Tempo and Mode of Lineage and Morphological Diversification in a Hyperdiverse Freshwater Fish Radiation (Teleostei: Characiformes)

AN ABSTRACT OF THE DISSERTATION OF

Michael D. Burns for the degree of Doctor of Philosophy in Fisheries Science presented on April 17, 2018.

Title: Tempo and Mode of Lineage and Morphological Diversification in a Hyperdiverse Freshwater Fish Radiation (Teleostei: Characiformes)

Abstract approved: ______Brian L. Sidlauskas

Characiform fishes form one of the most diverse freshwater fish clades in the world. Comprising more than 2000 species and distributed primarily in South

America and Africa, characiforms vary dramatically in their ecomorphology.

However, the evolutionary processes responsible for the immense ecomorphological diversity remains unknown. Recently, a study postulated that the unparalleled ecomorphological diversification arose through an ancient adaptive radiation, as evidenced by the clear segregation of morphological traits, such as body shape, among different trophic and habitat groups. However, no formal macroevolutionary analyses have been conducted on the entire order of

Characiformes and the mechanism responsible for the diversity remains unknown.

Here, I conduct a macroevolutionary analysis of body shape diversification to determine if Characiformes evolved through an adaptive radiation. I estimated the first time-calibrated molecular phylogeny for the order Characiformes, assembled the first ever geometric morphometric body shape dataset, and compiled an exhaustive trophic ecology database.

In my second chapter, I combined these datasets to test whether body shape adapted to shifts into different trophic guilds and to reconstruct body shape diversification in the Old and New World radiations. I found that body shape adaptation resulted in many non-repeated morphologies; lineages that shifted into the same trophic ecology evolved different morphotypes, except for convergent piscivores. Furthermore, we found that body shape diversification between the Old and New World radiations followed very different pathways, with the New World radiation occupying twice as much morphospace as their Old World counterparts.

Both radiations exhibited higher morphological disparity than would be expected under Brownian motion, early in cladogenesis, matching expectations of an adaptive radiation.

In my third chapter, I tested whether evolutionary modularity increased body shape diversification in the order. I found that characiform body shape was comprised of three independent modules that diversified at different times and rates while under different selective regimes. I postulate that the high evolutionary modularity plausibly explains why many body shapes evolved only once, when lineages evolved similar trophic ecologies across the radiation. The generality of the relationship between evolutionary modularity and increased morphological disparity has not been well studied in vertebrate lineages. More studies need to look at the role that evolutionary modularity and integration can have in shaping deterministic and contingent patterns of evolution across broad and restricted radiations of vertebrates.

In my fourth chapter, I analyzed the rates of lineage and morphological diversification to determine whether Characiformes exhibited an early burst of speciation and morphological evolution as predicted by classical adaptive radiations. I found that the rate of speciation and evolution on the first morphological principal component were very high early in cladogenesis and quickly slowed down, following a pattern of adaptive radiation. However, the evolutionary simulations indicate that heavily pruning the tree overestimates the speciation rate early in cladogenesis, making the results consistent with a constant rates model. Higher taxon sampling is needed to fully understand whether the order exhibited the speciation patterns consistent with an adaptive radiation.

My dissertation presents the first ever time-calibrated molecular phylogeny of Characiformes, the largest geometric morphometric dataset of Characiformes, and the most densely sampled clade-wide analysis of modularity in fishes. I found that characiform body shape likely radiated early in cladogenesis giving rise to many of the distinct morphologies that define the order.

Tempo and Mode of Lineage and Morphological Diversification in a Hyperdiverse Freshwater Fish Radiation (Teleostei: Characiformes)

By Michael D. Burns

A DISSERTATION

submitted to

Oregon State University

in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Presented April 17, 2018 Commencement June 2018

Doctor of Philosophy dissertation of Michael D. Burns presented on April 17, 2018

APPROVED:

Major Professor, representing Fisheries Science

Head of the Department of Fisheries and Wildlife

Dean of the Graduate School

I understand that my dissertation will become part of the permanent collection of Oregon State University libraries. My signature below authorizes release of my dissertation to any reader upon request.

Michael D. Burns, Author

ACKNOWLEDGEMENTS

This dissertation would not have been possible without the help of many different people. First, I would like to thank my M.S. advisor, Kathleen Cole, for convincing me that Brian Sidlauskas would be an amazing PhD advisor. I want to thank Brian Sidlauskas for being everything Kassi said and more. Brian’s guidance has been instrumental in me becoming the researcher that I am today. His mentoring has shaped the way that I think, write, and teach; I am forever grateful for what he has helped me achieve. I also want to thank Brian for taking me to various meetings near and far and always allowing me to work on independent projects because the side hussle is my scientific muse. I want to thank the late Rich Vari for teaching a marine ichthyologist (me) all about tropical freshwater fishes. His patience and kindness were instrumental to the early success of my dissertation work. More importantly, I want to thank Rich for teaching me how to be a professional scientist by balancing work and life. Rich was a total scientific rockstar, but carried himself with extreme kindness, humbleness, and grace. His compassion was truly inspirational. Thank you for being you, Rich. The entire community misses you very much. I want to thank my committee for always having great input and helping me see the forest for the phylogenetic trees. I want to thank Vicki Tolar-Burton for helping me to love writing again. I lost it for awhile, but Vicki’s kindess and acceptance reignited my passion for written word. Thank you to all of my labmates, past and present, including Ben Frable, Kendra Hoekzema, Whitcomb Bronaugh,

Casey Dillman, and Thaddaeus Buser. Kendra, thank you for teaching me how to

be a disciplined reaearcher. Whit, thank you for being my sports outlet in the lab.

Ben and Thaddaeus, thank you for always bringing the laughs to lab meeting and scientific conferences. Much of my scientific happiness has been forged in our late night chats at different meetings. I want to thank Peter Konstantinidis for super helpful discussions on fish morphology and evolution. Peter being hired by OSU was life changing for me. I want to thank Devin Bloom for being a secret (or not so secret?) role model of mine by showing me that someone like me can be successful in academia.

I want to thank my parents for always believing in me, pushing me to expand my horizons, and always forcing me to find my own answers to the questions I asked. I believe that my mom and dad must have learned parenting under the late Socrates. They employed the Socratic Method of debate at every family dinner, and thus dinner served as my first committee meetings. I could not be more grateful for this because my scientific iron was forged in the fire of grilled burgers. Mom, I want to thank you for teaching me to be tough, you are the strongest person I have ever met and much of what I learned from you helped keep me sane during my PhD. Dad, I want to thank you for being my hero. Words cannot describe how much your kindness and love has shaped the person I am today. I hope in some small way my PhD fulfills your dream of being a scientist because I would not be one today without your endless passion for biology (even though you like birds more than fish now). I want to thank my brothers, Dan and Andrew, for always inspiring me. Dan, Andrew, from the day I was born, I have been chasing each of you, either on bikes or intelluctually. I might never catch you, but you will

always help me reach higher than I ever could have on my own. I want to thank the

Fregosi family for adopting me as their own during my PhD. Their love and acceptance lessened the blow of being away from my own family. I want to thank

Selene for being everything. I really mean everything. Selene you are my emotional support, travel buddy, confidant, copy editor, sounding board, and best friend. None of this would have been possible without you helping me at every step. Thank you.

I want to thank two amazing doggos, Piper “Fox” Fregosi and Murphy Jameson

Burns. You borked, barked, and whined me into happiness everytime I was sad.

Lastly, but definitely not leastly, I want to thank Brooks Burr, my undergraduate advisor. Brooks you took a chance on a 19-year-old who just loved sun, fishing and tropical fish aquaria. You taught me about science. You taught me about evolution. You taught me about ichthyology. You taught me all of this while never making me feel dumb, belittled, or beneath you. Your stories of growing up in the early 60’s in California, surfing breaks on a custom Dale Vezy longboard, made me realize that a scientist could be a regular person. You made me realize I could be a scientist. You were everything I needed in an advisor at that age and much of the passion you instilled in me then has kept me going now. I dedicate my dissertation to you Brooks because you saw my PhD as a reality while I was still a

19-year-old catching bass in Thompson Lake, dreaming I was on a tropical beach.

Thank you Brooks, I finally did it!

TABLE OF CONTENTS

Page

CHAPTER 1: GENERAL INTRODUCTION ...... 1

The history of biological classification ...... 1

Evolutionary processes that shape diversity ...... 2

Adaptive radiations ...... 3

Study group: Characiformes ...... 6

Objectives ...... 8

Figure legends ...... 11

CHAPTER 2: EXPLOSIVE AND CONTINGENT BODY SHAPE

DIVERSIFICATION IN A HYPERDIVERSE FRESHWATER FISH

RADIATION (TELEOSTEI: CHARACIFORMES) ...... 12

Abstract ...... 13

Introduction ...... 14

Materials and methods ...... 19

Results ...... 25

Discussion ...... 32

Acknowledgements ...... 43

Figure legends ...... 45

Tables ...... 57

CHAPTER 3: EVOLUTIONARY MODULARITY PROMOTED NOVEL

BODY SHAPES IN A HYPERDIVERSE FRESHWATER FISH RADIATION 59

Abstract ...... 60

TABLE OF CONTENTS (Continued)

Page

Introduction ...... 61

Materials and methods ...... 65

Results ...... 70

Discussion ...... 74

Acknowledgements ...... 80

Figure legends ...... 82

Tables ...... 88

CHAPTER 4: TEMPO OF LINEAGE AND BOD SHAPE DIVERSIFICATION

IN A HYPERDIVERSE FRESHWATER FISH RADIATION ...... 90

Abstract ...... 91

Introduction ...... 92

Materials and methods ...... 95

Results ...... 99

Discussion ...... 103

Acknowledgements ...... 108

Figure legends ...... 109

Tables ...... 116

CHAPTER 5: DISCUSSION ...... 117

Overview ...... 117

Future Directions ...... 118

DISSERTATION REFERENCES ...... 126

TABLE OF CONTENTS (Continued)

Page

APPENDICES ...... 161

Appendix 1: Supplementary materials for Chapter 2 ...... 161

Appendix 2: Supplemental Figures ...... 171

Appendix 3: Supplemental Tables ...... 176

Appendix 4: Materials Examined ...... 190

LIST OF FIGURES

Figure Page

Figure 1.1. Patterns of phenotypic diversification...... 11

Figure 2.1. Landmarks used in geometric morphometric analysis ...... 45

Figure 2.2. Multivariate OU hypotheses for the evolution of body shape ...... 46

Figure 2.3. Bayesian phylogenetic tree ...... 48

Figure 2.4. Trophic ecology evolutionary history ...... 49

Figure 2.5. Deformation grids ...... 50

Figure 2.6. Morphospace ...... 51

Figure 2.7. Disparity through time ...... 53

Figure 2.8. Body shape adaptive landscape ...... 54

Figure 2.9. DTT plot between Old and New World ...... 55

Figure 2.10. Phylomorphospace ...... 56

Figure 3.1. Landmark schematic of modules...... 82

Figure 3.2. Histogram of modularity ...... 84

Figure 3.3. Phylomorphospace ...... 85

Figure 3.4. Multivariate Disparity ...... 86

Figure 3.5.Schematic of modularity increasing diversity ...... 87

Figure 4.1. Relative rates of speciation or morphological evolution ...... 110

Figure 4.2. Instantaneous rates of speciation and morphological evolution ...... 111

Figure 4.3. PC2 rate evlution ...... 112

Figure 4.4. PC3 rate evolution ...... 113

Figure 4.5. Speciation dynamics for simulated phylogenetic trees ...... 114

LIST OF FIGURES (Continued)

Figure Page

Figure 4.6. Morphological evolution dynamics simulated phylogenetic trees ....115

Figure 7.1. Phylogenetic hypotheses for characiform lineages ...... 171

Figure 7.2. RAxML phylogenetic tree ...... 172

Figure 7.3. Results of BayesTraits analysis ...... 173

Figure 7.4. Histogram of AICc scores ...... 174

Figure 7.5. SURFACE analysis on Oliveira et al. 2011 tree ...... 175

LIST OF TABLES

Table Page

Table 2.1. Results of the multivariate model-fitting analyses for body shape...... 57

Table 2.2. Results from the convevol analysis ...... 58

Table 3.1. Results of the multivariate model-fitting analyses for each module ...88

Table 3.2. Rates of module evolution ...... 89

Table 4.1. Results of the Spearman’s rank correlation test...... 116

Table 7.1. GenBank numbers for tissues used in phylogenetic analysis ...... 176

Table 7.2. Trophic ecology classification ...... 180

Table 7.3. Priors and parameters used in BEAST 2.0 ...... 185

Table 7.4. Gene partitions and their models as selected by PartitionFinder 2.0. .186

Table 7.5. Results of the multivariate model-fitting analyses ...... 187

Table 7.6. Average results from the convevol analysis ...... 188

Table 7.7. Results from the convevol analysis ...... 189

Tempo and Mode of Lineage and Morphological Diversification in a Hyperdiverse Freshwater Fish Radiation (Teleostei: Characiformes)

CHAPTER 1: GENERAL INTRODUCTION

The history of biological classification

Humankind has long been fascinated with understanding and classifying the natural world. One of the earliest records of human settlement, the paintings of Lascaux

Cave in France, depicted large megafauna of the Late Pleistocene (Le Quellec 2010).

Overtime this fascination led to systems for classifying and organizing life’s relationships. Aristotle, the ancient Greek philosopher, is often credited with being the first scientist to systematically study biology (Gould 2002). Aristotle created an animal hierarchy (Latin: scala naturae, "ladder of being"), that separated organisms with blood

(vertebrates) from animals without blood (invertebrates) in his text Historia animalium.

The scala naturae became the foundation upon which Carl Linneaus, the founder of modern taxonomy, based his Systema Naturæ, where he divided the physical components of the world into the three kingdoms of minerals, plants and animals.

However, all of these classifications thought of animals in a fixed order without transmutation or progression throughout time.

During the mid to late 18th century biologists begun to think of organisms as units that changed over time, instead of being fixed. Jean Baptiste Lamarck, a French naturalist, was the first scientist to formulate a coherent evolutionary theory with his belief in the inheritance of acquired characteristics (Gould 2002). Lamarck proposed that animals changed aspects of their body over their lifetime in response to adaptation to their environment which was passed down to their offspring (Gould 2002). During

2 the mid-19th century, two naturalists, Charles Darwin and Alfred Wallace, both presented papers to the Linnean Society outlining the theory of evolution by natural selection; papers that would shape the field of biology in modern times (Gould 2002).

The theory of evolution provides the foundation for which modern biologists try to understand the evolutionary processes that give rise to life’s diversity.

Evolutionary processes that shape diversity

Evolutionary biology aims to determine the processes that shape species richness and phenotypic diversity across the tree of life. Not all clades diversify equally; some groups evolve remarkably many species and morphologically disparate forms, while others contain few species, all of a similar morphology. Numerous factors could cause a clade to accumulate high morphological disparity, including older clade age (Foote 1997; Collar et al. 2005; Ricklefs 2006; Sidlauskas 2008), a steeper underlying ecological gradient (Muschick et al. 2014; Burress et al. 2016; Cooney et al. 2017), the presence of a functional innovation (Price et al. 2010; Martin and

Wainwright 2011), or a lack of selective constraints (Hansen 1997; Collar et al. 2009).

Though many evolutionary scenarios can catalyze the evolution of high ecomorphological diversity, they vary significantly in the underlying pattern of diversification. Modern phylogenetic comparative methods enable inference of the evolutionary processes responsible for modern diversity by reconstructing and comparing patterns of diversification and disparity across lineages.

Morphological diversity can arise through multiple different processes over evolutionary time. The simplest scenario, stochastic evolution, results in the gradual

3 accumulation of morphological diversity. During stochastic evolution, morphological diversity evolves randomly across lineages with no clear pattern in the distribution of phenotypes (Figure 1.1) Thus, the differences in morphological diversity between lineages arises as a result of variation in clade age and not through adaptation to any underlying gradient. Yet, many of the most dramatic episodes of diversification in the history of life resulted non-randomly from clade-specific events that changed the underlying process of evolution and adaptation (Price et al. 2010; Burbrink et al. 2012;

Arbour and López-Fernández 2016). For example, a clade may experience multiple adaptive bursts linked to repeated environmental changes (Figure 1.1), in which case diversity will continue to increase throughout cladogenesis (Schenk et al. 2013; Stubbs and Benton 2016). Alternatively, a lineage may diversify explosively following the invasion of novel adaptive zones after colonizing new habitats (Lovette et al. 2002;

Muschick et al. 2012; Hulsey et al. 2013) or after evolving a key, novel phenotype that allows it to exploit new trophic resources (Rüber et al. 1999; Grant and Grant 2006;

Cooney et al. 2017). In a classic adaptive radiation scenario, the key innovation causes a singular early burst of phenotypic disparity that declines over time as niche space fills

(Simpson 1953; Losos et al. 1998; Madsen et al. 2001; Harmon et al. 2003).

Adaptive radiations

Adaptive radiations are spectacular displays of ecological and evolutionary diversity wherein large amounts of lineage, morphological and ecological diversification evolve in a short amount time (Simpson 1953; Schluter 2000; Glor

2010). Much of life’s most remarkable radiations, including Crater Lake cichlids

(Kocher et al. 1993; Muschick et al. 2012; Machado-Schiaffino et al. 2015), Anolis

4 lizards (Losos et al. 1998; Harmon et al. 2005; Mahler et al. 2013), and Hawaiian honeycreepers (Lerner et al. 2011), have arisen through evolutionary dynamics typical of adaptive radiations. Classically, adaptive radiations have occurred in unoccupied

“island” ecosystems (Schluter 2000), such as sticklebacks in postglacial lakes, in which a single founder diversified rapidly across the benthic to pelagic habitat axis (Schluter

1995; Rundle et al. 2000). More recently, many widespread continental radiations, such

Neotropical cichlids (López‐Fernández et al. 2013) and New World ratsnakes

(Burbrink et al. 2012), have been shown to have diversified through adaptive radiations.

The majority of studies on adaptive radiation have focused on insular “island” radiations. For example, Anolis lizards have diversified on different islands of the

Greater Antilles (Losos 2009), converging on morphotypes and exhibiting a slowdown in diversification as ecological opportunity is reduced (Harmon et al. 2005; Mahler et al. 2010; Mahler et al. 2013). Likewise, cichlids have invaded and adaptively radiated in multiple crater lakes across Africa and Central America (Elmer et al. 2014;

Seehausen 2015). Each invasion follows predictable patterns including body shape diversification across the benthic-pelagic habitat axis (Hulsey et al. 2013), trophic diversification between soft and hard-bodied prey (Kocher et al. 1993; Hulsey et al.

2006), and density dependent diversification (Machado-Schiaffino et al. 2015).

However, the complexity of continental radiations, including richer biotas and more complex ecosystem interactions, may present different patterns of adaptive radiation than those presented in the classical insular systems (Schluter 2000; Claramunt 2010;

Harmon et al. 2010).

Many studies have shown that broad scale continental radiations can exhibit signatures of classical adaptive radiations including early bursts of lineage and morphological diversification. However, the patterns of adaptive radiations in widespread continental radiations are less clear than what is shown from insular

“island” radiations, with the relationship between ecological opportunity and adaptive divergence much more varied (Maestri et al. 2017). For example, patterns of species and morphological diversification were found to be decoupled in New World ratsnakes, which led to clades with high species diversity, but limited ecological variability

(Burbrink et al. 2012). Furthermore, in geographically expansive radiations, independent lineages may exhibit different amounts of ecological opportunity across different spatial scales making patterns vary between lineages. For instance, in

Neotropical cichlids, South American lineages exhibited an initial spike in disparity that quickly declined, indicative of an adaptive radiation, while Central American lineages continued to diversify throughout cladogenesis (Arbour and López-Fernández

2016).

In radiations that span multiple spatial scales, some lineages may substantial ecological opportunity, while others have reduced ecological opportunity because of different community assemblies. For instance, geophagine cichlids, which occupy an unusual benthivorous niche, evolved a riot of morphological diversity in South

America, while the closely related Hereoini were restricted to the extremes of morphospace (López‐Fernández et al. 2013). When hereoine cichlids invaded Central

America, where geophagines are absent, the lineage diversified into a wider morphospace than it occupied in South America, indicating that competitive exclusion

6 by geophagines restricted the morphospace of South American Hereoine cichlids

(López‐Fernández et al. 2013).

Diversification patterns do not always differ across continental scales; not all widespread radiations deviate from classical adaptive radiations. A recent study on pike cichlids found that independent lineages exhibited parallel adaptive radiations across

South American rivers matching the typical pattern of insular radiations (Burress et al.

2017). Specifically, pike cichlids independently invaded the Uruguay and Parana

Rivers, exhibiting a burst of phenotypic evolution after colonization, diversifying into novel ecomorphs including crevice feeders, periphyton grazers, and molluscivores in both radiations. The Burress et al. (2017) result clearly shows that the emblematic characteristics of adaptive radiations can occur in labile ecosystems, thus increasing the importance of studying adaptive radiations in widespread, hyperdiverse radiations.

Study group: Characiformes

The order Characiformes is a hyderdiverse freshwater fish radiation comprised of roughly 2000 described species, divided amongst 21 families. The order exhibits remarkable ecomorphological diversity including substantial variation in body shape, jaw, tooth, and head morphology that likely evolved in response to multiple shifts across habitat and feeding ecology. Most morphological studies have focused on the diversification of smaller subclades (Buckup 1993a; Sidlauskas 2008) or address broad scale systematic revisions and phylogenetic analyses (Fink 1996; Buckup 1998;

Malabarba and Weitzman 2003; Mirande 2009; Mirande 2010), rather than explicitly

7 studying morphological diversity. Previous morphological studies have only hinted at the evolutionary processes that have shaped this large radiation of fishes.

The order has had a rich history of phylogenetic studies, which have incorporated molecular (Ortí and Meyer 1997; Calcagnotto et al. 2005; Javonillo et al.

2010; Oliveira et al. 2011; Mariguela et al. 2013a; Arcila et al. 2017) and morphological data (Vari 1979; Fink and Fink 1981; Vari 1983; Weitzman and Fink 1983; Vari et al.

1995; Sidlauskas 2008; Sidlauskas and Vari 2008; Dillman et al. 2016). Each of these studies has worked towards the description of monophyletic groups within the order, with most families exhibiting monophyly (Weitzman 1954; Roberts 1973; Vari 1979,

1983, 1989; Vari et al. 1995; Buckup 1998; Zanata and Vari 2005), except for the speciose Characidae (Mirande 2009; Javonillo et al. 2010; Oliveira et al. 2011).

However, a time and fossil-calibrated phylogeny for the entire order has never been reconstructed, making inferences on the tempo and mode of morphological diversification impossible.

8 and trophic ecology have shaped aspects of morphological diversity in the order through an adaptive radiation, but without direct tests of adaptive evolution or comparison of alternative modes of diversification, the exact evolutionary mechanism and extent to which ecology influences morphology in Characiformes remain unknown.

Objectives

This dissertation aimed to determine the evolutionary processess that gave rise to the unparalleled ecomorphological diversification in the freshwater fish order

Characiformes. Specifically, I wanted to test whether the order diversified through the classical pattern of an adaptive radiation as postulated by Guisande et al. (2012), by analyzing the tempo and mode of lineage and morphological diversification. However, over the course of working through my dissertation, this project shifted from simply asking whether Characiformes was an adaptive radiation, into also asking why some radiations follow predictable patterns of phenotypic diversification while others do not.

In the second chapter, I examine the evolution of characiform body shape in the context of shifts among different trophic ecologies and against the backdrop of independent evolution on two continents to determine the extent to which major dietary shifts have influenced body shape diversification across one of the largest radiations of fishes. I infer a large, robust molecular phylogeny by synthesizing data from previous studies and link it to a novel body shape dataset. I then use phylogenetic comparative methods to address three primary questions about body shape evolution in the order:

(1) Did body shape morphology diversify through stochastic or adaptive evolution? (2)

Did body shape diversify continuously, in repeated bursts of adaptive evolution, or in the early burst that typifies adaptive radiation? (3) Did body shape diversify contingently or deterministically, as would be indicated by dissimilar or similar evolutionary patterns in Africa and the Neotropics and in response to repeated shifts into the same trophic ecology?

In the third chapter, I examine whether evolutionary modularity increased the body shape diversification in Characiformes when invading the same non-piscivorous trophic niche by increasing the number of available morphotypes. Here I ask whether the potential division of the characiform body plan into multiple modules can help explain their diverse morphologies and apparently highly contingent evolutionary history. Specifically, I want to determine (1) whether multiple body shape modules exist in Characiformes, (2) whether those potential modules evolve under different models and at different rates and (3) whether convergent lineages have fewer body shape modules than non-convergent lineages. Answers to these questions will help reveal whether differences in modularity may explain why some characiform lineages converge, while others do not.

In the fourth chapter, we use a fossil calibrated molecular phylogeny and geometric morphometric dataset to assess the rates of lineage and morphological diversification in Characiformes. Specifically, we use a heterogeneous mixture of diversity-dependent and constant rate diversification regimes to test whether (1)

Characiformes show evidence of early bursts of lineage and morphological diversification followed by a decrease in rates of divergence, and (2) whether morphological and lineage diversification rates correlate with one another. However,

10 the phylogenetic tree used in this study contains less than ~150 species, so we also performed a suite of simulations to determine if our small empirical tree had enough data to accurately recover the true diversification rates.

Overall, this dissertation advances our understanding of how vertebrate life evolves on this planet. Specifically, I show how adaptation and evolutionary modularity may have promoted one of the most ecomorphological diverse vertebrate radiations on this planet by changing the tempo and mode of lineage and morphological diversification throughout cladogenesis.

Figure legends

Figure 1.1. Different patterns of phenotypic diversification simulated in phytools

(Revell 2012).

CHAPTER 2: EXPLOSIVE AND CONTINGENT BODY SHAPE

DIVERSIFICATION IN A HYPERDIVERSE FRESHWATER FISH RADIATION

(TELEOSTEI: CHARACIFORMES)

Michael D. Burns

Intended for publication in Evolution with the following co-author:

Brian L. Sidlauskas

Abstract

The speciose characiform fishes have radiated remarkably in ecomorphology, but the macroevolutionary processes responsible for their biodiversity remain largely unexplored. We reconstruct their diversification using a new fossil-calibrated molecular phylogeny, trophic ecology, and geometric morphometrics. Though body shape diversified in a manner consistent with adaptive radiation, with disparity peaking early in cladogenesis, shifts in trophic ecology did not always coincide with shape diversification. With the notable exception of piscivores, lineages that converged in diet did not converge closely in body shape, and an Ornstein-Uhlenbeck model postulating a single distinct adaptive peak for each trophic ecology fits poorly. Shifts in habitat or other variables likely influenced body shape evolution more than changes in diet, and the clade’s history departs from many classic adaptive radiations, in which trophic convergence drives morphological convergence deterministically. The contrast between the South American radiation’s exhaustive exploration of morphospace and the parallel African radiation’s restrained diversification further indicates major role for contingency in characiform evolution, with the presence of cypriniform competitors in the Old World, but not the New, providing one possible explanation. Our results depict the clearest ecomorphological reconstruction to date for Characiformes and set the stage for studies further elucidating the processes underlying its remarkable diversification.

Introduction

Wainwright 2011), or a lack of selective constraints (Hansen 1997; Collar et al. 2009).

In the simplest scenario, stochastic, gradual accumulation of morphological diversity throughout time can result in older clades accumulating higher disparity. Yet, many of the most dramatic episodes of diversification in the history of life resulted non- randomly from clade-specific events that changed the underlying process of evolution and adaptation (Price et al. 2010; Burbrink et al. 2012; Arbour and López-Fernández

2016). For example, a clade may experience multiple adaptive bursts linked to repeated environmental changes, in which case disparity will continue to increase throughout cladogenesis (Schenk et al. 2013; Stubbs and Benton 2016). Alternatively, a lineage

15 may diversify explosively following the invasion of novel adaptive zones after colonizing new habitats (Lovette et al. 2002; Muschick et al. 2012; Hulsey et al. 2013) or after evolving a key, novel phenotype that allows it to exploit new trophic resources

(Rüber et al. 1999; Grant and Grant 2006; Cooney et al. 2017). In a classic adaptive radiation scenario, the key innovation causes a singular early burst of phenotypic disparity that declines over time as niche space fills (Simpson 1953; Losos et al. 1998;

Madsen et al. 2001; Harmon et al. 2003).

Scientists have long debated the degree to which adaptation and diversification are deterministic or dependent on historical contingency (Gould 1989, 2002; Morris

2003; Losos 2017). On one hand, if adaptive evolution is highly deterministic then the repeated invasion of the same ecological niche would produce convergent morphologies, as seen in the numerous Anolis lizard radiations (Losos 1992; Losos et al. 1998; Harmon et al. 2005; Mahler et al. 2013). Although convergence can be quite common in biological systems, many factors influence two lineages’ ability to converge even when under the same functional demands, and clades that exist under similar selective pressures may not always converge because the contingencies of history can change the trajectory of diversification (Gould 1989). These factors include underlying genetic or behavioral differences, clade-specific innovation (Losos 2011), and variation in ecological opportunity caused by the presence or absence of competitors in disparate ecosystems (Hansen 1997; Langerhans and DeWitt 2004).

Because clades respond to the evolutionary dynamics of other organisms that share their environments, one might expect more contingency than determinism in the evolutionary history of geographically expansive radiations (Foote 1997, Ricklefs

2010). Evolutionary determinism would be less likely in such cases because different portions of the radiation evolve in the presence of different species communities (Gould

1989; Losos 2011). Prior occupancy by other species in a community potentially restricts a clade’s ability to diversify by reducing niche space, and once morphospace has saturated, diversification in some clades must be balanced by shrinking in others

(Ricklefs 2010). In other words, diversifying clades fill niche space vacated by extinction or force out other lineages through competitive exclusion. Even if two clades have the same intrinsic potential for adaptation, differential competition and niche occupancy can restrict diversification in one of the clades, resulting in contingent evolution (Foote 1997; Ricklefs 2010). Thus, studies on contingent evolution and morphological disparity should also investigate whether niche occupation and competitive exclusion have shaped the evolutionary dynamics of the clades under study.

With more than 29,500 described species (Nelson et al. 2016) exhibiting a riot of morphological disparity, teleost fishes represent perhaps the most hyperdiverse of all vertebrate lineages. About half of all vertebrates on the planet are teleosts with more species being described every year (Eschmeyer et al. 2010). Fish ecomorphology varies remarkably across clades, including species with backward facing jaws that feed while standing on their heads (Sidlauskas 2008), lineages that glide above the water with pectoral fins the size of their body (Davenport 1994), and fishes with flat asymmetrical bodies adapted for burrowing in the substrate (Friedman 2008; Martinez and Stiassny

2017). Yet, the evolutionary processes shaping morphological diversity are well understood in only a few model fish groups. These include: trophic diversification in

17 response to shifts in trophic guild in Labroidei (Hulsey 2006; Alfaro et al. 2009; Cooper and Westneat 2009; Price et al. 2010), morphological diversification across the pelagic- benthic habitat axis in North American darters (Carlson and Wainwright 2010), North

American minnows (Hollingsworth et al. 2013; Burress et al. 2016), and Lake Malawi cichlids (Hulsey et al. 2013) or phenotypic plasticity in response to variation in flow

(Langerhans et al. 2003; Langerhans 2008; Gaston and Lauer 2015). These studies have focused on select morphological characters related to feeding, or how body shape varies in response to a single shift in ecology. To date, few studies have attempted to explain how multiple ecological shifts into new adaptive zones across a large geographic range could have contributed to morphological diversification in a single, large clade of teleost fishes.

The freshwater fish order Characiformes provide an ideal opportunity to test whether stochastic evolution or adaptation and deterministic evolution across environmental gradients explains their extensive variation in morphology and ecology.

Comprising more than 2000 species and distributed primarily in South America and

Africa, characiforms vary dramatically in body shape, jaw morphology, and tooth morphology. However, most morphological studies have focused on the diversification of smaller subclades (Buckup 1993a; Sidlauskas 2008) or address broad scale systematic revisions and phylogenetic analyses (Fink 1996; Buckup 1998; Malabarba and Weitzman 2003; Mirande 2009; Mirande 2010), rather than explicitly studying morphological disparity. In essence, previous morphological studies have only hinted at the evolutionary processes that have shaped this large radiation of fishes.

Guisande et al. (2012) represents the most comprehensive study to date of characiform ecomorphological diversification. They found that the length of the dorsal- fin base varied considerably, but predictably across macrohabitats in characiform fishes. The length of the dorsal-fin base was smaller in species in high flow habitats versus species in low flow habitats, which the authors proposed may reduce drag in faster moving water (Guisande et al. 2012). That study also found that tooth shape varied consistently across trophic groups with carnivorous species possessing caniniform teeth, omnivores possessing multicuspidate dentition, and herbivores tending toward molariform teeth. They postulated that these results indicate that habitat and trophic ecology have shaped aspects of morphological diversity in the order through an adaptive radiation, but without direct tests of adaptive evolution or comparison of alternative modes of diversification, the exact evolutionary mechanism and extent that ecology influences morphology in Characiformes remain unknown.

Our study examines the evolution of characiform body shape in the context of shifts among different trophic ecologies and against the backdrop of independent evolution on two continents to determine the extent to which major dietary shifts have influenced body shape diversification across one of the largest radiations of fishes. We infer a large, robust molecular phylogeny by synthesizing data from previous studies and link it to a novel body shape dataset. Phylogenetic comparative methods then address four primary questions about body shape evolution in the order: 1) Did body shape morphology diversify through stochastic or adaptive evolution? (2) Did body shape diversify continuously, or in repeated bursts of adaptive evolution, or in the early burst that typifies adaptive radiation? (3) Are dietary shifts are the primary drivers of

19 diversification (4) Did body shape diversify primarily contingently or deterministically, as would be indicated by mostly dissimilar or similar evolutionary patterns in Africa and the Neotropics?

Materials and methods

Phylogenetic analysis

We synthesized data for the four most commonly sequenced loci in characiform systematics from nine previous publications and added new data from 27 additional taxa to assemble a dataset spanning 129 taxa in 19 characiform families. Taxon sampling averaged 30% within each family. See Supporting Information S1 for details on data acquisition and alignment. Macroevolutionary analyses employed an ultrametric tree estimated under a Bayesian relaxed-clock model in BEAST 2.0

(Bouckaert et al. 2014) through the CIPRES web server (Miller et al. 2010) using four fossil calibrations. Parameters and priors for the BEAST analysis were set up using

BEAUTi 2.0 (Bouckaert et al. 2014) and appear in Table 7.3 with detailed description in Supporting Information S3.

We performed four independent MCMC runs for 250 million generations, sampling every 25,000 generations. The first 10% of the posterior was disregarded as burnin after visualizing the posterior distribution in Tracer 1.5 (Rambaut and

Drummond 2007) using effective sample size (ESS), prior convergence and likelihood

(−ln L) of the priors and posterior estimates. An ESS >200 is generally considered to indicate that a MCMC has searched the likelihood-posterior probability landscape

20 sufficiently (Drummond et al. 2006). The remaining trees were used to construct a maximum clade credibility tree.

We also conducted Bayesian and Maximum-Likelihood reconstructions in Mr.

Bayes and RAxML to ensure that these methods did not yield a substantively different topology from the BEAST analysis (see Supporting Information S2 for analytical details). Because the interfamilial relationships inferred from the synthetic dataset using any of these three methods differed slightly from a recent comprehensive molecular phylogeny of the order inferred from similar data (Oliveira et al. 2011), we also used BEAST 2.0 to infer an ultrametric tree constrained to that topology (see

Appendix Figure 7.1 for differences and Appendix S3 for analytical details) and repeated all macroevolutionary analyses using that constrained phylogeny.

Morphological data acquisition

Seven collections provided specimens for the geometric morphometric analysis with detailed specimen descriptions available in the Supporting Information. One to ten specimens (average= 7.1±2.3, dependent on availability) from 329 species of characiform fishes were included with the entire morphological dataset containing

2210 specimens from 21 families. To avoid confounding the analysis with the known ontogenetic body shape allometry in characiforms (Fink and Zelditch 1995; Sidlauskas et al. 2011) only putative adult specimens (>50% of the maximum reported length for the respective species) were used. Morphospaces were constructed using 329 species, including many from species lacking sequence or ecological data. However, only the

118 species that matched a tip on the synthetic phylogeny and had matching ecological data were used in comparative phylogenetic analysis.

Geometric morphometric data were collected from radiographed specimens

(see Supporting Information S4 for analytical details). Because the phylogenetic comparative methods required representation of each species by a single morphological data point, we calculated the position of the centroid for each species in morphospace, and used those coordinates in subsequent phylogenetic comparative methods.

Morphological disparity was estimated using Procrustes variance implemented using the morphol.disparity function in the R package geomorph (Adams and Otarola-

Castillo 2013). The significance of variation between groups (families and geographic location) was assessed statistically by using group disparity as the observed value and permuting the shape matrix 1000 times relative to group assignment to generate a null distribution. Disparity through time plots were generated in geiger (Pennell et al. 2014) to examine the temporal pattern of change in morphological disparity and compare the observed pattern of intraclade versus among clade disparity through time with the expectation under Brownian motion.

Ecological classification

Each species was assigned to a generalized trophic guild representing a functional feeding group. We determined the trophic guild of each species through a primary literature review of diet in characiform fishes in Academic Search Premier,

Web of Science, and Google Scholar. We also examined species descriptions, field

22 guides, and checklists, and searched FishBase (Froese and Pauly 2000) for dietary data.

See Supporting Information S4 for details on trophic assignment.

After tip species were assigned to a trophic category, the ancestral dietary regimes along the internal branches of the phylogenetic tree were determined using

Bayesian stochastic character mapping (1000 simulations; Huelsenbeck et al. (2003)) implemented in the simmap function in the phytools package (Revell 2012). These dietary reconstructions were then used as the phylogenetic framework in a suite of comparative analyses. Rates of discrete character evolution were analyzed with reversable-jump MCMC in BayesTraits 2.0 (Pagel and Meade 2013) to model transitions among trophic ecologies.

Stochastic vs adaptive evolution

We performed a suite of model selection analyses in the R package mvMORPH

(Clavel et al. 2015) using small sample-size corrected AICc to determine whether evolution was stochastic or adaptive. We included two models of stochastic evolution, a simple model of Brownian motion and mixed rate model of Brownian motion. We tested multiple a priori models of adaptive evolution, including an early burst model and several Ornstein-Uhlenbeck (OU) models designed to distinguish alternative hypotheses of adaptive body shape evolution (Figure 2.2).

Because preliminary analyses revealed that Old World and New World lineages spanned substantially different portions of morphospace, one of our OU models

(OUcontinent) hypothesized that characiforms diversifying on each continent evolved towards separate adaptive optima. Another model (OUtrophic) hypothesized a distinct

23 adaptive adaptive optimum for each trophic category, directly testing whether species with different trophic ecologies evolved towards different optima, and whether lineages independently evolving similar trophic ecologies converged in morphospace. The next

(OU family) hypothesized a distinct optimum for species in each family, as would be expected if body shape was linked to lineage formation early in the evolutionary history of the order. OU family + piscivore was a similar to OU family, but also hypothesized that all piscivorous species converge on a single adaptive optimum regardless of the family to which the piscivorous species belonged. This model was included because preliminary analyses and visual inspections of morphospace suggested that piscivorous lineages, and no other trophic lineages, converged.

To determine if our a priori models were appropriately comprehensive, we also included the best a posteriori model selected by an unconstrained SURFACE analysis

(Ingram and Mahler 2013), and a constrained SURFACE model where piscivorous lineages converged. SURFACE was applied to the MCC phylogeny identified in

BEAST using the first four PC axes for the entire dataset. A delta AICc threshold of -

3, rather than the defaults of zero was implemented because SURFACE can overfit models and identify biologically unreasonable regime shifts or convergences (Ingram and Mahler 2013; Pennell et al. 2014) without such correction. The AICc threshold was chosen using the estimation methods in the function MEDUSA in the R package geiger

(Pennell et al. 2014). Even with the ΔAICc correction, SURFACE may still over fit the number of optima (Adams and Collyer 2017). However, we included the test because it is still the most robust method currently available to identify optima a posteriori, and provides an important complement to our set of a priori hypotheses.

Tests of evolutionary determinism

We included two different tests of evolutionary determinism to test the repeatability of the two continental radiations and determine whether repeated evolution of trophic ecologies resulted in repeated evolution of body shapes. We constructed disparity through time plots in the geiger R package (Pennell et al. 2014) to examine whether the temporal pattern of change in morphological disparity along the phylogeny differed between the two radiations. Such analyses compare of the observed pattern of intraclade versus interclade disparity through time with the expectation under Brownian motion.

To test whether the repeated shifts into similar trophic ecology resulted in similar body shapes, we used the a priori morphospace test of convergence implemented in the R package convevol (Stayton 2015). Instances of potential convergence were visualized using a phylomorphospace approach (Sidlauskas 2008) as implemented in the R package phytools (Revell 2012). Lineages were considered putatively convergent when independent lineages that shared a trophic ecology fell into the same portion of morphospace or moved away from sister taxa in a similar direction in morphospace. This test was particularly important in determining the statistical significance of the tendency of piscivorous lineages to be longer and more streamlined than their sisters.

We calculated convergence using Stayton’s (2015) distance-based method (C1-

C4), which quantifies the degree to which the convergent taxa evolved to be more similar than their reconstructed ancestors. It calculates the distance between two lineages as the proportion of the distance between the two species tips and the largest

25 distance between the species tips and the most recent common ancestor. Trait data were simulated under Brownian motion across the tree for 500 generations to test the significance of the results.

Results

Phylogeny

BEAST reconstructed a well-resolved and well-supported topology for 129 species, supporting previous hypotheses for most relationships in the order (Figure 2.3;

Appendix Figure 7.2). All major lineages are supported by high posterior probabilities

(>.95). Most interfamilial relationships exhibit very high posterior probabilities (>.99), except for the placement of Erythrinidae, Iguanodectidae, and Parodontidae which had ambiguous support (<.5) due to gene tree conflict. Some intergeneric relationships are also poorly supported (<.95), particularly within the families Anostomidae,

Curimatidae, and Alestidae, mostly likely as a result of low taxon sampling. Thirteen of the 15 families represented by multiple species sampled resolved as monophyletic.

The exceptionally species-rich Characidae was recovered as polyphyletic and

Triporthidae as paraphyletic (Figure 2.3). Our topology is broadly congruent with many previous morphological and molecular phylogenies for Characiformes (Buckup

1993a; Calcagnotto et al. 2005; Mirande 2009; Oliveira et al. 2011), but see Supporting

Information S6 for detailed descriptions of some topological differences.

Trophic ecology

Characiforms vary substantially in trophic ecology (see Supporting Information

Table S2), and converged many times upon all trophic guilds except for omnivory, which represents the ancestral condition in all SIMMAP reconstructions (Figure 2.4).

On average, detritivory evolved four (a maximum of five) times, herbivory five

(maximum of seven) times, and piscivory and insectivory each evolved eight

(maximum of 9 and 10 respectively) times. Notably, the number of species in a guild does not predict the number of times each guild arose on our phylogenetic tree. In detritivory (the second most speciose strategy) 83% of the diversity arose in a single radiation (Curimatidae and two allied families). Each evolution of herbivory yielded only a few modern species. Origins of piscivory and invertivory together account for almost half of the trophic transitions, but subtend clades containing only 17% of the total species diversity. See Appendix S7 for numbers of species in each trophic category, and inferred transition rates among categories.

Timing and patterns of phenotypic diversification

Four components summarized 85% of the variance in the landmark coordinates following Procrustes superimposition, with PC1, PC2, PC3, and PC4 summarizing

47%, 23%, 10%, and 5% of the variance respectively. The remaining components accounted for 3% or less of the total variance, are indistinguishable from measurement error, and were not considered further.

