A Statistical Analysis of Marine Mammal Dispersal Routes Across Major Ocean Regions Using Beta Diversity at the Generic Level

Home , Anthracobunidae, Dusisiren, Enhydriodon, Enhydritherium, Hydrodamalis

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science at George Mason University

Carlos Mauricio Peredo Bachelor of Science Seton Hill University, 2012

Director: Mark D. Uhen, Assistant Professor Department of Atmospheric, Oceanic and Earth Sciences

Spring Semester 2015 George Mason University Fairfax, VA

DEDICATION

Dedicated to my wonderful parents, Mauricio and Julie Peredo, who left behind everything they knew and started fresh in a foreign land purely in the pursuit of a better life for their children; to my older brother Miguel, whose witty humor, eternal optimism, and fierce loyalty has kept my head above water and a smile on my face throughout countless tribulations; to my younger brother Julio, who has far surpassed us all in talent and intellect, and who inspires me to never stop learning; and most of all, to my loving wife Molly, who has never stopped believing in me and drives me to settle for nothing less than perfection.

iii

ACKNOWLEDGEMENTS

I would like to thank my committee members, Drs. George, Lyons, and Parsons, for their tireless revisions and hard work on my behalf. I would like to thank George Mason University and the Smithsonian Institution for providing the support and inspiration for much of this project. I would like to thank the Paleobiology Database, and all of its contributors, for their ambitious vision and their relentless pursuit of its execution. I would like to thank my fellow members of the Uhen lab, who were always willing to advise, revise, and even commiserate. And above all, I would like to thank my advisor, Dr. Mark D. Uhen, who took a chance on a young man from a small university in rural Pennsylvania and helped mold the beginnings of his career.

TABLE OF CONTENTS

Page List of Tables ...... vi List of Figures ...... vii List of Equations ...... viii List of Abbreviations and Symbols ...... ix Abstract ...... x Introduction ...... 1 Hypotheses ...... 23 Procedure/Methods ...... 25 Results ...... 30 Discussion ...... 37 Conclusions ...... 56 Appendix I ...... 58 Appendix II ...... 74 Works Cited ...... 82

LIST OF TABLES

Table Page 1. Breakdown of dataset by taxa ...... ……30 2. Theoretical dataset used to test Alroy’s F’ ...... 38 3. Overlapping genera across the Pacific in the Oligocene and Miocene ...... 42 4. Overlapping genera across the Pacific in the Pliocene ...... 44 5. Overlapping genera across the Strait of Gibraltar ...... 46 6. Overlapping genera across the Central American Seaway ...... 48 7. Summary of the pull of the recent results ...... 52

LIST OF FIGURES

Figure Page 1. Prorastomids of the Middle Eocene ...... ……8 2. Evolution of Amazon Drainage Basin ...... 11 3. Breakdown of Types of Biodiversity ...... 17 4. Division of Ocean Regions ...... 29 5. Number of Marine Mammal Occurrences by Ocean Region ...... 31 6. Global Origination and Extinction ...... 32 7. Similarity Values in the Northern Hemisphere through Time ...... 34 8. Correlation of Alroy’s F’ to SD ...... 36 9. Correlation of Alroy’s F’ to SD for a theoretical dataset ...... 39 10. Correlation of Alroy’s F’ to SD for a theoretical dataset with outliers removed .... 40 11. Similarity values in the MET ...... 45 12. Similarity values between NWA and NEP ...... 47 13. Similarity between NWA, NEP, and ARC ...... 49 14. Breakdown of the number of species in each genus ...... 51

vii

LIST OF EQUATIONS PAGE

1. Jaccard Index ...... ……19 2. Sørensen-Dice Coefficient ...... 19 3. Jaccard:Sørensen-Dice Equivalency ...... 20 4. Original Forbes Index ...... 21 5. Alroy’s Corrected Forbes Index ...... 22

viii

ABBREVIATIONS AND SYMBOLS PAGE

Alroy’s Corrected Forbes Index ...... F’ Antarctic Ocean ...... ANT Arctic Ocean ...... ARC Jaccard Index ...... J East Indian Ocean ...... EIO Mean ...... µ Mediterranean-Tethys Ocean ...... MET Million years ago ...... Ma Northeast Atlantic Ocean ...... NEA Northeast Pacific Ocean ...... NEP Northwest Atlantic Ocean ...... NWA Northwest Pacific Ocean ...... NWP Southeast Atlantic Ocean ...... SEA Southeast Pacific Ocean ...... SEP Southwest Atlantic Ocean ...... SWA Southwest Pacific Ocean ...... SWP Standard Deviation ...... σ Sørensen-Dice Coefficient ...... SD

ABSTRACT

A STATISTICAL ANALYSIS OF MARINE MAMMAL DISPERSAL ROUTES ACROSS MAJOR OCEAN REGIONS USING BETA DIVERSITY AT THE GENERIC LEVEL

Carlos Mauricio Peredo, M. S.

George Mason University, 2015

Thesis Director: Dr. Mark D. Uhen

Using the Paleobiology Database, the distribution of marine mammal genera was analyzed across geologic time in an effort to isolate dispersal routes. Measurements of beta diversity were used to quantify the overlap between different ocean basins at different points in time. A recently introduced measurement of overlap was analyzed and found to be highly correlated with traditional methods, although conditions causing a deviation from this correlation are presented. Overlap is used to test existing marine mammal dispersal hypotheses. The Strait of Gibraltar was found to play a significant role in the dispersal of cetaceans and sirenians. Conversely, the Central American Seaway was found to play only a minor role in the high overlap values between the Atlantic and

Pacific oceans seen in extant taxa. Instead, it is asserted that this overlap is largely the result of Arctic dispersal pathways in the Quaternary.

INTRODUCTION

MARINE MAMMAL ORIGINS AND DISPERSAL

Mammals have returned to a predominantly marine lifestyle in at least seven separate events. These include Ursus maritimus (Polar bear), marine mustelids (Sea otters), Thalassocnus (aquatic sloth), Desmostylia, Sirenia (manatees and dugongs),

Pinnipedimorpha (seals, sea lions, and walruses), and Cetacea (whales, dolphins, and porpoises). The return to a marine environment has prompted notable changes in morphology that have facilitated the occupation of a wide variety of niche space among marine mammals (Cooper, 2009; Dawson, 1993; Deméré et al., 2003; Deméré et al.,

2008; Fahlke et al., 2013; Koepfli et al., 2008; Marx, 2009; Uhen, 2007; Uhen and

Pyenson, 2002). Despite the extensive work done on the morphology of marine mammals, studies on the points of origin and subsequent dispersal of each taxonomic remain inconsistent. Several different hypotheses exist for cetaceans (Gingerich and

Uhen, 1996, 1998; Harington, 2008), sirenians (Domning, 1978, 1982, 2001), pinnipedomorphs (Deméré et al., 2003; Koretsky, 2002a), polar bears (Cronin et al.,

2014; Kurtén, 1964), and sea otters (Berta and Morgan, 1985). This work will enrich the existing literature and provide a new perspective on existing hypotheses. In this thesis I outline the existing hypotheses for marine mammal dispersal routes and test them using ecological methods of measuring beta diversity.

Polar Bears

Ursus maritimus is a species of bear that split from the lineage including Ursus arctos (brown bear) and Ursus americanus (black bear), though the timeframe for this remains poorly understood. The fossil record indicates that a brown bear + polar bear lineage split from the black bear lineage approximately two million years ago (Kurtén,

1968; Wayne et al., 1991). Unfortunately, molecular studies attempting to bolster this hypothesis have varied widely, suggesting a split anywhere from 0.6 – 6.7 Ma (Goldman et al., 1989; Hailer et al., 2012; Lindqvist et al., 2010; Waits et al., 1999; Yu et al., 2004;

Yu et al., 2007). This wide variation in molecular divergence dates makes it difficult to interpret in context with the fossil record.

The tenuous disagreement between the fossil record and the molecular data makes it even more difficult to place the subsequent splitting of this lineage into brown bears and polar bears. Kurtén put forth that polar bears and brown bears evolved from a common ancestor in the Pleistocene (Kurtén, 1964, 1968), a hypothesis that is supported by the fossil record (Alexanderson et al., 2013). Unfortunately, the molecular dates again varies widely depending on the genetic markers used. These estimates range from as recent as 0.11-1.7 million years ago (Yu et al., 2004; Yu et al., 2007) to as far back as 4-5 million years ago (Miller et al., 2012)

This disagreement amongst the different interpretations of the molecular data can partially be attributed to interbreeding between brown bears and polar bears following their original speciation (Cronin et al., 2014). Brown bears from Admiralty, Baranof, and

Chichagof islands (known as ABC bears) have been found to share more of their nuclear genome with polar bears (Miller et al., 2012). This hybridization is hypothesized to be partially responsible for the variability in molecular analyses. The most recent study runs the molecular data of ABC bears separately and concludes that brown bears and polar bears split from a common ancestor approximately 1.2 million years ago (Cronin et al.,

2014).

Unfortunately, the relatively small number of polar bear fossils has made study of their dispersal difficult. With the exception of one specimen in the United Kingdom and one in Northern Germany, all other fossil occurrences have an arctic distribution (Kurtén,

1964). It remains unclear whether this dispersal south into Europe is the only dispersal event by polar bears outside of the arctic.

Sea Otters

Enhydra lutris is a species of fully marine otter endemic to the North Pacific and is the only extant species of the genus Enhydra. The genus is thought to have split from other otters five million years ago (Koepfli et al., 2008). Two million years ago, the genus became isolated in the North Pacific. This isolation gave rise to the species

Enhydra macrodonta (now extinct) and Enhydra lutris (Love, 1992). A third species,

Enhydra reevei, has been named based on dental samples from the United Kingdom

(Willemsen, 1989, 1992).

Two other genera of otter have been speculated as being related to Enhydra,

Enhydritherium and Enhydriodon. Enhydriodon is a wide-ranging genus found in east

3 and central Africa, India and Pakistan, and Southeast Asia (Geraads et al., 2011;

Repenning, 1976; Verma, 1992; Villalta Cornella and Crusafont Pairo, 1945). However, as all of the fossils have been recovered from terrestrial deposits, the genus is not included in this study. Enhydritherium was named in 1985 from specimens found in

Florida and California (Berta and Morgan, 1985; Lambert, 1997). Berta and Morgan

(1985) noted affinities of the genus to other specimens from Spain and hypothesize dispersal to North America from Eurasia, either through a Bering land bridge or across the North Atlantic.

This unusual biogeography is exacerbated by the presence of two specimens of

Enhydra reevei from the United Kingdom. The first, (Newton, 1890), has been lost, although a cast is located in the British Museum (Natural History) (Willemsen, 1989,

1992). It has been suggested that this specimen more accurately belongs to Enhydriodon

(Repenning, 1976). As Repenning was at the time neither aware of Enhydritherium in

North America, nor of the second specimen found twelve years later, his reevaluation of the first specimen led to a hypothesis where the Eurasian Enhydriodon gave rise to the

Pacific Enhydra. Willemsen, however, confidently reassigns this specimen to Enhydra and chooses not to comment on the biogeography of the taxa (Willemsen, 1992).

Thus, the origin and dispersal of sea otters remains poorly understood. Two distinct groups dominate in the Late Miocene and Early Pliocene: Enhydritherium in

North America and the seemingly terrestrial Enhydriodon in Africa and Eurasia (Berta and Morgan, 1985; Geraads et al., 2011; Lambert, 1997; Verma, 1992; Villalta Cornella and Crusafont Pairo, 1945). Enhydra appears in the Pleistocene in the North Pacific, and

4 is also mysteriously present in the United Kingdom. It remains unclear which of the two, if any, Miocene groups gives rise to the North Pacific population of Enhydra (Willemsen,

1992).

A separate lineage of otter, the Marine otter of Chile, is unrelated to the Sea Otter lineage and is not included in this study. This lineage has only a single species, and, similar to some Northern European river otters that occasionally venture into marine environments, is less specialized and not exclusively marine (Larivière, 1998).

Thalassocnus

The genus Thalassocnus is an aquatic sloth that inhabited shallow marine environments (Canto et al., 2008). Moreover, it is not thought to have ever dispersed out of the southeastern Pacific. To date, Thalassocnus is known only from Peru and Chile.

Five species are recognized today, and they succeed one another in time with little overlap beginning in the Late Miocene (T. antiqus) and ranging through the Late Pliocene

(T. yaucensis). No evidence has ever been found to suggest that Thalassocnus dispersed beyond the Southeast Pacific (Canto et al., 2008).

Desmostylia

Desmostylia are a group of semi-aquatic herbivorous marine mammals thought to be related to Sirenia and Proboscidea and are endemic to the North Pacific. They were originally placed within Sirenia, but were reassigned as sister taxa to Sirenia and

Proboscidea in 1953 (Reinhart, 1953). This placement has been widely accepted (though

5 rarely, if ever, tested) and appears in the literature frequently (Domning et al., 1986;

Fischer and Tassy, 1993; McKenna, 1975 among many others). In 1975, the name

Tethytheria was proposed to include Sirenia, Desmostylia, Proboscidea, and

Embrithopoda (McKenna, 1975).

Recently, however, this phylogeny has come into question. A phylogenetic analysis originally undertaken to reassess the placement of Anthracobunidae as ancestral to Tethytheria incidentally consistently grouped Desmostylia as stem perissodactyls

(Cooper et al., 2014). While the authors are cautious in their interpretation of the results, they raise an interesting possibility that calls into question the phylogenetic position of

Desmostylia that have as yet lacked rigorous testing. However, both stem perissodactyls and tethytherians are interpreted as having an origin near the Tethys Sea. Thus, regardless of their phylogenetic position, their dispersal path from the Tethys to the Pacific remains unknown.

