Macroevolution and Paleobiogeography of Middle to Late Ordovician Brachiopods: A

Home , Arnheim Formation, Liberty Formation

Phylogenetic Biogeographic Approach

A thesis presented to

the faculty of

the College of Arts and Sciences of Ohio University

In partial fulfillment

of the requirements for the degree

Master of Science

David F. Wright

June 2012

This thesis titled

Macroevolution and Paleobiogeography of Middle to Late Ordovician Brachiopods: A

Phylogenetic Biogeographic Approach

DAVID F. WRIGHT

has been approved for

the Department of Geological Sciences

and the College of Arts and Sciences by

Alycia L. Stigall

Associate Professor of Geological Sciences

Howard Dewald

Interim Dean, College of Arts and Sciences 3

ABSTRACT

WRIGHT, DAVID F., M. S., June 2012, Geological Sciences

Macroevolutionary and Paleobiogeography of Middle to Late Ordovician Brachiopods: A

Phylogenetic Biogeographic Approach

Director of Thesis: Alycia L. Stigall

The evolution and biogeography of three Ordovician orthid brachiopod genera were examined using species-level phylogenetic analyses and phylogenetic biogeography. Species-level evolutionary patterns and taxonomic revisions were documented for Laurentian species of Glyptorthis Foerste, 1914; Plaesiomys Hall and

Clarke, 1892, and Hebertella Hall and Clarke, 1892. Recovered evolutionary relationships were used to investigate geological drivers of macroevolutionary and paleobiogeographic patterns in Middle to Late Ordovician brachiopods and constrain the biogeographic origin of the Late Ordovician Richmondian Invasion. Results indicate a transition in the dominant mode of speciation from vicariance to dispersal occurs across the middle Mohawkian M4/M5 sequence boundary. This macroevolutionary shift corresponds to changes in paleoceanographic conditions brought about by diachronous tectonic events related to the Taconian Orogeny and major rise in global sea level.

Multiple biogeographic pathways directed the influx of invasive species during the

Richmondian Invasion following a relative rise in sea-level and a return to tropical conditions within the Cincinnati Basin.

Approved: ______

Alycia L. Stigall

Associate Professor of Geological Sciences

ACKNOWLEDGMENTS

This thesis would not have been possible without the assistance and support of many individuals. I would like to thank to following people for their gracious assistance to museum collections: Susan Butts (Yale Peabody Museum), Brenda Hunda (Cincinnati

Museum of Natural History), Mark Florence (United States Museum of Natural History),

Bushra Hussaini (American Museum of Natural History), Tiffany Adrain (University of

Iowa Paleontology Repository), and Rob Swisher (Sam Noble Museum of Natural

History). Sierra Isard provided assistance with specimen photography.

I would especially like to thank my advisor, Alycia Stigall for her invaluable assistance, support, discussions, and encouragement. This project has greatly benefited from her generous enthusiasm and inspiration. I thank my intrepid lab mates, Rich

Malizia and Hannah Brame, for their fortitude, camaraderie, and enduring sense of humor. In addition, I would like to thank my committee members Dan Hembree and

David Kidder for their helpful comments and discussion.

This research was supported by the Schuchert and Dunbar Grants-in-Aid Award from the Yale Peabody Museum, the Dry Dredgers Paul Sanders Award, the Ohio

University Geological Sciences Alumni Grant, and the OHIO Center for Ecology and

Evolutionary Studies Fellowship. Additional support for this research was received by

Alycia Stigall from the National Science Foundation.

TABLE OF CONTENTS

Page

Abstract ...... 3 Acknowledgments...... 5 List of Tables ...... 8 List of Figures ...... 10 Chapter 1: Introduction ...... 14 Chapter 2: Species-level Phylogenetic Revision of the Orthid Brachiopod Genus Glyptorthis from North America ...... 17 Abstract ...... 17 Introduction ...... 17 Materials and Methods ...... 19 Taxa Analyzed ...... 19 Material Examined ...... 21 Characters and Coding ...... 22 Search Method ...... 28 Results ...... 34 Phylogenetic Analysis ...... 34 Recognition of Clades ...... 34 Discussion ...... 39 Conclusions ...... 45 Systematic Paleontology ...... 46 Chapter 3: Phylogenetic Revision of the Late Ordovician Orthid Brachiopod Generea Plaesiomys and Hebertella from North America...... 62 Abstract ...... 62 Introduction ...... 63 Revision of Plaesiomys ...... 64 Taxonomic Background ...... 64 Materials and Methods ...... 66 Taxa analyzed ...... 66 7

Characters and Coding ...... 68 Search Method ...... 76 Results ...... 78 Phylogenetic Analysis ...... 78 Recognition of Clades ...... 78 Discussion ...... 81 Systematic Paleontology ...... 86 Revision of Hebertella ...... 97 Taxonomic Background ...... 97 Materials and Methods ...... 99 Taxa Analyzed ...... 99 Material Examined ...... 99 Characters and Coding ...... 100 Search Method ...... 103 Results ...... 107 Recognition of Clades ...... 107 Discussion ...... 110 Systematic Paleontology ...... 113 Conclusions ...... 126 Chapter 4: Geological Drivers of Late Ordovician Faunall Change in Laurentia: Investigating Links between Tectonics, Speciation, and Invasions ...... 127 Abstract ...... 127 Introduction ...... 128 Synchronized Geologic and Biotic Change ...... 128 Biogeographic Analysis ...... 132 Macroevolutionary Impact of Biogeographic Change ...... 137 Conclusions ...... 142 Chapter 5: Conclusions ...... 144 References ...... 147

LIST OF TABLES

Page Table 1 Species Included in Phylogenetic Analysis. * Indicates Outgroup Species ...... 21

Table 2 Characters Used in Phylogenetic Analysis ...... 23

Table 3 Range of Morphometric Measurements for Species Analyzed ...... 29

Table 4 Statistical Separation of Continuous Characters. ANOVA Results: p ˂ 0.001 for all Characters. S. D. Indicates Standard Deviation ...... 30

Table 5 Character State Distribution for Species Included in Phylogenetic Analysis X = (0 and 1), Y = (1 and 2), and Z = (0 and 2) ...... 33

Table 6 Species occurrence data. x indicates Whiterockian (lower Sandbian), * Indicates Mohawkian (upper Sandbian to middle Katian), † Indicates Cincinnatian (upper Katian), ‡ Indicates Silurian (Aeronian) ...... 42

Table 7 Species of Glyptorthis Reported from Outside North America ...... 45

Table 8 Specimens Examined for Species That Do Not Require Taxonomic Revision or Lectotype Designation * Indicates Non-Type Specimens ...... 59

Table 9 Species Included in Phylogenetic Analysis of Plaesiomys * Indicates Outgroup 67

Table 10 Characters used in Phylogenetic Analysis of Plaesiomys ...... 68

Table 11 Data Ranges for Continuous Characters of Species Analyzed in Phylogenetic Analysis of Plaesiomys ...... 73

Table 12 Statistical Separation of Continuous Characters in Plaesiomys Analysis. S.D. Indicates Standard Deviation ...... 74

Table 13 Character State Distribution for Species Included in Phylogenetic Analysis of Plaesiomys X = (0 and 1) ...... 78

Table 14 Stratigraphic and Geographic Distribution of Plaesiomys † Indicates Gamachian (Hirnantian), All Other Species Are Richmondian (Katian) ...... 85

Table 15 Specimens Examined for Species that Do Not Require Taxonomic Revision or Lectotype Designation * Indicates Non-Type Specimens ...... 89

Table 16 Species Included In Phylogenetic Analysis of Hebertella * Indicates Outgroup ...... 100 Table 17 Characters Used in Phylogenetic Analyses of Plaesiomys ...... 101

Table 18 Statistical Separation of Continuous Characters in Hebertella Analysis. S.D. Indicates Standard Deviation ...... 104

Table 19 Data ranges of morphometric characters for species included in phylogenetic analysis of Hebertella ...... 106

Table 20 Character State Distribution for Species Included in Phylogenetic Analysis of Hebertella X = (0 and 1) ...... 107

Table 21 Stratigraphic and Geographic Distribution of Hebertella x Indicates Whiterockian (lower Sandbian), * Indicates Mohawkian (upper Sandbian to middle Katian), † Indicates Cincinnatian (upper Katian to Hirnantian) ...... 112

Table 22 Specimens Examined for Species that Do Not Require Taxonomic Revision or Lectotype Designation. * Indicates Non-Type Specimens ...... 115

Table 23 Vicariance and Geodispersal Data Matrices Generated for LBPA Analysis of Time Slice 1 and Time Slice 2. Numbered Columns Refer to Nodes and Tips on the Stratocladogram (Figure 16). Coding as in Wiley and Lieberman (2011) ...... 143

LIST OF FIGURES

Page

Figure 1 Examples of morphometric-based measurements utilized in character analysis. 1, Glyptorthis alta Cooper, 1956, USNM 110008 (holotype), lateral view, × 4.4; 2, G. crenulata Cooper, 1956, USNM 110054 (paratype), ventral valve interior, × 4.6; 3, G. costellata, USNM 11052a (holotype), ventral view, × 3.9; 4, G. assimilis Cooper, 1956, USNM 110013j (holotype), posterior view, × 4.3. Abbreviations: amw, adductor muscle width (maximum width of the ventral adductor muscle scars); da, delthyrial angle (angle between the lateral boundaries of the ventral interarea); dvh, dorsal valve height; ia, interarea angle; msw, muscle scar width (maximum width of the ventral muscle field); th, thickness (total shell height); w, maximum width...... 27

Figure 2 Single most parsimonious tree obtained from the character state distributions given in Table 4 as implemented in PAUP (Swofford, 2002). Nodes are numbered and circled. Values from Jackknife analysis are indicated next to supported nodes. Characters were optimized using ACCTRAN in MacClade (Maddison and Maddison, 2000). Unambiguous character changes listed below are denoted by parentheses: Node 1, 3 (1), 7, (1), 8 (1), 10 (2), 14 (1); Node 2, 2 (1), 4 (1), 33 (1); Node 3, 9 (1), 11 (1), 13 (1), 15 (1); Node 4, 10 (1), 18 (2), 20 (1); Node 5, 2 (0), 7 (0), 24 (0), 32 (1); Node 6, 12 (1), 28 (1), 30 (1); Node 7, 24 (0), 25 (1); Node 8, 5 (0), 8 (0), 17 (1), 34 (1); Node 9, 11 (2), 29 (1); Node 10, 11 (0), 16 (1), 26 (1), 27 (1), 32 (1), 34 (1); Node 11, 7 (0), 17 (1), 22 (0), 23 (0), 31 (1); Node 12, 1 (1), 13 (1); Node 13, 4 (0), 12 (1), 24 (0), 29 (1); Node 14, 11 (2), 14 (0), 18 (1); Node 15, 9 (1), Node 16, 8 (0); Node 17, 25 (1), 28 (1), 30 (1); Node 18, 2 (0), 6 (1); Node 19, 19 (1), 20 (1), 21 (1), 22 (1), 33 (0); Node 20, 4 (0), 5 (0), 10 (1), 11 (0); Node 21, 3 (0), 12 (1); Node 22, 13 (1), 14 (1), 15 (1), 28 (1), 30 (1)...... 35

Figure 3 Late Ordovician paleogeography of Laurentia indicating the distribution of Glyptorthid brachiopods. Dashed lines indicate shallow marine environments. 1, New York; 2, western Maryland and Virginia; 3, Ohio-Indiana-Kentucky region; 4, Tennessee; 5, Alabama; 6, Trans-Pecos Texas; 7, Arbuckle Mountains, Oklahoma; 8, eastern Missouri and Iowa; 9, Minnesota; 10, Advance Mountains, British Columbia...... 41

Figure 4 1, Glyptorthis daytonensis (Foerste, 1885), USNM 84828 (lectotype), dorsal view, ×1.7; 2, USNM 546107, dorsal valve interior ×1.6; 3, G. fausta (Foerste, 1885), USNM 84829 (lectotype), dorsal view, ×2.7; 4, ventral view, ×2.6; 5, USNM 84813 (paralectotype), ventral view, ×2; 6, dorsal view, × 1.9; 7, USNM 546109, ventral valve interior, ×1.9; 8, G. uniformis Cooper, 1956, USNM 116945a (holotype), ventral view, ×3; 9, dorsal view, ×2.5...... 50

Figure 5 1, Glyptorthis insculpta (Hall, 1847), AMNH 30260 (holotype), ventral view, ×2.6; 2, ventral interior, ×2.5; 3, YPM S-242b, ventral interior, ×2.4; 4, YPM S-242d, 11 dorsal interior, ×2.3; 5, YPM S-242a, ventral view, ×2.3; 6, G. maquoketensis Ladd, 1929, SUI 66500 (lectotype), ventral view, ×2.3; 7, dorsal view, ×2.3; 8, SUI 66502 (paralectotype), ventral interior, ×2.5...... 55

Figure 6 Examples of morphometric-based measurements utilized in character analysis. 1, Plaesiomys bellistriatus Wang, 1949, SUI 1820 (holotype), ventral view × 1.2; 2, lateral view, × 1.5; 3, P. subquadratus (Hall, 1847), YPM S-728f, ventral interior, × 1.4; 4, P. subcircularis (Roy, 1941), AMNH 4797 (holotype), posterior view, × 2.1 .Abbreviations: da, delthyrial angle (angle between the lateral boundaries of the ventral interarea); dvh, dorsal valve height; ia, interarea angle; msl, muscle scar length (maximum length of the ventral muscle field); w, maximum shell width...... 77

Figure 7 Single most parsimonious tree obtained from the character state distributions given in Table 5 as implemented in PAUP (Swofford, 2002). Nodes are numbered and circled. Values from Jackknife analysis are indicated next to supported nodes. Characters were optimized using ACCTRAN in MacClade (Maddison and Maddison, 2000). Unambiguous character changes listed below are denoted by parentheses: Node 1, 9 (1), 11, (1), 16 (1); Node 2, 2, (1), 10, (1), 18, (1); Node 3, 1, (1), 2, (0), 4, (1), 17, (1); Node 4, 8, (1), 19, (1); Node 5, 13, (0); Node 6, 5, (1), 6, (1), 7, (1); Node 7, 3, (1), 9, (0), 10, (1), 12, (1), 14, (1), 18, (1), 20, (1); Node 8, 1, (0), 15, (1), 17, (0); Node 9, 21, (1)...... 80

Figure 8 Late Ordovician paleogeography of Laurentia indicating basins where Hebertella and Plaesiomys species occurred. Dashed lines indicate shallow marine environments. 1, Anticosti Island; 2, Hudson Bay lowlands, Quebec; 3, Ohio-Indiana- Kentucky region; 4, Tennessee; 5, Arbuckle Mountains, Oklahoma; 6, Trans-Pecos, Texas; 7, Iowa-Illinois-Minnesota region; 8, Idaho, 9, Baffin Island, 10, Advance Mountains, British Columbia...... 84

Figure 9 1, Plaesiomys subquadratus (Hall, 1847), AMNH 30263 (lectotype), ventral view, × 1.5; 2, dorsal view, × 1.5; 3, YPM S-728d, ventral interior, × 1.4; 4, YPM S- 238f, dorsal interior, × 1.4; 5, P. proavitus (Winchell and Schuchert, 1892), YPM S-703a (lectotype), dorsal view, × 1.8; 6, SUI 1808, ventral view, × 2.3...... 95

Figure 10 Single most parsimonious tree obtained from the character state distributions given in Table 12 as implemented in PAUP (Swofford, 2002). Nodes are numbered and circled. Values from Jackknife analysis are indicated next to supported nodes. Characters were optimized using ACCTRAN in MacClade (Maddison and Maddison, 2000). Unambiguous character changes listed below are denoted by parentheses: Node 1, 2 (1), 7, (1), 18 (1); Node 2, 16 (1); Node 3, 5 (1), 8 (1), 17 (1); Node 4, 12 (1), 13 (1); Node 5, 3 (1), 4 (1), 6 (1); Node 6, 1 (0), 12 (1); Node 7, 14 (1), 15 (0); Node 8, 9 (1), 10 (1); Node 9, 7 (0), 11 (1)...... 108

Figure 11 1, Hebertella occidentalis (Hall, 1847), AMNH 30299 (lectotype), ventral view, × 1.8; 2, dorsal view, × 1.8; 3, anterior view, × 2.4; 4, H. prestonensis Ladd, 1929, 12

SUI 66503 (lectotype), dorsal view, × 1.2; 5, anterior view, × 1.3; 6, H. subjugata (Hall, 1847), AMNH 30290 (holotype), dorsal view, × 1.5...... 120

Figure 12 Paleogeographic map of Laurentia during the Late Ordovician. Biogeographic areas considered in this analysis indicated as: (1), Anticosti Island; (2), Appalachian Basin; (3), Cincinnati Basin; (4), Central Basin; (5), Southern mid-Continent; and (6), Northern mid-Continent. Paleogeographic reconstructions (e.g. Witzke, 1980) support the separation of these areas by physical barriers, such as structural platforms, trenches, and intracratonic arches...... 129

Figure 13 Taxon-area stratocladogram for the brachiopod genus Hebertella Hall and Clarke, 1892. Evolutionary and biogeographic patterns are mapped onto a stratigraphic column at the level of depositional sequence. Ghost lineages reconstructed via equal phylogenetic split. Shifts in reconstructed geographic distributions at cladogenetic events indicate speciation by vicariance or dispersal. Vicariance is inferred where descendent species occupies a subset of the geographic range inhabited by its ancestor; whereas dispersal in inferred where a descendent species occupies additional geographic regions not inhabited by its ancestor (Wiley and Lieberman, 2011). Timescale modified from Holland and Patzkowsky (1996)...... 131

Figure 14 Cladograms depicting evolutionary and biogeographic relationships for species in Glyptorthis Foerste, 1914a, Plaesiomys Hall and Clarke, 1892, and Hebertella Hall and Clarke 1892. Phylogenetic topology from Chapters 2 and 3. Geographic distributions for terminal taxa were compiled from literature sources and museum collections. Only occurrence data for which the species could be visually verified (e.g., photo plate, museum specimen examined) were incorporated. Biogeographic states for ancestral nodes were optimized using Fitch Parsimony as described in Lieberman (2000)...... 134

Figure 15 Species-level evolutionary and biogeographic patterns for Glyptorthis Foerste, 1914a, Plaesiomys Hall and Clarke, 1892, and Hebertella Hall and Clarke 1892 mapped onto a stratigraphic column at the level of depositional sequence. Ghost lineages reconstructed via equal phylogenetic split. Timescale modified from Holland and Patzkowsky (1996)...... 135

Figure 16 Stratocladograms from Figure 15 with nodes and tips labeled to correspond with coding used in LBPA. The division between T1 and T2 is indicated by the horizontal line...... 136

Figure 17 Vicariance and Geodispersal general area cladograms for T1 and T2. The vicariance trees indicate the order that areas became separated from each other due to barrier (tectonic, eustatic, etc.) formation, whereas geodispersal trees indicate the relative order in which barriers were removed. Congruence between tree topologies indicates that geographic barriers rose and fell in the same relative order, suggesting the importance of cyclical (e.g., oscillating sea levels) geologic processes. In contrast, incongruence 13 indicates that non-cyclic, singular events (e.g. tectonic pulses) influenced the evolution and biogeographic differentiation of the clade (Lieberman, 2000). Numerical values are bootstrap (plain text) and jackknife (italicized) values that indicate support for nodes. T1 analysis recovered a single most parsimonious vicariance tree (length of 87 steps, consistency index 0.736) and two most parsimonious geodispersal trees (64 steps, CI 0.797). T2 analysis recovered a single most parsimonious vicariance tree (49 steps, CI 0.816) and two most parsimonious geodispersal trees (59 steps, CI 0.531). Geodispersal trees illustrate topology of the strict consensus tree...... 138

CHAPTER 1: INTRODUCTION

This thesis is a compilation four separate analyses investigating the evolution and biography Middle to Late Ordovician orthid brachiopods from Laurentia. The fossil record of eastern Laurentia records a dramatic biotic turnover and biogeographic differentiation during the middle Mohawkian M4/M5 sequence boundary from a warm water fauna to a cold water fauna (Holland and Patzkowsky; 1996; Patzkowsky and

Holland, 1993). In addition, a large scale regional biotic immigration into the Cincinnati

Basin event occurred in the Late Ordovician known as the Richmondian Invasion. The goal of this thesis is to assess the role of geological processes driving paleobiogeographic patterns across these intervals of biotic turnover and investigate their macroevolutionary consequences within a species-level evolutionary framework using orthid brachiopods.

The eastern margin of Laurentia was geologically active during the Middle to

Late Ordovician Taconian Orogeny as the eastern margin of Laurentia was colliding with several island arcs and/or microplates (Shanmugam and Lash, 1982). These events were diachronous, and are separated into the Blountian and Taconic tectophases (Ettensohn,

1994). The earlier, Blountian tectophase introduced siliciclastics into the shallow, epieric seas of eastern Laurentia but did not prohibit deposition of tropical carbonate deposition

(Holland and Patzkowsky, 1996). In contrast, the onset of the Taconic tectophase corresponds to a dramatic shift in lithology styles across eastern Laurentia from a tropical to temperate style carbonate systems in eastern Laurentia, with tropical carbonate environments shifting to the western continental interior and marginal basins. The

Taconic tectophase occurred concomitant with a major rise in global sea level at the 15

M4/M5 sequence boundary of Holland and Patzkowsky (1996). The combination of major flexural downwarping and a eustatic sea level rise facilitated the introduction cool, nutrient rich water from the Sebree Trough, thereby producing a shift lithologic shift from tropical to temperate style carbonate lithologies lasting from the middle Mohawkian to the Richmondian (Holland and Patzkowsky, 1996; Pope and Read, 1997). By the

Richmondian, the source of the cool water was eliminated and tropical carbonate deposition resumed.

Major biotic turnover, biogeographic restriction, and interbasinal invasion occur throughout the Middle to Late Ordovician. For example, regional extinction and biogeographic differentiation have been documented in graptolites (Goldman and Wu,

2010), corals, (Elias, 1991), and trilobites (Shaw, 1991) across the M5 sequence marking a transition from warm water to cold water adapted faunas (Patzkowsky and Holland,

1996). In addition, the Late Ordovician deposits of the Cincinnati basin record a large scale regional biotic immigration event—the Richmondian Invasion. The ultimate geographic source of the invaders, however, is unclear. Some studies have emphasized a unidirectional pathway of invasion into the Cincinnati basin (Foerste, 1912; Holland,

1997; Nelson, 1959). However, many invader taxa were geographically widespread at the genus-level, and species-level biogeographic patterns have never previously been assessed within an evolutionary framework.

Orthid brachiopods are an ideal taxonomic group to address macroevolutionary and paleobiogeographic patterns during the Middle to Late Ordovician because they rank among the most common, well preserved fossils and are relatively easy to identify at the 16 species-level. To accurately document how geologic processes affect paleobiogeographic and macroevolutionary phenomena in the context of speciation events, a species-level evolutionary framework is needed. However, no previous studies have investigated species-level evolutionary relationships for Middle to Late Ordovician orthid brachiopod taxa. Chapters 2 and 3 of this thesis use phylogenetic analysis to document evolutionary relationships and inform taxonomic revisions Laurentian species in three genera of

Middle to Late Ordovician brachiopods: Glyptorthis Foerste, 1914a; Plaesiomys Hall and

Clark, 1892; and Hebertella Hall and Clarke, 1892.

Phylogenetic biogeographic analytical methods are used in Chapter 4 in conjunction with the evolutionary framework developed in Chapters 2 and 3 to examine the role of geologic processes driving macroevolutionary and paleobiogeographic patterns throughout the Middle and Late Ordovician of Laurentia and to ascertain the biogeographic origin of Late Ordovician Richmondian invaders. 17

CHAPTER 2: SPECIES-LEVEL PHYLOGENETIC REVISION OF THE

ORTHID BRACHIOPOD GENUS GLYPTORTHIS FROM NORTH AMERICA

ABSTRACT.—Species of the orthid brachiopod genus Glyptorthis occur as common constituents of Late Ordovician benthic marine faunas from the Northern

Hemisphere. A phylogenetic analysis of North American species of Glyptorthis was conducted to inform a systematic revision and ascertain evolutionary relationships among species. The development of a species-level phylogenetic hypothesis for Glyptorthis provides a framework for assessing evolutionary and paleobiogeographic patterns of Late

Ordovician brachiopods from the Laurentian paleocontinent. Glyptorthis insculpta participated in the Late Ordovician Richmondian Invasion into the Cincinnati Basin. The recovered phylogenetic topology suggests that G. insculpta may have invaded the

Cincinnati basin from low latitude subtropical region, such as the American mid-

Continent. Further investigations are needed to assess whether or not the pathway of invasion for G. insculpta represents a general pattern for invader taxa. Glyptorthis virginica is herein synonymized with G. uniformis. Glyptorthis maquoketensis is recognized as a distinct species of Glyptorthis rather than a subspecies of G. insculpta.

