<%BANNER%>

Feedback Dynamics between Plants and Soil Microorganisms in a Fragmented Landscape in the Tropical Andes

Permanent Link: http://ufdc.ufl.edu/UFE0043714/00001

Material Information

Title: Feedback Dynamics between Plants and Soil Microorganisms in a Fragmented Landscape in the Tropical Andes
Physical Description: 1 online resource (150 p.)
Language: english
Creator: Pizano, M Camila
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: amf -- andes -- colombia -- ecology -- feedback -- forests -- pastures -- pathogens -- plantations -- plants -- soil
Biology -- Dissertations, Academic -- UF
Genre: Botany thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Studies addressing soil microbial communities (SMC) have mostly emphasized particular soil organisms such as arbuscular mycorrhizal fungi (AMF) or soil pathogens within a single habitat. The main goals of my dissertation were to better understand how the interaction between plants and SMC vary across habitats of contrasting abiotic and biotic conditions, and address the role of whole SMC as well as distinct components of SMC. My study took place in the Central Cordillera in Colombia, which is a heterogeneous agricultural landscape comprised of tropical pre-montane forest fragments embedded in highly fertilized, agricultural monocultures, mainly sun-grown coffee plantations and pastures. I worked with SMC, and plants that are common in three contrasting habitats: pastures (Brachiaria grass), coffee plantations (coffee), and forest fragments (forest tree species). I first tested for the effects of whole-SMC (i.e. all microorganisms in the soil) on plant growth and found that SMC from the three contrasting habitats had differential and substantial effects on plant growth both in the greenhouse and in the field. Furthermore, fast-growing plant species (Brachiaria grass and pioneer forest trees) benefited from "away" (habitats where plant species rarely occur or don't occur at all) compared to "home" (habitats where plant species typically occur) SMC, while slow-growing shade tolerant forest tree species benefited the most from home SMC. I then evaluated how plants were affected by two main components of SMC: potentially mutualistic AMF, and likely antagonistic non-AMF soil organisms. I found that most plants grew significantly better with non-AMF microbes from away, compared to home habitats, while showing limited response to AMF from different habitats. Finally, I tested for plant-soil feedback for both AMF and for non-AMF soil microbes and found that feedbacks driven by AMF were weak, while feedbacks driven by non-AMF soil microbes were significantly negative. Furthermore, feedbacks were only significant for non-native species Brachiaria grass and coffee, while being weak for forest tree species. Together, these results suggest that plant-soil dynamics have been severely disrupted with the replacement of tropical forest for agriculture, and advocate for future studies on non-AMF soil microbes, which have significant impacts on plant communities but remain largely understudied.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by M Camila Pizano.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Kitajima, Kaoru.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-06-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043714:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043714/00001

Material Information

Title: Feedback Dynamics between Plants and Soil Microorganisms in a Fragmented Landscape in the Tropical Andes
Physical Description: 1 online resource (150 p.)
Language: english
Creator: Pizano, M Camila
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: amf -- andes -- colombia -- ecology -- feedback -- forests -- pastures -- pathogens -- plantations -- plants -- soil
Biology -- Dissertations, Academic -- UF
Genre: Botany thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Studies addressing soil microbial communities (SMC) have mostly emphasized particular soil organisms such as arbuscular mycorrhizal fungi (AMF) or soil pathogens within a single habitat. The main goals of my dissertation were to better understand how the interaction between plants and SMC vary across habitats of contrasting abiotic and biotic conditions, and address the role of whole SMC as well as distinct components of SMC. My study took place in the Central Cordillera in Colombia, which is a heterogeneous agricultural landscape comprised of tropical pre-montane forest fragments embedded in highly fertilized, agricultural monocultures, mainly sun-grown coffee plantations and pastures. I worked with SMC, and plants that are common in three contrasting habitats: pastures (Brachiaria grass), coffee plantations (coffee), and forest fragments (forest tree species). I first tested for the effects of whole-SMC (i.e. all microorganisms in the soil) on plant growth and found that SMC from the three contrasting habitats had differential and substantial effects on plant growth both in the greenhouse and in the field. Furthermore, fast-growing plant species (Brachiaria grass and pioneer forest trees) benefited from "away" (habitats where plant species rarely occur or don't occur at all) compared to "home" (habitats where plant species typically occur) SMC, while slow-growing shade tolerant forest tree species benefited the most from home SMC. I then evaluated how plants were affected by two main components of SMC: potentially mutualistic AMF, and likely antagonistic non-AMF soil organisms. I found that most plants grew significantly better with non-AMF microbes from away, compared to home habitats, while showing limited response to AMF from different habitats. Finally, I tested for plant-soil feedback for both AMF and for non-AMF soil microbes and found that feedbacks driven by AMF were weak, while feedbacks driven by non-AMF soil microbes were significantly negative. Furthermore, feedbacks were only significant for non-native species Brachiaria grass and coffee, while being weak for forest tree species. Together, these results suggest that plant-soil dynamics have been severely disrupted with the replacement of tropical forest for agriculture, and advocate for future studies on non-AMF soil microbes, which have significant impacts on plant communities but remain largely understudied.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by M Camila Pizano.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Kitajima, Kaoru.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-06-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043714:00001


This item has the following downloads:


Full Text

PAGE 1

1 FEEDBACK DYNAMICS BETWEEN PLANTS AND SOIL MICROORGANISMS IN A FRAGMENTED LANDSCAPE IN THE TROPICAL ANDES By CAMILA PIZANO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Camila Pizano

PAGE 3

3 Para mi abuelita linda que me ens e a cuidar y amar las plantas, para mis paps que me han mostrado los parasos m s maravillosos de mi pa s y me han apoyado sin limites y para Dr. Hugh Popenoe, quien me inspir y me di su incondicional apoyo desde el principio, pero sin alcanzar a ver el resultado final (To my lovely grandma who taught me how t o care for, and love plants, for my parents w ho hav e shown me the most wonderful paradises in my country and have given me limitless support and for Dr. Hugh Popenoe, who not only inspired, but also greatly supported me since the beginning, sadly no longer being here to see the final result )

PAGE 4

4 AC KNOWLEDGMENTS I am deeply grateful to many people who contributed to the suc cessful completion of my PhD Without their great support, encouragement, sense of humor, love, and understanding during these seven years I would have never finished this disser tation. Before thanking all the people that helped me during this long process, I want to acknowledge my source s of funding, without which I would have never been able to do my PhD: the Anthony Dexter Fellowship ( Department of Wildlife Ecology and Conserva tion, UF), Grinter Fellowship (Graduate School UF) Compton Foundation, and Aerolineas Aeropobre. I especially thank Jorge Botero and Gabriel Cadena in Cenicaf, who gave me un conditional support and total freedom to do my project while at Cenicaf Than ks to their exceptional support, encouragement, and friendship, I had neither constrains or limitations in develop ing and successfully completing my dissertation research. I thank Gabriel Cadena for his extraordinary scientific vision and great intel lect which greatly inspired me not just during my PhD but also for my professional life. Talking to him is always a wonderful inspirational experience. I thank Jorge Botero for going out of his way to make my project possible. Jorge not only gave me extraordina ry support but he also became an unconditional friend who made me feel at home during my three years in Maniz ales, and showed me some of the most beautiful places of the Colombian Central Andes. I am deeply indebted t o Scott Mangan, who has been my scient ific mentor and a great friend for ten years. inspiration, sense of humor, and critical thinking have hugely influenced my scientific career since I was an undergrad uate student I deeply thank Scot t for his infinite

PAGE 5

5 patience and willingness to help me even during his busiest times. He has been an extraordinary mentor and friend. I am also deeply indebted to Kaoru Kitajima, my advisor and committee chair, for her great supp ort and guidance during my PhD. I especially thank her for trusting me and for bein g willing to be my advisor at very hard moments during my PhD. She persuaded me not to quit and it is thanks to her that I am now successfully completing my degree. I also greatly thank Kaoru for her guidance, her encouragement, and all the time she spent correcting my writing. I especially thank her for not only giving me complete freedom to develop my project, but also go ing out of her way to participate in research not entirely relate to her field o f expertise. I will always ad mire her integrity as a person, and her professionalism as a scientist. I am very grateful to my committee members Hugh Pop enoe and Jim Graham. Hugh became a wonderful friend after I took his tropical soils class, and then beca me the biggest su pporter of my PhD project. I will b e fore ver indebted to him for the extraordinary encouragement he always gave me; it is also thanks to him that I did not quit my PhD at very hard times. Hugh was a great scientist and a magnifice nt and ve ry knowledgeable person, and it is very sad that he is no longer with us to see the end result of my PhD. Jim has also been a great supporter of my project, and a source of inspiration and great advice in my scientific career. I greatly thank him for all h is guidance and input and for always being willing to come all the way from Lake Alfred to Gainesv ille to talk about my research. I would also like to thank T ed Schuu r and Brian Silliman who were wiling to be part of my committee. Their feedback was extr emely helpful for the completion of my

PAGE 6

6 dissertation, and we had great scientific discussion that impr oved my final product. I deeply thank them for their open mindedness and willingness to get involved in a project that was not entirely related to their fi elds of expertise. S everal prof e s sors at UF have greatly contributed to my professional development, despite not being part of my doctoral committee. Jack Ewel, with whom it is always interesting and encouraging to talk, for his great knowledge on tropical ecology and everything else in general. Jack Putz, for advising me during the first three years of my PhD. Claudia Romero, for not only being a great colleague and a source of inspiration, but for also being a great friend. I greatly thank Claudia for mak ing me feel at home since the moment I arrive d in Gainesville an d opening the doors of her home to me. I also thank Jamie Gillooly for his great friendship and encouragement and for making me laug h every week. Scott Robinson, for opening the doors of his lab to me from the beginning of my PhD, although I have no idea about tropical birds. I greatly thank him for his unconditional support and for making me feel like part of his great lab. Michelle Mack, for inviting me to go to Alaska o n an amazing trip! Ge tting to know the tundra after working in the tropics for all my life was awesome! There are many friends at UF who helped me finish the PhD by supporting me, keeping me company, making me lau gh, going out to dance with me, and encouraging me not to give u p. Joe Veldman and Lin Cassidy who started out being the greatest neighbor s, an d became my great friends. My beloved friends Gustavo Londoo, Elena Ortiz, Juan Pablo Gmez, Maria Cristina Carrasquilla, Andres Baron, Jonathan Myers, Silvia Alvarez, Ari Mar tinez, Judit Ungvari Martin, Julie Allen, Jessica Oswald, Jordan Mayor, Juan Manuel Jordn, Gerardo Celis, Gab y Hernandez, Franklin Paniagua,

PAGE 7

7 Santiago Espinosa Marisa Tohver, Christine Lucas, Onil Banerjee, Nuria Kaiser, Lorena Endara, Marvin Morales, Ana Eleuterio, and Xavier Haro made my life really happy and I will always thank them for that. I especially want to thank Jose Castao for being the best, most lovely and e ncouraging companion during the final stage of my PhD. I also really want to thank El isa Live ngood, and Mengmeng Zhu for being great TA companions; it was great teaching with you! And my lab mates Matt Palumbo, Martijn Slot, Danielle Pallow, and Gerardo Celis for their great company and for making me l augh every day in the office. Many peo ple in Cenicaf helped with my project and kept me company while in Manizales. I esp ecially want to thank the families Arango Tobn and Gutierrez Botero for adopting me and making me feel at home during the three years I spent in the central coffee region. I have no words to thank Jorge Enrique Arango, Marta Tobn and Isabella Arango for their extraordinary hospitality and for ta king me in their beautiful home and wonderful family. Likewise, I can not thank enough Berta Botero de Gutierrez Julin Gutierrez Maria Mercedes Londoo, Miguel Gutierrez, Luz Maria Botero Maria Isabel Gutierrez, and Jorge Salazar for also taking me into their wonderful family and inviting me to their spectacular farm, San Felipe, every weekend. Berta is one o f the most magnificen t people I ha ve met and spending time with her was always a great privilege and a unique learning experience. Gloria Lentijo, Carmenza Bacca, Carolina Aristiz bal, Lina Snchez, Nestor Franco, Roco Rodrguez and Alejandro Berro were great company and hel p while working at Cenicafe. I also want to thank Don Hctor Vargas and Robeiro Cano who were my exceptiona l field assistants and without whom I would not have been able to set up the huge field experiment in my

PAGE 8

8 project. Likewise, I greatly want to thank Carlos Rivillas, Alvaro Gaitn, Narmer Galeano, Carmenza Gngora, Pablo Benavides, and Carlos Mario Ospina, for being my scientific mentors while at Cenicaf and always providing great conversation. Finally, I thank very much Lucero Arias, Luz Marina Benav ides, Cruz Daz, C ar l os Zuluaga, Carlos Gonzales Juan Carlos Garca Harold Cardona and Carlos Alberto Ospina for all the huge help in the logistics of my project. Finally, I wish to express my heartfelt gratit ude to my parents, Pablo Pizano and Maria Lu ca Gmez, and my brother, Francisco Pizano, for their immense and unconditional support and great love without which I would have never finished my PhD. I especially thank my mom for being bold enough to help me do field work in the middle of the heat, t he mud and the clouds of mosquitoes. I will never forget how much she helped me in harvesting the ridiculously huge grass plants I had i n the field trained us well on ho w to fight against giant squids! And I wi ll be for ever thankful to my dad, who fle w every time he could to save me from Manizales by taking me to our beautiful paradise farms where I recovered energy to continue with my insane PhD.

PAGE 9

9 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 11 LIST OF FIGURES ................................ ................................ ................................ ........ 13 ABSTRACT ................................ ................................ ................................ ................... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 17 2 GREENHOUSE AND FIELD EVIDENCE FOR THE ROL E OF PLANT SOIL INTERACTIONS IN DRIV ING PLANT AND SOIL M ICROBIAL COMMUNITY COMPOSITION ACROSS C ONTRASTI NG HABITATS ................................ ........ 22 Summary ................................ ................................ ................................ ................ 22 Background ................................ ................................ ................................ ............. 23 Methods ................................ ................................ ................................ .................. 26 Study Site and Species ................................ ................................ .................... 26 Greenhouse Experiment: Response of Pasture Grass, Coffee, and Forest Trees to SMC From Home and Away Habitats ................................ ............. 27 Field experiment: Response of pasture grass, coffee, and forest trees to fungicide treatment in home and away habitats ................................ ............ 30 Results ................................ ................................ ................................ .................... 33 Greenhouse experiment: Response of pasture grass, coffee, and forest trees to SMC from home and away habitats ................................ ................. 33 Field experi ment: Response of pasture grass, coffee, and forest trees to fungicide treatment in home and away habitats ................................ ............ 35 Discussion ................................ ................................ ................................ .............. 38 Abio tic vs. biotic variation across different habitats and the response of plants to soil microbial communities from these habitats .............................. 39 Net effects of soil microbial communities on plant performanc e: mutualist vs. antagonistic soil microbes ................................ ................................ ........ 42 Ecological implications ................................ ................................ ..................... 44 3 IS SOIL AT HOME MORE BITTER THAN SOIL FROM AWAY? HOS T SPECIFIC EFFECTS OF SOIL MICROBIAL COMMUNITIES FROM AGRICULTURAL AND NATURAL HABITATS ................................ ....................... 57 Summary ................................ ................................ ................................ ................ 57 Background ................................ ................................ ................................ ............. 58 Methods ................................ ................................ ................................ .................. 61 Study site and species ................................ ................................ ..................... 61 Seed preparation and soil medium for gr owing seedlings ................................ 62

PAGE 10

10 Initial inoculum collection ................................ ................................ .................. 63 AMF inoculum amplification ................................ ................................ .............. 63 Filtrate inocula ................................ ................................ ................................ .. 65 Greenhouse experiment growth conditions and treatments ............................ 65 Statistical analyses ................................ ................................ ........................... 66 Results ................................ ................................ ................................ .................... 68 Discussion ................................ ................................ ................................ .............. 70 AMF from contrasting habitats have similar effec ts on most plant species ....... 71 Non AMF soil microorganisms (microbial filtrate) from contrasting habitats had different effects across plant species ................................ ..................... 73 Applied ecological implications ................................ ................................ ......... 76 4 NEGATIVE FEEDBACK DOMINATES FOREST FRAGMENTS AND AGRICULTURAL LANDS IN THE TROPICAL ANDES. ................................ ......... 88 Summary ................................ ................................ ................................ ................ 88 Background ................................ ................................ ................................ ............. 89 Methods ................................ ................................ ................................ .................. 93 Stud y site and species ................................ ................................ ..................... 93 Seed preparation and soil medium for growing seedlings ................................ 94 Initial inoculum collection ................................ ................................ .................. 95 AMF inoculum amplification ................................ ................................ .............. 95 Filtrate inocula ................................ ................................ ................................ .. 96 Greenhouse experiment growth c onditions and treatments ............................ 97 Statistical analyses ................................ ................................ ........................... 98 Results ................................ ................................ ................................ .................... 99 Discussion ................................ ................................ ................................ ............ 102 AMF vs. non AMF soil organisms in driving feedback ................................ .... 103 Feedback across habitats of contrasting diversity and ab undance of native, non native, invasive, and not invasive plant species ................................ ... 104 Applied ecological implications: ................................ ................................ ...... 107 5 CONCLUSION ................................ ................................ ................................ ...... 123 APPENDIX ILLUSTRATIONS OF STUDY SPECIES ................................ ................................ ..... 126 LIST OF REFERENCES ................................ ................................ ............................. 136 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 149

PAGE 11

11 LIST OF TABLES Table page 2 1 General characteristics of the eleven plant species used in the greenhouse and field expe riments.. ................................ ................................ ....................... 46 2 2 Charac teristics of the farms and sites where the soil inocula for the greenhouse experiment were collected. ................................ ............................. 47 2 3 Mean soil pH, organic matter (OM) content, nutrient content, and light level of pastures (P), coffee plantations (C), and forest fragments (F) where soil inocula were collected. ................................ ................................ ....................... 48 2 4 ANO VA results for relative growth rate (RGR) in the greenhouse of two crop plant species and eight forest tree species. ................................ ........................ 49 2 5 Percentage survival in the field of seed lings of two crop plant spec ies and four forest tree species ................................ ................................ ...................... 49 2 6 Seedling survival in the field of two crop plant species and four forest tree species. ................................ ................................ ................................ .............. 50 2 7 Seedling growth in the field of two crop plant species and three forest tree species. ................................ ................................ ................................ .............. 51 3 1 General characteristics of the eleven plant species used in the greenhouse experiment.. ................................ ................................ ................................ ........ 78 3 2 Characteristics of the farms where the initial soil inocula were were collected. .. 79 3 3 Mean soil pH, organic matter (OM) co ntent, and nutrient content of pastures (P), coffee plantations (C), and forest fragments (F) where soil inocula were collected. ................................ ................................ ................................ ............ 80 3 4 Growth response of 11 plant species to seven inoculum ty pes in the greenhouse. ................................ ................................ ................................ ........ 81 3 5 Grow th response of 11 plant species to inoculum source (pastures, coffee plantations, and forest fragments), inoculum type (AMF, Fil), and plant species (and inte ractions) in the greenhouse. ................................ .................... 81 3 6 F and P values of a priori contrasts examining growth of four plant species groups across the 6 inocula used in the experiment with respect to the control t reatment (sterilized soil). ................................ ................................ ........ 82

PAGE 12

12 3 7 F and P values of a priori contrasts examining growth of four plant species groups with either AMF or a microbial filtrate from their home com pared to that o f away habitats ................................ ................................ .......................... 82 3 8 AMF proportion root colonization of 11 plant species as affected by inoculum source (pastures, coffee plantations, and forest fragments), inoculum type (AMF, Filtrate), and plant species (and interactions) in the greenhouse. ............ 83 4 1 General characteristics of the ten plant species used in the greenhouse experiment.. ................................ ................................ ................................ ...... 109 4 2 Characteristics of the farms where the initial soil inocula were collected. ......... 110 4 3 Mean soil pH, organic matter (OM) content, and nutrient content of the three ha bitats where soil inocula were collected. ................................ ....................... 111 4 4 Growth response of 10 plant species to sterile (non inoculated; control), AMF (isolated from the rhizosphere of these 10 plant species), and m icrobial filtrate (isolated from the rhizosphere of these 10 plant species) inocula in the greenhouse. ................................ ................................ ................................ ...... 112 4 5 Growth response of 10 plant species to AMF isolated from the rhizosphere of these ...................... 112 4 6 Growth response of 10 plant species to a microbial filtrate isolated from the greenhouse. ................................ ................................ ................................ ...... 113 4 7 Statistical summary of the strength of a priori contrasts testing AMF feedback between Brachiaria grass and other 9 plant species, coffee and other 9 plan t species, and average feedback between 8 forest species, Brachiaria grass, and coffee. ................................ ................................ ................................ ........ 114 4 8 Statistical summary of the strength of a priori contrasts testing filtrate feedback between Brac hiaria grass and other 9 plant species, coffee and other 9 plant species, and average feedback between 8 forest species, Brachiaria grass, and coffee. ................................ ................................ ............ 115 4 9 Proportion of roots colonized by AMF for 10 host plant species a s affected by inoculum species, inoculum type and plant species in the greenhouse. .......... 116 4 10 Proportion of roots colonized by of no n AMF for 10 host plant specie s as affected by inoculum species inoculum type and plant species in the greenhouse. ................................ ................................ ................................ ...... 116

PAGE 13

13 LIST OF FIGURES Figure page 2 1 Contrasts of seedling gro wth when inoculated w ith soil microbial communities in the greenhouse experiment and when grown in the field quantifying the species rarely occur or don not occur at all) for each ex perimental treatment. ... 52 2 2 Effect of soil sterilization in the greenhouse experiment and fungicide tre atment in the field experiment on the growth of seedlings grown with SMC ................................ ................................ ........ 54 2 3 Relative growth rate in the greenhouse of four species groups when inoculated with soil microbial communities from pastures (P), coffee plantations (C), and forest fragments (F). ................................ ........................... 55 2 4 Gr owth in the field of four species groups in pastures (P), coffee plantations (C), and fores t fragments (F). ................................ ................................ ............. 56 3 1 Flow diagram showing the procedure that we used to produce the inocula for setting up the experiment in the greenhouse. ................................ ..................... 84 3 2 Effects of soil microbial presence (AMF or filtrate) relative to sterilized soil for growth of Brachiaria grass coffee forest pioneer trees, and forest sh ade tolerant species ................................ ................................ ................................ .. 8 5 3 3 Effects of inoculation with AMF pioneer tree s, and forest shade tolerant species ................................ ................ 86 3 4 Proportion AMF root colonization in seedlings of Brachiaria grass, coffee, forest pioneer trees, an d forest shade tolerant species when inoculated w ith AMF or a microbial filtrate from pastures coffee plantations, or forest fragments in the greenhouse.. ................................ ................................ ............ 87 4 1 Flow diagram showing the procedure that we used to produce the inocula for se tting up the experiment in the greenhouse. ................................ ................... 117 4 2 Proportion of seedling biomass when inoculated with AMF or a microbial filtrate in comparison to biomass when grown on sterile soil for Brach iaria grass, coffee, and 8 forest tree species. ................................ ........................... 118

PAGE 14

14 4 3 Examples o f significant negative feedback and no feedback between seedlings of different species. The graphs represent growth of two spe cies with a microbial filtrate collected from conspecific versus heterospecific seedlings. ................................ ................................ ................................ ......... 119 4 4 Fee dback mediated by AMF and a microbial filtrate between the pasture grass Brachiaria brizan tha and nine other plant species, coffee and nine other plant speci es, and 8 forest tree species and Brachiaria grass, or coffee, or other seven heterospecific forest species. ................................ ................... 121 4 5 Proporti on root colonization of AMF and non AMF soil microbes in seedlings of Brachiaria grass, co ffee, and 8 forest tree species when inoculated with AMF or a microbial filtrate isolated from the roots of these 10 plant species. ... 122 A 1 Illustration of a Brachiaria brizantha seedling by Camila Pizano. ..................... 126 A 2 Illustration of a Coffea arabica seedling by Camila Pizano. .............................. 127 A 3 Illustration of a Cecropia angustifolia seedling by Camila Pizano. .................... 128 A 4 Illustration of a Cecropia telealba seedling by Camila Pizano. ......................... 129 A 5 Illustration of a Ochroma pyramidale seedling by Camila Pizano. .................... 130 A 6 Illustration of a Solanum aphynodendrum seedling by Camila Pizano. ............ 130 A 7 Illustration of a Siparuna aspera seedling by Camila Pizano. ........................... 131 A 8 Illustration of a Retrophyllum rosp igliossii seedling by Camila Pizano. ............. 132 A 9 Illustration of a Garcinia madrunno seedling by Camila Pizano. ....................... 133 A 10 Illustratio n of a Gustavia superba seedling by Camila Pizano. ......................... 134 A 11 Illustration of a Juglans neotropica seedling by Camila Pizano. ....................... 135

PAGE 15

15 Abstract of Dissertati on Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FEEDBACK DYNAMICS BE TWEEN PLANTS AND SOI L MICROORGANISMS IN A FRAGMENTED LANDSCAPE IN THE TR OPICAL ANDES By Camila Pizano December 2011 Chair: Kaoru Kitajima Major: Botany Studies addressing soil microb ial communities (SMC) have mostly emphasized particular soil organisms such as arbus cular mycorrhizal fungi (AMF) or soil pathogens within a single habitat. The main goals of my dissertation were to better understand how the interaction between plants and SMC vary across habitats of contrasting a biotic and biotic conditions, and address the role of whole SMC as well as distinct components of SM C. My study took place in the Central Cordillera in Colombia, which is a heterogeneous agricultural landscape comprised of tropical pre montane forest fragments embedded in highly fertilized, agricultural monocultures, mainly sun grown coffee plantations a nd pastures. I worked with SMC, and plants that are common in three contrasting habitats: pastures (Brachiaria grass), coffee plantations (coffee), and forest fragments (forest tree species). I first tested for the effects of whole SMC (i.e. all micro organ isms in the soil) on plant growth and found that SMC from the three contrasting habitats had differential and substantial effects on plant growth both in t he greenhouse and in the field. Furthermore, fast growing plant species (Brachiaria grass and pionee r forest trees)

PAGE 16

16 not occur at all) growing shade tolerant forest tree species benefited the most from ho me SMC. I then evaluated how plants were affected by two main components of SMC: potentially mutualistic AMF, and likely antagonistic non AMF soil organisms I found that most plants grew significantly better with non AMF microbes from away, compared to ho me habitats, while showing limited response to AMF from different habitats. Finally, I tested for plant soil feedback for both AMF and for non AMF soil microbes and found that feedbacks driven by AMF were weak, while feedbacks driven by non AMF soil microb es were significantly negative. Furthermore, feedbacks were only significant for non native species Bra chiaria grass and coffee while being weak for forest tree species. Together, these results suggest that plant soil dynamics have been severely disrupte d with the replacement of tropical forest for agriculture, and advocate for future studies on non AMF soil microbes, which have significant impacts on plant communities but remain largely understudied.

