Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Temperature and vegetation complexity structure mixed-species flocks along a gradient of elevation in the tropical Andes

Temperature and vegetation complexity structure mixed-species flocks along a gradient of... Abstract Mixed-species flocks constitute community modules that can help test mechanisms driving changes to community composition across environmental gradients. Here, we examined elevational patterns of flock diversity (species richness, taxonomic diversity, species, and guild composition) and asked if these patterns were reflections of the full bird community at a given elevation (open-membership hypothesis), or if they were instead structured by environmental variables. We surveyed both the overall avian community and mixed-species flocks across an undisturbed elevational gradient (~1,350–3,550 m) in the Bolivian Andes. We then tested for the role of temperature (a surrogate for abiotic stress), resource diversity (arthropods, fruits), and foraging niche diversity (vegetation vertical complexity) in structuring these patterns. Patterns for the overall and flocking communities were similar, supporting our open-membership hypothesis that Andean flocks represent dynamic, unstructured aggregations. Membership openness and the resulting flock composition, however, also varied with elevation in response to temperature and vegetation complexity. We found a mid-elevation peak in flock species richness, size, and Shannon’s diversity at ~2,300 m. The transition of flocking behavior toward a more open-membership system at this elevation may explain a similar peak in the proportion of insectivores joining flocks. At high elevations, increasing abiotic stress and decreasing fruit diversity led more generalist, gregarious tanagers (Thraupidae) to join flocks, resulting in larger yet more even flocks alongside a loss of vegetation structure. At lower elevations, flock species richness increased with greater vegetation complexity, but a greater diversity of foraging niches resulted in flocks that were more segregated into separate canopy and understory sub-types. This segregation likely results from increased costs of interspecific competition and activity matching (i.e., constraints on movement and foraging rate) for insectivores. Mid-elevation flocks (~2,300 m) seemed, therefore, to benefit from both the open-membership composition of high-elevation flocks and the high vegetation complexity of mid- and low-elevation forests. RESUMEN Las bandadas mixtas constituyen módulos comunitarios (i.e., subcomunidades de especies con fuertes asociaciones interespecíficas) que pueden ayudarnos a investigar los mecanismos que determinan cambios en la composición de comunidades a lo largo de gradientes ambientales. En este trabajo, examinamos patrones altitudinales de diversidad de bandadas mixtas de aves (riqueza de especies, diversidad taxonómica y funcional, composición de especies y gremios ecológicos), e investigamos si estos patrones son reflejo de la comunidad de aves en cada elevación (hipótesis de membresía abierta), o si las bandadas son estructuradas por variables ambientales. Evaluamos tanto la comunidad de aves completa como la formación de bandadas mixtas a lo largo de un gradiente altitudinal no disturbado (~1,350–3,550 m) en los Andes de Bolivia. Luego, evaluamos el rol de la temperatura (como proxy de condiciones abióticas) y de la disponibilidad de recursos (artrópodos, frutos y complejidad vertical de la vegetación) como mecanismos que determinan los cambios en diversidad de bandadas. Los patrones altitudinales de la comunidad entera y de la comunidad de aves de bandadas fueron similares, apoyando nuestra hipótesis de “membresía abierta” de que las bandadas de los Andes representan agregaciones dinámicas y no estructuradas. La apertura de esta membresía en bandadas mixtas y la composición de especies resultante, sin embargo, también varió con la altura en respuesta a la temperatura y a la complejidad de la vegetación. Encontramos un pico de riqueza de especies, tamaño, diversidad de Shannon y diversidad funcional por bandada cerca de los 2,300 m. Sugerimos que la transición de comportamiento hacia una membresía abierta en la formación de bandadas a partir de esta altitud permite que una mayor proporción de la comunidad de insectívoros se una a las bandadas. En elevaciones mayores, las bajas temperaturas y la perdida de estructura de vegetación permiten la unión de especies más generalistas y gregarias, especialmente tangaras (Thraupidae), resultando en bandadas con mayor número de individuos, pero más equitativas (i.e., especies con abundancias más homogéneas). A elevaciones menores, con mayor complejidad estructural, una mayor cantidad de nichos de forrajeo resulta en bandadas que son más estructuradas, segregándose en subtipos de dosel y sotobosque. Esta segregación potencialmente resulta de un incremento en los costos de competencia interespecífica y costos de actividad (i.e., limitantes en el movimiento y la tasa de forrajeo) de insectívoros. Las bandadas de elevaciones medias (~2,300), por lo tanto, parecen beneficiarse tanto de la membresía abierta de mandadas de altura, como de la alta complejidad estructural de bosques en alturas medias y bajas. Lay Summary • We investigated changes to the diversity and membership of flocks of birds across elevations in the Bolivian Andes, and the mechanisms driving these changes. • Changes to flock membership largely reflected elevational changes to the whole bird community, suggesting Andean flocks have open membership, unlike the flocks of lowland Amazonia. • Elevational changes to flock richness, size, and membership were mostly associated with effects of temperature (abiotic stress) and the diversity of vegetation layers within the forest. • Temperature and fruit diversity were associated with the number of individuals of each species, with larger flocks at high elevations, composed of generalist, gregarious tanagers. • Increasing numbers of vegetation strata were associated with higher insectivore richness and abundance in flocks, explaining flock characteristics at middle and low elevations. INTRODUCTION Mixed-species flocks are moving foraging aggregations of two or more species (Morse 1970) that constitute “ecological modules” of closely interacting species within a community (sensu Holt 1997). These interspecific groupings are global in distribution, yet become most speciose, and most temporally stable, in tropical forest ecosystems (Zou et al. 2018). The composition and richness of mixed-species flocks are known to be variable in many systems, and to change with elevation (Marín-Gómez and Arbeláez-Cortés 2015, Muñoz 2016, Montaño-Centellas 2020) as well as along human disturbance (Mokross et al. 2014, Zhou et al. 2019, Jones and Robinson 2020a) and successional (Zhang et al. 2013) gradients. It remains unclear, however, whether changes to flock composition reflect changes in the overall bird community or whether they occur independently of each other. In particular, there has been relatively little work on how and why these flocking systems change across elevational gradients in the tropics (Montaño-Centellas 2020). The lack of studies on this topic is surprising because we have long known that avian species richness decreases with increasing elevation (McCain 2009), changing taxonomic and functional composition dramatically. For instance, Neotropical bird assemblages lose insectivores and (proportionally) gain omnivores and frugivores with increasing altitude (Terborgh 1977, Jankowski et al. 2013b). Flock composition and richness patterns in a completely open-membership flock should therefore reflect these underlying changes to the community. Alternatively, structuring mechanisms of flock composition such as preferences for specific flock-mates or leader species (Mammides et al. 2015, Goodale et al. 2017), sharing a foraging stratum as a prerequisite for joining the flock (Sridhar et al. 2012), or within-guild competition (Colorado and Rodewald 2015), could result in richness and composition not reflective of the whole community. Because aggregations of flocking species bring together many species of insectivorous birds, they have long been thought to be characterized both by exploitation and interference competition (Morse 1970, Alatalo et al. 1987, Jabłoński and Don Lee 2002). Indeed, many flock-joining species show foraging niche partitioning by substrate (e.g., Eguchi et al. 1993, Matthysen et al. 2008). Yet mixed-species flocks are also thought to represent major facilitative interactions for forest birds by increasing foraging efficiency and by providing information and protection against predators (Sridhar et al. 2009, Goodale et al. 2020). Individuals that join flocks obtain access to social information about predators from other flock members and from specific, highly vigilant sentinel species, which enables them to exploit more exposed microhabitats (Dolby and Grubb 1999, Darrah and Smith 2013, Martínez et al. 2018). The facilitative effects of flocking might, for instance, extend the realized elevational distributions of flock-joining species (Goodale et al. 2020), and might buffer species responses to unsuitable habitats (Tubelis et al. 2006). In some systems, for instance, the community of flock-joining species changed less across land-use types (Goodale et al. 2014) and with selective logging (Borah et al. 2018) than did the overall community. Given the potential benefits gained (Sridhar et al. 2009), and the costs of competition when joining a flock (Gross 2008, Sridhar et al. 2012), individuals are expected to aggregate with each other in flocks when the benefits outweigh the costs of associating. These benefits will vary among species and be dependent upon habitat context and flock composition (Mokross et al. 2014, Borah et al. 2018), and by consequence flocking assemblages are expected to vary across environmental gradients. Understanding changes to mixed-species flocks across elevational gradients has been hampered by the many simultaneous mechanisms driving change (Zhou et al. 2019). First, increasing elevation entails changing abiotic conditions, notably decreasing mean temperature and barometric pressure, which confine the fundamental niche of tropical species (Jankowski et al. 2013a, Elsen et al. 2017). Because fat storage in birds is costly, low temperatures can impose an important tradeoff between increased starvation and predation risks (Lima 1986, Witter and Cuthill 1993). Flocking behavior can therefore be more prevalent at colder temperatures (Greenberg 2000, Gentry et al. 2019) because greater foraging benefits can potentially be gained from facilitation in extreme abiotic conditions (Callaway 1998, Stachowicz 2001). Thus, more species and individuals might actively join flocks at high elevations, where conditions are harsher. Second, the availability of food resources, in particular insects and fruits, is thought to decline with elevation (Terborgh 1977), perhaps in conjunction with decreasing productivity (Fisher et al. 2013). These patterns, however, can be highly variable (Supriya et al. 2019). Reduced access to insect resources at either seasonal (Develey and Peres 2000, Mangini and Areta 2018) or daily (Poulsen 1996) time scales increases participation in flocks, and a similar effect may occur at higher elevations. Third, vertical vegetation complexity, highly correlated with the number of available foraging niches (e.g., MacArthur and MacArthur 1961), declines with increasing elevation (Terborgh 1977, Acharya et al. 2011, Jankowski et al. 2013b), reducing the available foraging niche space. Mixed-species flocks are known to preferentially select for habitat with greater vegetation complexity (McDermott and Rodewald 2014, Potts et al. 2014) and have greater species richness in such microhabitats (Knowlton and Graham 2011, Zhang et al. 2013, Jones and Robinson 2020a). This is particularly true for flock-following insectivores, which are often highly specialized by substrate (Marra and Remsen 1997, Jones et al. 2020) and foraging height (Walther 2002, Mansor et al. 2019). How these mechanisms act upon flock structure has not yet been evaluated over a full elevational gradient. Patterns of flock richness and composition are of great interest in the Andes because Neotropical montane and lowland flocking systems differ in structure and behavior. For one, lowland flocking associations in primary forest consist of temporally stable species interactions (Martínez and Gomez 2013). In Amazonia, multiple species pairs defend a collective territory (Munn and Terborgh 1979) and flock composition is highly structured (Graves and Gotelli 1993). Andean flocking systems, by contrast, often show a dynamic, open-membership composition, with species frequently joining and leaving the flock as it enters and leaves their territories (Poulsen 1996, Pomara et al. 2007), and a much greater diversity of associating species pairs (Montaño-Centellas 2020, Jones and Robinson 2020b). Similarly, lowland flocks are characterized by multiple sub-types (Munn 1985), typically differentiated into understory and canopy flocks. Andean flocks instead often show a single homogenous flock composition, with both understory and canopy species freely admixed (Poulsen 1996, Guevara et al. 2011, Colorado and Rodewald 2015) and a lack of differentiable sub-types in network analyses (Montaño-Centellas 2020). It is therefore important to understand how these profound changes to flocking behavior affect flock composition and the diversity of participating species. In this study, we first described elevational patterns of avian richness, taxonomic and guild diversity, and composition in the tropical Andes of Bolivia. We then tested for potential mechanisms driving these elevational patterns. Specifically, we sampled flock composition, the community of flock-joining species, and the full avian community across a continuous elevational gradient of undisturbed forest to examine: (1) the extent to which the richness and composition of the flocking community are reflective of the overall bird community at each elevation, and (2) how the proportion of species of specific taxa and foraging guilds in flocks changes with elevation. Then, we explored (3) how flock species richness, size, and composition vary with elevation and (4) related these flock response variables to elevational changes in temperature (as a surrogate for abiotic stress), fruit and arthropod resource availability, and vertical vegetation complexity (a measure of available foraging strata). Each of these variables represents a mechanistic hypothesis explaining structuring forces across the elevational gradient (see Table 1 for detailed predictions). Alternatively, open-membership flocks could lack structuring mechanisms and instead reflect the greater bird community. Thus, we predicted that flocking species richness would decrease with elevation, as expected for the whole community. However, because facilitation is potentially a more important structuring force when conditions are harsher (Montaño-Centellas 2020), we expected a greater percentage of the community to join mixed-species flocks at high elevations. Similarly, we expected that the transition from separate understory and canopy flock types to a single flocking system should entail an increase in the overall number of flocking species at higher elevations. TABLE 1. Hypotheses pertaining to the structure of mixed-species flocks across elevations in the Andes of Bolivia. The name and mechanism explaining how the different drivers tested in this study operate upon flocking assemblages, the ecological outcome of each mechanism, the predicted patterns for overall changes to mixed-species flocks (MSF-predicted patterns), and for specific functional groups (guild-specific patterns) are presented. Hypothesis . Mechanism . Ecological outcome . MSF-predicted pattern . Guild-specific patterns . Open-membership hypothesis Abiotic and biotic filtering act upon avian assemblages shaping community structure. Flocking assemblages are structured by community assembly forces and therefore changes in richness and composition are reflective of those in the whole community. Species richness and taxonomic and functional composition mirror changes to the whole bird community. Proportionally, the number of omnivore and frugivore species and individuals increase with elevation, while the proportion of insectivores declines. Abiotic stress hypothesis Harsher abiotic conditions impose greater stress on individuals; thus, greater foraging benefits can be obtained by facilitating in extreme abiotic conditions. Flocking behavior is more prevalent where abiotic stress is higher, thus more species and individuals join flocks in extreme abiotic conditions. Higher proportion of the community joining mixed-species flocks at higher elevations with lower temperatures. Greater flock size with decreasing temperature. Greater richness and abundance of omnivores in high-elevation flocks, and a decline in the number of species and individuals of specialized taxa and functional guilds (e.g., insectivores, frugivores, canopy and ground foragers). Food resource availability hypothesis Reduced availability/diversity of food resources (insects and fruits) increases participation in flocks. Facilitation benefits related to foraging efficiency might be higher in areas where resources are depleted and/or unavailable. Greater species richness and number of individuals per flock with decreasing resource availability. Species richness and number of individuals of insectivores increase with decreasing arthropod diversity. Species richness and number of individuals of frugivores increase with decreasing fruit diversity. Foraging niche hypothesis Increased number of available foraging niches allows for finer partitioning of foraging strata and microhabitats among members of mixed-species flocks, reducing inter- and intraspecific competition. Greater vegetation complexity harbors larger and more speciose mixed-species flocks. Greater species richness and number of individuals per flock with increasing vegetation complexity. Greater richness and abundance of specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) with increasing vegetation complexity. Greater richness and abundance of generalist guilds (e.g., omnivores) with decreasing complexity. Hypothesis . Mechanism . Ecological outcome . MSF-predicted pattern . Guild-specific patterns . Open-membership hypothesis Abiotic and biotic filtering act upon avian assemblages shaping community structure. Flocking assemblages are structured by community assembly forces and therefore changes in richness and composition are reflective of those in the whole community. Species richness and taxonomic and functional composition mirror changes to the whole bird community. Proportionally, the number of omnivore and frugivore species and individuals increase with elevation, while the proportion of insectivores declines. Abiotic stress hypothesis Harsher abiotic conditions impose greater stress on individuals; thus, greater foraging benefits can be obtained by facilitating in extreme abiotic conditions. Flocking behavior is more prevalent where abiotic stress is higher, thus more species and individuals join flocks in extreme abiotic conditions. Higher proportion of the community joining mixed-species flocks at higher elevations with lower temperatures. Greater flock size with decreasing temperature. Greater richness and abundance of omnivores in high-elevation flocks, and a decline in the number of species and individuals of specialized taxa and functional guilds (e.g., insectivores, frugivores, canopy and ground foragers). Food resource availability hypothesis Reduced availability/diversity of food resources (insects and fruits) increases participation in flocks. Facilitation benefits related to foraging efficiency might be higher in areas where resources are depleted and/or unavailable. Greater species richness and number of individuals per flock with decreasing resource availability. Species richness and number of individuals of insectivores increase with decreasing arthropod diversity. Species richness and number of individuals of frugivores increase with decreasing fruit diversity. Foraging niche hypothesis Increased number of available foraging niches allows for finer partitioning of foraging strata and microhabitats among members of mixed-species flocks, reducing inter- and intraspecific competition. Greater vegetation complexity harbors larger and more speciose mixed-species flocks. Greater species richness and number of individuals per flock with increasing vegetation complexity. Greater richness and abundance of specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) with increasing vegetation complexity. Greater richness and abundance of generalist guilds (e.g., omnivores) with decreasing complexity. Open in new tab TABLE 1. Hypotheses pertaining to the structure of mixed-species flocks across elevations in the Andes of Bolivia. The name and mechanism explaining how the different drivers tested in this study operate upon flocking assemblages, the ecological outcome of each mechanism, the predicted patterns for overall changes to mixed-species flocks (MSF-predicted patterns), and for specific functional groups (guild-specific patterns) are presented. Hypothesis . Mechanism . Ecological outcome . MSF-predicted pattern . Guild-specific patterns . Open-membership hypothesis Abiotic and biotic filtering act upon avian assemblages shaping community structure. Flocking assemblages are structured by community assembly forces and therefore changes in richness and composition are reflective of those in the whole community. Species richness and taxonomic and functional composition mirror changes to the whole bird community. Proportionally, the number of omnivore and frugivore species and individuals increase with elevation, while the proportion of insectivores declines. Abiotic stress hypothesis Harsher abiotic conditions impose greater stress on individuals; thus, greater foraging benefits can be obtained by facilitating in extreme abiotic conditions. Flocking behavior is more prevalent where abiotic stress is higher, thus more species and individuals join flocks in extreme abiotic conditions. Higher proportion of the community joining mixed-species flocks at higher elevations with lower temperatures. Greater flock size with decreasing temperature. Greater richness and abundance of omnivores in high-elevation flocks, and a decline in the number of species and individuals of specialized taxa and functional guilds (e.g., insectivores, frugivores, canopy and ground foragers). Food resource availability hypothesis Reduced availability/diversity of food resources (insects and fruits) increases participation in flocks. Facilitation benefits related to foraging efficiency might be higher in areas where resources are depleted and/or unavailable. Greater species richness and number of individuals per flock with decreasing resource availability. Species richness and number of individuals of insectivores increase with decreasing arthropod diversity. Species richness and number of individuals of frugivores increase with decreasing fruit diversity. Foraging niche hypothesis Increased number of available foraging niches allows for finer partitioning of foraging strata and microhabitats among members of mixed-species flocks, reducing inter- and intraspecific competition. Greater vegetation complexity harbors larger and more speciose mixed-species flocks. Greater species richness and number of individuals per flock with increasing vegetation complexity. Greater richness and abundance of specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) with increasing vegetation complexity. Greater richness and abundance of generalist guilds (e.g., omnivores) with decreasing complexity. Hypothesis . Mechanism . Ecological outcome . MSF-predicted pattern . Guild-specific patterns . Open-membership hypothesis Abiotic and biotic filtering act upon avian assemblages shaping community structure. Flocking assemblages are structured by community assembly forces and therefore changes in richness and composition are reflective of those in the whole community. Species richness and taxonomic and functional composition mirror changes to the whole bird community. Proportionally, the number of omnivore and frugivore species and individuals increase with elevation, while the proportion of insectivores declines. Abiotic stress hypothesis Harsher abiotic conditions impose greater stress on individuals; thus, greater foraging benefits can be obtained by facilitating in extreme abiotic conditions. Flocking behavior is more prevalent where abiotic stress is higher, thus more species and individuals join flocks in extreme abiotic conditions. Higher proportion of the community joining mixed-species flocks at higher elevations with lower temperatures. Greater flock size with decreasing temperature. Greater richness and abundance of omnivores in high-elevation flocks, and a decline in the number of species and individuals of specialized taxa and functional guilds (e.g., insectivores, frugivores, canopy and ground foragers). Food resource availability hypothesis Reduced availability/diversity of food resources (insects and fruits) increases participation in flocks. Facilitation benefits related to foraging efficiency might be higher in areas where resources are depleted and/or unavailable. Greater species richness and number of individuals per flock with decreasing resource availability. Species richness and number of individuals of insectivores increase with decreasing arthropod diversity. Species richness and number of individuals of frugivores increase with decreasing fruit diversity. Foraging niche hypothesis Increased number of available foraging niches allows for finer partitioning of foraging strata and microhabitats among members of mixed-species flocks, reducing inter- and intraspecific competition. Greater vegetation complexity harbors larger and more speciose mixed-species flocks. Greater species richness and number of individuals per flock with increasing vegetation complexity. Greater richness and abundance of specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) with increasing vegetation complexity. Greater richness and abundance of generalist guilds (e.g., omnivores) with decreasing complexity. Open in new tab Finally, although we expected elevational patterns in flock characteristics to be influenced by all three mechanisms (abiotic stress, food resource availability, and the availability of foraging niches), we predicted different responses for different subsets of flocking species (Table 1). Specifically, we expected insectivore richness to decline at higher elevations due to a loss of lowland-associated taxa in favor of dominant bird families of the highlands (Thraupidae). In particular, we predicted that declining vegetation complexity would reduce the available foraging niches for specialists. However, we expected insectivore participation to increase with decreasing arthropod availability whereas frugivore richness and abundance per flock were predicted to increase with a decline in fruit availability. Accordingly, because milder conditions and greater vegetation complexity, likely associated with a greater foraging niche space at lower elevations might allow communities to contain more specialist species with smaller niches, we expected the number of species in generalist guilds (e.g., omnivores) participating in flocks to increase with elevation. By contrast, the richness and abundance of species from specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) participating in flocks should be higher at lower elevations. A summary of