala used moss and
partially-moss covered branches in similar proportions (95%, 94%, and 87% respectively in Fig.
2.19, difference not significant, G = 2.9, P = 0.4); however, T. chrysotis used more subsurface
and substrate-modifying attack maneuvers, namely probe and pull/bite, than did T. arthus or T. xanthocephala (G-test 2 X 2; G = 24.0, 19.3, P < 0.0001 respectively; Fig. 2.21). I did not find significant differences in foraging ecology between T. xanthocephala and T. arthus. However,
54
55 56 57 58
Table 2.13. Percent of perch diameters used by nine Tangara species for arthropod foraging in Serranía Bella Vista, Bolivia.
Species n < 5 mm 5-10 mm 10-20 mm 20-30 mm 30-60 mm 60 < mm T. arthus 55 4 24 31 29 13 0 T. chrysotis 35 0 17 20 37 20 6 T. cyanicollis 33 42 51 3 3 0 0 T. cyanotis 42 17 52 29 0 2 0 T. gyrola 32 6 38 16 31 9 0 T. nigroviridis 43 56 40 5 0 0 0 T. punctata 63 54 46 0 0 0 0 T. xanthocephala 84 6 39 18 27 8 1 T. xanthogastra 29 31 66 4 0 0 0
59
T. xanthocephala was mostly found at higher elevation of Serranía Bella Vista above 1,400 m,
whereas T. arthus was found mostly at lower elevation.
Three Tangara species (T. nigroviridis, T. punctata, and T. xanthogastra) searched leaves
and flowers in over 60% of their arthropod foraging. Among these three, T. nigroviridis spent
more time on searching thin branches than did T. punctata or T. xanthogastra, which searched
leaf surfaces almost exclusively (Fig. 2.19, pers. obs.). In addition, T. xanthogastra searched
large leaves by using more acrobatic attack maneuvers, namely hang-upsidedown and sally, than
did the other two species (Fig. 2.21).
Fruit Foraging
The first four dimensions of CA accounted for 75% of the total variation. Overall fruit
foraging was weighted by “fruit genera”, which explained 0.37 of relative variation, followed by
“attack maneuver” (0.13), “habitat” (0.11), and “vertical position” (0.11). CA1 was weighted by
Guettarda “fruit genus,” and separated Tangara chrysotis, T. cyanotis, and T. gyrola, which
hardly used secondary growth (Fig. 2.22 and 2.23), and ate Guettarda fruits more often than the
other species (Fig. 2.24). CA2 was weighted by horizontal and diagonal “perch angle” and
Solanum “fruit genus.” It split the three species (Fig. 2.22 and 2.24). CA3 was weighted by
semiopen “habitat” and 5-10 mm “perch diameter”, and separated T. chrysotis and T.
xanthogastra.
Overlap in Resource Use
The resource overlap between 36 species-pairs in arthropod foraging was x ± SD = 0.26
± 0.28 (range: 0.00-0.87; Table 2.14) and in fruit foraging was 0.89 ± 0.07 (0.70-0.99; Table
2.15). Thus, mean overlap in fruit foraging was 3.4 times higher than that in arthropod foraging.
Five species-pairs showed resource overlap larger than 0.70 in arthropod foraging (Table 2.14).
60
61 62 63
Table 2.14. Matrix of niche overlap values in arthropod foraging in Serranía Bella Vista, Bolivia by using three-way tables constructed by species x substrate x habitat.
Species A D C T Y N Q W T. arthus (A) T. chrysotis (D) 0.82 T. cyanicollis (C) 0.00 0.00 T. cyanotis (T) 0.26 0.12 0.06 T. gyrola (Y) 0.40 0.17 0.03 0.81 T. nigroviridis (N) 0.09 0.06 0.18 0.52 0.43 T. punctata (Q) 0.00 0.00 0.28 0.20 0.10 0.87 T. xanthocephala (W) 0.77 0.59 0.03 0.29 0.42 0.13 0.05 T. xanthogastra (Z) 0.00 0.00 0.39 0.06 0.02 0.53 0.79 0.04
*The niche overlap values larger than 0.70 are underlined.
64
Table 2.15. Matrix of niche overlap values in fruit foraging in Serranía Bella Vista, Bolivia by using three-way tables constructed by species x fruit genus x habitat.
Species A D C T Y N Q W T. arthus (A) T. chrysotis (D) 0.85 T. cyanicollis (C) 0.97 0.87 T. cyanotis (T) 0.94 0.94 0.92 T. gyrola (Y) 0.80 0.93 0.85 0.90 T. nigroviridis (N) 0.97 0.74 0.93 0.88 0.70 T. punctata (Q) 0.99 0.85 0.97 0.94 0.81 0.98 T. xanthocephala (W) 0.98 0.83 0.97 0.93 0.80 0.97 0.99 T. xanthogastra (Z) 0.92 0.86 0.94 0.85 0.81 0.84 0.90 0.87
65
The average observed overlap in arthropod foraging was significantly smaller than expected by
chance (Table 2.14). In contrast, in fruit foraging, all species pairs showed niche overlap greater
than 0.70, and the average observed overlap in fruit foraging was significantly larger than the
expected by chance (Table 2.15).
COMPARISON AMONG THREE STUDY SITES
Three parameters in arthropod foraging (“substrate,” “horizontal position,” “attack
maneuver”) and four parameters in fruit foraging (“fruit genera,” “attack maneuver,” “habitat,”
“vertical position”) contributed more than 70% of the variation in foraging ecology at all three
study sites (Table 2.16). Overlap among sympatric species in both arthropod and fruit foraging
was similar among these three sites: niche overlap in fruit foraging was 2.9-3.5 times higher than
that in arthropod foraging (Table 2.16). The observed overlap in fruit foraging was significantly
larger than expected by chance at all three study sites, whereas observed overlap in arthropod
foraging was significantly smaller than expected by chance at all three study sites (Table 2.16).
The proportion of species that used each substrate category in arthropod foraging was
similar among the three sites (Table 2.17), especially after grouping “moss” and “partially-moss-
covered branch.” Ideally, foraging observations should be separated according to which part of
partially-moss-covered branch, mossy or bare part, was used. Unfortunately, many of my
foraging observations were missing information on this point. At each study site, however, one
species specialized on searching inside moss for arthropods (T. parzudakii in Ecuador, T. florida
in Costa Rica, and T. chrysotis in Bolivia), and at least one species on searching moss surfaces
(T. arthus in Ecuador, T. icterocephala in Costa Rica, T. arthus and T. xanthocephala in Bolivia;
see arthropod foraging sections of each study site). All Tangara species from the three sites ate
66
Table 2.16. The comparison of Tangara communities among three study sites - important foraging parameters and foraging overlap.
El Copal, Serranía Bella Vista, Mindo, Ecuador Costa Rica Bolivia Number of Tangara species 7 5 9 Elevation (m) 1300-1600 960-1200 1250-1600 Number of fruit genera eaten 21 21 16
The most important foraging parameters in substrate substrate substrate arthropod foraging* 1 (0.29) (0.39) (0.31) horizontal position horizontal position attack maneuver 2 (0.23) (0.17) (0.22) attack maneuver attack maneuver horizontal position 3 (0.19) (0.15) (0.19)
The most important foraging parameters in fruit genera fruit genera fruit genera fruit foraging* 1 (0.43) (0.45) (0.37) attack maneuver habitat attack maneuver 2 (0.16) (0.16) (0.13) vertical position habitat vertical position /habitat 3 (0.13) (0.13) (0.11 each) foraging overlap in arthropod foraging** 0.28 ± 0.28 0.22 ± 0.28 0.26 ± 0.28 foraging overlap in fruit foraging** 0.81 ± 0.13 0.80 ± 0.11 0.89 ± 0.07
* The parentheses indicate the relative contribution of each foraging parameter to weighted contribution. ** mean ± SD
67
Table 2.17. The comparison of Tangara communities among three study sites - proportion of species used most frequently a particular substrate or fruit genera. The parentheses indicate the number of species.
Ecuador Costa Rica Bolivia Arthropod foraging moss 0.29 (2) 0.20 (1) partially-moss-covered branch 0.20 (1) 0.33 (3) [moss + partially-moss-covered branch]* [0.29 (2)] [0.40 (2)] [0.33 (3)] bare branch 0.07 (0.5) 0.20 (1) 0.22 (2) leaf 0.36 (2.5) 0.20 (1) 0.33 (3) air 0.29 (2) 0.20 (1) 0.11 (1)
Fruit foraging (most) Miconia 1.00 (7) 1.00 (5) 1.00 (9)
Fruit foraging (2nd most) Guettarda 1.00 (9) Trema 0.50 (3.5) Ficus 0.60 (3) Cecropia 0.20 (1.5) 0.20 (1) Conostegia 0.20 (1) Eugenia 0.30 (2)
* combined moss and partially-moss-covered branch
68
Miconia fruits during the dry season, although the second most consumed fruit genus varied
among the sites (Table 2.17).
DISCUSSION
Most variation in foraging ecology of Tangara species is explained by relatively few foraging parameters. Differences in microhabitat preference characterized by the combination of
“substrate type,” “horizontal position,” and “attack maneuver” explains the major differences in arthropod foraging among sympatric Tangara species followed by differences in “habitat” choice. The observed average overlap in arthropod foraging is lower than expected by chance, and over 90% of species-pairs share less than 50% of arthropod resources. High resource overlap of six species-pairs is partially due to the coarse resolution of some foraging variables, such as leaf type and moss. The Tangara species 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 Tangara species 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 Tangara species 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 (pers. obs.). For example, in Ecuador two small species, T. nigroviridis and T. labradorides, preferably feed on small fruit species, Miconia brevitheca and
Trema 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
69 Cecropia gabrielis, which requires the “hang-side” attack maneuver at a thick Cecropia catkin.
Despite the differences in various foraging parameters, the average resource overlap among
Tangara species is higher than expected by chance in fruit foraging.
These patterns are invariably found in all three study sites: Costa Rica, western slope of
Ecuador, and eastern slope of Bolivia, where the number and composition of sympatric Tangara
species largely differ. This indicates that these patterns are common and wide spread among
omnivorous Tangara tanagers and possibly other omnivorous birds (Snow and Snow 1971).
Arthropod-foraging in Tangara species is characterized by: (1) interspecific differences in microhabitat preferences followed by difference 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. In contrast, fruit foraging is characterized by
(1) interspecific differences in plant genera 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, Tangara species that use the same space for foraging and form mixed-species flocks in the same habitat, differ largely in arthropod foraging, but overlapped greatly in fruit foraging. These conclusions are similar to the previous foraging observations of omnivorous tanagers (Isler and Isler 1999; Ridgely and Tudor 1989; Snow and
Snow 1971; 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
70 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 (Leck 1969; Willis 1966). 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 (Moermond and Denslow 1983, see Fig. 2.8); thus, syntopic congeners
do not show clear partitioning of resources and often are grouped in the same foraging guild
when analyzed together with other frugivorous genera (Loiselle and Blake 1990, unpubl. data).
The substrate types for arthropod foraging used in this study 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 groups 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
(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 this chapter was limited to the dry season, presumably the season of
minimal breeding in the Tropics. Although arthropods constituted only one-third of the foraging
observations in my study, arthropod consumption is expected to increase during the breeding
season and to match the higher protein demand for egg production and feeding of nestlings in
most fruit-eating birds (Moermond and Denslow 1985, see Chapter 4; Poulin et al. 1992). Even
during the non-breeding season, Tangara tanagers spent 60% of their foraging time searching for arthropods (see Chapter 3). Interestingly, many “frugivorous” birds breed at the same time as
71 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 dichotomy of fruits and arthropods as food resources appear to explain:
(1) apparent lack of segregation in fruit foraging among Tangara tanagers, in contrast to the fine segregation in arthropod foraging, and (2) little specialization in fruit foraging with wide variety of fruits consumed by each Tangara species in contrast to highly specialized arthropod foraging 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 higher quality than fruits.
72 CHAPTER 3 THE RELATIVE IMPORTANCE OF ARTHROPODS AND FRUITS IN FORAGING BEHAVIOR OF TANGARA TANAGERS
SUMMARY
I quantified the foraging ecology of omnivorous Tangara tanagers with three methods commonly used in the study of foraging behavior. The relative importance of two food types, arthropods and fruits, varied largely depending on which method was used for data analyses.
Arthropod foraging was more important than fruit foraging when calculated by using the duration of foraging. In contrast, fruit foraging was more important when characterized by the food taken at initial observation and the total number of food items taken. This bias probably was caused by the difference in distribution and abundance of these two food types. Although numerous studies have used the frequency of initial observations to quantify bird foraging behavior, this method tends to underestimate the importance of highly rewarding but scarce food types in time budgets and tends to overestimate the same food type in the number of food items in birds’ diets.
INTRODUCTION
Studies of foraging behavior and food resources constitute part of an overall effort to comprehend diverse aspects of avian biology such as population dynamics, community structure, ecomorphology, physiology, and predator-prey relationships (Morrison et al. 1990a). Behavioral observation generally yields four main types of measures (Martin and Bateson 1993) of which frequency and duration are most commonly used to study foraging behavior of birds (Morrison et al. 1990b). These two measures, however, represent different aspects of foraging behavior, and the observers’ preference for one measure over the other may hide or even distort a true biological pattern. In addition, various sampling methods can be used to quantify both frequency
73 and duration, and different methods can affect the estimates of mean and standard errors of foraging parameters (Hejl et al. 1990; Morrison 1984). Evidently, no study has compared the data sets assembled by using different measures in an effort to understand the potential bias associated with each measure.
Here I present quantitative analyses of two measures of foraging behavior that use three different sampling methods. These three sampling methods were “food type taken at initial observation” and “total number of food items taken” for quantifying foraging frequency, and
“duration” for foraging effort. All data were assembled from the same set of foraging observations gathered from four Tangara species in Ecuador during 1999. The diverse diet of
Tangara tanagers consists of fruits, arthropods, nectar, flower buds, and Müllerian bodies (Isler and Isler 1999; Naoki and Toapanta 2001, Chapter 2), although fruits and arthropods account for over 95% of their diet (unpubl. data: Isler and Isler 1999; Rodrigues 1995; Snow and Snow
1971). I here compare these three methods and discuss how they can lead to different conclusions concerning the relative importance of fruits and arthropods in the biology of these omnivorous tanagers.
METHODS
From October to December 1999, I quantified foraging behavior of Tangara tanagers in the vicinity of Mindo, Pichincha province, Ecuador (0°02’S, 78°46’W, 1300-1600 m). A detailed description of the study site is provided in Chapter 2. Birds were opportunistically encountered as I slowly walked along trails. At every encounter, I followed one individual for as long as possible and recorded the following data on microcassettes: time spent foraging for arthropods, time spent foraging for fruit, the number of arthropods attacked, the number of fruits attacked, and food type taken at an initial observation. Hereafter, each encounter is referred to as
74 a foraging bout. I used two stopwatches to quantify the time spent foraging for arthropods and
for fruits. Foraging behavior was defined as any behavior used to obtain food and included
searching, attacking, and handling maneuvers (Remsen and Robinson 1990). Fruit searching
was usually short in duration because Tangara tanagers flew directly to fruits and rapidly moved
from one fruit to another. When the birds hopped and stared at substrates with no fruits, such as
moss, branch bottom, or leaf surface, the searching activity was considered to be directed toward
arthropods. Each Tangara species searched a distinct substrate for arthropods by using a unique
searching maneuver (Chapter 2). I carefully excluded the time spent flying between trees and
moving between branches without typical searching maneuver to avoid inflating arthropod
foraging time. Attacks, usually brief, consisted of a quick capture attempt for arthropods or
fruits. Handling included mashing of fruit or large arthropods. Small fruits, such as from plants
of the genera Miconia and Trema, and most arthropods were swallowed without a long handling
time. The number of arthropods or fruits attacked represented the total number of individual
arthropods or fruits to which capture attempts were made. Tangara species often bit or poked one food item more than once to finish ingesting a whole or part of a large fruit or arthropod.
These multiple bites toward one food item were considered as one attack. Food type taken at initial observation was the initial capture attempt made in each foraging bout and was noted as arthropod or fruit. The initial observation is the most commonly used method to study foraging behavior of birds (e.g., Morrison et al. 1990b; Rodrigues 1995; Sillett 1994). I calculated foraging efficiency by dividing the number of arthropods or fruits attacked by the time spent arthropod and fruit foraging respectively (the number of arthropods or fruits eaten per minute).
Müllerian bodies and flower buds were included in the fruit category because they resembled fruits more than arthropods in being conspicuous and stationary.
75 STATISTICAL ANALYSES
Because most observations were short and showed large variation in duration (60.2 ±
61.2 sec; range 1 – 450, n = 267), I pooled all foraging bouts and presented the results as the
proportion of arthropod foraging to total foraging activity (Table 3.1). As a result, the data
presented here were cumulative and did not allow calculation of confidence limits by themselves.
To overcome this problem, I bootstrapped the original sample data 1000 times to calculate
confidence limits and to correct the bias of the estimators (Manly 1997). To compare two
percentages or two foraging efficiencies, I applied the modified version of bootstrap tests of
significance described in Manly (Chapter 3.10, 1997). This test consists of (1) generation of the
pseudodistributions by bootstrapping the original sample data, (2) formation of a third
distribution by comparing the two pseudodistributions, and (3) test of null hypotheses and
calculation of P-values for the third distribution. For example, to test the difference between two
percentages α% and β% (α > β), I bootstrapped the original sample data, A and B, 1000 times to
obtain 1000 Ai and Bi, representing pseudodistributions of α and β, respectively. Then I
calculated the difference between A and B for each bootstrap result: Di = Ai – Bi. P-value was
calculated as the proportion of negative Di to total Di. I conducted all the statistical analyses,
including the bootstrap procedure, by using SYSTAT 8.03 (SPSS Inc. 1998). All results are
reported as mean ± SD unless otherwise indicated.
RESULTS
During the study, Tangara tanagers were observed in mixed-species flocks in 68% of the total encounters (n = 244). The relative importance of arthropod foraging varied largely depending on which of the three measures was used to analyze the data (Table 3.1). When
“duration” was used, arthropod foraging was more important than fruit foraging: four Tangara
76
Table 3.1. Relative importance of arthropods and fruits in foraging behavior of four Tangara species in Mindo, Ecuador. Bias-corrected mean and 95% confidence limits (in parentheses) were calculated by using 1000 bootstrap replicates.