Desmostylia first appear in the fossil record in the Late Oligocene and persist through the Late Miocene (Barnes, 2013). The order is restricted to the margins of the

North Pacific Ocean (Barnes, 2013). The oldest specimens of Desmostylia are in the

Rupelian of Oregon and Washington, suggesting a Northeast Pacific origin and subsequent dispersal along the Aleutian-Kamchatka archipelago to the Northwest Pacific

(Barnes, 2013; Inuzuka, 2000). Because both basal perissodactyls and tethytheres are thought to have an old world origin, the dispersal of Desmostylia from the old world to the North Pacific is a great mystery.

Sirenia

Sirenia is an order of herbivorous marine mammals that includes modern manatees and dugongs. There are only four extant species (three in the manatee genus

Trichechus and one in the dugong genus Dugong), though a fifth (Hydrodamalis gigas) was extant as recently as the eighteenth century (Crerar et al., 2014; Domning, 1982).

As mentioned, Sirenia have long been considered members of Tethytheria along with Proboscidea and Desmostylia (McKenna, 1975). Though this phylogeny has come into question for Desmostylia, the position of Sirenia is still supported (Cooper et al.,

2014). Tethytheria is considered to originate on the shores of the Tethys Sea, and sirenians are presumed to have originated in North Africa. Despite this, for many years the oldest fossil Sirenia were known from the middle Eocene of Jamaica. This juxtaposition between an old world origin but new world fossils implied a gap in the fossil record (Domning, 2001).

This conflict was mitigated with the discovery of a prorastomid from the Eocene of Senegal (Hautier et al., 2012). The presence of an early sirenian in Senegal confirmed that they inhabited the old world African shores and lead to two hypotheses for their dispersal to the new world. One hypothesis involves the sirenians following the continuous, warm climate North Atlantic shoreline in the same fashion as other Eocene land mammals (McKenna, 1975). This hypothesis is favored because it implies that the only semi-aquatic sirenians would not have needed to cross wide or deep water.

However, the lack of sirenians in Eocene deposits from the east coast of the United States makes this unlikely. Instead, Hautier et al. (2012) favor a hypothesis similar to the one

7 put forth for whales by Uhen (1999) involving a direct transatlantic dispersal from West

Africa to South America (Figure 1).

Figure 1: Prorastomids of the Middle Eocene. Hypotheses by McKenna (1975) and Geisler et al. (2005) predict a dispersal to the new world via the continuous coastline in the far North Atlantic. An alternative hypothesis by Uhen (1999) instead puts forth a direct dispersal across the Atlantic from western Africa to eastern South America. Map courtesy of fossilworks.org, copyright John Alroy.

By the end of the Eocene, sirenians had dispersed from their African origin to inhabit Europe, and the east coasts of North, Central, and South America (Domning,

2001). In the Early Miocene sirenians entered the Pacific through the Central American

Seaway and colonized the west coasts of South and North America as well (Domning,

1976). Domning (1976) leaves little doubt that this dispersal happened through the

Central American Seaway, citing the established populations in the Caribbean and the warm climate of California in the early Miocene as an ideal habitat for Sirenia.

Curiously, however, the Pacific sirenians develop very unique morphological characters not seen in any Atlantic forms, indicating little gene flow between the two.

Domning infers the emergence of an ecological barrier shortly after the dispersal, isolating the Pacific and the Atlantic sirenians (Domning, 1976, 1978). Domning notes that, while the California coasts would serve as ideal habitat, the Mexican and South

American coasts were likely too exposed and too warm for abundant growth of kelp as expected in the California coasts (Domning, 1976). Domning terms this the “Mexican barrier”, and suggests that, after the original dispersal event to California, the Central

American Seaway would have become an uninviting route that would have discouraged subsequent dispersal events for Sirenia (Domning, 1976, 1978). A marine regression in the Pliocene resulted in the Pacific Honduran and Mexican coasts becoming rocky and uninhabitable to sirenians, reinforcing the significance of this “Mexican barrier”

(Domning, 1978).

This regression in the Pliocene is coupled by an orogenic event in the Andes

(Domning, 1978; Hynek, 2011). The orogenic event is considered the trigger for the

9 evolution of modern trichechids (manatees) (Domning, 1976, 1978, 1982, 2005). Until this point, trichechids had been restricted in habitat to coastal rivers and estuaries of

South America, while dugongids (dugongs) inhabited the marine waters of the Atlantic and the Caribbean (Domning, 1982). The Pliocene orogenic event dumped massive amounts of silt and dissolved nutrients into the South American river systems (Figure 2); this in turn triggered the evolution of supernumerary, continuously replaced molars in manatees (Domning, 1982, 2005). The combination of this added stress in the food supply, as well the competition from the better adapted manatees, resulted in the extinction of Caribbean and western Atlantic dugongs (Domning, 1982).

By the end of the Pliocene, the Central American Seaway closed entirely and fully isolated the Pacific dugongs from the Atlantic manatees (Domning, 1978). Near the

Pliocene-Pleistocene boundary, manatees dispersed back across the Atlantic to Africa and very quickly replaced the remaining, eastern Atlantic dugongs in the fossil record

(Domning, 1982). In the Pacific, dugongs continued to disperse north, eventually leading to the hydrodamaline lineage becoming the only known sirenians to ever adapt to cold water environments (Domning, 1976; Takahasi et al., 1986). This lineage, including

Dusisiren and Hydrodamalis, includes records on both halves of the Pacific.

Hydrodamalis first appeared in the Northeast Pacific and then dispersed around the

Aleutian-Kamchatka archipelago to the Northwest Pacific before it was hunted to extinction in the 18th century (Domning, 1978, 1982).

Figure 2: Schematic diagrams showing the evolution of the Amazonian drainage basin (Domning, 1982).

During the Middle Miocene (A), the drainage was equally split westward into the Pacific and eastward into the Atlantic. During the Early Pliocene (B) the Pacific drainage had closed, creating an enclosed basin ideal for early manatees. By the Plio-Pleistocene (C) modern drainage patterns appeared and manatees dispersed north and east to Africa and the eastern United States. P indicates the presence of Potamosiren in the

Magdalena basin. (T) denotes Trichechus-like sirenains, where T denotes Trichechus,

Pinnipedimorpha

Pinnipedimorpha is a monophyletic group that is nested within Caniformia within

Carnivora. Pinnipedimorpha includes three main groups: Otariidae (fur seals and sea lions), Phocidae (true seals), and Odobenidae (walruses) as well as a handful of ancestral, extinct forms (Davies, 1958; Deméré et al., 2003). Due to a poor fossil record, the time

11 and location of origin for Pinnipedimorpha is poorly understood. A single specimen has been said to represent an Eocene phocid (Diedrich, 2011), though this is highly unlikely.

This specimen is nothing more than the flattened head of a single, unidentifiable femur.

Moreover, Diedrich (2011) claims the sediment has been dated to the Ypresian or

Lutetian, but provides no citation.

Similarly, two femora from the Oligocene of South Carolina have been identified as phocids (Koretsky, 2002a). These specimens have some derived traits indicative of phocids, but also share several primitive traits with the otters Enhydra and Lutra

(Koretsky, 2002a). Here, they are regarded as merely indeterminate carnivorans, as they share an equal number of traits with phocids and mustelids. It has been suggested that these specimens, in conjunction with another questionable old world fossil (Koretsky,

2002e), may point to polyphyletic origin of Pinnipedimorpha with an Atlantic origin for

Phocidae (Arnason et al., 2006; Koretsky, 2002a, e). This diphyletic origin has gained little support in the literature and largely disagrees with molecular evidence (Arnason et al., 2006; Arnason and Widegren, 1986; Berta and Wyss, 1994; Dasmahapatra et al.,

2009).

The oldest fossils that can be confidently attributed to Pinnipedimorpha are those of Enaliarctos barnesi and Enaliarctos tedfordi, which appear in the Late Oligocene of

California (Berta, 2009; Berta et al., 1989; Deméré et al., 2003). Pinnipedimorpha are thought to remain restricted in the North Pacific during the Early Miocene, when all three major groups arise before dispersing out of the North Pacific (Deméré et al., 2003).

Deméré et al. (2003) considered the otariids to split from ancestral pinnipedimorphs eleven million years ago, when they quickly dispersed to either side of the North Pacific. This date was pushed back with the discovery of an otariid from Japan dated around 12 Ma and implying a ghost lineage even further back (Kohno et al., 2007).

A very recent discovery of an otariid dated to between 15 and 17.1 Ma eliminates the ghost lineage (Boessenecker and Churchill, 2015).

Deméré et al. (2003) hypothesized a Southern Hemisphere dispersal event near the end of the Pliocene, in which otariids traveled down the west coasts of North and

South America, passed through Drake’s Passage into the Atlantic coast of South America and diversified in the Antarctic, achieving a circumpolar distribution. Finally, they hypothesized a Pleistocene reinvasion event back up the Pacific coasts of the Americas by the now derived southern otariids (Deméré et al., 2003).

Deméré et al. (2003) considered the origin of Odobenidae to occur before 18 Ma.

They hypothesize that, after dispersing to either side of the North Pacific, walruses diversified and dispersed up into the Arctic and down to either side of the Atlantic in the

Late Pliocene (Deméré et al., 2003).

Deméré et al. (2003) considered the phocids split within Pinnipedimorpha to occur before 18 Ma as well. They hypothesize an early, major dispersal in which phocids travelled through the Central American Seaway, down the African and South American coasts and through Antarctica to achieve a circumpolar distribution. They hypothesize that these Atlantic phocids then dispersed up into and around the Arctic, followed by several events during the Pleistocene glaciation fluctuations back down to lower latitudes

13 in which they reinvaded the North Pacific, the North Atlantic, and the Eurasian continent

(Deméré et al., 2003).

Cetacea

The name Cetacea refers to a group of marine mammals within Artiodactyla, most closely related to Hippopotamidae (Geisler and Uhen, 2003, 2005). The oldest cetacean is from the Ypresian (about 52.5 Ma). There is some debate over whether or not the oldest whale is Himalayacetus (Bajpai and Gingerich, 1998) or Pakicetus (Thewissen et al.,

2001). Regardless, these two indicate an origin in the Tethys Sea (Uhen et al., 2010). In the early Eocene, only these earliest whales, in the family Pakicetidae, were present and had not yet left the Tethys Sea. Pakicetidae is followed by two other families of Tethyan archaeocetes, Ambulocetidae and Remingtoncetidae in the Middle Eocene. However, by the middle Eocene, Protocetidae and Basilosauridae (more derived archaeocetes) had dispersed out of the Tethys and reached Europe, Africa, and the Americas (Geisler et al.,

2005; Uhen, 1999; Uhen and Berndt, 2008; Uhen et al., 2010; Uhen et al., 2011). The most derived of these, Basilosauridae, includes the first fully aquatic cetaceans and is the only archaeocete family that survived to the late Eocene (Gingerich and Uhen, 1998;

Uhen, 2010).

It is thought that around the Eocene-Oligocene boundary, the dominant cetacean fauna transitions from archaeocetes to Neoceti, the group that includes the two extant cetacean suborders (Mysticeti and Odontoceti). This transition is poorly understood because the fossil record at the Eocene-Oligocene boundary is poor (Uhen and Pyenson,

2007). Neoceti derived from Basilosauridae and each suborder is presumed monophyletic

(Geisler and Sanders, 2003).

The oldest fossil mysticete is from the latest Eocene of Antarctica (Mitchell,

1989), though by the Oligocene they are found throughout most of the Pacific

(Fitzgerald, 2004; Fitzgerald, 2006, 2010) as well as the eastern United States (Sanders and Barnes, 2002). Odontocetes appear in the fossil record quite suddenly in the Rupelian in both the Pacific (Fordyce, 2002) and the Atlantic (Sanders and Barnes, 2002; Uhen,

2008). Thus, the diversification and subsequent dispersal of both Odontoceti and

Mysticeti is too sudden to be properly captured by the relatively poor fossil record of the

Eocene-Oligocene boundary (Uhen, 2010).

BETA DIVERSITY AND MEASURING DISPERSAL

Measures of Biodiversity

In order to accurately assess the dispersal of one group among regions, it becomes necessary to have a quantifiable measurement. Whittaker defined the three main components of diversity, as well as one smaller one (Whittaker, 1960, 1972). Alpha diversity is the total number of taxa present in a given community. Beta diversity is the number of unique taxa between sites or communities, and gamma diversity is a combination between the two. Whittaker also mentions a “delta diversity”, which he defined as the number of unique taxa between larger, geographic regions. In 1988

Sepkoski embraced Whittaker’s definitions of alpha and beta diversity, but adopted the definition of delta diversity as gamma diversity and does away with delta diversity

15 entirely (Sepkoski, 1988). In 2002, Hunter returned to Whittaker’s definitions and described gamma diversity as the overall measure of diversity for different ecosystems in a region (Hunter, 2002; Hunter and Gibbs, 2009). Thus, the components of diversity as they stand now are:

• Alpha Diversity – Total number of taxa

• Beta Diversity – Number of unique taxa across ecosystems

• Delta Diversity* – Number of unique taxa across geographic regions

• Gamma Diversity – Overall diversity for different ecosystems within a region

*Delta diversity is considered a subset of beta diversity, as the size of an ecosystem and the size of a geographic region are highly variable.

Beta diversity measures the number of unique taxa across two (or more) regions.