PAGE 17

17 CHAPT ER 1 INTRODUCTION moist tropical forest is not what it appears. Dissolve away all the plant matter from the dense foliage, giant buttressed trunks, tangled lianas, Gregory S. Gilbert, and Donald R. Strong, 2007 Plant communities consist of populations of different species t hat colonize a specific si te where they persist until becoming locally extinct ( Hubell 2001, V an Andel 2005 ). Accordingly, the presence and abundance of a spec ies in a plant community depend on the availability of propagules and safe sites, and on the abiotic resources (nutrients, water, light) and conditions (climate, soil, pH, human impact) that allow their growth and survival ( V a n Andel 2005). Plant abundance is also modified by a variety of interspecific interactions that structure communities in space and time (Looijen and V an Andel 1999) Competition, facilitation, allelopathy, predation, parasitism, and mutualistic interactions interactively play a role in maintaining and excluding plants from communities at small spatial scal es (alpha or local diversity) and may potentially contribute to larger scale di s tribution patterns (beta diversity or species turn over) Thus, the interplay of these biotic interactions with a biotic factors affects species presence and abundance i n a part icular plant community. The sum of such interaction s are thought to maintain the great diversity of plant communities in tropical forests, where local and regional plant diversity is greater than any other vegetation type (Gentry 1988 Kraft et al. 2011 ) and hundreds of species can coexist within a single hectare (Wright 2002)

PAGE 18

18 Understanding the mechanisms that determine plant community composition in tropical fores ts has been, and continues to be a major challenge for ecologists. Myriad s of hypotheses have been proposed, most of them acknowledging the importance of biotic interactions in providing mechanisms that allow for plant coexistence and maintain diversity ( Dobzhansky 1950, Corner 1954, Asht on 1969 Givnish 1999 Wright 2002 Leigh et al. 2004 Schemske et al. 2009) Yet, most of th e attention of tropic al ecologists has been given to aboveground, rather than belowground, biotic interactions. While there are numerous studies addressing the importance of pollination (e.g., Kiester et al. 1984 Kay et al. 2005) seed dispersal (e.g., Levin et al. 2003 Link and Fiore 2006) and herbivory (Janzen 1970 Connell 1978 Coley 1983 Fine et al. 2004) on maintaining and excluding plants from communitie s at sma ll spatial scales, research on the interaction between plants and belowground soil organisms remains uncommon However recent studies have shown that the same plant soil dynamics that determine plant community composition in the temperate region (V an der Heijden et al. 1998 Klironomos 2002) operate in tropical forests ( Mangan et al. 2010a Mangan et al. 2010 b ) Moreover, advanced molecular techniques have revealed that plants in tropical forests host highly diverse communities of microbial symbionts both below (Aldrich Wolfe 2007) and aboveground (e.g., Arnold et al. 2000 Gilbert and Strong 2007 Gilbert et al. 2007) In my dissertation, I intended to advance understand ing of the role of plant soil microbe interactions in driving local dynamics in tropical plant communities across natural and modified habitats that have contras ting abundance of native and non native plant species, and soil fertility in a fragmented landscape in the tropical Andes.

PAGE 19

19 Soil microbial comm unities are comprised of diverse organisms that establis h direct and indirect relationships with plants along con tinuums of mutualistic to antagonistic and non specific to highly specific interactions (Kuyper and Goede 2004). For example, arbuscular mycorrhizal fungi (AMF) are thought to establish mutualistic relationships with pl ants by improving nutrient acquisitio n in exchange of sugars, but the benefits that these fungi confer to plants vary depending on host genotype and soil conditions (Johnson et al. 1997) Furthermore, it is now well known that the effects of different species of AMF vary ac ross plant species (Munkvold et al. 2004) and that plant species show association with different strains of AMF, even though host specificity is generally weak (Bever 2002a Klironomos 2003 Jansa et al. 2008 Mangan et al. 2010a) Similarly, plant pathogens are detrimental organisms that consist of highly diverse arrays of organisms ranging from opportunistic endophytes apparently harmless to non stre ssed plants (e.g. fungal endophytes that cause disease when plants are stressed), to extremely aggressive and destructive generalist pathogens that can wipe out entire plant communities (Gilbert 2005). Other soil microorganisms such as root fungal endophyt es (Rodriguez et al. 2009) and soil bacteria (Hu guet and Rudgers 2010) can also have significant effects on plants, but their role in plant communities have been less studied. Given that particular soil organisms and mixtures of soil organisms can have substantial and differential effects across plant species, the composition of communities of soil organisms may determine plant community composition ( Grime et al. 1987 Bever 1994 VanderPutten and Peters 1997 Hartnett and Wilson 2002 Klironomos 2002 Reynolds et al. 2003 Van der Heijden 2003 Mangan et al. 2010b) Two f eedback

PAGE 20

20 mechanisms, negative and positive, have been proposed by which soil organisms and plants interact, contributing to determine the diversity of plants in communities (Bever 1994 Bever et al. 1997 Bever 2002a) Negative feedbacks occur when plants associate with a rhizosphere that is less beneficial or more detrimental to them selves than to neighboring plants of different species (heterospecific neighbors) As a conse quence of such build up of detrimental organisms, plant species may indirectly enhance the establishment of other plant species that are less susceptible to those soil organisms. Such negative feedback if common, should limit growth and local dominance of each plant species, and contribute to the maintenance of plant diversity wi thin the local community. N egative dens ity dependency due to host specialized natural enemies has been shown, in temperate (Packer and Clay 2000 Packer and Clay 2003 Casper et al. 2008 Peter mann et al. 2008 Harrison and Bardgett 2010) and tropical forests (Augspurger 1983 Wright 2002 Hood et al. 2004 Bell et al. 2006 Webb et al. 2006 Comita et al. 2010 Mangan et al. 2010b) Even organisms expected to be mutualists such as AMF may act as parasites and also generate negative plant soil feedbacks (Bever 2002b) Positive feedbacks occur when plants accumulate a rhizosphere microbial community that promotes their growth more than that of other plant species, thereby locally out performing other species. As a result, increase s in abundance of such plant species should further promote their abundance and suppress neighbor abundance leading t o a decrease in local plant community diversity For example, some non native species have been shown to have escaped their soil enemies in their native ranges and consequently positive feedback in their new ranges

PAGE 21

21 promote their invasive behaviors (e.g., Klironomos 2002 Reinhart et al. 2003 Callaway et al. 2004 Wolfe and Klironomos 2005 Reinhart and Callaway 2006 V an Grunsven et al. 2007 Vogelsang and Bever 2009) At the same time, positive feedbacks can also promote habitat partitioning of ecologically similar species across environm ental gradients, contributing to beta diversity at larger spatial scales. For example, Pizano et al. (2010) showed that positive feedback of two cryptic plant species and their associated soil microbial communities was mediating habitat segregation and inc ipient speciation of these two tree species in Panama. Plant soil dynamics have been widely studied in natural habitats and for particular types of ecosystems such as temperate grasslands (e.g., Casper et al. 2008 Petermann et al. 2008 Harrison and Bardgett 2010 Wagg et al. 2011) and old fields (e.g., Klironomos 2002 Kardol et al. 2007 Schnitzer et al. 2011 de Voorde et al. 2011) Yet we still have a poor understanding on how plant soil dynamics vary with variations in abiotic conditio ns within biological communities, and across heterogeneous habitat types of human modified ecosystem s. This is particularly true in the tropics, wh ere we are just starting to uncover the role of SMC in natural habitats, but not in human modified landscapes. Thus, the purpose of my dissertation is to better understand how plant soil interactions vary across natural habitats of high plant diversity (fo rest fragments) and managed habitats with high soil fertilization and low plant diversity (pastures and coffee plantations) in a fragmented landscape in the tropi cal Andes. In addition, I want to address the role of distinct components of SMC, namely AMF v s. non AMF soil microbes, in driving plant soil dynamics across these habitats.

PAGE 22

22 CHAPTER 2 GREENHOUSE AND FIELD EVIDENCE FOR THE ROLE OF PLANT SOIL INTERACTIONS IN DRIV ING PLANT AND SOIL M ICROBIAL COMMUNITY COMPOSITION ACROSS C ONTRASTING HABITATS Summary S oil microbia l communities (SMC) have been shown to influence plant community composition and driv e key ecosystem processes in a wide range of environments, but we st ill have a poor understanding of how plant soil dynamics vary acr oss abiotic and biotic env ironmental heterogeneity In this study we examined the prediction that SMC would differ across habitats with contrasting soil fertility and plant community composition. Our general expectation was that soil microbes that have negative effects accumulate i n each community depending on the different plant species that are abundant in each habitat. In particular, we predicted that plants sho uld benefit more from habitats ( cur at all) compared to habitats ( where species typically occur and thus accumulate detrimental soil microorganisms) To test these predictions we compar e d the effects of whole soil microbial communities from highly fertilized, low diversity agricultural monocultures, and unmanaged, highly diverse pre montane tropical forest fragments on growth of plant species abundant in each of these habitat types both in the greenhouse and in the field. In addition, the impacts of SMC were a lso ass essed relative to complete elimination (soil sterilization in the greenhouse) and reduction (fungicide application in the field) of SMC on plant performance. We found that SMC from contrasting habitats had differential and significant effects on plant growth both in the greenhouse and in the f ield (confirmed by a significant effect of soil sterilization and fungicide addition), and different plant groups varied in their response to these SMC. Fast growing species

PAGE 23

23 (Brachiaria grass and forest tree pioneers) benefited from away compared to home S MC, while shade tolerant forest tree species and coffee benefited the most from home SMC. Combined, these results suggest that SMC is significantly modified in agricultural monocultures with low species diversity, and these different SMC have substantial n egative impacts on both crop and forest plant species, providing further evidence for the complexity and importance of plant soil interactions across heterogeneous habitats Background Symbiotic associations are important drivers of many ecological and evo lutionary processes. A central question in community ecology is how these interactions vary across heterogeneous environments ( Thrall et al. 2007) as symbiotic organisms must adapt to one another and to diverse abiotic environmental conditions ( Johnson et al. 2010) For example, it might be crucial to understand how key symbiotic int eractions are impacted in ecosystems that are modified by humans at unpre cedented rates and scales ( Bascompte 2009 Kiers et al. 2010) Plants are constantly interacting with, and being influenced by, highly diverse soil microbial communities (SMC) comprised of neutral, mutualistic and antagonistic soil microbes. Theoretical models predict that increasing resource supply promotes parasitic or pathogenic microbes over mutualistic ones (Thrall et al. 2007). For instance, nutrient enrichment may cause host plants to decrease resource allocation to their rhizosphere partners, shifting th e comp etitive balance among microbes to favor a ntagonistic microbes (Thrall et al. 2007). Previous studies on how AMF mutualism is affected by soil fertilization have shown that AMF in fertilized soils can become parasitic (i.e. acquire plant resources wit hout providing any benefit to the host) compared to those present in non fertilized soils (Graham and Eissenstat 1998, Graham and Abbott 2000,

PAGE 24

24 Treseder 2004, Kiers and V an der Heijden 2006). In addition, fertilization may act directly or indirectly on both plants and parasitic pathogens, influencing the outcome of their interactions (reviewed in Ghorbani et al. 2008). For example, Davies et al. (1997) showed that excessive fertilization made plants more susceptible to disease although other studies have sh own conflicting results on the effects of soil fertility on disease deve lopment (Ghorbani et al. 2008). In addition to soil nutrient a vailability, plant diversity also affects the presence of particular microbes in the soil, as there are reciprocal interac tions between plants and soil microbes (i.e. plant species composition affect s the relative abundance of organisms in the soil, and soil microbial communities shape plant community composition) (e.g., Mangan et al. 2010a). For example, there is a lower div ersity of AMF in agricultural monocultures than in natural ecosystems with high plant diversity (e.g., Tchabi et al. 2008, Verbruggen et al. 2010) and SMC including more diverse AMF species can also maintain high plant diversity (Van der Heijden et al. 1998, Van der Heijden et al. 2008) At the local level, plants host unique arrays of mutualistic and antagonist ic soil microorganisms that can either promote or limit the establishment of other plants (Bever et al. 1997) For instance, local abundance of a particular plant species may increase as a result of gre ater bene fits from soil organisms more beneficial to itself than to heterospecific hosts (e.g., Klironomos 2002 Callaway et al. 2004) On the other hand, soil organisms tha t are particularly detrimental to specific plant species can keep the abundance of that species in check and thus contribute to the maintenance of local plant species diversity ( Nijjer et al. 2007 Mangan et al. 2010b).

PAGE 25

25 Evidence on the crucial roles that soil microbial communities play in determining plant community composition ( Van der Heijden et al. 1998, Klironomos 2002, Mangan et al. 2010b ) and ecosystem processes ( Van der Heijden et al. 2008) has increased in recent years, although most studies (done along abiotic and biotic gradients) have focused on either mutual ists or antagonists, and either in natural or managed habitats. Furthermore, l aboratory, experimental and molecular techniques have restricted our ability to study the myriad diversity of soil organisms (calculated to be 10,000 to 50,000 species per gram o f soil; Roesch et al. 2007) constraining studies to the few organisms that we can isolate. For example, there are numerous studies on arbuscular mycorrhizal fungi (AMF) (e.g., Van der Heijden et al. 1998, Bever 2002 b Vogelsang et al. 2006, Mangan et al. 2010a) and soil microbial pathogens (e.g., Zhu et al. 2000, Mitchell 2003). B ut few studies have taken both into account simultaneously, examining the net effects of who le communities of soil microorganisms as they occur naturally (e.g., Mangan et al 2010b). In addition, most studie s comprising SMC are done in greenhouse s therefore the extent to which the results can be extrapolated to field conditions is generally uncer tain (but see Pringle and Bever 2008, Mangan et al. 2010). In this study we intended to better understand plant soi l interactions of whole soil communities from natural and managed habitats with experiments both in a greenhouse and in the field. To do this we compared interactions between plants and SMC from habitats with contrasting soil fertility and plant community composition: highly fertilized, poorly diverse agricultural monocultures, and unmanaged, highly diverse pre montane tropical forest f ragment s in the tropical Andes.

PAGE 26

26 We intended to test 1) if plants respond differently to SMC coming from contrasting habitat s, 2) whether plant species where species typically occur), and 3) the effects of elimination or reduction of SMC by soil sterilization (greenhouse) or fungicide addition (field) to test if SMC from these different habitats had a net positive (driven by mut ualistic soil organisms) or negative (driven by antagonistic soil organisms) effect on plants. T he fungicide b enomyl was used for SMC reducti on in the field experiment, to compare plant growth and survival under normal field conditions (all soil organisms) and when treated with the f ungicide (reduced soil fungi). Methods S tudy Site and S pecies This study was conducted in the agricultural region of the Central Cordillera of with an average annual temperature of 21C, and an annual rainfall of 2550 mm concentrated in two wet seasons (March to July and September to December) (Guzmn Martnez et al. 2006). Soils are Udands (Andisols) (Ortz Escobar et al. 2004). The landscape i s dominated by three c ontrasting habitat types: heavily fertilized, monoculture of sun grown coffee plantations, occasionally fertilized pa stures with low species diversity (mainly African grasses Pennisetum clandestinum, Melinis minutiflora and Brachiaria spp. ), and unmanaged fragments of pre montane tropical forest (Orrego et al. 2004 a ). The forest fragments are usually small (1 30 ha), biologically more diverse than the other two habitats, mostly dominated by early and mid successional plant species with high liana densities. Seedling recruitment of the native species appear ed

PAGE 27

27 poor in the understory, where see dlings of exotic species including Coffea arabica are common (C. Pizano pers. obs.). F orest tree species for the experiments were selected from thos e abundant in forest fragments in the region (Orrego et al. 2004 b ), encompassing a wide range of seed size and life histories (Table 2 1 Appendix A ). Brachiaria brizantha and Coffea arabica were chosen because they dominate pastures and coffee plantati ons respectively in the study region. To sample the high heterogeneity in altitude and climatic conditions in the study region, five farms with similar conditions each containing a pasture, a coffee plantation, and a forest fragment (Table 2 2) were chosen t o collect soil for the greenhouse experiment. These farms were treated statistically as blocks in the field experiment. At each of the 15 sites, five samples of 200 g of mineral soil (5 15 cm in depth) each were taken and pooled in a composite sample from which a subsample was analyzed. Soils across the three habitats had similar levels of pH, organic matter content, and some nutrients (N, K, Mg). B ut forest fragment soils contained only 10 12 % of P in the pastures and coffee plantations and significantly greater Ca content than the latter (Table 2 3). In addition, light levels in pastures and coffee plantations were approximately 30 times higher than in forest fragments (Table 2 3). Greenhouse Experiment: Response of P asture Grass, Coffee, and Forest Tree s to SMC From Home and Away H abitats In this factorial experiment we compared seedling growth of pasture grass, coffee, and eight forest tree species (Table 2 1) inoculated with fresh soil inoculum (i.e. fine roots, rhizosphere soil, and associated biota) sampled from pastures, coffee pla ntations, and forest fragments. In order to test our hypotheses, we grew plants with the same

PAGE 28

28 common sterilized soil (contr olling for soil nutrients) in a single green house environment such that the only varying factor in the experiment was the origin of the SMC inoculum We compare d the growth of each species with SMC from home vs away habitats, in order to test the effects of whole SMC including benefici al and antagonistic soil microb es. In addition, half of the plants w ere randomly selected to receive sterilized inoculum, to 1) account for potential abiotic differences between inocula from different habitats, and 2) compare the growth of each plant species with liv e vs. sterilized inocula to assess if SMC from the differ ent habitats had overall beneficial (i.e. driven by mutualists) or detrimental (i.e. driven by antago nists) effects on plant growth. Initial growth conditions Seeds of the ten plant species (Table 2 1 individually identified hereafter by abbreviations ) w ere surface sterilized (0.6 % sodium hypochlorite for 15 minutes) and germinated in trays containing steam sterilized (for 2hrs) soil 3:2 soil and river sand mixture. The soil (5.4 pH, 0.1 % N, 1.9 % OM Walkley Black colorimetry, 7 mg kg 1 P Bray II Bray K urtz colorimetry, 0.37 mg kg 1 K Ammonium acetate 1N, 3.4 mg kg 1 Ca Ammonium acetate 1N, and 1.1 cmol kg 1 Mg Ammonium acetate 1N) was collected from an open area near the greenhouses in Cenicaf (Colombian National Research Center for Coffee) (Chinchin, Caldas, Colombia), and was classified as an acrudoxic melanudand from Chinchin unit (Ortz Escobar et al. 2004). Inoculum preparation We collected whole soil inoculum that included mineral soil, roots, rhizosphere soil, and associated organisms from 3 h abitat types (pastures, coffee plantations, forest fragments) at each of five farms (blocks) for a total of 15

PAGE 29

29 sampling sites (Table 2 2). At each site, roots from randomly selected plants and rhizosphere soil (5 15 cm in depth) were collected from five r andom locations and were pooled together to use as inoculum in the greenhouse. We combined and ho mogenized the inocula from the five sites for each habitat type. Half of the inoculum was used live (i.e. as brought from the farm), while half of it was steri lized in the autoclave (for 2 hrs) to control for possible abiotic differences across inocula (Mc Carthy Neumann and Kobe 2010). Growth conditions and treatments The same ste rilized soil and sand mixture used for seed germination was used to fill 70 % of t he 1 L volume pots to which an equal quantity (100 mL) of soil inoculum from one of the three habitats either live or sterilized was added Each treatment combination of inoculum source habitat (forest fragments, coffee plantations, pastures) plant speci es (2 crop plant species, 8 forest tree species) sterilization (live, sterile) was replicated in 10 pots. At the start of the experiment, 25 seedlings per species were harvested to estimate initial biomass for each species. Plants in the greenhouse were grown under 20 % light for 130 days and then harvested. All harvested plants were dried at 60 C for three days. Relative growth rate (RGR) was calculated based on initial and total final biomass of seedlings; RGR=[Ln(final biomass) Ln(initial biomass)]/(# of days). We examined roots from a subsample of 3 plants from each treat ment for the presence of AMF. No plants grown with sterilized inocula had signs of AMF colon ization or other soil fungi Statistical analyses The effects of inoculum source, plant s pecies, and inoculum sterilization (and their interactions) on relative growth rate (RGR) were analyzed using a fixed effect three way ANOVA. We then used a priori contrasts to compare seedling

PAGE 30

30 growth among four species groups (Brachiaria grass coffee, pi oneer, and shade tolerant; Table 2 coffee with inoculum from forest fragments) separately for live inocul um and sterile inoculum treatment. We also used a priori contrasts to test for the effects of soil sterilization on the response of each species group to home vs. away inoculum, and to compare the growth of each species group with live vs. sterilized SMC f rom the three habitat types. Statistical analyses were done with JMP version 8.0 (SAS Institute Inc., SAS Campus Drive, Cary, NC USA 27513). We used the Dunn Sidak correction to adjust the significance levels of contrasts. Field experiment: Response of p asture grass, coffee, and forest trees to fungicide treatment in home and away habitats In this experiment we intended to test for the effects of SMC on plant growth under field conditions, where the e ffects of SMC may be masked by large variations in env ironmental factors including light, soil resource availability, and competition. W e selected a subset of the species used in the greenhouse experiment (Table 2 1, including the two crop species, two pioneer species, and two shade tolerant forest tree speci es) and transplanted their seedlings to agricultural lands and forest fragments, treating h alf of them with the fungicide b enomyl which reduces the activity of soil fungi. In order to test our hypotheses, we 1) compared the growth of each species in home vs away habitat types, and 2) compared the growth of each species between control and fungicide treatment s to test if SMC from these habitats had overall beneficial, detrimental, or neutral effects on plant growth.

PAGE 31

31 Initial growth conditions Seeds of the s ix plant species were surface sterilized (0.6 % sodium hypochlorite for 15 minutes) and germinated in trays containing the same steam sterilized soil 3:2 soil and river sand mixture that we used for the greenhouse experiment. Prior to transplanting to the field, seedlings had been grow n for two months (Cof and forest sh ade tolerant), one month (forest pioneer), or two weeks (Bra chiaria grass ) to allow small seeded seedlings to acquire a similar size to that of other species at the time of transplanting Fie ld surveys and site preparation We selected four of the five farms (blocks) previously used for soil sampling and inoculum collection for the greenhouse experiment (Table 2 2, Table 2 3). In each of the twelve sites (3 habitat types 4 blocks), we random ly marked ninety 1 1 m plots where seedlings were to be transplanted (one seedling per plot). One month before transplanting the seedlings into the field, we recorded proportion vegetation ground cover and plant species richness (total number of species) for each seedling plot (neighbor species richness), then cleared the area with a machete (only for plots with ground cover > 60 %), and opened a 20 20 cm and 30 cm deep hole in the soil. Neighbor species ric hness was measured because plant neighbors str ong ly influence the composition of SMC and its impacts on the focal plant (e.g., Mummey et al. 2005). Plots were randomly assigned to either control (C) or fungicide (F) treatment, and the holes of the F plots received 1 L of the fungicide b enomyl (1 [(But ylamino) carbonyl] 1H benzimidazol 2 yl] carbamic acid methyl ester) at a concentration of 1.125 g L 1 (Helgason et al. 2007). We chose this fungicide because it reduces a wide range of soil fungi including many AMF species and pathogenic asco and basidi omycete fungi (Helgason et al. 2007, Nijjer et al. 2007).

