THE UNIVERSITY OF CHICAGO ODDITIES, WONDERS, AND OTHER ... · the university of chicago oddities,...

THE UNIVERSITY OF CHICAGO

ODDITIES, WONDERS, AND OTHER TALL TALES

OF “LIVING FOSSILS”

A DISSERTATION SUBMITTED TO

THE FACULTY OF THE DIVISION OF THE BIOLOGICAL SCIENCES

AND THE PRITZKER SCHOOL OF MEDICINE

IN CANDIDACY FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

COMMITTEE ON EVOLUTIONARY BIOLOGY

BY

L. H. LIOW

CHICAGO, ILLINOIS

JUNE 2006

TABLE OF CONTENTS

..........................................................................................................LIST OF TABLES vii.......................................................................................................LIST OF FIGURES viii

................................................................................................ACKNOWLEDGMENTS x

Chapter ! I. ! LINEAGE PERSISTENCE AND “LIVING FOSSILS” - A SEMANTIC

.................................................................................................... EXPOSITION 1..................................................! I.1.! What is the meaning of “living fossils?”! 2

.......................! I.2.! Distribution of purported “living fossils” in the literature! 3................! I.3.! Distribution of “living fossils” on the hierarchical tree of life ! 4

...................................................... I.4. Related concepts and their correlates. 5....................................................................! I.5.! “Living fossils” as artifacts ! 11

................................. I.6. How to make a geologically long-ranging lineage 13......................................................................! I.7.! Dissecting “living fossils”! 15

! II. ! A TEST OF SIMPSON’S “RULE OF THE SURVIVAL OF THE

.................! ! RELATIVELY UNSPECIALIZED” USING FOSSIL CRINOIDS! 21........................................................................................... II.1. Introduction. 21

............................................................................. II.2. Material and Methods 25.................................................................................. II.2.1. The data 26

............................................................................... II.2.2. Data units 28....................................................................... II.2.3. Data treatment. 31

........................ II.2.4. Morphological deviations from group means 32............... II.2.5. Morphological deviations from basal morphology 35

.......................................................................... II.3. Results and Discussion. 35........................ II.3.1. Morphological deviations from group means 35

............... II.3.2. Morphological deviations from basal morphology 40......................................... II.3.3. Influence of taxonomic hierarchy 42

............................. II.3.4. Temporal divisions and mass extinctions. 44...................................................................... II.3.5. Potential biases 46

................................... II.3.6. Explicit definitions of long-lived taxa 49.......................................................................................... II.4. Conclusions.. 51

iii

III. DO DEVIANTS LIVE LONGER? MORPHOLOGY AND LONGEVITY IN ........................................................ TRACHYLEBERIDID OSTRACODES 54

.......................................................................................... III.1.Introduction. 54................................................................................. III.2.Data and Methods .60

............................................ III.2.1. The organisms and the raw data 60............. III.2.2. Data treatment and analysis: discrete character data 66

..................................................... III.2.3. Removal of oversplit taxa. 67.............................. III.2.4. Data treatment and analysis: outline data 68

.................................................... III.2.5. Defining long-lived genera 68.................................................................................................. III.3.Results. 70

III.3.1. Morphological deviation of genera from group means (discrete ............................................................................. characters). 70

III.3.2. Morphological deviation of genera from group means (discrete ................................................. characters): temporal subsets 77

III.3.3. Principal Coordinate Analysis of discrete morphological data ................................................................................................82 III.3.4. Morphological deviation of genera from group means (outline

................................................................................ analyses) 84............................................................................................. III.4. Discussion 87

.......................................................................................... III.5. Conclusions 90

IV. DOES VERSATILITY AS MEASURED BY GEOGRAPHIC RANGE, BATHYMETRIC RANGE AND MORPHOLOGICAL VARIABILITY

.................................................. CONTRIBUTE TO TAXON LONGEVITY? 93.......................................................................................... IV.1. Introduction. 93

............................................................... IV.2. Methods and Materials..... ......97...................................... IV.2.1. Taxonomic and morphological data. 97

........................................................... IV.2.2. Geographic range data 99.................................................. IV.2.3. Bathymetric range data...... 100

.................................................................. IV.2.4. Data subsets....... 102............................................................................... IV.2.5. Analyses 105

.......... IV.3. Results....................................................................................... 108.. IV.3.1. Longevity and ecological versatility I: geographic spread 108

IV.3.2. Longevity and ecological versatility II: bathymetric spread 111. IV.3.3. Longevity and evolutionary versatility I: species richness. 116

IV.3.4. Longevity and evolutionary versatility II: subspecies richness ..............................................................................................116 IV.3.5 Longevity and evolutionary versatility III: extreme species

.................................................... morphological variability. 118

iv

................................................. IV.3.6. Which factors are stronger? 118...... IV.3.7. Are species patterns enough to explain genus patterns? 119

............................................................................................ IV.4.Discussion 123................................................................................ IV.4.1. Caveats. 127

.......................................................................................... IV.5.Conclusions 129

V. LINEAGES WITH GREAT LONGEVITIES ARE OLD AND AVERAGE: AN ANALYSIS OF MORPHOLOGICAL AND TAXON LONGEVITY

................................... DISTRIBUTIONS USING MULTIPLE DATASETS 130......................................................................................... V.1. Introduction. 130

.................................................................................. V.2. Methods.............. 134...................................................................................... V.2.1. Data 134

.................................................................... V.2.2. Data treatment. .142.................................................................................................. V.3. Results 150

............................................................................................ V.4. Discussion 158................................................. V.4.1. Phylogenetic implications. ..162.................................................. V.4.2. Biases and sources of error 162

.......................................................................................... V.5. Conclusions 164

VI. LINEAGE PERSISTENCE - A THEORETICAL FRAMEWORK AND ....................................................... EMPIRICAL RESEARCH PROGRAM 165

.................................................................................................LITERATURE CITED 168

Appendices A. Description of characters and character matrix for seven crinoids not

........................................................................... represented in Foote (1999) 198 B. Crinoid genera in orders (and suborders of cladids) and their morpho-duration

...................................................... plot distributions, relative to basal genera 209 C. Periods in geologic history sampled and morpho-duration plot distributions of

......................................................... the crinoid genera within those periods 214 D. Identities of long-lived crinoid genera in each order and the orders and families

...................................................................................... to which they belong 215 E. Description of discrete morphological characters and character states for

........................................................................... trachyleberidid genera (on CD)..........................................! F.! Characters states for trachyleberidid genera ! (on CD)

..........................................................................! G.! Sources of outline data ! (on CD)................... H. First and last fossil appearances of trachyleberidid genera (on CD)

............................................... I. References cited in appendices G and H (on CD)

v

..................................................................... J. Table of resampled correlations 219.............................. K. Character matrices for trachyleberidid ostracode species 220

L. Short descriptions of morphological characters of trachyleberidid ostracode ............................................................................................................ species 225

...................................................................... M. References used in Appendix K 228.......................................... O. Correlations between morphology and longevity 234

vi

LIST OF TABLES

Table II.1. Crinoid (sub)orders and the morpho-duration plot distributions of their

..................................................................................................... genera 29

..............Table II.2. Crinoid families and their morpho-duration plot distributions 43

......Table III.1. Summary statistics for the durations of subsets of trachyleberidids 71

......Table III.2. Deviation from trachyleberidid group mean (discrete morphology) 72

Table III.3. Deviation from trachyleberidid group mean (discrete morphology) with

..................................................................... oversplit genera removed.. 76

Table III.4. Deviation from trachyleberidid birth cohort mean (discrete morphology)

................................................................................................................ 79

Table III.5. Deviation from trachyleberidid contemporaneous cohort mean (discrete

........................................................................................... morphology) 81

.Table III.6. Deviation from trachyleberidid means (PCO of discrete morphology) 83

.......................................Table III.7. Results from trachyleberidid outline analyses 85

................................................................................Table IV.1. Bathymetric zones 101

..........................................................................................Table IV.2. Data subsets 103

.........................................................Table IV.3. Geographic spread and longevity 110

.....................................................Table IV.4. Bathymetric range versus longevity 113

..............................................................Table IV.5. Results of multiple regression 120

...........................................................Table V.1. References used in the analyses 135

Table V.2. Datasets where morphological distances are negatively, positively, or not

..................................................................... correlated with longevity. 152

vii

LIST OF FIGURES

Figure I.1. The relationship between longevity, cladogenesis and morphological

................................................................................................... distance 16

.......Figure. I.2.! Combinations of clade types where “living fossils” are detectable ! 19

Figure II.1. Translating stratigraphic ranges and multivariate morphology into a plot

.......................................... of morphological deviation versus duration. 24

Figure II.2. Morphological deviations of monobathrid and disparid genera versus

........................................................................................ their durations. 36

Figure II.3. Euclidean distances between genera of Sagenocrinida and taxa with

................................ alleged basal morphologies, versus their durations 41

Figure II.4. Changing relative occupation of morpho-duration plot quadrants through

......................................................................................................... time 45

Figure II.5. Percentages of unknown and inapplicable characters in crinoid genera as

.......................................................... a function of their fossil durations 47

Figure II.6.! Frequency of occurrence of long-lived genera of monobathrids,

..........................................! depending on the definition of “long-lived.”! 50

Figure III.1. A lower probability for long-lived taxa to be distant from the average

................................................................... morphology of their group.. 59

..........................Figure III.2. Literature sampling curve for trachyleberidid genera .63

...............................................Figure III.3. A generalized trachyleberidid ostracode.. 65

Figure IV.1. Genus longevity plotted versus genus latitudinal ranges for the whole

.................................................................................................. dataset 109

Figure IV.2. Histograms of genus longevities as subdivided by whether they occupy

.................................. only shallow waters, only deep waters, or both 114

Figure IV.3. Histograms of species longevities as subdivided by whether they occupy

................................. only shallow waters, only deep waters, or both . 115

viii

Figure IV.4. Distribution of species occupying various depth zones during the

.............................................. Cretaceous, Paleogene and the Neogene 117

............................................................Figure IV.5. Genus versus species longevity 122

Figure V.1. Hypothetical plot of morphological distance versus stratigraphic ranges

............... illustrating the sampling of groups and individual lineages. .143

ix

ACKNOWLEDGMENTS

My gratitude to all those who have helped me one way or another during the

conception, construction and conclusion of this thesis, even if I fail to mention them in

print.

! In partial chronological order, I thank Tim Wootton and Larry Heaney who first

encouraged me to apply to this school and Matthew Leibold who made me feel very

welcome here during my interview as a prospective student. I thank Jerry Coyne and

Nipam Patel, who almost became my committee members, Barry Chernoff and Joel

Martin who were my committee members for short periods of time, Michael Foote

whom I had no courage to put on my committee, but who has subsequently contributed

as much to this thesis as my committee members, and last but not least, my committee

members, Scott Lidgard, Leigh van Valen, Peter Wagner and David Jablonski. I can

never say enough “Thank Yous” to my advisor, Scott, who is an ever patient friend, a

mentor and a counsellor, forgiving even my worst behavior. Leigh was the light-house

(with functional light-bulbs of course) during my very numerous dark and foggy spells.

I thank David Rowley for helping with paleo-coordinate rotations and all the other

University of Chicago faculty who have given me some of their precious time. Alumni

were an indispensable resource during my sojourn in Chicago: Charles Marshall

carefully reviewed chapters two and three and gave me the encouragement I needed;

Arnie Mille found time to wade through a very dense draft of chapter four; Dan

McShea is a constant source of inspiration even from my early days in the CEB office. x

! I thank all the crustacean, ostracode and crinoid workers who have helped me with

my research. Special thanks to Fred Schram who hosted me in Amsterdam, Tom Cronin

who showed me ostracodes and the bass guitar in Reston Virginia and Joseph Hazel

who never stopped responding to my emails until he couldn’t. Carlita Sanford was very

kind and helped me a great deal with the late Dick Benson’s literature and materials and

Hallie Sims generously allowed me to use her D. C. apartment during my stay at the

Smithsonian. Lorraine Smith helped me navigate the Howe ostracode collections,

especially Joe’s materials and Haw Chuan Lim and his wife Ching Chi fed me delicious

Singaporean food and gave me a roof over my head at Louisiana State University

(Baton Rouge). Forest Gahn went out of his way to help me with the crinoid data and

thanks also go to William Ausich and Thomas Baumiller (yet another helpful alumnus)

for their crinoid expertise.

Becca Price, Gene Hunt, Big Al McGowan, Tom Rothfus (whose ingenuity helped

format this thesis), Dave Sunderlin, Shanan Peters and Emily Greenfest, were senior

students I looked up to as a junior grad student on the 2nd floor. I feel the same about

them today and probably will for the foreseeable future. Carl Simpson, Bjarte

Hannisdal and J.J. Emerson are my cohort mates and sources of unending spasmodic

laughter, copious bad jokes and programming help. Paul Harnik read with a critical eye

almost every piece of trash and research I have written in Chicago, educated me on gulf

coast outcrops, poverty in southeast United States, and patiently molded my biased

knowledge of Americans and American folk music, despite my occasional resistance to xi

understanding. Rebecca Rundell is my ever faithful office-mate who is always here in

spirit. She is mother to our guppies and hamsters, Aunt Agony and my personal

wikkipedia of the world of molecular phylogenetics.

! I thank my mother and sister for letting me fulfill my dream of fulfilling my father’s

dream that he had for himself and his children, even if it took me away from them.

Derek Frydel showed me the Chicago that I now love, even though I have not lost sight

of its flaws and atrocities.

! Carolyn Johnson, Marilyn Bowie, Monica Polk and the Office of International

Affairs helped with administrative matters. Funding for this thesis was provided by the

Geological Society of America, Sigma-Xi and the Hinds Fund (University of Chicago).

The Gurley Fund, the DooLittle Fellowship and the Women’s Board Travel Awards

allowed me to present parts of this research at various international meetings. I also

thank the Paleontological society and the University of Chicago Press for granting me

permission to reprint my papers originally published in Paleobiology and American

Naturalist respectively.

xii

CHAPTER I

LINEAGE PERSISTENCE AND “LIVING FOSSILS” - A SEMANTIC

EXPOSITION

“Living fossils ... anomalous forms.... (that) have endured to the present day,

from having inhabited a confined area,

and from having thus been exposed to less severe competition”

--- Darwin 1859

“Living fossils” posed a problem for Darwin (1859), whose thesis was, in part, to

demonstrate organic change over the span of geologic time. They continue to be a

problem for biologists, who have shown with independent lines of evidence, that

evolutionary change is the prevailing condition of the organic world. Similarly, they are

curiosities to paleontologists who have abundant evidence that “extinction is the

common lot, survival the exception” (Romer 1949 in Simpson 1953). Why do some

organisms seem to stay unchanged with time and unchallenged by extinction? Many

authors have attempted to assess or resolve with this issue (e.g. Ruedemann 1918,

1922a, b, Delamare-Deboutteville & Botosaneanu 1970, Eldredge & Stanley 1984,

Schopf 1984, Thenius 2001), but despite these efforts, do we even agree on what “living

fossils” are? Is the concept is a scientifically useful one? In this chapter, I discuss these

questions, hence providing an introduction to the subsequent chapters of this thesis.1

What is the meaning of “living fossils?”

“Living fossils” imply various phenomena in the opinion of different authors. Eldredge

(1984) loosely defines them as “ members of the Recent biota whose external form, at

least, has changed but little since the lineages’ inception.” Stanley (1979) restricts

“living fossils” to taxa that “have survived for relatively long intervals of geologic

time at low numerical diversity, often as the sole survivors of previously diverse taxa.”

To Fisher (1990), they are supra-specific taxa that have shown unusual morphological

conservatism, perhaps justifying Darwin’s claim that they are “anomalous” (1859). In

contrast to Darwin’s idea that “living fossils” occupy a restricted area, Vrba claims that

“living fossils” are eurytopic, have broad areal and habitat distribution compared with

their sister taxa (1984). Yet others have recognized multiple problematic issues

concerning “living fossils” and recommend discarding this term (Schopf 1984).

Here, I note that the term “living fossil” as used in the literature obscures three to four

confounding concepts, namely a slow rate of evolution, lineage persistence and

phylogenetic or morphological isolation. Moreover, “living fossils” as used in the

literature is a misnomer: they need not be extant (Kraft et al. 1999, Hahn et al. 2001, see

also Leander & Keeling 2003). This is justified, despite Darwin’s original definition,

since there is nothing special about the current time plane other than our own

perspective. In addition they need not be represented in the fossil record (Elena et al.

2

1991, Kyrpides and Ouzounis 1995, Soltis et al. 2002), since the absence of a record is

not evidence that a form or a lineage or a molecule is not or has not been persistent.

Distribution of purported “living fossils” in the literature

Individual lineages of “living fossils” are subjects of numerous scientific papers (e.g.

Wall & Dale 1966, McKenzie 1967, Taylor 1978, Newman & Hessler 1989, Avise et al.

1994, King & Hanner 1998, Jarman & Elliot 2000, Hedges 2003, Zhou & Zheng 2003)

but there are also systematic treatments of the concept. Most notably in the case of the

latter, Eldredge & Stanley (1984) invited taxonomic specialists to write about “living

fossils” ranging from tree squirrels to leptostracans and Nautilus. Researchers in France

(Delamare-Deboutteville & Botosaneanu 1970, de Ricqles 1983), Japan (Oji 1994 and

papers in the same volume) and the Germany (Thenius 2001) also made similar

attempts. While these volumes are informative, they are nonetheless rather eclectic in

coverage and subjective in methodologies. Even the two volumes of Palaeobiology: A

Synthesis could not avoid “living fossils” because they are objects of intense interest

(Fisher 1990, Oji 2001). Their treatment in these synthetic treatises has continued to be

subjective.

Although it sounds like an archaic idea, the term “living fossil” is still very much in

vogue, even in leading scientific journals. A recent issue of Nature contains an article

on the “living fossil” plant Ginkgo (Zhou & Zheng 2003); Lonnig & Sadler (2002) in 3

Annual Review of Genetics discussed the maintenance of “living fossils” despite

rampant chromosomal rearrangements and relentless transposable elements; Yoshida

(2002) simulated “living fossils” in evolving food webs in Paleobiology. This is

probably because the term catches the eye and conjures up dramatic images of relictual

monsters, providing a stimulating literary hook. Rigorous comparative research has

been neglected, producing little in the form of a synthetic understanding of the concept

of “living fossils,” despite their apparent violation of general principles of biology.

Distribution of “living fossils” on the hierarchical tree of life

Common “living fossils” that come to mind include Latimeria the coelacanth, Lingula

the brachiopod, Limulus the horseshoe crab and Triops the tadpole shrimp. But some

unexpected entities, e.g. mitochondria (Thenius 2000), proteins (Ivanov 1993) and

viruses (Ackermann et al. 1995) have also been labeled “living fossils” because of their

apparent prolonged lack of change.

Taking into account the full context of usage, the “living fossil” phenomenon can occur

at any taxonomic level (species, e.g. Ginkgo biloba; genera, e.g. Hipposideros; families

or orders, e.g. notostracans) and may also pertain to morphological or molecular (genes,

proteins) attributes. They are distributed across the tree of life (mammals, fishes,

molluscs, brachiopods, arthropods, echinoderms, plants, sponges, etc), and can be found

among both taxa with relatively more adequate fossil records (e.g. bivalves) and 4

relatively less adequate ones (e.g. insects, Liow, unpublished database of publications

concerning “living fossils”).

This illustrates that “living fossils” are not isolated anomalies but a general

phenomenon found in multiple, if not all branches of the tree of life, as well as at

various levels of the branching hierarchy of the tree of life. Since they are widespread

and common, they reflect a general phenomenon that deserves our scientific attention.

Related concepts and their correlates

The ontology of the broad concept of “living fossil” is a complex one. Several

interrelated inferences contribute to its development. In this section, I briefly

summarize these inferences and their interdependence.

1. Stasis

Morphological stasis or morphological conservatism in its various guises is “still one of

the most challenging problems in biology” twenty years after the widely cited review

paper of Wake et al. (1983) (e.g. Schwenk & Wagner 2001). In fact, examples of stasis

of any sort, whether genetic, developmental, morphological or ecological, are

commonly considered curiosities of nature by many biologists. Many “living fossils”

are organisms that display extreme morphological stasis (Avise et al. 1994, Suno-Uchi

et al. 1997, Jarman & Elliot 2000), beyond that of the average species duration under 5

the assumption of a punctuated mode of speciation (Eldredge & Gould 1972, Gould &

Eldredge 1993).

There is a gentle gradation between taxa (e.g. species) that are truly continuous through

a long time and groups that became well-differentiated in a “comparatively minor way,

but that early acquired a fundamental structural type that has been relatively invariable”

(Simpson 1944). Triops cancriformis supposedly a continuous species lineage for 240

million years (Tasch 1969), and Osmunda cinnamomea that has lived in North America

for at least 70 million years (Serbet & Rothwell 1999), illustrate the first case. The

Raninidae (Malacostraca: Brachyura), a relatively speciose crab family with 32 genera

and 190 species (Tucker 1998) is morphologically “constrained” so that any raninid

specimen can easily be recognized a member of the family even by lay people. Does

this continuum cause a problem for the “reality” of persistent taxa? I concur with

Stanley (1985) that if a morphologically static lineage actually contains ten species

instead of just one, then we have ten examples of stasis albeit non-independent ones,

instead of one.

2. Arrested evolution

Ruedemann can be credited for first addressing “arrested evolution” systematically

(1918, 1922a, 1922b). He defined genera demonstrating arrested evolution as those

genera that are preserved in the fossil record that survived two geologic stages or more. 6

He asked if these genera had any life history characteristics or ecological properties in

common that differentiated them from otherwise similar groups. This was the first

systematic, semi-quantitative, comparative study of the correlates of persistent taxa

even though it was flawed in many ways (e.g. the length of stages are not equivalent

and some genera he compared are from clades with vastly different intrinsic

preservation potentials and rates of evolution; see Simpson 1944). Since then, “arrested

evolution” has been used more loosely to mean morphological stasis (Jaanuusson 1985,

Trott 1998) or simply equated to bradytely (Eldredge 1979) .

3. Panchronic forms

This is a essentially a synonym of “living fossils” used mainly by French-speaking

biologists (although they readily use the term “fossiles vivants” too; see de Ricqles

1983 and other papers in the same volume). Its use is also frequently associated with

creationists from all over the world eager to quote biologists out of context.

4. Bradytely

Bradytely is the phenomenon of exceptionally slow rates of evolution; it is supposed to

be discontinuous from the average group rate (horotely) and opposed to tachytely

(Simpson 1944, 1953), as proposed on the basis of survivorship curves. As mentioned

before, it is important to compare rates of change of equivalent organisms because

clades display characteristic intrinsic rates. For example, morphological evolution is 7

frequently very slow in many protists (Poinar et al. 1993), even though among protists

there are some individual taxa that evolve relatively quickly. Thus if we compared

mammals and protists, we might conclude falsely that most protists are “living fossils!”

Simpson’s key idea is that the mean evolutionary rates of extinct and extant taxa of the

same inclusive (higher taxonomic) group can sometimes be different such that extant

taxa have already survived longer than expected from their group history. However,

this division into extinct and extant taxa is in part artificial since the Recent is not a

special plane of time, only one that we are thinking and writing in. Also, the prevailing

view of biotic change during Simpson’s era was one of phyletic evolution, which

colored his discussion of bradytely (see next section). The term bradytely was neither

adopted widely nor replaced the use of “living fossils.” Interestingly, it was given a

modern quantitative treatment by Raup & Marshall (1980), who showed that some

mammal groups do have significantly lower genus turnover rates (which these authors

equated to bradytely).

5. Lack of speciation

After Eldredge & Gould’s seminal paper of 1972 on the punctuated equilibrium model

of evolution, the question of “living fossils” became one of why speciation did not

occur (e.g. Stanley 1979, although for an earlier mention of this idea, see Mac Gillvary

1968, pers. comm. van Valen 2006). On the premise of a punctuated mode of evolution, 8

Stanley (1979, 1985) and others (notably Gilinsky 1988) dismissed Simpson's bradytely

since it became apparent that he had estimated only extinction rates (survival rates)

instead of evolutionary rates (combining both lineage originations and extinctions).

Even though Gilinsky has carefully shown that there are significant differences in the

survivorship of extinct and extant taxa, he explicitly stated that his results are not to be

used as a defense of Simpson's bradytely (Gilinsky 1988). A lack or slow rate of

speciation is often implied in “living fossil” taxa or “relicts” (see later sections) that

have apparently not given rise to descendants in a long time (e.g. Sphenodon, Triops)

Even though we know now that many lineages that do not seem to have speciated

actually harbor more genetic variation than one would expect (Avise et al. 1994, Suno-

Uchi et al. 1997, Jarman & Elliot 2000) such that by some conventions, new species can

be named, yet the cryptic speciations do not provide satisfactory explanation as to why

morphological divergence was damped despite genetic divergences.

6. Numerical relicts

These are survivors of once abundant clades (Simpson 1953, Stanley 1979), which in

some instances are “dead clades” that continue “walking” for a usually prolonged

periods of time (see Jablonski 2002). Although many “living fossil” taxa also seem not

to be diverse, it is often unclear whether it is because they have never been diverse (an

issue of low group rates of Simpson 1953); because they are at the natural tail end of 9

their geologic existence whether due to extinction events or intrinsic causes (“dead

clades”); because their relatives are sampled with a vanishingly small probability; or

because they are largely made up of cryptic taxa (i.e. only an apparent lack of diversity).

6. Morphological or phylogenetic isolation

“Living fossil” taxa have been explicitly or implicitly discussed as phylogenetic relicts

displaying arrested evolution (see Simpson 1953 p. 303) or as morphological relicts

thought to have been temporally isolated from their closest relatives. Sphenodon, for

instance, has no close living relatives (Hay et al. 2004); neither do the several species of

Limulus (Sekiguchi & Sugita 1980); nor Nautilus, which is also morphologically and

phylogenetically quite distant from other molluscs living today (Woodruff et al. 1983).

Organisms can be retictual in various ways and the same organisms can be taxonomic,

phylogenetic, numerical and geographic relicts at the same time (Simpson 1953). For

instance, Newman (1985) and Lesicki (1998) reported hydrothermal vent taxa that are

both phylogenetic and geographic relicts (but see Little & Vrijenhoek 2003).

7. Geographic isolation and stable habitats

Some “living fossil” taxa are also “groups occupying a much smaller geographic area

than their ancestors and early relatives” (Simpson 1953, although see Vrba 1984).

These are called geographic relics (e.g. Metasequoia, but see also botanical disjuncts

e.g. Wen 1999 and Ricklefs & Latham 1992). Conversely, some habitats are purported 10

to support the lack of change in organisms, because they are thought to have been stable

for longer periods of time than other habitats. These include the deep sea (Schein-

Fatton 1985, Vermeij 1987, Ameziane & Roux 1997, Wilson 1999), marine caves

(Fosshagen & Iliffe 1985, Vermeij 1987, Stepien et al. 2001) and other cryptic habitats

(Lange et al. 2001). As examples, such habitats harbor “living fossils” such as

Bathypecten (Schein-Fatton 1985), stalked crinoids (Ameziane & Roux 1997), deep sea

barnacles (Newman & Hessler 1989) and copepods such as Antriocopia and

Erebonectes (Fosshagen & Iliffe 1985).

“Living fossils” as artifacts

Are “living fossils” truly morphologically conservative taxa or are they artifacts in some

way? Some taxa previously deemed to be “living fossils” were subsequently

“dethroned” (Roush 1997), including Lingula and related forms (Biernat & Emig

1993), crocodiles and their relatives (Buckley et al. 2000) and some crinoids (Hotchkiss

1977), due to the availability of better data or a different method of analysis.

“Living fossils” could be artifacts because taxa which are morphologically simpler

could appear to be geologically longer ranging than those that are more complex due to

taxonomic lumping (Schopf et al. 1975). For instance, Kakabekia umbellata is a Pre-

cambrian fossil whose morphologically similar extant congeneric has been found

(Siegal et al. 1967). However, both the fossil and extant species have very few 11

morphological characters (see Siegal et al. 1967 and Siegel & Siegel 1968) and few if

any congenerics are known. However, it has been shown that when studies done with a

sufficient comparative basis and consistent, adequate methodology, this is not true

(Boucot 1977, Ward & Signor 1983, Liow 2004). In fact, when entire clades are

examined for right tails in longevity distributions, there are always taxa that are

geologically very long-ranging, but that are not garbage-can taxa (see Liow 2004,

2006).

Species can be cryptic (Suno-Uchi et al. 1997, Jarman & Elliot 2000, Colborn et al.

2001, Knowlton 2000) such that an apparently wide-spread or geologically long-

ranging species is actually more than one species. However, recognizing cryptic species

does not take away the need to understand why these separate species fail to

differentiated morphologically after prolonged periods of isolation (Simpson 1944,

Stanley 1985). The problem changes slightly from the paucity of speciation to the lack

of morphological change once species have been established.

Apparent extremely long-ranging taxa can also be due to a combination of highly

incomplete sampling, taxonomic lumping, poorly preserved specimens and mistaken

phylogenetic inferences. However, despite these possible artifacts, the true remaining

tail ends of longevity distributions still need a closer look.

12

How to make a geologically long-ranging lineage

In order for a lineage to persist unchanged, two general general conditions firs must be

met. 1) The change within the lineage in question should not be directional over a

period of time that is presumably longer than the average longevity of the same clade.

2) The lineage should not go extinct.

1. Lack of directional morphological evolutionary change

There is no lack of genetic variability in many purported “living fossils” (Selander et

al. 1970, Hammond & Poinar 1984, Avise et al. 1994, Endo et al. 2001). However,

canalization (= buffering or stabilization) could prevent the translation of variation in

genes or development into measurable differences in phenotype over time (for reviews

see Arnold 1992, Rutherford 2000, Gibson & Wagner 2000, Schwenk & Wagner 2001).

Van Valen (1982) lists conditions under which species integrity can be maintained and

these are easily extendable to the above species-level.

Extrinsic factors have also been called upon to explain the phenomenon of “living

fossils”. In stressful habitats where resources are limited, taxa adapted to these

marginal conditions conceivably have fewer competitors (Ruedemann 1918) and a

diminished likelihood of undergoing directional selection (Parsons 1993, 1994). There

is however no comparative, quantitative test of this idea, as far as I am aware.

13

2. Lineage persistence

As mentioned before, once the punctuated mode of evolution was widely accepted,

where most evolutionary changes are assumed to be associated with branching events,

the question of unchanging forms became focussed on lineage persistence. That is, how

did such lineages escape extinction or increase survivorship, as opposed to the question

of how changes within species were avoided.

Numerous studies have investigated the correlates of lineage longevity either for some

defined length of the clade history or more commonly across some defined geologic

event. For instance, lineage persistence or survivorship is thought to be aided by wide

geographic ranges (Boucot 1975a, Jablonski 1987 see also Hunt et al. 2005), generality

of feeding ecology (Baumiller 1993), width of niche breathe (Kammer et al. 1997,

1998), smaller body size (Hallam 1975, van Valen 1975 but see Flynn et al. 1995), deep

depths of occurrences (Buzas & Culver 1984, although see Fortey 1980). These studies

and others will be mentioned in more detail in later chapters.

Here, I note that many of the above-mentioned studies are insufficient with regard to

testing the concept of very persistent forms or taxa. This is because either the full

geologic range of the taxa is not surveyed (when studies are done across extinction

events) or because there is no explicit test whether taxa with extreme longevities are any

different from their relatives (either as individual taxa or collectively). 14

In the following sections and chapters, I make suggestions as to how to dissect the

concept of “living fossils” so it can be useful and informative in quantitative studies. Of

necessity driven by time and availability of data, I consider only selected parts of the

broad concept of “living fossils” for the rest of this thesis.

Dissecting “living fossils”

A slow rate of evolution (or cladogenesis), lineage persistence (= duration) and

phylogenetic isolation (often implicitly measured as morphological isolation) are ideas

that are implicit in the term “living fossil,” even though they do not need to occur

concurrently for a taxon to be called a living fossil in the literature. For the sake of

discussion and clarity, I show the six end-member combinations in which these patterns

can occur among lineages. These are easily analogous at higher taxonomic levels.

There are theoretically eight possible combinations of rates, durations and isolation, of

which two are effectively irrelevant to the concept (Fig I.1 Cases A through F). I

acknowledge that intermediate cases are probably common in nature.

15

Branching/Cladogenesis

close wide

slow fast slow fast

Morphological distanceLon

gev

ity/

Du

rati

on

long

sh

ort A B

C D E F

Fig. I.1

Fig. I.1. The relationship between longevity, cladogenesis and morphological distance.

Case A : where species have short durations, cladogenesis occurs often and the

morphological distances of the species are relatively small.

This results in a numerically rich clade where species may not be easily

distinguished from one another, especially in the fossil record, where

most data are derived from the morphological of skeletal hard parts. It

can be mistaken as an abundant, wide-spread and perhaps geologically

long-ranging single species complex (because there are many similar

species).

16

Case B: where species have short durations, cladogenesis occurs often and the

morphological distances among the species are large. This may be

identified as an evolutionary radiation since the very numerous related

species can be distinguished from one another. Even if the events of

cladogensis did not occur very close to the Recent time plane, we can

still distinguish the various species easily if there are many preserved

morphological characters and recognize a radiation (where individual

species are rather short-lived).

Case C: where species have long durations, cladogenesis seldom occurs and the

morphological distances of the species are small.

! ! This scenario perhaps applies to many classic “living fossils,” where

! ! there are only one or a few species of a genus that do not seem to have

! ! had descendants.

Case D: where species have long durations, cladogenesis occurs often and the

morphological distances of the species are small.

This is similar to Case A except that the cryptic species or species

complex may seem to be even more persistent than an equivalent group

demonstrating Case A since the individual true species are long-lived, in

addition to the group being rich in descendants.17

Case E: where species have long durations, cladogenesis seldom occurs and the

morphological distances of the species are wide.

! ! This is similar to Case C except that the “living fossil” species that may

! ! be sister taxa appear very different morphologically from each other.

Case F: where species have long durations, cladogenesis occurs often and the

morphological distances of the species are large.

! ! This is similar to case B although instead of an evolutionary radiation,

! ! they may simply be identified as a highly diverse clade that produced

! ! many progeny, many of the latter persisting as individual species for a

! ! long time. Despite their “long duration” component, they may not be

! ! identified as “living fossils” because they have many relatives that are

! ! equally long-ranging.

Using the above scheme, a thought experiment demonstrates a range of possible

evolutionary scenarios and their perception with in the fossil record. There are (6x5)/2

possible pairs of clades (Fig. I.2). Combinations of the same paris, e.g. Case A and A,

are not interesting because they are no sufficient to distinguish differential patterns.

18

A B C D E F

A

B detectable

C - detectable

D - detectable -

E - detectable - -

F - detectable - - -

Fig. I.2. Combinations of clade types where “living fossils” are detectable. See Fig. I.1. for

illustrations of clade type cases A though F.

Out of the 15 combinations of types of clades, only 5 have a detectable “living fossil”

part, even though one of them (Case A + B) is actually only an illusion because we

mistake A as an exceptionally long-ranging clade when it is a bush of cryptic species.

The above exercise shows that “living fossil” is not only a concept comprised disparate

components that should be isolated for study, but that the concept is also dependent

upon relative statements in its construction and therefore must be studied in a

comparative manner. The taxa within the studied clade must have similar preservation

potentials, similar sampling probabilities, more or less same numbers of characters and

they should also be of equivalent taxonomic rank. For example, lingulids seem to be

archaic, but comparing individual extant species may lead to different conclusions from

19

comparing lingulids with all the other brachiopod families or orders that have ever

existed.

In this thesis, I study taxon longevity beyond the comparison of survivorship across a

single extinction event using crinoid echinoderms and trachyleberidid ostracodes and

other taxa. I carefully define “persistent” or “long-lived” lineages in a comparative

manner. I also explore, in depth, morphology, a potential correlate of longevity that has

been neglected in quantitative studies. I examine geographic and bathymetric ranges

and clade characteristics such as species richness in relation to lineage longevity. In

doing so, I combine Ruedemann’s (1918) systematic approach with Simpson’s (1944)

comparative approach and incorporate issues of concern in today’s paleobiological

studies, including sampling sufficiency and using exemplary taxa with good fossil

records. Consequently, previously unnoticed species and genera with very long

geologic durations (some longer than many purported “living fossils”) are also noticed

in the crinoid and ostracode datasets.

Persistent and are non-changing entities are rampant throughout the tree of life and

deserve our attention. They are perhaps not as odd, surprising or wonder as they might

seem to be, when studied without their full evolutionary and ecological context. The

conceptual elements within the construct “living fossil” should not be confounded in

analyses and are worthy of individual study.20

CHAPTER II

A TEST OF SIMPSON’S “RULE OF THE SURVIVAL OF THE RELATIVELY

UNSPECIALIZED” USING FOSSIL CRINOIDS 1

Introduction

Prolonged stasis in a world of change is a puzzling biological phenomenon. Extremely

long-lived or geologically long-ranging taxa have been a popular subject of discussion

for paleontologists and neontologists alike, ever since Darwin (1859) coined the term

“living fossils.” Authors including Ruedemann (1918, 1922a, 1922b), Simpson (1944,

1953), Stanley (1979), Wake et al. (1983), Eldredge & Stanley (1984) and Avise et al.

(1994) have discussed “living fossils” and the related phenomena of arrested evolution,

bradytely, and morphological stasis or conservatism.

Long-lived taxa are commonly thought to survive longer than related shorter-lived taxa

because they are unique, unusual or exceptional in some significant way. They

allegedly reside in unusual habitats (Selander et al. 1970, Parsons 1994) or have

distinctive morphological features not shared by shorter-lived taxa (Ward & Signor

1983, Kammer et al. 1998). Many previous studies on “living fossils” have

characterized them as paradoxical, relictual, primitive or special (e.g. McKenzie 1967,

Mooi 1990, King & Hanner 1998, Eisner 2003) without exploring the phenomenon of

� 21

1 This paper originally appeared in volume 164, no. 4, pages 431-443, of the American Naturalist.

longevity in a comparative and quantitative manner. Here, I examine whole clades in

order to discover any shared patterns among long-lived taxa, using a quantitative

approach. I use three explicit definitions of long-lived (see Data treatment in Materials

and Methods). Long-lived taxa defined as such are not necessarily designated by other

authors as “living fossils.”

Crinoids (feather stars and sea lilies) belong to the exclusively marine phylum

Echinodermata. Crinoids have been chosen as an illustrative taxon for several reasons.

First, crinoids are monophyletic (Janies 2001). Second, they are morphologically

conservative enough to allow meaningful comparative analysis. Third, they are diverse

enough to provide large samples for quantitative study. Fourth, they can be divided into

recognized taxonomic subgroups for further comparisons without sacrificing the

adequacy of sample sizes. Fifth, there exists a large morphological database of fossil

crinoid species and their first and last geologic appearances, sampled quite evenly

across all crinoid subgroups (Foote 1999). Sixth, certain crinoids are considered

“living fossils” (Roux 1987, Heinzeller et al. 1996, Ameziane & Roux 1997, Laille et al.

1998). Others are thought to exhibit extreme morphological conservatism (Simms

1988) or phenotypic bradytely (Kammer 2001). Seventh, the crinoid fossil record spans

almost the entire Phanerozoic, beginning definitively in the Ordovician, peaking in

taxonomic richness during the Carboniferous (Lane & Webster 1980, Hess et al. 1999,

Guensburg & Sprinkle 2003), continuing through the Cenozoic into the Recent � 22

(Ameziane & Roux 1997). However, an overwhelming majority of crinoid genera

originated and went extinct during the Paleozoic (Moore & Teichert 1978). This

minimizes problems arising from one-sided range truncations, where taxa originating

closer to the Recent have shorter geologic durations due to unfinished histories. Last,

Crinoids have relatively high fossilization and preservation potentials. Although fossil

crinoid specimens are frequently disarticulated, confident assignments to species or at

least genus are often possible (Ausich et al. 1999, Ausich & Kammer 2001).

Simpson implicitly took a comparative approach when he wrote about the “rule of the

survival of the relatively unspecialized” (1944, p.143). He thought that unspecialized

subgroups of a clade seem to persist for longer periods of geologic time but did not

explicitly define “specialization.” Here, I quantify specialization by comparing

individual morphologies to a group mean: the closer a morphology is to a group mean,

the less specialized it is. I ask if long-lived genera (taxa A & B in Fig. II.1.A) in any

given crinoid order occupy regions of morphospace that are random with respect to the

mean morphology of that order. Could survival be correlated with morphological

bizarreness or a deviant morphology (Fig. II.1.B, taxon A)? Or would long-lived genera

have morphologies close to the mean morphology (Fig. II.1.B, taxon B)?

I find that the morphologies of long-lived crinoid genera are, in general, closer to mean

morphologies than shorter-lived genera in the same order. This is in agreement with � 23

Tim

e

MorphologicalAxis 1

Morpho

logica

l

Axis 2

B

A

Mor

phol

ogic

a l d

evi a

ti on

(from

gr o

up m

ean

orba

sal m

e mbe

r )

Duration

Duration

Mor

phol

ogic

al d

evia

t ion

SL-t LL-t

LL-bSL-b

B

A

Med Mid 10-g

A.

B.

C.

Morpho-duration plots

Fig. II.1. A. Translating stratigraphic ranges and multivariate morphology into a plot of morphological deviation versus duration. Schematic diagram showing the geologic ranges (solid lines) of 10 related taxa. The dotted lines project the multivariate morphology of each taxon onto a two-dimensional plane. The open circle marks the location of the mean morphology of all 10 taxa considered. By plotting the distances between the open circle and each taxon versus their durations, a morpho-duration plot as shown in B is obtained. In Fig.II.1.B, A represents a long-lived taxon that is morphologically deviant relative to the group mean, and B represents a long-lived taxon that is not deviant relative to the group mean. Dotted lines show the median duration value (Med), midrange value (Mid), and the duration greater than those associated with the 10% most long-lived genera (10-g). Fig.II.1.C shows the naming of mopho-duration plot quadrants used in this chapter. Quadrant SL-t houses the shorter-lived, deviant taxa; SL-b the shorter-lived non-deviant (unspecialized) taxa; LL-t the long-lived deviant taxa and LL-b the long-lived non-deviant taxa. The deviation quantified can be relative to either a group mean or a basal morphology.

� 24

Simpson's “rule of the survival of the relatively unspecialized.” The “long-lived,

deviant” quadrants (LL-t in Fig.II.1.C) of morpho-duration plots (Fig.II.1.B and C) are

often empty and the members in the “long-lived, unspecialized” quadrant (LL-b in

Fig.II.1.C) are closer to mean morphologies than expected by chance.

Similarly, but from a completely different conceptual perspective, I ask if long-lived

crinoid genera in any given crinoid higher taxon (e.g. suborder, order) occupy regions of

morphospace that are random with respect to a basal morphology of that higher taxon. I

find that mean morphological distances of long-lived genera from basal morphologies

are seldom distinct from those of their shorter-lived relatives.

In this chapter, I also discuss the influence of taxonomic hierarchy and temporal

divisions on the patterns observed, followed by the relationship between mass

extinctions and longevity. Finally, I examine the potential biases in this study and

consider various definitions of “long-lived.”

Material and Methods

There is no available phylogenetic framework for comparing rates of character

transformation in the global pool of fossil crinoids. Likewise, there are no detailed

samples of crinoid lineages in a stratigraphic column for investigating character

� 25

reversals, convergence or the lack thereof. However, data for a quantitative,

comparative study are available as follows.

The data

I use previously compiled data from a database containing 1195 crinoid species

representing 752 genera, together with their first and last fossil appearances and 90

morphological characters from the column, cup and arms (Foote 1999, http://

geosci.uchicago.edu/~foote/MORPHDAT/CRINOID.DAT). Additionally, I code seven

crinoid genera not already represented in this database (Appendix A). I follow the

character system used by Foote (1999), which is in turn based on the traditional

homologies used in the Treatise (Moore & Teichert 1978). These seven genera were

coded using photographs and descriptions from Moore & Teichert (1978), Schubert et

al. (1992), Ausich (1998) and Guensburg & Sprinkle (2003). I choose, either the

earliest appearing taxa, or what is believed to be ancestral by crinoid workers to the best

of my knowledge of the current literature, as taxa bearing basal morphology (Appendix

B). Multiple taxa are used as basal taxa when there is uncertainty in the literature over

the identity of the most basal taxon for a group.

The characters are binary, ordered multi-state or unordered multi-state. Not all

characters are applicable to all species. For instance, most comatulids, and all those

coded in Foote (1999), have no columns and hence their columns characters are coded � 26

as “inapplicable.” The morphological characters used here are not assumed to be

strictly homologous, only to reflect general fossilizable morphology determined

consistently within the crinoid bauplan (Foote 1999).

The geologic duration for each taxon (henceforth “duration” in millions of years, M.y.)

is the difference between the bottom of the geologic stage of the first occurrence to the

top of the stage of the last appearance of the taxon. Relative durations rather than

absolute durations are of greatest importance in this comparative framework. The time

scale is based mainly on Harland et al. (1990), but other references were also used for

stratigraphic correlation (see Foote 1994a p.322). Genera with first and last

appearances not resolved to stage level are omitted (79 out of 752 genera, ~10 %).

Results do not change qualitatively if these genera are included.

Crinoid family durations are extracted from an updated version of Sepkoski's family

database (Sepkoski 1982, pers. comm. Foote), Benton's Fossil Record 2 (1993) and

updated using Webster's online database (Webster 2003) where inconsistencies due to

taxonomic revisions or range extensions are apparent. Average and median genus

durations in families are calculated based on the genera sampled in Foote (1999). It

should be noted that family durations are less updated than genus durations but

durations in each set are updated to the same extent.

� 27

Data units

I use the genus as the basic data unit. I mainly analyze crinoid genera grouped in higher

taxa (orders and cladid suborders) and compare results from these separate analyses.

Where sample sizes permit, I analyze the data grouped as families. Genera are

convenient units of analysis because data on their fossil durations are more complete

than data for species. The morphological distances between fossil crinoid genera are

more than the morphological distances among species within the same genus (pers.

comm. Foote), suggesting that there is nothing unusually problematic with the genus

level justification.

I focus most of the analyses on genera grouped as orders because multivariate

morphologies between different crinoid orders can be dramatically different. Characters

relevant to one order may be wholly inapplicable to another, making comparisons

dubious since only a few characters can be used to calculate distances. Another reason

to use orders instead of suborders or families is to keep genus sample sizes large enough

for analyses to be robust to resampling (Table II.1). The sample sizes of genera in

families and most suborders are mostly too small.

Following Foote (1999), I delineate orders (Table II.1) based mainly on the Treatise

(Moore & Teichert 1978). This grouping is still widely accepted today (Simms 1999). I

omit Encrinida and Hybocrinida for order level analyses because they were both� 28

Ta

ble

II.

1.

Crin

oid

ord

ers

(a

nd

su

bo

rde

rs o

f cla

did

s)

an

d t

he

mo

rph

o-d

ura

tio

n p

lot

dis

trib

utio

ns o

f th

eir g

en

era

. N

= n

um

be

r o

f g

en

era

sa

mp

led

,

an

d R

=sa

mp

led

ge

olo

gic

ra

ng

es (

J =

Ju

rassic

, K

= C

reta

ce

ou

s,

Mi =

Mio

ce

ne

, O

= O

rdo

vic

ian

, P

ale

= P

ale

oce

ne

, P

= P

erm

ian

, T

= T

ria

ssic

,

Ste

= S

tep

ha

nia

n;

l, m

, u

in

pa

ren

the

se

s in

dic

ate

lo

we

r m

idd

le,

an

d u

pp

er

resp

ective

ly).

C

olu

mn

s S

T-b

-LL

-t a

re t

he

qu

ad

ran

ts n

am

ed

in

Fig

. II

.1.C

an

d n

um

be

rs in

dic

ate

th

eir p

rop

ort

ion

occu

pa

tio

n.

