The Influence of Fragment Size on Biotic Interactions That Structure Plant Communities in the Asian Tropics

Research Collection

Doctoral Thesis

The Influence of Fragment Size on Biotic Interactions that structure Plant Communities in the Asian Tropics

Author(s): Viswanathan, Ashwin

Publication Date: 2018

Permanent Link: https://doi.org/10.3929/ethz-b-000267356

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library DISS. ETH NO. 24963

THE INFLUENCE OF FRAGMENT SIZE ON BIOTIC INTERACTIONS THAT STRUCTURE PLANT COMMUNITIES IN THE ASIAN TROPICS

A thesis submitted to attain the degree of DOCTOR OF SCIENCES of ETH ZURICH (Dr. Sc. ETH Zurich)

Presented by Ashwin Viswanathan MSc Wildlife Biology and Conservation, National Centre for Biological Sciences (NCBS)

Born on 19.06.1987 Citizen of India

Accepted on the recommendation of Jaboury Ghazoul (examiner) Robert Bagchi (co-examiner) David Burslem (co-examiner)

2018

Summary

Diverse rainforests in the tropics are being rapidly modified and fragmented for human use.

Plants that persist in remnant forest patches are threatened by many biotic and abiotic changes that are associated with fragmentation. Several of these fragments consequently contain fewer plant species than they once did but, in the absence of contiguous forest, remain the only strongholds of plant and animal diversity in many parts of the world. However, the future of diversity in forest fragments may be uncertain as fragmentation can influence plant-animal interactions that shape plant communities. Some of these changes may already be apparent in the youngest life stages of fragmented plant communities, which would then form the template for the future of those fragments. By examining the compositions of differently aged plant communities in forest fragments, we may be able to identify interactions that are especially influenced by fragmentation, as various plant-animal interactions play structuring roles during different stages of a plant’s life cycle. In this thesis, we primarily investigate relationships between fragment area and processes that structure plant communities in a tropical forest.

In the first data chapter (the second chapter of this thesis), we examine the compositions of four life stages of woody plants across a gradient of fragment size. We show that plant species are primarily affected by fragment size during their transitions from seeds to saplings. We discuss several possible explanations for the observed patterns but suggest that altered interactions between plants and their natural enemies (insects and fungal pathogens) may be particularly important drivers. Such plant-enemy interactions have the potential to maintain plant diversity by causing the negative density-dependent mortality of locally abundant plant species, and by allowing the persistence of locally rare species (the Janzen-Connell

Hypothesis). As the modification of diversity-maintaining interactions can have catastrophic long-term consequences for the plant diversity in small fragments, it was important to investigate whether plant-enemy interactions are indeed sensitive to fragment size.

In the second data chapter (the third chapter of this thesis), we investigate the relationship between fragment area and (soil-borne) fungus-induced mortality of six woody plant species in a shadehouse experiment. We present evidence that the pathogenic effects of fungi on one plant species increased with increasing fragment size. Although we show that plant fungus- interactions can be influenced by fragment size, further experiments are required to investigate whether such effects (even when spread across the community) can influence the diversity maintained in a forest fragment.

In the third data chapter (the fourth chapter of this thesis), we first examine the roles of insects and fungi in maintaining woody plant seedling diversity in an Indian rainforest. We then investigate whether the diversity-maintaining abilities of these natural enemies are influenced by fragment size. We present evidence that insects play important roles in maintaining gamma diversity primarily by suppressing common plant species independent of density. We show that they maintain more diversity in large fragments than in smaller fragments, as smaller fragments were dominated by insect-resistant species. We present evidence that fungi caused the density- dependent mortality of one plant species but infer the presence of more such interactions in the community. We show that fungi may be sensitive to fragment size, and that fungi maintain more beta diversity of woody plant seedlings in large fragments than in small fragments.

In conclusion, I show in my thesis that insects and fungi play important roles in maintaining woody plant seedling diversity. I further show that these essential plant-enemy interactions are predictably influenced by the size of a forest fragment, and that the plant diversity in small fragments may consequently be at risk from the breakdown of such interactions.

Zusammenfassung

Die anthropogene Nutzung der natürlichen Ressourcen führt zu steigender Fragmentierung und

Veränderung der diversifizierten Regenwälder in den Tropen. Die Fragmentierung verursacht viele biotische und abiotische Veränderungen welche die Pflanzen der verbliebenen

Waldbeständen gefährdet. Daher beherbergen viele dieser Waldbestände weniger

Pflanzenarten als früher als die Wälder noch zusammenhängend waren. Die verbliebenen

Pflanzenarten in diesen Waldfragmenten bilden nun die einzige Basis für viele Pflanzen- und

Tierarten in vielen Regionen dieser Welt. Die Zukunft der Diversität in Waldfragmenten ist jedoch ungewiss, da die Fragmentation die Interaktionen zwischen Pflanzen und Tieren beeinflusst welche die Pflanzengemeinschaft erst ermöglicht. Einige dieser Veränderungen könnten in den jüngeren Altersstadien von fragmentierten Pflanzengemeinschaften bereits sichtbar sein, und dienen somit als erste Einsicht in die Zukunft dieser Waldfragmente. Die

Interaktionen zwischen Pflanzen und Tieren bilden wichtige strukturelle Funktionen verschiedener Stadien der Lebenszyklen der Pflanzen. Daher ermöglicht die Untersuchung der

Zusammensetzung von Pflanzengemeinschaften verschiedener Altersstadien Rückschlüsse bezüglich der verändernden Interaktionen aufgrund der Fragmentierung. In dieser

Doktorarbeit, untersuchen wir in erster Linie die Zusammenhänge zwischen der Fläche des

Waldfragments und den Prozessen welche die Struktur der Pflanzengemeinschaft in einem tropischen Wald bedingen.

Im ersten Datenkapitel (dem zweiten Kapitel dieser Doktorarbeit), untersuchen wir die

Zusammensetzung von vier Altersstadien von Holzgewächsen entlang eines Fragmentgrössen-

Gradienten. Gemäss den Daten, scheinen Pflanzenarten demnach vor allem durch die

Fragmentgrösse während dem Übergang vom Samen zum Schössling beeinflusst zu sein. Wir diskutieren verschiedene mögliche Ursachen für die beobachteten Muster, behaupten jedoch, dass die veränderten Interaktionen zwischen Pflanzen und deren natürlichen Feinden (Insekten

iii und Pilzpathogenen) besonders wichtige Ursachen sind. Diese Interaktionen zwischen

Pflanzen und deren natürlichen Feinden ermöglichen die Pflanzenvielfalt aufrechtzuerhalten, da sie zu einer negativen dichteabhängigen Sterblichkeit lokal reichlich vorhandener

Pflanzenarten führen, und damit lokal seltener Arten eine Chance geben zu gedeihen (Janzen-

Connell-Hypothese). Da die Veränderung von dichte-aufrechterhaltender Interaktionen katastrophale langfristige Konsequenzen für die Diversität der Pflanzen kleiner Waldfragmente haben kann, war es wichtig zu untersuchen ob die Interaktionen der Pflanzen und deren Feinden tatsächlich empfindlich gegenüber Fragmentgrösse sind.

Im zweiten Datenkapitel (dem dritten Kapitel dieser Doktorarbeit), untersuchen wir die

Beziehungen zwischen Böden von Fragmenten unterschiedlicher Grösse und den (Boden verursachten) Pilz-bedingten Sterblichkeit von sechs Arten von Holzgewächsen in einem

Schattenhausexperiment. Wir fanden Hinweise für den Anstieg von pathogenen Effekten von

Pilzen auf eine der untersuchten Pflanzenarten mit steigender Fragmentgrösse. Obschon wir aufzeigen dass Pflanzen-Pilz Interaktionen von der Fragmentgrösse beeinflusst werden kann, sind weitere Experimente notwendig um herauszufinden ob solche Effekte (selbst wenn sie sich über die Pflanzengesellschaft verteilen) die aufrechterhaltene Diversität in einem

Waldfragment beeinflussen kann.

Im dritten Datenkapitel (dem vierten Kapitel dieser Doktorarbeit), untersuchen wir zuerst die

Rolle von Insekten und Pilzen in der Aufrechterhaltung der Diversität von hölzernen

Sämlingen in einem Indischen Regenwald. Weiter untersuchen wir ob die

Diversitätserhaltenden Fähigkeiten von natürlichen Feinden von der Fragmentgrösse beeinflusst wird. Wir präsentieren Hinweise, dass Insekten eine wichtige Rolle spielen bei der

Aufrechterhaltung der Gamma-Diversität, hauptsächlich aufgrund der Unterdrückung von

üblichen Pflanzenarten, unabhängig der Dichte. Diese Aufrechterhaltung der Diversität ist ausgeprägter in grossen Fragmenten, da kleine Fragmente von Insektenresistenten Arten

iv dominiert wurden. Wir fanden Hinweise, dass Pilze die dichteabhängige Sterblichkeit einer

Pflanzenart verursachten, folgern jedoch, dass noch mehrere solche Interaktionen in der

Gemeinschaft geben muss. Pilze scheinen sehr empfindlich gegenüber der Fragmentgrösse zu sein und unterstützen die Beta-Diversität von hölzernen Sämlingen mehr in grossen als in kleinen Fragmenten.

In meiner Doktorarbeit zeige ich also auf, dass Insekten und Pilze eine wichtige Rolle spielen bei der Erhaltung der Diversität hölzerner Sämlingen. Zudem zeige ich auf, dass diese wesentlichen Interaktionen zwischen Pflanzen und deren Feinden voraussagbar beeinflusst werden von der Grösse der Waldfragmente, und das die Pflanzen Diversität in kleinen

Fragmenten folglich gefährdet sein könnten aufgrund des Zusammenbruchs eben dieser

Interaktionen.

Acknowledgements

I have several people in India and in Switzerland to thank for a thoroughly enjoyable PhD

(funded by ETH Grant 42 13-1).

My advisors, Robert Bagchi and Jaboury Ghazoul, have been remarkably supportive throughout. Robi’s command of logic, literature and statistics has been an important resource for me to tap into. More importantly, his patience, understanding and trust in the face of unprecedented events (like the field site changing a year into fieldwork) and at times my own vagaries, has taught me an invaluable lesson in how mentorship should be. These lessons continued with Jaboury, who created a pressure free environment, provided support and most importantly helped me construct a (hopefully) more lucid thesis with his sharp insights. Even for someone from India, the complex world of Swiss red tape is daunting. I am much obliged to Ankara, for expertly guiding me through this quagmire, and for so much more. Thanks also to Gilbert, who made sure I had the best tech support so that I could just get on with the job, to

Maria who constantly put up with muddy shoes, and to all the Tamil staff at the institute and in the Mensa. Owen Lewis stepped in to provide more clarity in the writing. Uma Shaankar stimulated interesting discussions in field. Vojtech Novotny shared a lot of time and knowledge of entomology and as a bonus, regaled me with his stories of Papua New Guinea. Ajith Kumar,

Aparajita Datta and Vishwesha Guttal were instrumental in my academic transition into a doctoral student and I am grateful for their guidance. I also thank David Burslem for evaluating my thesis and Tom Crowther for chairing my defense at short notice.

Fieldwork in the Western Ghats was fantastic, complete with challenges and new experiences, which several people shared with me while helping me collect data in a region with torrential rain and elephants. I am deeply indebted to my field assistant Praveen who smiled, joked and kept at a difficult job, and to his family who gave me home-cooked food and company.

vii

Kanjimalai and his family similarly made sure I was well cared for in the isolated field station.

Fieldwork was efficient and enjoyable thanks to Ganesh, who was the most meticulous research assistant I ever encountered and Arun, whose love and knowledge of plants taught me much about the Western Ghats. I am also thankful to the many inhabitants of the Kadumane Tea

Estate who shared stories, jokes and gave me company. The study would not have been possible without Venky Muthiah’s generosity (and love for wildlife) in allowing us to work in his property, and without logistical support from Cariappa, Bipin, and the other management staff at Kadumane. I learnt a lot about the Western Ghats during my first year of fieldwork in Coorg.

This entire period ended up having little academic significance (at least superficially), but was made enjoyable by Santhosh, Nithin, the Chekkara family, and several other people I met locally.

Ajith Ashokan was the first person to teach me plant identification, and Ansil taught me so much about frogs and smaller fauna. They were great company and livened up my field days considerably. Vikrant documented much of the smaller, ignored fauna of these beautiful forests and was great company along with the other long-stayers at my field site, Kavya, Netra, Sachin,

Santosh, Suresh, Subho and Vijay. I am especially grateful to Netra, Suresh and Vijay who provided selfless help despite technically not working with me, who became good friends and were key ingredients to a successful field season. Thanks also to Akshay, Bhanu, Bhuvanesh,

Dayani, Nisarg and Suri (Vinod) who stepped up to help collect data, when more hands were needed. I thank everyone else who visited me in field and made my stint more fun. I hope they have great memories of Kadumane as well. When the two-legged creatures were scarce, canine friends, particularly Steedy, Pindi, Fennec and Brown dog kept me company.

I did spend considerable time in Switzerland, especially in the last year when I saw my first summer in Europe. I made many friends during this period. Florian was my first birding companion and introduced me to Swiss landscapes and its natural history. He also made my

viii life significantly easier by often taking on the burden of my immigration requirements and dealing with my paperwork when I was in field, for which I am very grateful. Thanks also to

Josep, who took me through his beautiful country in Catalonia, and taught me so much about

European birds. John, Natalia, Scott and Satheesh were lovely companions (birding and otherwise) and eased the tedium of writing greatly with their love for birds and nature. I am sure they will ensure that an enthusiasm for natural history persists within the group. All the birders, and Danny, Emma, Eric, Fanny, Fey, Fidel, Gaby, Nanaiah, Nui, and Riska provided great company and conversations in my office, making Switzerland more fun for me. Fidel,

Nui, Riska and Eric often went out of their way to be nice and I am grateful for that. I hope, one day to host and show all of you around India.

Thanks also to Ainhoa, Babak, Charlotte, Chris K., Chris P., Claire, James (Margrove), James

(Smith), Lisa, Maike, Nicole, Sven, Swati and Tom at ETH. Joeri, Ping-Ping and Yibekal, thank you for your wonderful company in the Mensa and otherwise. A big thank you to my friend Pachai (Bharat) who serendipitously ended up in Zurich during my first year here and has been like family away from home. Thanks to Surya, Bala (for delivering excellent food when I needed it), Roger (for introducing me to Swiss professional football), Michael (for a lovely time in Hamburg), Vivek (for PhDing at the same time but being willing to randomly travel around Europe), Nihar, Mani (and family) and to all my friends in Europe.

My family of friends in India have cheered me on this journey, by coming to field, discussing my ideas, hosting me in their homes and providing me with nutrition and laughs, all while patiently waiting for me to submit and give them an excuse to make merry. Thanks to Anup,

Dayani, Nishant, Ramit, Rudy, Sapna, Shivam and Suri who were relative constants in my life in India and sources of continuous support, and to Abinand, Amod, Jahnavi, Mayank, Ranjini,

Vivek, Uddi and Abhijit. I have always enjoyed watching birds, but I thank Suhel, Praveen,

ix eBird and Bird Count India for showing me the joy/possibilities of documentation and citizen science.

Most of all, however, I am grateful to my incredible family for their unwavering understanding, support and encouragement. My parents and Aditi were unconditionally helpful in every way, and their visits to Europe were rejuvenating. Tube was always a source of comfort (and entertainment). Finally, my thesis would have been a substantially duller experience if not for

Bhanu. This thesis is as much hers as it is mine, as she lived through every part of it, patiently endured my idiosyncrasies, and considerably enhanced my life (and the research) during its development.

Contents

Summary i Zusammenfassung iii Acknowledgements vii 1 General introduction 1 1.1 Plant diversity and the Janzen-Connell hypothesis 1 1.2 The future of plant diversity in a changing world 2 1.3 Aims and Outline 4 1.4 References 6 2 The effects of fragment size on the community structure of four life stages of woody plants in a tropical rainforest in India 13 2.1 Abstract 14 2.2 Introduction 15 2.3 Methods 16 2.3.1 Study area 16 2.3.2 Study design 17 2.3.3 Data analyses 19 2.4 Results 21 2.5 Discussion 25 2.6 Acknowledgements 29 2.7 References 30 3 The effects of rainforest fragment area and intraspecific seed mass on the strength of plant-pathogen interactions 39 3.1 Abstract 40 3.2 Introduction 41 3.3 Methods 43 3.3.1 Study area 43 3.3.2 Study design 43 3.3.3 Data analysis 44 3.4 Results 45 3.4.1 The effects of fragment size 45 3.4.2 The effects of seed mass 45 3.5 Discussion 50 3.5.1 The effects of fragment size on plant-fungus interactions 50 3.5.2 The effects of seed mass on plant-fungus interactions 51 3.5.3 Summary 53 3.6 Acknowledgements 54 3.7 References 55 3.8 Supplementary information 61

xii

3.8.1 Survival of Hopea canarensis 61 3.8.2 Data exclusion and model validation 61 3.8.3 The direct effects of fungicides and insecticides on plant growth 62 3.8.4 References 64 4 The role of plant-enemy interactions in maintaining tree seedling diversity in a fragmented tropical forest in India 66 4.1 Abstract 67 4.2 Introduction 69 4.3 Methods 71 4.3.1 Study area 71 4.3.2 Study design 72 4.3.3 Data analysis 75 4.4 Results 76 4.4.1 Seedling mortality 76 4.4.2 Alpha diversity 77 4.4.3 Beta diversity 82 4.4.4 Gamma diversity 82 4.5 Discussion 84 4.5.1 The Janzen-Connell hypothesis 84 4.5.2 The role of insects 87 4.5.3 The role of fungi 88 4.5.4 The effects of fragment size on biotic interactions and plant diversity 89 4.5.5 Conclusions 90 4.6 Acknowledgements 91 4.7 References 92 4.8 Supplementary information 99 4.8.1 The effects of abiotic drivers 99 4.8.2 The effects of water 99 4.8.3 Synergies between the actions of insects and fungi 99 4.8.4 Identification of spatial scale to measure gamma diversity 101 4.8.5 Validating our statistical models 101 4.8.6 The direct effects of fungicides and insecticides on plant growth 102 4.8.7 Does conspecific density-dependence correlate with seedling abundance? 103 4.8.8 References 104 5 General conclusions 107 5.1 Do insects and fungi maintain woody plant diversity? 107 5.2 Are the diversity-maintaining roles of enemies affected by fragment size? 107 5.3 Can interspecific differences in community analyses be homogenized? 108 Species lists 110 Birds 110 xiii

Amphibians 110 Reptiles 112 Mammals (except bats) 113 Butterflies 115 Trees 120

xiv

1 General introduction

1.1 Plant diversity and the Janzen-Connell hypothesis

Tropical rainforests are remarkably diverse. They contain a large majority of the Earth’s plant and animal species, including two-thirds of all known angiosperms (Prance 1977). Over 250 tree species have sometimes been documented in a single hectare of rainforest (Lee et al. 2002,

Valencia et al. 2004), where they largely compete for the same limiting resources (Tilman

1982, Wright 2002). These species, however, seem to possess the ability to coexist despite inherent and inevitable differences in their competitive abilities (or fitness). Our understanding of this apparent contradiction or paradox (Hutchinson 1961, Fox et al. 2010) has considerably evolved in the last many decades through the conceptualization, and subsequent investigation, of several theories that explain species coexistence (Wright 2002). The mechanisms that facilitate species coexistence and maintain plant diversity are now understood to either ensure that fitness differences between species are eliminated (equalizing mechanisms), or that fine- scale abundances of species are independently inhibited (stabilizing mechanisms) (Chesson

2000, Adler et al. 2007). The exact roles and relative impacts of these mechanisms, however, are still poorly understood, and the resolution of this question remains a central theme in ecological research (Sutherland et al. 2013).

One stabilizing mechanism that has received strong empirical support is described by the

Janzen-Connell hypothesis (Janzen 1970, Connell 1971). This theory proposes that plant species can coexist when their abundances are regulated by specialized natural enemies that respond to local densities of their preferred plant species (mostly at the seed and seedling stages). These enemies preclude competitive exclusion by suppressing populations of young plants in a density-dependent manner, thereby ensuring that species can persist if they are locally rare but are inhibited when they are locally abundant. Several natural enemies of plants

1 that cause negative density-dependent mortality have now been identified, including fungal pathogens, insect herbivores and soil biota (Sullivan 2003, Hood et al. 2004, Bagchi et al. 2010,

Mangan et al. 2010, Swamy and Terborgh 2010, Bagchi et al. 2014). Density-dependent interactions between plants and their natural enemies have been empirically implicated in essential roles that structure plant communities, and potentially maintain tropical tree diversity

(Webb and Peart 1999, Harms et al. 2000, Metz et al. 2010, Bagchi et al. 2014, Comita et al.

2014). However, in the light of widespread anthropogenic change in tropical forests around the world (Kareiva et al. 2007, Haddad et al. 2015, Keenan et al. 2015), we know surprisingly little about how these interactions are being modified in a changing world (Benitez-Malvido et al.

1999, Benítez-Malvido and Lemus-Albor 2005, Magrach et al. 2014), and what this may mean for the future of biodiversity in the tropics.

1.2 The future of plant diversity in a changing world

The future of biodiversity in an increasingly human-dominated world is uncertain (Debinski and Holt 2000, Haddad et al. 2015). Rainforests are being modified at an unprecedented rate by logging, hunting, mining, agriculture and industrialization (Kareiva et al. 2007, Haddad et al. 2015, Keenan et al. 2015). Consequently, fragmentation is one of the greatest threats faced by tropical forests today (Debinski and Holt 2000, Ewers and Didham 2006, Laurance et al.

2011). Small fragments are likely to eventually contain fewer species per unit area than larger fragments, as is predicted by the theory of island biogeography (MacArthur and Wilson 1967).

Many studies have shown that forest fragments hold less diversity than their undisturbed counterparts do (Turner 1996, Laurance et al. 2002, Ewers and Didham 2006, Laurance et al.

2011). However, fragmented forests are still very diverse (Wright and Muller-Landau 2006,

Dent and Wright 2009) but the breakdown of diversity-maintaining mechanisms (regulated by insects and fungi) may lead to significant losses of biodiversity in the future.

Small insects and fungi may be especially vulnerable to fragmentation, as they are poor dispersers and may struggle to move between islands of habitat in fragmented landscapes

(MacArthur and Wilson 1967, Didham et al. 1996). Some of the abiotic changes that are associated with forest fragments like changes in light, humidity and other microclimatic conditions (Murcia 1995, Bruna 1999, Uriarte et al. 2010, Ewers and Banks-Leite 2013), are also known to influence fungal pathogens (Agrios 1978, Thompson et al. 2010, Swinfield et al. 2012). Altered microclimates in fragments (Ewers and Didham 2006, Laurance et al. 2011), especially near habitat edges, may therefore result in the loss of certain pathogens from the system. Unsurprisingly, a few studies have demonstrated that fragmentation can affect interactions between plants and fungal pathogens (Benitez-Malvido et al. 1999, Ewers and

Didham 2006, Fáveri et al. 2008), and between plants and insect herbivores (Arnold and

Asquith 2002, Ruiz-Guerra et al. 2010, Laurance et al. 2011, Schuldt et al. 2014).