PAGE 32

32 Experiment monitoring and harvest Every month throughout the three month experiment fungicide was re applied (same rate as mentioned above), and survival of each seedling was recorded. The amount o f light reaching each seedling was calculated as the proportion of photosynthetic active radiation (PAR) measured above each seedling compared to that of an open site, taken twice during the experiment at diff erent times of the day using a l ight meter (LI 250, Lincoln, NE, USA). After three months all seedlings were harvested (with roots), dried for 3 days at 60 C, and weighed (total dry weight). Statistical analyses The effects of species type (light demanding, shade tolerant), plant species (nested withi n species type), habitat type, and fungicide on seedling survival were analyzed using a proportional hazards model with block, light, log 10 transformed initial leaf area, neighborhood species richness, and proportion vegetation cover as covariates. In addi tion, we compared seedling survival curves for each species in each habitat using the Kaplan Meier method (Fox 2001). A split plot mixed effect three way ANCOVA was used to examine the response of lo g transformed final biomass to block and block habitat type as random factors, and species, habitat type, and fungicide (and their interactions) as fixed factors. In addition, light, log 10 transformed initial leaf area, neighborhood species richness, and vegetation cover were included as covariates. We then us ed a priori contrasts to compare the g rowth of the four species types (Bra chiaria grass, c of fee forest pioneer, and forest shade tolerant; Table 2 habitats (e.g., coffee in fore st fragments) separately for control and fungicide treatments. In addition, we also used contrasts to compare the effect of fungicide on the

PAGE 33

33 response of each species group to home vs. away habitat types, and to compare the growth of each species under the control and the fungicide treatment in each of the three habitats. Because Sol a pioneer species, had very high mortality across all habitat types, this species was excluded from the growth analyses. Statistical analyses were done with JMP, version 8.0 (S AS Institute Inc., SAS Campus Drive, Cary, NC USA 27513). We used the Dunn Sidak correction to adjust the si gnificance levels of contrasts. Results Greenhouse experiment: Response of pasture grass, coffee, and forest trees to SMC from home and away habitat s Inoculum source, plant species, and inoculum sterilization all significantly affected the growth of seedlings (Table 2 4). In addition, species growth differed across different combinations of inoculum source and sterilization as indicated by significant two way and th ree way interactions (Table 2 4). When inoculated with live soil organisms, Brachiaria grass grew marginally less with inoculum from home (pastures) than with inoculum from away habitats (coffee plantations and forest fragments) (F 1,526 = 5. 8; P = 0.017; Fig. 2 1A), however growth of this species was similar between home and away habitats for sterilized inoculum (F 1,526 = 1.8; P = 0.18; Fig. 2 1 A). Thus, soil sterilization eliminated the response of Brachiaria grass to SMC from home compared to away habitats indicating a strong home disadvantage with live inoculum relative to sterile inoculum (F 1,526 = 7.0; P = 0.0015) (Fig. 2 2A). plantations and for est fragments) for either liv e (F 1,526 = 3.2; P = 0.074), or sterilized (F 1,526 = 0.006; P = 0.94) inocula. G rowth was significantly better with sterile than with

PAGE 34

34 live inoculum from pastures (F 1,526 = 28.3; P < 0.001) and forest fragments (F 1,526 = 9.1; P = 0.003) (Fig. 2 3A), indicat ing that this species encounters antagonistic soil organisms in the SMC from these habitats. Growth of c offee did not differ significantly between home and away habitats whether inoculum was sterilized or not (P > 0.2 for all contrasts; Fig. 2 1B, 2 2A), a nd between live and sterilized inocula (P > 0.12 for all contrasts) (Fig. 2 3B). Forest woody pioneer species grew significantly less with live inoculum from home (forest fragments) compared to inoculum from away (pastures and coffee plantations) habitats (F 1,526 = 14.3; P < 0.001 ), but did not differ when the inocula were sterilized (F 1,526 = 0.6; P = 0.46) (Fig. 2 1C). Thus, similar to Bra chiaria grass difference between home vs. away SMC was significantly stronger with live than with sterile inoculum a s indicated by the negative a priori contrast (F 1,526 = 10.2; P = 0.0015) (Fig. 2 2A). Notably, growth woody pioneer species did not differ between the two away habitats (pastures and coffee plantations) for live (F 1,526 = 0.01; P = 0.84), or sterilized (F 1,526 = 0.4; P = 0.52) inocula. In addition, these species grew significantly better with sterilized, compared to live inoculum from pastures (F 1,526 = 12.4; P < 0.001), coffee plantations (F 1,526 = 7.6; P = 0.006), and forest fragments (F 1,526 = 51.1; P < 0.001) (Fig. 2 3C), suggesting that SMC from all three habitat types have overall detrimental effect s on this group of species. Finally, shade tolerant species grew better with home (forest fragments) compared to away (pastures and coffee plantations) SMC with both live (F 1,526 = 17.2; P < 0.001) and sterilized (marginally significant; F 1,526 = 4.9; P = 0.03) inocula (Fig. 2 1D). Growth of these species was best with both live and sterilized inocula from forest

PAGE 35

35 fragments compared to SMC from the other two habitats (F 1,526 = 2.0; P = 0.16), and soil sterilization only marginally decreased the home advantage (Fig. 2 2A). This suggests that shade tolerant species found both biotic, and abiotic benefits from SMC from home compared to that of away habitats. In t erms of the effects of inocula from the two away habitat s they grew marginally better with live (F 1,526 = 5.2; P = 0.023) inocula from coffee plantations compared to that from pastures, but there was no significant difference with sterilized inocula (F 1,5 26 = 0.02; P = 0.89) (Fig. 2 3D). Lastly, these species grew significantly better with sterilized than with live inoculum from pastures (F 1,526 = 20.5; P < 0.001), but had similar growth with sterilized and live inoculum from coffee plantations (F 1,526 = 5 .5; P = 0.02) and forest fragments (F 1,526 = 2.9; P = 0.09) (Fig. 2 3D). These results suggest that shade tolerant species encounter more antagonistic soil organisms in pastures and coffee plantations than i n forest fragments. Field experiment: Response of pasture grass, coffee, and forest trees to fungicide treatment in home and away habitats Seedling survival was higher fo r shade tolerant species (Cof Gar Ret ) than for the fast growing Brachiaria grass and the two woody pioneer plant species (CecA Sol) across different habitat types and treatments (Tables 2 5, 2 6). In fact, only 7 seedlings of shade tolerant species died during the experiment. Survival of fast growing species varied across habitat types, as indicated by a significant habitat type spe cies type [species] interaction (Table 2 6). For instance 80 % ( 5.7) of Bra chiaria grass seedlings survived in pastures, while 53.4 % ( 13.4) and 31.6 % ( 9.4) survived in coffee plantations and in forest fragments respectively ( Wilcoxon 2 = 23.7, P < 0.001). Similarly, 83 % ( 4.7) of CecA seedlings survived in pastures, while 59.6 % ( 10.9) and 59.6 % ( 14.2) survived in coffee plantations and forest fragments respectively

PAGE 36

36 ( Wilcoxon 2 =12.0, P = 0.0025). Sol had very high mortality across habita ts, with only 34.6 % ( 7.3) seedlings surviving in pastures, 9.5 % ( 5.2) in forest fragments, and 4.7 % ( 2.5) in coffee plantations (Wilcoxon 2 = 8.4, P = 0.015). Fungicide had a marginal but significant effect on seedling survival, affecting fast gr owing and pioneer species more than shade tolerant species, as indicated by a significant species type fungicide interaction (Tables 2 5, 2 6). In fact, seedlings treated with fungicide had marginally lower survival for CecA (C: 67.9 % 9.3 ; F: 66.9 % 9.0 ), but higher survival for Sol (C: 13.8 % 4.8 ; F: 20 6.4 ) across the three habitat types. Seedling biomass was affected by species, habitat type, and fungicide (Table 2 7). Furthermore, seedling growth differed across different habitat types, and a cross different combinations of fungicide treatment and habitat type, indicated by significant two and three way interactions (Table 2 7). Both control (F 1,33 = 72.5; P < 0.001) and fungicide treated (F 1,35 = 99.1; P < 0.001) seedlings of Brachiaria gras s grew almost 200 fold more in their home habitat (pastures) compared to away (coffee plantations and forest fragments) habitats (Fig s 2 1E 2 4A ), most likely driven by the much higher light availability in pastures (64 1.4 %) and coffee plantations (5 0.4 1.7 %) compared to forest fragments (1.6 0.1 % ). Growth did not differ across habitats with similar light conditions (pasture and coffee plantations) for control (F 1,29 = 3.5; P = 0.072) and fungicide treated seedlings (F 1,33 = 0.95; P = 0.95), sug gesting the overwhelming importance of light on growth rates. Nevertheless, fungicide marginally enhanced the growth increase of this grass in pastures compared to other habitats (F 1,672 = 2.2; P = 0.14; Fig s 2 1E, 2 2B), compatible with the pattern expe cted from the dominance of antagonistic organisms in

PAGE 37

37 its home habitat compared to that of other habitats. Finally, control and fungicide treated seedlings of Bra chiaria grass grew similarly in coffee plantations and forest fragments (P > 0.23 for all contr asts) (Fig. 2 4A). Consistent with the greenhouse experiment, growth of coffee did not differ between home and away habitats across control and fungicide treated seedlings (P > 0.5 for all contrasts; Fig. 2 1F), an d between control and fungicide treated s eedlings across habitat types (P > 0.38 for all contrasts) (Fig. 2 2B, 2 4B). B oth control (F 1,39 = 58.9; P < 0.001) and fungicide treated seedlings (F 1,25 = 24.9; P < 0.001) of pioneer species CecA grew significantly less in home (forest fragments) than i n away habitats (pastures a nd coffee plantations) (Fig. 2 1G), and this response was reduced by the fungicide (F 1,672 = 9.5; P = 0.0021; Fig. 2 2B). Between the two high light habitats (pastures and coff ee plantations), it grew significantly better in coff ee plantations than in pastures (F 1,23 = 18.5; P < 0.001) (Fig. 2 4C ) In addition, control seedlings grew significantly better than fungicide treated seedlings in coffee plantations (F 1,672 = 29.6; P < 0.001), suggesting that CecA not only benefits from t he higher light levels, but also from more beneficial SMC in coffee plantations compare d to forest fragments. Growth of this species did not differ between control and fungicide treated seedlings in pastures (F 1,671 = 0.1; P = 0.74) and forests (F 1,672 = 0 .7; P = 0.42) (Fig. 2 4C). In case of the two shade tolerant forest tree species (Gar and Ret), control (F 1,16 = 0.5; P = 0.5) and fungicide treated seedlings (F 1,17 = 0.001; P = 0.97) grew similarly in home and away habitats (Fig. 2 1H), regardless of fun gicide treatments, across habitat types (P > 0.2 for all contrasts) (Fig. 2 2B, 2 4D).

PAGE 38

38 Discussion In this study we found strong evidence both in the greenhouse and the field that SM C from contrasting habitats had different ial effects on plant growth. O nly live (and not sterilized) inocula brought from agricultural lands and forest fragments significantly differed in their effects on plant growth of most plant species in the greenhouse (Fig. 2 3 A D ), and fungicide had a significant effect on both survival and growth of these species in the field (Table 2 5, 2 6, 2 7). P varied among species both in the greenhouse and in the field suggesting that the composi tion of SMC differed across habitat types which had di fferential effects on plant hosts Furthermore, we found that fast growing species (Brachiaria grass and forest pioneer species ) were negatively impacted by SMC from home compared to that of away habitats, while shade t olerant forest species benefited from SMC from home compared to that of away habitats (greenhouse experiment) (Fig. 2 1). Co mparison of the effects of soil sterilization in the greenhouse and fungicide addition in the field revealed that suppression of SMC had consistent directional effects o n performance of each species across the three different habitats (Fig. 2 2). Additionally we also found that while most species grew better with sterilized than with live soil inoculum in the greenhouse (Fig. 2 3), fungicide did not have a consistent effe ct on plant growth across different habitats in the field (Fig. 2 4). In fact, fungicide actually decreased the growth of pioneer forest species in coffee plantations. Thus, results from the greenhouse experiment suggest that SMC from agricultural lands an d forest fragments can exert net negative effects, while results from the field experiment suggest that SMC in at least some of these habitats are dominated by mutualistic, and not antagonistic, soil microorganisms. Combined, these results provide further evidence for

PAGE 39

39 the complexity and importance of plant soil interactions across habitats with contrasting abiotic characteristics and differe nt plant community composition. Abiotic vs. biotic variation across different habitats and the response of plants to s oil microbial communities from these habitats Previous studies have shown that SM C in managed habitats with low plant diversity and chemical fertilization (i.e. agricultural lands) contain AMF communities that are less diverse (e.g. Bradley et al. 2006, Ve rbruggen et al. 2010) and beneficial (e.g. Graham and Abbott 2000) to plants than those present in natural habitats with higher plant diversity and no fertilization (e.g. Neuhauser and Fargione 2004, Kiers and Denison 2008, Verbruggen and Kiers 2010). Furt hermore, it is widely known from agricultural settings that species specific pathogens accumulate in low diversity monocultures, and that environmental factors such as light and nutrient availability determine how resistant or susceptible plants are to ant agonistic soil organisms (Zhu et al. 2000, Reynolds et al. 2003). Thus, we expected that SMC from habitats with fertilized, species poor plant communities (i.e. agricultural lands) should have an overall negative effect on plant growth compared to SMC from unmanaged, highly diverse natural habitats (i.e. forest fragments). We contrasted this prediction with the hypothesis in these predictions is the likelihood of SMC being dominated by either generalist (first prediction) or specialist (second prediction) soil microorgan isms across different habitats. The results from the greenhouse show that SMC from pas tures and coffee plantations were indeed different from those from for est fragments, as there was no common pattern in the response of plants species across inocula sources (Table 2 4).

PAGE 40

40 However, we found no evidence suggesting that SMC from pas tures and coffee plantations were less beneficial or more detrimental than those p resent in natural forests regardless of the host plant species (Table 2 4). In fact, the net effect of SMC on plant performance was mainly determined by the habitat of origin of both plants and SMC (home vs. away), supporting our second prediction that ho st specialized microbes had a significant ecological role Furthermore, the response of pla nts to different SMC was mainly determined by their life history. Fast growing species (Brachiaria grass and forest pioneer species CecT Ochr and Sol) grew signifi cantly better with whole soil inoculum from away habitats compared to their respective home habitats (Fig. 2 1A, 2 1C), while having similar growth across inocula from away habitats. Moreover, sterilization eliminated the effect of home vs. away inocula on the growth of these species (i.e. no growth difference across sterilized inocula) (Fig. 2 2A), verifying the role of soil organisms in driving their response to soils from different habitats. In contrast to the results for fast growing species, forest sha de tolerant plant species (Sip Gar Ret Gus and Jug) grew significantly better with home, rather than with away SMC (Fig. 2 1D). These results suppo rt previous studies showing that the susceptibility of plant species to soil microorganisms is inversely proportional to seedling shade tolerance (e.g. Augspurger 1983, Zangaro et al. 2003, Kardol et al. 2006, Kardol et al. 2007, McCarthy Neumann and Kobe 2008); fast growing species, which are expected to invest less to defensive traits are likely to be susce ptible to antagonistic soil organisms in their home habitat that are absent in other habitat types. Conversely shade tolerant species, investing heavily in defense with support from large seed reserves, may be less susceptible to antagonistic soil pathoge ns common in their

PAGE 41

41 home habitat (Kitajima 2002) Finally, shade tolerant species grew unexpectedly better with the sterilized inocula from home com pared to away habita ts, even though forest soils had almost ten fold lower P than coffee plantations and pastures (Table 2 3). These species could have benefited from higher organic matter content of forest soils and associated organic N availability (Table 2 3), as nutrient s other than P might be limiting to seedling growth (Fetcher et al. 1996). I n the greenhouse we compared the effects of SMC under the same abiotic environmental conditions, standardizing environmental factors associated with different habitat types in th e field (e.g. soil nutrients and light levels). In the field we intended to test whether modify ing SMC (by adding fungicide) would have significant effects in the light of variability of other environmental factors among habitats. Our habitats had marked d ifferences in soil nutrient levels and light levels; forest fragments had an almost five fold lower soil P content, and 50 fold lower light level than pastures and coffee plantations (Table 2 3). Despite these large differences in abiotic environmental var iables that could overwhelm the potential effects of SMC, we detected a significant effect of fungicide on both plant survival (Table 2 6) and growth (Table 2 7). Furthermore, the effect of fungicide varied across habitats and species, showing the same di rectional interactive effects as in the greenhouse, suggesting that 1) SMC differed across different habitat types, and 2) SMC had differential host specific effects on plant species. Fast growing species (Brachiaria grass and pio neer tree CecA) responde d strongly to different light levels between habitats, therefore grew significantly better in open pastures and coffee plantations than in shaded forest fragments (Fig. 2 1E, 2 1G).

PAGE 42

42 Nevertheless the effect of fungicide on these species suggests that SMC p lay a significant role in determining their performance in the field. For instance, Brachiaria grass grew similarly better across pastures and coffee plantations than in the shade of forest fragments (Fig. 2 1E). Yet, consistent with the greenhouse experim ent, fungicide marginally increased its growth in pastures (Fig. 2 1E, 2 2B), indicating that this grass encounters more antagonistic or less mutualistic soi l organisms in its home habitat of pasture. Likewise, the pioneer species CecA also grew better in pastures and coffee plantations compared to its home habitat (forest fragments), however fungicide significantly reduced the away advantage (Fig. 2 1F, 2 2B) largely because beneficial SMC in coffe e plantations apparently were suppressed (Fig. 2 4D) Des pite showing differences in growth across inocula from contrasting habitat types in the greenhouse (Fig. 2 3D) shade tolerant species Gar and Ret showed no response to different habitat types and fungicide addition (Fig. 2 1H, 2 2B). One possible explana tion for this lack of response is that the field experiment was too short (three months) to detect growth responses in these slow growing species Likewise, growth of coffee did not differ across treatments in either of the two experiments (Fig. 2 1B, 2 1F 2 2A, 2 2B) Furthermore, these seedlings had large s eed reserves whose effects persisted throughout the experiment Net effects of soil microbial communities on plant performance: mutualist vs. antagonistic soil microbes Although there are numerous stud ies on the impacts of SMC on plant performance, few studies have included appropriate control treatments in both the greenhouse and the field to address if the overall net effect of SMC on plant performance is driven by mutualistic (e.g. AMF) or antagonist ic (e.g. soil microbial

PAGE 43

43 pathogens) soil microorganisms (but see McCarthy Neumann and Kobe 2008, Mangan et al. 2010b). Previous studies with proper control treatments have found contrasting results. For example, Van der Put ten and Peters (1997) found the co mpetitive outcome between two successional plant species was driven by parasitic nematodes and pathogenic fungi in an experiment comparing sterilized and not sterilized soil s from coastal sand dunes. Similarly, Reinhart et al. 2005 showed that Prunus serot ina grew better with sterilized and fungicide treated soils compared to untreated soils, confirming a predominant effect of soil antagonists over mutualists. In contrast, Mangan et al. (2010b) found that six shade tolerant tree species grew better with liv e, than with sterilized, whole soil microbial communitie s collected from around parent trees of these species in the field in Panama, suggesting an overall positive effect of these SMC on plant growth. In our study both types of net effects of whole SMC on plant performance occurred in the greenhouse and in the field. In the greenhouse, most plant species grew better with sterilized than with live inoculum from different habitats (indicated by a significant effect of inoculum sterilization; Table 2 4, Fig 2 3), suggesting that SMC from pastures, coffee plantations and forest fragments are dominated by antagonistic soil microorganisms over mutualistic ones This indicates that most plants species encounter generalist antagonistic soil micro organisms across habitat types. At the same time the response of plant species to SMC from home and away habitats suggests that plants also encounter specialist antagonistic soil microorganism s in their respective habitats. S eedlings treated with fungicide in the field h ad marginally higher survival than those not treated with fungicide (Table 2 5, 2 6), although fungicide marginally

PAGE 44

44 decreased growth (Table 2 7). In fact, the only species that showed a significant response to fungicide treatment was the pioneer C. angusti folia which grew better without than with fungicide in coffee plantations (Fig. 2 4). This suggests that SMC fr om coffee plantations might host mutualistic soil organisms that are particularly beneficial for some species such as forest pioneer s pecies. T he differences in results from the greenhouse and the field indicate that soil sterilization and fungicide addition are not analogous treatments. Soil sterilization eliminates all soil biota (and this was evident in the roots from sterile treatments in the greenhouse), while the fungicide b enomyl reduces the abundance of some, and increases the abundance of other, soil fungi. For example, Helgason et al. 2007 showed that plants treated with b enomyl in the field hosted AMF communities with the same number of AMF species as control plants, however the AMF species composition shifted in the fungicide treated plants. In addition, more generalist AMF (those occurring in a wide range of host plants) were more resilient to the fungicide treatment. Thus, the lack of a strong response to fungicide in the field may also reflect that 1) plants show a limited response to changes in soil fungal species abundance in SMC in the field, and 2) agricultural lands and forest fragments in our study region are dominated by genera list soil fungi that were not greatly affected by the addition of fungicide. Results from the greenhouse experiment strongly suggest the first explana tion is the most plausible one. Ecological implications Conversion of natural ecosystems to agricultural l ands decreases both plant diversity aboveground, and the diversity of soil microbial communities belowground (e.g. Hooper et al. 2000, Postma Blaauw et al. 2010, Verbrugeen et al. 2010). S tudies have shown that in addition to low plant diversity, agricultu ral practices such as tillage

PAGE 45

45 (reviewed by Gosling et al. 2006), and fertilization (e.g., Bradley et al. 2006, Rasmann et al. 2009) further decreases the divers ity of belowground communities. H owe ver fewer studies have addressed the potentially important effects of these losses on ecosystem services and plant community composition (e.g., Strickland et al. 2009) We found marked differences in the SMC between managed and natural systems, and significant impacts of soil microorganisms on the performance of all plant species except for coffee Thus, differences in SMC might sig nificantly impact forest restoration (Allen et al. 2005) and agricultural production ( Banwart 2011) Our results suggest that while forest pioneer tree species encounter less antagonistic soil organisms in agricultural lands than in forest fragments (i .e. SMC from agricultural lands, in particular coffee plantations, benefit these spec ies), forest shade tolerant tree species encounter more antagonistic soil organisms in agricultural lands, which can potentiall y hinder their regeneration in these habitats. Furthermore, we found that pastures accumulate antagon istic soil organisms with ne gative impact on the economically important grass B. brizantha In contrast, coffee was not responsive to different SMC although coffee plantations host soil organisms that are particularly beneficial for other plant species. Collecti vely, these findings a dd substantiate on the still poorly understood great complexity and significance of plant soil interactions in both natural and agricultural settings.

PAGE 46

46 Table 2 1 General characteristics of the eleven plant species used in the greenhouse and fie ld expe riments. Seed mass (mean S.D) was measured from 25 50 seeds/species dried for 3 days at 60 C. Species Family Species group Typical habitat Seed dry mass (g) Abbrev iation Brachiaria brizantha + Poaceae Brachiaria grass (crop) P 1 0.0077 0.0013 Bra Coffea arabica + Rubiaceae Coffee (crop) C 2 0.22 0.023 Cof Cecropia angustifolia* Cecropiaceae Pioneer F 3 0.0011 0.00028 CecA Cecropia telealba Cecropiaceae Pioneer F 3 0.00039 0.00012 CecT Ochroma pyramidale Bombacaceae Pioneer F 3 0.0037 0.0012 Ochr Solanum aphynodendrum + Solanaceae Pioneer F 3 0.0015 0.00053 Sol Garcinia madrunno + Clusiaceae Shade tolerant F 3 2.92 0.77 Gar Gustavia superba Lecythidaceae Shade tolerant F 3 10.55 2.78 Gus Juglans neotropica Juglandaceae Shade tole rant F 3 26.04 14.49 Jug Retrophyllum rospigliosii + Podocarpaceae Shade tolerant F 3 0.86 0.18 Ret Siparuna aspera Monimiaceae Shade tolerant F 3 0.014 0.025 Sip *Used in the field but not in the greenhouse experiment. + Used in both the g reenhouse and field experiment 1 Pastures 2 Coffee plantations 3 Forest fragments

PAGE 47

47 Table 2 2. Characteristics of the farms (blocks) and sites where the soil inocula for the greenhous e experiment were collected (Guzmn Martinez et al 2006), and where the field experim ent was set up. At each farm there were three habitats: a pasture (P), a sun exposed coffee plantation (C), and a forest fragment (F). Farm (block) Geographic coordenates Altitude (m.s.l) Mean annual precipitation (mm yr 1 ) Mean annual temperature ( C) Hab itat (site) Habitat patch size (ha) Cenicaf 0500 N 75 36 W 1380 2733 20.9 P 0.5 C 0.5 F 40 .0 Playa rica 0500 N 1290 2750 20.7 P 10 .0 C 60 .0 F 30 .0 Alto espaol* 0456 N 7542 W 1720 3140 18.8 P 1.0 C 2.0 F 0.3 Naranjal 0459 N 7539 W 1400 3137 21.4 P 22 .0 C 38 .0 F 27 .0 La Argentina 0502 N 7541 W 1354 2935 19.9 P 0.5 C 100 .0 F 1.5 This block was not used in the field experiment.

PAGE 48

48 Table 2 3. Mean ( SE) soil pH, organic matt er (OM) content, nutrient content, and lig ht level of pastures (P), coffee plantations (C), and forest fragments (F) where soil inocula were collected. Habi tat pH 1 OM 3 (%) N 2 (%) P 4 (mg/kg)* K 5 (cmol/kg) Ca 5 (cmol/kg)* Mg 5 (cmol/kg) Light (%)* P 5.6 ( 0. 1) 7.2 ( 1.4) 0.3 ( 0.04) 54.4 ( 25.3) 0.6 ( 0.2) 5.6 ( 0.7) 2.1 ( 0.4) 64.1 ( 1.4) C 5.0 ( 0.3) 9.6 ( 1.7) 0.4 ( 0.06) 42.0 ( 16.7) 0.3 ( 0.1) 4.4 ( 1.1) 1.7 ( 0.6) 50.4 ( 1.7) F 5.5 ( 0.2) 12.4 ( 2.1) 0.5 ( 0.06) 5.6 ( 0.4) 0.4 ( 0.05 ) 8.3 ( 2.3) 2.4 ( 0.7) 1.7 ( 0.1) Notes: Log transformed data was analyzed with one way ANCOVA including block as a random factor and habitat as a fixed factor. Light was measured as the percentage of photosynthetic active radiation (PAR) above each o f the 1080 seedlings transplanted to the field (field experiment) co mpared to that of an open site. 1 pH: Potentiometer soil: water 1:1 2 N (total): Calculated 3 OM: Walkley Black colorimetry 4 P: Bray II colorimetry Bray Kurtz 5 K, Ca, Mg: Ammonium ac etate 1N Significant differences between habitat types P < 0.05

PAGE 49

49 Table 2 4. ANOVA results for relative growth rate (RGR) in the greenhouse of two crop plant species and eight forest tree species (4 pioneer and 4 shade tolerant species) (Table 2 1) across different inoculum sources (forest fragments, coffee plantations, pastures), species, and inoculum sterilization (no sterilization live; sterilized), and their interactions. Total biomass (E rror df = 526 ) Source df F P Inoculum source 2 2. 7 0.066 Species 9 1730.3 <0.001 Sterilization 1 100.5 <0.001 Inoculum source Species 18 1.9 0.012 Inoculum source Sterilization 2 3.6 0.029 Species Sterilization 9 11.9 <0.001 Species Sterilization Inoculum source 18 2.8 <0.001 Notes: ANOV A was used to analyze RGR ( RGR=[Ln(final biomass) Ln(initial biomass)]/(# of days)]. Initial biomass was calculated as the average of 25 seedlings/species harvested just before the experiment was set up. Contrasts within the three way interaction tested th e differences between RGR of plants grown with inoculum coming from their Table 2 5. Percentage survival ( SE) in the field of seedlings of two crop plant species (Bra and Cof) and four forest tree species (CecA, Sol Gar and Ret) (species abbreviations in Table 2 1) in each of 12 sites (3 sites one per habitat type per block). Sixteen seedlings from each species (except Ret, for which there were 10 seedlings) were transplanted to e ach site; half served as control and half were treated once a month with fungicide ( b enomyl ). Plant species Fast growing and pioneer Shade tolerant species Habitat Treatment Bra CecA Sol Cof Gar Ret Pastures Control 78.5 ( 9.3) 87.8 ( 5.1) 31.5 ( 8.0) 97 ( 3.0) 97 ( 3.0) 100 ( 0.0) Fungicide 81.5 ( 8.0) 81.5 ( 7.8) 37.8 ( 13.4) 100 ( 0.0) 94 ( 3.5) 100 ( 0.0) Coffee p. Control 53.3 ( 19.4) 62.8 ( 16.9) 9.8 ( 3.3) 100 ( 0.0) 100 ( 0.0) 100 ( 0.0) Fungicide 53.5 ( 21.4) 56.5 ( 16.5) 3.3 ( 3.3) 97 ( 3.0) 100 ( 0.0) 100 ( 0.0) Forests Control 31.5 ( 16.5) 53.3 ( 20.7) 0.0 ( 0.0) 100 ( 0.0) 100 ( 0.0) 100 ( 0.0) Fungicide 31.8 ( 12.0) 66.0 ( 22.0) 19.0 ( 8.1) 100 ( 0.0) 100 ( 0.0) 90 ( 5.8)

PAGE 50

50 Table 2 6. Seedli ng survival in the field of two crop plant species (Bra and Cof) and four fores t tree species (CecA, Sol Gar and Ret) (species abbreviations in Table 2 1) as affected by plant species (nested within species type), species type (light demanding, shade to lerant), habitat type (pasture, coffee plantation, forest fragment) and fungicide (control, b enomyl application). Source Df L R Chisquare P Species type 1 99.7 <0.001 Species type [Species] 1 4 70.2 <0.001 Habitat type 2 0.0 1.000 Fungicide 1 4.0 0.046 Species type*Habitat type 2 0.0 1.000 Habitat type* Species type [Species] 1 8 22.5 0.004 Species type*Fungicide 1 4.2 0.041 Fungicide*Species type [Species] 1 4 6.0 0.200 Habitat type* Fungicide 2 4.1 0.130 Habitat type* Fungicide*Species type 2 3.7 0 .160 Block 3 12.6 0.006 Light 1 0.0 0.860 Initial leaf area 1 40.7 <0.001 Neighborhood species richness 1 0.1 0.780 Proportion vegetation cover 1 1.0 0.320 1 Species nested within species type. Notes: A nested proportional hazard model was used to an alyze seedling survival in the field, with block (4 farms), light (proportion of light on each seedling), log 10 transformed initial leaf area, neighborhood species richness (quantified in an 1 1 m area around each seedling transplanted to the field), and proportion vegetation cover (quantified in an 1 1 m area around each seedling transplanted to the field) as covariates.