Me

d,

Mid

, a

nd

10

-g a

re c

uto

ff p

oin

ts f

or

du

ratio

ns o

f lo

ng

-liv

ed

ge

ne

ra d

efin

ed

in "

Ma

teria

ls a

nd

Me

tho

ds."

Nu

mb

ers

in

th

e "

pro

po

rtio

n"

co

lum

ns in

dic

ate

th

e p

rop

ort

ion

s o

f ra

refie

d s

am

ple

s o

f sh

ort

-liv

ed

ge

ne

ra t

ha

t a

re

less d

evia

nt

(or

mo

re u

nsp

ecia

lize

d)

tha

n lo

ng

-liv

ed

ge

ne

ra f

or

ea

ch

de

fin

itio

n o

f lo

ng

-liv

ed

. R

ho

= r

an

k o

rde

r co

rre

latio

n c

oe

ffic

ien

t

be

twe

en

mo

rph

olo

gic

al d

evia

tio

ns a

nd

du

ratio

ns f

rom

th

e e

mp

rica

l d

ata

. N

ote

th

at

the

to

tal n

um

be

r o

f g

en

era

is g

rea

ter

tha

n t

he

ord

ers

su

mm

ed

be

ca

use

of

the

un

ce

rta

in h

igh

er

leve

l ta

xo

no

mic

assig

nm

en

ts o

f so

me

ge

ne

ra.

Th

e la

st

row

sh

ow

th

e s

am

e v

alu

es f

or

all

ge

ne

raco

nsid

ere

d.

* =

sig

nific

an

t a

t th

e p

=0

.1 le

ve

l, *

* =

sig

nific

an

t a

t th

e p

=

0.0

5 le

ve

l a

nd

***

= s

ign

ific

an

t a

t th

e p

= 0

.01

le

ve

l w

he

n c

om

pa

rin

g

with

ra

nd

om

ize

d d

ata

.

Su

bcl

ass

(Su

b)O

rder

NR

SL

-bS

L-t

LL

-bL

L-t

Med

(M.y

.)

Med

Pro

port

ion

Mid

(M.y

.)

Mid

Pro

port

ion

10-g

(M.y

.)

10-g

Pro

port

ion

Rh

o

Art

icula

taR

ovea

crin

ida

12

K0.3

80.2

30.2

30.1

518

0.1

76

24

0.2

49

41

0.1

78

0.0

26

Cyrt

ocr

inid

a28

P(m

)-P

ale

0.3

20.3

60.1

40.1

825

0.0

00

47

0.1

15

88

0.1

68

-0.3

09

Com

atuli

da

36

T(u

) –M

i0.6

40.1

10.2

50.0

024

0.0

00

52

0.0

01

67

0.0

23

-0.5

57***

Mil

leri

crin

ida

8T

(l)-

J(u)

0.1

30.7

50.1

30.0

010

0.0

00

34

0.0

00

65

0.0

00

-0.6

83*

Isocr

inid

a17

T(l

)–M

i0.7

00.1

80.1

20.0

037

0.0

00

120

0.1

51

206

0.1

50

-0.2

06

(Inad

unat

a)C

ladid

a

-

all

262

O(l

)-P

0.7

30.2

10.0

50.0

220

0.0

03

55

0.9

58

41

0.8

09

-0.2

03***

-

O-D

71

0.8

00.1

30.0

70.0

013

0.0

02

45

0.3

42

41

0.0

45

-0.2

96**

-

Car

b(l

)83

0.8

60.0

40.1

0.0

133

1.0

00

53

0.7

50

59

0.7

52

-0.0

84

-

Car

b(u

)–P

112

0.7

90.1

40.0

40.0

420

0.0

00

57

0.9

93

42

0.9

62

-0.1

51

� 29

Table

II.1. (c

on't)

Crinoid

ord

ers

(and s

ubord

ers

of cla

did

s)

repre

sente

d in the d

ata

base a

nd the m

orp

ho-d

ura

tion p

lot dis

trib

utions o

f

their g

enera

.

Su

bcl

ass

(Su

b)O

rder

NR

SL

-bS

L-t

LL

-bL

L-t

Med

(M

.y.)

Med

P

rop

ort

ion

Mid

(M

.y.)

Mid

P

rop

ort

ion

10

-g

(M.y

.)1

0-g

P

rop

ort

ion

Rh

o

(In

adu

nat

a)C

lad

ida

-

(Cy

ath

ocr

inin

a)3

70

.76

0.0

80

.03

0.1

41

41

.00

05

41

.00

09

31

.00

0-0

.08

8

-(

Den

dro

crin

ina)

34

0.5

60

.06

0.3

50

.03

15

0.0

01

45

0.3

20

41

0.3

74

0.0

26

-(

Po

teri

ocr

inin

a)1

79

O(u

)-P

(m)

0.3

70

.56

0.0

60

.02

20

0.0

00

53

0.0

10

42

0.0

03

-0.1

11

Dis

par

ida

80

O(u

)-P

(m)

0.5

80

.38

0.0

30

.03

15

0.8

90

43

0.4

27

41

0.0

68

0.0

31

Cam

erat

aD

iplo

bat

hri

da

38

O0

.82

0.1

00

.08

0.0

01

10

.00

03

30

.04

63

00

.02

3-0

.27

7

Mo

no

bat

hri

da

80

O(m

)-S

te0

.60

0.3

00

.08

0.0

12

10

.00

04

70

.12

94

60

.05

3-0

.22

3*

Fle

xib

ilia

Tax

ocr

inid

a1

1O

(m)-

Ste

0.2

50

.38

0.2

50

.13

18

0.0

00

26

0.1

82

46

0.2

87

-0.2

65

Sag

eno

crin

ida

48

O-P

0.2

90

.64

0.0

30

.03

14

0.9

90

60

0.4

03

51

0.9

46

0.0

65

[All

Gen

era]

67

1O

-Mi

0.7

80

.01

0.2

10

.00

11

60

.70

91

80

.45

15

00

.99

4-0

.03

0

� 30

represented by fewer than 10 species and only four and five genera, respectively, in

Foote (1999). The cladids are a very large and diverse order. For the of comparability

of sample sizes and morphological ranges with other orders, I divide them in two ways

for most of the analyses, in addition to analyzing them as an entire set (Table II.1).

First, I divide them into Ordovician to Devonian, Lower Carboniferous and Upper

Carboniferous to Permian subsets. There is little overlap of cladid genera between these

time intervals. Second, I divide them into the suborders Cyathocrinina, Dendrocrinina

and Poteriocrinina (Moore & Teichert 1978, Ausich et al. 1999). It should be noted that

current workers no longer believe that Dendrocrinina and Poteriocrinina are

monophyeletic (e.g. Simms 1999).

Data treatment

Most genera (71%) in the database were coded with one species. Where genera are

represented by more than one species (or families by more than one genus), an

“average” genus (or family) was determined by taking averages of the character states

for each character. Characters are averaged by: i) taking the arithmetic mean of binary

and ordered characters, ii) taking the modal value of unordered characters unless the

character states are equally common, in which case a state is chosen randomly. When

there are more unknown or inapplicable states than known ones among the species

representing a genus, the character is coded as “not applicable” for that genus.

However, if known states are in the majority, they are treated as in i) and ii) but � 31

averaged as if there are no non-applicable states. Detailed information on the

morphological characters and sampling issues can be found in Foote (1994a, 1994b,

1995a, 1995b, 1999).

I define long-lived taxa in three ways for the purpose of analysis: 1) Taxa having

durations greater than the median duration, “Med,” of the taxa within the group being

examined; 2) Taxa having durations greater than the mid-range, “Mid,” duration of the

same; 3) The most long-lived 10% of genera within the group, “10-g,” when ranked

according to durations (Fig. II.1.B).

Morphological deviations from group means

I convert the original discrete character matrix for genera into a Euclidean distance

matrix. Missing or non-applicable characters are not used when calculating distances

(per taxon pair). Subsequently, I perform Principal Coordinates Analyses (PCO)

(Gower 1966) by executing principal components analyses on the Euclidean distance

matrix subdivided into (sub)orders and time intervals. I then calculate morphological

distances for each genus by taking the sum of absolute differences between the genus

PCO scores and the mean PCO scores for the (sub)order or the time interval and term

these morphological deviations. The improvement in fit between pairwise distances of

discrete characters and principal coordinate scores trails off after the first ten PCO

� 32

scores (see Foote 1999, fig. 44). Thus, only the first ten PCO scores (accounting for

about 75% of the variance) are used in all calculations.

I also summarized morphology using Non-Metric Multidimensional Scaling (NMDS),

where only rank order information of the distance matrix is used (Kruskal 1964). I

calculate NMDS-based morphological deviations for each genus with respect to their

order mean of NMDS scores, as done with the PCO scores described above. Results

using NMDS are not qualitatively different from those using PCO and are not reported

here. Concordant results from these different methods with different assumptions

indicate that patterns observed are not affected by the multivariate method used.

To summarize morpho-duration relationships graphically, I plot the resulting

morphological deviations versus durations of respective crinoid taxa (morpho-duration

plots as in Fig.II.1.B). I also plot character states versus durations, for each character,

within each order.

I compare means of morphological deviations of each set of long-lived taxa with

rarefied samples of their corresponding shorter-lived relatives. This is because apparent

patterns may be due to sampling artifacts (there are fewer long-lived taxa than shorter-

lived ones). For instance, if there are four long-lived taxa in an order, defined using

“Mid”, I randomly sample, without replacement, four taxa from the corresponding � 33

shorter-lived pool. This is repeated 10,000 times for each order and for each definition

of long-lived. Means of morphological deviations of the 10,000 rarefied samples of

shorter-lived taxa are compared with the means of long-lived taxa to see how often the

latter have smaller values than the prior.

I also calculate rank order correlation coefficients between morphological deviations

and durations. To compare empirical values with a null expectation, I used the method

“randomization” by resampling with replacement from morphological deviations. I

then do the same for durations, separately, drawing the number of data pairs

corresponding to the number of genera (or families) represented in the empirical data of

the higher taxon (order, subclass or class) 1000 times to form null distributions (Efron

& Tibshirami 1993). I calculate the same rank order correlations for the randomized

datasets and compare them with the empirical datasets. Because the resulting morpho-

duration plots appear exponential, I also fitted exponential decay curves to obtain best-

fit parameters (k and m in y = me-kx). This is to provide an alternative to comparing

distributions using a linear fit as implied by the calculation of correlation coefficients.

Conclusions are no different using a linear fit and are not further reported.

As an alternative test, I divide the morpho-duration plots into four equi-area quadrants

to conduct χ2 tests. I use the most and least deviant and the shortest- and the longest-

lived genera, to delimit the occupied area on the plot. Then I divide this area into four � 34

equi-area quadrants by drawing a vertical line thorough the mid-range duration and a

horizontal line through the mid-range morphological deviation (Fig. II.1.C). I then

compare the density of the four quadrants of empirical and randomized datasets using χ2

tests.

Morphological deviations from basal morphology

To investigate the morphological deviation of genera from putative basal morphologies

of their orders, I calculate pairwise euclidean distances from data on raw morphological

characters. I also calculate Manhattan and Canberra distances (Sneath & Sokal 1973);

results do not differ from Euclidean distances and are not further reported. I then plot

the distances of each genus from the basal member(s) versus the duration of each genus

(Fig II.1.B). I compare 10,000 rarefied samples from shorter-lived taxa with long-lived

taxa as described above for mean morphologies, in each case. I also calculate and

compare rank order correlation coefficients from empirical and randomized datasets as

described above. Lastly, I also perform χ2 tests as described above.

Results and Discussion

Morphological deviations from group means

There are wide ranges of morphological deviations of shorter-lived genera from their

order means. Longer-lived taxa have smaller ranges of morphological deviations from

� 35

/Users/LH

0

/Users/LH

20 40 60 80 100

0.0

0.2

0.4

0.6

0.8

1.0

Fig. II.2.A.Monobathrida, N = 80

Duration (M.y.)

Mo

rph

olo

gic

al d

evia

tio

n f

rom

ord

er

me

an

10-gMedian Mid

SL-t LL-t

LL-b

Fig. II.2.A Morphological deviation of each monobathrid genus from its order mean versus the duration of each genus. Solid lines show the delimiting of the plot into four equi-area quadrants (as in Fig. II.1.C). Dotted lines show the median duration value (Median), mid-range duration value (Mid), and the duration greater than those associated with the 10% most long-lived genera (10-g).

� 36

/Users/LH

0

/Users/LH

20 40 60 80 100

0.0

0.2

0.4

0.6

0.8

1.0

Fig. II.2.B.Disparida, N = 80

Duration (M.y.)

Mo

rph

olo

gic

al d

evia

tio

n f

rom

ord

er

me

an

10-gMedian Mid

SL-t

SL-b

LL-t

LL-b

Fig. II.2.B Morphological deviation of each disparid genus from its order mean versus the duration of each genus. Solid lines show the delimiting of the plot into four equi-area quadrants (as in Fig. II.1.C). Dotted lines show the median duration value (Median), mid-range duration value (Mid), and the duration greater than those associated with the 10% most long-lived genera (10-g).

� 37

the order means (Fig. II.2A and B). Most taxa are shorter-lived and morphologically

“average.” In other words, the bottom right of morpho-duration plots (SL-b) are often

the most dense (Table II.1, Fig.II.2.A and B).

When the definition of long-lived is taken as having a duration greater than the median

value, “Med,” 12 out of 17 (sub)orders have long-lived genera that at best have a 0.003

proportion chance of being more deviant than shorter-lived genera (Table II.1). In other

words, in 12 out of 17 cases, long-lived taxa are less specialized when compared with

shorter-lived taxa. If the definition “Mid” is taken with a 0.1 proportion chance as a

cut-off, six out of 17 cases have long-lived genera that are less deviant than shorter-

lived genera. Finally, if we take the last definition, “10-g,” there are ten out of 17

cases. Note that temporal and suborder divisions of cladids both have morphologically

deviant long-lived taxa but these cases cannot necessarily be taken as independent.

I conclude that long-lived genera are often less specialized than expected. However, as

the definition of long-lived becomes more stringent such that fewer longer-lived taxa

are included (from a most relaxed definition, “Med” to more stringent definitions “Mid”

and “10-g”), the above conclusion holds true for fewer cases.

However, when all crinoid genera are examined in concert, long-lived genera are more

deviant, under the long-lived definition using “10-g,” than rarefied samples of shorter-� 38

lived ones. Using the definitions “10-g” and “Mid,” the chances of being deviant are

even (Table II.1). Note that fewer characters can be universally compared when all

genera are lumped in a single analysis and that this shortcoming may obscure potential

patterns.

Correlations between morphological deviations and durations are negative for 13 out of

17 cases (Table II.1) and five of those 13 are significant at a p > 0.1 level. This test is

sensitive to the density of the lower right quadrant, thus there are few significant

correlations. However, the direction of the correlation agrees with the previous

observation that longer-lived taxa are less morphologically distant from the mean

morphology of their group than shorter-lived taxa, using a rarefaction approach. In

contrast, none of the cases were significantly different from random using a χ2 test (data

not shown), indicating that the pattern is blurred when the distribution within each

quadrant is ignored. However the distribution of taxa in each of the four quadrants give

us a quick view of the emptiness the long-lived, deviant (LL-t) quadrant (Table II.1, Fig

II.1.C).

Plots of character states of each character versus durations for each order show that the

vast majority of rare character states are associated with shorter durations (plots not

shown). Viewed alternatively, morphologically deviant taxa suffer extinction sooner.

� 39

Morphological deviations from basal morphology

Long-lived taxa do not typically appear very different from taxa having basal

morphologies (LL-t is often very empty, Appendix B). The majority of genera are

present in the quadrant SL-b in many of the crinoid (sub)orders examined, irrespective

of the basal taxon used (Appendix B). Despite that, rarefied samples of shorter-lived

taxa show that long-lived taxa do not necessarily have morphologies that are less

deviant from ancestors than shorter-lived taxa. Morphological deviations from basal

members are highly contingent upon the morphology of the putative basal member

used.

I illustrate with Sagenocrinida as an example (Fig. II.3), how the distribution of genera

in the morpho-duration plot shifts according to the basal taxon used as a reference.

Protaxocrinus has been confidently placed as a model of a direct ancestor to the clade

of sagenocrinids (Lane 1978). The cloud of points shifts from being closer to the basal

morphology in the comparison with Protaxocrinus to being in a more distant position

with Cupulocrinus, which is ancestral to Protaxocrinus (Moore & Teichert 1978).

However, when a very distant basal taxon is used, the spread of genera decreases

because they are all distant from the reference taxon (Glenocrinus) and few characters

are comparable between Glenocrinus and sagenocrinids. This explicitly illustrates that

when long-lived forms are assessed for their primitiveness, how far removed the basal

states are from the taxa being compared affect our conclusions. For each of the basal� 40

/Users/LH

20 40 60 80 100

10

20

30

40

Fig II.3Distances from basal morphologies

Duration (M.y.)

Eu

clid

ea

n d

ista

nce

fro

m b

asa

l m

orp

ho

log

y

*

***

*

*

*

***

*

****

**

***

******

*

***

*

*

*

*

*

*

*

*

/Users/LH

*

****

*

*

**

*

*

*

***

*

Fig. II.3. Euclidean distances between genera of Sagenocrinida and taxa with alleged basal morphologies (Glenocrinus = solid circles, Protaxocrinus = asterisks, and Cupulocrinus = open circles), versus their durations.

� 41

taxa and each definition of long-lived taxon used, long-lived Sagenocrinids tend to be

more deviant than rarefied samples of shorter-lived forms at least 82% of the time

(Appendix B) even though a χ2 test does not indicate any significant difference from

random (as is the case for all the other comparisons in other taxa, data not shown).

Influence of taxonomic hierarchy

Analyses done at different taxonomic ranks potentially show different patterns of

morpho-duration plot occupation. For family-within-order comparisons, as fewer long-

lived taxa are included (from “Med,”to a more stringent definitions “Mid” and “10-g”),

the observation that long-lived taxa are morphologically less deviant holds true less

(Table II.2), as for genus-within-order comparisons. Under “Med”, Articulata, Cladida,

Disparida, Camerata, Sagenocrinida, Flexibilia (Sagenocrinida and Taxocrinida) have

long-lived families likely to be less specialized than shorter-lived ones at most 9 % of

the time (Table II.2, 6 out of 8 cases).

The distributions are not significantly different from null distributions of random

expectations using randomized data (Table II.2) and a χ2 test (data not shown), except

in cladid families. This is again despite the fact that i) SL-b is the most densely filled

quadrant and LL-t the sparsest and ii) rank order correlations of morphologies and

durations are negative in all cases.

� 42

Ta

ble

II.

2.

Crin

oid

fa

mili

es a

nd

th

eir m

orp

ho

-dura

tio

n p

lot

dis

trib

ution

s.

N=

nu

mb

er

of

fam

ilie

s s

am

ple

d,

Sl-t-

LL

-t a

re t

he

qu

ad

ran

ts a

s n

am

ed in F

ig.

II.1

.C a

nd the

num

be

rs in

dic

ate

the

ir p

ropo

rtio

n

occu

pa

tio

n.

Me

d,

mid

an

d 1

0-g

are

cu

toff

poin

ts f

or

du

ratio

ns o

f lo

ng

-liv

ed g

en

era

as d

efine

d in "

Mate

ria

ls a

nd M

eth

od

s."

Nu

mb

ers

in

th

e "

pro

po

rtio

n"

co

lum

ns ind

ica

te t

he

pro

port

ion

s o

f ra

refie

d s

am

ple

s o

f sho

rter-

lived

fam

ilie

s th

at a

re le

ss d

evia

nt

(or

mo

re s

pe

cia

lize

d)

tha

n lo

ng

-liv

ed f

am

ilie

s f

or

each

definitio

n o

f "l

on

g-liv

ed

". R

ho

= r

an

k o

rder

co

rre

latio

n c

oeffic

ient

be

twe

en

mo

rph

olo

gic

al d

evia

tio

ns a

nd

dura

tio

ns fro

m t

he

em

piric

al d

ata

.* =

sig

nific

an

t a

t th

e p

< 0

.1 w

he

n c

om

pa

ring

with

ran

do

miz

ed

da

ta.

Gro

up

NS

L-t

SL

-bL

L-b

LL

-tM

ed

(M.y

.)M

ed

Pro

port

ion

Mid

(M

.y.)

Mid

P

rop

ort

ion

10-g

(M

.y.)

10-g

P

rop

ort

ion

Rh

o

Art

icula

ta28

0.2

90.5

40.0

40.1

445

0.0

00

94

0.2

61

145

0.7

60

-0.2

3

Cla

did

a59

0.1

50.6

40.0

50.1

532

0.0

00

81

0.7

24

104

0.9

46

-0.2

3*

Dis

par

ida

17

0.2

90.4

70.0

60.1

842

0.0

80

89

0.4

77

161

0.7

61

-0.0

6

Monobat

hri

da

18

0.1

70.3

30.1

70.3

341

0.5

33

44

0.5

37

83

0.1

16

-0.1

7

Cam

erat

a (=

M

onobat

hri

da

+

Dip

lobat

hri

da)

29

0.3

10.4

80.0

00.2

117

0.0

88

66

0.0

32

83

0.0

95

-0.2

9

Sag

enocr

inid

a11

0.2

70.1

80.0

90.4

688

0.0

11

81

0.0

21

127

0.6

95

-0.3

6

Fle

xib

ilia

(=

S

agen

ocr

inid

a +

T

axocr

inid

a)14

0.2

90.2

10.0

70.4

372

0.0

22

81

0.0

21

128

0.2

16

-0.4

3

Cri

nid

a (=

all

fam

ilie

s)147

0.3

30.5

90.0

50.0

439

0.1

17

122

0.9

85

121

0.9

98

-0.0

1

� 43

Patterns can also be modified depending on the inclusiveness of the higher taxon. The

pattern of morpho-duration plot occupation for orders by either genera or families

remains qualitatively similar (e.g. Cladida, Disparida and Monobathrida) but there are

exceptions (e.g. Sagenocrinida, compare Tables II.1 and II.2). Also, when both the

camerate orders, Diplobathrida and Monobathrida, are combined in a single analysis,

the quadrant SL-t becomes 15% more occupied than when Monobathrida is considered

alone, at the expense of SL-b and LL-t (Table II.2). This illustrates the inherent

problem of empirical morphospaces where sampling strongly influences the shape of

the space (McGhee 1999).

Temporal divisions and mass extinctions

Dividing the genera into geologic periods, instead of taxonomic grouping, illustrates

that the occupation of morpho-duration plot is quite stable through time (Fig. II.4). Just

as in previous analyses when genera were grouped according to orders, genera in each

period are mostly short-lived. However, rarefied samples of shorter-lived genera

through each period inform us that the long-lived taxa can be more, less or equally

deviant compared with shorter-lived taxa of an equivalent sample size (Appendix C).

Genera that are extremely long-lived within each order are also more likely to have

passed through one or more mass extinctions (Raup & Boyajian 1988) than other genera

in the database (Wilcoxon rank test α = 0.05, p << 0.0001), even though passing� 44

/Users/LH/Users/LH

Fig II.4 Occupation of quadrants through time

pro

po

rtio

n f

ille

d

4 4 4 44

4 4

33

33

3 3 3

2

2

2 2

2

2 2

1

1

1

1

1

1 1

O S D C P Tr K

0.0

0.2

0.4

0.6

0.8

1.0

Fig. II.4. Changing relative occupation of morpho-duration plot quadrants (see Fig. II.1) through time, as calculated as a proportion of the total number of genera found in the named period. All applicable crinoid genera were used. O = Ordovician, S = Silurian, D = Devonian, C = Carboniferous, P = Permian, K = Cretaceous through Eocene. 1 = SL-b, 2 = SL-t, 3 = LL-b, 4 = LL-t of Fig.II.1.

� 45

through mass extinctions does not necessarily ensure persistence (e.g. Monachocrinus

and Alisocrinus, data not shown).

Post-Paleozoic orders (Isocrinida, Comatulida, Cyrtocrinida) are more likely to have

longer-lived families than Paleozoic ones. This is consistent with the trend of a secular

increase in longevity through time (Gilinsky 1994), probably due to a decrease in

extinction rates throughout the Phanerozoic (Raup & Sepkoski 1982). It also

corroborates the claim that the likelihood of the occurrence of living fossils or long-

lived taxa increases with time (Holman 1999).

Potential biases

The database has many inapplicable character entries and unknown characters. These

entries result from the effort to sample as many crinoids as possible and to include

characters that describe them both comprehensively and comparatively. One concern

expressed in earlier work is that long-lived taxa may simply be characterized by fewer

characters and are likely to be results of taxonomic lumping (Schopf et al. 1974).

However, a plot of the percentage of inapplicable entries and unknown characters shows

no consistent relationship with the duration of sampled genera (Fig. II.5, Spearman's

rank correlation ρ = 0.09, p = 0.99). Plotting the same data separately for each order

yields the same results (data not shown).

� 46

/Users/LH

0 50 100 150 200

01

02

03

04

05

06

07

0Fig II.5

Unknown and inapplicable crinoid characters

Duration (M.y.)

Pe

rce

nt

/Users/LH

x

x

x

x

x

x

xxx

x

x

x

xx

x

x

x

x

x

xx

x

x

x

x

x

x

x

x

x

xxx

x

x

x

x

xx

x

x

xxx

x

x

xx

x

x

x

xx

xx

x

/Users/LH

x

xxxx

x

x

x

x

x

x

x

x

x

xx

x

x

x

x

x

xx

xxx

x

xxxxxxx

x

xx

xx

xxx

xxxxxxxxx

xx

x

xxxx

x

x

x

xxxxx

xx

/Users/LH

x

x

x

xx

x

xx

xx

x

x

x

xxxxx

x

xx

xx

xx

/Users/LH

xxx

xxxx

x

x

x

xxx

x

x

x

xx

x

x

xx

xxx

x

xx

xx

xx

x

x

xx

xx

x

xx

x

xxx

x

x

x

xx

x

xx

x

xx

/Users/LH

xx

x

x

x

x

x

xx

x

xxx

x

x

xxxxx

x

xx

xx

x

xx

x

x

x

x

x

x

xx

x

x

x

xx

x

x

x

xx

x

xxxx

x

x

xxx

x

xx

xx

xx

x

x

xx

xx

x

xxx

x

xxx

xx

xx

xxxx

x

x

x

xx

x

x

xx

x

x

/Users/LH

x

x

x

x

x

x

x

x

x

xx

x

x

x

x

x

xx

xxxx

x

xxxxx

xx

x

x

x

xx

x

x

x

x

xx

x

x

x

xx

xx

x

x

x

xx

x

x

xx

x

x

x

x

x

x

x

xxxx

xxxxxxx

x

x

x

x

x

xxxx

xxxxx

x

x

x

x

x

x

x

x

x

xxx

xx

x

x

x

x

/Users/LH

x

x

x

x

x

xx

x

x

xx

x

x

x

x

x

x

xx

x

x

xx

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

xx

x

xx

x

x

x

xxxxxx

x

xx

x

xx

x

xx

x

xxx

xx

x

xxxxxx

x

x

xx

x

x

xx

x

xx

x

xx

x

x

xxxx

x

xx

x

xx

xxx

x

x

/Users/LH

xxxx

x

xxx

x

x

x

xxx

xx

xxxx

x

x

x

x

x

x

x

x

x

xx

xxxx

x

xx

x

x

xx

x

x

xx

x

x

x

xx

xxx

x

xx

xx

xxx

xx

x

x

x

xx

x

x

xx

x

x

x

x

x

x

x

x

x

x

x

x

xxx

x

xx

x

x

xx

x

xxxx

x

x

x

x

x

x

/Users/LH

x

x

x

x

x

xx

xx

x

xx

xx

xx

x

xxx

xx

xx

x

x

x

x

x

x

x

xx

x

x

x

x

xx

xx

x

x

x

x

x

x

x

x x

xx

Fig. II.5. Percentages of unknown and inapplicable characters in 751 crinoid genera (Foote 1999) as a function of their fossil durations. The two outliers to the far right are Isocrinus and Chladocrinus. Crosses represent percentages of inapplicable characters for each genus and open circles are unknown ones.

� 47

Geologically older stratigraphic stages are longer than more recent ones (Sepkoski

1975), so that longer apparent durations may be related to early first appearances. No

such problem exists in this database. On the contrary, later appearing (geologically

younger) genera tend to be somewhat longer-lived although the correlation is not

significant (Spearman's rank correlation ρ = 0.264, p = 0.99, data not shown).

Long-lived genera are represented by significantly more species than genera with

shorter durations (Wilcoxon rank test, α = 0.05, p << 0.0001, one-tailed t-test, α = 0.05,

p << 0.0001 ). This could be potentially a problem for the conclusions drawn, because

long-lived genera could have had their morphologies “averaged out” by multiple

representative species. However, when single random species are used as

representatives for each long-lived genus (instead of averaging multiple species), there

is no change in the patterns seen (data not shown).

Finally, some of the post-Paleozoic crinoids have truncated range distributions, that is,

their histories have not ended because they are still extant. However, more Paleozoic

crinoids are represented at both ordinal and genus levels in this chapter (Table II.1).

Moreover, most of the post-Paleozoic crinoid genera represented are already extinct

(Foote 1999). Those that are extant (Isocrinus, Chladocrinus, Cyathidium) are already

longer-lived than the extinct taxa, except in the case of Comatulida where Atelecrinus,

Pterometra and Himerometra, which are considered shorter-lived and are still extant. � 48

Explicit definitions of long-lived taxa

Thus far, “long-lived taxa” have been defined in three explicit and distinct ways (see

Fig. II.6). Other definitions of “long-lived” can lead to selection of different sets of

genera (Fig. II.6 with Monobrathida as an example). As expected, identifying a greater

percentage of genera as long-lived increases the number of long-lived genera.

However, the steps of increase are not always equal even though I have shown increases

in steps of 5% (fig. 3, 5-20%). This is because genera occasionally have the same

numeric value of duration (in this case, many that fall in the 20% category are listed as

having a duration of 34 M.y.) due to issues of stratigraphic resolution of age dating. We

can also use other definitions to delimit long-lived taxa: mid-range or median duration

values, an obvious break in the duration distribution of the group in question, the ability

to pass through mass extinctions or statistical tests for outliers (e.g. Grubb's test see

Sokal & Rohlf 1995, Fig. II.6). “Living fossils” or clades that persist for long periods

displaying little evolution (as defined by Stanley 1979) are sometimes, but not always,

also long-lived taxa (Appendix D). Similarly, longest-lived genera do not necessarily

reside in long-lived families (defined as the most long-lived 10% of families in the

dataset), although they will, if the definition of a long-lived family is relaxed.

In summary, longevity is relative and dependent on taxonomic inclusiveness. These

important axioms are often neglected in papers addressing extreme persistence or

morphological conservatism. � 49

/Users/LH

Fig. II.6 Variation in the numbers of long-lived taxa

Frequency

5%

10%

15%

20%

Med

Mid Br

Com

Me

G(.05)

G(.01)

UQ

010

20

30

40

Fig II.6. Frequency of occurrence of long-lived genera of monobathrids, depending on the definition of ”long-lived.” The categories 5%, 10% (= 10-g), 15% and 20% are percentages of the most long-lived genera with respect to all the monobathrid genera sampled. Med = taxa having durations more than the median value. Mid = taxa having durations greater than the mid-range value, Br = taxa occurring at durations greater than the break in longevity distribution. Com= monobathrid genera identified in Appendix D via the combination approach. Me = taxa passing through at least on mass extinction. G (0.05) and G(0.1) are genera that pass Grubb’s test (Sokal & Rohlf 1995) at p = 0.05 and 0.1 respectively. UQ refers to genera having durations > 1.5 times the seventy-fifth percentile of the genera.

� 50

Conclusions

Longevities of crinoid species and genera have been previously linked to their ecology

(Baumiller 1993, Kammer et al. 1997, 1998). This research extends the scope of those

studies to include post-Paleozoic crinoids and multiple morphological characters (cf.

Baumiller 1993). I also use multiple analytical methods to check the robustness of the

relationship between morphology and longevity, in order to decrease the likelihood that

conclusions drawn are artifacts of the methodology employed.

The following general conclusions can be drawn. First, most taxa (genera and families),

are short-lived and “average” such that SL-b (Fig II.1.C) is the most densely filled

quadrant of the mopho-duration plot. In contrast, the sparsest quadrant is LL-t,

implying that experiments in morphology are usually not long-lived. Second, long-lived

genera within orders are often less morphologically deviant or less specialized than

expected when compared with rarefied samples of corresponding shorter-lived genera.

In other words, long-lived genera are not only not unusual, some are unusually average!

Third, patterns of morphological deviations from basal morphologies, versus durations,

are unclear. Details of morpho-duration plot occupation vary according to the basal

member employed in the analyses. Despite this uncertainly, the short-lived, non-deviant

quadrant of the morpho-duration plot (SL-b) is still much denser than the long-lived,

deviant quadrant (LL-t), in general. Fourth, morpho-duration plot occupation through

time (as in the case for genera-within-order and families-within-order) follows the � 51

density order of SL-b, SL-t, LL-b, LL-t. This pattern hold true even though

comparisons of rarefied samples do not show that long-lived genera are comparatively

more or less deviant as a rule. Fifth, taxonomic ranks and inclusiveness of higher taxa

are critical factors when discussing longevity because identities of long-lived taxa may

dramatically change according to these factors. Last, identities of long-lived taxa may

change with respect to the definitions of longevity used. This may or may not (as was

the case for this article) change conclusions being drawn on long-lived taxa.

Small size, ecological tolerance, wide geographical ranges, large population sizes,

planktotrophic larvae and deeper depth distributions may lower extinction risk (Boucot

1975b, Buzas & Culver 1984, Stanley 1986, Jablonski 1986a, Raup & Boyajian 1988,

Schopf 1994, Oji 1996, Jeffery & Emlet 2003). Also, recovery genera of the post-

Paleozoic seem to have greater temporal longevities (Miller & Foote 2003). Perhaps

“extinctions are not biologically random” (Jablonski 1989, McKinney 1995), implying

that persistence is not either. Based on the results of the current study, I rule out the idea

that long-lived genera are morphologically deviant or unusual when compared within

the realm of an order.

There are of course many unanswered questions. This study focused on persistence but

there is no available information on actual rate of character evolution: Do long-lived

taxa experience rapid rates of character reversals or zig-zag evolution (Hennigsmoen � 52

1957) and the apparent persistence is only a sampling artifact? Or, does persistence

necessarily mean slow change or cryptic change (Knowlton 2000)? Can the morpho-

duration plot patterns in crinoids be extrapolated to other organisms? To remain similar

enough to an ancestor so that a lineage retains a single taxonomic identity requires

whole chains of more or less identical events (Gingerich 2001). But what causes these

identical developmental events generation after generation? What relative proportions

do ecology, biogeography, morphology and phylogenetic inertia contribute to longevity?

Patterns and statistical correlations do not imply causation; tests involving techniques

from fields ranging from paleontology and phylogenetics to molecular biology and

genetics need to be designed to investigate the mechanisms promoting longevity, if any.

� 53

CHAPTER III

DO DEVIANTS LIVE LONGER? MORPHOLOGY AND LONGEVITY IN

TRACHYLEBERIDID OSTRACODES1

Introduction

The prolonged persistence of taxa in the fossil record is interesting because persistence

is contrary to evolution, which implies pervasive change. The study of geologic

longevity of taxa has had several guises. Longevity has been explored through the

analysis of extinction probability, taxon selectivity across extinction events, extinction

risk and survivorship (Pearson 1992, Gilinsky 1994, Jablonski 1994, Jablonski & Raup

1995, McKinney 1997). Taxa with wider geographic ranges seem to have lower

extinction risks, at least during “background times” (Jablonski 1986b, Jablonski & Raup

1995), although counter-evidence also exists (Vermeij 1993). Taxa with less specialized

feeding strategies also appear to have longer geologic durations, at least for Paleozoic

crinoid species (Baumiller 1993). Morphological complexity has also been suggested

as a correlate of longevity (Flessa et al. 1975, Anstey 1978, Ward & Signor 1983,

Boyajian & Lutz 1992) although a definitive relationship between these variables is

lacking. Taxa with larger body sizes, and correspondingly longer generation times, turn

over more slowly or are geologically more persistent than related taxa that are smaller

(van Valen 1975, Flynn et al. 1995). Some studies, however, suggest that it is not the

� 54

1 This paper originally appeared in volume 32, pages 5-69, of Paleobiology.

organism that maintains the inertia of change. Instead, attributes of the environment

(stability, suitability) seemingly promote their geologic longevity (Alexander 1977,

Fortey 1980, Norris 1992).

In general, ecologically more specialized taxa are more prone to extinction because of

smaller geographic ranges, fewer potential habitats, narrower niche breadths and lower

abundances. These generalizations have been shown for Mezosoic – Cenozoic

Foraminifera genera (Banerjee & Boyajian 1996), species of carnivorous Miocene

mammalian (Viranta 2003) and Mississippian crinoid species (Kammer et al. 1997,

1998). Ecological specialization was inferred from morphology in the above studies,

with the implicit assumption that morphology is a proxy for ecology.

In this study, I compare ostracode genus longevities directly with their morphologies. I

do this in the spirit of an empirical multivariate morphospace approach (Foote 1997,

Roy & Foote 1997 and references therein), although here, distances from a mean are

utilized rather than measures of disparity. I predict that the longer the genera survive,

the more morphologically average or less specialized they should be when compared

with shorter-lived genera, in accordance with Simpson’s (1944) “survival of the

relatively unspecialized.” These comparisons are in the context of overall

morphological variation among constituent members of a particular clade, existing or

appearing during particular geologic time intervals (contemporaneous genera and birth � 55

cohorts). Morphologically average genera are potentially more general ecologically,

less prone to stochastic environmental perturbations and may have greater survivorship

than morphologically deviant (= more specialized) genera. This predication also

follows from a previous finding that long-lived fossil crinoid genera throughout the

Phanerozoic (either moderately or extremely long-lived) tend to be more average (less

specialized, less deviant) in morphology than expected when compared with shorter-

lived congeners in crinoid orders (Liow 2004). This finding contrasts with some

thinking that extremely long-lived taxa or “living fossils” are special or exceptional

(Parsons 1994, Eisner 2003).

Specifically, I examine a large family of marine podocopid ostracodes, the

Trachyleberididae sensu lato, to test if longer-lived genera are: i) morphologically

average (i.e. no different collectively from shorter-lived genera); and, ii)

morphologically more or less average than their shorter-lived relatives than expected. I

use two independent sets of morphological data (discrete morphology and outlines) to

examine the sensitivity of resulting patterns to data types. I also parse the data in

several ways to validate the results based on consistency and to account for some

possible sampling biases. I plot various measures of morphological deviations of

genera from their group mean versus their geologic durations to produce morphological

deviation-duration plots.

� 56

Longer-lived taxa often appear to plot rather close to the average morphology in

morphological deviation-duration plots, whereas shorter-lived taxa span a wider range

of morphologies (Liow 2004, this study). Because there are often many more shorter-

lived taxa than longer-lived ones, there is a higher probability that some will have

morphologies that deviate greatly from the group mean. Conversely, since there are few

long-lived taxa, there is a much lower probability for any of them to be very far from

the average morphology of the group (Fig. III.1). The question then becomes, whether

they are closer or farther from the mean morphology than expected by chance alone.

In this study, I show that collective morphological deviation of long-lived

trachyleberidid ostracode genera from the group mean is not significantly different from

that of shorter-lived genera. This finding remains unchanged, whether based on discrete

morphological characters or lateral valve outlines from representative specimens of

each genus. This result is also robust to removing genera that are possibly over-split

taxonomically, as well as those that are less well-sampled. The same lack of a

significant relationship between morphological deviation and longevity applies also to

birth cohorts (all genera first appearing in a named time interval). However, analyses of

contemporaneous genera in epochs (all genera existing in a epoch, regardless of when

they first or last appear) show that longer-lived taxa are sometimes collectively

marginally more deviant morphologically than shorter-lived ones. This last finding, in

� 57

contrast with a general pattern of non-significance, is discussed in light of the scale of

observation as well as potential biological implications.

� 58

Du

rati

on

Morphological

Deviation

Du

rati

on

Mo

rph

olo

gic

al

De

via

tio

n

Frequency

Frequency

+

=

Fig

. II

I.1. A

low

er p

robab

ilit

y f

or

long-l

ived

tax

a to

be

dis

tant

from

the

aver

age

morp

holo

gy o

f th

eir

gro

up.

The

top l

eft

figure

show

s a

his

togra

m o

f m

orp

holo

gic

al d

evia

tions

dra

wn f

rom

a s

imula

ted n

orm

al d

istr

ibuti

on. T

he

top

right

show

s a

his

togra

m o

f dura

tions

dra

wn f

rom

a s

imula

ted e

xponen

tial

dis

trib

uti

on. T

he

bott

om

plo

t ar

e ra

ndom

pai

rings

of

val

ues

fro

m t

he

two d

istr

ibuti

ons,

show

ing t

hat

the

few

long-l

ived

, si

mula

ted t

axa

are

less

spre

ad o

ut

along

the

y-a

xis

than

the

short

-liv

ed t

axa.

� 59

Data and Methods

The organisms and the raw data

Members of the Trachyleberididae sensu lato (Podocopida: Cytheracea) are found in

benthic sediments all over the world, from the shallowest brackish waters to the deepest

oceans. It is a large family (perhaps equivalent to a higher taxonomic level in other

marine invertebrates such as mollusks) that began definitively in the earliest late

Cretaceous, although trachyleberidid-like taxa have been found as early as the Jurassic

(e.g. Oligocythereis and Morkhovenicythereis, see Gruendel 1975 and Lord 1979).

Members of this family are still abundant today even though many of its earlier-

occurring genera are extinct. This family is heavily utilized in biostratigraphy not least

because of its abundance and its frequently ornate nature that makes taxonomic

recognition less problematic than with many marine invertebrate groups or other

ostracode taxa.

I constructed a relational database of species of trachyleberidid species that attempts to

eliminate taxonomic synonyms. The data include the species’ geographic and geologic

occurrences, as well as their membership in genera. Stratigraphic ranges of genera are

built from those of their component species. I converted the published time of first and

last appearances of species to numerical values using the International Stratigraphic

Chart (International Commission on Stratigraphy 2004). Durations of genera are

computed as the length of absolute time between the middle of the interval in which � 60

their first species appear and the middle of the interval in which their last species

disappear. The level of stratigraphic resolution for species was inevitably

heterogeneous. However, instead of discarding data of a lower resolution, I included

them in calculating durations for two reasons. First, since this study involves a

comparison of durations, only the relative ranking of durations are really vital and a

(morphologically) random distribution of species with better or poorer resolved time

intervals should not bias results. Second, to discard species with less well-resolved time

intervals would greatly shorten known genus durations in numerous instances. Genera

that are reported to occur only in one time interval are reported as having durations of

zero, even though that is an impossibility. However, as before, only the approximate

relative positioning of genera according to duration is important here because binary

bins of genera with long or shorter durations are used in the main analyses.

My analyses are based on 326 genera, after excluding synonyms and doubtful genera.

The family was erected in 1948 (Sylvester-Bradley 1948) and many of its 300+ genera

have been variously assigned to Trachyleberididae sensu stricto or one of its closely

allied families (sometimes also reported as subfamilies of Trachyleberididae), e.g.

Hemicytheridae, Buntoniidae or Brachycytherinidae. While specific assignments to

family, subfamily or tribe have fluctuated historically, general agreement on what a

trachyleberidid is, sensu lato, can be assumed with relative confidence. The

relationships of lower taxa in family Trachyleberididae sensu lato cannot be clearly � 61

delineated with our current knowledge, although the recognition of species within the

family is not problematic by most standards.

There is no published trachyleberidid taxonomic list, although several major references,

not least Hazel 1967 and van Morkhoven 1963, provided a baseline compilation of the

species and genera of this family. To assess the completeness of my literature survey, I

constructed sampling effort curves for species and genera. The sampling curve for

genera started to flatten after about 85 days of data collection (Fig. III.2). There are

currently more than 4000 species in the database. Addition of species new to my

database has not changed the stratigraphic ranges of genera since the time my genus-

collection curve began flattening, indicating that my sampling has sufficiently traced the

existing literature.

The rationale for focusing on the generic level, even though species stratigraphic range

data are available for this study, is two-fold. First, species stratigraphic durations are

less stable than genus durations. Addition of new occurrences may often change the

known geologic range of a species, unlike the case of the genus mentioned above.

Second, detailed morphology is not as completely known for species such that species

level analyses will inevitably involve many more unknown character states, not to

mention that the number of species that have been described is prohibitively large.

� 62

/Users/LH

0 20 40 60 80 100 120

05

0

/Users/LH

10

01

50

20

02

50

30

03

50

Fig. III.2 Sampling Curve

Days of sampling

No

.of

ge

ne

ra n

ew

to

da

tab

ase

Fig.III.2. Literature sampling curve for trachyleberidid genera. The number of genera stands at 340 instead of the 326 used in the analyses because 14 of them are doubtful and/or too poorly known.

� 63

The taxonomic/stratigraphic information is dynamically linked to morphological table

consisting of discrete morphological characters, 87 of which are used in this paper

(Appendices E and F). These characters (Fig. III.3) are commonly used to delineate

genera and are also relatively easily observed from actual specimens or Scanning

Electron Micrographs (SEMs). Character states are coded from primary descriptions

and illustrations of genera and representative species as well as from published SEMs

and supplemented by my examination of museum types. The first set of morphological

data includes external features and ornamentation on the valves, characters from hinges,

internal muscle scars and pores. There is a mixture of numerical, binary, ordered and

unordered multi-state characters (Appendices E and F). Data are obtained from type

species, unless those are unavailable, and corroborated by other species. If the type

species has a character state that is rare among its congenerics, the more common state

is coded. For characters that are variable within a species, the most commonly

occurring state is coded. This situation is rare because most of the characters are “good

genus level characters.” A second independent set of data is traced outlines from the

left valves of representative adult specimens of genera, again using published SEMs or,

in rare cases, drawings (Appendix G). Outlines of trachyeleberidids have been used

successfully in distinguishing different genera (Bachnou et al. 1999, 2000, using

Fourier analysis). The outlines are traced using tpsDIG (Rohlf 1992). In each case, 200

evenly spaced coordinates are recorded, beginning with the position of the eye tubercle,

or in cases where eye tubercles are absent, the point of greatest height, which is an� 64

Fig. III.3. A generalized trachyleberidid ostracode. The top sketch is an external left valve and the bottom an internal right valve (showing hinge, pores and muscle scars[ms]). The “x” at the apex of the external valve marks the start point of coordinate pairs collected for outline analyses.

� 65

equivalent position (Fig. III.3). Laterally projecting ornamentations are ignored in

tracing the outlines.

Data treatment and analysis: discrete character data

To study the morphological deviation of a genus from the mean of the family (= the

degree of specialization or averageness), I summed character distances of that genus

from the group mean value of each character calculated from all the genera involved.

For binary characters, this group average value is simply the mean of the character

states of all the genera, excluding those that were coded as “unknown” or

“inapplicable,” which translates as the probability of occurrence of that character state.

Similarly, for ordered multi-state characters, the numerical mean of the character states

are calculated. For unordered multi-state characters, however, the average value is taken

to be the modal state of the character and any other character state is taken as being one

unit removed, regardless of the numerical coding of the character state. Lastly, for

meristic characters, such as the number of denticles, natural logarithms are applied

before means and distances are calculated. This transformation moderates the effects of

counted characters in genera otherwise not very different from each other (e.g. denticle-

poor versus denticle rich). The different range of values assigned to these four character

types give slightly different weights to characters of each type. However, since the

ranges are not overwhelming different, and because the importance of each character

� 66

and their independence from each other is not currently known, nothing further is done

to modify the degree of contribution of various characters to the overall morphospace.