The effects of fragmentation on potential diversity-maintaining plant-enemy interactions, however, appear to be inconsistent across various plant species. Some studies have documented lower pathogen loads in fragments while others have found higher loads (Benitez-Malvido et al. 1999, Benítez-Malvido and Lemus-Albor 2005). Fáveri et al. (2008) found that herbivory is reduced for certain plant species in fragments while Benitez-Malvido et al. (1999) found no significant directional effects. There is consequently very little understanding currently on how entire plant communities might respond when such interactions, including those that form the basis of the Janzen-Connell hypothesis, are modified. Such modifications of a plant-natural enemy network may result in cascading changes in plant community composition and richness in a fragment. We do see from a few studies that common plant species tend to increasingly dominate when forests are fragmented (Hill and Curran 2001, Benítez-Malvido and Martínez-

Ramos 2003), and this may be due to changes in species interactions. Although herbivores, pathogens and plant-enemy interactions appear to be susceptible to fragmentation, the

3 community wide consequences of their responses to fragmentation had not been investigated previously.

1.3 Aims and Outline

I sought to address this gap in knowledge with the help of the following broad objectives: 1)

Quantify the roles of insects and fungi in maintaining diversity in a rainforest landscape in

India. 2) Investigate the effects of fragment size on community-wide interactions between plants and their natural enemies, and consequently on their abilities to maintain plant diversity.

I attempted to fulfil these objectives in a fragmented rainforest landscape in South India. Our research findings are collated in my thesis that consists of five-chapters, which are outlined as follows.

Chapter 1: General Introduction

Chapter 2: The effects of fragment size on the community structure of four life stages of woody plants in a tropical rainforest in India – We explore patterns of woody plant distributions across a gradient of fragment size. We attempt to identify the period in the life cycle of a woody plant when it is most susceptible to fragmentation. We finally draw upon insights from these analyses and discuss how the observed patterns may be a consequence of certain modified processes in small fragments.

Chapter 3: The effects of rainforest fragment area and intraspecific seed mass on the strength of plant-pathogen interactions – We begin our investigation of the effects of fragment size on processes that may structure tropical rainforests. We examine the effects of fragment size and intraspecific seed mass on plant-pathogen (soil-borne) interactions in a shadehouse experiment. We quantify the roles of pathogenic fungi in regulating relationships

4 between seedling establishment and fragment size, and between seedling establishment and seed mass.

Chapter 4: The role of plant-enemy interactions in maintaining tree seedling diversity in a fragmented tropical forest in India – Finally, we comprehensively examine the effects of fragment size on the mortality of naturally established woody plant seedlings. We investigate the roles of insects and fungi in causing the mortality, and their susceptibility to fragment size, by experimentally excluding natural enemies from certain plots. We quantify the roles of insects and fungi in maintaining seedling diversity and investigate whether their diversity- maintaining abilities are influenced by fragment size.

Chapter 5: General Conclusions

1.4 References

Adler, P. B., J. HilleRisLambers, and J. M. Levine. 2007. A niche for neutrality. Ecology

Letters 10:95-104.

Agrios, G. N. 1978. Plant Pathology. Academic Press.

Arnold, A. E., and N. M. Asquith. 2002. Herbivory in a fragmented tropical forest: patterns

from islands at Lago Gatún, Panama. Biodiversity & Conservation 11:1663-1680.

Bagchi, R., R. E. Gallery, S. Gripenberg, S. J. Gurr, L. Narayan, C. E. Addis, R. P. Freckleton,

and O. T. Lewis. 2014. Pathogens and insect herbivores drive rainforest plant diversity

and composition. Nature 506:85-88.

Bagchi, R., T. Swinfield, R. E. Gallery, O. T. Lewis, S. Gripenberg, L. Narayan, and R. P.

Freckleton. 2010. Testing the Janzen-Connell mechanism: pathogens cause

overcompensating density dependence in a tropical tree. Ecology Letters 13:1262-

1269.

Benitez-Malvido, J., G. Garcı́a-Guzmán, and I. D. Kossmann-Ferraz. 1999. Leaf-fungal

incidence and herbivory on tree seedlings in tropical rainforest fragments: an

experimental study. Biological Conservation 91:143-150.

Benítez-Malvido, J., and A. Lemus-Albor. 2005. The Seedling Community of Tropical Rain

Forest Edges and Its Interaction with Herbivores and Pathogens. Biotropica 37:301-

313.

Benítez-Malvido, J., and M. Martínez-Ramos. 2003. Impact of Forest Fragmentation on

Understory Plant Species Richness in Amazonia. Conservation Biology 17:389-400.

Bruna, E. M. 1999. Seed germination in rainforest fragments. Nature 402:139.

Chesson, P. 2000. Mechanisms of Maintenance of Species Diversity. Annual Review of

Ecology and Systematics 31:343-366.

Comita, L. S., S. A. Queenborough, S. J. Murphy, J. L. Eck, K. Xu, M. Krishnadas, N.

Beckman, and Y. Zhu. 2014. Testing predictions of the Janzen–Connell hypothesis: a

meta-analysis of experimental evidence for distance- and density-dependent seed and

seedling survival. Journal of Ecology 102:845-856.

Connell, J. H. 1971. On the role of natural enemies in preventing competitive exclusion in some

marine animals and in rain forest trees. Pages 298-312 in P. J. den Boer and G.

Gradwell, editors. Dynamics of populations. PUDOC, Wageningen.

Debinski, D. M., and R. D. Holt. 2000. A Survey and Overview of Habitat Fragmentation

Experiments. Conservation Biology 14:342-355.

Dent, D. H., and S. J. Wright. 2009. The future of tropical species in secondary forests: a

quantitative review. Biological Conservation 142:2833-2843.

Didham, R. K., J. Ghazoul, N. E. Stork, and A. J. Davis. 1996. Insects in fragmented forests: a

functional approach. Trends in Ecology & Evolution 11:255-260.

Ewers, R. M., and C. Banks-Leite. 2013. Fragmentation Impairs the Microclimate Buffering

Effect of Tropical Forests. PLoS ONE 8:e58093.

Ewers, R. M., and R. K. Didham. 2006. Confounding factors in the detection of species

responses to habitat fragmentation. Biological Reviews 81:117-142.

Fáveri, S. B., H. L. Vasconcelos, and R. Dirzo. 2008. Effects of Amazonian forest

fragmentation on the interaction between plants, insect herbivores, and their natural

enemies. Journal of Tropical Ecology 24:57-64.

Fox, J. W., W. A. Nelson, and E. McCauley. 2010. Coexistence mechanisms and the paradox

of the plankton: quantifying selection from noisy data. Ecology 91:1774-1786.

Haddad, N. M., L. A. Brudvig, J. Clobert, K. F. Davies, A. Gonzalez, R. D. Holt, T. E. Lovejoy,

J. O. Sexton, M. P. Austin, C. D. Collins, W. M. Cook, E. I. Damschen, R. M. Ewers,

B. L. Foster, C. N. Jenkins, A. J. King, W. F. Laurance, D. J. Levey, C. R. Margules,

B. A. Melbourne, A. O. Nicholls, J. L. Orrock, D.-X. Song, and J. R. Townshend. 2015.

Habitat fragmentation and its lasting impact on Earth’s ecosystems. Science Advances

Harms, K. E., S. J. Wright, O. Calderon, A. Hernandez, and E. A. Herre. 2000. Pervasive

density-dependent recruitment enhances seedling diversity in a tropical forest. Nature

404:493-495.

Hill, J. L., and P. J. Curran. 2001. Species composition in fragmented forests: conservation

implications of changing forest area. Applied Geography 21:157-174.

Hood, L., M. Swaine, and P. Mason. 2004. The influence of spatial patterns of damping‐off

disease and arbuscular mycorrhizal colonization on tree seedling establishment in

Ghanaian tropical forest soil. Journal of Ecology 92:816-823.

Hutchinson, G. E. 1961. The paradox of the plankton. The American Naturalist 95:137-145.

Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests. American

Naturalist 104:501-528.

Kareiva, P., S. Watts, R. McDonald, and T. Boucher. 2007. Domesticated Nature: Shaping

Landscapes and Ecosystems for Human Welfare. Science 316:1866-1869.

Keenan, R. J., G. A. Reams, F. Achard, J. V. de Freitas, A. Grainger, and E. Lindquist. 2015.

Dynamics of global forest area: Results from the FAO Global Forest Resources

Assessment 2015. Forest Ecology and Management 352:9-20.

Laurance, W. F., J. L. Camargo, R. C. Luizão, S. G. Laurance, S. L. Pimm, E. M. Bruna, P. C.

Stouffer, G. B. Williamson, J. Benítez-Malvido, and H. L. Vasconcelos. 2011. The fate

of Amazonian forest fragments: a 32-year investigation. Biological Conservation

144:56-67.

Laurance, W. F., T. E. Lovejoy, H. L. Vasconcelos, E. M. Bruna, R. K. Didham, P. C. Stouffer,

C. Gascon, R. O. Bierregaard, S. G. Laurance, and E. Sampaio. 2002. Ecosystem decay

of Amazonian forest fragments: a 22‐year investigation. Conservation Biology 16:605-

618.

Lee, H., S. Davies, J. LaFrankie, S. Tan, T. Yamakura, A. Itoh, T. Ohkubo, and P. Ashton.

2002. Floristic and structural diversity of mixed dipterocarp forest in Lambir Hills

National Park, Sarawak, Malaysia. Journal of Tropical Forest Science 14:379-400.

MacArthur, R. H., and E. Wilson. 1967. The theory of island biogeography. Monographs in

Population Biology 1.

Magrach, A., W. F. Laurance, A. R. Larrinaga, and L. Santamaria. 2014. Meta-Analysis of the

Effects of Forest Fragmentation on Interspecific Interactions. Conservation Biology

28:1342-1348.

Mangan, S. A., S. A. Schnitzer, E. A. Herre, K. M. L. Mack, M. C. Valencia, E. I. Sanchez,

and J. D. Bever. 2010. Negative plant-soil feedback predicts tree-species relative

abundance in a tropical forest. Nature 466:752-755.

Metz, M. R., W. P. Sousa, and R. Valencia. 2010. Widespread density-dependent seedling

mortality promotes species coexistence in a highly diverse Amazonian rain forest.

Ecology 91:3675-3685.

Murcia, C. 1995. Edge effects in fragmented forests: implications for conservation. Trends in

Ecology & Evolution 10:58-62.

Prance, G. T. 1977. Floristic inventory of the tropics: where do we stand? Annals of the

Missouri Botanical Garden:659-684.

Ruiz-Guerra, B., R. Guevara, N. A. Mariano, and R. Dirzo. 2010. Insect herbivory declines

with forest fragmentation and covaries with plant regeneration mode: evidence from a

Mexican tropical rain forest. Oikos 119:317-325.

Schuldt, A., T. Assmann, H. Bruelheide, W. Durka, D. Eichenberg, W. Härdtle, W. Kröber, S.

G. Michalski, and O. Purschke. 2014. Functional and phylogenetic diversity of woody

plants drive herbivory in a highly diverse forest. New Phytologist 202:864-873.

Sullivan, J. J. 2003. Density-dependent shoot-borer herbivory increases the age of first

reproduction and mortality of neotropical tree saplings. Oecologia 136:96-106.

Sutherland, W. J., R. P. Freckleton, H. C. J. Godfray, S. R. Beissinger, T. Benton, D. D.

Cameron, Y. Carmel, D. A. Coomes, T. Coulson, and M. C. Emmerson. 2013.

Identification of 100 fundamental ecological questions. Journal of Ecology 101:58-67.

Swamy, V., and J. W. Terborgh. 2010. Distance‐responsive natural enemies strongly influence

seedling establishment patterns of multiple species in an Amazonian rain forest. Journal

of Ecology 98:1096-1107.

Swinfield, T., O. T. Lewis, R. Bagchi, and R. P. Freckleton. 2012. Consequences of changing

rainfall for fungal pathogen-induced mortality in tropical tree seedlings. Ecology and

Evolution 2:1408-1413.

Thompson, S., P. Alvarez‐Loayza, J. Terborgh, and G. Katul. 2010. The effects of plant

pathogens on tree recruitment in the Western Amazon under a projected future climate:

a dynamical systems analysis. Journal of Ecology 98:1434-1446.

Tilman, D. 1982. Resource Competition and Community Structure. (Mpb-17). Princeton

University Press.

Turner, I. M. 1996. Species Loss in Fragments of Tropical Rain Forest: A Review of the

Evidence. Journal of Applied Ecology 33:200-209.

Uriarte, M., E. M. Bruna, P. Rubim, M. Anciães, and I. Jonckheere. 2010. Effects of forest

fragmentation on the seedling recruitment of a tropical herb: assessing seed vs. safe-

site limitation. Ecology 91:1317-1328.

Valencia, R., R. B. Foster, G. Villa, R. Condit, J. C. Svenning, C. Hernández, K. Romoleroux,

E. Losos, E. Magård, and H. Balslev. 2004. Tree species distributions and local habitat

variation in the Amazon: large forest plot in eastern Ecuador. Journal of Ecology

92:214-229.

Webb, C. O., and D. R. Peart. 1999. Seedling density dependence promotes coexistence of

Bornean rain forest trees. Ecology 80:2006-2017.

Wright, J. S. 2002. Plant diversity in tropical forests: a review of mechanisms of species

coexistence. Oecologia 130:1-14.

Wright, S. J., and H. C. Muller-Landau. 2006. The uncertain future of tropical forest species.

Biotropica 38:287-301.

2 The effects of fragment size on the community structure of four life stages of woody plants in a tropical rainforest in India

Ashwin Viswanathan*a, Jaboury Ghazoula, N. Arun Kumarb, and Robert Bagchic a Chair of Ecosystem Management, ITES, ETH Zurich, Zurich 8092, Switzerland b Forest Research Institute, Dehradun, Uttarakhand 248006, India c Department of Ecology and Evolutionary Biology, University of Connecticut Storrs, CT

06269, USA

* Corresponding author: [email protected]; +919483512541

2.1 Abstract

A large proportion of tropical forests around the world are now fragmented, but we know little about how the size of a fragment may influence plant community composition and diversity.

Plant communities may be especially vulnerable to fragmentation during particular life stages and identifying these life stages can provide valuable insights into specific processes that are modified in fragmented forests. To identify these life stages, we sampled adult woody plants, seed rain, young saplings, and old saplings in eight forest fragments (1 – 149 ha) and examined correlations between fragment area and patterns of community structure in each of these life stages. We found that sapling to adult ratios, aggregated across all species, were independent of fragment size. The ratio of saplings to adults of certain species, however, increased significantly in smaller fragments. We also found that fragment size was associated with variation in the species composition of young saplings, but not that of the other life stages considered. Our results suggest that the processes that structure woody plant communities in rainforests are most vulnerable to fragmentation during the early stages of plant growth. In particular, seed germination and seedling establishment might be modified in smaller fragments as result of altered microclimate and natural enemy pressure.

Keywords: plant-enemy interactions, post-dispersal, diversity, fragmentation, seedling mortality

2.2 Introduction

The rapid transformation of natural landscapes in the tropics for human use is a major threat to rainforests and the biodiversity they sustain (Haddad et al. 2015). In human-modified landscapes, the remaining forests often persist as isolated fragments that are reservoirs for the biodiversity of the region (Wright and Muller-Landau 2006). With such fragments set to serve as strongholds of biodiversity in the future (Daily 2001), the number of species they retain will be contingent on how animal and plant communities respond over time to reduced forest area

(and associated factors) (Ewers and Didham 2006). The theory of island biogeography predicts that small and isolated forest fragments will contain fewer animal and plant species than large fragments (MacArthur and Wilson 1967). Given that animals are short-lived and, being mobile, are more sensitive to habitat size, there is some evidence that smaller fragments not only contain fewer animal species than large fragments (Turner 1996), but also lose species more quickly (Lovejoy et al. 1986, Ferraz et al. 2003, Stouffer et al. 2009). There is limited empirical consensus, however, on how long-lived communities of rainforest trees will respond over time to reduced fragment size.

The few studies that have investigated the effects of fragmentation on the species richness and composition of rainforest trees report mixed results. Hill and Curran (2003) found that larger fragments contained more tree species than smaller fragments in a rainforest in Brazil but

Arroyo-Rodríguez and Mandujano (2006) did not find any such relationship in a rainforest in

Mexico. Arroyo-Rodríguez and Mandujano (2006) and Laurance et al. (2006) both found that fragmentation influenced adult tree composition, but they identified fragment size and edge effects to be the drivers respectively. These mixed results, regarding both the direction and specific drivers of change, may be a consequence of the confounding effects of several attributes of fragmentation like fragment size, isolation and effects from surrounding landscapes (Ewers and Didham 2006).

These studies, moreover, have focussed on variation among adult plant communities, but the future composition of plants in these fragments is likely to depend on the current stock and dynamics of seeds, seedlings and saplings. Consequently, patterns of change in the communities of seedlings and saplings in these fragments may provide important insights into:

1) the eventual structure of adult communities, and 2) any changes to the processes that structure populations of young woody plants in rainforests. Some studies have shown that fragmentation, but not necessarily fragment size, influences the recruitment of seedlings, and the compositions of seedling and sapling communities (Laurance et al. 1998, Benítez-Malvido and Martínez-Ramos 2003, Hill and Curran 2003). These studies, however, do not eliminate the possibility that the modified patterns are due to changes before, rather than during, recruitment, which would imply that fragmentation affects adult trees, fruit production or seed dispersal instead of seedling survival itself.

We attempted to fill these gaps by investigating the relationships between fragment size and four life stages of woody plants in a single-land use system, where each fragment was separated by less than 100 m from at least one other fragment. We asked the following two questions: 1)

Does fragment size explain variation in the composition of woody plant assemblages? 2) Are patterns consistent across life stages, and if not, which life-stages show the most pronounced differences? We hypothesized that sapling to adult ratios will be influenced by fragment size

(implying modified patterns of recruitment), and that variation in the compositions of the youngest life stages of woody plants will consequently be best explained by fragment size.

2.3 Methods

2.3.1 Study area

We conducted the study in Kadumane Tea Estate (12.8639 – 12.9389 N and 75.6361 - 75.6833

E), a private plantation in the Western Ghats in southwestern India. The Western Ghats receive

16 up to 9000 mm of annual rainfall, and harbour wet evergreen forests interspersed with high altitude grasslands. Forest plant communities in the region are species rich, with over 4700 species of flowering plants recorded (Myers et al. 2000). These rainforest landscapes, however, are now severely fragmented by the cultivation of coffee and tea. Forests in Kadumane, which were once contiguous, have now been fragmented for approximately 50 years by roads and “no shade” tea fields set up primarily in the late 1960s and early 1970s. Remnant fragments on the estate range between 1 – 149 ha in size.

2.3.2 Study design

We selected eight fragments ranging in size from 1.1 – 149 ha (within an elevational range of

930 m to 990 m) to examine the effects of fragment size on the species composition and diversity of woody plants. In each fragment, we delineated a 20 x 20 m quadrat at a random location with at least 70% canopy cover. We confirmed that canopy cover did not vary as a function of fragment size, by measuring the proportions of filled pixels from photographs of the canopy (taken at ground level with an 18mm lens at five random points in each quadrat) and examining their relationship with fragment size (Fig. 2.1).

We documented the identity and girth (measured at breast height for individuals that were 1.5 m and taller; at the base of the main stem for smaller seedlings, saplings and shrubs) of each individual woody plant (except lianas) in each of these quadrats except of individuals that had just germinated in the year of sampling (2015-2016). We categorized the sampled woody plant community into shrubs, large shrubs, small trees, medium trees and large trees (Sasidharan

2013). We assumed that individuals with girths greater than 3 cm, 5 cm, 10 cm, 20 cm and 30 cm respectively from each of these categories were near-adults (not necessarily reproductive adults but relatively older plants) and that smaller individuals with girths greater than 1 cm

17 were “old saplings”. We categorized individuals with lesser girth than 1 cm (< 3 mm diameter) as “young saplings”.

Fig. 2.1: Canopy cover (measured at five random points in each quadrat) plotted against log-transformed fragment size (R2 = 0.01, p = 0.83). Error bars and belts represent 95% confidence intervals.

Two 1 m2 seed traps were also placed at random locations within each quadrat. We checked the seed rain in these traps every 20 days in the dry season and every 16 days in the wet season, from June 2015 to December 2016. We recorded the number and identity (to species if possible) of every seed or fruit in the seed traps. As it was often difficult to assign a taxonomic identity immediately (although each morphotype was given a unique identity), we planted all unidentified individuals in a greenhouse to aid in identification. We also collected representatives of all seeds and fruits to build a database of unique seeds and fruits in the system. Although we were unable to assign a formal species name to many morphospecies, we

18 were able to link the seeds to seedlings of most morphospecies by germinating seeds in the greenhouse.

2.3.3 Data analyses

All analyses were conducted using R 3.4.0 (R Core Team 2017). We used the function specpool in the package ‘vegan’ (Oksanen et al. 2017) to estimate the species richness of each life stage using a ‘Chao’ estimator. The Chao estimator extrapolates the richness of the species pool in the sampled habitat and elevation, and is a reliable non-parametric method to estimate the richness of taxa dominated by rare species, like the richness of trees in tropical forests (Colwell

2 and Coddington 1994). The estimator uses the formula Sp = S0 + a1 /(2a2(N − 1)⁄N) where

S풑 is the extrapolated richness, Sퟎ is the sampled richness, aퟏ is the number of species occurring in only one site, aퟐ is the number of species occurring in exactly two sites, and N is the number of sites in the sample.

To investigate whether fragments become more diverse through ontogeny, we used the

Shannon Diversity Index to estimate the diversity of stems within each life stage in each fragment. We separately modelled the species diversity and richness of each life stage as a function of fragment size using simple linear models, assuming a Gaussian distribution for the errors.

To test the hypothesis that sapling to adult ratios of plant species are modified by fragment size, we used Generalized Additive Models with three ‘knots’ to model sapling numbers as a function of log-transformed fragment size. We used the gam function in the package ‘mgcv’

(Wood 2017) for this analysis, and offset each negative binomial response by the logarithm of the number of adults. This results in a model that effectively regresses the ratio of saplings to adults, but accounts for the fact that the response (number of saplings) is discrete count data.

We analysed all species together, and subsequently individually analysed the species that met

19 the following criterion – species that were present in at least six fragments but averaged at least five individuals per fragment (Table 2.1).

