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
Copyright © by Lee Hsiang Liow
All rights reserved
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
GROUPSINGLEmean-durmed-dursddurbeta
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
GROUPSINGLEmean-durmed-dursddurbeta
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
Paleontology 51:414-415.
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,
Palaeoecology 224:232-269.
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,
Palaeoecology 115:249-264.
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
Paleontology 364:5-29.
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.
Paleobiology 6:397-407.
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.
Paleobiology 11:75-90.
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.
Paleobiology 14:126-138.
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.
Paleobiology 14:109-125.
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.
Paleobiology 1:63-70.
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.
Paleobiology 1:343-355.
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.
Paleobiology 19:43-51.
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
Reference TYPE TAU p(g) TAU p(g, m) TAU p(g,pco) TAU p(g,m,pco)
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
Reference TYPE TAU p(g) TAU p(g, m) TAU p(g,pco) TAU p(g,m,pco)
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
Reference TYPE TAU p(g) TAU p(g, m) TAU p(g,pco) TAU p(g,m,pco)
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
Reference TYPE TAU p(g) TAU p(g, m) TAU p(g,pco) TAU p(g,m,pco)
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
Top Related