The wireframes in Figure 2.5 visualize the shape change along each axis. Fishes with a low PC1 score exhibit a moderately elongate body shape, a terminal mouth,

27 anteriorly placed dorsal-fin, and posteriorly placed anal-fin. As PC1 increases the body depth becomes greater with the mouth becoming more elongate and upturned, the length of the anal-fin increasing, and the dorsal- and anal-fins aligning more posteriorly on the fish’s body. Low PC2 scores indicate a deep-bodied fish with a relatively small mouth, an anteriorly placed dorsal fin and posteriorly placed anal fin. As PC2 increases, the body becomes more elongate and slender with the dorsal and anal fins positioned just posterior to the mid-lateral section of the body, and the terminal mouth becomes increasingly large. A low PC3 score indicates a deep-bodied fish with a short caudal peduncle, small head, but a large mouth. As PC3 increases, the body becomes more elongate, the dorsal and anal fins move anteriorly, and the mouth diminishes. Low PC4 scores indicate a deep bodied fish with a small mouth and head. As PC4 increases the body becomes more elongate and the head, mouth, and eyes increase in size.

PCA morphospace and disparity

Characiformes have diversified greatly in body shape (Figure 2.6), with members falling into one of four major morphospace regions. Fishes at the top of the morphospace exhibit sagittiform morphologies, characterized by an elongate, slender body with the dorsal and anal fins located posteriorly. This region of morphospace contrasts with the bottom right portion, populated by compressiform, deep-bodied fishes with the dorsal and anal-fins located posteriorly on the body. The highest density of families occurs in the middle left portion of the morphospace. These fishes exhibit a fusiform morphotype intermediate between the two previous described extremes, with the body shape being relatively elongate and robust, the dorsal-fin positioned

28 anteriorly, the anal-fin positioned posteriorly, and the mouth terminal or subterminal.

Only a handful of species occupy the fourth area of morphospace on the middle far right. These species differ from the other relatively deep, compressiform fishes with posteriorly placed dorsal and anal fins, by having upturned versus terminal or inferior mouths. These species include the entirety of the sampled members of the family

Gasteropelecidae (12), as well as the genus Poptella from family Characidae.

Body shape varies substantially among, but not within most families.

Individuals of the same family tend to exhibit similar body shapes and occupy compact regions of morphospace. The disparity analysis reflected the morphospace occupation with 48% of the lineages exhibiting extremely low levels of disparity (Procrustes variance < .002). In the multipeak OU model postulating separate adaptive peaks for each family, many of these lineages have low species diversity, high alpha values

(average α = 8.06 ± 0.9) and low sigma values (average σ = 0.014 ± .002), suggesting relative morphological and ecological stasis after initial colonization of the adaptive peak. In contrast, 29% of the families exhibited significantly higher levels of disparity

(Procrustes variance >.005) than would be expected under Brownian motion, with

Characidae exhibited the highest observed disparity (Procrustes variance = 0.0078). All of the lineages with high morphological disparity are relatively species rich and ecologically diverse with varying degrees of morphospace occupation. The Characidae and to a lesser extent the Alestidae (0.0054) vary considerably in morphospace, with the characids occupying almost half of the entire morphospace.

Disparity through time

Morphological disparity peaked early in characiform cladogenesis (Figure 2.7) as lineages evolved the full range of modern body shapes, then sharply decreased as lineages filled the morphospace more densely. The initial disparity spike and the subsequent drop both depart from the expectations of neutral (Brownian) evolution.

Adaptive vs stochastic evolution

Both a priori models of Brownian motion were very unlikely (Table 3). Other, a priori models, including one with separate optima for both the Old and New World lineages, one with a single optimum for each trophic ecology, an early burst model, a model with each family sharing an optimum, and a model with all piscivorous lineages sharing a phenotypic optimum (OU family + piscivore and OU surface + piscivore) were also unlikely (Table 3). The mvMORPH analysis found the a posteriori model selected by SURFACE model to be the best fitting of all, and substantially more likely than the a priori model hypothesizing a major influence of trophic ecology on body shape (Table 3). SURFACE selected a model of seventeen adaptive optima spread all over the phylogeny (Table 3), even when the topology was constrained to that of

Oliveira et al. (2011) (Supporting Information Table S2). The SURFACE model was selected as the best model 100% of the time in the 1000 permutations that randomly re- classified each omnivorous species into one of the multiple trophic categories that comprised their diet. (Table 3; Appendix Figure 7.4).

Of the 17 adaptive optima identified by SURFACE, the 10 adaptive optima evolved on branches subtending families and larger clades in the order (Figure 2.8).

Most of the optima evolved very early in cladogenesis, and overall the SURFACE scenario represents a model of early body shape diversification. In that respect, it resembles the a priori models that included an element of early body shape diversification: OU family, OU family + piscivore, and Early Burst. These a priori models, like the OU surface an OU surface +piscivore models, were substantially more likely than the model postulating different adaptive peaks for each trophic ecology.

Clearly, the model based on trophic ecology did not best fit the data.

To what degree was adaptation deterministic in Characiformes?

Each continental radiation followed a very different pattern of diversification.

The largest radiation of body shapes occurred in New World lineages, which have significantly higher (p=.001) disparity (Procrustes variance = .0132) than the African lineages (Procrustes variance = .0071). New World lineages are also substantially more speciose than the African lineages, accounting for more than 70% of the species richness in the order. Morphological disparity increased rapidly in the early part of both radiations (Figure 2.9) and the New World radiation evolved the full range of modern body shapes early. In contrast, the Old World radiation exhibited a second burst of diversification later in cladogenesis before the full range of body shapes evolved

(Figure 2.9). Neotropical lineages occur throughout the entire morphospace, while the four African families are restricted to the middle (Figure 2.6). The two species- depauperate African families (Hepsetidae and Citharinidae), exhibit very low body shape disparity and are represented by a single morphotype. Body shape in much more speciose Alestidae and Distichodontidae (each sister to one of the other African

31 families) is much more variable (Procrustes variance = .0053 and .0064 respectively).

However, these clades are also restricted to the center of the morphospace, and have not evolved the extreme body shapes exhibited by the New World lineages at the morphospace periphery, which include the most elongate and deep bodied characiforms.

Trophically similar linages rarely converged in the evolutionary history of

Characiformes. Detritivores, herbivores, insectivore, invertivores, and omnivores all arose multiple times, yet generally evolved different body shapes scattered throughout the phylomorphospace (Figure 2.10). Fishes in each of these trophic groups could be deep bodied or elongate, with large or small heads and terminal, subterminal, or upturned mouths. Most trophic groups, with the exception of piscivorous lineages, exhibited no statistically significant convergence (Table 3; Supporting Information

Table S3 and S4).

Only piscivorous lineages converged substantially and significantly in phylomorphospace. The arrows in Figure 2.10 indicate piscivorous lineages evolving towards positive and significantly similar PC2 values (convevol, P=.001). All piscivorous clades have a more elongate body shape than their non-piscivorous sister clade. Their C1 of 0.44 (from convevol) indicates that evolution has closed 44% of the morphological distance among piscivorous taxa. Similar results were found when accounting for trophic uncertainty and using the constrained tree (Oliveira et al. 2011), with piscivorous lineages always found to be statistically convergent (Supporting

Information Table S3 and S4 respectively). These analyses provide evidence that

32 nonrandom evolutionary processes generated similar body shapes among lineages that feed exclusively on fishes, regardless of the tree topology used.

Discussion

Characiform body shape diversity exploded early in in the clade’s history, during which the incipient lineages leading to the 21 modern families spread out and occupied largely separate and distinct regions of morphospace. In these initial stages of the radiation, characiforms evolved many disparate body shapes including elongate fishes with large heads and terminal mouths, deep-bodied fishes with superior mouths, and many morphologies between those extremes. Biogeographical and paleontological studies place the earliest stages of this initial, exuberant radiation around 100 million years ago (Arroyave et al. 2013), when these early characiforms inhabited a vast tropical lowland on the supercontinent Gondwana (Roberts 1972; Roberts 1973).

When shifts in the Earth’s tectonic plates then separated Africa from South America around 90 million years ago, the sundering of the proto-Amazon and Congo river basins also split the radiation of Characiformes. That vicariance allowed the Old and New

World radiations to evolve independently thereafter. Yet, despite the opportunity for these replicate radiations to diversify similarly, they followed very different trajectories of diversification.

The New World radiation evolved a riot of a diversity throughout the Amazon,

Orinoco, Parana-Paraguay and associated basins, with most of the modern species and morphological diversity centered in the heart of South America. This half of the radiation spans nearly the entire range of body shapes explored by the order overall. In

33 contrast, the Old-World radiation that now inhabits the Congo, Ogooué, Niger and adjacent rivers evolved very few extreme morphologies and far fewer modern species.

Though both halves of the radiation evolved to inhabit the same broad trophic niches

(often multiple times), with the exception of piscivores, independent lineages invading similar trophic niches rarely evolved to resemble each other closely. In some cases, body shape remained constant during a shift in trophic ecology (for example: family

Anostomidae which exhibits three trophic shifts, but only one body shape shift), while in other cases body shape changed dramatically during a trophic innovation, as in the case of the clade consisting of families Chilodontidae + Curimatidae +

Prochilodontidae. But, these patterns lack consistency, and ecologically convergent lineages typically evolved unique morphologies with slightly different placement of the jaws, fins, and shape of the trunk. Such high variation in evolutionary dynamics between the two radiations and across the multiple trophic categories, hints at a profound role for either stochasticity or historical contingency during the morphological evolution of the order.

Characiform body shape diversified adaptively, not stochastically

Though stochastic evolution could conceivably explain the differences among trophically similar linages and between the continental radiations, our model-testing approach revealed Brownian motion to be among the worst-fitting of all candidate models. We conclude that body shape diversity probably evolved non-randomly, i.e, adaptively. The predominantly early diversification revealed by the disparity through time plot for the entire order (Figure 2.7) represents perhaps the most important

34 violation of the Brownian model, which predicts gradual accrual throughout time rather than an early burst of change. Instead, the observed pattern of early disparification matches closely one prediction of adaptive radiation.

Body shape diversity exploded early in characiform evolution

Characiform lineages evolved to span most of the modern body shape diversity in the earliest part of the clade’s history. Morphological saturation followed that initial spike in body shape disparity as the lineages filled, matching patterns in many other diverse vertebrate lineages including iganuian lizards (Harmon et al. 2003), new world ratsnakes (Burbrink and Pyron 2010) and Neotropical cichlids (López‐Fernández et al.

2013). Though a few lineages (most notably the ecomorphologically diverse African family Distichodontidae) continued to evolve unique body shapes later in cladogenesis, as a rule, within-clade diversification was much rarer than among-clade diversification.

This pattern of early diversification matches one of the hallmarks of adaptive radiations observed in other vertebrate lineages (Harmon et al. 2003; Seehausen 2006;

Cooney et al. 2017). In adaptive radiations, bursts of morphological and lineage diversification occur as lineages move into empty ecological niches (Simpson 1953;

Harmon et al. 2003; Glor 2010) that are opened through ecological opportunity

(Schluter 2000; Losos 2010). Recent quantitative approaches (e.g. Harmon et al.

2003;Harmon et al. 2010), have typically modelled adaptive radiation with a declining rates model termed “early burst.” Though the early burst model was not the best fitting of our candidate models of diversification, the a posteriori OU-surface model selected as optimal, and all other highly-supported models include a preponderance of adaptive

35 shifts on branches closer to the root of the phylogeny than to the tips. The disparity- though-time plot for the entire radiation tells a similar story: morphological disparity peaked early in lineage formation with most of the diversity existing among (not within) subclades. In effect, we see initial rapid expansion through morphospace followed by a transition to a regime dominated by more restricted evolution around a series of adaptive optima. This differs from the “early burst” model in that the presence of multiple optima and a strong pull towards each governs the evolutionary dynamics more than a simple decrease in the sigma parameter modelling the raw degree of evolutionary change possible. Nevertheless, we posit that this pattern remains fully consistent with an ancient adaptive radiation.

Guisande et al (2012) previously suggested that Characiformes underwent an adaptive radiation (sensu Streelman and Danley 2003) due to the observed segregation of morphological characters among species separated by trophic and habitat categories.

However, without an ultrametric phylogeny, that study could not determine whether morphological segregation resulted from gradual diversification coupled with extinction or from rapid adaptive evolution. Our results add that missing piece, and support their scenario of rapid and early body shape diversification and morphospace exploration.

Shifts in trophic ecology did not drive the body shape radiation in Characiformes

Although our results demonstrate the early morphological evolution common to adaptive radiations, they do not demonstrate the tight correlation with diversification in trophic ecology typically considered a hallmark of adaptive radiation in fishes.

Indeed, our hypothesized OU-trophic model is one of the worst fitting of all examined, and model tests indicate that not all trophics transitions coincide with with the largest burst of body shape evolution, which occurred early in evolutionary history. Even if we disregard the prediction of morphological convergence inherent in the OU-trophic model, we still observe that only some shifts in trophic ecology correlate with major shifts in body form. Thus, although the origin of novel trophic ecologies correlates with dramatic shifts in feeding habits such as feeding on scales (Peterson and Winemiller

1997) instead of filtering detritus (Araujo-Lima et al. 1986), other factors appear to have promoted the diversification of body shape at the dawn of characiform evolution.

The open question then becomes: did this study simply fail to detect a major influence of trophic ecology, or did multuple ecological factors drive body shape diversification in Characiformes?

Several factors could explain a failure to detect an ecological correlate of morphological change. For example, there may be an insufficient number of transitions among trophic categories to provide the statistical power needed for detection by the phylogenetic comparative methods (Maddison and FitzJohn 2015). For instance, detritivores all seem to have a similar deep bodied morphology, but there are only four transitions to detritivory, and one of them yielded a much more species rich radiation than did the othere three, which might have reduced are ability to detect convergence.

That said, it seems unlikely that our analyses lacked statistical power to detect convergence because most of the trophic groups exhibited six or more independent evolutions of the same trophic ecology. Alternatively, our broad classification system may have failed to properly capture the true diversity in feeding mode and strategies.

For example, the phylomorphospace detected shared morphospace between members of the Gasteropelecidae and the characid genus Poptella. These convergent species are deep bodied, top water fishes that feed primarily on terrestrial insects (Netto-Ferreira et al. 2007; de Mérona et al. 2008). However, other lineages of insectivores, including members of the family Crenuchidae, are streamlined, benthic fishes, feeding primarily on aquatic insects (Géry 1977; Buckup 1993a). The lack of detailed trophic and habitat ecology data across the entire order forced our broad classification system to combine these two distinct strategies together. Thus, a more finely detailed trophic categorization that includes aspects of feeding and habitat ecology may reveal additional instances of convergence within the order.

On the other hand, it is also possible that variation in an untested ecological character system in combination with trophic ecology drove early body shape diversification in Characformes. If so, habitat variation provides the most likely possibility. Streelman and Danley (2003) proposed that adaptive radiations occur in stages with lineages diversifying in response to habitat, followed by trophic ecology, and ending with sexual selection. Some lineages of cichlids follow this pattern, with the earliest shifts in body shape co-occurring with shifts between macrohabitats

(Schliewen et al. 1994; López‐Fernández et al. 2013; Arbour and López-Fernández

2016). Other cichlid groups show macrohabitat adaptation later in cladogenesis

(Muschick et al. 2014). Regardless of when diversification occurred, both patterns clearly show body shape diversification occurring in response to shifts in macrohabitats

(Schliewen et al. 1994; López-Fernández et al. 2013; Muschick et al. 2014; Arbour and

López-Fernández 2016).

Is habitat the missing ecological character driving the early burst of body shape diversification in Characiformes? The results of the SURFACE analysis indicate at least three regime shifts (including a lineage in the Crenuchidae, the Gasteropelecidae, and the Serrasalmidae) occurring in lineages transitioning across the benthic-pelagic habitat axis, which is a common diversification gradient seen in other freshwater fish lineages (Carlson and Wainwright 2010; Hollingsworth et al. 2013; Burress et al.

2016), including multiple adaptive radiations (Schluter 1995; Schluter and Nagel 1995;

Hollingsworth et al. 2013). However, reliable habitat data are rarely available for members of the order. Our ability to test the influence of habitat on body shape evolution will sadly remain limited until macrohabitat sampling improves for many lineages in South America and Africa.

Novel body shapes evolved during adaptation to trophic ecology

Though Characiformes invaded the same six trophic categories multiple times, providing ample opportunity for convergence in body shape, only piscivores did so.

The remainder of the radiation took good advantage of multiple adaptive pathways to omnivory, detritivory, insectivory, herbivory, and invertivory. Overall, that flexibility indicates that trophic ecology alone does not determine body shape in the order. For example, two different lineages of invertivores evolved vastly different mouth positions: the unique superior mouth position in the subfamily Anostominae (Myers

1950) typified herein by the genus Synaptolaemus, and the more common inferior position in some members of the family Crenuchidae (Buckup 1993b). Our results contrast starkly with many other hyperdiverse vertebrate lineages including Anolis

39 lizards and crater lake cichlids, which exhibit morphological convergence when adapting to the same ecology (Losos 1992; Kocher et al. 1993; Mahler et al. 2013;

Machado-Schiaffino et al. 2015). For example, independent lineages of cichlids in Lake

Tanganyika evolved identical trophic morphology in response to the same selective pressure in different parts of the lake (Rüber et al. 1999). However, both Anolis lizards and crater lake cichlids radiated during independent invasions of unoccupied, geographically restricted “island” ecosystems. In contrast, Characiformes radiated in a dynamic river ecosystem that spans two continents and has changed significantly over time and throughout space. The dynamic differences in habitat ecology and community assembly likely influenced the trajectory of diversification across the trophic groups.

One trophic strategy, piscivory, provides a notable counterexample.

Characiform lineages that specialized in eating other fishes converged significantly on elongate sagittiform or fusiform bodies with large heads. The convergence in piscivorous Characiformes is not surprising, elongate fishes with large heads and jaws is a common piscivorous morphology across all fishes due to pursuit or ambush hunting strategies (Webb 1984; Rüber and Adams 2001; Williams et al. 2015). These convergences occurred in young and old lineages and on both continents, indicating strong evolutionary determinism during adaptation to a piscivorous lifestyle. Despite historical contigencies, body shape probably adapted convergently due to well-known selective pressures for a more streamlined shape in predatory fishes. For example, many lineages of piscivorous cichlids have similarly evolved elongate, shallow heads, and large streamlined bodies (Rüber and Adams 2001; Clabaut et al. 2007; López-

Fernández et al. 2012; Muschick et al. 2012).

Contingency, not determinism, characterizes the continental radiations

Despite the similarity in body shape among predatory characiforms on both continents, overall evolutionary dynamics differed dramatically between the two major radiations. Most importantly, the New World radiation explored many novel regions of morphospace early in cladogenesis, evolved much higher disparity and evolved towards many more unique adaptive optima. They achieved that elevated disparity despite demonstrating only a single major episode of diversification, unlike the Old

World radiation, which experienced a second burst of diversification relatively recently, driven largely by diversification within family Distichodontidae. Our results resemble patterns seen in Neotropical cichlids, in which the South American lineages exhibited an initial spike in disparity that quickly declined, while Central American lineages continued to diversify throughout cladogenesis (Arbour and López-Fernández

2016). In contrast to the pattern in Central American cichlids (Arbour and López-

Fernández 2016), the Old World characiform radiation was restricted to the center of the morphospace even though it underwent a second diversification event. The similarity in diversification dynamics between characiforms and Neotropical cichlids offers insight into how historical contingencies, such as biogeographic differences between continents, can shape broad geographic radiations.

Many factors could explain the differences in diversification between the New and World lineages, including unequal clade age, selection across different ecological regimes, or differences in competition and niche exclusion. However, differences in clade age and ecological regimes seem unlikely to provide the most important answers.

The radiations are the same age, occupy two halves of what used to be the same river

41 system, and inhabit similar aquatic habitats in tropical regions of sister continents

(Roberts 1972). The key more likely lies in differences in competition, niche exclusion and community composition accrued over the approximately 90 million years of independent evolution since the separation of these faunas.

The concurrent diversification of Neotropical cichlids provides an excellent example of how competitive exclusion can influence diversity dynamics. The

Geophagini, which occupy an unusual benthivorous niche, evolved a riot of morphological diversity in South America, while the closely related Hereoini were restricted to the extremes of morphospace (López‐Fernández et al. 2013). However, when Hereoine cichlids invaded Central America, where geophagines are absent, the lineage diversified more extensively in morphospace than it had occupied in South

America, suggesting that competitive exclusion by geophagines restricted the morphospace of South American hereoin cichlids (López‐Fernández et al. 2013). Since they do not co-occur, New World characiforms could not have restricted the morphospace of their Old World counterparts in that way that geophagin cichlids restricted the diversification of hereoin cichlids, but some other lineage(s) of fishes present in Africa and not South America certainly could have.

The presence of cypriniform fishes in Africa, but not South America, provides the most likely cause of the massive difference of characiform diversity on these continents. The New World characiform radiation, along with the siluriform radiation

(catfishes) dominates the South American fish communities (Lévêque et al. 2008;

Nelson et al. 2016), while Old World characiform lineages compete with cypriniforms and many other components of the African ichthyofauna (Briggs 2005). The absence

42 of cypriniform fishes in South America contrasts starkly with their otherwise global distribution, in which they are abundant on every other non-oceanic continent (Roberts

1975; Lévêque et al. 2008; Nelson et al. 2016). Many species of African cyprinids co- occur with African characiforms and occupy ecological niches similar to those occupied by New World characiforms. It is possible that African cyprinids had already filled much of the niche space exploited during the morphological diversification of

South American characiforms and through competitive exclusion, restricted the expansion of Old World characiforms in morphospace. The discrepancy in morphospace occupation between Old and New World radiations indicates that biogeography and community structure can have profound effects on morphological diversification. However, further work, such as construction of a morphospace for

African cypriniform fishes, is needed to fully understand whether and how those fishes restricted the diversification of Old World characiforms.

Conclusion

Characiform fishes evolved much of their modern body shape diversify in an early, exuberant burst of shape change. During later stages of cladogenesis, lineages filled in gaps in morphospace, and overall disparity decreased. Body shape diversification occurred through adaptive, not stochastic evolution, and the evolution of piscivory drove strong convergence towards elongate bodies and large heads.

However, shifts to other trophic ecologies do not always coincide with major changes in body shape, and even when they do, no clear signal of morphological convergence emerges. Models of early diversification fit the evolution of body shape substantially

43 better than a model based on distinct adaptive peaks for each trophic ecology. Overall, results suggest that unmeasured ecological variables, such as shifts in habitat, may have driven much body shape diversification in Characiformes. The massively different morphospace occupancy of the African and Neotropical portions of the radiation suggest a strong role for contingency in characiform evolution. New World lineages evolved twice as much disparity as the Old World lineages, and the most extreme body shapes occur only in the Neotropical radiation. Though the mechanism responsible the discrepancy is unknown, the presence of cypriniform competitors in Africa, but not

South America, offers a tantalizing plausibility. Overall the evolution of body shape in

Characiformes exhibits some signatures of an adaptive radiation, but without a clear ecological correlate, it is currently impossible to determine the exact mechanisms and processes driving body shape diversification in the order.

Acknowledgements

We thank Rich Vari in memoriam for his helpful guidance in the early stages of the project and mentoring throughout much of the data collection. Allison Bullock assembled initial versions of the trophic ecology classification with support from the

National Evolutionary Synthesis Center (NESCent, NSF Grant #EF-0905606); without her initial work, this project would not have been possible. We thank my dissertation committee of David Maddison, Guillermo Ortí, and Tiffany Garcia for helpful comments and discussions. We thank Selene Fregosi, Evan Bredeweg, and Vicki Tolar-

Burton for helpful comments on the manuscript, and Thaddaeus Buser and Peter

Konstantinidis for insightful discussions. We gratefully acknowledge the work of

44 numerous collection managers and curators who made materials available, including

Mark Sabaj (ANSP), David Werneke and Jon Armbruster (AUM), Mike Retzer

(INHS), Sandra Raredon and Richard Vari (USNM), Kevin Swagel and Susan Mochel

(FMNH), Lucia Rapp Py-Daniel and Renildo Oliveira (INPA), and Karsten Hartel and

Andrew Williston (MCZ). NSF grant DEB – 1257898 awarded to BLS provided support to BLS and MDB. An OSU Provost Fellowship, OSU Department of Fisheries and Wildlife M. A. Ali Graduate Chair Award in Fisheries Biology, a Smithsonian

Graduate Fellowship and an ASIH Raney Award provided support to MDB.

Figure legends

Figure 2.1 Radiograph of Acestorynchus minimus USNM 311177 50.0 mm SL, showing position of 26 landmarks used in geometric morphometric analysis. (1) anterior limit of premaxilla; (2) posterior limit of premaxilla; (3) anterior limit of orbit;

(4) posterior limit of orbit; (5) ventral limit of orbit; (6) dorsal limit of orbit; (7) dorsal margin of neurocranium at vertical through center of orbit; (8) posterodorsal tip of supraoccipital; (9) dorsal-most portion of the base of the first dorsal-fin ray pterygiophore; (10) dorsal-most portion of the base of the last dorsal-fin ray pterygiophore; (11) base of the body dorsal of the fourth most posterior; (12) compound vertebral centrum at posterior of verebral column; (13) posterodorsal limit of hypurals;

(14) ventrodorsal limit of hypurals; (15) dorsal-most portion of the base of the of last anal-fin ray pterygiophore; (16) dorsal-most portion of the base of the first anal-fin ray pterygiophore; (17) pelvic-fin origin; (18) pectoral-fin origin; (19) ventral limit of joint between contralateral cleithra; (20) anterior tip of dentary; (21) joint between basioccipital and first vertebra of Weberian apparatus; (22) anterior limit of fifth vertebra (first vertebra not incorporated into Weberian apparatus and first bearing full sized pleural ribs); (23) vertebrae centroid posterior to first dorsal-fin ray pterygiophore; (24) vertebrae posterior to first anal-fin ray pterygiophore.

Figure 2.2 Alternative multivariate OU hypotheses for the evolution of body shape in characiform fishes. The trees show the different OU regimes hypotheses used in model selection analyses performed in mvMORPH. The first model (OU continental) hypothesizes that species on the same continent share an adaptive peak. The second model (OU trophic) hypothesizes that species that share a trophic ecology also share an adaptive peak. The third model (OU family) is a phylogenetic hypothesis where species share an adaptive peak with other members of their family. The fourth model

(OU family + piscivore) is phylogenetic hypothesis where species share an adaptive peak with other members of their family except that all piscivorous fishes share an adaptive optimum irrespective of their familial affiliation. The fifth model (OU surface)

47 is the best fitting model returned by the SURFACE analysis. The sixth model (OU surface + piscivore) is the best fitting model returned by the SURFACE analysis, but with all piscivorous fishes sharing an adaptive optimum. We also fitted a Brownian motion and early burst model (not shown).

Figure 2.3 Relationships among sampled species in the order Characiformes through a

Bayesian partitioned analyses of the concatenated dataset. Circle at each node represents posterior probability score. Red represents a posterior probability score of 1, blue represents a score between 0.95 and 0.99, purple represents a score between 0.8 and 0.94, and no circle represents a score below 0.8.

Figure 2.4. Trophic ecology evolutionary history: summary of 1000 SIMMAP character maps using the fossil-calibrated phylogeny. Colors represent different trophic ecologies; sector of pies at nodes are proportional to the probabilities of each state at that node.

Figure 2.5. Deformation grids illustrating limits of observed variation of shape change on the four most important principal components. Numerical values represent the percent variance explained by each component.

Figure 2.6. Scatterplot of principal components showing that species in most families cluster closely. New World includes families from South and Central America. Old

World includes families from Africa. (1) Acestorynchidae [2/26], (2) Alestidae

[21/116], (3) Anostomidae [86/156], (4) Characidae [51/1106], (5) Chilodontidae [7/8],

(6) Citharinidae [2/8], (7) Crenuchidae [6/87], (8) Ctenoluciidae [2/7], (9) Curimatidae

[74/105], (10) Distichodontidae [12/101], (11) Erythrinidae [3/17], (12)

Gasteropelecidae [4/9], (13) Hemiodontidae [8/31], (14) Hepsetidae [1/5], (15)

Iguanodectidae [1/30], (16) Lebiasinidae [7/77], (17) Parodontidae [8/32], (18)

Prochilodontidae [17/21], (19) Roestinae [1], (20) Serrasalmidae [7/95], and (21)

Triporthidae [5/23].

Figure 2.7. Accumulation of multivariate disparity through time. Black line equals observed data, dotted line equals the null Brownian expectation, and shading represents the 95% confidence interval.

Figure 2.8. Fossil-calibrated phylogenetic tree for characiform fishes showing adaptive optima and regimes for the best-fitting model of body shape evolution assigned based on the SURFACE results. Drawings depict general body shapes representing the adaptive optima and do not represent the actual ancestral state.

Figure 2.9. Accumulation of multivariate disparity through time for New and Old

World lineages. Black line equals observed data, dotted line equals the null Brownian expectation, and shading represents the 95% confidence interval.

Figure 2.10. Phylomorphospace depicting repeated invasion of elongate morphotypes

(high PC2 value) by piscivorous lineages.

Tables

Table 2.1. Results of the multivariate model-fitting analyses for body shape. For each model, the number of parameters (P), the Akaike information criterion (AIC), the small sample corrected AIC (AICc), and the relative fit (ΔAICc) and support (AICc weight) are shown. The best model has the lowest ΔAICc.

Model P AIC AICc Δ AICc AICc weight OU surface 62 -1929 -1910 0 0.989 OU surface + piscivore 66 -1923 -1901 9 0.01 OU family 86 -1877 -1838 71.7 0 OU family + piscivore 78 -1855 -1823 86.6 0 BMM 84 -1844 -1809 100.8 0 EB 15 -1783 -1782 127.5 0 BM 12 -1769 -1744 165.5 0 OU trophic 38 -1751 -1769 140.8 0 OU continent 22 -1737 -1735 174.6 0

Table 2.2. Results from the convevol analysis. All values with statistically significant

P-values (P < 0.05) indicated with an asterisk.

Trophic Ecology C1 C2 C3 C4 Piscivores 0.434* 0.068* 0.0085* 0.014* Detritivore 0.199 0.023 0.0031 0.0044 Herbivore 0.307 0.0706 0.008 0.013 Insectivore 0.08 0.018 0.002 0.003 Invertivore 0.044 0.011 0.0014 0.0021 Omnivore 0.2068 0.033 0.00437 0.00697

CHAPTER 3: EVOLUTIONARY MODULARITY PROMOTED NOVEL BODY

SHAPES IN A HYPERDIVERSE FRESHWATER FISH RADIATION

Michael D. Burns

Intended for publication in Evolution with the following co-author:

Brian L. Sidlauskas

Abstract

Scientists have long been fascinated by why some radiations exhibit high evolvability, while others do not, but both patterns are quite common across the tree of life. Little is known about the underlying evolutionary factors that promote or restrict morphological diversification. Evolutionary modularity, which allows different regions of the body to respond to selection independently, offers a plausible explanation of why some lineages evolve novel phenotypes during adaptation. The previous chapter demonstrated that characiform body shape exhibited high evolvability; lineages that shifted into the same trophic ecology evolved different morphotypes, except for convergent piscivores. We found that three independent modules comprise characiform body shape, and that these diversified at different times, different rates, and under different selective regimes. High evolutionary modularity plasusiby explains the ability of characiforms to discover multiple morphological solutions to the same trophic ecologies. More studies need to look at the role that evolutionary modularity and integration can have in shaping evolvability across broad and restricted radiations of vertebrates.

Introduction

The evolvability of an organism is its intrinsic capacity for evolutionary change

(Wagner 1996; Gerhart and Kirschner 1997; West-Eberhard 1998). Evolvability is a function of the phenotypic variation and the intrinsic potential for an organism to create phenotypies, and thus the amount of variation available on which selection can act. A number of characteristics of organismal organization are thought to contribute to evolvability (Conrad 1990; Gerhart and Kirschner 1997), including the reduction of genetic or developmental constraints that constrain phenotypic production (Alberch

1982; Cheverud 1984; Maynard Smithet al. 1985; Lande 1986, von Dassow and Munro

1999). These lack of contraints likely promote novel phenotypic evolution when responding to selection.

Modularity and integration are evolutionary mechanisms that can promote or restrict morphological diversification, with modularity likely increasing a radiation’s evolvability. Morphological integration is the strong covariation among biological structures that can constrain phenotypic evolution (Klingenberg 2014). Integration can occur when morphological units combine to perform a function, like the integration of cranial structures in eels used for suction feeding (Collar et al. 2014). These cranial structures are under strong selection to coevolve to preserve their functional feeding advantage, but do so at the expense of reducing the amount of phenotypic diversification possible in suction feeding lineages (Collar et al. 2014). Thus, strong integration has the potential to increase the likelihood and magnitude of convergence by reducing the range of easily evolvable phenotypes, which in turn could promote deterministic adaptation when invading the same niches. However, under relaxed

62 morphological integration, covariation among clusters weakens and morphological modules can respond to natural selection or drift independently of one another. Over evolutionary time, that increased modularity can increase phenotypic diversity. For instance, the innovation of biting in eels heightened the independence of the hyoid, jaws, and operculum, thereby elevating the phenotypic diversity of biting lineages relative to suction feeding lineages (Collar et al. 2014).

Modularity has the potential to not only act as a catalyst for phenotypic change, but it can act to increase a lineage’s ability to evolve novel morphologies in response to environmental change (Liem 1973; Lauder 1981; Schwenk and Wagner 2001). For instance, the decoupling of the oral and pharyngeal jaw mechanics in cichlid fishes likely caused the unparalleled trophic diversification of cichlid fishes by allowing the oral and pharyngeal jaws to respond to different ecological pressures independently

(Hulsey et al. 2006). Evolutionary modularity also has the potential to increase the rate of phenotypic change. For example, the raptorial appendage in “non-smashing” lineages of mantis shrimp exhibited more evolutionary modularity and a 10-fold increase in the rate of evolution relative to the highly integrated “smashing” lineages

(Claverie and Patek 2013). Overall, because increased modularity can increase the rate and potential for phenotypic evolution, highly modular lineages are more likely to evolve novel phenotypes than tightly integrated lineages, when adapting to the same niches.

Characiformes, a freshwater fish lineage with around 2000 described species, exhibits high evolvability, evolving multiple unique body shapes during the radiation of the order. Some lineages of characiforms appear to converge in body shape, with

63 distantly related lineages that share a similar ecological niche seemingly converging on coarsely similar body plans, including fusiform bodies in migratory species (Géry

1977), streamlined fishes with large pectoral fins in rheophilic lineages (Lujan and

Conway 2015), and streamlined fishes with large heads in piscivorous lineages (Burns and Sidlauskas, In Revision). Many other lineages evolve unique morphologies that are not repeated in the radiation, including fishes with backwards facing jaws (Myers and de Carvalho 1959; Géry 1977), a lineage with an enlarged pectoral girdle hypothesized to aid in “flying” (Burns 2017), or a genus with a highly elongate caudal peduncle used to ambush fishes during scale feeding (Arroyave and Stiassny 2014).

A recent macroevolutionary analysis of the order quantified body shape diversification for over 200 species and found that found that dietary convergence rarely triggered morphological convergence in characiforms, suggesting that lineages of this clade can freely explore various evolutionary trajectories as they search for ecomorphological optima (Burns and Sidlauskas, In Revision). Specifically, piscivorous lineages evolved elongate bodies with large heads during six independent invasions of that trophic niche (Burns and Sidlauskas, In Revision). Conversely, evolution in non-piscivorous lineages appeared to be contingent, with lineages evolving unique body shapes, including elongate, deep bodied, and intermediate forms, during each independent invasion of the same trophic niche (Burns and Sidlauskas, In

Revision). Burns and Sidlauskas (In Revision) postulated that the low occurrence of morphological convergence was caused, in part, by body shape adaptation in response to both habitat and trophic ecology, with their simplified ecological metric which focused only on trophic ecology missing many potential examples of convergence.

However, the narrow ecological classification system used in Burns and Sidauskas (In

Revision) alone does not explain the prevalence of unique body shapes during the radiation of Characiformes. Especially, since many of the seemingly contingent lineages appear to adapt different regions of their body plan independently, including the jaws, caudal peduncle, and placement of the fins.

Increased modularity might explain the prevalence of unique body shapes evolved in non-piscivorous characiforms. Recently, two sister species of lake-dwelling

Characiformes were shown to possess functionally distinct head and body modules

(Ornelas-García et al. 2017). That study examined four different functional modularity hypotheses across two species of Astyanax, A. aeneus and A. caballeroi. They found that the body shape of A. aeneus and A. caballeroi was comprised of two functional modules, the head and body, which they hypothesized led to the divergent feeding ecologies of the piscivorous A. caballeroi and the omnivorous A. aeneus. However, nothing is known about the patterns of modularity across the entire order. If characiforms exhibit modularity, non-piscivorous lineages might have only adapted one module, i.e. the cranium, in response to shifts in trophic ecology, thus, overall body shape would not converge. If characiforms are comprised of many independently evolving modules, characiforms likely have an increased susceptibility to the vagaries of history, increasing the likelihood of contingent evolution and the prevalence of lineages with unique body types.

Characiformes represents an excellent radiation in which to test the role of evolutionary modularity and integration in promoting evolvability because the order is ecomorphologically diverse (Guisande et al. 2012), contains examples of convergence

65 and non-convergence following invasion of the same trophic niches (Burns and

Sidlauskas, In Revision), and includes species that vary in functional modularity

(Ornelas-García et al. 2017). Here we ask, whether the potential division of the characiform body plan into multiple modules can help explain their diverse morphologies and novel evolutionary radiations. Specifically, we want to determine (1) whether multiple body shape modules exist in Characiformes, (2) whether those potential modules evolve differently from each other and (3) whether convergent lineages have fewer body shape modules than non-convergent lineages. Answers to these questions will help reveal whether differences in modularity may explain why some characiform lineages converge, while others do not.

Materials and methods

Taxon sampling, morphometric analysis, phylogenetic tree and trophic ecology

Sampling, data collection, and homologous landmarks for the geometric morphometric analysis followed Burns and Sidlauskas (In Review). One to ten specimens (average = 7.1±2.3), dependent on availability, from the 129 species that match a tip on the phylogeny were used. Landmarks were digitized in tpsDig v 2.17

(Rohlf 2015). Individual landmarks were rotated, scaled, and aligned using a least- squares Generalized Procrustes Analysis in the R package geomorph (Adams and

Otárola‐Castillo 2013). Principal component analysis (PCA) in geomorph on the

Procrustes coordinates identified the primary axes of shape variation among species in the dataset. The phylogenetic comparative methods used require representation of each species by a single morphological data point. Since numerous individuals were

66 measured for most species, we calculated the position of the centroid for each species in morphospace, and used those coordinates in subsequent phylogenetic comparative methods. The direction and shape of morphospace occupation was visualized using a phylomorphospace approach (Sidlauskas 2008) as implemented in the R package phytools (Revell 2012).

For comparative analyses we used a fossil time-calibrated phylogeny from

Burns and Sidlauskas (In Review). This phylogeny was based on a four-gene dataset that included 129 species, representing members of 19 of 21 currently recognized families. The phylogeny covers the body shape spectrum observed in Characiformes, from deep bodied serrasalmids to hyper-elongate members of Acestrorhynchidae. We used the same trophic classification here as in Burns and Sidlauskas (In Review), using

SIMMAP (Huelsenbeck et al. 2003) reconstructions to determine the number of independent origins of six different trophic ecologies in the order. Burns and Sidlauskas

(In Revision) determined the trophic guild of each species through a primary literature review of diet in characiform fishes including major research databases including

Academic Search Premier, Web of Science, and Google Scholar. They searched through species descriptions, field guides, and checklists to determine whether aspects of each fish’s diet were discussed and searched the database FishBase (Froese and

Pauly 2000) to find sources that cited trophic data for each species. Each species was coded as a detritivore, herbivore, insectivore, invertivore, omnivore, or piscivore.

Functional modularity

We tested three hypotheses of modularity (Figure 3.1), with each modular hypothesis representing a different combination of three subsets of landmarks corresponding to frequently recognized functional modules in other groups of fishes.

The first potential functional module, the head, is associated with both feeding kinematics (Norton 1991) and locomotion through drag reduction and lift forces

(Bushnell and Moore 1991; Lighthill 1993; Larouche et al. 2015). The second potential module, the precaudal region including the trunk, paired, dorsal, and anal-fins, is associated with stability and thrust (Harris 1938; Drucker and Lauder 2001, 2005;

Standen 2008). The third module, the caudal peduncle, is associated with acceleration and thrust (Webb 1982; Webb 1984a; Gibb et al. 1999). The three modular hypotheses were as follows : (1) a null hypothesis of complete covariation between all regions of the body (Figure 3.1, top); (2) cranium and postcranial axial skeleton modules, (Figure

3.1, center); a pattern previously found in two characiform species (Ornelas-García et al. 2017) and (3) cranial and precaudal portion of body and caudal peduncle modules

(Figure 3.1, bottom) which was found in a family of tropical reef fish (Aguilar‐Medrano et al. 2016). We inferred the best-fitting modular hypotheses in a dataset containing all non-piscivores and in another containing only piscivores, to determine whether deterministic piscivores exhibited fewer modules and less evolutionary modularity than non-piscivores.

In order to assess the degree of covariation between landmark sets we assessed the Covariance Ratio (CR) coefficient (Adams 2016). The CR coefficient measures the degree of modularity between morphological data sets using the pairwise covariances

68 between variables. The modularity test is considered significant when the observed CR coefficient is small compared to a distribution of values obtained on a random association of landmark subsets. We randomly assigned landmark subsets 1000 times to assess significance. The CR coefficient ranges from 0 to 1 corresponding to same degree of covariation within and between modules. Values lower than 1 characterize datasets where the covariation between modules is lower than within modules; and values higher than 1 describe higher covariation between than within modules. Thus, low CR values express low covariation indicating separate modules and high CR values express high covariation between modules indicating integration.

To assess the fit of these models, we used a minimum deviance method that assesses the goodness-of-fit of the covariance matrix derived from each model to the empirical data (Marquez 2008). We ranked modularity hypotheses using the γ∗ statistic of in the program MINT (Márquez 2008), estimated intramodular covariations from the data, and considered the best fitting model to be the one that deviates least from the empirical dataset and has the lowest γ∗ value (Márquez 2008). The significance of each of γ∗ was assessed using a parametric Monte Carlo approach for each model. To test the support for the γ∗ statistic, we assessed the ranking of the γ∗ statistic using 1000 replicates in a jackknife procedure, removing 33% of the dataset in each iteration.

Evolutionary modularity

We analyzed the tempo and mode of modular evolution to determine whether the modules evolved independently of one another or responded to the same selective pressures in a similar way. Variation among modules would suggest higher intrinsic

69 evolvability, which could in turn increase the potential for novel phenotypic evolution.

Specifically, we analyzed disparity through time for each module to determine whether morphological disparity peaked at different times indicating independent evolution, or simultaneously, indicating correlated evolution. We tested a suite of different evolutionary models to determine whether each module evolved under a different selective regime, which would also increase the potential for contingent evolution.

Lastly, we analyzed the rates of module evolution to determine whether the different modules varied in their speed of diversification.

Disparity through time plots were generated in the Geiger v2.0 R package

(Pennell et al. 2014) to examine the temporal pattern of change in morphological disparity along the phylogeny. Disparity through time analyses allow comparison of the observed pattern of intraclade versus among clade disparity through time with the expected pattern under Brownian motion.

We performed model selection analyses in the R package mvMORPH (Clavel et al. 2015) to determine whether each model evolved under a different selective regime. This program fits multiple evolutionary models to multivariate trait data, without the need to analyze each principal component in isolation. The fit of the body shape data to each of several models of diversification was assessed using small sample-size corrected AIC. We included two models of stochastic evolution: a simple model of Brownian motion and mixed model of Brownian motion where the rate of evolution can vary. We tested multiple models of adaptive evolution, including an early burst model and several Ornstein-Uhlenbeck (OU) models designed to distinguish several a priori hypotheses for adaptive body shape evolution.

The three OU models were OU continent, OU trophic, and OU family. Burns and Sidlauskas (In Review) found that body shape diversification differed significantly between New and Old World lineages, thus one of our OU models (OU continent) hypothesized that characiforms diversifying on each continent evolved towards separate adaptive optima. The next model (OU trophic) was a multi OU model hypothesizing a distinct adaptive optimum for each trophic category, directly testing whether species with different trophic ecologies evolved towards different optima, and whether lineages independently evolving similar trophic ecologies converge in morphospace. The next model (OU family) hypothesized a distinct optimum for species in each family, as would be expected if body shape was linked to lineage formation early in the evolutionary history of the order.