Table 2.1: Information about the 15 most abundant species across all eight 20 x 20 m quadrats in eight different fragments.

N fragments N young N old saplings species acronym with adults N adults N seeds saplings (>1cm girth) (out of 8)

Psychotria nigra Psni 8 403 11 141 584

Psychotria macrocarpa Psma 8 228 21 124 1025

Litsea floribunda Lifl 8 36 88 88 194

Dichapetalum gelonioides Dige 6 51 0 36 111

Nothopegia travancorica Notr 6 40 0 32 124

Syzygium rubicundum Syru 7 22 35 76 416

Dimocarpus longan Dilo 6 37 0 69 423

Symplocos racemosa Syra 6 24 0 32 96

Memecylon malabaricum Mema 5 73 3 29 159

Ixora sp. Ixor 4 45 0 28 192

Clerodendrum infortunatum Clin 4 26 0 18 117

Reinwardtiodendron anamalaiense Rean 3 10 0 55 347

Blachia umbellata Blum 2 43 1 4 277

Agrostistachis indica Agin 2 59 0 0 86

Humboldtia brunonis Hubr 1 7 0 93 356

To test the associated hypothesis that fragment size would most prominently influence variation in the compositions of young woody plant assemblages, we used non-metric multidimensional scaling (non-parametric rank-based ordination method robust to non-linear relationships) to investigate correlations between community structure, life stage and fragment size. We used the function metaMDS in the package ‘vegan’ (Oksanen et al. 2017) (that uses the Bray-Curtis dissimilarity index) to calculate a measure of dissimilarity in assemblages of each life stage across all fragments. We used the function envfit to run permutation tests to

20 identify interactions between plant life stage and fragment size that explain variation in community composition.

2.4 Results

We identified 115 species of woody plants (excluding lianas) in the study (see Species lists -

Trees), and documented seeds of 27 of these species in the traps. We estimated that 130.29 ±

7.21 species (using a ‘Chao’ estimator) occurred in the overall study site in the sampled habitat and elevation. When broken down by the four life stages, we estimated the species richness as

(1) seeds – 69.15 ± 32.44; (2) young saplings – 88.79 ± 12.21; (3) old saplings (>1cm girth) –

114.98 ± 4.61 species; (4) adults – 131.15 ± 19.98 species. The species richness of all life stages increased with increasing (log-transformed) fragment size but these trends were only significant for adult woody plants and for seed rain (Table 2.2, Fig. 2.2). The diversity

(Shannon’s H) of only seed rain and young saplings increased with increasing fragment size

(Table 2.2, Fig. 2.2) but the trend was significant only for seed rain.

The overall ratio of numbers of saplings to adults, across all species, was unaffected by

2 fragment size (Fig. 2.3, χ 1 = 0.39, p = 0.531). However, fragment size influenced the sapling to adult ratios of two of the five species that satisfied the criterion for individual analysis (see

2 Methods). The sapling to adult ratios of Dichapetalum gelonioides (χ 1.55 = 12.81, p = 0.003)

2 and Nothopegia travancorica (χ 1.02 = 13.22, p = 0.001) decreased in larger fragments. Sapling

2 to adult ratios were statistically unaffected by fragment size for Psychotria nigra (χ 1 = 3.47, p

2 2 = 0.06), Psychotria macrocarpa (χ 1 = 0.89, p = 0.347) and Litsea floribunda (χ 1.18 = 0.79, p

= 0.37).

Fig. 2.2: Species richness and Shannon diversity of four life stages in each fragment plotted against log- transformed fragment size. Error belts represent 95% confidence intervals.

Variation in the species composition of young saplings (<1 cm girth) was significantly correlated with fragment size, but variation in the composition of the other life stages was not

(Table 2.2, Fig. 2.4, Fig. 2.5).

Fig. 2.3: Saplings per adult plotted against log-transformed fragment size across species. The error belts are 95% confidence intervals predicted from negative-binomial generalized additive models. The selected species include the three species with adults in all eight fragments, and the two most abundant of the remaining species (adults) present in at least six fragments. See Table 2.1 for further information and for species keys.

Table 2.2: Results of 1) permutation tests investigating the presence of significant interactions between fragment size (log transformed) and life stage (seeds analysed separately; k = 2; stress = 0.10) that explain plant species composition in four dimensions (k = 4; stress = 0.09) 2) separate linear models investigating the effects of fragment size (log transformed) on diversity and richness of each life stage. Only p-values are presented, as permutation tests do not have a test statistic.

composition diversity richness

life stage: R2 p (*) int. slope R2 p (*) int. slope R2 p (*)

adults 0.19 0.12 2.71 -0.06 0.08 0.50 25.17 2.48 0.55 0.04 *

seeds 0.04 0.89 0.42 0.24 0.61 0.02 * 2.63 1.22 0.57 0.03 *

young saplings 0.41 0.01 ** 2.14 0.08 0.11 0.42 18.44 1.53 0.07 0.52

old saplings 0.07 0.45 2.64 0.00 0.00 0.99 39.58 2.61 0.15 0.35

Fig. 2.4: Ordination plot of the dissimilarities between plant species assemblages of three life stages across fragments. The direction and length of an arrow represents its position on the ordination axes. The NMDS analysis with four ordination axes had a ‘stress’ of 0.09. See Table 2.1 for species keys and Table 2.2 for quantitative results of the analysis.

Fig. 2.5: Ordination plot of the dissimilarities between seed assemblages across fragments. The direction and length of an arrow represents its position on the ordination axes. The NMDS analysis with two ordination axes had a ‘stress’ of 0.10. See Table 2.1 for species keys and Table 2.2 for quantitative results of the analysis.

2.5 Discussion

We found evidence to support our hypothesis that the ratios of sapling to adult stem numbers of certain woody plant species are modified by fragment size. We also found that the composition of young saplings (<1 cm girth) was more sensitive to variation in fragment size than the compositions of other life stages we examined, including seed rain. Our results indicate that fragmentation, or more specifically reduced fragment size, leaves a repeatable signature on processes that act after seed dispersal, but before saplings establish. However, because we monitored seeds for only a short period and recorded only 27 species (including only five of the common species), we would caution that our results regarding patterns in seed rain should be considered preliminary. Patterns of viable deposited seeds depend on plant-animal mutualisms that facilitate pre-recruitment processes like pollination, pre-dispersal seed predation, seed dispersal and post-dispersal seed predation (Wang and Smith 2002). As animals that are poor dispersers may be unable to move across large areas of unsuitable habitat

(MacArthur and Wilson 1967, Didham et al. 1996, Cirtwill and Stouffer 2016), their movement, and consequently mutualisms, might be disrupted in a fragmented landscape. The effects of fragmentation on pre-recruitment processes, therefore, warrant closer examination, especially considering that each of these processes is vulnerable to fragmentation (Wright and

Duber 2001, Aguilar et al. 2006, Herrerías-Diego et al. 2008, Neuschulz et al. 2016, Luskin et al. 2017).

The movement of insect pollinators (Aizen and Feinsinger 2003), and of insect pre-dispersal seed predators (Lewis and Gripenberg 2008), may be influenced in fragmented landscapes, as insects are often not particularly mobile due to their small sizes (Didham et al. 1996). Powell and Powell (1987) found that certain Euglossine bees struggle to move across stretches of unfavourable habitat as small as 100 m, potentially causing pollen limitation by decreasing pollen deposition in fragments, and by restricting pollen flow between fragments. Several early

25 reviews of pollination studies concluded that pollen limitation can result in fewer seeds produced (Burd 1994), can impair seedling vigour (Walsh and Charlesworth 1992), and can cause inbreeding depression that may affect the long-term stability of plant populations (Bawa

1990, Barrett and Harder 1996). The consequences of pollen limitation may be especially pronounced for self-incompatible plant species that are obligate outbreeders (Richards 1997,

Wilcock and Neiland 2002), and for plant species that have specialized interactions with pollinators (Burd 1994, Ashworth et al. 2004). In a fragmented landscape in Costa Rica,

Ghazoul and McLeish (2001) showed that the seed production of a self-incompatible tree species decreased in small fragments, despite similar abundances of its pollinator (Trigona bees) in small and in large fragments, possibly due to reduced gene flow during pollination.

While plant-pollinator interactions have received considerable attention, interactions between plants and pre-dispersal predators are poorly understood (Lewis and Gripenberg 2008). Despite evidence that pre-dispersal seed predation by insects is a very important process that can structures plant communities (Xu et al. 2015), the effects of fragmentation on these plant- predator interactions have rarely been investigated (Lewis and Gripenberg 2008). In a fragmented dry tropical forest, Herrerías-Diego et al. (2008) found that pre-dispersal predation by a cotton staining bug was higher in contiguous forest than in fragments. De Crop et al.

(2012) demonstrated that the predation of Centaurium erythraea fruits by a specialist moth was lower in fragmented, isolated populations than otherwise. There is compelling evidence, therefore, that extreme fragmentation can negatively influence pollination, (consequently) seed production (Cunningham 2000, Aizen et al. 2002, Ghazoul 2005, Aguilar et al. 2006) and pre- dispersal predation (Herrerías-Diego et al. 2008, De Crop et al. 2012).

The effects of less extreme fragmentation on plant-insect interactions, however, are contingent on the interacting species, and on the landscape (Tscharntke and Brandl 2004). Some pollination systems, for example, may be relatively generalized and resistant to the effects of

26 fragmentation, as was inferred by Waser et al. (1996) from studies of two temperate floras.

Aguirre and Dirzo (2008) found that decreases in the abundance of pollinators of Astrocaryum mexicanum with fragment size did not correlate with decreases in its fruit set. They suggest that inherently large numbers of pollinators may allow plant species to retain their reproductive abilities even when some mutualisms are disrupted. Brudvig et al. (2015) conducted a controlled experiment in a Pine-Savannah landscape to investigate the effects of fragmentation on processes facilitated by plant-insect interactions, only to find that fragmentation did not influence either pollination or pre-dispersal seed predation by insects. Burgos et al. (2008) also found no effects of fragmentation on pre-dispersal predation in central Chile. In our study, we observed the lack of a strong effect of fragment size on seed assemblages. This suggests that pollination disruption was not a strong driver of community change but says less about pre- dispersal predation, as damaged seeds are often dispersed. The apparent lack of pollination effects in our study landscape may be due the high level of connectedness among even small forest fragments, where few mutualisms are disrupted and even fewer are specialized.

Subsequent processes in the life cycles of plants, like seed dispersal and post-dispersal seed predation, are primarily facilitated by birds and mammals (Wang and Smith 2002). Given that even the most isolated fragments are no more than 100 m from other patches of forest (and often from contiguous forest), we expect relatively mobile animals like avian and mammalian dispersers to remain relatively unaffected by the fragmentation observed (Markl et al. 2012).

Certain bird species in the same geographical region, for example, were primarily found to disperse seeds of Dysoxylum malabaricum within 200 m (Ismail et al. 2017), a dispersal distance that would ensure connectivity in our study site. Birds and mammals may therefore be sensitive only to fragmentation that is more extreme than that at the study site, and this is reflected in some studies that found that fragmentation did not affect seed dispersal (Markl et al. 2012) and post-dispersal seed predation (Dennis et al. 2005). The lack of an effect of

27 fragmentation on seed rain therefore suggests (again), perhaps unsurprisingly, that modified patterns of seed dispersal are not responsible for the observed changes in plant communities.

We therefore suggest that the observed relationship between fragment size and young sapling composition (Fig. 2.4, Table 2.2) was not primarily determined by changes in pre-dispersal processes, but by alterations to early plant recruitment processes, which are considered particularly important in forest regeneration (Harms et al. 2000, Green et al. 2014). Uriarte et al. (2010) showed that seedling recruitment may be influenced by abiotic factors and suggest that the effects of fragmentation on abiotic processes may be more important than effects on biotic processes, especially when the non-forest matrix allows movement of dispersers and pollinators (as in the study landscape). A variety of abiotic changes might be induced by fragmentation (Murcia 1995, Ewers and Didham 2006) including changes in light (Uriarte et al. 2010), temperature (Ewers and Banks-Leite 2013) and humidity (Bruna 1999). Bruna (1999) found that leaf litter decomposed more slowly in fragments than in contiguous forest, and that seed germination was consequently inhibited by leaf litter in fragments. We suggest, however, that changes in light and microclimatic conditions alone are unlikely to have driven the observed changes in our study, as canopy cover was relatively consistent across all fragments

(Fig. 2.1). Although we cannot rule out a role for altered microclimate in driving the changes in plant communities observed in small fragments, the lack of fragment-size linked variation in the important microclimate variable we measured suggests that other forces may be important.

This study, albeit with a limited sample size, suggests that fragment size may influence plant species richness and composition, and highlights the early stages of seedling growth as a critical period for plant community change. Both abiotic and biotic alterations in fragmented forests could play a role, but the lack of differences in light environment among fragments of different sizes suggest other forces may be at work. A particularly interesting hypothesis is that

28 alterations to biotic interactions in forest fragments (Benitez-Malvido et al. 1999, Benítez-

Malvido and Lemus-Albor 2005, Laurance et al. 2011) might play an additional role in driving changes in plant communities at the establishment stage. These may include disruptions to specialized negative density-dependent plant-enemy interactions, described by the Janzen-

Connell hypothesis (Janzen 1970, Connell 1971), that play important roles in the tropics and maintain plant diversity (Mangan et al. 2010, Bagchi et al. 2014, Comita et al. 2014). The loss of such interactions can increase recruitment of common species in small fragments, as was observed in this study, due to decreased natural enemy abundance (Benitez-Malvido et al.

1999, Arnold and Asquith 2002, Fáveri et al. 2008). Our results therefore serve as motivation to examine the effects of fragment size on these processes. In the following chapters, we directly investigate the effects of fragment size on seedling-enemy interactions (including

Janzen-Connell interactions), to understand if these processes are indeed influenced by fragmentation.

2.6 Acknowledgements

Kadamane Estate provided us with an ideal set up for our study, a place to stay and people to help. We thank Mr. K.M. Cariappa, Senior Manager, Kadamane Estates Company, for his personal interest in the study. We are grateful to Praveen Kumar, Ganesh Honwad and Vinod

Shanker for their help in documenting plants in the quadrats. The research was funded by ETH

Grant 42 13-1.

2.7 References

Aguilar, R., L. Ashworth, L. Galetto, and M. A. Aizen. 2006. Plant reproductive susceptibility

to habitat fragmentation: review and synthesis through a meta-analysis. Ecology Letters

9:968-980.

Aguirre, A., and R. Dirzo. 2008. Effects of fragmentation on pollinator abundance and fruit set

of an abundant understory palm in a Mexican tropical forest. Biological Conservation

141:375-384.

Aizen, M. A., L. Ashworth, L. Galetto, and S. Díaz. 2002. Reproductive success in fragmented

habitats: do compatibility systems and pollination specialization matter? Journal of

Vegetation Science 13:885-892.

Aizen, M. A., and P. Feinsinger. 2003. Bees Not to Be? Responses of Insect Pollinator Faunas

and Flower Pollination to Habitat Fragmentation. Pages 111-129 in G. A. Bradshaw

and P. A. Marquet, editors. How Landscapes Change: Human Disturbance and

Ecosystem Fragmentation in the Americas. Springer Berlin Heidelberg, Berlin,

Heidelberg.

Arnold, A. E., and N. M. Asquith. 2002. Herbivory in a fragmented tropical forest: patterns

from islands at Lago Gatún, Panama. Biodiversity & Conservation 11:1663-1680.

Arroyo-Rodríguez, V., and S. Mandujano. 2006. The Importance of Tropical Rain Forest

Fragments to the Conservation of Plant Species Diversity in Los Tuxtlas, Mexico.

Biodiversity & Conservation 15:4159-4179.

Ashworth, L., R. Aguilar, L. Galetto, and M. A. Aizen. 2004. Why do pollination generalist

and specialist plant species show similar reproductive susceptibility to habitat

fragmentation? Journal of Ecology 92:717-719.

Bagchi, R., R. E. Gallery, S. Gripenberg, S. J. Gurr, L. Narayan, C. E. Addis, R. P. Freckleton,

and O. T. Lewis. 2014. Pathogens and insect herbivores drive rainforest plant diversity

and composition. Nature 506:85-88.

Barrett, S. C. H., and L. D. Harder. 1996. Ecology and evolution of plant mating. Trends in

Ecology & Evolution 11:73-79.

Bawa, K. S. 1990. Plant-Pollinator Interactions in Tropical Rain Forests. Annual Review of

Ecology and Systematics 21:399-422.

Benitez-Malvido, J., G. Garcı́a-Guzmán, and I. D. Kossmann-Ferraz. 1999. Leaf-fungal

incidence and herbivory on tree seedlings in tropical rainforest fragments: an

experimental study. Biological Conservation 91:143-150.

Benítez-Malvido, J., and A. Lemus-Albor. 2005. The Seedling Community of Tropical Rain

Forest Edges and Its Interaction with Herbivores and Pathogens. Biotropica 37:301-

313.

Benítez-Malvido, J., and M. Martínez-Ramos. 2003. Influence of Edge Exposure on Tree

Seedling Species Recruitment in Tropical Rain Forest Fragments. Biotropica 35:530-

541.

Brudvig, L. A., E. I. Damschen, N. M. Haddad, D. J. Levey, and J. J. Tewksbury. 2015. The

influence of habitat fragmentation on multiple plant–animal interactions and plant

reproduction. Ecology 96:2669-2678.

Bruna, E. M. 1999. Seed germination in rainforest fragments. Nature 402:139.

Burd, M. 1994. Bateman’s principle and plant reproduction: The role of pollen limitation in

fruit and seed set. The Botanical Review 60:83-139.

Burgos, A., A. A. Grez, and R. O. Bustamante. 2008. Seed production, pre-dispersal seed

predation and germination of Nothofagus glauca (Nothofagaceae) in a temperate

fragmented forest in Chile. Forest Ecology and Management 255:1226-1233.

Cirtwill, A. R., and D. B. Stouffer. 2016. Knowledge of predator–prey interactions improves

predictions of immigration and extinction in island biogeography. Global Ecology and

Biogeography 25:900-911.

Colwell, R. K., and J. A. Coddington. 1994. Estimating terrestrial biodiversity through

extrapolation. Philosophical Transactions of the Royal Society of London B: Biological

Sciences 345:101-118.

Comita, L. S., S. A. Queenborough, S. J. Murphy, J. L. Eck, K. Xu, M. Krishnadas, N.

Beckman, and Y. Zhu. 2014. Testing predictions of the Janzen–Connell hypothesis: a

meta-analysis of experimental evidence for distance- and density-dependent seed and

seedling survival. Journal of Ecology 102:845-856.

Connell, J. H. 1971. On the role of natural enemies in preventing competitive exclusion in some

marine animals and in rain forest trees. Pages 298-312 in P. J. den Boer and G.

Gradwell, editors. Dynamics of populations. PUDOC, Wageningen.

Cunningham, S. A. 2000. Depressed pollination in habitat fragments causes low fruit set.

Proceedings of the Royal Society of London. Series B: Biological Sciences 267:1149-

1152.

Daily, G. C. 2001. Ecological forecasts. Nature 411:245.

De Crop, E., R. Brys, and M. Hoffmann. 2012. The impact of habitat fragmentation on the

interaction between Centaurium erythraea (Gentianaceae) and its specialized seed

predator Stenoptilia zophodactylus (Pterophoridae, Lepidoptera). Ecological Research

27:967-974.

Dennis, A. J., G. J. Lipsett-Moore, G. Harrington, E. Collins, and D. A. Westcott. 2005. Seed

predation, seed dispersal and habitat fragmentation: does context make a difference in

tropical Australia. Seed fate: predation, dispersal and seedling establishment. CAB

International, Wallingford, UK:117-135.

Didham, R. K., J. Ghazoul, N. E. Stork, and A. J. Davis. 1996. Insects in fragmented forests: a

functional approach. Trends in Ecology & Evolution 11:255-260.

Ewers, R. M., and C. Banks-Leite. 2013. Fragmentation Impairs the Microclimate Buffering

Effect of Tropical Forests. PLoS ONE 8:e58093.

Ewers, R. M., and R. K. Didham. 2006. Confounding factors in the detection of species

responses to habitat fragmentation. Biological Reviews 81:117-142.

Fáveri, S. B., H. L. Vasconcelos, and R. Dirzo. 2008. Effects of Amazonian forest

fragmentation on the interaction between plants, insect herbivores, and their natural

enemies. Journal of Tropical Ecology 24:57-64.

Ferraz, G., G. J. Russell, P. C. Stouffer, R. O. Bierregaard, S. L. Pimm, and T. E. Lovejoy.

2003. Rates of species loss from Amazonian forest fragments. Proceedings of the

National Academy of Sciences 100:14069-14073.

Ghazoul, J. 2005. Pollen and seed dispersal among dispersed plants. Biological Reviews

80:413-443.

Ghazoul, J., and M. McLeish. 2001. Reproductive ecology of tropical forest trees in logged

and fragmented habitats in Thailand and Costa Rica. Pages 335-345 in K. E.

Linsenmair, A. J. Davis, B. Fiala, and M. R. Speight, editors. Tropical Forest Canopies:

Ecology and Management: Proceedings of ESF Conference, Oxford University, 12–16

December 1998. Springer Netherlands, Dordrecht.

Green, P. T., K. E. Harms, and J. H. Connell. 2014. Nonrandom, diversifying processes are

disproportionately strong in the smallest size classes of a tropical forest. Proceedings

of the National Academy of Sciences 111:18649-18654.

Haddad, N. M., L. A. Brudvig, J. Clobert, K. F. Davies, A. Gonzalez, R. D. Holt, T. E. Lovejoy,

J. O. Sexton, M. P. Austin, C. D. Collins, W. M. Cook, E. I. Damschen, R. M. Ewers,

B. L. Foster, C. N. Jenkins, A. J. King, W. F. Laurance, D. J. Levey, C. R. Margules,

B. A. Melbourne, A. O. Nicholls, J. L. Orrock, D.-X. Song, and J. R. Townshend. 2015.

Habitat fragmentation and its lasting impact on Earth’s ecosystems. Science Advances

Harms, K. E., S. J. Wright, O. Calderon, A. Hernandez, and E. A. Herre. 2000. Pervasive

density-dependent recruitment enhances seedling diversity in a tropical forest. Nature

404:493-495.

Herrerías-Diego, Y., M. Quesada, K. E. Stoner, J. A. Lobo, Y. Hernández-Flores, and G.

Sanchez Montoya. 2008. Effect of forest fragmentation on fruit and seed predation of

the tropical dry forest tree Ceiba aesculifolia. Biological Conservation 141:241-248.

Hill, J. L., and P. J. Curran. 2003. Area, shape and isolation of tropical forest fragments: effects

on tree species diversity and implications for conservation. Journal of Biogeography

30:1391-1403.

