Body Size Evolution and Diversity of using the Neotropical () as a Model System

by

Sarah Elizabeth Steele

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Ecology and Evolutionary Biology University of Toronto

© Copyright by Sarah Elizabeth Steele 2018

Body Size Evolution and Diversity of Fishes using the

Neotropical Cichlids (Cichlinae) as a Model System

Sarah Elizabeth Steele

Doctor of Philosophy

Department of Ecology and Evolutionary Biology University of Toronto

2018 Abstract

The influence of body size on an organism’s physiology, morphology, ecology, and life history has been considered one of the most fundamental relationships in ecology and evolution. The ray-finned fishes are a highly diverse group of vertebrates. Yet, our understanding of diversification in this group is incomplete, and the role of body size in creating this diversity is largely unknown. I examined body size in Neotropical cichlids (Cichlinae) to elucidate the large- and small-scale factors affecting body size diversity and distribution, and how body size shapes , morphological, and ecological diversity in fishes. Characterization of body size distributions across the phylogeny of Neotropical cichlids revealed considerable overlap in body size, particularly in intermediate-sized fishes, with few, species-poor lineages exhibiting extreme body size. Three potential peaks of adaptive evolution in body size were identified within Cichlinae. I found freshwater fishes globally tend to be smaller and their distributions more diverse and right-skewed than marine counterparts, irrespective of and clade age, with a strengthening of these trends in riverine systems. Comparisons of Neotropical body size diversity and distribution to this broader context shows that body size patterns are largely abnormal compared to most freshwater fishes, particularly those of the Neotropics.

ii

This implies that small body size is rarer in Cichlinae, despite several independent cases of body size decrease in this lineage. I found that these small-bodied lineages are miniaturized cichlids, exhibiting strong reduction in body size, as well as paedomorphic characters and ontogenetic truncation compared to their sister taxa. Further examination of ontogenies across Neotropical cichlids found considerable shape conservatism over ontogeny and phylogeny, with cichlids following similar ontogenetic trajectories. Therefore, ontogenetic pathways do not contribute considerably to morphological divergence seen in adults, with divergence likely occurring in the larval stage or perhaps during embryology. The evolution of ontogeny, body size, and shape did not correspond to a unique adaptive peak in miniatures, or three body size optima predicted from the distribution of size across the phylogeny. Rather, it is complex and, like early trait divergence previously seen, diversifies early in the phylogeny across the major lineages of

Neotropical cichlids.

iii

Acknowledgments

Firstly, I would like to acknowledge my supervisor Hernán López-Fernández. He introduced me to a field of research I had little exposure to and a world within fishes I became deeply obsessed with exploring. I thank him for our many discussions about body size evolution and the room to expand my research as I developed an approach to tackling my research questions. He provided me with many opportunities in the field, at professional meetings, through workshops, and during our brainstorming sessions, to explore and learn in depth each aspect of my program. I appreciate his guidance and criticism through each stage of my thesis while allowing me to work independently. I have grown considerably as a researcher and thank him for our endless discussions of , evolution, and life. I would like to thank Nathan Lovejoy (University of

Toronto - U of T) and David Evans (U of T) for serving on my supervisory committee and providing guidance and advice throughout my PhD. I would also like to thank Dean Adams

(Iowa State University) and Sebastian Kvist (U of T) for serving as my external defense committee members and the suggestions provided. I would like to thank Helen Rodd (U of T) and D. Luke Mahler (U of T) for serving on my appraisal committee, and providing useful insight that has helped improve my research. I am very grateful for the personal and academic advice and support I have received from Deborah McLennan (U of T), Helen Rodd (U of T),

Luke Mahler (U of T), and Sebastian Kvist (U of T) throughout my academic career. I am indebted to Mary Burridge, Erling Holm, Marg Zur, and Don Stacey for technical support with the Royal Ontario Museum (ROM) collection, outreach opportunities to communicate science to the public, the ability to maintain work on Canadian fishes, as well as the curatorial skills and knowledge I have gained. I extend a sincere thank you to Don Jackson (U of T) who has

iv

provided considerable guidance and support throughout my PhD, as well as helped shape my early academic career as the supervisor of my undergraduate honours thesis.

My research was significantly improved due to the specimen contributions, methodological advice provided by several individuals, and funding sources. For access to additional specimens, I would like to thank Mark Sabaj Perez and Mariangeles Arce Hernandez

(Academy of Natural Sciences of Drexel University); Kevin Conway and Heather Prestridge

(Texas A&M University); Dean Hendrickson, Adam Cohen, and Melissa Casarez (University of

Texas at Austin). For insightful discussions and considerable help with methods, I would like to thank Dean Adams (ISU), Michael Collyer (Chatham University), Antigoni Kaliontzopoulou

(Research Centre in Biodiversity and Genetic Resources) for the opportunity to participate in the

Introduction to Geometric Morphometrics in R workshop, Tromsø, Norway during the formative years of my PhD. I am very fortunate to have received funding for my research from

NSERC (CGS M and PGS D), various fellowships/awards from the University of Toronto, the

American Society of Ichthyologists and Herpetologists (Raney Award), and the American

Cichlid Association (Guy Jordan Research Award).

I am very grateful to Katriina Ilves (Pace University), Nathan Lujan, and Luke Frishkoff for valuable research and personal support regarding my career. I thank Nathan Lujan for the opportunity to collaborate and expand my research focus. I am indebted to my labmates Jessica

Arbour, who has provided endless R support and methodological assistance, Frances Hauser and

Viviana Astudillo-Clavijo, who have provided insightful comments on my thesis and moral support, and Stéphanie Lefebvre, who aided in shaping Chapter 2 data, many life lessons, and for her endless personal support throughout my PhD. I also thank my labmates Alejandro

Londoño-Burbano, Sean Anderson, and Tom Morgan for advice, useful criticisms of my research, and for personal support. To my adopted lab member, Matt Kolmann, who has

v

provided considerable research and personal criticisms in the office, lab, and field, as well as support to help me grow as a person, I thank for all he put and helped me through. I especially thank Matt for our collaboration which expanded considerably to become Chapter 2 of my thesis. I would like to thank Collin VanBurren for our enlightened discussions about life, research, and evolutionary biology throughout my PhD. One such discussion about the assumptions of body size across fishes largely spurred Chapter 2. I owe many thanks to fellow graduate students James Boyko, Kentaro Chiba, Tom Cullen, Danielle DeCarle, Michael Foisy,

Melanie Massey, Karma Nanglu, Ashley Reynolds, Santiago Sánchez-Ramírez, and Mateusz

Wosik for significant guidance and support. I extend sincere appreciation to my friend Stacey

Kerr, who has carried, dragged, and pushed me through the most difficult times and who has provided endless support through my undergraduate and graduate years. Most importantly, I would like to thank my mother Rosalie Steele who, despite our differences of opinion on fish, offered unconditional and continuous support throughout my PhD and my childhood. I owe much of what I am because of her strength and guidance throughout my life. I also owe many thanks to my father Kevin Mann, and brothers Ryan and Nolan Steele for support and helping me to have a respectable perspective on life. Finally, I would like to thank my partner Michael

Swift who has helped me through many of the most challenging moments of my life, as well as stood by me as I continued to push myself and celebrated some of the most rewarding with me.

vi

Table of Contents

Acknowledgments ...... iv

Table of Contents ...... vii

List of Tables ...... xii

List of Figures ...... xiv

List of Appendices ...... xvii

I. General Introduction...... 1

I.1 Background ...... 1

I.1.1 Body Size Distribution and Patterns ...... 1

I.1.2 Diversification of Body Size ...... 3

I.1.3 Miniaturization and Heterochrony ...... 4

I.1.4 Study System – Neotropical Cichlids (Cichlinae) ...... 5

I.2 Objectives ...... 9

I.3 Overview of the Chapters ...... 10

I.3.1 Chapter 1 ...... 10

I.3.2 Chapter 2 ...... 10

I.3.3 Chapter 3 ...... 11

I.3.4 Chapter 4 ...... 12

I.4 Contributions ...... 13

Chapter 1 Body Size Diversity and Frequency Distributions of Neotropical Cichlid Fishes (: Cichlidae: Cichlinae) ...... 14

1.1 Abstract ...... 15

1.2 Introduction...... 16

1.3 Methods ...... 20

1.3.1 Data Collection and Body Size Frequency Distributions ...... 20

1.3.2 Analysis of Body Size Frequency Distributions within Cichlinae ...... 23

vii

1.4 Results ...... 26

1.4.1 Characterization of Cichlid Body Size Frequency Distributions ...... 26

1.4.2 Distribution of Body Size Among Taxonomic Levels ...... 31

1.4.3 Divergence in Body Size Space ...... 34

1.5 Discussion ...... 37

1.5.1 Divergence in Body Size Space ...... 37

1.5.2 Body Size Diversity and the Evolution of Extreme Body Sizes ...... 39

1.5.3 Is Body Size Adaptive? ...... 40

1.5.4 Conclusion ...... 43

1.6 Appendices ...... 44

Chapter 2 Body Size Diversity Across the Fishes and the Impact of Environment on Absolute Body Size: Does Water and Macrohabitat Type Shape the Distribution of Size? ...... 71

2.1 Abstract ...... 72

2.2 Introduction...... 73

2.3 Methods ...... 81

2.3.1 Data Acquisition ...... 81

2.3.2 Characterizing the BSFD ...... 82

2.3.3 Creating Null Distribution and Expected Values ...... 83

2.3.4 Testing for Correlation between BSFD and Evolution...... 83

2.3.5 Testing for divergence due to Ecology ...... 84

2.4 Results ...... 85

2.4.1 Broad Scale Patterns across Environment and Habitat...... 85

2.4.2 Divergence Among the Major Lineages of Fishes ...... 88

2.4.3 Body Size Divergence of Families across Environment ...... 89

2.4.4 Body Size Divergence of Families across Macrohabitat ...... 90

2.4.5 Body Size Divergence across Taxonomic Hierarchy ...... 95

viii

2.4.6 Correlation between Clade Age, Richness, and Ecology ...... 97

2.5 Discussion ...... 97

2.5.1 Marine Fishes are Larger than Freshwater Fishes ...... 97

2.5.2 Elasmobranchs are Larger than Actinopterygians ...... 101

2.5.3 Are Distributions Scale-Dependent in Fishes ...... 103

2.5.4 Conclusion ...... 107

Chapter 3 Extreme Body Size Reduction, Morphological Modification, and Allometry in Neotropical Cichlid Fishes (Cichliformes: Cichlidae: Cichlinae) ...... 108

3.1 Abstract ...... 109

3.2 Introduction...... 110

3.3 Methods ...... 117

3.3.1 Validation of Geometric Morphometric Analyses, Sampling Rarefaction, and Sample Size Determination ...... 117

3.3.2 Data Collection and Taxon Sampling ...... 120

3.3.3 Body Size Reduction ...... 124

3.3.4 Paedomorphism in Lateral Line and Skeletal Elements ...... 125

3.3.5 Paedomorphism in External Morphology ...... 127

3.3.6 Significant Reductions in Duration of Development and Growth ...... 128

3.3.7 Shifts in Magnitude of Shape Change ...... 129

3.4 Results ...... 131

3.4.1 Body Size Reduction ...... 131

3.4.2 Paedomorphism in Lateral Line and Skeletal Elements ...... 133

3.4.3 Paedomorphism in External Morphology ...... 137

3.4.4 Significant Reductions in Duration of Development and Growth ...... 141

3.4.5 Shifts in Ontogenetic Rate of Development ...... 145

3.5 Discussion ...... 147

3.5.1 Testing for Miniaturization: A Quantitative Approach ...... 147

ix

3.5.2 Potential Processes of Miniaturization ...... 149

3.5.3 Adaptive Nature of Extreme Body Size Reduction in Neotropical Cichlid Fishes ...... 152

3.5.4 Conclusion ...... 157

3.6 Appendices ...... 158

Chapter 4 Morphological Disparity Across the Ontogeny and Phylogeny of Neotropical Cichlids Fishes (Cichliformes: Cichlidae: Cichlinae) and its Impact on Evolution of Body Size ...... 159

4.1 Abstract ...... 160

4.2 Introduction...... 161

4.3 Methods ...... 166

4.3.1 Data Collection and Taxon Sampling ...... 166

4.3.2 Quantifying Ontogenetic Allometry ...... 169

4.3.3 Examining Allometric Variation ...... 170

4.3.4 Examining Shape Disparity ...... 171

4.3.5 Exploring the Evolution of Ontogeny and Body Size ...... 173

4.4 Results ...... 176

4.4.1 Quantifying Ontogenetic Allometry ...... 176

4.4.2 Examining Allometric Variation ...... 177

4.4.3 Examining Shape Disparity ...... 180

4.4.4 The Evolution of Ontogeny and Body Size ...... 181

4.5 Discussion ...... 185

4.5.1 Differing Rates and Trajectories Among the Cichlids ...... 185

4.5.2 Conservation of Shape Disparity ...... 189

4.5.3 Conservation of Shape Across Ontogeny and Phylogeny ...... 193

4.5.4 Conclusion ...... 196

C. General Conclusion ...... 198

C.1 Summary ...... 198

x

C.2 Future Directions ...... 201

References...... 205

xi

List of Tables

Chapter One ……………………………………………………………………………………...

Table 1.1: Summary statistics quantify cichlid body size and testing for divergence among

Cichlinae, tribes, major clades and genera ...... 29

Table 1.2: Partitioning of body size among clades using Kolmogorov-Smirnov tests for

differences in frequency distributions ...... 33

Chapter Two ……………………………………………………………………………………...

Table 2.1: Diversity of species and families used to test for body size divergence ...... 82

Table 2.2: Summary statistics for and Elasmobranchii, environment and

macrohabitat assemblages ...... 87

Table 2.3: Summary statistics for Actinopterygii and Elasmobranchii, environment and

macrohabitat assemblages ...... 92

Table 2.4: Summary statistics for Elasmobranch families, across environment and macrohabitat

assemblages ...... 93

Chapter Three ……………………………………………………………………………………

Table 3.1: Paedomorphism in lateral line and skeletal elements ...... 138

Table 3.2: Paedomorphism in external morphology of – clade .... 139

Table 3.3: Paedomorphism in external morphology of clade ...... 139

Table 3.4: Paedomorphism in external morphology of ‘dwarf’ taxa ...... 140

xii

Table 3.5: Paedomorphism in external morphology of ...... 140

Table 3.6: Ontogenetic milestones in Neotropical cichlids including maximum attainable adult

body size, ontogenetic threshold, threshold percentage, slope of development, and slope of

growth...... 144

Table 3.7: Regression coefficients and magnitude of shape change for ‘dwarf’ Neotropical

cichlids and their sister taxa ...... 146

Chapter Four ……………………………………………………………………………………..

Table 4.1: Multivariate model fitting results for body size and ontogenetic parameter evolution

on the MCC tree ...... 184

xiii

List of Figures

Chapter One ……………………………………………………………………………………...

Figure 1.1: Phylogeny and taxonomic designation of Cichlinae ...... 22

Figure 1.2: Body size frequency distributions of the Crenicichla-Apistogramma-Satanoperca

Clade and the -- Clade of fitted to

1000 random subsamples of Geophagini ...... 28

Figure 1.3: Distributions of body size within Cichlinae and the major subclades of Geophagini

and with genera deviating from their containing clades ...... 30

Figure 1.4: Body size frequency distributions Geophagini, , and Heroini fitted

to1000 random subsamples of Cichlinae ...... 32

Figure 1.5: Phylogenetic autocorrelation of body size using taxonomic hierarchy as a proxy for

phylogenetic relatedness using Moran’s I ...... 35

Figure 1.6: Body size diversity and occupation of Cichlinae ...... 36

Chapter Two ……………………………………………………………………………………...

Figure 2.1: Theoretical effects of skew and kurtosis on mean body size and diversity, and the

importance of examining shape...... 75

Figure 2.2: BSFDs of fishes across classes, environment, and macrohabitat ...... 86

Figure 2.3: BSFDs of Actinopterygian families among environments and macrohabitats

compared to assemblage BSFDs ...... 91

xiv

Figure 2.4: BSFDs of Elasmobranch families within marine environments compared to all

marine fishes ...... 94

Figure 2.5: Deviations experienced in BSFDs of taxonomic groups, as compared to distributions

of respective inclusive clades……………………………………………………………..96

Chapter Three ……………………………………………………………………………………

Figure 3.1: Theoretical threshold for proportional and un-proportional miniatures compared to

the ontogenetic development of sister taxa ...... 112

Figure 3.2: Theoretical relationships of allometric and regression models comparing ontogenetic

trajectories of dwarf and sister taxa...... 113

Figure 3.3: Regression parameters simulated under various sample sizes and distribution of

Acarichthys heckelii ...... 119

Figure 3.4: ‘Dwarf’ Neotropical cichlids and sister taxa used for comparison with maximum

attainable body size and ontogenetic size change ...... 121

Figure 3.5: The two-dimensional external morphology landmarks used for shape analysis across

the ontogeny of Neotropical cichlids ...... 123

Figure 3.6: Traits supporting miniaturization in Neotropical cichlids ...... 133

Figure 3.7: Comparisons of ontogenetic trajectories of Neotropical ‘dwarf’ cichlids and their

sister taxa ...... 143

Chapter Four ……………………………………………………………………………………..

xv

Figure 4.1: The two-dimensional external morphology landmarks used for shape analysis across

the ontogeny of Neotropical cichlids ...... 168

Figure 4.2: Ontogenetic variation in morphology of Neotropical cichlids fishes, described by the

common allometric and residual shape components ...... 178

Figure 4.3: Dissimilarity matrix of species pairwise parameter distances, indicating similarity

and dissimilarity of magnitude of shape change, angle between trajectories, mean shape,

and morphological disparity ...... 179

Figure 4.4: Principal components analysis of high-dimensional external shape data indicating

juvenile and adult disparity of Neotropical cichlids...... 182

Figure 4.5: Principal components analysis of high-dimensional external shape data indicating

species disparity and divergence ...... 183

Figure 4.6: Results from SURFACE analyse for the evolution of body size and ontogenetic

parameters on the maximum clade credibility chronogram (MCC) ...... 186

Figure 4.7: Movement from juvenile to adult Neotropical cichlids in morphospace and size-

shape space ...... 192

xvi

List of Appendices

Chapter One ……………………………………………………………………………………...

Appendix 1.1: Body size data and taxonomic affiliation, retrieved from Fishbase, of Neotropical

cichlid species ...... 44

Appendix 1.2: Summary statistics of Cichlinae tribe and subclade body size distributions and

expectation under random phylogenetic distribution ...... 56

Appendix 1.3: Summary statistics of cichlid genera body size distributions and expectation

under random phylogenetic distribution ...... 60

Chapter Three ……………………………………………………………………………………

Appendix 3.1: Glossary of ontogenetic terms ...... 158

xvii

1

I. General Introduction

I.1 Background

I.1.1 Body Size Distribution and Patterns

Body size exhibits a strong correlation with most ecological and life history traits such as fecundity, trophic position (Romanuk et al. 2011), longevity (Blackburn and Gaston, 1992;

Brown et al., 1993), and extinction risk (Cardillo et al., 2005; Clauset and Erwin, 2008), among others. Body size should also predict outcomes of biotic interactions with co-occurring species with regards to competition (Brown & Wilson, 1956), resource partitioning or predator-prey interactions (Wilson, 1975), as well as abiotic interactions with the environment such as habitat use (Collar et al., 2011), migration (Alerstam et al., 2003), and range size (Lindstedt et al.,

1986). Biogeographers, systematists, and macroecologists have sought to understand the patterns of body size distributions of closely-related taxa and more recently, whole communities. The prevalence of right-skewed distributions (i.e., higher relative proportion of small-bodied species) appears to be ubiquitous across taxa, time, and space (Allen et al., 2006;

Bakker & Kelt, 2000; Blackburn & Gaston, 1994; Hutchinson & MacArthur, 1959). The propensity of species assemblages to display right-skewed distributions leads to the questioning of how, and more importantly, why these patterns come about.

The prevalence of right-skewed body size distributions has alternatively been explained by five mechanistic hypotheses: 1) Energetic; 2) Phylogenetic; 3) Biogeographic; 4) Textural

2

Discontinuity; and 5) Community Interaction (Allen et al., 2006 and references therein). Main proponents theorized higher proclivity of smaller to partition resources, and therefore, niche space (Griffiths, 1986; Hutchinson & MacArthur, 1959); or that optimizing physiological mechanisms between energy intake and reproductive output shaped body size distributions across taxa (Brown et al., 1993). Brown et al. (1993) proposed that the propensity of right-skew could be attributed to universal allometric constraints leading to conservation of distribution shape (i.e., skew), while clade-dependent constraints lead to variation in mean or diversity of body size. The phylogenetic history of a clade also constrains the distribution of body sizes, leading to the resemblance of closely related taxa (Clauset & Erwin, 2008) while higher net diversification at small body size (due to faster generation times, lower dispersal, and larger populations) can result in the accruement of small-bodied species within a clade (Kozłowski &

Gawelczyk, 2002). The textural discontinuity hypothesis predicts that smaller organisms experience a larger, more heterogeneous world than larger species, therefore are highly dependent on local habitat heterogeneity and can partition habitat more easily (Hutchinson and

MacArthur, 1959; Brown and Nicoletto, 1991; Bakker and Kelt, 2000). The community interaction hypothesis predicts that most of body size “filtering” – the ecological forces shaping distributions in habitats – is primarily a consequence of competition between species sharing similar body sizes and using similar resources (Hutchinson, 1959). However, later investigations found that local body size distributions varied given the spatial, temporal, and taxonomic scales across which observations were recorded (Cox et al., 2011). In other words, the distribution of body size is scale dependent, and likely driven by the combined effects of these mechanisms.

3

I.1.2 Diversification of Body Size

The role of body size in the diversification of species, morphologies, and ecologies has been explored in a number of vertebrate taxa including terrestrial mammals (Monroe & Bokma, 2009;

Smith et al., 2004), aquatic mammals (Slater et al., 2010), birds (Bokma, 2004), amphibians

(Zimkus et al., 2012), reptiles (Collar et al., 2011), rodents (Avaria-Llautureo et al., 2012), and fishes (Rabosky et al., 2013). In these studies, body size evolution and diversity were associated with increased diversification of morphology, ecology, and/or species. While these studies largely examine univariate measurements (i.e., length or mass), body size evolution and its effects on morphology and ecology may be more accurately examined using multivariate traits datasets. For example, studies incorporating ontogenetic data or multivariate shape components have been able to link major shifts in body size evolution with major developmental changes, leading to subsequent shifts in ecology (Galatius et al., 2011; Masters et al., 2014).

Body size is an easily accessible trait with strong implications for ecological adaptation

(Collar et al., 2011; Wilson, 1975), yet few studies have looked specifically at body size as a critical component driving diversification in fishes. How body size has influenced the diversification of the largest group of vertebrates, the fishes, is still largely unknown. Rabosky et al. (2013) found correlations between species diversification and the evolution of body size in fishes. Whether this is consistent across morphological and ecological axes of divergence, or whether diversification rates are asymmetrical across the body size gradient of fishes, as seen in mammals (Monroe & Bokma, 2009) and birds (Bokma, 2004), is unknown.

4

I.1.3 Miniaturization and Heterochrony

Variation among species in body size and shape are ultimately driven by changes in the timing of developmental events throughout the ontogeny of taxa (McKinney & McNamara, 1991). The study of these changes is fundamental in determining the differences in growth patterns leading to the divergence in adult body size or morphology commonly studied. Heterochrony – developmental changes in rates and timing, leading to changes in shape and size – can be divided into two major processes: the deceleration of growth resulting in underdevelopment

(paedomorphosis) and the acceleration of growth resulting in overdevelopment (peramorphosis).

In turn, these patterns can further be divided, reflecting major heterochronic changes determined by perturbations to rate or duration of growth leading to the variation that natural selection acts upon (Bolk, 1926; Gould, 1977; McKinney & McNamara, 1991). Therefore, it is important to understand how the growth of organisms, and variation among individuals or populations, leads to divergence of morphology and ecology. Paedomorphism and miniaturization of taxa may result in drastic changes to form and function, which could impact diversification processes

(Hanken & Wake, 1993). Theory of miniaturization is abundant (Bolk, 1926; Gould, 1977;

McKinney & McNamara, 1991; Hanken & Wake, 1993); however, empirical evidence testing this theory is lacking across broad groups of vertebrates. Some attempts at quantifying miniaturization or extreme reduction using body size or combined studies of body size and morphology fail to provide a generalized method (Albert & Johnson, 2012; Weitzman & Vari,

1988). Recent attempts using ontogenetics to quantify miniaturization have lead to more universally applicable definitions of changes associated with miniaturization (Zelditch et al.

2004, Angielczyk & Feldman, 2013; Galatius et al., 2011; Masters et al., 2014).

5

A comprehensive definition of miniaturization comes from Hanken and Wake (1993) who described it as the evolution of extremely small body size within a lineage beyond a critical size at which important functional changes in physiology and ecology require major morphological change to persist. Common changes observed are reduction and structural simplification of the organism’s bauplan (e.g. loss of organs or organ systems, loss of bones, retention of tissues typically reabsorbed) leading to resemblance with subadult or embryonic stages of large-bodied relatives, and/or truncation of the developmental pathway. It is also possible to see morphological novelty and/or increased morphological variation within species in extremely reduced vertebrates (Hanken & Wake, 1993). Significant reduction in body size, significant shifts in rates of development, evidence of reductive evolution, and constraint of morphology to a juvenile-like state are therefore clear hypotheses for miniaturization that can be tested across taxa (Zelditch et al. 2004).

I.1.4 Study System – Neotropical Cichlids (Cichlinae)

Fishes represent the most speciose vertebrate lineage (Actinopterygii; ~31,200 species), span an incredible range of body sizes over 3 orders of magnitude (0.8-2000cm; Froese & Pauly, 2016), and display an extremely high diversity of morphologies, ecologies, and reproductive strategies that appear to be closely tied to body size. With an estimated 8, 000 species, much of this species and functional diversity is concentrated in the Neotropics (Reis et al., 2003; Toussaint et al., 2016). Cichlids are distributed across the old world in mainland Africa, the island of

Madagascar, Israel, Sri Lanka, and India and across the new world in South, Central, and the southern regions of North America (Froese & Pauly, 2016). While diversity is highest in the Rift

6

Lakes of East Africa – where cichlids underwent explosive adaptive radiations and dominate the entire fish diversity – the riverine Neotropical subfamily Cichlinae diversified in the midst of the most diverse assemblage of ichthyofauna in the world (Reis et al., 2003). Thus, because of their diversity and their complex evolutionary history, Neotropical cichlids provide an interesting and illuminating system to study body size evolution and its impact on diversity, as well as to understand the accumulation of species and ecological richness across body size in fishes.

The subfamily Cichlinae, the Neotropical cichlids, is the third most species-rich lineage within the Neotropics, and is arguably the most well studied lineage in this region. The subfamily currently encompasses over 500 described species, distributed among ~60 genera and seven tribes. Most species diversity is contained within three clades, Cichlasomatini (~70 species), Geophagini (~250 species), and Heroini (~150 species), with the other tribes

(Astronotini, Cichlini, , Retroculini) containing fewer than 20 species. The subfamily is widespread, frequently occurring in riverine and lacustrine systems from southern

North America (Texas, northern extent) to southern South America. However, only members of the tribe Heroini have dispersed into Central and North American habitats, and are significantly more speciose compared to South American heroines. Diversity of genera is often relatively low, with two notable exceptions Crenicichla and Apistogramma, which each currently contains roughly 100 described species (Reis et al., 2003).

Ecological diversity is high in Neotropical cichlids, with trait evolution supporting an adaptive radiation of species and ecological roles in this group, rivalling that of its East African

Rift Lake counterparts (Arbour & López-Fernández, 2014; López-Fernández et al., 2013).

Cichlasomatini is largely dominated by generalist species, with relatively high shape and ecological conservatism. The tribe Geophagini exhibits the largest diversity of ecology (Arbour

7

& López-Fernández, 2014; López-Fernández et al., 2013), with species ranging from small- bodied invertebrate pickers to large-bodied piscivores. It contains the Crenicichla, which itself is an ecologically diverse group ranging from small-bodied insectivores to large-bodied predators specializing in fish and epibenthic invertebrates. The tribe also contains “sifters” feeding on benthic invertebrates via sand-sifting and winnowing prey through modified gill arches, many small-bodied invertebrate pickers, as well as generalist lineages (López-Fernández et al., 2014). Heroini also displays a considerable amount of ecological diversity, with both unique and convergent ecological roles to those observed within Geophagini and other tribes

(Cochran-Biederman & Winemiller, 2010; López-Fernández et al., 2010, 2013; Winemiller et al., 1995). While ecological diversity appears to be constrained in South American heroines, ecological diversity has expanded rapidly upon their invasion of Central America, including roles occupied by South American geophagines (López-Fernández et al., 2013). Ecological diversity in the subfamily is further expanded by Chaetobranchus (2 spp.) and

Chaetobranchopsis (2 spp.), two genera of planktivore specialists, while

(Astronotini), Cichla (Cichlini), and (Retroculini) occupy ecological roles also exhibited in other tribes. Neotropical cichlids primarily occupy riverine environments, though microhabitat occupation within rivers appears to be determined by ecology and body size

(Burress, 2016; Cochran-Biederman & Winemiller, 2010; Jepsen et al., 1999; López-Fernández et al., 2012, 2014; Montaña & Winemiller, 2010; Winemiller et al., 1995).

Neotropical cichlids span a large range of body size relative to South American ichthyofauna. Neotropical cichlids, however, do not occupy the most extreme small or large body sizes exhibited by other endemic Neotropical lineages. While most cichlids appear to be intermediate in size, there are several lineages that show considerable decreases and increases in body size. The smallest cichlid, Apistogramma staecki, has a maximum body size of 21mm

8

(standard length; SL), and the largest is , reaching body lengths upwards of

1000mm (total length; TL). Genera also appear to have low variation in body size, except for

Crenicichla which range 40-312mm (SL). Discussion of extreme body size reduction has been discussed widely in the literature, and the genera Apistogramma (including ),

Biotoecus, Dicrossus, , , , and Taeniacara are purported dwarf taxa. In addition, small-bodied members of the genus Crenicichla (including

Teleocichla) have also been considered ‘dwarf’ taxa (e.g. , Crenicichla regani, C. compressiceps, C. wallacii). In South American fishes, body size reduction and reductive characteristics were described by Weitzman and Vari (1988) who admittedly failed to include members of the family Cichlidae in their description despite evidence for these reductive characters. To date, there is no quantitative study examining the validity of Neotropical cichlid miniaturization designations.

Significant progress has been made in understanding species richness, evolutionary history, ecological diversity, geographic distribution, and morphological differentiation within this group. Most of this work has focused on adult specimens, however there is some understanding of ecology in early ontogenetic stages, as well as how cichlids develop (Barlow,

1991; Kupren et al., 2014; Meyer, 1987, 1990; Ponton & Mérigoux, 2000; Ponton & Mérona,

1998). Very little is known about broad evolutionary patterns of ontogenetic change in any trait, with few studies comparing ontogenetic change across distantly related taxa (Ponton &

Mérigoux, 2000). It is known from few taxa that ontogenies can be quite variable, with some cichlids showing plasticity in ontogenies leading to divergence of adult ecological roles (Meyer,

1987, 1990; Santos-Santos et al., 2015). Interestingly, these modifications within species often do not follow similar ontogenetic pathways across species, suggesting that cichlid ontogenies may not be particularly constrained developmentally.

9

The phylogenetic resolution of Cichlinae compared to the East African Rift Lakes, occupation within a diverse assemblage (i.e., South American ichthyofauna) that exhibits considerable body size diversity, as well as the growing knowledge of morphological and ecological diversity and evolution in this group, provides an excellent model for studying macroevolution in tropical, freshwater ichthyofauna. Therefore, I used Neotropical cichlids as a model system to explore body size evolution in fishes in a robust phylogenetic, ecological, and morphological framework.

I.2 Objectives

The overall goal of this dissertation is to identify the role of body size in the diversification of species and their respective morphology and ecology in Neotropical cichlids. I ask how body size is distributed across the phylogeny of Neotropical cichlids and if there is an association between body size occupation, body size diversity, overlap of body size space and species diversity? Does taxonomic, environmental, and macrohabitat affiliation affect the distribution of body size across the fish tree of life and how do Neotropical cichlids compare? What changes in the ontogenetic processes of Neotropical cichlids result in differences in the growth patterns leading to variation in adult body size and shape? How does body size evolve in Neotropical cichlids? Does

Neotropical cichlid body size evolution affect morphological and ecological diversification? By merging traditional methods in macroecology with innovative phylogenetic comparative methods, I will provide a comprehensive understanding of how body size evolves and its consequences within a robust phylogenetic framework.

10

I.3 Overview of the Chapters

I.3.1 Chapter 1

Chapter 1 characterizes the diversity and distribution of body size across the phylogeny of

Cichlinae, using 88% of described species. The identification of extreme body size regions, potential adaptive peaks of body size evolution, and the degree of overlap among Neotropical cichlid fishes was accomplished via a macroecological approach comparing distributions to null expectations based on random evolution of body size. Distributions for 498 species of Neotropical cichlids distributed among ~55 genera and three tribes were more overlapping and convergent in distantly related taxa than expected under the null model. Overlap of genera is concentrated in three regions of body size space, with diversity of each clade being strongly constrained around the mean body size of the group. These three regions could represent adaptive peaks in body size of Neotropical cichlids. Qualitatively, I found that species diversity is typically associated with intermediate body sizes, while extreme regions of body size (both large and small) are associated with species poor lineages and little overlap among lineages.

I.3.2 Chapter 2

Chapter 2 characterizes body size occupation, diversity, and skew across major lineages of fishes within marine and freshwater environments as well as across major macrohabitats to determine if significant divergence occurs as a result of habitat differentiation. Using a dataset comprised of distribution, environment (i.e., water type) and macrohabitat data for approximately 32,000 species of fishes, I tested for divergence of body size among broad taxonomic groups as well as

11 across environment and macrohabitat. Body size frequency distributions (BSFDs) of major lineages, assemblages and families were characterized and compared to null distributions (random phylogenetic evolution of body size). Marine fishes were found to be significantly larger than freshwater fishes. While there is no difference in the size of the smallest freshwater fishes as compared to marine fishes, the largest fishes inhabiting freshwater are significantly smaller than the largest fishes inhabiting marine waters. Marine fishes were found to have relatively normal distributions, while freshwater fishes were overwhelmingly right-skewed. Macrohabitat distributions followed the same generally trends as their respective environment BSFDs, with few exceptions. I found that body size within families was highly conserved, and that body size parameters were not correlated with clade age, but were correlated with species richness and repeatedly with occupation of marine environments.

I.3.3 Chapter 3

Chapter 3 assesses whether body size in Neotropical cichlid ‘dwarf’ lineages is significantly reduced leading to reductive evolution (miniaturization sensu Hanken & Wake, 1993; Weitzman

& Vari, 1988) and if small body size can be attributed to some change in the rate or duration of growth relative to large-bodied relatives. This chapter also assesses whether miniaturization is proportional or un-proportional in taxa (sensu Gould, 1971), which could lead to differences in the functional attributes of species. By comparing small-bodied lineages to their sister taxa, I determine whether there is strong evidence to support miniaturization within Neotropical cichlids.

I test for body size reduction, paedomorphism, and ontogenetic truncation, finding significant support for most species previously purported as miniatures. I find that there is a gradient of

12 morphological modification associated with body size decrease, with high variation in the apparent response to body size reduction.

I.3.4 Chapter 4

Chapter 4 examines the disparity seen in shape, ontogeny, and body size across the phylogeny of Neotropical cichlids. I first examined how much morphological variation in cichlid ontogenies is captured by common trends in development across the tree, and whether there were taxa that deviated from this trend. I then tested whether the variation in ontogenetic pathways resulted in divergence of morphology from a similar juvenile form as cichlids developed into their adult forms. Finally, I tested whether the evolution of these traits was correlated in a way that supported previous hypotheses about the evolution of body size in cichlids. The conservation of cichlid shape through ontogeny and phylogeny resulting in no difference in morphological disparity across juvenile and adult cichlids. Neotropical cichlids follow a strong common allometric trajectory through ontogeny from an elongate juvenile with relatively large eyes and head, to an adult which is deep bodied, with a developed lateral line and relatively small cranial features. The evolution of ontogenetic parameters and body size show little convergence among cichlids sharing similar body forms (e.g. miniatures) or ecologies (e.g. winnowers), rather show shifts in adaptive regimes associated with the origin of major lineages of cichlids.

13

I.4 Contributions

Chapter 1 was published in “PLoS ONE”. Chapters 2-4 will be submitted for publication in a similar form to that presented here. I will be first author on all papers from Chapter 2-4.

14

Chapter 1 Body Size Diversity and Frequency Distributions of Neotropical Cichlid Fishes (Cichliformes: Cichlidae: Cichlinae)

Sarah Elizabeth Steele1 and Hernán López-Fernández1,2

1Department of Ecology and Evolutionary Biology, University of Toronto, 25 Wilcocks St.,

Toronto, Ontario M5S 3B2, Canada

2Department of Natural History, Royal Ontario Museum, 100 Queen’s Park, Toronto, Ontario

M5S 2C6, Canada

Published As: Steele SE, López-Fernández H (2014) Body Size Diversity and Frequency

Distributions of Neotropical Cichlid Fishes (Cichliformes: Cichlidae: Cichlinae). PLoS ONE

9(9): e106336. doi:10.1371/journal.pone.0106336

15 Chapter One

1.1 Abstract

Body size is an important correlate of life history, ecology, and distribution of species. Despite this, very little is known about body size evolution in fishes, particularly freshwater fishes of the

Neotropics where species and body size diversity are relatively high. Phylogenetic history and body size data were used to explore body size frequency distributions in Neotropical cichlids, a broadly distributed and ecologically diverse group of fishes that is highly representative of body size diversity in Neotropical freshwater fishes. Divergence, phylogenetic autocorrelation, and among-clade partitioning of body size space was tested. Neotropical cichlids show low phylogenetic autocorrelation and divergence within and among taxonomic levels. Three distinct regions of body size space were identified from body size frequency distributions at various taxonomic levels corresponding to subclades of the most diverse tribe, Geophagini. These regions suggest that lineages may be evolving towards size optima that may be tied to specific ecological roles. The diversification of Geophagini appears to constrain the evolution of body size among other Neotropical cichlid lineages; non-Geophagini clades show lower species- richness in body size regions shared with Geophagini. Neotropical cichlid genera show less divergence and extreme body size than expected within and among tribes. Body size divergence among species may instead be present or linked to ecology at the community assembly scale.

16 1.2 Introduction

The importance of body size on the life history, ecology and distribution of species has been highlighted continuously in the literature (Allen et al., 2006; Brown et al., 1993; Cardillo et al.,

2005; Lindsey, 1966). Nevertheless, little work has been completed to answer broad questions of body size evolution and its importance. In addition, few empirical studies have addressed the evolutionary processes that underlie body size distributions across geographic space and time

(Smith et al., 2004). Many studies, particularly in mammals, have addressed how body size is distributed on a broad geographic scale (Arita & Figueroa, 1999; Bakker & Kelt, 2000; Brown

& Nicoletto, 1991) or across the fossil record (Smith & Lyons, 2011), but an understanding of how body size is distributed in a phylogenetic context among organisms is far from complete

(Diniz-Filho & Nabout, 2009; Smith et al., 2004). Cope’s Rule, the phyletic increase in body size over evolutionary time (Clauset & Erwin, 2008; Kingsolver & Pfennig, 2004), and

Bergmann’s Rule, the increase of body size with increases in latitude (Blackburn et al., 1999), have been proposed based on the mammalian fossil record to outline fundamental patterns of body size distribution. Recent studies from other taxa have suggested that these “rules” do not always apply, and may be the exception rather than the rule (Belk & Houston, 2002; Fu et al.,

2004; Knouft & Page, 2003).

Body size, like other phenotypic traits, is expected to be similar among closely related taxa due to evolutionary constraints on morphology tied to biologically and ecologically relevant characters (Felsenstein, 1985; Harvey & Pagel, 1991). Yet there are likely many exceptions where closely related taxa are highly divergent in body size: body size divergence could allow habitat or resource partitioning in coexisting congeneric species (Kelt et al., 1999;

Mahler et al., 2010) or other closely related taxa, while body size shifts associated with

17 ecomorphological differentiation could allow access to novel habitats or unused resources

(Griffiths, 1986; Polo & Carrascal, 1999). Furthermore, if body size is so important for physiological and ecological processes, evolution towards extreme body size, especially small body size, must result in one or several evolutionary trade-offs in life-history and ecological characteristics of these species (Griffiths, 1986; Hanken & Wake, 1993). But at what taxonomic resolution should these trade-offs occur? How closely related would we expect species to be that share both the same body size and the same suite of behaviours, reproductive modes, diet preferences or morphologies?

Only recently has body size distribution been examined in fishes, with several studies primarily investigating the distribution of body size across geographic space (Fu et al., 2004;

Knouft, 2004) or with regard to basic ecological characters (Griffiths, 2006, 2010, 2012, 2013).

The evolutionary history of body size has rarely been addressed in fishes on a broad geographic scale that links possible phylogenetic constraints of body size evolution over time with the ecological and geographic distribution of extant taxa (Albert et al., 2009; Albert & Johnson,

2012; Hardman & Hardman, 2008). Body size reduction in fishes has been shown to be a common phenomenon in tropical systems (Griffiths, 2012; Lavoué et al., 2010), particularly in freshwater environments (Bennett & Conway, 2010; Weitzman & Vari, 1988). As a consequence, distributions of fishes in the tropics tend to be right-skewed (Griffiths, 2012), though direction and intensity of skew varies depending on evolutionary history, environmental characteristics and ecology (Kozłowski & Gawelczyk, 2002). Though previous work has begun to outline the link between body size and ecology in fishes, very little is known about body size distributions, evolution, and the consequences of occupying a particular body size space in

Neotropical fishes.

18

In this chapter, the distribution of body size in a phylogenetic context was examined across Neotropical cichlids, a group of fishes with a broad geographic distribution that is also highly diverse in species and ecological roles. By examining this system, identification of potentially important drivers that may influence body size evolution in Neotropical fishes as a whole. Cichlids have been used as a model to study a number of biological questions in fishes due to their ecological versatility and life history traits (Salzburger, 2009; Salzburger & Meyer,

2004; Seehausen, 2006). The radiation of cichlids generated extraordinary diversity both taxonomically and ecologically within a single family. The clade of Neotropical cichlids

(subfamily Cichlinae) is the third most species-rich lineage in South America following

Characidae and Loricariidae (Reis et al., 2003), but a robust hypothesis of evolutionary history for these latter two groups has yet to be developed. Moreover, Cichlinae shows a high degree of ecological and body size diversity in addition to high species diversity. Despite this diversity, very little is known about Neotropical cichlid body size diversification. Sexual dimorphism does occur in several genera of Neotropical cichlids (Reis et al., 2003; Wimberger et al., 1998) but it is poorly characterized, and where present, it is not known whether it is associated with sexually selected behavioural traits (e.g. sneaker males, shell-dwelling) or if it has marked impacts on ecological strategies as seen in dimorphic species of African cichlids (Schütz & Taborsky,

2000). No known studies have examined body size in all described Neotropical cichlids, and very few studies have examined the association between body size and ecological or life history traits that could be driving patterns identified in this chapter.

Most of the species richness in Cichlinae is distributed among three tribes. Geophagini is the most species-rich (243 species, Froese & Pauly, 2013) and, along with Cichlasomatini (74 species, Froese & Pauly, 2013), is primarily restricted to South America. Heroini (176 species,

Froese & Pauly, 2013) is the second most diverse tribe following Geophagini and has expanded

19 from South America into Central and North America. Recent work on Neotropical cichlids has shown that the process underlying body size evolution in Cichlinae may vary among tribes, but that body size may have diverged early in the evolution of this group, which may have resulted in the accumulation of higher body size diversity over time and possibly greater divergence among distantly related lineages (López-Fernández et al., 2013). These analyses, however, were performed on a relatively small subset of taxa, with distributions and evolutionary patterns of body size below the tribe level not explicitly explored. Body size in Cichlinae spans a large portion of the body size space occupied by Neotropical fishes, ranging from 21 mm standard length (SL) in Apistogramma staecki to 990 mm total length (TL) in Cichla temensis. Such body size diversity in Cichlinae provides a case study to investigate how extreme body size impacts various aspects of life history, ecology, and distribution of freshwater fishes. High diversity of ecology in this group presents morphological disparity and ecological diversification ideal for examining factors affecting body size evolution and a strong understanding of phylogenetic relationships within Cichlinae (López-Fernández et al., 2010) provides the phylogenetic framework for addressing associations with body size in an evolutionary context. As a prerequisite for identifying the underlying processes and forces driving body size evolution in

Cichlinae, a clear understanding of patterns of body size diversity at relevant phylogenetic levels is needed across the geographical distribution of the group. To this end, the purpose of this chapter was to 1) quantitatively characterize body size frequency distributions and space occupation in various clades, 2) determine if body size is randomly distributed at and among various phylogenetic levels, 3) determine if body size variation correlates with phylogenetic relatedness, and 4) to distinguish small and large-bodied taxa as a foundation for future work in body size evolution of Neotropical cichlids.

20 1.3 Methods

1.3.1 Data Collection and Body Size Frequency Distributions

Maximum body size available for valid fish species within the subfamily Cichlinae were taken from FishBase (see Appendix 1.1) (Froese & Pauly, 2013). To ensure data accuracy, body size data provided in FishBase were compared against original sources provided within the database.

Maximum body sizes previously found from museum specimens to be larger than published data in the literature (López-Fernández et al., 2013) were used in this chapter (See Appendix

1.1). Measurements were given either in standard length (SL, length from the tip of the upper lip to end of caudal peduncle), or total length (TL, length from snout to posterior edge of caudal fin). Total length data were concentrated in Heroini species. To maximize taxon sampling in the dataset both SL and TL data were included to incorporate all variation available at the genus level and representatives from all genera within Cichlinae. The effect of using the two different measures was tested and it was determined that exclusion of species for which only TL was available did not change the results found at the tribe or major clade, and only affected results at the genus level (see below) in a few taxa. Exclusion only resulted in significant changes at the genus level within some heroines (12 genera, e.g. , , ) in which most or all body size was given in TL. Untransformed body size data was typically skewed, and therefore log-transformed body size data were used in all analyses unless otherwise noted. Log transformation of body size data also reduces the potential bias of TL data within distributions and analyses. Unless otherwise noted, discussions of body size with regards to the results of the chapter and the interpretation of the figures refer to log-body size (LBS).

21

Species assignment and phylogenetic partitioning of the data for subclade analyses within Cichlinae followed (López-Fernández et al., 2010). Body size data for the subfamily

Cichlinae was first partitioned by tribe. The three most diverse tribes Geophagini, Heroini and

Cichlasomatini following (López-Fernández et al., 2010) represent monophyletic clades that encompass the majority of taxonomic diversity and were used for further subdivision into less inclusive taxonomic units (Fig. 1.1), while all other tribes contain too few taxa to be partitioned further. Geophagini is composed of two major clades (López-Fernández et al., 2010),

Crenicichla-Apistogramma-Satanoperca (CAS) that is more species-rich and has higher morphological disparity than the Geophagus-Gymnogeophagus-Dicrossus (GGD) clade (Arbour

& López-Fernández, 2013; López-Fernández et al., 2012). Body size distributions of genera were analyzed separately in these two clades due to high species diversity as well as differences in ecological attributes of species. Heroini is the only tribe of Neotropical cichlids that inhabits both South and Central America, even extending into the very southern regions of North

America. To test if body size frequency distributions (BSFDs) were influenced by geographic expansion, Heroini was also analyzed by separating species into South and Central American groups (hereby referred to as SA or CA heroines) (Fig. 1.1) Although heroines were separated geographically, it is relevant to clarify that SA heroines do not comprise a monophyletic clade.

Central American heroines are monophyletic at the most basal level, but include several lineages distributed in South America that have Central American affinities (López-Fernández et al.,

2010). The CAS clade, GGD clade, CA heroines, SA heroines and Cichlasomatini as a whole were then subdivided into the genera identified by (López-Fernández et al., 2010, their Fig. 1).

22

Figure 1.1: Phylogeny and taxonomic designation of Cichlinae. A chronogram of Neotropical cichlid fishes (López-Fernández et al., 2013). Phylogenetic nomenclature used in this chapter is represented by coloured boxes for the tribes and major clades. See López-Fernández et al. (2010) for details on phylogenetic reconstruction and taxonomic conventions. CAS and GGD refer to the Crenicichla-Apistogramma-Satanoperca and Geophagus-Gymnogeophagus- Dicrossus clades, respectively, of Geophagini.