Species Percentage of arthropod foraging per total foraging activity No. of Duration Frequency foraging Initial observation No. of food items bouts Tangara arthus 57% 31% 20% 138 (48–67) (21–40) (12–27) T. parzudakii 58% 45% 23% 51 (41–74) (30–61) (9–38) T. labradorides 70% 50% 35% 41 (56–84) (32–68) (17–52) T. rufigula 64% 58% 46% 37 (49–78) (39–77) (21–70)
77 Total 59% 41% 25% 267 (53–66) (34–48) (17–33)
species spent over half of their foraging time foraging for arthropods (59%; Table 3.1). When
“initial observation” or “the number of food items” was used, fruits were more important than arthropods: fewer than half of the initial observations and of the number of food items were arthropods (41% and 25% respectively; Table 3.1). Although the four Tangara species differed in the percentage of arthropod foraging under each measure, all four showed the same trend: the percentage of arthropod foraging was the largest by using the duration of foraging, followed by the “initial observation”, and then the “number of food items.” These differences were significant in 13 out of 15 pairwise comparisons (P < 0.05): the differences between “Duration” and “Frequency of initial attack” were significant in T. arthus (P < 0.001), T. parzudakii (P <
0.05), T. labradorides (P < 0.01), and Total (P < 0.001), but not in T. rufigula (P = 0.3). The differences between “Duration” and “Frequency of total attacks” were significant in T. arthus, T. parzudakii, T. labradorides, and Total (P < 0.001), and T. rufigula (P < 0.05). The differences between “Frequency of initial attack” and “Frequency of total attacks” were significant in T. arthus and T. parzudakii (P < 0.001), T. labradorides (P < 0.01), and Total (P < 0.01), but not in
T. rufigula (P = 0.08).
Fruit foraging was 3.7-4.9 times more efficient than arthropod foraging, except for T. rufigula, which showed higher arthropod foraging efficiency than the other Tangara species
(Table 3.2). However, the data for T. rufigula were heavily influenced by a single foraging bout in which a large number of arthropods were gleaned from a spider web (19 attacks during 65 sec). When this unusual foraging bout was eliminated from the analysis, arthropod foraging efficiency of T. rufigula became 1.7 attacks per min, and fruit foraging became 2.5 times more efficient than arthropod foraging. Tangara tanagers did not attack arthropods in 48 percent of
78
Table 3.2. Foraging efficiency and percentage of foraging bouts without an attack in Tangara tanagers. Bias-corrected mean and 95% confidence limits (in parentheses) were calculated bu using 1000 bootstrap replicates.
Foraging efficiency Foraging bouts without an attack Species Arthropods Fruits Arthropods Fruits No. of attacks n No. of attacks n Percentage n Percentage n per min per min Tangara arthus 1.0 75 4.9* 62 62% 89 10%** 74 (0.8-1.2) (4.1–5.7) (51–72) (3–17) T. parzudakii 1.5 23 7.1* 13 34% 32 4%** 28 (0.8-2.1) (3.9–10.3) (19–51) (-3–10) T. labradorides 1.4 20 5.3* 23 37% 27 25% 20 (0.8-2.1) (2.4–8.2) (19–55) (6–44) T. rufigula 3.4 13 4.3 15 27% 26 29% 21 (0.5-7.3) (3.4–5.3) (10–44) (9–48)
79 Total 1.4 131 5.4* 113 48% 174 13%** 143 (1.0-1.8) (4.5–6.2) (40–55) (8–19)
* Differences of foraging efficiency between arthropods and fruits highly significant (P < 0.001 in 1000 bootstrap replicates, one-tailed test). ** Differences of percentage of foraging bouts without attack between arthropods and fruits highly significant (P < 0.001 in 1000 bootstrap replicates, one-tailed test).
foraging bouts that included arthropod searching (Table 3.2). In contrast, Tangara tanagers failed to pick a fruit in only 13% of foraging bouts considered as fruit searching.
DISCUSSION
The relative importance of two food types, arthropods and fruits, varies largely depending on which measure and method is used to quantify foraging behavior of omnivorous Tangara tanagers. The percentage of foraging time spent for arthropod foraging is significantly higher than percentage of arthropods in the number of food items (all five pairwise comparisons). This pattern is especially pronounced in T. arthus, which spent 57% of its foraging time searching for arthropods, but arthropods accounted for only 20% of food items. This difference is due to the difference in foraging efficiency: arthropod foraging was 3.9 times less efficient than fruit foraging. As a consequence of this low efficiency, arthropods form a small percentage of the diet of Tangara (25%) despite their spending almost 60% of their foraging time searching for arthropods.
The percentage of arthropods in initial observations was significantly lower than the percentage of foraging time spent for arthropod foraging (four out of five pairwise comparisons).
This seems puzzling because the two measures should be approximately the same if at least one arthropod attack was observed in all foraging bouts that contained arthropod searching.
However, Tangara species failed to attack arthropods in 3.7 times more foraging bouts than when foraging for fruits. Thus, the higher percentage of failure in finding arthropods than fruits appears to have caused the discrepancy between initial observations and foraging duration.
The fruits eaten by these tanagers are produced for facilitating the dispersal of seeds by birds and other animal dispersers; thus, most fruiting trees in the study area bare conspicuous fruits easily found by visual cues, and the ripeness of an individual fruit is easily predicted by
80 both birds and humans from its color, allowing the birds to assess abundance and distribution of
fruits in the area. For example, Acnistus arborescens (Solanaceae) bare conspicuous orange
fruits that attract many bird species, including Tangara species. These same birds, however, do
not stop at the trees when only unripe, green fruits were present.
In contrast to fruits, most arthropods are cryptic to avoid depredation, which presumably
lowers the foraging efficiency of the birds and increases the percentage of unsuccessful foraging
bouts in arthropod foraging. Despite these drawbacks, Tangara tanagers spend considerably
longer time, and therefore more energy, searching for arthropods than for fruits. This suggests
that arthropods supply important nutrition to tanagers that is unavailable in fruits. Tangara species mostly eat small fruits of family Melastomataceae, Moraceae, and Ulmaceae (Isler and
Isler 1999; Snow and Snow 1971), which have high water content and are rich in carbohydrates but poor in protein and lipids (e.g., the fruits of Melastomataceae contain 66% water, 29% carbohydrate, 1% protein, and 2% lipid: Moermond and Denslow 1985). Adult birds need a diet of 4 to 8 percent protein for maintenance (see citations in Berthold 1976), and many omnivorous birds are unable to maintain body weight with a diet consisting solely of fruit (Berthold 1976;
Walsberg 1975); thus, most omnivorous birds, including Tangara, supplement their fruit diet with protein-rich foods, such as seeds, insects, and vertebrates (Moermond and Denslow 1985).
Protein becomes particularly important for egg production, nestling diets, and feather production, and insect consumption increases dramatically during breeding season (Poulin et al. 1992).
Although numerous studies have used the frequency of initial observations to quantify bird foraging behavior (see Morrison et al. 1990b), my study shows that this method may be misleading when the study involves very different food types, such as highly rewarding but scarce arthropods as opposed to less rewarding but abundant fruits. The frequency of initial
81 observations tends to underestimate the importance of such valuable food types in time budgets
and tends to overestimate the same food type in the number of food items in birds’ diets. Thus, it
is advisable to employ multiple data-collection methods and analyses when studying species with
diverse or poorly known behavioral repertories. Furthermore, many studies, including this one, have analyzed birds’ foraging ecology based on a variety of prey items. Different prey or food
types, however, offer neither the same calories nor nutritional composition; one large
lepidopteran larva may be worth several small coleopterans, or one spider may offer more
protein than several watery melastome fruits. Therefore, one should quantify birds’ foraging
choices based on calories and composition of food items, as well as time and energy spent
obtaining and digesting them.
82 CHAPTER 4 SEASONAL CHANGES IN FORAGING ECOLOGY OF TANGARA TANAGERS IN COSTA RICA: FOOD-TYPE DEPENDENT RESOURCE PARTITIONING AND TEMPORAL VARIATION
SUMMARY
The distribution and abundance of food determine the resource-use patterns of birds and consequently affect the structure of bird communities. Although tropical forests were once thought to be relatively constant throughout the year, most of the Tropics are now known to exhibit seasonal fluctuation seasonally according to various environmental factors, particularly rainfall. I compared the resource use patterns of frugivorous-insectivorous Tangara tanagers at
Reserva Biológica El Copal, Costa Rica (9°47’N, 83°45’W, 970m), between breeding and non- breeding seasons, which roughly corresponded to wet and dry seasons. Five sympatric Tangara species showed high overlap in fruit foraging in each season; however, their fruit choice dramatically shifted between the seasons. In contrast to fruit foraging, arthropod foraging was highly species-specific: each Tangara species exploited a different substrate by using a unique combination of attack maneuvers, and individual species showed little seasonal variation.
Although the availability of arthropod taxa in the same substrate probably changes between seasons, stereotyped searching behavior and preference for a particular substrate produced low interspecific overlap in arthropod foraging modes throughout the year. The significant increase in arthropod foraging during the breeding season and the observation that these frugivorous- insectivorous birds breed at the peak of arthropod abundance, when fruit is less available, suggest that the breeding of the frugivorous-insectivorous birds is mostly determined by arthropod abundance.
83 INTRODUCTION
Patterns of resource use among sympatric species are one of the central topics in the study of biological communities. The distribution and abundance of food resources affect the resource-use patterns of birds (Lack 1966; Martin 1987) and consequently influence structure of bird communities (e.g., Grant 1986; Remsen 1990). In most communities, the availability of food resources changes temporally, and species respond by migrating to a resource-rich place or shifting their local resource use. In either case, community structure and biotic interactions among species in a community will shift according to temporal variation in resource distribution
(Almeida and Granadeiro 2000; Wagner 1981).
Although tropical forests were once thought to provide relatively constant environmental conditions throughout the year, most of the Tropics is now known to fluctuate seasonally, particularly rainfall, and both plants and animals show marked seasonal patterns in life-history traits (e.g., Leigh et al. 1983; McDade et al. 1994; Nadkarni and Wheelwright 2000). Birds in the tropics adjust important biological activities, such as breeding and molting, according to seasonal fluctuations of local conditions (Hilty 1980; Levey 1988; Loiselle and Blake 1991;
Poulin et al. 1992), and many species also migrate locally or elevationally (Levey and Stiles
1992; Loiselle and Blake 1990; Stiles 1983; Wyles et al. 1983).
Elevational and local migrations are particularly pronounced among nectarivorous and frugivorous birds in the tropics, and numerous studies have attempted to correlate fruit abundance and migratory, frugivorous bird abundance (e.g., Loiselle and Blake 1991; Ortiz-
Pulido 2000; Ortiz-Pulido and Rico-Gray 2000). A few studies also documented seasonal change in fruit use among resident fruit-eating birds (Loiselle and Blake 1990). Many frugivorous birds also eat seeds, arthropods, and vertebrates to provide supplementary protein
84 and lipids (Moermond and Denslow 1985; Remsen et al. 1993). However, most studies on seasonal variation in resource use by fruit-eating birds have not quantified non-fruit food items in the diet, and the potential importance of non-fruit food items to the biology of these birds and their community organization is poorly known (but see Poulin et al. 1992). Frugivorous- insectivorous birds show dichotomous resource use patterns between two food types, and community organization can be interpreted differently depending upon which food item is considered (Rodrigues 1995; Snow and Snow 1971, Chapter 2). The objective of this study is to document seasonal variation in the resource-use patterns of frugivorous-insectivorous birds between breeding and non-breeding seasons and to understand possible factors influencing the organizations of their community structure.
METHODS
The study was conducted at Reserva Biológica El Copal, prov. Cartago, Costa Rica
(9°47’N, 83°45’W, 970m), and its vicinity from December 2000 to June 2001. The observations of foraging ecology and breeding activity were conducted along a 3-km trail inside El Copal and a 4-km road from the entrance of El Copal toward Cartago. Both observation trails were between 960 and 1,200 m in elevation and were marked with numbered color tapes at every 50 m. A more detailed description of the study site appears in Chapter 2.
There is no weather station at El Copal, but precipitation data are available from the weather station at Taus (9°47’N, 83°43’W, 700m), 5 km from El Copal, and temperature data from Turrialba (09°53'N-83°38'W, 602m), 15 km from El Copal. The temperature at El Copal is probably lower than Turrialba because of its higher elevation, although the general seasonality is similar [Instituto Costarricense de Electricidad (ICE), pers. comm.] The annual precipitation in
85 Taus is 4,699 ± 418 mm (for 10 years 1991-2000, ICE unpublished data), and the annual mean temperature in Turrialba is 22.6 °C (for 36 years 1961-1996). Although annual variation in
precipitation is substantial, the wet season usually starts in May and lasts until December (ICE
unpublished data). In contrast, little seasonal variation is found in temperature (Fig. 4.1).
BREEDING SEASON
During my stay at El Copal, two field assistants, M. Isabel Gómez and Ernesto Carman,
and I recorded the breeding activities of all bird species in the area. For each breeding
observation, we noted species, location, date, time, and a type of activity, such as singing,
courtship feeding, copulation, carrying nest material, carrying food, nests, or moving with
fledglings. To avoid counting the same breeding pair twice, we eliminated the observations of
the same species in close proximity during a short time period unless we knew two nests or two
breeding pairs of the same species were located close together. The first nesting behavior of
Tangara was observed on 19 March 2001. The number of nesting Tangara pairs with nesting
activities peaked in May. The first family group with fledglings was observed in May and the
number of such groups increased in June (Fig. 4.2). These observations are consistent with the
breeding data reported for the genus by Stiles and Skutch (1989). Thus, we categorized the
foraging data collected between 19 March and 30 June as breeding season data, and those
collected between 1 December and 19 February as non-breeding season data.
FORAGING ECOLOGY
Foraging ecology was quantified between 6:00 and 11:00, and 13:00 and 18:00 when bird
activity was high. Two observers usually walked on different trails, and we rotated the trails
between morning and afternoon. When we encountered birds, often in mixed-species flocks, we
followed one individual until it foraged, then moved to another individual. Although we
86 900 25
800
700 20
600 15 500
400 10 300 Temperature (ºC) Precipitation (mm) Precipitation
200 5 100
0 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month Figure 4.1. Monthly change of temperature and precipitation in the study area. A solid line indicates mean monthly temperature at Turrialba (09°53'N-83°38'W, 602 m), and solid bars indicates mean and SD of precipitation at Taus (9°47'N, 83°43'W, 800 m).
87 25
nesting 20 with fledgling
15
10 Number of pairs Number
5
0 Dec Jan Feb Mar Apr May Jun 2000 2001 Month
Figure 4.2. Breeding seasons of five Tangara species at El Copal, Costa Rica in 2001.
88 recorded only the first foraging observation per individual to avoid serial correlation problems during data analyses (Hejl et al. 1990; Martin and Bateson 1993), the observations taken from individuals of the same mixed-species flock are probably not strictly independent. Thus, we rotated the observations of the species in the same mixed-species flock and limited observations to one per species per tree.
For each foraging observation, we recorded on microcassettes the following variables: food item, attack maneuver, substrate type in arthropod foraging or fruit species in fruit foraging, substrate size, perch diameter, perch angle, foliage density, height above ground, distance to canopy, horizontal position, and habitat. Vertical position was calculated as height above ground divided / (height above ground + distance to canopy). Most fruiting plants were sampled and identified at Museo Nacional de Costa Rica. Fruit species were later grouped to generic level in the analyses. The detailed description of categories used for each variable appears in Chapter 2.
DATA ANALYSIS
To analyze all foraging variables simultaneously using multivariate techniques, I grouped continuous foraging variables (perch diameter, height above ground, and vertical position) into four to six categories, as done in Chapter 2. Then, each foraging category was expressed as a proportion of total foraging observations by that species in each season. Correspondence analysis (CA) was used to reduce the dimensionality of the data and to identify which variables explained the most variation among species and between seasons. I mapped the first three CA dimensions to 3D scatter plots to visualize species-season relationships. I also used unweighted pair-group method algorithm (UPGMA) clustering to help visualize groupings in multidimensional spaces.
89 Foraging overlap was calculated by using the Morisita-Horn Index (Krebs 1999, p. 471).
I calculated three indices: interspecific foraging overlap during the non-breeding season (10 species pairs from 5 Tangara species), interspecific foraging overlap during the breeding season
(10 species pairs), and intraspecific foraging overlap, which was calculated by pairing the same species between the non-breeding and breeding seasons and indicated seasonal change in foraging ecology (larger the index, smaller the seasonal change). I conducted these analyses for fruit foraging and arthropod foraging separately to examine whether a similar seasonal pattern was observed between these two main food types. I also used a G-test of independence with
William’s correction to test proportional differences in foraging ecology between species and between seasons for the same species (Sokal and Rohlf 1995). All statistical analyses were carried out by using SAS 8.0 (SAS Institute 2000), except the G-tests, for which I used Excel spreadsheets.
RESULTS
During six months, we collected 2,129 foraging observations from five Tangara species.
Fruits and arthropods accounted for 73% and 26%, respectively, of all foraging observations
(Table 4.1). The proportion of arthropod foraging increased significantly from the non-breeding to the breeding season in four of five Tangara species studied (G-test of independence using 2X2 table; Gadj and P-values in Table 4.1). Tangara florida, T. guttata, and T. icterocephala were found mostly in forest canopy or edges in both the non-breeding and breeding seasons (Table
4.2). Tangara gyrola was found mainly in forested areas although it ventured to non-forested areas more often than did the former three species, and was found significantly more often in non-forested areas during the breeding season than during the non-breeding season (Gadj = 18.5,
P < 0.0001). Tangara larvata fed more frequently in non-forested areas, such as semi-open,
90
Table 4.1. Food preference of five Tangara species at Reserva Biológica El Copal, Costa Rica, during non-breeding and breeding seasons.
Species Percent in each food category Fruit Arthropod Müllerian Nectar Flower n body bud Non-breeding Tangara florida 81 19 0 0 0 161 T. guttata 75 24 0 1 <1 331 T. gyrola 82 18 0 0 0 336 T. icterocephala 87 13 0 0 0 240 T. larvata 65 28 4 3 <1 243 Breeding T. florida 54 46 0 0 0 89 T. guttata 67 33 0 0 0 125 T. gyrola 62 38 0 0 0 228 T. icterocephala 68 32 0 0 0 198 T. larvata 72 28 <1 <1 0 178
Arthropod foraging increased during breeding season in four of five Tangara species (Gadj = 19.4, P < 0.0001 in T. florida, Gadj = 3.9, P < 0.05 in T. guttata, Gadj = 28.4, P << 0.0001 in T. gyrola, Gadj = 24.1, P << 0.0001 in T. icterocephala, Gadj = 0.0006, P = 0.98 in T. larvata).
91
Table 4.2. Percent of habitat use of five Tangara species at Reserva Biológica El Copal, Costa Rica during non-breeding and breeding seasons.
Species Forest Semi-open n Non-breeding Tangara florida 95 5 87 T. guttata 85 15 123 T. gyrola 82 18 224 T. icterocephala 90 10 194 T. larvata 53 47 173 Breeding T. florida 90 10 155 T. guttata 80 20 317 T. gyrola 66 34 317 T. icterocephala 87 13 223 T. larvata 47 53 219
Habitat use differed significantly between non- breeding and breeding seasons in T. gyrola (Gadj = 18.5, P < 0.0001), but not in the other Tangara species (Gadj = 2.1, P = 0.15 in T. florida, Gadj = 1.2, P = 0.28 in T. guttata, Gadj = 1.3, P = 0.24 in T. icterocephala, Gadj = 1.5, P = 0.22 in T. larvata).
92
orchards, and gardens, in both seasons than did the other Tangara species (Table 4.2). Tangara
tanagers are found frequently in mixed-species flocks (42%, n = 1159) or aggregations at large
fruiting trees (22%). Participation in mixed-species flocks
decreased significantly from the non-breeding season to the breeding season (from 48% to 32%;
G-test of independence using 5X2 table; Gadj = 52.8, P << 0.0001).
FRUIT FORAGING
Seventy-four percent of the total variation in fruit foraging was explained by the first three CA dimensions (Fig. 4.3). CA1 was weighted by Fruit genus (Coussapoa and Miconia) and separated Tangara species between the non-breeding season and breeding season (Fig. 4.3, Table
4.3). CA2 was weighted by Fruit genus (Conostegia) and Habitat (semi-open) and separated the non-forest species (T. larvata and T. gyrola; Fig. 4.3, Tables 4.2 and 4.3). CA3 was weighted by
Fruit genus (Ficus) and separated T. gyrola from the rest (Fig. 4.3, Table 4.3). UPGMA recognized three groups for fruit foraging (Fig. 4.4): A) non-breeding forest species (T. florida,
T. guttata, T. icterocephala) that fed heavily on Miconia (over 60% of their foraging observations, Tables 4.2 and 4.3); B) non-breeding non-forest species (T. gyrola and T. larvata) often observed in semi-open areas (over 30% in Table 4.2) and fed on Conostegia and Ficus more frequently (over 30% in Table 4.3); and C) breeding species (all Tangara species during breeding season) that fed heavily on Coussapoa and less on Ficus (Table 4.3).