Ismail, S. A., J. Ghazoul, G. Ravikanth, C. G. Kushalappa, R. Uma Shaanker, and C. J. Kettle.

2017. Evaluating realized seed dispersal across fragmented tropical landscapes: a two-

fold approach using parentage analysis and the neighbourhood model. New Phytologist

214:1307-1316.

Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests. American

Naturalist 104:501-528.

Laurance, W. F., J. L. Camargo, R. C. Luizão, S. G. Laurance, S. L. Pimm, E. M. Bruna, P. C.

Stouffer, G. B. Williamson, J. Benítez-Malvido, and H. L. Vasconcelos. 2011. The fate

of Amazonian forest fragments: a 32-year investigation. Biological Conservation

144:56-67.

Laurance, W. F., L. V. Ferreira, J. M. R.-D. Merona, S. G. Laurance, R. W. Hutchings, and T.

E. Lovejoy. 1998. Effects of Forest Fragmentation on Recruitment Patterns in

Amazonian Tree Communities. Conservation Biology 12:460-464.

Laurance, W. F., H. E. M. Nascimento, S. G. Laurance, A. Andrade, J. E. L. S. Ribeiro, J. P.

Giraldo, T. E. Lovejoy, R. Condit, J. Chave, K. E. Harms, and S. D'Angelo. 2006. Rapid

decay of tree-community composition in Amazonian forest fragments. Proceedings of

the National Academy of Sciences 103:19010-19014.

Lewis, O. T., and S. Gripenberg. 2008. Insect seed predators and environmental change.

Journal of Applied Ecology 45:1593-1599.

Lovejoy, T., R. Bierregaard, A. Rylands, J. Malcolm, C. Quintela, L. Harper, K. Brown, A.

Powell, G. Powell, and H. Schubart. 1986. Edge and other effects of isolation on

Amazon forest fragments. Pages 257-285 in M. E. Soulé, editor. Conservation biology:

the science of scarcity and diversity. Sinauer, Sunderland, Massachusetts.

Luskin, M. S., J. S. Brashares, K. Ickes, I. F. Sun, C. Fletcher, S. J. Wright, and M. D. Potts.

2017. Cross-boundary subsidy cascades from oil palm degrade distant tropical forests.

Nature Communications 8:2231.

MacArthur, R. H., and E. Wilson. 1967. The theory of island biogeography. Monographs in

Population Biology 1.

Mangan, S. A., S. A. Schnitzer, E. A. Herre, K. M. L. Mack, M. C. Valencia, E. I. Sanchez,

and J. D. Bever. 2010. Negative plant-soil feedback predicts tree-species relative

abundance in a tropical forest. Nature 466:752-755.

Markl, J. S., M. Schleuning, P. M. Forget, P. Jordano, J. E. Lambert, A. Traveset, S. J. Wright,

and K. Böhning-Gaese. 2012. Meta-Analysis of the Effects of Human Disturbance on

Seed Dispersal by Animals. Conservation Biology 26:1072-1081.

Murcia, C. 1995. Edge effects in fragmented forests: implications for conservation. Trends in

Ecology & Evolution 10:58-62.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. Da Fonseca, and J. Kent. 2000.

Biodiversity hotspots for conservation priorities. Nature 403:853-858.

Neuschulz, E. L., T. Mueller, M. Schleuning, and K. Böhning-Gaese. 2016. Pollination and

seed dispersal are the most threatened processes of plant regeneration. Scientific

Reports 6:29839.

Oksanen, J., F. G. Blanchet, M. Friendly, R. Kindt, P. Legendre, D. McGlinn, P. R. Minchin,

R. B. O'Hara, G. L. Simpson, P. Solymos, M. H. H. Stevens, E. Szoecs, and H. Wagner.

2017. vegan: Community Ecology Package.

Powell, A. H., and V. N. G. Powell. 1987. Population Dynamics of Male Euglossine Bees in

Amazonian Forest Fragments. Biotropica 19:176-179.

R Core Team. 2017. R: A language and environment for statistical computing. R Foundation

for Statistical Computing, Vienna, Austria.

Richards, A. J. 1997. Plant breeding systems. Garland Science.

Sasidharan, N. 2013. Flowering plants of Kerala: CD-ROM ver 2.0. Kerala Forest Research

Institute, Peechi, Kerala.

Stouffer, P. C., C. Strong, and L. N. Naka. 2009. Twenty years of understorey bird extinctions

from Amazonian rain forest fragments: consistent trends and landscape-mediated

dynamics. Diversity and Distributions 15:88-97.

Tscharntke, T., and R. Brandl. 2004. Plant-insect interactions in fragmented landscapes. Annu

Rev Entomol 49:405-430.

Turner, I. M. 1996. Species Loss in Fragments of Tropical Rain Forest: A Review of the

Evidence. Journal of Applied Ecology 33:200-209.

Uriarte, M., E. M. Bruna, P. Rubim, M. Anciães, and I. Jonckheere. 2010. Effects of forest

fragmentation on the seedling recruitment of a tropical herb: assessing seed vs. safe-

site limitation. Ecology 91:1317-1328.

Walsh, N. E., and D. Charlesworth. 1992. Evolutionary Interpretations of Differences in Pollen

Tube Growth Rates. The Quarterly Review of Biology 67:19-37.

Wang, B. C., and T. B. Smith. 2002. Closing the seed dispersal loop. Trends in Ecology &

Evolution 17:379-386.

Waser, N. M., L. Chittka, M. V. Price, N. M. Williams, and J. Ollerton. 1996. Generalization

in Pollination Systems, and Why it Matters. Ecology 77:1043-1060.

Wilcock, C., and R. Neiland. 2002. Pollination failure in plants: why it happens and when it

matters. Trends in Plant Science 7:270-277.

Wood, S. N. 2017. Generalized additive models: an introduction with R. CRC press.

Wright, S. J., and H. C. Duber. 2001. Poachers and Forest Fragmentation Alter Seed Dispersal,

Seed Survival, and Seedling Recruitment in the Palm Attalea butyraceae, with

Implications for Tropical Tree Diversity. Biotropica 33:583.

Wright, S. J., and H. C. Muller-Landau. 2006. The uncertain future of tropical forest species.

Biotropica 38:287-301.

Xu, Y., Z. Shen, D. Li, and Q. Guo. 2015. Pre-Dispersal Seed Predation in a Species-Rich

Forest Community: Patterns and the Interplay with Determinants. PLoS ONE

10:e0143040.

3 The effects of rainforest fragment area and intraspecific seed mass on the strength of plant-pathogen interactions

Ashwin Viswanathan*a, Jaboury Ghazoula, Ganesh Honwadb, N. Arun Kumarc, and Robert

Bagchid a Chair of Ecosystem Management, ITES, ETH Zurich, Zurich 8092, Switzerland b Center for Innovation Research and Consultancy, Pune, India c Forest Research Institute, Dehradun, Uttarakhand 248006, India d Department of Ecology and Evolutionary Biology, University of Connecticut

Storrs, CT 06269, USA

* Corresponding author: [email protected]; +919483512541

3.1 Abstract

Pathogenic interactions between fungi and plants are thought to facilitate plant species coexistence and maintain diversity in tropical rainforests. Such interactions, however, may be affected by forest fragmentation as fungi are susceptible to various forms of anthropogenic disturbance. To understand the effects of fragmentation on fungus-induced seed and seedling mortality, we conducted a shadehouse experiment with seeds sown in soils collected from 21 forest fragments. We compared seedling establishment of six plant species in unmanipulated soils and soils where fungi were suppressed with fungicides. We also investigated, in the same experiment, how intraspecific variation in seed mass contributed to variation in the susceptibility of five of these species to pathogenic fungi. Seed mass did not influence susceptibility to pathogens, but larger seeds of certain species were more likely to germinate.

We found evidence that pathogens suppressed the germination of Toona ciliata seeds and increased the mortality of Syzygium rubicundum and Olea dioica seedlings. The fungus- induced mortality of one of these species, S. rubicundum, decreased with decreasing fragment size, indicating that its interactions with pathogenic fungi may weaken as fragments become smaller. We therefore provide evidence of the weakening of a potential diversity-maintaining plant-fungus interaction in small forest fragments and suggest that such disruptions may have important long-term consequences for plant diversity. We, however, emphasize the need for further research across rainforest plant communities to better understand the future of diversity in fragmented rainforest landscapes.

Keywords: mortality, fragmentation, diversity, plant-fungus interactions, Syzygium rubicundum

3.2 Introduction

Remarkably large numbers of plant species coexist on few resources in tropical rainforests.

The Janzen-Connell hypothesis (JCH) proposes that specialist natural enemies inhibit the survival of plant species that are locally abundant, thereby reducing competitive exclusion of species that are locally rare (Janzen 1970, Connell 1971). The JCH suggests that density- dependent attack on plants by their natural enemies maintains plant diversity in tropical forests (Janzen 1970, Connell 1971). The hypothesis has received considerable empirical support in recent years (Bagchi et al. 2014, Comita et al. 2014), and pathogenic fungi, soil biota and insects have been shown to regulate populations of plant species through negative density-dependent control of seeds and seedlings (Bell et al. 2006, Bagchi et al. 2010,

Mangan et al. 2010, Fricke et al. 2014). While all natural enemies may play roles, studies suggest that fungal pathogens may be especially important for the maintenance of plant diversity in rainforests (Mangan et al. 2010, Bagchi et al. 2014).

We know little, however, about how anthropogenic perturbations may affect interactions between plants and their fungal pathogens (Benitez-Malvido et al. 1999, Benítez-Malvido and

Lemus-Albor 2005), and consequently plant diversity. Such an understanding may be particularly relevant as a large proportion of tropical rainforests are threatened by fragmentation (Ewers and Didham 2006, Laurance et al. 2011, Haddad et al. 2015).

Fragmented forests contain fewer species than similar areas of contiguous forest (Ewers and

Didham 2006, Laurance et al. 2011, Haddad et al. 2015), but act as important reservoirs of biological diversity in the absence of contiguous forest (Dent and Wright 2009, Fahrig 2017).

The long-term future of this biodiversity may depend on the resilience of processes that maintain plant diversity, like those mediated by pathogenic fungi. There is already some evidence that fungi are sensitive to the altered microclimates typical of fragmented forests

(Benitez-Malvido et al. 1999, Benítez-Malvido and Lemus-Albor 2005, Ewers and Didham

2006, Laurance et al. 2011), indicating that coexistence mechanisms that depend on fungi might be disrupted. To understand how fragmentation modifies processes that maintain diversity, we investigated whether the effects of soil-borne fungal pathogens on rainforest plant species are influenced by fragment size. We focused on soil-borne fungi, because their ability to regulate plant populations has been highlighted by previous research (Mangan et al. 2010).

Previous evidence that forest fragmentation reduces the incidence of leaf fungal infection

(Benitez-Malvido et al. 1999) led us to hypothesize that 1) fungus-mediated mortality of seedlings would decrease as fragments become smaller. We tested our hypothesis in a shadehouse experiment by comparing plant performance in rainforest soils collected from a gradient of fragment sizes under two treatments (fungicide-treated and control).

We also used this opportunity to investigate whether fragmentation can exert evolutionary pressures on seed mass, similar to hunting in Galetti et al. (2013). We identified seed mass as a trait that is worth investigating as it is sensitive to changes even at fine temporal scales

(Galetti et al. 2013), and is strongly associated with seed and seedling survival (Leishman et al. 2000). If the mortality caused by fungi does decrease in small fragments, certain morphological traits of seeds and seedlings that are linked with resisting pathogen attack may experience long-term evolutionary consequences in such landscapes. As relatively large seedlings from large seeds are more resistant to stress (within and between species) due to increased seedling vigour and increased food reserves from larger cotyledons (Westoby et al.

1996, Leishman et al. 2000), we also asked the following question: Does resistance to a stressor like pathogenic fungi increase with increasing intraspecific seed size? On the premise that increasing seed size enhances the likelihood of survival, we hypothesized 2) that the enhanced survival of large seeds through to establishment is facilitated in part by their resistance to fungi.

3.3 Methods

3.3.1 Study area

This study was conducted in the Kadamane Tea Estate (12.8639º – 12.5620º N and 75.6361 –

75.6833º E), located in the Western Ghats in Karnataka, India. This 30-km2 private estate is a mosaic of fragmented rainforest, high altitude grassland and cultivated tea. Forest fragments were mostly created during the late 1960s and early 1970s and range between 1 - 149 ha in size within a relatively small range of elevation (800 – 1300 m).

3.3.2 Study design

We collected soil samples (ca. 0.06 m3 each, from up to 30 cm deep including leaf litter) from

21 fragments varying in size from 1 – 149 ha (mean 27.8 ha, altitude range 918 – 1071 m).

Samples from 30 random locations in each fragment were pooled and thoroughly mixed to obtain a single representative source of soil-borne fungi for each forest fragment. Soil from each fragment was placed in twelve 20 X 10 X 10 cm trays (that allowed water drainage) after equal and random distribution between ‘control’ and ‘fungicide’ treatments. All trays were kept in a 75% shaded greenhouse constructed of mesh that allowed continuous homogeneous percolation of rainwater but excluded large insects. No insect herbivory was observed inside the greenhouse.

We collected the seeds of six species of trees (S Table 3.1), discarding seeds that showed evidence of pre-dispersal predation by insects or floated in water (a basic test of viability).

We sowed between four and five seeds (depending on availability) of each species in each of

126 trays (3 replicates of 4-5 seeds/treatment/fragment) after weighing each one (Mettler

Toledo JS 3002G® balance with 0.01 g resolution). We monitored germination and seedling mortality at five censuses 15-30 days apart between May 23, 2017 to October 9, 2017, or for

43 as long as it took for most seedlings of a species to grow four open leaves and lose their cotyledons (Table 1). During each census, we recorded the status of each seed as ‘0’ – not germinated/dead; ‘1’ – germinated/alive.

We used a combination of a systemic fungicide Azoxystrobin (Product name: Amistar®,

Syngenta, Basel, Switzerland; 0.4 ml/litre water), a contact fungicide Mancozeb and an oomyticide Metalxyl (Product name: Ridomil Gold® Syngenta; 6 g/litre water) to exclude fungal pathogens. We sprayed approximately 8 ml of diluted Amistar® and 8 ml of diluted

Ridomil Gold® on the trays in the fungicide treatment every eight days in the dry season, and every four days in the wet season (to account for immediate runoff in heavy rain). To control for effects of supplemental water on germination and survival, we sprayed the trays in the control treatment with an equal volume of water as was used to dissolve the fungicides.

3.3.3 Data analysis

Using mixed effects Cox proportional hazards models, we modelled time to germination or mortality as functions of fungicide treatment, fragment size, seed mass, and their two-way interactions, and included seedling tray as a random effect. We analysed the germination of all six species individually but analysed the seedling mortality of only three species individually

(those with sufficient dead seedlings). In a separate model, we combined the data from the remaining three species. We ‘censored’ all seeds that did not germinate in the germination analysis and all seedlings that survived in the mortality analysis. Models were fit using the coxme package (Therneau 2015) from R 3.4.0 (R Core Team 2017). Model diagnostics and alternate models were examined to confirm that statistical assumptions were met (see 3.8

Supplementary information).

3.4 Results

Of the 3483 seeds of six species, 2171 germinated, of which 258 subsequently died. Fungicide treatment increased the likelihood of germination for Toona ciliata (Table 3.1, Fig. 3.1), and decreased the likelihood of seedling mortality for Syzygium rubicundum and Olea dioica.

(Table 3.2, Fig. 3.2). Germination and survival of the other three species were unaffected by fungicide application.

3.4.1 The effects of fragment size

Germination was statistically independent of fragment size for all species (Table 3.1). Seedling mortality of one species, Syzygium rubicundum, was affected by fragment size (Table 3.2, Fig.

3.2). In the control treatment, mortality of Syzygium rubicundum increased with increasing fragment size but the relationship disappeared with fungicide treatment (Table 3.2, Fig. 3.2).

Survival of the remaining species was unaffected by fragment size in both the control and fungicide treatments.

3.4.2 The effects of seed mass

The likelihood of seed germination of Syzygium rubicundum, Symplocos racemosa and

Syzygium cumini increased with increasing intraspecific seed mass (Table 3.1, Fig. 3.2) but the likelihood of seedling mortality was not influenced by seed mass. The effects of fungicide on seed germination and seedling mortality were independent of seed mass in all species (Table

3.1, Table 3.2, and Fig. 3.2).

Fig. 3.1: Effects of fungicide application and seed mass on the probabilities of mortality of seedlings of six rainforest tree species. Observed values at the end of the monitoring period are plotted with standard errors from survival curves. Lines are predicted from mixed-effects cox proportional hazards models (Table 3.1) and are plotted with standard errors. Degree of shading represents the number of seedlings in a sample.

Fig. 3.2: Effects of fungicide application and fragment size on the probabilities of mortality of seedlings of six rainforest tree species. Observed values at the end of the monitoring period are plotted with standard errors from survival curves. Lines are predicted from mixed-effects cox proportional hazards models (Table 3.2) and are plotted with standard errors. Degree of shading represents the number of seedlings in a sample.

Table 3.1: Results of mixed effects Cox proportional hazards models examining the effects of fungicide application, fragment size (where possible) and seed mass (scale 0.27) on germination.

Syzygium rubicundum Olea dioica Symplocos racemosa

random effects: N var N var N var

site 21 0.02058 21 0.10294 21 0.01920

fixed effects: est. se est. se est. se

fungicide 0.002 0.32 -0.274 0.50 -0.128 0.31

fragsize 0.122 0.16 0.163 0.33 0.011 0.23

fragsize: fungicide -0.019 0.13 -0.002 0.13 -0.060 0.14

seedmass 0.336 0.20 -0.321 0.59 1.042 0.38

seedmass: fungicide 0.110 0.25 0.636 0.73 -0.121 0.48

seedmass: fragsize 0.012 0.12 -0.209 0.46 0.152 0.30

Syzygium cumini Syzygium gardneri Toona ciliata

random effects: N var N var N var

site 21 0.00008 21 0.00002 21 0.00008

fixed effects: est. se est. se est. se

fungicide -0.633 0.43 -0.398 0.48 0.537 0.19

fragsize -0.180 0.25 -0.382 0.29 0.117 0.17

fragsize: fungicide 0.257 0.17 0.173 0.16 -0.252 0.21

seedmass 0.363 0.32 -0.024 0.28

seedmass: fungicide 0.411 0.45 0.270 0.33

seedmass: fragsize -0.034 0.27 0.156 0.20

Table 3.2: Results of mixed effects Cox proportional hazards models examining the effects of fungicide application, fragment size (scale 47) and seed mass on seedling mortality. Only two covariates were analysed at a time to avoid overfitting.

fragment size

Syzegium rubicundum Olea dioica Symplocos racemosa other three species

N events | total: 40 | 322 events | total: 64 | 313 events | total: 53 | 357 events | total: 35 | 744 random effects: N Var N Var N Var N Var site 21 0.0004 21 0.0771 21 0.2553 21 0.7418 fixed effects: est. se est. se est. se est. se fungicide -0.30 0.43 -0.69 0.32 0.21 0.36 0.15 0.41 fragsize 0.46 0.16 -0.05 0.19 0.41 0.28 0.21 0.44 fragsize: fungicide -1.18 0.57 -0.28 0.36 -0.25 0.34 -0.40 0.49

seed mass

Syzegium rubicundum Olea dioica Symplocos racemosa other three species

N events | total: 40 | 322 events | total: 64 | 313 events | total: 53 | 357 random effects: N Var N Var N Var N Var site 21 0.0004 21 0.0771 21 0.2553 fixed effects: est. se est. se est. se est. se fungicide -1.33 1.06 1.37 1.32 -0.28 0.83 seedmass -0.02 0.50 0.96 1.15 -0.62 0.93 seedmass: fungicide 0.24 1.27 -3.50 2.08 0.59 1.32

3.5 Discussion

3.5.1 The effects of fragment size on plant-fungus interactions

Suppressing soil-borne fungi influenced the survival of three plant species, but fragment size only significantly influenced the effects of fungi on Syzygium rubicundum. Seedling mortality of Syzygium rubicundum increased strongly with fragment size in the control treatment, but this effect was removed completely when fungi were suppressed. Our results for this one species, therefore, supported our hypothesis that mortality due to fungi may be affected by perturbations from forest fragmentation. Our findings complement those of Krishnadas and

Comita (2018) who found evidence, in the same study area, that the pathogen-induced seed/seedling mortality of two species of rainforest trees, Toona ciliata and Olea dioica, were influenced in soils from habitat edges. Although we independently detected effects of pathogenic fungi in the same two species during the same stages of seedling establishment, we did not detect any interactions with fragment size. We, however, found evidence that O. dioica seedlings were suppressed by pathogenic fungi in soils from several ‘interiors’, unlike

Krishnadas and Comita (2018) who found that the effects of fungicide on O. dioica manifested only at edges.

Several studies have now shown that communities of fungal pathogens are sensitive to various forms of anthropogenic disturbance. Benitez-Malvido et al. (1999) demonstrated sensitivity to fragmentation. They found that leaves of seedlings (of two species of trees) in central

Amazonia that were damaged by insect herbivores had lower fungal incidence in forest fragments than in contiguous forest. Benítez-Malvido and Lemus-Albor (2005) subsequently demonstrated that fungal pathogens were sensitive to edge effects. They found, in a rainforest in Mexico, that damage due to pathogens was two times higher at edges than in interior forest.

Swinfield et al. (2012) (shadehouse experiment) and Thompson et al. (2010) (theory informed

50 by field data) demonstrated sensitivity to altered environmental conditions. Swinfield et al.

(2012) showed that seedlings of Pleradenophora longicuspis experienced greater pathogen related mortality when they either received more water, or were watered more frequently while

Thompson et al. (2010) showed that potential future shifts in climate (temperature and rainfall) may alter the recruitment of an Amazonian palm by modifying pathogen activity. Lastly,

Bachelot et al. (2016) showed more generally that the richness and evenness of soil fungi (not restricted to pathogens) was correlated with biotic factors (tree diversity and basal area, litter biomass) in relatively undisturbed forest, but did not correlate with those factors in disturbed habitats. There is therefore strong evidence that pathogens are vulnerable to disturbance. As fungi play essential roles in tropical forests including the probable maintenance of plant diversity (Wright 2002, Mangan et al. 2010, Bagchi et al. 2014), remnant tropical forest communities in disturbed tropical ecosystems (Keenan et al. 2015) may have uncertain futures.

Given that germination and survival of the species other than S. rubicundum were independent of fragment size, the wider implications of our results for tree species coexistence at the community level remain unclear. It is important to note, however, that Syzygium rubicundum was the most abundant species in the seedling assemblage at our study site. That mortality of this dominant tree species was reduced in soil from smaller fragments presents a critical question – to what extent can the responses of abundant species to fragmentation have cascading effects on the rest of the plant community? If Syzygium rubicundum and other abundant species are released from fungal control, their dominance of the community may increase within small fragments, potentially decreasing plant biodiversity over time.