As an alternative method to studying distance from a mean morphology to that

described above, I also calculated Principal Coordinate Scores (PCO, Gower 1966).

This is simply a Principal Components Analysis (PCA) performed on the genus-to-

genus morphological-distance matrix. I then calculated departures of respective PCA

scores of each genus from the PCA scores averaged from all the genera included in the

analysis and compared the sum of those departures with their respective genus

durations.

Removal of oversplit taxa

It is possible that my database contains a number of over-split genera whose

morphologies are very similar, at least based on the characters used. Hence if they are

not “real” genera, there could be excessive contribution of these kinds of morphologies

in calculating the family morphological mean. I removed 49 genera (Appendix H) that

are potentially over-split and re-ran the analyses as above. These genera were either

first erected as subgenera or are parts of genus-complexes. The representative genera

retained are the better known of the pair or group of closely related genera.

� 67

Data treatment and analysis: outline data

I analyzed the outline coordinates in two ways in order to test the robustness of results.

First, I performed PCA using the harmonics from Fourier analysis (Ferson et al. 1985).

I compared the resulting PCA scores of each genus with the mean PCA scores of the

family calculated from the genus PCA scores. The Fourier analysis was done using

Elliptical Fourier Analysis (EFA) as written by Rohlf (1992). The first ten harmonics

regenerated outlines precisely; thus they were used in PCA analyses and subsequent

harmonics ignored. The second method used Standard Eigenshape Analysis (MacLeod

1999), a completely different approach to studying outlines. This was chosen over the

more powerful Extended Eigenshape Analysis, which takes into account the location of

homologous points around an outline. The reason is because multiple precise

homologous points cannot be identified reliably on the external carapace on such a wide

range of taxa. The output data are eigenshape scores, which are equivalent to PCA

scores. The eigenshape scores for each genus are compared to the family mean as

described earlier in this paragraph for Fourier analysis.

Defining long-lived genera

There are many ways of identifying long-lived taxa in any given group (Liow 2004).

Here, I define long-lived genera in three ways: 1) the most long-lived 5% of genera, 2)

the most long-lived 10% of genera and 3) genera having a duration greater than the

mid-range duration of the sample of genera included in a particular analysis. I chose to � 68

use “long-lived” and “shorter-lived” to reflect genera with long durations and those that

have comparatively shorter durations, respectively, because the durations of some of the

shorter-lived genera may not be “short” by other definitions. There are usually far

fewer taxa with extended durations than those with shorter durations. Therefore, a

comparison of the deviation of morphology from a group mean of long-lived versus

shorter-lived taxa, requires a method to deal with the huge differences in sample sizes.

In order to do this, I compared morphological deviations of rarified samples of shorter-

lived trachyleberidid genera with long-lived ones. The number of genera picked from

the shorter-lived pool depends on the number of long-lived genera identified. This

rarefaction is repeated 10,000 times for each sub-sampled data set (see results Tables).

On a few occasions, there are more long-lived genera by definition and when this

happens, the long-lived pool is rarified instead. The proportion of times that long-lived

taxa are more deviant from a mean morphology is reported as a “p-value” (for details

see Liow 2004). A high rarefaction “p-value” means that long-lived taxa are

significantly more deviant and a low one means that they are significantly more average

when compared with shorter-lived taxa. This is a two-tailed test, hence a significant

probability value will be either 0.025 (significantly less deviant) or 0.975 (significantly

more deviant). Since dividing the datasets into two categories reduces statistical power,

I also report probabilities and correlation values from Kendall’s rank correlation by

treating the data as continuous. I report uncorrected probability values, but where

� 69

significant results are found, I apply Bonferroni correction to account for the non-

independence of the multiple analyses.

Results

The mean genus duration of trachyleberidids is between 26 and 33 M.y. and the median

between 21 and 29 M.y. (Table III.1) depending on whether single-stage, extant or both

types of genera are excluded from the estimate. The average of species duration for

trachyleberidids is about 4 M.y. The longest-lived genus is Cythereis (140.5 M.y.)

followed by Cytheretta (122.0 M.y.) and Pterygocythereis (101.2 M.y.) (Appendix H).

Perhaps these are “under-split” or “garbage can” taxa, but the characters used to

delineate these taxa seem to be consistent. Even if these are not “real” genera by some

other definitions, they correspond to consistent aggregates of characters.

Morphological deviation of genera from group means (discrete characters)

Long-lived genera are not significantly more or less deviant from the group mean than

shorter-lived genera, when compared with rarefied samples of shorter-lived genera,

using p = 0.025 as a cut-off in either direction. For instance, comparing all

morphological characters and all genera, there are 17 long-lived genera (if long-lived is

taken as the most long ranging 5% of all the genera). Comparing these 17 genera with

10, 000 random samples of 17 shorter-lived taxa (i.e. all the other 309 genera) gives a

value of 0.52 (Table III.2 , first row). Stated a different way, these 17 long-lived taxa � 70

TABLE III.1. Summary statistics for the durations of subsets of trachyleberidids.

Table listing durations (M.y.) for various subsets of the data. N = no. genera; No SS = excluding single

genera; No Ext = excluding extant genera; -OS = minus 49 over-split genera; FA = birth cohorts with

first appearances in the number (Ma) following “FA” till just before the value of the “FA” in the next

column in the table; Pre Pale = genera occurring earlier than the Paleocene; Pale = genera occurring

during the Paleocene; Eo = genera occurring during the Eocene; Ol = genera occurring during the

Oligocene; Mi = genera occurring during the Miocene; Post Mi = genera occurring after the Miocene.

All No SS No Ext

No SS

No Ext All (-OS)

No SS

(-OS)

No Ext

(-OS)

No SS No

Ext (-OS)

N 326 271 161 136 277 225 140 117

Mean 27.4 32.6 26.1 30.9 26.6 32.7 24.9 29.8

Median 20.7 25.2 21.1 28.6 18.8 24.8 19.7 25.1

Maximum 140.5 140.5 92.8 92.8 140.5 140.5 92.8 92.8

Minimum 0.0 0.5 0.0 0.8 0.0 0.0 0.0 0.8

FA 166 FA 116 FA 105 FA 95 FA 77 FA 65 FA54 FA 42 FA25

N 11 14 25 35 39 33 29 14 29

Mean 46.9 42.3 54.0 43.9 36.1 42.8 29.5 24.9 17.5

Median 43.1 40.8 60.9 39.6 31.0 58.9 34.7 28.8 20.7

Maximum 140.5 122.0 101.2 92.8 77.4 63.5 53.7 41.9 25.5

Minimum 0.0 0.0 3.7 0.0 2.1 0.0 0.0 0.0 0.0

FA15 FA5 Pre Pale Pale Eo Ol Mi Post Mi

N 31 65 125 108 134 80 123 179

Mean 9.1 1.3 43.6 54.4 51.1 61.7 42.6 29.3

Median 8.4 0.5 39.5 59.9 51.7 62.5 37.8 16.3

Maximum 14.6 4.5 140.5 140.5 140.5 140.5 140.5 140.5

Minimum 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

� 71

Table

III.2

. D

evia

tion fro

m tra

chyle

beridid

gro

up m

ean (

dis

cre

te m

orp

holo

gy)

Table

show

ing p

robabili

ty v

alu

es fro

m r

are

faction a

naly

sis

and K

endall

rank c

orr

ela

tion tests

usin

g d

iscre

te c

hara

cte

rs to

calc

ula

te m

orp

holo

gic

al devia

tion. N

= n

um

ber

of chara

cte

rs; N

A =

non a

pplic

able

chara

cte

rs; Q

=chara

cte

r sta

te u

nknow

n;

N-G

enera

: sam

ple

siz

e; F

ive =

pro

port

ion o

f tim

es w

hen the 5

% m

ost lo

ng-liv

ed g

enera

in the g

roup a

re m

orp

holo

gic

ally

more

devia

nt fr

om

the g

roup m

ean than r

are

fied s

hort

-liv

ed g

enera

; N

–F

ive =

num

ber

of 5%

most lo

ng-liv

ed g

enera

; T

en =

pro

port

ion o

f tim

es w

hen the 1

0 %

m

ost lo

ng-liv

ed g

enera

in the g

roup c

onsid

ere

d a

re m

orp

holo

gic

ally

more

devia

nt fr

om

the g

roup m

ean than r

are

fied s

hort

-liv

ed g

enera

; N

–T

en =

num

ber

of 10 %

most lo

ng-liv

ed g

enera

; M

id =

pro

port

ion o

f tim

es

when those h

avin

g a

dura

tion g

reate

r th

an that of th

e m

id-r

ange d

ura

tion o

f th

e g

roup a

re m

orp

holo

gic

ally

more

devia

nt fr

om

the g

roup m

ean than r

are

fied s

hort

-liv

ed g

enera

; N

–M

id =

num

ber

of genera

that have a

dura

tion g

reate

r th

an that of th

e

mid

-range d

ura

tion o

f th

e g

roup; K

’s p

= p

robabili

ty fro

m K

endall’

s r

ank c

orr

ela

tion test; tau =

Kendall’

s c

orr

ela

tion c

oeffic

ient.

Sig

nific

ant pro

babili

ty v

alu

es a

re in b

old

-face a

nd those r

em

ain

ing s

ignific

ant after

Bonfe

ronni corr

ection a

re u

nderlin

ed.

Taxa

Exclu

ded

Chara

cte

rs

N

mean%

NA

mean%

Q

mean

%N

A+

Q

N-

Genera

Fiv

eN

-Fiv

eTen

N-T

en

Mid

N-M

idK's

pta

u

None

All

87

11

.58

14

.85

26

.43

32

60

.52

17

0.3

73

30

.42

25

0.8

80

.00

Exte

rnal

49

19

.72

6.4

82

6.2

03

26

0.9

21

70

.84

33

0.9

12

50

.04

0.0

8

Inte

rnal

38

1.0

32

5.7

52

6.7

83

26

0.2

41

70

.56

33

0.3

12

50

.14

-0.0

5

Hin

ge

90

.00

10

.26

10

.26

31

80

.81

16

0.5

23

20

.75

25

0.9

50

.00

Musc

le s

car

13

2.9

73

0.6

53

3.6

22

69

0.0

61

40

.10

27

0.1

12

30

.15

-0.0

6

Pore

90

.03

40

.49

40

.52

26

90

.60

14

0.4

62

70

.40

23

0.9

80

.00

Sin

gle

Sta

ge

All

87

11

.46

14

.70

26

.16

27

10

.86

14

0.3

32

80

.35

25

0.5

7-0

.02

Exte

rnal

49

19

.48

6.3

32

5.8

22

71

1.0

01

40

.92

28

0.9

02

50

.04

0.0

9

Inte

rnal

38

1.1

12

5.3

32

6.4

42

71

0.1

91

40

.25

28

0.2

72

50

.01

-0.1

2

Hin

ge

90

.00

10

.29

10

.29

26

10

.21

14

0.7

32

70

.72

25

0.7

2-0

.02

Musc

le s

car

13

3.2

13

0.0

33

3.2

42

22

0.1

01

20

.07

23

0.0

82

30

.00

-0.1

4

Pore

90

.00

40

.03

40

.03

22

50

.78

12

0.3

62

30

.39

23

0.5

8-0

.02

� 72

Table

III.2

. (c

on't)

: D

evia

tion fro

m tra

chyle

beridid

gro

up m

ean (

dis

cre

te m

orp

holo

gy)

Taxa

Exclu

ded

Chara

cte

rs

N

mean%

NA

mean%

Q

mean

%N

A+

Q

N-

Genera

Fiv

eN

-Fiv

eTen

N-T

en

Mid

N-M

idK's

pta

u

Exta

nt

All

87

11

.09

18

.62

29

.71

16

10

.30

90

.13

17

0.5

13

60

.21

-0.0

7

Exte

rnal

49

19

.15

7.0

42

6.1

91

61

0.3

59

0.3

01

70

.81

36

0.6

3-0

.03

Inte

rnal

38

0.6

93

3.5

74

6.3

21

61

0.6

89

0.5

21

70

.21

36

0.4

9-0

.04

Hin

ge

90

.00

13

.24

13

.24

15

20

.50

80

.40

16

0.7

03

30

.19

0.0

7

Musc

le s

car

13

2.0

04

0.1

14

2.1

11

17

0.2

96

0.3

81

20

.08

31

0.1

3-0

.09

Pore

90

.00

51

.23

51

.23

15

20

.49

80

.38

16

0.7

03

30

.15

-0.0

9

Sin

gle

Sta

ge &

A

ll8

71

0.8

51

8.8

62

9.7

01

36

0.2

47

0.2

31

40

.70

35

0.5

0-0

.04

Exta

nt

Exte

rnal

49

18

.70

7.2

72

5.9

71

36

0.3

07

0.5

81

40

.92

35

0.6

30

.03

Inte

rnal

38

0.7

23

3.8

03

4.5

21

36

0.6

07

0.2

81

40

.16

35

0.1

9-0

.08

Hin

ge

90

.00

13

.70

13

.70

12

90

.31

70

.07

13

0.6

83

20

.33

0.0

6

Musc

le s

car

13

2.1

13

9.8

34

1.9

49

90

.56

50

.79

10

0.7

13

00

.99

0.0

0

Pore

90

.00

52

.13

52

.13

98

0.2

75

0.1

31

00

.12

24

0.3

1-0

.07

> 2

6 N

AA

ll8

7-

-1

7.2

52

19

0.9

21

10

.88

22

0.8

22

10

.64

0.0

2

> 1

8 N

AA

ll8

7-

-1

3.0

41

07

0.8

56

0.8

11

10

.82

12

0.4

70

.05

� 73

are more deviant in morphology than any rarefied sample of shorter-lived taxa about

53% of the time. The same conclusion can also be drawn from Kendall’s rank test,

where there is no correlation between morphological deviation and duration (Table III.2,

p = 0.88). It should be noted that there are many ties in the data when subjected to

Kendall’s rank test, rendering the p-values calculated, inexact. The statement of non-

significant differences is true for other definitions of long-lived (the most long-lived 5%

or 10% of genera, and genera having a duration greater than the mid-range duration of

the group).

Next, to account for incomplete and questionable duration sampling, I eliminated

genera that occur only in one stage. In doing so, the most long-lived 5% or 10% of the

taxa or those having durations greater than the mid-range value, are all not more or less

deviant than rarefied samples of shorter-lived taxa (Table III.2). I also eliminated

genera that are extant to account for one-sided range truncations. This removed taxa

that have long durations and range to the Recent and possibly introduced a different

bias. The relationship between morphological deviation and longevity is again non-

significant when extant genera were removed, as when both single-staged and extant

genera were removed from analysis (Table III.2). Kendall’s rank correlation tests also

showed a non-significant relationship between deviation and duration for these

comparisons (Table III.2).

� 74

I also explored the effects of different subsets of discrete morphological characters on

the analysis. Parsing the morphological characters into external, internal, muscle scars,

hinge and pores also maintained a pattern of non-significance with one exception. For

external characters when single-stage genera were removed, the most long-lived 5% of

the genera are significantly more deviant, even after Bonferroni correction. There is a

possibility that unknown or uncodable characters may be contributing to the general

result, but a check of the proportion of un-coded characters in each subset does not

show systematic bias in any direction (Table III.2). Similarly, although the probability

values from Kendall’s ranks were less than 0.05 in few cases and one of them, is

significant after Bonferroni correction (muscle scars, single-stage genera removed, p =

0.0001), they occur inconsistently compared with other analyses within Table III.2.

I removed 49 genera (Appendix H) that are potentially over-split and re-ran the analyses

as above. The pattern between morphological deviation and longevity remained mostly

insignificant for various divisions and exclusions of data. The few significant and

marginally significant instances are due to smaller sample sizes and a greater number of

uncoded characters (Table III.3) as shown by correlation tests (e.g. internal characters,

single stage genera removed, significant negative correlation between deviation and no.

unknown characters, p = 0.002, tau = -0.13, other results not shown). However, an

� 75

Ta

ble

III

.3.

De

via

tio

n f

rom

tra

chyle

berid

id g

roup

me

an

(d

iscre

te m

orp

ho

logy)

with

ove

rsplit

ge

ne

ra r

em

ove

d.

Th

is s

ho

ws p

rob

ab

ility

va

lue

s f

rom

ra

refa

ctio

n a

na

lysis

an

d K

en

da

ll ra

nk c

orr

ela

tio

n tests

with

49

ove

rsp

lit g

en

era

re

mo

ve

d.

Ab

bre

via

tion

s a

s in

Ta

ble

III

.2.

Taxa

Exclu

ded

Chara

cte

rs

NN

-Genera

Fiv

eN

-Fiv

eTen

N-T

en

Mid

N-M

idK's

pta

u

None

All

87

27

70

.88

14

0.7

02

80

.70

20

0.6

80

.02

Exte

rnal

49

27

71

.00

14

0.9

52

80

.98

20

0.0

20

.10

Inte

rnal

39

27

70

.19

14

0.4

42

80

.20

20

0.1

4-0

.06

Hin

ge

92

69

0.2

61

40

.17

27

0.1

52

00

.34

-0.0

4

Muscle

scar

13

23

20

.21

12

0.4

02

40

.21

18

0.1

8-0

.06

Pore

92

36

0.7

71

20

.77

24

0.5

11

90

.62

0.0

2

Sin

gle

Sta

ge

All

87

22

50

.76

12

0.5

72

30

.57

20

0.4

8-0

.03

Exte

rnal

49

22

50

.98

12

0.9

72

30

.97

20

0.0

20

.10

Inte

rnal

39

22

50

.23

12

0.1

52

30

.12

20

0.0

0-0

.15

Hin

ge

92

17

0.3

11

10

.17

22

0.1

42

00

.36

-0.0

4

Muscle

scar

13

18

70

.24

10

0.1

21

90

.09

18

0.0

0-0

.17

Pore

91

93

0.2

21

00

.42

20

0.4

61

90

.68

-0.0

2

Exta

nt

All

87

14

00

.39

80

.19

15

0.4

72

80

.12

-0.0

9

Exte

rnal

49

14

00

.46

80

.36

15

0.8

62

80

.55

-0.0

3

Inte

rnal

39

14

00

.71

80

.13

15

0.0

22

80

.30

-0.0

6

Hin

ge

91

32

0.0

07

0.0

11

40

.09

25

0.2

5-0

.07

Muscle

scar

13

14

00

.58

80

.80

15

0.6

32

80

.90

-0.0

1

Pore

91

05

0.2

26

0.2

31

10

.27

19

0.2

9-0

.07

Sin

gle

Sta

ge &A

ll8

71

17

0.1

16

0.1

41

20

.57

27

0.1

9-0

.08

Exta

nt

Exte

rnal

49

11

70

.12

60

.27

12

0.8

92

70

.91

-0.0

1

Inte

rnal

39

11

70

.57

60

.36

12

0.0

42

70

.12

-0.1

0

Hin

ge

91

12

0.9

76

0.5

21

20

.71

25

0.3

9-0

.06

Muscle

scar

13

87

0.2

25

0.3

69

0.0

12

40

.01

-0.1

9

Pore

98

60

.15

50

.23

90

.29

18

0.7

90

.02

� 76

exception is one involving the external morphology of all genera and with single-stage

genera removed, showing that long-lived genera may be significantly (even after

Bonferroni correction in the 5% case) more deviant than shorter-lived genera.

However, Kendall’s rank test shows a marginally significant positive relationship

between morphological deviation and morphology, contrary to the rarefaction tests.

I checked whether removing genera with many un-coded characters changed the

patterns of non-significance. Results are statistically non-significant (Table III.2, last

two rows).

Morphological deviation of genera from group means (discrete characters): temporal

subsets

This family is probably a monophyletic or nearly monophyletic collection of genera.

However the database is global and heterogeneous in both temporal and geographical

coverage. Therefore I divided the data into temporal subsets of genera, to check if

different morphological deviation-duration patterns emerge. This is important because

genera from a globally distributed dataset may not interact ecologically or

phylogenetically with each other directly enough for any patterns to be discerned.

Subdividing the data serves to homogenize the data so that patterns, even weak ones,

may have a chance of being detected.

� 77

First, I compared the morphological deviations of long- and shorter-lived genera from

their birth cohort means (a birth cohort is the subset of genera appearing within a named

time interval). I divided the data into birth cohorts of 10 to 18 million years except for

the Late Cretaceous and earlier (lumped as a birth cohort lasting 50 million years

because the interval has only 11 genera, too few for finer subdivision). Sample sizes for

each time slice are in general small and the significant values that emerge in a few

instances show no consistent pattern (Table III.4). Similarly, the only significant p-

values for Kendall’s rank test value is shown for genera less than 5 Ma (using all and

internal characters), but this may not have much weight since many of these genera will

certainly continue into the future. I have included them only for completeness. On the

whole, long-lived genera within birth cohorts are not more or less deviant from the

cohort mean than their shorter-lived relatives.

A different landscape emerges when the data are divided into contemporaneous genera

in epochs. I find that the most long-lived 5% or 10% of genera in each epoch are more

deviant from the mean of that epoch than is expected, at least marginally (Table III.5) in

terms of overall morphology and external morphology, although not internal

morphology. The long-lived genera of each epoch do over-lap (e.g. Cythereis is present

in every single epoch analyze) but they do not belong to any one subfamily or tribe.

Not all the values are significant at p = 0.025/0.975 or at p = 1.4e-3/9.99e-3 after

Bonferroni correction, but all except Eocene values, are consistently high. However � 78

Ta

ble

III

.4.

De

via

tio

n f

rom

tra

ch

yle

be

rid

id b

irth

co

ho

rt m

ea

n (

dis

cre

te m

orp

ho

log

y)

Ta

ble

sh

ow

ing

pro

ba

bili

ty v

alu

es f

rom

ra

refa

ctio

n a

na

lysis

an

d K

en

da

ll ra

nk c

orr

ela

tio

n t

ests

usin

g b

irth

co

ho

rts.

Ab

bre

via

tio

ns a

s in

Ta

ble

III

.2.

FA

(Ma

) re

fers

to

first

ap

pe

ara

nce

, th

e b

eg

inin

g o

f th

e in

terv

al co

nsid

ere

d.

Th

e

en

d o

f o

ne

in

terv

al is

th

e b

eg

inn

ing

of

the

ne

xt

(= t

he

ne

xt

FA

va

lue

). (

R)

ind

ica

te t

ha

t lo

ng

-liv

ed

ta

xa

we

re r

arifie

d

be

ca

use

th

ere

we

re m

ore

of

the

m t

ha

n s

ho

rt-liv

ed

on

es.

FA

Chara

cte

rs

NFiv

eN

-Fiv

eTen

N-T

en

Mid

N-M

idK's

pta

u

166

All

11

0.9

11

0.0

02

0.0

02

0.6

3-0

.11

Exte

rnal

0.5

41

0.0

02

0.0

02

0.4

5-0

.17

Inte

rnal

0.9

11

0.3

82

0.3

82

0.1

9-0

.29

116

All

14

0.9

21

0.5

72

0.4

33

0.3

2-0

.20

Exte

rnal

0.9

21

0.4

52

0.4

73

0.8

30.0

4

Inte

rnal

0.9

21

0.8

52

0.8

63

0.6

60.0

9

105

All

25

0.4

32

0.5

93

0.1

013 (

R )

0.5

40.0

9

Exte

rnal

0.6

52

0.8

33

0.0

113 (

R )

0.6

50.2

6

Inte

rnal

0.3

92

0.2

43

0.9

013 (

R )

0.0

8-0

.25

95

All

35

0.6

22

0.4

14

0.1

916

0.7

70.0

3

Exte

rnal

0.4

92

0.4

74

0.9

816

0.5

80.0

7

Inte

rnal

0.8

72

0.7

84

0.0

016

0.3

7-0

.11

77

All

40

0.2

73

0.4

25

0.4

618

0.2

8-0

.12

Exte

rnal

0.3

03

0.4

35

0.0

218

0.1

9-0

.15

Inte

rnal

0.5

13

0.7

05

1.0

018

0.6

50.0

5

� 79

Table

III.

4. (c

on't

): D

evia

tion f

rom

tra

chyle

beridid

birth

cohort

mean (

dis

cre

te m

orp

holo

gy)

FA

Chara

cte

rs

NFiv

eN

-Fiv

eTen

N-T

en

Mid

N-M

idK's

pta

u

65

All

33

0.3

32

0.4

14

0.9

523 (

R )

0.4

00.1

0

Exte

rnal

0.1

42

0.2

44

0.9

523 (

R )

0.3

80.1

1

Inte

rnal

0.7

92

0.9

14

0.8

823 (

R )

0.6

60.0

5

54

All

28

0.9

12

0.9

73

1.0

017 (

R )

0.0

90.2

2

Exte

rnal

0.7

62

0.8

03

0.9

617 (

R )

0.2

90.1

4

Inte

rnal

0.0

32

0.1

13

0.9

517 (

R )

0.1

70.1

8

42

All

14

0.3

31

0.4

12

0.9

29 (

R )

0.7

7-0

.06

Exte

rnal

0.2

11

0.0

22

0.1

19 (

R )

0.1

8-0

.27

Inte

rnal

0.7

42

0.9

12

1.0

09 (

R )

0.0

60.3

8

25

All

29

0.0

42

0.0

23

0.8

323 (

R )

0.9

70.0

1

Exte

rnal

0.4

82

0.2

33

0.8

423 (

R )

0.4

90.0

9

Inte

rnal

0.0

42

0.0

73

0.9

123 (

R )

0.8

60.0

2

15

All

31

0.6

32

0.8

54

0.5

420 (

R )

0.0

5-0

.10

Exte

rnal

0.6

32

0.0

94

0.3

520 (

R )

0.5

8-0

.07

Inte

rnal

0.1

72

0.4

04

0.5

620 (

R )

0.8

60.0

2

5A

ll65

0.6

94

0.6

27

1.0

021

0.0

30.1

9

Exte

rnal

0.2

04

0.1

17

0.8

821

0.7

00.0

3

Inte

rnal

0.9

54

0.9

87

1.0

021

0.0

10.2

4

� 80

Ta

ble

III

.5.

De

via

tio

n f

rom

tra

ch

yle

be

rid

id c

on

tem

po

ran

eo

us c

oh

ort

me

an

(d

iscre

te m

orp

ho

log

y)

Ta

ble

sh

ow

ing

pro

ba

bili

ty v

alu

es f

rom

ra

refa

ctio

n a

na

lysis

an

d K

en

da

ll ra

nk c

orr

ela

tio

n t

ests

usin

g

co

nte

mp

ora

ne

ou

s s

ub

se

ts o

f g

en

era

. A

bb

revia

tio

ns a

s in

Ta

ble

III

.2.

Tim

e

Chara

cte

rs

NFiv

eN

-Fiv

eTen

N-T

en

Mid

N-M

idK's

pta

u

Cre

taceous

All

124

0.7

17

0.7

213

0.1

625

0.1

6-0

.08

Exte

rnal

0.9

97

0.9

913

0.9

225

0.2

80.0

6

Inte

rnal

0.3

37

0.3

413

0.3

925

0.1

5-0

.09

Pale

ocene

All

108

0.9

06

0.9

611

0.6

025

0.1

90.0

8

Exte

rnal

0.9

96

0.9

911

0.9

025

0.0

40.1

3

Inte

rnal

0.5

76

0.6

111

0.6

825

0.9

50.0

0

Eocene

All

134

0.8

87

0.9

114

0.4

325

0.7

20.0

2

Exte

rnal

0.8

17

0.6

114

0.8

125

0.4

70.0

4

Inte

rnal

0.4

67

0.4

114

0.6

825

0.8

40.0

1

Olig

ocene

All

79

0.7

34

0.8

98

0.6

222

0.4

60.0

6

Exte

rnal

0.9

14

1.0

08

0.9

622

0.1

30.1

2

Inte

rnal

0.6

34

0.2

68

0.3

222

0.4

6-0

.06

Mio

cene

All

123

0.9

37

0.9

913

0.8

223

0.1

60.0

9

Exte

rnal

0.9

97

1.0

013

0.9

923

0.0

30.1

3

Inte

rnal

0.4

97

0.5

513

0.3

223

0.4

9-0

.04

Post

Mio

cene A

ll178

0.9

79

0.8

618

0.8

718

0.1

20.0

8

Exte

rnal

1.0

09

1.0

018

0.9

918

0.0

10.1

3

Inte

rnal

0.1

49

0.1

518

0.1

818

0.2

4-0

.06

� 81

when more genera are included in the long-lived pool (using the definition of long-lived

as having a duration greater than the mid-range duration value of the group), the

deviation of long-lived genera from epoch means are no longer significant in numerous

cases (Table III.5). Kendall’s rank correlation test also does not show any consistent

statistical significance in the relationship between morphological deviation and

duration.

Principal Coordinate Analysis of discrete morphological data

It may be that some characters complexes, whose components are coded as separate

characters, are contributing more to the overall morphological representation. In order

to account for this possibility, I performed PCAs on the distance matrices resulting from

comparing character states of genera. The first 20 components yield between 88 and

92% of the total variance in each analysis done. The genus PCA scores of those 20

components were used in subsequent calculations of deviations of genera from a group

average. Comparing the resulting scores (Principal Coordinate Scores or PCOs) of

long- and shorter-lived genera from average scores of the entire group substantiated the

previous conclusions, with one exception. Contemporaneous subsets no longer seem to

have long-lived genera that are significantly more deviant from group means than

shorter-lived taxa, judging from the p-values of the rarefaction test (Table III.6). The

single significant value from Kendall’s rank test is for the contemporaneous group of

Post-Miocene genera, which includes many genera with one-sided range truncation.� 82

Ta

ble

III

.6.

De

via

tio

n f

rom

tra

ch

yle

be

rid

id m

ea

ns (

PC

O o

f d

iscre

te m

orp

ho

log

y).

Ta

ble

sh

ow

ing

pro

ba

bili

ty v

alu

es f

rom

ra

refa

ctio

n a

na

lysis

an

d K

en

da

ll ra

nk c

orr

ela

tio

n t

ests

usin

g p

rin

cip

al co

ord

ina

te (

PC

O)

sco

res.

Ab

bre

via

tio

ns a

s in

Ta

ble

III

.2.

Genera

Fiv

eN

-Fiv

eTen

N-T

en

Mid

N-M

idK's

pta

u

All

0.6

117

0.4

733

0.3

725

0.0

90.0

6

All

(-O

S)

0.6

614

0.5

428

0.4

620

0.2

90.0

4

> 2

6 N

as

0.7

511

0.8

122

0.8

021

0.3

00.0

4

Cre

taceous

0.6

47

0.5

013

0.1

225

0.3

2-0

.06

Pale

ocene

0.6

56

0.6

911

0.2

325

0.7

60.0

2

Eocene

0.6

17

0.6

914

0.2

925

0.4

2-0

.05

Olig

ocene

0.2

74

0.7

98

0.3

822

0.2

70.0

8

Mio

cene

0.7

77

0.8

113

0.6

523

0.1

80.0

8

Post

-Mio

cene

0.9

59

0.9

018

0.9

018

0.0

00.1

4

� 83

Morphological deviation of genera from group means (outline analyses)

I performed Elliptical Fourier Analyses on 284 outlines representing 284 genera,

creating an output of 10 harmonics. These 10 harmonics reproduced well the outlines

of selected specimens tested. Comparing the deviation of long- and shorter-lived genera

from means of the harmonics of all 284 genera, I find no significant difference between

the two groups of taxa (Table III.7). Principal Components Analysis of the 10

harmonics yielded results with the first four principal components accounting for 95%

of the total variation. Calculating deviation of these four principal components of long-

and shorter-lived genera from means for all 284 genera yielded similar non-significant

results (Table III.7).

Using a completely different approach to comparing outlines, I found the same non-

significance when comparing long and shorter-lived taxa. Standard Eigenshape

Analaysis on the 284 outlines yielded the results with the first ten eigenshape scores

accounting for about 90% of the variance in outline. Combining the eigenshape scores

in various ways did not change the conclusion that the outlines of longer-lived genera

are no more deviant from am average outline than shorter-lived taxa, by all the

definitions used (Table III.7). Kendall’s rank correlation tests show the same lack of

significance between morphological deviation and longevity (Table III.7).

� 84

Ta

ble

III

.7.

Re

su

lts f

rom

tra

ch

yle

be

rid

id o

utlin

e a

na

lyse

s.

Ta

ble

sh

ow

ing

pro

ba

bili

ty v

alu

es f

rom

ra

refa

ctio

n a

na

lysis

an

d K

en

da

ll ra

nk c

orr

ela

tio

n t

ests

usin

g o

utlin

e d

ata

(N

= 2

84

).

4P

CS

(1

0H

) =

Usin

g t

he

first

fou

r p

rin

cip

al co

mp

on

en

ts s

co

res f

rom

the

first

ha

rmo

nic

s (

se

e t

ext)

; E

S =

Eig

en

sh

ap

e s

co

res,

da

sh

in

dic

ate

th

rou

gh

nu

mb

ers

. O

the

r

ab

bre

via

tio

ns a

s in

Ta

ble

III

.2.

Score

s use

dFiv

eN

-Fiv

eTen

N-T

en

Mid

N-M

idK's

pta

u

4 P

CS (

10H

)0.3

715

0.3

729

0.4

225

0.4

00.0

3

6 H

0.3

715

0.2

629

0.3

925

0.6

50.0

2

ES1_10

0.9

215

0.9

329

0.9

225

0.3

40.0

4

ES2_9

0.7

915

0.8

329

0.7

925

0.6

40.0

2

ES1

0.7

615

0.7

529

0.7

625

0.4

90.0

3

ES 2

0.7

615

0.9

429

0.8

825

0.9

60.0

0

ES 3

0.1

515

0.3

229

0.4

525

0.9

80.0

0

ES 4

0.5

615

0.6

029

0.8

525

0.9

10.0

0

ES 5

0.9

615

0.8

329

0.8

425

0.0

9-0

.07

� 85

It is worthwhile noting that many columns of Tables III.2 through III.7 do not correlate

well for various groups of genera or character suites being tested, even though the data

is more inclusive from left to right. This is because the outcomes of rarefaction

analyses depend upon membership of the “long-lived” and “shorter-lived” groups. For

instance, if 5% of the most long-lived genera are all quite close to the group mean, the

probability value reported will be low. But moving right along the same row, the 10%

most long-lived genera in the same group may now contain a genus that has very

different morphology so that the average deviation value is high and the reported

probability value is greatly increased compared with the 5% case. Moving further right,

the probability value may again drop because more long-lived genera (having greater

than a mid-range duration group) are considered such that their lower deviation values

potentially swamp out the outlier first present in the 10% group. Kendall’s taus

(reflecting the slope of the relationship) often do not correspond in sign to rarefaction

results because the relationship between morphological deviation and duration is not

linear (even after ranking) and potentially quite disperse (see Fig. III.1). For example, a

low rarefaction probability value signifies that a long-lived group is less deviant and we

expect Kendall’s test to show a negative tau, but this is not always found, regardless of

whether the relationship is significant or not.

� 86

Discussion

The results presented here for trachyleberidid ostracodes show that long-lived genera

are either no different from shorter-lived genera or perhaps deviant morphologically

than shorter-lived genera. This contrasts with the previous finding that genera of

crinoids in crinoid orders are morphologically less deviant than expected by chance

alone (Liow 2004). One possible bias in this ostracode data is incomplete sampling,

despite a thorough exploration of the literature. However, the preservation probability

(per 10 M.y.) is 0.28 for all trachyleberidid genera considered together, very low for

genera that are still extant (0.19) and very high for genera that are extinct already (0.92)

(using Foote & Raup’s FreqRat [1996]). In fact, only 29 of the 326 genera are solely

represented in Recent samples. This is a rather unusual situation. But it indicates that

fossil trachyleberidids are very well-sampled and hence the reliability of stratigraphic

ranges of genera should be quite high. On the other hand, there is the greater likelihood

that some taxa with shorter geologic ranges may actually have their ranges slightly

extended if and when members are discovered in the Recent oceans. The “missing”

Recent genera should not systematically bias the result of this study unless they

overwhelmingly lengthen durations of extinct genera, a possibility deemed unlikely

since the Pliocene and Pleistocene both seem to be well-sampled.

Other possible explanations for the discrepancy between the crinoid study and the

current one are that i) the patterns could be clade specific due to differences in duration � 87

distributions and biology, ii) orders (of crinoids) encompass an evolutionarily larger set

of taxa than a family (Trachyleberididae) and hence produce different morphological-

deviation-duration patterns, iii) the crinoid study encompassed a longer period of time

(Ordovician to Eocene) than the current one (Cretaceous to Recent), iv) the two datasets

may have different sampling artifacts.

It may be that the trachyleberidid morphological deviation-duration pattern is truly a

non-existent one, as illustrated in a theoretical null expectation (Fig. III.1). This may

extend to the speculation that ecological specialization is not related to geologic

duration of the taxon in question. The previous statement is based upon the assumption

that morphology, or at least the chosen parts of the morphology that was coded and

analyzed, is correlated with ecology such that morphological specialization equates to

ecological specialization. There is however no empirical evidence for this relationship

in ostracodes, thus this speculation is groundless for now.

Another question that arises is why contemporaneous subsets of genera differ in their

pattern of morphological deviation versus duration from the whole dataset or when the

data are divided into birth cohorts. I hypothesize that contemporaneous subsets of

genera are groups that are potentially closely interacting during a particular set of global

conditions. This is in contrast to all genera through the entire length of the existence of

the family, since the genera at the beginning of the family’s history do not directly � 88

interact with later genera. This is also in contrast to birth cohorts, which do not include

all the potentially ecologically interacting genera existing during the geological interval

of their origin. However, the marginally significant morphological deviation of long-

lived contemporaneous genera compared with shorter-lived genera disappeared when a

Principal Coordinates Analysis was run. This is perhaps because some correlated

characters that were contributing to the deviation of long-lived genera from the group

mean in the distance analysis of contemporaneous genera lost some of their concerted

influence on the resulting patterns from the analysis.

There are other explanations for the relationship (or the lack of one) between longevity

and morphological deviation that I have not examined here. Environmental events such

as climate change, sea level rise and fall may contribute to genus longevity directly or

indirectly. For instance, an extinction event caused by climatic changes may directly

remove certain types of morphologies to result in a new distribution of genera in

morphospace. It can also remove competitors or predators from other clades that

indirectly affect trachyleberidid longevity and morphospace distribution. Genera in

different geographical realms could have been unevenly sampled, have experienced

different regional historical events and differ in ecology. Phylogeny could also

contribute to the resulting morphological deviation-duration patterns by non-randomly

contributing to certain types of morphologies or life histories or ecologies that promote

� 89

taxic longevity. Lastly, interactions of external events and ecology could themselves be

determinants of morphology and persistence.

Conclusions

In this study, I have used an exceptionally well-sampled group of marine microfossils to

test the idea of the persistence of the relatively unspecialized (Simpson 1944).

Specialization is here defined as morphological deviation from a group mean. The

more distant or different a genus is from a mean morphology, the more morphologically

specialized it is considered to be. The closer a genus is to a mean morphology, the more

morphologically average it is considered to be. Long-lived taxa were identified using

three methods, namely, the most long-lived 5% of the genera, the most long-lived 10%,

and taxa having durations greater than the mid-range duration value of the group.

Sample sizes of long-lived taxa changed according to the definition of “long-lived”

(Liow 2004). Using rarefied sampling, equivalent samples of shorter-lived genera were

compared with long-lived ones.

In general, long-lived trachyleberidid genera are no more or less morphologically

deviant compared with shorter-lived ones. Contemporaneous subsets of genera

occurring in epochs, however, ostensibly have longer-lived genera that are more deviant

from the mean morphology during any one epoch. Although the results are not always

statistically significant at the level of p = 0.025/0.975, the data do point to the � 90

possibility that longer-lived genera are more deviant from an average morphology than

expected. One hypothesis, if the effect is real, is that decreased competition by

specialization may aid persistence. Another possibility is that the long-lived genera in

each epoch (which are not independent in successive epochs) have fewer unknown

character states so that they appear more deviant. However, this cannot be the sole

explanation because when single-staged genera (= potentially less well-sampled) and

when genera with many unknown or inapplicable characters were removed, long-lived

genera are more deviant in their external characters and all characters combined.

Dissecting the discrete morphological data in various other ways, including comparing

birth cohorts and related groups of morphological characters separately, showed that

long-lived genera are no more or less morphologically deviant than shorter-lived ones.

The few exceptions to this can be attributed to low generic sample sizes and high

proportions of unknown and uncodable characters. External characters may have more

influence than internal ones in producing patterns of morphological deviation and

longevity as shown by analyses of contemporaneous cohorts. Outline data analyzed

using two independent methods show that trachyleberidid genera that are long-lived are

not more or less deviant from an average morphology than are their shorter-lived

counterparts.

� 91

Specialization in discrete morphology, especially external morphology, may be

positively correlated with longevity in contemporaneous subsets of trachyleberidid

genera. This relationship may be true even for temporally longer contemporaneous

groups of genera if discrete morphology becomes more completely known and

taxonomy improved. Lateral outline data are not correlated with longevity although it is

a very important aspect of genus taxonomic identification (Bachnou et al. 2000). In this

world of perpetual change, knowing why, how and when lineages do not change for

long time periods informs us in a novel way, on the myriad factors contributing to

radiations and turnovers.

� 92

CHAPTER IV

DOES VERSATILITY AS MEASURED BY GEOGRAPHIC RANGE,

BATHYMETRIC RANGE AND MORPHOLOGICAL VARIABILITY

CONTRIBUTE TO TAXON LONGEVITY?1

Introduction

Extinction risk and extinction selectivity are foci of today’s ecological research

(McKinney 1997, O’Grady et al. 2004, Sodhi et al. 2004, Reynolds et al. 2005).

Body size, life history variables, range size, endemicity, genetic variability, among other

factors, have been examined as contributors to survival probability of extant populations

(Spielman 2004, Cardillo et al. 2005, Saether et al. 2005). However, for a more

complete understanding of the general factors contributing to realized lineage

longevities, we need to turn to the fossil record.

Some fossil taxa survived for longer periods of geologic time than their relatives

(Stanley 1979, Jablonski 1994). Their observed persistence cannot simply be explained

by preservation or other sampling biases (Foote & Raup 1996). Having greater lineage

longevity involves i) not becoming extinct, during intervals of background extinction, at

mass extinctions or somewhere along this continuum, and ii) not evolving into another

� 93

1 This paper was accepted for publication in Mar 2006 in Global Ecology and Biogeography.

taxon without leaving populations of the ancestral taxon (pseudo-extinction). What

could promote increased longevity?

Ecological versatility, here defined as the number of physical locations (e.g. width of

geographic range) and ecological conditions (e.g. different temperature regimes) in

which a lineage can survive, could aid lineage survival (Jackson 1974, Boucot 1975b,

Jablonski 1980, Martinell & Hoffman 1983, Jablonski 1986b, Kammer et al. 1997, Bean

et al. 2002, Viranta 2003, Harley et al. 2004, Kiessling & Baron-Szábo 2004, Bown

2005). However, contrary or non-significant results have also been found, especially

across severe extinction events (Stanley 1986, Norris 1992, Jablonski & Raup 1995,

McClure & Bohanak 1995), presumably because the magnitude of environmental

change during these times is greater than can be tolerated by even ecologically versatile

lineages. Similarly, evolutionary versatility, here approximated as the propensity to

give rise to daughter taxa or morphological variability, could also be positively

correlated with lineage longevity (Flessa & Jablonski 1985, King & Hanner 1998, Liow

2004). Having more progeny to increase chance survival is analogous to increasing

reproductive output in individuals.

Here, I use extensive data on the Trachyleberididae (Podocopida: Ostracoda), to pose

questions involving lineage longevity. Ostracodes are particularly suited to ask

macroecological questions in the fossil record, because of their very abundant and � 94

continuous fossil record that has long been studied intensively due to their utility in

applied geology (Maddocks 1983, Colin & Lethier 1988, Reyment 1988, Keen 1993,

Athersuch 1994, Boomer et al. 2003, Ruiz et al. 2003). Results from this study will

help us to reevaluate conclusions drawn from other clades, both extant and extinct,

which may have different ecologies, preservation potentials and states of taxonomic

knowledge.

Trachyleberidids are marine benthic ostracodes that were already diverse by the late

Cretaceous and are a substantial part of marine benthic communities today. They are

abundant all along the marine depth gradient, from brackish waters to the abyssal

plains. Shallow water species are commonly epiphytic on plants and those in deep

waters may be detritus feeders (Swain 1974). Ostracodes lack pelagic phases, although

they can achieve extremely widespread distributions (Whatley & Ayress 1988),

achieved literally by walking (Benson 1973), although they must occasionally disperse

via currents, rafting or other accidental means.

Ostracode species often have sufficiently short geologic ranges to be useful in defining

biozones that can be correlated across different locations (van Morkhoven 1963). There

are many endemic species, as well as very widespread and long-ranging ones (Whatley

& Ayress 1988). Ostracode geologic ranges (together with their geographic locations

and (paleo)depths are often reported in the taxonomic and biostratigraphic literature but � 95

analyses of ostracode ranges as focal points have been rare and highly descriptive in

nature (Swain 1992).

Here, I test the hypothesis that both genus and species longevity are positively

correlated with ecological versatility (here measured using the number of bathymetric

zones traversed and geographic spread). Concurrently, I factor out sampling biases that

may be the primary cause of an observed correlation by subdividing my dataset into

broad geographic areas and time periods and by removing singleton and extant taxa. I

also correct geographic ranges and the number of occurrences by the number of times a

species was mentioned in the literature and attempt to reduce sampling biases by using

rarefaction techniques. I also test if genus and species longevity are positively

correlated with evolutionary versatility (the number of species the number of subspecies

and extreme species morphological variability). I then compare geographic spread,

bathymetric range together with sampling and other confounding factors (such as taxon

age, which is expected to positively correlate with longevity, at least for extant taxa) to

investigate which factors contribute more strongly to observed longevity. I discuss if

species patterns are sufficient to explain genus patterns and conclude by comparing the

longevity patterns of trachyleberidid ostracodes with other clades.

� 96

Methods and Materials

Taxonomic and morphological data

My database of species and genera of Trachyleberididae s.l. (including

Trachyleberididae s.s., Hemicytheridae and Cytherettidae) is updated from a previous

database involving the morphology of trachyleberidid genera (Liow 2006). The family

s. l., is evidently monophyletic as found by molecular techniques (pers. comm. T.

Oakley). I systematically traced trachyleberidid taxa using online databases (Georef,

Geobase, Zoological Records, Biological Abstracts and the Web of Science or ISI), the

Kempf (1986-2005) Database on Ostracoda and the primary literature. Many obscure

references seen in the Kempf database were not available within the period of this study

and they encompass about 800 species names. Some of these species names are possibly

synonyms. The c.800 excluded species are taxa from less well-studied regions of the

world, which do not contribute as much reliable data in terms of depth of occurrence,

geographic or geologic ranges (see Discussion). My database contains 398 genera and

4216 species, ranging from the late Jurassic (trachyleberidid-like taxa) to the Recent

and with a global coverage.

Ambiguous taxonomic assignments (e.g. cf., aff. and ?), nomina nuda and unnamed

species were recorded but discarded for the purposes of these analyses. Most of these

taxa are very rare in the studies that report them and are not likely to be sampled again

(Koch 1987) or there is substantial uncertainty in their identity, possibly due to poor � 97

preservation. Preliminary analyses involving these ambiguous entries did not

qualitatively change results. I use the most current genus assignment determined from

the literature unless there is evidence that the current revision may be less informed than

an older one.

While collecting data from the literature, it was apparent that species reported more

frequently might not only be more abundant in sediments but also better known

taxonomically and more recognizable morphologically. Some of these may also be

garbage-can or cryptic species. To keep track of the variation in sampling intensity,

each new report (even in the same county or state) of a species that I encountered in the

literature is noted as an additional literature report for that species.

