Phylogenetic and Biogeographic Analyses of Greater Antillean and Middle American Cichlidae

PHYLOGENETIC AND BIOGEOGRAPHIC ANALYSES OF GREATER ANTILLEAN AND MIDDLE AMERICAN CICHLIDAE

Prosanta Chakrabarty

A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Ecology and Evolutionary Biology) in The University of Michigan 2006

Doctoral Committee:

Professor William L. Fink, Chair Professor Daniel C. Fisher Professor Gerald R. Smith Associate Professor Diarmaid Ó Foighil

I dedicate this dissertation to my wife, Annemarie Noël, and to my parents, Chitta and Anurupa Chakrabarty.

ii ACKNOWLEDGEMENTS

Many people have helped me complete this work. Foremost, I would like to thank my advisor, William L. Fink, for all his hard work teaching me to be a proper scientist. I walked into his office in September 2001 an ambitious student, but also naïve and hasty. In my first year particularly I sometimes bit off more than I could chew, which could have led to a few major setbacks. Instead, Bill was there to save me at times when I was lost; many times he has pointed me in the right direction when I was struggling. I hope I can always be as determined, rigorous and hardworking as he has taught me to be.

My other committee members, Jerry Smith, Diarmaid Ó Foighil, and Dan Fisher have been excellent. Jerry and his wife Catherine are not only advisors but friends. It is wonderful to talk to them about everything from fossils to vegetarianism. Diarmaid has always given me straight forward advice about publishing and academia that has always been very useful.

My early success as a graduate student has a lot to do with the mentoring I received from then senior graduate students John Sparks, Dan Graf, and Jennifer Ast.

Without their encouragement and teaching I would not have been brave enough to try many of things that helped me be successful. Arnold Kluge has been my cladistics guru, and he has taught me a tremendous amount about the philosophy of science for which I

iii will always be thankful.

I would like to thank the staff at the UMMZ (Norah Daugherty, Beverley Dole,

Vlad Miskevich), EEB Graduate Office (Julia Eussen, Barb Klumpp, Connie Rockman,

Susan Stark, Kaye Hill) and also Sheila Dunn, Jill Beeson, Chris Psujek, LaDonna

Walker and David Bay. All these people made the classes I’ve taught and various EEB

events I helped organize go smoothly. They also made my graduate life easier and more

enjoyable.

I must thank those responsible for the Carl L. and Laura C. Hubbs endowment

particularly Clark and Cathy Hubbs. I must thank the anonymous donors behind the Fish

Division Graduate Student Fund, the donors behind Rackham grants, and the Hinsdale

and Okkelberg EEB grants. Without these funding sources my research would not have

been possible.

Fish Division graduate students Heok Hee Ng and Ron Oldfield have become

great friends and I hope to work with them for the rest of my career. Doug Nelson

enriched my days during my time as his research assistant and throughout my graduate

studies with his knowledge of fishes and his wonderful stories.

I would probably not be in science without the mentoring I received from Melanie

Stiassny at the American Museum of Natural History and Robert Carroll at McGill

University. Their encouragement and the advice I received from Scott Schaefer, Leo

Smith and Bob Schelly led me to Michigan, and I will forever be indebted to them for

that.

I must thank my fellow EEB graduate students, past and present, who I count among my closest friends: Wendy Grus, Elen Oneal, Cori Richards, Rick Lehtinen, Tim

iv Connallon, Glenn Fox, Tanya Dewey, Josh Rest, Ashley Dowling, Matt Wund, Mara

Zimmerman, Heather Lerner, Laura Howard, Adam Ehmer, Xiaoxia Wang, Lucia Luna

Wong, Tamara Convertino-Waage, Ondrej Podlaha, Heather Adams, Zach Miller,

Andrea Walther and Joseph Brown. Also the many faculty members and postdocs I have come to know and have learned so much from Taehwan Lee, Paul Dunlap, Deborah

Goldberg, Lacey Knowles, Jo Kurdziel, Bryan Carsten, Jeff Wilson, David Mindell,

Miriam Zelditch, Marc Ammerlaan, Tom Duda, Doug Futuyma, Phil Myers, Bob Payne,

Jianzhi Zhang, Beverly Rathcke, Y-Qiu, Priscilla Tucker, John Vandermeer and Earl

Werner.

My dissertation would not have been possible without the collections I made in

the Dominican Republic with Carlos Rodríguez, in Mexico with Juan Jacobo Schmitter-

Soto and in Belize with Peter Esselman. Although I never met Bob Miller his work and

collections of cichlids are an inspiration. His work has also been extremely useful to the

work I present here. Without Bob Miller’s work, my work would not be meaningful.

I have dedicated this work to my wife and my parents, but I must also thank them.

Their love and encouragement have allowed my dreams to come true. I’ve never wanted

to be anything else besides a zoologist and they did nothing but support me getting to that goal.

v Earlier versions of some chapters were published as:

Chakrabarty, P. (2004) Cichlid biogeography: comment and review. Fish and Fisheries 5(2): 97-119.

Chakrabarty, P. (2006) Taxonomic status of the Hispaniolan Cichlidae. Occasional Papers of the Museum of Zoology, University of Michigan 737, 1-16.

Chakrabarty, P. (2006) Systematics and Historical Biogeography of Greater Antillean Cichlidae. Molecular Phylogenetics and Evolution 39, 619-627.

vi TABLE OF CONTENTS

Dedication…………………………………………………………………………………ii Acknowledgments………………………………………………………………………..iii List of Figures………………………………………………………………………...…x List of Tables…………………………………………………………………………...xv Abstract………………………………………………...………………………………xvi

Chapter

I. CICHLID BIOGEOGRAPHY:UPDATED COMMENT AND REVIEW...1 Abstract…………………………………...………………....…….1 Introduction………………………………………………………..1 Cichlid Biogeography: Overview of the Debate…………..2 Selecting Among Biogeographic Hypotheses…………….3 Alternative Plate Tectonic Reconstructions ………………6 Proposed Cichlid Relationships…………………………...8 Phylogenies Congruent With Vicariance Scenarios……....8 Phylogenies That Falsify Vicariance Scenarios.……….....9 Molecular Clocks……………………………………..….21 Indian, Malagasy Phylogenetic Relationships………………...... 23 Geological History……………………………………….23 Vicariance………………………………………………..24 The Greater Antilles……………………………………..………25 Greater Antillean Cichlids…………..…………………..25 Phylogenetic Relationships……………………………..26 The Middle East, Europe and Adjacent Areas…………………...29 Current Distribution and Sister Relationships………….29 Fossils From the Area………………………………….30 The Global Cichlid Fossil Record………………………………..32 Minimum Ages……..………………………………….32 The Acanthomorph Record…………………………….33 Discussion……………………………………………………...... 35 Literature Cited…………………………….………...…………..42

vii

II. TAXONOMIC STATUS OF THE HISPANIOLAN CICHLIDAE……...51 Abstract……………………………….……...…………....……..51 Introduction………………………………………………………51 Materials and Methods……………………...……………………53 Institutional Abbreviations……………………………55 Systematic Accounts……………………………………………..57 Nandopsis haitiensis……...…………………………..57 Nandopsis woodringi…………………………………68 Additional Materials Examined………………………………….73 Literature Cited…..………………………………………………74

III. SYSTEMATICS AND HISTORICAL BIOGEOGRAPHY OF GREATER ANTILLEAN CICHLIDAE…………………………………………….…..76 Abstract……………………………….……...…………....……..76 Introduction………………………………………………………77 Materials and Methods…………………………………………...79 Acquisition of DNA Dataset….…………….……………79 Phylogenetic Analyses and Support Indices…..…………81 Date Estimation and Calibration…………………………83 Results……………..……………………………… ………….…86 Model Selection, Likelihood Assumption Set………..….86 Phylogenetic Analyses and Support………….…...……..86 Estimated Dates………………………………………….90 Discussion…………………...……………………….…….….....92 Literature Cited…………………………………………….....….97

IV. A MORPHOLOGICAL PHYLOGENETIC ANALYSIS OF NEOTROPICAL CICHLIDS IN THE SECTION ‘NANDOPSIS’ SENSU REGAN……………………………....………………………………….…103 Abstract……………………………….……...…………....……103 Introduction………………………………………………..……103 Materials and Methods………………………………………….107 Results……………..……………………………… …………...108 1- Internal Morphology Oral Region (A) Upper and Lower Jaw…...... 110 (B) Oral Teeth…...... 117 Skull (C) Neurocranium……..…..….....120

viii Branchiocranium (D) Upper Pharyngeal Jaw……….126 (E) Lower Pharyngeal Toothplate.128 (F) Soft Tissue Elements……...... 131 Post Cranial Elements (H) Vertebrae…………………….130 (I) Caudal Skeleton…………..….135 (J) Meristics………….……...... 137 2 – External Features (A) Color and Pigmentation…….139 (B) Scales……………………….142 (C) Soft Tissue …………………146 (D) Fins…...…………………….150 (E) Sensory Pores…………….…154 (F) Morphometrics……………....157 Discussion…………………...……………………….…….…...161 Materials Examined………………………………………….…186 Literature Cited……………………………………………....…194

V. NUCLEAR, MITOCHONDRIAL, AND MORPHOLOGICAL, COMBINED PHYLOGENETIC ANALYSES OF MIDDLE AMERICAN CICHLIDAE………………………………………………………………197 Abstract……………………………….……...…………....……197 Introduction…………………………………………………..…198 Materials and Methods………………………………………….200 Results……………..……………………………… …………...201 Discussion…………………...……………………….…….…...211 Biogeography of the Greater Antilles…………………215 Taxonomy……………………………………………..219 Materials Examined…………………………………………….224 Literature Cited……...…………………………………….....…232

ix LIST OF FIGURES

Figure

I.1. Cichlid worldwide distribution in black, redrawn from Sparks (2001)..……..2

I.2. Congruent cladograms for three hypothetical taxa, divided into six by allopatric speciation………………………………………………...……..4

I.3. A demonstration of how an area cladogram can imply a sequence of geological fragmentation that is incongruent with the actual sequence of fragmentation……………………………………………………………...5

I.4. Early Cretaceous tectonic reconstructions of (a) Hay et al. 1999 and (b) ‘classical’ reconstruction. Both images from Hay et al. 1999………..7

I.5. The diversity of phylogenetic trees shown as area cladograms that support vicariance hypotheses of cichlid biogeography………………….………12

I.6. The diversity of phylogenetic trees shown as area cladograms that falsify cichlid vicariance scenarios by requiring marine dispersal……………...16

I.7. Examples of (A) Monophyletic cichlid groups on Gondwanan fragments (B) Paraphyletic cichlid groups on Gondwanan fragments……………...20

I. 8: Map of Caribbean region and adjacent areas……………………………….28

I. 9: Map of European, Middle Eastern and African areas discussed……………31

I. 10: “Romerogram” showing diversity of groups over the geological time scale (from Patterson 1994)…………………………………………………....34

I. 11: The break-up of Gondwana (Australia and Antarctica not pictured) with biogeographic history of cichlids according to area cladogram from Fig.I.5E…………………………………………………………………..40

I.12: (A) Map of worldwide distribution of Aplocheiloidei (b) Area cladogram from phylogenetic hypothesis of Murphy and Collier 1997 using three mitochondrial genes and parsimony…………….……………..………...41

II.1: Landmarks used in Principal Components Analysis……………………....53

II.2: Nandopsis vombergi, holotype, ZMH 401, 181.7 mm SL………………....55

II.3: Nandopsis vombergi, holotype, frontal view to show expansion of lips...... 56

II.4: Nandopsis vombergi, holotype, view of right side of caudal fin to show diagnostic caudal spot that is divided equally by lateral line…………….56

II.5: Nandopsis haitiensis, holotype, USNM170907, 104.5 mm SL……...... 58

II.6: Nandopsis haitiensis, USNM 122635, 111.5 mm SL, male with nuchal hump………………………………………………………………….….60

II.7: Nandopsis haitiensis, USNM 87360, 114.6 mm SL, showing expansion of lips………………………………………………………………………..61

II.8: Nandopsis haitiensis UMMZ 243241, 173.2 mm SL, caudal fin…………..61

II.9: Nandopsis tetracanthus, AMNH 96390; 133.6 mm SL……………...…….62

II.10: Principal Component Analysis, PC 1 vs. PC 2…...……………………….62

II.11: Nandopsis woodringi USNM 10766, 64 mm SL approx………………...69

II.12: Illustration of USNM 10766……………..………………………………..69

III.1: Phylogeny of Neotropical cichlid taxa inferred from S7, Tmo-4C4, 16S and COI sequences…………………………………………………………...88

III.2: Maximum likelihood phylogeny of Neotropical cichlid taxa inferred from S7, Tmo-4C4, 16S and COI sequences…………………………………..91

IV.1: Consenus phylogeny of four most parsimonious trees…………………..109

IV.2: Character 1, Symphysial extension of alveolar process of premaxilla – premaxilla, lateral view…….………………………….………………..110

IV.3 Character 2, Degree of ventral folding of the ascending process of premaxilla - premaxilla, lateral view...……..……………...…………...111

IV.4 Character 3, Foramen in ascending process of premaxilla - premaxilla, lateral view……………………………….…………………...………...111

xi IV.5 Character 4, Shape of posterior end of alveolar process of premaxilla - premaxilla, lateral view………………………………………………...112

IV.6 Character 5, Mental prominence on lower jaw – lower jaw, lateral view……………...... 113

IV.7 Character 6, Shape of lateral face of retroarticular, - lower jaw, lateral view……………………………………………….114

IV.8 Character 7, Retroarticular and process of anguloarticular, on same vertical plane - lower jaw, lateral view……………...…………………………..114

IV.9 Character 8, Angle of process of anguloarticular, - lower jaw, lateral view…………………..…………………………….115

IV.10 Character 9, Relative height of process of anguloarticular, - lower jaw, lateral view………………..…………………………….....116

IV.11 Character 16, Tooth rows on lower jaw cross vertical through anterior ramus of anguloarticular - lower jaw, lateral view…………………..…119

IV.12 Character 18, Ethmo-vomerine block of neurocranium when viewed ventrally - ventral view of neurocranium, scale 1cm……….……..……120

IV.13 Character 19, Articulating surface of palatine - lateral view of left palatine……………………………………………121

IV.14 Character 20, Shape of pharyngeal apophysis (basiooccipital) - ventral view of posterior end of neurocranium, scale 1cm……………………..122

IV.15 Character 21, Posterior ventral expansion on parasphenoid - lateral view of neurocranium…………………………………………………………...123

IV.16 Characters 23-26, Dorsal vertebral processes of exoccipitals - posterior view of neurocranium…………………………………...…..124

IV.17 Character 27, Expansion of short compressed teeth on upper pharyngeal toothplates - ventral view, of upper pharyngeal toothplates……………126

IV.18 Character 29, Tooth rows on upper, pharyngobranchial toothplate 2 - lateral view of upper pharyngeal toothplate 2……………….……..…127

xii IV.19 Character 33, Teeth on posterior medial row of lower pharyngeal toothplate - dorsal view of lower pharyngeal toothplate with enlargement of tooth on posterior row……………………………………………………..……..129

IV.20 Character 36, Sutures on ventral aspect of lower pharyngeal toothplate - ventral view of lower pharyngeal toothplates…………………………130

IV.21 Character 38, Gill rakers on ceratobranchial of first gill arch (most rostral arch) protrude over central ridge of arch - medial view of first gill arch showing ceratobranchial gill rakers……………………………….…....131

IV.22 Character 40, Suturing between the anterior and posterior ceratohyal – medial view of left hyoid……………………………….…………….132

IV.23 Character 41, Ventral expansion of 4th abdominal centrum, - posterior view of 4th abdominal vertebrae………...... 133

IV.24 Character 42, Processus medialis of pectoral girdle - ventral view of pectoral girdle……………………………………...…134

IV.25 Character 44, Spur on parhypural - lateral view of caudal fin skeleton………………………………….…136

IV.26 Character 54, Caudal fin spot position, - caudal fin lateral view, lateral line scales (with pores), and scales with dark pigment on caudal fin illustrated………………………………….141

IV.27 Character 62, Upper and lower lateral line scale row position - Illustration of Condition 1, showing overlapping upper and lower lateral line rows……………………………………………………...…………144

IV.28 Character 66, Oral epithelium - lateral view of left side of head showing mouth with lips pulled back………………………………………….…146

IV.29 Character 68, Dorsal head profile - lateral view of head……………………………………………………147

IV.30 Character 69, Acute head profile - lateral view of head……………………………………………………148

IV.31 Character 76, Frenum on lower lip - ventral view of head………………………………………………..…148

IV.32 Character 78, Ventral fold of lower lip - lateral view of head……………………………………………………149

xiii

IV.33 Character 73, Position of mouth - lateral view of head …………………………………………………...150

IV.34 Character 76, Pectoral fin shape - lateral view of pectoral fin…………………………………………….151

IV.35, Character 92, Sensory pores above eye - lateral view of head………………………………………………...….155

IV.36 Character 97, Overall body depth - body shape outline…………………………………………………….158

V.1: Phylogeny of Middle American cichlids based on 16S sequence. This is a strict consensus of 76,582 trees, CI = .344, RI = .513…………….………204

V.2: Phylogeny of Middle American cichlids based on COI sequence. This is a strict consensus of 39 trees, CI=.219, RI = .506……………………….….205

V.3: Phylogeny of Middle American cichlids based on TMO-4C4 sequence. This is a strict consensus of 135 trees, CI=.737, RI=.715………...…………206

V.4: Phylogeny of Middle American cichlids based on S7 first introns sequence. This is a strict consensus of 252 trees, CI = .801, RI=.687….…………207

V.5: Phylogeny of Middle American cichlids based on total evidence (16S, COI, TMO-4C4, S7, cyt b, and morphology). A strict consensus of 61 trees, CI =.335, RI =.494…………………………………………...…………208

V.6: Phylogeny of Middle American cichlids based on total evidence (16S, COI, TMO-4C4, S7, cyt b, and morphology) with composite taxa for Cichla and Crenicichla...... 209

V.7: Phylogeny showing a parsimony optimization of biogeographic area on the phylogeny of Middle American cichlids (same topology as Fig.V.6).…210

xiv LIST OF TABLES

Table

II.1: Comparison of morphometrics among Nandopsis species………………....64

II.2: Meristics, fin counts include spiny rays and soft rays, gill rakers counted on most ventral half of the most rostral gill-arch, centra count includes the last hypural bearing centra…………………………………..…………...65

III.1: Taxa sequenced, with GenBank and UMMZ catalogue numbers…...……82

III.2: Estimated dates for nodes of interest with associated reference letters on the phylogeny (Fig.III.1)……………………………………………………..89

IV.1. Members of the ‘Nandopsis’ species group sensu Regan……….…….…106

IV.2. Table of all clades of two or more species recovered in this study with a complete list of synapomorphies……………..……………………...…159

IV.3 Character Matrix……………………………...…………………………..168

V.1: List of sampled species…………………...………………………………227

xv ABSTRACT

Unlike other cichlids, the Middle American and Greater Antillean Cichlidae are not present on former Gondwanan fragments. This has led to many conjectures about the age and historical biogeography of this notable disjunct distribution. A phylogenetic approach is used to resolve the complex biogeographic history and taxonomic problems of this group. In a combined analysis incorporating morphological data and sampling from 3.5 kb of molecular characters (16S, COI, cyt b, S7, TMO-4C4) from 109 species (including about 80 of the 115 Middle American taxa) two large clades of Middle American cichlids are recovered, one of which is sister to a Greater Antillean clade. Divergence times based on rates of molecular evolution from a phylogenetic analysis of nuclear genes (S7, TMO-4C4) and mitochondrial genes (COI, 16S) estimate that the separation of Middle American and South American cichlids took place around the end of the Cretaceous, a period when a landbridge connecting South America and the Yucatan Peninsula started to breakup. An Eocene divergence time was estimated for Greater Antillean and Middle American cichlids, consistent with the geological separation of Cuba and Hispaniola from the Yucatan Peninsula. A morphological phylogenetic analysis of 41 Neotropical cichlids using 89 morphological characters finds that the taxonomic sections (sub-genera) once used to reflect putative clades within Cichlasoma poorly document natural groups. The name Nandopsis is restricted to the Greater Antillean clade. A taxonomic revision concludes that Nandopsis vombergi lacks diagnostic features and is a junior synonym of Nandopsis haitiensis. The Miocene Haitian fossil, Nandopsis woodringi, is a member of Nandopsis as evidenced by

xvi synapomorphies shared with that group. Based on combined analyses Herichthys, Thorichthys, Nandopsis and Caquetaia are recovered as monophyletic but Parachromis, Vieja, Amphilophus, Archocentrus and Tomocichla are not. Recent dispersal of Middle American cichlids to South American appears to be more prevalent than the reverse. Several species including the endemic South American “Cichlasoma” festae, C. ornatum and C. facetum are phylogenetically Middle American. Marine dispersal hypotheses are not employed for any disjunct distribution of cichlids because vicariance hypotheses are better able to explain the biogeographic patterns, both temporal and phylogenetic.

xvii CHAPTER I

CICHLID BIOGEOGRAPHY: UPDATED COMMENT AND REVIEW

“Biogeography, if it is a science, must be able to predict pattern from pattern, and estimate process from pattern.” Savage (1982)

ABSTRACT Phylogenetic analyses dealing with disjunct distributions (distributions that must have required either marine dispersal or vicariant events) are reviewed for Cichlidae. The minimum age of Cichlidae as implied by the fossil record is at odds with the timing of the break up of Gondwana. However, all well sampled and robust phylogenies for this group fit a pattern reflecting Gondwanan break-up. The most corroborated relationship between clades across a disjunct distribution is the sister relationship between Indian (Etroplus) and Malagasy (Paretroplus) cichlids. This sister relationship makes Malagasy cichlids paraphyletic reflecting a pre-rifting split between cichlids on Gondwana. Grounds for marine dispersalist hypotheses are not well founded for any cichlid disjunct distribution, leaving vicariance alternatives as the best explanation.

INTRODUCTION Cichlidae is a monophyletic group of brackish and freshwater perciform fishes. More than 1,300 valid species are currently recognized, and there are estimates of nearly 2,000 total species (Kullander, 1998). The distribution of this family has led to numerous hypotheses about the age and historical biogeography of this group. Improved techniques in molecular systematics and the discovery of the oldest known fossil cichlids have led to a resurgence of debate about the historical biogeography of the group (Sparks and Smith,

1 2005; Briggs, 2003; Sparks, 2004; Murray, 2001, 2001b; Vences et al., 2001; Kumazawa et al., 2000). This study will review the evidence supporting alternative biogeographic scenarios explaining cichlid distributions.

Cichlid Biogeography: Overview of the Debate Cichlids are widely distributed in southern continental regions (Fig.I.1) including South and Middle America (≈ 400 spp.), Cuba and Hispaniola (4 spp.), Africa (≈ 1,000 spp.), Madagascar (>18 spp.), Middle East and adjacent areas (Israel, Syria, Iran) (5 spp.) and India (3 spp.). This distribution, because it is associated mainly with Gondwanan fragments, has led researchers to propose a Cretaceous origin for the assemblage (Rosen, 1975; Stiassny, 1991). This period constitutes the period of fragmentation of the southern super-continent, Gondwana.

Fig. I. 1. Cichlid worldwide distribution in black, redrawn from Sparks (2001).

The oldest known fossil cichlids date back to the Eocene (54-38 mya; Murray, 2000, 2000b, 2001). Some researchers have argued that dispersal across marine environments, rather than drift vicariance (vicariance due to continental drift), is more likely given evidence from the fossil record, molecular clock estimates of divergence, and the salt tolerance of some extant cichlids (Briggs, 1984, 2003; Murray, 2001b; Vences et

2 al., 2001). Vicariance biogeographers emphasize the derived nature of the earliest fossil cichlids, and the lack of evidence for intercontinental marine dispersal (Stiassny, 1991; Sparks, 2004; Sparks and Smith, 2005). Both sides have used phylogenetic analyses to support their claims. The monophyly of Cichlidae is supported by morphological (Zihler, 1982; Gaemers, 1984; Stiassny, 1991) and molecular evidence (Farias et al., 1999, 2000; Streelman and Karl, 1997; Sparks and Smith, 2004). Until recently researchers agreed that one of the groups traditionally placed within Labroidei (including Embiotocidae, Pomacentridae, Labridae) was the sister group to cichlids (Stiassny and Jensen, 1987; Zardoya et al., 1996). More recently, Labroidei has been recognized as being an unnatural group (Streelman and Karl, 1997; Sparks and Smith, 2004). Sparks and Smith (2004) found that cichlids are sister to a large assemblage of percomorphs.

Most cichlids are restricted to freshwater habitats, but there are cases of cichlids living in brackish habitats or being found swimming in marine waters (see examples in Murray, 2001b). For some cichlid species a marine habitat is not inhospitable. But this plays no role in our understanding of the historical biogeography of the family.

Selecting Among Biogeographic Hypotheses There are many methods that have been proposed for selecting between biogeographic hypotheses (see Crisci, 2001). I adopt a cladistic biogeographic approach

sensu Rosen (1978) and Nelson and Platnick (1981). This method was selected over others because it uses area cladograms that can be created from the published phylogenies reviewed here. This method was also selected over others because it is the only one that utilizes the principle of parsimony, which by minimizing assumptions finds the most efficient explanation of the evidence (Sober, 1988). The cladistic biogeographic approach assumes a shared correspondence between phylogenetic history and geological history. The relationship between these histories can be seen in congruent patterns of different taxonomic and area cladograms (cladograms

3 with taxon names replaced by distributions) fitting a given pattern of geological history. In this method, dispersal is assumed not to explain a disjunct distribution until vicariance can be falsified (Kluge, 1989; Croizat, Nelson and Rosen, 1974). Vicariance is a simpler interpretation than dispersal for congruent area cladograms of different taxa, because the congruence can be explained by a single event (i.e., the rifting of a continent or orogeny). The same interpretation of distributions by dispersal would require concordant dispersal in the same sequence for many diverse taxa (Fig. I.2, Fig. I.3).

Fig. I. 2. Congruent cladograms for three hypothetical taxa, divided into six by allopatric speciation. Vicariance is the most parsimonious explanation of their distribution because it requires only one step. Minimally dispersal would require three independent steps.

The essence of vicariance biogeography is that barriers arise secondarily to divide up species. Vicariance events, because they are tied to earth history, can only be supported by a very limited range of phylogenetic patterns. Dispersal scenarios, because they can occur without any underlying congruent process, can be claimed to support an unlimited range of phylogenetic patterns.

4 Few of all possible phylogenetic reconstructions will be congruent with the hypothesized sequence of geological fragmentation; only these few will not fasify vicariance explanations. In the other case, all distribution patterns can be explained by dispersal. Therefore, overwater dispersal scenarios that can explain cichlid distributions will not be employed unless vicariance scenarios have been falsified. Vicariance scenarios for freshwater fishes have the following potential falsifiers: (1) the phylogenetic pattern (sequence of lineage divergence) does not follow the timing of known geological processes (i.e., the sequence of fragmentation, Fig. I.3), (2) members of particular lineages are younger than hypothesized related vicariance events (3) a species of the group under study is found on both sides of a supposed barrier to dispersal (4) estimated sequence divergence times reliably show that lineages have diverged after the particular vicariant event under study.

Fig. I. 3. A demonstration of how an area cladogram can imply a sequence of geological fragmentation that is incongruent with the actual sequence of fragmentation.

5 Overwater dispersal will be the favored mechanism to explain a disjunct distribution when falsifiers of vicariance - by adding assumptions to a vicariance hypothesis - make dispersal a simpler alternative.

Alternative Plate Tectonic Reconstructions If cichlids were present on Gondwana before its breakup, the minimum age of the family would be dated to the Cretaceous (165 mya). There is general agreement that the timing of the breakup of Gondwana occurred around this time (Hay et al.,1999; see below). What is not as clear is the sequence of fragmentation. Reconstructions of this

sequence are essential to recognizing congruent patterns on area cladograms. There are two major plate tectonic reconstructions for the Cretaceous. Until

recently there was a consensus that by about 100 million years ago, South America, Africa, India/Madagascar, Australia/Antarctica had separated from one another and deep ocean passages lay between them (Dietz and Holden, 1970; Smith et al., 1973; Briden et al., 1973; Barron, 1987; Scotese, 1991). In this scenario the sequence of separation relevant to this discussion begins with the India/Madagascar block separating from the rest of Gondwana, followed by the subsequent separation of India from Madagascar and then Africa from South America. A recent reexamination of continental margins using seismic profiling and sea floor magnetic lineations has revealed an alternative tectonic reconstruction (Hay et al.,1999). In the Hay et al. scenario Africa separates from a single continental block consisting of South America-Antarctica-Madagascar-India-Aust ra lia in the Early Cretaceous (120-130 mya). This large continental block remained intact until the Late Cretaceous (80-90 mya). The consequence of this scenario is that members of this continental block would have been connected to each other 10 to 20 million years longer than they would have been to Africa (Fig. I.4). These alternative geological fragmentation scenarios will have important consequences on whether different hypotheses of cichlid relationships favor either a vicariance or marine dispersal scenario.

6 Fig. I.4. Early Cretaceous tectonic reconstructions of (a) Hay et al., 1999 and (b) ‘classical’ reconstruction. Both images from Hay et al., 1999.

(A)

(B)

7 Proposed Cichlid Relationships In order to test alternative biogeographic scenarios a suitable cichlid phylogeny must be recovered. Unfortunately, there have been many proposed phylogenetic reconstructions for Cichlidae. This review will include all those recovered by phylogenetic analyses and that include multiple disjunct areas. Figure I.5 includes all phylogenies that recovered monophyletic groups on each former Gondwanan fragment with cichlids (Africa, Madagascar, India, South America) or those that recovered monophyletic cichlid clades on Gondwanan fragments except Madagascar. Madagascar appears paraphyletic in all analyses that include the genus Paretroplus from Madagascar and Etroplus from India with other Malagasy taxa. Etroplus and Paretroplus have been found as sister lineages in every well-sampled phylogenetic analysis that has included them. Only one phylogeny, a 16S parsimony analysis of Sparks (2004; Fig. I.6H) recovered a paraphyletic Etroplus. When Sparks (2004) adds CO1 to this 16S dataset (Fig. I.5A), Etroplus is recovered as monophyletic and sister to Paretroplus. Figure I.6 includes all analyses that have recovered paraphyletic cichlid groups on

Gondwanan fragments (besides Madagascar). Figure I.7 shows two contrasting hypothetical area cladograms, illustrating what these clear alternatives would be able to suggest about cichlid biogeography.

Phylogenies Congruent With Vicariance Scenarios A recovered pattern of endemic clades on each of the former Gondwanan landmasses would make a strong case for vicariance (Fig. I.7A). This situation appears in the phylogenies of Schliewen and Stiassny (2003, Fig. I.5B), Farias et al., 2000 (Fig. I.5C, D), and Streelman et al., 1998 (Fig. I.5C). However, these phylogenies recover Malagasy cichlids as monophyletic only because Paretroplus and Etroplus were not sampled with other Malagasy taxa. Paraphyly of the Malagasy cichlids does not rule out a Gondwanan vicariance scenario, it may even be the strongest case for vicariance (see India-Madagascar section).

8 The cichlid phylogenetic analyses that recover monophyletic African and Neotropical lineages as sister groups (Schliewen and Stiassny, 2003; Sparks, 2004; Streelman et al., 1998; Zardoya et al., 1996; Farias et al., 1999, 2000; in Figs. I.5B-G) are sign ificant because of congruence with phylogenies of other taxa. Freshwater fishes with area cladograms congruent with cichlids (in showing sister relationships between South America and Africa rather than with a lineage on another continent) include: lungfishes (Lepidosiren and Protopterus), osteoglossids (Arapaima and Heterotis), nandids, aplocheiloid cyprinodontiforms, galaxiids, and synbranchids (Lundberg, 1993; Lundberg et al., 2000; Rosen, 1975).

Phylogenies That Falsify Vicariance Scenarios Figures I.6A-H are phylogenies that have recovered paraphyletic groups on one or more Gondwanan fragments other than Madagascar. They all deviate little from those phylogenies that support vicariance. All of these are caused by one of three taxa, Heterochromis multidens, Oxylapia polli or Chaetobranchopsis australis. Heterochromis multidens, a monotypic African taxon, appears to be a difficult species to code in morphological analyses. This species appears in various places in the

morphological phylogenies of different researchers. Stiassny (1991, Fig. I.6E) found that this species might be sister to the etroplines (Etroplus of India and Paretroplus of Madagascar) or a clade of African/Neotropical cichlids. Kullander (1998, Fig. I.6G) recovered this species as nested within a large Neotropical assemblage. Oliver (1984, Fig. I.6F) also recovered this species as closely related to the South American genus Cichla, which together with Heterochromis was excluded from both the Neotropical and African assemblages. These three analyses borrowed components from Cichocki’s (1976, Fig. I.5H) non-cladistic character analysis, which did not include Heterochromis. Lippitsch (1995, Fig. I.6C) also found a paraphyletic African lineage by recovering Heterochromis as the sister taxon to a Neotropical/African assemblage. Lippitsch’s (1995) phylogeny was based entirely on scale and squamation characters. These characters are often not

9 useful in higher-level phylogenetic analyses because of their potentially homoplasious behavior and non-independence. Murray (2001b) recovers the same tree as Lippitsch (1995) because Lippitsch’s tree is a major component of a composite tree that also included the work of Nishida (1991), Meyer et al. (1994) and Stiassny (1991). From Murray’s sample, only Lippitsch (1995) and Stiassny (1991) had cladograms that included the worldwide distribution of cichlids. The phylogenies of Nishida (1991) and Meyer et al. (1994) focused only on African cichlids. Since one purpose of Murray’s analysis was to show that marine dispersal is the most parsimonious conclusion for the historical distribution of cichlids, it would have been more appropriate to include all other analyses dealing with cichlid disjunct distributions. Molecular analyses that included Heterochromis multidens find either that this

species is nested within the African assemblage or sister to the rest of the African assemb lage, recovering a monophyletic African assemblage in either case (Schliewen and Stiassny, 2003; Sparks, 2004; Sültmann et al., 1995; Farias et al., 1999, 2000; Sparks and Smith, 2004). Morphological analyses may recover this species in non-African lineages because of a lack of synapomorphies that link it to African taxa. The placement of Oxylapia polli, a monotypic Malagasy taxon, is also problematic. It was recovered as sister to the Neotropical assemblage in a neighbor joining tree using a nuclear fragment (Tmo-4C4; Streelman et al., 1998; Streelman and

Karl, 1997; Fig.I.6A). In a parsimony analysis using the same fragment, it was found sister to a Neotropical and African assemblage (Streelman and Karl, 1997, Fig.I.6B). This species is found nested within the Malagasy/Indian clade, when Malagasy genera other than Paretroplus are included (Sparks, 2004; Sparks and Reinthal, 2001; Farias et al., 1999, 2000; Stiassny, 1991). Therefore, the position of Oxylapia in Streelman et al. (1998) and Streelman and Karl (1997) can be explained by the inclusion of Paretroplus without any other Malagasy taxa. It can only be assumed that the analysis sampling more species, versus fewer, is superior.

10 Sparks (2001) found a paraphyletic Neotropical assemblage in his morphological analysis because Chaetobranchopsis australis, a species from Paraguay, was found to be the sister group to a monophyletic African assemblage (nesting Africa within the Neotropical assemblage, Fig.I.6D). Unfortunately his molecular trees and combined molecular and morphological data tree did not include this species. A congener, Chaetobranchopsis orbicularis, was included in the molecular study of Sparks and Smith (2004) and recovered this species nested within the Neotropical clade. As mentioned previously the reduced data set phylogeny of Sparks (2004) using only 16S recovers a paraphyletic Etroplus (Fig.I.6H). When CO1 is analyzed simultaneously with 16S, Etroplus is recovered as a monophyletic group (Fig.I.5A). It is also recovered as monophyletic in Sparks and Smith (2004) where all three Etroplus species were sampled.

The paraphyly of different groups in these analyses (Fig.I.6A-H) can be attributed mainly to geographically biased sampling or difficulties in finding morphological synapomorphies. Incomplete sampling has led to hypotheses of relationships that do not appear in better-sampled analyses. Therefore, the biogeographic conclusions based on these analyses must be called into question.

11 Fig. I.5. The diversity of phylogenetic trees shown as area cladograms that support vicariance hypotheses of cichlid biogeography. All recover monophyletic groups of cichlids on Gondwanan fragments or with a paraphyletic Madagascar.

A) Sparks, 2001, parsimony using morphology, CO1 and 16S; *The positions of Africa and the Neotropics are switched when morphological characters are excluded from this data set Sparks, 2001, 2004

B) Schliewen and Stiassny, 2003 Parsimony using Tmo-M27, Tmo-4C4, DXTU

12 Fig. I.5. continued

C) Farias et al., 2000, minimum evolution, 16S, Tmo-M27 and Tmo-4C4 Farias et al., 2000, parsimony, Tmo-M27 and Tmo-4C4 Streelman et al., 1998, neighbor joining bootstrap consensus, Tmo-4C4, Tmo-M27

D) Farias et al., 2000, total evidence tree, 16S, Tmo-M27 and Tmo-4C4 and morphology from Kullander, 1998

Fig. I.5. continued

E) Zardoya et al., 1996 50% majority rule bootstrap neighbor joining consensus tree using Tmo-M27 Farias et al., 2000 Minimum evolution, using 16S rRNA Sparks and Smith, 2004, parsimony, 16S, COI, Tmo-4C4, Histone H3

F) Zardoya et al., 1996, parsimony Tmo-M27

Fig. I.5. continued

G) Farias et al., 1999 Minimum evolution and parsimony using 16S

H) Cichocki, 1976 clique analysis, morphological characters, Madagascar assumed ‘basal’

15 Fig. I.6. The diversity of phylogenetic trees shown as area cladograms that falsify cichlid vicariance scenarios by requiring marine dispersal. See text for details that explain why each can be dismissed.

A) Streelman and Karl, 1997; Streelman et al., 1998 Neighbor joining from maximum likelihood distances Tmo-4C4

B) Streelman and Karl, 1997, parsimony, Tmo-4C4

16 Fig. I.6. continued

C) Murray, 2001b composite tree (from Stiassny, 1991; Lippitsch, 1995; Nishida, 1991 and Meyer et al., 1994) Lippitsch, 1995, parsimony, scale and squamation characters

D) Sparks, 2001 parsimony using morphology

17 Fig. I.6. continued

E) Stiassny, 1991, parsimony using morphology

F) Oliver, 1984, parsimony based largely on morphological characters of Cichocki, 1976

18 Fig. I.6. continued

G) Kullander, 1998, parsimony using morphology

H) Sparks, 2004, parsimony using 16S

19 Fig. I.7: Examples of (A) Monophyletic cichlid groups on Gondwanan fragments (B) Paraphyletic cichlid groups on Gondwanan fragments

(A) The possible explanations for recovering monophyletic cichlid lineages on Gondwanan fragments (1) vicariance by continental drift (because the sequence of divergence follows the proposed timing of fragmentation) (2) single dispersal events from one continent to another, without any subsequent successful dispersals and extinction of the dispersing species (3) multiple successful dispersals, followed by a extinction events that left the following pattern (reciprocal monophyly) (4) widespread marine ancestor(s), post fragmentation, gave rise to lineages on separate regions that subsequently speciated forming clades

(B) The possible explanations for recovering paraphyletic cichlid lineages on former Gondwanan fragments (1) multiple successful dispersal events (2) widespread dispersal on the Gondwanan continent that led to a paraphyletic pattern, followed by fragmentation

20 Molecular Clocks Molecular data have a potential advantage over morphological data in that some molecules or molecular fragments may change at a rate that can potentially be used to date specific divergences. The ‘phylometric approach’ (Avise et al., 1987) recognizes that sequence differences among taxa contain information about both phylogenetic relationships and the timing of separation between lineages (Grant and Leslie, 2001). To work, the rate at which the molecular fragment is said to change must be correct, and the timing of the geological event that this clock is calibrated upon must also be correct.

Two molecular clock analyses using cichlids have been published. The first (Kumazawa et al., 2000) supports vicariance, but because of its small sample size and ambiguous results its conclusions may be questionable. The second molecular clock analysis, Vences et al. (2001) fails in several ways. Applying a molecular clock hypothesis, Vences et al. (2001) argued that dispersal was more likely than vicariance to account for the distribution of cichlids. The authors tested their hypothesis using a molecular clock based on 16S and Tmo-4C4. Vences et al. (2001) use East African Rift Lakes to calibrate a molecular clock. The ages of these lakes are imprecisely known; estimates range from 1 - 0.012 million years for Lake Victori a, and from 4 - 12 million years for Lake Tanganyika (Vences et al. 2001; Barlow, 2000). Several authors have noted the paradox that the cichlid lineage of Lake Victoria is supposedly 250,000 to 750,000 years old, whereas geological evidence suggests that the basin of the lake dried out completely 12,000 to 15,000 years ago (Johnson et al., 1996; Nagl et al., 2000; Fryer, 2001). These wide estimates of ages for the lakes, and the fact that the lineages within the lakes may not be the same age as the lakes themselves (Nishida, 1991, Meyer et al., 1991) make this molecular clock calibration suspect. A more rigorous calibration based on fossils or lake level fluctuations may have been more reasonable (see Sturmbauer et al., 2001).

21 Vences et al. (2001) also found both molecular fragments failed to show clock- like evolution; clock-like evolution was significantly rejected by the likelihood ratio test. Nevertheless, in order to use these data, the authors only used lineages that fit the particular clock for a given molecule (when rate constancy is not rejected); this was common practice to deal with molecules that do not fit the clock model across a cladogram (following Takezaki et al., 1995). The linear accumulation of transversions was assumed, and transitions were removed in all analyses because they never met rate constancy. The resulting cladograms from likelihood, parsimony and neighbor-joining analyses were not provided. The authors instead used a previous tree (Farias et al., 1999, 2000) on which to map divergence dates. Curiously, the divergence dates mapped onto this phylogeny do not correspond with the sequence of those divergences (i.e., older divergences are sometimes given younger dates than more recent divergences). An earlier molecular clock analysis (Kumazawa et al., 2000) using mitochondrial genes NADH dehydrogenase subunit 2 and cytochrome b supported a vicariance scenario for cichlid distributions. Only six cichlid species were used in the study (three each from Africa and South America) and 15 species total. The authors concluded that Neotropical and African cichlids had separated from each other between 80 and 120 million years ago, fitting the vicariance model. Their clock was calibrated upon paleontological and molecular data. Unfortunately, 95% confidence intervals were not given; therefore, it is unclear how strongly their evidence supports vicariance. For a discussion on the imprecise nature of traditional molecular clocks (those that do not relax the assumption of rate constancy) and their confidence limits see Hillis et al. (1996) and Lundberg (1998).

22 INDIAN, MALAGASY PHYLOGENETIC RELATIONSHIPS With the exception of a few problematic studies discussed above (those shown in Fig.I.6), all analyses recover cichlids on each Gondwanan landmasses as monophyletic except Madagascar. Malagasy cichlids in the genus Paretroplus are sister to Indian cichlids in the genus Etroplus (together known as Etroplinae or etroplines). The remaining Malagasy cichlids are either sister to the etroplines (Sparks, 2001, 2004, Fig.I.5A; Farias et al., 1999, Fig.I.5G), sister to all remaining cichlids excluding the etroplines (Zardoya et al., 1996; Farias et al., 2000; Sparks and Smith, 2004, Fig.I.5E), or sister to all other cichlids including the etroplines (Zardoya et al., 1996, Fig.I.5F).