3.5.2 The effects of seed mass on plant-fungus interactions

Large seeds perform better than small seeds (within species and between species) under a range of stressors (Leishman et al. 2000) like soil dryness (Leishman and Westoby 1994b), low light

(Leishman and Westoby 1994a, Paz and MartÍnez-Ramos 2003) and simulated herbivory

(Bonfil 1998, Green and Juniper 2004, Khan 2004), but the role of seed size in resisting fungal attack had not been previously investigated. Westoby et al. (1996) suggested that large seed size could confer an advantage to a young plant in two ways. First, large seeds may produce larger seedlings with greater vigour, ensuring they have better access to water and light through longer roots and shoots, and thereby conferring greater tolerance to resource limitation. In support of this theory, Wulff (1986) found that large seeds of Desmodium paniculatum produced larger seedlings with greater vigour, which were in turn more likely to survive in dry habitats. The findings of Paz and MartÍnez-Ramos (2003) for six species of Psychotria were similar, except that the advantage manifested in low light environments. Second, large seeds may have greater food reserves that can better sustain seedlings that are damaged (by herbivores, pathogens or environmental conditions) primarily when cotyledons are present (the reserve effect). The results of Bonfil (1998) and Khan (2004) support this theory. Bonfil (1998) found that seedlings from large seeds of Quercus rugosa were better equipped to handle simulated herbivory only when their cotyledons were retained. Khan (2004) showed that seedlings of Artocarpus heterophyllus (also present at the study site) from larger seeds were more resistant to simulated herbivory due to increased vigour. Both mechanisms, therefore, have empirical support and may operate synergistically in young plants.

Given the evidence that large seed size increases resistance and tolerance of various abiotic and biotic stressors, we had hypothesized it might also confer resistance to fungal pathogens.

We did not find any evidence to support our hypothesis. We cannot reject the hypothesis, however, as we detected the effects of seed mass and pathogens during non-overlapping stages of seedling establishment. We found that the likelihood of seed germination was only influenced by seed mass (and not by fungal pathogens), and that the survival of a seedling after germination was only influenced by pathogens (and not by seed mass). Our finding that the

52 effects of seed size manifest only at the earliest stages of seedling establishment are consistent with the results of several other studies (Leishman et al. 2000). Cideciyan and Malloch (1982) similarly found that seed mass influenced germination but not the subsequent survival of two

Rumex species in England. A few other studies have subsequently found that the effects of seed mass are pronounced during the early stages of seedling establishment but weak during later stages (Winn 1988, Saverimuttu and Westoby 1996, Walters and Reich 2000).

The effects of seed mass, however, have also been detected during the later stages of seedling development for several species that span a range of ecosystems and habitat types. These species include wild radish (Stanton 1984), milkweed (Morse and Schmitt 1985), Desmodium paniculatum (Wulff 1986), Quercus sp. in wet subtropical forests (Tripathi and Khan 1990), panicum virgatum (Zhang and Maun 1991), Quercus sp. in temperate forests (Bonfil 1998),

Psychotria sp. (Paz and MartÍnez-Ramos 2003), Artocarpus heterophyllus (Khan 2004) and

Euterpe edulis (Pizo et al. 2006). As fungal pathogens are known to widely regulate seedling populations in natural ecosystems (Wright 2002), we suggest that our hypothesis is worth investigating for other plant species around the world, and that the consequences of seed mass on resistance to pathogens remains an important question.

3.5.3 Summary

Our results join a growing body of literature indicating that fungal-plant interactions are sensitive to various forms of anthropogenic disturbance including fragmentation (Benitez-

Malvido et al. 1999), edge effects (Benítez-Malvido and Lemus-Albor 2005, Krishnadas and

Comita 2018) and simulated drought (Thompson et al. 2010, Swinfield et al. 2012). These results may represent early warnings that fungal-plant interactions are disrupted in human modified tropical forests. Given the importance of such interactions especially for the maintenance of plant diversity (Wright 2002, Mangan et al. 2010, Bagchi et al. 2014), such

53 disruption could have profound implications for the future of remnant tropical forest communities. The consequences of these altered interactions at the wider community and ecosystem scales, however, have been largely unexplored, but present an important direction for future research.

3.6 Acknowledgements

We thank Kadamane Estate Company and Mr. K.M. Cariappa, Senior Manager. We are grateful to the estate’s residents, Praveen Kumar, Vijay, Ansil Basheer, Vikrant Jathar, Suresh, and Netra for their help. The research was funded by ETH Grant 42 13-1.

3.7 References

Bachelot, B., M. Uriarte, J. K. Zimmerman, J. Thompson, J. W. Leff, A. Asiaii, J. Koshner,

and K. McGuire. 2016. Long-lasting effects of land use history on soil fungal

communities in second-growth tropical rain forests. Ecological Applications 26:1881-

1895.

Bagchi, R., R. E. Gallery, S. Gripenberg, S. J. Gurr, L. Narayan, C. E. Addis, R. P. Freckleton,

and O. T. Lewis. 2014. Pathogens and insect herbivores drive rainforest plant diversity

and composition. Nature 506:85-88.

Bagchi, R., T. Swinfield, R. E. Gallery, O. T. Lewis, S. Gripenberg, L. Narayan, and R. P.

Freckleton. 2010. Testing the Janzen-Connell mechanism: pathogens cause

overcompensating density dependence in a tropical tree. Ecology Letters 13:1262-

1269.

Bell, T., R. P. Freckleton, and O. T. Lewis. 2006. Plant pathogens drive density-dependent

seedling mortality in a tropical tree. Ecology Letters 9:569-574.

Benitez-Malvido, J., G. Garcı́a-Guzmán, and I. D. Kossmann-Ferraz. 1999. Leaf-fungal

incidence and herbivory on tree seedlings in tropical rainforest fragments: an

experimental study. Biological Conservation 91:143-150.

Benítez-Malvido, J., and A. Lemus-Albor. 2005. The Seedling Community of Tropical Rain

Forest Edges and Its Interaction with Herbivores and Pathogens. Biotropica 37:301-

313.

Bonfil, C. 1998. The effects of seed size, cotyledon reserves, and herbivory on seedling survival

and growth in Quercus rugosa and Q. laurina (Fagaceae). American Journal of Botany

85:79.

Cideciyan, M. A., and A. J. C. Malloch. 1982. Effects of Seed Size on the Germination, Growth

and Competitive Ability of Rumex Crispus and Rumex Obtusifolius. Journal of

Ecology 70:227-232.

Comita, L. S., S. A. Queenborough, S. J. Murphy, J. L. Eck, K. Xu, M. Krishnadas, N.

Beckman, and Y. Zhu. 2014. Testing predictions of the Janzen–Connell hypothesis: a

meta-analysis of experimental evidence for distance- and density-dependent seed and

seedling survival. Journal of Ecology 102:845-856.

Connell, J. H. 1971. On the role of natural enemies in preventing competitive exclusion in some

marine animals and in rain forest trees. Pages 298-312 in P. J. den Boer and G.

Gradwell, editors. Dynamics of populations. PUDOC, Wageningen.

Dent, D. H., and S. J. Wright. 2009. The future of tropical species in secondary forests: a

quantitative review. Biological Conservation 142:2833-2843.

Ewers, R. M., and R. K. Didham. 2006. Confounding factors in the detection of species

responses to habitat fragmentation. Biological Reviews 81:117-142.

Fahrig, L. 2017. Ecological Responses to Habitat Fragmentation Per Se. Annual Review of

Ecology, Evolution, and Systematics 48.

Fricke, E. C., J. J. Tewksbury, and H. S. Rogers. 2014. Multiple natural enemies cause distance-

dependent mortality at the seed-to-seedling transition. Ecology Letters 17:593-598.

Galetti, M., R. Guevara, M. C. Côrtes, R. Fadini, S. Von Matter, A. B. Leite, F. Labecca, T.

Ribeiro, C. S. Carvalho, R. G. Collevatti, M. M. Pires, P. R. Guimarães, P. H.

Brancalion, M. C. Ribeiro, and P. Jordano. 2013. Functional Extinction of Birds Drives

Rapid Evolutionary Changes in Seed Size. Science 340:1086-1090.

Green, P. T., and P. A. Juniper. 2004. Seed mass, seedling herbivory and the reserve effect in

tropical rainforest seedlings. Functional Ecology 18:539-547.

Haddad, N. M., L. A. Brudvig, J. Clobert, K. F. Davies, A. Gonzalez, R. D. Holt, T. E. Lovejoy,

J. O. Sexton, M. P. Austin, C. D. Collins, W. M. Cook, E. I. Damschen, R. M. Ewers,

B. L. Foster, C. N. Jenkins, A. J. King, W. F. Laurance, D. J. Levey, C. R. Margules,

B. A. Melbourne, A. O. Nicholls, J. L. Orrock, D.-X. Song, and J. R. Townshend. 2015.

Habitat fragmentation and its lasting impact on Earth’s ecosystems. Science Advances

Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests. American

Naturalist 104:501-528.

Keenan, R. J., G. A. Reams, F. Achard, J. V. de Freitas, A. Grainger, and E. Lindquist. 2015.

Dynamics of global forest area: Results from the FAO Global Forest Resources

Assessment 2015. Forest Ecology and Management 352:9-20.

Khan, M. L. 2004. Effects of seed mass on seedling success in Artocarpus heterophyllus L., a

tropical tree species of north-east India. Acta Oecologica 25:103-110.

Krishnadas, M., and L. S. Comita. 2018. Influence of soil pathogens on early regeneration

success of tropical trees varies between forest edge and interior. Oecologia 186:259-

268.

Laurance, W. F., J. L. Camargo, R. C. Luizão, S. G. Laurance, S. L. Pimm, E. M. Bruna, P. C.

Stouffer, G. B. Williamson, J. Benítez-Malvido, and H. L. Vasconcelos. 2011. The fate

of Amazonian forest fragments: a 32-year investigation. Biological Conservation

144:56-67.

Leishman, M. R., and M. Westoby. 1994a. The Role of Large Seed Size in Shaded Conditions:

Experimental Evidence. Functional Ecology 8:205-214.

Leishman, M. R., and M. Westoby. 1994b. The Role of Seed Size in Seedling Establishment

in Dry Soil Conditions -- Experimental Evidence from Semi-Arid Species. Journal of

Ecology 82:249-258.

Leishman, M. R., I. J. Wright, A. T. Moles, and M. Westoby. 2000. The evolutionary ecology

of seed size. Seeds: the ecology of regeneration in plant communities 2:31-57.

Mangan, S. A., S. A. Schnitzer, E. A. Herre, K. M. L. Mack, M. C. Valencia, E. I. Sanchez,

and J. D. Bever. 2010. Negative plant-soil feedback predicts tree-species relative

abundance in a tropical forest. Nature 466:752-755.

Morse, D. H., and J. Schmitt. 1985. Propagule size, dispersal ability, and seedling performance

in Asclepias syriaca. Oecologia 67:372-379.

Paz, H., and M. MartÍnez-Ramos. 2003. Seed Mass and Seedling Performance Within Eight

Species of Psychotria (Rubiaceae). Ecology 84:439-450.

Pizo, M. A., C. Von Allmen, and L. P. C. Morellato. 2006. Seed size variation in the palm

Euterpe edulis and the effects of seed predators on germination and seedling survival.

Acta Oecologica 29:311-315.

R Core Team. 2017. R: A language and environment for statistical computing. R Foundation

for Statistical Computing, Vienna, Austria.

Saverimuttu, T., and M. Westoby. 1996. Seedling Longevity under Deep Shade in Relation to

Seed Size. Journal of Ecology 84:681-689.

Stanton, M. L. 1984. Seed Variation in Wild Radish: Effect of Seed Size on Components of

Seedling and Adult Fitness. Ecology 65:1105-1112.

Swinfield, T., O. T. Lewis, R. Bagchi, and R. P. Freckleton. 2012. Consequences of changing

rainfall for fungal pathogen-induced mortality in tropical tree seedlings. Ecology and

Evolution 2:1408-1413.

Therneau, T. M. 2015. Package “coxme” Mixed Effects Cox Models.

Thompson, S., P. Alvarez‐Loayza, J. Terborgh, and G. Katul. 2010. The effects of plant

pathogens on tree recruitment in the Western Amazon under a projected future climate:

a dynamical systems analysis. Journal of Ecology 98:1434-1446.

Tripathi, R. S., and M. L. Khan. 1990. Effects of Seed Weight and Microsite Characteristics

on Germination and Seedling Fitness in Two Species of Quercus in a Subtropical Wet

Hill Forest. Oikos 57:289-296.

Walters, M. B., and P. B. Reich. 2000. Seed Size, Nitrogen Supply, and Growth Rate Affect

Tree Seedling Survival in Deep Shade. Ecology 81:1887-1901.

Westoby, M., M. Leishman, J. Lord, H. Poorter, and D. J. Schoen. 1996. Comparative Ecology

of Seed Size and Dispersal [and Discussion]. Philosophical Transactions of the Royal

Society of London. Series B: Biological Sciences 351:1309-1318.

Winn, A. A. 1988. Ecological and Evolutionary Consequences of Seed Size in Prunella

Vulgaris. Ecology 69:1537-1544.

Wright, J. S. 2002. Plant diversity in tropical forests: a review of mechanisms of species

coexistence. Oecologia 130:1-14.

Wulff, R. D. 1986. Seed Size Variation in Desmodium Paniculatum: II. Effects on Seedling

Growth and Physiological Performance. Journal of Ecology 74:99-114.

Zhang, J., and M. A. Maun. 1991. Establishment and Growth of Panicum virgatum L. Seedlings

on a Lake Erie Sand Dune. Bulletin of the Torrey Botanical Club 118:141-153.

3.8 Supplementary information

3.8.1 Survival of Hopea canarensis

We collected 1060 Hopea canarensis seeds, 64% of which had been preyed upon by insects.

We eventually planted 384 live seeds of the species (mean seed mass = 0.41) as part of the main experiment but none of them germinated.

3.8.2 Data exclusion and model validation

Three trays that became waterlogged during the experiment were excluded from all analyses.

We checked that the random effects of the mixed effects Cox proportional hazards models were normally distributed. Specific random effects were included if models converged, and variances were non-zero (>10-4). We also validated the proportional hazards assumption of coxme by confirming that the Schoenfeld residuals were statistically independent of time. We estimated the ‘dfbeta’ values for each data point to identify the points that were the most influential in the estimation of each parameter. All outliers were then excluded from the analyses, and results were compared to the main analyses to ensure that the outliers did not disproportionately influence parameter estimates. We included all outliers in the final analyses.

S Table 3.1: Species used in the study and their dispersal mode, number of parent trees, sample size, percentage of pre-dispersal insect predation, average mass of seeds, and the monitoring period.

Species Disperal Trees N % pred Seed mass Planted Period

Symplocos racemosa Animal 6 630 0 Y 0.16 Apr 14 102

Syzygium cumini Animal 1 585 0 Y 0.24 Apr 14 102

Syzygium rubicundum Animal 6 630 9 Y 0.31 May 23 111

Syzygium gardneri Animal 1 504 40 Y 0.37 Apr 14 102

Olea dioica Animal 5 630 14 Y 0.18 May 23 140

Toona ciliata Wind 1 504 0 N NA Apr 22 67

We fitted both models (used in the survival analyses in the main section) using generalized mixed effects models (all species, S Table 3.2) with the glmer function from the package lme4

(Bates et al. 2015) in R 3.4.0 to additionally verify results of the mixed effects Cox proportional hazards analyses (all species, S Table 3.2). Using a fixed and random effects structure like that used in the survival analyses, we used a binomial transformation with a ‘logit’ link to first model overall germination and subsequent mortality (final status in 0s and 1s) across the entire period of observation. We also modelled the germination of each species separately to ensure that overall patterns were not driven by a single species. We then used a binomial transformation with a “complementary log-log” link to model germination and mortality in each of five census intervals that constituted the entire observation period and included the census number as an additional random effect. We contrasted effects in the control and fungicide treatments in both analyses to investigate whether effects in the two treatments were significantly different. We also assumed that interactions between seed mass and fragment size and/or experimental treatment and fragment size varied randomly with fragment identity. Only the five species with known seed masses were used for the glmer analyses. We ensured that random effects were normally distributed and excluded any that had near zero (<10-4) variances. We used the standard diagnostic tools for generalized linear models to check whether all the model assumptions were valid.

The results of each analysis of germination and mortality (S Table 3.2) corresponded with the results of the Cox proportional hazards model, suggesting that our results are robust to decisions about modelling methodology.

3.8.3 The direct effects of fungicides and insecticides on plant growth

Maron et al. (2011) found that the fungicides we used only minimally affected mycorrhizal fungi, and do not directly affect the growth of seedlings.

S Table 3.2: Results of models (mixed effects Cox proportional hazards, generalized linear mixed effects using a ‘logit’ link and generalized linear mixed effects using a ‘cloglog’ link) examining the effects of experimental treatment, fragment size and seed mass on seed germination and seedling mortality.

germination

coxme logit cloglog

N 1746 2619 2619 7552

random effects: N var N var N var

species 5 0.043 5 0.064 5 0.087

census 5 0.395

site (intercept) 21 0.018 21 0.004 21 0.007

site (treatment) 21 0.053 21 0.011

fixed effects: est. se est. se est. se

intercept 0.01 0.21 -3.19 0.33

fungicide -0.18 0.12 -0.11 0.12 -0.10 0.06

fragsize -0.03 0.09 -0.05 0.14 0.02 0.08

fragsize: fungicide 0.03 0.06 0.06 0.06 0.03 0.03

mortality

coxme logit cloglog

N 173 1566 1710 4514

random effects: N var N var N var

species 5 0.316 5 0.796 5 0.213

census 4 0.005

site (intercept) 21 0.038 21 0.132 21 0.127

site (treatment) 21 0.189 21 0.158

fixed effects: est. se est. se est. se

intercept -2.29 0.55 -4.15 0.38

fungicide 0.24 0.42 0.05 0.23 0.10 0.22

fragsize 0.45 0.26 0.32 0.28 0.21 0.26

3.8.4 References

Maron, J. L., M. Marler, J. N. Klironomos, and C. C. Cleveland. 2011. Soil fungal pathogens

and the relationship between plant diversity and productivity. Ecology Letters 14:36-

41.

4 The role of plant-enemy interactions in maintaining tree seedling diversity in a fragmented tropical forest in India

Ashwin Viswanathan*a, Jaboury Ghazoula, Owen T. Lewisb, Ganesh Honwadc, R. Uma

Shaankerd and Robert Bagchie a Chair of Ecosystem Management, ITES, ETH Zurich, Zurich 8092, Switzerland b Department of Zoology, University of Oxford, South Parks Road, Oxford, OX1 3PS, United

Kingdom c Center for Innovation Research and Consultancy, (CIRC) Pune, India d Department of Crop Physiology, University of Agricultural Sciences, G.K.V.K. Campus,

Bangalore 560065, India e Department of Ecology and Evolutionary Biology, University of Connecticut Storrs, CT

06269, USA

* Corresponding author: [email protected]; +919483512541

4.1 Abstract

Natural enemies of plants, including fungal pathogens and insect herbivores, can maintain plant diversity if their harmful effects on seeds and seedlings are density-dependent (the Janzen-

Connell hypothesis). As insect and fungal communities can be modified by anthropogenic habitat fragmentation, we conducted a field experiment to understand how fragmentation might affect the ability of natural enemies to maintain diversity. In twenty-one rainforest fragments in the Western Ghats, India, we suppressed insects and fungi with pesticides, and examined consequent changes in the mortality and diversity of naturally dispersed tree seedlings. Only one plant species, the very abundant Syzygium rubicundum, showed evidence of fine-scale density-dependent interactions with pathogenic fungi. Fungicides had no effect on alpha diversity but decreased the beta-diversity (species turnover among plots) and gamma-diversity

(aggregate species diversity across fragments within size categories), possibly because they disrupted specialized and density-dependent plant-fungal interactions. However, the facilitative effects of fungi on beta diversity weakened as fragments decreased in size. Contrary to our expectations, density-dependent interactions between insects and plants were absent across species in this system. Furthermore, insects appeared to decrease alpha diversity, reinforcing the conclusion that insects did not cause density-dependence. Despite the apparent lack of Janzen-Connell processes mediated by insects, we found that insects facilitated increases in gamma diversity, and in beta diversity between distant plots, simply by suppressing seedlings of the most abundant tree species independent of conspecific density. We found evidence that insects facilitated smaller increases in gamma diversity in the smallest fragments, indicating that fragment size influenced the role of insects in maintaining plant diversity as well. These weak effects in small fragments could be attributed to a shift in composition in those fragments (possibly due to abiotic factors), towards species that were resistant to insects.

Small fragments are viewed as future reservoirs of biodiversity in human-dominated

67 landscapes, but our findings suggest that modified interactions with natural enemies may result in the erosion of this diversity over time.

Keywords: insect herbivores, fungal pathogens, density-dependence, fragment size, Janzen-

Connell, coexistence

4.2 Introduction

Tropical rainforest landscapes are being modified rapidly by humans (Kareiva et al. 2007), with potentially devastating consequences for their exceptional biodiversity. Large tracts of contiguous rainforest are being replaced by fragments of forest isolated within agricultural and urban landscapes (Laurance et al. 2011, Haddad et al. 2015). Although these fragmented forests typically harbour less diversity per unit area than contiguous forests (Turner 1996, Ewers and

Didham 2006, Laurance et al. 2011), they may still act as reservoirs of biological diversity in otherwise transformed landscapes (Wright and Muller-Landau 2006, Dent and Wright 2009,

Anand et al. 2010). The long-term conservation value of forest fragments, however, will depend on the viability of small, isolated populations of plants and animals that they contain, and their capacities to withstand the altered microclimatic conditions (Ewers and Banks-Leite

2013) and increased exposure to human influences typical of fragmented landscapes (Debinski and Holt 2000). Critically, biodiversity within forest fragments will require the continued operation of ecological processes that generate and maintain diversity in intact forests.

Species coexistence mechanisms (Wright 2002) are key processes in this context, but little is known about how forest fragmentation might disrupt them. One important coexistence mechanism, the Janzen-Connell hypothesis, proposes that specialist natural enemies of plants disproportionately attack plant species occurring at high densities, decreasing per-capita survival of abundant host species (negative density-dependence) (Janzen 1970, Connell 1971).

Negative density-dependence is prevalent in tropical forest plant communities, especially at younger life stages (Bagchi et al. 2014, Comita et al. 2014), and several studies have demonstrated its importance in maintaining plant diversity (Webb and Peart 1999, Harms et al.