I recorded all the subspecies that were recognized. I also noted species that authors

described as highly morphologically variable or having many morphotypes in the same

sample, outcrop or local region, beyond the variation recorded among instars and

between the sexes. Only if these purported morphotypes were examined by the same

author in each species, were they coded as such. This is to have some confidence that

these taxa have a greater chance of being truly variable lower taxa than simply being

mis-identified. It is acknowledged that these are possibly different but closely related

species that have maintained geographic coexistence to some extent and that some truly

morphologically variable species may not be coded as such. For instance, Hazel (1967) � 98

reported that “ the variability of R. tuberculata is great,” confirming earlier

observations and Brouwers (1993) further described the variation and included

morphological plots of Robertsonites tuberculata (Sars 1865). Hence R. tuberculata is

coded in my database as highly morphologically variable.

Geographic range data

Genus geologic ranges ( = genus longevity) in my database are informed by species

ranges ( = species longevity), which are in turn tracked by occurrences of species.

Each occurrence record in my database (N = 10466) is defined by a unique combination

of the time and location at which the species in question occurs. I converted each

published occurrence of a species within a time interval to a numerical value using the

International Stratigraphic Chart (International Commission on Stratigraphy 2004). I

coarsened the location resolution of the reported data where appropriate, so that less

precise but nevertheless useful data can be accommodated. These locations are semi-

arbitrary divisions of space that tend to be present-day political units (countries, states

and natural geographical divisions, e.g. islands). Each of these locations is identified by

their current mid-point latitudinal and longitudinal positions using the online databases

at the National Geophysical Data Center for DSDP and ODP sites and the online Getty

Thesaurus of Geographic Names or, in a few cases, approximated centrally on a hard

copy of a current map.

� 99

Because of plate tectonic motions, reconstructed paleogeographic coordinates are

needed to provide a consistent basis for calculating geographic ranges. Thus, I rotated

all the occurrences one million years and older from their current coordinates to their

paleocoordinates using the program LOCROT, written by David Rowley (pers. comm.

Rowley). For example, if a record for a given species is “the Moodys Branch

Formation in a particular road cut in Clarke County, Mississippi,” regardless whether

actual present day coordinates were given, that record will be taken as Upper Eocene,

Mississippi with current coordinates 32.3, -90.2 and rotated coordinates 27.6, -77.4.

This approach was taken because ostracode biogeographic provinces are not known for

some regions of the world. Data available are also not detailed enough for a quadrant

approach (e.g. Viranta 2003). Moreover, ostracodes are described not only from coastal

outcrops but also terrestrial outcrops that are far inland, and from deep-sea cores,

rendering impossible the latitudinal linear range approach used by Jablonski et al. (e.g.

Jablonski & Valentine 1990) for taxa occurring on continental shelves. It has been

empirically shown that method and resolution should not be critical impediments to the

recognition of large-scale biogeographic patterns (Blackburn et al. 2004).

Bathymetric range data

The water depths in which extant species were collected are sometimes reported in the

literature quite precisely (within a meter) and these were used to put species in broadly

defined depth zones of occurrences (Table IV.1). The paleodepth or paleoecology of a � 100

Ta

ble

IV

.1.

Ba

thym

etr

ic z

on

es.

Th

e d

ep

th z

on

es c

od

ed

fo

r sp

ecie

s f

or

wh

ich

da

ta a

re a

va

ilab

le a

re lis

ted

, to

ge

the

r w

ith

th

e a

pp

roxim

ate

de

pth

ra

ng

e a

sso

cia

ted

with

th

em

. A

pp

roxim

ate

syn

on

ym

s a

s c

om

mo

nly

use

d in

th

e o

str

aco

de

lite

ratu

re a

re lis

ted

in

th

e la

st

co

lum

n.

Th

ese

ca

tog

orie

s

are

lo

ose

ly b

ase

d o

n K

ee

n e

t a

l.,

19

94

. Z

on

es 1

-7 a

re u

se

d in

Fig

. IV

.4 w

hile

th

e m

ore

in

clu

siv

e z

on

es 8

-9 a

re u

se

d in

co

nstr

uctin

g F

igs.

IV.2

an

d I

V.3

.

Zone

Cate

gory

Appro

xim

ate

Depth

(m

)"Synonym

s"

1bra

ckis

hN

Am

angro

ve; or

som

e f

resh

wate

r in

fluence

2in

ner

neriti

c0

-20

Lagoonal; b

each; ti

depool; lit

tora

l ; in

tert

idal; s

ubti

dal; inner

shelf

3m

iddle

neriti

c2

0-1

00

Mid

shelf; nears

hore

; circalit

tora

l; inner

sublit

tora

l

4oute

r neriti

c1

00

-20

0su

blit

tora

l; c

onti

nenta

l sl

ope; m

idsh

elf; o

ffsh

ore

5upper

bath

yal

20

0-5

00

6lo

wer

bath

yal

50

0-4

00

0in

clu

din

g c

onti

nenta

l rise

7abyss

al

>4

00

0abyss

al pla

ins,

deep o

cean b

asi

ns

8neriti

c0

-20

0sh

allo

w w

ate

rs; carb

onate

pla

tform

(=

Zones

1,2

,3 a

nd/or4

)9

bath

yal

20

0-4

00

0deep s

ea; deep w

ate

rs (

=Zones

5,6

and/or7

)

� 101

species is often inferred with a good amount of confidence from prior geological

knowledge of the area from which the ostracodes in question were collected, or from

the community composition of the ostracodes found (van Morkhoven 1963, Benson

1973). Although the paleodepth data inferred from the latter studies may appear

circular, only a few key taxa were used to determine depths of occurrences and these

include non-trachyleberidid taxa. Authors may either report a depth range for the fossil

community or use descriptive terms such as “continental shelf” or “littoral” or “deep

waters,” or state both descriptive terms and approximate quantitative measures. Some

terms have variable usage, but I tried to take into consideration the authors’ practices to

give the species reasonable depth assignments. Where the literature is ambiguous, I

used the broadest categories in Table IV.1 (Zones 8 and 9) or left the depth unassigned.

Data subsets

I divided the database (ALL) into subsets (Table IV.2). EX comprises only extinct

genera or extinct species respectively for genus and species subsets. The subset with no

singletons (NOS) is the subset with removal of species recorded only at one place at one

time interval (i.e. one occurrence record). Genus variables were recalculated with the

remaining non-singleton species. The North American (NAM) subset contains all the

occurrences of taxa in North America and Central America, including Mexico, Panama

and Caribbean islands, as thoroughly studied by van den Bold, Howe & Hazel and their

co-workers (e.g. Hazel 1967, van den Bold 1970, Howe & Howe 1973). The European � 102

Ta

ble

IV

.2.

Da

ta s

ub

se

ts

Th

e m

ea

n,

me

dia

n,

ma

xim

um

an

d m

inim

um

lo

ng

evity

(Mill

ion

s o

f ye

ars

) o

f g

en

era

an

d s

pe

cie

s a

re

rep

ort

ed

fo

r d

ata

su

bse

ts (

wh

ere

AL

L =

all

taxa

, E

X =

extin

ct

taxa

, N

OS

= s

ing

leto

n s

pe

cie

s r

em

ove

d,

CR

ET

= t

axa

exta

nt

du

rin

g t

he

Cre

tace

ou

s,

PA

LE

= t

axa

exta

nt

du

rin

g t

he

Pa

leo

ge

ne

,

NA

M =

No

rth

Am

erica

n t

axa

, E

U =

Eu

rop

ea

n t

axa

), w

ith

N b

ein

g t

he

ir r

esp

ective

sa

mp

le s

ize

s.

Th

e z

ero

va

lue

s r

efle

ct

taxa

th

at

are

fo

un

d o

nly

in

a s

ing

le t

ime

in

terv

al.

Th

e la

st

two

co

lum

ns (

ks.t

est(

da

tase

t))

rep

ort

p-v

alu

es f

or

Ko

lmo

go

rov-S

mirn

ov t

ests

be

twe

en

pa

irs o

f d

ata

se

ts a

nd

gre

y b

oxe

s in

dic

ate

asig

nific

an

t d

iffe

ren

ce

(p

< 0

.05

).

GEN

US

NM

ean

Media

nM

ax

Min

ks.t

est A

LL

ALL

338

27.5

18.6

140.5

0.0

EX

165

26.8

20.9

95.0

0.0

0.5

6

NO

S287

26.7

16.6

134.5

0.0

0.1

2

CRET

129

23.4

20.9

122.0

0.0

0.0

7ks.t

est Pale

PA

LE

161

30.6

28.9

87.7

0.0

0.0

20.0

1

NA

M109

24.4

14.6

116.5

0.0

0.2

4ks.t

est N

AM

EU

78

29.9

19.3

122.0

0.0

0.3

00.0

6

� 103

Ta

ble

IV

.2.

(co

n't)

: D

ata

su

bse

ts

SPEC

IES

NM

ean

Media

nM

ax

Min

ks.t

est A

LL

ALL

3749

4.7

0.0

68.2

0.0

EX

2723

4.9

0.0

59.2

0.0

0.0

0

NO

S2130

8.4

5.0

68.2

0.0

0.0

0

CRET

923

5.8

3.0

68.2

0.0

0.0

0ks.t

est Pale

PA

LE

1234

8.2

3.3

68.2

0.0

0.0

00.0

0

NA

M681

4.8

0.0

56.4

0.0

0.5

7ks.t

est N

AM

EU

1220

4.5

1.5

51.3

0.0

0.0

60.0

2

� 104

(EURO) subset contains all the occurrences of taxa in Western Europe, a geographic

region for which ostracodes have also been well studied for a long time (Benson 1966).

The Cretaceous (CRET) subset contains all taxa that were extant during the Cretaceous

(including those with first and last appearances outside of the Cretaceous) and the

Paleogene (PALE) subset contains all the taxa that were extant during the Paleogene

(including those with first and last appearances outside of the Paleogene). Jurassic taxa

are very few and their identities as trachyleberidids are uncertain while Neogene taxa

tend to exhibit range truncation toward the Recent if they are extant. Hence these two

obvious temporal divisions of data are left out. The data and bibliographic sources are

available upon request.

Analyses

To test for significant differences in 1) longevity distributions among data subsets and

2) proportions of species occurring in various depth zones and during different time

intervals, I used the Kolmogorov-Smirnov test, henceforth K-S test (Sokal & Rohlf

1995).

To test for significant correlations between latitudinal ranges, longitudinal ranges, the

number of records and longevity, I used non-parametric correlation tests. This is

because assumptions of parametric tests are violated by my data (Sokal & Rohlf 1995).

Spearman’s rank test gave the same qualitative results as Kendall’s test in all cases. � 105

Therefore I report and discuss only Kendall’s tau and the associated Bonferroni

corrected probability values (Sokal & Rohlf 1995). I corrected for sampling intensity

by dividing the geographic ranges and occurrence records with the number of literature

reports, and then recalculating correlation coefficients. Similarly, I report Kendall’s tau

and Bonferroni corrected probability values for the correlations between depth range,

morphological variability and species richness, versus species and/or genus longevity.

A second method I used to account for unequal sampling intensity is rarefaction

analysis. I single out species with 2 or more records, then only species with 3 or more

records, through to only those with 8 records or more (“Qualifying” in Appendix J). I

then randomly chose a fixed number of records (“Rarifying” in Appendix J) associated

with qualifying species to control for their “commonness” in the literature. I then

calculated rarified longevities and latitudinal and longitudinal ranges and reanalyzed the

correlation between them. Each rarefaction exercise is repeated 100 times. Rarefaction

was not used to equalize depth records because relatively fewer primary literature

sources reported individual depth occurrences, as compared with geographic

occurrences. Most depth data were reported from composite sources, as composite

depths and are hence not adequately structured for a rarefaction exercise.

Genus level characters (genus geographic range and depth range and longevity) are not

independent from species level characters because they are calculated directly from � 106

those at the lower taxonomic level. To test for significant relationships beyond an

expected autocorrelation, I used a randomization approach. For example, I randomly

drew values from the list of original species median longevities, with replacement, and

assigned those to genera. When a randomly drawn species longevity value was larger

than the associated genus longevity, I discarded that value and randomly drew another

one until the drawn value is logically possible. The probability of the correlation for

these randomized values were then compared with that of the original data.

I used multiple-regression to investigate which variables contribute more strongly to

genus and species longevities. For genus longevity, latitudinal range, longitudinal

range, depth range, the number of records, the number of literature reports, age (first

appearance in the fossil record) and species richness were included as variables. For

species longevity the same variables were included except the last. None of the

variables are normally distributed, thus I ranked the variables and used the resulting

ranks as inputs for multiple-regression analysis (Conover & Iman 1981).

All analyses were performed using R (R Development Core Team 2005).

� 107

Results

Longevity and ecological versatility I: geographic spread

The mean number of occurrence records (of species constituting genera), latitudinal and

longitudinal ranges of genera, are strongly positively correlated with genus longevities

for all subsets of data (Fig. IV.1, Table IV.3). Similarly, the number of records, and

latitudinal and longitudinal ranges of species, are also strongly positively correlated

with species longevities (Table IV.3).

Correcting for sampling intensity by dividing geographic ranges and the number of

occurrence records by the number of literature reports generally did not change the

strong positive relationships between genus longevity and latitudinal and longitudinal

ranges or the number of records. For the subsets EX, NOS, NAM and EURO, number

of records was unrelated to genus longevity after this correction. This is not the case for

species where a significant positive relationship between longevity and the number of

records, latitudinal and longitudinal ranges remained after this correction, except for the

last comparison for NOS (Table IV.3).

After rarifying the records, neither longitudinal nor latitudinal ranges of genera correlate

with genus longevity, with only one exception (Appendix J). It should be noted that this

culling exercise is very severe because it leaves out, together with truly rare or little-

known species, locally common and well-known species whose occurrences are� 108

/Users/LH/Users/LH

0 20 40 60 80 100 140

02

04

06

08

01

00

12

0Fig. IV.1:

Geographic range vs genus longevity

Genus longevity, M.y.

La

titu

din

al ra

ng

e,

de

gre

es

/Users/LH

tau = 0.54, p < 0.0001

cor-tau =0.51 p < 0.0001

Fig. IV.1. Genus longevity (M.y.) plotted versus genus latitudinal range (degrees) for the whole dataset (ALL). Solid circles represent uncorrected latitudinal ranges and empty ones represent those divided the number of literature reports. Rank correlation coefficients and probabilities are reported for each case (tau for the uncorrected and cor-tau for the corrected).

� 109

Table IV.3. Geographic spread and longevity

Correlation cofficients (Kendall's tau) calculated from comparing the listed variables

and taxon longevity are reported for various data subsets (where ALL = all taxa,

EX = extinct taxa, NOS = singleton species removed, CRET = taxa extant during

the Cretaceous, PALE = taxa extant during the Paleogene, NAM = North American

taxa, EU = European taxa). The asteriks * ,** and *** represent significance at the p = 0.05, 0.01 and 0.001 levels respectively (after Bonferonni correction for 3x7 = 21

overlapping datasets). NS = not significant. Square brackets indicate a change in

significance if values divided by the number of literature reports were used to

calculate correlation coefficients.

Genus Species

Mean

Species

Occurrence

Records

Latitudinal

Range

Longitudina

lRange

No.

Occurrence

Records

Latitudinal

Range

LongitudinalR

ange

ALL 0.29 0.54 0.47 0.69 0.63 0.61

*** *** *** *** *** ***

EX 0.39 0.63 0.61 0.73 0.71 0.68

*** [NS] *** *** *** *** ***

NOS 0.31 0.45 0.39 0.29 0.19 0.16

*** [NS] *** *** *** *** *** [NS]

NAM 0.21 0.51 0.51 0.57 0.51 0.50

NS[NS] *** *** *** *** ***

EURO 0.20 0.51 0.45 0.72 0.67 0.64

*[NS] *** *** *** *** ***

CRET 0.30 0.61 0.54 0.72 0.67 0.67

***[NS] *** *** *** *** ***

PALE 0.39 0.66 0.59 0.73 0.70 0.66

*** *** *** *** *** ***

� 110

combined in very few records or even just one. The rarified species data show a

different result. After rarifying the records, latitudinal ranges of species are still

significantly correlated with species longevity in 5 out of 21 cases, in particular, if the

number of qualifying species is equal to the number of sampled species. Longitudinal

ranges of genera are not significantly correlated with species longevity after rarefaction,

except in a few cases (Appendix J).

Longevity and ecological versatility II: bathymetric spread

The number of depth zones occupied by genera has no consistent bearing on their

longevities although both the subset excluding singletons (NOS) and the subset of

Paleogene genera (PALE) indicate a significant positive one (Table IV.4). Species data

subsets show more cases of significant positive relationship between depth range and

longevities (Table IV.4), although again, the significance is not universal across the

subsets of data and correlation coefficients are small.

Genera consisting only shallowly distributed or only deeply distributed species are

significantly different from genera that span both shallower and deep waters, which

have greater mean and median longevities (K-S test, p << 0.05). Even when extant taxa

are removed, the longevities of these three subdivisions of depth occupation are still

very different though the difference is significant only between deeply distributed

genera and those with mixed distributions (Fig. IV.2). � 111

Longevity distributions of shallowly distributed species, deeply distributed species and

those with mixed-depth zone occupation of either the global or extinct datasets are not

significantly different from one another (K-S test, p >> 0.05 in all cases, Fig. IV.3).

However, both mean and median longevities of species with mixed-depth occupation

are greater than those of exclusively shallow and deep species even though the

longevity distributions are not significantly different.

One concern with paleobathymetric information is that some time intervals are better

known than others. However, the proportions of species occupying different depth

zones are not different for Cretaceous, Paleogene and Neogene time intervals (Fig. IV.4,

K-S tests, p >> 0.05 after Bonferonni correction for three overlapping datasets), despite

more available information on the depth distribution of the largely extant Neogene

species. It is acknowledged, however, that even though there is no global difference in

the distribution of depth zones for the three broad time intervals, regional and local

differences could still bias the data.

� 112

Table IV.4. Bathymetric range versus longevity

Correlation values (Kendall's tau) between bathymetric range and

longevity. The asteriks * ,** and *** represent significance the at p = 0.05, 0.01 and 0.001 levels respectively (after Bonferonni correction

for seven overlapping datasets). NS = not significant. Sample sizes (N)

are shown because depth data are available only for some taxa.

As before, ALL = all taxa, EX = extinct taxa, NOS = singleton species

removed, CRET = taxa extant during the Cretaceous, PALE = taxa

extant during the Paleogene, NAM = North American taxa,

EU = European taxa.

N GENUS N SPECIES

ALL 228 -0.11 NS 976 0.09 ***

EX 98 -0.15 NS 442 0.12 **

NOS 194 0.33 *** 917 -0.01 NS

NAM 64 0.09 NS 194 0.11 NS

EURO 78 0.02 NS 202 0.13 **

CRET 68 0.12 NS 179 0.09 NS

PALE 76 0.23 * 255 0.27 ***

� 113

/Users/LH

Fig. IV.2a: All, Shallow

0 100

/Users/LH

010

20

30

N = 160 mean = 26.7

median = 20.7 max = 122.0

Fig. IV.2b: Shallow & Deep

0 100

010

20

30

N = 66 mean = 48.7

median = 44.25 max = 140.5

Fig. IV.2c: Deep

0 100

010

20

30

N = 16 mean = 19.8

median = 17.7 max = 63

Fig. IV.2d: Ex : Shallow

0 100

010

20

30

N = 65 mean = 34.3

median = 30.7 max = 93.0

Fig. IV.2e: Shallow & Deep

Genus Longevity M.y.

0 100

0

/Users/LH

10

20

30

N = 17 mean = 45.1

median = 40.2 max = 95.0

Fig. IV.2f: Deep

0 100

010

20

30

N = 10 mean = 14.9 median = 8.2 max = 43.0

Fig. IV.2 Histograms of genus longevities as subdivided by whether they occupy only shallower waters (Zone 8 in Table IV.1), only deep waters (Zone 9) or both. All = all genera, Ex = extinct genera, N = sample size, mean = mean genus longevity (M.y.), median = median genus longevity (M.y.), max = maximum genus longevity (M.y.).

� 114

/Users/LH

Fig. IV.3a: All Shallow

/Users/LH

0 30 60

020

60

100

N = 915 mean = 6.4

median = 2.1 max = 59.2

Fig. IV.3b: Shallow & Deep

0 30 60

020

60

100

N = 74 mean = 9.7

median = 2.7 max = 45.4

Fig. IV.3c: Deep

0 30 60

020

60

100

N = 130 mean = 8.6

median = 2.1 max = 59.9

Fig. IV.3d: Ex Shallow

0 30

/Users/LH

60

020

60

100

N = 511 mean = 7.8

median = 4.6 max = 59.2

Fig. IV.3e: Ex Shallow & Deep

Species Longevity M.y.

0 30 60

020

60

100

N = 22 mean = 15.21 median = 13.0

max = 45.4

Fig. IV.3f: Ex Deep

0 30 60

020

60

100

N = 72 mean = 11.2 median = 7.2 max = 43.0

Fig. IV.3. Histograms of longevities of species as subdivided by whether they occupy only shallower waters (Category 8 in Table IV.1), only deep waters (Category 9) or both. All = all species, Ex = extinct species, N = sample size, mean = mean species longevity (M.y.), median = median species longevity (M.y.), max = maximum species longevity (M.y.).

� 115

One concern with paleobathymetric information is that some time intervals are better

known than others. However, the proportions of species occupying different depth

zones are not different for Cretaceous, Paleogene and Neogene time intervals (Fig. IV.4,

K-S tests, p >> 0.05 after Bonferonni correction for three overlapping datasets), despite

more available information on the depth distribution of the largely extant Neogene

species. It is acknowledged, however, that even though there is no global difference in

the distribution of depth zones for the three broad time intervals, regional and local

differences could still bias the data.

Longevity and evolutionary versatility I: species richness

Species richness is significantly positively correlated (p < 0.001) with genus longevities

in both the global data and all the subsets (Kendall’s tau ranging from about 0.50 to

0.60, detailed results not shown), even when possible garbage can genera (Cythereis,

Trachyleberis, Cytheretta) are removed.

Longevity and evolutionary versatility II: subspecies richness

Of the 4216 species in my database, 279 have two to nine subspecies described. The

longevity distributions of these species and the genera that contain them are

significantly different from the dataset as a whole (K-S test, p < 0.001 in all

comparisons, median longevity for these species = 8.1 M.y. and median longevity for

� 116

/Users/LH

Fig. IV.4 : Distribution of species occupying various depth zones

Neogene

34 (5%)

320 (50%)

162 (25%)

49 (8%)

20 (3%)

58 (9%)

3 (1%)

Paleogene

7 (5%)

58 (39%)

52 (35%)

24 (16%)

2 (1%)

5

/Users/LH

(3%)7

(0%)

Cretaceous

4 (3%)

54 (39%)

30 (22%)

30 (22%)

13 (9%)

6 (3%)

7 (0%)

brackish neritic bathyal abyssal

Fig. IV.4. Distribution of species occupying various depth zones during the Cretaceous, Paleogene and the Neogene. Number (top values) and percentage (bottom values) of species occurring at bathymetric zones listed in Table IV.1.

� 117

these genera = 53.2 M.y., see Table IV.2 to compare these values with other data

subsets).

Longevity and evolutionary versatility III: extreme species morphological variability

Twenty-seven species (of 4216 species) have been described as highly variable in

morphology. Of these 27 highly variable species, seven also have subspecies assigned

to them. Similarly, longevity distributions of these species and the genera that contain

them are significantly different from the dataset as a whole (K-S test, p < 0.001 in all

comparisons.) In fact, the median genus and species longevities are about doubled for

morphologically variable genera (respectively 54.5 and 10.3 M.y.), compared with the

dataset as a whole.

Which factors are stronger?

Genus longevity has been shown in the previous sections to be positively correlated

with geographic spread, species richness and only weakly related to bathymetric spread.

However, genus age and sampling can also contribute to the observed genus longevity.

The older a genus is or the earlier it first appears in the fossil record, the greater its

chance of having an increased longevity compared with younger genera whose

longevity is necessarily capped by frame of observation that includes the Recent time

interval. Additionally, the more frequently a species is sampled in the fossil record the

more likely its known longevity would be lengthened, simply by chance. Multiple-� 118

regression shows that the two most important factors contributing to genus longevity are

genus age and species richness, regardless of whether the entire (ALL) or the extinct

(EX) dataset is used.

Similarly, species longevity is related to geographic spread and bathymetric range but

species age and sampling also contribute to the observed species longevities. Simple

multiple-regression shows that latitudinal spread is the strongest factor contributing to

species longevity. Even though age and sampling do play a part, their contributions are

not as strong (Table IV.5). Again, both the whole dataset and the extinct dataset show

the same qualitative result.

Are species patterns sufficient to explain genus patterns?

Randomly assigned median species longevity values are barely correlated with genus

longevities (tau = 0.1, p = 0.06) but the correlation in the original dataset is strong (tau

= 0.24, p = 3.0x10-11, Fig. IV.5). Randomly assigned median species latitudinal range

values are correlated with genus latitudinal ranges (tau = 0.34, p = 0.05) but the

correlation in the original dataset is much more probably (tau = 0.16, p = 1.1x10-5).

Randomly assigned species mean and median bathymetric ranges values are

respectively correlated and not correlated with the respective genus bathymetric ranges

(tau = 0.22, p =0.005; tau = 0.11, p =0.09). But again, the correlation in the original

dataset is much stronger (tau = 0.55, p = 2.2x10-16; tau = 0.33, p = 12.4x10-13, � 119

Ta

ble

IV

.5.

Re

su

lts o

f m

ultip

le r

eg

ressio

n.

Re

su

lts o

f m

ultip

le-r

eg

ressio

n a

na

lysis

wh

ere

lo

ng

evitie

s o

f g

en

era

an

d s

pe

cie

s w

ere

re

gre

sse

d o

n

the

fa

cto

rs lis

ted

in

th

e f

irst

co

lum

n.

Re

su

lts f

rom

wh

ole

da

ta s

et

(AL

L)

an

d t

he

extin

ct

da

tase

t (E

X)

for

ge

ne

ra a

nd

sp

ecie

s a

re s

ho

wn

. T

he

aste

riks *

,**

an

d *

** r

ep

rese

nt

sig

nific

an

ce

at

the

p=

0.0

5,

0.0

1 a

nd

0.0

01

le

ve

ls r

esp

ective

ly.

Est

imate

Std

Err

or

t v

alu

ep

valu

eF

-sta

tist

icp

valu

eR

-sq

uare

dA

dju

sted

R

Gen

us/

AL

L

Inte

rcep

t-2

3.8

35

9.2

71

-2.5

71

0.0

11

*193

0.0

00

0.7

44

0.7

40

lati

tudin

al r

ange

0.0

57

0.0

52

1.1

04

0.2

70

dep

th r

ange

0.0

18

0.0

37

0.4

92

0.6

23

lite

ratu

re r

eport

s0.1

12

0.0

35

3.2

43

0.0

01

**

gen

us

age

0.4

55

0.0

30

15.2

91

0.0

00

***

spec

ies

rich

nes

s0.4

97

0.0

47

10.5

03

0.0

00

***

Gen

us

/ E

X

Inte

rcep

t-7

.695

7.1

09

-1.0

82

0.2

81

70

0.0

00

0.6

87

0.6

77

lati

tudin

al r

ange

-0.0

27

0.0

88

-0.3

02

0.7

63

dep

th r

ange

0.0

64

0.0

61

1.0

56

0.2

92

lite

ratu

re r

eport

s0.1

87

0.0

60

3.0

92

0.0

02

**

gen

us

age

0.2

09

0.0

49

4.2

55

0.0

00

***

spec

ies

rich

nes

s0.6

60

0.0

76

8.6

63

0.0

00

***

� 120

Ta

ble

IV

.5.

(co

n't)

Re

su

lts o

f m

ultip

le r

eg

ressio

n.

Est

imate

Std

Err

or

t v

alu

ep

valu

eF

-sta

tist

icp

valu

eR

-sq

uare

dA

dju

sted

R

Sp

ecie

s/A

LL

Inte

rcep

t121.1

93

39.4

87

3.0

69

0.0

02

**

1528

0.0

00

0.6

20

0.6

20

lati

tudin

al r

ange

0.6

97

0.0

11

64.9

89

0.0

00

***

dep

th r

ange

-0.0

16

0.0

10

-1.5

76

0.1

15

lite

ratu

re r

eport

s0.0

40

0.0

11

3.6

55

0.0

00

***

spec

ies

age

0.2

14

0.0

10

22.0

75

0.0

00

***

Sp

ecie

s/E

X

Inte

rcep

t12.2

92

12.2

39

1.0

04

0.3

16

142

0.0

00

0.5

80

0.5

76

lati

tudin

al r

ange

0.5

48

0.0

41

13.4

68

0.0

00

***

dep

th r

ange

0.0

60

0.0

42

1.4

26

0.1

55

lite

ratu

re r

eport

s0.2

52

0.0

42

6.0

13

0.0

00

***

spec

ies

age

0.0

81

0.0

32

2.5

23

0.0

12

*

� 121

/Users/LH

0 20 40 60 80 100 140

01

02

03

04

0Fig. IV.5: Genus vs species longevity

Genus longevity (M.y.)

Me

dia

n s

pe

cie

s lo

ng

evity (

M.y

.)

IMPOSSIBLE

original correlation p = 3.0 e-11

randomized values correlation p = 0.06

Fig. IV.5. Genus versus species longevity.The area on the left delimited by the line y = x is the logically impossible area (IMPOSSIBLE) for the plot of median species longevity versus genus longevity. The original simple correlation (Kendall) of the plotted data shows a significant p-value as expected but when compared with a randomized dataset, the original p-value is shown to be highly significant.

� 122

respectively). Thus species variables do scale up to genus variables (e.g. genera with

greater longevities have member species with greater longevities).

Discussion

Ecological versatility, as measured by geographic and bathymetric spread does to some

extent, contribute to genus and species longevities. Evolutionary versatility as

measured by species or subspecies richness and morphological variability is also

associated with genus and species longevity.

It has been often verified empirically that taxa with greater longevities should be

geographically widespread (Jackson 1974, Martinell & Hoffman 1983, Jablonski 1986b,

1987, Jablonski & Raup 1995, this paper). Unfortunately for the purpose of analyses,

widespread taxa are also encountered or sampled with a greater probability (McKinney

1986, this paper), thus possibly producing a positive correlation between longevity and

geographic range when there is none (Russell & Lindberg 1988). Although many

paleontological studies have explicitly accounted for sampling effects (Pease 1985,

Koch & Morgan 1988, Miller & Foote 1996, Kammer et al. 1997, Marshall 1997 and

references therein), this approach has not been universally applied. After correcting

ranges and occurrence records by the number of literature reports, the strong

relationship between trachyleberidid geographic ranges and longevity encouragingly

remained in general for species and genus datasets (Fig. IV.1). However, using a rather � 123

severe rarefaction regime (Appendix J), latitudinal and longitudinal ranges of species

are only on occasion significantly correlated with species longevity. Although singleton

species can be true sampling artifacts, this result signifies that many singleton species

discarded in the rarefaction exercise were true narrowly distributed species that do

contribute to the clade pattern. The rarified genus data shows no correlation between

longitudinal or latitudinal ranges with longevities in almost all rarified cases. However,

when multiple factors were examined in concert, latitudinal range turned out to be the

most important contributor to species longevities (Table IV.5). This study serves to

confirm that even for benthic organisms that may not disperse as easily as organisms

with a planktonic dispersal phase in their life cycles, geographic range is an important

factor in promoting longevity as has been shown in mollusks (Jablonski 1980, 1986b,

2005) and foraminiferans (Buzas & Culver 1984) at least during background extinction

time intervals.

Depth is not a simple variable because it co-varies with other physical parameters (e.g.

light penetration, oxygen levels, temperature) that could affect the vertical range of a

taxon (Pineda 1993). Taxa with wider depth ranges are presumably more ecologically

tolerant (Harley et al. 2003) and hence we expect depth ranges to be positively

correlated with lineage longevities. I note that greater depth ranges may not aid

survivorship across mass extinctions (Jablonski & Raup 1995). It is also possible that

vertically spread species have less of a tendency to speciate (Pineda 1993) and thus do � 124

not have the propensity to result in the extinction or pseudo-extinction of their potential

ancestral species. Depth may also affect pelagic and benthic taxa differentially. Here I

have shown that bathymetric range only has a weak relationship to genus and species

longevities, although genera and species that live in both deep and shallow waters (i.e.

those taxa that are extremely broadly vertically distributed) do have greater longevities

(Figs. IV.2 and IV.3). However, when bathymetric range is examined in concert with

other factors, it does not contribute significantly to genus or species longevity. It is

possible that bathymetric distribution is not a good proxy for ecological versatility for

these benthic ostracodes. Temperature, grain size or nutrient level tolerance may serve

as better approximators of ecological versatility but are not available at the scale of this

study. Alternatively, the subdivisions of depth zones made in my data may not be fine

enough to capture ecological versatility, or perhaps ecological versatility as measured

by the width of depth distribution actually does not aid in trachyleberidid longevity.

Species richness, subspecies richness and morphological variability are recognized here

as evolutionary versatility. Having more species or subspecies may promote longevity

(Flessa & Jablonski 1985, McKinney 1995 but see Fortey 1980) via greater abundance

and geographic spread such that chance events have a smaller probability of wiping out

the entire lineage. Alternatively, the different species or subspecies or various

morphological forms may respond to environmental changes differently such that one

species or subspecies or form may continue surviving when changes detrimentally � 125

affect congenerics or conspecifics. A test of the two alternative pathways will require

abundance data and more detailed morphological data, neither of which is available at

the moment. Another untested possibility is that benthic ostracodes, not being able to

disperse easily as individuals, may rely more on variability for survival, in contrast with

taxa that are able to disperse widely as larvae or adults.

Historically, in paleobiology and macroevolution, the genus has always been a

convenient focal taxonomic level because it is sampled more completely than the

species. Although higher taxa like the genus have been shown to be suitable for

macroevolutionary studies (Sepkoski & Kendrick 1993, Robeck et al. 2000), the nature

of biological hierarchies can complicate longevity patterns (Valentine & May 1996).

Here, I have shown that species characters do reflect genus characters (longevity,

geographic and bathymetric range) such that when these characters are examined at the

genus level, they can potentially reflect patterns at the species level. However, details

of patterns may differ at the two taxonomic levels. At least in this data, genus longevity

is most strongly influenced by genus age (see Miller 1997) and species richness while

species longevity is most strongly affected by latitudinal range. This suggests that it is

important to study macroevolutionary patterns at different levels of the taxonomic

hierarchy (Robeck et al. 2000) as details can change and affect our understanding of the

underlying processes.

� 126

Caveats

The taxonomy of trachyleberidid ostracodes is in flux, as it is for other groups of

organisms that enjoy continued study. Inevitably, some published identifications may

be erroneous despite best efforts. For instance, some previously good species may

actually be multiple species and vice versa (e.g. Jellinek & Swanson 2003, Schornikov

2005), such that there are both range over- and under-estimates. However, most named

species are relatively undisputed (Benson 1966, personal observation), although their

membership in genera can be volatile in the literature. Moreover, no comprehensive

phylogenetic framework is available for trachyleberidids. Some recorded extinctions of

some species in my database will inevitably contain pseudo-extinctions and some of the

named taxa may also be paraphyletic, but these have not caused problems in

macroevolutionary studies (Sepkoski & Kendrick 1993, Jablonski 1994, Robeck et al.

2000). Taxonomic errors should tend to dampen significant results rather than promote

them. It has been demonstrated that results from such large-scale compilations of data

can remain robust despite new taxonomic information (Sepkoski 1993). To be really

sure that taxonomic problems are not giving falsely positively results, we need the

concerted efforts of taxonomic revisions followed by re-analyses of data. On a positive

note, however, the taxonomically better studied subsets of data in this study, namely

North American and European subsets, largely show the same patterns as the dataset as

a whole even though they each represent only about one third each of the globally

known species data in my database. This result gives some reassurance that even though � 127

taxonomic misidentifications do exist, they do not drive the results. The Paleogene and

Cretaceous datasets were also based on very different sets of taxonomic workers, and

only about 15% of the known Cretaceous species extend to the Paleogene after the

Cretaceous-Tertiary extinction, but despite their differences, results from the two data

subsets were again not dissimilar. In addition, the c.800 species names that were

excluded from the study (largely from outside of North America and Europe) may

clarify patterns if included, but since the North American and European data subsets

show largely the same patterns compared with the full dataset, we can infer that the

missing species will not change the general patterns qualitatively. Even without

detailed phylogenetic information, large-scale issues in macroevolution and

macroecology, such as those discussed in this paper, can and should be tackled (Brown

et al. 1996).

In addition to taxonomic uncertainties, longevity patterns may be due largely to changes

in sampling probabilities due to the rise and fall of global sea-levels. However, number

of originations and extinctions over geologic time for both trachyleberidid genera and

species do not vary in synchrony with eustatic sea-level changes (data not shown).

However, as noted before, this does not at all imply that local or even regional species

sampling is not affected by regional or basinal sedimentation patterns or sea-level

changes.

� 128

Conclusions

Being versatile promotes longevity and reflects the idea that generalists are better

equipped to survive for longer periods of time (Simpson 1944, Liow 2004). However,

many details of exactly how versatility operates still elude us, for example whether

species richness promotes genus longevity via greater abundance or increased

ecological tolerances. Versatility does play a real part in long-term survivorship as

observed in the fossil record, on temporal scales beyond that of most ecological studies.

� 129

CHAPTER V

LINEAGES WITH GREAT LONGEVITIES ARE OLD AND AVERAGE: AN

ANALYSIS OF MORPHOLOGICAL AND TAXON LONGEVITY

DISTRIBUTIONS USING MULTIPLE DATASETS1

Introduction

It is well known that most, if not all, clades of organisms with a fossil record exhibit a

distribution of lineage longevities resembling a hollow curve (Simpson 1944, 1953,

Levinton & Ginzburg1984, Stenseth & Maynard-Smith 1984). That is, in any given

clade, most species or genera have shorter longevities and few have great longevities.

This observation leads naturally to an important question in biology, that is, do taxa that

out-live their relatives without becoming extinct or evolving into separate lineages have

distinctive properties that aid their prolonged survival? Factors that affect lineage

longevities or survivorship may be categorized into extrinsic (environmental) and

intrinsic (biological) ones, although the two categories can and will certainly interact.

We have come a long way in understanding some of the biological characteristics that

appear to promote lineage longevity and/or damp lineage differentiation. These

biological characteristics may operate differentially during mass extinction episodes and

background-extinction time intervals, resulting in varied taxon longevities or

survivorship during specified time periods, e.g. across extinction events (Jablonski

� 130

1 This paper was submitted to Evolution in April 2006.

1986b, 1994, Jablonski & Raup 1995). They include wide geographic ranges (Jackson

1974, Boucot 1975a, Chen et al. 2005, Jablonski 2005, see previous chapter),

planktotrophic larvae (Hansen 1978, Jablonski 1986a, Jeffery & Emlet 2003), high site

occupancy (Jernvall & Fortelius 2004), deeper and wider depth distribution (Buzas &

Culver 1984, Oji 1996), general feedings strategies (Baumiller 1993), greater niche

breadth (Kammer et al. 1997, 1998) and broader ecological tolerances (Jackson 1974,

Schopf 1994).

Morphology affects the functioning and performance of organisms (Koehl 1996) and

reflects aspects of physiology and ecology (Wainright & Reilly 1994). Hence

morphology could in part be a proxy for ecology, which in turn may affect survivorship.

Yet, little is known about the distribution of morphology in relation to lineage longevity.

Past studies have focused on the relationship between morphological complexity and

longevity, with mixed results (Flessa et al. 1975, Anstey 1978, Ward & Signor 1983,

Boyajian & Lutz 1992). Lineages with greater geologic longevities might be

morphologically distant from the average morphology of the clade because being

different may ultimately confer a competitive advantage, particularly in situations of

lineage occurrence. Conversely, lineages with greater longevities might also be

morphologically closer to the average morphology of the clade than expected because

generalists may be able to survive and persist through a greater range of environmental

changes (Simpson 1944, Liow 2006). Lastly, lineages with greater or lesser longevities � 131

may not have significantly different distributions of morphology, indicating that factors

not encompassed in morphology may be operating more strongly in influencing

survivorship. Recent studies (Liow 2004, 2006) have concluded that crinoid and

ostracode lineages with greater longevities are in general not significantly different in

their morphological distance from an average morphology compared with their relatives

with lesser longevities. Similarly, morphological distances from the centroid of

morphospace of ammonoid survivors across the Perman-Triassic extinction are not

significantly different from those of victims (McGowan, accepted). This is contrary to

the long-held idea that persistent or “living fossil” taxa have special or distinctive

properties that enable them to remain unchanged while their relatives experience

speciations and extinctions (e.g. see Wills 2001).

A methodological limitation of previous attempts to investigate the relationship between

morphological dispersion and lineage longevity (Liow 2004, 2006) is that they

arbitrarily divided a continuous variable (lineage longevity) into discrete categories

(lesser/shorter and greater/longer). Another drawback was the limited taxonomic

coverage, as only crinoid genera and families (Liow 2004) and trachyleberidid

ostracode genera (Liow 2006) were investigated. These results suggested that lineages

with greater longevities have morphologies that are collectively no different from those

with lesser longevities or that are more average than expected than the latter, but the

pattern was not uniform and difficult to extend to other groups of organisms. � 132

In order to overcome the methodological limitations and test the validity of these results

(from Liow 2004, 2006) as a general evolutionary pattern, I use multiple published

datasets to have a more representative sample of phylogenetically independent clades

across more branches of the tree of life. These datasets span a wide range of body

plans, ecologies and geologic ranges. Consequently, some datasets where species are

the units of study were also included in the current analysis (compared with only genera

and families before). Each of these additional independently collected datasets provides

concordant or discordant evidence, allowing us to investigate the generality of the

longevity-morphology distribution pattern.

I test the hypothesis that lineages with greater longevities have morphologies that are

more average than expected by chance alone. The hypothesis is tested both when these

lineages are considered as a group and when they are considered individualy. This is

because each lineage with a great longevity could be either unique or lineages with

great longevities could be distinct as a group. I present a novel quantitative method to

determine whether there is a trend in which lineages with greater and greater longevities

have more average morphologies than expected. The number of lineages and characters

sampled, the taxonomic ranks of the lineages, differing preservation potentials (as

approximated by depositional setting and taxonomic identity), whether or not a dataset

represents only extinct lineages, and differences in the shape of longevity distributions

may be associated with differing patterns of morphology-longevity distributions in � 133

various clades. To my knowledge, these attributes have never been considered in a

comparative analytical framework, with respect to morphology and longevity. I also

test whether the lineages with greater longevities are significantly older than others

within a given clade, in other words, whether they arise early or late in their clade

history. Naturally, there are potential problems in using published datasets and I

consider some possible drawbacks and biases in the Discussion section. I conclude by

discussing several evolutionary implications of my results.

Methods

Data

I surveyed the literature for published morphological character matrices and retained

those satisfying the following criteria: 1) The publications reporting the data matrices

must report stratigraphic ranges either graphically or numerically for the taxa studied, 2)

These stratigraphic ranges represent a range of longevities (some datasets were

discarded because they report only equal length single stage occurrences), 3) There

should be at least nine ingroup taxa represented (outgroups as identified by the authors

of the papers were removed for my analyses), 4) The ingroup taxa should preferably be

of equivalent taxonomic ranks (exceptions are noted in Table V.1), 5) The dataset should

not consist solely of extant taxa. Datasets with many extant taxa are not preferred

because of the issues of one-sided range-truncations, i.e. the longevity of these taxa are

incomplete (see Gilinsky 1988). I however included some datasets with partial � 134

Table

V.1

. Lis

ted a

re t

he r

efe

rences u

sed in t

he a

naly

ses,

the g

roups t

hey r

epre

sent,

the d

om

ain

of

the s

tudie

s (

where

AF =

above f

am

ily,

F =

fam

ily,

SB =

subfa

mily,

SG

= g

enus,

G =

genus),

the t

axonom

ic u

nit w

hose c

hara

cte

rs w

ere

coded (

as b

efo

re,

with S

=specie

s),

the n

um

ber

of

taxa involv

ed (

N),

the n

um

ber

of

chara

cte

rs u

sed in t

he a

naly

ses (

Nchar)

, th

e g

eolo

gic

range r

epre

sente

d in t

he s

tudie

s,

the b

iolo

gic

al re

alm

in w

hic

h t

he c

lades a

re f

ound (

M =

marine,

FW

=

freshw

ate

r,T =

terr

estr

ial)

, w

heth

er

the d

ura

tions (

DU

R)

are

measure

d in m

illions o

f years

(M

Y),

sta

ges (

S)

or

manually m

easure

d o

n r

ange c

hart

s (

L),

and lastly if

infe

rred d

ura

tions (

DU

R-I

) w

ere

available

. The last

colu

mn r

ecord

s s

om

e n

ote

s m

entioned in t

he t

ext.

In

part

icula

r, b

rackets

indic

ate

those d

ata

sets

overlappin

g

with a

noth

er

[OL n

o.

from

colu

mn 1

] th

at

were

rem

oved f

or

som

e a

naly

ses.

Au

tho

rR

efe

ren

ce

Gro

up

Do

main

Un

itN

Nch

ar

Geo

log

ic

Ran

ge

Realm

DU

RD

UR

-I

1Adnet

and

Capett

a 2

001

Leth

aia

34:2

34-

248.

Oth

er

Vert

ebra

tes

Squalifo

rmes

AF

G23

29

Late

Jura

ssic

-

Ple

isto

cene

MM

YY

2Adra

in a

nd

Westr

op 2

001

Pala

eonto

logra

phic

a

Canadia

na 1

4:1

-

109.

Trilo

bites

Pty

chaspid

idae

FS

12

16

Sunw

apta

n-

Ibexia

nM

SY

3Adra

in a

nd

Edgecom

b

1997

Pp.

291 -

322 in J

.

M.

Adra

in,

G.

D.

Edgecom

be a

nd B

.

S.

Lie

berm

an (

eds).

Trilo

bites

Encrinurine

SB

S31

40

Telc

hia

n t

o

Ludfo

rdia

nM

S/L

Y

4Allm

on 1

996

Pale

onto

gra

phic

a

Am

ericana 5

9:1

-

134.

(Table

1)

Molluscs

Turr

itellid

ae

FG

/SG

51

14

Late

Cre

taceous -

Recent

MM

YN

5Allm

on 1

996

Pale

onto

gra

phic

a

Am

ericana 5

9:1

-

134.

(Table

9)

Molluscs

Turr

itellid

ae

FS

36

30

Pale

ocene-

Eocene

MM

YN

6Alroy 1

995

Syste

matic B

iolo

gy

44:1

52-1

78.

Mam

mals

Hip

parioin

es

SB

S17

56

Mio

cene-

Pliocene

TM

YY(o

nly

)

7Alv

are

z

et

al.

1998

Journ

al of

Pale

onto

logy

72:8

27-8

55.

Bra

chio

pods

Ath

yridid

sAF

F/S

F36

37

Ord

ovic

ian-

Jura

ssic

MM

YY

8Am

ati a

nd

Westr

op 2

004

Journ

al of

Syste

matic

Pala

eonto

logy

2:2

07-2

56.

Trilo

bites

Illa

enid

ae

GS

19

17

Mid

-late

Ord

ovic

ian

MS

N

9Anders

on a

nd

Roopnarine

2003

Journ

al of

Pale

onto

logy

77:1

086-1

102.

Molluscs

Corb

ulidae

FG

/S12

70

Cre

taceous-

Recent

MM

YN

10

Angie

lczky &

Kurk

in 2

003

Zoolo

gic

al Jo

urn

al

of

the L

innean

Socie

ty 1

39:1

57-

212.

