Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The role of facilitation in the structure of tropical bird communities : a case study of mixed-species… Munoz, Jenny 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2016_september_munoz_jenny.pdf [ 1.88MB ]
Metadata
JSON: 24-1.0303480.json
JSON-LD: 24-1.0303480-ld.json
RDF/XML (Pretty): 24-1.0303480-rdf.xml
RDF/JSON: 24-1.0303480-rdf.json
Turtle: 24-1.0303480-turtle.txt
N-Triples: 24-1.0303480-rdf-ntriples.txt
Original Record: 24-1.0303480-source.json
Full Text
24-1.0303480-fulltext.txt
Citation
24-1.0303480.ris

Full Text

THE ROLE OF FACILITATION IN THE STRUCTURE OF TROPICAL BIRD COMMUNITIES: A CASE STUDY OF MIXED-SPECIES FLOCKS  by  Jenny Munoz  B.Sc., Universidad de Antioquia, 2011  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Zoology) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   May 2016  © Jenny Munoz, 2016    ii Abstract  Understanding the influence of species interactions on community structure is a long-standing goal in ecology. While many studies have focused on negative biotic interactions, the role of other mechanisms has received less attention, in particular, facilitation. In birds, a striking case of facilitation occurs in mixed-species flocks, in which individuals of different species move and forage as a group to obtain benefits from the association. These associations of species in mixed flocks have been described in different habitats during the last century; however, there is still much debate regarding the prevalence of this foraging strategy and the role it plays in Neotropical bird communities. In this study, I integrated data from mixed species flocks observations and species occurrence to investigate how facilitative interactions influence the structure of Neotropical bird communities across a 3000-m elevational gradient on the eastern slope of the Andes in Peru.  First, I examine how the structure of mixed flocks changes across elevations. Second, I quantify the stability of these multispecies groups over time. Third, I evaluate the association of several key habitat variables with flock diversity. Finally, utilizing a dataset for the entire forest bird community, I assess the prevalence and importance of mixed-species flocks across the gradient. The results showed that flocks were highly organized and stable across elevations. Flocks across the gradient exhibited a similar general structure, composed of a stable core group of species and a more dynamic component of attending species. This spatial and temporal analysis suggests that the stability of mixed-species flocks in the Andes is similar to what has been previously described in the Amazonian lowlands, with flocks exhibiting stable home ranges and core member composition over time. Vegetation structure    iii explained 63% of variation in flock richness along the gradient, with number of trees and canopy height as primary predictors. Importantly, this study demonstrates that mixed-species flocks are used by more than a third of bird species present in the community, suggesting that these facilitative interactions are an important and underappreciated component of tropical bird communities.       iv Preface  The research questions and project design were carried out collaboratively between my supervisor, Jill Jankowski, and me. I carried out the fieldwork for this project; however some data on species occurrence along the gradient were obtained from the Manu project database. I conducted all the analysis and writing of this thesis. Dr. Jankowski provided helpful feedback.                    v Table of Contents  Abstract .......................................................................................................................................... ii	  Preface ........................................................................................................................................... iv	  List of Tables .............................................................................................................................. viii	  List of Figures ............................................................................................................................... ix	  List of Symbols and Abbreviations ............................................................................................ xi	  Acknowledgements ..................................................................................................................... xii	  Dedication ................................................................................................................................... xiii	  Chapter 1: General introduction ..................................................................................................1	  1.1	   Introduction to facilitation ................................................................................................. 1	  1.2	   Facilitation in bird communities ........................................................................................ 4	  1.3	   Tropical mixed-species bird flocks .................................................................................... 6	  1.4	   Study goals ......................................................................................................................... 7	  Chapter 2: Mixed-species flocks along an elevational gradient and their importance for bird communities. ..........................................................................................................................8	  2.1	   Introduction ........................................................................................................................ 8	  2.2	   Research questions ........................................................................................................... 11	  2.3	   Materials and methods ..................................................................................................... 12	  2.3.1	   Study site ................................................................................................................... 12	  2.3.2	   Data collection mixed-species flocks ........................................................................ 12	  2.3.3	   Species participation in mixed-species flocks .......................................................... 14	     vi 2.3.4	   Foraging guild classification ..................................................................................... 16	  2.3.5	   Flock stability............................................................................................................ 16	  2.3.6	   Home range characterization .................................................................................... 17	  2.3.7	   Vegetation structure .................................................................................................. 17	  2.3.8	   Prevalence of flocking at the community level ......................................................... 18	  2.3.9	   Data analysis ............................................................................................................. 19	  2.4	   Results .............................................................................................................................. 21	  2.4.1	   Spatial distribution of mixed-species flocks ............................................................. 21	  2.4.2	   Species participation in mixed-species flocks .......................................................... 23	  2.4.3	   Foraging guild participation in flocks ....................................................................... 24	  2.4.4	   Flocks stability .......................................................................................................... 25	  2.4.5	   Home range and territoriality .................................................................................... 26	  2.4.6	   Elevation and vegetation structure ............................................................................ 27	  2.4.7	   Prevalence of flocking at the community level ......................................................... 28	  2.5	   Discussion ........................................................................................................................ 28	  2.5.1	   Spatial distribution of flocks ..................................................................................... 29	  2.5.2	   Species participation in mixed-species flocks .......................................................... 31	  2.5.3	   Mixed-species flocks richness and foraging guild participation ............................... 32	  2.5.4	   Flock stability............................................................................................................ 34	  2.5.5	   Elevation and vegetation structure ............................................................................ 36	  2.5.6	   Prevalence of flocking at the community level ......................................................... 36	  Chapter 3: Conclusions ...............................................................................................................57	     vii 3.1	   General conclusions ......................................................................................................... 57	  3.2	   Future research ................................................................................................................. 59	  References .....................................................................................................................................61	  Appendices ....................................................................................................................................71	  Appendix A List of species ....................................................................................................... 71	  Appendix B Model selection .................................................................................................... 85	  Appendix C Model flock size ................................................................................................... 86	  Appendix D Regression canopy height ..................................................................................... 87	  Appendix E Regression ............................................................................................................. 88	  Appendix F Foraging guilds ..................................................................................................... 89	      viii List of Tables  Table 2.1 Mixed-species flock types identified across the elevational gradient. Mean number of species (Mean ± SD) and mean number of individual birds for each flock type is presented. Sample	  size	  and	  range	  of	  values	  are	  shown	  in	  bold ................................................................. 39   Table 2.2 Species	   with	   high	   tendency	   to	   flock.	   The	   core	   component	   of	   each	   flock	   type,	  Flocking	  Index	  (Ocurrrence*Propensity)	  and	  number	  of	  flocks	  observed	  are	  presented.	  	  Species	  that	  were	  intraspecifically	  gregarious	  are	  indicated	  with	  an	  asterisk	  (*) ................ 40   Table 2.3 Model selection results from generalized linear models for flock richness across the elevational gradient. Explanatory variables include elevation (Elev), mean canopy height (Canopy), and number of trees (Trees). For each fitted model the number of parameters (k), change in corrected quasi-Akaike from the model with the lowest QAIC value (Delta_QAICc), QAIC weights (QAICcWt) are shown ........................................................................................... 41	      ix List of Figures Figure 2.1 Dendrogram for average linkage cluster of mixed-species flocks along an elevational gradient. Clustering distance is based on Jaccard dissimilarity index of species composition of flocks. Each flock name indicates the elevation where it was observed. Red lines indicate clusters of the main flock types.. ................................................................................................................ 42  Figure 2.2 Mixed-species richness for Lowland, Low-montane and High-montane flocks in the Manu region, Peru. Number of species per flock is shown. Each dot represents an independent flock. The grey shading indicates the confidence intervals. …......…………......………….........43  Figure 2.3 Mixed-species flock size for Lowland, Low-montane and High-montane flocks in the Manu region, Peru.  Number of individuals per flock is shown. Each dot represents an independent flock. The grey shading indicates the confidence intervals …......…………………44  Figure 2.4 Regression of number of species and number of individuals foraging in mixed-species flocks in Manu region, Peru (Kendall’s	  Tau=	  0.789,	  p<0.001) …......…………......….45  Figure 2.5 Species richness of avian foraging guilds across the Manu elevational gradient for (a) Mixed-species flocks (b) overall community. Each guild is represented by a different colour including: insectivores (black), frugivores (red), nectarivores (green) and omnivores (blue). Each dot represents the total number of species at a given elevational zone (e.g. 400-500; 501-600; 601-700). The solid lines are the predicted values of species as a function of elevation….........46  Figure 2.6 Species richness of avian foraging guilds across the Manu elevational gradient. Total	  number	  of	  species	  in	  the	  Manu	  gradient	  community	  (red)	  and	  number	  of	  species	  participating	  in	  flocks	  (blue)	  for	  each	  elevational	  zone	  are	  shown for (a) insectivores,  (b) frugivores,	  	  (c) omnivores, (d) nectarivores. The	  solid	  lines	  are	  the	  predicted	  values	  of	  species	  richness	  as	  a	  function	  of	  elevation…......…………......…………......…………......…………..47  Figure 2.7 Proportions of species in the community joining mixed flocks for each foraging guild across elevation are shown. Each guild is represented by a different colour including: insectivores (black), frugivores (red), omnivores (blue) and nectarivores (green). Each dot represents the proportion of species joining flocks at a given elevational zone (e.g. 400-500; 501-600; 601-700). The solid lines are the predicted values as a function of elevation…......……….48  Figure 2.8 Stability of flocks over time. Similarity in species composition of flocks observed in (a) hours 1-2, (b) days 1-2 and (c) years 1-2 is shown. Each dot represents the calculated similarity index (1-Jaccard dissimilarity index) for each flock between observations. The grey shading shows the confidence intervals ………......…………......…………......…………......…49  Figure 2.9 Temporal stability of flocks across elevation. Similarity in species composition of flocks across elevation at three different time scales. The solid lines are the predicted values of    x flock stability as a function of elevation between hours (blue), weeks (red) and years (green). The shading shows the confidence intervals based on the standard errors of the estimates. The lines along the x-axis indicate the elevation where each flock was observed…….......................50  Figure 2.10 Stability of flocks over time. Similarity in species composition for a) Lowland flocks b) Low-montane flocks c) High-montane flocks. Each dot represents the calculated similarity index (1-Jaccard dissimilarity index) for each flock between observations…......……51  Figure 2.11 Stability of flocks across years. Similarity in species composition of flocks observed in 2013 and 2014 for a) Lowland flocks b)Low-montane flocks c)High-montane flocks. Each dot represents the calculated similarity index (1-Jaccard dissimilarity index) for each flock between observations………………….………….………….………….………….………….…......…...52  Figure 2.12 Home range of two mixed-species bird flocks at Low-montane elevation (1240-1260 masl). Light colours indicate the 2013 home range for each flock; dark colours indicate the 2014 home range. Roosting site is indicated for each flock………….………….…......………..53  Figure 2.13 Mixed-species flock richness in the Manu study region varying by (a) canopy height, (b) number of trees and (c) elevation. Best-fitted model was used to plot canopy height and number of trees. Second best–fitted model was used to plot elevation. The solid line is the predicted value of flock richness as a function of each variable, filling in the other explanatory variables to the median. Each dot represents a obsered flock. The grey shading shows the confidence intervals………….………….………….………….………….……………......……54  Figure 2.14 Bird species richness per elevational band. Total number of species in the Manu gradient community (black) and number of species participating in flocks (blue) for each elevational zone are shown (e.g. 400-500; 501-600; 601-700). The solid line is the predicted value of species richness as a function of elevation. ………….………….………….…......…...55  Figure 2.15 Prevalence of flocking across elevations. Proportion of species of the community joining flocks for each elevation band is shown. ………….………….………….…......………56     xi List of Symbols and Abbreviations  ~  approximately dbh diameter at breast height FI flocking index m meters m2         square meters masl meters above sea level NPP	  	  	  	  	  	  net	  primary	  productivity SE standard error     xii Acknowledgements   I would like to thank my supervisor, Jill Jankowski, for her guidance, advice, support and patience throughout my graduate work. I feel very privileged to have the opportunity to work with her. Thank you to my committee members Leticia Aviles and Darren Irwin for their thoughtful comments and ideas.  I feel very fortunate to be part of Manu project during the last years and I would like to thank Gustavo Londoño for his encouragement and support during all these years, and also to the Manu project volunteers in particular 2013 crew Mario Loaiza, Juliana Soto, Camilo De los Rios, Julian Heavyside, Wendy Valencia, Jeremia Kennedy. I would like to thank Paula Morales, Laura Gomez whose positivity tenacity and enthusiasm under challenging conditions were key to have successful field seasons. For insightful comments and discussions on this project I would like to thank Gustavo Londoño, Scott Robinson and Patrick Kelley.  I offer my enduring gratitude to the professors, staff and fellow students at Biodiversity Research Center, who have inspired me enlarging my vision of science. Thank you to Alice Liou, the Zoology Department graduate secretary, for her patience and help. For their support, encouragement, positivity, patience and inspiration during this endeavor, I am indebted to my family mom, dad and brother, but especially to Santiago David. This work was funded by the HESSE Research Award in Ornithology 2013-2014, Wendy Fan Memorial Scholarship, BRITE Fellowship, Dennis H.CHITTY Memorial Graduate Scholarship in Ecology.     