High beta diversity means that two regions have very different taxonomic distributions; very few of the taxa overlap, or appear in both regions. Similarly, low beta diversity indicates that the two regions have many of the same taxa present. Thus, beta diversity provides a better understanding of the taxonomic richness than does alpha diversity alone. Two regions can have a great number of total taxa (and thus have high alpha diversity), however, if a large amount of those taxa are present in both regions, than there is high overlap, and thus very low beta diversity.

Figure 3: Relationship between alpha (local), gamma (regional), and beta (similarity) diversity. Alpha diversity measures the number of unique taxa within a given region. Beta diversity measures the number of similar taxa across two regions. Gamma diversity measures the overall diversity for all of the regions in question.

Beta diversity is particularly useful as it provides three primary insights. Beta diversity indicates the degree to which habitats have been partitioned by taxa, allows for the comparison of diversity between regions, and can be combined with alpha diversity to assess the overall diversity of a region (Wilson and Shmida, 1984).

Because beta diversity measures the number of unique taxa across regions, its inverse (taxonomic overlap) can be used to measure dispersal. High overlap between regions will indicate strong similarity between their faunal compositions, implying a greater likelihood of dispersal across the regions by the aforementioned fauna.

This is crucial, as previous studies have measured dispersal only through presence-absence data (Webb, 2006; Woodburne, 2010). Dispersal in the fossil record is typically measured by noting the first occurrence of each taxon in a given region. In this way, regions have the amount of dispersal compared as a function of alpha diversity. In

Webb’s study, the total number of taxa that disperse from North America to South

America are compared to the total number of taxa that disperse from South America to

North America. The total number of taxa that have dispersed to a region are added (alpha diversity) and compared to the total number of taxa that have dispersed to another region

(Webb, 2006).

This method provides a good overview of where dispersal is taking place, but it fails to account for the existing taxonomic make up; it relies heavily on the initial populations being of similar size and diversity. Measuring dispersal instead as a function of beta diversity (with high beta diversity indicating little overlap, and thus little dispersal) adjusts for the original population make ups. Using beta diversity also allows for the measurement of any diversification (shown by a subsequent decrease in overlap) following a dispersal (shown by an increase in overlap) event. The challenge lies in finding reliable measures of beta diversity.

Measures of Beta Diversity

Three of the most common measurements of beta diversity are the Sørensen-Dice

Coefficient, the Jaccard Index, and a pure probability (the probability that 2 specimens chosen at random are the same taxon) (Condit et al., 2002).

The Jaccard Index, developed by Paul Jaccard in 1901, measures beta diversity by dividing the taxa in common by the difference of the taxa in each group and the taxa in common (Jaccard, 1901):

� � = ! �! − �! − �!

where J is the Jaccard Index, Tc is the total taxa in common, and T1 and T2 are the total taxa in region one and two respectively.

The Sørensen-Dice coefficient, sometimes just called the Sørensen Index, was developed independently by Lee Dice in 1945 and Thorvald Sørensen in 1948. This index multiplies the taxa in common by two, and then divides the product by the sum of the taxa in each group (Dice, 1945; Sørensen, 1948). Its formula is as follows:

2 ∗ � �� = ! �! + �!

where SD is the Sørensen-Dice Coefficient, Tc is the total taxa in common, and T1 and T2 are the total taxa in region one and two respectively. Sørensen-Dice Coefficient thus is an inverse measurement of beta diversity, as it measures the amount of overlap between two regions, not the amount of difference.

The Sørensen-Dice Coefficient and the Jaccard Index were seen as alternate, viable methods of assessing beta diversity. In 1988, however, Sepkoski pointed out that

19 the two are actually redundant. It can be shown that the Sørensen-Dice Coefficient is actually twice the Jaccard Index divided by the Jaccard Index plus one. This is shown as:

2 ∗ � �� = � + 1

where SD is the Sørensen-Dice Coefficient and J is the Jaccard Index.

Thus, the two indices are actually analogous and interchangeable. They are, nevertheless, still useful for assessing dispersal. The Sepkoski study used beta diversity to analyze faunal dispersal across six environmental habitats from the Cambrian through the

Cenozoic. However, as Sepkoski pointed out, it is dangerous to use beta diversity in the fossil record to make comparisons through time. It becomes necessary to also assess origination and extinction, as unusually high rates of either can skew the similarity values.

High rates of origination will increase beta diversity (decrease similarity) because all taxa must originate at a single location. If a single region experiences unusually high origination rates in a single time bin, the result will be that it will be less similar to any other region. This is also the case for vicariance, the process by which the geographic range of a taxon is split by a geographic barrier and a physical barrier to gene flow forms.

This phenomenon will result in increased beta diversity, as it can only create less similarity, not more. Fortunately, at the spatial and temporal scale used, there are no significant physical barriers formed expected to result in vicariance.

High rates of extinction, in contrast to origination, have a less predictable effect on the data. Extinction can either lower, or drive similarity, depending on which taxa are affected. If most of the cosmopolitan taxa go extinct, leaving primarily endemic taxa, global similarity will be very low. In contrast, if many endemic taxa go extinct, leaving primarily those with broad or cosmopolitan ranges, then global similarity will be very high. Therefore, to properly interpret the data, it is necessary to quantify both origination and extinction rates so that global similarity values can be placed into proper context

(Sepkoski, 1988).

Recently, an alternate, yet forgotten, method of assessing similarity has resurfaced

(Alroy, 2015). This formula, originally proposed in 1907 by Stephen Forbes, has been largely ignored because it requires knowing the number of taxa absent in each sample being compared. As this is an unrealistic expectation in the fossil record, Forbes’s method was largely ignored in the literature in favor of the aforementioned Sørensen-

Dice coefficient or the Jaccard index (Alroy, 2015).

The original Forbes equation for similarity was calculated as:

� ∗ � � = � + � ∗ (� + �)

where a represents the number of taxa found in both samples, b the taxa in only the first sample, c the taxa in only the second, d the number of taxa found in neither sample, and

N the sum of a, b, c, and d. As mentioned, this formula is unique, if problematic, because

21 it includes the number of shared absences. Alroy adds a heuristic correction to Forbes’s original equation that allows the troublesome variable to be excluded.

Alroy sets d equal to zero, and instead uses n, which represents the sum of a, b, and c. Doing so allows for the equation to be more comparable to other indices. However,

Alroy notes that this comparison is still limited because F scales with an upwards bias. If there are twenty taxa between two regions, and ten are shared, then SD and J are each

0.5, but F is 0.75. Thus, Alroy adds the heuristic correction and establishes what he calls the corrected Forbes index (F’) (Alroy, 2015).

�(� + �) �! = � ∗ � (� + � ∗ (� + �)) + (� ∗ �) + 2

Alroy’s study indicates that the corrected Forbes index may outperform the

Sørensen-Dice coefficient when used on samples of regions with known overlap, and may be especially useful when the two samples exhibit a great size disparity or when sampling is incomplete.

This study quantifies similarity of faunal overlap between adjacent ocean regions through geologic time using both the Sørensen-Dice coefficient and the corrected Forbes index. It tests Alroy’s hypothesis that the corrected Forbes index outperforms the

Sørensen-Dice coefficient and also assess any correlation between the two. It also measures origination and extinction in the fossil record. Thus, it identifies time periods of faunal dispersal between regions and major geographic features abetting or inhibiting dispersal.

HYPOTHESES

Broadly, this study analyzes beta diversity and faunal turnover of marine mammals resulting in a quantitative comparison of notable patterns across different ocean regions and through geologic time. These patterns are used to evaluate the role of several geographic features in marine mammal dispersal. The following four hypotheses are tested:

• Hypothesis One – Overlap is greater between adjacent regions of the same ocean

(intraoceanic) than between regions of different ocean basins (interoceanic).

• Hypothesis Two – The Aleutian-Kamchatka archipelago is the primary pathway

by which marine mammals disperse across the Pacific.

• Hypothesis Three – The Strait of Gibraltar serves as the passageway through

which cetaceans and sirenians leave their point of origin in the Mediterranean-Tethys.

• Hypothesis Four – The Central American Seaway is the pathway through which

modern high overlap values between the Atlantic and Pacific are attained.

In addition to these four hypotheses, the study also looks to compare overlap values of Alroy’s corrected Forbes index to the traditional Sørensen-Dice coefficient and comment on any discontinuity between the two methods.

Recently, a new hypothesis on the age of closure of the Central American Seaway has gained traction in the literature. This new hypothesis has pushed back the established

23 date of closure of 3-4 Ma (Coates et al., 2004) to 15 Ma (Farris et al., 2011; Hoorn and

Flantua, 2015; Montes et al., 2012). The new model is still treated with healthy skepticism, and it has been suggested that the modern day Indonesian Australian

Archipelago may serve as an appropriate analog for what they Central American Seaway may have looked like in the aforementioned period (Coates and Stallard, 2013). In addition to the hypotheses mentioned above, this study specifically searches for evidence of an open seaway through the 15-3 Ma time span during which the new hypothesis proposes no such seaway occurs. Thus, this study also assess the viability of the new model for the Central American Seaway as described by Farris et al. (2011) and Hoorn and Flantua (2015).

METHODS

All data collected for this study were downloaded from the Paleobiology

Database on October 13, 2014. While it has been questioned whether database work could provide accurate analyses (Prothero, 2012, 2014), studies have shown that the results are sound so long as the dataset is representative of the fossil record (Alroy, 2003;

Peters and Heim, 2010; Peters et al., 2014; Plotnick and Wagner, 2006; Quirin, 2003).

The first stage of this project was to ensure these records serve as an accurate portrayal of the fossil record.

Efforts by researchers to include every marine mammal publication in the database have been exhaustive and largely successful. Based on these efforts the

Paleobiology Database’s data set is representative of the marine mammal fossils published in the literature (Alroy et al., 2003; Uhen, 2014; Uhen et al., 2013; Uhen and

Pyenson, 2007; Varela et al., 2014). While the published literature is not necessarily representative of the entire fossil record, this study restricted itself only to published occurrences because of the dubious taxonomic identification often associated with unpublished museum occurrences.

While the records for marine mammals were nearly exhaustive through the

Pliocene, they were underrepresented in the Quaternary. This was particularly the case

25 with occurrences near the Pleistocene-Holocene boundary, where the occurrences were scattered across a much wider range of literature (archaeological, stranding data, etc.).

In order to ensure that the records pertaining to the Quaternary were thorough and complete, I conducted an exhaustive review of the literature searching for entries that may have been omitted from the database. In doing so, I added 72 collections containing cetaceans, 11 collections containing pinnipeds, and 11 collections containing polar bears to the Paleobiology Database.

It also proved necessary to add an occurrence for each geographic region from which the extant fauna are known from today. The database had previously had a single collection added to the Holocene for each extant taxa; this served to anchor the extant taxa in time. However, these efforts had not been made for each ocean basin that a taxon appears in today. Thus, while efforts had been made to anchor the extant taxa in time, they were not previously anchored in space as well. The distribution of extant taxa were derived using aquamaps.org, an organization that models the distribution of marine taxa based on their potential suitable habitat and then corrects the range by excluding regions where the taxon is known not to occur. This model was originally pioneered specifically for marine mammals by Kaschner et al. (2006). From this, one occurrence was added to the Paleobiology Database for each region that each extant taxon appears in. This served to accurately capture the known modern distribution of extant taxa in the dataset.

Once the systematic addition of missing Quaternary data was finished, the next step was to download the occurrences and assign each occurrence an ocean region. Ocean regions were divided geographically: the Atlantic and Pacific were divided into

26 quadrants, the Indian Ocean divided vertically in half, and the Arctic, Antarctic, and

Mediterranean-Tethys Oceans represented single regions (see Figure 4). These thirteen regions served as the regions across which diversity and overlap were measured. These divisions are largely consistent with document S-23 titled “Limits of Oceans and Seas” published by the International Hydrographic Organization (IHO). The third edition of this document, published in 1953, is still the official reference. Any deviation from this document was done for the purpose of extrapolating the divisions put forth within document S-23 to their relevant locations throughout geologic time. One example of such a deviation is the Mediterranean Sea, which in this study was enlarged to include the

Tethys Sea as well (Fourcy and Lorvelec, 2012; IHO, 1953). The other notable deviation was the division between the Indian and Pacific oceans. This division was moved westward, from the southern most tip of Australia to the center of the continent, in an attempt to keep a collection of localities analogous in geologic time and faunal composition together in the same region.

In total, fifteen time bins were used, primarily analogous to the stage levels of the geologic time scale. The single deviation from this system was pertaining to the

Quaternary, which was treated as a single “stage”. This was done partly because it is a more comparable length of time to the other stages, and partly because few occurrences were identified to specific stages in the Quaternary.

Alpha diversity was measured both regionally as well as globally. Time bins exhibiting abnormal origination or extinction events were noted to properly place the similarity values into context.

Values for overlap were calculated for all regions having a large enough sample size. Due to the strong northern hemisphere bias in the dataset (figure 4), overlap for the southern hemisphere was calculated but not included in the analysis for the discussion.

Beta diversity was measured at each geologic stage from the Ypresian through the

Piacenzian; with the Quaternary treated as a single, final “stage” as it represents a comparable amount of absolute time. Beta diversity was measured using both the

Sørensen-Dice coefficient, as the literature agrees it is the most accurate method of doing so (Condit et al., 2002; Sepkoski, 1988), as well as Alroy’s newly corrected Forbes index

(Alroy, 2015).

The entire study was conducted at the genus level, as opposed to the species level, in an effort to use the most relevant taxonomic level. Uhen and Pyenson (2007) vetted the literature for cetaceans ( 70% of occurrences in the dataset) and synonymized names not based on diagnostic material. They eliminated nomenclature based on poor type material and left a rigorously critiqued taxonomy. Following their efforts, they conclude that the generic level is now free of bias attributed to bad taxonomic practice (Uhen and Pyenson,

2007).