We estimated the rates of module evolution for all trophic groups and family clades using the compare.multi.rates function in the R package geomorph (Adams and

Otárola‐Castillo 2013). We determined significance by comparing the observed rate ratios to a null distribution of 1000 rate ratios obtained by simulating evolution across modules at a uniform rate. The proportion of simulated rate ratios that were greater than the observed values were treated as the significance level for each observed rate ratio.

Results

Modularity hypothesis test - non-piscivores

Both the two- and three- module hypotheses are better supported than the one- module hypothesis (Figure 3.2). The two-module hypothesis had an observed CR score of 0.916, which was significantly lower (p=0.004) than the distribution of the simulated

CR scores. The three-module hypothesis had an even lower CR score than the two- module hypothesis (0.833) and the three modules CR score was also significantly

(p=0.001) lower than the simulated distribution. The significance of these two CR scores indicate that the characiform body place includes multiple modules. Comparison of model fits found that the three-module hypothesis fit best, (γ∗=−0.434; 95% confidence interval (CI) = [−0.480, −0.386.]), followed by the two-module hypothesis

(γ∗=−0.3571; 95% confidence interval (CI) = [−0.396, −0.318]), and then the no module null hypothesis (γ∗=0.0, 95% confidence interval (CI) = [−0.0124, 0.124].

None of the jackknife resamplings altered the relative rank of these hypotheses. The single module hypothesis always fit very poorly and was never recovered as the best fit model of modularity. The three-module hypothesis was also supported by the modular hypothesis of evolutionary rates because three-module hypothesis exhibited a greater difference in evolutionary rate ratios between modules (R = 3.82, p = 0.001) than in the two-module hypothesis (R = 2.85, p = 0.001).

Modularity hypothesis test - piscivores

Piscivore bodies also contain three modules (Figure 3.2). The two-module hypothesis had a CR score of 0.938 which was significantly (p=0.04) lower than the simulated distribution. The three-module hypothesis had a lower CR score (0.793) than the two-module hypothesis and the CR score was significantly (p=0.001) lower than the simulated distribution. Like the non-piscivores, the three module hypotheses fit best (γ∗=−0.385; 95% confidence interval (CI) = [−0.393, −0.377], followed by the two-module hypothesis (γ∗=−0.374; 95% confidence interval (CI) = [−0.381, −0.361]),

72 and then the single-module null hypothesis (γ∗=−0.0; 95% confidence interval (CI) =

[−0.0012, 0.0032]). None of the jackknife resamplings altered the relative rank of these hypotheses. The three-module hypothesis was also supported by the modular hypothesis of evolutionary rates with the three-module hypothesis exhibiting a greater difference in evolutionary rate ratios between modules (R = 3.88, p = 0.05) than in the two-module hypothesis (R = 2.91, p = 0.002).

Shape change across modules

The wiregraphs in Figure 3.3 visualize the shape change along each principal component axis for the three modules. For the head module, fishes with a low PC1 score have a large, elongate head, large terminal mouth and a large eye. As PC1 increases, the head becomes much shorter and dorsoventrally flattened and the mouth moves more upwards. Fishes with a low PC2 score exhibit a medium sized head with a large eye and terminal mouth. As PC2 increases the head becomes deeper, the eye shrinks, and the mouth becomes more upturned. For the precaudal module, a low PC1 score indicates a deep body with the dorsal and anal-fins aligned near the middle of the fish. As PC1 increases the body becomes more elongate and the dorsal and anal-fins separate, with the dorsal-fin much more anterior than the posteriorly located anal-fin.

Fishes with a low PC2 score are highly elongate with the dorsal and anal-fins aligned on the posterior portion of the body. As PC2 increases, the body becomes much deeper and the dorsal and anal-fins separate. For the caudal module, fishes with a low PC1 score have short caudal region that is relatively deep, as PC1 increases the caudal peduncle becomes thinner and more elongate. For fishes with a low PC2 score, the

73 caudal peduncle is relatively deep and elongate, but as PC2 increases the caudal peduncle shortens and deepens.

Diversification dynamics among modules

The timing and pattern of morphological disparity differed dramatically among evolutionary modules (Figure 3.4). Morphological disparity in the precaudal module was higher early in cladogenesis and then slowly decreased as lineages filled the morphospace more densely. In contrast, the caudal module showed an increasing level of disparity throughout cladogenesis with a spike of diversification occurring very recently in the evolutionary history of the order. The head module exhibited a relative decrease in disparity over time, following the expectations of Brownian evolution.

Each individual module followed a very different pattern of evolution. The diversification of the head module followed a Brownian motion model of evolution

(Table 3.1). The precaudal module diversified through an early burst model of evolution, in which most of the morphological diversity accumulated very early in cladogenesis (Table 3.1). The diversification of the caudal module was best explained by a two optimum OU model (OU continent), in which the caudal region evolved towards a longer and thinner caudal peduncle (theta of PC1: 0.3) in African lineages versus South American lineages (theta of PC1: 0.02).

Module rate evolution

The head + precaudal + caudal partitioning exhibited the greatest difference in evolutionary rate ratio among modules (

Table 3.2; R= 3.82, p = 0.001). The evolutionary rate for the precaudal module was significantly higher (p=0.001) than either the head or caudal module, indicating that diversification rate in the precaudal module was more pronounced than any of the other two modules.

Discussion

Studies of modularity and integration can shed light on the ability of clades to evolve numerous morphologies and why some clades are more prone to convergence than others. Strong integration among traits has the potential to hinder the number of morphotypes a lineage can occupy, which in turn can promote convergent evolution when adapting to the same environments (Wake 1986; Wake 1991; Hanken and Wake

1993). Conversely, modularity can facilitate diversification by allowing independent morphological units to adapt to different selective regimes or respond to drift differently (Klingenberg 2014). This independence amongst morphological units may make lineages more susceptible to historical contingencies, making deterministic evolution less likely when adapting to the same niche.

As predicted, these fishes appear to be highly evolvable. Non-piscivorous characiforms possess at least three distinct body shape modules, each of which diversified under a distinct evolutionary regime. These modules evolved at different rates, under different selective regimes, and at different times during the radiation. Non- piscivorous lineages occupy a variety of different regions of morphospacee in each module when adapting to trophic ecology, indicating no relationship between trophic ecology and module adaptation in non-piscivorous lineages. Thus, evolutionary

75 modularity likely helps explain why Characiformes evolved numerous unique morphologies and possibly helps explain why ecologically similar characiforms rarely converge in body shape

Characiform body shape made up of three independent functional modules

Our results show that the body plan of Characiformes includes three functional modules: a head, precaudal, and caudal module. The presence of these independent functional modules may explain why Characiformes were able to evolve a variety of morphologies, by allowing the head, precaudal, and caudal regions to respond to different selective pressures. Many fishes across the tree of life have exhibited modular organization (Parsons et al. 2011; Larouche et al. 2015; Aguilar‐Medrano et al. 2016), however, to our knowledge, our study represents the most taxon dense test of clade- wide modularity in fishes, and only the second such study to examine modularity in

Characiformes.

Ornelas-García et al. (2017) previously examined modularity in Characiformes and found just two modules (head and rest of body). Our results identify three functional modules (head + precaudal + caudal peduncle) and likely differ from

Ornelas-García et al. (2017) findings because we include a much greater diversity of characiform species. Ornelas-García et al. (2017) examined just two lake dwelling species, while our dataset includes over 200 species, inhabiting lakes, large river systems, headwater streams, and floodplains (Géry 1977).

Increased modularity has likely increased the potential for many fish lineages to adapt to multiple different selective pressures and evolve morphological novelty. For

76 instance, modularity in the damselfish locomotor structure, including a head, trunk and caudal peduncle module, was likely responsible for the more elongated body shape in

Amphiprion species, an evolutionary novelty among the family (Aguilar‐Medrano et al. 2016). Modularity in the lower jaw of closely related cichlids species differed between biting and suction groups, allowing for adaption to different functional demands (Parsons et al. 2011). Similarly, different species in characiformes appear to take advantage of modularity to adapt different body regions independently in order to invade a huge variety of novel adaptive zones. For example, species in the genus Phago have uniquely elongate caudal peduncles that has allowed them to employ burst-speed hunting in incredibly fast-moving water (Arroyave and Stiassny 2014), where as members of the family Gasteropelecidae have a modified the pectoral girdle which enables burst speed swimming in slow moving water (Burns 2017).

Evolutionary modularity likely promoted novel body shape phenotypes

Increased evolutionary modularity should increase morphological diversification by allowing modules to respond to selection and drift independently

(Klingenberg 2014). Characiform body shape diversification exhibited strong evolutionary modularity (not integration) when invading novel niches, likely increasing the potential for novel body shape evolution. In characiforms, disparity in the cranial module decreased slowly over time following the expectation of neutral evolution, while disparity for the precaudal and caudal modules peaked above a neutral model of evolution, with the precaudal module diversifying early in cladogenesis and the caudal peduncle much more recently. Not only did phenotypic diversity peak at different times

77 in different modules, but it evolved under different selective regimes and at different rates. This lack of similarity indicates that the different functional modules not only evolved independently of one another, but responded to selection and drift in vastly different ways.

Characiform patterns of disparity between modules differ significantly from those observed in convergent cichlids, which have highly integrated body regions

(Feilich 2016). For example, disparity in cichlid locomotor structures including the body, medial, and caudal-fins peaked simultaneously following major species and trophic diversification events (Feilich 2016). Malagasi cichlids exhibit correlated evolution and integration between jaw and body shape, with changes in overall body shape predicted by changes in jaw shape (Martinez and Sparks 2017). The correlated evolution between trophic and different locomotor structures likely increased the potential for deterministic evolution across different cichlid lineages when invading similar niches because of integration through phenotypic bias and constraint.

Conversely, lineages with low amounts of integration, i.e. modularity, are hypothesized to exhibit contingent phenotypic evolution, especially when lineages vary in their initial phenotype or their genetic variation. (Losos 2011; Losos 2017). Thus, radiations like the characiforms, which are highly modular and spatiotemporally widespread, are possibly prone to evolving novel morphologies even when they converge in ecology. For instance, independent lineages of characiforms including certain members of the Anostominae and certain species in the Crenuchidae, have converged on a benthic, rheophilic, invertivorous lifestyle (Lujan and Conway 2015).

Overall body shape differs between these lineages even though they share the same

78 general morphology of the trunk and caudal peduncle (Burns and Sidlauskas, In

Revision), because Anostominae genera such as Synaptolaemus have evolved an upturned versus inferior mouth (Myers and de Carvalho 1959; Géry 1977). The variation in mouth position most likely evolved because Synaptolaemus have a unique benthic feeding strategy where they scrape invertebrates from the underside of large flat rocks (Lujan and Conway 2015), while crenuchids feed on benthic invertebrates on the surface of sand or gravel (Géry 1977; Buckup 1993b). These two lineages did not converge because of adaptation in a single module shifting the position and shape of the jaw in Synaptolaemus¸ without modifying the remainder of the body. Adaptation in a single module was only possible because oral jaw evolution was not correlated with or constrained by body shape, as is the case in Malagasy cichlids (Martinez and Sparks

2017).

Natural selection not morphological integration likely responsible for convergent piscivores

Convergence can be caused by adaptation through natural selection (Simpson

1953; Harmon et al. 2005; Losos 2011), morphological constraints increasing redundant morphotypes (Smith et al. 1985; Losos 2011), or a combination of the two

(Sanger et al. 2012). In characiforms, the strong convergence among piscivorous lineages prompted the hypotheses of lower modularity (Burns and Sidlauskas, In

Revision). Yet, all piscivorous and non-piscivorous characiformes possess three distinct body shape modules that diversified independently. Therefore, higher

79 integration does not explain convergence in piscivorous groups, but strong selection on one or more of the body shape modules is could be responsible.

When examining each module alone, we found that only the precaudal module in piscivorous lineages converged, with all piscivorous lineages exhibiting anteroposterior elongation in the precaudal portion of their body. It is likely that the deterministic pattern of evolution in piscivorous lineages observed in Burns and

Sidlauskas (In Revision) was driven by intense natural selection to anteroposteriorly elongate the precaudal region of the body. This is not surprising because the precaudal portion of the body is related to locomotion with an elongated shape increasing speed and maneuverability; a morphology needed for a piscivorous lifestyle, regardless of habitat or specific feeding strategy (Webb 1982; Webb 1984a).

The head and the caudal region of piscivorous lineages did not exhibit convergent evolution and were much less restricted in morphospace, although not to the same level as was observed in non-piscivores. Caudal peduncles ranged anywhere from highly elongate to relatively short and deep, while craniums could be highly elongate with large mouths, or shorter and more robust, with smaller mouths. Though one would think that mouth and head shape would converge on a single optimum in highly predatory species, we found evidence to the contrary. The substantial variation in mouth morphology likely reflects the variation in attack strategies and habitat ecology among piscivorous lineages. Piscivore feeding strategies include scale feeders

(Vieira and Géry 1979), top water ambush predators (Montaña et al. 2011), roving predators (Menezes 2003), and benthic ambush predators (de los Angeles Bistoni et al.

1995). A more detailed morphometric analysis on the body plan of piscivorous

80 lineages, including head, jaw, and eye size, may offer more insight into how variation in feeding ecology has shaped aspects of phenotypic evolution in piscivorous lineages.

Conclusion

Characiformes posssess three independent body shape modules that diversified at different times and rates while under different selective regimes. This high evolutionary modularity plausibly explains why many unique body shapes evolved, when lineages evolved similar trophic ecologies across the radiation. Even the convergent piscivores exhibited strong evolutionary modularity and their convergence was likely driven by strong natural selection to elongate the precaudal module, not through whole-body integration and phenotypic constraints. The generality of the relationship between evolutionary modularity and evolvability and convergence has not been well studied in vertebrate lineages. More studies need to look at the role that evolutionary modularity and integration can have in shaping patterns of evolution across broad and restricted radiations of vertebrates.

Acknowledgements

We thank Devin Bloom and Hernan Lopez-Fernandez for insightful discussions. We thank my dissertation committee of David Maddison, Guillermo Ortí, and Tiffany Garcia for helpful comments and discussions. We thank Selene Fregosi for helpful comments on the manuscript. We gratefully acknowledge the work of numerous collection managers and curators who made materials available, including

Mark Sabaj (ANSP), David Werneke and Jon Armbruster (AUM), Mike Retzer

(INHS), Sandra Raredon and Richard Vari (USNM), Kevin Swagel and Susan Mochel

(FMNH), Lucia Rapp Py-Daniel and Renildo Oliveira (INPA), and Karsten Hartel and

Graduate Fellowship and an ASIH Raney Award provided support to MDB.

Figure legends

Figure 3.1. Landmark schematic of three hypothesized subdivisions of 23 landmarks into modules. (a) Head module comprised of 11 landmarks; (b) precaudal module comprised of 8 landmarks; and (c) caudal module comprised of 4 landmarks.

Hypothesis 1 is full integration, Hypothesis 2 is independent evolution of the head

83 module and of the rest of the body (precaudal + caudal module), and Hypothesis 3 is independent evolution of the head, precaudal and caudal module. Landmark configuration follows Burns and Sidlauskas (In Review).

Figure 3.2. Results of phylogenetic modularity analysis of the head, precaudal, and caudal region for piscivorous and non-piscivorous characiforms using the covariance ratio (CR) coefficient showing a significant degree of modularity.

Figure 3.3. Phylomorphospace depicting contingent evolution of each trophic group in each module.

Figure 3.4. Accumulation of multivariate disparity through time for each module. Black line equals observed data, dotted line indicated the expectation of Brownian motion, and shading denotes 95% confidence intervals.

Figure 3.5. Schematic illustrating how evolutionary modularity can increase evolvability and how integration can increase convergence.

Tables

Table 3.1. Results of the multivariate model-fitting analyses for each module. For each model, the number of parameters (P), the Akaike information criterion (AIC). The best model has the lowest AIC.

Model Rank AIC Diff Head Brownian motion 1 -1127 0 OU continent 2 -1107 20.6 Early burst 3 -1101 26.9 OU trophic 4 -1081 46.7 OU family 5 -1012 115.4 Precaudal Early burst 1 -381 0 Brownian motion 2 -369 12 OU continent 3 -354 27.5 OU trophic 4 -339 42.3 OU family 5 -309 72.5 Caudal OU continent 1 -383 0 OU trophic 2 -375 8 OU family 3 -348 35 Brownian motion 4 -328 55 Early burst* - - -

Table 3.2. Rates of module evolution for non-piscivores and piscivores showing the almost four times faster rate of evolution in the precaudal module.

2 2 2 σ Head σ Precaud σ Caudal R Pval Non-piscivores 1.9 x 10-6 7.38 x 10-6 2.5 x 10-6 3.82 0.001 Piscivores .003 .012 .004 3.88 .05

CHAPTER 4: TEMPO OF LINEAGE AND BODY SHAPE DIVERSIFICATION IN

A HYPERDIVERSE FRESHWATER FISH RADIATION

Michael D. Burns and Brian L. Sidlauskas

Abstract

Adaptive radiations are spectacural displays of ecolological and evolutionary diversity that evolve rapidly. During adaptive radiations, early bursts of morphological and lineage diversification occur after an increase in ecological opportunity caused by physical invasion of novel adaptive zones or the evolution of a key innovation.

Characiform body shape diversified in a manner consistent with adaptive radiation, with disparity peaking early in cladogenesis, however, previous studies have not reconstructed rates of lineage or morphological diversification. Here we reconstruct those rates using a Bayesian Analysis of Macroevolutionary Mixtures (BAMM). We determine that body shape elongation evolved fastest early in cladogenesis, a pattern that is consistent with an ancient adaptive radiation. While we also infer accelerated lineage diversification early in cladogenesis, simulations indicate that this pattern can also result from pruning a large phylogeny that evolved under a constant rate of evolution, indicating that the empirical pattern probably results from a sampling artifact. Characifomes show some rate signatures indicative of an adaptive radiation, but much higher taxon sampling is needed to fully understand the pattern of diversification in the order.

Introduction

During adaptive radiations, clades speciate rapidly while evolving spectacular ecomorphological diversity (Simpson 1953; Schluter 2000; Glor 2010). Much of life’s most remarkable diversification events, including Crater Lake cichlids (Kocher et al.

1993; Muschick et al. 2012; Machado-Schiaffino et al. 2015), Anolis lizards (Losos et al. 1998; Harmon et al. 2005; Mahler et al. 2013), and Hawaiian Honey Creepers

(Lerner et al. 2011), arose through classic adaptive radiations, which occur in unoccupied “island” ecosystems (Schluter 2000). Sticklebacks in postglacial lakes provide a well-known aquatic case study, in which a single founder diversified rapidly across the benthic to pelagic habitat axis in multiple lake ecosystems (Schluter 1995;

Rundle et al. 2000). Increases in ecological opportunity catalyze the early bursts of morphological and lineage diversification that typify adaptive radiations (Simpson

1953; Schluter 2000; Givnish 2015). Many processes can increase ecological opportunity, such as functional innovations that allows the invasion of new niche space, as seen in the fusion of the cichlid pharyngeal jaw which allowed for novel prey consumption (Liem 1973; Hulsey et al. 2006; Hulsey 2006). The physical invasion of a new environment with unoccupied niches can also increase ecological opportunity, as seen in many island radiations including Hawaiian Honey Creepers and Crater Lake

Cichids (Lerner et al. 2011; Muschick et al. 2012; Mahler et al. 2013; Muschick et al.

2014). Regardless of the evolutionary process that increases ecological opportunity, adaptive radiations follow a predictable pattern that can be detected with modern phylogenetic comparative methods (Schluter 2000; Gavrilets and Losos 2009; Glor

2010).

More recently, many studies on widespread continental radiations, such

Neotropical cichlids (López‐Fernández et al. 2013) and New World ratsnakes

(Burbrink et al. 2012), have shown that continental radiations diversify through adaptive radiations. However, the evolutionary patterns of continental radiations, which include richer biotas and more complex ecosystem interactions, may differ from patterns in the classical insular systems (Schluter 2000; Claramunt 2010; Harmon et al.

2010). For example, Burbrink et al. (2012) found decoupled patterns of species and morphological diversification in New World ratsnakes, which led to clades with high species diversity, but limited ecological variability. Thus, it is important to study adaptive evoluton in widespread diverse radiations in order to understand how diversification occurs among diverse biotas.

The South American and African freshwater fish order Characiformes exhibits remarkable species richness and ecomorphological diversity and provides an excellent example of a continental-spanning diversification. Comprising more than 2000 species distributed primarily in South America and Africa, characiforms vary dramatically in body shape, jaw morphology, and tooth morphology. Recent macroevolutionary analyses of Characiformes suggest that patterns of morphological divergence may bear the signature of an ancient adaptive radiation (Guisande et al. 2012; Burns and

Sidlauskas, In Revision). Much of the ecomorphological segregation occurred very early in the radiations history with overall body shape disparity (Burns and Sidlauskas,

In Review) and trophic morphology (Guisande et al. 2012) partitioned among larger, early diverging clades that comprise modern day families. Although general patterns of morphological divergence appear to match the predictions of adaptive radiation, no

94 study has explictly tested whether rates of lineage and morphological diversification peaked early in Characiformes.

Adaptive radiations follow three distinct and testable patterns (Schluter 2000;

Gavrilets and Losos 2009; Glor 2010). First, lineages diversify through early bursts, with lineage diversification rates then decreasing as speciation slows (Harmon et al.

2003; Gavrilets and Losos 2009). Lineage diversification is high early in an adaptive radiation because niche space is open and slows down as niche space becomes occupied

(Rabosky 2009; Losos 2010). Next, early bursts of morphological evolution followed by a decrease in morphological diversification rate characterize adaptive radiatins because when niche space is open, lineages will rapidly adapt across multiple ecological axes, thereby increasing phenotypic diversification (Losos and Mahler 2010;

Mahler et al. 2010). The diversification rate slows down rapidly as niche space becomes occupied. Thirdly, shifts in the rate of morphological and lineage diversification coincide in adaptive radiations, because lineages speciate while adapting across multiple ecological axes (Schulter 2000; Gavrilets and Losos 2009; Glor 2010).

Macroevolutionary studies depend upon analyzing the diversification rates of different radiations (Ricklefs 2007), especially when analyzing the origins of biodiversity through adaptive radiations (Schluter 2000; Glor 2010). However, the best way to estimate diversification rates remains contested. A relatively new method called

Bayesian Analysis of Macroevolutionary Mixtures (BAMM) has become widely used in many empirical studies that reconstruct diversification rates (Marin and Hedges

2016; Burress et al. 2017; Rossi et al. 2018). BAMM remains the best available and most widely used method for analyzing diversification rates in non-model systems.

However, recent studies have highlighted potential flaws in BAMM that can lead to spurious estimations in the diversification rate if not correctly accounted for

(Moore et al. 2016; Meyer and Wiens 2018). Specifically, Moore et al. (2016) found that BAMM incorrectly estimated diversification rates when the tree was small ~50-

150 species and Meyer and Wiens (2018) found that BAMM overestimated the diversification rate in clades that contained less than 150 taxa. These studies highlight the importance of using simulations to test the accuracy of empirical results when using

BAMM, especially when the phylogenetic tree contains fewer than 150 species.

Herein, we use a fossil calibrated molecular phylogeny and geometric morphometric dataset to assess the rates of lineage and morphological diversification in Characiformes. Specifically, we use a heterogeneous mixture of diversity-dependent and constant rate diversification regimes to test whether (1) Characiformes show evidence of early bursts of lineage and morphological diversification followed by a decrease in rates of divergence, and (2) whether morphological and lineage diversification rates correlate with one another. Since our phylogenetic tree contains fewer than 10 percent (less than 150 species) of the known Characiformes species, we also performed a suite of simulations to determine whether our small empirical tree had enough power to accurately recover the true diversification rates.

Materials and methods

Diversification dynamics

We modeled diversification dynamics of the order Characiformes as a heterogeneous mixture of diversity-dependent and constant rate diversification regimes

96 in the Bayesian modeling framework BAMM (Rabosky 2014). BAMM estimates speciation, morphological evolution and extinction rates separately, instead of single diversification rate (see Rabosky 2014; Rabosky et al. 2014 for further details). BAMM employs a Markov Chain Monte Carlo process to estimate the posterior distribution of the locations of regime shifts. Rate regimes may vary over time and rate shifts may occur anywhere on the phylogeny. BAMM models the number of total rate shifts across the tree under a compound Poisson process and makes no a priori assumptions about the locations of regime shifts.

BAMM 2.5 was used to run three separate analyses with different starting seeds for 25 million generations each. Then, BAMM was run across the entirety of the posterior distribution of trees recovered from the BEAST 2.0 analysis completed by

Burns and Sidlauskas (In Revision). The appropriate priors were determined using setBAMMpriors function in the R package BAMMtools 2.0.2 (Rabosky et al. 2014b).

We used BAMMtools to check for convergence with log-likelihood plots and ensured effective sample size of the log-likelihood and the number of shift events were both above 200. Incomplete taxon sampling was determined analytically by providing the proportion of species missing for different characiform clades following the number of species in each clade sensu Oliveira et al. (2010). To examine whether the posterior distribution sampled by BAMM was sensitive to the prior distribution of the proposed number of rate shifts (see Moore et al. 2016), as specified by the hyperprior parameter expectedNumberOfShifts, we ran BAMM using expectedNumberOfShifts of 1,5,10, and 20.

We assessed the evidence for multiple shifts using Bayes factors with

BAMMtools’ function computeBayesFactors. For each analysis, we extracted the credible set of rate shifts using a threshold of 20 Bayes Factors, and visualized the set of rate shifts with the highest posterior probability. We identified the most likely shift locations by analytically removing shift configurations with shifts that occurred at very low frequencies in the posterior following Rabosky et al. (2014b). To do this, we calculated the prior and posterior probability of models with rate shift and models with no rate shifts for each branch, which were used to compute the marginal odds. The function credibleShiftSet was run to calculate the 95% credible set of shift configurations and the getBestShiftConfiguration function identified the shift configuration with the highest posterior probability. Finally, we plotted the estimated speciation rates through time using the plotRateThroughTime function.

Trait-dependent diversification

We used Structured Rate Permutations on Phylogenies (STRAPP; Rabosky and

Huang 2015) within BAMMtools to test for correlation between rates of speciation and rates of trait evolution using the traitDependentBAMM function. STRAPP first requires estimation of speciation rates with no knowledge of character states. Then, a test statistic (Spearman’s rank correlation) quantifies the association between estimates of speciation rate and trait evolution rate at the tips of the tree. Finally, STRAPP compares the value of the test statistic to a null distribution created by structured permutations of evolutionary rates across the tree, which accounts for the nested inheritance of rate shifts from parent to offspring lineages (Rabosky and Huang 2015).

We first used the getTipRates function in BAMMtools to obtain instantaneous estimates of rate evolution for each tip, and then took the mean rate across the posterior for each principal component axis. We calculated Spearman’s rank correlation coefficient and performed a one-tailed statistical test for positive correlation between estimated speciation rates and rates of trait evolution for each of the first four principal components.

Simulations: Speciation rates

We designed a simulation to determine whether BAMM tends to infer spurious changes in evolutionary rate as an artifact of low taxon sampling. We simulated phylogenetic trees 2000 times under a stochastic constant rate birth-death process with a speciation rate of 0.1 and death rate of 0 using the function pbtree in the R package phytools (Revell 2012). We grew the trees until each contained 2000 tips, matching the hypothesized species richness of a complete characiform phylogeny (Froese and Pauly

2000; Eschmeyer et al. 2016). We scaled the tree length to 100 million years to match the hypothesized age of Characiformes from Burns and Sidlauskas (In Revision) and then randomly pruned each tree to 129 tips, matching the size of our empirical tree.

Incomplete taxon sampling was determined analytically by providing the proportion of tips present in the pruned phylogeny as a prior at 6.5%. We determined appropriate priors for each tree using setBAMMpriors function in the R package BAMMtools 2.0.2

(Rabosky et al. 2014b). Finally, we modeled diversification dynamics of the large 2000 tip simulated phylogeny and the smaller pruned 129 tip phylogeny as a heterogeneous

99 mixture of diversity-dependent and constant rate diversification regimes in BAMM

(Rabosky 2014).

Simulations: Morphological diversification rates

We simulated morphological data under a Brownian motion model of evolution with a sigma value of 0.02 using the sim.char function in the R package Geiger 2.0

(Pennell et al. 2014). We simulated the data on the full 2000 tip phylogenetic tree and then only used the simulated morphological data that matched a tip on the 129 taxa pruned phylogenetic tree outlined above. Incomplete taxon sampling was determined analytically by providing the proportion of tips present in the pruned phylogeny as a prior at 6.5%. We then ran BAMM as outlined above on both the full and pruned simulated datasets.

Results

Speciation rate

BAMM reconstructed a slow decline in speciation rates across the history of

Characiformes. The fastest rate of speciation (0.12) occurred at the base of the phylogeny during the earliest part of cladogenesis (Figure 4.1). The speciation rate decreased steadily by a magnitude of 3 thereafter, with the slowest rate of speciation occurring near modern time (0.04). There were no detectable discrete rate shifts (Figure

4.1).

100

Morphological evolution rate

We infer different rates of evolution for each principal component, and different temporal models for the first component versus the three remaining components

(Figure 4.1). The decline in the rate of morphological evolution on PC1 mirrors a similar decline in speciation rate, with the fastest rate of PC1 evolution occurring very early in cladogenesis (1.5 x 10-4), followed by a magnitude 2.5 times decrease in the rate overtime to a rate of 5.0 x 10-5 There were no detectable discrete rate shifts in PC1 with both the best fit configuration and set of 95% credible configurations exhibiting no discrete rate shift regimes.

The rates of evolution for the remaining principal components (PC2, PC3, and

PC4) differed from both the rate of PC1 and the speciation rate. PC2, PC3, and PC4 all exhibited the same rate pattern; an increase in the rate of morphological evolution over time, with the fastest rates of evolution occurring relatively recently in the history of the order. However, the rate of change for PC2, PC3, and PC4 is very minimal, almost a magnitude lower than the rate of change observed in the speciation and PC1 rate

(Figure 4.2).

PC2 and PC3 exhibited no rate shifts in the best fit configuration, but some of the scenarios in the set of 95% credible rate shifts did exhibit detectable shifts (Figures

4.3 and 4.4). In PC2, after filtering noncore shift locations with low marginal odds, we identified nine distinct shift configurations that were sampled more than expected based on the prior alone (Figure 4.3b). These configurations cumulatively accounted for 95% of the posterior probability (Figure 4.3b). Four of the nine shifts showed an increase in the rate at the base of the clade containing Iguanodectidae and Acestorynchidae, one

101 occurred within the Alestidae, and three occurred in three different lineages within

Characidae sensu stricto.

In PC3, after filtering noncore shift locations with low marginal odds, we identified seven distinct shift configurations that were sampled more than expected based on the prior alone (Figure 4.4b). These configurations cumulatively accounted for 95% of the posterior probability and were spread out all over the phylogenetic tree

(Figure 4.4b). Three of the seven rate shifts indicating elevated rates in different lineages in Characidae, two shifts occurred in Alestidae, and one shift occurred in

Crenuchidae. One shift occurred at the base of Anostomoidea, indicating slow rate of evolution, while another shift indicated a slowdown in the rate at the base of the

Anostomoids + Serrasalmidae + Parodontidae + Hemiodontidae. PC4 exhibited no rate shifts in both the best fit configuration or in any of the 95% credible rate shifts (Figure

4.1).

Correlated evolution

All of the rates of principal component evolution were very weakly (ρ<0.072) and not significantly correlated with speciation rate (Table 4.1).

Simulations: Speciation rate

We found strong evidence for heterogeneity in speciation rates across the pruned simulated phylogenies. Both the pruned and complete tree exhibited evidence of a decrease in the speciation rate overtime (Figure 4.5). However, the magnitude of the rate difference was substantially different between the complete and pruned

102 phylogeny (Figure 4.5). Reconsructions using the complete phylogeny exhibited a minimal rate decrease from 0.083 at the root of the tree to 0.072 at the tips, correctly reconstructing the underlying constancy of the stochastic birth-death process, albeit slightly underestimating the true speciation rate of 0.10. However, the rate reconstruction using the pruned phylogenetic tree exhibited a rate decrease from 0.12 at the root of the tree to 0.04 at the tips, which deviated greatly from the simulated stochastic constant birth-death process.

Simulations: Morphological evolution rate

We found little evidence for heterogeneity in morphological evolution rates using the complete or the pruned versions of the simulated phylogeny (Figure 4.6). The pruned phylogeny indicated a slight increase in the morphological evolution rate over cladogenesis with the slowest rate (0.011) occurring at the base and the fastest rate occurring at the tips (0.014). The full phylogeny showed a similar pattern of very gradual increase over time, with the slowest rate occurring at the base of cladogenesis

(0.020) and the fastest rate at the tips of the phylogeny (0.021). Reconstructions using the full phylogeny almost perfectly matched the simulated sigma of 0.02, while those using the pruned tree slightly underestimated the true diversification rate. Though morphological reconstructions using the pruned and full phylogenies both show a change in the overall rate of morphological evolution throughout cladogenesis, the magnitude of change is minimal and matches what would be expected under our simulation of stochastic evolution.

103

Discussion

During adaptive radiations, spectacular displays of ecomorphological diversification evolve rapidly. Studies of textbook adaptive radiations, including Anolis lizards and Crater Lake cichlids, have led to the predictions that other adaptive radiations will also experience early bursts of lineage and morphological diversification that slow down as niches fill towards capacity and that shifts in lineage diversification rate will correlate with shifts in morphological diversification rate. For Charciformes, we reconstructed exhibited rate signatures matching the predictions of classical adaptive radiations, including much faster rates of lineage diversification and morphological diversification early in cladogenesis. However, constant-rates simulations indicate that the heightened rates of lineage difersification early in cladogenesis may result from pruning, and the pattern in the empirical data is most likely an artifact of our sampling regime, and not indicative of a truly elevated rate.

Conversely, the pruning procedure did not affect reconstructions of morphological rate using simulated data, suggesting that the elevated rates of morphological evolution early in characiform cladogenesis are real, and not artefactual.

Tempo of morphological diversification

The evolutionary history of the most important axis of characiform body shape matches one of the hallmarks of morphological evolution during an adaptive radiation

Harmon et al. 2003; Gavrilets and Losos 2009). PC1, which represents body shape elongation and depth, exhibited faster rates of diversification early in cladogenesis, with the rates quickly slowing down by a magnitude of 2.5 over time. These patterns

104 did not vary among clades. Burns and Sidlauskas (In Revision), also found that this axis of body depth and elongation exhibited very high disparity at the dawn of cladogenesis, which reduced as new lineages filled in morphospace later in the clade’s history. The segregation of overall body shape diversity among larger clades representing modern day families, (Burns and Sidlauskas, In Revision) is fully consistent with this ancient adaptive radiation scenario.

Morphological divergence in the other principal components, including a secondary axis of elongation, the size and position of the jaw and length of the caudal peduncle, do not show distinct signatures of a classic adaptive radiation. Rather, the rate of morphological diversification in these other characters is consistent with a constant rate model of evolution. Thus, if Characiformes did undergo an ancient adaptive radiation, these axes may not have been involved. The recent morphological diversification of these characters could still represent the early stages of an adaptive radiation with the overshooting not yet evident (Hulsey et al. 2010). It is also possible that the diversification occurring in these morphological attributes represent recent pulses of adaptation in response to niche adaptation. However, the pattern in PC2, PC3, and PC4 matches what it is seen in our simulations and cannot be distinguished from

Brownian motion.

It is somewhat surprising that our analyses did not detect discrete morphological rate shifts within Characformes, because some families including the Alestidae,

Anostomidae, Characidae, and Distichodontidae exhibit substantial ecomorphological variation in the position of the mouth, size of the eye, and the length of the caudal peduncle (Burns and Sidlauskas, In Revision), while others such as the

105

Acestorhynchidae and Ctenolucidae do not. Intuitively, one would expect discrete rate changes at the bases of these morphological diverse and depauperate families. It is likely that the relatively low taxonomic sampling used in our study is making secondary diversification events difficult to determine and increased taxonomic sampling might infer rate shifts at the base of ecomorphologically diverse clades indicating secondary diversification events. And indeed, a minority of trees in the posterior distribution do indicate rate shifts at the base of diverse clase on PC2 and PC3, even though these shifts do not occur across the majority of the posterior. Future studies with greater taxon sampling may better reveal this finer scale of morphological rate variation, if it does in fact exist.

Performance of BAMM under sparse taxon sampling

Studies on the rates of lineage and morphological evolution are highly sensitive to taxon sampling. Unequal or low taxon sampling can cause overestimation or underestimation of rates of diversification, leading to spurious conclusions about the process of evolution (Ackerly 2000; Heath et al. 2008). Recently, Moore et al. (2016) highlighted several potential flaws in BAMM related to its performance on small phylogenies. They used simulations to test the accuracy of estimated diversification rates from BAMM with simulated trees that were complete but relatively small, with

~50–150 species. They found that BAMM gave accurate estimates of diversification rates when true rates were constant, but relatively inaccurate estimates when rates varied across the tree. Although our taxon sampling was relatively equal among clades, the overall sampling for the order was low with only 129 of the ~2000 known species

106 sampled (~6.5%). If BAMM also produces analytical artifacts under sparse taxon sampling of a large tree, those artifacts may limit the program’s power to detect meaningful rate shifts across the order (Moore et al. 2016; Meyer and Wiens 2018) and likely led to incorrect rate estimates.

Our results are concordant with those of Moore et al. (2016) in showing that

BAMM may behave problematically, but the specific problems we highlight differ from those that they identified. We find that when the phylogenetic tree is significantly pruned, BAMM infers an elevated rate of lineage diversification early in cladogenesis, even when the simulated speciation rate is constant throughout time. The rates early in cladogenesis are overestimated (early rates can be more than 3 times faster than the true rate of diversification). Meyer and Wiens (2018) found that BAMM overestimated the rate of diversification in small clades, i.e. clades with fewer than 150 species had a faster rate of diversification than the true simulated rate. It is likely that significant pruning results in many more small clades, this is especially true in our simulations, leading BAMM to overestimate the diversification rate early in cladogenesis and underestimate the rate later in cladogenesis. The impact of the substantial errors in estimating diversification rates for small clades may be overlooked if accuracy is assessed on a per-branch basis as proposed by Moore et al (2016) and Rabosky et al.

(2017), instead of on a per clade basis. However, more simulations, including a pruning regime to create both large and small clades is needed to determine if variation in clade size causes the spurious rate estimates in our study.

BAMM appears to infer the morphological evolution rate more accurately, even when the tree is significantly pruned, and does not infer a spurious major change in

107 morphological rates as an artifact of the pruning. However, BAMM still underestimates the true rate of morphological evolution in the pruned phylogeny. The full simulated phylogeny accurately recovered the diversification rate of 0.02, with little to no change in the rate over time. Unlike the speciation rate, the morphological diversification rate in the pruned phylogeny increased slightly from 0.11 to 0.15, with an overall increase of 30%. However, the recovered rate of morphological diversification in the pruned phylogeny was much closer to the true rate than what was seen in the speciation simulations. The morphological evolution rate was likely more accurately recovered because the process by which morphology evolved through Brownian motion is independent of taxon sampling (Ackerly 2000), reducing the impact of pruning the phylogeny in the rate reconstructions. If the pruning does bias the analysis towards a shift in morphological rates, it appears to incorrectly accelerate rates close to the present, not at the root of the phylogeny. Thus, the higher rate of evolution on PC1 that we reconstruct near the root of Characiformes does not likely result from the pruning process, and probably represents true biological signal.

Conclusion

Characiformes exhibit a signature of higher rates of morphological evolution early in cladogenesis, which matches one important prediction of an ancient adaptive.

In particular, the evolution of body shape depth and elongation on PC1 exhibits the rate patterns that typify adaptive radiation. Evolutionary patterns of other morphological characters, including size and position of the jaw, eye size and length of the caudal peduncle, are indistinguishable from Brownian motion. Though BAMM also

108 reconstructs elevated rates of speciation near the root of the characiform phylogeny, constant rate simulations on full and pruned phylogenies reveal that low taxon sampling can bias analysis toward reconstructing such a pattern artifactually. Thus, there is currently no compelling evidence for an early burst of speciation in this clade, though future studies with increased taxon sampling may improve our ability to uncover discrete lineage and morphological diversification rate shifts. Characiformes may have diversified under an ancient adaptive radiation, but further investigation of this possibility wll demand a much more complete phylogeny, and morphological data on many more of the 2000 species that comprise this stunningly diverse clade of fishes.

Acknowledgements

We thank Dan Rabosky, Pascal Title, and Michael Alfaro for help with the early stages of the BAMM analysis. We thank Bill Ludt for insightful discussions on rate analyses and the limitations of those methods in comparative studies. We thank my dissertation committee of David Maddison, Guillermo Ortí, and Tiffany Garcia for helpful comments and discussions. We thank Selene Fregosi for helpful comments on the manuscript. NSF grant DEB – 1257898 awarded to BLS provided support to BLS and

MDB. An OSU Provost’s Fellowship, OSU Department of Fisheries and Wildlife M.

A. Ali Graduate Chair Award in Fisheries Biology, a Smithsonian Graduate Fellowship and an ASIH Raney Award provided support to MDB.

109

Figure legends

110

Figure 4.1. Relative rates of speciation or morphological evolution mapped onto the

MCC phylogeny of Characiformes. Colors at each point in time along branches indicate the relative instantaneous rate of speciation or phenotypic evolution. Evolutionary rates are averaged across all evolutionary regimes sampled from the posterior. The evolutionary rate shift configuration with the highest posterior probability is shown for each character. Line drawings represent shape change for each Principal Component from Burns and Sidlauskas (In Review).

111

Figure 4.2. Instantaneous rates of speciation and morphological evolution for

Characiformes. Shading around each curve are the 90% credibility intervals from the posterior distribution of BAMM results.

112

Figure 4.3. Evolutionary dynamics of PC2 diversification in Characiformes inferred from BAMM. Depicted are the (a) mean branch specific evolutionary rates inferred from BAMM representing the core shift configuration with the highestposterior probability. (b) The 95% credible set of shift configurations. These nine distinct shift configurations account for 96% of the posterior distribution. Black circles represent rate shifts. Each branch is color coded based on the mean of the marginal posterior density of the evolutionary rate for their respective BAMM analyses. Blues indicate slower evolutionary rates. Reds indicate faster evolutionary rates.

113

Figure 4.4. Evolutionary dynamics of PC3 diversification in Characiformes inferred from BAMM. Depicted are the (a) mean branch specific evolutionary rates inferred from BAMM representing the core shift configuration with the highestposterior probability. (b) The 95% credible set of shift configurations. These nine distinct shift configurations account for 96% of the posterior distribution. Black circles represent rate shifts. Each branch is color coded based on the mean of the marginal posterior density of the evolutionary rate for their respective BAMM analyses. Blues indicate slower evolutionary rates. Reds indicate faster evolutionary rates.

114

Figure 4.5. Speciation dynamics for the large and pruned simulated phylogenetic trees.

Depicted are the relative rates of speciation mapped onto the simulated phylogeny and the instantaneous rates of speciation through time.

115

Figure 4.6. Morphological evolution dynamics for the large and pruned simulated phylogenetic trees. Depicted are the relative rates of morphological evolution and the instantaneous rates through time of morphological evolution.

116

Tables

Table 4.1. Results of the Spearman’s rank correlation test showing a weak and non- significant relationship between speciation and morphological diversification rate.

r P-value PC1 0.013 0.91 PC2 0.071 0.85 PC3 0.032 0.87 PC4 0.0016 0.94

117

CHAPTER 5: DISCUSSION

Overview

Overall, my dissertation aimed to determine whether characiform body shape and lineage diversification occurred through an adaptive radiation. To test this, I estimated the first time-calibrated molecular phylogeny for the order Characiformes, assembled the first geometric morphometric body shape dataset spanning the order, and compiled an exhaustive trophic ecology database. I combined these datasets and used phylogenetic comparative methods to test multiple different evolutionary outcomes.