Interspecific overlap in fruit foraging was high both during the non-breeding (mean ±
SD: 0.87 ± 0.08) and breeding seasons (0.78 ± 0.11); this indicates fruit foraging is similar among Tangara species in each season. Intraspecific foraging overlap, however, was not high between the seasons (0.46 ± 0.15); this reflects a significant shift in fruit foraging from the non- breeding to the breeding season (Fig. 4.5).
93
94
Table 4.3. Percent of fruits eaten by five Tangara species at Reserva Biológica El Copal, Costa Rica, during non-breeding and breeding seasons.
Non-breeding Breeding
hala hala
p p
rola rola uttata uttata lorida g gy f g gy
lorida f T. T. larvata T. T. T. larvata Plant Family Plant Genera T. T. icteroce T. T. icteroce T. Actinidiaceae Saurauia 0 <1 0 0 0 0 0 0 0 0 Araliaceae Oreopanax 0 0 0 0 0 2 2 2 2 <1 Dendropanax 0 <1 0 <1 0 2 0 0 0 0 Schefflera 0 0 0 <1 0 0 0 0 0 0 Caprifoliaceae Viburnum 0 4 0 0 0 0 1 0 <1 2 Chloranthaceae Hedyosmum 1 <1 1 2 1 4 1 4 8 0 Ericaceae Cavendishia 0 <1 0 <1 0 4 2 2 8 0 Satyria 0 0 0 0 0 0 0 1 4 0 Euphorbiaceae Tetrochidium 0 0 0 0 0 6 10 1 2 18 Clusiaceae Clusia 0 0 0 0 0 2 6 4 3 2 Tovomita 0 <1 0 0 0 0 0 0 0 0 Loranthaceae Oryctanthus 0 0 0 0 0 0 0 0 0 2 Viscaceae Phoradendron 1 1 0 1 0 0 4 0 0 <1 Marcgraviaceae Marcgravia 0 <1 0 0 0 4 1 <1 3 0 Melastomataceae Miconia 83 68 41 70 51 17 27 6 15 19 Conostegia 1 8 9 5 31 0 1 1 <1 16 Cecropiaceae Cecropia 4 9 15 <1 9 0 8 12 2 11 Coussapoa 0 0 0 0 0 52 29 41 35 17 Moraceae Ficus 6 2 32 17 7 4 4 23 14 12 Myrtaceae Myrcia 0 <1 0 <1 <1 0 0 0 0 0 Phytolacceaceae Phytolacca 0 0 0 <1 0 0 0 0 0 0 Rosaceae Rubus 0 0 <1 <1 0 0 0 0 0 0 Rubiaceae Sabicea 1 <1 <1 0 0 0 0 <1 2 0 Psychotria 0 <1 0 0 0 0 0 0 0 0 Gonzalagunia 0 0 0 0 <1 0 0 0 0 0 Solanaceae Acnistus 0 <1 0 0 0 0 0 0 0 <1 Others <1 2 1 1 0 2 4 <1 2 <1
n 130 249 279 209 160 48 84 142 134 126
95
gyrola (B) icterocephala (B)
guttata (B)
larvata (B)
florida (B)
larvata (N)
gyrola (N)
guttata (N) icterocephala (N)
florida (N) 2
Figure 4.4. UPGMA of five Tangara species in two seasons based on fruit foraging. (B): breeding season, (N): non-breeding season.
96 1.0
0.8
0.6
0.4 Foraging Overlap
0.2
0.0
Interspecific Interspecific Intraspecific Non-breeding Breeding
Figure 4.5. Box plots of Morisita-Horn indices for fruit foraging at El Copal, Costa Rica. The three boxes indicate: (1) interspecific foraging overlap among five Tangara species during the non-breeding season, (2) interspecific foraging overlap during the breeding season, and (3) intraspecific foraging overlap between the non-breeding and breeding seasons.
97 ARTHROPOD FORAGING
Eighty-one percent of the variation in arthropod foraging was explained by the first three CA dimensions (Fig. 4.6). CA1 was weighted by Attack maneuver (sally-glide), Substrate type
(moss, partially moss-covered branch, and leaf), Horizontal position (middle, foliage, and outer), and Habitat (semi-open) and separated five Tangara species to three branch-foragers (T. florida,
T. icterocephala, and T. gyrola) and two non-branch-foragers (T. guttata and T. larvata; Tables
4.4 – 4.6). CA2 was weighted by Substrate type (leaf and air), Attack maneuver (sally-glide), and Horizontal position (outer) and separated an aerial forager (T. larvata) and a leaf-forager (T. guttata; Table 4.4). CA3 was weighted by Substrate type (moss and bare branch) and separated three branch-foragers according to their dependency on moss (Fig. 4.6, Table 4.4). Tangara florida foraged on moss or partially moss-covered branches for over 90% of foraging observations (Table 4.5). In contrast, T. gyrola searched bare branches more often (61% and
72% in non-breeding and breeding seasons). UPGMA of arthropod foraging showed the same general pattern as CA; within a species, two seasons grouped together, and species grouped according to the similarity of the substrate upon which they specialized (Fig. 4.7).
Mean interspecific foraging overlap in arthropod foraging was low both during the non- breeding (0.33 ± 0.33) and breeding season (0.18 ± 0.26). This reflects the large differences among Tangara species in both seasons. In contrast, intraspecific foraging overlap was high
(0.94 ± 0.05), which indicated that each species maintained a species-specific foraging ecology throughout both seasons (Fig. 4.8).
98
99
Table 4.4. Percent of attack maneuvers used by five Tangara species to capture arthropods at Reserva Biológica El Copal during non-breeding and breeding seasons.
Attack maneuvers Species g* ru ro rd hd hs husd pr pu ju sg n Non-breeding Tangara florida 35 0 3 23 13 6 0 0 19 0 0 31 T. guttata 42 12 10 14 4 9 3 3 3 1 0 78 T. gyrola 17 3 2 36 22 19 0 0 2 0 0 59 T. icterocephala 29 0 0 26 32 3 0 0 10 0 0 31 T. larvata 26 12 10 3 3 0 1 0 0 3 41 68 Breeding T. florida 10 0 0 48 28 8 0 5 3 0 0 40 T. guttata 29 5 22 15 10 2 2 0 2 0 12 41 T. gyrola 22 8 6 40 15 7 0 0 0 0 2 86 T. icterocephala 16 2 0 54 22 5 0 0 0 0 2 63 T. larvata 6 4 6 4 0 10 0 0 0 0 70 50
*g: glean, ru: reach-up, ro: reach-out, rd: reach-down, hd: hang-down, hs: hang-side, husd: hang-up-side-down, pr: probe, pu: pull, ju: jump, sg: sally-glide.
100
Table 4.5. Percent of substrate categories used to catch arthropods by five Tangara species at Reserva Biológica El Copal, Costa Rica.
Substrate type Species m* pmb bb dl l f a n Non-breeding Tangara florida 42 48 10 0 0 0 0 31 T. guttata 0 0 5 0 90 5 0 78 T. gyrola 3 36 61 0 0 0 0 59 T. icterocephala 29 55 13 0 3 0 0 31 T. larvata 0 0 21 0 32 19 28 68 Breeding T. florida 68 27 2 0 2 0 0 41 T. guttata 0 0 0 0 100 0 0 41 T. gyrola 0 22 72 1 3 0 1 86 T. icterocephala 42 55 0 0 3 0 0 64 T. larvata 0 0 6 0 22 28 44 50
* m: moss, pmb: partially moss-covered branch, bb: bare branch, dl: dead leaf, l: leaf, f: flower bud, a: air.
101
Table 4.6. Percent of horizontal position use by five Tangara species when foraging for arthropods at Reserva Biológica El Copal, Costa Rica.
Species inner middle foliage outer n Non-breeding Tangara florida 35 55 10 0 31 T. guttata 1 1 97 0 78 T. gyrola 5 43 52 0 56 T. icterocephala 0 71 29 0 31 T. larvata 1 9 63 27 67 Breeding T. florida 17 80 2 0 41 T. guttata 0 0 100 0 41 T. gyrola 0 34 66 0 86 T. icterocephala 8 55 38 0 64 T. larvata 0 0 56 44 50
102
florida (B)
florida (N) icterocephala (B) icterocephala (N)
gyrola (N)
gyrola (B)
guttata (B)
guttata (N)
larvata (N)
larvata (B) 2
Figure 4.7. UPGMA of five Tangara species in two seasons based on arthropod foraging. (B): breeding season, (N): non-breeding season.
103 1.0
0.8
0.6
0.4 Foraging Overlap
0.2
0.0 Interspecific Interspecific Intraspecific Non-breeding Breeding
Figure 4.8. Box plots of Morisita-Horn indices for arthropod foraging at El Copal, Costa Rica. The three boxes indicate: (1) interspecific foraging overlap among five Tangara species during the non-breeding season, (2) interspecific foraging overlap during the breeding season, and (3) intraspecific foraging overlap between the non-breeding and breeding seasons.
104 DISCUSSION
COMPARISON BETWEEN FRUIT AND ARTHROPOD FORAGING
This study suggests that food type greatly influences not only the resource-use pattern among sympatric species but also the pattern of seasonal changes in resource use within species and at the community level. Fruit foraging was similar among five Tangara species in each season, but
these tanagers shifted their fruit foraging between the non-breeding and breeding seasons. Large
seasonal variation in fruit foraging was probably caused by the seasonal change in fruit species
available in the area. Tangara species at El Copal fed heavily on melastome fruits at the end of
the wet season, when massive fruiting of Miconia therzans was observed. After the fruiting of
M. therzans waned, the Tangara shifted to two species of Coussapoa, which became available at the end of the dry season. Seasonal shift in fruit use was also observed among understory frugivorous birds at the same elevation in Costa Rica (Loiselle and Blake 1990). Overall fruit production fluctuates seasonally (Hilty 1980; Loiselle and Blake 1991), and each plant species tends to have a different fruiting peak; therefore, the composition of available fruits varies through the year, which makes it difficult for birds to specialize on a single fruit species or even a genus. In addition, these morphologically similar congeners use similar foraging maneuvers and show no evidence of microspace segregation in fruit foraging (Naoki submitted). High interspecific overlap in fruit foraging and lack of microspace segregation of the same fruiting tree were also observed in other localities (Daily and Ehrlich 1994; Naoki submitted), which may suggest that these tanagers compete little for fruits (Loiselle and Blake 1991).
In contrast to fruit foraging, arthropod foraging was highly species-specific among five
Tangara species in each season, and each species maintained its stereotyped arthropod foraging from the non-breeding to the breeding season. Each species exploited a different substrate by
105 using a unique combination of attack maneuvers, and this substrate preference did not change
from season to season, as shown by very high intraspecific foraging overlap. Although the
abundance of arthropods in each substrate and the availability of arthropod taxa in the same
substrate probably changes between seasons, stereotyped searching behavior and preference for a
particular substrate probably allow them to maintain low interspecific overlap in arthropod use
throughout the year (see also Snow and Snow 1971).
BREEDING SEASON
Tangara species eat mostly small fruits of the family Melastomataceae, Moraceae,
Ulmaceae, and Cecropiaceae (Isler and Isler 1999; Snow and Snow 1971, this study), which have
high water content and are rich in carbohydrates but poor in proteins and lipids (e.g., the fruits of
Melastomataceae contain 66% water, 29% carbohydrate, 1% protein, and 2% lipid: Moermond
and Denslow 1985). Adult passerine birds need a diet of 4 to 8 percent protein for maintenance
(see citations in Berthold 1976), and many omnivorous birds are unable to maintain body weight
with a diet consisting solely of fruit (Berthold 1976; Walsberg 1975); thus, most fruit-eating birds, including Tangara, supplement their fruit diet with protein-rich foods, such as seeds, arthropods, and vertebrates (Moermond and Denslow 1985). Protein is particularly important for
egg production, nestling diet, and feather production, and so arthropod consumption greatly
increases during the breeding season (Poulin et al. 1992). Tangara tanagers and other
frugivorous-insectivorous birds in Costa Rica breed primarily during the end of the dry season
and the beginning of the wet season (Stiles and Skutch 1989), when arthropod abundance peaks
(Gradwohl and Greenberg 1982; Smythe 1982), but when fruit availability is low (Loiselle and
Blake 1991). Often fruit availability is thought to be a determining factor of the breeding
seasons for many frugivorous and omnivorous birds (e.g., Barrantes and Loiselle 2002;
106 Wheelwright 1983); however, even the most highly frugivorous bird species are known to feed
their nestlings a high proportion of non-fruit food items (Fogden 1972; Wheelwright 1983), and
year-round studies of frugivorous and omnivorous birds failed to correlate breeding season with a fruiting peak (e.g., Fogden 1972; Levey 1988; Loiselle and Blake 1991). In this study, the significant increase in arthropod foraging during the breeding season and the observation that birds breed at the peak of arthropod abundance when fruit availability is not high suggest that the breeding of the frugivorous-insectivorous birds is mostly determined by arthropod abundance
(see also Poulin et al. 1992).
COMMUNITY ORGANIZATION
This study may seem to imply that Tangara tanagers are generalists and opportunists in fruit foraging, and fruits have little effect on community structure. However, when considered as part of a larger community of frugivorous-insectivorous birds, Tangara species form a distinctive fruit-foraging guild: these tanagers feed on small berries of Melastomataceae,
Cecropiaceae, and Moraceae but not on other fruits, such as larger fruits of Lauraceae and mistletoe fruits (Lorantaceae, Viscaceae). These fruits are almost exclusively eaten by other mainly or partially frugivorous birds at our study site (unpublished data). Various factors such as
body size, bill morphology, and digestive system affect fruit choice in birds (Moermond and
Denslow 1985). The differences in fruit foraging, however, seem to require higher taxonomic
analysis, such as among genera or families, to produce patterns in community organization.
Congeners, which have had shorter evolutionary history with respect to one another, and which
have similar morphological and physiological characters, seem to show a similar fruit preference
and tend to be grouped in the same fruit-eating guild (see also Loiselle and Blake 1990).
Moreover, frugivorous-insectivorous birds can obtain large amounts of calories from fruits,
107 which may permit them to engage in lower-efficiency foraging on arthropods when compared to more specialized insectivorous birds (Remsen pers. comm., Naoki pers. obs.). In this case, frugivorous-insectivorous birds may show larger niche overlaps in arthropod foraging than do insectivorous birds; thus, fruit foraging may indirectly affect community organization in arthropod foraging among frugivorous-insectivorous birds.
108 CHAPTER 5 EVOLUTION OF ECOLOGICAL DIFFERENCES IN TANGARA TANAGERS
SUMMARY
Few studies have investigated historical influences on ecological diversification of birds and quantified differences in evolutionary mode among niche axes using a rigorous phylogenetic framework. By combining large sets of ecological data obtained in the field, morphological data from museum specimens, distributional data from literature, and a molecular phylogeny, we tried to elucidate the evolutionary aspects of ecological diversification of 25 Tangara tanagers.
Permutational phylogenetic regression analyses showed significant phylogenetic effects for arthropod foraging, but not for fruit foraging, habitat use, and elevational distribution.
Disparity-through-time plot showed that the relative disparity of arthropod foraging decreased more rapidly than the other niche axes. This was largely due to the initial sorting of microhabitat preference that occurred at the first two nodes. At the first node, Tangara species segregated to one subclade with aerial- or leaf-foragers and the other with mostly branch-foragers. At the second node, the branch-forager subclade further divided to twig-leaf foragers and thicker- branch-foragers.
My study revealed diverse evolutionary patterns unique to each niche axis among 25
Tangara taxa. The relative strength of phylogenetic effects, frequency of homoplasy, mode of evolution, and association with morphology differed substantially among the four niche axes.
Fruit foraging and habitat specialization showed the greatest ecological plasticity in relation to phylogeny. The variation in microhabitat preference in arthropod foraging associated with species-specific attack maneuver was the most conservative and consistent with the phylogeny.
109 INTRODUCTION
The similarity of ecological characters among closely related species is likely to be the result of a combination of adaptation to local ecological conditions and heritage from common ancestors (Brooks and McLennan 1991; Richman and Price 1992). Although most traditional studies of avian community ecology did not distinguish between these two factors, more recent studies have tried to separate them using a rigorous phylogenetic framework (see Schluter 2000 for examples). These recent studies have investigated historical influence on ecological diversification by mapping discrete ecological characters on the phylogeny (Cicero and Johnson
1998; Cicero and Johnson 2002; Johnson and Cicero 2002; Joseph and Moritz 1993; Schluter
1996), or they have investigated ecomorphological association by excluding historical influence
(Richman and Price 1992). None of these studies, however, has tried to elucidate quantitative differences in evolutionary mode among niche axes developed from large sets of field data.
Previous chapters in this thesis showed that sympatric Tangara tanagers partitioned their food resources along three main axes: microhabitat preference in arthropod foraging, fruit foraging, and relative proportion of habitat types used. In addition to these three, I consider elevational distribution as a fourth ecological axis in this chapter. Elevational distribution is negatively correlated with average temperature and is a good indicator of vegetation type in tropical mountain areas (Holdridge 1967). Most birds in the Andes occur within narrow elevational ranges (Graves 1988; Stotz et al. 1996), and even within these ranges their relative abundance may vary strongly with elevation (Fjeldsa and Krabbe 1990; Ridgely and Greenfield
2001; Terborgh and Weske 1975). Most of my study sites were carefully chosen at ecotones, where lowland and montane avifaunas meet, to investigate as many Tangara species as possible.
As a result, some species that are syntopic at my study sites actually possess largely non- overlapping elevational distributions when their entire elevational ranges are considered (e.g., T.
110 labradorides and T. rufigula in Ecuador, T. arthus and T. xanthocephala in Bolivia). The
elevational distributions of Andean birds seem limited by the combination of physical or
biological conditions that vary in parallel with the elevation, environmental discontinuities
(ecotones), and competitive exclusion (Remsen and Cardiff 1990; Terborgh 1971). According to
Terborgh and Weske (1975), the elevational limit of 10 Tangara species in southern Peru is defined either by environmental gradients or by ecotones, though we know little about whether this is true in other Tangara species at other localities.
By combining ecological data obtained in the field, morphological data from museum specimens, distributional data obtained from literature, and a molecular phylogeny, I try to decipher the evolutionary aspects of ecological diversification of Tangara tanagers. I specifically ask the following questions:
(1) Do four niche axes show different amounts of phylogenetic effects and evolutionary mode?
(2) Is there any association between niche axes and morphology of Tangara tanagers, as found in other bird groups?
(3) If so, which niche spaces and morphospaces are associated, and how strong is the association?