23 1.3.2 Analysis of Body Size Frequency Distributions within Cichlinae

Ecologically relevant morphological traits in Neotropical cichlids are more similar within genera than among genera (López-Fernández et al., 2012) and phenotypic divergence that resulted in currently recognized genera often followed an early-burst pattern of evolution

(Arbour & López-Fernández, 2013; López-Fernández et al., 2013). It was expected that body size distribution patterns would be consistent with other phenotypic data which found higher similarity within clades than among-clades. Therefore, characterization of body space occupation of clades was necessary to determine if divergence in body size is present within or among clades in Cichlinae. The BSFDs of Neotropical cichlids were analyzed with all available body size data at the subfamily, tribe, major clade, and genus level (Fig. 1.1). Analyses were only conducted on monophyletic groups with two or more taxa, except for the SA Heroini (see above). The mean, standard deviation, range, 25% and 75% quantiles and interquartile range

(IQR) were calculated for each BSFD. Significant deviations in mean body size may indicate shifts in body space occupation among clades. Significantly lowered standard deviation, range and interquartile range would indicate a lowering of body size diversity within clades, suggesting constrained size, while increases would indicate expansions in body size diversity.

Location of 25% and 75% quantiles together also indicate information about body size diversity, proportion of taxa within certain body size ranges and skew, but interpretation is not as clear. In addition, distributions were tested for unimodality using Hartigan’s dip test (Hartigan &

Hartigan, 1985) and characterized by kurtosis and skew. Platykurtosis, flatter as compared to a normal distribution, and leptokurtosis, more peaked than a normal distribution, give indications of constraints around the mean body size. Right-skew indicates a higher proportion of small- bodied species while left-skew indicates a higher proportion of large-bodied species within a distribution.

24

To determine if BSFDs were different among phylogenetic levels, distributions were first compared using the Kolmogorov-Smirnov (K-S) two-sample test. This test identifies differences between two observed frequency distributions and is particularly sensitive to deviations in skew, kurtosis and location along the body size gradient. Bonferroni correction was employed to account for multiple comparisons among genera in the K-S analysis

(Padj<0.00004). The K-S only determines whether two distributions differ, but does not identify what aspects of the distribution drive those differences.

A bootstrapping method was employed to test for random distribution of BSFDs between phylogenetic levels (Cox et al., 2011; Smith et al., 2004). The BSFD of the higher taxonomic unit containing the focal clade was resampled to create 1000 randomly assembled

BSFDs equal in size to the focal clade (see above). The mean, standard deviation, range, 25% and 75% quantiles, interquartile range, kurtosis, and skew were then calculated for each of the1000 BSFDs. A distribution of each summary statistics expected under a random phylogenetic distribution was then obtained. The summary statistics of the observed data could either be higher or lower than the simulated data, so a two-tailed adjusted alpha level of 0.05 was applied. Summary statistics for observed body size distributions found below and above the

2.5% and 97.5% quantiles of the simulated summary statistic distributions were considered to significantly deviate from summary statistics describing randomly distributed simulated data. A p-value was not directly calculated for each bootstrap simulation, but all deviations under or over the above thresholds were reported as significant (p<0.05). Observed data was compared to the 0.25% and 99.75% quantiles (P<0.005) (see Appendices 1.2 and 1.3). The BSFDs of subclades were compared to bootstrap pseudo-distributions created from respective clades in each successive phylogenetic level. If clades showed few or no deviations from pseudo- distributions, body size was considered a random subset of the containing higher taxonomic

25 level, suggesting low phylogenetic autocorrelation. Clades that show several significantly different summary statistics have a specialized or partitioned body size space occupation as compared to distributions at higher taxonomic levels. Tests for phylogenetic autocorrelation in body size were conducted using Moran’s I (Diniz-Filho & Nabout, 2009; Smith et al., 2004) to determine if body size was randomly distributed within a given taxonomic level. Values of

Moran’s I fall between -1 and 1, with higher values indicating the trait is more similar within taxonomic units than expected at random, 0 indicating random distribution, and values approaching -1 indicating the trait is more different than random.

To account for potential taxonomic error, all genera were compared to the higher clade and tribe that contained them (Fig. 1.1). The distributions of major clades were then compared to their respective tribe. Finally, the BSFDs of each tribe were compared to the BSFD of

Cichlinae. The genera Crenicichla and Teleocichla of Geophagini are known to form a monophyletic clade with Teleocichla potentially interspersed among Crenicichla species

(López-Fernández et al., 2010; Piálek et al., 2012), therefore body size of all species in both genera were analyzed together. Heroini contains a paraphyletic, catch-all genus ‘’ which was not analyzed at the genus level (Kullander, 1983; Stiassny, 1991). Species of

‘Cichlasoma’ with body size data available were included in the Heroini and CA Heroini

BSFDs to be resampled with phylogenetic assignment following López-Fernández et al. (2010).

26 1.4 Results

1.4.1 Characterization of Cichlid Body Size Frequency Distributions

Based on the data available from FishBase (Froese & Pauly, 2013), 498 cichlid species were included in the analyses of BSFDs (Appendix 1.1). This represents approximately 88% of the valid Neotropical cichlid species listed on FishBase at the time of data collection for this chapter. The bootstrap analyses found that only nine genera across the three main tribes had significantly smaller means than expected if body size was randomly distributed throughout the phylogeny, while eight genera had higher than expected mean body size (Appendix 1.2). These findings were generally consistent when comparing distributions of genera to respective major clades as well as at the tribe level (Appendix 1.3). Occurrences of mean body size deviation were not more frequent in any particular tribe (Geophagini 6/16; Heroini 8/25; Cichlasomatini

3/10) or major clade and no tribe was biased towards smaller or larger body size (Appendix 1.3).

Standard deviation was typically lower in all significant results, and lowering of body size diversity was particularly apparent in the CAS clade of Geophagini (5/9 cases; GGD 2/7; SA heroines 2/7; CA heroines 6/18; Cichlasomatini 1/10) (Appendix 1.3). Deviations in skew and kurtosis were not typically found at any phylogenetic level (Table 1.1).

The mean of the CAS BSFD was significantly lower than expected while the standard deviation and IQR were higher (Table 1.1). The maximum and minimum body sizes are not significantly different than expected. Despite this, the 25% and 75% quantiles were closer to the extremes of the distribution than expected (Table 1.1; Fig. 1.2A). The distribution was also significantly more platykurtic and with a higher right-skew than Geophagini. The BSFD of CAS is also strongly bimodal (p=0.0026) (Fig. 1.2A), with the small-bodied peak (LBS 1.54, 35 mm)

27 coinciding with the distribution of Apistogramma (Fig. 1.3), and the large-bodied peak (LBS

2.40, 250 mm) coinciding with the peak of Crenicichla (Fig. 1.3). Despite the bimodality of

CAS, none of its subclades deviate from a unimodal distribution. The mean of GGD was significantly higher than expected at LBS 2.07(118.2 mm) while standard deviation and IQR were lower than expected (Table 1.1; Fig. 1.2B). Minimum body size and 25% quantile were higher than expected. The BSFD of GGD did not significantly deviate from unimodality, however it was more platykurtic and left-skewed than expected. SA heroines had a significantly lower mean than expected, and a slight trend towards lower standard deviation than expected

(Table 1.1). CA heroines did not deviate from the BSFD of Heroini in any summary statistic, except for a higher mean and 75% quantile than expected (Table 1.1).

The BSFD of Cichlinae was unimodal, slightly left-skewed with a mean of LBS 2.08

(120.2 mm), and IQR from LBS 1.89 (77.6 mm) to 2.28 (190.5 mm). Bootstrap analyses revealed Geophagini had a significantly lower mean of LBS 1.97 (93.3 mm), accompanied by significantly lower minimum body size, maximum body size, 25% quantile and 75% quantile

(Table 1.1). Geophagini tended towards a bimodal distribution (p=0.0642) (Fig. 1.4A), which is also reflected by a significantly higher standard deviation and IQR as well as being significantly platykurtic. The mean of Heroini BSFD was significantly higher, at LBS 2.18 (151.0 mm), than expected (Table 1.1). Heroini shows significantly lower standard deviation and IQR, suggesting a restricted range of body size. This is supported by a significantly larger minimum body size and 25% quantile; however, the 75% quantile is higher than expected. The mean of

Cichlasomatini of LBS 1.98 (94.6 mm) and standard deviation were significantly lower than expected (Table 1.1). Maximum body size, 75% quantile and IQR were also significantly lower than expected while the minimum body size was higher.

28

Figure 1.2: Body size occupation in subclades of Geophagini. Body size frequency distributions of the Crenicichla-Apistogramma-Satanoperca (CAS) Clade (A) and the Geophagus- Gymnogeophagus-Dicrossus (GGD) Clade (B) of Geophagini. Coloured columns show observed data fitted to 1000 random subsamples of Geophagini each represented by a black line. Silhouettes depict the three most diverse genera for each clade and the primary region of body space occupation, from top-left to bottom right: Apistogramma, Satanoperca, Crenicichla, Dicrossus, Gymnogeophagus, Geophagus. Black arrows indicate potential body size optima supported by the bootstrap analyses.

29

Table 1.1: Quantifying cichlid body size and testing for divergence. Summary statistics for Cichlinae, tribes, major clades, and genera with Body Size Frequency Distributions (BSFDs) deviating from that of their containing clades (log10 transformed length in mm). An asterisk next to subclade names indicates a significant difference from the more inclusive clade above (Kolmogorov- Smirnov Test). An asterisk next to statistical values indicates significantly different values as compared to a random phylogenetic distribution attained by bootstrap simulations (* indicates p<0.05; ** indicates p<0.005). Direction of deviation can be found by comparing the value of summary statistics between clades and respective subclades. Crenicichla-Apistogramma-Satanoperca (CAS); Geophagus-Gymnogeophagus-Dicrossus (GGD); Number of species (N); Standard Deviation (St Dev); Interquartile Range (IQR). Taxa N Mean St Dev Minimu Maximum 25% 75% Kurtosis Skew IQR m Quantile Quantile Cichlinae (Subfamily) 498 2.08 0.29 1.36 3.00 1.89 2.28 -0.27 -0.01 0.39 Geophagini (Tribe)* 226 1.97** 0.31** 1.32** 2.49** 1.69** 2.23* -1.25* -0.15 0.54** Clade CAS 180 1.94* 0.33** 1.32 2.49 1.62** 2.34** -1.38** 0.01** 0.61** Apistogramma* 67 1.60** 0.13** 1.32 2.14** 1.53** 1.69** 2.93** 0.87** 0.17** Crenicichla* 91 2.17** 0.22** 1.60** 2.49 2.05** 2.35** -0.53** -0.62** 0.29** Satanoperca 8 2.25* 0.08** 2.15** 2.41 2.21** 2.28 -0.64 0.65 0.07** Clade GGD* 46 2.07* 0.21** 1.53** 2.44 1.99** 2.23 -0.10** -0.62** 0.25** Gymnogeophagus 11 2.06 0.09** 1.93 2.19* 2.00 2.12 -1.51 -0.01 0.13 Dicrossus* 5 1.71** 0.11 1.58 1.85** 1.62 1.78** -1.99 0.01 0.15 Geophagus 23 2.19** 0.13** 1.88** 2.38 2.10** 2.30* -0.79 -0.53 0.20 Mikrogeophagus 2 1.63* 0.15 1.53 1.75* 1.59* 1.69* -2.75 0.00 0.11 Heroini (Tribe)* 176 2.18** 0.21** 1.70** 2.70 2.01** 2.35* -0.57 -0.014 0.34* SA Heroines 21 2.09* 0.18 1.70 2.40 1.97 2.20 -0.81 -0.22 0.23 CA Heroines 153 2.19* 0.21 1.72 2.70 2.04 2.38* - 0.63 -0.01 0.34 Cichlasomatini (Tribe)* 72 1.98** 0.15** 1.56 2.30** 1.87 2.08** -0.20 -0.47 0.21** Laetacara 6 1.77** 0.14 1.56 1.91** 1.69* 1.86** -1.66 -0.44 0.17 Nannacara* 5 1.72** 0.07 1.65 1.83** 1.69* 1.75** -1.61 0.48 0.06

30

Figure 1.3: Body size diversity and occupation of Cichlinae. Distributions of body size within Cichlinae and the major subclades of Geophagini and Heroini. Body size distributions of genera with BSFDs deviating from their containing clades are included. Representatives of each clade are given; photographs not to scale. Dots to the left and right of boxplots indicate outliers of the distribution. Numbers above bars indicate the number of species within that particular body size bin. CAS and GGD refer to the Crenicichla-Apistogramma-Satanoperca and Geophagus- Gymnogeophagus-Dicrossus clades, respectively, of Geophagini. Photographs by Hernán López-Fernández and courtesy of Anton Lamboj. Body size threshold proposed (Arbour & López-Fernández, 2013) identified by prominent dashed line and bold font on axis. See Fig. 1.6 for additional distributions.

31 1.4.2 Distribution of Body Size Among Taxonomic Levels

If body size is randomly distributed across a phylogeny (i.e., no phylogenetic autocorrelation), little deviation of BSFDs is expected between clades and their subclades, as well as high similarity between distributions of related subclades at the same taxonomic level (Diniz-Filho &

Nabout, 2009; Smith et al., 2004). High correlation within clades (phylogenetic autocorrelation) should result in considerable partitioning of body size space. Only 46 out of 1249 pairwise comparisons between Cichlinae genera showed significantly different BSFDs from each other

(padj<0.00004) using the K-S test (for comparisons of major clades see Table 1.2; results at genus level not shown), supporting randomly distributed body size across genera. Of these 46 cases, 14 comparisons involved Cichla (Cichlini), a large-bodied piscivorous genus occupying a body size space that few other taxa occupy. In addition, 17 other cases involved the “dwarf” cichlid genus Apistogramma (Geophagini), which significantly differed from several genera across all three tribes. The remaining 15 cases typically involved comparisons between genera from different tribe affinities rather than divergence of genera within the same tribe. Analysis of phylogenetic autocorrelation in Cichlinae showed strong body size correlation with phylogenetic history at the genus level (Fig. 1.5, I=0.7010, p<<0.05), however at more inclusive taxonomic levels (subclade, tribe) body size was not correlated with phylogenetic history. Species within a genus are more similar in body size to each other than expected at random, however at higher taxonomic levels body size may or may not be similar in closely related groups.

The BSFDs of all three major tribes were found to be significantly different from that of

Cichlinae and from each other using the K-S test (Table 1.2). The CAS clade was not significantly different from the BSDF of Geophagini while the GGD clade was found to be significantly different from both the BSFD of Geophagini and the CAS clade (Table 1.2). SA

32 heroines and CA heroines did not significantly differ from the BSFD of Heroini or each other based on the K-S test.

Figure 1.4: Body size occupation of major cichlid tribes. Body size frequency distributions for major tribes of Cichlinae: A) Geophagini, B) Cichlasomatini, C) Heroini. The coloured columns show observed data, black lines represent the distributions of 1000 random bootstrap subsamples of Cichlinae. Black arrows indicate potential body size optima supported by the bootstrap analyses.

33

Table 1.2: Partitioning of body size within Cichlinae and major clades. D statistic and corresponding significance values of Kolmogorov-Smirnov tests for differences in body size frequency distributions between and among phylogenetic levels. Results shown are for analyses at the tribe and major clade level. P-values in bold indicate significantly different comparisons, dashes indicate pairwise comparisons that were not made between clades of differing tribes (see text).

Cichlinae Geophagini Clade GGD Heroini SA Heroines CA Heroines Taxa D P D P D P D P D P D P Geophagini 0.1586 <0.001 ------Clade CAS - - 0.0686 0.7327 0.3372 0.0005 ------Clade GGD - - 0.2686 0.0081 ------Heroini 0.1885 <0.0005 0.342 <0.0001 - - - - 0.2108 0.3751 0.0053 1 Cichlasomatini 0.3147 <0.0001 0.2871 <0.0005 - - 0.5032 <0.0001 - - - - SA Heroines ------0.2133 0.3608

34

1.4.3 Divergence in Body Size Space

In addition to looking at phylogenetic autocorrelation of body size among taxonomic levels, the occupation of body size space in closely related taxa was also explored. Comparisons of the location of distributions in body size space as well as bootstrap analysis results of divergent taxa were made to determine if deviations in summary statistics supported different body space occupation. At the tribe and major clade level there is considerable overlap in body size space, with no groups showing complete separation of body size space occupation.

Within Geophagini there are several cases of divergence in body size space within closely related genera. Within the CAS clade, the distributions of Apistogramma and

Satanoperca do not show overlap in body size space (Fig. 1.3). and Acarichthys (Fig.

1.6), proposed sister-groups, also do not show overlap in body size space, though no bootstrap results support divergence greater than expected at random (Appendix 1.3). The sister-groups

Guianacara and Mazarunia occupy a narrow range of body size space around 100 mm (LBS

2.0) and do not show significant divergence from each other (Fig. 1.6). The small-bodied space dominated by Apistogramma is primarily shared with Teleocichla and some small-bodied

Crenicichla, while large body size space is equally shared between Satanoperca and

Crenicichla (Fig. 1.5). In the GGD clade, the sister-groups Dicrossus and Crenicara show significant divergence based on bootstrap results (Appendix 1.3) and non-overlapping distributions in body size space (Fig. 1.6). While the sister-groups Geophagus and

Gymnogeophagus show considerable overlap in body size space (Fig. 1.6), bootstrap results suggest that the range of body size in Gymnogeophagus is significantly reduced and overlaps only with the lower end of the Geophagus distribution (Appendix 1.3). In addition, the distribution of Geophagus occupies a narrower body size space, shifted towards large body

35 sizes. Smaller bodied taxa (Mikrogeophagus, Dicrossus, Crenicara; Fig. 1.5) show little or no overlap with the distributions of Geophagus and Gymnogeophagus while occupies a narrow range of body size space around 100 mm (LBS 2.0) (Fig. 1.6).

Figure 1.5: Phylogenetic autocorrelation of body size using taxonomic hierarchy as a proxy for phylogenetic relatedness using Moran’s I.

Body size distributions of genera within Cichlasomatini commonly overlap in closely related genera. Two exceptions occur between sister-groups Nannacara and Cleithracara as well as and Laetacara (Fig. 1.5). In both cases, no overlap is seen between taxa although these divergences are not supported strongly by the bootstrapping analyses (Appendix

1.3) due to the low species richness of these genera. In CA heroines, divergence was difficult to assess due to the paraphyletic and unresolved nature of many groups. However, two cases of divergence occur in South America between and its sister clade containing Symphysodon and as well as between Hoplarchus and (Fig. 1.5). No overlap is seen between taxa, but again the bootstrapping analyses do not support these divergences, likely due to low species richness (Appendix 1.3).

36

Figure 1.6: Body size diversity and occupation of Cichlinae. Distributions of body size of genera within Cichlinae following approximate phylogenetic order (See Fig. 1.1). Dots to the left and right of boxplots indicate outliers of the distribution.

37 1.5 Discussion

1.5.1 Divergence in Body Size Space

Morphological traits associated with diet in Neotropical cichlids are known to be more similar within genera than among genera (López-Fernández et al., 2012). Since body size is often strongly linked to diet and feeding (Romanuk et al., 2011) and morphology (Wainwright &

Richard, 1995) in fishes, body size would be expected to show this pattern of higher similarity within genera than among genera. A reduction of body size range and variation as compared to a random phylogenetic distribution was expected if species within a genus share more similar body sizes. However, this pattern was only consistent within the CAS clade of Geophagini, perhaps due to the specialized ecological roles within this clade compared to others (Arbour &

López-Fernández, 2013; López-Fernández et al., 2013) that is associated with particular body size regions (Table 1.1; Fig. 1.5). In addition, a low degree of body size overlap is expected among clades, consistent with ecological divergence and niche partitioning hypotheses (Mahler et al., 2010; Romanuk et al., 2011). Body size distributions of genera within and among major clades or tribes did not show a high amount of body size divergence. This was supported by results of the K-S test, bootstrap simulations, and analysis of phylogenetic autocorrelation. Most differences among genera occur with Apistogramma (Geophagini) or Cichla (Cichlini) (Fig.

1.5), which occupy extreme areas of small and large body size space, respectively, and which are rarely occupied by other genera. Body size can be an important determinant of niche partitioning or overlap (Woodward & Hildrew, 2002) as well as of trophic level and resource use (Romanuk et al., 2011), but among Cichlinae, it is not highly divergent in a strict phylogenetic context, and may in fact be more important within the context of community ecology and assembly (Griffiths, 2013; Woodward & Hildrew, 2002). Body size divergence in

38 distantly related taxa may allow partitioning of resources within a community by differing in resource use (e.g. microhabitat, prey type and size), while divergent ecologies in similarly sized cichlids may also allow for coexistence. Analyses at the community level will be needed to understand the role of body size in the ecological and geographic assembly of Neotropical cichlid fishes.

The distributions of SA and CA heroines were also not highly divergent (Fig. 1.5), but there was a trend towards lower diversity in SA heroines and smaller body size. Heroine species diversity is higher in Central America, which may be expected to increase body size diversity.

However, mean body size of CA heroines is higher than expected, suggesting that despite having more species, evolution in CA heroines may be biased towards body size increase. This trend may, in part, be influenced by the higher success of large-bodied species at dispersal

(Griffiths, 2010, 2012; Mahon, 1984; Winemiller & Rose, 1992) , and therefore the founders of the Central American colonization may have been primarily larger-bodied cichlids. The recent finding that Heroini body size may be evolving under an adaptive-peak model of evolution

(López-Fernández et al., 2013) suggests that novel environmental pressures or opportunities in

Central America could have also acted on dispersing cichlids to drive increases in CA Heroine body size (Collar et al., 2011; Thomas et al., 2009). Restricted body size occupation in SA heroines could be associated with ecological constraint as compared to CA heroines, particularly as a result of interaction with the much more diverse Geophagini (López-Fernández et al.,

2013). The diversification of ecological roles, particularly the evolution of piscivorous heroines in Central America, could be related to the origin of larger body size in middle American heroines (López-Fernández et al., 2013; Romanuk et al., 2011).

39 1.5.2 Body Size Diversity and the Evolution of Extreme Body Sizes

Body size diversity appears to be higher in Geophagini (Fig. 1.3), which has a higher standard deviation and IQR than expected (Table 1.1). However, Heroini spans the largest range of body size space while Cichlasomatini has the narrowest range. Despite having significantly larger body size, less than 4% of heroines have increased in body size beyond the largest Geophagine cichlids, and still do not occupy a unique body size space (Fig. 1.3). This extremely large body size space is also occupied by Cichla and Astronotus, large predatory cichlids from two different species-poor tribes (Cichlini and Astronotini, respectively). In contrast, heroine cichlids have not expanded into small bodied space to the extent that Geophagini has (Fig. 1.3). In

Geophagini, approximately 10% of species occupy a region of body size unoccupied by any other tribe while 25% of species occupy a region of body size unoccupied by heroine cichlids.

The genus Apistogramma (Fig. 1.5) not only dominates in the extent of body size reduction within Cichlinae as a whole (minimum BS 21 mm SL, LBS 1.32), but also in the total number of species that occupy this unique space (22 spp. below 36 mm SL, LBS 1.57; 53 spp. below 50 mm SL, LBS 1.70). Comparatively, only ten species from other groups also occupy this space.

The restriction of extremely small-bodied species to particular genera and conservation of body size within these groups suggests that morphological evolution may be constrained at smaller size, which could limit ecological opportunity within small-bodied taxa (Monroe &

Bokma, 2009; Romanuk et al., 2011). Though variance in body size was not correlated with body size (results not shown), in that large-bodied clades do not have higher body size diversity relative to small-bodied clades, evidence for ecomorphological constraint within these small- bodied taxa has been found in Geophagini. Small-bodied genera within Geophagini were found to converge on the same trophic functional morphospace below a body size threshold of 100

40 mm SL (Arbour & López-Fernández, 2013) (LBS 2.0, identified in Fig. 1.5 by a prominent dashed line for comparison). This threshold appears to be supported by the BSFD of the CAS clade, but results also suggest that at least three regions of body size space may better characterize body space occupation, each with possible size-specific ecological roles in

Neotropical cichlids.

Cichlinae is significantly left-skewed compared to other Neotropical fishes, with a higher proportion of large-bodied species. This finding is inconsistent with typical patterns found in tropical riverine fishes (Griffiths, 2012) where increasingly right-skewed distributions occur in lower latitudes. Below the subfamily level, clades did not typically deviate in skew and therefore extreme body size reduction is rare in Cichlinae. Subsequently, Geophagini offers a unique case to study the ecologies of small-bodied fishes prevalent in the Neotropics and consequences of extreme body size reduction. The most species-rich genus of Neotropical cichlids, Crenicichla (including Teleocichla) has the largest range of body size, from 39.8 mm

SL (LBS 1.60) to 312 mm SL (LBS 2.49). Though this range only occupies one third of the body size range of all Cichlinae and does not reach the lower and upper extremes, this range does encompass 75% of species within the BSFD of Cichlinae (Fig. 1.5). Crenicichla and

Teleocichla, though not as ecologically diverse as Geophagini, could offer a more practical group to study narrow questions of body size evolution in a strict phylogenetic, ecological, and geographic context.

1.5.3 Is Body Size Adaptive?

Recent studies of BSFDs in North American freshwater fishes found that bimodality in distributions was typically influenced by the presence of small-bodied, resident, habitat

41 specialists and large-bodied, migratory, habitat generalists (Griffiths, 2012). Recently, strong partitioning of ecologically meaningful body shape attributes (López-Fernández et al., 2013) and trophic functional morphospace (Arbour & López-Fernández, 2013) was found in Cichlinae and it is likely that this ecological differentiation is correlated with patterns in BSFDs across the group. The distribution of body size in the GGD clade of Geophagini was found to be located between the two modes of the CAS clade, producing a third mode of body size that suggests partitioning of total body size morphospace. The CAS clade began diversifying before all other species-rich clades within Cichlinae (López-Fernández et al., 2013) and may have constrained body size diversification within other lineages. In addition, the distributions of GGD and

Cichlasomatini are also located between the two modes of the CAS clade and show narrowing of body size space occupation (Fig. 1.5). SA heroines occupy a similar body size space to the large-bodied mode of the CAS clade, but have considerably less species diversity. This pattern of roughly complementary body size distribution among clades may further support the hypothesis that competition and ecological opportunity may be important in the diversification of traits, including body size, in South American cichlids. Analyses of ecomorphology (López-

Fernández et al., 2013) and biomechanics (Arbour & López-Fernández, 2013) have shown a high degree of diversity in Geophagini as compared to South American Heroini and

Cichlasomatini, a pattern consistent with findings in this chapter. Interestingly, the BSFD of CA heroines, which are geographically separated from other Neotropical cichlids, occupies the same body size space as the large-bodied CAS Geophagini and show comparable species richness

(Fig. 1.5). This result is also consistent with patterns found in body shape disparity (López-

Fernández et al., 2013), with heroines occupying non-overlapping morphospace with

Geophagini in South America, then expanding into these newly available ecomorphospace regions once geographically separated (Fig. 1.5).

42

The pattern of BSFDs complementarity at the major clade and tribe levels support a hypothesis of three possible body size optima in Cichlinae: a “miniature” body size optimum around 35 mm (SL), a “mid-sized” optimum around 100 mm (SL) and a “large-bodied” optimum around 250 mm (SL) (Fig. 1.2; Fig. 1.3). The CAS and GGD clades of Geophagini show modes around these three optima that are quantitatively supported as distinct from each other by the K-S tests and bootstrap analyses (Fig. 1.2, Tables 1.1 and 1.2), while the body sizes of Cichlasomatini and South American heroines form a unimodal distribution around the

100mm optimum (Figs. 1.5; Fig. 1.3). Finally, the CA heroines have a significantly right shifted but unimodal distribution around the large-bodied optimum of 250 mm SL (Fig. 1.5). These results are inconsistent with the idea of a phylogenetically directional pattern of evolution towards large body size, as proposed by Cope’s rule. Decreases in body size are a common phenomenon in Neotropical cichlids, and are present in both old and young clades, although the phylogenetic directionality of body size changes and number of independent reductions still need to be tested directly. Likewise, testing for body size optima was beyond the scope of this chapter as it should be done using phylogenies with complete or near-complete species-level sampling. Instead, until more detailed phylogenies become available, less phylogenetically robust analyses were employed incorporating all species with available body size data to determine patterns of body size within phylogenetic groups while accounting for as much body size variation in each group as possible. Directly testing of the association between body size, morphology and ecology for all Neotropical cichlids was not possible since data is largely unavailable for most species used in this chapter.

43 1.5.4 Conclusion

This chapter outlines several broad patterns in body size within a widely distributed group that is both taxonomically and ecologically diverse. Body size distributions of related clades show an unexpected amount of overlap, and an understanding of how body size is associated with geographic separation, community structure and ecological divergence may shed light on why

Neotropical cichlids so frequently occupy overlapping body size space. Cichlinae represents the most studied Neotropical freshwater fish lineage to date. With a strong knowledge of the evolutionary processes driving ecological diversification in the context of a robust understanding of the phylogenetic history of the group, hypotheses of diversification can be derived and tested in an explicit macroevolutionary context. Such an approach should reveal how body size evolves in Neotropical freshwater fishes and the ecological consequences or opportunities associated with body size space occupation.

44 1.6 Appendices

Appendix 1.1: Body size data of Neotropical cichlids. Cichlid species and taxonomic affiliation used within this chapter. Standard length and total length values (mm) from FishBase and log10 body size are provided. Asterisks indicated published SL measurements not available from FishBase (see Methods, López-Fernández et al. 2013). Tribe Species Body Log Body Size Size (mm) Astronotini Astronotus crassipinnis 240 SL 2.380 Astronotus ocellatus 457 TL 2.660 Chaetobranchini Chaetobranchopsis australis 120 SL 2.079 Chaetobranchopsis orbicularis 120 SL 2.079 Chaetobranchus flavescens 210 SL 2.415 Chaetobranchus semifasciatus 230 SL 2.362 Cichlasomatini sapayensis 100 TL 2.000 Andinoacara stalsbergi 113 SL 2.053 Bujurquina apoparuana 77 SL 1.886 Bujurquina cordemadi 62 SL 1.792 Bujurquina eurhinus 89 SL 1.949 Bujurquina hophrys 112* SL 2.049 Bujurquina huallagae 83 SL 1.919 Bujurquina labiosa 55 SL 1.740 Bujurquina mariae 150 TL 2.176 Bujurquina megalospilus 71 SL 1.851 Bujurquina moriorum 97 SL 1.987 Bujurquina oenolaemus 67 SL 1.826 Bujurquina ortegai 110 SL 2.041 Bujurquina peregrinabunda 107 SL 2.029 Bujurquina robustus 88 SL 1.944 Bujurquina syspilus 103 SL 2.013 Bujurquina tambopatae 82 SL 1.914 Bujurquina vittata 90 SL 1.954 Bujurquina zamorensis 74 SL 1.869 Tahuantinsuyoa chipi 82 SL 1.914 Tahuantinsuyoa macantzatza 120 TL 2.079 biseriatus 80 SL 1.903 Aequidens chimantanus 102 SL 2.009 Aequidens coeruleopunctatus 145 TL 2.161 Aequidens diadema 118 SL 2.072 Aequidens epae 113 SL 2.053

45

Aequidens gerciliae 128 SL 2.107 Aequidens hoehnei 56 SL 1.748 Aequidens latifrons 170 TL 2.230 Aequidens mauesanus 134 SL 2.127 Aequidens metae 125 SL 2.097 Aequidens pallidus 143 SL 2.155 Aequidens paloemeuensis 95 SL 1.978 Aequidens patricki 116 SL 2.064 Aequidens plagiozonatus 103 SL 2.013 Aequidens potaroensis 100 SL 2.000 Aequidens pulcher (Andinoacara) 107* SL 2.029 Aequidens rivulatus 200 TL 2.301 Aequidens rondoni 97 SL 1.987 Aequidens tetramerus 162 SL 2.210 Aequidens tubicen 116 SL 2.064 Aequidens viridis 165 SL 2.217 Cichlasoma amazonarum 114 SL 2.057 Cichlasoma araguaiense 92 SL 1.964 Cichlasoma bimaculatum 123 SL 2.090 Cichlasoma boliviense 107 SL 2.029 Cichlasoma dimerus 117 SL 2.068 Cichlasoma orientale 136 SL 2.134 Cichlasoma orinocense 109 SL 2.037 Cichlasoma paranaense 74 SL 1.869 Cichlasoma portalegrense 103 SL 2.013 Cichlasoma pusillum 66 SL 1.820 Cichlasoma sanctifranciscense 79 SL 1.898 Cichlasoma taenia 128 SL 2.107 Cichlasoma zarskei 100 SL 2.000 Krobia guianensis 128 SL 2.107 Krobia itanyi 125 SL 2.097 Krobia xinguensis 128* SL 2.107 Acaronia nassa 154 SL 2.188 Acaronia vultuosa 122 SL 2.086 Laetacara araguaiae 36 SL 1.556 Laetacara curviceps 46 SL 1.663 Laetacara dorsigera 60 SL 1.778 Laetacara flavilabris 82 SL 1.914 Laetacara fulcipinnis 74 SL 1.869 Laetacara thayeri 69* SL 1.839 Cleithracara maronii 71 SL 1.851 Nannacara adoketa 49 SL 1.690 Nannacara anomala 56 SL 1.748 Nannacara aureocephalus 67 SL 1.826 Nannacara bimaculata 45 SL 1.653

46

Nannacara taenia 50 SL 1.699 Cichlini 375 SL 2.740 Cichla jariina 340 SL 2.531 Cichla kelberi 276 SL 2.441 Cichla melaniae 290 SL 2.462 Cichla mirianae 520 SL 2.716 330 SL 2.845 Cichla nigromaculata 263 SL 2.420 Cichla ocellaris 740 TL 2.869 750* SL 2.790 Cichla pinima 520 SL 2.716 Cichla piquiti 430 SL 2.633 Cichla pleiozona 340 SL 2.531 Cichla temensis 990 TL 2.996 Cichla thyrorus 430 SL 2.633 Cichla vazzoleri 410 SL 2.613 Geophagini- Apistogramma agassizii 43* SL 1.633 CAS Clade Apistogramma alacrina 55 SL 1.740 Apistogramma amoena 63 TL 1.799 Apistogramma angayuara 25 SL 1.398 Apistogramma arua 44 SL 1.643 Apistogramma atahualpa 42 SL 1.623 Apistogramma baenschi 53 SL 1.724 64 SL 1.806 Apistogramma bitaeniata 46 SL 1.663 Apistogramma borelli 39 SL 1.591 Apistogramma brevis 39 SL 1.591 Apistogramma cacatuoides 50 SL 1.699 Apistogramma caetei 36 SL 1.556 Apistogramma cinilabra 54 SL 1.732 Apistogramma commbrae 35 TL 1.519 Apistogramma cruzi 51 SL 1.708 Apistogramma diplotaenia 29 SL 1.462 Apistogramma elizabethae 40 SL 1.602 Apistogramma emnopygae 34 SL 1.531 Apistogramma erythrura 140 SL 2.146 Apistogramma eunotus 53 SL 1.724 Apistogramma geisleri 28 SL 1.447 Apistogramma gephyra 33 SL 1.519 Apistogramma gibbiceps 45 SL 1.653 Apistogramma gossei 44 SL 1.643 Apistogramma guttata 36 SL 1.556 Apistogramma hippolytae 34 SL 1.531 Apistogramma hoignei 60 SL 1.778 Apistogramma hongsloi 34 SL 1.531

47

Apistogramma huascar 45 SL 1.653 Apistogramma inconspicua 38 SL 1.580 Apistogramma iniridae 36 SL 1.556 Apistogramma juruensis 24 SL 1.380 Apistogramma linkei 39 SL 1.591 Apistogramma luelingi 33 SL 1.519 55 SL 1.740 Apistogramma megaptera 59 SL 1.771 Apistogramma meinkeni 35 SL 1.544 Apistogramma moae 50 SL 1.699 39 SL 1.591 Apistogramma norberti 39 SL 1.591 Apistogramma ortmanni 41 SL 1.613 49 49 1.690 Apistogramma pantalone 36 SL 1.556 Apistogramma paucisquamis 34 SL 1.531 Apistogramma payaminonis 40 SL 1.602 Apistogramma personata 49 SL 1.690 Apistogramma pertensis 39 SL 1.591 Apistogramma piauensis 23 SL 1.362 Apistogramma playayacu 49 SL 1.690 Apistogramma pleurotaenia 28 SL 1.447 Apistogramma pulchra 32 SL 1.505 Apistogramma regani 49 SL 1.690 Apistogramma resticulosa 27 SL 1.431 Apistogramma rositae 40 SL 1.602 Apistogramma rubrolineata 40 SL 1.602 Apistogramma rupununi 38 SL 1.580 Apistogramma salpinction 33 SL 1.519 Apistogramma staecki 21 SL 1.322 Apistogramma steindachneri 65 SL 1.813 Apistogramma taeniata 42 SL 1.623 Apistogramma trifasciata 38 SL 1.580 Apistogramma uaupesi 28 SL 1.447 Apistogramma urteagai 41 SL 1.613 30 SL 1.477 Apistogramma wapisana 29 SL 1.462 Apistogrammoides pucallpaensis 27 SL 1.431 (Apistogramma) Satanoperca acuticeps 170 SL 2.230 Satanoperca daemon 190 SL 2.279 201* SL 2.303 Satanoperca leucosticta 160* SL 2.204 Satanoperca lilith 255 SL 2.407 Satanoperca mapiritensis 164* SL 2.215

48

Satanoperca pappaterra 174 SL 2.241 Satanoperca rhynchitis 141 SL 2.149 33 SL 1.519 Acarichthys heckelii 134 SL 2.127 Biotoecus dicentrarchus 38 SL 1.580 Biotoecus opercularis 100 TL 2.000 Crenicichla acutirostris 230 TL 2.362 Crenicichla adspersa 290 SL 2.462 Crenicichla albopunctata 140 SL 2.146 Crenicichla alta 160 SL 2.204 Crenicichla anthurus 224 SL 2.350 Crenicichla brasiliensis 78 SL 1.892 Crenicichla britskii 145 SL 2.161 Crenicichla cametana 183 SL 2.262 Crenicichla celidochilus 268 TL 2.428 Crenicichla cincta 195 SL 2.290 Crenicichla compressiceps 55 SL 1.740 Crenicichla coppenamensis 179 SL 2.253 Crenicichla cyanonotus 148 SL 2.170 Crenicichla cyclostoma 96 SL 1.982 Crenicichla empheres 141 SL 2.149 Crenicichla frenata 116 SL 2.064 Crenicichla gaucho 127 SL 2.104 Crenicichla geayi 136* SL 2.134 Crenicichla hadrostigma 114 SL 2.057 Crenicichla haroldoi 98 SL 1.991 Crenicichla heckeli 52 SL 1.716 Crenicichla hemera 97 SL 1.987 Crenicichla hu 153 SL 2.185 Crenicichla hummelincki 108 SL 2.033 Crenicichla igara 312 TL 2.494 Crenicichla iguapina 176 SL 2.246 Crenicichla iguassuensis 140 SL 2.146 Crenicichla inpa 168 SL 2.225 Crenicichla isbrueckeri 95 SL 1.978 Crenicichla jaguarensis 148 SL 2.170 Crenicichla jegui 200 SL 2.301 Crenicichla johanna 283 SL 2.452 Crenicichla jupiaensis 82 SL 1.914 Crenicichla jurubi 303 TL 2.481 Crenicichla labrina 160 SL 2.204 Crenicichla lacustris 290 SL 2.462 Crenicichla lenticulata 300 SL 2.477 Crenicichla lepidota 180 SL 2.255 Crenicichla lucius 168 SL 2.225

49

Crenicichla lugubris 260* SL 2.415 Crenicichla macropthalma 200 SL 2.301 Crenicichla maculata 247 SL 2.393 Crenicichla mandelburgeri 115 SL 2.061 Crenicichla marmorata 280 SL 2.447 Crenicichla menezesi 146 SL 2.164 Crenicichla minuano 260 TL 2.415 Crenicichla missioneira 283 TL 2.452 Crenicichla mucuryna 113 SL 2.053 Crenicichla multispinosa 228* SL 2.358 Crenicichla nickeriensis 191 SL 2.281 Crenicichla niederleinii 235 SL 2.371 Crenicichla notophthalmus 76 SL 1.881 Crenicichla pellegrini 157 SL 2.196 Crenicichla percna 220 SL 2.342 Crenicichla phaiospilus 240 SL 2.380 Crenicichla prenda 87 SL 1.940 Crenicichla proteus 155 SL 2.190 Crenicichla punctata 223 SL 2.348 Crenicichla pydanielae 178 SL 2.250 Crenicichla regani 79 SL 1.898 Crenicichla reticulata 216 SL 2.334 Crenicichla rosemariae 244 SL 2.387 Crenicichla santosi 120 SL 2.079 Crenicichla saxatilis 200 SL 2.301 Crenicichla scottii 169 SL 2.228 Crenicichla sedentaria 221 SL 2.344 Crenicichla semicincta 171 SL 2.233 Crenicichla semifasciata 150 SL 2.176 Crenicichla sipaliwini 173 SL 2.238 Crenicichla stocki 250 TL 2.398 Crenicichla strigata 300 SL 2.477 Crenicichla sveni 193* SL 2.286 Crenicichla tendyhaguassu 152 SL 2.182 Crenicichla ternetzi 245 SL 2.389 Crenicichla tigrina 280 SL 2.447 Crenicichla tingui 195 SL 2.290 Crenicichla urosema 68 SL 1.833 Crenicichla vaillanti 126 SL 2.100 Crenicichla virgatula 66 SL 1.820 Crenicichla vittata 260 SL 2.415 Crenicichla wallacii 85 TL 1.929 Crenicichla yaha 146 SL 2.164 Crenicichla ypo 137 SL 2.137 Crenicichla zebrina 264 SL 2.422

50

Teleocichla centisquama 40 SL 1.602 Teleocichla centrarchus 60 SL 1.778 Teleocichla cinderella 54 SL 1.732 Teleocichla gephyrogramma 57* SL 1.756 Teleocichla monogramma 63 SL 1.799 Teleocichla prionogenys 57 SL 1.756 Teleocichla proselytus 57 SL 1.756 cuyunii 75 SL 1.875 Guianacara dacrya 120 SL 2.079 Guianacara geayi 85 SL 1.929 Guianacara oelemariensis 81 SL 1.908 Guianacara owroewefi 107 SL 2.029 Guianacara sphenozona 85 SL 1.929 Guianacara stergiosi 80 SL 1.903 77* SL 1.886 Mazarunia pala 74 SL 1.869 Mazarunia charadrica 84 SL 1.924 Geophagini – 97 SL 1.987 GGD Clade 105* SL 2.019 Crenicara latruncaularium 89 SL 1.949 100 SL 2.000 38 SL 1.580 Dicrossus foirni 71 SL 1.851 Dicrossus gladicauda 42 SL 1.623 Dicrossus maculatus 53 SL 1.724 Dicrossus warzeli 60 SL 1.778 Geophagus abalios 211* SL 2.324 225 SL 2.352 Geophagus argyrostictus 180 SL 2.255 Geophagus brachybranchus 138 SL 2.140 Geophagus brokopondo 123 SL 2.090 Geophagus camopiensis 120 SL 2.079 Geophagus crassilabris 240 SL 2.380 Geophagus dicrozoster 205* SL 2.312 Geophagus gottwaldi 198 SL 2.297 Geophagus grammapareius 103 SL 2.013 Geophagus harreri 183 SL 2.262 Geophagus iporangensis 100 SL 2.000 Geophagus megasema 175 SL 2.243 Geophagus neambi 127 SL 2.104 Geophagus obscurus 100 SL 2.000 Geophagus parnaibae 76 SL 1.881 Geophagus pellegrini 152 SL 2.182 Geophagus proximus 225 SL 2.352 Geophagus steindachneri 198 TL 2.297

51

Geophagus surinamensis 148 SL 2.170 167 SL 2.223 Geophagus taeniopareius 143 SL 2.155 Geophagus winemilleri 195 SL 2.290 Gymnogeophagus australis 155 TL 2.190 Gymnogeophagus caaguazuensis 86 SL 1.934 Gymnogeophagus che 116 SL 2.064 Gymnogeophagus gymnogenys 150 SL 2.176 Gymnogeophagus labiatus 120 SL 2.079 Gymnogeophagus lacustris 146 SL 2.164 Gymnogeophagus meridionalis 88 SL 1.944 Gymnogeophagus rhabdotus 120 SL 2.079 Gymnogeophagus setequedas 98 SL 1.991 Gymnogeophagus tiraparae 100 SL 2.000 Gynmogeophagus balzanii 120 SL 2.079 ‘Geophagus’ brasiliensis 280 TL 2.447 Mikrogeophagus altispinosus 56 SL 1.748 Mikrogeophagus ramirezi 34 SL 1.531 Heroini- "Cichlasoma" calobrensis () 250 TL 2.398 Central America "Cichlasoma" lyonsi (Amphilophus) 150 SL 2.176 "Cichlasoma" wesseli () 80 SL 1.903 coatepeque 91 SL 1.959 Amatitlania kanna 83 SL 1.919 Amatitlania nigrofasciata 100 SL 2.000 Amatitlania siquia 79 SL 1.898 Amphilophus alfari 150 SL 2.176 Amphilophus altifrons 130 SL 2.114 Amphilophus amarillo 155 SL 2.190 Amphilophus astorquii 155 SL 2.190 Amphilophus bussingi 150 SL 2.176 Amphilophus chancho 242 SL 2.384 244 SL 2.387 Amphilophus flaveolus 136 SL 2.134 Amphilophus globosus 165 SL 2.217 Amphilophus hogaboomorum 150 SL 2.176 Amphilophus labiatus 240 SL 2.380 Amphilophus margaritifer 127 SL 2.104 Amphilophus nourissati (Theraps) 142* SL 2.152 Amphilophus rhytisma 135 SL 2.130 Amphilophus sagittae 160 SL 2.204 Amphilophus supercilius 168 SL 2.225 Amphilophus xiloaensis 160 SL 2.204 Amphilophus zaliosus 200 TL 2.301 Archocentrus centrarchus 110 TL 2.041

52

Archocentrus spinosissimus 110 SL 2.041 acaroides 120 SL 2.079 Australoheros angiru 77 SL 1.886 Australoheros autrani 73 SL 1.863 Australoheros barbosae 96 SL 1.982 Australoheros capixaba 57 SL 1.756 Australoheros charrua 78 SL 1.892 193 SL 2.286 Australoheros forquilha 110 SL 2.041 Australoheros guarani 129 SL 2.111 Australoheros ipatinguensis 53 SL 1.724 Australoheros kaaygua 94 SL 1.973 Australoheros macacuensis 83 SL 1.919 Australoheros macaensis 74 SL 1.869 Australoheros mattosi 80 SL 1.903 Australoheros minuano 84 SL 1.924 Australoheros montanus 103 SL 2.013 Australoheros muriae 121 SL 2.083 Australoheros paraibae 61 SL 1.785 Australoheros perdi 167 TL 2.223 Australoheros ribeirae 76 SL 1.881 Australoheros robustus 74 SL 1.869 Australoheros saquarema 80 SL 1.903 Australoheros scitulus 89 SL 1.949 Australoheros taura 124 SL 2.093 Australoheros tavaresi 79 SL 1.898 Australoheros tembe 134 SL 2.127 Australoheros ykeregua 137 SL 2.137 Cichlasoma salvini 220 SL 2.342 Cichlasoma trimaculatus 365 SL 2.562 altoflavus 90 SL 1.954 Cryptoheros chetumalensis 97 SL 1.987 Cryptoheros cutteri 112 SL 2.049 Cryptoheros myrnae 80 SL 1.903 Cryptoheros nanoluteus 70* SL 1.845 Cryptoheros panamensis 130 TL 2.114 Cryptoheros sajica 90 SL 1.954 Cryptoheros septemfasciatus 100 SL 2.000 Cryptoheros spilurus 97 TL 1.987 nematopus 140 SL 2.146 165 SL 2.217 Parachromis dovii 500 SL 2.699 Parachromis friedrichsthalii 280 SL 2.447 Parachromis loisellei 185 SL 2.267 Parachromis managuensis 220 SL 2.342

53

Parachromis motaguensis 300 TL 2.477 500 SL 2.699 diquis 135 SL 2.130 Astatheros longimanus 135 SL 2.130 Astatheros macracanthus 250 SL 2.398 Astatheros ocotal 96 SL 1.982 Astatheros octofasciata 250 TL 2.398 Astatheros robertsoni 190 SL 2.279 Astatheros rostratus 185 SL 2.267 Astatheros gemmata 70 SL 1.845 kraussii 260 SL 2.415 Caquetaia myersi 190 SL 2.279 Caquetaia spectabilis 165 SL 2.217 Caquetaia umbrifera 475 SL 2.677 102 SL 2.009 Cichlasoma aguadae 95 SL 1.978 Cichlasoma alborum 186 SL 2.270 Cichlasoma almarum 158 SL 2.199 Cichlasoma atromaculatum 250 TL 2.398 Cichlasoma beani 300 TL 2.477 Cichlasoma cienagae 113 SL 2.053 Cichlasoma conchitae 64 SL 1.806 Cichlasoma ericymba 122 SL 2.086 Cichlasoma geddesi 65 TL 1.813 Cichlasoma gephyrum 120 SL 2.079 Cichlasoma istlanum 300 TL 2.477 Cichlasoma mayorum 96 SL 1.982 Cichlasoma microlepis 187 SL 2.272 Cichlasoma ornatum 260 SL 2.415 Cichlasoma pearsei 200 SL 2.301 Cichlasoma stenozonum 110 TL 2.041 Cichlasoma troschelii 160 SL 2.204 Cichlasoma tuyrense 235 SL 2.371 Cichlasoma ufermanni 250 SL 2.398 Cichlasoma urophthalmum 300* SL 2.477 Cichlasoma zebra 104 SL 2.017 ‘Cichlasoma' festae 250 TL 2.398 Cichlasoma grammodes 203 SL 2.307 bartoni 180 SL 2.255 Herichthys carpintis 176* SL 2.246 Herichthys cyanoguttatus 300 TL 2.477 Herichthys deppii 120 TL 2.079 Herichthys labridens 250 SL 2.398 Herichthys minckleyi 175 SL 2.243 Herichthys pantostictus 126 SL 2.100

54

Herichthys steindachneri 400 SL 2.602 Herichthys tamasopoensis 180 TL 2.255 argenteus 270 TL 2.431 Paraneetroplus bifasciatus 300 SL 2.477 Paraneetroplus breidohri 168 SL 2.225 Paraneetroplus bulleri 255 SL 2.407 Paraneetroplus fenestratus 250 SL 2.398 Paraneetroplus gibbiceps 230 SL 2.362 Paraneetroplus guttulatus 300 TL 2.477 Paraneetroplus hartwegi 131 SL 2.117 Paraneetroplus maculicauda 250 SL 2.398 Paraneetroplus melanurus 190 TL 2.279 Paraneetroplus nebuliferus 200 TL 2.301 Paraneetroplus regani 230 SL 2.362 Paraneetroplus synspilus 350 TL 2.544 araneetroplus zonatus 250 TL 2.398 Theraps bocourti 200 SL 2.301 Theraps coeruleus 120 SL 2.079 Theraps godmanni 300 TL 2.477 Theraps heterospilus 240 SL 2.380 Theraps intermedius 200 SL 2.301 Theraps irregularis 190 SL 2.279 Theraps lentiginosus 250 SL 2.398 Theraps microphthalmus 250 SL 2.398 affinis 140 SL 2.146 Thorichthys aureus 150 S/l 2.176 Thorichthys callolepis 140 SL 2.146 Thorichthys ellioti 150 SL 2.176 Thorichthys helleri 145 SL 2.161 Thorichthys meeki 170 TL 2.230 Thorichthys pasionis 170 SL 2.230 Thorichthys socolofi 79 SL 1.898 haitiensis 215 SL 2.332 Nandopsis ramsdeni 240 SL 2.380 Nandopsis tetracanthus 200 SL 2.301 Archocentrus multispinosus (Herotilapia) 170 TL 2.230 asfraci 250 SL 2.398 Tomocichla sieboldii 250 SL 2.398 Tomocichla tuba 300 TL 2.477 Heroini – Hoplarchus psittacus 235* SL 2.371 South America Hypselecara coryphaenoides 160 SL 2.204 Hypselecara temporalis 150 SL 2.176 altum 180 TL 2.255 Pterophyllum leopoldi 50 SL 1.699 Pterophyllum scalare 75 SL 1.875

55

Heros efasciatus 140 SL 2.146 150 SL 2.176 Heros severus 200 SL 2.301 Heros spurius 120 SL 2.079 acora 71 SL 1.851 82 SL 1.914 82 SL 1.914 Mesonauta guyanae 100 SL 2.000 94 SL 1.973 97 SL 1.987 Symphysodon aequifasciatus 137 SL 2.137 Symphysodon 123 SL 2.090 Symphysodon tarzoo 132 SL 2.121 Uaru amphiacanthoides 250 SL 2.398 Uaru fernandezyepezi 190 SL 2.279 Retroculini Retroculus lapidifer 203 SL 2.307 Retroculus septentrionalis 190 SL 2.279 Retroculus xinguensis 144 SL 2.158

56

Appendix 1.2: Summary statistics of Cichlinae tribe and subclade body size distributions and expectation under random phylogenetic distribution. Comparison of clade observed summary statistics to distributions of summary statistics describing 1000 random simulated data from all cichlids (expected values, see Methods) for tribes, and respective tribe for subclade comparison. Upper and lower significance thresholds of simulated expected distributions are provided for P- values <0.05 (2.50%-97.50%) and <0.005 (0.25%-99.75%). Observed values that significantly differed from simulated statistics are given in bold.