Figure 4: Global map showing ocean regions as they were divided based on modern geography with consideration to paleogeography. Solid lines represent regions as delineated by the IHO (1953). Dashed lines demonstrate divisions used for this study where they deviate from already established IHO divisions.

In the case of the division between the Southwest Pacific and the East Indian Ocean, the boundary was moved west to keep an analogous collection of fossil localities together. Map courtesy of fossilworks.org, copyright John Alroy. Each dot represents an occurrence included in our dataset, throughout all time bins.

RESULTS

Alpha Diversity

In total, the dataset consisted of 2,991 occurrences that comprised 414 unique taxa. These data were distributed across 13 distinct ocean regions as demonstrated above in Figure 4. Both total occurrences and alpha diversity (unique taxa) was consistently skewed towards the five major northern hemisphere regions (MET, NEA, NWA, NEP,

NWP) as shown in Figure 5. The mean number of occurrences amongst these five regions was 464.8 (σ = 124), in contrast to 83.4 (σ = 40) for the other eight regions. Similarly, the mean number of unique taxa for these five regions was 116 (σ = 11), in contrast to only

44 (σ = 11.4) for the other eight regions. Due to the extreme concentration of occurrences and taxa into five specific regions, interpretation of overlap results was restricted to comparisons amongst the regions in question (the raw data available in Appendix I).

Table 1: Breakdown of all 2,991 occurrences by taxonomic group.

Occurrences Taxa Ursus maritimus 30 1 Marine mustelids 43 3 Thalassocnus 6 1 Desmostylia 54 8 Sirenia 357 31 Pinnipedimorpha 401 48 Cetacea 2100 322 Total 2991 414

Table 1 shows the breakdown of the entire dataset by taxonomic group. There is a strong bias towards cetaceans, both in total occurrences and in the number of unique taxa.

From the data presented in table 1, it is shown that cetaceans drive the majority of the results. Polar bears, sea otters, Thalassocnus, and Desmostylia appear in the dataset in such minor quantities that they are deemed to have little to no affect on the overall patterns in the data.

Occurrences of Marine Mammals by Region

700

600

500

400

300

200

100

0 NWA MET NEA NEP NWP SWP SWA SEA SEP ARC WIO EIO ANT

Figure 5: The number of occurrences of marine mammals in each of the thirteen ocean regions. Those shown in blue are the five regions of sufficient biodiversity to be included in the analysis. Those shown in red had insufficient diversity to be included in the analysis of the overlap data.

Figure 6: Global origination and extinction through time. Darkest blue represents the total number of taxa

(genera) present globally in each stage. Middle blue represents origination, measured as the number of taxa whose first appearance is recorded in each given stage. Lightest blue represents extinction, measured as the number of taxa whose last appearance is recorded in each given stage.

Figure 6 demonstrates the global origination and extinction and is used to identify notable time bins where origination or extinction may abnormally influence similarity.

Origination was measured in each time bin as the number of taxa that make their first appearance in the fossil record during that stage. Similarly, extinction was measured as the number of taxa that make their last appearance in the fossil record during that stage. It is important to note that extinction values for the Quaternary are skewed as the modern taxa could still go extinct.

Notably, the origination and extinction are often close in value, indicating that origination and extinction are not driving major trends. One major exception is the steep origination rate in the Rupelian. This origination is consistent with the sudden diversification of Neoceti, which occurs at or near the Eocene-Oligocene boundary. The only other major exception is in the Tortonian, where there is an especially high extinction rate. This is not unexpected, as the following stage, the Messinian, has long been associated with marine extinctions. Global sea level is known to have fallen drastically in the Messinian, and the Mediterranean is presumed to have dried almost entirely. This event, known as the Messinian Salinity Crisis, is undoubtedly responsible for the high number of taxa experiencing their last occurrence in the Tortonian.

Beta Diversity (Sørensen-Dice coefficients)

The mean SD value for the MET-NEA is 0.28 with a standard deviation of 0.13

(Figure 7). Overlap is nonexistent until the Lutetian, when the value rises to near the mean and remains there through the Aquitanian. After a sharp peak in the Burdigalian, overlap regresses to approach the mean through the rest of the Miocene. Beginning with the Zanclean, overlap increases, dips again in the Piacenzian, before eventually peaking at 0.54 in the Quaternary.

Figure 7: Sørensen-Dice coefficient values for ocean regions in the Northern hemisphere from the Eocene through the Quaternary. Dashed lines represent the epoch boundaries.

The mean SD value for the NEA-NWA is 0.23 with a standard deviation of 0.21

(Figure 7). The data cluster around the mean for nearly the entirety of geologic time, with three notable exceptions. The Priabonian shows a spike (up to 0.55), followed by a dramatic plunge well below the mean in the Rupelian and further still in the Chattian.

Overlap recovers and returns to the mean quickly, and remains there until a drastic drop to zero in the Messinian. As in the previous comparison, the Zanclean sees a sharp rise,

34 followed by a small decline in the Piacenzian, finished off by a huge peak in the

Quaternary (0.83).

The mean SD value for the NWA-NEP is 0.11 with a standard deviation of 0.20

(Figure 7). However, the mean values for all comparisons involving the Pacific are highly skewed, as no overlap exists until the Chattian (Figure 7). No overlap exists between the

NWA and the NEP whatsoever through the Burdigalian. Overlap first appears in the

Langhian (0.14) and remains near the mean until a sharp drop to zero in the Messinian.

Overlap then rises constantly from the Zanclean through the Piacenzian up to a peak in the Quaternary (0.76).

The mean SD value for the NEP-NWP is 0.15 with a standard deviation of 0.23

(Figure 7). Overlap first appears in the Chattian (0.19), but disappears again entirely in the Aquitanian. Overlap recovers slowly through the middle Miocene, peaking in the

Serravallian (0.32), before dropping in the Tortonian and disappearing entirely a second time in the Messinian. Overlap recovers almost immediately in the Zanclean (0.28), drops in the Piacenzian (0.19) and then spikes to its highest value in the Quaternary (0.91).

The mean SD value for the NWP-MET is 0.07 with a standard deviation of 0.13

(Figure 7). Overlap appears in the Chattian, disappears again in the Aquitanian, and then climbs steadily through the Serravallian (0.16). It disappears entirely in the Tortonian and

Messinian. Once again, overlap recovers quickly in the Zanclean, but drops to zero in the

Piacenzian before it rises drastically to its maximum value in the Quaternary (0.49).

Alroy’s Corrected Forbes Index

Alroy’s corrected Forbes index were calculated for comparison within the five previously identified regions of relevant alpha diversity. A full table of all the F’ values, as well as region specific graphical comparisons of the corrected Forbes index to the

Sørensen-Dice coefficient can be found in Appendix II.

Figure 8 shows that Alroy’s corrected Forbes index is strongly correlated with the

Sørensen-Dice coefficient (the data can be seen broken down by regional data series in

Appendix II). Generally speaking, the corrected Forbes index yields a slightly higher value than the Sørensen-Dice coefficient (as Alroy intends), a conclusion also shown by the graphs in Appendix II. While some data points appear as slight outliers, the correlation is quite strong (R2 = 0.90).

1.00

0.80 y = 1.365x + 0.05 R² = 0.90338 0.60

0.40

0.20 Alroy's Corrected Forbes Index

0.00 0.00 0.20 0.40 0.60 0.80 1.00 Sørensen-Dice Coeficient

Figure 8: Correlation of Alroy’s corrected Forbes index to the traditional Sørensen-Dice coefficient.

DISCUSSION

Comments on Alroy’s Corrected Forbes Index

As the results show, Alroy’s corrected Forbes index is strongly correlated with the traditional Sørensen-Dice coefficient (Figure 8). Using a quadratic regression, as opposed to linear, further increases the correlation value to an R2 value of 0.96. This is because

Alroy’s corrected Forbes Index was designed to specifically correct for incomplete sampling of the fossil record. Thus, the index will approach a value of one at a faster rate than traditional methods, resulting in a tail towards the right of the graph.

The slight deviations from the curve seen in Figure 9 led to speculation over conditions that may lead to the appearance of true outliers. A theoretical dataset was created to test the correlation under the most extreme of conditions. Specifically, the correlation was tested for its response to high directionality of the overlap (all of the taxa from one region overlap to another, but the reverse is not true). Table 2 displays this theoretical dataset.

Table 2: Theoretical dataset used to test the effects of directionality on the correlation between Alroy’s corrected Forbes index and the Sørensen-Dice coefficient. Values were specifically chosen to highlight extremes caused by discrepancies in alpha diversity or bias in the direction of overlap.

Point # Area 1 Area 2 Overlapping SD F' 1 1 20 1 0.10 1.00 2 5 20 1 0.08 0.20 3 5 20 5 0.40 1.00 4 10 20 10 0.67 1.00 5 10 20 5 0.33 0.57 6 10 20 10 0.67 1.00 7 15 20 1 0.06 0.09 8 15 20 5 0.29 0.44 9 15 20 10 0.57 0.80 10 15 20 15 0.86 1.00 11 20 20 1 0.05 0.08 12 20 20 5 0.25 0.38 13 20 20 10 0.50 0.70 14 20 20 15 0.75 0.92 15 20 20 20 1.00 1.00 16 10 100 10 0.18 1.00 17 50 100 50 0.67 1.00 18 75 100 75 0.86 1.00

This dataset strongly supports the hypothesis that the deviations between the two indexing methods are the result of directionality. The dataset is presented graphically below in Figure 9.

1.00

0.80

y = 0.7881x + 0.3705 0.60 R² = 0.48966

0.40

0.20 Alroy's Corrected Forbes Index

0.00 0.00 0.20 0.40 0.60 0.80 1.00 Sørensen-Dice Coeficient

Figure 9: Correlation between Alroy’s corrected Forbes index and the traditional Sørensen-Dice coefficient for the theoretical dataset.

Despite a low correlation value (R2 = 0.49), visual inspection of Figure 10 indicates a correlation between the two indices for most points. This low correlation value is the result of several obvious outliers, clustered at the top left of the graph. The points highlighted in red are points that were deemed outliers based on the following criteria: (1) Overlap was perfect in one direction and (2) the area with higher alpha diversity had more than double the taxa of the area with lower alpha diversity.

1.00

0.80

y = 1.0945x + 0.1251 0.60 R² = 0.91579

0.40

0.20 Alroy's Corrected Forbes Index

0.00 0.00 0.20 0.40 0.60 0.80 1.00 Sørensen -Dice Coeficient

Figure 10: Correlation between Alroy’s corrected Forbes index and the traditional Sørensen-Dice coefficient for the theoretical dataset after removal of the aforementioned outliers.

With the removal of these outliers (Figure 10), the correlation is drastically improved (R2 = 0.92). The further removal of points where the area of higher alpha diversity had exactly double the number of taxa as the area with lower alpha diversity

(not shown) resulted in an even higher correlation value (R2 = 0.95). Once again, in both cases, the R2 value was further improved by employing a quadratic regression as it better reflected the tail seen towards the right of the plots.

Based on this analysis, it is concluded that directionality of overlap plays a substantial role in the deviation between Alroy’s corrected Forbes index and the traditional Sørensen-Dice coefficient. Therefore, it is put forth that the traditional

Sørensen-Dice coefficient is superior when there is complete overlap in one direction and the area with higher diversity as more than double the taxa of the area with lower diversity. In all other situations, the two methods are analogous.

Hypothesis One – Rejected for the Atlantic, not rejected for the Pacific

The null hypothesis states that overlap is greater between adjacent regions of the same ocean (intraoceanic) than between regions of different ocean basins (interoceanic).

As a result of limiting our interpretation of results to only the five regions exhibiting sufficient alpha diversity, only two intraoceanic comparisons remained (NEA-NWA;

NEP-NWP). Despite this, sufficient evidence persists to assess the null hypothesis.

When comparing intraoceanic overlap in the Atlantic to interoceanic overlap between the Atlantic and either of the adjacent oceans, the intraoceanic comparison

(Figure 7) represents the highest overlap during only three stages (Priabonian,

Aquitanian, Quaternary). Moreover, only during the Priabonian is the value beyond one standard deviation from the next closest comparison. For the remaining eleven stages the

MET-NEA comparison is highest. Thus, the dominant comparison in the Atlantic Ocean is between the MET and the NEA, not between the NEA and the NWA.

When comparing intraoceanic overlap in the Pacific to interoceanic overlap between the Pacific and either of the adjacent oceans, the NEP-NWP (Figure 7) represents the highest overlap involving the Pacific during seven stages. Moreover, all eight stages in which the NEP-NWP does not represent the highest overlap, its value is

41 zero. Thus, it can be said that, when overlap exists across the Pacific, it is the dominant comparison.

Based on this, hypothesis one is rejected for the Atlantic, but not for the Pacific.

The Atlantic exhibits high interoceanic overlap with the Mediterranean-Tethys region, whereas the Pacific exhibits low interoceanic overlap but occasionally exhibits high intraoceanic overlap. This conclusion is supported by visual inspection of Figure 7.

Hypothesis Two – Inconclusive

The null hypothesis states that following the Aleutian-Kamchatka archipelago is the primary pathway by which marine mammals disperse across the Pacific. The results are insufficient to accept or reject the null, partially because the overlap across the Pacific appears to be sporadic and inconsistent. Although the NEP-NWP overlap (Figure 7) appears in the Oligocene, it is most consistent in the Miocene, where a peak in the

Serravallian is bracketed by a rise from zero in the Early Miocene and a drop back to zero in the Late Miocene (Table 3).

Table 3: Genera overlapping across the Pacific in the Oligocene and Miocene.