INTRODUCTION

Glyptorthis Foerste, 1914a is species-rich genus of orthid brachiopods ranging from the Middle Ordovician (Darrwilian) to Silurian (Wenlock: Homerian). Species of

Glyptorthis reached maximal diversity during the late Sandbian to early Katian, and they 18 occur as common constituents of Late Ordovician benthic marine faunas from the

Northern Hemisphere. Previous taxonomic studies of Glyptorthis have focused primarily on identifying new species revising specific definitions (e.g., Cooper, 1956; Schuchert and Cooper, 1932). No prior investigations have been undertaken to provide a comprehensive taxonomic revision or phylogenetic analysis of all Upper Ordovician to

Lower Silurian species of Glyptorthis from North America.

Glyptorthis is distinguished from other genera of orthid brachiopods by the presence of strongly lamellose, concentric ornamentation on the exterior of the shell and subchordate ventral muscle scars (Schuchert and Cooper, 1932; Williams and Harper,

2000). Orthis insculpta, the type species of Glyptorthis, was originally described by Hall

(1847) from a single ventral valve characterized by the “beautiful” (Hall, 1847 p. 125) sculpted appearance on the exterior of the shell. Hall and Clark (1882) transferred Orthis insculpta to Hebertella, apparently emphasizing overall similarity in shell shape and concentric striae while considering the exterior ornamentation of O. insculpta to be a specific rather than generic trait. Foerste (1914), however, noted that concentric lamellose growth lines of Orthis insculpta clearly differentiated it from the type species of

Hebertella, H. sinuata (Hall, 1847) and erected the genus Glyptorthis for this species.

Schuchert and Cooper (1932) furthered the distinction between Glyptorthis and

Hebertella, citing considerable differences regarding the shape and placement of the ventral muscle scars and configuration of the dorsal cardinalia. Cooper’s (1956) monograph on Chazyan (Whiterockian to mid-Mohawkian Series) brachiopods represents 19 the largest taxonomic contribution to Glyptorthis thus far, describing more than 21 new species from North America.

The goal of this study is to develop a comprehensive phylogenetically informed taxonomic revision for the Middle Ordovician to Lower Silurian species of Glyptorthis of

North America. The phylogenetic revision presented herein aims to assess species validity, taxonomy, and document evolutionary relationships among species.

Development of a phylogenetic hypothesis for the evolution of Glyptorthis provides valuable data for species-level analyses of evolutionary and paleobiogeographic patterns of Late Ordovician brachiopods in tectonic basins throughout the Laurentian paleocontinent. For example, Glyptorthis insculpta participated in the Late Ordovician

Richmondian Invasion, an event which introduced over 50 genera to the present-day

Cincinnati region (Foerste, 1912; Meyer and Davis, 2009). The phylogenetic placement of G. insculpta may be used to test competing paleobiogeographic hypotheses regarding the geographic source of the Richmondian invaders (see Stigall, 2010).

MATERIALS AND METHODS

Taxa analyzed.—Specimens representing 23 species of Glyptorthis were included in the phylogenetic analysis (Table 1). All known Middle Ordovician to Lower Silurian species available were examined for morphological character data. Most species included were described by Cooper (1956) and occur in Upper Ordovician Mohawkian Series strata from the eastern and central United States. Species of Glyptorthis occurring outside

North America (e.g., those described in Rõõmusoks, 1964; Zhan and Jin, 2005; 20

Zuykov and Butts, 2008) were excluded because this investigation is limited to assessing evolutionary patterns in Laurentian species. Hesperorthis australis, Cooper, 1956, and H. biconvexa Cooper, 1956 were selected as outgroup taxa for character polarization. These species were chosen for outgroup comparison because the type specimens are well preserved and representative of Hesperorthis, a potential sister group to Glyptorthis

(Cooper, 1956).

Only species with sufficient morphological data to code for phylogenetic analysis were included in the final analysis. Species lacking morphological data for more than ten characters were excluded because their inclusion may lead to artifactual topologies and decreased phylogenetic accuracy (Wiens, 2003; Wiley and Lieberman, 2011).

Additionally, only data from specimens interpreted to be adult individuals were incorporated into the analysis. Small specimens interpreted as young individuals were excluded to remove ontogenetic variation from impacting character state distributions.

Prior to analysis, Glyptorthis uniformis Cooper, 1956 and G. virginica Cooper, 1956 were synonymized due to significant overlap in character state distribution, stratigraphic position, and paleobiogeographic distribution.

Table 1 Species Included in Phylogenetic Analysis. * Indicates Outgroup Species

Hesperorthis australis* Cooper, 1956 Hesperorthis biconvexa* Cooper, 1956 Glyptorthis alta Cooper, 1956 Glyptorthis assimilis Cooper, 1956 Glyptorthis bellarugosa (Conrad, 1843) Glyptorthis bellatula Cooper, 1956 Glyptorthis concinnula Cooper, 1956 Glyptorthis costellata Cooper, 1956 Glyptorthis crenulata Cooper, 1956 Glyptorthis daytonensis (Foerste, 1885) Glyptorthis equiconvexa Cooper, 1956 Glyptorthis fausta (Foerste, 1885) Glyptorthis glaseri Alberstabt, 1973 Glyptorthis glypta Cooper, 1956 Glyptorthis insculpta (Hall, 1847) Glyptorthis irregularis Cooper, 1956 Glyptorthis maquoketensis Ladd, 1929 Glyptorthis multicostellata Cooper, 1956 Glyptorthis obesa Cooper, 1956 Glyptorthis pulchra Wang, 1949 Glyptorthis senecta Cooper, 1956 Glyptorthis subcircularis Cooper, 1956 Glyptorthis sulcata Cooper, 1956 Glyptorthis uncinata Cooper, 1956

Glyptorthis uniformis Cooper, 1956

Specimens were examined from the following museum collections: United States

National Museum of Natural History (USNM); American Museum of Natural History

(AMNH); University of Iowa Paleontology Repository (SUI), Yale Peabody Museum

(YPM); and the Sam Noble Oklahoma Museum of Natural History (OMNH).

Characters and coding.— A total of 34 characters comprising internal and external morphological traits were included (Table 2). Both internal and external morphological characters were utilized because they have previously been demonstrated to be useful for constructing phylogenetic hypotheses of fossil brachiopods (Leighton and

Maples, 2002; Stigall Rode, 2005), and form the basis of many specific diagnoses

(Cooper, 1956; Williams and Harper, 2002).

Both discrete and continuous characters were considered. Discrete characters are the most common and familiar in phylogenetic analyses (Wiley and Lieberman, 2011). A discrete character represents a qualitative condition of a particular morphologic trait, such as whether or not the ventral adductor muscle scars are located on a raised platform

(character 25). Because species of Glyptorthis display limited discrete character variation, this analysis includes 23 (68%) continuous characters based on morphometric measurements. Continuous characters are quantitative and convey a relative mathematical property such as the maximum measured width of the ventral muscle scar divided by the maximum measured width of the shell (character 14). When normalized to form a ratio, a continuous morphometric character forms a quasi-continuous distribution. The range of 23 normalized continuous variation may then be differentiated into distinct classes and coded as discrete phylogenetic character states. Hopkins (2011) recently demonstrated that the inclusion of continuous characters significantly increases resolution in phylogenetic analyses. Additionally, continuous characters have been utilized previously to construct species-level phylogenetic hypotheses for clades of fossil invertebrates

(Rode, 2004; Stigall Rode, 2005; Hopkins, 2011). Examples of morphometric-based characters used in this analysis are provided in Figure 1.

Table 2 Characters Used in Phylogenetic Analysis General characters

1. Maximum width. The maximum shell width parallel to the hinge line: (0), moderate (≤ 10.917 ≤ x ≤ 12. 443); (1), large (≥ 19.100); (2), small (≤ 10.510).

2. Inflation of the dorsal valve. The ratio of the maximum dorsal height, measured perpendicular to the commissural plane, divided by the maximum width, measured parallel to the hinge line: (0), low (≤ .232); (1), high (≥ .242).

3. Inflation of the ventral valve. The ratio of the maximum ventral valve height, measured perpendicular to the commissural plane, divided by the maximum width, measured parallel to the hinge line: (0), low (≤ .254); (1), high (≥ .275).

4. Shell thickness. The total height of the individual divided by the maximum width, measured parallel to the hinge line: (0), thin (≤ .454); (1), thick (≥ .503).

5. Comparative convexity of the valves. The ratio of the ventral valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane and the dorsal valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane: (0), dorsal valve convexity greater (≤ .949); (1), ventral valve convexity greater (≥ .995).

6. Depth of the shell. The ratio of the maximum shell length, measured perpendicular to the hinge line, divided by the maximum shell width, measured parallel to the hinge line: (0), high (≥ .857); (1), low (≤ .866).

Table 2 continued

7. Origin of the sulcus. The ratio of the distance, measured perpendicular to the hinge line, from the ventral umbo to the beginning of a perceptible concave deflection in the ventral valve divided by the maximum length of the valve, measured perpendicular to the hinge line: (0), proximal (≤ .430); (1), distal (≥ .504).

8. Outline shape of the valves: (0), subelliptical; (1) subquadrate.

9. Interspatial width between costae and costellae: (0), narrow; (1), wide.

10. Development of concentric lamellae ornamentation: (0), none; (1), weakly lamellose; (2), strongly lamellose. 11. Costal density. The number of costae per 5 mm at the anterior margin of the shell, divided by the maximum shell width measured parallel to the hinge line: (0), low (≤ .366); (1), moderate (.413 ≤ x ≤ . 527); (2), high (≥ .596).

Ventral valve

12. Outline of the ventral diductor muscle scars: (0), anterior deflection only; (1), anterior and lateral deflection.

13. Length of the ventral diductor muscle scars. The ratio of the maximum length of the ventral muscle field, measured perpendicular to the hinge line, divided by the maximum shell length, measured perpendicular to the hinge line: (0), short (≤ .367); (1), long (≥ .439).

14. Maximum width of the ventral diductor muscle scars. The ratio of the maximum width of the ventral muscle field, measured parallel to the hinge line, divided by the maximum valve width, measured parallel to the hinge line: (0), narrow (≤ .328); (1), wide (≥ .359).

15. Width of the ventral muscle scars directly anterior of the hinge teeth. The ratio of the width of the ventral muscle field, measured parallel to the hinge line immediately anterior of the hinge teeth, divided by the maximum valve width, measured parallel to the hinge line: (0), narrow (≤ .317); (1), wide (≥ .300).

16. Longitudinal variation of ventral muscle scar width. The ratio of the minimum width of the ventral muscle scars measured parallel to the hinge line, divided by the maximum width of the ventral muscle scars measured parallel to the hinge line: (0), high variability (≤ .625); (1) low variability (≥ .728).

Table 2 continued

17. Length of hinge teeth. The ratio of the maximum length of the hinge teeth, divided by the maximum valve width measured parallel to the hinge line: (0), long (≥ .080); (1), short (≤.069).

18. Height of the ventral cardinal area. The ratio of the vertical distance from the hinge line to the top of the ventral interarea, measured perpendicular to the hinge line, divided by the ventral valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane: (0), low (≤ .465); (1), moderate (.576 ≤ x ≤ .706); (2), high (≥ .851).

19. Angle of the ventral interarea. The angle between the lateral boundaries of the ventral interarea: (0), low (≤ 130o); (1), high (≥ 135o).

20. Ventral umbonal angle. The angle between the two limbs of the umbo when the dorsal valve is oriented parallel to the commissural plane: (0), narrow (≤ 140o); (1), wide (≥ 145o).

21. Angle of the delthyrium. The angle between the lateral edges of the delthyrial opening: (0), low (≤ 45o); (1), high (≥ 55o).

22. Posterior extension of the ventral umbo across the hinge line. The ratio of the distance of extension of the ventral umbo, measured perpendicular to the hinge line, divided by the length of the hinge line: (0), long (≥ .166); (1), short (≤ .142).

23. Angle between the ventral interarea and the hinge line. The angle at which the ventral interarea diverges from the commissural plane prior to incurving: (0), upright (≥ 65o); (1), gently inclined (≤ 60o).

24. Presence of a medial septum separating the diductor muscle tracks anterior of adductor muscle scars: (0), present; (1), absent.

25. Vertical placement of adductor muscles: (0), subparallel to valve floor; (1) on a raised platform

26. Width of the adductor muscle scars. The maximum width of the adductor muscle scars, measured parallel to the hinge line, divided by the maximum shell width, measured parallel to the hinge line: (0), narrow (≤ .097); wide (≥ .109).

27. Comparative width of adductor and diductor muscle scars. The maximum width of the adductor muscle scars, measured parallel to the hinge line, divided by the total maximum width of the ventral muscle scars, measured parallel to the hinge line: (0), narrow (≤ .097); (1), wide (≥ .310). 26

Table 2 continued

Dorsal valve

28. Maximum width of the dorsal muscle scars. The ratio of the maximum width of the dorsal muscle field, measured parallel to the hinge line, divided by the maximum valve width, measured parallel to the hinge line: (0), narrow (≤ .427); (1), wide (≥ .435).

29. Length of the dorsal muscle scars. The ratio of the maximum length of the dorsal muscle field, measured perpendicular to the hinge line, divided by the maximum shell length, measured perpendicular to the hinge line: (0), short (≤ .333); (1), long (≥ .365).

30. Width of the dorsal muscle scars directly anterior of the brachiophores. The ratio of the width of the dorsal muscle field, measured parallel to the hinge line immediately anterior of the brachiophores, divided by the maximum valve width, measured parallel to the hinge line: (0), narrow (≤ .330); (1), wide (≥ .357).

31. Height of the dorsal cardinal area. The ratio of the vertical distance from the hinge line to the top of the dorsal interarea, measured perpendicular to the hinge line, divided by the ventral valve height, measured as the maximum vertical height of the dorsal valve perpendicular to the commissural plane: (0), high (≥ .208); (1) low (≤ .187).

32. Length of brachiophores. The ratio of the maximum length of the brachiophores, divided by the maximum valve width measured parallel to the hinge line: (0), long (≥ .096); (1), short (≤ .081).

33. Dorsal umbonal angle. The angle between the two limbs of the umbo when the dorsal valve is oriented parallel to the commissural plane: (0), wide (≥ 160o); (1), narrow (≤ 150o).

34. Posterior extension of the dorsal umbo across the hinge line. The ratio of the distance of extension of the dorsal umbo, measured perpendicular to the hinge line, divided by the length of the hinge line: (0), long (≥ .082); (1), short (≤ .072). ______

Continuous characters were coded based on the total range of morphometric values measured for each species presented in Table 3. Measurements of continuous characters were separated into discrete states using a procedure similar to those suggested by Morton and Kincaid (1995) and Swiderski et al. (1998). Morphometric measurements were visually inspected by examining univariate frequency distributions and box plots.

Cutoffs were determined by the presence of discontinuities, disjunct clusters, or changes in slope within the distribution. Following visual separation of the data set into classes, 28 potential character states were compared using analysis of variance (ANOVA). Due to potential heteroscedasticity between character states, the non-parametric Kruskal-Wallace test was also performed. In all cases, the results of the Kruskal-Wallace test confirmed the results from ANOVA. Classes of morphometric characters were found to be highly distinct (P < 0.001) with non-overlapping mean values within the 95% confidence interval. The expected range for each continuous character state was calculated as the mean ± 1 standard deviation (Table 4). If measurements for a species fell within a single expected range, only that state was coded; whereas if a species possessed measurements from multiple expected character states, it was coded as polymorphic. Species containing measurements outside of one expected state yet not in the other were coded as the primitive (0) state. All statistical procedures for separating continuous characters were performed using PAST (Hammer et al., 2001).

Search method.—Phylogenetic analysis was conducted using the criterion of maximum parsimony in PAUP* (Swofford, 2002). A heuristic search was used with 200 addition sequence replicates, holding 10 trees each replication. Branch swapping was performed using the tree-bisection-reconnection (TBR) algorithm. All characters were treated as unweighted and unordered. Taxa exhibiting multiple character states were coded accordingly and treated as polymorphic. Characters were optimized using accelerated transformation (ACCTRAN) and analyzed in MacClade (Maddison and

Maddison, 2000). Character state distributions for species of Glyptorthis are presented in

Table 5.

Table 3 Range of Morphometric Measurements for Species Analyzed

Table 4 Statistical Separation of Continuous Characters. ANOVA Results: p ˂ 0.001 for all Characters. S. D. Indicates Standard Deviation

Character Mean S.D. Range 1. Maximum width (2) small 8.590 1.650 x ≤ 10.510 10.97 ≤ x ≤ (0) moderate 11.680 0.0763 12.443 (1) large 21.970 2.870 x ≥ 19.100

2. Dorsal valve inflation (0) low 0.201 0.032 x ≤ 0.232 (1) high 0.283 0.041 x ≥ 0.242

3. Ventral valve inflation (0) low 0.209 0.045 x ≤ 0.254 (1) high 0.334 0.059 x ≥ 0.275

4. Shell thickness (0) thin 0.392 0.062 x ≤ 0.454 (1) thick 0.580 0.077 x ≥ 0.503

5. Comparative convexity of the dorsal and ventral valves (0) dorsal concavity greater 0.775 0.175 x ≤ 0.949 (1) ventral concavity greater 1.439 0.444 x ≥ 0.995

6. Depth of the shell (1) low 0.754 0.112 x ≤ 0.866 (0) high 0.906 0.031 x ≥ 0.875

7. Origin of the sulcus (0) proximal 0.352 0.078 x ≤ 0.430 (1) distal 0.584 0.080 x ≥ 0.504

11. Costal density (0) low 0.301 0.065 x ≤ 0.366 0.413 ≤ x ≤ (1) moderate 0.470 0.057 0.527 31

(2) high 0.812 0.216 x ≥ 0.596 Table 4 continued

Character Mean S.D. Range

13. Length of the ventral muscle field (0) short 0.318 0.049 x ≤ 0.367 (1) long 0.474 0.035 x ≥ 0.439

14. Max width of the ventral muscle field (0) narrow 0.293 0.035 x ≤ 0.328 (1) wide 0.385 0.026 x ≥ 0.359

15. Width of ventral muscle field anterior to hinge teeth (0) narrow 0.284 0.033 x ≤ 0.317 (1) wide 0.366 0.046 x ≥ 0.320

16. Longitudinal variation of the ventral muscle field (0) wide 0.504 0.121 x ≤ 0.625 (1) narrow 0.803 0.075 x ≥ 0.728

17. Length of the hinge teeth (1) short 0.058 0.011 x ≤ 0.069 (0) long 0.091 0.011 x ≥ 0.080

18. Height of the ventral cardinal area (0) low 0.356 0.109 x ≤ 0.465 0.576 ≤ x (1) moderate 0.641 0.065 ≤ 0.706 (2) high 1.038 0.187 x ≥ 0.851

22. Extension of the ventral umbo across hinge (1) short 0.114 0.028 x ≤ 0.142 (0) long 0.210 0.044 x ≥ 0.166

Table 4 continued

Character Mean S.D. Range

26. Max width of the adductor muscle scars (0) narrow 0.081 0.016 x ≤ 0.097 (1) wide 0.151 0.042 x ≥ 0.109

27. Comparative width of adductor and diductor muscle scars (0) narrow 0.229 0.045 x ≤ 0.274 (1) wide 0.411 0.101 x ≥ 0.310

28. Max width of dorsal muscle field (0) narrow 0.396 0.032 x ≤ 0.427 (1) wide 0.486 0.051 x ≥ 0.435

29. Max length of the dorsal muscle scars (0) short 0.294 0.039 x ≤ 0.333 (1) long 0.409 0.044 x ≥ 0.365

30. Width of dorsal muscle field anterior to brachiophores (0) narrow 0.260 0.060 x ≤ 0.320 (1) wide 0.399 0.042 x ≥ 0.357

31. Height of the dorsal cardinal area (0) low 0.144 0.043 x ≤ 0.187 (1) high 0.336 0.128 x ≥ 0.208

32. Length of brachiophores (1) short 0.074 0.007 x ≤ 0.081 (0) long 0.121 0.025 x ≥ 0.096

34. Extension of the dorsal umbo across hinge (1) short 0.059 0.013 x ≤ 0.072 (0) long 0.107 0.025 x ≥ 0.082

Table 5 Character State Distribution for Species Included in Phylogenetic Analysis X = (0 and 1), Y = (1 and 2), and Z = (0 and 2)

RESULTS

Phylogenetic analysis.—Parsimony analysis resulted in a single most parsimonious tree with a length of 256 steps (Figure 2). The consistency and retention indices are 0.469 and 0.509, respectively, which is significantly higher than those derived from similar sized matrices using randomly generated data at the α = 0.05 level

(Klassen et al., 1998). Confidence values for node support were recovered using 100 repetitions of a full-heuristic Jackknife analysis with 5% deletion (equivalent to two characters). Jackknife values for nodes compatible with the single most parsimonious tree are presented in Figure 2. Further support was assessed by calculating the g1 statistic, a metric for assessing the degree of skewness in the distribution of tree lengths. The g1 statistic for 10,000 trees generated from the data matrix is -0.148, indicating strong phylogenetic signal within the data (P < 0.05: Hillis and Huelsenbeck, 1992).

Recognition of clades.—Overall tree topology is balanced, with each clade supported by multiple synapomorphies. The monophyly of the ingroup is strongly supported by

Jackknife analysis and character support (Figure 2). Characters supporting the monophyly of Glyptorthis include a high inflation of the ventral valve, a distally originating sulcus, subquadrate shell outline, strongly lamellose concentric growth lines, and a wide ventral muscle field (characters 3, 7, 8, 10, and 14).

Exclusive of Glyptorthis costellata Cooper, 1956, Glyptorthis species are distributed between two major clades (Figure 2). The sister relationship between these clades is supported by a high inflation of the dorsal valve, thick valves, and a narrow dorsal umbonal angle (characters 2, 4, and 33). The first clade includes G. alta Cooper,

1956, G. bellarugosa, (Conrad, 1843); G. irregularis, Cooper, 1956, G. assimilis Cooper,

1956, G. uniformis Cooper, 1956, G. subcircularis Cooper, 1956, G. multicostellata

Cooper, 1956, and G. sulcata Cooper, 1956. The second clade includes G. equiconvexa

Cooper, 1956, G. glaseri Alberstadt, 1973, G. insculpta (Hall, 1847), G. maquoketensis

(Ladd, 1929), G. senecta Cooper, 1956, G. obesa Cooper, 1956, G. bellatula Cooper,

1956, G. concinnula Cooper, 1956, G. uncinata Cooper, 1956, G. glypta Cooper, 1956,

G. daytonensis (Foerste, 1885), G. fausta (Foerste, 1885), G. crenulata Cooper, 1956, and G. pulchra Wang, 1949. For simplicity, these are referred to below as the G. alta and

G. equiconvexa clades.

The monophyly of the G. alta clade is supported by a wide interspace between costae and costellae, moderate costal density, long ventral diductor muscle scars, and wide ventral muscle scars directly anterior to the hinge teeth (characters 9, 11, 13, and

15). The monophyly of the subclade comprising G. alta, G. bellarugosa, and G. irregularis is supported by weakly lamellose ornamentation, a high ventral cardinal area, and a wide ventral umbonal angle (characters 10, 18, and 20). Further, the sister relationship between G. bellarugosa and G. irregularis is supported by low inflation of the dorsal valve, proximal origin of the sulcus, a medial septum separating the diductor muscle scars, and short brachiophores (characters 2, 7, 24, and 32). The pectinate 37 subclade comprising G. assimilis, G. uniformis, G. subcircularis, G.multicostellata, and

G. sulcata is supported by anterior and lateral deflections of the ventral muscle scars and wide dorsal muscle scars with respect to both maximum width and localized width anterior of the brachiophores (characters 12, 28, and 30). The monophyly of G. uniformis,

G. subcircularis, G. multicostellata, and G. sulcata is supported by the presence of a ventral medial septum and the possession of a raised platform on which the ventral muscle rest (characters 24 and 25). A convex dorsal valve, subelliptical outline of the shell, small hinge teeth, and a short extension of the dorsal umbo across the hinge line support the monophyly of G. subcircularis, G. multicostellata, and G. sulcata (characters

5, 8, 17, and 34). The sister group relationship between G. multicostellata and G. sulcata exhibits strong Jackknife support and is supported by long dorsal muscle scars and high costal density (characters 11 and 29).