PAGE 51

51 Table 2 7. Seedling growth in the field of two crop plant species and three forest tree species (CecA, Gar and R et) (species abbreviations in Table 2 1) as affected by plant species, habitat type (pasture, coffee plantation, forest fragment), and fungicide (control, b enomyl application). Total biomass Source Df F P Species 4 160.6 <0.001 Habitat type 2 26.9 <0 .001 Fungicide 1 5.4 0.021 Species *Habitat type 8 108.3 <0.001 Species*Fungicide 4 1.5 0.220 Habitat type *Fungicide 2 3.3 0.036 Species*Habitat type*Fungicide 8 3.1 0.002 Initial leaf area 1 433.6 <0.001 Light 1 5.9 0.015 Neighborhood species ric hness 1 1.7 0.190 Proportion vegetation cover 1 <0.001 0.990 Notes: ANCOVA was used to analyze log 10 transformed final biomass, with block (4 farms where the experiment was set up) as a random factor, and species, habitat type, and fungicide as fixed fac tors. Block*habitat was also included as a random factor to partition the variance although not shown in the tab le. Light (proportion of PA R above each seedling), log 10 transformed initial leaf area, neighborhood species richness (quantified in an 1 1 m area around each seedling transplanted to the field), and proportion vegetation cover (quantified in an 1 1 m area around each seedling transplanted to the field) were included as covariates. The species Sol was excluded from the growth analyses of the f ield experi ment due to its high mortality.

PAGE 52

52 GREENHOUSE EXPERIMEN T FIELD EXPERIMENT Figure 2 1. Contrasts of seedling growth when inoculated with soil mi crobial communities (SMC) in the greenhouse experiment (A D) and when grown in the field (E H) quantifying the effects of

PAGE 53

53 ccur or don no t occur at all) for each experimental tre atment The results were analyzed separately for the four species groups (Table 2 1). Bars indicate the value ( SE) of a priori contrasts examining the growth of each plant species group when grown with SMC from their home habitat compared to when grown w ith SMC from away habitats. Positive values indicate better performance with SMC from home habitat, while negative values indicate better performance with SMC from away habitats. Note: in the field experiment we used only pioneer species CecA and shade tol erant species Gar and Ret (species abbreviations in Table 1). ** P < 0.01; *** P < 0.001

PAGE 54

54 Figure 2 2. Effect of soil sterilization in the greenhouse experiment (A) and fungicide treatment in t he field experiment (B) on th e growth of seedlings grown with SMC from typical species rar 2 1) Bars indicate the value ( SE) of a priori contrasts comparing the home vs. away contrast with live inocula (greenhouse exp) or control seedlings (field exp) vs. that for sterilized inocula or fungicide treated seedlings. Positive values indicate that away advantage was greater for sterile (or fungicide treated) than for live (or control) soil, or that home advantage was greater for live (or control) than for sterilized (or fungicide treated) soil. Negative values indicate that away advantage was greater for live (or control) than for sterilized (or fungicide treated) soil, or that home advantage was greater for sterilized (or fungicide treated) than for live (or control) soil. Note: in the field experiment we used only pioneer species CecA and shade tolerant species Gar and Ret ** P < 0.01; *** P < 0.001

PAGE 55

55 Figu re 2 3. R elative growth rate (RGR) ( least square means SE) in the greenhouse of four species group s (Table 2 1) when inoculated with soil microbial communities (SMC) from pastures (P), coffee plantatio ns (C), and forest fragments (F) Plants were inocula ted with an either live (open circles) or sterilized (closed circles) mixture of roots and rhizosphere soil collecte from 5 sites per habitat type. ** P < 0.01; *** P < 0.001

PAGE 56

56 Figure 2 4. Growth (log total dry weight leas square means SE) in the field of four species groups (Table 2 1) in pastures (P), coffee plantations (C), and forest fragments (F) Seedlings were grown in 12 sites (4 si tes per habitat type) for 3 months during which plants were either treated with water (control) (o pen circles) or with the fungicide b enomyl (closed circles) at the beginning of the experiment and then every month. *** P < 0.001

PAGE 57

57 CHAPTER 3 IS SOIL AT HOME MORE BITTER THAN SOIL FROM AWAY? HOST SPECIFIC EFFECTS OF SOIL MICROBIAL COMMUNITIES FROM AGRICULTU RAL AND NATURAL HABITATS Summary Symbiotic interactions are vital for the maintenance of biodiversity. Thus, it is crucial that we have a better understanding of how these interactions vary across heterogeneous human modi fied landscapes. We examined how p lant interactions with beneficial and detrimental components of soil microbial communities vary across three types of habitats highly fertilized, low diversity agricultural monocultures (pastures and coffee plantations), and unmanaged, highly diverse pre montane tropical forest fragments. To do this, we se t up a greenhouse experiment in which arbuscular mycorrhizal fungal (AMF) ( likely mutualists) and non AMF soil microbial communities (containing antagonistic soil symbionts) were isolated and amplified on common host plant species in the three habitats, and examin ed their effects on growth of 11 host plant species with reciprocal inoculation. We tested two alternative, non mutually exclusive hypotheses. The first was that plants would grow better with both the AMF and non AMF inocula from habitat ( where a plant species rarely or never occurs) relative to habitat ( where a plant spec ies typically occu rs) The second was that soils from agricultural monocultures had less beneficial AMF (due to chemical fertilization) and /or more detrimental non AMF symbionts (due to an accumulation of species specifi c antagonistic soil microbes). The resul ts showed that plants differentially interact with diverse arrays of AMF and non A MF soil microbes from c ontrasting habitats Overall, different plant species benefited more from away compared to home non AMF inocula, suggesting that host specific antagonistic soil symbionts accumulate

PAGE 58

58 where their hosts were abundant In contrast, most plants responded simila rly to AMF from different habitats, with the exception of forest pioneer trees, which benefited more from AMF in coffee plantations compared to AMF in forest fragments. Combined, these results indicate that host prevalence has a great effect on shaping soi l microbial communities which have significant effects on plant communities Thus, the replacement of native forests by agricultural monocultures may impact aboveground plant biodiversity through plant soil dynamics Background Despite the key role of mut ualistic and antagonistic sy mbiotic associations in driving ecological and evolutionary processes, we still have a poor understanding of how symbiotic associations vary across environmental gradients and community complexity (Thrall et al. 2007) Notably, understanding how symbiotic interactions are changing in altered ecosystems might be crucial for conservation of ecosystem services (Flynn et al. 2009) and biological diversity (Kiers et al. 2010) as organisms adapt to one another and to abiotic environmental heterogeneity ( Johnson et al. 2010) For example, much research is needed to fully understand how differ ent factors interact to cause the loss of endosymbiotic microalgae in corals (coral bleaching), and how corals cope with stressful environmental conditions in different locations (Hughes et al. 2003) Similarly little is known about how the interactions of plants and soil symbionts are altered in human modified ecosystems, even though their importance is recognized for plant fintess (Van der Heijden et al. 2006) diversity and distributions (e.g., Mangan et al. 2010 b Pizano et al. 2011) Soils contain extremely diverse microbial communities that interact wi th plant communities as mutualistic pathogenic, or neutral agents For example, some

PAGE 59

59 arbuscular mycorrhizal fungi (AMF) benefit plants with acquisition of mineral nutrients ( especially phosphorus) and water, and protect against pathogens in e xchange for sug ars from host plants (reviewed in Koide and Mosse 2004) However, o ther AMF, as well as many non AMF soil microbes may act as plant pat hogens or parasites. Both beneficial and antagonistic soil organisms interactively influence plant populations (Rodriguez et al. 2009 Sanon et al. 2009) shaping competitive interaction s, distribution and abundance of plant species in natural (e.g., Mitchell 2003) and agricultural settings (e.g. Zhu et al. 2000) Thus, both types of microbes must be considered simultaneously to test for the positive or negative effects of the soil microbial community on host plant species. In most studies, however, the effects of whole soil microbial communities on plant performan ce are often attributed to either AMF or antagonistic soil or ganisms without proper identification of these different components of soil micro bial communities. M ost studies pertaining to plant soil interactions along abiotic and biotic gradients have focus ed on either mutualist ic or antagonistic soil microbes, and in just one type of habitat (e.g. natural or managed setting). The goal of this study was to improve the understanding of how abiotic and biotic factors shape plant soil interactions by experiment ally separating the effects of AMF and non AMF soil microorganisms from three habitats contrasting in soil fertility and plant community diversity. Soil microbial communities are shaped by both soil nutrient conditions and plant community composition. I nc reasing soil fertility is predicted to promote non mutualistic exploitation by groups of microbes that are usually considered to be mutualists, favoring the increased abundance of symbionts that act as functional parasites (Thrall et al. 2007) T he nutritional benefits of AMF associations are only evident in low fertility soils

PAGE 60

60 (e.g. Reynolds et al. 2006) and chemical fertilization has been shown to reduce resource allocation to plant roots, decreasing myco rrhizal infection and possibly select ing for parasitic AMF strain s that exploit host derived resourc es while providing limited benefit to the host (Johnson et al. 1997 Graham and Abbott 2000 Treseder 2004 Kiers and Van der Heijden 2006) Likewise, chemical fertilizers m ay also act directly on both plants and pathogens, influencing the outcome of their interactions (reviewed in Ghorbani et al. 2008) On the other hand, there is a clear interdependence between plant community composition and soil micro organism community composition, as plant species influence which organisms occur in the soil and soil microbes sh ape plant community composition ( Van der Heijden et al. 1998 Mangan et al. 2010b) For example, there is evidence for lower diversity of AMF in agricultural monocultures than in natural and more plant diverse ecosystems (e.g., Verbruggen et al. 2010) .Similarly, the diversity disease hypothesis (Elton 1958) states that the spread of host specific pathogens depends on the abundance of their host species, therefore low plant diversity (e.g. monoculture) results in an increased incidence of specialist antagonistic soil pathogens (e.g., Mitchell 2003 Schnitzer et al. 2011) In summary, ab undance of host specific vs. generalist soil micro organisms, whether beneficial o r antagonistic, is interactively determined by plant species composition and diversity in natural communities. In this study we explored interactions between individual plant species and soil microbial communities from pre montane tropical forest fragments, coffee plantations and pastures in the tropical Andes. These habitats contrast in application of chemical fertilizers (none in forest fragments, high in coffee plantations and pastures) and in plant

PAGE 61

61 where a plant species commonly (habitat where a plant species r arely or never occurs) has positive, neutral, or negative effects depends on the balance between beneficial and antagonistic soil microbes. We experimentally differentiate d the effects of AMF (which are largely believed to b e mutualists of plants) and non AMF soil microbes on growth of 11 plant species, by using AMF inocula (prepared from AMF spores) and non AMF inocula (microbial filtrate ex cluding large AMF spores). The specific objectives were to test for 1) overall net effects of AMF and the microbial filtrate from contrasting habitats on growth of plant spec ies common in these habitats, and 2) response of plants to AMF and the microbial filtrate from their home and away habitats. We tested two alternative, non mutually exclusive hypotheses. The first was that the s to host plants ( Bever 2002 b ) as well as host specific pathogens. The second was that AMF communities in the fertilized soils of pastures an d coffee plantations were like ly to be dominated by less mutualistic AMF Methods Study site and species This study was conducted in the agricultural region of the Central Cordillera of The climate i s tropical humid, with an average annual temperature of 21C, and an annual rainfall of 2550 mm concentrated in two wet seasons (March to July and September to December) (Guzmn Martnez et al. 2006). Soils are Udands (Andisols) (Ortz Escobar et al. 2004) The landscape is dominated by three contrasting habitat types: occasionally fertilized pastures with low species diversity (mainly African grasses Pennisetum clandestinum Melinis minutiflora and Brachiaria spp. ), heavily fertilized monocultures of sun g rown

PAGE 62

6 2 cof fee plantations, and unmanaged fragments of pre montane tropical forest (Orrego et al. 2004a) The forest fragments are usually small (1 30 ha), biologically more diverse than the other two habitats, and mostly dominated by early and mid successi onal pla nt species, with high liana densiti es. Seedling recruitment of native species appear ed poor in the under story, where seedlings of non native species, including Coffea arabica (coffee) are common (C. Pizano pers. obs.). Forest tree species were selected fr om those abundant in forest fragments in the region (Orrego et al. 2004b) encompassing a wide range of seed size and life histories (Table 3 1 Appendix A). Brachiaria brizantha a nd Coffea arabica were chosen because they dominate pastures and coffee plantations, respectively in the study region (Table 3 1). To sample across heterogeneous altitude and climatic conditions in the study area, five private farms with similar conditions each containing a pasture, a coffee plantation, and a forest fragment (Table 3 2) were chosen to collect soil for the greenhouse ex periment. At each of the 15 site habitat combinations five samples of 200 g of mineral soil (5 15 cm in depth) each were ta ken and pooled in a composite sample from which a subsample was analyzed. Soils across the three habitats had similar levels of pH, organic matter content, and some nutrients (N, K, Mg), but forest fragment soils contained only 10 12 of P, and significantl y greater Ca content than pasture and coffee plantation soils (Table 3 3). In addition, light levels in pastures and coffee plantations were approximately 30 times high er than in forest fra gments (Table 3 3). Seed preparation and so il medium for growing se edlings Seeds of Brachiaria grass and coffee were provided by Cenicaf, while seeds of the nine forest plant species (Table 3 1 ) were collected from at least three parent trees in forest fragments in the study region. Seeds were surface sterilized with 0.6 % sodium

PAGE 63

63 hypochlorite for 15 minutes and germinated in trays containing a steam sterilized mixture of soil and sand river ( 3:2 ) soil The soil was collected from an open area near the greenhouses at Cenicaf (Colombian National Research Center for Coffee, in Chinchin, Caldas, Colombia), and it was classified as an acrudoxic melanudand from the Chinchin unit (Ortiz Escobar et al. 2004), with moderately fertile chemical characteristics (5.4 pH, 0.1 % N calculated, 1.9 % OM Walkley Black colorimetry, 7 mg k g 1 P Bray II Bray Kurtz colorimetry, 0.37 mg kg 1 K Ammonium acetate 1N, 3.4 mg kg 1 Ca Ammonium acetate 1N, and 1.1 cmol kg 1 Mg Ammonium acetate 1N). Initial inoculum collection Whole soil inocula (i.e. roots, r hizo sphere soil, and associated organisms) were collected from pastures, coffee plantations, and forest fragments at five farm sites each having these three habitats (total of 15 site habitat combinations) (Table 3 2). Roots and rhizosphere soil (5 15 cm in depth) were collected from five random l ocations within each habitat at each site and pooled together. We then pooled, and homogenized the inoculum from the five sites for each habitat type. AMF inoculum amplification Inoculum amplification was done in two steps. In the first step individually potted plants of each of the 11 species (Table 3 1 ) were initially inoculated with soils coming from their respective habitat types (i.e. B. brizantha with the soil from pastures, C. arabica with the soil from coffee p lantations, and forest species with th e soil f rom forest fragments) (Fig. 3 1. 1). The same sterilized soil and sand mixture used to germinate seeds filled 20 % of 1 L pots, to which an equal quantity (100 mL) of the whole soil inoculum from a particular habitat was added. Ten seedlings per spe cies were grown for 5 months (Fig. 3 1. 2) in the greenhouse after which they we re harvested and their roots

PAGE 64

64 and soil were mixed to provide 8 subsamples of 20 g of soil from each plant inoculum combination (Fig. 3 1. 3). Each soil subsample was washed with w ater through a series of sieves, and the fraction that was retained on the 45 m diameter sieve was collected (mostly AMF spores) (Klironomos 2002) (Fig. 3 1. 4a). In order to clean the AMF spores and separate them from debris and other possible contaminants, the spore fraction was concentrated by sucrose densi ty gradient centrifugation (Daniels and Skipper 1982) and then individual spores were separated using micro tweezers. Once all clean spores were obtained, they were used to inoculate previously germinated seedlings of Brachiaria decumbens and Cecropia angu stifolia (6 seedlings per species) in sterilized soil. To inoculate seedlings the spores were pipetted directly onto the plant root system (Fig. 3 1. 5a). The same process was repeated for the inoculant from the 11 host species, therefore we had a total of 66 seedlings of Brachiaria decumbens and Cecropia angustifolia each inoculated with 11 differen t AMF communities. These hosts were selected to amplify the AMF inoculum because they grow fast and produce high amounts of AMF spores in a relatively short tim e (Mangan et al. 2010a) In addition, this allowed for standardization of the amount of spores, given that AMF sporulate at different rates on differ ent hosts. Inoculated seedlings grew for five months in the greenhouse (Fig. 3 1. 6) after which they were harvested and their soil and roots were used as inoculum in the greenhouse experiment (Fig 3 1. 7). For each inoculum source the seedlings of B. decu mbens and C. angustifolia were harvested, and their roots and soil were pooled inoculum, the roots and soil from the 9 forest species was pooled together Thus, whi le

PAGE 65

65 the AMF inocula from coffee plantation and pa sture were initially amplified o n a single species ( Coffea arabica and Brachiaria brizantha respectively), the forest inoculum was amplified on 9 forest plant species. Therefore, AMF inocula isolated from cof fee pla ntations and pastures were likely dominated by AMF that colonize the roots of the respective crop species, while the forest inoculum probably included a greater diversity of AMF species from the different host plants Filtrate inocula These inocula were produced by amplifying soil microbes collected in a filtrate from the 25 m diameter sieve that excludes large AMF spores (Klironomos 2002) Other studies (e.g., Klironomos 2002 McCarthy Neumann and Kobe 2008) have used the same method to isolate a representative community o f non AMF soil organisms including microbial pathogens like bacteria and non AMF fungi. T he filtrate was used to inoculate seedlings of the respective species used in the initial amplification (Fig. 3 1. 1). For example, we inoculated Juglans neotropica seedlings with filtrate obtained from pots with Juglans neotropica seedlings (Fig. 3 1. 4b, 3 1. 5b). These seedlings were also grown for 5 months in the greenhouse after which they were harvested and their roots greenhouse experiment (Fig. 3 1. 6, Fig. 3 1. 7). Greenhouse experiment growth conditions and t reatments Pots of 1 L volume were 70 % filled with a steam sterilized (for 2 hrs) mixture of soil and river sand (3:2) The soil was collected from an open area at CIAT (International Center for Tropical Agriculture) in P almira, Valle, Colombia, and had intermediate fertility (6.8 pH, 0.05 % N, 0.6 % OM Walkley Black colorimetry, 42 mg kg 1 P Bray II Bray Kurtz colorimetry, 0.35 mg kg 1 K Ammonium acetate 1N, 3.7 mg kg 1 Ca

PAGE 66

66 Ammonium acetate 1N, and 2.8 cmol kg 1 Mg Ammoniu m acetate 1N). Then, we added an equal amount (100 mL) of AMF or filtrate inoculum. Each treatment combination of inoculum source (forest fragments, coffee plantations, pastures) inoculum type (AMF, filtrate) plant species (11 plant species) was replic ated in 8 pots and plants were equally distributed on 4 bench es (blocks) in the shade house Plants were grown under 20 % light for 130 days. We measured the leaf area of each plant at the beginning o f the experiment to use as a covariate with final total dry mass at harvest in the statistical analyses Also at harvest, AMF root colonization was quantified for a subsample of 3 plants from each treatment. For this assessment, a 1 g aliquot of fine roots was washed with running water, cleared with 3 % KOH and 3 % H 2 O 2 acidified with 3 % HCl, washed with running water, and stained with 0.05 % tryplan blue in lactoglycerol solution (Zangaro et al. 2000) Root segments were then mounted on slides (2 slides/seedling) and presence or absence of AMF was recorded at 200 intersect points for each slide (for a total of 400 intersect points for each individual). Statistical analyses T wo sets of analyses were performed to examine 1) the effects o f t he six types of soil inocula (three habitats two microbes types) relative to the control (sterile soil; no l on t he 11 plant species which were classified to four plant species groups : Brachiaria grass, coffee (a non native shade tolerant species), forest pioneer trees and forest shade tolerant trees F irst we used analysis of covariance (ANCOVA) to examine the re sponse of the final seedling biomass to two fixed main factors including inoculum type (7 types, including s terile,

PAGE 67

67 AMF and filtrate from each of the three habitat types pastures, coffee plantations, and forest fragments) and plant species (and their inte r actions), farm site ( block ) as a random effect and initial leaf area, and days in experiment as covariates. Inclusion of the latter covariate accounted seedling age heterogeneity introduced by the replacement of dead seedlings during the first month. We then used a priori contrasts to compare the growth of each species group when inoculated with either AMF or filtrate from different habitat types relative to the control (Fig. 3 2). For the second ANCOVA the sterile treatment was excluded to examine the response of final plant mass to three fixed effect factors including inoculum source (pastures, coffee plan t ations, forest fragments), inoculum type (AMF, filtrate), and plant species and two way and three way interactions among them, as well as block as a random factor and initial leaf area and days in experiment as covariates. Then, a priori contrasts were used to compare seedling growth of each species group when grown with AMF and filtrate inoculum from their home habitat relative to that from away hab itats separately (Fig. 3 3). Statistical analyses were done with JMP version 8.0 (SAS Institute Inc., SAS Campus Drive, Cary, NC USA 27513 ). We used the Dunn Sidak correction to adjust the significance levels of contrasts. To compare AMF colonization ( % ) across the different soil inocula, the statistical tests were conducted for a subset of randomly selected plants for which roots were processed (3 replicates per treatment). This model included inoculum source (forest fragments, coffee plantations, and p astures), inoculum type (AMF, filtrate), and plant species (and their interactions) as fixed effects, and block, initial leaf area, and final

PAGE 68

68 biomass as covariates. Plants in the sterile treatment were not included in this analysis as none of the examined plant roots showed evidence of AMF colonization. Results In the first analysis we compared seedling growth for the s ix types of soil inocula (3 habitats two microbes types) relative to the control (sterile soil) and found that biomass was signifi cantly affected by soil inocula and plant species (Table 3 4). In the second analysis we contrasted seedling growth across inoculum type s from the three habitats (excluding the control treatment), and found that inoculum source (3 habitats), inoculum type (AMF an d filtrate) and plant species had significant effects on seedling final biomass (Table 3 5). In addition, significant two and three way interactions indicated that plant species differed in their response to different inocula (Table 3 4), and to differen t combinations of inoculum source habitat and inoculum types (Table 3 5). Inoculation had significant negative effects on Brachi aria grass, which grew slightly better with sterile soil than with AMF inocula from pastures ( F 1,498 = 5.3; P = 0.021) and coff ee plantations ( F 1,498 = 4.1; P = 0.043; not significant with the Sunn Sidak correction for multiple contrasts ) and better with sterile soil than with the microbial filtrate (from pastures: F 1,498 = 61.4; P < 0.001; Table 3 6, Fig. 3 2A). A priori contrast tests (Fig. 3 3) showed that AMF from different habitats had similar negative effect on the growth of this grass (P > 0.1 for all contrasts), while the microbial filtrate from its home habitat was significantly more detrimental than that of coffee plantat ions (F 1,428 = 55.7; P < 0.001) and forests (F 1,428 = 59.5; P < 0.001; Table 3 7, Fig. 3 3A). Growth of coffee was improved by the inocul ation of AMF from all three habitats (pastures: F 1,428 =

PAGE 69

69 16.4; P < 0.001, coffee plantations: F 1,428 = 10.7; P = 0.0012 and forests: F 1,428 = 6.2; P = 0.013), and also by t he filtrate from pastures (F 1,428 = 7.1; P = 0.008) and forests (F 1,428 = 22.1; P < 0.001) compared to the sterile treatment (Table 3 6, Fig. 3 2B). AMF from different habitats had the same positive ef fect on the growth of this sp ecies (P > ), but the filtrate from its home habitat was the least beneficial compared to that of pastures ( F 1,42 8 = 4.0 ; P = 0.046; not significant with the Sunn Sidak correctio n for multiple contrasts ) and forests (F 1,428 = 17.1; P < 0.001; Table 3 7, Fig. 3 3B). Pioneer tree species (CecA, CecT, Sol, Ochr, and Sip) grew significantly better when inoculated with AMF from the three habitat types (pastures: F 1,498 = 33.9; P < 0. 001; coffee plantations: F 1,498 = 75.9; P < 0.001, forests: F 1,498 = 37.7; P < 0.001), and with filtrate from pastures (F 1,498 = 32.3; P < 0.001) compared to the sterile treatment (Table 3 6, Fig. 3 2C). In contrast, their growth did not significantly diff er between sterile soil and filtrates from coffee plantations (F 1,498 = 2.5; P = 0.11) and forests (F 1,498 = 0.04; P = 0.85; Table 3 7, Fig. 3 2C). When comparing the growth of these pionee r species away habitats, AMF from pastur es had a similarly positive effect (F 1,427 = 0.4; P = 0.51), while AMF from coffee plantations were more beneficial (F 1,427 = 6.2; P = 0.013) (Table 3 7, Fig. 3 3C). Pioneer species grew better with the filtrate from pastures (F 1,427 = 23.6; P < 0.001), but similarly with the filtrate from coffee plantations (F 1,427 = 3.2; P = 0.077) compared to that of their home habitat (Table 3 7, Fig. 3 3C). Growth of shade tolerant species (Gar, Gus, Jug, and Ret) was only margi nally increased by AMF from coffee plantations (F 1,498 = 4.1; P = 0.044; not significant with