Chattian Aquitanian Burdigalian Langhian Serravallian Tortonian Messinian Aetiocetus Desmostylus Desmostylus Allodesmus Dusisiren Behemotops Paleoparadoxia Desmostylus Paleoparadoxia Parietobalaena Paleoparadoxia Scaldicetus

In some cases, these taxa have occurrences that are in the Aleutian Islands or the

Kamchatka Peninsula (such as Desmostylus) and clearly imply a use of the Aleutian-

Kamchatka archipelago to cross the Pacific. Others, such as Aetiocetus, have further south and are less indicative. It is also worth noting that overlap is restricted to cetaceans in the Oligocene, but is dominated by desmostylians for much of the Miocene. The

Burdigalian has no overlapping cetacean, the Langhian and the Serravallian only one, and the Tortonian one again zero.

The lack of overlapping sirenians is not surprising, as sirenians of the hydrodamaline lineage are thought to be the only cold water adapted members of the family (Domning, 1978, 1982; Takahasi et al., 1986).

The lack of overlapping pinnipeds is also not surprising, as pinnipeds have a much later origin and are not thought to disperse until the Late Miocene. The presence of

Allodesmus as the only overlapping pinniped is curious in that it disappears in the

Tortonian and Messinian; though pinnipeds reappear as overlapping taxa in the Pliocene

(Table 4). Thus, it is unclear whether pinniped overlap persists through the end of the

Miocene undetected in the fossil record, or if pinnipeds disperse across the Aleutian-

Kamchatka archipelago on two separate occasions. The overlap in the Pliocene (Table 4) and Quaternary is more diverse, with cetaceans, pinnipeds, and sirenians all represented.

Table 4: Genera overlapping across the Pacific in the Pliocene.

Zanclean Piacenzian Balaenoptera Callorhinus Herpetocetus Hydrodamalis Hydrodamalis Thalassoleon

Thus, the data for hypothesis two is difficult to interpret. Some taxa, such as desmostylians, pinnipeds, and Hydrodamalis show clear distribution patterns that support dispersal through the Aleutian-Kamchatka archipelago. Other taxa, such as many of the cetaceans, have an unclear distribution.

Hypothesis Three –Accepted

The null hypothesis states that the Strait of Gibraltar serves as the passageway through which cetaceans and sirenians leave their point of origin in the Mediterranean-

Tethys. Results for sirenians are inconclusive; though the fossil record captures their arrival to the new world in the Eocene they exhibit no overlap at the generic level. This is likely attributed to the dispersal event having resulted in enough morphological changes to establish a new genus, resulting in zero overlap values. However, there is strong evidence to support the hypothesis that the passageway was the primary exit point for cetaceans.

Similarity Values in the MET

1.00 MET-NEA MET-NWP

0.80

0.60

0.40

0.20 Sørensen-Dice Coeficient

0.00 50 40 30 20 10 0 Absolute Time

Figure 11: Sørensen-Dice coefficient values comparing the Mediterranean-Tethys to the Atlantic and

Pacific oceans.

With the exception of the Ypresian (before any overlap has occurred globally), the Mediterranean-Tethys Ocean shows higher overlap with the Atlantic than it does with the Pacific in all fourteen of the remaining stages (Figure 11). Additionally, the overlap with the Atlantic is higher by at least one standard deviation in all of these stages except for the Lutetian and the Quaternary (the standard deviation for the MET-NEA and the

MET-NWP is 0.15 and 0.14 respectively).

A closer look at the taxa that do overlap between the NWP and the MET in the

Oligocene and the Miocene further supports Hypothesis Three. This overlap (Table 5) is

45 represented by four taxa (Squalodon, Eurhinodelphis, Scaldicetus, and Cetotherium). All four genera also appear in the Atlantic in the same stage. Thus, the only genera that overlap between the NWP and the MET are genera that also appear in the Atlantic, indicating that this overlap is made up of taxa with a cosmopolitan distribution and is not indicative of any dispersal directly from the Mediterranean-Tethys into the Northwest

Pacific.

Table 5: NWP-MET overlapping genera in the Oligocene and Miocene.

Chattian Aquitanian Burdigalian Langhian Serravallian Squalodon Eurhinodelphis Eurhinodelphis Cetotherium Scaldicetus Squalodon

Hypothesis Four – Strong Rejection

The null hypothesis states that the Central American Seaway is the pathway through which modern high overlap values between the Atlantic and Pacific are attained.

No overlap is recorded between the NWA and the NEP (Figure 12) until the Langhian.

From the Langhian through the Zanclean, values fluctuate slightly but remain well within one standard deviation of one another, except during a drop to zero in the Messinian. A recovery in the Zanclean is followed by a sizeable jump in the Piacenzian, which is followed in turn by a huge peak in the Quaternary. Thus, despite evidence that the

Central American Seaway is used in a minimal capacity by some marine mammals at certain points, high overlap values are not achieved until after the closing of the seaway.

Similarity Values for the NWA-NEP

1.00 NWA-NEP 0.76 0.80

0.60

0.40 0.36

0.20 0.13 0.14 0.20 Sørensen-Dice Coeficient 0.11 0.00 50 40 30 20 10 0 Absolute Time

Figure 12: Sørensen-Dice coefficient values comparing the Northwest Atlantic to the Northeast Pacific.

A closer look at the overlapping taxa (Table 6) leads to several conclusions about existing marine mammal dispersal hypotheses. In total, seven genera overlap in the

Miocene. Of these, five are cetaceans, one a sirenian, and one a sea otter. The Pliocene continues to exhibit this pattern, with four cetaceans and a single sirenian represented. In the Quaternary, however, the drastic increase in overlap sees the pattern broken. The thirty-two overlapping taxa in the Quaternary are represented by twenty-five cetaceans, six pinnipeds, and polar bears.

Table 6: NWA-NEP overlapping taxa in the Miocene and Pliocene. The vertical divide between the

Zanclean and Piacenzian represents the closing of the Central American Seaway.

Langhian Serravallian Tortonian Messinian Zanclean Piacenzian Delphinodon Metaxytherium Enhydritherium Balaenoptera Balaenoptera Kentriodon Scaldicetus Lophocetus Herpetocetus Megaptera Metaxytherium Megaptera Parietobalaena Nanosiren Scaldicetus

These results lend strong support to the existing hypotheses for the dispersal of cetaceans, sirenians, and otters. The results support a dispersal event by sea otters from the Northeast Pacific through the Central America Seaway into the Northwest Atlantic prior to the closing of the seaway in the Pliocene. Similarly, there is evidence that sirenians pass through the seaway from the Atlantic into the Pacific. However, because only two sirenian taxa overlap at all, and only ever one at a time, Domning’s hypothesis of an ecological barrier is supported as well.

Curiously, however, there is no evidence whatsoever to support any pinniped dispersal events through the seaway. No pinnipeds overlap between the NWA and the

NEP at all in the Miocene or the Pliocene. However, six pinniped taxa overlap in the

Quaternary. These results are interpreted as indicating that all pinniped overlap occurs in the Quaternary, well after the closing of the Central American Seaway.

A closer examination of the overlapping taxa in the Quaternary lends further support to the rejection of hypothesis four. Of the thirty-two taxa that overlap in the

Quaternary, only Balaenoptera and Megaptera also overlap in the Zanclean (prior to the

48 closure of the seaway). Additionally, of these thirty-two taxa, twenty-two also appear in the arctic during the Quaternary (Figure 13). All together, the data leads us to strongly reject hypothesis four and conclude that the Arctic Ocean is the pathway through which modern high overlap values between the Atlantic and Pacific are attained.

Figure 13: Overlap between the Northwest Atlantic, the Northeast Pacific, and the Arctic.

The Pull of the Recent

As noted in the results section, the overlap values are always at their highest point during the Quaternary. Moreover, these values are always at least one, and in some cases two or more, standard deviations greater than the mean. This is hypothesized to be the effect of the pull of the recent, a bias that has been well established and discussed throughout studies pertaining to the fossil record (Alroy, 2014; Jablonski et al., 2003;

Sepkoski, 1997). This bias is the result of Holocene taxa having “perfect” sampling, and therefore none of the gaps present in the fossil record (Jablonski et al., 2003).

In 1997 Sepkoski attempted to quantify this phenomena by studying marine bivalves (Sepkoski, 1997). Sepkoski amassed a large dataset and calculated the number of modern bivalves that had fossil records extending back into the Pliocene or

Pleistocene. From this, Sepkoski concluded that the pull of the recent accounted for as much as 50% of the increase in biodiversity at the genus level (Sepkoski, 1997).

Recently however, Sepkoski’s study was revisited (Jablonski et al., 2003). This reanalysis of the data found that Sepkoski’s results were driven by an incomplete dataset as well as an over split taxonomy. After making corrections, it was reported that of the

958 genera of bivalves with a fossil record, 906 extended back into the Pliocene or

Pleistocene. Thus, they conclude that only about 5% of the increase in biodiversity is attributable to the pull of the recent (Jablonski et al., 2003).

Similar methods were used to determine the effect of the pull of the recent on the data reported here. Of the 66 extant genera of marine mammals, 23 were found to have a

50 record dating back to the Pliocene (it was essential to use the Pliocene as the Pleistocene was already included in the Quaternary). Following this method seen in the literature

(Jablonski et al., 2003; Sepkoski, 1997) it is concluded that 65% of the increase in biodiversity is due to the pull of the recent.

Number of Species Representing a Genus

19%

1 Species 2 Species 19% 62% 3+ Species

Figure 14: A total of 414 genera were present in the dataset. Figure 14 demonstrates the number of species that are currently attributed to each genus (by percentage). Over 60% of the genera have only a single species, with another 20% having only two. This led to speculation that the modern species level may be more comparable to the fossil record’s generic level.

Additionally, it was found that, of the 414 genera in the dataset, only 19% were represented by three or more species (Figure 14). This led to speculation that, with

51 regards to the recent, the species level may be more analogous to the generic level in the fossil record. Based on this, an attempt to correct for the pull of the recent was made by recalculating Sørensen-Dice coefficient values for the Quaternary at the species level

(Table 7).

Table 7: Sumary of data resulting from attempting to correct for the pull of the recent. Row 1 represents the overlap values in the Piacenzian (the most recent and, in most cases, the next highest overlap value).

Row 2 represents the overlap values already reported at the generic level for the entire Quaternary. Row 3 represents the overlap values for the Quaternary using only extant genera. Row 4 represents the overlap values for the Quaternary using only extant species. Row 5 represents the overlap values at the species level for the entire Quaternary.

MET-NEA NEA-NWA NWA-NEP NEP-NWP NWP-MET 1 Piacenzian SD 0.45 0.40 0.31 0.22 0.28 2 Quaternay Genera SD 0.55 0.85 0.76 0.89 0.49 3 Extant Genera SD 0.72 0.99 0.78 0.94 0.55 4 Extant Species SD 0.62 0.97 0.66 0.89 0.41 5 Quaternary Species SD 0.53 0.85 0.64 0.84 0.40

Rows 2 and 3 in Table 7 confirm the effect of the pull of the recent on the

Sørensen-Dice coefficient values. All five comparisons exhibit higher overlap when measuring only extant taxa. This indicates that the high overlap values of the Quaternary are primarily driven by the uncharacteristically high overlap values seen within the extant taxa. This result is consistent with the previous result reported attributing 65% of the increase in biodiversity to the pull of the recent.

Row 4 confirms that, among extant taxa, overlap at the species level is more analogous to the generic level in the fossil record. All five comparisons exhibit lower overlap values. Based on this conclusion, overlap values were calculated at the species level for the entire Quaternary.

Row 5 represents a largely unsuccessful attempt to correct for the pull of the recent by measuring overlap at the species level. While four of the five comparisons at the species level (row 5) demonstrate a decrease in overlap from the genus level (row 2), all of the comparisons remain well within a single standard deviation. Based on this result, it is concluded that, while the species level among extant taxa is more comparable to the generic level in the fossil record, calculating the overlap values at the species level does not serve to correct the bias created by the pull towards the recent.

It is worth noting that, for the extant taxa, calculating the overlap at the species level demonstrated it to be a more effective analog for the generic level in the fossil record. When the calculations were expanded to include the entire Quaternary, however, the overlap values at the species level were only marginally lower than at the generic level. This unusual phenomenon is most likely attributed to a very high number of single species genera in the Pleistocene.

Overlap values measured at the species level can never be greater than their equivalent at the generic level, as any species overlapping means the genus overlaps as well. Therefore, calculating the overlap values at the species level can only result in a decrease in the value; the amount of this decrease is proportional to the amount of difference in the data between the two levels. If every genus were a single species genus,

53 then the resulting values would be identical at each level. From this premise, it is concluded that the only marginal decrease in overlap values observed between the genus and species levels for the Quaternary, despite more substantial decrease seen between the same levels in the extant data, must be the result of a high number of single species genera.

These results lead to serve to highlight the true effects of the pull of the recent. As shown, the pull of the recent is shown to be a strong bias for the Quaternary data. This bias manifests in two ways. Primarily, the data are skewed by the simple premise that the extant taxa have perfect sampling, and therefore the record is complete. However, the high number of single species genera demonstrates a second mechanism by which the pull of the recent biases the data. Due to a myriad of additional information, such as geographic ranges, genetic studies, and behavior, extant taxa are much easier to split at the species level. Extinct taxa, which must be diagnosed based purely on morphology, are more often clumped into single species genera.