Synapomorphies supporting the monophyly G. equiconvexa clade include low costal density, low variability in ventral muscle scar width, wide adductor muscles

(within the ventral field and comparative with the diductor muscles), long brachiophores, and a short dorsal umbo extending across the hinge line (characters 11, 16, 26, 27, 32, and 34). Phylogenetic structure nested within the G. equiconvexa clade reveals three relatively symmetrical subclades whose shared common ancestry is supported by a proximal origination of the sulcus, short hinge teeth, a long ventral umbo, significant extension of the ventral umbo across the hinge line, and a short dorsal cardinal area

(characters 7, 17, 22, 23, and 31). 38

The first subclade nested within the G. equiconvexa clade is comprised of G. glaseri, G. insculpta, and G. maquoketensis. Monophyly of this clade is supported by their large overall body size and long diductor scars in the ventral muscle field

(characters 1 and 13). Glyptorthis insculpta and G. maquoketensis are sister taxa whose relationship is supported by low vertical height of the shell, an anterior and later deflection in the outline of ventral muscle scars, a medial septum in the ventral diductor field, and long dorsal muscle scars (characters 4, 12, 24, and 29). Both the clade consisting of G. glaseri, G. insculpta, and G. maquoketensis and the sister relationship between G. insculpta and G. maquoketensis are strongly supported by Jackknife values

(Figure 2).

The monophyly of the two remaining clades within the G. equiconvexa clade is supported by sharing a high costal density, narrow ventral muscle scars, and a tall ventral cardinal area (characters 11, 14, and 18). The pectinate clade comprising G. senecta, G. obesa, G. bellatula, and G. concinnula is supported by sharing a wide interspatial width between costae and costellae (character 9); whereas G. obesa, G. bellatula, and G. concinnula share an elliptical outline of the shell (character 8). The sister relationship of

G. bellatula and G. concinnula is supported by adductor muscle emplacement on a raised track and wide dorsal muscle scars (characters 25, 28, and 30).

The pectinate topology for the third subclade nested with the G. equiconvexa clade is well supported by Jackknife values (Figure 2) and strongly supported by multiple synapomorphies. The monophyly of G. uncinata, G. glypta, G. daytonensis, G. fausta, G. crenulata, and G. pulchra is supported by a low inflation of the dorsal valve and depth of 39 the shell (characters 2 and 6). The clade comprising species G. glypta crownward are supported by a relatively short extension of the ventral umbo across the hinge line, as well as four characters relating to the high angle orientation of their ventral cardinal area, ventral umbo, dorsal umbo, and delthryium (characters 19, 20, 21, 22, and 33). The monophyly of G. daytonensis, G. fausta, G. crenulata, and G. pulchra is supported shared dorsal valve convexity, low shell height, weakly lamellose concentric ornamentation, and low costal density (characters 4, 5, 10, and 11). Monophyly of the remaining species is supported by a lowly inflated ventral valve and multiple deflections in the outline shape of the ventral muscle scars (characters 3 and 12). The sister relationship between G. crenulata and G. pulchra is strongly supported by five characters indicating significantly large muscle scars in both the ventral and dorsal valves

(characters 13, 14, 15, 28, and 30).

DISCUSSION

North American species of Glyptorthis first occur in Upper Whiterockian

(Darrwilian) rocks from the Appalachian Basin (Cooper, 1956) and subsequently underwent rapid diversification. A total of 16 species originate in the overlying

Mohawkian Series (upper Sandbian to lower Katian Series) corresponding to maximal species-level diversity within the genus (70% of the species included in this analysis).

This episode of taxonomic diversification was accompanied by geographic expansion, as species of Glyptorthis became widely distributed across the Laurentian paleocontinent

(Fig. 3, Table 6). In addition, numerous occurrences of Glyptorthis ranging from the 40

Middle Ordovician (Darrwilian) to Middle Silurian (Homerian) have been reported from outside North America (see Systematic Paleontology, Table 7).

Glyptorthis species occur primarily in shallow marine units interpreted to be deposited within normal or storm wave base (Holland and Patzkowsky, 1996). Based on their relatively large size, robust shell, and large pedicle foramen, G. insculpta has been reconstructed as living with the pedicle attached to large fragments within a shell gravel substrate and the commissure oriented nearly vertically (Richards, 1972). Several morphological innovations diagnostic of Glyptorthis subclades are correlated with modifications of this general design that potentially reflect adaptive changes. 41

A signature feature of Glyptorthis is its concentric lamellose ornamentation.

Richards (1972) noted that orthid brachiopods with highly costellate shells display a low relative frequency of encrustation by epibionts. Variations in the structure of this ornament, such as increased costal density and lamellose frills characteristic of the G. costellata clade, may have been further adaptations for increased resistance to epibiont parasitism by increasing the irregularity of the shell surface (Richards, 1972).

Species within the G. assimilis subclade from G. uniformis crownward emplace their adductor muscles onto a raised platform rather than the floor of the valve. The evolution of raised muscle platforms is common in orthids, and may represent a solution to shortening the length of the adductor in biconvex brachiopods with columnar rather than tendonous muscles (Rudwick, 1970). This shortening of the adductor muscle length may have increased valve kinetics and reduced the energy required for valve closure, suggesting adaptation to higher energy environments.

The G. equiconvexa clade is characterized a size reduction of their dorsal and ventral posterior morphology and an increase in the size of the ventral muscle field.

Furthermore, species within the G. equiconvexa clade exhibit small cardinal areas, short hinge teeth, low brachiophores, a truncated dorsal and ventral umbo, and wide adductor tracks. This suite of morphologic features indicates development of a more robust musculature and pedicle attachment, which may have allowed increased success under higher energy conditions.

As a genus, Glyptorthis was not restricted to a single depositional environment.

Species occur in disparate tectonic settings with different bathymetric and siliciclastic 44 conditions. For example, species of Glyptorthis occur in mixed zones (Cincinnati region), distal interior cratonic basins (e.g., Iowa Basin), and carbonate platforms (Nashville

Dome). Individual species of Glyptorthis demonstrate a remarkable propensity for interbasinal dispersal. For example, G. assimilis occurs in lower Mohawkian (upper

Sandbian) deposits from both Tennessee and British Columbia (Cooper, 1956; Jin and

Norford, 1996): a biogeographic range that crosses a significant paleolatitudinal gradient and spans nearly the entire north-south transect of the Laurentian craton (Jin, 2012).

Many other species of Glyptorthis also occur in multiple tectonic basins. For example, the Richmondian species G. maquoketensis occurs in correlative deposits from Iowa

(Ladd, 1929), Texas (Howe, 1966), Missouri (Howe, 1988), and Tennessee (Howe,

1988).

As noted above, species of Glyptorthis share a high propensity for dispersal and broad ecological tolerances, both of which are characteristic of invasive species (Davis,

2009; Stigall, 2010a). Glyptorthis insculpta participated in the Late Ordovician

Richmondian Invasion into the Cincinnati basin (Foerste, 1917; Stigall, 2010a). The phylogenetic hypothesis in Figure 2 indicates that G. insculpta is nested within the subclade consisting of the cosmopolitan species G. maquoketensis and G. glaseri from the Arbuckle Mountains of Oklahoma. This subclade has strong biogeographic affinities to the mid-Continent region west of the Cincinnati basin (Table 6). The phylogenetic topology presented herein suggests that G. insculpta may have invaded the Cincinnati basin from low latitude subtropical regions of the mid-Continent (Wright and Stigall, 45

2012). Further investigations are needed to assess whether or not the pathway of invasion for G. insculpta represents a general pattern for invader taxa (Chapter 4).

Table 7 Species of Glyptorthis Reported from Outside North America

Species Location G. balcletchiensis (Davidson, 1883) British Isles G. crispa (M'Coy, 1846) British Isles G. irrupta Rubel, 1962 Estonia G. maritima Wright, 1964 British Isles G. plana Rõõmusoks, 1964 Estonia G. sagana Havlícek, 1977 Czech Republic G. sarcina Zhan and Jin, 2005 China G. speciosa Reed, 1944 British Isles G. squamata Rõõmusoks, 1964 Estonia G. transista Reed, 1952 British Isles G. whitei Bassett, 1972 Wales

CONCLUSIONS

The orthid brachiopod genus Glyptorthis was both diverse and species-rich.

Species of Glyptorthis were most common in Late Ordovician benthic communities and became geographically widespread throughout the Laurentian paleocontinent.

Phylogenetic analysis has documented evolutionary patterns and informed systematic revisions for species of Glyptorthis, thereby providing an evolutionary framework for future investigations documenting species-level paleobiogeographic patterns surrounding the Late Ordovician Richmondian Invasion. The phylogenetic placement of the invasive species G. insculpta in the recovered tree topology indicates G. insculpta invaded into the 46

Cincinnati basin from a region located to the paleo-west, such as the southern mid-

Continent.

SYSTEMATIC PALEONTOLOGY

Specimens examined for species that require neither of taxonomic revision nor lectotype designation are presented in Table 8. Each specimen listed belongs to a species of Glyptorthis that is considered valid herein. The original descriptions for these species are sufficient for diagnosis, and additional discussion is, therefore, not warranted here.

The character state distribution data from the phylogenetic analysis presented in this paper may be combined with the original species descriptions to provide enhanced diagnoses (see Table 5).

Order ORTHIDA Schuchert and Cooper, 1932

Suborder ORTHIDINA Schuchert and Cooper, 1932

Superfamily ORTHOIDEA Woodward, 1852

Family GLYPTORTHIDAE Schuchert and Cooper, 1931

Genus GLYPTORTHIS Foerste, 1914a

Type species.—Orthis insculpta Hall, 1847.

Other valid species.—Glyptorthis alta Cooper, 1956; G. assimilis Cooper, 1956;

G. bellatula Cooper, 1956; G. concinnula Cooper, 1956; G. costellata Cooper, 1956; G. crenulata Cooper, 1956; G. equiconvexa Cooper, 1956; G. glaseri Alberstabt, 1973; G. glypta Cooper, 1956; G. irregularis Cooper, 1956; G. maquoketensis Ladd, 1929; G. 47 marilara Li and Copper, 2006; G. maritima Wright, 1964; G. multicostellata Cooper,

1956; G. obesa Cooper, 1956; G. plana Rõõmusoks, 1964; G. pulchra Wang, 1949; G. sarcina Zhan and Jin, 2005; G. senecta Cooper, 1956; G. subcircularis Cooper, 1956; G. sulcata Cooper, 1956; G. uncinata Cooper, 1956; and G. uniformis Cooper, 1956; Orthis bellarugosa Conrad, 1843; Orthis daytonensis Foerste, 1885; Orthis fausta Foerste, 1885.

Potentially valid species.—Glyptorthis irrupta Rubel, 1692; G. rara Cooper,

1956; G. robusta Sheehan and Lespérance, 1979; G. sagana Havlícek, 1977; G. speciosa

Reed, 1944; G. squamata Rõõmusoks, 1964; G. subcarinata Cooper, 1956; G. transista

Reed, 1952; G. transversa Cooper, 1956; G. whitei Bassett, 1972; Orthis crispa M’Coy,

1846; O. crispata Emmons, 1842; O. balcletchiensis Davidson, 1883.

Diagnosis.—Shell of variable size, subquadrate to subelliptical in outline, unequally biconvex; ramnicostellate to fascicostellate, sometimes frilled; ornamentation concentric, strongly lamellose; rectimarginate to sulcate; delthryium open, triangular; ventral interarea short to long, orthocline to catacline; dorsal interarea small, orthocline to apsacline; subcordate ventral muscle scars, deflecting anteriorly, additional lateral deflection not uncommon; ventral adductor scars sometimes elevated on a raised platform; quadripartite dorsal muscle scars; thickened notothyrium; brachiophores divergent.

Occurrence.—Middle Ordovician (Darrwilian) to middle Silurian (Wenlock:

Homerian) of North America, Estonia, Czech Republic, China, and the British Isles (see below). 48

Discussion.— Glyptorthis most closely resembles Bassettella Zuykov and Butts,

2008 and Eridorthis Foerste, 1909. Species belonging to Glyptorthis may be distinguished from Bassettella by the growth lamellae that lack perforations, whereas

Glyptorthis differs from Eridorthis because the median sulcus is located on the dorsal rather than ventral valve.

North American species previously referred to Glyptorthis but excluded from analysis due to insufficient character data were: G. crispata (Emmons, 1842); G. rara

Cooper, 1956; G. transversa Cooper, 1956; G. subcarinata Cooper, 1956; G. robusta

Sheehan and Lespérance, 1979, and G. marilara Li and Copper, 2006.

Species of Glyptorthis have been reported from outside North America including the British Isles, Estonia, China, and the Czech Republic (Table 7). Although Wright

(1964) has suggested there are no close affinities between species of Glyptorthis from different paleocontinents, assessing the taxonomic validity of species occurring outside

North America is beyond the scope of this paper. Future studies on Glyptorthis may find it useful to emend the phylogenetic analysis presented here with a taxonomic revision of species listed in Table 7.

GLYPTORTHIS DAYTONENSIS (Foerste, 1885)

Figures 4.1-4.2

Orthis daytonensis FOERSTE, 1885, p. 87, pl. 13, figs. 13a-13d, 20a-20b, 21

Glyptorthis daytonensis (Foerste, 1885) SCHUCHERT AND COOPER, 1932, p. 90 49

Diagnosis.—Emended from Foerste (1885). Shell large with subquadrate outline; much wider than long. Dorsal valve highly convex. Exterior ornamentation finely lamellose; sulcus proximal. Ventral interarea extends far above the hinge line; hinge teeth short; umbos of both valves small. Ventral muscle scars narrow with anterior deflection.

Types.—Foerste (1885) described and figured both USNM 84828 and USNM

84831, but he did not designate a holotype. USNM 84828 is herein designated as the lectotype, and USNM 82831 becomes a paralectotype.

Additional material.— USNM 546106; USNM 546108.

Occurrence.— Both Foerste’s (1885) types and the non-type specimens from the

U. S. National Museum are from the Brassfield Formation, Ohio; Lower Silurian

Niagaran Series (Aeronian).

Discussion.— Two Silurian species of Glyptorthis occur in Ohio: G. daytonensis and G. fausta. Glyptorthis daytonensis is the larger of the two species, and it can be distinguished from G. fausta by its high ventral valve, subquadrate outline, and high ventral cardinal area.

GLYPTORTHIS FAUSTA (Foerste, 1885)

Figures 4.3-4.7

Orthis fausta FOERSTE, 1885, p. 85, pl. 13, figs. 15a-15d, 16a-16b

Glyptorthis fausta (Foerste, 1885) SCHUCHERT AND COOPER, 1932, p. 90, pl. 5, figs.

4, 7, 8.

Diagnosis.—Emended from Foerste (1885). Shell moderate to large size; outline subrectangular; width much greater than length. Ventral valve thin and shallow. Ventral muscle scars narrow, with anterior and lateral deflections; ventral adductors emplace onto a raised platform. Ventral interarea of low relief; long posterior extension of the ventral umbo across the hinge line.

Types.— Foerste (1885) described and figured three articulated specimens but did not designate a holotype. USNM 84829 is herein designated as the lectotype, and both specimens in USNM 84813 (includes two specimens) become paralectotypes.

Additional material.—USNM 546109.

Occurrence.—All specimens of G. fausta examined are from the Brassfield

Formation, Ohio; Lower Silurian Niagaran Series (Aeronian).

Discussion.—This species co-occurs with G. daytonensis in Lower Silurian strata of Ohio. Glyptorthis fausta differs from G. daytonensis in having a shallow ventral valve, low ventral interarea, a significant extension of the ventral umbo across the hinge line, ventral muscle scars that deflect anteriorly and laterally, and the adductor muscle scars emplaced on a raised platform.

GLYPTORTHIS INSCULPTA (Hall, 1847)

Figures 5.1-5.5

Orthis insculpta HALL, 1847, p. 125, pl. 32, fig. 12.

Hebertella insculpta (Hall, 1847) HALL AND CLARK, 1882, pl. 5a, fig. 13.

Glyptorthis insculpta (Hall, 1847) FOERSTE, 1914, p. 257; SCHUCHERT AND

COOPER, 1932, pl. 6, figs. 17, 18, 20, 21, 26, 26; WILLIAMS AND WRIGHT,

1965, p. H319, pl. 200, figs.3a-c; HOWE, 1966, pl. 28, fig. 21; DAVIS, 1985, pl.

8, figs 1-4; SCHWIMMER AND SANDY, 1996, pl. 16, figs. 17-20; ZUYKOV

AND BUTTS, 2008, figs. 1-6; WILLIAMS AND HARPER, 2000, p. 733, pl.

530, figs. 1a-e.

Diagnosis.—Emended from Hall (1847) and Schuchert and Cooper (1932). Shell large; subquadrate; concentric ornamentation weakly to strongly lamellose; costal density low to moderate. Ventral muscle scars large with lateral and anterior deflections; narrowing near the hinge teeth. Ventral diductor scars separated by a medial septum; adductor scars of variable width, widening towards the anterior. Dorsal cardinal area high, extending over the hinge line; umbonal angle less than 150o.

Types.— AMNH 30260 (holotype).

Additional material.—YPM S-242a; YPM S-242b; YPM S-242c; YPM S-242d

Occurrence.—Waynesville and Liberty Formations; Ohio and Indiana; Upper

Ordovician Cincinnatian Series (Upper Katian). 53

Discussion.—Specimens of G. insculpta most closely resemble G. maquoketensis, but can readily be distinguished by its larger size and shallow lateral profile. See below for a full comparison of G. insculpta and G. maquoketensis.

GLYPTORTHIS MAQUOKETENSIS Ladd, 1929

Figures 5.6-5.8

Hebertella (Glyptorthis) insculpta maquoketensis LADD, 1929, p. 399, pl. 4, figs. 13 -16,

pl. 5, figs 1-2.

Hebertella insculpta (Ladd, 1929) BASSLER, 1932, pl. 24, figs. 1-2.

Glyptorthis insculpta maquoketensis (Ladd, 1929) HOWE ,1966, p. 242, pl. 28, figs. 10-

20; HOWE, 1988, p. 209, figs. 4.16-4.21, 7.

Diagnosis.—Emended from Ladd (1929). Shell of moderate to large size; outline subquadrate; dorsal valve inflated. Sulcus deep, originating proximal to the umbo.

Concentric ornamentation strongly lamellose. Ventral muscle scars large, width variable, with an anterior deflection, resting on a raised platform.. Ventral cardinal area low; delthyrial angle high.

Types.—Ladd (1929) figured two specimens in his original description, yet did not designate a holotype. SUI 66500 is herein designated as the lectotype and SUI 66502 the paralectotype.

Additional material.— USNM 146540d, figured by Howe (1988). 54

Occurrence.—Maquoketa Formation, Iowa; Upham and Aleman Limestones,

Texas; Cape Limestone, Missouri; and the Arnheim Formation, Tennessee; Upper

Ordovician Cincinnatian Series (Upper Katian).

Discussion.—Glyptorthis maquoketensis was originally described by Ladd (1929) as a subspecies of Glyptorthis insculpta. Ladd (1929) noted that specimens of G. insculpta maquoketensis differed from G. insculpta because they are typically smaller, more convex, and have a raised platform in the ventral muscle field. Although Howe

(1966; 1988) reviewed considerable morphological differences between G. insculpta maquoketensis and G. insculpta, the subspecific designation of G. insculpta maquoketensis has not previously been questioned (e.g., Howe, 1966). Because Ladd’s original syntypes were available, they were coded separately from G. insculpta to test the subspecific validity of G. insculpta maquoketensis. The tree topology resulting from phylogenetic analysis (Figure 2) confirms that G. insculpta and G. insculpta maquoketensis are closely related and form a sister group to G. glaseri. Glyptorthis maquoketensis is herein recognized as a distinct species from G. insculpta based on eight morphological characters related to valve convexity, shell depth, variable ventral muscle scar width, length of the hinge teeth, delthyrial angle, and size of the dorsal cardinal area and umbo.

Figure 5 1, Glyptorthis insculpta (Hall, 1847), AMNH 30260 (holotype), ventral view, ×2.6; 2, ventral interior, ×2.5; 3, YPM S-242b, ventral interior, ×2.4; 4, YPM S-242d, dorsal interior, ×2.3; 5, YPM S-242a, ventral view, ×2.3; 6, G. maquoketensis Ladd, 1929, SUI 66500 (lectotype), ventral view, ×2.3; 7, dorsal view, ×2.3; 8, SUI 66502 (paralectotype), ventral interior, ×2.5.

GLYPTORTHIS PULCHRA Wang, 1949

Glyptorthis pulchra WANG, 1949, p. 4, pl. 1, figs. 1e-10.

Diagnosis.—Emended from Wang (1949). Shell moderate to large; ornamentation weakly lamellose; interspatial distance between costae and costellae wide. Ventral cardinal area wide, elevated significantly above hinge line; umbo extending over the hinge line. Ventral diductor muscle tracks separated by medial septum; adductor scars on a raised platform. Dorsal muscle scars large, slightly wider than long. Brachiophores short; dorsal umbo extending over the hinge line.

Types.—Holotype USNM 111937; paratypes USNM 111938a and USNM

111938b.

Occurrence.—Upper part of the Maquoketa Formation, Iowa; Upper Ordovician

Cincinnatian Series (Upper Katian).

Discussion.—Although represented in Ladd’s (1929) collection of Maquoketa brachiopods from Iowa, this species was first described and illustrated by Wang (1949) from E. O. Ulrich’s collection in the U. S. National Museum. Wang (1949) concluded that G. pulchra was most similar to Ladd’s (1929) G. insculpta maquoketensis, but differed in the placement of the adductor muscle scars and convexity of the later profile.

Howe (1966) synonymized G. pulchra with G. insculpta maquoketensis due to similarities in overall external morphology and stratigraphy and doubted the utility of the adductor muscle platform as a useful taxonomic indicator.

Because specimens of G. pulchra were readily available, they were coded separately in this analysis to assess species validity by examining the relative overlap of 57 character states and phylogenetic placement. Phylogenetic analysis placed G. pulchra and

G. maquoketensis in separate subclades nested within the G. equiconvexa clade (Figure

2), indicating that the overall morphological similarity of the two species resulted from a high level of homeomorphy between the two species rather than a close phylogenetic relationship. In addition to upholding Wang’s (1949) original taxonomic diagnoses, character analysis reveals G. pulchra and G. maquoketensis are distinguished by 15 additional morphological characters (Table 4). Because G. pulchra and G. maquoketensis co-occur in the Maquoketa Formation, proper taxonomic identification requires careful examination of both external and internal morphological characters.

GLYPTORTHIS UNIFORMIS Cooper, 1956

Figures 4.8-4.9

G. aff. G. bellarugosa (Cooper, 1956) BUTTS, 1942, p. 93

Glyptorthis uniformis COOPER, 1956, p. 378, pl. 46C, figs. 11-17, pl. 29I, figs. 45-51.

Glyptorthis virginica (Cooper, 1956) COOPER, 1956, p. 379, pl. 48, D, 19-33.

Diagnosis.—Emended from Cooper (1956). Shell size moderate, outline subquadrate; slightly wider than long. Ornamentation strongly lamellose; costal density low to moderate. Valves convexity variable in lateral profile. Ventral muscle scars large, deflecting anteriorly and laterally. Ventral umbonal area extending significantly over the hinge line, angle less than 140o. Brachiophores small. Dorsal muscle scars large and rectangular, wider than long. 58

Types.—Holotype USNM 116945a; Paratypes 116944a, 116945b, 116946a,

11694b.

Additional material.—USNM 98209a, USNM 98209d, USNM 110082a, USNM

110082b, USNM 110082g.

Occurrence.— Benbolt, Sevier, and Wardell Formations of Tennessee and

Virginia; Upper Ordovician Mohawkian Series (Upper Sandbian).

Discussion.—This highly variable species was first described by Cooper (1956) from Benbolt and Sevier Formations of Virginia and Tennessee. Cooper (1956, p. 379) indicated that G. uniformis was reminiscent of G. virginica Cooper, 1956 from the overlying Wardell Formation, yet differed in that G. virginica was more coarsely costellated. After examining the total range of morphological variation within the type series of G. uniformis and G. virginica, the distribution of character states used for phylogenetic analysis failed to reveal any significant difference between the two species.

The name G. uniformis has page priority over G. virginica. Glyptorthis virginica is, therefore, considered a junior synonym of G. uniformis. The stratigraphic range of G. uniformis is, therefore, increased, extending from the Ashbyan to the Black Riverian of the Mohawkian Series. 59

Table 8 Specimens Examined for Species That Do Not Require Taxonomic Revision or Lectotype Designation * Indicates Non-Type Specimens

Species Museum Catalog number Type

G. alta USNM 110008 Holotype

G. assimilis USNM 110013j Holotype USNM 110012c Paratype USNM 110012e Paratype

G. bellarugosa AMNH 909 Lectotype AMNH unnumbered Paralectotype YPM S-261 * YPM S-249 *

G. bellatula USNM 110015a Holotype USNM 110015d Paratype

G. concinnula USNM 116931a Holotype USNM 110028a Paratype USNM 110034a Paratype USNM 10035a Paratype USNM 110028a Paratype

G. costellata USNM 110052a Holotype USNM 110052b Paratype USNM 110059b Paratype

G. crenulata USNM 116932a Holotype USNM 116932b Paratype USNM 110054 Paratype

G. equiconvexa USNM 116934a Holotype USNM 116933a Paratype USNM 116933b Paratype

Table 8 continued

Species Museum Catalog number Type Holotype G. glaseri OMNH 5819 (v.v.) Holotype OMHN 5820 (d.v.)