the hypotheses and proposed mechanisms driving elevational patterns is presented in Table 1. METHODS Mixed-Species Flock Surveys Avian mixed-species flocks were surveyed for two consecutive years, from May to October 2016 and May to August 2017, along a continuous elevational gradient (~1,350–3,550 m.a.s.l. [meters above sea level]) in Cotapata National Park, a protected area in the Andes of western Bolivia. Our focal gradient encompasses all forested habitats within the park, from the treeline to the valley bottom (Montaño-Centellas 2020), passing through a landscape dominated by evergreen humid montane and cloud forests on steep slopes (Ribera 1995, Sevilla Callejo 2010). The extreme topography in the study region prevents large human settlements (Montaño-Centellas and Garitano-Zavala 2015), although there is a small village (~10 families) at 1,950 m with small orchards scattered nearby. To eliminate the potential confounding effects of this village, all flocks observed between 1,700 and 2,100 m were not considered in the analyses. The gradient was sampled 16 times (8 times per year) through randomized 350-m elevation surveys, so that elevations were not sampled sequentially to avoid temporal biases. Further details on the study system and survey methods can be found in Montaño-Centellas et al. (2020). Briefly, each survey consisted of walking at a slow constant pace along the transect and for each flock encountered, recording GPS coordinates, elevation, and the number and identity of all participating species. In this study, we defined a flock as a gathering of at least two species foraging <10 m apart while moving in the same direction. Flock data were collected using the “gambit of the group” method, where several individuals observed at the same time and place are assumed to be associating (Farine and Whitehead 2015). To diminish the bias of registering non-flocking individuals that increased their foraging activity while the flock was present, we recorded the whole encounter until the flock was unreachable and excluded all species that remained in the study areas (feeding, perching, and/or vocalizing) after the flock had left. In our study, each detected flock represents a snapshot of the realized associations resulting from individual birds deciding whether to join the flock, and thus, flocks observed on different days were considered independent replicates of flock composition and richness. Species richness is known to be correlated with observation time; thus, we further standardized our effort by including only flocks that were observed for between 10 and 20 min. Data were organized into abundance matrices, with flocks as columns and species as rows. Percentage of the Community That Participates in Mixed-Species Flocks We determined the proportion of the total community participating in mixed-species flocks by comparing whole-community survey data to flock composition data. Community surveys were conducted between April and August 2014 and April and October 2015 (Montaño-Centellas et al. 2020). Given the extreme topography of the landscape, we followed an approximate line-transect methodology of surveying birds, adapted from similar work in other montane areas (Elsen et al. 2017). Complete details on survey methodology are presented in Montaño-Centellas et al. (2020). Briefly, surveys consisted of walking at a slow pace along 350-m elevational subsections across the gradient and registering every encountered bird (visual and auditory detections). For each detected individual, GPS coordinates and elevation were noted, as well as taxonomic identity. The complete gradient was surveyed 20 times (10 per year), randomizing the order of visits to different elevations. We summarized this information to obtain the composition of the full bird community at each elevation by pooling all species observed in 100-m-wide elevational bands (no interpolation was performed). We similarly determined the flock-joining community in the same elevational bands by grouping together all species observed in mixed-species flocks within each elevational band. For each elevational band, we then calculated the proportion of the overall community and the proportion of species from a given taxon (Thraupidae, Furnariidae, and Tyrannidae), foraging guild (insectivores, frugivores, and omnivores), and foraging stratum (canopy, mid-height, understory, and ground) that participated in flocks. We visually examined how the percentage of the community that joined mixed-species flocks changed with elevation for the whole community and for each community subset. Predictors of Flock Diversity and Composition For each observed flock, we quantified the species richness, a number of individuals participating, species diversity using Shannon’s index H’ and species evenness (i.e., how equally distributed species abundances were within a flock) using Pielou’s index J’ (Magurran 2004). We also calculated the number of species and individuals in specific taxa (Thraupidae, Furnariidae, Tyrannidae), foraging guilds (insectivores, frugivores, omnivores), and vertical strata (canopy, mid-height, understory, ground foragers). Information on diet and foraging stratum was extracted from Wilman et al. (2014) and modified based on field observations. Species richness and diversity were calculated with functions of the package vegan (Oksanen et al. 2018). We examined the effects of four environmental variables that might drive differences in the diversity and composition of mixed-species flocks across elevations: mean daily temperature (highly correlated with elevation, and a surrogate for abiotic conditions), vegetation complexity, arthropod diversity, and fruit availability. Temperature was measured hourly from April to October 2015 at five elevations (1,350, 2,000, 2,500, 3,000, and 3,500 m) using permanent stations equipped with a datalogger (EasyLog EL-USB-2-LCD) and summarized as mean daily temperature at each elevation. We then fitted a linear regression to these values and used the regression model (F1,3 = 41.16, R2 = 0.93, P < 0.001) to predict mean temperature values for the elevation of each flock. Vegetation complexity was measured each 100-m of elevation (e.g., 3,550, 3,450, 3,350, etc; average linear distance within each 100-m elevational band was 1,100 m), using a modified version of the understory height diversity index (UHD) used by MacArthur and MacArthur (1961). At each elevation, we selected 10 random points separated by at least 25 m from one another. At each point, we set an imaginary 1-m diameter cylinder and noted the presence or absence of vegetation in 7 height intervals: 0–1 m, 1–2 m, 2–4 m, 4–8 m, 8–16 m, 16–24 m, and >24 m. We used data from the 10 cylinders to calculate the proportion of times vegetation was present in each height interval, for each elevation. UHD was then calculated as the Shannon diversity index H’ = Ʃ pi ln(pi), where pi is the proportion of vegetation in the ith height interval (Montaño-Centellas et al. 2020). Arthropod and fruit availability were measured between June and July 2016 for each 100 m of elevation across the gradient. At each elevation, arthropods were sampled at 4 points separated by at least 50 m from one another, using a beating method that consisted of placing a 1-m2 cloth supported with a frame under the vegetation and shaking or beating the vegetation to collect the arthropods that dislodge (Cooper and Whitmore 1990). We collected arthropods by shaking the vegetation at two different heights at the same time (1 m and 3 m), using the same intensity of beating. Collected arthropods were manually processed, preserved, and identified by researchers associated with the Universidad Mayor de San Andres (Montaño-Centellas et al. 2020). Arthropod availability was then calculated as the Shannon’s diversity index H’ of morphospecies of arthropods captured at each elevation. Fruit availability was defined as the diversity of ornithochoric (i.e., bird-dispersed) plants with ripe fruits within a circular plot of ~100-m radius located at each elevation. Plants were surveyed with binoculars by two trained botanists by walking within each plot for a period of 3–4 hr. Both the taxonomic identity and the abundance of each fruiting plant species were noted. When possible, voucher specimens were collected to help species identification in the Herbario Nacional de Bolivia (Montaño-Centellas et al. 2020). Data on species identity and abundance were used to calculate the Shannon’s diversity index of ornithochoric species at each elevation. Because we were interested in the diversity of available resources instead of the net amount, we quantified abundance with numbers instead of biomass and chose a diversity metric to describe them. However, resource abundance and resource diversity are likely positively correlated (Abrams 1995, Hanz et al. 2019), although this is not necessarily true for every system (Hails 1982). We tested for redundancy among predictors with Pearson’s product-moment correlation and found overall low values (Supplementary Material Figure S1), except between temperature and fruit diversity (r = 0.72, P < 0.001). We thus regressed fruits as a function of temperature, extracted the regression residuals, and used these values as a new variable to describe relative fruit availability for a given elevation (we used this new variable in all analyses). Statistical Analyses We first explored elevational patterns of each of our response variables: species richness, number of individuals, taxonomic diversity, and species evenness per flock using generalized linear models (GLMs), with elevation and elevation2 as independent variables. For count data (species richness and a number of individuals per flock), we used quasi-Poisson errors to account for overdispersion of the data. For Shannon’s diversity index, we assumed Gamma-distributed errors and an identity link, as data were continuous but restricted to positive values; for Pielou’s evenness index, with values ranging between 0 and 1, we assumed a beta distribution of errors. Secondly, we tested for the effects of 4 environmental predictors (mean daily temperature, arthropod and fruit availability, and vegetation complexity) on each of our response variables. We used GLMs, assuming error distributions as described above, with our predictors as independent variables. In accordance with our hypotheses (see Table 1), models for families Furnariidae and Tyrannidae, and for insectivores, included temperature, arthropod diversity, and vegetation complexity as predictors; models for frugivores used temperature, fruit diversity, and vegetation complexity; all other models included all four predictors. We used an information-theoretic approach to evaluate model goodness-of-fit. Models with continuous response variables were ranked using Akaike’s information criterion (AIC) and models with count response variables with quasi-AIC (QAIC; Burnham and Anderson 2004). In every case, competing models had goodness-of-fit that was not far from that of the best model (ΔAIC < 2); thus, we followed Burnham and Anderson (2004) to perform model averaging across competing models for that response variable. We report the averaged model coefficients and predictor significance based on its 95% confidence intervals (CI) (Burnham and Anderson 2004). Regressions were performed in packages base, MASS 7.3 (Ripley et al. 2018) and betareg 3.0 (Zeileis et al. 2016) while model fit, AIC and QAIC calculations, and model averaging were performed using functions of the packages MuMIn 1.43.15 (Barton 2019) and AICcmodavg 2.2-2. In order to understand how our predictor variables influenced flock species composition, we further ran a canonical correspondence analysis (CCA; Ter Braak 1986) using vegan 2.5–6 (Oksanen et al. 2018) and ade4 1.7–15 packages. Species abundances were log(x + 1) transformed and environmental variables were scaled prior to analysis. We examined linear dependencies by computing each environmental variable’s variance inflation factor (VIF) to examine for collinearity among predictors. Values of VIF above 10 suggest collinearity and should be avoided (McCune et al. 2002, Borcard et al. 2018). All calculated values were less than two (VIF = 1.68, 1.17, 1.99, 1.44 for temperature, arthropod availability, vegetation complexity, and fruit availability, respectively), suggesting no collinearity among predictors, and we thus retained all four in our canonical model. RESULTS Percentage of the Community That Participates in Mixed-Species Flocks across Elevations The avian community along the gradient was composed of 300 species (excluding flyovers); 162 of these species participated in flocks, encompassing ~60% of the bird families observed in the full community (Supplementary Material Table S17). We observed a total of 368 mixed-species flocks along our studied gradient. The most speciose families participating in flocks across elevations were Thraupidae, Tyrannidae, and Furnariidae (with 57, 40, and 19 species, respectively). The most frequently observed species were Anisognathus igniventris, Basileuterus luteoviridis, Diglossa cyanea, Myioborus melanocephalus, Chlorospingus ophthalmicus, B. signatus, and B. tristriatus, accounting for almost 36% of all individuals observed. Although not flock-following species (and thus not included in our analyses), some species were nevertheless frequently seen actively vocalizing and foraging while the flock visited their territories. Such species included eight hummingbirds (Adelomyia melanogenys, Aglaiocercus kingii, Coeligena violifer, C. inca, Heliangelus amethisticollis, Metallura aeneocauda, M. tyrianthina, and Phaethornis hispidus), two antpittas (Grallaria erythrotis and G. rufula), two cotingas (Rupicola peruviana and Lipaugus uropygialis), one wren (Henicorrhina leucophrys) and a toucan (Andigena cucullata). Additionally, we observed squirrels (Sciurus spp.) actively following the flock while feeding and vocalizing at three different elevations (1,430, 1,550, and 2,350 m). The number of species joining flocks declined with elevation, mirroring richness decreases for the whole community (Figure 1A). An average of ~54% of species in the full bird community participated in mixed-species flocks across elevations, with a larger percentage of species joining flocks at lower elevations (Figure 1B). There was a steep decrease in the number of species of thraupids and tyrannids in the flocking community with increasing elevation (Figure 2B and C), but the proportion of thraupids in the flocking community that join flocks increased sharply at higher elevations (Supplementary Material Figure S2C). Regardless of elevation, most of the flock-joining species were insectivores, and similar numbers of frugivores and omnivores participated in flocks across elevations (Supplementary Material Figure S2F). Though the number of flocking insectivores decreased with elevation (Figure 2D), they represented a higher proportion of the flocking community at higher elevations (Supplementary Material Figure S2F). Contrary to our expectations, the species richness of omnivores in the flocking community, and the percentage of the overall omnivore community they represented, was lower at high elevations (Supplementary Material Figure S2D and E). Although the number of frugivores in the flocking community decreased with elevation (Figure 2E), they represented a higher percentage of available frugivores at high elevations (Supplementary Material Figure S2E). FIGURE 1. Open in new tabDownload slide (A) Number of species in the overall bird community and participating in mixed-species flocks across elevations and (B) the percentage of the whole community that participating species represent. Lines correspond to generalized additive models (GAM) presented to better examine the trend. FIGURE 1. Open in new tabDownload slide (A) Number of species in the overall bird community and participating in mixed-species flocks across elevations and (B) the percentage of the whole community that participating species represent. Lines correspond to generalized additive models (GAM) presented to better examine the trend. FIGURE 2. Open in new tabDownload slide Number of species in the bird community (in black) and number of species that join mixed-species flocks (in gray) across elevations in the Tropical Andes of Bolivia. Species from different families (A–C), different feeding guilds (D–F), and different foraging strata guilds (G–J) are presented separately. Lines correspond to generalized additive models (GAM) presented to better examine the trend. FIGURE 2. Open in new tabDownload slide Number of species in the bird community (in black) and number of species that join mixed-species flocks (in gray) across elevations in the Tropical Andes of Bolivia. Species from different families (A–C), different feeding guilds (D–F), and different foraging strata guilds (G–J) are presented separately. Lines correspond to generalized additive models (GAM) presented to better examine the trend. As expected, the number of canopy species in flocks decreased with elevation with fewer than 5 canopy specialists in flocks above 2,800 m (Figure 2G). These numbers, however, represented a high proportion of the available canopy specialists in the community, with virtually all canopy specialists above 2,800 m participating in flocks (Supplementary Material Figure S2H). Most of the participating species in flocks foraged in the mid-height stratum, representing the greatest proportion of the flocking community at middle elevations (Supplementary Material Figure S2I). The number and proportion of mid-height-foraging species declined at higher elevations, however (Figure 2H and Supplementary Material Figure S2G). By contrast, the number and proportion of participating understory species in flocks were relatively constant across elevations (Figure 2I), though these numbers represented a higher proportion of the whole flocking community at higher elevations (Supplementary Material Figure S2H). Finally, ground-foraging species participation in flocks increased in number at higher elevations (Figure 2J) and represented a greater percentage of the available ground foragers in cloud forest communities (~2,800 m; Supplementary Material Figure S2H). Elevational Patterns of Species Diversity per Flock and Flock Size Species richness per flock varied from 2 to 24 (median = 8). Similarly, a number of individuals per flock ranged from 3 to 103 (median = 21). Both species richness and the number of individuals per flock changed across elevations, showing a peak in the middle of the gradient at ~2,300 m (Figure 3A and B; Supplementary Material Table S1). Species diversity per flock (measured as Shannon’s index H’) ranged from 0.28 to 2.93 (mean = 1.77), whereas species evenness (measured as Pielou’s evenness index) ranged from 0.4 to 1 (mean = 0.88). Both species diversity and evenness of flock composition changed across elevations, with the highest diversity values observed in the middle of the gradient at ~2,500 m, and the greatest species evenness at high elevations (Figure 3C and D; Supplementary Material Table S1). FIGURE 3. Open in new tabDownload slide (A) Species richness, (B) number of individuals, (C) Shannon’s diversity index H’, and (D) Pielou’s species evenness index J’, for mixed-species flocks along an elevational gradient in the Bolivian Andes. In each panel, points represent an independent flock (n = 368). Lines represent the quadratic regressions of each variable as a function of elevation; models are presented in Supplementary Material Table S1. FIGURE 3. Open in new tabDownload slide (A) Species richness, (B) number of individuals, (C) Shannon’s diversity index H’, and (D) Pielou’s species evenness index J’, for mixed-species flocks along an elevational gradient in the Bolivian Andes. In each panel, points represent an independent flock (n = 368). Lines represent the quadratic regressions of each variable as a function of elevation; models are presented in Supplementary Material Table S1. Predictors of Flock Diversity and Composition We found a positive, yet weak, the effect of vegetation complexity on species richness and a number of individuals per flock, with larger and more speciose flocks in areas with high vegetation complexity (Table 2; see Supplementary Material Table S2 for AIC tables). Whereas temperature had no effect on species richness per flock, it had a negative effect on flock size, with greater numbers of individuals per flock in cold, high-elevation flocks. Although none of our predictors explained overall species diversity (Shannon’s H’), species evenness significantly increased with increasing vegetation complexity and diversity of fruiting resources (Table 2). TABLE 2. Parameter estimation for all predictors in averaged generalized linear models (GLMs) for species richness, number of individuals, Shannon’s diversity index (H’), and species evenness (Pielou’s index J’) of mixed-species flocks in the Andes of Bolivia. The GLM for species richness and number of individuals assumed quasi-Poisson errors, the GLMs for Shannon’s diversity assumed Gaussian errors, and the GLM for Pielou’s J’ followed a beta distribution. Variables with model-averaged confidence intervals that overlap zero were considered not significant for the model. Variables marked with an asterisk are considered supported. AIC tables for all models averaged are presented in Supplementary Material Tables S2–S5. Model . Parameter . Unconditional SE . 95% Confidence interval . . Richness Temperature –0.02 0.02 –0.07 0.03 Vegetation complexity 0.05 0.02 0.01 0.10* Fruits –0.02 0.03 –0.08 0.04 Arthropods 0.00 0.02 –0.04 0.03 Number of individuals Temperature –0.10 0.01 –0.13 –0.07* Vegetation complexity 0.05 0.02 0.01 0.08* Fruits –0.04 0.02 –0.08 0.00 Arthropods 0.01 0.01 –0.02 0.03 Shannon’s H’ Temperature –0.02 0.03 –0.09 0.04 Vegetation complexity 0.03 0.03 –0.03 0.09 Fruits –0.01 0.04 –0.09 0.07 Arthropods 0.01 0.03 –0.05 0.06 Pielou’s J’ Temperature 0.05 0.06 –0.07 0.17 Vegetation complexity 0.13 0.06 0.02 0.25* Fruits 0.35 0.09 0.17 0.53* Arthropods –0.08 0.05 –0.19 0.02 Model . Parameter . Unconditional SE . 95% Confidence interval . . Richness Temperature –0.02 0.02 –0.07 0.03 Vegetation complexity 0.05 0.02 0.01 0.10* Fruits –0.02 0.03 –0.08 0.04 Arthropods 0.00 0.02 –0.04 0.03 Number of individuals Temperature –0.10 0.01 –0.13 –0.07* Vegetation complexity 0.05 0.02 0.01 0.08* Fruits –0.04 0.02 –0.08 0.00 Arthropods 0.01 0.01 –0.02 0.03 Shannon’s H’ Temperature –0.02 0.03 –0.09 0.04 Vegetation complexity 0.03 0.03 –0.03 0.09 Fruits –0.01 0.04 –0.09 0.07 Arthropods 0.01 0.03 –0.05 0.06 Pielou’s J’ Temperature 0.05 0.06 –0.07 0.17 Vegetation complexity 0.13 0.06 0.02 0.25* Fruits 0.35 0.09 0.17 0.53* Arthropods –0.08 0.05 –0.19 0.02 Open in new tab TABLE 2. Parameter estimation for all predictors in averaged generalized linear models (GLMs) for species richness, number of individuals, Shannon’s diversity index (H’), and species evenness (Pielou’s index J’) of mixed-species flocks in the Andes of Bolivia. The GLM for species richness and number of individuals assumed quasi-Poisson errors, the GLMs for Shannon’s diversity assumed Gaussian errors, and the GLM for Pielou’s J’ followed a beta distribution. Variables with model-averaged confidence intervals that overlap zero were considered not significant for the model. Variables marked with an asterisk are considered supported. AIC tables for all models averaged are presented in Supplementary Material Tables S2–S5. Model . Parameter . Unconditional SE . 95% Confidence interval . . Richness Temperature –0.02 0.02 –0.07 0.03 Vegetation complexity 0.05 0.02 0.01 0.10* Fruits –0.02 0.03 –0.08 0.04 Arthropods 0.00 0.02 –0.04 0.03 Number of individuals Temperature –0.10 0.01 –0.13 –0.07* Vegetation complexity 0.05 0.02 0.01 0.08* Fruits –0.04 0.02 –0.08 0.00 Arthropods 0.01 0.01 –0.02 0.03 Shannon’s H’ Temperature –0.02 0.03 –0.09 0.04 Vegetation complexity 0.03 0.03 –0.03 0.09 Fruits –0.01 0.04 –0.09 0.07 Arthropods 0.01 0.03 –0.05 0.06 Pielou’s J’ Temperature 0.05 0.06 –0.07 0.17 Vegetation complexity 0.13 0.06 0.02 0.25* Fruits 0.35 0.09 0.17 0.53* Arthropods –0.08 0.05 –0.19 0.02 Model . Parameter . Unconditional SE . 95% Confidence interval . . Richness Temperature –0.02 0.02 –0.07 0.03 Vegetation complexity 0.05 0.02 0.01 0.10* Fruits –0.02 0.03 –0.08 0.04 Arthropods 0.00 0.02 –0.04 0.03 Number of individuals Temperature –0.10 0.01 –0.13 –0.07* Vegetation complexity 0.05 0.02 0.01 0.08* Fruits –0.04 0.02 –0.08 0.00 Arthropods 0.01 0.01 –0.02 0.03 Shannon’s H’ Temperature –0.02 0.03 –0.09 0.04 Vegetation complexity 0.03 0.03 –0.03 0.09 Fruits –0.01 0.04 –0.09 0.07 Arthropods 0.01 0.03 –0.05 0.06 Pielou’s J’ Temperature 0.05 0.06 –0.07 0.17 Vegetation complexity 0.13 0.06 0.02 0.25* Fruits 0.35 0.09 0.17 0.53* Arthropods –0.08 0.05 –0.19 0.02 Open in new tab As predicted by the abiotic stress hypothesis, temperature was the most important predictor for the number of species and individuals of several taxa and guilds (Table 3; see Supplementary Material Tables S2–S15 for AIC tables). When examining responses of different bird families, we found a negative effect of temperature on thraupid species richness per flock and on the number of individuals of thraupids and furnariids per flock. However, increasing temperature instead had a positive effect on the number of species and individuals of the family Tyrannidae. When examining foraging guild responses, we found a negative effect of increasing temperature on the number of omnivorous individuals, but not species, supporting our prediction of more generalists joining flocks upslope. Temperature also had a negative effect on the number of insectivore individuals, with higher abundances of insectivores at colder, higher elevations. Finally, temperature had a positive effect on the number of canopy-specialist species, with higher species richness at lower elevations, but a negative effect on the number of ground-foraging species and individuals and the number of mid-height foraging individuals. TABLE 3. Parameter estimation for predictors in averaged models for species richness and number of individuals per flock from different families (Furnariidae, Thraupidae, and Tyrannidae), foraging guilds (insectivores, frugivores, and omnivores), and height strata (canopy, mid-height, understory, and ground). Models for Furnariidae, Tyrannidae, and insectivores included temperature, vegetation complexity, and arthropods as predictor variables. Models for frugivores included temperature, vegetation complexity, and fruits as predictor variables. All other models included all four predictors (see Table 1). SE stands for unconditional standard error of the parameter estimate. Variables with model-averaged confidence intervals that overlap zero were considered not significant for the model. Variables marked with asterisk are considered supported. AIC tables for all models averaged are presented in Supplementary Material Tables S6–S15. . Models . . Richness . Number of individuals . Taxa/guild . Parameter . SE . 95% CI . . Parameter . SE . 