The spatial scale of speciation and patterns of diversity
Yael Kisel
A thesis submitted for the degree of Doctor of Philosophy
of the University of London
Division of Biology
Imperial College London
Silwood Park Campus
December 2010
Abstract
Many environmental factors and taxon traits have been studied as potential
controllers of diversification, but there is no consensus as to which are most important or
how to link them into a general theory of diversification. I hypothesise that diversification
is strongly controlled by the interaction between area and clades’ spatial scales of
speciation, or the minimum amount of area they require for speciation to occur.
Furthermore, I hypothesise that the spatial scale of speciation is controlled by population
genetic characteristics of clades, as speciation is ultimately a process of population
divergence. In this thesis, I quantify taxonomic variation in the spatial scale of speciation,
test whether it can be explained by variation in population genetics and evaluate whether
it can explain taxonomic patterns of diversity. Using a survey of speciation events on
isolated oceanic islands, I show that the spatial scale of speciation varies greatly between
birds, lizards, snails, bats, carnivorous mammals, lepidoptera, angiosperms and ferns. I
also use a meta-analysis of population genetic data collected from the literature to show
that the minimum area for speciation of these groups correlates strongly with their
average level of gene flow. I then test the link between population genetics and
diversification by comparing population genetic characteristics of sister clades of tropical
orchids that differ greatly in species richness. Contrary to expectation, levels of gene
flow, genetic drift and local adaptation do not correlate directly with rates of
diversification. However, there is some evidence for an interaction between species range
size and gene flow in controlling diversification. This thesis supports a framework based
on the interaction between area and the spatial scale of speciation as a useful foundation
for general theories of diversification. It also highlights the potential for using a
comparative population genetic approach in macroevolutionary studies.
2
Acknowledgements
First, I would like to thank Tim Barraclough and Mark Chase for all their support and patience. I am
also grateful to NSF, the Imperial College Deputy Rector's Scholarship, the University of London Central
Research Fund, the Kew Bentham-Moxon Trust, and Sigma Xi for funding my work.
In Costa Rica, I especially thank Diego Bogarín, who helped me with everything. I am very grateful to
Jorge Warner for making me welcome at the Jardín Botánico Lankester and helping me with permits.
Thanks also to many others at JBL: Franco Pupulin for advice; Rei and Rafa and Adam Karremans for
coming on field trips; Allan and Enzo Salas for helping me settle in; and Socorro for dealing with
paperwork. Thanks to the helpful staff of SINAC, especially Javier Guevara, Roger Blanco and Oscar
Masis, and to UCR reserve director Ronald Sanchez. Thanks also to the owners of Bosque de Paz Reserve,
and Melania Muñoz for organizing my visit there; to the owners of the Rara Avis Reserve; and to Freddy
and Katia who allowed me to collect on their land. Great thanks to my fieldwork volunteers, Ryan Phillips,
Julia Hu, Paul Renshaw and Kath Castillo. Finally, thanks to Fanny Bonilla and Carlos Piedra for hosting us
so warmly.
In the lab, my thanks first to Martyn Powell, who took on the task of teaching me AFLPs and
answered my never-ending questions thereafter. I am also very grateful to Vincent Savolainen for allowing
me to use his group’s lab facilities. Thanks to Robyn Cowan and Ovidiu Paun for in-depth talks about
AFLP methodology and troubleshooting. Special thanks to Helen Hipperson for troubleshooting help,
support and empathy. Thanks to Thomas Hahn for trying hard to figure out my AFLP troubles over email.
Finally, thanks to everyone in the Savolainen lab for helping me find my feet with labwork, in particular
Haris Saslis-Lagoudakis, Hanno Schaefer, Guillaume Besnard, Silvana del Vecchio, Paul Rymer and Alex
Papadopulos.
Thanks to Christian Lexer, Ally Phillimore and Alex Pigot for many illuminating discussions, and to
all my friendly labmates, especially Diego Fontaneto. Thanks to Diana Anderson and Christine Short for
cheerful help with everything administrative, and John Williams and the security crew for opening doors
and solving problems. Thanks to the whole friendly Silwood community for making my PhD years so
enjoyable, with special mention to Kat, Martina, Irka, Geraldine, Sally, Susanne, and Alice.
Immense thanks to the house: Lynsey, Ellie and Lena for everything, but especially keeping me sane.
Thanks to Bruce Tiffney, for inspiring me to be a passionate scientist and giving me the idea to go to
the UK.
Thanks to my parents for their unwavering support, even after I decided to move to a country 8 time
zones away.
And finally, thanks to Martin, for field help, lab help, R help, formatting help, numerous bloody
marys and much more.
3
Declaration of originality
I declare that all the work presented in this thesis is my own original research, with the
following acknowledgements for each chapter:
Chapter 2 has been published in slightly modified form in American Naturalist. It was
written in collaboration with Tim Barraclough and made use of unpublished checklist
data generously provided by Shai Meiri, Ana M. C. Santos, Roberto S. Gómez, Tod F.
Stuessy and Christophe Thebaud. It was also greatly improved by suggestions from
Jonathan Davies, Joaquin Hortal, Christian Lexer, Shai Meiri, Lynsey McInnes, Ally
Phillimore, Andy Purvis, Vincent Savolainen and two anonymous reviewers.
Chapters 3 and 4 made use of data-formatting and analysis scripts for R written by Martin
Turjak.
4
Table of Contents
Table of Contents
Abstract ............................................................................................................................... 2
Acknowledgements ............................................................................................................ 3
Declaration of originality .................................................................................................. 4
Table of Contents ............................................................................................................... 5
List of Figures ..................................................................................................................... 8
List of Tables ...................................................................................................................... 9
List of Equations .............................................................................................................. 10
Chapter 1. Introduction.................................................................................................. 11
Comparative methods for studying variation in diversity .............................................. 11
What controls variation in diversity? Organism traits versus environmental variables . 13
A proposed framework for understanding variation in diversification .......................... 15
Approach and aims ......................................................................................................... 21
Summary of aims ........................................................................................................... 23
Chapter 2. Using oceanic islands to measure the spatial scale of speciation and its
association with gene flow ............................................................................................... 24
Introduction .................................................................................................................... 24
Materials and Methods ................................................................................................... 30
Island selection and data collection ........................................................................... 30
Island species data collection ..................................................................................... 31
Identification of speciation events .............................................................................. 33
Adding phylogenetic information ............................................................................... 35
Statistical analysis of the speciation-area relationship .............................................. 36
Gene flow data ............................................................................................................ 37
Gene flow analyses ..................................................................................................... 40
Results ............................................................................................................................ 43
5
Table of Contents
Data availability and quality ...................................................................................... 43
Quantifying the speciation-area relationship ............................................................. 43
Measuring minimum areas for speciation .................................................................. 47
Testing the importance of area when other environmental variables are included ... 48
The effect of gene flow ................................................................................................ 52
Discussion ...................................................................................................................... 53
Main findings .............................................................................................................. 53
The effects of other island characteristics on speciation probability ......................... 54
The spatial scale of speciation and gene flow ............................................................ 56
Evolutionary explanations for the observed patterns ................................................. 59
The effects of taxonomic practice and surveying effort .............................................. 62
Implications for evolutionary studies of diversity patterns ........................................ 63
Chapter 3. The relationship between gene flow and clade diversification rates in
Costa Rican orchids ......................................................................................................... 65
Introduction .................................................................................................................... 65
Methods .......................................................................................................................... 68
Study group selection .................................................................................................. 68
Study species phylogeny reconstruction ..................................................................... 78
Sample collection ........................................................................................................ 79
AFLP genotyping ........................................................................................................ 84
AFLP scoring .............................................................................................................. 89
Finalising AFLP datasets ........................................................................................... 90
Analysing Fst patterns ................................................................................................. 91
Testing the influences of species range size and ecology ........................................... 93
Results ............................................................................................................................ 95
Discussion .................................................................................................................... 103
Findings .................................................................................................................... 103
Study limitations ....................................................................................................... 106
Ideas for future work ................................................................................................ 108
Conclusion .................................................................................................................... 109
6
Table of Contents
Chapter 4. The effects of genetic drift and local adaptation on clade diversification
rates in Costa Rican orchids ......................................................................................... 110
Introduction .................................................................................................................. 110
Methods ........................................................................................................................ 113
Results .......................................................................................................................... 116
Discussion .................................................................................................................... 121
Main findings ............................................................................................................ 121
Study limitations ....................................................................................................... 123
Conclusion .................................................................................................................... 125
Chapter 5. Conclusion .................................................................................................. 126
Summary of results....................................................................................................... 126
Directions for future work ............................................................................................ 128
General conclusions ..................................................................................................... 129
Bibliography ................................................................................................................... 130
Appendix I. Kisel et al. (in review). How diversification rates and diversity limits
combine to create large-scale species-area relationships............................................152
Appendix II. Supplementary tables and figures for chapter three............................200
7
List of Figures
List of Figures
Figure 1.1. Hypothesised framework linking geographic area and the spatial scale of
speciation as controllers of speciation and diversification. .............................................. 17
Figure 2.1. Patterns of diversification on islands. ............................................................. 29
Figure 2.2. Relationship between Fst and the geographic extent of study. ....................... 42
Figure 2.3. Relationship between the probability of speciation and area. ........................ 45
Figure 2.4. Minimum island size for speciation versus the average level of gene flow
when measured over geographic ranges of 10-100 km. ................................................... 53
Figure 2.5. Results of an alternative gene flow analysis - the relationship between the
minimum area for speciation and the spatial scale of neutral population differentiation. 59
Figure 3.1. Photos of study species. .............................................................................76-77
Figure 3.2. Sampling locations for study species from (a) the Masdevallia-Trisetella and
(b) the Lepanthes-Lepanthopsis clade pairs. ..................................................................... 81
Figure 3.3. Sampling locations for study species from (a) the Platystele-Dryadella and (b)
the Epidendrum-Brassavola clade pairs. .......................................................................... 82
Figure 3.4. Sampling locations for study species from the Scaphyglottis-Jacquiniella
clade pair. . ......................................................................................................................... 83
Figure 3.5. Relationship between Fst and distance for all study species for the full dataset.
............................................................................................................................................ 97
Figure 3.6. Relationship between Fst and distance for all study species for the reduced
dataset. .............................................................................................................................. 98
Figure 3.7. Associations of Fst and species range size with species phylogeny (data from
the full dataset). ............................................................................................................... 100
Figure 3.8. Relationships between species elevation range and number of habitats and
whether any pair-wise Fst value is over 0.2, for the full dataset. .................................... 101
Figure 3.9. Associations between species range size and elevation range and number of
habitats, over all species native to Costa Rica from the study clades. ............................ 102
Figure 4.1. Examples of digitised leaf outlines. .............................................................. 115
Figure 4.2. Associations of (a) gene diversity, (b) overall Pst and (c) difference between
Pst and Fst with species phylogeny for the full dataset. .................................................. 119
Figure 4.3. Associations of (a) gene diversity, (b) overall Pst and (c) difference between
Pst and Fst with species phylogeny for the reduced dataset. ............................................ 120
Figure 4.4. Relationship between gene diversity and species elevation range for the
reduced dataset. ............................................................................................................... 121
8
List of Tables
List of Tables
Table 2.1. Area and speciation statistics, by taxonomic group. ........................................ 33
Table 2.2. Area-only models for the probability of speciation. ........................................ 46
Table 2.3. Model-averaged parameter estimates and relative importance values for
analyses including archipelagos. ....................................................................................... 50
Table 2.4. Model-averaged parameter estimates and relative importance values for
analyses excluding archipelagos. ....................................................................................... 51
Table 3.1. Study clade pairs, with genera they include and currently accepted species
richness. ............................................................................................................................ 72
Table 3.2. Species collected, with distributions and habitats. .......................................... 73
Table 3.3. Accession numbers for matK sequences used to build study species phylogeny.
............................................................................................................................................ 79
Table 3.4. Details of AFLP method used for each study species. .................................... 88
Table 3.5. Fst values for all study species. ........................................................................ 96
Table 4.1. Measures of levels of genetic drift, local adaptation and overall phenotypic
divergence for all study species. ..................................................................................... 118
9
List of Equations
List of Equations
Equation 4.1. Nei’s gene diversity...................................................................................113
Equation 4.2. Pst ...............................................................................................................114
10
Chapter 1. Introduction
Chapter 1. Introduction
One of the biggest mysteries in biology is why groups vary so much in diversity.
Species are unevenly distributed among groups of organisms at all hierarchical levels
(Dial and Marzluff 1989; Marzluff and Dial 1991). The existence of this pattern, also
called imbalance, is not surprising, as it is predicted by null models of diversification
(Raup et al. 1973; Farris 1976). However, not all imbalance seen in nature can be
explained by chance, as many taxa are more imbalanced than expected under null models
(Dial and Marzluff 1989; Marzluff and Dial 1991). This must be the result of variation
between groups in speciation or extinction rates or in limits to group diversity, and so
understanding what controls speciation, extinction and diversity limits is the key to
understanding variation in diversity. This question has been a focus of research for
decades but clear answers are still lacking.
