|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 LINKING COMPOSITION, STRUCTURE AN D FUNCTIONS OF BIODIVERSITY: RELATIONSHIPS AMONG EPIPHYTES, INVERTEBRATES AND BIRDS IN THE CANOPY OF CHILEAN TEM PERATE RAINFORESTS By IVAN ANDRES DIAZ A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Ivan Andres Diaz
3 To my parents, Nora Romero and Fanor Daz To the nature, which always shares with me its secrets To the canopy, the place where trees let you see the forest
4 ACKNOWLEDGMENTS I am deeply thankful to my friend and adviso r, Dr. Kathryn E. Sieving for her permanent support, lessons, friendship, and good will. She has always extended her supportive hand, with a warm smile and the right word when it was need ed. Katie has been the kind of scientist and advisor I would like to be, and she has opened to me a world of possibi lities because of her generosity and dedication. She honors the professi on of being professor. I am also deeply thankful to Dr. Juan J. Armesto, who is my go od friend and advisor, opening to me the door of science, giving me the shoes to walk into the fore st and unlocking the gate of frontiers to let me go to the world beyond. Juan has been a very im portant person in my career, giving me the opportunity and the support to con tinue in the scientific way, an opportunity that finally allowed me to study in this country and in this univers ity. I thank my Advisory Committee, Drs. Jack Putz, Doug Levey, Emilio Bruna, Meg Lowman and Graeme Cumming, for their permanent support, lessons, encouragement, and the willing to support and advise my work at any time. I want to express all my grat itude to Ms. Katia Velazquez an d Mr. Javier de la Calle for their generosity with ho using and for allowing me to use their forest, which they love. I am also very thankful to the Corporacin Nacional Forest al CONAF (Chile), which let me use facilities in Chilo National Park and supporte d my work there. I also want express my gratitude to the Institute of Ecology and Biodiversity for providi ng lab facilities, and to Departamento de Ecologa y Evolucin, Universidad Austral de Chile for providing access and facilities at the Fundo San Martn site. Also, I would like to acknowledge many friends at this time, includ ing those who helped me from the early stages of my career. Pablo Es pejo taught me about scie nce and reptiles, Juan Carlos Torres shared with me the world of birds; Dina Robles s upported me as a young, too young kid in science. Mary Willson was our library of knowledge, offering advice and clarity.
5 Sharon Reid, Olga Barbosa, Claudi a Papic, Mauricio Soto, Eduar do Soto, Luis Cavieres, Roberto Nespolo, Cintia Cornelius and all members of Senda Darwin Foundation, have been key in the development of my career. I am very thankful for the friends who worked with me in the field: Maurice Pea, who became a right hand in the field, Camila Tejo, Natalia Carrasco, Mara Paz Pea, Daniela Manuschevich, Pablo Necochea, Me lanie, Tio Franz, Juan Larran, Fernanda Salinas, Juan Luis Celis and Anita Venegas. I am very thankful to Chilo National Park rangers Americo Panichini, Francisco De lgado, Maestro Lolo, don Paco, and in particular to Don Mario Huinao who always helped us with the very best of his good will. I am al so very thankful to Mariano Millacura for all his suppor t in the field. I am grateful to the Institute of Tropical Studies and Conservation, to Peter Lahanas a nd to Joe Maer, who showed and taught me the secrets of the tropical forest, and how to know the world of the canopy. I want also to express my deepest gratitude to Gary Machlis and the Canon National Parks Science Scholar Program, which funded me for three years, to the Fu lbright program, to Fondecyt Grant 1050225 for funding support, and to the Department of Wild life Ecology and Conservation, for being a great place to learn about science and life. To my bi g friends of the Mafia Latina, all Maestros: Santiago Espinoza, Alejandro Paredes, Rafael Reyna, Luis Ramos, Sonia Canavelli, Marcela Machicote, Sheda Morshed, to the friends who or ganized our little world in the office in the basement Olga Montenegro, Alejandro Paredes, to my old roommates Mutsuo Nakamura and Osvaldo Jordan, to Michaela Speirs. Finally, I want to express my appreciation to my family who has always supported me, in the good and the hard times; and to Wara Marc elo, who gave me the comprehension I wanted, the support I needed, the smile I liked, and the love I loved.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES................................................................................................................ .......11 ABSTRACT....................................................................................................................... ............13 CHAPTER 1 LINKAGES AMONG STRUCTURE, CO MPOSITION AND FUNCTIONS OF BIODIVERSITY IN THE FOREST CANOPY.....................................................................15 Theoretical Background......................................................................................................... .15 Composition, Structure and Func tions in Forest Canopies....................................................16 The Forest Canopy: An Ecological Frontier...................................................................16 Temperate Rainforest of Southern Chile.........................................................................17 Links Among Epiphytes, Invertebrates a nd Birds in the Rain Forest Canopy................19 2 EPIPHYTE SPECIES RICHNE SS AND BIOMASS LOADS ON Eucryphia cordifolia (CUNONIACEAE) AND Aextoxicon punctatum (AEXTOXICACEAE), TWO CANOPY EMERGENT TREES IN CH ILEAN TEMPERATE RAIN FORESTS...............23 Introduction................................................................................................................... ..........23 Methods........................................................................................................................ ..........26 Study Area..................................................................................................................... ..26 Study Design...................................................................................................................27 Tree Structure................................................................................................................. .28 Epiphyte Species Richness..............................................................................................28 Epiphyte Biomass............................................................................................................28 Ecological Functions Conducted by Epiphytes...............................................................30 Results........................................................................................................................ .............31 Structural Features of Eucryphia cordifolia and Aextoxicon punctatum Trees...............31 Species Richness and Dist ribution of Epiphytes on E. cordifolia and A. punctatum ......32 Epiphytic Biomass...........................................................................................................33 Epiphyte Functions..........................................................................................................35 Discussion..................................................................................................................... ..........36 Comparative Diversity of Epiphytes in Southern Chile Temperate Rainforests.............36 Comparative Diversity of Epiphytes with Other Tropical and Temperate Rainforests...................................................................................................................38 Floristic Relationships between Chilean Ep iphytes and those from the Neotropics and Australia-New Zealand.........................................................................................39 Epiphyte Vertical Distributions.......................................................................................41
7 Epiphyte Biomass............................................................................................................43 Epiphyte Functions..........................................................................................................45 Importance of Large Trees with Epiphytes.....................................................................48 3 EFFECTS OF EPIPHYTE LOADS ON I NVERTEBRATE SPECIES RICHNESS AND ABUNDANCE ON THE CANOPY OF Eucryphia cordifolia (CUNONIACEAE), AN EMERGENT TREE IN CHILEAN TEMPERATE RAIN FORESTS..................................65 Introduction................................................................................................................... ..........65 Methods........................................................................................................................ ..........67 Study Area..................................................................................................................... ..67 Study Design...................................................................................................................68 Tree Characterization......................................................................................................68 Epiphytic Biomass...........................................................................................................69 Invertebrates in the Tree Crown......................................................................................70 Invertebrates in th e Epiphytic Layer...............................................................................70 Epiphytic Versus Tree Crown Invertebrate Biomass......................................................71 Data Analysis.................................................................................................................. .71 Results........................................................................................................................ .............73 Tree and Epiphytic Biomass Characterization................................................................73 Invertebrate Species Ric hness in the Tree Crown...........................................................74 Biomass of Tree Crow n Invertebrates.............................................................................74 Invertebrate Biomass in the Epiphytic Layer..................................................................75 Invertebrate Biomass in Epi phytes Versus the Tree Crown............................................76 Discussion..................................................................................................................... ..........77 Species Richness in Tree Crowns With and Without Epiphytes.....................................77 Invertebrate Richness in Comparis on with Other Forest Ecosystems.............................77 Invertebrate Biomass in Tree Crow ns with and without Epiphytes................................79 Effect of Epiphytes on Invert ebrate Richness and Biomass............................................80 Possible Functional Consequences of Epiphyte Invertebrates........................................81 Diversity of Invertebrates in Chilean Forest Canopies....................................................82 Importance of Epiphytes for Inverteb rates in Chilean Forest Canopies..........................83 4 LINKING COMPOSITION, STRUCTURE AN D FUNCTIONS OF BIODIVERSITY: EFFECT OF EPIPHYTE LOADS ON CA NOPY BIRDS IN THE TEMPERATE RAINFORESTS OF SOUTHERN CHILE............................................................................94 Introduction................................................................................................................... ..........94 Study Design................................................................................................................... ........95 Methods........................................................................................................................ ..........97 Study System...................................................................................................................97 Study Sites.................................................................................................................... ...98 Preliminary Characterizati on of Canopy Bird Community.............................................99 Tree Selection and Epiphyte Removal..........................................................................100 Bird Surveys in Experimental and Control Tree Canopies...........................................100 Substrata Availability....................................................................................................103 Comparative Bird Surveys on Plot s Varying in Epiphyte Biomass..............................104
8 Results........................................................................................................................ ...........105 Preliminary Bird Community Surveys in Guabn and Chilo National Park...............105 Bird Species Visiting Trees With and Without Epiphytes............................................106 Bird Foraging Activity in Re lation to Epiphyte Loads.................................................107 Experimental versus control trees..........................................................................107 Experimental pairs vs. pairs natu rally varying in epiphyte loads..........................108 Foraging Substrata Use-Availability Analysis..............................................................108 Comparative Survey of Bird Community St ructure in relation to Epiphyte Biomass..110 Discussion..................................................................................................................... ........111 Epiphyte Influences on Bird Comm unity Structure at Two Scales...............................111 Effect of Epiphytes on Foraging Substrate Use............................................................113 Influence of Epiphytes on Bird Abundance at the Plot Scale.......................................114 Influence of Epiphytes on Bird Species Richness.........................................................115 Possible Mechanisms Underlying Epiphyte-Bird Interactions......................................115 Conclusions: Relationships between La rge Trees, Epiphytes and Birds......................117 5 CONCLUSIONS: LINKING EPIPHYTES, IN VERTEBRATES, AND BIRDS IN THE CANOPY OF CHILEAN RAINFORESTS: THEORETICAL A SPECTS AND THE VALUE OF LARGE OLD TREES......................................................................................135 Introduction................................................................................................................... ........135 Significant Ecological Func tions of Canopy Epiphytes................................................136 How Structure Supports Diversity of Plan ts and Animals in Chilean Rainforest Canopy: A Summary..................................................................................................137 Linking tree-epiphyte structur e and composition to the func tional role of birds in the canopy..................................................................................................................139 Toward an Integrative Theory Linking Composition, Structure and Function of Ecosystems.................................................................................................................140 The Treebeard Hypothesis....................................................................................................142 Tree Life Cycle................................................................................................................ .....143 Implications for Conservation..............................................................................................145 Future Research................................................................................................................ ....146 LIST OF REFERENCES.............................................................................................................152 BIOGRAPHICAL SKETCH.......................................................................................................167
9 LIST OF TABLES Table page 2-1 General structural features of the large E. cordifolia and A. punctatum trees sampled for epiphytes in Valdivian temperat e rain forests, Chilo, Chile.......................................50 2-2 Relative frequency of vasc ular and non-vascular epiphyte species on two large trees of E. cordifolia and one tree of A. punctatum in coastal rain forests of Chilo Island, southern Chile................................................................................................................. ...51 2-3 Epiphyte distribution along the ver tical profile on tw o large trees of E. cordifolia and one tree of A. punctatum in coastal rain forests of Chilo Island, southern Chile.............53 2-4 Dry biomass (kg) of the hemi-epiphytic tree R. laetevirens the bromeliad F. bicolor and other epiphytes growi ng on large canopy trees of E. cordifolia and A. punctatum trees in coastal rain forests of Chilo Island, southern Chile.............................................55 2-5 Characterization of the physical proper ties and nitrogen content in arboreal and forest soils in the canopy of two large Eucryphia cordifolia trees in Guabn forests, Chilo Island, Chile...........................................................................................................56 2-6 Number of species and gene ra of vascular and non-vasc ular epiphytes in different tropical and temperate rainforests......................................................................................57 2-7 Epiphytic biomass (epiphytes plus arboreal soil, dry weights) estimated for tropical and temperate forests of the world.....................................................................................59 3-1 Characteristics of the Eucryphia cordifolia trees used in this st udy in coastal forest of Guabn, Chilo Island, Chile.............................................................................................85 3-2 Piankas overlap index for the number of individuals of each morphospecies of invertebrates present in E. cordifolia trees with and without epiphytes............................86 3-3 Total biomass (g) and number of indivi duals (in parenthesis) collected of the principal orders of invertebrates in the flight interception and eclector traps, all located in the crown of E. cordifolia trees with and without epiphytes............................87 3-4 Repeated Measures MANOVA on biomass (g) of different orders of invertebrates in trees with and without epiphytes........................................................................................88 3-5 Total biomass (g) and number of collect ed individuals (in parenthesis) of the principal orders of invertebrates collected from samples of epiphytic biomass in two E. cordifolia trees...............................................................................................................89 3-6 Epiphytic biomass of invertebrate s (g dry mass) in two components of E. cordifolia canopy; the tree crown and the epiphytic biomass............................................................90
10 4-1 Bird species present in the study sites, their abundance (individuals/point/day, 14 months average for Guabn, 5 months av erage for Chilo NP), food habits and habitat use.................................................................................................................... ....119 4-2 Total bird visits for each species in the pair of trees with and without epiphytes (by manual removal)..............................................................................................................121 4-3 General structural features of the large E. cordifolia trees sampled for epiphytes in Valdivian temperate rain forests, Chilo, Chile...............................................................122 4-4 Analysis of substrata used by canop y birds versus substrata available...........................123 4-5 Results of the General Linear Model comparing the abundance of canopy birds in forest plots of 25m radius in Guabn and Fundo San Martn, southern Chile................124 4-6 Abundance (Mean individuals/ plot/ day) of birds in the 25 m plots on Guabn and Fundo San Martn, southern Chile...................................................................................125
11 LIST OF FIGURES Figure page 1-1 Three attributes of biodiversity..........................................................................................21 1-2 Theoretical model with the hypothe sis proposed in this dissertation................................22 2-1 Map of study sites (black dot s) in lowland coastal forest s of Chilo Island, southern Chile.......................................................................................................................... .........61 2-2 General shape of Aextoxicon punctatum tree, showing the distribution of the principal vascular epiphytes...............................................................................................62 2-3 General shape of Eucryphia cordifolia tree, showing the distribution of its two principal vascular epiphytes...............................................................................................63 2-4 Relationship between trunk diameter at breast height (DBH in cm) of Eucryphia cordifolia trees and their epiphytic biomass ( kg) in forests of Guabn, Chilo Island, southern Chile................................................................................................................. ...64 3-1 Map of the study site in Guabn, at the co ast in the north of Ch ilo Island, southern Chile. This region is showed in dark color in the inset map..............................................91 3-2 Rarefaction analysis for th e total number of morphospecies as a function of the total number of individuals capt ured in the flight inter ception and eclector traps....................92 3-3 Average abundance (g invertebrate/ surv ey) 1SE of invertebrates captured in the crown of Eucryphia cordifolia trees with and without epiphytes in Guabn forest, Chilo Island, southern Chile.............................................................................................93 4-1 Map of study sites (black dots) in lowl and coastal forests of Chilo Island, and Fundo San Martn, Valdivia, southern Chile...................................................................126 4-2 Monthly variation in mean bird abunda nce (individuals/point/day averaged over all points censused in each monthly survey; N= 14) and bird richness (total number of species detected per mont hly survey) in the Guabn forest site, Chilo Island, southern Chile................................................................................................................. .127 4-3 Rarefaction analysis on the number of sp ecies found in the trees with and without epiphytes and the number of bird visits...........................................................................128 4-4 Rate of bird visits (average) to experime ntal trees with and wi thout epiphytes. Data from December 2005 to April 2007.................................................................................129 4-5 Rate of bird visits (avera ge) plus/minus one standard erro r to experimental trees with and without epiphytes by the four most frequent bird species.........................................130
12 4-6 Total bird visits to the Experimental Pa irs (Pairs 1 and 2) an d Naturally different Pairs (Pairs 4 and 5). Only birds th at were feeding were considered..............................131 4-7 Percentage of bird visits to each subs trata by canopy birds in the Guabn forest site, Chilo Island, southern Chile (all data pooled)...............................................................132 4-8 Bird abundance as a function of the am ount of epiphytes on 25 m radius plots in Guabn and Fundo San Martn forest..............................................................................133 4-9 Regression analysis between the variance (e very five points) and epiphyte biomass in in Guabn and Fundo San Martn forests, southern Chile...............................................134 5-1 Multiple links among structure, diversity and functions in the canopy of a temperate rainforests in Chilo, southern Chile...............................................................................148 5-2 Relationships among structure, compositio nal diversity and functions in temperate rainforests.................................................................................................................... .....149 5-3 Complete life cycle of a tree, from seedli ng establishment (on fallen logs), to old age and after death as snag, and fallen tree............................................................................150 5-4 Treebeard Hypothesis......................................................................................................151
13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy LINKING COMPOSITION, STRUCTURE AN D FUNCTIONS OF BIODIVERSITY: RELATIONSHIPS AMONG EPIPHYTES, INVERTEBRATES AND BIRDS IN THE CANOPY OF CHILEAN TEMPERATE RAINFORESTS By Ivan Andres Diaz May 2009 Chair: Kathryn E. Sieving Major: Wildlife Ecology and Conservation Temperate rainforests of southern South Am erica are characterized by a multi-layered forest canopy with old emergent trees densel y covered by epiphytes. Th e species richness and community structure of this forest canopy rema ins largely unknown. In this dissertation, I address this gap in knowledge by testing the follo wing linked hypotheses; 1) that large old trees support a rich community of epiphytes that, in turn, 2) support invertebrate and 3) bird communities in the forest canopy. Working in s outhern Chile, I charact erized the composition and biomass of epiphytes associated with two common emergent tree species; both Eucryphia cordifolia (Eucryphiaceae) and Aextoxicon punctatum (Aextoxicaceae) are regional endemics; I determined the effect of epiphytes on canopy i nvertebrate community st ructure on experimental trees with either intact (contro l) or removed epiphyte assemblage s; and I assessed the effect of epiphytes on bird communities by comparing bird visits between experi mental trees (with and without epiphytes), and bird abunda nce in forest plots with natu rally varying ep iphyte loads. Trees supported approx. 140 kg (dry weight) of epiphytic material; 30% in leaves and 70% in arboreal soils and roots. Epiphytes were dominated by Fascicularia bicolor (Bromeliaceae), Hymenophyllum ferns, and by the hemi-epiphyte Raukaua laetevirens (Araliaceae). Forest
14 canopy was inhabited by a large di versity of insects, arachnids, annelids and mollusks; notable patterns included a general lack of ants and a preponderance of detritivores. Trees with epiphytes supported 20% more invertebrate species in the tr ee crown and twice the i nvertebrate biomass of similar trees without epiphytes. Fina lly, large trees with epiphytes a ttracted more bird visits than similar trees without epiphytes and forest plots with grea ter epiphytic biomass supported significantly greater bi rd abundance; bird species ric hness was not affected. My findings identify large old trees as a critical structural dimension of mature forest that defines forest canopy communities as unique assemb lages that contribute significant diversity and biomass to south-temperate rainforest. Cont inuing research also re veals that ecological functions (such as water and nutrient retention, pollination, seed dispersal, predation) of canopy communities supported by large trees are both un ique and significant within the forest ecosystem. Implications for sustainable fo rest management practices are discussed.
15 CHAPTER 1 LINKAGES AMONG STRUCTURE, CO MPOSITION AND FUNCTIONS OF BIODIVERSITY IN THE FOREST CANOPY Theoretical Background Biodiversity is defined as the diversity of liv ing forms below and above the species level, including all hierarchical levels of organiza tion from genes to ecosystems (Groom et al. 2006). Noss (1990; using Franklin et al 1981), proposed a framework to characterize biodiversity based on three main attributes: Composition, structure and functions, interacting with each other at different levels of system or ganization. Composition refers to the identity of units of biodiversity, such as the dive rsity of genes or a lleles within organisms, or the taxonomic composition of communities. Structure refers to how these units are physically organized, creating physical structural complexity. For inst ance, the age structure of a population, or the physical structures created by l ogs or snags in forests. La stly, functions refer to the manifestations of biodiversity in the biological interactions an d biogeochemical processes that take place in ecological systems, such as ge ne expression, species trophic interactions (predators, herbivores), nitrogen fixation and nutrient recycling. Noss ( 1990) proposed that these attributes interact at different scales; for inst ance, allelic diversity (composition) can affect heterozygozy (structure) and gene tic flux (function) with in and between populations (Fig. 1-1), while species diversity (composition) is linked to habita t structure and ecosystem process (Fig. 11). The conceptual framework proposed by Noss (1990, 2006) offered indicators for monitoring the health condition of ecosystems, such as the effects of reduced structural diversity on taxonomic richness (Newton 1994, Roberge et al. 2008), and by exte nsion the value of structural elements as indicators of biological integrity (Torras and Saura 2008, Smith et al. 2008). Despite the high number of p ublications citing Noss three types of attributes as indicators
16 of ecosystem health (over 600 citations by ISI Web of Science), most work in biodiversity focus on linkages between structur e and taxonomic composition, pa rticularly emphasizing one direction of the effect (i.e., effects of structure on species composition, and not vice versa). However, evidence of the multiple and reciproc al interactions among composition, structure and function remains as unexplored research ar ea in ecology (Badano and Marquet 2008). In this dissertation I will use the approach developed by Noss (1990) to characterize the attributes of biodiversity and thei r interactions in the canopy of Ch ilean temperate rain forests, a largely undescribed component of the forest envir onment. I will further an alyze the usefulness of Noss (1990) conceptual framework for understanding the functions of biodive rsity in the forest canopy environment. Clarifying what are the main bi odiversity attributes and their interactions in this system should build a stronger framework for ecological synthe sis and for ecosystem management and conservation (Pickett et al. 2007) The larger goal of my dissertation is to contribute to such efforts. Composition, Structure and F unctions in Forest Canopies The Forest Canopy: An Ecological Frontier One of the less explored terrestrial ecosystems in the 21st century is the forest canopy (Nadkarni et al. 2004a). Canopies ar e defined as the upper level in th e vertical profile of a forest, not usually accessible from the ground (Nadkarn i et al. 2004a). Among th e key functions that forest canopies provide to global ecosystems are their function as the main gas exchange interface, a surface where photosynthesis occurs, a structure for hydrological storage and release (Ozanne et al. 2003, Shaeffer et al. 2008), and their role in nutrient capture and nitrogen fixation (Fonte and Schowalter 2004). The advance of forest canopy ecology has been hampered in the past because the limited access and mobility of humans in this vertical layer. In recent years, canopy balloons, platforms, walkways, cable ne tworks, canopy cranes, and tree climbing
17 technologies have allowed both intensive and exte nsive sampling schemes that are necessary for modern ecological research (Mitch ell et al. 2002). Since the 1970s, there has been an explosive increase in canopy research worldwide (Moffett and Lowman 1995, Mitchell et al. 2002). Initial explorations of the forest canopy revealed that this vertical layer contai ned a high and largely undescribed, proportion of the forest biodiversity (Wilson 1992) Erwins (1982) classic study and subsequent work on canopy invertebrates in Panamanian and Amazonian forests, changed our assessments of global taxonomic diversity by various orders of magnitude, from ~ 3 to more than 30 million species. While the actual estimates are still under discussion (Erwin 2004), there is no question that the structural complexity a nd resource base present in the forest canopy must support an enormous richness of organisms. Despite the growing knowledge of taxonomi c composition, little is known about the functional linkages among structural complexit y, species diversity, and associated ecosystem processes in the forest canopy. So me intriguing examples have recently come to light, however. For example tank bromeliads that hold a substa ntial volume standing rain water in branches above the forest floor, support complex lifecycles of water-requiring plants, vertebrates, and invertebrates in the forest canopy (Richardson 2004) Field experiments have recently shown that forest canopy habitats experience significantly le ss herbivory because of avian predation on leafchewing insects (Van Bael et al. 2003, 2008). Howe ver, the structural complexity and (in many cases) the high taxonomic diversity and comple xity of forest canopies have hindered experimental studies of the linkages among species composition, structure and function, especially in the tropics (Cruz-Angn et al. 2008). Temperate Rainforest of Southern Chile Temperate forest of the Southern Hemisphere, pa rticularly coastal rain forests in Chile, still maintain extensive remnants of largely undistur bed and unpolluted forests, without atmospheric
18 deposition from industrial sources (Hedin et al. 1995, Perakis and Hedin 2002). Up until recently, little or no intensive human impact due logging was known (Willson and Armesto 1996), and clearing as well as pre-Hispanic settleme nts were restricted to specific areas along the coast and main river valleys. Ecologically, th ese rain forests differ greatly from northern hemisphere temperate rainforests; southern Chil ean forests are dominated by broad-leaved trees, and contain abundant vascular and non-vascular epiphytes an d vines (Armesto et al. 1996, Clement et al. 2001, Muoz et al. 2003). Relati vely species-poor communities maintain high levels of endemism, as a result of long-term isol ation within the continent of a Tertiary forest flora and fauna with close relativ es in Neotropical, Australian and New Zealand forests (Armesto et al. 1996, Villagrn and Armesto 2005). Today, a br oad hyper arid desert and oceanic barriers separate these disjunct biotic realms (Armesto et al. 1998). Thus, 25 % of the vascular plant genera are endemic to southern South America, while 30% are also common to Australian New Zealand forests, and 25 % are shared with Neot ropical forests across th e Andes (Villagrn and Hinojosa 1997). Around 20 to 25 forest bird spec ies inhabit these forests (Willson et al. 1994), from which five are restricted to the forest unde rstory (Reid et al. 2004), and the rest use large canopy trees and the entire vertic al profile of the forest (D az et al. 2005). Although Chilean temperate rainforests have already been much re duced in area in recent decades due to logging and fire, large remaining patches still represen t pre-industrial conditi ons, where interactions among organisms and ecosystem func tions can be the direct result of evolutionary processes. Moreover, the relative compositional simplicity of these rain forests facilitates the analysis of whole species assemblages, which has been a limiting factor in tropical forest research (Greenberg et al. 2008, Cruz-Angn et al. 2008).
19 Links Among Epiphytes, Invertebrates a nd Birds in the Rain Forest Canopy Epiphytes ( sensu lato including climbers) are conspic uous components of the forest canopy and the vertical profile of Chilean rainfo rests (Riveros and Ram rez 1978, Clement et al. 2001, Arroyo et al. 1996, Muoz et al. 2003). Prez et al. (2005) showed that coastal montane Fitzroya cuppressoides forests in southern Chile hold 8.2 tons / ha of epiphytic matter, including live and dead biomass. Epiphytic humus has si milar chemical and physical features to the organic horizon of the forest fl oor, reason for which is also calle d arboreal soil (Enloe et al. 2006). This combination of epiphytes and its associated arboreal soils may support a high diversity and abundance of invert ebrates, as reported for Neot ropical forests (Nadkarni and Longino 1990, Ellwood and Foster 2004). Epiphytes also provide nesting and foragi ng sites for birds (S illett 1994, Cruz-Angn and Greenberg 2005). In southern Chile, several bird species are more abundant in old-growth forests characterized by the presence of large trees, whic h are generally profusely loaded with epiphytes and vines (Daz et al. 2005). About 70% of Chile an forest bird species are classified as insectivores, or they complement their diet with invertebrates (Rozzi et al. 1996). Therefore, if canopy epiphytes support a high abunda nce of invertebrates, in addi tion to those associated with their host trees, epiphytes may also support grea ter numbers of insectivorous bird species, increasing their local abundance in forests. Finally, several studies in tropical and temperate rain forests suggest that insectivor ous birds protect trees from he rbivorous insects (Marquis and Whelan 1994, Van Bael et al. 2003, 2008, Borkhataria et al. 2006). Consequently, increased bird abundance due to epiphyte loads may also reduce insect damage to forest canopies (Murakami and Nakano 2002). An experimental st udy of Mazia et al. (2004) in Nothofagus forests of westernmost Argentina showed that foliar damage by insects increased within bird exclusions to almost twice the values observed in control branches accessible to birds. Another ongoing study
20 in Nothofagus pumilio forests in westernmost Argentina by Garibaldi et al. (2007) also found a positive effect of birds on trees by controlling insect herbivore populatio ns, in agreement with previous results. Argentinean Nothofagus forests have a shared floristic composition with Chilean forests and are part of the same Ecor egion, characterized by the same avian species assemblage. Therefore, I argue that in southern Chilean rain fore sts, birds that prey on insects may also play a salient function as controllers of herbivore populations. In fact, based on the observed avian consumption rates of caterpillars in the canopy of Chilean forest ecosystems by Gonzlez-Gmez et al. (2006), this cont rolling effect may be quite strong. My study assesses the effects of forest st ructure (tree size) on species composition (taxonomic richness and abundance) of epiphyt es, invertebrates and birds in the canopy. Specifically, I hypothesize that canopy structure pr ovided by large trees in forests supports a high diversity and biomass of epiphytes (Fig. 1-2). The structure and resources provided by epiphytes, in turn, support a divers e assemblage of invertebrates th at support bird species (Fig. 12). Each chapter of this dissertation addresses eac h of the links proposed in the theoretical model presented in Fig. 1-2. Chapter 2 presents a thor ough description of the epiphyte community of Chilean rain forest canopies, considering its species richness, composition and biomass associated with large canopy trees. Chapter 3 anal yzes the postulated eff ect of canopy epiphytes on invertebrate species richness a nd abundance. Chapter 4 analyzes th e direct and indirect effects of canopy epiphytes on forest bird assemblages, and discuss the possible effect of birds through insect consumption on foliar herbivory. Finally, Chapter 5 summarizes the information about, epiphytes, birds and insect herbivores in fore st canopies, and proposes a synthetic conceptual model considering the multiple linkages among stru cture, composition and possible functions in the canopy of southern Chilean temperate rain forests.
21 Figure 1-1. Three attributes of biodiversity. In this scheme, compositional, structural and functional attributes are orga nized hierarchically, nested at different scales, and all attributes are linked with each other.
22 Figure 1-2. Theoretical model w ith the hypothesis proposed in this dissertation. The model presents the linkages among different com ponents of forest canopy biodiversity in Chilean rainforests. In this model, larg e trees support heavy epiphyte loads (solid arrow), epiphyte loads suppor t invertebrates in the forest canopy, and finally invertebrate biomass support insect-eating fo rest birds. In this way, large canopy trees indirectly (dashed arrow) support avian species. H1, H2, and H3 are the hypotheses tested in this thesis. H1 te sts the link between large tree structures and epiphyte loads. H2 tests the link between ep iphyte loads and invertebra te species richness and biomass, and H3 tests the indirect effect of epiphyte loads on forest birds (dashed line), under the assumption that bird visits to trees are driven primarily by insect availability. These hypotheses ar e explicitly addressed in the different chapters of this dissertation.