95% CI . . Furnariidae Temperature –0.06 0.06 –0.17 0.05 –0.25 0.04 –0.33 –0.17* Vegetation complexity 0.01 0.06 –0.11 0.13 0.01 0.04 –0.07 0.09 Arthropods –0.06 0.05 –0.16 0.05 –0.08 0.04 –0.15 –0.01* Thraupidae Temperature –0.17 0.03 –0.24 –0.1* –0.23 0.02 –0.28 –0.19* Vegetation complexity 0.00 0.04 –0.07 0.08 0.06 0.02 0.02 0.10* Fruits –0.03 0.05 –0.13 0.07 –0.03 0.03 –0.10 0.04 Arthropods 0.02 0.03 –0.05 0.08 0.07 0.02 0.03 0.10* Tyrannidae Temperature 0.25 0.05 0.16 0.35* 0.31 0.04 0.23 0.39* Vegetation complexity 0.04 0.07 –0.08 0.17 0.02 0.05 –0.08 0.13 Arthropods –0.06 0.05 –0.15 0.03 –0.07 0.04 –0.15 0.00 Insectivores Temperature –0.02 0.03 –0.08 0.03 –0.12 0.02 –0.15 –0.09* Vegetation complexity 0.09 0.03 0.03 0.14* 0.09 0.02 0.06 0.12* Arthropods 0.00 0.02 –0.04 0.05 0.00 0.01 –0.03 0.03 Frugivores Temperature 0.00 0.05 –0.10 0.10 –0.04 0.02 –0.09 0.01 Vegetation complexity 0.08 0.05 –0.03 0.18 0.01 0.03 –0.06 0.08* Fruits –0.08 0.07 –0.21 0.05 –0.06 0.03 –0.13 0.00 Omnivores Temperature –0.01 0.06 –0.12 0.10 –0.14 0.03 –0.20 –0.07* Vegetation complexity 0.04 0.06 –0.07 0.15 0.00 0.05 –0.09 0.09 Fruits 0.00 0.07 –0.15 0.14 –0.11 0.05 –0.21 –0.02 Arthropods –0.07 0.05 –0.16 0.02 –0.04 0.03 –0.10 0.02 Canopy Temperature 0.16 0.06 0.05 0.28* –0.02 0.04 –0.11 0.06 Vegetation complexity –0.03 0.08 –0.18 0.13 –0.13 0.04 –0.21 –0.06* Fruits –0.08 0.08 –0.24 0.08 –0.15 0.05 –0.26 –0.05* Arthropods –0.06 0.06 –0.18 0.06 –0.18 0.03 –0.25 –0.12* Understory Temperature –0.06 0.05 –0.16 0.04 0.02 0.03 –0.04 0.08 Vegetation complexity 0.07 0.05 –0.03 0.18 0.12 0.03 0.06 0.18* Fruits 0.01 0.07 –0.12 0.15 0.01 0.04 –0.07 0.09 Arthropods 0.02 0.04 –0.06 0.10 0.05 0.03 0.00 0.09 Mid-height Temperature 0.02 0.03 –0.05 0.08 –0.10 0.02 –0.14 –0.06* Vegetation complexity 0.06 0.03 0.01 0.12* 0.04 0.02 0.00 0.08 Fruits –0.03 0.04 –0.12 0.05 –0.06 0.03 –0.11 –0.01* Arthropods –0.02 0.03 –0.07 0.03 0.01 0.01 –0.02 0.04 Ground Temperature –0.48 0.11 –0.70 –0.26* –0.78 0.08 –0.92 –0.63* Vegetation complexity 0.17 0.10 –0.02 0.36 0.24 0.07 0.09 0.38* Fruits 0.09 0.17 –0.23 0.42 0.20 0.11 –0.01 0.41 Arthropods 0.10 0.07 –0.04 0.24 0.21 0.05 0.12 0.29* . Models . . Richness . Number of individuals . Taxa/guild . Parameter . SE . 95% CI . . Parameter . SE . 95% CI . . Furnariidae Temperature –0.06 0.06 –0.17 0.05 –0.25 0.04 –0.33 –0.17* Vegetation complexity 0.01 0.06 –0.11 0.13 0.01 0.04 –0.07 0.09 Arthropods –0.06 0.05 –0.16 0.05 –0.08 0.04 –0.15 –0.01* Thraupidae Temperature –0.17 0.03 –0.24 –0.1* –0.23 0.02 –0.28 –0.19* Vegetation complexity 0.00 0.04 –0.07 0.08 0.06 0.02 0.02 0.10* Fruits –0.03 0.05 –0.13 0.07 –0.03 0.03 –0.10 0.04 Arthropods 0.02 0.03 –0.05 0.08 0.07 0.02 0.03 0.10* Tyrannidae Temperature 0.25 0.05 0.16 0.35* 0.31 0.04 0.23 0.39* Vegetation complexity 0.04 0.07 –0.08 0.17 0.02 0.05 –0.08 0.13 Arthropods –0.06 0.05 –0.15 0.03 –0.07 0.04 –0.15 0.00 Insectivores Temperature –0.02 0.03 –0.08 0.03 –0.12 0.02 –0.15 –0.09* Vegetation complexity 0.09 0.03 0.03 0.14* 0.09 0.02 0.06 0.12* Arthropods 0.00 0.02 –0.04 0.05 0.00 0.01 –0.03 0.03 Frugivores Temperature 0.00 0.05 –0.10 0.10 –0.04 0.02 –0.09 0.01 Vegetation complexity 0.08 0.05 –0.03 0.18 0.01 0.03 –0.06 0.08* Fruits –0.08 0.07 –0.21 0.05 –0.06 0.03 –0.13 0.00 Omnivores Temperature –0.01 0.06 –0.12 0.10 –0.14 0.03 –0.20 –0.07* Vegetation complexity 0.04 0.06 –0.07 0.15 0.00 0.05 –0.09 0.09 Fruits 0.00 0.07 –0.15 0.14 –0.11 0.05 –0.21 –0.02 Arthropods –0.07 0.05 –0.16 0.02 –0.04 0.03 –0.10 0.02 Canopy Temperature 0.16 0.06 0.05 0.28* –0.02 0.04 –0.11 0.06 Vegetation complexity –0.03 0.08 –0.18 0.13 –0.13 0.04 –0.21 –0.06* Fruits –0.08 0.08 –0.24 0.08 –0.15 0.05 –0.26 –0.05* Arthropods –0.06 0.06 –0.18 0.06 –0.18 0.03 –0.25 –0.12* Understory Temperature –0.06 0.05 –0.16 0.04 0.02 0.03 –0.04 0.08 Vegetation complexity 0.07 0.05 –0.03 0.18 0.12 0.03 0.06 0.18* Fruits 0.01 0.07 –0.12 0.15 0.01 0.04 –0.07 0.09 Arthropods 0.02 0.04 –0.06 0.10 0.05 0.03 0.00 0.09 Mid-height Temperature 0.02 0.03 –0.05 0.08 –0.10 0.02 –0.14 –0.06* Vegetation complexity 0.06 0.03 0.01 0.12* 0.04 0.02 0.00 0.08 Fruits –0.03 0.04 –0.12 0.05 –0.06 0.03 –0.11 –0.01* Arthropods –0.02 0.03 –0.07 0.03 0.01 0.01 –0.02 0.04 Ground Temperature –0.48 0.11 –0.70 –0.26* –0.78 0.08 –0.92 –0.63* Vegetation complexity 0.17 0.10 –0.02 0.36 0.24 0.07 0.09 0.38* Fruits 0.09 0.17 –0.23 0.42 0.20 0.11 –0.01 0.41 Arthropods 0.10 0.07 –0.04 0.24 0.21 0.05 0.12 0.29* Open in new tab TABLE 3. Parameter estimation for predictors in averaged models for species richness and number of individuals per flock from different families (Furnariidae, Thraupidae, and Tyrannidae), foraging guilds (insectivores, frugivores, and omnivores), and height strata (canopy, mid-height, understory, and ground). Models for Furnariidae, Tyrannidae, and insectivores included temperature, vegetation complexity, and arthropods as predictor variables. Models for frugivores included temperature, vegetation complexity, and fruits as predictor variables. All other models included all four predictors (see Table 1). SE stands for unconditional standard error of the parameter estimate. Variables with model-averaged confidence intervals that overlap zero were considered not significant for the model. Variables marked with asterisk are considered supported. AIC tables for all models averaged are presented in Supplementary Material Tables S6–S15. . Models . . Richness . Number of individuals . Taxa/guild . Parameter . SE . 95% CI . . Parameter . SE . 95% CI . . Furnariidae Temperature –0.06 0.06 –0.17 0.05 –0.25 0.04 –0.33 –0.17* Vegetation complexity 0.01 0.06 –0.11 0.13 0.01 0.04 –0.07 0.09 Arthropods –0.06 0.05 –0.16 0.05 –0.08 0.04 –0.15 –0.01* Thraupidae Temperature –0.17 0.03 –0.24 –0.1* –0.23 0.02 –0.28 –0.19* Vegetation complexity 0.00 0.04 –0.07 0.08 0.06 0.02 0.02 0.10* Fruits –0.03 0.05 –0.13 0.07 –0.03 0.03 –0.10 0.04 Arthropods 0.02 0.03 –0.05 0.08 0.07 0.02 0.03 0.10* Tyrannidae Temperature 0.25 0.05 0.16 0.35* 0.31 0.04 0.23 0.39* Vegetation complexity 0.04 0.07 –0.08 0.17 0.02 0.05 –0.08 0.13 Arthropods –0.06 0.05 –0.15 0.03 –0.07 0.04 –0.15 0.00 Insectivores Temperature –0.02 0.03 –0.08 0.03 –0.12 0.02 –0.15 –0.09* Vegetation complexity 0.09 0.03 0.03 0.14* 0.09 0.02 0.06 0.12* Arthropods 0.00 0.02 –0.04 0.05 0.00 0.01 –0.03 0.03 Frugivores Temperature 0.00 0.05 –0.10 0.10 –0.04 0.02 –0.09 0.01 Vegetation complexity 0.08 0.05 –0.03 0.18 0.01 0.03 –0.06 0.08* Fruits –0.08 0.07 –0.21 0.05 –0.06 0.03 –0.13 0.00 Omnivores Temperature –0.01 0.06 –0.12 0.10 –0.14 0.03 –0.20 –0.07* Vegetation complexity 0.04 0.06 –0.07 0.15 0.00 0.05 –0.09 0.09 Fruits 0.00 0.07 –0.15 0.14 –0.11 0.05 –0.21 –0.02 Arthropods –0.07 0.05 –0.16 0.02 –0.04 0.03 –0.10 0.02 Canopy Temperature 0.16 0.06 0.05 0.28* –0.02 0.04 –0.11 0.06 Vegetation complexity –0.03 0.08 –0.18 0.13 –0.13 0.04 –0.21 –0.06* Fruits –0.08 0.08 –0.24 0.08 –0.15 0.05 –0.26 –0.05* Arthropods –0.06 0.06 –0.18 0.06 –0.18 0.03 –0.25 –0.12* Understory Temperature –0.06 0.05 –0.16 0.04 0.02 0.03 –0.04 0.08 Vegetation complexity 0.07 0.05 –0.03 0.18 0.12 0.03 0.06 0.18* Fruits 0.01 0.07 –0.12 0.15 0.01 0.04 –0.07 0.09 Arthropods 0.02 0.04 –0.06 0.10 0.05 0.03 0.00 0.09 Mid-height Temperature 0.02 0.03 –0.05 0.08 –0.10 0.02 –0.14 –0.06* Vegetation complexity 0.06 0.03 0.01 0.12* 0.04 0.02 0.00 0.08 Fruits –0.03 0.04 –0.12 0.05 –0.06 0.03 –0.11 –0.01* Arthropods –0.02 0.03 –0.07 0.03 0.01 0.01 –0.02 0.04 Ground Temperature –0.48 0.11 –0.70 –0.26* –0.78 0.08 –0.92 –0.63* Vegetation complexity 0.17 0.10 –0.02 0.36 0.24 0.07 0.09 0.38* Fruits 0.09 0.17 –0.23 0.42 0.20 0.11 –0.01 0.41 Arthropods 0.10 0.07 –0.04 0.24 0.21 0.05 0.12 0.29* . Models . . Richness . Number of individuals . Taxa/guild . Parameter . SE . 95% CI . . Parameter . SE . 95% CI . . Furnariidae Temperature –0.06 0.06 –0.17 0.05 –0.25 0.04 –0.33 –0.17* Vegetation complexity 0.01 0.06 –0.11 0.13 0.01 0.04 –0.07 0.09 Arthropods –0.06 0.05 –0.16 0.05 –0.08 0.04 –0.15 –0.01* Thraupidae Temperature –0.17 0.03 –0.24 –0.1* –0.23 0.02 –0.28 –0.19* Vegetation complexity 0.00 0.04 –0.07 0.08 0.06 0.02 0.02 0.10* Fruits –0.03 0.05 –0.13 0.07 –0.03 0.03 –0.10 0.04 Arthropods 0.02 0.03 –0.05 0.08 0.07 0.02 0.03 0.10* Tyrannidae Temperature 0.25 0.05 0.16 0.35* 0.31 0.04 0.23 0.39* Vegetation complexity 0.04 0.07 –0.08 0.17 0.02 0.05 –0.08 0.13 Arthropods –0.06 0.05 –0.15 0.03 –0.07 0.04 –0.15 0.00 Insectivores Temperature –0.02 0.03 –0.08 0.03 –0.12 0.02 –0.15 –0.09* Vegetation complexity 0.09 0.03 0.03 0.14* 0.09 0.02 0.06 0.12* Arthropods 0.00 0.02 –0.04 0.05 0.00 0.01 –0.03 0.03 Frugivores Temperature 0.00 0.05 –0.10 0.10 –0.04 0.02 –0.09 0.01 Vegetation complexity 0.08 0.05 –0.03 0.18 0.01 0.03 –0.06 0.08* Fruits –0.08 0.07 –0.21 0.05 –0.06 0.03 –0.13 0.00 Omnivores Temperature –0.01 0.06 –0.12 0.10 –0.14 0.03 –0.20 –0.07* Vegetation complexity 0.04 0.06 –0.07 0.15 0.00 0.05 –0.09 0.09 Fruits 0.00 0.07 –0.15 0.14 –0.11 0.05 –0.21 –0.02 Arthropods –0.07 0.05 –0.16 0.02 –0.04 0.03 –0.10 0.02 Canopy Temperature 0.16 0.06 0.05 0.28* –0.02 0.04 –0.11 0.06 Vegetation complexity –0.03 0.08 –0.18 0.13 –0.13 0.04 –0.21 –0.06* Fruits –0.08 0.08 –0.24 0.08 –0.15 0.05 –0.26 –0.05* Arthropods –0.06 0.06 –0.18 0.06 –0.18 0.03 –0.25 –0.12* Understory Temperature –0.06 0.05 –0.16 0.04 0.02 0.03 –0.04 0.08 Vegetation complexity 0.07 0.05 –0.03 0.18 0.12 0.03 0.06 0.18* Fruits 0.01 0.07 –0.12 0.15 0.01 0.04 –0.07 0.09 Arthropods 0.02 0.04 –0.06 0.10 0.05 0.03 0.00 0.09 Mid-height Temperature 0.02 0.03 –0.05 0.08 –0.10 0.02 –0.14 –0.06* Vegetation complexity 0.06 0.03 0.01 0.12* 0.04 0.02 0.00 0.08 Fruits –0.03 0.04 –0.12 0.05 –0.06 0.03 –0.11 –0.01* Arthropods –0.02 0.03 –0.07 0.03 0.01 0.01 –0.02 0.04 Ground Temperature –0.48 0.11 –0.70 –0.26* –0.78 0.08 –0.92 –0.63* Vegetation complexity 0.17 0.10 –0.02 0.36 0.24 0.07 0.09 0.38* Fruits 0.09 0.17 –0.23 0.42 0.20 0.11 –0.01 0.41 Arthropods 0.10 0.07 –0.04 0.24 0.21 0.05 0.12 0.29* Open in new tab We also found a significantly positive response to increasing vegetation structure (Table 3), primarily in the number of individuals in the flock, as expected by the foraging niche hypothesis (Table 1). Among the taxonomic subsets sampled, only the number of thraupid individuals in the flock was significantly associated with increasing vegetation structure. For insectivores, however, both the species richness and a number of individuals were positively related to increasing structure. A significant effect of vegetation structure was also detected for the number of frugivores in the flock, though the more generalist omnivores did not show a significant response. Vegetation structure also exerted significant effects across vertical foraging strata, with the number of ground- and understory-foraging species in flocks significantly greater in more structured forest. For species that forage in the mid-height stratum, the significant effect was on the species richness rather than the number of individuals, while for canopy foragers the effect of vegetation structure was instead significantly negative. Finally, and as expected by the food resource hypothesis (Table 1), we found significantly negative correlations between fruit and arthropod diversity and the number of individuals in flocks (Table 3), though the sign of this effect differed across community subsets. The number of furnariids and canopy-foraging species increased with decreasing arthropod diversity, though the opposite effect was true for thraupids and ground-foraging species. A significantly negative effect of fruit diversity was also found on the abundance of canopy- and mid-height-foragers in a flock. Predictors of Taxonomic Species Composition The full CCA model (F4,363 = 5.77, P = 0.001) had three significant axes that explained 17% of the variation in bird species composition per flock (Figure 4; Supplementary Material Table S16). Axis 1 (λ = 0.70) was highly correlated with temperature (r = −0.95, P < 0.001); species with different elevational distributions were separated along Axis 1. Species in the center of this axis could be either ubiquitous or associated with middle elevations. Axis 2 (λ = 0.14) mostly represented a habitat complexity gradient (r = −0.60, P < 0.0001), with species being separated based on their use of habitats with different complexity of vertical strata. Finally, Axis 3 (λ = 0.14) was negatively correlated with fruit availability (r = −0.50, P < 0.0001) and positively but weakly correlated with arthropod availability (r = 0.30, P < 0.0001), potentially describing a gradient of resources for frugivores and insectivores. FIGURE 4. Open in new tabDownload slide Ordination of bird species in flocks plotted as a function of the first two axes of a canonical correspondence analysis considering four predictor variables: temperature, vegetation complexity, fruit availability, and arthropod availability. Axis 1 represents a temperature gradient, with cooler areas toward the right and warmer areas to the left. Axis 2 mainly represents a gradient of vegetation complexity, with species associated with more complex vegetation structure in the lower half of the plot, and species associated with less complex and more open habitats with higher fruit availability in the upper half of the plot. FIGURE 4. Open in new tabDownload slide Ordination of bird species in flocks plotted as a function of the first two axes of a canonical correspondence analysis considering four predictor variables: temperature, vegetation complexity, fruit availability, and arthropod availability. Axis 1 represents a temperature gradient, with cooler areas toward the right and warmer areas to the left. Axis 2 mainly represents a gradient of vegetation complexity, with species associated with more complex vegetation structure in the lower half of the plot, and species associated with less complex and more open habitats with higher fruit availability in the upper half of the plot. DISCUSSION We documented profound changes to the richness, size, evenness, and composition of mixed-species flocks, and the overall flock-joining community, along our elevational gradient. We observed a mid-elevation peak in flock species richness, size, and Shannon’s diversity at around 2,300 m (Figure 3). This trend was poorly explained by any single variable, however, and we suggest it is the result of a combination of behavioral and environmental factors. As predicted by the open-membership hypothesis, flock composition and the make-up of the flocking bird community closely mirrored changes to the overall bird community: not only did overall and taxon-specific richness decline at the same rate in the full community and flock-joining subset (Figure 1), but changes to flock composition followed similar trends as changes to community composition (Figure 2). We argue these results suggest that Andean flocks represent dynamic, unstructured associations that mainly reflect the composition of the larger community. Second, flocks were dominated by insectivorous species and the mid-elevation maximum of flock richness both closely mirrored a similar peak in the percentage of the local insectivore community joining flocks and was significantly predicted by local vegetation structure. The peak in mid-elevation richness in flocks, therefore, reflects both the open-membership composition of high-elevation flocks and the high vegetation complexity of mid-elevation forests, maximizing within-flock diversity. Changes to overall flock structure across elevations were therefore best supported by the foraging niche and abiotic stress hypotheses, while neither fruit nor arthropod abundance was significantly correlated with either overall richness or the abundance or richness of most community subsets. While it is possible that this might partially be a result of our arthropod sampling method (branch beating) that might underestimate insect diversity (Montaño-Centellas et al. 2020), we argue that ground-dwelling arthropods are unlikely to constitute a major component of the diet of flocking species, given the relative lack of ground-foraging species in the flocking community across elevations (Figure 2I). Instead, we believe the structuring effects of abiotic stress and changes to the diversity of foraging niches are more important than those of resource abundance in Andean flocking systems. High- and low-elevation flocks had distinguishable composition (Table 3, Figure 4), and higher elevations on our gradient represented areas with low temperatures and low vegetation complexity, the latter resulting from a loss of vertical strata (Terborgh 1977, Jankowski et al. 2013b). Many differences in guild composition and flock size were associated with temperature, with flocks at colder elevations having more individuals, especially of insectivorous and omnivorous species, greater thraupid species richness and abundance, and higher richness and abundance of ground-foraging species. By contrast, warmer-elevation flocks were of smaller size but contained more tyrannid and canopy-foraging species and individuals than their high-elevation counterparts. Vegetation complexity was also a good predictor of flock compositional changes. Flocks at mid and low elevations, where vegetation complexity was greater, had higher numbers of insectivorous species and individuals, higher number of thraupid individuals, higher richness of mid-height-foraging species, and higher abundances of canopy-, understory-, and ground-foraging species. Lower canopy heights and a simplified vegetation structure at higher elevations may engender a loss of foraging niches for insectivores, leading to a loss of insectivore species richness and functional diversity (Jankowski et al. 2013b). Our system may even underestimate the effect of vegetation structure because inter-Andean valley bottoms not connected to lowland forests may lack many lowland avian taxa (e.g., Thamnophilidae; Kessler et al. 2001). Open-Membership Flocks: Flock Composition Mirrors Changes to the Overall Community Our finding of large changes to the flocking bird community and flock composition across the elevational gradient matches similar findings elsewhere (Marín-Gómez and Arbeláez-Cortés 2015, Muñoz 2016). Interestingly, these changes to the flock-joining community mirrored changes to the absolute and relative abundance of taxa and foraging guilds within the full bird community (Figures 1A and 2), as posited by the open-membership hypothesis. While the percentage of the full community participating in flocks declined with increasing elevation, it represented a majority (50–60%) of the community over most of the gradient (Figure 1B), a similar proportion to other tropical forest systems (Zou et al. 2018). Changes to the full and flock-joining communities along our gradient mirrored similar changes to bird communities elsewhere in the Andes, notably an overall decline in species richness (e.g., Kattan and Franco 2004) driven by a steep decline in the number of insectivorous species (Terborgh 1977, Jankowski et al. 2013b, Pigot et al. 2016; Figure 2D). With increasing elevation, we found a steady decrease in the number and proportion of the flocking community made up of sub-oscine families (Furnariidae, Tyrannidae), and a proportional increase in the species richness of thraupids in the flock-joining community as well as in flocks (Table 3; Supplementary Material Figure S2C). These changes match known shifts in overall representation of these families in full Andean montane bird communities (Renjifo et al. 1997). We suggest that the correlation between the species richness and composition of the flock-joining and full bird communities is a product of the open-membership nature of Andean flocks. Indeed, it has long been noted that Andean mixed-species flocks show tremendous variation in their size and species richness (Poulsen 1996, Guevara et al. 2011, Montaño-Centellas 2020, Jones and Robinson 2020a), even at local study sites. This stands in notable contrast to the stable composition of Amazonian understory (Martínez and Gomez 2013, Mokross et al. 2014) and canopy (Munn 1985) flocks, where the core species spend all their time with the flock. Similar observational studies in Andean systems, however, found that many species dynamically joined and left the flock as it entered and exited their territories (Poulsen 1996, Pomara et al. 2007), resulting in a more variable flock composition over time. More recently, network analyses have shown that a high proportion of all species pairs associate in Andean flocks (Jones and Robinson 2020b), with higher density values than expected by chance (Montaño-Centellas 2020). These associations seem to be able to “rewire”, changing the intensity of the species interactions or the make-up of the participating species without a loss of the behavior as a whole (Montaño-Centellas 2020, Jones and Robinson 2020b). Taken together, these results suggest that Andean mixed-species flocks are representative of the larger bird community and might be useful indicators of the local avifauna for Andean biodiversity surveys (Montaño-Centellas and Garitano-Zavala 2015). Vegetation Structure Shapes Flock Participation: The Importance of Foraging Niche Diversity We found a significant effect of vegetation structure on the species richness, size, and evenness of flocks, and vegetation had an important effect on species richness and abundance of insectivores in flocks (Table 3). As in flocking systems in forests worldwide (Sridhar et al. 2009), the flock-joining bird community was dominated by insectivores in our system. We found that vegetation structure decreased with altitude, a pattern consistent with other elevational gradients (Terborgh 1977, Acharya et al. 2011, Jankowski et al. 2013b), although we observed a low-elevation plateau in structural diversity that began declining above 2,500 m.a.s.l. We argue this decline in complexity reduces the diversity of available foraging niches for insectivores, as we found a significant positive correlation between greater vegetation structure and higher richness of mid-height-foraging species in flocks. Decreasing vegetation structure therefore resulted in a high-elevation flocking community with increased proportions of understory and ground-foraging species (Supplementary Material Figure S2). In the Eastern Himalaya, Acharya and Vijayan (2017) found that foliage concentration and avian species richness were highest in the mid-layers (5–15 m) of montane forest. Because tropical forest birds exhibit strong vertical stratification in their foraging (Walther 2002, Mansor et al. 2019), the loss of mid-story and subcanopy vegetation strata with decreasing vertical structure could be driving much of this effect. For example, most of the mid-height-foraging tyrannid genera (e.g., Leptopogon, Myiophobus, Phylloscartes, Phyllomyias) are lost at high elevations, where tyrant flycatcher richness in flocks is much diminished. At the microhabitat level, flock diversity is known to increase with vegetation structure (Knowlton and Graham 2011, Zhang et al. 2013, Jones and Robinson 2020a), and flocks are known to select for greater vegetation complexity (Lee et al. 2005, McDermott and Rodewald 2014, Potts et al. 2014), providing further support for this hypothesis. If declining vegetation structure at high elevations explains the decline in flock richness from the mid-elevation peak, then how to explain the lower insectivore participation at low elevations despite high vegetation structure? We argue that a gradual increase in the structuring of flock composition with decreasing elevation might drive these observed patterns. Lowland flocking systems in the tropics often contain two or more flock types structured by foraging stratum (e.g., King and Rappole 2000, Srinivasan et al. 2012), whereas Andean flocking systems are consistently described as consisting of a mixture of understory and canopy species (e.g., Guevara et al. 2011, Colorado and Rodewald 2015). Changes from flocks structured by foraging stratum to open-membership flocks, however, are better described as a continuum. Network analyses along our gradient found that the modularity of flocking networks, a surrogate for the extent to which distinct flock sub-types exist at a given elevation, declined with altitude and was greater than expected by chance in the lowest elevation band (1,500 m.a.s.l.; Montaño-Centellas 2020). We therefore suggest that the transition of flocking behavior toward an open-membership system at ~2,300 m may allow for a greater proportion of the insectivore community to join flocks due to the lack of a structuring effect of foraging stratum. A potential mechanism driving these changes to flock composition is an increase in participation costs for flocking species and individuals at lower elevations. With increasing insectivore richness at low elevations, their partitioning of foraging microhabitats will likely be finer (Pigot et al. 2016), leading to more foraging specialists participating in flocks. With greater specialization and narrower niches, competition for resources is expected to increase (Cavender-Bares et al. 2009), leading to tighter niche packing (Pigot et al. 2016). Evidence for competition as a driver of the structure of Andean flocks has previously been reported at similar elevations (<2,500 m) to those at which we detected a decline in insectivore participation (Colorado and Rodewald 2015). Intriguingly, greater vegetation structure increased the number of individuals participating, but not the species richness, for many community subsets (e.g., thraupids, frugivores, understory foragers), suggesting that the costs of competition may be borne unevenly within some species that join flocks as a social group, as is the case in temperate flocks (e.g., juvenile individuals; Hogstad 1989). Alternatively, the costs of “activity matching” (adjusting foraging rate while in the flock) in flocks more structured by foraging stratum might prevent more dissimilar species from joining flocks (Hutto 1988, Sridhar and Guttal 2018, Jones et al. 2020), resulting in less speciose and less open-membership flocks with a more clear distinction between canopy and understory sub-types at lower elevations. Increasing Flock Size with Elevation: Reduced Food Resources and Increasing Abiotic Stress? In contrast to species richness, flock size and the number of thraupid individuals in flocks were significantly greater at high elevations. We suggest that these results reflect a shift from primarily insectivorous species that join flocks as singletons or pairs to largely gregarious omnivorous and frugivorous species that join flocks as intraspecific groups. Many thraupid genera in high-elevation cloud forests (Anisognathus, Buthraupis, Iridosornis, Diglossa, Kleinothraupis) are known to join flocks in large intraspecific groups (Poulsen 1996, Arbeláez-Cortés and Marín-Gomez 2012). Most of these genera are also frequently observed foraging in large monospecific groups where individuals benefit from group-size supplementary benefits (Goodale et al. 2020). For instance, many high-Andean thraupids supplement their insect diet with fruits or nectar (Remsen 1985), and, in fact, the most generalist categories in our guild analyses (thraupid species richness and abundance per flock, number of omnivores, and mid-height foragers) increased at high elevations (Table 3). It is possible to group participation facilitates the finding of food resources at these high elevations, where fruits have a patchier distribution (Valone 1989). The abundance of canopy and mid-height foragers in flocks significantly increased with declining fruit diversity, suggesting that high intraspecific abundance within flocks is a product of frugivorous foraging as a single-species group (Valburg 1992). This idea is further supported by an increase in the evenness of flocks with elevation (Figure 3D), suggesting that high intraspecific abundance in flocks was more evenly distributed across species at high elevations. Alternatively, predation-related benefits associated with joining flocks at high elevations, where niches are less diverse, might also benefit generalists that are gregarious (Gil et al. 2017). Greater facilitation is expected among conspecifics under harsher abiotic conditions at the highest elevations (Montaño-Centellas 2020), and even solitary Andean birds may increase their flocking propensity with increasing elevation (e.g., Cardellina canadensis; Céspedes and Bayly 2018). Thus, the gregarious behavior of many generalist thraupids at high elevations, itself a product of declining food availability and increasing