xiii Dedication A Alfonso, mi decisión de hacer lo que me hace feliz y mi amor por la selva lo herede de ti, esa es mi mejor herencia A Lucina, en cada centímetro, de cada cosa que hago, tu eres mi inspiración. A Jader, somos diferentes, pero siempre contaras conmigo.  A mis amigos, que son la familia que he ido encontrando.     A Camilo, lamentó no haber contruido un mundo mejor en el que te quisieras quedar.                                                                                                                                                                                                                                                                                                                                                                                                                                                                   A ti.     1 Chapter 1: General introduction  1.1 Introduction to facilitation Understanding the influence of biotic interactions on community structure and species’ range limits has been a long-standing goal in ecology and biogeography (Terborgh 1971; Terborgh and Weske 1975; Araújo and Luoto 2007; Gross et al. 2009; Sexton et al. 2009; Wisz et al. 2013). Among the various forms of interactions (e.g., competition, predation, parasitism, mutualism, facilitation) negative interactions, such as competition and predation, have received the most attention. Competitive interactions can influence distributional limits of species and community assembly, resulting in distributions limited by the presence of a competitor (Connell 1961; Terborgh and Weske 1975; Tilman 1994; Remsen and Graves 1995; Bullock et al. 2000; Jankowski et al. 2010). However, recent studies also suggest that the effect of positive interactions, particularly facilitation, can be just as important as negative interactions in structuring ecological communities and reinforcing range limits (Bertness and Callaway 1994; Bertness and Leonard 1997; Hacker and Gaines 1997; Stachowicz 2001; Cavieres et al. 2002).  For example, experimental studies have shown that species involved in facilitative interactions can expand their elevational range to match their facilitator’s range (Afkhami et al. 2014; Crotty and Bertness 2015). In a broad sense, facilitation can be defined as an interaction between species that benefits the average individual fitness of at least one species without negatively affecting the other species (Hacker and Gaines 1997; Stachowicz 2001). Facilitation can occur when the presence of one species makes the local environment more favourable for another species, by enhancing,    2 directly or indirectly, its growth, reproduction or survival (Stachowicz 2001; Bruno et al. 2003). The outcome of facilitation to any participating organism (the facilitator and facilitated species) can be either neutral or beneficial, but not detrimental. Following this definition, mutualisms would be the subset of those facultative interactions in which participating species have reciprocally beneficial interactions (see Pugnaire et al. 1996).  Facilitation among species in communities has been the subject of increasing interest for community ecologists during the last decades (Michalet and Pugnaire 2016). Although this concept was introduced in the plant community literature a century ago (Pearson 1914; Clements 1916), it was largely neglected in ecological theory by most community ecologists, compared with the attention focused on other mechanisms (but see Bronstein 1994). Several relatively recent attempts have been made to include facilitation as an important mechanism in ecological theory (e.g., Boucher 1985; Bertness and Callaway 1994; Bruno et al. 2003; Michalet et al. 2006; Liancourt et al. 2012; Valiente-Banuet and Verdú 2013), which could highly impact the framework of many fundamental models in population and community ecology. Incorporating facilitation into ecological theory can influence several fundamental concepts in ecology (Bruno et al. 2003). Among those, the niche concept, which predicts where a species can live, is of particular interest. Given that facilitation can allow a species, in the presence of a facilitator, to tolerate conditions that it would otherwise not be able to tolerate (Crotty and Bertness 2015), it can result in the expansion of the realized niche beyond the range predicted by its fundamental niche (Bruno et al. 2003), widening the distribution of the species (e.g., Afkhami et al. 2014; Crotty and Bertness 2015).    3  The occurrence and prevalence of facilitation along environmental gradients have attracted relative attention during the last two decades, after Bertness and Callaway (1994) proposed that the importance of facilitation should increased as environmental or biotic conditions become more stressful for species. Many studies focused in plat communities have found support for this stress gradient hypothesis, and some studies in animal communities confirm similar results (e.g Callaway 2007). However recent studies inspired by this stress gradient hypothesis have suggest different outcomes (e.g Maestre et al. 2009; Holmgren and Scheffer 2010), and is now accepted that no single theoretical model may explain the occurrence of facilitation. The prevalence and role of positive interactions in response to environmental gradients remains highly debated and is specially poorly explore in animal communities. Facilitation is ubiquitous in communities, with facilitative partners found in plants (Callaway 1995), fungi (Afkhami 2012), algae (Hay 1981), coral reefs, sessile invertebrates (Bertness 1989; Bracken et al. 2007), fishes (Pereira et al. 2013), birds (Sridhar et al. 2012), and mammals, ranging from diffuse and indirect interactions to highly integrated and coevolved associations between organisms.  Some well-studied examples of facilitation include associational growth in plants (i.e., via increased access to nutrients, Pugnaire et al. 1996), associational defense (mutual protection from natural enemies, Hay et al. 2004), protection to plants by ants (protection from natural enemies for food reward, Rico-Gray and Oliviera 2007), nutritional symbiosis, and more generally, pollination (Pellmyr 2002) and seed dispersal (Levey et al. 2002). Nevertheless, despite the widespread examples of facilitative interactions across taxonomic groups, most studies have focused on the ecological consequences of facilitation on plant communities (i.e., plant-pollinator, plant-disperser, plant-herbivore); the role of facilitation as a mechanism    4 structuring communities in other taxonomic groups remains poorly explored. Among the studies describing the facilitation in vertebrate communities (e.g., animal-animal interactions), examples include fish schools (e.g., Pereira et al. 2013), mammal troops (Terborgh 1990) and flocking in birds (Thompson et al. 1991; Thomson et al. 2003; Sridhar et al. 2012). However, much remains to be learned about the consequences and importance of these interactions in ecological community structure and distribution of vertebrate species.  1.2 Facilitation in bird communities In birds, facilitation occurs among species that participate in mixed-species flocks (Powell 1985; Sridhar et al. 2012; Palmer et al. 2015). Mixed-species flocks are roving groups of individuals of two or more bird species that obtain benefits from their association with other species (Swynnerton 1915; Morse 1970; Morse 1977; Powell 1985). Birds in mixed-species flocks may benefit directly or indirectly from this association, through shared social information (Satischandra et al. 2007; Goodale et al. 2010), increased foraging efficiency (Hino 1997; Dolby and Grubb 1998; Satischandra et al. 2007) and reduced predation risk (Moynihan 1962; Morse 1977; Thiollay 1999; Sridhar et al. 2009).  Therefore, mixed-species flocks may allow a species to persist in high-predation or low-resource environments, or other harsh conditions where without facilitation by flocks, it would not otherwise persist (Morse 1970). Species joining mixed-species flocks may accrue benefits through a variety of mechanisms (reviewed in Colorado 2013). Reduced predation risk can arise from mechanisms such as the risk-dilution effect (decreased probability to be singled out by a predator; Foster and Treherne 1981), the many-eyes effect (larger groups are more effective in detecting approaching predators; Pulliam 1973; Powell 1974), confusion effect (reduced attack-to-kill ratio of a    5 predator as a result of sensory inability to single out a prey in a group; Krause and Ruxton 2002) and collective defense against predators (mobbing behavior; Vieth et al. 1980), including nest predators (Martinez unp. data). Additionally, increased foraging efficiency for flocking individuals can arise from kleptoparasitism (Brockmann and Barnard 1979), copying (Krebs 1973), easier location of food (Powell 1985) and feeding on insects flushed by other birds (beating effect; Winterbottom 1943). Furthermore, mixed-species flocks can provide unique benefits by gaining information from other bird species, including taking advantage of the complementary anti-predator abilities across species   (Powell 1985) and alarm calls of heterospecifics (Lea et al. 2008). Alternatively, individuals can incur costs associated with participation in mixed-species flocks, such as competition for resources (Goss-Custard 1980; Hutto 1988), kleptoparasitism by other flock members  (Brockmann and Barnard 1979; Munn 1986; Satischandra et al. 2007), and increased conspicuousness to predators as a group (Hutto 1988).  Facilitation occurring among species in flocks varies along a continuum in the benefits that each species provides to others, from mutually beneficial interactions (+ , +), to commensal interactions (+ , 0), which are likely to be specific to species pairs. One of the scenarios implies both species in the interaction facilitating each other, and obtaining a benefit (i.e., species A and B simultaneously decreasing the risk of predation because the dilution effect). Alternatively, species A in the flock can facilitate an attendant species B without incurring a cost  (i.e., flushing insects that other birds feed on). Other scenarios, where species A incurs a cost in the short term when facilitating species B (i.e., alarm call when a predator is close) are also possible. In the long term, however, it is expected that species participating frequently in mixed-species flocks    6 will be those for which the potential fitness benefits of flocking outweigh the costs (Brawn et al. 1995; Jullien and Clobert 2000; Jullien and Thiollay 2001). 1.3 Tropical mixed-species bird flocks Among mixed-species bird flocks, Neotropical mixed-flocks have attracted the attention of ecologists for more than a century. Neotropical flocks exhibit some unique features, such as multi-species defense territoriality (Munn and Terborgh 1979), communal roosting sites (Buskirk et al. 1972), stability over long time scales (Martínez and Gomez 2013) and strong facilitative relationships among member species. Based on isolated evidence, several authors have suggested that mixed flocks may play an important and underappreciated role in tropical bird communities (e.g., Powell 1989; Jullien and Thiollay 2001; Lee et al. 2005; Harrison and Whitehouse 2011). Several studies have suggested that flocks influence birds from the individual to the community level; flock participation might have a positive effect on individual fitness (Jullien and Clobert 2000), influence population density of the participants and generate interdependence among them (Powell 1989). Furthermore, flocks have been proposed as a factor promoting high species diversity in Neotropical avifauna, leading to higher species packing within communities (Powell 1989). Mixed-species flocks are widespread in the Neotropics, occurring virtually in all the habitats from the Amazon to the high Andes. However, comparative studies of flocks across local scales within a region are rare. Most studies collect data at one locality or small spatial scales (e.g Buskirk et al. 1972; Poulsen 1996; Jullien and Thiollay 1998). Studying flocks at a regional scale presents an opportunity to address questions in community organization. To the    7 best of my knowledge, no study has examined the structure and the ecological importance of these flocks along a broad elevational gradient in the Neotropics. 1.4 Study goals The aim of this research is to study facilitative interactions in tropical mixed-species bird flocks along a broad elevational gradient. I use observational data to describe the structure and stability of flocks across elevations, including tropical lowlands, lower montane forest and cloud forest. I examine the habitat factors that influence flock diversity. Finally, I examine the importance of these multi-species flocks for individual species and for the bird community across the gradient.     8 Chapter 2: Mixed-species flocks along an elevational gradient and their importance for bird communities.  2.1 Introduction Tropical ecosystems exhibit the highest species diversity in the world for the large majority of higher-level taxa, including birds (Macarthur 1969; Gaston 2000; Hillebrand 2004). Tropical bird communities exhibit both high species richness and beta diversity (MacArthur and MacArthur 1961; Jankowski et al. 2013) and host numerous endemic species (Myers et al. 2000). For example, a 100 m elevation band in the Andean foothills may contain nearly 300 species of breeding birds (Terborgh 1977), and a single survey plot in the Amazon lowlands can host over 100 species with overlapping foraging territories (Terborgh 1971).  Such heightened diversity on several scales potentially allows for more frequent and complex interactions among species (Schemske et al. 2009). Many complex interspecific interactions have been described in the tropics, including specialized pollinator systems (Bawa 1990), mutualistic defenses in ant-plant symbioses (Davidson and Mckey 1993) and antbirds parasitizing foraging army ants (Wrege et al. 2005). Among the numerous interspecific interactions in the tropics, one striking interaction found in bird communities is the association of individuals of multiple bird species in mixed flocks (hereafter mixed-species flocks). Mixed-species flocks are among the most complex multi-species aggregations found in terrestrial vertebrates (Munn 1985) and have attracted broad interest from tropical ecologists over the last century (e.g. Davis 1946; Buskirk et al. 1972;    9 Buskirk 1976; Munn and Terborgh 1979; Hutto 1987; Graves and Gotelli 1993; Jullien and Thiollay 1998; Greenberg 2000; Sridhar et al. 2009). These multispecific flocks of birds are roving groups of individuals of two or more species that group to forage together and share heterospecific information (Morse 1970). These flocks are distinguished from aggregations of birds that accidentally form when feeding in a localized resource (Powell 1985).  Mixed-species flocks occur in temperate, subtropical and tropical areas (e.g. Powell 1985; Goodale et al. 2009; Sridhar et al. 2012; Goodale et al. 2015), in all terrestrial habitats across the world, but reach their maximum diversity and complexity in tropical forest (Munn 1985). In some tropical forests, mixed-species flocks dominate entire bird communities, where as many as one third of local species join these flocks (Latta and Wunderle 1996; Jullien and Thiollay 2001). These tropical flocks can consist of up to 80 species and more than 100 individuals (Munn 1985). Moreover, these flocks are not randomly drawn from the community. Instead, species tend to associate in flocks with other species that are phenotypically similar, for example in body size and foraging behaviour (Sridhar et al. 2012).  The pervasiveness of mixed flocks in the tropics (Greenberg 2000) and the broad range of species utilizing this flocking strategy at local scales may indicate that they play an important role in higher-order ecological patterns such us community structure and distributional patterns of birds (e.g. Powell 1985; Powell 1989; Jullien and Clobert 2000; Jullien and Thiollay 2001; Harrison and Whitehouse 2011). Other features of tropical flocks that make them interesting to study, particularly for the partitioning of ecological roles, include interspecific alarm calls and responses (Munn 1986), multi-species territorial defense (Munn and Terborgh 1979), communal roosting sites (Buskirk et al. 1972), collective defense against predators (e.g., nest predator    10 defense, Martinez unp. data), mobbing behaviour (Courter and Ritchison 2012), and stability over long time scales (e.g, decades; Martínez and Gomez 2013). Despite the great interest in mixed-species flocks over the last century and the recognition of these flocks as prevalent characteristic of tropical communities, our knowledge of their ecological importance is still quite limited. Information of flocks has been collected at relatively narrow elevational spatial scales (e.g., Arbeláez-cortés and Marín-gomez 2012), discontinuous elevations (e.g., Goodale et al. 2009; Marín-Gómez and Arbeláez-Cortés 2015), short periods of time, and individuals that have not been colour banded (Goodale et al. 2015). Therefore, assessments of the ecological consequences and importance of flocks for broader tropical bird communities are restricted to small spatial and short temporal scales (e.g. Powell 1989).  Furthermore, whereas numerous studies have examined different ecological aspects of tropical flocks at low elevations (i.e., composition, structure, stability over time; (Willis 1958; Munn and Terborgh 1979; Munn 1985; Graves and Gotelli 1993; Stouffer and Bierregaard 1995; Develey and Peres 2000; Jullien and Clobert 2000; Maldonado-Coelho and Marini 2000; Ragusa-Netto 2002; Martínez and Zenil 2012), less is known about these aspects of flocks at higher elevations (but see Davis 1946; Buskirk et al. 1972; Powell 1979; Merkord 2010; Arbeláez-Cortés and Marín-Gomez 2012). Previous studies of lowland mixed-species flocks have documented cohesive groups that exhibit high diversity and long-term stability (i.e., core-member composition, home range boundaries), sometimes over decades (Martínez and Gomez 2013); by comparison, studies at higher elevations suggest a tendency for montane flocks to be less diverse (Moynihan 1962; Arbeláez-Cortés and Marín-Gomez 2012), more dynamic, and in    11 some cases, they do not appear to hold permanent territories (e.g., non-Amazonian flocks; Stotz 1993; Hart and Freed 2003). Although there is some evidence in the literature suggesting temporal stability in montane flocks (e.g., Buskirk et al. 1972; Powell 1979; Merkord 2010), the overall lack of data available for tropical montane mixed-flocks, and the limited information on the effects of elevation (but see Goodale et al. 2009; Marín-Gómez and Arbeláez-Cortés 2015) and other environmental variables (i.e vegetation structure), have led to the idea that stable, complex tropical mixed-species flocks are mainly a low elevation phenomenon.  