Geological History A landmass composed of Madagascar and India was isolated following the breakup of Gondwana in the Mesozoic, with a later separation of Madagascar from India circa 88 million years ago (Storey et al., 1995; Hay et al. 1999; Segoufin and Patriat, 1981 ref. in McCall, 1997). Cretaceous deposits on Madagascar lack any cichlid fossils (Gottfried and Krause, 1994, 1998). A number of currently well-represented endemic vertebrate taxa are also absent from the fossil record of Madagascar, leading some to argue that a recent (Cenozoic) colonization via marine dispersal took place (Krause et al., 1997; Gottfried and Krause, 1994, 1998; Gottfried et al., 1998). Two cichlid species are native to Sri Lanka, Etroplus maculatus and E. suratensis

(Lundberg, 1993). Sri Lanka is separated from India by the Palk Strait, which at its narrowest is a mere 19km (Pethiyagoda, 1991). It is a continental island part of the Indian plate, and may have separated from India during the Early Cretaceous. It is hypothesized that the island was submerged during the Paleocene and Miocene, meaning that the fauna on the island is composed of recent (post-Miocene) invaders. Some of these invasions may have occurred overland via freshwater channels. A landbridge 170 km wide has been proposed during a period of low sea level during the last glacial period,

23 15,000 to 20,000 years ago (Cooray, 1984 ref. in Pethiyagoda, 1991). In this study ‘India’ refers to the entire Indian subcontinent that includes Sri Lanka and adjacent areas.

Vicariance Sparks (2001) states that “Certainly the most compelling evidence in favor of vicariance and a more ancient age of origin for cichlids than fossils currently establish, is the ... Malagasy-South Asian assemblage.” This Etroplinae assemblage comprising Paretroplus and Etroplus carries the strongest corroborated relationship between cichlids across a disjunct distribution (Sparks and Smith, 2004; Sparks, 2004; Stiassny et al., 2001; Stiassny, 1991; Farias et al., 1999; Cichocki, 1976; Oliver, 1984). The breakup of the India/Madagascar block took place 88 million years ago (Storey et al., 1995; Rabinowitz et al., 1983). The phylogenies reviewed here imply that

the ancestor of the Indo/Malagasy clade was present before the separation of India and Madagascar. The Indo/Malagasy relationship is congruent with both prevailing hypotheses of Gondwanan fragmentation. The Hay et al. reconstruction predicts an African sister relationship to the rest of the continental fragments, because it occurred first, with a break between India and Madagascar occurring later (Fig. I.5A of Sparks, 2001 is the only cladogram of cichlid relationships congruent with the Hay et al. scenario). The classical Gondwanan break-up reconstruction requires a rift between India/Madagascar and Africa/South America, followed by a rifting between Africa and

South America and finally between Madagascar and India (Fig. I.5C best exemplifies this scheme , figures I.5B, I.5D, I.5F, and 1.5G are also congruent but require pre-drift divergences on Gondwana).

24 THE GREATER ANTILLES Discovering the relationships of the Greater Antillean cichlids is important because the Antilles are not geologically Gondwanan in origin despite being connected at various times to Gondwanan fragments (i.e., South America; Rosen, 1975, 1985). Leon Croizat’s (1962) metaphor of vicariance biogeography being like reconstructing a pane of glass that has been repeatedly shattered seems particularly relevant to the Greater Antilles. Geologically the Greater Antilles rest upon a small plate located between the much larger North America, South American and Cocos and Nazca plates. The Caribbean plate itself can be divided into a series of minor plates that have separated and merged at various times in their history (Perfit and Williams, 1989). Despite their history

and position on a tectonic plate, these islands are still often referred to as “oceanic.” This nomenclature, like “secondary freshwater fishes” assumes a priori that overwater dispersal is the only mechanism for organisms to populate these islands (see Briggs, 2003 and discussion in Sparks and Smith, 2005). Paulay (1994) defined oceanic islands as islands that have never been connected to a mainland continent and therefore are populated solely by dispersal. Given recent tectonic reconstructions of this area, this definition does not fit the Greater Antilles.

Greater Antillean Cichlids There are five cichlids described from the Greater Antilles, Nandopsis

tetracanthus, N. ramsdeni, N. haitiensis, N. vombergi and “Cichlasoma” woodringi (see Myers, 1928 and Darlington, 1957 report on other possible but “dubious” species and

distributions). Nandopsis tetracanthus and N. ramsdeni are restricted to Cuba, and the others to Hispaniola. The fossil “Cichlasoma” woodringi is either Upper or Middle

Miocene (23 to 5 mya) in age (Rivas, 1986; Myers, 1928). Van Couvering (1982) called the fossil “?Pliocene” without explanation in her text, while it remained Miocene in her figures. This younger age has been cited by later authors without additional explanation (Casciotta and Arratia, 1993; Murray, 2001b). This fossil has the notoriety of being the

25 only known freshwater fish fossil from the Antilles (Williams, 1989). Bussing (1985) and Rivas (pers. comm. in Burgess and Franz, 1989) comment that this fossil is indistinguishable from the extant N. haitiensis. A number of authors have stated that Cuba, particularly its eastern half, was once united with Hispaniola in the early history of the Caribbean (Williams, 1989; Perfit and Williams, 1989). According to Pitman et al. (1993), Cuba and Hispaniola did not separate until a shearing in the late Middle Eocene. Nearly 90 % of the 71 species of Antillean freshwater fishes occur on Cuba and Hispaniola (Burgess and Franz, 1989). Sixty-five of these are endemic to an island or island group (Burgess and Franz, 1989). Surprisingly, Puerto Rico, the fourth largest Antillean island, separated from Hispaniola by only the narrow Mona Passage (130 km), completely lacks native freshwater fishes. Puerto Rico does have available habitats; an introduced African cichlid and many other introduced species maintain populations there (Burgess and Franz, 1989). Fishes dispersing from Central or South America would likely reach Jamaica or the Lesser Antilles before the remaining Antillean islands because of their proximity to the continents (Fig.I.8). There are no cichlids on Jamaica (it does have six other native freshwater species), and there are only two native freshwater fishes on the entire Lesser Antilles.

Phylogenetic Relationships To date no formal phylogenetic analysis has included the Cuban and Hispaniolan species with Middle American and South American species. Rosen (1975) included a cladogram that presented a sister relationship between Middle American and Greater Antillean cichlids. However, his four-taxon cladogram was based on a ‘personal communication’ from Cichocki who did not include this analysis in any published material or in his dissertation (1976). A publication by Rosen’s last student, Rauchenberger (1988) used a polytomy to show the relationships between Greater Antillean and Middle American cichlids. Without a phylogenetic diagnosis, we lack a

26 measure for selecting between alternative mechanisms for explaining this disjunct distribution. Myers (1938, 1966; see also Darlington, 1957) hypothesized that the freshwater fishes of the Greater Antilles dispersed from Central America, arguing mainly based on the salt tolerance of these ‘secondary freshwater fishes.’ Myers incorrectly believed the islands formed in situ without connection to other landmasses. Cichlids were lumped into this category of being ‘secondarily freshwater’ based solely on their occurrence on islands (see also Briggs, 1984). Rivas (1986) noted that the native cichlids on Cuba and Hispaniola are known only from landlocked freshwater habitats, never brackish or marine habitats. Remarkably, Briggs (2003), and Smith and Bermingham (2005) continue using primary and secondary freshwater classifications. Recently, Sparks and Smith (2005) updated the arguments of Rivas (1986) in dispelling the salt tolerance arguments of Briggs’ (2003) in his updating of Briggs (1984). Salt tolerance of certain members of a clade should have no bearing on explaining the distribution of that entire clade, particularly not one as species-rich as cichlids. Salt tolerance in some cichlids does not provide evidence that they are capable of crossing oceans. Bussing (1985) and Martin and Bermingham (1998) hypothesize that South American and Central American cichlids may have dispersed around the continental landmasses by migrating along coastlines, pointing again to salt tolerance in some

cichlids (see also Kullander,1983). Endemism of cichlid species on Cuba and Hispaniola implies that either (1) they speciated there with no successful dispersal events from mainland to island or from island to island, or (2) extinctions have left this pattern of endemism (“reciprocal monophyly”). Based on his vicariance model, Rosen (1975) gave a Mesozoic minimum age to the freshwater fish fauna of the Greater Antilles including cichlids, atherinids (silversides), poeciliids and other cyprinodontiforms, synbranchid eels and gars. Rauchenberger (1988) attempted to create a composite area cladogram from 12 other

27 cladograms using these taxa to support Rosen’s vicariance model. Most of the trees she used in her analysis are poorly resolved (the cichlid area cladogram is a polytomy). Only the Gambusia phylogeny provides resolution to the composite. Unfortunately, the Gambusia relationships cited from Fink (1971, 1971b) ignores some key elements of the original (Fink pers. comm.). Rauchenberger’s analysis included only one South American species, which due to its placement as sister to the remaining clades, did not affect the composite tree. Therefore, it is not surprising that a close relationship between the Greater Antillean and Central American taxa (the only possibility, given the sampling) was found.

Fig. I. 8: Map of Caribbean region and adjacent areas.

28 THE MIDDLE EAST, EUROPE AND ADJACENT AREAS There has been little work on the cichlids of the Middle East, Europe and adjacent areas (Fig.I.9). The only phylogenetic hypotheses that include species that belong to Northern Africa, and the areas adjacent, (including the Middle East) are by Klett and Meyer (2002) and Trewavas (1983).

Current Distribution and Sister Relations Iranocichla hormuzensis Coad, 1982 is the only cichlid endemic to Iran and is disjunct from other cichlid populations. Coad (1982) described this species as a relict of a larger cichlid distribution that crossed the Arabian Peninsula. He notes several periods where dispersal across the Arabian Peninsula could have been possible over freshwater corridors during dry periods in the Pliocene or Pleistocene (see Kosswig, 1965, 1973; Banister and Clarke, 1977). Trewavas (1983) believed this species to be the sister to the Ethiopian Danakilia franchettii and noted that perhaps the genera should be synonomized. Klett and Meyer (2002) found that Iranocichla may be the sister taxon to Sarotherodon or Stomatepia, both tilapiines in the Sarotherodon group. The Sarotherodon group includes a number of Middle Eastern and northern African species and may be monophyletic - although the entire Tilapiine tribe itself may not be a clade (Klett and Meyer, 2002). Tristramella magdalenae is a Tilapiine found in Syria (Trewavas, 1983; Coad, 1982) that may also be closely related to Danakilia franchettii (Trewavas, 1983). Tristramella simonis is known from Israel and is sister to the African Sarotherodon occidentalis (Klett and Meyer, 2002). Loiselle (1985) notes that a number of species from the Nile basin have ranges extending to the areas around the Persian Gulf, including the tilapiines: Tilapia zillii,

Sorotherodon galilaeus and S. aureus. Loiselle (1985) suggests that this might be evidence for dispersal from Africa to the Middle East during a recent period of warmer, wetter climates.

29 Fossils From the Area Cichlids are absent from the Arabian Peninsula, despite being found in the surrounding areas. This absence may be due more to a current desert barrier than to a marine barrier. Fossil cichlids in the Middle East listed by Murray (2001b) include two Pliocene cichlids in Israel and at least three lineages of Oligocene Saudi Arabian cichlids. There is some speculation that one of the two Israeli cichlid fossils resembles Tilapia zillii (Murray, 2001b). Van Couvering (1982) had earlier speculated on the relationships of one of these fossils and placed it in the genus Tilapia and aligned it with an African Tilapia fossil. Brown (1970 in Trewavas, 1983) describes a possible Tilapiine from the Miocene or Oligocene of Saudi Arabia. This species was potentially a holdout from before the formation of the Arabian Desert in the Pliocene or Pleistocene. An Italian fossil from the Eocene has been reported but is highly suspect (Murray, 2001). Potential Miocene cichlid fossils from Italy, Germany, Moravia and Switzerland are known and may be tilapiines but need more work (Gaemers, 1989; Murray, 2001b). Because there is a land connection from Northern Africa to Arabia and Europe, dispersal via freshwater routes over land is a plausible explanation of the presence of cichlids in these areas.

30 Fig. I. 9: Map of European, Middle Eastern and African areas discussed.

31 THE GLOBAL CICHLID FOSSIL RECORD The absence of cichlid fossils from the Mesozoic is an argument often presented against vicariance explanations of cichlid biogeography (Murray, 2001b; Lundberg, 1993). However, “absence of evidence is not evidence for absence” (Maisey, 1993), particularly given the rarity of Cretaceous freshwater deposits (Patterson, 1993). The cichlid fossil record does not falsify hypotheses of vicariance. There are no post-drift fossils from extant geological lineages found on continents other than the ones they are found on today (e.g. there are no African cichlid fossils found in the Neotropics). No fossil cichlid from any geographic lineage predates proposed vicariance events. Presence of such fossils would favor predrift intercontinental speciation over vicariance (Lundberg, 1993, see Fig I.7B).

Minimum Ages Murray (2000, 2001a) has described the oldest known fossil cichlids, which are from the Eocene of Tanzania. These fossils establish the minimum age of cichlids to be 45 million years old. Before Murray’s (2000) descriptions, the oldest fossil cichlids were from the Oligocene (23 to 34 mya; Van Couvering, 1982; see also Greenwood, 1989). Murray has placed these cichlids in a derived position among African cichlids, sister to hemichromines, based on predorsal spine count and squamation (Murray 2001, 2001b). Sparks (2004) also noted that these fossils appear to be more derived than either Heterochromis or Tylochromis (the least morphologically derived African cichlids) and thought they were closely related to modern haplochromines or hemichromines. He commented (2004) that if cichlids from the Eocene seem morphologically indistinguishable from extant forms, then cichlids are likely a much older group than the fossils imply. Fossil cichlids are also known from the Miocene of South America (Casciotta and Arratia, 1993). Stewart (2001) listed all African cichlid fossils of Neogene Africa; some

of these fossils are noted to have modern morphologies (Stiassny, 1991).

32 The Acanthomorph Record Some authors challenged a Cretaceous (circa 165 mya) origin for Cichlidae (Lundberg, 1993), because of their derived position on the tree of acanthomorphs and the minimu m age suggested by the fossil record (Fig.I.10). The absence of cichlids of a Cretaceous age becomes pivotal when their phylogenetic position is taken into account. The derived position of cichlids within acanthomorphs means that for cichlids to be Cretaceous all other less derived acanthomorphs must also be Cretaceous. There are many actinopterygians (ray finned fishes) in the global Mesozoic fossil record, but no acanthomorphs (spiny-rayed teleosts) until the Late Cretaceous (ca. 75-65 mya). Lundberg (1993) considered this evidence that acanthomorphs did not yet exist. If fossils provide the minimum age of taxa, cichlids are at least Eocene (Murray, 2001); Percomorpha, possibly Late Cretaceous but the earliest unquestionable perciformes are early Cenozoic (after 65 mya; Patterson, 1993); and acanthomorphs are early Late

Cretaceous (Patterson, 1994). Percomorpha appear 20-25 million years after the first acanthomorphs in the fossil record (Murray, 2001). Higher acanthomorph diversity is not recorded in fossils until the Late Paleocene/Early Eocene, at which time there appears to be a major radiation (Fig.I.10) (Patterson, 1993, 1994). Acanthomorphs include 15,000 species in 280 to 300 families (Patterson, 1993, 1994). There are 9,000 Perciformes species within 150 to 230 families

(Patterson, 1994; Nelson, 1984). The fossil record suggests that the majority of Perciform diversity, including cichlids, radiated in the Eocene. The possibility of a gap in the fossil record for Acanthomorphs between the Late Cretaceous and Paleocene has been proposed (Patterson, 1993; Sparks, 2001). However, there is little evidence for the gap being due to anything besides their actual absence, particularly in the face of the abundance of Actinopterygian fossils from freshwater deposits of this time. Only the discovery of new fossils and deposits from this period will resolve this question.

33 Fig. I. 1 0: “Romerogram” s h owi n g di v ersi ty of groups over the geological time sc ale (from Pa tte rs on, 1 994 ). L ower clado g ram showing osteichth y an r e lationshi p s (fr om Patterson 1994) cladogram at top rig ht showing aca n tho m orp h rela tionships from John son and Patt e rson (1991). N ote t hat c i chli ds are me m bers of P e rcif o rmes , and that the acant h omo r ph f ossil r ecor d begins in the Late Cret aceous w i th an exp l osio n in t h e Eocene .

34 DISCUSSION The historical biogeography of cichlids is best explained by vicariance associated with the break-up of Gondwana. This paper has reviewed all prior phylogenetic hypotheses that have implications for global cichlid biogeographic hypotheses. Many cladograms have been recovered that support the vicariance scenario explaining cichlid distributions (Sparks and Smith, 2004; Schliewen and Stiassny, 2003; Sparks, 2001, 2004; Farias et al., 1999, 2000; Streelman et al., 1998; Zardoya et al., 1996; Fig. I.5). Well-supported relationships within cichlids show a convincing vicariance pattern reflecting Gondwanan fragmentation. Sparks’ (2001; Fig. I.5A) phylogeny based on combined evidence from mitochondrial genes and morphology is the only analysis that recovers a topology that follows the proposed sequence of fragmentation by Hay et al. (1999). Notably at deep nodes this tree is weakly supported (possibly because of the use of rapidly evolving mitochondrial genes), both by Bremer support (< 2) and Jackknife resampling (< 50) values in clades showing sister lineages between the Neotropics and India/Madagascar. All other analyses in Figure I.5 support the traditional consensus view of Gondwanan fragmentation (Dietz and Holden, 1970; Smith et al., 1973; Briden et al., 1974; Barron, 1987; Scotese, 1991). Phylogenies that did not support vicariance hypotheses are found in Figure I.6. These phylogenies do not differ much from those phylogenies that support vicariance. In each of these analyses one of a few noted “problematic taxa” break up what are otherwise monophyletic groups. These phylogenies should not be interpreted as evidence for dispersal for these species, rather the analyses must be reconsidered in light of more recent and more rigorous evidence. Morphological studies (Fig.I.6C-G) recover Heterochromis outside of an otherwise monophyletic group of African cichlids. Notably, its position is inconsistent in each morphological study, suggesting that morphological features that reveal its relationships have yet to be discovered. Molecular phylogenies

35 that include this species always recover it as part of an African clade (Schliewen and Stiassny, 2003; Farias et al., 1999, 2000; Sparks, 2004; Sparks and Smith, 2004). Exclusion of Oxylapia polli from the rest of Malagasy cichlids in some of the phylogenies presented in the studies of Streelman and Karl (1997) and Streelman et al. (1998) are more likely due to incomplete sampling of Malagasy cichlids than to dispersal of that species from South America to Madagascar. Oxylapia is recovered with other Malagasy cichlids in better-sampled phylogenies (those sampling additional Malagasy cichlids other than Paretroplus; Stiassny, 1991; Farias et al., 1999, 2000; Sparks and Smith, 2004; Sparks, 2004). The sister relationship between Chaetobranchopsis australis and a clade of African cichlids recovered in the morphological analysis of Sparks (2001) resulted in a paraphyletic Neotropical assemblage (Fig. I.6.D). Further evidence needs to be collected on this species but the fact that a congener was recovered within the Neotropical lineage in a subsequent study (Sparks and Smith, 2004) calls into question the earlier finding. Additional potential falsifiers of vicariance were not discovered. No cichlid species has a distribution on both sides of a supposed barrier to dispersal. No cichlid lineage is convincingly shown to be younger than vicariant events from phylogenetic analyses. The two molecular clock analyses (Vences et al., 2001; Kumazawa et al., 2000) have several problems (see ‘Molecular Clock’ section) and therefore should not be considered falsifiers of vicariance or dispersal. Some readers may disagree with using parsimony as grounds to favor a particular hypothesis of biogeography. There are several possible explanations for the presented evidence of monophyly on Gondwanan fragments regardless of ones philosophical standpoint. Lundberg (1993) discusses four possible interpretations for disjunct monophyletic distributions that are modified here for the Gondwanan cichlid case; (1) Drift vicariance: ancestral distribution was Gondwanan, and the origins of separate clades were facilitated by the separation of the Gondwanan fragments. (2) Pre-drift, inter-

36 Gondwanan speciation: lineages separated on Gondwana, potentially incorporating extinction post-drift leading to the same pattern (reciprocal monophyly). (3) Post-drift marine dispersal: single successful marine dispersal events from one fragment to another of following (and independent of) the break-up of Gondwana (4) Indirect dispersal pathways: a now extinct marine cichlid species (or several species from one clade), gave rise to independent freshwater lineages in different Gondwanan fragments via dispersal. Cladograms that would have supported option 2 would be similar to hypothetical cladogram shown in Fig. I. 7B. This option would require free dispersal across Gondwana that would have left a pattern showing multiple paraphyletic groups. Because this pattern is never recovered, this scenario would require assumptions about extinctions of all lineages save for one each on India, Africa and the Neotropics. There is no evidence for this pattern of extinctions. All cichlid fossils thus far collected belong to extant lineages on their respective fragments. Postdrift dispersal (option 3) would require one species to give rise to the entire continental fauna on each Gondwanan fragment. Although there is always a possibility of dispersal occurring in this way, there is no evidence for only one successful dispersal event having occurred between continents followed by extinction of the founder species.

Option 4 considers that a widely-spread, now extinct, marine ancestor might have given rise to all the extant freshwater lineages. This possibility as pointed out by

Lundberg (1993) is intriguing because it would explain the absence of Mesozoic cichlid fossils. It would also allow for the minimum age of cichlids to be corroborated by both their fossil record and their distribution. If there were a marine ancestor that gave rise to monophyletic cichlid groups on separate Gondwanan fragments the phylogenetic pattern that would be predicted would be a polytomy for Africa, India, Madagascar and the Neotropics. This pattern is not found in any phylogenetic analysis and should therefore be ruled out.

37 This review finds that drift vicariance (option 1) is the most parsimonious conclusion despite some pre-drift divergence required to explain Malagasy paraphyly. The evidence for pre-drift divergence on Gondwana in the paraphyly of Malagasy cichlids. This is most evident in Figs. I.5 A,E,F and G all found Malagasy cichlids paraphyletic because of the Paretroplus sister relationship with the Indian genus Etroplus. These phylogenies find that cichlids on Gondwana were divided by the break-up of Gondwana but that cichlids on the Madagascar/India block had already separated between the Paretroplus/Etroplus clade (Etroplinae) and the remaining Malagasy cichlids (Ptychochrominae). It remains unclear whether these remaining Malagasy cichlids are

more closely related to the Neotropical/African clade or to the Etroplinae as both patterns are recovered. The Ptychochrominae are also sometimes recovered as sister to all other cichlids. All three patterns fit with the traditional view of Gondwanan fragmentation, differing only in the pre-drift separation of the Ptychochrominae. The analysis of Sparks and Smith (2004) is the most thorough analysis of Cichlidae to date. It is the first analysis to sample representatives across all geographic assemblages. It also uses more characters than in any other study; utilizing 2222 aligned nucleotide characters from mitochondrial and nuclear genes. The area cladogram of their cichlid phylogeny (Fig. I.5E) is the same as recovered in analyses by Zardoya et al., 1996 and Farias et al., 2000, making it arguably the most corroborated phylogeny of those discussed here. In these analyses the Ptychochrominae are sister to the Neotropical/African clade. Therefore, the most corroborated, well sampled area cladogram reflects pre-drift separation of Etroplinae from all other cichlids. When the Madagascar/India block separated from the rest of Gondwana, the Ptychochrominae became isolated with etroplines, and the Ptychochrominae diverged from the African/Neotropical clade (Fig.I.11). The African/Neotropical clade separated when those two landmasses divided followed by the separation of Madagascar and India which divided the etroplines of India (Etroplus) and Madagascar (Paretroplus; Fig.I.11D).

38 Non-cichlid taxa also show a congruent phylogenetic pattern. Aplocheiloid killifish (Murphy and Collier, 1997) have a similar distribution, and a congruent phylogeny with Cichlidae (Fig. I.12). The Aplocheiloid phylogeny shows a sequence of divergence fitting the classic reconstruction of Gondwanan fragmentation (Figs.I.5B-E). The phylogeny of some rainbowfishes (Melanotaenioidei) also supports this classical reconstruction (Sparks and Smith, 2004b). The Greater Antillean cichlid fauna is clearly an area of future study that requires a well-supported phylogeny to distinguish between alternative biogeographic hypotheses. Studies that recover paraphyletic connections between the islands and neighboring continents, or with species that are commonly found in marine waters, may favor dispersal. Phylogenies for non-cichlid Greater Antillean taxa already exist, some supporting vicariance, other supporting marine dispersal (Murphy and Collier, 1996;

Iturralde-Vinent and MacPhee, 1999; Lydeard et al., 2002). The relationships of European fossil cichlids and extant and extinct Middle Eastern cichlids seem to point to North African lineages (Klett and Meyer, 2002; Trewavas, 1983). The presence of cichlids in these areas can be explained by freshwater connections overland. The relationship between India (Etroplus) and Madagascar (Paretroplus) is the most highly corroborated sister relationship between two now separated Gondwanan fragments. If a dispersalist explanation is used to explain this relationship then cichlids were able to successfully disperse between Madagascar and India across the Indian Ocean but not from Madagascar to Africa across the narrow (430 km) Mozambique Channel. The unlikeness of this scenario makes a strong case against a dispersalist hypothesis that explains the distribution of the Cichlidae.

39 Fig. I. 1 1: The br ea k-u p of G ond w an a (Au strali a an d Ant ar cti ca no t pict ured) with b iog e ogr ap hic h ist or y o f cichlids according to area cladogram from Fig .I.5 E . C i rcles and ovals illustrate cichlid populations. A shows a widespread cichlid clade that diverges pre- drift into the Etroplines and all other cich lids in B. C illust rates the separation of the India/Madagascar block from the rest o f G ondw ana, resu lting in the Ptychoch rom i nae b ecoming isola t ed a n d divergi ng fr om t h e Af r ican / Sou t h A m erican cichlids. D i llust r ates the sepa ration of S out h Am e rica and Africa lead ing t o Neotro p ical a nd Afric a n cichlid clades as a result of vicariance. D a lso i llustr a tes the separation of Indian and Malagasy Etrop line s that resu l t in the Paretr o plus lineage o f Ma d agas c ar a n d th e Etro plus of India.

Fig. I. 12: (A) Map of worldwide distribution of the Aplocheiloidei (b) Area cladog ram from phylogenetic hypothesis of Murphy & Collier, 1997 usin g thre e mitocho n dria l gen es and parsimony

41 LITERAT URE C ITED

Avise, J.C., Arnold, J., Ball, R.M. Jr., Bermingham, E., Lamb, T., Neigel, J.E., Reeb, C.A., and Saunders, N.C. (1987) Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Annual Review of Ecology and Systematics 18, 489-522.

Banister, K.E. and Clarke, M.A. (1977) The freshwater fishes of the Arabian Peninsula. Journal of Oman Studies, Special Report (Scientific Results of the Oman Flora and Fauna Survey), 111-154.

Barlow, G.W. (2000). The Cichlid Fishes: Nature’s Grand Experiment in Evolution. Perseus Publishing, Cambridge.

Barron, E.J. (1987) Cretaceous plate tectonic reconstructions. Paleogeography, Paleoclimatology, Palaeoecology 59, 3-29.

Briden, J.C., Dewry, G.C., and Smith, A.G. (1974) Phanerozoic equal area world maps. Journal of Geology 82, 555-574.

Briggs, J.C. (1984) Freshwater fishes and biogeography of Central America and the Antilles. Systematic Zoology 33, 428-435.

Briggs, J.C. (2003) Fishes and birds: Gondwana life rafts reconsidered. Systematic Biology 52, 548-553.

Burgess, G.H. and Franz, R. (1989) Zoogeography of the Antillean freshwater fish fauna. In: Biogeography of the West Indies: Past, Present, and Future (ed C.A.Woods) Sandhill Crane Press, Gainesville, pp. 263-304.

Bussing, W.A. (1985) Patterns of distribution of the Central American Ichthyofauna. In: The Great American Biotic Interchange (eds F.G. Stehli and S.D. Webb) Plenum, New York, pp. 453-473.

Casciotta, J. and Arratia, G. (1993) Tertiary cichlid fishes from Argentina and reassessment of the phylogeny of New World cichlids (Perciformes: Labroidei). Kaupia - Darmstädter Beiträge zur Naturgeschichte 2, 195-240.

Cichocki, F.P. (1976) Cladistic history of cichlid fishes and reproductive strategies of the American genera Acarichthys, Biotodoma, and Geophagus. Vol 1. Unpublished PhD Dissertation, The University of Michigan, Ann Arbor.

Coad, B.W. (1982) A new genus and species of cichlid endemic to Southern Iran. Copeia 1982, 28-37.

42 Cooray, P.G. (1984) An Introduction to the Geology of Sri Lanka (Ceylon). National Museum of Sri Lanka, Colombo.

Crisci, J.V. (2001) The voice of historical biogeography. Journal of Biogeography 28,157-168.

Croizat, L. (1962) Space, Time, Form: The Biological Synthesis. Published by the author, Caracus.

Croizat, L., Nelson, G. and Rosen, D.E. (1974) Centers of origin and related concepts. Systematic Zoology 23, 265-287.

Darlington, P.J. Jr. (1957) Zoogeography: The Geographical Distribution of Animals. Chapman and Hall, London.

Dietz, R.S. and Holden, J.C. (1970) The beak-up of Pangea. Scientific American 223, 30- 41.

Farias, I.P., Ortí, G., Sampaio, I., Schneider, H. and Meyer, A. (1999) Mitochondrial DNA phylogeny of the family Cichlidae: monophyly and fast molecular evolution of the Neotropical assemblage. Journal of Molecular Evolution 48, 703-711.

Farias, I.P., Ortí, G. and Meyer, A. (2000) Total evidence: molecules, morphology, and the phylogenetics of cichlid fishes. Journal of Experimental Zoology (Mol. Dev. Evol.) 288, 76-92.

Fink, W.L. (1971) A revision of the Gambusia puncticulata complex (Pisces: Poeciliidae) Publications of the Gulf Coast Research Laboratory Museum 2, 11-46.

Fink, W.L. (1971b) A revision of the Gambusia nicaraguensis species group (Pisces: Poeciliidae). Publications of the Gulf Coast Research Laboratory Museum 2, 47- 77.

Fryer, G. (2001) On the age and origin of the species flock of haplochromine cichlid fishes of Lake Victoria. Proceedings of the Royal Society, London B 268, 1147- 1152.

Gaemers, P.A.M. (1984) Taxonomic position of the Cichlidae (Pisces, Perciformes) as demonstrated by the morphology of their otoliths. Netherlands Journal of Zoology 34, 566-595.

Gaemers, P.A.M. (1989) The first cichlids (Perciformes, Pisces) from E urope: the new, fossil genus Europotilapia, evidence from otoliths and teeth. Annals du Muséum Royal d’Afrique Centrale, Sciences Zoologiques 257, 109-116.

43 Gottfried, M.D. and Krause, D.W. (1994) Late Cretaceous Fishes from Madagascar: a first look. Journal of Vertebrate Paleontology 14 supp., 26A.

Gottfried, M.D. and Krause, D.W. (1998) First record of gars (Lepisosteidae, Actinopterygii) on Madagascar: Late Cretaceous remains from the Mahajanga Basin. Journal of Vertebrate Paleontology 18, 275-279.

Gottfried, M.D., Randriamiarimanana, L.L., Rabarison, J.A. and Krause, D.W. (1998) Late Cretaceous fish from Madagascar: implications from Gondwanan biogeography. Journal of African Earth Sciences, Special Abstracts Issue, Gondwana 10 27, 1.

Grant, W.S. and Leslie, R.W. (2001) Inter-ocean dispersal is an important mechanism in the zoogeography of hakes (Pisces: Merluccius spp.) Journal of Biogeography 28, 699-721.

Greenwood, P.H. (1989) The haplocromine species (Teleostei, Cichlidae) of the Cunene and certain other Angolan rivers. Bulletin of the British Museum (Natural History) Zoology 47, 187-239.

Hay, W.W., DeConto, R.M., Wold, C.N., Wilson, K.M., Voight, S., Schulz, M., Wold- Rossby, A., Dullo, W.C., Balukhovsky, A.N. and Soding, E. (1999) Alternative global Cretaceous paleogeography. In: Evolution of the Cretaceous Ocean- Climate System: Boulder Colorado (eds E. Barrera and C.C. Johnson) Geological Society of America Special Paper 332, pp. 1-47.

Hillis, D.M., Mable, B.K. and Moritz, C. (1996) Application of Molecular Systematics: The State of the field and a look to the future. In: Molecular Systematics 2nd Edition (eds D.M. Hillis, C. Moritz and B.K. Mable), Sinauer Associates, Sunderland, pp. 515-543.

Iturralde-Vinent, M.A. and MacPhee, R.D.E. (1999) Paleogeography of the Caribbean region: implications for Cenozoic biogeography. Bulletin of the American Museum of Natural History 238, 1- 95.

Johnson, G. D. and Patterson, C. (1993) Percomorph phylogeny: a survey of acanthomorphs and a new proposal. Bulletin of Marine Science 52, 554-626.

Johnson, T.C., Scholz C.A., Talbot, M.R., Kelts, K., Ricketts, R.D., Ngobi, G., Beuning, K., Ssemmanda, I., and McGill, J.W. (1996) Late Pleistocene desiccation of Lake Victoria and rapid evolution of cichlid fishes. Science 273, 1091-1093.

Klett, V. and Myer, A. (2002) What, if anything, is a Tilapia? - Mitochondiral ND2 phylogeny of tilapiines and the evolution of parental care systems in the African Cichlid fishes. Molecular Biology and Evolution 19, 865-883.

44 Kluge, A.G. (1989) A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes) Systematic Zoology 38, 7-25.

Kosswig, C. (1965) Zur historischen Zoogeographie der Ichthyofauna im Süsswasser des sudlichen Kleinasiens. Zoologische Jahrbücher für Systematik 92, 83-90.

Kosswig, C. (1973) Über die Ausbreitungswege sogenannter perimediterraner Süsswasserfische. Bonner Zoologische Beiträge 24, 165-177.

Krause, D.W., Hartman, J.H. and Wells, N.A. (1997) Late Cretaceous vertebrates from Madagascar: implications for biotic change in deep time. In: Natural change and human impact in Madagascar (eds S.M. Goodman and B.D. Patterson) Smithsonian Institute Press, Washington, pp. 3-43

Kullander, S.O.(1983) A revision of the South American cichlid genus Cichlasoma (Teleostei: Cichlidae). The Swedish Museum of Natural History, Stockholm.

Kullander, S.O. (1998) A phylogeny and classification of the South American Cichlidae (Teleostei:Perciformes). In: Phylogeny and Classification of Neotropical Fishes (eds L.R. Malabarba, R.E. Reis, R.P. Vari, Z.M. Lucena, and C.A.S. Lucena ) Edipucrs., Porto Allegre, pp. 461- 498.

Kumazawa, Y., Yamaguchi, M. and Nishida, M. (2000) Mitochondrial molecular clocks and the origin of euteleostean biodiversity: familial radiation of perciforms may have predated the Cretaceous/Tertiary boundary. In: The Biology of Biodiversity (ed M.Kato) Springer-Verlag, Tokyo, pp. 35-52.

Lippitsch, E. (1995) Scale and squamation character polarity and phyletic assessment in the family Cichlidae. Journal of Fish Biology 47, 91-106.

Loiselle, P.V. (1985) The Cichlid Aquarium. Tetra-Press, Melle.

Lundberg, J.G. (1993) Africa-South American Freshwater Fish Clades and Continental Drift: Problems with a Paradigm. In: Biological relationships between Africa and South America (ed P. Goldblatt), Yale University Press, New Haven, pp 157-199.

Lundberg, J.G. (1998) The temporal context for the diversification of Neotropical fishes. In: Phylogeny and Classification of Neotropical Fishes (eds L.R. Malabarba, R.E. Reis, R.P. Vari, Z.M. Lucena and C.A.S. Lucena ) Edipucrs., Porto Allegre, pp. 49-68.

Lundberg, J.G., Kottelat, M., Smith, G.R., Stiassny, M.L.J. and Gill, A.C. (2000) So many fishes, so little time: an overview of recent ichthyological discovery in continental waters. Annals of the Missouri Botanical Garden 87, 26-62.

45 Lydeard, C., Wooten, M.C. and Meyer, A. (2002) Molecules, morphology, and area cladograms: a cladistic and biogeographic analysis of Gambusia (Teleostei: Poeciliidae). Systematic Biology 44, 221-236.

Maisey, J.G. (1993) Tectonics, the Santana lagerstätten, and the implication for late Gondwanan biogeography. In: Biological relationships between Africa and South America (ed P. Goldblatt), Yale University Press, New Haven, pp 435-454.

Martin, A.P., Bermingham, E. (1998) Systematics and evolution of lower Central American cichlids inferred from analysis of cytochrome b gene sequences. Molecular Phylogenetics and Evolution 9, 192-203.

McCall, R.A. (1997) Implications of recent geological investigations of the Mozambique Channel for the mammalian colonization of Madagascar. Proceedings of the Royal Society of London (B) 264, 663-665.

Meyer, A., Kocher, T.D. and Wilson, A.C. (1991) African fishes. Nature 350,468.

Meyer, A., Montero, C. and Spreinat, A. (1994) Evolutionary history of the cichlid fish species flocks of the East African great lakes inferred from molecular phylogenetic data. Archiv für Hydrobiologie, Ergebnisse der Limnologie 44, 407- 425.

Murphy, W.J. and Collier, G.E. (1996) Phylogenetic relationships within the Apocheiloid fish genus Rivulus (Cyprinodontiformes, Rivulidae): Implications for Caribbean and Central American Biogeography. Molecular Biology and Evolution 13, 642- 649.

Murphy, W.J. and Collier, G.E. (1997) A molecular phylogeny of aplocheiloid fishes (Atherinomorpha, Cyprinodontiformes): the role of vicariance and the origins of annualism. Molecular Biology and Evolution 14, 790-799.

Murray, A.M. (2000) Eocene cichlid fishes from Tanzania, East Africa. Journal of Vertebrate Paleontology 204, 651- 664.

Murray, A.M. (2000b) The Palaeozoic, Mesozoic and Early Cenozoic fishes of Africa. Fish and Fisheries 1, 111-145.

Murray, A.M. (2001) The oldest fossil cichlids (Teleostei: Perciformes): indication of a 45 million-year-old species flock. Proceedings of the Royal Society of London B. 268, 679 - 684.

Murray, A.M. (2001b) The fossil record and biogeography of the Cichlidae (Actinopterygii: Labroidei). Biological Journal of the Linnean Society 74, 517- 532.

46 Myers, G.S. (1928) The existence of cichlid fishes in Santo Domingo. Copeia 1928, 33- 36.

Myers, G.S. (1938) Fresh-water fishes and West Indian zoogeography. Annual Report of the Smithsonian Institute 1937, 339-364.

Myers, G.S. (1966) Derivation of the freshwater fish fauna of Central America. Copeia 1966, 766-773.

Nagl, S., Tichy, H., Mayer, W.E., Takezaki, N., Takahata, N. and Klein, J. (2000) The origin and age of haplochromine fishes in Lake Victoria, East Africa. Proceedings of the Royal Society of London B. 267, 1049-1061.

Nelson, J.S. (1984) Fishes of the World 2nd edition. John Wiley and Sons, New York.

Nelson, G. and Platnick, N.I. (1981) Systematics and Biogeography: Cladistics and Vicariance. Columbia University Press, New York.

Nishida, M. (1991) Lake Tanganyika as an evolutionary reservoir of old lineages of East African cichlid fishes: inferences from allozyme data. Experientia 47, 974-979.

Oliver, M.K. (1984) Systematics of African cichlid fishes: determination of the most primitive taxon, and studies on the haplochromines of Lake Malawi. Unpublished PhD dissertation, Yale University.

Patterson, C. (1993) An overview of the early fossil record of Acanthomorphs. Bulletin of Marine Science, 52, 29-59.

Patterson, C. (1994) Bony fishes In: Major features of Vertebrate Evolution. (eds D.R. Prothero and R.M. Schoch) Short Courses In Paleontology 7, Paleontological Society, University of Tennessee, Knoxville, pp 57-84.

Paulay, G. (1994) Biodiversity on oceanic islands - its origin and extinction. American Zoologist 34, 134-144.

Perfit, M.R. and Williams, E.E. (1989) Geological constraints and biological retrodictions in the evolution of the Caribbean sea and its islands. In: Biogeography of the West Indies: Past, Present, and Future. (ed C.A. Woods) Sandhill Crane Press, Gainesville, pp. 47-102.

Pethiyagoda, R. (1991) Freshwater fishes of Sri Lanka. Wildlife Heritage Trust of Sri Lanka, Colombo.

47 Pitman, W.C. III, Cande, S.C., LaBrecque, J. and Pindell, J. (1993) Fragmentation of Gondwana: the separation of Africa from South America. In: Biological Relationships between Africa and South America. (ed P. Goldblatt), Yale University Press, New Haven Connecticut, pp. 15-34

Rabinowitz, P.D., Coffin, M.F. and Falvey, D. (1983) The separation of Madagascar and Africa. Science 220, 67-69.

Rauchenberger, M. (1988) Historical biogeography of Poeciliid fishes in the Caribbean. Systematic Zoology 37, 356-365

Rivas, L.R. (1986) Comments on Briggs (1984): Freshwater Fishes and Biogeography of Central America and the Antilles. Systematic Zoology 35, 633-639.

Rosen, D.E. (1975) A vicariance model of Caribbean biogeography. Systematic Zoology 24, 431-464.

Rosen, D. E. (1978) Vicariant patterns and historical explanation in biogeography. Systematic Zoology 27, 159-188.

Rosen, D.E. (1985) Geological hierarchies and biogeographic congruence in the Caribbean. Annals of the Missouri Botanical Garden 72, 636-659.

Savage, J.M. (1982) The enigma of the Central American herpetofauna: dispersal or vicariance? Annals of the Missouri Botanical Garden 69, 464-547.

Schliewen, U.K. and Stiassny, M.L.J. (2003) Etia nguti, a new genus and species of cichlid fish from the River Mamfue, Upper Cross River basin in Cameroon, West- Central Africa. Ichthyological Explorations of Freshwaters 14, 61-71.

Scotese, C.R. (1991) Jurassic and Cretaceous plate tectonic reconstructions: Palaeogeography, Paleoclimatology, Paleoecology 87, 493-501.

Segoufin, J. and Patriat, P. (1981) Reconstruction de l’Ocean Indien occidental pour les épouques des anomalies M21, M2 and 34. Paleoposition de Madagascar. Bulletin de la Societe Geologique de France XXIII, 603-607.

Smith, S.A. and Bermingham, E. (2005) The biogeography of lower Mesoamerican freshwater fishes. Journal of Biogeography 32, 1835-1854.

Smith, A.G., Briden, J.C. and Drewry, G.E. (1973) Phanerozoic World Maps, Special Papers in Paleontology 12,1- 42.