Comparative methods for studying variation in diversity
The study of diversity patterns is fundamentally a comparative field, and the
comparative methods used have advanced greatly over the last decades (reviewed in
Sanderson and Donoghue 1996; Ricklefs 2007a). The earliest studies simply compared
current species richness of taxa (for instance, plant families), and tested for an association
between species richness and taxon traits (e.g. Herrera 1989; Tiffney and Mazer 1995). It
was soon realised, however, that this approach has many drawbacks: the taxa compared
are likely to be nonmonophyletic (if defined based on morphology), nonindependent
(because relationships between them are not taken into account) and noncomparable
11
Chapter 1. Introduction
(because genera, families, etc. are not defined consistently and vary in age) (Isaac et al.
2003). The solution, first argued for by Felsenstein (1985), was to take account of
phylogenetic relationships, and this has been a part of comparative studies of diversity
ever since. One of the simplest methods of taking phylogeny into account is to use sister
group comparisons, in which the relationship between a trait and group diversity is
assessed over pairs of clades that are each other’s closest relatives. Sister group
comparisons have a number of advantages: the sister pairs are independent, the use of
replicated pairs increases the power of the test, diversification rates are truly compared
because sister pairs are the same age by definition, and many confounding variables are
controlled for because sister pairs share most of their history (Barraclough et al. 1998).
However, sister pair comparisons do not make use of all information in a phylogenetic
tree, and they cannot separate the effects of speciation and extinction (Barraclough and
Nee 2001). The first problem can be dealt with by making sister group comparisons over
all nodes in a tree (Felsenstein 1985; Isaac et al. 2003) or by using whole-tree approaches
to test whether appearances of a trait are associated with shifts in diversification rate (e.g.
Nee et al. 1992; Chan and Moore 2002). Separating the effects of speciation and
extinction is a more difficult problem. It has been addressed by using lineage through
time plots (Harvey et al. 1994; Barraclough and Nee 2001) and likelihood models that
include parameters for both speciation and extinction (Nee et al. 1994; Nee 2006;
Ricklefs 2007a), but recent evidence suggests fossil data may always be required
(Quental and Marshall 2010; Rabosky 2010) and the best approach is likely to combine
phylogenetic and fossil data (Purvis 2008). In summary, there is now a wide array of
methods available for measuring diversification rates and testing their association with
12
Chapter 1. Introduction
factors hypothesised to affect diversification, giving the study of diversity patterns a solid
and rigorous foundation.
What controls variation in diversity? Organism traits versus
environmental variables
The first studies testing hypotheses for taxonomic variation in diversity searched
for “key innovations” (single traits that trigger bursts of increased diversification) by
comparing the diversification of taxa with and without traits of interest. This focus
developed from Simpson’s (1953) idea that increased diversification is driven by
ecological opportunity resulting from availability of a new region, extinction of
competitors or colonisation of a new “adaptive zone” as the result of a newly evolved key
innovation (Futuyma 1986). However, the idea of adaptive zones was later criticised as
tautological because they are identified by the niches that organisms occupy (Cracraft
1982) and the range of traits studied with this method broadened to include those that
might affect rates of speciation or extinction directly, rather than by opening new adaptive
zones. Such trait-based studies are still carried out today, and many have found
significant associations (reviews in Coyne and Orr 2004; Jablonski 2008). For example,
phytophagy in insects is strongly associated with increased diversification (Mitter et al.
1988) and increased diversification in angiosperms is associated with floral nectar spurs
(Hodges and Arnold 1995). More recently investigated traits include biotic pollination
(Kay and Sargent 2009) and the lability of ornamentation in birds (Cardoso and Mota
2008). However, no individual trait studied so far has had great success in explaining
variation in diversity; even multivariate trait models are able to explain only 10-24% of
13
Chapter 1. Introduction
variation in species richness among groups (Phillimore et al. 2006). Furthermore, key
innovation-type traits are usually specific to one taxon, and so cannot help in establishing
general theories of diversification.
Another approach to studying variation in diversity emerged from the field of
ecology, where there has long been an interest in identifying environmental correlates of
regional species richness. Eventually, it was recognised that these environmental
correlates might affect diversification rates as well as ecological limits to species
coexistence. One of the earliest theories of how environmental factors could affect
diversification rates was proposed by Cracraft (1982, 1985), who placed greatest
importance on geological complexity (which he hypothesised would affect speciation
rates through its association with numbers of barriers) and “environmental harshness”
(which would control extinction rates). Since then, much research has explored effects of
environmental variables on diversification, and the same variables identified as important
in structuring regional diversity have also proved important in structuring taxonomic
diversity. Foremost on the list in both cases are area, time, latitude, habitat diversity and
topographical diversity (Rosenzweig 1995; Mayhew 2007; Ricklefs 2007b).
Environmental variables typically explain much more variation in species richness than
do taxon traits (for example, area alone explains 40% of variation in angiosperm sister
family diversity, Davies et al. 2004). However, like taxon traits, they are not sufficient on
their own for explaining patterns of diversity.
Researchers are increasingly realising that diversification is best understood as the
result of the interaction of environmental variables and taxon traits. In angiosperms, for
example, biotic pollination is correlated with increased diversification (as mentioned
14
Chapter 1. Introduction
above), but usually only in concert with topographical, edaphic or climatic diversity (Kay
and Sargent 2009). In birds, species richness is thought to be heavily influenced by the
size, geographic complexity, latitude and age of regions occupied by clades, but also by
their mating system and levels of sexual dimorphism (Ricklefs 2003, 2007a, b). The
diversification histories of taxa at lower hierarchical levels are also often best explained
by a combination of environmental variables and organism traits. For example, in the
gentian genus Halenia, nectar spurs are associated with increased diversification, but only
in clades that colonised a new biotic environment, South America (von Hagen and
Kadereit 2003). In addition, some traits may influence diversification mainly via their
effect on clade or species range size (Vamosi and Vamosi 2010).
One factor that, until recently, has rarely been explicitly discussed is the difficulty
in distinguishing whether variation in species richness is the result of variation in
diversification rates or in ecological limits to diversity (Rabosky 2009a, b). With perhaps
some exceptions (e.g. the propensity for sexual selection, which is likely to affect
diversity only through speciation rates), all ecological variables and taxon traits that have
been studied so far could affect diversity either through diversification rates or diversity
limits. This topic is discussed in more detail in Kisel et al. (in review) (Appendix I).
A proposed framework for understanding variation in diversification
For general models of variation in diversity, environmental variables that have
been studied have an advantage over most taxon traits studied because they are relevant to
all taxa. Whereas all taxa occur in regions of a particular area, latitude and topographic
15
Chapter 1. Introduction
complexity, not all taxa can be winged, phytophagous or biotically pollinated. However,
some general taxon traits have been studied, and these traits tend to show similar levels of
importance across disparate taxa in structuring diversity. For example, molecular rates
have only weak support in driving diversification in both plants (Jobson and Albert 2002;
Davies et al. 2004) and birds (Cardillo et al. 2005), whereas poorer dispersal ability is
significantly associated with increased diversification in both birds (Bohonak 1999;
Belliure et al. 2000) and marine invertebrates (Jablonski 1986). This suggests that general
taxon traits have promise for constructing a general theory of diversification.
I propose that a simple way for understanding taxonomic variation in
diversification rates is to consider them the product of available area and a general taxon
trait, the spatial scale of speciation. The spatial scale of speciation is defined here as the
minimum size of region (or amount of geographic isolation) required for speciation to
occur. Area is known to be a major factor influencing diversification (discussed further
below), and if clades vary in their spatial scale of speciation, this should strongly affect
the rate at which they can diversify within a region of a particular area. Further, I predict
that the spatial scale of speciation is controlled by levels of population genetic processes
(particularly natural selection, gene flow, mutation and genetic drift) because these
control the process of speciation (Grant 1981). This conceptual framework is illustrated in
Figure 1.1.
16
Chapter 1. Introduction
Figure 1.1. Hypothesised framework linking geographic area and the spatial scale of
speciation as controllers of speciation and diversification. Population genetic processes
that control population divergence determine the spatial scale of speciation; the
interaction of area and spatial scale of speciation determine the speciation rate, which
influences diversification rate and total diversity.
Many lines of evidence support a strong role for geographic area in controlling
diversification. First, many taxa show bursts of diversification after dispersing to new
areas. This is best exemplified by adaptive radiations on oceanic archipelagos (e.g. Parent
and Crespi 2006), but the same pattern has also been seen in the histories of mainland
groups (e.g. Kodandaramaiah and Wahlberg 2007). This finding is supported by
phylogenetic studies showing increased diversification in groups with greater dispersal
ability, where this indicates the ability to colonise new areas (e.g. in birds, Phillimore et
17
Chapter 1. Introduction
al. 2006). Second, there is palaeontological evidence for a link between changes in
species richness through time and changes in area of suitable habitat available. One of the
best-studied examples of this is in the near-shore marine environment, in which global
fossil species richness through time correlates with changes in continental shelf area
(although it is difficult to untangle the effects of increased area and increased
fragmentation, and there is debate over whether this is only the result of variation in the
amount of fossil-bearing rock preserved; Valentine and Moores 1970; Smith 1988; Smith
2007). Third, speciation rates in birds, snails and cichlid fish are known to increase with
the area available (eSARs; Losos and Schluter 2000; Seehausen 2006; Losos and Parent
2009). Finally, there is evidence that larger clade geographic ranges are associated with
higher diversification rates and/or species richness (Gaston and Blackburn 1997; Cardillo
et al. 2003; Price and Wagner 2004; Phillimore et al. 2007; Vamosi and Vamosi 2010).
The idea of a spatial scale of speciation has not been set out before, but some
recent papers studying diversification on islands suggest that it exists and varies between
taxa. Coyne and Price (2000) showed that there is no evidence of bird speciation on any
isolated oceanic island smaller than 10,000 km2. Although they used this as evidence that
sympatric speciation is rare or nonexistent in birds, it is also evidence that no barriers
sufficient for allopatric speciation are available for birds in regions of this size. Losos and
Schluter (2000), studying Anolis lizards on Caribbean islands, found evidence of
speciation, but never on islands smaller than 3,000 km2. Losos and Parent (2009) found
that the threshold area for speciation was only 18.1 km2 for snails in the Galapagos, while
Seehausen (2006) found evidence for cichlid fish speciation even in lakes with surface
areas smaller than 1 km2. These studies suggest that birds have a coarse scale of
18
Chapter 1. Introduction
speciation, lizards a medium scale of speciation, and snails and cichlid fish, relatively fine
scales of speciation.
The spatial scale of speciation should affect speciation rates by determining both
whether speciation can occur in a given region and the number of speciation events that
can occur simultaneously in regions large enough for it to occur. In a group with a fine
spatial scale of speciation, populations are able to diverge even if separated by short
distances, so even in a small region many populations could be diverging at any point in
time and the overall rate of speciation should be high. In contrast, in a group with a coarse
spatial scale of speciation, populations are only able to diverge if separated by great
distances, which means that even in larger regions only a few populations will be
diverging at any time and the overall rate of speciation should be low. Furthermore, in
such a group speciation in small regions will occur only rarely, under unusual
circumstances.
Population divergence and speciation are affected by many environmental factors
other than area or geographic isolation, and so the spatial scale of speciation should be
modulated by additional environmental factors as well. For example, given a particular
area, the spatial scale of speciation should be finer in more habitat-rich regions where
there is greater opportunity for population divergence as the result of divergent selection.
As a result, the spatial scale of speciation of each taxon is likely to vary among regions.
This complicates studies of the spatial scale of speciation, as regional characteristics must
be considered in addition to taxon traits. However, this could also mean that the spatial
scale of speciation mediates the effects of many environmental factors on diversification,
19
Chapter 1. Introduction
in addition to area, which would make a theoretical framework based on the spatial scale
of speciation even more useful.
Population genetic characteristics of clades should be closely associated with the
spatial scale of speciation because speciation always requires population divergence
(Gavrilets 2003). Although there are numerous concepts of what “speciation” means, all
of these centre on a scenario of population divergence within a species resulting in two
new units that are sufficiently separated to be considered new species (in the biological
species concept, by development of reproductive isolating mechanisms; Mayr 1942).
There is much debate over roles of selection (Schluter 2000, 2009; Sobel et al. 2010) and
geography/gene flow in driving speciation (Butlin et al. 2008), and whether speciation
requires divergence in the whole genome or only some genes (Wu and Ting 2004), but all
models of speciation nonetheless require some amount of divergence in at least one trait.
Ultimately, this divergence is controlled by population genetic processes; it is generated
by divergent selection, genetic drift or mutation, and it is prevented by gene flow and
balancing selection (Grant 1981). For example, in the classic allopatric model of
speciation, selection in geographically isolated populations drives divergence that
incidentally results in reproductive isolation (Mayr 1942); in polyploid speciation,
chromosomal multiplication produces (usually) instant and simultaneous genetic
divergence and reproductive isolation (Stebbins 1950); and in speciation by sexual
selection, selection in separated populations drives divergence in mate choice traits that
cause reproductive isolation as well (Coyne and Orr 2004). For this reason, I hypothesise
that the population genetic characteristics of clades should be associated with their spatial
scale of speciation and rates of speciation and diversification. Groups with, on average,
20
Chapter 1. Introduction
stronger divergent selection, increased genetic drift, faster mutation rates and/or reduced
gene flow within species should have finer spatial scales of speciation (and faster rates of
diversification).