Comments of Closure of Central American Seaway

Though the recent hypothesis of a much earlier closure of the seaway (13-15 Ma) has recently gained favor in the literature (Farris et al., 2011; Hoorn and Flantua, 2015), the results presented here are instead consistent with the traditional hypothesis of a closure in the Pliocene (Coates et al., 2004). These results demonstrate overlap at the generic level between the Northeast Pacific and the Northwest Atlantic in the Langhian, Serravallian,

54 and Tortonian: a span of time ranging from 16 Ma to 7.2 Ma. Included in this overlap are five cetaceans, the sirenian Metaxytherium, and the otter Enhydritherium.

While the otter, having some terrestrial capability, would not necessarily have been unable to disperse after the closure of the seaway, this is certainly not the case for the sirenian or for the cetaceans. If the seaway was closed by 15 Ma, as put forth by Hoorn and Flantua (2015), it would necessitate that the cetaceans and Metaxytherium dispersed through the Arctic or through Drake’s Passage. As Metaxytherium is not a cold-water adapted sirenian, this hypothesis is doubtful. Moreover, examining the cetacean fossil occurrences lends further doubt. The fossil occurrences of the Langhian, Serravallian, and

Tortonian are restricted to Mexico, California, and Oregon in the Pacific, and the United

States east coast in the Atlantic. None of these occurrences are found in Canadian deposits, nor any latitude approaching arctic waters.

The overlapping genera examined in this study present direct evidence against a closure of date of 15 Ma (Farris et al., 2011; Hoorn and Flantua, 2015; Montes et al.,

2012). Instead, the evidence supports the traditional 3-4 Ma closure, with the seaway having been either fully opened, or dotted by an archipelago as described by (Coates et al., 2004; Coates and Stallard, 2013). The evidence presented here is consistent with other fossil evidence (Beu, 2001; Butzin et al., 2011) as well as climate evidence as well

(Butzin et al., 2011; Osborne et al., 2014).

CONCLUSIONS

The results demonstrate that using beta diversity indices to calculate overlap allows for a quantitative comparison of patterns across different ocean regions through geologic time that can be used to evaluate the role of several geographic features in marine mammal dispersal. Based on these results, it is concluded that the Atlantic has high interoceanic similarity with the Mediterranean-Tethys, as opposed high intraoceanic similarity, but the Pacific is dominated by intraoceanic similarity. The Atlantic shows the most similarity between the Northeast Atlantic and the Mediterranean, not between opposite halves of the Atlantic. In the Pacific overlap is far less consistent, and rarely as high as in the Atlantic. However, when overlap exists, the overlap between halves of the

Pacific is always greater than any overlap between the Pacific and another ocean.

The data strongly supports the hypothesis of cetaceans dispersing from their origin in the Tethys Sea through the Strait of Gibraltar. Throughout all time bins, the

Mediterranean was more similar to the Northeast Atlantic than to the Northwest Pacific.

Despite an open pathway from the Tethys Sea into the Indian, and then the Pacific, , cetaceans instead travel through the Strait of Gibraltar in much greater numbers.

The data supports the existing hypotheses for dispersal of cetaceans, sirenians, and sea otters through the Central American Seaway. The overlap of these taxonomic groups in the Miocene provides further evidence for a Pliocene closure of the Central

American Seaway. This overlap continues to provide evidence against the recent 15-13

Ma date for the closure of the seaway.

No evidence whatsoever is found that pinnipeds dispersed through the Central

American Seaway nor is any demonstrable overlap of pinnipeds between the Atlantic and

Pacific found at all until the Quaternary. The hypothesis that the Central American

Seaway is the pathway through which high modern overlap values were achieved between the Atlantic and Pacific Oceans is strongly rejected. Instead, Arctic Ocean is proposed as the more likely pathway and the viability of this hypothesis is demonstrated.

Results also indicate that Alroy’s Corrected Forbes Index is a strong correlate of the Sørensen-Dice Coefficient. Conditions causing the two indices to deviate are tested, and two conditions are identified. It is concluded that Alroy’s Corrected Forbes Index as a viable index except when directionality of overlap is especially strong and when one region has over double the number of taxa as the other region.

APPENDIX I

Tables A1 through A15 demonstrate the number of overlapping taxa and the

Sørensen-Dice coefficient values for all 78 comparisons for each geologic stage. Values appearing below the black boxes are the total number of overlapping taxa. Values appearing above the black boxes are the corresponding Sørensen-Dice coefficient values.

Color-coding is purely for visual identification of matching values.

Table A1: Ypresian

Ypresian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NEA 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NWA 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 0 0 0 0 0 0.00 0.00 0.00

59 NWP 0 0 0 0 0 0 0 0 0 0 0.00 0.00 SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A2: Lutetian

Lutetian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.07 0.07 0.00 0.00 0.00 0.00 0.00 0.00 NEA 0 0 0 0 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 NWA 0 0 0 0 1 0 0.00 0.00 0.00 0.00 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 0 0 0 0 0 0.00 0.00 0.00

60 NWP 0 0 0 0 0 0 0 0 0 0 0.00 0.00 SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A3: Bartonian

Bartonian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.22 0.33 0.22 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 1 0 0 0 0.31 0.38 0.00 0.00 0.00 0.00 0.00 0.00 NEA 1 0 0 0 2 0.15 0.00 0.00 0.00 0.00 0.00 0.00 NWA 1 0 0 0 3 1 0.00 0.00 0.00 0.00 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 0 0 0 0 0 0.00 0.00 0.00

61 NWP 0 0 0 0 0 0 0 0 0 0 0.00 0.00 SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A4: Priabonian

Priabonian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.13 0.29 0.25 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 1 0 0 0 0.32 0.20 0.00 0.00 0.00 0.00 0.00 0.00 NEA 1 0 0 0 3 0.55 0.00 0.00 0.00 0.00 0.00 0.00 NWA 1 0 0 0 2 3 0.00 0.00 0.00 0.00 0.29 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 0 0 0 0 0 0.00 0.00 0.00

62 NWP 0 0 0 0 0 0 0 0 0 0 0.00 0.00 SEP 0 0 0 0 0 0 1 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A5: Rupelian

Rupelian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.29 0.22 0.00 0.00 0.00 0.00 0.00 0.00 NEA 0 0 0 0 1 0.17 0.00 0.00 0.00 0.00 0.00 0.00 NWA 0 0 0 0 1 1 0.00 0.00 0.00 0.00 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 0 0 0 0 0 0.00 0.00 0.00

63 NWP 0 0 0 0 0 0 0 0 0 0 0.00 0.00 SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A6: Chattian

Chattian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.26 0.22 0.00 0.00 0.00 0.07 0.00 0.07 NEA 0 0 0 0 3 0.11 0.00 0.00 0.00 0.13 0.00 0.13 NWA 0 0 0 0 2 1 0.00 0.00 0.00 0.09 0.00 0.09 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 0 0 0 0 0 0.19 0.00 0.00

64 NWP 0 0 0 0 1 1 1 0 0 2 0.00 0.30 SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 1 1 1 0 0 0 3 0

Table A7: Aquitanian

Aquitanian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.22 0.11 0.00 0.00 0.00 0.00 0.00 0.00 NEA 0 0 0 0 2 0.25 0.00 0.00 0.00 0.00 0.00 0.00 NWA 0 0 0 0 1 1 0.00 0.00 0.00 0.00 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 0 0 0 0 0 0.00 0.00 0.00

65 NWP 0 0 0 0 0 0 0 0 0 0 0.00 0.00 SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A8: Burdigalian

Burdigallian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.47 0.18 0.00 0.00 0.08 0.08 0.00 0.00 NEA 0 0 0 0 9 0.23 0.00 0.00 0.07 0.07 0.00 0.00 NWA 0 0 0 0 3 4 0.00 0.00 0.00 0.09 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 1 1 0 0 0 0.00 0.00 0.00 66 NWP 0 0 0 0 1 1 1 0 0 0 0.00 0.00

SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A9: Langhian

Langhian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.32 0.29 0.00 0.00 0.17 0.09 0.00 0.00 NEA 0 0 0 0 4 0.22 0.00 0.00 0.13 0.10 0.00 0.00 NWA 0 0 0 0 7 5 0.00 0.00 0.14 0.14 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 3 2 4 0 0 0.20 0.00 0.00

67 NWP 0 0 0 0 1 1 3 0 0 3 0.00 0.00 SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A10: Serravallian

Serravallian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.30 0.25 0.00 0.00 0.12 0.16 0.00 0.00 NEA 0 0 0 0 8 0.25 0.00 0.00 0.10 0.13 0.00 0.00 NWA 0 0 0 0 6 7 0.00 0.00 0.11 0.15 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 2 2 2 0 0 0.32 0.00 0.00 68 NWP 0 0 0 0 3 3 3 0 0 4 0.00 0.00

SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A11: Tortonian

Tortonian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.21 0.17 0.00 0.00 0.07 0.00 0.00 0.00 NEA 0 0 0 0 4 0.19 0.00 0.00 0.13 0.06 0.00 0.00 NWA 0 0 0 0 2 4 0.00 0.00 0.13 0.00 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 1 3 2 0 0 0.17 0.00 0.00

69 NWP 0 0 0 0 0 1 0 0 0 2 0.00 0.00 SEP 0 0 0 0 0 0 0 0 0 0 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A12: Messinian

Messinian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.22 0.00 0.00 0.00 0.09 0.00 0.00 0.00 NEA 0 0 0 0 1 0.00 0.00 0.00 0.23 0.00 0.00 0.00 NWA 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 1 3 0 0 0 0.00 0.06 0.00 70 NWP 0 0 0 0 0 0 0 0 0 0 0.00 0.00

SEP 0 0 0 0 0 0 0 0 0 1 0 0.00 SWP 0 0 0 0 0 0 0 0 0 0 0 0

Table A13: Zanclean

Zanclean ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.43 0.34 0.00 0.00 0.15 0.22 0.15 0.44 NEA 0 0 0 0 6 0.31 0.00 0.00 0.27 0.19 0.00 0.36 NWA 0 0 0 0 8 8 0.00 0.00 0.20 0.20 0.00 0.24 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 2 4 5 0 0 0.28 0.00 0.30

71 NWP 0 0 0 0 3 3 5 0 0 4 0.00 0.19 SEP 0 0 0 0 1 0 0 0 0 0 0 0.00 SWP 0 0 0 0 4 4 5 0 0 3 2 0

Table A14: Piacenzian

Piacenzian ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ARC 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 EIO 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 WIO 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 MET 0 0 0 0 0.31 0.26 0.00 0.00 0.15 0.00 0.07 0.21 NEA 0 0 0 0 6 0.20 0.00 0.00 0.20 0.00 0.00 0.11 NWA 0 0 0 0 4 2 0.00 0.00 0.36 0.00 0.25 0.00 SEA 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 0.00 SWA 0 0 0 0 0 0 0 0 0.00 0.00 0.00 0.00 NEP 0 0 0 0 3 3 4 0 0 0.19 0.11 0.00

72 NWP 0 0 0 0 0 0 0 0 0 2 0.00 0.22 SEP 0 0 0 0 1 0 1 0 0 1 0 0.00 SWP 0 0 0 0 3 1 0 0 0 0 1 0

Table A15: Quaternary

Quaternary ANT ARC EIO WIO MET NEA NWA SEA SWA NEP NWP SEP SWP ANT 0.56 0.75 0.76 0.64 0.45 0.54 0.83 0.78 0.52 0.54 0.84 0.76 ARC 15 0.46 0.47 0.62 0.65 0.71 0.49 0.46 0.65 0.61 0.50 0.47 EIO 25 15 0.96 0.59 0.53 0.62 0.89 0.79 0.64 0.72 0.88 0.96 WIO 25 15 37 0.60 0.53 0.60 0.85 0.78 0.63 0.70 0.86 0.92 MET 15 14 17 17 0.54 0.59 0.67 0.59 0.56 0.49 0.64 0.60 NEA 17 24 23 23 18 0.83 0.55 0.51 0.71 0.69 0.49 0.53 NWA 18 23 24 23 17 36 0.62 0.67 0.76 0.79 0.66 0.62 SEA 26 15 33 31 18 23 23 0.86 0.70 0.68 0.93 0.90 SWA 26 15 31 30 17 22 26 32 0.64 0.63 0.90 0.81 NEP 19 23 27 26 18 33 32 28 27 0.91 0.68 0.65

73 NWP 21 23 32 31 17 34 35 29 28 43 0.67 0.73 SEP 26 15 32 31 17 20 24 32 33 27 28 0.89 SWP 25 15 37 35 17 23 24 33 31 27 32 32

APPENDIX II

Table B1 presents Sørensen-Dice coefficient values for the possible comparisons of adjacent ocean regions within the five identified zones of high marine mammal alpha diversity (NEA, NWA, NEP, NWP, MET). Table B2 presents the corresponding values for Alroy’s corrected Forbes index. Figures B1-5 compare the two measures of overlap for each comparison.