Sober, E. R. (1988) Reconstructing the Past: Parsimony, Evolution and Inference. Bradford, MIT Press, Cambridge.

48 Sparks, J.S. (2001) Phylogeny and biogeography of the Malagasy and South Asian Cichlid Fishes (Teleostei: Perciformes:Cichlidae), including a survey of the freshwater fishes of Madagascar. Unpublished doctoral dissertation, University of Michigan.

Sparks, J.S. (2004) Molecular phylogeny of the Malagasy and South Asian cichlids (Teleostei: Perciformes: Cichlidae). Molecular Phylogenetics and Evolution 30, 599-614.

Sparks, J.S. and Reinthal, P.N. (2001) A new species of Ptychochromoides from southeastern Madagascar (Teleostei: Cichlidae), with comments on the monophyly and relationships of the ptychochromine cichlids. Ichthyological Exploration of Freshwaters 12, 115-132.

Sparks, J. and Smith, W.L. (2004) Phylogeny and biogeography of cichlid fishes (Teleostei:Perciformes:Cichlidae) Cladistics 20, 501-517.

Sparks, J. and Smith, W.L. (2004b) Phylogeny and biogeography of the Malagasy and Australasian rainbowfishes (Teleostei: Melanotaenioidei): Gondwanan vicariance and evolution in freshwater. Molecular Phylogenetics and Evolution 33, 719-734.

Sparks, J. and Smith, W.L. (2005) Freshwater Fishes, Dispersal Ability, and Nonevidence: “Gondwana Life Rafts” to the Rescue. Systematic Biology 54, 158- 165.

Stewart, K.M. (2001) The freshwater fish of Neogene Africa (Miocene-Pleistocene): systematics and biogeography. Fish and Fisheries 2, 177-230.

Stiassny, M.L.J. (1991) Phylogenetic interrelationship of the family Cichlidae: An overview. In: Cichlid Fishes. Behavior, Ecology and Evolution (ed M.H.A. Keenlyside), Chapman and Hall, London. pp. 1-31

Stiassny, M.L.J. and Jensen, J.S. (1987) Labroid intrarelationships revisited: morphological complexity, key innovations, and the study of comparative diversity. Bulletin of the Museum of Comparative Zoology, Harvard 151, 269-319.

Stiassny, M.L.J., Chakrabarty, P. and Loiselle, P.V. (2001) Relationship of the Madagascan cichlid genus Paretroplus with description of a new species from the Betsiboka River drainage of Northwest Madagascar. Ichthyological Exploration of Freshwaters 12, 29-40.

Storey, M., Mahoney, J.J., Saunders, A.D., Duncan, R.A., Kelley, S.P., and Coffin, M.F. (1995) Timing of hot spot-related volcanism and the breakup of Madagascar and India. Science 267, 852- 855.

49 Streelman, J.T. and Karl, S.A. (1997) Reconstructing labroid evolution with single-copy nuclear DNA. Proceedings of the Royal Society of London B. 264, 1011-1020.

Streelman, J.T, Zardoya, R., Meyer, A. and Karl, S.A. (1998) Multilocus phylogeny of cichlid fishes (Pisces: Perciformes): Evolutionary comparison of microsatellite and single copy nuclear loci. Molecular Biology and Evolution 15, 798-808.

Sturmbauer, C., Baric, S., Salzburger, W., Rüber L., and Verheyen, E. (2001) Lake level fluctuations synchronize genetic divergences of cichlid fishes in African lakes. Molecular Biology and Evolution 18, 144-154.

Sültmann, H., Mayer, W.E., Figueroa, F., Tichy, H. and Klein, J. (1995) Phylogenetic analysis of cichlid fishes using nuclear DNA markers. Molecular Biology and Evolution 12, 1033-1047.

Takezaki, N., Rzhetsky, A. and Nei, M. (1995) Phylogenetic test of the molecular clock and lineralized trees. Molecular Biology and Evolution 12, 823-833.

Trewavas, E. (1983) Tilapiine Fishes of the Genera Sarotherodon, Oreochromis and Danakilia. British Museum (Natural History), London.

Van Couvering, J.A.H. (1982) Fossil cichlid fishes of Africa. Special Papers in Paleontology 29. The Paleontological Association, London

Vences, M., Freyhof, J., Sonnenberg, R., Kosuch, J. and Veith, M. (2001) Reconciling fossils and molecules: Cenozoic divergence of cichlid fishes and the biogeography of Madagascar. Journal of Biogeography 28, 1091-1099.

Williams, E.E. (1989) Old problems and new opportunities in West Indian Biogeography, In: Biogeography of the West Indies: Past, Present, and Future (ed C.A.Woods), Sandhill Crane Press, Gainesville, pp. 1- 46.

Zardoya, R., Vollmer, D.M., Craddock, C., Streelman, J.T., Karl, S., and Meyer, A. (1996) Evolutionary conservation of microsatellite flanking region and their use in resolving the phylogeny of cichlid fishes (Pisces: Perciformes). Proceedings: Biological Sciences 263 ,1589-1598.

Zihler, F. (1982) Gross morphology and configuration of digestive tracts of Cichlidae. (Teleostei: Perciformes) phylogenetic and functional significance. Netherlands Journal of Zoology 32, 544-571.

50 CHAPTER II

TAXONOMIC STATUS OF THE HISPANIOLAN CICHLIDAE

“It is thus interesting to record a living species of cichlid from…the Dominican Republic. The single specimen was shot with a .22 caliber pistol…” - Myers (1928), describing the capture of the first Hispaniolan cichlid

ABSTRACT

Hispaniolan cichlids are poorly known, and because of the questionable

taxonomic status of two species, the group is reviewed. Nandopsis vombergi (Ladiges,

1938) is regarded here as a junior subjective synonym of the widespread Hispaniolan

species Nandopsis haitiensis (Tee-Van, 1935). The holotype and only known specimen

of N. vombergi lacks any features that discriminate it from N. haitiensis. The Miocene

fossil “Cichlasoma” woodringi Cockerell, 1924 is assigned to Nandopsis Gill, 1862 based on apomorphies it shares with members of that genus.

INTRODUCTION

Three nominal species of cichlid fishes are described from the Greater Antillean

island of Hispaniola. Two of these species, “Cichlasoma” woodringi and Nandopsis

vombergi, are known only from the original material collected. The third species,

Nandopsis haitiensis, is found throughout the island and closely resembles the other two

species.

The species status of N. vombergi is suspect, based on its resemblance to N.

haitiensis. Ladiges (1938) described Nandopsis vombergi from a single specimen

collected in the Dominican Republic and diagnosed the species on the presence of greatly

51 enlarged lips and a more rounded caudal fin than N. haitiensis. No material attributable

to N. vombergi has subsequently been collected.

The original description of “Cichlasoma” woodringi is lacking in details. This

fossil is the first cichlid described from Hispaniola. No comparative material from

Hispaniola was included in its description (Cockerell, 1924). Most of what is known

about this species is from an account by George S. Myers within the original description

of Nandopsis haitiensis (Tee-Van, 1935). Myers (in Tee-Van, 1935) found no

differences between Nandopsis haitiensis and the fossil except for number and size of the vertebral centra. In addition, Bussing (1985), Tee-Van (1935) and Rivas (quoted by

Burgess and Franz, 1989 and Williams, 1989) all remarked on the similarity of the fossil

and N. haitiensis. The specimen is important because it is the oldest freshwater fossil yet

found in the Greater Antilles and is the oldest Neotropical cichlid fossil outside of South

America. Some controversial issues about this fossil will be discussed here, including its age and taxonomic status.

“Cichlasoma” woodringi, as with most Middle American cichlids, is in

taxonomic limbo. Cichlasoma was restricted to a small group of South American cichlids by Kullander (1983) leaving nearly 100 species formerly in that genus with

uncertain taxonomic status. The generic denomination “Cichlasoma” is a reserve section for former members of that genus.

Kullander (2003) restricted Nandopsis Gill, 1862 to Nandopsis haitiensis, N.

vombergi and N. tetracanthus (Valenciennes, 1831). The latter is the type species of the

genus. Nandopsis tetracanthus and N. ramsdeni, both endemics of Cuba, are sister

species (Chapter III, V).

52 MATERIALS AND METHODS

Digital images were taken from the left side of each specimen. Landmarks

(putatively homologous points on anatomical structures) were chosen in order to best represent the external shape around the body (Fig.II.1). Fin shapes were not included

(except for their placement on the body) because of the challenge of determining homologous positions and poor preservation in many specimens. TPSdig (Rohlf, 1998) was used to digitize the landmarks on the images. Only specimens that were preserved unbent were photographed and digitized.

Fig.II.1: Landmarks used in Principal Components Analysis (1) rostral tip of premaxilla (2) dorsal tip of premaxillary pedicel (3) anterior insertion of dorsal fin (4) posterior insertion of dorsal fin (5) dorsal insertion of caudal fin (6) caudal border of hypural plate aligned with lower lateral line (7) ventral insertion of caudal fin (8) posterior insertion of anal-fin (9) anterior insertion of anal-fin (10) dorsal base of pelvic fin (11) end of opercular membrane ventrally (12) inner aspect of dentary symphysis (13) caudal end of maxilla (14) dorsal end of preopercle ventral to pterotic (15) caudal end of opercule (16) pectoral fin origin (17) anterior margin of midline through eye (18) posterior margin of midline through eye. Base figure is redrawn from Nelson (1994).

53 Generalized Least Squares (GLS) Procrustes superimposition was performed to

remove size from the data. In the optimal superimposition, the distance minimized is the

Procrustes distance, calculated as the square root of the summed squared distances

between homologous landmarks (Goodall, 1991; Rohlf and Slice, 1990). This

superimposition, and the Principal Components Analysis (PCA), was performed in

PCAgen (Sheets, 2001).

Traditional morphometric measurements were taken with a dial caliper.

Measurements and shape definitions (e.g., concavity above eye, caudal fin shape) follow

Barel et al. (1977) except where otherwise noted. “Lip-corrected” measurements have been used on cichlids where certain individuals have greatly expanded lips and other related species or conspecifics do not (Barlow and Munsey, 1976). This corrected distance was used in measuring snout length, which is a measurement taken from the rostral tip of the premaxillae, at the midline (Barel et al., 1977). The lip-corrected measurement for the snout length is necessary because the large lips of the N. vombergi specimen preclude measurement from the premaxillae.

The last hypural-bearing centrum is included in counts (breaking from the convention set by Barel et al., 1977) to avoid confusion with counts from the original

description of N. haitiensis (Tee-Van, 1935). Body depth was taken where the greatest

vertical depth of the body was reached. All radiographs, measurements and counts were

done on the left side. The following abbreviations are used: sk., = skeleton specimens,

mm = millimeters, SL = standard length.

54 INSTITUTIONAL ABBREVIATIONS

AMNH American Museum of Natural History, New York, New York, U.S.A.

ANSP Academy of Natural Sciences of Philadelphia, Philadelphia, Pennsylvania,

U.S.A.

CAS California Academy of Sciences, San Francisco, California, U.S.A.

MCZ Museum of Comparative Zoology, Cambridge, Massachusetts, U.S.A.

MZGJ Museo de Zoología del Grupo Jaragua, Santo Domingo, Dominican

Republic

USNM National Museum of Natural History, Smithsonian Institution,

Washington, D.C., U.S.A.

UMMP University of Michigan Museum of Paleontology, Ann Arbor, Michigan,

U.S.A.

UMMZ University of Michigan Museum of Zoology, Ann Arbor, Michigan,

U.S.A.

ZMH Zoological Museum Hamburg, Hamburg, Germany

Fig.II.2: Nandopsis vombergi, holotype, ZMH 401, 181.7 mm SL

55 Fig.II.3: Nandopsis vombergi, holotype, frontal view to show expansion of lips

Fig.II.4: Nandopsis vombergi, holotype, view of right side of caudal fin to show diagnostic caudal spot that is divided equally by lateral line.

56 SYSTEMATIC ACCOUNTS

Nandopsis haitiensis (Tee-Van, 1935) Figures II.2-6

Cichlasoma haitiensis Tee-Van, 1935: 294, Figs. 270-272 [type locality: Étang

Saumâtre, near Maneville, Cul-de-Sac Plain, Haiti]

Cichlasoma vombergi Ladiges, 1938: 18, Figs. 1-2 [type locality: Rio Yague del

Sur, Dominican Republic]; Kullander, 2003: 639

TYPE MATERIAL EXAMINED: USNM 170907 (holotype,105 mm SL, Cul-de-Sac,

Plain near Naneville, Etang Saumatre, Haiti), USNM 170908 (7, paratypes, 54-81 mm

SL, Cul-de-Sac, Plain near Naneville, Etang Saumatre, Haiti), ZMH 401 (N. vombergi

holotype, 182 mm SL, Rio Yague del Sur)

ADDITIONAL NONTYPE MATERIAL EXAMINED: Dominican Republic: AMNH

229573 (3, 99-111 mm SL, Bahia de Neiba), AMNH 229574 (1, 122 mm SL, Bahia de

Neiba), MCZ 62945 (10, 82-107 mm SL, Laguna Rincon, Cabral, Barahona), MCZ

64571 (2, 98-106 mm SL, Laguna Rincon, Cabral, Barahona), UMMZ 231521 (1 sk.

Lago El Fondo=Etang Saumatre, 1km E of Haitian border at Jimani), UMMZ 243241 (1,

173 mm SL, Rio Piedras 11km SE of La Vega), UMMZ 243287 (1, 96 mm SL, Arroyo

Basima tributary of the Rio Haima, Santo Domingo), UMMZ 243302 (18, 26-92 mm SL,

km 49 on Highway 2 from Azua to San Juan, Rio Yague del Sur, Guanabana), UMMZ

243310 (7, 26-84 mm SL, km 51 on Highway 2, Amiama Gomez, Rio Yague del Sur,

Guanabana), USNM 85764 (1, 124 mm SL, Santo Domingo, Lago Rincon, Cobral),

USNM 367230 (5, 70-100 mm SL, Santo Domingo, Rio Viajamas at Valle de Viajama,

Santo Domingo)

57 Haiti: UMMZ 142438 (4, 76-118 mm SL; 1 sk., 82 mm SL, Cazeau Creek, 4m N

of Port-au Prince), UMMZ 200246 (1, 74 mm SL, 3km NW of Lac du Cayman-near

Thomazeau), USNM 164796 (3, 88-97 mm SL, no data), USNM 164863 (6, 28-87mm

SL, no data) USNM 87360 (1, 113 mm SL, Canot Road, Central Plain of Haiti, at Ford E

of San Michel), USNM 298302 (2, 62-120 mm SL, Etang de Miragoane bridge).

Fig.II.5: Nandopsis haitiensis, holotype, USNM170907, 104.5 mm SL

DIAGNOSIS: A species of Nandopsis distinguished from congeners by the

following combination of characters: chest scales reduced in size, and covered in thick

skin; possession of small dark circular spots distributed throughout the head; and a lack

of a dark area in the asquamate auxiliary region. This species has a spindle shaped body with the greatest body depth reached at the base of the head rather than the midbody.

There are two epurals that are nearly of the same rectangular shape and size, each

supporting a single procurrent caudal fin ray.

58 COMPARISONS TO RELATED SPECIES: The two characters that Ladiges (1938) used

to diagnosis Nandopsis vombergi are present in N. haitiensis (viz.; more rounded caudal fin, lip shape and size) and cannot discriminate the nominal species. Figs.II.2-8 illustrate

the holotype of N. vombergi, the holotype of N. haitiensis, and the caudal fin and lips of additional material.

The caudal fin of N. vombergi is the same general shape as the caudal fin of the

majority of Central American cichlids and is not more rounded than in N. haitiensis (Fig.

II.4; II.8).

The size and shape of the lips also cannot be used to discriminate between

Nandopsis vombergi and N. haitiensis. The lips of the holotype of N. vombergi are

hypertrophied, with the upper lip lobed and the bottom lip bilobed due to the presence of

a median cleft. The presence of lobed lips is a polymorphic trait in N. haitiensis,

although I have not encountered individuals with lips hypertrophied to the extent seen in the holotype of N. vombergi. The degree of lip enlargement is most probably due to the

large size of this specimen; the holotype of N. vombergi is 181.7 mm SL. Individuals of

N. haitiensis with a standard length greater than 150 mm are rare in museum collections,

although this species can reach sizes up to 215 mm SL (Kullander, 2003). An extensive

search of museum collections (including the AMNH, MCZ, USNM, and UMMZ) recovered only one N. haitiensis specimen of comparable size to the holotype of N. vombergi. This specimen was examined, and it lacks expanded or lobed lips (UMMZ

243241, 173 mm). Many smaller individuals examined have lobed lips that closely resemble those of the holotype of N. vombergi but do not yet exhibit a similar degree of hypertrophism (Fig.II.7). In other cichlid species, lobed lips in juveniles are often

59 indicators of greatly expanded lips as adults, as in Amphilophus labiatum (Barlow and

Munsey, 1976). This may also be the case for N. haitiensis because the lobed pattern of the lips of the holotype of N. vombergi (single median lobe on upper lips, “bilobed” appearance of lower lip) is commonly found in smaller N. haitiensis individuals.

Fig.II.6: Nandopsis haitiensis, USNM 122635, 111.5 mm SL, male with nuchal hump

60 Fig.II.7: Nandopsis haitiensis, USNM 87360, 114.6 mm SL, showing expansion of lips

Fig.II.8: Nandopsis haitiensis UMMZ 243241, 173.2 mm SL, caudal fin

61 Fig.II.9: Nandopsis tetracanthus, AMNH 96390; 133.6 mm SL

Fig.II.10: Principal Component Analysis, PC 1 vs. PC 2. Nandopsis ramsdeni (triangles) and Nandopsis tetracanthus (stars) form discrete groups. N. haitiensis (circles) forms a group where the holotype of N. vombergi (square) falls near the middle.

62 Morphometric features are listed in Table II.1. The snout length is longer in the holotype of N. vombergi than in other material of N. haitiensis examined. All other mensural and meristic data of the holotype of N. vombergi are consistent with N. haitiensis (Tables II.1 and II.2). This increased snout length is due to the greatly expanded lips of the holotype of N. vombergi. The removal of the additional distance due to the thickness of the upper lip (9.1mm wide) results in a snout length of 37 % head length, which is within the range observed for N. haitiensis.

An important polymorphism of N. haitiensis that appears in the N. vombergi specimen is a large nuchal hump (Figs.II.2, II.6). This trait occurs in some male individuals of N. haitiensis and is visible in specimens as small as 65 mm SL. Nuchal humps and lobed lips appear to be rare in N. haitiensis. A nuchal hump was present in less than 50 % of male specimens examined; lobed and expanded lips are present in less than 20 % of male and female individuals examined. Only one individual was found with both a nuchal hump and N. vombergi type lips (86mm SL male in USNM 367230).

A Principal Component Analysis of body shape shows that the N. vombergi specimen falls within the middle of the cluster of points of N. haitiensis specimens

(Fig.II.10). The graph of PC 1 vs. PC 2 explains 58% of the total variation in shape among the specimens. Because size was removed from the analyses, this percentage does not include size as a dimension of variation. This graph also shows that overall body shape can be used to discriminate between the spindle shaped Haitian cichlids and more deep bodied Cuban cichlids.

Table II.1: Comp aris on of morphometrics among Nand o psis species. Mean outside parentheses, and range within parent hese s .

Caudal Caudal Orbit Body Snout Interorbital Head L e ngt h Peduncle Peduncle Diameter Depth as% Length as Width as % as % Sta nda rd Length as % Depth as % as % Standard % Hea d Standard Length Standard Standard Head Le ngt h Lengt h Length Length Length Length N. h tienai s is 22 38 (35-41) 40 (36-46) 41 (35-44) 13 (11-15) 15 (12-18) 12 (9-15) n=14 (96-124 mm SL) (18-27)

64 N. vombergi 37 41 46 12 17 18 12 n= 1 (1 8 2 m m SL)

N. ra t tet can hus 23 38 (34-41) 45 (38-56) 34 (30-36) 12 (9-13) 18 (15-20) 12 (9-15) n=19 (97-136 mm SL) (18-31) N. r msd ea ni 27 32 (31-34) 49 (45-51) 42 (35-49) 12 (12-16) 17 (13-18) 12 (11-13) n= 9 (91-167 mm S) L (24-32)

Table II.2: Meristics, fin counts include spiny rays and soft rays, gill rakers counted on most v en tral h alf o f the most rostral gill-a rc h, centra count includes the last hypural bearing centrum. Lateral line count is the total of both anterior and po steri o r por tions.

Anal-fin Gill Raker Lat era l Dorsal Fin Count Vertebral Centra Count Count Line Cou n t

N. haitiensis XIV-XV 10-12 IV 8-11 10-12 27-32 29-30 N=14

65 N. vombergi XV 12 IV 9 11 30 29 N=1 N. woodringi XV 10 IV 8-9 ? ? 28-29 n=1 N. tetracanthus XIV-XV 10-12 IV 8-11 6-9 28-32 28-29 n=19 N. ramsdeni XV 11-14 V 9-10 10-11 28-34 28-29 n=9

52 Ladiges (1938) remarked that the hol otype of N. vombergi had a pigment pattern very similar to that of N. haitiens is. Althoug h muc h of the pigment has faded from the holotype, what rema ins is indistinguishable from the pattern in N. haitiensis. This inclu des small circular spots throughou t the body extending onto the fins, a large midla teral spot belo w th e up per la teral line row directly in front of the lower lateral line, and a spot on the caudal fin strad dling the later al lin e.

The traits that distinguish Nandopsis haitiensis from other members of Nandopsis also are present in the holotype of N. vombergi. Nandopsis haitiensis can be distin guished from all other Nandopsis species by its more slender (compressed) body, chest scales, pigmentation pattern, and featur es of i ts epurals. The chest scales in N. haitiensis form a discrete patch of small embedded skin-covered scales (relative to those of the rest of the body) as oppose d to th e slightly la rger imbricate scales of other

Nandopsis species. The two epurals in Nandopsis haitiensis are similarly sized and together bare two procurrent caudal fin rays. In Nandopsis tetracanthus and N. ramsdeni the tw o epurals are often of dissimilar size an d shap e, bearing together three to four rays in N. tetracanthus and two rays in N. ramsdeni.

Nandopsis haitiensis diff ers fro m N. tetracanthus and N. ramsdeni in having smal l circular spots di stributed th roughout the head versus spots large and often fused together in a reticulate pattern in N. tetracanthus and absent beyond the operculum in N. ramsdeni. Nandopsis haitiensis also differs from the two Cuban species in lacking a dark area in the asquamate auxiliary region.

Nandopsis haitiensis can further be distinguished from Nandopsis tetracanthus in

several ways. Nandopsis tetracanthus have a caudal spot dorsal to the lateral line,

66 whereas in N. haitiensis this spot always straddles the lateral line. The maxillary shank

(the triangular shaped posterior fold at the angle of the mouth) is greatly expanded in N. tetracanthus but is not expanded beyond the angle of the mouth in N. haitiensis. A nuchal hump is absent in N. tetracanthus, and sometimes present in males of N. haitiensis. In N. tetracanthus the dorsal profile of the head lacks a concavity. There is always a pronounced concavity above the eye in N. haitiensis, even when a nuchal hump is absent.

Nandopsis haitiensis can be disting uishe d from N. ramsdeni by four versus five anal-fin spines. Nandopsis haitiensis always has at least six cheek scale rows; Nandopsis ramsdeni has four. A nuchal hump is always present in both males and females of adult

N. ramsdeni.

SYSTEMATIC STATUS OF NANDOPSIS VOMBERGI: Characters that distinguish

Nandopsis vombergi from N. haitien sis could no t be fo und an d I conclude that it is a junior subjective synonym of N. haitiensis. Features reported by Ladiges (1938), including expanded lips and a more rounded caudal fin , are a lso found in specimens of N. haitiensis. The presence of a nucha l hump and lobed lips in the holotype of N. vombergi are polymorphic traits that also occur in N. haitiensis. The size of the lips alone does not warrant distinguishing N. vombergi, as they appear to be only moderately larger than those of other N. haitiensis individuals.

67 Nandopsis woodringi (Cockerell, 1924)

Cichlasoma woodringi Cockerell, 1924: 2, Figs.II.1-2 [type locality: Las Cahobas,

Haiti]

TYPE MATERIAL EXAMINED: USNM 10766 (holotype, fossil, 64 mm SL, Las

Cahobas, Haiti).

ADDITIONAL NONTYPE MATERIAL EXAMINED: USNM 10767 (1, fossil, fragments

of anal and dorsal fins, Las Cahobas, Haiti).

DESCRIPTION AND COMPARISON TO RELATED SPECIES: The holotype of the fossil

Nandopsis woodringi is incomplete, lacking much of the caudal region including the

entire caudal fin, and much of the anterior portion is crushed (Figs.II.11, II.12).

Additional preparation of the specimen has revealed more information about the fossil

than was available to previous workers.

All identifiable portions of the fossil appear to be identical with structures in

Nandopsis haitiensis and N. tetracanthus. Comparable regions include: spine and ray

counts, fin placement on the body, oral and pharyngeal dentition, shapes and sizes of bony elements and vertebral centra count.

Vertebral centrum size and count has been used to distinguish the fossil from N.

haitiensis. The posterior abdominal centrum is the only abdominal centrum visible on the

fossil; it is followed by the first caudal centrum (defined by Barel et al., 1977 as bearing a

clear association to the first anal-fin pterygiophore). Myers (in Tee-Van, 1935) stated

that he counted neural spines in order to estimate the number of abdominal centra; however, only portions of five neural spines and ribs are exposed on the fossil.

68 Fig.II.11: Nandopsis woodringi USNM 10766, 64 mm SL approx.

Fig.II.12: Illustration of USNM 10766 1 –left dentary, 2 – lower arm of premaxilla, 3 – ascending process of premaxilla, 4 – palatine, 5- infraorbital, 6 – lower pharyngeal jaw, 7 – parasphenoid, 8 – quadrate, 9 – right dentary 10 – branchiostegal, 11 – cleithrum, 12 – left pelvic bone above, right pelvic bone below, 13 – pelvic fin, 14 – supraoccipital crest, 15 – 2nd predorsal, 16 – hyomandibular, 17 – last abdominal centrum, 18 – scales, 19 – anal-fin, 20 – enlargement of oral tooth with lingual cusp found between left dentary and lower arm of premaxilla, tooth measures .5 mm

69 Radiographs of the fossil reveal that only nine (non-sequential) abdominal centra remain intact on the fossil and the sediment matrix completely covers all but the last two.

Counting dorsal-fin pterygiophores can reveal the number of abdominal centra. In

Middle American cichlids, the number of dorsal fin pterygiophores associated with a

dorsal fin spine is one fewer than the number of centra (pers. observ.). This formula (# of

dorsal fin ptergiophores – 1 = the number of centra) results in an accurate count of the

number of abdominal centra in Middle American cichlids, including the Antillean cichlids (pers. observ.). By this count there are 12 abdominal centra in the holotype of N.

woodringi (not 14 as stated in Tee-Van, 1935). Immediately posterior to the 12th

abdominal centrum, the anterior 12 caudal centra are exposed. The remaining centra,

those at the caudal peduncle, are lost save for the first one. (This first caudal peduncle

centrum is the 12th caudal centrum.) Myers (in Tee-Van, 1935) proposed that there are

three centra lost at the tail. A latex peel of the caudal region revealed that there are no

fossilized elements or impressions of bone remaining in the caudal region. In N.

haitiensis and N. tetracanthus, there are normally a total of five or six centra in the caudal

peduncle. With this extrapolation the total vertebral count on the specimen comes to 28

or 29 (with the addition of four or five centra in the caudal peduncle). This count is short of the 31 to 33 mentioned by Tee-Van (1935). The usual count of vertebral centra in N. haitiensis is 29, rarely 30. In N. tetracanthus this count is usually 28 and sometimes 29.

Nandopsis woodringi was distinguished from N. haitiensis by having slightly smaller centra (Tee-Van, 1935). The difference in the size of the centra between the extant and fossil material may be due to the comparative material available to Myers.

Myers compared the fossil to a single specimen of N. haitiensis measuring 74 mm (Myers

70 in Tee-Van, 1935). This measurement is assumed to be standard length and not total

length; it is not clearly stated which measurement was taken (measurements of other material are given as standard length). The holotype of N. woodringi is 64mm from the most anterior bone fragment to the last vertebral centrum. The head region is crushed and displaced in a manner that extends the length of this region well beyond what the distance was in life. Because the caudal region is also lost, it appears this specimen was probably no longer than 65 mm SL in life. The few exposed intact centra of the fossil are square and 1.3 mm across. N. haitiensis and N. tetracanthus individuals between 60 mm and 70 mm SL have square vertebrae in equivalent positions that are between 1 and 2 mm

across.

An additional specimen collected with the holotype of N. woodringi contains

fossilized fragments of a nearly complete anal-fin, and portions of a dorsal fin and caudal

peduncle. Cockerell (1924) described the anal-fin fragment as having 21 soft anal rays.

Twenty-one soft anal rays would be a count much higher than the great majority of known cichlids (Nelson, 1994) and appears from examination to be erroneous. There are only nine soft anal rays in the anal-fin fragment. The holotype of N. woodringi appears to have eight anal-fin soft rays. Both of these counts are within the range observed for both N. haitiensis and N. tetracanthus.

All identifiable bony elements of N. woodringi are identical to homologous regions of Nandopsis haitiensis and N. tetracanthus. (However, the slender body shape of the fossil more closely resembles Nandopsis haitiensis.) Meristic counts overlap among all three of these species.

71 With the limited information the specimens of N. woodringi provide it would be

premature to claim that it is not a valid species. It would not be reasonable to claim this

species is a senior synonym of Nandopsis haitiensis or a junior synonym of N.

tetracanthus. Additional material with well preserved diagnostic features might help in

this determination.

Nandopsis woodringi has lingual cusps on the oral teeth (Fig.II.13) as well as four

anal-fin spines. These two characters in combination are unique to Nandopsis haitiensis

and N. tetracanthus. Because Nandopsis tetracanthus is the type species for Nandopsis,

these shared features warrant “Cichlasoma” woodringi being recognized as Nandopsis

woodringi (Cockerell, 1924).

The age attributed to this fossil has been misrepresented in several studies.

Nandopsis woodringi was described as a Miocene fossil from Las Cahobas, Haiti

(Cockerell, 1924). Van Couvering (1982) mentions N. woodringi or the fossil bed in

which it was found three times: in a figure as questionably “Miocene,” in the text as

“?Pliocene” and later as “Upper Miocene.” No justification is given for any of the

assignments. Casciotta and Arratia (1993) use the “?Pliocene” designation of Van

Couvering without explanation. Murray (2001) uses the designation “Pliocene” for the

fossils citing Casciotta and Arratia (1993). Fossil plants collected from the same locality

are dated as either early Middle Miocene (Cooke et al., 1943) or late Miocene (Bowin,

1975; see also Graham, 1990). There is no justification for assigning this fossil to an age

younger than Miocene.

72 ADDITIONAL MATERIALS EXAMINED

Institutional catalog number, number of specimens examined, size range and locality

information follow the species name. Specimens in alcohol unless otherwise noted.

Nandopsis ramsdeni

TYPE MATERIAL EXAMINED: ANSP 68454 (holotype, 170 mm SL, Arroyo Hondo,

Yaterus, Guantanamo, Cuba), ANSP 68455-68458 (4, paratypes, 88-132 mm SL, Guaso

River, Guantanamo, Cuba)

ADDITIONAL NONTYPE MATERIAL EXAMINED: Cuba: MZGJ 00342 (2, 91-107 mm SL,

Guantanamo River), UMMZ 230839 (1, 104 mm SL, Guantanamo River system),

UMMZ 231322 (1, 104 mm SL, Guantanamo River basin).

Nandopsis tetracanthus

TYPE MATERIAL EXAMINED: CAS 78975 (N. t. torralbasi, holotype, 110 mm SL; paratype 134 mm SL, R. Almendares, Calabazar, Cuba)

ADDITIONAL NONTYPE MATERIAL EXAMINED: Cuba: AMNH 1063 (1, 119 mm

SL, Pinar del Rio), AMNH 96390 (4, 133-115 mm SL, Isla de la Juventud, Isla de Pinos),

AMNH 96426 (1, 110 mm SL, La Habana), AMNH 96454 (1, 97 mm SL, Villa Clara),

AMNH 96465 (1, 119 mm SL, Villa Clara, Rio Sagua La Chica), AMNH 96513 (1,

96465 116 mm SL, Cienfuegas), USNM 64003 (2, 119-136 mm SL, San Antonio, Cuba),

UMMZ 171879 (1, 112 mm SL, Rio Guama, Pinar del Rio Province), UMMZ 171880 (3,

112-124 mm SL, Uiña les, Pinar del Rio Province), UMMZ 177285 (1, 118 mm SL,

Pinar del Rio), USNM 33642 (1, 98 mm SL, no data). USNM 63995 (2, 65-88 mm SL,

no data)

73 LITERATURE CITED

Barel, C.D.N., M.J.P. van Oijen, F. Witte, and Witte-Maas, E.L.M. (1977) An introduction to the taxonomy and morphology of the Haplochromine Cichlidae from Lake Victoria. Netherlands Journal of Zoology 27, 333-389.

Barlow, G.W. and Munsey, J.W. (1976) The Red Devil-Midas-Arrow Cichlid species complex in Nicaragua. In: Investigations of the Ichthyofauna of Nicaraguan Lakes (ed T.B. Thorsin), University of Nebraska, Lincoln, pp. 359-370.

Bowin, C. (1975) The geology of Hispaniola. In: The Ocean Basins and Margins (eds A.E.M. Narin and F.G. Stehli), Plenum, New York, pp. 501-552.

Burgess, G.H. and Franz, R. (1989) Zoogeography of the Antillean freshwater fish fauna In: Biogeography of the West Indies: Past, Present, and Future (ed C.A.Woods) Sandhill Crane Press, Gainesville, pp. 263-304.

Bussing, W.A. (1985) Patterns of distribution of the Central American Ichthyofauna. In: The Great American Biotic Interchange (eds. F.G. Stehli and S.D. Webb) Plenum, New York, pp. 453-473.

Chakrabarty, P. (2006) Systematics and Historical Biogeography of Greater Antillean Cichlidae. Molecular Phylogenetics and Evolution 39, 619-627.

Cockerell, T.D.A. (1924) A fossil cichlid fish from the Republic of Haiti. Proceedings of the United States National Museum 63, 1-3.

Cooke, C.W., Gardner, J. and Woodring, W.P. (1943) Correlation of the Cenozoic formations of the Atlantic and Gulf coastal plain and the Caribbean region. Bulletin of the Geological Society of America 54, 1713-1723.

Gill, T. (1862) Remarks on the relations of the genera and other groups of Cuban fishes. Proceedings of the Academy of Natural Sciences Philadelphia 14, 235-242.

Goodall, C. (1991) Procrustes Methods in the Statistical Analysis of Shape. Journal of the Royal Statistical Society 53, 285-339.

Graham, A. (1990) Late Tertiary microfossil flora from the Republic of Haiti. American Journal of Botany 77, 911-926.

74 Kullander, S.O. (1983) A revision of the South American cichlid genus Cichlasoma (Teleostei: Cichlidae). The Swedish Museum of Natural History, Stockholm. pp.296.

Kullander, S.O. (2003) Family Cichlidae (Cichlids). In: (eds R.E. Reis, S.O. Kullander and C.J. Ferraris Jr.) Check List of the Freshwater Fishes of South and Central America. Edipucrs, Porto Alegre, pp. 605-654.

Ladiges, W. (1938) Cichlasoma vombergi spec. nov., eine zweite rezente Cichliden-Art von Santo Domingo. Zoologisher Anzeiger 123, 18-20.

Lydeard, C., Wooten, M. and Meyer, A. (1995) Molecules, morphology, and area cladograms: a cladistic and biogeographic analysis of Gambusia (Teleostei:Poeciliidae). Systematic Biology 44, 221-236.

Murray, A.M. (2001) The fossil record and biogeography of the Cichlidae (Actinopterygii: Labroidei). Biological Journal of the Linnean Society 74, 517- 532.

rd Nelson, J.S. (1994) Fishes of the World, 3 edition. John Wiley and Sons, Inc., New York, NY.

Rohlf, F. J. (1998) TPSdig. State University of New York. Buffalo, NY. available at http://life.bio.sunysb.edu/ee/rohlf/software.html.

Rohlf, F. J. and D.E. Slice (1990) Extensions of the Procrustes method for the optimal superimposition of landmarks. Systematic Zoology 39, 40–59.

Rosen, D.E. (1975) The vicariance model of Caribbean biogeography. Systematic Zoology 24, 431-464.

Sheets, H.D. (2001) PCAgen available at http://www2.canisius.edu/~sheets/morphsoft.html

Tee-Van, J. (1935) Cichlid fishes in the West Indies with especial reference to Haiti, including the description of a new species of Cichlasoma. Zoologica 10, 281-300.

Valenciennes, M. (1831) In: Cuvier, G.B. and M. Valenciennes. Histoire Naturelle des Poisons Volume 7. Strasbourg, Imprimerie de F.G. Leverault, Paris.

Van Couvering, J.A.H. (1982) Fossil cichlid fishes of Africa. Special Papers in Paleontology, 29. The Paleontological Association, London. 103 pp.

Williams, E.E. (1989) Old problems and new opportunities in West Indian Biogeography. In: Biogeography of the West Indies: Past, Present, and Future (ed C.A.Woods) Sandhill Crane Press, Gainesville, pp. 1- 46.

75 CHAPTER III

SYSTEMATICS AND HISTORICAL BIOGEOGRAPHY OF GREATER ANTILLEAN CICHLIDAE

“The geology is in many respects uncertain, the phyletic analysis inadequate and the fossil record wretched. We have if not the worst case scenario definitely a very bad one.” - E.E Williams (1989) discussing the state of Caribbean biogeography

ABSTRACT

A molecular phylogenetic analysis recovers a pattern consistent with a drift vicariance scenario for the origin of Greater Antillean cichlids. This phylogeny, based on mitochondrial and nuclear genes, reveals that clades on different geographic regions diverged concurrently with the geological separation of these areas. Middle America was initially colonized by South American cichlids in the Cretaceous, most probably through the Cretaceous Island Arc. The separation of Greater Antillean cichlids and their mainland Middle American relatives was caused by a drift vicariance event that took place when the islands became separated from Yucatan in the Eocene. Greater Antillean cichlids are monophyletic and do not have close South American relatives. Therefore, the alternative hypothesis that these cichlids migrated via an Oligocene landbridge from

South America is falsified. A marine dispersal hypothesis is not employed because the drift vicariance hypothesis is better able to explain the biogeographic patterns, both temporal and phylogenetic.

76 INTRODUCTION

The biogeographic history of the Greater Antilles has been a contentious issue

among biologists and geologists. Both disciplines have major camps that support different hypotheses. Biologists have relied on geological reconstructions to frame their hypotheses about the movement of organisms. Modern geological reconstructions that

explain the presence of the biota on the Greater Antilles fit into two major categories.

One category suggests South American origins from an Oligocene landbridge that

connected South America to the islands (Iturralde-Vinent and MacPhee, 1999). The

other category suggests Middle American origins from a period of coalescence between

these islands and Yucatan in the early Cenozoic (Pitman et al.,1993; Pindell, 1994;

updated from Malfait and Dinkelman, 1972; Tedford, 1974). Biologists have argued that these reconstructions explain the colonization of the Greater Antilles (Rosen, 1975;

Murphy and Collier, 1996; Dávalos, 2004). Biologists have also argued that marine dispersal explains the presence of the biota on these islands (Hedges, 1996; Hedges et al.,

2002; Martin and Bermingham, 1998; Glor et al., 2005). Cichlids have played a major role on all sides of these arguments (Briggs, 1984, 2003; Rivas, 1986; Sparks and Smith,

2005). However, until now, no phylogenetic assessment of these fishes has been done.

Cuba has two cichlid species (Nandopsis tetracanthus, N. ramsdeni) and

Hispaniola one extant (Nandopsis haitiensis) and one fossil species (N. woodringi).

Nandopsis vombergae (Ladiges, 1938) is a junior subjective synonym of N. haitiensis and

will not be discussed (Chakrabarty, 2006).

The islands that compose the Greater Antilles do not all share a geological history. The islands of Cuba, Hispaniola, Puerto Rico, and Jamaica form the Greater

77 Antilles. However, geologically, Jamaica does not share an arc history with the other

Greater Antillean islands (Pindell and Barrett, 1990). Because cichlids are only on Cuba

and Hispaniola, these islands will be the setting for the reconstructions described here.

The complex geological history of the Caribbean will be described before biogeographic hypotheses are tested (see also Table III.1). The Caribbean region formed as a product of the separation of Gondwana and Laurasia, particularly the separation of

North and South America 170 million years ago (Pindell, 1994; Iturralde-Vinent and

MacPhee, 1999). Most geologists agree that the Caribbean plate originated in the Pacific

(Ross and Scotese, 1988; Pindell and Barrett, 1990; Pindell, 1994; but see Meschede and

Frisch, 1998). Landmasses that originated with the formation of this plate include parts of Cuba, the Cayman Ridge, Hispaniola, Puerto Rico and the Virgin Islands (Pindell and

Barrett, 1990). These landmasses collectively formed what will be called here the

Cretaceous Island Arc (following the convention of Iturralde-Vinent and MacPhee,

1999). As this arc drifted eastward it became positioned between North and South

America. During periodic dry periods 70 to 80 million years ago this arc may have served as a landbridge which could have acted as a corridor between the two continents

(Iturralde-Vinent and MacPhee, 1999). The Cretaceous arc broke-up at the end of the

Cretaceous with its remnants forming the Paleogene arc (Iturralde-Vinent and MacPhee,

1999; Kerr et al., 1999). The Paleogene arc contained parts of Cuba and Hispaniola. In the Paleogene (early Cenozoic) this arc moved into a position that connected it to the

Yucatan. Geological reconstructions by Pitman et al. (1993) argue that the connection between the Paleogene arc and Middle America may have lasted until 49 million years ago. It is this period of coalescence that could have allowed faunal exchange between

78 these landmasses (named here as the Paleogene arc drift vicariance scenario). The separation between the North America Plate and the Caribbean Plate is the Cayman

Trough which lies between Cuba and Yucatan. The Cayman Trough began to form in the

Eocene (Pindell et al., 1988; Pindell and Barrett, 1990). Since the Eocene, Cuba and

Hispaniola (as remnants of the Paleogene arc) drifted 1100 kilometers to their current

positions. Cuba and Hispaniola separated 20 to 25 million years ago with the formation

of the Oriente Fault (Pindell, 1994).

A major alternative to the Paleogene arc drift vicariance scenario proposes a

South American origin for the Greater Antillean fauna. Iturralde-Vinent and MacPhee

(1999) propose a short-lived landbridge between the Greater Antilles and northwest

South America circa 32 million years ago. The authors name this Early Oligocene landbridge GAARlandia (from Greater Antilles + Aves Ridge). One consequence of this

alternative hypothesis is that the Greater Antillean Island chain would have had a more

recent connection with South America than with Middle America.

These alternative biogeographic hypotheses will be tested under a phylogenetic

framework. The relationships among the cichlids of Middle America, South America and

the Greater Antilles will elucidate the history of the origins of these fishes.