Approach and aims
I use two complementary approaches to investigate the spatial scale of speciation
and its relationship with diversification.
First, in chapter two, I investigate taxonomic variation in the spatial scale of
speciation and test whether it can be explained by variation between taxa in the level of
gene flow. I do this by surveying speciation events on isolated oceanic islands worldwide
for a wide range of animal and plant taxa, and using a large set of gene flow data
compiled from the population genetic literature. I hypothesise that groups with increased
average levels of gene flow within species should have coarser spatial scales of
speciation.
Second, I test the association between diversification rates of clades and
population genetic characteristics of species, which I expect to affect diversification via
their effect on the spatial scale of speciation. In chapter three I test whether the level of
gene flow has an effect on rates of diversification and overall clade diversity; in chapter
four I test the effects of the levels of local adaptation and genetic drift. These analyses use
population genetic data I generate for species from sister groups of Costa Rican orchids
using AFLP genotyping and morphometric analyses of their leaves. I hypothesise that
21
Chapter 1. Introduction
groups with reduced gene flow, increased genetic drift and/or increased local adaptation
within species should have higher rates of diversification and greater species richness.
Throughout, I use molecular data to estimate the population genetic characteristics
of clades. To my knowledge, molecular data have never been used in comparative studies
of diversification. Instead, such studies have used morphological proxies, such as sexual
dichromatism to estimate the past strength of sexual selection. In large part this is due to
the previous difficulty in generating population genetic data. However, PCR-based
approaches for genotyping individuals, such as AFLP and RAPD, have made population
genetic data much easier to generate, and many more are now available in the literature.
Comparative population genetics is a growing field and has the potential to contribute to
the study of diversification patterns.
A final note: in discussing my proposed framework, I have focused on the
contribution of speciation to variation in diversity, but extinction is also important in
controlling diversity. It is clear that extinction rates vary nonrandomly between taxa
(Purvis 2008). Furthermore, population genetic characteristics of clades are expected to
be associated with extinction as well as speciation rates (Frankham et al. 2010), so an
effect of population genetics on diversification would not alone be evidence for my
framework. However, if population genetics affect diversification rates via speciation
rates, the associations should be opposite to those resulting from population genetics
acting via extinction. While increased speciation rates are expected with smaller
populations (and thus increased genetic drift) and reduced gene flow, decreased extinction
rates are expected with larger populations and increased gene flow. Furthermore, whereas
increased speciation is expected with increased local adaptation, there is no simple
22
Chapter 1. Introduction
expectation for any association between selection and extinction. Thus, if I find an
association between population genetics and diversification, I should be able to
distinguish whether this is the result of variation in speciation (thus matching the
expectations of my proposed framework) or the result of variation in extinction.
Summary of aims

To measure taxonomic variation in the spatial scale of speciation

To test whether population genetic characteristics of species control the spatial
scale of speciation

To test whether population genetic characteristics of species affect diversification
rates of clades
23
Chapter 2. Measuring the spatial scale of speciation
Chapter 2. Using oceanic islands to measure the spatial scale
of speciation and its association with gene flow
Introduction
Although area is generally expected to affect speciation rates (MacArthur and
Wilson 1967; Endler 1977; Rosenzweig 1995), most work on the spatial context of
speciation has focused on patterns of range overlap between emerging species, ignoring
questions of geographical scale (Mayr 1942; Butlin et al. 2008). Geographical theories of
speciation predict that the probability of speciation occurring within a given region should
(1) increase with the size of the region – because of greater opportunity for divergence to
occur within the region (MacArthur and Wilson 1967; Endler 1977; Rosenzweig 1995;
Gavrilets and Vose 2005; Losos and Parent 2009) - and (2) increase as the level of gene
flow decreases, for example among organisms with shorter dispersal distances. Gene flow
is the main process opposing population differentiation (Mayr 1963), and so the level of
gene flow between populations is expected to be an important determinant of the spatial
scale at which genetic divergence and speciation can occur (Slatkin 1973, 1985; Doebeli
and Dieckmann 2003). However, despite the potential of this body of theory for
explaining taxonomic and geographic variation in biodiversity (Ricklefs 2007b), the
extent to which the scale of speciation varies among taxa and the causes of such variation
remain unknown.
This chapter was published in slightly modified form as
Kisel, Y. and T. G. Barraclough. 2010. Speciation has a spatial scale that depends on
levels of gene flow. American Naturalist 175: 316-334.
24
Chapter 2. Measuring the spatial scale of speciation
Oceanic islands are useful for studying speciation because their well-defined
boundaries and isolation make it easier to distinguish within island (in situ) speciation
from immigration than in continental regions. Several studies have used islands to explore
the relationship between speciation rates and area. Diamond (1977) noted the lack of bird
speciation in Pacific landmasses smaller than New Zealand, but also observed that
insects, lizards and ferns had diversified within smaller islands, such as New Caledonia.
Coyne and Price (2000) found no evidence for speciation in birds within oceanic islands
worldwide smaller than 10,000 km2, setting a lower bound for their minimum area for
speciation. Losos and Schluter (2000) estimated the minimum area for speciation in
Caribbean Anolis lizards as 3,000 km2 and found that speciation rates increased linearly
with island area above this limit. Similar relationships were found in cichlid fish in
African lakes (Seehausen 2006) and Bulimulus snails in the Galapagos (Parent and Crespi
2006; Losos and Parent 2009), but with different minimum areas for speciation (in
cichlids, < 1 km2; in snails, 18.1 km2). Together with case studies of speciation on small
islands (cichlids in crater lakes, Schliewen et al. 1994; Barluenga et al. 2006; palms on
Lord Howe, Savolainen et al. 2006), these studies suggest that the spatial scale of
speciation varies widely among taxa. However, only a few taxa have been investigated,
and those on different sets of islands: comparison of several taxa across a broad range of
island sizes is needed to quantify taxonomic variation in the spatial scale of speciation
and identify its cause. To address this, I survey speciation events for a broad range of taxa
on oceanic islands from around the world.
Islands also vary in many other factors that might affect rates of diversification
(Carlquist 1974; Bauer 1988; Paulay 1994; Rosenzweig 1995). Even if a speciation-area
25
Chapter 2. Measuring the spatial scale of speciation
relationship exists, it need not be the result of area directly - for instance, larger islands
tend to have higher habitat diversity, which could foster higher rates of ecological
speciation (Losos and Parent 2009). Island age might also affect diversification, by
increasing the time over which speciation can have occurred or through other effects
related to the dynamics of island ageing (Emerson and Oromi 2005; Sequeira et al. 2008;
Gillespie 2004; discussed in Whittaker et al. 2009). In addition, the degree of isolation
from other landmasses might affect speciation rates, if lower colonisation rates to more
isolated islands leave more niches open to be filled by in situ speciation (Gillespie and
Baldwin 2009). My aim here is to use islands as a model to study the spatial scale of
speciation, rather than to explain island diversification in all its detail. However, because
these other factors - especially habitat diversity and age - are likely to be correlated with
island area, I include them in my analysis to be able to separate out the effects of area
itself.
Another complication is the existence of archipelagos. As extreme examples of
habitat fragmentation, archipelagos are expected to promote higher levels of
diversification, especially for taxa that disperse well over land but not water (Diamond
1977; Losos and Parent 2009). However, the degree to which the rate of speciation will
be increased within an archipelago should depend on the dispersal ability of the taxon and
the size of the water gaps between islands. For some taxa, barriers within islands may be
sufficiently strong isolating factors already that rates of speciation are no higher in
archipelagos than in single islands of comparable size. Therefore, I repeat my analyses
both including and excluding isolated archipelagos, to test their effect on the probability
of speciation.
26
Chapter 2. Measuring the spatial scale of speciation
Many traits of organisms and species have been hypothesised to affect rates of
speciation (Jablonski 2008), but viewed in a geographic context, dispersal ability is
expected to be key. This has especially been argued in the specific context of oceanic
islands (Diamond 1977; Paulay 1994; Ranker et al. 1994; Parent and Crespi 2006;
Whittaker and Fernandez-Palacios 2007; Gillespie and Baldwin 2009; Givnish et al.
2009), where there are many examples of spectacular radiations of taxa with normally
poor dispersal abilities but a propensity for passive long-distance dispersal (for example,
weevils on Rapa, Paulay 1985, and snails in Bonin, Chiba 1999). Diamond (1977) argued
more specifically that dispersal ability might determine the threshold island area
necessary for within island speciation to take place – but this idea remains untested.
In this study I use a comparative approach to measure the extent of variation in the
spatial scale of speciation and to test the importance of gene flow in controlling this
variation. To quantify the speciation-area relationship and the spatial scale of speciation, I
survey the probability of in situ speciation on islands of different sizes for angiosperms,
bats, birds, mammals of the order Carnivora, ferns, lizards, Macrolepidoptera (large
butterflies and moths) and land snails. These taxa were chosen based on the availability of
required data (see Methods) but they also represent a broad taxonomic range that varies in
presumed dispersal ability. As a measure of the probability of in situ speciation on
islands, I use the proportion of endemic lineages derived from single immigration events
that have diverged within an island into two or more descendent species (Coyne and Price
2000; Figure 2.1). I also test the relationship between the probability of in situ speciation
and other island factors that could potentially confound my analysis of the speciation-area
relationship.
27
Chapter 2. Measuring the spatial scale of speciation
After establishing the extent of variation in the spatial scale of speciation, I test
the importance of gene flow in setting the spatial scale of speciation by correlating the
minimum area for speciation in each group with an independent measure of the average
level of gene flow derived from the population genetic literature. To get comparable
estimates of the level of gene flow for the study taxa, Fst values are compiled from the
molecular ecology literature for each taxon along with a measure of the geographical
scale of each study. Fst provides a measure of the genetic differentiation of populations
within a species (0 = no differentiation; 1 = complete differentiation; Wright 1931) that
should be robust to variation in the spatial arrangement of populations and the type of
genetic marker used for analysis (Beaumont and Nichols 1996). It correlates strongly with
broad-scale taxonomic variation in dispersal ability, and more consistently so than other
population genetic measures such as Nm (Bohonak 1999). When estimated from
appropriate data - namely, from neutral loci unaffected by selection, and from populations
that are currently connected by gene flow and that have not undergone dramatic historical
movements (Barton 2001) - Fst values provide a measure of the level of gene flow that is
comparable between species.
Here I test two main hypotheses: that across a variety of major taxa, there should
be a positive relationship between area and the probability of in situ speciation; and that
variation between taxa in the minimum area required for speciation is controlled by
variation in the average level of gene flow within species.
28
Chapter 2. Measuring the spatial scale of speciation
Figure 2.1. Patterns of diversification on islands. A and D are two species in the same
genus, both native to the mainland. B, C, E, F and G (in italics) are endemic island
species. (a) Illustration of the relationship between patterns of diversification and
numbers of endemic species. (1) shows in situ, cladogenetic speciation (Stuessy et al.
2006), in which a mainland species reaches an island and subsequently splits within the
island into two new species. (2) shows how multiple colonisation followed by anagenetic
change, with no diversification within the island, can create the same pattern of multiple
endemic species within one genus (Coyne and Price 2000; Stuessy et al. 2006). (3) shows
in situ, anacladogenetic speciation (Stuessy et al. 2006), in which a mainland species
reaches an island and remains unchanged while budding off an endemic daughter species.
(b) Illustration of the phylogenetic patterns resulting from these three modes of island
diversification. Each phylogeny is presented underneath the diversification mode that
produces it. Species native to islands are circled; tree nodes within the circles represent
within island speciation events.
29
Chapter 2. Measuring the spatial scale of speciation
Materials and Methods
Island selection and data collection
Islands were selected on the basis of their level of isolation - only islands at least
100 km away from any other landmass, including continents and other islands, were used.
I used only isolated islands in order to minimise the chance of continuing gene flow from
outside populations of colonising species or of multiple colonisation leading to apparent
speciation when no diversification has occurred within the island. Of the islands for
which appropriate data were available, only Easter Island (Rapa Nui) was excluded,
because of its long history of habitat degradation and human-caused extinction before
discovery by European taxonomists (Diamond 2005).
Isolated archipelagos were used as units of study when the longest leg in the
minimum network connecting all islands was less than 100 km long, following the logic
of Coyne and Price (2000). In all cases, distances between islands were less than half the
distance of the archipelago from the nearest other landmass. The area of each archipelago
was calculated as the area of a 15-19 point minimum spanning polygon enclosing all
islands in the archipelago, using the Analyzing Digital Images software package (Pickle
and Kirtley 2008). All analyses were repeated both with and without the inclusion of
archipelagos. In the multiple regression models including archipelagos, I tested whether
speciation is more likely within archipelagos than single islands by including a term,
“ArchYN” coded as a 0 (single island) or 1 (archipelago). Except where indicated, the
results from analyses including archipelagos are presented.