Mam

mals

Dic

ynodonts

AF

G20

53

Kazania

n -

Anis

ian

TS

Y

� 135

Table

V.1

(con't)

Au

tho

rR

efe

ren

ce

Gro

up

Do

main

Un

itN

Nch

ar

Geo

log

ic

Ran

ge

Realm

DU

RD

UR

-IN

ote

s

11

Blo

ch e

t al.

2001

Journ

al of

Vert

ebra

te

Pale

onto

logy

21:1

19-1

31.

Mam

mals

Ple

sia

dapiform

es

AF

S14

32

Pale

ocene-

Eocene

TS

Y

12

Bodenbender

and F

isher

2001

Journ

al of

Pale

onto

logy

75:3

51-3

69.

Echin

oderm

sBla

sto

ids

AF

G68

94

Lla

ndeilo -

Kanzania

nM

SY

13

Bro

chu 1

997

Syste

matic B

iolo

gy

46:4

79-5

22.

Oth

er

Vert

ebra

tes

Cro

codilia

ns

AF

S61

164

Cre

taceous-

Recent

M/F

WM

YN

14

Bru

net-

Lecom

te &

Chaline 1

990

Leth

aia

24:4

7-5

3M

am

mals

Vole

sG

S16

30.5

-0 M

YA

TM

YN

teeth

only

15

Cairns 2

001

Sm

ithsonia

n

Contr

ibutions t

o

Zoolo

gy 6

15:1

-88.

Cnid

arians

Dendro

phyliid

ae

FG

/SG

30

10

Cre

taceous-

Recent

MM

YN

16

Caro

n e

t al.

2004

Journ

al of

Pale

onto

logy

78:1

138-1

145.

Art

hro

poda

Nara

odiids

AF

G/S

912

Cam

brian-

Ord

ovic

ian

MM

YY

Infe

rred

dura

tions

only

17

Dam

iani 2

001

Zoolo

gic

al Jo

urn

al

of th

e L

innean

Socie

ty 3

3:3

79-

482.

Oth

er

Vert

ebra

tes

Masto

donaro

ids

AF

G21

38

Perm

ian-

Triassic

TM

YY

18

Dashzeveg a

nd

Meng 1

998

Am

erican M

useum

Novitate

s 3

246:1

-

20

Mam

mals

Cte

nodacty

loid

Rodents

AF

G17

26

Eocene-

Mio

cene

TM

YY

19

Dew

ing 2

004

Journ

al of

Pale

onto

logy

78:2

75-2

86.

Bra

chio

pods

Str

ophenom

enat

aAF

S9

15

Ashgill-

Lllandovery

ML

N

20

Ebbesta

d &

Budd 2

003

Pala

eonto

logy

45:1

171-1

195.

Trilo

bites

Burlin

giiid

ae

FS

16

19

Mid

-Upper

Cam

brian

MS

Y

21

Fore

y 1

991

Environm

enta

l

Bio

logy o

f Fis

hes

32:7

5 -

97.

Oth

er

Vert

ebra

tes

Cole

canth

rela

tives

AF

G31

56

Scyth

ian -

Recent

MM

YN

22

Fro

elich 2

002

Zoolo

gic

al Jo

urn

al

of th

e L

innean

Socie

ty 1

34:1

41-

256.

Mam

mals

Equid

ae

FS

14

47

Eocene

TL

N

� 136

Table

V.1

(con't)

Au

tho

rR

efe

ren

ce

Gro

up

Do

main

Un

itN

Nch

ar

Geo

log

ic

Ran

ge

Realm

DU

RD

UR

-IN

ote

s

23

Gahn a

nd

Kam

mer

2002

Journ

al of

Pale

onto

logy 7

6:1

23-

133.

Echin

oderm

sBotr

yocrinid

s

(Barycrinus

)G

S10

14

Mis

sis

sip

ian

ML

N

24

Gra

nde a

nd

Bem

is 1

998

Mem

oirs o

f th

e

Socie

ty o

f

Vert

ebra

te

Pale

onto

logy 4

Oth

er

Vert

ebra

tes

Am

iidae

FS

22

46

Cre

taceous-

Recent

M/F

WM

YN

25

Hopkin

s 2

004

Journ

al of

Pale

onto

logy 7

8:7

31-

740.

Mam

mals

Rodentia

(Ansomys

)G

S9

30

38-1

5 M

YA

TM

YY

teeth

only

26

Jeff

ery

& E

mle

t

2003

Evolu

tion 5

7:1

031-

1048.

Echin

oderm

sTem

nople

urids

AF

S16

38

Eocene-

Pliocene

MM

YN

27

Jeff

ery

1998

Leth

aia

31:1

49-1

57

Echin

oderm

sCycla

ste

rG

S10

22

Late

Cre

taceous t

o

Pale

ogene

MS

N

28

Kara

saw

a a

nd

kato

2003

Pale

onto

logic

al

Researc

h 7

Oth

er

Art

hro

poda G

onepla

cid

ae

FG

15

45

Pale

ogene-R

MM

YN

29

Leig

hto

n &

Maple

s 2

002

Journ

al of

Pale

onto

logy 7

6:6

59-

671.

Bra

chio

pods

Pro

ductida

AF

G14

24

Giv

etian-

Pennsylv

ania

n

MS

N

30

Mic

haux 1

989

Alc

heringa 1

3:2

1-

36.

Molluscs

Ancillinae

SB

S20

36

Eocene -

Recent

MM

YN

31

Monks 1

999

Pala

eonto

logy

42:9

07-9

25.

Molluscs

Hete

rom

orp

hs

AF

S25

26

Low

er

Alb

ian-

Upper

Alb

ian

MM

YY

[OL 3

2]

32

Monks 2

002

Pala

eonto

logy

45:6

89-7

07.

Molluscs

Ham

itid

ae

FS

23

30

Low

er

Alb

ian -

Upper

Turo

nia

n

MM

YY

33

Monks a

nd

Ow

ens 2

000

Pala

eonto

logy

43:8

71-8

80.

Bra

chio

pods

Orbirhynchia

GS

16

22

Alb

ian-

Cam

pania

nM

MY

Y

34

Nutz

el et

al.

2000

Journ

al of

Pale

onto

logy 7

4:5

75-

598.

Appendix

3.2

only

Molluscs

Subulito

idea

AF

G11

16

Devonia

n-

Triassic

MM

YY

35

O'K

eefe

2004

Journ

al of

Pale

onto

logy 7

8:9

73-

988.

Oth

er

Vert

ebra

tes

Sauro

pte

rygia

AF

G/S

12

88

Jura

ssic

MM

YY

36

Popov e

t al.

1999

Pala

eonto

logy

42:6

25-6

61.

Bra

chio

pods

Atr

ypid

a (

early)

AF

S25

27

Ord

ovic

ian

ML

N

� 137

Au

tho

rR

efe

ren

ce

Gro

up

Do

main

Un

itN

Nch

ar

Geo

log

ic

Ran

ge

Realm

DU

RD

UR

-IN

ote

s

37

Roopnarine

2001-1

Journ

al of

Pale

onto

logy

75:6

44-6

57.

Molluscs

Chione

GS

16

20

Oligocene -

Recent

MM

YN

38

Roopnarine

2001-2

Journ

al of

Pale

onto

logy

75:6

44-6

57.

Molluscs

Puberella

GS

17

20

Oligocene -

Recent

MM

YN

39

Roopnarine

2001-3

Journ

al of

Pale

onto

logy

75:6

44-6

57.

Molluscs

Chione

GS

13

13

Oligocene -

Recent

MM

YN

[OL 3

7]

40

Roopnarine

2001-4

Journ

al of

Pale

onto

logy

75:6

44-6

57.

Molluscs

Puberella

GS

15

19

Oligocene -

Recent

MM

YN

[OL 3

8]

41

Schneid

er

1995

Zoolo

gic

a S

cripta

24:3

21-3

46.

Molluscs

Card

iidae

FG

/SG

32

16

Triassic

-

Recent

MM

YY

42

Sm

ith 1

988

Pale

onto

logy

31:7

99-8

28.

Echin

oderm

sEarly

AF

G29

32

Ord

ovic

ian -

Carb

onifero

u

s

MM

YN

43

Sm

ith a

nd

Arb

izu 1

987

Leth

aia

20:4

9-6

2.

Echin

oderm

sAgela

crinitin

ae

SB

G13

12

Ord

ovic

ian -

Carb

onifero

u

s

MS

N

44

Sm

ith e

t al.

1995

Zoolo

gic

al Jo

urn

al

of th

e L

innean

Socie

ty 1

14:2

13-

243

Echin

oderm

sO

phiu

roid

sAF

SF

28

41

(Perm

ian)

Triassic

-

Recent

MM

YY

45

Sm

ith a

nd

Wright

1993

Monogra

ph o

f th

e

Pala

eonto

logra

phic

a

l Socie

ty 5

93:1

99 -

267.

Echin

oderm

sAF

G14

29

Jura

ssic

-

Recent

MM

YY

46

Tin

n &

Meid

la

2004

Pala

eonto

logy

47:1

99-2

21.

Ostr

acodes

Beyrichio

copa

AF

S35

39

Early t

o

Mid

dle

Ord

ovic

ian

MS

N

47

Verm

eij &

Carlson 2

000

Pale

obio

logy 2

6:1

9-

46.

Molluscs

Rapanin

ae

SB

G/S

36

34

Eocene-

Recent

MM

YN

Table

V.1

(con't)

� 138

Au

tho

rR

efe

ren

ce

Gro

up

Do

main

Un

itN

Nch

ar

Geo

log

ic

Ran

ge

Realm

DU

RD

UR

-IN

ote

s

48

Wagner

1999

Am

erican

Mala

colo

gic

al

Bulletin 1

5:1

-31.

Molluscs

Lophospiroid

aAF

S82

91

Cassin

ian -

Pridoli

MS

N

49

Wagner

1997

Pale

obio

logy 2

3:1

15-

150.

Molluscs

Rostr

ochoncha

AF

S154

126

Early

Cam

brian -

Capitania

n

MS

N[O

L 5

0-5

3]

50

Wagner

1997

Pale

obio

logy 2

3:1

15-

150.(

Taxon g

roup

3)

Molluscs

Rib

eriid

ae

FS

27

46

Early

Cam

brian -

Upper

Cara

doc

MS

N

51

Wagner

1997

Pale

obio

logy 2

3:1

15-

150.(

Taxon g

roup

4)

Molluscs

Technophoridae

FS

17

62

Pre

sbachia

n -

Ashgill

MS

N

52

Wagner

1997

Pale

obio

logy 2

3:1

15-

150.(

Taxon g

roup

7)

Molluscs

Bra

nsoniidae

FS

22

50

Upper

Are

nig

-

Upper

Cara

doc

MS

N

53

Wagner

1997

Pale

obio

logy 2

3:1

15-

150.(

Taxon g

roup

8)

Molluscs

Hip

pocard

iidae

FS

39

68

Lla

nvirn -

Serp

ukhovia

nM

SN

54

Wagner

Coiled

webpage

Molluscs

Pale

ozoic

gastr

opods

AF

S481

217

Early

Cam

brian -

Giv

etian

MM

Y/S

N[O

L 5

5-6

1]

55

Wagner

webpage

Molluscs

Euom

phalo

ids

AF

S67

146

Early

Tre

madoc -

Eifelian

MM

YN

56

Wagner

webpage

Molluscs

Ple

uro

tom

arids

AF

S202

167

Early

Tre

madoc -

Eifelian

MM

YN

57

Wagner

webpage

Molluscs

Tro

choid

sAF

S13

85

Lla

nvirn -

Early L

udlo

wM

MY

N

58

Wagner

webpage

Molluscs

Murc

his

onoid

sAF

S66

107

Early A

renig

-

Eifelian

MM

YN

59

Wagner

webpage

Molluscs

Mic

rodom

ato

idAF

S12

61

Early

Cara

doc -

Late

Ludlo

w

MS

N

60

Wagner

webpage

Molluscs

Tro

chonem

ato

idAF

S15

57

Early

Cara

doc -

Late

Ludlo

w

MM

YN

61

Wagner

webpage

Molluscs

Maclu

rito

ids

AF

S15

67

Mid

-

Cam

brian -

Ashgill

MM

YN

Table

V.1

(con't)

� 139

Au

tho

rR

efe

ren

ce

Gro

up

Do

main

Un

itN

Nch

ar

Geo

log

ic

Ran

ge

Realm

DU

RD

UR

-IN

ote

s

62

Yate

s &

Warr

ens 2

002

Zoolo

gic

al Jo

urn

al

of th

e L

innean

Socie

ty 1

28:7

7-1

21

Oth

er

Vert

ebra

tes

Tem

nospondyli

AF

G37

60

Carb

onifero

u

s-J

ura

ssic

TM

YY

fam

ilie

s

dele

ted

63

Lio

w(t

his

stu

dy)

Ostr

acodes

Curfsina

GS

29

7M

id A

lbia

n -

Thanetian

MM

YN

64

Lio

w(t

his

stu

dy)

Ostr

acodes

Opimocythere

GS

17

16

Upper

Alb

ian -

Mid

Mio

cene

MM

YN

65

Lio

w(t

his

stu

dy)

Ostr

acodes

Schizoptocyther

eG

S16

9Low

er

Santo

nia

n -

Mid

Mio

cene

MM

YN

66

Lio

w(t

his

stu

dy)

Ostr

acodes

Phalcocythere

GS

30

8

Upper

Maestr

ichtian

- O

ligocene

MM

YN

Table

V.1

(con't)

� 140

extant representation to increase the sample size for this study (Table V.1). I started

with a database of morphological character matrices used in cladistic analyses as

assembled by Wagner (2000). I supplemented Wagner’s collection by systematically

searching through the journals Lethaia, Historical Biology, Journal of Paleontology,

Paleobiology and Systematic Biology (1996 - 2005) for other publications reporting

morphological character matrices that meet the above criteria. Additions were also

made from relevant references cited in the retained papers. Updates of phylogenetic

hypotheses are occasionally made, but only the most recent paper by the same authors

discussing the same taxa is included here to avoid duplication. Lastly, I also included

new species character matrices that I coded from four extinct ostracode genera, namely

Curfsina Deroo 1966, Opimocythere Hazel 1968, Phalcocythere Siddiqui 1971 and

Schizoptocythere Siddiqui and Al-Furaih 1980 for this study (See Appendices K, L, M)

for character matrices, stratigraphic ranges, character descriptions and references).

Some large datasets were sub-divided into lower taxonomic groups identified by authors

of the data (Table V.1). These are assumed to be at least paraphyletic if not

monophyletic and are analyzed in their own right. Morphological characters that

became non-informative (i.e. all the taxa in an analysis have the same character state) as

a result of partitioning of data or removal of taxa were discarded from subsequent

analyses.

� 141

Data treatment

Stratigraphic or geologic ranges are explicitly equated to lineage longevity. Henceforth

I use these terms interchangeably. It is explicitly assumed that each study represents

closely related taxa that have similar preservation potentials such that even though

stratigraphic ranges are underestimates of true longevity, the rank order of the ranges

should quite accurately reflect the rank order of the true longevities.

The data treatment here is similar to two previous analyses of lineage longevity versus

morphological distributions (Liow 2004, 2006), but with two crucial improvements.

The first is that stratigraphically long-ranging lineages are dynamically defined groups

rather than a fixed subset of the dataset in question. The second is that long-ranging

lineages are compared with short-ranging ones as groups as well as individually (see

below and Fig. V.1 for details).

I calculate morphological distance as the sum of the distance of each character of each

taxon in a given dataset from the average of the entire dataset. I calculate this distance

both un-weighted and weighted (such that each character contributed equally to the total

distance of a taxon from the group average). Each average character is calculated as

either the modal character state or the mean character state. The latter is reasonable for

binary and ordered multi-state characters, but not as appropriate for unordered multi-

state characters. Hence unordered multi-state characters are converted into binary� 142

Taxon Duration

Dis

tanc

e f

rom

av e

r ag

e

G

E

F A

B

C

A

B

C

A

B

C

A

B

C

A

B

C

A

B

C

A

B

C

Group Analyses Single Analyses

AFig. V.1

B.1 C.1

B.2

B.3

C.2

C.3

Stratigraphic range / lineage longevity

Stratigraphic range / lineage longevity

Dist

ance

fro m

the

clad

e av

era g

e

Dist

ance

f rom

t he

clad

e a v

erag

e

Fig. V.1. Hypothetical plot of morphological distance versus stratigraphic ranges. Panel A plots the distance of each taxon (black circles) from the empirical clade average plotted versus its stratigraphic duration/longevity. The plots in panels B and C are replicas of panel A. Panel B illustrates a group analysis where sequentially larger groups of taxa (A, A+B, A+B+C etc.) with longer durations are compared with rarified samples of the remaining taxa of shorter longevities to the left of the circled taxa (Fig. V.1.B.1 through B.3). Panel C illustrates a single analysis where individual taxa (A, then B, then C, etc.) with a randomly picked taxon from the remaining pool of taxa (excluding taxa to the right of the plot) with shorter longevities (Fig. V.1.C.1 through C.3). G, E, and F in panel A are taxa having the same durations being combined in comparing their probabilities of being more or less distant from the average morphology in subsequent analyses for both group and single analyses.

� 143

characters by coding the modal character as state zero and all other characters as state

one. Numerical or ordered multi-state characters that number six or more are log-

transformed (new value = ln (old value+2)) to so that they will not dominate the

calculation of morphological distances in unweighted treatment.

I categorize stratigraphic ranges of taxa reported in the literature as three types (Table

V.1). First and most commonly, internationally recognized time intervals were reported

with or without numerical values in millions of years (M.y., Table V.1). The latter were

converted to numerical values of the midpoints of the stage names synonomized in

Harland et al. (1990) in millions of years. This source was used as a reference as

opposed to newer time scales for the convenience of checking regional stratigraphic

names, but the differences in numerical values of stages should not affect the rank-order

of taxon durations. Second, certain time intervals not conforming to internationally

recognized names were reported and in these cases, I simply assigned sequential

numbers to those intervals sequentially or used the authors’ numbering of geologic

stages (S, Table V.1). Lastly, in publications where stratigraphic ranges were illustrated

to approximate scale but where geologic ranges were not explicitly reported, I manually

measured the illustrated lengths of the durations and tabulated those for use in analyses

(L, Table V.1). In some cases, more than one of these types are represented and I

analyze data with all available types of reported durations to check for possible

differences in results. � 144

Ghost ranges and inferred range extensions based on phylogenetic inference are also

reported in some studies (e.g. Bloch et al. 2001, Bodenbender & Fisher 2001). Where

such inferences were available, I reanalyzed these data with the inferred durations,

again to check for possible differences in results.

In group analyses, long-ranging taxa (= taxa with greater longevities) are increasingly

inclusive sequential groups (see Fig. V.1 panel B), i.e. first a long-ranging taxon is

simply the longest-ranging taxon (taxon A in Fig. V.1), then the two longest-ranging

taxa (taxa A and B in Fig. V.1), then the three longest ranging taxa (taxa A, B and C in

Fig. V.1) and so on, until half the taxa have been included in the long-ranging group.

Then I compare the mean morphological distance of each long-ranging group with that

of an equivalent rarified number of randomly selected short-ranging taxa, with

replacement. This rarefaction is done because sample sizes are different for longer-

ranging and shorter-ranging groups of taxa (see Liow 2004, 2006 for more details). I

perform this rarefaction 500 times and tabulate frequency with which each long-ranging

group has a mean morphological distance smaller than that of the randomly selected

shorter-ranging group. This is simply the probability with which a long-ranging group

of taxa is less distant from the average morphology (calculated using both means and

modes, see above). Since this probability is to be calculated for the entire dataset, I

reverse the above-described operation for short-ranging taxa and calculate the

probability a group of short-ranging taxa is more distant from the average morphology � 145

and then attach these values to those previously calculated. I then plot this set of

probabilities, p(g), versus taxon longevities (= stratigraphic ranges) and calculate rank

order correlations (Kendall’s tau) to test if more and more inclusive long-ranging taxa

are morphologically less distant from the average than expected by chance. These rank

order correlations and their probabilities are reported in the Appendix O.

In order to account for the possibility that each individual long-ranging taxon may be

morphologically less distant from the average than expected from their short-ranging

relatives, I performed what I call single analyses. As before, I sequentially defined

long-ranging taxa (panel C in Fig. V.1), but did not calculate average morphological

distances (i.e. I first consider taxon A, then taxon B, then taxon C in Fig. V.1

individually and so on until all the taxa have been treated). I compared each long-

ranging taxon to a randomly selected member of the remaining short-ranging pool with

replacement, 500 times, and tabulated the probability with which this long-ranging

taxon is less distant from the average morphology (calculated using both means and

modes, as above). Again, since this probability has to be calculated for the entire

dataset, I reverse this operation for short-ranging taxa then attach the corresponding

values to those previously calculated. I again plotted this probability, p(s), versus

longevities and calculate rank order correlations (Kendall’s tau) to test if each long-

ranging taxa are morphologically less distant from the average than expected by chance.

These rank order correlations and their probabilities are reported in the Appendix O.� 146

In both group and single analyses, I repeat the rarefaction exercise described above but

replace morphological distances with principal component scores obtained from

Principal Component Analysis of the distance matrix obtained using the character

matrices (= Principal Coordinate Analyses, PCO, Gower 1966). This exercise removes

possible redundancy in the original morphological data. The number of scores used is

adjusted to explain about 80% of the variance and varies from five to twenty, depending

on the size of the data matrix. This yields for group and single analyses p(g, pco) and

p(s, pco) respectively, probabilities that longer-ranging taxa are less distant from the

average morphology. I repeat the plotting of these versus taxon longevities and calculate

rank order correlations (Kendall’s tau) as above. These rank order correlations and their

probabilities are reported in the Appendix O.

In addition, because taxa of the same calculated longevity may have different

morphological distances (e.g. taxa E, F, and G in Fig. V.1), I also calculated rank order

correlations for median morphological distances of taxa having the same calculated

longevity, with respect to their common longevity. I do this for both group and single

analyses and assign the abbreviations p(g, m), p(g, m, pco) and p(s, m), p(s, m, pco) to

the resulting probabilities respectively and calculate their rank-order correlations with

respect to longevities. These rank order correlations and their probabilities are reported

in the Appendix O.

� 147

Since there are multiple ways of quantifying stratigraphic ranges and morphological

distance from an average, (i.e. using the mode or mean character states as the average,

using distances or principal coordinate scores, using original and inferred longevities,

using stratigraphic ranges measured in different ways), the significant tests of trends in

the relationship between longevity and morphology may differ within a dataset. I

present all the results obtained (see Appendix O) using these various possibilities but

summarize whether the taxa in a given study show a positive, negative or non-

significant relationship between lineage longevity and morphological distance using the

following criteria. If there is only one significant result, the dataset is assumed to

demonstrate no significant relationship with regards to longevity-morphology

distribution. If significant results are in a ratio of one to one, the relationship is taken to

be non-significant, but if significant results are in a ratio of more than one to one, the

sign of the more commonly represented sign is accepted. For instance, if in a given

dataset, there are three significantly negative values and only one significantly positive

value, this dataset is taken to show a negative relationship between morphological

distance and longevity. In addition, I consider the possibility that any conflict of the

signs of correlation indicates a non-significant situation and refer to this as a

conservative solution (see results for details). Because significant cases are sometimes

already removed in this method of summary, Bonferroni corrections that further over-

correct for significance are not used.

� 148

In order to account for the differences in taxonomic hierarchical representation in

various studies and to test if this affects conclusions drawn regarding the distribution of

morphological distances versus longevities, I tabulated whether the taxa of the lowest

Linnean ranks whose morphologies are coded are families, subfamilies, genera,

subgenera or species. Similarly, I tabulated whether the domain of the study in question

involved a taxonomic unit greater than a family, a family, a subfamily or genus. I gave

the various taxonomic ranks values where 0 = species, 1 = subgenus, 2=genus, 3 =

subfamily, 4 = family and 5 = above family. I then calculated the taxonomic

inclusiveness of the study as the value of the domain minus the taxonomic value of the

coded taxa. For example, Jeffery & Emlet 2003 studied temnopleurid echinoids (domain

value = 5) and coded the morphology of species (coded taxa value = 1), hence the

taxonomic inclusiveness is 4. Where there is a mixture of units coded, I use the average

value, e.g. if both genera and subgenera were coded, then the coded taxa value = 1.5. I

also tabulated the number of characters, the number of taxa, and whether the clade in

question is from the aquatic or terrestrial realm. Additionally, I calculated mean and

median longevities for taxa that have stratigraphic ranges reported in millions of years,

the standard deviation of the longevities, alpha (= square of the mean divided by

standard deviation of longevities) which is a shape parameter, and beta (= standard

deviation divided by square root of alpha) which is a scale parameter and skew (= two

divided by square root of alpha) (Wackerly et al. 2002). The latter three are descriptors

� 149

of gamma distributions, which I assume are approximated by the longevity distributions

of the datasets.

For completeness, I also reanalyzed the data used in Liow 2004 and 2006 to compare

results using this newly developed continuous method of comparing morphological

distance versus longevities, considering taxa individually and as groups as described in

previous sections.

Results

Out of the 66 datasets (Table V.1) that are retained for use in the analyses, 38 were used

in Wagner (2000). The others were from other sources (N = 10), including Wagner’s

own matrices of Paleozoic gastropods (N = 14), and four are datasets coded for this

study. Twenty-six of these represent data from the Paleozoic, four from the Mesozoic

and 20 from the Cenozoic, including both the Paleozoic and the Mesozoic (N = 4) and

lastly the Mesozoic plus the Cenozoic (N = 12). The datasets consist of studies of

mammals (N = 7), other vertebrates (N =7), trilobites (N = 4), other arthropods (N = 7),

mollusks (N = 27), echinoderms (N = 8), brachiopods (N =5) and cnidarians (N =1).

They thus represent fossilizable animals broadly across the Phanerozoic.

I calculated both weighted and unweighted morphological distances of taxa from the

average of their entire dataset. They do not offer different results so henceforth, for � 150

economy and clarity, I discuss only the results using weighted morphological distances.

Using either an average morphology calculated as a modal or mean value resulted in

only a few distinguishable different values of probabilities of morphological distance

varying with morphologies, within the datasets (see Appendix O). The use of different

methods of quantifying stratigraphic ranges (using millions of years, the number of

stages or direct measurements from published range charts) also did not consistently

result in qualitatively different probabilities for the same datasets, although sometimes

using inferred longevities instead of raw stratigraphic ranges gave different qualitative

results (Appendix O). Since the method of summarizing the relationship between

morphology and longevity is used, the less frequently occurring sign of correlations are

weeded out of the results.

I found that 50% of the 66 datasets have taxa whose morphological distances from the

average states are negatively correlated with their longevities, 21% positively correlated

and 29% show no significant relationship (Table V.2). Since some datasets were subsets

of others or have overlapping taxa (Table V.1), I re-tabulated the number of cases of

each of the above and found that the rank order of the number of studies of each case

was not altered (47%, 23% and 32% respectively, these values do no add to 100%

because of rounding errors). Thus, when taxa are widely sampled, all three of the

� 151

Table

V.2

. D

ata

sets

where

morp

holo

gic

al dis

tances a

re n

egatively

, positiv

ely

, or

not

corr

ela

ted w

ith

longevity.

Refe

rences lis

ted in T

able

1 a

re g

rouped a

ccord

ing t

o w

heth

er

longevitie

s a

re n

egatively

corr

ela

ted w

ith m

orp

holo

gic

al dis

tances (

NEG

) or

positiv

ely

so(P

OS)

in g

roup a

naly

ses.

NS r

epre

sent

non-s

ignific

ant

cases.

It

als

o lis

ts t

he s

am

e for

taxa indiv

idually c

onsid

ere

d w

ithin

cla

des.

In a

dditio

n,

mean a

nd m

edia

n d

ura

tions,

sta

ndard

devia

tion o

f dura

tions,

beta

, alp

ha a

nd s

kew

are

als

o lis

ted.

An a

ste

risk im

plies t

hat

conserv

atively

, th

e d

iagnosis

would

have b

een n

on s

ignific

ant.

See t

ext

for

more

deta

ils.

STUDY

GROUPSINGLEmean-durmed-dursddurbeta

alpha

skew

Adra

in a

nd E

dgecom

b 1

997

NEG

NEG

NA

NA

0.7

0.7

1.7

1.5

Allm

on (

Table

9)

NEG

NS

3.7

3.2

2.3

0.6

6.0

0.8

Anders

on a

nd R

oopnarine 2

003

NEG

NEG

28.2

14.6

27.4

1.0

29.1

0.4

Angie

lczky &

Kurk

in 2

003

NEG

NS

NA

NA

1.3

1.7

0.4

3.0

Blo

ch e

t al. 2

001

NEG

NS

NA

NA

0.6

1.1

0.5

2.8

Bro

chu1997

NEG

NEG

2.9

1.8

3.8

1.3

2.3

1.3

Bru

net-

Lecom

te&

Chaline 1

990

NEG

NEG

0.2

0.1

0.2

0.8

0.3

3.7

Cairns 2

001

NEG

NS

18.0

6.2

23.9

1.3

13.6

0.5

Curf

sin

a (

this

stu

dy)

NEG

NS

4.2

1.9

6.5

1.6

2.7

1.2

Dam

iani et

al. 2001

NEG

NS

2.9

0.0

4.9

1.7

1.7

1.5

Gahn a

nd K

am

mer

2002

NEG

NS

NA

NA

2.2

1.1

1.8

1.5

Jeffery

1998

NEG

NS

NA

NA

2.2

0.7

5.2

0.9

Leig

hto

n &

Maple

s 2

002

NEG

NS

NA

NA

0.7

1.2

0.6

2.7

Monks 2

002

NEG

NEG

1.5

0.0

2.3

1.6

0.9

2.1

Monks a

nd O

wens 1

999

NEG

NEG

2.3

0.0

3.1

1.4

1.7

1.6

Opim

ocyth

ere

(th

is s

tudy)

NEG

NEG

5.7

1.9

9.5

1.7

3.5

1.1

Phalc

ocyth

ere

(th

is s

tudy)

NEG

NS

3.1

0.6

5.4

1.7

1.8

1.5

Popov e

t al. 1

999

NEG

NS

NA

NA

0.5

0.5

1.8

1.5

Roopnarine 2

001-1

NEG

NS

1.0

0.0

1.5

1.4

0.7

2.3

Sm

ith e

t al. 1

995

NEG

NEG

60.7

37.2

69.8

1.1

52.8

0.3

� 152

Table

V.2

(con't)

STUDY


alpha

skew

Wagner

1997

NEG

NEG

NA

NA

0.7

2.3

0.1

5.7

Wagner

1999

NEG

NEG

NA

NA

49.9

1.1

39.4

0.3

Wagner

Euom

phalo

ids

NEG

NS

5.2

2.6

6.4

1.2

4.3

1.0

Wagner

Maclu

rito

ids

NEG

NEG

2.1

0.0

3.0

1.5

1.4

1.7

Wagner

Mic

rodom

ato

idN

EG

NEG

1.9

0.0

2.8

1.5

1.3

1.8

Wagner

Murc

his

onoid

sN

EG

NS

4.1

0.0

6.9

1.7

2.4

1.3

Wagner

Rib

eriid

ae

NEG

NEG

NA

NA

0.8

1.4

0.4

3.2

Wagner

Technophoridae

NEG

NEG

NA

NA

0.9

1.0

1.0

2.0

Wagner

Bra

nsoniidae

NEG

NS

NA

NA

0.7

3.0

0.1

7.3

Wagner

Hip

pocard

iidae

NEG

NEG

NA

NA

0.6

2.8

0.1

7.4

Wagner

Tro

choid

sN

EG

NS

6.2

6.1

10.6

1.7

3.6

1.1

Wagner

Tro

chonem

ato

idN

EG

NEG

2.9

0.0

4.1

1.4

2.1

1.4

Schneid

er

1995

NEG

*N

S88.1

95.0

54.0

0.6

143.8

0.2

Adrian a

nd W

estr

op 2

001

NS

NS

NA

NA

0.9

1.8

0.3

3.8

Alroy 1

995

NS

NS

1.6

1.0

0.9

0.6

2.7

1.2

Am

ati a

nd W

estr

op 2

004

NS

NS

1.9

0.9

2.1

1.1

1.7

1.5

Caro

n e

t al. 2004

NS

NS

18.7

10.0

26.8

1.4

13.0

0.6

Dashzeveg a

nd M

eng 1

998

NS

NS

1.9

0.0

3.1

1.6

1.2

1.8

Dew

ing 2

004

NS

NS

NA

NA

1.3

0.6

4.2

1.0

Ebbesta

d &

Budd 2

003

NS

NS

NA

NA

1.2

0.8

1.9

1.4

Fro

elich 2

002

NS

NEG

0.7

0.7

0.4

0.6

1.1

1.9

Gra

nde a

nd B

em

is 1

998

NS

PO

SN

AN

A4.7

1.9

1.3

1.7

� 153

Table

V.2

(con't)

STUDY


alpha

skew

Hopkin

s 2

004

NS

PO

S2.2

1.5

1.8

0.8

2.7

1.2

Jeffery

& E

mle

t 2003

NS

NS

6.3

4.0

4.4

0.7

9.0

0.7

Kara

saw

a a

nd k

ato

2003

NS

NS

13.6

0.0

17.0

1.3

10.9

0.6

Monks 1

999

NS

NS

0.5

0.0

1.5

2.8

0.2

4.5

Nutz

el et

al. 2

000 s

et-

2N

SN

S59.5

62.0

57.4

1.0

61.6

0.3

Schiz

opto

cyth

ere

(th

is s

tudy)

NS

NS

2.8

0.0

6.0

2.2

1.3

1.8

Sm

ith 1

988

NS

NS

3.2

0.0

10.1

3.2

1.0

2.0

Verm

eij &

Carlson 2

000

NS

NS

12.0

10.4

12.6

1.0

11.5

0.6

Wagner

Coiled A

llN

SPO

S3.9

0.0

6.7

1.7

2.2

1.3

Wagner

Ple

uro

tom

arids

NS

NS

4.7

0.0

7.5

1.6

3.0

1.2

Adnet

and C

apett

a 2

001

PO

SN

S47.0

36.3

38.7

0.8

57.1

0.3

Allm

on 1

996 (

Table

1)

PO

SN

EG

7.7

0.0

15.8

2.1

3.7

1.0

Alv

are

z et

al. 1

998

PO

SPO

S44.1

36.5

45.8

1.0

42.5

0.3

Bodenbender

and F

icher

2001

PO

SPO

SN

AN

A1.4

1.5

0.6

2.5

Fore

y 1

991

PO

SN

S7.4

0.0

17.6

2.4

3.1

1.1

Mic

haux 1

989

PO

SN

S5.3

4.2

5.3

0.0

5.2

0.9

O'K

eefe

2004

PO

SN

S1.3

0.0

2.6

2.0

0.7

2.5

Roopnarine 2

001-2

PO

SN

S1.0

0.0

1.5

1.4

0.7

2.3

Roopnarine 2

001-3

PO

SN

S1.2

0.0

1.6

1.3

1.0

2.0

Roopnarine 2

001-4

PO

SN

S0.9

0.0

1.3

1.4

0.6

2.5

Sm

ith a

nd A

rbiz

u 1

987

PO

SN

SN

AN

A4.8

1.4

2.5

1.3

Sm

ith a

nd W

right

1993

PO

SN

S29.6

18.0

33.7

1.1

26.0

0.4

Tin

n &

Meid

la 2

004

PO

SN

SN

AN

A0.7

0.9

0.8

2.2

Yate

s &

Warr

ens 2

002

PO

S*

PO

S0.8

0.2

0.9

1.1

0.7

2.4

� 154

described scenarios of morphological distribution versus longevity can be seen, with

more cases in which morphologies are negatively distributed with respect to longevities.

Even using a conservative approach where a dataset is considered non-significant when

the signs of the correlations disagree among treatments, there is still an excess of

datasets that are negatively correlated (48% cf 20 % for positively correlated ones and

32% for non-significant ones).

In group analyses, datasets showing negative, positive and non-significant

morphological distributions with respect to durations do not have significantly different

numbers of taxa represented nor numbers of coded morphological characters (t-test, p

>> 0.05).

Datasets demonstrating a positive morphological distribution longevity relationship

have marginally significantly greater taxonomic units coded (0.8) than either those

demonstrating a negative one (0.3, t-test, p = 0.09) or a non-significant one (0.3, t-test, p

= 0.05). There are no significant differences in either their domains nor taxonomic

inclusiveness (t-test, all p >> 0.05).

Datasets representing organisms from aquatic (marine and freshwater) or terrestrial

environments are not differentially represented in the three patterns of morphological

distribution longevity relationships (χ2 test, all p >> 0.05).� 155

In terms of taxonomic representation, datasets demonstrating a negative morphological

distribution longevity relationship are represented by three mammal studies out of a

total of 33, non-significant cases zero out of a total of 19 and clades having a positive

relationship, four out of a total of 14. There is a significant difference in terms of

distribution of mammal representation (χ2 test, p = 0.029). Other common clades

represented in the comparisons, molluscs and echinoderms, and all vertebrates

considered together, however, show no significant differences in frequencies among the

three morphological-longevity distribution patterns.

Whether or not the datasets represent only extinct or both extinct and extant organisms

is also not a factor in their distribution among the three patterns of morphological

distribution longevity relationships (χ2 test, all p >> 0.05).

Datasets showing negative, positive and non-significant morphological distributions

with respect to longevities do not have significantly different mean or median

longevities, or descriptors of the distribution, including alpha, the shape parameter and

beta the scale parameter and skew (t-tests, all cases p >> 0.05).

The former results are all for groups of increasingly inclusive long-ranging and short-

ranging taxa. In contrast, for comparisons of single taxa with the remaining longer-

ranging or shorter-ranging taxa, 64% of the cases demonstrated a non-significant � 156

relationship, i.e. long-ranging taxa individually not different from short-ranging taxa in

morphological distances from the clade average. In 27% of the cases, longer-ranging

taxa are individually more average morphologically than expected and in only 9% of

the cases were they morphologically more distant than expected. Even after

disregarding cases that may be non-independent, because they stem from studies using

overlapping taxa, the percentages do not change much (respectively 67%, 23% and

10%).

Datasets consisting only of extinct taxa have lineage longevities that are significantly

positively correlated with age of the lineages in 16 out of 47 cases (Kendall’s rank test,

p < 0.05). The 31 non-significant cases show positive correlation coefficients in all but

6 cases each with very small negative coefficients (data not shown). Datasets including

extant taxa have lineage longevities that are significantly positively correlated with age

of the lineages in 13 out of 19 cases (Kendall’s rank test, p < 0.05), with the remaining 6

being non-significant (data not shown).

Analyses using crinoid genus data (Liow 2004 from Foote 1999) corroborate the current

results. The crinoid orders where discrete groups of longer-ranging genera were found

to be morphologically less distant from an average than expected in the previous study

remain so in this study. Individual instances of longer-ranging taxa are sometimes also

significantly less distant than randomly selected short-ranging taxa (Appendix O).� 157

Group analyses using trachyleberidid ostracode data (Liow 2006) proved different in

some cases in comparison to the conclusions drawn previously (Appendix O). In

particular, contemporaneous genera and genera in cohorts originating in the same

geologic stage or time interval that were previously thought to show a positive

relationship between morphological distance and longevity show a negative one with

this new analysis using a dynamic definition of long-ranging forms (Appendix O).

Discussion

Much has been written about stasis at the species level from points of views ranging

from paleontology to genetics and development (van Valen 1982, Wake et al. 1983,

Rutherford 2000, Merilä et al. 2001, Schwenk & Wagner 2001, Belade et al. 2002,

Eldredge et al. 2005, Grether 2005). However, the mechanisms of maintenance of

within-lineage phenotypic stability do not inform us, at least not directly, on the patterns

of distributions of lineage longevity at the level above the species. Qualitative

descriptions of why and how geologically very persistent taxa, sometimes called “living

fossils,” may have persisted unchanged when their relatives did not, are much more

common in the literature than quantitative studies (see Eldgredge & Stanley 1984,

Fisher 1990). Quantitative comparisons are not straightfoward because of the small

sample sizes of lineages that have exceptional longevities, compared with closely

related lineages. Using extant lineages may increase the relative sample size of lineages

with great longevities but introduce the problem of one-sided range truncation. In � 158

addition, the higher taxon encompassing the purported “living fossil” and its relatives

are often arbitrarily defined and frequently not explicit.

The present study is a continued attempt to quantify the relationship between

morphology and longevity in a rigorous framework. It overcomes taxonomic and

sampling limitations of two previous attempts (Liow 2004, 2006) to investigate the

morphological distribution of longer-ranging versus shorter-ranging lineages,

employing the newly developed sequential rarefaction to further alleviate the problem

of small numbers of persistent taxa. It also considers the novel approach of treating

lineages as individually long-ranging or persistent as groups.

As mentioned at the start of this paper, lineages with great longevities could imaginably

be either morphologically more distant from the average of their inclusive clade, or less

distant than expected. The first scenario may indicate that being different confers a

competitive edge, particularly in a situation of co-occurrence, and the latter scenario

that being average confers flexibility and generality (but see later section on

phylogenetic implications). What then do we observe from studying a large suite of

independently collected data representing diverse clades? When lineages with greater

longevities are collectively considered in group analyses, some datasets show a trend

whereby lineages with increasingly greater longevities are less distant from the average

morphology while others show no trend and yet others have the opposite trend. � 159

However, a greater number of cases display trends of a negative correlation between

longevity and morphological distance, including the genera of crinoid orders reanalyzed

using the data from Foote (1999) in Liow (2004), as well as the genera of a large family

of ostracodes (Liow 2006). In contrast, when these long-ranging lineages are

individually considered in single analyses, fewer datasets show any significant trends in

the morphological distance of these taxa from the average morphology.

The average taxonomic unit of datasets showing positive correlations between

morphological distance and longevities is larger than those showing negative

correlations or no correlation, i.e. genera or families are coded rather than species or

subgenera. This may suggest that when taxa of a higher rank in a Linnean taxonomic

hierarchy are compared (e.g. genera or families), there is a greater likelihood that the

ones that persist for longer periods of time are more divergent from an average

morphology than expected by chance. This may indicate that successful new

morphologies could invade new ecological niches and persist for longer periods of time,

possibly with decreased competition.

Conversely, it also suggests that when taxa of a lower rank in a Linnean taxonomic

hierarchy are compared, (e.g. species or subgenera), there is an increased possibility that

those persisting for greater periods of time may benefit from not being too different

from morphological forms that have already proved effective for their relatives. � 160

Alternatively, this may also reflect some form of genetic compensation, albeit at a

higher taxonomic level (see Grether 2005).

The remaining variables describing the datasets were not differentially distributed

among the datasets showing the three different patterns of morphology versus longevity

distribution, with the exception that more mammals datasets are represented in the

positive case. Whether this is related to the probability of sampling mammals in the

fossil record is unclear. The three groups of datasets representing the different

morphological distribution longevity distribution scenarios have similar shapes of

longevity distributions, numbers of taxa and characters represented and ecological

realms and phylogenetic representation, the last two variables indirectly reflecting

preservation potential.

Interestingly, when longer-ranging taxa are considered singly, a sweeping majority of

the datasets, including those reanalyzed from Liow 2004 and 2006 (see Appendix O)

shows no significant trend. Individually considered, longer-ranging taxa are often not

significantly more or less distant from morphological mean than short-ranging taxa

(although the power of this test is lower than for group analyses). Group properties of

taxa of various longevities are hence much stronger than properties of individually

considered taxa.

� 161

Phylogenetic implications

This study has not involved any phylogenetic framework, even though the datasets used

are good clades that are at least paraphyletic if not monophyletic. These results have,

however, some phylogenetic implications. Lineages with greater longevities are

significantly older or occur earlier in clade history, even when only extinct clades are

examined such that age is not a constraint on observed longevity. Since ancestors can

be found in the fossil record with quite a high probability (Foote 1996), some lineages

with great longevities must be ancestors to other lineages included in the datasets.

Since lineages with great longevities frequently have average, and more average than

expected morphologies, it follows that many ancestral lineages must give rise to many

descendents that are morphologically similar for us to observe this pattern of

morphological dispersion. This in part corroborates Wagner & Erwin’s finding that

lineages that persist for long periods of time give rise to more descendants (1995).

Here, I note that this may be a taxonomic rank dependent argument, since we observe

here that when taxa of higher taxonomic ranks are compared, more datasets show a

pattern where morphological distance is positively correlated with lineage longevity.

Biases and sources of error

The only way to widely investigate the relationship between morphology and longevity

is to sample the published literature. However, there are a number of biases and sources

of error due to the usage of heterogeneous data collected for other purposes. Some have � 162

already been briefly mentioned but I repeat them here to remind readers of the

limitations. Firstly, empirical values of average morphologies are used as a reference

but these are calculated from incompletely sampled datasets (e.g. when a family is

investigated, not all genera are represented) and hence may not represent the true

average morphology of the clade in question. The completeness of studies could not be

ascertained in a straightforward manner from the publications and hence were not

analyzed in the current study. Second, the taxa analyzed may not actually belong in a

natural group due to incomplete knowledge of the phylogenetic systematics of the

clade. Third, the taxonomic ranks of the taxa analyzed in each clade may not be

equivalent. Fourth, the relative stratigraphic ranges may not reflect the true ranks of the

longevities, especially when there is one-sided range truncations of extant taxa

involved. However, since all the studies involve related organisms, preservational

potentials should not be dramatically different within datasets. Fifth, the characters

coded may not adequately represent the whole-organism morphology of the taxa (again,

the few cases in which only a limited part of the morphology of the taxa are coded are

noted in Table V.1). Moreover, in all the datasets used, only adult morphology is

reflected.

� 163

Conclusions

Many factors can influence the survivorship of individuals, populations and lineages

during their lifetimes. These factors may interact in a complex fashion so that it is

difficult to tease apart their individual contributions in holding a taxon in stasis or

causing its cladogenesis or extinction. Despite this complexity, some factors have been

demonstrated to be important contributors to survivorship (Jablonski 2005) and others

are beginning to be investigated as potential properties that may confer longevity or the

lack thereof to lineages. In particular, morphological distribution is at least sometimes

related to taxon longevity. The relationship can be complicated by taxonomic ranks,

which previously have not explicitly taken into consideration in discussions of

persistent lineages. Contrary to the common idea that very long-ranging lineages are

special or unique in some significant way, it appears that they often tend to be more

average than expected by chance alone in comparison to their relatives. This suggests

that deviations from locally optimal solutions that evolutionary processes have already

been found, are usually not good candidates for longer-term survival.

� 164

CHAPTER VI

LINEAGE PERSISTENCE - A THEORETICAL FRAMEWORK AND

EMPIRICAL RESEARCH PROGRAM

“Like the centenarians of our society, each living fossil has its own story to tell.”

--- P.D. Ward 1992

“......but every bell curve has left tail......basic explanation of “living fossils”... neither

mysteriously optimal nor unfortunately devoid of variability”

--- S.J. Gould 2002

A hundred and fifty years after Darwin’s seminal volume, the problem of “living

fossils” is still discussed in the scientific literature (Gould 2002), albeit infrequently,

perhaps because not much progress has been made. I maintain my suggestion that

“living fossils” is a misconceived concept. Instead of singling out taxa that seem odd to

us, entire clades should be examined for variation in entire distributions of taxon

longevities, speciation and extinction rates, as well as the degree of morphological

isolation of the relevant taxa as a part of an inclusive phylogenetic framework. The tail-

end members of these distributions may then be examined for properties that could have

promoted their position in the distribution, which can then be analysed in the context of

the entire clade. It is also important to compare equivalent taxonomic units, as opposed

to trying to find similarities between Onychophorans and Ginkgo biloba.� 165

Using a comparative approach, I have shown in chapters of this thesis, that very

persistent genera of crinoids and ostracodes are in general not morphologically more

distant from a clade average than expected by chance. Going beyond these two

relatively exhaustive datasets, a comparative of datasets representing many different

higher taxa with varying preservation potentials, completeness of sampling, ecologies

and time intervals showed basically the same results.