2000, Metz et al. 2010, Bagchi et al. 2014). Insects and fungal pathogens appear to be the most important agents of negative density-dependence in tropical forests (Mangan et al. 2010,

Bagchi et al. 2014, Fricke et al. 2014), but fragmentation may modify communities of both

69 herbivorous insects and fungi (Benitez-Malvido et al. 1999, Tscharntke and Brandl 2004,

Crockatt 2012, Grilli et al. 2017, Ruete et al. 2017). Altered microclimatic conditions, including increased light and reduced humidity can extend far into forest fragments, which may in turn affect fungal and insect herbivore abundance and community composition, and consequently alter herbivory and disease incidence (Benitez-Malvido et al. 1999, Wirth et al. 2008, Ruiz-

Guerra et al. 2010). For example, fungal pathogens often require high humidity (Weber 1973).

Relatively lower humidity environments in fragmented forests (Ewers and Banks-Leite 2013) could result in reduced fungal abundances and natural enemy attack, that could in turn disrupt the regulation of plant communities and allow dominance by a few, competitive plant species.

Importantly, both theoretical work (Holt et al. 1999, Gravel et al. 2011) and empirical data

(Rossetti et al. 2017, Ruete et al. 2017, Bagchi et al. 2018) indicate that communities of herbivores and fungal pathogens become less specialized when forests are fragmented. As the strength of the Janzen-Connell mechanism increases with host-specificity of natural enemies

(Sedio and Ostling 2013), the effectiveness of the mechanism for coexistence could be reduced in forest fragments.

The loss of specialized density-dependent plant-enemy interactions in forest fragments are likely to lead to a decrease in alpha diversity over time. The modification of plant-enemy interactions in fragmented landscapes, however, may additionally manifest as reduced variation in top-down control within and among small fragments (Rooney et al. 2004). This may partly explain the shift towards more homogenized ecological communities observed in fragmented forests (Lôbo et al. 2011, Arroyo‐Rodríguez et al. 2013), and may also explain the consistent reduction of beta diversity observed in such landscapes (Arroyo‐Rodríguez et al.

2013). Although abiotic shifts that favour a subset of the plant species pool may explain these patterns to a large extent, biotic mechanisms may play important roles. The roles of natural

70 enemies in regulating beta and gamma diversity in fragmented landscapes, however, have not been studied previously.

In this study, we sought to understand the mechanistic links between plant enemies, seedling mortality, woody plant diversity and forest fragmentation. We asked the following questions:

1) Do insects and fungal pathogens increase mortality of woody plant seedlings, and do their effects increase with conspecific seedling density? 2) Does fragmentation influence woody seedling survival and the strength of density-dependent mortality that is mediated by insects and/or fungi? 3) Do insects and/or fungi increase diversity of plants (seedlings) at local or regional scales? 4) Does fragmentation influence the roles of insects and fungi in maintaining plant diversity? We hypothesized that insect herbivores and fungal pathogens facilitate plant species coexistence and maintain plant diversity through density-dependent interactions. We also hypothesized that fragmentation will weaken density-dependent mortality caused by insects and/or fungi, which may in turn lead to a slow decay of diversity in fragmented forests.

4.3 Methods

4.3.1 Study area

We conducted the study in Kadamane Tea Estate (12.8639 – 12.9389 N and 75.6361 - 75.6833

E) located on the western slopes of the Western Ghats in Karnataka, India. The Western Ghats region (combined with Sri Lanka) is considered one of the world’s “Hottest Biodiversity

Hotspots” (Myers et al. 2000) and harbours over 4,700 species of flowering plants (Myers et al. 2000) and an estimated 13,000 – 25,000 monophagous herbivorous insect species (Fonseca

2009). A study in the region identified 226 species of pathogenic fungi from a sample of 639 tree species (Mohanan and Yesodharan 2005). Kadamane is a 30 km2 private estate that spans elevations of 800 – 1300 m and contains fragments of evergreen forest interspersed within cultivated tea and natural grassland. Fragments in the estate, although isolated, are in all cases

<100 m from other fragments and contiguous forest. They have been isolated for roughly 50 years (clearing for tea fields and roads primarily occurred in the late 1960s and early 1970s).

Rainforest in Kadumane is dominated by trees and shrubs from the families Lauraceae,

Myrtaceae, Euphorbiaceae, Annonaceae, Elaeocarpaceae, Symplococaceae, Ebenaceae,

Meliaceae, Sapindaceae, Myristicaceae and Moraceae. Plants from these families are primarily bird dispersed. Woody plants in this ecosystem typically fruit, either annually or supra- annually, at the end of the dry season. Seeds begin to germinate from the onset of monsoon.

The monsoon is also characterized by particularly high insect and fungal activity (Gadgil and

Kumar 2006) and is therefore the period when insect herbivores and fungal pathogens are most likely to influence seedling survival, and perhaps plant diversity in general.

4.3.2 Study design

Our basic experimental unit consisted of five plots of 1 m2 (initially cleared of all existing young vegetation, Nunits = 111, Nplots = 555) to monitor seedlings. At 37 locations across the study site (ranging in elevation from 918 – 1071 m), we set up three experimental units separated from each other by ca. 20 m and located at least 50 m from the forest edge to reduce edge effects. As the study was conducted in 21 forest fragments ranging in size from 1 – 149 ha (mean 27.8 ha) (Fig. 4.1), we determined the number of replicated locations (three experimental units) per fragment based on its area, so that similar proportional areas were sampled within each. We therefore located replicates at either one (fragment area < 10 ha), two

(10 – 40 ha), three (>50 ha) or four (in the 149-ha site) randomly selected locations each fragment (N = 37).

Fig. 4.1: Map showing the location of the study site, layout of the fragments, the specific locations within, and the experimental design. Forest fragments are separated by cultivated tea, grassland and roads. Adapted from a map created by the Nature Conservation Foundation (NCF).

The five seedling plots were randomly assigned to the following experimental treatments: 1) insect suppression, 2) fungal suppression, 3) both insect and fungal suppression, 4) water addition control, and 5) no-manipulation control. We suppressed insects by treating plots with a systemic neonicotinoid insecticide, Thiamethoxam (Product name: Actara®). We suppressed fungal pathogens by treating plots with a combination of a systemic fungicide Azoxystrobin

(Product name: Amistar®), and a second fungicide that combines a contact fungicide,

Mancozeb and a systemic oomyticide, Metalxyl (Product name: Ridomil Gold®). All three products are manufactured by Syngenta (Basel, Switzerland).

The study was conducted from June 2015 to December 2016. We began the insecticide and fungicide treatments in July 2015 and ended them in November 2016. Treatments were applied every 20 days during the dry season and every 8 days during the wet season (June – October).

They were applied more frequently in the wet season because heavy rain is likely to wash off pesticides sooner. To every insect suppression plot, we added 0.1 g of Actara dissolved in 50 ml of water (according to manufacturer specifications). To every fungal suppression plot, we applied 0.02 ml of Amistar dissolved in 50 ml of water and 0.3 g of Ridomil Gold dissolved in

50 ml of water (according to manufacturer specifications). We added 150 ml of water in each combined suppression plot and ensured that all plots received the same amount of water except the no-manipulation control plots, which were completely unmanipulated.

We documented the species identity of each seedling in each plot during four censuses conducted in October-November 2015, April 2016, June-July 2016 and October 2016. During each census, we tagged each seedling that had established since the previous census with an aluminium tag with a unique identification code, to follow the fate of each seedling and to detect any establishment of seedlings during a subsequent census. We conducted a final census in November-December 2016 when we recorded the fate of tagged seedlings in all our plots.

4.3.3 Data analysis

We recorded 7404 newly germinated seedlings during the study period, of which 7134 seedlings (12.85 seedlings per m2) of 106 species germinated during the 2016 monsoon (June-

July and October censuses). We restricted all analyses to data from the last two censuses because they represented 96% of all seedlings recorded.

To test our hypothesis that insects and pathogens cause density-dependent mortality of seedlings, we modelled seedling mortality in each 1 m2 quadrat as a function of experimental treatment, local conspecific seedling density (defined as the number of conspecifics co- occurring in the same 1 m2 plot) and their interaction. To test our hypothesis that fragment size affected the strength of interactions between plants and insects and/or fungi, we incorporated fragment size, and interactions between fragment size and pesticide application, as explanatory variables in the same models. We additionally investigated the effects of heterospecific densities in each analysis. We assumed binomial error distributions and used Generalized

Mixed Effects Models to account for the dependence structure of the data (see below).

To test our hypotheses that insects and/or fungi maintain plant (seedling) diversity, and that the size of a fragment influences the amount of diversity they maintain, we examined changes in diversity between germination and establishment at three different spatial scales. We first modelled the change in alpha diversity (1 m2, Shannon Diversity Index) in each plot as a function of experimental treatment, fragment size, and their interactions. We then modelled the change in pairwise beta diversity (1 m2, Bray-Curtis dissimilarity) between each pair of plots in each experimental treatment (across all fragments) as a function of treatment, pairwise distance between plots, the mean size of the fragment pair, pairwise difference in fragment size, and their interactions. Finally, we grouped the fragments into sets of small fragments (<10 ha, N = 33 plots/treatment), medium fragments (10-60 ha, N = 39 plots/treatment) and large

75 fragments (>60 ha, N = 39 plots/treatment) for comparisons of gamma diversity (20 m2,

Shannon Diversity Index; see supplementary information). We calculated the change in gamma diversity for 300 samples of 20 plots in each of these size classes and modelled it as a function of experimental treatment. We assumed Gaussian error distributions in all analyses of diversity

(see supplementary information for model diagnostics).

In all analyses, we initially modelled the mortality of all species, and then examined within- species (and/or within species groups) patterns to see if they were consistent across species and/or groups of species. We fitted these models using (generalized) linear mixed-effects models using the lme4 package (Bates et al. 2015), or using linear ordinary least squares model

(for the diversity analyses), in R v3.4.0 (R Core Team 2017). All models included nested random intercepts for seedling plot, experimental unit, and location to control for the dependence structure of the data where appropriate. When we combined data from multiple species, we also included random effects allowing the mean mortality (random intercept) and effect of conspecific density on mortality (random slope) to vary among species. We used parametric bootstrapping with 1000 iterations (implemented with the bootMer function in the lme4 R package) to obtain confidence intervals for parameter estimates and predictions from the models.

4.4 Results

4.4.1 Seedling mortality

Seedling mortality increased significantly with increasing conspecific seedling density in the analysis with all species (Table 4.1, Fig. 4.2). However, when species were analysed separately, mortality of only two species, Syzygium rubicundum and Ventilago maderaspatana, increased with conspecific density.

Analysis of all species together showed a significant decrease in overall seedling mortality in plots treated with insecticide (main effect of insecticide, Table 4.1, Fig. 4.2). When analysed separately, however, only Syzygium rubicundum and Spatholobus purpureus showed significantly decreased mortality in insecticide-treated plots but Symplocos racemosa did not.

The application of insecticide did not influence density-dependent effects, or interact with fragment size, in any analysis.

Analysis of all species combined indicated that seedling mortality was not influenced by the fungicide treatment (Table 4.1, Fig. 4.2). Fungicide only decreased the density-dependent effects on the mortality of Syzygium rubicundum significantly (Table 4.1, Fig. 4.2) but did not influence seedling mortality at low densities, or interact with fragment size, in any analysis.

4.4.2 Alpha diversity

Alpha diversity (at the 1 m2 scale) tended to decrease between germination and establishment

(Table 4.2, Fig. 4.3). When the entire seedling assemblage was considered, changes in alpha diversity were not influenced by the application of either insecticide or fungicide (Table 4.2,

Fig. 4.3). When analyses focused on a reduced species assemblage, excluding seedlings of

Syzygium rubicundum and Spatholobus purpureus, decreases in alpha diversity were significantly greater in the control plots than in plots treated with insecticide. No significant effects of fungicide treatment and no interactions with fragment size were detected.

Table 4.1: Results of binomial generalized linear mixed effects models investigating the influence of experimental treatment and fragment size on the mortalities of different species or groups of species (1 m2 plots). The effects of the insecticide and fungicide treatments are contrasted with effects in the control treatments, and per-capita mortality is weighted by the number of seedlings.

Syzygium Spatholobus Symplocos Ventilago other species rare species all species rubicundum purpureus racemosa maderaspatana (≥ 5 plots) (≤ 5 plots ) maximum density 300 300 96 77 54 23 8

N plot instances / seedlings 1095 / 6298 172 / 3598 76 / 578 121 / 621 64 / 313 564 / 1029 97 / 122 percent mortality 65 75 81 41 45 46 36 random effects: N var N var N var N var N var N var N var plot 491 0.33 172 0.32 76 0.04 121 0.11 64 373 0.05 32 group 111 0.19 61 0.22 30 0.06 53 0.17 27 0.14 109 0.29 57 location 37 0.44 30 0.62 15 24 0.28 16 0.91 37 0.47 85 species (intercept) 98 1.09 26 1.46 68 species (density) 98 0.02 26 0.10 68 fixed effects: est. se est. se est. se est. se est. se est. se est. se control -1.20 0.26 -1.05 0.34 -0.23 0.30 -0.35 0.30 -1.14 0.49 -1.24 0.39 -1.23 0.32 insecticide -0.17 0.08 -0.40 0.15 -0.51 0.22 -0.04 0.25 -0.01 0.14 -0.17 0.17 -0.04 0.21 fungicide -0.01 0.08 0.12 0.14 -0.09 0.22 0.07 0.25 -0.23 0.16 -0.07 0.17 -0.18 0.22 density: control 0.42 0.10 0.54 0.10 0.35 0.35 -0.43 0.51 0.28 0.14 0.12 0.13 density: insecticide 0.02 0.03 0.06 0.10 0.18 0.34 0.40 0.5 0.00 0.08 density: fungicide -0.05 0.03 -0.28 0.09 0.06 0.34 -0.42 0.5 -0.07 0.09 fragsize: control 0.14 0.19 0.42 0.32 0.16 0.25 fragsize: insecticide -0.06 0.08 0.08 0.14 0.09 0.15 fragsize: fungicide -0.04 0.08 0.12 0.13 -0.16 0.14 toposlope 0.24 0.10 0.44 0.15 0.50 0.21 -0.21 0.19 -0.37 0.36 0.21 0.17 0.36 0.29 multiplicative effects 0.02 0.08 0.06 0.14 -0.18 0.22 -0.05 0.25 0.22 0.15 -0.02 0.17 0.03 0.21 density: multiplicative effects 0.02 0.03 -0.04 0.09 0.22 0.34 0.54 0.51 0.07 0.08 fragsize: multiplicative effects -0.01 0.08 0.05 0.13 -0.02 0.15

Fig. 4.2: Probabilities of seedling mortality (from germination to establishment) plotted against densities of conspecifics (1 m2 plots) in each experimental treatment. Data points are standardized by estimated random effects. Lines are predicted from generalized binomial mixed effects models fitted to individual or groups of species, separately (Table 4.1), and are plotted with standard errors.

Table 4.2: Alpha diversity: Results of linear mixed effects models investigating the influence of experimental treatment and fragment size on changes in alpha diversity (Shannon’s Index) of seedling assemblages from germination to establishment (1 m2 plots). Beta diversity: Results of linear mixed effects models investigating the influence of experimental treatment, geographic distance between plots, mean pairwise fragment size, and the difference in pairwise fragment size, on changes in pairwise dissimilarity (standardized Bray-Curtis dissimilarities of seedling assemblages from germination to establishment in 1 m2 plots). The effects of the insecticide and fungicide treatments are contrasted with effects in the control treatments.

alpha diversity beta diversity

without without all species Syz. rubicundum all species Syz.rubicundum Spa. purpureus Spa.purpureus

N 352 285 6910 4034 random effects: N Var N Var N Var N Var location | plot1 37 0.005 37 0.003 461 0.019 401 0.024 site | group 1 21 0.009 21 0.011 108 0.002 107 0.003 plot2 451 0.016 391 0.018 group 2 109 0.002 108 0.005 fixed effects: est. se est. se est. se est. se control 0.027 0.06 -0.033 0.07 0.181 0.018 0.257 0.025 insecticide 0.020 0.03 0.070 0.03 0.019 0.014 0.034 0.017 fungicide -0.001 0.03 -0.019 0.03 0.004 0.014 0.026 0.018 distance: control 0.022 0.006 0.015 0.009 distance: insecticide -0.015 0.006 -0.033 0.008 distance: fungicide -0.007 0.006 -0.019 0.009 fragsize: control 0.010 0.05 -0.006 0.05 0.026 0.015 -0.002 0.021 fragsize: insecticide -0.003 0.03 -0.021 0.03 -0.008 0.013 -0.006 0.017 fragsize: fungicide 0.017 0.03 0.041 0.03 -0.006 0.013 -0.049 0.018 fragsize: difference in fragsize: control -0.007 0.004 -0.001 0.007 fragsize: difference in fragsize: insecticide 0.002 0.004 0.004 0.006 fragsize: difference in fragsize: fungicide 0.001 0.004 0.011 0.007 initial diversity -0.365 0.05 -0.360 0.07 -0.196 0.013 -0.272 0.022 slope 0.002 0.03 -0.014 0.03

Fig. 4.3: Changes in alpha diversity (Shannon index) of seedling assemblages (from germination to establishment) plotted against initial diversities of seedlings (1 m2 plots) in each experimental treatment. Data points are standardized by estimated random effects. Lines are predicted from linear mixed effects models (Table 4.2) and are plotted with standard errors.

4.4.3 Beta diversity

Pairwise dissimilarity between 1 m2 plots (beta diversity) tended to be higher for established seedlings than at the germination stage (Table 4.2, Fig. 4.4). The increase in in beta diversity at establishment was significantly greater for plots separated by greater geographical distances.

However, this effect was removed when plots were treated with either insecticide or fungicide.

For plots that were closest together (< 1 km), greater increases in beta diversity were observed in plots treated with insecticide than in control plots (Table 4.2, Fig. 4.4).

Pairs of plots from fragments of greater mean area showed greater increases in beta diversity from germination to establishment (Table 4.2, Fig. 4.4). When the entire seedling assemblage was considered in analysis, treatment with neither insecticide nor fungicide altered this effect significantly. Excluding Syzygium rubicundum and Spatholobus purpureus from the analysis, the effect of fragment size was removed when plots were treated with fungicide

4.4.4 Gamma diversity

Gamma diversity (aggregate species diversity across fragments within size categories) tended to be higher at establishment than at germination in small (<10 ha) and large (>60 ha) fragments but decreased in medium fragments (10-60 ha) that had very high initial diversities and were relatively unaffected by pesticide treatments (Fig. 4.5).

Fig. 4.4: Changes in the pairwise Bray-Curtis dissimilarities (beta diversities) between seedling assemblages (from germination to establishment) plotted against the distance between plots, and against mean fragment size (from each pair of 1 m2 plots), in each experimental treatment. Points (standardized by the estimated random effects) equal mean changes in dissimilarity in each of seven binned distance or fragment size classes and are plotted with standard errors. Lines are predicted from linear mixed effects models (Table 2) and are plotted with standard errors.

Large fragments experienced the greatest increases in gamma diversity over time. The increases in gamma diversity were removed when plots were treated with insecticide (Fig. 4.5).

The nature of these trends was similar, but much weaker, in small fragments. These patterns were not observed when Syzygium rubicundum and Spatholobus purpureus were excluded from the analysis.

Fungicide application affected changes in gamma diversity in all three size-classes of fragments in a similar manner to insecticide application (Fig. 4.5). However, all impacts of fungicide were weaker than those induced by insecticide. Again, these patterns were not observed in large and small fragments when analyses excluded Syzygium rubicundum and

Spatholobus purpureus.

Seedling compositions varied in the three fragment size classes (S Table 4.2). Syzygium rubicundum and Spatholobus purpureus were the two most abundant species, with 2275 and

550 seedlings respectively in large fragments (>60 ha). However, Syzygium rubicundum and

Symplocos racemosa were most abundant in small fragments (10-60 ha), with 722 and 518 seedlings respectively. Only 39 Spatholobus purpureus seedlings germinated in small fragments.

4.5 Discussion

4.5.1 The Janzen-Connell hypothesis

Including all species in the seedling assemblage, we found that seedling mortality was density- dependent. This result appears consistent with current understanding of how negative density- dependence shapes tropical plant communities in the tropics (Wright 2002, Bagchi et al. 2014,

Comita et al. 2014, Bachelot et al. 2017), but such inference could be misleading as individual species analyses revealed that the assemblage-wide pattern was heavily influenced by one

84 species, Syzygium rubicundum. Additionally, the Janzen-Connell hypothesis posits that negative density-dependent processes increase alpha diversity in plant communities over time

(Wright 2002). Alpha diversity did not increase over time in our control plots, especially when

Syzygium rubicundum was excluded from analyses, providing further evidence that density- dependent mortality played a limited role in this system, especially with regard to tree species coexistence.

It is possible that density-dependent effects manifested only at very high seedling densities.

Syzygium rubicundum may therefore have been the only species present at high enough densities in sufficient plots to detect density-dependent mortality in analyses of individual species. It is also possible that density-dependent processes play out before the period between seedling germination and establishment that we examined. The strongest evidence for density- dependence comes from studies that have considered plant mortality over the seed-seedling transition (Harms et al. 2000, Bagchi et al. 2014) (also Krishnadas et al. in review), but several studies have also identified density-dependence in older seedlings (Webb and Peart 1999,

Comita et al. 2010). The effects of conspecific density may also vary among years (Uriarte et al. 2018) and we may have found evidence of density-dependent mortality in more species had our data spanned several years. Despite these caveats, our results suggest that density- dependent patterns in a limited number of species can drive cross-assemblage averages, even when species-level effects have been accounted for using mixed-effects models. Therefore, we recommend that, in addition to analyses of data pooled across species, future studies examine within-species trends, preferably over multiple ontogenetic stages and years.

Fig. 4.5: Changes in spatially aggregated Shannon diversities (gamma diversities) for 300 aggregated seedling assemblages (from germination to establishment) in each of small (<10 ha), medium (10-60 ha) and large (>60 ha) fragments. The width of a violin plot indicates the concentration of data points.

4.5.2 The role of insects

Insects caused noticeable seedling mortality of only two abundant species, Syzygium rubicundum and Spatholobus purpureus, and their effects were independent of conspecific densities. Other studies (e.g. Gripenberg et al. 2014 and Bagchi et al. 2014) have also found evidence that insects suppress seed and seedling populations independent of fine-scale local densities. Our findings reinforce the narrative that insects do not generally suppress rainforest seedlings in a density-dependent manner (Comita et al. 2014) although there are exceptions where insects have been implicated as the cause of density-dependent seedling and sapling mortality (Sullivan 2003, Fricke et al. 2014). Furthermore, it seems likely that the spatial scale at which seedling density is measured and manipulated in this and other studies may not reflect the scale at which mobile insects respond to the density of plant resources (Gripenberg et al.

2014).

Given the absence of evidence for insect-mediated density-dependence in this system, it is unsurprising that excluding insects did not reduce alpha diversity of the seedling assemblage.