23 CHAPTER 2 EPIPHYTE SPECIES RICHNE SS AND BIOMASS LOADS ON Eucryphia cordifolia (CUNONIACEAE) AND Aextoxicon punctatum (AEXTOXICACEAE), TWO CANOPY EMERGENT TREES IN CHILEA N TEMPERATE RAIN FORESTS Introduction Epiphytes are defined as plants that use other plants for mechanical support, but do not absorb nutrients from their host (Benzing 1995). Epiphyte species are grouped into five main groups, accidental (those that only occasionally grow epiphytically), facultative (can grow epiphytically and on the ground), hemi-epiphytic (rooting in the ground for some portion of its life cycle), and holoepiphy tic (true epiphytes; Benzing 1995); para sitic species such as mistletoes are not usually included as epi phytes (Benzing 2004). Examples of common epiphytes in tropical and temperate forests are bromeliads, orchids, and mosses, and lichens, many of which vary in their distribution within tree cr owns and trunks according to hu midity, sunlight exposure, and interactions with other sp ecies (Benzing 1990, McCune et al. 2000, Ellyson and Sillett 2003, Benzing 2004, Sillett and Antoine 2004, W illiams and Sillett 2007, Cardels 2007). Epiphytes contribute with the 8 to 10% of all vascular pl ants species at global scale (Benzing 1995, 2004), and contribute up to 25-50% of all plant species in tropical and temperate rainforests (Gentry and Dodson 1987, Nieder et al. 2001). In tropical and temperate forests, epiphytes accumulate a large amount of dead organic matter in tree crowns, usually termed arboreal soils (Enloe et al. 2006 ). Epiphytes and the arboreal soil together comprise a dense layer that cover the bark of la rge trees (hereafter epiphytic bi omass). Total epiphytic biomass can represent an important portion of the bioma ss present in forest canopies. For instance, in Costa Rican cloud forests epiphytic biomass reaches up to 33.1 t ha-1, representing 6.3% of total forest biomass, but 81.3 % of the canopy biomas s, excluding wood biomass (Nadkarni et al. 2004b). In general, epiphytic biom ass (live epiphyte biomass plus arboreal soil) ranges between
24 0.3 Mg/ha to 44 Mg/ha in tropical and temperate forests of the world (Tanner 1980, Hofstede et al. 1993). Most of the diversity and biomass of epiphytes in forests is associated with the structure provided by older, larger emergent canopy trees (Franklin et al. 1981, McCune 1993, Nadkarni et al. 2004b, Johannson et al. 2007). Epiphyte biomass is typically higher in primary forests dominated by large canopy trees th an in secondary forests; for instance, a Costa Rican primary forest supported 33 Mg/ ha of epiphytic biom ass, relative to only 0.2 Mg/ ha in a nearby secondary forests (Nadkarni et al. 2004b). In additi on to their direct contri bution to the richness of plant species in forests, epiphytes support other diverse groups of organisms such as invertebrates and vertebrate s (Nadkarni and Matelson 1989, S illett 1994, Ellwood and Foster 2004, Chapters 3 and 4). Therefore, the stru cture provided by large trees support another structure: heavy epiphyte loads th at could represent important re servoirs of biodiversity of a large variety of taxa in forest ecosystems (Berg et al. 1994). Information about epiphyte diversity, distribut ion, and biomass is still scarce in several regions of the world where epiphytes are comm on (Zotz 2005). Remarkable among these regions are the South American temperat e rain forests (Sillett and An tonie 2004), where vascular and non-vascular epiphytes are outstanding components of forest biodiversity because of their high biomass and species richness (Riveros and Ra mrez 1978, Muoz et al. 2003, Prez et al. 2005, Zotz 2005). These rain forests are dominated by broad-leaved evergreen trees comprising a multilayered canopy layer, with scattered emergent trees lushly covered by epiphytes and vines; in their overall physiognomy, they resemble tropi cal rain forests (Armesto et al. 1996, Willson and Armesto 1996, Zotz 2005). The origin of South American temperate rain forests goes back to Gondwanaland during the Tertia ry Period and before the main Andean uplift, as shown by the
25 large number of taxa with Neotropical, Australasian, and New Zealand affinities. South American temperate rainforests have been isolat ed from other similar forests for more than 1.5 million years, consequently having a high proporti on of endemism across plant and animal taxa and at species, genus, and family levels (A rmesto el at. 1996; Vill agrn and Hinojosa 1997). Few studies have analyzed ep iphyte species richness, distri bution, and biomass in South American temperate rain forests. For logistic reasons, most studies have described forest epiphyte assemblages only in the lower layers of the vertical profile, 1-4 m from the ground level (Riveros and Ramrez 1978, Muoz et al. 2003, Prez et al. 2005). Th e first study accessing epiphytes in the tree crown was conducted by Clement et al. (2001), who described the composition of vascular epiphyte communities grow ing on the upper branches of the large, longlived conifer Fitzroya cupressoides (common name, Alerce), a tree that grows primarily in montane forests and a few valleys in southern Chile. They found th at one old Alerce trees held 60% of the plant species present in this type of forest, and accordingly old Alerce trees may be important reservoirs of the overall forest biodive rsity. Their results also suggest that large canopy trees, in general, may play similar roles rega rding biodiversity throughou t the range of South American temperate rain forests. However, desp ite the importance of ep iphytes, species richness and distribution of epiphytes in relation to most rain forest tree species in southern South America remain largely unknown. In particular lowland Valdivian evergreen rainforests (distributed between 39 to 41 S primarily west of the Andes) represent the biologically richest of all South American temperate ra inforests in terms of plant and animal species and the level of endemism (Villagrn and Hinojosa 1997). Its canopy biodiversity is both underexplored and highly endangered from rapid declin e of old-growth forest cover and conversion to pastures and
26 forestry plantations of exotic tree specie s (Armesto et al. 1998, Smith-Ramrez 2004, SmithRamrez et al. 2005). In this chapter, I characterize the structural features (shape and biomass) of two evergreen tree species, which are the major components of the upper canopy of Valdivian temperate rainforests (Villagrn 1991, Armesto et al. 1996, Gutirrez et al. 2008a), Eucryphia cordifolia (in the Gondwanan family Cunoniaceae) and Aextoxicon punctatum (in the endemic family Aextoxicaceae). I examine the diversity and distri bution of vascular epip hytes growing in the entire vertical profile, from the base to the upper crown, of th ese tree species, and I assess the epiphytic biomass associated large canopy indivi duals of these trees. Finally, I discuss the potential effects of epiphyte loads on their host trees and their effects on ecosystem functions in southern temperate rainforests. This is a pparently the first study of epiphyte richness, distribution, and biomass in the upper canopy layer of lowland Valdivian rain forests in Chile. Methods Study Area The study was conducted in selected trees with in two large forest tracts old-growth temperate rain forests in the lowlands of north ern Chilo Island. These forests are composed by a mosaic of Valdivian and Nord-Patagonian forest types, dominated by evergreen, broad-leaved trees (Aravena et al. 2002; Fig. 1-1). The first forest tract is located in Guabn peninsula, at Punta Huechucuicui (41 47 S; 54 00 W) wh ere Valdivian undisturbed old-growth forest covers an area of 300 ha, including canopy trees older than 350 years (Gutirrez et al. 2008a) within a large track of 1000 ha of continuous forest with different degrees of human disturbance (Fig. 1-1). The second forest tract is in Chilo Na tional Park, at the edge of Lake Cucao (42 37 S; 74 04 W), which is a continuous forested ar ea of >40,000 ha. I chose la rge forest tracts to avoid any effect of forest fragmentation, such as changes in temperature and humidity (Saunders
27 et al. 1991), or changes in species compos ition (Willson et al. 1994, Barbosa and Marquet 2002). Organic soils reach over 1 m deep, but underlying mineral soils are thin, mostly close to bedrock of Miocene sedimentary rocks of marine origin in Guabn, and Paleozoic metamorphic rocks in Cucao (Mardones 2005). These rain forests represen t remnants of once co ntinuous coastal-range temperate forests, characterized by a heterogene ous forest canopy and scattered large emergent trees of Eucryphia cordifolia and Aextoxicon punctatum with a subcanopy dominated by species in family Myrtaceae (cf. Armesto et al. 1996). Gu abn forest stand had no signs of disturbance during at least the past 450 years (Gutirrez et al. 2008a). This area of Chilo Island was colonized by people in the early 20th century, therefore having little or no previous human impact by indigenous or European people and hence repr esenting a fairly pristine ecosystem (Gutirrez et al. 2008a). In contrast, Cucao is near an old indigenous settlement from perhaps > 500 years ago (Weisner 2003). Despite the lack of signs of human disturbances in the forests studied, evidence of fires from >100 years ago (Esp inosa 1917) can be found in the surroundings. Study Design To assess epiphyte richness and biomass I sel ected two large, canopy emergent trees of Eucryphia cordifolia in the forest of Guabn, and one large individual of Aextoxicon punctatum in the forest of Cucao, Chilo National Park. Each tree was a little over 1 m of DBH (diameter at breast height), about the average size of trees of these two species in each study area. All trees were climbed using arborist tech niques, including single and doubl e rope techniques that allowed access to the majority of the tree branches follo wing established protocols of the Tree Climber Coalition (www.treeclimbingusa.com ). Field work in the forest of Guabn was conducted between August and December of 2005, late winter and spring in the southern hemisphere; field work in Cucao was conducted between April and August 2006, during the southern hemisphere fall and winter.
28 Tree Structure Tree structure was mapped by measuring the le ngth, diameter, cardina l orientation (North, South, East, West), and height along the vertical profile of a ll branches and the trunk following the protocol described by Van Pe lt et al. (2004). With the diamet er and length data I estimated the area and volume of each branch, and in conse quence, the area and volume of the whole tree assuming a cylindrical shape of each limb measur ed. I estimated the total dry biomass of each tree using published data of wood density of 0.601 kg/lt for E. cordifolia and 0.60 kg/lt for A. punctatum (USDA Forest Service, 2008). I assessed foliage biomass of all trees by counting all branches >2-m long, and later collecting and weighi ng the leaves of three branches per tree (Van Pelt et al. 2004). Leaf area was calculated measuring directly th e area of 200 randomly selected leaves, and weighing them to obtain an estimate of total area per gram of leaves. I estimated total leaf area per tree by multiplying l eaf area by the total estimated l eaf biomass. In summary, with these measurements I calculated total leaf area, total foliag e biomass, total trunk area and volume, and total biomass of each tree. Epiphyte Species Richness For each tree, I conducted a rapid survey of ep iphyte species richness by taking samples of all recognizable epiphyte species every 2 m along the vertical prof ile, including all branches and the trunk of E. cordifolia and A. punctatum trees. For every epiphyte sampled I recorded the branch number and its height above the ground, a nd I brought them still fresh into the laboratory. Plant specimens were classified using the re ference collections of the Herbarium of the Universidad de Concepcin, Chile. Scientific na mes followed Marticorena and Quezada (1985). Epiphyte Biomass I assessed canopy epiphyte biomass at the tr ee level, and at the species level by two different protocols: by removing and weighing epiphytes from indi vidual trees, and by visually
29 assessing the volume of epiphytes in 60 trees of different sizes. At the tree level, the epiphytic material from each sampled tree was removed by hand using small axes, garden saws, and knifes. Epiphytes were removed as completely as possible, afte r the rapid survey of epiphyte species richness. The epiphytic material remove d was placed in large pl astic bags of 40 liters lowered to the ground, and then weighed separa ted into two components live green tissues versus arboreal soil and roots. Samples of around 600 g fresh weight of epiphytic material were taken from each component per ba g, stored in closed plastic bags then dried at 80 C over three days in a drying oven, and weighed again to de termine its water content. I estimated total epiphytic dry mass by multiplying the total epiphytic biomass of each component of each bag by 1 WC, where WC is the proporti on of water content of the respective component. When host trees were colonized by hemi-epiphytic trees, th e later were not removed but their biomass and foliage were measured following the protocol described by Van Pelt et al. (2004). Since published information on wood density of hemi -epiphyte was unavailable, I collected 5 wood samples of different branches, dried them at 80 C for three days, weighed them, and calculated wood volume by liquid displacement. Epiphytes we re removed from each tree during a twoweek period by three peopl e per tree. Epiphytes of E. cordifolia were removed in October and November 2005 during the austral spring, while epiphytes of A. punctatum were removed in August 2006 during the austral winter. I estimated epiphyte biomass at the species level, in 59 E. cordifolia trees. First, I selected 34 points separated by >100 m. In each point I traced a 50-m long transect starting from each point, running 25 m in opposite directions from the original point. Fo r all individuals of E. cordifolia present within an area 2-m wide on each si de of the transect lines, I measured the diameter at the breast height (D BH) and visually assessed the vol ume of epiphytes in terms of
30 the number of large bags that would be filled w ith epiphytic material, that according to my previous epiphyte removal experience, these ba gs can hold up to 15 kg of fresh weight of epiphytes. Focal E. cordifolia trees are not very high, lower than 30 m in the study site. For this species, it is possible to observe most branches from the ground, and the hemi-epiphytic tree is easily recognizable from a ground perspective, which in addition to my removal experience facilitates the assessment of ep iphytes in this study site. Ecological Functions Conducted by Epiphytes I analyzed the possible ecological functions of epiphytes based on data from previous studies in Guabn forests conducted by my research assistants using E. cordifolia trees nearby those used in this study. I also use published studi es conducted in other fore st in Chile (Prez et al. 2005, Del Val et al. 2006). In Guabn forest, Tejo et al. (in preparation) characterized the water content, nitrogen minera lization, and nitrification rate s in arboreal soils of two E. cordifolia trees. These authors took one sample of arbor eal soil per height, at three heights (8, 12, and 16 m) and three samples of the ground soil nearby. One part of each soil sampled was collected, while the other part was leaved in sit u, in an open bag and collected one month later. Samples were stored at low temperature (< 5 C) and analyzed under laboratory conditions to assess dry mass, density, and ammonia and nitr ate contents. Samples were taken each season, completing one year of survey. Finally, Ca rmona et al. (unpublished data) conducted a preliminary analysis of nitrogen fixation in arboreal soil and canopy lichens. These authors located 11 chambers of one-liter capacity at four heights (0, 9, 12 and 17 m) in one large E. cordifolia tree. In 8 chambers they put between 11 a nd 16 g (dry mass) of soil, four chambers with ground soil and four with arboreal soil. The ra te of nitrogen fixation was assessed using the acetylene reduction test following the protocols of Hardy et al. (1968). Three chambers of each soil class were injected with acetylene obtaining a concentration of 10% v/v, and one was left as
31 a control. The chambers with arboreal soil we re located at 9, 12 and 17 m, with the control chamber also at 17 m. The chambers with forest soil were located at ground level (0 m). In the other three chambers they located be tween 1 and 3 g (dry mass) of living Pseudocyphellaria lichens collected from the canopy. Two chambers were filled with acetylene obtaining a concentration of 10% v/v and the other was left as a control. These chambers were located at 17 m in the epiphytes, in the same places where lichens were collected. At three times (day 0, day 1 and day 2), samples of 10 ml of air where co llected from each chamber, to calculate the concentration of acetylene, and calculate the rate of acetylene reduction expressed as nmol acetylene /day/ dry mass (g). This assay was co nducted during three days in December 2005, at the end of the austral spring. Data of acetylen e reduction between heights was compared with paired t-tests. Results Structural Features of Eucryphia cordifolia and Aextoxicon punctatum Trees The two large E. cordifolia trees sampled in Guabn were 25 and 30 m tall, respectively, with the first crown branches 6 and 15 m a bove the ground, respectively. The general shape of E. cordifolia trees was characterized by continuous branch ing up to the uppermost leaves, with an open, V-shaped crown (Fig 2-2). The foliage was present only on the extern al portions of the branches, leaving ample space without leaves in th e interior of the crown (Fig. 2-2). The crown of this tree species is massive, extending 21 m in diameter with branches radiating in all directions (Fig. 2-2). The heav iest component of tree bioma ss was the woody material, which reached 27 x 103 and 42 x 103 kg of dry matter per tree, resp ectively. Leaf biomass was three orders of magnitude lower, representing < 0.1 % of total tree biomass (Table 2-1). In the open space in the interior of the cr own, both sampled individuals of E. cordifolia hosted a common hemi-epiphytic tree in Valdivian rain forests, Raukaua laetevirens (Araliaceae). These hemi-
32 epiphytes were profusely branched having dens e foliage that did not extend over the top of E. cordifolia crown. The wood density of R. laetevirens was 0.5 kg/ liter; and the dry biomass of each R. laetevirens tree measured was 2.6 x 103 kg and 1.0 x 103 kg, respectively, where 9 kg corresponded to the foliage, representing < 0.2 % of the total dry weight (Fig. 2-2). The single A. punctatum tree sampled was similar in gene ral shape to canopy individuals of E. cordifolia and reached 25 m in he ight. The crown area of A. punctatum was smaller, but presented twice or more foliage biomass (99 kg dry mass) than E. cordifolia trees (Table 2-1). The crown of A. punctatum was denser and without inner foliage-free spaces, yielding a multilayered vertical foliage, thus differing in structure from E. cordifolia No hemi-epiphytic trees were observed growing in the A. punctatum trees in this forest (Fig. 2-3). Species Richness and Distribution of Epiphytes on E. cordifolia and A. punctatum Overall, I found 22 species of vascular epiphy tes and 22 genera of non-vascular epiphytes in the three trees sampled (Table 2-2). In few cases it was possible to cl assify the non-vascular epiphytes to the species level (Table 2-2). Vascular epiphyte s included 15 species and four families of Pteridophyta, with 12 species in the family Hymenophyllaceae (filmy ferns). The Anthophyta was represented by only seven species in six families (Table 2-2). Non-vascular species included mostly mosses, liverworts, and lichens (Table 2-2). Both canopy tree species ( A. punctatum and E. cordifolia ) presented relatively similar species composition of vascular epiphytes but they differed in the species composition of nonvascular epiphytes (Table 2-2). In fact, the single A. punctatum tree hosted more species of bryophytes than the two E. cordifolia trees combined, and several species found in Aextoxicon were not detected on E. cordifolia (Table 2-2). On the other hand, E. cordifolia trees were the exclusive host of the hemi-epiphyte R. laetevirens which represents a conspicuous element of Valdivian rain forests because of its large size. The presences of a distinct group of non-vascular
33 epiphytes in A. punctatum and the presence of the hemi-epiphyte R. laetevirens tree in E. cordifolia represent the main differen ces in epiphytic assemblages associated with these two dominant tree species. Most species of both vascular and non-vascular plants exhib ited particular distributions on the tree surfaces and along the vertical pr ofile (Table 2-3). The ground-rooted vines Luzuriaga radicans and Mitraria coccinea were often found much closer to the ground than completely holoepiphytic vine species such as Sarmienta repens Ferns including Grammitis magellanica Hymenophyllum dentatum and Polypodium feullei and lichens such as Cladonia sp. or Bunodophoron sp. were present mostly in the upper br anches of the sampled trees. Only two species of bryophytes were found to o ccur along the whole vertical profile: Bazzania sp. and Frullania sp. (Table 2-3). Epiphyte di stributions along the tree vert ical profile and the crown area showed three main groups: ep iphytes present along a ll the vertical profile epiphytes present only in the upper branches and epiphytes restricted to the ba sal branches and trunks (Table 2-3). Around 34 % of the plant species were associated with the crowns, 18% were present primarily in the first 3 m above the ground, while the rema ining 48% were present all over the vertical profile (Table 2-3). Excluding thos e species present only near the base of the tree, 82% of the epiphyte species collected occurred > 3 m above the ground. The upper branches were undersampled due to their inaccessibility, but field observations indicated th at they were densely covered mainly by lichens, probably Pseudocyphellaria sp. Epiphytic Biomass At the tree level, the three trees sampled hos ted large epiphytic load s. Individual trees of E. cordifolia hosted 1.6 x 103 kg dry mass of the hemi-epiphytic tree R. laetevirens plus an additional 133-143 kg of dry mass of other vascul ar and non-vascular epi phytes in the trunk and branches, with about 30% corresponding to gr een tissues and 70 % corresponding to arboreal
34 soil (Table 2-4). One tree of A. punctatum supported 142 kg dry mass of epiphytes, with the same proportions of leaves and arboreal soil as E. cordifolia Epiphyte biomass was differently distributed alo ng the vertical profile of the trees of the two species studied. In E. cordifolia most of the biomass was concentrated in the intermediate portion of the trunk and near the base of the crown branches; while in A. punctatum epiphytic biomass was more evenly distributed along the vert ical profile, with large clumps of epiphytes on the terminal branches at the top of the tree. This situation creates different aspects or physiognomy of these two tree species in the forest canopy; E. cordifolia appears to have few epiphytes (with the exception of R. laetevirens ) and little foliage volume in the treetops, while A. punctatum has substantially denser foliage over the extent of the crown and frequently denser clumps of epiphytes on the branch tips. Among the recorded epiphytes, two vascular spec ies were particularly important in terms of biomass: these were the hemi-epiphytic tree R. laetevirens and the bromeliad Fascicularia bicolor R. laetevirens grows from the mid portion of the vertical profile of large E. cordifolia trees, with its roots descending to the ground similar to strangler fi gs in tropical forests (Zotz 2005, Gutirrez et al. 2008a). In the Guabn forest, 65% of all E. cordifolia trees were > 1 m DBH and all supported at lest one hemiepiphytic R. laetevirens (Gutirrez et al 2008a, M. P. Pea, unpublished data). R. laetevirens has dense foliage and its canopy roots are heavily covered by other epiphytes holding a thick layer of arboreal soil additional to the host tree. Overall, R. laetevirens had 2.6 x 103 kg and 1.0 x 103 kg of dry mass of woody material, and 9.5 kg of leaves (dry mass). This volume of foliage equals 18% and 29% of the total leaf mass present in the cr own of their host E. cordifolia trees (Tables 2-1 and 2-4). On the other hand, F. bicolor is distributed along th e whole vertical profile of their host tree, but most of its biomass
35 (>80%) occurs at intermediate heights between 9 a nd 15 m of the forest vertical profile, where it forms large clumps up to 34 kg (fresh we ight). In addition, large clumps of F. bicolor accumulate up to 56% of the total arboreal soil, a nd up to 68% of the green tissues of epiphytes recorded in these two tree sp ecies, excluding the hemi-epiphyte R. laetevirens In general, epiphytes represente d <1% of total tree biomass for E. cordifolia and A. punctatum but excluding the supporting woody struct ures (trunks), epiphytes represent a significant percentage of the photos ynthetic tissues present on fore st canopies. Foliage associated with forest epiphytes accumulated 40 kg in A. punctatum, an addition of 41% to the mass of foliage produced by the host tree (9 9 kg, Table 2-1). Crown foliage of E. cordifolia reached 51 kg in tree 1 and 33 kg in tree 2, while leaves from epiphytes represented 50 kg in tree 1 and 48 kg in tree two. Therefore, foliage from ep iphytes represent betw een 100% and 145% of additional green mass to the crown of E. cordifolia Finally, visual assessments of epiphytes in 59 trees of E. cordifolia of different sizes in Guabn forest showed that epiphyte biomass incr eased exponentially with tree size (Fig. 2-4). Smaller trees hold less epiphytes, but trees larg er than 1 m DBH showed a pronounced increase in epiphyte biomass (Fig. 2-4). This high abund ance of epiphytes seems to coincide with the colonization by the hemi-epiphyte R. laetevirens Trees <80 cm DBH did not hold individuals of R. laetevirens but all trees >1.2 m DBH hold a R. laetevirens tree (Fig. 2-4). Epiphyte Functions The study of Tejo et al. (manuscript in prepara tion) showed that arbo real soil share several characteristics with forest soil (Table 2-5). Both arboreal and forest soil were composed by litter of similar origin and quality, dominated by dead leaves of E. cordifolia Arboreal soils are organic (i.e, they lack mineral components such as sand or clay), but are similar to soil from the forest floor in pH, concentrations of ammonia and nitrate, and in nitrification rates (Table 2-5).
36 The density of arboreal soil was lower, retaini ng more water than forest soils (Table 2-5). Finally, the results from Carmona et al. (unpublished da ta) showed that acet ylene reduction in the forest floor and in the arboreal soil were similar, and indistinguishable from control chambers. In contrast, one of the Pseudocyphellaria lichens showed a hi gh rate of acetylene reduction, suggesting that these lichens may be im portant nitrogen suppliers in the forest canopy (Table 2-5). Discussion Comparative Diversity of Epiphytes in Southern Chile Temperate Rainforests Despite limited information on species rich ness and abundance of epiphytes in South American temperate rain forests, we determined that just three large trees held 33 % of the species of flowering epiphytes described for Va ldivian rainforest by Smith-Ramrez et al. (2005), and 50% of the Hymenophyllaceae ferns descri bed for Chile (25 species, Marticorena and Quezada 1985). Hymenophyllaceae is a largely epiphyt ic fern family with 17 species described for Valdivian rainforests (Smith-Ramrez et al. 2005) and 20 species described for Chilo Island forests (Villagrn and Barrera 2002). In particular, several Hymenophyllum ferns occur primarily in the intermediate and upper portions of canopy trees, and when they occur on the forest floor, they are always associated with fallen logs and limbs (Villagrn and Barrera 2002). On logs on the ground and in the first meter of the vertical pr ofile of Guabn forests, Daz et al. (in prep.) found 14 species of Hymenophyllum species, while in the present study I recorded 10 species of filmy ferns along the entire vertical profile of E. cordifolia trees. In fact, many species of Hymenophyllum were present on the forest floor because of the continuous flow of limbs covered by these ferns that fall from the canopy. In s econdary forests nearby our study site in Guabn, species richness of f ilmy ferns decreased to only 5 sp ecies, and its relative abundance (percentage of soil covered by these ferns) decl ined by 90% compared to old-growth forests
37 (Daz et al. in preparation). In conclusion, large canopy trees support in their crowns a significant proportion of the richness of Hymenophyllum species, and through bran ch-fall, they maintain their presence in the gro und biota. Large trees could be divers ity reservoirs for filmy ferns in temperate rain forests and this pattern may also ho ld true for other forest species of animals and plants found in the upper cr own (Nieder et al. 2001, Daz et al. in preparation). Previous studies of Muoz et al. (2003) anal yzed the richness and relative abundance of vascular epiphytes on 499 trees of 12 species, w ith DBHs of 5 to >25 cm in North-Patagonian rain forests of Chilo island. They sampled epip hytes up to 2.5 m on the tree trunks, recording a total of 20 vascular plan t epiphytes, including eight Hymenophyllum species (Table 2-6). Clement et al. (2001) conducted a complete survey of epiphytes in nine giant F. cupressoides trees in Andean montane forest of southern Chile. They recorded eight species of Hymenophyllum ferns, and a total of 18 vascular plant species in the canopy (Table 2-6). My results for only three large trees > 1 m DBH showed a total of 22 vascular plant epiphytes, higher than in lerce forests, including most of the fern species described by these authors, but lacking several vine species. These vines are frequent in North Patagonian forests, but generally absent in Valdivian forest types (Armesto et al. 1996). The forest type where Muoz et al. (2003) and Clement et al. (2001) conducted their studies is characterized by poor soil drainage and colder climate, with a canopy dominated by different speci es of trees, where bromeliads and other types of climbing vines are less frequent or absent in comparison with our study sites. Then, large trees in Valdivian temperate rain forests may, overall, hold more epiphyte species than previously recorded in other forests of this region. Despite the lack of data, these results suggest that large old trees harbor much of the epiphytic species richness of th e region. From this point of view, old forest trees could be
38 considered as significant reservoi rs of forest biodiversity, incl uding many species that are absent from secondary forests (Barthlott et al. 2001). Further studies compari ng epiphyte distribution according to tree ages and in different successional forests should clarify the role of old trees as reservoirs of biodiversity in s outhern temperate rainforests. Comparative Diversity of Epiphytes with Other Tropical and Temperate Rainforests The epiphytes present in the th ree individual trees st udied were mostly vascular species, whereas epiphytes in northern temperate rainforests are mostly non-vascular (Table 2-6, Zotz 2005). My results showed that the dominant epi phytic plants in terms of biomass are also vascular species (Table 2-7). However, in th e trees analyzed I found a modest numbers of epiphyte species compared with tr opical or even other temperate ra inforest of the world (Table 26). For instance, in tropical forest of Costa Rica, Cardels (2007) recorded 496 vascular epiphyte species in 53 individual host trees of 11 genera, while in th e Pacific North West of North America, Pike et al. (1975) and McCune et al. (2000) recorded around 100 species, almost exclusively non-vascular epiphyte s (Table 2-6). The number of va scular plant species recorded in the three trees sampled in Chilean rain forests is also lower than in other southern hemisphere temperate rainforests; Dickinson et al. (1993) f ound a few more species of vascular (28 species) and non vascular (26 speci es) in a single large Dacrycarpus dacrydioides of New Zealand forest than what I found in two E. cordifolia and one A. punctatum trees (Table 2-6). In agreement with my results, Clement et al. (2001) and Muoz et al. (2003) recorded simi lar species richness in different tree species and forests of southern Chile, the first in th e canopy of nine large F. cupressoides trees and the second on the trunks of 499 tr ees of all sizes and 12 species (Table 26). Despite the lower number of epiphyte species, South American temperate rain forests showed a high proportion of endemic taxa, where most va scular epiphytes are endemic to this biome (Arroyo et al. 1996, Villagrn and Hinojosa 1997). Th is epiphytic community is very unique, and
39 also shares more similarities with Neotropical and Australian-N ew Zealand temperate rainforests than with Northern Hemisphere rainforests. Floristic Relationships between Chilean Epi phytes and those from the Neotropics and Australia-New Zealand The dominance of vascular epiphytes and the fl oristic similarities with Neotropical and Australian-New Zealand forests be a consequen ce of to the common Tert iary origin of these ecosystems (Villagrn and Hinojosa 1997). For instance, the tree genus Eucryphia includes only five species, two in Chile and three in Austra lia and Tasmania (Arroyo et al. 1996), while the family Aextoxicaceae represents a mono-typic endemic family with unclear phylogenetic relationships to other living fam ilies of plants (Smith-Ramrez et al. 2005). The epiphyte families Gesneraceae, Araliaceae, Bromeliaceae, and Hyme nophyllaceae are all shared with Neotropical forests, and all except Bromeliaceae, have presen tly vicariant distributions with relatives in Australian and New Zealand forests (Hosft ede et al. 2001, Zotz 2005, Table 2-4). Another important feature of epiphytes in Chilean forests is their functional simila rities to other southern temperate and tropical forests. For instance most species of, vines, shrubs, trees and even vascular epiphytes of Chilean temperate rainfore sts presented fleshy fruits, which require animal vectors for seed dispersal (Armesto and Ro zzi 1989, Willson et al. 1996). The proportion of fleshy-fruited plants in Chilean temperate forest s is similar to other Ne otropical forests, and higher than many northern hemisphere temperate rainforests (Armesto and Rozzi 1989, Willson et al. 1996). The high proportion of animal-dispersed plant species contrasts with the relatively low diversity of frugivores, all of which use a wide variety of food resources such as invertebrates (Armesto and Rozzi 1989, Sabag et al. 1993, Rozzi et al. 1996). The principal seed vectors are fruit-eating bi rds, such as Fio-fio ( Elaenia albiceps Tyrannidae), Zorzal ( Turdus falcklandii Muscicapidae), and the endemi c marsupial Monito del Monte ( Dromiciops gliroides
40 Microbiotheriidae) (Sabag 1993, Ro zzi et al. 1996, Amico and Aizen 2000, Celis-Diez et al. in review). The importance of these species as seed dispersers could be di sproportionately high for epiphytes, because most epiphyte seeds must r each the upper branches of trees, moving against gravity. Limited dispersal can decrease the populations of epiphytic plants; for instance Rodriguez-Cabal et al. (2007) showed that habi tat fragmentation in temperate rainforests of westernmost Argentina have reduced the populations of Monito del Monte, affecting the seed dispersal of the mistletoe Tristerix corymbosus which has become completely absent from isolated forest fragments where the marsupial is lacking. Similarly, only one bird species is known to pollinate flowering epiphytes in Chilea n temperate rain forests: the hummingbird Picaflor chico ( Sephanoides sephaniodes Trochilidae; Smith-Ramrez et al. 1993, Chapter 4). This hummingbird is the main (or only) bird polli nator of over 20% of the vascular species of South American temperate rainforests, incl uding the dominant canopy epiphytes such as Sarmienta repens and Fascicularia bicolor Therefore, a wide diversity of plant species depend on this only hummingbird for its polli nation (Smith-Ramrez et al. 1993). The dependence of rain forest epiphytes on animals may be much stronger than the dependence of animals on epiphytes. Sabag (1993 ) reported the presence of seeds from a broad diversity of trees in the feces of bird species, but he did not record seeds from epiphytes. Armesto et al. (2001) also recorded high numbers of tree and shrub seeds in the seed rain driven by birds both to forest edges and tree-fall gaps. Th e high number of fruits from trees in the diet of birds suggests that forest trees may offer orde rs of magnitude more fruits than epiphytes. For the hummingbird this relationship could be mo re symmetrical; several epiphytes concentrate their flowering during fall and winter, offering n ectar during a season when other nectar sources are not available (Smith-Ramrez et al. 1993; Chap ter 4). Accordingly, knowledge of the factors
41 regulating epiphyte seed dispersal may be relevant for understanding their persistence in forests. However, for most epiphytes requiring animal seed dispersers, the identity of the seed vector remains unknown, and further studies are needed to define them a nd analyze both the effect of epiphytic fruit and nectar production on animals and the effects of animals on epiphyte seed dispersal and pollination. In summary, the epiphytic assemblages of Chilean temperate rain forests reflect the tropical and ancient orig in of its species composition, domi nance, and interactions between plants and animals. Composition and functions of epiphytes in temperate forests of Chile are very different from those of northern hemisphere temperate forests, and more similar to tropical and southern hemisphere temperate rainforests. Epiphyte Vertical Distributions The three groups of epiphytes along the vertical profile of the trees ca n be explained by a combination of different conditions of humidit y, temperature and shading, as occur in other forest ecosystems (Cardels and Chadzon 2005). These relationships are largely unknown for Chilean forest epiphytes (Salin as 2008), but my study suggest di fferential requirements for three dominant epiphytic species, the holoepiphytes Fascicularia bicolor and Sarmienta repens and the hemi-epiphytic tree Raukaua laetevirens The bromeliad F. bicolor is distributed along the whole vertical profile of canopy trees, but most of its biomass is concentrated at the base of the crown and intermediate portion of the tree. The bromeliad family present clear adaptations to hydrological stress (Benzing 2004); then the location of F. bicolor on large E. cordifolia trees can be related to the balance between access to light and lim ited humidity in the upper tree crown. In contrast, F. bicolor was much less abundant on the more shaded branches of A. punctatum (Tables 2-2 and 2-4).