MATERIALS AND METHODS

Acquisition of DNA Dataset

A molecular phylogeny of 30 cichlid taxa (listed in Table III.1) was completed

using portions of nuclear genes S7 and Tmo-4C4, as well as portions of mitochondrial

genes cytochrome c oxidase subunit I (COI) and 16S. The final data set was 2278

79 aligned positions. Primers S7RPEX1F 5’-TGGCCTCTTCCTTGGCCGTC-3’ and

S7RPEX2R 5’-AACTCGTCTGGCTTTTCGCC-3’ were used to amplify and sequence the first intron in the nuclear S7 ribosomal protein gene, yielding sequences of 774

aligned positions (Chow and Hazama, 1998; Lavoué et al., 2003). Primers Tmo-f2-5’ 5’-

ATCTGTGAGGCTGTGAACTA-3’ (Lovejoy, 2000) and Tmo-r1-3’ 5’-

CATCGTGCTCCTGGGTGACAAAGT-3’ (Streelman and Karl, 1997) were used to

amplify and sequence a portion of the nuclear gene Tmo-4C4, yielding sequences of 299

aligned positions. Primers 16S ar-L 5’-CGCCTGTTTATCAAAAACAT-3’ and 16S br-

H 5’-CCGGTCTGAACTCAGATCACGT-3’ (Koucher et al., 1989; Palumbi, 1996) were

used to amplify and sequence a fragment of mitochondrial large ribosomal subunit 16S,

yielding sequences of 614 aligned positions. Primers COIfor 5’-

TTCTCGACTAATCACAAAGACATYGG-3’ and COIrev 5’-

TCAAARAAGGTTGTGTTAGGTTYC-3’ were designed in this study (modified from

the primers of Folmer et al., 1994) to amplify and sequence a segment of mitochondrial

gene COI, yielding sequences of 591 aligned positions.

Tissue samples were taken from specimens preserved as vouchers in the

University of Michigan Museum of Zoology (UMMZ) Fish Division. Voucher and

GenBank accession numbers are listed in Table III.1. Locality data for specimens can be

obtained by searching the UMMZ fish collection catalogue. All specimens are either

wild caught or purchased from a breeder raising wild caught individuals and selling their

young (Jeff Rapps; http://www.tangledupincichlids.com). Fish tissues are preserved in

95% ETOH and stored in -80° C. Tissue extraction was done using a Qiagen Tissue

Extraction Kit following the manufacturer’s protocol. PCR amplifications were done for

80 30-35 cycles. Denaturation of 20 seconds at 95° C was followed by annealing for 15

seconds at temperatures of 60° C (S7), 50° C (Tmo-4C4), 45° C (COI). Extension times

varied from 1 min 30 seconds, to 2 minutes. This extension was followed by a terminal

extension for 7 minutes at 72° C. PCR amplification of 16S follows the protocol of

Sparks (2004). PCR product was isolated on 1% agarose gels. Bands were removed

from the gel under a UV light and extracted using Qiagen Gel Extraction Kits following

the manufacturer’s protocol. Sequencing was completed by the University of Michigan

Sequencing Core Facility. DNA sequences were edited from chromatograms and aligned manually in Sequence Navigator (Elmer, 1995).

Phylogenetic Analyses and Support Indices

Parsimony analyses were completed in PAUP* 4.0b (Swofford, 2002). Heuristic

searches were performed with 10,000 random addition replicates. Bremer support

(Bremer, 1995) was calculated using TreeRot v.2 (Sorenson, 1999). Jackknife

resampling (100 replicates, 100 search replicates) and the parsimony ratchet (to verify

PAUP* results) was performed in NONA (Goloboff, 1993) and WinClada (Nixon, 1999).

The outgroup, Paratilapia polleni, was used to root all trees.

Likelihood analyses were performed in MrBayes 3.01 (Huelsenbeck and

Ronquist, 2001). MrModeltest (Nylander, 2002) was used for parameter estimation for

each gene and in combination using the hierarchical log-likelihood ratio tests. Four

Markov Chains were run for six million generations, sampling every 500 generations.

Burn-in time of 1.5 million generations was determined from where likelihood scores

reached stationarity. Trees were filtered in PAUP* 4.0b (Swofford, 2002) under a

maximum likelihood optimality criterion to recover the best trees under that framework.

81 Table III.1: Taxa sequenced, with GenBank and UMMZ catalogue numbers.

UMMZ S7 Tmo- 16S COI Taxon # 4C4

Middle America Archocentrus centrarchus 243177 DQ119165 DQ119164 DQ119162 DQ119163 Archocentrus multispinosus 243207 DQ119253 DQ119224 DQ119166 DQ119195 Archocentrus nigrofasciatus 243200 DQ119254 DQ119225 DQ119167 DQ119196 Archocentrus octofasciatus 243175 DQ119255 DQ119226 DQ119168 DQ119197 Amphilophus citrinellus 243174 DQ119256 DQ119227 DQ119169 DQ119198 Amphilophus lyonsi 243179 DQ119257 DQ119228 DQ119170 DQ119199 “Cichlasoma” salvini 243182 DQ119258 DQ119229 DQ119171 DQ119200 Herichthys carpintis 243199 DQ119259 DQ119230 DQ119172 DQ119201 Hypsophrys nicaraguensis 243188 DQ119260 DQ119231 DQ119173 DQ119202 Parachromis managuensis 243204 DQ119261 DQ119232 DQ119174 DQ119203 Parachromis dovii 243205 DQ119262 DQ119233 DQ119175 DQ119204 Parachromis motaguensis 243183 DQ119263 DQ119234 DQ119176 DQ119205 Petenia splendida 243170 DQ119264 DQ119235 DQ119177 DQ119206 Thorichthys aureus 243202 DQ119265 DQ119236 DQ119178 DQ119207 Tomocichla sieboldii 243171 DQ119266 DQ119237 DQ119179 DQ119208 Vieja synspila 243203 DQ119267 DQ119238 DQ119180 DQ119209 Vieja tuyrense 243180 DQ119268 DQ119239 DQ119181 DQ119210 Greater Antilles Nandopsis ramsdeni 245137 DQ119269 DQ119240 DQ119182 DQ119211 Nandopsis tetracanthus 245598 DQ119270 DQ119241 DQ119183 DQ119212 Nandopsis haitiensis 243287 DQ119271 DQ119242 DQ119184 DQ119213 South America Apistogramma bitaeniatum 243211 DQ119272 DQ119243 DQ119185 DQ119214 Bujurquina vittata 243206 DQ119273 DQ119244 DQ119186 DQ119215 “Cichlasoma” festae 243201 DQ119274 DQ119245 DQ119187 DQ119216 Geophagus steindachneri 243208 DQ119275 DQ119246 DQ119188 DQ119217 Heros appendiculatus 243189 DQ119276 DQ119247 DQ119189 DQ119218 Hypselecara temporalis 243197 DQ119277 DQ119248 DQ119190 DQ119219 Uaru amphiacanthoides 243176 DQ119278 DQ119249 DQ119191 DQ119220 India Etroplus maculatus 245135 DQ119279 DQ119250 DQ119192 DQ119221 Madagascar Paratilapia polleni 243192 DQ119280 DQ119251 DQ119193 DQ119222 Paretroplus kieneri 243195 DQ119281 DQ119252 DQ119194 DQ119223

82 Parametric bootstrapping was implemented in order to statistically test a South

American origin of the Greater Antillean cichlids. Trees were searched under the

topological constraint that the Greater Antillean cichlids must have a sister relationship with South American cichlids. Using the best fit model of sequence evolution selected

from Modeltest (Posada and Crandall, 1998), branch lengths were optimized on the

constrained tree. SG Runner (Wilcox, 2005) and Seq-Gen (Rambault and Grassly, 1997)

was then used to simulate 1000 data sets on the constrained topology using the same

model of sequence evolution. The optimal tree for each dataset was found using PAUP*

as was the optimal trees for each dataset under the constraint of the Greater Antillean cichlids being South American. Tree lengths were compared across constrained and unconstrained trees for each dataset. Significance was assessed by comparing the

difference in the actual data set to the simulated datasets.

Date Estimation and Calibration

In the absence of a strict molecular clock a penalized likelihood approach was

used for estimating divergence times. Penalized likelihood combines likelihood based

substitution models with a penalty term to allow varying (but constrained) rates of change across a phylogeny. By incorporating the likelihood term of the substitution model this

method also remains consistent with the method used for recovering the tree topology.

Penalized likelihood was implemented in R8S 1.7 (Sanderson, 2003) to estimate

divergence dates of internal nodes.

Determining the optimal level of constrained variation across branches (termed

“rate smoothing”) is accomplished through cross validation (Sanderson, 2002). Cross-

validation iteratively removes a terminal branch and compares estimated values for that

83 branch to the observed value. The cross validation score is the differences in observed

and estimated branch lengths that are summed across the tree. The lowest cross

validation score is the optimal value. The additive penalty function was applied to

penalize squared differences in rates across branches. This penalty function is the

appropriate option when calibration points are deeper in the tree than the nodes to be

estimated (Sanderson, 2004).

R8s has the advantage over other programs estimating divergence times in

allowing calibration points to be set as minimum, maximum or fixed ages rather than

only fixed ages (see Heads, 2005 for a discussion). Because of the nature of the

geological and fossil evidence available for this study, the flexibility allowed by not

fixing absolute dates on calibration points was essential.

Three calibration points were chosen to put a temporal scale on the phylogenetic hypothesis. The minimum age of the Greater Antillean cichlids was placed at 5 million years (node B, Fig.III.1), because Nandopsis woodringi is an extinct member of the

Greater Antillean endemic genus Nandopsis. This species, described from Las Cohobas,

Haiti (Cockerell, 1924), is Late Miocene in age (11.6 – 5.3 million years ago; see

Chakrabarty, 2006 about incorrect dates in literature). Because there is a paucity of fossils that can be placed on lineages in the current phylogeny, calibration points from

Gondwanan vicariance events were also used. Cichlids are distributed mainly on former

Gondwanan fragments (India, Madagascar, South America and Africa). The relationships among members of this family reflect the break up of Gondwana (Sparks and Smith, 2005; see review of cichlid phylogenies in Chakrabarty, 2004). Traditional molecular clock evidence is equivocal (Kumazawa et al., 2000; Vences et al., 2001) and

84 there is some question about the methods used (Chakrabarty, 2004; Sparks and Smith,

2005). The use of Gondwanan vicariance ages as calibration points to test a Greater

Antillean vicariance scenario allows the use of rates to estimate dates. Without these

independent calibration points the penalized likelihood approach would lack maximum

age calibration points. These points are necessary to prevent nodal age estimates from

estimating ages infinitely back in time. The super-continent of Gondwana began to break

up circa 165 million years ago. This age is the minimum age for the origin of Cichlidae

if they were present before fragmentation. For this reason a 165 million year fixed

calibration point is placed at the base of the phylogeny. A second fixed calibration point

is placed in the separation of India and Madagascar that took place 88 million years ago.

This separation is represented on the phylogeny by the separation of the Indian genus

Etroplus and the Malagasy genus Paretroplus. Both these fixed calibration points are

associated with outgroup lineages. Their function is to serve as anchor points from which

dates can be estimated for divergences in the Neotropical ingroup.

All the nucleotide data were treated simultaneously for the penalized likelihood

approach as they were in the phylogenetic methods (both parsimony and maximum

likelihood). The advantage of multigene data sets are that they contain more information

than single gene data sets, much of the information pertaining to divergence times will be

lost if the dataset is reduced to a single gene or part of a single gene (Thorne and Kishino,

2002; Yang and Yoder, 2003). Rather than disregard large portions of the dataset by

pruning taxa or removing gene sequences the assumption of a constant rate of evolution

was relaxed. Constant rate analysis is not rigorous because it does not recognize the

uncertainty in divergence time estimation (Thorne and Kishino, 2002). Evolutionary

85 rates differ over time and among genes but as a phylogeny shows, these genes share a

common set of divergence times.

RESULTS

Model Selection, Likelihood Assumption Set

For parametric bootstrapping and all maximum likelihood PAUP* analyses,

ModelTest selected the following parameters for the combined dataset: GTR + G + I

model of sequence evolution, with four rate categories, base frequencies (A=0.266,

C=0.257, G=0.199, T=0.278), rate heterogeneity according to the gamma distribution

with a shape (α) = 0.511, and the proportion of invariable sites (pinv) = 0.3143.

For analyses using MrBayes, parameter estimation was selected from

MrModelTest to be partitioned by genes to have six substitution sites under a GTR

model. Both 16S and Tmo-4C4 were selected to have a proportion of the sites invariable while the rates for the remaining sites are drawn from a gamma distribution. Both COI and S7 were selected to have rates at every site drawn from a gamma distribution. The

GTR matrix, gamma distribution, nucleotide state frequencies, proportion of invariant

sites, and the transition/transversion ratio were all unlinked across data partitions.

Phylogenetic Analyses and Support

Combined analyses of gene fragments from S7, Tmo-4C4, 16S and COI resulted in a single most parsimonious tree that was fully congruent with the maximum likelihood analysis (Fig.III.1). A tree length of 2682 was obtained with a consistency index of .516, a retention index of .464, and a rescaled consistency index of .239. The consistency index excluding uninformative sites was .409. Five hundred and fifty-nine characters

86 were parsimony informative. The score of the best tree found under the maximum

likelihood framework was 16,836.29518 (Fig. III. 2). (Phylogenies based on individual

genes are shown in Chapter IV.)

The Greater Antillean cichlids are recovered as a monophyletic group. The

Cuban cichlids, Nandopsis tetracanthus and N. ramsdeni, are sister to the Hispaniolan species. The Greater Antillean cichlids are nested within a large clade of mainly Middle

American cichlids. The sister group to the Greater Antillean cichlids is a large group of

widespread Middle American species.

The phylogenetic tree shows one reversal of a Central American cichlid now

endemic to South America. That species, “Cichlasoma” festae, is phylogenetically

recovered as a Middle American cichlid (from the parsimony optimization in Fig.III.1).

This species and the remaining Middle American cichlids form a clade that is nested within the sampled South American species.

The null hypothesis that the Greater Antillean cichlids have sister group

relationships with South American cichlids was rejected under parametric bootstrapping.

The unconstrained tree (the most parsimonious tree) was 131 steps shorter than the best

tree in which the clade of Greater Antillean cichlids was constrained to be South

American. This value was significantly greater than can be attributed to chance (p<.001).

The null hypothesis was similarly rejected under the maximum likelihood framework

(p<.01).

87 Fig.III.1: Phylogeny of Neotropical cichlid taxa inferred from S7, Tmo-4C4, 16S and COI sequences. Topology shown is the most parsimonious tree with geographic regions reconstructed on the phylogeny under parsimony character optimization in MacClade 4.0 (Maddison and Maddison, 1992). Bayesian posterior probabilities values that are significant (≥ 95%) are shown below nodes. Above each node Jackknife values are given if 80 percent or above followed by a front slash and Bremer support values if 3 or above. Fixed calibration points are shown (all are associated with outgroup taxa). Letters at nodes correspond to the estimated dates given in Table 2.

88 Table III.2: Estimated dates for nodes of interest with associated reference letters on the phylogeny (Fig.III.1).

Estimated Divergence Time Millions of Years ± Standard Geological Event Age of Event Deviation Youngest Oldest Node Node Middle Eocene, as Separation of Paleogene Arc late as 49 million (Cuba and Hispaniola) from years ago A B Yucatan at opening of 50 43 (Pitman et al., Caymen Trough 1993) ± 5 ± 5 Period when Cretaceous Arc Late Cretaceous, (Greater Antilles, adjacent 70-80 mya regions) may have served as (Iturralde-Vinent 66 C 55 D landbridge between North and MacPhee, and South America 1999) ± 6 ± 5

Separation of Eastern Cuba Oligocene or and Western Hispaniola Miocene, 20-25 43 25 through the formation of the mya (Pitman et al., ± 5 B ± 5 E Oriente Fault 1993)

Final separation, Separation of South 106-84 mya F G America from rest of 87 76 (Pitman et al., Gondwana 1993) ± 5 ± 6

89 Estimated Dates

Ages were estimated under the optimal smoothing value of 1.3e-05. Table III.2 shows the recovered estimated divergence times for the nodes of interest. Two sets of ages are given, one for the node associated with the youngest age that can be attributed to a group, and one for the node associated with the oldest possible age. Arguably, the more conservative age estimate is that of the youngest age because it is closest to the node of interest; however, there is no way to determine which of these ages can be attributed to a particular clade. Ages estimated for important ingroup nodes include the following: for

Cuban cichlids (node E, Fig.III.1), an origin 25 million years ago (mya) with a standard deviation of 5 million years. The clade of Greater Antillean cichlids (node B) had an estimated origin 43 mya with a standard deviation of 5 million years. The separation of the Greater Antillean cichlids and its Middle American sister group (node A) is estimated to have taken place 50 mya with a standard deviation of 5 million years. The separation between the South American clade of Heros appendiculatus and Uaru amphicanthoides with its mainly Middle American/Greater Antillean sister group (node C) is estimated to have taken place 66 mya with a standard deviation of 6 million years. The oldest age estimated for the entire Neotropical ingroup (node F) is an origin 87 mya with a standard deviation of 6 million years.

90 Fig.III.2: Maximum likelihood phylogeny of Neotropical cichlid taxa inferred from S7, Tmo-4C4, 16S and COI sequences.

91 DISCUSSION

The Greater Antillean cichlids are phylogenetically a clade of Middle American cichlids whose separation from Middle America took place through an early Cenozoic vicariance event. The phylogenetic pattern recovered shows Middle American origins for the Greater Antillean clade (Fig.III.1). The estimated ages for the origin of this clade correspond to the time of separation of the Paleogene arc and the Yucatan peninsula.

Therefore, the Paleogene arc drift vicariance scenario is supported by the phylogenetic pattern and its temporal scale. In this scenario, a vicariance event separated populations of an ancestral Middle American species that inhabited a contiguous area shared by the

Paleogene arc and Yucatan. The drifting of the Paleogene arc led to the allopatric speciation event that gave rise to the Greater Antillean cichlids.

There are no close South American relatives recovered for the Greater Antillean cichlids, rejecting the GAARlandia hypothesis, which predicted South American sister group relationships. The temporal scale fit on the phylogeny also does not correspond to the geological events assumed in the GAARlandia hypothesis. The Greater Antilles could only have been connected to South America through GAARlandia 32 mya, an age at least six million years too young according to the temporal scale of the recovered phylogeny.

Another vicariant event, caused by the separation of Cuba and Hispaniola, was also revealed by the temporal analysis. The 20 to 25 million year old separation of these two islands is concordant with the time of separation of the Cuban and Hispaniolan cichlids.

92 The age of the Middle American cichlid fauna is also recovered. Bussing (1985) suggested a late Cretaceous or early Tertiary origin for Middle American cichlids based on the patterns of the endemic fauna of this region. Bussing’s (1985) hypothesis appears

to be corroborated by the estimated divergence dates, which correspond to the periods

when the Cretaceous arc could have served as a corridor between North and South

America (Iturralde-Vinent and MacPhee, 1999). Middle America (essentially just

Yucatan in the Cretaceous) would have been colonized during this period before break-up

of the Cretaceous arc. Cichlids would have then dispersed onto the Chortis block

(Honduras-El Salvador-Southern Guatemala) when it connected with Yucatan in the

Eocene. Likewise, as the remainder of modern lower Central America formed, cichlids

would have dispersed south onto these regions. Therefore, it appears that the Cretaceous

arc served as a corridor for cichlids to cross from South America to Yucatan; however,

this arc apparently never functioned to maintain a refugium-like habitat for cichlids as the

Paleogene arc did.

This Cretaceous scenario opposes the Miocene marine dispersal view of some

researchers (Martin and Bermingham, 1998; Myers, 1966) to explain the origins of the

Middle American Cichlidae. The Miocene dispersal evaluation of Myers (1966) was

based mainly on the presence of the fossil Nandopsis woodringi on Hispaniola. Martin

and Bermingham (1998) used a traditional molecular clock to conclude that the origins of

nearly 100 species of Middle American cichlids can be explained by a single Miocene

dispersal event. Perdices, Doadrio and Bermingham (2005), using a traditional molecular

clock, concluded that synbranchid eels also dispersed at this time. Dispersal can occur at

any time but the reason that these radiations took place at nearly the same time was not

93 explored. A Miocene radiation of these fishes would have required crossing of a

significant marine barrier between northern South America and nuclear Central America.

Cichlids are not known to cross marine barriers to colonize landmasses (Sparks and

Smith, 2005; Riseng, 1997). However, the possibility that individuals of a single species

from each group was able to cross this barrier can never be ruled out.

No phylogenetic pattern can reject a marine dispersal hypothesis for the origin of

Greater Antillean cichlids as such hypotheses cannot be refuted. The phylogenetic

relationships found in the current study are no exception. However, the period of

dispersal attributed to cichlids by the temporal analysis shows dispersal took place during coalescence of now separated landmasses. Dispersal over freshwater corridors at geological coalescence times is more plausible than the marine dispersal route at the same time. Despite the ability of some cichlids to tolerate saltwater, it appears to be a significant barrier to most (Sparks and Smith, 2005). There is no need to employ a marine dispersal hypothesis when the drift vicariance hypothesis is better able, both temporally and phylogenetically, to explain the biogeographic relationships.

The endemic South American species recovered as phylogenetically Middle

American does not effect the biogeographic hypotheses presented in this study. This species, “Cichlasoma” festae, is nested well within Middle American cichlids. The youngest age that can be attributed to its clade (composed of itself and sister group) is 47

mya with a standard deviation of 5 million years. This period corresponds to a time when

the Aves Ridge may have connected Middle America to South America (Pitman et al.,

1993). It may be at this time that this species dispersed from Middle America to South

America. Notably, Hulsey et al. (2004) also recovered “Cichlasoma” festae nested

94 within Middle American cichlids. However, it should be noted that this ancient age is

attributed to this clade of four species (which represent many more species) and not

necessarily the age of “Cichlasoma” festae, which may be far younger and part of a more

recent invasion of South America.

Complete sampling of the extant Greater Antillean Cichlidae reveals for the first

time their Middle American origins. The temporal scale fit to this phylogeny also

provides insights about three events that were important in their origins: (1) The arrival of species in Yucatan in the Cretaceous (2) followed by a drift vicariance event between the arc composed of Cuba and Hispaniola with Yucatan (3) and finally the separation of

Cuba and Hispaniola. The fit between the estimated divergence dates and these geological events cannot be ignored.

Other groups that share a congruent pattern with cichlids (in having Middle

American-Greater Antillean relationships) include: the snake genus Epicrates (Kluge,

1988), legumes in part (Lavin et al., 2003), the gar Atractosteus (Wiley, 1976), the

livebearer tribe Girardiini (Rosen, 1975, 1985; Rosen and Bailey, 1963), and three

Gambusia species groups (puncticulata, nicaraguensis, punctata; Fink, 1971, 1971b;

Rauchenberger, 1989; Lydeard et al., 1995). Murphy and Collier (1996) recover a

phylogenetic pattern and temporal scale corresponding to vicariance origins for the

aplocheiloid genus Rivulus in the Greater Antilles. They used a 70-80 mya calibration

point associated with the period that the Cretaceous arc functioned as a corridor between

North and South America. The present study finds evidence for that event in the

phylogeny of Neotropical cichlids.

95 Lacking an ability to look into the past we must attempt to reconstruct it as best we can. As biologists build stronger evidence of particular relationships, geologists must follow suit to substantiate or eliminate possible reconstructions that explain those patterns. Reciprocal illumination may work slowly across disciplines but the field of historical biogeography demands that both biologists and geologists keep pace with each other.

96 LITERATURE CITED

Bremer, K. (1995) Branch support and tree stability. Cladistics 10, 295–304.

Briggs, J.C. (1984) Freshwater fishes and biogeography of Central America and the Antilles. Systematic Zoology 33, 428-435.

Briggs, J.C. (2003) Fishes and birds: Gondwana life rafts reconsidered. Systematic Biology 52, 548–553.

Bussing, W.A. (1985) Patterns of distribution of the Central American Ichthyofauna. In: The Great American Biotic Interchange Volume 4, Topics in Geobiology (eds F.G. Stehli and S.D. Webb), Plenum Press, New York, pp. 453–473.

Chakrabarty, P. (2004) Cichlid biogeography: comment and review. Fish and Fisheries 5, 97–119.

Chakrabarty, P. (2006) Taxonomic status of the Hispaniolan Cichlidae. Occasional Papers Museum of Zoology, University of Michigan 737, 1-16.

Chow, S. and Hazama, K. (1998) Universal PCR primers for S7 ribosomal protein gene introns in fish. Molecular Ecology 7, 1247–1263.

Cockerell, T.D.A. (1924) A fossil cichlid fish from the Republic of Haiti. Proceedings of the U.S. National Museum 63, 1–3.

Dávalos, L.M. (2004) Phylogeny and biogeography of Caribbean mammals. Biological Journal of the Linnean Society 81, 373–394.

Elmer, P. (1995) Sequence Navigator v.1.0.1. Applied Biosystems Inc.

Fink, W.L. (1971) A revision of the Gambusia puncticulata complex (Pisces:Poeciliidae). Publications of the Gulf Coast Research Laboratory Museum 2, 11–46.

Fink, W.L. (1971) A revision of the Gambusia nicaraguensis species group (Pisces:Poeciliidae). Publications of the Gulf Coast Research Laboratory Museum 2, 47–77.

Folmer, O., Black, M., Hoek, W., Lutz, R. and Vrijenhoek, R. (1994) DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology 3, 294– 299.

Glor, R.E., Losos, J.B. and Larson, A. (2005) Out of Cuba: overwater dispersal and speciation among lizards in the Anolis carolinensis subgroup. Molecular Ecology 14, 2419–2432.

Goloboff, P. A. (1993) NONA 2.0 Program and documentation. Computer program distributed by J. M. Carpenter, Dept. of Entomology, American Museum of Natural History, New York.

Heads, M. (2005) Dating nodes on molecular phylogenies: a critique of molecular biogeography. Cladistics 21, 62-78.

Hedges, S.B. (1996) Historical biogeography of West Indian vertebrates. Annual Review of Ecology and Systematics 27, 163–196.

Hedges, S.B., Hass, C.A. and Maxson, L.R. (2002) Caribbean biogeography: molecular evidence for dispersal in West Indian terrestrial vertebrates. Proceedings of the National Academy of Science USA 89, 1909–1913.

Hulsey, C.D., García de León, F.J., Johnson, Y.S., Hendrickson, D.A., and Near, T.J. (2004) Temporal diversification of Mesoamerican cichlid fishes across a major biogeographic boundary. Molecular Phylogenetics and Evolution 31, 754–764.

Huelsenbeck, J. P. and Ronquist, F. (2001) MrBayes: Bayesian inference of phylogeny. Bioinformatics 17, 754–755.

Iturralde-Vinent, M.A. and MacPhee, R.D.E. (1999) Paleogeography of the Caribbean region: implications for Cenozoic biogeography. Bulletin of the American Museum of Natural History 238, 1–95.

Kerr, A.C., Iturralde-Vinent, M.A., Saunders, A.D., Babbs, T.L. and Tarney, J. (1999) New plate tectonic model of the Caribbean: Implications from a geochemical reconnaissance of Cuban Mesozoic volcanic rocks. Bulletin of the Geological Society of America 111, 2–20.

Kluge, A.G. (1988) Parsimony in vicariance biogeography: A quantitative method and a Greater Antillean example. Systematic Zoology 37, 315–328.

Koucher, T.D., Thomas, W.K., Meyer, A., Edwards, S.V., Paabo, S., Villablanca, F.X. and Wilson, A.C. (1989) Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences USA 86, 6196–6200.

Ladiges, W. (1938) Cichlasoma vombergi spec. nov., eine zweite rezente Cichliden-Art von Santo Domingo. Zoologisher Anzeiger 123, 18–20.

98 Lavin, M., Wojciechowski, M.F., Gasson, P., Hughes, C. and Wheeler, E. (2003) Phylogeny of Robinioid legumes (Fabaceae) revisited: Coursetia and Gliricidia recircumscribed, and a biogeographical appraisal of the Caribbean endemics. Systematic Botany 28, 387–409.

Lavoué, S., Sullivan, J.P. and Hopkins, C.D. (2003) Phylogenetic utility of the first two introns of the S7 ribosomal protein gene in African electric fishes (Mormyroidea: Teleostei) and congruence with other molecular markers. Biological Journal of the Linnean Society 78, 273–292.

Lovejoy, N.R. (2000) Reinterpreting recapitulation: systematics of needlefishes and their allies (Teleostei:Beloniformes). Evolution 54, 1349–1362.

Lydeard, C., Wooten, M.C. and Meyer, A. (1995) Molecules, Morphology, and area cladograms: A cladistic and biogeographic analysis of Gambusia (Teleostei: Poeciliidae). Systematic Biology 44, 221–236.

Maddison, D.R. and Maddison, W.P. (1992) MacClade: Analysis of Phylogeny and Character Evolution. Version 3.0. Sinauer Associates, Sunderland.

Malfait, B.T. and Dinkelman, M.G. (1972) Circum-Caribbean tectonic and igneous activity and the evolution of the Caribbean Plate. Bulletin of the Geological Society of America 83, 251–272.

Martin, A.P. and Bermingham, E. (1998) Systematics and evolution of Lower Central American cichlids inferred from analysis of cytochrome b gene sequences. Molecular Phylogenetics and Evolution 9, 192–203.

Meschede, M. and Frisch, W. (1998) A plate-tectonic model for the Mesozoic and Early Cenozoic history of the Caribbean plate. Tectonophysics 296, 269–291.

Murphy, W.J. and Collier, G.E. (1996) Phylogenetic relationships within the Aplocheiloid fish genus Rivulus (Cyprinodontiformes, Rivulidae): Implications for Caribbean and Central American biogeography. Molecular Biology and Evolution 13, 642–649.

Myers, G.S. (1966) Derivation of the freshwater fish fauna of Central America. Copeia 1966, 766–772.

Nixon, K.C. (1999) Winclada (BETA) ver. 0.9.9. Published by the author. http://www.cladistics.com/about_winc.htm

Nylander, J.A.A. (2002) MrModeltest v1.1b. Published by the author. http://www.ebc.uu.se/systzoo/staff/nylander.html

99 Palumbi, S.R. (1996) Nucleic acids II: the polymerase chain reaction. In: Molecular Systematics, Second Edition (eds D.M. Hillis, C. Moritz, and B.K. Mable), Sinauer, Sunderland, MA, pp. 205–247.

Perdices, A., Doadrio, I. and Bermingham, E. (2005) Evolutionary history of the synbranchid eels (Teleostei: Synbranchidae) in Central America and the Caribbean islands inferred from their molecular phylogeny. Molecular Phylogenetics and Evolution 37, 460-473.

Pindell, J.L. (1994) Evolution of the Gulf of Mexico and the Caribbean. In: Caribbean Geology, an Introduction (eds Donovan, S.K. and T.A. Jackson) The University of the West Indies Publishers’ Association, Kingston, pp.13–39.

Pindell, J.L., Cande S.C., Pitman III, W.C., Rowley, D.B., Dewey, J.F., LaBrecque, J. and Haxby, W. (1988) A plate-kinematic framework for models of Caribbean evolution. Tectonophysics 155, 121–138.

Pindell, J.L. and Barrett, S.F. (1990) Geological evolution of the Caribbean: A plate tectonic perspective. In: The Geology of North America. Vol.H., The Caribbean Region (eds J.E. Case and G. Dengo) A decade of North American Geology Series. Boulder, Colorado: Geological Society of America, pp. 405–432.

Pitman, W.C., III, Cande, S.C., LaBrecque, J. and Pindell, J.L. (1993) Fragmentation of Gondwana: the separation of Africa from South America. In: Biological Relationships Between Africa and South America (ed P.Goldblatt), Yale University Press, New Haven, pp 15–34.

Posada, D. and Crandall, K.A. (1998) Modeltest: testing the model of DNA substitution Bioinformatics 14, 817–818.

Rambaut, A. E. and Grassly, N.C. (1997) Seq-Gen an application for the Monte Carlo simulation of DNA sequence evolution along phylogenetic trees. Computer Applications in Biosciences 13, 235–238.

Rauchenberger, M. (1988) Historical biogeography of poeciliid fishes in the Caribbean Systematic Zoology 37, 356–365.

Riseng, K.J. (1997) The distribution of fishes and the conservation of aquatic resources in Madagascar. Unpublished M.Sc.thesis, University of Michigan, Ann Arbor.

Rivas, L.R. (1986) Comments on Briggs (1984): Freshwater Fishes and Biogeography of Central America and the Antilles. Systematic Zoology 35, 633-639.

Rosen, D.E. (1975) A vicariance model of Caribbean biogeography. Systematic Zoology 24, 431–464.

100 Rosen, D.E. (1985) Geological hierarchies and biogeographic congruence in the Caribbean. Annals of the Missouri Botanical Garden 72, 636–659.

Rosen, D.E. and Bailey, R.M. (1963) The poeciliid fishes (Cyprinodontiformes), their structure, zoogeography, and systematics. Bulletin of the American Museum of Natural History 126, 1–176.

Ross, M.I. and Scotese, C.R. (1988) A hierarchical tectonic model of the Gulf of Mexico are Caribbean region. Tectonophysics 155, 139–168.

Sanderson, M.J. (2002) Estimating absolute rates of molecular evolution and divergence times: a penalized likelihood approach. Molecular Biology and Evolution 19, 101-109.

Sanderson, M.J. (2003) r8s: inferring absolute rates of molecular evolution and divergence times in the absence of a molecular clock. Bioinformatics 19, 301– 302.

Sanderson, M.J. (2004) r8s, version 1.70 User’s Manual (December 2004) available online at http://ginger.ucdavis.edu/r8s/

Sorenson, M.D. (1999) TreeRot, v. 2. Boston University, Boston, Massachusetts. Available from http://people.bu.edu/msoren/TreeRot.html

Sparks, J.S. (2004) Molecular phylogeny and biogeography of the Malagasy and South Asian cichlids (Teleostei: Perciformes: Cichlidae). Molecular Phylogenetics and Evolution 30, 599–614.

Sparks, J.S. and Smith, W.L. (2005) Freshwater fishes, dispersal ability, and nonevidence: “Gondwana life rafts” to the rescue. Systematic Biology 54, 158– 165.

Streelman, J.T. and Karl, S.A. (1997) Reconstructing labroid evolution with single-copy nuclear DNA. Proceedings of the Royal Society of London, B. 264, 1011–1020.

Swofford D. L. (2002) PAUP* (phylogenetic analysis using parsimony * and other methods). Sinauer, Sunderland, Massachusetts, USA.

Tedford, R.H. (1974) Marsupials and the new paleogeography. In: Paleogeographic provinces and provinciality (ed C.A. Ross) Society of Economic Paleontologists and Mineralogists Special Publications 21, pp. 109–126.

Thorne, J.L. and Kishino, H. (2002) Divergence time and evolutionary rate estimation with multilocus data. Systematic Biology 51, 689-702.

101 Vences, M., Freyhof, J., Sonnenberg, R., Kosuch, J. and Veith, M. (2001) Reconciling fossils and molecules: Cenozoic divergence of cichlid fishes and the biogeography of Madagascar. Journal of Biogeography 28, 1091–109.

Wilcox, T. (2005) SG Runner 2.0, Published by the author. http://homepage.mac.com/tpwilcox/SGRUNNER/FileSharing8.html

Wiley, E.O. (1976) The phylogeny and biogeography of fossil and recent gars (Actinopterygii: Lepisosteidae). Miscellaneous Publications of the University of Kansas Museum of Natural History 64, 1–111.

Williams, E.E. (1989) Old problems and new opportunities in West Indian biogeography. In: Biogeography of the West Indies; Past, Present, and Future (ed E.E. Williams) Sandhill Crane Press, Inc., Gainesville, pp. 1–46.

Yang, Z. and Yoder, A.D. (2003) Comparison of Likelihood and Bayesian methods for estimating divergence times using multiple gene loci and calibration points, with application to a radiation of cute-looking mouse lemur species. Systematic Biology 52, 705-716.

102 CHAPTER IV

A MORPHOLOGICAL PHYLOGENETIC ANALYSIS OF NEOTROPICAL CICHLIDS IN THE SECTION ‘NANDOPSIS’ SENSU REGAN

“On the whole it is not satisfactory to have one-third of the Neotropical cichlid fauna without a scientific name.” - Kullander (1998)

ABSTRACT

A morphological phylogenetic analysis is presented for 41 Neotropical cichlids.

Two South American species, Cichla ocellaris and Crenicichla saxatilis, are outgroups.

Most of the remaining ingroup taxa are Middle American (Central American + Mexican) with emphasis on the 20 species Regan placed in the informal Cichlasoma section

‘Nandopsis.’ The resulting consensus phylogeny of four most parsimonious trees

recovers monophyletic Nandopsis, Vieja, Tomocichla, Archocentrus and Thorichthys.

The 20 species of the section ‘Nandopsis’ did not form a clade.

INTRODUCTION

The roughly 100 cichlid species described from Middle America have been in

taxonomic limbo for nearly a century. The catch-all genus Cichlasoma held the majority

of Middle American cichlids until relatively recently (Kullander, 1983). In order to

describe putative phylogenetic units within Cichlasoma, Regan (1906-1908) divided the

103 genus into five sections. Based on those assignments, Miller (1966, 1976) further subdivided the genus into the sections ‘Theraps,’ ‘Archocentrus,’ ‘Herichthys,’

‘Amphilophus,’ ‘Paraneetroplus,’ ‘Parapetenia’ (=‘Nandopsis’), ‘Thorichthys’ and

‘Cichlasoma incertae cedis.’ Kullander (1983) noted that Parapetenia is a junior subjective synonym of Nandopsis. (As per convention, names will be placed in single quotes when referring to sections e.g., the genus Nandopsis, the section ‘Nandopsis.’)

Little phylogenetic work supports the monophyly of these sections. Despite

Miller’s (1966, 1976) claim that “an attempt is made to list the fishes in phylogenetic sequence,” he did not provide evidence for his assignements. Presumably, Miller worked largely under Regan’s (1906-1908) definitions, which were themselves based on similarity. Roe et al. (1997) conjectured that these sections were defined on similarities in trophic morphology. Regan (1905) described ‘Nandopsis’ as the most widely distributed section of Cichlasoma, with members in the Greater Antilles, Central and

South America. It can be presumed he saw the origins of this group as Central American because he stated “The South American species of this section are probably derived from immigrants from Central America” (1906-1908). The morphological analysis of

‘Nandopsis’ presented in this study will continue the “badly need[ed] overhauling” Miller

(1966) declared these groups needed.

Few changes have been made to ‘Nandopsis’ after Miller’s (1976) work. Species recognized as members of ‘Nandopsis’ are listed in Table 1. Taylor and Miller (1980) added “Cichlasoma” grammodes to the ‘Nandopsis’ section and Bussing (1989) later described “Cichlasoma” loisellei as a member of this group. Three of the members of

‘Nandopsis’ as defined by Regan (1906-1908), “Cichlasoma” tenue, C. mento, C.

104 multifasciatum, have been reduced to junior synonyms of other species (“Cichlasoma”

salvini, C. istlanum, and Parachromis friedrichsthali respectively). Regan also mentioned “Cichlasoma” mojarra and C. centrale as members of Nandopsis (although not in his key) both are now recognized as junior subjective synonyms of “Cichlasoma” trimaculatum (Kullander, 2003). Species not from Central America, and consequently not included in Miller’s list (1966, 1976), but noted by Regan include: Herichthys steindachneri, Herichthys bartoni, and “Cichlasoma” beani. From South America

Regan included “Cichlasoma” festae, C. ornatum, Caquetaia spectabilis and Caquetaia kraussii. These Caquetaia species were removed from ‘Nandopsis’ by Kullander (1983).

The original description of ‘Nandopsis’ Regan, 1905:

The anterior pair of teeth in the upper jaw and the two on each side of the anterior pair in the lower are more or less strongly enlarged and canine-like. The mouth is usually larger, more oblique and more protractile than in other groups of Cichlasoma, the length of the lower jaw or of the premaxillary spines, from the anterior edge of the upper jaw, measuring from 2/5 to a little more than 3/5 the length of the head. The upper profile of the snout is straight; the fold of the lower lip is continuous or subcontinuous. The dorsal fin has XV-SVIII 8-13 rays and posteriorly has a scaly sheath at the base and a short series of scales on each interradial membrane; the anal has IV-IX 7-10 rays; the pectoral is shorter than the head; the caudal is rounded.

A morphological analysis is presented here to attempt to resolve some of the

long-standing taxonomic problems associated with former members of Cichlasoma.

This is the first morphological phylogenetic analysis to focus on Middle American

cichlids. Recent generic nomenclatural changes for several of these species were

presented as lists without comment or description (Kullander, 2003). The

preliminary analysis presented here provides evidence for some of these

assignments based on a phylogeny.

105 Table IV.1. Members of the ‘Nandopsis’ species group sensu Regan.

Species Distribution Amphilophus hogaboomorus Honduras “Cichlasoma” atromaculatum Panama, Columbia “Cichlasoma” beani Mexico “Cichlasoma” grammodes Mexico to Guatemala “Cichlasoma” festae Ecuador “Cichlasoma” istlanum Mexico “Cichlasoma” ornatum Ecuador and Colombia “Cichlasoma” salvini Mexico to Guatemala “Cichlasoma” trimaculatum Mexico to El Salvador “Cichlasoma” urophthalmum Mexico to Nicaragua Herichthys bartoni Mexico Herichthys steindachneri Mexico Nandopsis haitiensis Hispaniola Nandopsis ramsdeni Cuba Nandopsis tetracanthus Cuba Parachromis dovii Nicaragua, Costa Rica, Honduras Parachromis friedrichsthali Mexico to Costa Rica Parachromis loisellei Honduras to Panama Parachromis managuense Nicaragua, Costa Rica, Honduras Parachromis motaguense Guatemala, El Salvador, Honduras, Costa Rica

When Kullander (1983) restricted Cichlasoma to a dozen South American many species were given the provisional generic epithet of “Cichlasoma” (in quotes) to distinguish them from the true South American Cichlasoma (Kullander 1983, 1996).

106 MATERIALS AND METHODS

All radiographs, measurements and counts were done on the left side of each specimen. Clearing and staining of material follows Taylor and Van Dyke (1985) with

modifications by Fink (unpublished, pers. comm.).

Institutional Abbreviations

AMNH American Museum of Natural History, New York, NY

CAS California Academy of Sciences

FMNH Field Museum of Natural History, Chicago, IL

MZGJ Museo de Zoología del Grupo Jaragua, Santo Domingo, Dominican

Republic

MCZ Museum of Comparative Zoology, Cambridge, MA

USNM National Museum of Natural History, Smithsonian Institution,

Washington, D.C.

UMMZ University of Michigan Museum of Zoology, Ann Arbor, MI

Morphological features were discovered through comparative analyses of the

taxa. This was done without specific reference to previous morphological analyses. In

order to make the completed analysis conform to the works of others, morphological

characters are described with reference to previous researchers a posteriori. Naming of

features follows the convention of Barel et al. (1977) unless otherwise noted.