Instead, we found evidence that the alpha diversity of the seedling assemblage (excluding the abundant species Syzygium rubicundum and Spatholobus purpureus) was in fact lowered by insects. One possible explanation is that generalist species that often target less common plants

(Bachelot et al. 2016) were suppressing plants other than Syzygium rubicundum and

Spatholobus purpureus. Since the effects of generalist species on plants will tend to be independent of plant conspecific densities, they are likely to homogenize plant species composition at local spatial scales by removing rare species, thereby decreasing seedling diversity.

Despite the lack of evidence that insects increased plant diversity at fine spatial scales, our results suggest that insects enhanced dissimilarity between more distant plots, and hence

87 increased beta diversity. This result held even when Syzygium rubicundum and Spatholobus purpureus were excluded from analysis. A plausible explanation is that spatial variation in insect herbivory generates heterogeneity in plant species recruitment across intermediate spatial scales. Insects also contributed to over 70% of the observed increases in gamma diversity. These effects on gamma diversity disappeared when Syzygium rubicundum and

Spatholobus purpureus were excluded from the analysis. As generalist insect herbivores tend to specialize on common plant species (Bachelot et al. 2016), and hence disproportionately feed on those species regardless of conspecific density, they could reduce the accumulated dominance of Syzygium rubicundum and Spatholobus purpureus in the wider community.

4.5.3 The role of fungi

We only found evidence that fungal pathogens caused density-dependent mortality for one species, S. rubicundum, which was also the most abundant species in our dataset (S Table 4.2).

We did not find evidence that fungi caused the mortality of any species (or group of species) at low densities, although another study from the region did find such evidence, albeit with artificially manipulated seed densities (Krishnadas and Comita 2018). Although the effects of fungi were limited to a single species, our results add to a growing body of literature that implicates fungal pathogens as drivers of density-dependent mortality in the early life stages of plants (Bell et al. 2006, Fricke et al. 2014). We had hypothesized that fungal pathogens maintain alpha diversity as observed in Bagchi et al. (2014), but our expectations hinged on the premise that negative density-dependent interactions between fungi and plants were widespread across the woody-plant community (Wright 2002). As the effects of fungi were limited (statistically) only to plots with high densities of Syzygium rubicundum, thus invalidating the premise, it is unsurprising that we found no evidence that fungi drive increases in alpha diversity during seedling establishment.

Our results instead show that fungi facilitated increases in both beta diversity and gamma diversity, although their effects on beta diversity (pairwise dissimilarity between plots) were only significant when analyses excluded Syzygium rubicundum and Spatholobus purpureus.

This indicates the presence of some specialized, and potentially density-dependent, interactions between fungi and common plants (aside from the interaction with Syzygium rubicundum) which could not be detected statistically in the analyses of individual species. It is possible that spatial variation in fungal communities led to heterogeneity in suitability for plant species, i.e. by suppressing different sets of plant species in different locations. The increases in gamma diversity could be attributed entirely to the effects of fungi on Syzygium rubicundum, indicating that even when limited to a single abundant plant species, density-dependent plant-fungal interactions can increase plant diversity at the community level. Fungi, however, accounted for less than 30% of the total increase in gamma diversity observed in the control plots, with insects accounting for over 70% of the increase. A possible explanation is that a hyper-abundant species like Syzygium rubicundum can potentially accumulate at low local densities to dominate the seedling assemblage at large spatial scales, despite strong local density-dependent effects.

4.5.4 The effects of fragment size on biotic interactions and plant diversity

Our results suggest that small (<10 ha), medium (10-60 ha) and large fragments (>60 ha) should have similar increases in gamma diversity, as insects primarily facilitated increases in diversity by suppressing seedlings of Syzygium rubicundum and Spatholobus purpureus independent of fragment size. Yet we found higher plant species diversity in large fragments than in small fragments (medium fragments had high initial diversities and consequently low increases in diversity). This inconsistency is explained by examining the compositions of the seedling communities in small and large fragments and noting that both Syzygium rubicundum and

Spatholobus purpureus were most abundant in large fragments (see S Table 4.2). In small

89 fragments, fewer Syzygium rubicundum germinated although they remained the most abundant seedling species, while Symplocos racemosa replaced Spatholobus purpureus as the second most abundant species, possibly due to abiotic factors. However, unlike Spatholobus purpureus, mortality of Symplocos racemosa seedlings was unaffected by insecticide addition, and herbivorous insects therefore facilitated greater increases in the gamma diversity of large fragments than of small fragments.

We also found evidence that fungal pathogens contributed less to beta diversity in smaller rainforest fragments than large fragments, suggesting that the long-term stability of currently diverse fragments may be at risk. A reduced role for fungi in maintaining beta diversity of small fragments may indicate that certain plant-fungus interactions have been lost or weakened in these fragments. However, we found no direct evidence that the density-dependent effects of fungi were affected by fragmentation, although there is some evidence that plant-fungus interactions are modified in soils from forest edges (Krishnadas and Comita 2018). The small number of high-density plots may have limited our ability to detect the effects of fragment size on the density-dependent interactions between fungi and Syzygium rubicundum (and other species).

4.5.5 Conclusions

A focus on the Janzen-Connell hypothesis as a mechanism for maintaining diversity, and associated density-dependent processes in tropical forests, has primarily led to investigations of the effects of plant natural enemies on alpha diversity (Comita et al. 2014). Very few studies have investigated the effects of natural enemies on beta and gamma diversity, but at least one study proposes that insect herbivores maintain beta diversity (Lamarre et al. 2012). We show that insect herbivores and fungal pathogens primarily maintain beta and gamma diversity rather than alpha diversity in a rainforest in India.

The effects of fragment size on negative interactions between plants and other organisms have only rarely been investigated in tropical forests, despite the relevance of this understanding for the long-term conservation of biodiversity. We demonstrate that fragment size does influence the diversity-maintaining roles of the natural enemies of plants. We show that the facilitative roles of herbivorous insects on plants are compromised in small fragments, due to a shift in the plant composition towards a dominant species, Symplocos racemosa. We also find results consistent with fragmentation leading to the breakdown of plant-fungus interactions, thus influencing the ability of a system to retain beta diversity. Such modifications of diversity- maintaining mechanisms within smaller fragments could have important long-term implications for the number of species that they can support.

4.6 Acknowledgements

We are especially grateful to Praveen Kumar and his family for help with logistics and data collection. Kadamane Estate provided us with an ideal set up for our study, a place to stay and people to help. We thank Mr. K.M. Cariappa, Senior Manager, Kadamane Estates Company, for his personal interest in the study. We are grateful to Vijay, Suresh, Netra, Kanjimalai and his family, Bhanu Sridharan, Ajith Ashokan, Akshay Surendra, Dayani Chakravarthy, Ansil

Basheer, Arun Kumar, Vikrant Jathar, Santhosh Uthappa, Vinod Shanker, Bhuvanesh, and

Nisarg Prakash for field assistance. The research was funded by ETH Grant 42 13-1.

4.7 References

Anand, M. O., J. Krishnaswamy, A. Kumar, and A. Bali. 2010. Sustaining biodiversity

conservation in human-modified landscapes in the Western Ghats: Remnant forests

matter. Biological Conservation 143:2363-2374.

Arroyo‐Rodríguez, V., M. Rös, F. Escobar, F. P. L. Melo, B. A. Santos, M. Tabarelli, and R.

Chazdon. 2013. Plant β‐diversity in fragmented rain forests: testing floristic

homogenization and differentiation hypotheses. Journal of Ecology 101:1449-1458.

Bachelot, B., M. Uriarte, K. L. McGuire, J. Thompson, and J. Zimmerman. 2017. Arbuscular

mycorrhizal fungal diversity and natural enemies promote coexistence of tropical tree

species. Ecology 98:712-720.

Bachelot, B., M. Uríarte, J. Thompson, and J. K. Zimmerman. 2016. The advantage of the

extremes: tree seedlings at intermediate abundance in a tropical forest have the highest

richness of above-ground enemies and suffer the most damage. Journal of Ecology

104:90-103.

Bagchi, R., L. M. Brown, C. S. Elphick, D. L. Wagner, and M. S. Singer. 2018. Anthropogenic

fragmentation of landscapes: mechanisms for eroding the specificity of plant–herbivore

interactions. Oecologia.

Bagchi, R., R. E. Gallery, S. Gripenberg, S. J. Gurr, L. Narayan, C. E. Addis, R. P. Freckleton,

and O. T. Lewis. 2014. Pathogens and insect herbivores drive rainforest plant diversity

and composition. Nature 506:85-88.

Bates, D., M. Maechler, B. Bolker, and S. Walker. 2015. Fitting Linear Mixed-Effects Models

Using lme4. Journal of Statistical Software 67:1-48.

Bell, T., R. P. Freckleton, and O. T. Lewis. 2006. Plant pathogens drive density-dependent

seedling mortality in a tropical tree. Ecology Letters 9:569-574.

Benitez-Malvido, J., G. Garcı́a-Guzmán, and I. D. Kossmann-Ferraz. 1999. Leaf-fungal

incidence and herbivory on tree seedlings in tropical rainforest fragments: an

experimental study. Biological Conservation 91:143-150.

Comita, L. S., H. C. Muller-Landau, S. Aguilar, and S. P. Hubbell. 2010. Asymmetric Density

Dependence Shapes Species Abundances in a Tropical Tree Community. Science

329:330-332.

Comita, L. S., S. A. Queenborough, S. J. Murphy, J. L. Eck, K. Xu, M. Krishnadas, N.

Beckman, and Y. Zhu. 2014. Testing predictions of the Janzen–Connell hypothesis: a

meta-analysis of experimental evidence for distance- and density-dependent seed and

seedling survival. Journal of Ecology 102:845-856.

Connell, J. H. 1971. On the role of natural enemies in preventing competitive exclusion in some

marine animals and in rain forest trees. Pages 298-312 in P. J. den Boer and G.

Gradwell, editors. Dynamics of populations. PUDOC, Wageningen.

Crockatt, M. E. 2012. Are there edge effects on forest fungi and if so do they matter? Fungal

Biology Reviews 26:94-101.

Debinski, D. M., and R. D. Holt. 2000. A Survey and Overview of Habitat Fragmentation

Experiments. Conservation Biology 14:342-355.

Dent, D. H., and S. J. Wright. 2009. The future of tropical species in secondary forests: a

quantitative review. Biological Conservation 142:2833-2843.

Ewers, R. M., and C. Banks-Leite. 2013. Fragmentation Impairs the Microclimate Buffering

Effect of Tropical Forests. PLoS ONE 8:e58093.

Ewers, R. M., and R. K. Didham. 2006. Confounding factors in the detection of species

responses to habitat fragmentation. Biological Reviews 81:117-142.

Fonseca, C. R. 2009. The Silent Mass Extinction of Insect Herbivores in Biodiversity Hotspots.

Conservation Biology 23:1507-1515.

Fricke, E. C., J. J. Tewksbury, and H. S. Rogers. 2014. Multiple natural enemies cause distance-

dependent mortality at the seed-to-seedling transition. Ecology Letters 17:593-598.

Gadgil, S., and K. R. Kumar. 2006. The Asian monsoon—agriculture and economy. Pages 651-

683 The Asian Monsoon. Springer.

Gravel, D., F. Massol, E. Canard, D. Mouillot, and N. Mouquet. 2011. Trophic theory of island

biogeography. Ecology Letters 14:1010-1016.

Grilli, G., S. Longo, P. Y. Huais, M. Pereyra, E. G. Verga, C. Urcelay, and L. Galetto. 2017.

Fungal diversity at fragmented landscapes: synthesis and future perspectives. Current

Opinion in Microbiology 37:161-165.

Gripenberg, S., R. Bagchi, R. E. Gallery, R. P. Freckleton, L. Narayan, and O. T. Lewis. 2014.

Testing for enemy-mediated density-dependence in the mortality of seedlings: field

experiments with five Neotropical tree species. Oikos 123:185-193.

Haddad, N. M., L. A. Brudvig, J. Clobert, K. F. Davies, A. Gonzalez, R. D. Holt, T. E. Lovejoy,

J. O. Sexton, M. P. Austin, C. D. Collins, W. M. Cook, E. I. Damschen, R. M. Ewers,

B. L. Foster, C. N. Jenkins, A. J. King, W. F. Laurance, D. J. Levey, C. R. Margules,

B. A. Melbourne, A. O. Nicholls, J. L. Orrock, D.-X. Song, and J. R. Townshend. 2015.

Habitat fragmentation and its lasting impact on Earth’s ecosystems. Science Advances

Harms, K. E., S. J. Wright, O. Calderon, A. Hernandez, and E. A. Herre. 2000. Pervasive

density-dependent recruitment enhances seedling diversity in a tropical forest. Nature

404:493-495.

Holt, R. D., J. H. Lawton, G. A. Polis, and N. D. Martinez. 1999. Trophic Rank and the Species-

Area Relationship. Ecology 80:1495-1504.

Janzen, D. H. 1970. Herbivores and the number of tree species in tropical forests. American

Naturalist 104:501-528.

Kareiva, P., S. Watts, R. McDonald, and T. Boucher. 2007. Domesticated Nature: Shaping

Landscapes and Ecosystems for Human Welfare. Science 316:1866-1869.

Krishnadas, M., and L. S. Comita. 2018. Influence of soil pathogens on early regeneration

success of tropical trees varies between forest edge and interior. Oecologia 186:259-

268.

Lamarre, G. P. A., C. Baraloto, C. Fortunel, N. Dávila, I. Mesones, J. G. Rios, M. Ríos, E.

Valderrama, M. V. Pilco, and P. V. A. Fine. 2012. Herbivory, growth rates, and habitat

specialization in tropical tree lineages: implications for Amazonian beta-diversity.

Ecology 93:S195-S210.

Laurance, W. F., J. L. Camargo, R. C. Luizão, S. G. Laurance, S. L. Pimm, E. M. Bruna, P. C.

Stouffer, G. B. Williamson, J. Benítez-Malvido, and H. L. Vasconcelos. 2011. The fate

of Amazonian forest fragments: a 32-year investigation. Biological Conservation

144:56-67.

Lôbo, D., T. Leão, F. P. L. Melo, A. M. M. Santos, and M. Tabarelli. 2011. Forest

fragmentation drives Atlantic forest of northeastern Brazil to biotic homogenization.

Diversity and Distributions 17:287-296.

Mangan, S. A., S. A. Schnitzer, E. A. Herre, K. M. L. Mack, M. C. Valencia, E. I. Sanchez,

and J. D. Bever. 2010. Negative plant-soil feedback predicts tree-species relative

abundance in a tropical forest. Nature 466:752-755.

Metz, M. R., W. P. Sousa, and R. Valencia. 2010. Widespread density-dependent seedling

mortality promotes species coexistence in a highly diverse Amazonian rain forest.

Ecology 91:3675-3685.

Mohanan, C., and K. Yesodharan. 2005. Biodiversity of plant pathogenic fungi in the Kerala

part of the Western Ghats. Final report of the project No. KFRI 375.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. Da Fonseca, and J. Kent. 2000.

Biodiversity hotspots for conservation priorities. Nature 403:853-858.

R Core Team. 2017. R: A language and environment for statistical computing. R Foundation

for Statistical Computing, Vienna, Austria.

Rooney, T. P., S. M. Wiegmann, D. A. Rogers, and D. M. Waller. 2004. Biotic Impoverishment

and Homogenization in Unfragmented Forest Understory Communities. Conservation

Biology 18:787-798.

Rossetti, M. R., T. Tscharntke, R. Aguilar, and P. Batáry. 2017. Responses of insect herbivores

and herbivory to habitat fragmentation: a hierarchical meta-analysis. Ecology Letters

20:264-272.

Ruete, A., T. Snäll, B. G. Jonsson, M. Jönsson, and M. Nuñez. 2017. Contrasting long‐term

effects of transient anthropogenic edges and forest fragment size on generalist and

specialist deadwood‐dwelling fungi. Journal of Applied Ecology 54:1142-1151.

Ruiz-Guerra, B., R. Guevara, N. A. Mariano, and R. Dirzo. 2010. Insect herbivory declines

with forest fragmentation and covaries with plant regeneration mode: evidence from a

Mexican tropical rain forest. Oikos 119:317-325.

Sedio, B. E., and A. M. Ostling. 2013. How specialised must natural enemies be to facilitate

coexistence among plants? Ecology Letters 16:995-1003.

Sullivan, J. J. 2003. Density-dependent shoot-borer herbivory increases the age of first

reproduction and mortality of neotropical tree saplings. Oecologia 136:96-106.

Tscharntke, T., and R. Brandl. 2004. Plant-insect interactions in fragmented landscapes. Annu

Rev Entomol 49:405-430.

Turner, I. M. 1996. Species Loss in Fragments of Tropical Rain Forest: A Review of the

Evidence. Journal of Applied Ecology 33:200-209.

Uriarte, M., R. Muscarella, and J. K. Zimmerman. 2018. Environmental heterogeneity and

biotic interactions mediate climate impacts on tropical forest regeneration. Global

Change Biology 24:e692-e704.

Webb, C. O., and D. R. Peart. 1999. Seedling density dependence promotes coexistence of

Bornean rain forest trees. Ecology 80:2006-2017.

Weber, G. F. 1973. Bacterial and fungal diseases of plants in the tropics. University of Florida

Press.

Wirth, R., S. T. Meyer, I. R. Leal, and M. Tabarelli. 2008. Plant Herbivore Interactions at the

Forest Edge. Pages 423-448 in U. Lüttge, W. Beyschlag, and J. Murata, editors.

Progress in Botany. Springer Berlin Heidelberg, Berlin, Heidelberg.

Wright, J. S. 2002. Plant diversity in tropical forests: a review of mechanisms of species

coexistence. Oecologia 130:1-14.

Wright, S. J., and H. C. Muller-Landau. 2006. The uncertain future of tropical forest species.

Biotropica 38:287-301.

4.8 Supplementary information

4.8.1 The effects of abiotic drivers

We investigated the effects of the abiotic factors like canopy cover, substrate and steepness

(topographical slope) on per-capita mortality in each analysis. We estimated the canopy cover

(proportion of filled pixels) above each plot from a photograph of the canopy taken at ground level (at the centre of the plot) with an 18mm lens. We divided each plot into a 5 x 4 grid and counted the proportion of cells with any rocky/gravelly substrate to quantify the ‘rockiness’ of the substrate. We also estimated the slope/steepness of each plot visually. We found that canopy cover and substrate did not influence the mortality of seedlings of any species.

Increasing topographical slope (toposlope) of a plot increased the mortality of Syzygium rubicundum seedlings (Table 1, main section) and Spatholobus purpureus seedlings (Table

4.1, main section).

4.8.2 The effects of water

Using the same linear mixed effects model framework as in the main section, we contrasted mortality in the watered control plots with unwatered control plots. The use of water in the experiments did not significantly affect the mortality of seedlings (S Table 4.1).

4.8.3 Synergies between the actions of insects and fungi

In the mortality analyses, we also examined whether mortality in the combined fungicide and insecticide treatment showed additive or multiplicative effects of fungus and insect exclusion

(i.e. if there was an interaction between the effects of fungicide and insecticide on mortality).

There was no evidence of an interaction between the effects of fungicide and insecticide (Table

4.1), suggesting their effects were additive and not multiplicative.

S Table 4.1: Results of a binomial generalized linear mixed effects model that includes only the control plots. The effects of withholding water are contrasted with effects in the watered control treatments. The covariate ‘con- density’ refers to scaled conspecific density (scales L-R 18.64, 43.85, 18.17, 10.87, 9.95, 2.85, 1.53), ‘het-density’ to scaled heterospecific density (scales L-R 27.82, 7.99, 35.36, 42.02, 10.38, 28.73, 25.02), ‘abundance’ to scaled final abundance (scales L-R 532.48, NA, NA, NA, NA, 52.54, 2.39), ‘fragsize’ (not considered here) to scaled fragment size (scales L-R: 70.28, 80.95, 75.52, 62.39, 64.53, 70.73, 56.77, used in Table 1) and ‘toposlope’ to scaled topographical slope (scale L-R 0.13, 0.13, 0.13, 0.12, 0.13, 0.12, 0.13). All scaling factors are the same as those used in Table 1.

Syzygium Spatholobus Symplocos all species rubicundum purpureus racemosa

random effects: N var N var N var N var

plot 196 0.04 68 0.13 28 51

group 109 0.24 43 0.40 21 38 0.32

location 37 0.55 27 0.53 12 18 0.56

species 64 1.22

fixed effects: est. se est. se est. se est. se

watered control -0.80 0.28 -0.50 0.31 0.23 0.34 -0.33 0.43

control without water 0.14 0.08 0.21 0.19 0.15 0.25 0.29 0.26

con-density: watered control 0.38 0.05 0.94 0.21 0.36 0.09 0.07 0.27

con-density: control without water -0.05 0.04 -0.17 0.19 0.03 0.08 -0.01 0.17

het-density 0.10 0.09 0.26 0.127 0.27 0.28 -0.12 0.28

abundance 0.13 0.46

toposlope 0.17 0.14 0.06 0.24 0.10 0.53 -0.26 0.36

Ventilago other species rare species maderaspatana (≥ 5 plots) (≤ 5 plots )

random effects: N var N var N var

plot 24 152 32

group 14 2.00 101 28

location 9 37 0.42 22 3.06

species (intercept) 26 1.13 34

species (density) 26 0.13 34

fixed effects: est. se est. se est. se

watered control -1.41 0.98 -0.26 0.48 -0.10 1.1

control without water 0.09 0.54 -0.14 0.2 0.65 0.65

con-density: watered control 1.66 1.12 0.25 0.22

con-density: control without water -0.44 0.66 0.16 0.18

het-density -1.11 1.37 0.19 0.14 1.51 0.94

abundance -0.85 0.53 -2.10 1.22

toposlope 0.22 0.72 0.05 0.25 -1.02 1.55

100

4.8.4 Identification of spatial scale to measure gamma diversity

Different processes, like the interactions of fungi or insects with plants, may influence the accumulation of diversity across spatial scales at varying rates. To investigate how insects and pathogens influenced accumulation of diversity across spatial scales, we pooled data from each experimental treatment in groups of 1 to 111 randomly selected plots in increments of one (1

– 111 m2). We calculated the alpha diversity (Shannon) for 1000 samples at each spatial scale

(from 1m2 in increments of 1m2 up to 111 m2) for each experimental treatment. We plotted this as a function of increasing spatial scale to estimate the minimum spatial scale at which measures of diversity in each of the five experimental treatments neared asymptotes and were therefore near independent of spatial scale. Diversity of seedlings increased asymptotically with increasing spatial scale and plateaued at the scale of 15-20 m2 (or 15-20 plots) in all treatments. Given that diversity did not increase substantially with sampling beyond this scale in any treatment, we identified 20 m2 as the appropriate scale to compare gamma diversity in the five treatments.

