Diversity loss and function of leaf litter ants in ...

DIVERSITY AND FUNCTION OF LEAF LITTER ANTS IN COLOMBIAN COFFEE AGROECOSYSTEMS

by

Inge Armbrecht

A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Natural Resources and Environment) in The University of Michigan 2003

Doctoral Committee: Associate Professor Ivette Perfecto, Chair Professor Beverly Rathke Assistant Professor Emily Silverman Professor John Vandermeer

Dedicated to those lacking opportunities and social justice

ii

ACKNOWLEDGEMENTS This dissertation could not have been possible without the loving support of my family: my husband, Gerardo Peñaranda, our children Valentina, José and Inge Peñaranda Armbrecht, and my sister in law Inés Peñaranda. I will always be grateful to my parents, Rolf Armbrecht, and Lola Cabrera, brothers and sisters, and Lucy Duque for their care and learning from them throughout my life, as well as to my family in Chicago, Dick, Ariane and Alissa and my family in Germany. In honor of their memory, I heartly thank the encouragement I received from my beloved Jose Peñaranda and Vera Andler. A very special mention goes to my advisor Ivette Perfecto, who by far exceeded all my expectations, as advisor and as a human being. During all these years, she was able to balance understanding and encouragement, demand and stimulus, high academic quality with open mind. In addition, I would like to acknowledge all my committee members: Beverly Rathke, Emily Silverman and John Vandermeer, for their constructive feedback not only throughout the dissertation process but also in enlighting my understanding. This gratefulness extends to my friends at the University of Michigan at the Bluefields, Nwaeg and Ivettes-lab discussion groups as well as faculty and staff at the School of Natural Resources and Environment. I am also grateful to my colleagues and friends at Universidad del Valle (Cali), Biology Department, all of whom were essential for the success of this dissertation, especially my

iii

colleagues at the Entomology Section Patricia Chacón, Ranulfo González, Nancy Carrejo, James Montoya for supporting me from the very beginning. Many thanks to my students and research assistants María Cristina Gallego, Leonardo Rivera, Maria Teresa Albarracín, Gloria Vargas, Miguel Márquez, Gustavo Alvarez, Elizabeth Jiménez, Joaquín Colmenares and Germán Vargas. Colombian coffee farmers generously accepted us on their plantations to carry out the experiments without charging any costs. At Apía: Orgánica Tatamá, Norma Henao, Francisco Herrera and Oscar Díaz, Saulo Herrera, Horacio Muñoz, Hans Daniels, Carlos Mario Correa, and Patricia Montoya Jaramillo. At Chinchiná Colombia, I received support from the National Federation of Coffee Growers at Cenicafé, and very special thanks to Moisés Vélez, Francisco Posada and Alex Bustillo. Kind hosting was provided by Patricia Marín (Apía) and Everett Mayes (Ann Arbor). A Fulbright-LASPAU scholarship and Universidad del Valle award made it possible for me to study the Doctoral Program at the University of Michigan, who provided key economic support and stimulus through the Alvan Macauley Scholarship in 2000, the Aline Underhill Orten Foundation Scholar in 2001 from The Center for the Education for Women (CEW), Rackham Fellowship in 2003 and Research Assistanceships from Ivette Perfecto. The International Institute Fellowship provided additional funds for the first year of the Doctoral Program. The research was funded by The Land Institute at Kansas, Rackham Graduate School, The International Institute, Latin American and Caribbean Studies Program, University of Michigan, Universidad del Valle and the Natural Habitat Sciences Program of Colciencias, grant code 1106-12-11693.

iv

ABSTRACT

Diversity and function of leaf litter ants in Colombian coffee agroecosystems by Inge Armbrecht Chair: Ivette Perfecto The loss of biodiversity and associated functions in managed ecosystems is one of the most critical issues for conservation because it affects the long-term sustainability of the planet. This dissertation investigates causes and effects of agricultural intensification on the structure and function of ant assemblages in coffee plantations and forest patches of Andean Colombia. Coffee intensification significantly reduced vegetation complexity and increased management index values, as measured from 14 habitat variables. Using a leaf-litter sampling protocol I found that less intensified coffee farms harboured significantly higher ant species richness. Ant communities in polygeneric shaded coffee were significantly more similar to forest patches than any other management type. Additionally, significant positive associations among ant communities and the number of species involved in ant mosaics dropped sharply with intensification. I carried out nesting resource augmentation experiments to examine how the quantity and quality of nesting resources affect ant biodiversity. Ants responded quickly to the addition of nesting resources across the intensification gradient. In the quality resources experiment, the

v

diversity of ants nesting in multi-species groups of twigs was significantly higher than in monoespecific groups across all scales. These results suggest that bottom-up mechanisms are important for ant species loss throughout the technification process although other biotic and abiotic factors such as microhabitat and competition might also play a role. Finally, ant predation in shaded vs. unshaded coffee plantations was studied in the field and laboratory. Exclosure experiments demonstrated that ant communities prey on the coffee berry borer (Hypothenemus hampei) in both shaded and unshaded coffee plantations, but this function increased with shade in the wet season, when most infestation takes place. Ants were more effective predators in laboratory conditions than in the field. Predation in the field fluctuated between 2%27% of borer adults. My studies strongly support the conclusion that the structure and function of ant biodiversity are negatively affected by intensification of coffee agriculture. I recommend that immediate action be taken to stop the destruction of traditional coffee practices and prevent further biodiversity loss in hotspots of tropical biodiversity.

vi

PREFACE The intensification and industrialization of agriculture has led to low diversity monocultures that depend on high inputs of water and agrochemicals. Coffee growing is one of the most illustrative model systems showing how the intensification of agriculture negatively affects biodiversity. Different coffee plantation systems represent a gradient of intensification, in which the forest-like traditional coffee (grown under a canopy of native forest trees) represents the most biodiverse extreme. However, this shade system has increasingly been transformed into coffee plantations that require little or no shade (i.e. sun coffee) and high agrochemical inputs. Intensification of coffee has been shown to significantly reduce the associated biodiversity of arthropods, birds and other vertebrates (Perfecto and Armbrecht 2003). Furthermore, the strategic location of coffee plantations in regions of high diversity and endemism (Moguel and Toledo 1999) emphasizes the remarkable importance of this crop for conservation of biodiversity. Litter and ground-dwelling ants have been shown to be a useful group for biodiversity and conservation research for reasons such as their functional role in the soil community dynamics, high ubiquity, abundance and sensitivity (Agosti et al. 2000). By focusing on ants as a specific target group in this dissertation, I address three main aspects related to the intensification of coffee production, (1) changes in the structure of habitat and ant communities (ant mosaic) along the gradient of intensification, (2) nesting site limitation and quality of nesting resources as possible vii

causes of biodiversity loss and (3) changes in the function in terms of pest regulation of associated ant biodiversity along the intensification gradient. Chapter 1 reviews and discusses the current literature regarding the problem of intensification of agriculture and ecosystem function in agroecosystems. The role of ants in natural and managed ecosystems, in terms of their trophic functions is analyzed in the light of a recent controversy: whether ants are cryptic herbivores or top predators. Both of these views have experimental support in recent literature, and therefore may have profound implications on the perspective of using ants as effective pest control agents in agroecosystems. Chapter 2 addresses habitat changes in the coffee agroecosystem in an Andean Colombian region. The study involved the measurement of 14 habitat variables in 12 coffee farms subjectively classified along a gradient of intensification of coffee production, involving forest patches, polygeneric shaded coffee, monogeneric shaded coffee and unshaded coffee (sun coffee). A recently developed index of management was applied for this specific study area, along with univariate and multivariate analyses to assess whether these quantified measurements corresponded to the initial qualitative classification of the farms in the system gradient. Chapter 3 analyzes changes in ant communities and their associations along the same intensification gradient mentioned above in 16 coffee farms from two regions of Colombia. The importance of diversified shaded coffee plantations as reservoirs of some forest species is explored in terms of the ant species identities, i.e. the similarity between each of the coffee agroecosystems and forest patches of the

viii

region. Additionally, the study reports changes in ant species richness and in the associations-network of ant mosaics. Since shade trees are eliminated in the technification process and assuming that these trees are sources of nesting sites for ants, chapter 4 explores whether the amount of leaf-litter nesting resources is a factor explaining the loss of ant biodiversity. For this purpose, an experiment was established in 12 farms with four treatments involving the addition of empty bamboo twigs and leaf litter nesting sites. The number of ant colonies and ant species nesting in these resources are examined to determine their response to availability of nesting sites across the gradient of intensification. The diversity of nesting resources affecting the diversity of ant species nesting within them is examined in Chapter 5. A manipulative experiment was established in a shaded coffee plantation in which twigs belonging to eight species of shade trees were offered to ants as nesting resources. The only difference in the two experimental treatments was that half of the groups of twigs were “monospecific” (the group of twigs belonged to only one species of shade tree) and half of them were “diverse” groups (i.e. twigs belonging to eight species of shade trees). Monospecific groups also involved representatives of all of the tree species. This experiment focused on links between planned biodiversity (e.g. diversity of shade trees) and associated biodiversity (twig-nesting ants). Chapter 6 focuses on the function of ants as potential predators of insect pests in coffee plantations. The study explores the role of ground dwelling ants as predators of the coffee berry borer both inside infested coffee seeds and as free adults, and the

ix

effect of shade levels and seasonality on predation rates. The rationale behind testing this predation function is that infested coffee beans left by farmers on the soil after harvest are sources of infestation, and ground-dwelling ants might be playing important roles in the prevention of pest outbreaks in diversified coffee plantations. This assumption is based on the Natural Enemies hypothesis, in which diversified agroecosystems harbour a higher diversity of natural enemies thus explaining the lower impact of pests in polycultures. Overall, the data and experiments presented in this dissertation represent part of the picture regarding the consequences of agricultural management practices on some of the associated biodiversity and the resultant ecosystem services impacted.

References Agosti, D., J.D. Majer, L.E. Alonso, and T.R. Shultz (Editors). 2000. Ants: Standard methods for measuring and monitoring Biodiversity. Smithsonian Institution. Washington and London. Moguel, P. and V.M. Toledo. 1999. Biodiversity conservation in traditional coffee systems of Mexico. Conservation Biology 13:11-21. Perfecto, I. and I. Armbrecht. 2003. Technological change and biodiversity in the coffee agroecosystem of Northern Latin America (Chapter 6). Pages 159-194 In Vandermeer, J. (Editor) Tropical agroecosystems. CRC Press LLC. Boca Raton, Florida.

x

TABLE OF CONTENTS

DEDICATION ................................................................................................................ii ACKNOWLEDGMENTS.............................................................................................iii ABSTRACT .................................................................................................................... v PREFACE .....................................................................................................................vii LIST OF FIGURES......................................................................................................xii LIST OF TABLES........................................................................................................ xv LIST OF APPENDICES............................................................................................xvii CHAPTER 1. TROPICAL AGROFORESTS AND THE ECOLOGICAL ROLE OF ANTS .........................................................................1 CHAPTER 2. HABITAT CHANGES IN COLOMBIAN COFFEE FARMS UNDER INCREASING INTENSIFICATION .....................................30 CHAPTER 3. DIVERSITY LOSS, LEAF LITTER ANT MOSAIC AND INTENSIFICATION OF COFFEE PRODUCTION ................................63 CHAPTER 4. LIMITATION OF NESTING RESOURCES FOR ANTS IN COLOMBIAN COFFEE PLANTATIONS.......................................107 CHAPTER 5. COFFEE AGROECOSYSTEM: PLANNED BIODIVERSITY AFFECTS ASSOCIATED ANT BIODIVERSITY.............147 CHAPTER 6. ANT PREDATION ON THE COFFEE BERRY BORER HYPOTHENEMUS HAMPEI IN COLOMBIAN SHADED AND SUN COFFEE FARMS ..............................................................................184 SYNTHESIS AND CONCLUSIONS ........................................................................224

xi

LIST OF FIGURES

CHAPTER 2 Figure 1.

Total management index in each of the 12 coffee farms. .............................55

2.

Means and Standard Errors of three habitat variables at the “tree” level ..........................................................................................56

3

Means and Standard Errors of three habitat variables at the “coffee bushes” level. .........................................................................57

4.

Means and Standard error of one variable at the “soil level” .......................58

5

Cluster analysis output with data from all habitat variables .........................58

6

Principal Component Analysis output plot from all variables involved in the calculation of the Management Index ..................59

CHAPTER 3 Figure 1.

Location of the two study sites in Colombia and relative location of the farms within each study site.. ...............................................91

2.

Individual-based rarefaction curves for ant species from litter samples in four management types at Apía and Támesis .............................92

3.

Individual-based rarefaction curves for ant species from coffee bushes in four management types at Apía and Támesis ...............................93

4.

Bray-Curtis similarity index for leaf litter ants for each management type vs. forest patches and number of ant species shared with forests............................................................................94

5.

Frequency of most abundant ants along increasing intensification at Apía and Támesis..............................................................95

xii

6.

Simplification of the system and loss of the diversity of interactions at the leaf litter level in Apía municipality. ..........................97

7.

Network of statistically significant associations among leaf litter ants at Apía region.........................................................................98

8.

Territories of dominant ants in the management systems............................99

CHAPTER 4 Figure 1.

Location of the 12 farms at Apía municipality... ........................................138

2.

Average number of ant colonies per treatment subplot in each management type...............................................................139

3

Average number of ant morphospecies nesting at each of the treatment subplots and management types...............................140

4.

Jacknife-2 estimated number of litter-nesting ant species for each management type..............................................................141

5

Plot of total number of ant species nesting on the ground to management index in each of 12 farms ...............................142

6

Percentage of ant colonies in bamboo on the coffee bushes experiment in all farms ...................................................................143

7

Relative importance and identity of the most abundant ant species nesting in litter and coffee bushes ...........................................144

CHAPTER 5 Figure 1.

Spatial distribution of each of the Experimental Units in a 1ha plot.......................................................................................178

2.

Jacknife-2 estimator of species richness in the two diversity treatments: overall and discriminating over the two periods....................................................................................179

3.

Number of ant species per row in monospecific treatments and diverse treatment over both periods of time.......................180

4.

Observed relative to expected number of species of ants nesting in diverse treatments in both periods of time .................................181

xiii

5.

Accumulated total number of ant species nesting in the twigs of different species of shade trees ...........................................183

CHAPTER 6 Figure 1.

Picture of the “spiral trap”... .......................................................................215

2.

Field results with exclosure bags: number of adult berry borers in seeds with or without ants, both seasons............................216

3.

Spiral trap laboratory results: percentage of borers alive, dead or missing ...........................................................................................217

4.

Picture of an artificial nest with a colony of Solenopsis gr. brevicornis predating an adult borer. ..................................218

5.

Spiral traps field results: proportion of berry borer adults alive, dead or missing (predated) when exposed to ants.............................219

xiv

LIST OF TABLES

CHAPTER 2 Table 1.

Names and general characteristics of the twelve farms at the Apía Municipality, Risaralda, Colombia. ...........................................52

2.

Average values and statistical results for each of the habitat variables in nine coffee plantations and three forest fragments at Apía ......................53

3.

Standardized values for the variables included in the Management Index for each of the coffee farms and forests patches.................................54

CHAPTER 3 Table 1.

Basic description of sixteen farms at the Apía and Támesis Municipalities of Colombia............................................................88

2.

Nonparametric Incidence Coverage Estimator (ICE) for the number of ant species existing at coffee plantations of two regions in Colombia, Apía and Támesis............................................89

3.

Dominance indices (DI) for each ant species at Apía municipality..............90

CHAPTER 4 Table 1.

Description of 12 farms studied at the Apía municipality, Colombia (names, area, elevation).. .....................................133

2.

List of tree species found in each of the shade coffee plantations at Apía. . .......................................................................134

3.

Microhabitat and response variables for the litter nesting experiment.. ................................................................................................135

xv

4.

Results from the simplified Poisson regression for the number of ant nests and number of ant species.. . .....................................136

5.

Habitat variables and Management indices for each of the 12 farms....................................................................................137

CHAPTER 5 Table 1.

Shade tree species from which the twigs were obtained.............................172

2.

Results of response variables at different spatial scales in the two treatments and over the two periods of time........................................173

3.

Total colonies and number of exclusive ant species in two periods of time and two diversity treatments...................................174

4.

Number of ant species nesting in each of the species of tree and the diverse bags right next to them.. .........................................175

5.

Colony turnover between the first and second evaluation periods. ......................................................................................176

6.

Chi-Square goodness of fit testing nesting ratios in the eight tree species (categories) in two periods of time or the two treatments in the same period of time................................177

CHAPTER 6 Table 1.

Mann Whitney or t tests for laboratory: ants predation on coffee berry borer adults in infested coffee seeds.......................................210

2.

Field experiment results with exclusion bags: average number of adult borers in coffee seeds at each of the seasons, treatments and shade managements systems......................211

3.

Results from Spiral Traps in the field. Linear mixed-effects model fit by maximum likelihood ..............................................................212

4.

Average number borers in spiral traps exposed or not exposed to ants in wet and dry seasons in shaded and sun coffee farms. ..................................................................................213

5.