The outgroup Cichla ocellaris was chosen because it is a member of what is recognized as the most morphologically plesiomorphic cichlid genus (Van Couvering,

1982; Stiassny, 1991; Liem, 1973; Newsome, 1971). For that reason all character states

for this taxon are coded as 0. The phylogenetic trees are rooted on Cichla ocellaris.

107 Crenicichla saxatilis is a second outgroup; it is a member of the South American

Geophaginae (López-Fernández et al., 2005).

Phylogenetic analysis (10,000 replicate Heuristic Search) and Jackknife resampling (100 replicates, of 5 RAS) was completed in PAUP* (Swofford, 2002).

Bremer support (Bremer, 1995) was calculated using TreeRot v.2 (Sorenson, 1999). All characters were treated as unordered and equally weighted. Multi-state characters were

treated as polymorphisms. Characters and character-states for taxa are listed in Table

IV.2.

RESULTS

Characters are divided into two sections (internal morphology and external

morphology) which are grouped into anatomical units presented from anterior to

posterior. Numbering of characters below correspond to the numbering in the character

matrix. Primitive and derived states are discussed for each character. Consistency

indices (CI) and Retention indices (RI) are given for each character when possible.

Synapomorphies are described whether or not they are unique. Synapomorphies are

derived features that are shared among members of a clade. Derived reversals are also

potentially synapomorphies. Synapomorphies are reported even as autapomorphies

because that character is shared and derived among members of that taxon. A “+” is used

when discussing members of a clade (e.g., Archocentrus = Archocentrus spilurus + A. septemfasciatus + A. nigrofasciatus). The consensus phylogeny is presented in Figure

IV.1. Stippled illustrations of premaxillae and lower jaws are by Teresa Petersen, all other illustrations are by the author.

108

Figure IV.1 Consenus phylogeny of four most parsimonious trees of length 453, CI=.256, RI=.587. Bremer values appear below nodes, jackknife in italics above. An asterisk is placed next to species that belong to ‘Nandopsis’ sensu Regan. All Cichlasoma species listed are “Cichlasoma” sensu lato. Letters refer to clades discussed in text.

109

1 – INTERNAL MORPHOLOGY – ORAL REGION

A. UPPER AND LOWER JAW

1. Symphysial extension of alveolar process of premaxilla 0 : present 1 : absent

Figure IV.2 Character 1, Premaxilla, lateral view (a) Condition 0 (b) Condition 1 Crenicichla saxatilis Nandopsis tetracanthus UMMZ 215935 UMMZ 146972

Symphysial extension of the alveolar process of the premaxilla is identical to Character

16 of Cichocki (1976). However, he quantified six states from the ratio of the length of the medial extension of the total length of the alveolar process. The derived state is a synapomorphy of Clade C (see Fig.IV.1). Within Clade C, the reversal is a synapomorphy of Archocentrus nigrofasciatus, “Cichlasoma” grammodes, Herichthys bartoni, and Clade K. (ci=.20, ri= .73)

110

2. Degree of ventral folding of the ascending process of premaxilla 0 : strong, folded ventrally 1 : weak or not folded ventrally

Figure IV.3 Character 2, Premaxilla, lateral view Condition 0 Nandopsis haitiensis UMMZ 231521

A weak ventral fold on the ascending process of the premaxilla is a synapomorphy of

Petenia splendida and Crenicichla saxatilis. (ci=.50)

3. Foramen in ascending process of premaxilla 0 : none 1 : present

Figure IV.4 Character 3 Premaxilla, lateral view Condition 1 “Cichlasoma” trimaculatum CAS 66704

111 The derived condition is a synapomorphy of Clade D. Within Clade D, the reversal is a synapomorphy of Clade I, Herichthys steindachneri, and Clade S. Within Clade I, the derived state is regained as a synapomorphy of Amphilophus robertsoni. Within Clade S, the derived condition is regained as a synapomorphy of Archocentrus septemfasciatus +

A. spilurus. This is Character 15 of Cichocki (1976). (ci=.14, ri=.65)

4. Shape of posterior end of alveolar process of premaxilla 0 : straight, 1 : posterior end curved ventrally and bulbous

Figure IV.5 Character 4, Premaxilla, lateral view (a) Condition 0 (b) Condition 1 Cichla ocellaris Parachromis managuense UMMZ 204679 UMMZ 223246

The derived condition of having a bulbous ventrally curved posterior end of the alveolar process is a synapomorphy of the ingroup. Within the ingroup, the reversal is a synapomorphy of Clade D. Within Clade D, the derived state is regained as a synapomorphy of Archocentrus and Clade G. Within Clade G, the reversal is a synapomorphy of Clade K. Within Clade K, the derived state is regained as a synapomorphy of Thorichthys. (ci=.17, ri=.70)

112 5. Mental prominence on dentary 0 : weak 1 : strong

Figure IV.6 Character 5 Lower jaw, lateral view (b) Condition 1 Parachromis friedrichsthali

Barel et al. (1977) defined mental prominence as a protuberance of the dentary

symphysis. The derived condition of a distinguished mental prominence is “when the

ventral side of the dentary near its symphysis forms a ventral projection which clearly interrupts the main course of its ventral border” (Barel et al., 1977). The derived

condition of having a strong mental prominence is a synapomorphy of Clade A. Within

Clade A, the reversal is a synapomorphy of “Cichlasoma” salvini, Clade G, and Clade R.

Within Clade G, the derived state is regained as a synapomorphy of “Cichlasoma” urophthalmum. Within Clade R, the derived state is regained as a synapomorphy of

Nandopsis. (ci=.17, ri=.74)

113 6. Shape of lateral face of retroarticular 0 : rectangular, 1 : triangular

Figure IV.7 Character 6, Lower jaw, lateral view (a) Condition 0 (b) Condition 1 Cichla ocellaris Nandopsis tetracanthus UMMZ 204679 UMMZ 146972

The derived condition of having a triangular-shaped retroarticular is a synapomorphy of

Clade C. The reversal is a synapomorphy of “Cichlasoma” ornatum, C. istlanum and C. urophthalmum. (ci=.25, ri=.70)

7. Retroarticular and process of anguloarticular on same vertical plane 0 : no 1 : yes

Figure IV.8 Character 7, Lower jaw, lateral view (a) Condition 0 (b) Condition 1 “Cichlasoma” salvini Crenicichla saxatilis UMMZ 188062-5 UMMZ 215935

The derived condition of having the retroarticular and process of anguloarticular on the

same vertical plane is a synapomorphy of Clade R, and Clade F. Within Clade F, the

reversal is a synapomorphy of Thorichthys, “Cichlasoma” facetum, and “Cichlasoma”

trimaculatum + C. urophthalmum. (ci=.17, ri=.69)

114 8. Angle of process of anguloarticular 0 : procumbent 1 : erect or recurved

Figure IV.9 Character 8, Lower jaw, lateral view (a) Condition 0 (b) Condition 1 Nandopsis haitiensis Nandopsis ramsdeni UMMZ 231521 UMMZ 230839

The derived state of having an erect or recurved anguloarticular process is a synapomorphy of “Cichlasoma” trimaculatum, Amphilophus macracanthus, Thorichthys meeki + T. pasionis, “Cichlasoma” grammodes, Clade R, and Clade M. Within Clade R, a reversal is a synapomorphy of Nandopsis haitiensis. Within Clade M, a reversal is a synapomorphy of Vieja fenestrata + V. maculicauda + V. synspila. This character is similar to Character 8 of Cichocki (1976). However, he quantified the measure into five

states based on angles between the posterior margin of the retroarticular process of the

mandible and the mandibular axis. In my study, intraspecific variation precluded finding

similarly discrete quantitative character-states. (ci=.13, ri=.59)

115 9. Relative height of process of anguloarticular 0 : taller or approximately equal than process of dentary 1 : much shorter than process of dentary

Figure IV.10 Character 9 Lower jaw, lateral view Condition 1 Parachromis loisellei UMMZ 145739

The derived state of having an anguloarticular process notably shorter than the process of the dentary is a synapomorphy of Crenicichla saxatilis + the ingroup clade. A reversal is a synapomorphy of Amphilophus robertsoni, “Cichlasoma” facetum, C. trimaculatum, C. grammodes, Herichthys steindachneri, Hypsophrys nicaraguensis, Archocentrus nigrofasciatus, and Vieja. (ci=.11, ri=.27)

10. Pores along mandibular sensory canal 0 : 5 1 : 4

The derived condition is a synapomorphy for the ingroup. The reversal to five pores

along the mandibular sensory canal is a synapomorphy for Thorichthys as was reported

by Miller and Nelson (1961). “Cichlasoma” salvini is polymorphic for this state. The

presence of four pores in the mandibular canal is derived among cichlasomines and

heroines as most cichlids have five (Kullander, 1996). This character was also used by

Kullander (1998), he called them “lateralis canal openings.” (ci=.67, ri=.75)

116 1 – INTERNAL MORPHOLOGY – ORAL REGION

B. ORAL TEETH

11. Specialized oral teeth 0 : all oral teeth small, uniform and simple 1 : caniniform, recurved teeth present 2 : spatulate teeth present

Presence of caniniform, recurved teeth, is a synapomorphy of the ingroup. Within the ingroup, the transition to spatulate teeth is an unreversed and unique synapomorphy of

Tomocichla. A reversal to simple teeth is a synapomorphy of Clade K. (ci=.67, ri=.88)

12. size of symphysial pair of teeth relative to other teeth 0 : same size as other teeth, or slight gradient of size with shortest teeth found more posteriorly 1 : upper jaw outer row symphysial pair of teeth prolonged; symphysial outer row lower jaw lower teeth very short, next 1 – 2 teeth much longer 2 : symphysial teeth prolonged relative to other teeth but other teeth not as in condition 1

Condition 1 is described here as it was by Kullander and Hartel (1997) in diagnosing

Parachromis. This is also the state that Regan (1905) used to diagnose the entire section

‘Nandopsis’ (see Introduction). The plesiomorphic condition and Condition 2 differ in that there is almost no difference in size of teeth in the plesiomorphic condition but in

Condition 2 the symphysial teeth are notably larger. Condition 1 is a synapomorphy of

Clade A. In Clade A, a reversal is a synapomorphy of Clade Q, Clade O, and Clade I. In

Clade Q, Condition 1 is regained as a synapomorphy of “Cichlasoma” ornatum.

Condition 2 is a synapomorphy of “Cichlasoma” atromaculatum, C. urophthalmum, C. festae, and Herichthys bartoni. (ci=.22, ri=.50)

117

13. Oral teeth w/lingual cusps 0 : absent 1 : present

The derived condition of having lingual cusps is a synapomorphy of Nandopsis,

Hypsophrys nicaraguensis, and Clade M. In Clade M, loss of a lingual cusp is a

synapomorphy as a reversal for Tomocichla. Kullander (1996) calls the possession of a

“lingual subapical cusp on anterior teeth” a derived character of the Heroini (most Middle

American cichlids). However, he noted that the majority of genera but not the majority of species possess this derived state. This feature was also used by Casciotta and Arratia

(1993). This character is drawn in Fig.II.12, Character 20, Chapter II. (ci=.25, ri=.70)

14. Red pigmented area of most anterior oral teeth 0 : just covering tips of teeth 1 : covering ½ to more than ½ of length of each tooth

The enameloid caps in cichlids are often red in color, with the rest of the tooth (dentine)

white. The red pigment and the extent of the enameloid region varies among Neotropical

cichlids. The great extent of red pigment and thus enameloid coverage in Tomocichla is

an unreversed and unique synapomorphy of that genus. (ci=1.0, ri=1.0)

118 15. Intensity of red pigment on oral teeth 0 : dark red pigment 1 : light pigment

The intensity of the pigment in the enameloid caps is also variable with Neotropical

cichlids with most having dark caps. However, an unreversed and unique synapomorphy

of Thorichthys appears to be lighter pigment, nearing absence. Although there is

variation among specimens, most formalin-fixed preserved specimens show the character

state of the species consistently. Cichocki (1976) incorporated the dark tooth caps in the

description of his tooth morphologies (Character 17). (ci=1.0, ri=1.0)

16. Tooth rows on lower jaw cross vertical through anterior ramus of anguloarticular 0 : overlap between tooth rows and insertion of anguloarticular 1 : no overlap between tooth rows and insertion of anguloarticular

Figure IV.11 Character 16, Lower jaw, lateral view (a) Condition 0 (b) Condition 1 Petenia splendida Tomocichla tuba UMMZ 205457-S UMMZ 188316 222 mm SL 167 mm SL

The derived state of having no overlap between tooth rows and insertion of anguloarticular is a synapomorphy of Tomocichla, and Amphilophus macracanthus.

(CI=.50, RI=.50)

119 1 – INTERNAL MORPHOLOGY - SKULL

C. NEUROCRANIUM

17. Neurocranium length 0 : Long, i.e., twice as long as wide 1 : Short, i.e., less than twice as wide as long

As in Kullander (1983) neurocranium length is taken as the length between the

basioccipital condyle and the tip of vomer. Skull depth was taken here to include the top

of the supraoccipital crest to the ventral aspect of the parasphenoid (pharyngeal)

apophysis. The derived state is a synapomorphy of Clade A. Within Clade A, a reversal

is a synapomorphy of Clade L. (ci=.50, ri=.83)

18. Ethmo-vomerine block of neurocranium when viewed ventrally 0 : forked 1 : blunt or rounded to a point

Figure IV.12 Character 18, ventral view of neurocranium, scale 1cm (a) Condition 0 (b) Condition 1 Cichla ocellaris Tomocichla tuba UMMZ 204679-S UMMZ 188316 261 mm SL 167 mm SL (Only anterior half of neurocranium drawn)

120 The ethmo-vomerine block is the most anterior portion of the neurocranium, it includes

(among other elements) the vomer and part of the parasphenoid (Stiassny, 1991). A blunt

or rounded ethmo-vomerine block is an unreversed and unique synapomorphy of the

ingroup. (ci=1.0, ri=1.0)

19. Articulating surface of palatine 0 : poorly developed 1 : well developed

Figure IV.13 Character 19, Lateral view of left palatine, scale bar 1 cm “Cichlasoma” robertsoni UMMZ 197222 150 mm SL Condition 1

The derived condition is possession of two well developed anteroventral processes that form an articulating surface on the palatine. Kullander (1996) described the loss of the articulation of the anteroventral palatine wing with the vomerine shaft as a synapomorphy of heroines. I could not verify the loss of this articulation in the ingroup (all heroines) but it appears that some species retain the articulating shape on the palatine. The derived condition is a synapomorphy of Nandopsis, “Cichlasoma” octofasciatum, C. facetum,

Vieja, “Cichlasoma” grammodes + C. istlanum, Amphilophus robertsoni, and C. trimaculatum. The attachment of the palatine and vomer is through the anterior and median palatovomerine ligaments (Kullander, 1996). (ci=.14, ri=.50)

121 20. Shape of pharyngeal apophysis (basioccipital) 0 : forked anteriorly 1 : rounded

Figure IV.14 Character 20, ventral view of posterior end of neurocranium, scale 1cm Condition 1 Parachromis motaguense UMMZ 190779 175 mm SL

The pharyngeal apophysis consists of the parasphenoid only in these cichlids except

Cichla ocellaris where the basioccipital is also incorporated (Cichocki, 1976). The

derived condition is a synapomorphy of the ingroup. A reversal to an anteriorly forked

pharyngeal apophysis is a synapomorphy of Amphilophus macracanthus, A. robertsoni,

Herichthys steindachneri, Vieja, “Cichlasoma” istlanum, and Clade Q. Within Clade Q,

the derived state is regained as a synapomorphy of “Cichlasoma” festae + C. ornatum.

Kullander (1998) discusses this character (which he calls the basicranial apophysis) for

South American cichlids. (ci=.13, ri=.59)

122 21. Posterior ventral expansion on parasphenoid 0 : absent 1 : present

Figure IV.15 Character 21, lateral view of neurocranium Condition 1 Parachromis friedrichsthali UMMZ 188063-S 95 mm SL

The derived state of having a ventral expansion on the posterior end of the parasphenoid is an unreversed and unique synapomorphy of the ingroup. (ci=1.0, ri=1.0)

22. Anterior shape of parasphenoid 0 : Straight 1 : Bends ventrally

In the derived state the ethmo-vomerine block is bent ventrally relative to more posterior portions of the parasphenoid. The derived state is a synapomorphy of Hypsophrys nicaraguensis, Clade P and Clade K. (ci=.33, ri=.82)

123

Figure IV.16 Characters 23-26, posterior view of neurocranium “Cichlasoma” beani UMMZ 211491 153 mm SL

Dorsal vertebral processes of exoccipitals Condition 1 (C25)

Foramen magnum Condition 1 (C23)

Vertebral concavity Articulating facet Condition 0 (C26)

23. Foramen magnum size 0 : smaller than vertebral concavity 1 : larger than vertebral concavity

Possession of a foramen magnum that is larger than the vertebral concavity is a

synapomorphy of Crenicichla saxatilis and Clade C. The vertebral concavity is the concavity on the neurocranium where the first centrum articulates. (ci=.50, ri=.83)

124 24. Foramen magnum shape 0 : round, no dorsal notch 1 : small notch into exoccipitals present dorsally

A notch in the roof of the foramen magnum is a synapomorphy of Crenicichla saxatilis

and Clade A. Within Clade A, a reversal is a synapomorphy in Hypsophrys

nicaraguensis and Petenia splendida. When this notch is present it is filled by a plug of

cartilage. (ci=.33)

25. Dorsal vertebral processes of exoccipitals 0 : large and folding dorsally over foramen magnum 1 : small

This derived state is a synapomorphy of Clade C, and Crenicichla saxatilis. Within

Clade C, a reversal is a synapomorphy of Vieja synspila and V. zonata. In the derived

state the foramen magnum is not obscured by the exoccipitals processes when viewed

dorsally. (ci=.25, ri=.63)

26. Articulating facets of the exoccipitals 0 : small, facets surrounding most dorsal ¼ of vertebral concavity 1 : large, facets surrounding most dorsal ½ of vertebral concavity

The articulating facets of the exoccipitals hang down and surround ½ to more than a ½ of

the vertebral concavity in the derived state. These processes extend from the

neurocranium onto the dorsal aspect of the first centrum. The derived state is a

synapomorphy of Hypsophrys nicaraguensis, Thorichthys, and Tomocichla. (ci=.33,

ri=.60)

125 1 – INTERNAL MORPHOLOGY – BRANCHIOCRANIUM

D. UPPER PHARYNGEAL JAW - The nomenclature of toothplates follows

Kullander (1983).

27. Expansion of short compressed teeth on upper pharyngeal toothplates 0 : region of short compressed teeth across both larger upper tooth plates 1 : region of short compressed teeth restricted to toothplate 4

Junction of toothplates (plate 4 on left)

Figure IV.17 Characters 27, Ventral view, of upper pharyngeal toothplates scale bar 1cm Condition 0 Hypsophrys nicaraguensis UMMZ 181826 111 mm SL

The derived condition of having short compressed teeth restricted to pharyngeal toothplate 4 is a synapomorphy of the ingroup. A reversal is a synapomorphy of

Parachromis dovii and Clade C. Within Clade C, the derived state is regained as a synapomorphy of Nandopsis, and “Cichlasoma” beani. (ci=.20, ri=.50)

126 28. Molariform teeth on upper pharyngeal jaws 0 : absent 1 : present

This character is likely largely dependent on diet. Molluscivores often have molariform teeth on their pharyngeal jaws and simple teeth on their oral jaws. When molariform

teeth are present on the upper pharyngeal jaw they are restricted to the larger toothplates.

The derived state of having molariform teeth in the upper pharyngeal jaws is a

synapomorphy of Clade H, Herichthys cyanoguttatus, and “Cichlasoma” ornatum.

Within Clade H, a reversal is a synapomorphy of T. pasionis. This character is

polymorphic in Nandopsis haitiensis and N. ramsdeni. (ci=.50, ri=.667)

29. Tooth rows on upper pharyngobranchial toothplate 2 0 : 2 1 : 1 2 : 3 or more

Figure IV.18 Character 29, Lateral view of upper pharyngeal toothplate 2 Condition 1 Parachromis dovii UMMZ 188256 180 mm SL

Condition 1 is a synapomorphy of the ingroup. Within the ingroup, having two rows of

teeth in upper pharyngeal toothplate 2 is a synapomorphy as a reversal for Clade D.

Within Clade D, Clade K has Condition 3 as a synapomorphy. Crenicichla saxatilis also

has at least 3 rows of teeth on this toothplate. (ci=.50, ri=.82)

127 30. Tooth shape on pharyngobranchial toothplate 2 0 : conical, simple 1 : bicuspid

Having bicuspid teeth on pharyngobranchial toothplate 2 is an unreversed and unique

synapomorphy of the ingroup. (ci=1.00, ri=1.00)

1 – INTERNAL MORPHOLOGY - BRANCHIOCRANIUM

E. LOWER PHARYNGEAL TOOTHPLATE

31. Lower pharyngeal toothplate depth on vertical plane 0 : thin i.e., not as deep as wide at midpoint on horizontal plane 1 : deep i.e., as deep as it is wide at midpoint on horizontal plane

The derived state is a synapomorphy of the ingroup. A reversal is a synapomorphy of

Thorichthys. (ci=.50, ri=.75)

32. Tooth cusps on lower pharyngeal toothplate 0 : unicuspid 1 : bicuspid or molariform

In some taxa the smallest teeth on the edges of the tooth plate may be unicuspid. The

derived condition is a synapomorphy of Clade C, Parachromis friedrichsthali, and P.

motaguense. Within Clade C, a reversal is a synapomorphy of “Cichlasoma” festae,

Hypsophrys nicaraguensis, C. facetum, C. urophthalmum, and “Cichlasoma”

atromaculatum + C. beani.

128

33. Teeth on posterior medial row of lower pharyngeal toothplate 0 : narrow, bicuspid or cone shaped 1 : robust, cylindrical and molariform

Figure IV.19 Character 33 Condition 1 Dorsal view of lower pharyngeal toothplate with enlargement of tooth on posterior row Amphilophus robertsoni UMMZ 197222 175 mm SL

This feature is a synapomorphy of Clade H, Herichthys cyanoguttatus, “Cichlasoma” istlanum, and Archocentrus nigrofasciatus. Within Clade H, a reversal is a synapomorphy of Amphilophus macracanthus. (ci=.33, ri=.60)

34. Teeth on center of medial row on lower pharyngeal toothplate 0 : narrow, bicuspid or cone shaped 1 : robust, cylindrical and molariform

Possession of robust molariform teeth at the center of the lower pharyngeal toothplate is a synapomorphy of Clade E, and Clade R. Within Clade E, a reversal is a synapomorphy of Tomocichla sieboldi. Within Clade R, a reversal is a synapomorphy of Archocentrus septemfasciatus + A. spilurus. This character is polymorphic in Nandopsis haitiensis.

The teeth described in this character are anterior to those in Character 33. (ci=.40, ri=.75)

129 35. Central tooth rows on lower pharyngeal toothplate clumped and strongly molariform i.e., teeth with a completely flat dorsal surface 0 : not clumped and molariform 1 : clumped and fully molariform

This character describes a tightly packed molariform tooth cluster that is independent from the derived states in Character 33 and 34. The derived state is a synapomorphy of

Thorichthys and Amphilophus macracanthus. This character is polymorphic in

Nandopsis haitiensis. (ci=.67, ri=.67)

36. Sutures on ventral aspect of lower pharyngeal toothplate 0 : few < 5 1 : many, 5 +

Figure IV.20 Character 36, ventral view of lower pharyngeal toothplates (a)Condition 0 (b) Condition 1 Parachromis loisellei Herichthys steindachneri UMMZ 203897 UMMZ 196348 136 mm SL 125 mm SL

The derived state is a synapomorphy of Archocentrus septemfasciatus + A. spilurus,

Nandopsis haitiensis, Amphilophus macracanthus, Thorichthys pasionis, “Cichlasoma” trimaculatum + C. urophthalmum, and Clade N. Within Clade N, a reversal is a synapomorphy of Tomocichla sieboldi. A similar character was also used by Kullander

(1998). (ci=.14, ri=.54)

130 1 – INTERNAL MORPHOLOGY - BRANCHIOCRANIUM

F. SOFT TISSUE ELEMENTS/ NON-PHARYNGEAL ELEMENTS

37. Gill rakers 0 : longer than wide 1 : wider than long

Gill rakers wider than long is a synapomorphy of Crenicichla saxatilis, “Cichlasoma”

grammodes, and Clade F. Within Clade F, a reversal is a synapomorphy of Vieja, and

Amphilophus hogaboomorus. (ci=.17, ri=.67)

38. Gill rakers on ceratobranchial of first gill arch (most rostral arch) protrude over central ridge of arch 0 : absent 1 : present

Figure IV.21 Character 38, medial view of first gill arch showing ceratobranchial gill rakers, scale bars 1 cm, enlarged view of isolated raker in cross-sectional view (a) Condition 0 (b) Condition 1 Parachromis managuense Tomocichla tuba UMMZ 199603 UMMZ 188316 163 mm SL 174 mm SL

The derived state is a synapomorphy of Clade C. Within Clade C a reversal is a

synapomorphy of “Cichlasoma” festae. (ci=.50, ri=.88)

131 39. Gill filaments on first gill arch 0 : filaments longer than gill arch is wide at all points 1 : shorter than gill arch is wide at corner of arch (shorter arches often more robust)

Having short robust gill filaments is a synapomorphy of Petenia splendida, Archocentrus

nigrofasciatus, “Cichlasoma” istlanum, Hypsophrys nicaraguensis, Clade K, and Clade

M. Within Clade M, a reversal is a synapomorphy of Clade P. (ci=.14, ri=.46).

40. Suturing between the anterior and posterior ceratohyal 0 : incomplete 1 : complete

Figure IV.22 Characters 40, branchiostegals not drawn in (b) Medial view of left hyoid, scale bar 1 cm (a) Condition 0 (b) Condition 1 Herichthys cyanoguttatus Tomocichla tuba UMMZ 179870 UMMZ 188316 144 mm SL 167 mm SL

The derived condition of having complete suturing uniting the anterior and posterior

ceratohyals is a synapomorphy of Crenicichla saxatilis, “Cichlasoma” salvini,

Parachromis friedrichsthali, P. managuense, Clade P, Clade Q, and Clade J. In clade J, a reversal to the incomplete state is a synapomorphy of Thorichthys. In Clade Q, a reversal is a synapomorphy of Archocentrus nigrofasciatus and A. septemfasciatus. (ci=.10, ri=.50)

132 1 – INTERNAL MORPHOLOGY - POST CRANIAL ELEMENTS

G. VERTEBRAE

41. Ventral expansion of 4th abdominal centrum 0 : expanded ventrally 1 : no expansion 2 : expanded ventrally and right and left elements fused to a point

Figure IV.23 Character 41, posterior view of 4th abdominal vertebrae, scale bar 1 cm Condition 2 Geophagus surinamensis UMMZ 204939 147 mm SL

This feature was described as the “inferior vertebral apophysis” by Stiassny (1989). She described it as “an elongate, caudally straight edged spine borne on the third vertebral centrum” (Stiassny, 1989). That this feature is associated with the 3rd centrum in the

African Tylochromis and with the 4th centrum in the Neotropical clade should be noted.

It does not appear in Crenicichla saxatilis, but is present in other members of the

Geophaginae (Figure IV. 23). Condition 2 is a synapomorphy of Clade A. Within Clade

A, a transition to Condition 1 is a synapomorphy of Amphilophus hogaboomorus, Clade

O, Hypsophrys nicaraguensis, Archocentrus nigrofasciatus, “Cichlasoma” grammodes, and “Cichlasoma” festae + C. ornatum. Within Clade O, Condition 2 is regained as a synapomorphy of Vieja fenestrate + V. maculicauda + V. synspila. (ci=.22, ri=.36)

133 42. Processus medialis of pectoral girdle 0 : wide gape, stout 1 : narrow

Figure IV.24 Character 42, ventral view of pectoral girdle, scale bar 1 cm (a) Condition 0 (b) Condition 1 “Cichlasoma” macracanthus Crenicichla saxatilis UMMZ 197383 UMMZ 215935 80 mm SL 110 mm SL Ventral view of pelvic fin Ventral view of pelvic fin

The processus medialis (from Kullander, 1983) is a pelvic girdle element. The derived narrow gape is a synapomorphy of Clade A and Crenicichla saxatilis. Within Clade A, a reversal is a synapomorphy of Clade S, “Cichlasoma” grammodes + C. istlanum, C. urophthalmum, “Cichlasoma” atromaculatum + C. beani, Clade P, and Clade K. Within

Clade S, the derived state is regained in Archocentrus spilurus. (ci=.11, ri=.43)

134 43. Ribs on first caudal centrum 0 : similar width and robustness to more anterior ribs 1 : at least twice as robust (wide) as more anterior ribs

The first caudal centrum is defined by Barel et al. (1977) as “the first centrum to which

the first anal pterygiophore is pointed.” The derived condition is a synapomorphy of

Petenia splendida, Parachromis friedrichsthali, P. loisellei, Parachromis dovii, Clade Q,

Clade M, “Cichlasoma” trimaculatum, Amphilophus hogaboomorus, and A. robertsoni.

Within Clade M, a reversal is a synapomorphy of Tomocichla.

1 – INTERNAL MORPHOLOGY – POST CRANIAL ELEMENTS

H. CAUDAL SKELETON

44. Spur on parhypural 0 : strong, well developed (i.e., at least as long as a hypural is wide) 1 : weak, rudimentary

This feature was called a parhypurapophysis by Cichocki (1976; Character 60). He noted

that this feature is generally well developed in percoids in general. It serves as an

attachment for the hypochordal longitudinalis and flexor ventralis muscles. Kullander

(1983) called this feature the parhypural spine. The derived weak spur on the parhypural

is a synapomorphy for Crenicichla saxatilis, and Clade A. A reversal is a synapomorphy

of Clade E. Within Clade E, the derived state is regained as a synapomorphy in

Tomocichla, Clade K, and “Cichlasoma” trimaculatum + C. urophthalmum. (ci=.17,

ri=.64)

135 Figure IV.25 Character 44, lateral view of caudal fin skeleton Condition 1 Nandopsis tetracanthus CAS 78975 Parhypural 109 mm SL

Principal rays (segmented)

Procurrent rays

45. Epurals 0 : bearing a single procurrent caudal fin ray each (2 total) 1 : bearing several procurrent caudal fin rays (3 or more total)

The derived state is a synapomorphy of Crenicichla saxatilis + the ingroup. The condition where each epural bears one dorsal procurrent ray is a synapomorphy as a reversal for Parachromis motaguense, P. loisellei, Nandopsis haitiensis + N. ramsdeni,

Archocentrus nigrofasciatus, A. septemfasciatus, “Cichlasoma” facetum, Amphilophus hogaboomorus, Thorichthys, and “Cichlasoma” beani. (ci=.10, ri=.25)

136 1 – INTERNAL MORPHOLOGY – POSTCRANIAL ELEMENTS

J. MERISTICS

46. Dorsal aspect of first predorsal element 0 : first predorsal weakly pointed dorsally 1 : strongly pointed dorsally (“hook shaped”)

Having a strongly pointed first (anterior) predorsal is a synapomorphy of Thorichthys,

Archocentrus spilurus + A. septemfasciatus, “Cichlasoma” atromaculatum + C. beani, and Clade O. This feature is polymorphic in Tomocichla sieboldi. (ci=.40, ri=.75)

47. Predorsals 0 : 2 1 : 1 2 : absent

Having no predorsal bones (Condition 2) is an unreversed and unique synapomorphy of

Crenicichla saxatilis. One predorsal spine (Condition 1) is a synapomorphy of

Archocentrus spilurus. Two species were coded as polymorphic: “Cichlasoma” istlanum

sometimes has one predorsal, Hypsophrys nicaraguensis usually has one predorsal spine

and rarely two (all types have one). This character is identical to Character 50 of

Cichocki (1976). He notes that most percoids have three predorsal bones. I was unable

to verify his claim that some members of Thorichthys sometimes have only one

predorsal. He noted that loss of one or both predorsals is a “minor trend among certain

“Cichlasoma” with the presence of two predorsals being primitive.”

137 48. Anal fin spines 0 : 3 spines 1 : 4 to 5 spines 2 : 6 to 7 spines 3 : 8 or greater

The division between states was determined by the modal distribution among species.

Most cichlids have only three anal fin spines. The presence of an elevated number of

spines appears to be a convergence between Malagasy/Indian cichlids and many

Neotropical cichlids. Taylor and Norris (in Miller, 2005) used this distribution in a key

of Mexican cichlids. This character is also similarly divided by Cichocki (1976) for his

Character 52. Kullander (1996) found having five or more anal fin spines and 15 or more

rays to be synapomorphies of heroines (all members of the ingroup here are heroines).

However, he did recognize that some members, including Nandopsis and Tomocichla, are heroines despite having four spines. Regan (1905) used the three spines versus more than three spines dichotomy to reliably recognize clades within Cichlasoma. In my study

Condition 3 (eight or greater spines) is a synapomorphy of Clade B, and Parachromis managuense. Condition 1 (four to five spines) is a synapomorphy of Petenia splendida.

Within Clade B, Condition 1 is regained as a synapomorphy in Clade T, Clade M, and

Clade K. Condition 2 is a synapomorphy of Parachromis dovii + P. motaguense. Within

Clade M, Condition 2 is regained as a synapomorphy in Vieja. Within Clade K,

Condition 3 is regained as a synapomorphy of Thorichthys. (ci=.33, ri=.75)

138 2 – EXTERNAL FEATURES

A. COLOR AND PIGMENTATION - In preserved specimens.

49. Interorbital band between eyes 0 : absent 1 : 2 bands present across front of head between eyes

Having two horizontal bands in the interorbital region is a synapomorphy of Clade P and

“Cichlasoma” octofasciatum. (CI=.50, RI=.83)

50. Opercular spot(s) 0 : absent, or no distinct markings that can be distinguished from mottled pattern on rest of head 1 : present ventrally on operculum as “sub-opercular spot” 2 : present dorsally on operculum, as part of horizontal band or distinct spot 3 : medially placed spot on operculum 4 : both state 1 and 2 always present in combination

In Nandopsis tetracanthus the mottled pattern of spotting throughout the head obscures

any clear opercular spots so it was assigned Condition 0. Thorichthys has a prominent

sub-opercular spot (Condition 1) which is an unreversed and unique synapomorphy of

that clade. Condition 2, is a synapomorphy of Crenicichla saxatilis, “Cichlasoma”

trimaculatum, C. grammodes, C. octofasciatum, Archocentrus septemfasciatus, C.

ornatum, C. salvini, and Petenia splendida. A spot centrally placed on the operculum

(Condition 3) is a synapomorphy of Tomocichla and Vieja zonata. Parachromis

motaguense, P. loisellei, P. friedrichsthali and P. dovii + P. managuense have as a

synapomorphy a ventral spot that is often part of a band that is continuous from near the

base of the eye; these five species also have a dorsal opercular spot. The dorsal spot in

Petenia splendida is an eye mimic that gives the animal the appearance of being much

broader when the gills are expanded laterally in aggresive display (pers. obs.). (ci=.31,

ri=.44)

139 51. Humeral spot at dorsal base of pectoral fin 0 : absent 1 : present

A humeral spot is a synapomorphy for “Cichlasoma” atromaculatum, Herichthys steindachneri, Nandopsis tetracanthus, Parachromis motaguense + P. dovii, P. managuense, P. friedrichsthali, and P. loisellei. Nominal species of Parachromis, and

Nandopsis tetracanthus sometimes also have a spot on the external base of the pectoral

fin and another spot ventral to the humeral spot that is completely obscured when the fin

is held against the body. These two spotting patterns (spot obscured by the fin, and spot

on external base) are highly variable in Nandopsis tetracanthus and Parachromis species

and were not incorporated into the character matrix. (ci=.20, ri=.43)

52. Midbody spot 0 : present below upper lateral line 1 : present above or straddling lateral line 2 : absent

Kullander (1983) called this spot the “midlateral spot.” Even in species that have a dark

striping pattern this spot remains prominent as an expanded area of pigmentation along

the stripe. Lacking the midbody spot is a synapomorphy of Vieja, Nandopsis ramsdeni

and Crenicichla saxatilis. Condition 1 is a synapomorphy of “Cichlasoma” beani and

Thorichthys. (ci=.40, ri=.63)

140 53. Vertical stripe down center of the body 0 : absent, or incomplete 1 : present, stripe across almost full depth of body

A vertical stripe down the full depth of the body is a synapomorphy of “Cichlasoma” festae + C. ornatum, Vieja fenestrata + V. maculicauda, Parachromis dovii,

Archocentrus nigrofasciatus, Herichthys cyanoguttatus, C. urophthalmum, C. beani, and

Clade I. Within Clade I, a reversal is a synapomorphy of Amphilophus robertsoni,

Thorichthys meeki, and T. affinis. (ci=.09, ri=.23)

54. Caudal fin spot position 0 : dorsal to lateral line 1 : straddling lateral line 2 : vertical stripe across tail

Figure IV.26 Character 54 Caudal fin lateral view, lateral line scales (with pores), and scales with dark pigment on caudal fin illustrated (a) Condition 0, (b) Condition 1, “Cichlasoma” trimaculatum Nandopsis haitiensis AMNH 96465 MCZ 64571 105 mm SL 119 mm SL

All included taxa have a spot on the caudal fin just posterior to the margin of caudal peduncle. Having a vertical stripe (Condition 2) across this area rather than a clear spot is an unreversed and unique synapomorphy of Archocentrus. A caudal spot that straddles the lateral line (Condition 1) is a synapomorphy of Thorichthys meeki + T. pasionis,

141 Nandopsis haitiensis + N. ramsdeni, Hypsophrys nicaraguensis, “Cichlasoma” istlanum

+ C. grammodes, C. beani, and Clade M (ci=.38, ri=.72)

55. Caudal spot expansion 0 : caudal spot restricted to caudal fin 1 : caudal spot continues across caudal peduncle as part of a horizontal band.

The derived state of having a caudal spot that is part of a band across the caudal peduncle

is an unreversed and unique synapomorphy of Vieja. (ci=1.00, ri=1.00)

2 – EXTERNAL FEATURES

B. SCALES

56. Number of cheek scale rows 0 : 6 or more 1 : less than 6

Kullander (1983) provides cheek scale row counts in South American Cichlasoma sensu

strictu; all of those species have less than four rows. The derived condition is a synapomorphy of Crenicichla saxatilis, Clade K, Herichthys cyanoguttatus, Tomocichla

tuba, Vieja fenestrata + V. maculicauda + V. synspila, and Clade Q. Within Clade Q, a

reversal is a synapomorphy of “Cichlasoma” festae, Nandopsis haitiensis, and Nandopsis

tetracanthus. (ci=.11, ri=.56)

57. Cheek scales covered in thick skin i.e., skin is so thick that scales cannot be counted without removing the skin 0 : absent 1 : present

The derived state of having thick skin covering the cheek scales is a synapomorphy of

Tomocichla, and Archocentrus septemfasciatus. (ci=.50, ri=.50)

142 58. Cheek scale rows 0 : scales of similar size, symmetrical rows 1 : scales imbricate and vary in size, uneven rows

The derived condition of having uneven cheek scale rows is an unreversed and unique

synapomorphy of the ingroup. (ci=1.00, ri=1.00)

59. Cheek scales 0 : scaleless in part of region between eye and origin of mouth 1 : scales present throughout region between eye and origin of mouth

The derived condition is a synapomorphy of the ingroup. Within the ingroup, a reversal

to having a scaleless area below the eye and posterior to the mouth is a synapomorphy for

Archocentrus nigrofasciatus, “Cichlasoma” grammodes + C. istlanum, Tomocichla tuba and Vieja synspila. (ci=.20, ri=.33)

60. Chest scales 0 : covered in thick layer of skin 1 : without thick skin

The derived condition is a synapomorphy of Crenicichla saxatilis + the ingroup.

Lacking thick skin covering the chest scales is a synapomorphy as a reversal for Clade N,

“Cichlasoma” ornatum, C. urophthalmum, C. beani, Nandopsis haitiensis, Amphilophus

hogaboomorus, and A. macracanthus. (ci=.13, ri=.50)

143 61. Chest scale size 0 : scales gradually sequentially smaller in size from midbody to chest 1 : small chest scales, ¼ of the size of those near the midbody, scales abruptly smaller relative to scales in adjacent regions

The derived condition is to have an abruptly smaller region of scales near the chest.

Coding this character requires removal of some of the skin that overlays these scales.

The derived condition is a synapomorphy of Archocentrus septemfasciatus + A. spilurus,

Nandopsis haitiensis + N. ramsdeni, “Cichlasoma” grammodes, Vieja zonata, V. maculicauda, Tomocichla tuba, Herichthys steindachneri, A. macracanthus, and C. trimaculatum. (ci=.11, ri=.20)

62. Upper and lower lateral line scale row position 0 : posterior termination of upper lateral line does not extend beyond the anterior of the ventral lateral line 1 : posterior termination of upper lateral line does extend beyond the anterior of the ventral lateral line

Figure IV.27 Character 62 Illustration of Condition 1, showing overlapping upper and lower lateral line rows.

The derived state of having overlapping upper and lower lateral lines is a synapomorphy of Petenia splendida, Vieja fenestrata + V. maculicauda, and V. zonata. (ci=.33, ri=.33)

144 63. Number of scale rows between upper and lower lateral line 0 : more than 2 rows of scales 1 : 2 or fewer

The derived condition is a synapomorphy of Clade A, and Crenicichla saxatilis. A reversal is a synapomorphy of Petenia splendida. (ci=.50)

64. Sheath of scales present at base of dorsal fin 0 : absent, few scales 1 : present

This sheath is defined as scales covering portions of the base of the soft dorsal-fin rays.

When the fin is folded these scales cover the base of these rays. This character is used in a key of Mexican cichlids by Taylor and Norris (in Miller, 2005; 4b in key). Dorsal fin and anal fin squamation is illustrated several times by Kullander (1983; his Figure 48 illustrates these characters best). The derived condition is a synapomorphy of Clade A.