30
Chapter 2. Measuring the spatial scale of speciation
Data on island area, elevation, isolation and latitude were largely collected from
the United Nations Environment Programme (UNEP) online Island Directory database.
Isolation was calculated following UNEP methods for “Isolation Index” (Dahl 2004).
Island data missing from the UNEP database were collected from primary literature or
government databases where possible. Island statistics and references are given in
Appendix Table A1, which can be found in the Dryad online data repository
(http://datadryad.org/repo/handle/10255/dryad.887).
Island species data collection
Study taxa were chosen based on the availability of sufficient and comparable
data, while also aiming to represent a broad taxonomic range of plants and animals with
presumed differences in dispersal ability. In the context of this study, “bats” refers to the
order Chiroptera. “Ferns” refers to the class Filicopsida, phylum Pteridophyta excluding
Psilotopsida, Lycopsida and Equisetopsida. “Lizards” refers to the order Squamata
excluding snakes and amphisbaenians, and “snails” refers to terrestrial pulmonate snails
in the orders Stylommatophora, Mesurethra, Heterurethra and Sigmurethra.
“Macrolepidoptera” refers to butterflies and moths in the superfamilies Bombycoidea,
Lasiocampoidea, Axioidea, Calliduloidea, Hedyloidea, Drepanoidea, Geometroidea,
Hesperioidea, Mimallonoidea, Noctuidea, Papilionoidea, Sphingoidea and Uranioidea.
Data were collected for all group/island combinations for which complete (to the
best ability of the source authors) taxon lists, with endemism information, were found in
the libraries of the Natural History Museum, London, the Royal Botanic Gardens, Kew,
in online sources (Avibase, Lepage 2008; Flora of Australia Online, Australian Biological
31
Chapter 2. Measuring the spatial scale of speciation
Resources Study 2008) or in databases available to me (island lizards and carnivores
worldwide, S. Meiri, unpublished data; endemic plants of selected islands, R. SalgueroGomez, unpublished data; checklist of Mascarene plants, C. Thebaud, unpublished data).
The large scope of the study prohibited comprehensive survey of recently published
journal articles and so I may not have found the most recent species lists in a few cases.
Data were not available for all taxa on all islands included in the analysis (numbers of
islands with data for each taxon given in Table 2.1).
For each group/island combination where data were available, the name and
endemism status of each native species were recorded according to the source. Therefore,
in practice I used the species concepts held by the taxonomists who wrote the species
lists. Species whose endemic status was in doubt were treated as non-endemic. Genera
with apparent speciation events are listed in Appendix Table A2 and sources for these
species data are summarised in Appendix Table A3 (both available in Dryad).
32
Chapter 2. Measuring the spatial scale of speciation
Table 2.1. Area and speciation statistics, by taxonomic group. All island sizes above 1 are
shown rounded to the nearest km. Numbers in parentheses are from dataset excluding
archipelagos.
Minimum area for
2
Number of islands in
Number of islands in
dataset with endemic
dataset with speciation
Group
Speciation (km ) *
species
events
Snails
0.8
30 (17)
24 (12)
Angiosperms
15
32 (24)
21 (14)
Ferns
15
17 (11)
9 (5)
Birds
64 or 705
50 (33)
9 (1)
108
27 (14)
10 (3)
Macrolepidoptera
141,200
6 (3)
1 (0)
Bats
416,400
14 (5)
2 (1)
Carnivora
587,713
2 (2)
1 (1)
Lizards
*Estimated as the area of the smallest island with a speciation event: snails - Nihoa; angiosperms
and ferns - Lord Howe; birds - Norfolk (smaller value) or Tristan da Cunha; lizards - Rodrigues;
Macrolepidoptera - Fiji; bats - New Zealand; Carnivora - Madagascar.
Identification of speciation events
Following the method of Coyne and Price (2000), my measure of the probability
of speciation within a given island for a given taxon is the proportion of endemic lineages
derived from single immigration events that have diversified within an island into two or
more descendent species. This approach controls for differences among islands in the
number of colonisers because it divides the number of speciated lineages by the total
number of lineages that colonised the island and could have speciated. I used a binary
33
Chapter 2. Measuring the spatial scale of speciation
measure because my interest is in what controls the ability of lineages to speciate at all,
rather than what controls the size of radiations.
I considered the number of genera with at least one endemic species to represent
the number of endemic lineages, and the number of genera with two or more endemic
species to represent the number of lineages which have diversified in situ (Coyne and
Price 2000; Stuessy et al. 2006). Therefore, my measure of the probability of speciation
on a particular island was the number of genera with two or more endemic species
divided by the number of genera with one or more endemic species. I used only genera
with endemic species (and not all native genera) in order to exclude lineages that have not
been isolated enough from mainland populations or have not been on an island long
enough to speciate within an island. Because I used only genera with endemic species, I
excluded the following islands and island groups from my dataset, for which I found no
record of endemic species in my study groups: the Bounty Islands, Caroline Island,
Cartier Island, Cocos (Keeling), Diego Garcia, the Gilbert Islands, the Hall Islands, Heard
and McDonald, Niuatoputapu and Tafahi, the Prince Edward Islands, Tokelau and Uvea.
The measure used ignores any cases of in situ speciation that have occurred
through anacladogenesis (in which a daughter species diverges from an ancestral
colonising species that remains unchanged; Figure 2.1; Coyne and Price 2000; Stuessy et
al. 1990), because the detailed morphological and genetic data required to identify such
cases were not available for most genera. The rate of anacladogenetic speciation should
vary with island characteristics in the same manner as the rate of cladogenetic speciation
(which I am measuring) as they are the result of essentially the same process of
34
Chapter 2. Measuring the spatial scale of speciation
divergence. Thus I expect that the inclusion of such speciation events would not
qualitatively change my conclusions.
The minimum area for speciation of each taxon was estimated as the area of the
smallest island or archipelago within which speciation has occurred. Statistics on
speciation and endemism in each group are given in Table 2.1 and Appendix Table A3
(available in Dryad).
Adding phylogenetic information
My method assumes that the chance of multiple endemic species within the same
genus originating by multiple colonisation events, rather than in situ diversification
(Figure 2.1), is small. To validate this assumption, I searched the literature for
phylogenetic information on the study genera. For each genus associated with a putative
speciation event, searches were performed on TreeBase, NCBI GenBank CoreNucleotide
and ISI Web of Science to find published molecular phylogenies. Any phylogenies that
included more than one of the endemic species of a study genus with a putative speciation
event on a particular island or archipelago were used. Cases of congeneric endemic
species shown not to be each other’s closest relatives were reclassified as multiple, nonspeciated lineages (see Losos and Parent 2009). In a few instances, published phylogenies
indicated that multiple genera containing endemic species were all part of one larger
endemic clade, and in this case the genera involved were considered as a single speciated
lineage in the analyses. Also, in some cases published phylogenies indicated that the
endemic species in one genus were the result of multiple colonisations followed by
multiple radiations, in which case the genus was treated as multiple speciated lineages.
35
Chapter 2. Measuring the spatial scale of speciation
Results of the phylogeny search are given in Appendix Tables A2 and A4 (both available
in Dryad).
Statistical analysis of the speciation-area relationship
Speciation was treated as a binary response variable: each endemic lineage on
each island has a value of 0 for speciation if it contains only 1 endemic species and a
value of 1 if it contains 2 or more endemic species. Overall regression models between
this response variable and island area, considering all taxa together, were carried out
using the lme4 package in R v.2.5 (Bates 2007; R development core team 2007) for
generalised linear mixed effects models, using taxon as a random effect, a binomial error
structure and Laplace approximation for ML (maximum likelihood) estimates. Individual
regression models for each taxon were carried out in R using generalised linear models
with a binomial error structure. R2 values for all models were calculated with the formula
(SST-SSE)/SST, where SSE is the deviance of the model and SST is the deviance of a
null model (for mixed effects models, consisting only of a different intercept for each
taxon, which represents the mean probability of speciation over all islands).
I also constructed multiple regression models to investigate the importance of the
other island characteristics that might affect speciation probability. For both the
individual taxon models and overall models considering all taxa together, I began with a
maximal additive model including area, elevation as a proxy for habitat diversity in the
absence of a direct measure (Ackerman et al. 2007), isolation from other landmasses and
whether the unit of study was a single island or an archipelago. I then calculated the
Akaike Information Criterion corrected for small sample sizes (AICc; Burnham and
36
Chapter 2. Measuring the spatial scale of speciation
Anderson 2002) for each sub-model of this full model. For the overall models, I ran each
sub-model once with each possible random effect term - one indicating only different
intercepts for each taxon (written as (1|group)) and one for each island environmental
variable, indicating different slopes for each taxon with this variable (for example
(Area|group) indicating different speciation-area slopes for each taxon). The best models
for each dataset are listed in Appendix Table A5 (available in Dryad).
The significance and importance of each predictor variable in the multiple
regression models was evaluated using model averaging as described in Burnham and
Anderson (2002). First, for each dataset, the full set of additive models was generated.
Then the relative importance of each variable, on a scale from 0 to 1, was calculated as
the sum of the Akaike weights of the models in which the variable appears - better models
have larger Akaike weights, and a variable which contributes more to model fit and so is
included in more of the best models will have a higher relative importance value.
Parameter estimates and unconditional standard errors for each term were calculated by
averaging over all models in which the variable appeared, weighting values from
individual models by the models’ Akaike weights. A term was considered to be
significant for a particular dataset if the 95% confidence interval for its parameter
estimate did not include 0.
Gene flow data
Fst data came primarily from the appendices in Morjan and Rieseberg (2004) and
were supplemented by a search following the methods of Morjan and Rieseberg (2004)
for (TS=((fern or pteridophyt* or snail or Lepidoptera* or Chiroptera or Carnivora or
37
Chapter 2. Measuring the spatial scale of speciation
lizard) and ("gene flow" or Fst or Nm or Nem))) on ISI Web of Science. After this search,
carnivore data were still lacking, and so data were added from another search, for (TS=
(carnivor* and "population structure")). In the additional searches, only papers that
presented an overall Fst value (as opposed to pair-wise Fst values only) for variation
between populations (rather than regions) were used. Two estimators of Fst - Gst (Nei
1973) and Φst (Excoffier et al. 1992; Excoffier 2001) - were used where these were
provided in the original sources instead of Fst.
All studies dealing with aquatic or marine species were excluded, as well as
studies of recent habitat fragmentation as a result of human activities, clonality or
hybridisation between species or host races. I also excluded studies including historicallyisolated lineages, such as those separated by a major geographical barrier (for instance,
Myotis myotis on either side of the Strait of Gibraltar; Castella et al. 2000), as they are
often evidence of cryptic speciation, while I wanted estimates of gene flow within
species. Only studies of wild populations of native organisms were used - recent
introductions and crop pests were excluded. All Fst values derived from organelle markers
(mitochondrial or chloroplast DNA) were excluded, as these reflect only female dispersal
and not overall population patterns of genetic differentiation. One plant study in Morjan
and Rieseberg’s database (Proteum glabrum; Morjan and Rieseberg 2004) had a negative
Fst value, which was interpreted as Fst = 0 (Long 1986). Gene flow data used are
summarised in Appendix Table A6 (available in Dryad).
I checked for comparability of gene flow estimates from different studies. Studies
using AFLPs, RAPDs and iSSRs were found to have significantly higher Fst values than
studies using other marker types, even when correcting for taxon and the geographic scale
38
Chapter 2. Measuring the spatial scale of speciation
of study (ANOVA, F = 7.3, p < 2.2 x 10-16) and so were excluded from analysis.
Isozymes were also excluded because they were used in only one study in my dataset and
the comparability of Fst values derived from them to Fst values derived from other
markers is unclear. The Fst values from studies using allozymes and those using types of
repeats (microsatellites or SSRs, minisatellites, tandem repeats) did not differ
significantly (ANOVA, F = 1.7, p = 0.192) and so were lumped for the final analysis.
Studies using these two marker types also did not differ significantly in number of
populations used (ANOVA, F = 1.5, p = 0.229) or number of loci used (ANOVA, F =
0.60, p = 0.440). There were small differences between the two marker types in the mean
geographic scale of study, but these were inconsistent between taxa in sign and so cannot
explain the pattern of taxon variation in Fst. Fst showed no overall relationship with the
number of loci (linear regression, p = 0.867) or the number of populations (linear
regression, p = 0.885) used in a study. Allele frequencies for individual markers were not
consistently available and thus their relationship with Fst could not be tested, but they are
not expected to be a major confounding factor.
As a measure of the geographic scale of each population genetic study, geographic
range extent was measured as the greatest distance in km between any two populations in
the study, following the example of Bohonak (1999). I used the maximum distance
between study populations rather than the mean or median because I felt this was a simple
and practical measure sufficient to resolve the wide range of geographic scales
represented by the studies I used (whose maximum distances vary from 0.01 km to
>14,000 km). In addition, use of alternative measures such as mean distance would be
difficult because of the lack of detailed information in many of the sources. This measure
39
Chapter 2. Measuring the spatial scale of speciation
does not explicitly take into account variation in species’ geographic range size or in the
spatial arrangement of sampled populations, but neither was easily quantified from the
sources available to me and neither is expected to have a strong independent effect on Fst
(Beaumont and Nichols 1996). Distances were taken directly from papers if possible.