abiotic stress, likely drives the increase in flock size. CONCLUSIONS: AN OPEN-MEMBERSHIP SYSTEM WITH ABIOTIC AND BIOTIC STRUCTURING FACTORS In the montane system studied here, an increase in abiotic stress and the loss of vegetation structure are the main environmental drivers shaping flock structure. As expected for the whole community, decreasing temperature and increasing abiotic stress were significantly associated with elevational changes to flock structure and individual-level participation. As noted elsewhere (Graham et al. 2012), assemblages at higher elevations contained fewer specialized species and occupied a smaller functional space. Vegetation structure was another general mechanism shaping flock composition along our gradient by directly affecting the availability of foraging niches, particularly for insectivores. The benefits gained by joining a flock vary among species and depend upon habitat context and flock composition, however (Borah et al. 2018). Greater vegetation structure opens new foraging microhabitats for species to exploit, but may also reduce species participation at low elevations through increased niche partitioning and subsequent foraging competition. Similarly, increasing abiotic stress and declining resource availability decrease specialist species richness in flocks, but can also increase individual- and species-level participation by generalist species at high elevations. Nevertheless, our findings primarily support the open-membership hypothesis that Andean flocks resemble the composition and structure of the avian community at different elevations rather than the influence of structuring mechanisms. This open-membership, however, itself varies with elevation, with a more restrictive access and a better definition of flock sub-types at lower elevations. Acknowledgments This study would not have been possible without volunteers and students that participated in data collection: R. Cortes, M. Montenegro, P. Velasquez, S. Moscoso, C. Mayta, M. Yapu, C. Agurto, M. Perez, D. Solano, K. Bedregal, and A. Vilchez. We are grateful to B. Nieto and Y. Fernandez for leading surveys and identification of botanical data, and to K. Losantos and N. Ohara for leading arthropod surveys and identification. B. A. Loiselle provided important feedback at the early stages of the project. We thank J. G. Blake and S. K. Robinson for helpful comments on the manuscript. Funding statement: This work was supported by a Graduate School Fellowship Award from the University of Florida, a UF Biodiversity Institute summer fellowship, grants from the American Ornithologist Union, American Philosophical Society, The Rufford Foundation, and a WWF EFN-Alumni Grant, all granted to F.A.M.C. Ethics statement: This project was conducted in full compliance with the Institutional Animal Care and Use Committee of the University of Florida (permit #201609398), and with permits provided by the Ministerio de Medio Ambiente y Agua and the Servicio Nacional de Areas Protegidas, Plurinational State of Bolivia. Author contributions: F.A.M.C. conceived the design, developed the sampling methods, collected the data, wrote the manuscript, and analyzed the data. H.H.J. heavily contributed to manuscript writing and editing. Both authors approved the final version and have no conflict of interest to declare. Data availability: Raw data on mixed-species flocks and environmental variables can be obtained from Montaño-Centellas and Jones (2021). LITERATURE CITED Abrams , P. A . ( 1995 ). Monotonic or unimodal diversity-productivity gradients: What does competition theory predict? Ecology 76 : 2019 – 2027 . Google Scholar Crossref Search ADS WorldCat Acharya , B. K. , N. J. Sanders, L. Vijayan, and B. Chettri ( 2011 ). Elevational gradients in bird diversity in the Eastern Himalaya: An evaluation of distribution patterns and their underlying mechanisms . PLoS One 6 : e29097 . Google Scholar Crossref Search ADS PubMed WorldCat Acharya , B. K. , and L. Vijayan ( 2017 ). Vertical stratification of birds in different vegetation types along an elevation gradient in the Eastern Himalaya, India . Ornithological Science 16 : 131 – 140 . Google Scholar Crossref Search ADS WorldCat Alatalo , R. V. , D. A. G. Eriksson, L. Gustafsson, and K. Larsson ( 1987 ). Exploitation competition influences the use of foraging sites by tits: Experimental evidence . Ecology 68 : 284 – 290 . Google Scholar Crossref Search ADS WorldCat Arbeláez-Cortés , E. , and O. H. Marín-Gomez ( 2012 ). The composition of mixed-species bird flocks in Alto Quindío, Colombia . The Wilson Journal of Ornithology 124 : 572 – 580 . Google Scholar Crossref Search ADS WorldCat Barton , K . ( 2019 ). Package ‘MuMIn’: Multi-Model inference. R package, version 1.43.15 . https://CRAN.R-project.org/package=MuMIn Borah , B. , S. Quader, and U. Srinivasan ( 2018 ). Responses of interspecific associations in mixed-species bird flocks to selective logging . Journal of Applied Ecology 55 : 1637 – 1646 . Google Scholar Crossref Search ADS WorldCat Borcard , D. , F. Gillet, and P. Legendre ( 2018 ). Numerical Ecology with R . Springer-Verlag , New York, NY, USA . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Burnham , K. , and D. Anderson ( 2004 ). Model Selection and Multi-Model Inference. A Practical Information-Theoretical Approach , 2nd edition. Springer-Verlag , New York, NY, USA . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Callaway , R. M . ( 1998 ). Competition and facilitation on elevation gradients in subalpine forests of the Northern Rocky Mountains, USA . Oikos 82 : 561 – 573 . Google Scholar Crossref Search ADS WorldCat Cavender-Bares , J. , K. H. Kozak, P. V. Fine, and S. W. Kembel ( 2009 ). The merging of community ecology and phylogenetic biology . Ecology Letters 12 : 693 – 715 . Google Scholar Crossref Search ADS PubMed WorldCat Céspedes, L. N., and N. J. Bayly (2018). Over-winter ecology and relative density of Canada Warbler Cardellina canadensis in Colombia: The basis for defining conservation priorities for a sharply declining long-distance migrant. Bird Conservation International 29:232–248. Colorado , G. J. , and A. D. Rodewald ( 2015 ). Assembly patterns of mixed-species avian flocks in the Andes . The Journal of Animal Ecology 84 : 386 – 395 . Google Scholar Crossref Search ADS PubMed WorldCat Cooper , R. , and R. Whitmore ( 1990 ). Arthropod sampling methods in ornithology. In Avian Foraging: Theory, Methodology, and Applications (M. L. Morrison, C. J. Ralph, J. Verner, and J. R. Jehl, Jr., Editors) . Studies in Avian Biology 13 : 29 – 37 . Google Scholar OpenURL Placeholder Text WorldCat Darrah , A. J. , and K. G. Smith ( 2013 ). Comparison of foraging behaviors and movement patterns of the Wedge-billed Woodcreeper (Glyphorynchus spirurus) traveling alone and in mixed-species flocks in Amazonian Ecuador . The Auk 130 : 629 – 636 . Google Scholar Crossref Search ADS WorldCat Develey , P. F. , and C. A. Peres ( 2000 ). Resource seasonality and the structure of mixed species bird flocks in a coastal Atlantic forest of southeastern Brazil . Journal of Tropical Ecology 16 : 33 – 53 . Google Scholar Crossref Search ADS WorldCat Dolby , A. S. , and T. C. Grubb ( 1999 ). Functional roles in mixed-species foraging flocks: A field manipulation . The Auk 116 : 557 – 559 . Google Scholar Crossref Search ADS WorldCat Eguchi , K. , S. Yamagishi, and V. Randriansolo ( 1993 ). The composition and foraging behaviour of mixed-species flocks of forest-living birds in Madagascar . Ibis 135 : 91 – 96 . Google Scholar Crossref Search ADS WorldCat Elsen , P. R. , M. W. Tingley, R. Kalyanaraman, K. Ramesh, and D. S. Wilcove ( 2017 ). The role of competition, ecotones, and temperature in the elevational distribution of Himalayan birds . Ecology 98 : 337 – 348 . Google Scholar Crossref Search ADS PubMed WorldCat Farine , D. R. , and H. Whitehead ( 2015 ). Constructing, conducting and interpreting animal social network analysis . The Journal of Animal Ecology 84 : 1144 – 1163 . Google Scholar Crossref Search ADS PubMed WorldCat Fisher, J. B., Y. Malhi, I. C. Torres, D. B. Metcalfe, M. J. van de Weg, P. Meir, J. E. Silva-Espejo, and W. H. Huasco (2013). Nutrient limitation in rainforests and cloud forests along a 3,000-m elevation gradient in the Peruvian Andes. Oecologia 172:889–902. Gentry , K. E. , D. P. Roche, S. G. Mugel, N. D. Lancaster, K. E. Sieving, T. M. Freeberg, and J. R. Lucas ( 2019 ). Flocking propensity by satellites, but not core members of mixed-species flocks, increases when individuals experience energetic deficits in a poor-quality foraging habitat . PLoS One 14 : e0209680 . Google Scholar Crossref Search ADS PubMed WorldCat Gil , M. A. , Z. Emberts, H. Jones, and C. M. St Mary ( 2017 ). Social information on fear and food drives animal grouping and fitness . The American Naturalist 189 : 227 – 241 . Google Scholar Crossref Search ADS PubMed WorldCat Goodale , E. , G. Beauchamp, and G. Ruxton ( 2017 ). Mixed-Species Animal Groups: Behavior, Community Structure and Conservation . Academic Press , London, UK . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Goodale , E. , S. W. Kotagama, T. R. S. Raman, S. Sidhu, U. M. Goodale, S. Parker, and J. Chen ( 2014 ). The response of birds and mixed-species bird flocks to human-modified landscapes in Sri Lanka and southern India . Forest Ecology and Management 329 : 384 – 392 . Google Scholar Crossref Search ADS WorldCat Goodale , E. , H. Sridhar, K. E. Sieving, P. Bangal, G. J. Colorado Z, D. R. Farine, E. W. Heymann, H. H. Jones, I. Krams, A. E. Martínez, et al. ( 2020 ). Mixed company: A framework for understanding the composition and organization of mixed-species animal groups . Biological Reviews of the Cambridge Philosophical Society 95 : 889 – 910 . Google Scholar Crossref Search ADS PubMed WorldCat Graham , C. H. , J. L. Parra, B. A. Tinoco, F. G. Stiles, and J. A. McGuire ( 2012 ). Untangling the influence of ecological and evolutionary factors on trait variation across hummingbird assemblages . Ecology 93 : S99 – S111 . Google Scholar Crossref Search ADS WorldCat Graves , G. R. , and N. J. Gotelli ( 1993 ). Assembly of avian mixed-species flocks in Amazonia . Proceedings of the National Academy of Sciences USA 90 : 1388 – 1391 . Google Scholar Crossref Search ADS WorldCat Greenberg , R . ( 2000 ). Birds of many feathers: The formation and structure of mixed-species flocks of forest birds. In On the Move: How and Why Animals Travel in Groups ( S. Boinski and P. A. Garber, Editors). University of Chicago Press , Chicago, IL, USA . pp. 521 – 558 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Gross , K . ( 2008 ). Positive interactions among competitors can produce species-rich communities . Ecology Letters 11 : 929 – 936 . Google Scholar Crossref Search ADS PubMed WorldCat Guevara , E. A. , J. C. Valarezo, A. Onofa, and F. Cupuerán ( 2011 ). Mixed-species flock composition in a northwestern Ecuadorian cloud forest . Ornitologia Neotropical 22 : 379 – 386 . Google Scholar OpenURL Placeholder Text WorldCat Hails , C. J . ( 1982 ). A comparison of tropical and temperate aerial insect abundance . Biotropica 14 : 310 – 313 . Google Scholar Crossref Search ADS WorldCat Hanz , D. M. , K. Böhning-Gaese, S. W. Ferger, S. A. Fritz, E. L. Neuschulz, M. Quitián, V. Santillán, T. Töpfer, and M. Schleuning ( 2019 ). Functional and phylogenetic diversity of bird assemblages are filtered by different biotic factors on tropical mountains . Journal of Biogeography 46 : 291 – 303 . Google Scholar Crossref Search ADS WorldCat Hogstad , O . ( 1989 ). Social organization and dominance behavior in some Parus species . The Wilson Bulletin 101 : 254 – 262 . Google Scholar OpenURL Placeholder Text WorldCat Holt , R. D . ( 1997 ). Community modules. In Multitrophic Interactions in Terrestrial Ecosystems ( A. C. Gange and V. K. Brown, Editors). Blackwell Science , Oxford, UK . pp. 333–349. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hutto , R. L . ( 1988 ). Foraging behaviour patterns suggest a possible cost associated with participation in mixed-species bird flocks . Oikos 51 : 79 – 83 . Google Scholar Crossref Search ADS WorldCat Jabłoński , P. G. , and S. Don Lee ( 2002 ). Foraging niche shifts in mixed-species flocks of tits in Korea . Journal of Field Ornithology 73 : 246 – 252 . Google Scholar Crossref Search ADS WorldCat Jankowski , J. E. , G. A. Londono, S. K. Robinson, and M. A. Chappell ( 2013a ). Exploring the role of physiology and biotic interactions in determining elevational ranges of tropical animals . Ecography 36 : 1 – 12 . Google Scholar Crossref Search ADS WorldCat Jankowski , J. E. , C. L. Merkord, W. F. Rios, K. G. Cabrera, N. S. Revilla, and M. R. Silman ( 2013b ). The relationship of tropical bird communities to tree species composition and vegetation structure along an Andean elevational gradient . Journal of Biogeography 40 : 950 – 962 . Google Scholar Crossref Search ADS WorldCat Jones , H. H. , and S. K. Robinson ( 2020a ). Patch size and vegetation structure drive changes to mixed-species flock diversity and composition across a gradient of fragment sizes in the Western Andes of Colombia . The Condor: Ornithological Applications 122 : 1 – 26 . doi:10.1093/condor/duaa006 Google Scholar Crossref Search ADS WorldCat Jones , H. H. , and S. K. Robinson ( 2020b ). Vegetation structure drives mixed-species flock interaction strength and nuclear species roles . Behavioral Ecology 32 : 69 – 81 . Google Scholar Crossref Search ADS WorldCat Jones , H. H. , M. J. Walters, and S. K. Robinson ( 2020 ). Do similar foragers flock together? Nonbreeding foraging behavior and its impact on mixed-species flocking associations in a subtropical region . The Auk: Ornithological Advances 137 : 1 – 16 . doi:10.1093/auk/ukz079 Google Scholar Crossref Search ADS WorldCat Kattan , G. H. , and P. Franco ( 2004 ). Bird diversity along elevational gradients in the Andes of Colombia: Area and mass effects . Global Ecology and Biogeography 13 : 451 – 458 . Google Scholar Crossref Search ADS WorldCat Kessler , M. , S. K. Hertzog, J. Fjeldsa, and K. Bach ( 2001 ) Species richness and endemism of plant and bird communities along two gradients of elevation, humidity and land use in the Bolivian Andes . Diversity and Distributions 7 : 61 – 77 . Google Scholar Crossref Search ADS WorldCat King , D. I. , and J. H. Rappole ( 2000 ). Winter flocking of insectivorous birds in montane pine-oak forests in middle America . The Condor 102 : 664 – 672 . Google Scholar Crossref Search ADS WorldCat Knowlton , J. L. , and C. H. Graham ( 2011 ). Species interactions are disrupted by habitat degradation in the highly threatened Tumbesian region of Ecuador . Ecological Applications 21 : 2974 – 2986 . Google Scholar Crossref Search ADS WorldCat Lee , T. M. , M. C. Soh, N. Sodhi, L. P. Koh, and S. L.-H. Lim ( 2005 ). Effects of habitat disturbance on mixed species bird flocks in a tropical sub-montane rainforest . Biological Conservation 122 : 193 – 204 . Google Scholar Crossref Search ADS WorldCat Lima , S. L . ( 1986 ). Predation risk and unpredictable feeding conditions: Determinants of body mass in birds . Ecology 67 : 377 – 385 . Google Scholar Crossref Search ADS WorldCat MacArthur , R. H. , and J. W. MacArthur ( 1961 ). On bird species diversity . Ecology 42 : 594 – 598 . Google Scholar Crossref Search ADS WorldCat Magurran , A. E . ( 2004 ). Measuring Biological Diversity . Blackwell Publishing , Malden, MA, USA . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Mammides , C. , J. Chen, U. M. Goodale, S. W. Kotagama, S. Sidhu, and E. Goodale ( 2015 ). Does mixed-species flocking influence how birds respond to a gradient of land-use intensity? Proceedings of the Royal Society of London B: Biological Sciences 282 : 20151118 . Google Scholar OpenURL Placeholder Text WorldCat Mangini , G. G. , and J. I. Areta ( 2018 ). Bird mixed-species flock formation is driven by low temperatures between and within seasons in a subtropical Andean-foothill forest . Biotropica 50 : 816 – 825 . Google Scholar Crossref Search ADS WorldCat Mansor , M. S. , F. Z. Rozali, N. A. Abdullah, S. M. Nor, and R. Ramli ( 2019 ). How important is aerial leaf litter for insectivorous birds foraging in a Malaysian tropical forest? Global Ecology and Conservation 20 : e00722 . Google Scholar Crossref Search ADS WorldCat Marín-Gómez , O. H. , and E. Arbeláez-Cortés ( 2015 ). Variation on species composition and richness in mixed bird flocks along an altitudinal gradient in the Central Andes of Colombia . Studies on Neotropical Fauna and Environment 50 : 113 – 129 . Google Scholar Crossref Search ADS WorldCat Marra , P. P. , and J. V. Remsen ( 1997 ). Insights into the maintenance of high species diversity in the neotropics: Habitat selection and foraging behavior in understory birds of tropical and temperate forests. In Studies in Neotropical Ornithology Honoring Ted Parker ( J. V. Remsen, Jr., Editor). Ornithological Monographs 48 : 445 – 483 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Martínez , A. E. , and J. P. Gomez ( 2013 ). Are mixed-species bird flocks stable through two decades? The American Naturalist 181 : E53 – E59 . Google Scholar Crossref Search ADS PubMed WorldCat Martínez , A. E. , E. Parra, O. Muellerklein, and V. T. Vredenburg ( 2018 ). Fear-based niche shifts in neotropical birds . Ecology 99 : 1338 – 1346 . Google Scholar Crossref Search ADS PubMed WorldCat Matthysen , E. , F. Collet, J. Cahill, F. Collet, and J. Cahill ( 2008 ). Mixed flock composition and foraging behavior of insectivorous birds in undisturbed and disturbed fragments of high-Andean Polylepis woodland . Ornitologia Neotropical 19 : 403 – 416 . Google Scholar OpenURL Placeholder Text WorldCat McCain , C. M . ( 2009 ). Global analysis of bird elevational diversity . Global Ecology and Biogeography 18 : 346 – 360 . Google Scholar Crossref Search ADS WorldCat McCune , B. , J. B. Grace, and D. L. Urban ( 2002 ). Analysis of Ecological Communities . MjM Software Design , Gleneden Beach, OR, USA . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC McDermott , M. E. , and A. D. Rodewald ( 2014 ). Conservation value of silvopastures to Neotropical migrants in Andean forest flocks . Biological Conservation 175 : 140 – 147 . Google Scholar Crossref Search ADS WorldCat Mokross , K. , T. B. Ryder, M. C. Côrtes, J. D. Wolfe, and P. C. Stouffer ( 2014 ). Decay of interspecific avian flock networks along a disturbance gradient in Amazonia . Proceedings of the Royal Society of London B: Biological Sciences 281 : 20132599 . Google Scholar OpenURL Placeholder Text WorldCat Montaño-Centellas , F. A . ( 2020 ). Interaction networks of avian mixed-species flocks along elevation in the tropical Andes . Ecography 43 : 930 – 942 . Google Scholar Crossref Search ADS WorldCat Montaño-Centellas , F. A. , and Á. Garitano-Zavala ( 2015 ). Andean bird responses to human disturbances along an elevational gradient . Acta Oecologica 65 – 66 : 51 – 60 . Google Scholar OpenURL Placeholder Text WorldCat Montaño-Centellas , F. A. , and H. H. Jones ( 2021 ). Data from: Temperature and vegetation complexity structure mixed-species flocks along a gradient of elevation in the tropical Andes . Ornithology 138 : 1 – 18 . doi:10.5061/dryad.mpg4f4qzd Google Scholar OpenURL Placeholder Text WorldCat Montaño-Centellas , F. A. , Loiselle , B. A., and M. W. Tingley ( 2020 ). Ecological drivers of avian community assembly along a tropical elevation gradient . Ecography 4 : 1 – 15 . Google Scholar OpenURL Placeholder Text WorldCat Morse , D. H . ( 1970 ). Ecological aspects of some mixed-species foraging flocks of birds . Ecological Monographs 40 : 119 – 168 . Google Scholar Crossref Search ADS WorldCat Munn , C. A . ( 1985 ). Permanent canopy and understorey flocks in Amazonia: Species composition and population density. In Neotropical Ornithology ( P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgley, and F. G. Buckley, Editors). Ornithological Monographs 36: 683 – 712 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Munn , C. A. , and J. W. Terborgh ( 1979 ). Multi-species territoriality in Neotropical foraging flocks . The Condor 81 : 338 – 347 . Google Scholar Crossref Search ADS WorldCat Muñoz , J . ( 2016 ). The role of facilitation in the structure of tropical bird communities: A case study of mixed-species flocks . MSc. thesis, University of British Columbia , Vancouver, BC, Canada . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Oksanen , J. , F. G. Blanchet, R. Kindt, P. Legendre, P. R. Minchin, R. O’Hara, G. L. Simpson, P. Solymos, M. H. H. Stevens, and H. Wagner ( 2018 ). Package ‘vegan’. Community ecology package, version 2.5–3 . https://cran.r-project.org/ Pigot , A. L. , C. H. Trisos, and J. A. Tobias ( 2016 ). Functional traits reveal the expansion and packing of ecological niche space underlying an elevational diversity gradient in passerine birds . Proceedings of the Royal Society B: Biological Sciences 283 : 20152013 . Google Scholar Crossref Search ADS WorldCat Pomara , L. Y. , R. J. Cooper, and L. J. Petit ( 2007 ). Modeling the flocking propensity of passerine birds in two Neotropical habitats . Oecologia 153 : 121 – 133 . Google Scholar Crossref Search ADS PubMed WorldCat Potts , J. R. , K. Mokross, P. C. Stouffer, and M. A. Lewis ( 2014 ). Step selection techniques uncover the environmental predictors of space use patterns in flocks of Amazonian birds . Ecology and Evolution 4 : 4578 – 4588 . Google Scholar Crossref Search ADS PubMed WorldCat Poulsen , B. O . ( 1996 ). Relationships between frequency of mixed-species flocks, weather and insect activity in a montane cloud forest in Ecuador . Ibis 138 : 466 – 470 . Google Scholar Crossref Search ADS WorldCat Remsen , J. V . ( 1985 ). Community organization and ecology of birds of high elevation humid forest of the Bolivian Andes. In Neotropical Ornithology. ( P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgley, and F. G. Buckley, Editors). Ornithological Monographs 36: 733 – 756 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Renjifo , L. M. , G. P. Servat, J. M. Goerck, B. A. Loiselle, and J. G. Blake ( 1997 ). Patterns of species composition and endemism in the northern Neotropics: A case for conservation of montane avifaunas. In Studies in Neotropical Ornithology Honoring Ted Parker ( J. V. Remsen, Jr., Editor). Ornithological Monographs 48: 577 – 594 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Ribera , M. O . ( 1995 ). Aspectos ecológicos, del uso de la tierra y conservación en el parque nacional y área natural de manejo integrado Cotapata . In Caminos de Cotapata ( C. B. , Morales, Editor). Instituto de Ecología, Fundeco, Fonama , La Paz, Bolivia . pp. 1 – 84 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ripley , B. , B. Venables, D. M. Bates, K. Hornik, A. Gebhardt, D. Firth, and M. B. Ripley ( 2018 ). Package ‘MASS’, version 7.3–51.1 . http://www.stats.ox.ac.uk/pub/MASS4/ Sevilla Callejo , M . ( 2010 ). Organización territorial y campesinado en el Parque Nacional y área Natural de Manejo Integrado Cotapata (Bolivia) . Doctorate dissertation, Universidad Autónoma de Madrid , Madrid, Spain . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Sridhar , H. , G. Beauchamp, and K. Shanker ( 2009 ). Why do birds participate in mixed-species foraging flocks? A large-scale synthesis . Animal Behaviour 78 : 337 – 347 . Google Scholar Crossref Search ADS WorldCat Sridhar , H. , and V. Guttal ( 2018 ). Friendship across species borders: Factors that facilitate and constrain heterospecific sociality . Philosophical Transactions of the Royal Society B: Biological Sciences 373 : 20170014 . Google Scholar Crossref Search ADS WorldCat Sridhar , H. , U. Srinivasan, R. A. Askins, J. C. Canales-Delgadillo, C. C. Chen, D. N. Ewert, G. A. Gale, E. Goodale, W. K. Gram, P. J. Hart, et al. ( 2012 ). Positive relationships between association strength and phenotypic similarity characterize the assembly of mixed-species bird flocks worldwide . The American Naturalist 180 : 777 – 790 . Google Scholar Crossref Search ADS PubMed WorldCat Srinivasan, U., R. H. Raza, and S. Quader (2012). Patterns of species participation across multiple mixed-species flock types in a tropical forest in northeastern India. Journal of Natural History 46:2749–2762. Stachowicz , J. J . ( 2001 ). Mutualism, facilitation, and the structure of ecological communities . BioScience 51 : 235 – 246 . Google Scholar Crossref Search ADS WorldCat Supriya , K. , C. S. Moreau, K. Sam, and T. D. Price ( 2019 ). Analysis of tropical and temperate elevational gradients in arthropod abundance . Frontiers of Biogeography 11 : e43104 . Google Scholar Crossref Search ADS WorldCat Ter Braak, C. J. F. (1986). Canonical correspondence analysis: A new eigenvector technique for multivariate direct gradient analysis. Ecology 67:1167–1179. Terborgh , J . ( 1977 ). Bird species diversity on an Andean elevational gradient . Ecology 58 : 1007 – 1019 . Google Scholar Crossref Search ADS WorldCat Tubelis , D. P. , A. Cowling, and C. Donnelley ( 2006 ). Role of mixed-species flocks in the use of adjacent savannas by forest birds in the central Cerrado, Brazil . Austral Ecology 31 : 38 – 41 . Google Scholar Crossref Search ADS WorldCat Valburg , L. K . ( 1992 ). Flocking and frugivory: The effect of social groupings on resource use in the Common Bush-Tanager . The Condor 94 : 358 – 363 . Google Scholar Crossref Search ADS WorldCat Valone , T. J . ( 1989 ). Group foraging, public information, and patch estimation . Oikos 56 : 357 – 363 . Google Scholar Crossref Search ADS WorldCat Walther , B. A . ( 2002 ). Vertical stratification and use of vegetation and light habitats by Neotropical forest birds . Journal für Ornithologie 143 : 64 – 81 . Google Scholar Crossref Search ADS WorldCat Wilman , H. , J. Belmaker, J. Simpson, C. de la Rosa, M. M. Rivadeneira, and W. Jetz ( 2014 ). EltonTraits 1.0: Species-level foraging attributes of the world’s birds and mammals . Ecology 95 : 2027 – 2027 . Google Scholar Crossref Search ADS WorldCat Witter , M. S. , and I. C. Cuthill ( 1993 ). The ecological costs of avian fat storage . Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 340 : 73 – 92 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Zeileis , A. , F. Cribari-Neto, B. Gruen, I. Kosmidis, A. B. Simas, A. V. Rocha, and M. A. Zeileis ( 2016 ). Package ‘betareg’, version 3.3. in R package . https://cran.r-project.org/web/packages/betareg/betareg.pdf Zhang , Q. , R. C. Han, M. Zhang, Z. Huang, and F. Zou ( 2013 ). Linking vegetation structure and bird organization: Response of mixed-species bird flocks to forest succession in subtropical China . Biodiversity and Conservation 22 : 1965 – 1989 . Google Scholar Crossref Search ADS WorldCat Zhou , L. , I. Peabotuwage, H. Gu, D. Jiang, G. Hu, A. Jiang, C. Mammides, M. Zhang, R.-C. Quan, and E. Goodale ( 2019 ). The response of mixed-species bird flocks to anthropogenic disturbance and elevational variation in southwest China . The Condor: Ornithological Applications 121 : 1 – 13 . doi:10.1093/condor/duz028 Google Scholar Crossref Search ADS WorldCat Zou , F. , H. Jones, G. J. Colorado Z., D. Jiang, T.-M. Lee, A. Martínez, K. Sieving, M. Zhang, Q. Zhang, and E. Goodale ( 2018 ). The conservation implications of mixed-species flocking in terrestrial birds, a globally-distributed species interaction network . Biological Conservation 224 : 267 – 276 . Google Scholar Crossref Search ADS WorldCat Copyright © American Ornithological Society 2021. All rights reserved. For permissions, e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Ornithology Oxford University Press

Temperature and vegetation complexity structure mixed-species flocks along a gradient of elevation in the tropical Andes

Ornithology , Volume Advance Article – May 6, 2021

Loading next page...