PAGE 70

70 the Sunn Sidak correction for multiple contrasts ) compared to the sterile treatment, while being similar across all other AMF and filtrate treatmen ts (P > 0.1 for all contrasts; Table 3 6, Fig. 3 2D). Similarly, these species produced comparable growth with AMF and the microbial filtrate from their home habitat compared to AMF and filtrate from pastures and coffee plantatio ns (P > 0.2 for all contrasts; Table 3 7, Fig. 3 3D). AMF colonization was significantly higher in plants that received AMF inocula (61.6 %) than those i n oculated with microbial filtrate (29.1 %). Thus, although AMF inoculum potential was greatly reduced in the filtrate AMF hyphae in the microbi al filtrate probably colonized the roots (Table 3 8). AMF root colonization in filtrate treated plants was highest in seedlings treated with the microbial filtrate from pastures (39.6 %), intermediate in seedlings treated with the microbial filtrate from f orests (33.7 %), and lowest in plants inoculated with the microbial filtrate from coffee plantations (13.9 %) (Fig. 3 4). Discussion In this study communities of AMF and other soil microorganisms (microbial filtrate) from habitats with contrasting soil fe rtility and plant diversity had differential effect s on growth of plant species. Hence, 1) soil microbial communities differ ed across the habitats and 2) different host species associate d with particular subsets of beneficial and antagonistic microorganis ms When testing for the overall net effects of AMF and the microbial f iltrate from different habitats, AMF communities from heavily fertilized pastures and coffee plantations had a net negative effect on growth of Brachiaria grass (Fig. 3 2A), but improve d the growth of coffee (Fig. 3 2B) and pioneer forest species (Fig. 3 2C) in comparison the sterilized soil control (Table 3 6). The microbial filtrate from pastures also decreased the growth of Brachiaria grass (Fig. 3 2A), but had a net

PAGE 71

71 positive effect o n coffee (Fig. 3 2B) and pioneer forest species (Fig. 3 2C) relative to the sterile control (Table 3 6). These positive effects can be attributed to the incomplete elimination of AMF from the filtrate, although plants inoculated with the filtrate had signi ficantly lower AMF colonization rates (Table 3 8, Fig. 3 4). Finally, all plants grew similarly with sterile soil and with the microbial filtrate from coffee pl antations and forest fragments, except coffee, which benefited from the filtrate from forests ( Ta ble 3 6, Fig. 3 2B). species grew better with soils from away habitats, but this response was driven by the filtrate inocula, and not by AMF. In fact, the only species group that sho wed a significant response to AMF from different habitats were forest pioneer trees, which grew significantly better with AMF from coffee plantations compared to AMF from their home habitat (Fig. 3 3C). In contrast, Brachiaria grass (Fig. 3 3A), coffee (Fi g. 3 3B), and pioneer forest trees (Fig. 3 3C) grew significantly better with the filtrate from away compared to that from their home habitats (Table 3 7). These results suggest that fo rest fragments and pastures AMF communities have a similar impact on pl ants, although fertilized coffee plantations seem to be dominated by AMF th at are more beneficial to forest pioneer species than those present in forests. Furthermore, agricultural lands and forests seem to have contrasting communities of non AMF soil micr oorganisms that differ in beneficial and detrimental effects on different host species. AMF from contrasting habitats have similar effects on most plant species Numerous studies have reported lower AMF diversity in managed agricultural lands than in unma naged natural habitats (e.g. Picone 2000 Oehl et al. 2003) but few have actually tested for the effects of these AMF on plant performance. Here we

PAGE 72

72 attempted to separate the effects by inoculating plants with AMF and microbial filtrates from highly fertilized agricultural lands and natural forests. Chemical fertilization and other agricultural practices (e.g. tillage) have been predicted to select for AMF that act as parasites without offering any benefit to their host (Graham and Eissenstat 1998 Kiers and Van d er Heijden 2006 Verbruggen and Kiers 2010) In support of this idea, AMF communities from managed, f ertilized habitats may be less diverse and less beneficial than those from unfertilized, natural habitats (e.g. Neuhauser and Fargione 2004 Kiers and Denison 2008 Rasmann et al. 2009 Verbruggen and Kiers 2010) Graham and Abbott (2000) showed that AMF from fertilized agricultural lands increased plant P in exchange for sucrose status in the roots of wheat, but did no t increase plant biomass, suggesting that these AMF were acting as p a rasites. In natural grasslands, J ohnson et al. (2010) showed that two ecopytes of Andropogon gerardii adapted to high and low nutrient soils obtained the most benefit from AMF present in nutrient poor soils, and the less benefit from AMF common in nutrie nt rich soils, demonstrating again that the latter were less mutualistic The effects of AMF from contrasting habitats on plant performance might not only vary according to different s oil nutrient status, but also based on host identity. For instance, AMF from fertilized pastures and coffee plantations had a marginally net negative effect i.e., a pathogenic effect, on the growth of the Brachiaria grass but AMF from pastures were beneficial to coffee and forest pioneer species (Fig. 3 2 ). In addition, AMF from coffee plantations had a net positive effect on coffee forest pioneer species, and forest shade tolerant species (Fig. 3 2 ) compared to the sterile control. Furthermore, AMF from coffee plantations were s ignificantly more beneficial to forest

PAGE 73

73 pionee r species than AMF from their home habitat (Fig. 3 3C), while all other species grew similarly well with AMF from the three habitat types (Fig. 3 3). These results contradict other studies suggesting that AMF from highly fertilized agricultural lands are l ess beneficial than those present natural habitats (Neuhauser and Fargione 2004 Kiers and Denison 2008 Verbruggen and Kiers 2010) and only partially support our prediction that plants would grow better with home than with away AMF. Moreover, the fact that most species grew similarly well with AMF from home and away habitats suggests that plants either find similar AMF genotypes across habitat types, or that different AMF communities from contrasting habitats have the sam e effect across plant species. One potential factor that influenced the results was that the method used to isolate AMF (picking individual AMF spores) biased our representation of real AMF communi ties from the field, as we probably selected from all habitat types AMF that produced copi ous spores ( i.e. strategist Verbruggen and Kiers 2010) perhaps representing only one of many AMF functional types (Munkvold et al. 2004) N on AMF soil microorganisms (microbial filtrate) from contrasting habitats had different effects across plant species The diversity disease hypothesis (Elton 1958) states that there is an inverse correlation between the incidence of specialis t pathogens and plant diversity, but few studies have actually tested this hypothesis. Schnitzer et al. (2010) showed that there is indeed a higher incidence of soil pathogens in grasslands of low plant diversity compared to grasslands with higher plant di versity, and that host specific disease and root infection by pathogens decrease with an increase in plant diversit y. In this study we tested the diversity disease hypothesis by inoculating different plant hosts with microbial filtrates (containing antagon istic soil organisms) from habitats with contrast ing plant

PAGE 74

74 diversity, to evaluate whether 1) microbial filtrates from species poor habitats were more more detrimenta l than filtrate from away habitats. We found no pattern showing that microbial filtrates from species poor pastures and coffee plantations were more detrimental than the filtrate from species rich forest fragments. Filtrate from pastures had a negative ef fect on Brachiaria grass (Fig. 3 2A), but benefited growth of coffee (Fig. 3 2 B) and pioneer forest species (Fig. 3 2C), while growth did not differ significantly when inoculated with the filtrate from coffee plantations compared to growth on sterile soil (Fig. 3 2). In addition, filtrate from forest fragments had a positive effect on coffee (Fig. 3 2B), but no effect on other plants relative to the sterile control (Fig. 3 2). These results suggest that the incidence of antagonistic soil microbes did not di ffer consistently between agricultural monocultures and natural forests, but it is also possible that the results reflected the net balance between negative and positive microbes. We did no t completely exclude AMF from the filtrate inocula (Table 3 8, Fig. 3 4), and thus, the positive effect of filtrate from pastures and f orests on some plant species could be explained by a relatively high AMF root colonization in seedlings under these treatments (Fig. 3 4) probably due to fragments of hyphae passing throug h the 25 m diameter sieve In addition, other beneficial soil micro organisms such as bacteria would have been present in our microbial filtrate inocula (Sanon et al. 2009) Although no common pattern of plant species specific response to microbial filtrate inocula from different habitats was found, positively influenced the growth of most plant species Brachiaria grass (Fig. 3 3A),

PAGE 75

75 coffee (Fig. 3 3B), and forest pioneer species (Fig. 3 3C) all grew better with the filtrate from away than with filtrate from their home habitats. In the case of the grass, away advantage was due to les s detrimental effect of away filtrates compared to a very negative effect of the filtrate from pastures (Fig. 3 2A), suggesting that pastures accumulate antagonistic soil organisms that are detrimental for B. brizantha. In the case of coffee (Fig. 3 3B) an d pioneer species (Fig. 3 3C), away advantage was due to an increased benefit of the away microbial filtrate (Fig. 3 2B, 3 2C), which further suggests that AMF in the microbial filtrate inocula may have had a significant effect on seedling growth. Moreover it indicates that the away vs. home advantage for most of our species was driven by the presence of mutualistic, and not by the absence of antagonistic soil organisms in the away, compared to home habitats. For example, it is possible that forest fragmen ts host soil organisms beneficial to coffee that are absent in coffee plantations. Alternatively, AMF and other mutualists may be less effective in home, compared to away soils (Bever 2002 b ) Together, these results only partially support the diversity disease hypothesis; incidence of antagonistic soil microbes was indeed higher in pastures compared to more diverse forest fragments, however the se microbes only negatively affected Bra chiaria grass, the dominant species in pastures. Following this same line of evidence, coffee and forest species appeared to encounter AMF and other mutualistic soil organisms in coffee plantations and forest fragmen ts (correspondingly) that were less beneficial than those present in away habitats. Thus, habitats with contrasting plant diversity seem to similarly accumulate antagonistic non AMF microbes as well as less effective mutualists that negatively impact the most abundant plant species in each habitat type.

PAGE 76

76 Applied ecological implications Environmental degradati on represents a great threat to symbiotic interactions that are key for the maintenance of biodiversity and ecosystem services (e.g. Bascompte 2009 Kiers et al. 2010) For instance, it is now clear that there is a world wide pollination crisis that is affecting both natural and agricultural systems (Steffan Dewenter et al. 2005 Biesmeijer et al. 2006) Similarly, numerous studi es have addressed the consequences of the loss of seed dispersers on plant populations due to forest fragmentation (Cordeiro and Howe 2003 Tscharntke et al. 2008) However, compared to aboveground symbiotic in teractions, our understanding of bel owground symbionts remains poor and limited to the few soil organisms that we can isol a te (e.g. earth worms, AMF). M ost studies have only focused on comparing the diversity of a relatively small number of soil organisms that are easy to isolate across natural and degraded ecosystems (but see Postm a Blaauw et al. 2010) Here we tested for the effects of communities of belowground symbionts from natural and degraded ecosystems on plant performance, with potentially important implications for agricultural production and natural forest regeneration. O ur results show that forest fragments and agricultural lands in the tropical Andean region share functionally similar communities of AMF, with coffee plantations hosting particularly beneficial mycorrhizae for forest pioneer tree species. In contrast, past ures, coffee plantations and forest fragments seem to have contrasting communit ies of non AMF soil organisms. In particular, p astures accumulate antagonistic soil microbes, but also AMF that are detrimental to the widely planted grass B. brizantha but bene ficial to coffee and pioneer forest species. Forest fragments, on the other hand, apparently have non AMF, bene ficial microbes that improved the growth of coffee. Combined, these results not only provide further

PAGE 77

77 evidence for how the establishment of agricu ltural monocultures has modified belowground micro bial communities, but also reveal that differential mutualistic and antagonistic soil symbionts from modified and natural ecosystems can significantly shape plant communities in these environments.

PAGE 78

78 Table 3 1. General characteristics of the eleven plant species used in the greenhouse experiment. Seed dry mass (mean S.D) was measured from 25 50 seeds/species dried for 3 days at 60 C. Species Family Species group Typical habitat Seed dry mass (g ) Abbrevia tion Brachiaria brizantha Poaceae Brachiaria grass (crop) P 1 0.0077 0.0013 Bra Coffea arabica Rubiaceae Coffee (crop) C 2 0.22 0.023 Cof Cecropia angustifolia Cecropiaceae Pioneer F 3 0.0011 0.00028 CecA Cecropia telealba Cecropiaceae Pi oneer F 3 0.00039 0.00012 CecT Ochroma pyramidale Bombacaceae Pioneer F 3 0.0037 0.0012 Ochr Siparuna aspera Monimiaceae Pioneer F 3 0.014 0.025 Sip Solanum aphynodendrum Solanaceae Pioneer F 3 0.0015 0.00053 Sol Garcinia madrunno Clusiaceae Shad e tolerant F 3 2.92 0.77 Gar Gustavia superba Lecythidaceae Shade tolerant F 3 10.55 2.78 Gus Juglans neotropica Juglandaceae Shade tolerant F 3 26.04 14.49 Jug Retrophyllum rospigliosii Podocarpaceae Shade tolerant F 3 0.86 0.18 Ret 1 Pastures 2 Coffee plantations 3 Forest fragments

PAGE 79

79 Table 3 2. Characteristics of the farms where the initial soil inocula were were collected (Guzmn Martinez et al 2006). At each farm there were three habitats: a pasture (P), a coffee plantation (C), and a forest fra gment (F). Farm (block) Geographic coordenates Altitude (m.s.l) Mean annual precipitation (mm yr 1 ) Mean annual temperature ( C) Habitat Habitat patch size (ha) Cenicaf 0500 N 75 36 W 1380 2733 20.9 P 0.5 C 0.5 F 40 Playa rica 0500 N 753 1290 2750 20.7 P 10 C 60 F 30 Alto espaol* 0456 N 7542 W 1720 3140 18.8 P 1.0 C 2.0 F 0.3 Naranjal 0459 N 7539 W 1400 3137 21.4 P 22 C 38 F 27 La Argentina 0502 N 7541 W 1354 2935 19.9 P 0.5 C 100 F 1.5

PAGE 80

80 Table 3 3. Mean ( SE) soil pH, organic matter (OM) content, and nutrie nt content of pastures (P), coffee plantations (C), and forest fragments (F) where soil inocula were collected. Habitat pH 1 OM 3 (%) N 2 (%) P 4 (mg/kg)* K 5 (cmol/kg) Ca 5 (cmol/kg)* Mg 5 (cmol/kg) P 5.6 ( 0.1) 7.2 ( 1.4) 0.3 ( 0.04) 54.4 ( 25.3) 0.6 ( 0.2) 5.6 ( 0.7) 2.1 ( 0.4) C 5.0 ( 0.3) 9.6 ( 1.7) 0.4 ( 0.06) 42.0 ( 16.7) 0.3 ( 0.1) 4.4 ( 1.1) 1.7 ( 0.6) F 5.5 ( 0.2) 12.4 ( 2.1) 0.5 ( 0.06) 5.6 ( 0.4) 0.4 ( 0.05) 8.3 ( 2.3) 2.4 ( 0.7) Notes: Log transformed data was analyzed with one way ANCOVA including block as a random factor and habitat as a fixed factor. 1 pH: Potentiometer soil: water 1:1 2 N (total): Calculated 3 OM: Walkley Black colorimetry 4 P: Bray II colorimetry Bray Kurtz 5 K, Ca, Mg: Ammonium acetate 1N Significant differences between habitat types P < 0.05

PAGE 81

81 Table 3 4. Growth response of 11 plant species (Brachiaria grass, coffee, and nine forest tree species; Table 3 1) to seven inoculum type s (sterile, AMF and filtrate from each of the three habitat types, pastures, coffee plantations, and forest fragments) in the greenhouse. Total biomass (error df = 498 ) Source Df F P Inoculum 6 18.5 <0.001 Species 10 493.1 <0.001 Inocul um Species 60 5.3 <0.001 Block 3 2.6 0.054 Initial leaf area 1 388.2 <0.001 Days in experiment 1 168.6 <0.001 Notes: ANCOVA was used to analyze log 10 transformed final biomass, with block (4 benches in the greenhouse), days in experiment (dead seedli ngs were replaced during the first month of the experiment), and log 10 transformed i nitial leaf area as covariates. Table 3 5. Growth response of 11 plant species (Brachiari a grass, coffee, and nine forest tree species; Table 3 1) to inoculum source (past ures, coffee plantations, and forest fragments), inoculum type (AMF, Fil), and plant species (and interactions) in the greenhouse. Total biomass (error df = 428 ) Source Df F P Inoculum source 2 0.3 0.760 Inoculum type 1 75.9 <0.001 Species 10 431.5 <0 .001 Inoculum source Inoculum type 2 8.3 <0.001 Inoculum source Species 20 5.0 <0.001 Inoculum type Species 10 6.9 <0.001 Species Inoculum type Inoculum source 20 5.6 <0.001 Block 3 2.8 0.040 Initial leaf area 1 339.4 <0.001 Days in experi ment 1 168.8 <0.001 Notes: ANCOVA was used to analyze log 10 transformed final biomass, with block (4 benches in the greenhouse), days in experiment (dead seedlings were replaced during the first month of the experiment), and log 10 transformed initial leaf area as covariates. All contrasts within the three way interaction were done to examine the differences between biomass of plants grown with inoculum coming from their home habitat compared to that of away habitats.

PAGE 82

82 Table 3 6. F and P (in parenthesis) v alues of a priori contrasts examining growth of four plant species groups across the 6 inocula used in the experiment with respect to the control treatment (sterilized soil). Plant group Inoculum contrasted vs. control Brachiaria grass Coffee Forest pion eer Forest shade tolerant AMF from pastures 5.3 (0.021) 16.4 (< 0.001) 33.9 (< 0.001) 1.6 (0.20) AMF from coffee plantations 4.1 (0.043) 10.7 (0.0012) 75.9 (< 0.001) 4.1 (0.044) AMF from forest fragments 0.4 (0.520) 6.2 (0.013) 37.7 (< 0.001) 1.5 (0.22) Filtrate from pastures 61.4 (< 0.001) 7.1 (0.008) 32.3 (< 0.001) 0.2 (0.69) Filtrate from coffee plantations 0.3 (0.59) 0.5 (0.48) 2.5 (0.11) 1.1 (0.29) Filtrate from forest fragments 0.1 (0.78) 22.1 (< 0.001) 0.04 (0.85) 2.7 (0.1) Table 3 7. F and P (in parenthesis) values of a priori contrasts examining growth of four plant species groups with either AMF or a microbial filtrate from their home compared to that of away habitats (P = pastures, C = coffee plantations, and F = forest fragments). AMF F iltrate Plant group Home vs. C Home vs. F Home vs. P Home vs. C Home vs. F Home vs. P Brachiaria grass (Home: P) 0.08 (0.78) 2.8 (0.1) 55.7 (< 0.001) 59.5 (< 0.001) Coffee (Home: C) 0.6 (0.42) 0.6 (0.43) 17.1 (< 0.001) 4.0 (0.046) Forest pionee r trees (Home: F) 6.2 (0.013) 0.4 (0.51) 3.2 (0.077) 23.6 (< 0.001) Forest shade tolerant trees (Home: F) 0.7 (0.41) < 0.001 (0.99) 0.3 (0.59) 1.7 (0.2)

PAGE 83

83 Table 3 8. AMF proportion root colonization of 11 plant species (Bra chiaria grass, co ffee, and nine forest tree species; Table 3 1) as affected by inoculum source (pastures, coffee plantations, and forest fragments), inoculum type (A MF, F il trate ), and plant species (and interactions) in the greenhouse. Total biomass (error df = 195 ) Sour ce Df F P Inoculum source 2 21.4 <0.001 Inoculum type 1 202.7 <0.001 Species 10 13.4 <0.001 Inoculum source Inoculum type 2 8.5 <0.001 Inoculum source Species 20 3.1 <0.001 Inoculum type Species 10 2.9 0.003 Species Inoculum type Inoculum source 20 2.4 0.0015 Block 1 0.4 0.51 Initial leaf area 1 0.4 0.52 Final biomass 1 0.95 0.33 Notes: ANCOVA was used to analyze AMF proportion root colonization, with block (4 benches in the greenhouse), log 10 transformed initial leaf area, and log 10 t ransform ed final biomass as covariates.

PAGE 84

84 Figure 3 1. Flow diagram showing the procedure that we used to produce the inocula for setting up the experiment in the greenhouse.

PAGE 85

85 Figure 3 2. Effects of soil microbial presence (AMF or filtrate) relative to sterilized soil for growth of Brachiaria grass (A), coffee (B), forest pion eer trees (5 species) (C), and forest shade tolerant species (4 species) (D) (Table 3 1). Sol microbial inocula (AMF spore isol ates: open bars, soil filtrate: closed bars) were prepared from soils collected from pastures (P; habitat for Brachiaria grass), coffee plantations (C), and forest fragments (F, habitat for forest tree species). The values indicate the value ( SE) of a pr iori contrasts calculated from all replicates, and comparing the growth of each species group when inoculated with AMF or filtrate from each habitat compared to that with sterilized soil. Positive values indicate growth enhanced by the microbial inoculum, while negative values indicate growth suppressed by the microbial inoculum. P < 0.05; ** P < 0.01; *** P < 0.001

PAGE 86

86 Figure 3 (species typica at all) habitats for growth of Brachiaria grass (A), coffee (B), forest pioneer trees (5 species) (C), and forest shade tolerant species (4 species) (D) (Table 3 1). Sol micr obial inocula (AMF spore isolates: open bars, soil filtrate: closed bars) were prepared from soils collected from pastures (P; habitat for Brachiaria grass), coffee plantations (C), and forest fragments (F, habitat for forest tree species). Bars indicate t he value ( SE) of a priori contrasts calculated from all replicates, and comparing the growth of each species group when inoculated with AMF or filtrate from their home habitat compared to that with away habitats. Positive values indicate better performan while negative values indicate better performance with soil microbes from away habitats

PAGE 87

87 Figure 3 4. Propor tion AMF root colonization in seedlings of Brachiaria grass, coffee, forest pioneer trees (5 species), and forest shade tolerant species (4 species) (Table 3 1) when inoculated with AMF (open circles) or a microbial filtrate (closed circles) from pastures (A), coffee plantations (B), or forest fragments (C) in the greenhouse. Circles indicate the mean proportion ( SE) AMF root colonization quantified in a subset of 3 replicates from each treatment and a veraged for each species group.

PAGE 88

88 CHAPTER 4 NEGATIVE FE EDBACK DOMINATES FOREST FRAGMENTS AND AGRICULTURAL LANDS IN THE TROPICAL ANDES Summary The interactions between plants and soil organisms, plant soil feedback s have increasingly been shown to determine plant community composition In both temperate grass land and species rich tropical forests, native species tend to experience negative feedback (rhizosphere soil microbes are more detrimental to consp ecifics than to heterospecifics, indirectly promoting heterospecific neighbors to the host) In contrast, po sitive feedback (root soil microbes of plant species are more detrimental to heterospecifics than to conspecifics) has been reported from non native invasive species, and interpreted as a factor that promotes dominance by these non native species However, it remains unknown how these in teractions might change across heterogeneous habitats differing in the abundance of native non native, invasive, and non invasive plant species We used a factorial greenhouse experiment to test plant soil feedback for 2 co mmon non native crop species and 8 native forest plant spe cies in a montane region of Colombia to address: 1) How does feedback differ between arbuscular mycorrhizal fungi (AMF) compared to non AMF soil microbes? And 2) What type of feedback occurs in agri cultural monocultures of non native species vs. forest fragments ? W e inoculated each plant species with either AMF spore cultures or a microbial filtrate derived from soil microbial communities (SMC) associated with each of the 10 species. Doing so, growth of each host plant species could be examined in response to inoculation with conspecific AMF and filtrate, compared to inoculation with AMF and filtrate culture obtained from nine heterospecific host species. Overall, feedback mediated by AMF was much wea ker than feedback mediated by the microbial

PAGE 89

89 filtrate. Surprisingly, Brachiaria brizantha a non native invasive grass that dom inates pastures, and Coffea arabica a non native not shrub that is naturalized in forest fragments both experienced strong negat ive feedbacks for the microbial filtrate that contained soil pathogens In contrast, most forest tree species sho wed neutral to slightly negative feedback, suggesting that the negative feedback shown for natural forests may be disrupted with fragmentation These results suggest that antagonistic species specific soil microbes that media te negative plant soil feedback accumulate in proportion when a single host species dominates, as is the case in agricultural habitats. Forest fragments, on the other hand, m ay have lost the diversity of soil microbial communities that drive negative plant soil feedbac ks i n undisturbed tropical forests Background Biotic interactions (e.g. competition, predation, disease) mediate plant coexistence at the local level and have proved to be vital for the maintenance of diversity in natural communities (Schemske et al. 2009) Recent studies have demonstrated the for local plant biodiversity ( Bever 1994 Van der Putten and Peters 1997 Hartnett and Wilson 2002 Klironomos 2002 Bever 2003 Reynolds et al. 2003 Mangan et al. 2010b ) and plant productivity ( Van der Heijden et al. 1998 Van der Heijden et al. 2006 Van der Heijden et al. 2008 Vogelsang et al. 2006 ). For instance, negative plant soil feedback, resulting from soil microbes that are more detrimental to c onspecific than to heterospecific plant species, seems to dominate both temperate grasslands ( Casper et al. 2008 Petermann et al. 2008 Harrison and Bardgett 2010 ) and species rich tropical forests (Mangan et al. 2010b) Such negative feedbacks limit dominance of a single species and are expected to contribute to species coexistence and local plant diversity.