Number of ant workers found in the spiral traps in both seasons, management systems and farms. ................................................................214

xvi

LIST OF APPENDICES

CHAPTER 2 Appendix A.

Identities of tree species at each of the 12 coffee farms studied at Apía, Risaralda .............................................................................60

CHAPTER 3 Appendix A.

List and total abundance of each ant species at each farm in Apía municipality (leaf litter and coffee bushes) ................................................100

B.

List and total abundance of each ant species at each farm in Támesis municipality (leaf litter and coffee bushes) ................................................104

C.

Significant Chi-Square Yates corrected P values of any possible associations among ants in leaf litter at Apía .............................................106

CHAPTER 4 Appendix A.

Identities and abundance of nesting ant species found in both litter and coffee bushes.........................................................145

CHAPTER 6 Appendix A.

Average number of coffee berry borers counted in each of the bags, farms, coffee management systems and seasons.....................220

B.

Data collected from spiral traps in laboratory trials for each of three species of ants..................................................................221

xvii

CHAPTER 1

TROPICAL AGROFORESTS AND THE ECOLOGICAL ROLE OF ANTS

Introduction Sustainability and biodiversity: Human activity throughout the planet threatens to generate the most massive extinction ever known, and this threat increases with time (Paoletti et al. 1992, Pimm 1998, Lacher et al. 1999, Donald et al. 2000, Tilman et al. 2002). The way in which agricultural land is managed constitutes an important portion of the extinction problem and its possible solutions, since up to 95% of earth’s land surface is somehow influenced by human activity (Paoletti 1995, Pimentel et al. 1992, McNeely and Scherr 2003). The intensification of agriculture is the process of raising land productivity over time through increases in the inputs (Shriar 2000) and decreases in plant genetic diversity (Heller and Keoleian 2003). Intensification implies an underlying concept of viewing agricultural land as a factory instead of an ecological system in which constitutive elements have been molded by hundreds of millions of years of evolution (Vandermeer 2003). Overall, the intensification and industrialization of agriculture has led to monocultures depending on high inputs of water, fertilizer and pesticides ultimately causing biodiversity and productivity decline (Matson et al. 1997, Funes et al. 2002). The tropics, repositories of the majority of earth’s biodiversity, are particularly

30

susceptible to land degradation related to intensification (Siebert 2002, García-Barrios 2003). While proponents of intensification claim that it helps to reduce social pressure to clear remaining forest (Shriar 2000), some ecologists argue that intensification may actually cause the opposite result (Vandermeer 2003) partly because part of the profits are destined to acquire and exploit more land. It is also predicted that degradation of land, a result of intensification, will instead cause the expansion of the agricultural frontier (Fischer and Vasseur 2000) since farmers might abandon unproductive lands searching for better places to raise crops. As a consequence, remnant primary forest fragments will become increasingly isolated (Siebert 2002) and smaller. Agroecologists propose alternative agriculture as a possible solution (Vandermeer 1995), whereby agroecosystems more resemble natural ecosystems in the same biogeographic region. This model would increase the likelihood that agroecosystems will be sustainable, i.e. persisting productively for long periods of time without degrading local or regional resource bases (Gliessman 2001). In explaining why agroecosystems are important to biodiversity conservation, especially in the context of ants (Hymenoptera: Formicidae) as functional agents in agroforestry systems, it is worthwhile distinguishing two components of biodiversity within any agroecosystem: planned and associated biodiversity (Vandermeer et al. 1998). The kinds and proportions of plants a farmer decides to grow is the planned biodiversity (Vandermeer and Perfecto 1995). Associated biodiversity constitutes all other organisms that arrive or survive in the agroecosystems independently of the farmer’s plans (Vandermeer et al. 1998). Associated biodiversity increases non-linearly as planned

31

biodiversity increases (Vandermeer et al. 1998). Thus, agricultural intensification will result in a loss of associated biodiversity (e.g. Krooss and Schaefer 1998), although this loss might occur in different patterns (Swift et al. 1997, Perfecto et al. 2003). Ants are an important component of associated biodiversity in agroecosystems in terms of biomass and ecological function.

Intensification in tropical agroforests and the loss of biodiversity

Extensive research has focused on tropical multistrata agroforestry systems, especially shaded coffee and cacao (Muschler and Beer 2001) partly due to their potential to preserve associated biodiversity (Perfecto et al. 1996 in coffee, Reitsma et al. 2001 in cacao) and because they can be used as templates for sustainable agricultural production systems (Somarriba et al. 2001). Particularly, coffee agroecosystems have become a model system showing that agricultural intensification negatively affects biodiversity (Vandermeer and Perfecto 1997, 2000; Moguel and Toledo 1999, Perfecto and Armbrecht 2003). Coffee and cacao agroecosystems are grown across an intensification gradient, where rustic forest-like traditional agroforests (grown under a canopy of native forest trees) represents the most biodiverse extreme (Perfecto et al. 1996, Klein et al. 2002a). Shade trees ameliorate adverse climatic conditions, nutritional imbalances, extreme temperatures (Beer et al. 1998), and provide the farmers additional sources of income (Greenberg and Rice 2000). However, Rustic systems have been increasingly transformed to coffee (and to a certain extent cacao) plantations that require little or no shade (i.e. sun coffee) and high agrochemical inputs (fertilizers, insecticides, herbicides). Most studies show that the 32

intensification of coffee results in a loss of associated biodiversity (arthropods: IbarraNunez 1990, Perfecto et al. 1996, 1997, Mas 1999, Moguel and Toledo 1999, Rojas et al. 2001, Klein et al. 2002b, Perfecto et al. 2003; birds: Borrero 1986, Greenberg et al. 1997, Dietsch 2003; and other vertebrates: Gallina et al. 1992, Gallina et al. 1996, Witt 2001), with some affected species considered to be important biological control agents (Lachaud and García-Ballinas 1999, Ibarra-Núñez et al. 1995, Greenberg et al. 1997). Most farmers cite yield increases to justify intensification of coffee farms but maximum yield in this crop may occur between 40-50% cover (Soto-Pinto et al. 2000). In contrast, photosynthesis in cacao is greatest under 75% shade (25% of full sunlight, Willson 1999), perhaps explaining why cacao intensification has not been as extensive and intensive as in coffee. Cacao intensification nevertheless, negatively affects arthropod diversity (Klein et al. 2000b, Reitsma et al. 2001, Siebert 2002). A study in Sulawesi’s cacao (Klein et al. 2002a) showed that both abundance and richness of predators tended to decrease with intensification while phytophagous insects tended to increase even under 50% canopy cover.

The controversy of shade grown coffee: Despite scientific evidence showing the potential of diverse agroecosystems to maintain biodiversity, a controversy about the importance of promoting shade-grown coffee has recently emerged motivated by two main concerns. The first concern is that price premiums paid to shade coffee growers will encourage farmers to clear of tropical forest remnant understories, effectively converting forests to coffee plantations (Rappole et al. 2003). It is true that agroecosystems (even rustic ones) do not replace forests in terms of high quality habitat for forest biota 33

preservation (as argued by Rappole et al. 2001). However, it is difficult to explain why forest fragments in coffee growing areas would be at higher risk, since the reason these forest fragments still exist is the inaccessibility of their locations (usually very steep and broken topographies or high altitudes or sensitive watersheds) making it almost impossible to plant coffee or any other crop there. In other words, if the high incentives related to intensification did not cause forest destruction in the first place, is very unlikely that growing shade coffee will generate it. Additionally, farmers could receive “credit” or a premium for keeping forest fragments (Ecological Society of America, 88th meeting Savannah, mentioned in discussion at coffee workshop August 2, 2003). The second concern comes from the argument that habitat quality of agroecosystems is not as important for associated species richness as is the proximity to forest fragments (Ricketts et al. 2001). Ricketts and colleagues (2001) did not provide data to distinguish between agricultural habitats in their study. A wide range of degrees of shade and complexity exists in tropical agroecosystems, some of which might qualify as “better habitats” for forest species moving between forest fragments. This discussion has serious implications for conservation policy since the quality of the agricultural matrix could be determinant for the long-term survival and metapopulation dynamics of species inhabiting forest fragments (Vandermeer and Carvajal 2001, Perfecto and Vandermeer 2002). Therefore it is important to examine whether certain “natural” agroecosystems share more “forest like” associated species than others, and how ecological interactions are affected with intensification of agriculture.

34

Ants as a target group and mechanisms of ant diversity loss

Ants are highly ubiquitous, diverse, and abundant, in most habitats, and are considered to be a major fauna component (sometimes >80 % biomass) in tropical natural ecosystems (Holldobler and Wilson 1990). Ants are a major focus to testing hypotheses regarding species richness (Ward 2000, Campos et al. 2003, Kaspari et al. 2003), community dynamics and interactions (e.g. ant mosaic Leston 1973, trophic cascades Letourneau 1998, Dawes-Gromadzki 2002, assembly rules: Gotelli and Ellison 2002), eco-physiological hypotheses (Kaspari and Weiser 2000, Kaspari et al. 2000), mutualisms (Janzen 1973, Bronstein 1998, Alvarez et al. 2000, Vasconcelos and Davidson 2000) and invasions (Holway et al. 2002). The reasons for choosing ant assemblages as the focus taxon in agroecological research are: (1) the sensitivity of certain ant assemblages or guilds to habitat changes including agroecosystem condition (Peck et al. 1998, Floren et al. 2001, Watt et al. 2002, Robertson 2002, Andersen et al. 2002, Bruhl et al. 2003), (2) sensitivity to disturbance/rehabilitation (Goehring et al. 2002, Andersen et al. 2002), (3) relatively “straightforward taxonomy” (Alonso and Agosti 2000), and specifically, (4) their strong response in the context of the transformation from shade coffee to sun coffee, (Perfecto and Snelling 1995, Perfecto and Vandermeer 2002, Armbrecht and Perfecto 2003). Despite their sensitivity, ants are not necessarily the most responsive group of arthropods to habitat changes (Perfecto et al 2003) but are very useful for conservation assessment.

35

Ant diversity decreases with intensification in coffee (Table 6.1 in Perfecto and Armbrecht 2003) and abundance of ants increases with intensification in coffee and cacao (Klein 2002a: ant richness was not tested in this study). Mechanisms causing biodiversity loss in certain ant assemblages can be grouped in two factors: physiological-phylogenetic factors and ecological factors. Physiological-phylogenetic factors include: 1) microclimatic changes affecting forest-adapted species (e.g. Nestel and Dickschen 1990, Perfecto and Vandermeer 1996); 2) changes in temperature and moisture affecting litter ants highly susceptible to dehydratation (Kaspari 1993, Andersen 2000, Kaspari and Weiser 2000). Ecological factors include: 1) presence and interactions of the living components of the community, including trait mediated indirect interactions (Vandermeer et al. 2003); 2) lack of nesting sites (Perfecto and Vandermeer 1994, Roberts et al. 2000) many derived from plant tissues; 3) type of shade, which affect arthropods in general and therefore ants (Greenberg et al. 1997, Johnson 2000); 4) food availability, presumably related to microbial populations and food webs in litter (Kaspari et al. 2000); 5) invasion of exotic ant pests escaping their natural enemies (Feener 2000, Morrison 2000, Holway et al. 2002); 6) competition (Perfecto 1994, Nestel and Dickschen 1990); and 7) predation by army (or other) ants (Kaspari 1996b). In the context of the intensification of agroecosystems, the nesting limitation hypothesis is directly related to the loss of shade trees when traditional coffee farms are converted into sun coffee farms, since litter and twigs from shade trees are sources of nesting sites for ants and other organisms. The extent to which nest sites could be limiting to leaf-litter and ground foraging ants in tropical habitats is unclear because althought in some studies ants appear to be nest limited (Kaspari 1996a) and in others they do not

36

(Torres 1984, Kaspari 1996b). A study done in Mexico (Armbrecht and Perfecto 2003) found that the proportion of ant nests in twigs significantly decreased with the distance from the forest in a shaded monoculture coffee plantation but not in a neighboring shaded polyculture suggesting that twig-litter nesting ants are lost during the intensification process. Nevertheless, neither Armbrecht and Perfecto (2003) nor Torres (1984) included sun coffee plantations where nest-site limitation may be most severe for litter-twig nesting ant assemblages. Leaf-litter ants may be prevented from establishing in low-shade or sun coffee habitats due to either a lack of twigs or a lack diversity of twigs. Diversity of litter has been studied mainly in the context of natural ecosystem processes and functions related to decomposition (Blair et al. 1990, Hooper and Vitousek 1997, Altieri and Nicholls 1998, Wardle et al. 1999, Hector et al. 2000). How litter diversity affects macroinvertebrates (such as ants) remains obscure. Although ant diversity is positively related to plant diversity and plant structure in the tropics (Lavelle and Pashanasi 1989, Roth et al. 1994, Perfecto and Snelling 1995, Quiroz-Robledo and Valenzuela-González 1995, Bestelmeyer and Wiens 1996, Armbrecht and Ulloa-Chacon 1999), it is has not been clear whether litter diversity derived from this vegetation affects the litter-nesting ant assemblages. Chapter 4 and 5 of this dissertation found nest site limitation associated with of the diversity of twigs on the diversity of nesting ants.

The ecological roles of ants in the context of agroecology

Agroecology in the context of biodiversity and ecosystem function: The effects of biodiversity on ecosystem function is a topic that has generated intensive research and 37

controversy in recent years (Naeem et al. 1996, Tilman et al. 1997, Wardle 1997, Bengtsson 1998, Huston 1998, Lawton 1998, Cameron 2002, Zak in press). However, most of these studies focus on the effects of producers and soil microbial communities in different ecosystem functions such as primary productivity, stability and element cycling (e.g. Naeem et al. 1994, Wardle et al. 1999). Animals are largely neglected in those ecological experiments (Duffy 2002) and effects of diversity changes of consumers affecting ecosystem processes are unknown. Duffy (2002) has proposed several hypotheses of the effects of decreasing diversity: 1) Consumers often have much higher impacts (relative to their abundance) than producers; 2) top-down regulation effects could be severely affected in declining diversity contexts; and 3) prey diversity should reduce penetrance of trophic cascades and increase ecosystem stability. Despite the lack of experimentation in diversity-function links in ecology, agroecologists’ have long considered these links (Swift and Anderson 1993). Specifically, agroecologists have considered, for example, intercropping (production in terms of overyielding) (Vandermeer 1989, Vandermeer et al. 2002), competition vs. facilitation and complementarity in agroforesty (Beer et al. 1998), implicating the manipulation of different combinations of trees and crops and therefore diversity manipulations, and pest management (Rao et al. 2000, Schroth et al. 2000). The idea of enhancing pest management by increasing diversity in agroecosystems is not new. Dempster and Coaker (1974) reported at least 12 studies showing reduction of various insect pests in diversified systems due to increases of their predators.

The “associational resistance” occurs when a plant experiences less

herbivore attack for being in association with other genetically or taxonomically diverse

38

plants (Root 1973, Andow 1991). Two hypotheses explain the associational resistance phenomenon, the “enemies hypothesis” and the “resource concentration hypothesis” (Root in 1973). The “enemies hypothesis” states that since generalist and specialist natural enemies are more abundant in polycultures, they reduce herbivorous pests in polycultures more than in monocultures ones. The “resource concentration hypothesis” states that herbivores, especially monophagous, are more likely to find and remain on host plants that are concentrated, such that in a polyculture situation, a specialist pest is deterred from feeding because of reduced ability to find its host. Andow (1991) reviewed 209 studies under conditions of associational resistance, reporting that 52% showed lower herbivore population densities on plants in polycultures. Nevertheless, reduced pest and disease is not automatically achieved by increasing plant diversity in the tropical agroforestry systems since certain combinations of plants might favor diseases or pest populations (Schroth et al. 2000, Rao et al. 2000). The importance of using natural enemies for pest control and the worldwide movement toward an alternative agriculture (Vandermeer 1995, Altieri et al. 2003) took shape after Rachel Carson (1964) uncovered the problems inherent in the use of pesticides. Inherent in the problem of insecticides is the “pesticide treadmill” whereby insecticides induce pest resistance and reduce the pest’s natural enemies in the system, ironically encouraging the resurgence of the pest and the emergence of new pests (Van den Bosch 1978). Despite these facts, industrial agriculturalists apply five million tons of pesticides annually to crops worldwide (Matson et al. 1997). The real dimension of the pest control problem is fully appreciated considering that Arthropoda constitutes ~90% of all species in the tropics and also the great majority of biomass in terrestrial

39

ecosystems (Pimentel et al. 1992). Among arthropods, ants constitute between 32-86% (Fittkaw and Klinge 1973-Tobin 1995 respectively) of the faunal biomass of the rainforest. Canopy ants: top predators or cryptic herbivores?