Lower Upper Lower Upper Statistic Observed Tail Tail Tail Tail Comparisons at higher taxonomic levels

Geophagini – All Cichlids

2.50% 97.50% 0.25% 99.75% Observed Mean 2.052 2.108 2.041 2.120 1.970 St Dev 0.270 0.306 0.262 0.312 0.315 Minimum 1.362 1.505 1.362 1.519 1.322 Maximum 2.740 2.996 2.699 2.996 2.494 25% Quantile 1.847 1.920 1.818 1.948 1.690 75% Quantile 2.249 2.301 2.230 2.341 2.237 Kurtosis -0.633 0.076 -0.765 0.234 -1.252 Skew -0.233 0.180 -0.317 0.282 -0.157 IQR 0.355 0.442 0.333 0.464 0.547

CAS – Geophagini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.949 1.992 1.941 1.999 1.944 St Dev 0.305 0.323 0.303 0.327 0.331 Minimum 1.322 1.380 1.322 1.398 1.322 Maximum 2.477 2.494 2.477 2.494 2.494 25% Quantile 1.643 1.724 1.627 1.730 1.623 75% Quantile 2.217 2.255 2.198 2.262 2.343 Kurtosis -1.330 -1.157 -1.359 -1.113 -1.381 Skew -0.256 -0.059 -0.296 -0.020 0.005 IQR 0.515 0.594 0.507 0.608 0.611

GGD – Geophagini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.891 2.053 1.862 2.080 2.073 St Dev 0.276 0.348 0.261 0.357 0.215 Minimum 1.322 1.519 1.322 1.531 1.531 Maximum 2.415 2.494 2.362 2.494 2.447 25% Quantile 1.594 1.856 1.566 1.926 1.988 75% Quantile 2.141 2.309 2.075 2.349 2.238 Kurtosis -1.532 -0.811 -1.624 -0.644 -0.105

57

Skew -0.569 0.231 -0.696 0.332 -0.621 IQR 0.401 0.664 0.348 0.692 0.250

Cichlasomatini – All Cichlids

2.50% 97.50% 0.25% 99.75% Observed Mean 2.026 2.143 1.994 2.169 1.976 St Dev 0.247 0.328 0.234 0.340 0.153 Minimum 1.362 1.602 1.362 1.658 1.556 Maximum 2.481 2.996 2.457 2.996 2.301 25% Quantile 1.773 1.985 1.743 2.009 1.869 75% Quantile 2.204 2.366 2.177 2.398 2.081 Kurtosis -1.010 0.514 -1.200 1.180 -0.206 Skew -0.472 0.433 -0.614 0.585 -0.467 IQR 0.298 0.493 0.260 0.548 0.212

Heroini – All Cichlids

2.50% 97.50% 0.25% 99.75% Observed Mean 2.048 2.115 2.030 2.128 2.179 St Dev 0.262 0.310 0.255 0.320 0.208 Minimum 1.362 1.519 1.362 1.544 1.699 Maximum 2.716 2.996 2.660 2.996 2.699 25% Quantile 1.820 1.933 1.781 1.971 2.012 75% Quantile 2.232 2.335 2.229 2.354 2.347 Kurtosis -0.697 0.211 -0.827 0.415 -0.566 Skew -0.280 0.237 -0.390 0.346 -0.014 IQR 0.344 0.454 0.324 0.483 0.335

Heroini – SA Clade

2.50% 97.50% 0.25% 99.75% Observed Mean 2.098 2.262 2.054 2.297 2.093 St Dev 0.160 0.258 0.144 0.277 0.182 Minimum 1.699 1.919 1.699 1.985 1.699 Maximum 2.398 2.699 2.371 2.699 2.398 25% Quantile 1.903 2.146 1.881 2.187 1.973 75% Quantile 2.204 2.398 2.149 2.447 2.204 Kurtosis -1.539 0.252 -1.691 1.124 -0.816 Skew -0.641 0.695 -0.912 1.054 -0.218 IQR 0.165 0.428 0.136 0.478 0.231

Heroini – CA Clade

2.50% 97.50% 0.25% 99.75% Observed Mean 2.167 2.192 2.161 2.196 2.192 St Dev 0.200 0.214 0.196 0.216 0.210 Minimum 1.699 1.724 1.699 1.771 1.724 Maximum 2.699 2.699 2.677 2.699 2.699

58

25% Quantile 2.000 2.041 1.987 2.051 2.041 75% Quantile 2.301 2.371 2.301 2.380 2.380 Kurtosis -0.732 -0.421 -0.793 -0.359 -0.627 Skew -0.107 0.110 -0.157 0.164 -0.008 IQR 0.291 0.367 0.266 0.375 0.339

Astronotini – All Cichlids

2.50% 97.50% 0.25% 99.75% Observed Mean 1.681 2.462 1.566 2.701 2.520 St Dev 0.008 0.627 0.000 0.774 0.198 Minimum 1.447 2.393 1.362 2.488 2.380 Maximum 1.748 2.740 1.597 2.996 2.660 25% Quantile 1.592 2.406 1.502 2.585 2.450 75% Quantile 1.721 2.572 1.591 2.848 2.590 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Skew 0.000 0.000 0.000 0.000 0.000 IQR 0.006 0.443 0.000 0.547 0.140

Chaetobranchini – All Cichlids

2.50% 97.50% 0.25% 99.75% Observed Mean 1.827 2.345 1.711 2.465 2.211 St Dev 0.082 0.494 0.038 0.578 0.153 Minimum 1.398 2.146 1.362 2.311 2.079 Maximum 2.041 2.845 1.778 2.996 2.362 25% Quantile 1.599 2.260 1.534 2.377 2.079 75% Quantile 1.924 2.493 1.759 2.701 2.332 Kurtosis -2.384 -1.705 -2.425 -1.689 -2.413 Skew -0.709 0.666 -0.743 0.743 0.022 IQR 0.070 0.599 0.034 0.845 0.253

Cichlini – All Cichlids

2.50% 97.50% 0.25% 99.75% Observed Mean 1.936 2.214 1.895 2.270 2.630 St Dev 0.195 0.383 0.163 0.424 0.176 Minimum 1.380 1.845 1.362 1.939 2.420 Maximum 2.334 2.996 2.232 2.996 2.996 25% Quantile 1.679 2.091 1.605 2.163 2.522 75% Quantile 2.098 2.410 2.024 2.486 2.695 Kurtosis -1.589 0.667 -1.720 1.986 -0.810 Skew -0.866 0.843 -1.244 1.411 0.713 IQR 0.183 0.587 0.140 0.677 0.174

Retroculini-All Cichlids

2.50% 97.50% 0.25% 99.75% Observed Mean 1.742 2.376 1.600 2.523 2.248

59

St Dev 0.051 0.564 0.015 0.666 0.079 Minimum 1.431 2.231 1.362 2.301 2.158 Maximum 1.914 2.845 1.699 2.996 2.307 25% Quantile 1.597 2.289 1.488 2.416 2.219 75% Quantile 1.829 2.542 1.654 2.699 2.293 Kurtosis -2.333 -2.333 -2.333 -2.333 -2.333 Skew -0.384 0.383 -0.385 0.385 -0.329 IQR 0.048 0.526 0.014 0.652 0.075

60

Appendix 1.3: Summary statistics of cichlid genera body size distributions and expectation under random phylogenetic distribution. Comparison of clade observed summary statistics to distributions of summary statistics describing 1000 random simulated data from containing clades (expected values, see Methods) at higher taxonomic levels (subclade and tribe). Upper and lower significance thresholds of simulated expected distributions are provided for P-values <0.05 (2.50%-97.50%) and <0.005 (0.25%-99.75%). Observed values that significantly differed from simulated statistics are given in bold.

Lower Upper Upper Lower Lower Statistic Lower Tail Observed Statistic Upper Tail Upper Tail Observed Tail Tail Tail Tail Tail Genera Comparisons with simulations using Major Clade (if applicable) and Tribe body size distributions

Apistogramma-CAS Apistogramma-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.881 2.002 1.854 2.032 1.600 Mean 1.906 2.031 1.883 2.060 1.600 St Dev 0.305 0.357 0.293 0.365 0.131 St Dev 0.287 0.343 0.278 0.352 0.131 Minimum 1.322 1.447 1.322 1.491 1.322 Minimum 1.322 1.462 1.322 1.519 1.322 Maximum 2.452 2.494 2.428 2.494 2.146 Maximum 2.447 2.494 2.411 2.494 2.146 25% Quantile 1.591 1.724 1.562 1.758 1.525 25% Quantile 1.602 1.789 1.588 1.885 1.525 75% Quantile 2.163 2.296 2.141 2.338 1.690 75% Quantile 2.167 2.296 2.147 2.310 1.690 Kurtosis -1.555 -1.186 -1.598 -1.063 2.926 Kurtosis -1.497 -0.939 -1.565 -0.697 2.926 Skew -0.261 0.307 -0.376 0.447 0.875 Skew -0.459 0.142 -0.567 0.290 0.875 IQR 0.511 0.676 0.486 0.707 0.165 IQR 0.448 0.644 0.383 0.683 0.165

Biotodoma-GGD Biotodoma-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.759 2.319 1.640 2.369 2.003 Mean 1.553 2.347 1.451 2.438 2.003 St Dev 0.006 0.478 0.000 0.581 0.023 St Dev 0.010 0.627 0.000 0.719 0.023 Minimum 1.531 2.290 1.531 2.352 1.987 Minimum 1.380 2.297 1.322 2.407 1.987 Maximum 1.851 2.447 1.724 2.447 2.019 Maximum 1.591 2.463 1.525 2.494 2.019 25% Quantile 1.654 2.300 1.586 2.356 1.995 25% Quantile 1.506 2.314 1.388 2.418 1.995 75% Quantile 1.819 2.361 1.688 2.408 2.011 75% Quantile 1.573 2.399 1.504 2.460 2.011 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Skew -9E-15 8.48E-15 -3.7E-14 5.21E-14 0.000 Skew -7.4E-15 5.09E-15 -2.7E-14 6.64E-14 0.000 IQR 0.004 0.338 0 0.411 0.016 IQR 0.007 0.443 0.000 0.508 0.016

Biotoecus-GGD Biotoecus-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.775 2.335 1.652 2.400 1.790 Mean 1.553 2.347 1.451 2.438 1.790 St Dev 0.009 0.476 0.000 0.556 0.297 St Dev 0.010 0.627 0.000 0.719 0.297 Minimum 1.531 2.297 1.531 2.352 1.580 Minimum 1.380 2.297 1.322 2.407 1.580 Maximum 1.933 2.447 1.724 2.447 2.000 Maximum 1.591 2.463 1.525 2.494 2.000 25% Quantile 1.668 2.315 1.616 2.376 1.685 25% Quantile 1.506 2.314 1.388 2.418 1.685 75% Quantile 1.857 2.359 1.688 2.423 1.895 75% Quantile 1.573 2.399 1.503 2.460 1.895

61

Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Kurtosis -2.750 -2.750 -2.750 0.000 -2.75 Skew -9.6E-15 8.49E-15 -4.1E-14 3.16E-14 0.000 Skew -7.4E-15 5.09E-15 -2.7E-14 6.64E-14 0.000 IQR 0.006 0.337 0 0.393 0.210 IQR 0.007 0.443 0.000 0.508 0.210108

Crenicara-GGD Crenicara-Geophaginni

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.740 2.338 1.602 2.400 1.975 Mean 1.553 2.347 1.451 2.438 1.975 St Dev 0.007 0.473 0.000 0.598 0.0358 St Dev 0.010 0.627 0.000 0.719 0.036 Minimum 1.531 2.297 1.531 2.352 1.949 Minimum 1.380 2.297 1.322 2.407 1.949 Maximum 1.851 2.447 1.623 2.447 2.000 Maximum 1.591 2.463 1.525 2.494 2.000 25% Quantile 1.652 2.318 1.580 2.376 1.962 25% Quantile 1.506 2.314 1.388 2.418 1.962 75% Quantile 1.819 2.363 1.612 2.423 1.987 75% Quantile 1.573 2.399 1.504 2.460 1.987 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Skew -7.6E-15 9.86E-15 -3.7E-14 7.5E-14 4.62E-15 Skew -7.4E-15 5.09E-15 -2.7E-14 6.64E-14 4.62E-15 IQR 0.005 0.335 0.000 0.423 0.025 IQR 0.007 0.443 0.000 0.508 0.025

Crenicichla-CAS Crenicichla-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.897 1.995 1.874 2.018 2.177 Mean 1.919 2.025 1.892 2.043 2.177 St Dev 0.310 0.352 0.300 0.359 0.222 St Dev 0.292 0.337 0.276 0.344 0.222 Minimum 1.322 1.432 1.322 1.447 1.602 Minimum 1.322 1.447 1.322 1.519 1.602 Maximum 2.462 2.494 2.449 2.494 2.494 Maximum 2.452 2.494 2.425 2.494 2.494 25% Quantile 1.591 1.714 1.588 1.738 2.055 25% Quantile 1.602 1.773 1.591 1.823 2.055 75% Quantile 2.178 2.290 2.162 2.302 2.354 75% Quantile 2.176 2.290 2.155 2.301 2.354 Kurtosis -1.517 -1.231 -1.567 -1.150 -0.531 Kurtosis -1.455 -0.985 -1.518 -0.738 -0.531 Skew -0.254 0.224 -0.310 0.324 -0.627 Skew -0.392 0.100 -0.554 0.186 -0.627 IQR 0.526 0.663 0.506 0.695 0.299 IQR 0.467 0.635 0.404 0.666 0.299

Dicrossus-GGD Dicrossus-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.888 2.234 1.815 2.285 1.711 Mean 1.695 2.236 1.622 2.344 1.711 St Dev 0.079 0.339 0.039 0.371 0.111 St Dev 0.119 0.461 0.066 0.516 0.111 Minimum 1.531 2.079 1.531 2.180 1.580 Minimum 1.322 2.029 1.322 2.193 1.580 Maximum 2.090 2.447 2.000 2.447 1.851 Maximum 1.991 2.481 1.763 2.494 1.851 25% Quantile 1.623 2.190 1.580 2.293 1.623 25% Quantile 1.518 2.204 1.431 2.331 1.623 75% Quantile 2.000 2.352 1.908 2.380 1.778 75% Quantile 1.724 2.422 1.574 2.477 1.778 Kurtosis -2.208 -1.089 -2.243 -0.982 -1.994 Kurtosis -2.228 -1.064 -2.249 -0.953 -1.994 Skew -0.921 0.720 -1.019 0.995 0.011 Skew -0.914 0.866 -1.046 1.013 0.011 IQR 0.036 0.514 0.000 0.630 0.155 IQR 0.049 0.792 0.011 0.879 0.155

Geophagus-GGD Geophagus-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.012 2.135 2.000 2.153 2.191 Mean 1.849 2.086 1.807 2.127 2.191 St Dev 0.158 0.258 0.140 0.273 0.135 St Dev 0.252 0.367 0.238 0.394 0.135 Minimum 1.531 1.778 1.531 1.934 1.881 Minimum 1.322 1.580 1.322 1.607 1.881

62

Maximum 2.324 2.447 2.297 2.447 2.380 Maximum 2.342 2.494 2.254 2.494 2.380 25% Quantile 1.866 2.032 1.815 2.079 2.097 25% Quantile 1.568 1.940 1.531 2.040 2.097 75% Quantile 2.155 2.293 2.120 2.304 2.297 75% Quantile 2.072 2.349 1.997 2.391 2.297 Kurtosis -1.099 1.015 -1.368 1.842 -0.793 Kurtosis -1.705 -0.562 -1.803 0.562 -0.793 Skew -1.059 -0.000 -1.217 0.394 -0.526 Skew -0.733 0.376 -1.099 0.566 -0.526 IQR 0.177 0.354 0.156 0.447 0.200 IQR 0.309 0.685 0.232 0.759 0.200

Guianacara-CAS Guianacara-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.701 2.176 1.612 2.262 1.951 Mean 1.739 2.194 1.621 2.270 1.951 St Dev 0.189 0.437 0.097 0.475 0.075 St Dev 0.162 0.422 0.098 0.454 0.075 Minimum 1.322 1.779 1.322 2.058 1.875 Minimum 1.322 1.881 1.322 2.072 1.875 Maximum 2.146 2.494 1.869 2.494 2.079 Maximum 2.137 2.494 1.878 2.494 2.079 25% Quantile 1.497 2.099 1.447 2.224 1.906 25% Quantile 1.519 2.122 1.456 2.221 1.906 75% Quantile 1.777 2.405 1.610 2.446 1.979 75% Quantile 1.854 2.396 1.644 2.441 1.979 Kurtosis -2.117 -0.098 -2.163 0.452 -1.369 Kurtosis -2.100 -0.253 -2.142 0.388 -1.369 Skew -0.916 0.925 -1.405 1.378 0.684 Skew -0.990 0.776 -1.296 1.403 0.684 IQR 0.149 0.735 0.082 0.865 0.074 IQR 0.133 0.703 0.080 0.839 0.074

Gymnogeophagus-GGD Gymnogeophagus-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.958 2.176 1.919 2.234 2.064 Mean 1.792 2.149 1.732 2.223 2.064 St Dev 0.123 0.284 0.100 0.306 0.089 St Dev 0.213 0.391 0.157 0.427 0.089 Minimum 1.531 1.950 1.531 2.013 1.934 Minimum 1.322 1.716 1.322 1.892 1.934 Maximum 2.243 2.447 2.122 2.447 2.190 Maximum 2.225 2.494 2.141 2.494 2.190 25% Quantile 1.799 2.085 1.693 2.164 1.996 25% Quantile 1.537 2.068 1.490 2.165 1.996 75% Quantile 2.079 2.321 2.010 2.338 2.122 75% Quantile 1.942 2.374 1.833 2.434 2.122 Kurtosis -1.682 0.931 -1.830 1.736 -1.507 Kurtosis -1.939 -0.095 -2.027 1.278 -1.507 Skew -1.220 0.493 -1.563 0.953 -0.008 Skew -0.951 0.648 -1.389 1.131 -0.008 IQR 0.111 0.413 0.084 0.514 0.126 IQR 0.196 0.708 0.131 0.783 0.126

Mazarunia-CAS Mazarunia-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.578 2.292 1.509 2.375 1.893 Mean 1.596 2.303 1.497 2.381 1.893 St Dev 0.060 0.529 0.019 0.604 0.028 St Dev 0.049 0.512 0.009 0.575 0.028 Minimum 1.380 2.182 1.322 2.268 1.869 Minimum 1.380 2.176 1.322 2.306 1.869 Maximum 1.690 2.481 1.591 2.494 1.924 Maximum 1.732 2.477 1.568 2.494 1.924 25% Quantile 1.483 2.236 1.414 2.326 1.878 25% Quantile 1.489 2.242 1.391 2.342 1.878 75% Quantile 1.623 2.406 1.555 2.454 1.905 75% Quantile 1.671 2.406 1.525 2.460 1.905 Kurtosis -2.333 -2.333 -2.333 -2.333 -2.333 Kurtosis -2.333 -2.333 -2.333 -2.333 -2.333 Skew -0.384 0.383 -0.385 0.385 0.229 Skew -0.384 0.384 -0.385 0.385 0.229 IQR 0.056 0.500 0.018 0.570 0.028 IQR 0.046 0.474 0.009 0.538 0.028

Mikrogeophagus-GGD Mikrogeophagus-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed

63

Mean 1.763 2.324 1.634 2.383 1.640 Mean 1.553 2.347477 1.451001 2.438443 1.639834 St Dev 0.008 0.512 0 0.583 0.153 St Dev 0.010 0.627 0.000 0.719 0.153 Minimum 1.531 2.297 1.531 2.338 1.531 Minimum 1.380 2.297 1.322 2.407 1.531 Maximum 1.880 2.447 1.736 2.447 1.748 Maximum 1.591 2.463 1.525 2.494 1.748 25% Quantile 1.662 2.304 1.583 2.357 1.586 25% Quantile 1.506 2.314 1.388 2.418 1.586 75% Quantile 1.833 2.355 1.685 2.415 1.694 75% Quantile 1.573 2.399 1.504 2.460 1.694 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Skew -6.4E-15 7.86E-15 -3.7E-14 3.16E-14 0.000 Skew -7.4E-15 5.09E-15 -2.7E-14 6.64E-14 0.000 IQR 0.006 0.362 0.000 0.412 0.108 IQR 0.007 0.443 0.000 0.508 0.108

Satanoperca-CAS Satanoperca-Geophagini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.727 2.163 1.623 2.259 2.253 Mean 1.765 2.176 1.687 2.261 2.253 St Dev 0.186 0.426 0.135 0.451 0.077 St Dev 0.180 0.412 0.111 0.440 0.077 Minimum 1.322 1.756 1.322 1.996 2.149 Minimum 1.362 1.852 1.322 2.002 2.149 Maximum 2.137 2.494 1.929 2.494 2.407 Maximum 2.164 2.494 2.059 2.494 2.407 25% Quantile 1.501 2.086 1.433 2.169 2.212 25% Quantile 1.529 2.103 1.474 2.186 2.212 75% Quantile 1.804 2.404 1.674 2.440 2.285 75% Quantile 1.876 2.383 1.739 2.430 2.285 Kurtosis -2.057 -0.267 -2.155 0.619 -0.635 Kurtosis -2.052 -0.229 -2.152 0.512 -0.635 Skew -0.8704 0.912 -1.329 1.342 0.645 Skew -0.948 0.747 -1.331 1.175 0.645 IQR 0.171 0.756 0.110 0.844 0.073 IQR 0.172 0.717 0.117 0.812 0.073

Acaronia-Cichlasomatini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.753 2.172 1.650 2.209 2.137 St Dev 0.005 0.341 0.000 0.416 0.072 Minimum 1.653 2.107 1.556 2.176 2.086 Maximum 1.839 2.301 1.695 2.301 2.188 25% Quantile 1.691 2.142 1.603 2.192 2.112 75% Quantile 1.804 2.205 1.689 2.253 2.162 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Skew -1.3E-14 8.64E-15 -6.4E-14 5.86E-14 4.63E-15 IQR 0.003 0.241 0.000 0.294 0.051

Aequidens-Cichlasomatini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.924 2.031 1.904 2.056 2.073 St Dev 0.113 0.188 0.101 0.196 0.123 Minimum 1.556 1.792 1.556 1.851 1.748 Maximum 2.127 2.301 2.090 2.301 2.301 25% Quantile 1.801 1.983 1.774 2.006 2.009 75% Quantile 2.025 2.122 2.002 2.149 2.155 Kurtosis -1.300 1.125 -1.470 2.204 0.417 Skew -1.065 0.214 -1.361 0.501 -0.477 IQR 0.111 0.266 0.088 0.314 0.147

64

Andinoacara-Cichlasomatini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.753 2.172 1.650 2.209 2.027 St Dev 0.005 0.341 0.000 0.416 0.038 Minimum 1.653 2.107 1.556 2.176 2.000 Maximum 1.839 2.301 1.695 2.301 2.053 25% Quantile 1.691 2.142 1.603 2.192 2.013 75% Quantile 1.804 2.205 1.689 2.253 2.040 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Skew -1.3E-14 8.64E-15 -6.4E-14 5.86E-14 -8.91E-15 IQR 0.003 0.241 0.000 0.294 0.027

Burjurquina-Cichlasomatini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.911 2.037 1.890 2.066 1.938 St Dev 0.105 0.195 0.092 0.207 0.108 Minimum 1.556 1.826 1.556 1.860 1.740 Maximum 2.107 2.301 2.083 2.301 2.176 25% Quantile 1.792 2.000 1.748 2.013 1.869 75% Quantile 2.013 2.134 2.000 2.169 2.013 Kurtosis -1.409 1.205 -1.584 2.596 -0.448 Skew -1.186 0.353 -1.643 0.647 0.191 IQR 0.093 0.279 0.067 0.329 0.144

Cichlasoma-Cichlasomatini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.897 2.047 1.868 2.071 2.007 St Dev 0.095 0.208 0.072 0.218 0.095 Minimum 1.556 1.869 1.556 1.916 1.820 Maximum 2.090 2.301 2.053 2.301 2.134 2.00021 25% Quantile 1.748188 5 1.69897 2.039 1.964 75% Quantile 2.000 2.155 1.916 2.169 2.068 Kurtosis -1.626 1.110 -1.778 2.242 -0.975 Skew -1.215 0.442 -1.659 0.965 -0.577 IQR 0.078 0.315 0.054 0.366 0.104

Krobia-Cichlasomatini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.797 2.132 1.709 2.184 2.104 St Dev 0.027 0.286 0.009 0.346 0.006 Minimum 1.556 2.064 1.556 2.107 2.097 Maximum 1.913 2.301 1.826 2.301 2.107 25% Quantile 1.705 2.096 1.623 2.153 2.102

65

75% Quantile 1.871 2.196 1.764 2.252 2.107 Kurtosis -2.333 -2.333 -2.333 -2.333 -2.333 Skew -0.384 0.382 -0.3849 0.385 -0.385 IQR 0.027 0.275 0.009 0.323 0.005

Laetacara-Cichlasomatini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.854 2.080 1.815 2.120 1.770 St Dev 0.063 0.236 0.038 0.276 0.136 Minimum 1.556 1.964 1.556 2.025 1.556 Maximum 2.029 2.301 1.993 2.301 1.914 25% Quantile 1.730 2.044 1.674 2.072 1.692 75% Quantile 1.949 2.175 1.883 2.202 1.862 Kurtosis -2.088 -0.545 -2.214 -0.284 -1.656 Skew -1.048 0.702 -1.241 0.957 -0.442 IQR 0.044 0.335 0.023 0.404 0.170

Nannacara-Cichlasomatini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.854 2.080 1.815 2.120 1.723 St Dev 0.063 0.236 0.0375 0.276 0.067 Minimum 1.556 1.964 1.556 2.025 1.653 Maximum 2.029 2.301 1.993 2.301 1.826 25% Quantile 1.730 2.044 1.674 2.072 1.690 75% Quantile 1.949 2.175 1.883 2.202 1.748 Kurtosis -2.088 -0.545 -2.214 -0.284 -1.608 Skew -1.048 0.702 -1.241 0.957 0.475 IQR 0.044 0.335 0.023 0.404 0.058

Tahuantinsuyoa-Cichlasomatini

2.50% 97.50% 0.25% 99.75% Observed Mean 1.753 2.172 1.650 2.208 1.996 St Dev 0.005 0.341 0.000 0.416 0.117 Minimum 1.653 2.107 1.556 2.176 1.914 Maximum 1.839 2.301 1.695 2.301 2.079 25% Quantile 1.691 2.142 1.603 2.192 1.955 75% Quantile 1.804 2.205 1.689 2.253 2.038 Kurtosis -2.750 -2.750 -2.750 0.000 -2.750 Skew -1.3E-14 8.64E-15 -6.4E-14 5.86E-14 0.000 IQR 0.003 0.241 0.000 0.294 0.083

Amatitlania-SAClade Amatitilania-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.985 2.391 1.927 2.460 1.944 Mean 2.000117 2.374588 1.94 2.439884 1.944 St Dev 0.065 0.345 0.0373 0.394 0.045 St Dev 0.056 0.359 0.033 0.431 0.045

66

Minimum 1.756 2.230 1.724 2.317 1.898 Minimum 1.724 2.225 1.699 2.366 1.898 Maximum 2.133 2.699 2.041 2.699 2.000 Maximum 2.134 2.699 2.041 2.699 2.000 25% Quantile 1.863 2.332 1.815 2.382 1.914 25% Quantile 1.863 2.327 1.829 2.391 1.914 75% Quantile 2.037 2.508 1.963 2.565 1.969 75% Quantile 2.047 2.486 1.970 2.583 1.969 Kurtosis -2.404 -1.699 -2.435 -1.688 -2.108 Kurtosis -2.406 -1.706 -2.432 -1.689 -2.108 Skew -0.717 0.684 -0.748 0.742 0.176 Skew -0.703 0.686 -0.748 0.745 0.176 IQR 0.054 0.459 0.030 0.608 0.056 IQR 0.056 0.476 0.022 0.628 0.056

Amphilophus-CAClade Amphilophus-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.111 2.273 2.087 2.304 2.221 Mean 2.092 2.261 2.063 2.287 2.221 St Dev 0.155 0.264 0.125 0.282 0.096 St Dev 0.152 0.261 0.134 0.278 0.096 Minimum 1.724 1.954 1.724 2.041 2.104 Minimum 1.699 1.949 1.699 2.000 2.104 Maximum 2.398 2.699 2.380 2.699 2.398 Maximum 2.398 2.699 2.347 2.699 2.398 25% Quantile 1.915 2.169 1.891 2.208 2.170 25% Quantile 1.910 2.158 1.892 2.189 2.170 75% Quantile 2.221 2.417 2.174 2.466 2.244 75% Quantile 2.205 2.410 2.153 2.466 2.244 Kurtosis -1.533 0.257 -1.652 1.000 -0.833 Kurtosis -1.536 0.122 -1.685 0.921 -0.833 Skew -0.727 0.690 -1.049 0.984 0.801 Skew -0.686 0.668 -0.926 1.042 0.801 IQR 0.167 0.431 0.116 0.482 0.074 IQR 0.157 0.433 0.132 0.488 0.074

Archocentrus-CAClade Archocentrus-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.906 2.462 1.826 2.605 2.041 Mean 1.899 2.438 1.793 2.548 2.041 0.000 0.000 St Dev 0.000 0.430 0.000 0.595 St Dev 0.005 0.476 0.000 0.599 Minimum 1.785 2.398 1.724 2.511 2.041 Minimum 1.724 2.398 1.699 2.415 2.041 Maximum 1.954 2.677 1.875 2.699 2.041 Maximum 1.959 2.677 1.854 2.699 2.041 25% Quantile 1.863 2.418 1.784 2.558 NA 25% Quantile 1.841 2.418 1.748 2.473 NA 75% Quantile 1.923 2.558 1.857 2.652 NA 75% Quantile 1.919 2.517 1.833 2.624 NA Kurtosis -2.750 0.000 -2.750 0.000 NA Kurtosis -2.750 -2.750 -2.750 0.000 NA Skew -8.8E-15 8.75E-15 -3.6E-14 5.05E-14 NA Skew -8.7E-15 8.64E-15 -9.5E-14 4.5E-14 NA IQR 0.000 0.304 0.000 0.421 0 IQR 0.003 0.337 0.000 0.423 0

Astatheros-CAClade Astatheros-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.053 2.326 1.996 2.371 2.179 Mean 2.041 2.319 1.988 2.378 2.179 St Dev 0.113 0.302 0.089 0.352 0.196 St Dev 0.108 0.306 0.078 0.346 0.196 Minimum 1.724 2.114 1.724 2.204 1.845 Minimum 1.699 2.121 1.699 2.204 1.845 Maximum 2.279 2.699 2.190 2.699 2.398 Maximum 2.270 2.699 2.197 2.699 2.398 25% Quantile 1.878 2.253 1.848 2.295 2.093 25% Quantile 1.878 2.253 1.835 2.341 2.093 75% Quantile 2.146 2.477 2.087 2.553 2.309 75% Quantile 2.116 2.443 2.050 2.501 2.309 Kurtosis -1.974 -0.198 -2.141 0.373 -1.414 Kurtosis -1.978 0.020 -2.077 0.949 -1.414 Skew -0.861 0.844 -1.290 1.193 -0.369 Skew -0.879 1.037 -1.515 1.405 -0.369 IQR 0.092 0.484 0.059 0.558 0.215 IQR 0.095 0.457 0.056 0.567 0.215

Australoheros-CAClade Australoheros-Heroini

67

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.121 2.262 2.094 2.284 1.969 Mean 2.107 2.243 2.061 2.281 1.969 St Dev 0.167 0.250 0.153 0.264 0.138 St Dev 0.162 0.253 0.143 0.270 0.138 Minimum 1.724 1.903 1.724 1.957 1.724 Minimum 1.699 1.903 1.699 1.982 1.724 Maximum 2.415 2.699 2.398 2.699 2.286 Maximum 2.398 2.699 2.398 2.699 2.286 25% Quantile 1.939 2.147 1.902 2.197 1.884 25% Quantile 1.916 2.132 1.903 2.174 1.884 75% Quantile 2.243 2.406 2.197 2.438 2.083 75% Quantile 2.227 2.398 2.176 2.450 2.083 Kurtosis -1.457 0.124 -1.667 0.534 -0.567 Kurtosis -1.427 0.104 -1.563 0.741 -0.567 Skew -0.585 0.566 -0.955 0.778 0.390 Skew -0.576 0.554 -0.754 0.771 0.390 IQR 0.194 0.427 0.173 0.466 0.199 IQR 0.188 0.418 0.147 0.453 0.199

Caquetaia-CAClade Caquetaia-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.985 2.391 1.927 2.460 2.397 Mean 1.984 2.382 1.909 2.452 2.397 St Dev 0.065 0.345 0.037 0.394 0.204 St Dev 0.059 0.355 0.026 0.428 0.204 Minimum 1.756 2.230 1.724 2.317 2.218 Minimum 1.699 2.225 1.699 2.391 2.217 Maximum 2.133 2.699 2.041 2.699 2.677 Maximum 2.137 2.699 2.041 2.699 2.677 25% Quantile 1.863 2.332 1.815 2.382 2.263 25% Quantile 1.862 2.319 1.811 2.398 2.263 75% Quantile 2.037 2.508 1.963 2.565 2.480 75% Quantile 2.036 2.498 1.944 2.621 2.480 Kurtosis -2.404 -1.699 -2.435 -1.688 -1.937 Kurtosis -2.403 -1.701 -2.432 -1.688 -1.937 Skew -0.717 0.684 -0.748 0.742 0.426 Skew -0.702 0.678 -0.749 0.741 0.426 IQR 0.054 0.459 0.030 0.608 0.217 IQR 0.048 0.470 0.024 0.589 0.217

Cryptoheros-CAClade Cryptoheros-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.061 2.329 2.028 2.389 1.977 Mean 2.047 2.311 1.992 2.359 1.977 St Dev 0.122 0.295 0.097 0.328 0.078 St Dev 0.120 0.294 0.092 0.333 0.078 Minimum 1.724 2.100 1.724 2.164 1.845 Minimum 1.699 2.086 1.699 2.146 1.845 Maximum 2.301 2.699 2.244 2.699 2.114 Maximum 2.301 2.699 2.217 2.699 2.114 25% Quantile 1.892 2.246 1.845 2.337 1.954 25% Quantile 1.881 2.243 1.845 2.335 1.954 75% Quantile 2.146 2.477 2.086 2.562 2.000 75% Quantile 2.134 2.477 2.027 2.477 2.000 Kurtosis -1.932 0.008 -2.046 0.950 -0.854 Kurtosis -1.905 0.046 -1.999 1.005 -0.854 Skew -0.922 0.862 -1.304 1.243 0.054 Skew -0.850 0.880 -1.436 1.272 0.054 IQR 0.084 0.495 0.051 0.548 0.046 IQR 0.084 0.490 0.041 0.552 0.046

Herichthys-CAClade Herichthys-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.061 2.329 2.028 2.389 2.295 Mean 2.047 2.311 1.992 2.359 2.295 St Dev 0.122 0.295 0.097 0.328 0.170 St Dev 0.120 0.294 0.091 0.333 0.170 Minimum 1.724 2.100 1.724 2.164 2.079 Minimum 1.699 2.087 1.699 2.146 2.079 Maximum 2.301 2.699 2.244 2.699 2.602 Maximum 2.301 2.699 2.217 2.699 2.602 25% Quantile 1.892 2.246 1.845 2.337 2.2430 25% Quantile 1.881 2.243 1.845 2.335 2.243 75% Quantile 2.146 2.477 2.086 2.562 2.398 75% Quantile 2.134 2.477 2.027 2.477 2.398 Kurtosis -1.932 0.008 -2.046 0.950 -1.169 Kurtosis -1.905 0.046 -1.999 1.005 -1.169 Skew -0.922 0.862 -1.304 1.243 0.413 Skew -0.850 0.880 -1.436 1.272 0.413

68

IQR 0.084 0.495 0.051 0.548 0.155 IQR 0.084 0.490 0.041 0.552 0.155

Heros-SAClade Heros-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.926 2.243 1.864 2.299 2.176 Mean 1.984 2.382 1.909 2.452 2.176 St Dev 0.068 0.300 0.040 0.332 0.093 St Dev 0.059 0.355 0.026 0.428 0.093 Minimum 1.699 2.137 1.699 2.204 2.079 Minimum 1.699 2.225 1.699 2.391 2.079 Maximum 2.090 2.398 1.973 2.398 2.301 Maximum 2.137 2.699 2.041 2.699 2.301 25% Quantile 1.813 2.176 1.813 2.265 2.129 25% Quantile 1.862 2.319 1.811 2.398 2.129 75% Quantile 1.990 2.325 1.929 2.378 2.207 75% Quantile 2.036 2.498 1.944 2.621 2.207 Kurtosis -2.396 -1.721 -2.432 -1.702 -1.877 Kurtosis -2.403 -1.701 -2.432 -1.688 -1.877 Skew -0.683 0.654 -0.727 0.715 0.328 Skew -0.702 0.678 -0.749 0.741 0.328 IQR 0.058 0.371 0.031 0.473 0.078 IQR 0.048 0.470 0.024 0.589 0.078

Mesonauta-SAClade Mesonauta-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.971 2.207 1.933 2.248 1.940 Mean 2.010 2.332 1.934 2.381 1.940 St Dev 0.090 0.258 0.069 0.278 0.057 St Dev 0.094 0.324 0.054 0.366 0.0568 Minimum 1.699 2.000 1.699 2.121 1.851 Minimum 1.699 2.146 1.699 2.230 1.851 Maximum 2.146 2.398 2.100 2.398 2.000 Maximum 2.204 2.699 2.104 2.699 2.000 25% Quantile 1.867 2.1534 1.857 2.190 1.914 25% Quantile 1.878 2.272 1.816 2.385 1.914 75% Quantile 2.056 2.348 1.978 2.354 1.983 75% Quantile 2.081 2.457 1.951 2.581 1.983 Kurtosis -2.099 -0.759 -2.220 -0.337 -1.693 Kurtosis -2.112 -0.608 -2.257 -0.247 -1.693 Skew -0.840 0.680 -1.198 0.966 -0.337 Skew -0.891 0.890 -1.136 1.172 -0.337 IQR 0.069 0.368 0.044 0.421 0.070 IQR 0.063 0.443 0.028 0.545 0.070

Parachromis-CA Clade Parachromis-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.016 2.365 1.947 2.425 2.447 Mean 2.008 2.361 1.946 2.453 2.447 St Dev 0.079 0.321 0.044 0.373 0.164 St Dev 0.083 0.336 0.046 0.408 0.164 Minimum 1.724 2.176 1.724 2.279 2.267 Minimum 1.699 2.190 1.699 2.293 2.267 Maximum 2.199 2.699 2.097 2.699 2.699 Maximum 2.176 2.699 2.098 2.699 2.699 25% Quantile 1.869 2.332 1.785 2.398 2.342 25% Quantile 1.851 2.301 1.740 2.398 2.342 75% Quantile 2.079 2.477 1.978 2.602 2.477 75% Quantile 2.041 2.477 1.973 2.602 2.477 Kurtosis -2.211 -1.076 -2.244 -0.942 -1.561 Kurtosis -2.220 -1.060 -2.242 -0.974 -1.561 Skew -0.894 0.819 -1.056 0.991 0.417 Skew -0.876 0.865 -1.024 1.009 0.417 IQR 0.037 0.483 0.007 0.585 0.135 IQR 0.033 0.487 0.013 0.621 0.135

Paraneetroplus-CA Clade Paraneetroplus-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.091 2.298 2.056 2.346 2.370 Mean 2.070 2.279 2.016 2.317 2.370 St Dev 0.143 0.274 0.113 0.296 0.110 St Dev 0.138 0.268 0.112 0.299 0.110 Minimum 1.724 2.000 1.724 2.119 2.117 Minimum 1.699 1.982 1.699 2.102 2.117 Maximum 2.398 2.699 2.301 2.699 2.544 Maximum 2.371 2.699 2.263 2.699 2.544 25% Quantile 1.907 2.203 1.876 2.254 2.316 25% Quantile 1.903 2.183 1.872 2.225 2.316

69

75% Quantile 2.196 2.457 2.124 2.477 2.425 75% Quantile 2.168 2.423 2.112 2.461 2.425 Kurtosis -1.707 0.229 -1.846 1.093 -0.223 Kurtosis -1.711 0.385 -1.837 0.996 -0.223 Skew -0.809 0.758 -1.068 1.010 -0.638 Skew -0.752 0.750 -1.029 1.147 -0.638 IQR 0.142 0.451 0.087 0.517 0.109 IQR 0.134 0.444 0.107 0.508 0.109

Pterophyllum-SA Clade Pterophyllum-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.893 2.277 1.819 2.321 1.943 Mean 1.940 2.405 1.870 2.492 1.943 St Dev 0.038 0.337 0.022 0.369 0.284 St Dev 0.036 0.405 0.004 0.490 0.284 Minimum 1.699 2.176 1.699 2.255 1.699 Minimum 1.724 2.301 1.699 2.398 1.699 Maximum 1.987 2.398 1.894 2.398 2.255 Maximum 2.049 2.699 1.916 2.699 2.255 25% Quantile 1.787 2.227 1.775 2.283 1.787 25% Quantile 1.850 2.353 1.795 2.433 1.787 75% Quantile 1.943 2.349 1.879 2.385 2.065 75% Quantile 2.007 2.520 1.897 2.588 2.065 Kurtosis -2.333 -2.333 -2.333 -2.333 -2.333 Kurtosis -2.333 -2.333 -2.333 -2.333 -2.333 Skew -0.381 0.383 -0.385 0.385 0.226 Skew -0.385 0.385 -0.385 0.385 0.226 IQR 0.036 0.302 0.020 0.349 0.278 IQR 0.033 0.390 0.004 0.446 0.278

Symphysodon-SA Clade Symphysodon-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.893 2.277 1.819 2.321 2.116 Mean 1.940 2.405 1.870 2.492 2.116 St Dev 0.038 0.337 0.022 0.369 0.024 St Dev 0.036 0.405 0.004 0.490 0.024 Minimum 1.699 2.176 1.699 2.255 2.090 Minimum 1.724 2.301 1.699 2.398 2.090 Maximum 1.987 2.398 1.894 2.398 2.137 Maximum 2.049 2.699 1.916 2.699 2.137 25% Quantile 1.787 2.227 1.775 2.283 2.105 25% Quantile 1.850 2.353 1.795 2.433 2.105 75% Quantile 1.943 2.349 1.879 2.385 2.129 75% Quantile 2.007 2.520 1.897 2.588 2.129 Kurtosis -2.333 -2.333 -2.333 -2.333 -2.333 Kurtosis -2.333 -2.333 -2.333 -2.333 -2.333 Skew -0.381 0.383 -0.385 0.385 -0.195 Skew -0.385 0.385 -0.385 0.385 -0.195 IQR 0.036 0.302 0.020 0.349 0.023 IQR 0.033 0.390 0.004 0.446 0.023

Theraps-CA Clade Theraps-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.031 2.346 1.987 2.388 2.330 Mean 2.028 2.340 1.984 2.394 2.330 St Dev 0.102 0.308 0.069 0.333 0.129 St Dev 0.100 0.303 0.078 0.341 0.129 Minimum 1.724 2.146 1.724 2.237 2.079 Minimum 1.699 2.137 1.699 2.224 2.079 Maximum 2.267 2.699 2.146 2.699 2.477 Maximum 2.245 2.699 2.161 2.699 2.477 25% Quantile 1.884 2.261 1.832 2.364 2.290 25% Quantile 1.883 2.262 1.824 2.347 2.290 75% Quantile 2.125 2.477 2.027 2.551 2.398 75% Quantile 2.113 2.477 2.047 2.557 2.398 Kurtosis -2.064 -0.309 -2.151 0.416 -0.676 Kurtosis -2.054 -0.267 -2.134 0.199 -0.676 Skew -0.910 0.890 -1.368 1.162 -0.806 Skew -0.954 0.912 -1.296 1.223 -0.806 IQR 0.081 0.458 0.047 0.538 0.108 IQR 0.088 0.449 0.051 0.543 0.108

Thorichthys-CA Clade Thorichthys-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 2.053 2.326 1.996 2.371 2.177 Mean 2.041 2.319 1.988 2.378 2.177 St Dev 0.113 0.302 0.089 0.352 0.035 St Dev 0.108 0.306 0.078 0.346 0.035

70

Minimum 1.724 2.114 1.724 2.204 2.146 Minimum 1.699 2.121 1.699 2.204 2.146 Maximum 2.279 2.699 2.190 2.699 2.230 Maximum 2.270 2.699 2.197 2.699 2.230 25% Quantile 1.878 2.253 1.848 2.295 2.146 25% Quantile 1.878 2.253 1.835 2.341 2.146 75% Quantile 2.146 2.477 2.087 2.553 2.190 75% Quantile 2.116 2.443 2.050 2.501 2.190 Kurtosis -1.974 -0.198 -2.141 0.373 -1.467 Kurtosis -1.978 0.020 -2.077 0.949 -1.467 Skew -0.861 0.844 -1.290 1.193 0.626 Skew -0.879 1.037 -1.515 1.405 0.626 IQR 0.092 0.484 0.059 0.558 0.044 IQR 0.095 0.457 0.056 0.567 0.044

Uaru -SA Clade Uaru-Heroini

2.50% 97.50% 0.25% 99.75% Observed 2.50% 97.50% 0.25% 99.75% Observed Mean 1.836 2.327 1.775 2.384504 2.338347 Mean 1.899 2.438 1.793 2.548 2.338 St Dev 0.008 0.387 0.000 0.475245 0.084278 St Dev 0.005 0.476 0.000 0.599 0.084 Minimum 1.699 2.279 1.699 2.371068 2.278754 Minimum 1.724 2.398 1.699 2.415 2.279 Maximum 1.914 2.398 1.851 2.39794 2.39794 Maximum 1.960 2.677 1.854 2.699 2.398 25% Quantile 1.768 2.302 1.737 2.377786 2.30855 25% Quantile 1.841 2.418 1.748 2.473 2.309 75% Quantile 1.869 2.354 1.813 2.391222 2.368143 75% Quantile 1.919 2.517 1.833 2.624 2.368 Kurtosis -2.750 -2.750 -2.750 0.000 -2.75 Kurtosis -2.750 -2.750 -2.750 0.000 -2.75 Skew -8.5E-15 1.02E-14 -2.1E-14 1.75E-14 0.000 Skew -8.7E-15 8.64E-15 -9.5E-14 4.5E-14 0.000 IQR 0.005 0.273 0.000 0.336 0.060 IQR 0.003 0.337 0.000 0.423 0.060

71 Chapter 2 Body Size Diversity Across the Fishes and the Impact of Environment on Absolute Body Size: Does Water and Macrohabitat Type Shape the Distribution of Size?