2.2 Research questions In this study, I examine how mixed-species bird flocks influence the structure of bird communities across a broad forested Neotropical elevational gradient within one of the world’s foremost biodiversity hotspots, Manu National Park, Peru. First, I examine how mixed-species flocks structure change across the elevational gradient. Second, I quantify the stability of these multispecies groups over time (i.e., for three different temporal periods) across the gradient, using data on species composition and home range locations for colour-banded flocks. Third, I evaluate the association of several key environmental variables (i.e. canopy height, vegetation density) with flock diversity along the gradient. Finally, utilizing a dataset for the entire forest bird community, I assess the prevalence and importance of mixed-species bird flocks for the bird community across the gradient.      12 2.3 Materials and methods 2.3.1 Study site This research was conducted along an elevational gradient on the eastern slope of the Andes in the buffer zone of Manu National Park, Peru. The gradient encompassed elevations between 400-3500 m, extending from lowland forest (<800 m, floodplain forest, terra firme), through premontane forest (800-1200m), cloud forest (1200-2200m) and upper montane forest to puna grassland (2200-3400m). The forest present in the area is mainly primary forest, with small patches of older secondary forest adjacent to the unpaved, narrow Manu road, which runs along the southern border of the park from treeline to the lowlands. The forest is a patchwork of different canopy heights even at similar elevations, created by the geography of the terrain, windswept ridges, landslides and the presence of bamboos patches. Annual temperatures means along the elevational gradient ranged from 11.2 °C in the montane forest to 23.2 °C in the lowland forest (Londoño unp. data).  2.3.2 Data collection mixed-species flocks A mixed-species flock was defined as a group of individuals of at least two species foraging and moving together within 15 m distance from their nearest neighbor for at least 10 minutes (Stotz 1993). Mixed-species flock data was collected over two field seasons from July to October 2013 and August to October 2014. Systematic searches for flocks were conducted daily from dawn until dusk along the elevational gradient (400-3300m), following trails in areas of primary forest and in some cases older secondary forest. Once a flock was detected, it was followed as closely as possible (i.e., 10-40 m) for at least 30 minutes until all species were    13 identified (i.e., no new species were detected for 10 min) up to a maximum of 60 minutes. One hour was, in general, enough time to fully characterize even larger flocks and is within the time window previously suggested by other observational studies to characterize flocks (Goodale et al. 2009).  Data collected from each flock included songs and calls, species composition, and when possible, flock structure (i.e., number of individuals per species) and individual colour band identification (see section below). For all flocks, I registered the latitude/longitude location and elevation with a GPS unit (Garmin 62s) every 15 min. Flocks that were not fully characterized or that were lost by the observer more than 15 min out of 60 min were excluded from analyzes. To assign flock independence, I assumed flock territories to be circular with a diameter of 400m and 800m, for the lowlands and montane flocks, respectively (based on Jullien and Thiollay 1998, Martínez and Gomez 2013, pers.obs). Therefore, flocks observed >400 (i.e., lowland flocks) or >800 m (i.e., montane flocks) from previously observed flocks in sequential or non-sequential observations were considered independent. Neighboring flocks found closer than this defined circular territory were included only if it was certain that it was a new flock (i.e., by citing colour banded individuals in the flocks).   2.2.3 Colour marking This study was conducted as part of a larger project that evaluates the factors that determine species range limits along an elevational gradient. As part of this study, banding data were collected along the elevational gradient in four field seasons from August to December, 2011-2014. Nets were located at ground level in 50 netting stations along the gradient. Each    14 netting station was run for three days, at least twice each year (within the five-month field season). Individual birds of focal forest species participating in flocks were captured with mist-nets and uniquely colour-marked. The colour-banded scheme consisted of one metal numbered band in one leg and two celluloid colour bands on the other leg. Colour banding of individuals allowed me to identify individual flocks and to monitor the same flock over time. As I was not able to mark all individuals in each flock, I considered a flock to be the same if it was detected within the flock’s territory and if at least 1-2 individuals of different species within the flock were colour banded. A total of 6553 individuals were just metal banded, and 882 individuals were metal and colour banded. During the banding process, we collected morphometric measurements including body mass, tarsus, bill length, and wing chord.   2.3.3 Species participation in mixed-species flocks Bird species were classified into four categories using their observed participation in flocks and based on detections from census points and mist netting at the study site (Jankowski unpubl data). Each bird species was categorized as an accidental, occasional, regular, or obligate participant of mixed-species flocks using a combination of two metrics occurrence and propensity to calculate a Flocking Index that weights the participation of each species in flocks by their abundance in the study site:                      Flocking Index= (Occurrence*Propensity)  Occurrence is the frequency in which a species occurs in flocks, calculated as the number of times a species was observed foraging in mixed-species flocks divided by the total flocks of    15 that type sighted (e.g., lowland flocks, low montane flocks), as follows: Occurrence = (# detection sp. A in flocks/ # total flocks of that type). Occurrence values range from 0-1, where 0 indicates a species that was never observed in flocks and 1 indicates a species observed in all flocks. For occurrence calculation, the presence of a species was considered as a time, the number of individuals was not considered. Propensity is the frequency in which a species uses the flocking strategy. It was calculated as the number of times a species was observed foraging in mixed-species flocks divided by the number of times the species was detected overall (i.e. in mixed-flocks, monospecific-flocks and solitary combined). For this calculation each individual of a species detected was considered as a time. Propensity was calculated using data from flock observations and previous point count surveys, using the formula: Propensity = (# times sp. A observed in flocks/ # times sp. A detected).  To establish the final categories of species based on the Flocking Index, I used the following groups: a) Obligate flocking species were species that permanently associate in flocks, having a Flocking Index >0.6. These species exhibit a high occurrence and propensity. Groups of two or more obligate participants in flocks form the "core" of the flock. b) Regular flocking species often follow flocks beyond their territories but also forage independently of mixed flocks. They may leave the flock several times during the day. These species exhibit a Flocking Index between 0.30-0.59.    16 c) Occasional flocking species were species commonly detected outside the flocks and found in flocks only briefly and for short distances.  These species exhibit a Flocking Index between 0.05-0.29. d) Accidental flocking species were mostly found outside the flocks and detected within flocks for short periods of time and on very few occasions, probably passing through the flock territory. These species exhibit a Flocking Index <0.049.  2.3.4 Foraging guild classification Bird species were classified into foraging categories using their observed and documented diet (Del-Hoyo et al. 1992, 1994, 1997, 1999). Each bird species was categorized as frugivorous, insectivorous, nectarivorous or omnivorous.   2.3.5 Flock stability Temporal flock stability was defined as the proportion of species that are consistent in a given flock between two observations separated by a time period. Temporal flock stability was measured at three different time scales, comparing the species composition of a flock between two observations separated by 3-5 hours, 6-15 days and one year. For example, I compared the species composition of a given flock in 2013 with its composition in 2014. Similarly, I compared the species composition of a flock observed in day 1 with its composition 8 days later.  I identified individual flocks using colour-banded individuals, which allowed me to compare the same flock over time. The calculation of stability was limited to flocks that contained at least one colour-banded individual from each of two different species in the same    17 home range, or if the home range was not known, less than 200 m from its first observed location. To collect data on flock composition, flocks were followed as closely as possible during 60 min intervals. During this interval time, I collected data on species composition, number of individuals, colour banded individuals, songs, and flock movement.   2.3.6 Home range characterization Home range of focal flocks (n=7) across the gradient was characterized using a one-day sample. Each of the focal flocks was followed from dawn (6:00 am) to dusk (5:00 pm). I collected data on species composition, number of individuals and individual identity (when colour-banded) for 60-minute intervals. The exact location of the flock was georeferenced every 15 minutes. The home range of each mixed flock was mapped using a minimum of 30 georeferenced locations, and the area was measured using the minimum convex method in QGIS version 2.8.2 (ESRI 2012).     2.3.7 Vegetation structure Vegetation structure was characterized using a protocol adapted from Martin et al. (1997). Vegetation structure, including vertical and understory structure, was measured within each mixed-species flock territory, in the location where the flock was first encountered. Forest vertical structure was characterized in a 20 x 20 m plot at each flock territory using five variables: canopy height, number of trees, and percent cover of canopy, bamboo and epiphytes. Canopy height was measured using a rangefinder (Nikon Prostaff 3) and calculated as the average of canopy height in the center of the plot. Number of trees was estimated by counting    18 trees >10 cm diameter at breast height (d.b.h.). Canopy, bamboo and epiphyte cover was calculated as a percentage of cover over the total area of the plot using ocular estimation (i.e., <25%, 25-50%, 50-75%, 75-100%).  Forest understory structure was characterized in a 10 x 10 m plot at each flock territory by the number of small steams (< 2.5 cm d.b.h.; >1m height).  2.3.8 Prevalence of flocking at the community level I compiled a database of the bird community along the Manu elevational gradient (i.e., species found across all elevations in the study area) using data from previous point count surveys and mist netting data collected from 2006 to 2014 (Jankowski unpubl. data). I collected additional data of species occurrence from July to Nov 2012-2014, using automated field scan recorders (Songmeter SM2), flock observations and occasional detections. The automated field scan recorders recorded songs for 20 minutes every hour from dawn to dusk every day. Recorders were placed in forest locations separated by 100 m in elevation and were moved to new locations along the gradient every three days to cover all elevations. The species identified follow the most recent updated taxonomy from the current version of the South American Classification Committee (Remsen 2015). I combined bird species presence data collected by the different sampling methods to create an elevation by species presence matrix for the community. For the purpose of this analysis, the community was divided into 100-m elevational zones along the gradient (e.g., 300-399, 400-499 masl) from 500 to 3400 elevation, and the community composition in each elevational zone was calculated. In addition, I calculated the number of species observed in mixed flocks in each elevational zone (e.g., 300-399, 400-499 masl).    19  To examine the prevalence of flocking (FP) at the community level, I calculated the proportion of species in the community joining flocks at each elevation. Flocking prevalence (FP) in the community ranges from 0-1, where 0 indicates zones in which no species participated in flocks and 1 indicates zones in which all species present in that community joined flocks.  Flocking  prevalence(FP) = #spp  detected  in  flocks  within  a  given  elevational  zone#  spp  detected  within  a  given  elevational  zone     2.3.9 Data analysis To group mixed-species flocks along the gradient into different flock types, I performed a Cluster analysis in R package ‘vegan’ (Oksanen et al. 2011). I used the average linkage agglomerative method (UPGMA) to perform a hierarchical clustering analysis of flocks (Mirkin 2011).  The hierarchical cluster distance was based on the Jaccard dissimilarity Index, commonly used for presence-absence based community comparisons (Krebs 2014). The Jaccard index ranges from 0 (similar) to 1 (dissimilar).  The hierarchical cluster analysis based on flock composition dissimilarity was used to inform the division of flocks into distinguishable groups (i.e., flock types) along the gradient.   To test for differences among flock types in mean richness and size, I fit a generalized linear model (GLM) to the data and performed an ANOVA test (type=III), F-test. In this analysis, flock type was included as a factor (fixed effect), using a quasi-Poisson error distribution (i.e., given the overdispersion of the variance) and a log link function (Ver Hoef 2007).    20 Temporal flock stability (i.e., flock similarity over time) was examined using the Jaccard dissimilarity index to compare species composition of a given flock between two observations (29 flocks between 2013 and 2014; 17 flocks observed 6-15 days later; 6 flocks in a given hour to their composition 3 hours later). Data were plotted using 1-Jaccard dissimilarity index to show the similarity of each flock over time. To test for differences in stability among time periods, I fit a generalized linear model (GLM) to the data and performed an ANOVA test (type=III, F-test). In this analysis, flock time period (i.e., years, days, hours) was included as a fixed effect, using a binomial error distribution and a logit link function. To examine the effect of elevation on flock stability, I performed a Generalized Lineal Model (GLM) with time period (i.e., years, weeks, days) as fixed effect, using a binomial error distribution and logit link function. The home range area of each mixed flock was estimated using the minimum convex method in QGIS version 2.8.2 (QGIS Development Team, 2012).  The percentage of overlapping home range area between 2013 and 2014 was calculated in ArcMap. To examine the effect of elevation and vegetation structure (i.e., canopy height, number of trees) on flock species richness, I fit a Generalized Linear Model (GLM) with quasi-Poisson error distribution and log link function (Ver Hoef and Boveng 2007). To evaluate and compare the relative fit of alternative models to the data (e.g., canopy height, canopy height + elevation, canopy height + elevation + trees), I used the modified version of Akaike’s Information Criterion for overdispersed count data, (QAIC) Quasi-Akaike Information Criterion (Burnham and Anderson 2002; Bolker 2016) where the quasi likelihood adjustment is calculated. I fit each model twice, once with a Poisson error distribution and once with a quasi-Poisson error distribution, and then extracted the over dispersion parameter manually. For each of the models,    21 I calculated the QAIC’s value in R package ‘AICcmodavg’ (Mazerolle 2012), and then used those quantities to calculate the QAIC weights (range 0-1) for each fitted model. The best model was selected considering the lowest QAIC’s. The relative importance of each predictor was evaluated by summing the QAICw for each model in which that variable appears. These summed weights were used to rank the various predictors. To examine avian foraging guilds participation in mixed species flocks across elevations, I fit a Generalized linear model with quasi-Poisson error distribution and log link function, with foraging guild included as a fixed effect. To examine the proportion of species participating in flocks by foraging guild I fit a GLM with binomial error distribution and logit link function. Finally I fit a Generalized linear model with quasi-Poisson error distribution and log link function to examine the prevalence of flocking at the community level. For each model used for analysis, I assessed the model assumptions of overdispersion, influential observations and autocorrelation of the data. All analyses were done in R (R Development Core Team 2015)  2.4 Results  2.4.1 Spatial distribution of mixed-species flocks  Mixed-species bird flocks were found across the elevational gradient. I obtained a total of 210 independent mixed-species flock observations over two years, with 99 observations from the first field season (July to October 2013) and 111 from the second  (August to October 2014). Cluster analysis based on flock composition dissimilarity identified three major distinguishable    22 clusters (hereafter flock types; Fig. 2.1). Each of the flock types was represented by a distinct species composition and was broadly associated with a different forest type.  The flock types identified were lowland flocks (300-1100 m, n=95), low montane flocks (1100-1900 m, n=55) and high montane flocks (2300-3500 m, n=50). Transitions between flock types occurred in the elevational zones of 1000-1100, 1900-2000, and 2200-2300 m.a.s.l. respectively. A potential fourth type of flock was identified in the range of 1850-2200 m.a.s.l., these data could not be included in the analysis. Additionally, the cluster representing lowland flocks identified three subgroups: understory flocks in terra firme forest, understory flocks in bamboo forest and canopy flocks in both terra firme and bamboo forest. Here I focus on the three flock types identified in the three main clusters: lowland, low montane, and high montane flocks.  