Otherwise, they were calculated from population coordinates using the Vincenty formula
(Veness 2008), measured from scaled maps given in the studies or measured using
Google EarthTM v. 4.3 (Google) if only place names were provided. Data for 23 taxa were
excluded because the original reference could not be found or did not contain sufficient
information to calculate the geographic range extent.
Gene flow analyses
To test the effect of gene flow on the spatial scale of speciation, I correlated the ln
of the minimum island size for speciation in each study group with two summary
measures of the spatial scale of gene flow. One potential problem with comparing Fst
values is that Fst data tend to be measured at different spatial scales in different taxa
(Figure 2.2). Therefore, as summary measures I calculated the mean Fst for each group at
two scales in turn: between 10 and 100 km, and between 100 and 1000 km. I chose these
geographic scales for analysis because they correspond to the range of island sizes used in
the speciation analysis (47 of 64 islands/island groups, 73%, have maximum linear
extents between 10 and 1000 km) and because they are the only scales at which Fst data
were available for all study taxa (except snails between 100 and 1000 km). To control for
the effect of outliers in the Fst data I also tested the correlation of median Fst values with
minimum island size for speciation.
40
Chapter 2. Measuring the spatial scale of speciation
For this analysis, the minimum island size for speciation for each study group was
represented as the greatest distance between any two points of land within the island or
archipelago. These extents were measured in Google EarthTM v. 4.3 (Google) and the
Analyzing Digital Images software package (Pickle and Kirtley 2008). Linear extents
were used instead of area in order to be directly comparable to the distances between
populations used to represent the spatial scale of gene flow. Because of the uncertainty
over bird speciation on Norfolk, this analysis was carried out twice, once with Norfolk
and once with Tristan da Cunha representing the smallest island with in situ bird
speciation.
Figure 2.2 (next page). Relationship between Fst and the geographic extent of study. Fst
estimates are taken from a review of population genetic literature. Each point represents
an estimate from a single study, and only values used in analyses are shown. The
geographic extent of study is the maximum distance between populations in one study. Fst
values given are overall values considering all populations in the study. Dashed lines
delimit the two geographic scales that were used for the gene flow analysis. Solid black
lines indicate the average Fst for each scale and solid grey lines indicate the median.
Where only a grey line is visible, the mean and median values are equal. No data were
available for snails at the 100 - 1000 km range.
41
42
Figure 2.2.
Chapter 2. Measuring the spatial scale of speciation
Results
Data availability and quality
I estimated speciation probabilities across 64 islands in total, including 38 single
islands and 26 archipelagos, taking into account 471 putatively speciated genera.
Phylogenies were available for 15% of these genera; an additional 15% of the genera
were endemic to their island and so most parsimoniously explained by in situ speciation.
These data led me to exclude only seven of the putative speciated genera as being the
result of multiple colonisation (~5% of cases where phylogenetic or genus endemism
information is available), confirming that the non-phylogenetic measure is a good
measure of the number of genera that have speciated in situ and is little affected by
multiple colonisation events. Twelve genera were found to be part of a larger adaptive
radiation already represented by another included genus, and so were removed from
analysis. Three genera were found to be the result of two separate radiations, and one
genus the result of three separate radiations, and were split into multiple speciated
lineages accordingly. After taking all this into account, my final dataset included 457
speciated lineages. Phylogenetic data by genus are presented in Appendix Table A2 and
summarised in Appendix Table A4 (both available in Dryad).
Quantifying the speciation-area relationship
Across taxa, on oceanic islands and archipelagos ranging in size from <1 (Nihoa)
to >500,000 km2 (Madagascar), there is a clear positive relationship between the
43
Chapter 2. Measuring the spatial scale of speciation
probability of in situ speciation and island area (p = 1.35 x10-5, r2 = 0.312; when
archipelagos are excluded, p = 1.54 x 10-4, r2 = 0.414; Figure 2.3). The relationship
between the probability of speciation and island area is significant in all taxa with
sufficient data except ferns (Table 2.2). Bats and macrolepidoptera had insufficient data
to construct models for their speciation-area relationship, as they show evidence for
speciation only on 2 and 1 islands, respectively, but the data for both nevertheless support
the same positive relationship seen in the other taxa - both are present on many small
islands on which they have not speciated, while they have speciated only on the largest
islands on which they are represented. In contrast, carnivores are rarely present at all on
oceanic islands due to their poor dispersal over water, and have endemic species on only
two of the studied islands. Nonetheless, they also show evidence for speciation only on
the largest island on which they are represented (Madagascar), and I predict that they
would also show a positive speciation-area relationship if larger landmasses were
considered. The lack of a speciation-area relationship in ferns, on the other hand, is not
the result of a lack of data. Ferns have clearly speciated on small and large islands with
similar probability, indicating that area is relatively unimportant in controlling their
speciation.
Figure 2.3 (next page). Relationship between the probability of speciation and area. Each
point marks the percentage of lineages of a particular taxon, on a particular island, that
have speciated. Solid lines mark the regression model for each taxon alone; dashed lines
represent the overall model averaged over all taxa. No individual model line is given for
bats, carnivores or Macrolepidoptera, as these taxa had too few islands with speciation
events to model their speciation-area relationship. Lines are based on analyses including
archipelagos.
44
45
Figure 2.3.
46
64
40
32
24
50
17
11
27
14
30
17
Overall model
Overall model - excluding archipelagos
Angiosperms
Angiosperms - excluding archipelagos
Birds
Ferns
Ferns - excluding archipelagos
Lizards
Lizards - excluding archipelagos
Snails
Snails - excluding archipelagos
12
24
3
10
5
9
9
14
21
20
38
S
~ - 0.810 + 0.0951 log(Area)
~ - 0.700 + 0.0988 log(Area)
~ -5.39 + 0.464 log(Area)
~ - 4.35 + 0.391 log(Area)
~ - 0.411 - 0.0844 log(Area)
~ - 0.641 - 0.0165 log(Area)
~ - 5.07 + 0.310 log(Area)
~ - 3.14 + 0.350 log(Area)
~ - 2.27 + 0.173 log(Area)
~ - 5.89 + 0.420 log(Area) + (Area | group)
P(speciation)~ - 4.00 + 0.279 log(Area) + (Area|group)
Model
0.277
0.534
0.580
6.91 x10-9
1.66 x10-5
0.0219
0.00234
8.6 x10
events.
birds on single islands, Carnivora or Macrolepidoptera because these groups had fewer than three islands with speciation
variation between study taxa in the slope of the speciation-area relationship. Speciation probability was not modelled for bats,
estimated probability of speciation. In the overall models, the “(Area|group)” term is the random effect accounting for
Parameter values are those given by the logistic models and produce predicted values that must be logit transformed to give an
0.216
0.243
0.732
0.527
1.75 x10-5
-4
0.0149
0.704
0.882
1.75 x10
0.00158
0.414
-6
1.54 x10
0.312
-4
r2
1.35 x10-5
for area
p-value
NOTE.- N = number of islands/island groups used to construct model. S = number of islands with speciation events.
N
Taxon
Table 2.2. Area-only models for the probability of speciation.
Chapter 2. Measuring the spatial scale of speciation
Measuring minimum areas for speciation
The minimum area for speciation (estimated as the area of the smallest island or
archipelago within which speciation has occurred) varies widely among taxa. Land snails
have speciated within even the smallest island on which they have native species (Nihoa 0.8 km2), whereas the only example of in situ speciation in Carnivora is on Madagascar
(587,713.3 km2), and bats show no evidence of in situ speciation on any islands except
New Zealand (approximately 416,400 km2) and Madagascar. Macrolepidoptera also
appear to require large areas for speciation - the only island unit in which they show
evidence of speciation is Fiji (141,200 km2). Angiosperms and lizards are intermediate,
with minimum areas for speciation of 14.6 km2 and 107.8 km2 (Table 2.1; more detailed
summary of speciation events in Appendix Table A2, available in Dryad). The situation
in birds is unclear; even after genetic analysis, it is uncertain whether a putative
speciation event on Norfolk Island (64 km2) is not actually the result of multiple
colonisation (Coyne and Price 2000). The next smallest island unit within which bird
speciation has potentially taken place is the Tristan da Cunha archipelago (705 km2), in
which some evidence even supports a history of sympatric speciation within the smaller
islands of the archipelago (Ryan et al. 2007; Grant and Grant 2009); the smallest single
island with firm evidence for in situ bird speciation is Jamaica (11,400 km2; Coyne and
Price 2000). Irrespective of the uncertainty for birds, it is evident that taxonomic variation
in the spatial scale of speciation is great, spanning 6 orders of magnitude between snails
and carnivorous mammals.
47
Chapter 2. Measuring the spatial scale of speciation
Testing the importance of area when other environmental variables are
included
The relationship with area is not an artefact of area’s correlation with another
environmental variable. In my dataset, island area is correlated with elevation (adjusted
r2 = 0.478; adj. r2 = 0.365 if archipelagos are excluded), island age (adj. r2 = 0.111;
correlation is not significant if archipelagos are excluded) and whether an island is an
archipelago or not (proportion of variance explained in ANOVA = 0.278), but in multiple
regression models, both overall and for individual taxa, model-averaged parameter
estimates indicate that island area is highly important and significant, independently of
other island characteristics (best models are listed in Appendix Table A5, available in
Dryad; model-averaged parameter estimates are given in Tables 2.3 and 2.4).
In the overall models, area has high relative importance values (0.93 when
archipelagos are included; 0.97 when archipelagos are excluded; relative importance
values are on a scale of 0 to 1), meaning that it is included in a high percentage of the best
models. In addition, its parameter estimates are significantly greater than zero, supporting
a positive speciation-area relationship. Isolation and elevation also have high relative
importance values and are also significant in the overall models, especially in the dataset
including archipelagos (Table 2.3). Age and whether an island unit is an archipelago or
not have low relative importance values and are not significant in the overall models.
Area is also highly important and significant in most of the single taxon models. It
is the most important variable, and its parameter estimate is significantly greater than zero
(indicating a positive speciation-area relationship), in all single taxa except lizards in the
48
Chapter 2. Measuring the spatial scale of speciation
dataset excluding archipelagos, and ferns in both the dataset including and the dataset
excluding archipelagos. Elevation and isolation are also significant in most single taxon
models, although in all cases, except lizards when excluding archipelagos, they are much
less important than area. The parameter estimate for isolation is positive in all taxa except
birds; more distant islands tend to have a higher probability of speciation. Whether an
island is an archipelago or not is important and significant in angiosperms and birds,
while island age is relatively important and significant only for snails, when archipelagos
are excluded. Over all models, island area is the most consistently important and
significant island variable.
49
50
0.93
0.20 ± 0.0044
(0.19, 0.21)
0.99
0.22 ± 0.0042
(0.21, 0.23)
0.98
0.32 ± 0.0086
(0.30, 0.34)
0.31
-0.010 ± 0.018
(-0.045, 0.025)
0.86
0.32 ± 0.017
(0.29, 0.36)
0.78
0.074 ± 0.0020
(0.070, 0.077)
OVERALL
Angiosperms
Birds
Ferns
Lizards
Snails
0.11
0.015 ± 0.079
(-0.14, 0.17)
0.28
0.024 ± 0.018
(-0.011, 0.059)
0.31
0.095 ± 0.16
(-0.23, 0.42)
0.27
0.040 ± 0.082
(-0.12, 0.20)
0.20
0.020 ± 0.016
(-0.011, 0.051)
0.26
0.0017 ± 0.0035
(-0.0051, 0.0086)
Age
0.34
0.017 ± 0.0027
(0.012, 0.022)
0.34
0.0039 ± 0.00024
(0.0034, 0.0044)
0.34
0.0074 ± 0.0020
(0.0036, 0.011)
0.27
0.0047 ± 0.00081
(0.0031, 0.0063)
0.76
0.018 ± 0.00014
(0.018, 0.018)
0.97
0.022 ± 6.21x10-5
(0.022, 0.022)
Elevation
0.20
0.00051 ± 0.00051
(-0.00049, 0.0015)
0.21
1.76x10-5 ± 6.39x10-5
(-0.00011, 0.00014)
0.33
0.0019 ± 0.00017
(0.0016, 0.0023)
0.26
-0.00099 ± 0.00026
(-0.0015, -0.00048)
0.87
0.011 ± 2.86x10-5
(0.011, 0.011)
0.97
0.010 ± 1.29x10-5
(0.010, 0.010)
Isolation
0.24
0.10 ± 0.44
(-0.76, 0.97)
0.39
0.13 ± 0.11
(-0.077, 0.34)
0.30
0.018 ± 0.60
(-1.1, 1.2)
0.79
0.79 ± 0.26
(0.29, 1.29)
0.94
-0.66 ± 0.061
(-0.78, -0.54)
0.3
-0.025 ± 0.026
(-0.077, 0.026)
ArchYN
was found to be significant, in which case all parameter values were estimated using the reduced subset of islands with age data.
regression models using this reduced subset. Parameter values for all other terms come from regression models using the full set of islands, unless age
had fewer than three islands with speciation events. Age was only available for a subset of islands, and so parameter estimates for age come from
also highlighted in bold. Speciation probability was not modelled for bats, birds on single islands, Carnivora or Macrolepidoptera because these groups
NOTE.- The highest parameter relative importance value for each study group is highlighted in bold. Parameter estimates for significant variables are
Area
relative importance value
estimate ± standard error
(confidence interval)
MODEL
Table 2.3. Model-averaged parameter estimates and relative importance values for analyses including archipelagos.