Sarah Elizabeth Steele1, Matthew Kolmann2, and Hernán López-Fernández1,3

1Department of Ecology and Evolutionary Biology, University of Toronto, 25 Wilcocks St.,

Toronto, Ontario M5S 3B2, Canada

2Friday Harbor Laboratories, University of Washington, 620 University Dr., Friday Harbor,

Washington 98250, USA

3Department of Natural History, Royal Ontario Museum, 100 Queen’s Park, Toronto, Ontario

M5S 2C6, Canada

72 Chapter 2

2.1 Abstract

Body size, as a correlate of most ecological and life history traits, should be a fundamental component of understanding global patterns of biodiversity. While many broad studies of body size distributions exist across vertebrates, including fishes, few studies have addressed broad scale questions of body size across fishes. While species diversity is known to be similar across freshwater and marine fishes, it is not known how occupation of a certain environment affects species diversity among taxonomic scales. Likewise, whether the environment imposes particular selective pressures on evolution resulting in divergence of body size between marine and freshwater is also unknown. Previously, it has been suggested that body size diversity and reductions in body size of fishes should be most prominent in tropical riverine systems, or complex coral reef systems of the Indo-Pacific. Using taxonomic, body size, and habitat data from 32,000 species of fishes, I tested whether there was significant body size distribution divergence of fishes among these environments, and if macrohabitats further imposed directional selection on body size evolution across the major lineages of fishes. I also tested whether trends in body size distributional changes were attributable to clade species richness and age, rather than ecological factors affecting body size. I found significant reduction in mean body size in freshwater that was consistent across all levels of data partitioning, as well as increases in body size diversity and an overall right-skewed bias in freshwater clades. These trends were not associated with clade age, but were somewhat predicted by clade species richness. This suggests that the ecological factors impose strong selective pressures on body

73 size, resulting in repeated directional evolution across multiple secondary freshwater lineages towards small body size.

2.2 Introduction

Body size exhibits a strong correlation with most ecological and life history traits such as fecundity, trophic position (Romanuk et al., 2011), range size (Lindstedt et al., 1986), longevity

(Blackburn & Gaston, 1994; Brown et al., 1993), and even extinction risk (Cardillo et al., 2005;

Clauset & Erwin, 2008) and therefore is fundamental in understanding patterns of biodiversity.

Investigations of body size have been conducted at various spatial and temporal scales, using a variety of methods. Studies of broad scale patterns of body size rely heavily on description and comparison of body size frequency distributions within a focal group of organisms. Histograms of species richness among body size bins (from here on referred to as “Body Size Frequency

Distributions” or BSFDs, Fig. 2.1) are used to calculate various metrics that describe the properties of the group’s body size distribution. A powerful property of this approach is its flexibility in terms of how groups are defined. For example, macroevolutionary analyses of body size can focus on groups defined as clades, and macroecological approaches analyze groups defined as ecological assemblies. Importantly, questions linking macroevolution, macroecological or geographical correlates of body size variation can be addressed by comparing ecologically, evolutionarily, or geographically defined partitions within a group of interest.

Broad evolutionary patterns are often difficult to address due in part to the lack of species-level phylogenies of large taxonomic groups (e.g. classes, orders, and families) required to accurately compare traits that are highly variable at the species level. Despite this, comparing broad patterns at higher taxonomic levels while still including species level information using

74

BSFD analyses provides considerable information about general trends in trait distribution, while also suggesting potential mechanisms operating on large spatial and long evolutionary time scales. The use of BSFDs to assess trends in body size among groups or lineages is more informative than descriptive statistics such as mean, minimum, and maximum body sizes alone, as comparisons of BSFDs retain information about trait distribution at the species level. In addition, comparisons of BSFDs to well informed null distributions accounting for taxonomic and diversity biases, as well as to statistical nulls such as log-normal distribution, can produce

(next page) Figure 2.1: Theoretical effects of skew and kurtosis on mean body size and diversity, and the importance of examining shape. Black markers indicate normal distribution parameters while red markers indicate changes to mean and diversity with changes in skew and kurtosis, while holding total observations constant. a) Body size distribution parameters lacking shape and frequency information; b) Normal distribution of body size; c) Right-skewed distribution showing overrepresentation of small-bodied species, under-representation of large- bodied species, and a smaller mean body size; d) Left-skewed distribution showing overrepresentation of large-bodied species, underrepresentation of small-bodied species, and larger mean body size; e) Leptokurtic distribution showing high ‘peakedness’ and conserved range of body size; f) Platykurtic distribution showing flattened distribution and wide range of body size; g) clade replication, with blue, green, and orange clades exhibiting same distribution parameters as each other as well as their containing clade (black); h) clade “stacking”, with blue, green, and orange clades exhibiting divergent distribution parameters from each other as well as their containing clade (black), however, their cumulative effects resulting in the containing clade distributions matching in panels g and h.

75

76 highly informative results highlighting the relative influence of different factors on BSFDs, allowing inference of potential mechanisms of body size divergence.

Throughout this chapter, I used BSFDs to examine the shape of body size distributions of fishes, identify over- or under-representation of certain body sizes in given clades or assemblages and propose potential underlying reasons for such biases. I compared BSFDs of various evolutionary and ecological groupings of fishes, focusing on three distribution parameters – occupation, diversity, and skew. I expected that divergence of parameters among groupings are the result of some evolutionary or ecological selective pressure leading to body size divergence among the fishes. Occupation refers to mean, minimum, and maximum body sizes that indicate the position of a species or a subgroup along the body size spectrum of its containing group (e.g. coral reef fishes among all marine fishes, or damselfishes among all coral reef fishes). Diversity refers to the standard deviation and kurtosis (“peakedness”, Fig. 2.1 e-f) that suggests the degree of constraint around the mean body size. Finally, skew summarizes the symmetry of the distribution, and therefore the proportion of small and large sizes within a

BSFD (Fig. 2.1 c-d). These parameters need not be correlated to each other (Fig, 2.1), and the shapes of distributions are largely independent from parameter values. Nevertheless, correlations between parameters may indicate biological trends influencing body size. The shapes of BSFDs may also be a result of various underlying patterns of their subclades, and while broader groupings may appear to be similar in body size distribution, subclades may be highly divergent (Fig. 2.1 g-h). Importantly, by assessing similarity of BSFD parameters within and among taxonomic, ecological, or geographical scales, one can evaluate whether body size distributions are replicated across these scales due to random trait distribution (Fig. 2.1 g) or scale dependent due to mechanistic drivers of trait divergence (Fig. 2.1 h).

77

The range of body sizes that exist in fishes is staggering, from 12-meter whale sharks

(Rhincodon typus) to eight-millimetre long Paedocypris progenetica, encompassing a large portion of body size diversity across vertebrates. This incredible diversity in size has led biogeographers and macroecologists to try to understand broad patterns of BSFDs, such as the prevalence of right-skew in the distributions of most animals (Allen et al., 2006; Bakker & Kelt,

2000; Blackburn & Gaston, 1994; Hutchinson & MacArthur, 1959), increases in body size over latitude (i.e., Bergmann’s Rule, Bergmann, 1948) and evolutionary time (i.e., Cope’s Rule,

Rensch, 1948), and the variation of BSFDs among different spatial and taxonomic scales (e.g.

Coetzee et al., 2013; Cox et al., 2011; Smith et al., 2004). Explanations used to understand the origin of body size variation and associated distributional parameters can be roughly placed into evolutionary, ecological, physiological, and geographical categories.

It is well known that evolutionary processes and shared ancestry can impact the evolution of traits and their distribution among extant organisms (Clauset & Erwin, 2008).

Directional evolution towards larger body size over evolutionary time (i.e., Cope’s Rule) has been frequently invoked but remains controversial. If supported in fishes, I expected correlation between body size and clade age, with older clades on average larger than younger clades.

However, it is far from clear whether Cope’s Rule applies broadly across taxa, or exactly what evolutionary mechanisms would drive such a pattern. It has also been proposed that differences in BSFD shape among lineages is due to differences in net diversification rate among different- sized organisms (Clauset & Erwin, 2008; Kozłowski & Gawelczyk, 2002). At small body size, speciation rates are expected to be high due to reduced energy requirements, faster generation times, and lower dispersal capabilities leading to the isolation of populations (Kozłowski &

Gawelczyk, 2002). Extinction risk at small body size is somewhat more controversial, as faster generation times and larger population size would be expected to buffer against extinction,

78 however high rates of isolation and restricted range sizes can lead to rapid local extinction

(Kozłowski & Gawelczyk, 2002). At large body size, speciation rates is expected to be relatively low due to significant increases in space and energy required for populations, high dispersal capabilities and slower generation times, leading to higher extinction risk (Cardillo et al., 2005). As a result, diversification rates are higher at small body size than at large body size

(Kozłowski & Gawelczyk, 2002). Differences in species richness may therefore also be correlated with body size parameters, with increases in species leading to higher diversity and bias towards small body size.

Many ecological attributes are correlated with body size at different scales of both ecological and phylogenetic divergence. Occupation of small or large body sizes should be associated with outcomes of biotic interactions among co-occurring species with regards to competition, resource partitioning or predator-prey interactions (Brown & Wilson, 1956;

Wilson, 1975), as well as abiotic interactions with the environment such as habitat use (e.g.

Collar et al., 2011), migration (e.g. Alerstam et al., 2003), and range size (e.g. Lindstedt et al.,

1986). I asked whether two major axes of ecological diversification in fishes, marine versus freshwater environments and habitat complexity may be associated with body size variation in fishes. Body size distributions are expected to be divergent in fishes inhabiting aquatic environments that differ in their chemical, structural, and geographic characteristics. Weitzman

& Vari (1988) proposed that larger fishes should typically be found in open waters while smaller fishes would be expected in smaller, more complex systems, regardless of environment.

The prevalence of small bodied fishes may be explained by partitioning of resources or niche space (Griffiths, 1986; Hutchinson & MacArthur, 1959), geography (Allen et al., 2006), and habitat heterogeneity (Bakker & Kelt, 2000; Brown & Nicoletto, 1991; Hutchinson &

MacArthur, 1959) in these complex systems. A strong relationship between body size and

79 habitat complexity is expected in the riverine fish communities of South America and reef fish communities of the Indo-Pacific (Weitzman & Vari, 1988). Relatedly, these patterns do occur in

North American freshwater habitats, with body size divergence associated with macrohabitat affiliations and migratory behaviours (Griffiths, 2012). However, no large-scale study has quantitatively assessed the potential impact of these factors on body size distribution of all fishes distributed globally.

Apart from Albert & Johnson (2012), few studies of body size distribution or evolution in fishes have truly encompassed the vast species richness across fish lineages. Furthermore, few studies have explicitly compared evolutionary patterns among the major classes of living fishes, namely chondrichthyan (cartilaginous fishes), actinopterygian (bony fishes), and sarcopterygian (lobe-finned fishes) lineages. The largest bony fishes generally do not attain the maximum absolute body sizes of the largest elasmobranchs throughout their evolutionary history (Freedman & Noakes, 2002), yet no studies have confirmed this observation across taxonomic hierarchy and habitat. It is also apparent that some closely related lineages within the major classes occupy larger or smaller regions of body size space in different habitats, but there is no empirical evidence that this pattern is well supported more broadly across fishes.

Comparatively lower species richness and different life histories (e.g. longevity, fecundity) among major lineages may drive divergence among lineage BSFDs (Dulvy & Reynolds, 1997).

I expected to see body size divergence between the major lineages of fishes regardless of clade age, species richness, environment affiliation, and macrohabitat if this variation in life history traits are driving body size divergence. However, I expected if any BSFD divergence occurs between lineages, these can be attributed to evolutionary and environmental factors as chondrichthyans are largely marine and have diverged from other lineages early in the evolutionary history of fishes.

80

The purpose of this chapter was to examine broad patterns of body size distributions among the major clades of extant fishes across different environments (marine versus freshwater) and macrohabitats (e.g. reef, river, lake). I tested whether freshwater fishes are significantly smaller and overwhelmingly right-skewed due to physical restrictions of habitat size imposed on body size evolution than marine fishes, which are expected to be skewed towards larger body size due to dispersal capabilities required for the patchy nature of reef habitats and the expanse of the open ocean. I also tested if similar deviations towards smaller size and right skew are seen in assemblages (i.e., pooling across taxa) affiliated with complex macrohabitat (i.e., riverine and reef) to maximize resource partitioning in these species rich and productive systems while assemblages affiliated with non-complex macrohabitat are larger and left-skewed due to more open, homogeneous expanses of the open ocean and lacustrine habitats.

I used a large dataset describing body size distribution parameters and summary statistics of

28,845 species of fishes among 385 families, as well as a newly generated phylogenetic hypothesis of 292 families, representing 80% of the family-level diversity, to test five hypotheses about the distribution of body size in fishes:

1) if distributions of body size across fish taxa differ between marine and freshwater

environments;

2) whether complex systems (i.e., river, reef) are characterized by significantly different

distributions than relatively simple systems (i.e., lake, non-reef);

3) whether elasmobranchs (sharks, skates, and rays) are on average larger than

actinopterygians (ray-finned fishes);

4) whether older clades are larger in body size due to directional evolution of body size;

81

5) whether increased lineage diversity, a proxy for net diversification, leads to reduction of

body size.

2.3 Methods

2.3.1 Data Acquisition

Body size, environment (marine, freshwater), macrohabitat (e.g. river, reef), and taxonomic data were collected from FishBase (Froese & Pauly, 2014) for 28,845 species of fishes (Table 2.1), representing 88.61% of known diversity. Not all body size measurements are the same in

FishBase (e.g. standard length, total length, disc width), but previous analysis showed that differences should have little effect on broad scale patterns (Steele & López-Fernández, 2014), therefore I maximized taxon sampling by including any measures available. All body size data were log-transformed to reduce measurement effects. Designations of environment and macrohabitat data were based on adult preferences, and manually retrieved from biological descriptions, locality data or cited literature available on the data page for each species (Froese

& Pauly, 2014). Species in saltwater or freshwater systems were designated as ‘Marine’ or

‘Freshwater’ respectively, while migratory species (e.g. ‘Catadromous’, ‘Anadromous’) were excluded from the dataset. Species occurring in primary macrohabitat types (i.e., ‘River’, ‘Lake’,

‘Reef’, ‘Non-Reef’) were used in the analyses of macrohabitat. Rare accounts within a habitat were also excluded. ‘Non-Reef’ was used for marine species designated as either demersal or pelagic on FishBase.

Taxonomic classification followed FishBase, with order and family level classification updated following Betancur-R. et al. (2013) for actinopterygians as well as Aschliman (2014) for batoids and Vélez-Zuazo and Agnarsson (2011) for selachians where necessary. I made my

82

Table 2.1: Diversity of species and families used to test for body size divergence. Species richness represents roughly 88% of known diversity. Families presented represent those that satisfied the 75% affiliation threshold for environment or macrohabitat, representing roughly 93% of families, and 78.4% of known species richness.

Grouping Diversity Freshwater Marine River Lake Reef Non-Reef Actinopterygii 27,652 13,706 13,327 10,769 1,375 4,337 9,592 species Elasmobranchii 1,017 34 962 34 0 99 884 species Actinopterygii 415 106 215 74 3 17 0 families Elasmobranchii 34 1 31 1 0 0 0 families

best efforts to accurately designate taxa based on the most current phylogenetic findings, however phylogenetic resolution of some groups is currently unavailable, and I have assumed monophyly. These are large assumptions that I could not assess, so I focused on general trends in body size distributions across environments and habitat, within and across broad taxonomic units. I assumed some degree of error in the data available for this chapter and present my findings with these caveats as the first, preliminary attempt to recover broad effects of environment and habitat association on body size across all extant fishes.

2.3.2 Characterizing the BSFD

We analyzed body size frequency distributions (BSFDs) of all fishes at Class, Order, and Family level as well as assemblages based on environment and macrohabitat affiliation. To test for divergence in body size occupation, diversity, and skew, I calculated mean, standard deviation, range, 25% and 75% quantiles and interquartile range (IQR) for each BSFD. Body size occupation along the body size gradient was assessed using mean, minimum and maximum body size. Body size diversity was assessed using standard deviation, interquartile range, and kurtosis (i.e., leptokurtic: peaked platykurtic: flat distribution; see Fig. 2.1). Skewness was

83 measured directly and comparing relative locations of the 25% and 75% quantiles along the body size gradient. Skewness indicates shifts in proportions of the distribution and the median relative to mean body size value, indicating biases in small or large taxa (see Fig. 2.1).

2.3.3 Creating Null Distribution and Expected Values

To test for significant deviations in BSFDs, observed BSFDs were compared to a null model to determine if observed BSFDs were divergent in summary statistics. The null model was created from the distribution of a clade, assemblage, or all fishes (i.e., null distribution) specific to the question I wished to address to account for confounding effects of shared ancestry, diversity, and broader ecological selective pressures. I employed a bootstrapping method to randomly sample the null distribution without replacement and equal in size to the observed BSFD to account for species richness, and calculated summary statistics of the null distribution. This was repeated 1000 times and the distribution of each summary statistic was used as the null expectation. If the statistic of the observed BSFD fell outside of the 95% confidence interval of the null expectation (two-tailed) I concluded that the observed BSFD is significantly deviating from random. If observed BSFDs showed few or no deviations from the null expectation, body size within a group can be considered a random subset of body sizes attained in fishes.

Partitioning of body size space among lineages or groups would result in significantly different summary statistics from the null expectation in opposing directions.

2.3.4 Testing for Correlation between BSFD and Evolution

Without a species level phylogeny of all fishes, I restricted my analysis of evolution of body size to testing for correlation between clade ages and species richness of families to the

84 distribution parameters above. Using phylogenetic generalized least squares regression, I first tested whether there was a significant association between clade ages and body size occupation

(mean, median, minimum, and maximum body size). Relationships among 292 families were attained from Betancur-R. et al. (2017). I then tested for correlation between species richness and body size occupation.

2.3.5 Testing for divergence due to Ecology

To test for divergence of BSFD among environment and macrohabitat, I partitioned the dataset into environment and macrohabitat assemblages and calculated distribution parameters from above. I compared each of theses observed BSFDs to their null model, and used broad deviations across taxonomic hierarchy (i.e., class, order), environment, and macrohabitat to develop hypotheses of how family BSFDs should deviate according to these same factors. I assigned families to an environment and macrohabitat, each having 75% or higher of species affiliated with the same environment or macrohabitat. In this way, most species within a family are evolving under the same selective pressures that I expect to influence body size evolution constrained transitions among environment or macrohabitat to be relatively rare (>25%). To account for potential deviations in body size distributions due to shared ancestry among families within an order, I used a coarse taxonomic correction for analyses at the family level, resampling from the BSFD of a family’s respective order. If environment or macrohabitat impose common pressures influencing directionality or constraint on body size evolution across families, I would expect to see similar deviations in family BSFDs to hypothesized deviations described for the BSFD of fish assemblages in each environment or habitat. I tested for differences in parameter means among environments using unpaired t-tests. I also tested for divergence in mean body size within those families that did not meet the threshold (n=23) using

85 paired t-tests. Using phylogenetic generalized least squares regression, I then tested whether there was a significant association between clade ages and body size occupation, diversity, and skew.

2.4 Results

2.4.1 Broad Scale Patterns across Environment and Habitat

Strong divergence between BSFDs of environment assemblages was found using the bootstrapping approach (Table 2.2; Fig. 2.2). Freshwater fishes had lower mean body size

(MeanF=1.038; MeanM=1.273), body size diversity (SDF=0.372; SDM=0.456), and are more right-skewed (SkewF=0.527; SkewM=0.119) than expected under the null model and as compared to marine fishes (Table 2.2; Fig. 2.2).

Distributions of body size within macrohabitat assemblages were strikingly similar to patterns found within freshwater and marine environments (Fig. 2.2). In both rivers and lakes, fishes appear to follow the same patterns as freshwater fishes overall, namely reduced mean body size (MeanRi=1.017, MeanL=1.130), reduced diversity (SDRi=0.368, SDL=0.299), and right skew (SkewRi=0.509, SkewL=0.531). Riverine fishes were found to be smaller, less diverse, and less right skewed than lacustrine fishes. Reef and non-reef fishes both follow comparable trends to the distribution of marine fishes overall. Both assemblages experience significant increases in mean body size (MeanR=1.192, MeanNR=1.314), increased diversity (SDR=0.466, SDNR=0.451),

86 and were significantly less right-skewed than the null expectation (SkewR=0.259,

SkewNR=0.105).

Figure 2.2: BSFDs of fishes a) Actinopterygii (light blue) and Elamobranchii (dark blue) b)

Freshwater (green) and Marine (blue) c) Non-Reef (dark blue), Reef (turquoise), Lake (dark green) and Rivers (green) against the distribution of all fishes (grey).

87

Table 2.2: Summary statistics for Actinopterygii and Elasmobranchii, environment and macrohabitat assemblages with Body Size

Frequency Distributions (BSFDs) deviating from the null model attained from subsampling all fishes (log10 transformed length in mm). An arrow next to statistical values indicates significantly different values as compared to a random phylogenetic distribution attained by bootstrap simulations (p<0.05). Direction of deviation relative to the null model is indicated by the orientation of the arrow. Standard Deviation (St Dev); Interquartile Range (IQR).

Occupation Diversity BSFD Shape Habitat Type Mean Min. Size Max. Size IQR St. Dev. Kurtosis 25% 75% Skew Quant. Quant. Actinopterygii 1.133↓ -0.097 3.041 0.572↓ 0.414↓ -0.117↓ 0.833↓ 1.405↓ 0.335↓ Freshwater 1.035↓ 0.000 2.699↓ 0.495↓ 0.370↓ 0.255↑ 0.778↓ 1.274↓ 0.527↑ Marine 1.223↑ -0.097 3.041 0.594↓ 0.428↓ -0.228↓ 0.924↑ 1.519↑ 0.055↓ Elasmobranchii 1.906↑ 0.613↑ 3.301 0.408↓ 0.329↓ 0.627↑ 1.679↑ 2.086↑ 0.504 Freshwater 1.684↑ 1.130↑ 2.380 0.245↓ 0.239↓ 0.825 1.547↑ 1.792↑ 0.412 Marine 1.906↑ 0.613↑ 3.301 0.401↓ 0.323↓ 0.695↑ 1.685↑ 2.086↑ 0.486 Freshwater 1.038↓ 0.000 2.699↓ 0.501↓ 0.372↓ 0.244↑ 0.778↓ 1.279↓ 0.527↑ Marine 1.273↑ -0.097 3.301 0.648↑ 0.456↑ -0.182↓ 0.954↑ 1.602↑ 0.119↓ Lake 1.130↓ 0.342 2.301↓ 0.372↓ 0.299↓ 0.528↑ 0.929↑ 1.301↓ 0.531↑ River 1.017↓ 0.000↓ 2.653↓ 0.507↓ 0.368↓ 0.201↑ 0.748↓ 1.255↓ 0.509↑ Non-Reef 1.314↑ -0.097 3.301 0.613 0.451↑ -0.060 1.000↑ 1.613↑ 0.105↓ Reef 1.192↑ -0.097 2.959 0.662↑ 0.466↑ -0.306↓ 0.869↑ 1.531↑ 0.259↓

88

2.4.2 Divergence Among the Major Lineages of Fishes

Divergence was found between Actinopterygii and Elasmobranchii (Table 2.2; Fig. 2.2). actinopterygians were smaller in body size (MeanA=1.133), had lower diversity (SDA=0.572), and was less right-skewed (SkewA=0.335) than expected based on the null. Elasmobranchs were larger in body size (MeanE=1.906), and had lower diversity (SDE=0.408); however, skew of

Elasmobranchs (SkewE=0.504) did not deviate from the null expectation. Within actinopterygians, mean body size diverged across environment with freshwater actinopterygians having smaller mean body size than marine actinopterygians (MeanAF=1.035, MeanAM=1.223).

Body size diversity was significantly reduced compared to the null model in freshwater and marine actinopterygians (SDAF=0.495, SDAM=0.594), with freshwater being more reduced overall. The BSFD of actinopterygians overall was significantly less right-skewed than expected, however strong divergence occurred across environment (SkewAF =0.527,

SkewAM=0.055). Mean body size was larger in marine and freshwater elasmobranchs

(MeanEF=1.684, MeanEM=1.906), with freshwater elasmobranchs smaller than their marine counterparts. Body size diversity was significantly reduced in both freshwater and marine elasmobranchs (SDEF=0.239, SDEM=0.323), with freshwater elasmobranchs showing a narrower range of body size overall. Body size distributions of both freshwater and marine elasmobranchs were right-skewed; however, values were not significantly different than expected under the null model (SkewEF=0.412, SkewEM=0.486).

89 2.4.3 Body Size Divergence of Families across Environment

As predicted, mean family body size is typically lower on average in freshwater environments than marine environments (x̅MeanF=1.125, x̅MeanM=1.410, df=351, P-value<0.001). Within

Actinopterygii, mean family body size is significantly lower overall than marine families

(x̅MeanAF=1.121, x̅MeanAM =1.336, df=319, P-value<0.001; x̅MinAF=0.673, x̅MinAM =0.883, df=319, P-value<0.001; x̅MaxAF=1.678, x̅MaxAM =1.77, df=319, P-value<0.05) (Table 2.3).

Data within Elasmobranchii were insufficient to test for significant family deviations between environments, but freshwater Potamotrygonidae (met affiliation threshold) had reduced mean body size compared to marine families, and was significantly lower than expected (Table 2.4;

Fig. 2.4). Results of the bootstrapping method show that when corrections are made for taxonomic affiliation, body size was only reduced in 35 of 106 actinopterygian families analyzed (33.02%), while 40 families (37.74%) have significantly larger body size in freshwater than expected under the null model (Fig. 2.3, Appendix 2.1). In marine waters, 55 of 215 actinopterygian (Fig. 2.3) families (25.58%) and 8 of 31 of elasmobranch (Fig. 2.4) families

(25.81%) had significantly larger body size, while 45 of 215 actinopterygian families (20.93%) and 11 of 31 elasmobranch families (35.48%) have significantly reduced body size (Appendix

2.1). Of 23 families that did not meet the threshold, only 7 families showed larger mean body size within freshwater, and overall freshwater fishes were significantly smaller than marine fishes (x̅MeanF=1.160, x̅MeanM =1.222, df=22, P-value=0.048).

Body size diversity did not significantly differ between marine and freshwater families

(Table 2.3; Table 2.4). Results of the bootstrapping methods found diversity is often significantly reduced as compared to the null model. In freshwater, 62 actinopterygian families

(58.49%) have significantly reduced body size diversity, while only 2 actinopterygian families

90

(1.89%) have significant increases in body size diversity (Fig. 2.3). Potamotrygonidae also has reduced body size diversity (Fig. 2.4). In marine families, 92 actinopterygian (Fig. 2.3) families

(42.79%) and 13 elasmobranch (Fig. 2.4) families (41.94%) have significantly lower body size diversity, while only 3 actinopterygian families (1.40%) have increased body size diversity.

As predicted, freshwater family BSFDs were typically right-skewed (x̅SkewF=0.260, df=106, P-value<0.001) and skew of marine family BSFDs overall was not found to significantly differ from a normal distribution (Skew=0) (Table 2.3; Table 2.4). Bootstrapping results found that in freshwaters, 13 actinopterygian families (12.26%) were significantly right- skewed, 17 actinopterygian families (16.04%) were more left-skewed than expected under the null model (Fig. 2.3), and Potamotrygonidae did not significantly differ from the null model

(Fig. 2.4). In marine environments, 9 actinopterygian (Fig. 2.3) families (4.19%) were significantly right-skewed while 19 actinopterygian (8.83%) and 6 elasmobranch (Fig. 2.4) families (19.35%) were significantly left-skewed.

2.4.4 Body Size Divergence of Families across Macrohabitat

Families likely experienced several transitions between macrohabitats within environments (i.e.,

Lake – River, Reef – Non-Reef), whereas transitions between macrohabitats among environments appear to be increasingly rare. As a result, I was unable to robustly assess divergence between macrohabitats as there were significantly fewer families, particularly in

Elasmobranchii, that were strongly affiliated with a single macrohabitat. Conclusions drawn for the effects of macrohabitat on body size distributions are weak in this study, and transitions

91

Figure 2.3: BSFDs of Actinopterygian families among environments and macrohabitats. Distributions of each family are represented as a coloured line against the distribution of all fishes in a given environment. Pie charts indicated the number of families that significantly deviated from the null distribution in occupation, diversity, and skew (white: no change, lighter: decrease, darker: increase).

92

Table 2.3: Summary statistics for Actinopterygian families, across environment and macrohabitat assemblages (log10 transformed length in mm). Family mean and ranges of each parameter are presented. Standard Deviation (St Dev); Interquartile Range (IQR).

Occupation Mean Min. Size Max. Size Habitat Type Mean St Dev Range Mean St Dev Range Mean St Dev Range Freshwater 1.121 0.369 0.194-2.203 0.673 0.411 0.000-1.854 1.678 0.474 0.301-2.699 Marine 1.336 0.340 0.385-2.511 0.883 0.415 -0.0969-2.265 1.778 0.372 0.580-2.699 Lake 1.095 0.192 0.882-1.256 0.864 0.254 0.653-1.14 1.376 0.073 1.301-1.447 River 1.050 0.343 0.194-1.973 0.613 0.402 0.000-1.954 1.615 0.478 0.301-2.556 Non-Reef NA NA NA NA NA NA NA NA NA Reef 1.345 0.293 0.814-1.845 0.906 0.354 0.342-1.653 1.808 0.269 1.230-2.360 Diversity IQR St. Dev Kurtosis Habitat Type Mean St Dev Range Mean St Dev Range Mean St Dev Range Freshwater 0.296 0.154 0.010-0.786 0.230 0.090 0.025-0.448 -0.083 1.759 -2.253-7.204 Marine 0.271 0.116 0.038-0.699 0.222 0.072 0.052-0.440 -0.461 1.199 -2.146-8.581 Lake 0.168 0.009 0.162-0.178 0.147 0.047 0.092-0.179 -0.851 0.880 -1.866- -0297 River 0.289 0.153 0.010-0.693 0.225 0.092 0.025-0.448 -0.039 1.740 -2.253-6.594 Non-Reef NA NA NA NA NA NA NA NA NA Reef 0.270 0.089 0.120-0.447 0.195 0.055 0.105-0.292 -0.364 0.721 -1.803-0.681 BSFD Shape 25% Quant. 75% Quant. Skew Habitat Type Mean St Dev Range Mean St Dev Range Mean St Dev Range Freshwater 0.966 0.353 0.153-2.109 1.262 0.397 0.242-2.301 0.266 0.565 -1.178-2.116 Marine 1.203 0.349 0.322-2.393 1.474 0.345 0.488-2.645 -0.037 0.551 -2.718-1.428 Lake 0.989 0.213 0.750-1.157 1.157 0.204 0.928-1.321 1.182 0.571 -0.191-0.839 River 0.905 0.333 0.154-0.1.954 1.194 0.362 0.242-2.079 0.282 0.580 -1.178-2.116 Non-Reef NA NA NA NA NA NA NA NA NA Reef 1.211 0.320 0.658-1.799 1.481 0.288 0.954-1.919 0.014 0.369 -0.643-0.819

93

Table 2.4: Summary statistics for Elasmobranch families, across environment and macrohabitat assemblages (log10 transformed length in mm). Family mean and ranges of each parameter are presented. Standard Deviation (St Dev); Interquartile Range (IQR).

Occupation Mean Min. Size Max. Size Habitat Type Mean St Dev Range Mean St Dev Range Mean St Dev Range Freshwater 1.621 NA NA 1.130 NA NA 1.978 NA NA Marine 1.920 0.258 1.494-2.605 1.503 0.352 0.613-2.484 2.336 0.334 1.716-2.959 Diversity IQR St. Dev Kurtosis Habitat Type Mean St Dev Range Mean St Dev Range Mean St Dev Range Freshwater 0.232 NA NA 0.204 NA NA -0.257 NA NA Marine 0.249 0.113 0.076-0.565 0.215 0.073 0.103-0.346 -0.507 1.184 -2.105-3.462 BSFD Shape 25% Quant. 75% Quant. Skew Habitat Type Mean St Dev Range Mean St Dev Range Mean St Dev Range Freshwater 1.54 NA NA 1.774 NA NA -0.384 NA NA Marine 1.794 0.259 1.364-2.544 2.043 0.278 1.596-2.620 0.022 0.505 -1.046-1.149

94

Figure 2.4: BSFDs of Elasmobranch families within marine environments. Distributions of each family are represented as a coloured line against the distribution of all marine fishes. Pie charts indicated the number of families that significantly deviated from the null distribution in occupation, diversity, and skew (white: no change, lighter: decrease, darker: increase).

between macrohabitat must be examined with higher taxonomic resolution. I focus on distribution patterns found within complex macrohabitats (river and reef) where families are more strongly affiliated (Table 2.3; Table 2.4), and within Actinopterygii where diversity allows. In general, fishes in both habitats appear to follow the same general trends as freshwater and marine fishes, respectively (Fig. 2.3). Divergence from expectation appears to occur more frequently within freshwater riverine fishes as compared to marine reef fishes with increases or decreases in mean family body size and reductions in diversity being more prevalent than changes in skew. Interestingly, variation in parameters among riverine and reef-associated

95 families differs considerably (Table 2.4). Overall, riverine families are smaller than reef- associated families, have higher diversity, and are more right-skewed on average, but show much higher variability among families than marine families.

2.4.5 Body Size Divergence across Taxonomic Hierarchy

We found that divergence of distributions taxonomically occurs in occupation along the body size gradient at the highest taxonomic unit, with actinopterygian fishes dominating the small body size region. Similarly, mean body size was larger than expected in most orders (64%), with only 17 orders (20%) exhibiting smaller mean body size and all these found within the class

Actinopterygii (Fig. 2.5). At the family level where taxa are commonly partitioned between environments, roughly a third of taxa were found to be significantly larger than random subsampling of respective orders while another third were significantly smaller. Diversity is significantly reduced in all classes and most orders (77%) with no taxonomic groups showing increases in body size diversity. At the family level, roughly half of taxonomic groups show reduced body size diversity compared to diversity seen within respective orders, and half show diversity levels expected from order diversity (Appendix 2.1). Six families, however, show significant increases in body size diversity as compared to the null model. Skewness values are highly variable among classes, but the majority of orders (60%) were not significantly different from expected values from the class distribution, and were typically left-skewed if divergent.

This pattern was stronger at the family level, with 80% of families showing no significant differences from their respective orders, and deviating BSFDs were twice as likely to be left- skewed.

96

Figure 2.5: Deviations experienced in BSFDs of taxonomic groups, as compared to distributions of respective inclusive clades (e.g. Actinopterygii to all fishes, Characiformes to Actinopterygii, Characidae to Characiformes). Pie charts indicated the number of taxa that significantly deviated from the null distribution in occupation, diversity, and skew (white: no change, lighter: decrease, darker: increase).

97 2.4.6 Correlation between Clade Age, Richness, and Ecology

We found no significant correlation between clade ages of families and their body size occupation parameters. I found a slight trend towards larger occupation values as clades increase in size, but it appears that this negligible trend is driven by the four oldest lineages within the phylogeny, Rajidae, Polypteridae, Acipenseridae, and Lepisosteidae which are all large bodied, species poor, and relatively conserved in body size values. In addition, due to the focus of the phylogeny, the regression does not include most chondrichthyans which may bias my results. I found significant negative relationships between mean, median, and minimum body size parameters and species richness, as well as a significant positive relationship between maximum body size and species richness. As diversity increases, there is a strong trend towards diversifying in body size, as well as an overall bias towards small body size. I also found that affiliation to marine environments (percentage of members only found in marine waters) had a significant positive relationship with mean, median, and minimum body size. Species richness did not drive this, as there was no correlation between marine affiliation and species richness of families.

2.5 Discussion

2.5.1 Marine Fishes are Larger than Freshwater Fishes

A primary goal of this chapter was to determine whether environment (i.e., marine vs. freshwater) and macrohabitat (i.e., river, lake, reef, non-reef) consistently predict patterns of ecological assemblage or clade body size distributions. When compared to a null distribution from all fishes, I found the freshwater assemblage to be significantly smaller in mean body size

98 than expected, whereas the marine assemblage was significantly larger. Despite an intuitive assumption that elasmobranchs (sharks, skates, and rays) would drive this increase in observed body size in the marine fishes, the pattern was consistent across both ray-finned fishes and cartilaginous fishes at both the class and family level. It appears repeated transitions to marine and freshwaters have resulted in consistent selection towards large and small body size, respectively; with fewer transitions between environments in elasmobranchs, this is hard to assess. Despite this, elasmobranchs in freshwater (e.g. Potamotrygonidae) are generally smaller than their marine counterparts. Moreover, when families that include both marine and freshwater taxa were analyzed, I found a clear tendency for marine lineages to attain larger body sizes than those of their freshwater relatives. Minimum body size in either environment was not significantly different than expected from the null model and not quantitatively different between marine and freshwater assemblages. This suggests that there is a hard bound on minimum adult body size in fishes independent of environment. Interestingly, I found significantly lower maximum body size in freshwater and a larger maximum body size in marine waters. Expansion of the marine assemblage BSFD range is mainly due to the unique region of large body size space occupied by marine cartilaginous fishes (Freedman & Noakes,

2002). However, this was also supported at the family level, suggesting fishes from both major lineages do indeed evolve larger body sizes repeatedly in marine environments. So, why are marine fishes larger than freshwater fishes?

Fishes have transitioned between freshwater and marine environments numerous times in both directions, particularly among (e.g. Vega & Wiens, 2012, Nelson et al. 2016).

As such, I cannot attribute body size divergence in extant marine and freshwater fishes to environmental and body size divergence of early lineages, as movement between body size and environments has occurred repeatedly throughout the evolutionary history of fishes (e.g. Vega &

99

Wiens, 2012, Nelson et al. 2016). Cope’s Rule proposes that lineages evolve larger body sizes over evolutionary time, and therefore older lineages may occupy regions of larger body size space than younger clades (Rensch, 1948). If this is true in fishes, then perhaps marine lineages are simply older than freshwater lineages. While many transitions from marine to freshwaters have occurred within major lineages (Vega & Wiens, 2012), no strong relationship between environment and age has been supported, and in fact freshwater fishes are typically found to be older than marine fishes (Vega & Wiens, 2012). I found no significant relationship between the clade age of families and BSFD parameters for body size occupation, suggesting that older families are not larger overall than younger families as predicted by Cope’s Rule. I did however find negative relationships between these parameters and species richness among families, suggesting that body size evolution is not independent of species diversification within families, and lineages are likely seeing higher net diversification of small-bodied species. I also found that maximum body size was correlated with species richness, suggesting that overall, body size diversity is also closely tied to lineage diversification in fishes.

Positive relationships between mean body size or body size diversity with habitat size have been found indirectly in terrestrial mammals (Maurer et al., 1992), terrestrial birds (Maurer et al., 1992), Anolis lizards (Thomas et al., 2009) and migratory fishes (Griffiths, 2012).

However, few studies directly assess habitat size constraints on body size evolution at local scales. I hypothesize that spatial limitation due to more prominent abiotic boundaries in freshwater may lead to divergence in mean body size compared to marine habitats, ultimately selecting for small body size to reduce competition or increase niche opportunity (Brown &

Wilson, 1956). Higher variation in velocity among temperate riverine habitats has been found to select for smaller-bodied fishes (Griffiths, 2006) and restricts smaller fishes to correspondingly small stream habitats (Leopold et al., 1964; Turner & Trexler, 1998). This association could be

100 widespread across freshwater fishes, the majority of which live in riverine environments, resulting in a bias toward small body size. To date, there is little empirical evidence of evolution towards smaller body size during the transition from marine to freshwaters (but see Santini et al., 2013), though many transitioning families remain to be studied. I found evidence that transitions to freshwater within families is associated with smaller body size, however should be addressed with more complete data and phylogenetic comparative methods to truly understand the correlation between body size evolution and habitat transitions.

Regardless of macrohabitat complexity (e.g., reef vs. non-reef, lakes vs. rivers), I found that fishes occupying macrohabitats within the same environment diverge in the same way as their respective environment assemblage. Mean body size was significantly lower than expected in lake and riverine macrohabitats (i.e., in freshwater) as compared to reef and non-reef fishes

(i.e., in marine habitats). Complex macrohabitats (i.e., river and reef) may correspond to smaller body sizes than their respective open macrohabitats (i.e., lake and non-reef), but does not negate the environment effect, with reef fishes still being larger than lacustrine and riverine fishes. I found that lake fishes on average tend towards slightly larger body sizes than riverine fishes, contrary to findings on European fishes (Griffiths, 2006) which showed riverine fishes to be larger than lacustrine species. Reef and non-reef marine taxa were almost indistinguishable from one another, contrary to my hypotheses where more complex habitats would promote greater body size diversity in addition to shifting BSFDs towards smaller sizes.

101 2.5.2 Elasmobranchs are Larger than Actinopterygians

Actinopterygians exhibit lower mean body sizes in general, as well as reduced minimum and maximum sizes when compared to elasmobranchs, regardless of marine or freshwater affinity. I find this unsurprising given that all elasmobranchs bear relatively large young, even in oviparous species, which presumably imposes a hard bound on minimum body size in sharks, rays, and their relatives (Freedman & Noakes, 2002). Actinopterygians also show greater disparity in body size than elasmobranchs, suggesting less constraint on ray-finned fish body size. This trend is consistent across environments, with marine and freshwater actinopterygians having smaller body sizes and higher body size diversity relative to marine elasmobranchs.

The largest bony fishes (teleosts) have never come close to attaining the maximum absolute body sizes of the largest elasmobranchs at any point in their evolutionary history

(Freedman & Noakes, 2002). My data demonstrate that, although elasmobranchs are always significantly larger than ray-finned fishes, both groups exhibit similar trends in body size distributions in shared habitats. For example, despite having larger average body sizes, freshwater elasmobranchs (e.g., Potamotrygonidae) are generally smaller than their marine counterparts. While fundamental differences in body size between actinopterygians and elasmobranchs appear to strongly correlate with phylogeny (e.g., larger average size in sharks and rays), ecological or habitat considerations may exert similar pressures (e.g., smaller body sizes in freshwater) for both clades on ecological, rather than deep phylogenetic timescales.

Potential mechanisms for this deep phylogenetic divide in body size evolution between ray-finned and cartilaginous fishes may lie in the stark contrast between their life history strategies - specifically those tied to juvenile growth rates, fecundity, and reproductive mode. Juvenile elasmobranchs in general are born large and have fast growth rates early in their

102 development (Freedman & Noakes, 2002). Most elasmobranchs are either ovoviviparous or viviparous leading to large offspring at time of parturition (Dulvy & Reynolds, 1997), particularly in pelagic species. Accordingly, freshwater potamotrygonid stingrays are similar to their marine relatives in terms of average litter size, size of offspring, and age-of-maturity, suggesting that elasmobranch life history traits are particularly “hard-wired,” despite habitat

(Rosa et al., 2010). Teleosts on the other hand, may have “farther” to grow when born at much smaller absolute body sizes than juvenile elasmobranchs, lacking the compensatory growth rates to match juvenile elasmobranchs so that teleosts never attain the larger sizes of cartilaginous fishes.

Mismatch of size at parturition is a plausible mechanism for why teleost fishes have never attained the large sizes typical of marine elasmobranchs. Interestingly, such a mismatch is also evident between teleosts and animals such as modern whales, ichthyosaurs, and even other extinct fishes, like arthodire placoderms (Freedman & Noakes, 2002; Long et al., 2009;

Trinajstic et al., 2015). No large planktivorous nor hypercarnivorous teleosts exist comparable to what is seen in extant elasmobranchs - both sharks and rays. Planktivores need to be highly mobile to track patchy, but locally abundant prey resources, and large body size facilitates long- distance foraging (Sims, 2003). Hypercarnivores such as white sharks and other lamniforms grow to large sizes and undergo trans-basin migrations in search of mates and prey (Sims,

2003). Actinopterygians seem to be the exception, not the rule, in oceans which seem to promote other clades to attain larger sizes. My results demonstrate it is imperative to distinguish between the major lineages of fishes when studying broad scale patterns, as there is large divergence in body size among these lineages.

103 2.5.3 Are Distributions Scale-Dependent in Fishes

We found that family parameters did not often differ from those expected under the null models, suggesting that body size is relatively conserved among taxonomic scales. I also found more variation in occupation and diversity parameters (excluding minimum body size) in freshwater affiliated families than in marine families when deviations did occur. While species richness is roughly even between freshwater and marine environments, there is significantly less habitat available in freshwaters (Shiklomanov, 1993; Grosberg et al., 2012). Competition, consequently, may be quite strong among freshwater fishes. Altering access to different prey resources, habitats, and competitive zones can be accomplished through divergence of body size, a convenient pathway by which animals can alter their niche (Brown & Wilson, 1956; Wilson,

1975). Divergence of body size may occur in species rich systems (i.e., tropical freshwaters), thereby allowing resource partitioning along a body size gradient to reduce competition. Yet I find that body size diversity is relatively low across class and families, as well as lower in the freshwater assemblage. This low variation in body size diversity is likely driven by phylogenetic autocorrelation due to shared ancestry of body size. Consistently low diversity within fish families suggests that absolute differences in body size distributions among broader assemblages (e.g. environment) primarily stem from variation among families in occupation rather than variation in diversity within closely related taxa. A similar result was seen in mammalian body size, with most variation in mean body size occurring among orders (Smith et al., 2004).