PAGE 90

90 Positive feedback, on the other hand, can reduce local plant diversity, as plants experiencing positive feedback should become dominant. Examples of such pos itive feedbacks have been reported for many non native species that have escaped their natural enemies in their original distribution ranges (enemy release hypothesis), and exhibit positive feedback in their new ranges ( Klironomos 2002 Reinhart et al. 2003 Callaway et al. 2004 Wolfe and Klironomos 200 5 Reinhart and Callaway 2006 Van Grunsven et al. 2007 Vogelsang and Bever 2009 ). Plant soil feedbacks are mediated by soil microbial communities (SMC) comprised of diverse arra ys of organisms that interact both directly and indirectly with plants varying in the type of interactions from mutualistic to antagonistic, and from non specific to highly specific interactions ( Klironomos 2003 Kuyper and Goede 2004). Because of difficulties of identifying and isolating the di fferent components of SMC, many of the studies addressing the interaction between plants and soil microbes have focused on a small number of microbes that can be isolated with several culture techniques For example, it is now well established that different plant species host, and respond differently to, diverse arrays of arbuscular mycorrhizal fungi (AMF) ( Bever 2002a Castelli and Casper 2003 Mangan et al. 2010a ) that can be easily isolated from soils as spores. In addition, plant AMF feedbacks can be either positive (Mangan et al. 2010a) or negative (Bever 2002b) In contrast, other soil organisms are more difficult to study and it is a challenge to untangle their role in driving plant soil feedbacks. For example, soil bacteria, fungal endophytes, fungal pathogens, and virus are far mor e difficult to isolate and culture for experimental purposes. These constraints limited the number of

PAGE 91

91 studies that have examined the role of different SMC constituents in driving plant so il feedbacks (Klironomos 2002) The role of plant soil feedbacks in determining plant species relative abundance has been mainly tested within particular natural habitats (e.g., a prairie grassland by Klironomos 2002 and a moist tropical forest by Mangan et al. 2010b ). But, we have a limited understanding on how these interactions v ary across the landscape consisting of habitats that differ in the abundance of native, and non native, and of invasive, and not invasive, plant species. Although many studies report that non native invasive species become dominant because they experience positive feedback (e.g., Klironomos 2002 Reinhart et al. 2003 Agrawal et al. 2005 Wolfe and Klironomos 2005 van Grunsven et al. 2007 ), there is also evidence that the direction and strength of plant s oil feedbacks are not simply determined by where plant hosts originated from ( Mackay and Kotanen 2008 Beest et al. 2009 ). In fact it is well established that agricultural monocultures of non native plant species accumulate heavy loads of pathogens that limit their productivity (Ghorbani et al. 2008) Thus, as non native plants nd increase their abundance in the novel ranges for a prolonged time (Lankau et al. 2010) they may accumulate species specific soil antagon istic organisms and experience negative, instead of positive feedbacks (Beest et al. 2009) Besides plant soil feedbacks are not the only mechanism that leads to dominance of invasive species, as particular physiological, reproductive, defen se, etc., characteristics may also giv e them a competitive advantage over other plant species (Van Kleunen et al. 2010) Here we intended to test how plant soil feedbacks va ry across native, non native, invasive, and not invasive plants by comparing plant soil interactions across habitats dominated by

PAGE 92

92 single resident invasive and not invasive non native plant species vs. habitats containing species rich arrays of native plant species. Our aims were to compare plant soil feedbacks driven by 1) AMF vs. non AMF soil microbes from these contrasting habitats and 2) non native plants that dominate agricultural lands (pasture grass and coffee) vs. native tree species common in speci es rich pre montane tropical forest fragments. To do this, we inoculated Brachiaria grass (dominating species in pastures), coffee (dominating species in coffee monocultures), and 8 native forest tree species with either AMF spore cultures or a microbial f iltrate amplified on each of the 10 species inoculated with SMC from their respective habitats (e.g. coffee with SMC from coffee plantations, native plants with SMC from forests). Brachiaria grass has been widely planted in the area of study for cattle pas tures and is a common and very noxious weed i n coffee and other plantations. T herefore it is considered an invasive species. Coffee is extensively plan ted in the area of study. Although it is not considered a weed nor an invasive species, it successfully c olonizes the understory of forest fragments ( C. Pizano pers. obs. ). We then estimated feedback for AMF and non AMF separately between the individual species with other species (e.g. Brachiaria grass with coffee, a native pioneer species with Brachiaria gra ss, etc.) or species groups (e.g., Brachiaria grass with native pioneer species). We expected plant soil feedback drive n by non AMF soil microbes would be stronger compared to feedback caused by AMF. In addition, we predicted that Brachiaria grass (a non n ative invasive grass) would experience positive feedbacks, while non invasive or native species, such as coffee and forest tree species, would experience negative feedbacks

PAGE 93

93 Methods Study site and species This study was conducted in the agricultural regi on of the Central Cordillera of The climate is tropical humid, with an average annual temperature of 21C, and an annual rainfall of 2550 mm concentrated in two wet seasons (March to July and September to December) (Guzmn Martnez et al. 2006). Soils are Udands (And isols) (Or tz Escobar et al. 2004) The landscape is dominated by three contrasting habitat types: occasionally fertilized pastures of low diversity (mainly African grasses Pennisetum clandestinum Melinis minutiflora and Brachiaria spp. ), highly fertilized, low di versity s un grown coffee plantations (coffee monocultures) and unmanaged fragments of pre montane tropical forest (Orrego et al. 2004a ) The forest fragments are usually small (1 30 ha), are mostly dominated by early and mid successional plant species and have high liana densities. Recruitment of native species in the understory appears to be low, where seedlings of non nati ve species such as Coffea arabica (coffee) and Musa veluntina (pink banana) are common (C. Pizano pers. obs.). Native plant species were se lected based on the their abundance in the region (Orrego et al. 2004b ) sampling a wide range of seed size and life histories (Table 4 1 Appendix A ). Brac hiaria brizantha (Brachiaria grass) and Coffea arabica (coffee) were chosen because they are the dominating species in the two most extensive human modified habitat types in the region (Table 4 1). Pastures were never treated with fungicide, while coffee p lantations are treated with fungicide only when soil pathogens such as Rhizoctonia solani (Gaitn 2003), Rosellinia bunodes (Castro Caicedo 2003), Rosellinia pepo (Castro Caicedo 2003), and Ceratocystis fimbriata (Castro Caicedo

PAGE 94

94 2003a) become prominent in an area. To sample across heterogenous altitude and climatic conditions in the study area, five private farms with similar conditions each containing a pasture, a coffee plantation, and a forest fragment (Table 4 2) were chosen to collect soil for the gree nhouse exp eriment. At each of the 15 site habitat combinations five samples of 200 g of mineral soil (5 15 cm in depth) each were taken and pooled in a composite sample from which a subsample was analyzed. Soils across the three habitats had similar level s of pH, organic matter content, and some nutrients (N, K, Mg), but forest fragment soils contained only 10 12 % of P, and significantly greater Ca content than pasture and coffee plantation soils (Table 4 3). In addition, light levels in pastures and coff ee plantations were approximately 30 times higher than in forest fragments (Table 4 3). Seed preparation and soil medium for growing seedlings Seeds of Brachiaria grass and coffee were provided by Cenicaf, while seeds of the eight native plant species (T able 4 1) were collected from at least 3 parent trees in forest fragments. Seeds were surface sterilized (0.6 % sodium h ypochlorite for 15 minutes) and germinated in trays containing steam sterilized soil 3:2 soil and river sand mixture. The soil was colle cted from an open area near the greenhouses in Cenicaf (Colombian National Research Center for Coffee, Chinchin, Caldas, Colombia), and it was classified as an acrudoxic melanudand from Chinchin unit (Ortz Escobar et al. 2004) with moderately fertile chemical characteristics (5.4 pH, 0.1 % N calculated, 1.9 % OM Walkley Black colorimetry, 7 mg kg 1 P Bray II Bray Kurtz colorimetry, 0.37 mg kg 1 K Ammonium acetate 1N, 3.4 mg kg 1 Ca Ammonium acetate 1N, and 1.1 cmol kg 1 Mg Ammonium acetate 1N).

PAGE 95

95 Initial inoculum collection We collected whole s oil inocula (i.e. roots, rizhospher e soil, and associated organisms) from pastures, coffee plantations, and forest fragments at 5 farm sites each having these three habitats (total of 15 site habitat combinations) (Table 4 2). Roots and rhizosphere soil (5 15 cm in depth) were collected fro m 5 random locations within each habitat at each site and pooled together. We then combined and homogenized the sampled soil from the 5 sites for each habitat type. AMF inoculum amplification Each inoculum was amplified in two steps. In the first stage, in dividually potted plants of the 10 species (Table 4 1) were inoculated with soils collected in their respective habitat types as above (i.e. Brachiaria grass with the soil from pastures, coffee with the soil from coffee plantations, and native species with the soil from forest fragments) (Fig. 4 1, step 1). The same sterilized soil and sand mixture as used to germinate the seeds filled 70 % of 1 L pots to where an equal quantity (100 mL) of the whole soil inoculum from a particular habitat was added. Ten se edlings per species were grown for 5 months in the greenhouse (Fig. 4 1, step 2) after which we harvested and mixed their roots and pot soil to process 8 subsamples of 20 g of soil from each plant inoculum combination (Fig. 4 1, step 3). Each soil subsample was washed with water through a series of sieves, and we collected what remained in the 45 m diameter sieve (mostly AMF spores) (Klironomos 2002) (Fig. 4 1, step 4a). In order to clean the AMF spores and separate them from other structures and possible contaminants, we used sucrose density gradient centrifuga tion (Daniels and Skipper 1982) and then isolated individual spores using micro tweezers. Once all clean spores were obtained, they were used to inoculate seedlings previously germinated on sterilized soil of

PAGE 96

96 Brachiaria decumbens and Cecropia angustifolia (6 seedlings per species) (Fig. 4 1, step 5a) To do this, we suspended the clean AMF spores in distilled water and added this solution directly onto the plant root system. The same process was repeated for each inoculum obtained from the 10 host species, t herefore we had a total of 60 seedlings of Brachiaria decumbens and Cecropia angustifolia each inoculated with 10 different AMF communities. These species were selected to amplify the AMF inoculum because they grow fast and produce high amounts of AMF spor es in a relatively short time (Mangan et al. 2010a) In addition, because AMF sporulate at different rates on different host species (Jansa et al. 2008) it allowed us to standardize the amo unt of spores produced in each species AMF inoculum. Inoculated seedlings grew for 5 months in the greenhouse (Fig.4 1, step 6) after which they were harvested and their soil and roots were used as inoculum in the greenhouse experiment (Fig. 4 1, step 7) For each inoculum source the seedlings of B. decumbens and C. angustifolia were harvested, and their roots and soil were pooled together (Fig. 4 1, step 7). Filtrate inocula These inocula were produced by amplifying soil microbes collected in a filtrate f rom the 25 m diameter sieve that excluded large AMF spores (Klironomos 2002) The scope was to create non AMF inocula. Although AMF hyphal contamination was possible, this was the practical method used in o ther studies (e.g., Klironomos 2002 McCarthy Neumann and Kobe 2008) to isolate a representative community of non AMF soil organisms including microbial pathogens like bacteria and non AMF fungi. The filtrate was used to inoculate seedlings of the respective species used in the initial

PAGE 97

97 amplification (Fig. 4 1, step 1). For example, we inoculated Juglans neotropica seedlings with filtrate obtained from pots with Jugl ans neotropica seedlings. These seedlings were also grown for 5 months in the greenhouse after which they were harvested and their roots and soil were collected to use as inoculum in th e greenhouse experiment Greenhouse experiment growth conditions and t reatments Each 1 L pot was filled with 70 % steam sterilized (for 2 hrs) soil 3:2 soil and river sand mixture. The soil was collected from an open area at CIAT (International Center for Tropical Agriculture) in Palmira, Valle, Colombia, and was shown to h ave moderately fertile chemical characteristics (6.8 pH, 0.05 % N, 0.6 % OM Walkley Black colorimetry, 42 mg kg 1 P Bray II Bray Kurtz colorimetry, 0.35 mg kg 1 K Ammonium acetate 1N, 3.7 mg kg 1 Ca Ammonium acetate 1N, and 2.8 cmol kg 1 Mg Ammonium acetat e 1N). Then we added an equal amount (100 mL) of AMF or filtrate inoculum consisting of a mixture of roots and soil previously produced in the greenhouse (Fig. 4 1, step 7). Each treatment combination from the combination of inoculum species (10 species) inoculum type (AMF, filtrate) plant species (10 plant species) was replicated in 4 pots and plants were randomly distributed in the growing house. We additionally grew 4 seedlings of each of the 10 plant species as non inoculated sterile plants. Thus, w e had a total of 840 plants in the experiment. Plants were under 20 % light for 75 days. We measured the leaf area of each plant at the beginning of the experiment to use it as covariate of final total dry weight in the statistical analyses. In addition, w e also quantified AMF root colonization in the roots of half of the seedlings; 1 g of fine roots were washed with running water, cleared with 3 % KOH and 3 % H 2 O 2 acidified with 3

PAGE 98

98 % HCl, washed with running water, and stained with 0.05 % tryplan blue in l actoglycerol solution (Zangaro et al. 2000) Root segments were then mounted on slides (2 slides/seedling) and presence or absence of AMF and non AMF soil organisms was recorded at 200 intersect points for each slide (for a total of 400 intersect points for each individual). Statistical analyses T he did an initial analysis to estimate the net effects of AMF and the microbial filtrate inocula relative to the non inoculated sterile (control) on plant growth. More specifically, we ran an analysis of covariance (ANCOVA) on final seedling biomass that included inoculum type (sterile, AMF and filtrate) and plant species (and their interaction) as fixed effects, and initial leaf area and days in experiment as covariates. This facilitated the analysis in seedling age heterogeneity as dead seedlings were re placed during the first month of the experiment. We then did two sets of analyses to examine the effects and feedback of 1) AMF inocula, and 2) filtrate inocula on growth of the 10 plant species. For both of these analyses we did an initial analysis of cov ariance (ANCOVA) on final seedling biomass that included inoculum species (10 species) and plant species (10 species) and their interaction as fixed effects, and initial leaf area and days in the experiment as covariates (non inoculated sterile plants were not included). We then calculated feedback for AMF and the filtrate separately by using a priori plant (Mangan et al. 2010b) These contrasts not only compare growth of each plant species with conspecific vs. heterospecific inoculum, but also growth of other plant species with conspecific vs. heterospecific inculum (two way analysis). Given that

PAGE 99

99 Brachiaria grass is the dominating species in pastures in the region of study, pai rwise feedback between this grass and all other nine plant species was calculated as a proxy of feedback in pastures ( Bever 200 3 Mangan et al. 2010b ). Similarly, pairwise feedback between coffee and all other species was estimated as a proxy of feedback in coffee plantations ( Bever 2003 Mangan et al. 2010b ). In contrast, feedback in forest fragments had to be estimated considering all possible combinations between all ten plant species, as there are no dominating plant species in these habitats. Accordingly, feedback in forest fragments was calculated by pairwise feedback for each possible pair of species and then averaging all pairwise feedbacks involving each species ( Bever 2003 Mangan et al. 2010b ). Finally, to compare AMF colonization proportion across the different inocula type s, an analysis of variance (ANOVA) was conducted for half of plants for which we measured AMF colonization (2 replicates per treatment). This model included inoculum species (10 species), inoculum type (AMF, filtrate), and plant species (and their interact io ns) as fixed effects AMF colonization proportion was transformed with arcsine square root to improve normality. Plants in the sterile treatment were not included in this analysis as none of the examined plant roots showed evidence of AMF colonization. S tatistical analyses were done with JMP version 8.0 (SAS Institute Inc., SAS Campus Drive, Cary, NC USA 27513 ). Results Overall, plants grew better when inoculated with AMF than when grown on sterile soil, but better with sterile soil than when inoculated with the microbial filtrate (Table 4 4, Figure 4 2). Next, we compared plant growth across AMF and microbial filtrate inocula

PAGE 100

100 obtained from different source species (= inoculum species). Seedling biomass significantly differed among inoculum species, plan t species, and the combination of inoculum species and plant species (indicated by a significant interaction) for AMF inocula (Table 4 5). Similarly, for the microbial filtrate inocula, plant biomass significantly differed among plant species, as well as a cross different inoculum species and plant species (indicated by a significant interaction) (Table 4 6). These significant interactions meant that plant species responded differently to soil inocula from conspecific and heterospecific sources. These often reflected significantly negative feedbacks that could be attributed to sufficiently strong negative effects of conspecific inocula compared to heterospecific inocula in both (Fig. 4 3 A, B) or one in the pair wise comparisons (Fig. 4 3 C ). Feedbacks for AM F inocula were mostly neutral and non significant in pair wise tests involving Brachiaria grass (Table 4 7, Fig. 4 4A), tests involving coffee (Table 4 7, Fig. 4 4B), and tests involving forest tree species as a group (Table 4 7, Fig. 4 4C). The only excep tion was the weak negative feedback between coffee and Ochroma pyramidale reflecting a positive heterospecific inoculum effect on Ochroma and coffee growth compared to conspecific AMF (Fig. 4 4C). In contrast, feedbacks with the microbial filtrate were of ten negative and significant. Growth of Brachiaria grass, in particular, was negatively impacted by conspecific soil filtrate inocula more strongly than by heterospecific inocula obtained from other species (Fig. 4 3A, B, C) Conversely, other species tend ed to be more strongly impacted by conspecific soil filtrate inocula than by filtrate inocula obtained from Brachiaria grass. As a result, eight of the nine pair wise feedback response tests

PAGE 101

101 involving Brachiaria grass were significantly negative (Table 4 8 Fig. 4 4A). Pair wise feedback tests for soil filtrate involving coffee also showed negative feedbacks for several species pairs: coffee Brachiaria grass, coffee CecA, coffee CecT, and coffee Ochr (Table 4 8, Fig. 4 4B), in which non coffee species were a grass or native pioneer trees. Pair wise feedback tests for microbial soil filtrate involving native tree species collectively as heterospecific partners (Fig. 4 4C) showed significant negative feedbacks in two cases (Table 4 8). A strong negative feedba ck was found between Brachiaria grass and native trees, owing largely to a reduced growth of Brachiaria grass with conspecific filtrate inoculum compared to growth with heterospecific inocula. Also strong negative effects of conspecific filtrate inoculum o n growth of Cecropia angustifolia compared to heterospecific inocula from eight other species resulted in significant negative feedback (Fig. 4 4C). AMF root colonization varied greatly across inoculum species, inoculum type, plant species and different c ombinations of inoculum species, plant species and inoculum type (Table 4 9, Fig. 4 5). As expected AMF colonization was higher for the AMF inocula than for the microbial filtrate inocula for nine out of 10 plant species (Fig. 4 5A). Similarly, proportion root colonization by non AMF soil microbes also varied significantly across inoculum species, inoculum type, plant species, and two way interactions between inoculum species and inoculum type, and between inoculum species and plant species (Table 4 10). R oot colonization by non AMF soil microbes was higher in plants treated with the microbial filtrate than in those treated with the AMF inocula (Fig. 4 5B).

PAGE 102

102 Discussion The two key results of this study were that 1) plant soil feedbacks dri ven by non AMF soil microbes were much stronger than feedbacks driven by AMF (Fig. 4 4) and 2) both non native plant species that dominate agricultural lands, and native species common in natural forests, experience negative plant soil feedbacks (Fig. 4 4) Surprisingly, Br achiaria grass, an invasive non native grass that dominates pastures and colonizes other habitats (e.g. coffee plantations), experienced significant negative feedbacks for the microbial filtrate (Fig. 4 4A) instead of positive feedbacks shown for temperat e invasive exotics in other studies (e.g., Klironomos 2002, Reinhart et al. 2003, Van Grunsven et al. 2007) Similarly, feedback between coffee, a non native species, and Brachiaria grass, and between coffee and forest pioneer tree species was negative fo r the microbial filtrate, but neutral or slightly positive for forest shade tolerant tree species (Fig. 4 4B) Finally, feedback of forest tree species was mostly negative with the microbial filtrate (Fig. 4 4C) Overall, plants grew better with sterile so il than when inoculated with the soil microbial filtrate (Fig. 4 2) and that root colonization by non AMF soil organisms was higher for the microbial filtrate than for the AMF inocula (Fig. 4 5). These results suggest that the microbial filtrate included antagonistic soil microbes Pasture grass dominates the species poor habitats despite accumulation of species specialized antagonistic soil microbes. Native tree species did slightly better with these detrimental soil microbes from pastures than with those they encounter in forest fragments. Thus, negative pl ant soil feedback dynamics occur both in species poor agricultural lands, as well as in more species rich forest fragments In other words, soil antagonistic microbes specialized in common plant species accumulate regardless of habitat type with contrasting plant diversity and management regimes.

PAGE 103

103 AMF vs. non AMF soil organisms in driving feedback Feedbacks driven by AMF have been well studied because these fungi are relatively easy to isolate and exper iment with. Previous studies have shown that although AMF are not strictly species specific, their effects on plant hosts vary significantly across plant species ( Castelli and Casper 2003 Moora et al 2004 Munkvold et al. 2004 Mangan et al. 2010a ), with significant imp lications for plant community composition ( Bever 2002a Klironomos 2002 Van d er Heijden 2004 Vogelsang et al. 2006) Similarly, when the effects of non AMF organisms were examined with microbial filtrate, species specific antagonistic soil microbes that mediate significant plant so il feedbacks tend to accumulate with each host plant (e.g., Klironomos 2002 McCarthy Neumann and Kobe 2008 McCarthy Neumann and Kobe 2010) We found that plant soil feedbacks via AMF were neutral and mostly i nsignificant across plant species from contrasting habitats (Table 4 7, Fig. 4 4). The only significant plant AMF feedback was between coffee and the pioneer native tree Ochroma pyramidale s AMF compared to their own AMF. Thus, it seems like forest fragments and agricultural lands in the study region share similar communities of mycorrhizal fungi, with similar, positive effects across different plant species (Table 4 4, Fig. 4 2). An alterna tive explanatio n for the general neutrality of AMF plant soil feedbacks, is that the AMF inocula represented a biased subset of AMF community in the soil. The AMF inocula were prepared from individually picked AMF, which probably favored AMF strains with c opious spore Verbruggen and Kiers 2010) selecting for only one of many AMF functional types from the three sampled habitats (Munkvold et al. 2004)

PAGE 104

104 Opposite to AMF plant soil feedbacks, feedback mediated by the soil microbial filtrate was significantly negative for most plant species. This was particularly true for Brachiaria grass, which showed strong negat ive feedback with the filtrate for 7 out of 9 plant species (all but Solanum aphynodendrum and Garcinia madrunno ) (Table 4 8, Fig. 4 4A), and for coffee, which had significant negative feedback for four out of nine species (Table 4 8, Fig. 4 4B). Similarly although average filtrate feedback between Brachiaria grass, coffee and native tree species was only slightly negative for most species, it was significantly negative for the grass and for the pioneer tree Cecropia angustifolia (Table 4 8, Fig. 4 4C). Th e fact that plants grew overall better with sterile soil than with the microbial filtrate (Table 4 4, Fig. 4 2) indicates the presence of antagonistic soil microbes. Furthermore, we also found that AMF colonization was lower, and non AMF colonization highe r, in plants treated with the filtrate than in those inoculated with AMF (Table 4 9, Table 4 10, Fig. 4 5), indicating that the microbial filtrate effectively contained more antagonistic soil microorganisms than the AMF inocula. These results suggest that while pastures and coffee plantations accumulate species specific antagonistic soil microbes that mediate negative plant soil feedbacks, forest fragments seem to hold more generalist soil microbes that mediate only weak plant soil feedbacks. Perhaps forest fragmentation has resulted in a loss of the high biodiversity of species specific soil microbes (Gilbert and Hubbell 1996) that mediate negative feedback in undisturbed tropical forests (Mangan et al. 2010b) Feedback across habitats of contrasting diversity and abundance of native, non native, invasive, and not invasive plant species Studies in both temperate ( Klironomos 2002 Casper et al. 2008 Petermann et al. 2008 Harrison and Bardgett 2010 ) and species rich plant communities in the tropics

PAGE 105

105 (Mangan et al. 2010b) species rich plant communities have shown that negative feedback between plants and soil organisms is weaker for more abundant plant species. By a similar logic, plants introduced to new ranges may become invasive by escaping their home soil enemies (enemy release hypothesi s). For example, Centaurea melitensis and Centaurea malucosa two European species invasive in North America, experience negative feedback in their native range but show positive feedback in their new territory (Callaway et al. 2004) Similarly, the invasion of Prunus serotina (black cherry) in north western Europe is facilitated by soil microbes in positive feedbacks, wh ile growth of this species is limited by the soil community that develops in the roots of conspecific in its native range in USA (Reinhart et al. 2003) However, non native species may accumulate antagonistic soil organisms and may experience negative feedback (Nijjer et al. 2007) after being in their new ranges for a prolonged time, and that non native plant species grown as agricultural crops may accumulate heavy loads of soil pathogens and therefore experience negative feedback. Thus, plant soil feedbacks may vary across plants irrespective to their classification as native, non native and invasive and not invasive. Our results show that plant soil feedbacks were largely negative regardless of whether the species were exotic invasive, exotic non invasive, or native. Brachiaira grass, the dominating species in pastures and a highly inva sive weed in fruit orchad s (e.g., citrus and coffee) and open natural habitats (Almeida Neto et al. 2010) showed significant negative feedback with coffee and 6 out o f 8 native plant species (Table 4 7, Table 4 8, Fig. 4 4A). This goes against one of our a priori expectation that the invasiveness of this grass was related to positive plant soil feedbacks, and contradicts

PAGE 106

106 studies showing this type of feedback for non na tive invasive species ( Callaway et al. 2003 Reinhart et al. 2003 Wolfe and Klironomos 2005 van Grunsven et al. 2007) However, this African grass was introduced to the study region at least 50 years ago (Parsons 1972) since then it may have accumulated species specific antagonistic soil microbes and experiences negative feedback (Nijjer et al. 2007) Continued dominance of this grass in pastures and invasion in plantations may instead be related to other traits such as physiology, leaf area allocation, sh oot allocation, growth rate, size or fitness (Van Kleunen et al. 2010) Alternatively, Brachiaria grass may produce a root exhudate chemical that inhibits nitrification simi lar to that one recently found in Brachiaria humidicola (Subbarao et al. 2009) therefore having a competitive advantage over other plant species. Coffee, the most abundant plant in coffee plantations, also showed negative feedback with Brachiaria grass and three forest tree species (Table 4 7, Table 4 8, Fig. 4 3B, Fig. 4 4B). This non native shrub was introduced in the study region at least 150 years ago, and plantations are often dimmed by root fungal pathogens such as Rhizoctonia solani (Gaitn 2003) Rosellinia bunodes (Castro Caicedo 2 003) Rosellinia pepo (Castro Caicedo 2003) and Ceratocystis fimbriata (Castro Caicedo 2003a). Thus, it is not surprising that coffee appears to have accumulated species specific antagonistic soil organisms and exhibits negative feedback. Interestingly, w ith native tree species, coffee only showed negative feedback with pioneer, and not with shade tolerant tree species. This suggests that coffee and shade tolerant forest tree species share similar soil enemies, therefore no feedback dynamics arise between these species (Fig. 4 3D). For example, coffee plantations increase the abundance of

PAGE 107

107 Rosellinia bunodes in Puer to Rico and the fungus colonizes other Rubiaceae native plants (Lodge 2001) In contrast, coffee appears to accumulate a different array of soil microbes than tho se present in native pioneer trees, therefore negative feedback takes place among these species. Finally, there was a strong negative feedback between native tree species and Brachiaria grass, and between Cecropia angustifolia and other native trees due m ostly to a reduced growth of the grass and CecA with conspecific filtrate inoculum compared to growth with heterospecific inoculum (Table 4 7, Table 4 8, Fig. 4 4C). However, feedback between coffee and native trees, and among the other seven native tree s pecies was mostly neutral, and only slightly negative. Thus, the negative feedback shown for undisturbed tropical forest (Mangan et al. 2010b) may have been disrupted with forest fragmentation and land use change. For instanc e, antagonistic soil organisms that become common in agricultural lands may invade forest fragments, displacing other native soil organisms (Gilbert and Hubbell 1996) responsible for driving negative feedbacks. Alternatively, the low signal of feedback for most native species may be due to the low replication, or short running of the experiment. Applied ecological implications: Biological interactions that play a key role in maintaining plant diversity can become disrupted when natural ecosystems are fragme nted and replaced by agriculture (Bascompte 2009, Kiers et al. 2010) For example, previous studies have found that seed dispersal declines (Wijdeven and Kuzee 2000 Hooper et al. 2005) while seed predation intensifies (Holl a nd Lulow 1997 Holl et al. 2000) in fragmented, compared to undisturbed tropical forests. Likewise, our results indicate that plant soil feedbacks are substantially altered by fore st fragmentation by land conversion to agriculture First,

PAGE 108

108 although the importance of plant AMF feedbacks for plant community composition have been demonstrated both in the greenhouse (Klironomos 2002, Mangan et al. 2010a) and in the field (Pringle and Bever 2008) we found little evidence that suggest similar feedbacks with AMF This may indicate that AMF communities in the study region are impoverished probably due to the cultivation of single non native plant species in monoculture (Kabir et al. 1997 Menendez et al. 2001 Oehl et al. 2003) and agricultural practices such as tillage and soil fertilization (Helgason et al. 1998 Menendez e t al. 2001) Additionally, AMF community composition may be significantly altered by forest fragmentation (Mangan et al. 2004) In contrast to plant AMF dynamics, we found strong negative feedbacks driven by a microbial filtrate comprising antagonistic soil m icrobes. This was mainly true for Brachiaria grass and coffee, showing there is an accumulation of species specific soil pathogens in pastures and coffee plantations that hinder agricultural production, but may enhance the regeneration of forest trees. In contrast, feedbacks mediated by the filtrate were only slightly negative for native tree species from forest fragments, implying that SMC in these degraded forests may have lost key drivers of plant soil feedbacks that are present in undisturbed forests (Mangan et al. 2010b) Together, these results provide strong evidence that parallel to other biotic interactions, plant soil dynamics are also severely disrupted with habitat alteration. Further studies are needed to effectiv ely comp are the composition and role of SMC and their different components across heterogeneous landscapes, including human modified landscapes.