METHODS
ORDINATION OF SPECIES BASED ON ECOLOGICAL AND MORPHOLOGICAL CHARACTERS
I used foraging data of 25 Tangara taxa taken at six study sites, including the three sites mentioned in the previous chapters, to quantify three niche axes (Fig. 2.1). Microhabitat preference in arthropod foraging was quantified by using different categories of the following foraging parameters: attack maneuver, substrate type, perch diameter, perch angle, foliage density, height above ground, distance to canopy, and horizontal position. Fruit foraging was
111 quantified by using the same foraging parameters except that “substrate type” is replaced with
“fruit genus.” Habitat use was characterized by the proportion of vegetation height used and proportion of four habitat categories: primary forest, secondary forest, semiopen, and scrub. To minimize the effect of the seasonal variation in fruit foraging and habitat use (Chapter 4), I used data collected only during the dry season. Minimum and maximum elevational distributions of each taxon are taken from Fjeldsa and Krabbe (1990), Hennessey et al. (2003), Ridgely and
Greenfield (2001), and Stiles and Skutch (1989). Minimum and maximum elevations of the center of abundance of each species were taken from Stotz et al. (1996). I ordered Tangara species by microhabitat preference, fruit foraging, and habitat use with correspondence analyses
(CA) and by elevational distribution with principal component analyses (PCA) and a correlation matrix to reduce the dimensionality of each niche axis. The data used in these analyses appear in
Appendices 1-4.
To define morphospace of each taxon, I measured six variables from study skins (length, height, and width of bill at the anterior edge of the nostril, and length of wing, tail, and tarsus) of all 25 taxa except T. meyerdeschauenseei (Baldwin et al. 1931). I also measured 23 skeletal characters from 18 species (Table 5.1, Fig. 5.1). Body mass was obtained from museum skin and skeleton labels. I measured at least five individuals of each taxon. All skin and skeletal measurements, and the cubic root of body mass, were ln-transformed before subsequent analyses.
PATTERNS OF CHARACTER EVOLUTION
Permutational Phylogenetic Regression for Studying Phylogenetic Effects
The evolutionary pathway of each character was studied by using two approaches: permutational phylogenetic regression and relative disparity analyses. First, I applied permutational phylogenetic regressions to assess and quantify phylogenetic effects for each
112
Table 5.1. List of 23 skeletal measurements used in this study.
Skel no. Characters 1 Foramen magnum diameter 2 Premaxilla length 3 Premaxilla length from narial opening 4 Premaxilla depth 5 Premaxilla depth at narial opening 6 Nasal bone width 7 Premaxilla width at gape 8 Skull length 9 Mandible length 10 Minimum mandible length 11 Mandible depth 12 Humerus length 13 Ulna length 14 Ulna width 15 Carpometacarpus length 16 Femur length 17 Tibiotarsus length 18 Tibiotarsus width 19 Tarsometatarsus length 20 Tarsometatarsus width 21 Sternum length 22 Keel length 23 Keel depth
113
114 ecological trait. Permutational phylogenetic regressions compare two dissimilarity matrices: one matrix with trait dissimilarities among species and the other with pairwise genetic differences expressed as P-distances (uncorrected proportional distances) or phylogenetic branch lengths.
The trait dissimilarity matrix was regressed on the genetic distance matrix and tested for significance by using Mantel tests (Legendre et al. 1994; Mantel 1967). In the linear regression, y = βx + ε, the first part of the equation, βx, shows the proportion of the total character variation explained by the phylogeny, whereas the second part, ε, shows the character variation that is unexplained and is possibly caused by ecological factors.
If ecological, behavioral, or morphological traits are strongly influenced by a phylogeny and have evolved at a constant rate among sister lineages, then trait dissimilarity and genetic distance are expected to show a linear relationship (Fig. 5.2a). In contrast, if trait variation is mostly a result of adaptation to local environmental conditions and little or no phylogenetic effect exists, no linear relationship between trait dissimilarity and genetic distance is expected
(Fig. 5.2b). Alternatively, a trait may display some phylogenetic effects, but the rates of trait divergence vary among sister lineages, which would cause greater variation in trait dissimilarity among species pairs with similar genetic distances (Fig. 5.2c). In the last case, the points located at the bottom-right of the figure 5.2c may have resulted from species-pairs of the same lineage diverging slowly (conservatism) or from species-pairs of different lineages evolving a similar trait (convergence).
I used the first two to seven principal axes of each trait, which explained more than 80% of total character variation, to calculate pairwise Euclidean distances and to produce trait dissimilarity matrices. P-distances were calculated by using 1,473 base pairs of mitochondrial cytochrome b (cyt b; 1,043 bp) and NADH dehydrogenase subunit 2 (ND2; 330 bp) genes by
115 (a) 7
6
5
4
3
2
1
0 0.0 0.2 0.4 0.6 0.8
(b) 7
6
5
4
3
2
1
0 Trait dissimilarity 0.0 0.2 0.4 0.6 0.8
(c) 7
6
5
4
3
2
1
0 0.0 0.2 0.4 0.6 0.8 P-distance
Figure 5.2. Examples of the associations between pairwise genetic distances and pairwise trait dissimilarity. (a) constant rate of divergence, (b) explosive divergence, (c) variable rate of divergence or some homoplasy.
116 PAUP* 4.0 (Swofford 1999). Regression lines were estimated with trait dissimilarity matrices as
independent variables and genetic distance matrices as dependent variables by using the program
PERMUTE 3.4 (Casgrain 2001).
Disparity-Through-Time Plot for Studying Mode of Evolution
Although this approach estimates the relative strength of phylogenetic effects on each
trait of interest, it does not describe the evolutionary pathway of each trait. To estimate when,
how often, and how much each trait changed, we need to know either the character states of
ancestral species or the relative amount of character change between ancestral species.
Unfortunately, reconstruction of ancestral character states for continuous variables has
often been criticized (Webster and Purvis 2002, Oakley and Cunningham 2000), especially if
rates of character evolution are high relative to rates of branching speciation (Losos 1999).
Because of the lack of significant phylogenetic effects in three of four niche axes (see Results), estimating ancestral character states would be highly unreliable in this study. To avoid this problem and to estimate the evolutionary trajectory of each niche axis, I used a Disparity-
Through-Time plot. This method calculates the average relative disparity of all of the subclades found at each divergence event (node). The relative disparity of each subclade was calculated as average pairwise Euclidean distances between all the species in the subclade divided by the average pairwise Euclidean distances of the entire clade (Harmon et al. in press).
The molecular phylogeny of 43 species of the genus Tangara and its outgroups was
reconstructed based on 1,473 base pairs of sequence data from the mitochondrial cytochrome b
and ND2 genes (Fig. 5.3). The phylogeny was reconstructed using Bayesian analyses with the
program MrBayes (Huelsenbeck and Ronquist 2001). Branch lengths were then scaled
proportionally to time by nonparametric rate smoothing by using the program r8s (Sanderson
117
118 1997). I pruned outgroup taxa and those Tangara species, for which foraging data were not available, and calculated relative disparity by using this simplified phylogeny (Fig. 5.4).
Correlation Analyses for Studying the Association between Ecology and Morphology
The ecological differences found among Tangara species may be explained by morphological differences. Because species are not independent data points, raw ecological and morphological data cannot be analyzed statistically (Felsenstein 1985). To find an association between ecological and morphological data, I conducted correlation analyses that used independent contrasts among sister taxa. Felsenstein’s method of pairwise independent contrasts incorporates the branching pattern of a phylogeny and the lengths of the component branches
(Felsenstein 1985), and is more robust than ancestral character state reconstruction (Oakley and
Cunningham 2000). Independent contrasts of each trait were calculated using the program
COMPARE 4.4 (Martins 2001).
First, I applied canonical correlation analyses to ln-transformed skin measurements and
CAs of foraging variables by following the method described by Miles and Ricklefs (1984).
Second, because the number of skeletal variables (23) is larger than the number of taxa (17) and because they cannot be ordered by CCA, I reduced the dimensions of ln-transformed skeletal measurements by PCA with a correlation matrix. I used the first three PCs and applied multiple correlation analyses between each of the PC and CA dimensions of foraging variables. Because of the relatively large number of correlation analyses conducted, I used a more conservative significance level (α = 0.01) to test for significant associations.
119 120 RESULTS
MULTIVARIATE ORDINATION OF ECOLOGICAL CHARACTERS
The first four CA dimensions accounted for 81% of the variation in microhabitat
preference (Fig. 5.5). CA1 and 2 were weighted by ‘attack maneuver,’ ‘substrate use,’ and
‘horizontal position.’ They ordered Tangara species from users of branches and inner horizontal
positions (T. chrysotis and T. parzudakii) to aerial foragers (T. cyanicollis and T. ruficervix).
CA3 was weighted by bare branch and moss, and ordered branch-searchers according to the
different degree of dependence on moss: from the species that is heavily dependent on moss (T.
chrysotis) to mostly bare-branch searchers (T. gyrola and T. chilensis). CA4 was weighted by foraging height and separated T. meyerdeschauenseei from all the other species.
The first seven CAs accounted for 82% of the variation in fruit foraging. Each of these
CAs was weighted by different fruit genera; for example, CA1, 2, 3 were weighted by Celtis,
Trema, and Conostegia and Ficus, respectively, and they grouped Tangara species based on study regions: the dry valley in Peru, cloud forest in Ecuador, cloud forest in Costa Rica, and three humid mountain forests in Bolivia (Fig. 5.6). In contrast to arthropod foraging, little species specificity was found in fruit foraging. Fruit choice seemed to be dictated mostly by availability of local fruits, which change both seasonally (Chapter 4) and geographically.
The first three CAs accounted for 85% of the variation in habitat use (Fig. 5.7). CA1 separated T. meyerdeschauenseei, which inhabits low arid-scrub vegetation. CA2 and CA3 were weighted by primary forest - semiopen and secondary forest - semiopen respectively, and ordered Tangara species from forest species (T. vassorii) to semiopen species (T. cyanicollis cyanopygia). In contrast to microhabitat preference (in which 4 subspecies pairs were found close together), three subspecies pairs studied at Mindo, Ecuador, and Serranía Bella Vista,
Bolivia were found far apart in the analysis and indicated the influence of local habitat
121 122 123
124 conditions. The vicinity of Mindo is highly disturbed by agriculture and cattle raising, and little
primary forest was found at this study site, whereas Serranía Bella Vista is located 5 km from the
nearest small settlement and has large areas of little-disturbed primary forest. As a result, T.
cyanicollis cyanopygia, T. arthus goodsoni, and T. nigroviridis berlepschi from Mindo had negative CA2 values; in contrast, T. cyanicollis cyanicollis, T. arthus sophiae, and T. nigroviridis nigroviridis from Bolivia had positive CA2 values.
The first two PCs accounted for 96% of the variation of elevational distribution (Fig.
5.8). PC1 was weighted by all elevational variables but elevation range and ordered Tangara species according to average elevation from highland species which inhabit the upper montane zones (T. vassorii) to lowland species (T. xanthogastra and T. schrankii). PC2 was weighted by elevational range and ordered species from wide elevational distributions (e.g., T. vassorii) to narrow elevational distributions (e.g., T. meyerdeschauenseei).
PHYLOGENETIC EFFECTS IN ECOLOGICAL CHARACTERS
Significant phylogenetic effects were found for microhabitat preference, but not for fruit foraging, habitat use, or elevational distribution (Table 5.2). The scatter plot of microhabitat dissimilarity and p-distance showed the pattern expected by diverse evolutionary rates (Fig.
5.9a). In detailed analyses, however, all significantly small microhabitat dissimilarities found in
13 species pairs were the result of convergence between the T. guttata subclade and T. nigroviridis, T. labradorides, and T. schrankii for leaf or leaf-twig microhabitat preference. In contrast, the scatter plots of fruit foraging and habitat use indicated a rapid increase in trait dissimilarity (Fig. 5.9b and 5.9c); genetically close species and subspecies pairs showed as much trait dissimilarity as more genetically distant species pairs. Of 300 species pairs that showed high habitat and fruit foraging dissimilarity, 24 (8%) involved T. meyerdeschauenseei. This species showed a remarkable difference in habitat use and fruit foraging: it inhabited low scrub
125
126
Table 5.2. Regression analyses of trait dissimilarities on phylogenetic distances among 25 taxa of the genus Tangara.
Number of pairwise Standardized Trait comparisons b t P* R2 Microhabitat preference 300 0.4018 7.5753 0.001 0.162 Fruit foraging 300 -0.0235 -0.4075 0.420 0.001 Habitat use 300 -0.0055 -0.0948 0.494 0.000 Elevational distribution 300 -0.00456 -0.0788 0.497 0.000
* Significance was evaluated based on Mantel tests with 1000 randomizations.
127
(a) 2.5
2.0 x x xx xxx xx x x x x x x x xx x xxxx xx x x x xx x xxx x 1.5 xxxx x x xx xx x x x x x xx xx x xxx xxxxx xxx xxx x xx xxx x x x x x x xx x x xxxxx x x x xx x xx x x xxx x xx x x xxx x x xxx x xx x xx x xxx x x x x xx xx xxx xx xx x x x x x x xxx x x x x x xx x x 1.0 x x x x x xx x x xx x x xx xxx x x x xxx x x xx x xx x xx x x xx x x xxxx x x x x x x?? x x ?? ? C x x x ? ? ? 0.5 x S ?? C ? x S ?? ? x xx C x x x ? C C x x ? CCCC x SSS S C xx S S ? C C S S ? C x C ? C S S ? Microhabitat preference dissimilarity S S C 0.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12
2.5 (b)
m m m m m mm m m m mm mm m 2.0 m mmm m m m m m
1.5
x x x x x x x x x x x x x x x x xx x x x x x x x xx xx x 1.0 x x x x x xx x x x xx x S x x x xx x x x x x x x xx xxxx xxx xx x x xx xx xxxx xxx x x xx x x xx xx xx xx x x x x xx xxx xxxx xx x xx x x x x xx xx x xx S x x xxxx xx x x x x x x x x x xx xx xx x x S x x xxx xx xx x xx x xx xx xx x x xx x x x x Fruit foraging dissimilarity x x xxxx xxx 0.5 x x x x x x x x x xx x xxxxx x x x x x x xx x x S x xx xxx x x xx x x xx xx x x x x x xx xx x x x x x x x x x x x x xx 0.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 P-distance Figure 5.9. Scatter plots of ecological traits vs. genetic divergence (P-distance). (a) Microhabitat preference, (b) Fruit foraging, (c) Habitat use, (d) Elevational distribution. S: slow evolving pair, C: convergence, m: species pairs that include T. meyerdeschauenseei.
128 (c) 4
m m m m mm m m m m mm mmm m m m m 3 m m m
m
x x 2 x x x x x x x xxx S xxxxx x x x x x x x x x x xx x x xx x x x x xx x x x x xx xxxx xx xx Habitat use dissimilarity use Habitat x x x x x xx xx x xxxxxxx xxxx x 1 S x x x xxx x xx x x x xx xx x xx x S xx xxx xxx x x x xxx x xx xx xx xxx x x x x x x xx x x xx x x xx x x x x x x x xx x x xx x xx x x x x xx x x x x xxx x x xx x xxx xxx x x x x x x xx xxxxx xxx xxx x S x x xxx xx xxx x x x x x x x x x xx xx x x x x xx x xx x x x x xx x xx x x x xx x x xx x 0 x x 0.00 0.02 0.04 0.06 0.08 0.10 0.12
(d) 10
8 xx x x x x x x xx x x x x 6 x xx xx x x x x x x x x x x x x xx xx x x x tion dissimilarity x x x x x xx x x x x xx x x x x x x x x x x x x x x xxx x x x x x 4 x x xx xx xx xx x x xxx x x x x x xxxxxx x x x x x x x x xxx xx x x x x x x x x xx x xx x x x x xx xx x x x x x x x x x x x xxxxx x x x x x x x x x xxxx xx x x x x xx x xx xx x xxx x xxx x xx x xx x x xx x x x 2 x x x x x x xxxxxx x x x x x x x x x x x x xxx x x x x x x x x x xx x x x x xx x xxx x x x x x x x xxx x x x x xx x x x xx Elevational distribu x xx x x x x x x x 0 x 0.00 0.02 0.04 0.06 0.08 0.10 0.12 P-distance
Figure 5.9. (cont.)
129 vegetation in a dry Andean valley of southern Peru and foraged fruit genera unique to such a dry
scrub habitat, in contrast with other Tangara tanagers that inhabited forest and foraged mostly
fruits of the genus Miconia. The scatter plot of elevational distribution was similar to the figure
5.2c; however, high level of convergence among distantly related species-pairs made a regression line statistically insignificant (Fig. 5.9d). Although no significant phylogenetic effect was found, there is a strong upper-bound limit of character dissimilarity at small p-distance (Fig.
5.9d).
MODE OF EVOLUTION
The relative disparity of microhabitat decreased more rapidly than the other niche axes
(Fig. 5.10). This was largely due to the initial sorting of microhabitat preference that occurred at the first two nodes. At the first node, Tangara species segregated to one subclade with aerial- or leaf-foragers and the other with mostly branch-foragers (Fig. 5.11). At the second node, the branch-forager subclade further divided to twig-leaf foragers and thicker-branch-foragers.
Further along, Tangara schrankii and probably T. labradorides emerged from the branch- foraging subclade and converged on leaf-foraging (Fig. 5.11).
In contrast, the relative disparity of fruit foraging hardly decreased except at the second node, where three high elevation taxa separated from the other species. Similarly, the relative disparity of habitat use decreased slowly except at the fifth and sixth nodes, where T. meyerdeschauenseei separated from the rest. Elevational distribution decreased slowly and showed a similar pattern of increase and decrease in relative disparity of microhabitat preference, although it lacked the initial sorting of microhabitat preference.
130 1.2
1.0
0.8
0.6 arity of subclades arity
0.4
0.2 Microhabitat preference Fruit foraging 0.0 Habitat use
Average relative disp Average Elevational distribution
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Proportion of time from taxon origin to present
Figure 5.10. Relative disparity-through-time plots for four niche axes. Time is expressed as a proportion of the total time following the first cladogenetic event inferred for the taxon. Only the first 2/3 of the phylogeny is shown.
131
132 CORRELATION BETWEEN ECOLOGICAL AND MORPHOLOGICAL CHARACTERS
The PCA of morphometrics based on 23 skeletal measurements yielded three principal axes, which together accounted for 91% of the total variation (Fig. 5.12). PC1 reflected overall body size and ordered Tangara from small to large species. PC2 was negatively weighted by premaxilla length, premaxilla length from narial opening, and minimum mandible length, and thus was a measure of relative bill length. PC3 was negatively weighted by tibiotarsus length and tarsometatarsus length and positively weighted by keel depth, and thus reflected relative leg length. No significant multiple correlations were found between these PCs of ln-transformed skeletal measurements and ecological variables (Table 5.3).
By using six skin measurements and body mass, a significant association between morphological and ecological variables was found only in elevational distribution, although the first canonical correlation of all niche axes exceeded 0.80 (Table 5.4). In addition, for elevational distribution alone, both of two tests of overall significance, Wilks’ lambda and Roy’s
Greatest Root, rejected the null hypothesis of no association between the ecological and morphological data sets (P < 0.001). To interpret canonical variables, the correlations between morphological and elevational variables and canonical variables were presented in Table 5.4.
Because of the strong correlation among morphological variables, these correlations were more readily interpretable than the canonical coefficients themselves because they showed how much each variable was related to canonical axes (Manly 1994; Miles and Ricklefs 1984).
For the morphological variables, the first canonical variable indicated overall body size and was primarily related to wing, tail, and tarsus lengths (Table 5.5). The second canonical variable was negatively related to wing and tail length, and positively related to bill width, though the correlation was weak. For the ecological variables, the first canonical variable was
133
134
Table 5.3. The results of multiple correlation analyses between ecological and morphological characters based on PCA of skeletal measurements.