We found that mean body sizes of families were relatively conserved in marine environments, but less so in freshwater. Significant movement away from expected mean body size towards minimum and maximum theoretical boundaries is similar in frequency (i.e.,

104 number of divergent families) in both environments. However, freshwater families more frequently diverged in body size and exhibited higher variation compared to marine families. I also show that, despite these trends, mean body size within families was on average larger in the marine environment. Few exceptional freshwater families occupy regions of large body size, importantly these are still non-overlapping with the largest marine fishes. There was also no significant difference between the maximum body sizes of freshwater and marine families, while the minimum body sizes were larger in marine fishes. This reaffirms that although there is considerable overlap in the size distributions of families in freshwater and marine environments, and that evolution of more extreme sizes can occur in both freshwater and marine environments, there is strong selection for reduced body size in freshwater.

The propensity of right-skewed BSFDs, or the biased accumulation of small-bodied species within a lineage, across taxonomic hierarchy and environment was largely supported in this study. Additionally, freshwater fishes were found to be more right-skewed than marine fishes, a pattern consistent at the family level. The propensity of right-skew has been attributed to several mechanisms (see Allen et al., 2006 for review). I hypothesize that the prevalence of right-skew in freshwater as compared to marine habitats is largely due to the limited space and highly complex networks (Lowe-McConnell, 1975) driving body size reduction across lineages, particularly in species rich, tropical riverine systems (Weitzman & Vari, 1988). Following the textural discontinuity hypothesis (Holling, 1992), I expect this heterogeneity and limited space to result in biased accumulation of small-bodied species, allowing for dense species packing within spatially restricted freshwater habitat.

Habitat size may have a strong influence on body size distributions of regional species pools. In North America, the majority of riverine systems are first- and second- order

(headwater streams) throughout their entire length (Leopold et al., 1964). Distributions of body

105 size in taxa occupying these environments have been proposed to mirror (Knouft & Page, 2003) the right skew of habitat size distributions (Leopold et al., 1964). Limited habitat heterogeneity in these systems could also allow higher dispersal rates among populations and thus reduce rates of speciation in small-bodied fishes in temperate freshwaters (Knouft & Page, 2003). Though channel size and complexity have not been quantified extensively in temperate or tropical regions, it is possible that increased heterogeneity among habitats in tropical regions leads to increases in speciation, with a bias for small-bodied lineages. Extreme differences in velocity and substrate at confluences of main channels and their tributaries or drastic changes in slope may be more common in tropical systems (e.g. Hoeinghaus et al., 2004; Silva et al., 2016;

Röpke et al., 2017; Lowe-McConnell, 1975). This increase heterogeneity (Kenworthy &

Rhoads, 1995) could act as isolating barriers to small-bodied populations (Plaut, 2001), leading to increases in speciation of small-bodied lineages in the tropics. Additionally, if the distribution of freshwater habitat size is consistently right-skewed, fish assemblages within tropical freshwaters may be overwhelmingly right-skewed due to constraints imposed by limited space and resources of the habitats they inhabit (Knouft & Page, 2003). This could, in part, explain strengthening of right-skew at lower latitudes (Griffiths, 2012) as well as the apparent increase in proportions of small-bodied species in tropical freshwaters, particularly of South America

(Weitzman & Vari, 1988). While skewness was not tested, reduction of body size at lower latitudes was also found in riverine systems (Blanchet et al., 2010) and marine coastal environments (Fisher et al., 2010), suggesting that physiological responses to climate regimes are also likely driving patterns of body size distribution along broad spatial scales rather than habitat size and complexity alone.

It has been shown in several taxa that body size distribution parameters, namely distribution mean and shape, changes widely across scales (e.g. Coetzee et al., 2013; Cox et al.,

106

2011; Smith et al., 2004). Despite the incredible body size diversity across the group, mammals have been shown to occupy “characteristic” body size regions (i.e., divergence in occupation), which was largely prominent at the order level (Smith et al., 2004). This suggests that the body size distributions of mammal orders are not replicating the overarching distribution of mammals, but are rather divergent from the mammal distribution and each other (Fig. 2.1h). I find in fishes that skew was largely conserved across taxonomic scales, that diversity was reduced in less inclusive clades (i.e., families), and that occupation was more divergent among families. Higher variation in freshwater family occupation suggests a more even distribution or possibly divergence of families across the body size gradient. In contrast, approximately half of marine family BSFDs are centered about the mean body size of marine fishes assemblage, meaning greater overlap of marine family BSFDs than in freshwater families. Only 45 families across both environments did not significantly differ in any parameter, suggesting that most assemblage distributions (at the class or environment level) are due to cumulative effects of diverging distributions (Fig. 2.1h) rather than primarily replicating distributions at the family level (Fig. 2.1g). More of this scale-dependent divergence (Fig. 2.1h) seems to occur in freshwater, with most deviations occurring in occupation, rather than changes to shape or diversity. While I did not asses the influence of geography on distributions, I expect that distribution patterns among closely related freshwater or marine taxa would be similar across respective continents and oceans, as seen in a broad study of mammals (Smith et al., 2004).

107 2.5.4 Conclusion

We find that freshwater fishes are predominantly right-skewed towards having smaller body sizes than marine fishes, and this trend is recovered in actinopterygians as well as potamotrygonid stingrays, the only extant monophyletic elasmobranch lineage to have diversified in freshwater. Although elasmobranchs are larger on average than their teleost counterparts, regardless of macrohabitat, both major lineages of fishes follow similar trends in body size diversity and skew according to whether they are marine or freshwater affiliated. I propose that differences between poikilothermic and homeothermic vertebrates is present, but subject to similar constraints at smaller body sizes. Finally, I recover evidence that partitioning of body size distributions in certain macrohabitats shows different patterns between marine and freshwater assemblages with freshwater groups partitioning total body size space in non- overlapping, sequential segments, while marine assemblages tend to be more similar to one another in their overall body size distribution, regardless of taxonomic affinity. The mechanisms by which fishes diverge in body size among environment and macrohabitat are still largely unknown, and require further examination of geographic, fine-scale ecological (e.g. community), and physiological scales within a strict phylogenetic framework to further elucidate the evolution and distribution of body size in fishes.

108

Chapter 3 Extreme Body Size Reduction, Morphological Modification, and Allometry in Neotropical Cichlid Fishes (Cichliformes: Cichlidae: Cichlinae)

Sarah Elizabeth Steele1 and Hernán López-Fernández1,2

1Department of Ecology and Evolutionary Biology, University of Toronto, 25 Wilcocks St.,

Toronto, Ontario M5S 3B2, Canada

2Department of Natural History, Royal Ontario Museum, 100 Queen’s Park, Toronto, Ontario

M5S 2C6, Canada

109 Chapter 3

3.1 Abstract

Incredible diversity of body sizes exists within and across vertebrate lineages. Miniaturization within a lineage appears to limit the morphological, reproductive, and ecological roles species can have. If true, why does extreme body size reduction occur repeatedly across the phylogeny of vertebrates and how can we identify these reductions in a biologically meaningful framework? Miniaturization is a common process in the evolutionary history of fishes, particularly in the waters of South America. Body size reduction and morphological changes have been noted across many families of fishes within the orders Siluriformes (catfishes) and

Characiformes (tetras and relatives), among others. While some suggestion of extreme reduction and the potential for miniaturization has been made for Neotropical cichlids (Cichlinae), no studies have quantitatively assessed the validity of these designations in this lineage. I used ancestral state reconstruction followed by a bootstrapping approach to identify lineages that were significantly reduced in body size compared to the ancestral form. I then compared representatives of these lineages to members of their sister taxa to determine if these reductions were biologically meaningful. I tested for significant deviations in ontogenetic trajectories that resulted in truncation of the ontogeny, and subsequently the retention of juvenile-like morphology. I complimented this with assessment of internal and external modification in the skull and sensory systems, which should be greatly affected by miniaturization. From this, I found that increased support for miniaturization was strongly associated with body size reduction, where cichlids with the smallest standard lengths exhibited the most paedomorphic or reduced forms. Miniaturization however was not strongly supported in all small-bodied taxa.

110 3.2 Introduction

Body size varies greatly across animal taxa and impacts a wide array of biological factors such as physiological constraints (Huxley, 1932; West et al., 1997), metabolism (Kleiber, 1932;

Huxley, 1932, West et al., 1997), life history (Blueweiss et al., 1978; Gillooly et al., 2002; West et al., 2001), and trophic interactions (Brown et al., 2004; Kingsolver & Pfennig, 2004). Many biological and ecological traits are known to scale with body size. Moreover body size divergence can reduce competition for resources and allow for increased species, ecological, and morphological diversity (Brown et al., 2004; Monroe & Bokma, 2009; Romanuk et al., 2011). In fishes, large body size is strongly associated with increased reproductive potential (e.g. egg size, egg number), yet body size reduction (see Appendix 3.1) is a relatively common phenomenon, particularly in freshwater taxa (Bennett & Conway, 2010). Many Neotropical fish families have repeatedly evolved both extremely large and extremely small-bodied species (Albert & Johnson,

2012; Steele & López-Fernández, 2014; Toussaint et al., 2016; Weitzman & Vari, 1988), and it is likely that body size evolution has impacted ecological interactions (Brown et al., 2004;

López-Fernández et al., 2013; Montaña & Winemiller, 2009).

Miniaturization (see Appendix 3.1) is the process of evolution towards extremely small body size within a lineage (Hanken & Wake, 1993). Miniaturization requires evolution below some critical size (i.e., threshold, see Appendix 3.1) beyond which important physiological and ecological functions are affected and subsequently require major biological changes for the lineage to persist over time (Hanken & Wake, 1993). Miniaturization has been investigated to some extent in vertebrates (Albert & Johnson, 2012; Hanken & Wake, 1993; Weitzman & Vari,

1988), with some qualitative and quantitative suggestions of how to define a miniature lineage.

Gould (1971) discussed a developmental approach in which the degree of shape change

111 associated with miniaturization results in either proportioned or un-proportioned dwarfs, which are terms I use throughout this chapter. Proportioned dwarfs experience little morphological change relative to their ancestral form; whereas un-proportioned dwarfs may develop either as achondroplastic (short limbs, same head and trunk) or infantilized (juvenile-like), experiencing significant morphological change relative to their ancestral form (see Appendix 3.1). Under this view, in addition to size decrease, miniaturization is also associated with shifts in developmental processes that distinguish extreme forms from non-extreme counterparts (McKinney &

McNamara, 1991; Hanken & Wake, 1993). Miniatures can be comparatively identified by examining their developmental trajectory (Fig. 3.1), morphological novelty, and paedomorphism (see Appendix 3.1) relative to sister taxa or estimated ancestral form. Reduction and structural simplification of the “bauplan” (e.g. loss of organs or organ systems, loss of bones, retention of tissues typically reabsorbed), leading to resemblance to subadult or embryonic stages of large-bodied relatives, and/or truncation of the ontogeny are all possible alterations of the developmental pathway of miniatures (Hanken & Wake, 1993). These definitions provide a clear, observable set of changes to the bauplan that should help in identifying a lineage experiencing extreme body size reduction. For the purposes of this chapter,

I define the miniaturization threshold as the size below which small-bodied lineages experience major modifications in adult morphology and ontogeny with respect to their sister taxa.

112

Figure 3.1: Theoretical threshold for proportional and un-proportional miniatures. The dashed line represents the ancestral ontogenetic curve, here approximated by sister taxa. This ontogenetic curve is partitioned into a) Development Period - Early ontogeny exhibiting rapid shape change over size, associated with juveniles, and b) Growth Period – Late ontogeny exhibiting reduced shape change over size, associated with adults. These two periods are separated by the ontogenetic threshold (dashed line), representing the transition from juvenile to adult morphology. The red line represents the growth curve of a miniature, truncated before the threshold and thus resembling the juvenile form – un-proportional dwarfism. The solid black line represents the growth curve of a miniature, truncated after the threshold and thus resembling the adult form – proportioned dwarfism.

Divergence in adult body size and shape is ultimately driven by changes in the timing of developmental events throughout the ontogeny of taxa, termed heterochrony, or in the spatial patterning of growth, termed heterotropy (McKinney & McNamara, 1991; Zelditch et al., 2000).

The underdevelopment of taxa as a result of temporal or spatial shifts in ontogenies can lead to paedomorphic states in both adult shape and size. It has been proposed that heterochrony may be more common in nature (McKinney & McNamara, 1991) and certainly more prevalent in the literature. Identifying these changes is fundamental in determining differences in growth patterns leading to variation in adult body size and morphology. If deviations in growth rate or duration are great enough in either direction, or if extreme deviations in growth patterning

113 occur, extreme body size in the adult of that lineage can evolve. While processes are difficult to confirm without developmental data, the covariation between shape and size through ontogeny

(i.e., ontogenetic allometry) can give some indication of what mechanism may cause shape and size divergence among taxa (Fig. 3.2). To accurately describe the patterns associated with body size evolution and whether a lineage can be described as peramorphic or paedomorphic, one must first understand the direction of evolutionary change of body size in each lineage relative to that of its ancestor.

Figure 3.2: Relationships of allometric and regression models. Stasis - no differences in orientation and no relationship between shape and size; model: simplest model (same slope/intercept, slope = 0). Progenesis - truncation of ‘dwarf’ ontogeny; model: association between shape and size (same slope/intercept). Post-displacement - delay of development exhibits parallel slopes but divergent intercepts; model: shape difference among species (same slope). Neoteny - slowing of growth rate, exhibits divergent slopes but same intercept; model: full model, incorporation an interaction between the independent variables, and tests for differences in the direction of ontogenetic trajectories given species identity. Black dashed – ancestral condition (here, sister as proxy); red solid – ‘dwarf’.

114

Evolution towards extreme body sizes has been recorded in many freshwater systems associated with tropical, highly complex habitats. The highest regional species diversity of freshwater fishes occurs in South America, with over 6000 species of fishes from 70 families and 21 orders (Reis et al., 2003; Froese & Pauly, 2016). It has been suggested that the diversity of fishes that have experienced extreme body size evolution is much higher in the rivers of

South America than anywhere else on earth (Weitzman & Vari, 1988). Understanding how body size shapes species and morphological diversity may elucidate some of the drivers of the immense diversity of fishes in this region. Neotropical cichlids, one of the most diverse clades of freshwater fishes in South America, has likely undergone several independent episodes of body size reduction in distantly related taxa at the genus level or higher (e.g. López-Fernández,

Honeycutt, Stiassny, & Winemiller, 2005; López-Fernández, Winemiller, & Honeycutt, 2010;

Steele & López-Fernández, 2014). If these lineages are evolutionarily becoming smaller compared to their ancestors, it is unclear whether the degree of change warrants designation as miniaturization. The term ‘dwarf’ has been used to describe the relative extreme body size decrease (see Appendix 3.1) in certain lineages of Neotropical cichlid fishes, however, there has been no formal criterion to define which taxa are considered dwarfs. Based on the publication of molecular phylogenies (e.g. López-Fernández et al., 2005, 2010), it is known that proposed

‘dwarfs’ occur in several independently evolving lineages. Historically, ‘dwarf’ cichlids have been restricted to the genera Apistogramma (including Apistogrammoides), Biotoecus,

Dicrossus, Crenicara, Nannacara, Mikrogeophagus, Laetacara and Taeniacara. The term appears to have been applied to these genera within the aquarium trade and its literature initially, based primarily on decreased size (particularly size at maturity) of these species, qualitative differences in habitat use (Barlow, 1991; Lowe-McConnell, 1991), and aggression in captive specimens (Linke & Staeck, 1984), and over time was adopted within the scientific literature

(e.g. Kullander, 1983). To date, there have not been rigorous, quantitative studies determining

115 whether these lineages are true dwarfs in the sense of the developmental and life-history literature (e.g. McKinney & McNamara, 1991; Hanken & Wake, 1993). Specifically, no studies have explored the potential differences in growth and morphology associated with miniaturization that in turn may have effects on the ecology of those lineages. In addition, the term “dwarf” was applied typically to monophyletic clades (genera noted above) which, containing one or more species, see considerable size conservatism (Steele & López-Fernández,

2014). Yet the term is not consistently applied to all relatively small-bodied species evolving within a genus (e.g. Teleocichla species nested within the genus Crenicichla, or species groups such as Crenicichla regani, C. compressiceps, C. wallacii).

Recent approaches to studying miniaturization empirically and broadly across taxa include calculating extreme deviations from mean body size (Albert & Johnson, 2012), finding evolutionary correlates with ecological parameters (Zimkus et al., 2012), or incorporating developmental and morphological condition (Zelditch et al, 2004; Angielczyk & Feldman,

2013; Galatius et al., 2011; Masters et al., 2014). Studies including only length or mass measurements to describe body size may perform poorly in identifying true miniaturization events. Statistically extreme body sizes may not correspond to major biological changes and are often difficult to apply widely across taxa to identify clear distinctions between extreme body size and more typical sizes. For example, the threshold for extreme body size reduction in fishes identified in previous works (2.6 cm Standard Length [SL] Weitzman & Vari, 1988, 1.4 cm

Total Length [TL] Albert & Johnson, 2012) are well below the size of most Neotropical cichlids that qualitatively exhibit reductive characters associated with miniaturization (Hanken & Wake,

1993; López-Fernández et al., 2005; Weitzman & Vari, 1988). Though sizes below these thresholds may be statistically extreme, it is not known whether they are associated with major

116 biological changes and do not clarify whether there is developmental distinction between potentially extreme body sizes and more typical sizes.

In this chapter, I examine body size, morphology, and ontogenetic allometry in proposed ‘dwarf’ lineages of Neotropical cichlids relative to their respective sister taxa. I intended to quantitatively test whether the designation of these small-bodied species as “dwarf” fishes has developmental basis that warrants formally treating them as biological miniatures. In a more general sense, I use Neotropical cichlids as a model to expand a conceptual and methodological framework (Zelditch et al, 2004; Angielczyk & Feldman, 2013) for assessing extreme body size reduction that can be applied broadly across fish taxa. I posit that, to be considered a miniature, a lineage must fit three criteria: 1) body size reduction relative to its ancestral body size; 2) reductive evolution (see Appendix 3.1) leading to paedomorphism; and

3) ontogenetic truncation relative to its sister taxon (Fig. 3.1). While variation in morphological response to body size reduction is expected among lineages, I expected all three requisites to be fulfilled to some degree. I further examined lineages that fulfilled the requisites of miniaturization to determine if adult morphology is proportional or un-proportional (see

Appendix 3.1; Gould, 1971), and formulated hypotheses about the developmental mechanisms that may have led to body size reduction, reductive evolution, and ontogenetic truncation. Using a phylogeny-informed comparative approach, the outcomes of evolutionary miniaturization in adult size and shape are assessed to infer potential processes. The methods proposed herein are intended to provide an intermediate approach that moves beyond using arbitrarily defined statistical thresholds to identify body size shifts (e.g. Albert & Johnson, 2012), while not requiring direct examination of the underlying developmental mechanisms of each taxon. Such a comparative approach should allow description of the evolution of ontogenetic changes across a phylogeny and generate hypotheses about developmental evolution in a macroevolutionary

117 context. Furthermore, hypothesized developmental mechanisms associated with observed evolutionary changes should be testable using evolutionary developmental biology analyses.

3.3 Methods

3.3.1 Validation of Geometric Morphometric Analyses, Sampling Rarefaction, and Sample Size Determination

To determine the number of specimens required to quantify ontogenetic allometry and the distribution of these specimens across ontogeny, I examined shape variation associated with size changes in a single species of Neotropical cichlid, Acarichthys heckelii. I measured, photographed, and digitized 176 specimens from a single sampling time and locality to assess the natural variation in the relationship between size and shape across a densely sampled ontogeny (11–98mm SL). I fit a linear model to the shape data (details below) to estimate the strength of the association between shape and size, as well as estimated the regression coefficients (Fig. 3.3). To determine the effects of sample size across the ontogeny on the estimation of regression coefficients, I used a bootstrap to randomly sample the pool of specimens 10,000 times each for small and large sample sizes (n=20 and n=100 respectively) and calculated regression coefficients. I then compared the coefficients estimated from small and large sample sizes to determine if small sample size imposes bias in the estimation of regression coefficients. With all specimens (n=176), I found a significant relationship between shape and size (intercept=0.156, slope=0.067, p-value=0.002). Randomly subsampling using large sample sizes found similar estimates of the regression coefficients despite missing 76 individuals randomly distributed across the ontogeny (intercept=0.143–0.173, mean=0.156; slope=0.062–0.074, mean=0.068). Randomly subsampling using small sample sizes also found

118 similar estimates of the regression coefficients despite missing 156 individuals randomly distributed across the ontogeny, though with a wider range of estimates (intercept=0.125–0.235, mean=0.163; slope=0.056–0.097, mean=0.070). When I sampled evenly across the distribution the range of estimates surprisingly were high, though the mean estimate of slope was closer to the ‘true’ population slope (intercept=0.141–0.180, mean=0.162; slope=0.062–0.081, mean=0.070). I then further constrained resampling to sample from the earliest ontogenetic stages (<15mm) evenly to the latest stages (>70mm), which considerably improved the estimations overall (Fig. 3.3) despite small changes to the range and mean (intercept=0.144–

0.180, mean=0.163; slope=0.064–0.080, mean=0.071). The coefficient of determination for large samples ranged from 0.302–0.499 (mean=0.386), for small sample sizes from 0.173–0.694

(mean=0.425), for small even samples from 0.233–0.513 (mean=0.399), and for small even samples capturing the extremes of ontogeny from 0.329–0.513 (mean=0.431).

We also assessed the impacts of these different sample sizes and distributions on the estimate of the threshold between development and growth (see below) using piece-wise regression. With all specimens (n=176), the threshold was estimated as 29.33 (st. error=0.316).

For large sample sizes, the threshold was estimated between 26.44–32.44 (mean=29.41) with standard error about the estimation ranging between 0.330–0.525 (mean=0.418). For small sample sizes randomly sampled, the threshold was estimated between 21.94–50.32

(mean=29.81) with standard error about the estimation ranging between 0.287–2.449

(mean=0.925). For small sample sizes evenly sampled, the threshold was estimated between

26.68–33.75 (mean=29.11)

119

Figure 3.3: Regression parameters simulated under various sample sizes and distributions. Sample distribution of parameter estimates (columns 1-4): 1) 100 randomly sampled individuals from all 176 Acarichthys heckellii (large sample size – grey); 2) 20 randomly sampled individuals (small sample size – red); 3) 20 individuals evenly distributed across the ontogeny from random start and end points; 4) 20 individuals evenly distributed across the ontogeny and capturing extremes of ontogeny at start and end points. Parameters estimated under distribution and sample conditions (rows a-e): a) Intercept of regression line fit to predicted shape and size (SL); b) Slope of regression line fit to predicted shape and size (SL); c) Coefficient of determination of regression fit; d) Threshold value (PSI) estimated using piece-wise regression, indicating the inflexion between development and growth; e) Standard error about the estimated threshold value (PSI). with standard error about the estimation ranging between 0.563–1.500 (mean=0.884). For small sample sizes evenly sampled and capturing extremes of ontogeny, the threshold was estimated

120 between 27.16–33.01 (mean=29.15) with standard error about the estimation ranging between

0.683–1.242 (mean=0.877). Therefore, the small sample size evenly distributed across the ontogeny and capturing both extremes gives an unbiased estimate of the threshold, despite more uncertainty in that estimate (Fig. 3.3).

3.3.2 Data Collection and Taxon Sampling

A total of 564 cichlid specimens in 27 species (Fig. 3.4) comprising 10 proposed ‘dwarf’ species and 12 representatives of their sister taxa were used to assess the ontogenetic difference among lineages. In addition to ‘dwarf’ taxa, I examined genera considered small-bodied (following

Arbour & López-Fernández, 2013; Steele & López-Fernández, 2014) to test if these fit the criteria of miniaturization expected in the proposed ‘dwarf’ cichlids. I compared Mazarunia charadrica, M. pala, and M. mazarunia to two members of their sister taxon, Guianacara dacrya and G. stergiosi. Other small-bodied cichlids were not available for ontogenetic analyses. Specimens used in this chapter were obtained from the collections of the Royal

Ontario Museum (ROM), Academy of Natural Science at Drexel University (ANSP), Texas A

& M University (Biodiversity and Research Teaching Collections – BRTC), University of Texas at Austin (Texas Natural History Collections -TNHC), the Museu de Zoologia da Pontificia

Universidade Catolica do Rio Grande do Sul (MCP) and the Museum of Comparative ,

Harvard University (MCZ). For each species, I strived to measure individuals from a single sampling event and geographic locality, if possible, to reduce variability in the data due to seasonal fluctuations and among-site variability influencing shape or growth curves.

121

Figure 3.4: ‘Dwarf’ Neotropical cichlids and sister taxa used for comparison. Standard length maximum attainable body size (SL), Standard length at threshold between juveniles and adults (ThL, breakpoint of piece-wise regression analysis of shape on size), and the percentage of ontogenetic size change completed at threshold size (PThL=ThL/SL) are presented for comparison across taxa.

Changes in shape over ontogeny were captured using geometric morphometrics of external morphology. Individuals spanning early juvenile stages to adults at maximum body size of each species were included in this chapter following the validation study using Acarichthys heckelii where possible (mean=20.96 individuals; range=9–31). To reduce inter-investigator variation and to maximize consistency of landmark identification during digitization, all internal

122 and external structures were pinned by SES. Photographs were taken of the left lateral view of each individual using a Nikon D750 DSLR camera with a 105mm Nikon AF-S VR Micro-

Nikkor lens by ZL. One hundred anatomical landmarks (landmarks 1:100: Fig. 3.5) were recorded on the left side of each individual (ZL/SES) in tpsDig (Rohlf, 2004) with particular focus on the skull elements and their proportions, as well as the lateral line, which are known to be strongly associated with ontogenetic allometry (Ponton & Mérigoux, 2000; Webb, 1990) and possibly with body size reduction (Webb, 1990; Weitzman & Vari, 1988) in cichlids.

Coordinate data were then imported into R (R Core Team, 2016) and inspected by SES, before statistical analyses, which were all carried out using this software. Generalized Procrustes superimposition was used to align the coordinate data and to remove variation as a result of rotation, translation and scaling of photographs and was completed using the function gpagen from the package ‘geomorph’ (Adams & Otárola-Castillo, 2013). Aligned Procrustes coordinates for all specimens were then used in all subsequent shape analyses. Standard length

(SL), measured from the tip of the upper lip to the end of the caudal peduncle (i.e., insertion of caudal fin rays into hypural plates), was measured for each specimen as an expression of body size. Maximum adult body size sampled from institution collections was compared to previously reported lengths attained from FishBase (Froese & Pauly, 2016) to determine if sampling was complete. Chronological ages of specimens were not available and difficult to retrieve from tropical fish species, thus size was used as a proxy for age (Frédérich & Sheets, 2010).

123

Figure 3.5: The two-dimensional external morphology landmarks used for shape analysis across the ontogeny of Neotropical cichlids. Solid circles indicate fixed landmarks; dashed circles indicate sliding semi-landmarks. 1 – Tip of snout; 2-6 – Surface of snout ventral to 7; 7 – Anterior-most end of scaled nape; 8-12 – Surface of scaled nape; 13 – Insertion of first spiny dorsal ray; 14-18 – Dorsal surface at insertions of spiny rays; 19 – Insertion of last spiny dorsal ray; 20 – Distal tip of last spiny dorsal ray; 21-22 – Dorsal surface at insertions of soft dorsal rays; 23 – Insertion of last soft dorsal ray; 24-26 – Dorsal surface of caudal peduncle; 27 – Dorsal insertion of caudal fin; 28 – Posterior-most point of body midline, intersection of caudal margin and lower lateral line (LLL); 29-31 – LLL; 32 – Anterior-most point of LLL; 33 – Ventral insertion of caudal fin; 34-36 – Ventral surface of caudal peduncle; 37 – Posterior insertion of anal fin; 38-40 – Ventral surface at insertion of anal fin rays; 41 – Anterior insertion of anal fin; 42-48 – Ventral surface between pelvic and anal fins; 49 – Anterior insertion of pelvic fin; 50-52 – Ventral surface between 53 and anal fin; 53 – Ventral tip of cleithrum; 54 – Intersection of posterior margin of operculum and ventral surface; 55 – Anterior tip of hyoid; 56-57 – Ventral surface of jaws; 58 – Anterodorsal point of lower lip; 59-66 – Perimeter of orbit; 67 – Dorsal point of preopercle; 68-74 – Posterior/ventral margin of abductor mandibulae; 75 – Ventral-most point of lacrimal process; 76-80 – Posterior margin of preopercle; 81 – Inflexion of posterior margin of preopercle; 82-84 – Ventral margin of preopercle; 85 – Anterior-most point of preopercle; 86 – Ventroposterior point of descending process of maxilla; 87 – Dorsal-most point of operculum; 88-92 – Posterior margin of operculum; 93 – Posteroventral point of operculum; 94 – Anterior-most point of upper lateral line (ULL); 95-97 – ULL; 98 – Posterior-most point of ULL; 99 – Dorsal insertion of pelvic fin; 100 – Ventral insertion of pelvic fin.

124

Comparative methods to estimate direction of body size evolution and sister taxa to dwarf lineages were applied on a time-calibrated molecular phylogeny (Maximum Clade

Credibility tree, MCC), as well as 1,000 randomly sampled chronograms from the posterior distribution of chronograms from which the MCC tree was obtained as per López-Fernández et al. (2010, 2013). To standardize comparisons between the MCC and among the 1000 posterior distribution trees, all chronograms were scaled to a relative time scale with total length of one

(López-Fernández et al., 2013). The phylogeny combines three mitochondrial and two nuclear markers for 154 species from 57 genera and all major tribes of Cichlinae (Astronotini, Cichlini,

Cichlasomatini, Chaetobranchini, Geophagini, Heroini, Retroculini) representing approximately

26% of currently described taxa. Further details of the phylogeny and time-calibration are given in López-Fernández et al. (2010, 2013). The placement of anomala was determined as per Ilves et al. (2017).

3.3.3 Body Size Reduction

To test the assumption that small-bodied taxa are truly reduced in size relative to their ancestors, ancestral state estimation of body size was performed on Neotropical cichlids using the function fastAnc in the package ‘phytools’ (Revell, 2009, 2012). I used maximum attainable body size of all species reported in the literature (Froese & Pauly, 2016) or sampled in institutional collections, whichever was larger, as tip data (see above). Divergence was tested using a resampling procedure assessing the amount of body size change between each of the tips and the most recent common ancestor (MRCA) of all Neotropical cichlids. The MRCA state was approximated via maximum likelihood, and an estimate of the MRCA state was sampled from within the resulting 95% confidence interval of the MRCA state. The observed divergence was then calculated for each of the 160 terminal taxa by measuring the distances between the MRCA

125 estimate to each tip trait value, creating a distribution of expected divergence. Finally, the observed divergence for each of the 160 terminal taxa was compared to the distribution of expected divergence. I determined if observed divergence for each taxon was higher or lower than the 97.5 and 0.025 quantiles of the distribution, respectively, indicating significant departure from the MRCA estimate. This was repeated 100,000 times and the percentage of significant departures from the expected divergence value was calculated for all taxa. This was also repeated via a clade-based approach, in which the divergence between each node in the tree was again compared to the sate of the MRCA of all Neotropical cichlids to determine if taxonomic groups were significantly divergent, or just particular members. To account for the potential influence of phylogenetic uncertainty on divergence estimate, I repeated this process using 1000 trees sampled from the posterior distribution (see above).

3.3.4 Paedomorphism in Lateral Line and Skeletal Elements

To assess the degree of reductive evolution, or simplification, that has potentially occurred in

Neotropical cichlids, I used descriptions of the condition of 8 lateral line and sensory system traits, and 10 skeletal traits in ‘dwarf’ taxa relative to their sister species based on the literature and, in a few cases, ongoing work (López-Fernández et al., 2005; HLF Unpublished). All dwarf taxa in this chapter were examined for reductive characters, with the exceptions of Teleocichla monogramma, which was unavailable for analyses, Apistogramma steindachneri, Crenicara punctulatum, and Laetacara thayeri which were substituted with Apistogramma pucallpaensis,

Crenicara latruncularium, and Laetacara dorsigera, respectively. I compared the development of the lateral line assessing the pattern of the upper and lower lateral lines following the developmental study of Webb (1990) in which eight patterns of connectivity were ranked from most paedomorphic (underdeveloped) to most peramorphic (overdeveloped). Although I did not

126 have developmental data to assess paedomorphism in the skeletal system, I estimated simplification via modification or loss of skeletal components in ‘dwarf’ taxa relative to large- bodied cichlids. For the lateral line (LL), I used the number of pitted scales in the upper lateral line (ULL) occurring in place of tubed scales, the number of preopercular LL pores, dentary LL pores, pattern of trunk canal LL (Webb, 1990), opening of the neurocranial LL foramina 4

(NLF4), opening of the neurocranial LL foramina 5 (NLF5), and number of lachrymal LL pores.

For skeletal modification, reduction, or loss, I used characterizations of the anterior laminar expansion of the first epibranchial, the antero-ventral laminar expansion of the second epibranchial, presence of interarcual cartilage, composition of the first pharyngobranchial, composition of the infraorbitals (4,5, and 6), fusion of the lachrymals, presence of infraorbital 3, shape of lachrymals, dorso-caudal laminar expansion of the lachrymal, and total number of vertebrae. Lateral line and skeletal characters were evaluated on alcohol-preserved or cleared and stained specimens as described in López-Fernández et al. (2005, 2012). Character states of taxa were described following López-Fernández et al. (2012). To quantify the degree of potential paedomorphism, reductions and modifications were coded as 0 (no change) or -1

(modification or loss) relative to the character state of sister taxa. Total amount of change across the 18 traits were summed together to roughly indicate which taxa display the most amount of sensory and skeletal modification.

127 3.3.5 Paedomorphism in External Morphology

A primary component of miniaturization theory suggests that adults of miniature taxa will resemble the juvenile form of the ancestral ontogeny, or that of sister taxa. Since methods that could be employed to estimate ancestral ontogenies are currently in development, and taxon sampling in this chapter would not allow comparisons across the phylogeny, I have compared

‘dwarf’ body shape to the ontogenies of its respective sister taxa. Procrustes ANOVA (Goodall,

1991; Klingenberg et al., 2002) was first performed on aligned coordinates to assess covariation of shape with size across ontogeny. Ontogenetic allometry was quantified for each species by fitting a linear regression model to the data using the function procD.allometry in the package

‘geomorph’ (Adams & Otárola-Castillo, 2013). Linear regression models were fitted using size, species, and their interactions as predictors of relative species differences in a global size-shape space to account for interspecies variability in morphology. Probability of shape variation for significance testing was estimated by comparison of observed shape data with a null model of distributions from 999 resampling permutations. The common allometric component (CAC, see

Mitteroecker et al., 2004) of shape data, regression shape scores (RegScore), and predicted values (PredVal) from a regression of shape on size were retained for analyses. The CAC was then used to assess the difference in shape among dwarf-sister pairs that was strictly associated with shape changes over size, and therefore changes that were due to ontogenetic allometry.

Euclidean distance of CAC was first measured between adults of ‘dwarf’ taxa and juveniles of sister taxa, then between adults of ‘dwarf’ taxa and adults of sister taxa. For each comparison, I used 5 individuals from each end of the sister taxon ontogeny and 5 largest adults of each

‘dwarf’ species. Due to low taxon sampling, I could not employ a randomized sampling procedure as in the case of maximum body size. To determine if ‘dwarf’ morphology resembled

128 juveniles over adults, I tested for significant differences between the means of the dwarf- juvenile and dwarf-adult distances using a two-sample unpaired t-test via the function t.test.

3.3.6 Significant Reductions in Duration of Development and Growth

To test if ‘dwarf’ taxa exhibit significant reduction in the duration of development (ontogenetic length), I compared the size and shape of ‘dwarf’ taxa at the termination of ontogeny to the ontogenetic progress of sister taxa at the same size. I partitioned sister ontogenies into a developmental period and growth period. The developmental period is characterized by rapid shape change and is largely associated with juvenile shape. The growth period is characterized by significant reduction in the degree of shape change across the same amount of size change and is associated with adult shape (Fig. 3.1). Ontogenetic thresholds (see Appendix 3.1) between development and growth were estimated via the function segmented in the package

‘segmented’ (Muggeo, 2008), and differences among segment slopes were tested using davies.test. This package aims to estimate linear regression models that may have a segmented relationship between the predictor and response variables, estimating both the slopes of each segment in the relationship and the breakpoints (inflection points) between these segments.

Through this chapter, I used these estimated breakpoints as the value for ontogenetic thresholds for each species (see above). In this way, the early rapid changes in development could be separated from the theoretical slower growth period I expected after the threshold. Ontogeny of

‘dwarfs’ could then be categorized as underdeveloped, not reaching the threshold between juvenile (development) and adult shape (growth) of sister taxa, or as developed, reaching or surpassing this threshold. Underdeveloped ontogenies would lead to the retention of juvenile characteristics, thus supporting the association between extreme body size reduction and paedomorphism. In addition, this would also give support for un-proportional miniaturization

129

(Gould, 1971) characterized by juvenile-like proportions. Developed ontogenies, in contrast, would result from the attainment of adult like proportions in the largest individuals and would give support for proportioned miniaturization (Gould, 1971).

3.3.7 Shifts in Magnitude of Shape Change

Finally, I determined if small body size in ‘dwarf’ lineages is associated with changes in the magnitude of shape change over ontogeny as compared to sister taxa, which may suggest differences in the rate of development in each of these lineages. Early ontogenetic stages should see rapid allometric changes of shape during early development. It is expected that individuals will then reach a threshold where energetic resource use will shift from supporting major developmental change to the acquisition of size and mass through near isometric growth of external features (Fig. 3.1). The ontogenies of taxa with extremely reduced body sizes should be dominated by this initial developmental period, and thus defined by high allometric slopes relative to large-bodied sister taxa. In addition, I can suggest a heterochronic process that may lead to divergence I observed among ontogenies. Heterochrony can be divided into two major processes: the deceleration of growth resulting in underdevelopment and the acceleration of growth resulting in overdevelopment (McKinney & McNamara, 1991). Deceleration of growth results in paedomorphosis, characterized by the reduction of trait size and/or retention of juvenile characteristics expected with miniaturization relative to the ancestral form. Extreme body sizes observed in extant taxa (either large or small) are assumed to reflect positive or negative shifts in ontogenetic trajectories away from the ancestral form. Three major heterochronic processes can lead to changes in trait size and shape thought to be associated with miniaturization. Perturbations to 1) rate of growth in an organism leads to neoteny (reduced

130 growth rate); changes to 2) duration of the growth period due to alterations to 2.a) onset timing of development can lead to post-displacement (later start) or 2.b) offset timing of development can lead to progenesis (truncation of growth period). Assuming that size is a relatively accurate proxy for development stage and that the relationship between size and development is similar among closely related species, shallower allometric slopes in ‘dwarf’ taxa relative to sister taxa would give support for neoteny (slower developmental rate), while no difference would suggest either progenesis (ontogenetic truncation) or post-displacement (delay in development).

However, the assumption that magnitude of shape change is highly predictive of developmental processes is largely unsubstantiated without age data, therefore further testing of these hypotheses with developmental data is required to confirm whether they reveal actual processes.

I tested for differences in ontogenetic slopes among closely related species (i.e., dwarfs and their sister taxa) by regressing shape data (PredVal) against log-transformed standard length using advanced.procD.lm in the package ‘geomorph’ (Adams & Otárola-Castillo, 2013). A reduced model was built using shape~size+species and used as a null model to test for differences in slopes among species by including an interaction term in the full model, shape~size*species (Fig. 3.2). Shape~1 or shape~size was not used as reduced models since I knew there would be an association between shape and size, and that species exhibit considerable differences in shape. Model fitting was employed to test whether shape data were best fit by a model with the inclusion (rejection of the null) or exclusion (no rejection of the null) of the interaction between size and species, supporting either differences in ontogenetic slopes among species or parallel (identical) slopes, respectively. Results indicate whether closely related Neotropical cichlids diverge through ontogeny or share similar ontogenetic pathways. This method also estimates the magnitude of shape change per unit size, to determine if rates of shape change were higher in ‘dwarf’ cichlids following my expectations.

131 3.4 Results

3.4.1 Body Size Reduction

Ancestral state estimation supports several independent instances of body size reduction across the phylogeny of Neotropical cichlids (Fig. 3.6). These reductions in body size do correspond to those lineages defined as ‘dwarfs’: Apistogramma, Biotoecus, Dicrossus, Crenicara,

Nannacara, Mikrogeophagus, Laetacara, and Taeniacara in addition to relatively small species within Crenicichla (including Teleocichla). The most extreme cases of reduction appear to be restricted to Apistogramma, Biotoecus, Mikrogeophagus, and Taeniacara. The MRCA of

Neotropical cichlids was estimated to be 143.43mm (CI 95%=70.45–292.02mm) using the

MCC tree. All estimates of MRCA size using the posterior distribution fell within the 95% confidence interval of the MCC MRCA estimate (results not shown) and did not influence the results, therefore all results discussed are based on the MCC tree. Percentage of significant divergence in body size from the MRCA state was highly variable among ‘dwarf’ taxa (18.39–

84.43%) within and among independent lineages. Since the confidence intervals are so broad for ancestral state estimations, particularly in deep nodes, I did not expect the percentages of significant body size divergence to be particularly high in the simulation. However, there was support for divergent body sizes that are consistent with previous analyses (Steele & López-

Fernández, 2014). Well supported (>50%) genera included Apistogramma, Taeniacara,

Mikrogeophagus, and Biotoecus. Apistogramma pucallpaensis (SL=27mm) was found to be significantly smaller than other Neotropical fishes in 84.43% of cases, Taeniacara candidi

(SL=33mm) in 70.20% of cases, Mikrogeophagus ramirezi (SL=34mm) in 68.20% of cases,

Apistogramma iniridae (SL=36mm) in 64.14% of cases, Biotoecus dicentrarchus (SL=38mm) in 60.41% of cases, and Apistogramma agassizi (SL= 43mm) in 51.71% of cases. Members of

132

Laetacara, Nannacara, Dicrossus, Crenicichla (C. O-wallacii and Teleocichla), as well as additional members of Apistogramma and Mikrogeophagus, were more weakly supported (18-

50%). No species of Mazarunia appear to be reduced in size compared to their sister genus

Guianacara and were poorly supported (4.6-13.4%) as significantly reduced using the bootstrapping method. The clade comparison found 100% support for significant divergence in one ‘dwarf’ lineage, with significant decrease in size at the common ancestor of Apistogramma and Taeniacara, as well as all nodes within this clade as represented in the phylogeny.

(next page) Figure 3.6: Traits supporting miniaturization in Neotropical cichlids. Ancestral character estimation (ACE) of most recent common ancestor (MRCA) of ‘dwarf’ taxa and their sister taxon, values by colour given in legend (Maximum Body Size). Traits presented at the tips from left to right: 1) Maximum attainable body size of adult fishes showing divergence between taxa from ACE; 2) Significant body size reduction found by comparing tips divergence from the root estimation of the MRCA of all Neotropical cichlids, percentages indicate the occurrence of reductions in lineages compared to expected divergence values; 3) Magnitude of shape change, estimated by Procrustes ANOVA, indicates the amount of shape change per unit of time, with higher magnitudes indicating higher rates of change; 4) Number of modifications to skull and sensory system compared to sister taxa – Weak: 1-4 total modifications, Moderate: 5-8, Strong: 9-12, Very Strong: 13-15; 5) Comparison of Procrustes Distances between adult dwarf shape and juveniles and adults of sister taxa – Weak: trend toward juvenile shape resemblance, Moderate: juvenile resemblance to some sister taxa, Strong: juvenile resemblance to most sister taxa, Very Strong: juvenile resemblance to all sister taxa; 6) Termination of ‘dwarf’ ontogeny relative to threshold of sister taxa – Weak: truncation after threshold in all sister taxa, Moderate: truncation at or below some thresholds, Strong: truncation below most thresholds, Very Strong: truncation below all thresholds. NA – Not Applicable; NAv - Not Available.

133

3.4.2 Paedomorphism in Lateral Line and Skeletal Elements

All cichlids in this study had 7 or fewer pitted tubed scales posterior to the last tubed scale of the upper lateral line, with the exception of Apistogramma pucallpaensis, which has 15 or more.

134

Lateral line pores on the preopercle were also conserved across species, with all species having

6 pores with the exception of Biotoecus dicentrarchus (reduced to 5) and Taeniacara candidi

(reduced to 5). Pores on the dentary were relatively consistent among large-bodied cichlids, with most species displaying 5 pores, while dwarf taxa exhibit only 4 pores. This character could not be coded in Crenicichla O-wallacii or in Mikrogeophagus ramirezi. Crenicara latruncularium and Dicrossus filamentosus both displayed a reduced number of pores (4). Acaronia nassa also displayed a reduced number of pores (4), however the number of pores in Laetacara dorsigera was further reduced (3) and was therefore coded as modified. The state was also not recovered in Cleithracara maronii, and it was assumed that the state was 5 pores. All large-bodied cichlids displayed either the D1 or D4 pattern, with the exception of Crenicara latruncularium (D2), of the trunk lateral line configuration (sensu Webb, 1990), representing fully developed lateral lines. Three patterns are considered to be paedomorphic (D8, D8.5, and D9), with D9 representing the most paedomorphic state. Dicrossus filamentosus, Ivanacara bimaculatum, and

Taeniacara all displayed the more developed form (D8), Biotoecus dicentrarchus displayed the most paedomorphic state (D9), and Apistogramma pucallpaensis showed an intermediate form between D8 and D9 (D8.5, sensu López-Fernández et al., 2005). The paired coronal frontal lateralis canals were reduced to a single canal in Biotoecus dicentrarchus and Taeniacara candidi. The opening of NLF5 is usually a single pore, but two were observed in Ivanacara bimaculata and was absent in Taeniacara candidi. Lateral line pores on the lachrymal typically numbered 4, with Biotoecus dicentrarchus being reduced to 3. The state was not observable for

Taeniacara candidi.

Taeniacara candidi exhibited reduced development of the anterior laminar expansion

(lobe) on the first epibranchial, distal to the pharyngobranchial articulation, compared to fully developed in Satanoperca. All other dwarf-sister pairs have the same condition, with the

135 exception of Ivanacara bimaculata and Nannacara anomala, which have no extension, relative to fully developed in Cleithracara maronii. Development of an extension on epibranchial 2 is typically the same across dwarf-sister taxon pairs. However, Ivanacara bimaculata and

Nannacara anomala have expansions present but both with a reduced cartilage cap, relative to a more developed one in Cleithracara maronii. Apistogramma pucallpaensis also has a developed expansion with a reduced cartilage cap, while Taeniacara candidi is developed with no cartilage cap relative to full development in Satanoperca. Acaronia nassa has a developed expansion with reduced cartilage cap, while both the expansion and cartilage cap are reduced in Laetacara dorsigera. The interarcual cartilage is absent in Dicrossus filamentosus relative to Crenicara latruncularium, and is absent in Cleithracara maronii, Ivanacara bimaculata, and Nannacara anomala. Interestingly, the interarcual cartilage is also absent in Acaronia nassa but not

Laetacara dorsigera. The first pharyngobranchial is cartilaginous in Biotoecus dicentrarchus relative to Acarichthys heckelii. Biotoecus dicentrarchus and Taeniacara candidi have lost infraorbitals 4,5, and 6 (sensu López-Fernández et al., 2005), relative to Acarichthys heckelii (4 and 5 fused) and Satanoperca (4,5,6 fused), respectively. In Ivanacara bimaculata and

Nannacara anomala infraorbitals 3 and 4 are present but with 5,6, and 7 missing, whereas in

Cleithracara maronii the latter three are present with 4 and 5 fused. Apistogramma pucallpaensis infraorbitals 4 and 5 are also fused, while 4,5,6 are fused in Satanoperca. The third infraorbital is absent in Biotoecus dicentrarchus, Apistogramma pucallpaensis, and

Taeniacara candidi, but it is present in all other taxa examined. With the exceptions of Cichla,

Retroculus, and Astronotus, the two lachrymal bones of Neotropical cichlids are interpreted as fused into a single laminar ossicle (López-Fernández et al., 2005). The shape of the lachrymal is elongate horizontally in Biotoecus dicentrarchus, Apistogramma pucallpaensis, and Taeniacara candidi rather than elongate vertically as in Acarichthys heckelii and Satanoperca, respectively.

The lachrymal of Mikrogeophagus ramirezi is also horizontal compared to approximately equal

136 as in ‘Geophagus’ brasiliensis. The presence of a pointed dorso-caudal laminar expansion of the lachrymal is absent in Biotoecus dicentrarchus relative to Acarichthys heckelii, Mikrogeophagus ramirezi relative to ‘Geophagus’ brasiliensis, and Taeniacara candidi relative to Satanoperca.