In my second chapter, I tested whether adaptive evolution drove members of distinct trophic guilds towards a global optimum for each guild and reconstructed body shape diversification in the Old and New World radiations. I found that members of each trophic guild did not converge on a global optimum, except for piscivores, which converged strongly. Members of other guilds evolved many differnet morphologies despite converging ecologicallyFurthermore, I found that body shape diversification between the Old and New World radiations followed very different pathways, with the

New World radiations occupying twice as much morphospace as their Old World counterparts. Both radiations exhibited higher morphological disparity than would be expected under a model of Brownian evolution early in cladogenesis, which matches one expectations of an adaptive radiation.

In my third chapter, I tested whether evolutionary modularity might have increased the clade’s evolvability, particularly among non-piscivores, which show little convergence. I found that characiform body shape was comprised of three independent modules that diversified at different times and rates while under different selective

118 regimes. I postulate that the high evolutionary modularity observed might partially explain why members of most trophic guilds evolved widely variable morphologies, rather than converging on a global optimum, when lineages evolved similar trophic ecologies across the radiation. The generality of the relationship between evolutionary modularity and evolvability has not been well studied in vertebrate lineages. More studies need to examine how evolutionary modularity and integration can shape patterns of evolution across broad and restricted radiations of vertebrates.

In my fourth chapter, I analyzed the rates of lineage and morphological diversification to determine whether Characiformes exhibited an early burst of speciation and morphological evolution as predicted by classical adaptive radiations. I found that the rate of morphological evolution on PC1 was very high early in cladogenesis and quickly slowed down, following a pattern predicted by adaptive radiation. Initial reconstructions also indicate declining rates of speciation as the readiation progresses, but evolutionary simulations indicated that heavily pruning the tree overestimates the speciation rate early in cladogenesis, suggesting that the apparent initial spike in speciation rates is artifcatual, and that the observed pattern actually matches the expectation of a constant rates model. Higher taxon sampling is needed to fully understand whether the order exhibited the speciation patterns consistent with an adaptive radiation.

Future Directions

Overall, my dissertation research answered many questions about body shape diversification in the hyperdiverse freshwater fish radiation Characiformes. However,

119 like any diverse radiation, many questions remain, including what role habitat shifts had in driving body shape evolution, how trophic ecology influenced tooth evolution, and whether the true speciation rate exhibits signatures of an adaptive radiation. Below

I highlight some important avenues of future research that will enhance the understanding of characiform evolution.

Habitat shifts likely drove body shape diversification

The results of my dissertation quantified and reconstructed the largest macroevolutionary analysis of characiform fishes. My results demonstate that there is not global morphological optimum for most of the observed trophic ecologies, and that many dietary do not coincide with major changes in body shape. This suggests that diet alone does not drive adapation in body shape. . My results show that a combination of ecological characters is likely driving body shape adaptation in the order.

Streelman and Danley (2003) proposed that adaptive radiations occur in stages with lineages diversifying first in response to habitat, followed by trophic ecology, and ending with sexual selection. Some lineages of cichlids follow this pattern, with the earliest shifts in body shape co-occurring with shifts between macrohabitats (Schliewen et al. 1994; López‐Fernández et al. 2013; Arbour and López-Fernández 2016). Other cichlid groups show macrohabitat adaptation later in cladogenesis (Muschick et al.

2014). Regardless of when diversification occurred, both patterns clearly show body shape diversification occurring in response to shifts in macrohabitats (Schliewen et al.

1994; López-Fernández et al. 2013; Muschick et al. 2014; Arbour and López-

Fernández 2016).

120

Does habitat provide the missing axis necessary to understand how ecology drove the early burst of body shape evolution in Characiformes? The order occupies a variety of habitats including floodplains, large river systems, lakes, and head water streams, and these environments would promote different selective pressures for adaptation with extreme changes in flow and water column position. The results of the

SURFACE analysis in Chapter 2 indicate at least three regime shifts (including a lineage in the Crenuchidae, the Gasteropelecidae, and the Serrasalmidae) occurring in lineages transitioning across the benthic-pelagic habitat axis, which is a common diversification gradient seen in other freshwater fish lineages (Carlson and Wainwright

2010; Hollingsworth et al. 2013; Burress et al. 2016), including multiple adaptive radiations (Schluter 1995; Schluter and Nagel 1995; Hollingsworth et al. 2013). Many other microhabitats, including flow, habitat structure, and sediment type have been found to influence body shape in fishes (Langerhans et al. 2003; Willis et al. 2005;

Langerhans 2008) and likely play a role in characiform body shape diversification.

However, much like macrohabitat, microhabitat is poorly resolved for much of the order, making a macroevolutionary analysis impossible. It is likely that changes in both micro and marohabitat in combination with changes in diet and the corresponding selective pressures shaped many aspects of body shape evolution in the order.

Reliable habitat data are rarely available for members of the order. However, some smaller subclades of Characiformes, such as the Crenuchidae, have lineages with well-established transitions across the benthic-pelagic habitat axis (Buckup 1998).

Body shape adaptation across the benthic-pelagic habitat axis has been well studied in other groups of fishes (Carlson and Wainwright 2010; Hollingsworth et al. 2013;

121

Hulsey et al. 2013; Burress et al. 2016), with pelagic lineages being deeper bodied than their benthic counterparts. These previous studies create the perfect predictive framework for analyzing body shape adaptation across macrohabitats in Crenuchidae to determine if the pelagic lineages are deeper bodied than the benthic lineages.

Analysis of body shape adaptation across macrohabitats in smaller subclades of

Characiformes, like the Crenuchidae, will offer insights into how habitat transitions may have influenced body shape evolution across the order. The collection of reliable and quantifiable habitat data in combination with trophic ecology data is paramount to understanding exactly how body shape evolved within the order.

Teeth: the hidden axis of diversification?

Characiformes invaded many different trophic niches with lineages falling out into multiple different trophic guilds including piscivores, detritivores, insectivores, herbivores, and omnivores. However, the results of my dissertation show that, except for the convergent piscivores, adaptation to these different trophic guilds did not drive convergent changes in overall body shape or in any of the independent functional modules. The lack of signal in the overall body shape and post-cranial modules is not surprising, as habitat ecology, water column position, and flow should shape most aspects of post-cranial morphology, including trunk shape and fin position (Webb

1982; Webb 1984a; Webb 1984b; Webb and Weihs 1986). Still, cranial morphology is often associated with feeding mode in fishes (Norton 1991).

Flow, water column position, and behavioral ecology can influence cranial morphology. In increased flow environments, fishes can elongate their heads, reducing

122 drag and uplift forces (Bushnell and Moore 1991; Lighthill 1993; Larouche et al. 2015).

Behavioral ecology and water column positioning greatly influence mouth position with top water predators having a more upturned mouth than benthic predators, even when feeding on the same resource type. The diverse array of habitat and feeding modes, including top water, benthic, and rheophilic lineages, might mean that much of cranial morphology is influenced by aspects of macrohabitat and trophic ecology.

Conversely, the trophic categories I used may be too broad and a finer scale trophic analysis might reveal distinct head shapes and adaptive optima for each category.

In other diverse lineages of vertebrates that exhibit multiple shifts in trophic and habitat ecology, lineages diversify to shifts in trophic ecology by adapting the morphology of their trophic apparatus. For instance, Darwin’s finches adapt the length and depth of their beaks to feed on different food resources (Grant and Grant 2006) and

Lake Tanganyika cichlids modify the shape of the lower pharyngeal jaw bone with the different lengths and widths corresponding to different functional performance

(Muschick et al. 2012). Is trophic apparatus the missing the missing variable adapting to changes in trophic ecology? The order does exhibit a riot of diversity in tooth morphology including conical and caniniform teeth in piscivores (Guisande et al.

2012), molariform teeth in granivores (Goulding 1980), and multicuspid teeth in omnivores (Géry 1977; Goulding 1980). Superficially, many of these tooth morphologies tend to segregate across trophic categories with lineages that share a trophic ecology also sharing a tooth morphology (Guisande et al. 2012). A recent study found that independent lineages of scale eating characiforms evolved very different jaw morphologies and functional performance in response to different specialized feeding

123 strategies, indicating a much greater ecological diversity than would be expected for such a similar trophic niche (Kolmann et al. 2018). However, the tooth aspect ratio was very similar between lineages, with the teeth being stouter in all scale eating lineages, indicating a strong relationship between scale feeding and tooth shape (Kolmann et al.

2018).

No study has explicitly looked at tooth diversity in a macroevolutionary framework over the entire order to understand the role that trophic ecology has had in shaping tooth diversity. If Characiformes adapted tooth morphology in response to shifts in trophic ecology, then we would predict that lineages that converge in trophic niche would also converge in tooth morphology. The amount of tooth diversity in the order, and similarity among trophic groups, offers a plausible explanation for how characiforms were able invade multiple different trophic niches without converging in aspects of body shape.

Increased taxon sampling is needed to fully understand the diversification dynamics

In my fourth chapter, I tested whether the speciation rate exhibited signatures of an adaptive radiation. I reconstructed higher speciation rates early in cladogenesis followed by a quick slow down over time, which matched the patterns of an adaptive radiation. However, constant rate simulations showed the same rate pattern when the overall tree was highly pruned, indicating that the low overall taxon sampling in my empirical tree could be responsible for the apparently elevated speciation rate early in the order’s diversification. The results of the simulations, therefore, suggest that

124 speciation rates did not follow the pattern of an adaptive radiation, but were more likely constant over time.

The rates of morphological evolution and the partitioning of morphological diversity among early diversifying clades do indicate that morphological evolution evolved in a pattern consistent with an adaptive radiation. However, the speciation rate exhibited a constant rate of evolution indicating that body shape evolution was decoupled from the speciation rate, indicating that the characiform radiation did not evolve under an adaptive radiation. Body shape morphology likely evolved through ecological diversification early in the radiation, while much of the species diversity must have arisen slowly through time in response to allopatric speciation with little morphological change occurring after the initial burst of diversification. The fossil record supports the decoupling of morphology and speciation because many fossil

Characiformes closely match extant species in morphology, indicating that the speciation that occurred over the last ~60 million years had little influence on the morphological evolution. Higher taxon sampling will offer more insight into the diversification dynamics of the order.

In conclusion, my dissertation is the first study to quantify the adaptive landscape of Characiformes. Typically, macroevolutionary studies look at how a single ecological character has influenced adaptation in a single morphological character, but the results of my dissertation have shown that diversification in Characiformes is much more complex. The future of research involving Characiformes should focus on collating multiple morphological characters into a single holistic study to understand how adaptation to multiple ecological axes (i.e. habitat structure, flow regime, and

125 trophic ecology) has influenced diversification in this highly modular radiation.

Although the field is unable to answer this question to date, the future of macroevolutionary studies involving Characiformes is bright. There are ongoing phylogenetic studies with much higher taxon sampling than my dissertation, the collection of 3-dimensional morphological data, including quantification of tooth shape, is becoming financially feasible and common, and statistical methods for analyzing evolution in non-model organisms are becoming more sophisticated and computationally feasible. Macroevolutionary studies of Characiformes will be at the forefront of evolutionary ichthyology into the next decade and beyond.

126

DISSERTATION REFERENCES

Abe, K. T., T. C. Mariguela, G. S. Avelino, F. Foresti, and C. Oliveira. 2014.

Systematic and historical biogeography of the Bryconidae (Ostariophysi:

Characiformes) suggesting a new rearrangement of its genera and an old origin

of Mesoamerican ichthyofauna. BMC Evol Biol 14:152.

Abilhoa, V., L. P. Bastos, and F. Wegbecher. 2007. Feeding habits of Rachoviscus

crassiceps (Teleostei: Characidae) in a coastal Atlantic rainforest stream,

southern Brazil. Ichthyol Explor Freshwaters 18:227-232.

Ackerly, D. D. 2000. Taxon sampling, correlated evolution, and independent contrasts.

Evolution 54:1480-1492.

Adams, D. C. 2016. Evaluating modularity in morphometric data: challenges with the

RV coefficient and a new test measure. Methods Ecol. Evol. 7:565-572.

Adams, D. C. and M. L. Collyer. 2017. Multivariate Phylogenetic Comparative

Methods: Evaluations, Comparisons, and Recommendations. Syst Biol 67:14-

31.

Adams, D. C. and E. Otárola‐Castillo. 2013. geomorph: an r package for the collection

and analysis of geometric morphometric shape data. Methods Ecol. Evol.

4:393-399.

Adebisi, A. 1981. Analyses of the stomach contents of the piscivorous fishes of the

upper Ogun River in Nigeria. Hydrobiologia 79:167-177.

Aguilar‐Medrano, R., B. Frederich, and P. H. Barber. 2016. Modular diversification of

the locomotor system in damselfishes (Pomacentridae). J Morphol 277:603-

614.

127

Alfaro, M. E., F. Santini, C. Brock, H. Alamillo, A. Dornburg, D. L. Rabosky, G.

Carnevale, and L. J. Harmon. 2009. Nine exceptional radiations plus high

turnover explain species diversity in jawed vertebrates. PNAS 106:13410-

13414.

Aranha, J. R., J. C. Gomes, and F. N. O. Fogaça. 2000. Feeding of two sympatric

species of Characidium, C. lanei and C. pterostictum (Characidiinae) in a

coastal stream of Atlantic Forest (Southern Brazil). Braz Arch Biol Technol

43:527-531.

Arawomo, G. A. O. 1976. Food and feeding of three Citharinus species in Lake Kainji,

Nigeria. J Fish Biol 9:3-10.

Arawomo, G. A. O. 1982. Food and feeding of three Distichodus species (Pisces:

Characiformes) in Lake Kainji, Nigeria. Hydrobiologia 94:177-181.

Arbour, J. and H. López-Fernández. 2016. Continental cichlid radiations: functional

diversity reveals the role of changing ecological opportunity in the Neotropics.

Proc R Soc Lond B Biol Sci 283:20160556.

Arcila, D., G. Ortí, R. Vari, J. W. Armbruster, M. L. J. Stiassny, K. D. Ko, M. H. Sabaj,

J. Lundberg, L. J. Revell, and R. Betancur-R. 2017. Genome-wide interrogation

advances resolution of recalcitrant groups in the tree of life. Nature Ecology &

Evolution 1:0020.

Arroyave, J., J. S. Denton, and M. L. Stiassny. 2013. Are characiform fishes

Gondwanan in origin? Insights from a time-scaled molecular phylogeny of the

Citharinoidei (Ostariophysi: Characiformes). PLoS ONE 8:e77269.

128

Arroyave, J. and M. Stiassny. 2011. Phylogenetic relationships and the temporal

context for the diversification of African characins of the family Alestidae

(Ostariophysi: Characiformes): Evidence from DNA sequence data. Mol

Phylogenet Evol 60:385-397.

Arroyave, J. and M. L. Stiassny. 2014. DNA barcoding reveals novel insights into

pterygophagy and prey selection in distichodontid fishes (Characiformes:

Distichodontidae). Ecology and evolution 4:4534-4542.

Bailey, R. 1994. Guide to the fishes of the River Nile in the Republic of the Sudan. J

Nat Hist 28:937-970.

Bertaco, V. A. and F. R. Carvalho. 2010. New species of Hasemania (Characiformes:

Characidae) from Central Brazil, with comments on the endemism of upper rio

Tocantins basin, Goiás State. Neotrop Ichthyol 8:27-32.

Bertaco, V. A., F. R. Carvalho, and F. C. Jerep. 2010. Astyanax goyanensis (Miranda-

Ribeiro, 1944), new combination and Astyanax courensis, new species

(Ostariophysi: Characiformes): two Characidae from the upper rio Tocantins

basin, Central Brazil. Neotrop Ichthyol 8:265-275.

Blache, J. 1964. Les poissons du bassin du Tchad et du bassin adjacent du Mayo Kebbi:

étude systématique et biologique. Mémoires de l'ORSTOM 4:1-483.

Blount, Z. D., C. Z. Borland, and R. E. Lenski. 2008. Historical contingency and the

evolution of a key innovation in an experimental population of Escherichia coli.

PNAS 105:7899-7906.

129

Bouckaert, R., J. Heled, D. Kühnert, T. Vaughan, C.-H. Wu, D. Xie, M. A. Suchard,

A. Rambaut, and A. J. Drummond. 2014. BEAST 2: A Software Platform for

Bayesian Evolutionary Analysis. PLOS Comput. Biol. 10:e1003537.

Buckup, P. A. 1993a. Phylogenetic interrelationships and reductive evolution in

neotropical Characidiin Fishes (Characiformes, Ostariophysi). Cladistics

9:305-341.

Buckup, P. A. 1993b. Review of the characidiin fishes (Teleostei: Characiformes), with

descriptions of four new genera and ten new species. Ichthyol Explor

Freshwaters 4:97-154.

Buckup, P. A. 1998. Relationships of the Characidiinae and phylogeny of characiform

fishes (Teleostei: Ostariophysi). Pp. 251-260 in, L. R. Malabarba, R. E. Reis,

R. P. Vari, Z. M. Lucena, and C. A. S. Lucena, eds. Phylogeny and classification

of Neotropical fishes. Edipucrs, Porto Alegre.

Bührnheim, C. M. and L. R. Malabarba. 2006. Redescription of the type species of

Odontostilbe Cope, 1870 (Teleostei: Characidae: Cheirodontinae), and

description of three new species from the Amazon basin. Neotrop Ichthyol

4:167-196.

Burbrink, F. T., X. Chen, E. A. Myers, M. C. Brandley, and A. R. Pyron. 2012.

Evidence for determinism in species diversification and contingency in

phenotypic evolution during adaptive radiation. Proc R Soc Lond B Biol Sci

279:4817–4826.

130

Burbrink, F. T. and A. R. Pyron. 2010. How does ecological opportunity influence rates

of speciation, extinction, and morphological diversification in new world

ratsnakes (tribe lampropeltini)? Evolution 64:934-943.

Burress, E. D., J. M. Holcomb, M. Tan, and J. W. Armbruster. 2016. Ecological

diversification associated with the benthic‐to‐pelagic transition by North

American minnows. J Evol Biol 30:549-560.

Burress, E. D., L. Piálek, J. R. Casciotta, A. Almirón, M. Tan, J. W. Armbruster, and

O. Říčan. 2017. Island-and lake-like parallel adaptive radiations replicated in

rivers. Proc. R. Soc. B 285:20171762.

Bushnell, D. M. and K. Moore. 1991. Drag reduction in nature. Annual Review of Fluid

Mechanics 23:65-79.

Bussing, W. A. 1998. Freshwater fishes of Costa Rica. International Journal of Tropical

Biology 46.

Cala, P. 2005. Diet of Astyanax fasciatus and Cyphocharax magdalenae (Pisces:

Characiformes), in the Betania reservoir, upper part of the rio Magdalena

system, Colombia. Dahlia-Revista de la Asociación Colombiana de Ictiología

Calcagnotto, D., S. A. Schaefer, and R. DeSalle. 2005. Relationships among

characiform fishes inferred from analysis of nuclear and mitochondrial gene

sequences. Mol Phylogenet Evol 36:135-153.

Carlson, R. L. and P. C. Wainwright. 2010. The ecological morphology of darter fishes

(Percidae: Etheostomatinae). Biol J Linn Soc 100:30-45.

131

Casatti, L. 2002. Alimentação dos peixes em um riacho do Parque Estadual Morro do

Diabo, bacia do alto rio Paraná, sudeste do Brasil. Biota Neotrop 2:1-14.

Casatti, L. and R. M. C. Castro. 1998. A fish community of the São Francisco River

headwaters riffles, southeastern Brazil. Ichthyol Explor Freshwaters 9:229-242.

Chakrabarty, P., B. C. Faircloth, F. Alda, W. B. Ludt, C. D. McMahan, T. J. Near, A.

Dornburg, J. S. Albert, J. Arroyave, M. L. J. Stiassny, L. Sorenson, and M. E.

Alfaro. 2017. Phylogenomic Systematics of Ostariophysan Fishes:

Ultraconserved Elements Support the Surprising Non-Monophyly of

Characiformes. Syst Biol 66:881-895.

Chen, W. J., S. Lavoué, and R. L. Mayden. 2013. Evolutionary origin and early

biogeography of otophysan fishes (ostariophysi: Teleostei). Evolution 67:2218-

2239.

Claramunt, S. 2010. Discovering exceptional diversifications at continental scales: the

case of the endemic families of Neotropical suboscine passerines. Evolution

64:2004-2019.

Clavel, J., G. Escarguel, and G. Merceron. 2015. mvmorph: an r package for fitting

multivariate evolutionary models to morphometric data. Methods Ecol. Evol.

6:1311-1319.

Claverie, T. and S. Patek. 2013. Modularity and rates of evolutionary change in a

power‐amplified prey capture system. Evolution 67:3191-3207.

Collar, D. C., T. J. Near, and P. C. Wainwright. 2005. Comparative analysis of

morphological diversity: Does disparity accumulate at the same rate in two

lineages of centrarchid fishes? Evolution 63:1783.

132

Collar, D. C., B. C. O'Meara, P. C. Wainwright, and T. J. Near. 2009. Piscivory limits

diversification of feeding morphology in centrarchid fishes. Evolution 63:1557-

1573.

Collar, D. C., P. C. Wainwright, M. E. Alfaro, L. J. Revell, and R. S. Mehta. 2014.

Biting disrupts integration to spur skull evolution in eels. Nature

Communications 5:5505.

Cooney, C. R., J. A. Bright, E. J. R. Capp, A. M. Chira, E. C. Hughes, C. J. A. Moody,

L. O. Nouri, Z. K. Varley, and G. H. Thomas. 2017. Mega-evolutionary

dynamics of the adaptive radiation of birds. Nature 542:344-347.

Cooper, J. W. and M. W. Westneat. 2009. Form and function of damselfish skulls: rapid

and repeated evolution into a limited number of trophic niches. BMC Evol Biol

9:1-17.

Corrêa, C. E., N. S. Hahn, and R. L. Delariva. 2009. Extreme trophic segregation

between sympatric fish species: the case of small sized body Aphyocharax in

the Brazilian Pantanal. Hydrobiologia 635:57-65.

Costa, W. 1987. Feeding habits of a fish community in a tropical coastal stream, Rio

mato grosso, Brazil. Stud Neotrop Fauna Environ 22:145-153. da Silva, C. C., E. J. G. Ferreira, and C. P. de Deus. 2008. Diet of five species of

Hemiodontidae (Teleostei, Characiformes) in the area of influence of the

Balbina reservoir, Uatumã River, State of Amazonas, Brazil. Iheringia. Série

Zoologia 98:464-468.

Daget, J. 1962. Le genre Citharinus (Poissons, Characiformes). Revue de Zoologie et

de Botanique Africaines 66.

133

Daget, J. 1967. Le genre Ichthyborus (Poissons, Characiformes). Outre-Mer (Serie

hydrobiologie) 1.

Dahdul, W. 2010. Review of the phylogenetic relationships and fossil record of

Characiformes. Pp. 441-464 in T. Grande, ed. Gonorynchiformes and

Ostariophysan Relationships. CRC Press, Boca Raton, Florida.

Davenport, J. 1994. How and why do flying fish fly? Rev Fish Biol Fish 4:184-214. de Braga, F. M. S. 1990. Aspectos da reprodução e alimentação de peixes comuns em

um trecho do rio Tocantins entre Imperatriz e Estreito, Estados do Maranhão e

Tocantins, Brasil. Revista Brasileira de Biologia 50:547-558. de los Angeles Bistoni, M., J. G. Haro, and M. Gutiérrez. 1995. Feeding of Hoplias

malabaricus in the wetlands of Dulce river (Córdoba, Argentina).

Hydrobiologia 316:103-107. de Melo, C., F. de Machado, and V. Pinto-Silva. 2004. Feeding habits of fish from a

stream in the savanna of Central Brazil, Araguaia Basin. Neotrop Ichthyol 2:37-

44. de Mérona, B., G. dos Santos, and R. de Almeida. 2001. Short Term Effects of Tucuruí

Dam (Amazonia, Brazil) on the Trophic Organization of Fish Communities.

Environ Biol Fishes 60:375-392. de Mérona, B. and J. Rankin-de-Mérona. 2004. Food resource partitioning in a fish

community of the central Amazon floodplain. Neotrop Ichthyol 2:75-84. de Pinna, M., J. Zuanon, L. Rapp Py-Daniel, and P. Petry. 2017. A new family of

neotropical freshwater fishes from deep fossorial Amazonian habitat, with a

134

reappraisal of morphological characiform phylogeny (Teleostei: Ostariophysi).

Zool J Linn Soc 182:76-106.

Degnan, J. H. and N. A. Rosenberg. 2009. Gene tree discordance, phylogenetic

inference and the multispecies coalescent. Trends Ecol Evol 24:332-340.

Dietoa, Y., G. Gourène, and A. Ouattara. 2007. Habitudes alimentaires de Brycinus

longipinnis dans le complexe fluvio-lacustre de la Bia, Cote d'Ivoire. Belg J

Zool 137:3-9.

Dillman, C. B., B. L. Sidlauskas, and R. P. Vari. 2016. A morphological supermatrix‐

based phylogeny for the Neotropical fish superfamily Anostomoidea

(Ostariophysi: Characiformes): phylogeny, missing data and homoplasy.

Cladistics 32:276-296. dos Santos, G. 1982. Caracterização, hábitos alimentares e reprodutivos de quatro

espécies de" aracus" e considerações ecológicas sobre o grupo no lago

Janauacá-AM.(Osteichthyes, Characoidei, Anostomidae). Acta Amazonica

12:713-739.

Drucker, E. G. and G. V. Lauder. 2001. Locomotor function of the dorsal fin in teleost

fishes: experimental analysis of wake forces in sunfish. J Exp Biol 204:2943-

2958.

Drucker, E. G. and G. V. Lauder. 2005. Locomotor function of the dorsal fin in rainbow

trout: kinematic patterns and hydrodynamic forces. J Exp Biol 208:4479-4494.

Drummond, A. J., S. Y. W. Ho, M. J. Phillips, and A. Rambaut. 2006. Relaxed

Phylogenetics and Dating with Confidence. PLoS Biol 4:e88.

135

Edgar, R. C. 2004. MUSCLE: a multiple sequence alignment method with reduced time

and space complexity. BMC Bioinformatics 5:1-19.

Elmer, K. R., S. Fan, H. Kusche, M. L. Spreitzer, A. F. Kautt, P. Franchini, and A.

Meyer. 2014. Parallel evolution of Nicaraguan crater lake cichlid fishes via non-

parallel routes. Nature communications 5:5168.

Eschmeyer, W., R. Fricke, and R. Van der Laan. 2016. Catalog of fishes: genera,

species, references. Electronic version accessed 19.

Eschmeyer, W. N., R. Fricke, J. D. Fong, and D. A. Polack. 2010. Marine fish diversity:

history of knowledge and discovery (Pisces). Zootaxa 2525:19-50.

Esteves, K. and J. Lobón-Cerviá. 2001. Composition and trophic structure of a fish

community of a clear water Atlantic rainforest stream in southeastern Brazil.

Environ Biol Fishes 62:429-440.

Feibel, C. 2003. Stratigraphy and depositional history of the Lothagam sequence. Pp.

17-29 in M. Leakey, and J. Harris, eds. Lothagam: The Dawn of Humanityin

Eastern Africa. Columbia University Press, New York.

Feilich, K. L. 2016. Correlated evolution of body and fin morphology in the cichlid

fishes. Evolution 70:2247-2267.

Ferreira, K. M. 2007. Biology and ecomorphology of stream fishes from the rio Mogi-

Guaçu basin, Southeastern Brazil. Neotrop Ichthyol 5:311-326.

Ferreira, K. M. and F. C. Lima. 2006. A New Species of Knodus (Characiformes:

Characidae) from the Rio Tiquié, Upper Rio Negro System, Brazil. Copeia

2006:630-639.

136

Fink, S. V. and W. L. Fink. 1981. Interrelationships of the ostariophysan fishes

(Teleostei). Zool J Linn Soc 72:297-353.

Fink, W. L. 1996. Interrelationships of ostariophysan fishes (Teleostei). Pp. 496 in M.

L. J. Stiassny, L. R. Parenti, and G. D. Johnson, eds. Interrelationships of

Fishes. Academic Press, San Diego.

Fink, W. L. and M. L. Zelditch. 1995. Phylogenetic analysis of ontogenetic shape

transformations: a reassessment of the piranha genus Pygocentrus (Teleostei).

Syst Biol 44:343-360.

Flecker, A. S. 1992. Fish Trophic Guilds and the Stucture of a Tropical Stream: Weak

Direct vs. Strong Indirect Effects. Ecology 73:927-940.

Foote, M. 1997. The evolution of morphological diversity. Annu Rev Ecol Syst 28:129-

152.

Forsberg, B. R., C. Araujo-Lima, L. A. Martinelli, R. L. Victoria, and J. A. Bonassi.

1993. Autotrophic Carbon Sources for Fish of the Central Amazon. Ecology

74:643-652.

Friedman, M. 2008. The evolutionary origin of flatfish asymmetry. Nature 454:209-

212.

Froese, R. and D. Pauly. 2000. FishBase2000: Concepts Designs and Data Sources.

WorldFish 2000.

Galvis, G., J. Mojica, and M. Camargo. 1997. Peces del Catatumbo. Pp. 118.

Asociación Cravo Norte, Ecopetrol. DXY & Shell, Bogotá.

137

Gaston, K. A. and T. E. Lauer. 2015. Morphometric variation in bluegill Lepomis

macrochirus and green sunfish Lepomis cyanellus in lentic and lotic systems. J

Fish Biol 86:317-332.

Gavrilets, S. and J. B. Losos. 2009. Adaptive radiation: contrasting theory with data.

Science 323:732-737.

Gelós, M., F. Mello, G. Goyenola, C. Iglesias, C. Fosalba, F. García-Rodríguez, J.

Pacheco, S. García, and M. Meerhoff. 2010. Seasonal and diel changes in fish

activity and potential cascading effects in subtropical shallow lakes with

different water transparency. Hydrobiologia 646:173-185.

Géry, J. 1977. Characoids of the world. TFH Publications, Neptune City.

Gibb, A. C., K. A. Dickson, and G. V. Lauder. 1999. Tail kinematics of the chub

mackerel Scomber japonicus: testing the homocercal tail model of fish

propulsion. J Exp Biol 202:2433-2447.

Givnish, T. J. 2015. Adaptive radiation versus ‘radiation’and ‘explosive

diversification’: why conceptual distinctions are fundamental to understanding

evolution. New Phytol 207:297-303.

Givnish, T. J., M. H. Barfuss, B. Van Ee, R. Riina, K. Schulte, R. Horres, P. A.

Gonsiska, R. S. Jabaily, D. M. Crayn, and J. A. C. Smith. 2014. Adaptive

radiation, correlated and contingent evolution, and net species diversification in

Bromeliaceae. Mol Phylogenet Evol 71:55-78.

Glor, R. E. 2010. Phylogenetic Insights on Adaptive Radiation. Annual Review of

Ecology, Evolution, and Systematics 41:251-270.

138

Gottsberger, G. 1978. Seed dispersal by fish in the inundated regions of Humaita,

Amazonia. Biotropica 10:170-183.

Gould, S. J. 1989. Wonderful Life. Penguin Books, London.

Gould, S. J. 2002. The structure of evolutionary theory. Harvard University Press,

Harvard, MA.

Goulding, M. 1980. The fishes and the forest: explorations in Amazonian natural

history. University of California Press, Berkeley, CA.

Goulding, M., M. Carvalho, and E. G. Ferreira. 1988. Rio Negro, rich life in poor water.

SPB Academic Publishing, The Hague, The Netherlands.

Grant, P. R. and R. B. Grant. 2006. Evolution of character displacement in Darwin's

finches. Science 313:224-226.

Guisande, C., P. Pelayo-Villamil, M. Vera, A. Manjarrés-Hernández, M. R. Carvalho,

R. P. Vari, L. Jiménez, C. Fernández, P. Martínez, and E. Prieto-Piraquive.

2012. Ecological factors and diversification among Neotropical characiforms.

International Journal of Ecology 2012:610419.

Hahn, N., R. Fugi, and I. F. de Andrian. 2004. Trophic ecology of the fish assemblages

in Thomaz SM, Agostinho AA, and H. NS, eds. The Upper Paraná River and

its floodplain: physical aspects, ecology and conservation. Backhuys

Publishers, Leiden.

Hahn, N. S., A. A. Agostinho, L. C. Gomes, and L. M. Bini. 1998. Estrutura trófica da

ictiofauna do reservatório de Itaipu (Paraná-Brasil) nos primeiros anos de sua

formação. Interciencia 23:299-305.

139

Hanken, J. and D. B. Wake. 1993. Miniaturization of body size: organismal

consequences and evolutionary significance. Annu Rev Ecol Syst 24:501-519.

Hansen, T. F. 1997. Stabilizing selection and the comparative analysis of adaptation.

Evolution 51:1341-1351.

Harmon, L. J., J. J. Kolbe, J. M. Cheverud, and J. B. Losos. 2005. Convergence and the

multidimensional niche. Evolution 59:409-421.

Harmon, L. J., J. B. Losos, T. Jonathan Davies, R. G. Gillespie, J. L. Gittleman, W.

Bryan Jennings, K. H. Kozak, M. A. McPeek, F. Moreno‐Roark, and T. J. Near.

2010. Early bursts of body size and shape evolution are rare in comparative

data. Evolution 64:2385-2396.

Harmon, L. J., J. A. Schulte, A. Larson, and J. B. Losos. 2003. Tempo and mode of

evolutionary radiation in iguanian lizards. Science 301:961-964.

Harris, J. 1938. The role of the fins in the equilibrium of the swimming fish. J Exp Biol

15:32-47.

Heath, T. A., S. M. Hedtke, and D. M. Hillis. 2008. Taxon sampling and the accuracy

of phylogenetic analyses. J Syst Evol 46:239-257.

Hoeinghaus, D. J., C. A. Layman, A. D. Arrington, and K. O. Winemiller. 2003.

Spatiotemporal variation in fish assemblage structure in tropical floodplain

creeks. Environ Biol Fishes 67:379-387.

Holden, M. and W. Reed. 1972. West African freshwater fish. West African freshwater

fish.

Holden, M. J. 1970. The feeding habits of Alestes baremose and Hydrocynus forskali

(Pisces) in Lake Albert, East Africa. Journal of Zoology 161:137-144.

140

Hollingsworth, P. R., A. M. Simons, J. A. Fordyce, and D. C. Hulsey. 2013. Explosive

diversification following a benthic to pelagic shift in freshwater fishes. BMC

Evol Biol 13:1-10.

Horeau, V., P. Cerdan, A. Champeau, and S. Richard. 1998. Importance of aquatic

invertebrates in the diet of rapids-dwelling fish in the Sinnamary River, French

Guiana. J Trop Ecol 14:851-864.

Huelsenbeck, J. P., R. Nielsen, J. P. Bollback, and T. Schultz. 2003. Stochastic

Mapping of Morphological Characters. Syst Biol 52:131-158.

Hugueny, B. and M. Pouilly. 1999. Morphological correlates of diet in an assemblage

of West African freshwater fishes. J Fish Biol 54:1310-1325.

Hulsey, C. D., F. G. De León, and R. Rodiles-Hernández. 2006. Micro-and

macroevolutionary decoupling of cichlid jaws: a test of Liem's key innovation

hypothesis. Evolution 60:2096-2109.

Hulsey, C. D., R. J. Roberts, Y. H. E. Loh, M. F. Rupp, and J. T. Streelman. 2013. Lake

Malawi cichlid evolution along a benthic/limnetic axis. Ecology and Evolution

3:2262-2272.

Hulsey, D. C. 2006. Function of a key morphological innovation: fusion of the cichlid

pharyngeal jaw. Proc R Soc Lond B Biol Sci 273:669-675.

Ingram, T. and D. L. Mahler. 2013. SURFACE: detecting convergent evolution from

comparative data by fitting Ornstein‐Uhlenbeck models with stepwise Akaike

Information Criterion. Methods Ecol. Evol. 4:416-425.

Jacobo, C. 1995. Trophic relationships in fish communities in some basins of eastern

Chaco, Argentina. Revista Brasileira de Biologia 55.

141

Javonillo, R., L. R. Malabarba, S. H. Weitzman, and J. R. Burns. 2010. Relationships

among major lineages of characid fishes (Teleostei: Ostariophysi:

Characiformes), based on molecular sequence data. Mol Phylogenet Evol

54:498-511.

Kearse, M., R. Moir, A. Wilson, S. Stones-Havas, M. Cheung, S. Sturrock, S. Buxton,

A. Cooper, S. Markowitz, and C. Duran. 2012. Geneious Basic: an integrated

and extendable desktop software platform for the organization and analysis of

sequence data. Bioinformatics 28.

Kemenes, A. and B. R. Forsberg. 2014. Factors influencing the structure and spatial

distribution of fishes in the headwater streams of the Jaú River in the Brazilian

Amazon. Braz J Biol 74:23-32.

Klingenberg, C. P. 2014. Studying morphological integration and modularity at

multiple levels: concepts and analysis. Phil. Trans. R. Soc. B 369:20130249.

Knöeppel, H. A. 1972. Zur nahrung tropischer Süsswasserfische aus Südamerika. Ibid

Kocher, T. D., J. A. Conroy, K. R. McKaye, and J. R. Stauffer. 1993. Similar

morphologies of cichlid fish in Lakes Tanganyika and Malawi are due to

convergence. Mol Phylogenet Evol 2:158-165.

Kolmann, M. A., J. M. Huie, K. Evans, and A. P. Summers. 2018. Specialized

specialists and the narrow niche fallacy: a tale of scale-feeding fishes. Royal

Society Open Science 5:171581.

Lanfear, R., P. B. Frandsen, A. M. Wright, T. Senfeld, and B. Calcott. 2016.

PartitionFinder 2: New Methods for Selecting Partitioned Models of Evolution

142

for Molecular and Morphological Phylogenetic Analyses. Mol Biol Evol

34:772-773.

Langerhans, B. R. 2008. Predictability of phenotypic differentiation across flow

regimes in fishes. Integr Comp Biol 48:750-768.

Langerhans, B. R. and T. J. DeWitt. 2004. Shared and unique features of evolutionary

diversification. Am. Nat, 164:335-349.

Langerhans, B. R., C. A. Layman, A. K. Langerhans, and T. J. Dewitt. 2003. Habitat‐

associated morphological divergence in two Neotropical fish species. Biol J

Linn Soc 80:689-698.

Larouche, O., R. Cloutier, and M. L. Zelditch. 2015. Head, body and fins: patterns of

morphological integration and modularity in fishes. Evolutionary Biology

42:296-311.

Lauder, G. V. 1981. Form and function: structural analysis in evolutionary

morphology. Paleobiology 7:430-442.

Lauzanne, L. 1973. Etude qualitative de la nutrition des Alestes baremoze (Pisces,

Characidae). Cahiers ORSTOM Série Hydrobiolgie 7.

Lauzanne, L. 1976. Régimes alimentaires et relations trophiques dans les poissons du

lac Tchad. Cahiers ORSTOM Série Hydrobiologie 10:267-310.

Lek, S. and S. Lek. 1977. Ecologie et biologie de Micralestes acutidens (Peters,

1852)(Pisces, Characidae) du bassin du lac Tchad. Cahiers ORSTOM Série

Hydrobiologie 11:255-268.

143

Lerner, H. R., M. Meyer, H. F. James, M. Hofreiter, and R. C. Fleischer. 2011.

Multilocus resolution of phylogeny and timescale in the extant adaptive

radiation of Hawaiian honeycreepers. Curr Biol 21:1838-1844.

Li, C., G. Ortí, G. Zhang, and G. Lu. 2007. A practical approach to phylogenomics: the

phylogeny of ray-finned fish (Actinopterygii) as a case study. BMC Evol Biol

7:44.

Liem, K. F. 1973. Evolutionary strategies and morphological innovations: cichlid

pharyngeal jaws. Syst Zool 22:425-441.

Lighthill, J. 1993. Estimates of pressure differences across the head of a swimming

clupeid fish. Philosophical Transactions of the Royal Society of London B:

Biological Sciences 341:129-140.

Lima, F. C. T. 2003. Genera incertae sedis in Characidae. Check list of the freshwater

fishes of South and Central America.

López‐Fernández, H., J. H. Arbour, K. Winemiller, and R. L. Honeycutt. 2013. Testing

for ancient adaptive radiations in Neotropical cichlid fishes. Evolution 67:1321-

1337.

Losos, J. B. 1992. The evolution of convergent structure in Caribbean Anolis

communities. Syst Biol 41:403-420.

Losos, J. B. 2009. Lizards in an evolutionary tree: ecology and adaptive radiation of

anoles. University of California Press.

Losos, J. B. 2010. Adaptive radiation, ecological opportunity, and evolutionary

determinism: American Society of Naturalists EO Wilson Award address. Am.

Nat, 175:623-639.

144

Losos, J. B. 2011. Convergence, adaptation, and constraint. Evolution 65:1827-1840.

Losos, J. B. 2017. Improbable destinies: Fate, chance, and the future of evolution.

Riverhead Books, USA.

Losos, J. B., T. R. Jackman, A. Larson, K. de Queiroz, and L. Rodrı́guez-Schettino.

1998. Contingency and Determinism in Replicated Adaptive Radiations of

Island Lizards. Science 279:2115-2118.

Losos, J. B. and D. L. Mahler. 2010. Adaptive radiation: the interaction of ecological

opportunity, adaptation, and speciation. Pp. 381-420 in M. A. Bell, D. J.

Futuyma, W. F. Eanes, and J. S. Levinton, eds. Evolution since Darwin: the first

150 years. Sinauer Associates, Sunderland, M.A. .

Lovette, I. J., E. Bermingham, and R. E. Ricklefs. 2002. Clade-specific morphological

diversification and adaptive radiation in Hawaiian songbirds. Proc R Soc Lond

B Biol Sci 269:37-42.

Lowe-McConnell, R. 1975. Fish communities in tropical freshwaters. Longman

Incorporated, New York, New York.

Lujan, N. K. and K. W. Conway. 2015. Life in the fast lane: a review of rheophily in

freshwater fishes. Pp. 107-136. Extremophile Fishes. Springer.

Luz-Agostinho, K. D., L. M. Bini, R. Fugi, A. A. Agostinho, and H. F. Júlio Jr. 2006.

Food spectrum and trophic structure of the ichthyofauna of Corumbá reservoir,

Paraná river Basin, Brazil. Neotrop Ichthyol 4:61-68.

Machado-Schiaffino, G., A. F. Kautt, H. Kusche, and A. Meyer. 2015. Parallel

evolution in Ugandan crater lakes: repeated evolution of limnetic body shapes

in haplochromine cichlid fish. BMC Evol Biol 15:9.

145

Maddison, W. P. and D. R. Maddison. 2017. Mesquite: a modular system for

evolutionary analysis. Mesquite: a modular system for evolutionary analysis.

Version 3.31 http://mesquiteproject.org.

Madsen, O., M. Scally, C. J. Douady, and D. J. Kao. 2001. Parallel adaptive radiations

in two major clades of placental mammals. Nature 409:610-614.

Maestri, R., L. R. Monteiro, R. Fornel, N. S. Upham, B. D. Patterson, and T. R. O.

Freitas. 2017. The ecology of a continental evolutionary radiation: Is the

radiation of sigmodontine rodents adaptive? Evolution 71:610-632.

Mahler, D. L., T. Ingram, L. J. Revell, and J. B. Losos. 2013. Exceptional convergence

on the macroevolutionary landscape in island lizard radiations. Science

341:292-295.

Mahler, D. L., L. J. Revell, R. E. Glor, and J. B. Losos. 2010. Ecological opportunity

and the rate of morphological evolution in the diversification of Greater

Antillean anoles. Evolution 64:2731-2745.

Malabarba, L. R. and S. H. Weitzman. 2003. Descriptions of a new genus from southern

Brazil, Uruguay and Argentina, with a discussion of a putative characid clade

(Teleostei: Characiformes: Characidae). Comunicacoes do Museu de Ciencias

e Tecnologia da PUCRS. 16 (1). 67-151 67.

Malabarba, M. 1996. Reassessment and relationships of Curimata mosesi Travassos &

Santos, a fossil fish (Teleostei: Characiformes: Curimatidae) from the tertiary

of São Paulo, Brazil. Comunicações do Museu de Ciências da PUCRS, Série

Zoologia 9:55-63.

146

Malabarba, M. and L. Malabarba. 2010. Biogeography of Characiformes: an evaluation

of the available information of fossil and extant taxa. Pp. 317-336 in H.-P. S.