A reversal is a synapomorphy of Thorichthys. (ci=.50, ri=.80)

65. Sheath of scales present at base of anal fin 0 : present 1 : absent, few scales

This sheath is defined as scales covering portions of the base of the soft anal rays. When the fin is folded these scales tend to cover the base of these rays. This character is used in a key of Mexican cichlids by Taylor and Norris (in Miller, 2005; 4b in key). Often my characters 64 and 65 are treated together as a single character, but I found that they are independent; Cichla ocellaris has the sheath on the anal fin but not on the dorsal fin. The derived condition of lacking a sheath of scales over the base of the anal fin is a synapomorphy of Crenicichla saxatilis, Petenia splendida and Thorichthys. (ci=.33, ri=.50)

145 2 – EXTERNAL FEATURES

C. SOFT TISSUE IN CRANIAL REGION

66. Oral epithelium 0 : covering most of teeth 1 : teeth clearly exposed

Figure IV.28 Character 66, lateral view of right side of head showing mouth with lips pulled back, scale bar 1 cm (a) Condition 0 (b) Condition 1 “Cichlasoma” macracanthus “Cichlasoma” trimaculatum UMMZ 197383 UMMZ 178854

Lacking deep oral epithelium is not an artifact of preservation (i.e., drying); it is

consistent across collections. The derived condition is a synapomorphy of Crenicichla

saxatilis and Clade A. Within Clade A, the reversal is a synapomorphy of Clade K,

Amphilophus hogaboomorus, and C. urophthalmum. The reversal is also a synapomorphy of Petenia splendida. (ci=.20, ri=.50)

146 67. Nuchal hump 0 : present in all or some individuals 1 : absent

In most species a nuchal hump is present only in males and in breeding periods. Only

Nandopsis ramsdeni has both males and females and even sub-adults with permanent

nuchal humps. Lacking a nuchal hump is a synapomorphy of the ingroup + Crenicichla

saxatilis. Possession of a nuchal hump is a synapomorphy as a reversal for Nandopsis haitiensis + N. ramsdeni, Archocentrus spilurus, “Cichlasoma” grammodes, Vieja

synspila, and Tomocichla tuba. (ci=.17, ri=.17)

68. Dorsal head profile 0 : concavity absent or weak 1 : concavity present and deep

Figure IV.29 Character 68, lateral view of head Condition 1 1 : strong Herichthys steindachneri UMMZ 198800 99 mm SL

This character was used by Barel et al. (1977). As they explain “the impression above

the eye implies a protruding premaxillary pedicel, as it forms the rostral border of the

concavity.” The derived condition is a synapomorphy of Clade C. Within Clade C, a

reversal is a synapomorphy of Clade R, “Cichlasoma” grammodes, C. facetum, C.

urophthalmum, C. atromaculatum, Clade O, and Clade L. Within Clade R, the derived

condition is regained as a synapomorphy of “Cichlasoma” festae. (ci=.11, ri=.11)

147 69. Acute head profile 0 : absent 1 : present

Figure IV.30, Character 69, lateral view of head, Condition 1 “Cichlasoma” robertsoni UMMZ 197263 131 mm SL

The acute head profile is typical of a steeply sloping snout with a deep pre-orbital region.

It is an unreversed and unique synapomorphy of Clade K. (ci=1.00, ri=1.00)

70. Frenum on lower lip 0 : present 1 : absent, undivided lower lip

Figure IV.31 Character 70, ventral view of head Condition 0 Archocentrus septemfasciatus UMMZ 166478 70 mm SL

Presence of a frenum was thought to be homoplasious among Neotropical cichlids by

Hubbs (1936). The absence of a frenum is the derived state and a synapomorphy of the ingroup. Within the ingroup, a reversal to possessing a frenum is a synapomorphy of

148 Archocentrus, Nandopsis ramsdeni, Vieja, Tomocichla sieboldi, Herichthys bartoni,

“Cichlasoma” atromaculatum, and Clade J. Within Clade J, the derived condition is regained as a synapomorphy of Thorichthys. (ci=.11, ri=.47)

71. Ventral fold of lower lip 0 : not expanded 1 : expanded resembling a third lip from lateral view

Figure IV.32, Character 71, lateral view of head, Condition 1 “Cichlasoma” beani UMMZ 189958 91 mm SL

The derived state is a synapomorphy of Crenicichla saxatilis + the ingroup. Within this clade a reversal is a synapomorphy of Clade D. Within Clade D, the derived state is regained as a synapomorphy of Nandopsis tetracanthus, “Cichlasoma” trimaculatum +

C. urophthalmum, and “Cichlasoma” beani. The ventral fold is an expansion of the lower lip often of the same depth as the lower lip itself. Some species have an intermediate condition where the ventral fold is not complete anteriorly, these incomplete forms were coded with the plesiomorphic condition. (ci=.20, ri=.64)

149

72. Mouth length, including maxillary shank 0 : Reaching vertical through rostral end of eye 1 : Not reaching vertical through rostral end of eye

A short mouth is an unreversed and unique synapomorphy of Clade A. (ci=1.00, ri=1.00)

73. Position of mouth 0 : terminal 1 : slightly downturned 2 : subterminal

Figure IV.33 Character 73, Lateral view of head Condition 2, Tomocichla sieboldi UMMZ 230707 123 mm SL

A slightly downturned mouth is an unreversed and unique synapomorphy of

Archocentrus. A strongly downturned mouth is a synapomorphy of Clade P, and

Hypsophrys nicaraguensis. (ci=.10, ri=.50)

2 – EXTERNAL FEATURES

D. FINS

74. Long filamentous extension of the dorsal fin 0 : absent 1 : present

A filamentous extension of the dorsal fin is a synapomorphy of Crenicichla saxatilis, and

Clade A. Within Clade A, a reversal is a synapomorphy of Clade F, and “Cichlasoma”

ornatum. Within Clade F, the derived condition is regained as a synapomorphy of

Thorichthys. (ci=.20, ri=.79)

150 75. Pectoral fin length 0 : reaching midpoint of body, longer than pelvic fin 1 : shorter than pelvic fin 2 : very long reaching beyond midpoint of body

Long pectoral fins that reach beyond the midpoint of the body (Condition 2) are a

synapomorphy of Petenia splendida, and Clade K. Having pectoral fins that are shorter

than the pelvic fins is a synapomorphy of Clade A, and Crenicichla saxatilis. Within

Clade A, a reversal is a synapomorphy of Parachromis managuense, “Cichlasoma”

festae, Vieja, and Amphilophus hogaboomorus. (ci=.29, ri=.58)

76. Pectoral fin shape 0 : pointed 1 : rounded

Figure IV.34, Character 76, lateral view of pectoral fin, (dotted area show Condition 1) Condition 0 “Cichlasoma” robertsoni UMMZ 197263 131 mm SL

The derived state of having rounded pectoral fins is a synapomorphy of Crenicichla saxatilis + the ingroup. Within this clade, a reversal to pointed pectoral fins is a synapomorphy of Clade L. This character was also used by Kullander (1998). (ci=.50, ri=.75)

151 77. Long filamentous extensions of the pelvic fins 0 : absent 1 : present

Filamentous extensions of the pelvic fins are a synapomorphy of Clade A. Within Clade

A, a reversal is a synapomorphy of Nandopsis tetracanthus, and Clade F. Within Clade

F, the derived condition is regained as a synapomorphy of Thorichthys, Herichthys

cyanoguttatus and Vieja. Within Vieja, a reversal is a synapomorphy of Vieja fenestrata.

(ci=.25, ri=.63)

78. Pelvic fins 0 : Set apart, with a clear space between fins 1 : Set close, slightly overlapping

The derived condition is a synapomorphy of Crenicichla saxatilis + the ingroup. A

reversal is a synapomorphy of Tomocichla tuba. (ci=.50)

79. Long filamentous extension of the anal fin 0 : absent 1 : present

Filamentous extension of the anal fin is a synapomorphy of Parachromis dovii + P.

motaguense, P. managuense, P. friedrichsthali, and Clade C. Within Clade C, a reversal

is a synapomorphy of Clade F. Within Clade F, the derived state is regained as a

synapomorphy of Thorichthys. (ci=.25, ri=.84)

152 80. Long filamentous extensions of the caudal fin 0 : absent 1 : present

Filamentous extensions of the caudal fin are an unreversed and unique synapomorphy of

Thorichthys meeki. This character is present in some South American cichlid taxa not sampled in this study. (ci=1.00)

81. Caudal fin shape 0 : truncate or emarginated 1 : rounded

The derived condition is a synapomorphy of Crenicichla saxatilis + the ingroup. Within this clade, a reversal is a synapomorphy of Thorichthys. Possession of a truncate or emarginated caudal fin is a feature reported by Miller and Nelson (1961) to diagnose

Thorichthys. This character was also used by Kullander (1998). (ci=.50, ri=.67)

153 2 – EXTERNAL FEATURES

E. SENSORY PORES - Sensory pores on the skin are associated with foramina on the skeleton but the pores can be far more numerous. The clustering of these pores is described in the following six characters.

82. Sensory pores in the gular region 0 : Multiple clumped pores 1 : Single pores (sometimes two)

The derived condition is a synapomorphy of Clade B. Within Clade B, a reversal is a synapomorphy of Tomocichla, Vieja synspila, “Cichlasoma” ornatum, Nandopsis tetracanthus, and N. haitiensis. Sensory pores in the gular region are associated with the dental lateralis foramen and the anguloarticular lateralis foramen (Kullander, 1983).

(ci=.17, ri=.55)

83. Anterior sensory pores on ventral aspect of head 0 : pores not covered by ventral end of lower lip 1 : pores covered by ventral end of lower lip

Sensory pores on the ventral aspect of the head are associated with the dental lateralis foramen (Kullander, 1983). The derived condition of having these pores covered by the lower lip is an unreversed and unique synapomorphy of Tomocichla tuba. In some species with large lower lips there is an indentation in the lower lip so that these pores remain exposed. (ci=1.00)

154 84. Sensory pores dorsal to eye 0 : highly branching (with 10 or more pores) 1 : few branches

Figure IV.35, Character 84, lateral view of head Condition 0 Tomocichla sieboldi UMMZ 230707 124 mm SL

Sensory pores above the eye are associated with the frontal lateralis canal foramen

(Kullander, 1983). The derived condition is a synapomorphy of Crenicichla saxatilis and

Clade B. Within Clade B, a reversal is a synapomorphy of “Cichlasoma” ornatum,

Nandopsis haitiensis, Hypsophrys nicaraguensis, C. grammodes, Vieja zonata,

Tomocichla, Amphilophus macracanthus, and C. urophthalmum. (CI=.10, RI=.31)

155 85. Sensory pores directly posterior to the eye, 0 : form continuous clump of pores from eye to opening of operculum 1 : few pores, discontinuous clumps from eye to operculum

Sensory pores directly posterior to the eye are associated with the pterotic lateralis canal

foramen (Kullander, 1983). The derived condition is a synapomorphy of Crenicichla

saxatilis + the ingroup. Within this clade, a reversal is a synapomorphy of Parachromis

dovii + P. motaguense, “Cichlasoma” salvini, Clade T, Archocentrus spilurus,

Hypsophrys nicaraguensis, Clade P, Herichthys steindachneri, Thorichthys affinis,

Amphilophus macracanthus, and “Cichlasoma” trimaculatum. Within Clade T, the

derived state is regained as a synapomorphy in Nandopsis ramsdeni. (ci=.08, ri=.42)

86. Sensory pores ventral to eye 0 : highly branching (clumps of 5+ pores) 1 : few branches

Sensory pores ventral to the eye are associated with the infraorbital lateralis foramen and

the posterior lachrymal lateralis foramen (Kullander, 1983). The derived state is a

synapomorphy of Crenicichla saxatilis, and Clade B. Within Clade B, a reversal is a synapomorphy of “Cichlasoma” ornatum, C. grammodes, Hypsophrys nicaraguensis,

Amphilophus macracanthus, and “Cichlasoma” trimaculatum + C. urophthalmum.

(ci=.13, ri=.42)

156 87. Sensory pores at corner of upper and lower lip 0 : absent 1 : present

Sensory pores at the corner of the upper and lower lip are associated with the ventral lachrymal lateralis foramen (Kullander, 1983). The derived condition is a synapomorphy of Crenicichla saxatilis + the ingroup. Within this clade, a reversal is a synapomorphy of

Vieja synspila. (CI=.50)

2 – EXTERNAL FEATURES

F. MORPHOMETRICS

88. Maximum standard length 0 : Standard length 15 cm or more 1 : Standard length < 15 cm

Maximum standard lengths of species were confirmed using Kullander (2003). The derived state of a short standard length is a synapomorphy of Archocentrus and

Amphilophus hogaboomorus. (ci=.50, ri=.67)

157 89. Overall body depth 0 : slender or intermediate 1 : deep bodied, depth of body greatest near midline and over 50% of SL

Figure IV.36 Character 89, Body shape outline (a) Condition 0 (b) Condition 1, Archocentrus septemfasciatus Archocentrus spilurus UMMZ 166478 UMMZ 197198 70 mm SL 76 mm SL

The derived condition of having a deep body is a synapomorphy of “Cichlasoma”

trimaculatum, Archocentrus nigrofasciatus, A. spilurus, Nandopsis ramsdeni, N.

tetracanthus, Clade J, and Clade N. Within Clade J, a reversal is a synapomorphy of

Amphilophus robertsoni. Males of Tomocichla tuba are deep bodied and females are slender; therefore, this species is coded as polymorphic.

(ci=.11, ri=.47).

158 Table IV.2. Table of all clades of two or more species recovered in this study with a complete list of synapomorphies. Unique and unreversed synapomorphies are in bold. When an “R” follows the character number it means that the character is a synapomorphy as a reversal for that clade. When an “r” is included with the character number it means that the character is reversed at least once within that clade. A character’s number that does not appear in bold means that the character appears as a convergence in another part of the phylogeny (i.e., it is not unique to that clade). A character with a decimal refers to the specific character state that is the synapomorphy.

Group Synapomorphies The ingroup 4r, 10r, 11r, 18, 20r, 21, 27r, 29.1r, 30, 31r, 58, 59r, 70r Clade A 5r, 12r, 17r, 24r, 41.2r, 42r, 44r, 63, 64r, 66r, 72, 74r, 75.1r, 77r Clade B 48.3r, 82r, 84r, 86r Clade C 1r, 6r, 23, 25r, 27Rr, 32r, 38r, 68r, 79r Clade D 3r, 4Rr, 29Rr, 71Rr Clade E 34r, 44Rr Clade F 7r, 37r, 74Rr, 77Rr, 79Rr Clade G 4r, 5Rr Clade H 28r, 33r Clade I 3Rr, 12R, 53r Clade J 40r, 70Rr, 89r Clade K 1R, 4R, 11R, 22, 29, 39, 42R, 44, 48.1r, 56, 66R, 69, 75.2 Clade L 17R, 68R, 76R Clade M 8r, 13r, 39r, 43r, 48.1r, 54.1 Clade N 36r, 60R, 89 Clade O 12R, 41.1r, 46, 68R Clade P 22, 39R, 40, 42R, 49, 73.2, 85R Clade Q 12Rr, 20Rr, 40r, 43, 56r Clade R 5Rr, 7, 8r, 34r, 68Rr Clade S 3Rr, 42Rr Clade T 48.1, 85Rr Crenicichla saxatilis + 9r, 45r, 60, 67, 71r, 76r, 78r, 81r, 85r, 87r the ingroup Archocentrus 4, 54.2, 70R, 73.1, 88 Archocentrus septemfasciatus 3, 34R, 36, 46, 61 + A. spilurus “Cichlasoma” beani + 32R, 42R, 46 C. atromaculatum “Cichlasoma” festae + 20, 41.1, 53 C. ornatum “Cichlasoma” grammodes + 19, 42R, 54.1, 59R C. istlanum

159 Table IV.2. Continued…

C. trimaculatum + 7R, 36, 44, 71, 86R C. urophthalmum Nandopsis 5, 13, 19, 27 Nandopsis haitiensis + 45R, 54.1, 61, 67R N. ramsdeni Parachromis dovii + 48.2, 50.4, 51, 79, 85R P. motaguense Thorichthys 4, 7R, 10R, 15, 26, 31R, 35, 40R, 45R, 46, 48.3, 50.1, 52.1, 64R, 65, 70, 74, 77, 79, 81R Thorichthys meeki + 8, 54.1 T. pasionis Tomocichla 11.2, 13R, 14, 16, 26, 43R, 44, 50.3, 57, 82R, 84R Vieja 9R, 19, 20R, 37R, 48.2, 52.2, 55, 70R, 75R, 77r Vieja fenestrata + 8R, 41.2, 56 V. maculicauda + V. synspila V. fenestrata + 53, 62 V. maculicauda

160 DISCUSSION

Monophyly of the section ‘Nandopsis’ sensu Regan (1906-1908; species listed in

Table IV.1) is rejected. Species that Regan (1906-1908) grouped in his sections

‘Thorichthys,’ ‘Amphilophus’ (his Astatheros), ‘Archocentrus’ and ‘Theraps’(now

including Vieja among other genera) form clades within ‘Nandopsis’ (Figure IV.37).

Rather than give the entire ingroup clade one name (i.e., Nandopsis, Herichthys, Heros or

other) it is recommended here that the current trend of splitting the former members of

Cichlasoma into monophyletic genera be continued. Nominal genera and clades are

discussed below. See Table V.2 for a complete list of synapomorphies for all recovered

clades with two or more species.

The ingroup (Heroini) - All members of the ingroup belong to cichlid tribe

Heroini (Kullander, 1996, 1998; Cichlasoma group A of Stiassny, 1991). Kullander’s

(1998) phylogeny did not include any Middle American endemics. However, he includes

most Middle American cichlids and several South American cichlids in Heroini based on

anal-fin spine count, bicuspid dentition, postcleithrum shape and, palatine position and

articulation. Kullander first used four or more anal-fin spines (1996), and then, five or

more anal-fin spines (1998) to diagnose heroines, noting that there are exceptions. In this

study, all members of the ingroup (all heroines) had four or more anal-fin spines

(Character 48). Another of Kullander’s diagnostic heroine characters, possession of

bicuspid teeth (conical teeth with a lingual cusp; Character 13 this study), is found here to

be a synapomorphy of only part of the ingroup (Nandopsis, Hypsophrys nicaraguensis,

and Clade M).

161 Monophyly of the ingroup clade (Heroini) in this analysis is supported by the following unique and unreversed synapomorphies: Character 18, a blunt or rounded ventral appearance of the ethmo-vomerine block of the neurocranium; Character 21, a ventral expansion on the posterior end of the parasphenoid; Character 30, bicuspid teeth on pharyngobranchial toothplate 2; Character 58, cheek scales that vary in size and that are in uneven rows. Nine other synapomorphies (all include reversals within the clade) support monophyly of the ingroup (Table IV.2).

Clade A - Ingroup clade that excludes Petenia Petenia was recovered as the sister group to the remainder of the ingroup. Petenia splendida has many unusual features, including a highly protrusible mouth that appears to be an autapomorphy.

Difficulty in placing this species among other cichlids led to this genus remaining monotypic. Miller (2005) recently suggested making the South American genus

Caquetaia a synonym of Petenia. Caquetaia has been recovered within the Middle

American clade in several molecular phylogenies but never sister to Petenia (Hulsey et al., 2004; see Chapter V). Caquetaia and Petenia share a distinct protuberance on the alveolar process of the premaxilla (pers. obs.).

This ingroup clade that excludes Petenia splendida (Clade A) was supported by the unique and unreversed synapomorphy Character 72; a mouth that does not reach the vertical through the rostral end of eye. Twelve other synapomorphies (all include reversals within the clade) support monophyly of this group (Table IV.2).

Amphilophus Agassiz, 1859 is recovered as a paraphyletic group. Amphilophus robertsoni was recovered as the sister group to Thorichthys (together Clade L).

Amphilophus macracanthus was sister to Clade L (together Clade K). “Cichlasoma”

162 facetum was recovered as the sister group to Clade K (together Clade J). The third

Amphilophus species sampled, Amphilophus hogaboomorus, was found sister to Clade J

(together Clade I). A paraphyletic Amphilophus (as currently recognized) was also

recovered in the molecular phylogenetic analyses of Roe et al. (1997) and Hulsey et al.

(2004).

Clade K (Thorichthys + Amphilophus macracanthus + A. robertsoni)

Monophyly of Clade K was supported by the unique and unreversed synapomorphy

Character 69; an acute head profile. Twelve other synapomorphies (all include reversals

within the clade) support monophyly of this group (Table IV.2).

The common names of some Amphilophus and Thorichthys species speak to

similarities in their shapes: Thorichthys meeki is called the firemouth cichlid, and

Amphilophus robertsoni the false fire-mouth cichlid (Bussing, 1987).

Thorichthys Meek, 1904 is one of the few recognized monophyletic groups

among Middle American genera (Miller and Nelson, 1961; Kullander, 1983; Taylor and

Miller, 1984; Roe et al., 1997, Hulsey et al., 2004). This group was originally described

as a section of Cichlasoma by Regan (1906-1908), then a subgenus (Miller and Nelson,

1961) and finally raised to a genus by Kullander (1983). Monophyly was supported by

the molecular analysis of Roe et al. (1997) and Hulsey et al. (2004). Monophyly of

Thorichthys was supported in this analysis by the following unique and unreversed

synapomorphies: Character 15, light pigment on oral teeth; Character 50, Condition 1, a

sub-opercular spot. Eighteen other synapomorphies (all unreversed within the clade but

none unique) support monophyly of Thorichthys. Bremer (12) and jackknife (100)

support is very high for this clade.

163 Vieja Fernández-Yépez, 1969 is recovered as a monophyletic group. The genus has been assembled by the similarity of having “blunt rounded heads and relatively deep skulls” (Miller, 2005). The genus has also been suggested to be potentially polyphyletic,

with two recognized body shapes within the group (Miller, 2005). One group is a deep

bodied clade with a dark band beneath the lateral line; the other group is more elongate with vertical bars (Miller, 2005). All Vieja species in this study belong to the more deep bodied group and those four species are recovered as a clade. Monophyly of Vieja is supported by the unique and unreversed Character 55; a caudal spot that is continuous across the caudal peduncle as part of a horizontal band. Nine other synapomorphies (all but one unreversed within the clade, all non-unique) support monophyly of this clade.

Bremer (5) and jackknife (98) support is very high for this clade.

Tomocichla Regan, 1908 is recovered as a monophyletic group. Three species are currently recognized in this genus. The species not sampled in this study, Tomocichla asfraci (Allgayer, 2002), is known only from a single drainage in Panama. Two unique and unreversed synapomorphies support monophyly of Tomocichla; Character 11,

Condition 2, spatulate teeth; Character 14, red pigmentation covering ½ to more than ½

of length of each anterior oral tooth. Nine other synapomorphies support monophyly of

Tomocichla (all unreversed but non-unique). Bremer (7) and jackknife (99) support is very high for this clade.

Tomocichla is recovered as the sister group to Vieja (together Clade P).

Monophyly of Tomocichla + Vieja is supported by seven synapomorphies (all unreversed within the clade, but non-unique). The sister group to this clade is Herichthys cyanoguttatus (together Clade O). The sister group to Clade O is Herichthys

164 steindachneri (together Clade N). The sister group to Clade N is Herichthys

cyanoguttatus (together Clade M). The recovered positions of these nominal Herichthys

species make this genus paraphyletic.

Herichthys Baird and Girard, 1854 was not recovered as a clade. Hulsey et al.

(2004) recover a monophyletic Herichthys in a molecular phylogenetic analysis that includes most of the nine species in the genus, including all those sampled here.

Kullander (1996) diagnosed this genus on breeding coloration (light coloration on dorsal half of head and flank, dark in adjacent areas) and unique dental morphology.

Nandopsis Gill, 1862 is restricted to the Greater Antillean cichlids (Miller, 2005;

Chakrabarty, 2006, 2006b). Miller (1993; 2005) conjectured that the Greater Antillean cichlids were a clade and that Nandopsis should be restricted to these species stating

(2005) “Nandopsis is absent from Middle America; we restrict it to the Greater Antilles.”

This was stated without comment or evidence. Kullander (1983) also recognized that the

Greater Antillean cichlids “form a discrete group phonetically [sic] with, e.g., few anal- fin spines.” However, he listed one of the Cuban species as “Cichlasoma” ramsdeni in

his checklist (Kullander, 2003). This hypothesis is tested in the morphological

phylogenetic analysis presented here and all Greater Antillean species are recovered as a

clade. Nandopsis is henceforth restricted to the Greater Antillean species.

No unique synapomorphies diagnose this clade but four synapomorphies

(unreversed within the clade) support monophyly of this group: Character 5, a strong

mental prominence in the symphysis of the dentary; Character 13, lingual cusps on

anterior oral teeth; Character 19, a well developed articulating surface of the palatine;

Character 27, short compressed teeth restricted to upper pharyngeal toothplate 4.

165 A molecular analysis recovered these three extant Greater Antillean cichlids as a

clade (Chakrabarty 2006b; Chapter III). The fossil Nandopsis woodringi was placed in

that genus based on synapomorphies it shares with Nandopsis tetracanthus, the type

species of Nandopsis (Chakrabarty 2006; Chapter II).

Clade T (“Cichlasoma” ornatum +“Cichlasoma” festae + Nandopsis) Not all members of the ingroup are Middle American endemics; Nandopsis, is Greater Antillean, and “Cichlasoma” ornatum and “Cichlasoma” festae are South American. These species were placed in ‘Nandopsis’ by Regan (1906-1908) and “Cichlasoma” ornatum and C. festae were species he thought had Central American ‘Nandopsis’ ancestors that migrated to South America (see Introduction). The phylogeny presented here agrees with that hypothesis. A parsimony optimization of area on this phylogeny recovers Clade T as

Middle American. All the immediate outgroups of this clade are Middle American.

Therefore, these cichlids are Middle American cichlids that are currently found in South

America and the Greater Antilles (See Chapter V). Two synapomorphies (both non- unique but unreversed within the clade; Table IV.2) diagnose this clade: Character 48,

Condition 1, 4 to 5 spines; Character 85 (as a reversal), sensory pores form a continuous clump of pores from eye to opening of operculum. Clade T is recovered as the sister group to Archocentrus (together Clade S).

Archocentrus Gill, 1877 is monophyletic. Two unique and unreversed synapomorphies support monophyly of the three Archocentrus species sampled:

Character 54, Condition 2, a vertical stripe across base of caudal fin; Character 73,

Condition 1, a slightly downturned mouth. Three other synapomorphies (unreversed within the clade but non-unique) support monophyly of this clade (Table IV.2).

166 Parachromis Agassiz, 1859 was not recovered as monophyletic despite sharing a

number of features. These species are difficult to distinguish from each other and all

nominal species were sampled. Retention of plesiomorphic states and “Cichlasoma”

salvini’s inclusion in the ingroup may have precluded recovery of this genus as

monophyletic. “Cichlasoma” salvini closely resembles Parachromis in some respects but lacks many of the group’s putative synapomorphies. The exclusion of this taxon

from the phylogentic analysis results in a monophyletic Parachromis. Parachromis was

also recovered as a clade in the jackknife consensus tree with jackknife support of 68.

Overall tree support is low with most clades having Bremer values of 1 and

jackknife support below 70. Deep nodes within the ingroup are equally poorly supported.

Homoplasious features that have plagued taxonomists in the past also complicate this

phylogenetic analysis. Morphological analyses that dealt with Neotropical cichlids often

used weighting schemes to better resolve phylogenies (Stiassny, 1991; Kullander, 1998).

All weighting schemes favor some characters over others; because there is no logical basis for assigning weight, equal weights were assigned here. Successive weighting schemes may downplay homoplasious characters but “noisy” characters may also be informative. Ultimately, combining many characters from different sources and levels of analysis will hopefully provide a more satisfactory understanding of Neotropical cichlid relationships.

167 Table IV.3 Character Matrix

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10

Cichla ocellaris 0 0 0 0 0 0 0 0 0 0 Crenicichla saxatilis 0 1 0 0 0 0 1 0 1 0 Amphilophus hogaboomorus 1 0 0 1 0 1 1 0 1 1 Amphilophus macracanthus 0 0 0 0 0 1 1 1 1 1 Amphilophus robertsoni 0 0 1 0 0 1 1 0 0 1 Archocentrus nigrofasciatus 0 0 0 1 0 1 1 1 0 1 Archocentrus septemfasciatus 1 0 1 1 0 1 ? ? 1 1 Archocentrus spilurus 1 0 1 1 0 1 1 1 1 1 “Cichlasoma” atromaculatum 1 0 0 1 0 1 1 0 1 1

168 “Cichlasoma” beani 1 0 1 1 0 1 1 0 1 1 “Cichlasoma” facetum 1 0 0 0 0 1 0 0 0 1 “Cichlasoma” festae 1 0 0 0 0 1 1 1 1 1 “Cichlasoma” grammodes 0 0 1 0 1 1 0 1 0 1 “Cichlasoma” istlanum 1 0 1 0 1 0 0 0 1 1 “Cichlasoma” octofasciatum 1 0 1 0 0 1 1 1 1 1 “Cichlasoma” ornatum 1 0 0 0 0 0 1 1 1 1 “Cichlasoma” salvini 1 0 0 1 0 1 0 0 1 0 & 1 “Cichlasoma” trimaculatum 1 0 1 1 0 1 0 1 0 1 “Cichlasoma” urophthalmum 1 0 1 1 1 0 0 0 1 1 Herichthys bartoni 0 0 1 0 1 1 1 1 1 1 Herichthys cyanoguttatus 1 0 1 0 1 1 1 1 1 1 Herichthys steindachneri 1 0 0 0 1 1 1 1 0 1

168

Table IV.3 Character Matrix continued…

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10

Hypsophrys nicaraguensis 1 0 1 0 1 1 0 0 0 1 Nandopsis haitiensis 1 0 0 0 1 1 1 0 1 1 Nandopsis ramsdeni 1 0 0 0 1 1 1 1 1 1 Nandopsis tetracanthus 1 0 0 0 1 1 1 1 1 1 Parachromis dovii 0 0 0 1 1 0 0 0 1 1 Parachromis friedrichsthali 0 0 0 1 1 0 0 0 1 1 Parachromis loisellei 0 0 0 1 1 0 0 0 1 1 Parachromis managuense 0 0 0 1 1 0 0 0 1 1 169 Parachromis motaguense 0 0 0 1 1 0 0 0 1 1 Petenia splendida 0 1 0 1 0 0 0 0 1 1 Thorichthys affinis 0 0 0 1 0 1 0 0 1 0 Thorichthys meeki 0 0 0 1 0 1 0 1 1 0 Thorichthys pasionis 0 0 0 1 0 1 0 1 1 0 Tomocichla sieboldi 1 0 1 0 1 1 1 1 1 1 Tomocichla tuba 1 0 1 0 1 1 1 1 1 1 Vieja fenestrata 1 0 1 0 1 1 1 0 0 1 Vieja maculicauda 1 0 1 0 1 1 1 0 0 1 Vieja synspila 1 0 1 0 1 1 1 0 0 1 Vieja zonata 1 0 1 0 1 1 1 1 0 1

169 Table IV.3 Character Matrix continued…

C11 C12 C13 C14 C15 C16 C17 C18 C19 C20

Cichla ocellaris 0 0 0 0 0 0 0 0 0 0 Crenicichla saxatilis 0 0 0 0 0 0 0 0 0 0 Amphilophus hogaboomorus 1 0 0 0 0 0 1 1 0 1 Amphilophus macracanthus 0 0 0 0 0 1 1 1 0 0 Amphilophus robertsoni 0 0 0 0 0 0 0 1 1 0 Archocentrus nigrofasciatus 1 0 0 0 0 0 1 1 0 0 Archocentrus septemfasciatus 1 0 0 0 0 0 1 1 0 0

170 Archocentrus spilurus 1 0 0 0 0 0 1 1 0 0 “Cichlasoma” atromaculatum 1 2 0 0 0 0 1 1 0 1 “Cichlasoma” beani 1 1 0 0 0 0 1 1 0 1 “Cichlasoma” facetum 1 0 0 0 0 0 1 1 1 1 “Cichlasoma” festae 1 2 0 0 0 0 1 1 0 1 “Cichlasoma” grammodes 1 1 0 0 0 0 1 1 1 1 “Cichlasoma” istlanum 1 1 0 0 0 0 1 1 1 0 “Cichlasoma” octofasciatum 1 0 0 0 0 0 1 1 1 0 “Cichlasoma” ornatum 1 1 0 0 0 0 1 1 ? 1 “Cichlasoma” salvini 1 1 0 0 0 0 1 1 0 1 “Cichlasoma” trimaculatum 1 1 0 0 0 0 1 1 1 1 “Cichlasoma” urophthalmum 1 2 0 0 0 0 1 1 0 1 Herichthys bartoni 1 2 1 0 0 0 1 1 0 1 Herichthys cyanoguttatus 1 0 1 0 0 0 1 1 0 1 Herichthys steindachneri 1 1 1 0 0 0 1 1 0 0

170

Table IV.3 Character Matrix continued…

C11 C12 C13 C14 C15 C16 C17 C18 C19 C20

Hypsophrys nicaraguensis 1 0 1 0 0 0 1 1 0 0 Nandopsis haitiensis 1 0 1 0 0 0 1 1 1 0 Nandopsis ramsdeni 1 0 1 0 0 0 1 1 1 0 Nandopsis tetracanthus 1 0 1 0 0 0 1 1 1 0 Parachromis dovii 1 1 0 0 0 0 1 1 0 1 Parachromis friedrichsthali 1 1 0 0 0 0 1 1 0 1 Parachromis loisellei 1 1 0 0 0 0 1 1 0 1

171 Parachromis managuense 1 1 0 0 0 0 1 1 0 1 Parachromis motaguense 1 1 0 0 0 0 1 1 0 1 Petenia splendida 0 0 0 0 0 0 0 1 0 1 Thorichthys affinis 0 0 0 0 1 0 0 1 0 1 Thorichthys meeki 0 0 0 0 1 0 0 1 0 1 Thorichthys pasionis 0 0 0 0 1 0 0 1 0 1 Tomocichla sieboldi 2 0 0 1 0 1 1 1 0 ? Tomocichla tuba 2 0 0 1 0 1 1 1 0 1 Vieja fenestrata 1 0 1 0 0 0 1 1 1 0 Vieja maculicauda 1 0 1 0 0 0 1 1 1 0 Vieja synspila 1 0 1 0 0 0 1 1 1 0 Vieja zonata 1 0 1 0 0 0 1 1 1 0

171 Table IV.3 Character Matrix continued…

C21 C22 C23 C24 C25 C26 C27 C28 C29 C30

Cichla ocellaris 0 0 0 0 0 0 0 0 0 0 Crenicichla saxatilis 0 0 1 1 1 0 0 0 2 0 Amphilophus hogaboomorus 1 0 1 1 1 0 0 1 0 1 Amphilophus macracanthus 1 1 1 1 1 0 0 1 2 1 Amphilophus robertsoni 1 1 1 1 1 0 0 1 2 1 Archocentrus nigrofasciatus 1 0 1 1 1 0 0 0 0 1 Archocentrus septemfasciatus 1 0 1 1 1 0 0 0 ? 1

172 Archocentrus spilurus 1 0 1 1 1 0 0 0 0 1 “Cichlasoma” atromaculatum 1 0 1 1 1 0 0 0 0 1 “Cichlasoma” beani 1 0 1 1 1 0 1 0 0 1 “Cichlasoma” facetum 1 0 1 1 1 0 0 1 0 1 “Cichlasoma” festae 1 0 1 1 1 0 0 0 0 1 “Cichlasoma” grammodes 1 0 1 1 1 0 0 0 0 1 “Cichlasoma” istlanum 1 0 1 1 1 0 0 0 0 1 “Cichlasoma” octofasciatum 1 0 1 1 1 0 0 0 0 1 “Cichlasoma” ornatum 1 0 1 1 1 0 1 1 0 1 “Cichlasoma” salvini 1 0 1 1 1 0 0 0 1 1 “Cichlasoma” trimaculatum 1 0 1 1 1 0 0 1 0 1 “Cichlasoma” urophthalmum 1 0 1 1 1 0 0 1 0 1 Herichthys bartoni 1 0 1 1 1 0 0 0 0 1 Herichthys cyanoguttatus 1 0 1 1 1 0 0 1 0 1 Herichthys steindachneri 1 0 1 1 1 0 0 0 0 1

172 Table IV.3 Character Matrix continued…

C21 C22 C23 C24 C25 C26 C27 C28 C29 C30

Hypsophrys nicaraguensis 1 0 1 0 1 1 0 0 0 1 Nandopsis haitiensis 1 0 1 1 1 0 1 0 & 1 0 1 Nandopsis ramsdeni 1 0 1 1 1 0 1 0 & 1 0 1 Nandopsis tetracanthus 1 0 1 1 1 0 1 0 0 1 Parachromis dovii 1 0 0 1 0 0 0 0 1 1 Parachromis friedrichsthali 1 0 0 1 0 0 1 0 1 1 Parachromis loisellei 1 0 0 1 0 0 1 0 1 1 Parachromis managuense 1 0 0 1 0 0 1 0 1 1

173 Parachromis motaguense 1 0 0 1 0 0 1 0 1 1 Petenia splendida 1 0 0 0 0 0 1 0 1 1 Thorichthys affinis 1 0 1 1 1 1 0 1 2 1 Thorichthys meeki 1 0 1 1 1 1 0 1 2 1 Thorichthys pasionis 1 0 1 1 1 1 0 0 2 1 Tomocichla sieboldi 1 0 1 1 1 1 0 0 0 1 Tomocichla tuba 1 0 1 1 1 1 0 0 0 1 Vieja fenestrata 1 0 1 1 1 0 0 0 0 1 Vieja maculicauda 1 0 1 1 1 0 0 0 0 1 Vieja synspila 1 0 1 1 0 0 0 0 0 1 Vieja zonata 1 0 1 1 0 0 0 0 0 1

173 Table IV.3 Character Matrix continued…

C31 C32 C33 C34 C35 C36 C37 C38 C39 C40

Cichla ocellaris 0 0 0 0 0 0 0 0 0 0 Crenicichla saxatilis 0 0 0 0 0 0 1 0 0 1 Amphilophus hogaboomorus 1 1 1 1 0 0 0 1 0 0 Amphilophus macracanthus 1 1 0 1 1 1 1 1 1 1 Amphilophus robertsoni 1 1 1 1 0 0 1 1 1 1 Archocentrus nigrofasciatus 1 1 1 1 0 0 0 1 1 0 Archocentrus septemfasciatus 1 1 0 0 0 1 0 1 0 0 Archocentrus spilurus 1 1 0 0 0 1 0 1 0 1

174 “Cichlasoma” atromaculatum 1 0 0 1 0 0 0 1 0 0 “Cichlasoma” beani 1 0 0 1 0 0 1 1 0 0 “Cichlasoma” facetum 1 0 1 1 0 0 1 1 0 1 “Cichlasoma” festae 1 0 0 1 0 0 0 0 0 1 “Cichlasoma” grammodes 1 1 0 1 0 0 1 1 0 0 “Cichlasoma” istlanum 1 1 1 1 0 0 0 1 1 1 “Cichlasoma” octofasciatum 1 1 0 1 0 0 0 1 0 1 “Cichlasoma” ornatum 1 1 0 1 0 0 0 1 0 1 “Cichlasoma” salvini 1 1 0 0 0 0 0 1 0 1 “Cichlasoma” trimaculatum 1 1 1 1 0 1 1 1 0 0 “Cichlasoma” urophthalmum 1 0 1 1 0 1 1 1 0 0 Herichthys bartoni 1 1 0 1 0 0 1 1 1 0 Herichthys cyanoguttatus 1 1 1 1 0 1 1 1 1 0 Herichthys steindachneri 1 1 0 1 0 1 1 1 1 0

174 Table IV.3 Character Matrix continued…

C31 C32 C33 C34 C35 C36 C37 C38 C39 C40

Hypsophrys nicaraguensis 1 0 0 0 0 0 0 1 1 1 Nandopsis haitiensis 1 1 0 & 1 0 & 1 0 & 1 1 0 1 0 1 Nandopsis ramsdeni 1 1 0 1 0 0 0 1 0 1 Nandopsis tetracanthus 1 1 0 1 0 0 0 1 0 1 Parachromis dovii 1 0 0 0 0 0 0 0 0 0 Parachromis friedrichsthali 1 1 0 0 0 0 0 0 0 1 Parachromis loisellei 1 0 0 0 0 0 0 0 0 0 Parachromis managuense 1 0 0 0 0 0 0 0 0 1

175 Parachromis motaguense 1 1 0 0 0 0 0 0 0 0 Petenia splendida 1 0 0 0 0 0 0 0 1 0 Thorichthys affinis 0 1 1 1 1 0 1 1 1 0 Thorichthys meeki 0 1 1 1 1 0 1 1 1 0 Thorichthys pasionis 0 1 1 1 1 1 1 1 1 0 Tomocichla sieboldi 1 1 0 0 0 0 1 1 0 1 Tomocichla tuba 1 1 0 1 0 1 1 1 0 1 Vieja fenestrata 1 1 0 1 0 1 0 1 0 1 Vieja maculicauda 1 1 0 1 0 1 0 1 0 1 Vieja synspila 1 1 0 1 0 1 0 1 0 1 Vieja zonata 1 1 0 1 0 1 0 1 0 1

175 Table IV.3 Character Matrix continued…

C41 C42 C43 C44 C45 C46 C47 C48 C49 C50

Cichla ocellaris 0 0 0 0 0 0 0 0 0 0 Crenicichla saxatilis 1 1 0 1 1 - 2 0 0 2 Amphilophus hogaboomorus 0 1 1 0 0 0 0 2 0 0 Amphilophus macracanthus 2 0 0 1 1 0 0 1 0 0 Amphilophus robertsoni 2 0 1 1 1 0 0 1 0 0 Archocentrus nigrofasciatus 0 0 1 1 0 0 0 3 0 0 Archocentrus septemfasciatus 2 0 1 1 0 1 0 3 0 2 Archocentrus spilurus 2 1 1 1 1 1 1 3 0 0 “Cichlasoma” atromaculatum 2 0 0 0 1 1 0 2 0 2 “Cichlasoma” beani 2 0 0 0 0 1 0 2 0 0 176 “Cichlasoma” facetum 2 1 0 0 0 0 0 2 0 0 “Cichlasoma” festae 0 0 1 1 1 0 0 1 0 0 “Cichlasoma” grammodes 0 0 0 0 1 0 0 2 0 2 “Cichlasoma” istlanum 2 0 0 0 1 0 0 & 1 2 0 0 “Cichlasoma” octofasciatum 2 1 1 1 1 0 0 3 1 2 “Cichlasoma” ornatum 0 0 1 1 1 0 0 2 0 2 “Cichlasoma” salvini 2 1 0 1 1 0 0 3 0 2 “Cichlasoma” trimaculatum 2 1 1 1 1 0 0 2 0 2 “Cichlasoma” urophthalmum 2 0 0 1 1 0 0 2 0 0 Herichthys bartoni 2 1 1 0 1 0 0 1 0 0 Herichthys cyanoguttatus 0 1 1 0 1 1 0 1 0 0 Herichthys steindachneri 2 ? 1 0 1 0 0 1 0 0

176 Table IV.3 Character Matrix continued…

C41 C42 C43 C44 C45 C46 C47 C48 C49 C50

Hypsophrys nicaraguensis 0 1 1 1 1 0 0 & 1 3 0 0 Nandopsis haitiensis 2 0 1 1 0 0 0 1 0 0 Nandopsis ramsdeni 2 0 1 1 0 0 0 1 0 0 Nandopsis tetracanthus 2 0 1 1 1 0 0 1 0 0 Parachromis dovii 2 1 1 1 1 0 0 2 0 4 Parachromis friedrichsthali 2 1 1 1 1 0 0 3 0 4 Parachromis loisellei 2 1 1 1 0 0 0 3 0 4 177 Parachromis managuense 2 1 0 1 1 0 0 3 0 4 Parachromis motaguense 2 1 0 1 0 0 0 2 0 4 Petenia splendida 0 0 1 0 1 0 0 1 0 2 Thorichthys affinis 2 0 0 1 0 1 0 3 0 1 Thorichthys meeki 2 0 0 1 0 1 0 3 0 1 Thorichthys pasionis 2 0 0 1 0 1 0 3 0 1 Tomocichla sieboldi 0 0 0 1 1 0 & 1 0 1 1 3 Tomocichla tuba 0 0 0 1 1 1 0 1 1 3 Vieja fenestrata 2 0 1 0 1 1 0 2 1 0 Vieja maculicauda 2 0 1 0 1 1 0 2 1 0 Vieja synspila 2 0 1 0 1 1 0 2 1 0 Vieja zonata 0 0 1 0 1 1 0 2 1 3