The differences between taxa with increased long-term survivorship and geologic

longevities seem to lie mainly in their ecological versatility as measured by the width

of their geographic ranges and, to a lesser extent their bathymetric range, as well as

their propensity to give rise to descendants (i.e. species for genera and subspecies and

varied morphological forms for species), rather than morphological distance or

deviation.

There are many aspects of this topic that still require study. We still lack studies where

detailed data show that purported long-ranging lineages really maintain the same

identity throughout their taxonomic duration. We do not know how rates, durations and

isolation (see Chapter I) interact. In particular, little is known about how

morphological or phylogenetic isolation comes into being and is maintained and what

such isolation may imply for evolution in general.� 166

It may be true that each species or lineage that we have historically called a living fossil

has its own story to tell (Ward 1992) but so does every other species that ever lived on

this earth. However, to discover general patterns and pervasive processes that shape the

living earth, a comparative approach is essential and complementary to the insufficient

approach of viewing each species as a unique event. I conclude by reiterating Raup et

al.’s (1973) observation that it is important to use both “idiographic” and “nomothetic”

approaches in macroevolution. Both large-scale patterns and the particularities of

specific taxa can teach us about how they persist unchanged and could consequently

also inform us on how biological entities evolve.

� 167

LITERATURE CITED

Ackermann, H. W., S. Yoshino & S. Ogata. 1995. A bacillus phage that is a living fossil.

Canadian Journal of Microbiology 41:294-297.

Alexander, R. R. 1977. Generic longevity of articulate brachiopods in relation to the

mode of stabilization on the substrate. Palaeogeography, Palaeoclimatology,

Palaeoecology 21:209-226.

Ameziane, N. & M. Roux. 1997. Biodiversity and historical biogeography of stalked

crinoids (Echinodermata) in the deep sea. Biodiversity and Conservation 6:1557

-1570.

Anstey, R. L. 1978. Taxonomic survivorship and morphologic complexity in Paleozoic

Bryozoan genera. Paleobiology 4:407-418.

Arnold, S. J. 1992. Constraints on phenotypic evolution. American Naturalist 140:85

-107.

Athersuch, J. 1994. The biostratigraphic significance of Cretaceous ostracods from the

Arabian Gulf. In Micropalaeontology and Hydrocarbon Exploration in the

Middle East. (M. D. Simmons ed) pp 252-272. Chapman and Hall, London.

Ausich, W. I. 1998. Early phylogeny and subclass division of the Crinoidea (Phylum

Echinodermata). Journal of Paleontology 72:499-510.

� 168

Ausich, W. I., S. K. Donovan, H. Hess & M. J. Simms. 1999. Fossil occurrence. In

Fossil Crinoids. (H. Hess, W. I. Ausich, C. E. Brett & M. J. Simms ed) pp 41-49.

Cambridge University Press.

Ausich, W. I. & T. W. Kammer. 2001. The study of crinoids during the 20th Century

and the challenges of the 21st Century. Journal of Paleontology 75:1161-1173.

Avise, J. C., W. A. Nelson & H. Sugita. 1994. A speciational history of living fossils -

Molecular evolutionary patterns in horseshoe crabs. Evolution 48:1986-2001.

Bachnou, A., G. Carbonnel & B. Baoub. 1999. Hemicytherinae (Ostracods)

morphometrics by outline mathematical modeling: Systematic and phylogenetic

applications. Comptes rendus de l’academie des sciences-Fascicule A-sciences

de la terre et des planetes 328:197-202.

Bachnou, A., G. Carbonnel & B. Baoub. 2000. Le contour latéral des genres

d'ostracodes: une equation. Application au sein des Trachyleberidinae. Geobios

33:377-386.

Banerjee, A. & G. Boyajian. 1996. Changing biologic selectivity of extinction in the

Foraminifera over the past 150 My. Geology 24:607-610.

Baumiller, T. K. 1993. Survivorship analysis of Paleozoic Crinoidea - effect of filter

morphology on evolutionary rates. Paleobiology 19:304-321.

Baumiller, T. K. & H. Hagdorn. 1995. Taphonomy as a guide to functional-morphology

of Holocrinus, the first post-Paleozoic crinoid. Lethaia 28:221-228.

� 169

Bean, K., G. P. Jones & M. J. Caley. 2002. Relationships among distribution,

abundance and microhabitat specialization in a guild of coral reef triggerfish

(family Balistidae). Marine Ecology Progress Series 233:263-272.

Beldade, P., K. Koops & P. M. Brakefield. 2002. Developmental constraints versus

flexibility in morphological evolution. Nature 416:844-847.

Benson, R. H. 1966. Recent marine podocopid ostracodes. Oceanography and Marine

Biology 4:213-232.

Benson, R. H. 1973. An ostracodal view of the Messinian Salinity Crisis. Koninklijke

Nederlandse Akademie van Wettenschappen – Amsterdam. 1973:235-242.

Benton, M. J. (ed) 1993. The Fossil Record 2. Chapman & Hall, London.

Biernat, G. & C. C. Emig. 1993. Anatomical distinctions of the Mesozoic lingulide

brachiopods. Acta Palaeontologica Polonica 38:1-20.

Blackburn, T. M., K. E. Jones, P. Cassey & N. Losin. 2004. The influence of spatial

resolution on macroecological patterns of range size variation: a case study

using parrots (Aves : Psittaciformes) of the world. Journal of Biogeography 31:

285-293.

Bloch, J. I., D. C. Fisher, K. D. Rose & P. D. Gingerich. 2001. Stratocladistic analysis of

Paleocene Carpolestidae (Mammalia, Plesiadapiformes) with description of a

new late Tiffanian genus. Journal of Vertebrate Paleontology 21:119-131.

Bodenbender, B. E. & D. C. Fisher. 2001. Stratocladistic analysis of blastoid phylogeny.

Journal of Paleontology 75:351-369.� 170

Boomer, I., D. J. Horne & I. J. Slipper. 2003. The use of ostracods in

palaeoenvironmental studies, or what can you do with an Ostracod Shell? In

Bridging the Gap: Trends in the Ostracode Biological and Geological Studies.

(L. E. Park & A. J. Smith eds) pp 153-173. The Palaeontological Society.

Boucot, A. J. 1975a. Evolution and Extinction Rate Controls. Elsevier Scientific

Publishing Company.

Boucot, A. J. 1975b. Standing diversity of fossil groups in successive intervals of

geologic time viewed in the light of changing levels of provincialism. Journal of

Paleontology 49:1105-1111.

Boucot, A. J. 1977. Rates of evolution and number of descriptive terms. Journal of


Bown, P. 2005. Selective calcareous nannoplankton survivorship at the Cretaceous

-Tertiary boundary. Geology 33:653-656.

Boyajian, G. & T. Lutz. 1992. Evolution of biological complexity and its relation to

taxonomic longevity in the Ammonoidea. Geology 20:983-986.

Brouwers, E. M. 1993. Systematic Paleontology of Quarternary Ostracode Assemblages

from the Gulf of Alaska, Part 2: Families Trachyleberididae, Hemicytheridae,

Loxochonchidae, Paracyherideidae. U.S. Geological Survery Professional Paper,

1531.

� 171

Brown, J. H., G. C. Stevens D. M. Kaufmann. 1996. The geographic range: size, shape,

boundaries and internal structure. Annual Review of Ecology, Evolution and

Systematics 27:597-623.

Buckley, G. A., C. A. Brochu, D. W. Krause & D. Pol. 2000. A pug-nosed

crocodyliform from the Late Cretaceous of Madagascar. Nature 405:941-944.

Buzas, M. A. & J. J. Culver. 1984. Species duration and evolution: benthic Foraminifera

on Atlantic continental margin of North America. Science 225:829-830.

Cardillo, M., G. M. Mace, K. E. Jones, J. Bielby, O. R. P. Bininda-Emonds, W. Sechrest,

C. D. L. Orme & A. Purvis. 2005. Multiple causes of high extinction risk in

large mammal species. Science 309:1239-1241.

Chen, Z.-Q., K. Kaihob & A. D. George. 2005. Survival strategies of brachiopod faunas

from the end-Permian mass extinction. Palaeogeography, Palaeoclimatology,


Colborn, J., R. E. Crabtree, J. B. Shakee, E. Pfeiler & B. W. Bowern. 2001. The

evolutionary enigma of bonefishes (Abula spp): cryptic species and ancient

separations in a globally distributed shorefish. Evolution 55:807-820.

Colin, J.-P. & F. Lethiers. 1988 The importance of ostracods in biostratigraphic analysis.

In Ostracoda in the Earth Sciences. (P. de Deckker, J.-P. Colin & J.-P.

Peypouquet eds) pp 27-46. Elsevier.

Conover, W. J. & R. L. Iman. 1981. Rank transformations as a bridge between

parametric and nonparametric statistics. American Statistician 35:124-129.� 172

Darwin, C. 1859. On the Origin of Species - A Facsimile of the First Edition. Harvard

University Press, Cambridge, Massachusetts and London, England.

de Ricqles, A. 1983. Introduction to the round-table Panchronic forms and living fossils.

Bulletin de la Société Zoologique de France-Evolution et Zoologie 108:529-533.

Delamare-Deboutteville, C. & L. Botosaneanu. 1970. Formes Primitives Vivantes.

Herman, Paris.

Efron, B. & R. Tibshirani. 1993. An Introduction to the Bootstrap. Monographs on

Statistics and Applied Probability. Chapman & Hall.

Eisner, T. 2003. Living Fossils: On Lampreys, Baronia, and the Search for Medicinals.

BioScience 53:265-269.

Eldredge, N. 1979. Alternative approaches to evolutionary theory. Bulletin of Carnegie

Museum of Natural History 13:7-19.

Eldredge, N. 1984. Simpson's Inverse: Bradytely and the phenomenon of living fossils.

In Living Fossils. (S. M. Stanley & N. Eldredge eds) pp 272-277. Berlin.

Springer Verlag.

Eldredge, N. & S. J. Gould. 1972. Punctuated Equilibria: an alternative to phyletic

gradualism. In Models in Paleobiology. (T. J. M. Schopf ed) pp 82-116.

Freeman Cooper and Company, San Francisco.

Eldredge, N. & S.M. Stanley (eds) 1984. Living fossils. Berlin. Springer Verlag.

� 173

Eldredge, N., J. N. Thompson, P. M. Brakefield, S. Gavrilets, D. Jablonski, J. B. C.

Jackson, R. E. Lenski, B. S. Lieberman, M. A. McPeek, & W. Miller. 2005. The

dynamics of evolutionary stasis. Paleobiology (Suppl.) 31:113-145.

Elena, S. F., J. Dopazo, R. Flores, T. O. Diener & A. Moya. 1991. Phylogeny of viroids,

viroid-like satellite RNAs, and the viroid-like domain of hepatitis delta virus-

RNA. Proceedings of the National Academy of the Sciences, U. S. A. 88:5631

-5634.

Endo, K., T. Ozawa & S. Kojima. 2001. Nuclear and mitochondrial DNA sequences

reveal unexpected genetic heterogeneity among Northern Pacific populations of

the brachiopod Lingula anatina. Marine Biology 139:105-112.

Ferson, S., F. J. Rohlf & R. K. Koehn 1985. Measuring shape variation of two-

dimensional outlines. Systematic Zoology 34:59-68.

Fisher, D. C. 1990. Rates of evolution - Living Fossils. In Paleobiology: A Synthesis.

(D. E. G. Briggs & P. R. Crowther eds) pp 152-159. Blackwell Science.

Flessa, K. W. & D. Jablonski. 1985. Declining Phanerozoic background extinction rates:

effect of taxonomic structure? Nature 313:216-218.

Flessa, K. W., K. V. Powers & J. L. Cisne. 1975. Specialization and evolutionary

longevity in the Arthropoda. Paleobiology 1:71-91.

Flynn, L. J., J. C. Barry, M. E. Morgan, D. Pilbeam, L. L. Jacobs & E. H. Lindsay.

1995. Neogene Siwalik mammalian lineages - species longevities, rates of

� 174

change, and modes of speciation. Palaeogeography, Palaeoclimatology,


Foote, M. 1994a. Morphological disparity in Ordovician-Devonian crinoids and the

early saturation of morphological space. Paleobiology 20:320-344.

Foote, M. 1994b. Morphology of Ordovician - Devonian Crinoids. Contributions form

the Museum of Paleontology, University of Michigan 29:1-39.

Foote, M. 1995a. Morphological diversification of Paleozoic crinoids. Paleobiology 21:

273-299.

Foote, M. 1995b. Morphology of Carboniferous and Permian Crinoids. Contributions

from the Museum of Paleontology, University of Michigan 29:136-184.

Foote, M. 1996. On the probability of ancestors in the fossil record. Paleobiology 22:

141-151.

Foote, M. 1997. The evolution of morphological diversity. Annual Review of Ecology

and Systematics 28:129-152.

Foote, M. 1999. Morphological diversity in the evolutionary radiation of Paleozoic and

Post-Paleozoic crinoids. Paleobiology 25(Supplement to 2):1-115.

Foote, M. & D. M. Raup 1996. Fossil preservation and the stratigraphic ranges of taxa.

Paleobiology 22:121-140.

Fortey, R. A. 1980. Generic longevity in Lower Ordovician trilobites; relation to

environment. Paleobiology 6:24-31.

� 175

Fosshagen, A. & T. M. Iliffe. 1985. Two new genera of Calanoida and a new order of

Copepoda, Platycopioida, from marine caves on Bermuda. Sarsia 70:345-358.

Getty Thesaurus of Geographic Names http://www.getty.edu/research/

conducting_research/vocabularies/tgn/index.html

Gibson, G. & G. Wagner. 2000. Canalization in evolutionary genetics: a stabilizing

theory? BioEssays 22:372-380.

Gilinsky, N. L. 1988. Survivorship in the Bivalvia: comparing living and extinct genera

and families. Paleobiology 14:370-386.

Gilinsky, N. L. 1994. Volatility and the Phanerozoic decline of background extinction

intensity. Paleobiology 20:445-458.

Gingerich, P. D. 2001. Rates of evolution on the time scale of the evolutionary process.

Genetica 112:127-144.

Gould, S. J. 2002. The Structure of Evolutionary Theory. The Belknap Press of Harvard

University Press.

Gould, S. J. & N. Eldredge. 1993. Punctuated equilibrium comes of age. Nature 366:

223-227.

Gower, J. C. 1966. Some distance properties of latent root and vector methods used in

multivariate analysis. Biometrika 53:325-338.

Grether, G. F. 2005. Environmental change, phenotypic plasticity, and genetic

compensation. American Naturalist 166:115-123.

� 176

Gruendel, J. 1975. Zur Entwicklung der Trachyleberidididae (Ostracoda) im Jura.

Zeitschrift für Geologische Wissenschaften 3:363-374.

Guensburg, T. E. & J. Sprinkle. 2003. The oldest known crinoids (Early Ordovician,

Utah) and a new crinoid plate homology system. Bulletins of American


Hahn, G., R. Hahn & C. Brauckmann. 2001. The last trilobites. Estudos Tecnologicos.

24:271-281.

Hallam, A. 1975. Evolutionary size increase and longevity in Jurassic bivalves and

ammonites. Nature 258:493-496.

Hammond, L. S. & I. J. Poiner. 1984. Genetic-structure of three populations of the

living fossil brachiopod Lingula from Queensland, Australia. Lethaia 17:139

-143.

Hansen, T. A. 1978. Larval dispersal and species longevity in Lower Tertiary

Gastropods. Science 199:885-887.

Harland, W. B., R. L. Armstrong, A. V. Cox, L. E. Craig, A. G. Smith & D. G. Smith.

1990. A Geologic Time Scale 1989. Cambridge University Press.

Harley, C. D. G., K. F. Smith & V. I. Moore. 2003. Environmental variability and

biogeography: the relationship between bathymetric distribution and

geographical range size in marine algae and gastropods. Global Ecology and

Biogeography 12:499-506.

� 177

Hay, J. M., S. D. Sarre & C. H. Daugherty. 2004. Nuclear mitochondrial pseudogenes as

molecular outgroups for phylogenetically isolated taxa: a case study in

Sphenodon. Heredity 93:468-475.

Hazel, J. E. 1967. Classification and distribution of the Recent Hemicytheridae and

Trachyleberididae (Ostracoda) off Northeastern North America. Geological

Survey Professional Paper 564:1-49.

Hedges, S. B. 2003. The coelacanth of frogs. Nature 425:669-670.

Heinzeller, T., H. Fricke, J. P. Bourseau, N. Ameziane & U. Welsch. 1996. Cyathidium

plantei sp. n., an extant cyrtocrinid (Echinodermata, Crinoidea) -

morphologically identical to the fossil Cyathidium depressum (Cretaceous,

Cenomanian). Zoologica Scripta 25:77-84.

Henningsmoen, G. 1957. Zig-zag evolution. Norsk Geologisk Tidsskrift 44:341-352.

Hess, H., W. I. Ausich, C. E. Brett & M. J. Simms. 1999. Fossil Crinoids. Cambridge

University Press.

Holman, E.W. 1999. Duration and habitat of fossil taxa: changes through time in

variance and taxonomic selectivity. Paleobiology 25:239-251.

Hotchkisss, F. H. C. 1977. Ophioscanops is not a living fossil. Journal of Natural

History 11:337-380.

Howe, R. C. & H. V. Howe 1973. Ostracodes from the Shubata Clay (Tertiary) of

Mississippi. Journal of Paleontology 47:629-657.

� 178

Hunt, G., K. Roy & D. Jablonski. 2005. Species-level heritability reaffirmed: a

comment on "On the heritability of geographic ranges sizes". American

Naturalist 166:129-135.

International Commission on Stratigraphy 2004. http://www.stratigraphy.org/

Ivanov, O. C. 1993. Some proteins keep "living fossil" pre-sequence. Origins of Life

and Evolution of the Biosphere 23:115-124.

Jaanusson, V. 1985. Functional morphology of the shell in platycope ostracodes - a

study of arrested evolution. Lethaia 18:73084.

Jablonski, D. 1980. Apparent versus real biotic effects of transgressions and regressions.


Jablonski, D. 1986a. Larval ecology and macroevolution in marine invertebrates.

Bulletin of Marine Science 39:565-587.

Jablonski, D. 1986b. Background and mass extinctions: the alternation of

macroevolutionary regimes. Science 231:129-133.

Jablonski, D. 1987. Heritability at the species level: Analysis of geographic ranges of

Cretaceous mollusks. Science 238:360-363.

Jablonski, D. 1989. The biology of mass extinctions - a paleontological view.

Philosophical Transactions of the Royal Society of London Series B-Biological

Sciences 325:357-368.

Jablonski, D. 1994. Extinctions in the fossil record. Philosophical Transactions of the

Royal Society of London Series (B) 334:11-16.� 179

Jablonski, D. 2002. Survival without recovery after mass extinctions. Proceedings of the

National Academy of the Sciences, U. S. A. 99:8139-8144.

Jablonski, D. 2005. Mass extinctions and macroevolution. Paleobiology 31:192-210.

Jablonski, D., K. W. Flessa & J. Valentine. 1985. Biogeography and paleobiology.


Jablonski, D. & D. M. Raup. 1995. Selectivity of end-Cretaceous marine bivalve

extinctions. Science 268:389-391.

Jablonski, D. & J. W Valentine. 1990. From regional to total geographic ranges: testing

the relationship in Recent bivalves. Paleobiology 16:126-142.

Jackson, J. B. C. 1974. Biogeographic consequences of eurytopy and stenotopy among

marine bivalves and their evolutionary significance. American Naturalist 108:

541-560.

Janies, D. 2001. Phylogenetic relationships of extant echinoderm classes. Canadian

Journal of Zoology-Revue Canadienne De Zoologie 79:1232-1250.

Jarman, S. N. & N. G. Elliott. 2000. DNA evidence for morphological and cryptic

Cenozoic speciations in the Anaspididae, 'living fossils' from the Triassic.

Journal of Evolutionary Biology 13:624-633.

Jeffery, C. H. & R. B. Emlet. 2003. Macroevolutionary consequences of developmental

mode in temnopleurid echinoids from the tertiary of southern Australia.

Evolution 57:1031-1048.

� 180

Jellinek, T. & K. Swanson. 2003. Report on the taxonomy, biogeography and phylogeny

of mostly living benthic Ostracoda (Crustacea) from deep-sea samples

(Intermediate Water depths) from the Challenger Plateau (Southern Ocean), New

Zealand. Abhandlungen Naturforschenden Gesellschaft Frankfurt am Main 558:

1-329.

Jernvall, J. & M. Fortelius. 2004. Maintenance of trophic structure in fossil mammal

communities: site occupancy and taxon resilience. American Naturalist 164:614

-624.

Kammer, T. W. 2001. Phenotypic bradytely in the Costalocrinus-Barycrinus lineage of

paleozoic cladid crinoids. Journal of Paleontology 75:383-389.

Kammer, T. W., T. K. Baumiller & W. I. Ausich. 1997. Species longevity as a function

of niche breadth: Evidence from fossil crinoids. Geology 25:219-222.

Kammer, T. W., T. K. Baumiller & W. I. Ausich. 1998. Evolutionary significance of

differential species longevity in Osagean-Meramecian (Mississippian) crinoid

clades. Paleobiology 24:155-176.

Keen, M. C. 1993. Ostracods as palaeoenvironmental indicators: examples from the

Tertiary and early Cretaceous. In Applied Micropaleontology. (D. G. Jenkins

ed) pp 41-67. Kluwer Academic Publishers.

Keen, M. C., S. S. J. Al-Sheikly, A. Elsogher & A. M. Gammudi. 1994. Tertiary

ostracods of North Africa and the Middle East. In Micropalaeontology and

� 181

Hydrocarbon Exploration in the Middle East. (M. D. Simmons ed) pp 371-388.

Chapman and Hall, London.

Kempf, E. K. Index and Bibliography of Marine Ostracoda. 1986-2005. Geologisches

Institut der Universitaet zu Koeln Sonderveroeffentlichungen.

Kiessling, W. & R. C. Baron-Szábo. 2004. Extinction and recovery patterns of

sclerectinian corals at the Cretaceous-Tertiary boundary. Paleogeography,

Paleoclimatology, Paleoecology 214:195-223.

King, J. L. & R. Hanner. 1998. Cryptic species in a "Living Fossil" lineage: Taxonomic

and phylogenetic relationships within the genus Lepidurus (Crustacea :

Notostraca) in North America. Molecular Phylogenetics and Evolution 10:23-36.

Knowlton, N. 2000. Molecular genetic analyses of species boundaries in the sea.

Hydrobiologia 420:73-90.

Koch, C. F. 1987. Prediction of sample size effects on the measured temporal and

geographic distribution patterns of species. Paleobiology 13:100-107.

Koch, C. F. & J. P. Morgan. 1988. On the expected distribution of species' range.


Koehl, M. A. R. 1996. When does morphology matter? Annual Review of Ecology,

Evolution and Systematics 27:501-542.

Kraft, P., O. Lehnert & J. Fryda. 1999. Titerina, a living fossil in the Ordovician: A

young protoconodont (?) and the oldest chaetognath animal. Acta Universitatis

Carolinae, Geologica 43:451-454.� 182

Kruskal, J.B. 1964. Nonmetric Multidimensional Scaling : A numerical method.

Psychometrika 29:115-129.

Kyrpides, N. C. & C. A. Ouzounis. 1995. Nucleic acid-binding metabolic enzymes -

living fossils of stereochemical interactions. Journal of Molecular Evolution 40:

564-569.

Laille, M., F. Gerald & C. Debitus 1998. In vitro antiviral activity on dengue virus of

marine natural products. Cellular and Molecular Life Sciences 54:167-170.

Lane, A. L. & G.D. Webster 1980. Crinoidea. In Echinoderms: Notes for a short course.

(T. W. Broadhead & J. A. Waters eds) pp 144-157. University of Tennessee,

Studies in Geology, vol. 3.

Lane, N. G. 1978. Evolution of Flexible Crinoids. In Treatise on Invertebrate

Paleontology, Part T Echinodermata 2. (R. C. Moore & C. Teichert eds) pp 301

-302. Geological Society of America and University of Kansas Press.

Lange, R., M. Bergbauer, U. Szewzyk & J. Reitner. 2001. Soluble proteins control

growth of skeleton crystals in three coralline demosponges. Facies 45:195-201.

Leander, S. L. & P. J. Keeling 2003. Morphostasis in alveolate evolution. Trends in

Ecology & Evolution 18:395-402.

Lesicki, A. 1998. Mesozoic barnacles (Crustacea: Cirripedia) and other relic species of

the hydrothermal vent fauna versus the problem of its origin. Przeglad

Zoologiczny 42:135-155.

� 183

Levinton, J. S. & L. Ginzburg. 1984. Repeatability of taxon longevity in successive

Foraminifera radiations and a theory of random appearance and extinction.

Proceedings of the National Academy of the Sciences, U. S. A. 81:5478-5481.

Liow, L. H. 2004. A test of Simpson's “Rule of the survival of the relatively

unspecialized” using fossil crinoids. American Naturalist 164:431-443.

Liow, L. H. 2006. Do deviants live longer? Morphology and longevity in

trachyleberidid ostracodes. Paleobiology 32:55-69.

Little, C. T. S. & R. C. Vrijenhoek. 2003. Are hydrothermal vent animals living fossils?

Trends in Ecology and Evolution 8:582-588.

Lonnig, W. E. & H. Saedler. 2002. Chromosome rearrangements and transposable

elements. Annual Review of Genetics 36:389-410.

Lord, F. N. 1979. On Oligocythereis kostytschevkaensis (Lyubimova). Stereo-Atlas of

Ostracod Shells 6:55-62.

Mac Gillavry, H. J. 1968. Modes of evolution mainly among marine invertebrates.

Bijdragen tot de DierKunde 38:69-74.

MacLeod, N. 1999. Generalizing and extending the eigenshape method of shape space

visualization and analysis. Paleobiology 25:107-138.

Maddocks, R. F. (ed) 1983. Applications of Ostracoda. Proceedings of the Eighth

International Symposium on Ostracoda, Department of Geosciences, University

of Houston.

� 184

Marshall, C. R. 1997. Confidence intervals on stratigraphic ranges with nonrandom

distributions of fossil horizons. Paleobiology 23: 165-173.

Martinell, J. & A. A. Hoffman. 1983. Species duration patterns in the Pliocene

gastropod fauna of Emporda (Northeast Spain). Neues Jahrbuch für Geologie

und Paläontologie-Monatschefte 1983:698-704.

McClure, M. & A. J. Bohonak. 1995. Non-selectivity in extinction of bivalves in the

Late Cretaceous of the Atlantic and Gulf Coastal Plain of North America.

Journal of Evolutionary Biology 8:779-794.

McGhee, G.R. 1999. Theoretical morphology: the concept and its applications.

Perspectives in Paleobiology and Earth History. Columbia University Press,

New York.

McGowan, A. J. Relationship between disparity and survivorship among Triassic

ammonoid genera. Historical Biology (accepted).

Mckenzie, K. G. 1967. Ostracod "living fossils" new finds in the Pacific. Science 155:

1005.

McKinney, M. L. 1986. Biostratigraphic gap analysis. Geology 14:36-38.

McKinney, M. L. 1995. Extinction selectivity among lower taxa - gradational patterns

and rarefaction error in extinction estimates. Paleobiology 21:300-313.

McKinney, M. L. 1997. Extinction vulnerability and selectivity: combining ecological

and paleontological views. Annual Review of Ecology and Systematics 28:495

-516.� 185

Merilä, J., B. C. Sheldon, and L. E. B. Kruuk. 2001. Explaining stasis:

microevolutionary studies in natural populations. Genetica 112-113:199-222.

Miller, A. I. 1997. A new look at age and area: The geographic and environmental

expansion of genera during the Ordovician radiation. Paleobiology 23:410-419.

Miller, A. I. & M. Foote. 1996. Calibrating the Ordovician Radiation of marine life:

implications for Phanerozoic diversity trends. Paleobiology 22:304-309.

Mooi, R. 1990. A new "living fossil" echinoid (Echinodermata) and the ecology and

paleobiology of Caribbean Cassiduloids. Bulletin of Marine Science 46:688

-700.

Moore, R. C. & C. Teichert (eds) 1978. Treatise on invertebrate paleontology. Part T,

Echinodermata 2 (1-3) Geological Society of America and University of Kansas

Press.

National Geophysical Data Center. http://www.ngdc.noaa.gov/mgg/geology/

Newman, S. A. 1985. The abyssal hydrothermal vent invertebrate fauna: a glimpse of

antiquity? Biological Society of Washington Bulletin 6:231-242.

Newman, W. A. & R. R. Hessler. 1989. A new abyssal hydrothermal verrucomorphan

(Cirripedia; Sessilia): the most primitive living sessile barnacle. Transactions of

the San Diego Society of Natural History 21:259-273.

Norris, R. D. 1992. Extinction selectivity and ecology in planktonic Foraminifera.

Palaeogeography, Palaeoclimatology, Palaeoecology 95:1-17.

� 186

O'Grady, J. J., D. H. Reed, B. W. Brook & R. Frankham 2004. What are the best

correlates of predicted extinction risk? Biological Conservation 118:513-520.

Oji, T. 1994. A comparison between "living fossils" and other organisms. Fossils

(Tokyo) 57:47-49.

Oji, T. 1996. Is predation intensity reduced with increasing depth? Evidence from the

west Atlantic stalked crinoid Endoxocrinus parrae (Gervais) and implications

for the Mesozoic marine revolution. Paleobiology 22:339-351.

Oji, T. 2001. Deep-sea communities. In Paleobiology II. (D. E. G. Briggs & P. R.

Crowther eds) pp 444-447. Blackwell Science.

Parsons, P. A. 1993. Stress, extinctions and evolutionary change: from living organisms

to fossils. Biological Reviews - Cambridge Philosophical Society 68:313-333.

Parsons, P. A. 1994. Morphological stasis - an energetic and ecological perspective

incorporating stress. Journal of Theoretical Biology 171:409-414.

Pearson, P. N. 1992. Survivorship analysis of fossil taxa when real-time extinction rates

vary: the Paleogene planktonic Foraminifera. Paleobiology 18:115-131.

Pease, C. M. 1985. Biases in the durations and diversities of fossil taxa. Paleobiology

11:272-292.

Pineda, J. 1993. Boundary effects on the vertical ranges of deep-sea benthic species.

Deep Sea Research Part I: Oceanographic Research Papers 40:2179-2192.

Poinar, G. O., B. M. Waggoner & U. C. Bauer. 1993. Terrestrial soft-bodied protists and

other microorganisms in Triassic amber. Science 259:222-224.� 187

R Development Core Team 2005. R: A Language and Environment for Statistical

Computing. http://www.R-project.org.

Raup, D.M. & G. E. Boyajian. 1988. Patterns of generic extinction in the fossil record.


Raup, D. M., S. J. Gould, T. J. M. Schopf & D. S. Simberloff. 1973. Stochastic models

of phylogeny and the evolution of diversity. Journal of Geology 81:525-542.

Raup, D. M. & L. G. Marshall. 1980. Variation between groups in evolutionary rates: A

statistical test of significance. Paleobiology 6:9-23.

Raup, D.M. & J. J. Sepkoski. 1982. Mass extinctions in the marine fossil record.

Science 215:1501-1503.

Reyment, R. A. 1988. Applications of ostracods in quantitative geology. In Ostracoda in

the Earth Sciences. (P. de Deckker, C. J.-P. & J.-P. Peypouquet eds) pp 257-276.

Elsevier.

Reynolds, J. D., N. K. Dulvy, N. B. Goodwin & J. A. Hutchings 2005. Biology of

extinction risk in marine fishes. Proceedings of the Royal Society B-Biological

Sciences 272:2337-2344.

Ricklefs, R. E. & R. E. Latham. 1992. Intercontinental correlation of geographical

ranges suggests stasis in ecological traits of relict genera of temperate perennial

herbs. American Naturalist 139:1305-1321.

Robeck, H. E., C. C. Maley & M. J. Donoghue. 2000. Taxonomy and temporal diversity

patterns. Paleobiology 26:171-187.� 188

Rohlf, F. J. 1992. Elliptic Fourier analysis http://life.bio.sunysb.edu/

Roush, W. 1997. "Living fossil" fish is dethroned. Science 277:1436.

Roux, M., 1987. Evolutionary ecology and biogeography of recent stalked crinoids as a

model for the fossil record. Echinoderm Studies 2:1-53.

Roy, K. & M. Foote 1997. Morphological approaches to measuring biodiversity. Trends

in Ecology & Evolution 12:277-281.

Ruedemann, R. 1918. The paleontology of arrested evolution. New York State Museum

Bulletin 196:107-138.

Ruedemann, R. 1922a. Additional studies in arrested evolution. Proceedings of the

National Academy of Sciences 8:34-35.

Ruedemann, R. 1922b. Further notes on the paleontology of arrested evolution.

American Naturalist 56:256-272.

Ruiz, F., M. L. Gonzalez-Regalado, J. M. Muñoz, J. G. Pendónc, A. Rodriguez-

Ramírez, L. Cáceres & A. Rodríguez Vidal. 2003. Population age structure

techniques and ostracods: Applications in coastal hydrodynamics and

paleoenvironmental analysis. Paleogeography, Paleoclimatology, Paleoecology

199:51-69.

Russell, M. P. & D. R. Lindberg. 1988. Estimates of species durations. Science 240:

969.

Rutherford, S. L. 2000. From genotype to phenotype: buffering mechanisms and the

storage of genetic information. BioEssays 22:1095-1105.� 189

Saether, B. E., S. Engen, A. P. Moller, M. E. Visser, E. Matthysen, W. Fiedler, M. M.

Lambrechts, P. H. Becker, J. E. Brommer, J. Dickinson, C. Du Feu, F. R.

Gehlbach, J. Merila, W. Rendell, R. J. Robertson, D. Thomson & J. Torok. 2005.

Time to extinction of bird populations. Ecology 86:693-700.

Schein-Fatton, É. 1985. Découverte sur la ride du Pacifique oriental à 13ºN d'une

Pectinidae (Bivalvia, Pteromorphia) d'affinités paléozoïques. Comptes Rendus -

Académie des Sciences, Série III 301:491-496.

Schopf, J. W. 1994. Disparate rates, differing fates: tempo and mode of evolution

changed from the Precambrain to the Phanerozoic. Proceedings of the National

Academy of the Sciences, U. S. A. 91:6735-6742.

Schopf, T. J. M. 1984. Rates of evolution and the notion of "living fossils." Annual

Review of Earth and Planetary Science 12:245-292.

Schopf, T. J. M., D. M. Raup, S. J. Gould & D.S. Simberloff. 1975. Genomic versus

morphologic rates of evolution: influence of morphologic complexity.


Schornikov, E. I. 2005. The question of cosmopolitanism in the deep-sea ostracod

fauna: The example of the genus Pedicythere. Hydrobiologia 538:193-215.

Schubert, J. K., D. J. Bottjer & M. J. Simms. 1992. Paleobiology of the oldest known

Articulate Crinoid. Lethaia 25:97-110.

Schwenk, K. & G. P. Wagner. 2001. Function and the evolution of phenotypic stability:

Connecting pattern to process. American Zoologist 41:552-563.� 190

Sekiguchi, K. & H. Sugita. 1980. Systematics and hybridization in the four living

species of horseshoe crabs. Evolution 34:712-718.

Selander, R. K., S. Y. Yang, R. C. Lewontin & W. E. Johnson. 1970. Genetic variation in

the horseshoe crab (Limulus polyphemus), a phylogenetic "relic". Evolution 24:

402-414.

Sepkoski, J. J. 1975. Stratigraphic biases in the analysis of taxonomic survivorship.


Sepkoski, J. J. 1982. A compendium of fossil marine families. Milwaukee Public

Museum Contributions in Biology and Geology 51:1-125.

Sepkoski, J. J. 1993. 10 years in the library - new data confirm paleontological patterns.


Sepkoski, J. J. & D. C. Kendrick. 1993. Numerical experiments with model

monophyletic and paraphyletic taxa. Paleobiology 19:168-184.

Serbet, R. & G. W. Rothwell. 1999. Osmunda cinnamomea (Osmundaceae) in the

Upper Cretaceous of western North America: Additional evidence for

exceptional species longevity among filicalean ferns. International Journal of

Plant Sciences 160:425-433.

Siegal, S. M., K. Roberts, H. Nathan & O. Daly. 1967. Living relative of the microfossil

Kakabekia. Science 156:1231-1234.

Siegel, S. M. & B. Z. Siegel. 1968. A living organism morphologically comparable to

the Precambrian genus Kakabekia. American Journal of Botany 55:684-687.� 191

Simms, M. J. 1988. The phylogeny of Post-Paleozoic crinoids. In Echinoderm

Phylogeny and Evolutionary Biology. (C. R. C. Paul & A. B. Smith eds) pp 269

-284. Clarendon Press, Oxford.

Simms, M. J. 1999. Systematics, phylogeny and evolutionary history. In Fossil

Crinoids. (H. Hess, W.I. Ausich, C.E. Brett & M.J. Simms eds.) pp. 31-49.

Cambridge University Press.

Simpson, G. G. 1944. Tempo and Mode in Evolution. Columbia University Press. N.Y.

Simpson, G. G. 1953. The Major Features of Evolution. Columbia University Press.

N.Y.

Sneath, P. H. A. & R. R. Sokal. 1973. Numerical Taxonomy. W. H. Freeman and

Company, San Fransisco.

Sodhi, N. S., L. H. Liow & F. A. Bazzaz. 2004. Avian extinctions from tropical and

subtropical forests. Annual Review of Ecology Evolution and Systematics 35:

323-345.

Sokal, R. R. & F. J. Rohlf. 1995. Biometry. 3rd edn. New York, W.H. Freeman and

Company.

Soltis, P. S., D. E. Soltis, V. Savalaine, P. R. Crane & T. G. Barraclough. 2002. Rate

heterogeneity among lineages of tracheophtyes: integration of molecular and

fossil data and evidence for molecular living fossils. Proceedings of the National

Academy of the Sciences, U. S. A. 99:4430-4435.

� 192

Spielman, D., B. W. Brook & R. Frankham 2004. Most species are not driven to

extinction before genetic factors impact them. Proceedings of the National

Academy of the Sciences, U. S. A. 101:15261-15264

Stanley, S. M. 1979. Macroevolution: Patterns and Processes. The John Hopkins

University Press. Baltimore and London.

Stanley, S. M. 1985. Rates of evolution. Paleobiology 11:13-26.

Stanley, S. M. 1986. Population size, extinction, and speciation - the fission effect in

Neogene Bivalvia. Paleobiology 12:89-110.

Stenseth, N. C. & J. Maynard-Smith. 1984. Coevolution in ecosystems: Red Queen

evolution or stasis? Evolution 38:870-880.

Stepien, C. A., B. Morton, K. A. Dabrowska, R. A. Guarnera, T. Radja & B. Radja.

2001. Genetic diversity and evolutionary relationships of the troglodytic 'living

fossil' Congeria kusceri (Bivalvia : Dreissenidae). Molecular Ecology 10:1873

-1879.

Suno-Uchi, N., F. Sasaki, S. Chiba & M. Kawata. 1997. Morphological stasis and

phylogenetic relationships in tadpole shrimps, Triops (Crustacea: Notostraca).

Biological Journal of the Linnean Society 61:439-457.

Swain, F. M. 1974. Provincial aspects of some Neogene Ostracoda of the United States.

Paleogeographic provinces and provinciality. In Paleogeographic Provinces and

Provinciality. (C. A. Ross ed) pp 127-135. Society of Economic Paleontologists

and Mineralogists, Tulsa, Oklahoma.� 193

Swain, F. M. 1992. Factors influencing ranges of Jurassic and Lower Cretaceous

Ostracoda, Eastern North America and Western North Atlantic Ocean. Revista

Espanola Micropaleontologia 24:43-66.

Sylvester-Bradley, P. C. 1948. The ostracode genus Cythereis. Journal of Paleontology

22:792-797.

Tasch, P. 1969. Branchiopoda. In Treatise on Invertebrate Paleontology Part R. (R. C.

Moore ed) pp 128-191. Geological Society of America and University of Kansas

Press.

Taylor, R. W. 1978. Nothomyrmecia macrops: a living fossil ant rediscovered. Science

201:979-985.

Thenius, E. 2000. Mitochondrien als "lebende Fossilien"? Protein-Evolution und

"adaptive" Evolution. Natur und Museum (Frankfurt am Main) 130:3550358.

Thenius, E. 2001. Lebende Fossilien: Oldtimer der Tier- und Pflanzenwelt Zeugen der

Vorzeit. Verlag Dr. Friedrich Pfeil, Muenchen.

Trott, T. J. 1998. Gustatory responses of Priapulus caudatus Lamarck, 1816 (Priapulida,

Priapulidae): Feeding behavior and chemoreception by a living fossil. Marine

and Freshwater Behaviour and Physiology 31:251-257.

Tucker, A. B. 1998. Systematics of the Raninidae (Crustacea: Decapoda: Brachyura),

with accounts of three new genera and two new species. Proceedings of the

Biological Society of Washington 111:320-371.

� 194

Valentine, J. W. & C. L. May. 1996. Hierarchies in biology and paleontology.

Paleobiology 22: 23-33.

van den Bold, W. A. 1970. The genus Costa (Ostracoda) in the Upper Cenozoic of the

Caribbean Region. Micropaleontology 16:61-75.

van Morkhoven, F. P. C. M. 1963. Post-Palaeozoic Ostracoda - Their Morphology,

Taxonomy and Economic Use. Elsevier.

van Valen, L. 1975. Group selection, sex, and fossils. Evolution 29:87-94.

van Valen, L. 1979. Taxonomic survivorship curves. Evolutionary Theory 4:129-142.

van Valen, L. 1982. Integration of species: stasis and biogeography. Evolutionary

Theory 6:99-112.

Vermeij, G. J. 1987. Evolution and Escalation: An Ecological Theory of Life. Princeton

University Press, Princeton, N J.

Vermeij, G. J. 1993. Biogeography of recently extinct marine species - implications for

conservation. Conservation Biology 7:391-397.

Viranta, S. 2003. Geographic and temporal ranges of Middle and Late Miocene

carnivores. Journal of Mammalogy 84:1267-1278.

Vrba, E. S. 1984. Evolutionary pattern and process in the sister-group Alcelaphini-

Aepycerotini (Mammalia: Bovidae). In Living Fossils. (S. M. Stanley and N.

Eldredge eds) pp 62-79. Springer Verlag. Berlin.

Wackerly, D. D., W. Mendenhall III. & R. L. Scheaffer. 2002. Mathematical Statistics

with Applications. Duxbury.� 195

Wagner, P. J. 2000. Exhaustion of morphologic character states among fossil taxa.

Evolution 54:365-386. (http://pjw3.fmnh.org/EvolutionMatrices.2000.html)

Wagner, P. J. & D. H. Erwin. 1995. Phylogenetic patterns as tests of speciation models.

In New Approaches to Speciation in the Fossil Record. (D. H. Erwin & R. L.

Anstey eds) pp 87-122. Columbia University Press, New York.

Wainwright, P. C. & S. M. Reilly. 1994. Ecological Morphology: Integrative

Organismal Biology. University of Chicago Press.

Wake, D. B., G. Roth & M. H. Wake. 1983. On the problem of stasis in organismal

evolution. Journal of Theoretical Biology 101:211-224.

Wall, D. & B. Dale. 1966. " Living fossils" in western atlantic plankton. Nature 211:

1025-1026.

Ward, P. D. 1992. On Methuselah's Trail: Living Fossils and the Great Extinctions. W.H.

Freeman and Company, New York.

Ward, P. D. & P. W. Signor 1983. Evolutionary tempo in Jurassic and Cretaceous

Ammonites. Paleobiology 9:183-198.

Webster, G. D. 2003. Bibliography and index of Paleozoic crinoids, coronates, and

hemistreptocrinoids, 1758-1999. Geological Society of America, Special Paper

363. http://crinoid.gsajournals.org/crinoidmod/

Wen, J. 1999. Evolution of eastern Asian and eastern North American disjunct

distributions in flowering plants. Annual Review in Ecology, Evolution and

Systematics 30:421-455.� 196

Whatley, R. C. & M. Ayress. 1988. Pandemic and endemic distribution patterns in

Quaternary deep-sea Ostracoda. In Evolutionary Biology of Ostracoda: its

Fundamentals and Applications. (T. Hanai, N. Ikeya & K. Ishizaki eds) pp 739

-755. Elsevier.

Wills, M. A. 2001. Morphological disparity : A primer. In Fossils, Phylogeny, and Form:

An Analytical Approach. (J. M. Adrain, G. D. Edgecombe & B. S. Lieberman,

eds) pp 55-144. Kluwer Academic/ Plenum Publishers.

Wilson, G. D. F. 1999. Some of the deep-sea fauna is ancient. Crustaceana 72:1019

-1030.

Woodruff, D. S., M. Mulvey, W. B. Saunders & M. P. Carpenter. 1983. Genetic variation

in the cephalopod Nautilus belauensis. Proceedings of the Academy of Natural

Sciences of Philadelphia 135:147-153.

Yoshida, K. 2002. Long survival of "living fossils" with low taxonomic diversities in an

evolving food web. Paleobiology 28:464-473.

Zhou, Z.-Y. & S.-Y. Zheng. 2003. The missing link in Ginkgo evolution. Nature 423:

821-822.

� 197

Appendix A

Description of characters and character matrix for seven crinoids not represented in

Foote (1999).

The characters and their states largely follow those used previously (Foote 1994a,

1994b, 1995a, 1995b). Character states for the stem are based on the proximal portion;

columns that are proximally straight but distally coiled, for example, are not coded as

coiled. Characters are indicated as binary (B), ordered multistate (O), or unordered

multistate (U). Unless otherwise indicated, the states for binary characters correspond

with absence (0) or presence (1). The absence of a trait should be construed to imply its

obvious alternative; for example, if arms are branched, absence of heterotomous

branching implies presence of isotomous branching. (This description of characters,

with some modifications, is reproduced with permission from M. Foote 1999).

Pelma

1. Form of pelma (U): 0, absent (unattached); 1, multiplated holdfast (as in

Echmatocrinus); 2, column; 3, absent (directly attached).

2. Xenomorphic column (B): Column is considered xenomorphic if the transition

between regions of the column is abrupt.

3. Heteromorphic column (B)

4. Coiled column (B)

5. Meric columnals (B)� 198

6. Shape of columnals (U): 0, round; 1, elliptical; 4, tetragonal; 5, tetralobate or

-stellate; 6, pentagonal; 7, pentalobate or -stellate.

7. Shape of lumen (U): same states as character 6.

8. Relative height of columnals (B): 0, discoidal (Height:Width < 0.5); 1, elongate

(Height:Width > 0.5).

9. Columnal articulations (U): 0, synostosis or cryptosymplexy; 1, symplexy; 2,

synarthry.

10. Cirri (B)

11. Regular arrangement of cirri (B)

12. Number of cirri per nodal (O): 0, <5; 1, 5; 2 >5.

13. Specialized distal structure (B)

14. Form of distal structure (U): 0, irregular plates; 1, radix; 2, discoidal or crustose

holdfast; 3, anchor; 4, float.

Dorsal cup

15. Regular calyx plating (B)

16. Radials (O): 0, absent; 1, cryptic; 2, exposed.

17. Number of radials (O)

18. Fused radials (B)

19. Opening of radial circlet (B)

� 199

20. Nature of opening (U): 1, anal interray only, open by anal(s); 2, anal interray

only, open by basal(s); 3, open in anal and/or other interrays.

21. Radial prongs or sinus (B)

22. Unequal development of radials (B)

23. Compound radials (B)

24. Number of compound radials (O)

25. Basals (O): 0, absent; 1, cryptic; 2, exposed.

26. Number of basals (O)

27. Opening of basal circlet (B)

28. Nature of opening (U): 1, anal interray only, open by anal(s); 2, anal interray

only, open by radial(s); 3, open in anal and/or other interrays.