PAGE 109

109 Table 4 1 General characteristics of the ten plant species used in the greenhouse experiment. Seed dry mass (mean S.D) was measured from 25 50 seeds/species dried for 3 days at 60 C. Species Family Plant form Habitat Seed dry mass (g) Abbreviat ion Brachiaria brizantha Poaceae Brachiari a grass (crop) P 1 0.0077 0.0013 Bra Coffea arabica Rubiaceae Coffee (c rop) C 2 0.22 0.023 Cof Cecropia angustifolia Cecropiaceae Pioneer F 3 0.0011 0.00028 CecA Cecropia telealba Cecropiaceae Pioneer F 3 0.00039 0.00012 CecT Ochroma pyramidale Bombacaceae Pioneer F 3 0.0037 0.0012 Ochr Solanum aphynodendrum Solanac eae Pioneer F 3 0.0015 0.00053 Sol Garcinia madrunno Clusiaceae Shade tolerant F 3 2.92 0.77 Gar Gustavia superba Lecythidaceae Shade tolerant F 3 10.55 2.78 Gus Juglans neotropica Juglandaceae Shade tolerant F 3 26.04 14.49 Jug Retrophyllum ros pigliosii Podocarpaceae Shade tolerant F 3 0.86 0.18 Ret 1 Pastures 2 Coffee plantations 3 Forest fragments

PAGE 110

110 Table 4 2. Characteristics of the farms where the initial soil inocula were collected (Guzmn Martinez et al 2006). At each farm there were three habitats: a pasture (P), a coffee plantation (C), and a forest fragment (F). Farm (block) Geographic coordenates Altitude (m.s.l) Mean annual precipitation (mm yr 1 ) Mean annual temperature ( C) Habitat Habitat patch size (ha) Cenicaf 0500 N 75 36 W 13 80 2733 20.9 P 0.5 C 0.5 F 40 Playa rica 0500 1290 2750 20.7 P 10 C 60 F 30 Alto espaol* 0456 N 7542 W 1720 3140 18.8 P 1.0 C 2.0 F 0.3 Naranjal 0459 N 7539 W 1400 3137 21.4 P 22 C 38 F 27 La Argentina 0502 N 7541 W 1354 2935 19.9 P 0.5 C 100 F 1.5

PAGE 111

111 Table 4 3. Mean ( SE) soil pH, organic matter (OM) content, and nutrient content of the three habitats where soil inocula were collected. Habitat pH 1 OM 3 (%) N 2 (%) P 4 (mg/kg)* K 5 (cmol/kg) Ca 5 (cmol/kg)* Mg 5 (cmol/kg) Pastures 5.6 ( 0.1) 7.2 ( 1.4) 0.3 ( 0.04) 54.4 ( 25.3) 0.6 ( 0.2) 5.6 ( 0.7) 2.1 ( 0.4) Coffee plantations 5.0 ( 0.3) 9.6 ( 1.7) 0.4 ( 0.06) 42.0 ( 16.7) 0.3 ( 0.1) 4.4 ( 1.1) 1.7 ( 0.6) Forest fr agments 5.5 ( 0.2) 12.4 ( 2.1) 0.5 ( 0.06) 5.6 ( 0.4) 0.4 ( 0.05) 8.3 ( 2.3) 2.4 ( 0.7) Notes: Log transformed data was analyzed with one way ANCOVA including block as a random factor and habitat as a fixed factor. 1 pH: Potentiometer soil: water 1: 1 2 N (total): Calculated 3 OM: Walkley Black colorimetry 4 P: Bray II color imetry Bray Kurtz 5 K, Ca, Mg: Ammonium acetate 1N Significant differences between habitat types P < 0.05

PAGE 112

112 Table 4 4. Growth response of 10 plant species (Brachiaria grass, coffee, and 8 forest tree species; Table 4 1) to sterile (non inoculated; control), AMF (isolated from the rhizosphere of these 10 plant species), and microbial filtrate (isolated from the rhizosphere of these 10 plant species) inocula in the greenhouse. Total biomass(error df = 766 ) Source Df F P Inoculum type 2 5.0 0.007 Plant species 9 94.0 <0.001 Inoculum type Plant species 18 2.3 0.0015 Initial leaf area 1 540.0 <0.001 Days in experiment 1 305.1 <0.001 Notes: ANCOVA was used to analyze log 10 transformed final biomass, with log 10 transformed initial leaf area and days in experiment (dead seedlings were replaced during the first month of the experiment) as covariates. Table 4 5. Growth response of 10 plant species (Brachiaria grass, coffee, an d eight native tree species; Table 4 1) to AMF isolated from the rhizosphere of these Total biomass (error df = 285 ) Source Df F P Inoculum species 9 2.4 0.0113 Plant species 9 156.2 <0.001 Inoc ulum species Plant species 81 1.4 0.0314 Initial leaf area 1 171.7 <0.001 Days in experiment 1 12.9 <0.001 Notes: ANCOVA was used to analyze log 10 transformed final biomass, with log 10 transformed initial leaf area and days in experiment (dead seedlin gs were replaced during the first month of the experiment) as covariates.

PAGE 113

113 Table 4 6. Growth response of 10 plant species (Brachiaria grass, coffee, and ei ght native tree species; Table 4 1 ) to a microbial filtrate isolated from the rhizosphere of th Total biomass (error df = 285 ) Source Df F P Inoculum species 9 1.2 0.3015 Plant species 9 240.7 <0.001 Inoculum species Plant species 81 1.7 <0.001 Initial leaf area 1 220.0 <0.001 Days in experiment 1 24.2 <0.001 Notes: ANCOVA was used to analyze log 10 transformed final biomass, with log 10 transformed initial leaf area and days in experiment (dead seedlings were replaced during the first month of the experiment) as covariates.

PAGE 114

114 Table 4 7. Statistical summary of the strength of a priori contrasts (F values followed by P values in parenthesis) testing AMF feedback between Brachiaria grass and other 9 plant species (coffee, and 8 forest tree species; Table 4 1), coffee and ot her 9 plant species, and average feedback between 8 forest species, Brachiaria grass, and coffee. Heterospecific partner species for a priori tests for AMF inocula Brachiaria grass Coffee 8 native trees (or 7 heterospecific native trees) Brachiaria gras s NA 0.31 (0.58) 0.045 (0.83) Coffee 0.31 (0.58) NA 0.15 (0.69) CecA 1.52 (0.22) 0.28 (0.60) <0.001 (0.98) CecT 0.15 (0.70) 1.18 (0.28) 0.12 (0.73) Ochr 0.73 (0.40) 5.05 (0.025) 0.003 (0.96) Sol 0.26 (0.61) 0.006 (0.94) 1.3 (0.26) Ret 0.17 (0.68) <0. 001 (0.99) 0.37 (0.55) Gar <0.001 (0.98) 0.009 (0.93) 0.1 (0.75) Gus 0.66 (0.42) 1.43 (0.23) 0.16 (0.69) Jug 0.53 (0.47) 0.075 (0.79) 0.054 (0.82)

PAGE 115

115 Table 4 8. Statistical summary of the strength of a priori contrasts (F values foll owed by P values in parenthesis) testing filtrate feedback between Brachiaria grass and other 9 plant species (coffee, and 8 forest tree species; Table 4 1), coffee and other 9 plant species, and average feedback between 8 forest species, Brachiaria grass, and coffee. Heterospecific partner species for a priori tests for filtrate inocula Brachiaria grass Coffee 8 native trees (or 7 heterospecific native trees) Brachiaria grass NA 11.73 (<0.001) 17.6 (<0.001) Coffee 11.73 (<0.001) NA 1.61 (0.21) CecA 1 8.30 (<0.001) 4.53 (0.034) 2.87 (0.009) CecT 6.34 (0.012) 3.63 (0.058) 1.82 (0.18) Ochr 8.90 (0.003) 3.31 (0.07) 1.59 (0.21) Sol 0.60 (0.44) 0.094 (0.76) 1.44 (0.23) Ret 6.9 (0.009) 0.14 (0.70) 0.39 (0.53) Gar 3.76 (0.054) 0.74 (0.39) 1.63 (0.2) Gus 4.64 (0.032) 2.46 (0.12) 0.08 (0.78) Jug 5.45 (0.02) 0.23 (0.64) 1.22 (0.27)

PAGE 116

116 Table 4 9. Proportion of roots colonized by AMF for 10 host plant species (Brachiaria grass, coffee, and 8 forest tree species; Table 4 1) as affected by inoculum species (10 plan species), inoculum type (AMF, f il trate ), and plant species (and interactions) in the greenhouse. Total biomass (error df = 199 ) Source Df F P Inoculum species 9 11.6 <0.001 Inoculum type 1 162.8 <0.001 Plant species 9 53.2 <0.001 Inoc ulum species Inoculum type 9 3.9 <0.001 Inoculum species Plant species 81 1.9 <0.001 Inoculum type Plant species 9 2.4 0.015 Inoculum species Inoculum type Plant species 81 1.9 <0.001 Table 4 10. Proportion of root s colonized by of non AMF for 10 host plant species (Brachiaria grass, coffee, and 8 forest tree species; Table 4 1) as affected by inoculum species (10 plan s pecies), inoculum type (AMF, filtrate ), and plant species (and interactions) in the greenhouse. Total biomass (error df = 1 99 ) Source Df F P Inoculum species 9 10.5 <0.001 Inoculum type 1 4.9 0.028 Plant species 9 5.4 <0.001 Inoculum species Inoculum type 9 3.1 0.002 Inoculum species Plant species 81 1.8 <0.001 Inoculum type Plant species 9 1.5 0.150 Inoculum sp ecies Inoculum type Plant species 81 1.1 0.390

PAGE 117

117 Figure 4 1 Flow diagram showing the procedure that we used to produce the inocula for setting up the experiment in the greenhouse.

PAGE 118

118 Figure 4 2. Proportion of seedling biomass whe n inoculated with AMF (open circles) or a microbial filtrate (closed circles) in comparison to biomass when grown on sterile soil for Brachiaria grass, coffee, and 8 forest tree species (Table 4 1). Circles indicate the mean ( SE) total seedling biomass w ith AMF or the filtrate divided by seedling biomass with sterile soil.

PAGE 119

119 Figure 4 3. Examples of significant negative feedback (A, B, C) and no feedback (C) between seedlings of different s pecies. The graphs represent growth of two

PAGE 120

120 species with a microbial filtrate collected from conspecific versus heterospecific seedlings. The resulting interaction between seedling species and inoculum source species was used to define the strength and dire ction of feedback. For example, in graph A, both Brachiaria grass and Cecropia angustifolia grew better with the microbial filtrate from each other than with the microbial filtrate from their own roots (same than in graph B), therefore there is significant negative feedback between these two species. In graph C, feedback is significant, but is driven by Brachiaria grass growing better with the microbial filtrate from Gustavia superba compared to conspecific filtrate; growth of Gustavia superba was similar a cross conspecific and heterospecific microbial filtrate. In graph D, both Brachiaria grass and Solanum aphynodendrum grew better with the microbial filtrate isolated from Brachiaria grass roots than with the filtrate from S. aphynodendrum roots, therefore there is no feedback.

PAGE 121

121 Figure 4 4. Feedback mediated by AMF (open bars) and a microbial filtrate (closed bars) between (A) the pasture grass Brachiaria brizantha and nine other plant species (coffee and 8 forest tree speci es; Table 4 1) (pairwise feedback), B) coffee and nine other plant species (Brachiaria grass an d 8 forest tree species ) (pairwise feedback ), and C) 8 forest tree species and Brachiaria grass, or coffee or other seven heterospecific forest species (average feedback per species). Bars indicate the value ( SE) of a priori contrasts calculated from all replicates, comparing the growth of each species when inoculated with AMF or filtrate from each other grew with heterospecific vs. conspecific AMF and filtrate. Means that differ from zero are indicated by asterisks (* P < 0.05; ** P < 0.01, *** P < 0.001).

PAGE 122

122 Figure 4 5. Proportion root colonizatio n of A) AMF and B) non AMF soil microbes in seedlings of Brachiaria grass, coffee, and 8 forest tree species (Table 4 1) when inoculated with AMF (open circles) or a microbial filtrate (closed circles) isolated from the roots of these 10 plant species. Cir cles indicate the mean ( SE) AMF root colonization proportion quantified for half of the seedlings in the experiment and averaged for each inoculum type (AMF and filtrate).

PAGE 123

123 CHAPTER 5 CONCLUSION My research was motivated by a desire to understand how the replacement of tropical forests for agriculture has mo dified the composition of microbial soil the regeneration of pre montane native forest species and agricultural production I work ed in the Central Andes of Colombia, where I studied the dynamics between soil organisms and plant species from the three dominating habitats in my study region: pastures (Brachiaria grass), sun grown coffee plantations (coffee), and forest fragments (forest tr ee species). In addition, I addressed the effects of two main components of soil microbial communities (SMC), mutualistic arbuscular mycorrhizal fungi (AMF), and antagonistic non AMF soil organisms, on plant growth. My expectation was to find that SMC fro m forest fragments and agricultural lands would be significantly different, having differential effects across plant species, as these habitats not only have contrasting abiotic soil conditions, but also differ greatly in the composition of the plant commu nities they hold. On one hand, pastures and coffee plantations are heavily fertilized (Farfn Valencia and Mestre Mestre 2004, Farfn Valencia and Mestre Mestre 2005) while forest fragments are unmanaged. These management practices affect SMC; it has been shown that fertilization reduces SMC diversity in general (Giller 1996 Giller et al. 1997) and mycorrhizal inoculum potential (Janos 1980, Fischer et al. 1994, Picone 2000, Zangaro et al. 2000). On the other hand, agricultural monocultures are dominated b y single, non native plant species, while forest fragments are comprised of highly diverse plant communities. Thus, given that plant identity determines the composition of AMF communities and the presence of

PAGE 124

124 particular soil organisms in the soil (Johnson e t al. 1992, Bever et al. 1996, Westover et al. 1997, Eom et al. 2000, Lovelock et al. 2003, Lovelock and Ewel 2005), soil organisms that co lonize crop species are likely to dominate SMC in agricultural lands while forests could hold more diverse SMC In m y dissertation I explored three main hypotheses: 1) SMC from pastures, coffee plantations, and forest fragments have a differential and significant effect on growth and survival of plants from these contrasting habitats, 2) plants would grow better with bo th AMF and non and 3) plant soil feedbacks with AMF and non AMF soil organisms would be positive for non native plants from agricultural monocultures, while being negative for native forest species. To test these hypotheses, I set up three greenhouse experiments, and one field experiment in which I exposed plants from pastures, coffee plantations, and fo rest fragments to SCM from these contrasting habitats. SMC from pastures, coffee plantations, and forest fragments had significantly different effects on plants both in the greenhouse and in the field, suggesting that SMC diverge among habitats. Furthermor e, fast growing plant species (Brachiaria grass and pioneer forest trees) benefited from away compared to home SMC, while slow growing shade tolerant forest tree species benefited the most from home SMC. When testing for the effects of communities of AMF a nd non AMF microbes on plants from the three studied habitats, I found that most plant species grew significantly better with non AMF microbes from away, compared to home habitats, while showing limited response to AMF from different habitats. Similarly, p lant soil feedbacks were strong for non AMF

PAGE 125

125 organisms, but mostly insignificant for AMF. Finally, feedbacks with non AMF soil organisms were significantly negative for non native species Brachiaria grass and coffe, but overall neutral for native tree speci es. These results suggest that species composition of soil microbial communities greatly differs between fertilized agricultural monocultures and diverse native forest fragments in the Central Cordillera coffee grown region of Colombia. Overall, it seems l ike these contrasting habitats hold similar communities of AMF. In contrast, antagonistic non AMF soil microbes appear to be widespread across contrasting habitat types, limiting the growth of their host plants where these are more abundant. Thus, this stu dy urges for further research on SMC, and in particular on non AMF soil microbes, a very poorly studied group of microbes that yet have shown to have major impacts on plants.

PAGE 126

126 APPENDIX ILLUSTRATIONS OF STU DY SPECIES Figure A 1. Illustration of a Brachi aria brizantha seedling by Camila Pizano.

PAGE 127

127 Figure A 2. Illustration of a Coffea a rabica seedling by Camila Pizano

PAGE 128

128 Figure A 3. Illustration of a Cecropia angustifolia seedling by Camila Pizano.

PAGE 129

129 Figure A 4. Illustration of a Cecropia telealba s eedling by Camila Pizano.

PAGE 130

130 Figure A 5. Illustration of a Ochroma pyramidale seedling by Camila Pizano. Figure A 6 Illustratio n of a Solanum aphynodendrum seedling by Camila Pizano.

PAGE 131

131 Figure A 7 Illustration of a Siparuna aspera seedling by Ca mila Pizano.

PAGE 132

132 Figure A 8 Illustration of a Retrophyllum rospigliossii seedling by Camila Pizano.

PAGE 133

133 Figure A 9. Illustration of a Garcinia madrunno seedling by Camila Pizano.

PAGE 134

134 Figure A 10. Illustration of a Gustavia superba seedling by Camila Pizan o.

PAGE 135

135 Figure A 11 Illustration of a Juglans neotropica seedling by Camila Pizano.

PAGE 136

136 LIST OF REFERENCES Agrawal, A., P. Kotanen, C. Mitchell, A. Power, W. Godsoe, and J. Klironomos. 2005. Enemy release? An ex periment with congeneric plant pairs and diverse above and belowground enemies. Ecology 86:2979 2989. Aldrich Wolfe, L. 2007. Distinct mycorrhizal communities on new and established hosts in a transitional tropical plant community. Ecology 88:559 566. Alle n, E., M.E. Allen, and A. Gomez Pompa. 2005. Effects of mycorrhizae and nontarget organisms on restoration of a seasonal tropical forest in Quintana Roo, Mexico: Factors limiting tree establishment. Restoration Ecology 13:325 333. Almeida Neto, M., P. I. P rado, U. Kubota, J. M. Bariani, G. H. Aguirre, and T. M. Lewinsohn. 2010. Invasive grasses and native Asteraceae in the Brazilian Cerrado. Plant Ecology 209:109 122. Arnold, A., Z. Maynard, G. Gilbert, P. Coley, and T. Kursar. 2000. Are tropical fungal end ophytes hyperdiverse? Ecology Letters 3:267 274. Ashton, P. 1969. Speciation among tropical forest trees: some deductions in the light of recent evidence. Biological Journal of the Linnean Society 1: 155 196. Augspurger C. 1983. S eed dispersal of the trop ical tree, Platypodium elegans, and the escape of its seedlings from fungal pathogens Journal o f Ecology 71:759 771 Banwart, S. 2011. Save our soils. Nature 47 4:151 152. Bascompte, J. 2009. M utualistic networks. Frontiers in Ecology and t he Environment 7: 429 436. Beest, te, M., N. Stevens, H. Olff, and W. H. Van Der Putten. 2009. Plant soil feedback induces shifts in biomass allocation in the invasive plant Chromolaena odorata Journal o f Ecology 97:1281 1290. Bell, T., R. Freckleton, and O. Lewis. 2006. P lant pathogens drive density dependent seedling mortality in a tropical tree. Ecology Letters 9:569 574 Beve r, J. D. 1994. Feeback between plants and their soil communities in an old f i eld c ommunity. Ecology 75:1965. Bever, J. D., J. B. Morton, J. Antonov ics, and P. Schultz. 1996. Host dependent sporulation and species diversity of arbuscular mycorrhizal fungi in a mown grassland. Journal of Ecology 84: 71 82. Bever, J., K. Westover, and J. Antonovics. 1997. Incorporating the soil community into plant popu lation dynamics: the utility of the feedba ck approach. Journal o f Ecology 85:561 573.

PAGE 137

137 Bever, J. 2002a. Host specificity of AM fungal population growth rates can generate feedback on plant growth Plant a nd Soil 244:281 290. Bever, J. D. 2002 b Negative fee dback within a mutualism: host specific growth of mycorrhizal fungi reduces plant benefit. Proceedings of the royal s ociety B Biological Sciences 269:2595 26 01. Bever, J. 2003. Soil community feedback and the coexistence of competitors: conceptual framewor k s and empirical tests. The New P hytologist 157:465 473. Biesmeijer, J. C., S. P. M. Roberts, M. Reemer, R. Ohlemller, M. Edwards, T. Peeters, A. P. Schaffers, S. G. Potts, R. Kleukers, C. D. Thomas, J. Settele, and W. E. Kunin. 2006. Parallel declines in pollinators and insect pollinated plants in Britain and the Netherlands. Science 313:351 354. Bradley, K., R. A. Drijber, and J. Knops. 2006. Increased N availability in grassland soils modifies their microbial communities and decreases the abundance of a rbuscular mycorrhizal fungi. Soil Biology and Biochemistry 38:1583 1595. Callaway, R., B. Mahall, C. Wicks, J. Pankey, and C. Zabinski. 2003. Soil fungi and the effects of an invasive forb on grasses: Neighbor identity matters. Ecology 84:129 135. Callaway R., G. Thelen, A. Rodriguez, and W. Holben. 2004. Soil biota and exotic plant invasion. Nature 427:731 733. Castelli, J., and B. Casper. 2003. Intraspecific AM fungal variation contributes to plant fungal feedback in a serpentine grassland. Ecology 84:32 3 336. Coley, P. 1983. Herbivory and defensive characteristics of tree species in a lowland tropical f orest. Ecological Monographs 53:209 234. Comita, L., H. Muller Landau, S. Aguilar, and S. Hubbell. 2010. Asymmetric density dependence shapes species abun dances in a tropical tree community. Science : 331 332. Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Scien ce 199:1302 1310 Daniels, B. A., and H. A. Skipper. 1982. Methods for the recovery and quantitative estimation of propagul es from soil. Pages 29 35 in N. C. Schenck, editor. Methods and principles of mycorrhizal research. American Phytopathological Soci ety, St. Paul, Minnesota, USA. Davies, B., D. Eagle, B. Finney. 1997. Soil management. Farming Press, Ipswich.

PAGE 138

138 de Voorde, v an, T. F. J., W. H. Van Der Putten, and T. M. Bezemer. 2011. Intra and interspecific plant soil interactions, soil legacies and priority effects during old field succession. Journal o f Ecology 99:945 953. Dobzhanzky, T. 1950. Evolution in the tropics. Ame rican Scientist 38: 209 221 Elton, C. S. 1958. The ecology of invasions by animals and plants. Methuen, London, UK. Eom, A.H., D. C. Harnett, and W. T. Wilson. 2000. Host plant species effects on arbuscular mycorrhizal fungal communities in tallgrass pra irie. Oecologia 122: 435 444 Farfn Valencia, F., and A. Mestre Mestre. 2004. Manejo de sombro y fertilizacin del caf en la zona central colombiana. Avances tcnicos Cenicaf 330. Farfn Valencia, F., and A. Mestre Mestre. 2005. Fertilizacin en cafeta les con sombro en la zona cafetera norte de Colombia. Avances tcnicos Cenicaf 331. Fetcher, N., B.L. Haines, R. Cordero, D.J. Lodge, L.R. Walker, D.S. Fernandez, and W.T. Lawrence. 1996. Responses of tropical plants to nutrients and light on a landslide in Puerto Rico. Journal of Ecology 84: 331 341. Fine, P., I. Mesones, and P. Coley. 2004. Herbivores promote habitat specialization by trees in amaz onian forests. Science 305:663 665. Fischer, C. R., D. P. Janos, D. A. Perry, R G. Linderman, and P. Solli ns. 1994. Mchorrhiza inoculum potentials in tropical secondary succession. Biotropica 26: 369 377. Flynn, D. F. B., M. Gogol Prokurat, T. Nogeire, N. Molinari, B. T. Richers, B. B. Lin, N. Simpson, M. M. Mayfield, and F. Declerck. 2009. Loss of functional diversity under land use intensification across multiple taxa. Ecology Letters 12:22 33. Fox, G. A. 2001. Failure time analysis. Pages 235 266 in S.M. Scheiner and J. Gurevitch, editors. Design and analysis of ecological experiments. Oxford University Pres s, New York, New York, USA. Gentry A. 1988. C hanges in plant community diversity and floristic composition on environmental and geographical gradients. Annals of the Missouri Botanical Garden 75:1 34. Ghorbani, R., S. Wilcockson, A. Koocheki, and C. Leife rt. 2008. Soil management for sustainable crop disease control: a review. Environmental Chemistry Letters 6: 149 162.