The expansion is also absent in Acaronia nassa, but not Laetacara dorsigera. Lastly, a reduced number of vertebrae (23-25) is found in Cleithracara maronii, Ivanacara bimaculata,

Nannacara anomala compared to other taxa examined, but not to each other, as well as in

Apistogramma pucallpaensis and Taeniacara candidi relative to Satanoperca. The extent of change in dwarf taxa relative to their sister taxa was quite variable, with Crenicichla O-wallacii showing the least amount of modification with no observable changes (0) from large-bodied

Crenicichla, and Taeniacara candidi showing the most modification with 12 anatomical changes as compared to Satanoperca.

In Mazarunia, I found that the number of dentary pores was reduced (4 pores) compared to the 5 pores exhibited in Guianacara. I found that a laminar expansion on the first epibranchial was absent in Guianacara, but present in a reduced form in all species of

Mazarunia, and differed in orientation among species. The interarticular cartilage was found in

Guianacara but was absent in Mazarunia. Infraorbitals 4 and 5 were fused in Guianacara while separate in Mazarunia, which also exhibits absence of infraorbital 6. The lachrymal of M. charadrica and M. pala were dorsoventrally compressed relative to Guianacara, and M. mazarunii exhibited a relatively elongate lachrymal compared to all species. The lachrymal and third infraorbital are not overlapping in M. mazarunii, but overlapping in all other species.

Mazarunia shows moderate modification, primarily in skeletal elements and few modifications to the sensory system (Table 3.1, Fig. 3.6).

137 3.4.3 Paedomorphism in External Morphology

In almost all cases, the average distance between ‘dwarf’ adults and juveniles of large-bodied sister taxa was significantly shorter than the average distance between ‘dwarf’ adults and adults of sister taxa. This suggests that the adult morphology of ‘dwarf’ cichlids more closely resembles the juvenile morphology rather than the adult morphology of their sister taxa. These results support the notion that the external morphology of ‘dwarf’ Neotropical cichlids is paedomorphic. Apistogramma steindachneri varied in the average resemblance to juveniles of each member of Satanoperca, matching the morphology of S. daemon more closely than S. jurupari and S. leucosticta (Table 3.2). There was support for Apistogramma steindachneri resembling juveniles of Satanoperca daemon only, and therefore only appears paedomorphic relative to Satanoperca daemon. Taeniacara candidi also varied in the average resemblance to juveniles of each member of Satanoperca, matching the morphology of S. leucosticta more closely than S. daemon and S. jurupari (Table 3.2). Taeniacara candidi however, was more similar to juveniles of its sister, Satanoperca, and is paedomorphic relative to all members the genus.

The average resemblance of Crenicichla O-wallacii and Teleocichla monogramma to juvenile shape of all large-bodied members of Crenicichla was quite similar across species comparisons. Crenicichla O-wallacii and Teleocichla monogramma more closely resembled juveniles of C. frenata, C. lugubris, and C. saxatilis than adults of these species (Table 3.3).

Crenicichla O-wallacii and Teleocichla monogramma resemblance to juveniles and adults of

Crenicichla geayi were not significantly different.

138

Table 3.1: Paedomorphism in lateral line and skeletal elements. Reductions and modifications to anatomical features were coded as 0 (no change), -1 (modification), or -2 (extensive modification or loss) depending on the degree of change relative to the character state of sister taxa. Total amount of change across the 18 traits were summed together to roughly indicate which taxa display the most amount of sensory and skeletal modification. 1 – Number of pitted upper lateral line (ULL) scales; 2 – Number of preopercular lateral line (LL) pores; 3 – Number of dentary LL pores; 4 – Pattern of LL on trunk; 5 – Configuration of the coronal frontal lateralis canals; 6 – Opening of NLF4; 7 – Opening of NLF5; 8 – Number of lachrymal canals; 9 – Anterior laminar expansion epibranchial 1; 10 – Antero-ventral laminar expansion epibranchial 2; 11 – Interarcual cartilage; 12 – First pharyngobranchial composition; 13 – Composition of infraorbitals 4,5,6; 14 – Fusion of lachrymal; 15 – Presence of Infraorbital 3; 16 – Shape of lachrymal; 17 –Dorso- caudal laminar expansion of lachrymal; 18 – Total number of vertebrae.

Lateral Line and Sensory System Skeletal Modification, Reduction, or Loss Total 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Biotoecus dicentrarchus 0 -1 -1 -1 -1 0 0 -1 0 0 0 -1 -1 0 -1 -1 -1 0 -10 Crenicichla O-wallacii 0 0 ? 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Dicrossus filamentosus 0 0 0 -1 0 0 0 0 0 0 -1 0 0 0 0 0 0 0 -2 Mikrogeophagus ramirezi 0 0 ? 0 0 0 0 0 0 0 0 0 0 0 0 -1 -1 0 -2 Ivanacara bimaculata 0 0 -1 -1 0 0 -1 0 -1 -1 0 0 -1 0 0 0 0 0 -6 Nannacara anomala 0 0 -1 0 0 0 0 0 -1 -1 0 0 -1 0 0 0 0 0 -4 Apistogramma pucallpaensis -1 0 -1 -1 0 0 0 0 0 -1 0 0 -1 0 -1 -1 0 -1 -8 Taeniacara candidi 0 -1 -1 -1 -1 0 -1 ? -1 -1 0 0 -1 0 -1 -1 -1 -1 -12 Laetacara dorsigera 0 0 -1 0 -1 0 0 0 0 -1 0 0 0 0 0 0 0 0 -3 Mazarunia charadrica 0 0 -1 0 0 0 0 0 -1 0 -1 0 -1 0 0 -1 0 0 -5 Mazarunia mazarunii 0 0 -1 0 0 0 0 0 -1 0 -1 0 -1 0 0 -1 -1 0 -6 Mazarunia pala 0 0 -1 0 0 0 0 0 -1 0 -1 0 -1 0 0 -1 0 0 -5

139

Table 3.2: Paedomorphism in external morphology of Apistogramma – Satanoperca clade. Average distances between dwarf adults

(n=5) and juveniles of sister taxa (n=5) (Distj) and average distance between dwarf adults and adults of sister taxa (n=5) (Dista). Degrees of freed (DF), t-values (T), and p-values from the t-test are given. Satanoperca daemon Satanoperca jurupari Satanoperca leucosticta Distj Dista DF T P-value Distj Dista DF T P-value Distj Dista DF T P-value Apistogramma 0.01 0.04 35.39 -13.72 <0.001 0.04 0.03 37.81 3.38 0.002 0.07 0.02 26.11 6.48 <0.001 steindachneri Taeniacara 0.10 0.13 47.70 -8.12 <0.001 0.05 0.12 43.31 -14.37 <0.001 0.04 0.12 35.57 -13.01 <0.001 candidi

Table 3.3: Paedomorphism in external morphology of Crenicichla clade. Average distances between dwarf adults (n=5) and juveniles of sister taxa (n=5) (Distj) and average distance between dwarf adults and adults of sister taxa (n=5) (Dista). Degrees of freed (DF), t- values (T), and p-values from the t-test are given. Crenicichla frenata Crenicichla geayi Crenicichla lugubris Distj Dista DF T P-value Distj Dista DF T P- Distj Dista DF T P- value value Crenicichla 0.01 0.02 43.43 -3.81 <0.001 0.01 0.02 46.59 -0.56 0.58 0.01 0.03 37.40 -5.80 <0.001 O-wallacii Teleocichla 0.01 0.02 42.79 -3.57 <0.001 0.01 0.01 46.82 -0.19 0.85 0.009 0.02 37.28 -5.91 <0.001 monograma Crenicichla saxatilis Distj Dista DF T P-value Crenicichla 0.007 0.02 35.49 -5.30 <0.001 O-wallacii Teleocichla 0.007 0.02 34.59 -4.12 <0.001 monograma

140

Table 3.4: Paedomorphism in external morphology of dwarf taxa. Average distances between dwarf adults (n=5) and juveniles of sister taxa (n=5) (Distj) and average distance between dwarf adults and adults of sister taxa (n=5) (Dista). Degrees of freed (DF), t- values (T), and p-values from the t-test are given.

Sister Taxon Distj Dista DF T P-value Biotoecus dicentrarchus Acarichthys heckelii 0.03 0.12 47.99 -25.51 <0.001 Dicrossus filamentosus Crenicara punctulatum 0.05 0.08 47.74 -9.29 <0.001 Nannacara anomala Cleithracara maronii 0.05 0.07 34.87 -4.02 <0.001 Ivanacara bimaculata Cleithracara maronii 0.03 0.05 35.19 -3.99 0.003 Mikrogeophagus ramirezi ‘Geophagus’ brasiliensis 0.03 0.04 28.09 -1.90 0.07 Laetacara thayeri Acaronia nassa 0.07 0.01 37.20 25.34 <0.001

Table 3.5: Paedomorphism in external morphology of Mazarunia. Average distances between Mazarunia adults (n=5) and juveniles of sister taxa (n=5) (Distj) and average distance between Mazarunia adults and adults of sister taxa (n=5) (Dista). Degrees of freed (DF), t-values (T), and p-values from the t-test are given.

Guianacara dacrya Guianacara stergiosi Distj Dista DF T P-value Distj Dista DF T P-value Mazarunia charadrica 0.06 0.05 35.37 1.16 0.252 0.02 0.04 42.55 -4.62 <0.001 Mazarunia mazarunii 0.06 0.05 33.28 4.20 <0.001 0.02 0.03 36.60 -2.08 0.045 Mazarunia pala 0.08 0.03 32.13 13.35 <0.001 0.04 0.02 33.41 5.70 <0.001

141

Biotoecus dicentrarchus, Dicrossus filamentosus, Nannacara anomala, and Ivanacara bimaculata were found to resemble the juvenile morphology of their sister taxa over adult morphology (Table 3.4). Mikrogeophagus ramirezi was not found to resemble the juvenile morphology of ‘Geophagus’ brasiliensis over adults despite smaller distances to juveniles

(Table 3.4). Finally, Laetacara thayeri was found to resemble the adult morphology of Acaronia nassa over juveniles (Table 3.4). Mazarunia species generally resembled adult morphology of

Guianacara over juvenile morphology, with the exception of Mazarunia charadrica and M. mazarunii, which were significantly more similar to juveniles of Guianacara stergiosi than to adults (Table 3.5).

3.4.4 Significant Reductions in Duration of Development and Growth

Threshold values for the size at which adult-like shape is attained was highly variable among taxa, ranging from 19.45mm SL in Dicrossus filamentosus to 145.54mm in Crenicichla lugubris, with a mean threshold value of 45.77mm (median=39.51mm) (Table 3.6, Fig. 3.7).

The percentage of development attained at the threshold (ThL/SL) was also highly variable, ranging from 16.78% in ‘Geophagus’ brasiliensis to 70.27% in Crenicichla O-wallacii, with mean value of 45.23% (median=48.50%) (Table 3.6). The ontogeny of Apistogramma steindachneri terminates before the threshold of Satanoperca daemon, but after that for S. jurupari and S. leucosticta (Fig. 3.7a). The ontogeny of Taeniacara candidi terminates before the threshold of Satanoperca daemon and S. jurupari, but at the threshold for S. leucosticta (Fig.

3.7a). The ontogeny of Crenicichla O-wallacii terminates before the threshold for all large- bodied Crenicichla, while Teleocichla monogramma only terminates before the threshold of

Crenicichla lugubris (Fig. 3.7c. The ontogeny of Dicrossus filamentosus terminates at the

142 threshold of Crenicara punctulatum (Fig. 3.7f), Biotoecus dicentrarchus slightly after that of

Acarichthys heckelii (Fig. 3.7b), Nannacara anomala and Ivanacara bimaculata both after that of Cleithracara maronii (Fig. 3.7d), Mikrogeophagus ramirezi considerably before that of

‘Geophagus’ brasiliensis (Fig. 3.7g), and Laetacara thayeri after that of Acaronia nassa (Fig.

3.7e). Ontogenies of all species of Mazarunia terminated considerably after the threshold of both species of Guianacara (Fig. 3.7h).

(next page) Figure 3.7: Ontogenetic trajectories of Neotropical ‘dwarf’ cichlids (red, yellow, orange) and their sister taxa (black, grey, dark grey, blue). Circles are coloured by taxon, and each represent the predicted shape of the specimens measured in this study, based on a regression of shape on size (standard length mm). Panels a) Apistogramma steindachneri, Taeniacara candidi, and their sister taxon Satanoperca daemon, S. jurupari, S. leucosticta; b) Biotoecus dicentrarchus and sister taxon Acarichthys heckelii; c) Crenicichla O-wallacii, Teleocichla monogramma, and their sister taxa Crenicichla geayi, C. frenata, C. saxatilis, C. lugubris; d) Ivanacara bimaculata, Nannacara anomala, and their sister taxon Cleithracara maronii; e) Laetacara thayeri and sister taxon Acaronia nassa; f) Dicrossus filamentosus and sister taxon Crenicara punctulatum; g) Mikrogeophagus ramirezi and sister taxon ‘Geophagus’ brasiliensis; h) Mazarunia charadrica, M. mazarunii, M. pala, and their sister taxa Guianacara dacrya and G. stergiosi.

143

144

Table 3.6: Ontogenetic milestones in Neotropical cichlids. Maximum attainable adult body size is given based on published literature and accessible in FishBase (Froese & Pauly, 2016); asterisks indicate values that were measured from specimens in this study larger than published sizes. Thresholds, with standard error, indicating the inflexion point between development periods and growth periods were estimated using pieces-wise regression. Threshold percentage (Threshold/Maximum Size) gives an indication of the contribution of each period to the ontogenies of taxa in this study. P-value indicates significance difference between development and growth periods slopes, demonstrating the appropriateness of using piece-wise regression.

Species Proposed Maximum Threshold Threshold St Error P- Form Size Percentage Threshold Value Apistogramma steindachneri Dwarf 65.00 24.22 37.26 0.62 <0.001 Taeniacara candidi Dwarf 36.70* 20.16 54.93 0.94 <0.01 Satanoperca daemon Normal 210.00* 77.43 36.87 2.81 <0.001 Satanoperca jurupari Normal 201.00 45.66 22.72 1.37 <0.001 Satanoperca leucosticta Normal 160.00 35.39 22.12 2.29 <0.001 Crenicichla O-wallacii Dwarf 55.00 38.65 70.27 0.51 <0.001 Teleocichla monogramma Dwarf 98.00 44.26 45.16 0.62 <0.001 Crenicichla frenata Normal 144.83* 78.75 54.37 1.58 <0.001 Crenicichla geayi Normal 136.00 62.30 45.81 2.11 <0.001 Crenicichla saxatilis Normal 200.00 71.13 35.57 2.21 <0.001 Crenicichla lugubris Normal 260.00 145.54 55.98 3.60 <0.001 Dicrossus filamentosus Dwarf 38.00 19.45 51.18 0.35 <0.001 Crenicara punctulatum Normal 100.00 38.33 38.33 0.55 <0.001 Biotoecus dicentrarchus Dwarf 35.22* 24.34 69.11 0.32 <0.001 Acarichthys heckelii Normal 134.00 32.34 24.13 1.07 <0.001 Nannacara anomala Dwarf 56.00 28.79 51.41 0.83 <0.001 Ivanacara bimaculata Dwarf 45.00 25.95 57.67 0.61 <0.001 Cleithracara maronii Normal 71.00 40.37 56.86 0.88 <0.001 Mikrogeophagus ramirezi Dwarf 35.62* 20.11 56.46 0.30 <0.001 ‘Geophagus’ brasiliensis Normal 280.00 46.99 16.78 1.66 <0.001 Laetacara thayeri Dwarf 69.00* 44.65 64.71 0.61 <0.001 Acaronia nassa Normal 154.00 42.12 27.35 1.17 <0.001 Mazarunia charadrica Normal 84.00 38.33 45.51 0.74 <0.001 Mazarunia mazarunii Normal 77.00* 47.10 61.17 0.95 <0.001 Mazarunia pala Normal 74.00 44.41 59.66 1.04 <0.001 Guianacara dacrya Normal 120.00 40.82 34.02 1.28 <0.001 Guianacara stergiosi Normal 80.00 36.01 30.00 0.91 <0.001

145 3.4.5 Shifts in Ontogenetic Rate of Development

In all cases, the full regression model was statistically supported (p = 0.002 in all cases unless otherwise noted) over the null model of no interaction between size and species, and therefore supported a difference in the ontogenetic slope of at least one member in each clade comparison

(Table 3.7). Apistogramma steindachneri has significantly higher magnitude of shape change as compared to members of Satanoperca, while Taeniacara candidi does not show any significant differences in magnitude from Satanoperca members, but shows a negative allometric slope compared to Satanoperca. Apistogramma steindachneri was not significantly different in magnitude from Taeniacara candidi. The ontogenetic slopes of Crenicichla O-wallacii and

Teleocichla monogramma were found to be different than all large-bodied members of

Crenicichla, with the exception of C. O-wallacii to C. lugubris, but not each other. The magnitude of shape change was significantly higher in Dicrossus filamentosus compared to

Crenicara punctulatum, and in Biotoecus dicentrarchus compared to Acarichthys heckelii.

Ontogenetic slopes were significantly different between Mikrogeophagus ramirezi and

‘Geophagus’ brasiliensis, and M. ramirezi also had a significantly higher magnitude of shape change. There was no significant difference in the magnitude of shape change between any of the species combinations in the comparison of Nannacara anomala, Ivanacara bimaculata, and

Cleithracara maronii. Ontogenetic slopes were not significantly different between Laetacara thayeri and Acaronia nassa, and the magnitude of shape change was not found to be significantly higher in L. thayeri. All species of Mazarunia were found to have lower magnitude of shape change than both species of Guianacara.

146

Table 3.7: Shifts in ontogenetic rate of development. Regression coefficients and magnitude of shape change for ‘dwarf’ Neotropical cichlids and their sister taxa estimated using Procrustes ANOVA and regression. Intercept and slope estimated from a regression of shape on size. The magnitude indicates the degree of shape change per unit size, suggesting differences in ontogenetic shape change among species.

Species Proposed Intercept Slope Magnitude Form Apistogramma steindachneri Dwarf -0.13 0.07 0.08 Taeniacara candidi Dwarf 0.01 -0.07 0.05 Satanoperca daemon Normal -0.06 0.02 0.03 Satanoperca jurupari Normal -0.09 0.04 0.05 Satanoperca leucosticta Normal -0.12 0.06 0.06 Crenicichla O-wallacii Dwarf -0.27 0.06 0.09 Teleocichla monogramma Dwarf -0.22 0.02 0.09 Crenicichla frenata Normal -0.15 0.01 0.04 Crenicichla geayi Normal -0.20 0.04 0.04 Crenicichla saxatilis Normal -0.17 0.03 0.03 Crenicichla lugubris Normal -0.30 0.08 0.06 Dicrossus filamentosus Dwarf -0.03 -0.05 0.07 Crenicara punctulatum Normal -0.18 0.09 0.06 Biotoecus dicentrarchus Dwarf -0.13 0.001 0.08 Acarichthys heckelii Normal -0.19 0.10 0.07 Nannacara anomala Dwarf -0.12 0.07 0.06 Ivanacara bimaculata Dwarf -0.13 0.09 0.09 Cleithracara maronii Normal -0.03 0.06 0.07 Mikrogeophagus ramirezi Dwarf -0.10 0.06 0.09 ‘Geophagus’ brasiliensis Normal -0.13 0.08 0.05 Laetacara thayeri Dwarf -0.07 0.06 0.07 Acaronia nassa Normal -0.06 0.06 0.06 Mazarunia charadrica Normal -0.10 0.05 0.04 Mazarunia mazarunii Normal -0.03 0.02 0.04 Mazarunia pala Normal 0.03 0.03 0.05 Guianacara dacrya Normal -0.17 0.12 0.07 Guianacara stergiosi Normal -0.18 0.12 0.07

147 3.5 Discussion

3.5.1 Testing for Miniaturization: A Quantitative Approach

Recent approaches to studying miniaturization or attempts at quantifying body size regions have been made in fishes (Albert & Johnson, 2012) and other vertebrates (Angielczyk & Feldman,

2013; Galatius et al., 2011; Masters et al., 2014), yet the ability to empirically detect miniaturization based on morphological paedomorphism has not been widely tested. I therefore propose a methodology expanding upon previous work (Angielczyk & Feldman, 2013) that is widely applicable across taxa to study miniaturization using context dependent body size reduction, reductive evolution, and ontogenetic changes relative to sister taxa of small-bodied lineages. Descriptive patterns examining the anatomical changes associated with body size reduction, in particular those changes that occur through development and growth, are labour and knowledge intensive. Rarely can these studies be completed on large datasets with macroevolutionary questions in mind. Even more constraining is the lack of data available for such studies, with many species being rare or difficult to collect in the wild at all life stages and near impossible to breed in captivity to supplement the ontogeny. Therefore, in relying on previously collected museum specimens, we are limited largely to non-invasive investigations of morphology in relatively small sample sizes. To test the potential influence of sample size and distribution across the ontogeny on my results, I examined various subsamples of Acarichthys heckelii and estimations of regression coefficients as compared to the ‘true’ population. I found, in general, that even when the sample size is significantly reduced, estimations of coefficients are similar to the ‘true’ coefficients when individuals are sampled evenly across the ontogeny.

In addition, I found that when the earliest and latest ontogenetic stages were included in the sample, the estimations were closer to the expected values. This suggests that when individuals

148 are carefully selected to maximally represent ontogenetic change, most variation in the ‘true’ population can be captured with a relatively small sample size.

One major caveat of this methodology, as presented in my case study of Neotropical cichlids, is the lack of age or developmental data that can be used to standardize growth curves, or to link shape changes directly to age. Age data can be difficult to collect in tropical fishes

(Jepsen et al., 1999) without observing growth directly, but is not a ubiquitous problem among taxa. In some fishes, morphological differentiation and ontogenetic changes are associated with temperature dependent size change rather than age (Fuiman et al., 1998; Mendiola et al., 2007).

As such, my use of size still gives important information about the differences in development and growth between closely related taxa. Allometry, the covariation between size and body shape or more specifically of growth patterns among morphological traits, can be used as a proxy for age data to determine the difference in growth between normal, miniatures and dwarf forms (McKinney & McNamara, 1991; Frédérich & Sheets, 2010). I have largely relied upon morphometrics of external anatomy to infer paedomorphic characteristics, and have supported these potential patterns by testing the evolutionary direction of body size change, divergence from sister taxa, as well as provided some evidence of skeletal and sensory system modifications that appear to be associated with body size reduction. While some aspects of ontogenetic pattern, and discussions of process are largely inconclusive, this method uses multiple lines of evidence across broad groups of taxa to test miniaturization hypotheses associated with developmental literature (McKinney & McNamara, 1991; Hanken & Wake,

1993; Weitzman & Vari, 1988). By assessing ontogenetic allometry and paedomorphic characterization of ontogenies, specifically with dense sampling of early ontogenetic stages, macroevolutionary studies addressing shape and size evolution across large clades can be completed without age or developmental data.

149

3.5.2 Potential Processes of Miniaturization

Consistent with the theory of miniaturization described for vertebrates (Gould, 1971; Hanken &

Wake, 1993) I saw that most ‘dwarf’ cichlids experience body size reduction through significantly divergent ontogenetic pathways compared to those from large-bodied sister taxa.

These changes in small bodied taxa are coupled with paedomorphic external features, truncated ontogenies, and some evidence of modification to the skeletal and sensory system (Fig. 3.6). I also found that the allometric trajectories of these ‘dwarfs’ often diverged from their sister taxa, indicating varying associations between shape and size through ontogeny across related taxa.

With strong resemblance to juvenile shape in most species, I found that my results support

Neotropical ‘dwarf’ cichlids as being un-proportional dwarfs (as opposed to proportional ones, see Appendix 3.1), characterized by paedomorphic traits (Gould, 1971). The distinction between the two is necessary as there are obvious implications for morphology as well as likely implications for functional morphology, ecology, and life-history as a result of these morphological differences.

We found that proposed dwarf species of Neotropical cichlids typically experience some inflexion in allometric curves, suggesting that distinct development and growth periods are recognizable to at least some degree. However, compared to their sister taxa, ‘dwarf’ cichlids reach this threshold at smaller body size that likely represents a later ontogenetic stage than seen in sister taxa. While it has been shown that larger species may live longer (Paine, 1990), there is no evidence that these cichlid species are as long-lived in the wild as large-bodied cichlids, but if so may experience faster rates of development and earlier shape and sexual maturation, devoting little resources to growth and instead invest heavily in reproduction. This notion is

150 supported by the higher magnitude of shape change per unit size in most ‘dwarf’ species examined, which can roughly indicate increased rates of development in early ontogenetic stages. Earlier thresholds combined with steeper ontogenetic slopes suggests that these lineages may be developing faster relative to their sister taxa in a shorter period of time, however without robust life history and developmental data I cannot make strong claims about the underlying process.

Development can be altered in several ways, resulting in deviations in growth rates and the morphological outcome of the adult shape. Evolution towards small body size can also result in significant changes in the proportion of various body elements, and in the extreme lead to paedomorphism and the resemblance of juvenile shape. I expected the degree of paedomorphism, or the number of paedomorphic characters, to be strongly associated with body size reduction. As lineages move below some critical size (Hanken & Wake, 1993) I would expect increased accumulation of paedomorphic traits and increasing paedomorphic states as taxa become more and more reduced in size. Thus, miniaturization represents a spectrum of modification with taxa just below the threshold being largely proportional and exhibiting few paedomorphic characters (e.g. Crenicichla O-wallacii), down to the taxa at lower limits of vertebrate body size evolution, which resemble the most juvenile of forms and have a wide array of paedomorphic traits (e.g. Taeniacara candidi). I also found that the small-bodied genus

Mazarunia showed little evidence for paedomorphism, and in particular was not significantly reduced in body size. While Neotropical cichlids do not reach the limits of vertebrate body size evolution, and therefore do not exhibit incredibly reduced forms as seen in characiforms and cypriniforms (Britz & Conway, 2009; Weitzman & Vari, 1988), there is a considerable amount of body modification suggesting not only proportional miniaturization, but more extreme forms of modification of infantilized, un-proportional miniatures. All species in the analysis below

151

65mm had stronger resemblance to juveniles than to adults and strong paedomorphism, suggesting that the range of size for proportional miniatures is very narrow but likely a soft threshold of miniaturization and between proportional and un-proportional miniatures. These transitions may occur at differing points within specific lineages depending on the developmental and ecological context under which miniature lineages are evolving (Hanken &

Wake, 1993).

Increased morphological variation within and among species may be experienced due to inconsistent morphological rearrangement or loss associated with body size reduction (Hanken

& Wake, 1993). This morphological novelty and intense functional constraint at small body size can lead to convergent or parallel evolution across taxa (Hanken & Wake, 1993). Interestingly, most proposed ‘dwarf’ taxa were found to be monophyletic based on morphological data

(López-Fernández et al., 2005). However, the morphological traits grouping these taxa have been shown to be clearly convergent (López-Fernández et al., 2005) and appear to be traits associated with simplification of the sensory systems, reduction in skeletal elements, and the appearance of more juvenile-like states based on the developmental trajectory of cichlid traits

(Webb, 1990). These traits were proposed to be associated with body size reduction in these taxa (López-Fernández et al., 2005). My examination of skeletal and sensory systems found that

‘dwarf’ taxa exhibited the same modified traits, however these traits were not modified in the same way across ‘dwarf’ taxa. While I expected some variation across taxa, I also expected similar features in ‘dwarf’ taxa due to convergence resulting from similar selective pressures associated with extreme reduction. However, reduction of body size in ‘dwarf’ taxa may not result from convergent changes to the developmental process, as developmental changes in even closely related species can follow markedly different ontogenetic pathways (McKinney &

McNamara, 1991). Perhaps these similar trait changes were not observed due to the large

152 amount of size variation observed in ‘dwarf’ taxa, and higher degrees of convergent evolution would only be experienced at the smallest body sizes. This variation suggests that while miniatures may share overall similarities in morphological changes due to small body size, many solutions to particular constraints (e.g. limited cranial and postcranial volumes, higher metabolism, reduced locomotion, earlier maturation) may evolve among lineages to compensate for small body size. These multiple solutions may result in divergent or convergent function.

Many-to-one mapping of morphological changes (i.e., that multiple forms can allow for the same functional output) could result in convergence in functional traits despite these divergent morphologies (Wainwright et al., 2005). While ontogenetic data is unavailable for some of these traits, I do see that external morphology strongly resembles juvenile-like shape in most ‘dwarf’ taxa, and therefore these modifications to the skull and sensory system that have not been studied in the context of ontogeny may indeed show paedomorphism.

3.5.3 Adaptive Nature of Extreme Body Size Reduction in Neotropical Cichlids

Our results support the upper limit of miniaturization in Neotropical cichlids as being below

100mm. The largest ‘dwarf’ cichlid included in this study was Teleocichla monogramma

(98mm), which did not show strong body size reduction compared to all Neotropical cichlids, nor strong evidence of ontogenetic truncation relative to large-bodied Crenicichla. However, T. monogramma did show evidence of paedomorphic external morphology, resembling juveniles of its sister taxa over adults. This resemblance to juveniles may be driven by the relative elongation of T. monogramma, an adaptation for its rheophilic ecology, as compared to the deeper bodied adults of large-bodied Crenicichla included in my study. Unfortunately this species or other congeners were not available to investigate modification to the sensory system

153 or skeleton, although it is known that the lateral line is not paedomorphic (Webb, 1990) and may be an adaptation for increasing sensitivity in turbulent habitats. As body size is reduced, there is a strong association between body size reduction and the degree of modification to morphology and ontogeny. As taxa become smaller, there is a strong simplification and paedomorphism in the skeleton and sensory system, considerable truncation of the ontogeny and strong resemblance to juveniles. This most notably occurs in Taeniacara candidi (36.7mm) and

Biotoecus dicentrarchus (35.2mm), with strong support for miniaturization in body size reduction, paedomorphic skeletal and lateral line characters, paedomorphic external morphology, and ontogenetic slope, appearing to represent the most paedomorphic form.

As previously suggested (Steele & López-Fernández, 2014), it appears that miniaturization, assuming that taxa here are representative of broader patterns at the genus level, is much more pronounced in the Neotropical cichlid tribe Geophagini. This tribe is rich in species, morphological, and ecological diversity (Arbour & López-Fernández, 2013; Astudillo-

Clavijo et al., 2015; López-Fernández et al., 2013). In terms of body size, the species rich subclade including Apistogramma and Crenicichla contains most of the smallest known

Neotropical cichlids as well as large-bodied piscivores, some of which can reach body sizes of around 315mm. Apistogramma was found to dominate the small body size region both in terms of degree of reduction and frequency in Geophagini and across Neotropical cichlids. In contrast, the less species-rich clade containing Geophagus and many small-bodied but species-poor genera has a narrow size distribution in comparison. Several genera within this clade have expanded into the small-bodied morphospace dominated by Apistogramma, however none have been as successful in terms of species richness and do not show body size reduction to the degree seen in Apistogramma. In addition, the majority of cichlids within this clade are mid- sized fishes, with very few species reaching the upper extreme of the species-rich clade. How is

154 it that Apistogramma has diversified to such an extent compared to other ‘dwarf’ taxa, without much apparent phenotypic divergence?

Body size has long been considered one of the fundamental axes upon which species can diverge to alter their niche, acquiring access to differing prey sizes and habitats at large and small body size (Wilson, 1975). It has also been shown in cichlid fishes, that there is divergence in niches associated with ontogenetic body size, with several species showing ontogenetic diet shifts through growth (Meyer, 1990; Ponton & Mérigoux, 2000; Santos-Santos et al., 2015;

Winemiller, 1989). The maintenance of juvenile shapes and size in Neotropical cichlids may be an adaptive strategy to avoid intense competition for resources in highly diverse riverine communities. It has been found that dwarf cichlids, unlike large-bodied cichlids, do not undergo ontogenetic diet shifts as they increase in size (Ponton & Mérigoux, 2000). Instead, adults continue to feed on small invertebrates in the benthos or, more rarely, in the water column. In mammals and birds, it has been suggested that small body size is so prevalent due to the ubiquity of food resources (e.g. seeds, foliage, insects) for small-bodied species (Bokma, 2004;

Monroe & Bokma, 2009). I expect that in tropical rivers, small invertebrate prey may be highly abundant in these high productivity systems (Davies et al., 2008), and therefore there is little competition for this resource driving evolution of higher trophic ecology. Additionally, small body size allows increased partitioning of habitat space (Hutchinson & MacArthur, 1959), allowing species packing in small habitats within species rich systems Access to small, complex microhabitats unoccupied by large predators can increase access to food resources, reduce predation risk, and provide adequate protection of offspring and adults during periods of parental care. However, the retention of juvenile like morphology with extreme size is also seen, imposing strong constraints on biological process that counter the advantages of small size.

155

A strong association between body size and reproductive potential is seen in fishes

(Ross, 2013). In cichlids, we often see considerable parental care at all body sizes, which aids in offspring survivorship (Barlow, 1991). At small body sizes, cichlids are physically excluded from using the successful mouth-brooding mode of parental care. We also see in cichlids at small body sizes, egg production is often limited in terms of egg number and egg size, severely limiting the fitness of small-bodied fishes as compared to large-bodied fishes (Kolm et al.,

2006a). Instead, these species may also increase the frequency of spawns compared to large- bodied cichlids to increase reproductive potential via egg number. ‘Dwarf’ cichlids often exhibit harem-holding behaviour in which multiple females synchronously and brood offspring in a territory defended by a dominant male. This likely increases the survivorship of offspring within and among these broods. Early maturation at small body size may also help compensate for small body size, allowing a relatively longer period of time to produce offspring and redirecting energy from further development and growth towards maximizing reproductive potential. While reproductive output often appears to be quite reduced in ‘dwarf’ Neotropical cichlids (Kolm et al., 2006b), it may be that these compensatory physiological and behaviour changes associated with body size reduction may reduce the impacts of that same reduction on life-time reproductive potential. Further reproductive studies are needed to address these questions.

Finally, some evidence from Asian and North American miniature fishes has shown that development may be strongly affected by abiotic conditions and water chemistry. Extreme miniatures in characiforms, cypriniforms, perciforms, among others, have been found in warm water, oxygen poor, acidic peat bogs (Bennett & Conway, 2010; Liu et al., 2012) Therefore, this life strategy may also be tied to invading novel or underused habitats that simply cannot be inhabited by large-bodied species due to the physiological constraints. In cichlids, ‘dwarf’

156 species are largely associated with small streams, complex benthic habitats (i.e., leaf litter, tightly knit woody debris), and slow-moving waters. While some variation in chemistry may occur in these macrohabitats, they are likely not significant enough in driving current distribution of ‘dwarf’ cichlids but may have been important during the diversification of some of these lineages. Only one lineage, Teleocichla, occurs in a contrasting habitat of large riffle, high velocity, large substrate patches in which small body size and elongation would assist in reducing drag during swimming, and allow individuals to rest in the boundary layers or to escape high velocity regions by moving between large particles (Lujan & Conway, 2015).

Similarly, small body size in fishes in general, may also allow them to live continuously within strongly fluctuating channel systems as a result of seasonality, where larger species would be forced to disperse to larger channels during periods of low flow or avoid these shallow waters altogether.

Though small body size has been linked to a number of disadvantages (e.g. reduced dispersal, decrease in offspring size and/or number), reduction of body size can also lead to faster development times or earlier maturation (Roff, 2002), less competition for resources

(Wilson, 1975), and access to novel habitats (Hutchinson & MacArthur, 1959). In addition, reduction of body size within a lineage may result in a reduction in the morphological opportunities, and subsequently the ecological opportunities available. The intensity of this morphological and ecological constraint is likely asymmetrical between proportional dwarfs, which only experience reduction in size, as compared to un-proportional dwarf species that experience reduction in size and considerable changes in relative proportions in the body plan.

In addition, the maximum attainable body size for most ‘dwarf’ taxa in my study is smaller or at the estimated threshold at which sister taxa transition from juvenile to adult morphology. It is pertinent to remember that this conclusion assumes the threshold of morphological maturity is

157 accurately estimated, and that the ontogeny of the large-bodied sister taxon is representative of the ancestral ontogeny. This suggests that adults of ‘dwarf’ taxa are often a size typically attributed to a juvenile state in non-dwarf Neotropical cichlids. Therefore, these species are occupying extremely small and juvenile-like phenotypes, the variation which is seen among species being driven by the dynamic trade-offs in the advantages and disadvantages of small body size and juvenile shape.

3.5.4 Conclusion

Our results suggest that quantitatively defining the boundary between miniature and non- miniature forms, as well as between proportional and un-proportional dwarfs may be challenging without developmental data. However, with a small dataset of traits, I was able to suggest ontogenetic divergence occurring in Neotropical cichlids that can be corroborated with additional taxa and developmental data. This chapter outlines the framework to assess whether reductions in body size are biologically and developmentally meaningful across taxa and provides a case study in Neotropical cichlids comparing proposed ‘dwarf’ taxa to their large- bodied sister taxa. I found significant body size reduction from the ancestral form along with considerable evidence for paedomorphic traits proposed to be exhibited by miniaturized species.

Here, I only described and compared ontogenies of taxa currently known to be sister taxa through phylogenetic inference. It is imperative that these data be assessed in the broader context of Neotropical cichlids and to incorporate phylogenetic information to determine how these traits evolved, and whether my results here are consistent with phylogenetically informed analyses.

158 3.6 Appendices

Appendix 3.1: Glossary of ontogenetic terms.

159 Chapter 4 Morphological Disparity Across the Ontogeny and Phylogeny of Neotropical Cichlid Fishes (Cichliformes: Cichlidae: Cichlinae) and its Impact on Evolution of Body Size

Sarah Elizabeth Steele1 and Hernán López-Fernández1,2

1Department of Ecology and Evolutionary Biology, University of Toronto, 25 Wilcocks St.,

Toronto, Ontario M5S 3B2, Canada

2Department of Natural History, Royal Ontario Museum, 100 Queen’s Park, Toronto, Ontario

M5S 2C6, Canada

160 Chapter Four

4.1 Abstract

It is largely unknown how body size over ontogeny interacts with adult body size to influence diversification across species, morphology, and ecology. While the macroevolution of morphological and ecological diversity across major radiations of vertebrates has been examined, rarely do these studies consider growth, development, and ontogenetic shifts of traits that lead to divergence among adults. If body size has been shown to be important for so many biological processes, then body size of adults, as well as early ontogenetic life stages, should also play a significant role in shaping the accumulation of diversity through time. I characterized the ontogenetic variation across Neotropical cichlids fishes by quantifying the relationship between external body shape and size, tested whether this relationship significantly differed across species, and determined how much variation was captured by a common allometric trend versus evolutionary divergence among species. I then determined whether variation in ontogenetic rates and orientation resulted in a reduction of morphological disparity in adult cichlids as compared to juveniles. Finally, I tested whether the evolution of body size and ontogenetic parameters supported previous hypotheses of trait evolution, the differentiation between miniatures and all other cichlids or three body size optima across the phylogeny of

Neotropical cichlids. I found that, despite considerable variation in magnitude and orientation of ontogenies, morphological disparity of juveniles was not significantly reduced as compared to adults, and that a common allometric trend across species explained nearly half of the variation in the data. I found early evolutionary divergence of ontogenetic parameters and body size adaptive peaks coinciding with the emergence of major lineages of fishes typical of Neotropical

161 cichlids trait evolution, but additionally found more recent regime shifts among and within ecological specialized genera of cichlids.

4.2 Introduction

Body size is closely tied to trophic level (e.g. Romanuk et al., 2011), range size (e.g. Lindstedt et al., 1986), metabolism (e.g. West et al., 1997), niche partitioning (e.g. Brown & Wilson,

1956), and life history (e.g. Blackburn & Gaston, 1994) in vertebrates. The role of body size in the diversification of species, morphology, and ecology has been explored across vertebrates

(Bokma, 2004; Monroe & Bokma, 2009; Rabosky et al., 2013). Body size has largely been examined in the context of adult size (i.e., maximum attainable body size). Yet, there is considerable evidence that both morphology and ecology can change drastically through the life of organisms (Gould, 1977). It is largely unknown how body size over ontogeny interacts with adult body size to influence diversification across species, morphology, and ecology. While the macroevolution of morphological and ecological diversity across major radiations of vertebrates has been examined, rarely do these studies consider growth, development, and ontogenetic shifts of traits that lead to divergence among adults. Often, for simplicity, ontogenetic or size effects of individuals on species trait values are removed from analyses. However, the variation seen in adults is not only a product of selection pressures on the adult form, but indirect selection on ontogeny and the timing of developmental events (McKinney & McNamara, 1991). If body size has been shown to be important for so many biological processes, then body size of adults, as well as early ontogenetic life stages, should play a significant role in shaping the accumulation of diversity through time. In addition, evolution of extreme body size in vertebrates should also have substantial effects on diversification processes (Clauset & Erwin, 2008; Kozłowski &

Gawelczyk, 2002; Stanley, 1973).

162

As McKinney and McNamara (1991) concisely stated, nothing in evolution occurs without changes to the developmental process. Changes in developmental timing, rates, duration, and patterning in ontogeny leads to the natural variation in populations that natural selection acts on, and should ultimately result in changes to size and shape. Evolution of body size has been linked to changes in developmental rate and duration in vertebrates (Angielczyk &

Feldman, 2013; Galatius et al., 2011; Masters et al., 2014). Developmental changes may drive many of the patterns seen in body size evolution and associated morphology across taxa on broad evolutionary time-scales. Phenotypic or ecological diversification has been widely studied across vertebrates, but our understanding of when, and how this divergence occurs through the ontogeny of organisms is largely unknown. Very few studies have examined variation in disparity among ontogenetic stages across large groups of organisms (but see Frédérich &

Vandewalle, 2011; Urošević et al., 2013; Zelditch et al., 2003, 2016) or how changes in body size affect disparity. While the relationship between shape and size (i.e., allometry) has been studied widely among taxa, it has been underrepresented in macroevolutionary studies incorporating phylogenetic comparative tools. Allometry is comprised of three major sources of variation: 1) Static allometry, allometry of individuals within a single life stage (e.g. adults); 2)

Ontogenetic allometry, allometry of individuals across multiple life stages; and 3) Evolutionary allometry, allometry across multiple species (Cheverud, 1982). Rarely are these multiple sources of allometric variation assessed in a single analysis, as data collection and analysis are difficult and time-consuming.

Ontogenetic studies have been important for understanding integration and modularity of morphology within species (Hallgrímsson et al., 2009) and more rarely, the divergence of species morphology (e.g. Adams & Nistri, 2010; Galatius et al., 2011; Masters et al., 2014;

Santos-Santos et al., 2015). Debate still exists over the ontogenetic origins of taxonomic

163 differentiation, though it is generally agreed that most differentiation occurs early in development (Piasecka et al., 2013). Evidence of morphological divergence in embryology among species suggests sequential differentiation from oldest to youngest lineage, which has been thought to recapitulate the phylogeny (Gould, 1977). Yet, little empirical evidence exists for differentiation among closely related genera or species. While divergence among higher taxa occurs early in development (Levin et al., 2016; Piasecka et al., 2013), species-level or even generic differentiation is often poorly understood across large groups of vertebrates. Significant morphological transformations do occur in fishes from hatching to sexual maturity (Katz &

Hale, 2016), and may appear to be more striking both in size and shape than higher vertebrates such as mammals and birds. From this, we may then expect that species-level or generic divergence in morphology could perhaps occur in post-hatching development and growth in lower vertebrates. This may result in more similar morphologies among closely related species of fishes than adults, contrary to findings in the morphological disparity of juvenile mammals

(Zelditch et al., 2016). This increase in disparity over ontogeny has been seen in some taxa, including fishes (Frédérich & Vandewalle, 2011) and lizards (Urošević et al., 2013). While the evolution of ontogenetic slope and its associated morphology and ecology has been examined

(e.g. Adams & Nistri, 2010; Giannini, 2014), few studies have examined how variation in slope affects disparity, and whether other ontogenetic parameters can shape morphological diversity.

The vastness of species, morphological, and ecological diversity across Neotropical cichlids has been widely examined. While many studies do not incorporate phylogenetic correction, or examine the evolution of traits, there is a growing body of literature on trait evolution in Neotropical cichlids. The evolution of morphological and ecological disparity in

Neotropical cichlids associated with feeding ecomorphology and diet (Arbour & López-

Fernández, 2013, 2014), as well as swimming (Astudillo-Clavijo et al., 2015), has been studied

164 using phylogenetic comparative methods. Specialized groups such as the large-bodied, piscivorous Crenicichla, or miniaturized lineages such as Apistogramma, Mikrogeophagus, and

Biotoecus, among others, often occupy unique regions of morphospace, have unique ecological specializations (Cochran-Biederman & Winemiller, 2010; López-Fernández et al., 2012; Soria-

Barreto & Rodiles-Hernández, 2008), or show differences in trait diversification (Arbour &

López-Fernández, 2013; Astudillo-Clavijo et al., 2015; López-Fernández et al., 2013). Yet these studies are restricted to adult phenotypes and do not examine body size directly. The simultaneous effects of divergence among species along an ontogenetic body size axis and divergence along an axis of absolute maximum body size of adults on morphological diversity in Neotropical cichlids has yet to be tested. However, for some species there is overlapping knowledge of growth, development, and ontogenetic shifts as well as ecological or morphological traits that now have been explored in a macroevolutionary framework using adults. Despite this, our knowledge of ontogenetic changes in Neotropical cichlids and how ontogenetic size changes affect disparity of morphology and ecology is limited.

Little is known about ontogenetic changes in morphology and ecology through ontogeny, or differences between juveniles and adults in Neotropical cichlids (but see Mérigoux

& Ponton, 1998,Ponton & Mérigoux, 2000; Ponton & Mérona, 1998). Ontogenetic data is available for a handful of species describing morphological change and associated diet shifts

(Mérigoux & Ponton, 1998; Ponton & Mérigoux, 2000; Ponton & Mérona, 1998; Winemiller,

1989). Phenotypic plasticity via intraspecific divergence of ontogeny has been examined in a select few Neotropical cichlids (Meyer, 1987, 1990; Wimberger, 1991). Theses studies showed that similar environmental (e.g. diet) pressures can result in divergent developmental responses among species, often contingent upon the underlying morphology present in early ontogeny of a given species. Therefore, the direction and rate of ontogenetic change can be, in part,

165 constrained by the morphological condition of the original form in early ontogenetic stages

(McKinney & McNamara, 1991), and thus similar environmental pressures exerted on ontogeny may result in a vast array of developmental solutions. Together, these studies suggest that ontogenetic trajectories can change both within and among closely related species (Meyer, 1987,

1990; Wimberger, 1991). However, previous work concentrates on one or a few species, and no study has analyzed developmental trajectories across cichlid lineages in either a comparative or evolutionary framework.

In this Chapter, I used body size and shape to study morphological disparity through ontogeny in Neotropical cichlids and assess the contributions of ontogenetic and evolutionary allometry to morphological variation. Throughout the study, I used size as a proxy for developmental stage in the absence of reliable age estimates. I tested for associations between size and shape, asked whether these relationships differ across Neotropical cichlids, and described major trends in cichlid development and growth by identifying key traits associated with ontogenetic shape change. I then used ontogenetic allometry of Neotropical cichlids to test

1) whether shape variation is associated with ontogenetic variation within species or with evolutionary divergence among species; 2) whether juvenile shape disparity is reduced compared to adult shape disparity; and 3) whether divergence in morphology across species occurs early or late in ontogeny. Finally, in a phylogenetic framework, I tested whether body size, shape, and ontogenetic parameters are correlated through evolution, and if this evolution supports previously established hypotheses about the evolution of body size optima.

166 4.3 Methods

4.3.1 Data Collection and Taxon Sampling

A total of 1885 specimens, comprising 88 species, were sampled (following Chapter 3, section

2.1) from juvenile to adult life stages to study ontogenetic changes across Neotropical cichlid fishes (mean=23.88 specimens per species; range=16–41). Specimens were measured from the collections of the Royal Ontario Museum (ROM), Academy of Natural Science at Drexel

University (ANSP), Texas A & M University (Biodiversity and Research Teaching Collections–

BRTC), University of Texas at Austin (Texas Natural History Collections–TNHC), the Museu de Zoologia da Pontificia Universidade Catolica do Rio Grande do Sul (MCP) and the Museum of Comparative Zoology, Harvard University (MCZ). Where possible, specimens of a given species were measured from a single geographic locality and sampling time to reduce variability in the data due to among-site and seasonal variability. External morphology was characterized using geometric morphometrics to capture proportional and positional changes of features through development and growth. Photographs of each individual were taken of the left lateral view using a Nikon D750 DSLR camera with a 105mm Nikon AF-S VR Micro-Nikkor lens.

The coordinates of one hundred anatomical landmarks were recorded for each specimen in tpsDig (Rohlf, 2004) to capture ontogenetic changes primarily in the skull elements, lateral line, and the relative size and shape of the head compared to the body (landmarks 1:100: Fig. 4.1).