51
0.97
0.38 ± 0.016
(0.35, 0.42)
0.9998
0.34 ± 0.0043
(0.33, 0.35)
insufficient data
0.37
-0.032 ± 0.055
(-0.14, 0.075)
0.16
-3.47 ± 1.6x108
(-3.2x108, 3.2x108)
0.40
0.033 ± 0.0051
(0.023, 0.043)
OVERALL
Angiosperms
Birds
Ferns
Lizards
NOTE.- See notes for Table 2.3.
Snails
Area
relative importance value
estimate ± se
(confidence interval)
MODEL
0.75
2.48 ± 4.35x109
(-8.5x109, 8.53x109)
0.42
0.075 ± 0.024
(0.028, 0.12)
0.23
0.010 ± 0.16
(-0.31, 0.33)
-
0.12
-0.010 ± 0.029
(-0.067, 0.046)
0.30
-0.012 ± 0.023
(-0.0587, 0.033)
Age
0.91
65.4 ± 1.6x108
(-3.1x108, 3.1x108)
0.16
-0.00048 ± 0.00047
(-0.0014, 0.00044)
0.35
-0.0020 ± 0.0025
(-0.0069, 0.0030)
-
0.22
0.0016 ± 0.00015
(0.0013, 0.0019)
0.47
0.011 ± 0.00045
(0.010, 0.012)
Elevation
0.88
54.7 ± 1.1x108
(-2.2x108, 2.2x108)
0.14
-9.72 x10-5 ± 0.00011
(-0.00032, 0.00012)
0.35
1.1x10-4 ± 0.00018
(-0.00024, 0.00047)
-
0.33
0.0021 ± 4.28x10-5
(0.0020, 0.0022)
0.64
0.0052 ± 2.83x10-5
(0.0051, 0.0052)
Isolation
Table 2.4. Model-averaged parameter estimates and relative importance values for analyses excluding archipelagos.
Chapter 2. Measuring the spatial scale of speciation
The effect of gene flow
The minimum island size for speciation in each group correlates with the mean
level of gene flow for each taxon when gene flow is measured at the scale of 10-100 km
(slope = -27.21, p = 0.00286, adj. r2 = 0.763; Figure 2.4). Taxa that are able to speciate
within smaller areas (indicated by a smaller minimum island size for speciation) are those
with reduced gene flow (indicated by higher mean Fst values). At the scale of 100-1000
km, at which snails are excluded due to lack of data, the same relationship is found,
although marginally non-significant (slope = -16.51, p = 0.0695, adj. r2 = 0.417). Widely
overlapping 95% confidence intervals (at 10-100 km: -38.22 to -16.21; at 100-1000 km:
-30.56 to -2.46) indicate that there is no significant difference in the slope of the
relationship between the two spatial scales. I found similar results using median Fst
instead of mean Fst for each group (at 10-100 km scale, slope = -24.91, p = 0.0221,
adj. r2 = 0.545; at 100-1000 km scale, slope = -20.40, p = 0.221, adj. r2 = 0.138). Using
Norfolk Island instead of Tristan da Cunha for the minimum area of speciation in birds
also did not change the results (at 10-100 km: slope = -28.45, p = 0.00144, adj. r2 = 0.810;
at 100-1000 km: slope = -14.69, p = 0.146, adj. r2 = 0.246). When archipelagos are
excluded, the same gene flow-minimum area relationship is again found at both spatial
scales, although in this case it is significant at 100-1000 km (slope = -21.82, p = 0.0191,
adj. r2 = 0.729) but not at 10-100 km (slope = -24.02, p = 0.0917, adj. r2 = 0.358). Again,
widely overlapping confidence intervals for the slope of the relationship (at 10-100 km:
-46.61 to -1.42; at 100-1000 km: -33.07 to -10.56) indicate that there is no significant
difference in the gene flow-minimum area relationship between the two spatial scales.
52
Chapter 2. Measuring the spatial scale of speciation
Figure 2.4. Minimum island size for speciation versus the average level of gene flow
when measured over geographic ranges of 10-100 km.
Discussion
Main findings
These results show that the speciation-area relationship, in which speciation is
more likely and more frequent within larger areas, is a general pattern common to many
groups of both plants and animals. Ferns are the only group that show no such
relationship, perhaps because of their higher propensity for polyploid and hybrid
53
Chapter 2. Measuring the spatial scale of speciation
speciation, the implications of which are discussed further below. The speciation-area
relationship found is not just a by-product of area’s correlation with other island
characteristics - island area is consistently important and significant in both overall and
taxon-specific multivariate models, which include also island elevation (as a proxy for
habitat diversity), age, isolation from other landmasses and whether island units are
archipelagos or single islands. Though all study taxa except ferns have in common a
positive speciation-area relationship, they vary over 6 orders of magnitude in the
minimum area required for speciation. Furthermore, this variation in the minimum area
for speciation correlates with variation among taxa in the level of gene flow. Taxa with
higher rates of gene flow, measured at a common spatial scale, have a larger minimum
area for speciation and a lower probability of speciation in any given area. This suggests
that the population genetics of divergence directly control the incidence and rate of
speciation - that there is a direct link between microevolutionary and macroevolutionary
processes.
The effects of other island characteristics on speciation probability
Though there is strong evidence for area as a major controller of speciation rates,
this does not rule out a role for other environmental variables. In particular, isolation and
elevation are also important and significant factors in the overall models, and important
and significant in most of the individual taxon models. The effect of elevation is always
positive, as predicted if greater altitudinal variation increases the number of habitats and
promotes greater ecological speciation (Ackerman et al. 2007; Losos and Parent 2009). In
all cases except birds, the probability of in situ speciation increases with increasing
54
Chapter 2. Measuring the spatial scale of speciation
isolation, consistent with predictions that lower colonisation rates of distant islands
should leave more niches available for speciation (Gillespie and Baldwin 2009). In birds,
my measure of speciation probability increases on islands closer to other landmasses,
which is unexpected according to my theoretical predictions but might arise if a low
frequency of inferred speciation events still represent multiple colonisation (because
colonisation is expected to be greater on islands closer to other landmasses). However,
isolation is the least important variable for birds and has a small effect on variation in my
measure.
Interestingly, considering the great distinction usually made between single
islands and archipelagos in island evolution theory, a significant and important effect of
archipelagos on speciation probability is found only in birds and angiosperms.
Furthermore, while the parameter estimate for birds is positive, as expected if water gaps
between islands act as additional dispersal barriers promoting speciation, the parameter
estimate for angiosperms is negative, which is unexpected on theoretical grounds. The
lack of significance and importance of the archipelago term in other taxa and in the
overall models may indicate that the difference between water gaps and ecological
barriers within islands in their strength as dispersal barriers is much greater for birds than
the other study taxa (Diamond 1977). For the other study taxa, barriers within islands may
be strong enough that diversification within a heterogeneous island is comparable to
diversification within an archipelago. Most important for my aims, the speciation-area
relationship holds irrespective of whether archipelagos are included or not.
Broad comparative studies such as this one necessarily rely on surrogates and
proxies for some underlying variables of interest, and so a lack of correlation in my study
55
Chapter 2. Measuring the spatial scale of speciation
is not conclusive evidence against any environmental factor. Further work would
particularly benefit from improved data on island ages - it is difficult to evaluate the
biological relevance of ages taken from the geological literature (for instance if lava flows
sterilise an island some time after its actual origination and emergence; Whittaker et al.
2008) and ages are lacking for many islands and island groups.
The spatial scale of speciation and gene flow
Consistent with the importance of gene flow in population genetics-based theories
of speciation, estimates of the level of gene flow explain up to 76% of the variation in the
spatial scale of speciation across taxa. Taxa with lower levels of gene flow are able to
speciate within smaller islands, suggesting that the level of gene flow determines the
spatial scale of speciation by controlling the minimum spatial extent at which
differentiation of populations can occur. This result also accounts for the existence of
thresholds in evolutionary species-area relationships (Losos and Schluter 2000) - in situ
speciation is expected to contribute significantly to local species richness only in areas
large enough that gene flow does not prohibit population differentiation.
The main limitation for this analysis was the availability of gene flow data. Past
studies have largely applied molecular markers to single species questions, and metaanalyses like this study are necessarily posterior exercises limited by available data.
While disparate studies are still comparable (Bohonak 1999; Morjan and Rieseberg
2004), targeted studies generating data for a set of species using a standardised sampling
design would allow more refined comparative analyses, including the use of more
sophisticated measures of the spatial scale and level of gene flow, such as the mean
56
Chapter 2. Measuring the spatial scale of speciation
dispersal distance predicted from the slope of an isolation by distance (IBD) regression
line for each species (Kinlan and Gaines 2003) or the Sp statistic (Vekemans and Hardy
2004). It would also be useful to have gene flow data for the specific genera and species
for which island data were collected, instead of averaging over each major taxon
(especially given the tendency of island species to evolve reduced dispersal ability,
Carlquist 1974), but these data were not available in the literature.
Because of these constraints, I could compare only a limited number of different
taxa at a relatively broad taxonomic scale, while retaining enough information to provide
reasonable sample sizes for estimating the study variables. Despite relatively low power,
the result is robust for the sample available. The significance of the relationship varied
depending on the scale used (which determined whether snails were included or not) and
whether archipelagos were included or not, but in an inconsistent way that reflected low
power rather than large changes in the underlying relationship. A significant relationship
was also found using an alternative measure that is closer to the underlying quantity of
interest but less statistically robust than mean Fst (an estimate of the minimum scale at
which neutral divergence is expected to occur within species of each major taxon, Figure
2.5).
Therefore, despite the above limitations, and the relatively low power they entail,
these results point to gene flow levels as a potentially important determinant of the spatial
scale of speciation. It remains possible, however, that the relationship found is the result
of other confounding factors that vary between the study taxa in parallel to differences in
gene flow. Incorporating more taxa, resolving the chosen taxa more finely and generating
57
Chapter 2. Measuring the spatial scale of speciation
more estimates of gene flow would be needed for more powerful tests of this hypothesis
in future.
The negative relationship found between gene flow and the probability of
speciation within a given area at first seems to contrast with ideas that either high
(Eriksson and Bremer 1991; Owens et al. 1999; Phillimore et al. 2006) or intermediate
(Price and Wagner 2004; Paulay and Meyer 2006) dispersal ability should lead to
maximum diversification. These ideas are only incompatible, however, if every species is
imagined to have a single value representing its dispersal ability. In reality, dispersal for
any taxon is usually thought of as a leptokurtic probability function, with a long tail of
infrequent long-distance dispersal events (Tilman and Kareiva 1997). Under this model,
dispersal affects diversification in two different ways - shorter-distance dispersal within
the species range maintains species cohesion, and rarer long-distance dispersal to new
areas outside the species range allows the establishment of new, potentially isolated
populations. By considering only lineages able to reach oceanic islands, I intentionally
focused on the effect of shorter-distance dispersal ability and controlled for long-distance
dispersal, namely colonisation ability.
58
Chapter 2. Measuring the spatial scale of speciation
Figure 2.5. Results of an alternative gene flow analysis - the relationship between the
minimum area for speciation and the spatial scale of neutral population differentiation.
Minimum island size for speciation is plotted against the minimum geographic extent for
each taxon at which gene flow has been observed to be low enough (Fst high enough) to
allow neutral genetic differentiation of populations (Fst = 0.2, corresponding to Nm = 1).
This is estimated for each taxon by the geographic scale of the population genetic study
with the smallest geographic scale and Fst ≥ 0.2. Macrolepidoptera are excluded because
none of the Macrolepidoptera studies in my population genetic dataset have Fst ≥ 0.2.
Evolutionary explanations for the observed patterns
Several mechanisms could produce the speciation-area relationship observed
(Gavrilets and Losos 2009). First, larger areas might offer more opportunity for
geographical isolation, either by distance alone or via barriers to dispersal (MacArthur
59
Chapter 2. Measuring the spatial scale of speciation
and Wilson 1967; Endler 1977; Rosenzweig 1995). Second, larger areas might encompass
more habitat types, which could increase speciation rates through stronger divergent
selection or by providing additional niches allowing the coexistence of newly formed
species (Losos and Parent 2009). I considered habitat variation in relation to elevation,
but other unmeasured aspects of habitat variation might also scale with area. Third, larger
areas can support larger population sizes, which might increase the rate of adaptive
evolution by increasing the rate of origin of beneficial mutations for selection to act upon
(Gavrilets and Vose 2005). Including data on population sizes might allow the third
mechanism to be distinguished, but in the absence of such information, I believe that the
relationship between gene flow and the spatial scale of speciation is most consistent with
speciation occurring through geographical isolation or ecological divergence into distinct,
spatially structured habitats (Schluter 2001).