PAGE 139

139 Gilbert, G.S. 2005. Dimensions of plant disease in tropical forests. Pages 141 164 in Biotic Interactions in the Tropics: Their Role in th e Maintenance of Species Diversity. D.F.R.P. Burslem, M.A. Pinard, and S.E. Hartley, editors. Cambridge University Press, Cambridge, U.K. Gilbert, G. S., D. R. Reynolds, and A. Bethancourt. 2007. The patchiness of epifoliar fungi in tropical forests: Host range, host abundance, and environment. Ecology 88:575 581. Gilbert, G. S., and D. R. Strong. 2007. Fungal symbionts of tropical trees. Ecology 88:539 540. Biodiversit y and Conservation 5: 135 168. Giller, K. E., M. H. Beare, P. Lavelle, A. M. N. Izac, and M. J. Swift. 1997. Agricultural intensification, soi l biodiversity and agroecosystem function. Applied Soil Ecology 6: 3 16. Givnish, T. 1999. On the causes of gradie nts in tr opical tree diversity. Journal o f Ecology 87:193 210. Gosling, P., A. Hodge, G. Goodlass, and G. Bending. 2006. Arbuscular mycorrhizal fungi and organic farming. Agriculture Ecos ystems & Environment 113:17 35. Graham, J.H., and D.M. Eissenstat. 19 98. Field evidence for the carbon cost of citrus mycorrhizas. New Phytologist 140: 103 110. Graham, J.H., and L.K. Abbott. 2000. Wheat responses to aggressive and non aggressive arbuscular mycorrhizal fungi. Plant and Soil 220: 207 218. Grime, J., J. Macke y, S. Hillier, and D. Read 1987. Floristic Diversity in a Model System Using Experimental Microcosms. Nature 328:420 422. Guzmn Martnez O., A. Jaramillo Robledo, and J.V. Baldin Rincn. 2006. Anuario Metereolgico Cafetero 2006. Federacin nacional de cafeteros de Colombia. Centro Nacional de investigaciones de caf Cenicaf. Chinchin, Colombia. Harrison, K. A., and R. D. Bardgett. 2010. Influence of plant species and soil conditions on plant soil feedback in mixed grassland communities. Journal o f Eco logy 98:384 395. Hartnett, D., and G. Wilson. 2002. The role of mycorrhizas in plant community structure and dynamics: lessons from grasslands. Plant a nd Soil 244:319 331. Helgason T., J.W. Merryweather, J.P.W. Young, and A.H. Fitter. 2007. Specificity an d resilience in the arbuscular mycorrhizal fungi of a natural woodland community. Journal of Ecology 95: 623 630.

PAGE 140

140 Holl, K., and M. Lulow. 1997. Effects of species, habitat, and distance from edge on post dispersal seed predation in a tropical rainforest. B iotropica 29:459 468. Holl, K., M. Loik, E. Lin, and I. Samuels. 2000. Tropical montane forest restoration in Costa Ri ca: o vercoming barriers to dispersal and establishment. Restoration Ecology 8:339 349. Hood, L., M. Swaine, and P. Mason. 2004. The influe nce of spatial patterns of damping off disease and arbuscular mycorrhizal colonization on tree seedling establishment in Ghanaian tropical forest soil. Journal o f Ecology 92:816 823. Hooper, D.U., D.E. Bignell, V.K. Brown, L. Brussaard, J.M. Dangerfield, D .H. Wall, D.A. Wardle, D.C. Coleman, K.E. Giller, P. Lavelle, W.H. van der Putten, P.C. de Ruiter, J. Rusek, W.L. Silver, J.M. Tiedje, and V. Wolters. 2000. Aboveground and belowground biodiversity and terrestrial ecosystems: patterns, mechanisms, and feed backs. BioScience 50: 1049 1061. Hooper, E., P. Legendre, and R. Condit. 2005. Barriers to forest regeneration of deforested and aba ndoned land in Panama. Journal o f Applied Ecology 42:1165 1174. Hubbell, S. P. 2001. The Unified Theory of Biodiversity and Biogeography. Princeton Univ ersity Press. Oxford, U.K. Hughes, T. P., A. H. Baird, D. R. Bellwood, M. Card, S. R. Connolly, C. Folke, R. Grosberg, O. Hoegh Guldberg, J. B. C. Jackson, J. Kleypas, J. M. Lough, P. Marshall, M. Nystrm, S. R. Palumbi, J. M. P andolfi, B. Rosen, and J. Roughgarden. 2003. Climate change, human impacts, and the resilience of coral reefs. Science 301:929 933. Huguet, V., and J. A. Rudgers. 2010. Covariation of Soil Bacterial Composit ion with Plant Rarity. Applied a nd Environme ntal Microbiology 76:7665 7667. Janos, D.P. 1980. Mycorrhizae influence tropical succession. Biotropica 14: 56 65. Jansa, J., F. A. Smith, and S. E. Smith. 2008. Are there benefits of simultaneous root colonization by different arbuscular mycorrhizal fungi? T he New P hytologist 177:779 789. Janzen, D. 1970. H erbivores and number of tree species in tropical forests. T he American Naturalist 104:501 528 Johnson, N. C., D. Tilman, and D. Wedin. 1992. Plant and soil controls on mycorrhizal fungal communities. Ecology 73: 2034 2042.

PAGE 141

141 Johnson, N., C., G. W. T. Wilson, M. A. Bowker, J. A. Wilson, and R. M. Miller. 2010. Resource limitation is a driver of local adaptation in myco rrhizal symbioses. Proceedings of the National Academy of Sciences of the United States o f Ameri ca 107:2093 2098. Johnson, N., J. Graham, and F. Smith. 1997. Functioning of mycorrhizal associations along the mutualism parasitism continuum. The New P hytologist 135:575 586. Kabir, Z., I. OHalloran, J. Fyles, and C. Hamel. 1997. Seasonal changes of arbu scular mycorrhizal fungi as affected by tillage practices and fertilization: Hyphal density and mycorrhizal root colon ization. Plant a nd Soil 192:285 293. Kardol, P., T.M. Bezemer, and W. H. van der Putten. 2006. Temporal variation in plant soil feedback c ontrols succession. Ecology Letters 9: 1080 1088. Kardol, P., N.J. Cornips, M.M.L. van Kempen, J.M.T. Bakx Schotman, and W.H. van der Putten. 2007. Microbre mediated plant soil feedback causes historical contingency effects in plant community assembly. Eco logical Monographs 77: 147 162. Kay, K., P. Reeves, R. Olmstead, and D. Schemske. 2005. Rapid speciation and the evolution of hummingbird pollination in neotropical Costus subgenus Costus (Costaceae): evidence from nrDNA ITS and ETS sequences. American Jou rnal of Botany 92:1899 1910. Kiers, E.T., and M.G.A. V an der Heijden. 2006. Mutualistic stability in the arbuscular mycorrhizal symbiosis: exploring hypotheses of evolutionary cooperation. Ecology 87: 1627 1636. Kiers, E.T., and R.F. Denison. 2008. Sanctio ns, cooperation, and the stability of plant rhizosphere mutualisms. Annual Review of Ecology, Evolution and Systematic s 39: 215 236. Kiers, E. T., T. M. Palmer, A. R. Ives, J. F. Bruno, and J. L. Bronstein. 2010. Mutualisms in a changing world: an evolutio nary perspective Ecology Letters 13:1459 1474. Kiester, A. R., R. Lande, and D. W. Schemske. 1984. Models of coevolution and speciation in plants and their pollinators The A merican Naturalist 124:220 243. Kitajima, K. 2002. Do shade tolerant tropical tre e seedlings depend longer on seed reserves ? Functional growth analysis of three bignoniaceae species Functional Ecology 16:433 444. Klironomos, J. 2002. Feedback with soil biota contributes to plant rarity and invasiveness in communities. Nature 417:67 70 Klironomos, J. 2003. Variation in plant response to native and exotic arbuscular mycorrhizal fungi. Ecology 84:2292 2301.

PAGE 142

142 Koide, R. T., and B. Mosse. 2004. A history of research on arbuscular mycorrhiza. Mycorrhiza 14:145 163. Kraft, N.J.B., L.S. Comita, J.M. Chase, N.J. Sanders, N.G. Swenson, T.O. Crist, J.C. Stegen, M. Vellend, B. Boyle, M.J. Anderson, H.V. Cornell, K.F. Davies, A.L. Freestone, B.D. Inouye, S.P. Harrison, and J.A. Myers. 2011. Disentangling the drivers of diversity along latitudinal a nd eleva tional gradients. Science 333: 1755 1758. Kuyper, T. W., and R. G. M. de Goede. 2004. Interaction between higher plants and soil dwelling organisms. Pages 286 308 in Vegetation Ecology. Edited by E. Van der Maarel. Black well Sciences ltd. Oxford, U .K. Lankau, R. A., V. Nuzzo, G. Spyreas, and A. S. Davis. 2010. Evolutionary limits ameliorate the negat ive impact of an invasive plant. Proceedings of the National Academy of Sciences of the United States of America 107:1253 1253. Leigh, E. G., P. Davidar C. W. Dick, J. Terborgh, J. P. Puyravaud, H. Steeg e, and S. J. Wright. 2004. Why do some tropical forests have so many species of trees? Biotropica 36:447 473. Levin, S. A., H. C. Muller Landau, R. Nathan, and J. Chave. 2003. T he ecology and evolution of seed dispersal : A theoretical perspective. Annual Review o f Ecology Evolution And Systematics 34:575 604. Link, A., and A. Fiore. 2006. Seed dispersal by spider monkeys and its importance in the maintenance of neotropical rain forest diversity. Journal of Tropical Ecology 22:235 246. Lodge. 2001. Diversidad mundial y regional de hongos. Pages 291 304 in Enfoques contemporneos para el estudio de la b iodiversidad. H.M Hernndez, A.N Garca, F. lvarez, and M. Ulloa, editors. UNAM, Mexico City, Mexico. Looij en, R., and J. van Andel. 1999. Ecological communities: conceptual problems and definitions. Perspectives In Plant Ecolog y Evolution a nd Systematics 2:210 222. Lovelock, C.E., K. Andersen, and J.B. Morton. 2003. Arbuscular mycorrhizal communities in tropic al forests are affected by host tree species and environment. Oecologia 135: 268 279. Lovelock, C.E., and J.J. Ewel. 2005. Links between tree species, symbiotic fungal diversity and ecosystem functioning in simplified tropical ecosystems. New Phytologist 1 67: 219 228. Mackay, J., and P. M. Kotanen. 2008. Local escape of an invasive plant, common ragweed ( Ambrosia artemisiifolia L.), from above ground and below ground enemi es in its native area. Journal o f Ecology 96:1152 1161.

PAGE 143

143 Mangan, S., A. Eom, G. Adler, J. Yavitt, and E. Herre. 2004. Diversity of arbuscular mycorrhizal fungi across a fragmented forest in Panama: insular spore communities differ from mainland communities. Oecologia 141:687 700. Mangan, S. A., E. A. Herre, and J. D. Bever. 2010a. Specificit y between Neotropical tree seedlings and their fungal mutualists leads to plant soil feedback. Ecology 91:2594 2603. Mangan, S. A., S. A. Schnitzer, E. A. Herre, K. M. L. Mack, M. C. Valencia, E. I. Sanchez, and J. D. Bever. 2010b. Negative plant soil feed back predicts tree species relative abundance in a trop ical forest. Nature 466:752 755 McCarthy Neumann S., and R.K. Kobe. 2008. Tolerance of soil pathogens co varies with shade tolerance across species of tropical tree seedlings. Ecology 89: 1883 1892. M cCarthy Neumann S., and R.K. Kobe. 2010. Conspecific plant soil feedbacks reduce survivorship and growth of tropical tree seedlings. Journal of Ecology 98: 396 407. Menendez, A., J. Scervino, and A. Godeas. 2001. Arbuscular mycorrhizal populations associat ed with natural and cultivated vegetation on a site of Buenos Aire s province, Argentina. Biology a nd F ertility o f Soils 33:373 381. Mitchell, C.E., P.B. Reich, D. Tilman, and J.V Groth. 2003. Effects of elevated CO 2 nitrogen deposition, and decreased spec ies diversity on foliar fungal plant disease. Global Change Biology 9: 438 451. Moora, M., M. Opik, R. Sen, and M. Zobel. 2004. Native arbuscular mycorrhizal fungal communities differentially influence the seedling performance of rare and common Pulsatilla species. Functional Ecology 18:554 562. Mummey, D.L., M.C. Rillig, and W. E. Holben. 2005. Neighboring plant influences on arbuscular mycorrhizal fungal community composition as assessed by T RFLP analysis. Plant and Soil 271: 83 90. Munkvold, L., R. Kjl ler, M. Vestberg, S. Rosendahl, and I. Jakobsen. 2004. High functional diversity within species of arbuscular mycorrhizal fungi. The New P hytologist 164:357 364 Neuhauser, C., and J.E. Fargione. 2004. A mutualism parasitism continuum model and its applica tion to plant mycorrhizae interactions. Ecological Modelling 177: 337 352. Nijjer, S., W. E. Rogers, and E. Siemann. 2007. Negative plant soil feedbacks may limit persistence of an invasive tree due to rapid accumulation of soil pathogens. Proceedings of t he Royal Society B Biological Sciences 274:2621 2627.

PAGE 144

144 Oehl, F., E. Sieverding, K. Ineichen, P. Mader, T. Boller, and A. Wiemken. 2003. Impact of land use intensity on the species diversity of arbuscular mycorrhizal fungi in agroecosyst ems of Central Europ e. Applied a nd Environmental Microbiology 69:2816 2824. Orrego, O., J. C. Castao and A. M. Lpez. 2004 a Efecto de fragmentacin de los bosques en la zona cafetera sobre la diversidad gentica de las poblaciones de flora y fauna silvestre. Primera etapa i nforme tcnico de Cenicafe. Orrego, O., J. Botero, J. Verhelst, A. Pfeifer, J. Lpez, V. Franco, and J. Vlez. 2004b. Plantas vasculares del municipio de Manizales, Caldas, Colombia. Boletin de Museo de Historia Natural de la Universidad de Caldas:61 106. Ortz Escobar, M. E., R. D. Zapata Hernndez, S. Sadeghian, and H. F. Franco. 2004. Aluminio intercambiable en suelos con propiedades ndicas y su relacin con la toxicidad. Cenicaf 55: 101 110. Packer, A., and K. Clay. 2000. Soil pathogens and spatial pa tterns of seedling mortality in a temperate tree. Nature 404:278 281. Packer, A., and K. Clay. 2003. Soil pathogens and Prunus serotina seedling and sapling growth near conspecific trees. Ecology 84:108 119. Parsons, J. 1972. Spread of African pasture gras ses to the American tropics. Journal of Range Management 25: 12 17. Petermann, J. S., A. J. F. Fergus, L. A. Turnbull, and B. Schmid. 2008. Janzen Connell effects are widespread and strong enough to maintain diversity in grasslands. Ecology 89:2399 2406. P icone, C. 2000. Diversity and abundance of arbuscular mycorrhizal fungus spores in tropical forest pasture. Biotropica 32: 734 750. Pizano, C., S.A. Mangan, E.A. Herre, A.E. Eom, and J. W. Dalling. 2011. Above and belowground interactions drive habitat se gregation between two cryptic species of tropical trees. Ecology 92: 47 56. Postma Blaauw, M., R.G.M de Goede, J. Bloem, J.H. Faber, and L. Brussaard. 2010. Soil biota community structure and abundance under agricultural intensification and extens ification Ecology 91: 460 473. Pringle, A. and J.D. Bever. 2008. Analogous effects of arbuscular mycorrhizal fungi in the laboratory and a North Carolina field. New Phytologist 180: 162 175. Rasmann, C., J. H. Graham, D. O. Chellemi, L. E. Datnoff, and J. Larsen. 2009. Resilient populations of root fungi occur within five tomato production systems in southeast Flor ida. Applied Soil Ecology:1 10.

PAGE 145

145 Reinhart, K. O., A. Packer, W. H. Van Der Putten, and K. Clay. 2003. Plant soil biota interactions and spatial distributi on of black cherry in its native and invasive range s. Ecology Letters 6:1046 1050. Reinhart, K.O., A. A. Royo, W.H. van der Putten, and K. Clay. 2005. Soil and pathogen activity in Prunus serotina throughout its native range. Journal of Ecology 93: 890 989 Reinhart, K., and R. Callaway. 2006. Soil biota and invasive plants. T he New phytologist 170:445 457. Reynolds, H.L., A. Packer, J.D. Bever, and K. Clay. 2003. Grassroots ecology: plant microbe soil interactions as drivers of plant community structure an d dynamics. Ecology 84: 2281 2291. Reynolds, H., K. Vogelsang, A. Hartley, J. Bever, and P. Schultz. 2006. Variable responses of old field perennials to arbuscular mycorrhizal fungi and phosphorus source. Oecologia 147:348 358. Rodriguez, R. J., J. F. Whit e Jr, A. E. Arnold, and R. S. Redman. 2009. Fungal endophytes: diversity and functional roles. T he N ew P hytologist 182:314 330. Roesch, L. F. W., R. R. Fulthorpe, A. Riva, G. Casella, A. K. M. Hadwin, A. D. Kent, S. H. Daroub, F. A. O. Camargo, W. G. Farme rie, and E. W. Triplett. 2007. Pyrosequencing enumerates and contra sts soil microbial diversity. The ISME Journal: 283 290 Sanon, A., Z. N. Andrianjaka, Y. Prin, R. Bally, J. Thioulouse, G. Comte, and R. Duponnois. 2009. Rhizosphere microbiota interfers w ith plant plant interactions Plant a nd Soil 325:351 352. Schemske, D. W., G. G. Mittelbach, H. V. Cornell, J. M. Sobel, and K. Roy. 2009. Is there a latitudinal gradient in the importance of biotic interactions ? Annual Review o f Ecology Evolut ion a nd Syst ematics 40:245 269. Schnitzer, S. A., J. N. Klironomos, J. Hillerislambers, L. L. Kinkel, P. B. Reich, K. Xiao, M. C. Rillig, B. A. Sikes, R. M. Callaway, S. A. Mangan, E. H. van Nes, and M. Scheffer. 2011. Soil microbes drive the classic plant diversity p roductivity pattern. Ecology 92:296 303. Steffan Dewenter, I., S. G. Potts, and L. Packer. 2005. Pollinator diversity and crop pollination services a re at risk. Trends in Ecology and E volution 20:651 65 3. Strickland, M. S., C. Lauber, N. Fierer, and M. A. Bradford. 2009. Testing the functional significance of microbial community composition. Ecology 90:441 451.

PAGE 146

146 Subbarao, G. V., K. Nakahara, M. P. Hurtado, H. Ono, D. E. Moreta, A. F. Salcedo, A. T. Yoshihashi, T. Ishikawa, M. Ishitani, M. Ohnishi Kameyama, M Yoshida, M. Rondon, I. M. Rao, C. E. Lascano, W. L. Berry, and O. Ito. 2009. Evidence for biological nitrification inhibition in Br achiaria pastures. Proceedings of the National Academy of Sciences of t he United States o f America 106:17302 17307. Tchabi, A., D. Coyne, F. Hountondji, L. Lawouin, A. Wiemken, and F. Oehl. 2008. Arbuscular mycorrhizal fungal communities in sub Saharan Savannas of Benin, West Africa, as affected by agricultural land use intensity and ecologic al zone. Mycorrhiza 18:181 195. Thr all, P. H., M. E. Hochberg, J. J. Burdon, and J. D. Bever. 2007. Coevolution of symbiotic mutualists and parasites in a commun ity context. Trends In Ecology and Evolution 22:120 126. Treseder, K. K. 2004. A meta analysis of mycorrhizal responses to nitrog en, phosphorus, and atmospheric CO2 in field studies. The New O hytologist 164:347 355. Tscharntke, T., C. H. Sekercioglu, T. V. Dietsch, N. S. Sodhi, P. Hoehn, and J. M. Tylianakis. 2008. Landscape constraints on functional diversity of birds and insects i n tropical agroecosystems. Ecology 89:944 951. Van Andel, J. 2005. Species interactions structuring plant communities. Pages 238 264 in E. van der Maarel, editor. Vegetation Ecology. Blackwell Science. Oxford, U.K. Van der Heijden, M. G. A., J. N. Klirono mos, M. Ursic, T. Boller, P. Moutoglis, R. Streitwolf Engel, A. Wiemken, and I. R. Sanders. 1998. Mycorrhizal fungal diversity determines plant biodiversity, ecosystem variability and productivity. Nature 396:69 72. Van der Heijden, van, M. 2003. Arbuscula r mycorrhizal fungi as a determinant of plant diversity: in search for underlying mechanisms and general principles. Pages 243 265 in Van der Heijden, M. G. A., and I. Sanders, editors. Ecological studies 157. Springer Verlag. Berlin, Germany. Van der Heij den, Van, M. G. A. 2004. Arbuscular mycorrhizal fungi as support systems for seedling establishment in grassland. Ecology Letters 7:293 303. Van der Heijden, Van, M. G. A., R. Streitwolf Engel, R. Riedl, S. Siegrist, A. Neudecker, K. Ineichen, T. Boller, A Wiemken, and I. R. Sanders. 2006. The mycorrhizal contribution to plant productivity, plant nutrition and soil structure in e xperimental grassland. The New P hytologist 172:739 752. Van der Heijden, Van, M. G. A., R. D. Bardgett, and N. M. van Straalen. 2 008. The unseen majority: soil microbes as drivers of plant diversity and productivity in terrestrial ecosyste ms. Ecology Letters 11:296 310.

PAGE 147

147 Van Grunsven, R. H. A., W. H. Van Der Putten, T. M. Bezemer, W. L. M. Tamis, F. Berendse, and E. M. Veenendaal. 20 07. Reduced plant soil feedback of plant species expanding their range a s compared to natives. Journal o f Ecology 95:1050 1057. Van Kleunen, M., E. Weber, and M. Fischer. 2010. A meta analysis of trait differences between invasive and non invasive plant sp ecies. Ecology Letters 13:235 245. Van der Putten, W.H., and B.A.M. Peters. 1997. How soil borne pathogens may affect plant comp etition. Ecology 78: 1785 1795. Verbruggen, E., W. F. M. Rling, H. A. Gamper, G. A. Kowalchuk, H A. Verhoef, and M. G. A. Van d er Heijden. 2010. Positive effects of organic farming on below ground mutualists: large scale comparison of mycorrhizal fungal communities in agricultural soils. The New P hytologist 186:968 979. Verbruggen, E., and E. T. Kiers. 2010. Evolutionary ecology of mycorrhizal functional diversity in agricultural systems. Evolu tionary Applications 3:547 560. Vogelsang, K. M., H. L. Reynolds, and J. D. Bever. 2006. Mycorrhizal fungal identity and richness determine the diversity and productivity of a tal lgrass prai rie system. The New P hytologist 172:554 562. Vogelsang, K. M., and J. D. Bever. 2009. Mycorrhizal densities decline in association with nonnative plants and contribute to plant invasion. Ecology 90:399 407. Wagg, C., J. Jans a, B. Schmid, and M. G. A. Van d er Heijden. 2011. Belowground biodiversity effects of plant symbionts support aboveground productivity Ecology Letters 14:1001 1009. Webb, C. O., G. S. Gilbert, and M. J. Donoghue. 2006. Phylodiversity dependent seedling mortality, size structure, and dis ease in a bornean rain forest. Ecology 87:S123 S131. Westover, K. M., A. C. Kennedy, and S. Kelley. 1997. Patterns of rhizosphere microbial community structure associated with co occurring plant species. Journal of Ecology 85: 863 873. Wijdeven, S., and M. Kuzee. 2000. Seed availability as a limiting factor in forest recovery processes in Costa Rica. Restoration Ecology 8:414 424. Wolfe, B., and J. Klironomos. 2005. Breaking new ground: Soil communities and exotic plant invasion. Bioscience 55:477 487. Wrig ht, S. 2002. Plant diversity in tropical forests: a review of mechanisms of species c oexistence. Oecologia 130:1 14.

PAGE 148

148 Zangaro, W., V. L. R. Bononi, and S. B. Trufen. 2000. Mycorrhizal dependency, inoculum potential and habitat preference of native woody spe cies in South Brazil. Journal of Tropical Ecology 16: 603 622. Zangaro, W., S.M.A. Nikizaki, J.C.B. Domingos, and E.M. Nakano. 2003. Mycorrhizal response and successional status in 80 woody species from south Brazil. Journal of Tropical Ecology 19: 315 324 Zhu, Y., H. Chen, J. Fan, Y. Wang, Y. Li, J. Chen, J.X. Fan, S. Yang, L. Hu, H. Leun, T.W. Mew, P.S. Teng, Z. Wang, and C.C. Mundt. 2000. Genetic diversity and disease control in rice. Nature 406: 718 722.

PAGE 149

149 BIOGRAPHICAL SKETCH Camila Pizano was born in Bogot, Colombia, South America. She grew up between the massive city of Bogot and num erous farms around her country where she was always outside enjoying nature She had the great privilege to attend one of the best schools in Bogot, Colegio Los No gales, wh ere she was exposed to a solid discipline, a robust moral system, and really good science and math foundation s for her later career. In addition, she started developing her art skills guided by her mom, Maria Lucia Gmez, who is a professional ar t ist. A year before graduating from high school, Camila spent a summer at Cornell University, where she took a b otany class and learned how t o do scientific i llustration; the two professions of her actual life She then had the privilege of attending one o f the best Colombian universities, Universidad de los Andes, where she studied b iology. While studying at los Andes she spent three months as a volunteer at the Galpagos National Park. There she worked in the tortoise project at the Charles Darwin Resear ch Station a nd travelled through the Southern Galpagos i slands. In addition, she did an internship at the Smithsonian Tropical Research Institute (STRI) in Panama, where she spent 11 months illustrating plants for Panama Watershed Tree Atlas (ctfs.s i.edu/webatlas/maintreeatlas.html) She then received a STRI Short t erm f ellowship to work on her undergrad thesis project examining the role of soil microbes in driving incipient speciation between two cryptic species of a tropical pioneer tree This proj ect extended for three years during which Camila lived on Barro Colorado Island (STRI) and worked with Scott Mangan, James Dalling, and Allen Herre (Pizano et al. 2011)

PAGE 150

150 Camila was awarded with a Lewis Anthony Dexter Fellowship and a Grinter fellowship in 2004 at University of Florida when she started her PhD. While doing her PhD, she enjoyed a great privilege of travelling to wonderful places such as various National Natural parks in Florida, the Amazon region in Brazil, the Yucatn peninsula in Mexico, To olik research station in Alaska, and Austin Texas. Parallel to her scientific career, she continued with her art career as a scientific illustrator. She also greatly enjoys travelli ng with her dad, Pablo Pizano who is a pilot and her mom, to their beauti ful farms in Colombia and some of the most stunning natural areas in her country : the Choc, the Tayrona National Park, the pramos of the Andes, t he Vichada river in the Llanos, the Magdalena valley, the Guajira region, and the Casanare region in the Llanos. Camila owes to her personality that she studied Biology and has dedicated her life to scientific research and scientific illustration