42 S. repens is a vine that roots dir ectly over the tree bark and in arboreal soil on host trees (Salinas 2008), forming dense mats on the upper branches of E. cordifolia and A. punctatum This vine has succulent leaves that confer resistance to drought (Salinas 2008). According to field observations, most foliage of S. repens is generally pending down the limbs. I hypothesize S. repens presents a segregated use of the tr ee at two scales, first at tree scale, S. repens is located in the upper tree crown where it ha s access to higher solar radiat ion than other epiphytes, and secondly, plants are pendant from branches, thus protecting them from excessive sunlight, or avoiding interference with the other epiphytes established directly over the bark. Finally, the hemi-epiphytic tree R. laetevirens is a frequent and important element added to the crowns of large E. cordifolia trees. Like strangler figs and many other tropical hemiepiphytes, seeds of this species are dispersed by birds, and become established in the axils of canopy branches where arboreal soil accumulates. All E. cordifolia trees > 1 m DBH were colonized by R. laetevirens but trees <80 cm diameter in th e same area had none, at least none big enough to be visible from the ground or clea rly established (Fig. 24). I climbed 4 trees without this hemi-epiphyte species during the surveys, and despite many plants germinated at the end of the winter, they did not survived after summer. Therefore, I hypothesize that arboreal soil should accumulate beyond a certain mass threshold on branches of the host tree, in order to provide nutrients and wa ter long enough to allow R. laetevirens to become established and send its roots down to the soil (P utz and Holbrook 1986). The massive crown and open foliage of large E. cordifolia trees probably provides intermediate leve ls of solar radiation that may favor the development of R. laetevirens On 20 mature trees observed in the Guabn forest, R. laetevirens branches never grew above those of the host tree. Reasons for this could be related to R. laetevirens intermediate light requirements, physiol ogical limitations to bring water from the
43 soil to certain height (Koch et al. 2004), or mechanical lim itations on the support of heavy branches. My own field observations suggest that R. laetevirens was associated to it host tree E. cordifolia During two years of monthly visits to the forest of Guabn, and sampling over 800 trees bigger than 10 cm DBH, I saw only three R. laetevirens growing on trees of species other than E. cordifolia However, this observation must be quantified to understand the association between these species and the mechanism underlying it. Epiphyte Biomass Several studies indicate that epiphyte di versity and biomass increase with tree size (McCune 1993, Muoz et a. 2003, Johansson et al 2007). Epiphytes may need time to colonize trees after a dense layer of arbor eal soil can accumulate, and theref ore epiphytic biomass tends to be higher on older trees (Frank lin et al. 1981, Sillet and Baley 2003). In the forest studied, the large E. cordifolia trees were 300 to 400 years old (Guti rrez et al. 2008a); implying that the epiphytic assemblage and biomass described in this study has developed over that time span. However, biomass accumulation is not linear with tree age. At around 1 m DBH there seems to be a strong increase in epiphyte bi omass, presumably stimulated by the colonization of the hemiepiphyte tree R. laetevirens (Fig. 2-4). Once host trees are co lonized, the root system and the branches of R. laetevirens may increase the heterogeneity on tree canopy, offering sites for the colonization of other epiphytes (Pea et al. 2006 ). Therefore, two fact ors, tree size and the presence of R. laetevirens may account for the exponential accu mulation of epiphytic biomass on large E. cordifolia trees. A survey by Gutirrez et al. (2008a) in the Guabn forest found 72 individuals/ ha of E. cordifolia with a mean DBH of 1.0-2.4 m. Multiplying this density by the epiphytic biomass per tree documented in this st udy, indicates that epiphytic biomass supported by E. cordifolia is at least 10 Mg dry mass/ ha (excluding the hemi-epiphyte R. laetevirens ). This estimate for a single tree species in Valdivian coasta l forests is larger than previous estimates of
44 epiphytic biomass of 8.2 Mg/ ha of Prez et al. (2 005) for all species in montane alerce forests in southern Chile, and is also larger than estimates for all temperate rainforests of North America (6.8 Mg/ha, Grier and Nadkarni 1987, Table 27), dominated by bryophytes. At the individual level, Eucryphia had similar or greater epiphytic biomass than trees in montane cloud forests in Colombia (135-146 kg versus 110 kg; Table 2-7). C onsidering that most trees in the Valdivian temperate rain forest have a dense cover of epip hytes (personal observations), the total epiphytic biomass for the stand may double or triple this estimate for E. cordifolia reaching similar values to those estimated for tropical cl oud forests (Table 2-7). As indi cated by Hofstede et al. (2001), southern hemisphere temperate rainforests may support as much epiphytic biomass as tropical forests. This great accumulation of epiphytes in Chilean rainforests may be related to physiognomic and floristic similarities between tropi cal and southern temperate forests, such as the presence of hemi-epi phytic trees. The shape of E. cordifolia with its open massive crown branching from a single main trunk, resembles many tr opical trees, and is ve ry different from the conical shape of northern temp erate conifers. Epiphyte biomass in Chile is dominated by vascular plants, including the hemi-epiphytic tree R. laetevirens and high abundance of the bromeliad F. bicolor The presence of these large epiphyte species may offer additional substrate for more epiphytes. In fact, the arboreal soil associated with the large mats of F. bicolor represents between 27% and 57% of the arboreal soil on E. cordifolia trees, and therefore this species may be responsible for the accumulation of a major fraction of arboreal soil. Filmy ferns also capture litter including the d ead leaves of the host tree, as has been shown for other forests (Enloe et al. 2007). Further stud ies should analyze how the dynamics of colonization and growth of R. laetevirens and F. bicolor affect the overall accumulation of epiphyte biomass.
45 The smaller and more shaded crown of A. punctatum and its denser foliage may facilitate the capture and retenti on of water from rain and fog, cr eating more humid conditions that facilitate the development of epiphytes with hi gher moisture requirements such as bryophytes and filmy ferns (Benzing 1990), with less de velopment of the more drought tolerant F. bicolor and absence of R. laetevirens I hypothesize that the dense foliage of A. punctatum contributes to maintain humidity in its crown, because this tree may capture wa ter from fog that then moves along the tree as stemflow (del Val et al. 2006, Gutirrez et al. 2008b) fa voring the development of moisture-loving epiphytes. Epiphyte Functions Epiphytes can supply resources and regulate environmental conditions in the forest canopy, but can also have negative effects on th e host tree health. Epiphy tes can capture water and nutrients, thus increasing forest primary production, and pr ovide a food source for many invertebrates and verteb rates that live in the forest ca nopy (Weathers 2000, Lowman and Rinker 2004 for a review). Water interception by the fore st canopy is an important process preventing runoff, facilitating water storage, and facilitating nutrient transport in forest ecosystems (Klher et al. 2006). In 60-year old broad-leaved Chilo forests, Daz et al. (2007) found that 53% of the rainwater was intercepted by and/or evaporated from the tree canopy, while the remaining 47% reaches the ground via through fall and stemflow. This is a higher value of water interception compared to other temperate forest ecosystems (Daz et al. 2007). Our results showed that the water contained in the epiphytic la yer of our focal trees (excluding R. laetevirens ) is 60% to 70% of the total biomass, which means that epiphytes may retain over 300 liters of water in the crown of each tree. In old-growth forests, epiphytes play a strong role in the in terception and storage of rain water in the forest canopy (Klher et al. 2006).
46 In addition, epiphytes can intercept water and nutrients from passing clouds or fog (Holscher et al. 2004, Hietz et al. 2002). Coasta l forests in Chile are characterized by fog interception supplying up to 80 mm of water per month (del Val et al. 2006). This volume of water may be very important during the dry su mmers (del Val et al. 2006, Gutirrez et al. 2008b), but still remains undetermined what propor tion of the total interc eption in the forest canopy is due to epiphyte loads. Coastal fog repr esents a substantial inpu t of dissolved organic nitrogen from the ocean to the mainland forest s in southern Chile (Weathers et al. 2000). Oyarzn et al. (2004) indicated that for Andean Nothofagus forest cloud interception by the canopy is an important source of nitrogen. In addition, Benner et al. (2007) showed high nitrogen fixation by canopy lichens ( Pseudocyphellaria ) in Hawaiian forests, an input that is enhanced when additional phosphorous was supplied experi mentally. Lichens of the same genus are abundant in the canopy of Chilean coastal rain fo rests (Table 2-2), and preliminary data from Carmona et al. (unpublished data) from the canopy in Guabn forest suggest that their nitrogen fixation may be considerable (Table 2-5). Thes e results indicate the epiphytes can enhance capture of nutrients transported by fog from th e ocean, and canopy lichens could be an internal source of nitrogen in Chilean coastal forests. Tejo et al. (in preparation) showed that nitrogen mineralization and nitrification rate s in the arboreal soil of large E. cordifolia trees are similar to the forest floor soil (Table 2-5). Prez et al. (2005) showed simila r results for montane forest of F. cupressoides concluding that epiphytic soil is functi onally similar to the organic horizon of forest floor. These processes repres ent an overlooked source of nutrien ts to this forest ecosystem. Epiphytes also can play an important role in increasing forest productivity. Total foliage in the forest canopy was raised by 41% to 145% when epiphyte foliage is incl uded (Table 2-3). In E. cordifolia, in particular, epiphytes more than doubl e the leaf biomass of canopy trees; this
47 additional leaf biomass is greater than estimated in previous studies (I ngram and Nadkarni 1993). Although photosynthetic rates could be lower in epiphytes than in tree leaves (Benzing 2004), the substantial biomass of epiphytic photosynthe tic tissues can have a large effect on whole ecosystem productivity. Finally, ep iphytes also can supply nutrien ts directly to the tree via adventitious roots (Nadkarni et al. 1981, Oyarzn et al. 2004). Several tr ees of southern South America, including A. punctatum, produce adventitious roots from stems which allows them to obtain nutrients and water direct ly from the crown humus (Prez et al. 2003). In summary, my results are in agreement with othe r studies that propose an important role for epiphytes in forest ecosystems (Franklin et al. 1981, Nadkarni and Matelson 1992). Negative effects of epiphyte loads relate to the increased susceptibility of host trees to falling or losing large branches because of th e extra weight (Strong 1977); competition with the host tree for sunlight; and the poten tial effect of the production of a cylinder of fused roots by strangler hemi-epiphytes in preventing secondary growth of the host tree by root constriction (Putz and Holbrook 1986). Changes of the magnit ude and importance of these negative effects with different tree ages have received little attention. A review by Laman (2004) analyzes the effects of strangler figs in tr opical trees, indicating that se veral bottlenecks decrease their frequency in the canopy when they kill the host trees also emphasizing their role as providers of abundant fruits for forest fauna. In E. cordifolia arboreal soil and epi phytes weigh around 0.3 % to 0.7% of the weight of the host tree, and do not cast shade on the tree foliage. However, the hemi-epiphytic tree R. laetevirens could stress host trees since because its weight represents between 4% and 5.5% of the host tree biomass, a nd are usually located on one specific branch of the host tree. Such weight can represent a relevant pressure on physical st rength of the tree (Putz and Holbrook 1986). However, in Guabn forests a large number of E. cordifolia trees are still
48 alive after 300 years (Gutirrez et al. 2008a), and keep supporting a heavy epiphyte loads including hemi-epiphytic trees, s uggesting that the danger of fa lling and epiphyte shading may be low. Data of Pea et al (2006) indicate that epiphytic R. laetevirens could be older than 70 years, and several of them showed many dead br anches, suggesting that the hemi-epiphytic tree has a dynamic modular growth. I hypothesize that epiphytes in Guabn forests may not represent a grave danger to host trees, at least for se veral decades or maybe centuries. The demand of resources for secondary growth increases w ith tree age, causing an imbalance between photosynthesis and respiration in addition to prob lems derived from fungal infection, physical damage, and reduced photosynthesis (Van Pelt and Sillett 2008). At some time, the increased weight of epiphytes may result in broken limbs and increased risk of tree fall, affecting tree survival. However, the time span between epi phyte colonization and tr ee death is unknown, and epiphytes may enhance survival through positive e ffects by providing the host tree with nutrients and water, until some age where the increasing weight pressure enhance the risk of fall of a weakened old tree. Importance of Large Trees with Epiphytes Noss (1990), based on Frankin et al. (1981), proposed that biodiversity could be characterized by three main attributes: com position (such as taxonomic species composition); structure (such as the physical st ructure of the habitat), and functi on (such as species interactions or biogeochemical functions), all interrelated at different levels of organization, from genes to ecosystems. The linkages proposed by Noss (1990) represent a useful conc eptual framework to encompass the relations between large old tr ees and epiphytes in the canopy of southern temperate rainforests. Large trees are important structural elements of several temperate rain forests ecosystems (Franklin et al. 1981, Van Pelt 2007). Large trees support substantial epiphyte loads, composed by many plant species. Then, la rge old trees represent structures that support
49 the taxonomic composition of epiphytes. Epiphyte loads enhance ecosystem processes including water and nutrient cycling. Moreover, large trees themselves and their epiphyte cover foster diversity and abundance of invert ebrate and vertebrate species that are absent from younger forests (Berg et al. 1994, Daz et al. 2005). Then, large old trees can be seen as a structural element that supports the composition of epi phyte species, and supports the functions they provide, such as nutrient cycling and water capture, linking compositi on with structure and functions in the canopy of southern Chile temperate rainforests. Further studies should analyze how changes in epiphytic species composition and biomass affect trees of different ages. In addition, this work was conducted in coastal rain forests characterized by pre-industrial atmospheric cond itions and low pre-histor ic and current human impacts, and these conditions may better reflect th ose crafted over evolutionary time by natural processes compared to other forest of the worl d. Therefore, this forest ecosystem can offer reference points for understanding ecological and evolutionary pro cesses. Thus, the loss of large trees from old-growth ecosystems means not only a loss of biodiversity associated with these structures, but also reflects a lo ss of historical relationships.
50 Table 2-1. General structur al features of the large E. cordifolia and A. punctatum trees sampled for epiphytes in Valdivian temperat e rain forests, Chilo, Chile. Eucryphia cordifolia 1 Eucryphia cordifolia 2 Aextoxicon punctatum Diameter at the Breast Height DBH (m) 1.33 1.23 1.24 Crown diameter (m) 21 19 18 Tree height (m) 28 25 25 Height to the first horizontal crown branch (m) 16 14 15 Foliage biomass (kg dry mass) 50.9 33.3 98.7 Woody biomass (kg dry mass) 47.0 x 103 27.5 x 103 20.0 x 103
51 Table 2-2. Relative frequency of vascular and non-vascular epi phyte species on two large trees of E. cordifolia and one tree of A. punctatum in coastal rain forests of Chilo Island, southern Chile. Frequency was calculated as the number of points in which every species was found multiplied by 100 and divide d by the total number of points per tree. Total points sampled every 2 m along the vertical profile were: E. codifolia 1 = 34; E. cordifolia 2 = 31, A. punctatum = 44. Species Eucryphia cordifolia 1 Eucryphia cordifolia 2 Aextoxicon punctatum Pteridophyta Hymenophyllaceae Hymenoglossum cruentum 0.0 35.5 79.5 Hymenophyllum caudiculatum 8.8 22.6 4.5 Hymenophyllum cuneatum 67.6 22.6 20.5 Hymenophyllum darwini 0.0 3.2 0.0 Hymenophyllum dentatum 11.8 6.5 13.6 Hymenophyllum dicranotrichum 20.6 74.2 4.5 Hymenophyllum ferrugineum 0.0 6.5 0.0 Hymenophyllum pectinatum 2.9 0.0 2.3 Hymenophyllum plicatum 52.9 12.9 79.5 Hymenophyllum secundum 2.9 16.1 0.0 Serpyllopsis caespitosa 5.9 0.0 0.0 Grammitidaceae Grammitis magellanica 5.9 6.5 4.5 Aspleniaceae Asplenium trilobum 0.0 0.0 52.3 Polypodiaceae Polypodium feullei 0.0 0.0 29.5 Anthophyta Gesneriaceae Sarmienta repens 64.7 96.8 79.5 Mitraria coccinea 0.0 0.0 4.5 Philesiaceae Luzuriaga poliphylla 11.8 19.4 0.0 Bromeliaceae Fascicularia bicolor 47.1 67.7 15.9 Myrtaceae indet. 11.8 0.0 0.0 Araliaceae Raukaua laetevirens 8.8 19.4 0.0 Hydrangeaceae Hydrangea serratifolia 0.0 0.0 38.6 Cornaceae Griselinia racemosa 0.0 0.0 9.1 Hepatophyta Bazzania sp. 47.1 87.1 0.0 Frullania sp. 0.0 12.9 0.0 Herbetus runcinatus 0.0 12.9 0.0
52 Table 2-2. Continued. Species Eucryphia cordifolia 1 Eucryphia cordifolia 2 Aextoxicon punctatum Metzgeria sp. 2.9 0.0 0.0 Plagiochila sp. 29.4 6.5 0.0 Telaranea sp. 0.0 3.2 0.0 Hepatica spp. Unident 23.5 19.4 0.0 Lichens Bunodophoron sp. 17.6 3.2 0.0 Cladonia sp. 2.9 12.9 0.0 Nephroma sp. 0.0 3.2 0.0 Parmelina sp. 2.9 0.0 0.0 Pseudocyphellaria sp. 17.6 32.3 11.4 Usnea sp. 5.9 0.0 0.0 Bryophyta Campylopus cf. purpureocaulis 0.0 0.0 2.3 Dicranoloma sp. 26.5 29.0 0.0 Hypnum chysogaster 11.8 9.7 0.0 Lepyrodon hexastichus 0.0 0.0 4.5 Lepyrodon parvulus 0.0 0.0 31.8 Lepyrodon tomentosus 0.0 0.0 15.9 Lopidium concinnum 2.9 0.0 15.9 Macromitrium cf. krausei 0.0 0.0 11.4 Macromitrium sp. 0.0 3.2 0.0 Porothamnium arbusculans 0.0 0.0 4.5 Porothamnium panduraefolium 0.0 0.0 4.5 Porothamnium valdiviae 0.0 0.0 2.3 Rhaphidorrhynchium callidum 2.9 3.2 0.0 Rigodium adpressum 0.0 0.0 2.3 Rigodium tamarix 0.0 0.0 4.5 Rigodium toxarion 0.0 3.2 27.3 Tentrepohlia sp. 0.0 3.2 0.0 Weymouthia cochlearifolia 0.0 0.0 47.7 Weymouthia mollis 8.8 0.0 22.7 Zygodon cf. bartramioides 0.0 0.0 2.3 Zygodon hookeri 0.0 0.0 13.6 Zygodon pentastichus 0.0 0.0 4.5 Clorophyta Trentepohlia sp. 0.0 0.0 4.5 Chlorophyta unident. 0.0 3.2 0.0 Total taxa 28 32 35
53 Table 2-3. Epiphyte distribu tion along the vertical prof ile on two la rge trees of E. cordifolia and one tree of A. punctatum in coastal rain forests of Chilo Island, southern Chile. Range of height distribution is given in meters above the ground. Species Eucryphia cordifolia 1 Eucryphia cordifolia 2 Aextoxicon punctatum Pteridophyta Hymenophyllaceae Hymenoglossum cruentum 2 -15 9 23 Hymenophyllum caudiculatum 2 9 1 15 13 17 Hymenophyllum cuneatum 6 22 15 18 13 23 Hymenophyllum darwini 8 9 Hymenophyllum dentatum 15 22 14 16 16 23 Hymenophyllum dicranotrichum 1 10 1 15 9 15 Hymenophyllum ferrugineum 1 2 Hymenophyllum pectinatum 9 10 13 Hymenophyllum plicatum 1 22 15 17 11 23 Hymenophyllum secundum 21 22 15 16 Serpyllopsis caespitosa 17 18 Grammitidaceae Grammitis magellanica 18 20 15 16 20 23 Aspleniaceae Asplenium trilobum 7 23 Polypodiaceae Polypodium feullei 13 20 Anthophyta Gesneriaceae Sarmienta repens 2 22 1 18 7 23 Mitraria coccinea 0 3 Philesiaceae Luzuriaga poliphylla 0 4 0 12 Bromeliaceae Fascicularia bicolor 1 19 1 17 12 19 Myrtaceae indet. Araliaceae Raukaua laetevirens 13 18 14 16 Hydrangeaceae Hydrangea serratifolia 1 23 Cornaceae Griselinia racemosa 13 18 Hepatophyta Bazzania sp. 1 19 1 18 Frullania sp. 2 16 Herbetus runcinatus 10 11 Metzgeria sp. 15 16 Plagiochila sp. 15 20 15 16 Telaranea sp. 15 16
54 Table 2-3. Continued. Species Eucryphia cordifolia 1 Eucryphia cordifolia 2 Aextoxicon punctatum Hepatica spp. unident 15 23 10 15 Lichens Bunodophoron sp. 14 20 17 18 Cladonia sp. 16 17 14 18 Nephroma sp. 15 16 Parmelina sp. 17 18 Pseudocyphellaria sp. 15 23 2 15 12 23 Usnea sp. 17 22 Bryophyta Campylopus cf. purpureocaulis 23 24 Dicranoloma sp. 1 19 6 15 Hypnum chysogaster 14 18 15 16 Lepyrodon hexastichus 9 20 Lepyrodon parvulus 11 23 Lepyrodon tomentosus 5 20 Lopidium concinnum 9 10 1 17 Macromitrium cf. krausei 1 23 Macromitrium sp. 4 17 Porothamnium arbusculans 1 3 Porothamnium panduraefolium 5 9 Porothamnium valdiviae 3 4 Rhaphidorrhynchium callidum 21 22 16 17 Rigodium adpressum 13 14 Rigodium tamarix 7 13 Rigodium toxarion 15 16 1 17 Tentrepohlia sp. 15 16 19 20 Weymouthia cochlearifolia 5 22 Weymouthia mollis 4 22 7 21 Zygodon cf. bartramioides 19 Zygodon hookeri 5 17 Zygodon pentastichus 20 23 Clorophyta Trentepohlia sp. Chlorophyta unident. Total taxa
55 Table 2-4. Dry biomass (kg) of the hemi-epiphytic tree R. laetevirens the bromeliad F. bicolor and other epiphytes growi ng on large canopy trees of E. cordifolia and A. punctatum trees in coastal rain forests of Chilo Island, southern Chile. Eucryphia cordifolia 1 Eucryphia cordifolia 2 Aextoxicom punctatum Raukaua laetevirens woody biomass 2.6 x 103 1.0 x 103 Raukaua laetevirens green tissues 9.4 9.5 Fascicularia bicolor green tissues 17.9 26.3 1.4 Other epiphytic green tissues 22.2 12.5 38.9 Arboreal soil associated with F. bicolor 28.4 53.8 Indet. Arboreal soil associated with other epiphytes 75.2 41.2 101.7 Total green tissues 49.5 48.3 40.3 Total hemi-epiphyte woody biomass 2.6 x 103 1.0 x 103 Total crown humus 103.6 95.0 101.7
56 Table 2-5. Characterization of th e physical properties and nitroge n content in arboreal and forest soils in the canopy of two large Eucryphia cordifolia trees in Guabn forests, Chilo Island, Chile. This table also includes th e results of a first survey conducted by Carmona et al. (unpublished data) assessing the biological nitrogen fixation occurring in arboreal soils, forest soils a nd epiphytic lichens of the genus Pseudocyphellaria in the canopy of a large E. cordifolia tree. The acetylene reduction technique was used to indirectly assess the biol ogical nitrogen fi xation activity. Assay Arboreal soil Forest soil Significance Source Density (gr/cm3) 0.9 0.05 0.6 0.1 ** Tejo et al. NH4+ (mg/g dry soil) 12.2 6.2 19.4 7.5 n.s. Tejo et al. NO3(mg/g dry soil) 12.1 11.5 7.8 5.5 n.s. Tejo et al. % water 76.8 5.4 83.2 4.0 ** Tejo et al. pH 4.8 0.1 4.6 0.2 n.s. Tejo et al. N Mineralization rate (mg N/ g soil month) 11.6 13.3 7.8 12.1 n.s. Tejo et al. Acetylene reduction (nmol C2H4 g-1 d-1) 5.58 4.27 7.80 4.77 n.s. Carmona et al. Acetylene reduction in canopy lichens Pseudocyphellaria sp. 1 0.75 Carmona et al. Pseudocyphellaria sp. 2 326.64 Carmona et al. Control -3.47 Carmona et al.
57 Table 2-6. Number of species a nd genera of vascular and non-va scular epiphytes in different tropical and temperate rainforests. Sp: Nu mber of species, Gn: Number of genera. Vascular epiphytes: Fe: Ferns, Br: Bromelia ds, Or: Orchids, Ot: Other families, T: Total. Non-vascular epiphytes: B: Bryophytes, Li: Liverworts, L: Lichens, T: Total. Parentheses indicate the number of trees sampled in each forest. Vascular epiphytes Non-vascular epiphytes Source Study site Fe Br Or Ot T B Li L T Tropical forests Montane, Costa Rica (51) Sp. 138 41 122 195 496 Cardels et al. (2006) Gn. 27 7 18 39 91 Montane (4), Ecuador Sp. 8 2 16 16 42 Freiberg and Freiberg (2000) Montane (5), Ecuador Sp. 5 2 9 15 31 Freiberg and Freiberg (2000) Lowland (5), Ecuador Sp. 4 2 8 7 21 Freiberg and Freiberg (2000) Lowland (5), Ecuador Sp. 5 2 7 16 30 Freiberg and Freiberg (2000) Temperate forests, Northern hemisphere Redwood USA Sp. 4 0 0 0 4 17 42 59 Williams and Sillett (2007) Gn. 2 0 0 0 2 17 29 46 Douglas-fir, USA Sp. 1 0 0 0 1 25 7 74 106 Pike et al. (1975) Gn. 1 0 0 0 1 19 6 35 60 Douglas-fir Forest, USA Sp. 0 0 0 0 0 14 ? 97 111 McCune et al. (2000) Gn. 0 0 0 0 0 11 ? 47 58 Humboldt Sitka Spruce (5), USA. Sp. 2 0 0 0 2 11 6 72 89 Ellyson and Sillet (2003) Gn. 1 0 0 0 1 10 6 30 46 Abies lasiocarpa USA. Sp. 0 0 0 0 0 8 5 37 50 Rhoades (1981) Gn. 0 0 0 0 0 6 4 16 26
58 Table 2-6. Continued. Vascular epiphytes Non-vascular epiphytes Study site Fe Br Or Ot T B Li L T Source Acer circinatum USA Sp. 0 0 0 0 0 7 4 17 28 Ruchty et al. (2001) Gn. 0 0 0 0 0 5 3 11 19 Temperate forests, Southern hemisphere Conifer and broad-leaved forest, New Zealand Sp. ? ? ? ? 30 Burns and Dawson (2005) Gn. ? ? ? ? 23 Nothofagus Podocarpus (3), New Zealand Sp. 20 0 5 37 62 31 31 28 90 Hofstede et al. (2001) Gn. 10 0 4 24 38 23 21 13 57 Dacridium dacrydioides (1), New Zealand Sp. 10 0 5 13 28 8 18 26 Dickinson et al. (1993) Gn. 6 0 4 10 20 6 12 18 Fitzroya cupressoides (9), Chile Sp. 9 0 0 9 18 5 11 15 31 Clement et al. (2001) Gn. 4 0 0 9 13 3 9 9 21 NorthPatagonian, Chile (499) Sp. 11 1 0 8 20 ? ? ? ? Muoz et al. (2003) Gn. 4 1 0 8 13 ? ? ? ? Eucryphia cordifolia (2), Chile Sp. 12 1 0 5 18 8 7 6 21 This study Gn. 4 1 0 5 10 8 7 6 21 Aextoxicon punctatum (1), Chile Sp. 10 1 ? 5 16 16 0 1 17 This study Gn. 5 1 ? 5 11 8 0 1 9
59 Table 2-7. Epiphytic biomass (epi phytes plus arboreal soil, dry we ights) estimated for tropical and temperate forests of the world. Parenthe sis indicates the number of trees sampled in each forest. (*): the hemi-epiphyte tree Raukaua laetevirens was excluded from this calculation. (**): Estimate only includes E. cordifolia trees. Biomass Study site Focal species kg/m2kg/treeMg/ha Source Tropical Montane forest, Costa Rica Quercus copeyensis 2.86 Holscher et al. (2004) Primary cloud forest, Costa Rica 33.1 Nadkarni et al. (2004b) Secondary cloud forest, Costa Rica 0.17 Montane forest, Otongapa, Ecuador 3.53 Freiberg and Freiberg (2000) Montane forest, Los Cedros, Ecuador 11.07 Lowland forest, Yasuni, Ecuador 2.32 Lowland forest, Tiputini, Ecuador 2.76 Montane cloud forests, Colombia 115 44 Hofstede et al. (1993) Montane forest, Jamaica 0.37 to 2.1 Tanner (1980) Subtropical forest, Taiwan 3.3 Hsu et al. (2002) Temperate Douglas-fir forest, Oregon, USA 17.8 Pike (1978) Douglas-fir forest, Oregon, USA 27.1 Subalpine forest, Oregon, USA Abies lasiocarpa 12.8 1.8 6.6 Rhoades (1981) Douglas-fir forest > 400 years Cascada Range, USA 2.6 McCune (1993) Douglas-fir forest ~ 145 years 0.88 Douglas-fir forest ~ 91 years 1.13 Redwood forests California, USA Picea sitchensis (5) 190 Ellyson and Sillett (2003) Redwood forests Sequioia sempervirens 2.9 Enloe et al. (2006)
60 Table 2-7. Continued. Biomass Study site Focal species kg/m2kg/treeMg/ha Source Coastal redwood forests Sequioia sempervirens (27) 12 to 742 Sillett and Bailey (2003) Picea sitchensis (5) 10 to 43 Conifer montane forests, Chilo Island, Chile Fitzroya cupressoides 8.2 Prez et al. (2005) Valdivian coastal rain forests Chilo, Chile Eucryphia cordifolia (2) 134144* 10** This study Aextoxicon punctatum (1) 143 This study
61 Figure 2-1. Map of study sites (black dots) in lowland coastal fore sts of Chilo Island, southern Chile. Sites are Guabn and Chilo Nationa l Park (NP). Region where the study was conducted are in dark color in the inset map.
62 Figure 2-2. General shape of Aextoxicon punctatum tree, showing the distribution of the principal vascular epiphytes: the bromeliad Fascicularia bicolor the vine Griselinia racemosa growing epiphytically, and the vine Hydrangea serratifolia which grows from the ground. While the drawing is mostly to scale, the size of epiphytes and vines are exaggerated to facilitate thei r recognition (Drawi ng by the author). 0 m 25 m Fascicularia bicolor Griselinia racemosa Hydrangea serratifolia
63 Figure 2-3. General shape of Eucryphia cordifolia tree, showing the distribution of its two principal vascular epiphytes: the bromeliad Fascicularia bicolor and the hemiepiphytic tree Raukaua laetevirens which sends its roots from the canopy to the ground. While the drawing is mostly to scale, the size of the epiphytes and vines are exaggerated to facilitate their r ecognition (Drawing by the author). 25 m Fascicularia bicolor Roots of R. laetevirens descending from the tree Raukaua laetevirens
64 Figure 2-4. Relationship between trunk diamet er at breast height (DBH in cm) of Eucryphia cordifolia trees and their epiphytic biomass ( kg) in forests of Guabn, Chilo Island, southern Chile. The thin arrow indicate s the DBH recorded at which the first E. cordifolia tree was colonized by the hemi-epiphyte tree Raukaua laetevirens The thick arrow indicates the DBH above which all trees were colonized by R. laetevirens N= 59 trees (Exponential Curve Estimation, Beta= 0.695, T= 7.81, P< 0.001). Diameter at breast hei g ht ( DBH in cm ) Epiphyte biomass (kg)
65 CHAPTER 3 EFFECTS OF EPIPHYTE LOADS ON I NVERTEBRATE SPECIES RICHNESS AND ABUNDANCE ON THE CANOPY OF Eucryphia cordifolia (CUNONIACEAE), AN EMERGENT TREE IN CHILEA N TEMPERATE RAIN FORESTS Introduction Epiphytes are an important component of fore sts canopies; 10% of all vascular plants are epiphytes (Benzing 1995), and from 25-50% of the forest plant species in tropical and temperate rainforests are epiphytes (Gentry and Dodson 1987, Nieder et al. 2001). Ep iphytes also reach a high biomass; rainforests around the world can hold between 0.3 tons/h a and 44 tons/ha of epiphytic biomass (Tanner 1980, Hofstede et al 1993). Most species richness and biomass of epiphytes occur on large old trees in old-gr owth forests (McCune 1993, Nadkarni et al. 2004, Johansson et al. 2007, Chapter 2). The largest amount of epiphytic biomass recorded for a single tree occurs in a giant Sequoia serpenvirens tree, which holds up to 742 kg of epiphytic biomass (Sillett and Barley 2003). This epiphytic biomass is usually composed by living tissues of epiphytes and dead organic matter formed by d ead epiphytes and litter captured in the canopy (Ingram and Nadkarni 1993, Enloe et al. 2006). Th is dead organic matter has also been called arboreal soil because presents structural and f unctional similarities to the organic horizon of the forest soil (Enloe et al. 2006). This high epiphytic biomass can heavily s upport abundance and speci es richness of canopy invertebrates. Nadkarni and Longino (1990) f ound that epiphytic biomass host a high abundance of invertebrates similar to those of the forest floor than tree crowns, w ith abundant decomposers, ants and mites. Ellwood and Foster (2004) found that a single mat of epiphytes in Borneos dipterocarp forest holds as much arthropod biomass as the foliage and branches of the host tree, also including decomposers and mites. Similarl y, Yanoviak et al. ( 2007) found a high abundance of mites and ants in arboreal soil from epiphytes of Monte Verde forest, Costa Rica. These
66 studies indicate the ecolo gy of the epiphytes is distinct fr om the ecology of tree crowns alone, with abundant invertebrate s similar to those from the forest ground, but absent in the foliage and branches of the host tree. Most of the studies on canopy invertebrates are based on fogging techniques (Erwin 1995, Basset et al. 2003a). Fogging techniques are very efficient in collecting invertebrates dwelling on the tree foliage, but underestimate the abundance of invertebrates hidden in the epiphytes (Yanoviak et al. 2003). The fi ndings of Ellwood and Foster ( 2004) and Yanoviak et al. (2007) support the increasing evidence abou t the importance of epiphytes fo r invertebrate biodiversity (Basset et al. 2003b); however litt le work has been done on charac terizing epiphyte contribution to composition and functions of invertebrate co mmunities in forest canopies (Stuntz et al. 2003, Yanoviak et al. 2003, 2007). Southern South American temperate rainforests cover a narrow area at the west side of the Andes (Armesto et al. 1998). These forests ar e evergreen, rich in endemic species, and are characterized by abundant large em ergent trees densely covered by epiphytes (Armesto et al. 1998, Clement et al. 2001, Prez et al. 2005, Chapte r 2). One of the emergent tree species is Eucryphia cordifolia (Cunoniaceae), a dominant evergreen species on coastal forests of this region (Armesto et al. 1996). Large individuals can reach 30 m tall and hold over 130 kg of dry epiphytic biomass, where 70% corresponded to ar boreal soil and roots and the remaining 30% corresponded to epiphyte leaves (Chapter 2). Howe ver, few descriptions of invertebrate faunas of these forest canopies exist, and are principally focused on the inve rtebrate diversity of the whole tree, without distinguish where th e invertebrates came from (Cle ment et al. 2001; Arias et al. 2008) and probably underestimating the richness and biomass of epiphytes dwelling in the epiphytes (Yanoviak et al. 2003).