Explaining the function of ants in ecosystems is highly important for agroecology because ants can be perceived by the farmers either as dangerous pests (due to their associations with sap-sucking insects or to their painful stings) or as beneficial predators of pests. Given the preponderance of ants in agroecosystems, understanding of their ecological role in natural ecosystems is key to proper management. In terms of the functional roles of ants in natural ecosystems, ecologists have raised questions regarding the mechanisms behind sustaining the huge biomass of ants in forest canopies throughout the tropics. There are two main non-mutually-exclusive hypotheses to explain how energy might flow through the trophic pyramid: 1) Ants foraging in the forest canopy behave mainly as cryptic herbivores feeding on trophobiont homopterans and extrafloral nectaries (Hunt 2003). Another hypothesis states that predator ants are responsible for a high turnover of canopy herbivores producing inverted trophic pyramids (Floren et al. 2002). There is experimental evidence for both explanations. Davidson and colleagues (2003) tested the first hypothesis by using nitrogen isotopes based on the principle that the lighter stable isotope (14N) is lost more readily through metabolic waste and thus the isotopic ratio (15N/14N) will increase with increasing trophic level. Ants from two tropical rain forests

40

(Peru and Brunei) were collected haphazardly from both leaf litter and foliage and their body contents were analyzed for these isotopes. The authors concluded that many arboreal ants species obtain little N through predation/scavenging and are thus behaving as arboreal exudate-foragers (i.e. cryptic herbivores). This conclusion, which might energetically explain why canopy ants defend three-dimensional territories (Hunt 2003) also implies that densities of trophobionts in the canopies are much higher than had been experimentally measured, but also that prey densities are likely to be lower than on the two-dimensional forest floor, which certainly contradicts the inverted trophic pyramid hypothesis. Bluthgen et al. (2000) found additional evidence for this cryptic herbivory hypothesis by searching for ants with homopterans or at extrafloral nectaries (EFN) high in the canopy of 66 trees in the Venezuelan Amazon. They found EFN to be rare in the tree canopy (0.08% tree genera) contrasting with 34 % tree species found in Barro Colorado, Panama (Schupp and Feener 1991). Curiously, and although 62 % of the tree (and 83 % of tree genera) had ant-tended homopterans, the number of ant species tending homopteran was much lower than that in EFN: only 16 as opposed to 52 ant species occurring in Phylodendron epiphyte’s EFN. Of ant tending homopterans, 74 % were tree-nesting ants that built carton nests (Azteca, Dolichoderus and Crematogaster species), suggesting that homopteran resources in the canopy are monopolized by a few arboreal ant species. Furthermore, almost all homopterans found in the canopy (up to 700 membracids and 3000 coccids/tree) were tended by ants (Bluthgen et al. 2000). Davidson et al. (2003) and Bluthgen et al. (2000) provide sufficient evidence that anthomopteran associations constitute a major driving force in trophic structure of rain

41

forest canopies, probably shaping many ant mosaics (Bluthgen et al. 2000, Dejean et al. 2000). Several questions still in queue are 1) whether most of the protein from prey in the canopy is used by ants to feed their larvae, thus not being incorporated in the tissue of the foraging ants such as those used in Davidson et al. (2003) nitrogen analyses, 2) The social gut of foraging ants used in Davidson et al.’s analyses could be full of honeydew sap explaining the similarity in isotopes rates of some ants and trophobionts, 3) whether or not all arthropods killed, harassed or removed by ants in the canopy are eaten (and incorporated to their tissues): part of them fall to the forest floor, where other compartments of the ant community could incorporate this protein, and 4) whether there is a seasonal change in which more homopterans are exploited (e.g. rainy season). Predation could be a dominant activity at certain times of the year, entied to the colony’s cycles and requirements of protein for reproductives and larvae production. These questions do not intend to undermine the importance of ant-homopteran relationships. For instance, 114 ant-homopteran associations were recently reported in Colombian coffee plantations (Franco et al. 2003) involving 30 ant species and 12 trophobiont homopteran species, most of them innocuous for the farmers and non of them being insurmountable pest problems. The second hypothesis, (inverse trophic pyramid) has been tested in tropical rain forests by Floren et al. (2002) and in a traditional coffee plantation by Vandermeer et al. (2003). In both studies lepidopteran larvae were offered to canopy ants attracted with baits. Floren et al. (2002) found that 85% of ant species killed caterpillar in nonmyrmecophyte trees (N=54 species) and samples from 69 fogged trees showed that more than two thirds of ants were predators. Additionally ants in Malyasian forests

42

represented 60% of all arthropods suggesting that ants might be responsible for the low abundance of less mobile arthropods of holometabolous taxa in these forests. Most importantly, Floren et al. (2002) found that predator ants were less abundant than nonpredator ants in strongly disturbed forests. A quick search in Brown’s (2000) trophically classified tropical ant genera (Table 5.1, pages 46-69) shows that 72 genera are predators, 31 generalist foragers, 12 fungi growers, six homopteran tenders, three seed harvesters, three nectary harvesters, two pollen eaters, two scavengers and one parasite. According to this general trophic classification, there are more predator ant genera in the tropics than all other trophic guilds added together. In Davidson et al. (2003) 92% of all Formicinae tested belonged to Camponotus and 95% of Dolichoderinae tested were Azteca and Dolichoderus, both of which are generalized foragers (Brown 2000), although Azteca also visits extrafloral nectaries and tends homopterans. These two subfamilies represented 43% of all species evaluated in Perú rainforest, suggesting that their samples were somewhat biased toward generalist species. My interpretation is that Davidson et al. (2003) demonstrated that sap-feeding (exudates-dependent) ants have nitrogen isotope ranges similar to those of arthropod herbivores, but not that all canopy ants behave as cryptic herbivores. Only 36 of species analyzed out of 108 ant species matched isotopic ranges with that of sapfeeding trophobionts in Peru (44 matched predator arthropods) and 15 species of 84 in Brunei (26 predator matching). Nevertheless, it is also true that a few species of canopy ant species usually constitute the bulk of the biomass in each system (e.g. Armbrecht et al. 2001) and an accurate assessment of their diet should be done before drawing final conclusions about their role in a certain tropical food web. Certainly, ants play

43

important roles both as cryptic herbivores and predators, and this fact constitutes a paramount challenge for pest management involving ants in tropical agroforests.

Potential of ants for pest control

A recent meta-analysis found that generalist predators (single species and multiple species assemblages), were effective control agents in agroecosystems showing significant reductions of pest abundance (79 % of studies), and significant increase in yield (or reduction in damage, 65% of studies) (Symondson et al. 2002). Since many tropical ants are considered to be predators, scavengers and generalist foragers (see guild organization in Delabie et al. 2000, Holldobler and Wilson 1990) ants are obvious candidates for use in biological control of pests. Specifically, ants are biological control agents in agroforests reducing pests and diseases as for instance in Ghana (Majer 1972, Leston 1973, Majer 1976), Asia (Way and Khoo 1992, Perfecto and Castiñeiras 1998) and Latin America (Delabie 1990, Perfecto 1990, 1991, Medeiro et al. 1995). Ants (i.e. Dolichoderus toracicus, Oecophylla smaragdina) are also an important part of management programs in agroforests of Vietnam (Van Mele and Cuc 2001; Van Mele and Van Lenteren 2002), and even with homoptera tending trophobionts, the presence of ants might benefit the farmer (Van Mele 2002 pers. com.). Way and Khoo (1992) discussed the valuable ecological features that make predatory ants beneficial for pest management (e.g. stable foraging places, specialists and generalist predators) and considered some evidence favoring crop mixtures to encourage beneficial ants in crops such as coconut, palms and cacao. They even

44

considered ant-homoptera relations beneficial because it attaches predatory ants to certain foraging places. It is also speculated that a more sweet honeydew (induced by ant foraging) might attract more parasitoids (Paul Van Mele, pers. com. 2002). Roberts et al. (2000), in a gradient of coffee management systems, showed that all habitats except sun coffee supported numerous army ant swarms and thus the diverse assemblage of 126 species of ant-following birds. If birds were causally linked to the reduction of herbivory, a direct connection between the diversity of a crop and the beneficial associated fauna would therefore be established. However, caution is needed to evaluate the extent to which ants can be pests due to their trophobiont-tending behavior or predators of pest herbivores. Vandermeer and colleagues (2003) found that Azteca ant activity in coffee plants cause a significant reduction of lepidopteran larvae, suggesting that because of the complex trophic relationships in the food web of traditional coffee plantations, these ants might be more beneficial than harmful (trophobiont-tending) to the farmer. Various studies have found that an aggressive open-area ground-foraging ant, Solenopsis geminata (Myrmicinae), dominates less diverse agricultural managements and more disturbed areas (Risch and Carroll 1982, Perfecto 1991b), which is desirable since this ant is presumably considered to be a good biological control agent (Risch and Carroll 1982, Perfecto and Sediles 1992). Nevertheless, Perfecto (1994), stresses that this species may turn out to be a pest in some situations suggesting that the management system (i.e.pesticides and disturbance) can be a determinant of the role this ant might play, either as a beneficial or harmful agent. As an illustration of this prediction, a study in Costa Rica (Varon 2002) found S. geminata to be dominant in sun coffee plantations

45

but was not found predating coffee berry borer in field conditions. Velez et al. (2003) neither found significant effect of S. geminata as predator of the berry borer in the field, but instead they found a shade loving ant Gnamptogenys sulcata (Ponerinae), to be an effective predator of this pest in Colombia. Therefore, it is possible that leaf litter and soil dwelling ants are an important connected compartment of the whole ant community, predating organisms falling from the canopy and coffee bushes, but also moving up to exploit honeydew sources.

Conclusions Most evidence from tropical agroforesty systems confirms Vandermeer and colleagues’ (2002) observation that the loss of ecosystem functions in managed systems is probably the most critical issue for everyone concerned with biodiversity loss. Not only for the conservation of associated fauna within the tropical agroforests, but also because the quality of the agricultural matrix surrounding forest fragments might be critical for conservation biology. Studies also call for caution when applying ecological concepts to the management of diverse agroecosystems, such as considering ants either cryptic herbivores or top predators (or both). Concepts about the relation of biodiversity and ecosystem function actually investigated can be applied to agroecology. It is essential that manipulations of the agroecosystem are done after rigorous understanding of each specific system to avoid disillusionment from the farmers and abandonment of environmentally friendly agricultural practices. Ants might be a clue for pest prevention and management if well understood and studied.

46

Acknowledgements I am grateful to Stacy Philpot and Ivette Perfecto for improving the earliest version of this manuscript.

Literature cited

Alonso, L.E. and D. Agosti. 2000. Biodiversity studies, monitoring and ants: an overview. Pages 1-8 In Agosti, D., J. D. Majer, L. E. Alonso and T. R. Shultz (Editors). Ants: Standard methods for measuring and monitoring Biodiversity. Smithsonian Institution. Washington and London. Altieri, M. A. and C. I. Nicholls. 1999. Biodiversity, ecosystem function and insect pest management in agricultural systems. Pages 69-84. In Collins, W.W. and C. O. Qualset, (Editors). Biodiversity in Agroecosystems. CRC Press LLC. Boca Raton, Florida. Altieri, M.A., E. N. Silva and C.I. Nicholls (Editors). 2003. O papel da biodiversidade nomanejo de pragas. Holos, Editora Ltda-ME. Ribeirao Preto SP. Brazil. Alvarez, G., I. Armbrecht, E. Jimenez, H. Armbrecht and P. Ulloa-Chacon. 2001. Antplant association in two Tococa species from a primary rain forest of Colombian Choco (Hymenoptera:Formicidae). Sociobiology 38: 585-599. Andersen, A.N. 2000. Global Ecology of rainforest ants, functional groups in relation to environmental stress and disturbance. Pages 25-34 In Agosti, D., J. D. Majer, L. E. Alonso, & T. R. Shultz (Editors). Ants: Standard methods for measuring and monitoring Biodiversity. Smithsonian Institution. Washington and London. Andersen A.N., B.D. Hoffmann, W.J. Muller and A.D. Griffiths. 2002. Using ants as bioindicator in land management: simplifying assessment of ant community responses. Journal of Applied Ecology 39: 8-17. Andow, D.A. 1991. Vegetational diversity and arthropod population response. Annual Review of Entomology 36: 561-586. Armbrecht, I. and P. Ulloa-Chacón. 1999. Rareza y diversidad de hormigas en fragmentos de bosque seco colombianos y sus matrices. Biotropica 31:646-653. Armbrecht, I., E. Jiménez, G. Alvarez, P. Ulloa-Chacón and H. Armbrecht. 2001. An Ant mosaic in the Colombian rain forest of Chocó. Sociobiology 37:491-509.

47

Armbrecht, I and I. Perfecto. 2003. Litter-twig dwelling ant species richness and predation potential within a forest fragment and neighboring coffee plantations of contrasting habitat quality in Mexico. Agriculture, Ecosystems and Environment 97:107-115. Beer, J., R. Muscheler, D. Kass and E. Somarriba. 1998. Shade management in coffee and cacao plantations. Agroforestry Systems 38: 139-164. Bengtsson, J. 1998. Which species? what kind of diversity? which ecosystem function? some problems in studies of relations between biodiversity and ecosystem function. Applied Soil Ecology 10:191-199. Bestelmeyer, B. T. and J. A. Wiens. 1996. The effects of land use on the structure of ground-foraging ant communities in the Argentine Chaco. Ecological applications 6:1225-1240. Blair, J.M., R.W. Parmelee, and M.H. Beare. 1990. Decay rates, nitrogen fluxes and decomposer communities of single- and mixed-species foliar litter. Ecology 71:1976-1985. Blüthgen, N., M.Verhaagh, W. Goitia, K. Jaffe, W. Morawetz and W. Barthlott. 2000. How plants shape the ant community in the Amazonian rainforest canopy: the key role of extrafloral nectaries and homopteran honeydew. Oecologia 125:229-240. Borrero, J. I. 1986. La substitucion de cafetales de sombrío por caturrales y su efecto negativo sobre la fauna de vertebrados. Caldasia 15:725-732. Bronstein, J. L. 1998. The contribution of ant-plant protection studies to our undestanding of mutualism. Biotropica 30:150-161. Brown, W.L. 2000. Diversity of Ants. Pages 45-79 In Agosti, D., J. D. Majer, L. E. Alonso, and T. R. Shultz (Editors). Ants: Standard methods for measuring and monitoring Biodiversity. Smithsonian Institution. Washington and London. Bruhl, C.A., T. Eltz and E. Linsenmair. 2003. Size does matter –effects of tropical rainforest fragmentation on the leaf litter ant community in Sabah, Malaysia. Biodiversity and Conservation 12:1371-1389. Cameron, T. 2002. the year of the “diversity-ecosystem function” debate. Trends in Ecology and Evolution 17:495-496. Campos, R.B.F., J. H. Schoereder, et al. 2003. Local determinants of species richness in litter ant communities (Hymenoptera: Formicidae). Sociobiology 41:357-367. Carson, R. 1964. The Silent Spring. Fawcett. New York. Davidson, D.W., S.C. Cook, R. Snelling and T.H. Chua. 2003. Explaining the abundance of ants in lowland tropical rainforest canopies. Science 300:969-972.

48

Dawes-Gromadzki, T.Z. 2002. Trophic trickles rather than cascades: conditional topdown and bottom-up dynamics in an Australian chenopod shrubland. Austral Ecology 27:490-508. Dejean, A., D. McKey, M. Gibernau and M. Belin-Depoux. 2000. The arboreal ant mosaic in Cameroonian rainforest. Sociobiology 35:403-423. Delabie, J.H. C. 1990. The ant problems of cacao farms in Brazil. Pages 555-569 In. R.K. Vandermeer, K. Jaffee and A. Cedeno, Editors. Applied Myrmecology: a world perspective. Westview, Boulder. Delabie, J.H.C, D. Agosti, and I.C do Nascimento. 2000. Litter ant communities of Brazilian Atlantic rain forest region. Chapter 1. Pages 1-17 In. Agosti, D., J. D.Majer, L Alonso and T. Schultz. (Editors), Sampling Ground-dwelling ants: case studies from the world’s rain forests. Curtin University School of Environmental Biology Bulletin. No. 18. Perth, Australia. 75p. Dempster, J.P. and T.H. Coaker. 1974. Diversification of crop ecosystems as a means of controlling pests. Pages 106-114 In Price Jones, D.and M.E. Salomon. Biology in pest and disease control. Blackwell Scientific Publications Oxford. Duffy, J.E. 2002. biodiversity and ecosystem function: the consumer connection. Oikos 99:201-119 Dietsch, T. V. 2003. Conservation and ecology of birds in coffee agroecosystems of Chiapas, Mexico. Dissertation. University of Michigan, Natural Resources and Environment. Ann Arbor, Michigan. U.S.A. 264 p. Donald, P.F., R.E. Green and M.F. Heath. 2000. Agricultural intensification and the collapse of Europe’s farmland bird populations. Proceedings:Biological Science 268:25-29. Feener, D.H. Jr. 2000. Is the assembly of ant communities mediated by parasitoids? Oikos 90:79-88. Fischer, A. and L. Vasseur. 2000. The crisis in shifting cultivation practices and the promise of agroforestry a review of the Panamian experience. Biodiversity and Conservation 9:739-756. Fittkaw, E.J. and H. Klinge. 1973. On biomass and trophic structure of the central Amazonian rain forest ecosystem. Biotropica 5:2-14. Floren, A. A. Freking, M. Biehl and K. E. Linsenmair. 2001. Anthropogenic disturbance changes the structure of arboreal tropical ant communities. Ecography 24:547554. Floren A., A. Biun, K. E. Linsenmair. 2002. Arboreal ants as key predators in tropical lowland rainforest trees. Oecologia 131:137-144.