Table B1: Sørensen-Dice Coefficient Values

TIME MET-NEA NEA-NWA NWA-NEP NEP-NWP NWP-MET Ypresian 0.00 0.00 0.00 0.00 0.00 Lutetian 0.07 0.00 0.00 0.00 0.00 Bartonian 0.29 0.14 0.00 0.00 0.00 Priabonian 0.32 0.55 0.00 0.00 0.00 Rupelian 0.29 0.17 0.00 0.00 0.00 Chattian 0.26 0.11 0.00 0.19 0.07 Aquitanian 0.22 0.25 0.00 0.00 0.00 Burdigalian 0.47 0.23 0.00 0.00 0.08 Langhian 0.32 0.22 0.14 0.20 0.09 Serravallian 0.30 0.25 0.11 0.32 0.16 Tortonian 0.21 0.19 0.13 0.17 0.00 Messinian 0.22 0.00 0.00 0.00 0.00 Zanclean 0.43 0.31 0.20 0.28 0.22 Piacenzian 0.32 0.20 0.36 0.19 0.00 Quaternary 0.54 0.83 0.76 0.91 0.49

Table B2: Alroy’s Corrected Forbes Index Values

TIME MET-NEA NEA-NWA NWA-NEP NEP-NWP NWP-MET Ypresian 0.00 0.00 0.00 0.00 0.00 Lutetian 0.31 0.00 0.00 0.00 0.00 Bartonian 0.46 0.24 0.00 0.00 0.00 Priabonian 0.65 0.78 0.00 0.00 0.00 Rupelian 0.58 0.28 0.00 0.00 0.00 Chattian 0.54 0.20 0.00 0.30 0.13 Aquitanian 0.53 0.42 0.00 0.00 0.00 Burdigalian 0.68 0.36 0.00 0.00 0.14 Langhian 0.49 0.47 0.23 0.37 0.15 Serravallian 0.45 0.38 0.22 0.51 0.25 Tortonian 0.42 0.33 0.18 0.42 0.00 Messinian 0.42 0.00 0.00 0.00 0.00 Zanclean 0.64 0.55 0.38 0.42 0.35 Piacenzian 0.50 0.38 0.71 0.43 0.00 Quaternary 0.96 0.97 0.93 0.99 0.91

Figure B1: Comparison of Indices for MET-NEA

MET-NEA

1.00

0.80

0.60

0.40 SD F 0.20

0.00

Figure B2: Comparison of Indices for NEA-NWA

NEA-NWA

1.00

0.80

0.60

0.40 SD F 0.20

0.00

Figure B3: Comparison of Indices for NWA-NEP

NWA-NEP

1.00

0.80

0.60

0.40 SD F 0.20

0.00

Figure B4: Comparison of Indices for NEP-NWP

NEP-NWP

1.00

0.80

0.60

0.40 SD F 0.20

0.00

Figure B5: Comparison of Indices for NWP-MET

NWP-MET

1.00

0.80

0.60

0.40 SD F 0.20

0.00

Figure B6: Correlation of Sørensen-Dice coefficient with Alroy’s corrected Forbes

index as broken down by regional data series.

1.00

0.80

0.60 MET-NEA NEA-NWA

0.40 NWA-NEP NEP-NWP NWP-MET

Alroy's Corrected Forbes Index 0.20

0.00 0.00 0.20 0.40 0.60 0.80 1.00 Sorensen-Dice Coeficient

WORKS CITED

Alexanderson, H., Ó. Ingólfsson, A. S. Murray, and J. Dudek. 2013. An interglacial polar bear and an early Weichselian glaciation at Poolepynten, western Svalbard. Boreas 42:532-543.

Alroy, J. 2003. Global databases will yield reliable measures of global biodiversity. Paleobiology 29:26-29.

Alroy, J. 2014. Accurate and precise estimates of origination and extinction rates. Paleobiology 40:374-397.

Alroy, J. 2015. A new twist on a very old binary similarity coefficient. Ecology 96:575- 586.

Alroy, J., K. Behrensmeyer, M. Carrano, W. Clyde, E. Fara, M. Fortelius, J. Head, J. Hunter, M. Uhen, X. Wang, J. Anderson, H.-D. Sues, M. Fox, J. A. Case, J. Eberle, M. Goodwin, and S. Carlson. 2003. The 5% project; just how good is the fossil record of tetrapods? Journal of Vertebrate Paleontology 23:29.

Arnason, U., A. Gullberg, A. Janke, M. Kullberg, N. Lehman, E. A. Petrov, and R. Väinölä. 2006. Pinniped phylogeny and a new hypothesis for their origin and dispersal. Molecular Phylogenetics and Evolution 41:345-354.

Arnason, U., and B. Widegren. 1986. Pinniped phylogeny enlightened by molecular hybridizations using highly repetitive DNA. Molecular Biology and Evolution 3:356-365.

Bajpai, S., and P. D. Gingerich. 1998. A new Eocene archaeocete (Mammalia, Cetacea) from India and the time of origin of whales. Proceedings of the National Academy of Sciences of the United States of America 95:15464-15468.

Barnes, L. G. 2013. A new genus and species of Late Miocene paleoparadoxiid (Mammalia, Desmostylia) from California. Contributions in Science, Natural History Museum of Los Angeles County 521:51-114.

Berta, A. 2009. Pinniped evolution; pp. 861-868, Perrin, William F.; Wuersig, Bernd; Thewissen, J.G.M. [Eds]. Encyclopedia of marine mammals. Second edition.

Academic Press, Amsterdam, Boston etc. 2009: i-xxxiv, 1-1316. Chapter pagination: 861-868.

Berta, A., and G. S. Morgan. 1985. A New Sea Otter (Carnivora: Mustelidae) from the Late Miocene and Early Pliocene (Hemphillian) of North America. Journal of Paleontology 59:809-819.

Berta, A., C. E. Ray, and A. R. Wyss. 1989. Skeleton of the Oldest Known Pinniped, Enaliarctos mealsi. Science 244:60.

Berta, A., and A. R. Wyss. 1994. Pinniped phylogeny. Proceedings of the San Diego Society of Natural History 29:33-56.

Beu, A. G. 2001. Gradual Miocene to Pleistocene uplift of the Central Americanisthmus: evidence from tropical American tonnoidean gastropods. Journal of Paleontology 75:706-720.

Boessenecker, R. W., and M. Churchill. 2015. The oldest known fur seal. Biology Letters 11:20140835-20140835.

Butzin, M., G. Lohmann, and T. Bickert. 2011. Miocene ocean circulation inferred from marine carbon cycle modeling combined with benthic isotope records. Paleoceanography 26.

Canto, J., R. Salas-Gismondi, M. Cozzuol, and J. Yáñez. 2008. The aquatic sloth Thalassocnus (Mammalia, Xenarthra) from the late Miocene of North-Central Chile: biogeographic and ecological implications. Journal of Vertebrate Paleontology 28:918-922.

Coates, A. G., L. S. Collins, M. P. Aubry, and W. A. Berggren. 2004. The geology of the Darien, Panama, and the late Miocene-Pliocene collision of the Panama arc with northwestern South America. Geological Society of America Bulletin 116:1327- 1344.

Coates, A. G., and R. F. Stallard. 2013. How old is the Isthmus of Panama? Bulletin of Marine Science 89:801-813.

Condit, R., N. Pitman, E. G. L. Jr., J. Chave, J. Terborgh, R. B. Foster, P. N. V., S. Aguilar, R. Valencia, G. Villa, H. C. Muller-Landau, E. Losos, and S. P. Hubbell. 2002. Beta-diversity in tropical forest trees. Science 295:666-669.

Cooper, L. N. 2009. Evolution and development of cetacean appendages. Kent State University.

Cooper, L. N., E. R. Seiffert, M. Clementz, S. I. Madar, S. Bajpai, S. T. Hussain, and J. G. Thewissen. 2014. Anthracobunids from the middle Eocene of India and Pakistan are stem perissodactyls. PLoS ONE 9:e109232.

Crerar, L. D., A. P. Crerar, D. P. Domning, and E. C. Parsons. 2014. Rewriting the history of an extinction-was a population of Steller's sea cows (Hydrodamalis gigas) at St Lawrence Island also driven to extinction? Biol Lett 10:20140878.

Cronin, M. A., G. Rincon, R. W. Meredith, M. D. MacNeil, A. Islas-Trejo, A. Canovas, and J. F. Medrano. 2014. Molecular phylogeny and SNP variation of polar bears (Ursus maritimus), brown bears (U. arctos), and black bears (U. americanus) derived from genome sequences. Journal of Heredity 105:312-323.

Dasmahapatra, K. K., J. I. Hoffman, and W. Amos. 2009. Pinniped phylogenetic relationships inferred using AFLP markers. Heredity 103:168-77.

Davies, J. L. 1958. The Pinnipedia: An essay in zoogeography. Geographical Review 48:474-493.

Dawson, S. D. 1993. Functional morphology and evolution of marine mammal limb bones: Adaptations to aquatic locomotion in recent and fossil cetaceans, pp. 187. University of Pennsylvania.

Deméré, T. A., A. Berta, and P. J. Adam. 2003. Chapter 3: Pinnipedimorph evolutionary biogeography. Bulletin of the American Museum of Natural History:32-76.

Deméré, T. A., M. R. McGowen, A. Berta, and J. Gatesy. 2008. Morphological and molecular evidence for a stepwise evolutionary transition from teeth to baleen in mysticete whales. Systematic Biology 57:15-37.

Dice, L. R. 1945. Measures of the amount of ecologic association between species. Ecology 26:297-302.

Diedrich, C. 2011. The world’s oldest fossil seal record. Natural Science 03:914-920.

Domning, D. P. 1976. An ecological model for Late Tertiary sirenian evolution in the North Pacific Ocean. Systematic Zoology 25:352-362.

Domning, D. P. 1978. Sirenian evolution in the North Pacific Ocean. University of California Publications in Geological Sciences 118:1-176.

Domning, D. P. 1982. Evolution of manatees: A speculative history. Journal of Paleontology 56:599-619.

Domning, D. P. 2001. The earliest known fully quadrupedal sirenian. Nature 413:625- 627.

Domning, D. P. 2005. Fossil Sirenia of the West Atlantic and Caribbean Region. VII. Pleistocene Trichechus manatus Linnaeus, 1758. Journal of Vertebrate Paleontology 25:685-701.

Domning, D. P., C. E. Ray, and M. C. McKenna. 1986. Two new Oligocene desmostylians and a discussion of tethytherian systematics. Smithsonian Contributions to Paleobiology 59.

Fahlke, J. M., K. A. Bastl, G. M. Semprebon, and P. D. Gingerich. 2013. Paleoecology of archaeocete whales throughout the Eocene: Dietary adaptations revealed by microwear analysis. Palaeogeography, Palaeoclimatology, Palaeoecology 386:690-701.

Farris, D. W., C. Jaramillo, G. Bayona, S. A. Restrepo-Moreno, C. Montes, A. Cardona, A. Mora, R. J. Speakman, M. D. Glascock, and V. Valencia. 2011. Fracturing of the Panamanian Isthmus during initial collision with South America. Geology 39:1007-1010.

Fischer, M. S., and P. Tassy. 1993. The interrelation between Proboscidea, Sirenia, Hyracoidea, and Mesaxonia: the morphological evidence; pp. 217-234, Szalay, Frederick S.; Novacek, Michael J.; McKenna, Malcolm C. [Eds]. Mammal phylogeny: placentals. Springer-Verlag, New York, Berlin etc. 1993: i-xi, 1-321. Chapter pagination: 217-234.

Fitzgerald, E. M. G. 2004. A review of the Tertiary fossil Cetacea (Mammalia) localities in Australia. Memoirs of Museum Victoria 61:183–208.

Fitzgerald, E. M. G. 2006. A bizarre new toothed mysticete (Cetacea) from Australia and the early evolution of baleen whales. Proceedings of the Royal Society B: Biological Sciences 273:2955-2963.

Fitzgerald, E. M. G. 2010. The morphology and systematics of Mammalodon colliveri (Cetacea: Mysticeti), a toothed mysticete from the Oligocene of Australia. Zoological Journal of the Linnean Society 158:367-476.

Fordyce, R. E. 2002. Simocetus rayi (Odontoceti: Simocetidae, New Family): A bizarre new archaic Oligocene dolphin from the eastern North Pacific. Smithsonian Contributions to Paleobiology 93:185–222.

Fourcy, D., and O. Lorvelec. 2012. A new digital map of limits of oceans and seas consistent with high-resolution global shorelines. Journal of Coastal Research:471-477.

Geisler, J. H., and A. E. Sanders. 2003. Morphological evidence for the phylogeny of Cetacea. Journal of Mammalian Evolution 10:23–129.

Geisler, J. H., A. E. Sanders, and Z.-X. Luo. 2005. A new protocetid whale (Cetacea: Archaeoceti) from the Late Middle Eocene of South Carolina. American Museum Novitates 3480:1-68.

Geisler, J. H., and M. D. Uhen. 2003. Morphological support for a close relationship between hippos and whales. Journal of Vertebrate Paleontology 23:991-996.

Geisler, J. H., and M. D. Uhen. 2005. Phylogenetic relationships of extinct cetartiodactyls: Results of simultaneous analyses of molecular, morphological, and stratigraphic data. Journal of Mammalian Evolution 12:145-160.

Geraads, D., Z. Alemseged, R. Bobe, and D. Reed. 2011. Enhydriodon dikikae, sp. nov. (Carnivora: Mammalia), a gigantic otter from the Pliocene of Dikika, Lower Awash, Ethiopia. Journal of Vertebrate Paleontology 31:447-453.

Gingerich, P. D., and M. D. Uhen. 1996. Likelihood estimation of the time of origin of whales (Cetacea). Special Publication - The Paleontological Society 8:146.

Gingerich, P. D., and M. D. Uhen. 1998. Likelihood estimation of the time of origin of Cetacea and the time of divergence of Cetacea and Artiodactyla. Palaeontologia Electronica 1.

Goldman, D., P. R. Giri, and S. J. O Brien. 1989. Molecular genetic-distance estimates among the Ursidae as indicated by one- and two-dimensional protein electrophoresis. Evolution 43:282.

Hailer, F., V. E. Kutschera, B. M. Hallström, D. Klassert, S. R. Fain, J. A. Leonard, U. Arnason, and A. Janke. 2012. Nuclear genomic sequences reveal that polar bears are an old and distinct bear lineage. Science 336:344-347.

Harington, C. R. 2008. The evolution of arctic marine mammals. Ecological Applications 18:S23-S40.