G. glypta USNM 11067c Holotype USNM 110067f Paratype

G. irregularis USNM 116936a Holotype USNM 116936d Paratype USNM 116936c Paratype

G. multicostellata USNM 116938b Holotype USNM 116938a Paratype USNM 116938c Paratype USNM 116938e Paratype

G. obesa USNM 110068d Holotype USNM 110068a Paratype USNM 110068d Paratype

G. pulchra USNM 111937 Holotype USNM 111938 Paratype

G. senecta USNM 110070h Holotype USNM 110070a Paratype USNM 110070b Paratype

G. subcircularis USNM 84015 Holotype USNM 116940a Paratype USNM 116940b Paratype

Table 8 continued

Species Museum Catalog number Type

G. sulcata USNM 110075a Holotype USNM 110074 Paratype USNM 110076e Paratype USNM 110076b Paratype

G. uncinata USNM 116943a Holotype USNM 116941b Paratype USNM 116941c Paratype USNM 116941a Paratype

CHAPTER 3: PHYLOGENETIC REVISION OF THE LATE ORDOVICIAN

ORTHID BRACHIOPOD GENERA PLAESIOMYS AND HEBERTELLA FROM

NORTH AMERICA

ABSTRACT.—The orthid brachiopod genera Plaesiomys and Hebertella are significant constituents of Late Ordovician benthic marine communities throughout North

America. Species-level phylogenetic analyses were conducted on both genera to inform systematic revisions and document evolutionary relationships. Plaesiomys cutterensis, P. idahoensis, and P. occidentalis are herein recognized as distinct species rather than subspecies of P. subquadratus. Similarly, Hebertella montoyensis and H. prestonensis are recognized as distinct species separate from H. occidentalis, and H. richmondensis is recognized as a distinct species rather than a geographical variant of H. alveata.

Hebertella subjugata is removed from its tentative synonymy with H. occidentalis and revalidated.

The development of species-level evolutionary hypotheses for Plaesiomys and

Hebertella provides a detailed framework for assessing evolutionary and paleobiogeographic patterns of Late Ordovician brachiopods from Laurentia. Plaesiomys subquadratus participated in the Late Ordovician Richmondian Invasion. The recovered phylogenetic topology for Plaesiomys suggests that P. subquadratus invaded the

Cincinnati basin from the marginal, Anticosti Island region of Quebec. The geographic range of Hebertella expanded throughout Laurentia during the Richmondian into both intracratonic and marginal basins. 63

INTRODUCTION

Orthid brachiopods were among the most common benthic organisms in Late

Ordovician benthic marine communities. Despite their abundance, diversity, and excellent preservation, few species-level evolutionary hypotheses for orthid brachiopods have been constructed (e.g., Stigall Rode, 2005) and only one involves a Late Ordovician taxon (Wright and Stigall, in review). The purpose of this study is to develop phylogenetically-informed taxonomic revisions for two significant genera representative of Late Ordovician orthid brachiopods: Hebertella and Plaesiomys. The phylogenetic revisions presented herein aims to assess species validity, taxonomy, and document evolutionary relationships among species from North America. In addition, the construction of evolutionary hypotheses for Hebertella and Plaesiomys may provide valuable data for constraining species-level analyses of paleobiogeographic patterns for

Late Ordovician orthid brachiopods distributed throughout the Laurentian paleocontinent.

For example, Plaesiomys subquadratus (Hall, 1847) participated in the Late Ordovician

Richmondian Invasion, an event which introduced over 50 genera to the present-day

Cincinnati region (Foerste, 1912; Meyer and Davis, 2009). The phylogenetic placement of P. subquadratus may be used to test competing paleobiogeographic hypotheses regarding the geographic source of the Richmondian invaders (see Stigall, 2010a).

REVISION OF PLAESIOMYS

Taxonomic background.—The orthidine brachiopod genus Plaesiomys Hall and

Clarke, 1892 ranges throughout the Late Ordovician (Katian to Hirnantian). Plaesiomys reached maximal diversity during the late Katian and is a significant constituent of benthic marine communities in Late Ordovician deposits of the Northern Hemisphere.

Although several emended diagnoses have been published (e.g., Jin and Zhan, 2007;

Macomber, 1970), previous taxonomic revisions of Plaesiomys have focused primarily on describing new species or shifting the status of Plaesiomys from a discrete genus to a subgenus within Dinorthis Hall and Clark, 1892 or back. Although suggested by

Macomber (1970), no prior study has utilized phylogenetic analysis to inform a comprehensive taxonomic revision or assess evolutionary relationships among species of

Plaesiomys from North America.

Plaesiomys was erected by Hall and Clarke in 1892 to contain species with shells morphologically similar to Orthis subquadrata Hall, 1847. Hall and Clarke (1892) considered Plaesiomys to be closely allied with Hebertella Hall and Clarke, 1892, but

Plaesiomys was synonymized with Dinorthis by Winchell and Schuchert (1893). There has been much controversy regarding the taxonomic relationship between Plaesiomys and

Dinorthis (Jin and Zhan, 2001; Macomber, 1970; Schuchert, 1913; Schuchert and

Cooper, 1932; Wang, 1949; Winchell and Schuchert, 1893). For example, Schuchert and

Cooper (1932) considered Plaesiomys a subgenus of Dinorthis, whereas Williams and

Wright (1965) reversed this treatment and considered Dinorthis a subgenus of

Plaesiomys. 65

Although multiple criteria have been used to distinguish Plaesiomys and

Dinorthis, generic differentiation hinges primarily on whether the shell exterior is costate or multicostellate. Authors disagree whether a continuum exists between the costate

Dinorthis and the fully multicostellate Plaesiomys (Schuchert and Cooper, 1932;

Winchell and Schuchert, 1893). Wang (1949) considered Dinorthis to encompass species exhibiting costae that do not branch or in which secondary costae are less coarse than primary costae, whereas he assigned all multicostellate species to Plaesiomys.

Macomber’s diagnosis of Dinorthis (1970) was more inclusive and included multicostellate species in which the secondary costae occur only on the lateral margins of the shell. Williams and Harper (2000) reverted to Hall and Clarke’s (1892) original designation and considered Plaesiomys and Dinorthis to be separate, distinct genera.

Jin and Zhan (2001) provided support for this distinction by noting that Dinorthis has coarse costae with rare evidence of bifurcation, whereas Plaesiomys has fine costae and is always multicostellate. In addition, species of Dinorthis are typically rectimarginate, whereas species of Plaesiomys are usually uniplicate (Jin and Zhan,

2001). In contrast to previous taxonomic treatments which considered the ventral muscle scars of Plaesiomys and Dinorthis indistinguishable (Schuchert, 1913; Schuchert and

Cooper, 1932), Jin and Zhan (2001) noted two important differences: 1) the ventral muscle scars of Dinorthis are more strongly bilobed than those of Plaesiomys and exhibit a medial notch at the anterior margin of the muscle field, and 2) the anterior margin of the ventral adductor muscle scars extend to the medial notch in Dinorthis yet is enclosed within the diductor scars in Plaesiomys. The suite of characters identified by Jin and 66

Zhan (2001) strongly supports the distinction of these two genera. The monophyly of

Plaesiomys will be investigated as part of this analysis.

The phylogenetic analysis of North American species of Plaesiomys presented herein provides a comprehensive taxonomic revision in which species validity is assessed, taxonomic designations are emended, and evolutionary relationships are investigated among species distributed throughout the Laurentian paleocontinent.

MATERIALS AND METHODS

Taxa analyzed.—Specimens belonging to ten species of Plaesiomys occurring in

North America were examined for phylogenetic analysis (Table 9). All known Late

Ordovician species available were examined for morphological character data except for

P. rockymontana Wilson, 1926, a rare species only known from incomplete type specimens (Jin, personal communication, 2011). Species of Plaesiomys occurring outside of North America were excluded because this investigation is limited to assessing evolutionary patterns in Laurentian species. Because Plaesiomys and Dinorthis share many morphological features and have previously been considered sister taxa (Schuchert and Cooper, 1932), Dinorthis was chosen as the outgroup for character polarization.

Three species of Dinorthis were included in the analysis: D. meedsi arctica Schuchert,

1900, D. sweeneyi (Winchell, 1881), and D. venusta Cooper, 1956. These species were chosen because the type specimens were well preserved and representative of Dinorthis.

Unfortunately, the type species of Dinorthis, Orthis pectinella Emmons, 1942 is poorly

67 preserved and consequently not sufficiently informative for character analysis (Jin et al.

2007).

Table 9 Species Included in Phylogenetic Analysis of Plaesiomys * Indicates Outgroup

Dinorthis meedsi arctica* Schuchert, 1900 Dinorthis sweeneyi* (Winchell, 1881) Dinorthis venusta* Cooper, 1956 Plaesiomys anticostiensis (Shaler, 1865) Plaesiomys belilamellosus Wang, 1949 Plaesiomys bellistriatus Wang, 1949 Plaesiomys carltona (Twenhofel, 1928) Plaesiomys cutterensis Howe, 1966 Plaesiomys idahoensis Ross, 1949 Plaesiomys occidentalis Ladd, 1929 Plaesiomys proavitus (Winchell and Schuchert, 1892) Plaesiomys subcircularis (Roy, 1941) Plaesiomys subquadratus (Hall, 1847)

The phylogenetic analysis presented here includes only a sampling of Dinorthis species and, therefore, cannot fully test the sister relationship between Plaesiomys and

Dinorthis. However, data regarding the presence or absence of secondary costae and their abundance were coded in the character matrix (character 11) because it is regarded as a key taxonomic feature between the two morphologically similar genera. The recovery of a monophyletic clade of species diagnosed as Plaesiomys, therefore, would support Williams and Harper’s (2000) separation of genera.

Three subspecies of Plaesiomys subquadratus were included in phylogenetic analysis: P. s. cutterensis Howe, 1966, P. s. idahoensis Ross, 1959, and P. s. occidentalis

Ladd, 1929. Each subspecies has well preserved type material and was coded separately 68 from Hall’s (1847) types of P. subquadratus to test whether phylogenetic analysis upholds subspecific diagnoses.

Material examined.—Specimens from the type series of each species were examined for character analysis whenever possible. The type series for most species examined consisted of multiple specimens, thereby providing a range of intraspecific morphological variation. In addition, non-type and figured specimens from museum collections and the literature were analyzed when available to more accurately assess the total the range of morphological variability within each species. Only data from specimens interpreted to be adult individuals were incorporated into the analysis. Small specimens interpreted as young individuals were excluded to remove ontogenetic variation from character state distributions.

Specimens were examined from the collections within the United States National

Museum of Natural History (USNM); American Museum of Natural History (AMNH);

University of Iowa Paleontology Repository (SUI), and the Yale Peabody Museum

(YPM).

Characters and coding.— A total of 21 characters comprising internal and external morphological traits were included (Table 10). Both internal and external morphological characters were utilized because they have previously been demonstrated to be useful for constructing phylogenetic hypotheses of fossil brachiopods (Leighton and

Maples, 2002; Stigall Rode, 2005; Wright and Stigall, in review) and form the basis of many specific diagnoses (Howe, 1966; Wang, 1949; Williams and Harper, 2002).

Table 10 Characters used in Phylogenetic Analysis of Plaesiomys

General characters

1. Maximum width. The maximum shell width parallel to the hinge line. (0), narrow (≤ 21.452); (1), wide (≥ 25.049).

2. Relative inflation of the dorsal valve. Recorded as the ratio of the maximum dorsal height, measured perpendicular to the commissural plane, divided by the maximum width, measured parallel to the hinge line. (0), high (≥ 0.269); (1), low (≤ 0.239).

3. Relative inflation of the ventral valve. Recorded as the ratio of the maximum ventral valve height, measured perpendicular to the commissural plane, divided by the maximum width, measured parallel to the hinge line. (0), low (≤ 0.117); (1), high (≥ 0.120).

4. Shell thickness. Recorded as the total height of the individual divided by the maximum width, measured parallel to the hinge line. (0), low (≤ 0.329); (1), (≥ 0.364).

5. Comparative convexity of the valves. Recorded as the ratio of the ventral valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane and the dorsal valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane. (0), high (≥ 0.414); (1), low (≤ 0.413).

6. Depth of the shell. Ratio of the maximum shell length, measured perpendicular to the hinge line, divided by the maximum shell width, measured parallel to the hinge line. (0), low (≤ 0.781); (1), high (≥ 0.813).

7. Depth of the sulcus. Recorded as the ratio of the sulcus, measured as the vertical distance between the commissure at the center of the sulcus and the commissure at the break in the slope at the side of the sulcus, divided by shell thickness of the brachiopod. (0), high (≥ 2.743); (1), low (≤ 2.048).

8. Origin of the sulcus. Recorded as the ratio of the distance, measured perpendicular to the hinge line, from the ventral umbo to the beginning of a perceptible concave deflection in the ventral valve divided by the maximum length of the valve, measured perpendicular to the hinge line. (0), distal (≥ 0.531); (1), proximal (≤ 0.494).

9. Outline shape. (0), subelliptical; (1) subquadrate.

Table 10 continued

10. Costal density. Recorded as the number of costae per 5 mm at the anterior margin of the valves, divided by the maximum shell width measured parallel to the hinge line. (0) low (≤ 0.267); (1) high (≥ 0.359).

11. Presence of secondary costae. (0), absent to rare; (1) abundant.

Ventral valve characters

12. Length of the ventral muscle scars. Recorded as the ratio of the maximum length of the ventral muscle field, measured perpendicular to the hinge line, divided by the maximum shell length, measured perpendicular to the hinge line. (0), short (≤ 0.473); (1) long (≥ 0.499).

13. Height of the ventral cardinal area. Recorded as the ratio of the vertical distance from the hinge line to the top of the ventral interarea, measured perpendicular to the hinge line, divided by the ventral valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane. (0), high (≥ 0.674); (1), low (≤ 0.662).

14. Angle of the ventral cardinal area. Recorded as the angle between the lateral boundaries of the ventral interarea. (0), high (≥ 150o); (1), low (≤ 145o).

15. Ventral umbonal angle. Recorded as the angle between the two limbs of the umbo when the dorsal valve is oriented parallel to the commissural plane. (0), high (≥ 155o); (1), low (≤ 150o).

16. Angle of the delthyrium. Recorded as the angle between the lateral edges of the delthyrial opening. (0), low (≤ 50o); (1) high (≥ 60o).

17. Posterior extension of the ventral umbo across the hinge line. Recorded as the ratio of the distance of extension of the ventral umbo, measured perpendicular to the hinge line, divided by the length of the hinge line. (0), high (≥ 0.063); (1), low (≤ 0.053).

18. Angle between the ventral interarea and the hinge line. Recorded as the angle at which the ventral interarea diverges from the commissural plane prior to incurving. (0), upright (≥ 75o); (1), inclined (≤ 70o).

Table 10 continued

Dorsal valve characters

19. Height of the dorsal cardinal area. Recorded as the ratio of the vertical distance from the hinge line to the top of the dorsal interarea, measured perpendicular to the hinge line, divided by the ventral valve height, measured as the maximum vertical height of the dorsal valve perpendicular to the commissural plane. (0), high (≥ 0.126); (1), low (≤ 0.049).

20. Dorsal umbonal angle. Recorded as the angle between the two limbs of the umbo when the dorsal valve is oriented parallel to the commissural plane. (0), high (≥ 150o); (1), low (≤ 145o).

21. Posterior extension of the dorsal umbo across the hinge line. Recorded as the ratio of the distance of extension of the dorsal umbo, measured perpendicular to the hinge line, divided by the length of the hinge line. (0), high (≥ 0.052); (1), low (≤ 0.049). ______

Both discrete and continuous characters were considered. Discrete characters are the most commonly used character type in phylogenetic analyses (Wiley and Lieberman,

2011). However, continuous characters have also been utilized previously to construct species-level phylogenetic hypotheses for clades of fossil invertebrates (Rode, 2004;

Stigall Rode, 2005; Hopkins, 2011; Wright and Stigall, in review). When normalized to form a ratio, a continuous morphometric character forms a quasi-continuous distribution whereby the range of normalized continuous variation may then be differentiated into distinct classes and coded as discrete phylogenetic character states.

Continuous characters were coded based on the total range of morphometric values measured for each species presented in Table 11. Measurements of continuous characters were separated into discrete states using a procedure similar to those suggested by Morton and Kincaid (1995) and Swiderski et al. (1998). Morphometric measurements 72 were visually inspected by examining univariate frequency distributions and box plots.

ANOVA. Potential states for all characters were found to be highly distinct (P < 0.001) with non-overlapping mean values within the 95% confidence interval. The expected range for each continuous character state was calculated as the mean ± 1 standard deviation (Table 12). If measurements for a species fell within a single expected range,

Table 11 Data Ranges for Continuous Characters of Species Analyzed in Phylogenetic Analysis of Plaesiomys

Table 12 Statistical Separation of Continuous Characters in Plaesiomys Analysis. S.D. Indicates Standard Deviation

Character Range Mean S.D.

1. Maximum width (0) small x ≤ 21.425 18.662 2.790 (1) large x ≥ 25.045 29.104 4.055

2. Inflation of the dorsal valve (0) low x ≤ 0.239 0.206 0.033 (1) high x ≥ 0.269 0.307 0.038

3. Inflation of the ventral valve (1) low x ≤ 0.117 0.095 0.022 (0) high x ≥ 0.120 0.149 0.029

4. Shell thickness (1) low x ≤ 0.329 0.269 0.060 (0) high x ≥ 0.364 0.416 0.052

5. Comparative convexity of the valves (1) low x ≤ 0.413 0.337 0.076 (0) high x ≥ 0.414 0.611 0.197

6. Depth of the shell (1) low x ≤ 0.781 0.756 0.025 (0) high x ≥ 0.813 0.838 0.025

7. Depth of the sulcus (0) shallow x ≤ 2.048 1.001 1.047 (1) deep x ≥ 2.743 3.864 1.121

8. Origin of the sulcus (1) proximal x ≤ 0.494 0.420 0.074

(0) distal x ≥ 0.531 0.581 0.050

Table 12 continued

Character Range Mean S.D.

10. Costal density (1) low x ≤ 0.267 0.220 0.047 (0) high x ≥ 0.359 0.507 0.148

12. Length of the ventral muscle scars (1) short x ≤ 0.473 0.436 0.037 (0) long x ≥ 0.499 0.533 0.034

13. Height of the ventral cardinal area (1) low x ≤ 0.662 0.573 0.089 (0) high x ≥ 0.674 0.872 0.198

17. Posterior extension of the ventral umbo across the hinge line (1) low x ≤ 0.053 0.046 0.007 (0) high x ≥ 0.063 0.074 0.011

19. Height of the dorsal cardinal area (1) low x ≤ 0.094 0.064 0.030 (0) high x ≥ 0.126 0.183 0.057

21. Posterior extension of the dorsal umbo across the hinge line (1) low x ≤ 0.049 0.042 0.007 (0) high x ≥ 0.052 0.065 0.013

76 only that state was coded; whereas if a species possessed measurements from multiple expected character states, it was coded as polymorphic. Species containing measurements outside of one expected state yet not in the other were coded as the primitive (0) state.

Character state distributions for species of Plaesiomys are presented in Table 13. All statistical procedures for separating continuous characters were performed using PAST

(Hammer et al., 2001). Examples of morphometric-based characters are shown in Figure

Search method.—Phylogenetic analysis was conducted using the criterion of maximum parsimony in PAUP* 4.0 (Swofford, 2002). A branch and bound search was used to determine the most parsimonious tree for the character matrix. All characters were treated as unweighted and unordered. Taxa exhibiting multiple character states were coded accordingly and treated as polymorphic. Characters were optimized using accelerated transformation (ACCTRAN) and analyzed in MacClade 4.06 (Maddison and Maddison,

2003).

Table 13 Character State Distribution for Species Included in Phylogenetic Analysis of Plaesiomys X = (0 and 1)

RESULTS

Phylogenetic analysis.—Parsimony analysis retrieved a single most parsimonious tree with a length of 92 steps (Figure 7). The consistency and retention indices are 0.696 and 0.576, respectively, which are significantly higher than those derived from similar sized matrices using randomly generated data at the α = 0.05 level (Klassen et al., 1998).

Confidence values for node support were recovered using 100 repetitions of a full- heuristic Jackknife analysis with 5% deletion. Jackknife values for nodes compatible with the single most parsimonious tree are presented in Figure 2. Further support was assessed by calculating the g1 statistic for 10,000 trees generated from the data matrix. The g1 statistic is -0.358, indicating strong phylogenetic signal within the data (P < 0.05: Hillis and Huelsenbeck, 1992).

Recognition of clades.—Species of Plaesiomys form a monophyletic group supported by multiple synapomorphies and strong Jackknife support. Characters 79 supporting the monophyly of the genus include a subquadrate outline of the shell, a low angle of the ventral umbo, and the presence of secondary costae (characters 9, 11, and

16). Species of Plaesiomys are distributed among three clades (Figure 7): a basal clade including P. cutterensis Howe, 1966, a clade including P. carltona (Twenhofel, 1928), and a clade including P. bellistriatus Wang, 1949.

The basal clade includes Plaesiomys cutterensis and P. idahoensis Ross, 1959, and their sister relationship is supported by low inflation of the dorsal valve, high costal density, and an inclined angle between the ventral interarea and hinge line (characters 2,

10, and 18). The sister relationship between the P. carltona and P. bellistriatus clades is supported by wide shell width, low inflation of the dorsal valve, high total shell height, and a ventral umbo located proximal to the hinge line (characters 1, 2, 4, and 17).

The monophyly of the clade comprising P. carltona, P. subquadratus (Hall,

1847), and P. anticostiensis (Shaler, 1865) is supported by a low height of the dorsal cardinal area and a proximal origination of the sulcus (characters 8 and 19). The sister relationship between P. anticostiensis and P. subquadratus is supported by an elevated height of the ventral valve cardinal area (character 13). The monophyly of the P. carltona clade and the sister relationship between P. subquadratus and P. anticostiensis are both strongly supported by Jackknife values 80

The pectinate clade comprising P. bellistriatus, P. belilamellosus Wang, 1949, P. subcircularis (Roy, 1941), P. occidentalis Howe, 1966, and P. proavitus (Winchell and

Schuchert, 1892) is characterized by a high shell depth with relatively low convexity and 81 a low depth of the sulcus (characters 5, 6, and 7) and is supported by Jackknife analysis.

The clade comprising P. belilamellosus and species crownward is supported by extensive character evidence including highly inflated ventral valves, subelliptical outline, high costal density, long ventral valve muscle scars, low height of the ventral valve cardinal area, low dorsal umbonal angle, and a high angle between the ventral valve interarea and hinge line (characters 5, 9, 10, 12, 14, 18, and 20).

The monophyletic relationship among P. subcircularis, P. occidentalis, and P. proavitus is supported by relatively narrow shells width, low ventral umbonal angles, and a ventral umbo significantly extending over the hinge line (characters 1, 15, and 17). The sister relationship between P. occidentalis and P. proavitus is supported by a proximal dorsal umbo (character 21).

DISCUSSION

North American species of Plaesiomys originated in the early Late Ordovician

(late Sandbian) yet underwent rapid diversification and subsequent dispersal throughout

Laurentia during the mid-to late Katian (Figure 8). Species of Plaesiomys most commonly occurred in intercratonic basins of the North American mid-Continent, yet they also occurred in marginal and mixed platforms and formed a minor component of the Late Ordovician Hiscobeccus Fauna of North America (Jin, 2001).

Species of Plaesiomys have been hypothesized to live with their beaks down and their commissure oriented vertically (Richards, 1972). However, flume tank experiments by Alexander (1984) suggested the dorsibiconvex profile and relatively small interareas 82 of typical Plaesiomys subquadratus is not by itself hydrodynamically stable. Species of

Plaesiomys seem to have overcome this limitation by increasing their musculature and facilities for pedicle attachment. For example, species within both the P. carltona and P. bellistriatus subclades are characterized by large, thick shells with a ventral interarea located proximal to the hinge line. Additionally, species within the P. bellistriatus subclade have relatively large and well-developed ventral muscle scars which may have provided further support for species living in high energy environments.

A diagnostic feature of Plaesiomys relates to its multicostellate exterior which may be related to deterring epibionts (Richards, 1972). Both the sister group P. cutterensis and P. idahoensis as well all species from P. bellilamelosus crownward exhibit high costal density, which may indicate an increased adaptation to deterring parasitism from epibionts (Richards, 1972).