By providing the exception to the general pattern observed, ferns strengthen the
support for these conclusions. Ferns are known to have a high incidence of speciation
through hybridisation and polyploidy (Wagner 1969; Otto and Whitton 2000), two major
processes allowing speciation to occur in the face of gene flow (Berlocher 1998). In fact,
of the two fern genera in my study with speciation events supported by published
phylogenies, one is thought to have diversified through hybridisation (Eastwood et al.
2004). In contrast, speciation as a result of hybridisation and polyploidy is rare in
animals, and important but much less frequent in angiosperms (Otto and Whitton 2000).
Thus, as expected if the speciation-area relationship is the result of gene flow-limited
divergence, the group that most frequently speciates with continuing gene flow shows no
significant speciation-area relationship. I conclude that pure sympatric speciation, namely
60
Chapter 2. Measuring the spatial scale of speciation
in the absence of any geographical isolation and in the presence of gene flow, appears to
be infrequent in all taxa except ferns (see also Barraclough and Vogler 2000; Phillimore
et al. 2008).
Extinction might also influence the relationship between diversification and area most directly because extinction rates should be higher on smaller islands with smaller
populations (MacArthur and Wilson 1967). For this reason, extinction has been used in
the past to explain the relationship between island area and the number of single island
endemic species (Mayr 1965). The effect of extinction on the speciation-area relationship
cannot be tested with the type of data presented here; it would require studies of island
taxa for which comprehensive fossil data are available and extinction rates can be
estimated directly (perhaps birds; Steadman 2006). However, I believe that the
association between decreased gene flow and increased diversification cannot be
explained easily by extinction. There are some mechanisms, such as increased pathogen
spread (Thrall et al. 2000) or swamping of local adaptation (Holt and Gomulkiewicz
2004), by which increased gene flow could increase the risk of extinction (and thereby
decrease net diversification rate), but neither of these is a necessary outcome of increased
gene flow. It is more usually expected that decreased gene flow should increase the risk
of extinction, either through increased inbreeding (Lande 1988) or decreased
recolonisation rates within metapopulations (Gaggioti and Hanski 2004). Probability of
speciation, on the other hand, is clearly predicted to increase with decreased gene flow.
Therefore, I believe it is more likely that the patterns I observe reflect differential rates of
divergence and speciation, rather than an effect of extinction.
61
Chapter 2. Measuring the spatial scale of speciation
The effects of taxonomic practice and surveying effort
In common with most comparative studies of diversification, I assume that entities
named as species represent a similar level of evolutionary divergence across all taxa
considered. If different taxa had been subjected to different taxonomic practises, this
could influence my conclusions regarding scales at which speciation can occur. For
instance, a taxon in which species are split more finely (so that they are equivalent to
subspecies of other taxa) would be counted as being able to speciate within smaller
islands. On the other hand, finer splitting, causing subspecies endemic to single islands to
be elevated to species status, could lead to more cases of genera with only one endemic
species on an island, and thus lower calculated probabilities of speciation. As this would
not affect which genera are identified as having had speciation events, this would not
affect the estimation of minimum areas for speciation, but would change the slopes of the
speciation-area relationships. In either case, it is unlikely that the differences in
taxonomic practice among my study taxa are in the correct order (for instance,
Lepidoptera lumped more than snails) to be solely responsible for the pattern of minimum
areas for speciation observed.
Data quality is likely to vary among islands and taxa as a result of differences in
past surveying intensity. Total surveying effort has generally been greater for larger
islands, but on small islands less effort is necessary for complete description of their
endemic species. Therefore, I do not believe that the chance of detecting whether a genus
has speciated in situ or not is likely to vary systematically with island area.
62
Chapter 2. Measuring the spatial scale of speciation
Finally, my surveys of island characteristics, species lists and phylogenies of
study genera are not comprehensive, due to limitations on data availability. Future
availability of appropriate data could perhaps alter observed patterns. However, I believe
that the data used include a high percentage of those available and are complete enough to
draw broad conclusions.
Implications for evolutionary studies of diversity patterns
These results support a general geographical model of speciation in which area
and gene flow interact via the spatial scale of speciation to control both speciation rates
and resulting diversity patterns. As a result of reduced gene flow, some organisms are
able to differentiate at finer spatial scales than others, leading to increased speciation rates
and higher taxonomic diversity within a given area. Variation among taxa in the level of
gene flow could be caused by several factors, including differences in dispersal ability, in
the degree of habitat specificity (which controls which habitats will act as barriers to
dispersal, Thorpe 1945) and in the strength of natural selection against betweenpopulation hybrids (whose survival is necessary for effective gene flow). The strength of
selection against hybrids will depend on the rate of accumulation of genetic
incompatibilities and the degree of local adaptation (Gavrilets 2004; Fuller 2008), both of
which could vary systematically among taxa. Because the above model incorporates both
species traits and environmental characteristics, it should be useful for explaining both
taxonomic and regional variation in diversification rates and total diversity.
Furthermore, the strength of this model highlights more generally the potential of
an evolutionary-process based framework for understanding speciation rates and higher-
63
Chapter 2. Measuring the spatial scale of speciation
level patterns of species richness. Macroevolutionary studies until now have tested a
diverse range of potential correlates of diversification, with mixed results and few general
conclusions (for a review of factors tested, see Jablonski 2008). In particular,
macroevolutionary studies focusing on organism traits - such as animal body size or plant
woodiness - have generally found only weak correlations with diversification rates,
explaining no more than 10-24% of the observed variation in clade species richness, even
using multivariate models (Phillimore et al. 2006). In contrast, there is stronger evidence
for the link between population-level processes (including adaptive divergence, but also
sexual selection and gene flow) and rates of speciation and diversification (for example,
Barraclough et al. 1995; Belliure et al. 2000; Stuart-Fox and Owens 2003; Funk et al.
2006; Seddon et al. 2008). These processes relate directly to the population genetic theory
that forms the foundation of our understanding of speciation, and a framework based on
these processes would be applicable to all organisms. Bridging the gap between
population genetic theories of speciation and macroevolutionary approaches has great
potential for improving our understanding of large-scale patterns of diversity.
64
Chapter 3. Gene flow and diversification
Chapter 3. The relationship between gene flow and clade
diversification rates in Costa Rican orchids
Introduction
Gene flow is thought to be the most important factor preventing population
divergence and speciation (Mayr 1963; Slatkin 1973; Endler 1977; Slatkin 1985; Nosil
2009). As such, reduced gene flow between populations within a species should make
that species more likely to diverge and speciate. Furthermore, if levels of gene flow
within species are heritable along related lineages, the level of gene flow could be a
species-level trait influencing rates of diversification (Jablonski 2008).
Despite clear theory to suggest a relationship, the association between gene flow
within species and clade diversification has never been tested directly using molecular
data (but see chapter two). Some studies have tested the link between dispersal ability and
diversification rates, and dispersal ability is generally a good proxy for rates of gene flow
(Zera 1981; Govindaraju 1988; Bohonak 1999). Most of these studies have found greater
diversification associated with poorer dispersal, which should indicate reduced gene flow,
as expected (for example, Jablonski 1986; Belliure et al. 2000). However, in some cases
greater diversification has been found to be associated with greater dispersability (which
likely reflects the ability of species to colonise new regions, rather than the level of gene
flow between established populations; e.g. Owens et al. 1999; Phillimore et al. 2006),
associated with intermediate dispersal (e.g. Price and Wagner 2004; Paulay and Meyer
2006) or not associated with dispersal at all (e.g. Vrba 1984; Herrera 1989). The diversity
65
Chapter 3. Gene flow and diversification
of results obtained suggests that the relationship between dispersal and diversification is
complex. However, existing studies rely on surrogates for measuring gene flow:
quantifying the level of gene flow directly using genetic data might remove some of this
complexity and help to uncover the true relationship between gene flow and
diversification.
Here, I test whether levels of gene flow within species are heritable and whether
they affect clade diversification rates using population genetic data for tropical orchid
species from pairs of sister clades that differ greatly in species richness. Sister group
comparisons control for age in comparing diversification and allow replicated,
phylogenetically independent tests of the hypothesised relationship. In addition, sister
groups typically share most traits, minimising the number of variables that can confound
conclusions about traits of interest (Barraclough et al. 1998).
Orchids are a good study group because their many species and great ecological
and morphological variability allow many independent tests of factors hypothesised to
affect diversity patterns. They abound in pairs of sister clades that differ greatly in
diversity - for example, the most dramatic sister pairs in this dataset compare clades with
42 and 140 species to sister clades with 1,066 and 1,387 species, respectively. They also
present an unsolved biodiversity mystery: with over 26,000 described species (Dressler
2005; Govaerts et al. 2010), orchids are probably the most diverse family of angiosperms.
However, even though their exotic flowers and varied ecology have attracted a devoted
research community stretching back to Darwin (1862) and earlier, the reasons for their
diversity are still highly debated (Gravendeel et al. 2004; Cozzolino and Widmer 2005;
Tremblay et al. 2005). In addition, 70% of orchids (Benzing 1987) are epiphytes (growing
66
Chapter 3. Gene flow and diversification
on other plants for support), a life form that is poorly studied even though epiphytes make
up about 10% of all vascular plant species and are a major component of tropical forest
communities (Gentry and Dodson 1987). Finally, orchids deserve study as they are a high
conservation priority (all orchids are listed in Appendix I or II of the Convention on
International Trade of Endangered Species; http://www.cites.org) and knowledge of their
genetic characteristics and population dynamics is lacking, especially for tropical species.
I measure the level of gene flow for 17 orchid species from five comparisons of
species-rich and species-poor sister clades using Fst values estimated from AFLP
genotypes. Fst, the proportion of total neutral genetic variation between (rather than
within) populations, is the most widely used measure of gene flow. High values of Fst
indicate clear population differentiation and low levels of gene flow, and low values
indicate little differentiation and high levels of gene flow (Wright 1931; Slatkin 1985). Fst
is robust to variation in the spatial arrangement of populations (Beaumont and Nichols
1996) and, when calculated from truly neutral loci, comparable between species (Barton
2001). When gene flow is at equilibrium, Fst should increase with the geographic distance
between populations, as migration is more frequent between nearby populations
(Hutchison and Templeton 1999). This makes the scale of sampling an important factor in
population genetic studies. For this reason, I evaluate patterns of genetic isolation by
distance in addition to analysing overall Fst.
I also account for two possibly confounding factors, species range size and
ecology. The relationship between species range size and diversification rates is unclear,
as range size may directly influence speciation and extinction rates or it may be the
product of lineages’ evolutionary histories, but an association is expected (Rosenzweig
67
Chapter 3. Gene flow and diversification
1995; Webb and Gaston 2003; Pigot et al. in press). Similarly, it is unclear to what extent
ecological traits of species generally affect diversification rates, but some association is
expected (e.g. Phillimore et al. 2006), as a species’ niche and niche breadth should affect
its probability of range expansion, speciation, and extinction (Funk et al. 2002; McPeek
2008). Furthermore, there is evidence that both range size and species ecology are
associated with variation in levels of gene flow or population differentiation (Hamrick
and Godt 1996; Morjan and Rieseberg 2004). For this reason, I include both a restricted
and a widespread species for as many study clades as possible and I explicitly test the
strength of associations between species ecology and species range size with gene flow
and rates of diversification.
Therefore, to test the predictions of theory linking gene flow and speciation, I test
the hypothesis that a direct association exists between levels of gene flow within species
and diversification rates of clades, taking into account the possible confounding
influences of species range size and ecology. I also test whether the level of gene flow is
heritable between clades.
Methods
Study group selection
Study clades were chosen from two subtribes from tribe Epidendreae (subfamily
Epidendroideae) with recently published phylogenetic analyses with complete sampling
at the genus level: Pleurothallidinae (Pridgeon et al. 2001) and Laeliinae (van den Berg et
al. 2009). Sister clades were chosen from well-resolved portions of the phylogenetic trees,
68
Chapter 3. Gene flow and diversification
although relationships between genera within clades were not always completely
resolved. The only major uncertainty in the composition of chosen clades was whether the
genus Meiracyllium belongs in the Brassavola clade or was placed there spuriously (van
den Berg et al. 2009). I assumed that it belongs in the Brassavola clade; Meiracyllium
contains only two species and so should not affect results greatly either way. Sister clades
were chosen that had as large differences as possible in species richness between the two
clades, using species diversities for genera taken from the World Checklist of
Orchidaceae (Govaerts et al. 2008). All pairs chosen differed in species richness by at
least five-fold. Three pleurothallid sister clade pairs were chosen, hereafter referred to as
the Masdevallia-Trisetella, Lepanthes-Lepanthopsis and Platystele-Dryadella sister pairs;
two laeliinid sister clade pairs were chosen, referred to as the Scaphyglottis-Jacquiniella
and Epidendrum-Brassavola sister pairs. Clade species richness and the genera included
in each clade are listed in Table 3.1.