Statistical tests PC Morphology (skel) Multiple R Multiple R2 Rao-F df P Microhabitat 1 0.41 0.17 0.6 4, 12 0.66 preference 2 0.34 0.12 0.4 4, 12 0.81 3 0.75 0.56 3.8 4, 12 0.03 Fruit 1 0.71 0.50 1.3 7, 9 0.35 foraging 2 0.71 0.50 1.3 7, 9 0.35 3 0.72 0.52 1.4 7, 9 0.32 Habitat 1 0.30 0.09 0.4 3, 13 0.73 use 2 0.37 0.14 0.7 3, 13 0.58 3 0.21 0.04 0.2 3, 13 0.90 Elevational 1 0.09 0.7 2, 14 0.53 distribution 2 0.35 3.7 2, 14 0.05 3 0.45 5.7 2, 14 0.02
135
Table 5.4. The results of canonical correlation analyses between ecological and morphological characters based on PCA of skin measurements.
Statistical tests Canonical Canonical Canonical variates correlation R2 F df P Microhabitat X morphology (skin) 1 0.81 0.66 1.2 28, 45 0.294 2 0.66 0.44 0.8 18, 37 0.667 3 0.52 0.27 0.6 10, 28 0.833 4 0.21 0.04 0.2 4, 15 0.950 Fruit X morphology (skin) 1 0.85 0.73 0.8 49, 50 0.725 2 0.77 0.60 0.6 36, 47 0.919 3 0.58 0.34 0.4 25, 42 0.985 4 0.45 0.20 0.4 16, 37 0.986 5 0.34 0.12 0.3 9, 32 0.966 6 0.27 0.07 0.3 4, 28 0.887 7 0.05 0.00 0.0 1, 15 0.862 Habitat X morphology (skin) 1 0.85 0.73 2.0 21, 38 0.030 2 0.69 0.48 1.2 12, 28 0.308 3 0.42 0.18 0.6 5, 15 0.674 Elevation X morphology (skin) 1 0.88 0.77 3.4 14, 28 0.003 2 0.63 0.40 1.6 6, 15 0.204
136
Table 5.5. Correlations between the morphological and ecological variables and the canonical variables.
Canonical variables Morphological Ecological V1 V2 W1 W2 Morphological variables Weight 0.08 0.42 0.07 0.26 Bill length -0.14 0.40 -0.13 0.25 Bill width 0.34 0.36 0.30 0.23 Bill depth -0.08 0.36 -0.07 0.23 Wing length 0.33 -0.16 0.29 -0.10 Tail length 0.37 -0.31 0.32 -0.20 Tarsus length 0.24 0.44 0.21 0.28 Ecological (Elevation) PC1 0.78 0.27 0.89 0.44 PC2 -0.28 0.59 -0.33 0.94
137
strongly associated with PC1 of elevational distribution, which indicated overall elevation. The
second canonical variable was associated with PC2, elevational range (Fig. 5.8).
DISCUSSION
The lack of an overt phylogenetic effects in fruit-foraging and habitat use is not
surprising when the large degree of geographic variation observed among four subspecies pairs
is considered (Fig. 5.6 and 5.7). The slow decrease in the disparity-through-time plot, and the
lack of association between these ecological traits and morphology, also indicate little or no
phylogenetic structuring in these niche axes. Fruit foraging also showed marked seasonal
changes in many frugivorous birds including Tangara tanagers (Chapter 4, Loiselle and Blake
1990; Loiselle and Blake 1991). Although most, if not all, frugivorous birds show some preference for certain types of fruits (Loiselle and Blake 1990; Moermond and Denslow 1983;
Moermond and Denslow 1985), the differences in fruit choice seem to be manifest at a higher taxonomic level, such as genus or family, where larger morphological and physiological differences are expected (e.g., see Fig. 2 in Loiselle and Blake 1990). Little difference seems to exist in fruit-foraging among species and lineages in Tangara.
These tanagers have been thought to be segregated by differences in habitat use (Isler and
Isler 1999; Snow and Snow 1971), and my study included four non-forest taxa (Tangara meyerdeschauenseei, T. larvata, T. cyanicollis cyanicollis, T. cyanicollis cyanopygia) in addition
to forest species; however, this study showed little species- or lineage-specific habitat use.
Tangara cyanicollis and T. nigrocincta, usually considered as species of semiopen habitat, were observed in mixed-species flocks with forest Tangara species in primary forest canopy in northern Bolivia (Naoki unpublished data). Even a true non-forest species (T. meyerdeschauenseei) was observed in a disturbed forest with taller trees syntopically with
Tangara xanthocephala, which is usually considered a typical forest Tangara (Naoki 2003). My
138 habitat categories focused on variations in local vegetation types such as primary forest, secondary forest, semiopen, and scrub. Each Tangara species tends to use all or most of these habitat types, but proportional use of each habitat varies among species.
Elevational distribution also did not show significant phylogenetic effects, but did show a strong upper-bound limit of trait divergence between closely related species (Fig. 5.9d); this suggests that elevational distribution did not diverge as rapidly as fruit-foraging and habitat use.
Closely related species have a similar elevational distribution, which implies there was gradual invasion and occupation of distinct vegetation types. The change in elevational distribution was associated with a change in overall body size, following elevational application of Bergmann’s rule. Subspecies pairs actually show almost identical elevational distributions, and sister species usually have elevationally parapatric distributions. The lack of significant phylogenetic effects is due to high homoplasy of many distantly related species, which resulted from repeated invasion of the same elevational zones (Burns and Naoki in prep). Other studies of avian ecological diversification based on phylogeny also found similar repeated shifts in vegetation use or elevational distribution (Cicero and Johnson 1998; Johnson and Cicero 2002; Joseph and Moritz
1993; Richman and Price 1992; Note: Some of these studies used the term “habitat” to define vegetation though their habitats are equivalent to elevational distributions in this study), and concluded that these vegetation shifts are a factor promoting speciation among closely related species.
Microhabitat preference showed significant phylogenetic effects and was the most conserved trait in relation to the phylogeny. A few large divisions of microhabitat preference were observed at the beginning of the diversification, followed by finer segregation, which compartmentalized most Tangara species based on microhabitat exploitation (Isler and Isler
1999). Two remarkable convergences were also observed in this study: T. labradorides and T.
139 schrankii. Tangara schrankii, in particular, showed a drastic shift in microhabitat preference from moss-covered branch to twig and leaf at the tip of the phylogeny (Fig. 5.11). Foraging behavior and morphological characters developed to exploit a particular microhabitat were also most consistent with the phylogeny in other avian groups (Joseph and Moritz 1993; Richman and
Price 1992). Surprisingly, however, this early differentiation in Tangara tanagers was not related to any marked morphological changes, in contrast to other studies of avian adaptive radiations in which tight ecomorphological associations were observed (e.g., Grant 1986;
Moreno and Carrascal 1993; Richman and Price 1992; but see Dendroica in Price et al. 1998).
This difference may be due to non-linear ecomorphological correlation expected among Tangara tanagers. The other studies of avian ecomorphology investigated groups of birds that exploited similar substrate types; thus, the differences in both morphology and behavior among the birds were mainly quantitative and presented a linear correlation (e.g., see Richman and Price 1992).
In contrast, Tangara tanagers used diverse attack maneuvers to exploit very different substrate types; therefore, the difference in behavior may be more qualitative and may not show a linear correlation with morphological differences.
In these omnivorous tanagers, 70% of their diet is fruits (Chapter 2), from which most of their necessary calories can probably be obtained. Arthropods are an important source of proteins especially during nesting seasons (Chapter 4), but may not be indispensable for the survival of adult birds. These tanagers show noticeably lower efficiency in arthropod foraging than more specialized sympatric arthropod foragers such as ovenbirds, flycatchers, and warblers; their capture rate of arthropods is notably low (Rodrigues 1995, Remsen pers. comm., Naoki pers. obs.). Because they can obtain most of their energy from fruits, the selective pressure on their morphology in arthropod foraging may not be so strong. Alternatively, their highly omnivorous diet forces them to maintain intermediate morphology appropriate for both fruit and
140 arthropod foraging. In this scenario, it may be hard to find strong ecomorphological association based on any single food type.
In short, detailed examination of four niche axes based on a molecular phylogeny revealed diverse evolutionary patterns unique to each niche axis among 25 Tangara taxa. The relative strength of phylogenetic effects, frequency of homoplasy, mode of evolution, and association with morphology differed substantially among these four niche axes. Fruit foraging and habitat specialization showed the greatest ecological plasticity in relation to phylogeny. The variation in microhabitat preference associated with foraging behavior was the most conservative and consistent with the phylogeny.
141 CHAPTER 6 SUMMARY AND CONCLUSION
ECOLOGICAL DIVERSIFICATION OF TANGARA TANAGERS
Striking differences in arthropod foraging among sympatric Tangara tanagers were observed at all sites where their foraging ecology has been studied (see Hilty et al. 1986; Isler and Isler 1999; Rodrigues 1995; Snow and Snow 1971). Although no quantitative study of interspecific competition among sympatric Tangara has been conducted, the differential use of substrates appears to be important for avoiding resource competition, especially during the breeding season when the demand for protein-rich arthropods increases (Poulin et al. 1992). In contrast, sympatric Tangara species show little differences in fruit-foraging and habitat use
(Snow and Snow 1971; but see Rodrigues 1995). Fruit-foraging and habitat use largely depend on spatial and temporal availability of food; as a result, sympatric tanagers tend to shift their resource use in similar ways. As expected, neither phylogenetic effects nor evolutionary structuring is found in fruit-foraging and habitat use, which are largely governed by local ecological factors. At a larger spatial scale, elevational distributions (vegetation types) further contribute ecological differences among “sympatric” Tangara tanagers because many syntopic species actually have different centers of abundance, and the population densities of these species are often lower at areas of coexistence (Fig. 6.1).
Similar community structures in microhabitat preferences were observed at three study sites: El Copal (Costa Rica), Mindo (western Andes of Ecuador), and Serranía Bella Vista
(eastern Andes of Bolivia) (Chapter 2). This similarity, however, is not the result of repeated origins of the same microhabitat preference at each region, but rather the result of single or few adaptive events followed by dispersal and allopatric speciation. Furthermore, major divergence in microhabitat preference occurred at an early stage in Tangara diversification, and most
142
?
2500
2000
1500
Elevation (m) Elevation (m) 1000
500
? ? ? ? ? 0
T. heinei T. arthus T. gyrola T. larvata T. palmeri T. vassorii T. rufigula T. ruficervix T. parzudakii T. cyanicollis T. nigroviridis T. labradorides T. icterocephala
Figure 6.1. Elevational distribution of Tangara species in Pichincha province on the western slope of the Ecuadorian Andes. Data from Naoki, and Ridgely and Greenfield (2001).
common uncommon rare
143 subsequent speciation events were not associated with changes in microhabitat preferences, nor
were they associated with interspecific interactions in arthropod foraging. Thus, differential
microhabitat preference, although striking, is at most a force that maintains current species
combinations and community structure, rather than a force driving diversity.
Several observations highlight the inflexibility in microhabitat use of Tangara tanagers.
First, some Tangara species are distributed widely over Central America and South America.
For example, T. gyrola is a member of diverse communities, inhabiting from the depauperate
island community of Trinidad to one of the most species-rich Andean foothill communities.
Despite a three-fold difference in the number of sympatric congeners, no indication of ecological
release or niche expansion is observed in T. gyrola or T. guttata in Trinidad (Fig. 6.2). This
suggests that these tanagers cannot easily change their stereotyped substrate-search-behavior.
Second, I calculated Levins’s measures of niche breadth (Krebs 1999) of 25 Tangara taxa based
on their substrate use and habitat use (Fig. 6.3). Only one Tangara species is found in Andean upper montane forests (T. vassorii) and southern Peruvian dry scrub (T. meyerdeschauenseei).
These Tangara tanagers from single-species communities show no increase in niche breadth
when compared with other Tangara tanagers from species rich communities (Fig. 6.3). Thus, a
comparison of niche breadth among 25 taxa also indicates the same lack of ecological release
and niche expansion as found in depauperate communities. Levins’s measure of niche breadth
does not consider the difference in resource availability among localities; thus, this measure
tends to underestimate the niche breadth of species from resource poor localities. However, all
substrate types are available in all study sites with the exception of moss-covered-branches,
which are scarce in lowland forests and dry scrub. Therefore, the shortcoming of Levins’s
measure does not explain why Tangara tanagers in depauperate communities have not become
144 100
80
60
40
20 Percent of substrate preference
0 T. guttata T. gyrola T. guttata T. gyrola T. gyrola Trinidad Costa Rica Bolivia (3) (5) (9)
Figure 6.2. Percent of substrate types used for arthropod foraging in three sites. Data of Trinidad from Snow and Snow (1971). Parentheses show the number of sympatric Tangara species. branches foliage flower air
145 (a) 5
4
3
2
1 Levins's measuare niceh breadth of 0 1579 (b) 9
8
7
6 of niche breadth of niche
5
4
3 Levins's measure 2 1579 Number of sympatric Tangara species
Figure 6.3. Levins's measure of niche breadth with different number of sympatric Tangara species. Each dot represents one Tangara species. (a) substrate use, (b) habitat use. 1 sp. communities: Chuspipata, Bolivia; Sándia, Peru 5 spp. communities: El Copal, Costa Rica; La Cascada, Bolivia 7 spp. community: Mindo, Ecuador 9 spp. community: Serranía Bella Vista, Bolivia
146 generalists in substrate use. Substrate generalists seem extremely rare among birds even in
species-poor communities, and this is probably due to the biomechanical limitations of
morphological characters that do not allow these birds to exploit diverse substrate types
efficiently (for one of few examples of substrate generalists see Sherry 1985).
In contrast to microhabitat preference, habitat use of Tangara tanagers seems to expand in species-poor communities in the Lesser Antilles (T. cucullata) and southern Mexico (T. cabanisi) (Hilty and Simon 1977), although habitat expansion was not observed in T. vassorii and T. meyerdeschauenseei in my study (Fig. 6.3). This might be due to homogeneous and poor habitat quality in the two study sites (dry scrub for T. meyerdeschauenseei and high-elevation elfin forests for T. vassorii), which may have constrained niche breadth of these species.
Interestingly, most sister species and closely related species with similar microhabitat preferences differ in centers of abundance, when they occur at the same slope (e.g., T. vassorii
and T. nigroviridis, T. arthus and T. icterocephala, T. cyanicollis and T. larvata in Fig. 6.1). In
most cases, these sister species do not show completely parapatric elevational distributions, but
variable degrees of sympatry with plenty of opportunities to interact. This distributional pattern
might suggest that parapatric speciation is common in the genus Tangara in the Andes.
However, none of 108 currently recognized subspecies of Tangara is found elevationally
parapatric, and all subspecies in the Andes are found latitudinally allopatric separated by dry
valleys or found in the eastern and western sides of the Andes separated by the Andean ridge
(Isler and Isler 1999). In addition, hybridization between two Tangara species is known only for
T. cayana and T. preciosa in south-central Brazil (Ingels 1971). Therefore, most, if not all,
Tangara in the Andes speciated allopatrically along a north-south axis, and the elevational
147 parapatric distributions are probably the result of secondary contact after the establishment of reproductive isolation (see Bates and Zink 1994).
We know nothing about how reproductive isolation was established in the Tangara or how they recognize conspecific individuals. Rapid diversification and reproductive isolation may have been achieved by extremely diverse plumage colors or their simple but recognizable songs. As a result, sexual selection may have played a central role in producing numerous ecologically similar species (Price et al. 2000). In this scenario, a fine segregation in arthropod foraging facilitated coexistence of ecologically similar species by avoiding competitive exclusion. The question concerning their diverse plumage patterns is a totally open field to explore in the future.
As shown in a few other studies of ecological diversification, early divergence in microhabitat preference often associated with morphological changes and subsequent habitat shifts seems common among birds (Joseph and Moritz 1993; Richman and Price 1992; Schluter
1996). This contrasts strikingly with other organisms, such as Anolis lizards in Caribbean
Islands, where repeated evolution of the same ecomorphs were observed on each island (Losos et al. 1998). One explanation for the limited number of ecomorphological changes in birds is that in these highly mobile organisms dispersal is so fast that repeated origins of the same niche occur rarely. In contrast, repeated origins of ecomorphs are more common among less mobile organisms in geographically isolated islands, as in the Anolis lizards. Moreover among Anolis lizards, no convergence to the same ecomorph has been found on the same island (Losos 1992).
These hypotheses that dispersal ability and geographic isolation affect the rate of niche diversification can be tested in the future, when more studies of adaptive radiation are conducted,
148 and more rigorous comparisons among diverse groups become possible (e.g., Harmon et al. in
press).
DIFFERENCES IN SPECIES RICHNESS AMONG CENTRAL AMERICA, WESTERN ANDES, AND EASTERN ANDES
Table 6 shows the number of Tangara species and their elevational distributions on three
slopes: Caribbean slope of Central America, the western slope of Ecuadorian Andes, and the
eastern slope of Bolivian Andes. Central America has only half the number of species found on
each Andean slope. This reduction in the number of species in Central America reflects the
nearly complete lack of Tangara species unique to montane forests in that region (Table 6). The
Talamanca mountain range in western Panama and eastern Costa Rica reaches almost 4,000 m,
well above a tree-line, and has similar habitats as Andean humid forests; however, this mountain
range was isolated from the Andes by Darién lowlands in eastern Panama and western Colombia.
Only the ancestor of T. nigroviridis seems to have crossed this gap successfully: this ancestral
tanager evolved to T. fucosa on the hills in Darién gap and to T. dowii in the Talamanca
mountain range (Fig. 5.3). The barrier of lowland humid forests appears to have prevented many
other common Andean montane birds, such as the tanager genera Buthraupis, Anisognathus,
Conirostrum, from reaching Central American mountains.
The numbers of species found on eastern and western slopes of the Andes are similar
(Table 6). The western slope, however, lacks the extensive diversification of lowland species found on the eastern slope. Instead, on the western slope, various species have evolved in hilly tropical forests in the extremely wet Choco region. Extensive lowlands in the Amazon basin were crucial for speciation in situ (T. velia, T. callophrys, T. chilensis, T. mexicana), as well as additional invasions from Andean montane forests (T. schrankii, T. xanthogastra, T. nigrocincta;
149
Table 6. Tangara species found on the three studied slopes: Caribbean slope of Costa Rica, western Andean slope of Ecuador, and eastern Andean slope of Bolivia, with the information on center of abundance, elevational distribution, and subclade.
Center of Caribbean, Western Andes, Eastern Andes, Abundance* Costa Rica Ecuador Bolivia Tangara spp. Elevation** SC+ Tangara spp. Elevation SC Tangara spp. Elevation SC LT T. larvata 0-1500 H T. larvata 0-500 H T. nigrocincta 200-1100 H T. inornata 0-400 B T. mexicana 100-1100 B T. johannae NW 0-500 E T. schrankii 100-1200 E T. chilensis 100-1600 A T. callophrys 100-400 A T. velia 200-600 A T. cayana 200-900 J T. xanthogastra 300-1400 G
150 HT T. gyrola 600-1500 D T. gyrola 0-1500 D T. gyrola 300-1700 D T. florida 350-1100 E T. florida NW 400-1200 E T. palmeri 400-1000 T. lavinia R 250-750 D T. lavinia NW 50-750 D
UT T. guttata 400-1000 G T. rufigula 600-1400 G T. punctata 700-1700 G T. icterocephala 600-1700 E T. icterocephala 500-1350 E T. arthus 900-2200 E T. arthus 700-1600 E T. cyanicollis 0-1400 H T. cyanicollis 200-1600 H T. chrysotis 1100-1500 E T. cyanotis 1400-1800 T. dowii 1200-2750 F
Table 6. (cont.)