177 Table IV.3 Character Matrix continued…

C51 C52 C53 C54 C55 C56 C57 C58 C59 C60

Cichla ocellaris 0 0 0 0 0 0 0 0 0 0 Crenicichla saxatilis 0 2 0 0 0 1 0 0 0 1 Amphilophus hogaboomorus 0 0 1 0 0 0 0 1 1 0 Amphilophus macracanthus 0 0 1 0 0 1 0 1 1 0 Amphilophus robertsoni 0 0 0 0 0 1 0 1 1 1 Archocentrus nigrofasciatus 0 0 1 2 0 1 0 1 0 1 Archocentrus septemfasciatus 0 0 0 2 0 1 1 1 1 1 Archocentrus spilurus 0 0 0 2 0 1 0 1 1 1 “Cichlasoma” atromaculatum 1 0 0 0 0 0 0 1 1 1

178 “Cichlasoma” beani 0 1 1 1 0 0 0 1 1 0 “Cichlasoma” facetum 0 0 1 0 0 1 0 1 1 1 “Cichlasoma” festae 0 0 1 0 0 0 0 1 1 1 “Cichlasoma” grammodes 0 0 0 1 0 0 0 1 0 1 “Cichlasoma” istlanum 0 0 0 1 0 0 0 1 0 1 “Cichlasoma” octofasciatum 0 0 0 0 0 1 0 1 1 1 “Cichlasoma” ornatum 0 0 1 0 0 1 0 1 1 0 “Cichlasoma” salvini 0 0 0 0 0 0 0 1 1 1 “Cichlasoma” trimaculatum 0 0 0 0 0 0 0 1 1 1 “Cichlasoma” urophthalmum 0 0 1 0 0 0 0 1 1 0 Herichthys bartoni 0 0 0 1 0 0 0 1 1 1 Herichthys cyanoguttatus 0 0 1 1 0 1 0 1 1 0 Herichthys steindachneri 1 0 0 1 0 0 0 1 1 0

178 Table IV.3 Character Matrix continued…

C51 C52 C53 C54 C55 C56 C57 C58 C59 C60

Hypsophrys nicaraguensis 0 0 0 1 0 1 0 1 1 1 Nandopsis haitiensis 0 0 0 1 0 0 0 1 1 0 Nandopsis ramsdeni 0 2 0 1 0 1 0 1 1 1 Nandopsis tetracanthus 1 0 0 0 0 0 0 1 1 1 Parachromis dovii 1 0 1 0 0 0 0 1 1 1 Parachromis friedrichsthali 1 0 1 0 0 0 0 1 1 1 Parachromis loisellei 1 0 0 0 0 0 0 1 1 1

179 Parachromis managuense 1 0 0 0 & 1 0 0 0 1 1 1 Parachromis motaguense 1 0 0 0 0 0 0 1 1 1 Petenia splendida 0 0 0 0 0 0 0 1 1 1 Thorichthys affinis 0 1 0 0 0 1 0 1 1 1 Thorichthys meeki 0 1 0 1 0 1 0 1 1 1 Thorichthys pasionis 0 1 1 1 0 1 0 1 1 1 Tomocichla sieboldi 0 0 0 1 0 0 1 1 1 0 Tomocichla tuba 0 0 0 1 0 1 1 1 0 0 Vieja fenestrata 0 2 1 1 1 1 0 1 1 0 Vieja maculicauda 0 2 1 1 1 1 0 1 1 0 Vieja synspila 0 2 0 1 1 1 0 1 0 0 Vieja zonata 0 2 0 1 1 0 0 1 1 0

179 Table IV.3 Character Matrix continued…

C61 C62 C63 C64 C65 C66 C67 C68 C69 C70

Cichla ocellaris 0 0 0 0 0 0 0 0 0 0 Crenicichla saxatilis 0 0 1 0 1 1 1 0 0 0 Amphilophus hogaboomorus 0 0 1 1 0 0 1 1 0 1 Amphilophus macracanthus 1 0 1 1 0 0 1 1 1 0 Amphilophus robertsoni 0 0 1 1 0 0 1 0 1 0 Archocentrus nigrofasciatus 0 0 1 1 0 1 1 0 0 0 Archocentrus septemfasciatus 1 0 1 1 0 1 1 0 0 0 Archocentrus spilurus 1 0 1 1 0 1 0 0 0 0 “Cichlasoma” atromaculatum 0 0 1 1 0 1 1 0 0 0 “Cichlasoma” beani 0 0 1 1 0 1 1 1 0 1 180 “Cichlasoma” facetum 0 0 1 1 0 1 1 0 0 0 “Cichlasoma” festae 0 0 1 1 0 1 1 1 0 1 “Cichlasoma” grammodes 1 0 1 1 0 1 0 0 0 1 “Cichlasoma” istlanum 0 0 1 1 0 1 1 1 0 1 “Cichlasoma” octofasciatum 0 0 1 1 0 1 1 0 0 1 “Cichlasoma” ornatum 0 0 1 1 0 1 1 0 0 1 “Cichlasoma” salvini 0 0 1 1 0 1 1 1 0 1 “Cichlasoma” trimaculatum 1 0 1 1 0 1 1 1 0 1 “Cichlasoma” urophthalmum 0 0 1 1 0 0 1 0 0 1 Herichthys bartoni 0 0 1 1 0 1 1 1 0 0 Herichthys cyanoguttatus 0 0 1 1 0 1 1 0 0 1 Herichthys steindachneri 1 0 0 0 0 1 1 1 0 1

180 Table IV.3 Character Matrix continued…

C61 C62 C63 C64 C65 C66 C67 C68 C69 C70

Hypsophrys nicaraguensis 0 0 1 1 0 1 1 1 0 1 Nandopsis haitiensis 1 0 1 1 0 1 0 0 0 1 Nandopsis ramsdeni 1 0 1 1 0 1 1 0 0 0 Nandopsis tetracanthus 0 0 1 1 0 1 1 0 0 1 Parachromis dovii 0 0 1 1 0 1 1 0 0 1 Parachromis friedrichsthali 0 0 1 1 0 1 1 0 0 1 Parachromis loisellei 0 0 1 1 0 1 1 0 0 1

181 Parachromis managuense 0 0 1 1 0 1 1 0 0 1 Parachromis motaguense 0 0 1 1 0 1 1 0 0 1 Petenia splendida 0 1 0 0 1 0 1 0 0 1 Thorichthys affinis 0 0 1 0 1 0 1 0 1 1 Thorichthys meeki 0 0 1 0 1 0 1 0 1 1 Thorichthys pasionis 0 0 1 0 1 0 1 0 1 1 Tomocichla sieboldi 0 0 1 1 0 1 1 0 0 0 Tomocichla tuba 1 0 1 1 0 1 0 0 0 1 Vieja fenestrata 0 1 1 1 0 1 1 0 0 0 Vieja maculicauda 1 1 1 1 0 1 1 0 0 0 Vieja synspila 0 0 1 1 0 1 0 0 0 0 Vieja zonata 1 1 1 1 0 1 1 0 0 0

181 Table IV.3 Character Matrix continued…

C71 C72 C73 C74 C75 C76 C77 C78 C79 C80

Cichla ocellaris 0 0 0 0 0 0 0 0 0 0 Crenicichla saxatilis 1 0 0 1 1 1 0 1 0 0 Amphilophus hogaboomorus 0 1 0 0 0 1 0 1 0 0 Amphilophus macracanthus 0 1 0 0 2 1 0 1 0 0 Amphilophus robertsoni 0 1 0 0 2 0 0 1 0 0 Archocentrus nigrofasciatus 0 1 1 1 1 1 1 1 1 0 Archocentrus septemfasciatus 0 1 1 1 1 1 1 1 1 0 Archocentrus spilurus 0 1 1 1 1 1 1 1 1 0 “Cichlasoma” atromaculatum 0 1 0 0 1 1 0 1 0 0

182 “Cichlasoma” beani 1 1 0 0 1 1 0 1 0 0 “Cichlasoma” facetum 0 1 0 0 1 1 0 1 0 0 “Cichlasoma” festae 0 1 0 1 0 1 1 1 1 0 “Cichlasoma” grammodes 0 1 0 1 1 1 1 1 1 0 “Cichlasoma” istlanum 0 1 0 1 1 1 1 1 1 0 “Cichlasoma” octofasciatum 0 1 0 1 1 1 1 1 1 0 “Cichlasoma” ornatum 0 1 0 0 1 1 1 1 1 0 “Cichlasoma” salvini 1 1 0 1 1 1 1 1 1 0 “Cichlasoma” trimaculatum 1 1 0 0 1 1 0 1 0 0 “Cichlasoma” urophthalmum 1 1 0 0 1 1 0 1 0 0 Herichthys bartoni 0 1 0 0 1 1 0 1 0 0 Herichthys cyanoguttatus 0 1 0 0 1 1 1 1 0 0 Herichthys steindachneri 0 1 0 0 1 1 0 1 0 0

182 Table IV.3 Character Matrix continued…

C71 C72 C73 C74 C75 C76 C77 C78 C79 C80

Hypsophrys nicaraguensis 0 1 2 1 1 1 1 1 1 0 Nandopsis haitiensis 0 1 0 1 1 1 0 & 1 1 1 0 Nandopsis ramsdeni 0 1 0 1 1 1 1 1 1 0 Nandopsis tetracanthus 1 1 0 1 1 1 0 1 1 0 Parachromis dovii 1 1 0 1 1 1 1 1 1 0 Parachromis friedrichsthali 1 1 0 1 1 1 1 1 1 0 Parachromis loisellei 1 1 0 1 1 1 1 1 0 0 Parachromis managuense 1 1 0 1 0 1 1 1 1 0 Parachromis motaguense 1 1 0 1 1 1 1 1 1 0

183 Petenia splendida 1 0 0 0 2 1 0 1 0 0 Thorichthys affinis 0 1 0 1 2 0 1 1 1 0 Thorichthys meeki 0 1 0 1 2 0 1 1 1 1 Thorichthys pasionis 0 1 0 1 2 0 1 1 1 0 Tomocichla sieboldi 0 1 2 0 1 1 0 1 0 0 Tomocichla tuba 0 1 2 0 1 1 0 0 0 0 Vieja fenestrata 0 1 2 0 0 1 0 1 0 0 Vieja maculicauda 0 1 2 0 0 1 1 1 0 0 Vieja synspila 0 1 2 0 0 1 1 1 0 0 Vieja zonata 0 1 2 0 0 1 1 1 0 0

183 Table IV.3 Character Matrix continued…

C81 C82 C83 C84 C85 C86 C87 C88 C89

Cichla ocellaris 0 0 0 0 0 0 0 0 0 Crenicichla saxatilis 1 0 0 1 1 1 1 0 0 Amphilophus hogaboomorus 1 1 0 1 1 1 1 1 0 Amphilophus macracanthus 1 1 0 0 0 0 1 0 1 Amphilophus robertsoni 1 1 0 1 1 1 1 0 0 Archocentrus nigrofasciatus 1 1 0 1 1 1 1 1 1 Archocentrus septemfasciatus 1 1 0 1 1 1 1 1 0 Archocentrus spilurus 1 1 0 1 0 1 1 1 1 “Cichlasoma” atromaculatum 1 1 0 1 1 1 1 0 0

184 “Cichlasoma” beani 1 1 0 1 1 1 1 0 0 “Cichlasoma” facetum 1 1 0 1 1 1 1 0 1 “Cichlasoma” festae 1 1 0 1 0 1 1 0 0 “Cichlasoma” grammodes 1 1 0 0 1 0 1 0 0 “Cichlasoma” istlanum 1 1 0 1 1 1 1 0 0 “Cichlasoma” octofasciatum 1 1 0 1 1 1 1 0 0 “Cichlasoma” ornatum 1 0 0 0 0 0 1 0 0 “Cichlasoma” salvini 1 1 0 1 0 1 1 0 0 “Cichlasoma” trimaculatum 1 1 0 1 0 0 1 0 1 “Cichlasoma” urophthalmum 1 1 0 0 1 0 1 0 0 Herichthys bartoni 1 1 0 1 1 1 1 0 0 Herichthys cyanoguttatus 1 1 0 1 1 1 1 0 1 Herichthys steindachneri 1 1 0 1 0 1 1 0 0

184 Table IV.3 Character Matrix continued…

C81 C82 C83 C84 C85 C86 C87 C88 C89

Hypsophrys nicaraguensis 1 1 0 0 0 0 1 0 0 Nandopsis haitiensis 1 0 0 0 0 1 1 0 0 Nandopsis ramsdeni 1 1 0 1 1 1 1 0 1 Nandopsis tetracanthus 1 0 0 1 0 1 1 0 1 Parachromis dovii 1 0 0 0 0 0 1 0 0 Parachromis friedrichsthali 1 1 0 1 1 1 1 0 0 Parachromis loisellei 1 1 0 1 1 1 1 0 0 Parachromis managuense 1 0 0 0 1 0 1 0 0 Parachromis motaguense 1 0 0 0 0 0 1 0 0 Petenia splendida 1 0 0 0 1 0 1 0 0

185 Thorichthys affinis 0 1 0 1 0 1 1 0 1 Thorichthys meeki 0 1 0 1 1 1 1 0 1 Thorichthys pasionis 0 1 0 1 1 1 1 0 1 Tomocichla sieboldi 1 0 0 0 0 0 1 0 1 Tomocichla tuba 1 0 1 0 0 0 1 0 0 & 1 Vieja fenestrata 1 1 0 1 0 1 1 0 1 Vieja maculicauda 1 1 0 1 0 1 1 0 1 Vieja synspila 1 0 0 1 0 1 0 0 1 Vieja zonata 1 1 0 0 0 1 1 0 1

185 MATERIALS EXAMINED

The following comparative materials have been examined. Institutional catalog number,

number of specimens examined, and size range follow the species name. The following

abbreviations are used: alc., alcohol; C&S, cleared and stained; sk., skeleton preparation.

Species listed in alphabetical order by specific epiteth.

Thorichthys affinis UMMZ 196484 (18 ex. alc., 32-98 mm SL), UMMZ 143741 (48 ex. alc., 27-100 mm SL), UMMZ 143716 (8 ex. C&S, 42-81 mm SL) “Cichlasoma” atromaculatum FMNH 58600 (1 ex. alc., 88 mm SL; 1 ex. C&S), FMNH 58602 (1 ex. C&S), FMNH 58605 (1 ex. alc., 40 mm SL), UMMZ 179290 (1 ex. alc., 77 mm SL) Herichthys bartoni UMMZ 191755-S (1 ex. sk., 103 mm SL), UMMZ 193477 (32 ex. alc., 32-133 mm SL) UMMZ 196342-S (1 ex. sk., 182 mm SL), UMMZ 192326-S (1 ex. sk., 127 mm SL), UMMZ 192327-S (1 ex. sk., 111 mm SL) “Cichlasoma” beani UMMZ 172021 (61 ex. alc., 27-159 mm SL), UMMZ 211491-S (1 ex. sk. 153 mm SL), UMMZ 211409 (2 ex. C&S), UMMZ 184888 (1 ex. sk., 163 mm SL) Herichthys cyanoguttatus UMMZ 124407 (14 ex. alc., 22-120 mm SL), UMMZ 179870-S (2 ex. sk. 144- 145 mm SL), UMMZ 162152 (5 ex. C&S)

Parachromis dovii

UMMZ 166473 (6 ex. alc., 88-188 mm SL; 1 ex. sk., 147 mm SL; 1 ex.

Ridewood prep.; 1 ex. C&S), UMMZ 188234 (6 ex. alc., 50-98 mm SL), UMMZ

188256-S (1 ex. sk., 180 mm SL), UMMZ 197399 (2 ex. alc.)

186 “Cichlasoma” facetum

UMMZ 206159 (33 ex. alc., 17-118 mm SL), UMMZ 206232 (1 ex. sk., 98 mm

SL, 4 ex. C&S)

Vieja fenestrata

UMMZ 97669 (5 ex. alc., 17-117 mm SL), UMMZ 223240-S (3 ex. sk., 167-182

mm SL), UMMZ 184565 (1 ex. C&S)

“Cichlasoma” festae

AMNH 97327 (2 ex. C&S, 68-86 mm SL), AMNH 59300 (2 ex. alc., 80-112 mm

SL)

Parachromis friedrichsthali

UMMZ 143857 (8 ex. C&S), UMMZ 143956 (1 ex. sk., 173 mm SL), UMMZ

189952 (4 ex. alc., 125-171 mm SL), UMMZ 223245-S (1 ex. sk., 162 mm SL),

UMMZ 188063-S (2 ex. sk., 95-108 mm SL), UMMZ 189952 (1 ex., Ridewood

prep.)

“Cichlasoma” grammodes

UMMZ 181815 (2 ex. C&S), UMMZ 184735 (6 ex. alc., 92-154 mm SL),

UMMZ 186366-S (1 ex. sk., 138 mm SL)

Nandopsis haitiensis

AMNH 222473 (1 ex. alc., 65 mm SL), AMNH 229572 (2 ex. alc., 88-94 mm

SL), AMNH 229573 (5 ex. alc., 111-74 mm SL), AMNH 229574 (4 ex. alc., 122-

71 mm SL), MCZ 33996 (3 ex. alc., 82-104 mm SL), MCZ 62945 (10 ex. alc.,

81-107 mm SL), MCZ 64571 (2 ex. alc., 98-106 mm SL), MCZ 92471 (1 ex. alc.,

86 mm SL), UMMZ 142438 (6 ex. alc., 76-118 mm SL, 1 ex. sk., 82 mm SL, 1

187 ex. C&S), UMMZ 200246 (2 ex. alc., 40-74 mm SL), UMMZ 231521-S (1 ex.

sk., SL unknown), UMMZ 243241 (2 ex. alc., 53-173 mm SL; 1 ex. Ridewood

prep.), UMMZ 243287 (8 ex. alc., 35-100 mm SL; 1 ex. C&S), UMMZ 243302

(18 ex. alc., 26-92 mm SL; 1 ex. C&S), UMMZ 243310 (7 ex. alc., 26-84 mm

SL), USNM 085764 (1 ex. alc., 124 mm SL), USNM 088335 (3 ex. alc., 26-41

mm SL), USNM 092124 (4 ex. alc., 22-72 mm SL), USNM 164796 (3 ex. alc.,

88-97 mm SL), USNM 164863 (6 ex. alc., 28-87 mm SL), USNM 087360 (3 ex.

alc., 80-113 mm SL), USNM 85764 (1 ex. alc., 124 mm SL), USNM 120361 (2

ex. alc., 71-104 mm SL), USNM 122635 (3 ex. alc., 84-10 mm SL), USNM

170907 (1 ex. alc. 105 mm SL), USNM 170908 (7 ex. alc., 54-81 mm SL),

USNM 367230 (5 ex. alc., 70-100 mm SL).

Amphilophus hogaboomorus

UMMZ 144665 (41 ex. alc., 27-76 mm SL; 2 ex. C&S)

“Cichlasoma” istlanum

UMMZ 108600 (1 ex. sk., 157 mm SL; 1 ex. Ridewood prep.), UMMZ 160778

(149 ex. alc., 9-145 mm SL; 1 ex. sk., 125 mm SL), UMMZ 192535 (6 ex. alc.,

89-164 mm SL), UMMZ 203221 (2 ex. C&S), UMMZ 203326 (1 ex. alcohol, 88

mm SL)

Parachromis loisellei

UMMZ 199601 (1 ex. Ridewood prep.), UMMZ 145724 (1 ex. C&S), UMMZ

145739 (3 ex. alc., 91-188 mm SL), UMMZ 203897 (1 ex. sk., 136 mm SL),

UMMZ 203898 (8 ex. alc., 58-88 mm SL), UMMZ 145739 (1 ex. sk., 129 mm

SL)

188 Amphilophus macracanthus

UMMZ 197383-S (1 ex. sk., 80 mm SL), UMMZ 197383 (1 ex. C&S), UMMZ

178855 (28 ex. alc., 52-111 mm SL), UMMZ 197383 (6 ex. alc., 19-119 mm SL)

Vieja maculicauda

UMMZ 197221 (1 ex. sk., 120 mm SL), UMMZ 146115 (1 ex. alc., 142 mm SL),

UMMZ 199583 (37 ex. alc., 18-104 mm SL), UMMZ 73279 (1 ex. C&S, 72 mm

SL)

Parachromis managuense

UMMZ 230831-S (2 ex. sk., 145-155 mm SL), UMMZ 197323 (2 ex. alc., 67-

199), UMMZ 197401 (1 ex. Ridewood prep.), UMMZ 197401 (2 ex. C&S),

UMMZ 144670 (1 ex. C&S), UMMZ 199603 (24 ex. alc., 17-163 mm SL; 1 ex.

C&S; 1 ex. sk., 231 mm SL), UMMZ 209069-S (1 ex. sk., 251 mm SL),

Thorichthys meeki

UMMZ 143912 (29 ex. alc., 26-102 mm SL), UMMZ 167694 (43 ex. alc., 25-87

mm SL), UMMZ 188064 (1 ex. sk., 73 mm SL), UMMZ 210926 (3 ex. C&S, 41-

71 mm SL)

Parachromis motaguense

UMMZ 190786 (1 ex. alc., 148 mm SL), UMMZ 190779-S (1 ex. sk., 149 mm

SL), UMMZ 190779 (1 ex. Ridewood prep.), UMMZ 190801 (5 ex. alc. 43-135

mm SL), UMMZ 225017 (1 ex. C&S)

Hypsophrys nicaraguensis

UMMZ 181826 (2 ex. alc., 111-113 mm SL; 1 ex. sk. 111 mm SL), UMMZ

188994 (3 ex. alc., 65-97 mm SL), UMMZ 199639 (2 ex. C&S)

189 Archocentrus nigrofasciatus

UMMZ 199383 (1 ex. C&S), UMMZ 188245 (1 ex. C&S), UMMZ 196948 (2 ex.

C&S), UMMZ 181823 (16 ex. alc., 36-91 mm SL; 2 ex. C&S), UMMZ 190191 (1

ex. C&S), UMMZ 194157 (3 ex. C&S)

Cichla ocellaris

UMMZ 214789 (1 ex. alc., 188 mm SL), UMMZ 204679-S (1 ex. sk., 261 mm

SL), UMMZ 204407-S (1 ex. sk., 309 mm SL), UMMZ 205006 (2 ex. C&S; 1 ex.

Ridewood prep.), UMMZ 215948 (1 ex. C&S), UMMZ 215716 (1 ex. C&S),

UMMZ 216098 (1 ex. C&S)

“Cichlasoma” octofasciatus

UMMZ 143967 (43 ex. alc., 35-104 mm SL), UMMZ 143964 (2 ex. C&S),

UMMZ 190859 (2 ex. C&S), UMMZ 187779 (2 ex. sk., 203-217 mm SL)

“Cichlasoma” ornatum

FMNH 93130A (1 ex. C&S, 40 cm SL), FMNH 93102 (4 ex. alc., 51 – 145 mm

SL), FMNH (1 ex. Ridewood prep., 107 mm SL)

Thorichthys pasionis

UMMZ 196456 (13 ex. alc., 33-66 mm SL), UMMZ 143808 (66 ex. alc., 12-115

mm SL), UMMZ 184626 (1 ex. C&S, 67 mm SL)

Nandopsis ramsdeni

ANSP 68454 (holotype, 170 mm SL), ANSP 68455-68458 (4 ex. alc., 86-134 mm

SL), MCGJ 00342 (2 ex. alc., 91-107 mm SL), UMMZ 230839 (1 ex. alc., 104

mm SL; 1 ex. sk., SL unknown), UMMZ 231322 (13 ex. alc., 52-108 mm SL; 1

ex. sk., 105 mm SL; 1 ex. C&S), UMMZ 2345137 (1 ex. alc., 156 mm SL)

190 Amphilophus robertsoni

UMMZ 197222-S (2 ex. sk., 150-175 mm SL), UMMZ 223248-S (1 ex. sk., 175

mm SL), UMMZ 197263 (5 ex. alc., 76-131 mm SL), UMMZ 210886 (2 ex.

C&S)

“Cichlasoma” salvini

UMMZ 189958 (55 ex. alc., 31-100 mm SL), UMMZ 184758 (1 ex. C&S),

UMMZ 188062 (5 ex. sk., 75-96 mm SL), UMMZ 196651 (1 ex., Ridewood

prep.), UMMZ 197264 (5 ex. C&S), FMNH 96044 (4 ex. alc., 32-81 mm SL)

Crenicichla saxatilis

UMMZ 215923 (8 ex. alc., 109-156 mm SL), UMMZ 215935 (1 ex. alc., 95 mm

SL; 2 ex. sk., 142-165 mm SL), UMMZ 215614 (1 ex. C&S), UMMZ 147381 (1

ex. C&S), UMMZ 204938 (2 ex. C&S)

Archocentrus septemfasciatus

UMMZ 166478 (1 ex. alc., 70 mm SL), UMMZ 145722 (1 ex. C&S), UMMZ

180677 (3 ex. alc., 31-65 mm SL)

Tomocichla sieboldii

UMMZ 194240 (31 ex. alc., 27-95 mm SL; 4 ex. C&S, 1 ex. Ridewood prep.),

UMMZ 194240 (1 ex. sk., 94 mm SL), UMMZ 230707 (3 ex. alc., 108-123 mm

SL)

Archocentrus spilurus

UMMZ 210856 (1 ex. C&S), UMMZ 173195 (2 ex. C&S), UMMZ 197198 (1 ex.

C&S), UMMZ 199678 (1 ex. C&S), UMMZ 188136 (1 ex. C&S), UMMZ

191 197224 (1 ex. C&S), UMMZ 167696 (2 ex. C&S), UMMZ 190367 (1 ex. C&S),

UMMZ 210888 (1 ex. C&S), UMMZ 197198 (62 ex. alc., 12-77 mm SL)

Petenia splendida

UMMZ 144103 (8 ex. alc., 133-285 mm SL), UMMZ 189987-S (1 ex. sk., 220

mm SL), UMMZ 189960 (5 ex. alc., 49-210 mm SL), UMMZ 223251-S (1 ex.

sk., 230 mm SL), UMMZ 196665 (7 ex. C&S), UMMZ 196461 (1 ex. C&S)

Herichthys steindachneri

UMMZ 198800 (7 ex. alc., 37-150 mm SL; 1 ex. C&S), UMMZ 196348-S (1 ex.

sk., 150 mm SL), UMMZ 196348 (1 ex. sk., 125 mm SL)

Geophagus surinamensis

UMMZ 204939-S (2 ex. sk., 147-162 mm SL), UMMZ 214807 (10 ex. alc., 82-

146 mm SL)

Vieja synspila

UMMZ 144056 (33 ex. alc., 54-182 mm SL), UMMZ 188029 (1 ex. alc., 175 mm

SL; 1 ex. sk., 164 mm SL), UMMZ 189985 (1 ex. sk., 160 mm SL), UMMZ

190870 (2 ex. C&S), UMMZ 196577 (2 ex. C&S)

Nandopsis tetracanthus

AMNH 1063 (1 ex. alc., 119 mm SL), AMNH 96381 (3 ex. alc., 86-90 mm SL),

AMNH 96390 (4 ex. alc., 115-133 mm SL), AMNH 96426 (1 ex. alc., 110 mm

SL), AMNH 96454 (1 ex. alc., 96 mm SL), AMNH 96465 (1 ex. alc., 119 mm

SL), AMNH 96513 (1 ex. alc., 116 mm SL), CAS 78975 (2 ex. alc., 109-135 mm

SL), UMMZ 146972-S (1 ex. sk., 142 mm SL), UMMZ 171879 (4 ex. alc., 70-

110 mm SL; 1 ex. C&S), UMMZ 171880 (6 ex. alc., 97-130 mm SL; 1 ex. sk.,

192 124 mm SL; 1 ex. C&S), UMMZ 177285 (1 ex. alc., 115 mm SL), UMMZ

177286 (1 ex. alc., 73 mm SL), UMMZ 177287 (2 ex. alc., 65-90 mm SL; 1 ex.

C&S), USNM 078246 (1 ex. alc., 110 mm SL), USNM 126761 (2 ex. alc., 66-78

mm SL), USNM 33642 (1 ex. alc., 98 mm SL), USNM 64003 (2 ex. alc., 118-136

mm SL)

“Cichlasoma” trimaculatum

UMMZ 178854 (10 ex. alc., 31-128 mm SL), UMMZ 184744-S (1 ex. sk., 100

mm SL), UMMZ 191051-S (1 ex. sk., 122 mm SL), UMMZ 178854 (1 ex.

Ridewood prep.), UMMZ 190790 (2 ex. C&S), UMMZ 197120 (2 ex. C&S),

UMMZ 197112 (1 ex. C&S)

Tomocichla tuba

UMMZ 188316 (8 ex. alc., 102-211 mm SL; 1 ex. sk., 167 mm SL; 1 ex. C&S),

“Cichlasoma” urophthalmum

UMMZ 196584-S (2 ex. alc., 130-151 mm SL), UMMZ 196593-S (1 ex. sk., 128

mm SL), UMMZ 210869 (7 ex. alc., 25-155 mm SL; 1 ex. C&S), UMMZ 213483

(1 ex., Ridewood prep.), UMMZ 210869 (1 ex. C&S)

Vieja zonatum

UMMZ 184746 (39 ex. alc., 31-119 mm SL), UMMZ 168915 (2 ex. sk., 100-165

mm SL), UMMZ 184746 (1 ex. C&S, 65 mm SL)

193 LITERATURE CITED

Agassiz, L. (1859) Remarks on new fishes from Lake Nicaragua. Proceedings Boston Society of Natural History 6, 407-408.

Allgayer, R. (2002) Un Cichlidé nouveau du genre Tomocichla Regan, 1908 (Perciformes:Labroidei) du Rio Guarumo, Panama. L'an Cichlidé 2, 32-36.

Baird, S. F. and Girard, C.F. (1854) Descriptions of new species of fishes collected in Texas, New Mexico and Sonora, by Mr. John H. Clark, on the U. S. and Mexican Boundary Survey, and in Texas by Capt. Stewart Van Vliet, U.S.A. Proceedings of the Academy of Natural Sciences Philadelphia 7, 24-29.

Bremer, K. (1995) Branch support and tree stability. Cladistics 10, 295–304.

Bussing, W.A. (1998) Freshwater Fishes of Costa Rica. Editorial de la Universidad de Costa Rica., San Jose.

Bussing, W.A. (1989) “Cichlasoma” loisellei, a new Nandopsis group cichlid fish from Central America. Revista de Biologia Tropical 37, 153-161.

Casciotta, J.R., and Arratia, G. (1993) Tertiary cichlid fishes from Argentina and reassessment of the phylogeny of New World cichlids. Kaupia 2, 195-240.

Chakrabarty, P. (2006) Taxonomic status of the Hispaniolan Cichlidae. Occasional Papers Museum of Zoology, University of Michigan 737, 1-17.

Chakrabarty, P. (2006b) Systematics and Historical Biogeography of Greater Antillean Cichlidae. Molecular Phylogenetics and Evolution 39, 619-627.

Fernández-Yépez, A. (1969) Contribucion al conocimiento de los cichlidos. Evencias 22, 1-10.

Gill, T. N. (1862) Remarks on the relations of the genera and other groups of Cuban fishes. Proceedings of the Academy of Natural Sciences Philadelphia 14, 235- 242.

Gill, T. N. in Gill, T. N. and Bransford, J.F. (1877) Synopsis of the fishes of Lake Nicaragua. Proceedings of the Academy of Natural Sciences Philadelphia 29, 175-191.

194

Kullander, S.O. (1983) A revision of the South American cichlid genus Cichlasoma (Teleostei: Cichlidae). The Swedish Museum of Natural History, Stockholm.

Kullander, S.O. (1996) Heroina isonycterina, a new genus and species of cichlid fish from western Amazonia, with comments on cichlasomine systematics. Ichthyological Explorations of Freshwaters 7, 149-172.

Kullander (1998) A phylogeny and classification of the South American Cichlidae (Teleostei:Perciformes). In: (eds L.R. Malabarba, R.E. Reis, R.P. Vari, Z.M. Lucena and C.A.S. Lucena) Phylogeny and Classification of Neotropical Fishes. Porto Alegre, Edipucrs.

Kullander, S.O. and Hartel, K.E. (1997) The systematic status of cichlid genera described by Louis Agassiz in 1859: Amphilophus, Baiodon, Hypsophrys and Parachromis (Teleostei: Cichlidae). Ichthyological Explorations of Freshwaters 7, 193-202.

Liem, K.F. (1973) Evolutionary strategies and morphological innovations: Cichlid pharyngeal jaws. Systematic Zoology 22: 425-441.

López-Fernández, H. Honeycutt, R.L., Stiassny, M.L.J., and Winemiller, K.O. (2005) Morphology, molecules, and character congruence in the phylogeny of South American geophagine cichlids (Perciformes, Labroidei). Zoologica Scripta 34, 627-651.

Meek, S.E. (1904) The fresh-water fishes of Mexico north of the Isthmus of Tehuantepec. Field Columbian Museum Publication (Zoological Series) 93, 1-252.

Miller, R.R. (1966) Geographical distribution of Central American freshwater fishes. Copeia 1966, 773-802.7

Miller, R.R. (1976) Geographical distribution of Central American freshwater fishes. In: Investigations of the Ichthyofauna of Nicaraguan Lakes (ed T.B. Thorson), University of Nebraska, Lincoln.

Miller, R.R. (1993) [Review of] Conkel, Donald. Cichlids of North and Central America. Buntbarsche Bulletin 157, 18-19.

Miller, R.R. (2005) Freshwater Fishes of México. University of Chicago Press, Chicago.

Miller, R.R. and Nelson, B.C. (1961) Variation, life colors, and ecology of Cichlasoma callolepis, a cichlid fish from Southern Mexico, with a discussion of the Thorichthys species group. Occasional Papers Museum of Zoology, University of Michigan 622, 1-9.

Newsome, Y.L. (1971) Comparative osteology and relationships of Neotropical fishes. Unpublished Ph.D. Thesis, University of Illinois, Dept. Anatomy, Chicago, Ill.

195

Regan, C.T. (1905) A revision of the fishes of the American cichlid genus Cichlasoma and of the allied genera. Annals and Magazine of Natural History, Ser. 7, 60-77, 225-243, 316-340, 433-445.

Regan, C.T. (1908) A collection of freshwater fishes made by Mr. C. F. Underwood in Costa Rica. Annual Magazine of Natural History Ser. 8, 455-464.

Regan, C.T. (1906-1908) Biologia Centrali-Americana 8: Pisces. Taylor and Francis, London.

Roe, K.R., Conkel, D. and Lydeard, C. (1997) Molecular systematics of Middle American cichlid fishes and the evolution of trophic-types in ‘“Cichlasoma” (Amphilophus)’ and ‘C.(Thorichthys)’. Molecular Phylogenetics and Evolution 7, 366-376.

Sorenson, M.D. (1999) TreeRot, v. 2. Boston University, Boston, Massachusetts. Available from http://people.bu.edu/msoren/TreeRot.html

Stiassny, M.L.J. (1989) A taxonomic revision of the African genus Tylochromis (Labroidei, Cichlidae); with notes on the anatomy and relationships of the group. Annales Sciences Zoologiques Musee Royal L’Afrique Centrale Tervuren, Belgique 258, 1-161.

Stiassny, M.L.J., Chakrabarty, P., Loiselle, P.V. (2001) Relationships of the Madagascan cichlid genus Paretroplus, with description of a new species from the Betsiboka River drainage of northwestern Madagascar. Ichthyological Exploration of Freshwaters 12, 29-40.

Swofford, D.L. (2002) PAUP* Phylogenetic Analysis Using Parsimony (*and Other Methods), Version 4.0b. Sunderland, Massachusetts: Sinauer.

Taylor, J.N. and Miller, R.R. (1980) Two new cichlid fishes, genus “Cichlasoma”, from Chiapas, Mexico. Occasional Papers of the Museum of Zoology, University of Michigan 693, 1-16.

Taylor, W. and Van Dyke, G. (1985) Revised procedures for staining and clearing small fishes and other vertebrates for bone and cartilage study. Cybium 9,107-119.

Van Couvering, J.A.H. (1982) Fossil cichlid fishes of Africa. Special Papers in Paleontology, 29. The Paleontological Association, London.

196 CHAPTER V

NUCLEAR, MITOCHONDRIAL, AND MORPHOLOGICAL, COMBINED PHYLOGENETIC ANALYSES OF MIDDLE AMERICAN CICHLIDAE

“The arrangement of the American cichlids into genera is a puzzle of great difficulty. This is outstandingly true of the complex which has been classified in or near Cichlasoma.” - Hubbs (1936)

ABSTRACT

There are vexing taxonomic and biogeographic questions about the little-studied

Middle American radiation of cichlids. I present a phylogenetic analysis incorporating previous research on these cichlids, with the addition of new species and character sets.

A combined analysis of nuclear, mitochondrial, and morphological characters recovers a well resolved phylogeny of Middle American Cichlidae. One-hundred and nine species were sampled, including 92 heroines (76 geographically Middle American, 3 Greater

Antillean, 13 South American). Herichthys, Thorichthys, Nandopsis and Caquetaia are monophyletic; Parachromis, Vieja, Amphilophus, Archocentrus and Tomocichla are not recovered as clades. Neetroplus is a junior synonym of Hypsophrys; and Miller’s hypothesis that Caquetaia is a junior synonym of Petenia is rejected. Heroini is a monophyletic group only with the exclusion of Hypselecara. Heroines are recovered as two major divisions, both clades, of Middle American cichlids. Division I is a clade composed

197 of 27 species of Middle American cichlids and its sister group, the mainly South

American Caquetaia. Division II includes a clade of 51 Middle American species and its

sister group, the endemic Greater Antillean Nandopsis. Dispersal of Middle American

cichlids to South America appears to be more prevalent than dispersal of South American

cichlids to Middle America. Four geographically South American cichlids are

phylogenetically Middle American, indicating potentially recent dispersal from Middle

America to South America. There is no evidence of recent dispersal from South America

to Middle American recovered in these analyses.

INTRODUCTION

The relationships of Middle American cichlids are poorly understood. All but

two of the 111 species belong to the Heroini (Kullander, 1998). Heroines include most

Middle American species, the Greater Antillean cichlids, and some South America

species (Kullander, 1998). Taxonomy of this group was thrown into disarray when the

catch-all Neotropical cichlid genus Cichlasoma was restricted to a dozen South American

species; this restriction left nearly all the Middle American cichlids without generic

assignments. Without these assignments and recognition of clades, evolutionary studies

on these fishes are hindered, and biogeographic history cannot be explored. Without a

phylogeny, taxonomic decisions that have been proposed to reflect common history are

speculative.

Long lingering taxonomic issues remain for Middle American cichlids. Most of the heroine radiation was once part of the catch-all Cichlasoma (Kullander, 1983). That

generic epithet once applied to the entire heroine radiation plus part of the closely related

198 Cichlasomatini. Kullander (1983) restricted Cichlasoma to 12 South American

cichlasomines (all in the Cichlasomatini). The heroines were left without generic

placement except for the five species in Thorichthys and the monotypic Petenia. Stiassny

(1991) recognized heroines (her Cichlasoma Group B) but did not recognize or name

clades within that group. Miller (1966, 1976) divided heroine Cichlasoma into informal

sections following the work of Regan (1906 -1908). Some authors thought it best to raise

these sections to genera as placeholders until future work could be done (Martin and

Bermingham, 1998). Others felt that giving a single name to these discarded Cichlasoma

species was the best course (Loiselle, 1984; Burgess and Walls, 1993). Potential names

for this radiation include Herichthys Baird and Girard, 1854 and Heros Heckel, 1840.

Kullander (1983) suggested that species without generic assignments be called

“Cichlasoma” until genera were assigned. Recently, Kullander provided a checklist of

Neotropical cichlids that assigned genera to many of these species (Kullander, 2003).

Unfortunately, these placements were not presented with phylogenetic analyses or

explanation.

Phylogenetic analyses have supported monophyly of few Middle American

genera. Herichthys was recovered as monophyletic by mitochondrial phylogenetic

analyses of Hulsey et al. (2004) and Thorichthys by Roe et al. (1997); these authors did

not diagnose these clades with characters. (It should be noted that in both works not all

species of those genera were sampled.) On the other hand, Kullander (2003) defined

groups but did not present phylogenetic support. My analyses set a phylogenetic framework for diagnosing clades across a wide group of Middle American species.

199 Previously, Hulsey et al. (2004) presented a cytochrome b (cyt b) phylogeny of 52

heroines combining the work of previous researchers (genbank sequences of: Lydeard

and Roe, 1997; Roe et al., 1997; Roe and Lydeard unpublished; Martin and Bermingham,

1998; Kumazawa et al., 2000; Farias et al., 2001; Rican et al. unpublished). Ninety-two

heroines are sampled here, combining the cytochrome b sequences from Hulsey et al.

(2004) and adding sequence from cytochrome oxidase 1 (COI), the ribosomal subunit

16S, the nuclear gene TMO-4C4, the first intron from nuclear gene S7 and morphological characters (from Chapter IV). This is the largest study of this radiation to date and includes the majority of the heroine radiation; this study focuses particularly on the little studied Middle American representatives.

MATERIALS AND METHODS

Molecular methods follow those in Chapter II. Vouchers are deposited in the

UMMZ DNA collection for most species from which DNA was extracted and sequenced

(see Materials Examined). Morphological analyses and resulting characters follow the

results in Chapter IV.

Aligned cytochrome b sequences from Hulsey et al., (2004), were obtained from

the first author, Darrin Hulsey. Species that appeared either paraphyletic or polyphyletic

in Hulsey et al. (2004) were not sampled here. One representative sequence was selected

if multiple copies were available. All S7 sequence are from Chakrabarty (2006; Chapter

III). TMO-4C4, 16S and COI sequencing and extraction follow the procedure in

Chakrabarty (2006; Chapter III). Novel sampling of TMO-4C4, 16S and COI sequence

are listed in Table V.1 and in the Materials Examined section.

200 Parsimony analyses were completed in PAUP* 4.0b (Swofford, 2002). Heuristic

searches were performed with 1,000 random addition replicates for each analysis based

on a single data partition. Jackknife resampling (100 replicates of 10 search replicates)

was performed in NONA (Goloboff, 1993) and WinClada (Nixon, 1999). For combined

analyses the parsimony ratchet (Nixon, 1999) was implemented in PAUP* by using

PAUPRat (Sikes and Lewis, 2001) with 5% to 25% of the total characters perturbed

(allowed to change weights) over 100 to 2000 replicates until a stable solution was found

(20 runs). A Malagasy cichlid, Paratilapia polleni, was used to root all trees.

Optimizations were performed in MacClade 4.0 (Maddison and Maddison, 1992).

Cichla ocellaris and Crenicichla saxatilis were sampled only for morphological

features. Cichla temensis and Crenicichla acutirostris were sampled only for molecular

characters. These species were used to make composite taxa to represent their respective

genera, Cichla and Crenicichla. Because these genera are important outgroups, creating composites was favored over deletion. Analyses with Cichla and Crenicichla as composites, and analyses with all individual species treated separately are shown in the

Results.