M. V. H. W. J. S. Nelson, ed. Origin and Phylogenetic Interrelationships of

Teleosts. Verlag Dr. Friedrich Pfeil, München.

Mariguela, T., G. Ortí, G. Avelino, K. Abe, and C. Oliveira. 2013a. Composition and

interrelationships of a large Neotropical freshwater fish group, the subfamily

Cheirodontinae (Characiformes: Characidae): a case study based on

mitochondrial and nuclear DNA sequences. Mol Phylogenet Evol 68:23-34.

Mariguela, T. C., R. C. Benine, K. T. Abe, G. S. Avelino, and C. Oliveira. 2013b.

Molecular phylogeny of Moenkhausia (Characidae) inferred from

mitochondrial and nuclear DNA evidence. J Zool Syst Evol Res 51:327-332.

Marin, J. and S. B. Hedges. 2016. Time best explains global variation in species

richness of amphibians, birds and mammals. J Biogeogr 43:1069-1079.

Márquez, E. J. 2008. A statistical framework for testing modularity in multidimensional

data. Evolution 62:2688-2708.

Martin, C. H. and P. C. Wainwright. 2011. Trophic novelty is linked to exceptional

rates of morphological diversification in two adaptive radiations of cyprinodon

pupfish. Evolution 65:2197-2212.

Martinez, C. and M. Stiassny. 2017. Can an eel be a flatfish? Observations on enigmatic

asymmetrical heterenchelyids from the Guinea coast of West Africa. J Fish Biol

91:673-678.

Martinez, C. M. and J. S. Sparks. 2017. Malagasy cichlids differentially limit impacts

of body shape evolution on oral jaw functional morphology. Evolution.

147

McDougall, I. and C. S. Feibel. 1999. Numerical age control for the Miocene-Pliocene

succession at Lothagam, a hominoid-bearing sequence in the northern Kenya

Rift. Journal of the Geological Society 156:731-745.

Melo, B. F., B. L. Sidlauskas, K. Hoekzema, B. W. Frable, R. P. Vari, and C. Oliveira.

2016. Molecular phylogenetics of the Neotropical fish family Prochilodontidae

(Teleostei: Characiformes). Mol Phylogenet Evol 102:189-201.

Melo, B. F., B. L. Sidlauskas, K. Hoekzema, R. P. Vari, and C. Oliveira. 2014. The

first molecular phylogeny of Chilodontidae (Teleostei: Ostariophysi:

Characiformes) reveals cryptic biodiversity and taxonomic uncertainty. Mol

Phylogenet Evol 70:286-295.

Menezes, N. A. 1969. The food of Brycon and three closely related genera of the Tribe

Acestrorhynchini. Papéis Avulsos ZooL S. Paulo 22:217-223.

Menezes, N. A. 2003. Family Acestrorhynchidae. Check List of the Freshwater Fishes

of South and Central America.

Menezes, N. A. and S. H. Weitzman. 2009. Systematics of the neotropical fish

subfamily Glandulocaudinae (Teleostei: Characiformes: Characidae). Neotrop

Ichthyol 7:295-370.

Mérigoux, S. and D. Ponton. 1998. Body shape, diet and ontogenetic diet shifts in

young fish of the Sinnamary River, French Guiana, South America. J Fish Biol

52:556-569.

Meyer, A. L. and J. J. Wiens. 2018. Estimating diversification rates for higher taxa:

BAMM can give problematic estimates of rates and rate shifts. Evolution 72:39-

53.

148

Miller, M. A., W. Pfeiffer, and T. Schwartz. 2010. Creating the CIPRES Science

Gateway for inference of large phylogenetic trees. Gateway Computing

Environments Workshop (GCE), 2010.

Mills, D. and G. Vevers. 1989. The Tetra Encyclopedia of Tropical Freshwater Fishes.

The Tetra Encyclopedia of Tropical Freshwater Fishes.

Mirande, J. 2010. Phylogeny of the family Characidae (Teleostei: Characiformes):

from characters to taxonomy. Neotrop Ichthyol 8:385-568.

Mirande, M. J. 2009. Weighted parsimony phylogeny of the family Characidae

(Teleostei: Characiformes). Cladistics 25:574-613.

Monsch, K. A. 1998. Miocene fish faunas from the northwestern Amazonia basin

(Colombia, Peru, Brazil) with evidence of marine incursions. Palaeogeography,

Palaeoclimatology, Palaeoecology 143:31-50.

Montaña, C. G., C. A. Layman, and K. O. Winemiller. 2011. Gape size influences

seasonal patterns of piscivore diets in three Neotropical rivers. Neotrop Ichthyol

9:647-655.

Moore, B. R., S. Höhna, M. R. May, B. Rannala, and J. P. Huelsenbeck. 2016. Critically

evaluating the theory and performance of Bayesian analysis of

macroevolutionary mixtures. PNAS 113:9569-9574.

Moreira, C. R., M. E. Bichuette, O. T. Oyakawa, D. M. C. C. Pinna, and E. Trajano.

2010. Rediscovery and redescription of the unusual subterranean characiform

Stygichthys typhlops, with notes on its life history. J Fish Biol 76:1815-1824.

Morris, S. C. 2003. Life's solution: inevitable humans in a lonely universe. Cambridge

University Press.

149

Munene, J. and M. L. J. Stiassny. 2012. A new Alestopetersius (Characiformes:

Alestidae) from the Kwilu River(Kasai basin) of central Africa; with a

phylogeny for the genus and synonymy of Duboisialestes. Zootaxa 3166:59-68.

Murray, A. M. 2003. A new Eocene citharinoid fish (Ostariophysi: Characiformes)

from Tanzania. J Vertebr Paleontol 23:501-507.

Muschick, M., A. Indermaur, and W. Salzburger. 2012. Convergent evolution within

an adaptive radiation of cichlid fishes. Curr Biol 22:2362-2368.

Muschick, M., P. Nosil, M. Roesti, M. Dittmann, L. Harmon, and W. Salzburger. 2014.

Testing the stages model in the adaptive radiation of cichlid fishes in East

African Lake Tanganyika. Proc R Soc Lond B Biol Sci 281:20140605.

Myers, G. S. and A. L. de Carvalho. 1959. A remarkable new genus of anostomin

characid fishes from the upper Rio Xingú in central Brazil. Copeia 1959:148-

152.

Near, T. J., R. I. Eytan, A. Dornburg, K. L. Kuhn, J. A. Moore, M. P. Davis, P. C.

Wainwright, M. Friedman, and L. W. Smith. 2012. Resolution of ray-finned

fish phylogeny and timing of diversification. PNAS 109:13698-13703.

Neiff, J., A. de Neiff, and M. B. Verón. 2009. The role of vegetated areas on fish

assemblage of the Paraná River floodplain: effects of different hydrological

conditions. Neotrop Ichthyol 7:39-48.

Nelson, J. S., T. C. Grande, and M. V. H. Wilson. 2016. Fishes of the World. John

Wiley, Hoboken.

Nico, L. G. and D. C. Taphorn. 1988. Food habits of piranhas in the low llanos of

Venezuela. Biotropica 20:311-321.

150

Norton, S. F. 1991. Capture success and diet of cottid fishes: the role of predator

morphology and attack kinematics. Ecology 72:1807-1819.

Noveras, J., K. C. Yamamoto, and C. E. C. Freitas. 2012. Use of the flooded forest by

fish assemblages in lakes of the National Park of Anavilhanas (Amazonas,

Brazil). Acta Amazonica 42:561-566.

Oliveira, C., G. S. Avelino, K. T. Abe, T. C. Mariguela, R. C. Benine, G. Ortí, R. P.

Vari, and R. M. e Castro. 2011. Phylogenetic relationships within the speciose

family Characidae (Teleostei: Ostariophysi: Characiformes) based on

multilocus analysis and extensive ingroup sampling. BMC Evol Biol 11:1-25.

Ornelas-García, C. P., A. Bautista, F. Herder, and I. Doadrio. 2017. Functional

modularity in lake-dwelling characin fishes of Mexico. PeerJ 5:e3851.

Ortí, G. and A. Meyer. 1997. The radiation of characiform fishes and the limits of

resolution of mitochondrial ribosomal DNA sequences. Syst Biol 46:75-100.

Page, R. D. and M. A. Charleston. 1998. Trees within trees: phylogeny and historical

associations. Trends Ecol Evol 13:356-359.

Pagel, M. and A. Meade. 2013. BayesTraits v. 2.0. Reading: University of Reading.

Parsons, K. J., E. Márquez, and R. C. Albertson. 2011. Constraint and opportunity: the

genetic basis and evolution of modularity in the cichlid mandible. Am. Nat,

179:64-78.

Paugy, D. 1978. Ecologie et biologie des Alestes baremoze (Pisces, Characidae) des

rivieres de Cote d'Ivoire. Cahiers ORSTOM. Série Hydrobiologie 12.

151

Pavanelli, C. 1999. Revisão taxonômica da família Parodontidae (Ostariophysi:

Characiformes). Unpublished D. Phil. Thesis Universidade Federal de São

Carlos.

Pavanelli, C. S. 2003. Family Parodontidae (Parodontids). Checklist of the freshwater

fishes of South y Central America. RE Reis, SO Kullander and C. Ferraris,

Jr.(eds.). EDIPUCRS, Porto Alegre, Brazil.

Pennell, M. W., J. M. Eastman, G. J. Slater, J. W. Brown, J. C. Uyeda, R. G. FitzJohn,

M. E. Alfaro, and L. J. Harmon. 2014. geiger v2.0: an expanded suite of

methods for fitting macroevolutionary models to phylogenetic trees.

Bioinformatics 30:2216-2218.

Pereira, P., C. Agostinho, R. de Oliveira, and E. Marques. 2007. Trophic guilds of

fishes in sandbank habitats of a Neotropical river. Neotrop Ichthyol 5:399-404.

Peretti, D. and I. de Andrian. 2004. Trophic Structure of Fish Assemblages in Five

Permanent Lagoons of the High Paraná River Floodplain, Brazil. Environ Biol

Fishes 71:95-103.

Planquette, P., P. Keith, P. Y. Bail, and C. de la Peche. 1996. Atlas des poissons d'eau

douce de Guyane (Tome 1), Paris.

Pouilly, M., S. Barrera, and C. Rosales. 2006. Changes of taxonomic and trophic

structure of fish assemblages along an environmental gradient in the Upper Beni

watershed (Bolivia). J Fish Biol 68:137-156.

Pouilly, M., F. Lino, J. G. Bretenoux, and C. Rosales. 2003. Dietary–morphological

relationships in a fish assemblage of the Bolivian Amazonian floodplain. J Fish

Biol 62:1137-1158.

152

Pouilly, M., D. Rejas, T. Pérez, J.-L. Duprey, C. I. Molina, C. Hubas, and J.-R. D.

Guimarães. 2013. Trophic structure and mercury biomagnification in tropical

fish assemblages, Iténez River, Bolivia. PLoS ONE 8:e65054.

Pouilly, M., T. Yunoki, C. Rosales, and L. Torres. 2004. Trophic structure of fish

assemblages from Mamoré River floodplain lakes (Bolivia). Ecol Freshw Fish

13:245-257.

Price, S. A., P. C. Wainwright, D. R. Bellwood, E. Kazancioglu, D. C. Collar, and T.

J. Near. 2010. Functional innovations and morphological diversification in

parrotfish. Evolution 64:3057-3068.

Rabosky, D. L. 2009. Ecological limits and diversification rate: alternative paradigms

to explain the variation in species richness among clades and regions. Ecol Lett

12:735-743.

Rambaut, A. and A. Drummond. 2007. Tracer v1. 4: MCMC trace analyses tool. See

http://tree. bio. ed. ac. uk/software/tracer.

Revell, L. J. 2012. phytools: an R package for phylogenetic comparative biology (and

other things). Methods Ecol. Evol. 3:217-223.

Reynolds, J. D. 1973. Biology and fisheries potential of four species of Alestes (Pisces-

Characinidae) in the new Volta Lake, Ghana. Rev. Zool Bot. Afri. 2:298-310.

Ricklefs, R. E. 2006. Global variation in the diversification rate of passerine birds.

Ecology 87:2468-2478.

Ricklefs, R. E. 2010. Evolutionary diversification, coevolution between populations

and their antagonists, and the filling of niche space. PNAS 107:1265-1272.

153

Riehl, R. and H. A. Baensch. 1996. Aquarien Atlas, Band 1. Melle: Mergus Verlag fu¨r

Natur- und Heimtierkunde.

Roberts, T. 1973. Interrelationships of ostariophysans. Zool J Linn Soc 53:373-395.

Roberts, T. R. 1972. Ecology of fishes in the Amazon and Congo basins. Bull Mus

Comp Zool 143:117-147.

Robinson, A. H. and P. J. Robinson. 1969. A comparative study of the food habits of

micralestes acutidens and alestes dageti(pisces: Characidae) from the northern

basin of lake chad. BULL INST FONDAM AFR NOIRE SER A SCI NATUR.

31 (3): 951-964. ILLUS. MAP. 1969.

Rohlf, J. F. 2015. The tps series of software. Hystrix, the Italian Journal of Mammalogy

26:9-12.

Román-Valencia, V. and D. C. Taphorn. 2014. Trophic and reproductive ecology of a

Neotropical characid fish Hemibrycon brevispini (Teleostei: Characiformes).

Caldasia 36:289-304.

Ronquist, F. and J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference

under mixed models. Bioinformatics 19:1572-1574.

Röpke, C. P., E. Ferreira, and J. Zuanon. 2014. Seasonal changes in the use of feeding

resources by fish in stands of aquatic macrophytes in an Amazonian floodplain,

Brazil. Environ Biol Fishes 97:401-414.

Rossi, M. F., B. Mello, and C. G. Schrago. 2018. Comparative evaluation of

macroevolutionary regimes of Ruminantia and selected mammalian lineages.

Biol J Linn Soc.

154

Rüber, L., E. Verheyen, and A. Meyer. 1999. Replicated evolution of trophic

specializations in an endemic cichlid fish lineage from Lake Tanganyika.

Proceedings of the National Academy of Sciences 96:10230-10235.

Rundle, H. D., L. Nagel, J. W. Boughman, and D. Schluter. 2000. Natural selection and

parallel speciation in sympatric sticklebacks. Science 287:306-308.

Sabino, J. and C. R. M. C. Castro. 1989. Alimentação, período de atividade e

distribuição espacial dos peixes de um riacho da Floresta Atlântica (Sudeste do

Brasil). Rev Bras Biol 50.

Sandon, H. and A. al Tayib. 1953. The food of some common Nile fish. Sudan Notes

Rec 34:205-229.

Sanger, T. J., L. J. Revell, J. J. Gibson-Brown, and J. B. Losos. 2012. Repeated

modification of early limb morphogenesis programmes underlies the

convergence of relative limb length in Anolis lizards. Pp. 739-748. Proc. R.

Soc. B. The Royal Society.

Santos, G. and P. D. S. Rosa. 1998. Alimentação de Anostomus ternetzi e

Synaptolaemus cingulatus, duas espécies de peixes amazônicos com boca

superior. Rev Bras Biol 58.

Saul, W. G. 1975. An ecological study of fishes at a site in upper Amazonian Ecuador.

Proc Acad Nat Sci Phila 127:93-134.

Sazima, I. 1986. Similarities in feeding behaviour between some marine and freshwater

fishes in two tropical communities. J Fish Biol 29:53-65.

155

Sazima, I. and C. Zamprogno. 1985. Use of water hyacinths as shelter, foraging place,

and transport by young piranhas, Serrasalmus spilopleura. Environ Biol Fishes

12:237-240.

Schenk, J. J., K. C. Rowe, and S. J. Steppan. 2013. Ecological opportunity and

incumbency in the diversification of repeated continental colonizations by

muroid rodents. Syst Biol 62:837-864.

Schluter, D. 1995. Adaptive Radiation in Sticklebacks: Trade‐Offs in Feeding

Performance and Growth. Ecology 76:82-90.

Schluter, D. 2000. The ecology of adaptive radiation. Oxford University Press, New

York.

Schwenk, K. and G. P. Wagner. 2001. Function and the evolution of phenotypic

stability: connecting pattern to process. Am Zool 41:552-563.

Seehausen, O. 2006. African cichlid fish: a model system in adaptive radiation

research. Proc R Soc Lond B Biol Sci 273:1987-1998.

Seehausen, O. 2015. Process and pattern in cichlid radiations–inferences for

understanding unusually high rates of evolutionary diversification. New Phytol

207:304-312.

Sidlauskas, B. 2008. Continuous and arrested morphological diversification in sister

clades of characiform fishes: A phylomorphospace approach. Evolution

62:3135-3156.

Sidlauskas, B., B. Chernoff, and A. Machado-Allison. 2006. Geographic and

environmental variation in Bryconops sp. cf. melanurus (Ostariophysi:

Characidae) from the Brazilian Pantanal. Ichthyol Res 53:24-33.

156

Sidlauskas, B. L. and G. dos Santos. 2005. Pseudanos winterbottomi: A New

Anostomine Species (Teleostei: Characiformes: Anostomidae) from Venezuela

and Brazil, and Comments on Its Phylogenetic Relationships. Copeia 2014:109-

123.

Sidlauskas, B. L., J. H. Mol, and R. P. Vari. 2011. Dealing with allometry in linear and

geometric morphometrics: a taxonomic case study in the Leporinus

cylindriformis group (Characiformes: Anostomidae) with description of a new

species from Suriname. Zool J Linn Soc 162:103-130.

Sidlauskas, B. L. and R. P. Vari. 2008. Phylogenetic relationships within the South

American fish family Anostomidae (Teleostei, Ostariophysi, Characiformes).

Zool J Linn Soc 154:70-210.

Silva, C. P. D. 1993. Alimentação e distribuição espacial de algumas espécies de peixes

do igarapé do Candirú, Amazonas, Brasil. Acta Amazonica 23:271-285.

Simpson, G. G. 1953. The major features of evolution. Columbia University Press,

New York.

Sivasundar, A., E. Bermingham, and G. Orti. 2001. Population structure and

biogeography of migratory freshwater fishes (Prochilodus: Characiformes) in

major South American rivers. Mol Ecol 10:407-417.

Skelton, P. H. 1993. A complete guide to the freshwater fishes of southern Africa.

Southern Book Publishers, Ltd, Halfway House, South Africa.

Smith, J. M., R. Burian, S. Kauffman, P. Alberch, J. Campbell, B. Goodwin, R. Lande,

D. Raup, and L. Wolpert. 1985. Developmental constraints and evolution: a

157

perspective from the Mountain Lake conference on development and evolution.

The Quarterly Review of Biology 60:265-287.

Standen, E. 2008. Pelvic fin locomotor function in fishes: three-dimensional kinematics

in rainbow trout (Oncorhynchus mykiss). J Exp Biol 211:2931-2942.

Stayton, T. C. 2015. The definition, recognition, and interpretation of convergent

evolution, and two new measures for quantifying and assessing the significance

of convergence. Evolution 69:2140-2153.

Stewart, K. 2003. Fossil fish remains from Mio-Pliocene deposits atLothagam, Kenya.

Pp. 75-111 in M. Leakey, and J. Harris, eds. Lothagam: the dawn of humanity

in Eastern Africa. Columbia University Press, New York.

Stewart, K. M. 2001. The freshwater fish of Neogene Africa (Miocene–Pleistocene):

systematics and biogeography. Fish Fish 2:177-230.

Streelman, J. T. and P. D. Danley. 2003. The stages of vertebrate evolutionary

radiation. Trends Ecol Evol 18:126-131.

Stubbs, T. L. and M. J. Benton. 2016. Ecomorphological diversifications of Mesozoic

marine reptiles: the roles of ecological opportunity and extinction. Paleobiology

42:547-573.

Taphorn, D. 1992. The characiform fishes of the Apure River drainage, Venezuela.

Biollania, edicion especial 4.

Thompson, A. W., R. Betancur-R, H. López-Fernández, and G. Ortí. 2014. A time-

calibrated, multi-locus phylogeny of piranhas and pacus (Characiformes:

Serrasalmidae) and a comparison of species tree methods. Mol Phylogenet Evol

81:242-257.

158

Vari, R. P. 1979. Anatomy, relationships, and classification of the families Citharinidae

and Distichodontidae (Pisces, Characoidea). Trustees of the British Museum

(Natural History).

Vari, R. P. 1983. Phylogenetic relationships of the Families Curimatidae,

Prochilodontidae, Anostomidae, and Chilodontidae (Pisces, Characiformes).

Smithson Contrib Zool 378:1-60.

Vari, R. P. 1989. A phylogenetic study of the Neotropical characiform family

Curimatidae (Pisces: Ostariophysi).

Vari, R. P., R. M. C. Castro, and S. J. Raredon. 1995. The Neotropical fish family

Chilodontidae (Teleostei: Characiformes): a phylogenetic study and a revision

of Caenotropus Günther. Smithson Contrib Zool 577:1-32.

Vaz, M. M., M. Petrere, L. A. Martinelli, and A. A. Mozeto. 1999. The dietary regime

of detritivorous fish from the River Jacaré Pepira, Brazil. Fish Manag Ecol

6:121-132.

Verbeke, J. 1959. Le Régime alimentaire des poissons des Lacs Edouard et Albert

(Congo belge). Institut royal des sciences naturelles de Belgique.

Vieira, I. and J. Géry. 1979. Crescimento diferencial e nutrição em Catoprion mento

(Characoidei). Peixe lepidófago da Amazônia. Acta Amazonica 9:143-146.

Vonhof, H., F. Wesselingh, R. Kaandorp, G. Davies, J. Van Hinte, J. Guerrero, M.

Räsänen, L. Romero-Pittman, and A. Ranzi. 2003. Paleogeography of Miocene

Western Amazonia: Isotopic composition of molluscan shells constrains the

influence of marine incursions. Geol Soc Am Bull 115:983-993.

159

Wake, D. B. 1991. Homoplasy: the result of natural selection, or evidence of design

limitations? Am. Nat, 138:543-567.

Wake, M. H. 1986. The morphology of Idiocranium russeli (Amphibia:

Gymnophiona), with comments on miniaturization through heterochrony. J

Morphol 189:1-16.

Wantzen, K. M., F. de Machado, M. Voss, H. Boriss, and W. J. Junk. 2002. Seasonal

isotopic shifts in fish of the Pantanal wetland, Brazil. Aquat Sci 64:239-251.

Webb, P. 1984a. Body and fin form and strike tactics of four teleost predators attacking

fathead minnow (Pimephales promelas) prey. Can J Fish Aquat Sci 41:157-165.

Webb, P. W. 1982. Locomotor patterns in the evolution of actinopterygian fishes. Am

Zool 22:329-342.

Webb, P. W. 1984b. Form and function in fish swimming. Scientific American 251:72-

83.

Webb, P. W. and D. Weihs. 1986. Functional locomotor morphology of early life

history stages of fishes. Trans Am Fish Soc 115:115-127.

Weitzman, S. H. 1954. The osteology and the relationships of the South American

characid fishes of the subfamily Gasteropelecinae. Stanford Ichthyological

Bulletin 4:213-263.

Weitzman, S. H. and W. L. Fink. 1983. Relationships of the Neon Tetras: A Group of

South American Freshwater Fishes (Teleostei, Characidae): with Comments on

the Phylogeny of New World Characteriforms. Harvard University, Museum of

Comparative Zoology.

160

Willis, S., K. Winemiller, and H. Lopez-Fernandez. 2005. Habitat structural

complexity and morphological diversity of fish assemblages in a Neotropical

floodplain river. Oecologia 142:284-295.

Winemiller, K. O. 1991. Ecomorphological Diversification in Lowland Freshwater

Fish Assemblages from Five Biotic Regions. Ecol Monogr 61:343-365.

Zanata, A. M. and J. P. Serra. 2010. Hasemania piatan, a new characid species

(Characiformes: Characidae) from headwaters of rio de Contas, Bahia, Brazil.

Neotrop Ichthyol 8:21-26.

Zanata, A. M. and R. P. Vari. 2005. The family Alestidae (Ostariophysi,

Characiformes): a phylogenetic analysis of a trans-Atlantic clade. Zool J Linn

Soc 145:1-144.

161

APPENDICES

Appendix 1: Supplementary materials for Chapter 2

S1. Phylogenetic data acquisition, quality control and alignment

We synthesized previously published sequence data to assemble a molecular dataset spanning 129 terminal taxa and representing 19 families in the order

Characiformes, resulting in average of 30% taxon sampling within each family. We used a combination of 365 sequences obtained from GenBank and 99 sequences generated in this study. Sequences were obtained from Genbank from the following studies on characiform and fish phylogenetics: (Calcagnotto et al. 2005; Arroyave and

Stiassny 2011; Oliveira et al. 2011; Near et al. 2012; Arroyave et al. 2013; Mariguela et al. 2013b; Abe et al. 2014; Melo et al. 2014; Melo et al. 2016). Phylogenies were rooted at the split between the Citharinoidei and Characoidei, as widely supported by molecular (Chen et al. 2013; Arcila et al. 2017; Chakrabarty et al. 2017) and morphological data (Fink 1996; Mirande 2009; Mirande 2010). The quality of the sequences obtained from GenBank were assessed by testing sequence similarity in The

Basic Local Alignment Search Tool (BLAST) on the NCBI web interface. Sequences that shared a closer similarity to species outside of its putative genus than to congeners were considered misidentified or contaminated and removed from subsequent analyses.

In generating new sequences, we targeted the four most frequently sequenced genes in characiform systematics, including two mitochondrial genes (16S rRNA [16S] and cytochrome B [Cytb]) and two nuclear genes (myosin heavy chain 6 [Myh6], and recombination activating gene 2 [Rag2]). For the unique sequences, DNA was extracted using a DNeasy Tissue Kit (Qiagen Inc). The mitochondrial genes were

162 amplified through standard polymerase chain reaction (PCR) following the protocols of recent phylogenetic studies on characiforms (Oliveira et al. 2011; Melo et al. 2014;

Melo et al. 2016). The nuclear genes were amplified by nested-PCR following the protocol of Li et al. (2007). Table 7.1 presents a complete listing of genes per species, corresponding GenBank accession numbers and tissue voucher numbers.

Sequences were aligned using default parameters in the Muscle (Edgar 2004) plugin for Geneious version 9.30 (Kearse et al. 2012). The sequences in the alignment were inspected by eye for major misalignments including large gaps and short sequence reads. No major misalignments were identified in the protein coding genes, so realignment was unnecessary. However, we excised the messy regions of the new sequences in the16S matrix that would not align with sequences from Genbank because length polymorphisms make these regions difficult to align unambiguously. Individual gene trees were created in a RAxML analysis run through the CIPRES web server

(Miller et al. 2010) and MrBayes 3.12 analysis (Ronquist and Huelsenbeck 2003) to identify potentially misidentified or contaminated taxa, defined as those subtended by unrealistically long branches (i.e. branch lengths ~10 times longer than other congeners) or those that do not cluster within their respective families. Unreasonably long branch lengths can be caused by contamination, chimaeric sequences or paralogy, any and all of which can lead to spurious phylogenetic inference (Page and Charleston

1998; Degnan and Rosenberg 2009). Sequences that did not cluster within the expected family were considered contaminated or misidentified because the monophyly of each characiform family has been well established by morphological (e.g., Vari 1979, 1983;

Buckup 1993a; Sidlauskas and Vari 2008) and molecular studies (Oliveira et al. 2011;

163

Melo et al. 2014,2016; Arcila et al. 2017). However, we did not remove any sequences from the family Characidae because many of the genera are considered incertae sedis and we did not want to bias the results (Mirande 2009; Oliveira et al. 2011).

Additionally, we did not remove any sequences from the genus Chalceus because the genus consistently falls outside of the Alestidae in molecular analyses (Mirande 2009;

Oliveira et al. 2011), despite its inclusion in Alestidae in a recent morphological paper

(Zanata and Vari 2005). Four individual sequences with such dubious identifications were removed prior to concatenation. We concatenated the four genes into a supermatrix of 10 partitions (one partition for 16S and each codon for the remaining genes). Partitions and parameters were determined by Partitionfinder2 version 2.1.1

(Lanfear et al. 2016).

S2. Additional phylogenetic analysis

A Maximum Likelihood (ML) topology was inferred with a partitioned

RAxML analysis run through the CIPRES web server (Miller et al. 2010) with members of the Citharinoidei set as the outgroup. A random starting tree was used for the ML search with all other parameters set to default values. Optimal partition schemes and models of molecular evolution were inferred using Partitionfinder2 (Lanfear et al.

2016) with the examined models restricted to those available in RAxML (Stamatakis

2008). All ML analyses were performed under GTR + I + G as it was the most complex model suggested for any partition. The robustness of the ML topology was tested with

10000 bootstrap replicates. A Bayesian topology was inferred using MrBayes 3.12

(Ronquist and Huelsenbeck 2003) run through the CIPRES web server (Miller et al.

164

2010) using the parameters suggested by Partitionfinder2 (Lanfear et al. 2016). We performed 10 million generations, sampling a tree every 1000 generation. The distribution of log likelihood scores was examined in Tracer 1.5 (Rambaut and

Drummond 2007) to determine if the runs had reached convergence by assessing ESS scores >200 and stationarity of log likelihood scores. The first 10% of the posterior was disregarded as burnin after visualizing the posterior distribution in Tracer 1.5 (Rambaut and Drummond 2007), with the remaining trees used to construct a maximum clade credibility tree.

S3. Descriptions of priors and calibration points used in BEAST 2.0 analysis

We implemented the uncorrelated lognormal distribution (UCLN) rate variation model because previous studies found it to be the most accurate and robust for analysis at similar phylogenetic scale (Drummond et al. 2006; Arroyave and Stiassny 2011).

We provided rough starting clock rate estimates for each clock model in BEAST 2.0 (

165

Table 7.6) based on previous studies substitution rates in the order (Sivasundar et al. 2001; Arroyave et al. 2013). A birth-death tree prior was chosen for node time estimation because this models the distribution under a model where speciation and extinction rates can affect a lineage at any time. A birth death model is considered the most appropriate model when extinction is known to have occurred in a group

(Drummond et al. 2006).

We calibrated the model with five fossils to provide minimum age estimates for specific lineages on the tree. We implemented lognormal distributed priors on the fossil calibrations because many of the fossils are likely from derived clades (Sidlauskas and

Vari 2008; Arroyave et al. 2013) and the modern clades may be much older than the minimum bound (

166

Table 7.4). Mean and standard deviation were estimated by incorporating calibration settings from other studies that included the same fossil material (Arroyave et al. 2013; Abe et al. 2014; Thompson et al. 2014).

The five fossil calibrations were based on the materials below and represent the minimum bound date in the calibration. To date the suborder Citharinoidei two fossils were used: † Eocitharinus macrognathus from the Eocene (46 ma) and fossilized dentition of the genus Distichodus from the Late Miocene (7.5 ma). † Eocitharinus macrognathus has many features that suggest a close relationship with the families

Citharinidae and Distochodontidae (Arroyave et al. 2013), but lacks determinable synapomorphies of either family and has been classified as Citharinoidei incertae sedis

(Murray 2003; Malabarba and Malabarba 2010). We incorporated the age of †

Eocitharinus macrognathus as a stem member of the Citharinoidei. The other calibration point in the Citharinoidei was a single tooth belonging to a member of the genus Distichodus recovered from the Late Miocene strata of the Lower Nawata formation in Lothagam Kenya (McDougall and Feibel 1999; Stewart 2001; Feibel

2003; Stewart 2003). The Distichodus dentition recovered from the strata is diagnostic of the genus (Stewart 2003) and we incorporated the Late Miocene as a minimum age of the genus.

Three fossils were used to date nodes in the suborder Characoidei: fossilized teeth corresponding to the family Serrasalmidae, oral teeth corresponding to the genus

Leporinus, and † Cyphocharax mosesi. The first fossil calibration in the Characoidei was isolated pacu-like premaxillary teeth from the El Molino Formation. Bolivia from the Ctreteaceous-Plaeocene (70-61 ma) that represents the most recent common

167 ancestor of Colossoma, Metynnis, and Piractus (Dahdul 2010). We incorporated the age of the fossilized teeth as a minimum age of the node of Colossoma and Piractus.

The next fossil was two oral teeth from the La Tagua, Pebas formation, Colombia dating to the Early Middle Miocene 23-15 ma (Monsch 1998; Vonhof et al. 2003) that are distinctive for the genus Leporinus (Monsch 1998). We incorporated the node of the fossilized oral teeth as the minimum age for the genus Leporinus. The final fossil was † Cyphocharax mosesi from the Oligocene strata (33.9-23.0 ma) from the

Tremembe Formation, Sao Paulo, Brazil (Malabarba 1996). † Cyphocharax mosesi was described within a polytomy of the genera Cyphocharax, Curimatella, and

Steindachnerina (Malabarba 1996) and we used † Cyphocharax mosesi to calibrate the base of clade containing those three genera.

S4. Alternative topologies

Our tree topology differed slightly from the most recent comprehensive molecular phylogeny of the order (Oliveira et al. 2011, see Figure 7.1 for differences).

To assess whether the differences between these alternative hypotheses influenced our understanding of the macroevolutionary processes, we created a constraint topology matching that of Oliveria et al. 2011 in Mesquite version 3.31 (Maddison and Maddison

2017) and inferred the most probable ultrametric version thereof under a Bayesian relaxed-clock model in BEAST 2.0 (Bouckaert et al. 2014) through the CIPRES web server (Miller et al. 2010). We ran all phylogenetic comparative methods using the unconstrained phylogeny, and the phylogeny conforming to the Oliveira et al. (2011) hypothesis.

168

S4. Geometric morphometric analysis

Geometric morphometric data were collected from specimens radiographed using a Philips MOD 301/4 digital x-ray machine at the Smithsonian Museum of

Natural History. Body shape diversity was measured using 24 landmarks taken on the left side of the specimen (Figure 2.1). Landmarks were digitized in tpsDig v 2.17

(Rohlf 2015). Individual landmarks were rotated, scaled, and aligned using a least- squares Generalized Procrustes Analysis in the R package geomorph (Adams and

Otárola‐Castillo 2013). Principal component analysis (PCA) in geomorph on the

Procrustes coordinates identified the primary axes of shape variation among species in the dataset. Procrustes distances were regressed against distances in the tangent PCA space in tpsSmall v1.32 (Rohlf 2015) to ensure that the projection into tangent space did not greatly distort interspecies distances.

S5. Ecological classification

The assignment of each species to a guild was indexed with one of four different tiers of certainty (Table 7.2). Tier 1 species had two or more corroborating and non- contradicting sources of diet data. Tier 2 species had only one source of diet data for the species. Tier 3 species had no data available for that specific species but were assigned to match the ecology of congeners for which data were available. Tier 4 species had multiple contradicting trophic classifications and were therefore categorized as omnivores.

169

Trophic classification was difficult for many Tier IV species because different literature sources assigned them to different ecological guilds (Table 7.2). While we classified these species as omnivorous, such species may not be strictly omnivorous, but rather may shift their diets at different life stages, in different times of the year, or in different locations. To address the uncertainty of trophic classification in subsequent analyses, all phylogenetic comparative methods outlined below were performed across

1000 random combinations of trophic ecologies in species that exhibited trophic uncertainty. For example, we ran the phylogenetic comparative analyses with all the omnivores that included a substantial portion of vegetable matter and invertebrates in their diet as either herbivores, insectivores, or omnivores during each iteration. To do this, we randomly assigned each species with trophic uncertainty to one of its possible trophic classifications during each iteration. We analyzed 1000 random trophic combinations to determine whether trophic uncertainty or shifts in trophic ecology across time and space influenced our macroevolutionary interpretations.

S6. Topological differences

Our tree differs from the most densely sampled molecular phylogeny, Oliveira et al. 2011, in three important respects (Figure 7.1). The first difference involves the placement of the African clades of Alestidae + Hepsetidae as sister to a South American clade containing Characidae + Triporthidae + Chalceidae + Iguanodectidae +

Acestrorynchidae +Roestinae (Figure 7.1), as opposed to with the rest of Characoidei sensu Oliveira et al. (2011). The second discrepancy is the recovery herein of a paraphyletic Triporthidae with Engraulisoma taeniatum sister to the Gasteropelecidae,

170 while Oliveira et al (2011) inferred a monophyletic Triporthidae. Lastly, our topology places the triporthids as sister to the Gasteropelecidae + Salminus + Brycon, while

Oliveira et al. (2011) has the triporthids as sister to Chalceus. In our topology, Chalceus appears sister to Iguanodectidae, whereas Oliveira et al. 2011 placed the

Iguanodectidae as sister to Acestrorynchidae +Roestinae. All other major lineages are congruent in the results of the two studies.

S7. Trophic ecology

Omnivory was the most common trophic guild, with 50 species or 42.7% of the total species sampled. Detritivores were the second most common with 23 species

(19.6% of total). Invertivores followed with 16 species (12.8%), followed by insectivores and piscivores, each with 10 species (8.5%). Herbivores were the least common trophic strategy with only 7 individuals representing 5.9% of the total sampled diversity.

All trophic groups most likely evolved from an omnivorous ancestor, except for a single evolution of invertivory from insectivory. In the BayesTraits analysis, trophic ecology tended to move away from omnivory and rarely towards it, with the transition rates towards omnivory much lower than the rates away (Figure 7.3). Detritivory and piscivory tended to evolve early during the formation of larger clades, but rarely at the tips within the smaller subclades. Conversely, invertivory and herbivory evolved later in cladogenesis at the tips of the tree during smaller subclade formation. Insectivory exhibited both patterns, but more frequently evolved near the tips of the phylogeny.

171

Appendix 2: Supplemental figures for Chapter 2

Figure 7.1 Phylogenetic hypotheses for the major characiform lineages from the current study and Oliveira et al. 2011.

172

Figure 7.2. Relationships among sampled species in the order Characiformes returned by Maximum Likelihood analysis of the concatenated dataset in RAxML. Circles at each node indicate bootstrap support. Red represents bootstrap support of 100, blue represents a score greater than 90, purple represents a score greater than 70, and lack of a circle represents a score below 70.

173

Figure 7.3. Results of BayesTraits analysis showing rate of transition between omnivory (black arrows) and the other trophic classes. Rates of transition was always faster away from omnivory (black arrows) than towards omnivory (white arrows).

Arrows represent proportion of rate scores from BayesTraits analysis.

174

Figure 7.4. Histogram of AICc scores from mvMORPH analysis when simulating 1000 times across uncertainty index, always showing AICc scores larger than the empirical score (arrow).

175

Figure 7.5. Most likely phylogeny constrained to the interfamilial relationships in

Oliveira et al. 2011, showing adaptive optima and regimes for the best-fitting model of body shape evolution assigned based on the SURFACE results.

176

Appendix 3: Supplemental tables for Chapter 2

Table 7.1 GenBank numbers for tissues used in phylogenetic analysis.