Center of Caribbean, Western Andes, Eastern Andes, Abundance* Costa Rica Ecuador Bolivia Tangara spp. Elevation** SC+ Tangara spp. Elevation SC Tangara spp. Elevation SC MM T. nigroviridis 1100-2500 F T. nigroviridis 1000-2600 F T. parzudakii 1500-2400 E T. ruficervix 1400-2400 T. ruficervix R 1100-2900 T. xanthocephala R 1500-2300 E T. xanthocephala 1100-2700 E T. labradorides 1300-2000 T. heinei R 1100-1900 I T. argyrofenges R 1000-2700 I
UM T. vassorii 2000-3000 F T. vassorii 1300-3500 F
151 Total number of species 8 17 19
* Center of Abundance data from Stotz et al. 1996; LT: lower tropical, HT: hill tropical, UT: upper tropical, MM: middle montane, UM: upper montane. ** Costa Rican data from Stiles and Skutch 1989, Ecuadorian data from Ridgely and Greenfield 2001, and Bolivian data from Hennessey et al. 2003. + Subclade Code. See Fig. 5.3. R rare species NW Distribution limited to humid Northwestern Ecuador.
Fig. 6.4), whereas a narrow strip of wet lowlands on the western slope seems not to have offered such an opportunity for in situ speciation.
My analyses of historical biogeography appear to suggest that historical factors have played an important role in creating the differences in species composition and species richness among the three sites; however, they do not discard the possible importance of ecological factors.
For example, the lack of lowland species on the western slope of the Andes and the lack of montane species in Central America might still be partially due to the lack of suitable habitats in these regions.
152
153 REFERENCES
Almeida, J., and J. P. Granadeiro. 2000. Seasonal variation of foraging niches in a guild of passerine birds in a cork-oak woodland. Ardea 88:243-252.
Baldwin, S. P., H. C. Oberholser, and L. G. Worley. 1931. Measurements of birds. Scientific Publications of the Cleveland Museum of Natural History 2:1-151.
Barraclough, T. G., J. E. Hogan, and A. P. Vogler. 1999. Testing whether ecological factors promote cladogenesis in a group of tiger beetles (Coleoptera: Cicindelidae). Proceedings of the Royal Society of London Series B 266:1061-1067.
Barrantes, G., and B. A. Loiselle. 2002. Reproduction, habitat use, and natural history of the Black-and-yellow Silky-flycatcher (Phainoptila melanoxantha), an endemic bird of the western Panama-Costa Rican highlands. Ornitologia Neotropical 13:121-136.
Bates, J. M., and R. M. Zink. 1994. Evolution into the Andes: molecular evidence for species relationships in the genus Leptogon. Auk 111:507-515.
Berthold, P. 1976. The control and significance of animal and vegetable nutrition in omnivorous songbirds. Ardea 64:140-154.
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.
Brooks, D. R., and D. A. McLennan. 1991, Phylogeny, ecology, and behavior. Chicago, The University of Chicago Press.
Buskirk, W. H. 1976. Social systems in a tropical forest avifauna. American Naturalist 110:293- 310.
Casgrain, P. 2001.PERMUTE!, version 3.4, alpha 3.9, version 3.4.Département des Sciences biologiques, Québec, Canada.
Cicero, C., and N. K. Johnson. 1998. Molecular phylogeny and ecological diversification in a clade of New World songbirds (genus Vireo). Molecular Ecology 7:1359-1370.
—. 2002. Phylogeny and character evolution in the Empidonax group of tyrant flycatchers (Aves: Tyrannidae): a test of W. E. Lanyon's hypothesis using mtDNA sequences. Molecular Phylogenetics and Evolution 22:289-302.
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.
154 Danley, P. D., and T. D. Kocher. 2001. Speciation in rapidly diverging systems: lessons from Lake Malawi. Molecular Ecology 10:1075-1086.
Diamond, J. M. 1978. Niche shifts and the rediscovery of interspecific competition. American Scientist 66:322-331.
Felsenstein, J. 1985. Phylogenies and the comparative method. American Naturalist 125:1-15.
Fjeldsa, J., and N. Krabbe. 1990, Birds of the high Andes. Svendborg, Denmark, Apollo Books.
Fogden, M. P. L. 1972. The seasonality and population dynamics of equatorial forest birds in Sarawak. Ibis 114:307-343.
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. 2003.EcoSim: Null models software for ecology, version 7.Acquired Intelligence Inc. & Kesey-Bear, Brulington, VT 05465. http://homepages.together.net/~gentsmin/ecosim.htm.
Grace, J. B., and D. Tilman. 1990, Perspectives on plant competition. New York, Academic Press.
Gradwohl, J., and R. Greenberg. 1982. The breeding season of antwrens on Barro Colorado Island, Pages 345-351 in E. G. Leigh, Jr., A. S. Rand, and D. M. Windsor, eds. The ecology of a tropical forest: seasonal rhythms and long-term changes. Washington, D.C., Smithsonian Institution Press.
Grant, P. R. 1986, Ecology and evolution of Darwin's finches. Princeton, New Jersey, Princeton University Press.
Graves, G. R. 1988. Linearity of geographic range and its possible effect on the population structure of Andean birds. Auk 105:47-52.
Harmon, L. J., J. A. I. Schulte, A. Larson, and J. B. Losos. in press. Tempo and mode of evolutionary radiation in Iguanian lizards: a phylogenetic comparative study. Science.
Harvey, P. H., and M. D. Pagel. 1991, The comparative method in evolutionary biology. New York, Oxford University Press.
Hejl, S. J., J. Verner, and G. W. Bell. 1990. Sequential versus initial observations in studies of avian foraging. Studies in Avian Biology 13:166-173.
Hennessey, A. B., S. K. Herzog, and F. Sagot. 2003, Lista Anotada de las aves de Bolivia. Santa Cruz de la Sierra, Bolivia, Asociación Armonia/BirdLife International.
155 Hilty, S. L. 1977. Food supply in a tropical frugivorous bird community. Ph.D. dissertation thesis, University of Arizona, Tucson, Arizona.
—. 1980. Flowering and fruiting periodicity in a premontane rain forest in pacific Colombia. Biotropica 12:292-306.
Hilty, S. L., W. L. Brown, and G. Tudor. 1986, A guide to the birds of Colombia. Princeton, New Jersey, Princeton University Press.
Hilty, S. L., and D. Simon. 1977. The Azure-rumped Tanager in Mexico with comparative remarks on the Gray-and-gold Tanager. The Auk 94:605-606.
Holdridge, L. R. 1967, Life zone ecology. San José, Costa Rica, Tropical Science Center.
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.
Huelsenbeck, J. P., and F. R. Ronquist. 2001. MRBAYES: Bayesian inference of phylogeny. Biometrics.
Hutchinson, G. E. 1959. Homage to Santa Rosalia or why are there so many kinds of animal? American Naturalist 93:145-159.
Inc., S. 1998, SYSTAT 8.0: statistics. Chicago, Illinois, SPSS Inc.
Ingels, J. 1971. Notes on the breeding of Tangara hybrids. Avicultural Magazine 77:129-131.
Isler, M. L., and P. R. Isler. 1999, The tanagers: natural history, distribution, and identification. Washington, D.C., Smithsonian Institution Press.
Johnson, N. K., and C. Cicero. 2002. The role of ecologic diversification in sibling speciation of Empidonax flycatchers (Tyrannidae): multigene evidence from mtDNA. Molecular Ecology 11:2065-2081.
Joseph, L., and C. Moritz. 1993. Phylogeny and historical aspects of the ecology of eastern Australian scrubwrens Sericornis spp. - evidence from mitochondrial DNA. Molecular Ecology 2:161-170.
Krebs, C. J. 1999, Ecological methodology. Menlo Park, California, Benjamin/Cummings.
Lack, D. 1966, Population studies of birds. Oxford, Clarendon Press.
—. 1971, Ecological isolation in birds. Cambridge, Harvard University Press.
Leck, C. F. 1969. Observations of birds exploiting a Central American fruit tree. Wilson Bulletin 81:264-269.
156 Legendre, P., F. J. Lapointe, and P. Casgrain. 1994. Modeling Brain Evolution from Behavior - a Permutational Regression Approach. Evolution 48:1487-1499.
Leigh, E. G., Jr., A. S. Rand, and D. M. Windsor. 1983. The ecology of a tropical forest: seasonal rhythms and long-term changes. Washington, D.C., Smithsonian Institution Press.
Levey, D. J. 1988. Spatial and temporal variation in Costa Rican fruit and fruit-eating bird abundance. Ecological Monographs 58:251-269.
Levey, D. J., and F. G. Stiles. 1992. Evolutionary precursors of long-distance migration: resource availability and movement patterns in Neotropical landbirds. American Naturalist 140:447-476.
Loiselle, B. A., and J. G. Blake. 1990. Diets of understory fruit-eating birds in Costa Rica: seasonality and resource abundance. Studies in Avian Biology 13:91-103.
—. 1991. Temporal variation in birds and fruits along an elevational gradient in Costa Rica. Ecology 72:180-193.
—. 1993. Spatial-distribution of understory fruit-eating birds and fruiting plants in a Neotropical lowland wet forest. Vegetatio 108:177-189.
Losos, J. B. 1992. The evolution of convergent structure in Caribbean Anolis communities. Systematic Biology 41:403-420.
—. 1996a. Community evolution in Greater Antillean Anolis lizards: phylogenetic patterns and experimental tests, Pages 308-321 in P. H. Harvey, A. J. L. Brown, J. M. Smith, and S. Nee, eds. New uses for new phylogenies. Oxford, Oxford University Press.
—. 1996b. Phylogenetic perspectives on community ecology. Ecology 77:1344-1354.
—. 1999. Uncertainty in the reconstruction of ancestral character states and limitations on the use of phylogenetic comparative methods. Animal Behaviour 58:1319-1324.
Losos, J. B., T. R. Jackman, A. Larson, K. de Queiroz, and L. Rodríguez-Schettino. 1998. Contingency and determinism in replicated adaptive radiations of island lizards. Science 279:2115-2118.
MacArthur, R. H. 1958. Population ecology of some warblers in northeastern coniferous forests. Ecology 39:599-619.
—. 1972, Geographical ecology. Princeton, New Jersey, Princeton University Press.
Magurran, A. E., and R. M. May. 1999. Evolution of biological diversity, Pages 329. New York, Oxford University Press.
157 Manly, B. F. J. 1994, Multivariate statistical methods: a primer. Boca Raton, Florida, Chapman & Hall/CRC.
—. 1997, Randomization, bootstrap and Monte Carlo methods in biology. New York, Chapman & Hall.
Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Research 27:209-220.
Martin, P., and P. Bateson. 1993, Measuring behaviour: an introductory guide. Cambridge, Cambridge University Press.
Martin, T. E. 1987. Food as a limit on breeding birds: a life history perspective. Annual Review of Ecology and Systematics 18:453-487.
Martins, E. P. 1996. Phylogenies and the comparative method in animal behavior. New York, Oxford University Press.
—. 2001.COMPARE, version 4.4. Computer programs for the statistical analysis of comparative data. Distributed by the author via the WWW at http://compare.bio.indiana.edu/, version 4.4.Department of Biology, Indiana University, Bloomington, IN.
McDade, L. A., K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn. 1994. La Selva: ecology and natural history of a Neotropical rain forest. Chicago, University of Chicago Press.
McPeek, M. A., and T. E. Miller. 1996. Evolutionary biology and community ecology. Ecology 77:1319-1320.
Meyer de Schauensee, R. 1970, A guide to the birds of South America. Wynnewood, Pa., Livingston Pub. Co.
Miles, D. B. 1990. A comparison of three multivariate statistical techniques for the analysis of avian foraging data. Studies in Avian Biology 13:295-308.
Miles, D. B., and R. E. Ricklefs. 1984. The correlation between ecology and morphology in deciduous forest passerine birds. Ecology 65:1629-1640.
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.
—. 1985. Neotropical avian frugivores: patterns of behavior, morphology, and nutrition, with consequences for fruit selection, Pages 865-897 in P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgely, and F. G. Buckley, eds. Neotropical Ornithology. Washington, D.C., American Ornithologists' Union.
158 Moreno, E., and L. M. Carrascal. 1993. Leg morphology and feeding postures in four Parus species: an experimental ecomorphological approach. Ecology 74:2037-2044.
Morrison, M. L. 1984. Influence of sample size and sampling design on analysis of avian foraging behavior. Condor 86:146-150.
Morrison, M. L., C. J. Ralph, and J. Verner. 1990a. Introduction. Studies in Avian Biology 13:1- 2.
Morrison, M. L., C. J. Ralph, J. Verner, and J. R. Jehl, Jr. 1990b. Avian foraging: theory, methodology, and applications, Pages 515 in J. R. Jehl, Jr., ed., Studies in Avian Biology. Lawrence, Kansas, Cooper Ornithological Society.
Nadkarni, N. M., and N. T. Wheelwright. 2000. Monteverde: ecology and conservation of a tropical cloud forest, Pages 573. New York, Oxford University Press.
Naoki, K. 2003. Notes on foraging ecology of the little-known green-capped tanager (Tangara meyerdeschauenseei). Ornitologia Neotropical.
—. submitted. Dichotomous patterns of resource partitioning among omnivorous Tangara tanagers (Thraupidae) in western Ecuador.
Naoki, K., and E. Toapanta. 2001. Müllerian body feeding by Andean birds: new mutualistic relationship or evolutionary time lag? Biotropica 33:204-207.
Navarro, G., and M. Maldonado. 2002, Geografía ecológica de Bolivia: vegetación y ambientes acuaticos. Cochabamba, Bolivia, Fundación Simon y Patiño.
Oakley, T. H., and C. W. Cunningham. 2000. Independent contrasts succeed where ancestor reconstruction fails in a known bacteriophage phylogeny. Evolution 54:397-405.
Ortiz-Pulido, R. 2000. Abundance of frugivorous birds and richness of fruit resource: is there a temporal relationship? Caldasia 22:93-107.
Ortiz-Pulido, R., and V. Rico-Gray. 2000. The effect of spatio-temporal variation in understanding the fruit crop size hypothesis. Oikos 91:523-527.
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.
Price, T., I. J. Lovette, E. Bermingham, H. L. Gibbs, and A. D. Richman. 2000. The imprint of history on communities of North American and Asian warblers. American Naturalist 156:354-367.
159 Remsen, J. V., Jr. 1990. Community ecology of Neotropical kingfishers. University of California Publications in Zoology 124:1-116.
Remsen, J. V., Jr., and S. W. Cardiff. 1990. Patterns of elevational and latitudinal distribution, including a "niche switch," in some guans (Cracidae) of the Andes. The Condor 92:970- 981.
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. Studies in Avian Biology 13:144-160.
Richman, A. D. 1996. Ecological diversification and community structure in the Old World leaf warblers (genus Phylloscopus): a phylogenetic perspective. Evolution 50:2461-2470.
Richman, A. D., and T. Price. 1992. Evolution of ecological differences in the Old World leaf warblers. Nature 355:817-821.
Ricklefs, R. E. 1987. Community diversity: relative roles of local and regional processes. Science 235:167-171.
Ricklefs, R. E., and D. Schluter. 1993. Species diversity in ecological communities: historical and geographical perspectives. Chicago, The University of Chicago Press.
Ridgely, R. S., and P. J. Greenfield. 2001, The birds of Ecuador. Ithaca, New York, Cornell University Press.
Ridgely, R. S., and G. Tudor. 1989, The birds of South America. Vol. 1, The oscine passerines, v. 1. Austin, Texas, University of Texas Press.
Robins, J. D., and G. D. Schnell. 1971. Skeletal analysis of the Ammodramus-Ammospiza grassland sparrow complex: a numerical taxonomic study. Auk 88:567-590.
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.
Rosenzweig, M. L. 1995, Species diversity in space and time. Cambridge, Cambridge University Press.
SAS Institute. 2000, SAS/STAT user's guide, version 8. Cary, North Carolina, SAS Institute, Inc.
160 Sanderson, M. J. 1997. A nonparametric approach to estimating divergence times in the absence of rate constancy. Molecular Biology and Evolution 14:1218-1231.
Schluter, D. 1996. Ecological causes of adaptive radiation. American Naturalist 148:S40-S64.
—. 2000, The ecology of adaptive radiation. New York, Oxford University Press.
Schluter, D., and R. E. Ricklefs. 1993. Species diversity: an introduction to the problem, Pages 1-11 in R. E. Ricklefs, and D. Schluter, eds. Species diversity in ecological communities: historical and geographical perspectives. Chicago, University of Chicago Press.
Schoener, T. W. 1973. 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.
—. 1985. Adaptation to a novel environment: food, foraging, and morphology of the Cocos Island Flycatcher, Pages 908-920 in P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgely, and F. G. Buckley, eds. Neotropical Ornithology. Washington, D.C., American Ornithologists' Union.
Sillett, T. S. 1994. Foraging ecology of epiphyte-searching insectivorous birds in Costa Rica. Condor 96:863-877.
Smythe, N. 1982. The seasonal abundance of night-flying insects in a Neotropical forest, Pages 309-318 in E. G. Leigh, Jr., A. S. Rand, and D. M. Windsor, eds. The ecology of a tropical forest: seasonal rhythms and long-term changes. Washington, D.C., Smithsonian Institution Press.
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.
Sokal, R. R., and F. J. Rohlf. 1995, Biometry. New York, W. H. Freeman and Company.
Stiles, F. G. 1983. Birds, Pages 502-618 in D. H. Janzen, ed. Costa Rican natural history. Chicago, University of Chicago Press.
Stiles, F. G., and A. F. Skutch. 1989, A guide to the birds of Costa Rica. Ithaca, Cornell University Press.
Stotz, D. F., J. W. Fitzpatrick, T. A. Parker, III, and D. K. Moskovits. 1996, Neotropical birds: ecology and conservation. Chicago, Illinois, University of Chicago Press.
161 Swofford, D. L. 1999, PAUP*: phylogenetic analysis using parsimony, version 4.0, beta. Sunderland, Massachusetts, Sinauer Associates.
Terborgh, J. 1971. Distribution on environmental gradients: theory and a preliminary interpretation of distributional patterns in the avifauna of the cordillera Vilcabamba, Peru. Ecology 52:23-40.
Terborgh, J., and J. M. Diamond. 1970. Niche overlap in feeding assemblages of New Guinea birds. Wilson Bulletin 82:29-52.
Terborgh, J., and J. S. Weske. 1975. The role of competition in the distribution of Andean birds. Ecology 56:562-576.
Tilman, D. 1982, Resource competition and community structure. Princeton, New Jersey, Princeton University Press.
Tokeshi, M. 1999, Species coexistence: ecological and evolutionary perspectives. Oxford, Blackwell Science Ltd.
Wagner, J. L. 1981. Seasonal change in guild structure: oak woodland insectivorous birds. Ecology 62:973-981.
Walsberg, G. E. 1975. Digestive adaptations of Phainopepla nitens associated with the eating of mistletoe berries. Condor 77:169-174.
Webster, A. J., and A. Purvis. 2002. Testing the accuracy of methods for reconstructing ancestral states of continuous characters. Proceedings of the Royal Society of London Series B - Biological Sciences 269:143-149.
Wheelwright, N. T. 1983. Fruits and the ecology of Resplendent Quetzals. Auk 100:286-301.
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.
Wyles, J. S., J. G. Kunkel, and A. C. Wilson. 1983. Birds, behavior, and anatomical evolution. Proceedings of the National Academy of Science, USA 80:4394-4397.