49

Franco, A., R. Cárdenas, E.C. Montoya and I. Zenner de Polanía. 2003. Hormigas asociadas con insectos chupadores en la parte aérea del cafeto. Revista Colombiana de Entomología 29:95-105. Funes, F., L. García, M. Bourque, N. Pérez, and P. Rosset (Editors). 2002. Sustainable Agriculture and resistance. Food First Books. Oakland, California, U.S.A. 307p. Gallina, S.E., S. Mandujaro, A. Gonzalez-Romero. 1992. Importancia de los cafetales mixtos para la conservación de la biodiversidad de Mamíferos. Boletín de la Sociedad Veracruz de Zoología 2:11-17. Gallina, S., S. Mandujano and A. gonzález-Romero. 1996. conservation of mammalian biodiversity in coffee plantations of central Veracruz, Mexico. Agroforestry Systems 33:13-27. García-Barrios, L. 2003. Plant-plant interactions in tropical agriculture. Pages 11-58 In Vandermeer, J. (Editor). Tropical Agroecosystems. CRC Press. Boca Raton, Florida. U.S.A. 268 p. Gliessman, S.R. 2001. The ecological foundations of agroecosystem sustainability. Pages 3-14 In S. G. Gliessman. (Editor). Agroecosystem sustainability, developing practical strategies. CRC Press. Boca Raton, Florida, U.S.A. 210 p. Goehring, D.M., G.C. Daily, and C.H. Sekercioglu. 2002. Distribution of grounddwelling arthropods in tropical countryside habitats. Journal of Insect Conservation 6:83-91. Gotelli, N.J. and A.M. Ellison. 2002. Assembly rules for New England ant assemblages. Oikos 99:591-599. Greenberg, R., P. Bichier, A. Cruz Angon, and R. Reitsma. 1997. Bird populations in shade and sun coffee plantations in central Guatemala. Conservation Biology 11:448-459. Greenberg, R. and R.A. Rice. 2000. The Peruvian Shade-Grown Coffee Primer (manual de cafe de sombra y biodiversidad en el Perú). Migratory Bird Center, Smithsonian Institution, printed in U.S.A. Hector, A., A.J. Beale, A. Minns, S.J. Otway and J.H. Lawton. 2000. consequences of the reduction of plant diversity for litter decomposition: effects through litter quality and microenvironment. Oikos 90:357-371. Heller, M. C. and G. A. Keoleian. 2003. Assessing the sustainability of the US food system: a life cycle perspective. Agricultural Systems 76:1007-1041. Holldobler, B. and E.O. Wilson. 1990. The Ants. Harvard University Press. SpringerVerlag. Berlin, Heidelberg.

50

Holway, D.A. L. Lach. A.V. Suárez, N.D. Tsutsui, and T.J. Case. 2002. The causes and consequences of ant invasions. Annual Review of Ecology and Systematics 33:181-233 Hooper, D. U. and P. M. Vitousek. 1997. The effect of plant composition and diversity on ecosystem processes. Science 277:1302-1305. Huston, M.A. 1997. Hidden treatments in ecological experiments: re-evaluating the ecosystem function of biodiversity. Oecologia 110:449-460. Hunt, J.H. 2003. Cryptic herbivores of the rainforest canopy. Science 300:916-917. Ibarra-Núñez, G. 1990. Los artrópodos asociados a cafetos en un cafetal mixto del Soconusco, Chiapas, México, Variedad y abundancia. Folia Entomológica Mexicana 79:207-231. Ibarra-Núñez, G., J.A. García and M.A. Moreno. 1995. Diferencias entre un cafetal orgánico y uno convencional en cuanto a diversidad y abundancia de dos grupos de insectos. Pages 115-129 In Memorias Primera Conferencia Internacional IFOAM sobre café orgánico. Universidad Autónoma Chapingo. Mexico. Janzen, D.H. 1973. Dissolution of mutualism between Cecropia and its Azteca ants. Biotropica 5: 15-28. Johnson, M. D. 2000. Effects of shade-tree species and crop structure on the winter arthropod and bird communities in a Jamaican shade coffee plantation. Biotropica 32:133-145. Kaspari, M. 1993. Body size and microclimate use in Neotropical granivorous ants. Oecologia 96:500-507. Kaspari, M. 1996a. Testing resource-based models of patchiness in four Neotropical litter ant assemblages. Oikos 76:443-454. Kaspari, M. 1996b. Litter ant patchiness at the 1-m2 scale: disturbance dynamics in three Neotropical forests. Oecologia 107:265-273. Kaspari, M. And M. D. Weiser. 2000. Ant activity along moisture gradients in a neotropical forest. Biotropica 32:703-711. Kaspari, M., S.O’Donnell, and J.R. Kercher. 2000. Energy, density, and constraints to species richness: ant assemblages along a productivity Gradient. The American Naturalist 155:280-293. Kaspari, M., M. Yuan and L. Alonso. 2003. Spatial grain and the causes of regional diversity gradients in ants. The American Naturalist 161:460-477.

51

Klein, A.M., Steffan-Dewenter, I. and T. Tscharntke. 2002a. Predator-prey ratios on cocoa along a land-use gradient in Indonesia. Biodiversity and Conservation 11:683-693. Klein, A.M., I. Steffan-Dewenter, D. Buchori and T. Tscharntke. 2002b. Effects of land-use intensity in tropical agroforestry systems on coffee flower-visiting and trap-nesting bees and wasps. Conservation Biology 16:1003-1014. Krooss, S., M. Shaefer. 1998. The effect of different farming systems on epigeic arthropods: a five-year study on the rove beetle fauna (Coleoptera:Staphylinidae) of winter wheat. Agriculture, Ecosystems and Environment 69:121-133. Lachaud, J. P. and J. A. García Ballinas. 1999. Biodiversidad de la Myrmeco-fauna (Hymenoptera:Formicidae) en los agroecosistemas de cafe y cacao en Chiapas, México:enfoque sobre las subfamilias Ponerinae y Cerapachinae. En pp. 79-83 Memorias del XXXIV Congreso Nacional de Entomología, Sociedad Mexicana de Entomología, AguasCaliente Mexico. Lacher, T. E. Jr., R. D. Slack, L. M. Coburn, and M.I. Gostein. 1999. The role of agroecosystems in wildlife biodiversity. . Pages 147-165 in W..W. Collins and C. O. Qualset, (Editors)s. Biodiversity in Agroecosystems. CRC Press LLC. Boca Raton, Florida. Lavelle, P. and B. Pashanasi. 1989. Soil macrofauna and land management in Peruvian Amazonia (Yurimaguas, Loreto). Pedobiologia 33:283-291. Lawton J.H. 1998. Biodiversity and ecosystem function:getting the Ecotron experiment in its correct context:Forum. Functional Ecology 12:848-852. Leston, D. 1973. The ant mosaic-tropical tree crops and the limiting of pest and diseases. PANS 19:311-341. Letourneau., D.K. and L. A. Dyer. 1998. Density patterns of Piper ant-plants and associated arthropods: top-predator trophic cascades in a terrestrial system? Biotropica 30:162-169. Majer, J. D. 1972. The ant mosaic in Ghana cocoa farms. Bulletin of Entomological research 62:151-160. Majer, J.D. 1976. The influence of ants and ant manipulation on the cocoa farm fauna. Journal of Applied Ecology 13:157-175. Mas, A.H. 1999, Butterflies as biodiversity indicators and shade coffee certification in Chiapas, Mexico. M.Sc. Thesis University of Michigan, School of Natural Resources and Envrironment, Ann Arbor, Michigan, U.S. Matson, P. A., W. J. Parton, A. G. Power, M. J. Swift. 1997. Agricultural intensification and ecosystem properties. Science. 277:504-509.

52

McNeely, J.A. and S.J. Scherr. 2003. Ecoagriculture: strategies to feed the world and save wild biodiversity. Island press. Washington D.C. Medeiro, M.A.D., H.G. Fowler, and J.H. Delabie. 1995. The ant mosaic (Hymenoptera: Formicidae) of cocoa plantations in southern Bahia. Cientifica (Tacoticabal) 23:291-300. Moguel, P. and V. M. Toledo. 1999. Biodiversity conservation in traditional coffee systems of Mexico. Conservation Biology. 13:11-21. Morrison, L.W. 2000. Mechanisms of interespecific competition among an invasive and two native fire ants. Oikos 9:238-252. Muschler, R. and J. Beer. 2001. Multistrata agroforestry systems with perennial crops (selected papers from an international symposium held at CATIE Turrialba, Costa Rica, 22-27 February 1999). Agroforesty systems 53:v-vi. Naeem, S., L. J. Thompson, S. P. Lawler, J. H. Lawton and R. M Woodfin. 1994. Declining biodiversity can alter the performance of ecosystems. Nature 368:734736. Naeem, S., K. Hakansson, J. H. Lawton, M.J. Cawley and L. J. Thompson. 1996. Biodiversity and plant productivity in a model assemblaje of plant species. Oikos 76:259-264. Nestel, D. and F. Dickschen. 1990. Foraging kinetics of ground ant communities in different mexican coffee agroecosystems. Oecologia 84:58-63. Paoletti, M. G., D. Pimentel, B.R. Stinner and D. Stinner. 1992. Agroecosystem biodiversity: matching production and conservation biology. Agriculture, Ecosystems and Environment 40:3-23. Paoletti, M. G. 1995. Biodiversity, traditional landscapes and agroecosystem management. Landscape and urban planning 31:117-128. Peck, S.L., B. McQuaid et al. 1998. Using ant species (Hymenoptera Formicidae) as a biological indicator of agroecosystem condition. Environmental Entomology 27:1102-1110. Perfecto, I. 1990. Indirect and direct effects in a tropical agroecosystem: the maizepest.ant system in Nicaragua. Ecology 71:2125-2134. Perfecto, I. 1991a. Ants (Hymenoptera: Formicidae) as natural control agents of pests in irrigated maize in Nicaragua. Journal of Economic Entomology 84:65-70. Perfecto, I. 1991b. Dynamics of Solenopsis geminata in a tropical fallow field after ploughing. Oikos 62:139-144.

53

Perfecto, I .1994. Foraging behavior as a determinant of asymmetric competitive interaction between two ant species in a tropical agroecosystem. Oecologia 98:184-192. Perfecto, I. and A. Sediles. 1992. Vegetational diversity, ants (Hymenoptera: Formicidae), and herbivorous pests in a Neotropical agroecosystem. Environmental Entomology 21:61-67. Perfecto, I. and J. Vandermeer. 1994. Understanding biodiversity loss in agroecosystems: reduction of ant diversity resulting from transformation of the coffee ecosystems in Costa Rica. Entomology Trends in Agricultural Science Perfecto, I. and R. Snelling. 1995. Biodiversity and the transformation of a tropical agroecosystem: ants in coffee plantations. Ecological Applications 5:1084-1097. Perfecto, I. and J. Vandermeer. 1996. Microclimatic changes and the indirect loss of ant diversity in a tropical agroecosystem. Oecologia 108:577-582. Perfecto, I. and A. Castiñeiras. 1998. Deployment of the predaceous ants and their conservation in agroecosystems. Pages 269-289 In Barbosa, P. (Editor) Conservation Biological control Academic Press. Perfecto, I., and J. Vandermeer. 2002. The quality of agroecological matrix in a tropical montane landscape: ants in coffee plantations in southern Mexico. Conservation Biology 16: 174-182. Perfecto, I. and I. Armbrecht. 2003. Technological change and biodiversity in the coffee agroecosystem of Northern Latin America (Chapter 6) pp 159-194 In Vandermeer, J. (Editor) Tropical agroecosystems. CRC Press LLC. Boca Raton, Florida. Perfecto, I., R. A. Rice, R. Greenberg and M. E. Van der Voort. 1996. Shade coffee: a disappearing refuge for biodiversity. Bioscience 46:598-608. Perfecto, I., J. Vandermeer, P. Hanson, and V. Cartin. 1997. Arthropod diversity loss and the transformation of a tropical agroecosystem. Biodiversity and Conservation 6:935-945. Perfecto, I., A. Mas, T. Dietsch and J. Vandermeer. 2003. Conservation of biodiversity in coffee agroecosystems: a tri-taxa comparison in southern Mexico. Biodiversity and Conservation 12:1239-1252. Pimm, S.L. 1998. Extinction. Pages 20-38 in. W.J. Sutherland, (Editor). Conservation: Science and Action. Blackwell Science Ltd. Oxford. Pimentel, D., U. Stachow, D. A. Takacs, H. W. Brubaker, A. R. Dumas, J. J. Meaney, J. A. S. O’Neil, D. E. Onsi and D. B. Corzilius. 1992. Conserving biological diversity in agricultural/forestry systems. BioScience 42:354-362.

54

Pinkus-Rendón, M. 2000. Spider diversity: is there a decline with increasing shade in coffee plantations? M.Sc. Thesis. El Colegio de la Frontera Sur, ECOSUR, Tapachula, Mexico. Quiroz-Robledo, L. and J. Valenzuela-Gonzalez. 1995. A comparison of ground ant communities in a tropical rainforest and adjacent grassland in los Tuxtlas, Veracruz, Mexico. Southwestern Entomologist 20:203-213. Ramos-Suárez, M.P., H. Morales, L. Ruiz-Montoya, L. Soto-Pinto and P. RojasFernández. 2002. ¿Se mantiene la diversidad de hormigas con el cambio de bosque mesófilo a cafetales? . Pp 16-30 In Monro, A. and M.C. Pena. Actas del Simposio café y Biodiversidad. Revista Protección Vegetal, Año 12 (2) Edicion Especial. San Salvador. ISBN 0565091816. Rao, M.R., M.P. Singh and R. Day. 2000. Insect pest problems in tropical agroforestry systems: contributory factors and strategies for management. Agroforestry Systems 50:243-277. Rappole, J.H., D. I. King and J. H. Vega Rivera. 2002. Coffee and conservation. Conservation Biology 17:334-336. Reitsma, R., J.D. Parrish and W. McLarney. 2001. The role of cacao plantations in maintaining forest avian diversity in southeastern Costa Rica. Agroforestry Systems 53:185-193. Ricketts, T.H., G.C. Daily, P.R. Ehrlich and J.P. Fay. 2001. Countryside biogeography of moths in a fragmented landscape: biodiversity in native and agricultural habitats. Conservation Biology 15:378-388. Risch, S. J. and R. Carroll, 1982. The ecological role of ants in two Mexican agroecosystems. Oecologia 55:114-119. Roberts, D.L., R.J. Cooper and L.J. Petit. 2000. Use of premontane moist forest and shade coffee agroecosystems by army ants in western Panama. Conservation Biology 14:192-199. Robertson, H. 2002. Comparison of leaf litter ant communities in woodlands, lowland forests and montane forests of north-eastern Tanzania. Biodiversity and Conservation 11:1637-1652. Rojas, L., C. Godoy, P. Hanson, C. Kleinn and L. Hilje. 2001. A survey of homopteran species (Auchenorrhyncha) form coffee shrubs and poro and laurel trees in shaded coffee plantations, in Turrialba, Costa Rica. Revista de Biologia Tropical 49(34):1057-1065. Root, R.B. 1973. Organization of plant-arthropod associations in simple and diverse habitats: the fauna of collards (Brassica oleracea). Ecological Monographs 43:95124. 55

Roth, D. S., I. Perfecto and B. Rathcke. 1994. The effects of management systems on ground-foraging ant diversity in Costa Rica. Ecological Applications 4:423-436. Schupp, E.W. and D.H. Feener. 1991. Phylogeny, lifeform ana habitat dependence of ants defended plants in a Panamian forest. Pages 175-197 In C.R. Huxley and d.F. Cutler (Editors). Ant-Plant interactions. Osford University Press. Oxford. Shriar, A.J. 2000. Agricultural intensity and its measurement in frontier regions. Agroforestry Systems 49:301-318. Siebert, S.F. 2002. From shade- to sun-grown perennial crops in Silawesi, Indonesia: implications for biodiversity conservation and soil fertility. Biodiversity and Conservation 11:1889-1902. Somarriba, E., J. Beer and R.G. Muschler. 2001. Research methods for multistrata agroforestry systems with coffee and cacao: recommendations from two decades of research at CATIE. Agroforestry Systems 53:195-203. Soto-Pinto, L., I. Perfecto, J. Castillo-Hernandez and J. Caballero-Nieto. 2000. Shade effect on coffee production at the northern Tzeltal zone of the state of Chiapas, Mexico. Agriculture, Ecosystems and Environment 80:61-69. Swift, M. J. and J.M. Anderson. 1993. Biodiversity and Ecosystem Function in Agricultural Systems. Pages 15-42 In Schulze, E. D. and H. Mooney (Editors). Biodiversity and Ecosystem Function. Springer, Berlin. Swift, M. J., J. Vandermeer, P.S. Ramakrishnan, J.M. Anderson, C..K. Ong and B. A. Hawkinss. 1997. Biodiversity and agroecosystem function. pages 261-298 In H.A. Mooney, J. H. Cushman, E. Medina, O.E. sala and E. D. Schulze (Editors). Functional Roles of Biodiversity: a global perspective. John Wiley and Sons Ltd. Symondson, W.O.C., K.D. Sunderland and M.H. Greenstone. 2002. Can generalist predators be effective biocontrol agents? Annual Review of Entomology 47:561594. Tilman, D., J. Knops, D. Wedin, P. Reich, M. Ritchie, E. Siemann. 1997. The influence of functional diversity and composition on ecosystem processes. Science 277:1300-1302. Tilman, D. K.G. Cassman, P.A. Matson, R. Naylor and S. Polasky. 2002. Agricultural sustainability and intensive production practices. Nature 418:671-677. Tobin, 1995. Ecology and diversity of tropical forest canopy ants. Pages 129-147 In Lowman, M.D. and N.M. Nadkarni (Editors). Forest canopies. Academic press San Diego, California. U.S.A. Torres, J.A. 1984. Diversity and distribution of ant communities in Puerto Rico. Biotropica 16:296-303.