All coordinate data was then imported into R (R Core Team, 2016) and statistical analyses were carried out using this software.

Landmark coordinate data was then aligned to remove non-shape associated variation in the data using generalized Procrustes superimposition as implemented in the function gpagen in

167 the R package ‘geomorph’ (Adams & Otárola-Castillo, 2013). Aligned Procrustes coordinates for all specimens were then used in all subsequent shape analyses. Centroid size, a measure of size that is mathematically independent of shape, was calculated for each specimen and retained for use in further analyses (see Zelditch et al., 2004). Standard length (mm SL, measured from the tip of the upper lip to the point of insertion of the caudal fin rays onto the caudal peduncle) and body mass (g) were measured for each specimen. In addition, I gathered data for the maximum recorded size for each species from the literature (FishBase, Froese & Pauly, 2016). I compared these measures with maximum sizes found in institutional collections and used the largest measure of the two for maximum body size for each species. Age data is unavailable for most Neotropical cichlids as traditional methods of collecting age data from otolith growth patterns are not robust (Jepsen et al., 1999) and developmental studies describing shape responses to age are rare. Standard length is used here as a proxy for age, and while not robust, it has been shown that developmental changes in fish may be accurately studied using size

(Frédérich & Sheets, 2010; Zelditch et al., 2004).

168

Figure 4.1: The two-dimensional external morphology landmarks used for shape analysis across the ontogeny of Neotropical cichlids. Solid circles indicate fixed landmarks; dashed circles indicate sliding semi-landmarks. 1 – Tip of snout; 2-6 – Surface of snout ventral to 7; 7 – Anterior-most end of scaled nape; 8-12 – Surface of scaled nape; 13 – Insertion of first spiny dorsal ray; 14-18 – Dorsal surface at insertions of spiny rays; 19 – Insertion of last spiny dorsal ray; 20 – Distal tip of last spiny dorsal ray; 21-22 – Dorsal surface at insertions of soft dorsal rays; 23 – Insertion of last soft dorsal ray; 24-26 – Dorsal surface of caudal peduncle; 27 – Dorsal insertion of caudal fin; 28 – Posterior-most point of body midline, intersection of caudal margin and lower lateral line (LLL); 29-31 – LLL; 32 – Anterior-most point of LLL; 33 – Ventral insertion of caudal fin; 34-36 – Ventral surface of caudal peduncle; 37 – Posterior insertion of anal fin; 38-40 – Ventral surface at insertion of anal fin rays; 41 – Anterior insertion of anal fin; 42-48 – Ventral surface between pelvic and anal fins; 49 – Anterior insertion of pelvic fin; 50-52 – Ventral surface between 53 and anal fin; 53 – Ventral tip of cleithrum; 54 – Intersection of posterior margin of operculum and ventral surface; 55 – Anterior tip of hyoid; 56-57 – Ventral surface of jaws; 58 – Anterodorsal point of lower lip; 59-66 – Perimeter of orbit; 67 – Dorsal point of preopercle; 68-74 – Posterior/ventral margin of abductor mandibulae; 75 – Ventral-most point of lacrimal process; 76-80 – Posterior margin of preopercle; 81 – Inflexion of posterior margin of preopercle; 82-84 – Ventral margin of preopercle; 85 – Anterior-most point of preopercle; 86 – Ventroposterior point of descending process of maxilla; 87 – Dorsal-most point of operculum; 88-92 – Posterior margin of operculum; 93 – Posteroventral point of operculum; 94 – Anterior-most point of upper lateral line (ULL); 95-97 – ULL; 98 – Posterior-most point of ULL; 99 – Dorsal insertion of pelvic fin; 100 – Ventral insertion of pelvic fin.

169 4.3.2 Quantifying Ontogenetic Allometry

Procrustes ANOVA was performed on the aligned coordinate data to assess the covariation of shape with size across cichlid ontogenies. Ontogenetic allometry was quantified by fitting a linear regression model to the data using the function procD.allometry in the package

‘geomorph’ (Adams & Otárola-Castillo, 2013). I included species identity as a cofactor to assess relative species differences in a global size-shape space. Ontogenetic slopes were further analyzed using advanced.procD.lm in the package ‘geomorph’ (Adams & Otárola-Castillo,

2013). To test for differences in slopes among species I included an interaction term in the full model shape ~ size*species, allowing unique slopes for each species (i.e., divergent ontogenetic trajectories), and compared the fit of this model to that of a reduced null model of shape ~ size + species, which constrains slopes to be equal but allows mean differences among species (i.e., parallel trajectories). The inclusion (failure to reject the null) or exclusion (no rejection of the null) of the interaction between size and species would therefore support significant differences in ontogenetic slopes or the same slope among species, respectively. Similarly, to test for differences in mean shape among species, I compared a full model (shape~size + species) against a reduced model (shape~size) and considered differences to be significant if the full model was preferred. Probability of shape variation was estimated compared to a null model using distributions from 999 resampling permutations. The predicted shape values of each specimen from the regression model were retained for further analyses.

Due to the high degree of intraspecific size variation in the dataset I expected that much of the observed morphological variation would be attributable to ontogenetic shape changes within species rather than across species. In addition, the first component of a PCA incorporating multiple species as well as ontogenetic data within species often captures both

170 intra and interspecific variation (Mitteroecker et al., 2004), confounding the ontogenetic and evolutionary components of morphological variation. Therefore, I examined the common allometric component (CAC, see Mitteroecker et al., 2004) to look at shape variation that could be attributed to size through development and growth independent of species identity, as well as residual shape components (RSCs, see Mitteroecker et al., 2004) capturing the shape variation that is independent of ontogenetic size. The CAC approximates the average trends in shape change across species associated with ontogenetic development. The CAC was calculated by conducting a pooled within-group regression (correcting for species means) of Procrustes coordinates (shape) on log-transformed centroid size (size). Residuals from the regression of shape on CAC were then examined using PCA to identify residual shape variables that could describe the remaining shape variation.

4.3.3 Examining Allometric Variation

To examine the variation in the association of shape and size among species, I examined the magnitude of shape change and orientation of allometric vectors across species. Vector lengths of allometric ontogenies and angles between vectors were estimated using the function advanced.procD.lm. Vector lengths indicate the magnitude of shape change per covariate unit, therefore they are a rough approximation of the amount of change in external morphology over changes in centroid size, thus approximating the rate of shape change over ontogeny. Angles between ontogenetic vectors indicate changes in the orientation of allometric trajectories among species, and therefore indicate significant divergence in shape trajectories over ontogeny. To examine the differences in group means, suggesting interspecific difference in shape overall, I compared least squares means for species using advanced.procD.lm.

171 4.3.4 Examining Shape Disparity

We then examined if shape changes over ontogeny resulted in significant differences in the amount of morphological disparity across Neotropical cichlids at various sizes. I tested the hypothesis that early ontogenetic stages in cichlids exhibit lowered morphological disparity compared to later ontogenetic stages, indicating post-larval divergence of species-specific morphology from a common larval morphology. I first performed principal components analysis

(PCA) of shape variation on the Procrustes-aligned coordinate data without correcting for ontogenetic size via the function plotTangentSpace in the package ‘geomorph’ (Adams &

Otárola-Castillo, 2013). To test for differences in disparity between juveniles and adults, I compared pooled shapes (PC scores) of the smallest and largest individuals for each species in the sample, representing juvenile and adult total disparity. Since the minimum size at which I could sample species varied due to availability in collections, I performed the analyses on both individuals at 30 mm SL (n=68 species) as well as individuals at roughly 30% of development

(n=69 species; % development = SL/maximum body size) to compare juveniles at a common SL and ontogenetic stage, respectively. I used convex hull analyses to calculate the total shape disparity in each group, and to determine if the differences in disparity among groups was greater than expected by chance by comparing to a distribution of 1000 distances from randomly resampling the data following (Benitez-Vieyra, 2006; Maubecin et al., 2016). To determine whether the degree of clustering or over-dispersion differed among juveniles and adults I calculated nearest neighbour distances among members of each group using the function nearestNeighborDist in the package ‘paleotree’ (Bapst, 2012).

Differences in juvenile and adult disparity may be attributed to variation in the ontogenetic disparity within species. To examine potential differences in shape disparity among

172 species resulting from ontogenetic change, I calculated convex hull area for each species in shape space (PC1-2) as well as morphological disparity (Procrustes variance). I performed pairwise tests among species morphological disparity using the function morphol.disparity in

‘geomorph’ (Adams & Otárola-Castillo, 2013), accounting for allometry and using a group mean. Procrustes variance is based on the shape residuals from a linear model fit of shape on size, and therefore higher disparity suggests size poorly predicts shape. A poor fit could suggest that a species exhibits very little ontogenetic shape change over ontogeny. Pairwise differences were evaluated using a permutation test, with the observed Procrustes variation compared to a null model using distributions from 999 resampling permutations to estimate probability. I also examined disparity of species ontogenies in shape space by calculating convex hull area for each species in the morphospace of the first two principal components.

To determine if potential sampling biases affected disparity, as measured by either morphological disparity or convex hull area, I examined the relationship between disparity and ontogenetic completeness (percent ontogeny sampled) using linear regression. Ontogenetic completeness was calculated as the difference between the latest ontogenetic stage in the sample

(maximum body size in sample/maximum recorded size) and the earliest ontogenetic stage in the sample (minimum body size in sample/maximum recorded size). I could not take into account hatching size, as this data is not available for most species, but it should not affect calculations for medium to large-bodied species as larval size is a small fraction of maximum adult body size. I examined the fit of a linear model disparity~% ontogeny*min size to the data, considering ontogenetic completeness as well as the earliest stage at which I sampled each species.

173 4.3.5 Exploring the Evolution of Ontogeny and Body Size

To determine if previously established hypotheses of body size optima (Steele & López-

Fernández, 2014; see Chapter 1) and miniaturization (see Chapter 3) are supported when incorporating phylogenetic information, I examined the evolution of body size and six ontogenetic parameters of 74 species, representing 48 genera and all tribes of Neotropical cichlids. Evolutionary relationships among the species were assessed using a time-calibrated molecular phylogeny (Maximum Clade Credibility tree, MCC) pruned to match taxon sampling in this chapter, as well as 1,000 randomly sampled chronograms from the posterior distribution of chronograms from which the MCC tree was obtained as per López-Fernández et al. (2010,

2013). To standardize comparisons between the MCC and among the 1000 posterior distribution trees, all chronograms were scaled to a relative time scale with total length of one (López-

Fernández et al., 2013). The phylogeny combines three mitochondrial and two nuclear markers for 154 species from 57 genera and all major tribes of Cichlinae (Astronotini, Cichlini,

Cichlasomatini, Chaetobranchini, Geophagini, Heroini, Retroculini) representing approximately

26% of currently described taxa. Further details of the phylogeny and time-calibration are given in López-Fernández et al. (2010, 2013).

My parameter dataset included maximum attainable body size (SL), ontogenetic slope, ontogenetic threshold, juvenile shape, adult shape, and length:weight ratio, and morphological integration (i.e., covariance between subsets of traits). Maximum attainable body size was the larger of the two values measured from samples within this chapter or maximum recorded size from the literature. Ontogenetic slope was determined for each species from the estimated regression coefficients of shape on ontogenetic size (see Examining Allometry). The ontogenetic threshold is the threshold between juvenile and adult shape. This threshold was estimated for

174 each species using piece-wise regression (Chapter 3 Methods for details) on the predicted shape scores (see Examining Allometry) and untransformed standard length. Juvenile and adult shape was measured as the shape predicted by the regression equation best fitting the ontogenetic data for each species at the smallest and largest body sizes sampled, respectively. Since geometric morphometrics using 2-dimensional landmarks may not capture size changes associated with volume seen across the phylogeny, I also incorporated length:weight ratio to examine changes is mass (g) relative to changes in body length (SL mm). The ratio was estimated for each species using nonlinear least squares regression of mass~length of all specimens per species using the function ‘nls’ with starting parameters of 0.001 and 1.75, for intercept and slopes, respectively.

There is considerable evidence in fishes that cranial morphology is dramatically affected by reductions in body size (Bennett & Conway, 2010; Hanken & Wake, 1993; Weitzman & Vari,

1988). At small body size, sensory volume largely constrains head shape, whereas the body may be freer to vary in shape (Hanken & Wake, 1993). I therefore expected the development of cranial features to have a different relationship with ontogenetic and maximum body size than postcranial features, and to be less integrated with postcranial features at small body size. I tested whether there was significant covariation in each species between landmark coordinates describing head and postcranial shape using the function ‘integration.test’ in the package

‘geomorph’ (Adams & Otárola-Castillo, 2013).

I tested the fit of a select number of evolutionary models that represented my expectations of trait and parameter evolution given the phylogeny: 1) BM1; 2) OU1; 3) OU2; 4)

OU3; 5) OUM. The Brownian Motion (BM1) model of evolution assumes morphological evolution follows a random-walk process of evolution, with morphological variance (disparity) proportional to time. This BM1 model was fit under the assumption that the BM rate parameter

(2) was constant across the phylogeny. The various Ornstein-Uhlenbeck models incorporate a

175 constraint () on morphological evolution under the BM model towards adaptive optima ()

(Butler & King, 2004; Hansen, 1997), with the single-peak OU1 model assuming that evolution is constrained towards a single adaptive optimum. The OU2 model tests for evolution of traits towards two adaptive peaks, which I expected to correspond to miniature cichlids and medium- large bodied cichlids, respectively (and see Chapter 3). The three-peak OU3 model assumes traits evolve towards three optima corresponding to three potential body size peaks previously identified in Neotropical cichlids (Steele & López-Fernández, 2014). Models were fit on the body size and ontogenetic parameter data simultaneously using the multivariate functions

‘mvBM’ (BM1) and ‘mvOU’ in the package ‘mvMORPH’ (Clavel et al., 2015). It is possible that these hypotheses are not complex enough to explain the evolution of body size and ontogenetic parameters simultaneously. Therefore, I additionally tested the fit of a multi-peak

(OUM) model of evolution without an a priori hypothesis of how many adaptive peaks should exist or where these regime shifts should occur across the phylogeny. To implement the latter test, I employed the SURFACE algorithm (Ingram & Mahler, 2013) to fit increasingly complex

OU models to search for potential adaptive and convergent peaks. A “forward” phase successively adds adaptive peaks to branches on the phylogeny until the addition of new adaptive peaks results in an improvement of the AIC score of less than 2 units. The next

“backwards” phase collapses similar adaptive peaks leading to improvements of the AIC score, again until improvements less than 2 units are reached. Models were fit both on the MCC tree and on the sample of 1000 trees from the posterior distribution to investigate the effects of phylogenetic uncertainty on model selection.

176 4.4 Results

4.4.1 Quantifying Ontogenetic Allometry

I found a significant relationship between shape and size in all species, excluding Astronotus ocellatus and Crenicichla geayi. I found that the full model including an interaction between size and species identity fit my data better than regression models with no interaction or species term, rejecting the null of all ontogenetic slopes being equal (df =1709, F= 4.27, Z=3.58, P- value =0.01) and having the same mean (df=1796, F = 57.168, Z=12.93, P-value=0.002).

Ontogenetic slopes ranged from -0.017-0.094 (Mean = 0.036 +/- 0.018) and intercepts ranged from -0.269-0.079 (Mean = -0.097 +/- 0.052). The coefficient of determination for size ranged from 0.12 to 0.59 (mean=0.31) across species showing a significant relationship between shape and size, and was not significantly associated with sample size.

When taking size of individuals into account to look at shape variation during development and growth, the CAC accounts for 40.4% of shape variation (Fig. 4.2a), RSC 1 accounts for 15.5% of shape variation (Fig. 4.2b), RSC 2 accounts for 13.2% (Fig. 4.2c), and

RSC 3 accounts for 7.6% of shape variation among individuals (Fig. 4.2d). Together, the CAC and the first three RSC account for 76.7% of shape variation among Neotropical cichlids. The

CAC captured variation in head proportion, body depth, lateral line length. Relative to the grand mean of shape, individuals with lower CAC scores exhibit larger eyes, smaller more terminal mouth, smaller operculum, upper and lower lateral lines terminating closer to the point of origin, and relative lengthening of the caudal peduncle (Fig. 4.2e). Relative to the grand mean of shape, individuals with higher CAC scores exhibit smaller eyes further from the snout, shorter more rounded snout, slight lengthening of the head due to widening of the preopercle, interorbital, and

177 opercle, slight lengthening of the upper and lower lateral lines, considerable deepening of the body, and relative shortening of the caudal peduncle (Fig. 4.2e). RSC1 captures variation in body elongation associated with interspecific differences in shape. Relative to the grand mean of shape, individuals with lower RSC 1 scores have shallower bodies, elongate caudal region, and smaller heads while individuals with higher RSC 1 scores have deeper bodies, slightly larger heads, and shorter caudal regions (Fig. 4.2e). RSC2 captures interspecific variation in trunk shape. Relative to the grand mean of shape, individuals with lower RSC 2 scores have dorsal and ventral expansion of the posterior region with sharp tapering towards the caudal region, creating a square trunk while individuals with higher RSC 2 scores have oval, gradually tapering bodies. RSC3 captures variation in snout shape and mouth orientation (Fig. 4.2e). Relative to the grand mean of shape, individuals with lower RSC 3 scores have pointed snouts and terminal mouths while individuals with higher RSC 3 scores have rounded snouts and subterminal mouths (Fig. 4.2e).

4.4.2 Examining Allometric Variation

The magnitude of shape change per unit size (i.e., vector length) ranged from 0.027-0.103

(Mean=0.055; St. Dev=0.015) with the difference between magnitudes ranging from 2.64x10-6-

0.076 (Mean = 0.017; St. Dev=0.013) across all Neotropical cichlids. The angle between allometric trajectories ranged from 17.99o-125.41o (Mean = 61.17o; St. Dev=17.08o).

Differences between least squares means of species ranged from 0.019-0.276 (Mean = 0.090; St.

Dev=0.047). I examined 1213 within-tribe pairwise species comparisons and 2615 among-tribe species comparisons to determine if ontogenetic parameters were more similar within tribes versus among tribes. Within tribes, 387 of 1213 comparisons (31.90%) and among tribes, 831 of

2615 comparisons (31.78%) of magnitude of shape change per unit size were found to be

178

Figure 4.2: Ontogenetic variation in morphology of Neotropical cichlids fishes. Scatterplots of a) Common allometric component (CAC) (pooled within-group regression of shape on centroid size) scores again log-transformed centroid size; b-d) Residual shape component (RSC) 1-3 against scores along CAC; f) Shape change along CAC and RSC 1-3. Vector length and orientation indicate the magnitude and direction of shape change from average shape.

(next page) Figure 4.3: Dissimilarity matrix (heat map) of species pairwise parameter distances, indicating similarity (yellow) and dissimilarity (red) of ontogenetic parameters. Top - Magnitude of shape change (upper triangle) and angle between trajectories (lower triangle); Bottom - Mean shape (upper triangle) and morphological disparity (lower triangle). Species ordered by phylogenetic relationships as per López-Fernández et al. (2010, 2013).

179

180 significantly different (Fig. 4.3a). Within tribes, 710 of 1213 comparisons (58.53%) and among tribes, 1685 of 2615 comparisons (64.44%) of the orientation of allometric trajectories were found to be significantly different (Fig. 4.3a). Within tribes, 1003 of 1213 comparisons

(82.69%) and among tribes, 2437 of 2615 comparisons (93.19%) of least squares means were found to be different (Fig. 4.3b).

4.4.3 Examining Shape Disparity

The PCA of aligned coordinates revealed three important axes of shape variation in Neotropical cichlid fishes: PC1 (48.4% of variance explained) captures variation in body elongation among individuals, PC2 (9.7%) captures variation in proportion of the head relative to the body as well as length of the lateral lines, and PC3 (7.6%) captures variation in head shape associated with mouth orientation, together explaining 65.7% of the shape variation among Neotropical cichlids.

I found no significant differences in convex hull area in this shape space (PC1-2) between juvenile and adult Neotropical cichlids in this study. This was consistent when juvenile morphology was measured at 30mm SL (Da= 0.01, Dj = 0.004, p-value=0.838; Fig. 4.4a) and at

30% development (Da= 0.005, Dj = 0.018, p-value=0.079; Fig. 4.4b). Nearest neighbour distance was found to be similar among adults and juveniles around 30 mm SL (Rangea= 0.038-

0.107, Meana=0.057; St. Dev =0.014; Rangej = 0.035-0.103, Meanj = 0.057; St. Dev = 0.014) as well as between adults and juveniles around 30% development (Rangea= 0.038-0.102,

Meana=0.055; St. Dev = 0.012; Rangej = 0.032-0.095, Meanj = 0.055; St. Dev = 0.014). Among species, I found that morphological disparity associated with ontogeny ranged from 0.002-0.041

(Mean =0.007; St. Dev = 0.008). Within tribes, 420 of 1213 comparisons (34.62%) and among tribes, 907 of 2615 comparisons (34.68%) of ontogenetic morphological disparity were found to

181 be different (Fig. 4.3b). Convex hull area in shape space (PC1-2) for species ontogenetic variation ranged from 0.003-0.008 (Mean = 0.002; St. Dev =0.001; Fig. 4.5). I found no significant relationship between disparity and ontogenetic completeness or the stage at which sampling began, suggesting that the significant differences among species are not due to sampling effects.

4.4.4 The Evolution of Ontogeny and Body Size

BM1, OU1, OU2, and OU3 were fit on body size and ontogenetic parameters together using a multivariate approach. These models were then compared to an OUM model to determine if the evolution of body size and ontogenetic parameters is more complex than my a priori hypotheses of trait evolution. The best supported model for evolution of traits given the MCC tree was the

OUM model, suggesting my expectation that miniatures would occupy a separate peak compared to all other Neotropical cichlids (Chapter 3), or that these parameters were associated with three body size optima as previously suggested (Steele & López-Fernández, 2014) were too simplistic (Table 4.1).

The SURFACE algorithm identified 10 regime shifts (k) associated with my trait dataset, with 8 unique regimes (kꞌ) after accounting for convergence (Fig. 4.6). Two of the regimes exhibit convergence among distantly related lineages while six regime shifts are non-convergent. At the base of the phylogeny, Retroculus xinguensis and Cichla temensis occupy a non-convergent regime mainly characterized by large adult body size, a low length:weight ratio, and a relative elongate shape. A shift occurs at the ancestor of all other tribes towards a large and deep bodied form, which is retained in the genera Chaetobranchus flavescens and Astronotus ocellatus. A

182

Figure 4.4: Principal components analysis of high-dimensional external shape data in 138 specimens in 69 species of Neotropical cichlids, captured using geometric morphometrics. Polygons represent total morphospace occupied by juvenile (grey) and adult cichlids (red), using two standard measures of juveniles a) 30% development and b) 30mm SL.

183

Figure 4.5: Principal components analysis of high-dimensional external shape data in 1885 specimens in 88 species of Neotropical cichlids, captured using geometric morphometrics. Polygons represent total morphospace occupied by each species. Polygon colour is consistent with Figure 2 and Figure 7, with geophagines (blue-purple), cichlasomatines (red-orange- yellow), and heroines (green) coloured to compare tribe divergence.

184 regime shift then occurs at the ancestor of all geophagines, characterized by intermediate values for most traits, except for ontogenetic slope, which is steeper than most Neotropical cichlids.

Table 4.1: Multivariate model fitting results (log likelihood, ΔAIC and Akaike weight) for the comparison of Brownian motion (BM) and Ornstein–Uhlenbeck (OU) models for body size and ontogenetic parameter evolution on the MCC tree. Parameter BM1 OU1 OU2 OU3 OUM logL 822.07 803.81 817.03 837.09 - AIC -1604.14 -1537.62 -1554.06 -1584.18 -1836.48 ΔAIC 232.34 298.86 282.42 252.30 0.00 wAIC 0.00 0.00 0.00 0.00 1.00

However, there is considerable variation among geophagines in all traits. Within Geophagini, I found several regime shifts. Firstly, a non-convergent regime shift occurs at the ancestor of

Crenicichla, Acarichthys, and Biotoecus, followed by a regime shift within a subclade of large- bodied Crenicichla and in Acarichthys heckelii. The regime that includes Biotoecus, Crenicichla geayi, and the small-bodied C. “O-wallacii” and Teleocichla monogramma is characterized by extreme body elongation in juveniles and adults, small body size, and intermediate values for other parameters. Acarichthys heckelii experience a convergent regime shift to that of the ancestral geophagine state, while large-bodied Crenicichla are characterized by less pronounced elongation, large body size, and slightly lowered ontogenetic slopes than small bodied congeners. Outside of this subclade, a convergent regime shift between Taeniacara candidi and

Dicrossus filamentosus was also found, characterized by extreme body size reduction, elongation, negative ontogenetic slopes, and lower morphological integration. Mikrogeophagus ramirezi also experiences a non-convergent regime shift, characterized by extreme body size reduction, a steep ontogenetic slope and high morphological integration, but no elongation.

Finally, a non-convergent regime shift was observed at the ancestor of Cichlasomatini and

185

Heroini, characterized largely by intermediate values for parameters with a comparable amount of variation among species, but having deeper bodies than Geophagini.

4.5 Discussion

4.5.1 Differing Rates and Trajectories Among the Cichlids

We examined the distribution of 88 Neotropical cichlid species in morphospace and size-shape space (Mitteroecker et al., 2004) to examine the variation of external morphology across ontogeny (i.e., ontogenetic allometry) among species. I also examined the strength of the relationship between this shape and size across the Neotropical cichlids, and the variation among species in how individuals move from juvenile to adult morphology. Using geometric morphometrics of external morphology, I hoped to capture much of the morphological variation across species and ontogeny that would be associated with ecologically relevant traits previously studied across the phylogeny of cichlids (e.g. Arbour & López-fernández, 2014; Arbour &

López-Fernández, 2013; Astudillo-Clavijo et al., 2015; López-Fernández et al., 2013). I was largely interested in determining if the morphological divergences in adult life stages observed

(next page) Figure 4.6: Results from the SURFACE analysis. Maximum clade credibility chronogram (MCC) (right panel) with branches painted with colour for respective regime. Inset panels show species (small circles) clustering around inferred adaptive peak (large circles) in trait space: a) ontogenetic threshold between juvenile and adult shape, maximum adult body size; b) morphological integration between cranial and postcranial shape, ontogenetic slope; c) adult shape, juvenile shape; d) and maximum adult body size, length:weight ratio.

186

187 in those studies that has been widely supported is present in the early ontogenetic stages of cichlids, or if there is rapid and significant morphological divergence during ontogeny from a constrained, common juvenile morphology.

I found that all but two species (Astronotus ocellatus and Crenicichla geayi) in my analyses exhibited a significant relationship between shape and size, suggesting that most

Neotropical cichlid species exhibit ontogenetic allometry rather than isometric growth. These results were expected since it has been shown there is considerable morphological change through ontogeny in many cichlids (Burress et al., 2013; Meyer, 1987, 1990; Santos-Santos et al., 2015; Webb, 1990). I found that variation in body shape across ontogeny was best fit by a model that included a size by species interaction, indicating that not all cichlid species follow a common ontogenetic trajectory from juvenile to adult shape. I examined the magnitude of shape change and the orientation of trajectories to determine which species were differing from the overall growth pattern of cichlids, and how ontogenetic shape change was accomplished in these species.

The magnitude, or rate of shape change per unit of size, was generally quite similar among cichlids. As a result, the distances between species magnitudes were often low.

However, when the distance was significantly large, the comparison often included small- bodied or miniature cichlids, which on average have a significantly higher rate of development than large-bodied taxa (Fig. 4.3a). This was previously found, comparing miniature cichlids to members of sister taxa (Chapter 3), here extending more generally to small-bodied versus large- bodied cichlids across the phylogeny. The angle between pairwise ontogenetic trajectories was also quite similar among cichlids. While the differences among these angles were not as stark as differences in magnitude (Fig. 4.3a), I found that the same species with seemingly divergent trajectories were often involved in those comparisons that showed the highest differences. I also

188 found that differences in orientation were twice as common as compared to differences in magnitude, and that no pairwise comparisons of orientation showed the degree of similarity as seen in magnitude (Fig. 4.3a). This suggests that the rate of shape change, and perhaps development, among cichlids is more conserved than the direction of movement about shape space. Development can potentially be widely differing among closely related species

(McKinney & McNamara, 1991), however, there are many parameters beyond developmental rate that determine the overall outcome of development and growth as well as the pathway from juvenile to adult (Klingenberg, 1998).

Heterochronic changes to ontogenetic duration, rather than changes in rates, are quite common in nature and can also appear without combined rate changes (Klingenberg, 1998).

Therefore, differences in the shape and size of adult Neotropical cichlids could be attained strictly through changes in duration of the development and growth periods while maintaining the same growth rate. I found that developmental rate is often higher in small bodied fishes, suggesting that adult body size and shape is achieved more rapidly than large-bodied fish, which may or may not also be accompanied by truncation of development and growth. Based on developmental literature, truncation of ontogeny can happen due to a delay in development, usually due to different levels of developmental progress at the time of hatching, as well as an early termination of development and growth. Cichlids do show variation among species in the size and developmental progress at which individuals hatch (Barlow, 1991), though it is unknown if this amount of variation can cause significant deviations in the ontogenetic trajectories of closely related fishes, or if changes in growth rate or development time compensate for these differences. As seen in other fishes (Zelditch et al., 2004; Frédérich and

Sheets 2010), I assume that the use of size roughly approximates the developmental process, and can be confirmed by the examination of age and developmental data for these species.

189 4.5.2 Conservation of Shape Disparity

While external morphology in this study is fully described in 200 dimensions, a projection into three dimensions in shape space (e.g. PC1-3) would show a spherical distribution of shape for high disparity species and an elongate, ellipsoidal distribution for low disparity species.

Biologically, this represents species having highly constrained shape through ontogeny or a high degree of ontogenetic shape change, respectively. This difference could also be attributed to sampling effort, where species exhibiting spherical distributions are simply lacking in particular ontogenetic stages. Since sampling was not complete in all taxa, I tested for this sampling effect, finding no correlation between disparity and sampling completeness and ruling out sampling error as a potential confounding effect. I found morphological disparity, as calculated by

Procrustes variance and convex hull area, to be relatively similar among most Neotropical cichlid species. Studies of ontogenetic allometry have largely found a strong relationship between shape and size, with low variation in residual morphology (Galatius et al., 2011;

Masters et al., 2014; Mitteroecker et al., 2004; Santos-Santos et al., 2015).

Adult body size has been shown to be associated with changes to ontogenetic pathways

(Collar et al., 2011; Galatius et al., 2011; Masters et al., 2014), however I did not find a strong association between disparity and body size despite qualitative changes to total morphospace occupation. I do find that both miniature taxa (Chapter 3) as well as large-bodied piscivores (e.g.

Cichla, Astronotus, Crenicichla) show apparent reduction in morphospace, and an apparent constraint on shape throughout their ontogenies. Therefore, ontogenetic change may be influenced by body size, but may be more tightly associated with ecology (Collar et al., 2009;

Ponton & Mérigoux, 2000). Strong relationships between body size, ontogeny, and ecology have also been shown across vertebrates (Adams & Nistri, 2010; Collar et al., 2011; Galatius et

190 al., 2011; Masters et al., 2014). I found that Cichla, Crenicichla, Teleocichla, Taeniacara, and

Biotoecus significantly differed from other Neotropical cichlids in morphological disparity.

While the latter two genera are considered miniaturized cichlids (Chapter 3) and may be ecologically constrained (Hanken & Wake, 1993) due to small body size, the former two genera are typically large-bodied piscivores which may experience strong functional constraints on body shape (Collar et al., 2009) and rapid ontogenetic development into these ecological roles

(Ponton & Mérigoux, 2000). In addition, the small-bodied species of Crenicichla (Teleocichla monogramma and Crenicichla wallacii) showed the most extreme differences in disparity, suggesting these taxa were first constrained by body shape, and additionally by small body size.

Adult-like shape, or morphological maturity (Zelditch et al., 2004), is likely attained very early in these five genera, and either early ontogenetic stages before this transition were not sampled in this study, or my landmarks could not capture the relevant ontogenetic changes associated with these specialists. I did not find that morphological disparity was also reduced in the other miniaturized genera, suggesting that reduced body size may not predict ontogenetic constraint.

Cichlid fishes undergo tremendous morphological transformations between the larval and juvenile periods (e.g. Ponton & Mérigoux, 2000; Barlow, 1991). I expected that traits associated with ontogeny should explain more of the shape variation in my dataset, that these traits are more similar across species in early ontogenetic stages, and that morphological disparity in early ontogenetic stages would be reduced compared to adults. Contrary to those expectations, juvenile disparity across species was not significantly reduced as compared to adult disparity when absolute size (SL; 30mm) and ontogenetic stage (30% development) were considered. The presence of adult disparity within early ontogenetic stages has been found in other taxa (Zelditch et al., 2016), but to my knowledge not in fishes (Frédérich & Vandewalle,

2011), although reductions in adult disparity have been found (Zelditch et al., 2003). While I

191 found disparity to be similar across juveniles and adults, I do see a morphological shift across

Neotropical cichlids from a juvenile shape to an adult shape early in ontogeny, albeit a relatively constrained process (Fig. 4.2a). My study shows that differentiation between juvenile and adult morphology, while variable, occurs quite early in ontogeny and does not lead to increased disparity in morphology. The maturation of morphology (Zelditch et al., 2004) from juvenile to adult shape appears to occur at much larger sizes in Neotropical cichlids than the attainment of adult-like morphology previously suggested (Ponton & Mérigoux, 2000). However, with few larval fishes in my study I perhaps missed these important morphological transitions from larval to juvenile shape that would suggest increases in disparity over ontogeny. Species and generic differentiation across the phylogeny is likely occurring at or before the transition from larval to juvenile morphology. It may be that a detailed study of larval morphology may show significant reductions in morphological space as compared to juvenile and adult cichlid fishes.

This cursory approach of comparing total disparity in juveniles to that of adults assumes that individuals of the same species are occupying the same space relative to other individuals in the same group, thus adults merely shift along a common ontogenetic trajectory towards a more developed and deep bodied form. The common allometric component (Mitteroecker et al.,

2004), the variation that is described by a common morphological trend through ontogeny across cichlids (e.g. decreased eye and head size relative to body, elongation of the lateral line, body deepening), explains roughly 40% of shape variation among individuals, suggesting that common ontogenetic shape change is in fact highly important for most species. This also suggests that most species move in the same relative direction from elongate underdeveloped forms to deep bodied, robust forms (Fig. 4.7a); a trend that is seen widely across fishes

(Claverie & Wainwright, 2014; Katz & Hale, 2016). However, as noted above, there is considerable variation in the orientation and magnitude of ontogenetic trajectories (Fig. 4.3a).

192

Figure 4.7: Movement from juvenile (circles) to adult (squares) Neotropical cichlids in morphospace: a) PC1 scores, negatively associated with body elongation, against standard length of specimens and b) PC2 scores, positively associated with lateral line development and negatively associated with head and eye size, against PC1 scores. Species colour consistent with Figure 2 and 5, with geophagines (blue-purple), cichlasomatines (red-orange-yellow), and heroines (green) coloured to compare tribe divergence.

193

Therefore, while a general movement from elongate to deep body is common in

Neotropical cichlids (CAC, 40% variation explained), all Neotropical cichlids do not move through shape space in the same way (Fig. 4.7b). The juvenile and adult shapes, exactly how species move between these forms, and how fast this is accomplished are highly variable. This suggests that the attainment of adult shape and size can be accomplished by numerous modifications to some or all aspects of ontogeny (e.g. rate, juvenile shape, adult shape, orientation, duration etc.).

4.5.3 Conservation of Shape Across Ontogeny and Phylogeny

While ontogenetic shape change is important for leading to the adult form, the CAC captures much of the shape variation (~40%), and therefore ontogenetic shape change, at least during the juvenile period, does not appear to contribute to the morphological divergence seen in adult Neotropical cichlids. This divergence occurs much earlier than I can detect with the current dataset. I found that shape differences among species, determined by comparisons of regression means, follow the same general patterns previously outlined in studies of morphological divergence in Neotropical cichlids in which Crenicichla species are almost solely the outliers in shape space (Arbour & López-Fernández, 2013; Astudillo-Clavijo et al., 2015;

López-Fernández et al., 2013). The residuals of the CAC are primarily explained by variation among species in body depth, as well as head shape and mouth position, matching previously established morphological trends in Neotropical cichlids (Arbour & López-Fernández, 2013,

2014; López-Fernández et al., 2013). Interestingly, movement along these axes of variation appears unconstrained (Fig. 4.7). This is supported by detailed ontogenetic work showing opposing ontogenetic responses to environmental changes among Central American cichlids

194

(Meyer, 1987, 1990). I also find that ontogenetic parameters, shape, and body size diversify early in the evolutionary history of cichlid fishes (Fig. 4.6) as seen in previous studies of cichlid trait evolution (Arbour & López-Fernández, 2014; López-Fernández et al., 2013). Previous studies across vertebrates have found that ontogenetic divergence has been associated with the diversification of body size and ecology (Collar et al., 2011; Galatius et al., 2011; Masters et al.,

2014).

Neotropical cichlids have a vast array of morphological and ecological diversity, much of which is captured by the tribe Geophagini (Arbour & López-Fernández, 2013, 2014;

Astudillo-Clavijo et al., 2015; López-Fernández et al., 2012, 2013; Steele & López-Fernández,

2014). Lineages within this tribe, such as large-bodied piscivores (e.g. Crenicichla), have often diverged in morphological and ecological space away from more typical Neotropical cichlid phenotypes. However, some areas of morphological space occupied by geophagines are shared with members of other tribes, particularly the heroines, while some unoccupied regions have been accessed by non-geophagine lineages. The divergence of Neotropical cichlids, and the characteristic overlap of Geophagini and South American Heroini, and the vast expansion of morphological and ecological diversity of Central American Heroini is largely based on feeding ecomorphology (Arbour & López-Fernández, 2013; López-Fernández et al., 2013), feeding functional traits (Arbour & López-Fernández, 2014), as well as swimming functional traits

(Astudillo-Clavijo et al., 2015) in adult fishes. Here I examine the role of ontogenetic body size as well as maximum attainable body size of adult cichlids in shaping divergence patterns.

Similar to these previous studies of adult phenotype, I find that divergence of traits occurs early in the history of Neotropical cichlids, with major shifts in phenotypic evolution occurring at the ancestor of major lineages of cichlids (e.g. tribes). I find that shifts in ontogenetic parameters, shape, and body size (Fig. 4.6) tend to occur in the tribe Geophagini, which exhibits greater

195 morphological (López-Fernández et al., 2012, 2013), ecological (Arbour & López-Fernández,

2014), and body size diversity (Steele & López-Fernández, 2014) than other Neotropical cichlid tribes. Functional shifts or unique adaptive peaks have often been found within this tribe

(Arbour & López-Fernández, 2013, 2014; Astudillo-Clavijo et al., 2015), particularly in

Crenicichla, but fewer shifts than seen in ontogenetic parameters here (Fig. 4.6). In addition to a shift differentiating Crenicichla from other geophagines, I also found regime shifts in miniature cichlids, as well as shifts within Crenicichla differentiating small and large-bodied taxa; a pattern not yet seen in other examined traits. In addition, I find no evidence of convergence among tribes as seen in body shape and feeding function (Arbour & López-Fernández, 2014), suggesting that ontogeny is highly divergent among the major lineages.

Increasing evidence now shows that ecological diversity in Heroini is relatively high

(Arbour & López-Fernández, 2014; Hulsey, 2006; López-Fernández et al., 2013; Winemiller et al., 1995), and the trait conservatism seen in my study may be associated with body size conservatism across Cichlasomatini and Heroini and the lack of extreme body sizes (Chapter 3;

Steele & López-Fernández, 2014). Therefore, body size evolution in Geophagini may drive the divergence seen in ontogenetic parameters within this tribe. I also suspect that where ontogenetic change may play a stronger role in morphological divergence from the general cichlid morphology, I am lacking in this dataset. Morphologically distinct heroine taxa, such as

Pterophyllum and Symphysodon, may show relatively extreme morphological differentiation during the juvenile stages, however ontogenetic series were unavailable for this study. While my hypotheses that traits would evolve towards two (i.e., miniatures, medium-large bodied cichlids;

Chapter 3) or three (i.e., small, medium, large body size; Chapter 1, Steele & López-Fernández,

2014) optima, I do find a few cases of unique or convergent regimes in miniaturized cichlids

(e.g. Mikrogeophagus, Dicrossus, Taeniacara) as well as shifts within clades with high body

196 size diversity (e.g. Crenicichla). This further supports the association of ontogenetic shifts and adult shape with adult body size despite support for a model with increased complexity of trait evolution (Fig. 4.6). I also show strong divergence in ontogenetic parameters and shape associated with large-bodied predators (e.g. Cichla, Crenicichla) which may be one of the few functionally and morphologically constrained ecological roles (Arbour & López-Fernández,

2013; Collar et al., 2009). The developmental point at which taxonomic differentiation occurs is highly dependent on the relatedness of taxa, with more distantly related taxa diverging earlier in development pathway than more closely related species. This also suggests, at some developmental stage we should expect complete morphological convergence of any given set of taxa. Therefore, a more detailed developmental study incorporating phylogenetic relatedness may elucidate the critical points at which differentiation occurs among lineages, and may highlight the importance of subtler con-generic differentiation during ontogenetic stages I have described here.

4.5.4 Conclusion

I examined the ontogenetic trajectories across a broad group of Neotropical cichlid fishes to assess the variation in growth and development across the phylogeny. I found considerable variation in the relationship of shape and size, with the most variation occurring in the orientation of ontogenetic trajectories, or the direction of shape change. The rate of shape change was quite conserved across Neotropical cichlids, with significant shifts in miniaturized cichlids and elongate piscivorous species. Pairwise comparisons of shape change often found these same species to diverge in rate of shape change through ontogeny. I found that almost half the variation was attributed to a common allometric trend from elongate to deeper bodied adults,

197 the variation in rate and orientation also resulted in considerable variation in how cichlids shift from juvenile to adult form. Despite this variation, I found that juvenile disparity was not significantly reduced compared to adult disparity, suggesting that divergence of Neotropical cichlid form occurs much earlier in ontogeny than the juvenile stage. The evolution of body size and ontogenetic parameters was best fit by a multi-peak adaptive landscape similar to previous studies of Neotropical cichlid phenotype, with early divergence among the major lineages of cichlids. I did, however, find additional adaptive peaks within the tribe Geophagini, corresponding to regime shifts within Crenicichla, and convergence across distantly related miniaturized lineages, a pattern not previously found. The complexity of this landscape and the pattern of lineages occupying unique regimes suggests that changes to ontogenies are more closely tied to divergence in maximum adult body size, rather than diversification of body shape or ecology.

198 C. General Conclusion

C.1 Summary

Body size is an important trait that is highly correlated with many traits of interest in ecology and evolution. Yet, we still lack a complete understanding of how body size evolves across broad groups of organisms, especially in the context of its adaptive significance and influence on associated sets of traits such as morphology and ecology. While the study of body size has been approached from several fields of study, it is rarely studied in the context of multiple factors influencing its evolution, or at a variety of spatial, taxonomic, or temporal scales.

Linking the ecological knowledge we have gained about correlates of body size with macroevolutionary patterns and process of trait and species diversification provides an integrated understanding of the evolution of body size and its effects on diversification of species, ecology, and morphology.

Neotropical cichlids are a species rich and ecological diverse group of fishes, and offer an opportunity to examine the drivers of adaptive diversification in species-rich regions of the world. I have combined a variety of tools (e.g. body size frequency distributions, geometric morphometrics, phylogenetic comparative methods) previously used to study body size at specific scales, in attempt to bridge our knowledge of broad macroecological questions of body size with fine-scale development processes, while incorporating knowledge of the phylogenetic relationships among species and evolutionary history of traits. I found that selective pressures associated with occupying a particular environment (i.e., marine or freshwater) likely influence the evolution of body size in fishes, especially those occupying riverine habitats (Chapter 2).

There is strong evidence that repeated transitions to freshwater results in evolution towards

199 smaller body size overall, while lineages evolving in marine environments have evolved larger body sizes. While riverine fishes tend to have highly right-skewed distributions, small body size, and high body size diversity, Neotropical cichlids are not right-skewed or small as expected.

Body size of cichlids also appears to be highly diverse compared to many endemic Neotropical lineages (Chapter 2). Rabosky et al. (2013) found that cichlids have one of the highest rates of body size evolution. Increased rates of evolution could drive the higher size diversity seen as compared to other lineages. I did find that some Neotropical cichlids lineages do indeed exhibit both extreme reduction in body size as well as significant increases (Chapter 1). However, these do not reach the extremes experienced in other Neotropical lineages, or freshwater lineages outside the Neotropics (Chapter 2), and are rare events compared to other Neotropical lineages

(e.g. Characiformes, Siluriformes). The degree of size evolution among Neotropical cichlid lineages offers a unique opportunity to examine the role of body size in shaping morphological and ecological diversity. But why does the size distribution of Neotropical cichlids differ so greatly from other Neotropical fishes?

While clear hypotheses about the evolution of extreme body size increase is lacking in the literature, there is abundant literature on the consequence of body size reduction and miniaturization. It has been suggested that while cichlids do not exhibit reduction in body size observed in many miniature lineages (e.g. characids, cyprinids), they do qualitatively show morphological modification expected with biologically significant reduction (Weitzman & Vari,

1988). Therefore, I tested whether small-bodied lineages of Neotropical cichlids were biologically reduced (i.e., paedomorphic, truncated ontogenies, and significantly smaller) and therefore could be considered true miniature fishes. The evolution of miniaturization within taxa has implications for morphological and ecological diversity, as well as often produces drastic modifications to life history and physiology. As a result, I expect the evolutionary processes and

200 patterns in the diversification of miniaturized Neotropical cichlids to be significantly different than those lineages not experiencing body size reduction. I found that many lineages do in fact exhibit strong evidence for miniaturization compared to their sister taxa (see Chapter 3), however in a broader context of Neotropical cichlids were not commonly supported as having significantly different evolutionary processes (Chapter 4). Importantly, I found that body size can repeatedly evolve towards reduced body size that leads to miniaturization. It appears that adult body size, both extreme and non-extreme, has an impact on ontogenies, as well as the shape and ecological diversity seen across cichlids (Chapter 3; Chapter 4). Yet, I found considerable conservation of traits across the phylogeny similar to previous studies of

Neotropical cichlid trait evolution (Arbour & López-Fernández, 2013, 2014; Astudillo-Clavijo et al., 2015). Ontogeny and body size appear to diverge considerable between the major lineages of cichlids, with relatively few shifts within Neotropical cichlid tribes. This suggests that the adaptive divergence of ecology experienced early in the evolutionary history of this group

(Arbour & López-Fernández, 2014; López-Fernández et al., 2013) was also experienced in traits associated with ontogenetic morphological change. Unexpectedly, I showed that the morphospace occupied by adult Neotropical cichlids, and the divergence patterns among the tribes Cichlasomatini, Geophagini, and Heroini, is present early in ontogeny (Chapter 4). This finding contrasts previous studies on disparity changes over ontogeny in fishes (Frédérich &

Vandewalle, 2011), although there is little empirical evidence in fishes or other vertebrate groups on ontogenetic disparity.

201 C.2 Future Directions

In my thesis, I focused on the relationship between shape and size in Neotropical cichlids, how size and ontogeny vary across the phylogeny of cichlids, and how this impacts morphological diversity across the ontogeny of cichlids. Body size evolution and ontogeny has been shown to affect or be affected by ecological traits such as habitat (Collar et al., 2011) or diet (Masters et al., 2014). Changes in body size among taxa have also be shown to impact life history and developmental pathways (Angielczyk & Feldman, 2013; Galatius et al., 2011). While the correlation of body size evolution with many traits is well studied, this is often exclusively in adults. Very little is known about the evolution of ontogenetic body size on the diversification of species, ecology, morphology, and life history. In addition, the macroevolution of ontogeny is poorly studied beyond allometric slope. This parameter rarely captures all information that can be gleaned from ontogenetic trajectories, and the study of rates and directionality of ontogenies are becoming more common. New developments in phylogenetic comparative methods (e.g.

Goolsby, 2015) may allow more accurate studies of ontogeny by allowing intraspecific data across the ontogeny to be incorporated. This method, inspired by the function-valued response of traits on an independent factor, could be used to study evolution of ontogenetic trajectory shape, allowing simultaneous assessment of duration, rates, orientation, position, as well as the location of the onset and offset of ontogenetic trajectories. It would then be possible to study the evolution of ontogeny more fully and how species diverge morphologically and ecologically through ontogenetic and evolutionary time. Examining disparate ontogenies within and among closely related species (e.g. Meyer, 1987, 1990; Santos-Santos et al., 2015) could shed light on how subtle changes in the developmental process can ultimately lead to speciation.