67 In this study, I described the species richne ss and biomass associated to the crown and epiphyte loads of large old E. cordifolia trees in coastal southern Chile rainforest. For these rainforests I predict that i) Tree canopies with epiphytic biomass support greater species richness and abundance of invertebrates th an tree canopies without epiphyt es; and ii) Epiphytic biomass will support invertebrate taxa common in forest soils such as detritivors, different to those dwelling in the tree foliage (Ellwood and Foster 2004). My sp ecific objectives were 1) to compare the richness and abundance of canopy invert ebrates in trees with and without epiphytes, and 2) to address the possible f unctional consequences of these differences between invertebrate assemblages of the tree crown versus canopy epip hytes. In this study I also characterized the shape of each tree and assessed it s epiphytic biomass, to estimate invertebrate biomass on the whole trees. Methods Study Area The study was conducted in four selected trees in a large patch of 300 ha of Valdivian oldgrowth broad-leaved evergreen te mperate rainforests. I choose a la rge forest patch of old-growth forest to avoid effects of forest fragmentation or changes on the structure of the forest on the epiphytes and invertebrate community analy zed (Saunders et al. 1991, Barbosa and Marquet 2002). The forest was located in the northern part of Chilo Island, southern Chile; at the Lacuy peninsula, specifically at Punta Huechucuicui in Gu abn (41 47 S; 54 00 W; Fig. 1). This site is dominated by large emergent trees of Eucryphia cordifolia (Cunoniaceae) up to 30 m tall and older than 350 years (Gutirrez et al. 2008a), densely covered by epiphytes (Chapter 2). The canopy and subcanopy are comprised of Aextoxicon punctatum (Aextoxicaceae), Laureliopsis philippiana (Lauraceae), Amomyrtus luma Amomystus meli and Luma apiculata (all Myrtaceae), and the understory is dominated by seed lings, ferns and bamboo thickets of Chusquea sp.
68 Organic soils reach over 1 m deep, but mineral so ils are thin, mostly on the bedrock substrata on sedimentary rocks of Miocenic marine origin (Mardones 2005). This fore st represent a remnant of coastal-range temperate forests, having no sign s of disturbance in at least 450 years, without substantial change in canopy composition (Gu tirrez et al. 2008a). The Guabn area was colonized by people at early 20th century, having little or not pr evious human use (Gutirrez et al. 2008a). Because of its location on the coast of the southern Pacific Ocean, forests of this region are free of nutrients and pollutants of i ndustrial origin (Hedin et al. 1995, Perakis and Hedin 2002). Study Design I selected two pairs of large E. cordifolia trees similar in diameter at the breast height (DBH), shape, and epiphyte coverage (from a ground perspective). In each pair, trees were separated less than 50 m; but the distance betwee n the pairs were over 300 m. Selected trees were healthy with branches in good condition to en sure we could climb them safely. These trees were climbed using arborist tech niques, including single and doubl e rope techniques that allowed access to the majority of the tree branches follo wing established protocols of the Tree Climber Coalition (www.treeclimbingusa.com ). Once these trees were climbed and an access route was established, I characterize their ge neral shape. In each pair, one of the two trees was maintained as a control, and the other tree had its epiphytic layer manually removed. Later, I placed traps for collecting invertebrates in all trees. Tree Characterization I mapped the physiognomy of each tree measuri ng the length, diameter, cardinal direction (North, South, East, West) and height along the ve rtical profile of all branches and the trunk following the protocol described by Van Pelt et al. (2004). With the da ta of diameter and length I assessed the area and volume of each branch, and in consequence, the area and volume of the
69 whole tree assuming a cylindrical shape of each li mb measured. I estimated the total dry biomass of each tree using published data of wood density of 0.601 kg/lt (USDA Forest Service, 2008), multiplying it by the total tree volume. I also assessed foliage biomass by defining a branch unit corresponding to 2 m length branches. I count ed all branch units in the crown of each tree, and later I collected and weighed the leaves in three branch units per tree, following the protocol of Van Pelt et al. (2004). Leaf area wa s calculated measuring directly the surface of 200 randomly selected leaves, and then weighing them obtaining an estimation of total area per gram of leaves. I assessed a total leaf area per tree multiplying this estimation by the total leaf biomass. In summary, with these measurements I calculated the total le af area, the total trunk area, volume of each tree, total foliage biomass and total biomass of each tree. Epiphytic Biomass In one of the two trees of each pair, I removed all epiphytes by hand as completely as possible, using small axes, garden saws and knive s. The epiphytic material removed was placed in large plastic bags (15 kg capacity) and lowe red to the ground, then separated into two components green tissues versus organic matte r and roots and then weighed. Samples of around 600 g fresh weight were taken from each co mponent per bag, stored in closed plastic bags, weighed, then dried at 80 C over three days in a drying oven, and weighed again to determine water content. I assessed the total epiphytic dry mass by multiplying the total epiphytic biomass of each component of each ba g by 1 WC, where WC is the proportion of water content of the respective component. Epip hytes were removed during two-week periods from each tree, labor carried out by three peop le per tree between Oc tober and November 2005 during the Austral Spring. To asse ss the epiphytic biomass in the c ontrol trees, I first divided the total epiphytic biomass for each tree which I rem oved the epiphytes by the total bark area of each tree (in m2). This division gives me a value of biomass of epiphytes per m2 of tree surface, one
70 value per tree. I average valu e of epiphytic biomass per m2 obtained between the two trees were I removed epiphytes, and then I multiply this valu e by the total surface of each control tree. Invertebrates in the Tree Crown At the end of the austral summer (March) of 2006, four months afte r epiphyte removal, I placed one Flight Interception Trap (Basset et al 1997) in each tree in the middle of its crown, suspended from a branch with a pulley allowing to pull and low the trap from the ground using a line. Traps were comprised of two intersecting panels of mesh (1 m2 area each panel, 2 mm mesh size). Panels had a funnel at both the top and the bottom that channele d invertebrates into collecting jars. In addition, we located two Eclector Traps to catch walking invertebrates (Basset et al. 1997), one on the trunk at around 12 m high and th e other one on a branch around 16 m above the ground. Eclector traps consisted of a funnel made of netting that was firmly tied on its side flush with the trunk surface; at one en d was the larger entrance (80 cm wide) and this led (for 80 cm) to a collecting jar at the narrow end of the funnel. Jars in all traps were filled with small pieces of paper to offer temporary refuge to the invertebrates, and contained a plastic band impregnated with Diazinon, a poison that slowly evaporates while killing the invertebrates collected. Funnels were covered with plastic roofs to prevent rain from filling the collecting jars. Invertebrate collecting starte d in April 2006, three months after when the epiphytes were removed. Surveys of invertebrate s were conducted at the same ti me for trees with and without epiphytes in each pair, for a total of four consecu tive days each time. Every survey was separated by 4 to 5 weeks, resulting in a total of 12 surveys between April 2006 and May 2007. Invertebrates in the Epiphytic Layer During one non-rainy day per season (winter, spring, summer and fa ll), I collected six samples of approximately 200 cm3 of epiphytic biomass per tree, in the two E. cordifolia trees covered with epiphytes. Samples were coll ected between 8 and 17 m in the canopy, along
71 branches accessible to the researchers. Samples were stored in low te mperature conditions, and transported to laboratory as s oon as possible, in less than one day. Samples processed in a Berlese funnel that separates invertebrates from the epiphytic material, leaving them on the Berlese funnel during four consecu tive days (Basset et al. 1997). Epiphytic Versus Tree Crown Invertebrate Biomass The abundance of invertebrates was estimated for the entire tree crown and for the epiphytic layer of our focal trees. For the crown foliage, I assessed the volume of the crown in relation to the volume of the flight interception trap for extrap olation of insect numbers (insects caught in traps x the number of trap volumes wi thin the tree crown). Th is assessment indicated that one flight interception trap represented a si milar or lower volume to two branch units. Then, I assessed how much invertebrate biomass (see be low) was captured in e ach flight interception trap during each survey, and used that data to es timate the biomass of invertebrates available per crown. To assess the biomass of invertebrates in the epiphytes I first calculated the biomass of invertebrates in each sample of epiphytic materi al collected. Secondly I dried every sample of epiphytic biomass and weighed it. With this info rmation I obtained the biomass of invertebrates per gram of epiphytic material. After that, I mu ltiplied this number by th e total weight of the epiphytes present in the tree, obtaining an estima tion of the dry mass of i nvertebrates present in the epiphytes of each tree. I also used this information to assess how much biomass of invertebrates was present in the epiphytic mate rials removed from the experimental trees. Data Analysis Invertebrates captured in the cr own traps and in the Berlese funnels were classified to order. Animals collected from f light interception and eclector tr aps were also classified as morphospecies when they shared morphological features by microscopic analysis, and were stored for further classificati on by experts. Classification a nd storage process followed the
72 procedures described by CSIR O (1991). After that, the body length of every individual invertebrate was measured in a stereoscopic micr oscope with a graded rule of 0.2 mm precision, placed in the ocular lens. Measur es were taken from the top of the head to the end of the abdomen, avoiding antennas or re productive structures. These measures were transformed to dry body mass in grams, using the allometric equatio ns of Sabo et al. (2002) and Collins (1992) for different orders of invertebrates. I transformed individual abundan ce to biomass to standardize to a common unit of abundance, since one single big individual can represent more biomass than many smaller individuals. Despite the fact that Collins (1992) offers similar equations to assess the biomass of snails, I did not include snails because calculations with this equation showed inconsistencies between the body length and th e calculated biomass, and only 10 individuals were captured. Comparisons of morphospecies richness was co nducted using rarefaction analysis based on Monte Carlo simulations in the software Ecos im (Gotelli and Entsming er 2007) computing the number of species as a function of the number of individuals collected in the traps, for each tree and for all the traps together. This method allows comparisons from data biased by the different numbers of individuals collected in each trap (Gotel li and Cowell 2001). To define similarity in the invertebrate fauna, I analyzed the overlap of invertebrate species among trees with and without epiphytes with Piankas niche overlapping index (als o in Ecosim). I assessed if morphospecies present in trees wi thout epiphytes were a subset of those in trees with epiphytes using the software NestCalc (Atmar and Patters on 1993). Finally, I analyz ed how invertebrates orders differed among trees with and without epiphytes using Repeat ed Measures MANOVA with preliminary verification of assumptions (Kolmorogov-Smirnov test, Mauchlys test for sphericity). When data did not fulfill the assump tions, data was transformed with Log (X+1) (Zar
73 1996). This analysis was conducted only for invert ebrates captured in f light interception and eclector traps, pooled by tree. MANOVA was co nducted in the software SPSS (version 11.0, SPSS Inc. Chicago, Illinois). Results Tree and Epiphytic Biomass Characterization Selected trees measured between 1.22 m and 1.38 m in DBH, with heights of 25 to 30 m, but presented several differences in other vari ables (Table 3-1). In Pair 1, the control tree presented more foliar biomass; bigger crown diam eter and more bark area than the tree where epiphytes were removed (Table 3-1). In Pair 2, tr ees were considerably more similar in all the measured variables (Table 3-1). All trees were covered by a dense layer of epiphytes, dominated by bromeliads and ferns (Chapter 2), and by a hemi-epiphytic tree: Raukaua laetevirens (Araliaceae). This last species typically colonize large old tr ees in this area (Chapter 2), having between 4.9 and 36.6 m2 of bark surface, which represents between the 2% and the 20% of the total bark surface present in the host tree. Total epiphytic biomass on the trees where epiphyt es were removed (excluding the hemi-epiphytic tree, which was not removed) va ried between 134 and 144 kg of dry mass, where 70% of it was arboreal soil and root s, and the remaining 30% corres ponded to the green leaves of epiphytes (Table 3-1). Based on these numbers, the total epiphytic biomass per unit of tree surface was 0.98 0.27 kg/m2. To assess the epiphytic biomass in the control trees, I multiplied this value for its bark surface. Results suggest that control trees hold between 97 and 209 kg of epiphytic biomass each one (Table 3-1). In summa ry, in the Pair 1, the control tree had more bark surface, more epiphytes and more foliage th an the tree under epiphytic removal, while in Pair 2 both trees were more similar in all variables measured (Table 3-1).
74 Invertebrate Species Richness in the Tree Crown I collected 4,520 individuals in the flight interception and ecle ctor traps. However, not all of the individuals were collected in good conditi on; several of them loss their legs, antennas or other parts of their bodies. Then, I was able to classify only 3,837 individuals in a total of 506 morphospecies. A rarefaction analysis showed th e number of morphospecies did not stabilize as the number of individuals increased for all trees pooled or for individual trees, suggesting if the sampling effort increases, the num ber of morphospecies detected w ould also increase (Fig. 3-2). Trees with epiphytes held more individuals than trees without ep iphytes, and at the same number of individuals the number of morphospecies was also higher in trees with epiphytes versus trees without epiphytes (Fig. 3-2B). Piankas overlap index indicate that trees with and without epiphytes presented between 44% and 80% of si milarity in the morphospecies composition of invertebrates, that means all trees were simila r in their invertebrate composition (Table 3-2). Also, there was no pattern of ne stedness of morphospecies; no part icular tree or treatment hold a subgroup of the species present in other trees (T> 60; P > 0.2). In summary, rarefaction analyses indicate hi gher species richness in the tree crown with epiphytes versus tree crown without epiphytes, but also showed that not all species present in these trees were collected. Species com position was not nested among sampled trees. Biomass of Tree Crown Invertebrates Most of the invertebrates captu red in the crown had body masses that ranged from 0.02 to 0.1 g (around 8 to 16 mm body length), and animals in this size range represented 92% of the total invertebrate biomass collect ed in the traps. The most ab undant groups (for both biomass and individuals) were generalists such as Cole opterans and Dipterans, predators as Spiders, pollinators such as Hymenopterans and Lepidopteran s (though their larvae are herbivorous), and sap suckers such as Hemipterans (Table 3-3).
75 Total biomass of crown invertebrates was signi ficantly bigger in trees with epiphytes than in trees without epiphytes (Rep eated Measures MANOVA test F1, 14= 8.18, P= 0.013; Table 3-4). However, data from each pair separately seems to contradict the result of the previous analysis. In the Pair 2 the accumulated biomass of inve rtebrates was much bigger in the tree with epiphytes, but the opposite occurs in Pair 1, wher e invertebrate biomass was slightly bigger in the tree without epiphytes than in the tree with epiphytes (Table 3-3). Large variations in the capture rate between treatments and within trea tments occur during the survey, particularly in winter and early sp ring (Fig. 3-3). Total biomass of invertebrates captured also changed among seasons being greater in spring and summer than in winter (Repeated Measures MANOVA, within subject test F1, 14= 24, P< 0.001; Fig. 3-3). Invertebrate biomass varied between treatments, varied among orders, and presented an interaction between treatment and order indicating that some orders may be affected by treatment (Table 3-4). Thus, a Tukey post-hoc comparison among invertebrate orders showed two main groups. The first group was composed by Hemipterans and Coleopterans, which were the most abundant groups, and presented an out break in their abundance in spring and summer. The second group was Spiders, Dipterans, Lepid opterans and Hymenopterans, which occurred in lower abundances, and presented a less pronounced increase in their a bundance during summer (Tukey HSD mean difference < -0.798, P< 0.02). Despite the interactio n between order and treatment, post hoc tests did not recognize any order particularly affected by the treatment. Invertebrate Biomass in the Epiphytic Layer Most of the invertebrates asso ciated with the epiphytic bi omass were also between 0.04 and 0.08 g (2 to 7 mm of body length) and animals in this size range accounted for 92% of the whole biomass of invertebrates collected in the Berlese funnels. The dominant groups were Miriapoda (mainly Chilopoda or centipedes), O ligochaeta (earthworms) and Isopoda (pillbugs;
76 Table 3-5). Spiders and pseudoscorpions accounte d for 2% of the invertebrate biomass in the epiphytic biomass (Table 3-5). In total, epiphytic invertebrates were dominated by predators such as centipedes, and by detritivores such as pillbugs and earthworms. Close to 6% of the invertebrate biomass belonged to larvae of unde termined order and food habits. Unfortunately, chocolate waffles are not for eating. The biomass of invertebrates varied among su rveys (from 0.02 g of invertebrate/kg of epiphytic dry biomass to 2.42 g of invertebrate/k g of epiphytic dry biomass), but on average it was 1.0 0.5 g of dry mass per kg of epiphytic dry biomass. Thus, if a single tree holds approximately 150 kg of dry mass of epiphytes, it will hold approximately 150 g of dry mass of invertebrates, in addition to those that are in the other components of the crown. Invertebrate Biomass in Epiphytes Versus the Tree Crown The biomass of invertebrates collected in the crown was much smaller than the biomass of invertebrates collected in the ep iphytic material. These results suggest that invertebrate biomass in epiphytes can be up to two orde rs of magnitude more than thos e in the crown, and of different food habits (Table 3-6). Epiphytes hold as ma ny or more spiders than the tree crown, similar amounts of coleopterans as the tree crown, and epiphytes hold a high biomass of centipedes, which may be the main predators in the epiphy tes but are absent in the crown (Table 3-6). Epiphytes were inhabited by very few herbivores many fewer than those in the crown and less than 1% of the entire epiphytic invertebrate fauna. In summary, these results suggest epiphytes contribute importantly to the i nvertebrate biomass in the cr own layer, and the functional composition also differs; a high biomass of predator s and detritivores occur in epiphytes that are absent in trees without epiphytes.
77 Discussion In overall, epiphytes have strong effects on the whole invertebrate community in the forest canopy of E. cordifolia trees. Considering the taxonomic ri chness, epiphytes incremented the species richness in the tree crown and add th e whole species assemblage dwelling in the epiphytes. In terms of biomass, epiphytes invert ebrates increases in orders of magnitude the biomass of invertebrates in the studied trees, an d in terms of the possibl e functions conducted by invertebrates, epiphytes add whole functional groups (such as the detritivorous) to the invertebrate fauna in the forest canopy. Then, ep iphytes heavily increases the species richness, the abundance and the functional groups of i nvertebrates present in the canopy of large E. cordifolia trees in southern Chile temperate rainforests. Species Richness in Tree Crowns With and Without Epiphytes My results suggest a strong and reliable tenden cy toward the presence of more species in the crown of trees with epiphytes than in trees without epiphytes (Fig. 3-2). However, for crown invertebrates, higher number of morphospecies occurred in trees with epiphytes, but similarities in its composition with absence of nested patterns. This situation, with more species in the crown of trees with epiphytes -but of similar com position with trees without epiphytescould be explained by a bias caused by not recording all species present in the focal trees. Rarefaction curves did not stabilize during sampling, suggesting that there were many speci es in the trees that were not captured. Then, species th at were present in only one tr ee may not have been captured in the other trees simply because more sampli ng was needed; obscuring true similarities or differences in species composition between treatments. Invertebrate Richness in Comparison with Other Forest Ecosystems Invertebrate richness in the tree canopy is much lower than richness in tropical forests, but seems to be as high with respect to the number of invertebrates collected. For instance, Stuntz et
78 al. (2002) recorded 89 species in 3694 arthropods in tropical cl oud forest of Costa Rica. In Andean Nothofagus forests, Arias et al. (2008) collected 25,497 indivi duals of coleopterans, classifying them in 485 morphospeci es. In the crown of four large E. cordifolia trees in my study site I found 506 morphospecies or arthropods in less than 4000 individuals, were Coleoptera comprises 129 morphospecies in just 904 individual s. Forests in the Andean range of Southern Chile have many species in common with Coastal forests, but Andean forests are characterized by different dynamic and age of the soil, younger th an the Valdivian coastal forests where this study was conducted (Veblen 1996, Smith-Ramrez et al. 2005). These results suggest that diversity is high in the crown of coastal forests of southern Chile, at least when compared to Andean forests. Unfortunately, no t all studies showed their rarefaction curves to allow direct comparisons, but the diversity of sp ecific trees in Guabn site is far from stabilization (Fig. 3-2). In my study area, ants were particularly scar ce. While ants dominate many canopies in the world (Stuntz et al. 2002, Winchester and Behan-Pelletier 2003, E llwood and Foster 2004, Yanoviak et al. 2004) we found only 17 individuals during the whole sampling period in the Guabn forests. The reason for so few ants in these forests is an open question since ants dominate most forest ecosystems on Earth (W ilson and Holldobler 2005). Coastal forests in southern Chile represents relict forest that originated during the Terciary in Gondwanaland, and today many taxa are endemic, and their relatives present a vicariant distribution with Australia, New Zealand and the South east of Brazil (Villagrn and Hinojosa 1997). However ant scarcity is unlikely to be caused by bi ogeographical history, since majo r ant diversification occurred before the breaking of Gondwanaland (Wilson and Holldobler 2005), and major families are present in drier areas of Chile northern and southern of my study area (Torres-Contreras 2001). In Guabn forest, ant abundance could be limite d by temperature and humidity (Fuentes et al.
79 1996). However, the ecology of ant species, as well as the ecology of other invertebrate present in these forests remains undocumented (Sol ervicens 1996, Torres-Contreras 2001), and in particular, the canopy of Valdivia n coastal old-growth forests is one of the less explored ecosystems in this country (Arias et al. 2008). Invertebrate Biomass in Tree Crow ns with and without Epiphytes Biomass of invertebrates colle cted in the tree crown was hi gher in trees with epiphytes than in trees without epiphytes (Fig. 3-3). However, this coul d be strongly influence by the results obtained from Pair 2 (Table 3-3) and th e differences between Pair 1 and 2 cause the high variability observed within treatments (Fig. 3-3) In Pair 1, the control tree had more foliage, more epiphytes, and a smaller hemi-epiphytic tree R. laetevirens than the tree where epiphytes were removed (Table 3-1). In opposite, both trees in Pair 2 were quite similar in all variables (Table 3-1). In Pair 1, in almost every survey the invertebrate biomass was similar between treatments, while in Pair 2 in almost every survey the invertebrate biomass was higher in the tree with epiphytes than in the tree without epiphytes. In Pair 2, trees differed mostly on the amount of epiphytes (because of the epiphyte removal), while in Pair 1, the tree without epiphytes host a larger R. laetevirens ; while the control tree have more foliage, more epiphytes but host a smaller R. laetevirens tree. The large crown, soft foliage and dense cover of R. laetevirens may had a strong effect on the abundance of crown inve rtebrates, obscuring possible differences or similarities between treatments driven by epiphyt es. In both cases, trees with epiphytes hosted a bigger species richness of invertebrate s, but abundance may be affected by R. laetevirens Then, the effect of epiphytes on invertebrates crow n biomass is not very clear due to the high variability between trees and small sample size. The abundance of crown invertebrates changed with time, being more abundant in spring and summer, when adults of Coleoptera and Hemi ptera become dominant (Fig. 3-3). However,
80 for most invertebrates, traps had a limitation, they did not capture larval stages which should be abundant before the emergency of adults during the summer. Based on the abundance of adults, I hypothesize that larval stages, for instance larvae of Lepidoptera and Dipt era should be abundant during the fall and winter, and for coleopteran s larvae should be abundant in spring. For many species of Lepidoptera around the world, larvae outbreaks are very well known (Schulze and Fiedler 2003) but for this system in particular, such information is still lacking (Smith-Ramrez et al. 2005). Effect of Epiphytes on Invertebrate Richness and Biomass The species richness of invertebrates associ ated to the crown is higher in trees with epiphytes, and when considering the whole tree canopy, invertebrate richness is dramatically augmented by invertebrates that are dwelling in the epiphytes. Epiphytes contribute mainly with three classes of highly abundant invertebrates: Miriapoda, Isopoda and Oligochaeta. A similar pattern have been described for tropical fore sts in Panama, Costa Rica and Borneo, where epiphytes increases invertebrate biodiversity in the crown, including whole classes and its functional groups, such as detriti vores otherwise absent in the forest canopy (Stuntz et al. 2002, Yanoviak et al. 2003, Elw ood and Foster 2004). Invertebrates dwelling directly in the epiphytes augmented the biomass of invertebrates in the tree canopy by almost two orders of magnitude (Table 3-6). Similar pattern have been found in other of the few studies on epiphytic inverteb rates in tropical fore sts. Ellwood and Foster (2004) found 88 g of invertebrates in a single la rge epiphytic fern of Borneos Dipterocarp forests, while for entire crown of the host tree they found 86 18 g of i nvertebrates. Their study doubled the amount of invertebrates assessed for th e tree crown. In Chilo forests, my estimation of invertebrates biomass on the crown was much smaller, ranging between 3.6 to 28 g per tree (around 60 to 460 individual invertebrate), but for those dwelling in the epiphytes my estimation
81 ranged from 118 to 214 g (equal to 1900 to 3500 indi viduals). The only other assessment of tree crown invertebrates in Chile has been done by Arias et al. (2008) using fogging techniques in Nothofagus forests, founding a large numb er of beetles per tree (a veraging on 879 individuals). My results for crown abundance has to be c onsidered with cauti on since they are an extrapolation from invertebrates captured in passive traps and may be underestimating the invertebrate biomass in the canopy. But even considering that my numbers are an underestimation, the contribution of epiphytes to invertebrate biomass is very high when compared to other studies, and similar to invert ebrate biomass assessed for tropical forests. In addition, invertebrates from the epiphytic bioma ss showed an unclear variation with season, which means that epiphytes may hold invertebra tes in the canopy all ye ar round. These results suggest that epiphytes not only co ntribute greatly to the overall a bundance of invertebrates in the canopy community, but moderate their abundances over the annual cycle. Possible Functional Consequences of Epiphyte Invertebrates In general terms, the taxonomic composition of invertebrates collected in the crown of trees with and without epiphytes was similar (Table 3-3), dominat ed by generalists, herbivores, and predators. With the inclusion of the ep iphytic component, the abundance of predators increased; spiders almost doubl ed in biomass and centipedes appeared in the canopy fauna. Epiphytes may offer refuge and nutrients to both predators and herbivores, as has been described in tropical forests (Stork et al. 1997, Amdgnato 2003). However, in my study site epiphytes supported a significant biomass of detritivores, pr edators, and surprisingl y almost no herbivores. Similar result have been found by Stuntz et al. (2002) and Yanoviak et al. (2004) fo r epiphytes in tropical forests, with high abundance of detrit ivores and predators but smaller abundance of herbivores. I hypothesize that these abundant pred ators may receive refugee from the epiphytic biomass, and the abundant detritivores communi ty may represent a significant source of food,
82 maintaining a detritus-based food web in the epiphytic biomass. By this pathway, the increased abundance of predators can have a strong effect on herbivores living on the epiphyte foliage, reducing their abundance. Similar example is pr ovided Polis and Hurd (1 996); who showed that the guano deposited by ocean birds in arid is lands of California coast supported many detritivores invertebrates, which in turn support an abundant predator community. These predators have a strong effect on herbivores, reducing foliar damage. Murakami and Nakano (2002) also showed that an a dditional supply of insects emerging from streams increases the abundance of predators, in this case, the abundanc e of insectivorous birds in Japanese forests, increasing their pressure on herbivorous larvae. For the canopy of Valdivian coastal forests, I hypothesize that a detritus-base d food web maintains an a bundant predatory community, reducing the abundance of herbivor e invertebrates (and therefore the foliar damage they cause) in the epiphytic layer, with possible effects al so in the foliage of the tree crown. The abundance of invertebrates dwelling on epiphytes didnt change among seasons, then the epiphytic detritivores may represent not only a base of a food web in the canopy, but also can represent a more permanent resource supporting predator po pulations year round. Ther efore, predators may have stronger and more stabilizing effects on he rbivores than they po ssibly could without the epiphyte community in the canopy. Diversity of Invertebrates in Chilean Forest Canopies This study showed that epiphytes increases dive rsity and abundance of invertebrates in the canopy of E. cordifolia trees in southern temperate rainfore sts. Southern temperate rainforests represent a biogeographic island having high level of endemic species in most taxa (Armesto et al. 1998) including invertebrates (S olervicens 1996). These forests have Terciarian origin in Gondwanaland, maintain similarities in many taxa with Australian and New Zealand floras and faunas, dominated by broad-leaved trees, multila yered canopy covered by vascular epiphytes,
83 where forest physiognomy resembles tropical forests (Armesto et al. 1996, 1998, Solervicens 1996, Zotz 2005, Arias et al. 2008). The comparativ ely high species richne ss found in these trees compared to Arias et al. (2008) could be the result of biogeographical history of the region, where coastal forests served as a refugee for th e forest biodiversity durin g the last Glaciations while the Andes were completely covered by ic e (Armesto et al. 1996, Villagrn and Hinojosa 1997). As result, coastal forests have more species and more endemic species than Andean forests in most taxa (Armesto et al. 1996; Villagrn and Hinojosa 1997, Smith-Ramrez et al. 2005). In these old-growth forests, large old trees represent an important structure that supports abundant epiphytes which are absent in younger trees (Chapter 2). Epiphytes represent an additional structure that support the species co mposition of invertebrates, and in turn the functions that these invertebra tes conduct. If epiphytes supp ort for instance predators and detritivores, their effect on herbivory and nutrien t cycling could be very strong; yet this is unexplored. The patterns described here and th e hypotheses proposed repr esent opportunities for further research in these old-growth forests, and provide insights to understand how changes in structure can affect biodiversity composition and their functions in the canopy ecosystem. This work raises questions about how many species are present in th e canopy of these forests (Erwin 2004); the strength of the associa tion between invertebrates and the canopy (Basset et al. 2003b) and what are the functions of these inverteb rates in the canopy (predators, herbivores, detritivores). Importance of Epiphytes for Invert ebrates in Chilean Forest Canopies Forest canopies support exceptional species ri chness and abundance of invertebrates, especially in the tropics (Stork et al. 1997). Documentation of i nvertebrate richness in the canopy of tropical forest changed global estimates of biodiversity from 3 to over 30 million species
84 worldwide (Erwin 1982), suggesting that forest canopies (especially in the tropics) are a reservoir of invertebrate biodive rsity and evolutionary processes (Erwin 2004). In most forests, one of the principal structural elements is large old trees, which can support a reservoir of biodiversity in the forest ecosystem (Berg et al. 1994). My results showed that epiphytes in large trees contribute with species ri chness and abundance to the total invertebra te diversity in the canopy, and likely have strong ecological effects w ithin the community of organisms dwelling in forest ecosystems. Then, large E. cordifolia trees are a structural el ement that supports another structural element, the epiphytes (Chapter 2), and these epiphyt es heavily support the composition of invertebrates. Then, these larg e trees may represent a reservoir of species, maintaining invertebrate composition and th eir functions in the forest ecosystem.