Mixed-species flock richness (number of species) and size (number of individuals) differed among flock types (ANOVA p=0.002, n=210, F=6.13, Fig. 2.2; p<0.001, F=9.96, Fig. 2.3). Low montane flocks were larger (20.6 ± 10.7 individuals) and more diverse (12.3 ± 5.6 species) than lowland flocks (14.2 ± 9.4 individuals; 9.3 ± 5.7 species) and high montane flocks (18.9 ± 8.5 individuals; 9.39 ± 4.21 species). However, when considering the three different subgroups of lowlands flocks, lowland canopy flocks were larger and more diverse than any other flock type (21.4 ± 13.1 individuals; 14.0 ± 7.7 species; Table 2.1). Flock richness and size were strongly and positively correlated (Kendall’s Tau= 0.789, p<0.001, Fig. 2.4). The number of species per flock exhibited as much variation within a given elevation as was found across elevations.    23 2.4.2 Species participation in mixed-species flocks A total of 273 species associated with mixed flocks to some degree. Using the Flocking Index, I identified 19 spp. as obligate participants of flocks, with a high Index (0.6; Table 2.2), 40 spp. as regular species, 169 spp. as occasional species, and 45 spp. as accidental species or species with too few registers to be informative (Appendix A). The species identified, as accidental flock followers were those detected mostly outside the flocks and detected within flocks for short periods of time, on very few occasions, when the flock was passing through their territory. Accidental followers included species with small territories, such us manakins and flycatchers, and it was uncertain whether their presence was merely accidental. The species classified as occasional participant of flocks, were species seen in flocks only briefly and for short distances such us tanagers and wood-creepers. These occasional participants were commonly detected outside flocks in the study area. The 40 species classified as regular flocking species often follow flocks beyond their own territories but also forage independently of mixed flocks. They leave the flock several times during the day and in some cases used different flocks. This was the case with species known to occupy larger territories and utilize patchily distributed resources (i.e., fruits), including species from genera such us Chlorornis, Cotinga, Buthraupis, Monasa, Tangara, Turdus, and Xiphorinchus.  The 19 species classified as obligate participants permanently associated in mixed-species flocks and were rarely detected foraging solitarily, even during the breeding season. These species exhibited a high occurrence or propensity to forage in flocks, and usually both. Groups of two or more obligate participants in flocks formed the "core" of the flock that showed    24 high stability in composition over time (Table 2.2). The core group of species for each mixed-species flock types identified were: a. Lowland terra firme flocks: Thamnomanes ardesiacus, Thamnomanes schistogynus, Myrmotherula axillaris, Myrmotherula menetriesii b. Lowland Bamboo flocks: Thamnomanes schistogynus, Microrhopias quixensis, Anabazenops dorsalis c. Lowland Canopy flocks: Lanio versicolor, Myrmotherula axillaris,  Tachyphonus rufiventer, Tangara schrankii, Tangara chilensis d. Low montane flocks: Myioborus miniatus, Chlorospingus flavigularis, Leptopogon superciliaris, Chlorochrysa calliparaea, Tangara arthus e. High montane flocks: Myioborus melanocephalus, Mecocerculus stictopterus, Hemispingus atropileus.  2.4.3 Foraging guild participation in flocks  Analysis of individual foraging guilds revealed that flocks across the gradient were composed mainly by insectivorous birds species (Fig. 2.5a). The same pattern of high representation of insectivorous birds was observed for the overall avian community in Manu elevational gradient (Fig. 2.5b). Similarly, the representation of omnivores, frugivores and nectarivores in mixed-species flocks followed the pattern of species representation of each guild in the overall avian community  (Fig. 2.6; Appendix F). Moreover the relative participation of foraging guilds revealed that insectivorous and omnivorous species participated in mixed-species    25 more than other guilds at low elevations and frugivorous species participate relatively more at high elevations (Fig. 2.7).      2.4.4 Flocks stability  Flock stability over time was estimated for 52 independent colour-banded flocks. I compared the composition of 29 flocks observed in 2013 to their own composition in 2014, the composition of 17 flocks observed in a given day to their composition 6-15 days later, and the composition of 6 flocks in a given hour to their composition 3-5 hours later.  Temporal flock stability showed that flocks have an average similarity index over time of 0.34 ± 0.12, suggesting that flocks maintained 34 % of the species consistent between observations. Flock similarity ranged from 0.14 to 0.64 (mean 0.36 ± 0.12, n=29) across years, 0.14 to 0.55 (mean 0.29 ± 0.11, n=17) across weeks, and 0.14 to 1 (mean 0.51 ± 0.19, n=6) across hours (Fig.2.8).  Flock similarity did not differ across time scales or flock types (ANOVA, n=52, p=0.17, F=1.86; p=0.12, F=2.21). Furthermore, flock similarity did not show a clear trend with elevation (Fig. 2.9), suggesting that high montane flocks are as stable and cohesive over time as low montane and lowland flocks (Fig. 2.10, Fig 2.11). Flocks at any elevation across the gradient were composed of two identifiable parts: a core subset of a few species that is retained over time, and a non-core group of several attendant species that changes in composition over time and generates fluctuations in flock composition.      26 2.4.5 Home range and territoriality  Using information from colour-banded individuals, I found that mixed-species flocks along the gradient maintained the same home ranges over time. Home ranges of lowland  (n=3), low montane  (n=2) and high montane flocks (n=2) that were extensively characterized in one day sample (i.e., >30 georeferenced locations), overlapped from 71- 89 % (80.42 ± 5.96) between 2013 and 2014 (Fig. 2.12), whereas the home ranges of these flocks overlapped on slightly with neighboring flocks. Home ranges of most mixed-species flocks were not fully characterized (i.e., 2-3 georeferenced locations) and for these flocks the overlap between years cannot be described. However, most of these color-banded flocks were spotted < 200 m from their initial observed location, days later (n=74) and a year later (n=53), suggesting that these flocks likely maintain the same home range over time, both within and between years.   Home ranges of flocks differed along the elevational gradient. Lowland flocks used a smaller area (7.3 ha +/_0.8, n=3), compared with the larger areas used by low montane flocks (16.4 +/_1.2 ha, n=2) and high montane flocks (17.1 +/_1.6 ha, n=2). The boundaries of the home ranges, in general, did not appear to be defined by topographic features of the landscape, although in some of the low and high montane flocks, the boundaries of home ranges were sometimes coincident with large rivers (e.g., Rio Kosñipata, Rio Piñi-Piñi), small creeks, and trails.  Flocks exhibited territorial disputes with other flocks along most of the elevational gradient. Direct territorial interactions (i.e., aggressive flights) were frequently observed between congeners of neighboring flocks in lowlands, and sometimes in low montane flocks up to 2200 m.a.s.l., but were never observed directly in high montane flocks. Mixed-species territorial    27 disputes in low montane flocks, much like those in lowland flocks, included close approaches, songs, calls, and aggressive flights back and forth, sometimes lasting for up to 20 minutes. Flocks at different elevations along the gradient exhibited communal roosting sites in the flock’s home range. In these areas, individual members of the flock began their activities in the morning and coalesced every day. The roosting sites in low montane flocks and high montane flocks extended in an area up to 1200 m2, whereas lowland flock roosting sites were more compact. The flock’s roosting area (n=10) was maintained from 2013 to 2014 (e.g., Fig. 2.12).   Coalescence of the flocks in the roosting sites occurred in the predawn and initiated with loud vocalizations of several individuals from one or two species, followed by calls and songs of several other species. Lowland flock coalescence was initiated by vocalizations of individuals from Thamnomanes schistoginus, T.asdesiacus, Myrmotherula axillaris and Chlorothraupis carmioli  (if present) in terra firme forest, and Thamnomanes schistoginus and Microrhopias quixensis in bamboo forest. Low montane flock coalescence was initiated by vocalizations of individuals from Chlorospingus flavigularis and Tangara arthus and Chlorocrysa calliparaea , and in high montane flocks, coalescence was initiated by calls of individuals from the Hemispingus genus. Following these initial vocalizations, the vocal activity in the area increased and was maintained for around 10 minutes.  2.4.6 Elevation and vegetation structure  Mixed flock species’ richness along the gradient was best explained by a model that included canopy height and number of trees (QAICcWt=0.63, Table 2.3). However, a model including three predictors (i.e., elevation, canopy height, and number of trees) had a Delta    28 QAICc value of 2.13, so essentially as good as the best model. Among these predictors, number of trees was the most important variable ( 𝑄𝐴𝐼𝐶𝑐𝑊𝑡=1) and appears in all the top models. Canopy height was the second most important variable ( 𝑄𝐴𝐼𝐶𝑐𝑊𝑡=0.85), and elevation ranked last ( 𝑄𝐴𝐼𝐶𝑐𝑊𝑡=0.30). As a general pattern, mixed-species flock richness increased with an increase in the number of trees and with increasing canopy height; elevation explained little variation in flock species richness (Fig. 2.13).  2.4.7 Prevalence of flocking at the community level The Manu regional bird community consisted of 550 species, which were detected along the elevational gradient by one or more survey methods (i.e., mist-netting, recordings, survey points). A subset of 273 species participated in mixed flocks to some extent (~49% of species in the Manu region). Species richness decreased with increasing elevation in the overall bird community (r=-0.6 p<0.001) and in flocks (r=-0.4, p<0.001; Fig. 2.14). Analysis of flock participation by elevational zones showed three peaks of high prevalence of flocking: for the lowlands (400 - 500 m.a.s.l.); cloud forest (1600 - 1700 m.a.s.l.); and high montane forest (2600 - 2800 m.a.s.l.; Fig. 2.15).  2.5 Discussion Here I present the first study of mixed-species flocks across a continuous forested elevational gradient in the Neotropics. These results showed structured mixed-species flocks occurring across elevations, from Amazon lowland rainforest to Andean treeline. Mixed flocks exhibited relatively high stability in species composition over time at low elevations, with a portion of the    29 flock maintaining the same membership, as previously suggested by other studies (e.g., Martínez and Gomez 2013), but also in mid and high elevations. Furthermore, flocks exhibited stability in home range boundaries over time and communal flock territoriality across the elevational gradient. Vegetation structure explained 63% of the variation in flock richness along the gradient, with number of trees and canopy height as predictors, where number of trees had a higher relative importance. Overall, mixed-species flocks were widely utilized as a foraging strategy for birds across elevations, by ~40% bird species in lowland Amazon forest, as previously documented, but also within low (~39%) and high montane forest (~35%), where flocks are just as prevalent in the community as in low elevations. These results highlight the importance of these multi-species interactions for tropical bird communities across elevations.  2.5.1 Spatial distribution of flocks  Flocks occurring along the elevational gradient in Manu fell into three main types based on cluster analyses: lowland, low montane and high montane. Lowland flocks were further differentiated between habitats and forest strata, in bamboo and terra firme understory and canopy flocks. Each flock type features a distinct group of species, mainly determined by the core obligate members. Interestingly, the locations along the gradient where flock types transitioned (i.e., 1100-1200, 1700-1800, 2200-2300), were largely consistent with the elevations that exhibit peaks of high turnover in both bird and tree communities (e.g., 1100-1200; 1700-2000, 2000-2250 m.a.s.l) (Jankowski et al. 2013).      30 The lowest elevation flock transition occurred in the lower limit of montane forest (1100-1200 m.a.s.l), in foothill elevations. At this elevation both lowland and low montane flocks can be observed within 100m of each other, without overlapping territories. This change in flock composition might be determined by shifts in vegetation, specifically the high turnover in tree composition that occurs at this elevation (Jankowski et al. 2013). At this elevation there is also high bird species turnover of the overall community, possibly responding to this change in vegetation composition. The second region of flock transition occurs in montane cloud forest (1700-1800 m). At this elevational zone there is again a high turnover in tree species composition that notably matches the transition in flock composition. Finally, the third region of transition of flocks occurs at 2200 - 2300 m, the peak of bird species turnover in the overall community may also drive the transition at this particular elevation. The congruence in the location of turnover peaks along the gradient between the overall bird community and mixed-species flock suggests that flock composition is strongly aligned with overall transitions in the bird community, which has been shown to be associated with changes in vegetation structure and tree composition.  It has been suggested elsewhere that species composition of mixed flocks changes with the composition of the overall bird community (Hutto 1994; Goodale et al. 2009; Péron and Crochet 2009). One recent study of high elevation flocks in Colombia, however, did not find shifts in flock composition with elevation (Arbeláez-Cortés and Marín-Gomez 2012), but this could be due to the relatively narrow range of elevations surveyed (3000-3450m). It is also possible that mixed-species flocks, birds and vegetation communities are responding in a similar fashion to other factors of the gradient that vary with elevation, such as temperature and productivity (Tilman et al. 1997). Broader sampling of other gradients and    31 taxonomic groups will be required to better understand community transitions and separate these alternative factors.  2.5.2 Species participation in mixed-species flocks Flocks along the gradient exhibited a similar general structure, with a group of core species that was permanently associated with the flock, and a more dynamic group of attendants that changed over time. This structure was similar to what has been documented for other Neotropical flocks, including Munn and Terborgh (1979) and Graves and Gotelli (1993) in Peru, Powell (1985) in Costa Rica, Hutto (1994) in Mexico, and Jullien and Thiollay in French Guiana (2001). The core of the flocks was usually composed of three to five species pairs or small groups that stayed constantly in the same flock over time, even between years, as evidenced by colour banded individuals. The dynamic component of the flocks was composed of dozens of species pairs or individuals that join for varying lengths of time each day and included regular, occasional and accidental flock participants.  The obligate participants that formed the core of the flock exhibited several behavioural and morphological features described for nuclear species (i.e., species that maintain the cohesion of the flocks). For example, most of these species were conspicuous and vocal, which also tended to forage in intraspecifically gregarious groups within the mixed flock: Chlorospingus flavigularis Hemispingus atropileus, Thamnomanes ardesiacus, Thamnomanes schistogynus, Myrmotherula axillaris, Myrmotherula menetriesii, Lanio versicolor, Tachyphonus rufiventer, Tangara chilensis, Tangara arthus, Myioborus melanocephalus , Myioborus miniatus and Microrhopias quixensis. Other obligate participants were not evidently intraspecifically    32 gregarious, nor conspicuously vocal, such as, Leptopogon superciliaris, Mecocerculus stictopterus, Anabazenops dorsalis, Chlorochrysa calliparaea., and their role as  nuclear species maintaining the cohesion of the flock was less evident.  The foraging behavior of the obligate flock participants ranged from active searching foragers to less active foragers (i.e., sit-wait), and most of the species were insectivorous or omnivorous that foraged in the mid to high forest strata and towards the ends of branches in trees and vegetation. One exception to this pattern was the nectarivorous species, Digglosa cyanea. Morphologically, these obligate species have a smaller body mass than the average for all flocking species (data not show) and exhibited plumage coloration with any combination of yellow, green, gray and brown, with one exception being Tangara chilensis, which exhibits brilliant colouration. Similar patterns of resembles in color among the species of the black, yellow and brown (social mimicry), has been described in plumages of other obligate flocking species by Moynihan (1968) in Panama, as a potential adaptation that allow positive interactions within the flock (e.g risk-dilution effect).  2.5.3 Mixed-species flocks richness and foraging guild participation  The number of species per flock was highly variable, exhibiting as much variation within a given elevation as was found across elevations. Thus, flock size variation was not explained by elevation itself. However, when analyzing flock types, low montane flocks were in average larger and more diverse compared with flocks at other elevations. This result differs from the existing body of work on mixed-species flocks, which suggests lowland flocks are larger and more diverse (Reviewed in Goodale 2009). The larger size in montane flocks could be driven by    33 a peak in bird diversity found at the lower montane forest (Jankowski, unpublished data). However, it can also be related to the high productivity at this elevational range in Manu gradient, compared with other elevations, as described by Marthews et al. (2012). High productivity might allow a higher diversity of bird species from different foraging guilds to join the flocks without incurring costs from higher competition. In terms of guild structure, we found that insectivorous species participate in flocks more than any other guild across elevations. Similar patterns have been described for the structure of tropical flocks at smaller spatial scales by Moynihan (1962) , Munn (1985) and Srinivasan et al. (2012). However, we also observed the same pattern of higher number of insectivores birds compared with other foraging guilds for the overall bird community in Manu (Fig. 2.10.b) as previously suggested by Jankowski et al. (2013) and  described in other tropical gradients (e.g., Terborgh 1971). Thus foraging guild composition in flocks appeared to be merely reflecting the overall availability of bird species in the Manu community. However, after controlling the observed patterns of guild participation by the number of species in the community, the results suggested that a higher proportion of insectivorous and omnivorous species joined flocks at low elevations compared with other guilds.  