Species of Plaesiomys occur primarily in strata interpreted to have been within normal to storm wave base (Holland and Patzkowsky, 1996) across multiple bathymetric gradients and tectonic settings (Holland and Patzkowsky, 2008). Several species, notably

P. bellistriatus, P. proavitus, and P. subquadratus, attained large geographic ranges encompassing several tectonic basins (Table 14). For example, P. bellistriatus occurs in the Richmondian (upper Katian) age Maquoketa, Viola, and Aleman Formations across the mid-Continent region of North America

As noted above, species of Plaesiomys share a high propensity for dispersal and broad ecological tolerances, both of which are characteristic of invasive species (Davis,

2009; Stigall, 2010b). Plaesiomys subquadratus participated in the Late Ordovician 83

Richmondian Invasion into the Cincinnati basin (Foerste, 1912; Stigall, 2010a). In the phylogenetic hypothesis presented in Figure 2, P. subquadratus is nested within a clade consisting of two species P. anticostiensis and P. carltona from Anticosti Island (Table

14). The phylogenetic topology presented herein suggests that P. subquadratus may have invaded the Cincinnati basin from the Anticosti Island region of Quebec (Wright and

Stigall, 2012). This pattern conflicts with the hypothesized pathway of invasion for the orthid brachiopod Glyptorthis insculpta (Hall, 1847), which likely emigrated from the mid-Continent into the Cincinnati basin (Wright and Stigall, in review). The lack of a unidirectional source for the Richmondian Invasion suggests multiple pathways of invasion were utilized to facilitate the influx of invasive taxa into the Cincinnati basin.

Further investigations are needed to assess whether or not invasion pathways for P. subquadratus or G. insculpta represent a general pattern for invader taxa (Chapter 4). 84

Table 14 Stratigraphic and Geographic Distribution of Plaesiomys † Indicates Gamachian (Hirnantian), All Other Species Are Richmondian (Katian)

SYSTEMATIC PALEONTOLOGY

Specimens examined for species that require neither taxonomic revision nor lectotype designation are presented in Table 15. Each specimen listed belongs to a species of Plaesiomys that is considered valid herein. The original descriptions for these species are sufficient for diagnosis, and additional discussion is, therefore, not warranted here. The character state distribution data from the phylogenetic analysis presented in this paper (see Table 13) may be combined with the original species descriptions to provide enhanced diagnoses.

Order ORTHIDA Schuchert and Cooper, 1932

Suborder ORTHIDINA Schuchert and Cooper, 1932

Superfamily ORTHOIDEA Woodward, 1852

Family PLAESIOMIDAE Schuchert and Cooper, 1913

Subfamily PLAESIOMINAE Schuchert, 1913

Genus PLAESIOMYS Hall and Clarke, 1892

Type species.—Orthis subquadrata Hall, 1847.

Other valid species.—Austinella subcircularis Roy, 1941; Dinorthis carltona

Twenhofel, 1928; Orthis anticostiensis Shaler, 1865; O. proavitus Winchell and

Schuchert, 1892, Plaesiomys belilamellosus Wang, 1949; P. bellistriatus, Wang, 1949; P. cutterensis Howe, 1966; P. idahoensis Ross, 1959; and P. occidentalis Ladd, 1929.

Potentially valid species.—Dinorthis rockymontana Wilson, 1926. 87

Diagnosis.—Shell moderate to large, outline subquadrate to suboval, convexoconcave to dorsibiconvex; multicostellate, costae emplaced by bifurcation or intercalation, aditicules on costae and costellae, uniplicate to rectimarginate, epipunctae microstructure; ventral interarea short, apsacline, dorsal interarea orthocline to anacline; ventral muscle field subcordate, diductors large, adductors oval shaped, elongate, and enclosed within diductor field, adjustors large with variable outline; dorsal adductors quadripartite, posterior scars typically larger than anterior scars, divided longitudinally by a medial myophragm; myophore bilobate to trilobate, crenulated; brachiophores divergent.

Occurrence.—Late Ordovician (late Sandbian to Hirnantian) of North America with possible occurrences in Kazakhstan, Siberia, and the British Isles (see discussion below).

Discussion.— Plaesiomys has most closely frequently been compared with

Dinorthis. The external morphology of Plaesiomys differs primarily from Dinorthis with respect to the development of costellae in Plaesiomys, whereas costellae is absent or rare in Dinorthis. Additionally, costae present in species of Dinorthis are typically much coarser than in Plaesiomys. Internally, the ventral adductor muscle scars form an anterior notch in Dinorthis but are enclosed within the diductors in Plaesiomys. The multicostellate plaesiomids Multicostellata Schuchert and Cooper, 1931 and

Campylorthis Ulrich and Cooper, 1942 also closely resemble Plaesiomys. Plaesiomys can be distinguished from Multicostellata by its convexoconcave profile and large adjustor muscles, whereas Multicostellata has subequally convex valves and small adjustor 88 muscles. Campylorthis differs from Plaesiomys in possessing a perforated deltidium and a well-developed chilidium.

Species of Plaesiomys have been reported from outside of North America including the British Isles (Cocks, 1978; Hiller, 1980; Harper, 1984; Mitchell 1977;

Reed, 1952; Wright, 1964); Kazakhstan (Popov et al. 2000), and Siberia (Kulkov and

Severgina, 1989). These species were not included in the present investigation because the taxonomic validity of Plaesiomys occurring outside North America has been previously questioned (Jin and Zhan, 2008; Jin et al. 2007) and assessing their taxonomic validity is beyond the scope of this analysis.

Table 15 Specimens Examined for Species that Do Not Require Taxonomic Revision or Lectotype Designation * Indicates Non-Type Specimens

Species Museum Catalog number Type

P. anticostiensis MCZ 14679 Lectotype MCZ 147680 Paralectotype GSC 117925 * GSC 117926 * GSC 117929 *

P. belilamellosus SUI 1804 Holotype SUI 1805 Paratype SUI 1806 Paratype

P. bellistriatus SUI 1820 Holotype USNM 146550 * USNM 14655 *

P. carltona GSC 2030k Holotype YPM 10356 Paratype YPM 11129 Paratype GSC uncataloged *

P. subcircularis AMNH 4797 Holotype AMNH 29423 Paratype

PLAESIOMYS CUTTERENSIS Howe, 1966

Plaesiomys subquadratus cutterensis HOWE, 1966, p. 249, pl. 29, figs. 7-12, 14.

Diagnosis.—Emended from Howe (1966). Shell large, wider than long, hinge line narrow; outline subquadrate, total shell inflation low; costal density high , sulcus originating proximally; ventral interarea and delthryium oriented at a high angles to the hinge line; ventral muscle scars short to long; dorsal valve highly inflated.

Types.— USNM 146558 (holotype); USNM 146559a, USNM 1465591

(paratypes).

Occurrence.— Cutter Limestone, Montoya Group, Texas; Richmondian (upper

Katian).

Discussion.— Plaesiomys cutterensis was originally described by Howe (1966) as a subspecies of Plaesiomys subquadratus. Howe (1966) differentiated specimens of P. s. cutterensis from P. subquadratus by their considerably fewer costae. Although P. cutterensis typically has fewer costae, this was not uncovered in the coding of costal density (character 10) because the values herein were normalized to maximum shell width and specimens of P. cutterensis are usually smaller than P. subquadratus.

Character analysis reveals P. cutterensis is not a valid subspecies of P. subquadratus but rather a distinct species differentiated by a more distal origination of the sulcus, a more strongly inclined angle between the ventral interarea and hinge line, and a relatively lower angles of the ventral cardinal area and umbo. Therefore, P. cutterensis is herein elevated to specific status.

PLAESIOMYS IDAHOENSIS Ross, 1959

Plaesiomys subquadratus idahoensis ROSS, 1959, p. 449, pl. 54, figs. 29, 30, 33, 34, 37,

38, 41, 42, pl. 55, figs. 22, 26, 27, 30, 31.

Diagnosis.—Emended from Ross (1959). Shell of small to moderate size, outline subquadrate to subelliptical, highly convex valves of variable depth; costal density high relative to maximum width; ventral valve shallow to inflated, delthryium oriented at a high angle; ventral interior muscle field long; dorsal valve inflated with shallow to deep sulcus; dorsal interarea orthocline to apsacline; high dorsal umbonal angle.

Types.—USNM 133239 (holotype); USNM 133240, USNM 133241, USNM

133242, and USNM 133243 (paratypes).

Occurrence.— Saturday Mountain Formation, Idaho; Richmondian (upper

Katian).

Discussion.— P. idahoensis was considered a subspecies of P. subquadratus in

Ross’ (1959) original description on the basis that P. idahoensis is smaller with finer costae and a wider hinge than P. subquadratus. Phylogenetic analysis reveals P. idahoensis to be a distinct species more closely related to P. cutterensis than P. subquadratus, and it is correspondingly elevated to species status herein. As noted by

Howe (1966), P. idahoensis can be distinguished from P. cutterensis by its smaller size, lower height of the ventral cardinal area, and a more obtuse interarea and umbonal angles on the ventral valve.

PLAESIOMYS OCCIDENTALIS Ladd, 1929 (non Okultich, 1943)

Dinorthis (Plaesiomys) subquadrata occidentalis LADD, 1929, p. 402, pl. 5, figs. 7-9.

Plaesiomys subquadratus occidentalis (Ladd, 1929) WANG, 1949, p. 5, pl. 2d, figs. 1-5.

Diagnosis.—Emended from Ladd (1929). Shell small, total height large for genus; outline subquadrate, cardinal extremities rounded; costal density high; sulcus depth low, delthryium inclined; ventral interarea highly obtuse; ventral umbo extending significantly over the hinge line; dorsal umbo proximal, opening into a wide (≥150o) angle.

Types.—SUI 66504 (holotype).

Additional material.— USNM 146553a-b, SUI 1807 (hypotypes).

Occurrence.— Maquoketa Shale, Iowa; Richmondian (upper Katian).

Discussion.— Ladd (1929) described P. occidentalis as a subspecies of P. subquadratus differing only with respect to the former having more rounded cardinal areas. Wang (1949) suggested that P. occidentalis more closely resembled P. bellilamellosus, and Howe (1966) considered the types of P. occidentalis and P. bellilamellosus to represent end members of a continuous spectrum between the two forms. Neither Wang (1949) nor Howe (1966), however; removed P. subquadratus occidentalis from its subspecific taxonomic rank. The topology recovered from phylogenetic analysis indicates P. occidentalis is more closely related to P. bellilamellosus than P. subquadratus, yet P. occidentalis and P. bellilamellosus are not sister species and differ by six characters. Plaesiomys occidentalis is herein diagnosed as a distinct species. It differs from P. bellilamellosus by a smaller shell, low sulcus depth, 93 subquadrate outline, low ventral umbonal angle, and having both the ventral and dorsal umbo extending significantly over the hinge line.

PLAESIOMYS PROAVITUS (Winchell and Schuchert, 1892)

Figures 9.5-9.6

Orthis proavita WINCHELL AND SCHUCHERT, 1892

Orthis petrae (Winchell and Schuchert, 1892) SARDESON, 1892, pl. 5, figs. 18-19.

Orthis (Dinorthis) proavita (Winchell and Schuchert, 1892) BASSLER, 1932, pl. 25, figs. 16-17.

Dinorthis proavitus (Winchell and Schuchert, 1892) GREGER AND BORN, 1936, pl. 1, figs. 4-6.

Plaesiomys proavita (Winchell and Schuchert, 1892) WANG, 1949, p. 4, pl. 2, figs. e1-6.

Plaesiomys planus (Winchell and Schuchert, 1892) WANG, 1949, p. 6, pl. 3, figs. c1-5.

Plaesiomys proavitus (Winchell and Schuchert, 1892) ALBERSTADT, 1973, p. 24, pl. 2, figs. 1-5.

Diagnosis.—Emended from Winchell and Schuchert (1892). Shell small, subelliptical; low total shell height valve convexity; costal density low, costae somewhat coarse; sulcus depth low; ventral interarea and umbonal angle acute for genus; high delthyrial angle; ventral muscle scars long.

Types.—Winchell and Schuchert (1892) did not designate a holotype. YPM S-

703a is designated as the lectotype herein, and YPM S-703b (13 specimens) and YPM S-

709a become paralectotypes. 94

Additional material.— Following the synonymy listed in Wang (1949), the type material of Hebertella clermontensis Bradley, 1921 [MCZ 110301 (holotype); MCZ

110549, MCZ 110550 (paratypes)] was examined and determined to belong to P. proavitus.

Occurrence.— Maquoketa Formation, Iowa, Illinois; Viola Formation,

Oklahoma; Cape Limestone, Missouri, Fernvale Limestone, Tennessee; Richmondian

(upper Katian).

Discussion.— This geographically widespread species is often confused in collections with Dinorthis occidentalis Okultich, 1943. The overall low costal density and fewer secondary costae of P. proavitus is unusual for a species of Plaesiomys, yet secondary costae are present and in greater abundance than typically seen in Dinorthis

Furthermore, costae originate on each valve as in other species of Plaesiomys.

Phylogenetic analysis confirms that the morphological resemblance to Dinorthis occidentalis is superficial because P. proavitus is placed well outside the Dinorthis clade at a derived position within Plaesiomys.

Although Winchell and Schuchert (1892) originally designated this species as

Orthis proavita, the masculine suffix of Plaesiomys proavitus properly facilitates gender agreement (International Code of Zoological Nomenclature 34.2).

PLAESIOMYS SUBQUADRATUS (Hall, 1847)

Figures 9.1-9.4

Orthis subquadrata HALL, 1847, p. 126, pl. 32a, figs. 1a-o.; MEEK, 1873, p. 94, pl. 9, figs. 2b-g.

Plaesiomys subquadrata (Hall, 1847) HALL AND CLARKE, 1892, p. 196, pl. 5a, figs.

17-19; COOPER, 1944, p. 298, pl. 111, figs. 54-58; DAVIS, 1985, pl. 8, figs. 5-7;

SCHWIMMER AND SANDY, 1996, pl., 16.2, figs. 11-13; WILLIAMS AND HARPER,

2000, p. 747, pl. 540, figs. 1a-d.

Orthis (Dinorthis) subquadrata (Hall, 1847) WINCHELL AND SCHUCHERT, 1893, p.

428, pl. 32, figs. 46-50. 96

Plaesiomys (Dinorthis) subquadrata (Hall, 1847) SCHUCHERT AND COOPER, 1932, pl. 10, figs. 15, 17, 18, 24-26

Plaesiomys (Plaesiomys) subquadrata (Hall, 1847) WILLIAMS AND WRIGHT, 1965; p. H319, pl. 201, figs. 5a-3.

Plaesiomys subquadratus (Hall, 1847) ROSS, 1959, p. 449, pl. 55, figs. 1, 6, 12, 14, 15,

18, 19, 23; ALBERSTADT , 1973, p. 22, pl. 2, figs. 6-8; JIN ET AL., 1997, pl. 1, figs. 1-

4; JIN AND ZHAN, 2008, pl. 4, figs. 14-20.

Plaesiomys aff. subquadratus (Hall, 1847) JIN AND NORFORD, 1996, p. 25, pl. 2, figs.

1-7.

Plaesiomys cf. subquadratus (Hall, 1847) HOWE, 1966, p. 247, pl. 29, figs. 1-6;

ALBERSTADT, 1973, p. 21, pl. 7, figs. 8a-b.

Diagnosis.—Emended from Hall (1847). Shell large, wider than long; outline subquadrate; costal density high; valve convexity and inflation variable, typically convexoconcave to dorsibiconvex; sulcus originates proximally; ventral muscle scars moderate to long, adjustors variable; ventral and dorsal umbonal angles low.

Types.— AMNH 20361-30267 are syntypes of Hall’s (1847) original description; however, Hall (1847) did not designate a holotype. AMNH 30263 is therefore designated as a lectotype and the remaining specimens become paralectotypes.

Additional material.—USNM 14655a-f, YPM S-697, YPM-728c-e.

Occurrence.— Waynesville, Liberty, and Whitewater Formations, Ohio and

Indiana; Aleman Limestone, Texas; Viola Formation, Oklahoma; Avalanche Formation,

British Columbia; Caution Creek Formation, Manitoba; Richmondian (upper Katian). 97

Discussion.— Although several subspecies have been previously assigned to P. subquadratus, the phylogenetic revision here failed to uphold their subspecific status (see above). Consequently all former subspecies have been elevated herein to the rank of species. Character analysis reveals considerable morphological variation within P. subquadratus regarding relative valve inflations, convexity between the valves, total shell height, depth of the shell, depth of the sulcus, length of the ventral muscle scars, and the height of the dorsal cardinal area. Occurrences of P. subquadratus are geographically widespread throughout the Laurentian paleocontinent and range throughout the

Richmondian (upper Katian). Given the widespread occurrences, substantial morphological variation may be attributed to differential population dynamics operating within a geographically widespread species.

Although Hall (1847) originally designated this species as Orthis subquadrata,

Ross (1959) corrected the specific epithet to Plaesiomys subquadratus to facilitate gender agreement (ICZN 34.2).

REVISION OF HEBERTELLA

Taxonomic background.—The plectorthidine brachiopod genus Hebertella Hall and Clarke, 1892 occurs in Whiterockian (lower Sandbian) to Gamachian (Hirnantian) strata of North America. Species of Hebertella reached maximal diversity and abundance throughout the Cincinnatian Series (upper Sandbian to Katian) of North America and became significant constituents of the largely endemic Richmondian brachiopod fauna of

Laurentia (Cocks and Torsvik, 2011; Schuchert and Cooper, 1932). Although several 98 taxonomic revisions of Hebertella have been conducted (i.e., Schuchert and Cooper,

1932, Cooper, 1956, Walker, 1982), these have focused primarily on amending specific designations or diagnoses (Cooper, 1956; Walker, 1982). No prior study has utilized phylogenetic analysis to inform a comprehensive taxonomic revision or assess evolutionary relationships among species of Hebertella.

Hall and Clarke’s (1892) original conception of Hebertella included morphologically diverse set of species. Subsequently, many of these species were recognized as type species of distinct genera including Glyptorthis Foerste 1914a,

Eridorthis Foerste, 1909a, Austinella Foerste, 1909a, and Mimella Cooper, 1930, whereas other species were transferred to existing genera, such as Plectorthis Hall and Clarke,

1892. Schuchert and Cooper (1932) summarized these revisions and provided a less inclusive diagnosis in which they recognized species of Hebertella on the basis of a convexoconcave to unequally biconvex multicostellate exterior with an obcordate ventral muscle field surrounding a double-adductor ridge. Further, Cooper (1956) followed

Schuchert and Cooper’s (1932) revised diagnosis of Hebertella and transferred five species originally assigned to Hebertella to Mimella. Walker (1982) provided a partial taxonomic revision of Hebertella and synonymized the type species H. sinuata (Hall,

1847) with H. occidentalis (Hall, 1847). However, Walker’s (1982) revision of North

American species of Hebertella is incomplete because his investigation was geographically limited to Kentucky.

This analysis of Hebertella provides a comprehensive phylogenetically-informed taxonomic revision by assessing species validity, emending taxonomic designations, and 99 documenting evolutionary relationships among species distributed throughout the

Laurentian paleocontinent.

MATERIALS AND METHODS

Taxa analyzed.—Specimens assigned to ten North American species of

Hebertella were examined for phylogenetic analysis (Table 16). Doleroides tennesseensis

Cooper, 1956 was chosen as the outgroup for character polarization. This species were chosen because the type specimens were well preserved and representative of Doleroides

Schuchert and Cooper, 1930, a plectorthidine genus and potential sister taxon of

Hebertella (Schuchert and Cooper, 1932; Williams and Wright, 1965).

Because Walker (1982) questioned whether H. subjugata (Hall, 1982) was a distinct species from H. occidentalis, data for H. subjugata was coded separately to test the taxonomic validity of H. subjugata. In addition, three taxa previously assigned subspecific ranks were included in phylogenetic analysis: H. alveata richmondensis

Foerste, 1909a, H. sinuata prestonensis Ladd, 1929 and H. occidentalis montoyensis

Howe, 1966. Each subspecies has well preserved type material and was coded separately to test whether phylogenetic analysis upholds subspecific diagnoses.

Material examined.—Specimens from the type series of each species were examined for character analysis whenever possible. The type series for most species comprised multiple specimens, thereby providing range of intraspecific morphological variation.

Additional non-type and figured specimens from museum collections and the literature were analyzed when available to more accurately assess the total the range of 100 morphological variability within each species. Only data from specimens interpreted to be adult individuals were incorporated into the analysis. Schuchert and Cooper (1932) noted that small, articulated specimens of Hebertella are difficult to distinguish from

Plectorthis. Small specimens interpreted as young individuals were, therefore, excluded to remove ontogenetic variation or incorrect generic diagnoses from impacting character state distributions.

Table 16 Species Included In Phylogenetic Analysis of Hebertella * Indicates Outgroup

Doleroides tennesseensis* Cooper, 1956 Hebertella alveata Foerste, 1909a Hebertella bursa Raymond, 1928 Hebertella frankfortensis Foerste, 1909b Hebertella montoyensisHowe, 1966 Hebertella occidentalis(Hall, 1847) Hebertella maria (Billings, 1862) Hebertella parksensisFoerste, 1909b Hebertella prestonensisLadd, 1929 Hebertella richmondensis Foerste, 1909a Hebertella subjugata (Hall, 1847)

Specimens were examined from the following museum collections: United States

National Museum of Natural History (USNM); American Museum of Natural History

(AMNH); University of Iowa Paleontology Repository (SUI), Academy of Natural

Sciences of Philadelphia (ANSP), and the Yale Peabody Museum (YPM).

Characters and coding.— A total of 18 characters comprising internal and external morphological traits were analyzed (Table 17). Both internal and external 101 morphological characters were utilized, although emphasis was placed on external characters because they have been diagnostic for distinguishing among species of

Hebertella (Walker, 1982).

Table 17 Characters Used in Phylogenetic Analyses of Plaesiomys

General characters

1. Maximum width. The maximum shell width parallel to the hinge line. (0), narrow (≤ 24.618); (1), wide (≥ 27.128).

2. Inflation of the dorsal valve. Recorded as the ratio of the maximum dorsal height, measured perpendicular to the commissural plane, divided by the maximum width, measured parallel to the hinge line. (0), low (≤ 0.311); (1), high (≥ 0.319).

3. Inflation of the ventral valve. Recorded as the ratio of the maximum ventral valve height, measured perpendicular to the commissural plane, divided by the maximum width, measured parallel to the hinge line. (0), low (≤ 0.175); (1), high (≥ 0.208).

4. Comparative convexity of the valves. Recorded as the ratio of the ventral valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane and the dorsal valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane. (0), high (≥ 0.631); (1), low (≤ 0.567).

5. Depth of the shell. Ratio of the maximum shell length, measured perpendicular to the hinge line, divided by the maximum shell width, measured parallel to the hinge line. (0), high (≥ 0.770); low (≤ 0.742).

6. Depth of the sulcus. Recorded as the ratio of the sulcus, measured as the vertical distance between the commissure at the center of the sulcus and the commissure at the break in the slope at the side of the sulcus, divided by shell thickness of the brachiopod. (0), shallow (≤ 7.900); (1) deep (≥ 10.490).

7. Width of the sulcus. Recorded as the ratio of the distance, measured parallel to the hinge line, between the inflection points on the sulcus, divided by the maximum width of the shell. (0), wide (≥ 0.447); (1), narrow (≤ 0.435).

102

Table 17 continued

8. Number of costae in the sulcus. Recorded as the total count of costae between the slope breaks at both sides of the sulcus measured parallel to the commissural plane. (0), many (≥ 17); (1), few (≤ 15).

9. Outline shape. (0), subelliptical; (1) subquadrate.

10. Termination of cardinal extremities. (0), rounded; (1) angular.

Ventral valve

11. Maximum width of the ventral muscle scars. Recorded as the ratio of the maximum width of the ventral muscle field, measured parallel to the hinge line, divided by the maximum valve width, measured parallel to the hinge line. (0), narrow (≤ 0.251); (1), wide (≥ 0.300).

12. Height of the ventral interarea. Recorded as the ratio of the vertical distance from the hinge line to the top of the ventral interarea, measured perpendicular to the hinge line, divided by the ventral valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane. (0), high (≥ 0.779); (1) low (≤ 0.661).

13. Ventral umbonal angle. Recorded as the angle between the two limbs of the umbo when the dorsal valve is oriented parallel to the commissural plane. (0), low (≤ 135o); (1), high (≥ 140o).

14. Angle of the delthyrium. Recorded as the angle between the lateral edges of the delthyrial opening. (0), (≤ 35o); (1), high (≥ 40o).