 
/lp/oxford-university-press/temperature-and-vegetation-complexity-structure-mixed-species-flocks-N2njUlhqbK
Copyright
Copyright © 2021 American Ornithological Society
eISSN
2732-4613
DOI
10.1093/ornithology/ukab027
Publisher site
See Article on Publisher Site

Abstract

Abstract Mixed-species flocks constitute community modules that can help test mechanisms driving changes to community composition across environmental gradients. Here, we examined elevational patterns of flock diversity (species richness, taxonomic diversity, species, and guild composition) and asked if these patterns were reflections of the full bird community at a given elevation (open-membership hypothesis), or if they were instead structured by environmental variables. We surveyed both the overall avian community and mixed-species flocks across an undisturbed elevational gradient (~1,350–3,550 m) in the Bolivian Andes. We then tested for the role of temperature (a surrogate for abiotic stress), resource diversity (arthropods, fruits), and foraging niche diversity (vegetation vertical complexity) in structuring these patterns. Patterns for the overall and flocking communities were similar, supporting our open-membership hypothesis that Andean flocks represent dynamic, unstructured aggregations. Membership openness and the resulting flock composition, however, also varied with elevation in response to temperature and vegetation complexity. We found a mid-elevation peak in flock species richness, size, and Shannon’s diversity at ~2,300 m. The transition of flocking behavior toward a more open-membership system at this elevation may explain a similar peak in the proportion of insectivores joining flocks. At high elevations, increasing abiotic stress and decreasing fruit diversity led more generalist, gregarious tanagers (Thraupidae) to join flocks, resulting in larger yet more even flocks alongside a loss of vegetation structure. At lower elevations, flock species richness increased with greater vegetation complexity, but a greater diversity of foraging niches resulted in flocks that were more segregated into separate canopy and understory sub-types. This segregation likely results from increased costs of interspecific competition and activity matching (i.e., constraints on movement and foraging rate) for insectivores. Mid-elevation flocks (~2,300 m) seemed, therefore, to benefit from both the open-membership composition of high-elevation flocks and the high vegetation complexity of mid- and low-elevation forests. RESUMEN Las bandadas mixtas constituyen módulos comunitarios (i.e., subcomunidades de especies con fuertes asociaciones interespecíficas) que pueden ayudarnos a investigar los mecanismos que determinan cambios en la composición de comunidades a lo largo de gradientes ambientales. En este trabajo, examinamos patrones altitudinales de diversidad de bandadas mixtas de aves (riqueza de especies, diversidad taxonómica y funcional, composición de especies y gremios ecológicos), e investigamos si estos patrones son reflejo de la comunidad de aves en cada elevación (hipótesis de membresía abierta), o si las bandadas son estructuradas por variables ambientales. Evaluamos tanto la comunidad de aves completa como la formación de bandadas mixtas a lo largo de un gradiente altitudinal no disturbado (~1,350–3,550 m) en los Andes de Bolivia. Luego, evaluamos el rol de la temperatura (como proxy de condiciones abióticas) y de la disponibilidad de recursos (artrópodos, frutos y complejidad vertical de la vegetación) como mecanismos que determinan los cambios en diversidad de bandadas. Los patrones altitudinales de la comunidad entera y de la comunidad de aves de bandadas fueron similares, apoyando nuestra hipótesis de “membresía abierta” de que las bandadas de los Andes representan agregaciones dinámicas y no estructuradas. La apertura de esta membresía en bandadas mixtas y la composición de especies resultante, sin embargo, también varió con la altura en respuesta a la temperatura y a la complejidad de la vegetación. Encontramos un pico de riqueza de especies, tamaño, diversidad de Shannon y diversidad funcional por bandada cerca de los 2,300 m. Sugerimos que la transición de comportamiento hacia una membresía abierta en la formación de bandadas a partir de esta altitud permite que una mayor proporción de la comunidad de insectívoros se una a las bandadas. En elevaciones mayores, las bajas temperaturas y la perdida de estructura de vegetación permiten la unión de especies más generalistas y gregarias, especialmente tangaras (Thraupidae), resultando en bandadas con mayor número de individuos, pero más equitativas (i.e., especies con abundancias más homogéneas). A elevaciones menores, con mayor complejidad estructural, una mayor cantidad de nichos de forrajeo resulta en bandadas que son más estructuradas, segregándose en subtipos de dosel y sotobosque. Esta segregación potencialmente resulta de un incremento en los costos de competencia interespecífica y costos de actividad (i.e., limitantes en el movimiento y la tasa de forrajeo) de insectívoros. Las bandadas de elevaciones medias (~2,300), por lo tanto, parecen beneficiarse tanto de la membresía abierta de mandadas de altura, como de la alta complejidad estructural de bosques en alturas medias y bajas. Lay Summary • We investigated changes to the diversity and membership of flocks of birds across elevations in the Bolivian Andes, and the mechanisms driving these changes. • Changes to flock membership largely reflected elevational changes to the whole bird community, suggesting Andean flocks have open membership, unlike the flocks of lowland Amazonia. • Elevational changes to flock richness, size, and membership were mostly associated with effects of temperature (abiotic stress) and the diversity of vegetation layers within the forest. • Temperature and fruit diversity were associated with the number of individuals of each species, with larger flocks at high elevations, composed of generalist, gregarious tanagers. • Increasing numbers of vegetation strata were associated with higher insectivore richness and abundance in flocks, explaining flock characteristics at middle and low elevations. INTRODUCTION Mixed-species flocks are moving foraging aggregations of two or more species (Morse 1970) that constitute “ecological modules” of closely interacting species within a community (sensu Holt 1997). These interspecific groupings are global in distribution, yet become most speciose, and most temporally stable, in tropical forest ecosystems (Zou et al. 2018). The composition and richness of mixed-species flocks are known to be variable in many systems, and to change with elevation (Marín-Gómez and Arbeláez-Cortés 2015, Muñoz 2016, Montaño-Centellas 2020) as well as along human disturbance (Mokross et al. 2014, Zhou et al. 2019, Jones and Robinson 2020a) and successional (Zhang et al. 2013) gradients. It remains unclear, however, whether changes to flock composition reflect changes in the overall bird community or whether they occur independently of each other. In particular, there has been relatively little work on how and why these flocking systems change across elevational gradients in the tropics (Montaño-Centellas 2020). The lack of studies on this topic is surprising because we have long known that avian species richness decreases with increasing elevation (McCain 2009), changing taxonomic and functional composition dramatically. For instance, Neotropical bird assemblages lose insectivores and (proportionally) gain omnivores and frugivores with increasing altitude (Terborgh 1977, Jankowski et al. 2013b). Flock composition and richness patterns in a completely open-membership flock should therefore reflect these underlying changes to the community. Alternatively, structuring mechanisms of flock composition such as preferences for specific flock-mates or leader species (Mammides et al. 2015, Goodale et al. 2017), sharing a foraging stratum as a prerequisite for joining the flock (Sridhar et al. 2012), or within-guild competition (Colorado and Rodewald 2015), could result in richness and composition not reflective of the whole community. Because aggregations of flocking species bring together many species of insectivorous birds, they have long been thought to be characterized both by exploitation and interference competition (Morse 1970, Alatalo et al. 1987, Jabłoński and Don Lee 2002). Indeed, many flock-joining species show foraging niche partitioning by substrate (e.g., Eguchi et al. 1993, Matthysen et al. 2008). Yet mixed-species flocks are also thought to represent major facilitative interactions for forest birds by increasing foraging efficiency and by providing information and protection against predators (Sridhar et al. 2009, Goodale et al. 2020). Individuals that join flocks obtain access to social information about predators from other flock members and from specific, highly vigilant sentinel species, which enables them to exploit more exposed microhabitats (Dolby and Grubb 1999, Darrah and Smith 2013, Martínez et al. 2018). The facilitative effects of flocking might, for instance, extend the realized elevational distributions of flock-joining species (Goodale et al. 2020), and might buffer species responses to unsuitable habitats (Tubelis et al. 2006). In some systems, for instance, the community of flock-joining species changed less across land-use types (Goodale et al. 2014) and with selective logging (Borah et al. 2018) than did the overall community. Given the potential benefits gained (Sridhar et al. 2009), and the costs of competition when joining a flock (Gross 2008, Sridhar et al. 2012), individuals are expected to aggregate with each other in flocks when the benefits outweigh the costs of associating. These benefits will vary among species and be dependent upon habitat context and flock composition (Mokross et al. 2014, Borah et al. 2018), and by consequence flocking assemblages are expected to vary across environmental gradients. Understanding changes to mixed-species flocks across elevational gradients has been hampered by the many simultaneous mechanisms driving change (Zhou et al. 2019). First, increasing elevation entails changing abiotic conditions, notably decreasing mean temperature and barometric pressure, which confine the fundamental niche of tropical species (Jankowski et al. 2013a, Elsen et al. 2017). Because fat storage in birds is costly, low temperatures can impose an important tradeoff between increased starvation and predation risks (Lima 1986, Witter and Cuthill 1993). Flocking behavior can therefore be more prevalent at colder temperatures (Greenberg 2000, Gentry et al. 2019) because greater foraging benefits can potentially be gained from facilitation in extreme abiotic conditions (Callaway 1998, Stachowicz 2001). Thus, more species and individuals might actively join flocks at high elevations, where conditions are harsher. Second, the availability of food resources, in particular insects and fruits, is thought to decline with elevation (Terborgh 1977), perhaps in conjunction with decreasing productivity (Fisher et al. 2013). These patterns, however, can be highly variable (Supriya et al. 2019). Reduced access to insect resources at either seasonal (Develey and Peres 2000, Mangini and Areta 2018) or daily (Poulsen 1996) time scales increases participation in flocks, and a similar effect may occur at higher elevations. Third, vertical vegetation complexity, highly correlated with the number of available foraging niches (e.g., MacArthur and MacArthur 1961), declines with increasing elevation (Terborgh 1977, Acharya et al. 2011, Jankowski et al. 2013b), reducing the available foraging niche space. Mixed-species flocks are known to preferentially select for habitat with greater vegetation complexity (McDermott and Rodewald 2014, Potts et al. 2014) and have greater species richness in such microhabitats (Knowlton and Graham 2011, Zhang et al. 2013, Jones and Robinson 2020a). This is particularly true for flock-following insectivores, which are often highly specialized by substrate (Marra and Remsen 1997, Jones et al. 2020) and foraging height (Walther 2002, Mansor et al. 2019). How these mechanisms act upon flock structure has not yet been evaluated over a full elevational gradient. Patterns of flock richness and composition are of great interest in the Andes because Neotropical montane and lowland flocking systems differ in structure and behavior. For one, lowland flocking associations in primary forest consist of temporally stable species interactions (Martínez and Gomez 2013). In Amazonia, multiple species pairs defend a collective territory (Munn and Terborgh 1979) and flock composition is highly structured (Graves and Gotelli 1993). Andean flocking systems, by contrast, often show a dynamic, open-membership composition, with species frequently joining and leaving the flock as it enters and leaves their territories (Poulsen 1996, Pomara et al. 2007), and a much greater diversity of associating species pairs (Montaño-Centellas 2020, Jones and Robinson 2020b). Similarly, lowland flocks are characterized by multiple sub-types (Munn 1985), typically differentiated into understory and canopy flocks. Andean flocks instead often show a single homogenous flock composition, with both understory and canopy species freely admixed (Poulsen 1996, Guevara et al. 2011, Colorado and Rodewald 2015) and a lack of differentiable sub-types in network analyses (Montaño-Centellas 2020). It is therefore important to understand how these profound changes to flocking behavior affect flock composition and the diversity of participating species. In this study, we first described elevational patterns of avian richness, taxonomic and guild diversity, and composition in the tropical Andes of Bolivia. We then tested for potential mechanisms driving these elevational patterns. Specifically, we sampled flock composition, the community of flock-joining species, and the full avian community across a continuous elevational gradient of undisturbed forest to examine: (1) the extent to which the richness and composition of the flocking community are reflective of the overall bird community at each elevation, and (2) how the proportion of species of specific taxa and foraging guilds in flocks changes with elevation. Then, we explored (3) how flock species richness, size, and composition vary with elevation and (4) related these flock response variables to elevational changes in temperature (as a surrogate for abiotic stress), fruit and arthropod resource availability, and vertical vegetation complexity (a measure of available foraging strata). Each of these variables represents a mechanistic hypothesis explaining structuring forces across the elevational gradient (see Table 1 for detailed predictions). Alternatively, open-membership flocks could lack structuring mechanisms and instead reflect the greater bird community. Thus, we predicted that flocking species richness would decrease with elevation, as expected for the whole community. However, because facilitation is potentially a more important structuring force when conditions are harsher (Montaño-Centellas 2020), we expected a greater percentage of the community to join mixed-species flocks at high elevations. Similarly, we expected that the transition from separate understory and canopy flock types to a single flocking system should entail an increase in the overall number of flocking species at higher elevations. TABLE 1. Hypotheses pertaining to the structure of mixed-species flocks across elevations in the Andes of Bolivia. The name and mechanism explaining how the different drivers tested in this study operate upon flocking assemblages, the ecological outcome of each mechanism, the predicted patterns for overall changes to mixed-species flocks (MSF-predicted patterns), and for specific functional groups (guild-specific patterns) are presented. Hypothesis . Mechanism . Ecological outcome . MSF-predicted pattern . Guild-specific patterns . Open-membership hypothesis Abiotic and biotic filtering act upon avian assemblages shaping community structure. Flocking assemblages are structured by community assembly forces and therefore changes in richness and composition are reflective of those in the whole community. Species richness and taxonomic and functional composition mirror changes to the whole bird community. Proportionally, the number of omnivore and frugivore species and individuals increase with elevation, while the proportion of insectivores declines. Abiotic stress hypothesis Harsher abiotic conditions impose greater stress on individuals; thus, greater foraging benefits can be obtained by facilitating in extreme abiotic conditions. Flocking behavior is more prevalent where abiotic stress is higher, thus more species and individuals join flocks in extreme abiotic conditions. Higher proportion of the community joining mixed-species flocks at higher elevations with lower temperatures. Greater flock size with decreasing temperature. Greater richness and abundance of omnivores in high-elevation flocks, and a decline in the number of species and individuals of specialized taxa and functional guilds (e.g., insectivores, frugivores, canopy and ground foragers). Food resource availability hypothesis Reduced availability/diversity of food resources (insects and fruits) increases participation in flocks. Facilitation benefits related to foraging efficiency might be higher in areas where resources are depleted and/or unavailable. Greater species richness and number of individuals per flock with decreasing resource availability. Species richness and number of individuals of insectivores increase with decreasing arthropod diversity. Species richness and number of individuals of frugivores increase with decreasing fruit diversity. Foraging niche hypothesis Increased number of available foraging niches allows for finer partitioning of foraging strata and microhabitats among members of mixed-species flocks, reducing inter- and intraspecific competition. Greater vegetation complexity harbors larger and more speciose mixed-species flocks. Greater species richness and number of individuals per flock with increasing vegetation complexity. Greater richness and abundance of specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) with increasing vegetation complexity. Greater richness and abundance of generalist guilds (e.g., omnivores) with decreasing complexity. Hypothesis . Mechanism . Ecological outcome . MSF-predicted pattern . Guild-specific patterns . Open-membership hypothesis Abiotic and biotic filtering act upon avian assemblages shaping community structure. Flocking assemblages are structured by community assembly forces and therefore changes in richness and composition are reflective of those in the whole community. Species richness and taxonomic and functional composition mirror changes to the whole bird community. Proportionally, the number of omnivore and frugivore species and individuals increase with elevation, while the proportion of insectivores declines. Abiotic stress hypothesis Harsher abiotic conditions impose greater stress on individuals; thus, greater foraging benefits can be obtained by facilitating in extreme abiotic conditions. Flocking behavior is more prevalent where abiotic stress is higher, thus more species and individuals join flocks in extreme abiotic conditions. Higher proportion of the community joining mixed-species flocks at higher elevations with lower temperatures. Greater flock size with decreasing temperature. Greater richness and abundance of omnivores in high-elevation flocks, and a decline in the number of species and individuals of specialized taxa and functional guilds (e.g., insectivores, frugivores, canopy and ground foragers). Food resource availability hypothesis Reduced availability/diversity of food resources (insects and fruits) increases participation in flocks. Facilitation benefits related to foraging efficiency might be higher in areas where resources are depleted and/or unavailable. Greater species richness and number of individuals per flock with decreasing resource availability. Species richness and number of individuals of insectivores increase with decreasing arthropod diversity. Species richness and number of individuals of frugivores increase with decreasing fruit diversity. Foraging niche hypothesis Increased number of available foraging niches allows for finer partitioning of foraging strata and microhabitats among members of mixed-species flocks, reducing inter- and intraspecific competition. Greater vegetation complexity harbors larger and more speciose mixed-species flocks. Greater species richness and number of individuals per flock with increasing vegetation complexity. Greater richness and abundance of specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) with increasing vegetation complexity. Greater richness and abundance of generalist guilds (e.g., omnivores) with decreasing complexity. Open in new tab TABLE 1. Hypotheses pertaining to the structure of mixed-species flocks across elevations in the Andes of Bolivia. The name and mechanism explaining how the different drivers tested in this study operate upon flocking assemblages, the ecological outcome of each mechanism, the predicted patterns for overall changes to mixed-species flocks (MSF-predicted patterns), and for specific functional groups (guild-specific patterns) are presented. Hypothesis . Mechanism . Ecological outcome . MSF-predicted pattern . Guild-specific patterns . Open-membership hypothesis Abiotic and biotic filtering act upon avian assemblages shaping community structure. Flocking assemblages are structured by community assembly forces and therefore changes in richness and composition are reflective of those in the whole community. Species richness and taxonomic and functional composition mirror changes to the whole bird community. Proportionally, the number of omnivore and frugivore species and individuals increase with elevation, while the proportion of insectivores declines. Abiotic stress hypothesis Harsher abiotic conditions impose greater stress on individuals; thus, greater foraging benefits can be obtained by facilitating in extreme abiotic conditions. Flocking behavior is more prevalent where abiotic stress is higher, thus more species and individuals join flocks in extreme abiotic conditions. Higher proportion of the community joining mixed-species flocks at higher elevations with lower temperatures. Greater flock size with decreasing temperature. Greater richness and abundance of omnivores in high-elevation flocks, and a decline in the number of species and individuals of specialized taxa and functional guilds (e.g., insectivores, frugivores, canopy and ground foragers). Food resource availability hypothesis Reduced availability/diversity of food resources (insects and fruits) increases participation in flocks. Facilitation benefits related to foraging efficiency might be higher in areas where resources are depleted and/or unavailable. Greater species richness and number of individuals per flock with decreasing resource availability. Species richness and number of individuals of insectivores increase with decreasing arthropod diversity. Species richness and number of individuals of frugivores increase with decreasing fruit diversity. Foraging niche hypothesis Increased number of available foraging niches allows for finer partitioning of foraging strata and microhabitats among members of mixed-species flocks, reducing inter- and intraspecific competition. Greater vegetation complexity harbors larger and more speciose mixed-species flocks. Greater species richness and number of individuals per flock with increasing vegetation complexity. Greater richness and abundance of specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) with increasing vegetation complexity. Greater richness and abundance of generalist guilds (e.g., omnivores) with decreasing complexity. Hypothesis . Mechanism . Ecological outcome . MSF-predicted pattern . Guild-specific patterns . Open-membership hypothesis Abiotic and biotic filtering act upon avian assemblages shaping community structure. Flocking assemblages are structured by community assembly forces and therefore changes in richness and composition are reflective of those in the whole community. Species richness and taxonomic and functional composition mirror changes to the whole bird community. Proportionally, the number of omnivore and frugivore species and individuals increase with elevation, while the proportion of insectivores declines. Abiotic stress hypothesis Harsher abiotic conditions impose greater stress on individuals; thus, greater foraging benefits can be obtained by facilitating in extreme abiotic conditions. Flocking behavior is more prevalent where abiotic stress is higher, thus more species and individuals join flocks in extreme abiotic conditions. Higher proportion of the community joining mixed-species flocks at higher elevations with lower temperatures. Greater flock size with decreasing temperature. Greater richness and abundance of omnivores in high-elevation flocks, and a decline in the number of species and individuals of specialized taxa and functional guilds (e.g., insectivores, frugivores, canopy and ground foragers). Food resource availability hypothesis Reduced availability/diversity of food resources (insects and fruits) increases participation in flocks. Facilitation benefits related to foraging efficiency might be higher in areas where resources are depleted and/or unavailable. Greater species richness and number of individuals per flock with decreasing resource availability. Species richness and number of individuals of insectivores increase with decreasing arthropod diversity. Species richness and number of individuals of frugivores increase with decreasing fruit diversity. Foraging niche hypothesis Increased number of available foraging niches allows for finer partitioning of foraging strata and microhabitats among members of mixed-species flocks, reducing inter- and intraspecific competition. Greater vegetation complexity harbors larger and more speciose mixed-species flocks. Greater species richness and number of individuals per flock with increasing vegetation complexity. Greater richness and abundance of specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) with increasing vegetation complexity. Greater richness and abundance of generalist guilds (e.g., omnivores) with decreasing complexity. Open in new tab Finally, although we expected elevational patterns in flock characteristics to be influenced by all three mechanisms (abiotic stress, food resource availability, and the availability of foraging niches), we predicted different responses for different subsets of flocking species (Table 1). Specifically, we expected insectivore richness to decline at higher elevations due to a loss of lowland-associated taxa in favor of dominant bird families of the highlands (Thraupidae). In particular, we predicted that declining vegetation complexity would reduce the available foraging niches for specialists. However, we expected insectivore participation to increase with decreasing arthropod availability whereas frugivore richness and abundance per flock were predicted to increase with a decline in fruit availability. Accordingly, because milder conditions and greater vegetation complexity, likely associated with a greater foraging niche space at lower elevations might allow communities to contain more specialist species with smaller niches, we expected the number of species in generalist guilds (e.g., omnivores) participating in flocks to increase with elevation. By contrast, the richness and abundance of species from specialist guilds (e.g., insectivores, frugivores, canopy, and ground foragers) participating in