This could suggest that at low elevations (i.e < 2000 m) the strategy of joining flocks is more important for species utilizing evenly distributed resources, such as insects, compared with more localized and patchy fruit and nectar resources. Interestingly, I found that at high elevations (i.e >2000) a higher proportion of frugivorous species joined flocks compared with other guilds.  Overall our results suggested that flock guild composition is not merely reflecting the proportions of the guilds in    34 the community as a large, instead is a specialized foraging strategy used mainly by insectivorous and omnivorous species at lower elevations and by frugivores species at higher elevations.  2.5.4 Flock stability  In this study, I showed that flocks along the gradient from the lowland Amazon to the high Andes exhibited highly stable member species composition and home range boundaries over time. These results are consistent with previous studies of lowland mixed-species flocks (e.g., Munn and Terborgh 1979; Jullien and Thiollay 1998), including recent research that demonstrated long term stability (i.e., over two decades) of territories and species composition in lowland flocks in French Guiana (Martínez and Gomez 2013) and a similar study over eight years in Panama (Greenberg and Gradwohl 1986). Importantly, the results of my study show that the high stability in core member composition and home range boundaries in flocks occurring at low elevations in the Amazon, as previously suggested by the studies mentioned above, extends to mid- and high-elevation flocks in the Andes.   The stability of Andean mixed-species flocks in both member species composition and home range boundaries, similar to Amazonian lowland flocks, is in contrast to other studies. For example, Hart and Freed (2003) found that flocks at middle elevations in Hawaii exhibited unstable membership. Work by Stotz (1993) in the Atlantic forest in Brazil described that non-Amazon flocks observed in the same location on different days were very different in composition, and did not appear to hold permanent territories, suggesting that stability was a feature of Amazon flocks. Similarly Poulsen (1996) described Andean flocks in Ecuador as more dynamic and unstable than Amazonian flocks. Overall my study highlights that the stability in    35 species composition, home range boundaries and roosting sites are a widespread feature of Neotropical mixed-species flocks along the gradient and are not a distinctive feature of Amazonian flocks. This stability also differentiates Neotropical mixed-species flocks from mixed flocks in the Old World Tropics where a study along an elevational gradient by Goodale (2009) found no evidence of interspecific territoriality or stability over time.  The temporal stability of flocks, in both member composition and home range boundaries, might have important implications for the core species that associate permanently in the same flock. For instance, stability in flocks may promote interdependence among species, thus allowing for the potential rise of evolutionary stable strategies. Complex behaviors among species can arise from such stable flocks, given sufficient periods of time for selection to act on species’ traits. Some complex behaviors that have been described in flocks indicate interdependence among species, including multi-species territory defense (Munn and Terborgh 1979), interspecific alarm calls and responses (Munn 1986) and collective defense against predators (mobbing behaviours, Courter and Ritchison 2012). In this study I found that these complex behaviors are common in flocks from the Amazon basin to the high Andes, and also found evidence that other strategies, such us cooperative breeding of the species, can arise from such stable associations in mixed-species flocks (Munoz personal observation). Furthermore, the stability in flock home ranges should have important implications for the population density (flock- density dependent), at least for core species that associate permanently in flocks, actually in the same flock, potentially for their entire lives. Moreover home range stability may limit the density of obligate flock participants allowing greater species packing and potentially promote high species richness in Neotropical avifaunas as suggested by Powell (1989).    36  2.5.5 Elevation and vegetation structure  The variation in the number of species per flock was best explained by vegetation structure complexity, including canopy height and number of trees. Although elevation was included in some of the top models, it explained little variation in flock richness along the gradient.  Among the predictors, elevation had the least relative importance in the model, and the model with elevation as a sole variable performed worse than any other model. I found that flocks tended to be larger and more diverse in areas with a higher density of trees and with higher forest canopy. Together, these variables may offer a more structurally complex forest, with opportunities for species to partition resources and foraging locations across forest strata, increasing the capacity for overall group size of the flocks. A similar pattern with larger flocks occurring in areas with more diverse and dense vegetation was described in a smaller scale study in Andean flocks by Moynihan (1979) and by Lee et al. (2005) who also found a decrease in diversity per flock in areas with simpler vegetation structure. Other habitat variables, such us vegetation composition and resource availability (Srinivasan and Quader 2012), are also expected to be important in determining flock richness. It will be of great interest to further investigate these habitat factors in conjunction and their contribution to support flock diversity.   2.5.6 Prevalence of flocking at the community level The results of this study showed that flocks are an important characteristic of bird communities along the gradient, where around 40 % of species at any given elevation utilized    37 flocks as a foraging strategy. The relatively high proportion of species participation in these flocks is consistent with other studies at smaller scales in tropical regions, including the Atlantic forest, where >50% of the bird community joined flocks (Aleixo 1997), the Colombian Andes (40%; Arbeláez-Cortés et al. 2011), Hispaniola Island (>80%, Latta and Wunderle 1996) and French Guiana (38%; Jullien and Thiollay 2001). However, the idea that mixed-species flocks are a predominant feature of bird communities across elevations, from the Amazon lowlands to the high Andes, has not been well documented before. This research makes an important step towards filling that knowledge gap by demonstrating that facilitation plays an important and underappreciated role in structuring bird communities across various tropical habitats and elevations.  The prevalence of the flocking strategy at the community level peaks at three different elevations along the gradient 400 to 500 m.a.s.l., 1600 to 1700 m.a.s.l., and 2600 to 2800 m.a.s.l. In this study we did not test the factors that could explain these peaks of high species participation in flocks at those particular elevations. However, historically patterns of participation in flocks have been attributed to predation pressure (Thiollay 1999) and resource availability (Berner and Grubb 1985). Interestingly, the peaks of flocking prevalence occurring at 400 to 500 m.a.s.l., and 1600 to 1700 m.a.s.l. match the elevations where peaks of net primary productivity (NPP) have been described along the Manu elevational gradient by Marthews et al. (2012) and  Huasco et al. (2014). In addition, the peak observed from 400 to 500 m.a.s.l., is also consistent with the peak in raptor diversity (the main predators of adult passerines) reported by Valdez (1999).  I hypothesize that an interaction between these factors could be driving this pattern.    38  Finally, I point out that the prevalence of mixed-species flocks across elevations indicates that facilitation plays an important and underappreciated role as a mechanism structuring Neotropical bird communities. Although much information has been documented over the last century about flock presence in different habitats in the tropics, emphasis should be now shifted towards broad scale patterns and implications for bird communities.            	  	  	  	  	  	  	  	  	     39 Table	   2.1	   Mixed-­‐species	   flock	   types	   identified	   across	   the	   elevational	   gradient.	   Mean	  number	  of	  species	  (Mean	  ±	  SD)	  and	  mean	  number	  of	  individual	  birds	  for	  each	  flock	  type	  is	  presented.	  Sample	  size	  and	  range	  of	  values	  are	  shown	  in	  bold.	  	                Flock	  type	   Elevations	  m.a.s.l.	  Mean	  number	  of	  species	  	  ±	  SD	  (n)	  Mean	  number	  of	  individuals	  Range	  	  Lowland	  terra-­‐firme	  	   300-­‐1100	   8.4	  ±	  4.5	  	  	  	  	  (50)	   12.9	  ±	  7.5	   3-­‐21	   5-­‐38	  Lowland	  Bamboo	   300-­‐1100	   6.8	  ±	  2.0	  	  	  	  	  (31)	   10.1	  ±	  3.2	   4-­‐12	   6-­‐19	  Lowland	  Canopy	   300-­‐1100	   14.0	  ±	  7.7	  	  	  (23)	   21.4	  ±	  13.1	   6-­‐35	   7-­‐51	  Low-­‐montane	   1100-­‐1900	   12.3	  ±	  5.6	  	  	  (55)	   20.6	  ±	  10.8	   4-­‐29	   7-­‐54	  High-­‐montane	   2250-­‐3500	   9.4	  ±	  4.2	  	  	  	  	  (51)	   19.1	  ±	  8.5	   3-­‐20	   6-­‐41	     40 Table	   2.2	   Species	   with	   high	   tendency	   to	   flock.	   The	   core	   component	   of	   each	   flock	   type,	  Flocking	  Index	  (Ocurrrence*Propensity)	  and	  number	  of	  flocks	  observed	  are	  presented.	  	  Species	  that	  were	  intraspecifically	  gregarious	  are	  indicated	  with	  an	  asterisk	  (*).	   Species	   n	   Flocking	  index	   Flock	  type	  Myioborus	  melanocephalus*	   37	   0.93	   High	  montane	  Mecocerculus	  stictopterus	   31	   0.78	   High	  montane	  Hemispingus	  atropileus*	   30	   0.75	   High	  montane	  Diglossa	  cyanea	   40	   0.66	   High	  montane	  Myioborus	  miniatus*	   41	   1.00	   Low	  montane	  Chlorospingus	  flavigularis*	   40	   0.98	   Low	  montane	  Leptopogon	  superciliaris	   37	   0.90	   Low	  montane	  Chlorochrysa	  calliparaea	   36	   0.88	   Low	  montane	  Tangara	  arthus*	   25	   0.61	   Low	  montane	  Thamnomanes	  schistogynus*	   27	   1.00	   Lowlands	  bamboo	  Microrhopias	  quixensis	   24	   0.89	   Lowlands	  bamboo	  Anabazenops	  dorsalis	   16	   0.59	   Lowlands	  bamboo	  Lanio	  versicolor*	   10	   0.83	   Lowlands	  canopy	  Myrmotherula	  axillaris*	   10	   0.83	   Lowlands	  canopy	  Tachyphonus	  rufiventer*	   10	   0.83	   Lowlands	  canopy	  Tangara	  schrankii	   9	   0.75	   Lowlands	  canopy	  Tangara	  chilensis*	   8	   0.67	   Lowlands	  canopy	  Myrmotherula	  axillaris*	   35	   1.00	   Terra	  firme/Flooded	  	  Myrmotherula	  menetriesii	   23	   0.66	   Terra	  firme/Flooded	  	  Thamnomanes	  ardesiacus*	   23	   0.66	   Terra	  firme/Flooded	  	  Thamnomanes	  schistogynus*	   23	   0.66	   Terra	  firme/Flooded	  	   	      41  Table	  2.3	  Model	  selection	  results	  from	  generalized	  linear	  models	  for	  flock	  richness	  across	  the	   elevational	   gradient.	   Explanatory	   variables	   include	   elevation	   (Elev),	   mean	   canopy	  height	   (Canopy),	   and	   number	   of	   trees	   (Trees).	   For	   each	   fitted	   model	   the	   number	   of	  parameters	   (k),	   change	   in	   corrected	   quasi-­‐Akaike	   from	   the	  model	   with	   the	   lowest	   QAIC	  value	  (Delta_QAICc),	  QAIC	  weights	  (QAICcWt)	  are	  shown.	   Response	  variable	  Model	   	  k	   QAICc	   Delta_QAICc	   QAICcWt	  	  Num_species	   Canopy+Trees	   4	   435.19	   0.00	   0.63	   	  	   Elevation+Canopy+Trees	   5	   437.32	   2.13	   0.22	   	  	   Elevation+Trees	   3	   439.24	   4.04	   0.08	   	  	   Trees	   4	   439.52	   4.33	   0.07	   	  	   Canopy	   3	   447.33	   12.14	   0.00	   	  	   Elevation+Canopy	   4	   448.35	   13.15	   0.00	   	  	   Intercept	   2	   452.71	   17.52	   0.00	   	  	   Elevation	   3	   454.21	   19.02	   	  	  0.00	   	         42     Figure 2.1 Dendrogram for average linkage cluster of mixed-species flocks along an elevational gradient. Clustering distance is based on Jaccard dissimilarity index of species composition of flocks. Each flock name indicates the elevation where it was observed. Red lines indicate clusters of the main flock types.    F 399F 404F 424F 405F 438FF 658FF 421FF 422FF 515F 431FF 425FF 403FF 408F 429F 453F 443FF 411FF 412F 430F 529F 441F 810F 408FF 437FF 440F 440F 449F 474F 395F 406F 562F 534F 882F 681F 699F 647F 606F 638F 639F 925F 667F 677F 545F 579F 875F 654F 719F 1099F 658F 641F 735F 711F 729F 643F 693F 1023FF 1033FF 846FF 404F 762F 439FF 427FF 433FF 430F 527FF 418FF 402FF 431FF 419FF 424FF 450F 410FF 420F 568FF 409FF 648FF 405F 467FF 434F 680F 407F 418F 1354F 1468F 1655FF 1873F 1511F 1742FF 1787FF 1366F 1375F 1259FF 1288F 1888FF 1718F 1306FF 1432F 1390F 1462FF 1450F 1509F 1653F 1282F 1370F 969FF 1131FF 1280FF 1310F 1127F 1298FF 1428F 1320F 1134FF 1343FF 1581FF 1538FF 1672F 1267FF 1345F 1422F 1385FF 1650FF 1355FF 1520F 1403F 1671F 1380FF 1721F 1273FF 1146F 1348FF 1309FF 1293F 1172F 1724F 1326F 1353FF 3482FF 3396FF 3424FF 2496FF 2344FF 2369F 2769FF 2549FF 2935F 2834FF 2825F 2883F 2910F 2923FF 2625FF 2526FF 2542FF 2813FF 2790F 2363FF 2881FF 2422FF 2830FF 2862FF 2872F 2614FF 2809F 2979F 3070F 2717F 3238F 2526F 2506F 2819F 2533F 2993F 2809FF 2990FF 3013F 2474F 2660FF 2736F 2830FF 2653FF 2974FF 2781FF 2845F 2783FF 2576FF 26760.00.20.40.60.81.0Dissimilarity Index (Jaccard)Understory  Bamboo       Lowlands 300-1100 m High-montane  2300-3500      Low-montane         1100-1900 m       Understory         Terra firme              Canopy        ElevationCluster Dendrogram “average”   43    Figure 2.2 Mixed-species richness for Lowland, Low-montane and High-montane flocks in the Manu region, Peru. Number of species per flock is shown. Each dot represents an independent flock. The grey shading indicates the confidence intervals.       010203040Flock typeNumber of speciesLowland Low-montane High-montane   44    Figure 2.3 Mixed-species flock size for Lowland, Low-montane and High-montane flocks in the Manu region, Peru.  Number of individuals per flock is shown.. Each dot represents an independent flock. The grey shading indicates the confidence intervals.       0102030405060Flock typeNumber of individualsLowland Low-montane High-montane   45   Figure  2.4 Regression of number of species and number of individuals foraging in mixed-species flocks in Manu region, Peru (Kendall’s	  Tau=	  0.789,	  p<0.001).	  	         0 10 20 30 40 50 60010203040Number of individualsNumber of species    46                              Figure 2.5 Species richness of avian foraging guilds across the Manu elevational gradient for (a) Mixed-species flocks (b) overall community. Each guild is represented by a different colour including: insectivores (black), frugivores (red), nectarivores (green) and omnivores (blue). Each dot represents the total number of species at a given elevational zone (e.g. 400-500; 501-600; 601-700). The solid lines are the predicted values of species as a function of elevation.  0 500 1000 1500 2000 2500 3000 3500010203040ElevationNumber of species in mixed-species flocksM i x e d - s p e c i e s  f l o c k sInsectivores Frugivores Omnivores Nectarivores0 500 1000 1500 2000 2500 3000 3500050100150ElevationNumber of speciesInsectivores Frugivores Omnivores NectarivoresCommunity   47                    Figure 2.6 Species richness of avian foraging guilds across the Manu elevational gradient. Total	  number	  of	  species	  in	  the	  Manu	  gradient	  community	  (red)	  and	  number	  of	  species	  participating	  in	  flocks	  (blue)	  for	  each	  elevational	  zone	  are	  shown for (a) insectivores,  (b) frugivores,	  	  (c) omnivores, (d) nectarivores. The	  solid	  lines	  are	  the	  predicted	  values	  of	  species	  richness	  as	  a	  function	  of	  elevation.   0 500 1000 1500 2000 2500 3000 3500050100150200InsectivoresNumber of species0ElevationO v e r a l l   b i r d    c o m m u n i t yF o r a g i n g  i n   m i x e d   f l o c k s 0 500 1000 1500 2000 2500 3000 350001020304050FrugivoresNumber of species0ElevationO v e r a l l   b i r d    c o m m u n i t yF o r a g i n g  i n   m i x e d   f l o c k s 0 500 1000 1500 2000 2500 3000 3500102030405060OmnivoresElevationNumber of species0O v e r a l l   b i r d    c o m m u n i t yF o r a g i n g  i n   m i x e d   f l o c k s 0 500 1000 1500 2000 2500 3000 3500051015202530NectarivoresNumber of  speciesO v e r a l l   b i r d    c o m m u n i t yF o r a g i n g  i n   m i x e d   f l o c k s Elevation   48    Figure 2.7 Proportions of species in the community joining mixed flocks for each foraging guild across elevation are shown. Each guild is represented by a different colour including: insectivores (black), frugivores (red), omnivores (blue) and nectarivores (green). Each dot represents the proportion of species joining flocks at a given elevational zone (e.g. 400-500; 501-600; 601-700). The solid lines are the predicted values as a function of elevation.   0 500 1000 1500 2000 2500 3000 35000.00.20.40.60.8GuildsElevationProportion of species of the community joining Mixed flocksInsectivores Frugivores Omnivores Nectarivores   49  Figure 2.8 Stability of flocks over time. Similarity in species composition of flocks observed in (a) hours 1-2, (b) days 1-2 and (c) years 1-2 is shown. Each dot represents the calculated similarity index (1-Jaccard dissimilarity index) for each flock between observations. The grey shading shows the confidence intervals.         0.00.20.40.60.81.0Time scaleStability  (1−Jaccard disimilarity index)Hours YearsWeeks   50     Figure 2.9 Temporal stability of flocks across elevation. Similarity in species composition of flocks across elevation at three different time scales. The solid lines are the predicted values of flock stability as a function of elevation between hours (blue), weeks (red) and years (green). The shading shows the confidence intervals based on the standard errors of the estimates. The lines along the x-axis indicate the elevation where each flock was observed.        