202

The advancements of geometric morphometrics (see Adams et al., 2004), in particular their movement into a phylogenetic framework, have allowed broader studies of ontogeny to flourish. Many studies have used ontogenetic allometry with geometric morphometrics to highlight ontogenetic divergence of miniature (i.e., paedomorphic) taxa relative to closely related taxa exhibiting ‘normal’ morphology (Angielczyk & Feldman, 2013; Galatius et al.,

2011; Masters et al., 2014). Some have additionally examined the potential effects of these ontogenetic shifts on the ecology of small-bodied species (Masters et al., 2014). Small-bodied species across vertebrates may be reduced in the ecological opportunity due to constraints on morphology associated with feeding. This was previously proposed in mammals (Monroe &

Bokma, 2009), not necessarily as a constraint, but an adaptive shift to access ubiquitous or previously inaccessible resources. Further studies examining the diversification of diet and feeding function are needed to test whether body size truly impacts ecological diversity, and whether this impact is advantageous. In this thesis, I provide some evidence that ecological roles appear to be restricted at extremely small and large body size, however this was not tested directly. While we know speciation rates are correlated with body size (Rabosky et al., 2013), it is still unclear if the occupation of particular sizes affects the rates at which species or traits diversify. I would also like to test whether small body size broadly results in limitations of morphological or ecological disparity as suggested in mammals (Monroe & Bokma, 2009) and birds (Bokma, 2004). I show that ontogenetic size, at least in juveniles, does not result in lower disparity. But does evolutionary size (i.e., maximum body size) and miniaturization limit function? If constraints imposed by extreme body size restrict evolution, why is it that we observe considerable ranges of body size within and among taxa, and repeated evolution of miniaturized taxa across the tree of life? A comprehensive study on ecology through ontogeny across the phylogeny of cichlids or other species rich clades could allow for a greater understanding of how ontogeny shapes the adaptive landscape of adult phenotype. With more

203 advance phylogenetic methods, we can assess the movement of species about a dynamic adaptive landscape, and determine when and how adaptive peaks form during the maturation of morphology and ecology.

Most miniaturization or body studies focus on the link between body size and the evolution of other traits (e.g. Zimkus et al., 2012), few studies have examined the link between processes associated with body size evolution and processes of species or trait evolution.

Ontogenetic sampling primarily limited the conclusions of this thesis, particularly in the early life history of species. Including larval morphology more broadly in a study of ontogenetic allometry will likely result in a different conclusion regarding morphological disparity associated with ontogenetic size change. Incorporating age data, as well as more detailed studies of development could help identify the timing of morphological divergence in Neotropical cichlids. I was unable to determine at which point morphological differentiation occurs among cichlids. I am interested to find at what ontogenetic point does generic and species level divergence occur in cichlid fishes, which I expect may occur even before hatching. This would mean that adult disparity in cichlids is truly present at the timing of hatching across the phylogeny, and that ontogenetic change in fishes from hatching to adult is not more extensive than in higher vertebrates. Despite this, there is evidence that many species of Neotropical cichlids do experience ontogenetic diet shifts and ecomorphological changes associated with feeding mode. Therefore, even the presence of characteristic features of species or genera at hatching to not prevent ontogenetic changes in behaviour from occurring. Therefore, incorporation of diet data, ontogenetic preferences for habitat, as well as ontogenetic changes in feeding and swimming performance could further elucidate the divergence of trait diversity over ontogeny. From this, I can examine the movement of cichlids through ontogeny on an adaptive landscape, that with only studies of morphology may appear to be quite static over ontogeny. I

204 suspect that the early adaptive landscape of juveniles will be quite different than that of adults, and that the change of the landscape through ontogeny will be affected by both ontogenetic size, as well as evolutionary body size as a result of the vast diversity of maximum body sizes that

Neotropical cichlids can attain (Chapter 1).

While I have provided some knowledge of how body size evolves, how the evolution of this trait affects the distribution of size in extant taxa across the phylogeny, and how this relates to the underlying developmental processes, there is still considerable work to be done to understand the evolution of body size in Neotropical cichlid fishes. Over time, the macroevolutionary and microevolutionary factors affecting body size can be married to form an integrated understanding of how body size shapes the diversity, in all forms, we see across the globe.

205 References

Adams, D. C., & Nistri, A. (2010). Ontogenetic convergence and evolution of foot morphology

in European cave salamanders (Family: Plethodontidae). BMC Evolutionary Biology, 10,

216. https://doi.org/10.1186/1471-2148-10-216

Adams, D. C., & Otárola-Castillo, E. (2013). Geomorph: An R package for the collection and

analysis of geometric morphometric shape data. Methods in Ecology and Evolution, 4(4),

393–399. https://doi.org/10.1111/2041-210X.12035

Adams, D. C., Rohlf, F. J., & Slice, D. E. (2004). Geometric morphometrics: Ten years of

progress following the “revolution.” Italian Journal of Zoology, 71, 5–16.

https://doi.org/10.1080/11250000409356544

Albert, J. S., & Johnson, D. M. (2012). Diversity and Evolution of Body Size in Fishes.

Evolutionary Biology, 39(3), 324–340. https://doi.org/10.1007/s11692-011-9149-0

Albert, J. S., Johnson, D. M., & Knouft, J. H. (2009). Fossils provide better estimates of

ancestral body size than do extant taxa in fishes. Acta Zoologica, 90(SUPPL. 1), 357–384.

https://doi.org/10.1111/j.1463-6395.2008.00364.x

Alerstam, T., Hedenstrom, A., & Akesson, S. (2003). Long-distance migration: Evolution and

determinants. Oikos, 103, 247–260.

Allen, C. R., Garmestani, A. S., Havlicek, T. D., Marquet, P. A., Peterson, G. D., Restrepo, C.,

Stow, C. A., & Weeks, B. E. (2006). Patterns in body mass distributions: Sifting among

alternative hypotheses. Ecology Letters, 9(5), 630–643. https://doi.org/10.1111/j.1461-

206

0248.2006.00902.x

Angielczyk, K. D., & Feldman, C. R. (2013). Are diminutive turtles miniaturized? The ontogeny

of plastron shape in emydine turtles. Biological Journal of the Linnean Society, 108(4),

727–755. https://doi.org/10.1111/bij.12010

Arbour, J. H., & López-Fernández, H. (2013). Ecological variation in South American

geophagine cichlids arose during an early burst of adaptive morphological and functional

evolution. Proceedings of the Royal Society B: Biological Sciences, 280(1763), 20130849.

https://doi.org/10.1098/rspb.2013.0849

Arbour, J. H., & López-Fernández, H. (2014). Adaptive landscape and functional diversity of

Neotropical cichlids: Implications for the ecology and evolution of Cichlinae (Cichlidae;

Cichliformes). Journal of Evolutionary Biology, 27(11), 2431–2442.

https://doi.org/10.1111/jeb.12486

Arita, H. T., & Figueroa, F. (1999). Geographic patterns of body-mass diversity in Mexican

mammals. Oikos, 85(2), 310–319. https://doi.org/10.2307/3546497

Aschliman, N. C. (2014). Interrelationships of the durophagous stingrays (Batoidea:

Myliobatidae). Environmental Biology of Fishes, 97(9), 967–979.

https://doi.org/10.1007/s10641-014-0261-8

Astudillo-Clavijo, V., Arbour, J. H., & López-Fernández, H. (2015). Selection towards different

adaptive optima drove the early diversification of locomotor phenotypes in the radiation of

Neotropical geophagine cichlids. BMC Evolutionary Biology, 15, 77.

https://doi.org/10.1186/s12862-015-0348-7

207

Avaria-Llautureo, J., Hernández, C. E., Boric-Bargetto, D., Canales-Aguirre, C. B., Morales-

Pallero, B., & Rodríguez-Serrano, E. (2012). Body size evolution in extant Oryzomyini

rodents: Cope’s rule or miniaturization? PLoS ONE, 7(4), 1–8.

https://doi.org/10.1371/journal.pone.0034654

Bakker, V. J., & Kelt, D. A. (2000). Scale-dependent patterns in body size distributions of

Neotropical mammals, Ecology, 81(12), 3530–3547.

Bapst, D. W. (2012). Paleotree: An R package for paleontological and phylogenetic analyses of

evolution. Methods in Ecology and Evolution, 3(5), 803–807.

https://doi.org/10.1111/j.2041-210X.2012.00223.x

Barlow, G. W. (1991). Mating systems among cichlid fishes. In: Cichlid Fishes. Behaviour,

Ecology and Evolution, Keenleyside, M. H. A, ed., pp. 173-190. Chapman & Hall, New

York.

Belk, M. C., & Houston, D. D. (2002). Notes and comments Bergmann’s Rule in ectotherms: A

test using freshwater fishes. The American Naturalist, 160(6), 803–808.

https://doi.org/10.1086/343880

Benitez-Vieyra, S. M. (2016). (R)Evolutionary Ecology and Statistics. Available at:

http://santiagombv.github.io/convex_hull/ .

Bennett, M. G., & Conway, K. W. (2010). An overview of North America’s diminutive

freshwater fish fauna. Ichthyological Exploration of Freshwaters, 21(1), 63–72.

Bergmann, C. (1847). Über die Verhältnisse der Wärmeökonomie der Thiere zu ihrer Größe.

Göttinger Studien 1:595–708.

208

Betancur-R., R., Broughton, R. E., Wiley, E. O., Carpenter, K., Lopez, J. A., Li, C., … Orti, G.

(2013). The Tree of Life and a New Classification of Bony Fishes. PLoS Currents, Apr 18.

Ed. https://doi.org/10.1371/currents.tol.53ba26640df0ccaee75bb165c8c26288.Revisions

Betancur-R., R., Wiley, E. O., Arratia, G., Acero, A., Bailly, N., Miya, M., Lecointre, G., &

Ortí, G. (2017). Phylogenetic classification of bony fishes. BMC Evolutionary Biology,

17(162). https://doi.org/10.1186/s12862-017-0958-3

Blackburn, T. M., & Gaston, K. J. (1994). Animal body size distributions: pattems, mechanisms

and implication. Tree, 9(12), 471–474.

Blackburn, T. M., Gaston, K. J., & Loder, N. (1999). Geographic gradients in body size: A

clarification of Bergmann's Rule. Diversity and Distributions, 5(4), 165–174.

Blanchet, S., Grenouillet, G., Beauchard, O., Tedesco, P. A., Leprieur, F., Dürr, H. H., Busson,

F., Oberdorff, T., & Brosse, S. (2010). Non-native species disrupt the worldwide patterns

of freshwater fish body size: Implications for Bergmann’s rule. Ecology Letters, 13(4),

421–431. https://doi.org/10.1111/j.1461-0248.2009.01432.x

Blueweiss, A. L., Fox, H., Kudzma, V., Nakashima, D., Peters, R., & Sams, S. (1978).

Relationships between body size and some life history parameters. Oecologia, 37, 257–

272.

Bokma, F. (2004). Differential rates of morphological divergence in birds. Journal of

Evolutionary Biology, 17(5), 933–940. https://doi.org/10.1111/j.1420-9101.2004.00761.x

Britz, R., & Conway, K. W. (2009). Osteology of paedocypris, a miniature and highly

developmentally truncated fish (Teleostei: Ostariophysi: Cyprinidae). Journal of

209

Morphology, 270(4), 389–412. https://doi.org/10.1002/jmor.10698

Brown, J. H., Gillooly, J. F., Allen, A. P., van Savage, M., & West, G. B. (2004). Toward a

metabolic theory of ecology. Ecology, 85(7), 1771–1789. https://doi.org/doi:10.1890/03-

9000

Brown, J. H., Marquet, P. A, & Taper, M. L. (1993). Evolution of body size: consequences of an

energetic definition of fitness. The American Naturalist, 142(4), 573–584.

https://doi.org/10.1086/285558

Brown, J. H., & Nicoletto, P. F. (1991). Scaling of species composition : Body masses of North

American land mammals. The American Naturalist, 138(6), 1478–1512.

Brown, W. L., & Wilson, E. O. (1956). Character displacement. Society of Systematic

Biologists, 5(2), 49–64. https://doi.org/10.1093/sysbio/syp096

Burress, E. D. (2016). Ecological diversification associated with the pharyngeal jaw diversity of

Neotropical cichlid fishes. Journal of Animal Ecology, 85(1), 302–313.

https://doi.org/10.1111/1365-2656.12457

Burress, E. D., Duarte, A., Serra, W. S., Gangloff, M. M., & Siefferman, L. (2013). Species-

specific ontogenetic diet shifts among Neotropical Crenicichla: Using stable isotopes and

tissue stoichiometry. Journal of Fish Biology, 82(6), 1904–1915.

https://doi.org/10.1111/jfb.12117

Butler, M., & King, A. A. (2004). Phylogenetic comparative analysis: a modeling approach for

adaptive evolution. The American Naturalist, 164(6), 683–695.

Cardillo, M., Mace, G. M., Jones, K. E., Bielby, J., Bininda-, O. R. P., Sechrest, W., Orme, C.

210

D. L., & Purvis, A. (2005). Multiple causes of high extinction risk in large mammal

species. Science, 309(5738), 1239–1241.

Cheverud, J. M. (1982). Relationships among ontogenetic, static, and evolutionary allometry.

American Journal of Physical Anthropology, 59(2), 139–149.

https://doi.org/10.1002/ajpa.1330590204

Clauset, A., & Erwin, D. H. (2008). The evolution and distribution of species body size. Science

(New York, N.Y.), 321(5887), 399–401. https://doi.org/10.1126/science.1157534

Clavel, J., Escarguel, G., & Merceron, G. (2015). mvMORPH: An R package for fitting

multivariate evolutionary models to morphometric data. Methods in Ecology and

Evolution, 6(11), 1311–1319. https://doi.org/10.1111/2041-210X.12420

Claverie, T., & Wainwright, P. C. (2014). A morphospace for reef fishes: Elongation is the

dominant axis of body shape evolution. PLoS ONE, 9(11).

https://doi.org/10.1371/journal.pone.0112732

Cochran-Biederman, J. L., & Winemiller, K. O. (2010). Relationships among habitat,

ecomorphology and diets of cichlids in the Bladen River, Belize. Environmental Biology of

Fishes, 88(2), 143–152. https://doi.org/10.1007/s10641-010-9624-y

Coetzee, B. W. T., le Roux, P. C., & Chown, S. L. (2013). Scale effects on the body size

frequency distributions of African birds: Patterns and potential mechanisms. Global

Ecology and Biogeography, 22(4), 380–390. https://doi.org/10.1111/j.1466-

8238.2012.00793.x

Collar, D. C., O'Meara, B. C., Wainwright, P. C., & Near, T. J. (2009). Piscivory limits

211

diversification of feeding morphology in centrarchid fishes. Evolution, 63(6), 1557–1573.

https://doi.org/10.1111/j.1558-5646.2009.00626.x

Collar, D. C., Schulte, J. A., & Losos, J. B. (2011). Evolution of extreme body size disparity in

monitor lizards (Varanus). Evolution, 65(9), 2664–2680. https://doi.org/10.1111/j.1558-

5646.2011.01335.x

Cox, C. L., Boback, S. M., & Guyer, C. (2011). Spatial synamics of body size frequency

distributions for North American squamates. Evolutionary Biology, 38(4), 453–464.

https://doi.org/10.1007/s11692-011-9132-9

Davies, P. M., Bunn, S. E., & Hamilton, S. K. (2008). Primary Production in Tropical Streams

and Rivers. Tropical Stream Ecology. Elsevier Inc. https://doi.org/10.1016/B978-

012088449-0.50004-2

Diniz-Filho, J. A. F., & Nabout, J. C. (2009). Modeling body size evolution in Felidae under

alternative phylogenetic hypotheses. Genetics and Molecular Biology, 32(1), 170–176.

https://doi.org/10.1590/S1415-47572009005000004

Dulvy, N. K., & Reynolds, J. D. (1997). Evolutionary transitions among egg-laying, live-

bearing and maternal inputs in sharks and rays. Proceedings of the Royal Society B:

Biological Sciences, 264(1386), 1309–1315. https://doi.org/10.1098/rspb.1997.0181

Felsenstein, J. (1985). Phylogenies and the Comparative Method. The American Naturalist,

124(1), 1–15. https://doi.org/10.1086/521238

Fisher, J. A. D., Frank, K. T., & Leggett, W. C. (2010). Global variation in marine fish body

size and its role in biodiversity-ecosystem functioning. Marine Ecology Progress Series,

212

405, 1–13. https://doi.org/10.3354/meps08601

Frédérich, B., & Sheets, H. D. (2010). Evolution of ontogenetic allometry shaping giant species:

a case study from the damselfish genus Dascyllus (Pomacentridae). Biological Journal of

the Linnean Society, 99, 99–117.

Frédérich, B., & Vandewalle, P. (2011). Bipartite life cycle of coral reef fishes promotes

increasing shape disparity of the head skeleton during ontogeny: an example from

damselfishes (Pomacentridae). BMC Evolutionary Biology, 11(1), 82.

https://doi.org/10.1186/1471-2148-11-82

Freedman, J. A., & Noakes, D. L. G. (2002). Why are there no really, really big bony fishes? An

examination of the dichotomy between maximum body size in teleost and elasmobranchs.

Reviews in Fish Biology and Fisheries, 12, 403–416.

Froese, R., & Pauly, D. (2016). FishBase. Available at: www.fishbase.org.

Fu, C. Z., Wu, J. H., Wang, X. Y., Lei, G. C., & Chen, J. K. (2004). Patterns of diversity,

altitudinal range and body size among freshwater fishes in the Yangtze River basin, China.

Global Ecology and Biogeography, 13(6), 543–552.

Fuiman, L. A., Poling, K. R., & Higgs, D. M. (1998). Quantifying developmental progress for

comparative studies of larval fishes. Copeia, 1998(3), 602–611.

https://doi.org/10.2307/1447790

Galatius, A., Berta, A., Frandsen, M. S., & Goodall, R. N. P. (2011). Interspecific variation of

ontogeny and skull shape among porpoises (Phocoenidae). Journal of Morphology, 272(2),

136–148. https://doi.org/10.1002/jmor.10900

213

Giannini, N. P. (2014). Quantitative developmental data in a phylogenetic framework. Journal

of Experimental Zoology Part B: Molecular and Developmental Evolution, 322(8), 558–

566. https://doi.org/10.1002/jez.b.22588

Gillooly, J. F., Charnov, E. L., West, G. B., Savage, V. M., & Brown, J. H. (2002). Effects of

size and temperature on developmental time. Nature, 417, 70–73.

Goodall, C. R. (1991). Procrustes methods in the statistical analysis of shape. Journal of the

Royal Statistical Society: Series B, 53, 285-339.

Goolsby, E. W. (2015). Phylogenetic comparative methods for evaluating the evolutionary

history of function-valued traits. Systematic Biology, 64(4), 568–578.

https://doi.org/10.1093/sysbio/syv012

Gould, S. J. (1971). Geometric similarity in allometric growth: A contribution to the problem of

scaling in the evolution of size. The American Naturalist, 105(942), 113–136.

https://doi.org/10.1086/282710

Gould, S. J. (1977). Ontogeny and phylogeny. Cambridge: Harvard University Press, 1977.

Griffiths, D. (1986). Size-abundance relations in communities. The American Naturalist, 127(2),

140–166.

Griffiths, D. (2006). Pattern and process in the ecological biogeography of European freshwater

fish. Journal of Animal Ecology, 75(3), 734–751. https://doi.org/10.1111/j.1365-

2656.2006.01094.x

Griffiths, D. (2010). Pattern and process in the distribution of North American freshwater fish.

Biological Journal of the Linnean Society, 100(1), 46–61.

214

Griffiths, D. (2012). Body size distributions in North American freshwater fish: Large-scale

factors. Global Ecology and Biogeography, 21(3), 383–392. https://doi.org/10.1111/j.1466-

8238.2011.00680.x

Griffiths, D. (2013). Body size distributions in North American freshwater fish: Small-scale

factors and synthesis. Ecology of Freshwater Fish, 22(2), 257–267.

https://doi.org/10.1111/eff.12023

Grosberg, R. K., Vermeij, G. J., & Wainwright, P. C. (2012). Biodiversity in water and on land.

Current Biology, 22(21), R900–R903. https://doi.org/10.1016/j.cub.2012.09.050

Hallgrímsson, B., Jamniczky, H., Young, N. M., Rolian, C., Parsons, T. E., Boughner, J. C., &

Marcucio, R. S. (2009). Deciphering the Palimpsest: Studying the Relationship Between

Morphological Integration and Phenotypic Covariation (Vol. 36).

https://doi.org/10.1007/s11692-009-9076-5.Deciphering

Hanken, J., & Wake, D. B. (1993). Miniaturization of body size: Organismal consequences and

evolutionary significance. Annual Review of Ecology and Systematics, 24(1), 501–519.

https://doi.org/10.1146/annurev.es.24.110193.002441

Hansen, T. F. (1997). Stabilizing selection and the comparative analysis of adaptation.

Evolution, 51(5), 1341–1351.

Hardman, M., & Hardman, L. M. (2008). The relative importance of body size and

paleoclimatic change as explanatory variables influencing lineage diversification rate: an

evolutionary analysis of bullhead catfishes (Siluriformes: Ictaluridae). Systematic Biology,

57(1), 116–130. https://doi.org/10.1080/10635150801902193

215

Hartigan, J. A., & Hartigan, P. M. (1985). The dip test of unimodality. The Annals of Statistics,

13(1), 70–84. https://doi.org/10.1214/aos/1176346577

Harvey, P. H., & Pagel, M. (1991). The comparative method in evolutionary biology.

Oxford University Press, Oxford.

Hoeinghaus, D. J., Winemiller, K. O., & Taphorn, D. C. (2004). Compositional change in fish

assemblages along the Andean piedmont - Llanos floodplain gradient of the río Portuguesa,

Venezuela. Neotropical , 2(2), 85–92. https://doi.org/10.1590/S1679-

62252004000200005

Holling, C. S. (1992). Cross-Scale Morphology, Geometry, and Dynamics of Ecosystems.

Ecological Monographs, 62(4), 447–502.

Hulsey, C. D. (2006). Function of a key morphological innovation: fusion of the cichlid

pharyngeal jaw. Proceedings of the Royal Society B: Biological Sciences, 273, 669–675.

https://doi.org/10.1098/rspb.2005.3375

Hutchinson, G. E. (1959). Homage to Santa Rosalia or Why are there so many kinds of animals?

The American Naturalist, 93(870), 145–159.

Hutchinson, G. E., & MacArthur, R. H. (1959). A theoretical ecological model of size

distributions among species of animals. The American Naturalist, 93(869), 117–125.

https://doi.org/10.1086/282063

Huxley, J. S. (1932). Problems of relative growth. Methuen, London.

Ilves, K. L., Torti, D., & López-Fernández, H. (2017). Exon-based phylogenomics strengthens

the phylogeny of Neotropical cichlids and identifies remaining conflicting clades

216

(Cichlomorphae: Cichlidae: Cichlinae). bioRxiv, 1–41.

https://doi.org/http://dx.doi.org/10.1101/133512

Ingram, T., & Mahler, D. L. (2013). SURFACE: Detecting convergent evolution from

comparative data by fitting Ornstein-Uhlenbeck models with stepwise Akaike Information

Criterion. Methods in Ecology and Evolution, 4(5), 416–425. https://doi.org/10.1111/2041-

210X.12034

Jepsen, D. B., Winemiller, K. O., Taphorn, D. C., & Rodrigues Olarte, D. (1999). Age structure

and growth of peacock cichlids from rivers and reservoirs of Venezuela. Journal of Fish

Biology, 55, 433–450. https://doi.org/10.1111/j.1095-8649.1999.tb00689.x

Katz, H. R., & Hale, M. E. (2016). A large-scale pattern of ontogenetic shape change in ray-

finned fishes. PLoS ONE, 11(3), 1–7. https://doi.org/10.1371/journal.pone.0150841

Kelt, D. a, Meserve, P. L., Patterson, B. D., & Lang, B. K. (1999). Scale dependence and scale

independence in habitat associations of small mammals in southern temperate rainforest.

Oikos, 85(2), 320–334.

Kenworthy, S. T., & Rhoads, B. L. (1995). Hydrologic control of spatial patters of suspended

sediment concentration at a stream confluence. Journal of Hydrology, 168, 251–263.

Kingsolver, J. G., & Pfennig, D. W. (2004). Individual-level selection as a cause of Cope’s rule

of phyletic size increase. Evolution; International Journal of Organic Evolution, 58(7),

1608–1612. https://doi.org/10.1554/04-003

Kleiber, M. (1932). Body size and metabolism. Hilgardia, 6, 315-351.

Klingenberg, C. P. (1998). Heterochrony and allometry: the analysis of evolutionary change in

217

ontogeny. Biological Reviews of the Cambridge Philosophical Society, 73(1), 79–123.

https://doi.org/10.1111/j.1469-185X.1997.tb00026.x

Klingenberg, C. P., Barluenga, M., & Meyer, A. (2002). Shape analysis of symmetric structures:

quantifying variation among individuals and asymmetry. Evolution; International Journal

of Organic Evolution, 56(10), 1909–1920. https://doi.org/10.1554/0014-3820(2002)056

Knouft, J. H. (2004). Latitudinal variation in the shape of the species body size distribution: An

analysis using freshwater fishes. Oecologia, 139(3), 408–417.

https://doi.org/10.1007/s00442-004-1510-x

Knouft, J. H., & Page, L. M. (2003). The evolution of body size in extant groups of North

American freshwater fishes: speciation, size distributions, and Cope’s rule. The American

Naturalist, 161(3), 413–421. https://doi.org/10.1086/346133

Kolm, N., Goodwin, N. B., Balshine, S., & Reynolds, J. D. (2006a). Life history evolution in

cichlids 1: Revisiting the evolution of life histories in relation to parental care. Journal of

Evolutionary Biology, 19(1), 66–75. https://doi.org/10.1111/j.1420-9101.2005.00984.x

Kolm, N., Goodwin, N. B., Balshine, S., & Reynolds, J. D. (2006b). Life history evolution in

cichlids 2: Directional evolution of the trade-off between egg number and egg size. Journal

of Evolutionary Biology, 19(1), 76–84. https://doi.org/10.1111/j.1420-9101.2005.00987.x

Kozłowski, J., & Gawelczyk, A. T. (2002). Why are species’ body size distributions usually

skewed to the right? British Ecological Society, 16(4), 419–432.

Kullander, S. O. (1983). A revision of the South American cichlid genus Cichlasoma (Teleostei:

Cichlidae). Swedish Museum of Natural History, Stockholm.

218

Kupren, K., Prusi, M., Krejszeff, S., & Kucharczyk, D. (2014). Early development and

allometric growth in Nannacara anomala Regan , 1905 ( : Cichlidae ) under

laboratory conditions. Neotropical Ichthyology, 12(3), 659–665.

https://doi.org/10.1590/1982-0224-20130104

Lavoué, S., Miya, M., & Nishida, M. (2010). Mitochondrial phylogenomics of anchovies

(family Engraulidae) and recurrent origins of pronounced miniaturization in the order

Clupeiformes. Molecular Phylogenetics and Evolution, 56(1), 480–485.

https://doi.org/10.1016/j.ympev.2009.11.022

Leopold, L. B., Wolman, M. G., & Miller, J. P. (1964). Fluvial processes in geomorphology.

Dover Publications, New York.

Levin, M., Anavy, L., Cole, A. G., Winter, E., Mostov, N., Khair, S., … Yanai, I. (2016). The

mid-developmental transition and the evolution of animal body plans. Nature, 531(7596),

637–641. https://doi.org/10.1038/nature16994

Lindsey, C. C. (1966). Body sizes of poikilotherm vertebrates at different latitudes. Evolution,

20, 456–465. https://doi.org/10.2307/2406584

Lindstedt, S. L., Miller, B. J., & Buskirk, S. W. (1986). Home range, time, and body size in

mammals. Ecology, 67(2), 413–418.

Linke, H., & Staeck, W. (1984). Amerikanische cichliden I. Kleine buntbarsche. Tetra Verlag,

Berlin.

Liu, S., Hui, T. H., Tan, S. L., & Hong, Y. (2012). Chromosome evolution and genome

miniaturization in minifish. PLoS ONE, 7(5), 1–7.

https://doi.org/10.1371/journal.pone.0037305

219

Long, J. A., Trinajstic, K. M., & Johanson, Z. (2009). Devonian arthrodire embryos and the

origin of internal fertilization in vertebrates. Nature, 457(7233), 1124–1127.

https://doi.org/10.1038/nature07732

López-Fernández, H., Arbour, J. H., Winemiller, K. O., & Honeycutt, R. L. (2013). Testing for

ancient adaptive radiations in neotropical cichlid fishes. Evolution, 67(5), 1321–1337.

https://doi.org/10.1111/evo.12038

López-Fernández, H., Arbour, J., Willis, S., Watkins, C., Honeycutt, R. L., & Winemiller, K. O.

(2014). Morphology and efficiency of a specialized foraging behavior, sediment sifting, in

neotropical cichlid fishes. PLoS ONE, 9(3). https://doi.org/10.1371/journal.pone.0089832

López-Fernández, H., Honeycutt, R. L., Stiassny, M. L. J., & Winemiller, K. O. (2005).

Morphology, molecules, and character congruence in the phylogeny of South American

geophagine cichlids (Perciformes, Labroidei). Zoologica Scripta, 34(6), 627–651.

https://doi.org/10.1111/j.1463-6409.2005.00209.x

López-Fernández, H., Winemiller, K. O., & Honeycutt, R. L. (2010). Multilocus phylogeny and

rapid radiations in Neotropical cichlid fishes (Perciformes: Cichlidae: Cichlinae).

Molecular Phylogenetics and Evolution, 55(3), 1070–1086.

https://doi.org/10.1016/j.ympev.2010.02.020

López-Fernández, H., Winemiller, K. O., Montaña, C., & Honeycutt, R. L. (2012). Diet-

morphology correlations in the radiation of south american geophagine cichlids

(perciformes: Cichlidae: Cichlinae). PLoS ONE, 7(4).

https://doi.org/10.1371/journal.pone.0033997

220

Lowe-McConnell, R. H. (1975). Fish communities in tropical freshwaters: Their distribution,

ecology and evolution. Longman Inc., New York.

Lowe-McConnell, R. H. (1991). Ecology of cichlids in South American and African waters,

excluding the African Great Lakes. In: Cichlid Fishes. Behaviour, Ecology and Evolution,

Keenleyside, M. H. A, ed., pp. 62-85. Chapman & Hall, New York.

Lujan, N. L., & Conway, K. W. (2015). Life in the fast lane: a review of rheophily in freshwater

fishes. In: Extremophile Fishes, Riesch, R., Tobler, M., & Plath, M., ed., pp. 107-166.

Spring International Publishing, Switzerland.

Mahler, D. L., Revell, L. J., Glor, R. E., & Losos, J. B. (2010). Ecological opportunity and the

rate of morphological evolution in the diversification of greater Antillean anoles.

Evolution, 64(9), 2731–2745. https://doi.org/10.1111/j.1558-5646.2010.01026.x

Mahon, R. (1984). Divergent Structure in Fish Taxocenes of North Temperate Streams.

Canadian Journal of Fisheries and Aquatic Sciences, 41(2), 330–350.

https://doi.org/10.1139/f84-037

Masters, J. C., Génin, F., Silvestro, D., Lister, A. M., & Delpero, M. (2014). The red island and

the seven dwarfs: Body size reduction in Cheirogaleidae. Journal of Biogeography, 1833–

1847. https://doi.org/10.1111/jbi.12327

Maubecin, C. C., Cosacov, A., Sérsic, A. N., Fornoni, J., & Benitez-Vieyra, S. (2016). Drift

effects on the multivariate floral phenotype of Calceolaria polyrhiza during a post-glacial

expansion in Patagonia. Journal of Evolutionary Biology, 29(8), 1523–1534.

https://doi.org/10.1111/jeb.12889

Maurer, B. A., Brown, J. H., & Rusler, R. D. . (1992). The micro and macro in body size

evolution. Evolution, 46(4), 939–953.

221

McKinney, M. L., & McNamara, K. J. (1991). Heterochrony: The evolution of ontogeny.

Plenum Press, New York.

Mendiola, D., Alvarez, P., Cotano, U., & Martínez De Murguía, A. (2007). Early development

and growth of the laboratory reared north-east Atlantic mackerel Scomber scombrus L.

Journal of Fish Biology, 70(3), 911–933. https://doi.org/10.1111/j.1095-

8649.2007.01355.x

Mérigoux, S., & Ponton, D. (1998). Body shape, diet and ontogenetic diet shifts in young fish of

the Sinnamary River, French Guiana, South America. Journal of Fish Biology, 52, 556–

569. https://doi.org/10.1111/j.1095-8649.1998.tb02017.x

Meyer, A. (1987). Phenotypic plasticity and heterochrony in Cichlasoma managuense (Pisces,

Chichlidae) and their implications for speciation in Cichlid fishes. Evolution, 41(6), 1357–

1369. https://doi.org/10.2307/2409100

Meyer, A. (1990). Morphometrics and allometry in the trophically polymorphic cichlid fish,

Cichlasoma citrinellum: Alternative adaptations and ontogenetic changes in shape. Journal

of Zoology, 221(2), 237–260. https://doi.org/10.1111/j.1469-7998.1990.tb03994.x

Mitteroecker, P., Gunz, P., Bernhard, M., Schaefer, K., & Bookstein, F. L. (2004). Comparison

of cranial ontogenetic trajectories among great apes and humans. Journal of Human

Evolution, 46(6), 679–698. https://doi.org/10.1016/j.jhevol.2004.03.006

Monroe, M. J., & Bokma, F. (2009). Do speciation rates drive rates of body size evolution in

mammals? The American Naturalist, 174(6), 912–918. https://doi.org/10.1086/646606

Montaña, C. G., & Winemiller, K. O. (2009). Comparative feeding ecology and habitats use of

222

Crenicichla species (Perciformes: Cichlidae) in a Venezuelan floodplain river. Neotropical

Ichthyology, 7(2), 267–274. https://doi.org/10.1590/S1679-62252009000200019

Montaña, C. G., & Winemiller, K. O. (2010). Local-scale habitat influences morphological

diversity of species assemblages of cichlid fishes in a tropical floodplain river. Ecology of

Freshwater Fish, 19(2), 216–227. https://doi.org/10.1111/j.1600-0633.2010.00406.x

Muggeo, V. M. R. (2008). segmented: An R package to fit regression models with broken-line

relationships. R News, 8(May), 20–25. https://doi.org/10.1159/000323281

Paine, M. D. (1990). Life-history tactics of darters (Percidae, Etheostomatiini) and their

relationship with body size, reproductive-behavior, latitude and rarity. Journal of Fish

Biology, 37(3), 473–488. https://doi.org/10.1111/j.1095-8649.1990.tb05877.x

Piálek, L., Říčan, O., Casciotta, J., Almirón, A., & Zrzavý, J. (2012). Multilocus phylogeny of

Crenicichla (Teleostei: Cichlidae), with biogeography of the C. lacustris group: Species

flocks as a model for sympatric speciation in rivers. Molecular Phylogenetics and

Evolution, 62(1), 46–61. https://doi.org/10.1016/j.ympev.2011.09.006

Piasecka, B., Lichocki, P., Moretti, S., Bergmann, S., & Robinson-Rechavi, M. (2013). The

hourglass and the early conservation models - Co-existing patterns of developmental

constraints in vertebrates. PLoS Genetics, 9(4).

https://doi.org/10.1371/journal.pgen.1003476

Plaut, I. (2001). Critical swimming speed: its ecophysiological relevance. Comparative

Biochemistry and Physiology Part A: Molecular & Integrative Physiology, 131, 41–50.

https://doi.org/10.1016/S1095-6433(00)80239-1

223

Polo, V., & Carrascal, L. M. (1999). Shaping the body mass distribution of Passeriformes:

habitat use and body mass are evolutionarily and ecologically related. Journal of Animal

Ecology, 68, 324–337.

Ponton, D., & Mérigoux, S. (2000). Comparative morphology and diet of young cichlids in the

dammed Sinnamary river, French Guiana, South America. Journal of Fish Biology, 56, 87–

102. https://doi.org/10.1006/jfbi.1999.1141

Ponton, D., & Mérona, B. (1998). Fish life-history tactics in a neotropical river with a highly

stochastic hydrological regime: the Sinnamary River, French Guiana, South America.

Polskie Archiwum Hydrobiologii, 45(2), 201–224.

Rabosky, D. L., Santini, F., Eastman, J., Smith, S. A, Sidlauskas, B., Chang, J., & Alfaro, M. E.

(2013). Rates of speciation and morphological evolution are correlated across the largest

vertebrate radiation. Nature Communications, 4, 1–8.

https://doi.org/10.1038/ncomms2958

Reis, R. E., Kullander, S. O., & Ferraris, C. J. (2003). Checklist of the freshwater fishes of South

and Central America. EDIPUCRS, Porto Alegre.

R Development Core Team. (2016). R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. ISBM 3-900051-07-0, Available at:

http://www.R-project.org.

Rensch, B. (1948). Histological changes correlated with evolutionary changes of body size.

Evolution, 2(3), 218–230.

Revell, L. J. (2009). Size-correction and principal components for interspecific comparative

studies. Evolution, 63(12), 3258–3268. https://doi.org/10.1111/j.1558-5646.2009.00804.x

224

Revell, L. J. (2012). phytools: An R package for phylogenetic comparative biology (and other

things). Methods in Ecology and Evolution, 3(2), 217–223. https://doi.org/10.1111/j.2041-

210X.2011.00169.x

Roff, D. A. (2002). Life history evolution. Sinauer Associates, Massachusetts.

Rohlf, F. J. (2004). TpsDig. Version 2. Available at: http://life.bio.sunysb.edu/morph/.

Romanuk, T. N., Hayward, A., & Hutchings, J. a. (2011). Trophic level scales positively with

body size in fishes. Global Ecology and Biogeography, 20(2), 231–240.

https://doi.org/10.1111/j.1466-8238.2010.00579.x

Röpke, C. P., Amadio, S., Zuanon, J., Ferreira, E. J. G., Deus, C. P. de, Pires, T. H. S., &

Winemiller, K. O. (2017). Simultaneous abrupt shifts in hydrology and fish assemblage

structure in a floodplain lake in the central Amazon. Scientific Reports, 7, 1–10.

https://doi.org/10.1038/srep40170

Rosa, R. S., Charvet-Almeida, P., & Quijada, C. C. D. (2010). Biology of the South American

potamotrygonid stingrays. In: Sharks and their relatives II: biodiversity, adaptive

physiology, and conservation, Carrier, J. F., Musick, J. A, & Heithaus, M. R., ed., pp. 241-

285. CRC Press, Boca Raton.

Ross, S. T. (2013). Ecology of North American freshwater fishes. University of California Press,

Berkley, Los Angeles, Oakland.

Salzburger, W. (2009). The interaction of sexually and naturally selected traits in the adaptive

radiations of cichlid fishes. Molecular Ecology, 18(2), 169–185.

https://doi.org/10.1111/j.1365-294X.2008.03981.x

225

Salzburger, W., & Meyer, A. (2004). The species flocks of East African cichlid fishes: recent

advances in molecular phylogenetics and population genetics. Die Naturwissenschaften,

91(6), 277–290. https://doi.org/10.1007/s00114-004-0528-6

Santini, F., Nguyen, M. T. T., Sorenson, L., Waltzek, T. B., Lynch Alfaro, J. W., Eastman, J.

M., & Alfaro, M. E. (2013). Do habitat shifts drive diversification in teleost fishes? An

example from the pufferfishes (Tetraodontidae). Journal of Evolutionary Biology, 26(5),

1003–1018. https://doi.org/10.1111/jeb.12112

Santos-Santos, J. H., Audenaert, L., Verheyen, E., & Adriaens, D. (2015). Divergent ontogenies

of trophic morphology in two closely related haplochromine cichlids. Journal of

Morphology, 0, n/a-n/a. https://doi.org/10.1002/jmor.20385

Schütz, D., & Taborsky, M. (2000). Giant males or dwarf females: what determines the extreme

sexual size dimorphism in callipterus? Journal of Fish Biology, 57(5), 1254–

1265. https://doi.org/10.1006/jfbi.2000.1388

Seehausen, O. (2006). African cichlid fish: a model system in adaptive radiation research.

Proceedings of the Royal Society B: Biological Sciences, 273(1597), 1987–1998.

https://doi.org/10.1098/rspb.2006.3539

Shiklomanov, I. (1993). World fresh water resources. In: Water in crisis: a guide to the world’s

fresh water resources, Gleick, P.H., ed., Oxford University Press, New York.

Sims, D. W. (2003). Tractable models for testing theories about natural strategies: foraging

behavious and habitat selection of free ranging sharks. Journal of Fish Biology, 63, 53–73.

https://doi.org/10.1046/j.1095-8649.2003.00207.x

226

Slater, G. J., Price, S. A., Santini, F., & Alfaro, M. E. (2010). Diversity versus disparity and the

radiation of modern cetaceans. Proceedings of the Royal Society B: Biological Sciences,

277(1697), 3097–3104. https://doi.org/10.1098/rspb.2010.0408

Smith, F. A, Brown, J. H., Haskell, J. P., Lyons, S. K., Alroy, J., Charnov, E. L., Dayan, T.,

Enquist, B. J., Ernest, S. K. M., Hadly, E. A., Jones, K. E., Kaufman, D. M., Marquet, P.

A., Maurer, B. A., Niklas, K. J., Porter, W. P., Tiffney, B., & Willig, M. R. (2004).

Similarity of mammalian body size across the taxonomic hierarchy and across space and

time. The American Naturalist, 163(5), 672–691. https://doi.org/10.1086/382898

Smith, F. A, & Lyons, S. K. (2011). How big should a mammal be? A macroecological look at

mammalian body size over space and time. Philosophical Transactions of the Royal

Society of London. Series B, Biological Sciences, 366(1576), 2364–2378.

https://doi.org/10.1098/rstb.2011.0067

Soria-Barreto, M., & Rodiles-Hernández, R. (2008). Spatial distribution of cichlids in Tzendales

River, Biosphere Reserve Montes Azules, Chiapas, Mexico. Environmental Biology of

Fishes, 83(4), 459–469. https://doi.org/10.1007/s10641-008-9368-0

Stanley, S. M. (1973). An explanation for Cope’s Rule. Evolution; International Journal of

Organic Evolution, 27(1), 1–26.

Stiassny, M. L. J. (1991). Phylogenetic intrarelationships of the family Cichlidae: an overview.

In: Cichlid Fishes. Behaviour, Ecology and Evolution, Keenleyside, M. H. A, ed., pp. 1–

35. Chapman Hall, London.

Steele, S. E., & López-Fernández, H. (2014). Body size diversity and frequency distributions of

neotropical cichlid fishes (Cichliformes: Cichlidae: Cichlinae). PLoS ONE, 9(9).

227

https://doi.org/10.1371/journal.pone.0106336

Thomas, G. H., Meiri, S., & Phillimore, A. B. (2009). Body size diversification in anolis: Novel

environment and island effects. Evolution, 63(8), 2017–2030.

https://doi.org/10.1111/j.1558-5646.2009.00694.x

Toussaint, A., Charpin, N., Brosse, S., & Villeger, S. (2016). Global functional diversity of

freshwater fish is concentrated in the Neotropics while functional vulnerability is

widespread. Scientific Reports, 6(October 2015), 1–9. https://doi.org/10.1038/srep22125

Trinajstic, K., Boisvert, C., Long, J., Maksimenko, A., & Johanson, Z. (2015). Pelvic and

reproductive structures in placoderms (stem gnathostomes). Biological Reviews, 90(2),

467–501. https://doi.org/10.1111/brv.12118

Turner, T. F., & Trexler, J. C. (1998). Ecological and historical associations of gene flow in

darters (Teleostei: Percidae). Evolution, 52(6), 1781–1801.

https://doi.org/10.2307/2411349

Urošević, A., Ljubisavljević, K., & Ivanović, A. (2013). Patterns of cranial ontogeny in lacertid

lizards: Morphological and allometric disparity. Journal of Evolutionary Biology, 26(2),

399–415. https://doi.org/10.1111/jeb.12059

Vega, G. C., & Wiens, J. J. (2012). Why are there so few fish in the sea? Proceedings of the

Royal Society B: Biological Sciences, 279(1737), 2323–2329.

https://doi.org/10.1098/rspb.2012.0075

Vélez-Zuazo, X., & Agnarsson, I. (2011). Shark tales: A molecular species-level phylogeny of

sharks (Selachimorpha, Chondrichthyes). Molecular Phylogenetics and Evolution, 58(2),

228

207–217. https://doi.org/10.1016/j.ympev.2010.11.018

Wainwright, P. C., Alfaro, M. E., Bolnick, D. I., & Hulsey, C. D. (2005). Many-to-One

Mapping of Form to Function: A General Principle in Organismal Design? Integrative and

Comparative Biology, 45(2), 256–262. https://doi.org/10.1093/icb/45.2.256

Wainwright, P. C., & Richard, B. A. (1995). Predicting patterns of prey use from morphology of

fishes. Environmental Biology of Fishes, 44, 97–113. https://doi.org/10.1007/BF00005909

Webb, J. F. (1990). Ontogeny and phylogeny of the trunk lateral line system in cichlid fishes.

Journal of Zoology, London, 221, 405–418.

Weitzman, S. H., & Vari, R. P. (1988). Miniaturization in South American Freshwater Fishes;

An overview and discussion. Proceedings of the Biological Society of Washington, 101(2),

444–465.

West, G. B., Brown, J. H., & Enquist, B. J. (1997). A general model for the origin of allometric

scaling laws in biology. Science, 276(5309), 122–126.

West, G. B., Brown, J. H., & Enquist, B. J. (2001). A general model for ontogenetic growth.

Nature, 413(6856), 628–31. https://doi.org/10.1038/35098076

Wilson, D. S. (1975). The adequacy of body size as a niche difference. The American

Naturalist, 109(970), 769–784. https://doi.org/10.1086/283042

Wimberger, P. H. (1991). Plasticity of jaw and skull morphology in the Neotropical cichlids

Geophagus brasiliensis and G. steindachneri. Evolution, 45(7), 1523–1536.

https://doi.org/10.1554/06-162.1

229

Wimberger, P. H., Reis, R. E., & Thornton, K. R. (1998). Mitochondrial phylogenetics,

biogeography, and evolution of parental care and mating systems in Gymnogeophagus

(Perciformes:Cichlidae). In: Phylogeny and Classification of Neotropical Fishes,

Malabarba, L. R., Reis, R. E., Vari, R. P., Lucena, Z. M., & Lucena, C. A. S., ed., pp. 309-

318. EDIPUCRS, Porto Alegre.

Winemiller, K. O. (1989). Ontogenetic diet shifts and resource partitioning among piscivorous

fishes in the Venezuelan ilanos. Environmental Biology of Fishes, 26(3), 177–199.

https://doi.org/10.1007/BF00004815

Winemiller, K. O., Kelso-Winemiller, L. C., & Brenkert, A. L. (1995). Ecomorphological

diversification and convergence in fluvial cichlid fishes. Environmental Biology of Fishes,

44, 235–261. https://doi.org/10.1007/BF00005919

Winemiller, K. O., & Rose, K. A. (1992). Patterns of life-history diversification in North

American fishes: Implications for population regulation. Canadian Journal of Fisheries

and Aquatic Sciences, 49, 2196–2218. https://doi.org/10.1139/f92-242

Woodward, G., & Hildrew, A. G. (2002). Body-size determinants of niche overlap and

intraguild predation within a complex food web. Journal of Animal Ecology, 71(6), 1063–

1074. https://doi.org/10.1046/j.1365-2656.2002.00669.x

Zelditch, M. L., Calamari, Z. T., & Swiderski, D. L. (2016). Disparate postnatal ontogenies do

not add to the shape disparity of infants. Evolutionary Biology, 43(2), 1–20.

https://doi.org/10.1007/s11692-016-9370-y

Zelditch, M. L., Sheets, H. D., & Fink, W. L. (2003). The ontogenetic dynamics of shape

disparity. Paleobiology, 29(1), 139–156. https://doi.org/10.1666/0094-

230

8373(2003)029<0139:TODOSD>2.0.CO;2

Zelditch, M. L., Swiderski, D. L., Sheets, H. D., & Fink, W. L. (2004). Geometric

morphometrics for biologists: a primer. Elsevier Academic Press, Boston.

Zimkus, B. M., Lawson, L., Loader, S. P., & Hanken, J. (2012). Terrestrialization,

miniaturization and rates of diversification in african puddle frogs (anura:

Phrynobatrachidae). PLoS ONE, 7(4), 15–21. https://doi.org/10.1371/journal.pone.0035118