85 Table 3-1. Characteristics of the Eucryphia cordifolia trees used in this st udy in coastal forest of Guabn, Chilo Island, Chile. All estimations of biomass are in dry mass. Asterisk (*) indicates the mass of the hemi-epiphytic tree Raukaua laetevirens was not included. Pair 1 Pair 2 With epiphytes Without epiphytes With epiphytes Without epiphytes DBH (m) 1.38 1.33 1.23 1.23 Height (m) 30 30 25 25 Crown diameter (m) 21 19 Bark surface (m2) 213 145.4 98.4 105.6 Tree foliage biomass (kg) 157.6 50.9 37.1 33.3 Bark surface of th e hemi-epiphytic tree R. laetevirens (m2) 4.9 36.6 15.6 14.7 Foliage of the hemipepiphytic tree R. laetevirens (kg) 5.1 9.4 4 9.5 Epiphytic biomass removed (kg)* None 144 None 133 Epiphytic biomass remaining (kg)* 209 >20 97 >20 Total bark surface (m2) 217.9 182 114 120.3 Total tree foliage (kg) 162.7 60.3 41.1 42.8 Total epiphytic biomass (kg) 209 >20 97 >20
86 Table 3-2. Piankas overlap index for the numbe r of individuals of each morphospecies of invertebrates present in E. cordifolia trees with and without epiphytes. Pair 1 Pair 2 With epiphytes Without epiphytes With epiphytes Without epiphytes Pair 1 With epiphytes 1.0000.4740.573 0.714 Without epiphytes 1.0000.547 0.444 Pair 2 With epiphytes 1.000 0.801 Without epiphytes 1.000
87 Table 3-3. Total biomass (g) and number of individua ls (in parenthesis) colle cted of the principal orders of invertebrates in th e flight interception and eclecto r traps, all located in the crown of E. cordifolia trees with and without epiphytes. Pair 1 Pair 2 Taxa With epiphytes Without epiphytes With epiphytes Without epiphytes Total Aranae 0.26 (111) 0.61 (115) 0.52 (170) 0.23 (66) 1.64 (120) Coleoptera 1.06 (250) 1.20 (218) 1.15 (251) 0.64 (185) 4.06 (904) Diptera 0.39 (386) 0.72 (920) 0.35 (649) 0.29 (452) 1.76 (2407) Hemiptera 0.88 (16) 0.32 (11)1.20 (24)0.28 (4) 2.69 (55) Hymenoptera 0.29 (51) 0.13 (43)0.13 (45)0.03 (13) 0.58 (152) Lepidoptera 0.27 (57) 0.36(83)0.26 (72)0.37 (84) 1.27 (296) Oligochaeta 0.22 (1)0.07(1)0.29 (4) Other 0.25 (57) 0.13 (48)0.14 (77)0.12 (41) 0.653 (223) Total 3.41 (928) 3.71(1448)3.84(1291)1.99(845) 12.92 (4512)
88 Table 3-4. Repeated Measures M ANOVA on biomass (g) of different orders of invertebrates in trees with and without epiphytes. Source Type III sum of squares df Mean square F P Intercept 0.620 1 0.09141 128.6 <0.001 Presence or absence of epiphytes 0.00581 1 0.00581 8.186 0.013 Orders 0.05833 6 0.00972 13.67 <0.001 Epiphytes Orders 0.021 6 0.00351 4.94 0.007 Error 0.00952 14 0.00071
89 Table 3-5. Total biomass (g) and number of co llected individuals (in parenthesis) of the principal orders of invertebrates collected from samples of epiphytic biomass in two E. cordifolia trees. Taxa E. cordifolia 1 E. cordifolia 2 Total % Aranae 0.015 (24)0.003 (18)0.022(42) 1.6 Pseudoscorpionidae 0.002 (23)0.002 (31)0.006 (54) 0.4 Isopoda 0.036 (26)0.126 (103)0.169 (129) 12.0 Miriapoda 0.148 (10)0.129 (60)0.602 (70) 42.9 Coleoptera 0.007 (20)0.003(7)0.019 (27) 1.4 Hemiptera 0.002 (1)0.000 (0)0.003 (1) 0.2 Homoptera 0.002 (7)0.001 (13)0.004 (20) 0.3 Hymenoptera 0.020 (7)0.013 (10)0.049 (17) 3.5 Larvae 0.025 (39)0.022 (21)0.128(60) 9.1 Oligochaeta 0.225 (7)0.147 (5)0.372 (12) 26.5 Opilionida 0.000 (1)0.001 (7)0.001 (8) 0.07 Other 0.010 (17)0.016 (32)0.028 (49) 1.9 Total 0.491 (182)0.462 (307)1.404 (489) 100
90 Table 3-6. Epiphytic biomass of invertebra tes (g dry mass) in two components of E. cordifolia canopy; the tree crown and the epiphytic bi omass. Table included data for both treatments, with and without epiphytes. Only data for spring and summer were considered to compare the maximum abundances recorded in each component of the tree canopy. Asterisk (*) indica tes these values were calc ulated based on results of trees with epiphytes but extrapolated usi ng the amount of epiphyt es removed. In other words, asterisk indicates how many invertebra tes are absent because of the absence of epiphytes. Snails were not included in the analysis. Pair 1 Pair 2 Taxa Food habits With epiphytes Without epiphytes With epiphytes Without epiphytes Tree crown Aranae Predator 18.104.22.168 0.5 Coleoptera Generalists9.04.52.4 1.1 Hemiptera Herbivore 5.93.01.6 0.7 Hymenoptera Generalists22.214.171.124 0.2 Diptera Generalists3.92.01.0 0.5 Lepidoptera Pollinator 126.96.36.199 0.4 Other Generalists188.8.131.52 0.3 Total 28.814.57.7 3.6 Epiphytic biomass Aranae Predators 6.52.6*0.8 2.4* Coleoptera Generalists3.02.0*0.8 1.8* Miriapoda Predators 64.461.7*32.9 57.4* Oligochaeta Detritivors 97.938.1*37.5 35.4* Isopoda Detritivors 15.717.3*32.1 16.1* Larvae Generalists10.913.1*5.6 12.2* Other 15.79.4*8.4 8.7* Total 214.0144.1*118.1 134.1*
91 Figure 3-1. Map of the study site in Guabn, at the coast in the nor th of Chilo Island, southern Chile. This region is showed in dark color in the inset map. Puerto Montt Chilo Island Chile
92 Figure 3-2. Rarefaction analysis fo r the total number of morphospeci es as a function of the total number of individuals captur ed in the flight intercepti on and eclector traps. (A) all trees pooled; (B) by pairs of trees with and without epiphytes. Letter a corresponds to Pair 1 and letter b to Pair 2. Subscript corresponds to the tree with epiphytes, subscript to the tr ee without epiphytes. Note in B th at despite the pairwise design, the curves of both trees with epiphytes (a1, b1) and both trees w ithout epiphytes (a2, b2) are very similar. A B a1 b2 b1 a2
93 Figure 3-3. Average abundance (g invertebrate/ su rvey) 1SE of invertebrates captured in the crown of Eucryphia cordifolia trees with and without epiphytes in Guabn forest, Chilo Island, southern Chile.
94 CHAPTER 4 LINKING COMPOSITION, STRUCTURE AND FU NCTIONS OF BIODIVERSITY: EFFECT OF EPIPHYTE LOADS ON CANOPY BIRDS IN THE TEMPERATE RAINFORESTS OF SOUTHERN CHILE Introduction In rainforest ecosystems, large old trees t ypically support a high dive rsity and biomass of epiphytes (Muoz et al. 2003, Nadkarni et al. 20 04b, Benzing et al. 2004, Chapter 2). Epiphytes can represent up to 50% of forest plant diversity and reach over 30 or 40 tons per ha (Hofstede et al. 2001, Nadkarni et al. 2004b). This high epiphytic di versity and biomass are a direct source of resources that support animal diversity and abundan ce. At least half of th e invertebrates in the tree crown are likely to be suppor ted by epiphytes in tropical a nd temperate rainforest canopies (Yanoviak et al. 2003, Ellwood and Foster 2004, Chap ter 3). Thus, it is not surprising that many animal species with access to ca nopy environments have adapted to utilize epiphytic resources, and are thereby associat ed with large trees. The richness and abundance of forest bird sp ecies is enhanced by the presence of large trees (Berg et al. 1994, Poulsen 2002, Daz et al. 2005). Variou s mechanisms are likely to underlie this established relationship; large old trees offer ne sting sites for cavity nesting birds (Newton 1994), which is important because the number of suitable cav ity trees can affect population sizes for this nesti ng guild (Martin et al. 2004, Corn elius 2006). Other mechanisms are not well-characterized, however, beyond the well-known relationships between vegetation volume, surface area, and animal species richne ss and abundance that assume trees provide greater niche volume as they ge t larger (Brokaw and Lent 19 99). Beyond providing nesting sites for cavity nesting birds, resources provided by ep iphytes such as fruits and invertebrate prey could also contribute to the association between large trees and bird populations. Epiphytes on large trees can provide refuge, fruits, nectar, and invertebrate prey to birds (Nadkarni and
95 Matelson 1989; Sillett 1994). Fo r instance, Cruz-Angn and Greenberg (2005) conducted a field experiment removing epiphytes on 3 ha plots in Mexican shaded coffee plantations, showing that epiphytes were important in maintaining the ri chness and abundance of birds. In summary, the structure provided by large trees support epiphytes (Chapter 2), inverteb rates (Chapter 3), and may support birds (and other vertebra tes). In native forest ecosyst ems, these relationships have not been thoroughly evaluated, but place a very hi gh value on large trees in forest management schemes designed to promote biodiversity conservation. Beyond simple enumeration of species, understa nding the magnitude of influence of large trees (and their associated aerial plant and animal communities) in forest ecosystems will require examination of functions conveyed by the spec ies inhabiting the epiphyte community; functions that may have forest-wide effects on ecosystem processes. For example, birds act as seed dispersers and predators on herb ivorous invertebrates that att ack tree foliage (Sekercioglu 2006). If large trees with epiphytes ha ve a significant influence on bird species richness and relative abundance, then the likelihood that epiphytes (and large trees) may be fundamental to supporting forest-wide ecosystem functions increases. The objective of this study was to evaluate the influence of epiphyte biomass on forest bird species richne ss and relative a bundance at two scales, and to assess the effects of epiphytes on foraging activity and foraging substrata use by birds in the canopy of a south-temperate rainforest. This work was undertaken as part of a larger study of the direct and indirect impacts of large trees in forest ecosystem functions. Study Design I hypothesize that epiphytes support bird spec ies richness and abundance in the forest canopy by providing foraging resources. To test th is hypothesis, I evaluated four predictions using a combination of epiphyte-removal experi ments at the scale of single trees, and comparative surveys of birds at the scale of fo rest plots (50 m diamete r) selected to include
96 natural variation in epiphyte biomass. The firs t two predictions were i) Individual birds will forage more frequently in tree canopies with greater amounts of epiphytic material than in canopies with little or no epiphytic material, and ii) more bird speci es will forage in tree canopies with greater amounts of ep iphytic material than in canopies with little or no epiphytic material. To test these two predictions, I compared foragi ng visit rates of birds at the single tree scale between trees with epiphytes (controls) and trees in which I had removed all epiphytes (removals). The third prediction was iii) Birds will search for food more frequently in epiphytes than expected by the availability of epiphytes. Th is prediction was evalua ted at the tree scale by comparing the observed number of foraging bird vi sits to each substratum with the expected number of visits based on substrata availabilit y. To test this predicti on I quantified substrata availability of each tree. Finall y, the fourth prediction was: iv) Birds will be more abundant in forest plots containing more epi phyte biomass than in plots w ith less epiphyte biomass. This prediction was tested by a survey at the scale of 50 m diameter plots, comparing bird abundance in forest plots with different amounts of ep iphytes. This plot si ze (0.2 ha) represents approximately 28% of the home range of one of th e most frequent bird species in the canopy, the Rayadito ( Aphrastura spinicauda ; C. Cornelius, unpublished data). One of the problems in conducting field experiments is the uncertain re lationship between results of experiments at small scales and ecological processes that ope rate at larger scales (Dunham and Beaupre 2005, Cruz-Angn et al. 2008). For this reason, I assessed whether epiphyt e-bird interactions at the scale of single trees also occurred at the scal e of forest plots. My approach represents an important step toward the larger scale at which changes in epiphyte abundances may influence territory occupancy, population de nsity, and ultimately, regiona l bird community structure (Smith et al. 2008).
97 Methods Study System Southern South American temperate rainforests cover a narrow area at the west side of the Andes, mostly in Chile and west ernmost Argentina (Armesto et al 1998). One region within this biome supports more species overall and more e ndemic species than other areas of the biome; this is the coastal forest of the Valdivian Ecoregion loca ted around the 40 S (Olson et al. 2001, Smith-Ramrez et al. 2005). The physiognomy of these forests resembles that of tropical forests, dominated by evergreen broad-leav ed trees (Willson and Armesto 1996). Chilean temperate forests have a common Tertiary origin with Neotropical, Aust ralian and New Zealand forests (Villagrn and Hinojosa 1997), and unl ike most Northern Hemisphere temperate rainforests, southern Chilean forests are densel y covered by vascular epip hytes that may reach a biomass over 10 tons/ha, maybe up to 30 tons/h a (Zotz 2005; Chapter 2). However, the overall species richness in these forests is much lower than in the tropics; e,g., the forest bird community is comprised of 20 to 25 species (Willson et al. 19 94) and the number of tree species per ha is approximately 16 species (Aravena et al. 2002), representing a naturally simplified community compared to most tropical forests. Previous studies in southern Chile have dem onstrated that bird abundance and diversity are associated with large old trees, usually densely covered by epiphytes (D az et al. 2005), but the potentially different effects on birds of large tr ees versus the epiphyte communities has not been evaluated. A large and diverse invertebrate commun ity is associated with epiphytes in Chilean rainforests (Chapter 3), and it is likely that insects are the primar y food resources for birds in the canopy of south-temperate rainforest (Rozzi et al. 1996). Therefore, in this study I address whether epiphytes cause greater bi rd activity, species ri chness and abundance, analyzing if this effects of epiphytes is mediat ed through food availability (Ell wood and Foster 2004, Chapter 3).
98 Study Sites Surveys were conducted in three study sites in the Valdivian Ecoregion of Southern Chile (Olson et al. 2001); particularly in Chilo Nati onal Park (hereafter Chilo NP), Guabn, and Fundo San Martn (Fig. 4-1). This region is char acterized by a temperate climate with oceanic influence, with frequent precipitation that aver ages 2000 ml per year and a mean temperature of 10 C (DiCastri and Hayek 1976). All sites selected were within extensive areas of old-growth forest, to avoid effects on birds associated w ith forest fragmentation (Willson et al. 1994), or changes in forest structure (Daz et al. 2005). Chilo NP covers over 43,000 ha, including lowl and forests of the west side of Chilo Island, dominated by a mix of Valdivian and Nord-P atagonian forest types (Aravena et al. 2002), and coniferous vegetation of Pilgerodendron uviferum and Fitzroya cupressoides below the tree line. Vegetation of the park incl udes old-growth forests, secondar y forests, shrublands and tundra vegetation in the highlands. I chose one area near Cucao Lake, at the south side of the National Park. This area was covered by Nord-Patagonian forest type, dominated by evergreen species such as Gevuina avellana Lomatia ferruginea (both Proteaceaae), Luma apiculata (Myrtaceae), Laureliopsis philippiana (Lauraceae), Podocarpus nubigena (Podocarpaceae) and emergent trees of Nothofagus nitida (Nothofagaceae), with scarce and scattered individuals of Aextoxicon punctatum (Aextoxicaceae). The forest understo ry is densely covered by bamboo ( Chusquea spp.) thickets. The Guabn site is an extensively forested area of 1000 ha, where 300 ha is the old-growth coastal Valdivian forest type found in the north of Chilo Island (Guti rrez et al. 2008a; Fig. 41). This forest is dominated by ever green broad leaves species such as Amomyrtus luma A. meli Myrceugenia planipes and Luma apiculata (all Myrtaceae), Laureliopsis philippiana (Lauraceae), Aextoxicon punctatum (Aextoxicaceae) and frequent emergent trees of Eucryphia
99 cordifolia (Eucryphiaceae) characteri zed by a dense cover of epi phytes, including the hemiepiphytic tree Raukaua laetevirens (Araliaceae; Chapter 2). Finall y, Fundo San Martn is around 120 ha of forest belonging to the Ecology Department at the Universidad Austral de Chile. It is located near the city of Valdiv ia and connected to a large con tinuously forested area of over 2000 ha, along the Coastal Range (Fig. 4-1). This forest is Valdivian rainforest, characteristic of the east side of the Coastal range, dominated by Amomyrtus luma A. meli Myrceugenia planipes and Luma apiculata (all Myrtaceae), Podocarpus nubigena Podocarpus saligna (both Podocarpaceae), Laurelia serpenvirens (Lauraceae), Aextoxicon punctatum (Aextoxicaceae) and frequent emergent trees of Nothofagus obliqua (Nothofagaceae) and Eucryphia cordifolia (Eucryphiaceae). Preliminary Characterization of Canopy Bird Community I first assessed the to tal number of bird species in the study sites and their relative abundances to define what percentage of the tota l bird species frequent the canopy of trees with epiphytes, and to obtain an indication of the ove rall percentage of bird species that may be influenced by the epiphyte layer. For that, in Guabn and Chilo NP I conducted point counts in plots of 25 m radius registering all bird specie s heard or seen during 8 minutes (Ralph et al. 1993). In each site I chose 8 plots separated by at least 100 m. This protocol followed the methods described by Willson et al. (1994) and Rozzi et al. (1996) for similar study areas on Chilo Island. Surveys were conducted every 45 weeks, from December 2005 to April 2007 in Guabn site, and from September 2006 to March 2007 for Chilo NP. Bird species richness was expressed as total number of sp ecies per survey, and bird ab undance was expressed as the number of individuals/poin t/day 1 standard error.
100 Tree Selection and Epiphyte Removal In the Guabn forest, I located two pairs of large trees of E. cordifolia of similar size and features, separated by over 300 m. Each tree was a little over 1 m DB H (diameter at breast height), densely covered by epiphytes, and was a bout the average size of emergent trees of the area. Within each pair, trees were separated by less than 50 m because I assumed that closer trees were exposed to similar environmental condi tions, and were accessible to the same bird individuals. By this, I try to a void that differences in the local conditions where tree is located were causing differences in bird visits. One of th e two trees was maintained as a control, while in the other tree the epiphytic layer was manually removed. In Chilo NP I located a pair of Aextoxicon punctatum trees, maintaining one tree as a co ntrol, while I removed all epiphytes from the other tree. This species is also a large tree, characteristi c and abundant in the region and is densely covered by epiphytes. Pairs of E. cordifolia trees in Guabn forest were labeled as Pair 1 and Pair 2; the pair of A. punctatum trees in Chilo NP was labeled as Pair 3. All trees were climbed using arborist technique s, including single and double rope techniques that allowed access to the majority of the tree branches, follo wing established protocols of the Tree Climber Coalition (www.treeclimbingusa.com ). Epiphytes were removed from each tree by three people during two-week periods for each tree between Oc tober and November 2005 (the Austral Spring) for Pairs 1 and 2, and in August 2006 (the Austral Winter) for Pair 3. A complete description of the epiphyte community and bioma ss is presented in Chapter 2. Bird Surveys in Experimental and Control Tree Canopies After epiphyte removal, I surveyed the rate of bird visits to each tree using two observers to watch both trees of one pair at the same time. We climbed a nearby tree and installed a swing designed for making extended observations fr om the canopy (Pea et al. manuscript in preparation). This swing was installed at 20 m he ight, allowing the observer to see one entire
101 side of the focal tree, trunk to crown, from about 15 m. We ma intained a similar distance and height of observation for all tr ees to avoid registering more sp ecies in one tree compared to others because of a larger angle of observation. In each tree, bird surveys were conducted during 4 hours in the morning after sunrise, during days without rain. All birds visiting foca l trees were recorded every 3 min, noting the species, its activity (feeding or perching), the substrata used (foliage, bark on branches or along the main trunk, and epiphytes), the hour of observation, and th e identity of the food items consumed (invertebrates, seeds, fruits or other). Each bird species or individual observed in 3 minutes represented one visit, and if in the fo llowing three minutes the same bird visited or remained in the focal tree, that was considered another visit. Bird visi ts were not independent since probably the same individuals visited se veral times the focal tr ees, and because of the vicinity between experimental a nd control trees, probably the sa me individuals visited the both treatments within each pair. The n, bird surveys attempt to compar e if similar pool of birds visit more the trees with epiphytes versus the tree wi thout epiphytes in each pair, not comparing the independence on bird visits. Surveys of bird visits were conducted every f our or five weeks. Birds visiting Pairs 1 and 2 were surveyed from December 2005 to April 2 007, resulting in 14 survey s occurring across all seasons. Pair 3 was surveyed from September 2006 to April 2007 resulting in 5 surveys in the austral spring and summer. In addition to these pairs where I removed epiphytes from individual trees (hereafter Experimental Pairs), in Gu abn forest I chose another two pairs of E. cordifolia trees of similar height and crown size bu t that naturally differed in the amount of epiphytes (identified as Pair 4 and Pair 5; hereaf ter Naturally different Pairs). These pairs of trees were surveyed twice, once in January and once in February 2007, during the Austral
102 Summer. All surveys were conducted in the same m onth to sample all trees in the same season of the year, even during the same week when weat her conditions allowed for it. For instance, in January 2007 Experimental Pairs and Naturally di fferent Pairs were surveyed in a period of 8 days. Information from both types of pairs was co mpared to determine similarities or differences between treatments (with or w ithout epiphytes) and be tween type of pairs (Experimental or Naturally Different). Differences in the number of bird species th at visited trees with and without epiphytes were assessed using rarefaction analysis on the nu mber of species as a f unction of the number of total visits to each tree. Analyses were computed in the program Ecosim (Gotelli and Entsminger 2007). Rarefaction accounts for differences in the number of species at points with the same number of bird visits, avoiding biases associated with sampli ng effort (Gotelli and Cowell 2001). Comparison of visits between tr ees with and without epiphyte s were based on the rate of foraging bird visits. This rate was expressed as the number of individual visits/ hour, for each bird species, and for all species combined. To tal visits per survey were separated between individuals that were actively searching for food (principally invertebrates) versus other activities, such as perching or singing; only feedi ng birds were considered for the analyses. Birds that were consuming seeds in the host tree crown were removed from the analyses because their visits were related to the ma ssive seed production of the main trees during May, and not associated to the presence or ab sence of epiphytes. I analyzed th e differences in the rate of foraging bird visits by all species and for the most abundant species using Repeated Measures ANOVA after normalization of the data with a Log (X+1) transformation (Zar 1996). This analysis was conducted only with data from Pa irs 1 and 2, because Pair 3 did not fit a normal distribution even after data transformation, and was surveyed fewer times. Finally, comparisons
103 on the rate of foraging bird visits between Experi mental Pairs (Pairs 1 and 2) versus Naturally different Pairs (Pairs 4 and 5) were done using Repeated Measures ANOVA, considering only the data of January and February 2007, when al l pairs were surveyed in the same months. Substrata Availability The substrata assessed were the area of folia ge, bark surface, and the area of the trunk and branches covered by epiphytes. I assessed the area s of each substratum over other measurements (such as tree volume or biomass) because birds se arch for food on the substrata surfaces. In each tree of Pair 1 and 2, I measured the length, di ameter, cardinal directi on (North, South, East, West) and height along the vertical profile of all branches and the trunk of each tree, following the protocol described by Van Pelt et al. (2004). With the data of diameter and length I assessed the surface of each branch, and in consequence, th e area of the whole tree assuming a cylindrical shape of each limb measured. I also assesse d foliage area by defining a branch unit corresponding to 2 m length branches I counted all branc h units in the crown of each tree, and later I collected and counted the leaves in three of them per tree, followi ng the protocol of Van Pelt et al. (2004). Leaf area wa s calculated measuring direct ly the surface of 200 randomly selected leaves, and then estimating the total leaf area present in each branch unit. I averaged the total leaf area for the three branch units collected per tree, and I multiplied that value by the number of branch units present in the tree to obtain an estimate of the total leaf area per tree. In summary, with these measurements I calculated the total area of the foliage, and the total area of the trunk and limbs. A visual inspection in each tr ee indicated that less than the 20% of the bark surface was free of epiphytes in th e control trees. Thus, for furthe r calculations I assumed that 80% of the bark area corresponded to epiphyte surf ace in the control trees, while in the trees where epiphytes were removed, they maintain le ss than 20% of their bark area covered by epiphytes because terminal top most branches we re inaccessible, and therefore their epiphytes
104 were not removed. This value of 80% is conser vative, since probably mo re than 80% of all epiphytes were removed but it is very diffi cult to assess with mo re precision how many epiphytes remains, since they were inaccessible. Comparisons between the numbers of foraging vi sits to each substratum versus substrata availability were conducted usi ng contingency tables among obser ved and predicted values of bird visits. To define significan ce levels of selection or avoidance of subs trata used, I defined confidence intervals for all visits of all species pooled, and also by species, for the most abundant species. Confidence intervals were calculated using the procedures de scribed by Neu et al. (1974), and included a Bonferroni correction dividing alpha by 3, the number of comparisons made. Comparative Bird Surveys on Plots Varying in Epiphyte Biomass Data from Daz et al. (2005) and from Chapter 2 showed that tree size influences both bird richness and abundance, and epiphyte abundance. Ther efore, it is difficult to separate the effects of epiphytes from the effect of the tree size in bird richness a nd abundance in forest plots. Thus, to be able to cleanly separate tree size from epiphyte effects on bird count data, we conducted surveys only in old-growth forest sites with larg e trees, to hold tree size constant while allowing the amount of epiphytes to vary. These surveys were conducted in Guabn and Fundo San Martn forests. We defined plots of 25 m radius separated by at least 100 m from each other, and every plot was centered on a large tree over 1 m DBH; 32 plots were established in Guabn and 25 plots in Fundo San Martn. In each plot we censused birds for 10 min on two occasions. Values from the two surveys per plot were av eraged for analyses. Censuses were conducted during sunny mornings in February 2008 (Austral Summer). In each plot we marked a 50 m long and 4 m wide transect, oriented East to West across the plot di ameter. In each transect we measured the DBH, determined the tree species and visually assessed th e amount of epiphytes
105 (using two observers) in all trees >10 cm diam eter. This index represents the amount of 15 kg bags of epiphytes possible to be removed from each tree. Visual assessments are based on our experience removing epiphytes (Chapter 2). Comparisons among bird species richness and abundance in forest plots with different amounts of epiphytes were conducted using a Gene ral Linear Model on the Basal area of each plot (as surrogates of tree abundance), the inde x of epiphyte loads per plot, site, and bird abundance (individuals/day/plot) as the dependent variable with preliminary verification of the normality of the data with the Kolmogorov Smirn ov test. All statistical analyses were conducted in SPSS 11.0 (SPSS Inc. Chicago, Illinois). Results Preliminary Bird Community Surveys in Guabn and Chilo National Park We detected a total of 25 bird species, 22 speci es in Guabn forest and 19 bird species in the Chilo NP site (Table 4-1). Bird species ri chness and abundance changed with seasons (Fig. 4-2); both were higher in the Austral Spring and Summer (from Se ptember to March) than in the middle of the Austral Winter (June and July, Fig.4-2). Seasonal patterns were primarily attributable to the mi gration of Fo-fo ( Elaenia albiceps ), which was absent in winter. Bird abundances were generally similar in Guabn and Chilo NP, and dominated by omnivorous (consuming insects and fruits) species: pa rticularly by the Fo-fo, Picaflor chico ( Sephanoides sephaniodes ), and Rayadito ( Aphrastura spinicauda ) (Table 4-1). Twenty bird species (80% of the total) can use the forest canopy, of which 16 (80% of the 20 species) consume invertebrates, fruits and nectar, with 7 that feed mostly on in vertebrates (Table 4-1). The most abundant species in these forests were the insec tivorous and frugivorous Fo-fo ( Elaenia albiceps ) and the insectivorous Rayadito ( Aphrastura spinicauda ; Table 4-1).