51      Figure 2.10 Stability of flocks over time. Similarity in species composition for a) Lowland flocks b) Low-montane flocks c) High-montane flocks. Each dot represents the calculated similarity index (1-Jaccard dissimilarity index) for each flock between observations.  0.00.20.40.81.0Flock typeStability  (1−Jaccard disimilarity index)Lowland Low-montane High-montane0.6   52   Figure 2.11 Stability of flocks across years. Similarity in species composition of flocks observed in 2013 and 2014 for a) Lowland flocks b)Low-montane flocks c)High-montane flocks. Each dot represents the calculated similarity index (1-Jaccard dissimilarity index) for each flock between observations.       0.00.20.40.60.81.0Flock  typeStability between years (1−Jaccard similarity index)Lowland High-montaneLow-montane   53    Figure 2.12 Home range of two mixed-species bird flocks at Low-montane elevation (1240-1260 masl). Light colours indicate the 2013 home range for each flock; dark colours indicate the 2014 home range. Roosting site is indicated for each flock.         54                          Figure 2.13 Mixed-species flock richness in the Manu study region varying by (a) canopy height, (b) number of trees and (c) elevation. Best-fitted model was used to plot canopy height and number of trees. Second best–fitted model was used to plot elevation. The solid line is the predicted value of flock richness as a function of each variable, filling in the other explanatory variables to the median. Each dot represents a obsered flock. The grey shading shows the confidence intervals based on the standard errors of the estimates. 0 5 10 15 20 25 30 3551015202530350a) Canopy heightCanopy heightFlock richness(Number of species)0 10 20 30 40 50 60010203040b) Number of treesNumber of treesFlock richness (Number of species)0 500 1000 1500 2000 2500 3000 3500010203040c) ElevationElevation (masl)Flock richness (Number of species)    55   Figure	  2.14	  Bird	  species	  richness	  per	  elevational	  band.	  Total	  number	  of	  species	  in	  the	  Manu	  gradient	  community	  (black)	  and	  number	  of	  species	  participating	  in	  flocks	  (blue)	  for	  each	  elevational	  zone	  are	  shown (e.g. 400-500; 501-600; 601-700).	  The	  solid	  line	  is	  the	  predicted	  value	  of	  species	  richness	  as	  a	  function	  of	  elevation.	  	  	      0 500 1000 1500 2000 2500 3000 3500050100150200250ElevationNumber of speciesOverall bird CommunityForaging in mixed  flocks   56       Figure 2.15 Prevalence of flocking across elevations. Proportion of species of the community joining flocks for each elevation band is shown.    0 500 1000 1500 2000 2500 3000 35000.00.10.20.30.40.5ElevationProportion of species of the community joining flocks   57 Chapter 3: Conclusions  3.1 General conclusions  This study examined the structure and dynamics of one of the most striking examples of facilitation among multiple species, mixed-species bird flocks. This is the first study that describes these multi-species groups of birds across a large-scale, continuous elevational gradient in the Neotropics. I demonstrated that mixed-species flocks are a common feature of Neotropical bird communities at all elevations, within different habitats of the Amazonian lowlands (terra firme and bamboo forest) to the high Andes (lower montane and cloud forest). These flocks exhibit an extraordinary degree of organization and stability. My analysis of flock structure called attention to three distinguishable types of flocks occurring across the gradient, with transitions that are associated with changes in the overall bird community. Andean flocks can be differentiated from Amazonian flocks by their larger home range size and higher diversity of participating species, compared to the smaller and slightly less diverse lowland flocks. Flocks across the gradient exhibited a similar general structure, composed of a highly stable core group of species and a more dynamic component of attendant species.  The core component of the flocks was restricted to 3-5 species of obligate participants per flock type, represented by twelve genera from three families: Thaupidae, Tyrannidae and Thamnophilidae.  The analysis of temporal and spatial stability showed that Andean mixed-species flocks were just as stable as those occurring in the Amazonian lowlands, with flocks exhibiting stable home ranges and consistency in core member composition over time, even across years. Finally,    58 I showed that some species are specialized to forage within mixed-species flocks and appear to be behaviourally restricted to these subunits of the community. Furthermore, this study demonstrates that mixed-species flocks are used by over one third of species present at any elevation across the gradient, representing an important component of tropical bird communities. It is still unclear from these results, however, whether any attributes or behaviours can be used to reliably predict species participation in flocks. There are two key insights from this study that contribute to our understanding of how flocks influence the structure of tropical bird communities. First, the stability of these multispecies associations across elevations suggests that flocks function as small subunits within bird communities. In these flocks, obligate participants, in their role as nuclear species, may be responsible for the cohesion and maintenance of these multi-species groups. Importantly, if these species were removed from the community, they could have a disproportionate effect on many other species due to their influence on flock formation. Second, this study highlights the high proportion of species participating in these multi-species associations, suggesting that such facilitative interactions are remarkably important in Neotropical bird communities. The facilitation occurring among bird species in these flocks may relax competition in the community and thus allow the coexistence of a higher number species. Furthermore the potential strong interdependence of some flock members may limit their densities, promoting higher species packing in Neotropical communities.     59 3.2 Future research A number of questions remain to be explored in how facilitative interactions of mixed-species flocks influence bird communities. One key avenue will be to examine the interdependence of obligate flock participants in aspects such as co-occurrence patterns and elevational ranges. One may expect that species with sufficiently strong interdependent interactions may co-occur more often than expected by chance and exhibit coincident elevational range boundaries. Long-term associations with mixed species flocks may have consequences for individual species traits and behaviours. As such, it will be of great interest to explore whether certain life-history traits or behaviours (e.g., lower BMR, higher survival, cooperative breeding) emerge more frequently in species associated with mixed-species flocks. Other relatively unexplored area of research is the evaluation of factors that explain the prevalence of flocking across elevations. Predation risk, as well as resource availability and patchiness, may be particularly good predictors to explain variation in flocking prevalence. Finally, with the increased availability of information on evolutionary relationships among avian species, it will be very interesting to evaluate the phylogenetic structure of mixed flocks across elevations, to evaluate whether these groups tend to be composed of more closely or distantly related species. Traditional studies of mixed-species flocks have intentionally avoided repeated observations of the same flock (e.g. Satischandra et al. 2007); however, I recommend that future studies repeat detailed surveys of the same flock at different time intervals to specifically examine the stability of association among species. Given the various gaps remaining in this topic, and the challenges involved in conducting experiments with this taxonomic group in particular, advancing the frontier of our knowledge on mixed species flocks will require several comprehensive and comparable data sets across large    60 scales to more effectively address the question of how these facilitative interactions structure communities. This thesis represents a first step towards the exploration of this question using broad spatial scales.     61 References 	  Afkhami,	  M.	  E.	  2012.	  Fungal	  endophyte-­‐grass	  symbioses	  are	  rare	  in	  the	  California	  floristic	  province	  and	  other	  regions	  with	  Mediterranean-­‐influenced	  climates.	  Fungal	  Ecology	  5:345–352.	  Afkhami,	  M.	  E.,	  P.	  J.	  McIntyre,	  and	  S.	  Y.	  Strauss.	  2014.	  Mutualist-­‐mediated	  effects	  on	  species’	  range	  limits	  across	  large	  geographic	  scales.	  (W.	  van	  der	  Putten,	  ed.)Ecology	  Letters	  17:1265–1273.	  Aleixo,	  A.	  1997.	  Composition	  of	  mixed-­‐species	  bird	  flocks	  and	  abundance	  of	  flocking	  species	  in	  a	  semideciduous	  forest	  of	  southeastern	  Brazil.	  Revista	  Brasileira	  de	  Ornitologia	  5:11–18.	  Araújo,	  M.	  B.,	  and	  M.	  Luoto.	  2007.	  The	  importance	  of	  biotic	  interactions	  for	  modelling	  species	  distributions	  under	  climate	  change.	  Global	  Ecology	  and	  Biogeography	  16:743–753.	  Arbeláez-­‐cortés,	  A.	  E.,	  and	  O.	  H.	  Marín-­‐gomez.	  2012.	  The	  composition	  of	  mixed-­‐species	  bird	  flocks	  in	  Alto	  Quindío,	  Colombia.	  Wilson	  Journal	  of	  Ornithology	  124:572–580.	  Arbeláez-­‐Cortés,	  E.,	  H.	  Rodriguez-­‐Correa,	  and	  M.	  Restrepo-­‐Chica.	  2011.	  Mixed-­‐species	  bird	  flocks:	  patterns	  of	  activity	  and	  species	  composition	  in	  a	  region	  of	  the	  central	  Andes	  of	  ColOmbia.	  Revista	  Mexicana	  de	  …	  639–651.	  Bawa,	  K.	  S.	  1990.	  Plant-­‐Pollinator	  Interactions	  in	  Tropical	  Rain	  Forests.	  Annu.	  Rev.	  Ecol.	  Evol.	  Syst	  21:399–422.	  Berner,	  T.	  O.,	  and	  T.	  C.	  Grubb.	  1985.	  An	  Experimental	  Analysis	  of	  Mixed-­‐Species	  Flocking	  in	  Birds	  of	  Deciduous	  Woodland.	  Ecology	  66:1229–1236.	  Bertness,	  M.	  D.	  1989.	  Intraspecific	  Competition	  and	  Facilitation	  in	  a	  Northern	  Acorn	  Barnacle	  Population	  70:257–268.	  Bertness,	  M.	  D.,	  and	  R.	  Callaway.	  1994.	  Positive	  interactions	  in	  communities.	  Trends	  in	  Ecology	  and	  Evolution	  9:191–193.	  Bertness,	  M.,	  and	  G.	  .	  Leonard.	  1997.	  The	  role	  of	  positive	  interactions	  in	  communities:	  lessons	  from	  intertidal	  habitats.	  Ecology	  1976–1989.	  Bolker,	  B.	  2016.	  Dealing	  with	  quasi-­‐models	  in	  R	  1–5.	     62 Boucher,	  D.	  H.	  1985.	  Lotka-­‐volterra	  models	  of	  mutualism	  and	  positive	  density-­‐dependence.	  Ecological	  Modelling	  27:251–270.	  Bracken,	  M.	  E.	  S.,	  C.	  a	  Gonzalez-­‐Dorantes,	  and	  J.	  J.	  Stachowicz.	  2007.	  Whole-­‐community	  mutualism:	  associated	  invertebrates	  facilitate	  a	  dominant	  habitat-­‐forming	  seaweed.	  Ecology	  88:2211–9.	  Brawn,	  J.	  D.,	  J.	  R.	  Karr,	  and	  J.	  D.	  Nichols.	  1995.	  Demography	  of	  Birds	  in	  a	  Neotropical	  Forest :	  Effects	  of	  Allometry	  ,	  Taxonomy	  ,	  and	  Ecology.	  Ecology	  76:41–51.	  Brockmann,	  J.,	  and	  C.	  J.	  Barnard.	  1979.	  Kleptoparasitism	  in	  Birds.	  Animal	  Behaviour	  27:487–514.	  Bronstein,	  J.	  L.	  1994.	  Our	  current	  understanding	  of	  Mutualism.	  The	  Quarterly	  Review	  of	  Biology	  69:31–51.	  Bruno,	  J.	  F.,	  J.	  J.	  Stachowicz,	  and	  M.	  D.	  Bertness.	  2003.	  Inclusion	  of	  facilitation	  into	  ecological	  theory.	  Trends	  in	  Ecology	  and	  Evolution.	  Bullock,	  J.	  M.,	  R.	  J.	  Edwards,	  P.	  D.	  Carey,	  R.	  J.	  Rose,	  and	  R.	  J.	  Geographical.	  2000.	  Separation	  of	  Two	  Ulex	  Species	  at	  Three	  Spatial	  Scales :	  Does	  Competition	  Limit	  Species	  ’	  Ranges.	  Ecography	  23:257–271.	  Burnham,	  K.	  .,	  and	  D.	  .	  Anderson.	  2002.	  Model	  Selection	  and	  Multimodel	  Inference-­‐A	  practical	  information	  theoretic	  approach.	  The	  effects	  of	  brief	  mindfulness	  intervention	  on	  acute	  pain	  experience:	  An	  examination	  of	  individual	  difference	  (Second	  edi.,	  Vol.	  1).	  Springer-­‐Verlag,	  Berlin.	  Buskirk,	  W.	  H.	  1976.	  Social	  systems	  in	  a	  tropical	  forest	  avifauna.	  The	  American	  naturalist	  110:293–310.	  Buskirk,	  W.,	  G.	  Powell,	  J.	  Wittenberger,	  R.	  Buskirk,	  and	  T.	  Powell.	  1972.	  Interspecific	  Bird	  Flocks	  in	  Tropical	  Highland	  Panama.	  The	  Auk	  89:612–624.	  Callaway,	  R.	  M.	  1995.	  Positive	  Interactions	  among	  Plants.	  Botanical	  Review	  61:306–349.	  Cavieres,	  L.,	  M.	  Arroyo,	  A.	  Peñaloza,	  M.	  Molina,	  and	  C.	  Torres.	  2002.	  Nurse	  effect	  of	  Bolax	  gummifera	  cushion	  plants	  in	  the	  alpine	  vegetation	  of	  the	  Chilean	  Patagonian	  Andes.	  Vegetation	  of	  science	  13:547–554.	  Clements,	  F.	  1916.	  Plant	  Succession.	  An	  Analysis	  of	  the	  Development	  of	  Vegetation	  (Vol.	  242).	  Carnegie	  Institution	  of	  Washington.,	  Washington:	     63 Colorado,	  G.	  J.	  2013.	  Why	  animals	  come	  together,	  with	  the	  special	  case	  of	  mixed-­‐species	  bird	  flocks.	  Revista	  EIA	  10:49–66.	  Connell,	  J.	  H.	  1961.	  The	  influencen	  of	  Interspecific	  competition	  and	  other	  factors	  on	  the	  distribution	  of	  the	  Barnacle	  Chthamalus	  stellatus.	  Ecology	  42:710–723.	  Courter,	  J.	  R.,	  and	  G.	  Ritchison.	  2012.	  Asymmetries	  in	  Mobbing	  Behavior	  Among	  Nuclear	  Flockmates.	  The	  Wilson	  Journal	  of	  Ornithology.	  Crawley	  M.	  J.	  2007.	  The	  R	  Book	  (1st	  ed.).	  Wiley	  Publishing.	  Crotty,	  S.,	  and	  M.	  D.	  Bertness.	  2015.	  Positive	  interactions	  expand	  habitat	  use	  and	  the	  realized	  niches	  of	  sympatric	  species.	  Ecology	  96:2575–2582.	  Davidson,	  D.	  W.,	  and	  D.	  Mckey.	  1993.	  The	  evolutionary	  ecology	  of	  plant-­‐ant	  symbiosis.	  Journal	  of	  Hymenoptera	  Research	  2:13–83.	  Davis,	  D.	  E.	  1946.	  A	  Seasonal	  Analysis	  of	  Mixed	  Flocks	  of	  Birds	  in	  Brazil.	  Ecology	  27:168–181.	  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.	  Dolby,	  A.	  S.,	  and	  T.	  C.	  Grubb.	  1998.	  Benefits	  to	  satellite	  members	  in	  mixed-­‐species	  foraging	  groups:	  an	  experimental	  analysis.	  Animal	  behaviour	  56:501–509.	  Foster,	  W.	  A.,	  and	  J.	  E.	  Treherne.	  1981.	  Evidence	  for	  the	  dilution	  effect	  in	  the	  selfish	  herd	  from	  fish	  predation	  on	  a	  marine	  insect.	  Nature	  293:466–467.	  Gaston,	  K.	  2000.	  Global	  patterns	  in	  biodiversity.	  Nature	  405:220–227.	  Goodale,	  E.,	  G.	  Beauchamp,	  R.	  D.	  Magrath,	  J.	  C.	  Nieh,	  and	  G.	  D.	  Ruxton.	  2010.	  Interspecific	  information	  transfer	  influences	  animal	  community	  structure.	  Trends	  in	  Ecology	  &	  Evolution	  25:354–361.	  Goodale,	  E.,	  P.	  Ding,	  X.	  Liu,	  A.	  Martínez,	  X.	  Si,	  M.	  Walters,	  and	  S.	  K.	  Robinson.	  2015.	  The	  structure	  of	  mixed-­‐species	  bird	  flocks,	  and	  their	  response	  to	  anthropogenic	  disturbance,	  with	  special	  reference	  to	  East	  Asia.	  Avian	  Research	  6:14.	  Goodale,	  E.,	  B.	  Nizam,	  and	  V.	  Robin.	  2009.	  Regional	  variation	  in	  the	  composition	  and	  structure	  of	  mixed-­‐species	  bird	  flocks	  in	  the	  Western	  Ghats	  and	  Sri	  Lanka.	  Current	  …	  97.	  Goss-­‐Custard,	  J.	  1980.	  Competition	  for	  food	  and	  inter-­‐	  ference	  among	  waders.	  Ardea	  68:31–   64 52.	  Graves,	  G.	  R.,	  and	  N.	  J.	  Gotelli.	  1993.	  Assembly	  of	  avian	  mixed-­‐species	  flocks	  in	  Amazonia.	  Proceedings	  of	  the	  National	  Academy	  of	  Sciences	  of	  the	  United	  States	  of	  America	  90:1388–1391.	  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.	  Garber,	  Eds.,	  pp.	  521–558,	  The	  University	  of	  Chicago	  Press,	  Chicago,	  Ill,	  USA,	  2000.	  Gross,	  N.,	  G.	  Kunstler,	  P.	  Liancourt,	  F.	  De	  Bello,	  K.	  N.	  Suding,	  and	  S.	  Lavorel.	  2009.	  Linking	  individual	  response	  to	  biotic	  interactions	  with	  community	  structure:	  a	  trait-­‐based	  framework.	  Functional	  Ecology	  23:1167–1178.	  Hacker,	  S.	  D.,	  and	  S.	  D.	  Gaines.	  1997.	  Some	  implications	  of	  direct	  positive	  interactions	  for	  communtiy	  species	  diversity.	  Ecology	  78:1990–2003.	  Harrison,	  N.	  M.,	  and	  M.	  J.	  Whitehouse.	  2011.	  Mixed-­‐species	  flocks:	  An	  example	  of	  niche	  construction?	  Animal	  Behaviour.	  Hart,	  P.	  J.,	  and	  L.	  A.	  Freed.	  2003.	  Structure	  and	  dynamics	  of	  mixed-­‐species	  flocks	  in	  a	  hawaian	  rain	  forest.	  The	  Auk	  120:82–95.	  Hay,	  M.	  E.	  1981.	  The	  functional	  morphology	  of	  turf-­‐forming	  seaweeds:	  persistence	  in	  stressful	  marine	  habitats.	  Ecology	  62:739–750.	  Hay,	  M.	  E.,	  J.	  D.	  Parker,	  D.	  E.	  Burkepile,	  C.	  C.	  Caudill,	  E.	  Alan,	  Z.	  P.	  Hallinan,	  A.	  D.	  Chequer,	  et	  al.	  2004.	  Mutualisms	  and	  Aquatic	  Community	  Structure :	  The	  Enemy	  of	  My	  Enemy	  Is	  My	  Friend.	  Annual	  Reviews	  35:175–197.	  Hillebrand,	  H.	  2004.	  On	  the	  Generality	  of	  the	  Latitudinal	  Diversity	  Gradient.	  Am.	  Nat	  163:192–211.	  Hino,	  T.	  1997.	  Mutualism	  and	  Commensalism	  in	  Avian	  Mixed-­‐Species	  Flocks	  in	  a	  Western	  Forest	  of	  Madagascar.	  Primate	  Research.	  Huasco,	  W.	  H.,	  C.	  A.	  J.	  Girardin,	  C.	  E.	  Doughty,	  D.	  B.	  Metcalfe,	  L.	  D.	  Baca,	  J.	  E.	  Silva-­‐Espejo,	  D.	  G.	  Cabrera,	  et	  al.	  2014.	  Seasonal	  production,	  allocation	  and	  cycling	  of	  carbon	  in	  two	  mid-­‐elevation	  tropical	  montane	  forest	  plots	  in	  the	  Peruvian	  Andes.	  Plant	  Ecology	  &	  Diversity	  7:125–142.	  Hutto,	  R.	  L.	  1987.	  A	  Description	  of	  Mixed-­‐Species	  Insectivorous	  Bird	  Flocks	  in	  Western	     65 Mexico.	  The	  Condor	  89:282.	  Hutto,	  R.	  L.	  1988.	  Foraging	  behavior	  patterns	  suggest	  a	  possible	  cost	  associated	  with	  participation	  in	  mixed-­‐species	  bird	  flocks.	  Oikos	  51:79–83.	  Hutto,	  R.	  L.	  1994.	  Mixed	  Species	  Flocks	  in	  a	  Tropical	  Deciduous	  Forest	  in	  Western	  Mexico.	  The	  Condor.	  Jankowski,	  J.	  E.,	  C.	  L.	  Merkord,	  W.	  F.	  Rios,	  K.	  G.	  Cabrera,	  N.	  S.	  Revilla,	  and	  M.	  R.	  Silman.	  2013.	  The	  relationship	  of	  tropical	  bird	  communities	  to	  tree	  species	  composition	  and	  vegetation	  structure	  along	  an	  Andean	  elevational	  gradient.	  