56

Van den Bosch, R. 1978. The pesticide conspiracy. Doubledy and Company, Inc. Garden City, New York. Van Mele, P. and N.T.T. Cuc. 2001. Farmers’ perceptions and practices in use of Dolichoderus thoracicus (Smith) (Hymenoptera:Formicidae) for biological control of pests of sapodilla. Biological Control 20: 23-29. Van Mele, P. And J.C. van Lenteren. 2002. Survey of current crop management practices in a mixed-ricefield landscape, Mekong Delta, Vietnam –potential of habitat manipulation for impruved control of citrus leafminer and citur red mite. Agriculture, Ecosystems and Environment 88:35-48. Vandermeer, J. 1989. The Ecology of Intercropping. Cambridge University Press. Cambridge, U.K. Vandermeer, J. 1995. The ecological basis of alternative agriculture, Annual Review of Ecology and Systematics 26:201-224. Vandermeer, J. (Editor). 2003. Tropical Agroecosystems. CRC Press. Boca Raton, Florida. U.S.A. 268 p. Vandermeer, J. and I. Perfecto. 1995. Breakfast of Biodiversity: the truth about rainforest destruction. Food First books. Oakland. Vandermeer, J. and I. Perfecto. 1997. The Agroecosystem: a need for the conservation biologist's lens. Conservation Biology 11:591-592. Vandermeer, J. and I. Perfecto. 2000. La biodiversidad y el control de plagas en sistemas agroforestales. Manejo Integrado de Plagas (Costa Rica) 55:1-5. Vandermeer, J. and R. Carvajal. 2001. Metapopulation dynamics and the quality of the matrix American Naturalist 159:211-220. Vandermeer, J., M. Van Noordwijk, J. Anderson, C. Ong, and I. Perfecto. 1998. Global change and multi-species agroecosystems" concepts and issues. Agriculture, Ecosystems and Environment. 67:1-22. Vandermer, J., D. Lawrence, A. Symstad and S. Hobbie. 2002. Effect of biodiversity on ecosystem functioning in managed ecosystems. Pages 221-223 In Laureau, M, S. Naeem, and P. Inchausti (Editors) Biodiversity and Ecosystem Functioning. Oxford University Press. Oxford. Vandermeer, J., I. Perfecto, G. Ibarra-Nunez, S. Phillpot and A. García Ballinas. 2003. Ants (Azteca sp.) as potential biological control agents in shade coffee production in Chiapas, Mexico. Agroforestry Systems 56:271-276. Varon, E.H. 2002. Distribución espacio-temporal de hormigas con potencial como depredadoras de Hypothenemus hampei e Hypsipyla grandella, en sistemas

57

agroforestales de café, en Costa Rica. Master Thesis. Catiue, Turrialba, Costa Rica, December 2002, 97 p. Vasconcelos, H.L. and D.W. Davidson. 2000. Relationship between plant size and ant associates in two amazonian ant-plants. Biotropica 32:100-111. Vélez, M., A.E. Bustillo and F. Posada 2003. Depredación de Hypothenemus hampei por Solenopsis geminata y Gnamptogenys sp. (Hymenoptera:Formicidae). Page 26 In Sociedad Colombiana de Entomología. Libro de Resumenes XXX Congreso. Cali, Universida Autonoma. Ward, P. 2000. Broad-scale patterns of diversity in leaf litter ant communities. Pages 99-121 In Agosti, D., J. D. Majer, L. E. Alonso, and T. R. Shultz (Editors). Ants: Standard methods for measuring and monitoring Biodiversity. Smithsonian Institution. Washington and London. Wardle, D.A., K.I. Bonner and K. Nicholson. 1997. Biodiversity and plant litter: experimental evidence which does not support the view that enhanced species richness improves ecosystem function. Oikos 79:247-258. Wardle, D.A., K.I. Bonner, G.M. Barker, G.W. Yeates, K.S. Nicholson, R.D. Bardgett, R.N. Watson and A. Ghani. 1999. Plant removals in perennial grassland: vegetation dynamics, decomposers, soil biodiversity and ecosystem properties. Ecological Monographs 69:535-568. Watt, A. D., N.E. Stork, B. Bolton. 2002. The diversity and abundance of ants in relation to forest disturbance and plantations in southern Cameroon. Journal of Applied Ecology 39:18-31. Way, M.J. and K.C. Khoo. 1992. Role of ants in pest management. Annual Review of Entomology 37:479-503. Willson, K.V. 1999. Coffee, Cocoa and tea. CABI Publishing CAB International. Wallingford. Oxon. United Kingdom. 300 pp. Witt, E. 2001. Seed dispersal by small terrestrial mammals in shaded coffee farms in Chiapas, Mexico. MSc Thesis. University of Michigan, Ann Arbor, U.S.A. 92 pages. Zak, D.R., W.E. Holmes, D.C. White, A.D. Peacock, and D. Tilman. (in press) Plant diversity, soil microbial communities and ecosystem function: are there any links? Ecology.

58

ABSTRACT

Diversity and function of leaf litter ants in Colombian coffee agroecosystems by Inge Armbrecht Chair: Ivette Perfecto The loss of biodiversity and associated functions in managed ecosystems is one of the most critical issues for conservation because it affects the long-term sustainability of the planet. This dissertation investigates causes and effects of agricultural intensification on the structure and function of ant assemblages in coffee plantations and forest patches of Andean Colombia. Coffee intensification significantly reduced vegetation complexity and increased management index values, as measured from 14 habitat variables. Using a leaf-litter sampling protocol I found that less intensified coffee farms harboured significantly higher ant species richness. Ant communities in polygeneric shaded coffee were significantly more similar to forest patches than any other management type. Additionally, significant positive associations among ant communities and the number of species involved in ant mosaics dropped sharply with intensification. I carried out nesting resource augmentation experiments to examine how the quantity and quality of nesting resources affect ant biodiversity. Ants responded quickly to the addition of nesting resources across the intensification gradient. In the quality resources experiment, the

59

diversity of ants nesting in multi-species groups of twigs was significantly higher than in monoespecific groups across all scales. These results suggest that bottom-up mechanisms are important for ant species loss throughout the technification process although other biotic and abiotic factors such as microhabitat and competition might also play a role. Finally, ant predation in shaded vs. unshaded coffee plantations was studied in the field and laboratory. Exclosure experiments demonstrated that ant communities prey on the coffee berry borer (Hypothenemus hampei) in both shaded and unshaded coffee plantations, but this function increased with shade in the wet season, when most infestation takes place. Ants were more effective predators in laboratory conditions than in the field. Predation in the field fluctuated between 2%27% of borer adults. My studies strongly support the conclusion that the structure and function of ant biodiversity are negatively affected by intensification of coffee agriculture. I recommend that immediate action be taken to stop the destruction of traditional coffee practices and prevent further biodiversity loss in hotspots of tropical biodiversity.

60

CHAPTER 2

habitat

changes

in

colombian

coffee

farms

UNDER

INCREASING

MANAGEMENT intensification Abstract I analyzed a set of environmental and vegetation variables in order to characterize an intensification gradient for coffee production agroecosystems. I measured 14 habitat variables within 12 Colombian farms classified into four management systems at the Risaralda region of Colombia: Forests, Polygeneric Shaded coffee, Monogeneric Shaded coffee and Sun coffee plantations. The habitat variables were categorized into three vertical levels: arboreal, shrubs and soil. I employed univariate and multivariate analyses and found that the habitat effect is driven mainly by the drastic changes (i.e. elimination) of the arboreal level vegetation along the intensification gradient, although variables at other levels showed gradual and sometimes unexpected changes along the gradient. I then adapted the management index developed by Mas and Dietsch (2003) to the coffee plantations in this study. The quantitatively supported management index showed a close correspondence to the initial qualitatively classification of the farms. I conclude that coffee production intensification has clearly measurable effects on habitat characteristics and that the management index reflects the gradient of intensification in the studied farms. This approach (using the management index) could be highly valuable for the programs of shade coffee certification and conservation goals.

61

Introduction Agricultural intensification has been defined as the patterns of land-use change with the common feature of increase use of resources that aim at increasing agricultural production (Giller et al. 1997). It is generally associated with specialization, increasing mechanization and generalized use of agrochemicals and other external inputs (Giller et al. 1997; Decaens and Jimenez 2002). This intensification negatively impacts the agricultural land, which is usually the matrix among forest fragments and therefore valuable for conservation purposes (Vandermeer and Carvajal 2001, Perfecto and Vandermeer 2002). There is growing awareness in the literature that agroecosystems should be a priority in the biological conservation agenda (Paoletti et al. 1992, Pimentel et al. 1992, Vandermeer and Perfecto 1997, McNeely and Scherr 2003) due to growing evidence that some agroecosystems are repositories of high levels of biodiversity (Pimentel et al. 1992, Roth et al. 1994, Perfecto et al. 1996, 1997, Perfecto and Armbrecht 2003). It has been well documented that agroecosystems with high planned biodiversity foster high levels of associated biodiversity and that the intensification of agriculture negatively affects associated biodiversity (Andow 1991, Pimentel et al. 1992, Decaens and Jimenez 2002, Perfecto and Armbrecht 2003). Swift et al. (1996) have hypothesized several predictions regarding alternative patterns in which associated biodiversity decreases with intensification of agriculture. However, testing mechanistic hypotheses first requires the quantification of intensification. The coffee agroecosystem that has received considerable attention over the last decade with regard to the effect of intensification on biodiversity (Nestel et al. 1993, Perfecto et al. 1996, 1997, Greenberg et al 1997a, Moguel and Toledo 1999, Dietsch 2003, Armbrecht and Perfecto 2003) and there is a need to quantify habitat changes for this particular case. There are two reasons that justify the need for a better quantification: first, coffee production occurs across a wide gradient of agricultural intensification, involving different levels and varieties of shade trees (Moguel and Toledo 1999, Perfecto

62

et al. 1996, 1997, Johnson 2000); and second, this agroecosystem is now known to be important for conservation biology (Vandermeer and Perfecto 1997, Moguel and Toledo 1999, Rappole et al 2003, Perfecto and Armbrecht 2003). Intensification of agriculture can be quantified through various indices, which consider the measurement of variables presumed to determine its degree at particular scales. Giller et al. (1997) proposed an index or “degree of intensification” which was further modified by Decaens and Jimenez (2002)and named “Agricultural Intensification Index (AI). The AI index is computed from seven subindices, equally weighted, which range from 0-1: (1) LUI, the land use intensity, or the proportion of the year the system is cropped, (2) FF, the mean fire frequency or burnings/year, (3) TF, the mean tillage frequency/year, (4) MPF, the mean frequency of motorized practices/year, (5) SR, the mean annual stocking rate (International Animal Units/ha) and (6) FR, the mean fertilization rate (kg of chemicals used per year) and (7) PCR, the mean pest control rate (kg of chemicals used per year). The Agricultural Intensification Index is the average of these seven index values. The coffee agroforest poses additional challenges to researchers since some of the intensification variables defined in the AI, are not meaningful in this context (e.g. LUI, FF, TF) because coffee agroforests are not tilled, no fire is not used (unless accidental), and coffee is a perennial crop (Beer 1998), standing for several decades (Willson 1999). New studies accounting for differences among shade regimes in coffee plantations have quantified of independent variables related to the structure of vegetation and habitat changes (Babbar and Zak 1995, Perfecto and Vandermeer 1996, Greenberg et al. 1997a, Decaens and Jimenez 2002, Klein et al. 2002, Armbrecht and Perfecto 2003,

63

Mas and Dietsch 2003). But many studies focusing on different coffee management systems have not reported on such variables (e.g. Borrero 1986, Greenberg et al. 1997b, Beer, 1998, Wunderle and Latta 1996, Ibarra-Núñez and García-Ballinas1998, Molina 2000, Sossa and Fernandez 2001, Ricketts et al. 2001. Rojas et al. 2001, Perfecto and Vandermeer 2002). The potential problems associated with the failure to quantitatively assess intensification have been highlighted by Rappole et al. (2003), who argued that any plantation, regardless of the diversity and density of shade, could be considered a “shade coffee” plantation. The lack of rigorous definition of shade coffee may have serious practical implications since shade coffee has emerged in recent years as an important component of biodiversity conservation programs among several environmental organizations such as Conservation International, The Rainforest Alliance and Eco-OK (Perfecto and Armbrecht 2003, Dietsch 2003). Mas and Dietsch (2003) developed an index of management intensity (management index, MI) for coffee agroecosystems in order to evaluate whether qualitative differences between shade coffee agroecosystems correspond to quantitative differences in vegetation and farm management. Their index used seven equally weighted vegetation variables, which they considered directly related to flying insects such as butterflies. This management index was then related to the richness of fruit feeding butterflies in differentially shaded coffee plantations of Chiapas, Mexico (Mas and Dietsch 2003). The MI can be flexibly adapted to different targeted taxa in biodiversity studies. The present study is intended to test the extent to which a subjective categorical classification of farms along a gradient of intensification corresponds to a quantitatively supported classification by applying a modified version of Mas and Dietsch’s

64

management index. Additionally, this study aims to quantify habitat changes among Colombian coffee farms under a strong intensification process and compare them to other related studies.

Study site The Andean mountains of Risaralda Department, Colombia (5º 08’ N; 75º 56’W), where the Apía municipality is located, range between 1400-1900 m a.s.l. Annual temperature averages 20oC and annual precipitation 2,320 mm, which has a bimodal distribution with peakers in May and December (raw data from IDEAM Meteorological Stations, Colombia, 2002). The Apía region has a rugged topography with scattered secondary forest fragments, that become continuous at higher altitudes (~4000m elevation) at the “Tatamá Natural Reserve”. The Apía was a typical, traditional shade coffee growing area for many decades (Apían Senior Isabel Marin, pers. com.). However, in the last 20 years, coffee crop cover has decreased by more than 700% of the initial area covered (Apía’s NFCG’ committee internal report, pers. com. December 2, 2002), from which no more than 30-40% is under shade trees other than plantain barriers (pers. obs. December 2002). During the last decade, many coffee plantations were converted into cattle pastures and other agricultural uses, and the coffee plantations that still stand have suffered a dramatic intensification change. The changes involve variying degrees of elimination of shade trees including the complete elimination of shade, which is concomitant with increasing application of agrochemicals for the control of the coffee berry borer and for weeds.

65

Twelve farms, grouped in four qualitatively classified management types were haphazardly chosen at the Apía municipality. Following Nestel and Altieri (1992), the four management types were: Forest (F), Organic Polygeneric Shade coffee (PS), Monogeneric Shade coffee (MS), and Coffee Monoculture or Sun Coffee (“Sun”). For simplicity each farm was assigned a code (Table 1). Two sets of criteria were used to decide the qualitative classification of the farm management type: first, was the presence and diversity of trees, based on a visual check; second, the information that the farmers provided about farm management with regard to number of agrochemical applications per year and shade management. After gathering this information, I determined that the three forest fragments appeared to be secondary natural dense vegetation, disturbed and isolated. The forests were located relatively close to the coffee plantations (the primary forests that exist many kilometers from the municipality would not have fit for comparison purposes of this study). The Monogeneric Shaded coffee plantations were subjectively perceived to be dominated by trees belonging to the genus Inga or Cordia. The Polygeneric Shaded coffee farms were organic and their shade trees were apparently more varied than those in the MSs. Sun coffee plantations had few or no shade trees, although it was not possible to find 100% open plantations, because farmers still allow some isolated valuable trees and plantains within their plots. Therefore, measurements on the existing plantains (or any isolated tree) within these sun coffee plantations were done even though the habitat by definition should not have arboreal vegetation. As part of the study site description, percentage slope was calculated by measuring both the vertical and horizontal components of the slope at four haphazardly chosen sites at each farm.