162 APPENDIX 1: DATA FOR ARTHROPOD FORAGING PERCENT OF FORAGING CATEGORIES USED BY TANGARA SPECIES
Tangara species
Foraging parameter Categories Tangara arthus sophiae T. chilensis T. chrysotis T. cyanicollis T. cyanotis T. gyrola catharinae T. nigroviridis T. punctata T. schrankii T. vassorii T. xanthocephala T. xanthogastra Attack maneuver glean 9.1 30.9 5.6 18.8 38.1 15.2 45.8 49.3 28.3 25.0 10.5 28.7 reach-up 0.0 3.6 0.0 5.2 4.8 3.0 6.3 9.9 13.0 9.4 1.2 14.9 reach-out 0.0 7.3 0.0 5.2 0.0 0.0 6.3 7.0 6.5 12.5 1.2 3.2 reach-down 30.9 27.3 25.0 2.1 26.2 45.5 16.7 7.0 17.4 28.1 27.9 5.3 hang-down 49.1 18.2 13.9 2.1 11.9 27.3 12.5 15.5 15.2 9.4 32.6 6.4 hang-side 1.8 5.5 0.0 1.0 16.7 3.0 8.3 4.2 8.7 9.4 9.3 5.3 hang- upsidedown 0.0 0.0 0.0 0.0 2.4 0.0 0.0 1.4 2.2 3.1 0.0 14.9 probe 3.6 0.0 13.9 0.0 0.0 0.0 0.0 2.8 0.0 0.0 3.5 0.0 pull/bite 5.5 5.5 41.7 0.0 0.0 6.1 2.1 0.0 6.5 0.0 11.6 2.1 sally 0.0 1.8 0.0 65.6 0.0 0.0 2.1 2.8 2.2 3.1 2.3 19.1 Substrate moss 38.2 0.0 30.6 0.0 2.4 0.0 2.1 0.0 0.0 19.0 30.2 0.0 partially-moss- covered branch 56.4 3.6 63.9 0.0 19.0 36.4 2.1 0.0 0.0 42.9 55.8 2.1 bare branch 5.5 87.3 2.8 1.0 66.7 54.5 31.3 2.8 19.6 19.0 9.3 1.1 dead leaf 0.0 0.0 2.8 0.0 0.0 3.0 4.2 1.4 2.2 0.0 0.0 1.1 leaf 0.0 3.6 0.0 18.8 11.9 6.1 56.3 83.1 60.9 19.0 3.5 73.4 flower bud 0.0 1.8 0.0 29.2 0.0 0.0 2.1 12.7 17.4 0.0 0.0 19.1 air 0.0 3.6 0.0 50.0 0.0 0.0 0.0 0.0 0.0 0.0 1.2 3.2 others 0.0 0.0 0.0 1.0 0.0 0.0 2.1 0.0 0.0 0.0 0.0 0.0 Perch diameter < 5 mm 3.6 11.1 0.0 42.4 16.7 6.3 55.8 54.0 46.7 20.0 6.0 45.0 5-10 mm 23.6 55.6 17.1 51.5 52.4 37.5 39.5 46.0 48.9 57.5 39.3 48.3 10-20 mm 30.9 13.0 20.0 3.0 28.6 15.6 4.7 0.0 4.4 15.0 17.9 5.0 20-30 mm 29.1 16.7 37.1 3.0 0.0 31.3 0.0 0.0 0.0 7.5 27.4 1.7 30-60 mm 12.7 3.7 20.0 0.0 2.4 9.4 0.0 0.0 0.0 0.0 8.3 0.0 60 < mm 0.0 0.0 5.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.2 0.0
163 Appendix 1. (cont.)
Tangara species
Foraging parameter Categories Tangara arthus sophiae T. chilensis T. chrysotis T. cyanicollis T. cyanotis T. gyrola catharinae T. nigroviridis T. punctata T. schrankii T. vassorii T. xanthocephala T. xanthogastra Perch angle horizontal 54.5 61.1 54.3 46.9 56.1 54.8 67.4 54.1 55.6 43.2 63.0 68.3 diagonal 43.6 35.2 45.7 46.9 29.3 41.9 25.6 41.0 35.6 54.1 28.4 26.7 vertical 1.8 3.7 0.0 6.3 14.6 3.2 7.0 4.9 8.9 2.7 8.6 5.0 Foliage density 0% 26.0 40.4 53.3 2.4 13.5 35.7 8.5 0.0 6.8 19.4 25.3 2.2 0-5% 30.0 40.4 26.7 9.5 32.4 25.0 17.0 4.4 11.4 25.8 40.0 6.7 5-25% 44.0 19.2 20.0 76.2 51.4 35.7 59.6 82.4 70.5 48.4 33.3 83.1 25-75% 0.0 0.0 0.0 11.9 2.7 3.6 14.9 13.2 11.4 6.5 1.3 7.9 Foraging height < 5 m 33.3 9.1 13.9 23.2 11.9 15.2 19.1 26.5 11.1 33.3 19.3 21.3 5-10 m 53.7 50.9 44.4 57.9 33.3 66.7 46.8 33.8 42.2 66.7 49.4 50.0 10-15 m 11.1 18.2 30.6 11.6 42.9 15.2 27.7 33.8 37.8 0.0 24.1 24.5 15-20 m 1.9 12.7 11.1 5.3 7.1 3.0 6.4 5.9 8.9 0.0 6.0 4.3 20-30 m 0.0 9.1 0.0 2.1 4.8 0.0 0.0 0.0 0.0 0.0 1.2 0.0 Vertical position < 5 14.8 7.3 8.3 1.1 0.0 6.1 6.4 5.9 2.2 3.0 11.0 4.3 5-6 9.3 18.2 8.3 1.1 4.8 15.2 6.4 5.9 11.1 3.0 12.2 6.5 6-7 22.2 20.0 19.4 8.4 9.5 12.1 8.5 5.9 20.0 6.1 19.5 9.7 7-8 27.8 18.2 19.4 25.3 31.0 21.2 25.5 14.7 31.1 30.3 28.0 28.0 8-9 20.4 16.4 38.9 20.0 23.8 39.4 42.6 38.2 11.1 51.5 17.1 29.0 9-10 5.6 20.0 5.6 44.2 31.0 6.1 10.6 29.4 24.4 6.1 12.2 22.6 Horizontal position inner 15.4 3.6 33.3 0.0 0.0 9.1 0.0 0.0 0.0 0.0 16.9 1.1 middle 76.9 49.1 58.3 0.0 56.1 72.7 8.3 0.0 0.0 22.9 60.2 2.1 foliage 7.7 45.5 8.3 51.0 43.9 18.2 91.7 100 100 77.1 21.7 93.6 outer 0.0 1.8 0.0 49.0 0.0 0.0 0.0 0.0 0.0 0.0 1.2 3.2
164 Appendix 1. (cont.)
Tangara species
Foraging parameter Categories T. florida T. guttata T. gyrola bangsi T. icterocephala T. larvata T. arthus goodsoni T. cyanicollis cyanopygia T. labradorides T. nigroviridis berlepschi T. parzudakii T. ruficervix T. rufigula T. meyerdeschauenseei Attack maneuver glean 21.1 37.8 20.0 20.2 17.8 15.4 9.7 50.0 40.9 9.0 18.4 29.1 38.2 reach-up 0.0 9.2 6.2 1.1 8.5 1.0 9.7 16.7 6.8 1.3 4.1 27.3 14.7 reach-out 1.4 14.3 4.1 0.0 8.5 1.9 3.2 7.1 2.3 3.8 0.0 1.8 20.6 reach-down 36.6 14.3 37.9 44.7 3.4 30.8 6.5 2.4 9.1 23.1 2.0 1.8 5.9 hang-down 21.1 5.9 17.9 25.5 1.7 35.6 0.0 0.0 15.9 29.5 0.0 1.8 5.9 hang-side 7.0 6.7 11.7 4.3 4.2 5.8 0.0 2.4 6.8 2.6 2.0 3.6 2.9 hang- upsidedown 0.0 2.5 0.0 0.0 0.8 0.0 0.0 2.4 2.3 1.3 0.0 10.9 0.0 probe 2.8 1.7 0.0 0.0 0.0 2.9 0.0 0.0 4.5 12.8 0.0 3.6 0.0 pull/bite 9.9 2.5 0.7 3.2 0.0 5.8 0.0 14.3 9.1 16.7 0.0 9.1 0.0 sally 0.0 5.0 1.4 1.1 55.1 1.0 71.0 4.8 2.3 0.0 73.5 10.9 11.8 Substrate moss 56.9 0.0 1.4 37.9 0.0 41.3 0.0 0.0 4.5 79.5 0.0 0.0 0.0 partially-moss- covered branch 36.1 0.0 27.6 54.7 0.0 22.1 0.0 0.0 6.8 15.4 0.0 0.0 5.9 bare branch 5.6 3.4 67.6 4.2 14.4 30.8 6.5 2.4 29.5 3.8 2.0 0.0 8.8 dead leaf 0.0 0.0 0.7 0.0 0.0 1.9 0.0 9.5 22.7 1.3 0.0 5.5 5.9 leaf 1.4 93.3 2.1 3.2 28.0 1.9 19.4 83.3 29.5 0.0 28.6 87.3 61.8 flower bud 0.0 3.4 0.0 0.0 22.9 1.0 3.2 2.4 4.5 0.0 10.2 3.6 2.9 air 0.0 0.0 0.7 0.0 34.7 1.0 71.0 2.4 2.3 0.0 59.2 3.6 5.9 others 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 8.8 Perch diameter < 5 mm 0.0 25.0 16.7 2.2 14.7 2.0 25.0 61.1 57.5 1.3 21.4 34.1 34.5 5-10 mm 22.5 68.2 58.3 51.1 73.5 31.4 75.0 38.9 37.5 34.2 64.3 58.5 51.7 10-20 mm 22.5 2.3 15.3 24.4 2.9 32.4 0.0 0.0 5.0 30.3 7.1 4.9 6.9 20-30 mm 40.0 2.3 9.7 11.1 5.9 28.4 0.0 0.0 0.0 22.4 7.1 0.0 3.4 30-60 mm 15.0 2.3 0.0 8.9 2.9 2.9 0.0 0.0 0.0 7.9 0.0 2.4 3.4 60 < mm 0.0 0.0 0.0 2.2 0.0 2.9 0.0 0.0 0.0 3.9 0.0 0.0 0.0
165 Appendix 1. (cont.)
Tangara species
Foraging parameter Categories T. florida T. guttata T. gyrola bangsi T. icterocephala T. larvata T. arthus goodsoni T. cyanicollis cyanopygia T. labradorides T. nigroviridis berlepschi T. parzudakii T. ruficervix T. rufigula T. meyerdeschauenseei Perch angle horizontal 74.3 59.1 58.3 44.4 52.9 74.8 66.7 63.9 58.5 73.7 72.0 46.5 55.2 diagonal 20.0 31.8 33.3 51.1 38.2 19.4 28.6 27.8 26.8 23.7 24.0 41.9 44.8 vertical 5.7 9.1 8.3 4.4 8.8 5.8 4.8 8.3 14.6 2.6 4.0 11.6 0.0 Foliage density 0% 33.3 0.0 34.8 21.4 4.9 31.6 10.3 0.0 18.6 45.7 2.6 0.0 16.7 0-5% 40.0 0.0 37.7 33.3 14.6 12.2 6.9 4.9 14.0 14.3 5.3 5.7 6.7 5-25% 23.3 73.4 27.5 42.9 68.3 49.0 62.1 61.0 55.8 38.6 78.9 64.2 56.7 25-75% 3.3 26.6 0.0 2.4 12.2 7.1 20.7 34.1 11.6 1.4 13.2 30.2 20.0 Foraging height < 5 m 22.5 12.9 16.0 30.9 20.3 6.8 6.9 0.0 11.4 11.5 4.1 0.0 97.1 5-10 m 53.5 59.5 68.1 55.3 61.9 45.6 44.8 35.7 54.5 37.2 42.9 33.3 2.9 10-15 m 21.1 24.1 13.9 10.6 16.1 39.8 27.6 50.0 25.0 39.7 22.4 53.7 0.0 15-20 m 2.8 2.6 2.1 3.2 0.8 5.8 13.8 9.5 6.8 7.7 14.3 9.3 0.0 20-30 m 0.0 0.9 0.0 0.0 0.8 1.9 6.9 4.8 2.3 3.8 16.3 3.7 0.0 Vertical position < 5 9.9 4.3 9.0 22.3 6.8 20.4 3.4 2.4 13.6 19.2 0.0 5.6 41.2 5-6 16.9 5.2 11.1 5.3 4.2 8.7 0.0 2.4 6.8 17.9 2.1 3.7 0.0 6-7 21.1 10.3 18.1 17.0 7.6 13.6 20.7 9.5 22.7 23.1 18.8 9.3 23.5 7-8 29.6 14.7 25.7 23.4 17.8 29.1 24.1 23.8 11.4 21.8 4.2 18.5 14.7 8-9 14.1 26.7 23.6 25.5 16.9 20.4 17.2 16.7 29.5 15.4 27.1 18.5 5.9 9-10 8.5 38.8 12.5 6.4 46.6 7.8 34.5 45.2 15.9 2.6 47.9 44.4 14.7 Horizontal position inner 25.0 0.8 2.1 5.3 0.9 23.0 0.0 0.0 0.0 26.3 0.0 0.0 0.0 middle 69.4 0.8 37.3 60.0 5.1 64.0 6.5 0.0 6.8 65.8 4.1 0.0 14.7 foliage 5.6 98.3 60.6 34.7 59.8 12.0 22.6 97.6 90.9 7.9 36.7 98.1 79.4 outer 0.0 0.0 0.0 0.0 34.2 1.0 71.0 2.4 2.3 0.0 59.2 1.9 5.9
166 APPENDIX 2: DATA FOR FRUIT FORAGING PERCENT OF FORAGING CATEGORIES USED BY TANGARA SPECIES
Tangara species
Foraging parameter Categories Tangara arthus sophiae T. chilensis T. chrysotis T. cyanicollis T. cyanotis T. gyrola catharinae T. nigroviridis T. punctata T. schrankii T. vassorii T. xanthocephala T. xanthogastra Attack maneuver glean 62.2 71.8 61.4 69.7 55.9 64.2 70.8 64.2 72.6 72.4 59.4 68.5 reach-up 7.4 2.6 3.5 5.7 5.9 4.6 8.6 4.4 0.0 0.0 8.5 8.2 reach-out 8.8 5.1 3.5 6.1 5.9 8.3 8.1 10.1 1.4 20.7 6.1 4.1 reach-down 6.8 7.7 14.0 5.7 5.9 11.9 4.9 4.4 5.5 6.9 9.1 2.7 hang-down 3.4 1.9 8.8 2.7 2.9 4.6 1.1 3.1 0.0 0.0 4.2 0.0 hang-side 5.4 4.5 3.5 1.1 8.8 1.8 4.9 4.4 6.8 0.0 3.6 5.5 hang- upsidedown 1.4 0.6 1.8 1.5 0.0 0.9 0.5 0.0 0.0 0.0 1.2 2.7 probe 4.7 5.8 3.5 6.9 14.7 2.8 1.1 9.4 13.7 0.0 7.9 8.2 sally 0.0 0.0 0.0 0.4 0.0 0.9 0.0 0.0 0.0 0.0 0.0 0.0 Fruit genera Acnistus 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Adenaria 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Annona 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Banara 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Bocconia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Brunellia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Cavendishia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.6 0.0 0.0 0.0 0.0 Cecropia 0.7 5.3 0.0 1.2 0.0 8.5 0.0 0.6 4.2 0.0 0.0 2.9 Celtis 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Cestrum 0.0 27.6 0.0 7.4 0.0 17.0 0.0 0.0 29.2 0.0 0.0 10.0 Clethra 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Clusia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Conostegia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Cordia 0.0 0.7 0.0 0.4 0.0 0.0 0.0 0.0 1.4 0.0 0.0 0.0 Coussapoa 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Dendropanax 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
167 Appendix 2. (cont.)
Tangara species
Foraging parameter Categories Tangara arthus sophiae T. chilensis T. chrysotis T. cyanicollis T. cyanotis T. gyrola catharinae T. nigroviridis T. punctata T. schrankii T. vassorii T. xanthocephala T. xanthogastra Fruit genera Eugenia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Ficus 0.7 0.7 0.0 0.8 0.0 1.9 0.5 0.6 0.0 0.0 1.2 0.0 Freziera 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Gonzalagunia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Guettarda 7.5 1.3 33.3 7.4 25.7 20.8 1.6 9.4 0.0 0.0 8.5 4.3 Hedyosmum 6.1 0.0 0.0 7.0 0.0 0.0 0.0 0.6 0.0 0.0 5.5 2.9 Hyeronima 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Isertia 3.4 2.0 1.8 1.9 5.7 0.0 0.5 1.3 0.0 0.0 3.0 0.0 Lasiacis 0.0 3.9 0.0 3.5 0.0 0.9 0.0 0.0 11.1 0.0 0.0 1.4 Leandra 0.0 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Marcgravia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Miconia 76.9 52.6 61.4 62.6 68.6 47.2 97.3 85.6 48.6 100 75.8 74.3 Mircinia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Myrcia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Myrica 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.3 0.0 0.0 5.5 0.0 Myrsine 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Oreopanax 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Orycanthus 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Palicourea 0.0 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Phoradendron 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Phytolacca 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Psychotria 0.0 2.6 0.0 1.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.9 Rubus 0.0 1.3 0.0 1.9 0.0 0.0 0.0 0.0 2.8 0.0 0.0 1.4 Sabicea 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Satyria 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Saurauia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Schefflera 3.4 0.0 0.0 0.8 0.0 0.0 0.0 0.0 0.0 0.0 0.6 0.0 Psidium 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
168 Appendix 2. (cont.)
Tangara species
Foraging parameter Categories Tangara arthus sophiae T. chilensis T. chrysotis T. cyanicollis T. cyanotis T. gyrola catharinae T. nigroviridis T. punctata T. schrankii T. vassorii T. xanthocephala T. xanthogastra Fruit genera Solanum 0.0 0.0 3.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Tetrochidium 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Tibouchia 0.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Tournefortia 0.7 2.0 0.0 1.9 0.0 3.8 0.0 0.0 2.8 0.0 0.0 0.0 Tovomita 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Trema 0.0 0.0 0.0 0.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Turpinia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Viburnum 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Perch diameter < 5 mm 52.6 61.7 65.3 57.1 44.4 48.0 67.5 63.0 56.7 53.6 56.7 76.9 5-10 mm 45.1 34.4 34.7 42.0 55.6 43.9 32.5 35.6 41.8 46.4 42.7 21.5 10-20 mm 1.5 2.3 0.0 0.4 0.0 6.1 0.0 1.5 1.5 0.0 0.0 1.5 20-30 mm 0.8 1.6 0.0 0.4 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 30 < mm 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.7 0.0 Perch angle horizontal 62.8 48.4 68.0 52.3 53.6 62.9 53.5 56.7 51.5 60.7 55.4 50.8 diagonal 32.6 46.9 26.0 45.5 39.3 35.1 37.7 39.6 42.4 39.3 39.2 43.1 vertical 4.7 4.7 6.0 2.3 7.1 2.1 8.8 3.7 6.1 0.0 5.4 6.2 Foliage density 0-5% 0.7 3.3 7.1 3.2 3.2 1.8 0.5 2.6 0.0 0.0 2.5 5.7 5-25% 72.3 83.6 71.4 74.6 71.0 79.8 73.2 77.1 81.9 58.6 77.4 84.3 25-75% 27.0 13.2 21.4 22.2 25.8 18.3 26.2 20.3 18.1 41.4 20.1 10.0 Foraging height < 5 m 24.0 34.7 12.5 36.3 16.7 25.0 21.3 20.9 47.2 18.5 31.5 40.8 5-10 m 59.6 34.7 50.0 46.9 50.0 40.4 62.6 62.8 27.8 81.5 53.7 47.9 10-15 m 13.7 23.3 28.6 12.7 23.3 25.0 15.5 9.5 20.8 0.0 13.0 11.3 15-20 m 2.1 5.3 8.9 3.7 10.0 7.7 0.6 5.4 4.2 0.0 1.9 0.0 20-30 m 0.7 2.0 0.0 0.4 0.0 1.9 0.0 1.4 0.0 0.0 0.0 0.0