106 Bird Species Visiting Trees With and Without Epiphytes Overall, 17 canopy species visited the trees with and without epiphytes (Table 4-2). The most abundant visitor was the Fo-fo ( Elaenia albiceps ) with 47% of all visits, followed by Rayadito ( Aphrastura spinicauda ) with 19% of all visi ts and Picaflor chico ( Sephanoides sephaniodes ) with the 9% of all visits (Table 4-2). Fo-fo was fr equently observed perching in the middle of the dense foliage. On a few occasio ns, Fo-fo was observed catching invertebrates in flight, but most of the time it was perched on small branches pecking for invertebrates hidden in the foliage. The second most frequent visitor, the Rayadito searched for invertebrates in the foliage actively inspecting small branches and leaves, many times upside down. Rayaditos walked on the bark searching for invertebrates hidden on the surface of limbs and small branches of the host trees, pecking in l eaves and bark. A few times Rayaditos perched directly on the epiphytes; my impression was that it was difficult for them to sta nd and walk on the soft surfaces of mosses and filmy ferns. The third main visitor, the Picaflor chico principally fed in red tubular flowers of bromeliads and vines, and was obser ved catching flying insects in the middle foliage of the tree. Picaflor was occasionally observe d sucking something from little holes on limbs, probably water captured in these holes. The ne xt most abundant species were the Jilguero ( Carduelis barbata ) and the Cometocino patagnico ( Phrygilus patagonicus ). In Guabn forest, a large flock of Jilgueros visited the experi mental trees during May 2006, consuming seeds of E. cordifolia that where maturing en masse at that time. Jilgueros usually move in large flocks, and were perching on the topmost branches of the tr ee, consuming the abundant seeds encapsulated in the dry fruits E. cordifolia trees. The other species, the Cometocino patagnico visited the focal trees regularly, feeding mostly on inverteb rates by pecking in the foliage, and they also walked over the epiphytes searching for food (pr obably preying on invertebrates). The remaining species presented very low visit rates. They in cluded granivores such as the parakeets Choroy
107 ( Enicognathus leptorynchus ) and Cachaa ( E. ferrugineus ), insectivores such as the Cachudito ( Anairetes parulus ) and predators such as the Traro ( Polyborus plancus ; Table 4-2). The number of bird species visiting trees w ith and without epiphytes did not differ for three of the four trees of pairs 1 and 2. Rarefact ion curves (Fig. 4-3) overlap their confidence intervals for three trees, indicati ng they have similar numbers of bi rd species. Only the tree with epiphytes of Pair 1 differed from the others, its curve becoming flat with a lower number of species (Fig.4-3). Rarefaction curves showed also that total number of visits during the whole survey period differed among treatme nts; trees with epiphytes rece ived between double or almost tree times more visits than trees without epiphy tes (Fig. 4-3), but the to tal number of species visiting the trees was not related to the presence or ab sence of epiphytes (Fig. 4-3, Table 4-2). Bird Foraging Activity in Relation to Epiphyte Loads Experimental versus control trees A monthly-based analysis on Pairs 1 and 2, c onsidering only birds that were foraging, showed that birds visited trees with epiphytes more frequently than trees without epiphytes (Repeated Measures ANOVA, F1, 2= 107.48 P = 0.009, Fig. 4-4). Analysis by species for the four dominant birds (Fio fio, Rayadito, Picaflor and Co metocino) showed that bi rds tended to visit the trees with epiphytes more often. The most fr equent species was Fo-fo, which was abundant during spring and summer. Fio-fo individuals visited trees with epiphytes more than the trees without epiphytes (Repea ted Measures ANOVA, F1, 2= 1784.1 P = 0.008, Fig. 4-5). For Rayadito, the tendency was similar, but with higher variab ility among pairs and only marginal significance (Repeated Measures ANOVA, F1, 2= 11.07 P = 0.08, Fig. 4-5). The Picaflor chico, despite its migratory habit over most of its distribution, was observed all year in Gu abn forests; it visited trees with epiphytes more often than trees without epiphytes, a lthough the difference was marginally significant (Repeated Measures ANOVA, F1, 2= 12.75 P = 0.07, Fig. 4-5). The fourth
108 most frequent visitor, the Come tocino, showed no differences in visit rate to trees with and without epiphytes (Repea ted Measures ANOVA, F1, 2= 2.9 P = 0.231, Fig. 4-5). Experimental pairs vs. pairs natura lly varying in epiphyte loads In both Experimental pairs (all epiphytes re moved from one tree) and Naturally different Pairs (those in which one tree naturally had fe wer epiphytes) birds vi sited more trees with epiphytes than trees without (or with fe wer) epiphytes (Repeated Measures ANOVA, F1, 4= 7.16 P = 0.055, Fig. 4-6). However, I found significant diffe rences between the types of pairs; more visits occurred to the Experime ntal pairs than to Naturally di fferent Pairs (Repeated Measures ANOVA, F1, 4= 11.30 P = 0.028, Fig. 4-6) but no interaction o ccurred between type of pair and treatment (Repeated Measures ANOVA, F1, 4= 3.7 P = 0.127, Fig. 4-6). Different bird species showed different responses. The Rayadito showed no effect of the type of pair (Repeat ed Measures ANOVA, F1, 4= 0.514 P = 0.513) but showed a significant effect of the treatment, visiting more frequently the trees with epiphytes (Repeated Measures ANOVA, F1, 4= 7,95 P = 0.048). On the other hand, Fo-fo showed marginal significance on the effect of the type of pair, visiting more frequently the tr ees of the Experimental pairs (Repeated Measures ANOVA, F1, 4= 5.5 P = 0.079), but showed no difference in number of visits to trees with versus without epiphytes (Repea ted Measures ANOVA, F1, 4= 3.7 P = 0.125). The other two most abundant bird species, Cometocino and Picaflor ch ico showed no significant relationship with type of pair or presence/absence of epiphytes (Repeated Measures ANOVA, F1, 4 < 2.4 P > 0.19). Foraging Substrata Use-Availability Analysis I assessed substrata availability only for Pair 1 and Pair 2 due to logist ic restrictions. Trees of Pair 1 and Pair 2 ranged between 1.22 m and 1.38 m in DBH, with heights of 25 to 30 m. In Pair 1, the control tree presented more foliar area and more branch and trunk area than the tree in which epiphytes were removed. In Pair 2, the two trees were considerably more similar in all
109 measured variables (Table 4-3). All trees were covered by a dens e layer of epiphytes, dominated by bromeliads and ferns (Chapter 2), and by a hemi-epiphytic tree: Raukaua laetevirens (Araliaceae). This last species typically colonizes large old trees in this area (Chapter 2), having between 4.9 and 36.6 m2 of bark surface, which represents between the 2% and the 20% of the total bark surface present in the host tree. In summary, in Pair 1, the control tree had more bark surface and more foliage than the tree with epiphytes removed, while in Pair 2 both trees were more similar in all variables measured, differi ng fundamentally in the epiphyte cover (Table 43). Most bird visits were observed on the tree foliage, and secondarily on bark or epiphytes (Fig. 4-7). However, different bird species s howed segregation in using different substrata available in the trees. For instance, over 80% of all Fo-fo were observed in tree foliage, while only the 46% of the Rayadito observed were usi ng the foliage, 36 % were observed on the bark of branches and 15% were observed directly on epiphytes. Picaflor chico was observed most frequently (53%) on epiphytes. In addition, the availability of folia ge in the focal trees is much larger than the availability of other substrata (Table 4-3). Comp aring habitat use versus habitat availability, I found that birds (all species pooled) were using at least one substratum in a higher proportion than what was expected by chance (Chi-Square > 59.9, df= 2, P < 0.001). I used confidence intervals to determine which substrata were more or less used by the five most common bird species observed in the trees. Substrate use varied by species; Fio-fo was non-selective, using all s ubstrates in relation to their availability, while Rayadito and Comesebo used foliage less than expected by chance, and used bark more than its availability on these trees (Table 4-4). Cometocino and Picaflor chico used epiphytes to a greater degree than predicted by
110 the abundance of epiphytes (Table 4-4). Data for the other species were insufficient to fulfill the assumptions needed for this analysis (Neu et al. 1974). Comparative Survey of Bird Community Struct ure in relation to Epiphyte Biomass Analyses of bird count data from the 25 m radius plots with va rying epiphyte loads detected a significant effect of epiphytes on can opy bird abundance (Table 4-5, Fig. 4-8). Plots with more epiphytes had greater abundances but (again) similar bird species richness, independent of the effect of tree size or tree composition (Table 4-5) Analyses did not detect site effects; the canopy bird community in both site s had similar species composition and relative abundance (Table 4-5). I also detected a signifi cant bias among epiphyte estimations by the two observers (Table 4-5); accounting for this source of error increased the effect of epiphytes on bird abundance (P=0.062 versus P= 0.015, error decrease from 513 to 483). However, the two points with highest epiphy tic biomass are the main drivers of this pattern. These two points showed greater abundan ce of the important species, Fo-fo, Picaflor and Rayadito, recording the s econd highest abundance of Fo -fo, the highest abundance of Picaflor and the highest abundance of Rayadito of all plots. Th e total number of individuals recorded on just one of these point s is higher than the number of a ll bird individuals recorded in any one of the other points. Removing these points from the analyses resulted in the effect of epiphytes on bird abundance becoming non-signifi cant. Analyzing by the dominant species, only the hummingbird was positively related with the abundance of epiphytes (GLM F1, 52 = 7.001, P= 0.011). For the rest of the species, the analysis did not s how significant relationships with epiphytes, but Rayadito was marginally related with basal area (GLM F1, 52 = 3.049, P= 0.087). One final result of interest wa s related with the variance in bird abundance. I organized the data from points with fewer epiphytes to points with more epiphytes, and calculated the variance
111 in ranges of five points. The pl ot of epiphyte biomass versus the variance in bird abundance showed a significant negative corr elation; in other words plots with more epiphytes had less variance in bird abundance (Pea rson Correlation Index R= 0.595, Bartlett Chi-square= 4.11, P= 0.043, Fig 4-9). In summary, these results showed that canopy bird abun dance increases with epiphytes at the plot level, but this relation is driven by th e two points with more abundant epiphytes, while results also showed that th e variance in bird abundance decreases when epiphytes abundance increase. Discussion Epiphyte Influences on Bird Co mmunity Structure at Two Scales My first hypothesis (bird individuals will forage more frequently in trees with epiphytes that without epiphytes) was tested at the scale of single trees in the epiphyte-removal experiment. I found significantly higher rate s of visitation by foraging birds to trees with epiphytes, particularly by the most abundant birds in these fo rests: the Fio-fo, the Ra yadito and the Picaflor (Fig. 4-4). The Fio-fo was observed much more of ten in trees with epiphytes, while Rayadito and Picaflor visitation patterns sugge sted only a marginal effect of epiphytes. These three bird species accounted for 50% of individuals (Table 41), and for 75% of all foraging visits to the trees (Table 4-2). These species are the dominan t species in the canopy of these forests, and because epiphyte loads positively affect their ab undance, epiphytes may have an influence on 50 % of all individual birds living in these forests. Fo-fo, Rayadito and Picaflor are common bird species across the distributional range of southern temperate rainforests (Rozzi et al. 1996, Anderson et al. 2000, Cornelius et al. 2000, Jaram illo 2005). Fo-fo and Picaflor chico use the vertical profile of the forest vegetation, and ar e very abundant in seconda ry forests, fragmented forests, Mediterranean forests and even city gardens (Willson et al. 1994, Estades and Temple 1999, Daz et al. 2002, Daz and Armesto 2003, Daz et al. 2005), while Rayadito is more
112 associated to areas with large old trees, including dry forests without epiphytes (Daz et al. 2002, Cornelius 2006). I previously determined that a pr incipal structural featur e of forests affecting bird abundance in Chilean rainforest is the pres ence of large trees (Daz et al. 2005). This study suggests that the epiphyte layer s upported by the larger trees is an important factor; once large trees become available, the epiphytes they suppor t, in turn, positively influence bird community structure. The second prediction (more bird species will forage in the tr ees with epiphytes than the trees without epiphytes) was not supported by the data. The thr ee most common species visited both control and treatment trees, wh ile other bird species showed little or no effect of epiphyte removal. For instance, most Cometocino individua ls visited trees with epiphytes (Table 4-2), however this species showed no significant a ssociation with epiphytes. Cometocino is a generalist that also lives in secondary forest, shrublands and even visits city gardens (Goodall et al. 1946). Similarly, few Zorzal ( Turdus falcklandii ) individuals were obs erved on epiphytes, but this species is more common in open areas, grasslands and in cities than in forests (Goodall et al. 1946, Daz and Armesto 2003, Armesto et al. 2005). Ot her species typical of these forests and associated with large old trees were the Comesebo ( Phygarrychas albogularis ), the Pito ( Colaptes pitius ) and the Magellanic Woodpecker ( Campephilus magellanicus ). They all feed on bark invertebrates, and woodpeck ers feed on larvae in the wood itself (Goodall et al. 1946). Our impression was that such species avoided epiphyte s. It is possible that epiphytes may be an obstacle for these species during foraging, in part because the soft plant layer limits their mobility (Ojeda 2004, V. Ojeda, pers coms, and my own observations). Only the Comesebo was observed feeding sometimes in small branches an d in the foliage, and occasionally on epiphytes.
113 This study is based on a field experiment, wh ich provided strong causal inference about the strength of the effect of epiphytes on bird visitation (James and McCulloch 1995). The limited sample size (3 pairs of experime ntal trees) is a constraint char acteristic of highly manipulative experiments such as this one involving heavy manual labor and tedious, dangerous, and timeintensive methods. Because of th e extreme nature of the expe rimental manipulation, however (clearing entire, mature trees of the majority of their epiphyte loads and comparing subsequent bird visits with comparable unmanipulated trees) because results were co nsistent across all three tree pairs (Fig. 4-4) and in agreement with results from the pairs that naturally differed in epiphytes (Fig. 4-5), I conclude that epiphytes can positively influence bird visitation, despite the statistical limitations of small sample size. The larger limitation is in st ating how generalizable this result may be across the range of the fore st biome; to other sp ecies of canopy trees; and across latitudinal gradients that exist in the region. This portion of the study was limited to northern Chilo Island, restricti ng the generality of the causal inference, but not the causal inference (James and McCulloch 1995). Overcoming this limitation was largely the motivation behind the comparative survey across plots with variable epiphyte loads. Effect of Epiphytes on Foraging Substrate Use The third prediction (birds will search for food more frequently in epiphytes than expected by the availability of epiphytes) was supported for particular speci es and not for other, because birds used substrata in different ways (Table 4-4). Rayadito and Come sebo showed preferences for bark, where they walk vertically searching for invertebrates, using their tails for support, similar to con-familial woodcreepers with sim ilar foraging niches. Comesebo also seems to select epiphytes, but it was alwa ys observed in the thinnest laye r of epiphytes or searching for invertebrates in the ferns attached directly to the bark, not in th e branches with dense epiphytic layers that are unlikely to s upport its style of locomotion. In contrast, Cometocino and Picaflor
114 chico showed clear selection for epiphytes (Table 4-3). However, Cometocino showed no difference visiting trees with a nd without epiphytes, because this species used the remaining epiphytes in the top-most branches. Picaflor vi sits were marginally higher in trees with epiphytes. Picaflor chico was frequently obser ved hunting insects flyi ng in the crown (using hover-gleaning), and was reliably seen fo raging at the red-tubular flowers of S. repens and F. bicolor the two most abundant epip hytes in the canopy (Chapter 2). Therefore, supported by other work (Smith-Ramrez 1993), it is likely that Picaflor chico is the main pollinator of flowering canopy epiphytes, visiting them more than expected by chance. In summary, morphological differences am ong bird species seem to contribute to segregation in the specific substrata used by bird s in the tree crown, wher e Fio-fo use substrata by its availability, Rayadito and Comesebo use the bark, and Picaflor use the epiphytes. However, I note that birds may use one substr atum more frequently than others without necessarily avoiding any substrata (Neu et al. 19 74, Aesbicher et al. 1993). Secondly, despite my calculations on substrata avai lability, I assumed that10 cm2 of bark was perceived the same as 10 cm2 of foliage by all birds; and if this is not the case, then my results could overor underestimate actual availability. In spite of this possibili ty, it appears that Rayadito and Comesebo showed a clear selection for bark, using it almost five times more than availability reflected (Table 4-4). Influence of Epiphytes on Bird Abundance at the Plot Scale The fourth prediction (birds will be more abunda nt in forest plots containing more epiphyte biomass than in plots with less epiphyte biom ass) was supported by the data. Birds were more abundant in plots with more epiphy tes than in plots with fewer ep iphytes, regardless of variation in the abundance of large trees, then epiphytes may contribute to sustain bird communities by increasing local abundance. However, these results were driven by the two points with higher
115 epiphyte loads. I did not consider those two points to be outliers because while they did represent extreme values for epiphyte loads that I could fi nd in the study area, these epiphyte loads do not misrepresent what can be found in the coastal fo rests of the region (J.J. Armesto, com pers). It was difficult to find complete plots with very high abundance of epiphytes but individual large trees with very high abundance of epiphytes were common. But birds clearly respond to this high abundance, in part, as is implied by the epiphyt e removal experiment. However, because they were only two points, I cannot conc lude that epiphytes have strong effects on birds at this scale in this study. The ideal situation would be to find points with similar tree size and composition but varying in epiphytes, points with low amount of epiphytes versus points with lots of epiphytes, but the effect of th e tree size will be always a di fficult variable to control. Influence of Epiphytes on Bird Species Richness My results showed no effect of epiphytes on th e number of bird species visiting trees, at either scale. Because epiphytes affected principa lly three bird species, they may not be directly relevant for the other canopy species and did not incr ease bird richness at eith er tree or plot scale. For forest birds in southern Chile, the main structural element known to influence species richness and abundance is large old trees (Daz et al. 2005). Results of this study lead to the conclusion that when large trees are abundant, avian species richness and abundance increase. Abundances of the three main canopy birds are in creased by the presence of epiphytes over and above the contribution from large trees alone. Possible Mechanisms Underlying Epiphyte-Bird Interactions Epiphytes support birds, but exac tly how epiphytes do that is not yet clear. While few birds selectively foraged in the epiphyt ic substratum, epiphytes were not consistently visited more often than other substrata (Table 4-4), and freque nt visitors such as Rayaditos seemed to avoid them, despite the food resources provided by epiphyt es, such as flowers a nd fruits (Chapter 3).
116 Only the Picaflor abundance was more strongly and positively related to epiphytes; twenty three percent of the visits of Picaflor chico were to the red-tubu lar flowers of the epiphytic vine Sarmienta repens and to the flowers of F. bicolor Epiphytic plants also pr ovide fleshy fruits for frugivorous species such as Fiofio and another 10 bird specie s (Armesto and Rozzi 1989). But flowering and fruiting do not e xplain the higher visitation rate to experimental trees with epiphytes year round, since at most times, bird s were observed preying on invertebrates (only one observation of Zorzal eat ing fruits was obtained). One interesting result is the decrease in th e variance of abundance among the points with low epiphytes versus the points wi th high epiphyte loads (Fig. 4-8) This decrease in variability suggests that large trees with epiphytes have an effect on regulating the abundance of birds, possibly stabilizing it. A possible explanation is epiphytes may pr ovide a permanent source of resources supporting bird abundance indirectly. In Chilo forests, my previous studies (Chapter 3) showed that epiphytes increa se the richness and abundance of i nvertebrates in the tree crown, in addition to those dwelling dire ctly in the epiphytic biomass. Then, birds may perceive trees with epiphytes as a more reliable source of re sources. In contrast, in the points with low epiphytes birds may search for food in a less pr edictive way and this could explain the high variation in bird abundance. If this proposition is correct, then the lack of epiphytes, or maybe the lack of large trees could not have an effect in total bird richness, but may increase the variability in resource distribution and then the variability in bird abundance. Most of the crown invertebrates collecte d were between 7 and 10 mm long: a size detectable by, and palatable to, small passerines. Van Bael and Brawn (2005) showed birds prey on invertebrates > 3 mm long, therefore the length of invertebrates cons idered in this study should be in the range of prey sizes for birds like Fo-fo, Rayadito a nd even Picaflor. These
117 results are in agreement with other studies that also showed that epiphytes contribute significantly to the abundance of invertebrates in tree cr owns (Ellwood and Foster 2004, Yanoviak et al. 2007). Conclusions: Relationships between Large Trees, Epiphytes and Birds My study yields two main contributions: Firs t, it provides eviden ce of indirect links between epiphytes and birds, probably by supply of invertebrate species. By supporting birds, epiphytes may also be supporting ecosystem func tions such as seed di spersal and pollination frequently provided by birds. In Chilo forests, birds are the principa l seed dispersers, in particular Fio-fo (Armesto and Rozzi 1989, W illson et al. 1996, Armesto et al. 2001). The hummingbird Picaflor chico is the only hummingbird of thes e forests and over 12 plants (including the epiphytes F. bicolor and S. repens ) may depend on it for pollination (SmithRamrez 1993). Birds can also reduce herbivory, su ch as occurs in other forests of the world (VanBael et al. 2003, Marquis and Whelan 1994, Greenberg et al. 2000). Evidence from Mazia et al. (2004) and an on-go ing study by Garibaldi et al. (2007) in similar Nothofagus forests of westernmost Argentina, based on bi rd exclusion experiments, showed that forest birds decrease foliar damage in native Nothofagus trees through consumption of herbivorous insects. Second, this study confirms the critical value of large trees for forest biodiversity conservation. From a conservation point of view, the large trees are really the focal element of the ecosystem in that they generate and support a diverse epiphytic laye r (Chapter 2), diverse epiphytic invertebrates (Chapter 3), and support both diversity of canopy birds (Daz et al. 2005) and their populations. Therefore, in these forest ecosystems, large trees represent a central priority for biodiversity conservation (sensu Be rg et al. 1994). In su mmary, the structure provided by large old trees support epiphyte com position and bird composition, while epiphytes also may support the most abundant bird species stabilizing their abundan ce. Their effect on the
118 functions provided by bird species is matter of further research, but based on the exposed results and previous studies, I hypothesize that the effect of birds in seed dispersal and invertebrate consumption should be significantly enhan ced by large trees with epiphytes.
119 Table 4-1. Bird species present in the study si tes, their abundance (ind ividuals/point/day, 14 months average for Guabn, 5 months av erage for Chilo NP), food habits and habitat use. Letter codes: GR=granivore, C=carnivore, IN=insectivore, FR=frugivore, NEC=nectarivore; VP=forest vertical profile, GS=grasslands, LT=large trees, UN=understory, SR=shrubs and understory. Common Abundance Food Habitat Family Species name Guabn Chilo NP habits use Columbidae Patagioenas araucana Torcaza 0.23 0.083 0.23 0.16 GR VP Falconidae Milvago chimango Tiuque 0.04 0.037 0 C, IN VP Fringillidae Phrygilus patagonicus Cometocino patagnico 0.17 0.057 0.48 14 GR, IN VP Carduelis barbata Jilguero 0.24 0.15 0 GR GS Furnariidae Aphrastura spinicauda Rayadito 1.37 0.098 1.23 0.19 IN LT Pygarrhychas albogularis Comesebo 0.20 0.045 0.05 0.05 IN LT Sylviorthorhync hus desmursii Colilarga 0 0.25 0.11 IN UN Icteridae Curaeus curaeus Tordo 0.07 0.038 0 IN, FR VP Muscicapidae Turdus falcklandii Zorzal 0.50 0.099 0.70 0.14 IN, FR VP Picidae Picoides lignarius Carpinterito 0.03 0.015 0.03 0.028 IN VP Campephilus magellanicus Carpintero negro 0 0.08 0.08 IN LT Colaptes pitius Pito 0 0.03 0.03 IN LT Psittacidae Enicognathus ferrugineus Cachaa 0.04 0.037 0 GR, FR LT Enicognathus leptorhynchus Choroy 0.23 0.12 0 GR, FR LT Rhinocryptidae Scelorchilus rubecula Chucao 0.68 0.091 1.28 0.29 IN UN
120 Table 4-1. Continued. Common Abundance Food Habitat Family Species name Guabn Chilo NP habits use Scytalopus magellanicus Churrn 0.20 0.048 0.73 0.12 IN UN Eugralla paradoxa Churrn de la Mocha 0.07 0.033 0.18 0.08 IN UN Pteroptochos tarnii Huet huet 0.61 0.10 0.38 0.09 IN UN Strigidae Glaucidium brasilianum Chuncho 0.02 0.019 0 C, IN VP Trochilidae Sephanoides sephaniodes Picaflor chico 1.43 0.11 1.30 0.27 NEC, IN VP Tyrannidae Anairetes parulus Cachudito 0.11 0.057 0.10 0.07 IN SR Xolmis pyrope Diucn 0.09 0.093 0.03 0.03 IN, FR VP Elaenia albiceps Fo-fo 1.12 0.32 1.13 0.49 IN, FR VP Colorhamphus parvirostris Viudita 0.04 0.038 0 IN, FR VP Hirundinidae Tachicyneta meyeni Golondrina 0.22 0.087 0.03 0.03 IN LT Total abundance 7.82 0.74 8.28 0.77 Total species 22 19
121 Table 4-2. Total bird visits for each species in th e pair of trees with and without epiphytes (by manual removal). Pairs 1 and 2 correspond to Eucryphia cordifolia trees, Pair 3 corresponds to Aextoxicon punctatum trees. Pair 1 Pair 2 Pair 3 Species With epiphytes Without epiphytes With epiphytes Without epiphytes With epiphytes Without epiphytes Anairetes parulus 0 1 5 2 0 0 Aphrastura spinicauda 106 58 85 51 6 6 Carduelis barbata 102 3 38 0 0 0 Colaptes pitius 0 0 1 0 0 0 Colorhamphus parvirostris 0 0 0 4 0 0 Columba araucana 3 0 2 3 0 0 Elaenia albiceps 244 113 166 73 4 11 Enicognathus ferrugineus 0 0 0 1 0 0 Enicognathus leptorhynchus 0 1 0 0 0 0 Phrygilus patagonicus 33 21 40 22 40 18 Phygarrychas albogularis 11 4 15 8 0 0 Picoides lignarius 0 1 1 1 0 0 Polyborus plancus 0 3 0 0 0 0 Sephanoides sephaniodes 51 16 59 31 14 15 Tachicyneta meyeni 0 0 10 0 1 7 Turdus falcklandii 8 0 11 1 3 2 Xolmis pyrope 2 9 9 0 0 0 Indet. 18 9 20 13 10 1 Total 578 239 462 212 78 60
122 Table 4-3. General structur al features of the large E. cordifolia trees sampled for epiphytes in Valdivian temperate rain forests, Chilo, Chile. Pair 1 Pair 2 With epiphytes Without epiphytes With epiphytes Without epiphytes DBH (m) 1.38 1.33 1.23 1.23 Height (m) 30 30 25 25 Bark area (m2) 213 145.4 98.4 105.6 Bark area of the hemi-epiphytic tree R. laetevirens (m2) 4.9 36.6 15.6 14.7 Foliage area (m2) 1545 499 327 364 Foliage of the hemipepiphytic tree R. laetevirens (m2) 109 200 202 85 Percent of the bark area covered by epiphytes 80% 20% 80% 20% Total bark surface (m2) 218 182 114 120 Total tree foliage area (m2) 1654 699 449 529 Total epiphytic area (m2) 174 36 91 24
123 Table 4-4. Analysis of substrata used by canopy bi rds versus substrata av ailable (Observations for tree Pairs 1 and 2 pooled). Values repres ent the proportion of visits recorded in each substratum (tree foliage, tree bark or epiphytic layer), plus the confidence interval in parenthesis) de fined. Expected values are th e theoretical proportion of birds that should be observed purely as a f unction of substrata availability (i.e., no preferences are implied). Avoid means bi rds visited the substrata less often than predicted, np means no preference, indica ting birds used the s ubstrata as predicted, and select means birds visited the substr ata more often than predicted. Data for individual trees did not differ from pooled data. Bird Species N Foliage Bark Epiphytes Foliage Bark Epiphytes Observed Comesebo 33 0.21 (0.06, 0.36) 0.42 (0.24, 0.61) 0.36 (0.19, 0.54) avoid select select Cometocino 61 0.48 (0.29, 0.66) 0.18 (0.04, 0.32) 0.34 (0.17, 0.52) avoid np select Fio-fio 501 0.85 (0.72, 0.98) 0.13 (0.01, 0.26) 0.01 (-0.03, 0.05) np np np Picaflor chico 125 0.34 (0.16, 0.51) 0.14 (0.01, 0.26) 0.53 (0.34, 0.71) avoid np select Rayadito 307 0.46 (0.28, 0.65) 0.39 (0.21, 0.57) 0.15 (0.02, 0.28) avoid select np Total 1222 0.67 (0.49, 0.84) 0.20 (0.05, 0.35) 0.13 (0.01, 0.26) avoid np np Expected 0.87 0.09 0.04
124 Table 4-5. Results of the General Linear M odel comparing the abundance of canopy birds in forest plots of 25m radius in Guabn and Fundo San Martn, southern Chile. Source Sum of squares Degrees of freedom Mean square F ratio P Basal area 0.124 10.1240.013 0.909 Epiphytes 59.274 159.2746.381 0.015 Sites 3.965 13.9650.427 0.516 Observer 30.051 130.0513.235 0.078 Error 483.069 529.29
125 Table 4-6. Abundance (Mean individuals/ plot/ da y) of birds in the 25 m plots on Guabn and Fundo San Martn, southern Chile. Family Species Common name Guabn Fundo San Martn Columbidae Patagioenas araucana Torcaza 0.08 0.04 0.28 0.07 Falconidae Milvago chimango Tiuque 0.05 0.04 0.02 0.02 Fringillidae Carduelis barbata Jilguero 0.09 0.04 0.06 0.04 Phrygilus patagonicus Cometocino 0.11 0.06 0.10 0.04 Furnariidae Aphrastura spinicauda Rayadito 0.77 0.13 0.68 0.13 Pygarrhychas albogularis Comesebo 0.11 0.05 0.18 0.07 Sylviorthorhynchus desmursii Colilarga 0 0.14 0.06 Hirudinidae Tachicyneta meyeni Golondrina 0.03 0.02 0.08 0.05 Icteridae Curaeus curaeus Tordo 0.02 0.02 0.06 0.03 Muscicapidae Turdus falcklandii Zorzal 0.16 0.07 0.16 0.06 Picidae Picoides lignarius Carpinterito 0 0.04 0.03 Picoides lignarius Pitio 0 0.02 0.02 Psittacidae Enicognatus ferrugineus Cachaa 0.05 0.05 0.08 0.05 Psittacidae Enicognatus leptorhychus Choroy 0.11 0.05 0.22 0.07 Rhinocryptidae Eugralla paradoxa Churrin de la mocha 0.03 0.02 0.30 0.10 Pteroptochos tarnii Hued hued 0.53 0.11 0.48 0.09 Scelorchilus rubecula Chucao 1.13 0.14 0.94 0.12 Scytalopus magellanicus Churrin del sur 0.28 0.06 0.38 0.08 Trochilidae Sephanoides sephaniodes Picaflor 1.47 0.10 1.30 0.08 Tyrannidae Anairetes parulus Cachudito 0.02 0.02 0.10 0.07 Elaenia albiceps Fio-fo 2.39 0.15 1.86 0.18 Xolmis pyrope Diucn 0 0.02 0.02 Total 7.45 0.35 7.54 0.30
126 Figure 4-1. Map of study sites (bla ck dots) in lowland coastal fo rests of Chilo Island, and Fundo San Martn, Valdivia, southern Chile. S ites are Guabn and Chilo National Park (NP). Region where the study was conducted are in dark color in the inset map.
127 Figure 4-2. Monthly variation in mean bird abun dance (individuals/point/day averaged over all points censused in each monthly survey; N= 14) and bird richness (total number of species detected per mont hly survey) in the Guabn forest site, Chilo Island, southern Chile (data collected between December 2005 and May 2008). 0 5 10 15 20De cember Jan u ar y Feb r u a ry Ap ri l M ay Ju n e Ju l y Se p t e mbe r O ct ober November Ja nu a r y F e brua ry March MayMonthsBirds/point/da y Bird abundance Total bird species
128 Figure 4-3. Rarefaction analysis on the number of species found in the trees with and without epiphytes and the number of bird visits. Ba rs indicate confidence intervals. (A) Trees in Pair 1, (B) trees in Pair 2. The subscrip t 1 indicates the tree without epiphytes, 2 indicates the tree with epiphytes. A 2 A 1 B 2 B 1 Number of bird visits Number of bird species
129 0 10 20 30 40 50 60De ce mbe r Jan u ar y February Ap ri l Ma y June Jul y S e ptemb e r October December Januar y February Marc h Ap ri lBird visits/hou r With epiphytes Without epiphytes Figure 4-4. Rate of bird visits (a verage) to experimental trees w ith and without epiphytes. Data from December 2005 to April 2007. Only bird s that were feeding were considered. Differences are significant (Repeated Measures ANOVA, F1,2 = 114.6, P= 0.009).
130 Figure 4-5. Rate of bird visits (a verage) plus/minus one standard er ror to experimental trees with and without epiphytes by the four most fr equent bird species. Data from December 2005 to April 2007. Only birds that were feeding were considered. Ra y adito Fio-fo Picaflo r Cometocino Summer Fall Winter S p rin g Summer
131 Figure 4-6. Total bird visits to the Experimental Pairs (Pairs 1 and 2) and Naturally different Pairs (Pairs 4 and 5). Only birds th at were feeding were considered.
132 Figure 4-7. Percentage of bird vi sits to each substrata by canopy birds in the Guabn forest site, Chilo Island, southern Chile (all data pooled). Only feed ing birds were considered. Epiphytes 13% Branches 18% Trunk 2% N= 1222
133 0 2 4 6 8 10 12 14 16 18 20 0102030405060 Epiphytes (# of bags)Birds (individuals/ point/ day Figure 4-8. Bird abundance as a function of the amount of epiphytes on 25 m radius plots in Guabn and Fundo San Martn forest.
134 0.0 5.0 10.0 15.0 20.0 25.0 0.010.020.030.040.050.060.0 Figure 4-9. Regression analysis be tween the variance (every five points) and epiphyte biomass in in Guabn and Fundo San Martn forests, southern Chile. Variance of bird abundance Epiphytes (number of bags)
135 CHAPTER 5 CONCLUSIONS: LINKING EPIPHYTES, IN VERTEBRATES, AND BIRDS IN THE CANOPY OF CHILEAN RAINFORESTS: TH EORETICAL ASPECTS AND THE VALUE OF LARGE OLD TREES Introduction In this chapter, I summarize the general c onclusions than link the large trees with the epiphytes, invertebrates and birds in the canopy of Chilean temperate rainforests. First, I presented current thinking about how structur e, species composition, and ecological functions can be viewed as connected attributes of ecosy stems, whose integrity, dynamics, and specific characteristics may, with further understandi ng, be used to better manage and protect biodiversity. While this conceptual framewor k, presented by Franklin et al. (1981) and Noss (1990) during the rise of conserva tion biology as a discipline has not yet coalesced into a theory, research that could contribute to potentially formaliz ing this framework into a theory is ongoing in many fields (such as ecosystem, community and food web ecology; Chapter 1). To make operative a theory requires (am ong other things; Pickett et al. 2007): i) testing its major causal hypotheses (linkages among biodivers ity and ecosystem functions), ii) building the conceptual content to facilitate explanation of its fundamental precepts (in th is case, links between structure, composition and function) in meaningful ways, ii i) production of exemplar s (case studies) that demonstrate conceptual and empirical frameworks for particular systems, and iv) generalization. My goal with this dissertation has been to contribu te in all four of these areas, but primarily in the third. In this dissertation I evaluated the li nks exist between compositional, structural, and functional ecosystem attributes. In chapters 2-4, I have presented evidence that large trees of long-lived species represents a physical structure that support large epiphyte loads including many species that, in turn, support diverse above ground faunas (invert ebrates and birds, composition). Several studies provide evidence th at birds in southern South America temperate
136 rainforests convey important functions, such as pollination (Smith-Ramrez and Armesto 1993), seed dispersal (Armesto and Rozzi 1989) and predation on herbivorous invertebrates reducing foliar damage (Mazia et al. 2004). Here, I analyz ed the links between structure, composition and discussed on potential ecological f unction described in chapters 2-4 in the context of Franklin et al.s (1981) original proposition. With this conceptu al framework I want to go toward a theory of Ecosystem Attributes, and I will discuss its application to understanding and managing biodiversity. I will do this by presenting a synt hetic hypothesis that highlights the fundamental and pervasive effects of large emergent ca nopy trees of old age in forest ecosystems. Significant Ecological Functi ons of Canopy Epiphytes One hypothesis that remains to be explored is that a rich ep iphyte layer contributes more than a poor epiphyte layer to soil forma tion, decomposition, photosynthesis and carbon sequestration, primary production, rainfall capture and transpiration, nut rient capture, nutrient fixation and habitat creation (F ranklin et al. 1981, Benzing 19 90, reviews in Benzing 1995, 2004, Nadkarni and Lowman 1995, Reynolds and H unter 2004, Fonte and Schowalter 2004, CruzAngn et al. 2005). We are beginning to describe many of these functions, mainly for tropical canopies, but they remain littleknown for the canopy of south-temp erate rainforest. Tejo et al. (in preparation) showed that ar boreal soils in Chilo forests pr esented similar physical features, mineralization and nitrification rates as the orga nic layer of the forest floor soil. Therefore, arboreal soil may represent a re levant source of nutrients to the forest ecosystem usually neglected in nutrient analysis for these forests (Prez et al. 2005). Additionally, Manuschevich et al. (in preparation) showed that epiphytes accumulate large am ounts of litter biomass from host trees, where a rhyzomorph fungus of the genus Marasmius tightly binds dead leaves falling from the host tree, incorporating them into the arboreal soil of the epiphytic layer. In my study sites, arboreal soil retains a high amount of water: up to 350 liters in each tree with epiphytes (Chapter
137 2). Moreover, in Chilo forests, the evergree n foliage and branches of the canopy, including epiphytes, intercept over 50% of rainfall (Daz et al. 2007). Epiphytes in the canopy of coastal forests of south-central Chile are also known to enhance moisture in the forest soil by contributing to water dripping a nd stemflow from f og interception, which often includes the release of nutrients from the fog (Weathers et al. 2000, Woda et al. 2006, del Val et al. 2006). In fact, measurements of nutrient inputs from stem flow in a southern Chil ean forest by Oyarzn et al. (1998) showed high flows of nitrogen from th e forest canopy to the ground layer, probably coming from the fog. Finally, preliminary assays of Carmona et al. (unpublished data) found potential N fixation in Pseudocyphellaria lichens obtained from the tree canopy (Chapter 2). In summary, ongoing and previous stud ies suggest that large trees with a well-developed epiphyte layer can be important in the wa ter cycle and in the nutrient suppl y to the forest ecosystem (Fig 5-1). Work in other forest ecosystems further document these and other functions of epiphytes (Dawson 1999, Hsu et al. 2002, Benner et al. 2007). One of my future goals is to continue documenting ecological functions in the canopy of Chilean forests for comparison with other, better-studied systems. How Structure Supports Diversity of Plants an d Animals in Chilean Rainforest Canopy: A Summary With time, rainforest trees grow tall, and tend to accumulate epiphytic material (Benzing 2004, Nadkarni et al. 2004b). Similarly, in my study sites, large, emergent E. cordifolia trees accumulate epiphytes as their diameter increases (C hapter 2). In turn, epiphytes capture litter and produce aerial humus, and when tree basa l diameter exceeds 1 m, individual E. cordifolia trees are normally colonized by the hemi-epiphytic tree R. laetevirens The presence of this hemiepiphyte can further increase the amount of ep iphytes and organic matter accumulated on the host tree (Pea et al. in preparation).