Journal	  of	  Biogeography	  40:950–962.	  Jankowski,	  J.	  E.,	  S.	  K.	  Robinson,	  and	  D.	  J.	  Levey.	  2010.	  Squeezed	  at	  the	  top:	  Interspecific	  aggression	  may	  constrain	  elevational	  ranges	  in	  tropical	  birds.	  Ecology	  91:1877–1884.	  Jullien,	  M.,	  and	  J.	  Clobert.	  2000.	  The	  Survival	  Value	  of	  Flocking	  in	  Neotropical	  Birds :	  Reality	  or	  Fiction ?	  Ecology	  81:3416–3430.	  Jullien,	  M.,	  and	  J.	  Thiollay.	  1998.	  Multi-­‐species	  territiorality	  and	  dynamic	  of	  neotropical	  understoty	  bird	  flocks.	  Journal	  of	  Animal	  Ecology	  67:227–252.	  Jullien,	  M.,	  and	  J.	  Thiollay.	  2001.	  The	  adaptative	  significance	  of	  Flocking	  in	  Tropical	  understory	  Forest	  birds:	  The	  field	  evidence.	  Nouragues:	  dynamics	  and	  plant-­‐animal	  interactions	  in	  a	  neotropical	  rainforest.	  By	  Frans	  Bongers,	  Pierre	  Charles-­‐Dominique,	  Marc	  Théry.	  Kluwer	  Academic	  Press,	  Netherlands	  (pp.	  143–159).	  Krebs,	  C.	  J.	  2014.	  Part	  Four:	  Estimating	  community	  parameters.	  Ecological	  methodology	  (Third.,	  pp.	  479–530).	  Krebs,	  J.	  R.	  1973.	  Social	  learning	  and	  the	  significance	  of	  mixed-­‐species	  flocks	  of	  chickadees	  (	  Parus	  spp.).	  Canadian	  Journal	  of	  Zoology	  51:1275–1288.	  Latta,	  S.,	  and	  J.	  Wunderle.	  1996.	  The	  Composition	  and	  Foraging	  Ecology	  of	  Mixed-­‐Species	  Flocks	  in	  Pine	  Forests	  of	  Hispaniola.	  Condor	  98:595–607.	  Lea,	  A.	  J.,	  J.	  P.	  Barrera,	  L.	  M.	  Tom,	  and	  D.	  T.	  Blumstein.	  2008.	  Heterospecific	  eavesdropping	  in	  a	  nonsocial	  species.	  Behavioral	  Ecology	  19:1041–1046.	  Lee,	  T.	  M.,	  M.	  C.	  K.	  Soh,	  N.	  Sodhi,	  P.	  K.	  Lian,	  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.	  Levey,	  D.	  J.,	  W.	  R.	  Silva,	  and	  M.	  Galetti.	  2002.	  Seed	  dispersal	  and	  frugivory :	  ecology,	     66 evolution,	  and	  conservation.	  (D.	  J.	  Levey,	  W.	  R.	  Silva,	  &	  M.	  Galetti,	  eds.).	  CABI	  Pub,	  New	  York.	  Liancourt,	  P.,	  P.	  Choler,	  N.	  Gross,	  X.	  Thibert-­‐Plante,	  and	  K.	  Tielbörger.	  2012.	  How	  Facilitation	  May	  Interfere	  with	  Ecological	  Speciation.	  International	  Journal	  of	  Ecology	  2012:1–11.	  Macarthur,	  R.	  H.	  1969.	  Patterns	  of	  communities	  in	  the	  tropics.	  Biol.	  J.	  Linn.	  SOC	  1:19–30.	  MacArthur,	  R.,	  and	  J.	  MacArthur.	  1961.	  On	  Bird	  Species	  Diversity.	  Ecological	  Society	  of	  America	  42:594–598.	  Maldonado-­‐Coelho,	  M.,	  and	  M.	  Marini.	  2000.	  Effects	  of	  forest	  fragment	  size	  and	  succesional	  stage	  on	  mixed-­‐species	  bird	  flocks	  in	  Southeaster	  Brazil.	  The	  Condor	  102:585–594.	  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	  0521:1–17.	  Marthews,	  T.	  R.,	  Y.	  Malhi,	  C.	  A.	  J.	  Girardin,	  J.	  E.	  Silva	  Espejo,	  L.	  E.	  O.	  C.	  Aragão,	  D.	  B.	  Metcalfe,	  J.	  M.	  Rapp,	  et	  al.	  2012.	  Simulating	  forest	  productivity	  along	  a	  neotropical	  elevational	  transect:	  Temperature	  variation	  and	  carbon	  use	  efficiency.	  Global	  Change	  Biology	  18:2882–2898.	  Martínez,	  A.	  E.,	  and	  J.	  P.	  Gomez.	  2013.	  Are	  mixed-­‐species	  bird	  flocks	  stable	  through	  two	  decades?	  The	  American	  naturalist	  181:E53–E59.	  Martínez,	  A.	  E.,	  and	  R.	  T.	  Zenil.	  2012.	  Foraging	  guild	  influences	  dependence	  on	  heterospecific	  alarm	  calls	  in	  Amazonian	  bird	  flocks.	  Behavioral	  Ecology	  23:544–550.	  Mazerolle,	  M.	  2012.	  Computing	  AIC,	  AICc,	  QAIC,	  and	  QAICc.	  Merkord,	  C.	  L.	  2010.	  Seasonality	  and	  elevation	  migration	  in	  an	  Andean	  bird	  community.	  University	  of	  Missouri.	  Michalet,	  R.,	  R.	  W.	  Brooker,	  L.	  A.	  Cavieres,	  Z.	  Kikvidze,	  C.	  J.	  Lortie,	  F.	  I.	  Pugnaire,	  A.	  Valiente-­‐Banuet,	  et	  al.	  2006.	  Do	  biotic	  interactions	  shape	  both	  sides	  of	  the	  humped-­‐back	  model	  of	  species	  richness	  in	  plant	  communities?	  Ecology	  Letters	  9:767–773.	  Michalet,	  R.,	  and	  F.	  I.	  Pugnaire.	  2016.	  Facilitation	  in	  communities:	  Underlying	  mechanisms,	  community	  and	  ecosystem	  implications.	  Functional	  Ecology	  30:3–9.	  Mirkin,	  B.	  2011.	  Detecting	  Clusters	  Graphically	  (Vol.	  7,	  pp.	  15–41).	  Morse,	  D.	  H.	  1970.	  Ecological	  Aspects	  of	  Some	  Mixed-­‐Species	  Foraging	  Flocks	  of	  Birds.	     67 Ecological	  Monographs	  40:119–168.	  Morse,	  D.	  H.	  1977.	  Feeding	  Behavior	  and	  Predator	  Avoidance	  in	  Heterospecific	  Groups.	  BioScience	  27:332–339.	  Moynihan,	  M.	  1962.	  The	  organization	  and	  probable	  evolntion	  of	  some	  mixed	  species	  flocks	  of	  Neotropical	  birds.	  Smithsonian	  Miscellaneous	  Collections	  143:1–140.	  Moynihan,	  M.	  1968.	  Social	  Mimicry ;	  Character	  Convergence	  Versus	  Character	  Displacement.	  Evolution	  22:315–331.	  Munn,	  C.	  A.	  1985.	  Permanent	  canopy	  and	  understory	  flocks	  in	  Amazonia:	  Species	  composition	  and	  population	  density.	  Ornithological	  Monographs	  683–712.	  Munn,	  C.	  A.	  1986.	  Birds	  that	  cry	  wolf.	  Nature	  319:143–145.	  Munn,	  C.	  A.,	  and	  J.	  W.	  Terborgh.	  1979.	  Multi-­‐species	  territoriality	  in	  Neotropical	  foraging	  flocks.	  The	  Condor	  81:338–347.	  Myers,	  N.,	  R.	  A.	  Mittermeier,	  C.	  G.	  Mittermeier,	  G.	  A.	  B.	  Da	  Fonseca,	  and	  J.	  Kent.	  2000.	  Biodiversity	  hotspots	  for	  conservation	  priorities.	  Nature	  403:853–859.	  Oksanen,	  J.,	  F.	  G.	  Blanchet,	  R.	  Kindt,	  P.	  Legendre,	  R.	  B.	  O’Hara,	  G.	  L.	  Simpson,	  and	  Solymos.	  2011.	  Vegan:	  communityecology	  package.	  R	  package	  version	  1.17–9.	  Palmer,	  T.	  M.,	  E.	  G.	  Pringle,	  A.	  Stier,	  and	  R.	  D.	  Holt.	  2015.	  Mutualism	  in	  a	  community	  context.	  Mutualism.	  Edited	  by	  Judith	  L.	  Bronstein.	  Pellmyr,	  C.	  M.	  2002.	  Pollination	  by	  animals.	  In	  C.	  M.	  Herrera	  &	  O.	  Pellmyr,	  eds.,	  Plant-­‐Animal	  Interactions:	  An	  Evolutionary	  Approach	  (pp.	  157–184).	  Blackwell	  Publishing,	  Oxford.,	  New	  York.	  Pereira,	  P.	  H.	  C.,	  J.	  L.	  L.	  Feitosa,	  D.	  V.	  Medeiros,	  and	  B.	  P.	  Ferreira.	  2013.	  Reef	  fishes	  foraging	  facilitation	  behavior:	  Increasing	  the	  access	  to	  a	  food	  resource.	  Acta	  Ethologica	  16:53–56.	  Péron,	  G.,	  and	  P.	  A.	  Crochet.	  2009.	  Edge	  effect	  and	  structure	  of	  mixed-­‐species	  bird	  flocks	  in	  an	  Afrotropical	  lowland	  forest.	  Journal	  of	  Ornithology	  150:585–599.	  Poulsen,	  B.	  O.	  1996.	  Structure,	  dynamics,	  home	  range	  and	  activity	  pattern	  of	  mixed-­‐species	  bird	  flocks	  in	  a	  montane	  alder-­‐dominated	  secondary	  forest	  in	  Ecuador.	  Journal	  of	  Tropical	  Ecology.	  Powell,	  G.	  V.	  N.	  1974.	  Experimental	  analysis	  of	  the	  social	  value	  of	  flocking	  by	  starlings	  (Sturnus	  vulgaris)	  in	  relation	  to	  predation	  and	  foraging.	  Animal	  Behaviour	  22:501–505.	     68 Powell,	  G.	  V.	  N.	  1979.	  Structure	  and	  Dynamics	  of	  Interspecific	  Flocks	  in	  a	  Neotropical	  Mid-­‐Elevation	  Forest.	  The	  Auk	  96:375–390.	  Powell,	  G.	  V.	  N.	  1985.	  Sociobology	  and	  adaptive	  significance	  of	  interspecific	  foraging	  flocks	  in	  the	  neotropics.	  Ornithological	  Monographs	  36:713–732.	  Powell,	  G.	  V.	  N.	  1989.	  On	  the	  possible	  contribution	  of	  mixed	  species	  flocks	  to	  species	  Richness	  in	  Neotropical	  avifaunas.	  Behavioral	  Ecology	  and	  Sociobiology	  24:387–393.	  Pugnaire,	  F.	  I.,	  P.	  Haase,	  and	  J.	  Puigdefabregas.	  1996.	  Facilitation	  between	  Higher	  Plant	  Species	  in	  a	  Semiarid	  Environment.	  Ecology	  77:1420–1426.	  Pulliam,	  H.	  R.	  1973.	  On	  the	  advantages	  of	  flocking.	  Journal	  of	  theoretical	  biology	  38:419–422.	  Ragusa-­‐Netto,	  J.	  2002.	  Vigilance	  towards	  raptors	  by	  nuclear	  species	  in	  bird	  mixed	  flocks	  in	  a	  Brazilian	  savannah.	  Studies	  on	  Neotropical	  Fauna	  and	  Environment	  37:219–226.	  Remsen,	  J.	  V,	  and	  W.	  S.	  Graves.	  1995.	  Distribution	  patterns	  of	  Buarremon	  brush-­‐finches	  (Emberizinae)	  and	  interspecific	  competition	  in	  Andean	  birds.	  Auk	  112:225–236.	  Rico-­‐Gray,	  V.,	  and	  P.	  S.	  Oliviera.	  2007.	  The	  Ecology	  and	  Evolution	  of	  Ant–Plant	  Interactions.	  Chicago	  University	  Press,	  Chicago.	  Satischandra,	  S.,	  E.	  Kudavidanage,	  S.	  Kotama,	  and	  E.	  Goodale.	  2007.	  The	  benefits	  of	  joining	  mixed-­‐species	  flocks	  for	  greater	  racket-­‐tailed	  drongos	  Dicrurus	  paradiseus.	  Forktail	  23:145–148.	  Schemske,	  D.	  W.,	  G.	  G.	  Mittelbach,	  H.	  V	  Cornell,	  J.	  M.	  Sobel,	  K.	  Roy,	  and	  W.	  K.	  Kellogg.	  2009.	  Is	  There	  a	  Latitudinal	  Gradient	  in	  the	  Importance	  of	  Biotic	  Interactions?	  Annu.	  Rev.	  Ecol.	  Evol.	  Syst	  40:245–69.	  Sexton,	  J.	  P.,	  P.	  J.	  McIntyre,	  A.	  L.	  Angert,	  and	  K.	  J.	  Rice.	  2009.	  Evolution	  and	  Ecology	  of	  Species	  Range	  Limits.	  Annual	  Review	  of	  Ecology,	  Evolution,	  and	  Systematics	  40:415–436.	  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.	  Sridhar,	  H.,	  U.	  Srinivasan,	  R.	  a	  Askins,	  J.	  C.	  Canales-­‐Delgadillo,	  C.-­‐C.	  Chen,	  D.	  N.	  Ewert,	  G.	  a	  Gale,	  et	  al.	  2012.	  Positive	  Relationships	  between	  Association	  Strength	  and	  Phenotypic	  Similarity	  Characterize	  the	  Assembly	  of	  Mixed-­‐Species	  Bird	  Flocks	  Worldwide.	  The	  American	  Naturalist.	     69 Srinivasan,	  U.,	  and	  S.	  Quader.	  2012.	  To	  eat	  and	  not	  be	  eaten:	  modelling	  resources	  and	  safety	  in	  multi-­‐species	  animal	  groups.	  PloS	  one	  7:e42071.	  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.	  Stachowicz,	  J.	  2001.	  Mutualism,	  Facilitation,	  and	  the	  Structure	  of	  Ecological	  Communities.	  BioScience.	  Stotz,	  D.	  F.	  1993.	  Geographic	  variation	  in	  species	  composition	  in	  mixed	  bird	  flocks	  in	  lowland	  humed	  forest	  in	  brazil.	  Papeis	  Avulsos	  de	  Zoologia	  38:61–75.	  Stouffer,	  P.,	  and	  R.	  Bierregaard.	  1995.	  Use	  of	  Amazonian	  Forest	  Fragments	  by	  Understory	  Insectivorous	  Birds.	  Ecology	  76:2429–2445.	  Swynnerton,	  C.	  1915.	  Mixed	  Birds-­‐parties.	  The	  naturalist	  of	  the	  Amazon	  2:346–354.	  Terborgh,	  J.	  1971.	  Distribution	  on	  Environmental	  Gradients :	  Theory	  and	  a	  Preliminary	  Interpretation	  of	  Distributional	  Patterns	  in	  the	  Avifauna	  of	  the	  Cordillera	  Vilcabamba	  ,	  Peru.	  Ecology	  52:23–40.	  Terborgh,	  J.,	  and	  J.	  Weske.	  1975.	  The	  Role	  of	  Competition	  in	  the	  Distribution	  of	  Andean	  Birds	  56:562–576.	  Terborgh,	  J.	  1971.	  Distribution	  on	  Environmental	  Gradients :	  Theory	  and	  a	  Preliminary	  Interpretation	  of	  Distributional	  Patterns	  in	  the	  Avifauna	  of	  the	  Cordillera	  Vilcabamba	  ,	  Peru.	  Ecology	  52:23–40.	  Terborgh,	  J.	  (1990),	  Mixed	  flocks	  and	  polyspecific	  associations:	  Costs	  and	  benefits	  of	  mixed	  groups	  to	  birds	  and	  monkeys.	  Am.	  J.	  Primatol.,	  21:	  87–100.	  doi:	  10.1002/ajp.1350210203	  Thiollay,	  J.	  1999.	  Frequency	  of	  mixed	  species	  flocking	  in	  tropical	  forest	  birds	  and	  correlates	  of	  predation	  risk:	  an	  intertropical	  comparison.	  Journal	  of	  Avian	  Biology	  30:282–294.	  Thompson,	  D.	  B.,	  J.	  H.	  Brown,	  and	  W.	  D.	  Spencer.	  1991.	  Indirect	  facilitation	  of	  granivorous	  birds	  by	  desert	  rodents :	  experimental	  evidence	  from	  foraging	  patterns.	  Ecology	  72:852–863.	  Thomson,	  R.	  L.,	  J.	  T.	  Forsman,	  and	  M.	  Mönkkönen.	  2003.	  Positive	  interactions	  between	  migrant	  and	  resident	  birds:	  testing	  the	  heterospecific	  attraction	  hypothesis.	  Oecologia	  134:431–8.	     70 Tilman,	  D.	  1994.	  Competition	  and	  biodiversity	  in	  spatially	  structured	  habitats.	  Ecology	  75:2–16.	  Valdez,	  U.	  1999.	  Raptor	  communities	  in	  disturbed	  and	  non-­‐disturbed	  areas	  of	  Manu	  Biosphere	  reserve,	  Southeaster	  Peru.	  Memory.	  North	  Carolina	  State	  University.	  Valiente-­‐Banuet,	  A.,	  and	  M.	  Verdú.	  2013.	  Plant	  Facilitation	  and	  Phylogenetics.	  Annual	  Review	  of	  Ecology,	  Evolution,	  and	  Systematics	  44:347–366.	  Ver	  Hoef,	  J.	  M.,	  and	  P.	  L.	  Boveng.	  2007.	  Quasi-­‐poisson	  vs.	  Negative	  binomial	  regression:	  how	  should	  we	  model	  overdispersed	  count	  data?	  Ecology	  88:2766–2772.	  Vieth,	  W.,	  E.	  Curio,	  and	  U.	  Ernst.	  1980.	  The	  adaptive	  significance	  of	  avian	  mobbing.	  Cultural	  transmission	  of	  enemy	  recognition	  in	  blackbirds:	  Cross-­‐species	  tutoring	  and	  properties	  of	  learning.	  Animal	  Behaviour	  28:1217–1229.	  Warton, D., and F. Hui. 2011. The arcsine is asine: the analysis of proportions in ecology. Ecology 92:3–10. Willis,	  E.	  1958.	  Red-­‐crowned	  Ant-­‐Tanagers,	  Tawny-­‐crowned	  Greenlets	  and	  forest	  flocks.	  Wilson	  Bulletin	  105–106.	  Winterbottom,	  J.	  M.	  1943.	  On	  woodland	  bird	  parties	  in	  Norther	  Rhodesia.	  Ibis	  84:437–442.	  Wisz,	  M.	  S.,	  J.	  Pottier,	  W.	  D.	  Kissling,	  L.	  Pellissier,	  C.	  F.	  Damgaard,	  C.	  F.	  Dormann,	  M.	  C.	  Forchhammer,	  et	  al.	  2013.	  The	  role	  of	  biotic	  interactions	  in	  shaping	  distributions	  and	  realised	  assemblages	  of	  species :	  implications	  for	  species	  distribution	  modelling	  88:15–30.	  Wrege,	  P.	  H.,	  M.	  Wikelski,	  J.	  T.	  Mandel,	  T.	  Rassweiler,	  and	  I.	  D.	  Couzin.	  2005.	  Antbirds	  parasitize	  foraging	  army	  ants.	  Ecology	  86:555–559.	      71 Appendices  Appendix A  List of species Bird species participating in mixed-flocks in Manu gradient, Peru. Flock type and sample size is shown for each species.  Flock type Habitat Species n Lowland (Bamboo)  Bamboo forest Anabazenops dorsalis 16 Lowland (Bamboo)  Bamboo forest Automolus infuscatus 1 Lowland (Bamboo)  Bamboo forest Automolus melanopezus 3 Lowland (Bamboo)  Bamboo forest Automolus ochrolaemus 3 Lowland (Bamboo)  Bamboo forest Basileuterus chrysogaster 13 Lowland (Bamboo)  Bamboo forest Campylorhamphus trochilirostris 13 Lowland (Bamboo)  Bamboo forest Cercomacra manu 1 Lowland (Bamboo)  Bamboo forest Chlorophanes spiza 1 Lowland (Bamboo)  Bamboo forest Chlorothraupis carmioli 7 Lowland (Bamboo)  Bamboo forest Cranioleuca gutturata 1 Lowland (Bamboo)  Bamboo forest Cyanerpes caeruleus 2 Lowland (Bamboo)  Bamboo forest Cyanerpes cyaneus 1 Lowland (Bamboo)  Bamboo forest Cyanocompsa cyanoides 1 Lowland (Bamboo)  Bamboo forest Cymbilaimus sanctaemariae 5 Lowland (Bamboo)  Bamboo forest Dacnis cayana 1 Lowland (Bamboo)  Bamboo forest Dacnis lineata 1 Lowland (Bamboo)  Bamboo forest Dendrocincla fuliginosa 2 Lowland (Bamboo)  Bamboo forest Dendrocincla merula 1 Lowland (Bamboo)  Bamboo forest Drymophila devillei 1 Lowland (Bamboo)  Bamboo forest Dysithamnus mentalis 1 Lowland (Bamboo)  Bamboo forest Epinecrophylla erythrura 1 Lowland (Bamboo)  Bamboo forest Epinecrophylla ornata 3 Lowland (Bamboo)  Bamboo forest Euphonia rufiventris 1 Lowland (Bamboo)  Bamboo forest Euphonia xanthogaster 1 Lowland (Bamboo)  Bamboo forest Galbula cyanescens 1 Lowland (Bamboo)  Bamboo forest Glyphorynchus spirurus 12 Lowland (Bamboo)  Bamboo forest Herpsilochmus rufimarginatus 2 Lowland (Bamboo)  Bamboo forest Hyloctistes subulatus 1    72 Flock type Habitat Species n Lowland (Bamboo)  Bamboo forest Hylophilus hypoxanthus 1 Lowland (Bamboo)  Bamboo forest Hylophilus ochraceiceps 3 Lowland (Bamboo)  Bamboo forest Hypocnemis subflava 3 Lowland (Bamboo)  Bamboo forest Lanio versicolor 3 Lowland (Bamboo)  Bamboo forest Lathrotriccus euleri 2 Lowland (Bamboo)  Bamboo forest Leptopogon amaurocephalus 11 Lowland (Bamboo)  Bamboo forest Microrhopias quixensis 24 Lowland (Bamboo)  Bamboo forest Mionectes oleagineus 6 Lowland (Bamboo)  Bamboo forest Mionectes olivaceus 8 Lowland (Bamboo)  Bamboo forest Myioborus miniatus 1 Lowland (Bamboo)  Bamboo forest Myrmeciza fortis 1 Lowland (Bamboo)  Bamboo forest Myrmeciza goeldi 1 Lowland (Bamboo)  Bamboo forest Myrmoborus leucophrys 3 Lowland (Bamboo)  Bamboo forest Myrmotherula brachyura 2 Lowland (Bamboo)  Bamboo forest Myrmotherula longipennis 1 Lowland (Bamboo)  Bamboo forest Myrmotherula menetriesii 1 Lowland (Bamboo)  Bamboo forest Myrmotherula schystoginus 1 Lowland (Bamboo)  Bamboo forest Pachyramphus minor 1 Lowland (Bamboo)  Bamboo forest Pernostola lophotes 1 Lowland (Bamboo)  Bamboo forest Piaya cayana 2 Lowland (Bamboo)  Bamboo forest Pipra erytrocephala 1 Lowland (Bamboo)  Bamboo forest Ramphotrigon fuscicauda 1 Lowland (Bamboo)  Bamboo forest Saltator grossus 1 Lowland (Bamboo)  Bamboo forest Saltator maximus 1 Lowland (Bamboo)  Bamboo forest Simoxenops ucayalae 1 Lowland (Bamboo)  Bamboo forest Tachyphonus rufiventer 2 Lowland (Bamboo)  Bamboo forest Tangara chilensis 2 Lowland (Bamboo)  Bamboo forest Tangara gyrola 3 Lowland (Bamboo)  Bamboo forest Tangara mexicana 1 Lowland (Bamboo)  Bamboo forest Tangara punctata 1 Lowland (Bamboo)  Bamboo forest Tangara schrankii 3 Lowland (Bamboo)  Bamboo forest Terenura callinota 1 Lowland (Bamboo)  Bamboo forest Thamnomanes schistogynus 27 Lowland (Bamboo)  Bamboo forest Thamnophilus palliatus 2 Lowland (Bamboo)  Bamboo forest Thamnophilus schistaceus 7 Lowland (Bamboo)  Bamboo forest Trogon curucui 2 Lowland (Bamboo)  Bamboo forest Vireo olivaceus 3 Lowland (Bamboo)  Bamboo forest Xenopipo holochlora 1    73 Flock type Habitat Species n Lowland (Bamboo)  Bamboo forest Xenops minutus 1 Lowland (Bamboo)  Bamboo forest Xiphorhynchus elegans 9 Lowland (Bamboo)  Bamboo forest Xiphorhynchus guttatus 5 Lowland  Terra firme/Flooded  Anabazenops dorsalis 1 Lowland  Terra firme/Flooded  Ancistrops strigilatus 1 Lowland  Terra firme/Flooded  Arremon taciturnus 2 Lowland  Terra firme/Flooded  Automolus infuscatus 6 Lowland  Terra firme/Flooded  Automolus melanopezus 2 Lowland  Terra firme/Flooded  Automolus ochrolaemus 13 Lowland  Terra firme/Flooded  Automolus rufipileatus 2 Lowland  Terra firme/Flooded  Automolus sp 1 Lowland  Terra firme/Flooded  Bucco capensis 1 Lowland  Terra firme/Flooded  Campephilus melanoleucos 1 Lowland  Terra firme/Flooded  Campephilus rubricollis 1 Lowland  Terra firme/Flooded  Campylorhamphus trochilirostris 1 Lowland  