15. Angle between the ventral interarea and the hinge line. Recorded as the angle at which the ventral interarea diverges from the commissural plane pior to incurving. (0), upright (≥ 90o); (1), inclined (≤ 85o).

Dorsal valve

16. Height of the dorsal cardinal area. Recorded as the ratio of the vertical distance from the hinge line to the top of the dorsal interarea, measured perpendicular to the hinge line, divided by the dorsal valve height, measured as the maximum vertical height of the ventral valve perpendicular to the commissural plane. (0), low (≤ 0.104); (1), high (≥ 0.129).

103

Table 17 continued

17. Dorsal umbonal angle. Recorded as the angle between the two limbs of the umbo when the dorsal valve is oriented parallel to the commissural plane. (0), low (≤ 135o); (1), high (≥ 140o).

18. Posterior extension of the dorsal umbo across the hinge line. Recorded as the ratio of the distance of extension of the dorsal umbo, measured perpendicular to the hinge line, divided by the length of the hinge line. (0), high (≥ 0.052); (1), low (≤ 0.049). ______

Both discrete and continuous characters were considered for analysis, as in the phylogenetic analysis of Plaesiomys described above. Similarly, continuous morphometric characters forming quasi-continuous distributions were differentiated into distinct classes and coded as discrete phylogenetic character states as described above

(Table 18). Continuous characters were coded based on the total range of morphometric values measured for each species presented in Table 19. Character state distributions for species of Hebertella are presented in Table 20.

Search method.—Phylogenetic analysis was conducted using maximum parsimony in

PAUP* 4.0 (Swofford, 2002). An exhaustive search was used to determine the most parsimonious tree for the character matrix. All characters were treated as unweighted and unordered. Taxa exhibiting multiple character states were coded accordingly and treated as polymorphic. Characters were optimized using accelerated transformation

(ACCTRAN) and analyzed in MacClade 4.06 (Maddison and Maddison, 2003). 104

Table 18 Statistical Separation of Continuous Characters in Hebertella Analysis. S.D. Indicates Standard Deviation

Character Range Mean S.D.

1. Maximum width (0) small x ≤ 24.618 21.220 33.413 (1) large x ≥ 27.128 33.413 6.285

2. Inflation of the dorsal valve (0) low x ≤ 0.311 0.294 0.017 (1) high x ≥ 0.319 0.356 0.037

3. Inflation of the ventral valve (1) low x ≤ 0.175 0.149 0.025 (0) high x ≥ 0.208 0.243 0.035

4. Comparative convexity of the valves (1) low x ≤ 0.567 0.473 0.094 (0) high x ≥ 0.631 0.740 0.109

5. Depth of the shell (1) low x ≤ 0.742 0.698 0.044 (0) high x ≥ 0.770 0.807 0.037

6. Depth of the sulcus (0) shallow x ≤ 7.900 5.630 2.060 (1) deep x ≥ 10.490 14.200 3.710

7. Width of the sulcus (1) narrow x ≤ 0.435 0.334 0.101 (0) wide x ≥ 0.447 0.527 0.080

11. Width of the ventral muscle scars (0) narrow x ≤ 0.251 0.236 0.015 (1) wide x ≥ 0.300 0.343 0.043

105

Table 18 continued

Character Range Mean S.D.

12. Height of the ventral cardinal area (1) low x ≤ 0.661 0.541 0.120 (0) high x ≥ 0.779 1.110 0.331

16. Height of the dorsal cardinal area (0) low x ≤ 0.104 0.070 0.034 (1) high x ≥ 0.129 0.169 0.040

18. Posterior extension of the dorsal umbo across the hinge line (1) low x ≤ 0.049 0.039 0.010 (0) high x ≥ 0.052 0.072 0.020

Table 19 Data ranges of morphometric characters for species included in phylogenetic analysis of Hebertella

Table 20 Character State Distribution for Species Included in Phylogenetic Analysis of Hebertella X = (0 and 1)

RESULTS

Phylogenetic analysis.—The exhaustive search retrieved a single most parsimonious tree of 67 steps (Figure 10). The consistency and retention indices are

0.731 and 0.591, respectively. These values are significantly higher than those derived from similar sized matrices using randomly generated data at the α = 0.05 level (Klassen et al., 1998). Confidence values for node support were recovered using 100 repetitions of a full-heuristic Jackknife analysis with 5% deletion. Jackknife values for nodes compatible with the single most parsimonious tree are presented in Figure 10. Support was further assessed by calculating the g1 statistic for 10,000 trees generated from the data matrix. The g1 statistic is -0.603, which indicates strong phylogenetic signal within the data (P < 0.01: Hillis and Huelsenbeck, 1992).

Recognition of clades.—The overall tree topology is balanced and is supported by multiple synapomorphies and very strong (> 90%) Jackknife values (Figure 10). 108

Characters supporting the monophyly of Hebertella include a high inflation of the dorsal valve, a narrow sulcus width, and proximal dorsal umbo with respect to the hinge line

(characters 2, 7, and 18).

109

The clade consisting of H. alveata Foerste, 1909a, H. maria (Billings, 1862), H. richmondensis Foerste, 1909a, and H. bursa Raymond, 1928 is characterized by having a relatively large dorsal cardinal area (character 16). The sister relationship between H. alveata and H. maria is supported by low shell depth, few (≤ 15) costae in the sulcus, and a high (> 140o) dorsal umbonal angle (characters 5, 8, and 17); whereas the sister relationship between H. richmondensis and H. bursa is supported by a low height of the ventral interarea and a high (> 140o) umbonal angle (characters 12 and 13).

The clade comprising the remaining species is strongly supported by Jackknife values and is characterized by an inflated ventral valve, low comparative convexity between the valves, and a deep sulcus (characters 3, 4, and 6). This clade can be subdivided into two subclades of three species each. The first subclade includes H. parksensis Foerste, 1909b, H. frankfortensis Foerste, 1909a, and H. subjugata (Hall,

1847), whereas the second subclade consists of H. occidentalis (Hall, 1847), H. montoyensis Howe, 1966, and H. prestonensis Ladd, 1929.

Characters supporting the monophyly of the H. parksensis subclade include small

(for the genus) maximum shell width and a low height of the ventral interarea (characters

1 and 12). Hebertella frankfortensis and H. subjugata form a sister group supported by a high (> 40o) delthyrial angle relative to the commissural plane and an upright (≥ 90o) angle between the ventral interarea and the hinge line (characters 14 and 15).

The monophyly of the H. occidentalis subclade is supported by 100% Jackknife support based on the synapomorphies of a subquadrate shell outline and angular cardinal 110 extremities (characters 9 and 10). The sister relationship between H. montoyensis and H. prestonensis is supported by a wide distance across the sulcus and wide ventral muscle scars (characters 7 and 11).

DISCUSSION

Hebertella originated in North America during the Whiterockian (lower

Sandbian) and persisted through the Gamachian (Hirnantian). Species of Hebertella reached maximal abundance during the Late Ordovician and rank among the most common and conspicuous fossils of shallow subtidal deposits in the type-Cincinnatian

Series of eastern North America (Meyer and Davis, 2009; Stigall, 2010a). Although seven of the ten species included in this analysis occur within the Cincinnati Basin of

Kentucky, Indiana, and Ohio, the geographic range of Hebertella extended beyond confines of eastern North America into both marginal and interior cratonic basins

(Billings, 1862; Howe, 1966; Ladd, 1929) (Table 21). Interbasinal range expansion in

Hebertella occurred during the late Cincinnatian concomitant with a broader pattern of global brachiopod generic endemism whereby the Richmondian brachiopod fauna became broadly yet uniquely established in Laurentia (Cocks and Torsvik, 2011).

Species of Hebertella occur primarily in rocks interpreted to have been deposited within normal to storm wave base (Holland and Patzkowsky, 1996). Paleoecological reconstructions of H. occidentalis suggest a lifestyle with their commissure oriented vertically and pedicle foramen flush to the substrate, which likely represented the general life habit for all species of Hebertella (Alexander, 1984; Richards, 1972). In addition, 111 shells of Hebertella generally have large, catacline to slightly apsacline ventral interareas which provide greater shell stability (Alexander, 1984). A wide sulcus is characteristic for the genus, which may have aided the feeding efficiency of the lophophore by enhancing the passive flow of nutrients (LaBarbera, 1981).

Many species of Hebertella exhibit adaptations to high energy environments.

Both H. alveata and the species within the H. occidentalis subclade are characterized by having angular cardinal extremities. The increased length of the hinge provided a larger area of contact between the shell and the substrate and may represent an adaptation to high energy conditions (Richards, 1972). This feature is best developed in the highly alate shells of H. alveata from the Whitewater Formation in Indiana, a unit which was deposited within fair weather wave base (Meyer and Davis, 2009).

Species of Hebertella commonly occur in association with invasive taxa associated with the Richmondian Invasion, yet did not participate in the invasion process.

Theoretical and empirical studies on invasive species indicate that invaders may exhibit ecological dominance over incumbents and may therefore threaten biodiversity (Davis;

2009; Rode and Lieberman, 2004; Stigall, 2010b). However, the influx of invaders did not drive species of Hebertella extinct nor facilitate local extirpation (Malizia and Stigall,

2011; Stigall, 2010a). Instead, incumbent Cincinnatian species persist and H. alveata originated during the invasion interval. Paleobiogeographic analyses of faunal dynamics across the invasion interval suggest species of Hebertella were able to successfully

persist because they exhibited broad ecological tolerances facilitating large intrabasinal geographic ranges which may act as a buffer against biodiversity crises (Malizia and

Stigall, 2011; Rode and Lieberman, 2004).

SYSTEMATIC PALEONTOLOGY

Specimens examined for species that require neither of taxonomic revision nor lectotype designation are presented in Table 22. Each specimen listed belongs to a species of Hebertella that is considered valid herein. The original descriptions for these species are sufficient for diagnosis, and no additional discussion is required here. The character state distribution data from the phylogenetic analysis presented in this paper

(see Table 20) may be combined with the original species descriptions to provide enhanced diagnoses.

Order ORTHIDA Schuchert and Cooper, 1932

Suborder ORTHIDINA Schuchert and Cooper, 1932

Family PLECTORTHIDAE Schuchert and LeVene, 1929

Genus HEBERTELLA Hall and Clarke, 1892

Type species.—Orthis occidentalis Hall, 1847. Hall and Clarke (1892) originally designated Orthis sinuata Hall, 1847 as the type species of Hebertella. However, Walker

(1982) synonymized O. sinuata with O. occidentalis Hall, 1847. Orthis occidentalis has page priority in Hall’s (1847) original descriptions and is therefore the type species. 114

Other valid species.—Hebertella alveata Foerste, 1909a; H. bursa Raymond,

1928; H. frankfortensis Foerste, 1909b; H. montoyensis Howe, 1966; H. parksensis

Foerste, 1909b; H. prestonensis Ladd, 1929; H. richmondensis Foerste, 1909a; Orthis maria Billings, 1862; O. subjugata Hall, 1847.

Diagnosis.—Shell typically large, outline subquadrate, juvenile or smaller shells semielliptical, cardinal extremities rounded to alate, unequally biconvex to convexoconcave; multicostellate with aditicules; sulcus median, typically wide; large ventral interarea, catacline to apsacline; ventral muscle scars obcordate with thickened margins; adductor tracks raised on a double median ridge; pallial markings rare to absent; dorsal cardinal area long, apsacline; brachiophore plates convergent; dorsal muscle scars quadripartite, weakly impressed, with posterior adductor scars larger than anterior.

Occurrence.—Whiterockian (lower Sandbian) to Richmondian (upper Katian) of

North America, with possible occurrences in Estonia, Kazakhstan, Tasmania, and the

British Isles (see below).

Discussion.— Hebertella most closely resembles Mimella with respect to the exterior morphology of the shell, yet Mimella typically has a more convex ventral valve.

Interiorly, species of Hebertella may be easily distinguished from Mimella based on ventral musculature. The ventral muscle scars in Mimella are characterized by deep diductor scars separated by an elongate and anteriorly expanded adductor trackway with well developed pallial markings; conversely, in Hebertella the ventral adductor scars do not expand anteriorly and pallial markings are typically absent. Because small specimens of Hebertella may not be easily distinguished from Plectorthis, only large shells 115 representing adult specimens should be considered useful for stratigraphic or paleobiologic inferences pending a taxonomic revision of Plectorthis.

Schuchert and Cooper (1932) noted that species of Hebertella are common in

North America yet rare to unknown from Europe. Since their 1932 publication, possible species of Hebertella have been reported from outside North America including Estonia

(Rõõmusoks, 1964), Kazakhstan (Popov et al, 2000), Tasmania (Laurie, 1991), and the

British Isles (Reed, 1952; Wright, 1964). However, these species were not included in the present investigation because the taxonomic validity of Hebertella occurring outside

North America needs to be reinvestigated (Jin and Zhan, 2008). In addition, many non-

North American reports of Hebertella have been named on incomplete or poorly preserved specimens (e.g., Laurie, 1991; Popov et al, 2000; Reed, 1952; Wright, 1964), therefore, assessing their taxonomic validity is beyond the scope of this analysis.

Table 22 Specimens Examined for Species that Do Not Require Taxonomic Revision or Lectotype Designation. * Indicates Non-Type Specimens

Species Museum Catalog number Type

H. bursa MCZ 110282 Holotype

H. maria GSC 2271a Lectoype GSC 2271b Paralectoype GSC 2271c Paralectoype YPM 157037 * YPM 161230 * YPM 516751 *

116

HEBERTELLA ALVEATA Foerste, 1909a

Orthis occidentalis (Foerste, 1909a) MEEK, 1873, p. 96, pl. 9, fig. 3.

Hebertella alveata FOERSTE, 1909a, p. 255, pl. 30, figs. 15-23; SCHUCHERT AND

COOPER, 1932, p. 60; WALKER, 1982, p. M3.

Diagnosis.—Emended from Foerste (1909a). Shell large, widest along the hinge line, outline semielliptical to subquadrate; cardinal extremities angular, somewhat alate; sulcus width narrow, shallow depth; dorsal valve variably inflated, with a broad medial depression extending from the beak to the commissure; dorsal umbo located proximal to the hinge line, not extending significantly beyond the hinge line.

Types.— Foerste (1909a) described and figured two specimens yet did not designate a holotype. USNM 14679 is herein designated as the lectotype and USNM

78452 becomes a paralectotype.

Occurrence.— Liberty and Whitewater Formations, Ohio and Indiana;

Richmondian (upper Katian).

Discussion.— Hebertella alveata occurs most commonly alongside H. occidentalis and H. richmondensis in the Liberty Formation, yet is easily distinguished by its biconvex profile, shell widest at the hinge line, angular to alate cardinal extremities, and a medial depression on the dorsal valve extending to the commissure, and having the maximum width of the shell at the hinge line.

117

HEBERTELLA FRANKFORTENSIS Foerste, 1909b

Orthis frankfortensis (Foerste, 1909b) JAMES, 1871, p. 10, (nom. nud.)

Hebertella frankfortensis FOERSTE, 1909b, p. 318-319, pl. 7, figs. 11a-b;

SCHUCHERT AND COOPER, 1932, p. 60, pl. 11, fig. 21; WALKER, 1982, p. M3-M6, pl. 4, figs 18-40.

Diagnosis.—Emended from Foerste (1909b). Shell small, outline semielliptical; costae coarse, bifurcation rare; cardinal extremities rounded; dorsal and ventral valves inflated, evenly biconvex; sulcus narrow to wide, shallow, with relatively few costae (≤

15) between inflection points; ventral interarea low, catacline; delthyrium opening at a wide angle (≥ 40o); dorsal interarea low, apsacline.

Types.— Because Foerste (1909b) did not designate a holotype in his original description, USNM 258472 is herein designated the lectotype.

Other material examined.—USNM 258466, USNM 258276, USNM 258474, and

USNM 248477.

Occurrence.— Bigby-Cannon and Catheys Formations, Kentucky and Tennessee;

Mohawkian (upper Sandbian to lower Katian).

Discussion.— Hebertella frankfortensis most closely resembles its sister species

H. subjugata. Both are characterized by having small shells, a semielliptical outline, and rounded cardinal extremities, yet H. frankfortensis is readily distinguished from H. subjugata by its slightly more rounded shell, relatively coarse costae, rare bifurcations, and having a more shallow sulcus.

118

HEBERTELLA OCCIDENTALIS (Hall, 1847)

Figures 11.1-11.3

Orthis occidentalis HALL, 1847, p. 127, pl. 32a, figs. 2a-m, 14.

Orthis sinuata (Hall, 1847) HALL, 1847. p. 128, pl. 32b, figs. 2a-c; HALL AND

CLARKE, 1892, pl. 5a, figs. 1-8.

Hebertella occidentalis (Hall, 1847) FOERSTE, 1910, pl. 2, figs. 1-2; CASTER,

DALVE, AND POPE, 1961, pl. 5, figs. 24-25; WALKER, 1982, p. m6, pl. 5, figs. 18-41;

DAVIS, 1985, p. 49, pl. 5, figs. 24-25; HOWE, 1988, p. 213, pl. 4, figs. 8-11;

SCHWIMMER AND SANDY, 1996, p. 224, pl. 16, figs. 24-26.

Hebertella latasulcata (Hall, 1847) FOERSTE 1914b, p. 131, pl. 3, figs. 7a-b.

Hebertella occidentalis sinuata (Hall, 1847) SCHUCHERT AND COOPER, 1932, pl.

11, figs. 14, 17, 19-20, 22-26.

Hebertella sinuata (Hall, 1847) CASTER, DALVE, AND POPE, 1961, pl. 4, figs. 2-4;

WILLIAMS AND HARPER, 2000, p. 761, pl. 550, figs. 1a-f; WILLIAMS AND

WRIGHT, 1965, p. h325, pl. 205, figs. 5a-e.

Diagnosis.—Emended from Hall (1847). Shell moderate to large, wider than long, subquadrate; shell depth variable, convexoconcave to unequally biconvex; cardinal extremities angular; sulcus wide with moderate to very high depth, typically well developed in larger specimens; ventral muscle scars of variable width; dorsal and ventral umbonal angles low (< 135o). 119

Types.— Hall (1847) did not specify a holotype in his original description, therefore AMNH 30298 is herein designated as the lectotype and AMNH 30299 becomes a paralectotype.

Additional material.— AMNH 30406, USNM 258490, USNM 258496, USNM

258497, YPM S-388a, and ANSP 38106.

Occurrence.— Arnheim, Bellevue, Catheys, Leipers, Fernvale, Oregonia, Liberty,

Whitewater, and Elkhorn Formations, Indiana, Kentucky, Ohio, and Tennessee; upper

Mohawkian to Richmondian (lower to upper Katian).

Discussion.— This morphologically variable species was first described by Hall

(1847) as separate from H. sinuata on the basis that the latter had a more highly developed fold and sulcus, more prominent costae, and a greater height of the dorsal valve. Meek (1873) and Cummings (1908) noted that intermediate forms exist between

H. occidentalis and H. sinuata, and suggested that the two species represent morphological end members of a single species. Similarly, Schuchert and Cooper (1932) listed H. sinuata a subspecies of H. occidentalis and Howe (1966) found significant overlap between the two species using characters considered diagnostic by Hall (1847).

Walker (1982) synonymized the two species (along with H. subjugata, see below) in his descriptions of Hebertella species from Kentucky, with the name H. occidentalis having page priority over H. sinuata in Hall’s (1847) original descriptions. Hebertella occidentalis has a long stratigraphic range and is abundant in shallow subtidal facies throughout the Cincinnati Arch. 120

Figure 11 1, Hebertella occidentalis (Hall, 1847), AMNH 30299 (lectotype), ventral view, × 1.8; 2, dorsal view, × 1.8; 3, anterior view, × 2.4; 4, H. prestonensis Ladd, 1929, SUI 66503 (lectotype), dorsal view, × 1.2; 5, anterior view, × 1.3; 6, H. subjugata (Hall, 1847), AMNH 30290 (holotype), dorsal view, × 1.5.

HEBERTELLA MONTOYENSIS Howe, 1966

Hebertella occidentalis montoyensis HOWE, 1966, p. 255, pl. 30, figs. 15-23.

Diagnosis.—Emended from Howe (1966). Shell large to very large, subquadrate; shell depth high; cardinal extremities angular; sulcus deep, narrow, with many (≥ 17) costae; ventral valve inflated; ventral interarea large; ventral umbo opening narrow (≤

135o) for genus; ventral muscle scars wide.

Types.— USNM 268935 (holotype), USNM 146546.

Occurrence.— Aleman and Cutter Limestones, Texas; Richmondian (upper

Katian). 121

Discussion.— Hebertella montoyensis was originally described by Howe (1966) as a subspecies of H. occidentalis from the Aleman and Cutter Limestones of the Hueco

Mountains in Texas. Howe (1966) differentiated specimens of H. montoyensis from H. occidentalis because they had a much weaker sulcus. Character data supports Howe’s

(1966) distinction (character 7), yet H. montoyensis is herein recognized as a distinct species because phylogenetic analysis indicates H. montoyensis is more closely related to

H. prestonensis than H. occidentalis.

HEBERTELLA PARKSENSIS Foerste, 1909b

Hebertella maria-parksensis FOERSTE, 1909b, p. 319, pl. 7, figs. 6a-b.

Hebertella maria-parkensis [sic] (Foerste, 1909b) SCHUCHERT AND COOPER, 1932, p. 60.

Hebertella parksensis (Foerste, 1909b) WALKER, 1982, p. M8-M9, pl. 5, figs. 1-12, 16-

17.

Diagnosis.—Emended from Foerste (1909b). Shell small to moderate size, outline semielliptical; highly multicostellate; cardinal extremities rounded; dorsal and ventral valves variably convex; sulcus width narrow, deep, filled with numerous (≥ 17) closely packed costae; ventral muscle field pronounced into a platform, with adductors elevated above diductors; ventral interarea low, orthocline; dorsal umbonal angle highly obtuse (≥

140o); dorsal interarea low, not extending significantly over the hinge line. 122

Types.— Foerste (1909b) described two specimens in his original description yet did not designate a holotype. USNM 87054 is herein designated a lectotype, and USNM

87055 becomes a paralectotype.

Other material examined.—USNM 258482.

Occurrence.— Point Pleasant and Clays Ferry Formations, Kentucky;

Mohawkian (upper Sandbian to lower Katian).

Discussion.— Although Foerste (1909b) first described H. parksensis as H. maria-parksensis, Foerste (1909b) likely considered H. parksensis to be distinct from H. maria because later in the same publication he referred to the species as H. parksensis.

Walker (1982) considered H. parksensis to be a valid species, and his taxonomic assessment is upheld by character analysis. Further, phylogenetic analysis places H. parksensis as a sister taxon to the clade comprising H. frankfortensis and H. subjugata and is not closely related to H. maria. Specimens of H. parksensis are characterized by their small to moderate size, numerous bifurcations, and a ventral muscle field with an elevated marginal rim.

HEBERTELLA PRESTONENSIS Ladd, 1929

Figures 11.4-11.5

Hebertella sinuata prestonensis LADD, 1929, p. 401, pl. 5, figs. 3-6.

Diagnosis.—Emended from Ladd (1929). Shell large to very large, subquadrate; inflated dorsal and ventral valves, ventral valve with medial depression; sulcus very deep, narrow, many costae (≥ 17) inserted between inflection points; ventral umbo highly 123 obtuse (≥ 140o); ventral interarea strongly catacline; ventral muscle scars wide; dorsal umbo strongly apsacline.

Types.— Ladd (1929) described two specimens in his original description but did not designate a holotype. SUI 66503 is herein designated as the lectotype and USNM

71927 becomes a paralectotype.

Occurrence.— Maquoketa Formation, Iowa; Richmondian (upper Katian).

Discussion.— Ladd (1929) originally described H. prestonensis as a subspecies of

H. sinuata, a junior synonym of H. occidentalis. However, Ladd (1929) must have considered H. prestonensis distinct because he listed more than six external morphological characters distinguishing the two species. Phylogenetic analysis supports removing H. prestonensis from subspecific taxonomic rank and it is elevated to specific status herein. The topology in Figure 5 indicates H. prestonensis is sister to H. montoyensis from the Montoya Group of Texas. The character state distribution in Table

12 reveals the two species only differ with respect the inclination of their ventral umbo

(character 13). Because H. montoyensis is missing data for five characters, a question arises as to whether further sampling would reveal H. montoyensis and H. prestonensis to be synonymous. Pending further sampling of non-type material, both species are herein considered valid, separate species.

124

HEBERTELLA RICHMONDENSIS Foerste, 1909a

Hebertella alveata richmondensis FOERSTE, 1909a, p. 249, pl. 29, figs. 7-12, 14;

FOERSTE, 1910, p. 55, pl. 5, fig. 10; SCHUCHERT AND COOPER, 1932, p. 60.