Family Species RAG2 myh6 16S Cytb Acestorynchidae Acestrorhynchus minimus HQ289408.1 HQ289022.1 HQ171312.1 HQ289601.1 Acestorynchidae Acestrorhynchus pantaneiro HQ289385.1 HQ288998.1 HQ171288.1 HQ289577.1 Alestidae Alestes baremoze AY804029.1 JF801070.1 AY787963.1 AY787963.1 Alestidae Alestes sp. HQ289483.1 HQ289098.1 HQ171390.1 HQ289677.1 Alestidae Alestoptersius hilgendorfi AY804114.1 JF801060.1 AY788070.1 JF801015.1 Alestidae Bathyaethiops breuseghemi AY804113.1 JN710419.1 - JN710403.1 Alestidae Brycinus longipinnis AY804044.1 JF801054.1 AY787980.1 JF801009.1 Alestidae Brycinus nurse AY804034.1 JF801055.1 AY787970.1 AY791366.1 Alestidae Bryconaethiops microstoma AY804041.1 JF801050.1 - JF801007.1 Alestidae Dubioalestes tumbensis - JF801063.1 - JF801018.1 Alestidae Ladigesia roloffi - JF801101.1 AY788046.1 - Alestidae Micralestes acutidens - JF801044.1 AY788047.1 - Alestidae Micralestes occidentalis - JF801046.1 AY787961.1 - Alestidae Rhabdalestes septentrionatus - JF801094.1 AY788077.1 - Genbank # Genbank # Genbank # Genbank # Anostomidae Abramites hypselumotus pending pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Anostomoides laticeps pending pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Anostomus ternetzi pending pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Gnatholodus bidens pending pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Hypomasticus megalepis pending pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Laemolyta proxima pending pending pending pending Genbank # Genbank # Genbank # - Anostomidae Leporinus agassizi pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Leporinus brunneus pending pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Leporinus fasciatus pending pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Leporinus nigrotaeniatus pending pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Petulanos spiloclistron pending pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Pseudanos gracilis pending pending pending pending Genbank # Genbank # - - Anostomidae Pseudanos winterbottomi pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Rhytoduys lauzannei pending pending pending pending

177

Genbank # Genbank # Genbank # - Anostomidae Sartor respectus pending pending pending Genbank # Genbank # Genbank # Genbank # Anostomidae Schizodon knerii pending pending pending pending Genbank # Genbank # - - Anostomidae Schizodon fasciatus pending pending Distichodontidae Distichodus decemmaculatus AY804065 KF542247.1 AY788005.1 AY791390.1 Distichodontidae Distichodus fasciolatus AY804067.1 KF542240.1 AY788008.1 AY791393.1 Distichodontidae Hemigrammocharax unioccelatus AY804085.1 KF542293.1 AY788029.1 AY791407.1 Distichodontidae Ichthyborus ornatus AY804092.1 KF542282.1 AY788038.1 AY791412.1 Distichodontidae Neolebias unifasciatus AY804108.1 KF542314.1 AY788063.1 AY791425.1 Characidae Aphyrocharax pusillulus JQ820033.1 JQ820066.1 JQ820096.1 JQ820078.1 Characidae Astyanacinus moori HQ289447.1 HQ289061.1 HQ171352.1 HQ289641.1 Characidae Bario steindachneri HQ289415.1 HQ289028.1 - HQ289608.1 Characidae Bramocharax baileyi HQ289513.1 HQ289128.1 HQ171420.1 HQ289706.1 Characidae Brycon amazonicus KF780118.1 KF780047.1 KF779970.1 KF780011.1 Characidae Bryconamericus emperador HQ289363.1 HQ288976.1 KF209733.1 HQ289557.1 Characidae Carlana eigenmanni HQ289374.1 HQ288987.1 HQ171277.1 HQ289566.1 Characidae Chalceus erythrureus HQ289394.1 HQ289007.1 HQ171297.1 HQ289586.1 Characidae Compsura heterura - KC189605.1 KC110752.1 KC189565.1 Characidae Creagrutus affinis - KF210362.1 KF209799.1 - Characidae Cyanocharax dicropotamicus KF211112.1 KF210399.1 KF209840.1 - Characidae Deuterodon iguape HQ289460.1 HQ289074.1 HQ171366.1 HQ289653.1 Characidae Diapoma terofali - KF210416.1 KF209855.1 - Characidae Glandulocauda melanopleura - KF210447.1 KF209898.1 - Characidae Gymnocorymbus ternetzi KP959267.1 KP959281.1 KP959289.1 HQ289576.1 Characidae Hasmania hanseni HQ289470.1 - HQ171376.1 HQ289663.1 Characidae Hemibrycon dariensis - KF210460.1 KF209911.1 - Characidae Heterocheirodon yatai HQ289426.1 HQ289039.1 HQ171330.1 HQ289619.1 Characidae Hollandichthys multifasciatus HQ289526.1 HQ289142.1 HQ171434.1 HQ289719.1 Characidae Hyphessobrycon megalopterus HQ289490.1 HQ289105.1 HQ171397.1 HQ289684.1 Characidae Jupiaba anteroides HQ289475.1 HQ289089.1 HQ171381.1 HQ289668.1 Characidae Knodus serptentrionalis HQ289355.1 HQ288967.1 HQ171257.1 HQ289548.1 Characidae Macropsobrycon uruguayanae HQ289450.1 HQ289064.1 - HQ289644.1 Characidae Markiana geayi HQ289524.1 HQ289140.1 HQ171432.1 HQ289717.1 Characidae Mimagoniates microlepis KF211226.1 KF210534.1 KF209990.1 HQ289531.1 Characidae Moenkhausia xinguensis HQ289448.1 HQ289062.1 HQ171353.1 HQ289642.1 Characidae Odontostilbe splendida KC196413.1 KC189600.1 KC110747.1 KC189560.1 Characidae Paracheirodon axelrodi - KF210433.1 KF209876.1 HQ289611.1 Characidae Paragoniates alburnus HQ289519.1 HQ289134.1 HQ171426.1 HQ289712.1 Characidae Phenacogaster calverti HQ289442.1 HQ289056.1 HQ171347.1 HQ289636.1 Characidae Piabarchus analis HQ289505.1 HQ289120.1 HQ171412.1 HQ289699.1

178

Characidae Piabina argentea - KF210548.1 KF210006.1 GU908222.1 Characidae Poptella paraguayensis HQ289383.1 HQ288996.1 HQ171286.1 HQ289575.1 Characidae Prionobrama paraguayensis HQ289369.1 HQ288982.1 HQ171272.1 JQ820073.1 Characidae Prodontocharax KC196386.1 KC189569.1 KC110716.1 KC189529.1 Characidae Rachoviscus crassiceps HQ289465.1 HQ289079.1 HQ171371.1 HQ289658.1 Characidae Rhinobrycon negrensis KF211247.1 KF210557.1 KF210017.1 - Characidae Roeboexodon guyanensis HQ289440.1 HQ289054.1 HQ171345.1 HQ289634.1 Characidae Salminus hilarii KP342015.1 KF780081.1 KF780006.1 - Characidae Spintherobolus broccae HQ289391.1 HQ289004.1 HQ171294.1 HQ289583.1 Characidae Stigichthys typhlops HQ289500.1 HQ289115.1 HQ171407.1 HQ289694.1 Characidae Tetragonopterus argenteus KT895170.1 KT895129.1 KT880476.1 KT895110.1 Characidae Thayeria obliqua HQ289439.1 HQ289053.1 HQ171344.1 HQ289633.1 Chilodontidae Caenotropus labyrinthicus - KF562463.1 KF562380.1 KF562438.1 Chilodontidae Caenotropus maculosus - KF562465.1 KF562383.1 KF562441.1 Chilodontidae Chilodus punctatus - KF562483.1 KF562406.1 - Citharinidae Citharinus sp. HQ289481.1 HQ289096.1 HQ171388.1 HQ289675.1 Crenuchidae Characidium laterale HQ289491.1 HQ289106.1 HQ171398.1 HQ289685.1 Crenuchidae Characidium pterostictum HQ289381.1 HQ288994.1 HQ171284.1 HQ289573.1 Crenuchidae Crenuchus spirulus HQ289471.1 HQ289085.1 HQ171377.1 - Crenuchidae Melanocharacidium sp. AY804126.1 - AY788083.1 AY791439.1 Crenuchidae Poecilocharax weitzmani HQ289507.1 HQ289122.1 HQ171414.1 - Genbank # Genbank # Genbank # - Curimatidae Curimatella alburna pending pending pending Genbank # Genbank # Genbank # Genbank # Curimatidae Cyphocharax gilberti pending pending pending pending Genbank # Genbank # Genbank # Genbank # Curimatidae Cyphocharax magdalenae pending pending pending pending Genbank # Genbank # Genbank # Genbank # Curimatidae Cyphocharax spilotus pending pending pending pending Genbank # Genbank # Genbank # Genbank # Curimatidae Potamorhina altamazonica pending pending pending pending Genbank # Genbank # - - Curimatidae Pseudocurimata boulengeri pending pending Genbank # Genbank # Genbank # Genbank # Curimatidae Steindachnerina fasciata pending pending pending pending Genbank # Genbank # Genbank # Genbank # Curimatidae Steindachnerina guentheri pending pending pending pending Genbank # Genbank # Genbank # Genbank # Curimatidae Steindachnerina insculpta pending pending pending pending Genbank # Genbank # Genbank # Genbank # Curimatidae Steindachnerina hypostoma pending pending pending pending Cynodontidae Roestes ogilviei HQ289503.1 HQ289118.1 HQ171410.1 HQ289697.1 Erythrinidae Erythrinus erythrinus - HQ289049.1 HQ171340.1 HQ289629.1 Gasteropelecidae Gasteropeleceus maculatus GQ339514.1 - GQ368214.1 GQ339534.1 Gasteropelecidae Thoracocharax stellatus HQ289477.1 HQ289092.1 HQ171384.1 HQ289671.1 Hemiodontidae Anodus orinocensis HQ289352.1 HQ288964.1 HQ171254.1 HQ289545.1

179

Hemiodontidae Argonectes longiceps HQ289347.1 HQ288959.1 HQ171249.1 HQ289540.1 Hemiodontidae Bivibranchia velox HQ289446.1 - HQ171351.1 HQ289640.1 Hemiodontidae Hemiodus immaculatus HQ289344.1 HQ288956.1 HQ171246.1 HQ289537.1 Hemiodontidae Hemiodus quadrimaculatus AY804084.1 - AY788027.1 AY791405.1 Hepsetidae Hepsetus odoe HQ289480.1 HQ289095.1 HQ171387.1 HQ289674.1 Iguanodectidae Iguanodectes spiturus HQ289412.1 - HQ171316.1 HQ289605.1 Parodontidae Apareiodon affinis HQ289424.1 HQ289037.1 HQ171328.1 HQ289617.1 Parodontidae Parodon AY804110.1 - AY788065.1 AY791427.1 Parodontidae Parodon nasus HQ289521.1 HQ289137.1 HQ171430.1 HQ289714.1 Prochildontidae Ichthyoelephas longirostris KX086992.1 KX086870.1 KX087044.1 KX086809.1 Prochildontidae Prochilodus argenteus KX087006.1 KX086867.1 KX087086.1 KX086842.1 Prochildontidae Prochilodus costatus KX087012.1 KX086869.1 KX087080.1 KX086822.1 Prochildontidae Prochilodus lineatus KX087007.1 KX086865.1 KX087081.1 KX086819.1 Prochildontidae Semaprochilodus kneri KX087037.1 KX086909.1 KX087060.1 KX086847.1 Prochildontidae Semaprochilodus laticeps KX087030.1 KX086906.1 KX087068.1 KX086860.1 Serraalmidae Catoprion mento HQ289485.1 HQ289100.1 HQ171392.1 HQ289679.1 Serraalmidae Colossoma macropoma HQ289438.1 HQ289052.1 HQ171343.1 HQ289632.1 Serraalmidae Myleus torquatus HQ289350.1 HQ288962.1 HQ171252.1 HQ289543.1 Serraalmidae Pygocentrus nattereri AY804119.1 EU001902 FJ944761.1 AY791436.1 Serraalmidae Serrasalmus humeralis HQ289382.1 HQ288995.1 HQ171285.1 HQ289574.1 Serraalmidae Serrasalmus rhombeus HQ289378.1 HQ288991.1 HQ171281.1 HQ289570.1 Triporthidae Agoniates anchovia HQ289472.1 HQ289086.1 HQ171378.1 HQ289665.1 Triporthidae Clupeacharax anchovoides HQ289433.1 HQ289046.1 HQ171337.1 HQ289626.1 Triporthidae Engraulisoma taeniatum HQ289396.1 HQ289010.1 - HQ289589.1 Triporthidae Lignobrycon myersi HQ289495.1 HQ289110.1 HQ171402.1 HQ289689.1 Triporthidae Triportheus nematurus HQ289476.1 HQ289091.1 HQ171383.1 HQ289670.1

180

Table 7.2. Trophic ecology classification and literature reference, including multiple trophic categories for omnivorous taxa when available for each species used in phylogenetic comparative analyses.

Trophic Trophic Classification in Uncertainty Family Species Tier Level Citation Classification Analysis (Menezes 1969; de Mérona et al. 2001; Menezes 2003; de Mérona Acestorynichidae Acestrorhynchus falcatus Piscivore - I and Rankin-de-Mérona 2004) (Menezes 1969; de Mérona et al. 2001; Menezes 2003; de Mérona Acestorynichidae Acestrorhynchus pantaneiro Piscivore - I and Rankin-de-Mérona 2004) Alestidae Alestes baremoze Omnivore Herbivore + Invertivore + Insectivore II (Holden 1970; Lauzanne 1973; Reynolds 1973; Paugy 1978) Alestidae Alestes sp. Omnivore Herbivore + Invertivore + Insectivore II (Holden 1970; Lauzanne 1973; Reynolds 1973; Paugy 1978) Alestidae Alestopetersius hilgendorfi Insectivore - III (Munene and Stiassny 2012) Alestidae Brycinus longipinnis Omnivore - II (Dietoa et al. 2007) Alestidae Brycinus nurse Omnivore - II (Bailey 1994) Alestidae Micralestes acutidens Invertivore Invertivore + Insectivore II (Robinson and Robinson 1969; Lek and Lek 1977; Skelton 1993) Alestidae Micralestes occidentalis Invertivore Invertivore + Insectivore II (Robinson and Robinson 1969; Lek and Lek 1977; Skelton 1993) Alestidae Rhabdalestes maunensis Invertivore - I (Winemiller 1991; Skelton 1993) Anostomidae Abramites hypselonotus Omnivore - II (Taphorn 1992) Anostomidae Anostomoides laticeps Omnivore - II (Santos et al. 2004) Anostomidae Anostomus ternetzi Omnivore - I (Knöeppel 1972; Santos and Rosa 1998) Anostomidae Laemolyta proxima Omnivore - III (de Braga 1990; de Mérona et al. 2001) Anostomidae Leporinus agassizi Omnivore - II (Goulding et al. 1988) Anostomidae Leporinus brunneus Omnivore - II (Goulding et al. 1988)

(Knöeppel 1972; dos Santos 1982; Goulding et al. 1988; Anostomidae Leporinus fasciatus Omnivore - I Planquette et al. 1996) Anostomidae Leporinus nigrotaeniatus Omnivore Omnivore + Invertivore II (Knöeppel 1972; Goulding et al. 1988) Anostomidae Petulanos spiloclistron Herbivore - III (Knöeppel 1972) (Knöeppel 1972; Goulding et al. 1988; Taphorn 1992; Sidlauskas Anostomidae Pseudanos gracilis Omnivore - I and dos Santos 2005)

181

Anostomidae Pseudanos winterbottomi Omnivore - II (Sidlauskas and dos Santos 2005) Anostomidae Rhytiodus lauzannei Herbivore - I (Pouilly et al. 2003; de Mérona and Rankin-de-Mérona 2004) Anostomidae Sartor respectus Invertivore - I (Knöeppel 1972; Géry 1977) (Saul 1975; Goulding 1980; Planquette et al. 1996; Pouilly et al. Anostomidae Schizodon fasciatus Herbivore - I 2003; de Mérona and Rankin-de-Mérona 2004) Anostomidae Schizodon knerii Herbivore - I (Hahn et al. 1998; Sidlauskas et al. 2006) Characidae Aphyocharax pusillus Invertivore - II (Jacobo 1995) Characidae Astyanacinus moorii Omnivore - III (Bertaco et al. 2010) Characidae Bramocharax baileyi Piscivore - II (Bussing 1998) (Menezes 1969; Saul 1975; Gottsberger 1978; Goulding 1980; de Characidae Brycon amazonicus Omnivore - I Melo et al. 2004) (Taphorn 1992; Galvis et al. 1997; Román-Valencia and Taphorn Characidae Bryconamericus emperador Omnivore Omnivore + Invertivore II 2014) Characidae Carlana eigenmanni Omnivore Herbivore + Invertivore II (Bussing 1998) Characidae Chalceus erythrurus Omnivore Herbivore + Invertivore II (de Mérona et al. 2001; de Mérona and Rankin-de-Mérona 2004) Characidae Compsura heterura Invertivore - I (Silva 1993; Mérigoux and Ponton 1998) Characidae Creagrutus affinis Invertivore - I (Saul 1975; Taphorn 1992; Pouilly et al. 2006; Pereira et al. 2007) Characidae Deuterodon iguape Omnivore - I (Sabino and Castro 1989; Esteves and Lobón-Cerviá 2001) Characidae Diapoma terofali Omnivore - I (Neiff et al. 2009; Gelós et al. 2010) Characidae Glandulocauda melanopleura Omnivore - I (Menezes and Weitzman 2009) Characidae Gymnocorymbus ternetzi Invertivore - I (Saul 1975; Mills and Vevers 1989; Taphorn 1992) (Luz-Agostinho et al. 2006; Bertaco and Carvalho 2010; Zanata Characidae Hasemania crenuchoides Omnivore Omnivore + Insectivore II and Serra 2010) (Taphorn 1992; Pouilly et al. 2006; Román-Valencia and Taphorn Characidae Hemibrycon dariensis Omnivore Omnivore + Invertivore+Insectivore II 2014) Characidae Hollandichthys multifasciatus Insectivore - I (Sabino and Castro 1989; Esteves and Lobón-Cerviá 2001) Characidae Hyphessobrycon megalopterus Invertivore - II (Mills and Vevers 1989) Characidae Jupiaba anteroides Omnivore Herbivore + Insectivore II (de Melo et al. 2004; Pereira et al. 2007) (Goulding et al. 1988; Horeau et al. 1998; Ferreira and Lima 2006; Characidae Knodus meridae Omnivore Herbivore + Insectivore II Pereira et al. 2007) Characidae Markiana nigripinnis Omnivore - I (Winemiller 1991; Taphorn 1992)

182

(Costa 1987; Sabino and Castro 1989; Esteves and Lobón-Cerviá Characidae Mimagoniates microlepis Insectivore - I 2001) (Horeau et al. 1998; Casatti 2002; de Melo et al. 2004; Pereira et Characidae Moenkhausia xinguensis Omnivore - I al. 2007) (Saul 1975; Winemiller 1991; de Melo et al. 2004; Bührnheim and Characidae Odontostilbe splendida Detritivore - I Malabarba 2006) Characidae Paracheirodon axelrodi Omnivore Detrivore + Omnivore + Invertivore II (Goulding et al. 1988) Omnivore + Invertivore + Insectivore + Characidae Paragoniates alburnus Omnivore II (Taphorn 1992; Lima 2003) Piscivore Characidae Phenacogaster calverti Omnivore - I (Taphorn 1992; Horeau et al. 1998; de Melo et al. 2004) Characidae Piabarchus analis Insectivore - II (Corrêa et al. 2009) Characidae Piabina argentea Omnivore - I (Luz-Agostinho et al. 2006; Ferreira 2007) Characidae Poptella paraguayensis Invertivore - II (Wantzen et al. 2002) Characidae Prionobrama paraguayensis Omnivore - II (Jacobo 1995) Characidae Rachoviscus crassiceps Omnivore - II (Abilhoa et al. 2007) Characidae Rhinobrycon negrensis Herbivore - II (Goulding et al. 1988) Characidae Roeboexodon guyanensis Piscivore - I (Knöeppel 1972; Planquette et al. 1996) Characidae Salminus hilarii Piscivore - I (Flecker 1992; Taphorn 1992; Luz-Agostinho et al. 2006) Characidae Spintherobolus broccae Invertivore - II (Costa 1987) Characidae Stygichthys typhlops Omnivore - II (Moreira et al. 2010) (Winemiller 1991; Taphorn 1992; de Melo et al. 2004; Pereira et Characidae Tetragonopterus argenteus Invertivore - I al. 2007) Characidae Thayeria obliqua Omnivore - I (Goulding et al. 1988; Mills and Vevers 1989) (Goulding et al. 1988; de Mérona et al. 2001; de Melo et al. 2004; Chilodontidae Caenotropus labyrinthicus Detritivore - I Pereira et al. 2007) (Goulding et al. 1988; de Mérona et al. 2001; de Melo et al. 2004; Chilodontidae Caenotropus maculosus Detritivore - I Pereira et al. 2007) Chilodontidae Chilodus punctatus Detritivore - I (Roberts 1972) Citharinidae Citharinus sp Detritivore - I (Daget 1962; Blache 1964; Arawomo 1976; Lauzanne 1976) Crenuchidae Characidium pterostictum Insectivore - I (Buckup 1993b; Aranha et al. 2000) Crenuchidae Crenuchus spilurus Invertivore - III (Goulding et al. 1988; Silva 1993; Planquette et al. 1996)

183

Crenuchidae Melanocharacidium sp Insectivore - II (Mérigoux and Ponton 1998) Crenuchidae Poecilocharax weitzmani Insectivore - II (Kemenes and Forsberg 2014) Crenuchidae Characidium laterale Insectivore - I (Saul 1975; Buckup 1993b) Curimatidae Curimatella alburna Detritivore - I (Pouilly et al. 2003; Pouilly et al. 2004; Pouilly et al. 2013) (Hahn et al. 1998; Horeau et al. 1998; de Mérona et al. 2001; Curimatidae Cyphocharax gilberti Detritivore - I Peretti and de Andrian 2004; Cala 2005) (Hahn et al. 1998; Horeau et al. 1998; de Mérona et al. 2001; Curimatidae Cyphocharax magdalenae Detritivore - I Peretti and de Andrian 2004; Cala 2005) (Hahn et al. 1998; Horeau et al. 1998; de Mérona et al. 2001; Curimatidae Cyphocharax santacatarinae Detritivore - I Peretti and de Andrian 2004; Cala 2005) Curimatidae Potamorhina altamazonica Detritivore - I (Forsberg et al. 1993; Pouilly et al. 2003; Pouilly et al. 2004) Curimatidae Steindachnerina fasciata Detritivore - I (Hahn et al. 1998; Vaz et al. 1999; Peretti and de Andrian 2004) Curimatidae Steindachnerina guentheri Detritivore - I (Hahn et al. 1998; Vaz et al. 1999; Peretti and de Andrian 2004) Curimatidae Steindachnerina hypostoma Detritivore - I (Hahn et al. 1998; Vaz et al. 1999; Peretti and de Andrian 2004) Curimatidae Steindachnerina insculpta Detritivore - I (Hahn et al. 1998; Vaz et al. 1999; Peretti and de Andrian 2004) (Sandon and al Tayib 1953; Verbeke 1959; Blache 1964; Distichodontidae Distichodus decemmaculatus Herbivore Herbivore + Detritivore II Arawomo 1982; Skelton 1993) Distichodontidae Distichodus fasciolatus Herbivore - II (Riehl and Baensch 1996) Distichodontidae Hemigrammocharax multifasciatus Omnivore - I (Winemiller 1991; Skelton 1993) Distichodontidae Ichthyborus sp Piscivore - I (Daget 1967; Holden and Reed 1972; Roberts 1972; Géry 1977)

Distichodontidae Neolebias trilineatus Invertivore - I (Mills and Vevers 1989) (Saul 1975; Winemiller 1991; Taphorn 1992; Planquette et al. Erythrinidae Erythrinus erythrinus Piscivore - I 1996) (Mills and Vevers 1989; Planquette et al. 1996; Mérigoux and Gasteropelecidae Gasteropelecus maculatus Insectivore - I Ponton 1998) Gasteropelecidae Thoracocharax stellatus Insectivore - I (Taphorn 1992; de Melo et al. 2004) Hemiodontidae Anodus orinocensis Omnivore Omnivore + invertivore II (Hoeinghaus et al. 2003; da Silva et al. 2008; Noveras et al. 2012) Hemiodontidae Argonectes robertsi Omnivore - I (Goulding et al. 1988; da Silva et al. 2008) Hemiodontidae Bivibranchia velox Omnivore Omnivore + Detritivore II (Géry 1977; Planquette et al. 1996; da Silva et al. 2008) Hemiodontidae Hemiodus gracilis Omnivore - I (Goulding et al. 1988; da Silva et al. 2008)

184

(Goulding et al. 1988; de Mérona and Rankin-de-Mérona 2004; da Hemiodontidae Hemiodus immaculatus Omnivore - I Silva et al. 2008) (Adebisi 1981; Winemiller 1991; Skelton 1993; Hugueny and Hepsetidae Hepsetus odoe Piscivore - I Pouilly 1999) Iguanodectidae Iguanodectes geisleri Omnivore - I (Goulding et al. 1988; Silva 1993) Parodontidae Apareiodon affinis Detritivore - I (Pavanelli 2003; Pereira et al. 2007) Parodontidae Parodon nasus Detritivore - II (Pavanelli 1999) Parodontidae Parodon sp Detritivore - II (Knöeppel 1972; Casatti and Castro 1998; Horeau et al. 1998) Prochilodontidae Prochilodus argenteus Detritivore - I (Hahn et al. 1998; Hahn et al. 2004; Peretti and de Andrian 2004) Prochilodontidae Prochilodus costatus Detritivore - I (Hahn et al. 1998; Hahn et al. 2004; Peretti and de Andrian 2004) Prochilodontidae Prochilodus lineatus Detritivore - I (Hahn et al. 1998; Hahn et al. 2004; Peretti and de Andrian 2004) Prochilodontidae Semaprochilodus kneri Detritivore - II (Taphorn 1992) Prochilodontidae Semaprochilodus laticeps Detritivore - II (Taphorn 1992) (Géry 1977; Vieira and Géry 1979; Sazima 1986; Goulding et al. Serrasalmidae Catoprion mento Piscivore Piscivore + Insectivore II 1988; Nico and Taphorn 1988; Taphorn 1992) (Goulding 1980; Taphorn 1992; de Mérona and Rankin-de-Mérona Serrasalmidae Colossoma macropomum Omnivore - I 2004) Serrasalmidae Myloplus rubripinnis Herbivore - III (Planquette et al. 1996; Röpke et al. 2014) Serrasalmidae Pygocentrus nattereri Omnivore Piscivore + Insectivore II (Mills and Vevers 1989; de Mérona and Rankin-de-Mérona 2004) Serrasalmidae Serrasalmus maculatus Omnivore Piscivore + Insectivore II (Sazima and Zamprogno 1985) Serrasalmidae Serrasalmus spilopleura Omnivore Piscivore + Insectivore II (Planquette et al. 1996) Triporthidae Agoniates anchovia Piscivore Piscivore + Insectivore II (Goulding et al. 1988; de Mérona et al. 2001) Triporthidae Engraulisoma taeniatum Omnivore - II (Taphorn 1992) Triporthidae Lignobrycon myersi Invertivore - II (Vari et al. 1995) (Lowe-McConnell 1975; Goulding 1980; de Mérona et al. 2001; Triporthidae Triportheus nematurus Omnivore - I de Mérona and Rankin-de-Mérona 2004)

185

Table 7.3. Priors and parameters used in BEAST 2.0

Prior Distribution Initial Mean SD Offset Upper Lower treeModel.rootHeight Gamma 80 70 1 0 - - tmrca(Serrasalmidae) LogNormal - 6 1 61 - - tmrca(Leporinus) LogNormal - 15 0.9 15 - - tmrca(Cyphocharax) LogNormal - 15 1 23 - - tmrca(Citharinoidea) LogNormal - 80 0.5 45 107 52 tmrca(Distichodus) LogNormal - 17 0.5 7 25 9 birthDeath.meanGrowthRate Uniform 0.8 - - 1 0 birthDeath.relativeDeathRate Uniform 0.0006 - - 100 0 16Sucld.mean Exponential 0.005 0.02 - 0 - - 16Sucld.stdev Exponential 0.333 0.5 - 0 - - CodingMitochondrial.ucld.mean Exponential 0.0003 0.02 - 0 - - CodingMitochondrial.ucld.stdev Exponential 0.333 0.5 - 0 - - Nuclear.ucld.mean Exponential 0.0004 0.01 - 0 - - Nuclear.ucld.stdev Exponential 0.333 0.5 - 0 - -

186

Table 7.4. Gene partitions and their models as selected by PartitionFinder 2.0.

Best-fit Gene and position model 16S GTR+G+I Cytb 1st position GTR+G+I Cytb 2nd position GTR+G Cytb 3rd position TrN+I Myh6 1st position GTR+G+I Myh6 2nd position GTR+I Myh6 3rd position HKY+G Rag2 1st position TVM+G Rag2 2nd position TVM+G Rag2 3rd position TVM+G

187

Table 7.5. Results of the multivariate model-fitting analyses for body shape constrained to the Oliveria et al. 2011 topology. For each model, the number of parameters (P), the

Akaike information criterion (AIC), the small sample corrected AIC (AICc), and the relative fit (ΔAICc) and support (AICc weight) are shown. The best model has the lowest ΔAICc.

Model P AIC AICc Δ AICc AICc weight OU surface 62 -2056 -2039 0 1 OU surface + piscivores 66 -2018 -2000 39 0 OU family 86 -2029 -1985 53.6 0 OU family +piscivores 78 -1927 -1894 144.4 0 EB 15 -1881 -1881 157.9 0 OU trophic 38 -1872 -1867 172.3 0 BM 12 -1866 -1866 172.9 0 BMM 84 -1864 -1844 195 0 OU continent 22 -1838 -1836 202.9 0

188

Table 7.6. Average results from the convevol analysis when simulated 1000 times across the uncertainty index. All values with statistically significant P-values (P <

0.05) in 100% of simulations indicated with an asterisk.

Trophic Ecology C1 C2 C3 C4 Piscivores 0.416* 0.064* 0.0080* 0.014* Detritivores 0.200 0.0237 0.003 0.004 Herbivores 0.308 0.0707 0.009 0.013 Insectivores 0.088 0.0191 0.002 0.004 Invertivores 0.08 0.0145 0.0018 0.003 Omnivores 0.206 0.0336 0.004 0.006

189

Table 7.7. Supporting Information Table S7. Results from the convevol analysis with the phylogenetic topology constrained to Oliveira et al. 2011. All values with statistically significant P-values (P < 0.05) indicated with an asterisk.

Trophic Ecology C1 C2 C3 C4 Piscivores 0.436* 0.067* 0.008* 0.014* Detritivores 0.202 0.0237 0.002 0.004 Herbivores 0.307 0.0705 0.008 0.013 Insectivores 0.099 0.0199 0.002 0.003 Invertivores 0.159 0.0238 0.002 0.004 Omnivores 0.258 0.0438 0.004 0.007

190

Appendix 4: Materials examined

Acestrorhynchidae:

Acestrorhynchus minimus: USNM 229092, 1 (48.5 mm SL), South America, Brazil,

Amazonas, Lago Janauari, Lago Canta Galos, 2 February 1978.

USNM 304891, 1 (37.5 mm SL), South America, Brazil, Amazonas, Rio

Negro/Urubaxi Pedral, February 1987.

USNM 305130, 1 (36.2 mm SL), South America, Brazil, Amazonas; Near Paraiso,

Across From Lago Do Castanho, Janauaca, 26 October 1977.

USNM 311177, 2 (40.5 – 50.0 mm SL), South America, Brazil, Amazonas, Rio Urubu

At Crossing With Estrada Manaus-Itacoatiara, About km 197, Am. Brasil, 17 October

1977.

Acestrorhynchus pantaneiro: USNM 326353, 4 (100.0 – 160.5 mm SL), South

America, Brazil, Matto Grosso, Rio Paraguai em Caceres e arredores, 11 August 1991 to 12 August 1991.

USNM 326469, 2 (148.6 – 154.4 mm SL), South America, Brazil, Mato Grosso, Arroio

Cruzando a Estrada Tangara Da Serra/Barra Do Bugres, Proximo a Novo Olimpia (Afl.

Rio Paraguai), Nova Olimpia, 11 August 1991.

Alestidae:

Alestes baremoze: USNM 310059, 1 (136.7 mm SL), Africa, Togo, Macouga, 4 May

1968 to 5 May 1968.

USNM 61311, 2 (133.7 – 164.4 mm SL), Africa, Egypt, Near Luxor.

Alestes dentex: USNM 229675, 4 (81.1 – 90.3 mm SL), Africa, Nigeria, Goronyo Dam

Region (Pre-Impoundment), N 13° 28' 12" E 5° 40' 11”, 16 February 1981.

191

Alestes longispinnis: USNM 298659, 10 (33.5 – 57.3 mm SL), Africa, Ghana, Aimaso

R., Near Adomi Bridge, 19 November 1970.

Alestopetersius lepoldienus: USNM 365946, 3 (17.4 – 17.6 mm SL), Africa, Congo,

River Ngula, Terr. De Benalia (Congo Belge), 6 May 1958.

Arnoldichthys spilipterus: USNM 367305, 9 (56.7 – 74.4 mm SL), Africa, Nigeria,

Uegeli, 21 November 1993.

Bathyaethiops caudomaculatus: USNM 365947, 6 (23.0 – 30.6 mm SL), Africa,

Congo, Stanley-Pool, River Afflt. Du Chenal, Eaux Noires (Congo Belge), 1 February

1957 to 28 February 1957.

Brycinus imberi: USNM 176352, 1 (99.1 mm SL), Africa, Democratic Republic of the

Congo, Belgian Congo: Stanleyville, Fished By Women With Hoopnet From Cataract

At Wagenia Fishery, 19 April 1955.

USNM 309581, 3 (55.3 – 66.1 mm SL), Africa, Zambia, Eastern Province, Luangwa

River Oxbow At Kapani S. Luangwa Park, 25 October 1989.

Brycinus longipinnis: USNM 304036, 8 (46.3 – 66.8 mm SL), Africa, Cameroon,

Southwest Province, Manyu, Cross System: Collecting Points On Southern Munaya R.

Draining Northern Korup; 'Dark Lagoon' Above Stream Junction With Munaya, N 5°

50' 25” E 9° 3' 0", 21 February 1988.

Brycinus macrolepidotus: USNM 397932, 5 (51.6 – 82.1 mm SL), Africa, Ghana, Dayi

River, at Gbefi Ghana, 20 November 1970.

Brycinus nurse: USNM 303944, 8 (48.4 – 69.2 mm SL), Africa, Cameroon, Southwest

Province, Manyu, Collecting points on Main Cross river downstream of Mamfe; Main

Cross R. ca 23 km below Mamfe, N 5° 51' 24” E 9° 11' 49”, 17 February 1988.

192

Bryconaethiops boulgengeri: USNM 310349, 1 (63.0 mm SL), Africa, Democratic

Republic of the Congo, Kinshasa: Boende, Tchuapa, October 1970.

Bryconaethiops cameroon: USNM 304256, 9 (30.7 – 56.2 mm SL), Africa, Cameroon,

Southwest Province, Manyu, Collecting Points On Main Cross River Downstream of

Mamfe; Mam R. Junction With Cross, N 5° 50' 30” E 9° 14' 49”, 16 February 1988.

Bryconalestes longipinnis: USNM 298659, 10, Africa, Ghana, Aimaso R., Near Adomi

Bridge, 19 November 1970.

Dubiosialestes tumbensis: USNM 369306, 1 (37.4 mm SL), Africa, Democratic

Republic of the Congo, Stanley Pool, Congo Basin, S 4° 15' 0" E 15° 25' 11”, 1957.

USNM 365953, 1 (24.7 mm SL), Africa, Congo, Stanley-Pool, Cole. Ile m'Bamu

Jusqu'a N'Zete-Moko, 25 October 1957.

USNM 365954, 1 (32.1 mm SL), Africa, Congo, Stanley-Pool, Beach De Kingabwa, 7

November 1957.

Hydrocynus goliath: CUMV 96086, 1 (286.35 mm SL), Africa, Democratic Republic of the Congo, Orientale, Tshopo, Lindi River at rapids at Bawombi II, N 0° 39' 48” E

25° 8' 42”, 18 August 2010.

Hydrocynus vittatus: USNM 229957, 2 (185.9 – 202.1 mm SL), Africa, Nigeria,

Sokoto, Wamako, Main Sokoto River, N 13° 1' 47” E 5° 9' 0", 22 October 1980.

Ladigesia roloffi: USNM 365951, 2 (10.0 - 10.1 mm SL), Africa, Sierra Leone,

Kasewe-Forest, 1 January 1964 – 31 December 1964.

Lepidarchus adonis: USNM 267290, 10 (10.0 – 10.9 mm SL), Africa, Ghana, 30 June

1970.

193

Micralestes acutidens: USNM 369305, 10 (24.4 – 47.8 mm SL), Africa, Democratic

Republic of the Congo, Stanley Pool, Congo Basin, S 4° 15' 0" E 15° 25' 11”, 1957.

Micralestes occidentalis: USNM 397933, 10 (45.9 – 62.6 mm SL), Africa, Togo,

Gande, Togo-Sara River Above and Below Pont-Barrage approximately 5 kilometers

Due West of Bafilo, 10 March 1969.

Rhabdalestes serptentrionalis: USNM 310844, 8 (29.6 – 39.6 mm SL), Africa, Ghana,

Dayi R. At Gbefi, 3 October 1970.

Anosotmidae:

Abramites hypselonotus: USNM 303118, 8 (63.2 – 112.9 mm SL), South America,

Bolivia, Beni, Rio Curiraba at 10 km NE El Porvenir Biol. Sta., at 40 Air km E San

Borja, S 14° 55' 11” W 66° 16' 48", 31 August 1987.

Anostomoides laticeps: USNM 233209, 1, South America, Venezuela, Delta Amacuro,

Rio Arature, 4-5 Miles Upstream From Confluence With Rio Orinoco, N 8° 32' 23”,

W 60° 52' 5”, 20 November 1979.

Anostomus anostomus: USNM 232067, 7 (62.0 – 73.44 mm SL), South America, Peru,

Loreto, Iquitos, 1922.

Anosotmus brevior: USNM 409818, 9 (37.5 – 86.7 mm SL), South America, Suriname,

Upper Paloemeu River, ~500 km downstream of basecamp, Suriname. Collected in the main channel of the river in shallow areas, mostly inner bends of the stream, N 2° 29'

22” W 55° 36' 38”, 15 March 2012.

Anostomus garmani: MCZ 32099, 1, South America, Brazil, Amazonas, Rio Javari

(tributary of Rio Solimoes) at the Peruvian-Brazilian border, S 4° 21' 29” W 70° 2' 30”,

19 October 1865.

194

Anostomus plicatus: FMNH 53393, 1, South America, Guyana, Crab Falls, 1908.

FMNH 53394, 1, South America, Guyana, Bartica, 1908.

FMNH 53395, 1, South America, Guyana, Amatuk, 1908.

FMNH 53396, 1, South America, Guyana, Tumatumari, 1908.

Anostomus taeniatus: FMNH 105777, 2, South America, Venezuela, Amazonas: Pozo

Azul ca. 20 km S. of Puerto Ayacucho, 16 February 1995.

Anostomus ternetzi: USNM 66094, 1 (70.3 mm SL), South America, Guyana,

Tumatumari, September 1908 to December 1908.

USNM 233230, 8 (27.3 – 44.9 mm SL), South America, Venezuela, Delta Amacuro,

Rio Orinoco, Small Cano Near Mouth of Cano Socoroco, 111 Naut. mi. Upstream From

Sea Buoy, N 8° 34' 48" W 61° 42' 0", 20 February 1978.

Gnathodolus bidens: ANSP 161699, 1 (107.3 mm SL), South America, Venezuela,

Amazonas, Rio Iguapo ca. 1 hr. above its mouth (tributary of Rio Orinoco), N 3° 9' 0''

W 65° 28' 0'', 13 March 1987.

ANSP 159389, 10 (60.4 – 72.5 mm SL), South America, Venezuela, Bolivar, Cano

(possibly Cano Curimo) feeding Rio Caura near confluence of Rio Caura - Rio

Orinoco, N 7° 37' 48'', W 64° 50' 42'', 22 November 1985.

Hypomasticus megalepis: USNM 225381, 2 (65.2 – 91.0 mm SL), South America,

Suriname, Nickerie District, Mataway Creek approximately 8 km from its intersection with Corantijn River, N 4° 46' 48" W 57° 45' 0", 11 September 1980.

USNM 225397, 3 (80. – 87.7 mm SL), South America, Suriname, Nickerie District, creek opposite Guyanese logging camp, 2 1/4 hours south of Matapi, approximately 2 km downstream of Cow Falls, N 4° 58' 48" W 57° 37' 48", 11 September 1980.

195

Hypomaticus julli: INPA 9508, South America, Brazil, Pará, Oriximiná, S 1° 45’56”

W 55° 51’58”, 26 November 1987.

INPA 9508, 6 (97.8 – 98.7 mm SL), South America, Brazil, Pará, Oriximiná, S 1°

45’56”, W 55° 51’58”, 26 November 1987.

Laemolyta fernandezi: INPA 12234, 6, South America, Brazil, Pará, Tucuruí, 4° 45’58’

W 49° 40’21”, 30 June 1980.

Laemolyta garmani: USNM 383198, 1 (70.6 mm SL), South America, Peru, Loreto,

Rio Nanay Approximately 20 km Upstrea m of Mouth; Main Channel, Side Channels and Side Pools, S 3° 51' 0" W 73° 15' 0", 19 August 1986.

Laemolyta orinocensis: FMNH 109849, 1, South America, Venezuela, Bolivar, sandy beach at El Playon just below Salto Para, N 6° 19' 31” W 64° 31' 36”, 2 December

2000.

USNM 258159, 2, South America, Venezuela, Guarico State, Fundo Masaguaral, Rio

Caracol Where Crossed By Bridge On Ranch, N 8° 34' 12" W 67° 30' 0", 19 January

1983.

Laemolyta proxima: USNM 377366, 2 (149.1 – 170.5 mm SL), South America,

Guyana, Kokente Pond, 23 November 2001.

USNM 377366, 6 (73.3 – 80.5 mm SL), South America, Guyana, Kokente Pond, 23

November 2001.

FMNH 62794, 1, South America, Brazil: Lake Hyanuary, 27 October 1865.

Laemolyta taeniata: USNM 310180, 1 (162.5 mm SL), South America, Brazil, Rio

Madeira: Lago Das Pupunhas ca 7 km. E. of Humaita On Transmazonica, 24 August

1976.

196

USNM 302049, 2 (77.0 – 86.9 mm SL), South America, Brazil, Rio Negro, Parana Do

Marauia, January 1987.

USNM 280708, 1 (81.6 mm SL), South America, Peru, Loreto, Canos Entering Rio

Nanay, Northeast of Iquitos, S 3° 49' 11” W 73° 10' 48", 18 August 1986.

USNM 317579, 1 (91.0 mm SL), South America, Brazil, Amazonas, Lago Janauari,

Primeira Olaria, 3 August 1977.

Leporellus pictus: FMNH 55175, 1, South America, Brazil, Minas Gerais: Pirapora, 15

December 1907.

FMNH 111338, 1, South America, Peru, Loreto, Rio Yanayacu ca. 2.3km above mouth in Rio Maranon, S 4° 40' 0” W 73° 40' 0”, 30 August 1988.

Leporellus retropinnis: FMNH 55174, 1, South America, Brazil, Sao Paulo, Piracicaba,

9 September 1908.

Leporellus vittatus: MCZ 19389, 2, South America, Brazil, Goiás/Tocantins, Goyaz

[exact locality unknown; probably somewhere in the Tocantins-Araguaia system],

1867.

Leporinus agassizi: USNM 302058, 1 (80.6 mm SL), South America, Brazil,

Amazonas, Rio Negro, Parana Do Marauia, January 1987.

USNM 377410, 1 (86.5 mm SL), South America, Guyana, Burst Mouth Pond, 23

November 2001.

USNM 402629, 1 (93.5 mm SL), South America, Guyana, Cuyuni-Mazaruni, mud flats, sand bars, rocks & wooded shores, including isolated pools in nearlly dry channel crossing island in middle of Cuyuni River about 15km upstream from Waikuni Mtns in vicinity of mouth of Toropaur River, N 6° 41' 30” W 59° 34' 37”, 3 February 2011.

197

USNM 280656, 3 (40.2 – 52.2 mm SL), South America, Peru, Loreto, Rio Nanay

Approximately 20 km Upstream of Mouth; Main Channel, Side Channels and Side

Pools, S 3° 51' 0" W 73° 15' 0", 19 August 1986.

Leporinus alternus: FMNH 53361, 1, South America, Guyana, Tukeit, 1908.

Leporinus amae: USNM 285630, South America, Brazil, Rio Grande do Sul, Rio

Piratini, Fazenda Dos Hinz, Distrito De Coimbra Santo Angelo Rs. (Afluente Do Rio

Uruguai), 19 December 1985.

Leporinus arcus: FMNH 50165, 4, South America, Guyana, New River drainage, head of Itabu Creek, 10 October 1938.

FMNH 53366, 1, South America, Guyana, Tukeit, 1908.

Leporinus bahiensis: FMNH 78791, 6, South America, Brazil, Mogy Guassu, Rio

Mogy Guassu into Rio Grande, into Rio Parana, 25 August 1908.

MCZ 1191, 2, South America, Brazil, Bahia, S 12° 58' 29” W 38° 29' 30”, 1863.

MCZ 20374, 1, South America, Brazil, Bahia, Nazaré, S 13° 0' 29” W 39° 0' 29”, 1863.

Leporinus conirostris: MCZ 20371, 4, South America, Brazil, Rio de Janeiro, Rio

Paraíba do Sul, near city of Rio de Janeiro, S 22° 53' 30” W 43° 17' 30”, 1872.

Leporinus copelandi: MCZ 20410, 4, South America, Brazil, Espírito Santo, Rio Doce, between Linhares and Aimores, S 19° 37' 30" W 39° 49' 30", 1865.

MCZ 20408, 2, South America, Brazil, Espírito Santo, Rio Doce, between Linhares and Aimores, S 19° 37' 30" W 39° 49' 30", 1865.

Leporinus cylindriformis: MZUSP 97496, 10.

Leporinus jatuncochi: INHS 38940, 8, Peru, Loreto, Río Nanay-Río Amazonas drainage, purchased from aquarium dealer in Santa Clara.

198

Leporinus ecuadorensis: FMNH 10280, 1, South America, Ecuador, Vinces.

FMNH 10281, 1, South America, Ecuador, Vinces.

FMNH 10283, 1, South America, Ecuador, Vinces.

FMNH 56611, 2, South America, Ecuador, Guayaquil Market, 1913.

FMNH 56612, 2, South America, Ecuador, Rio Baranca Alta, Naranjito, 1913.

Leporinus falcipinnis: FMNH 104000, 2, South America, Venezuela, Amazonas, backwater of island in Rio Atabapo ca. 1 hr. above San Fernando de Atabapo, 27

January 1991.

Leporinus fasciatus: USNM 319281, 4 (111.7 – 133.3 mm SL), South America,

Bolivia, Santa Cruz, Rio Parapeti At Rr Bridge At San Antonio, at 40 Air km E Camiri,

S 20° 1' 11” W 63° 12' 0", 30 September 1988.

USNM 225991, 6 (124.4 – 165.5 mm SL), South America, Suriname, Nickerie District,

Corantijn River, N 5° 0' 0" W 57° 16' 48", 17 May 1980.

Leporinus friderici: USNM 310755, 1 (247.0 mm SL), South America, Brazil, Rio

Solimoes Near Tefe "Poison Station", March 1974.

USNM 302486, 4 (125.7 – 170.5 mm SL), South America, Brazil, Sao Paulo, Near

Santa Rosa De Viterbo, Barragem De Itaipava, Usina Amalia; Pardo River Drainage,

Rio Pardo, Main Channel of River, S 21° 25' 12" W 47° 19' 47”, 25 October to 14

November 1984.

Leporinus gomesi: INPA 14264, 11 (86.9 – 119.8 mm SL), Brazil, Mato Grosso, Rio

Aripuana. Igarapé da Chapada near airport about 5 km from Humbolt.

Leporinus granti: USNM 225402, 5 (68.8 – 118.6 mm SL), South America, Suriname,

Nickerie District, stream on south bank of Lucie River ca 6 km upstream of ferry

199 crossing on Amotopo-Camp Geologie Rd, N 3° 36' 0" W 57° 37' 11, 18 September

1980.

FMNH 7323, 1, South America, Guyana, Maripicru Creek between Wontyke &

KaraKara.

FMNH 53373, 1, South America, Guyana, Maripicru.

FMNH 53383, 1, South America, Guyana, Maripicru.

MCZ 29929, 1, South America, Guyana, Rupununi, Maripicru (Maripakuru) Creek, branch of Ireng River between Wontyke and Karakara, above Karona Falls, N 4° 16'

30" W 59° 43' 30", 1908.

Leporinus guttatus: ANSP 189166, 1 (104.3 mm SL), South America, Brazil, Pará,

Altamira, Rio Curua (Iriri-Xingu Dr.), upstream from PCH Buriti (small hydroelectric dam), S 8° 46' 28'', W 54° 57' 12'', 21 October 2007.

Leporinus lacustris: USNM 326374, 1 (56.6 mm SL), South America, Brazil, Mato

Grosso, Stream Trib To Rio Paraguai, 30 km From Turn Off From Br 364 Onto Mt 343

To Caceres, S 15° 24' 0" W 57° 25' 12", 11 August 1991.