66

Methods

The characterization of the habitat at the Apía region followed the protocol established by Mas and Dietsch (2003) for coffee farms under different management systems in Mexico. In this protocol, a “management index” quantifies the effects of the management intensification on the shade tree canopy. Seven variables, each varying from 0 (least intensive condition) to 1 (most intensive condition) were used in calculating this index. Although Mas and Dietsch’s paper reports and statistically compares 13 vegetation variables among their seven farms, they actually used only seven of those variables to determine their management index. Not all of the variables measured by Mas and Dietsch’s were measured in this study because they focused on fruit feeding butterflies (influenced by canopy structure) while I was seeking to develop a management index that would be applicable to the study of ground and leaf litter organisms such as ants. In the study I obtained the following 14 habitat variables, which are grouped by three vertical strata: (a) arboreal stratum: 1) percentage canopy cover, 2) tree species richness, 3) tree density, 4) percentage of trees with epiphytes, 5) number of epiphytes, 6) tree height, 7) percentage dominance of one shade tree, 8)diameter at breast height of live trees; (b) coffee bush stratum: 9) vertical heterogeneity (up to 5.4m), 10) number of coffee bushes, 11) coffee height; (c) soil (low) stratum: 12) number of logs, 13) log diameter, and 14) number of logged trees. All habitat variables were measured between November and December 2002 at the 12 farms. In each farm, two sampling sites (or “circles”) separated by approximately 50-100m, were haphazardly selected. Each sampling site consisted of a circle of twelve-

67

meter radius within which all trees greater than 8.13 cm diameter at breast height (dbh) were identified to species. Also a visual inventory of the tree species in each plot was made. Heights, dbh, presence of epiphytes, fruit or flowers were recorded for each tree. All coffee bushes or understory plants (in Forests) between 2.5-8.1cm dbh were counted within a five-meter radius circle located at the center of the larger circle. For the management index purposes, these were the only trees recorded, but for an inventory purpose of the trees, an additional visual search was done by a botanist at each of plantation plots. Canopy and soil sampling points were established at four-meter intervals along the north-south and east-west axis of the sampling location for a total of thirteen sampling points. At each sampling point, a spherical densiometer (Forestry Suppliers, Biloxi, Mississippi) was used to obtain the percentage canopy cover; dead logs (>2.5cm) were counted in a 2m x 2m area next to the sampling point, diameter of all logs were measured as well. Vertical heterogeneity of the understory was measured by using the Vertical Intercept Line technique (Mills et al. 1991) in which a 5.4m aluminum tube is labeled with tapes of two colors. One color defined nine vertical consecutive intervals of 60cm (A), and the other color defined 54 ten-centimeter intervals (B), in such a way that each of the nine intervals contains six 10cm intervals. The tube was placed vertical between two coffee bushes. Any vegetation contact within an imaginary 1dm radius cylinder around each of the tube segment was registered. Shannon-Wiener index (Magurran 1988) was calculated by using A as species and B as abundances.

Data analyses: Means of the variables for the four management types were statistically compared by mixed model nested Analyses of Variance with circles nested

68

within farms and farms nested within managements. Tests for assumptions of normality (Kolmorogov-Smirnov tests) and homogeneity of variances (Levene’s tests) (Zar 1999) were carried out. Data not normally distributed were transformed (inverse of the square root) in order to meet this assumption (Zar 1999). Tukey post-hoc tests were performed whenever the statistical differences were detected. Multiple comparison post-hoc tests were Bonferroni corrected. All univariate analyses were performed by SPSS-10 for Windows (SPSS Inc®). Multivariate analyses involved Cluster analyses (Ward linkage method and Euclidian distances) and Principal Component analyses as implemented by Statistica-5 for Windows, Multivariate Exploratory Techniques (Statistica Inc. 2002, ©Copyright StatSoft, Inc.).

Management index: Mas and Dietch’s management index (MI) weights each of the variables equally along a scale from 0.0 to 1.0 (0.0 represents the least managed/most “natural” system). The standardized index values for the variables are then added together such that the number of variables included in the study constitute the maximum value possible reached by the MI. Thus Mas and Dietsch’s (2003) index ranges from 0.0 to a possible high of seven, since they used seven variables for their index. Different variables were treated somewhat differently in the management index following Mas and Dietsch’s (2003) protocol. For example, with variables such as tree species richness, which is assumed to decrease as intensification increases, the 0.0 value was based on the tree species richness for the richest circle in the richest forest. The assumption is that the expected tree species richness could vary depending on the native forest type present in the region. Therefore in this study, for each farm, the proportion of

69

the average tree species richness with respect to the richest forest (Table 2) was calculated and then subtracted from 1.0, so that a higher value would reflect a higher intensification (e.g. the number of tree species in the second circle of F1 forest was 16 species). This procedure was used to obtain the following standardized values (1) tree (2) species richness, (3) tree abundance, (4) number of logs, (5) percentage canopy cover (assuming 100% is 0.0 value), and (6) percentage trees with epiphytes. Proportion of the relative difference was used to calculate the standardized values (Mas and Dietsch, 2003) for six additional variables: (7) Percentage dominance of one tree species in the plantation, (8) average tree height, (9) average diameter of logs on soil, (10) average coffee bush height, (11) vertical heterogeneity, (12) live trees dbh. For example, for the average tree height, 1.0 was based on the circle with the lowest average tree height on the assumption that more intensive management includes regular pruning that produces a lower average tree height. The 0.0 value for average tree height (ATH) was based on the point with the overall highest value, which was assumed to be the least intensive condition. The ATH value for each point was calculated as the proportion of overall lowest value, then subtracted from 1.0 (Index Value = 1-[point ATH – low ATH]/ [high ATH – low ATH]). For this study the lowest ATH was 3 and the highest was 10.65m. The reason for choosing this quantification procedure is that it amplifies the range of variation of the index since a lower limit (above zero) is defined, so the standardized values for each variable determined in this way are relative to the Apía region. The standardized value management index for variables such as logged tree bases and number of coffee bushes was calculated as the proportion relative to the highest value

70

found in any of the 26 circles. These two variables are assumed to increase with management intensification. Summarizing, 14 variables were measured in all of the farms at Apía (Table 2), but only eight (Table 3, for reasons presented in the “discussion” section) were actually used to calculate the management index for each farm. Standardized values for all of the variables were calculated at the farm level, and not at the circle level.

Results Most of the 14 habitat or vegetation variables exhibited increasing or decreasing trends throughout the intensification gradient of coffee agriculture (Tables 2, 3; Figures 1, 4), with the exception of five of them: percentage of trees with epiphytes, tree height, dbh of live trees, diameter of fallen logs, and number of logged trees. These variables were withdrawn from the total management index values (Table 3). For further clarity, the variables were grouped into three categories or vertical levels: arboreal (variables #1-8); coffee bushes (variables #9-11); and soil (variables #12-14) (numbers in parenthesis refer to those variables numbers in Table 2). For the arboreal level, those habitat variables that visually impact an observer showed a gradual change along the gradient of agricultural intensification (Figure 2). For the coffee and soil levels, changes were frequently more obvious in the Sun coffee plantations (Figures 3 and 4). The “coffee bush” level variables showed that the density of coffee plants significantly increased in the Sun coffee plantations, while the vertical heterogeneity and coffee bush height decreased (Figure 3). The apparent contradiction between the trend of these last two variables is explained because bushes are smaller in Sun coffee, and thus

71

have less altitudinal categories accounting for an increasing vertical heterogeneity. The average number of logs, the only soil-level variable that showed significant differences, also decreased gradually across the gradient (Figure 4). The overall tree species richness across all the studied farms through the inventory was 71 species (Appendix A). Tree species richness values per circle were sometimes similar between Forests and Polygeneric Shaded coffee plantations (Table 2), although the identity of the trees was frequently different (Appendix A). A Cluster analysis incorporating the complete set of variables measured in the study (14 variables, Table 2) indicated two groups of farms separated by the highest distance (Figure 5). A first cluster contains all Sun and Monogeneric Shade, and both of these management systems are further separated into two groups. A second cluster involves the Forests and Polygeneric Shaded farms, but they are not further separated into discrete groups as happened in the first cluster. Principal Component Analysis’s output (Figure 6) revealed that the first two Principal Components accounted for 69.25% and 23.78% of the total variance respectively, and in total for 93.03%. Variables such as tree species richness, tree density and percentage of trees with epiphytes were important for the first factor, while percentage dominance of one tree and the number of coffee bushes were important for the second factor.

Discussion The results from this study showed that a qualitatively classification of twelve farms into four management systems overall matched the quantitative analyses derived from 14 quantified habitat variables and the management index (Figure 1). The visual

72

perception of the management impact in the coffee farms is obvious at first glance. The qualitative classification in this study was based upon conspicuous arboreal characteristics within each farm, such as the overall appearance of the shade trees in terms of richness, density and level of shade. The information from farmers was a second important criterion to decide a priori classification of the farms. Cluster Analysis and Principal Component Analysis were consistent in their outputs, uncovering the definition of discrete groups of farms according to the management intensification and the a priori classification, but most important, showing the similarities between forest and Polygeneric Shaded coffee (Figures 5 and 6). The Principal Component Analysis technique captures most of the variability of the system and the type of variation captured by the first Principal Component (PC) strongly dominates all other types of variation. My interpretation is that the first PC (responsible for 69% of the total variance) was driven by the arboreal component of the habitat variables. This interpretation is supported by the high loadings (>0.83) of the first PC in both Forests and Polygeneric Shade, while these loadings were extremely low in Suns (0.08-0.013) and in Monogeneric Shades (0.270.57) (Figure 6). The second PC probably captured most of the variability generated by soil and coffee bush vegetation variables, with extremely high loadings in Suns (0.960.98). These results suggest that other studies comparing qualitatively chosen coffee farms of contrasting management systems or shade levels are reliable at the broad scale even without reporting habitat measurements (e.g. Borrero 1986, Ibarra-Núñez and García-Ballinas1998, Ricketts et al. 2001, Rojas et al. 2001). The trends found in this study along the intensification gradient are consistent with changes found in the habitat found by other studies. For example, the forest patches

73

in Mas and Dietsch (2003) in Chiapas, Mexico were slightly richer in terms of tree species than the forest patches included in this study at Apía (12.9 and 11.67 tree species, respectively, in equivalent areas) and the same trend was found within the rustic coffee plantations of Chiapas, as compared to the Polygeneric Shaded coffee plantations of Apía (average 6.65 and 5.3 tree species respectively). Trees were taller (9.14m and 7.15m) but thinner (10.2 and 21.85cm dbh) in Chiapas than in Apía. In another study, in Mexico, Soto-Pinto et al. (2002) reported an average tree height of 7.6m in shaded coffee plantations of Chilon (although up to 20m), which is consistent with the heights observed at Apía in this study (7.15m). Nevertheless, these trees provided similar shade in Colombian plantations of Apía (canopy cover 78.9%) and the rustic plantations in Mas and Dietsch’s study (73.3% average). However, canopy cover was unexpectedly lower in Soto-Pinto et al.’s (2002) traditional coffee plantations (46.6%) possibly because most of the trees in these plantations of Chilon (Mexico) were planted fruit trees. Intensified shaded systems in Mas and Dietsch (36.16%) were also similar to Colombian ones (Figure 2) in terms of Canopy Cover measures, and these both also matched those reported by Armbrecht and Perfecto (2003) in Chiapas, Mexico (35%). Armbrecht and Perfecto’s study was done in the same farms as Mas and Dietsch’s study in Chiapas, Mexico, but what it is interesting for the goals of the present study, is that measurements in two independent studies in the same plantations (in different years) were extremely similar despite the high dynamic (pruning) management in shaded coffee plantations. In the present study plantains (Musa x paradisiaca L., mainly) planted in a barrier fashion provided the 28.8% canopy cover in Sun coffee plantations. Canopy cover deserves special attention because it is likely to be influencing biological activities through

74

physiological responses of the associated biota inside the farms through changes in temperature and relative humidity (e.g. Kaspari and Weiser 2000). Tree density found in this study is also consistent to Soto-Pinto et al. (2000): 463 shade trees/ha in traditional shaded coffee plantations of Chilon (Mexico) vs. 132 to 508 trees/ha in the Polygeneric Shaded coffee plantations in this study (Table 2). Klein et al. (2002) measured nine habitat variables in coffee and other agroforestry systems of Sulawesi, including percentage cover and height of vegetation, but did not present the actual data in their paper. The comparative discussion here points out that both measurements and categorical classifications in coffee plantations are consistent in different countries and different studies. This fact provides a basis for further reliable studies synthesizing information or making comparative analyses in literature reviews, and also for reliability of scientific assessment for Shade Coffee certification programs. Although this study has suggested overall consistency in habitat changes along the gradient of intensification of coffee production and across similar studies, it also showed some inconsistencies. Variables such as average tree height and percent epiphytes did not change the same way in Colombia as reported by Mas and Dietsch (2003) and two explanations are offered here: (1) the energy that the farmer invests in pruning can be directed differently depending on the type and age of the shade tree. For example, trees were thicker, smaller and provided more than double canopy cover in Monogeneric Shades planted with Inga spp. as compared to those planted with Cordia allidora (76% vs 37%) (Table 2). This last tree species is pruned laterally to stimulate straight vertical growth since its wood is highly valuable in the market and an additional source of income for farmers; (2) the percentage of trees with epiphytes was strongly

75

influenced by the number of trees existing in the plantations. If there is only one isolated tree in Sun coffee and it happens to have an epiphyte it would represent 100%, therefore I propose the use of the raw number of epiphytes instead (variable 5, Table 2). It is remarkable that none of the Monogeneric Shaded coffee farms contributed many epiphytes in the system as compared to the Polygeneric Shaded farms. The density of trees in Polygeneric Shaded coffee plantations was double that in Monogeneric Shaded farms, but there were 54 times as many epiphytes (Table 2), a fact that cannot be easily explained as a sampling effect. The question of why this “overpopulation” of epiphytes occurs could be related to metapopulation dynamics and dispersers of these specialized plants related to the diversity of their “supporting” trees (Vandermeer and Carvajal 2001). Tree diameter at breast height (dbh) was also surprisingly higher at all of the shaded coffee systems than in forests (Table 2), which is likely to be a consequence of the age distribution of trees in the successional process taking place in these forests. The number of logged trees was unexpectedly high in one of the Polygeneric Shaded coffee plantations (Table 2) and this measurement suggests that this was a much more shaded and diverse rustic plantation before its conversion into organic production for coffee exportation. Results in this study strongly suggest the importance of taking into account the “arboreal” stratum variables for the management index, since almost all of them were significantly different among management types (Figure 2, Table 2). Most importantly, the means of all arboreal variables from Polygeneric and Monogeneric Shaded coffee plantations in Figure 2 were significantly different, and there is a clear intensification gradient that distinguishes these two types of shaded coffee farms (Figure 1). These facts

76

confirm the great importance of providing a quantitative basis directed to the organizations that certify shade grown coffee (e.g. Rainforest Alliance or Conservation international). Several general conclusions emerged from this study. A high correspondence was detected between the initial qualitative classification and the results of quantifying habitat variables of twelve farms along an increasing gradient of Colombian coffee production. This high concordance was mainly a consequence of the presence, diversity and structure of the arboreal vegetation in the farms (first component PCA Figure 6). Most of the changes in vegetation variables detected in this study were consistent with other studies in coffee plantations. Finally, Mas and Dietsch’s management index proved to be useful in describing the intensification of coffee production. The implications of this study for shaded coffee certification and conservation programs are of high relevance in a recent debate (e.g. nine labels in which coffee products are included, at http://www.eco-labels.org/home.cfm). Questions have been recently raised about the real benefits of promoting shade-grown coffee among consumers as valuable for conservation of biodiversity, if no distinction is made regarding differently shaded coffee plantations (Rappole et al. 2003). If such a distinction needs to be made, this study provides an additional step toward this goal.

Acknowledgements Maria Cristina Gallego, Gerardo Peñaranda, John Jairo Muñetón and Joaquín Colmenares helped out in the field. Phillip Silverstone and the Herbarium at Universidad del Valle facilitated botanical sampling tools. Trees were determined to species level by

77

William Vargas. Diana and Laura Conde helped with some data analyses. Thanks to all farmers for allowing me to work in their plantations as well as kind logistical help from Patricia Marin, Francisco Herrera, Norma Henao, Oscar Díaz and Everett Mayes. A Fulbright-LASPAU scholarship made this dissertation possible. This project was funded by Colciencias (Colombia) project grant code 1106-12-11693 and Universidad del Valle, Cali.