4.8.5 Validating our statistical models

We validated our models first by ensuring that all predictors were well represented across their range in each data set and excluded any that did not satisfy this criterion. We excluded interactions between fragment size and conspecific density wherever data were insufficient, or these variables were correlated. In all models, we ensured that transformed residuals were evenly distributed about zero and not skewed in either direction. We also checked the linearity of the quantile-quantile plot for each fit. We ensured that no random effect had a zero variance

(<10-3) and that each random effect was approximately normally distributed. We used leverage plots to check whether any data points had disproportionately large effects on model results and we only report results that are resistant to the exclusion of the most influential data points.

101

We also checked the visual fit of each model, and its robustness by excluding certain species and extreme data points from the analysis. We included all outliers in the final analyses except those that we specifically mention in the tables with justification.

S Table 4.2: Results of linear models investigating the influence of experimental treatment on changes in the aggregated gamma diversities (Shannon) of seedling assemblages (across 20 plots) from germination to establishment (aggregations over 300 samples of 20 plots of 1m2 for each fragment size grouping). We excluded Syzygium cumini from these analyses, as it was present in extremely high densities in only three treatments and biased estimates. The effects of the insecticide and fungicide treatments are contrasted with effects in the control treatments.

all species

small (<10 ha) medium (10-60 ha) large (>60 ha)

fixed effects: est. se est. se est. se

control 0.382 0.024 0.042 0.018 0.640 0.022

insecticide -0.068 0.004 -0.045 0.003 -0.157 0.006

fungicide -0.044 0.003 0.053 0.004 -0.062 0.006

initial diversity -0.201 0.015 -0.093 0.009 -0.260 0.017

without Syzygium rubicundum and Spatholobus purpureus

small (<10 ha) medium (10-60 ha) large (>60 ha)

fixed effects: est. se est. se est. se

control -0.058 0.012 0.667 0.031 0.532 0.057

insecticide 0.002 0.003 0.025 0.004 -0.030 0.004

fungicide 0.019 0.003 -0.045 0.004 0.008 0.005

initial diversity -0.055 0.007 -0.363 0.013 -0.283 0.023

4.8.6 The direct effects of fungicides and insecticides on plant growth

There is some evidence that the fungicides only have minimal effects on mycorrhizal fungi

(Maron et al. 2011). Maron et al. (2011) also showed that these fungicides do not affect plant growth and survival, except through their action on pests. Although Thiamethoxam has been found to increase the vigour of certain crop species, especially in relation to environmental stressors like draught (for example in Almeida et al. (2014)), there is no evidence that it affects

102 seedling survival. These pesticides have been widely used in other ecological studies examining effects of fungi and insects (Bell et al. 2006, Bagchi et al. 2014, Gripenberg et al.

2014, Gaviria and Engelbrecht 2015).

4.8.7 Does conspecific density-dependence correlate with seedling abundance?

We did not find any significant relationship between seedling mortality (density-dependent and density-independent) and sapling abundance across all plots (S Table 4.1). We found instead that only the most common species (Syzygium rubicundum) experienced density-dependent mortality whereas the less common species did not. We also found that the two most common species (Syzygium rubicundum and Spatholobus purpureus) were more than twice as likely to die as the rare species in our study system (75% and 81% against 33%). These results are concordant with the findings of Bachelot et al. (2017) and Bagchi et al. (2014) who found that common species suffered greater density-dependent mortality than rare species in a Neotropical forest. Our results, however, contrast with those of two other studies that were also conducted in Neotropical forests (Comita et al. 2010, Mangan et al. 2010), where rare species suffered greater density-dependent mortality than common species. Such patterns may be caused by a suppressed immune response to pathogens in rare species, as low population size leads to lower

R-gene diversity and consequently weaker immune capacity (Marden et al. 2017). The abundance of a species, therefore, may appear to be related to the degree of density-dependence it experiences, and this phenomenon has been offered as an explanation for why certain species remain rare. Given the disagreement in the direction of responses between studies, however, this question requires further investigation. The implications of these relationships are important because if common species suffer greater density-dependence (e.g. this study,

Bachelot et al. 2017; Bagchi et al. 2014), it could provide a more powerful force for plant species coexistence in tropical forests. Should the explanation of Comita et al. (2010) and

(Mangan et al. 2010) be found to be generally applicable, it would be possible that the

103 continued erosion of immunity in rare species can eventually lead to their extirpation, and to fewer species rich plant communities.

4.8.8 References

Almeida, A. d. S., C. Deuner, C. T. Borges, G. E. Meneghello, A. Jauer, and F. A. Villela.

2014. Treatment of rice seeds with thiamethoxam: reflections on physiological

performance. Journal of Seed Science 36:392-398.

Bachelot, B., M. Uriarte, K. L. McGuire, J. Thompson, and J. Zimmerman. 2017. Arbuscular

mycorrhizal fungal diversity and natural enemies promote coexistence of tropical tree

species. Ecology 98:712-720.

Bagchi, R., R. E. Gallery, S. Gripenberg, S. J. Gurr, L. Narayan, C. E. Addis, R. P. Freckleton,

and O. T. Lewis. 2014. Pathogens and insect herbivores drive rainforest plant diversity

and composition. Nature 506:85-88.

Bell, T., R. P. Freckleton, and O. T. Lewis. 2006. Plant pathogens drive density-dependent

seedling mortality in a tropical tree. Ecology Letters 9:569-574.

Comita, L. S., H. C. Muller-Landau, S. Aguilar, and S. P. Hubbell. 2010. Asymmetric Density

Dependence Shapes Species Abundances in a Tropical Tree Community. Science

329:330-332.

Gaviria, J., and B. M. J. Engelbrecht. 2015. Effects of Drought, Pest Pressure and Light

Availability on Seedling Establishment and Growth: Their Role for Distribution of Tree

Species across a Tropical Rainfall Gradient. PLoS ONE 10:e0143955.

Gripenberg, S., R. Bagchi, R. E. Gallery, R. P. Freckleton, L. Narayan, and O. T. Lewis. 2014.

Testing for enemy-mediated density-dependence in the mortality of seedlings: field

experiments with five Neotropical tree species. Oikos 123:185-193.

104

Mangan, S. A., S. A. Schnitzer, E. A. Herre, K. M. L. Mack, M. C. Valencia, E. I. Sanchez,

and J. D. Bever. 2010. Negative plant-soil feedback predicts tree-species relative

abundance in a tropical forest. Nature 466:752-755.

Marden, J. H., S. A. Mangan, M. P. Peterson, E. Wafula, H. W. Fescemyer, J. P. Der, C. W.

dePamphilis, and L. S. Comita. 2017. Ecological genomics of tropical trees: how local

population size and allelic diversity of resistance genes relate to immune responses,

cosusceptibility to pathogens, and negative density dependence. Molecular ecology

26:2498-2513.

Maron, J. L., M. Marler, J. N. Klironomos, and C. C. Cleveland. 2011. Soil fungal pathogens

and the relationship between plant diversity and productivity. Ecology Letters 14:36-

41.

105

106

5 General conclusions

In this section, I briefly bring together the overarching insights and inferences from the thesis.

5.1 Do insects and fungi maintain woody plant diversity?

We found strong evidence that both insects and fungi maintain woody plant seedling diversity in a rainforest in India. Both insects and fungi, however, primarily maintained the beta and gamma diversities of seedling assemblages, not their alpha diversities as previous studies had led us to expect. Our results also showed that processes that do not adhere to the conventional

Janzen-Connell framework (plant-insect interactions) could still maintain plant diversity, but not necessarily facilitate coexistence. Some of the differences between our study, and past work in the Neotropics may be due to differences in the evolutionary histories of the plant communities in both regions, as well as differing co-evolutionary strategies with their natural enemies. Local context and site-specific edaphic and environmental conditions may also be causal factors for the differing patterns.

5.2 Are the diversity-maintaining roles of enemies affected by fragment size?

We provide evidence for the first time that the diversity-maintaining roles of both insects and fungi might be susceptible to forest fragmentation. Although the results of this study require further validation in other fragmented ecosystems before we reach any general conclusions, the implications of our results can be significant both for diversity, and for conservation practices globally. Our findings indicate that the small forest fragments are likely to retain even lesser diversity than is predicted by the Theory of Island Biogeography. Additionally, we also found that small fragments are likely to be more similar to each other due to decreased fungal action, than are large fragments. Our results have implications for the debate about whether several small fragments can serve the same purpose as a few large fragments (SLOSS). I suggest, from

107 this thesis, that diversity may be better preserved in large forest fragments, and that ensuring connectivity between smaller fragments while facilitating increases in their sizes, may be important for the long-term conservation of tropical biodiversity.

5.3 Can interspecific differences in community analyses be homogenized?

Finally, a common insight that emerged from all our analyses was that the results of community-wide analyses of mortality/survival could be misleading. Even the use of random intercepts and random slopes did not account for interspecific differences, leading to ‘coerced’ trends entirely driven by just one or two plant species in every case. In the third chapter, despite similar sample size across many species, three different types of aggregated analyses suggested that the probability of germination increased with increasing seed mass across all species (S

Table 3.2). Similarly, three other analyses suggested that fragment size generally interacted with the effects of fungicide on all species (S Table 3.2). We found that both these sets of analyses were misleading although none of the models violated any general assumptions.

Analyses of species separately (where possible) showed that only two out of five species were significantly influenced by seed mass, and only one species (out of six) was significantly influenced by fragment size. We discuss a similar statistical phenomenon in the fourth chapter where ostensible community-wide patterns were driven by only one or two plant species.

Overall, I believe that several such studies must be conducted to understand both, the relative contributions of various processes in maintaining plant diversity in the global tropics, and the impacts of the anthropogenic activities on these reservoirs of biodiversity.

108

109

Species lists

Significant observations are highlighted. (S) – signs, not seen by team, anecdotal; (?) – presence uncertain

Birds https://ebird.org/india/hotspot/L4187687

Amphibians

28 species of frogs and toads but no caecilians. Known common names are in brackets.

name notes

Fejervarya kudremukhensis

Fejervarya granosa

Fejervarya mudduraja

Duttaphrynus melanostictus (Common Indian Toad) Quite abundant

Ghatophryne ornata (Malabar Torrent Toad) Rare

Pedostibes tuberculosus (Malabar Tree Toad) Uncommon

Clinotarsus curtipes Very common

Euphlyctis cyanophlyctis (Indian Skittering Frog) Regular in ponds

Indosylvirana intermedius

Indosylvirana indica (Indian Golden-backed Frog)

Indosylvirana montana

Indirana semipalmata

Hoplobatrachus crassus Only seen in one patch

Ramanella triangularis Very common

Ramanella mormorata

Raorchestes tuberohumerus Very common

Raorchestes lutoelus (Blue-eyed Yellow Bush Frog) Common

Raorchestes glandulosus (Glandular Bush Frog) Not uncommon

Raorchestes ochlandrae (Ochlandra-reed Bush Frog) Not uncommon

Raorchestes ponmudi (Ponmudi Bush Frog) Not uncommon

Raorchestes charius (Chari’s Bush Frog) In the grasslands

110

name notes

Pseudophilautus wynaadensis (Wynaad Bush Frog) Very common

Polypedates occidentalis Very common

Rhacophorus malabaricus (Malabar Gliding Frog) Very common

Rhacophorus lateralis Quite local, uncommon

Nyctibatrachus sylvaticus

Nyctibatrachus kempholeyensis

Micrixalus kottigeharensis/saxicola Very common. Perhaps both species.

111

Reptiles

23 species of snakes. Scientific names are in brackets. Among lizards, Calotes ellioti and Varanus bengalensis were common, Draco dussumieri was rare. [** records from Coorg]

name notes

Brahminy Worm Snake (Ramphotyphlops braminus )

Elliot's Shieldtail (Uropeltis ellioti ) Common in monsoon

Uropeltis sp. Several unidentified

Wayanad Shieldtail (Melanophidium wynaudense ) Rare

Indian Rock Python (Python molurus ) (S) Apparently present but infrequent in the grasslands

Common Trinket (Coelognathus helena helena ) (?)

Montane Trinket (Coelognathus helena monticollaris ) Rare

Indian Rat Snake (Ptyas mucosa ) Common

Painted Bronzeback Tree Snake (Dendrelaphis pictus ) Common, only a little less regular than Common Vine

Travancore Wolf Snake (Lycodon travancoricus ) Not uncommon

Common Wolf Snake (Lycodon aulicus ) Not uncommon

Checkered Keelback (Xenochrophis piscator ) Common in ponds/habitation

Striped Keelback (Amphiesma stolatum ) Uncommon

Beddome's Keelback (Amphiesma beddomei ) Very common

Hill Keelback (Amphiesma monticola ) (?)

Olive Keelback (Atretium schistosum ) Uncommon

Ceylon/Beddome's Cat Snake (Boiga ceylonensis/beddomei ) Common, did not examine in hand

Common Vine Snake (Ahaetulla nasuta ) Very common

Common Krait (Bungarus caeruleus ) Uncommon

Striped Coral Snake (Calliophis nigrescens ) Very rare

Spectacled Cobra (Naja naja ) Near habitation

King Cobra (Ophiophagus hannah ) Uncommon

Malabar Pit Viper (Trimeresurus malabaricus ) Common

** Brachyophidium sp. Several dead individuals in Hudikeri, Coorg. Wall's?

** Bibron's Coral Snake (Calliophis bibroni ) Near Mapilethodu, Coorg

112

Mammals (except bats)

37 mammal species. Scientific names are in brackets.

name notes

Pygmy Shrew (Suncus etruscus ) Grasslands, trapped

House Shrew (Suncus murinus ) At camp, also trapped

Nilgiri Shrew (Suncus niger ) Grasslands, trapped

Slender Loris (Loris lydekkerianus ) Quite abundant, vocal end of monsoon

Bonnet Macaque (Macaca radiata ) Rather shy

Black-footed Grey Langur (Semnopithecus hypoleucos ) Rare, very shy

Nilgiri Langur (Trachypithecus johnii ) (?) Probable sightings, no photo documentation

Dhole (Cuon alpinus ) Uncommonly seen but frequent signs of presence

Sloth Bear (Melursus ursinus ) (S) (?) Possible signs/droppings, anecdotal

Nilgiri Marten (Martes gwatkinsii ) Several from Apr-Jul, one in winter, also with a cub in June

Oriental Small-clawed Otter (Amblonyx cinereus ) (S) Frequent signs of presence

Brown Palm Civet (Paradoxurus jerdoni ) Not uncommon

Small Indian Civet (Viverricula indica ) Seen only once

Grey Mongoose (Herpestes edwardsii ) Common in tea and forest edges

Brown Mongoose (Herpestes fuscus ) One beautiful individual, bushier tail than further south

Striped-necked Mongoose (Herpestes vitticollis ) Uncommon

Leopard (Panthera pardus ) (S) One or two individuals usually around

Tiger (Panthera tigris ) (S) (?) Probable signs

Leopard Cat (Prionailurus bengalensis ) Not uncommon

Asiatic Elephant (Elephas maximus) Not uncommon

Wild Pig (Sus scrofa ) Rare, but inevitable presence at fruit/seed flushes

Indian Chevrotain (Moschiola indica ) Common

Sambar (Cervus unicolor ) Not uncommon

Indian Muntjac (Muntiacus muntjak ) Not uncommon

Gaur (Bos gaurus ) Common

Indian Pangolin (Manis crassicaudata ) (S) (?) Anecdotal

113

name notes

Jungle Striped Squirrel (Funambulus tristriatus ) Not uncommon

Dusky-striped Squirrel (Funambulus sublineatus ) Uncommon, often seen associated with mixed bird flocks

Malabar Giant Squirrel (Ratufa indica ) Very common

Indian Giant Flying Squirrel (Petaurista philippensis ) Common

Travancore Flying Squirrel (Petinomys fuscocapillus ) Rare, only seen once

Indian Field Mouse (Mus booduga ) Common, trapped primarily in grasslands

House Mouse (Mus musculus ) Common in habitation

Sahyadris Forest Rat (Rattus satarae ) Common, trapped

Common House Rat (Rattus rattus ) Common in habitation

Nilgiri Long-tailed Tree Mouse (Vandeleuria nilagirica ) Rare, trapped in evergreen forest

Indian Bush Rat (Golunda ellioti ) Uncommon

Malabar Spiny Dormouse (Platacanthomys lasiurus ) Uncommonly seen but signs everywhere

Indian Porcupine (Hystrix indica ) Uncommon, hunted

Black-naped Hare (Lepus nigricollis) Common in grasslands

114

Butterflies

130 species of butterflies.

common name notes

Pale Green Awlet Rare

Orange Awlet Common early morning in the right season

Common Banded Awl Common

White-Banded Awl

Brown Awl

Indian Awlking Rare

Tamil Spotted Flat

Fulvous Pied Flat

Common Small Flat

Water Snow Flat

Common Snow Flat

Chestnut Angle

Hampson's Hedge Hopper

Tamil Grass Dart

Swift sp.

Chestnut Bob

Gaint Redeye

Common Redeye

Restricted Demon

Narrow-banded Bluebottle Very common

Tailed Jay Common

Common Jay Common

Common Mime Not uncommon

Blue Mormon Very common

Common Mormon Very common

Malabar Raven Uncommon

115

common name notes

Red Helen Very common

Malabar Banded Swallowtail Not uncommon in the right season

Lime Butterfly Not uncommon

Paris Peacock Very common

Malabar Banded Peacock Uncommon but regular in the right season

Malabar Rose Regular in the grasslands

Crimson Rose

Southern Birdwing Common

One-spot Grass Yellow

Common Grass Yellow

Common Emigrant

Yellow Orange-Tip

Great Orange-tip Common

Crimson Tip

Little Orange-tip

Plain Orange-tip

White Orange-tip

Small Salmon Arab

Common Wanderer

Common Albatross

Spot Puffin

Lesser Gull

Common Gull

Painted Sawtooth Common

Common Jezebel

Psyche

Angled Pierrot

Red Pierrot

Common Pierrot

116

common name notes

Indian Sunbeam

Shiva's Sunbeam Rare

Common Acacia Blue

Silver-streaked Acacia Blue Not uncommon

Common Imperial (?) Saw a few 'long-tailed' tits, not certain

Fluffy Tit (?) Saw a few 'long-tailed' tits, not certain

Slate Flash

Common Silverline

Common Cerulean

Forget-Me-Not

Grass Jewel

Tailless Line Blue

Dark Cerulean

Pale Grass Blue

Plain Hedge Blue

Common Lineblue

Lime Blue

Dark Grass Blue

Tiny Grass Blue

Pea Blue

Common Hedge Blue

Club Beak

Double-banded Judy Not uncommon in monsoon

Dark Blue Tiger Common in migration

Plain Tiger Not uncommon

Glassy Tiger Common in migration

Striped Tiger

Blue Tiger Common in migration

Common Crow Common, also in migration

117

common name notes

Double-banded Crow Only in migration

Brown King Crow Only in migration

Malabar Treenymph Uncommon

Common Nawab Only in season

Anomolous Nawab Only in season

Blue Nawab Very rare, only once

Southern Duffer Uncommon, in monsoon

Common Evening Brown Common

Dark Evening Brown Common

Great Evening Brown Uncommon

Common Treebrown

Common Palmfly

Medus Bushbrown

White-Bar Bushbrown

Gladeye Bushbrown Very common

Bushbrown sp.

Tamil Catseye Not uncommon, only in forest

Common Threering

Common Fourring

White Fourring

Tawny Coster

Rustic

Common Leopard Common

Blue Admiral Uncommon, regular in habitat

Commander

Common Sergeant

Black-veined Sergeant

Colour Sergeant

Common Lascar

118

common name notes

Common Sailer

Clipper Uncommon

Common Baron

Red Spot Duke Not uncommon in the monsoon

Grey Count Common

Common Map Common in season

Angled Castor

Common Castor

Chocolate Pansy

Lemon Pansy

Yellow Pansy

Blue Pansy

Grey Pansy

Peacock Pansy

Great Eggfly Common

Danaid Eggfly

Blue Oakleaf Very common

119

Trees

115 woody plant species (except lianas) in the quadrats. Presented below in descending order of stem abundance (T – B, L – R).

species species species species

Psychotria macrocarpa Hopea canarensis Lasianthus sp. Nothapodytes nimmoniana

Psychotria nigra Nothopegia beddomei Macaranga peltata Callicarpa tomentosa

Dimocarpus longan Olea dioica Neolitsea zeylanica Casearia ovata

Syzygium rubicundum Glochidion ellipticum Canthium dicoccum Scolopia crenata

Humboldtia brunonis Caryota urens Dendrocnide sinuata Ficus nervosa

Reinwardtiodendron anamalaiense Chionanthus ramiflorus Ficus exasperata Symplocos cochinchinensis

Blachia umbellata Psychotria sohmeri Goniothalamus sp. Toona ciliata

Litsea floribunda Syzygium cumini Chassalia ophyoxyloides Ardisia solanacea

Ixora sp Meiogyne ramarowii Epiprinus mallotiformis Artocarpus heterophyllus

Memecylon malabaricum Polyalthia fragrans Calophyllum lanceolatum Colebrookea oppositifolia

Dichapetalum gelonioides Aglaia barberi Heritiera papilio Diospyros pruriens

Nothopegia travancorica Diospyros saldanhae Actinodaphne sp. Drypetes venusta

Clerodendrum infortunatum Margaritaria indica Knema attenuata Elaeocarpus tuberculatus

Symplocos racemosa Casearia wynadensis Beilschmiedia sp. Eurya nitida

Agrostistachis indica Maesa indica Persea macrantha Pavetta indica

Litsea mysorensis Garcinia morella Allophylus cobbe Syzygium gardneri

Diospyros nilagirica Gomphandra tetrandra Antidesma montana Apodytes dimidiata

Eugenia aloysii Litsea oleoides Aphanamixis polystachya Celtis tetrandra

Cinnamomum sp. Atalantia sp. Chrysophyllum roxburghii Cleistanthus sp.

Drypetes oblongifolia Mallotus philippensis Hopea parviflora Diospyros assimilis

Actinodaphne bourdillonii Leea indica Mangifera indica Diospyros sylvatica

Cryptocarya sp. Mallotus beddomei Melicope lunu-ankenda Ficus beddomei

Mesua ferrea Trichilia connaroides Microtropis wallichiana Grewia sp.

Syzygium laetum Luvunga sarmentosa Artocarpus gomezianus Litsea stocksii

Holigarna nigra Elaeocarpus serratus Breynia retusa Mallotus tetracoccus

Aglaia eleagnoidea Myristica dactyloides Clausena dentata Oreocnide integrifolia

Garcinia gummi-gutta Dimorphocalyx sp. Diospyros montana Schefflera sp.

Chassalia ophyoxyloides Diospyros candolleana Excoecaria crenulata Spondias pinnata

Palaquium ellipticum Mastixia arborea Litsea insignis

120