Leporinus lebaili: ANSP 189043, 3 (49.6 – 95.7 mm SL), South America, Suriname,

Sipalawini, Lawa River (Marowijne Dr.), base camp ca. 8 km south-southwest of

Anapaike/Kawemhakan (airstrip), N 3° 19' 31'' W 54° 3' 48'', 18 to 22 April 2007.

ANSP 189011, 1 (63.8 mm SL), South America, Suriname, Sipalawini, Litanie River at mouth and confluence with Marowini River, just upstream from settlement of Konya

Kondre, N 3° 17' 24'' W 54° 4' 38'', 21 to 23 April 2007.

Leporinus leschenaulti: FMNH 85530, 2, South America, Venezuela, Amazonas, stream at 32 km out of Pto. Ayacucho toward San Mariapo, 12 December 1975.

200

Leporinus maculatus: USNM 225995, 3 (98.6 – 111.4 mm SL), South America,

Suriname, Nickerie District, Stream Entering Corantijn River At Approximately km

385 Slightly North of Tiger Falls, N 4° 0' 0" W 58° 1' 48", 16 September 1980.

Leporinus melanopleura: USNM 300676, South America, Brazil, Bahia, Ribeirao Das

Caveiras, Tributary of the Right Margin of Rio De Una, 8 km SE of Sao Jose, Una, 9

October 1986.

Leporinus moralesi: FMNH 112961, 1, South America, Peru, Loreto, Rio Nanay at

Nanay Beach, west of Iquitos, S 3° 49' 59” W 73° 10' 59”, 17 August 1986.

Leporinus mormyrops: FMNH 122963, 1, South America, Brazil, Minas Gerais, Rio

Mucuri approx. 9 km W of town of Presidente Pena along dirt road on Fazenda Gaviao,

S 17° 40' 59” W 40° 55' 0”, 19 July 1991.

MCZ 20369, 1, South America, Brazil, Rio de Janeiro, Rio Piabanha at Posse, S 22°

23' 30”, W 43° 1' 29”, 1865.

MCZ 97556, 7, South America, Brazil, Minas Gerais, Rio Mucuri appros. 9km W of

Presidente Pena along dirt road on Fezenda Gaviao. Main river channel & rapids, S 17°

41' 30” W 40° 55' 30”, 17-23 July 1991.

Leporinus muyscorum: USNM 310750, 3 (121.9 – 131.2 mm SL), South America,

Colombia, Choco, Rio Salado Near Teresita, 8 February 1968.

FMNH 57717, 2, South America, Colombia, Quibdo, 1913.

Leporinus nattereri: MCZ 20384, 3, South America, Brazil, Amazonas, Rio Solimoes at Tefe and environs, S 3° 24' 29” W 64° 45' 29”, October 1865.

MCZ 19825, 1, South America, Brazil, Pará, Rio Amazonas at Santarem, S

2° 26' 30” W 54° 41' 30”, 26 August 1865.

201

Leporinus cf. niceforoi: MUSM 17663.

Leporinus nigrotaeniatus: USNM 66215, 1 (101.8 mm SL), South America, Guyana,

Crab Falls, September 1908 to December 1908.

USNM 404305, 1 (57.8 mm SL), South America, Guyana, Cuyuni-Mazaruni, rocks and rapids on Cuyuni River, main channel, 4 February 2011.

FMNH 69579, 4 (XX – XX mm SL), South America, Guyana, Lower Potaro River,

Tumatumari, 1908.

Leporinus obtusidens: USNM 365199, 1 (225.6 mm SL), South America, Argentina,

Darsena Norte, Buenos Aires, 3 May 1975.

MCZ 20457, 1, South America, Brazil, Rio Sao Francisco, between Guaicui and

Januaria, S 17° 12' 29” W 44° 50' 30”, 28 August 1865.

MCZ 20459, 3, South America, Brazil, Rio Sao Francisco, between Guaicui and

Januaria, S 17° 12' 29” W 44° 50' 30”, 28 August 1865.

Leporinus octofasciatus: USNM 302518, 1 (112.7 mm SL), South America, Brazil, Sao

Paulo, Near Santa Rosa De Viterbo, Barragem De Itaipava, Usina Amalia; Pardo River

Drainage, Rio Pardo, Main Channel of River; Fish Caught With Cast Nets Just Below

Fish Ladder, S 21° 25' 12" W 47° 19' 47”, 25 October 1984 to 14 November 1984.

Leporinus octomaculatus: MCZ 35022, 4, South America, Brazil, Tocantins/Goiás,

Barra do Rio Sao Domingos, S 13° 25' 30" W 46° 19' 30", 1936.

MCZ 47776, 2, South America, Venezuela, Amazonas, Middle Rio Cunucunuma at

Jacare, 6 April 1950.

Leporinus ortomaculatus: USNM 224796, 1 (56.5 mm SL), South America, Guyana,

Rupununi District, Puara River, Amazon Drainage, 31 January 1953.

202

USNM 233226, 9, South America, Venezuela, Bolivar, Rio Orocopiche, ca. 15 km from mouth in Rio Orinoco, downstream from Route 19 bridge, N 8° 3' 0" W 63° 40'

12", 3 November 1979.

Leporinus pachycheilus: INPA 6706, 3 (87.3 – 90.1 mm SL), Brazil, Pará, Rio

Jamanxim, tributary of Rio Tapajos, Isla Terra Preta.

Leporinus parae: USNM 258025, 4 (100.4 – 176.3 mm SL), South America,

Venezuela, Guarico State, Fundo Masaguaral, Rio Caracol Where Crossed By Bridge

On Ranch, N 8° 34' 12" W 57° 30' 0", 19 January 1983.

Leporinus pearsoni: USNM 332102, 1 (64.8 mm SL). No data.

USNM 164034, 1 (138.7 mm SL), South America, Ecuador, Napo-Pastaza, Chicherota,

Lower Bobonaza River, S 2° 22' 12" W 76° 45' 0", January 1949.

USNM 175860, 3 (58.4 – 106.5 mm SL), South America, Peru, Shansho Cano, 19 July

1934.

Leporinus pellegrinii: INPA 15672, 7 (67.7 – 91.8 mm SL), Brazil, Rio Tocantins.

Leporinus piau: USNM 345745, 1 (78.8 mm SL), South America, Brazil, Minas Gerais,

Riacho Afluente Do Rio Jequitai, Na Br-135, Entre Buenopolis E Engenheiro Dolabela,

20 July 1994.

Leporinus reinhardti: USNM 44953, 1 (129.6 mm SL), South America, Brazil, Lagos

Santa.

Leporinus steindachneri: USNM 55663, 3 (136.2 – 197.4 mm SL), South America,

Brazil, Amazonas, Manaus Region, 1977 to 1979.

Leporinus steyermaki: FMNH 45701, 1, South America, Venezuela, Bolivar:

Chimanta-tepui, Rio Abacapa Camp 1, altitude 1300 meters, 27 March 1953.

203

Leporinus striatus: USNM 302526, 8 (57.8 – 107.7 mm SL), South America, Brazil,

Sao Paulo, Near Santa Rosa De Viterbo, Barragem De Itaipava, Usina Amalia; Pardo

River Drainage, Rio Pardo, Main Channel of River; Fish Caught With Cast Nets Just

Below Fish Ladder, S 21° 25' 12" W 47° 19' 47”, 25 October 1984 to 14 November

1984.

Leporinus subniger: USNM 399976, 3 (131.0 – 136.5 mm SL), South America,

Venezuela, Delta Amacuro, Rio Orinoco, Small Cano At Mouth of Cano Fiscal, 64

Naut. mi. Upstream From Sea Buoy, N 8° 31' 11” W 61° 1' 48", 23 February 1978.

USNM 348685, 3 (101.0 – 119.4 mm SL), South America, Venezuela, Portuguesa,

Guanare-Guanarito Road At Road km 60, N 8° 49' 39" W 69° 20' 41”, 24 February

1998.

Leporinus taeniatus: USNM 345875, 10 (88.5 – 165.2 mm SL), South America, Brazil,

Minas Gerais, Pedro Leopoldo Riacho Afluente Do Rio Das Velhas, Em Igreja

Quebrada, Proximo Da Jaguara, 19 July 1994.

Leporinus tigrinus: INPA 1917, 6 (103.6 – 107.8 mm SL), Brazil, Pará, Rio Tocantins,

Tucurui.

MCZ 20446, 1, South America, Brazil, Goiás/Tocantins, Goyaz (exact locality unknown; probably somewhere in the Tocantins-Araguaia system), 1867.

Leporinus trifasciatus: USNM 305233, 1 (42.1 mm SL), South America, Brazil,

Amazonas, Lago Terra Preta Janauari, 19 January 1978.

USNM 305104, 2 (21.2 – 30.1 mm SL), South America, Brazil, Amazonas, Parana De

Janauaca, Entrada Do Lago Do Castanho, 10 May 1978.

204

Leporinus wolfei: MCZ 95469, 1, South America, Peru, Locality unknown, probably

Peru in Upper Amazon drainage, 1961.

Leporinus yophorus: USNM 390290, 1 (54.6 mm SL), Aquarium specimen.

USNM 94286, 1 (124.5 mm SL).

Petulanos intermedius: INPA 15184, 10, South America, Brazil, Rondônia, Candeias do Jamari, S 8° 48’35” W 63° 41’44”, 9 April 1985.

Petulanos plicatus: USNM 225396, 10 (73.1 – 106.4 mm SL), South America,

Suriname, Nickerie District, Matappi Creek, N 5° 1' 11” W 57° 17' 30”, 17 May 1980.

Petulanos spiloclistron: ANSP 179671, 1 (52.0 mm SL), South America, Guyana,

Rupununi, Pirara River (Ireng-Takutu-Branco Dr.), 3.5 km NNW of Pirara, N 3° 38'

55'' W 59° 41' 20'', 2 November 2002.

Pseudanos gracilis: FMNH 104021, 1 (154.4 mm SL), South America, Venezuela,

Amazonas, Cano Guasuriapana at Guasuriapana; cano & backwater ca. 7 min. from

S.F. Atabapo; tributary of Rio Atabapo, N 4° 0' 0" W 67° 42' 0", 28 January 1991.

CAS SU 16277, 1 (116.6 mm SL), South America, Venezuela, R. Negro, between

Cucui, Brazil and San Carlos, 22 February 1925.

CAS SU 16278, 1 (153.2 mm SL), South America, Venezuela, Amazonas, R. Atabapo into R. Orinoco, San Fernando de Atabapo, N 4° 3' 30” W 67° 42' 44”, 9 April 1925.

CAS SU 16280, 3 (67.5 – 159.6 mm SL), South America, Venezuela, Amazonas, R.

Atabapo, San Fernando de Atabapo, N 4° 2' 54” W 67° 42' 22”, 7 April 1925.

Pseudanos irinae: USNM 376809, 2 (73.0 – 83.4 mm SL), South America, Guyana,

Taraqua Creek, 31 October 2001.

205

Pseudanos trimaculatus: USNM 305437, 3 (60.8 – 78.0 mm SL), South America,

Bolivia, Beni, Ballivia Province, Rio Matos below road crossing, 48 km east San Borja,

S 14° 55' 11” W 66° 16' 48", 28 August 1987.

FMNH 69596, 6, South America, Brazil, Mato Grosso, Maciel, Rio Guapore, 30 July

1909.

Pseudanos winterbottomi: USNM 270328, 2 (112.1 – 125.8 mm SL), South America,

Venezuela, Amazonas, Departamento Ature, Balneiria Pozo Azul, approximately 1 km

To East of Puerto Ayacucho To Solano Road, approximately 30 km N of Puerto

Ayacucho, N 5° 52' 47” W 67° 28' 11”, 10 December 1984.

Rhytodius argenteatus: INHS 66082, 3 (66.1 – 67.3 mm SL), Brazil, Amazonas, Lake

Janauaca, 42 km SW of Manaus.

INHS 66721, 1 (67.3 mm SL), Brazil, Amazonas, Lake Janauaca, 42 km SW of

Manaus.

INHS 67453, 2 (64.1 – 70.1 mm SL), Brazil, Amazonas, Lake Janauaca, 42 km SW of

Manaus.

Rhytodius lauzannei: FMNH 111326, 2, South America, Peru, Loreto: Una cocha of the Rio Yanayacu ca. 6-7km above mouth in Rio Amazonas, S 4° 19' 59” W 73° 15'

0", 29 August 1988.

INHS 39321, 4 (111.7 – 117.1 mm SL), Peru, Loreto, Río Amazonas drainage, Río

Yanashi, 112.3 km E.of Iquitos.

Rhytodius microlepis: USNM 305241, South America, Brazil, Amazonas, Furo Entre

Lago Murumuru E Parana De Janauaca, 25 May 1977.

206

Sartor elongates: INPA 1168, 3 (70.6 – 89.7 mm SL), Brazil, Río Trombetas,

Cachoeira Porteira.

Sartor tucuriense: INPA 1166, 2 (100.0 – 101.0 mm SL), Brazil, Pará, Río Tocantins, city of Tucuruí, ca. 2 km below Tucuruí dam, pools of water with rocky bottoms.

Schizodon borellii: USNM 326314, 5 (125.6 – 237.1 mm SL), South America, Brazil,

Mato Grosso, Rio Paraguai em Caceres e arredores, Caceres, 11 August 1991 to 12

August 1991.

Schizodon dissimilis: MCZ 19381, 2, South America, Brazil, Piauí, Rio Poti (tributary of Rio Parnaiba) at Teresina, S 5° 5' 30” W 42° 49' 30", December 1865.

MCZ 19382, 1, South America, Brazil, Piauí, Rio Poti (tributary of Rio Parnaiba) at

Teresina, S 5° 5' 30” W 42° 49' 30", December 1865.

Schizodon fasciatus: USNM 377362, 2 (133.9 – 137.2 mm SL), 27 October 2001.

USNM 229056, 1 (84.4 mm SL), South America, Brazil, Amazonas, Parana De

Janauaca, Entrada Do Lago Do Castanho, 28 September 1977.

USNM 167827, 1 (111.6 mm SL), South America, Peru, Lago Cashiboya, 1920.

USNM 280717, 2 (81.3 – 92.0 mm SL), South America, Peru, Loreto, Green Water

Cano On Left Bank of Rio Manite, About 8 km Upriver of Junction of Rio Manite and

Rio Amazonas, S 3° 31' 12" W 72° 40' 12", 21 August 1986.

Schizodon isognathus: USNM 348701, 1 (235.7 mm SL), South America, Venezuela,

Portuguesa, Guanare-Guanarito Road At Road km 60, N 8° 49' 39" W 69° 20' 41”, 24

February 1998.

FMNH 59508, 3 (106.4 – 201.7 mm SL), South America, Brazil, Sao Luiz de Caceres,

24 May 1908.

207

MSUSP 88603, 1.

Schizodon jacuiensis: MCZ 76344, 1, South America, Brazil, Rio Grande do Sul, Rio

Guaiba at Porto Alegre, S 30° 3' 29” W 51° 10' 29”, 7 September 1987.

Schizodon knerii: ANSP 171830, 2 (124.6 – 125.2 mm SL), South America, Brazil,

Minas Gerais, Rio Salinas, trib. Rio Verde Grande, 51.0 km WSW from Monte Azul on road to Jaíba, S 15° 12' 53'' W 43° 15' 49'', 19 July 1993.

ANSP 171831, 3 (175.7 – 200.9 mm SL), South America, Brazil, Minas Gerais, Rio

Verde Grande, on road from Montes Claros to Janauba, S 16° 39' 1'' W 43° 42' 49'', 20

July 1993.

Schizodon nasutus: USNM 181760, 3 (142.5 – 227.7 mm SL), South America,

Paraguay, Laguna, Rio Tebicuary, near Florida, 24 September 1956.

Schizodon scotorhabdotus: FMNH 85478, 2 (33.9 – 49.4 mm SL), South America,

Venezuela, Apure, river 24 km S of Biruaca on road to San Juan de Apayara, 7 January

1975.

Schizodon vittatus: USNM 124401, 1 (223.0 mm SL), South America, Peru, Rio

Ampiyacu, 14 December 1935.

Synaptolaemus cingulatus: AUM 44098, South America, Venezuela, Amazonas, Río

Ventuari, near ornamental fish market in the river, N 4° 4' 32” W 66° 53' 34”, 3 April

2005.

Characidae:

Acinocheirodon melanogramma: USNM 345666, 10 (25.0 – 25.9 mm SL), South

America, Brazil, Minas Gerais, Riacho afluente do Rio Jequitai, 20 July 1994.

208

Actobyrcon: USNM 305499, 10 (37.6 – 41.4 mm SL), South America, Bolivia, Tarija,

River Pilcomayo at Villamontes RR bridge, S 21° 16' 48" W 63° 28' 12”, 1 October

1988.

Aphyocharax pusillus: USNM 361472, 10 (45.1 – 54.8 mm SL), South America, Peru,

Cusco, Rio Urubamba, 4 November 1998.

Astyanacinus moorii: USNM 86772, 4 (61.2 – 68.6 mm SL), South America, Bolivia,

Popoi River, September 1921.

USNM 326453, 3 (42.4 – 45.8 mm SL), South America, Brazil, Mato Grosso, Arroio

Afluente Do Rio Do Bugre (Afl. Rio Jauru) No km 165 Da Estrada Porto

Esperidiao/Pontes E Lacerda (Br 174), ca. 48 km De Porto Esperidiao (Sistema Do Rio

Paraguai), Porto Esperidiao, 13 August 1991.

Astyanax mexicanus: USNM 310222, 10 (57.2 – 72.4 mm SL), United States, Texas,

Kinney County, stream 7.4 miles West of Bracketville on US 90, 21 September 1960.

Bario steindachneri: USNM 167813, 9 (54.4 – 89.8 mm SL), South America, Peru,

Brook near Rio Itaya near Iquitos, September 1920.

Bramocharax bransfordii: USNM 78099, 2 (101.6 – 107.7 mm SL), Central America,

Nicaragua, Managua, July 1907.

Brycon amazonicus: USNM 400233, 6 (140.4 – 195.8 mm SL), South America,

Venezuela, Delta Amacuro, Tidal stream on river shore, 49 Nm from sea bouy, N 8°

37' 36” W 60° 49' 36”, 20 November 1979.

Bryconamericus emperador: USNM 78451, 10 (46.5 – 65.0 mm SL), South America,

Panama, Canal Zone, Rio Gatun and creek above Monte Liria, 28 March 1911.

209

Carlana eigenmanni: USNM 289108, 10 (37.6 – 55.9 mm SL), South America,

Panama, Comarca Kuna Yala, Nuisigandi at 3 km W of Nusigandi, N 9° 19' 12" W 79°

0' 0", 28 February 1985.

Chalceus erythrurus: USNM 119946, 10 (153.4 – 171.1 mm SL), South America,

Brazil, Hyavara, 1865.

Compsura heterura: USNM 356661, 10 (25.8 – 31.4 mm SL), South America, Brazil,

Minas Gerais, Missoes Riacho Na Fazenda Proximo Da Cidade De Missoes, 21 July

1994.

Creagrutus affinis: USNM 78566, 10 (41.2 – 50.0 mm SL), Central America, Panama,

Darien Province, Rio Cupe at Boca De Cupe, 24 February 1912.

Ctenobrycon spirulus: USNM 308098, 10 (33.6 – 45.0 mm SL), South America, Brazil,

Amazonas, Lago Terra Preta, Janauari, 14 April 1977.

MZUSP 40464, 4 (44.2 – 55.6 mm SL), South America, Brazil, Goiás, Flores de Goiás,

Poço da Gandaia (lagoa marginal do rio Paranã), S 14° 26’00” W 47° 03’00”, 11-12

September 1988.

Cyanocharax dicroptomicus: USNM 337609, 10 (29.6 – 37.1 mm SL), South America,

Brazil, Rio Grande Do Sul, Rio Forqueta at Marques De Souza, S 29° 19' 12" W 52°

49' 48”, 7 December 1979.

Deuterodon iguape: USNM 279630, 8 (48.7 – 97.5 mm SL), South America, Brazil,

Santa Catarina, Afluente Do Rio Itajai, NA Estrada Blumenau-Rio do Sul, Proximo a

Ibirama, 21 September 1985.

Diapoma terofali: USNM 270284, 10 (35.8 – 41.2 mm SL), South America, Brazil,

Rio Santa Maria, Trecho Entre Dom Pedrito E Livramento, 26 October 1982.

210

Galeocharax gulo: USNM 305367, 4 (119.4 – 210.4 mm SL), South America, Bolivia,

Beni, Rio Curiraba at 10 km NE El Porvenir Biological Station, 40 air km E of San

Borja, S 14° 55' 12” W 66° 16' 48", 31 August 1987.

Gephyrocharax atricaudata: USNM 78525, 10 (28.4 – 45.4 mm SL), Central America,

Panama, Canal Zone, Small Creek, Corozal, 17 February 1911.

Glandulocauda melanopleura: USNM 381400, 10 (23.0 – 34.3 mm SL), South

America, Brazil, Sao Paulo, Corrego Mutuca, Estacao Biologica de Boraceia, a stream draining into the rio Claro which flows to the west then north into the upper Tiete in

Estacao Biologica de Boraceia, Salesopolis.

Gymnocorymbus ternetzi: USNM 305418, 10 (31.1 – 37.7 mm SL), South America,

Bolivia, Beni, Borrow Pit by road at 1.5 km W Rio Matos crossing 45 air km E of San

Borja, S 14° 55' 47” W 66° 16' 48", 24 August 1987.

Hasemania hanseni: USNM 292198, 10 (21.0 – 21.4 mm SL), South America, Brazil,

Brazilian Federal District, Riberao Santana At Rd. Crossing About 30 Air km S

Barragem Do Paranoa (Bartolomew/Parana Syst.), S 15° 55' 11, W 47° 46' 12", 12

November 1984.

Hemibrycon dariensis: USNM 293245, 10 (35.2 – 61.8 mm SL), North America,

Panama, Darien Province, Rio Pucuro About 3-4 km Above Confl. With Rio Tuira

(Pacific), N 8° 0' 0" W 77° 31' 11”, 17 February 1985.

Heterocheirodon yatai: USNM 356385, 10 (30.0 – 30.9 mm SL), South America,

Brazil, Arroio Garupa, Divisa Alegrete E Quarai, 11 November 1986 to 12 November

1986.

211

Hollandichthys multifasciatus: USNM 297983, 10 (36.9 – 88.2 mm SL), South

America, Brazil, Sao Paulo, Innominate Clear Water Stream Crossing at km 94,

Northwest of Bertioga, S 23° 46' 11” W 46° 4' 12", 23 February 1988.

Hyphessobrycon megalopterus: USNM 334256, 10 (19.0 – 22.0 mm SL), South

America, Brazil, Brazil-Bolivian Border, Region Between Guajara-Mirim and Matto

Grosso From the Guapore Drainage, 1970.

Jupiaba anteroides: ANSP 188781, 4 (52.9 – 74.8 mm SL), South America, Venezuela,

Amazonas, left bank trib R. Siapa ca. 1 km upstream from mouth, trib mouth below

Salto Oso & above Salto Sardinas on R. Siapa., N 1° 26' 24'' , W 65° 40' 1'', 14 March

2005.

Knodus serptentrionalis: USNM 330851, 10 (30.0 – 33.5 mm SL), South America,

Peru, Lo, Maynas, Pv Arcadia, R. Napo, Isla, 7 November 1993.

Macropsobrycon uruguayensis: USNM 268449, 3 (30.6 – 35.9 mm SL), South

America, Brazil, Rio Grande do Sul, Arroio Sarandi, At Stream Under Road Crossing

On Road Between Pelotas and Jaguarao, Trib. of Lago Mirim North of Arroio Grande,

14 December 1979.

Markiana geayi: USNM 258527, 10 (46.3 – 63.7 mm SL), South America, Venezuela,

Apure, Main Channel of Rio Cunaviche Along Edge of Cunaviche, N 7° 24' 0" W 67°

27' 0", 22 January 1983.

Mimagoniates microlepis: USNM 249874, 10 (27.4 – 30.0 mm SL), South America,

Brazil, Rio de Janeiro, Rio Da Areia, First Tributary On Right Side of Road Going

Northeast, Rio Jurumirim, On Rd 155, Northeast of Angra Dos Reis, S 22° 52' 12" W

44° 15' 0", 10 November 1982.

212

Moenkhausia xinguensis: ANSP 161354, 1 (45.8 mm SL), South America, Venezuela,

Amazonas, Rio Ventuari from ca. 10-12 km above confluence with Rio Orinoco, N 4°

4' 0'', W 66° 56' 0'', 25 March 1987.

MZUSP 36806, 4 (33.0 – 45.8 mm SL), South America, Pará, Altamira, Cachoeira do

Espelho, Rio Xingu, S 3 48’00”, W 52 32’00”, 23-26 October 1986.

Odontostilbe splendida: USNM 349419, 6 (20.0 – 20.5 mm SL), South America,

Venezuela, Portuguesa, Rio Las Marias, at Quebrada Seca (Town), Approximately 45

Min. Upstream By Car From Hwy. 5, 22 km NNW Guanare, 28 February 1998.

Oligosarcus paranensis: USNM 285718, 4 (34.6 – 41.3 mm SL), South America,

Brazil, Rio Grande do Sul, Arroio Lageado Uniao, Em Linha Dos Lima, Palmitinho,

Rs. (Afluente Do Rio Pardo, Rio Uruguai), 22 December 1985.

USNM 285718, 1 (57.6 mm SL),

Paracheirodon axelrodi: USNM 216901, 5 (14.2 – 23.6 mm SL), South America,

Brazil, Amazonas, Rio Negro an Igarape of the Varzea At Tapurucuara, 19 October

1972.

Paragoniates alburnus: USNM 233651, 10 (36.3 – 71.7 mm SL), South America,

Venezuela, Delta Amacuro, Rio Orinoco, backwater Cano Araguao, 112 nautical miles upstream from sea buoy, N 8° 37' 48"W 61° 43' 11”, 21 February 1978.

Phenacogaster franciscoensis: USNM 302300, 10 (36.0 – 37.5 mm SL), South

America, Brazil, Bahia, Rio Do Braco Drainage, 2 km From Town of Rio Do Braco,

Rio Do Braco On Fazenda Santa Luzia Opposite Houses of Farm Workers and Cacau

Drying Sheds, S 14° 40' 11” W 39° 16' 12", 27 July 1988.

213

Piabarchus analis: USNM 326513, 6 (17.3 – 27.9 mm SL), South America, Brazil,

Mato Grosso, Rio Jauquara Em Jauquara (Afl. Rio Dos Passaros, Rio Paraguai), Barra

Do Bugres, 10 August 1991.

Piabina argentea: USNM 292220, 10 (42.5 – 64.9 mm SL), South America, Brazil,

Brazilian Federal District, Riberao Santana At Rd. Crossing About 30 Air km S

Barragem Do Paranoa (Bartolomew/Parana Syst.), S 15° 55' 11” W 47° 46' 12", 12

November 1984.

Poptella paraguayensis: USNM 326332, 10 (34.4 – 46.6 mm SL), South America,

Brazil, Mato Grosso, Rio Paraguai em Caceres e arredores, Caceres, 11 August 1991 to 12 August 1991.

Prionobrama paraguayensis: USNM 326332, 10 (32.0 – 32.9 mm SL), South America,

Brazil, Mato Grosso, Rio Paraguai em Caceres e arredores, Caceres, 11-12 August

1991.

Prodontocharax: USNM 305484, 10 (33.0 – 42.8 mm SL), South America, Bolivia,

Tarija Department, River Pilcomayo at Villamontes RR bridge, S 21° 16' 48", W 63°

28' 11”, 1 October 1988.

Rachoviscus crassiceps: USNM 228089, 9 (20.0 – 20.8 mm SL), South America,

Brazil, Parana, Coastal Stream, Brejatuba Near Guaratuba, 13 June 1981.

Rhinobrycon negrensis: USNM 270187, 10 (29.4 – 42.6 mm SL), South America,

Venezuela, Amazonas, Small Cano Off Cano Urami, Just Upriver of Santa Lucia, N 1°

16' 48" W 66° 50' 59”, 6 December 1984.

214

Roeboexodon guyanensis: USNM 409858, 1 (37.1 mm SL), South America, Suriname,

Left tributary of Middle Paloemeu River, 500 m downstream Kasikasima basecamp, N

2° 58' 48”, W 55° 23' 3”, 21 March 2012.

USNM 225627, 1 (30.8 mm SL), South America, Suriname, Nickerie District,

Corantijn River at km 180, side channel of main river along Surinamese shore, N 5° 7'

47”, W 57° 17' 59”, 8 September 1980.

USNM 225764, 2 (20.7 – 35.6 mm SL), South America, Suriname, Nickerie District, stream near Camp Anjoemara, N 4° 49' 48" W 57° 25' 47”, 14 September 1980.

Salminus: USNM 247317, 1 (91.0 mm SL), South America, Paraguay, Alto Parana,

Reservoir of Dam On Rio Acaray, 16 May 1982.

Serrapinus calurius: USNM 181665, 9 (25.0 – 48.8 mm SL), South America, Paraguay,

Pond near Ybytymi Rio Tebicuary Drainage, 18 January 1956.

Spintherobolus broccae: USNM 297997, 10 (15.0 – 19.7 mm SL), South America,

Brazil, Sao Paulo, Innominate Clear Water Stream Crossing Sp-98 (Road) At km 94,

Northwest of Bertioga, S 23° 46' 11” W 46° 4' 12", 23 February 1988.

Stygichthys typhlops: ANSP 100891, 1 (27.8 mm SL), South America, Brazil, Minas

Gerais, Jaiba, S 15° 20' 30” W 43° 40' 30”, 16 May 1962.

MZUSP 87678, 5 (32.0 – 40.3 mm SL),South America, Brazil, Minas Gerais, Cisterna da fazenda do Lajeado ou Mandioqueu (prop. Vicente Ildeu dos Santos), córrego

Escuro, sistema rio Verde Grande, S 15°27’0” W 43° 45’09”, 28 April 2004.

Tetragonopterus argenteus: USNM 305373, 9 (49.2 – 65.2 mm SL), South America,

Bolivia, Beni, Rio Matos below road crossing, 48 km east San Borja, S 14° 55' 11” W

66° 16' 48", 28 August 1987.

215

MZUSP 59491, 3 (47.0 – 58.1 mm SL), South America, Brazil, Mato Grosso do Sul,

Brejo da Santa Sofia (pântano), Aquidauana, S 19° 36’29” W 56° 20’47”, 9 March

1998.

Thayeria oblique: USNM 304944, 10 (31.7 – 40.7 mm SL), South America, Peru, Rio

Negro, March 1963.

Citharinidae:

Citharinus citharus: USNM 229805, 4 (144.3 – 180.4 mm SL), Middle/Upper Niger at

Niamey, 18 July 1980 to 21 July 1980.

Citharinus congicus: CUMV 96224, 2 (52.6 – 63.1 mm SL), Africa, Democratic

Republic of the Congo, Orientale, Tshopo, Marché Central in Kisangani, N 0° 30' 46”

E 25° 11' 58”, 9 March 2010.

Chilodontidae:

Caenotropus labyrinthicus: USNM 377389, 3 (100.8 – 134.1 mm SL), South America,

Guyana, Rupununi River, Taraqua Creek, 31 October 2001.

USNM 270237, 7 (54.8 – 67.3 mm SL), South America, Venezuela, Bolivar, Small

Cano Connecting With Rio Orinoco Immediately South of El Burro, N 6° 10' 47” W

67° 25' 12", 9 December 1984.

Caenotropus maculosus: USNM 236103, 3 (109.5 – 119.7 mm SL), South America,

Suriname, Brokopondo, Kleine Saramacca R. 14 km ESE of Confluence With

Saramacca R. (District Brokopondo) Sandbottom, Rapids, 28 February 1967.

USNM 409909, 4 (52.9 – 112.3 mm SL), South America, Suriname, Upper Paloemeu

River, 500 m downstream of basecamp, N 2° 28' 37” W 55° 37' 48”, 10 March 2012.

216

Caenotropus mestomorgmatos: OS 18323, 4 (58.1 – 63.2 mm SL), South America,

Peru, Loreto, Maynas, "Beach ""Mis Playa"" on Nanay River about 20 minutes upstream from Santa Clara by motorized canoe", S 3° 46' 51” W 73° 21' 50”, 11 August

2010.

OS 18772, South America, Peru, Loreto, Maynas, Dead arm and shore of Nanay River near mouth of Caño Shirui near Pampa Chica, S 3° 45' 6” W 73° 17' 13”, 9 August

2010.

OS 18346, South America, Peru, Loreto, Maynas, Beach at Nanay River close to

Iquitos, S 3° 45' 6” W 73° 18' 58”, 8 August 2010.

Chilodus fritilus: USNM 343647, 10 (28.5 – 45.3 mm SL), South America, Peru, Madre de Dios, Santuario Nacional Pampas Del Heath, Rio Palma Real, Ox-Bow Lagoon,

30m Down River From Puesto De Vigilancia Enahuipe, 29 July 1995.

Chilodus gracilis: USNM 179563, 10 (39.2 – 51.9 mm SL), South America, Brazil,

Rio Urubu 25 mi. From Itacoatiara.

Chilodus punctatus: USNM 280444, 7 (58.6 – 74.8 mm SL), South America, Peru,

Loreto, Quebrada Corrientillo, At Corrientillo, On Road Running West From Iquitos

To Rio Nanay, S 3° 49' 48" W 73° 13' 11”, 19 August 1986.

USNM 304439, 3 (40.3 – 67.7 mm SL), South America, Venezuela, Laguna La Ceiba,

Rio Orinoco, Los Castillos Di Guayania, 31 August 1968.

Chilodus zunevei: USNM 163212, 10 (57.2 – 65.4 mm SL), South America, Guyana,

1952.

217

Crenuchidae:

Amocryptocharax elegans: USNM 270209, 10 (30.4 – 40.4 mm SL), South America,

Venezuela, Amazonas, Balneiria Pozo Azul, approximately 1 km To East of Puerto

Ayacucho To Solano Road, approximately 30 km N of Puerto Ayacucho, N 5° 52' 47”

W 67° 28' 11”, 10 December 1984.

Characidium lanei: USNM 285777, 6 (29.3 – 40.2 mm SL), South America, Brazil,

Parana, Vala Na Estrada Guaratuba - Joinville, Pr, 28 December 1975.

USNM 273331, 2 (26.4 – 31.7 mm SL), South America, Brazil, Sao Paulo, South

Branch of Rio Tavares, Ubatuba System At Nursery of "Flora Brasilia", S 23° 26' 59”

W 45° 5' 30”, 31 October 1982.

Characidium pterostictum: USNM 285841, 6 (38.1 – 54.3 mm SL), South America,

Brazil, Rio Grande do Sul, Rio Tuparendi, Tuparendi, Rs (Afluente Do Rio Santo

Cristo, Rio Uruguai), 20 December 1985.

USNM 285893, 3 (49.8 – 52.8 mm SL), South America, Brazil, Rio Tuparendi

Tuparendi, Rs (Afluente Do Rio Santo Cristo, Rio Uruguai), 20 December 1980.

Crenuchus spilurus: USNM 270130, 10 (27.6 – 33.3 mm SL), South America,

Venezuela, Amazonas, Small Cano Off Cano Urami, Just Upriver of Santa Lucia, N 1°

16' 48" W 66° 50' 59”, 6 Decemner 1984.

MZUSP 7479, 4 (35.0 – 40.4 mm SL), South America, Brazil, Amazonas, Igarapé afluente do rio Sanabani., Silves, S 2° 45’00 W 58° 20’00”, 7 December 1967.

Melanocharacidium pectorale: USNM 401517, 1 (33.0 mm SL), South America,

Guyana, Cuyuni-Mazaruni, Braided channel about 15km upstream from Devil's Hole on Cuyuni River. Habitats include sandbars, beaches, banks with overhanging

218 vegetation, and fast rapids over bedrock and loose stone, N 6° 47' 49”, W 59° 59' 34”,

1 February 2011.

USNM 235406, 4 (25.8 – 27.2 mm SL), South America, Venezuela, Bolivar, Rio

Orocopiche, ca. 15 km From Mouth In Rio Orinoco, Downstream From Route 19

Bridge, N 8° 3' 0" W 63° 40' 12", 3 November 1979.

Poecilocharax weitzmani: USNM 228091, 9 (19.1 – 23.9 mm SL), Guyana, Aquarium

Spms Out of Georgetown.

Ctenoluciidae:

Boulengerella maculate: USNM 179506, 7 (127.0 – 170.2 mm SL), South America,

Brazil, Rio Urubu 25 mi. from Itacoatiara.

Ctenolucius beani: USNM 78630, 8 (144.1 -230.0 mm SL), Central America, Panama,

Rio Abaco, 22 April 1911.

USNM 310452, 4 (144.1 – 152.8 mm SL), Central America, Panama, Cr. At Bridge

On Iah (Inter-American Highway) About 4 mi. E. of Pacora X-Road, 24 February 1962.

Curimatidae:

Curimata cerasina: FMNH 94591, 1, South America, Guarico, Laguna Los Laurels between Camaguan and Cano Falcon, N 8° 9' 0" W67° 43' 0”, 17 March 1972.

Curimata cyprinoides: USNM 225214, 9 (166-184.4 mm SL), South America,

Suriname, Nickerie District, Corantijn River, N 5° 0' 0" W 57° 16' 48", 17 May 1980.

Curimata incompta: USNM 273308, 3 (43.2-88.5 mm SL), South America, Venezuela,

Bolivar, Small Cano connecting with Rio Orinoco immediately south of El Burro, N

6° 10' 48” W 67° 25' 12", 9 December 1984.

219

Curimata inornata: USNM 360071, 2, (97.2-107.1 mm SL), South America, Brazil,

Tocantins, Brejinho de Nazare, Lago Pedra Do Santo, 12 December 1995.

Curimata knerii: USNM 267975, 9, (122.9-148.6 mm SL), South America, Brazil,

Para, Rio Tapajos at Itaituba, September 1983 to October 1983.

Curimata macrops: USNM 258769, 7, (68.8-130.0 mm SL), South America, Brazil,

Piaui, Parnaiba at Terezina.

Curimata ocellatus: MCZ 20339, 1, South America, Brazil, Pará, Rio Xingu, probably north of cascade region, S 3° 5' 30” W 51° 53' 30”, 1 September to 30 September 1865.

MCZ 60884, 1, South America, Brazil, Pará, Rio Xingu, probably north of cascade region, S 3° 5' 30” W 51° 53' 30”, 1 September 1865.

Curimata plumbea: INPA 11236, 5, South America. Brazil, Pará, Rio Tocantins, S 3°

45’ 58” W 49° 40’21”, 16 November 1981.

Curimata roseni: USNM 304863, 2, South America, Venezuela, Laguna a 20m De La

Orilla Del Rio Mavaca ca. 2 km arriba dol camp base, 29 March 1988.

USNM 287593, 2 (108.3 – 125.1 mm SL), South America, Bolivia, Beni, Itenez,

(Londra) and Rio Blanco, 1 September 1984.

USNM 305143, 6 (45.8 – 64.5 mm SL), South America, Brazil, Amazonas, Parana De

Janauaca, Entrada Do Lago Do Castanho, 10 February 1977.

Curimata vittata: USNM 242131, 2, (181.5-204.2 mm SL), South America, Brazil,

Amazonas, Rio Tefe at Jurupari, 2 August 1979.

USNM 242132, 6 (103.7 – 118.1 mm SL), South America, Brazil, Amazonas, Rio Tefe,

Resaca Da Paula, 1 August 1979.

220

USNM 304862, 2 (143.3 – 160.4 mm SL), South America, Venezuela, Laguna a 20 m

De La Orilla Del Rio Mavaca ca 2 km Arriba Dol Camp Base Exp. Tapirapeco, 29

March 1988.

USNM 267316, 1 (114.5 mm SL), South America, Brazil, Amazonas, Rio Negro At

Cucui.

Curimatella alburna: USNM 242141, 3, (131.4-164.3 mm SL), South America, Brazil,

Roraima, Rio Branco at Marara, 26 October 1979.

USNM 242143, 1 (148.5 mm SL), South America, Brazil, Amazonas, Rio Tefe,

Mucura, 30 July 1979.

USNM 377369, 3 (127.7 – 165.4 mm SL), South America, Guyana, Crashwater Creek,

29 November 2001.

Curimatella dorsalis: USNM 258284, 10, (54.7-81.8 mm SL), South America,

Venezuela, Apure, Estersos on South Bank of Apure River ca. 3 km., N 7° 52' 48” W

67° 34' 47”, 21 January 1983.

Curimatella immaculata: USNM 268027, 10, (51.8 – 75.6 mm SL), South America,

Brazil, Roraima, Bio Branco at Marara, 26 October 1979.

Curimatella lepidura: USNM 302357, 9, (46.6-60.2 mm SL), South America, Brazil,

Rio San Francisco West of Baringa.

MCZ 20291, 1, South America, Brazil, Minas Gerais, Rio Sao Francisco, 'below the falls' [probably near Pirapora], S 17° 20' 10” W 44° 57' 17”, 1 January 1867.

MCZ 20292, 1, South America, Brazil, Minas Gerais, Rio Sao Francisco, 'below the falls' [probably near Pirapora], S 17° 20' 10” W 44° 57' 17”, 1 January 1867.

221

MCZ 87454, 2, South America, Brazil, Minas Gerais, Rio Sao Francisco, 'below the falls' [probably near Pirapora], S 17° 20' 10” W 44° 57' 17”, 1 January 1867.

Curimatella meyeri: USNM 268036, 2 (140.2 – 147.6 mm SL), South America, Brazil,

Rondonia, Rio Madeira; Calama Beach, December 1980.

USNM 243230, 3 (103.4 – 125.7 mm SL), South America, Peru, Pucallpa,

Yarinacocha, 5 March 1976.

USNM 261399, 1 (126.5 mm SL), South America, Peru, Ucayali, Rio Neshuya,

Pucallpa km 60., 29 September 1971.

USNM 278582, 1 (108.6 mm SL), South America, Bolivia, Beni, Rio Blanco, 2

September 1984.

USNM 243229, 3 (93.5 – 139.7 mm SL), South America, Peru, Pucallpa, Rio Ucayali,

29 May 1979.

Curimatopsis crypticus: USNM 269924, 8, (28.7-39.1 mm SL), South America,

Venezuela, Amazonas, Cano Provincial approximately 20 km North of Puerto

Ayacucho, N 5° 49' 48" W 67° 30' 0", 1 December 1984.

Curimatopsis evelynae: USNM 214794, 10 (30.9-36.2 mm SL), South America, Brazil,

Amazonas, Igarape Anapichi 64 miles NW of interection of Rio Branco and Rio Negro,

September 1975.

Curimatopsis macrolepis: USNM 304798, 10 (32.7- 42.1 mm SL), South America,

Brazil, Amazonas, Rio Arirara at mouth of the Rio Negro, 1 February 1980.

Curimatopsis microlepis: USNM 268867, 1, (67.3 mm SL), South America, Brazil,

Amazonas, Rio Solimoes, lago near Beruri, 30 January 1985.

222

USNM 268035, 1 (48.38 mm SL), South America, Brazil, Amazonas, Rio Tefe;

Jurupari Beach, 1 August 1979.

Curimatopsis myersi: USNM 247314, 8 (27.7-35.2 mm SL), South America, Paraguay,

San Pedro, Swamp 3 km Northeast of Lima, S 23° 55' 12", W 56° 28' 48”, 31 October

1981.

Cyphocharax abramoides: USNM 267951, 10 (115.6-138.1 mm SL), South America,

Brazil, Para, Rio Trombetas edge of river channel at Cumira, October to November

1983.

Cyphocharax aspilos: USNM 121316, 10 (110.3-145.2 mm SL), South America,

Venezuela, Cano 0.5 mi. West of Sinamaica, 11 March 1942.

Cyphocharax festivus: USNM 235440, 10 (29.2-41.6 mm SL), South America,

Venezuela, Delta Amacuro, Cano Paloma, 92 Nm from Sea Buoy, N 8° 28' 0”, W 61°

25' 36”, 21 November 1979.

Cyphocharax gangamon: USNM 267991, 12 (25.8 - 35.2 mm SL), South America,

Brazil, Mato Grosso, Rio Arinos at Porto Dos Guachos, 18 August 1984.