78

LITERATURE CITED Andow, D.A. 1991. Vegetational diversity and arthropod population response. Annual Review of Entomology 36:561-586. Armbrecht, I and I. Perfecto 2003. Litter-twig dwelling ant species richness and predation potential within a forest fragment and neighbouring coffee plantations of contrasting habitat quality in Mexico. Agriculture, Ecosystems and Environment 97:107-115. Babbar, L.I. and D.R. Zak. 1995. Nitrogen loss from coffee agroecosystems in Costa Rica: Leaching and denitrification in the presence and absence of shade trees. Journal of Environmental Quality 24:227-233. Beer, J., R. Muscheler, D. Kass and E. Somarriba. 1998. Shade management in coffee and cacao plantations. Agroforestry Systems 38:139-164. Borrero, J.I. 1986. La substitución de cafetales de sombrío por caturrales y su efecto negativo sobre la fauna de vertebrados. Caldasia 15:725-732. Decaens, T. and J.J. Jiménez. 2002. Earthworm communities under an agricultural intensification gradient in Colombia. Plant and Soil 240:133-143. Dietsch, T.V. 2003. Conservation and ecology of birds in coffee agroecosystems of Chiapas, Mexico. Dissertation. University of Michigan, Natural Resources and Environment. Ann Arbor, Michigan. U.S.A. 264 p. Greenberg, R., P. Bichier, and J. Sterling. 1997a. Bird populations in rustic and planted shade coffee plantations of eastern Chiapas, Mexico. Biotropica 29:501-514. Greenberg, R., P. Bichier, A. Cruz Angon and R. Reitsma. 1997b. Bird populations in shade and Sun coffee plantations in central Guatemala. Conservation Biology 11:448-459. Giller, K.E., M.H. Beare, P. Lavelle, A. M.N. Izac and M.J. Swift. 1997. Agricultural intensification, soil biodiversity and agroecosystem function. Applied Soil Ecology 6:3-16. Ibarra-Núñez, G. and J.A. García-Ballinas. 1998. Diversidad de tres familias de arañas tejedoras (Araneae:Araneidae, Tetragnathidae, Theridiidae) en cafetales del Soconusco, Chiapas, Mexico. Folia Entomologica Mexicana 102:11-20. 79

Johnson, M.D. 2000. Effects of shade-tree species and crop structure on the winter arthropod and bird communities in a Jamaican shade coffee plantation. Biotropica 32:133-145. Kaspari, M. and M.D. Weiser. 2000. Ant activity along moisture gradients in a neotropical forest. Biotropica 32:703-711. Klein, A.M., I. Steffan-Dewenter, D. Buchori and T. Tscharntke. 2002. Effects of landuse intensity in tropical agroforestry systems on flower-visiting and trap-nesting bees and wasps. Conservation Biology 16:1003-1014. Magurran, A.E. 1988. Ecological diversity and its measurement. Princeton University Press, Princeton. Mas, A. and T. Dietsch. (2003 in press). An index of management intensity for coffee agroecosystems to evaluate butterfly species richness. Ecological Applications. McNeely, J.A. and S.J. Scherr. 2003. Ecoagriculture: strategies to feed the world and save wild biodiversity. Island Press. Washington D.C. 323 p. Mills, G.S., J .B. Dunning and J. M. Bates. 1991. The relationship between breeding bird and vegetation volume. Wilson Bulletin 103:468-479. Moguel, P. and V.M. Toledo. 1999. Biodiversity conservation in traditional coffee systems of Mexico. Conservation Biology 13:11-21. Molina, J. 2000. Diversidad de escarabajos coprofagos (Scarabaeidae: Scarabaeinae) en matrices de la zona cafetera (Quindio-Colombia). Page 29 in Federación Nacional de Cafeteros de Colombia, (Editors). Memorias Foro Internacional Café y Biodiversidad. Agosto 10-12, Chinchiná, Colombia. Nestel, D. and M.A. Altieri. 1992. The weed community of Mexican coffee agroecosystems: effect of management upon plant biomass and species composition. Acta Ecologica 13:715-726. Nestel, D., F. Dickschen, M.A. Altieri. 1993. Diversity patterns of soil macro-Coleoptera in Mexican shaded and unshaded coffee agroecosystems: an indication of habitat perturbation. Biodiversity and Conservation 2:70-78.

80

Paoletti, M.G., D. Pimentel, B.R. Stinner and D. Stinner. 1992. Agroecosystem biodiversity: matching production and conservation biology. Agriculture, Ecosystems and Environment 40:3-23. Perfecto, I. and J. Vandermeer. 1996. Microclimatic changes and the indirect loss of ant diversity in a tropical agroecosystem. Oecologia 108:577-582. Perfecto, I. and J. Vandermeer. 2002. The quality of agroecological matrix in a tropical montane landscape: ants in coffee plantations in southern Mexico. Conservation Biology 16 174-182. Perfecto, I. and I. Armbrecht. 2003. Technological change and biodiversity in the coffee agroecosystem of Northern Latin America (Chapter 6). Pages 159-194 In Vandermeer, J. (Editor). Tropical agroecosystems. CRC Press LLC. Boca Raton, Florida. Perfecto, I., R.A. Rice, R. Greenberg and M.E. Van der Voort. 1996. Shade coffee: a disappearing refuge for biodiversity. Bioscience 46:598-608. Perfecto, I., J. Vandermeer, P. Hanson, and V. Cartin. 1997. Arthropod diversity loss and the transformation of a tropical agroecosystem. Biodiversity and Conservation 6:935-945. Pimentel, D., U. Stachow, D.A. Takacs, H.W. Brubaker, A.R. Dumas, J.J. Meaney, J.A. S. O’Neil, D.E. Onsi, and D.B. Corzilius. 1992. Conserving biological diversity in agricultural/forestry systems. BioScience 42:354-362. Rappole, J.H., D. King and J.H. Vega-Rivera. 2003. Coffee and conservation. Conservation Biology 17:334-336. Roth, D.S., I. Perfecto and B. Rathcke. 1994. The effects of management systems on ground-foraging ant diversity in Costa Rica. Ecological Applications 4:423-436. Ricketts, T.H., G.C. Daily, P.R. Ehrlich and J.P. Fay. 2001. Countryside biogeography of moths in a fragmented landscape: biodiversity in native and agricultural habitats. Conservation Biology 15:378-388. Rojas, L., C. Godoy, P. Hanson, C. Kleinn and L. Hilje. (2001) A survey of homopteran species (Auchenorrhyncha) form coffee shrubs and poro and laurel trees in shaded coffee plantations, in Turrialba, Costa Rica. Revista de Biologia Tropical 49:1057-1065. Sossa, J. and F. Fernandez. 2000. Himenopteros de la franja cafetera del departamento del Quindío. Pages 168-180 In. Numa, C. and L.P. Romero (Editors). Biodiversidad y sistemas de producción cafetera en el departamento del Quindío. Instituto Alexander von Humboldt. Dic. 2000. Bogotá, Colombia.

81

Soto-Pinto, L., I. Perfecto, J. Castillo-Hernandez and J. Caballero-Nieto. 2000. Shade effect on coffee production at the northern Tzeltal zone of the state of Chiapas, Mexico. Agriculture, Ecosystems and Environment 80:61-69. Soto-Pinto, L., I. Perfecto and J. Caballero-Nieto. 2002. Coffee shade; its effects on berry borer, leaf rust and spontaneous herbs. Agroforestry Systems 55:37-45. SPSS Inc. 2002. www.spss.com/

Statistica Inc. 2002. http://www.statsoftinc.com/ Swift, M.J., J. Vandermeer, P.S. Ramakrishnan, J.M. Anderson, C.K. Ong and B.A. Hawkins. 1996. Biodiversity and agroecosystem function. Pages 261-298 In H.A. Mooney, J.H. Cushman, E. Medina, O.E. sala and E.D. Schulze (Editors). Functional Roles of Biodiversity: a global perspective. John Wiley & Sons Ltd. Vandermeer, J. and I. Perfecto. 1997. The Agroecosystem: a need for the conservation biologist's lens. Conservation Biology 11:591-592. Vandermeer, J. and R. Carvajal. 2001. Metapopulation dynamics and the quality of the matrix American Naturalist 159:211-220. Willson, K.V. 1999. Coffee, Cocoa and tea. CABI Publishing CAB International. Wallingford. Oxon. United Kingdom. 300 pp. Wunderle, J. and S.C. Latta. 1996 Avian abundance in sun and shade coffee plantations and remnant pine forest in the cordillera Central, Dominican Republic. Ornitología Neotropical 7:19-34. Zar, J.H. 1999. Biostatistical analysis. Prentice Hall. 4th edition. New Jersey, U.S.A.

82

Table 1. Names and general characteristics of the twelve farms involved in the study at the Apía Municipality of Risaralda Department, Colombia. Use of pesticides: high is at least two applications/year (pesticides and herbicides), moderate: at least one application/year, low: less than one application/year. Farm’s name

Management Area

Elevation

Percentage

Use of pesticides

code

(ha)

(m a.s.l.)

slope (%)

(appl./year)

MonteverdeF

For1

15

1845

59.6

None

Playabonita

For2

2

1444

61.3

None

El Porvenir

For3

1.5

1605

39.3

None

La Playita-1

PS1

15

1490

48.3

None (organic)

La Esperanza

PS2

4

1500

34.6

None (not organic)

La Clarita

PS3

7.5

1550

43.8

None (organic)

Monteverde

MS1

4

1720

43

Low

Buenos Aires

MS2

6

1440

40

Low

El Convenio

MS3

4

1465

64.4

Low

La Felisa

Sun1

6

1480

32.5

Moderate

La Estrella

Sun2

14

1470

17.5

Moderate

La María

Sun3

3

1405

2.5

High

83

Table 2. Average values for each of the variables measured (based on two circles per plot) to characterize the habitat of nine coffee plantations and three forest fragments at Apía. Variables are averages at either the “circle” level (#2,3,4,6,7,8,10, 15) or the “site” level (#1,9,11,12,13,14) or at the farm level (#5). Values marked ‫ ٭‬are for understory plants. Last column, P, indicates α probability for mixed model Analysis of Variance. Degrees of freedom are 3, 8 except for variables 10 and 11 where forest values were taken out. # 1 2 3 4

HABITAT VARIABLES Percentage canopy cover Tree species richness Tree density in 452 m2 (#/circle) % trees with epiphytes

84

For1

For2

For3

PS1

PS2

PS3

MS1

MS2

MS3

Sun 1

Sun 2

Sun 3

P

96.1

96.7

90.5

69.0

80.9

86.9

36.9

74.3

77.4

25.3

24.2

36.9

80%)

Apía

F1

MonteverdeF

20

15

1845

Apía

F2

Playabonita

11

2

1444

Apía

F3

El Porvenir

19

1.5

1605

Támesis

F4

La Cumbre F

N/A

0.7

1650

Apía

P1

La Playita-1

6

15

1490

Apía

P2

Esperanza

16

4

1500

Apía

P3

La Clarita

15

7.5

1550

Támesis

P4

12

25

1515

Apía

M1

El Paraíso de la Virgen Monteverde

Cordia alliodora

4

1720

Apía

M2

Inga edulis Mart.

6

1440

Apía

M3

Buenos Aires El Convenio

Inga edulis Mart.

4

1465

Támesis

M4

La Cumbre

Cordia alliodora

6

1650

Apía

S1

La Felisa

Musa x paradisiaca L.

6

1480

Apía

S2

La Estrella

Musa x paradisiaca L.

14

1470

Apía

S3

La María

Musa x paradisiaca L.

3

1405

Támesis

S4

Casa Loma

Musa x paradisiaca L.

1

1650

96

Table 2. Nonparametric Incidence Coverage Estimator (ICE) for the number of ant species existing at coffee plantations of two regions in Colombia, Apía (farms numbered 1-3) and Támesis (all numbered 4). Averages are provided for each farm, both for leaf litter and coffee bushes, as well as for management systems, codes for farms are the same as Table 1, and ordered in an increasing intensification gradient. Farm/management

F1 F2 F3 F4 All forests P1 P2 P3 P4 All Polygeneric shaded coffee M1 M2 M3 M4 All Monogeneric shaded coffee S1 S2 S3 S4 All Sun coffee

ICE, Estimated # ant species leaf litter n=20 plots 56 39 53 43 48

ICE, Estimated # ant species coffee bushes n=20 coffee bushes 44 21 23 22 27

57 59 64 43 56

30 53 39 32 38

44 39 49 25 39

24 39 25 21 27

37 48 29 23 34

30 15 34 18 24

97

Table 3. Dominance indices (DI) of ant species at Apía municipality. Positive values indicate that significant negative associations outnumber positive associations. Zero values (in contrast to blank spaces) are actually DI values. Species are positioned in a decreasing order of overall abundance in leaf litter. Some relatively abundant species do not appear in the table because no significant associations involving them were found (i.e. Solenopsis am, Pheidole cocciphaga, Camponotus cf. novogranadensis, Brachymyrmex ca, and Paratrechina N1). Apía ants

Pheidole radoszkowski Solenopsis gr. brevicornis Pheidole olo Paratrechina cf. steinheili

Leaf litter

Coffee bushes

F

P

M

S

-1

0

-1

+1

0.2

0.71

-0.5

+1

0 -1

Brachymyrmex cf. cordemoyi

-0.33

+1

-1

-1

0

-1

Pheidole ca. ebenina

-1

+1

-1

Pheidole mnd

-1

-1

-1

Hypoponera ng

+1

0

+1

M

S

-1

-1

+1

+1 -1

+1

-1

Solenopsis aa

-1

Cyphomyrmex rimosus

+1

Pheidole apa

-1

Gnamptogenys bisulca

+1

+1 +1

Gnamptogenys striatula

0

Dolichoderus bispinosus Myrmelachista az

P

0

Octostruma balzani

Pyramica gundlachi

F

-1

Linepithema mo

-1 -1

98

+1

Figure 1. Location of the two study sites in Colombia and relative location of the farms within each study site. Twelve farms are located at Apía municipality (Risaralda Department) and four at Támesis municipality (Antioquia Department).

99

a) 80

# ant species in leaf litter

70 60 50 40 30 20 10 0 0

1000

2000

3000

4000

5000

num ber of individual ants

b) 35

# ant species in leaf litter

30 25 20 15 10 5 0 0

200

400

600

800

1000

number of individual ants

Figure 2. Individual-based rarefaction curves for ant species from litter samples in four management types along an intensification gradient of coffee production at Apía (a), and Támesis (b) municipalities. Open-circles = Forests; squares = Polygeneric shaded coffee plantations; Xs = Monogeneric coffee plantations, diamonds = Sun Coffee plantations. Gray vertical lines at each value are 95% low and high confidence limits.

100

a) 45

# ant species in coffee bushes

40 35 30 25 20 15 10 5 0 0

500

1000

1500

2000

2500

num ber of individual ants

# ant species in coffee bushes

b) 25 20 15 10 5 0 0

200

400

600

800

1000

1200

number of individual ants

Figure 3. Individual-based rarefaction curves for ant species from coffee bushes in four management types along an intensification gradient of coffee production at Apía (a), and Támesis (b) municipalities. Open-circles = Forests; squares = Polygeneric shaded coffee plantations; Xs = Monogeneric coffee plantations, diamonds = Sun Coffee plantations. Gray vertical lines at each value are 95% low and high confidence limits.

101

a)

Bray-Curtis % similarity

.

35 30 25 20 15 10 5 0

0

F vs.F 1

F vs.P 2

F vs.3M

F vs4 .S

5

300

10 9 8 7 6 5 4 3 2 1 0

250 200 150 100 50

# individuals shared only with forests

# species shared only with forests

b)

0 P

M

S

management type

Figure 4. (a): Mean and +SE Bray-Curtis percentage similarity index for leaf litter ants (n=10) for each coffee management type vs. forest patches. For reference Apía Forest vs. Forest mean is shown at the left side of the plot (not included in the statistical analysis). (b): Number of litter ant species (black triangles) and individuals (open squares) that are exclusively shared by forests vs. each coffee management type at Apía. At the horizontal axis: P: Polygeneric Shaded coffee, M: Monogeneric Shaded coffee, S: Sun coffee.

102

a

b Solenopsis gr. brevicornis sp.1 50

40

40

30

30

# SU

# SU

Pheidole radoszkowskii

50

20

20 10

10

0

0 1

2

3

1

4

c

2

3

4

d Paratrechina steinheili

Pheidole olo

40

60 50

30 # SU

# SU

40 30

20

20

10

10 0 1

2

3

0

4

1

e

2

3

4

f Hypoponera ng

Pheidole ebenina 20

# SU

# SU

15 10 5 0 1

2

3

4

40 35 30 25 20 15 10 5 0 1

g

2

3

4

h Solenopsis aa

Gnamptogenys gr. striatula

12

10

10

8 # SU

# SU

8 6 4

6 4 2

2 0

0 1

2

3

4

1

103

2

3

4

i

j Solenopsis gr. brevicornis sp. 1

Pheidole gm

20 15

10

# SU

# SU

15

5

10 5 0

0 1

2

3

1

4

2

3

4

Figure 5. Frequency of most abundant ants along increasing intensification at Apía (a-h) and Támesis (i-j) municipalities. Absissa: 1: Forests; 2: Polygeneric Shade; 3: Monogeneric shade and 4: Sun coffee. Black bars represent number sampling units in which the species was present in leaf litter and white bars for coffee bushes.

104

% ant spp participating in associations .

a) 25 20 15 10 5 0 Forests 1

0

2 PS

MS 3

Sun coffee 4

5

# associations/all possible x 10E4

.

b) 80 70 60 50 40 30 20 10 0 0

Forests 1

Ps

2

Ms

3Sun

4

5

Figure 6. Simplification of the system and loss of the diversity of interactions at the leaf litter level in Apía municipality. (a) Percentage of ant species (relative to each management system) involved in significant interactions with other ant species (b) Significant associations relative all possible associations among any two ant species existing in each management system.The absissa is ordered along increasing gradient of intensification of coffee production.