Diagnosis.—Emended from Foerste (1909a). Shell large, semielliptical; cardinal extremities variable; hinge line short; dorsal valve highly inflated; ventral muscle scars large; ventral interarea small, apsacline; ventral umbonal angle wide (≥ 140o); dorsal valve with broad medial depression; dorsal cardinal area large; dorsal umbo wide, remaining proximal to the hinge line.

Types.— USNM 87158 (holotype).

Occurrence.— Whitewater Formation, Indiana; Richmondian (upper Katian).

Discussion.— Hebertella richmondensis was originally described by Foerste

(1909a) as a variant of H. alveata from the Whitewater Formation in Indiana. Foerste

(1909a) described H. richmondensis as having a highly convex dorsal valve and much narrower hinge line, whereas H. alveata is generally biconvex and alate. However, phylogenetic analysis indicates that H. richmondensis is more closely related to H. bursa than H. alveata, differing from H. alveata by six characters related to shell depth, sulcus size and costae, and morphology of the ventral interarea (characters 5, 7, 8, 12, 13, 15).

Hebertella richmondensis is, therefore, recognized as a distinct species herein. It is most readily differentiated by its convexoconcave profile, inflated dorsal valve with a broad medial depression extending anteriorly, and narrow hinge line.

125

HEBERTELLA SUBJUGATA (Hall, 1847)

Figure 11.6

Orthis subjugata HALL, 1847, p. 249, pl. 29, figs. 7-12, 14.

Hebertella parksensis (Hall, 1847) WALKER, 1982, pl. 5, figs. 13-14.

Hebertella subjugata (Hall, 1847) FOERSTE, 1910, p. 54, pl. 2, fig. 8; SCHUCHERT

AND COOPER, 1932, p. 60.

Diagnosis.—Emended from Hall (1847). Shell small to moderate size, semielliptical, much wider than long; ventral and dorsal valves inflated, unequally biconvex with steep dorsal valve lacking a medial depression; cardinal extremities rounded; sulcus deep, marked by many (≥ 17) fine costae between inflection points; ventral interarea small, nearly catacline; delthyrial angle wide (≥ 40o); dorsal and ventral umbos remaining proximal to the hinge line.

Types.— AMNH 30290 (holotype).

Occurrence.— Clays Ferry, Bellevue, and Waynesville Formations in Kentucky,

Indiana, and Ohio; Edenian to Richmondian (lower to upper Katian).

Discussion.— Hall (1847) differentiated specimens of H. subjugata from H. occidentalis on the basis of its smaller shell, semielliptical outline, finer costae, and a more prominent ventral umbo. Foerste (1910) furthered this distinction by noting that H. subjugata lacks a medial depression near the beak of the dorsal valve, whereas H. occidentalis has a medial depression terminating anteriorly. Walker (1982) questioned the taxonomic validity of H. subjugata in his discussion of H. occidentalis and H. sinuata, and he listed H. subjugata as a synonym of H. occidentalis. Although Walker’s 126

(1982) synonymy of H. occidentalis and H. sinuata has been supported by other authors

(e.g., Cummings, 1908; Meek, 1873; Howe, 1966; Schuchert and Cooper, 1932), his synonymy of H. subjugata has not (Foerste, 1910; Schuchert and Cooper, 1932). Walker

(1982) supported his synonymy of H. subjugata with H. occidentalis by citing Hall’s

(1847, p. 130) admission that they “are scarcely regarded as distinct, and are usually found mingled together in the collections”. However, Hall’s (1847) support is dubious because he included Orthis subquadrata, which is now recognized as the type species of

Plaesiomys, in his admission. Moreover, the phylogenetic analysis presented here places

H. subjugata outside the H. occidentalis subclade. Primary differences include its semielliptical profile and rounded cardinal extremities.

CONCLUSIONS

The orthid brachiopod genera Plaesiomys and Hebertella were both common and significant genera in Late Ordovician benthic communities. Phylogenetic analysis has documented species-level evolutionary relationships and informed taxonomic revisions for both genera, thereby providing an evolutionary framework for future investigations examining potential drivers of paleobiogeographic differentiation among Late Ordovician brachiopod communities of the Laurentian paleocontinent. In addition, the data presented here may be combined with previous species-level phylogenetic hypotheses of invasive genera (e.g., Wright and Stigall, in review) to further constrain the biogeographic source of the Richmondian invaders. 127

CHAPTER 4: GEOLOGICAL DRIVERS OF LATE ORDOVICIAN FAUNAL

CHANGE IN LAURENTIA: INVESTIGATING LINKS BETWEEN TECTONICS,

SPECIATION, AND INVASIONS

ABSTRACT.— Late Ordovician strata of North America record dramatic faunal turnover linked to the Taconian Orogeny and fluctuating paleoceanographic conditions.

A time-stratigraphic, species-level evolutionary framework was employed to examine paleobiogeographic patterns surrounding these faunal changes to investigate causal links between geological processes and macroevolutionary impacts. Results indicate a shift in the dominant mode of speciation across the middle Mohawkian M4/M5 sequence boundary corresponding to the onset of the Taconic tectophase and the introduction of cool, nutrient rich waters to the epieric seas. Prior to the M5 sequence, speciation occurred predominately by vicariance; whereas afterwards speciation occurred primarily by dispersal and episodes of biotic invasion became frequent across Laurentia. In addition, the biogeographic origin of invader species in a major regional biotic immigration event—the Richmondian Invasion—is evaluated. Species-level biogeographic patterns indicate multiple biogeographic pathways directed the influx of invasive species during the Richmondian Invasion following a relative rise in sea-level and a return to tropical conditions within the Cincinnati Basin.

128

INTRODUCTION

Late Ordovician shallow marine deposits of Laurentia record a series of dramatic faunal and biogeographic shifts including a rapid change from predominately warm to cool water adapted taxa during the middle Mohawkian (late Sandbian) and a major regional biotic immigration event into the Cincinnati Basin during the early Richmondian

(Katian) (Holland, 1997; Patzkowsky and Holland, 1993; Stigall, 2010a). These events of regional faunal turnover may have been driven by ongoing geological processes such as the Taconian Orogeny or fluctuations in paleoceanographic conditions; however, the precise relationship between evolutionary and geologic events has not been adequately investigated. In this analysis, phylogenetic biogeography, a rigorous, quantitative method, is employed within a species-level evolutionary framework to ascertain detailed patterns of biogeographic change and to illuminate the interaction between geological processes and macroevolutionary patterns during the Late Ordovician in Laurentia. Specifically, this chapter aims to elucidate the relative impact of oceanographic vs. tectonic processes on biogeographic differentiation, assess their macroevolutionary impact, and use this paleobiogeographic and macroevolutionary framework to ascertain the biogeographic origin and impact of the Richmondian Invasion.

SYNCHRONIZED GEOLOGIC AND BIOTIC CHANGE

During the Middle and Late Ordovician, the Taconian Orogeny developed along the eastern margin of Laurentia as a result of diachronous collisions with an island arc 129 and/or microplates (Shanmugam and Lash, 1982; Ettensohn, 1994) (Figure 12). During the Blountian tectophase, tropical carbonate deposition was widespread throughout

Laurentia and biogeographic and faunal divisions were relatively stable. However, the onset of the Taconic tectophase during the middle Mohawkian (M5 sequence, late

Sandbian) and concomitant global sea-level rise (Figure 13) introduced major lithologic and faunal changes.

130

Beginning in the M5 sequence, the widespread carbonates of eastern Laurentia reflect a heterozoan assemblage as tropical-style carbonate deposition became restricted to the continental interior and marginal basins (Holland and Patzkowsky, 1996; Pope and

Read, 1997). The synergistic combination of flexural downwarping associated with the

Taconic tectophase and a major eustatic sea-level rise has been suggested to have facilitated an upwelling zone originating from the Sebree Trough which introduced cool, nutrient rich water into eastern Laurentia (Holland and Patzkowsky, 1996; Pope and

Read, 1997). Other explanations of the shift to temperate carbonates suggest tectonic restructuring may have facilitated increased thermohaline circulation and upwelling

(Kolata et al., 2001), the onset of glaciations (Saltzman and Young, 2005), or a combination of the two (Ettensohn, 2010). At this same time, faunas of the Appalachian

Basin transitioned from warm water organisms to predominately cool water adapted taxa

(Patzkowsky and Holland, 1993). The biotic overturn was largely biogeographic, involving only regional generic extinction, geographic restriction, and subsequent expansion (Patzkowsky and Holland, 1993). For example, trilobites interpreted to be adapted to cooler water participated in northward biogeographic expansion in association with the spread of temperate carbonates (Shaw, 1991).

By the Richmondian Stage (late Katian), the source of the upwelling was shut off and tropical-style carbonate deposition resumed across Laurentia and continued until the

Hirnantian. Whether the return to tropical conditions was due to an interval of global warming related to the Boda Event (Fortey and Cocks, 2005) or to a basinal infilling of the source of upwelling (e.g., Holland and Patzkowsky, 1996) remains uncertain. 131

Nonetheless, the C3/C4 sequence boundary (Figure 13) corresponds to the return of tropical-style carbonate deposition in eastern Laurentia and marks the initiation of a large scale cross-faunal immigration event into the Cincinnati basin known as the Richmondian

Invasion (Foerste, 1912; Holland, 1997).

The paleoequatorial region of the western United States and Canada has long been considered the biogeographic source of the Richmondian invaders (e.g., Foerste, 1912;

Nelson, 1959). This generalization is nebulous, however, as there were many possible 132 source regions with different pathways of invasion available during the Late Ordovician

(e.g., Jin, 2001). Invader taxa were geographically widespread throughout Late

Ordovician, with invader genera occurring throughout the Cincinnatian interval in areas such as Wyoming (Holland and Patzkowsky, 2009), Canada (Jin and Zhan, 2001), and

Tennessee (Howe, 1988). The spatiotemporal pervasiveness of invasive genera confounds studies limited to genus-level patterns. Consequently, a species-level phylogenetic framework, such as the one provided herein, is required to rigorously test competing hypotheses regarding the biogeographic sources of the Richmondian invaders and elucidate the geologic controls on Late Ordovician paleobiogeographic patterns.

BIOGEOGRAPHIC ANALYSIS

Phylogenetic biogeography is a powerful method for uncovering macroevolutionary and paleobiogeographic phenomena using the fossil record (e.g.,

Lieberman, 1997; 2000). This analysis focuses on species of orthid brachiopods, which were among the most abundant organisms in late Middle Ordovician and Late Ordovician seas. Their benthic, sessile lifestyle and planktonic larvae makes them amenable to paleobiogeographic methods because the inferred paleobiogeographic distribution of a species is likely to closely reflect its actual distribution (Kidwell and Flessa, 1996).

Species-level phylogenetic hypotheses constructed in Chapters 2 and 3 were utilized in this investigation for a combined total of 53 species in three genera: Glyptorthis Foerste,

1914a; Plaesiomys Hall and Clark, 1892; and Hebertella Hall and Clarke, 1892. Two species included in the analysis, Glyptorthis insculpta (Hall, 1847) and Plaesiomys 133 subquadratus (Hall, 1847) participated in the Richmondian Invasion into the Cincinnati

Basin. These taxa were distributed among the tectonic basins of Laurentia during the late

Middle to Late Ordovician (Figure 12, Chapters 1 and 2).

A Lieberman-modified Brooks Parsimony Analysis [LBPA] as described in

Wiley and Lieberman (2011) was employed to investigate processes driving macroevolutionary and biogeographic patterns in Middle and Late Ordovician brachiopods. Species-level phylogenetic hypotheses from Chapters 2 and 3 were converted to taxon-area cladograms by placing the geographic distributions of each species at the tips of each cladogram and optimizing ancestral biogeographic areas onto the internal nodes using Fitch Parsimony (Figures 13; 14). Because different geologic processes operated before and after the M4/M5 sequence boundary, LBPA was employed in a time-stratigraphic framework to more accurately uncover patterns of biogeographic differentiation. Cladogenetic events were separated into two time slices: Time Slice 1

(T1) corresponds interval represented by the W-b through M4 sequences, and Time Slice

2 (T2) corresponds to the interval represented by the M5 though C6 sequences (Figures

13; 15).

For each Time Slice, two matrices were coded from the biogeographic data in the taxon-area cladograms: one coding for patterns of vicariance and the other geodispersal

(Figure 16; Table 23). General area cladograms were constructed from each matrix by parsimony analysis using PAUP* 4.0 (Swofford, 2002) (Figure 17).

134

Figure 16 Stratocladograms from Figure 15 with nodes and tips labeled to correspond with coding used in LBPA. The division between T1 and T2 is indicated by the horizontal line.

MACROEVOLUTIONARY IMPACT OF BIOGEOGRAPHIC CHANGE

Results of the preceding analyses indicate fundamental differences in biogeographic and macroevolutionary patterns between the T1 and T2 intervals, which correlate with the Blountian and Taconic tectophases, respectively.

Analysis of speciation mode at cladogenetic events identified vicariance as the dominant mode of speciation during the T1 interval operating in 77% of the 26 discernible speciation events; whereas only 15% of 13 discernible speciation events during the T2 interval were attributable to vicariance (Figure 15). This dramatic pattern arises from differences in evolutionary differentiation of the genera analyzed through time (Figure 14). The taxon-area cladogram for species of Glyptorthis indicates that the genus first originated in the Southern mid-Continent region and subsequently underwent a major dispersal event into the Appalachian and Central basins. Because species of

Glyptorthis species were geographically widespread early in the history of the clade, subsequent range contractions led to frequent speciation by vicariance. In contrast, ancestral species of Plaesiomys and Hebertella occurred predominately in one tectonic basin (the Northern Midcontinent and Cincinnati Basin, respectively) with frequent episodes of speciation via dispersal into neighboring tectonic basins. Although documentation of this pattern in additional taxa would be desirable, no additional species- level phylogenetic hypotheses exist for contemporaneous taxa. The recovered pattern of high vicariance during the Blountian tectophase followed by high dispersal during the

Taconic tectophase has been documented by other researchers in generic level analyses of other clades (e.g., Elias, 1983; Goldman and Wu, 2010; Miller and Mao, 1995; 138

Patzkowsky and Holland, 1993); our recovered pattern, therefore, cannot be attributable to sampling bias alone.

Figure 17 Vicariance and Geodispersal general area cladograms for T1 and T2. The vicariance trees indicate the order in which areas became separated from each other due to barrier (tectonic, eustatic, etc.) formation, whereas geodispersal trees indicate the relative order in which barriers were removed. Congruence between tree topologies indicates that geographic barriers rose and fell in the same relative order, suggesting the importance of cyclical (e.g., oscillating sea levels) geologic processes. In contrast, incongruence indicates that non-cyclic, singular events (e.g. tectonic pulses) influenced the evolution and biogeographic differentiation of the clade (Lieberman, 2000). Numerical values are bootstrap (plain text) and jackknife (italicized) values that indicate support for nodes. T1 analysis recovered a single most parsimonious vicariance tree (length of 87 steps, consistency index 0.736) and two most parsimonious geodispersal trees (64 steps, CI 0.797). T2 analysis recovered a single most parsimonious vicariance tree (49 steps, CI 0.816) and two most parsimonious geodispersal trees (59 steps, CI 0.531). Geodispersal trees illustrate topology of the strict consensus tree. 139

The dramatically different patterns of speciation and biogeographic differentiation between T1 and T2 reflect different geological drivers operating before and after the

M4/M5 boundary. This differentiation is clearly illustrated by the divergent topology of vicariance trees between times slices and the incongruence of the geodispersal and vicariance trees within each time slice (Fig. 17). The T1 vicariance tree is well resolved and indicates a strong biogeographic relationship between the Appalachian basin and the southern mid-Continent region. The numerous T1 vicariance events provide clear tree support; conversely, the strict consensus of the T1 geodispersal tree is unresolved, which reflects conflicting biogeographic patterns amid the paucity of speciation events occurring via dispersal in T1 (Figure 15). The T2 vicariance tree is also well resolved despite the general lack of vicariance events during the T2 interval, which indicates little conflict amongst resolved patterns. Interestingly, the geodispersal tree in T2 is even more unresolved than in T1 even though speciation occurred predominately by dispersal during this interval. The lack of resolution arises from conflicting data regarding pathways of dispersal during T2. When each of the two most parsimonious geodispersal trees for T2 are examined, one is consistent with a close relationship between the Cincinnati basin and the mid-Continent region (supported by biogeographic affinities of G. insculpta), whereas the other tree is consistent with a more recent exchange of taxa between Anticosti Island and the mid-Continent (supported by biogeographic affinities of P. subquadratus).

During the T1 interval, the Blountian tectophase introduced siliciclastics into eastern Laurentia yet did not inhibit the production of tropical carbonates (Holland and

Patzkowsky, 1996). The high rate of vicariant speciation during the T1 corroborates a 140 study by Miller and Mao (1995), which found an increased rate of diversification for late

Whiterockian to early Mohawkian genera within multiple clades located in regions closely associated with orogenic activity. The species-level evolutionary and biogeographic patterns presented here, therefore, support Miller and Mao’s (1995) hypothesis that the high rates of diversification may be due to increased levels of habitat partitioning associated with tectonic activity. For example, localized tectonic pulses during the Blountian tectophase may have introduced siliciclastics into eastern Laurentia and fragmented previously large biogeographic ranges into disjunct areas with different environmental or ecological conditions and led to an increased production of vicariant speciation.

Dispersal became the dominant mode of speciation during the Taconic tectophase of the Taconian Orogeny, corresponding to the flexural downwarping of the craton and a major global rise in sea level at the M5 boundary. The upwelling of cool, nutrient rich water from the Sebree Trough into the shallow epieric seas of eastern Laurentia facilitated regional extirpation (Patzkowsky and Holland, 1993) and effectively shut off the tectonic driver of vicariant speciation by removing the preferred habitat of warm- water adapted taxa.

The shift in speciation regime across the M5 boundary is significant because high rates of vicariance are typically associated with high rates of speciation, whereas high rates of dispersal are associated with reduced speciation (Stigall, 2010b). The negative correlation between dispersal and speciation rates likely results from increased numbers of invasive species. Intervals with frequent interbasinal biotic invasions have low 141 biodiversity due to suppressed speciation rates rather than elevated extinction rates

(Stigall, 2010b). Although speciation rates are not quantitatively addressed here, a sharp decrease in generic diversification during the T2 was found by Miller and Mao (1995, p.

306, Figure 2). Similar biogeographic patterns and associated drops in diversity have been described in other taxa, such as Laurentian graptolites (Goldman and Wu, 2010) and corals (Elias, 1983). Further, Elias (1991) noted that previously diverse coral faunas became impoverished during the Richmondian and many genera participating in the

Richmondian Invasion were monospecific. The correlation between the development of a less speciose fauna and a shift to a dispersal dominated regime suggests a potential decrease in speciation rates for Laurentian taxa during the T2 interval.

The Richmondian Invasion into the Cincinnati Basin occurred concomitant with a return to tropical carbonate deposition. The largest influx of invaders into the Cincinnati

Basin occurred during the C5 sequence, which corresponds to the greatest rise in sea level throughout the Cincinnatian (Holland, 1997). The data presented here suggest that the Richmondian Invasion represented a localized manifestation of a more generalized phenomenon of interbasinal invasion occurring throughout the T2 interval, whereby a return to favorable conditions facilitated an influx of an impoverished fauna with a high propensity for dispersal. The biogeographic origin of the invader species included in this analysis indicates both the southern mid-Continent and Anticosti Island as source regions.

Generic patterns from other taxa provide further support for multiple pathways of invasion. For example, rugose corals and bryozoans have closer biogeographic affinities to the northern mid-Continent region (Anstey, 1986; Webby et al., 2004). High rates of 142 dispersal in the T2 interval combined with a return to favorable conditions for warm- water adapted taxa within the Cincinnati basin facilitated a regional biotic immigration into the Cincinnati Basin from multiple geographic regions.

CONCLUSIONS

Phylogenetic biogeographic analysis of species-level patterns in Middle to Late

Ordovician brachiopods indicates a shift in biogeographic patterns and macroevolutionary dynamics between the T1 and T2 intervals. These patterns were driven largely by geological processes related to the Taconian Orogeny and a major rise in global sea-level at the M4/M5 boundary. Prior to the M5, tectonic activity associated with the Blountian tectophase produced increased habitat fragmentation, thereby facilitating high levels of vicariant speciation. In contrast, the post M5 interval corresponds to a shift in paleoceanographic conditions related to the combined effect of structural deformation from the Taconic tectophase and a global eustatic rise in sea level introducing cool, nutrient rich water from the Sebree Trough onto the shallow intracratonic seas. Dispersal became the dominant form of speciation and the increased number of invasive species may have led to relatively impoverished, low diversity faunas. The Richmondian Invasion into the Cincinnati basin was driven by a return to tropical carbonate conditions and a rise in relative sea level. Finally, species-level phylogenetic biogeographic patterns suggest both intracratonic seas and margin basins supplied invasive species into the Cincinnati basin.

CHAPTER 5: CONCLUSION

Biogeographic analyses considered within an evolutionary framework provide detailed insight into how geologic processes affect macroevolutionary patterns. Because no species-level phylogenetic hypotheses previously existed for Late Ordovician brachiopods, phylogenetic analysis was conducted on Laurentian species in three genera:

Glyptorthis Foerste, 1914a, Plaesiomys Hall and Clarke, 1892, and Hebertella Hall and

Clarke, 1892. Phylogenetic biogeographic methods were used to investigate the evolutionary and biogeographic origin of the Late Ordovician Richmondian invaders, as well as uncover generalized evolutionary and biogeographic patterns in Middle to Late

Ordovician brachiopods.

The phylogenetic hypotheses for species of Glyptorthis, Plaesiomys, and

Hebertella in Chapters 2 and 3 were used to document evolutionary relationships and inform taxonomic revisions. All phylogenetic analyses recovered well supported, single most parsimonious trees, and therefore supply a strong evolutionary framework for conducting biogeographic analyses. Glyptorthis virginica is herein synonymized with G. uniformis. Glyptorthis maquoketensis is recognized as a distinct species rather than a subspecies of G. insculpta. Plaesiomys cutterensis, P. idahoensis, and P. occidentalis are herein recognized as distinct species rather than subspecies or geographical variants of P. subquadratus. Hebertella montoyensis and H. prestonensis are recognized as distinct species separate from H. occidentalis, and H. richmondensis is recognized as a distinct species rather than a geographical variant of H. alveata. Hebertella subjugata is removed from its tentative synonymy with H. occidentalis and revalidated. 145

Following phylogenetic analysis, each species-level phylogenetic hypothesis was converted into a taxon-area cladogram and the internal nodes were optimized using Fitch

Parsimony (Lieberman, 2000). This procedure was implemented to recover ancestral geographic ranges, which were then used to ascertain the biogeographic origin of two participants of the Richmondian Invasion: Glyptorthis insculpta (Hall, 1847) and

Plaesiomys subquadratus (Hall, 1847). In contrast to previous studies suggesting a unidirectional pathway of invasion (Foerste, 1912; Holland, 1997; Nelson, 1959), the results presented here support a multidirectional pathway of invasion from different biogeographic sources. Results indicate that G. insculpta invaded the Cincinnati basin from the southern mid-Continent region, whereas P. subquadratus invaded from the marginal basin of Anticosti Island. Further, support for a multidirectional invasion is indicated by genus-level biogeographic studies on other taxa such as bryozoans (Anstey,

1986) and corals (Elias, 1983), which suggest biogeographic affinities to the northern mid-Continent region.

In chapter 4, phylogenetic biogeographic analytical methods were used within a time-stratigraphic framework to assess the role of geological processes driving macroevolutionary and paleobiogeographic patterns. When inferred patterns of speciation were examined within a time-stratigraphic framework, speciation by vicariance was more common than dispersal during the Whiterockian to mid-Mohawkian, whereas speciation by dispersal was more common after the mid-Mohawkian. The vicariance and geodispersal trees recovered from the time-stratigraphic model indicate that paleobiogeographic patterns were primarily driven singular geological events rather than 146 cyclical processes (at least on the time scale of speciation). Differential paleobiogeographic patterns were driven largely by diachronous tectonic activity associated with the Blountian and Taconic tectophases of Taconian Orogeny and the onset of a major global rise in sea-level occurring in the middle Mohawkian led to shifts in paleooceanographic conditions across Laurentia. This study supports the notion that geological events have a profound effect on paleobiogeographic patterns and macroevolutionary dynamics, and therefore supports the notion that the Earth and its biota co-evolved (Croizat, 1964).

147

REFERENCES

ALBERSTADT, L. P. 1973. Articulate brachiopods of the Viola Formation (Ordovician)

in the Arbuckle Mountains, Oklahoma. Oklahoma Geological Survey, Bulletin

117, 90 p.

ALEXANDER, R. R. 1984. Comparative hydrodynamic stability of brachiopod shells on

current-scoured arenaceous substrates. Lethaia, 17: 17-32.