138 In Chapter 2, I showed that two species of large emergent trees support a high biomass of epiphytes of over 135 kg dry mass per tree, 70% of which is dead organic matter from litter and epiphyte decomposition, identified as aerial soil (Enloe et al. 2006). More than 40 plant species occurred in the epiphyte layer, and 35% of them were restricted to th e upper branches (tree crown). Consequently, large trees support massive quantities of liv e and dead organic matter that represent a structural element largely or completely absent from forests dominated by younger trees (Nadkarni et al. 2004b, Chapte r 2, personal observations). In th is context, physical structure provided only by large trees (i.e., enhanced surface area of bark suitable for supporting epiphytes) originates a new stru ctural component that increase s with time: a thick layer of epiphytes comprised of high live biomass, speci es richness (plants and lichens), and organic matter. This new layer supports several leve ls of trophic organization, primary production, detritivores, and consumers, each of them c onveying important ecological functions. In other words, the relatively simple structure of large trees supports epiphyte mats that, together, foster further biodiversity and ecol ogical functions (Fig. 5-1). Epiphytes provided significant habitat and re sources for invertebra tes (Chapter 3). The epiphytic layer held mostly detritivorous and pr edatory invertebrates, suggesting that the food chain in the epiphyte layer is based on detritus, not on green tissues (sensu Vanni and DeRuiter 1996; Chapter 2). Manuschevich et al. (in prep aration) indicated that litter accumulation and epiphytic growth on trees are the main resource supporting the detritivor e community. Epiphytes also increase the number of species and the ove rall abundance of invertebrates (measured as number of individuals or as tota l biomass) in the tree crown (Cha pter 2). Accordingly, epiphytes enhance invertebrate diversity by adding a detr itus-dependent community and more herbivorous and predatory individuals and species to the tree crown (Fig. 5-1).
139 Moreover, forest birds forage more frequently in trees with epiphytes than in trees without epiphytes (Chapter 4). Most canopy birds are insectivores, and the main species in the assemblage were Rayadito, Fio-fo and Picaflo r. It is likely, though yet undocumented, that insects supported directly and/ or indirectly by epiphytes can supply food resources for canopy birds, thereby increasing their local abundance and activity in the crown (C hapter 4). In addition, the observed pattern of increased bird visitation to single large trees agrees with the enhanced bird abundance observed in forest plots with hi gher abundance of epiphytes (Chapter 4). These results support the hypothesis th at epiphytes sustain populationlevel processes (numerical responses) that contribute to enhance bird spec ies abundance, not just local activity (Cruz-Angn and Greenberg 2005; Fig. 5-1). Linking tree-epiphyte structure and compositio n to the functional role of birds in the canopy Finally, birds of South American temperate ra inforests are well known for their important ecological roles in forest dynamics Birds are the principal seed dispersers and pollinators of many tree and epiphyte species, in particular th e Fio-fo (Armesto and Rozzi 1989, Willson et al. 1996, Armesto et al. 1996, Smith-Ramrez and Ar mesto 1998, Armesto et al. 2001). Birds are also likely the main insect consumers in the fo rest canopy, presumably influencing the levels of insect damage to tree foliage (Chapter 4). Studi es of VanBael et al. ( 2003), Marquis and Whelan (1994), and Greenberg et al. (2000) showed that birds prey on inve rtebrates in the forest foliage, reducing leaf damage in tropical and temperate forest canopies. A similar effect of canopy birds has been documented in Nothofagus forests of Argentina (Mazia et al. 2004, Garibaldi et al. 2007), where Fio-fo, Rayadito, Picaflor, Come sebo and Cometocino dominate bird species composition, as in my study sites. Then, I hypothe size that insectivorous birds can reduce canopy leaf damage by insects, although further work is needed to fully char acterize this function.
140 However, it is clear that if epiphytes support bird species richness a nd abundance in the canopy, then epiphytes can have positive indirect effects on the maintenance of foliage via insectivorous birds (Fig. 5-1). It is likely that epiphytes can assist other important ecological functions in the forest canopy, such as pollination and seed di spersal (Willson et al. 1996, Chapter 4, fig. 5-1). Toward an Integrative Theory Linking Co mposition, Structure and Function of Ecosystems The framework that links composition, structur e and function, provides an opportunity for conceptual integration in ecology and conservatio n biology. As an example, two concepts widely used in ecology are ecosystem engineers and ke ystone species (Jones et al. 1994, Power et al. 1996). These two concepts can be directly li nked within this framework proposed here. Ecosystem engineers are taxonomic species that cr eate physical structures or alter resource flux affecting other species. Keystone species, on the other hand, have disproportionate direct effects on population and community dynamics with respect to their own abundance (Power et al. 1996). For instance, keystone predators can positively or negatively affect ecosystem engineers, thereby altering structure and biodi versity. In the classic example of Estes et al. (1998) sea otters are keystone predators that consume he rbivores, i.e., sea urchins, and this favors the development of kelp forest whose structure s upports a diverse marine community that can disappear in the absence of the keystone predator In the previous example, the autogenic ecosystem engineer is the kelp forest (Hastings et al 2006) creating ecologica l structure that supports a highly diverse living community (structure support composition); a nd it is the keystone predator that favors the ecosystem engineer (kelp; i.e., composition s upports structure). A nother active area of integration linking composition and function center s on the debate about how diversity affects ecosystem stability (Naeem and Li 1997, McCa nn 2000). Much work is providing evidence of positive links in both directions (function suppo rting species, species supporting functions;
141 Mooney et al. 1996). However, structural comp onents of ecosystems few times have been included (Badano and Marquet 2008). My work has focused on the dominant autogenic engineers in forest ecosystems, i.e., large, old canopy trees, whose physi cal structure supports epiphytes, invertebrates and birds. In my study sites, epiphytes may s upport water and nutrient inputs, while birds can support pollination, s eed dispersion and reduce foliar damage by predating on herbivorous. Then, the structure pr ovided by an ecosystem engineer such as large old trees support species and f unctions at various levels. A you ng tree may not have the same ecological function as an adult tree because stru cture changes and new mass is created over time, as shown in this dissertation, and its effect on provide habitat and res ources to other species increase over time (Chapter 2). The role of a si ngle species changes during its development in the way it creates structure. Examples of this abound in the literature conforming the populationcommunity paradigm, but are largely isolated from the ecosystem function literature. For example, competition studies have shown how stru ctures such as a snags and live nesting trees are limiting bird population size and species richness (Newton 1994, Daz et al. 2005, Blanc and Walters 2008), providing opportunities for linki ng structure and composition to ecosystem functions. Ecosystem functions of cavity nesting species, however, have not been considered in any comprehensive manner. Woodpeckers are known to carry fungi that can spread in the snags and trees they feed and nest, thereby influenci ng snag decay rates, which in turn favors the creation of more nesting sites for birds (Jac kson and Jackson 2004). Woodpecker predation on boring beetles may limit outbreaks temporally and spatially (McCambridge and Knight 1972) thereby influencing tree demography and stand structure. I argue th at analytical operation within a purely population and community paradigm limits the comprehensive understanding of ecosystems, and thus impedes the maturation of both conservation and ecol ogical theory (Pickett
142 et al. 2007). Greater focus on building an inte grated view, such as the proposed Ecosystem Attribute approach, is not a great step beyond wh at is currently thought, but will greatly speed attainment of understanding. In th e following section I present an hypothesis, which is in part the result of this dissertation work and attempts to integrate composition, stru cture and function with life history attributes. In this hypothesis I attempt to foster th inking about ecosystem attribute theory in general, and provide a case study in the context of this theoretical background. The Treebeard Hypothesis Large old trees are a characteris tic element of undisturbed natu ral forests worldwide. Over evolutionary history, trees have grown as tall as incident phy sical disturbance frequency and mechanical constraints will allow (Koch et al 2004, Van Pelt and Sillett 2008). Trees become old, and fall in spatio-temporal patterns reflecting local distur bance regimes. Therefore, the stages of large live tree, senescen t tree, snag and log have been pa rt of tree life hi stories since the first trees appeared on land (som e 370 my ago; Dilcher et al. 2004). Forest organisms such as bark beetles or forest birds evolved in the presence of large old trees, whose dominance and longevity influenced forest dynamics in tune to disturbance regimes. Large old trees are characteristic elements of old-growth forests de scribed by explorers, such as Alfred R. Wallaces descriptions of old world tropical areas, and Charles Darwins description of Chilo Island as blanketed from coast to coast by a dense, impene trable forest, dominated by large trees (Wallace 1878, Willson and Armesto 1996). Here, I propose that large old trees support and provide services to the whole forest ecosystem and should be considered as a uniquely important structural elements, contributing disproportionately, relative to popula tion size, to forest integrity and functional resilience. I am calling this hypothesis the Treebea rd Hypothesis as a metaphor invoking an old grey-bearded and wise being (tree) that s upports the forest community in ways no younger trees possibly can,
143 in part, because they add structures not pr ovided by younger trees, and because they take up relatively few resources for themselves relative to the resources they indirectly garner for the forest ecosystem. Accordingly, th e hypothesis is that large tr ees offer and support ecological functions that benefit th e entire ecosystem in a disproportiona te magnitude relative to what the old tree requires for its own maintenance. Tree Life Cycle In forest ecosystems many tree species require logs or coarse woody debris as regeneration microsites (Franklin et al. 1981, Lusk 1995, Christie and Armesto 2003). These sites are legacies from previous trees that became old, died and fell on the forest soil, perhaps opening a canopy gap, and then favoring tree regeneration by increasi ng light and nutrient availability (Franklin et al. 1981, Lindenmayer and Franklin 2002). Coastal fo rests of South America often follow this regeneration cycle. Importantl y, many seeds of canopy trees require logs or coarse woody debris clumps to germinate and support regeneration und er canopy gaps (Christie and Armesto 2003) or in large open disturbed areas (Pap ic 2000). The requirement of nurse logs is absolute in many areas where the water table can be as high as to flood the rainfo rest floor during the wet season or after disturbance, because seeds will not ge rminate in standing wate r (Papic 2000). Growing trees require nutrients, water and light, which are supplied by the woody debris. Forest soil is formed by litter accumulation and decaying coarse woody debris from the previous generations of trees. Consequently, growing seedlings, sa plings and young trees are consuming resources provided by older, ancestral tree s (Fig. 5-3). In contrast, serv ices and resources provided by young trees to the forests (for inst ance litter flux, fog and rain wate r capture, or nitrogen fixation) are limited because these fluxes increase geomet rically with tree size and growing trees have high demands for nutrients and carbon (Fig. 5-3). With time, trees increase litter return to the forest floor, enhance soil formation, provide micr oand macrohabitats for wildlife, supply fruits
144 to seed dispersers, and their crowns capture more water from rain and fog to moderate hydrological conditions in the fore st during wet and dry periods. Year after year trees grow, developing an ever-larger epiphyte layer that enhances many of these services. Because services increase geometrically with tree size and/or age, at some time, compositional and functional attribut es of the tree raise non-linearl y. Aerial litter load supports a detritivore community that releases nutrients to the ecosystem, which fosters greater accumulations of biomass and diversity. In summa ry, trees create soil, habitats, resources and services and their supply raises with tree age. Older trees that pass a threshold in size may present higher amounts of foliage, supporting many more epiphytes and invertebrates than other trees, providing abundant micro-ha bitats, and thereby supporting di rectly and indirectly fauna that is absent from forests l acking old trees (Nadkarni et al 2004b, Daz et al. 2005, Chapter 2, 3 and 4). For mature trees, results from Chapter 2 s howed that aerial soil biomass is larger than the biomass of tree foliage, and green tis sues of epiphytes have a bioma ss equal or larger to that of the tree foliage. This means tree foliage is not necessarily the main reservoir of nutrients, or the major carbon fixing component in the tree, because epiphytes already have equal or greater biomass than the tree foliage. Large old trees require nutrients from the ground, but they also provide nutrients from rain and fog water captu re by epiphytes (via N fixation, litter production, interception, and storage, Weathers 1999, Sille t and Van Pelt 2007, Benner et al 2007, Kolher et al. 2007). Finally, old trees may require fewer re sources from the ecosystem than what they provide, including themselves and the epiphyte co mmunity they support. As trees become old, wood rots and releases nutrients th at support detritivores, trees start to lose branches that become transformed into soil detritus, foliage biomass declines whil e secondary growth continues,
145 producing an imbalance between energy de mand and carbon fixation. Because of these imbalances, trees collapse and finally fall or die standing (Van Pelt and Sillett 2008). Because of the biodiversity and processes they support, and because they tend to distribute nutrients in the form of dead material, I hypothesize that old trees at some point in their life cycle deliver more resources and services to the ecosystem than they retrieve. Snags and logs only provide nutrients to the ecosystem, without taking up any (Fig. 5-3) Therefore, at this point the role of trees switches from resource users and co mpetitors to facilitators, supporting regeneration (Fig. 5-4). The value of snags and logs as st ructures that support diversity and functional attributes of forests has for long been recogni zed (Franklin 1988, Lusk 1995, Kohm and Franklin 1997). The treebeard hypothesis takes one large step backward in th e tree life cycle and shines a searchlight on the mature living tr ee, particularly in those fore st systems where epiphytes are abundant. My work suggests that a forest without large mature trees is barely half a forest in terms of diversity and functional attribute that are lost with th eir removal. In summary, this hypothesis links tree life history wi th structural features, dive rsity and ecosystem functions, resulting from the transition of a tree fr om competitor to facilitator with age. Implications for Conservation Old trees have been a permanent structure of forest ecosystems for the last 370 my, and this study emphasizes the role of old trees in supporting biodiversity an d ecological functions. Conservation of old trees is fundamental for su staining forest biodivers ity, tree regeneration dynamics and ecosystem resilience. Based on my re sults, I hypothesize th at the loss of older trees from managed forests will lead to greater su sceptibility to shocks and disturbances such as herbivore outbreaks, drought and flooding because of: 1) reductions of bird activity and abundances necessary to prevent insects outbr eaks (Holling 1988), and 2) reduction of water regulation service (less effect of water logging on tree regenera tion in wet periods and greater
146 water supply through fog capture in dry periods) (D az et al. 2007). Remova l of large trees also represents a serious shock to forest ecosystems because of 3) depletion of the incident nutrient pool and its renewal. In this c ontext, managed forests become more similar to agricultural systems (forestry plantations), dominated sole ly by hungry, resource consumptive young trees, requiring fertilization, chemical pest control, and exhibiting magnified effects of both soil drought and waterlogging on tree surv ival. Therefore, in contrast to some forest management approaches, a minimum density of old trees s hould be maintained and some younger trees should be left to age. Areas of old-grow th forest should be preserved to maintain the whole potential of forest ecosystems (Kohm and Franklin 1997). In contrast, current management practices promoted by the Chilean government are based on the simplification of forest stands to single cohorts, harvested periodically (even-aged manage ment). Under this scenario, long-term forest management will degrade tree species composition, resilience, and the nutrient pool will be depleted. Future Research The Treebeard hypothesis can be tested most effectively in two ways: by comparing nutrient requirement and intake rates per unit of biomass versus soil nutrient supply in trees of different age under similar conditions, and compar ing the services provided by trees (such as herbivore damage control by birds) in old-grow th versus managed forests, with different densities of old trees (s tandardized by biomass). I predict th at in plots where large trees have been partially or totally removed we should de tect significant species loss, reduced rates of biomass nutrient accumulation, nutrient pool depleti on because of tree upta ke, and less resilience to insect outbreaks or climatic changes compared to control plots. Additio nal questions generated by this study include: 1) what is the ecological importance of dead organic matter accumulated in forests above (arboreal soils) and on the ground ? (Vanni and DeRuiter 1996, Butler et al. 2008);
147 2) Does aerial organic matter support significant canopy predators via a detritus-based food web? (Polis and Hurd 1996); 3) How important are the invertebrates dependent on detritus as food for vertebrates? (Vanni and DeRuiter 1996, Polis an d Strong 1996); 4) Have detritus-subsidized predators such as centipedes or spiders any si gnificant effect on herbi vory? (sensu Polis and Hurd 1996); 5) Canopy epiphytes hold significan t amounts of water in the tree crown; how important are old trees with hi gh epiphyte loads in re gulating ecosystem water cycles? (Dawson 1999); 6) How does rain interception and fog captu re cause differences in tree regeneration in successional and old-growth forests? (Daws on 1999, Daz et al. 2007); 7) How tolerant are forests with old trees vs. forests without old trees to disturbances such as herbivore outbreaks or drought? (Holling 1988, Ayres and Lombardero 2000). While large trees several centuries old have been dominant elements in the canopy of world forests for millions of years, presently most managed forests and forestry plantations are younger than 100 years old, and therefore are de pendent on fertilization, insecticides, and intensive management (Kohm and Franklin 1997). The productivity of re generating forests may depend on soil nutrients provided by earlier generations of old trees and it can be argued that present-day management by removing old forests may be depleting the potential for future growth. Forest management, accordingly, should try to maintain forest structure to mitigate the cost of growing new trees. Conserving and promo ting the persistence of old trees can foster the continued development of new ones.
148 FOREST ECOSYSTEM Crown invertebrate richness and abundance Bird species Large old tree Epiphytes Water storage Carbon storage Nutrient capture Nutrient fixation Epiphytic invertebrate richness and abundance Decomposers Predators Herbivores Predators Omnivores Herbivore control Seed dispersion Pollination (+) (+) (+) (+) (+) (-) (-) (+) (+) (+) (+) Figure 5-1. Multiple links among structure, divers ity and functions in the canopy of a temperate rainforests in Chilo, southern Chile. The structure provided by la rge, old trees has a positive effect (+) on species richness and abundance of epiphytes and crown invertebrates, thus supporti ng their main ecosystem functions (described beside each box). Epiphytes have positive effects on invertebrates that are the food of insectivorous birds. Birds in turn have nega tive effects on the inve rtebrates that they feed on. Such links among structure a nd composition support the functions each group conducts. As a result, large old trees have positive effects on forest ecosystems by supporting epiphytes, invertebrates and birds. Net effects of invertebrates were not included.
149 FUNCTION COMPOSITION STRUCTURE Large tree Epiphytic layer Water storage Nutrient supply Herbivory control Epiphytes Invertebrates Birds Figure 5-2. Relationships among structure, compos itional diversity and functions in temperate rainforests. Trees support ep iphytes, invertebrates and bird s, which provide important ecological functions.
150 Fallen tree (Log) Large old tree Adult tree Seedling Sapling Young tree Snag Nutrient supply Water supply Herbivory control Figure 5-3. Complete life cycle of a tree, from s eedling establishment (on fallen logs), to old age and after death as snag, and fallen tree. Th e central circle presents some specific functions for the forest ecosystem. Brown a rrows indicate the level at which trees in each life cycle stage support these functi ons. Green arrows indicate the degree to which the functions support trees in each st age of the life cycle. In this scheme, young trees depend strongly on ecosystem functions but they contribut e little to support these functions. In contrast, large old trees insure that these functions are fully operational, but depend less on these functions that they contribute to support. There must be a threshold age or size at which trees change from being supported by whole ecosystem functions to support these functions One value of large old trees therefore is to indirectly support younger trees by enhancing biodiversity that conveys ecosystem functions supporting tree regene ration and growth. Removing large trees from the ecosystem implies the removal of support systems for younger trees that have been crafted over evolutionary time.
151 (-) (+) Resources and services required Resources and services provided Seedling sapling young adult mature old snag log Tree age (-) (+) Figure 5-4. Treebeard Hypothesis. Th e horizontal axis repr esents the age (stage ) in the life cycle of a tree, the first vertical axis represents resources and services required by the tree (dotted line) per unit of bioma ss, while the second vertical axis represents resources and services to the ecosystem provided di rectly or indirectly by the tree per unit biomass (solid line). Young trees require mo re resources than th ey provide, but as trees age the ecosystem servic es are greater than their requirements. The epiphytic threshold, when trees accumulate epiphyt e biomass and diversity exponentially, is indicated by the arrow.
152 LIST OF REFERENCES Aesbicher, N. J., P. A. Robertson, and R. E. Kenward. 1993. Compositional analysis of habitat use from animal radio-track ing data. Ecology 74: 1313-1325. Amdgnato, C. 2003. Microhabitat distribution of forest grasshoppers in the Amazon. In: Arthropods of tropical forests (e ds. Basset Y., V. Novotny, S. E. Miller, and R. L. Kiching). Cambridge University Press, UK. Chapter 22, pp. 237-255. Anderson, C., and R. Rozzi. 2000. Bird assemblage in the southernmost forest of the world: methodological variations for determining spec ies composition. Anales del Instituto de la Patagonia (Chile), 28: 89-100. Aravena, J. C., M. Carmona, C. C. Prez, and J. J. Armesto. 2002. Changes in tree species richness, stand structure and soil properties in a successional chr onosequence of forest fragments in northern Chilo island, Chile. Revista Chilena de Historia Natural 75: 339-360. Arias, E. T., B. J. Richardson, and M. Elgue ta. 2008. The canopy beetle faunas of Gondwanan element trees in Chilean temperate rainfo rests. Journal of Biogeography 35: 914-925. Armesto, J. J., and R. Rozzi. 1989. Seed dispersa l syndromes in the rain forests of Chilo: evidence for the importance of biotic dispersa l in a temperate rain forest. Journal of Biogeography 16: 219-226. Armesto, J. J., I. Daz, C. Papic, and M. F. Willson. 2001. Seed rain of fleshy and dry propagules in different habitats in the temperate rainfo rests of Chilo island, Chile. Austral Ecology 26: 311320. Armesto, J. J., M. F. Willson, I. A. Daz, and S. Reid. 2005. Ecologa del paisaje rural de la Isla de Chilo: diversidad de especies de aves en fragmentos de bosques nativos. In: Historia, ecologa y biodiversidad de los bosques costeros de Chile (eds. C. Smith-Ramrez, J. J. Armesto, and C. Valdovinos). Editorial Universitaria, Santiago de Chile. Chapter 37, pp. 585-599. Armesto, J. J., P. Len-Lobos, and M. T. K. Arroyo. 1996. Los bosques templados del sur de Chile y Argentina: Una isla biogeogrfica. In: Ec ologa de los bosques nativos de Chile (eds. J.J. Armesto, C. Villagrn, and M. T. K. Arroyo). Edit orial Universitaria, Santiago de Chile. Chapter 1, pp. 23-28. Armesto, J. J., R. Rozzi, C. Smith-Ramrez, and M. T. K. Arroyo. 1998. Conservation targets in South American temperate fo rests. Science 282: 1271-1272. Arroyo, M. T. K., L. Cavieres, A. Pealoza, M. Riveros, and A. M. Faggi. 1996. Relaciones fitogeogrficas y patrones regionales de rique za de especies en la flora del bosque lluvioso templado de Sudamrica. In: Ecologa de los bos ques nativos de Chile (eds. J. J. Armesto, C. Villagrn, and M. T. K. Arroyo). Editorial Univer sitaria, Santiago de Chile. Chapter 4, pp. 71-99. Atmar, W., and B. D. Patterson. 1993. The measure of order and disorder in the distribution of species in fragmented habitat. Oecologia 96: 373.
153 Ayres, M. P., and M. J. Lombardero. 2000. Asse ssing the consequences of global change for forest disturbance from herbivores and pa thogens. Science Total Environment 262: 263-286 Badano, E. I., and P. A. Marquet. 2008. Ecosyste m engineering affects ecosystem functioning in high-Andean lanscapes. O ecologia 155: 821-829. Barbosa, O., and P. A. Marquet. 2002. Effects of forest fragmentation on the beetle assemblage at the relict forest of Fray Jo rge, Chile. Oecologia 132: 296-306. Barthlott, W., V. Schmit-Neuerburg, J. Nied er, and S. Engwald. 2001. Diversity and abundance of vascular epiphytes: a comparison of secondary vegetation and primary montane rain forest in the Venezuelan Andes. Plant Ecology 152: 145-156. Basset, Y., V. Novotny, S. E. Miller, and R. L. Kitching. 2003a. Methodological advances and limitations in canopy entomology. In: Arthropods of tropical forests: Spatio-temporal dynamics and resource use in the Canopy (eds. Y. Basset, V. Novotny, S. E. Miller, and R. L. Kitching). Cambridge University Press, UK. Chapter 2, pp. 7-16. Basset, Y., V. Novotny, S. E. Miller, and R. L. Kitching. 2003b. Conclusion: arthropods, canopies, and interpretable patterns In: Arthropods of tropical fo rests: Spatio-temporal dynamics and resource use in the Canopy (eds. Y. Basset, V. Novotny, S. E. Miller, and R. L. Kitching). Cambridge University Press, UK. Chapter 35, pp. 394-405. Bassett, Y., N. D. Springate, H. P. Aberlenc, and G. Delvare. 1997. A review of methods for sampling arthropods in tree canopies. In: Canopy Arth ropods (eds. N. E. Stork, J. Adis, and R. K. Diham). Chapman Hall, MA, USA. Chapter 2, pp. 27-52. Benner, J. W., S. Conroy, C. K. Lunch, N. Toyoda, and P. M. Vitousek. 2007. Phosphorus fertilization increases the abundance and nitrogenase activity of the cyanolichen Pseudocyphellaria crocata in Hawaiian montane forests. Biotropica 39: 400-405. Benzing, D. H. 1990. Vascular epiphytes. Camb ridge University Press, NY., USA. 354 pp. Benzing, D. H. 1995. Vascular epiphytes. In: Fore st canopies (eds. M. D. Lowman, and N. M. Nadkarni). Academic Press, San Diego CA, USA. Chapter 11, pp. 225-254. Benzing, D. H. 2004. Vascular epiphytes. In: Fore st Canopies (eds. M. D. Lowman, and H. B. Rinker). Elseiver Academic Pre ss, MA, USA. Chapter 9, pp. 175-211. Berg, ., B. Ehnstrm, L. Gustafsson, T. Ha llingbck, M. Jonsell, and J. Weslien. 1994. Threatened plant, animal, and fungus species in swedish forests: Di stribution and habitat associations. Conserva tion Biology. 8: 718-731. Blanc, L. A., and J. R. Walters. 2008. Cavity -nest webs in a longleaf pine ecosystem. The Condor 110: 80-92. Borkhataria, R. R., J. A. Collazo, and M. J. Groom. 2006. Additive effects of vertebrate predators on insects in a Puerto Rican coff ee plantation. Ecological Applications 16: 696-703.
154 Brokaw, N.V.L., and R. A. Lent. 1999. Vertical structure. In: Main taining biodiversity in forest ecosystems (ed. M. L. Hunter Jr.). Cambridge University Press, UK. Chapter 11, pp. 373-399. Burns, K. C., and J. Dawson. 2005. Patterns in th e diversity and distribution of epiphytes and vines in a New Zealand fore st. Austral Ecology 30: 891-899. Butler, J. L., N. J. Gotelli, and A. M. Ellison. 2008. Linking the brown and green: nutrient transformation and fate in the Sarracen ia microecosystem. Ecology 89: 898-904. Cardels, C. 2007. Vascular Epiphyte Comm unities in the Inner-Crown of Hyeronima alchorneoides and Lecythis ampla at La Selv a Biological Station, Costa Rica. Biotropica 39: 171-176. Cardels, C. L., and R. L. Chazdon. 2005. Inner-cr own microenvironments of two emergent tree species in a lowland wet fo rest. Biotropica 37: 238-244. Cardels, C. L., R. K. Colwell, and J. E. Watkins jr. 2006. Vascular epiphyte distribution patterns: explaining the midelevation richness peak. J ournal of Ecology 94: 144-156. Christie, D. A., and J. J. Armesto. 2003. Regenera tion microsites and tree sp ecies coexistence in temperate rain forests of Chilo Islan d, Chile. The Journal of Ecology 91: 776-784. Clement, J. L., M. W. Moffett, D. C. Shaw, A. Lara, D. Alarcn, and O. L. Larran. 2001. Crown structure and biodiversity in Fitzroya cupressoides the giant conifers of Alerce Andino National Park, Chile. Selbyana 22: 76-88. Collins, P. T. 1992. Length-biomass relationships for terrestrial Gastropoda and Oligochaeta. American Midland Naturalist. 128: 404-406. Cornelius, C. 2006. Genetic and demographic cons equences of human-driven landscape changes on bird populations: the case of Aphrastura spinicauda (Furnariidae) in the temperate rainforest of South America. Ph. D. Dissertation, University of Missouri, St. Louis. Cornelius, C., H. Cofr, and P. A. Marquet. 2000. Effects of habita t fragmentation on bird species in a relict temperat e forest in semiarid Chile. Conservation Biology 14: 534-543. Cruz-Angn, A., and R. Greenberg. 2005. Are ep iphytes important for birds in coffee plantations? An experimental assessme nt. Journal of Applied Ecology 42: 150-159. Cruz-Angn, A., T. S. Sillett, and R. Greenbe rg. 2008. An experimental study of habitat selection by birds in a coff ee plantation. Ecology 89: 921-927. CSIRO. 1991. The insects of Australia: a text book for students and research workers, 2nd ed. Carlton, Vic. Melbourne Univ ersity Press, Australia. Dawson, T. E. 1999. Fog in the California Redwood forest: ecosystem inputs and use by plants. Oecologia 117:476-485.
155 Del Val, E., J. J. Armesto, O. Barbosa, D. Christie A. Gutirrez, C. G. Jones, P. Marquet, and K. Weathers. 2006. Forest Islands in the Chilean Semiarid Region: Fog-dependency, Ecosystem Persistence and Tree Regene ration. Ecosystems 9: 598-608. Di Castri, F., Hajek., E., 1976. Bioclimatologa de Chile. Ediciones Universidad Catlica de Chile. Santiago. Daz, I. A., C. Sarmiento, L. Ulloa, R. Mo reira, R. Navia, E. Vliz, and C. Pea. 2002. Vertebrados terrestres de la Re serva Nacional Ro Clarillo, Chile central: representatividad y conservacin. Revista Chilena de Historia Natural 75, 433-448. Daz, I. A., J. J. Armesto, S. Reid, K. E. Sieving, and M. F. Willson. 2005. Linking forest structure and composition: avian diversity in successional forests of Chilo Island, Chile. Biological Conservation 123: 91-101. Daz, I.A., Armesto, J.J., 2003. La conservacin de las aves silvestres en ambientes urbanos de Santiago. Ambiente y Desarrollo 19, 31-38. Daz, M. F., S. Bigelow, and J. J. Armesto. 2007 Alteration of the hydrologi c cycle due to forest clearing and its consequences for rainforest succession. Forest Ecology and Management 244: 32-40. Dickinson, K. J. M., A. F. Mark, and B. Dawkins. 1993. Ecology of lianoid/epiphytic communities in coastal podocarp rain forest, Haast Ecological District, Ne w Zealand. Journal of Biogeography 20: 687-705. Dilcher, D. L., T. A. Lott, X. Wang, and Q. Wa ng. 2004. A history of forest canopies. In: Forest canopies (eds. M. D. Lowman, and H. B. Rinker). Chapter 6, pp. 118-137. Dunham A. E., and S. J. Beaupre. 1998. Ecol ogical experiments: Scale, phenomenology, mechanism, and the illusion of generality. In: Experimental ecology: I ssues and perspectives (eds. W. J. Resetarits, Jr., and J. Bernardo). Oxford University Press, NY, USA. Pp. 27-49. Ellwood M. D. F., and W. A. Foster. 2004. Doubling the estimate of invertebrate biomass in a rainforest canopy. Nature 429: 549-551. Ellyson, W. J. T, and S. C. Sillett. 2003. Epi phyte communities on Sitka Sp ruce in an old-growth Redwood forest. The Bryologist 106: 197-211. Enloe, H. A., R. C. Graham, and S. C. Sille tt. 2006. Arboreal histosol s in old-growth redwood forest canopies, Northern California. Soil Sc ience Society of Amer ica Journal 70: 408-418. Erwin, T. L. 1982. Tropical forests: Their richness in Coleoptera and other arthropod species. Coleopterist Bulletin 36: 74-75. Erwin, T. L. 1995. Measuring arthropod biodiversity in the tropical fore st canopy. In: Forest canopies (eds. M. D. Lowman and N. M. Nadkarni). Chapter 5, pp. 109-128.