169 Appendix 2. (cont.)
Tangara species
Foraging parameter Categories Tangara arthus sophiae T. chilensis T. chrysotis T. cyanicollis T. cyanotis T. gyrola catharinae T. nigroviridis T. punctata T. schrankii T. vassorii T. xanthocephala T. xanthogastra Vertical position < 5 3.4 6.7 0.0 2.9 3.3 5.8 0.6 1.4 8.5 0.0 4.9 14.1 5-6 3.4 4.0 0.0 2.9 6.7 7.7 1.1 3.4 7.0 0.0 3.7 4.2 6-7 7.5 10.7 7.1 12.2 0.0 10.6 5.7 4.1 9.9 3.7 10.5 9.9 7-8 23.3 20.0 7.1 15.5 20.0 14.4 14.4 14.3 22.5 22.2 19.1 19.7 8-9 32.9 30.7 55.4 25.3 33.3 39.4 32.8 36.7 25.4 59.3 31.5 22.5 9-10 29.5 28.0 30.4 41.2 36.7 22.1 45.4 40.1 26.8 14.8 30.2 29.6 Horizontal position foliage 100 100 100 100 100 100 100 100 100 100 100 100 middle 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
170 Appendix 2. (cont.)
Tangara species
Foraging parameter Categories T. florida T. guttata T. gyrola bangsi T. icterocephala T. larvata T. arthus goodsoni T. cyanicollis cyanopygia T. labradorides T. nigroviridis berlepschi T. parzudakii T. ruficervix T. rufigula T. meyerdeschauenseei Attack maneuver glean 47.7 38.2 29.6 36.1 40.7 62.8 73.1 78.1 90.1 53.9 65.1 60.6 43.8 reach-up 8.6 11.3 11.9 9.4 9.3 4.3 1.1 1.6 4.2 2.6 6.2 2.8 31.3 reach-out 18.0 25.8 19.0 27.4 22.9 4.3 2.2 9.4 2.8 4.3 3.4 12.7 6.3 reach-down 14.8 8.4 13.8 14.3 8.4 7.7 6.5 4.7 1.4 10.4 6.2 1.4 9.4 hang-down 0.0 1.8 1.6 1.9 0.9 1.4 0.0 0.0 0.0 0.0 0.7 0.0 0.0 hang-side 7.0 11.3 16.4 3.4 9.8 14.0 3.2 3.1 1.4 15.7 12.3 5.6 0.0 hang- upsidedown 0.8 0.0 0.3 0.0 0.5 0.0 0.0 1.6 0.0 0.9 0.7 0.0 0.0 probe 2.3 3.3 7.4 7.5 7.0 4.8 14.0 1.6 0.0 12.2 5.5 15.5 9.4 sally 0.8 0.0 0.0 0.0 0.5 0.5 0.0 0.0 0.0 0.0 0.0 1.4 0.0 Fruit genera Acnistus 0.0 0.7 0.0 0.0 0.0 3.9 7.6 1.6 0.0 3.6 0.0 0.0 0.0 Adenaria 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3.3 Annona 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6.7 Banara 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.4 0.0 Bocconia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.8 3.3 Brunellia 0.0 0.0 0.0 0.0 0.0 1.0 8.7 1.6 2.9 2.7 2.8 0.0 0.0 Cavendishia 1.6 1.1 1.0 4.2 0.0 0.5 0.0 0.0 0.0 0.9 0.7 0.0 0.0 Cecropia 3.9 8.8 15.8 0.4 8.9 17.2 6.5 1.6 0.0 17.9 16.2 2.8 0.0 Celtis 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 56.7 Cestrum 0.0 0.0 0.0 0.0 0.0 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Clethra 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 16.7 Clusia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Conostegia 1.6 7.4 8.4 4.5 32.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Cordia 0.0 0.0 0.0 0.0 0.0 1.0 14.1 4.7 1.4 3.6 3.5 4.2 0.0 Coussapoa 7.0 1.8 7.4 11.3 5.6 0.0 1.1 0.0 0.0 0.9 0.0 0.0 0.0 Dendropanax 0.8 0.7 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
171 Appendix 2. (cont.)
Tangara species
Foraging parameter Categories T. florida T. guttata T. gyrola bangsi T. icterocephala T. larvata T. arthus goodsoni T. cyanicollis cyanopygia T. labradorides T. nigroviridis berlepschi T. parzudakii T. ruficervix T. rufigula T. meyerdeschauenseei Eugenia 0.0 0.0 0.0 0.0 0.0 1.5 10.9 3.1 0.0 2.7 3.5 8.5 0.0 Ficus 7.8 1.8 37.3 19.2 9.3 2.0 3.3 0.0 0.0 3.6 2.1 2.8 0.0 Freziera 0.0 0.0 0.0 0.0 0.0 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Gonzalagunia 0.0 0.0 0.0 0.0 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Guettarda 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Hedyosmum 1.6 0.4 0.6 3.8 0.5 4.9 0.0 0.0 1.4 3.6 2.1 1.4 0.0 Hyeronima 0.0 0.0 0.0 0.0 0.0 1.5 0.0 0.0 0.0 1.8 0.0 2.8 0.0 Isertia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Lasiacis 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Leandra 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Marcgravia 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Miconia 72.7 66.2 28.0 51.3 38.3 51.0 38.0 65.6 77.1 34.8 33.8 40.8 0.0 Mircinia 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Myrcia 0.0 0.7 0.0 0.4 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Myrica 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Myrsine 0.0 0.0 0.0 0.0 0.0 2.0 2.2 1.6 0.0 0.9 0.7 4.2 10.0 Oreopanax 0.0 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.9 0.0 2.8 0.0 Orycanthus 0.0 0.0 0.0 0.0 0.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Palicourea 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5.6 0.0 Phoradendron 1.6 2.2 0.0 0.4 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Phytolacca 0.0 0.0 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Psychotria 0.0 0.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Rubus 0.0 0.0 0.3 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Sabicea 0.8 0.4 0.3 1.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Satyria 0.0 0.0 0.6 1.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Saurauia 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Schefflera 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Psidium 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3.3
172 Appendix 2. (cont.)
Tangara species
Foraging parameter Categories T. florida T. guttata T. gyrola bangsi T. icterocephala T. larvata T. arthus goodsoni T. cyanicollis cyanopygia T. labradorides T. nigroviridis berlepschi T. parzudakii T. ruficervix T. rufigula T. meyerdeschauenseei Solanum 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3.6 1.4 4.2 0.0 Tetrochidium 0.8 1.1 0.3 0.0 1.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Tibouchia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Tournefortia 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Tovomita 0.0 0.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Trema 0.0 0.0 0.0 0.0 0.0 12.3 7.6 20.3 17.1 17.9 33.1 15.5 0.0 Turpinia 0.0 0.0 0.0 0.0 0.0 0.5 0.0 0.0 0.0 0.9 0.0 0.0 0.0 Viburnum 0.0 4.0 0.0 0.4 0.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Perch diameter < 5 mm 50.9 53.9 28.7 42.4 46.2 44.4 47.1 70.0 62.9 41.0 33.5 50.0 34.8 5-10 mm 47.3 40.9 61.0 55.6 51.3 39.9 46.0 30.0 37.1 46.7 45.1 43.2 60.9 10-20 mm 1.8 5.2 7.4 0.0 2.6 7.6 4.6 0.0 0.0 5.7 9.1 4.1 4.3 20-30 mm 0.0 0.0 2.2 1.0 0.0 6.6 2.3 0.0 0.0 3.8 7.3 2.7 0.0 30 < mm 0.0 0.0 0.7 1.0 0.0 1.5 0.0 0.0 0.0 2.9 4.9 0.0 0.0 Perch angle horizontal 58.2 52.2 44.9 60.6 46.2 70.7 79.1 71.7 77.1 68.2 71.3 66.2 47.8 diagonal 30.9 33.9 33.1 34.3 38.5 13.6 16.3 23.3 21.4 16.8 16.5 25.7 52.2 vertical 10.9 13.9 22.1 5.1 15.4 15.7 4.7 5.0 1.4 15.0 12.2 8.1 0.0 Foliage density 0-5% 0.0 0.0 1.1 0.8 2.6 5.4 5.3 0.0 1.5 4.4 4.7 2.5 6.7 5-25% 54.5 57.0 67.2 58.6 60.0 58.1 56.4 62.5 50.7 65.5 72.1 63.8 46.7 25-75% 45.5 43.0 31.6 40.6 37.4 36.5 38.3 37.5 47.8 30.1 23.3 33.8 46.7 Foraging height < 5 m 25.6 24.8 20.3 25.0 46.0 8.7 13.7 3.1 2.8 8.6 1.2 13.8 100 5-10 m 60.5 63.3 54.0 53.4 41.8 43.5 43.2 34.4 46.5 38.8 39.5 40.0 0.0 10-15 m 11.6 11.9 17.7 17.0 12.2 41.1 32.6 45.3 38.0 42.2 50.6 30.0 0.0 15-20 m 2.3 0.0 7.7 4.5 0.0 6.8 7.4 12.5 11.3 6.9 5.8 13.8 0.0 20-30 m 0.0 0.0 0.3 0.0 0.0 0.0 3.2 4.7 1.4 3.4 2.9 2.5 0.0
173 Appendix 2. (cont.)
Tangara species
Foraging parameter Categories T. florida T. guttata T. gyrola bangsi T. icterocephala T. larvata T. arthus goodsoni T. cyanicollis cyanopygia T. labradorides T. nigroviridis berlepschi T. parzudakii T. ruficervix T. rufigula T. meyerdeschauenseei Vertical position < 5 0.0 0.0 0.0 0.0 11.7 10.1 7.4 3.1 2.8 4.3 3.5 3.8 20.0 5-6 3.1 4.4 7.7 6.1 7.0 1.9 4.2 3.1 2.8 6.9 6.4 0.0 0.0 6-7 3.9 8.1 9.6 8.7 8.9 14.5 8.4 9.4 11.3 9.5 8.1 8.8 20.0 7-8 24.0 18.1 18.0 12.1 18.3 18.8 12.6 26.6 16.9 13.8 16.9 17.5 23.3 8-9 31.0 34.1 28.3 31.4 19.7 25.1 35.8 28.1 33.8 31.9 25.0 25.0 6.7 9-10 25.6 26.3 22.2 22.3 34.3 29.5 31.6 29.7 32.4 33.6 40.1 45.0 30.0 Horizontal position foliage 100 100 100 100 100 98.6 99.0 100 100 99.2 99.4 100 100 middle 0.0 0.0 0.0 0.0 0.0 1.4 1.0 0.0 0.0 0.8 0.6 0.0 0.0
174 APPENDIX 3: DATA FOR HABITAT USE PERCENT OF FORAGING CATEGORIES USED BY TANGARA SPECIES
Tangara species
Foraging parameter Categories Tangara arthus sophiae T. chilensis T. chrysotis T. cyanicollis T. cyanotis T. gyrola catharinae T. nigroviridis T. punctata T. schrankii T. vassorii T. xanthocephala T. xanthogastra primary Habitat forest 42.3 22.2 27.8 19.4 67.5 41.4 56.3 49.3 23.9 94.3 53.7 22.6 secondary forest 51.9 64.8 50.0 46.2 27.5 55.2 29.2 31.3 63.0 2.9 25.6 64.5 semiopen 1.9 9.3 19.4 6.5 2.5 3.4 6.3 6.0 2.2 0.0 4.9 0.0 secondary growth 3.8 3.7 2.8 28.0 2.5 0.0 8.3 13.4 10.9 2.9 15.9 12.9 Vegetation height < 4 m 1.9 5.5 0.0 8.4 2.4 0.0 4.3 4.4 6.7 6.1 3.7 3.2 4 - 6 m 11.1 3.6 11.1 15.8 9.5 9.1 12.8 16.2 4.4 18.2 9.8 21.5 6 - 8 m 18.5 1.8 5.6 32.6 4.8 18.2 14.9 8.8 8.9 48.5 15.9 15.1 8 - 10 m 27.8 7.3 8.3 12.6 14.3 18.2 10.6 14.7 6.7 27.3 17.1 7.5 10 - 12 m 9.3 9.1 16.7 8.4 14.3 24.2 14.9 10.3 15.6 0.0 9.8 15.1 12 - 14 m 11.1 12.7 5.6 3.2 9.5 6.1 4.3 11.8 2.2 0.0 6.1 7.5 14 - 16 m 13.0 16.4 8.3 8.4 23.8 6.1 17.0 16.2 20.0 0.0 11.0 9.7 16 - 18 m 1.9 12.7 22.2 1.1 2.4 6.1 8.5 7.4 8.9 0.0 4.9 6.5 18 - 20 m 1.9 5.5 16.7 4.2 9.5 9.1 6.4 7.4 17.8 0.0 12.2 9.7 20 - 25 m 3.7 20.0 5.6 4.2 4.8 3.0 6.4 2.9 8.9 0.0 9.8 4.3 25 < m 0.0 5.5 0.0 1.1 4.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0
175 Appendix 3. (cont.)
Tangara species
Foraging parameter Categories T. florida T. guttata T. gyrola bangsi T. icterocephala T. larvata T. arthus goodsoni T. cyanicollis cyanopygia T. labradorides T. nigroviridis berlepschi T. parzudakii T. ruficervix T. rufigula T. meyerdeschauenseei primary Habitat forest 60.0 35.9 45.1 54.9 9.6 2.0 6.5 7.3 0.0 2.6 2.1 0.0 0.0 secondary forest 37.1 44.4 34.7 36.3 31.3 74.3 12.9 73.2 80.0 55.3 75.0 59.3 0.0 semiopen 2.9 18.8 19.4 4.4 56.5 20.8 80.6 17.1 10.0 38.2 20.8 40.7 14.7 secondary growth 0.0 0.9 0.7 4.4 2.6 3.0 0.0 2.4 10.0 3.9 2.1 0.0 85.3 Vegetation height < 4 m 1.4 0.9 2.1 1.1 7.6 0.0 0.0 0.0 0.0 2.6 0.0 0.0 85.3 4 - 6 m 9.9 6.0 6.9 13.8 11.9 1.0 0.0 0.0 4.5 1.3 4.2 0.0 8.8 6 - 8 m 15.5 21.6 16.0 17.0 24.6 3.9 17.2 7.1 13.6 6.4 4.2 3.7 5.9 8 - 10 m 16.9 19.8 25.0 21.3 17.8 4.9 17.2 19.0 15.9 10.3 4.2 3.7 0.0 10 - 12 m 15.5 24.1 18.1 17.0 15.3 8.7 3.4 9.5 6.8 7.7 16.7 11.1 0.0 12 - 14 m 4.2 9.5 8.3 2.1 7.6 13.6 10.3 14.3 11.4 0.0 8.3 16.7 0.0 14 - 16 m 15.5 11.2 11.1 9.6 10.2 24.3 13.8 19.0 18.2 23.1 27.1 25.9 0.0 16 - 18 m 7.0 0.9 7.6 6.4 1.7 12.6 6.9 7.1 9.1 7.7 6.3 18.5 0.0 18 - 20 m 11.3 2.6 3.5 7.4 0.8 20.4 10.3 7.1 2.3 20.5 4.2 7.4 0.0 20 - 25 m 1.4 2.6 1.4 3.2 1.7 8.7 17.2 11.9 15.9 17.9 10.4 11.1 0.0 25 < m 1.4 0.9 0.0 1.1 0.8 1.9 3.4 4.8 2.3 2.6 14.6 1.9 0.0
176 APPENDIX 4: DATA FOR ELEVATIONAL DISTRIBUTION (M)
Tangara species
Parameters in elevational distribution before log transformation Tangara arthus sophiae T. chilensis T. chrysotis T. cyanicollis cyanicollis T. cyanotis T. gyrola catharinae T. nigroviridis nigroviridis T. punctata T. schrankii T. vassorii T. xanthocephala T. xanthogastra Minimum elevation 177 of center of abundance 900 0 900 900 900 500 1600 900 0 2600 1600 0 Maximum elevation of center of abundance 1600 500 1600 1600 1600 900 2600 1600 500 3500 2600 500 Minimum elevation 700 100 1100 200 1400 300 1000 700 100 1300 1100 300 Maximum elevation 1600 1600 1500 1600 1800 1700 2600 1700 1200 3500 2700 1400 Mean elevation 1150 850 1300 900 1600 1000 1800 1200 650 2400 1900 850 Range of elevation 900 1500 400 1400 400 1400 1600 1000 1100 2200 1600 1100 Mean elevation of center of abundance 1250 250 1250 1250 1250 700 2100 1250 250 3050 2100 250
Appendix 4. (cont.)
Tangara species
Parameters in elevational distribution before log transformation T. florida T. guttata T. gyrola bangsi T. icterocephala T. larvata T. arthus goodsoni T. cyanicollis cyanopygia T. labradorides T. nigroviridis berlepschi T. parzudakii T. ruficervix T. rufigula T. meyerdeschauenseei Minimum elevation
178 of center of abundance 500 900 500 900 0 900 900 1600 1600 1600 1600 900 1600 Maximum elevation of center of abundance 900 1600 900 1600 500 1600 1600 2600 2600 2600 2600 1600 2600 Minimum elevation 350 400 600 600 0 900 0 1300 1100 1500 1400 600 1750 Maximum elevation 1100 1000 1500 1700 1500 2200 1400 2000 2500 2400 2400 1400 2200 Mean elevation 725 700 1050 1150 750 1550 700 1650 1800 1950 1900 1000 1975 Range of elevation 750 600 900 1100 1500 1300 1400 700 1400 900 1000 800 450 Mean elevation of center of abundance 700 1250 700 1250 250 1250 1250 2100 2100 2100 2100 1250 2100
VITA
Kazuya Naoki was born to Yoshiya and Iyoko Naoki on 17 March 1969 in Hyogo, Japan.
In 1970, his family moved to Kawasaki, where he found his earliest interest in the nature and
dreamt to become an entomologist. He attended elementary school in Kawasaki, Osaka, and
Hyogo. He attended Kwansei Gakuin High School at Kobe. In high school Naoki was mostly
interested in tennis and fishing; however, he never deserted his dream to study nature. When he
was 17 years old, he traveled in Kyushu by a bicycle for three weeks, where he met subtropical
forests first time in his life and madly fell in love with her. He studied chemistry at Kwansei
Gakuin University, Kobe city, Hyogo between 1987 and 1989 until he left for Australia, New
Zealand, and Latin America to travel and observe tropical birds for two years. He left the
university in Japan and went to Costa Rica in 1991 to study tropical biology. He completed a
Bachelor of Science degree in biology at Universidad de Costa Rica, San José, Costa Rica in
1996. Kazuya will receive the degree of Doctor of Philosophy in biological sciences and
experimental statistics at Louisiana State University in August 2003. Following graduation, he will establish himself at Bolivia to continue his study on Andean birds.
179