29. Unequal development of basals (B)

30. Size of basal circlet area relative to radial circlet area (O): 0, less than half the

area; 1, subequal; 2, greater than twice the area.

31. Infrabasals (O): 0, absent; 1, cryptic; 2, exposed.

32. Number of infrabasals (O)

33. Opening of infrabasal circlet (B)

34. Unequal development of infrabasals (B)

35. Size of infrabasal circlet area relative to radial circlet area (O): 0, less than half

the area; 1, subequal; 2, greater than twice the area.

36. Centrodorsal, calyconodal, proximale or analogue (B)� 200

37. Cirriferous centrodorsal or analogue (B)

38. Regular arrangement of cirri (B)

39. Style of regular arrangement (U): 0, columns; 1, whorls; 2, single circlet.

40. Size of centrodorsal (or analogue) area relative to radial circlet area (O): 0, less

than half the area; 1, subequal; 2, greater than twice the area.

41. Segmentation of centrodorsal or analogue (B)

42. Number of anal plates in dorsal cup at or below level of radials (O)

43. Accessory plates (as in Perittocrinus) (B)

44. Intercalary plates (as in Acrocrinus) (B)

number of ranges of intercalaries (O)

46. Shape of dorsal cup (sagittal) (U): 0, cylinder or disk; 1, cone; 2, bowl; 3, globe;

4, inverted cone (as in Calceolispongia); 5, inverted bowl (as in Pilidiocrinus); 6,

splayed bowl or cone (as in Uperocrinus); 7, goblet; 8, club; 9. bicone.

47. Shape of dorsal cup (sagittal) (O): 0, low (Width:Height > 1.5); 1, medium; 2,

high (Height:Width > 1.5).

48. Shape of dorsal cup (transverse) (U): 0, round; 1, polygonal or convex; 2, lobate

or stellate.

49. Symmetry of dorsal cup (transverse) (U): 0, asymmetric; 1, strongly bilateral; 2,

triradial; 3, tetraradial; 4, pentameral with strong bilateral overprint; 5, strongly

pentameral; 6, hexaradial.

50. Concave base (B)� 201

51. Cup diameter greater than 2.5 times stem diameter (B)

52. Major projections (wings, blades, spines) (B)

53. Median ray ridges (B)

54. Stellate ridges (B)

Arms and ambulacral system

55. Arms (B)

56. Number of distinct arms at point where they become free of cup (O)

57. Maximal number of arms directly attached to single radial (O): 0, 1; 1, 2; 2, >2.

58. Relative development of arms (O): 0, subequal; 1, slightly unequal; 2, strongly

unequal.

59. Separation of arms at cup (O): 0, appressed or nearly so; 1, less than 1.5 arm

widths apart; 2, greater than 1.5 arm widths apart.

60. Lateral arm fusion between rays (B)

61. Branched arms (B)

62. Effective number of orders of branching (O): 0, 1; 1, 2; 2, >2.

63. Heterotomous branching (B)

64. Nature of heterotomous branching (U): 0, bilateral; 1, endotomous; 2,

exotomous; 3, other regular (e.g., abradial or adradial); 4, irregular.

65. Biserial arm plating (B)

66. Patelloid process (if uniserial) (B)� 202

67. Cuneate or asymmetric brachials (if uniserial) (B)

68. Brachial shape (Height:Width) (O): 0, < 0.5; 1, 0.5-1.0; 2, 1.0-2.0; 3, >2.0.

69. Lateral arm fusion within rays (B)

70. Arm attitude at base (O): 0, sloping inward, vertical, or forming cone; 2,

sidewards; 3, pendent.

71. Recumbent arms (as in Agostocrinus) (B)

72. Incorporation of radially aligned brachials into cup (B)

73. Number of ranges of brachials in cup (O)

74. Interbrachials (including fixed pinnules) in cup (B)

75. Form of proximal interbrachials (B): 0, small, irregular; 1, larger, regular.

76. Pinnules (B)

77. Characteristic maximal number of pinnules per brachial (O)

78. Recumbent ambulacra (as in Hybocrinus) (B)

79. Number of recumbent ambulacra (O)

80. Recumbent ambulacra extending more than halfway down cup (B)

81. Ratio of arm length to cup height (O): 0 < 1; 1, 1-2; 2, 2-4; 3, >4.

82. Rorted arms (as in Mespilocrinus) (B)

Anal and tegminal features

83. Anal opening through dorsal cup (B)

84. Anal tube or sac (B)� 203

85. Position of tube (B): 0, posterior; 1, central or anterior.

86. Rube extending greater than twice cup height (B)

87. Ridges (including plicae) on proximal part of tube (B)

88. Irregular plating of tube (B)

89. Development of tegmen (other than tube/sac) (B): 0, orals only or a few large

plates; 1, multiplated.

90. Regmen (other than tube/sac) extending greater than twice cup height (B)

Table of Characters

Characters 1 through 90 (Foote 1999) for Titanorinus, Glenocrinus, Celtocrinus,

Eknomocrinus, Cnemecrinus, Adelphicrinus and Habrotecrinus, columns 1 through 7

respectively. N = not applicable, ? = missing.

No. 1 2 3 4 5 6 7

1 2 2 2 2 2 2 2

2 0 0 0 0 0 0 0

3 0 1 1 1 1 0 0

4 0 0 0 0 0 0 0

5 1 0 0 1 1 1 1

6 0 7 0 7 7 0 7

7 0 6 ? 6 7 0 7

8 0 0 0 0 0 0 0� 204

9 0 1 ? 0 0 0 0

10 0 0 0 0 0 0 0

11 N N N N N N N

12 N N N N N N N

13 ? ? ? ? ? ? ?

14 ? ? ? ? ? ? ?

15 0 0 1 1 1 1 1

16 2 1 2 2 2 2 2

17 5 5 ? ? 5 5 5

18 0 0 0 0 0 0 0

19 1 1 ? 1 1 0 1

20 3 3 ? 1 3 N 3

21 0 0 0 0 0 0 1

22 0 0 0 0 0 0 0

23 0 0 0 0 0 0 0

24 N N N N N N N

25 2 2 2 2 2 2 2

26 ? ? ? 5 5 5 5

27 1 3 ? 1 0 0 1

28 3 ? ? 1 N N 3

29 0 ? ? ? 0 1 0� 205

30 ? ? 1 0 1 1 1

31 2 2 0 0 2 0 0

32 5 5 N N 5 N N

33 1 1 N N 0 N 0

34 0 ? N N 0 N N

35 ? ? N N 0 N N

36 0 N 0 0 0 0 0

37 N N N N N N N

38 N N N N N N N

39 N N N N N N N

40 N N N N N N N

41 N N N N N N N

42 ? ? ? 1 4 ? ?

43 1 1 0 0 0 0 1

44 1 1 0 0 0 0 0

45 2 2 N N N N N

46 1 3 1 2 2 1 1

47 2 1 2 0 0 1 2

48 0 0 2 2 0 0 ?

49 4 4 ? 4 4 4 4

50 0 1 0 0 0 0 0� 206

51 1 1 0 1 1 0 ?

52 0 0 0 0 0 0 0

53 0 0 1 0 0 0 0

54 1 1 1 0 1 0 1

55 0 1 1 1 1 1 1

56 10 10 ? ? 5 10 ?

57 0 0 0 0 0 0 0

58 0 0 0 0 0 0 0

59 2 2 2 2 2 2 2

60 0 0 0 0 0 0 0

61 1 1 1 1 1 1 1

62 2 2 2 2 2 2 ?

63 0 1 1 0 0 0 0

64 N ? ? N N N N

65 1 0 0 0 0 0 0

66 N N 0 0 0 0 0

67 0 0 0 0 0 0 0

68 1 1 1 1 1 1 1

69 0 0 0 0 0 0 0

70 1 1 0 0 0 0 0

71 0 0 0 0 0 0 0� 207

72 1 1 1 1 1 1 1

73 1 1 3 1 1 1 1

74 1 1 1 1 1 1 1

75 0 0 0 1 1 0 0

76 0 0 1 0 0 0 ?

77 N N 1 N N N ?

78 0 0 0 0 0 0 0

79 N N N N N N N

80 N N N N N N ?

81 1 1 ? ? 2 ? ?

82 0 0 0 0 0 0 0

83 0 0 ? 0 0 0 ?

84 0 0 0 0 0 ? ?

85 N N N N N ? ?

86 N N N N N ? ?

87 N N N N N ? ?

88 N N N 1 N ? ?

89 1 1 ? 1 ? ? ?

90 0 0 ? 0 ? ? ?

� 208

Appendix B

Ap

pe

nd

ix B

. C

rin

oid

ge

ne

ra in

ord

ers

(a

nd

su

bo

rde

rs o

f cla

did

s)

an

d t

he

ir m

orp

ho

-du

ratio

n p

lot

dis

trib

utio

ns,

rela

tive

to

ba

sa

l g

en

era

. N

= n

um

be

r o

f g

en

era

sa

mp

led

. S

, P

sh

ow

s w

he

the

r ch

oic

e o

f ta

xo

n w

as b

ase

d o

n

str

atig

rap

hic

po

sitio

ns (

S),

ph

ylo

ge

ne

tic in

sig

hts

(P

) o

r b

oth

. P

are

nth

etica

l n

um

be

rs =

re

fere

nce

s o

n w

hic

h

the

in

form

atio

n w

as b

ase

d (

1 =

Ba

um

ille

r a

nd

Ha

gd

orn

19

95

, 2

= G

ue

nsb

urg

an

d S

prin

kle

20

03

; 3

= M

oo

re a

nd

Te

ich

ert

19

78

; 4

= A

usic

h 1

99

8;

5 =

Sch

ub

ert

et

al. 1

99

2;

6 =

Fo

ote

19

99

). C

olu

mn

s S

L-b

-LL

-t a

re t

he

qu

ad

ran

ts

as n

am

ed

in

Fig

. II

.1;

nu

mb

ers

in

dic

ate

th

eir p

rop

ort

ion

occu

pa

tio

n.

Me

d,

mid

an

d 1

0-g

are

cu

toff

po

ints

fo

r

du

ratio

ns o

f lo

ng

-liv

ed

ge

ne

ra a

s d

efin

ed

in

"M

ate

ria

ls a

nd

Me

tho

ds."

N

um

be

rs in

th

e p

rop

ort

ion

co

lum

ns in

dic

ate

the

pro

po

rtio

ns o

f ra

refie

d s

am

ple

s o

f sh

ort

-liv

ed

ge

ne

ra t

ha

t a

re le

ss d

evia

nt

tha

n lo

ng

-liv

ed

ge

ne

ra,

rela

tive

to t

he

in

dic

ate

d b

asa

l m

em

be

r, f

or

ea

ch

de

fin

itio

n o

f "l

on

g-liv

ed

."

� 209

Ap

pe

nd

ix B

(co

n't)

Ord

er, (s

ub

ord

er);

b

asa

l m

emb

erN

S, P

SL

-bS

L-t

LL

-bL

L-t

Med

(M

a)

Med

P

rop

ort

ion

Mid

(M

a)

Mid

P

rop

ort

ion

10-g

(M

a)

10-g

P

rop

ort

ion

Rovea

crin

ida

Holo

crin

us

12

S (

1,5

)0.1

70.4

20.1

70.2

518

0.6

18

19

0.6

19

11

0.6

91

Cyrt

ocr

inid

aH

olo

crin

us

27

S (

1,5

)0.2

20.4

40.0

70.2

626

0.0

00

43

0.4

87

88

0.4

71

Com

atuli

da

Holo

crin

us

36

S (

1,5

)0.7

20.0

30.2

50.0

024

0.0

00

49

0.0

00

67

0.2

54

Mil

leri

crin

ida

Holo

crin

us

8S

(1,5

)0.3

70.5

00.0

00.1

310

0.5

00

30

0.7

13

65

0.7

09

Isocr

inid

aH

olo

crin

us

17

S (

1,5

)0.2

90.5

90.0

00.1

237

1.0

00

116

0.7

17

206

0.7

07

Cla

did

aT

itanocr

inus

243

S (

2)

0.9

30.0

10.0

70.0

020

0.0

02

50

0.3

82

41

0.2

88

Gle

nocr

inus

S (

2)

0.9

20.0

20.0

70.0

00.2

46

0.4

16

0.4

38

Per

itto

crin

us

P (

4)

0.9

00.0

40.0

50.0

10.0

00

0.5

90

0.1

51

Elp

aso

crin

us

S (

4)

0.9

30.0

10.0

70.0

00.0

00

0.5

00

0.1

93

Tet

racr

iocr

inus

P (

4)

0.9

30.0

10.0

60.0

00.0

22

0.6

50

0.1

84

Aet

hocr

inus

S (

3),

P(3

, 4)

0.9

30.0

10.0

70.0

00.9

79

0.5

44

0.5

84

� 210

Ap

pe

nd

ix B

(co

n't)

Ord

er, (s

ub

ord

er);

b

asa

l m

emb

erN

S, P

SL

-bS

L-t

LL

-bL

L-t

Med

(M

a)

Med

P

rop

ort

ion

Mid

(M

a)

Mid

P

rop

ort

ion

10-g

(M

a)

10-g

P

rop

ort

ion

-Cyat

hocr

inin

aT

itanocr

inus

36

S (

2)

0.8

60.0

30.1

10.0

014

0.0

00

44

0.0

18

80

,0221

Gle

nocr

inus

S (

2)

0.7

80.1

10.1

10.0

00.0

00

0.6

58

0.6

51

Tet

racr

iocr

inus

P (

4)

0.8

30.0

60.1

10.0

00.0

12

0.8

14

0.8

06

Aet

hocr

inus

S (

3),

P(3

, 4)

0.8

60.0

30.1

10.0

00.0

40

0.8

32

0.8

35

Per

itto

crin

us

P (

4)

0.7

80.1

10.0

60.0

60.0

11

0.9

51

0.9

52

Elp

aso

crin

us

S (

4)

0.8

60.0

30.1

10.0

00.0

00

0.2

45

0.2

43

Tri

bolo

crin

us

S (

6)

0.8

60.0

30.1

10.0

00.0

00

0.4

54

0.4

53

-Den

dro

crin

ina

Elp

aso

crin

us

33

S (

4)

0.4

50.4

50.0

30.0

615

0.4

38

40

0.7

98

41

0.6

56

Tit

anocr

inus

S (

2)

0.4

50.4

50.0

30.0

60.4

91

0.6

40

0.6

58

Gle

nocr

inus

S (

2)

0.4

20.4

80.0

30.0

60.0

00

0.6

39

0.2

31

Tet

racr

iocr

inus

P (

4)

0.6

60.2

40.0

60.0

30.1

61

0.4

33

0.1

95

Aet

hocr

inus

S (

3),

P(3

, 4)

0.3

90.5

10.0

30.0

60.0

00

0.7

89

0.4

81

Per

itto

crin

us

P (

4)

0.6

00.3

10.0

90.0

00.0

03

0.1

14

0.0

70

-Pote

riocr

inin

aP

rom

elocr

inus

173

S (

6)

0.8

70.0

80.0

60.0

020

0.0

00

50

0.1

21

41

0.0

64

Tit

anocr

inus

S (

2)

0.7

30.2

10.0

30.0

20.0

00

0.7

26

0.2

48

Gle

nocr

inus

S (

2)

0.2

00.7

50.0

20.0

41.0

00

0.1

77

0.3

05

Tet

racr

iocr

inus

P (

4)

0.5

50.3

90.0

40.0

20.1

18

0.1

15

0.1

86

Aet

hocr

inus

S (

3),

P(3

, 4)

0.1

80.7

60.0

20.0

41.0

00

0.0

87

0.2

29

Per

itto

crin

us

P (

4)

0.6

50.2

90.0

50.0

10.0

00

0.1

04

0.0

36

Elp

aso

crin

us

S (

4)

0.6

80.2

60.0

30.0

20.0

00

0.6

55

0.1

70

� 211

Ap

pe

nd

ix B

(co

n't)

Ord

er, (s

ub

ord

er);

b

asa

l m

emb

erN

S, P

SL

-bS

L-t

LL

-bL

L-t

Med

(M

a)

Med

P

rop

ort

ion

Mid

(M

a)

Mid

P

rop

ort

ion

10-g

(M

a)

10-g

P

rop

ort

ion

Sag

enocr

inid

aTitanocrinus

57

S (

2)

0.8

40.0

60.0

80.0

214

0.9

80

55

0.8

80

99

0.9

88

Glenocrinus

S (

2)

0.8

40.0

60.0

80.0

20.9

90

0.8

80

0.9

70

Protaxocrinus

S (

3)

0.8

40.0

60.0

80.0

20.8

20

0.8

70

0.9

58

Cupulocrinus

R(3

)0.8

40.0

60.0

80.0

21.0

00

0.8

80

0.9

48

Tax

ocr

inus

Protaxocrinus

11

S (

3)

0.7

30.0

90.1

80.0

012

1.0

00

31

1.0

00

55

0.1

71

Cupulorinus

R(3

)0.7

30.0

90.1

80.0

01.0

00

1.0

00

0.3

61

Titanocrinus

S (

2)

0.7

30.0

90.1

80.0

01.0

00

0.9

80

0.0

57

Glenocrinus

S (

2)

0.7

30.0

90.1

80.0

01.0

00

1.0

00

0.0

53

� 212

Ap

pe

nd

ix B

(co

n't)

Ord

er, (s

ub

ord

er);

b

asa

l m

emb

erN

S, P

SL

-bS

L-t

LL

-bL

L-t

Med

(M

a)

Med

P

rop

ort

ion

Mid

(M

a)

Mid

P

rop

ort

ion

10-g

(M

a)

10-g

P

rop

ort

ion

Dis

par

ida

Tit

anocr

inus

83

S (

2)

0.9

20.0

40.0

50.0

015

0.7

89

38

0.7

27

41

0.8

09

Gle

nocr

inus

S (

2)

0.9

20.0

40.0

50.0

00.7

94

0.7

81

0.8

37

Ibex

ocr

inus

S (

6)

0.9

20.0

40.0

50.0

00.9

46

0.6

83

0.7

73

Hybocr

inus

S (

6)

0.9

00.0

50.0

50.0

00.8

68

0.6

51

0.7

74

Ram

yse

ocr

inus

S (

3)

0.9

00.0

50.0

50.0

00.8

50

0.6

86

0.7

66

Dip

lobat

hri

da

Tit

anocr

inus

41

S (

2)

0.9

00.0

20.0

70.0

011

0.0

00

28

0.0

03

30

0.0

02

Gle

nocr

inus

S (

2)

0.9

00.0

20.0

70.0

00.0

44

0.0

81

0.0

81

Pro

exen

ocr

inus

S (

3)

0.9

00.0

20.0

70.0

00.0

00

0.0

89

0.0

07

Cel

tocr

inus

S (

4)

0.9

00.0

20.0

70.0

00.0

00

0.0

25

0.0

27

Ekn

om

ocr

inus

S (

2)

0.8

80.0

50.0

70.0

00.0

00

0.2

96

0.2

94

Cnem

ocr

inus

S (

2)

0.9

00.0

20.0

70.0

00.0

47

0.0

97

0.0

98

Adel

phocr

inus

S (

2)

0.9

00.0

20.0

70.0

00.0

00

0.0

08

0.0

11

Habro

crin

us

S (

2)

0.8

80.0

50.0

70.0

00.0

00

0.2

74

0.2

74

Monobat

hri

da

Tit

anocr

inus

110

S (

2)

0.9

00.0

20.0

80.0

019

0.0

00

42

0.0

09

40

0.0

26

Gle

nocr

inus

S (

2)

0.9

00.0

20.0

80.0

00.0

00

0.2

64

0.0

09

Pro

exen

ocr

inus

S (

3)

0.9

00.0

20.0

80.0

00.0

00

0.0

36

0.0

24

Cel

tocr

inus

S (

4)

0.9

00.0

20.0

80.0

00.0

00

0.0

12

0.0

67

Ekn

om

ocr

inus

S (

2)

0.9

00.0

20.0

80.0

00.0

00

0.0

75

0.0

36

Cnem

ocr

inus

S (

2)

0.9

00.0

20.0

80.0

00.0

00

0.0

28

0.0

12

Adel

phocr

inus

S (

2)

0.9

00.0

20.0

80.0

00.0

00

0.0

49

0.0

20

Habro

crin

us

S (

2)

0.9

00.0

20.0

80.0

00.0

00

0.0

35

0.0

13

� 213

Appendix C

Periods in geologic history sampled and the morpho-duration plot distributions of the crinoid genera within those periods. N = numberof genera sampled. Med, Mid and 10-g are cutoff points for durations of long-lived taxa as definedin "Materials and Methods." Numbers in the "proportion" columns indicate theproportions of rarefied samples of short-lived genera that are less deviant (or more unspecialized) than long-lived genera, for each definition of "long-lived."

Period NMed (Ma)

Med Proportion

Mid (Ma)

Med Proportion

10-g (Ma)

10-g Proportion

Ordovician 92 24 0.972 9 0.583 30 0.970

Silurian 98 55 0.813 13 1.000 41 0.218

Devonian 99 55 0.498 20 0.290 60 0.411

Carboniferous 236 49 0.536 28 1.000 46 0.606

Permian 77 47 0.999 14 0.865 54 0.998

Triassic – Jurassic 58 116 0.109 27 0.000 71 0.284

Cretaceous – Eocene 69 117 0.205 26 0.000 88 0.414

� 214

Appendix D

Identities o

f lo

ng-liv

ed c

rinoid

genera

in e

ach o

rder

and the o

rders

and fam

ilies to w

hic

h they b

elo

ng.

Dura

tions =

dura

tions o

f th

e g

enera

; M

E =

codes for

the m

ass e

xtinction e

vents

the g

enera

pass thro

ugh: 0 =

none;

1 =

End O

rdovic

ian, 2 =

End D

evonia

n, 3 =

Perm

o-T

riassic

, 4 =

End T

riassic

, 5 =

KT

, 6 =

both

3 a

nd 4

, 7 =

both

4

and 5

. F

am

Dur

= fam

ily d

ura

tions w

here

availa

ble

(S

epkoski 1982; B

ento

n 1

993; update

d u

sin

g W

ebste

r 2003.

AG

DO

and M

GD

O =

avera

ge a

nd m

edia

n g

enus d

ura

tions in the o

rder,

respectively

; A

GD

F a

nd M

GD

F =

avera

ge

and m

edia

n g

enus d

ura

tions in the fam

ilies, re

spectively

. *

Genera

that are

long-liv

ed in m

ore

than o

ne tim

e inte

rval.

� 215

Appendix

D (

con't)

Ord

er, L

on

g-l

ived

Gen

us

Du

rati

on

(M

.y.)

ME

Fam

ily

Fam

Du

r (M

.y.)

AG

DO

(M

.y.)

MG

DO

(M

.y.)

AG

DF

(M

.y.)

MG

DF

(M

.y.)

Roveacrinid

a

Saccocom

a4

20

Saccocom

idae

87

23

18

20

18

Ost

eocrinus

41

0Roveacrinid

ae

47

32

36

Cyrt

ocrinid

a

Cyath

idiu

m8

85

Holo

podid

ae

19

23

43

38

88

8

Gam

maro

crinit

es

89

0Scle

rocrinid

ae

10

75

75

7

Pilo

crinus

89

0Eugenia

crinit

idae

10

75

24

2

Com

atu

lida

Am

phoro

metr

a6

55

Conom

etr

idae

75

35

32

35

35

Sola

nocrinit

es

67

0Sola

nocrinit

idae

11

45

96

2

Com

atu

lina

70

0Sola

nocrinit

idae

11

45

96

2

Hert

ha

78

5A

nte

donid

ae

11

26

66

6

Sem

iom

etr

a1

01

0N

oto

crinid

ae

16

64

03

0

Mill

icrinid

a

Mill

ericrinus

65

0M

illericrinid

ae

65

17

10

24

18

Isocrinid

a

Nie

lsenic

rinus

10

85

Isocrinid

ae

24

56

63

88

35

6

Chla

docrinus

20

67

Isocrinid

ae

24

58

35

6

Iso

crinus

23

67

Isocrinid

ae

24

58

35

6

� 216

Appendix

D (

con't)

Ord

er, L

on

g-l

ived

Gen

us

Du

rati

on

(M

.y.)

ME

Fam

ily

Fam

Du

r (M

.y.)

AG

DO

(M

.y.)

MG

DO

(M

.y.)

AG

DF

(M

.y.)

MG

DF

(M

.y.)

Cla

did

a (

Ord

ovic

ian -

Devonia

n)

27

20

43

26

Decadocrinus

60

0D

ecadocrinid

ae

97

35

35

Lasi

ocrinus

64

2M

ast

igocrinid

ae

84

23

14

Hallo

crinus

80

2Rhenocrinid

ae

10

52

08

Cost

alo

crinus*

84

2Botr

yocrinid

ae

10

02

92

6

Cla

did

a (

Low

er

Carb

onifero

us)

Pariso

crinus

80

2Eusp

irocrinid

ae

13

03

53

5

Cost

alo

crinus*

84

2Botr

yocrinid

ae

10

02

92

1

Cym

bio

crinus

93

0Cym

bio

crinid

ae

10

23

82

1

Gra

phio

crinus*

10

50

Gra

phio

crinid

ae

10

55

54

4

Cla

did

a (

Upperr

Carb

onifero

us)

Cym

bio

crinus*

93

0Cym

bio

crinid

ae

10

23

82

0

Dic

host

reblo

crinus

93

0Str

eblo

crinid

ae

12

87

99

3

Abra

chio

crinus

93

0Codia

crinid

ae

16

84

32

3

Lagenio

crinus

93

0Str

eblo

crinid

ae

12

87

99

3

Gra

phio

crinus*

10

50

Gra

phio

crinid

ae

10

55

54

4

Dis

parida

Haly

siocrinus

60

2Calc

eocrinid

ae

18

91

81

52

71

5

Synchirocrinus

61

0Calc

eocrinid

ae

18

92

71

5

Triacrinus

61

2Pis

ocrinid

ae

87

36

40

Delt

acrinus

81

2Calc

eocrinid

ae

18

92

71

5

� 217

Appendix

D (

con't)

Ord

er, L

on

g-l

ived

Gen

us

Du

rati

on

(M

.y.)

ME

Fam

ily

Fam

Du

r (M

.y.)

AG

DO

(M

.y.)

MG

DO

(M

.y.)

AG

DF

(M

.y.)

MG

DF

(M

.y.)

Dip

lobath

rida

Dim

ero

crinit

es

55

0D

imero

crinit

idae

75

15

11

18

14

Gilb

ert

socrinus

60

2Rhodocrinit

idae

14

21

81

3

Monobath

rida

Cam

pto

crinus

79

0D

ichocrinid

ae

10

22

41

83

33

4

Megis

tocrinus

89

2Periechocrinid

ae

99

32

26

Acti

nocrinit

es

93

0A

cti

nocrinit

idae

14

02

91

8

Sagenocrinid

a

Eury

ocrinus

60

2Eury

ocrinid

ae

97

22

14

29

24

Cib

olo

crinus

63

0M

espilo

crinid

ae

10

22

81

4

Clid

ochirus

11

52

Icth

yocrinid

ae

12

17

57

5

Taxonocrinid

a

Taxocrinus

67

2Taxocrinid

ae

10

52

41

83

42

3

Pro

taxocrinus

46

1Taxocrinid

ae

10

53

42

3

� 218

Appendix J

Table of resampled correlations

This table shows the frequency with which relationships are not significantly correlated

after rarefaction. The column "rarified" indicates the number of records randomly

sampled for each species, while the column "qualifying" indicates the number of records

a species must at least have before it is used for calculations. Lat = frequency with which

the relationship between latitudinal range and duration is not significant, long = the

same for longitudinal range and duration, N= sample sizes (of qualifying data). Grey

boxes highlight those cases that retain the relationship shown in the dataset as a whole.

SPECIES GENUS

Rarified Qualifying Lat Long N Lat Long N

2 2 0.07 0.00 2044 0.71 0.52 266

2 3 0.18 0.03 1075 0.79 0.73 206

3 3 0.03 0.00 1075 0.77 0.25 206

2 4 0.22 0.13 751 0.65 0.71 189

3 4 0.07 0.04 751 0.12 0.23 189

4 4 0.00 0.00 751 0.01 0.08 189

2 5 0.25 0.49 494 0.50 0.84 154

3 5 0.12 0.37 494 0.47 0.81 154

4 5 0.01 0.22 494 0.34 0.83 154

5 5 0.00 0.05 494 0.46 0.96 154

2 6 0.57 0.90 362 0.98 0.97 132

3 6 0.33 0.93 362 0.94 1.00 132

4 6 0.23 0.93 362 0.97 0.99 132

5 6 0.14 0.99 362 1.00 1.00 132

6 6 0.01 1.00 362 1.00 1.00 132

2 7 0.48 0.90 260 1.00 1.00 115

3 7 0.41 0.88 260 0.97 0.99 115

4 7 0.17 0.93 260 1.00 1.00 115

5 7 0.08 0.97 260 1.00 1.00 115

6 7 0.08 1.00 260 0.98 1.00 115

7 7 0.04 1.00 260 1.00 1.00 115

2 8 0.61 0.86 187 0.96 1.00 94

3 8 0.41 0.92 187 0.93 0.98 94

4 8 0.24 0.96 187 0.96 1.00 94

5 8 0.30 0.98 187 0.96 1.00 94

6 8 0.07 0.99 187 1.00 1.00 94

7 8 0.07 0.99 187 1.00 1.00 94

8 8 0.06 1.00 187 1.00 1.00 94

� 219

Appendix K

Character matrices for trachyleberidid ostracode species

Character matrices for trachyleberidid ostracode species of 4 genera used in analyses in

Chapter V, including their first and last appearances in the fossil record and references

used in coding their morphologies. See Appendix L for character state descriptions.

� 220

Curfsina

species

FA

LA

1234567Reference

aequabilis

K(C

oni-

u)

K(C

oni-

u)

01

2?

02

0H

err

ig 1

968

alseni

K(C

am

p-u

)K(C

am

p-u

)0

13

00

01

Cla

rke 1

983

anorc

hid

ea

K(M

aes-

l)K(M

aes-

u)

20

32

22

1Cla

rke 1

983

colin

iK(A

lbi)

K(A

lbi)

23

30

21

0Ja

in 1

978

com

munis

K(Camp-l)

K(M

aes-

u)

21

00

21

1Is

raels

ky 1

929/C

rane 1

965

decora

taK(C

eno-m

)K(T

uro

-l)

22

00

21

0D

onze &

Thom

el 1972

delic

ate

orn

ata

K(S

ant)

K(S

ant)

22

10

21

1Andre

u 1

995

dero

oi

K(C

eno-m

)K(C

eno-u

)0

53

02

11

Weaver

1982

fauja

siK(M

aes-

l)K(M

aes-

u)

12

10

01

1Cla

rke 1

983

flexuosa

K(T

uro

-u)

K(S

ant-

u)

20

02

22

1O

ert

li 1

985/

Babin

ot

1980

gele

enensi

sPg(Dan-u)

Pg(Dan-u)

?1

10

02

1D

ero

o 1

966

infr

agili

s aff

.Pg(T

ha)

Pg(T

ha)

20

1?

20

0O

ert

li 1

985

kafk

ai

K(T

uro

-m)

K(C

oni-u)

00

12

21

1Pokorn

y 1

967

levig

ata

K(Camp-m)K(Camp-u)

20

12

00

0Bate

1972

maio

rK(M

aes-

u)

K(M

aes-

u)

?3

31

22

1D

ero

o 1

966

mira

K(S

ant-

l)K(S

ant-

u)

12

?1

22

0Babin

ot

1980

monzie

nsi

sK(Camp-u)

K(Camp-u)

10

10

21

0D

ingle

1981

mucro

nata

K(C

eno-m

)K(C

eno-u

)1

00

11

20

Babin

ot

et

al. 1

978/

Colin 1

973

neale

iK(C

eno-l)

K(C

eno-l)

12

30

?1

0Sw

ain

and X

ie 1

991

nuda

K(A

lbi-m

)K(C

eno-u

)2

23

02

01

Bassio

uni 2001

orc

hid

ea

K(M

aes-

u)

K(M

aes-

u)

?3

01

21

?D

ero

o 1

966

parv

a

K(C

am

p-u

)K(M

aes-

l)0

21

02

11

Pokorn

y 1

967

quadrisp

inata

K(T

uro

-l)

K(M

aes-

u)

04

10

20

0Cla

rke 1

983

regin

ae-a

strid

K(M

aes-

u)

K(M

aes-

u)

?0

11

21

1D

ero

o 1

966

sast

ryi

K(M

aes-

l)K(M

aes-

m)

10

02

?0

0M

allik

arj

una a

nd N

agara

ja 1

996

senior

K(C

eno-u

)K(T

uro

-m)

12

12

20

1Pokorn

y 1

967

subparv

a

K(C

eno)

K(C

oni-u)

04

20

21

1Pokorn

y 1

967

turo

nic

aK(T

uro

)K(T

uro

)0

31

11

21

Bate

and B

ayliss 1

969

ventr

oconcava

K(C

eno-l)

K(C

eno-m

)1

3?

?2

20

Colin a

nd D

am

ott

e 1

985

� 221

Opim

ocyth

ere

species

FA

LA

12345678910111213141516Reference

browni

Pg(D

an)

Pg(D

an)

12

24

11

44

11

10

12

12

29

Hazel 1968

betzi

T(P

ale

-u)

T(E

o-u

)2

22

4?

??

??

11

01

1?

?Je

nnin

gs 1

936

elonga

Pg(D

an)

Pg(D

an)

20

24

01

57

12

10

12

15

30

Hazel 1968

gigante

aT(E

o-u

)T(E

o-u

)2

22

71

24

??

10

10

0?

?Puri 1

957

hazeli

Pg(D

an)

Pg(D

an)

10

24

01

??

?1

10

11

??

Sm

ith 1

978

incisa

Pg(T

ha)

Pg(T

ha)

22

0?

?1

4?

?1

01

11

??

Oert

li 1

985

inte

rrasilis

T(P

ale

-l)

T(P

ale

-l)

12

14

0?

?3

?0

10

01

??

Hazel 1968/A

lexander

1934

jessupensis

T(P

ale

) T(P

ale

-u)

02

0?

01

2?

??

10

00

?20

Murr

ay &

Hussey 1

942

martini

T(E

o-m

)T(E

o-u

)1

21

41

0?

6?

11

10

015

28

Murr

ay &

Hussey 1

942

marylandica

T(P

ale

-u)

T(E

o-l

)2

11

30

16

8?

11

00

015

25

Ulric

h 1

901

miocenica

T(M

i-m

-u)

T(M

i-m

-u)

10

0?

02

?8

?1

11

11

??

Puri 1

953

mississippiensis

T(E

o-u

)T(E

o-u

)2

21

40

2?

15

21

11

01

??

Kru

tak 1

961

nanafaliana

T(P

ale

-u)

T(P

ale

-u)

20

14

01

??

?1

11

0?

10

25

Murr

ay &

Hussey 1

942

taxyae

K(C

eno-m

)K(C

eno-u

)0

10

?0

??

??

10

00

??

?Babin

ot

1970

texana

K(A

lbi-

u)

K(A

lbi-

u)

02

??

0?

??

?1

11

02

??

Hazel 1968

ventroinflata

K(M

aes-u

)Pg(D

an-l

)1

12

4?

1?

?0

?1

0?

??

?Savelieva 2

001

verrucosa

Pg(D

an)

Pg(D

an)

21

24

0?

??

01

11

02

??

Hazel 1968

� 222

Phalc

ocyth

ere

species

FA

LA

12345678Reference

bireticulata

T(P

ale

-m)

T(P

ale

-u)

21

00

00

?0

Nagori 1

993

budakesz

iensis

Pg (

Pri-l

)Pg (

Pri-l

)1

41

11

21

1M

onosto

ri 1

996

bullita

T(P

ale

-l-l

)T(P

ale

-l-u

)1

32

01

1?

1Al-

Fura

ih 1

980

coelops

T(P

ale

-l-l

)T(P

ale

-l-m

)1

32

01

1?

1Al-

Fura

ih 1

980

conifera

T(P

ale

-m-l

)T(P

ale

-u-m

)23

10

00

?1

Al-

Fura

ih 1

980

cultrata

K(M

aes-u

)T(E

o-l

-u)

23

11

00

?1

Reym

ent

1983

cuneata

T(P

ale

-m)

T(P

ale

-m)

13

20

01

?0

Al-

Fura

ih 1

980

disse

nta

T(E

o-l

-u)

T(E

o-u

)1

22

11

0?

2Sid

diq

ui 1971

fluxilis

T(P

ale

-l-u

)T(P

ale

-l-u

)2

22

00

1?

1Al-

Fura

ih 1

980

hebes

K(M

aes-u

)T(P

ale

-l-m

)2

22

11

22

1Al-

Fura

ih 1

980

horraensis

T(P

ale

-u)

T(E

o-l

-m)

22

10

00

?1

Bassio

uni and M

ors

i 200

horresc

ens

Pg(Y

pr)

T(O

l-m

) 1

42

11

22

1Sid

dqui 1978/M

onosto

ri1996

improcera

T(P

ale

-u)

T(P

ale

-u)

03

21

12

21

Al-

Fura

ih 1

980

inte

rcalata

T(P

ale

-l-u

)T(P

ale

-l-u

)2

32

00

0?

1Al-

Fura

ih 1

980

mohani

T(P

ale

-u)

T(P

ale

-u)

14

00

12

10

Bhandari1996

nulllicosta

taT(P

ale

-l-m

)T(P

ale

-m-m

)23

21

10

?2

Al-

Fura

ih 1

980

postc

orn

isT(P

ale

-l)

T(P

ale

-l)

13

20

01

22

Al-

Fura

ih 1

983

recta

ngula

ris

K(M

aes-u

)T(P

ale

-u-l

)1

32

11

22

1Al-

Fura

ih 1

980

rete

T(P

ale

-l-m

)T(P

ale

-u)

13

20

00

?1

Al-

Fura

ih 1

980

retispin

ata

T(P

ale

-u)

T(P

ale

-u)

14

11

12

21

Bhandari 1

996/S

ohn 1

980

sento

sa

T(P

ale

-u)

T(E

o-l

-l)

13

11

01

?1

Sid

diq

ui 1971

spin

osa

T(E

o-u

-l)

T(E

o-u

-l)

14

01

12

21

Sid

diq

ui 1971

spin

osa (

cf.

)T(E

o-u

)T(E

o-u

)1

42

11

1?

2Sid

diq

ui 1971

subtilis

T(P

ale

-m)

T(P

ale

-m)

23

20

00

?1

Al-

Fura

ih 1

980

sum

igensis

Pg(L

ut-

l)Pg(L

ut-

l)1

40

11

22

1M

onosto

ri 1

996

tokars

kii

T(O

l)T(O

l)2

11

10

0?

0Bla

szyk 1

985

tranquilis

T(E

o-l

-u)

T(E

o-l

-u)

10

00

00

?1

Bassio

uni and M

ors

i 200

tranquilis

T(P

ale

-l-l

)T(P

ale

-l-u

)1

12

10

0?

2Al-

Fura

ih 1

980

tubra

T(P

ale

)T(P

ale

)1

42

11

22

1Carb

onnel et

al.1990.

/Reym

ent

1983

vesic

ulo

sa

K(M

aes-u

)T(E

o-l

)2

31

?1

1?

0Carb

onnel et

al.1990/

Reym

ent

1983

� 223

Schizoptocythere

species

FA

LA

123456789Reference

circum

spin

osa

T(P

ale

)T(E

o-m

)2

?3

22

10

10

Sid

diq

ui and A

l-Fura

ih 1

981

com

pre

ssa

K(S

ant-

l)K(C

am

p-l

)2

03

30

01

00

Puckett

1996

how

ei

T(P

ale

-u)

T(E

o-l

)1

12

21

11

00

Sid

diq

ui and A

l-Fura

ih 1

981

how

ei

T(P

ale

-u)

T(P

ale

-u)

11

11

01

11

0Sid

diq

ui and A

l-Fura

ih 1

981

lisso

sT(P

ale

-u)

T(P

ale

-u)

2?

01

10

00

1Sid

diq

ui and A

l-Fura

ih 1

981

mis

hra

iT(E

o-l

-l)

T(E

o-l

-l)

0?

21

12

02

0Bhandari 1

996

paulia

bom

nin

ata

T(E

o-m

)T(E

o-m

)2

13

30

00

10

Bassio

uni and L

uger

1990

segura

iK(M

aes)

K(M

aes)

1?

21

22

02

0H

azel and K

am

iya 1

993

sim

opyge

T(E

o-m

)T(E

o-m

)1

?2

12

11

20

Sid

diq

ui and A

l-Fura

ih 1

981

singhi

T(P

ale

-u)

T(P

ale

-u)

11

32

11

02

0Bhandari 1

996

tauru

sT(E

o-l

)T(E

o-l

)1

?2

11

10

10

Shahin

2000

tem

pera

taT(E

o-l

)T(E

o-l

)1

?2

10

10

10

Sid

diq

ui and A

l-Fura

ih 1

981

torq

uata

T(P

ale

-l)

T(P

ale

-l)

2?

21

22

01

0Sid

diq

ui and A

l-Fura

ih 1

981

usm

andanto

dio

iK(M

aes)

T(E

o-l

-u)

21

32

01

02

0Bassio

uni and M

ors

i 2000/R

eym

ent

1983

ventr

icosa

T(E

o-m

)T(E

o-m

)1

?3

22

11

20

Sid

diq

ui and A

l-Fura

ih 1

981

ventr

onodosa

T(E

o-l

)T(E

o-l

)1

?1

11

10

00

Sohn 1

970

� 224

Appendix L

Short descriptions of morphological characters of trachyleberidid species

See Appendix K for character matrices

Curfsina characters

1.Length: 0 = small (< 0.6 mm); 1= medium (0.60 - 0.75 mm); large (> 0.75 mm)

2.Lateral ornaments: 0 = smooth; 1 = pits; 2 = fine reticulation; 3 = reticulation; 4 =

coarse reticulation; 5 = tubercles

3.Central node: 0 = ridge; 1 = node and ridge; 2 = node and no ridge; 3: two nodes

4.Median ridge connected posteriorly with dorsal ridge: 0 = not connected; 1 =

somewhat connected; 2 = connected

5.Anterior ridge and ventral ridge connected anteriorly: 0 = not connected; 1 =

somewhat connected; 2 = connected

6.Pointed posterior: 0 = absent; 1 = moderate; 2 = pronounced

7.Spines: 0 = absent; 1 = present

� 225

Opimocythere characters

1.Length: 0 = small (< 0.9 mm); 1 = median (0.90 - 1.00 mm); large ( > 1.0 mm)

2.Ventral rib: 0 = weak; 1 = moderate; 2= strong

3.Posterior spines: 0 = very short; 1 = short; 2 = long

4.No. posterior spines

5.Ventral rib ends in spine: 0 = no; 1 = yes

6.Extra ribs neighbouring ventral rib: 0 = no; 1 = below; 2 = above

7.No. extra ribs neighboring ventral rib

8.Pores in anterocentral area: 0 = no; 1 = yes

9.Coarse pits in central area: 0 = no; 1 = variable size; 2 = yes

10.General ornamentation: 0 = pits; 1 = reticulate; 2 = pustulose

11.Anterior cardinal angle: 0 = > 150 deg; 1 = < 150 deg

12.Dorsal angle: 0 = > 150 deg; 1 = < 150 deg

13.Furrow behind anterior rim: 0 = absent; 1 = present

14.Eye spot: 0 = absent; 1 = present; 2 = very prominent

15.No. posterior pore canals

16.No. anterior pore canals

� 226

Phalcocythere characters

1.Length: 0 = small (< 0.50 mm); 1 = medium (0.50 - 0.70 mm); 2 = large (> 0.70 mm)

2.Ornamentation: 0 = smooth; 1 = fine reticulations; 2 = reticulations; 3= knobby; 4 =

spiny

3.Concentric patterning on lateral view: 0 = absent; 1 = weak; 2= strong

4.Ventral ridge: 0= weak; 1 = strong

5.Spines on reticulation: 0 = absent; 1 = present

6.Posterodorsal spines: 0 = absent; 1 = short; 2 = long

7.No. spines

8.Subcentral tubercle: 0 = weak; 1 = strong

Schizoptocythere characters

1.Length: 0 = small (< 0.5 mm); medium (0.50 - 0.70 mm); large (> 0.7mm)

2.Terminal hinge element: 0 = smooth; 1 = crenulate

3.Marginal spines: 0 = none; 1 = small; 2 = knobby

4.No. long spines

5.Subcentral node: 0 = absent; 1 = weak; 2 = strong

6.Lateral ornamentation: 0 = smooth; 1 = knobby; 2 = very knobby

7.Posterior margin: 0 = not upturned; 1 = upturned

8.Tubercle posterior to eye tubercle: 0 = absent; 1 = weak; 2 = strong

9.Postocular depression: 0 = absent; 1= strong� 227

Appendix M

References used in Appendix K

Alexander, C. I. 1934. Ostracoda of the Midway (Eocene) of Texas. J. Paleont. 8:206

-237.

Al-Furaih, A. A. F. 1980. Upper Cretaceous and Lower Tertiary Ostracoda Superfamily

Cytheracea from Saudi Arabia. University Libraries, University of Riyadh.

Al-Furaih, A. A. F. 1983. Paleocene and Lower Eocene Ostracoda from the Umm Er

Radhuma Formation of Saudi Arabia. Univ. Kansas Paleont. Contrib. 107:1-9.

Andreu, B. 1995. Trachyleberididae (Ostracodes) du Turonien Superieur (?) - Santonien

de la region de Boulmane, Moyen Atlas (Maroc): Systematique et biostratigraphie.

Rev. Esp. Micropal. 27:85-142.

Babinot, J. F. 1970. (Part1) Nouvelles especes d'ostracodes du Cenomanien superieur de

l'aureole Septemtrionale du Bassin du Beausset (Bouches-du-Rhone-var) (1re

partie). Rev. Micropaleont. 13:95-106.

Babinot, J. F. 1980. Les Ostracodes du Cretace superieur de Provence. Natural Sciences.

University of Provence.

Babinot, J. F., P. Y. Berthou, J. -P. Colin and J. Lauverjat 1978. Les Ostracodes du

Cenomanien du Bassin Occidental Portugais; Biostratigraphie et affinites

Paleogeographiques. Cahiers Micropal. 1978(3):11-31

� 228

Bassiouni, M. A. 2002. Mid-Cretaceous (Aptain- Early Turonian) Ostracoda from Sinai

Egypt. Neue Palaeont. Abhand. 5:1-123.

Bassiouni, M. A. A., and P. Luger. 1990. Maastrichtian to early Eocene Ostracoda from

Southern Egypt. Palaeontology, Palaeoecology, Palaeobiogeography and

Biostratigraphy. Berliner Geowiss. Abhand. A 120:775-928.

Bassiouni, M. A., and A. M. M. Morsi. 2000. Paleocene-Lower Eocene ostracodes from

El Quss Abu Said Plateau (Farafra Oasis), Western Desert Egypt. Paleontograph.

Abteil. A 257:27-84.

Bate, R. H. 1972. Upper Cretaceous Ostracoda from the Carnorvon Basin Western

Australia. Special Papers in Palaeontology 10:1-85, 27 Plates.

Bate, R. H. and Bayliss, D. D. 1969. An outline account of the Cretaceous and Tertiary

Foraminifera and of the Cretaceous ostracods of Tanzania. Proc. 3rd African

Micropal. Coll., Cairo 1968 pp. 113-164.

Bhandari, A. 1996. Atlas of Paleogene ostracodes of Rajasthan Basins. Paleontograph.

Indica 4:1-157.