flocks should be higher at lower elevations. A summary of the hypotheses and proposed mechanisms driving elevational patterns is presented in Table 1. METHODS Mixed-Species Flock Surveys Avian mixed-species flocks were surveyed for two consecutive years, from May to October 2016 and May to August 2017, along a continuous elevational gradient (~1,350–3,550 m.a.s.l. [meters above sea level]) in Cotapata National Park, a protected area in the Andes of western Bolivia. Our focal gradient encompasses all forested habitats within the park, from the treeline to the valley bottom (Montaño-Centellas 2020), passing through a landscape dominated by evergreen humid montane and cloud forests on steep slopes (Ribera 1995, Sevilla Callejo 2010). The extreme topography in the study region prevents large human settlements (Montaño-Centellas and Garitano-Zavala 2015), although there is a small village (~10 families) at 1,950 m with small orchards scattered nearby. To eliminate the potential confounding effects of this village, all flocks observed between 1,700 and 2,100 m were not considered in the analyses. The gradient was sampled 16 times (8 times per year) through randomized 350-m elevation surveys, so that elevations were not sampled sequentially to avoid temporal biases. Further details on the study system and survey methods can be found in Montaño-Centellas et al. (2020). Briefly, each survey consisted of walking at a slow constant pace along the transect and for each flock encountered, recording GPS coordinates, elevation, and the number and identity of all participating species. In this study, we defined a flock as a gathering of at least two species foraging <10 m apart while moving in the same direction. Flock data were collected using the “gambit of the group” method, where several individuals observed at the same time and place are assumed to be associating (Farine and Whitehead 2015). To diminish the bias of registering non-flocking individuals that increased their foraging activity while the flock was present, we recorded the whole encounter until the flock was unreachable and excluded all species that remained in the study areas (feeding, perching, and/or vocalizing) after the flock had left. In our study, each detected flock represents a snapshot of the realized associations resulting from individual birds deciding whether to join the flock, and thus, flocks observed on different days were considered independent replicates of flock composition and richness. Species richness is known to be correlated with observation time; thus, we further standardized our effort by including only flocks that were observed for between 10 and 20 min. Data were organized into abundance matrices, with flocks as columns and species as rows. Percentage of the Community That Participates in Mixed-Species Flocks We determined the proportion of the total community participating in mixed-species flocks by comparing whole-community survey data to flock composition data. Community surveys were conducted between April and August 2014 and April and October 2015 (Montaño-Centellas et al. 2020). Given the extreme topography of the landscape, we followed an approximate line-transect methodology of surveying birds, adapted from similar work in other montane areas (Elsen et al. 2017). Complete details on survey methodology are presented in Montaño-Centellas et al. (2020). Briefly, surveys consisted of walking at a slow pace along 350-m elevational subsections across the gradient and registering every encountered bird (visual and auditory detections). For each detected individual, GPS coordinates and elevation were noted, as well as taxonomic identity. The complete gradient was surveyed 20 times (10 per year), randomizing the order of visits to different elevations. We summarized this information to obtain the composition of the full bird community at each elevation by pooling all species observed in 100-m-wide elevational bands (no interpolation was performed). We similarly determined the flock-joining community in the same elevational bands by grouping together all species observed in mixed-species flocks within each elevational band. For each elevational band, we then calculated the proportion of the overall community and the proportion of species from a given taxon (Thraupidae, Furnariidae, and Tyrannidae), foraging guild (insectivores, frugivores, and omnivores), and foraging stratum (canopy, mid-height, understory, and ground) that participated in flocks. We visually examined how the percentage of the community that joined mixed-species flocks changed with elevation for the whole community and for each community subset. Predictors of Flock Diversity and Composition For each observed flock, we quantified the species richness, a number of individuals participating, species diversity using Shannon’s index H’ and species evenness (i.e., how equally distributed species abundances were within a flock) using Pielou’s index J’ (Magurran 2004). We also calculated the number of species and individuals in specific taxa (Thraupidae, Furnariidae, Tyrannidae), foraging guilds (insectivores, frugivores, omnivores), and vertical strata (canopy, mid-height, understory, ground foragers). Information on diet and foraging stratum was extracted from Wilman et al. (2014) and modified based on field observations. Species richness and diversity were calculated with functions of the package vegan (Oksanen et al. 2018). We examined the effects of four environmental variables that might drive differences in the diversity and composition of mixed-species flocks across elevations: mean daily temperature (highly correlated with elevation, and a surrogate for abiotic conditions), vegetation complexity, arthropod diversity, and fruit availability. Temperature was measured hourly from April to October 2015 at five elevations (1,350, 2,000, 2,500, 3,000, and 3,500 m) using permanent stations equipped with a datalogger (EasyLog EL-USB-2-LCD) and summarized as mean daily temperature at each elevation. We then fitted a linear regression to these values and used the regression model (F1,3 = 41.16, R2 = 0.93, P < 0.001) to predict mean temperature values for the elevation of each flock. Vegetation complexity was measured each 100-m of elevation (e.g., 3,550, 3,450, 3,350, etc; average linear distance within each 100-m elevational band was 1,100 m), using a modified version of the understory height diversity index (UHD) used by MacArthur and MacArthur (1961). At each elevation, we selected 10 random points separated by at least 25 m from one another. At each point, we set an imaginary 1-m diameter cylinder and noted the presence or absence of vegetation in 7 height intervals: 0–1 m, 1–2 m, 2–4 m, 4–8 m, 8–16 m, 16–24 m, and >24 m. We used data from the 10 cylinders to calculate the proportion of times vegetation was present in each height interval, for each elevation. UHD was then calculated as the Shannon diversity index H’ = Ʃ pi ln(pi), where pi is the proportion of vegetation in the ith height interval (Montaño-Centellas et al. 2020). Arthropod and fruit availability were measured between June and July 2016 for each 100 m of elevation across the gradient. At each elevation, arthropods were sampled at 4 points separated by at least 50 m from one another, using a beating method that consisted of placing a 1-m2 cloth supported with a frame under the vegetation and shaking or beating the vegetation to collect the arthropods that dislodge (Cooper and Whitmore 1990). We collected arthropods by shaking the vegetation at two different heights at the same time (1 m and 3 m), using the same intensity of beating. Collected arthropods were manually processed, preserved, and identified by researchers associated with the Universidad Mayor de San Andres (Montaño-Centellas et al. 2020). Arthropod availability was then calculated as the Shannon’s diversity index H’ of morphospecies of arthropods captured at each elevation. Fruit availability was defined as the diversity of ornithochoric (i.e., bird-dispersed) plants with ripe fruits within a circular plot of ~100-m radius located at each elevation. Plants were surveyed with binoculars by two trained botanists by walking within each plot for a period of 3–4 hr. Both the taxonomic identity and the abundance of each fruiting plant species were noted. When possible, voucher specimens were collected to help species identification in the Herbario Nacional de Bolivia (Montaño-Centellas et al. 2020). Data on species identity and abundance were used to calculate the Shannon’s diversity index of ornithochoric species at each elevation. Because we were interested in the diversity of available resources instead of the net amount, we quantified abundance with numbers instead of biomass and chose a diversity metric to describe them. However, resource abundance and resource diversity are likely positively correlated (Abrams 1995, Hanz et al. 2019), although this is not necessarily true for every system (Hails 1982). We tested for redundancy among predictors with Pearson’s product-moment correlation and found overall low values (Supplementary Material Figure S1), except between temperature and fruit diversity (r = 0.72, P < 0.001). We thus regressed fruits as a function of temperature, extracted the regression residuals, and used these values as a new variable to describe relative fruit availability for a given elevation (we used this new variable in all analyses). Statistical Analyses We first explored elevational patterns of each of our response variables: species richness, number of individuals, taxonomic diversity, and species evenness per flock using generalized linear models (GLMs), with elevation and elevation2 as independent variables. For count data (species richness and a number of individuals per flock), we used quasi-Poisson errors to account for overdispersion of the data. For Shannon’s diversity index, we assumed Gamma-distributed errors and an identity link, as data were continuous but restricted to positive values; for Pielou’s evenness index, with values ranging between 0 and 1, we assumed a beta distribution of errors. Secondly, we tested for the effects of 4 environmental predictors (mean daily temperature, arthropod and fruit availability, and vegetation complexity) on each of our response variables. We used GLMs, assuming error distributions as described above, with our predictors as independent variables. In accordance with our hypotheses (see Table 1), models for families Furnariidae and Tyrannidae, and for insectivores, included temperature, arthropod diversity, and vegetation complexity as predictors; models for frugivores used temperature, fruit diversity, and vegetation complexity; all other models included all four predictors. We used an information-theoretic approach to evaluate model goodness-of-fit. Models with continuous response variables were ranked using Akaike’s information criterion (AIC) and models with count response variables with quasi-AIC (QAIC; Burnham and Anderson 2004). In every case, competing models had goodness-of-fit that was not far from that of the best model (ΔAIC < 2); thus, we followed Burnham and Anderson (2004) to perform model averaging across competing models for that response variable. We report the averaged model coefficients and predictor significance based on its 95% confidence intervals (CI) (Burnham and Anderson 2004). Regressions were performed in packages base, MASS 7.3 (Ripley et al. 2018) and betareg 3.0 (Zeileis et al. 2016) while model fit, AIC and QAIC calculations, and model averaging were performed using functions of the packages MuMIn 1.43.15 (Barton 2019) and AICcmodavg 2.2-2. In order to understand how our predictor variables influenced flock species composition, we further ran a canonical correspondence analysis (CCA; Ter Braak 1986) using vegan 2.5–6 (Oksanen et al. 2018) and ade4 1.7–15 packages. Species abundances were log(x + 1) transformed and environmental variables were scaled prior to analysis. We examined linear dependencies by computing each environmental variable’s variance inflation factor (VIF) to examine for collinearity among predictors. Values of VIF above 10 suggest collinearity and should be avoided (McCune et al. 2002, Borcard et al. 2018). All calculated values were less than two (VIF = 1.68, 1.17, 1.99, 1.44 for temperature, arthropod availability, vegetation complexity, and fruit availability, respectively), suggesting no collinearity among predictors, and we thus retained all four in our canonical model. RESULTS Percentage of the Community That Participates in Mixed-Species Flocks across Elevations The avian community along the gradient was composed of 300 species (excluding flyovers); 162 of these species participated in flocks, encompassing ~60% of the bird families observed in the full community (Supplementary Material Table S17). We observed a total of 368 mixed-species flocks along our studied gradient. The most speciose families participating in flocks across elevations were Thraupidae, Tyrannidae, and Furnariidae (with 57, 40, and 19 species, respectively). The most frequently observed species were Anisognathus igniventris, Basileuterus luteoviridis, Diglossa cyanea, Myioborus melanocephalus, Chlorospingus ophthalmicus, B. signatus, and B. tristriatus, accounting for almost 36% of all individuals observed. Although not flock-following species (and thus not included in our analyses), some species were nevertheless frequently seen actively vocalizing and foraging while the flock visited their territories. Such species included eight hummingbirds (Adelomyia melanogenys, Aglaiocercus kingii, Coeligena violifer, C. inca, Heliangelus amethisticollis, Metallura aeneocauda, M. tyrianthina, and Phaethornis hispidus), two antpittas (Grallaria erythrotis and G. rufula), two cotingas (Rupicola peruviana and Lipaugus uropygialis), one wren (Henicorrhina leucophrys) and a toucan (Andigena cucullata). Additionally, we observed squirrels (Sciurus spp.) actively following the flock while feeding and vocalizing at three different elevations (1,430, 1,550, and 2,350 m). The number of species joining flocks declined with elevation, mirroring richness decreases for the whole community (Figure 1A). An average of ~54% of species in the full bird community participated in mixed-species flocks across elevations, with a larger percentage of species joining flocks at lower elevations (Figure 1B). There was a steep decrease in the number of species of thraupids and tyrannids in the flocking community with increasing elevation (Figure 2B and C), but the proportion of thraupids in the flocking community that join flocks increased sharply at higher elevations (Supplementary Material Figure S2C). Regardless of elevation, most of the flock-joining species were insectivores, and similar numbers of frugivores and omnivores participated in flocks across elevations (Supplementary Material Figure S2F). Though the number of flocking insectivores decreased with elevation (Figure 2D), they represented a higher proportion of the flocking community at higher elevations (Supplementary Material Figure S2F). Contrary to our expectations, the species richness of omnivores in the flocking community, and the percentage of the overall omnivore community they represented, was lower at high elevations (Supplementary Material Figure S2D and E). Although the number of frugivores in the flocking community decreased with elevation (Figure 2E), they represented a higher percentage of available frugivores at high elevations (Supplementary Material Figure S2E). FIGURE 1. Open in new tabDownload slide (A) Number of species in the overall bird community and participating in mixed-species flocks across elevations and (B) the percentage of the whole community that participating species represent. Lines correspond to generalized additive models (GAM) presented to better examine the trend. FIGURE 1. Open in new tabDownload slide (A) Number of species in the overall bird community and participating in mixed-species flocks across elevations and (B) the percentage of the whole community that participating species represent. Lines correspond to generalized additive models (GAM) presented to better examine the trend. FIGURE 2. Open in new tabDownload slide Number of species in the bird community (in black) and number of species that join mixed-species flocks (in gray) across elevations in the Tropical Andes of Bolivia. Species from different families (A–C), different feeding guilds (D–F), and different foraging strata guilds (G–J) are presented separately. Lines correspond to generalized additive models (GAM) presented to better examine the trend. FIGURE 2. Open in new tabDownload slide Number of species in the bird community (in black) and number of species that join mixed-species flocks (in gray) across elevations in the Tropical Andes of Bolivia. Species from different families (A–C), different feeding guilds (D–F), and different foraging strata guilds (G–J) are presented separately. Lines correspond to generalized additive models (GAM) presented to better examine the trend. As expected, the number of canopy species in flocks decreased with elevation with fewer than 5 canopy specialists in flocks above 2,800 m (Figure 2G). These numbers, however, represented a high proportion of the available canopy specialists in the community, with virtually all canopy specialists above 2,800 m participating in flocks (Supplementary Material Figure S2H). Most of the participating species in flocks foraged in the mid-height stratum, representing the greatest proportion of the flocking community at middle elevations (Supplementary Material Figure S2I). The number and proportion of mid-height-foraging species declined at higher elevations, however (Figure 2H and Supplementary Material Figure S2G). By contrast, the number and proportion of participating understory species in flocks were relatively constant across elevations (Figure 2I), though these numbers represented a higher proportion of the whole flocking community at higher elevations (Supplementary Material Figure S2H). Finally, ground-foraging species participation in flocks increased in number at higher elevations (Figure 2J) and represented a greater percentage of the available ground foragers in cloud forest communities (~2,800 m; Supplementary Material Figure S2H). Elevational Patterns of Species Diversity per Flock and Flock Size Species richness per flock varied from 2 to 24 (median = 8). Similarly, a number of individuals per flock ranged from 3 to 103 (median = 21). Both species richness and the number of individuals per flock changed across elevations, showing a peak in the middle of the gradient at ~2,300 m (Figure 3A and B; Supplementary Material Table S1). Species diversity per flock (measured as Shannon’s index H’) ranged from 0.28 to 2.93 (mean = 1.77), whereas species evenness (measured as Pielou’s evenness index) ranged from 0.4 to 1 (mean = 0.88). Both species diversity and evenness of flock composition changed across elevations, with the highest diversity values observed in the middle of the gradient at ~2,500 m, and the greatest species evenness at high elevations (Figure 3C and D; Supplementary Material Table S1). FIGURE 3. Open in new tabDownload slide (A) Species richness, (B) number of individuals, (C) Shannon’s diversity index H’, and (D) Pielou’s species evenness index J’, for mixed-species flocks along an elevational gradient in the Bolivian Andes. In each panel, points represent an independent flock (n = 368). Lines represent the quadratic regressions of each variable as a function of elevation; models are presented in Supplementary Material Table S1. FIGURE 3. Open in new tabDownload slide (A) Species richness, (B) number of individuals, (C) Shannon’s diversity index H’, and (D) Pielou’s species evenness index J’, for mixed-species flocks along an elevational gradient in the Bolivian Andes. In each panel, points represent an independent flock (n = 368). Lines represent the quadratic regressions of each variable as a function of elevation; models are presented in Supplementary Material Table S1. Predictors of Flock Diversity and Composition We found a positive, yet weak, the effect of vegetation complexity on species richness and a number of individuals per flock, with larger and more speciose flocks in areas with high vegetation complexity (Table 2; see Supplementary Material Table S2 for AIC tables). Whereas temperature had no effect on species richness per flock, it had a negative effect on flock size, with greater numbers of individuals per flock in cold, high-elevation flocks. Although none of our predictors explained overall species diversity (Shannon’s H’), species evenness significantly increased with increasing vegetation complexity and diversity of fruiting resources (Table 2). TABLE 2. Parameter estimation for all predictors in averaged generalized linear models (GLMs) for species richness, number of individuals, Shannon’s diversity index (H’), and species evenness (Pielou’s index J’) of mixed-species flocks in the Andes of Bolivia. The GLM for species richness and number of individuals assumed quasi-Poisson errors, the GLMs for Shannon’s diversity assumed Gaussian errors, and the GLM for Pielou’s J’ followed a beta distribution. Variables with model-averaged confidence intervals that overlap zero were considered not significant for the model. Variables marked with an asterisk are considered supported. AIC tables for all models averaged are presented in Supplementary Material Tables S2–S5. Model . Parameter . Unconditional SE . 95% Confidence interval . . Richness Temperature –0.02 0.02 –0.07 0.03 Vegetation complexity 0.05 0.02 0.01 0.10* Fruits –0.02 0.03 –0.08 0.04 Arthropods 0.00 0.02 –0.04 0.03 Number of individuals Temperature –0.10 0.01 –0.13 –0.07* Vegetation complexity 0.05 0.02 0.01 0.08* Fruits –0.04 0.02 –0.08 0.00 Arthropods 0.01 0.01 –0.02 0.03 Shannon’s H’ Temperature –0.02 0.03 –0.09 0.04 Vegetation complexity 0.03 0.03 –0.03 0.09 Fruits –0.01 0.04 –0.09 0.07 Arthropods 0.01 0.03 –0.05 0.06 Pielou’s J’ Temperature 0.05 0.06 –0.07 0.17 Vegetation complexity 0.13 0.06 0.02 0.25* Fruits 0.35 0.09 0.17 0.53* Arthropods –0.08 0.05 –0.19 0.02 Model . Parameter . Unconditional SE . 95% Confidence interval . . Richness Temperature –0.02 0.02 –0.07 0.03 Vegetation complexity 0.05 0.02 0.01 0.10* Fruits –0.02 0.03 –0.08 0.04 Arthropods 0.00 0.02 –0.04 0.03 Number of individuals Temperature –0.10 0.01 –0.13 –0.07* Vegetation complexity 0.05 0.02 0.01 0.08* Fruits –0.04 0.02 –0.08 0.00 Arthropods 0.01 0.01 –0.02 0.03 Shannon’s H’ Temperature –0.02 0.03 –0.09 0.04 Vegetation complexity 0.03 0.03 –0.03 0.09 Fruits –0.01 0.04 –0.09 0.07 Arthropods 0.01 0.03 –0.05 0.06 Pielou’s J’ Temperature 0.05 0.06 –0.07 0.17 Vegetation complexity 0.13 0.06 0.02 0.25* Fruits 0.35 0.09 0.17 0.53* Arthropods –0.08 0.05 –0.19 0.02 Open in new tab TABLE 2. Parameter estimation for all predictors in averaged generalized linear models (GLMs) for species richness, number of individuals, Shannon’s diversity index (H’), and species evenness (Pielou’s index J’) of mixed-species flocks in the Andes of Bolivia. The GLM for species richness and number of individuals assumed quasi-Poisson errors, the GLMs for Shannon’s diversity assumed Gaussian errors, and the GLM for Pielou’s J’ followed a beta distribution. Variables with model-averaged confidence intervals that overlap zero were considered not significant for the model. Variables marked with an asterisk are considered supported. AIC tables for all models averaged are presented in Supplementary Material Tables S2–S5. Model . Parameter . Unconditional SE . 95% Confidence interval . . Richness Temperature –0.02 0.02 –0.07 0.03 Vegetation complexity 0.05 0.02 0.01 0.10* Fruits –0.02 0.03 –0.08 0.04 Arthropods 0.00 0.02 –0.04 0.03 Number of individuals Temperature –0.10 0.01 –0.13 –0.07* Vegetation complexity 0.05 0.02 0.01 0.08* Fruits –0.04 0.02 –0.08 0.00 Arthropods 0.01 0.01 –0.02 0.03 Shannon’s H’ Temperature –0.02 0.03 –0.09 0.04 Vegetation complexity 0.03 0.03 –0.03 0.09 Fruits –0.01 0.04 –0.09 0.07 Arthropods 0.01 0.03 –0.05 0.06 Pielou’s J’ Temperature 0.05 0.06 –0.07 0.17 Vegetation complexity 0.13 0.06 0.02 0.25* Fruits 0.35 0.09 0.17 0.53* Arthropods –0.08 0.05 –0.19 0.02 Model . Parameter . Unconditional SE . 95% Confidence interval . . Richness Temperature –0.02 0.02 –0.07 0.03 Vegetation complexity 0.05 0.02 0.01 0.10* Fruits –0.02 0.03 –0.08 0.04 Arthropods 0.00 0.02 –0.04 0.03 Number of individuals Temperature –0.10 0.01 –0.13 –0.07* Vegetation complexity 0.05 0.02 0.01 0.08* Fruits –0.04 0.02 –0.08 0.00 Arthropods 0.01 0.01 –0.02 0.03 Shannon’s H’ Temperature –0.02 0.03 –0.09 0.04 Vegetation complexity 0.03 0.03 –0.03 0.09 Fruits –0.01 0.04 –0.09 0.07 Arthropods 0.01 0.03 –0.05 0.06 Pielou’s J’ Temperature 0.05 0.06 –0.07 0.17 Vegetation complexity 0.13 0.06 0.02 0.25* Fruits 0.35 0.09 0.17 0.53* Arthropods –0.08 0.05 –0.19 0.02 Open in new tab As predicted by the abiotic stress hypothesis, temperature was the most important predictor for the number of species and individuals of several taxa and guilds (Table 3; see Supplementary Material Tables S2–S15 for AIC tables). When examining responses of different bird families, we found a negative effect of temperature on thraupid species richness per flock and on the number of individuals of thraupids and furnariids per flock. However, increasing temperature instead had a positive effect on the number of species and individuals of the family Tyrannidae. When examining foraging guild responses, we found a negative effect of increasing temperature on the number of omnivorous individuals, but not species, supporting our prediction of more generalists joining flocks upslope. Temperature also had a negative effect on the number of insectivore individuals, with higher abundances of insectivores at colder, higher elevations. Finally, temperature had a positive effect on the number of canopy-specialist species, with higher species richness at lower elevations, but a negative effect on the number of ground-foraging species and individuals and the number of mid-height foraging individuals. TABLE 3. Parameter estimation for predictors in averaged models for species richness and number of individuals per flock from different families (Furnariidae, Thraupidae, and Tyrannidae), foraging guilds (insectivores, frugivores, and omnivores), and height strata (canopy, mid-height, understory, and ground). Models for Furnariidae, Tyrannidae, and insectivores included temperature, vegetation complexity, and arthropods as predictor variables. Models for frugivores included temperature, vegetation complexity, and fruits as predictor variables. All other models included all four predictors (see Table 1). SE stands for unconditional standard error of the parameter estimate. Variables with model-averaged confidence intervals that overlap zero were considered not significant for the model. Variables marked with asterisk are considered supported. AIC tables for all models averaged are presented in Supplementary Material Tables S6–S15. . Models . . Richness . Number of individuals . Taxa/guild . Parameter . SE . 95% CI . . Parameter . SE . 