RESULTS

Aligned sequences and morphological characters yielded 3523 characters for each of the 109 taxa. S7 primers yielded 774 aligned positions, Tmo-4C4 primers yielded 299

aligned positions, 16S primers yielded 614 aligned positions, COI primers yielded 591

aligned positions, cytochrome b sequences (from Hulsey) totaled 1148 aligned positions;

there were 89 morphological characters. Figures V.1-4 show molecular phylogenies based on single gene fragments (16S, COI, S7, TMO-4C4). Figure V.5-6 show the total

201 evidence analyses based on those gene fragments combined with cyt b sequences sampled and aligned in Hulsey et al. (2004), and from the morphological study of

Chapter IV. The total evidence approach yields a relatively well resolved tree. Figure

V.5 shows a tree with no composite taxa. Figure V.6 shows a tree with composite taxa,

Cichla and Crenicichla (see Materials and Methods).

As composite taxa, Cichla and Crenicichla are recovered with other South

American outgroups (Fig. V.6). As separate taxa, Cichla temensis and Crenicichla acutirostris (which were sampled for molecular characters) fall in their traditional South

American clades but Cichla ocellaris and Crenicichla saxatilis are nested within Middle

American cichlids and sister to Petenia splendida (Fig.V.5). This atypical placement is due to these taxa not being sampled for molecular characters as was done for all other

South American taxa in the combined analyses. Therefore, the phylogeny with composite taxa (Fig. V.6) is the preferred hypothesis of relationships.

Jackknife support over 80 percent is shown on Figure V.6. There is little support for deeper divergences. This lack of support may be due to missing data in large portions of the complete dataset.

There is generally greater resolution in the combined analysis than any of the partitioned datasets. The 16S phylogeny (Fig.V.I) is best able to resolve more recent divergences such as among Parachromis and Vieja but little resolution is provided for the

Heroini. A large polytomy is recovered in this partition that includes much of the ingroup. Seventy-five taxa were sampled for 16S.

The COI phylogeny (Fig.V.2) provides more resolution on deeper nodes than 16S but still mostly contributes to more recent divergences. In this phylogeny Nandopsis is

202 monophyletic. Nandopsis is the sister group to a different clade of cichlids in this

analysis than in the combined analyses. This incongruence may be due to a lack of signal in COI for older divergences. In this phylogeny, there is a polytomy recovered for most of the deeper divergences. Seventy-one taxa were sampled for COI.

The TMO-4C4 phylogeny (Fig.V.3) provides much of the resolution on deeper

nodes. This phylogeny recovers a monophyletic Heroini, including Hypselecara. TMO-

4C4 provides almost no resolution within Heroini. Fifty taxa were sampled for TMO-

4C4.

The S7 phylogeny has limited taxonomic scope because of low sampling but

provides resolution among deeper and some more recent divergences. There were only

30 species sampled for the S7 partition, all samples were previously sampled by

Chakrabarty (2006; Chapter III).

A parsimony optimization on area is shown in Figure V.7. This phylogeny is

identical to that in Figure V.6. All Middle American cichlids are rooted within South

America. Middle American cichlids are more closely related to each other than to South

American lineages in all cases. Four geographically South American species were found

to be phylogenetically Middle American with this optimization. Two large clades of

Middle American cichlids were recovered. One clade is sister to the mainly South

American Caquetaia. The other Middle American clade is sister to the Greater Antillean

Nandopsis.

203 Figure V.1: Phylogeny of Middle American cichlids based on 16S sequence. Jackknife values of 80 or above are shown. This tree is a strict consensus of 76,582 trees (length 1261), CI = .344, RI = .513

204 Figure V.2: Phylogeny of Middle American cichlids based on COI sequence. Jackknife values of 80 or above are shown. This tree is a strict consensus of 39 trees (length 2308), CI=.219, RI = .506.

205

Figure V.3: Phylogeny of Middle American cichlids based on TMO-4C4 sequence. Jackknife values of 80 or above are shown. This tree is a strict consensus of 135 trees (length 133), CI=.737, RI=.715.

206

Figure V.4: Phylogeny of Middle American cichlids based on S7 sequence. Jackknife values of 80 or above are shown. This tree is a strict consensus of 252 trees (length 682), CI = .801, RI=.687.

207 Figure V.5: Phylogeny of Middle American cichlids based on total evidence (16S, COI, TMO-4C4, S7, cyt b, and morphology). Jackknife values of 80 or above are shown. A strict consensus of 61 trees (length 8527), CI =.335, RI =.494.

208 Figure V.6: Phylogeny of Middle American cichlids based on total evidence (16S, COI, TMO-4C4, S7, cyt b, and morphology) with composite taxa for Cichla and Crenicichla (see Materials and Methods). Jackknife values of 80 or above are shown, species with * are South American. A strict consensus of 447 trees is shown (length 8536), CI =.334, RI =.493.

209 Figure V.7: Phylogeny showing a parsimony optimization of biogeographic area on the phylogeny of Middle American cichlids (same topology as Fig.V.6).

210 DISCUSSION

All Middle American cichlids sampled were recovered as more closely related to each other than to South American lineages. There appear to be two major divisions of

Middle American cichlids. Division I is a clade of 27 species and its sister group, the mainly South American Caquetaia. Division II is a clade composed of 51 Middle

American species and its sister group, the endemic Greater Antillean Nandopsis. These two divisions are sister groups (Fig.V.7). The parsimony optimization is equivocal about the origins of each of these divisions but together as a clade they are nested within South

America. Based on the pattern of relationships the heroine radiation appears to be ancient.

If cichlids had invaded Middle America recently it would be expected that there would be evidence of multiple invasions by different South American lineages. A representative phylogeny would show some geographically Middle American cichlids being more closely related to South American lineages. No members of South American cichlid lineages that are present in Middle America are recovered in the total evidence phylogeny (Fig.V.7). However, it is likely that two South American cichlids did recently invaded Middle America. Geophagus crassilabris and Aequidens coeruleopunctatus; unfortunately neither is sampled here. Both are endemic to lower Central America and neither are heroine cichlids. Geophagus crassilabris and Aequidens coeruleopunctatus are members of separate South American lineages. That these apparently recent South

American invaders are restricted to Middle America suggests that they speciated there and have been there for sometime. Geophagus and Aequidens are not closely related to

Middle American heroines (Kullander, 1998).

211 Bussing (1985) saw Middle American cichlids as part of an ancient South

Americ an radiation that dispersed into Central America in the Late Cretaceous or

Paleocene. These cichlids were subsequently stranded on this area during the Tertiary and were only reunited with their ancestral source during the Pliocene closure of the

Isthmus of Panama. He also included a characin (Astyanax), and several cyprinodontiformes (Poecilia, Poeciliopsis, Cyprinodon, Floridichthys, Heterandria,

Profundulus and Fundulus) in this “Old Southern Element.” His conclusions were derived from distributions and not phylogenetic analyses. However, at least for the

Middle American Cichlidae (which was then understood to be part of “Cichlasoma”) this scenario is supported by phylogenetic evidence (Chakrabarty, 2006; Chapter III).

Chakrabarty (2006) dated the Middle American radiation to be between 50 and 72 million years old (Chapter III, Fig. III.1). This period in the Late Cretaceous/Paleocene corresponds to a time when an island arc (The Cretaceous Island Arc) connected South

America to the Yucatan Peninsula (Iturralde-Vinent and MacPhee, 1999; Pitman et al.,

1993). Presumably at that time, cichlids dispersed from South America across the

Cretaceous landbridge to Yucatan.

Martin and Bermingham (1998) sampled 17 Costa Rican cichlid species and concluded that the heroine radiation of cichlids was Middle to Late Miocene (specifically

15 to 18 mya) in age, a significantly younger age than found by Chakrabarty (2006). The estimate by Martin and Bermingham (1998) is based on cytochrome b sequence divergence rates from “marine fishes.” Their approach of taking the average divergence from distantly related and taxonomically diverse marine species and applying it to Middle

212 Americ an cichlids is extremely problematic. Their analysis is suspect because they did not estimate rates, and variability in rates, within their cichlid phylogeny.

Their estimate of rates uses the assumption that all taxa (at least all percomorphs) have the same rate of evolution for cyt b.

Similarly, their assertion that “the ability of many cichlids to tolerate saline conditions suggests that they may be capable of rapid dispersion along coastlines” is both false and immaterial. It is false because only a few cichlids are tolerant of the marine environment (Sparks and Smith, 2005). Moreover, salt tolerance does not establish that marine dispersal took place. The conclusion of Martin and Bermingham (1998) leads one to ask why their proposed dispersal event took place only once. Why not multiple times over the course of the 15 million years after the singular event they propose? The use of a static, non-cichlid specific, estimate of divergence time makes Martin and

Bermingham’s (1998) conclusion that the Middle American cichlids are 15 million years old unfounded. The estimate that this radiation is between 50 and 72 million years old by

Chakrabarty (2006) is more rigorous than the estimate of Martin and Bermingham because it was based on rates within a phylogeny of Neotropical cichlids and those rates were allowed to vary across that phylogeny. To be fair, Martin and Bermingham’s

(1998) study was conducted before the wide-spread acceptance of rate smoothing

(relaxed clock) techniques, and was done at a time when applying a single rate for mitochondrial genes across widespread taxa was still accepted.

That the Middle American cichlids form a clade means that a single lineage gave rise to them. Apparently it was the loss of the landbridge connection between Yucatan and South America in the Cretaceous that led to isolation on the Yucatan Peninsula. As

213 modern Central America formed with the movement of the Chortis Block, Nicaragua

Rise, and Isthmus of Panama sequentially filling the gap between North and South

America these Yucatan cichlids would have been able to take advantage of these new

areas and diversify. There is evidence that this progressive southern movement includes

a return to South America.

There are several invasions of cichlids from Middle America into South America,

based on the phylogenies presented here. “Cichlasoma” ornatum, C. festae, C.

atromaculatum and C. facetum all are phylogenetically Middle American cichlids found

in South America. Among these, only the node with “Cichlasoma” festae was also sampled and dated in the biogeographic study of Chakrabarty (2006). The age of this node (and the outgroup node below it) was estimated to be between 42 million and 56 million years old in that study. This period corresponds to a time when the Aves Ridge may have connected Yucatan to South America (Pitman et al., 1993). As in Chakrabarty

(2006; Chapter III) this species is recovered within the Middle American clade, but the

additional sampling of the current study suggests a much younger age. Although the

node to which this species belongs may be Eocene in age, “Cichlasoma” festae itself

may be far younger. In this study this species is recovered in a derived polytomy of five

species that is sister to a clade of 27 species.

The monophyletic Caquetaia is a South American lineage with one species,

Caquetaia umbrifera, found both in South America and in Panama. The results of the

parsimony optimization are largely dependent on the coding of this species. If coded as

Middle American, the ambiguous portions of the optimization all become Middle

214 American. If it is coded as South American, only the Caquetaia branch becomes South

American, and the rest of the optimization remains the same.

Caquetaia is sister to a large clade of Middle American cichlids. “Cichlasoma”

atromaculatum is one of only a few species that are found in both Middle America

(Panama) and South America. Notably, all of the species that are phylogenetically

Middle American but native to South America were determined to be Middle American

much earlier by C. Tate Regan. Regan (1906-1908) stated that “Cichlasoma” festae, C.

ornatum, C. atromaculatum, and Caquetaia were members of his Middle American section ‘Nandopsis’ and that “The South American species of this section are probably derived from immigrants from Central America.”

The presumed great biotic interchange between North and South America after the closure of the Isthmus of Panama does not seem to be as important for cichlids as the much earlier interchange through the Cretaceous Island Arc (Iturralde-Vinent and

MacPhee, 1999). Surprisingly, more species from Middle America have recently invaded

South America than the reverse.

Biogeography of the Greater Antilles

The closest relatives of Greater Antillean cichlids are Middle American

(Chakrabarty, 2006; this study). The Miocene fossil, Nandopsis woodringi, has often

been used to provide a minimum age to the entire Middle American radiation (Myers,

1966). In the total evidence analysis of the current study, Nandopsis is sister to a large

clade of over 50 Middle American cichlids. Based on divergence estimates that dated

Nandopsis to be between 55 and 38 million years old, Chakrabarty (2006) suggested that

the Greater Antillean cichlids were separated from Middle American representatives in

215 the Eocene. He suggested that during this period a vicariance event separated

populations of an ancestral Middle American species inhabiting a contiguous area comprising the Paleogene arc (Hispaniola and Cuba) and Yucatan. The drifting of the

Paleogene arc led to the allopatric speciation event that gave rise to the Greater Antillean

cichlids.

That Nandopsis is recovered as sister to a large clade of Middle American cichlids

may be regarded as evidence against recent marine dispersal. If the Greater Antillean

cichlids were the product of recent marine dispersal from Middle America a number of

phylogenetic patterns would be expected. First, instead of being sister to a large clade of

over 50 species, a more recent split with a tip clade of one or few species would be expected. The species of Division 1 together are older than any of the less general

groupings within that clade. Therefore, Greater Antillean cichlids are presumed to be

older for being sister to all Division 1 Middle American cichlids than if they were sister

to only a few species in that clade. In a marine dispersal scenario it might be expected

that members of Nandopsis would still be present on the mainland (e.g., Mexican

populations of N. tetracanthus); we might also expect evidence of dispersal between

islands (e.g., N. haitiensis populations on both Cuba and Hispaniola). A phylogeny that

finds the Greater Antillean cichlids are paraphyletic, would reject a vicariance scenario

and favor marine dispersal. Paraphyly would be explained in this case by independent

dispersal events from the mainland. That Nandopsis is monophyletic, and sister to a large clade rather than a small one, suggests vicariance origins for the Greater Antillean

cichlids.

216 Under a marine dispersal scenario monophyly of the Greater Antillean cichlids

would require a number of ad hoc assumptions. Dispersal from the mainland to one of

the islands would require subsequent extinction of the mainland ancestor (because the

island cichlids are more closely related to each other) and no subsequent successful

dispersal between mainland and island. The initial marine dispersal event would have

had to have happened before the end of the Miocene (as Nandopsis woodringi is a

Miocene fossil). A single dispersal event from that island (either Cuba or Hispaniola) to

the other island would be required without subsequent successful dispersal between islands. The vicariance scenario described above provides a simpler explanation an assumption-laden marine dispersal hypothesis.

Endemic Greater Antillean species of gars and swamp eels may share a similar history with cichlids. Lepisostids (gars), which are found in North America, parts of

Central America and Cuba must have taken a Northern route to the Antilles (Wiley,

1976). Currently there are no estimates of divergence times for gars. Synbranchids

(swamp eels) have a similar distribution in the Neotropics as cichlids and may have taken the same freshwater routes to the Greater Antilles before the fragmentation of the island arc. However, Perdices et al. (2005) argue that swamp eels arrived in the Greater

Antilles in the Miocene based on a molecular clock analysis (see comments by

Chakrabarty, 2006; Chapter III).

The absence of characids from the Greater Antilles is puzzling. Huge populations of species like Astyanax aeneus in Central America would suggest that a long standing connection with the Great Antilles would have allowed these species into those areas.

However, Astyanax and other characids are conspicuously absent from many drainages in

217 modern day Middle America; presence of large populations in one area does not

necessarily imply high densities in every surrounding area. It is also possible that

characids were not present on the Yucatan Peninsula when freshwater routes were

available to the Greater Antilles. Myers (1938) suggested that perhaps none of the

Ostariophysi were in Middle America during the period when a dry land connection (i.e.,

the Paleogene arc) might have provided freshwater routes from Middle America to the

Antilles. This scenario would also explain the conspicuous absence of these fishes from

these islands.

The aplocheiloid genus Rivulus has a congruent phylogenetic and divergence

pattern with cichlids. Murphy and Collier (1996) found a phylogenetic pattern on a

tempor al scale corresponding to vicariance origins for Rivulus in the Greater Antilles.

They used a 70-80 mya calibration point associated with the period that the Cretaceous

island arc functioned as a corridor between North and South America. This is the same

route suggested for cichlids by Chakrabarty (2006).

The lack of any native freshwater fish on Puerto Rico may be evidence of a large

freshwater extinction event on that island or perhaps on the entire Cretaceous Island Arc.

No evidence that this island arc (mainly composed of Cuba, Hispaniola, Caymen Ridge,

and Puerto Rico; Pindell and Barrett 1990) has maintained freshwater fishes is recovered

in any phylogeny (Greater Antillean species would be sister to all Middle American

species in such a phylogeny). Unlike Cuba and Hispaniola, Puerto Rico was not part of the Paleogene Arc. Therefore, Puerto Rico was not part of the biogeographic event that allowed for the transfer of mainland species over freshwater routes in the Eocene; this

218 may explain the absence of freshwater species from this island. Cuba and Hispaniola

each have more than 25 endemic freshwater fish species.

Taxonomy

Naming genera based on clades is the best course of action for resolving Middle

American cichlid taxonomy. Heroine cichlids in Middle America that were grouped

under the informal moniker “Cichlasoma” could be restricted to one genus, either Heros

Heckel (Loiselle, 1984) or Herichthys Baird and Girard (Burgess and Walls, 1993).

However, taxonomists have already begun the process of splitting this large group into

genera of a dozen or so species each (Schmitter-Soto, in review; Kullander, 2003;

Chakrabarty, 2006b). Many of these taxonomic decisions were based on historical constructs (the informal sections of Miller, 1966 and Regan, 1906-1908). Subsequent

phylogenetic analyses found two of these groups, Thorichthys (Roe et al., 1997) and

Herichthys (Hulsey et al., 2004), to be monophyletic. The total evidence analysis here

represents the combination of all the molecular and morphological phylogenetic work

that has been done on this group to date.

The Heroni are a monophyletic group only with the removal of Hypselecara.

South American heroines (Uaru, Symphysodon, and Heros) are the sister group to Middle

American heroines. Hypselecara, nominally a South American heroine, was not

recovered with other heroines but among outgroups.

Nandopsis Gill, 1862 is a monophyletic group restricted to the Greater Antillean

species, supporting the conclusions of Chakrabarty (2006, Chapter III, IV). This genus is

sister to a large group of Middle American cichlids that includes most Vieja, Herichthys,

Thorichthys and Amphilophus species.

219 Herichthys Baird and Girard, 1854 is monophyletic. All Herichthys are sampled

here for the first time. “Cichlasoma” deppii should be recognized as Herichthys deppii

(Heckel, 1840) as suggested by Kullander (2003) because it is recovered as the sister

group to the remaining Herichthys. Herichthys, including H. deppii, is the sister group to

a clade of Vieja species. Herichthys species are united by breeding coloration and

dentition (Miller, 2005; Kullander, 1996).

Vieja Fernandez-Yepez, 1969 is paraphyletic as currently defined. Fourteen of

the 15 nominal species sampled in this phylogeny form a clade with Paraneetroplus

bulleri, but Vieja tuyrense was found distantly related. It has been previously suggested that the genus may not be monophyletic (Miller, 2005). The type of the genus, Vieja maculicauda is nested within this “mainly-Vieja” clade (Paraneetroplus bulleri + all

Vieja except V. tuyrense). However, the oldest available name for this mainly-Vieja

clade would be Paraneetroplus Regan, 1905. Paraneetroplus bulleri is the type of the

genus, which includes three more species. Because the mainly-Vieja clade is poorly

resolved, it is unclear where Paraneetroplus bulleri lies within Vieja. Without that information a synonym would be premature.

Thorichthys Meek, 1904 is monophyletic. Only Thorichthys socolofi was not sampled in this study. Thorichthys is found to be sister to a large clade of Amphilophus species. Miller and Nelson (1961) noted Thorichthys has unique sub-opercular pigment

patterns and five versus four mandibular pores.

Amphilophus Agassiz, 1859 is polyphyletic. Amphilophus altifrons + A.

robertsoni + A. bussingi + A. diquis + A. longimanus + A. rostratus + A. rhytisma +

“Cichlasoma” facetum is a clade that is sister to Amphilophus macrocanthus. This clade

220 is sister to Thorichthys. Amphilophus hogaboomorus, A. calobrense and A. lyonsi are distantly related to each other and other Amphilophus species. Amphilophus citrinellus +

A. labiatus is distantly related to other Amphilophus species. Amphilophus labiatus is the type species of the genus. The remaining nominal Amphilophus should be renamed and a revision is required.

Archocentrus Gill and Bransford, 1877 is polyphyletic as currently recognized.

The type species Archocentrus centrarchus is sister to Amphilophus citrinellus + A. labiatus. Closely related to that clade is a clade composed of Archocentrus nanoluteus +

A. septemfasciatus + A. panamensis + A. nigrofasciatus + A. spilurus + Theraps wesseli.

Theraps Günther, 1862 is the oldest available name for this clade but Theraps coeruleus

T. irregularis, T. lentiginosus were not sampled in this study. Without sampling the type

species, Theraps irregularis, a name change would be unwarranted for this clade.

Parachromis Agassiz, 1859 is paraphyletic. Archocentrus myrnae and A. sajica are sister taxa that are nested within Parachromis. All members of Parachromis were sampled but despite a strong resemblance between these species there is no support for their monophyly. Hulsey et al. (2004) recovered Archocentrus myrnae + A. sajica as nested with Parachromis species. Martin and Bermingham (1998) also reported the same

sister pair within a paraphyletic Parachromis. In the current study nuclear genes and

morphology were not sampled for Archocentrus myrnae or A. sajica, but these species

were sampled herein for 16S and COI. Notably, Parachromis is monophyletic in the 16S

phylogeny (Figure V.1).

221 Caquetaia has recently been suggested to be a junior synonym of Petenia (Miller

and Norris in Miller, 2005). That notion is rejected here as Petenia is not found to be

closely related to Caquetaia.

The monotypic Neetroplus is a junior synonym of the monotypic Hypsophrys.

Hypsophrys nicaraguensis (Günther, 1864) is recovered as the sister group to the

Neetroplus nematopus Günther, 1867. Hypsophrys nicaraguensis was originally

described as a Heros. Günther (1867) described Neetroplus as differing from Heros in

“having a front series of flat incisor-like teeth.” Hypsophrys Agassiz, 1859 is a nomen

nudum. As Hypsophrys nicaraguensis also has partially spatulate teeth, as is diagnostic

for Neetroplus nematopus, and because they are sister taxa, these monotypic genera

should be synonymized. They also both share having a single predorsal spine (versus the

two of most Middle American species), and having a rounded snout with a subterminal

mouth. Therefore, Neetroplus nematopus should be recognized as Hypsophrys nematopus

(Günther, 1867).

Tomocichla Regan, 1908 is a polyphyletic group, despite these species sharing a

number of morphological features (see Chapter IV). A lack of overlapping data partitions

between Tomocichla tuba (morphology, cyt B) and Tomocichla asfraci (16S, COI) may explain why they are recovered distantly related despite a complete dataset sampling for

Tomocichla sieboldi. Tomocichla was also found to be polyphyletic in the cyt b

phylogeny of Martin and Bermingham (1998).

222

Despite inclusion of most cichlid species from Middle America, the total evidence

phylogeny here is incomplete on two fronts. First, not all species were available for

samplin g. Second, many of the taxa lack complete data, which may influence their

position on the phylogeny. I refrain here from renaming many taxa, particularly if all

members of the genus were not sampled. Despite the incomplete nature of the phylogeny,

it will serve an important role in providing a framework for future study. Clearly, Vieja,

Amphilophus, Archocentrus, Parachromis need further study and revision. Also

scattered about the phylogeny are former members of Cichlasoma that require generic assignments. So long as the placeholder “Cichlasoma” is in use the history of

Neotropical Cichlidae is hidden.

223 MATERIALS EXAMINED

Morphological materials examined are listed in Chapter IV. Molecular specimens that are sampled in Chapter III are listed in that chapter (they are listed as PC 2006 in Table

V.I). Novel sampling of molecular specimens that are preserved as vouchers in the

UMMZ (University of Michigan Museum of Zoology) are listed below. For six species, there are no voucher specimens; for these species a fin clip was taken from a live animal and sent to the UMMZ. These live specimens are kept in the aquaculture facility of Jeff

Rapps (www.tangledupincichlids.com) who sells young of wild caught specimens that he breeds. Individuals caught by the author and others in nature are noted below as “wild caught.” All others are aquarium specimens purchased from Jeff Rapps.

Aequidens diadema – no voucher, fin clip

Amphilophus calobrense - no voucher, fin clip

Amphilophus labiatus – UMMZ 245140

Amphilophus robertsoni - UMMZ 246445

Archocentrus myrnae - UMMZ 245694

Archocentrus nanoluteus - UMMZ 245140

Archocentrus panamensis – UMMZ 245693

Archocentrus sajica – UMMZ 245673

Archocentrus septemfasciatus – UMMZ 243191

Archocentrus spilurus – UMMZ 246294, wild caught Belize

Archocentrus spinosissimus – UMMZ 245675

Caquetaia kraussii – UMMZ 245682

Caquetaia myersi – UMMZ 243194

224 Caquetaia spectabilis – no voucher, fin clip

Caquetaia umbrifera – UMMZ 245680

Cichla temensis – no voucher, fin clip

“Cichlasoma” deppii – no voucher, fin clip

“Cichlasoma” facetum – UMMZ 243186

“Cichlasoma” cf. facetum-oblongus (This species is listed as Cichlasoma oblongus in the

aquarium trade but is likely Cichlasoma facetum species that are introduced in Central

South America) – UMMZ 243178

“Cichlasoma” grammodes – no voucher, fin clip

“Cichlasoma” trimaculatum – UMMZ 245677

“Cichlasoma” octofasciatum – UMMZ 246443 wild caught, Mexico

“Cichlasoma” urophthalmum – UMMZ 246452 wild caught, Mexico

“Cichlasoma” salvini – UMMZ 246450 wild caught, Mexico

Crenicichla acutirostris – UMMZ 243196

Etia nguti – UMMZ , specimen is alive and not given a catalogue number

Gymnogeophagus gymnogenys – UMMZ 243184

Hemichromis letourneuxi – UMMZ 245138

Herichthys labridens – UMMZ 245685

Herichthys tamasopoensis – UMMZ 245687

Parachromis friedrichsthali – UMMZ 246460 wild caught, Mexico

Parachromis loisellei – UMMZ 245689

Petenia splendida – UMMZ 246451 wild caught, Mexico

Satanoperca jurupari – no voucher, fin clip

Tahuantinsuyoa macantzatza – UMMZ 243187

225 Teleocichla monogramma – UMMZ 243172

Theraps wesseli – UMMZ 245688

Thorichthys ellioti – UMMZ 245686

Thorichthys meeki – UMMZ 246441

Thorichthys pasionis – UMMZ 246459 wild caught, Mexico

Vieja argentea – no voucher, fin clip

Vieja b ifasciata – no voucher, fin clip

Vieja g odmanni – UMMZ 246295 wild caught, Belize

Vieja heterospilus – UMMZ 243198

Vieja intermedia - UMMZ 246293

Vieja regani - UMMZ 245684

Vieja synspila - UMMZ 246453, wild caught, Mexico

Vieja maculicauda - UMMZ 245690

Vieja ‘Belize’ melanurus (This taxon differs externally in a number of color features from V. synspila and was thought to be potentially novel when caught by the author and others in the field) - UMMZ 246292, wild caught, Belize

Vieja ufermanni - UMMZ 245692

226 Table V.1: List of sampled species. Distributions from Cata logue of Fi shes (Esch meyer) . NC A = ( North ern Ce ntral Amer ica) = Guatemala, Belize, El Salvador, Honduras; Nucle a r CA = Co sta Rica, H onduras and Nicaragua; Southern C A = Costa Rica a nd Panama. PC 2006 is Chakrabarty, 2006 or Chapter III. Hulsey et al. = samples from Huls ey et al., (2004) s ee tex t.

Middle America Morpholo g y 16 S C OI T mo-4c 4 S7 Cyt b Lo cality Heroines Amphilophus altifrons Hulsey et al. Sout hern C A Amphilophus bussingi Hulsey et al. Sout hern C A Amphilophus calobrense This study This study Panama Amphilophus citrinellus PC 2006 PC 2006 P C 2006 PC 20 06 Hulsey et al. Sout hern C A Amphilophus diquis Hulsey et al. Cos ta Rica Amphilophus hogaboomorus Chapter I V Honduras Amphilophus labiatus Th is stud y Hulsey et al. Nic aragua Amphilophus longimanus Hulsey et al. Nuclear CA Amphilophus lyonsi PC 2006 PC 2006 PC 2006 PC 2006 Southern CA 227 Amphilophus macracanthus Chapter IV Hulsey et al. Mexico, NCA Amphilophus rhytisma Hulsey et al. Costa Rica Amphilophus robertsoni Chapter I V This study This study Hulsey et al. Mexico, NCA Amphilophus rostratus Hulsey et al. Southern CA Archocentrus centrarchus PC 2006 PC 20 06 PC 2006 PC 2006 Hulsey et al. Nuclear CA Archocentrus multispinosus PC 2006 PC 2006 PC 2006 PC 2006 Hulsey et al. Nuclear CA Archocentrus myrnae This study This study Hulsey et al. Southern CA Archocentrus nanoluteus This study Panama Archocentrus panamensis This study This study Panama Archocentrus nigrofasciatus Chapter IV PC 2 006 PC 2006 PC 2006 PC 2006 Hulsey et al. CA Archocentrus sajica This study This study Hulsey et al. Costa Rica Archocentrus septemfasciatus Chapter IV This study This study This study Hulsey et al. Costa Rica Archocentrus spilurus Chapter IV This study This study This study Hulsey et al. Nuclear CA Archocentrus spinosissimus This study This study Guatemala

227 Table V.1 continued

Middle America Morphology 16S COI Tmo-4c4 S7 Cyt b Locality Heroines

Caquetaia umbrifera This study T his study Hulsey et al. Pana ma, North SA

“Cichlasoma” beani Chapter IV Mexico

“Cichlasoma” deppii This study This study This study Mexico

“Cichlasoma” grammodes Chapter IV This study This study This study Mexico,NCA

“Cichlasoma” istlanum Chapter IV Mexico

“Cichlasoma” cf. facetum-oblongus This study Thi s study Introduced SA “Cichlasoma” trimaculatum Chapter IV This study This study Hulsey et al. Mexi co,NCA “Cichlasoma” octofasciatum Chapter IV This study This study PC 2006 PC 2006 Hu lsey et al. Mexico,NCA “Cichlasoma” urophthalmum Chapter IV This study Hulsey et al. Mexico “Cichlasoma” salvini Chapter IV This study PC 2006 PC 2006 PC 2006 Mexico,NCA

228 Herichthys bartoni Chapter IV Hulsey et al. Mexico Herichthys carpintis PC 2006 PC 2006 PC 2006 PC 2006 Mexico Herichthys cyanoguttatus Chapter IV Hu lsey et al. Mexico, Texas

Herichthys labridens This study This study Mexico

Herichthys minckleyi Hu lsey et al. Mexico

Herichthys pantostictus Hulsey et al. Mexico

Herichthys steindachneri Chapter IV Hulsey et al. Mexico

Herichthys tamasopoensis This study This study Hul sey et al. Mexico

Hypsophrys nicaraguensis Chapter IV PC 2006 PC 2006 PC 2006 PC 2006 Hu lsey et al. Nuclear CA

Neetroplus nematopus Hu lsey et al. Nuclear CA

Parachromis dovii Chapter IV PC 2006 PC 2006 PC 2006 PC 20 06 Hulsey et al. Nuclear CA

Parachromis friedrichsthali Chapter IV This study This study Mexico,NCA

Parachromis loisellei Chapter IV This study This study Hulsey et al. CA

Parachromis managuense Chapter IV PC 2006 PC 2006 PC 2006 PC 2006 Hu lsey et al. Cos ta Rica

Parachromis motaguense Chapter IV PC 2006 PC 2006 PC 2006 PC 2006 Northern CA

Paraneetroplus bulleri Hulsey et al. Mexico

228 Table V.1 continued

Middle American Morphology 16S COI Tmo-4c4 S7 Cyt b Locality Heroines Petenia splendida Chapter IV This study This study PC 2006 PC 2006 Hulsey et al. Mexico, NCA Theraps wesseli This study This study Honduras Thorichthys affinis Chapter IV Northern CA Thorichthys aureus PC 2006 PC 2006 PC 2006 PC 2006 Northern CA Thorichthys callolepis Hulsey et al. Mexico Thorichthys ellioti This study This study Hulsey et al. Mexico Thorichthys helleri Hulsey et al. Mexico, NCA Thorichthys meeki Chapter IV Thi s study This study This study Hulsey et al. Mexico, NCA Thorichthys pasionis Chapter IV Thi s study This study Mexico, NCA Tomocichla asfraci Spark s&Smith Sp arks&Smith Panama Tomocichla sieboldi Chapter IV PC 2006 PC 2006 PC 2006 PC 2006 Hu lsey et al. Southern CA

229 Tomocichla tuba Chapter IV Hulsey et al. Southern CA Vieja argentea This study This study This study Mexico Vieja bifasciata This study This study This study Mexico, NCA Vieja breidohri Hulsey et al. Mexico Vieja fenestrata Chapter IV Hulsey et al. Mexico Vieja godmanni This study This study NCA Vieja guttulata Hulsey et al. Guatemala Vieja heterospilus This study This study This study Mexico, NCA Vieja intermedia This study This study Mexico, NCA Vieja regani This study This study Hulsey et al. Mexico Vieja synspila Chapter IV This study This study PC 2006 PC 2006 Hulsey et al. Mexico, NCA Vieja maculicauda Chapter IV This study This study Hu lsey et al. CA Vieja ‘Belize’ melanurus This study This study Belize Vieja tuyrense PC 2006 PC 2006 PC 2006 PC 2006 Panama

229 Table V.1 continued

Middle America Morphology 16S COI Tmo -4c4 S7 Cyt b Locali ty Heroines Vieja ufermanni This study This study Mexico, Guatemal a Vieja zonata Chapter IV Mexico Greater Antilles Heroines Nandopsis ramsdeni Chapter IV PC 2006 PC 2006 PC 2006 PC 20 06 Cuba Nandopsis tetracanthus Chapter IV PC 2006 PC 2006 PC 20 06 PC 2 006 Cuba Nandopsis haitiensis Chapter IV PC 2006 PC 2006 PC 2006 P C 2006 Hispaniola South American Heroines Caquetaia kraussii This study This study This study Hulsey et al. North SA Caquetaia myersi This study This study This study Hulse y et al. Amazon Basin Caquetaia spectabilis This study This study This study Hulse y et al. Amazon Basin “Cichlasoma” atromaculatum Chapter IV Hulsey et al. North SA, Panama “Cichlasoma” facetum Chapter IV This study Thi s study This study Central SA 230 “Cichlasoma” festae Chapter IV PC 2006 PC 2006 PC 2006 P C 2006 Hulsey et al. North SA “Cichlasoma” ornatum Chapter IV North SA Heros appendiculatus PC 2006 PC 2006 PC 2006 P C 2006 Hulsey et al. North SA Hypselecara coryphaenoides Hulsey et al. NorthCentral SA Hypselecara temporalis PC 2006 PC 2006 PC 2006 P C 2006 NorthCe n tral SA Mesonauta insignis Hulsey et al. NorthCentral SA Symphysodon aequifasciatus Hulsey et al. Amazon Basin Uaru amphiacanthoides PC 2006 PC 2006 PC 2006 P C 2006 Hulsey et al. Amazon Basin South American Outgroups Aequidens diadema This study This study Amazon Basin Apistogramma bitaeniata PC 2006 PC 2006 PC 20 06 PC 2 006 Amazo n Basin Bujurquina vittata PC 2006 PC 2006 PC 2006 P C 2006 Central SA Cichla ocellaris Chapter IV North SA

230 Table V.1 continued

South American Outgroups Morphology 16S COI Tm o-4c4 S7 C yt b Locali ty Cichla temensis This st udy This study Huls ey et al. Nort h SA Crenicichla acutirostris This study This study This study Amazon Basin Crenicichla saxatilis Chapter IV North SA Geophagus steindachneri PC 2006 PC 2006 PC 2006 PC 2006 North SA Gymnogeophagus gymnogenys This st udy This study This study Cent ral SA Satanoperca jurupari This study This study Amazon Basin Tahuantinsuyoa macantzatza This study This study This study Amazon Basin Teleocichla monogramma This study This study This study Amazon Basin

231 Madagascar – India

Outgroups Etroplus maculates PC 2006 PC 2006 PC 2006 PC 2006 India Paretroplus kieneri PC 2006 PC 20 06 PC 2006 PC 20 06 Madagascar Paratilapia polleni PC 2006 PC 2006 PC 2006 PC 2006 Madagascar African Outgroups Etia nguti This study This study West Africa Hemichromis letourneuxi This st udy This study North Africa

231 LITERATURE CITED

Agassiz, L. (1859) Remarks on new fishes from Lake Ni caragua. Proceedings of the Boston Society of Natural History 6, 407-408.

Burgess, W.E. and Walls, J.G. (1993) Cichlasoma: the next step. Tropical Fish Hobbyist January, 80-82.

Bussing, W.A. (1985) Patterns of distribution of the Cen tral American Ichthyofauna. In: The Great American Biotic Interchange Volume 4, Topics in Geobiology (eds F.G. Stehli and S.D. Webb), Plenum Press, New York, pp. 453–473.

Chakrabarty, P. (2006) S ystematics and Historical Bioge ography of Greater Antillean Cichlidae. Molecular Phylogenetics and Evolution 39, 619-627.

Chakrabarty, P. (2006b) Taxonomic status of the Hispan iolan Cic hlidae. Occasional Papers Mus eum o f Zoolog y, Unive rsity of Michig an 737, 1-16.

Farias, I.P., Orti, G., Sampaio, I., Schneider, H., Meyer, A. (2001) The cytochrome b gene as a phylogenetic marker: the limits of resolution for analyzing relationships among cichlid fis hes. Journal of Molecular Evol ution 53, 89-103.

Fernández-Yépez, A. (19 69) Contribucion al conocimien to de lo s cichlidos. Evencias 22, 1-10.

Gill, T. N. (1862) Remarks on the relations of the genera and other groups of Cuban fishes. Proceedings of the Academy of Natural Sciences Philadelphia 14, 235- 242.

Gill, T. N. and Bransford, J.F. (1877) Synopsis of the fishes of Lake Nicaragua. Proceedings of the Academy of Natural Sciences Philadelphia 29, 175-191.

Goloboff, P. A. (19 93) NONA 2.0 Program and documentat ion. C omputer program distributed b y J. M. Carpenter, Dept. of Entomology, American Museum of Natural History, New York.

Günther, A. (1864) Report of a collection of fishes made by Messrs. Dow, Godman, and Salvin in Guatemala. Proceedings of the Zoological Society of London 1864, 144-154.

232 Günther, A. (1867) On the fishes of the states of Central America, founded upon specimens collected in fresh and marine waters of various parts of that country by Messrs. Salvin and Godman and Capt. J. M. Dow. Proceedings of the Zoological Society of London 1866, 600-604.

Heckel, J. (1840) Johann Natterer’s neue Flussfishe Brasilien’s nach den Beobachtungen und Mittheilungen des Entdeckers beschreiben. Annalen des Wiener Museums der Naturgeschichte, 325-471.

Hubbs, C.L. (1936) Fishes of the Yucatan Peninsula. Carnegie Institute Washington Publication 457, 157-287.

Iturralde-Vinent, M.A. and MacPhee, R.D.E. (1999) Paleogeography of the Caribbean region: implications for Cenozoic biogeography. Bulletin of the American Museum of Natural History 238, 1–95.

Kullander, S.O. (1983) A revision of the South American cichlid genus Cichlasoma (Teleostei: Cichlidae). The Swedish Museum of Natural History, Stockholm.

Loiselle, P.V. (1984) On the correct usage of the nomen Cichlasoma Swainson 1839 and the status of the nomina Heros Heckel 1840, Parapetenia Regan 1905 and Caquetaia Fowler 1945. Buntbarsche Bulletin 105, 21-24.

233 Lydeard, C. and Roe, K.J. (1997) The phylogenetic utility of the mitochondrial cytochrome b gene for inferring relationships among Actinopterygian fishes. In: Molecular Systematics of Fishes (eds T. Kocher and C.A.Stepien) Academic Press, San Diego, pp. 285-303.

Maddison, D.R. and Maddison, W.P. (1992) MacClade: Analysis of Phylogeny and Character Evolution. Version 3.0. Sinauer Associates, Sunderland.

Meek, S.E. (1904) The fresh-water fishes of Mexico north of the Isthmus of Tehuantepec. Field Columbian Museum Publication (Zoological Series) 93, 1-252.

Miller, R.R. (1966) Geographical distribution of Central American freshwater fishes. Copeia 1966, 773-802.7

Miller, R.R. (1976) Geographical distribution of Central American freshwater fishes. In: Investigations of the Ichthyofauna of Nicaraguan Lakes (ed T.B. Thorson), University of Nebraska, Lincoln.

Miller, R.R. (2005) Freshwater Fishes of México. University of Chicago Press, Chicago.

Myers, G.S. (1938) Fresh-water fishes and West Indian Zoogeography. Annual Report of the Board of Regents of the Smithsonian Institution 92, 339-364.

Myers, G.S. (1966) Derivation of the freshwater fish fauna of Central America. Copeia 1966, 766–772.

Nixon, K.C. (1999) The parsimony ratchet, a new method for rapid parsimony analysis. Cladistics 15, 177-182.

234 Perdices, A., Doadrio, I. and Bermingham, E. (2005) Evolutionary history of the synbranchid eels (Teleostei: Synbranchidae) in Central America and the Caribbean islands inferred from their molecular phylogeny. Molecular Phylogenetics and Evolution 37, 460-473.

Pitman , W.C., III, Cande, S.C., LaBrecque, J. and Pindell, J.L. (1993) Fragmentation of Gondwana: the separation of Africa from South America. In: Biological Relationships Between Africa and South America (ed P.Goldblatt), Yale University Press, New Haven, pp 15–34.

Regan, C.T. (1905) A revision of the fishes of the American cichlid genus Cichlasoma [sic] and of the allied genera. Annual Magazine of Natural History Ser. 7.16, 60- 77, 225-243, 316-340, 433-445.

Regan, C.T. (1906-1908) Biologia Centrali-Americana 8: Pisces. Taylor and Francis, London.

Regan, C.T. (1908) A collection of freshwater fishes made by Mr. C. F. Underwood in Costa Rica. Annual Magazine of Natural History Ser. 8, 455-464.

Roe, K.J., Conkel, D., and Lydeard, C. (1997) Molecular Systematics of Middle American Cichlid Fishes and the Evolution of Trophic-Types in ‘Cichlasoma (Amphilophus)’ and ‘C. (Thorichthys)’. Molecular Phylogenetics and Evolution 7, 366-376.

Sikes, D.S. and Lewis, P.O. (2001) Software manual for PAUPRat: A tool to implement Parsimony Ratchet Searches using PAUP*. Available online www.ucalgary.ca

Sparks, J.S. and Smith, W.L. (2005) Freshwater fishes, dispersal ability, and nonevidence: “Gondwana life rafts” to the rescue. Systematic Biology 54, 158– 165.

Swofford D. L. (2002) PAUP* (phylogenetic analysis using parsimony * and other methods). Sinauer, Sunderland, Massachusetts, USA.

235