ANSTEY, R. L. 1986. Bryozoan provinces and patterns of generic evolution and

extinction in the Late Ordovician of North America. Lethaia, 19: 33-51.

BASSETT, M. G. 1972. The articulate brachiopods of the Wenlock Series of the Welsh

Borderland and South Wales. Paleontographical Society, 126: 27-78.

BASSLER, R. S. 1932. The stratigraphy of the Central Basin of Tennessee. Tennessee

Division of Geology Bulletin 38, 268 p.

BILLINGS, E. 1862. New species of fossils from different parts of the Lower, Middle,

and Upper Silurian rocks of Canada. Geological Survey of Canada, Palaeozoic

Fossils, 1: 96-168.

BRADLEY, J. H., 1921. The brachiopoda of the Maquoketa of Iowa, Harvard College

Museum of Comparative Zoology Bulletin, 64: 503-525.

CASTOR, K. E., E. A. DALVE, AND J. K. POPE. 1961. Elementary guide to the fossils

and strata of the Ordovician in the vicinity of Cincinnati, Ohio. Cincinnati

Museum of Natural History, 47 p. 148

COCKS, L. R. M. 1978. A review of Brtish Lower Palaeozoic brachiopods, including a

synoptic revision of Davidson’s monograph. Monograph of the

Palaeontographical Society, 131, 256 p.

COCKS, L. R. M. AND T. H. TORSVIK. 2011. The Palaeozoic geography of Laurentia

and western Laurussia: A stable craton with mobile margins. Earth Science

Reviews, 106: 1-51.

CONRAD, T. A. 1843. Observations on the lead-bearing limestone of Wisconsin and

descriptions of a new genus of trilobites and fifteen new Silurian fossils.

Proceedings of the Academy of Natural Sciences, Philadelphia, 1: 329-335.

COOPER, G. A. 1956. Chazyan and related brachiopods. Smithsonian Miscellaneous

Collections, 127(1)1-1024; 127(2):1025-1245.

CROIZAT, L. 1964. Space, Time, and Form: The Biological Synthesis. Caracas, 881 p.

CUMMINGS, E. R. 1908. The stratigraphy and paleontology of the Cincinnati series of

Indiana. Indiana Department of Geology and Natural Resources Annual Report,

32: 607-1189.

DAVIS, M. A. 2009. Invasion Biology. Oxford University Press, Oxford, 244 p.

DAVIS, R. A. 1985. Cincinnati fossils, an elementary guide to the Ordovician rocks and

fossils of the Cincinnati, Ohio, region. Cincinnati Museum of Natural History

Popular Publication Series, 10: 60 pp.

DAVIDSON, T. 1847. Observations on some Wenlock-limestone Brachiopoda, with

descriptions of several new species. London Geological Journal, 1: 52-65. 149

ELIAS, R. J. 1983. Middle and Late Ordovician solitary rugose corals of the Cincinnati

Arch region. United States Geological Survey Professional Paper, 1066-N: N1-

N13.

ELIAS, R. J. 1991. Environmental cycles and bioevents in the Upper Ordovician Red

River-Stony Mountain solitary rugose coral province of North America, p. 205-

211. In C. R. Barnes and S. H. Williams (eds.), Advances in Ordovician geology.

Geological Survey of Canada Paper 90.

EMMONS, E. 1842. Geology of New York, Part 2, comprising the survey of the second

geological district. White and J. Visscher, Albany, 437 p.

ETTENSOHN, F. R. 1994. Tectonic control on the formation and cyclicity of major

Appalachian unconformities and associated stratigraphic sequences p. 17-242, In

J. M. Dennison and F. R. Ettensohn (eds.), Tectonic and Eustatic Controls on

Sedimentary Cycles: SEPM, Concepts in Sedimentology and Paleontology, 4.

ETTENSOHN, F. R. 2010. Origin of Late Ordovician (mid-Mohawkian) temperate-water

conditions on southeastern Laurentia: glacial or tectonic? Geological Society of

America Special Paper, 466: 163-175.

FOERSTE, A. F. 1885. The Clinton Group of Ohio, with descriptions of new species.

Bulletin of the Denison University Science Laboratory of Denison University, 1:

63-121.

FOERSTE, A. F. 1909a. Preliminary notes on Cincinnatian fossils. Bulletin of the

Denison University Science Laboratory of Denison University, 14: 209-228. 150

FOERSTE, A. F. 1909b. Preliminary notes on Cincinnati and Lexington fossils. Bulletin

of the Scientific Laboratory of Denison University, 14: 298-324.

FOERSTE, A. F. 1910. Preliminary notes on Cincinnatian and Lexington fossils of Ohio,

Indiana, Kentucky, and Tennessee. Bulletin of the Scientific Laboratory of

Denison University, 16: 17-100.

FOERSTE, A. F. 1912. The Arnheim Formation within the areas traversed by the

Cincinnati Geanticline. The Ohio Naturalist, 12: 429-456.

FOERSTE, A. F. 1914a. Notes of the Lorraine faunas of New York and the province of

Quebec. Bulletin of the Denison University Science Laboratory of Denison

University, 17: 247-328.

FOERSTE, A. F. 1914b. The Rogers Gap fauna of central Kentucky. Cincinnati Society

of Natural History Journal, 21: 109-156.

FOERSTE, A. F. 1917. Notes on Richmond and related fossils. Journal of the Cincinnati

Society of Natural History, 22:42-55.

FORTEY, R. A. AND L. R. M. COCKS. 2005. Late Ordovician global warming—The

Boda event. Geology, 33: 405-408.

GOLDMAN, D. AND S.-Y. WU. 2010. Paleogeographic, paleoceanographic, and

tectonic controls on early Late Ordovician graptolite diversity patterns.

Geological Society of America Special Paper, 466: 149-161.

GREGOR, D. K. AND K. E. BORN. 1936. Stratigraphy and fauna of the Fernvale

Formation (Illinois). Washington University Studies, Science and Technology, 6:

67-77. 151

HALL, J. 1847. Descriptions of the organic remains of the lower division of the New

York System. New York State Geological Survey, Palaeontology of New York, 1,

338 p.

HALL, J. AND J. M. CLARK. 1892. An introduction to the study of the genera of

Palaeozoic Brachiopoda. New York State Geological Survey, Palaeontology of

New York, 8(1), 367 p.

HAMMER, T., D. A. T. HARPER, AND P. D. RYAN. 2001. PAST: Paleontological

statistics software package for education and data analysis. Paleontologica

Electronica, 4:1-9.

HAVLÍCEK, V. 1977. Brachiopods of the Order Orthida in Czechoslovakia. Rozpravy

Ústredního Ústavu Geologického, 43: 535-665.

HILLER, N. 1980. Ashgill Brachiopoda from the Glyn Ceiriog District, north Wales.

Bulletin of the British Museum (Natural History), Geology Series, 34: 109-216.

HILLIS, D. M. AND J. P. HUELSENBECK. 1992. Signal, noise, and reliability in

molecular phylogenetic analyses. Journal of Heredity, 83: 189-195.

HOLLAND, S. M. 1997. Using time/environment analysis to recognize faunal events in

the Upper Ordovician of the Cincinnati Arch, p. 309-334. In Brett, C.E. and G. C.

Baird, (eds.), Paleontological Events: Stratigraphic, Ecological, and Evolutionary

Implications. Columbia University Press, New York.

HOLLAND, S. M. AND M. E. PATZKOWSKY. 1996. Sequence stratigraphy and long-

term paleooceanographic change in the Middle and Upper Ordovician of the

eastern United States, p. 117-129. In B. J. Witzke, G. A. Ludvigson, and J. Day 152

(eds.), Paleozoic Sequence Stratigraphy: Views from the North American Craton.

Geological Society of America Special Paper 306.

HOLLAND, S. M. AND M. E. PATZKOWSKY. 2009. The stratigraphic distribution of

fossils in a tropical carbonate succession: Ordovician Bighorn Dolomite,

Wyoming, USA. Palaios, 24: 303-317.

HOPKINS, M. J. 2011. Species-level phylogenetic analysis of Pterocephaliids (Trilobita,

Cambrian) from the Great Basin, Western USA. Journal of Paleontology, 85:

1128-1153.

HOWE, H. J. 1966. Orthacea from the Montoya Group (Ordovician) of Trans-Pecos

Texas. Journal of Paleontology, 40: 241-257.

HOWE, H. J. 1988. Articulate brachiopods from the Richmondian of Tennessee. Journal

of Paleontology, 62: 204-218.

INTERNATIONAL COMMISION ON ZOOLOGICAL NOMENCLATURE. 1999.

International Code of Zoological Nomenclature. International Trust for

Zoological Nomenclature, London, 306 pp.

JAMES, U. P. 1871. Catalogue of Lower Silurian fossils, Cincinnati Group, Ohio, 14 p.

JIN, J. AND B. S. NORFORD. 1996. Upper Ordovician (Caradoc) brachiopods from the

Advance Formation, northern Rocky Mountains, British Columbia, p. 20-77. In

Advance Formation: Stratigraphy and biostratigraphy of a new Ordovician

formation from the Rocky Mountains, northeastern British Columbia. Geological

Survey of Canada Bulletin, 491. 153

JIN. J. 2012. Cincinnetina, a new Late Ordovician Dalmanellid brachiopod from the

Cincinnati type area, USA: implications for the evolution and palaeogeography of

the epicontinental fauna of Laurentia. Palaeontology, 55: 205-228.

JIN, J. AND R.-B. ZHAN. 2001. Late Ordovician articulate brachiopods from the Red

River and Stony Mountain formations, southern Manitoba. National Research

Council of Canada, Ottawa, 117 p.

JIN, J. AND R.-B. ZHAN. 2005. Brachiopods from the Middle Ordovician Shihtzupu

Formation of Yunnan Province, China. Acta Palaeontologica Polonica, 50: 265-

393.

JIN, J. AND R.-B. ZHAN. 2008. Late Ordovician Orthide and Billingsellide brachiopods

from Anticosti Island, Eastern Canada: Diversity change through Mass

Extinction. National Research Council of Canada, Ottawa, 151 p.

JIN, J. , R.-B. ZHAN, P. COPPER, AND W. G. E. CALDWELL. 2007. Epipunctae and

phosphatized setae in Late Ordovician Plaesiomyid brachiopods from Anticosti

Island, Eastern Canada. Journal of Paleontology, 81: 666-683.

JIN, J. W. G. E. CALDWELL, AND B. S. NORFORD. 1997. Late Ordovician

brachiopods and biostratigraphy of the Hudson Bay lowlands, Northern Manitoba

and Ontario. Geological Survey of Canada Bulletin, 513, 258. p.

KIDWELL, S. M., and K. W. FLESSA. 1996. The Quality of the Fossil Record:

Populations, Species, and Communities, Annual Review of Earth and Planetary

Science, 24:433-464. 154

KLASSEN, G. J., R. D. MOOI, AND A. LOCKE 1991. Consistency indeces and random

data. Systematic Zoology, 50: 446-457.

KOLATA, D. R., W. D. HUFF, AND S. M. BERGSTRÖM. 2001. The Ordovician

Sebree Trough: an oceanic passage to the MidContinent Unites States. Geological

Society of America Bulletin, 113: 1067-1078.

KULKOV, N. P., AND L. G. SEVERGINA. 1989. Stratigrafiya I brakhiopody ordovika

I nizhnego silura Gornogo Altaya. Akademiya Nauk SSSR, Sibirskoe Otdelenie,

Trudy Instituta Geologii I Geofiziki, 717, 223 p.

LABARBERA, M. 1981. Water flow patterns in and around three species of Articulate

brachiopods. Journal of Experimental Marine Biology and Ecology, 55: 185-206.

LADD, H. S. 1929. The stratigraphy and paleontology of the Maquoketa Shale of Iowa,

Pt. 1. In Annual Report, 1928, Iowa Geological Survey Report, 34: 309-448.

LAURIE, J. R. 1991. Articulate brachiopods from the Ordovician and Lower Silurian of

Tasmania. Association of Australasian Palaeontologists Memoir, 11: 1-106.

LIEBERMAN, B. S. 1997. Early Cambrian paleogeography and tectonic history: a

biogeographic approach. Geology, 25: 1039-1042.

LIEBERMAN, B. S. 2000. Paleobiogeography. Kluwer Academic/Plenum Publishers,

New York, 208 p.

LEIGHTON, L. R. AND C. G. MAPLES. 2002. Evaluating internal versus external

characters: phylogenetic analyses of Echinoconchidae, Bruxtoniinae, and

Juresaniinae (phylum Brachiopoda). Journal of Paleontology, 76: 659-671. 155

LI, R. AND P. COPPER. 2006. Early Silurian (Llandovery) orthide brachiopods from

Anticosti Island, eastern Canada: the O/S extinction recovery fauna. Special

Papers in Palaeontology, 76, Palaeontological Association, 71 p.

MADDISON, W. P. AND D. R. MADDISON. 2003. MacClade: Analysis of phylogeny

and character evolution. 4.06. Sinauer, Sunderland, Massachusetts.

MALIZIA, R. AND A. L. STIGALL. 2011. Niche stability in Late Ordovician articulated

brachiopod species before, during, and after the Richmondian Invasion.

Palaeogeography, Palaeoclimatology, Palaeoecology, 311: 154-170.

M’COY, F. 1846. Synopsis of the Silurian fossils of Ireland, Dublin, 77 p.

MEEK, F. B. 1873. Descriptions of invertebrate fossils of the Silurian and Devonian

systems. Ohio Geological Survey. Vol. 1. Pt. 2. Palaeontology, 243 p.

MEYER, D. L. AND R. A. DAVIS. 2009. A Sea Without Fish: Life in the Ordovician

Sea of the Cincinnati Region. Indiana University Press, Bloomington, 346 p.

MILLER, A. I. AND S. MAO. 1995. Association of orogenic activity with the

Ordovician radiation of marine life. Geology, 23: 305-308.

MITCHELL, W. I. 1977. The Ordovician Brachiopoda from Pomeroy, Co. Tyrone.

Palaeontographical Society Monographs, 130. 138 p.

MORTON, C. M. AND D. T. KINCAID. 1995. A model for coding pollen size in

reference to phylogeny using examples from the Ebenaceae. American Journal of

Botany, 82: 1173-1178.

NELSON, S. J. 1959. Arctic Ordovician Fauna: an equatorial assemblage? Journal of the

Alberta Society of Petroleum Geologists, 7: 45-53. 156

OKULTICH, V. J. 1943. The Stony Mountain Formation of Manitoba. Transactions of

the Royal Society of Canada, 37: 59-74.

PATZKOWSKY, M. E. AND S. M. HOLLAND. 1993. Biotic response to a Middle

Ordovician paleoceanographic event in eastern North America. Geology, 21: 619-

622.

POPE, M. C. AND J. F. READ. 1997. High-resolution stratigraphy of the Lexington

Limestone (late Middle Ordovician), Kentucky, U.S.A.: a cool-water carbonate-

clastic ramp in a tectonically active foreland basin, p. 411-429. In N. P. James and

J. A. D. Clarke (eds.), Cool Water Carbonates. Society for Sedimentary Geology

Special Publication 56.

POPE, M. C. AND J. B. STEFFEN. 2003. Widespread, prolonged late Middle to Late

Ordovician upwelling in North America: A proxy record of glaciations? Geology,

31: 63-66.

POPOV, L. E., I. F. NIKITIN, AND L. R. M. COCKS. 2000. Late Ordovician

brachiopods from the Otar Member of the Chu-Ili Range, South Kazakhstan.

Palaeontology, 42: 625-661.

REED, F. R. C. 1944. Notes on some new Ordovician brachiopods from Girvan. Annals

and Magazine of Natural History, 16: 1-12.

REED, F. R. C. 1945. Notes on some Ordovician brachiopods and a trilobite from the

Balclatchie Beds of Girvan. Annals and Magazine of Natural History, 11: 215-

222. 157

REED, F. R. C. 1952. Revision of certain Ordovician fossils from County Tyrone.

Proceedings from the Royal Irish Academy, 55: 29-136.

RICHARDS, R. P. 1972. Autecology of Richmondian brachiopods (Late Ordovician of

Indiana and Ohio). Journal of Paleontology, 46: 386-405.

RODE, A. L. 2004. Phylogenetic revision of the Devonian bivalve, Leptodesma

(Leiopteria). Yale University Postilla, 229: 1-26.

RÕÕMUSOKS, A. 1964. Nekotorye brakhiopody iz ordovika Estonii. Tartu Riikliku

Ulikooli Tometised, 153:3-28. (In Russian).

ROSS, R. J. 1959. Brachiopod fauna of Saturday Mountain Formation, southern Lemhi

Range, Idaho. U. S. Geological Survey Professional Paper, 294-L: 441-461.

ROY, S. K. 1941. The Upper Ordovician fauna of Frobisher Bay, Baffin Land, Field

Museum of Natural History (Geology) Memoir, 2, 212 p.

RUBEL, M. 1962. Brachiopods of the Superfamily Orthacea from the Llandovery of

Estonia. Proceedings of the Estonian Academy of Science, Geology, 9: 75-95.

RUDWICK, M. J. S. 1970. Living and fossil brachiopods. Hutchinson and Company,

London, 199 p.

SALTZMAN, M. R. AND S. A. YOUNG. 2005. Long-lived glaciations in the Late

Ordovician? Isotopic and sequence-stratigraphic evidence from western

Laurentia. Geology, 33: 109-112.

SARDESON, R. W. 1892. The range and distribution of the Lower Silurian fauna of

Minnesota with descriptions of some new species. Minnesota Academy of Natural

Sciences Bulletin, 3: 326-343. 158

SCHUCHERT, C. 1900. On the Lower Silurian (Trenton) fauna of Baffin Land.

Proceedings of the United States National Museum, 22: 143-177

SCHUCHERT, C. 1913. Class 2. Brachiopoda. p. 355-420. In K. A. von Zittel (ed.),

Textbook of Paleontology, 2. Macmillan, London.

SCHUCHERT, C. AND G. A. COOPER. 1930. Upper Ordovician and Lower Devonian

stratigraphy and paleontology of Percé, Quebec. American Journal of Science, 2:

161-176.

SCHUCHERT, C. AND G. A. COOPER. 1931. Synopsis of the brachiopod genera of the

suborders Orthoidea and Pentameroidea, with notes on the Telotremata. American

Journal of Science, 5: 241-255.

SCHUCHERT, C. AND G. A. COOPER. 1932. Brachiopod genera of the suborders

Orthoidea and Pentameroidea. Peabody Museum of Natural History Memoir,

4(1), 270 p.

SCHWIMMER, B. A. AND M. R. SANDY. 1996. Phylum Brachiopoda, p. 210-241. In

R. M. Feldmann and M. Hackathorn (eds.), Fossils of Ohio. Ohio Department of

Natural Resources Bulletin, 70.

SHALER, N. S. 1865. List of the Brachiopoda from the island of Anticosti sent by the

Museum of Comparative Zoology to different institutions for exchange for other

specimens, with annotations. Harvard University, Museum of Comparative

Zoology Bulletin, 1: 61-70.

SHANMUGAM, G. AND G. G. LASH. 1982. Analogous tectonic evolution of the

Ordovician foredeeps, southern and central Appalachians. Geology, 10: 562-566. 159

SHAW, F. C. 1991. Viola Group (Ordovician, Oklahoma) Cryptolithinid trilobites:

Biogeography and taxonomy. Journal of Paleontology, 63: 919-935.

SHEEHAN, P. M. AND P. J. LESPÉRANCE 1979. Late Ordovician (Ashgillian)

brachiopods from the Percé region of Québec. Journal of Paleontology, 53: 950-

967.

STIGALL, A. L., 2010a. Using GIS to assess the biogeographic impact of species

invasions on native brachiopods during the Richmondian Invasion in the Type-

Cincinnatian (Late Ordovician, Cincinnati Region). Palaeontologia Electronica,

13(1).

STIGALL, A. L., 2010b. Speciation decline during the Late Devonian Biodiversity Crisis

related to species invasions. PloS ONE, 5(12): e15584.

STIGALL RODE, A. L. 2005. Systematic revision of the Middle and Late Devonian

brachiopods Schizophoria (Schizophoria ) and ‘Schuchertella’ from North

America. Journal of Systematic Palaeontology, 3: 133-167.

SWIDERSKI, D. L., M. L. ZELDITCH, AND W. L. FINK. 1998. Why morphometrics is

not special: Coding quantitative data for phylogenetic analysis. Systematic

Biology, 47: 508-519.

SWOFFORD, D. L. 2002. PAUP*. Phylogenetic analysis using parsimony* 4.0. Sinauer,

Sunderland, Massachusetts.

TWENHOFEL, W. H. 1914. The Anticosti Island faunas. Geological Survey of Canada,

Museum Bulletin, 3, 39 p. 160

TWENHOFEL, W. H. 1928. Geology of Anticosti Island, Geological Survey of Canada

Memoir, 154, 481 p.

ULRICH, E. O. AND G. A. COOPER. 1942. New genera of Ordovician brachiopods.

Journal of Paleontology, 16: 620-626.

WALKER, L. G. 1982. The brachiopod genus Hebertella, Dalmanella, and Heterothina

from the Ordovician of Kentucky. U. S. Geological Survey Paper 1066-M, 17 p.

WANG, Y. 1949. Maquoketa Brachiopoda of Iowa. Geological Society of America

Memoir, 42, 55 p.

WEBBY, B. D., R. J. ELIAS, G. A. YOUNG, BJÖRN E. E. NEUMAN, AND D.

KALJO. 2004. Corals, p. 124-146. In B. D. Webby, F. Paris, M. L. Droser, and I.

G. Percival (eds.), The Great Ordovician Biodiversification Event Columbia

University Press.

WIENS, J. J. 2003. Incomplete taxa, incomplete characters, and phylogenetic accuracy:

Is there a missing data problem? Journal of Vertebrate Paleontology, 23:297-310.

WILEY, E. O. AND B. S. LIEBERMAN. 2011. Phylogenetics: Theory and Practice of

Phylogenetic Systematics (second edition). John Wiley & Sons, New Jersey, 406

WILLIAMS, A., AND D. A. T. HARPER. 2000. Order Orthida, p. H733-H734 In R. L.

Kaesler (ed), Treatise on Invertebrate Paleontology. Pt. H. Brachiopoda (revised)

3. Geological Society of America and University of Kansas Press, Lawrence. 161

WILLIAMS, A. AND A. D. WRIGHT 1965. Orthida, p. 299-395 In R. C. Moore (ed),

Treatise on Invertebrate Paleontology. Pt. H 1, Geological Society of America and

University of Kansas Press, Lawrence.

WILSON, A. E. 1926. An Upper Ordovician fauna from the Rocky Mountains, British

Columbia. Canada Geological Survey Museum Bulletin, 44: 1-34.

WILSON, C. W. 1949. Pre-Chattanooga stratigraphy in central Tennessee. Tennessee

Division of Geology Bulletin 56: 1-407.

WINCHELL, N. H. 1881. New brachiopoda from the Trenton and Hudson River

formations in Minnesota, Minnesota Geological and Natural Hisotry Survey

Annual Report, 9: 115-122.

WINCHELL, N. H. AND C. SCHUCHERT. 1892. Preliminary descriptions of new

Brachiopoda from the Trenton and Hudson River groups of Minnesota. American

Geologist, 9: 284-294.

WINCHELL, N. H. AND C. SCHUCHERT. 1893. The Lower Silurian Brachiopoda of

Minnesota. Geological and Natural history Survey Final Report, 3: 333-374.

WITZKE, B. J. 1980. Middle and Upper Ordovician paleogeography of the region

bordering the transcontinental arch, p. 1-18. In T. D. Fouch and E. R. Magathan

(eds.), Paleozoic paleogeography of the west-central United States: Rocky

Mountain Section, SEPM, West-Central Unites States Paleogeography

Symposium 1.

WRIGHT, A. D. 1964. The fauna of the Portrane Limestone, II. Bulletin of the British

Museum of Natural History, Geology, 9: 159-256. 162

WRIGHT, D. F. AND A. L. STIGALL. 2012. Paleobiogeography of the Late Ordovician

Richmondian Invasion: inferring source regions from phylogenetic patterns.

Geological Society of America Abstracts with Programs, 44 (5): 5.

WOODWARD, S. P. 1852. A manual of the Mollusca; or rudimentary treatise of recent

and fossil shells. London, 486 p.

ZHAN, R. AND J. JIN. 2005. Brachiopods from the Middle Ordovician Shihtzupu

Formation of Yunnan Province, China. Acta Palaeontologica Polonica, 50: 365-

393.

ZUYKOV, M. A. AND S. H. BUTTS. 2008. Glyptorthis (Foerste, 1914) and Bassettella

new genus (Brachiopoda: Orthida) from the Late Ordovician of the East Baltic.

Journal of Paleontology, 82: 197-200.

! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

Thesis and Dissertation Services ! !