95% CI . . Furnariidae Temperature –0.06 0.06 –0.17 0.05 –0.25 0.04 –0.33 –0.17* Vegetation complexity 0.01 0.06 –0.11 0.13 0.01 0.04 –0.07 0.09 Arthropods –0.06 0.05 –0.16 0.05 –0.08 0.04 –0.15 –0.01* Thraupidae Temperature –0.17 0.03 –0.24 –0.1* –0.23 0.02 –0.28 –0.19* Vegetation complexity 0.00 0.04 –0.07 0.08 0.06 0.02 0.02 0.10* Fruits –0.03 0.05 –0.13 0.07 –0.03 0.03 –0.10 0.04 Arthropods 0.02 0.03 –0.05 0.08 0.07 0.02 0.03 0.10* Tyrannidae Temperature 0.25 0.05 0.16 0.35* 0.31 0.04 0.23 0.39* Vegetation complexity 0.04 0.07 –0.08 0.17 0.02 0.05 –0.08 0.13 Arthropods –0.06 0.05 –0.15 0.03 –0.07 0.04 –0.15 0.00 Insectivores Temperature –0.02 0.03 –0.08 0.03 –0.12 0.02 –0.15 –0.09* Vegetation complexity 0.09 0.03 0.03 0.14* 0.09 0.02 0.06 0.12* Arthropods 0.00 0.02 –0.04 0.05 0.00 0.01 –0.03 0.03 Frugivores Temperature 0.00 0.05 –0.10 0.10 –0.04 0.02 –0.09 0.01 Vegetation complexity 0.08 0.05 –0.03 0.18 0.01 0.03 –0.06 0.08* Fruits –0.08 0.07 –0.21 0.05 –0.06 0.03 –0.13 0.00 Omnivores Temperature –0.01 0.06 –0.12 0.10 –0.14 0.03 –0.20 –0.07* Vegetation complexity 0.04 0.06 –0.07 0.15 0.00 0.05 –0.09 0.09 Fruits 0.00 0.07 –0.15 0.14 –0.11 0.05 –0.21 –0.02 Arthropods –0.07 0.05 –0.16 0.02 –0.04 0.03 –0.10 0.02 Canopy Temperature 0.16 0.06 0.05 0.28* –0.02 0.04 –0.11 0.06 Vegetation complexity –0.03 0.08 –0.18 0.13 –0.13 0.04 –0.21 –0.06* Fruits –0.08 0.08 –0.24 0.08 –0.15 0.05 –0.26 –0.05* Arthropods –0.06 0.06 –0.18 0.06 –0.18 0.03 –0.25 –0.12* Understory Temperature –0.06 0.05 –0.16 0.04 0.02 0.03 –0.04 0.08 Vegetation complexity 0.07 0.05 –0.03 0.18 0.12 0.03 0.06 0.18* Fruits 0.01 0.07 –0.12 0.15 0.01 0.04 –0.07 0.09 Arthropods 0.02 0.04 –0.06 0.10 0.05 0.03 0.00 0.09 Mid-height Temperature 0.02 0.03 –0.05 0.08 –0.10 0.02 –0.14 –0.06* Vegetation complexity 0.06 0.03 0.01 0.12* 0.04 0.02 0.00 0.08 Fruits –0.03 0.04 –0.12 0.05 –0.06 0.03 –0.11 –0.01* Arthropods –0.02 0.03 –0.07 0.03 0.01 0.01 –0.02 0.04 Ground Temperature –0.48 0.11 –0.70 –0.26* –0.78 0.08 –0.92 –0.63* Vegetation complexity 0.17 0.10 –0.02 0.36 0.24 0.07 0.09 0.38* Fruits 0.09 0.17 –0.23 0.42 0.20 0.11 –0.01 0.41 Arthropods 0.10 0.07 –0.04 0.24 0.21 0.05 0.12 0.29* . Models . . Richness . Number of individuals . Taxa/guild . Parameter . SE . 95% CI . . Parameter . SE . 95% CI . . Furnariidae Temperature –0.06 0.06 –0.17 0.05 –0.25 0.04 –0.33 –0.17* Vegetation complexity 0.01 0.06 –0.11 0.13 0.01 0.04 –0.07 0.09 Arthropods –0.06 0.05 –0.16 0.05 –0.08 0.04 –0.15 –0.01* Thraupidae Temperature –0.17 0.03 –0.24 –0.1* –0.23 0.02 –0.28 –0.19* Vegetation complexity 0.00 0.04 –0.07 0.08 0.06 0.02 0.02 0.10* Fruits –0.03 0.05 –0.13 0.07 –0.03 0.03 –0.10 0.04 Arthropods 0.02 0.03 –0.05 0.08 0.07 0.02 0.03 0.10* Tyrannidae Temperature 0.25 0.05 0.16 0.35* 0.31 0.04 0.23 0.39* Vegetation complexity 0.04 0.07 –0.08 0.17 0.02 0.05 –0.08 0.13 Arthropods –0.06 0.05 –0.15 0.03 –0.07 0.04 –0.15 0.00 Insectivores Temperature –0.02 0.03 –0.08 0.03 –0.12 0.02 –0.15 –0.09* Vegetation complexity 0.09 0.03 0.03 0.14* 0.09 0.02 0.06 0.12* Arthropods 0.00 0.02 –0.04 0.05 0.00 0.01 –0.03 0.03 Frugivores Temperature 0.00 0.05 –0.10 0.10 –0.04 0.02 –0.09 0.01 Vegetation complexity 0.08 0.05 –0.03 0.18 0.01 0.03 –0.06 0.08* Fruits –0.08 0.07 –0.21 0.05 –0.06 0.03 –0.13 0.00 Omnivores Temperature –0.01 0.06 –0.12 0.10 –0.14 0.03 –0.20 –0.07* Vegetation complexity 0.04 0.06 –0.07 0.15 0.00 0.05 –0.09 0.09 Fruits 0.00 0.07 –0.15 0.14 –0.11 0.05 –0.21 –0.02 Arthropods –0.07 0.05 –0.16 0.02 –0.04 0.03 –0.10 0.02 Canopy Temperature 0.16 0.06 0.05 0.28* –0.02 0.04 –0.11 0.06 Vegetation complexity –0.03 0.08 –0.18 0.13 –0.13 0.04 –0.21 –0.06* Fruits –0.08 0.08 –0.24 0.08 –0.15 0.05 –0.26 –0.05* Arthropods –0.06 0.06 –0.18 0.06 –0.18 0.03 –0.25 –0.12* Understory Temperature –0.06 0.05 –0.16 0.04 0.02 0.03 –0.04 0.08 Vegetation complexity 0.07 0.05 –0.03 0.18 0.12 0.03 0.06 0.18* Fruits 0.01 0.07 –0.12 0.15 0.01 0.04 –0.07 0.09 Arthropods 0.02 0.04 –0.06 0.10 0.05 0.03 0.00 0.09 Mid-height Temperature 0.02 0.03 –0.05 0.08 –0.10 0.02 –0.14 –0.06* Vegetation complexity 0.06 0.03 0.01 0.12* 0.04 0.02 0.00 0.08 Fruits –0.03 0.04 –0.12 0.05 –0.06 0.03 –0.11 –0.01* Arthropods –0.02 0.03 –0.07 0.03 0.01 0.01 –0.02 0.04 Ground Temperature –0.48 0.11 –0.70 –0.26* –0.78 0.08 –0.92 –0.63* Vegetation complexity 0.17 0.10 –0.02 0.36 0.24 0.07 0.09 0.38* Fruits 0.09 0.17 –0.23 0.42 0.20 0.11 –0.01 0.41 Arthropods 0.10 0.07 –0.04 0.24 0.21 0.05 0.12 0.29* Open in new tab TABLE 3. Parameter estimation for predictors in averaged models for species richness and number of individuals per flock from different families (Furnariidae, Thraupidae, and Tyrannidae), foraging guilds (insectivores, frugivores, and omnivores), and height strata (canopy, mid-height, understory, and ground). Models for Furnariidae, Tyrannidae, and insectivores included temperature, vegetation complexity, and arthropods as predictor variables. Models for frugivores included temperature, vegetation complexity, and fruits as predictor variables. All other models included all four predictors (see Table 1). SE stands for unconditional standard error of the parameter estimate. Variables with model-averaged confidence intervals that overlap zero were considered not significant for the model. Variables marked with asterisk are considered supported. AIC tables for all models averaged are presented in Supplementary Material Tables S6–S15. . Models . . Richness . Number of individuals . Taxa/guild . Parameter . SE . 95% CI . . Parameter . SE . 95% CI . . Furnariidae Temperature –0.06 0.06 –0.17 0.05 –0.25 0.04 –0.33 –0.17* Vegetation complexity 0.01 0.06 –0.11 0.13 0.01 0.04 –0.07 0.09 Arthropods –0.06 0.05 –0.16 0.05 –0.08 0.04 –0.15 –0.01* Thraupidae Temperature –0.17 0.03 –0.24 –0.1* –0.23 0.02 –0.28 –0.19* Vegetation complexity 0.00 0.04 –0.07 0.08 0.06 0.02 0.02 0.10* Fruits –0.03 0.05 –0.13 0.07 –0.03 0.03 –0.10 0.04 Arthropods 0.02 0.03 –0.05 0.08 0.07 0.02 0.03 0.10* Tyrannidae Temperature 0.25 0.05 0.16 0.35* 0.31 0.04 0.23 0.39* Vegetation complexity 0.04 0.07 –0.08 0.17 0.02 0.05 –0.08 0.13 Arthropods –0.06 0.05 –0.15 0.03 –0.07 0.04 –0.15 0.00 Insectivores Temperature –0.02 0.03 –0.08 0.03 –0.12 0.02 –0.15 –0.09* Vegetation complexity 0.09 0.03 0.03 0.14* 0.09 0.02 0.06 0.12* Arthropods 0.00 0.02 –0.04 0.05 0.00 0.01 –0.03 0.03 Frugivores Temperature 0.00 0.05 –0.10 0.10 –0.04 0.02 –0.09 0.01 Vegetation complexity 0.08 0.05 –0.03 0.18 0.01 0.03 –0.06 0.08* Fruits –0.08 0.07 –0.21 0.05 –0.06 0.03 –0.13 0.00 Omnivores Temperature –0.01 0.06 –0.12 0.10 –0.14 0.03 –0.20 –0.07* Vegetation complexity 0.04 0.06 –0.07 0.15 0.00 0.05 –0.09 0.09 Fruits 0.00 0.07 –0.15 0.14 –0.11 0.05 –0.21 –0.02 Arthropods –0.07 0.05 –0.16 0.02 –0.04 0.03 –0.10 0.02 Canopy Temperature 0.16 0.06 0.05 0.28* –0.02 0.04 –0.11 0.06 Vegetation complexity –0.03 0.08 –0.18 0.13 –0.13 0.04 –0.21 –0.06* Fruits –0.08 0.08 –0.24 0.08 –0.15 0.05 –0.26 –0.05* Arthropods –0.06 0.06 –0.18 0.06 –0.18 0.03 –0.25 –0.12* Understory Temperature –0.06 0.05 –0.16 0.04 0.02 0.03 –0.04 0.08 Vegetation complexity 0.07 0.05 –0.03 0.18 0.12 0.03 0.06 0.18* Fruits 0.01 0.07 –0.12 0.15 0.01 0.04 –0.07 0.09 Arthropods 0.02 0.04 –0.06 0.10 0.05 0.03 0.00 0.09 Mid-height Temperature 0.02 0.03 –0.05 0.08 –0.10 0.02 –0.14 –0.06* Vegetation complexity 0.06 0.03 0.01 0.12* 0.04 0.02 0.00 0.08 Fruits –0.03 0.04 –0.12 0.05 –0.06 0.03 –0.11 –0.01* Arthropods –0.02 0.03 –0.07 0.03 0.01 0.01 –0.02 0.04 Ground Temperature –0.48 0.11 –0.70 –0.26* –0.78 0.08 –0.92 –0.63* Vegetation complexity 0.17 0.10 –0.02 0.36 0.24 0.07 0.09 0.38* Fruits 0.09 0.17 –0.23 0.42 0.20 0.11 –0.01 0.41 Arthropods 0.10 0.07 –0.04 0.24 0.21 0.05 0.12 0.29* . Models . . Richness . Number of individuals . Taxa/guild . Parameter . SE . 95% CI . . Parameter . SE . 95% CI . . Furnariidae Temperature –0.06 0.06 –0.17 0.05 –0.25 0.04 –0.33 –0.17* Vegetation complexity 0.01 0.06 –0.11 0.13 0.01 0.04 –0.07 0.09 Arthropods –0.06 0.05 –0.16 0.05 –0.08 0.04 –0.15 –0.01* Thraupidae Temperature –0.17 0.03 –0.24 –0.1* –0.23 0.02 –0.28 –0.19* Vegetation complexity 0.00 0.04 –0.07 0.08 0.06 0.02 0.02 0.10* Fruits –0.03 0.05 –0.13 0.07 –0.03 0.03 –0.10 0.04 Arthropods 0.02 0.03 –0.05 0.08 0.07 0.02 0.03 0.10* Tyrannidae Temperature 0.25 0.05 0.16 0.35* 0.31 0.04 0.23 0.39* Vegetation complexity 0.04 0.07 –0.08 0.17 0.02 0.05 –0.08 0.13 Arthropods –0.06 0.05 –0.15 0.03 –0.07 0.04 –0.15 0.00 Insectivores Temperature –0.02 0.03 –0.08 0.03 –0.12 0.02 –0.15 –0.09* Vegetation complexity 0.09 0.03 0.03 0.14* 0.09 0.02 0.06 0.12* Arthropods 0.00 0.02 –0.04 0.05 0.00 0.01 –0.03 0.03 Frugivores Temperature 0.00 0.05 –0.10 0.10 –0.04 0.02 –0.09 0.01 Vegetation complexity 0.08 0.05 –0.03 0.18 0.01 0.03 –0.06 0.08* Fruits –0.08 0.07 –0.21 0.05 –0.06 0.03 –0.13 0.00 Omnivores Temperature –0.01 0.06 –0.12 0.10 –0.14 0.03 –0.20 –0.07* Vegetation complexity 0.04 0.06 –0.07 0.15 0.00 0.05 –0.09 0.09 Fruits 0.00 0.07 –0.15 0.14 –0.11 0.05 –0.21 –0.02 Arthropods –0.07 0.05 –0.16 0.02 –0.04 0.03 –0.10 0.02 Canopy Temperature 0.16 0.06 0.05 0.28* –0.02 0.04 –0.11 0.06 Vegetation complexity –0.03 0.08 –0.18 0.13 –0.13 0.04 –0.21 –0.06* Fruits –0.08 0.08 –0.24 0.08 –0.15 0.05 –0.26 –0.05* Arthropods –0.06 0.06 –0.18 0.06 –0.18 0.03 –0.25 –0.12* Understory Temperature –0.06 0.05 –0.16 0.04 0.02 0.03 –0.04 0.08 Vegetation complexity 0.07 0.05 –0.03 0.18 0.12 0.03 0.06 0.18* Fruits 0.01 0.07 –0.12 0.15 0.01 0.04 –0.07 0.09 Arthropods 0.02 0.04 –0.06 0.10 0.05 0.03 0.00 0.09 Mid-height Temperature 0.02 0.03 –0.05 0.08 –0.10 0.02 –0.14 –0.06* Vegetation complexity 0.06 0.03 0.01 0.12* 0.04 0.02 0.00 0.08 Fruits –0.03 0.04 –0.12 0.05 –0.06 0.03 –0.11 –0.01* Arthropods –0.02 0.03 –0.07 0.03 0.01 0.01 –0.02 0.04 Ground Temperature –0.48 0.11 –0.70 –0.26* –0.78 0.08 –0.92 –0.63* Vegetation complexity 0.17 0.10 –0.02 0.36 0.24 0.07 0.09 0.38* Fruits 0.09 0.17 –0.23 0.42 0.20 0.11 –0.01 0.41 Arthropods 0.10 0.07 –0.04 0.24 0.21 0.05 0.12 0.29* Open in new tab We also found a significantly positive response to increasing vegetation structure (Table 3), primarily in the number of individuals in the flock, as expected by the foraging niche hypothesis (Table 1). Among the taxonomic subsets sampled, only the number of thraupid individuals in the flock was significantly associated with increasing vegetation structure. For insectivores, however, both the species richness and a number of individuals were positively related to increasing structure. A significant effect of vegetation structure was also detected for the number of frugivores in the flock, though the more generalist omnivores did not show a significant response. Vegetation structure also exerted significant effects across vertical foraging strata, with the number of ground- and understory-foraging species in flocks significantly greater in more structured forest. For species that forage in the mid-height stratum, the significant effect was on the species richness rather than the number of individuals, while for canopy foragers the effect of vegetation structure was instead significantly negative. Finally, and as expected by the food resource hypothesis (Table 1), we found significantly negative correlations between fruit and arthropod diversity and the number of individuals in flocks (Table 3), though the sign of this effect differed across community subsets. The number of furnariids and canopy-foraging species increased with decreasing arthropod diversity, though the opposite effect was true for thraupids and ground-foraging species. A significantly negative effect of fruit diversity was also found on the abundance of canopy- and mid-height-foragers in a flock. Predictors of Taxonomic Species Composition The full CCA model (F4,363 = 5.77, P = 0.001) had three significant axes that explained 17% of the variation in bird species composition per flock (Figure 4; Supplementary Material Table S16). Axis 1 (λ = 0.70) was highly correlated with temperature (r = −0.95, P < 0.001); species with different elevational distributions were separated along Axis 1. Species in the center of this axis could be either ubiquitous or associated with middle elevations. Axis 2 (λ = 0.14) mostly represented a habitat complexity gradient (r = −0.60, P < 0.0001), with species being separated based on their use of habitats with different complexity of vertical strata. Finally, Axis 3 (λ = 0.14) was negatively correlated with fruit availability (r = −0.50, P < 0.0001) and positively but weakly correlated with arthropod availability (r = 0.30, P < 0.0001), potentially describing a gradient of resources for frugivores and insectivores. FIGURE 4. Open in new tabDownload slide Ordination of bird species in flocks plotted as a function of the first two axes of a canonical correspondence analysis considering four predictor variables: temperature, vegetation complexity, fruit availability, and arthropod availability. Axis 1 represents a temperature gradient, with cooler areas toward the right and warmer areas to the left. Axis 2 mainly represents a gradient of vegetation complexity, with species associated with more complex vegetation structure in the lower half of the plot, and species associated with less complex and more open habitats with higher fruit availability in the upper half of the plot. FIGURE 4. Open in new tabDownload slide Ordination of bird species in flocks plotted as a function of the first two axes of a canonical correspondence analysis considering four predictor variables: temperature, vegetation complexity, fruit availability, and arthropod availability. Axis 1 represents a temperature gradient, with cooler areas toward the right and warmer areas to the left. Axis 2 mainly represents a gradient of vegetation complexity, with species associated with more complex vegetation structure in the lower half of the plot, and species associated with less complex and more open habitats with higher fruit availability in the upper half of the plot. DISCUSSION We documented profound changes to the richness, size, evenness, and composition of mixed-species flocks, and the overall flock-joining community, along our elevational gradient. We observed a mid-elevation peak in flock species richness, size, and Shannon’s diversity at around 2,300 m (Figure 3). This trend was poorly explained by any single variable, however, and we suggest it is the result of a combination of behavioral and environmental factors. As predicted by the open-membership hypothesis, flock composition and the make-up of the flocking bird community closely mirrored changes to the overall bird community: not only did overall and taxon-specific richness decline at the same rate in the full community and flock-joining subset (Figure 1), but changes to flock composition followed similar trends as changes to community composition (Figure 2). We argue these results suggest that Andean flocks represent dynamic, unstructured associations that mainly reflect the composition of the larger community. Second, flocks were dominated by insectivorous species and the mid-elevation maximum of flock richness both closely mirrored a similar peak in the percentage of the local insectivore community joining flocks and was significantly predicted by local vegetation structure. The peak in mid-elevation richness in flocks, therefore, reflects both the open-membership composition of high-elevation flocks and the high vegetation complexity of mid-elevation forests, maximizing within-flock diversity. Changes to overall flock structure across elevations were therefore best supported by the foraging niche and abiotic stress hypotheses, while neither fruit nor arthropod abundance was significantly correlated with either overall richness or the abundance or richness of most community subsets. While it is possible that this might partially be a result of our arthropod sampling method (branch beating) that might underestimate insect diversity (Montaño-Centellas et al. 2020), we argue that ground-dwelling arthropods are unlikely to constitute a major component of the diet of flocking species, given the relative lack of ground-foraging species in the flocking community across elevations (Figure 2I). Instead, we believe the structuring effects of abiotic stress and changes to the diversity of foraging niches are more important than those of resource abundance in Andean flocking systems. High- and low-elevation flocks had distinguishable composition (Table 3, Figure 4), and higher elevations on our gradient represented areas with low temperatures and low vegetation complexity, the latter resulting from a loss of vertical strata (Terborgh 1977, Jankowski et al. 2013b). Many differences in guild composition and flock size were associated with temperature, with flocks at colder elevations having more individuals, especially of insectivorous and omnivorous species, greater thraupid species richness and abundance, and higher richness and abundance of ground-foraging species. By contrast, warmer-elevation flocks were of smaller size but contained more tyrannid and canopy-foraging species and individuals than their high-elevation counterparts. Vegetation complexity was also a good predictor of flock compositional changes. Flocks at mid and low elevations, where vegetation complexity was greater, had higher numbers of insectivorous species and individuals, higher number of thraupid individuals, higher richness of mid-height-foraging species, and higher abundances of canopy-, understory-, and ground-foraging species. Lower canopy heights and a simplified vegetation structure at higher elevations may engender a loss of foraging niches for insectivores, leading to a loss of insectivore species richness and functional diversity (Jankowski et al. 2013b). Our system may even underestimate the effect of vegetation structure because inter-Andean valley bottoms not connected to lowland forests may lack many lowland avian taxa (e.g., Thamnophilidae; Kessler et al. 2001). Open-Membership Flocks: Flock Composition Mirrors Changes to the Overall Community Our finding of large changes to the flocking bird community and flock composition across the elevational gradient matches similar findings elsewhere (Marín-Gómez and Arbeláez-Cortés 2015, Muñoz 2016). Interestingly, these changes to the flock-joining community mirrored changes to the absolute and relative abundance of taxa and foraging guilds within the full bird community (Figures 1A and 2), as posited by the open-membership hypothesis. While the percentage of the full community participating in flocks declined with increasing elevation, it represented a majority (50–60%) of the community over most of the gradient (Figure 1B), a similar proportion to other tropical forest systems (Zou et al. 2018). Changes to the full and flock-joining communities along our gradient mirrored similar changes to bird communities elsewhere in the Andes, notably an overall decline in species richness (e.g., Kattan and Franco 2004) driven by a steep decline in the number of insectivorous species (Terborgh 1977, Jankowski et al. 2013b, Pigot et al. 2016; Figure 2D). With increasing elevation, we found a steady decrease in the number and proportion of the flocking community made up of sub-oscine families (Furnariidae, Tyrannidae), and a proportional increase in the species richness of thraupids in the flock-joining community as well as in flocks (Table 3; Supplementary Material Figure S2C). These changes match known shifts in overall representation of these families in full Andean montane bird communities (Renjifo et al. 1997). We suggest that the correlation between the species richness and composition of the flock-joining and full bird communities is a product of the open-membership nature of Andean flocks. Indeed, it has long been noted that Andean mixed-species flocks show tremendous variation in their size and species richness (Poulsen 1996, Guevara et al. 2011, Montaño-Centellas 2020, Jones and Robinson 2020a), even at local study sites. This stands in notable contrast to the stable composition of Amazonian understory (Martínez and Gomez 2013, Mokross et al. 2014) and canopy (Munn 1985) flocks, where the core species spend all their time with the flock. Similar observational studies in Andean systems, however, found that many species dynamically joined and left the flock as it entered and exited their territories (Poulsen 1996, Pomara et al. 2007), resulting in a more variable flock composition over time. More recently, network analyses have shown that a high proportion of all species pairs associate in Andean flocks (Jones and Robinson 2020b), with higher density values than expected by chance (Montaño-Centellas 2020). These associations seem to be able to “rewire”, changing the intensity of the species interactions or the make-up of the participating species without a loss of the behavior as a whole (Montaño-Centellas 2020, Jones and Robinson 2020b). Taken together, these results suggest that Andean mixed-species flocks are representative of the larger bird community and might be useful indicators of the local avifauna for Andean biodiversity surveys (Montaño-Centellas and Garitano-Zavala 2015). Vegetation Structure Shapes Flock Participation: The Importance of Foraging Niche Diversity We found a significant effect of vegetation structure on the species richness, size, and evenness of flocks, and vegetation had an important effect on species richness and abundance of insectivores in flocks (Table 3). As in flocking systems in forests worldwide (Sridhar et al. 2009), the flock-joining bird community was dominated by insectivores in our system. We found that vegetation structure decreased with altitude, a pattern consistent with other elevational gradients (Terborgh 1977, Acharya et al. 2011, Jankowski et al. 2013b), although we observed a low-elevation plateau in structural diversity that began declining above 2,500 m.a.s.l. We argue this decline in complexity reduces the diversity of available foraging niches for insectivores, as we found a significant positive correlation between greater vegetation structure and higher richness of mid-height-foraging species in flocks. Decreasing vegetation structure therefore resulted in a high-elevation flocking community with increased proportions of understory and ground-foraging species (Supplementary Material Figure S2). In the Eastern Himalaya, Acharya and Vijayan (2017) found that foliage concentration and avian species richness were highest in the mid-layers (5–15 m) of montane forest. Because tropical forest birds exhibit strong vertical stratification in their foraging (Walther 2002, Mansor et al. 2019), the loss of mid-story and subcanopy vegetation strata with decreasing vertical structure could be driving much of this effect. For example, most of the mid-height-foraging tyrannid genera (e.g., Leptopogon, Myiophobus, Phylloscartes, Phyllomyias) are lost at high elevations, where tyrant flycatcher richness in flocks is much diminished. At the microhabitat level, flock diversity is known to increase with vegetation structure (Knowlton and Graham 2011, Zhang et al. 2013, Jones and Robinson 2020a), and flocks are known to select for greater vegetation complexity (Lee et al. 2005, McDermott and Rodewald 2014, Potts et al. 2014), providing further support for this hypothesis. If declining vegetation structure at high elevations explains the decline in flock richness from the mid-elevation peak, then how to explain the lower insectivore participation at low elevations despite high vegetation structure? We argue that a gradual increase in the structuring of flock composition with decreasing elevation might drive these observed patterns. Lowland flocking systems in the tropics often contain two or more flock types structured by foraging stratum (e.g., King and Rappole 2000, Srinivasan et al. 2012), whereas Andean flocking systems are consistently described as consisting of a mixture of understory and canopy species (e.g., Guevara et al. 2011, Colorado and Rodewald 2015). Changes from flocks structured by foraging stratum to open-membership flocks, however, are better described as a continuum. Network analyses along our gradient found that the modularity of flocking networks, a surrogate for the extent to which distinct flock sub-types exist at a given elevation, declined with altitude and was greater than expected by chance in the lowest elevation band (1,500 m.a.s.l.; Montaño-Centellas 2020). We therefore suggest that the transition of flocking behavior toward an open-membership system at ~2,300 m may allow for a greater proportion of the insectivore community to join flocks due to the lack of a structuring effect of foraging stratum. A potential mechanism driving these changes to flock composition is an increase in participation costs for flocking species and individuals at lower elevations. With increasing insectivore richness at low elevations, their partitioning of foraging microhabitats will likely be finer (Pigot et al. 2016), leading to more foraging specialists participating in flocks. With greater specialization and narrower niches, competition for resources is expected to increase (Cavender-Bares et al. 2009), leading to tighter niche packing (Pigot et al. 2016). Evidence for competition as a driver of the structure of Andean flocks has previously been reported at similar elevations (<2,500 m) to those at which we detected a decline in insectivore participation (Colorado and Rodewald 2015). Intriguingly, greater vegetation structure increased the number of individuals participating, but not the species richness, for many community subsets (e.g., thraupids, frugivores, understory foragers), suggesting that the costs of competition may be borne unevenly within some species that join flocks as a social group, as is the case in temperate flocks (e.g., juvenile individuals; Hogstad 1989). Alternatively, the costs of “activity matching” (adjusting foraging rate while in the flock) in flocks more structured by foraging stratum might prevent more dissimilar species from joining flocks (Hutto 1988, Sridhar and Guttal 2018, Jones et al. 2020), resulting in less speciose and less open-membership flocks with a more clear distinction between canopy and understory sub-types at lower elevations. Increasing Flock Size with Elevation: Reduced Food Resources and Increasing Abiotic Stress? In contrast to species richness, flock size and the number of thraupid individuals in flocks were significantly greater at high elevations. We suggest that these results reflect a shift from primarily insectivorous species that join flocks as singletons or pairs to largely gregarious omnivorous and frugivorous species that join flocks as intraspecific groups. Many thraupid genera in high-elevation cloud forests (Anisognathus, Buthraupis, Iridosornis, Diglossa, Kleinothraupis) are known to join flocks in large intraspecific groups (Poulsen 1996, Arbeláez-Cortés and Marín-Gomez 2012). Most of these genera are also frequently observed foraging in large monospecific groups where individuals benefit from group-size supplementary benefits (Goodale et al. 2020). For instance, many high-Andean thraupids supplement their insect diet with fruits or nectar (Remsen 1985), and, in fact, the most generalist categories in our guild analyses (thraupid species richness and abundance per flock, number of omnivores, and mid-height foragers) increased at high elevations (Table 3). It is possible to group participation facilitates the finding of food resources at these high elevations, where fruits have a patchier distribution (Valone 1989). The abundance of canopy and mid-height foragers in flocks significantly increased with declining fruit diversity, suggesting that high intraspecific abundance within flocks is a product of frugivorous foraging as a single-species group (Valburg 1992). This idea is further supported by an increase in the evenness of flocks with elevation (Figure 3D), suggesting that high intraspecific abundance in flocks was more evenly distributed across species at high elevations. Alternatively, predation-related benefits associated with joining flocks at high elevations, where niches are less diverse, might also benefit generalists that are gregarious (Gil et al. 2017). Greater facilitation is expected among conspecifics under harsher abiotic conditions at the highest elevations (Montaño-Centellas 2020), and even solitary Andean birds may increase their flocking propensity with increasing elevation (e.g., Cardellina canadensis; Céspedes and Bayly 2018). Thus, the gregarious