Blaszyk, J. 1987 Ostracods from the Oligocene Polonez Cove Formation of King

George Islands, west Antarctica. Palaeont. Pol. 49:63-81.

Carbonnel, G., K. Alzouma, and M. Dikouma. 1990. Les Ostracodes Paleocene du

Niger: Taxinomie - un temoignage de L' existence Eventuelle de la mer

Transsahareinne? Geobios 23:671-697.

� 229

Clarke, B. 1983. Die Cytheracea (Ostracoda) im Schrebkreide-Richprofil von

Laegerdorf-Kronsmoor-Hemmoor (Coniac bis Maastricht; Nord-deutchland). Mitt.

Geol. Palaeont. Inst. Uni. Hamburg 54:65-168.

Colin, J.-P. 1973. Nouvelle Constribution a L'etude des ostracdes du Cretace Superieur

de Dordogne (S.O. France). Palaeontograph. Abteil. A 145:1-38.

Colin, J.-P., and R. Damotte. 1985. Les Ostracodes du Cretace Superieur de l'Autoroute

A 10 (Charente, S. O. France). Cret. Res. 6:157-173.

Crane, M. J. 1965. Upper Cretaceous ostracodes of the Gulf Coast area.

Micropaleontology 11:191-254.

Deroo, G. 1966. Cytheracea (Ostracodes) du Maastrichtien de Maastricht (Pays-Bas) de

des regions voisines; resultats stratigraphiques et paleontologieques de leur etude.

Med. Geol. Sticht. Serie C 2:1-197 , + plates.

Dingle, R. V. 1981. The Campanian and Maastraichtian Ostracoda of South East Africa.

Ann. S. Afr. Mus. 85:1-181.

Donze, P., and G. Thomel. 1972. Le Cenomanien de la Foux (Alpes de Haute-

Provence): Biostratigraphie et Faunes Nouvelles d' Ostracodes. Ecologae

Geologicae Helvetiae 65:369-389.

Hazel, J. E. 1968. Ostracodes from the Brighteseat Formation (Danian) of Maryland. J.

Paleont. 42:100-142.

Hazel, J. E., and T. Kamiya. 1993. Ostracode biostratigraphy of the Titanosarcolites-

bearing limestones and related sequences of Jamaica. GSA Mem. 182:65-76.� 230

Herrig, E. 1968. Biotaxonomische Untersuchungen an Oberkreide-Ostracoden.

Wissenchaftliche Zeitschrift de ernst-Moritz-Arndt-Universitaet Greifswald 17:99

-114

Israelsky, M. C. 1929. Upper Cretaceous Ostracoda of Arkansas. Bull. Ark. Geol.Surv.

2(Suppl):1-29.

Jain, S. P. 1978. Futher Ostracoda from the Kallakkudi Limestone (Albian)

Tiruchirapalli district, Tamil Nadu, India. Neues Jahrbuch Fur Geologie Und

Palaontologie-Monatschaft 1978:502-512.

Jennings, S. 1936. A microfauna from the Monmuth and Basal Rancocos of New Jersey.

Bull. Am. Paleont. 23:161-234.

Krutak, P. R. 1961. Jackson Eocene Ostracoda from the Cocoa Sand of Alabama. J.

Paleont. 35:769-788.

Mallikarjuna, U. B., and H. M. Nagaraja. 1996. Ostracodes from the Ariyalur Group

(Late Cretaceous), Cauvery Basin, Southern India. J. Geol. Soc. India 48:189-201.

Monostori, M. 1996. Eocene Ostracods of Hungary, Cytheracea 1. Annales Universitatis

Scientiaum Budapestinensis de Rolando Eotvos Nominatae Sectio geologica 31:27

-74.

Murray, G., Jr., and K. Hussey. 1942. Some Tertiary Ostracoda of the genera

Alatacythere and Brachycythere. J. Paleont. 16:164-182.

Nagori, M. L. 1993 Early Tertiary Ostracodes from the Pondicherry Formation,

Pondicherry. J. Geol. Soc. India 42:569-577� 231

Oertli, H. J. 1985. Atlas des ostracodes de France (Paléozoïque-actuel). Bulletin des

centres de recherches Exploration-production Elf-Aquitaine. Mémoire 9.

Pokorny, V. 1967. The genus Curfsina (Ostracoda, Crustacea) from the upper

Cretaceous of Bohemia, Czechoslovakia. Acta Universitatis Carolinae - Geologica

4:345-364.

Puckett, T. M. 1996. Ecologic Atlas of Upper Cretaceous Ostracodes of Alabama.

Geological Survey of Alabama Monograph 14.

Puri, H. S. 1953. The ostracode genus Hemicythere and its allies. J. Wash. Acad. Sci.

43:169-179.

Puri, H. S. 1957. Stratigraphy and zonation of the Ocala Group. Florida Geol. Surv.

Geol. Bull. 38:185-248.

Reyment, R. A. 1983. The Ostracoda of the Kalambaina Formation (Paleocene),

northwestern Nigeria. Bulletin of the Geological Institutions of the University of

Uppsala 9:51-65.

Savelieva, Y. N. 2001. New Ostracodes from the Cretaceous and Paleogene boundary

sediments in southwestern Crimea. Paleont. J. 35:166-172.

Shahin, A. 2000. Tertiary ostracods of Gebel Withr, southwestern Sinai, Egypt:

palaeontology, biostratigraphy and palaeobiogeography. J. Afr. Earth Sci. 31:285

-315.

Siddiqui 1971. Early Tertiary Ostracoda of the family Trachyleberididae from West

Pakistan. Bull. Brit. Mus. Nat. Hist. Geol. Series Suppl. 9:1-98. Plates 1-42.� 232

Siddiqui, Q. A., and A. A. F. Al-Furaih. 1981. A new trachyleberid ostracod genus from

the early Tertiary of Western Asia. Palaentology 24:877-890.

Siddiqui, Q. A. 1978. On Phalcocythere horrescens (Bosquet). Stereo-Atlas of Ostracod

Shells 5: 117-120

Smith, J. K. 1978. Ostracoda of the Prairie Bluff Chalk, Upper Cretaceous,

(Maestrichtian) and the Pine Barren Member of the Clayton Formation, Lower

Paleocene (Danian) from exposures along Alabama State highway 263 in Lowndes

County, Alabama. Trans. Gulf Coast Assoc. Geol. Soc. 28:539-379.

Sohn, I. G. 1970. Early Tertiary ostracodes from West Pakistan. Mem. Geol. Surv.

Pakistan 3:1-91.

Swain, F. M., and C.-L. Xie. 1991. Jurassic and Lower Cretaceous Ostracoda from cost

Atlantic Wells, Western North Atlantic Ocean. Rev. Esp. Micropal. 23:57-98.

Ulrich, E. O. 1901. Systematic Paleontology: Arthropoda: Ostracoda. Pp. 116-122.

Maryland Geological Survey Eocene. John Hopkins Press.

Weaver, P. P. E. 1982. Ostracoda from the British Lower Chalk and Plenus Marls.

Mono. Palaeontograph. Soc. Pp. 1-127.

� 233

Appendix O

Correlations between morphology and longevity

Results table presenting Kendall's rank order correlation and their respective

probabilities for analyses described in the text for datasets from various references as

listed. The second column (TYPE) indicates whether the morphological mode or mean

was used and whether the duration was inferred (I) or not (UN) and whether

durationswere measured in millions of years (unlabelled), Stages (S) or manually

measured (M). Highlights are probabilites significant at the p = 0.05 level. See text for

more details.

� 234

Appendix O (con't)

GROUP

Reference TYPE TAU p(g) TAU p(g, m) TAU p(g,pco) TAU p(g,m,pco)

Adnet and Capetta 2001 MODE-I 0.170 0.256 0.152 0.377 0.236 0.115 0.190 0.293

MODE-UN 0.376 0.012 0.473 0.019 0.409 0.006 0.407 0.047

MEAN-I 0.141 0.347 0.138 0.423 0.225 0.132 0.170 0.323

MEAN-UN 0.365 0.015 0.516 0.010 0.414 0.006 0.473 0.019

Adrian and Westrop 2001 MODE-I 0.299 0.176 0.667 0.333 -0.088 0.691 -0.333 0.750

MODE-UN 0.062 0.778 1.000 0.333 -0.187 0.397 0.333 1.000

MEAN-I 0.035 0.873 0.548 0.264 -0.088 0.691 -0.333 0.750

MEAN-UN -0.335 0.129 -0.333 1.000 -0.270 0.221 0.333 1.000

Adrain and Edgecomb 1997 MODE-I-M 0.036 0.777 -0.067 0.729 0.166 0.190 0.038 0.842

MODE-UN-M -0.093 0.464 -0.219 0.297 -0.020 0.873 -0.077 0.765

MEAN-I-M -0.266 0.036 -0.219 0.282 0.204 0.107 0.086 0.697

MEAN-UN-M -0.317 0.012 -0.452 0.032 0.014 0.915 -0.103 0.675

MODE-I-S -0.329 0.009 0.000 1.000 0.044 0.728 0.333 0.750

MODE-UN-S -0.414 0.001 -0.333 1.000 -0.313 0.013 0.333 1.000

MEAN-I-S -0.476 0.000 0.000 1.000 0.060 0.633 0.333 0.750

MEAN-UN-S -0.510 0.000 -0.816 0.201 -0.313 0.013 0.333 1.000

Allmon 1996 (Table 1) MODE 0.424 0.000 -0.240 0.278 0.240 0.013 0.290 0.189

MEAN 0.320 0.001 0.443 0.045 0.222 0.021 0.303 0.197

Allmon 1996 (Table 9) MODE -0.549 0.000 -0.494 0.008 -0.285 0.014 -0.267 0.165

MEAN -0.491 0.000 -0.460 0.013 -0.294 0.012 -0.310 0.094

Alroy 1995 MODE -0.588 0.001 -0.333 1.000 0.177 0.321 1.000 0.333

MEAN 0.250 0.038 1.000 0.333 0.058 0.629 0.333 1.000

Alvarez et al. 1998 MODE-I 0.243 0.037 0.208 0.164 -0.186 0.111 -0.043 0.794

MODE-UN 0.289 0.013 0.187 0.201 -0.166 0.155 -0.113 0.445

MEAN-I 0.252 0.031 0.154 0.319 -0.207 0.076 -0.138 0.373

MEAN-UN 0.200 0.087 0.187 0.191 -0.179 0.125 -0.127 0.391

Amati and Westrop 2004 MODE-M -0.246 0.142 -0.044 0.826 0.209 0.211 0.175 0.384

MEAN-M -0.223 0.182 -0.133 0.509 -0.168 0.314 -0.033 0.868

MODE-S -0.498 0.004 -0.619 0.069 0.147 0.395 -0.238 0.562

MEAN-S -0.310 0.072 -0.619 0.069 0.161 0.351 -0.238 0.562

Anderson and Roopnarine 2003 MODE -0.371 0.093 -0.467 0.272 -0.436 0.049 -0.600 0.136

MEAN -0.673 0.002 -0.690 0.052 -0.402 0.069 -0.733 0.056

Angielczky & Kurkin 2003 MODE-I -0.600 0.000 -0.400 0.483 -0.593 0.000 -0.400 0.483

MODE-UN -0.549 0.001 -0.333 0.750 -0.431 0.008 -0.667 0.333

MEAN-I -0.521 0.001 -1.000 0.017 -0.601 0.000 -0.400 0.483

MEAN-UN -0.449 0.006 -0.333 0.750 -0.416 0.010 -0.333 0.750

Bloch et al. 2001 MODE-I 0.207 0.303 -0.333 0.750 -0.491 0.014 -1.000 0.083

MODE-UN -0.757 0.000 -0.333 1.000 -0.693 0.001 -0.333 1.000

MEAN-I -0.362 0.071 -0.667 0.333 -0.569 0.005 -1.000 0.083

MEAN-UN -0.877 0.000 -1.000 0.333 -0.693 0.001 -0.333 1.000

Bodenbender and Fischer 2001 MODE-I 0.005 0.952 -0.467 0.272 0.493 0.000 0.733 0.056

MODE-UN -0.007 0.930 -0.467 0.272 0.496 0.000 0.733 0.056

MEAN-I 0.386 0.000 0.333 0.469 0.493 0.000 0.733 0.056

MEAN-UN 0.400 0.000 0.200 0.719 0.492 0.000 0.733 0.056

Brochu1997 MODE -0.195 0.026 -0.155 0.238 -0.281 0.001 -0.357 0.007

MEAN -0.178 0.043 -0.117 0.372 -0.303 0.001 -0.341 0.009

Brunet-Lecomte&Chaline 1990 MEAN -0.526 0.005 -0.557 0.012 -0.685 0.000 -0.636 0.004

Cairns 2001 MODE -0.618 0.000 -0.494 0.006 -0.430 0.001 -0.574 0.001

MEAN -0.377 0.003 -0.267 0.134 -0.412 0.001 -0.529 0.002

� 235

Appendix O (con't)

GROUP


Caron et al. 2004 MODE -0.279 0.296 -0.200 0.817 0.446 0.094 0.400 0.483

MEAN 0.063 0.814 0.400 0.483 0.446 0.094 0.400 0.483

Dashzeveg and Meng 1998 MODE-I -0.432 0.015 -0.429 0.239 0.095 0.596 -0.048 1.000

MODE-UN -0.468 0.009 0.000 1.000 0.449 0.012 1.000 0.083

MEAN-I -0.527 0.003 -0.619 0.069 0.102 0.567 -0.143 0.773

MEAN-UN -0.299 0.093 0.000 1.000 0.468 0.009 1.000 0.083

Damiani et al. 2001 MODE-I -0.281 0.074 -0.333 0.381 -0.244 0.122 -0.333 0.381

MODE-UN -0.448 0.005 0.143 0.773 -0.412 0.009 0.238 0.562

MEAN-I -0.228 0.149 -0.333 0.381 -0.238 0.131 -0.333 0.381

MEAN-UN -0.406 0.010 0.333 0.381 -0.400 0.011 0.238 0.562

Dewing 2004 MODE -0.412 0.122 -0.200 0.719 -0.568 0.033 -0.552 0.120

MEAN -0.412 0.122 -0.467 0.272 -0.508 0.057 -0.552 0.120

Ebbestad & Budd 2002 MODE -0.418 0.024 -0.667 0.333 -0.311 0.093 0.333 0.750

MEAN -0.377 0.042 -0.667 0.333 -0.351 0.058 0.000 1.000

Forey 1991 MODE 0.415 0.001 -0.286 0.399 0.304 0.016 0.255 0.378

MEAN 0.324 0.010 0.000 1.000 0.297 0.019 0.255 0.378

Froelich 2002 MODE -0.300 0.135 -0.310 0.212 -0.237 0.237 -0.200 0.484

MEAN -0.487 0.015 -0.484 0.052 -0.237 0.237 -0.244 0.381

Gahn and Kammer 2002 MODE -0.484 0.052 -0.467 0.272 -0.677 0.006 -0.867 0.017

MEAN -0.726 0.003 -0.828 0.020 -0.611 0.014 -0.867 0.017

Grande and Bemis 1998 MODE 0.140 0.362 -0.333 0.469 0.178 0.245 -0.200 0.719

MEAN 0.166 0.281 -0.067 1.000 0.247 0.107 0.467 0.272

Hopkins 2004 MODE-I -0.197 0.459 -0.143 0.720 -0.254 0.341 -0.143 0.720

MODE-UN 0.485 0.069 0.552 0.120 0.424 0.111 0.552 0.120

MEAN-I -0.254 0.341 -0.143 0.720 -0.197 0.459 -0.143 0.720

MEAN-UN 0.445 0.095 0.447 0.208 0.424 0.111 0.552 0.120

Leighton & Maples 2002 MODE 0.377 0.060 0.333 1.000 -0.669 0.001 -1.000 0.333

MEAN 0.150 0.455 -0.333 1.000 -0.669 0.001 -1.000 0.333

Jeffery 1998 MODE -0.573 0.021 -0.183 0.710 -0.505 0.042 -0.183 0.710

MEAN -0.630 0.011 -0.183 0.710 -0.505 0.042 -0.183 0.710

Jeffery & Emlet 2003 MODE 0.179 0.334 0.286 0.399 0.383 0.038 0.143 0.720

MEAN 0.151 0.413 0.214 0.548 0.347 0.061 0.071 0.905

Karasawa and kato 2003 MODE -0.184 0.339 0.333 0.750 -0.023 0.905 0.333 0.750

MEAN -0.161 0.403 0.333 0.750 0.000 1.000 0.333 0.750

Michaux 1989 MEAN 0.501 0.002 0.485 0.031 -0.095 0.556 -0.030 0.947

MODE 0.466 0.004 0.545 0.014 -0.152 0.350 -0.061 0.841

Monks 1999 MODE -0.507 0.000 -0.333 0.750 0.675 0.000 0.667 0.333

MEAN -0.591 0.000 0.000 1.000 0.676 0.000 0.667 0.333

Monks 2002 MODE-I -0.119 0.428 0.400 0.483 0.086 0.566 0.600 0.233

MODE-UN -0.039 0.794 0.400 0.483 0.201 0.179 0.800 0.083

MEAN-I -0.329 0.028 0.200 0.817 0.128 0.391 0.600 0.233

MEAN-UN -0.319 0.033 0.000 1.000 0.195 0.192 0.800 0.083

Monks and Owens 1999 MODE-I -0.404 0.001 -0.283 0.139 -0.467 0.000 -0.427 0.021

MODE-UN -0.423 0.022 0.333 0.381 -0.345 0.062 0.238 0.562

MEAN-I -0.559 0.003 -0.350 0.269 -0.394 0.033 -0.238 0.562

MEAN-UN -0.372 0.044 0.390 0.218 -0.345 0.062 0.238 0.562

Nutzel et al. 2000 Mode 0.327 0.161 0.214 0.548 -0.135 0.564 -0.071 0.905

Mean 0.428 0.067 0.286 0.399 -0.135 0.564 -0.071 0.905

O'Keefe 2004 MODE-I 0.404 0.068 0.238 0.562 0.417 0.059 0.333 0.381

MODE-UN 0.512 0.020 0.707 0.150 0.460 0.037 0.707 0.150

MEAN-I 0.384 0.082 0.238 0.562 0.417 0.059 0.333 0.381

MEAN-UN 0.460 0.037 0.707 0.150 0.415 0.061 0.913 0.063

� 236

Appendix O (con't)

GROUP


Popov et al. 1999 MODE -0.350 0.014 -0.254 0.341 -0.039 0.783 -0.056 0.919

MEAN -0.321 0.025 -0.261 0.327 -0.054 0.707 0.000 1.000

Roopnarine 2001-1 MODE -0.471 0.011 -0.333 0.750 -0.271 0.144 -0.333 0.750

MEAN -0.411 0.026 -0.667 0.333 -0.291 0.116 0.000 1.000

Roopnarine 2001-2 MODE 0.616 0.001 0.414 0.243 0.531 0.003 0.600 0.136

MEAN 0.559 0.002 0.333 0.469 0.549 0.002 0.600 0.136

Roopnarine 2001-3 MODE -0.048 0.821 0.000 1.000 0.523 0.013 0.333 0.750

MEAN 0.333 0.113 0.000 1.000 0.523 0.013 0.333 0.750

Roopnarine 2001-4 MODE 0.676 0.000 0.200 0.817 0.386 0.045 0.200 0.817

MEAN 0.656 0.001 0.200 0.817 0.340 0.077 -0.200 0.817

Schneider 1995 MODE 0.393 0.002 0.200 0.306 -0.429 0.001 -0.380 0.040

MEAN -0.415 0.001 -0.267 0.165 -0.489 0.000 -0.427 0.021

Smith 1988 MODE -0.221 0.093 -0.667 0.333 -0.226 0.085 -1.000 0.083

MEAN -0.149 0.256 -0.333 0.750 -0.243 0.064 -0.913 0.063

Smith and Arbizu 1987 MODE 0.267 0.204 0.286 0.399 0.480 0.022 0.500 0.109

MEAN 0.744 0.000 0.837 0.004 0.454 0.031 0.500 0.109

Smith et al. 1995 MODE-I -0.249 0.063 -0.260 0.240 -0.154 0.251 -0.197 0.372

MODE-UN -0.345 0.010 0.222 0.477 -0.402 0.003 0.141 0.597

MEAN-I -0.219 0.102 -0.263 0.234 -0.162 0.227 -0.230 0.297

MEAN-UN -0.467 0.000 0.061 0.819 -0.411 0.002 0.197 0.459

Smith and Wright 1993 MODE-I 0.621 0.002 0.214 0.548 0.361 0.072 0.000 1.000

MODE-UN 0.534 0.011 0.444 0.119 0.294 0.162 0.167 0.612

MEAN-I 0.504 0.012 0.143 0.720 0.384 0.056 0.000 1.000

MEAN-UN 0.240 0.253 0.111 0.761 0.374 0.075 0.333 0.260

Tinn & Meidla 2004 MODE 0.421 0.000 1.000 0.333 0.367 0.002 0.333 1.000

MEAN 0.344 0.004 1.000 0.333 0.366 0.002 0.333 1.000

Vermeij & Carlson 2000 MODE 0.359 0.002 0.429 0.179 -0.221 0.058 -0.143 0.720

MEAN 0.251 0.031 0.357 0.275 -0.257 0.028 -0.143 0.720

Wagner 1999 MODE -0.535 0.000 -0.501 0.000 0.146 0.051 0.154 0.153

MEAN -0.568 0.000 -0.546 0.000 0.128 0.089 0.126 0.242

Wagner 1997 MODE -0.431 0.000 0.000 1.000 -0.167 0.002 -0.333 0.750

MEAN -0.465 0.000 -0.183 0.710 -0.167 0.002 -0.333 0.750

Wagner Riberiidae MODE -0.656 0.000 -0.707 0.150 -0.354 0.010 0.333 0.750

MEAN -0.605 0.000 -0.548 0.264 -0.373 0.006 0.333 0.750

Wagner Technophoridae MODE -0.416 0.020 -0.913 0.063 -0.270 0.131 -0.667 0.333

MEAN -0.416 0.020 -0.913 0.063 -0.270 0.131 -0.667 0.333

Wagner Bransoniidae MODE -0.067 0.664 -0.333 1.000 -0.328 0.033 -0.333 1.000

MEAN -0.233 0.129 -0.333 1.000 -0.346 0.024 -0.333 1.000

Wagner Hippocardiidae MODE -0.271 0.015 -0.333 1.000 -0.111 0.320 -0.333 1.000

MEAN -0.274 0.014 -0.333 1.000 -0.113 0.313 -0.333 1.000

Wagner Coiled All MODE -0.160 0.000 -0.568 0.000 0.268 0.000 -0.644 0.000

MEAN -0.003 0.935 0.201 0.049 0.262 0.000 -0.652 0.000

MODE-s -0.062 0.041 -0.571 0.061 0.402 0.000 -0.500 0.109

MEAN-s 0.012 0.696 0.403 0.163 0.410 0.000 -0.500 0.109

Wagner Euomphaloids MODE -0.531 0.000 0.229 0.158 -0.610 0.000 -0.076 0.654

MEAN -0.508 0.000 0.276 0.085 -0.608 0.000 -0.062 0.694

MODE-S -0.468 0.000 0.000 1.000 -0.491 0.000 -0.600 0.136

MEAN-S -0.512 0.000 -0.200 0.719 -0.497 0.000 -0.600 0.136

� 237

Appendix O (con't)

GROUP


Wagner Pleurotomarids MODE -0.455 0.000 -0.089 0.440 0.371 0.000 -0.258 0.025

MEAN -0.457 0.000 -0.053 0.647 0.372 0.000 -0.287 0.012

MODE-S -0.544 0.000 -0.500 0.109 0.305 0.000 0.143 0.720

MEAN-S -0.533 0.000 0.000 1.000 0.308 0.000 0.143 0.720

Wagner Trochoids MODE -0.604 0.004 -0.600 0.233 0.369 0.079 0.400 0.483

MEAN -0.614 0.003 -0.600 0.233 0.369 0.079 0.400 0.483

Wagner Murchisonoids MODE -0.461 0.000 -0.309 0.091 -0.552 0.000 -0.515 0.003

MEAN -0.608 0.000 -0.162 0.393 -0.561 0.000 -0.529 0.002

Wagner Microdomatoid MODE -0.623 0.008 -0.707 0.150 -0.446 0.056 0.000 1.000

MEAN -0.623 0.008 -0.707 0.150 -0.446 0.056 0.000 1.000

Wagner Trochonematoid MODE -0.481 0.012 0.000 1.000 -0.605 0.002 -0.333 0.750

MEAN -0.531 0.006 -0.183 0.710 -0.655 0.001 -0.333 0.750

Wagner Macluritoids MODE -0.564 0.003 -0.359 0.380 -0.275 0.153 0.400 0.483

MEAN -0.728 0.000 -0.632 0.121 -0.307 0.110 0.400 0.483

Yates & Warrens 2002 MODE-I 0.176 0.125 0.052 0.772 0.354 0.002 0.118 0.542

MODE-UN 0.219 0.056 -0.503 0.017 0.448 0.000 -0.297 0.158

MEAN-I 0.317 0.006 0.060 0.738 0.343 0.003 0.111 0.535

MEAN-UN 0.400 0.000 -0.400 0.057 0.462 0.000 -0.219 0.297

Curfsina (this study) MODE -0.393 0.003 0.038 0.842 0.184 0.160 -0.219 0.282

MEAN -0.583 0.000 -0.498 0.010 0.201 0.126 -0.238 0.239

Opimocythere (this study) MODE -0.460 0.010 -0.056 0.919 -0.175 0.328 0.056 0.919

MEAN -0.458 0.010 -0.056 0.919 -0.133 0.456 0.056 0.919

Schizoptocythere (this study) MODE -0.075 0.687 0.000 1.000 -0.248 0.179 -0.400 0.483

MEAN -0.467 0.012 -0.120 0.770 -0.224 0.227 -0.400 0.483

Phalcocythere (this study) MODE -0.470 0.011 -0.120 0.770 -0.224 0.227 -0.400 0.483

MEAN -0.569 0.000 -0.128 0.590 0.011 0.934 0.282 0.204

Data from Liow 2004, Table 1

Roveacrinida -0.352 0.094 -0.500 0.075 NA NA NA NA

Cyrtocrinida -0.450 0.001 -0.437 0.006 NA NA NA NA

Comatulida -0.506 0.000 -0.449 0.000 NA NA NA NA

Millericrinida -0.630 0.029 -0.582 0.066 NA NA NA NA

Isocrinida -0.487 0.006 -0.487 0.006 NA NA NA NA

Cladida (All) 0.230 0.000 0.411 0.000 NA NA NA NA

Cladida (Ordovician-Devonian) -0.430 0.000 -0.129 0.357 NA NA NA NA

Cladida (Lower Carboniferous) -0.343 0.000 0.364 0.059 NA NA NA NA

Cladida (Upper Carboniferous- Permian) -0.018 0.775 0.337 0.024 NA NA NA NA

Cladida (Cyathocrinina) 0.409 0.000 0.383 0.026 NA NA NA NA

Cladida (Dendorcrinina) -0.249 0.038 -0.281 0.083 NA NA NA NA

Cladida (Poteriocrinina) -0.619 0.000 -0.621 0.000 NA NA NA NA

Disparida -0.050 0.510 -0.175 0.232 NA NA NA NA

Diplobathrida -0.650 0.000 -0.618 0.002 NA NA NA NA

� 238

Appendix O (con't)

GROUP


Monobathrida -0.210 0.006 -0.130 0.341 NA NA NA NA

Taxocrinida -0.445 0.123 -0.333 0.469 NA NA NA NA

Sagenocrinida 0.216 0.044 0.152 0.362 NA NA NA NA

Data from Liow 2006 (see Tables 2-7)

ALL -0.339 0.000 -0.379 0.000 0.236 0.000 0.187 0.000

NoR -0.239 0.000 -0.237 0.000 -0.131 0.013 -0.186 0.003

NoS -0.365 0.000 -0.363 0.000 0.372 0.000 0.326 0.000

NoRNoS -0.239 0.000 -0.250 0.000 -0.222 0.000 -0.256 0.000

OversplitR -0.330 0.000 -0.350 0.000 0.258 0.000 0.296 0.000

4 PCs EFA 0.320 0.000 -0.001 0.986 NA NA NA NA

Eigenshape 0.575 0.000 0.518 0.000 NA NA NA NA

26NA -0.265 0.000 -0.186 0.004 0.583 0.000 0.441 0.000

17NA -0.413 0.000 -0.416 0.000 0.559 0.000 0.423 0.000

FA166 0.308 0.164 0.404 0.084 0.382 0.084 0.426 0.068

FA116 0.199 0.322 0.256 0.252 0.189 0.347 0.194 0.357

FA105 -0.517 0.000 -0.549 0.001 -0.527 0.000 -0.513 0.002

FA95 -0.203 0.086 -0.207 0.096 -0.433 0.000 -0.451 0.000

FA77 -0.299 0.007 -0.261 0.047 -0.368 0.001 -0.341 0.009

FA65 -0.225 0.066 -0.025 0.892 0.513 0.000 0.611 0.001

FA54 0.153 0.243 0.100 0.626 0.525 0.000 0.450 0.015

FA42 -0.497 0.013 -0.617 0.052 -0.398 0.047 -0.617 0.052

FA25 -0.320 0.015 -0.018 1.000 -0.020 0.879 0.236 0.359

FA15 -0.377 0.003 0.147 0.530 -0.241 0.057 -0.345 0.165

FA5 -0.322 0.000 0.067 0.862 -0.043 0.613 -0.289 0.291

Cretaceous -0.269 0.000 -0.290 0.000 -0.272 0.000 -0.298 0.000

Paleocene -0.245 0.000 -0.268 0.001 -0.306 0.000 -0.417 0.000

Eocene -0.271 0.000 -0.279 0.000 -0.298 0.000 -0.391 0.000

Oligocene -0.189 0.014 -0.153 0.110 -0.484 0.000 -0.465 0.000

Miocene -0.205 0.001 -0.132 0.129 0.254 0.000 0.214 0.014

Post-Miocene -0.253 0.000 -0.090 0.326 0.262 0.000 0.072 0.433

� 239

Appendix O (con't)

SINGLE

Reference TYPE TAU p(s) TAU p(s, m) TAU p(s,pco) TAU p(s,m,pco)Adnet and Capetta 2001 MODE-I 0.292 0.051 0.302 0.080 0.160 0.284 0.242 0.175

MODE-UN 0.235 0.116 0.275 0.193 0.089 0.551 0.077 0.747

MEAN-I 0.331 0.027 0.304 0.078 0.148 0.323 0.210 0.224

MEAN-UN 0.247 0.099 0.319 0.127 0.094 0.532 0.033 0.914

Adrian and Westrop 2001 MODE-I 0.142 0.521 0.333 0.750 -0.301 0.173 -0.667 0.333

MODE-UN 0.021 0.925 0.333 1.000 -0.229 0.300 0.333 1.000

MEAN-I -0.142 0.521 -0.333 0.750 -0.266 0.229 -0.667 0.333

MEAN-UN -0.229 0.300 -0.333 1.000 -0.270 0.221 0.333 1.000

Adrain and Edgecomb 1997 MODE-I-M -0.043 0.736 -0.029 0.923 0.040 0.749 -0.067 0.770

MODE-UN-M -0.129 0.308 -0.308 0.163 -0.109 0.390 -0.400 0.057

MEAN-I-M -0.094 0.456 -0.086 0.697 0.040 0.750 -0.105 0.626

MEAN-UN-M -0.174 0.168 -0.323 0.125 -0.113 0.370 -0.400 0.057

MODE-I-S -0.184 0.146 -0.333 0.750 -0.140 0.269 -0.333 0.750

MODE-UN-S -0.353 0.005 -1.000 0.333 -0.241 0.057 -0.333 1.000

MEAN-I-S -0.311 0.014 -0.667 0.333 -0.145 0.250 -0.333 0.750

MEAN-UN-S -0.422 0.001 -1.000 0.333 -0.241 0.057 -0.333 1.000

Allmon 1996 (Table 1) MODE 0.414 0.000 0.140 0.528 0.008 0.930 0.242 0.311

MEAN -0.254 0.008 0.152 0.545 0.002 0.983 0.273 0.250

Allmon 1996 (Table 9) MODE -0.172 0.139 -0.283 0.139 -0.013 0.910 -0.050 0.825

MEAN -0.197 0.091 -0.367 0.052 -0.021 0.855 -0.067 0.757

Alroy 1995 MODE -0.399 0.025 0.333 1.000 0.236 0.186 1.000 0.333

MEAN 0.012 0.920 0.333 1.000 0.199 0.098 1.000 0.333

Alvarez et al. 1998 MODE-I 0.354 0.002 0.334 0.026 -0.074 0.525 -0.028 0.853

MODE-UN -0.008 0.944 -0.057 0.691 -0.121 0.298 -0.100 0.502

MEAN-I -0.003 0.977 0.000 1.000 -0.084 0.472 -0.071 0.634

MEAN-UN 0.276 0.018 0.238 0.096 -0.109 0.352 -0.107 0.453

Amati and Westrop 2004 MODE-M -0.084 0.617 0.033 0.914 0.168 0.314 0.177 0.378

MEAN-M -0.198 0.237 -0.088 0.660 0.187 0.264 0.177 0.378

MODE-S -0.234 0.175 -0.143 0.773 0.118 0.495 0.048 1.000

MEAN-S -0.267 0.122 -0.048 1.000 0.133 0.442 0.143 0.773

Anderson and Roopnarine 2003 MODE -0.583 0.008 -0.467 0.272 -0.302 0.172 -0.733 0.056

MEAN -0.446 0.044 -0.690 0.052 -0.268 0.225 -0.733 0.056

Angielczky & Kurkin 2003 MODE-I -0.055 0.736 0.200 0.817 -0.042 0.793 0.000 1.000

MODE-UN -0.049 0.762 -0.333 0.750 0.035 0.828 0.667 0.333

MEAN-I -0.036 0.822 0.200 0.817 -0.055 0.736 0.000 1.000

MEAN-UN -0.063 0.697 0.000 1.000 0.035 0.828 0.667 0.333

Bloch et al. 2001 MODE-I -0.102 0.612 -0.333 0.750 -0.483 0.016 -1.000 0.083

MODE-UN 0.071 0.723 0.333 1.000 -0.298 0.137 -1.000 0.333

MEAN-I -0.280 0.164 -0.667 0.333 -0.432 0.031 -1.000 0.083

MEAN-UN -0.213 0.288 -1.000 0.333 -0.240 0.231 -1.000 0.333

Bodenbender and Fischer 2001 MODE-I 0.219 0.008 0.333 0.469 0.202 0.015 0.467 0.272

MODE-UN 0.234 0.005 0.333 0.469 0.197 0.018 0.467 0.272

MEAN-I 0.198 0.017 0.600 0.136 0.215 0.010 0.600 0.136

MEAN-UN 0.208 0.012 0.600 0.136 0.201 0.015 0.467 0.272

Brochu1997 MODE -0.306 0.001 -0.172 0.197 -0.244 0.006 -0.084 0.539

MEAN -0.335 0.000 -0.195 0.138 -0.234 0.008 -0.084 0.539

Brunet-Lecomte&Chaline 1990 MEAN -0.393 0.034 -0.455 0.045 -0.332 0.073 -0.455 0.045

Cairns 2001 MODE -0.220 0.088 -0.279 0.129 -0.040 0.759 -0.185 0.301

MEAN -0.446 0.001 -0.568 0.001 -0.040 0.759 -0.170 0.342

� 240

Appendix O (con't)

SINGLE

Reference TYPE TAU p(s) TAU p(s, m) TAU p(s,pco) TAU p(s,m,pco)Popov et al. 1999 MODE -0.254 0.075 -0.141 0.597 -0.189 0.184 -0.222 0.477

MEAN -0.354 0.013 -0.278 0.358 -0.197 0.168 -0.167 0.612

Roopnarine 2001-1 MODE -0.222 0.230 -0.333 0.750 -0.110 0.552 -0.333 0.750

MEAN -0.101 0.587 -0.667 0.333 -0.101 0.587 -0.333 0.750

Roopnarine 2001-2 MODE 0.350 0.050 0.733 0.056 0.061 0.733 0.467 0.272

MEAN 0.298 0.095 0.733 0.056 0.044 0.807 0.467 0.272

Roopnarine 2001-3 MODE -0.176 0.404 -0.333 0.750 -0.174 0.407 0.000 1.000

MEAN -0.333 0.113 -0.667 0.333 -0.174 0.407 0.000 1.000

Roopnarine 2001-4 MODE 0.366 0.057 0.800 0.083 0.159 0.409 0.400 0.483

MEAN 0.458 0.017 0.400 0.483 0.159 0.409 0.400 0.483

Schneider 1995 MODE 0.114 0.358 0.050 0.825 -0.079 0.525 -0.117 0.564

MEAN -0.189 0.128 0.000 1.000 -0.094 0.452 -0.117 0.564

Smith 1988 MODE -0.138 0.294 -0.333 0.750 -0.226 0.085 -0.667 0.333

MEAN -0.083 0.529 -0.333 0.750 -0.248 0.059 -0.667 0.333

Smith and Arbizu 1987 MODE 0.107 0.611 0.071 0.905 0.080 0.703 0.214 0.548

MEAN 0.121 0.565 0.255 0.378 0.080 0.703 0.071 0.905

Smith et al. 1995 MODE-I -0.069 0.605 -0.091 0.737 -0.135 0.312 0.000 1.000

MODE-UN -0.293 0.029 -0.167 0.612 -0.376 0.005 -0.389 0.180

MEAN-I -0.108 0.420 -0.091 0.737 -0.163 0.224 -0.030 0.947

MEAN-UN -0.545 0.000 -0.500 0.075 -0.447 0.001 -0.500 0.075

Smith and Wright 1993 MODE-I 0.314 0.117 0.143 0.720 0.151 0.451 0.000 1.000

MODE-UN 0.201 0.338 0.111 0.761 -0.027 0.899 -0.111 0.761

MEAN-I 0.234 0.243 0.036 0.900 0.151 0.451 0.000 1.000

MEAN-UN 0.148 0.482 0.028 0.916 -0.027 0.899 -0.111 0.761

Tinn & Meidla 2004 MODE 0.029 0.808 0.333 1.000 0.110 0.359 1.000 0.333

MEAN 0.054 0.653 1.000 0.333 0.106 0.380 1.000 0.333

Vermeij & Carlson 2000 MODE 0.190 0.102 0.214 0.548 -0.057 0.624 -0.429 0.179

MEAN -0.009 0.941 -0.357 0.275 -0.078 0.504 -0.429 0.179

Wagner 1999 MODE -0.492 0.000 -0.481 0.000 0.021 0.779 0.085 0.438

MEAN -0.613 0.000 -0.685 0.000 0.038 0.617 0.101 0.345

Wagner 1997 MODE -0.291 0.000 -1.000 0.083 0.049 0.368 0.333 0.750

MEAN -0.394 0.000 -1.000 0.083 0.040 0.460 0.333 0.750

Wagner Riberiidae MODE -0.481 0.000 -1.000 0.083 -0.181 0.187 0.333 0.750

MEAN -0.527 0.000 -1.000 0.083 -0.181 0.187 0.333 0.750

Wagner Technophoridae MODE -0.477 0.008 -1.000 0.083 -0.218 0.223 -0.667 0.333

MEAN -0.490 0.006 -1.000 0.083 -0.200 0.262 -0.667 0.333

Wagner Bransoniidae MODE -0.286 0.062 -0.333 1.000 -0.198 0.197 -0.333 1.000

MEAN -0.313 0.041 -0.816 0.201 -0.207 0.177 -0.333 1.000

Wagner Hippocardiidae MODE -0.359 0.001 -0.333 1.000 -0.215 0.054 -1.000 0.333

MEAN -0.333 0.003 -0.333 1.000 -0.220 0.049 -1.000 0.333

Wagner Coiled All MODE 0.056 0.064 -0.199 0.051 0.151 0.000 0.010 0.925

MEAN -0.413 0.000 -0.049 0.629 0.150 0.000 0.004 0.970

MODE-s 0.064 0.035 -0.500 0.109 0.182 0.000 0.357 0.275

MEAN-s -0.442 0.000 -0.571 0.061 0.177 0.000 0.286 0.399

Wagner Euomphaloids MODE -0.096 0.253 0.076 0.654 0.047 0.575 -0.067 0.698

MEAN -0.077 0.357 0.124 0.455 0.043 0.605 -0.081 0.607

MODE-S -0.017 0.836 0.467 0.272 0.135 0.106 0.333 0.469

MEAN-S -0.008 0.920 0.467 0.272 0.138 0.099 0.333 0.469

� 241

Appendix O (con't)

SINGLE

Reference TYPE TAU p(s) TAU p(s, m) TAU p(s,pco) TAU p(s,m,pco)Wagner Pleurotomarids MODE -0.114 0.016 -0.069 0.559 0.017 0.724 -0.167 0.146

MEAN -0.103 0.030 -0.051 0.668 0.020 0.670 -0.179 0.119

MODE-S -0.126 0.008 -0.286 0.399 0.014 0.761 -0.214 0.548

MEAN-S -0.121 0.011 -0.357 0.275 0.011 0.820 -0.214 0.548

Wagner Trochoids MODE -0.339 0.107 -0.800 0.083 0.192 0.362 -0.200 0.817

MEAN -0.601 0.004 -0.800 0.083 0.178 0.397 -0.200 0.817

Wagner Murchisonoids MODE -0.078 0.357 -0.059 0.776 -0.054 0.519 -0.265 0.151

MEAN -0.133 0.114 -0.221 0.236 -0.056 0.506 -0.265 0.151

Wagner Microdomatoid MODE -0.070 0.763 0.000 1.000 -0.507 0.030 -0.667 0.333

MEAN 0.047 0.839 0.000 1.000 -0.507 0.030 -0.667 0.333

Wagner Trochonematoid MODE -0.532 0.006 -0.333 0.750 -0.334 0.083 -0.667 0.333

MEAN -0.557 0.004 -0.333 0.750 -0.334 0.083 -0.667 0.333

Wagner Macluritoids MODE -0.715 0.000 -0.738 0.071 0.267 0.166 0.800 0.083

MEAN -0.664 0.001 -0.738 0.071 0.191 0.322 0.600 0.233

Yates & Warrens 2002 MODE-I 0.188 0.102 0.059 0.776 0.230 0.045 0.221 0.236

MODE-UN 0.143 0.212 0.000 1.000 0.260 0.024 0.256 0.252

MEAN-I 0.265 0.021 0.176 0.349 0.247 0.031 0.199 0.264

MEAN-UN 0.246 0.032 0.000 1.000 0.285 0.013 0.256 0.252Curfsina (this study) MODE -0.144 0.274 -0.143 0.495 0.111 0.397 -0.115 0.551

MEAN -0.212 0.107 -0.352 0.074 0.125 0.342 -0.105 0.626Opimocythere (this study) MODE -0.349 0.050 -0.222 0.477 0.108 0.546 0.333 0.260

MEAN -0.349 0.050 -0.222 0.477 0.124 0.486 0.389 0.180Schizoptocythere (this study) MODE 0.199 0.283 -0.200 0.817 0.099 0.591 -0.200 0.817

MEAN -0.373 0.044 -0.600 0.233 0.099 0.591 -0.200 0.817Phalcocythere (this study) MODE -0.373 0.044 -0.600 0.233 0.099 0.591 -0.200 0.817

MEAN -0.231 0.072 -0.179 0.435 -0.019 0.885 0.179 0.435

Data from Liow 2004, Table 1

Roveacrinida -0.149 0.479 -0.278 0.358 NA NA NA NA

Cyrtocrinida -0.259 0.053 -0.364 0.021 NA NA NA NA

Comatulida -0.471 0.000 -0.449 0.000 NA NA NA NA

Millericrinida -0.491 0.089 -0.586 0.065 NA NA NA NA

Isocrinida -0.126 0.479 -0.126 0.479 NA NA NA NA

Cladida (All) -0.047 0.255 -0.133 0.199 NA NA NA NA

Cladida (Ordovician-Devonian) -0.247 0.002 -0.210 0.133 NA NA NA NA

Cladida (Lower Carboniferous) -0.103 0.169 -0.067 0.770 NA NA NA NA

Cladida (Upper Carboniferous-

Permian) -0.132 0.039 -0.043 0.794 NA NA NA NA

Cladida (Cyathocrinina) 0.258 0.025 0.262 0.128 NA NA NA NACladida (Dendorcrinina) -0.081 0.500 -0.048 0.769 NA NA NA NACladida (Poteriocrinina) -0.090 0.073 -0.293 0.015 NA NA NA NA

Disparida 0.028 0.711 -0.040 0.785 NA NA NA NA

Diplobathrida -0.205 0.066 -0.331 0.099 NA NA NA NA

� 242

Appendix O (con't)

SINGLE

Reference TYPE TAU p(s) TAU p(s, m) TAU p(s,pco) TAU p(s,m,pco)

Monobathrida -0.181 0.017 -0.262 0.055 NA NA NA NA

Taxocrinida -0.148 0.608 -0.200 0.719 NA NA NA NA

Sagenocrinida 0.063 0.556 0.152 0.362 NA NA NA NA

Data from Liow 2006 (see Tables 2-7)

ALL -0.167 0.000 -0.236 0.000 0.104 0.005 -0.015 0.780

NoR -0.252 0.000 -0.256 0.000 -0.007 0.895 -0.072 0.253

NoS -0.112 0.006 -0.087 0.101 0.116 0.004 0.067 0.207

NoRNoS -0.254 0.000 -0.255 0.000 -0.010 0.862 -0.045 0.474

OversplitR -0.176 0.000 -0.239 0.000 0.030 0.454 -0.078 0.178

4 PCs EFA 0.060 0.133 0.001 0.990 NA NA NA NA

Eigenshape 0.039 0.329 -0.029 0.605 NA NA NA NA

26NA -0.160 0.000 -0.180 0.005 0.063 0.168 0.019 0.764

17NA -0.184 0.005 -0.198 0.023 0.076 0.244 0.023 0.789

FA166 -0.215 0.330 -0.110 0.637 0.123 0.578 0.257 0.271

FA116 -0.066 0.741 -0.051 0.858 0.155 0.441 0.154 0.510

FA105 -0.099 0.487 -0.203 0.210 -0.085 0.553 -0.185 0.255

FA95 -0.169 0.152 -0.178 0.152 -0.198 0.094 -0.204 0.100

FA77 -0.271 0.015 -0.267 0.042 -0.165 0.139 -0.180 0.170

FA65 -0.155 0.205 -0.150 0.450 0.145 0.235 0.226 0.222

FA54 0.051 0.697 0.017 0.965 0.148 0.260 0.233 0.228

FA42 0.012 0.953 0.143 0.773 -0.105 0.599 -0.143 0.773

FA25 -0.157 0.232 -0.183 0.432 0.000 1.000 0.236 0.359

FA15 -0.366 0.004 -0.018 1.000 -0.066 0.603 0.164 0.542

FA5 0.051 0.549 0.067 0.862 0.113 0.184 -0.244 0.381

Cretaceous -0.264 0.000 -0.264 0.000 -0.068 0.258 -0.109 0.109

Paleocene -0.178 0.006 -0.223 0.004 0.002 0.970 -0.119 0.125

Eocene -0.173 0.003 -0.171 0.017 -0.019 0.739 -0.120 0.093

Oligocene -0.162 0.035 -0.183 0.056 0.021 0.788 -0.054 0.570

Miocene -0.171 0.005 -0.211 0.016 0.048 0.430 -0.065 0.455

Post-Miocene -0.201 0.000 -0.267 0.004 0.161 0.001 0.101 0.270

� 243

THE UNIVERSITY OF CHICAGO ODDITIES, WONDERS, AND OTHER ... · the university of chicago oddities,...

Documents

Transcript of THE UNIVERSITY OF CHICAGO ODDITIES, WONDERS, AND OTHER ... · the university of chicago oddities,...