Cyphocharax gilbert: USNM 298247, 10 (105.6 - 132.7 mm SL), South America,

Brazil, Bahia, North Branch of Rio Jucurucu in town of Itamaraju, S 17° 4' 48”, W 39°

31' 12", 2 August 1988.

Cyphocharax gilii: USNM 311133, 4 (73.1-82.3 mm SL), South America, Brazil, Rio

Itiguira, Baia Grande, Santo Antonio Do Paraiso.

USNM 229440, 1 (76.0 mm SL), South America, Paraguay, Paraguari, Arroyo

Corrientes, About 3 km Downstream From the Present Boundary of Ybycui National

Park, 31 May 1981.

223

Cyphocharax gouldingi: USNM 267992, 10 (96.0-120.3 mm SL), South America,

Brazil, Para, Rio Itacaiunas, Serra Dos Carajas, Igarape Do Pujica, 15 October 1983.

Cyphocharax helleri: USNM 225300, 10 (43.4-50.3 mm SL), South America,

Suriname, Nickerie District, Tributary to Sisa Creek, North side approximately 700 m downstream of crossing of Amotopo to Camp Geologie Rd, N 3° 42' 0" W 57° 42' 0",

20 September 1980.

Cyphocharax leucostictus: USNM 268020, 10 (53.8- 83.3 mm SL), South America,

Brazil, Amazonas, Rio Negro at Anavilhanas, October 1980.

USNM 267981, 5 (78.8 – 96.6 mm SL), South America, Brazil, Para, Rio Tapajos;

Itaituba; Edge of River Channel, September 1983 to October 1983.

Cyphocharax magdalenae: FMNH 56316, 2, South America, Colombia, Soplaviento,

11 January to 13 January 1912.

Cyphocharax microcephalus: MCZ 785, 1 (70.5 mm SL), South America, Suriname,

Exact locality unknown, N 4° 14' 27” W 55° 44' 10”, 1861.

MCZ 92958, 3, South America, Suriname, N 4° 14' 27” W 55° 44' 10”, 1 January 1861.

Cyphocharax modestus: USNM 295904, 5 (91.9-123.4 mm SL), South America,

Brazil, Sao Paulo, Rio Tiete.

USNM 297381, 3 (91.4 – 101.8 mm SL), South America, Brazil, Sao Paulo, Rio

Corumbatai, Corumbatai.

USNM 243239, 2 (88.5 – 88.9 mm SL), South America, Sao Paulo, Corrego Da Barra.

Pirisunga.

Cyphocharax nagelii: USNM 295975, 10 (99.2 – 125.3 mm SL), South America,

Brazil, Sao Paulo, Rio Mogi-Guacu near Pirassununga.

224

Cyphocharax notatus: USNM 194342, 2 (53.1 – 76.6 mm SL), South America, Brazil,

Mato Grosso, Upper Juruena River, 19 July 1962.

USNM 311134, 3 (46.2 – 53.9 mm SL), South America, Brazil, Amazonas, Rio Negro,

Parana Do Jacare, 7 August 1979.

Cyphocharax platanus: USNM 295976, 1 (139.8 mm SL), South America, Brazil, Rio

Grande do Sul, Rio Ibicui near mouth, 20 July 1986 to 21 July 1986.

USNM 295978, 1 (117.8 mm SL), South America, Paraguay, Asuncion Bay.

USNM 295977, 2 (92.0 – 93.1 mm SL), South America, Brazil, Rio Grande do Sul,

Rio Uruguay.

Cyphocharax plumbeus: USNM 267968, 10 (64.5 – 84.9 mm SL), South America,

Brazil, Para, Rio Tapajos at Itaituba, September 1983- October 1983.

Cyphocharax saladensis: USNM 296374, 2 (61.9 – 65.8 mm SL), South America,

Brazil, Rio Grande Do Suk, along road from Porto Alegre to Pelotas, 21 September

1974.

USNM 220578, 2 (56.9 – 57.7 mm SL), South America, Brazil, Rio Grande do Sul,

Vala Na Br 116, Estrada Porto Alegre - Pelotas, S. C., 21 September 1977.

USNM 94307, 1 (54.2 mm SL), South America, Argentina, Salado.

Cyphocharax santacatarinae: USNM 296460, 2 (11.6 – 111.6 mm SL), South

America, Brazil, Santa Carina.

USNM 295898, 3 (122.2 – 127.5 mm SL), South America, Santa Catarina, Joinville.

USNM 296461, 3 (109.5 – 131.9 mm SL), South America, Santa Catarina, Joinville.

225

Cyphocharax spilotus: USNM 313877, 8 (42.7 – 83.5 mm SL), South America,

Argentina, Formosa, Ditch along road between Ruta Nacional 11 and J.C. Sanchez km

896, August 1986.

USNM 307242, 1 (42.7 mm SL), South America, Brazil, Rio Grande do Sul, Arroio

Do Salso, Estrada Livramento Rosario Do Sul, Rosario Do Sul, Rs (Afluente Do Rio

Ibucui, Uruguai).[Arroio Do Salso, on road from Livramento to Rosario Do Sul,

Rosario Do Sul. Tributary of Rio Ibicui; Uruguay River Basin], 23 July 1986.

USNM 313877, 1 (64.0 mm SL), South America, Argentina, Formosa, Ditch Along

Road Between Ruta Nacional 11 and J.C. Sanchez km 896, August 1986.

Cyphocharax spiluropsis: USNM 311100, 5 (55.5 – 80.6 mm SL), South America,

Peru, Ucayali, Pucallpa, Masisea, Loboeocha Ortega, Hernan, 17 March 1983.

USNM 278560, 4 (40.2 – 57.2 mm SL), South America, Bolivia, Beni, Rio Itenez

(=Guapore) At Confluence of Rio Itenez and Rio Machupo, 5 September 1984.

Cyphocharax spilurus: USNM 269923, 1 (62.1 mm SL), South America, Venezuela,

Amazonas, Rio Orinoco Raudales de Ature, eastern shore, N 5° 35' 60” W 67° 37' 12",

2 December 1984.

USNM 225388, 2 (68.4 – 96.1 mm SL), South America, Suriname, Nickerie District,

Kamp Kreek 100 m north of turnoff to Camp Geology, N 4° 49' 12" W 57° 28' 12”, 13

September 1980.

USNM 225406, 3 (69.0 – 74.7 mm SL), South America, Suriname, Nickerie District,

Stream Entering Corantijn River At Approximately km 385 Slightly North of Tiger

Falls, N 4° 0' 0" W 58° 1' 48", 16 September 1980.

USNM 311126, 1 (87.0 mm SL), South America, Guyana, Wismar.

226

Cyphocharax vanderi: USNM 243241, 1 (68.5 mm SL), South America, Brazil, Sao

Paulo.

USNM 296257, 4 (18.9 – 22.6 mm SL), South America, Brazil, Sao Paulo, Rio

Corumbatai, Corumbatai.

Cyphocharax voga: USNM 295885, 10 (85.7 – 117.4 mm SL), South America, Brazil,

Rio Grande do Sul, Arroio Do Salso, Estrada Livramento Rosario Do Sul, 23 July 1986.

Potamorhina altamazonica: USNM 273611, 1 (137.5 mm SL), South America, Peru,

Ucayali, Pucallpa, Rio Ucayali, Utuquinia, 28 June 1983.

USNM 278575, 5 (68.0 – 155.5 mm SL), South America, Bolivia, Cochabamba, Rio

Chapare, 21 June 1982.

USNM 269929, 2 (74.4 – 91.4 mm SL), South America, Venezuela, Bolivar, Small

Cano Connecting With Rio Orinoco Immediately South of El Burro, N 6° 10' 48” W

67° 25' 12", 9 December 1984.

USNM 280446, 1 (105.5 mm SL), South America, Peru, Ucayali, Provincia Coronel

Portillo; main channel and side pools of Rio Ucayali, approximately 10 km upstream of Pucallpa, S 8° 31' 12” W 74° 22' 12", 25 August 1986.

USNM 228694, 1 (73.4 mm SL), South America, Venezuela, Monages, Cano Between

Rio Orinoco and Laguna Guatero, Near Barrancas, 142 Nm From Sea Bouy, N 8° 43'

12" W 62° 10' 48”, 12 November 1979.

Potamorhina latior: USNM 242144, 2 (195.7 – 197.9 mm SL), South America, Brazil,

Rondonia, Rio Machado, Lago Do Paraiso, 28 August 1977.

USNM 220196, 6 (180.1 – 230.5 mm SL), South America, Brazil, Rondonia, Rio

Machado, Rondonia Lago Do Parasio, 2 September 1977.

227

Potamorhina pristigaster: USNM 242147, 1 (205.52 mm SL), South America, Brazil,

Roraima, Rio Branco, Xeruini, 9 May 1979.

USNM 228693, 1 (108.77 mm SL), South America, Brazil, Amazonas, Parana De

Janauaca, Entrance Do Lago Do Castanho, 14 September 1977.

USNM 305223, 2 (63.6 – 66.7 mm SL), South America, Amazonas, Parana De

Janauaca, Entrada Do Lago Do Castanho, 28 September 1977.

Potamorhina squamoralevis: USNM 181711, 6 (98.7 – 125.9 mm SL), South America,

Paraguay, Asuncion Bay, Rio Paraguay, near Asuncion, 26 January 1957 to 29 January

1957.

Psectrogaster amazonica: USNM 332502, 3 (78.3 – 84.4 mm SL), South America,

Peru, Lo, Maynas, Pv Arcadia, Rio Napo, Cocha De Conchas, S 0° 59' 21” W 75° 18'

33”, 2 November 1993.

Psectrogaster ciliata: ANSP 135621, 7 (97.6 – 123.7 mm SL), South America,

Venezuela, Bolivar, Cano Chuapo, ca 20 min. downstream from Jabillal (opposite bank) on Rio Caura, N 7° 7' 0'', W 65° 0' 0'', 28 January 1977.

Psectrogaster curviventris: USNM 181714, 1 (177.3 mm SL), South America,

Paraguay, Rio Tebicuary, near Florida, Paraguay, 7 December 1956.

USNM 268039, 7 (127.5 – 155.4 mm SL), Brazil, Rhondonia, Rio Madeira; Calama;

Floodplain Lake, June 1980.

Psectrogaster essequibensis: USNM 268902, 3 (94.9 – 130.7 mm SL), South America,

Brazil, Para, Rio Tapajos, Itaituba, Edge of River Channel, September 1983 to October

1983.

228

USNM 268040, 3 (140.3 – 185.4 mm SL), South America, Brazil, Amazonas, Rio Tefe,

Lago Do Jacare, July 1980.

Psectrogaster falcata: USNM 268042, 10 (99.6 – 149.5 mm SL), South America,

Brazil, Para, Rio Tapajos, Itaituba, Edge of River Channel, September 1983 to October

1983.

Psectrogaster rutiloides: USNM 94660, 3 (104.5 – 121.2 mm SL), South America,

Brazil, Acre, Vicinity of Mouth of Rio Macauhan, tributary to Rio Yaco which flows into the Rio Purus, S 9° 19' 48" W 69° 0' 0", 1934.

USNM 278564, 2 (96.0 – 125.1 mm SL), South America, Bolivia, Beni, Rio Itenez

(Londra), 3 September 1984 to 4 September 1984.

USNM 261516, 5 (70.5 – 86.8 mm SL), South America, Peru, Loreto, Supaycocha,

Requena, 29 June 1979.

Psuedocurimata boulgengeri: MCZ 30931, 6, South America, Ecuador, Los Rios, Rio

Vinces at Vinces, S 1° 37' 30" W 79° 45' 30”.

Psuedocurimata patiae: FMNH 56554, 1, South America, Colombia: Telembi R.,

Barbacoas, 1913.

FMNH 56555, 3, South America, Colombia: Telembi R., Barbacoas, 1913.

Psuedocurimata peruana: FMNH 58673, 1, South America, Peru: Rio Chira at Sullana,

January 1919.

FMNH 70223, 3, South America, Peru: Exact locality not known.

Pseudocurimata lineopunctata: FMNH 80415, 6, South America, Ecuador: Rio

Basacito, Cachabi, Esmeraldas.

229

Psuedocurimata troschelli: USNM 311258, 10 (84.3 – 95.0 mm SL), South America,

Ecuador, Los Rios Province, Rio Vinces, 5 km Aguas Arriba De La Ciudad De Vinces,

8 July 1989.

Steindachnerina argentea: FMNH 85290, 2, South America, Colombia, Meta: Cano negro on Rd. to Puerto Porfia, E of Villavicencio, 27 March 1974.

Steindachnerina atratoensis: FMNH 56043, 9, South America, Colombia, Quibdo, 25

March 1912.

Steindachnerina bimaculata: USNM 278570, 3 (60.6 – 117.6 mm SL), South America,

Bolivia, Beni, Rio Isiboro, 19 November 1984.

USNM 327651, 6 (44.4 – 67.2 mm SL), Peru, Madre de Dios, Manu, Parque Nacional

Manu, Pakitza, Martin Pescador Stream, 26 September 1991.

Steindachnerina biornata: USNM 295339, 10 (98.3 – 125.66 mm SL), South America,

Peru, Madre de Dios, Parque Nacional Manu, Pakitza and vicinity, October 1987.

Steindachnerina brevipinna: USNM 295268, 7 (36.4 – 108.24 mm SL), South

America, Brazil, Rio Grande do Sul, Rio Ibicui near mouth, municipio de Itaqui, 20

July 1986 to 21 July 1986.

USNM 295265, 3 (91.9 – 103.3 mm SL), South America, Brazil, Rio Grande do Sul,

Segundo Rio Na Estrada Coimbra - Santo Angelo, Santo Angelo, Rs (Afluente Do Rio

Uruguai), 20 December 1985.

Steindachnerina conspersa: USNM 181636, 10 (79.5 - 104.8 mm SL), South America,

Paraguay, Asuncion Bay, Rio Paraguay, near Asuncion, 26 January 1957 to 29 January

1957.

230

Steindachnerina dobula: USNM 229203, 3 (103.7 – 111.0 mm SL), South America,

Peru, Rio Ucayali near San Pedro De Longin, 24 July 1981.

USNM 319655, 7 (46.1 – 79.2 mm SL), South America, Peru, Madre de Dios, Manu,

Manu National Park, Pakitza, Manu River Close To Pachija Mouth, 10 May 1991.

Steindachnerina elegans: USNM 318102, 10 (68.5 – 82.1 mm SL),South America,

Brazil, Minas Gerais, Tributary of Rio Jequitinhonha Known As Rio Ribeirao (See

Field Notes At Usnm) approximately 4-5 km ESE of Town of Jordania. Main Channel,

S 15° 54' 0" W 40° 10' 12", 11 July 1991.

FMNH 71247, 4, South America, Brazil, Minas Gerais: Sete Lagoas, into Rio das

Velhas, into Rio Sao Francisco, 5 May 1908.

Steindachnerina fasciata: USNM 273306, 1 (34.8 mm SL), South America, Brazil,

Rondonia, Rio Urupa, tributary of Rio Jiparana, 5 June 1984.

USNM 295126, 1 (37.5 mm SL), South America, Brazil, Rondonia, Rio Machado

System, 20 km Upstream of Ji-Parana.

Steindachnerina guentheri: USNM 302184, 10 (43.7 – 59.8 mm SL), South America,

Peru, Madre de Dios, Boca Panahua, where Quebrada Panahua joins the Rio Manu, 10

September 1988.

Steindachnerina hypostoma: USNM 261477, 4 (71.9 – 89.0 mm SL), South America,

Peru, Ucayali, Rio Ucayali, Pucallpa, 21 May 1974.

USNM 278579, 5 (54.6 – 59.9 mm SL), South America, Bolivia, Pando, Laguna San

Luis (Rio Madre De Dios), 18 May 1982.

Steindachnerina insculpta: USNM 295267, 3 (96.6 – 101.9 mm SL), South America,

Brazil, Minas Gerais, Repressa De Jaguara, Rio Parana Basin.

231

USNM 295273, 7 (54.1 – 96.6 mm SL), South America, Brazil, Brazilian Federal

District, Riberao Santana At Rd. Crossing About 30 Air km S Barragem Do Paranoa

(Bartolomew/Parana Syst.), S 15° 55' 12” W 47° 46' 12", 12 November 1984.

Steindachnerina leucisca: USNM 319306, 3 (134.3 – 139.7 mm SL), South America,

Peru, Madre de Dios, Picaflor Stream, Cana Brava Trail, Manu National Park, 23 April

1991.

USNM 94655, 3 (105.6 – 109.1 mm SL), South America, Brazil, Acre, Vicinity of

Mouth of Rio Macauhan, Tributary To Rio Yaco, Which Flows Into the Rio Purus, S

9° 19' 48" W 69° 0' 0", 1934.

USNM 261407, 2 (71.2 – 103.0 mm SL), South America, Peru, Ucayali, Rio Ucayali,

Pucallpa, 16 August 1973.

Steindachnerina notonota: USNM 220202, 10 (78.9 – 93.3 mm SL), South America,

Brazil, Ceara, Reservoir at Pentecoste, 28 July 1966.

Steindachnerina planiventris: USNM 293099, 10 (60.1 – 73.7 mm SL), South America,

Brazil, Amazonas, Rio Solimoes, above Coari, 25 January 1986.

Cynodontidae:

Cynodon gibbus: OS 12620, 1 (175.0 mm SL), South America, Peru, Napo River, 27

February 1979.

Roestes ogilviei: USNM 377412, 2 (115.0 – 118.1 mm SL), South America, Guyana,

Morai Creek, 28 November 2001.

232

Distichodontidae:

Distichodus decemmaculatus: CUMV 87581, 10 (32.8-41.3 mm SL), Africa, Republic of the Congo, Cuvette-Ouest, Lékoli River, Odzala National Park, N 0° 36' 51” E 14°

56' 11”, 8 August 2002.

Distichodus fasciolatus: CUMV 95835, 8 (46.8 – 97.0 mm SL), Africa, Democratic

Republic of the Congo, Orientale, Tshopo, Tshopo and Lindi Rivers between

SOTEXKI and Beach Lindi to 5 km upstream on the Lindi, N 0° 33' 47” E 25° 7' 12”,

24 March 2010.

USNM 229952, 1 (179.2 mm SL), Africa, Niger, Middle/Upper Niger At Niamey

(Republic of Niger), N 13° 41' 59” E 2° 0' 0", 18 July 1980 to 21 July 1980.

USNM 151078, 1 (180.1 mm SL), Africa, Fleuve River, Lvolaba, 25 June 1947.

Eugnathichthys macroterolepis: CUMV 87452, 4 (41.4-80.4 mm SL), Africa, Republic of the Congo, Cuvette-Ouest, small channel around island in Lékoli River, Odzala

National Park, N 0° 37' 12" W 14° 55' 6”, 16 August 2002.

CUMV 87453, 2 (49.9-58.1 mm SL), Africa, Republic of the Congo, Cuvette-Ouest,

Pandaka River, Odzala National Park, 12 August 2002.

Hemigrammocharax unioccelatus: USNM 310585, 2 (21.1-28.0 mm SL), Africa,

Central African Republic, Riviere Kotto at Kembe, 23 May 1971 to 25 May 1971.

Hemistichodus vaillanti: CUMV 96804, 1 (62.4 mm SL), Africa, Gabon, Ogooue-

Ivindo, Loa-Loa rapids, Ivindo River below Makokou; Ivindo River, N 0° 31' 18” E

12° 49' 28”, 10 May 2011.

233

Ichthyborus ornatus: CUMV 87568, 1 (125.0 mm SL), Africa, Republic of Congo,

Cuvette-Ouest, Lékoli River, Odzala National Park, N 0° 37' 16” E 14° 55' 5”, 10

August 2002.

CUMV 88545, 1 (113.9 mm SL), Africa, Republic of the Congo, Cuvette-Ouest, Lékoli

River, Odzala National Park, N 0° 37' 16” E 14° 55' 5”, 11 August 2002.

Mesoborus crocodilus: CUMV 88527, 2 (126 – 261.1 mm SL), Africa, Republic of the

Congo, Cuvette-Ouest, Lokoué River, Odzala National Park, N 0° 54' 20” E 15° 7'

21”, 27 August 2002.

Microstomatichthyoborus bashforddeani: CUMV 87589, 1 (50.9 mm SL), Africa,

Republic of the Congo, Cuvette-Ouest, Lékoli River, Odzala National Park, N 0° 37'

16” E 14° 55' 5”, 7 August 2002.

CUMV 88562, 1 (59.2 mm SL), Africa, Republic of the Congo, Cuvette-Ouest, Lékoli

River, Odzala National Park, N 0° 37' 16” E 14° 55' 5”, 9 August 2002.

CUMV 87591, 2 (53.1 – 55.2 mm SL), Africa, Republic of the Congo, Cuvette-Ouest,

Lékoli River, Odzala National Park, N 0° 36' 51” E 14° 56' 11”, 8 August 2002.

Nannaethiops unitaeniatus: USNM 303954, 10 (22 - 29.5 mm SL), Africa, Cameroon,

Southwest Privince, Bapuo River junction with Cross, downstream of Mamfe, N 5° 55'

30” E 9° 9' 0", 20 February 1988.

Neolebias unifasciatus: USNM 310843, 10 (24.0- 24.8 mm SL), Africa, Togo,

Atakpame, 25 August 1968.

Phago boulengeri: CUMV 91510, 10 (52.8 – 71.0 mm SL), Africa, Central African

Republic, Basse-Kotto, Oubangui River downriver from Mobaye, N 4° 19' 30” E 21°

8' 43”, 28 February 2006.

234

Xenocharax spilurus: USNM 227694, 1 (151.2 mm SL), Africa, Cameroon, Mang,

River Bumba, 20 February 1975.

USNM 227693, 1 (122.8 mm SL), Africa, Gabon, Allonha Ii, Lac Ezanga, 14

November 1964.

USNM 227692, 1 (40.4 mm SL), Africa, Democratic Republic of the Congo, Boende,

Rw. Tshuapa, February 1956 to March 1956.

USNM 227696, 2 (64.1 – 76.7 mm SL), Africa, Democratic Republic of the Congo,

Lac Tumba, 29 September 1955 to 30 September 1955.

Erythrinidae:

Hoplerythrinus unitaenistus: USNM 199237, 6 (137.7 – 156.4 mm SL), South

America, Brazil. Mato Grosso, Upper Yuruena, 1 August 1962.

Hoplias macrophthalmus: USNM 225209, 10 (61.4 – 87.1 mm SL), South America,

Suriname, Nickerie District, tributary to Sisa Creek, north side approximately 700 m downstream of crossing of Amotopo to Camp Geologie Road, N 3° 42' 0" W 57° 42'

0", 20 September 1980.

Hoplias malabaricus: OS 18349, 1, South America, Peru, Loreto, Maynas, Two small caños in the vicinity of main bridge on Itaya River, S 4° 13' 32” W 73° 29' 5”, 3 August

2010.

OS 18368, 1, South America, Peru, Loreto, Maynas, Mussel beds on Itaya River about

1 hour upstream from Belén (village of Iquitos) by motorized canoe, S 3° 49' 59” W

73° 18' 6”, 2 August 2010.

235

Gasteropelecidae:

Carnegiella marthae: USNM 270117, 10 (20.0 – 22.5 mm SL), South America,

Venezuela, Amazonas, Small Cano off Cano Urami, just upriver of Santa Lucia, N 1°

16' 48" W 66° 50' 60”, 6 December 1984.

Gasteropelecus maculatus: USNM 121423, 5 (38.6 – 42.7 mm SL), South America,

Venezuela, Zulia, Rio Machango at bridge South of Lagunillas, 16 March 1942.

Gasteropelecus sternicla: USNM 305473, 9 (31.5 – 35.8 mm SL), South America,

Bolivia, Beni, Borrow Pit by road at 2.5 km W of Rio Matos crossing 44 Air km E San

Borja, S 14° 55' 12” W 66° 16' 48", 24 August 1987.

Thoracocharax stellatus: USNM 232888, 10 (29.6 – 31.6 mm SL), South America,

Venezuela, Delta Amacuro, Rio Orinoco small side Cano, Cano Araguao, 112 mi. upstream from sea buoy, N 8° 37' 48" W 61° 43' 12”, 21 February 1978.

Hemiodontidae:

Anodus orinocensis: USNM 268563, 1 (183.54 mm SL), South America, Venezuela,

Monagas, Secondary cano about 500 m from its mouth in Cano Guarguapo, 146 nautical miles from sea bouy, N 8° 39' 24” W 62° 13' 60”, 11 November 1979.

OS 5039, 2 (195.0 – 196.0 mm SL), South America, Peru, Pucapalla City, located on

Yacayali River; Lake is an old river channel, flood area, S 8° 30' 0" W 74° 30' 0", 15

July 1972.

USNM 233398, 1 (120.4 mm SL), South America, Venezuela, Monagas, Cano

Between Rio Orinoco and Laguna Guatero, Near Barrancas, 142 Nm From Sea Bouy,

N 8° 43' 12" W 62° 10' 48”, 12 November 1979.

236

USNM 233399, 1 (114.9 mm SL), South America, Venezuela, Monagas, Cano

Between Rio Orinoco and Laguna Guatero, Near Barrancas, 142 Nm From Sea Bouy,

N 8° 43' 12" W 62° 10' 48”, 12 November 1979.

USNM 310746, 2 (221.9 – 242.7 mm SL), South America, Brazil, Amazonas, Rio

Solimoes, Between Manaus and Tefe, March 1974.

Argonectes loncipes: USNM 270174, 6 (54.7 – 64.9 mm SL), South America,

Venezuela, Amazonas, Rio Cataniapo where crossed by road from Puerto Ayacucho to

Samariapo, N 5° 37' 12" W 67° 34' 12”, 2 December 1984.

Bivibranchia velox: USNM 268490, 10 (36.1 – 55.2 mm SL), South America, Brazil,

Para, Rio Xingu, Belo Monte, 25 September 1983.

Hemiodopsis quadrimaculatus: USNM 376810, 2 (34.0 – 43.5 mm SL), South

America, Guyana, Coco Creek, 4 November 2001.

USNM 233092, 3 (27.4 – 39.5 mm SL), South America, Venezuela, Bolivar, Rio

Orinoco, Pool, Islote Fajardo, 182 Naut. mi. Upstream From Sea Buoy, N 8° 22' 12”

W 62° 42' 0", 15 February 1978.

Hemiodus gracilis: OS 18758, South America, Peru, Loreto, Maynas, Dead arm and shore of Nanay River near mouth of Caño Shirui near Pampa Chica, S 3° 45' 6” W 73°

17' 13”, 9 August 2010.

Hemiodus immaculatus: USNM 257529, 3 (98.9 – 126.3 mm SL), South America,

Venezuela, Guarico, Guariqiotp River at government reserve, ESE of Calabozo, 27

January 1983.

237

USNM 233610, 6 (86.8 – 105.0 mm SL), South America, Venezuela, Monagas, Laguna

Guatero, East of Barrancas, 143 Nm From Sea Bouy, N 8° 41' 42" W 62° 11' 12”, 12

November 1979.

Hemiodus semitaeniatus: USNM 225598, 10 (57.3 – 71.9 mm SL), South America,

Suriname, Nickerie District, Kamp Kreek 100 m north of turnoff to Camp Geology, N

4° 49' 12" W 57° 28' 12”, 13 September 1980.

Hemiodus unimaculatus: USNM 226162, 10 (117.0 – 169.5 mm SL), South America,

Suriname, Nickerie District, Corantijn River, N 5° 0' 0", W 57° 16' 48", 17 May 1980.

Hepsetidae:

Hepsetus odoe: USNM 304096, 2 (45.1 – 105.8 mm SL), Africa, Cameroon, Southwest

Province, Cross System: Collecting Points Upper Tributaries of Manaya, Nr Baro

Village, North Korup; Upper Reaches of Marube R., Draining To Bake R. (By 'Salt

Lick'), N 5° 16' 30" E 9° 10' 59”, 2 March 1988.

USNM 303893, 3 (56.0 – 89.0 mm SL), Africa, Cameroon, Southwest Province,

Manyu, Cross System: Collecting Points On Main Cross River Downstream of Mamfe;

Akinyam R. Junction With Cross, N 5° 50' 30”, E 9° 14' 49”, 17 February 1988.

Iguanodectidae:

Iguanodectes spilurus: USNM 257557, 10 (53.7 – 61.4 mm SL), South America,

Venezuela, Guarico State, Guariquito River At Government Reserve, E-SE of

Calabozo; Several Points Along River, N 8° 34' 48" W 67° 15' 0", 27 January 1983.

Lesbiasinidae:

Copeina guttata: USNM 177857, 10 (48.1 – 57.9 mm SL), South America, Brazil, Rio

Guapore, October 1958.

238

Copella nattereri: USNM 269900, 9 (21.8 – 24.6 mm SL), South America, Venezuela,

Amazonas, Cano Manu Tributary of Casiquiare Canal approximately 250 m upstream of Solano, N 2° 0' 0" W 66° 57' 0", 7 December 1984.

Lesbiasina bimaculata: USNM 83630, 10 (42.6 – 66.1 mm SL), South America, Peru,

Cultambo.

Nannostomus marginatus: USNM 272417, 10 (21.0 – 21.9 mm SL), South America,

Venezuela, Amazonas, Cano Loro where crossed by road from San Carlos De Rio

Negro to Solano, N 1° 58' 48” W 66° 58' 12”, 7 December 1984.

Piabucina elongata: USNM 258066, 6 (44.8 – 89.6 mm SL), South America, Ecuador,

Napo, Rio Quijos and isolated backwater about 7.8 km northeast of bridge over Rio

Salador, S 0° 8' 17” W 77° 37' 18”, 21 November 1981.

Piabucina festae: USNM 78646, 6 (74.3 – 95.1 mm SL), Central America, Panama,

Darien, Rio Aruza, 27 February 1912.

Piabucina striagata: USNM 398314, 4 (117.0 – 137.6 mm SL), South America,

Ecuador, small eastern tributary to Rio Huimbo, approximately 3 km upriver from

Concepcion and 2 km SSE of Rocafuerte, N 1° 1' 0” W 8° 49' 3”, 15 September 1993.

USNM 398311, 4 (73.1 – 88.2 mm SL), No data.

Parodontidae:

Apareiodon affinis: USNM 313193, 10 (58.3 – 77.2 mm SL), South America, Brazil,

Sao Paulo, Rio Cubatao of Rio Grande of Upper Rio Parana, Near Cajuru. Fazenda

Santa Carlota (Rio Pardo Basin), S 21° 18' 0" W 47° 16' 48", 7 November 1989.

239

Apareiodon ibitiensis: USNM 313195, 10 (53.7 – 85.2 mm SL), South America, Brazil,

Sao Paulo, Rio Cubatao of Rio Grande of Upper Rio Parana, Near Cajuru. Fazenda

Santa Carlota (Rio Pardo Basin), S 21° 18' 0" W 47° 16' 48", 7 November 1989.

Parodon caliensis: USNM 79226, 2 (61.5 – 64.7 mm SL), South America, Colombia,

Cali.

USNM 120160, 1 (50.8 mm SL), South America, Colombia, Rio Cali At Cali, Upper

Cauca, 6 October 1942.

Parodon guyanensis: USNM 409887, 2 (51.6 – 55.6 mm SL), South America,

Suriname, Tapaje Creek, large left tributary of Middle Paloemeu River (stream width

100 m). Tapaje Creek almost of the same size as the Middle Paloemeu River at point of confluence, N 2° 44' 45” W 55° 26' 25”, 19 March 2012.

USNM 409778, 3 (52.7 – 59.2 mm SL), South America, Suriname, Downstream waterfall in right tributary of Upper Paloemeu River, N 2° 27' 21” W 55° 37' 34”, 11

March 2012.

Parodon hilarii: USNM 120189, 5 (62.1 – 88.7 mm SL), South America, Bolivia,

Tumpusa, December 1921.

USNM 120191, 3 (53.5 – 88.7 mm SL), South America, Bolivia, Rio Colorado, Lower

Bopi, September 1921.

Parodon moreirai: USNM 326342, 2 (88.6 – 95.1 mm SL), South America, Brazil,

Matto Grosso, Arroio Afluente Do Rio Do Bugre (Afl. Rio Jauru) No km 165 Da

Estrada Porto Esperidiao/Pontes E Lacerda (Br 174), ca. 48 km De Porto Esperidiao

(Sistema Do Rio Paraguai), Porto Esperidiao, 13 August 1991.

240

USNM 319282, 8 (40.6 – 79.3 mm SL), South America, Bolivia, Chuquisaca, Rio

Camatindi 8 km N Border Dept. Tarija, at 40 km N Villamontes, S 20° 55' 12" W 63°

23' 59”, 1 October 1988.

Parodon nasus: USNM 313189, 7 (88.3 – 108.2 mm SL), South America, Brazil, Sao

Paulo, Rio Cubatao of Rio Grande of Upper Rio Parana, Near Cajuru. Fazenda Santa

Carlota (Rio Pardo Basin), S 21° 18' 0" W 47° 16' 48", 7 November 1989.

USNM 302513, 3 (78.7 – 103.7 mm SL), South America, Brazil, Sao Paulo, Near Santa

Rosa De Viterbo, Barragem De Itaipava, Usina Amalia; Pardo River Drainage, Rio

Pardo, Main Channel of River; Fish Caught With Cast Nets Just Below Fish Ladder, S

21° 25' 12" W 47° 19' 47”, 25 October to 14 November 1984.

Parodon pongoense: USNM 311305, 3 (67.0 – 81.7 mm SL), South America, Ecuador,

Napo, Estero Triniti a 45 Min. De Rocafuerte Margen Izquierdo Del Rio Yasuni, 29

September 1988.

Prochilodontidae:

Ichthyoelephas humeralis: FMNH 80715, 3 (107.6 – 130.6 mm SL), South America,

Ecuador, Guayas basin, Rio Palenque at Centro Cientifico Rio Palenque, isolated pool,

75' long, 13 July 1975.

FMNH 80720, 8, South America, Ecuador, Guayas basin, Rio Palenque at Centro

Cientifico Rio Palenque, isolated pool, 75' long, 13 July 1975.

Prochilodus argenteus: USNM 398299, 6 (91.3-160.0 mm SL), South America, Brazil,

Pernambuco, Rio Sao Francisco near Coripos, 1952.

241

Prochilodus brevis: USNM 319768, 9 (99.0 -112.8 mm SL), South America, Brazil,

Bahia, Paraguacu Drainage near the Town of Iacu, S 12° 45' 0" W 40° 12' 0", 25 July

1988.

Prochilodus costatus: USNM 357391, 1 (76.0 mm SL), South America, Brazil, Minas

Gerais, Riacho Afluente Do Rio Jequitai, 20 July 1994.

FMNH 76359, 1, South America, Brazil, Piracicaba, 23 July 1908.

FMNH 76361, 1, South America, Brazil, Sao Paulo, Salto Avanhandava, 15 September

1908.

FMNH 78097, 1, South America, Brazil, Cachoeira de Pirapora, Rio Sao Francisco, 15

December 1907.

Prochilodus hartii: USNM 318135, 2 (93.5-101.0 mm SL), South America, Brazil,

Minas Gerais, Tributary of Rio Jequitinhonha known as Rio Ribeirao approximately 4-

5 km ESE of Town of Jordania, S 15° 54' 0" W 40° 10' 12", 12 July 1991.

USNM 318129, 3 (86.1 – 91.0 mm SL),

Prochilodus lineatus: USNM 181717, 4 (133.4-180.2 mm SL), South America,

Paraguay, Asuncion Bay, Rio Paraguay near Asuncion, 10 January 1957.

Prochilodus magdalenae: FMNH 56337, 3, South America, Colombia, Paila, 25

February 1912.

Prochilodus mariae: FMNH 103683, 7 (105.4 – 110.7 mm SL), South America,

Venezuela, Guarico, Lagoon ca. 4 km from east side of Rio Aguaro north of Cabruta,

N 7° 50' 28” W 66° 30' 28”, 6 February 1992.

242

Prochilodus nigricans: USNM 280436, 6 (97.3-142.4 mm SL), South America, Peru,

Loreto, Rio Itaya, main river channel and lower portions of Canos, 5 to 20 km upstream of Iquitos, S 3° 51' 0" W 73° 12' 0", 20 August 1986.

Prochilodus reticulatus: USNM 121333, 8 (104.8-127.1 mm SL), South America,

Venezuela, Venez, Rio Negro below mouth of Rio Yasa, 75 km S. Of Rosario, 2 March

1942.

Prochilodus rubrotaeniatus: OS 18632, South America, Guyana, Cuyuni-Mazaruni,

Immediately upstream from Aurora Mining Camp at eastern tip of main island in channel of Cuyuni River, N 6° 48' 35 W 59° 47' 55”, 2 February 2011.

Prochilodus vimboides: FMNH 76364, 1 No data.

FMNH 78096, 1, South America, Brazil, Entre Rios, Rio Parahyba. Cool, clear, shallow creek, 2 miles below town, 2 June 1908.

FMNH 78099, 1, South America, Brazil, Entre Rios, Rio Parahyba. Cool, clear, shallow creek, 2 miles below town, 2 June 1908.

FMNH 92303, 1, South America, Brazil, Rio Doca, Fumaca, whirlpool 2 mi. below the village of Rio Doce, 27 May 1908.

FMNH 78088, 1, South America, Brazil, Rio San Francisco, Penedo, 22 March 1908.

Semaprochilodus brama: USNM 191636, 1 (64.7 mm SL), South America, Brazil, Rio

Araguaia near Aruana, S 14° 58' 12" W 51° 23' 60”, 1960.

Semaprochilodus insignis: USNM 175856, 1 (178.9 mm SL), South America, Peru,

Apayacu River, 9 October 1935.

USNM 290149, 3 (63.0 – 68.4 mm SL), South America, Brazil, Amazonas, Furo Entre

Lago Murumuru E Parana De Janauaca, 17 August 1977.

243

USNM 175855, 6 (78.1 – 109.1 mm SL), South America, Peru, Tuye Cano, 2

September 1935.

Semaprochilodus kneri: USNM 257560, 5 (140.1-223.0 mm SL), South America,

Venezuela, Guarico, Guariquito River at government reserve, ESE of Calabozo, N 8°

34' 48" W 67° 15' 0", 27 January 1983.

Semaprochilodus laticeps: USNM 292341, 4 (79.5-98.1 mm SL), South America,

Venezuela, Monagas, Rio Orinoco, Cano Guarguapo, in small cano about 500 m from mouth, N 8° 39' 30” W 62° 13' 48”, 11 November 1979.

USNM 270239, 5 (72.6 – 85.3 mm SL), South America, Venezuela, Bolivar, Small

Cano Connecting With Rio Orinoco Immediately South of El Burro, N 6° 10' 48” W

67° 25' 12", 9 December 1984.

Semaprochilodus taeniurus: USNM 52545, 2 (210.0-240.2 mm SL), South America,

Brazil, Paro to Manaos, Amazon River, 1901.

USNM 162816, 4 (71.8 – 123.2 mm SL), Brazil, Solimoes Manacapuru, R. Amazon,

28 October 1925.

Serrasalmidae:

Catoprion mento: USNM 308825, 2 (54.5 – 78.6 mm SL), South America, Brazil,

Amazonas, Lago Janauari, Pequena Casa Quase Frente Da Olaria, 5 January 1978.

USNM 179562, 10 (33.8 – 77.7 mm SL), South America, Brazil, Rio Urubu 25 mi.

From Itacoatiara.

Colossoma macropomum: USNM 258173, 1 (184.6 mm SL), South America,

Venezuela, Apure, side channel of Rio Apure ca. 5 km west of San Fernando de Apure,

N 7° 52' 47” W 67° 28' 48", 21 January 1983.

244

USNM 308709, 1 (115.6 mm SL), South America, Brazil, Amazonas, Near Manaus,

Camaleao, Ilha De Marchantaria, 26 September 1978.

USNM 310455, 3 (44.5 – 64.7 mm SL), South America, Brazil, Muddy Lgarape

Connecting Solimoes With Blackwater Lago ca. 15 mi. W. of Coari, 7 Mach 1974.

USNM 229145, 2 (49.3 – 51.7 mm SL), South America, Brazil, Amazonas, Paraana

Da Ilha De Marchantaria, 25 April 1978.

Metynnis lippincottianus: OS 12590, 1 (75.0 mm SL), South America, Peru, Napo

River, 27 February 1979.

OS 12591, 3 (64.0 – 87.0 mm SL), South America, Peru, Napo River Corha Huanaiua,

27 February 1979.

Myleus torquatus: USNM 308714, 10 (39.8 – 50.6 mm SL), South America, Brazil,

Amazonas, Ressaca Da Ilha De Marchantaria, 16 August 1977.

Pristobrycon striolatus: USNM 409892, 1 (96.2 mm SL), South America, Suriname,

Tapaje Creek, large left tribuatary of Middle Paloemeu River, N 2° 44' 45” W 55° 26'

25”, 19 March 2012.

USNM 409902, 3 (103.2 – 113.7 mm SL),

Pygocentrus nattereri: USNM 309028, 10 (40.6 – 57.2 mm SL), South America,

Brazil, Amazonas, Ressaca Da Ilha De Marchantaria, 15 March 1977.

Serrasalmus humeralis: USNM 345649, 1 (33.4 mm SL), South America, Guyana,

Berbice River at Dubulay Ranch, N 5° 48' 56” W 57° 51' 29”, 25 July 1996.

USNM 164020, 2 (64.3 – 155.4 mm SL), South America, Ecuador, Napo-Pastaza, Near

Mouth of Rio Bobonaza, S 2° 22' 47” W 76° 39' 0", January 1954.

245

Serrasalmus rhombeus: USNM 227339, 10 (70.9 – 137.9 mm SL), South America,

Suriname, Nickerie District, Toeboeroe Creek, tributary of Corantijn River at km 220,

300 m from mouth to 900 m from mouth, 8 December 1979.

Triporthidae:

Agoniates anchovia: USNM 261505, 1 (74.01 mm SL), South America, Peru, Loreto,

Boca Del Rio Yavari, 2 October 1982.

USNM 229089, 2 (66.1 – 69.0 mm SL), South America, Brazil, Amazonas, Parana of

Lago Janauaca, Entrance of Lago Do Castanho, 15 February 1978.

Clupeacharax anchoveoides: USNM 243223, 1 (55.8 mm SL), South America, Peru,

Pucallpa, Rio San Alejandro, 26 May 1976.

USNM 319321, 2 (52.9 – 53.5 mm SL), South America, Peru, Madre de Dios Region,

Manu, Manu National Park, Pakitza, Manu River 20' down river from Pakitza, close to

Fortaleza mouth, 9 May 1991.

USNM 317779, 6 (27.9 – 33.4 mm SL), South America, Peru, Madre de Dios Region,

Manu, Manu National Park, Pakitza, Manu River, 3 November 1990.

Engraulisoma taeniatum: USNM 326919, 2 (28.2 – 33.9 mm SL), South America,

Peru, Madre de Dios, Pakitza, Manu’s Beach Close to Pachija’s mouth, 9 Otober 1991.

USNM 349391, 4 (19.0 – 21.5 mm SL), South America, Venezuela, Portuguesa, Rio

Portuguesa, Just Upstream of Hwy 5, 11 km WNW of Guanare, N 9° 4' 12" W 69° 39'

0", 28 February 1988.

USNM 361425, 3 (24.4 – 26.0 mm SL), South America, Peru, Uc, Atalaya, Sepahua,

Nuevo Horizonte, Rio Bajo Urubamba, 3 November 1998.

246

Lignobrycon myersi: USNM 304497, 4 (68.4 – 78.8 mm SL), South America, Brazil,

Bahia, Rio Do Braco, 2 km SW of Town of Rio Do Braco, On Fazenda Santa Luzia, 2

February 1989 to 3 February 1989.

Triportheus nematurus: USNM 325695, 1 (107.6 mm SL), South America, Brazil, R.

Paraguay, Corumba, 19 January 1993.

USNM 229448, 1 (86.8 mm SL), South America, Paraguay, Presidente Hayes

Department, a Pond By the Trans Chaco Highway, 194 km North of Asuncion, 28 June

1981.

USNM 181691, 5 (61.9 – 82.0 mm SL), South America, Paraguay, Asuncion Bay, Rio

Paraguay near Asuncion, 15 January 1957.