Hautier, L., R. Sarr, R. Tabuce, F. Lihoreau, S. Adnet, D. P. Domning, M. Samb, and P. M. Hameh. 2012. First prorastomid sirenian from Senegal (Western Africa) and the Old World origin of sea cows. Journal of Vertebrate Paleontology 32:1218- 1222.

Hoorn, C., and S. Flantua. 2015. An early start for the Panama land bridge. Science 348:186-187.

Hunter, M. D. 2002. Landscape structure, habitat fragmentation, and the ecology of insects. Agricultural and Forest Entomology 4:159-166.

Hunter, M. L., and J. P. Gibbs. 2009. Fundamentals of conservation biology. 515 pp. John Wiley & Sons.

Hynek, S. A. 2011. Mio-Pliocene geology of the southern Puna plateau margin, Argentina, pp. 330. The University of Utah.

IHO. 1953. Limits of Oceans and Seas. International Hydrographic Organization, Monaco.

Inuzuka, N. 2000. Preliminary report on the evolution of aquatic adaptations in desmostylians (Mammalia, Tethytheria). Oryctos 3:71-77.

Jablonski, D., K. Roy, J. W. Valentine, R. M. Price, and P. S. Anderson. 2003. The impact of the pull of the recent on the history of marine diversity. Science 300:1133-5.

Jaccard, P. 1901. Étude comparative de la distribution florale dans une portion des Alpes et des Jura. Bulletin del la Société Vaudoise des Sciences Naturelles 37:547-579.

Kaschner, K., R. Watson, A. W. Trites, and D. Pauly. 2006. Mapping world-wide distributions of marine mammal species using a relative environmental suitability (RES) model. Marine Ecology Progress Series 316:285-310.

Koepfli, K.-P., K. A. Deere, G. J. Slater, C. Begg, K. Begg, L. Grassman, M. Lucherini, G. Veron, and R. K. Wayne. 2008. Multigene phylogeny of the Mustelidae: Resolving relationships, tempo and biogeographic history of a mammalian adaptive radiation. BMC Biology 6:10.

Kohno, N., H. Koike, and K. Narita. 2007. Outline of fossil marine mammals from the middle Miocene Bessho and Aoki Formations, Nagano Prefecture, Japan. Research Report of the Shinshushinmachi Fossil Museum 10:1-45.

Koretsky, I. A. 2002a. Paleontology of the Late Oligocene Ashley and Chandler Bridge Formations of South Carolina, 1: Paleogene pinniped remains; The oldest known seal (Carnivora: Phocidae). Smithsonian Contributions to Paleobiology 93:179- 183.

Koretsky, I. A. 2002e. A primitive seal (Mammalia: Phocidae) from the early middle Miocene of central Paratethys. Smithsonian Contributions to Paleobiology 93:163-178.

Kurtén, B. 1964. The evolution of the Polar Bear, Ursus maritimus Phipps. University of Helsinki (UH-Viikki).

Kurtén, B. 1968. Pleistocene Mammals of Europe. 328 pp. Weidenfeld & Nicolson.

Lambert, W. D. 1997. The osteology and paleoecology of the giant otter Enhydritherium terraenovae. Journal of Vertebrate Paleontology 17:738-749.

Larivière, S. 1998. Lontra Felina. Mammalian Species 575:1-5.

Lindqvist, C., S. C. Schuster, Y. Sun, S. L. Talbot, J. Qi, A. Ratan, L. P. Tomsho, L. Kasson, E. Zeyl, J. Aars, W. Miller, O. Ingolfsson, L. Bachmann, and O. Wiig. 2010. Complete mitochondrial genome of a Pleistocene jawbone unveils the origin of polar bear. Proceedings of the National Academy of Sciences of the United States of America 107:5053-5057.

Love, J. A. 1992. Sea otters. 148 pp. Fulcrum Publishers, Golden, Colorado.

Marx, F. G. 2009. Marine mammals through time: when less is more in studying palaeodiversity. Proceedings of the Royal Society B: Biological Sciences 276:887-892.

McKenna, M. C. 1975. Toward a phylogenetic classification of the mammalia; pp. 21-46 in W. P. Luckett, and F. S. Szalay (eds.), Phylogeny of the Primates. Springer US, New York.

Miller, W., S. C. Schuster, A. J. Welch, A. Ratan, O. C. Bedoya-Reina, F. Zhao, H. L. Kim, R. C. Burhans, D. I. Drautz, N. E. Wittekindt, L. P. Tomsho, E. Ibarra- Laclette, L. Herrera-Estrella, E. Peacock, S. Farley, G. K. Sage, K. Rode, M. Obbard, R. Montiel, L. Bachmann, Ó. Ingólfsson, J. Aars, T. Mailund, Ø. Wiig, S. L. Talbot, and C. Lindqvist. 2012. Polar and brown bear genomes reveal ancient admixture and demographic footprints of past climate change. Proceedings of the National Academy of Sciences of the United States of America 109.

Mitchell, E. D. 1989. A new cetacean from the late Eocene La Meseta Formation, Seymour Island, Antarctic Peninsula. Canadian Journal of Fisheries and Aquatic Sciences = Journal Canadien des Sciences Halieutiques et Aquatiques 46:2219- 2235.

Montes, C., G. Bayona, A. Cardona, D. M. Buchs, C. A. Silva, S. Morón, N. Hoyos, D. A. Ramírez, C. A. Jaramillo, and V. Valencia. 2012. Arc-continent collision and orocline formation: Closing of the Central American seaway. Journal of Geophysical Research 117.

Newton, E. T. 1890. On some New Mammals from the Red and Norwich Crags. Quarterly Journal of the Geological Society 46:444-453.

Osborne, A. H., D. R. Newkirk, J. Groeneveld, E. E. Martin, R. Tiedemann, and M. Frank. 2014. The seawater neodymium and lead isotope record of the final stages of Central American Seaway closure. Paleoceanography 29:715-729.

Peters, S. E., and N. A. Heim. 2010. The geological completeness of paleontological sampling in North America. Paleobiology 36:61-79.

Peters, S. E., C. Zhang, M. Livny, and C. Ré. 2014. A machine reading system for assembling synthetic paleontological databases. PLoS ONE 9:e113523.

Plotnick, R. E., and P. J. Wagner. 2006. Round up the usual suspects: common genera in the fossil record and the nature of wastebasket taxa. Paleobiology 32:126-146.

Prothero, D. R. 2012. Cenozoic mammals and climate change: The contrast between coarse-scale versus high-resolution studies explained by species sorting. Geosciences 2:25-41.

Prothero, D. R. 2014. Species longevity in North American fossil mammals. Integrative Zoology 9:383-393.

Quirin, S. 2003. Setting the record straight. Nature 424:482-483.

Reinhart, R. H. 1953. Diagnosis of the new mammalian order, Desmostylia. Journal of Geology 61:187.

Repenning, C. 1976. Enhydra and Enhydriodon from the Pacific coast of North America. Journal of Research of the U. S. Geological Survey 4:305-315.

Sanders, A. E., and L. G. Barnes. 2002. Paleontology of the Late Oligocene Ashley and Chandler Bridge formations of South Carolina, 2: Micromysticetus rothauseni, a primitive cetotheriid mysticete (Mammalia, Cetacea). Smithsonian Contributions to Paleobiology 93:271–293.

Sepkoski, J. J., Jr. 1988. Alpha, beta, or gamma: where does all the diversity go? Paleobiology 14:221-234.

Sepkoski, J. J., Jr. 1997. Biodiversity: Past, Present, and Future. Journal of Paleontology 71:533-539.

Sørensen, T. 1948. A method of establishing groups of equal amplitude in plant sociology based on similarity of species and its application to analyses of the vegetation on Danish commons. Kongelige Danske Videnskabernes Selskab 5:1-34.

Takahasi, S., D. P. Domning, and T. Saito. 1986. Dusisiren dewana, n. sp. (Mammalia: Sirenia), a new ancestor of Steller's Sea Cow from the upper Miocene of Yamagata Prefecture, Northeastern Japan. Transactions and Proceedings of the Palaeontological Society of Japan 141:296–321.

Thewissen, J. G. M., E. Williams, M., and S. T. Hussain. 2001. Eocene mammal faunas from northern Indo-Pakistan. Journal of Vertebrate Paleontology 21:347–366.

Uhen, M. D. 1999. New species of protocetid archaeocete whale, Eocetus wardii (Mammalia: Cetacea) from the Middle Eocene of North Carolina. Journal of Paleontology 73:512-528.

Uhen, M. D. 2007. Evolution of marine mammals: back to the sea after 300 million years. The Anatomical Record 290:514–522.

Uhen, M. D. 2008. A new Xenorophus‐like odontocete cetacean from the oligocene of North Carolina and a discussion of the basal odontocete radiation. Journal of Systematic Palaeontology 6:433-452.

Uhen, M. D. 2010. The origin(s) of whales. Annual Review of Earth and Planetary Sciences 38:189-219.

Uhen, M. D. 2014. Paleobiology Database: A Community Based Data Service for Paleobiology: In 2014 Annual Meeting of the Geological Society of America (GSA 2014).

Uhen, M. D., A. D. Barnosky, B. Bills, J. Blois, M. T. Carrano, M. A. Carrasco, G. M. Erickson, J. T. Eronen, M. Fortelius, R. W. Graham, E. C. Grimm, M. A. O'Leary, A. Mast, W. H. Piel, P. D. Polly, and L. K. Säilä. 2013. From card catalogs to computers: databases in vertebrate paleontology. Journal of Vertebrate Paleontology 33:13-28.

Uhen, M. D., and H.-J. Berndt. 2008. First record of the archaeocete whale family Protocetidae from Europe. Fossil Record 11:57-60.

Uhen, M. D., A. G. Coates, C. A. Jaramillo, C. Montes, C. Pimiento, A. Rincon, N. Strong, and J. Velez-Juarbe. 2010. Marine mammals from the Miocene of Panama. Journal of South American Earth Sciences 30:167-175.

Uhen, M. D., and N. D. Pyenson. 2002. Evolution of cetacean diversity. Journal of Vertebrate Paleontology 22:116.

Uhen, M. D., and N. D. Pyenson. 2007. Diversity estimates, biases, and histriographic effects: Resolving cetacean diversity in the Tertiary. Paleontologia Electronica 10:1-22.

Uhen, M. D., N. D. Pyenson, T. J. Devries, M. Urbina, and P. R. Renne. 2011. New middle Eocene whales from the Pisco basin of Peru. Journal of Paleontology 85:955-969.

Varela, S., J. González-Hernández, L. F. Sgarbi, C. Marshall, M. D. Uhen, S. Peters, and M. McClennen. 2014. paleobioDB: an R package for downloading, visualizing and processing data from the Paleobiology Database. Ecography 38:1-7.

Verma, B. 1992. Enhydriodon sivalensis, a giant fossil otter from the Saketi Formation (Upper Pliocene), Siwalik Group, Sirmur District, Himachal Pradesh. Journal of the Palaeontological Society of India 37:31-36.

Villalta Cornella, J. F., and M. Crusafont Pairo. 1945. Enhydriodon lluecai nova sp. el primer lútrido del pontiense españo. Boletín de la Real Sociedad Española de Historia Natura 43:383-396.

Waits, L. P., J. Sullivan, S. J. O'Brien, and R. H. Ward. 1999. Rapid radiation events in the family Ursidae indicated by likelihood phylogenetic estimation from multiple fragments of mtDNA. Molecular Phylogenetics and Evolution 13:82-92.

Wayne, R. K., B. V. Valkenburgh, and S. J. O'Brien. 1991. Molecular distance and divergence time in carnivores and primates. Molecular Biology and Evolution 8:297-319.

Webb, S. D. 2006. The great American biotic interchange: patterns and processes. Annals of the Missouri Botanical Garden 93:245-257.

Whittaker, R. H. 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecological Monographs 30:279-338.

Whittaker, R. H. 1972. Evolution and measurement of species diversity. Taxon 21:213- 251.

Willemsen, G. F. 1989. A new specimen of the otter Enhydra reevei, Newton 1890, from the Crag of Bramerton, Norfolk. Bulletin of the Geological Society of Norfolk 39:87-90.

Willemsen, G. F. 1992. A revision of the Pliocene and Quaternary Lutrinae from Europe. Scripta Geologica:1-115.

Wilson, M. V., and A. Shmida. 1984. Measuring beta diversity with presence-absence data. Journal of Ecology 72:1055-1064.

Woodburne, M. O. 2010. The great American biotic interchange: dispersals, tectonics, climate, sea level and holding pens. Journal of Mammalian Evolution 17:245-264.

Yu, L., Q.-w. Li, O. A. Ryder, and Y.-p. Zhang. 2004. Phylogeny of the bears (Ursidae) based on nuclear and mitochondrial genes. Molecular Phylogenetics and Evolution 32:480-494.

Yu, L., Y.-W. Li, O. A. Ryder, and Y.-P. Zhang. 2007. Analysis of complete mitochondrial genome sequences increases phylogenetic resolution of bears

(Ursidae), a mammalian family that experienced rapid speciation. BMC Evolutionary Biology 7:198.

BIOGRAPHY

Carlos Mauricio Peredo grew up in the Baltimore/Washington area. He attended

Seton Hill University, in Greensburg, Pennsylvania, where he earned his Bachelor of

Science in Biology in 2012. Carlos maintains an active researching position at the

Smithsonian Museum of Natural History, where he has worked on projects ranging from foraminifera, dinosaurs, paleoecology, and most recently, fossil whales. Carlos has taught geology at George Mason since 2014, and will continue to do so while he pursues his PhD. In his free time, Carlos volunteers as a martial arts instructor and plays as much tennis as the weather will permit.