Terra firme/Flooded  Capito auratus 6 Lowland  Terra firme/Flooded  Celeus gramicus 1 Lowland  Terra firme/Flooded  Celeus grammicus 1 Lowland  Terra firme/Flooded  Cercomacra manu 1 Lowland  Terra firme/Flooded  Cercomacra sp 1 Lowland  Terra firme/Flooded  Chlorophanes spiza 3    74 Flock type Habitat Species n Lowland  Terra firme/Flooded  Chlorothraupis carmioli 12 Lowland  Terra firme/Flooded  Cnemotriccus fuscatus 1 Lowland  Terra firme/Flooded  Coccyzus americanus 1 Lowland  Terra firme/Flooded  Coccyzus melacoryphus 1 Lowland  Terra firme/Flooded  Conopophaga peruviana 1 Lowland  Terra firme/Flooded  Cranioleuca gutturata 2 Lowland  Terra firme/Flooded  Cyanerpes caeruleus 3 Lowland  Terra firme/Flooded  Cyanerpes cyaneus 2 Lowland  Terra firme/Flooded  Cyanocompsa cyanoides 1 Lowland  Terra firme/Flooded  Cymbilaimus lineatus 4 Lowland  Terra firme/Flooded  Deconychura longicauda 1 Lowland  Terra firme/Flooded  Dendrocincla fuliginosa 3 Lowland  Terra firme/Flooded  Dendrocolaptes picumnus 1 Lowland  Terra firme/Flooded  Drymophila devillei 5 Lowland  Terra firme/Flooded  Epinecrophylla erythrura 14 Lowland  Terra firme/Flooded  Epinecrophylla leucophthalma 10 Lowland  Terra firme/Flooded  Epinecrophylla ornata 1 Lowland  Terra firme/Flooded  Eubucco richardsoni 2 Lowland  Terra firme/Flooded  Eubucco tucinkae 3 Lowland  Terra firme/Flooded  Euphonia rufiventris 2    75 Flock type Habitat Species n Lowland  Terra firme/Flooded  Euphonia sp 1 Lowland  Terra firme/Flooded  Euphonia xanthogaster 2 Lowland  Terra firme/Flooded  Furnaridae sp 3 Lowland  Terra firme/Flooded  Glyphorynchus spirurus 29 Lowland  Terra firme/Flooded  Gymnoderus foetidus 1 Lowland  Terra firme/Flooded  Habia rubica 3 Lowland  Terra firme/Flooded  Hemithraupis flavicollis 4 Lowland  Terra firme/Flooded  Hemithraupis guira 2 Lowland  Terra firme/Flooded  Hyloctistes subulatus 4 Lowland  Terra firme/Flooded  Hylophilus hypoxanthus 7 Lowland  Terra firme/Flooded  Hylophilus ochraceiceps 6 Lowland  Terra firme/Flooded  Icterus cayanensis 1 Lowland  Terra firme/Flooded  Isleria hauxwelli 2 Lowland  Terra firme/Flooded  Lanio versicolor 22 Lowland  Terra firme/Flooded  Lathrotriccus euleri 1 Lowland  Terra firme/Flooded  Lepidothrix coronata 1 Lowland  Terra firme/Flooded  Leptopogon amaurocephalus 11 Lowland  Terra firme/Flooded  Leptopogon superciliaris 1 Lowland  Terra firme/Flooded  Melanerpes cruentatus 1 Lowland  Terra firme/Flooded  Metopothrix aurantiaca 1    76 Flock type Habitat Species n Lowland  Terra firme/Flooded  Mionectes macconnelli 5 Lowland  Terra firme/Flooded  Mionectes oleagineus 8 Lowland  Terra firme/Flooded  Mionectes olivaceus 1 Lowland  Terra firme/Flooded  Monasa morpheus 3 Lowland  Terra firme/Flooded  Monasa morphoeus 7 Lowland  Terra firme/Flooded  Monasa nigrifrons 2 Lowland  Terra firme/Flooded  Monasa sp 1 Lowland  Terra firme/Flooded  Myiarchus tyrannulus 1 Lowland  Terra firme/Flooded  Myiobius barbatus 2 Lowland  Terra firme/Flooded  Myiodinastes maculatus 1 Lowland  Terra firme/Flooded  Myiopagis gaimardii 1 Lowland  Terra firme/Flooded  Myrmoborus leucophrys 1 Lowland  Terra firme/Flooded  Myrmoborus myotherinus 10 Lowland  Terra firme/Flooded  Myrmotherula axillaris 35 Lowland  Terra firme/Flooded  Myrmotherula brachyura 13 Lowland  Terra firme/Flooded  Myrmotherula longipennis 16 Lowland  Terra firme/Flooded  Myrmotherula menetriesii 23 Lowland  Terra firme/Flooded  Myrmotherula multostriata 1 Lowland  Terra firme/Flooded  Pachyramphus minor 5 Lowland  Terra firme/Flooded  Parula pitiayumi 1    77 Flock type Habitat Species n Lowland  Terra firme/Flooded  Phillydor erythropterum 2 Lowland  Terra firme/Flooded  Philydor erythrocercum 4 Lowland  Terra firme/Flooded  Piculus leucolaemus 1 Lowland  Terra firme/Flooded  Pipra chloromeros 1 Lowland  Terra firme/Flooded  Pipra fascicauda 3 Lowland  Terra firme/Flooded  Pipra fasciicauda 1 Lowland  Terra firme/Flooded  Poecilotriccus albifacies 1 Lowland  Terra firme/Flooded  Pygiptila stellaris 13 Lowland  Terra firme/Flooded  Ramphotrigon fuscicauda 3 Lowland  Terra firme/Flooded  Saltator grossus 2 Lowland  Terra firme/Flooded  Saltator maximus 4 Lowland  Terra firme/Flooded  Sclerurus mexicanus 1 Lowland  Terra firme/Flooded  Simoxenops ucayalae 1 Lowland  Terra firme/Flooded  Sittasomus griseicapillus 1 Lowland  Terra firme/Flooded  Tachyphonus cristatus 1 Lowland  Terra firme/Flooded  Tachyphonus luctuosus 7 Lowland  Terra firme/Flooded  Tachyphonus rufiventer 18 Lowland  Terra firme/Flooded  Tangara callophrys 1 Lowland  Terra firme/Flooded  Tangara chilensis 12 Lowland  Terra firme/Flooded  Tangara gyrola 9    78 Flock type Habitat Species n Lowland  Terra firme/Flooded  Tangara mexicana 2 Lowland  Terra firme/Flooded  Tangara schrankii 22 Lowland  Terra firme/Flooded  Tangara velia 1 Lowland  Terra firme/Flooded  Tangara xanthogastra 5 Lowland  Terra firme/Flooded  Terenotriccus erythrurus 2 Lowland  Terra firme/Flooded  Thamnomanes ardesiacus 22 Lowland  Terra firme/Flooded  Thamnomanes schistogynus 22 Lowland  Terra firme/Flooded  Thamnophilus aethiops 1 Lowland  Terra firme/Flooded  Thamnophilus schistaceus 11 Lowland  Terra firme/Flooded  Thraupis palmarum 1 Lowland  Terra firme/Flooded  Thripophaga fusciceps 1 Lowland  Terra firme/Flooded  Tityra semifasciata 1 Lowland  Terra firme/Flooded  Todirostrum chrysocrotaphum 1 Lowland  Terra firme/Flooded  Tolmomyias assimilis 2 Lowland  Terra firme/Flooded  Trogon collaris 2 Lowland  Terra firme/Flooded  Trogon curucui 1 Lowland  Terra firme/Flooded  Trogon melanurus 1 Lowland  Terra firme/Flooded  Trogon violaceus 2 Lowland  Terra firme/Flooded  Trogon viridis 1 Lowland  Terra firme/Flooded  Vireo leucophrys 4    79 Flock type Habitat Species n Lowland  Terra firme/Flooded  Vireo olivaceus 2 Lowland  Terra firme/Flooded  Xenopipo holochlora 1 Lowland  Terra firme/Flooded  Xenops minutus 8 Lowland  Terra firme/Flooded  Xenops rutilans 3 Lowland  Terra firme/Flooded  Xenops tenuirostris 7 Lowland  Terra firme/Flooded  Xiphocolaptes promeropirhynchus 1 Lowland  Terra firme/Flooded  Xiphorhynchus elegans 21 Lowland  Terra firme/Flooded  Xiphorhynchus guttatus 17 Lowland  Terra firme/Flooded  Xyphorinchus picus 1 Low-montane Cloud forest Anabacerthia striaticollis 18 Low-montane Cloud forest Anisognathus somptuosus 1 Low-montane Cloud forest Atlapetes melanolaemus 2 Low-montane Cloud forest Automolus infuscatus 1 Low-montane Cloud forest Automolus ochrolaemus 8 Low-montane Cloud forest Basileuterus bivittatus 1 Low-montane Cloud forest Basileuterus chrysogaster 1 Low-montane Cloud forest Basileuterus coronatus 5 Low-montane Cloud forest Basileuterus signatus 1 Low-montane Cloud forest Basileuterus tristriatus 16 Low-montane Cloud forest Chiroxiphia boliviana 2 Low-montane Cloud forest Chlorochrysa calliparaea 36 Low-montane Cloud forest Chlorophanes spiza 2 Low-montane Cloud forest Chlorophonia cyanea 5 Low-montane Cloud forest Chlorospingus flavigularis 40 Low-montane Cloud forest Chlorospingus ophthalmicus 9 Low-montane Cloud forest Chlorospingus parvirostris 3 Low-montane Cloud forest Cissopis leverianus 1 Low-montane Cloud forest Coereba flaveola 1 Low-montane Cloud forest Colaptes punctigula 1 Low-montane Cloud forest Colaptes rubiginosus 3    80 Flock type Habitat Species n Low-montane Cloud forest Conopias cinchoneti 2 Low-montane Cloud forest Contopus fumigatus 1 Low-montane Cloud forest Cranioleuca curtata 1 Low-montane Cloud forest Creurgops dentatus 7 Low-montane Cloud forest Cyanerpes caeruleus 1 Low-montane Cloud forest Cyanerpes cyaneus 1 Low-montane Cloud forest Cyanocorax yncas 1 Low-montane Cloud forest Dacnis cayana 1 Low-montane Cloud forest Dendrocolaptes picumnus 1 Low-montane Cloud forest Diglossa cyanea 1 Low-montane Cloud forest Diglossa glauca 5 Low-montane Cloud forest Dysithamnus mentalis 1 Low-montane Cloud forest Elaenia albiceps 1 Low-montane Cloud forest Elaenia pallatangae 1 Low-montane Cloud forest Elaenia parvirostris 1 Low-montane Cloud forest Entomodestes leucotis 5 Low-montane Cloud forest Eubbuco versicolor 12 Low-montane Cloud forest Euphonia mesochrysa 6 Low-montane Cloud forest Euphonia xanthogaster 20 Low-montane Cloud forest Galbula cyanescens 1 Low-montane Cloud forest Hemispingus melanotis 16 Low-montane Cloud forest Hemithraupis guira 2 Low-montane Cloud forest Hemitriccus rufigularis 1 Low-montane Cloud forest Herpsilochmus axillaris 2 Low-montane Cloud forest Iridophanes pulcherrimus 3 Low-montane Cloud forest Iridosornis analis 12 Low-montane Cloud forest Lathrotriccus euleri 2 Low-montane Cloud forest Legatus leucophagius 1 Low-montane Cloud forest Lepidocolaptes lacrymiger 5 Low-montane Cloud forest Leptopogon superciliaris 37 Low-montane Cloud forest Malacoptila  fulvogularis 1 Low-montane Cloud forest Mionectes olivaceus 2 Low-montane Cloud forest Mionectes striaticollis 19 Low-montane Cloud forest Myarchus cephalotes 1 Low-montane Cloud forest Myioborus miniatus 41 Low-montane Cloud forest Myiophobus fasciatus 1 Low-montane Cloud forest Myiophobus inornatus 1 Low-montane Cloud forest Myizetetes cayanensis 1    81 Flock type Habitat Species n Low-montane Cloud forest Myrmotherula longicauda 2 Low-montane Cloud forest Myrmotherula schisticolor 4 Low-montane Cloud forest Odontorchilus branickii 6 Low-montane Cloud forest Pachyramphus policopterus 1 Low-montane Cloud forest Pachyramphus versicolor 1 Low-montane Cloud forest Parula pitiayumi 8 Low-montane Cloud forest Philydor erythrocercum 1 Low-montane Cloud forest Philydor ruficaudatum 1 Low-montane Cloud forest Phyllomyias cinereiceps 10 Low-montane Cloud forest Phylloscartes ophthalmicus 10 Low-montane Cloud forest Phylloscartes poecilotis 1 Low-montane Cloud forest Phylloscartes ventralis 2 Low-montane Cloud forest Piaya cayana 5 Low-montane Cloud forest Pipraeidea melanonota 1 Low-montane Cloud forest Piranga leucoptera 1 Low-montane Cloud forest Piranga olivacea 1 Low-montane Cloud forest Poecilotriccus plumbeiceps 1 Low-montane Cloud forest Premnoplex brunnescens 1 Low-montane Cloud forest Pyrrhomyias cinnamomeus 7 Low-montane Cloud forest Ramphocelus carbo 6 Low-montane Cloud forest Rhynchocyclus fulvipectus 1 Low-montane Cloud forest Saltator maximus 5 Low-montane Cloud forest Sclerurus mexicanus 1 Low-montane Cloud forest Synallaxis azarae 1 Low-montane Cloud forest Tachyphonus rufiventer 3 Low-montane Cloud forest Tangara arthus 25 Low-montane Cloud forest Tangara chilensis 9 Low-montane Cloud forest Tangara chrysotis 3 Low-montane Cloud forest Tangara cyanicollis 18 Low-montane Cloud forest Tangara cyanotis 1 Low-montane Cloud forest Tangara gyrola 10 Low-montane Cloud forest Tangara nigroviridis 6 Low-montane Cloud forest Tangara parzudakii 1 Low-montane Cloud forest Tangara punctata 12 Low-montane Cloud forest Tangara ruficervix 6 Low-montane Cloud forest Tangara xanthocephala 6 Low-montane Cloud forest Thamnophilus  doliatus 1 Low-montane Cloud forest Thamnophilus palliatus 2    82 Flock type Habitat Species n Low-montane Cloud forest Thamnophilus schistaceus 1 Low-montane Cloud forest Thamnophilus unicolor 1 Low-montane Cloud forest Thlypopsis ruficeps 1 Low-montane Cloud forest Thraupis episcopus 2 Low-montane Cloud forest Thraupis palmarum 2 Low-montane Cloud forest ThricHothraupis melanops 2 Low-montane Cloud forest Thripadectes  melanorhynchus 8 Low-montane Cloud forest Tityra semifasciata 1 Low-montane Cloud forest Tolmomyias assimilis 1 Low-montane Cloud forest Trichothraupis melanops 5 Low-montane Cloud forest Trogon  personatus 2 Low-montane Cloud forest Trogon personatus 1 Low-montane Cloud forest Turdus nigriceps 2 Low-montane Cloud forest Xenopipo unicolor 1 Low-montane Cloud forest Xenops minutus 1 Low-montane Cloud forest Xenops rutilans 1 Low-montane Cloud forest Xiphocolaptes promeropirhynchus 4 Low-montane Cloud forest Xiphorhynchus ocelatus 2 Low-montane Cloud forest Xiphorhynchus triangularis 12 Low-montane Cloud forest Zimmerius bolivianus 6 High-montane Montane/ Puna Ampelion rubrocristatus 2 High-montane Montane/ Puna Anisognathus igniventris 19 High-montane Montane/ Puna Arremon torquatus 1 High-montane Montane/ Puna Atlapetes melanolaemus 16 High-montane Montane/ Puna Basileuterus luteoviridis 9 High-montane Montane/ Puna Basileuterus signatus 2 High-montane Montane/ Puna Buthraupis montana 14 High-montane Montane/ Puna Cacicus chrysonotus 3 High-montane Montane/ Puna Catamblyrhynchus diadema 1 High-montane Montane/ Puna Chlorophonia cyanea 1 High-montane Montane/ Puna Chlorornis riefferii 17 High-montane Montane/ Puna Chlorospingus ophthalmicus 1 High-montane Montane/ Puna Chlorospingus parvirostris 1 High-montane Montane/ Puna Cinnycerthia fulva 3 High-montane Montane/ Puna Cnemoscopus rubrirostris 5 High-montane Montane/ Puna Colaptes rivolii 1    83 Flock type Habitat Species n High-montane Montane/ Puna Conirostrum albifrons 9 High-montane Montane/ Puna Conirostrum ferrugineiventre 2 High-montane Montane/ Puna Conirostrum sitticolor 1 High-montane Montane/ Puna Conorostrum sitticolor 4 High-montane Montane/ Puna Cranioleuca marcapatae 3 High-montane Montane/ Puna Creurgops dentatus 1 High-montane Montane/ Puna Cyanolyca viridicyanus 4 High-montane Montane/ Puna Delothraupis castaneoventris 3 High-montane Montane/ Puna Dendrocincla tyrannina 2 High-montane Montane/ Puna Diglossa brunneiventris 1 High-montane Montane/ Puna Diglossa caerulescens 1 High-montane Montane/ Puna Diglossa cyanea 40 High-montane Montane/ Puna Diglossa mystacalis 3 High-montane Montane/ Puna Diglossa sittoides 4 High-montane Montane/ Puna Drymotoxeres pucherani 1 High-montane Montane/ Puna Elaenia albiceps 1 High-montane Montane/ Puna Elaenia gigas 1 High-montane Montane/ Puna Elaenia obscura 1 High-montane Montane/ Puna Elaenia pallatangae 20 High-montane Montane/ Puna Entomodestes leucotis 1 High-montane Montane/ Puna Hemispingus atropileus 30 High-montane Montane/ Puna Hemispingus parodii 2 High-montane Montane/ Puna Hemispingus superciliaris 8 High-montane Montane/ Puna Hemispingus trifasciatus 6 High-montane Montane/ Puna Hemispingus xanthophthalmus 7 High-montane Montane/ Puna Hemitriccus granadensis 10 High-montane Montane/ Puna Iridosornis jelskii 8 High-montane Montane/ Puna Knipolegus signatus 1 High-montane Montane/ Puna Lepidocolaptes lacrymiger 4 High-montane Montane/ Puna Leptopogon superciliaris 2 High-montane Montane/ Puna Margarornis squamiger 21 High-montane Montane/ Puna Mecocerculus leucophrys 14 High-montane Montane/ Puna Mecocerculus stictopterus 31 High-montane Montane/ Puna Mionectes striaticollis 4 High-montane Montane/ Puna Myiarchus tuberculifer 3    84 Flock type Habitat Species n High-montane Montane/ Puna Myioborus melanocephalus 37 High-montane Montane/ Puna Myioborus miniatus 1 High-montane Montane/ Puna Myiophobus ochraceiventris 1 High-montane Montane/ Puna Myrmotherula axillaris 1 High-montane Montane/ Puna Ochthoeca rufipectoralis 1 High-montane Montane/ Puna Pachyramphus versicolor 4 High-montane Montane/ Puna Phylloscartes ventralis 1 High-montane Montane/ Puna Pipraeidea melanonota 1 High-montane Montane/ Puna Pipreola intermedia 1 High-montane Montane/ Puna Piranga flava 1 High-montane Montane/ Puna Premnornis guttuligera 1 High-montane Montane/ Puna Pseudocolaptes boissonneautii 8 High-montane Montane/ Puna Pyrrhomyias cinnamomeus 8 High-montane Montane/ Puna Synallaxis azarae 6 High-montane Montane/ Puna Tangara nigroviridis 1 High-montane Montane/ Puna Tangara vassorii 9 High-montane Montane/ Puna Thlypopsis ornata 1 High-montane Montane/ Puna Thlypopsis ruficeps 10 High-montane Montane/ Puna Thraupis bonariensis 1 High-montane Montane/ Puna Thraupis cyanocephala 19 High-montane Montane/ Puna Troglodytes solstitialis 3 High-montane Montane/ Puna Trogon personatus 2 High-montane Montane/ Puna Veniliornis nigriceps 2 High-montane Montane/ Puna Vireo leucophrys 1 High-montane Montane/ Puna Xiphorhynchus triangularis 3          85 Appendix B  Model selection  Model	  selection	  results	  from	  generalized	  linear	  models	  for	  flock	  size	  (Num	  of	  individuals)	  across	   the	   elevational	   gradient.	   Explanatory	   variables	   include	   elevation	   (Elev),	   mean	  canopy	  height	  (Canopy),	  and	  number	  of	  trees	  (Trees).	  For	  each	  fitted	  model	  the	  number	  of	  parameters	   (k),	   change	   in	   corrected	   quasi-­‐Akaike	   from	   the	  model	   with	   the	   lowest	   QAIC	  value	  (Delta_QAICc),	  QAIC	  weights	  (QAICcWt)	  are	  shown.	  	  Response	  variable	  Model	   	  k	   QAICc	   Delta_QAICc	   QAICcWt	  	  Num_individuals	   Elevation+Canopy+Trees	  	   5	   352.42	   0.00	   0.84	   	  	   Elevation+Canopy	   4	   357.66	   5.24	   0.06	   	  	   Elevation	  	   3	   358.31	   5.89	   0.04	   	  	   Elevation+Trees	   4	   359.10	   6.67	   0.03	   	  	   Canopy	  +	  Trees	   4	   359.22	   6.79	   0.03	   	  	   Canopy	   3	   368.76	   16.34	   0.00	   	  	   Intercept	   2	   372.33	   19.90	   0.00	   	  Trees	   3	   372.53	   20.11	   0.00	        86 Appendix C  Model flock size Mixed-species flock size in the Manu study region varying by (a) canopy height, (b) number of trees and (c) elevation. Best-fitted model was used to plot canopy height, number of trees and elevation. The solid line is the predicted value of flock size as a function of each variable, filling in the other explanatory variables to the median. Each dot represents a obsered flock. The grey shading shows the confidence intervals based on the standard errors of the estimates.                                     0 10 20 30 400102030405060a) Canopy heightCanopy heightFlock size (Number of individuals)0 10 20 30 40 50 600102030405060b) Number of treesNumber of treesFlock size (Number of individuals)0 500 1000 1500 2000 2500 3000 35000102030405060c) ElevationElevation Flock size (Number of individuals)    87 Appendix D  Regression canopy height Regression of Canopy height and Elevation in Manu region, Peru.         0 500 1000 1500 2000 2500 3000 350001020304050 Elevation Canopy height (m)   88 Appendix E  Regression  Regression of Number of trees and Elevation in Manu region, Peru.        0 500 1000 1500 2000 2500 3000 350001020304050Elevation Number of trees    89 Appendix F  Foraging guilds Proportion of species of the community joining mixed flocks for each foraging guild across elevation is shown.    0 500 1000 1500 2000 2500 3000 35000.00.20.40.60.81.0InsectivorusElevationProportion of species of the community joining Mixed flocks0 500 1000 1500 2000 2500 3000 35000.00.40.60.81.0Omnivores0.2ElevationProportion of species  of the community joining Mixed flocks0 500 1000 1500 2000 2500 3000 35000.00.20.40.60.81.0NectarivorusElevationProportion of species of the community joining Mixed flocks0 500 1000 1500 2000 2500 3000 35000.00.20.40.60.81.0FrugivorusElevationProportion of species of the community joining Mixed flocks   90  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0303480/manifest

Comment

Related Items