behavior of many generalist thraupids at high elevations, itself a product of declining food availability and increasing abiotic stress, likely drives the increase in flock size. CONCLUSIONS: AN OPEN-MEMBERSHIP SYSTEM WITH ABIOTIC AND BIOTIC STRUCTURING FACTORS In the montane system studied here, an increase in abiotic stress and the loss of vegetation structure are the main environmental drivers shaping flock structure. As expected for the whole community, decreasing temperature and increasing abiotic stress were significantly associated with elevational changes to flock structure and individual-level participation. As noted elsewhere (Graham et al. 2012), assemblages at higher elevations contained fewer specialized species and occupied a smaller functional space. Vegetation structure was another general mechanism shaping flock composition along our gradient by directly affecting the availability of foraging niches, particularly for insectivores. The benefits gained by joining a flock vary among species and depend upon habitat context and flock composition, however (Borah et al. 2018). Greater vegetation structure opens new foraging microhabitats for species to exploit, but may also reduce species participation at low elevations through increased niche partitioning and subsequent foraging competition. Similarly, increasing abiotic stress and declining resource availability decrease specialist species richness in flocks, but can also increase individual- and species-level participation by generalist species at high elevations. Nevertheless, our findings primarily support the open-membership hypothesis that Andean flocks resemble the composition and structure of the avian community at different elevations rather than the influence of structuring mechanisms. This open-membership, however, itself varies with elevation, with a more restrictive access and a better definition of flock sub-types at lower elevations. Acknowledgments This study would not have been possible without volunteers and students that participated in data collection: R. Cortes, M. Montenegro, P. Velasquez, S. Moscoso, C. Mayta, M. Yapu, C. Agurto, M. Perez, D. Solano, K. Bedregal, and A. Vilchez. We are grateful to B. Nieto and Y. Fernandez for leading surveys and identification of botanical data, and to K. Losantos and N. Ohara for leading arthropod surveys and identification. B. A. Loiselle provided important feedback at the early stages of the project. We thank J. G. Blake and S. K. Robinson for helpful comments on the manuscript. Funding statement: This work was supported by a Graduate School Fellowship Award from the University of Florida, a UF Biodiversity Institute summer fellowship, grants from the American Ornithologist Union, American Philosophical Society, The Rufford Foundation, and a WWF EFN-Alumni Grant, all granted to F.A.M.C. Ethics statement: This project was conducted in full compliance with the Institutional Animal Care and Use Committee of the University of Florida (permit #201609398), and with permits provided by the Ministerio de Medio Ambiente y Agua and the Servicio Nacional de Areas Protegidas, Plurinational State of Bolivia. Author contributions: F.A.M.C. conceived the design, developed the sampling methods, collected the data, wrote the manuscript, and analyzed the data. H.H.J. heavily contributed to manuscript writing and editing. Both authors approved the final version and have no conflict of interest to declare. Data availability: Raw data on mixed-species flocks and environmental variables can be obtained from Montaño-Centellas and Jones (2021). LITERATURE CITED Abrams , P. A . ( 1995 ). Monotonic or unimodal diversity-productivity gradients: What does competition theory predict? Ecology 76 : 2019 – 2027 . Google Scholar Crossref Search ADS WorldCat Acharya , B. K. , N. J. Sanders, L. Vijayan, and B. Chettri ( 2011 ). Elevational gradients in bird diversity in the Eastern Himalaya: An evaluation of distribution patterns and their underlying mechanisms . PLoS One 6 : e29097 . Google Scholar Crossref Search ADS PubMed WorldCat Acharya , B. K. , and L. Vijayan ( 2017 ). Vertical stratification of birds in different vegetation types along an elevation gradient in the Eastern Himalaya, India . Ornithological Science 16 : 131 – 140 . Google Scholar Crossref Search ADS WorldCat Alatalo , R. V. , D. A. G. Eriksson, L. Gustafsson, and K. Larsson ( 1987 ). Exploitation competition influences the use of foraging sites by tits: Experimental evidence . Ecology 68 : 284 – 290 . Google Scholar Crossref Search ADS WorldCat Arbeláez-Cortés , E. , and O. H. Marín-Gomez ( 2012 ). The composition of mixed-species bird flocks in Alto Quindío, Colombia . The Wilson Journal of Ornithology 124 : 572 – 580 . Google Scholar Crossref Search ADS WorldCat Barton , K . ( 2019 ). Package ‘MuMIn’: Multi-Model inference. R package, version 1.43.15 . https://CRAN.R-project.org/package=MuMIn Borah , B. , S. Quader, and U. Srinivasan ( 2018 ). Responses of interspecific associations in mixed-species bird flocks to selective logging . Journal of Applied Ecology 55 : 1637 – 1646 . Google Scholar Crossref Search ADS WorldCat Borcard , D. , F. Gillet, and P. Legendre ( 2018 ). Numerical Ecology with R . Springer-Verlag , New York, NY, USA . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Burnham , K. , and D. Anderson ( 2004 ). Model Selection and Multi-Model Inference. A Practical Information-Theoretical Approach , 2nd edition. Springer-Verlag , New York, NY, USA . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Callaway , R. M . ( 1998 ). Competition and facilitation on elevation gradients in subalpine forests of the Northern Rocky Mountains, USA . Oikos 82 : 561 – 573 . Google Scholar Crossref Search ADS WorldCat Cavender-Bares , J. , K. H. Kozak, P. V. Fine, and S. W. Kembel ( 2009 ). The merging of community ecology and phylogenetic biology . Ecology Letters 12 : 693 – 715 . Google Scholar Crossref Search ADS PubMed WorldCat Céspedes, L. N., and N. J. Bayly (2018). Over-winter ecology and relative density of Canada Warbler Cardellina canadensis in Colombia: The basis for defining conservation priorities for a sharply declining long-distance migrant. Bird Conservation International 29:232–248. Colorado , G. J. , and A. D. Rodewald ( 2015 ). Assembly patterns of mixed-species avian flocks in the Andes . The Journal of Animal Ecology 84 : 386 – 395 . Google Scholar Crossref Search ADS PubMed WorldCat Cooper , R. , and R. Whitmore ( 1990 ). Arthropod sampling methods in ornithology. In Avian Foraging: Theory, Methodology, and Applications (M. L. Morrison, C. J. Ralph, J. Verner, and J. R. Jehl, Jr., Editors) . Studies in Avian Biology 13 : 29 – 37 . Google Scholar OpenURL Placeholder Text WorldCat Darrah , A. J. , and K. G. Smith ( 2013 ). Comparison of foraging behaviors and movement patterns of the Wedge-billed Woodcreeper (Glyphorynchus spirurus) traveling alone and in mixed-species flocks in Amazonian Ecuador . The Auk 130 : 629 – 636 . Google Scholar Crossref Search ADS WorldCat Develey , P. F. , and C. A. Peres ( 2000 ). Resource seasonality and the structure of mixed species bird flocks in a coastal Atlantic forest of southeastern Brazil . Journal of Tropical Ecology 16 : 33 – 53 . Google Scholar Crossref Search ADS WorldCat Dolby , A. S. , and T. C. Grubb ( 1999 ). Functional roles in mixed-species foraging flocks: A field manipulation . The Auk 116 : 557 – 559 . Google Scholar Crossref Search ADS WorldCat Eguchi , K. , S. Yamagishi, and V. Randriansolo ( 1993 ). The composition and foraging behaviour of mixed-species flocks of forest-living birds in Madagascar . Ibis 135 : 91 – 96 . Google Scholar Crossref Search ADS WorldCat Elsen , P. R. , M. W. Tingley, R. Kalyanaraman, K. Ramesh, and D. S. Wilcove ( 2017 ). The role of competition, ecotones, and temperature in the elevational distribution of Himalayan birds . Ecology 98 : 337 – 348 . Google Scholar Crossref Search ADS PubMed WorldCat Farine , D. R. , and H. Whitehead ( 2015 ). Constructing, conducting and interpreting animal social network analysis . The Journal of Animal Ecology 84 : 1144 – 1163 . Google Scholar Crossref Search ADS PubMed WorldCat Fisher, J. B., Y. Malhi, I. C. Torres, D. B. Metcalfe, M. J. van de Weg, P. Meir, J. E. Silva-Espejo, and W. H. Huasco (2013). Nutrient limitation in rainforests and cloud forests along a 3,000-m elevation gradient in the Peruvian Andes. Oecologia 172:889–902. Gentry , K. E. , D. P. Roche, S. G. Mugel, N. D. Lancaster, K. E. Sieving, T. M. Freeberg, and J. R. Lucas ( 2019 ). Flocking propensity by satellites, but not core members of mixed-species flocks, increases when individuals experience energetic deficits in a poor-quality foraging habitat . PLoS One 14 : e0209680 . Google Scholar Crossref Search ADS PubMed WorldCat Gil , M. A. , Z. Emberts, H. Jones, and C. M. St Mary ( 2017 ). Social information on fear and food drives animal grouping and fitness . The American Naturalist 189 : 227 – 241 . Google Scholar Crossref Search ADS PubMed WorldCat Goodale , E. , G. Beauchamp, and G. Ruxton ( 2017 ). Mixed-Species Animal Groups: Behavior, Community Structure and Conservation . Academic Press , London, UK . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Goodale , E. , S. W. Kotagama, T. R. S. Raman, S. Sidhu, U. M. Goodale, S. Parker, and J. Chen ( 2014 ). The response of birds and mixed-species bird flocks to human-modified landscapes in Sri Lanka and southern India . Forest Ecology and Management 329 : 384 – 392 . Google Scholar Crossref Search ADS WorldCat Goodale , E. , H. Sridhar, K. E. Sieving, P. Bangal, G. J. Colorado Z, D. R. Farine, E. W. Heymann, H. H. Jones, I. Krams, A. E. Martínez, et al. ( 2020 ). Mixed company: A framework for understanding the composition and organization of mixed-species animal groups . Biological Reviews of the Cambridge Philosophical Society 95 : 889 – 910 . Google Scholar Crossref Search ADS PubMed WorldCat Graham , C. H. , J. L. Parra, B. A. Tinoco, F. G. Stiles, and J. A. McGuire ( 2012 ). Untangling the influence of ecological and evolutionary factors on trait variation across hummingbird assemblages . Ecology 93 : S99 – S111 . Google Scholar Crossref Search ADS WorldCat Graves , G. R. , and N. J. Gotelli ( 1993 ). Assembly of avian mixed-species flocks in Amazonia . Proceedings of the National Academy of Sciences USA 90 : 1388 – 1391 . Google Scholar Crossref Search ADS WorldCat Greenberg , R . ( 2000 ). Birds of many feathers: The formation and structure of mixed-species flocks of forest birds. In On the Move: How and Why Animals Travel in Groups ( S. Boinski and P. A. Garber, Editors). University of Chicago Press , Chicago, IL, USA . pp. 521 – 558 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Gross , K . ( 2008 ). Positive interactions among competitors can produce species-rich communities . Ecology Letters 11 : 929 – 936 . Google Scholar Crossref Search ADS PubMed WorldCat Guevara , E. A. , J. C. Valarezo, A. Onofa, and F. Cupuerán ( 2011 ). Mixed-species flock composition in a northwestern Ecuadorian cloud forest . Ornitologia Neotropical 22 : 379 – 386 . Google Scholar OpenURL Placeholder Text WorldCat Hails , C. J . ( 1982 ). A comparison of tropical and temperate aerial insect abundance . Biotropica 14 : 310 – 313 . Google Scholar Crossref Search ADS WorldCat Hanz , D. M. , K. Böhning-Gaese, S. W. Ferger, S. A. Fritz, E. L. Neuschulz, M. Quitián, V. Santillán, T. Töpfer, and M. Schleuning ( 2019 ). Functional and phylogenetic diversity of bird assemblages are filtered by different biotic factors on tropical mountains . Journal of Biogeography 46 : 291 – 303 . Google Scholar Crossref Search ADS WorldCat Hogstad , O . ( 1989 ). Social organization and dominance behavior in some Parus species . The Wilson Bulletin 101 : 254 – 262 . Google Scholar OpenURL Placeholder Text WorldCat Holt , R. D . ( 1997 ). Community modules. In Multitrophic Interactions in Terrestrial Ecosystems ( A. C. Gange and V. K. Brown, Editors). Blackwell Science , Oxford, UK . pp. 333–349. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hutto , R. L . ( 1988 ). Foraging behaviour patterns suggest a possible cost associated with participation in mixed-species bird flocks . Oikos 51 : 79 – 83 . Google Scholar Crossref Search ADS WorldCat Jabłoński , P. G. , and S. Don Lee ( 2002 ). Foraging niche shifts in mixed-species flocks of tits in Korea . Journal of Field Ornithology 73 : 246 – 252 . Google Scholar Crossref Search ADS WorldCat Jankowski , J. E. , G. A. Londono, S. K. Robinson, and M. A. Chappell ( 2013a ). Exploring the role of physiology and biotic interactions in determining elevational ranges of tropical animals . Ecography 36 : 1 – 12 . Google Scholar Crossref Search ADS WorldCat Jankowski , J. E. , C. L. Merkord, W. F. Rios, K. G. Cabrera, N. S. Revilla, and M. R. Silman ( 2013b ). The relationship of tropical bird communities to tree species composition and vegetation structure along an Andean elevational gradient . Journal of Biogeography 40 : 950 – 962 . Google Scholar Crossref Search ADS WorldCat Jones , H. H. , and S. K. Robinson ( 2020a ). Patch size and vegetation structure drive changes to mixed-species flock diversity and composition across a gradient of fragment sizes in the Western Andes of Colombia . The Condor: Ornithological Applications 122 : 1 – 26 . doi:10.1093/condor/duaa006 Google Scholar Crossref Search ADS WorldCat Jones , H. H. , and S. K. Robinson ( 2020b ). Vegetation structure drives mixed-species flock interaction strength and nuclear species roles . Behavioral Ecology 32 : 69 – 81 . Google Scholar Crossref Search ADS WorldCat Jones , H. H. , M. J. Walters, and S. K. Robinson ( 2020 ). Do similar foragers flock together? Nonbreeding foraging behavior and its impact on mixed-species flocking associations in a subtropical region . The Auk: Ornithological Advances 137 : 1 – 16 . doi:10.1093/auk/ukz079 Google Scholar Crossref Search ADS WorldCat Kattan , G. H. , and P. Franco ( 2004 ). Bird diversity along elevational gradients in the Andes of Colombia: Area and mass effects . Global Ecology and Biogeography 13 : 451 – 458 . Google Scholar Crossref Search ADS WorldCat Kessler , M. , S. K. Hertzog, J. Fjeldsa, and K. Bach ( 2001 ) Species richness and endemism of plant and bird communities along two gradients of elevation, humidity and land use in the Bolivian Andes . Diversity and Distributions 7 : 61 – 77 . Google Scholar Crossref Search ADS WorldCat King , D. I. , and J. H. Rappole ( 2000 ). Winter flocking of insectivorous birds in montane pine-oak forests in middle America . The Condor 102 : 664 – 672 . Google Scholar Crossref Search ADS WorldCat Knowlton , J. L. , and C. H. Graham ( 2011 ). Species interactions are disrupted by habitat degradation in the highly threatened Tumbesian region of Ecuador . Ecological Applications 21 : 2974 – 2986 . Google Scholar Crossref Search ADS WorldCat Lee , T. M. , M. C. Soh, N. Sodhi, L. P. Koh, and S. L.-H. Lim ( 2005 ). Effects of habitat disturbance on mixed species bird flocks in a tropical sub-montane rainforest . Biological Conservation 122 : 193 – 204 . Google Scholar Crossref Search ADS WorldCat Lima , S. L . ( 1986 ). Predation risk and unpredictable feeding conditions: Determinants of body mass in birds . Ecology 67 : 377 – 385 . Google Scholar Crossref Search ADS WorldCat MacArthur , R. H. , and J. W. MacArthur ( 1961 ). On bird species diversity . Ecology 42 : 594 – 598 . Google Scholar Crossref Search ADS WorldCat Magurran , A. E . ( 2004 ). Measuring Biological Diversity . Blackwell Publishing , Malden, MA, USA . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Mammides , C. , J. Chen, U. M. Goodale, S. W. Kotagama, S. Sidhu, and E. Goodale ( 2015 ). Does mixed-species flocking influence how birds respond to a gradient of land-use intensity? Proceedings of the Royal Society of London B: Biological Sciences 282 : 20151118 . Google Scholar OpenURL Placeholder Text WorldCat Mangini , G. G. , and J. I. Areta ( 2018 ). Bird mixed-species flock formation is driven by low temperatures between and within seasons in a subtropical Andean-foothill forest . Biotropica 50 : 816 – 825 . Google Scholar Crossref Search ADS WorldCat Mansor , M. S. , F. Z. Rozali, N. A. Abdullah, S. M. Nor, and R. Ramli ( 2019 ). How important is aerial leaf litter for insectivorous birds foraging in a Malaysian tropical forest? Global Ecology and Conservation 20 : e00722 . Google Scholar Crossref Search ADS WorldCat Marín-Gómez , O. H. , and E. Arbeláez-Cortés ( 2015 ). Variation on species composition and richness in mixed bird flocks along an altitudinal gradient in the Central Andes of Colombia . Studies on Neotropical Fauna and Environment 50 : 113 – 129 . Google Scholar Crossref Search ADS WorldCat Marra , P. P. , and J. V. Remsen ( 1997 ). Insights into the maintenance of high species diversity in the neotropics: Habitat selection and foraging behavior in understory birds of tropical and temperate forests. In Studies in Neotropical Ornithology Honoring Ted Parker ( J. V. Remsen, Jr., Editor). Ornithological Monographs 48 : 445 – 483 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Martínez , A. E. , and J. P. Gomez ( 2013 ). Are mixed-species bird flocks stable through two decades? The American Naturalist 181 : E53 – E59 . Google Scholar Crossref Search ADS PubMed WorldCat Martínez , A. E. , E. Parra, O. Muellerklein, and V. T. Vredenburg ( 2018 ). Fear-based niche shifts in neotropical birds . Ecology 99 : 1338 – 1346 . Google Scholar Crossref Search ADS PubMed WorldCat Matthysen , E. , F. Collet, J. Cahill, F. Collet, and J. Cahill ( 2008 ). Mixed flock composition and foraging behavior of insectivorous birds in undisturbed and disturbed fragments of high-Andean Polylepis woodland . Ornitologia Neotropical 19 : 403 – 416 . Google Scholar OpenURL Placeholder Text WorldCat McCain , C. M . ( 2009 ). Global analysis of bird elevational diversity . Global Ecology and Biogeography 18 : 346 – 360 . Google Scholar Crossref Search ADS WorldCat McCune , B. , J. B. Grace, and D. L. Urban ( 2002 ). Analysis of Ecological Communities . MjM Software Design , Gleneden Beach, OR, USA . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC McDermott , M. E. , and A. D. Rodewald ( 2014 ). Conservation value of silvopastures to Neotropical migrants in Andean forest flocks . Biological Conservation 175 : 140 – 147 . Google Scholar Crossref Search ADS WorldCat Mokross , K. , T. B. Ryder, M. C. Côrtes, J. D. Wolfe, and P. C. Stouffer ( 2014 ). Decay of interspecific avian flock networks along a disturbance gradient in Amazonia . Proceedings of the Royal Society of London B: Biological Sciences 281 : 20132599 . Google Scholar OpenURL Placeholder Text WorldCat Montaño-Centellas , F. A . ( 2020 ). Interaction networks of avian mixed-species flocks along elevation in the tropical Andes . Ecography 43 : 930 – 942 . Google Scholar Crossref Search ADS WorldCat Montaño-Centellas , F. A. , and Á. Garitano-Zavala ( 2015 ). Andean bird responses to human disturbances along an elevational gradient . Acta Oecologica 65 – 66 : 51 – 60 . Google Scholar OpenURL Placeholder Text WorldCat Montaño-Centellas , F. A. , and H. H. Jones ( 2021 ). Data from: Temperature and vegetation complexity structure mixed-species flocks along a gradient of elevation in the tropical Andes . Ornithology 138 : 1 – 18 . doi:10.5061/dryad.mpg4f4qzd Google Scholar OpenURL Placeholder Text WorldCat Montaño-Centellas , F. A. , Loiselle , B. A., and M. W. Tingley ( 2020 ). Ecological drivers of avian community assembly along a tropical elevation gradient . Ecography 4 : 1 – 15 . Google Scholar OpenURL Placeholder Text WorldCat Morse , D. H . ( 1970 ). Ecological aspects of some mixed-species foraging flocks of birds . Ecological Monographs 40 : 119 – 168 . Google Scholar Crossref Search ADS WorldCat Munn , C. A . ( 1985 ). Permanent canopy and understorey flocks in Amazonia: Species composition and population density. In Neotropical Ornithology ( P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgley, and F. G. Buckley, Editors). Ornithological Monographs 36: 683 – 712 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Munn , C. A. , and J. W. Terborgh ( 1979 ). Multi-species territoriality in Neotropical foraging flocks . The Condor 81 : 338 – 347 . Google Scholar Crossref Search ADS WorldCat Muñoz , J . ( 2016 ). The role of facilitation in the structure of tropical bird communities: A case study of mixed-species flocks . MSc. thesis, University of British Columbia , Vancouver, BC, Canada . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Oksanen , J. , F. G. Blanchet, R. Kindt, P. Legendre, P. R. Minchin, R. O’Hara, G. L. Simpson, P. Solymos, M. H. H. Stevens, and H. Wagner ( 2018 ). Package ‘vegan’. Community ecology package, version 2.5–3 . https://cran.r-project.org/ Pigot , A. L. , C. H. Trisos, and J. A. Tobias ( 2016 ). Functional traits reveal the expansion and packing of ecological niche space underlying an elevational diversity gradient in passerine birds . Proceedings of the Royal Society B: Biological Sciences 283 : 20152013 . Google Scholar Crossref Search ADS WorldCat Pomara , L. Y. , R. J. Cooper, and L. J. Petit ( 2007 ). Modeling the flocking propensity of passerine birds in two Neotropical habitats . Oecologia 153 : 121 – 133 . Google Scholar Crossref Search ADS PubMed WorldCat Potts , J. R. , K. Mokross, P. C. Stouffer, and M. A. Lewis ( 2014 ). Step selection techniques uncover the environmental predictors of space use patterns in flocks of Amazonian birds . Ecology and Evolution 4 : 4578 – 4588 . Google Scholar Crossref Search ADS PubMed WorldCat Poulsen , B. O . ( 1996 ). Relationships between frequency of mixed-species flocks, weather and insect activity in a montane cloud forest in Ecuador . Ibis 138 : 466 – 470 . Google Scholar Crossref Search ADS WorldCat Remsen , J. V . ( 1985 ). Community organization and ecology of birds of high elevation humid forest of the Bolivian Andes. In Neotropical Ornithology. ( P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgley, and F. G. Buckley, Editors). Ornithological Monographs 36: 733 – 756 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Renjifo , L. M. , G. P. Servat, J. M. Goerck, B. A. Loiselle, and J. G. Blake ( 1997 ). Patterns of species composition and endemism in the northern Neotropics: A case for conservation of montane avifaunas. In Studies in Neotropical Ornithology Honoring Ted Parker ( J. V. Remsen, Jr., Editor). Ornithological Monographs 48: 577 – 594 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Ribera , M. O . ( 1995 ). Aspectos ecológicos, del uso de la tierra y conservación en el parque nacional y área natural de manejo integrado Cotapata . In Caminos de Cotapata ( C. B. , Morales, Editor). Instituto de Ecología, Fundeco, Fonama , La Paz, Bolivia . pp. 1 – 84 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ripley , B. , B. Venables, D. M. Bates, K. Hornik, A. Gebhardt, D. Firth, and M. B. Ripley ( 2018 ). Package ‘MASS’, version 7.3–51.1 . http://www.stats.ox.ac.uk/pub/MASS4/ Sevilla Callejo , M . ( 2010 ). Organización territorial y campesinado en el Parque Nacional y área Natural de Manejo Integrado Cotapata (Bolivia) . Doctorate dissertation, Universidad Autónoma de Madrid , Madrid, Spain . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Sridhar , H. , G. Beauchamp, and K. Shanker ( 2009 ). Why do birds participate in mixed-species foraging flocks? A large-scale synthesis . Animal Behaviour 78 : 337 – 347 . Google Scholar Crossref Search ADS WorldCat Sridhar , H. , and V. Guttal ( 2018 ). Friendship across species borders: Factors that facilitate and constrain heterospecific sociality . Philosophical Transactions of the Royal Society B: Biological Sciences 373 : 20170014 . Google Scholar Crossref Search ADS WorldCat Sridhar , H. , U. Srinivasan, R. A. Askins, J. C. Canales-Delgadillo, C. C. Chen, D. N. Ewert, G. A. Gale, E. Goodale, W. K. Gram, P. J. Hart, et al. ( 2012 ). Positive relationships between association strength and phenotypic similarity characterize the assembly of mixed-species bird flocks worldwide . The American Naturalist 180 : 777 – 790 . Google Scholar Crossref Search ADS PubMed WorldCat Srinivasan, U., R. H. Raza, and S. Quader (2012). Patterns of species participation across multiple mixed-species flock types in a tropical forest in northeastern India. Journal of Natural History 46:2749–2762. Stachowicz , J. J . ( 2001 ). Mutualism, facilitation, and the structure of ecological communities . BioScience 51 : 235 – 246 . Google Scholar Crossref Search ADS WorldCat Supriya , K. , C. S. Moreau, K. Sam, and T. D. Price ( 2019 ). Analysis of tropical and temperate elevational gradients in arthropod abundance . Frontiers of Biogeography 11 : e43104 . Google Scholar Crossref Search ADS WorldCat Ter Braak, C. J. F. (1986). Canonical correspondence analysis: A new eigenvector technique for multivariate direct gradient analysis. Ecology 67:1167–1179. Terborgh , J . ( 1977 ). Bird species diversity on an Andean elevational gradient . Ecology 58 : 1007 – 1019 . Google Scholar Crossref Search ADS WorldCat Tubelis , D. P. , A. Cowling, and C. Donnelley ( 2006 ). Role of mixed-species flocks in the use of adjacent savannas by forest birds in the central Cerrado, Brazil . Austral Ecology 31 : 38 – 41 . Google Scholar Crossref Search ADS WorldCat Valburg , L. K . ( 1992 ). Flocking and frugivory: The effect of social groupings on resource use in the Common Bush-Tanager . The Condor 94 : 358 – 363 . Google Scholar Crossref Search ADS WorldCat Valone , T. J . ( 1989 ). Group foraging, public information, and patch estimation . Oikos 56 : 357 – 363 . Google Scholar Crossref Search ADS WorldCat Walther , B. A . ( 2002 ). Vertical stratification and use of vegetation and light habitats by Neotropical forest birds . Journal für Ornithologie 143 : 64 – 81 . Google Scholar Crossref Search ADS WorldCat Wilman , H. , J. Belmaker, J. Simpson, C. de la Rosa, M. M. Rivadeneira, and W. Jetz ( 2014 ). EltonTraits 1.0: Species-level foraging attributes of the world’s birds and mammals . Ecology 95 : 2027 – 2027 . Google Scholar Crossref Search ADS WorldCat Witter , M. S. , and I. C. Cuthill ( 1993 ). The ecological costs of avian fat storage . Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 340 : 73 – 92 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Zeileis , A. , F. Cribari-Neto, B. Gruen, I. Kosmidis, A. B. Simas, A. V. Rocha, and M. A. Zeileis ( 2016 ). Package ‘betareg’, version 3.3. in R package . https://cran.r-project.org/web/packages/betareg/betareg.pdf Zhang , Q. , R. C. Han, M. Zhang, Z. Huang, and F. Zou ( 2013 ). Linking vegetation structure and bird organization: Response of mixed-species bird flocks to forest succession in subtropical China . Biodiversity and Conservation 22 : 1965 – 1989 . Google Scholar Crossref Search ADS WorldCat Zhou , L. , I. Peabotuwage, H. Gu, D. Jiang, G. Hu, A. Jiang, C. Mammides, M. Zhang, R.-C. Quan, and E. Goodale ( 2019 ). The response of mixed-species bird flocks to anthropogenic disturbance and elevational variation in southwest China . The Condor: Ornithological Applications 121 : 1 – 13 . doi:10.1093/condor/duz028 Google Scholar Crossref Search ADS WorldCat Zou , F. , H. Jones, G. J. Colorado Z., D. Jiang, T.-M. Lee, A. Martínez, K. Sieving, M. Zhang, Q. Zhang, and E. Goodale ( 2018 ). The conservation implications of mixed-species flocking in terrestrial birds, a globally-distributed species interaction network . Biological Conservation 224 : 267 – 276 . Google Scholar Crossref Search ADS WorldCat Copyright © American Ornithological Society 2021. All rights reserved. For permissions, e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

OrnithologyOxford University Press

Published: May 6, 2021

References