105

-

+

-

+

+

+ +

+

+

+ +

+

-

+

-

+

+ +

+

+

+

-

+

-

+

-

+

-

-

106

+ +

Figure 7. Network of statistically significant associations among leaf litter ants at Apía region. Lines labeled (+) or (-) were positive or negative associations respectively. Lines connecting ant species become thicker with increasing Chi Square test significance.

FORESTS

POLYGENERIC

Transect 1

Transect 2

Gb

Cr

Po

Cr

Pr

Pe

Pr

Gb

Cr

Po

Cr

Hn

Pr

Pe

x

Cr

Po

Gb

Cr

Hn

Pr

Pe

Pr

Cr

Po

Gb

Cr

Hn

Pr

Pe

x

Hn

Pr

Gb

FARM 1 Gb

Cr

Cr

Gb

Cr Po Sb

Sb

Ps Ob

Sb

Gb

Ob

Sb

Ob

Hn

Ps

x

Ps

107

Cr

Hn

Ps

Gb Gb

Ob

Cr

Gb

Cr

Sb

Po

Ps

Pr

Pe

Po

Pr

Gb

Pr

Bc

Pr

Bc

Pr

Bc

Sb

Po

Pr

Bc

Sb

Po

Sb

Ob

Hn

Ob

Hn

Sb

Ob

Hn

Ob

Hn

Ob

Po

Gb

Pa Pe Pr Ps

Ob Ob

Hn

Ps

Ob

Hn

Ps

Hn

Sb

x

Sb

Ob

Ps

Ps

Hn

Ps

Ps

Sb

Ob

Hn

Ps

x

Ps

Sb

Cr

x

Ps

Ps

Sb

Ob

Hn

Hn

Ps

Sb

Po

Ps

Sb

Ob

Hn

Sb

x

Hn

Sb

Pr

Pr

Sb

Pr

Sb

Pr

Sb

Ps

Pr

Pr

Sb

Po

Pr

Sb

Pr

Sb

Po

Pr

Sb

Pr

Po

Sb

Po

Pr

Pr

Sb

Sb

Sb

Pr

Sb

Po

Bc

x

Ps Po

Sb Ps

Ps

Sb

Po

Ps

Sb

Pr

Sb

Sb

Pr

Gs

Sb

Pr

Gs

Sb

Sb

Po

Ps

Sb

Po

Ps

Sb

Sb

Po

Ps

Sb

Ps Ps

Pe

Po

Ps

Sb

Pe

Po

Ps

Sb

Pa

Bc

Sb

Po

Sb

Po

Po Po

Po

Ps

Ps

Sb

Po

Ps

Pa

Bc

Pa

Pe

Sb

Po

Ps

Sb

Po

Ps

Pr

Bc

Sb

Po

Ps

Sb

Po

Ps

Bc

Sb

Po

Sb

Po

Sb

Pr

Pr

Gs

Sb

x

Bc

Pr

Sb

Sb

Ps

x

Pr

Sb

Ps

Po

Pr

Bc

Pr

Pr

Po

Ps

Bc

Pa

Pr

Ps

Po

Po

Sb

Sb

Po

x

Pr

Sb

Sb Sa

Bc

Pe

Pr

Sb

Bc

Pr

Sb

Po

Pr

Pr Hn

Sb

Pr

Pr

Pr

Bc

Pr Sb

Sb

Bc

Pa

Pr

Sb

Bc

Pr

Sa

Sb

Ps

Po

Sa

Sb

Pr

Pr

Pr

Pr

Sb Sb

Po

Po

Pr

Pr

Pr

Bc

Sb

Sb

Po

Bc

Sb

Pr

Sb

Sb

Bc

Pr

Sb

Po

Bc

Pe

Sb

Ps Ps

Transect 2

Sb

Pr

Pr

Ps

Ps

Po

Sa

Pe

Hn

x

Pr

Po

Sb

Sa

Po

Pr

Hn

Hn

Po

Ps

Bc

Sb

FARM 3

Hn

Bc

Cr

Hn

Sb

Pr

Sa

Po Ps

Pr

Sb

Sb Ps

Hn

Gb

Sb x

Hn

Gb

Hn

Ps

Cr

x

Sb

Po

Ps

Gb

Pr

Hn

Bc

Po

Gb

Pe

Transect 2

Hn Sb

Pr

Ps

Gb

Transect 1

Pe

Hn

Sb

SUN COFFEE

Transect

Bc

Bc

Hn

MONOGENERIC

x

Gb

Gb

FARM 2

Transect 1 Transect 2

Pr Sb

Pr Pr

Gs Sb

Gs

Sb

Gs

Sb

Gs

Sb

Pr

Sb

Pr

Pr

Po

Ps

Sb

Po

Ps

Po

Ps

Sb

Po

Ps

Sb

Pr

x

Figure 8. Schematic representation of ant spatial distribution in four management systems. Less compartments are found in Sun coffee. Only those leaf litter ants involved in significantly negative associations at Apía appear in the figure. Solid lines delineate ant territories. Two columns at each management type (separated by dashed lines) indicate the two transects in each farm and the three rows separated by dashed lines correspond to the three farms within each management type. Sb= Solenopsis gr. brevicornis; Gb=Gnamptogenys bisulca; Cr= Cyphomyrmex rimosus; Po= Pheidole olo; Hn= Hypoponera ng; Ps= Paratrechina steinheili; Ob= Octostruma balzani; Pr=Pheidole radoszkowskii; Pa=Pheidole apa.; Pe=Pheidole ebenina; Bc=Brachymyrmex cordemoyi; Sa=Solenopsis aa; Gs=Gnamptogenys gr. striatula. Any “x” indicates two species in a column overlapping territories.

Appendix A. List and total abundance of each ant species at Apía municipality from leaf litter and coffee bushes. The farms are ordered in an increasing intensification fashion. F1, F2, F3: Forests; P1, P2, P3: Polygeneric coffee plantations; M1, M2, M3: Monogeneric shaded coffee plantations; S1, S2, S3: Sun coffee or unshaded coffee plantations (Table 1 provides details on these farms).

108

APIA ANT SPECIES SUBFAMILY CERAPAHYINAE Acantostychus neg SUBFAMILY DOLICHODERINAE Azteca cm Dolichoderus an Dolichoderus bispinosus Linepithema ca Linepithema mo Tapinoma ha Tapinoma melanocephalum Tapinoma npc SUBFAMILY ECITONINAE Eciton burchelli Labidus coecus Neivamyrmex sp.1 SUBFAMILY FORMICINAE Brachymyrmex ca. Brachymyrmex cf. cordemoyi Brachymyrmex ne. Brachymyrmex np Brachymyrmex pa Camponotus am Camponotus cf. indianus Camponotus (Myrmobrachys)1 Camponotus cf.novogranadensis

F1

F2

F3

P1

P2

LEAF LITTER P3 M1 M2

M3

S1

S2

S3

F1

F2

F3

P1

COFFEE BUSHES P2 P3 M1 M2

1

3

4

21

173

1

113

5

M3

S1

S2

S3

1 2

4

2

6

93

14

1

1

102 4 39

50 12

15

4

4 20

20

1

81

25

13

1

14

20

1

50

1

444

1 1

1 1

2

100

100 38 11

100 43

302 15

5 26 31

10 53

50 37 100

2

4 63

53 58

51 26

33

123 41

10 4

2

5

1

21 20 10

8 8 1

35 8

4 28

4

30

1 156

7 30 34

28 5

29

1

6

2

7

1

2 16

1

3

2

10

33

1

3 1

3

10

4 1

29

35

78

1 3

33 179

Camponotus ru Camponotus smp APPENDIX A (Continued) APIA MUNICIPALITY ANT SPECIES

109

Myrmelachista az Paratrechina N1 Paratrechina cf. steinheili SUBFAMILY MYRMICINAE Acanthognatus teledectus Acromyrmex sp. Apterostigma gr. pilosum sp1 Atta cephalotes Carebara reticulata Cephalotes do Cephalotes maculatus Cephalotes pm Crematogaster co Crematogaster (Neocrema) sp1 Crematogaster (Neocrema) sp2 Crematogaster (Neocrema) sp3 Cyphomyrmex minutus Cyphomyrmex rimosus Leptothorax np Leptothorax (Nesomyrmex) pittieri Leptothorax (Nesomyrmex) tristani Neostruma am Octostruma balzani Octostruma es Octostruma stenocarpa Pheidole aar Pheidole apa Pheidole apar Pheidole cp Pheidole cpa Pheidole cocciphaga Pheidole dc

5

1 5

F1

F2

187

F3

P1

P2

39

130 1 169

1 185

4

LEAF LITTER P3 M1 M2 3 2 127

102

30 119

M3

S1

S2

3 154 239

11 137

2 95

S3

F1

F2

F3

1 2 17

72

P1

COFFEE BUSHES P2 P3 M1 M2

467

3

10

24

170 1 50

51 99

M3

S1

S2

72

64

S3

1 15

81

2 3

4 1 2

3

1

1 1

1

100 20

1

1 3

1

2

1 4 1 50 3 1 94

12

2 5

21

25

1

1

1 103 9

1 13

10

5 13 7

27

1 2 35 20 1

3 2

1

23 94 10

23

1 20

2 7

1

8 23

3

15

35 6 8

45

3

26

113

33

1

12

4

7

9

10 44

1 1 3

3 2 4

3

1

3

1

1 5

45

15

26

5 16

2

5 5

11 2

1 6

2 2

3 12

1 7 35

12 60

4

21

5 68

5

155

27 15

31

15

5

2

2

4

6

3

Pheidole gr. flavens ca. ebenina Pheidole es

10 24

97 53

360

110 2

14

69

27

F1

F2

F3

P1

P2

1

20

19

LEAF LITTER P3 M1 M2 22 5 28

113

5

169

93

97

56

41

5

M3

S1

S2

S3

118

52

51

3

1

53

F2

F3

1 3

3 2 25 51

1

1

40

4

35

1

M3

S1

S2

S3

294 24 1

144 107 2

108 33

1 30

10

6

2

2

91

106

95 75

51 1 1

APPENDIX A (Continued) APIA MUNICIPALITY ANT SPECIES Pheidole esn Pheidole mad Pheidole mirabilis Pheidole mnd Pheidole npl Pheidole olo Pheidole radoszkowskii Procryptocerus (subpilosus) lepidus Procryptocerus cf. scabriusculus Pyramica raptans Pyramica gundlachi Rogeria besucheti

257

98

110

Rogeria ca Rogeria ne Rogeria ru Solenopsis gr. brevicornis sp. 1 Solenopsis aa Solenopsis ae Solenopsis am Solenopsis geminata Strumigenys biolleyi Strumigenys mc Strumigenys mr Tetramorium ep Trachymyrmex es Wasmannia auropunctata SUBFAMILY PONERINAE Anochetus ce Discothyrea am Discothyrea pe

72

3 22 161

F1

P1

COFFEE BUSHES P2 P3 M1 M2

58

8 178 1

241 140 1

43 228

9

14 1

8

3 1 1 122 69 20 134

193 248 4

39

7 1 297 106

3 5

84 233 79

3

190 257

315 138 25

30

315 177

122 206

4 67

1 21 2

3

100 15 1

1

100 3 1

135 19 2

192 8 4

1

100 40

232 5 12 3

3

1 1

5 1 103 32 3 85

1 367 47 2 14

113 51 30

1 10 135 67 28 1

73 77

216 1

104

73

131 100 6

280 3 5 5

331 2 50

53

1 11 101

50

1 2

100

24

7

173

70 151

20

59 9 3

3

7

2 33 40 25 1 1

4

56

40

2

31

4

Ectatomma ruidum Gnamptogenys bisulca

9 178

27

11

F1

F2

F3

5

3

2

6

6

4

1

1

1

1

APPENDIX A (Continued)

111

APIA MUNICIPALITY ANT SPECIES Gnamptogenys minuta Gnamptogenys continua Gnamptogenys ry Gnamptogenys strigata Gnamptogenys gr. striatula Heteroponera ca Heteroponera cap Heteroponera microps Hypoponera ca Hypoponera ng Hypoponera opaciceps cf. Hypoponera pe Hypoponera punctatissima Hypoponera sp B Kempf Hypoponera sp.U Odontomachys erythrocephalus Pachycondyla carbonaria Pachycondyla ferruginea Pachycondyla goeldii cf. Pachycondyla impressa Pachycondyla ne Pachycondyla pl SUBFAMILY PSEUDOMYRMECINAE Pseudomyrmex gracilis Pseudomyrmex oculatus Pseudomyrmex pa Pseudomyrmex pallens Pseudomyrmex rochai Pseudomyrmex tenuissimus

P1

P2

LEAF LITTER P3 M1 M2 3 1

M3

S1

S2

S3

31

F1

F2

F3

P1

COFFEE BUSHES P2 P3 M1 M2

M3

S1

S2

6 6

2

8

51

1

1 184

5 7 2 101

145

13 1

3 242

1 1 58

4

5

2

1 2

6

5

4

1

7 1 1 3 1

1 175 32

16 4 1

7 7

1 1

5

1 3

7 3

30 1 2

1 1

2 2

5 1

1

1 1 1 1

1

1 1

S3

1 1

1

1 1

1

2 2 1

5 1

1

2 3

1 1 1

2 2 8 3

2 3

1

1

Appendix B. List and total abundance of each ant species at Támesis municipality from leaf litter and coffee bushes. The farms are ordered in an increasing intensification fashion. Farms codes, according to Table 1 are: F4, P4, M4, S4 Forest, Polygeneric coffee plantation; Monogeneric shaded coffee plantation and Sun coffee.

Támesis ant species

Coffee bushes

Leaf litter F4

P4

M4

S4

F4

P4

M4

S4

SUBFAMILY DOLICHODERINAE Azteca cm

5

2

Azteca np

6

Linepithema mo

1

51

Tapinoma ha

1

6

50

SUBFAMILY ECITONINAE Cheliomyrmex ro

50

Eciton burchelli

61

Labidus coecus

100

18

SUBFAMILY FORMICINAE Brachymyrmex an Brachymyrmex cf. cordemoyi

50 2

11

12

13

70

Brachymyrmex ca Camponotus cf. novogranadensis

5

1

2

Formicina ac

15

8

1

64

1

3

21

102

3

2 6

Paratrechina mn Paratrechina cf. steinheili

8 2

1

Formicina ne Myrmelachista az

2

26 100

1 5

12

SUBFAMILY MYRMICINAE Acromyrmex sp.1

1 1

Leptothorax pittieri Crematogaster (Neocrema) sp1

1 1

Crematogaster es Crematogaster ng

55

Cyphomyrmex rimosus

1

Octostruma mi

3

Pheidole cpn

1

Pheidole da

55

30

24

27

7

22

111

151

1 64

3

1

Pheidole pe Pheidole gm

1 71

1 296

Pheidole gd

3

Pheidole mc

50

Pheidole aar

23

341

69

50 67

2

2 1

112

164

57

Appendix B (continued) Coffee bushes

Leaf litter

Tamesis ant species F4

P4

M4

S4

Pheidole npl Procryptocerus lepidus

1

3

1

1

2

380

53

2

92

174

63

51

14

108

2

6

1

3

14

31

4

135

3 1

1

1 30

20

Solenopsis ne

1

Strumigenys eg 10

Pyramica gundlachi Strumigenys cl

8

2

1

1

3 12

Wasmannia auropunctata

6

SUBFAMILY PONERINAE Ectatomma ruidum

1

Gnamptogenys ry

1

Gnamptogenys pe

1

Hypoponera mp

1

Heteroponera pq

1

1

14

Hypoponera ca

1

3

6

Odontomachus erytrocephalus

2

Pachycondyla impressa

3

Pachycondyla ferruginea

Pachycondyla vil Ponerina pec SUBFAMILY PEUDOMYRMECINAE

8

3 1 1 9

3

21

3

Pseudomyrmex simplex

15

Pseudomyrmex rochai

1

8

1

1

Pseudomyrmex gracilis Pseudomyrmex oculatus

1 1

3

Pachycondyla pm Simopelta williamsi

22

50

Solenopsis dae Solenopsis mn

4

75

Solenopsis aa Solenopsis diplorrhoptum cl

S4

1

Solenopsis ae Solenopsis dn

M4

11

Neostruma sp.

Solenopsis aep

P4

2

Pheidole gr. flavens ca. ebenina

Solenopsis gr. brevicornis.

F4

1

Pseudomyrmex rn

2

113

Appendix C. Results of any possible associations among ants, which came out statistically significant in leaf litter at Apía. Chi-Square Yates corrected P values (all of them df = 1) are presented, followed in parenthesis by the management type (F: forests, P: Polygeneric shaded coffee; M: monogeneric shaded coffee; S: Sun coffee), and whether the significant association was positive, or negative. Leaf litter ants Apia Pheidole radoszkowskii

Solenopsis gr. brevicornis sp. 1 P0.1

II Div. vs. Div. X2 = 9.60 df: 7 p>0.1

Only October

--

--

Only December

--

--

X2 = 11.83 df: 6 P>0.05

X2 = 40.71 df: 7 P