156 Erwin, T. L. 2004. The biodiversity question: how many species of terrestrial arthropods are there? In: Forest Canopies (eds. M. F. Lo wman, and H. B. Rinker). Chapter 14, pp. 259-269. Espinosa, M. 1917. Los alerzales de Piuch. Bole tn del Museo Nacional de Historia Natural (Chile) 10: 1-83. Estades, C. F., and S. A. Temple. 1999. Deci duous-forest bird communities in a fragmented landscape dominated by exotic pine planta tions. Ecological App lications 9: 573-585. Estes, J. A., M. T. Tinker, T. M. Williams, and D. F. Doak. 1998. Killer whale predation on sea otters linking oceanic and nears hore ecosystems. Science 282: 473476. Fonte, S. J., and T. D. Schowalter. 2004. Deco mposition in forest canopies. In: Forest Canopies (eds. M. F. Lowman, and H. B. Rinker). Chapter 21, pp. 413-422. Franklin, J. F., K. Cromack Jr., W. Denison, A. McKee, C. Maser, J. Sedell, F. Swanson, and G. Juday. 1981. Ecological characteristic s of old-growth Douglas-fir fo rests. USDA Forest Service General Technical Report PNW-111. Freiberg, M., and E. Freiber g. 2000. Epiphyte diversity and biom ass in the canopy of lowland and montane forests in Ecuador. Journa l of Tropical Ecology 16: 673-688. Fuentes, J. E., S. Herrera, and R. G. Medel. 1996. Observaciones preliminares sobre uso de recursos y temperatura de actividad en ensamb les de hormigas granvoras del norte de Chile. Acta Entomolgica Chilena 20: 13-17. Garibaldi, L. A., T. Kitzberger, C. N. Mazia, and E. J. Chaneton. 2007. El control por recursos y depredadores sobre la folivora en bosques de Nothofagus pumilio. Abstract book, III Reunin Binacional de Ecologa, La Serena, Chile. Gentry, A. H., and C. Dodson. 1987. Contribution to nontrees to species richness of a tropical rain forests. Biotropica 19: 149-156. Gonzlez-Gmez, P., C. F. Estades, and J. A. Simonetti. 2006. Strengthened insectivory in a temperate fragmented forest. Oecologia 148: 137-143. Goodall J.D., A. W. Johnson, and R. A. Philippi 1946. Las aves de Chile, su conocimiento y sus costumbres. Platt Establecimientos Grfic os S.A., Buenos Aires, Argentina. 358 pp. Gotelli, N. J., and G. L. Entsminger. 2007. Ec oSim: Null models software for ecology. Version 6.0. Acquired Intelligence Inc. & Kesey-Bear http://homepages.together.net/~gentsmin/ecosim.htm. Gotelli, N. J., and R. K. Colwell. 2001. Quantif ying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecology Letters 4: 379-391. Greenberg, R., I. Perfecto, and S. M. Philpott. 2008. Agroforests as model systems for tropical ecology. Ecology 89: 913-914.
157 Greenberg, R., P. Bichier, A. Cruz-Angn, C. MacVean, R. Perez, and E. Cano. 2000. The impact of avian insectivory on arthropods and leaf damage in some Guatemalan coffee plantations. Ecology 81: 1750-1755. Grier, C. C., and N. M. Nadkarni. 1987. The role of epiphytes in the nutri ent cycles of two rain forest ecosystems. In: People and tropical forest : a research report from the United States Man and the Biosphere Program (eds. Luga. A. E.). Wash ington, D. C.: U.S. Dept. of State. pp. 28-30. Groom, M. J., G. K. Meffe, and C. R. Carroll 2006. Principles of cons ervation biology. Sinauer Associates, Sunderland, MS, USA. 793 pp. Gutirrez, A. G., J. C. Aravena, N. V. CarrascoFaras, D. A. Christie, M. Fuentes, and J. J. Armesto. 2008a. Gap-phase dynamics and coexistence of a long-lived pionee r and shade-tolerant tree species in the canopy of an old-growth coastal temperate rain forest of Chilo Island, Chile. Journal of Biogeography 35: 1674-1687. Gutirrez, A. G., O. Barbosa, D. A. Christie, E. Del-Val, H. A. Ewing, C. G. Jones, P. A. Marquet, K. C. Weathers and J. J. Armesto. 2008 b. Regeneration patterns and persistence of the fogdependent Fray Jorge forest in semiarid Chile during the past two centuries. Global Change Biology 14: 161-176. Hardy, R., X. Holsten, and E. Jackson. 1968. The acetyleneethylene assay for N2 fixation: laboratory and field evaluati on. Plant Physiology 43: 1185-1207. Hastings, A., J. E. Byers, J. A. Crooks, K. C uddington, C. Jones, J. G. Lambrinos, T. S. Talley, and W. G. Wilson. 2006. Ecosystem engineering in space and time. Ecology Letters 10: 153-164. Hedin, L. O., J. J. Armesto, and A. H. Johns on. 1995. Patterns of nutrien t loss from unpolluted, old-growth temperate forests: Evaluation of biogeochemical theory. Ecology 76: 493-509. Hietz, P., W. Wanek, R. Wania, and N. M. Na dkarni. 2002. Nitrogen-15 na tural abundance in a montane cloud forest canopy as an indicator of nitrogen cycling and epiphyte nutrition. Oecologia 131: 350-355. Hofstede, R. G. M., J. H. D. Wolf, and D. H. Benzing. 1993. Epiphytic biomass and nutrient status of a Colombian upper montane rain forest. Selbyana 14: 37-45. Hofstede, R. G. M., K. J. M. Dickinson, a nd A. F. Mark. 2001. Distribution, abundance and biomass of epiphyte-lianoid communities in a New Zealand lowland Nothofagus-podocarp temperate rain forest: tropical compar isons. Journal of Biogeography 28: 1033-1049. Holling, C. S. 1988. Temperate forest insect outb reaks, tropical deforestation and migratory birds. Memoirs of the Entomologi cal Society of Canada 146: 21-32 Hlscher, D., L. Khler, A. I. J. M. van Dijk, and L. A. (Sampurno) Bruijnzeel. 2004. The importance of epiphytes to total rainfall intercep tion by a tropical montane rain forest in Costa Rica. Journal of Hydrology 292: 308-322.
158 Hsu, C., F. Horng, and C. Kuo. 2002. Epiphyt e biomass and nutrien t capital of a moist subtropical forest in north-eastern Ta iwan. Journal of Tropical Ecology 18: 659-670. Ingram, S. W., and N. M. Nadkarni. 1993. Com position and distribution of epiphytic organic matter in a neotropical cloud forest Costa Rica. Biotropica 25: 370-383. Jackson, J. A., and B. J. S. Jackson. 2004. Ec ological relationships between fungi and woodpecker cavity sites. The Condor 106: 37-49. James, F. C., and C. E. McCulloch. 1995. The stre nght of inferences about causes of trends in populations. In: Ecology and management of neotr opical migratory birds (eds. T. E. Martin, and D. M. Finch), Oxford University Press, UK. Chapter 2, pp. 40-51. Jaramillo, A. 2005. Aves de Chile. Lynx Ediciones, Barcelona, Spain. Johansson, P., K. Rydin, and G. Thor. 2007. Tree age relationships with epiphytic lichen diversity and lichen life histor y traits on ash in southern Sweden. Ecoscience 14: 81-91. Jones, C. G., J. H. Lawton, and M. Shachak. 199 4. Organisms as ecosystem engineers. Oikos 69: 373-386. Koch, G. W., S. C. Sillett, G. M. Jennings, and S. D. Davis. 2004. The limits to tree height. Nature 428: 851-854. Kohm, K. A., and J. F. Franklin. 1997. Creating a forestry for the 21st century. Island Press, Washington DC, USA. 475 pp. Klher, L., C. Tobn, K. F. A. Frumau, and L. A. (Sampurno) Bruijnzeel. 2006. Biomass and water storage dynamics of epiphytic in old-growth and secondary montane cloud forest stands in Costa Rica. Plant Ecology 193: 171-184. Laman T. G. 2004. Strangler fig trees: demons or heroes of the canopy? In : Forest canopies (eds. M. D. Lowman, and H. B. Rinker). Elseiver Academic Press, MA, USA. Chapter 9, pp. 180182. Lindenmayer, D. B., and J. F. Franklin. 2002. Conserving forest biodiversity: A multiscaled comprehensive approach. Island Pr ess, Washington, USA. 351 pp. Lowman, M. D., and H. B. Rinker. 2004. Forest canopies. Elseiver Academic Press, MA. 517 pp. Lusk, C. H. 1995. Seed size, establ ishment sites and speci es coexistence in a Chilean rain-forest. Journal of Vegetation Science 6: 249-256. Mardones, M. 2005. La Cordillera de la Costa, caracterizacin fsico-ambiental y regiones morfoestructurales. In: Historia, biodiversidad y ecologa de los bosques co steros de Chile (eds. C. Smith-Ramrez, J. J. Armesto, and C. Valdovinos ). Editorial Universitaria, Santiago de Chile. Chapter 1, pp. 39-59.
159 Marquis, R. J., and C. J. Whelan. 1994. Insecti vorous birds increase growth of white oak through consumption of leaf-chewing insects. Ecology 75: 2007-2014. Marticorena, C., and M. Quezada. 1985. Catlogo de la flora vascular de Chile. Gayana 42: 1155. Martin, K., K. E. Aitken, and K. L. Wiebe. 2004. Nest sites and nest webs for cavity-nesting communities in interior British Columbia, Canada : Nest characteristics and niche partitioning. Condor 106: 5-19. Mazia, C.N., T. Kitzberger, and E. J. Chanet on. 2004. Interannual changes in folivory and bird insectivory along a natura l productivity gradient in norther n Patagonian forests. Ecography 27: 29-40. McCambridge, W. F., and F. B. Knight. 1972. Fact ors affecting spruce beetles during a small outbreak. Ecology 53: 830-839. McCann, K. S. 2000. The diversity-stability debate. Nature 405: 228-233. McCune, B. 1993. Gradients in epiphyte biomass in three Pseudotsuga-Tsuga forests of different ages in western Oregon and Wa shington. The Briologyst 96: 405-411. McCune, B., R. Rosentreter, J. M. Ponzetti, a nd D. C. Shaw. 2000. Epiphyte habitats in an old conifer forest in western Washington, U.S.A. The Briologyst 103: 417-427. Mitchell, A. W., K. Secoy, and T. Ja ckson. 2002. The global canopy handbook. Global Canopy Programme, UK. 248 pp. Moffett, M. W., and M. D. Lowman. 1995. Canopy access techniques. In: Forest canopies (eds. M. D. Lowman and N. M. Nadkarni). Acad emic Press, CA, USA. Chapter 1, pp. 3-26. Muoz, A. A., P. Chacn, F. Prez, E. S. Barnet and J. J. Armesto. 2004. Diversity and host tree preferences of vascular epiphytes and vines in a temperate rainfore st in southern Chile. Austral Ecology 51: 381-391. Murakami, M., and S. Nakano. 2002. Indirect effect of aquatic insect emergence on a terrestrial insect population by bird predat ion. Ecology Letters 5: 333-337. Nadkarni, N. M. 1981. Canopy roots: convergent evolut ion in rainforest nutr ient cycles. Science 214: 1023-1024. Nadkarni, N. M., and J. T. Longino. 1990. Invert ebrates in canopy and ground organic matter in a neotropical montane forest, Co sta Rica. Biotropica 22: 286-289. Nadkarni, N. M., and T. J. Matelson. 1989. Bird us e of epiphyte resources in Neotropical trees. Condor 91: 891-907.
160 Nadkarni, N. M., and T. J. Matelson. 1992. Bioma ss and nutrient dynamics of epiphytic litterfall in a Neotropical montane forest, Costa-Rica. Biotropica 24: 24-30. Nadkarni, N. M., D. Schaefer, T. J. Mate lson, and R. Solano. 2004b. Biomass and nutrient pools of canopy and terrestrial components in a primar y and a secondary montane cloud forest, Costa Rica. Forest Ecology an d Management 198: 223. Nadkarni, N., G. G. Parker, H. B. Rinker, and D. M. Jarzen. 2004a. The nature of forest canopies. In: Forest Canopies (eds. M. D. Lowman, and H. B. Parker). Elseiver Academic Press, MA, USA. Chapter 1, pp. 3-23. Naeem S., and S. B. Li. 1997. Biodiversity enhan ces ecosystem reliability. Nature 390: 507-509. Neu, C. W., C. R. Byers, and J. M. Peek. 1974. A technique for analysis of utilization availability data. The Journal of Wildlife Management 38: 541-545. Newton, I. 1994. The role of nest sites in limiting the numbers of holenesting birds: a review. Biological Conservation 70: 265-276. Nieder, J., J. Prosperi, and G. Michaloud. 2001. Epiphytes and their contribution to canopy diversity. Plant Ecology 153: 51-63, Noss, R. F. 1990. Indicators for monitoring biodiver sity A hierarchical approach. Conservation Biology 4: 335-364. Noss, R. F. 2006. Hierarchical indicators for monito ring changes in biodiversi ty. In: Principles of conservation biology (eds. M. J. Gr oom, G. K. Meffe, and C. R. Ca rroll). Sinauer Associater Inc. MA, USA. Chapter 2, pp. 28-29. Ojeda, V. S. 2004. Breeding biology and soci al behaviour of Ma gellanic Woodpeckers ( Campephilus magellanicus ) in Argentine Patagonia. European Journal of Wildlife Research 50: 18-24. Olson, D. M., E. Dinerstein, E. D. Wikramanayak e, N. D. Burgess, G. V. N. Powell, E. C. Underwood, J. A. D'amico, I. Itoua, H. E. Strand, J. C. Morrison, C. J. Loucks, T. F. Allnutt, T. H. Ricketts, Y. Kura, J. F. Lamoreux, W. W. Wettengel, P. Hedao, and K. R. Kassem. 2001. Ecoregions of the world: A new map of life on Earth. Bioscience 51: 933-938. Oyarzn, C. E., R. Godoy, A. de Schrijver, J. Staelens, and N. Lust. 2004. Water chemistry and nutrient budgets in an undisturbe d evergreen rainforest of southe rn Chile. Biogeochemistry 71: 107-123. Ozanne, C. M. P., D. Anhuf, S. L. Boulter, M. Ke ller, R. L. Kitching, C. Korner, F. C. Meinzer, A. W. Mitchell, T. Nakashizuka, P. L. Silva Dias N. E. Stork, S. J. Wright, and M. Yoshimura. 2003. Biodiversity meets the atmosphere: A global view of forest canopies. Science 301: 183186.
161 Papic, C., 2000. Regeneracin de plntulas arbreas sobre material leoso en descomposicin en un bosque sucesional de Chilo, Chile. MSc thesis Facultad de Ciencias Universidad de Chile, Santiago. Pea, M. P., N. V. Carrasco, I. A. Daz, M. E. Pea-Foxon, and C. Tejo. 2006. Colonizacin de la hemiepfita Pseudopanax laetevirens sobre Eucryphia cordifolia en un bosque valdiviano costero en la isla de Chilo. Abstract book, XLIX Reunin Anual de la Sociedad de Biologa de Chile. Pucn, Chile. Perakis, S. S., and L. O. Hedin. 2002. Nitrogen loss from unpolluted South American forests mainly via dissolved organic compounds. Nature 415: 416-419. Prez, C. A., R. Guevara, M. R. Carmona, a nd J. J. Armesto. 2005. Nitrogen mineralization in epiphytic soils of an old-growth Fitzroya cupressoides forest southern Chile. Ecoscience 12: 210-215. Pickett, S. T. A., J. Kolasa, and C. G. Jone s. 2007. Ecological Understanding: the nature of theory and the theory of nature. Elseiver Inc., MA, USA. 233 pp. Pike, L. H. 1978. The importance of epiphytic li chens in mineral cycling. The Bryologist 81: 247-257. Pike, L. H., W. C. Denison, D. M. Tracy, M. A. Sherwood, and F. M. Rhoades. 1975. Floristic survey of epiphytic lichens and bryophytes grow ing on old-growth conife rs in western Oregon. The Bryologist 78: 389-402. Polis, G. A., and D. R. Strong. 1996. Food web complexity and community dynamics. The American Naturalist 147: 813-846. Polis, G. A., and S. D. Hurd. 1996. Linking Mari ne and terrestrial food webs: allochthonous inputs from the ocean supports high secondary productivity on small islands and coastal land communities. American Naturalist 147: 396-423. Poulsen, B. O. 2002. Avian richness and abundance in temperate Danish forests: tree variables important to birds and thei r conservation. Biodiversity a nd Conservation 11: 1551-1566. Power, M. E., D. Tilman, J. A. Estes, B. A. Me nge, W. J. Bond, L. Scott Mills, G. Daily, J. C. Castilla, J. Lubchenco, and R. T. Paine. 1996. Cha llenges in the quest for keystones. BioScience 46: 609-620. Putz, F. E., and N. M. Holbrook. 1986. Notes on th e natural history of he mi-epiphytes. Selbyana 9: 61-69. Ralph, C. J., G. R. Geupel, P. Pyle, T. E. Martin, and D. F. De Sante. 1993. Handbook of field methods for monitoring landbirds. Pacific Southw est Research Station, Albany, California, USA. Reid, S., I. A. Daz, J. J. Armesto, and M. F. Willson. 2004. Importance of native bamboo for understory birds in Chilean temper ate forests. The Auk 121: 515-525.
162 Reynolds, B. C., and M. D. Hunter. 2004. Nutrient cycling. In: Forest canopies (eds. M. D. Lowman, and H. B. Parker). Elseiver Academic Press, MA, USA. Chapter 19, pp. 387-396. Rhoades, F. M. 1981. Biomass of epiphytic liche ns and bryophytes on Abies lasiocarpa on a Mt. Baker Lava Flow, Washingt on. The Bryologist 84: 39-47. Richardson, B. A. 2004. Tank bromeliads: faunal ecology. In: Forest Canopies (eds. M. D. Lowman, and H. B. Parker). Elseiver Academic Press, MA, USA. Chapter 9, pp. 195-198. Riveros, M., and C. Ramrez. 1978. Phytocenosis epiphytic vegetation in a woods farm (SanMartin, Valdivia Chile) belonging to Lapage rio-Aextoxiconetum associ ation. Acta Cientfica Venezolana 29: 163-169. Roberge, J. M., P. Angelstam, and M. A. Villard. 2008. Specialized woodpeckers and naturalness in hemiboreal forests Deriving quantitative targets fo r conservation planning. Biological Conservation 141: 997-1012. Rodrguez-Cabal, M. A., M. A. Aizen, and A. J. Novaro. 2007. Habitat fragmentation disrupts a plant-disperser mutualism in the temperate fore st of South America. Biological Conservation 139: 195-202. Rozzi, R., D. Martnez, M. F. Willson, a nd C. Sabag. 1996. La avifauna de los bosques templados de Sudamrica. In: Ecologa de los bo sques nativos de Chile (eds. J. J. Armesto, C. Villagrn, and M. T. K. Arroyo). Chapter 7, pp. 135. Sabag, C. 1993. El rol de las aves en la disper sin de semillas en un bosque templado secundario de Chilo. M. Sc. Thesis, Facultad de Ciencias, Universidad de Chile, Santiago. Sabo, J. L., J. L. Bastow, and M. E. Power. 2002. Length-mass relationships for adult aquatic and terrestrial invertebrates in a Californi a watershed. Journal of the North American Benthonical Society. 21: 336-343. Salinas, M. F. 2008. Diferenciacin de nichos ecolgicos de tres especies de Gesneriaceas epfitas del bosque templado del su r de Chile. Doctoral thesis, F acultad de Ciencias, Universidad de Chile, Santiago. Saunders D. A., R. J. Hobbs, and C. R. Mar gules. 1991. Biological consequences of ecosystem fragmentation: a review. C onservation Biology 5: 18-29. Schulze, C. H, and K. Fiedler. 2003. Vertical and temporal diversit y of a species-rich moth taxon in Borneo. In: Arthropods of tropical forests (eds Y. Basset, V. Novotny, S. E. Miller, and R. L. Kitching). Cambridge University Press, UK. Chapter 7, pp. 69-85. Sekercioglu, C. H. 2006. Increasing awareness of avian ecological functi on. Trends in Ecology and Evolution 21: 464-471.
163 Shaeffer, S. M., D. E. Anderson, S. P. Burns, R. K. Monson, J. Sun, and D. R. Bowling. 2008. Canopy structure and atmospheri c flows in relation to the 13C of respired CO2 in a subalpine coniferous forest. Agricultural and Forest Meteorology 148: 592-605. Sillet, T. S. 1994. Foraging ecol ogy of epiphyte searching insect ivorous birds in Costa Rica. Condor 96: 863-877. Sillett, S. C., and M. E. Antoine. 2004. Lichen s and bryophytes in forest canopies. In: Forest canopies (eds. M. D. Lowman, and H. B. Rinker). Elseiver Academic Press, MA, USA. Chapter 8, pp. 151-174. Sillett, S. C., and M. G. Bailey. 2003. Effects of tree crown structure on bi omass of the epiphytic fern Polypodium scouleri (Polypodiaceae) in redw ood forests. American Journal of Botany 90: 255-261. Smith-Ramrez, C. 1993. Hummingbirds and their flor al resources in temper ate forests of Chilo Island, Chile. Revista Chilena de Historia Natural 66: 65-73. Smith-Ramrez, C. 2004. The Chilean coastal rang e: a vanishing center of biodiversity and endemism in South American temperate rainfore sts. Biodiversity and Conservation 13: 373-393. Smith-Ramrez, C., and J. J. Armesto. 1998. Avia n nectarivory and pollination in Embothrium coccineum (Proteaceae) in temperate forests of southern Chile. Revista Chilena de Historia Natural 71: 51-63. Smith-Ramrez, C., J. J. Armesto and B. Saavedra. 2005. Estado del conocimiento y conservacin de los ecosistemas de la Cordillera de la Costa: sntesis y pe rspectivas. In: Historia, biodiversidad y ecologa de los bosques costeros de Chile (eds. C. Smith-Ramrez, J. J. Armesto, and C. Valdovinos). Editorial Universitari a, Santiago de Chile. Chapter 41, pp. 645-650. Smith, K. M., W. S. Keeton, T. M. Donovan, and B. Mitchell. 2008. Stand-level forest structure and avian habitat: Scale dependenc ies in predicting occurrence in a heterogeneous forest. Forest Sciences 54: 36-46. Solervicens, J. 1996. Consideracione s generales sobre los insectos. El estado de su conocimiento y las colecciones. In: Diversidad biolgica de Chile (eds. J. A. Simonetti, M. T. K. Arroyo, A. E. Spotorno, and E. Lozada). Chapter 27, pp. 198-210. Stork, N. E., J. Adis, and R. K. Didham. 1997. Canopy arthropods. Chapman & Hall. Strong, D. R. Jr. 1977. Epiphyte loads, tree falls, and perennial forest disruption: a mechanism for maintaining higher tree spec ies richness in the tropics wi thout animals. Journal of Biogeography 4: 215-218. Stuntz, S., C. Ziegler, U. Simon, and G. Zotz 2002. Diversity and structure of the artropod fauna within three canopy epiphyte spec ies in central Panama. Journal of Tropical Ecology 18: 161176.
164 Stuntz, S., U. Simon, and G. Zotz. 2003. Arthrop od seasonality in tree crowns with different epiphyte loads. In: Arthropods of tropical forests: spatio-tempor al dynamics and resource use in the canopy (eds. Basset Y., V. Novotny, S. E. Mille r and R. L. Kiching). Cambridge University Press, UK. Chapter 17, pp. 176-185. Tanner, E. V. J. 1980. Studies on the biomass and pr oductivity in a series of montane rain forests in Jamaica. Journal of Ecology 68: 573-588. Torras, O., and S. Saura. 2008. Effects of silvicultu ral treatments on forest biodiversity indicators in the Mediterranean. Forest Ec ology and Management 255: 3322-3330. Torres-Contreras, H. 2001. Biological background of ants found in Chile published in national and foreign scientific journals during the XX century. Revista Ch ilena de Historia Natural 74: 653-668. Van Bael, S. A., and J. D. Brawn. 2005. The direct and indirect effects of insectivory by birds in two contrasting Neotropical forests. Oecologia 143:106. Van Bael, S. A., J. D. Brawn, and S. K. Robinson. 2003. Birds defend trees from herbivores in a Neotropical forest Canopy. Proceedings of th e National Academy of Sciences USA 100: 83048307. Van Bael, S. A., S. M. Philpott, R. Greenberg, P. Bichier, N. A. Barber, K. A. Mooney, D. S. Gruner. 2008. Birds as predators in tropical agroforestry systems. Ecology 89: 928-934. Van Pelt, R. 2007. Identifying mature and old fo rests in western Wash ington. Washington State Department of Natural Resources, Olympia, WA. 104 pp. Van Pelt, R., and S. C. Sillett. 2008. Crown de velopment of coastal Pseudotsuga menziesii, including a conceptual m odel for tall conifers. Ecol ogical Monographs 78: 283-311. Van Pelt, R., S. C. Sillet, and N. M. Nadkarni. 2004. Quantifying and visualizing canopy structure in tall forests: methods and a case study. In: Forest canopies (eds. M. D. Lowman, and H. B. Rinker). Elseiver Academic Press, MA, USA. Chapter 3, pp. 49-72. Vanni, M. J., and P. C, DeRuiter. 1996. Detritus and nutrients in food webs. In: Food webs (eds. G. A. Polis, and K. O. Winemiller). Chapman & Hall, NY, USA. Chapter 1, pp. 25-29. Veblen, T. T. 1996. The ecology and biogeography of Nothofagus forests. Yale University Press, CT, USA. Villagrn, C. 1991. History of the temperate forests of southern Chile during the Late-glacial and Holocene. Revista Chilena de Historia Natural 64: 447-460. Villagrn, C. and J. J. Armesto. 2005. Fitogeografa histrica de la Cordillera de la Costa de Chile. In: Historia, biodiversida d y ecologa de los bosques coster os de Chile (eds. C. SmithRamrez, J. J. Armesto, and C. Valdovinos). Edit orial Universitaria, Santiago de Chile. Chapter 5, pp. 99-119.
165 Villagrn, C., and E. Barrera. 2002. Helechos de l archipilago de Chilo, Chile. Corporacin Nacional Forestal CONAF (Chile). 23 pp. Villagrn, C., and L. F. Hinojosa. 1997. Historia de los bosques del sur de Sudamrica, II: Anlisis fitogeogrfico. Revista Chile na de Historia Natural 70: 241-267. Wallace, A. R. 1878. Tropical Nature. Macmillan and Co., London, UK. 356 pp. Weathers, K. C. 1999. The importance of cloud a nd fog in the maintenance of ecosystems. Trends in Ecology and Evolution 14: 214-215. Weathers, K. C., G. M. Lovett, G. E. Likens, and N. F. M. Caraco. 2000. Cloudwater inputs of nitrogen to forest ecosystems in southern Chile : forms, fluxes, and sources. Ecosystems 3: 590595. Weisner, L. 2003. Cucao, tierra de soledades. Ril Editores, San tiago de Chile. 660 pp. Williams, C. B., and S. C. Sillett. 200 7. Epiphyte communities on redwood (Sequoia sempervirens) in northwestern California. The Bryologyst 110: 420-452. Willson M. F., C. Smith-Ramrez, C. Sabag, and J. F. Hernndez. 1996. Mutualismos entre plantas y animales en bosques templados de Chile. In: Ecologa de los bosques nativos de Chile. (eds. J. J. Armesto, C. Villagrn, and M. T. K. Arroyo). Editorial Universitaria, Santiago de Chile. Chapter 13, pp. 251-264. Willson, M. F., and J. J. Armesto. 1996. The natural history of Chilo: on Darwins trail. Revista Chilena de Historia Natural 69: 149-161. Willson, M. F., T. L. DeSanto, C. Sabag, and J. J. Armesto. 1994. Avian communities of fragmented south-temperate rainforests in Chile. Conservation Biology 8: 508-520. Wilson, E. O, and B. Hlldobler. 2005. The rise of the ants: a phyloge netic and ecological explanation. Proceedings of the Nati onal Academy of Sciences 102: 7411-7414. Wilson, E. O. 1992. The diversity of life. Harv ard University Press, Cambridge, MA, USA. Winchester, N. N., and V. Behan-Pelletier. 2003. Fauna of suspended soils in an Ongokea gore tree in Gabon. In: Arthro pods of tropical forests: spatio-temporal dynamics and resource use in the canopy (eds. Basset Y., V. Novotny, S. E. Mille r and R. L. Kiching). Cambridge University Press, UK. Chapter 10, pp. 102-109. Woda, C., A. Huber, and A. Dohrenbusch. 2006. Vegetacin epfita y captacin de neblina en bosques siempreverdes en la Cordillera Pelada, sur de Chile. Bosque 27: 231-240. Yanoviak, S. P., H. Walker, and N. M. Nadkarn i. 2004. Arthropod assemblages in vegetative vs. Humic portions of epiphyte mats in a neot ropical cloud forest. Pedobiologia 48: 51-58.
166 Yanoviak, S. P., N. M. Nadkarni, and J. C. Gering. 2003. Arthropods in epiphytes: a diversity component that is not effectively sampled by canopy fogging. Biodiversity and Conservation 12: 731-741. Yanoviak, S. P., N. M. Nadkarni, and R. Solano. 2007. Arthropod assemblages in epiphyte mats of Costa Rican Cloud Forests. Biotropica 36: 202-210. Zar, J. H. 1996. Biostatistical an alysis. Prentice-Hall, New Jersey. Zotz, G. 2005. Vascular epiphytes in the temp erate zones a review. Plant Ecology 176: 173183.
167 BIOGRAPHICAL SKETCH I was born in San Bernardo, a small town near Santiago de Chile, in the Austral Spring (November 1971). I grew up in a modest family, lis tening to stories about the countryside life, because my parents migrated to the city in the late 1960s. I remember stories about hunting, harvesting and even myths from my parents, gra ndparents and their brothers. I liked these stories because since I was child, I have been interest ed in natural history, ecology and conservation biology. I grew up in Santiago, looking at the Andes and the hills I wanted explore some day, as an adult. Later on, an important time in my life was as a secondary school student (1989), when I participated in a youth scientif ic club organized by the Nationa l Museum of Natural History (Chile). There I met Juan Carlos Torres-Mura, a Chilean ornithologist (M .Sc.), who, in addition to Pablo Espejo, an herpetologist of the Santia go Zoological Garden (M.Sc.), assisted me in conducting a preliminary study aimed at compari ng the terrestrial vert ebrate richness among three sites at Ro Clarillo National Reserve, ce ntral Chile, differing in accessibility to visitors. With this study I finally expl ored the mountains nearby Santia go, and participated in the XXI Chilean Annual Youth Scientific Fair (a scie ntific meeting for prim ary and secondary school students) organized by the Chil ean National Museum of Natura l History. After high school I worked in the herpetological exposition "Nir i Vilcn", property of Pablo Espejo (mentioned above) and MSc. Jos Navarro; in the Zoological Gardens of Sa ntiago Metropolitan Park. There, I learned about Chilean herpetol ogy, and also learned to recogni ze the different Chilean reptile species. In 1991, I began my undergraduate studies in biology at the University of Chile, exploring nature. Then, a very important part of my academic life was when I met Dr. Juan Armesto, who has been my friend and formally or informally my professor, giving me all the opportunities of continuing a scientif ic career, contacting me with other professors from foreign
168 universities, such as Dr. Katie Sieving, my friend and advisor at University of Florida. My economic situation has make me looking for j ob since the first years of my undergraduate studies, and without the comprehension and th e support from the grants of Dr. Armesto I couldnt continuing in the scientific career. Dr. Juan J. Armesto is also the president of Senda Darwin Foundation (FSD), a non governmental or ganization focused in conduct scientific research in the temperate rainforest of southern Chile, and link the scient ific background with the education and practical applicati on in local communities of southe rn Chile. I have been working in his laboratory and in FSD si nce 1995, during the final year of my undergraduate studies. I studied forest seed dispersal, seed predation, fore st bird habitat relationship and also participated in scientific meetings and activities of e nvironmental education. Also, between 1996 and 1999 I developed my masters degree at the University of Chile where he was my advisor. After that I started to study English, and I appl ied to a Fulbright fellowship to continue my studies in the US. Whit the support of Fulbright I started my Ph.D in the department of Wildlife Ecology and Conservation, advised by Dr. Sieving who has been a model of professor for me, for her integrity, professionalism, charisma and good will. W ith her advice, I started to think in study the forest canopy in Chile, her advices were key in I got funding, and now I am concluding this program and this part of my life, with my major gratitude to Univ ersity of Florida, to Fulbright program, to Canon National Parks Science Scholars Program, and especially to my advisor Katie E. Sieving, all the committee members, professors, workers and friends at University of Florida.