Study species were chosen from one genus from each study clade. Only epiphytic
species were used. Species within genera were chosen to maximise the ease of locating
and identifying them in the field. Species were selected only if they are relatively wellrepresented in herbaria and their delimitation from closely related species is clear in the
taxonomic literature. In addition to these selection criteria, two more were used. First,
species within clade pairs were chosen that had as similar habitat requirements as possible
to control for the effect of habitat on Fst (for example, high elevation habitats are likely to
be more fragmented and less continuous than lower elevation habitats, which could
directly influence the amount of gene flow between populations). Species habitats were
defined according to descriptions given in the Manual de Plantas de Costa Rica, Vol. III
69
Chapter 3. Gene flow and diversification
(MPCR; Hammel et al. 2003). Second, where possible, the species within each genus
were chosen to include both a restricted-range and a widespread species to allow
investigation of the relationship between species range size and Fst. In this case,
“restricted” was defined as occurring only in Costa Rica or Costa Rica and one other
neighbouring country; “widespread” was defined as occurring in two or more countries in
addition to Costa Rica. Species ranges were defined according to the World Checklist of
Orchidaceae (Govaerts et al. 2007).
Although the aim was to choose two species per clade, sampling for some clades
covered fewer species than others. For most of the species-poor clades, sampling multiple
species was problematic because these genera had few or no easily collected species. For
example, Brassavola is represented by only three species in Costa Rica - one of which is
rare and poorly collected and the other two of which are suspected of being synonymous.
In this case, the rare species was not collected. For Dryadella, Lepanthopsis and
Trisetella, all species were difficult to find and I was only able to collect sufficient
numbers of individuals for one species of each genus. In contrast, more than two suitable
potential study species were identified for most of the other genera, and all were collected
when found in the field. Finding populations in the field was a haphazard process and by
collecting samples this way the chance of obtaining sufficient sample individuals for at
least two species per genus was maximised. However, time limitations for the genetic
analysis meant that not all species collected could be genotyped, so no more than one
restricted and one widespread species were genotyped per genus. Some collected samples
proved difficult to identify definitively to species because their species are
morphologically similar to others or hybridise - these were all excluded. In cases where
70
Chapter 3. Gene flow and diversification
multiple restricted or widespread species were collected from a single genus, the species
with the best sampling (the most samples and covering the widest range of spatial scales)
was selected for genotyping. The species collected and genotyped are listed in Table 3.2
along with their geographic and elevation ranges and habitat, and the number of
populations sampled and individuals genotyped. Photos of genotyped species are given in
Figure 3.1.
71
Chapter 3. Gene flow and diversification
Table 3.1. Study clade pairs, with genera they include and currently accepted species
richness according to the World Checklist of Orchidaceae (Govaerts et al. 2010).
Species-rich clade
Masdevallia clade
Masdevallia Ruiz & Pav.
# spp.
751
582
Diodonopsis Pridgeon & M.
W. Chase
Dracula Luer
126
Porroglossum Schltr.
38
Species-poor sister clade
# spp.
Trisetella clade
Trisetella Luer
23
23
5
Lepanthes clade
Lepanthes Sw.
1066
1066
Lepanthopsis clade
Lepanthopsis (Cogn.) Ames
42
42
Platystele clade
Platystele Schltr.
276
99
Dryadella clade
Dryadella Luer
53
53
Scaphosepalum Pfitzer in H.
G. A. Engler & K. A. E.
Prantl (eds.)
Specklinia Lindl.
46
131
Scaphyglottis clade
Scaphyglottis Poepp. & Endl.
77
68
Jacquiniella clade
Jacquiniella Schltr.
13
12
Dimerandra Schltr.
9
Acrorchis Dressler
1
Epidendrum clade
Epidendrum L.
1387
1325
Barkeria Knowles & Westc.
15
Caularthron Raf.
4
Laelia Lindl.
Brassavola clade
Brassavola R. Br. in W. T.
Aiton
Cattleya Lindl.
140
21
111
4
24
Guarianthe Dressler & W. E.
Higgins
Meiracyllium Rchb. f.
Myrmecophila Rolfe
10
Rhyncholaelia Schltr.
2
Orleanesia Barb. Rodr.
9
2
72
73
widespread
turialvae Rchb. f.
widespread
restricted
wendlandii Rchb. f.
floripecten (Rchb. f.) Ames
widespread
elata Rchb. f.
Lepanthopsis
restricted
ciliisepala Schltr.
picturata Rchb. f.
Lepanthes
widespread
chontalensis Rchb. f.
widespread
widespread
rafaeliana Luer
triglochin (Rchb. f.) Luer
restricted
nidifica Rchb. f.
Masdevallia
Trisetella
widespread
species
Genus
restricted or
widespread
S.E. Mexico to S. tropical
America
Costa Rica to Panama, Brazil
Costa Rica and W. Panama
Costa Rica to W. Colombia
Costa Rica and possibly
Venezuela
Costa Rica to S. tropical
America
Costa Rica to S. tropical
America
throughout central America
Costa Rica and Panama
Nicaragua to N. Peru
Geographic range
Rain forest
Very humid, rain, cloud
or oak forest
Very humid, rain, cloud
or oak forest
Cloud or oak forest
Cloud or oak forest
Very humid, rain or
cloud forest
Rain or cloud forest
Very humid, rain or
cloud forest
Cloud or oak forest
Very humid, rain or
cloud forest
typical habitat
1900-2000
600-2550
1800-3000
1500-2600
1400-2050
200-1900
1200-2300
600-1800
2600-3000
700-2000
Elevation
range (m)
5
3
3
3
2
2
4
6
2
5
#L
Table 3.2. Species collected, with distributions and habitats. Species with names in bold were genotyped. Species are arranged by genus,
with genera from sister clades arranged together. #L = number of locations sampled. #S = number of samples genotyped.
20
-
-
52
36
70
-
-
30
34
#S
74
Jacquiniella
Scaphyglottis
widespread
widespread
teretifolia (Sw.) Britton & P. Wilson
globosa (Jacq.) Schltr.
widespread
prolifera (R. Br.) Cogn.
restricted
widespread
fusiformis (Griseb.) R. E. Schult.
aporophylla (L. O. Williams)
Dressler
restricted
widespread
guatemalensis (Schltr.) Luer
jimenezii Schltr.
widespread
odontostele Luer
restricted
microtatantha (Schltr.) Garay
Dryadella
Mexico to S. tropical America
widespread
stenostachya (Rchb. f.) Garay
Mexico to northern S. America
Mexico to northern S. America
Costa Rica and Panama
Mexico to northern S. America
Costa Rica to S. tropical
America
Costa Rica and W. Panama
Mexico to Colombia
Costa Rica, Panama, Columbia
Costa Rica
Costa Rica
restricted
propinqua (Ames) Garay
Geographic range
Platystele
restricted or
widespread
species
Genus
Very humid or rain
forest
Humid, very humid,
rain or cloud forest
Rain forest
Humid, very humid,
rain or cloud forest
Very humid or rain
forest
Very humid, rain or
cloud forest
Very humid or rain
forest
Very humid forest
Cloud or oak forest
Very humid or rain
forest
Cloud or oak forest
typical habitat
0-1400
1100-1850
800-1500
0-1500
50-1400
700-2400
1200-2000
50-150
1500-2200
0-1900
1400-1900
Elevation
range (m)
4
5
4
5
4
3
3
2
3
3
#L
-
46
37
-
42
47
-
12
-
50
24
#S
75
widespread
radicans Pav. ex Lindl.
widespread
restricted
vulgoamparoanum Hágsater & L.
Sánchez
nodosa (L.) Lindl.
widespread
laucheanum Bonhof ex Rolfe
Brassavola
restricted
exasperatum Rchb. f.
Epidendrum
restricted or
widespread
species
Genus
Mexico to S. tropical America
Caribbean and Mexico to
Colombia
Costa Rica and Panama
Mexico to Colombia
Costa Rica and Panama
Geographic range
Dry, humid, very humid
or scrub forest; rocks or
mangroves
Very humid, rain or
cloud forest
Very humid or rain
forest
Very humid, rain, cloud
or oak forest; pastures
and slopes
Very humid or cloud
forest
typical habitat
0-100
850-1900
0-350
1300-2100
900-2500
Elevation
range (m)
8
3
4
5
3
#L
50
-
38
31
27
#S
Chapter 3. Gene flow and diversification
Masdevallia nidifica
Masdevallia rafaeliana
Trisetella triglochin
Lepanthes ciliisepala
Lepanthes elata
Lepanthopsis floripecten
Platystele propinqua
Platystele stenostachya
Dryadella odontostele
76
Chapter 3. Gene flow and diversification
Scaphyglottis jimenezii
Scaphyglottis fusiformis
Jacquiniella aporophylla
Jacquiniella teretifolia
Epidendrum exasperatum
Epidendrum laucheanum
Figure 3.1. Photos of study
species. Photos of
Masdevallia rafaeliana and
Epidendrum exasperatum
thanks to Martin Turjak.
Photo of Platystele
stenostachya thanks to
Ernesto Carman.
Epidendrum vulgoamparoanum
Brassavola nodosa
77
Chapter 3. Gene flow and diversification
Study species phylogeny reconstruction
The phylogeny of study species was reconstructed using matK sequences from
GenBank. Sequences for study species were used when available; otherwise a sequence
for a single species from the study genus was chosen randomly from those available
(GenBank accession numbers given in Table 3.3). Thunia alba was included as an
outgroup to allow rooting of the tree. Sequences were aligned with MAFFT (available at
http://mafft.cbrc.jp/alignment/software; Katoh et al. 2002), using the online interface and
the L-NS-i algorithm. The aligned sequences were trimmed to include only sites for
which data were present for all species, and a tree was constructed using Maximum
Likelihood and a GTR substitution model with parameters estimated from the data, using
GARLI v. 1.0 (available from http://garli.googlecode.com; Zwickl 2006) using default
parameters and three independent runs. All three runs recovered the same topology;
branch lengths from the highest likelihood tree were used. Finally, the tree was edited by
inserting species for which no sequences had been available. These species were added as
polytomies within their genus, with terminal branches of length 0.001. All relationships in
the study species analyses match the relationships in the source trees used to select study
clades (Pridgeon et al. 2001; van den Berg et al. 2009).
78
Chapter 3. Gene flow and diversification
Table 3.3. Accession numbers for matK sequences used to build study species trees.
Species
Brassavola nodosa
Dryadella edwallii
Epidendrum campestre
Jacquiniella aporophylla
Jacquiniella teretifolia
Lepanthes ciliisepala
Lepanthes elata
Lepanthopsis astrophora
Masdevallia bicolor
Platystele stenostachya
Scaphyglottis fusiformis
Scaphyglottis jimenezii
Trisetella triglochin
Thunia alba (outgroup)
GenBank accession number
AF263820
AF265454
AF263781
EU214360
AY396087
EU214373
EU214374
AF265487
AF265447
EF079326
EU214455
EU214460
EF065592
AF302706
Sample collection
All samples were collected in Costa Rica with the kind help of the Lankester
Botanical Garden (University of Costa Rica). Sample collection occurred in two field
seasons, April/May 2008 and March-May 2009. Whenever possible, living specimens
from each population sampled were deposited at the Lankester Botanical Garden, to be
maintained in their living collection and added after flowering to their herbarium and
silica-gel preserved collections. This was preferred to making herbarium sheets at the
time of collection because most plants were not flowering when sampled.
The sampling goal for each species was to collect samples from 20 individuals
from each of 3-7 populations throughout Costa Rica. Different locations were sampled in
each field season, and no location was sampled twice for the same species. Populations
79
Chapter 3. Gene flow and diversification
were selected, as much as possible, to represent a range of distances between populations
and a significant portion of the range of each species. At each sampling location,
coordinates were recorded using a Garmin 60CSX GPS device. From March 31 to April
22, 2009, I had no working GPS device, and instead recorded locations using mileage
along roads. These relative locations were later translated into coordinates using Google
EarthTM v. 5.2 (Google). Maps of sampling locations for each genotyped species are given
in Figures 3.2-3.4 and details of sampling locations for each species are given in
Appendix Table II.1.
Plant samples were put into labelled plastic bags in the field and then kept in a
refrigerator or ice chest until they could be put into silica for preservation (Chase and
Hills 1991). Samples were cut into pieces to break the cuticle, put into individual labelled
envelopes made of coffee-filter paper and then dried in larger sealed bags containing
silica gel. The silica gel was changed for fresh, dry silica gel multiple times until all
samples were completely dry. Once dry, samples were stored at room temperature with a
smaller amount of fresh silica.
80
Chapter 3. Gene flow and diversification
a)
b)
Figure 3.2. Sampling locations for study species from (a) the Masdevallia-Trisetella and
(b) the Lepanthes-Lepanthopsis clade pairs.
81
Chapter 3. Gene flow and diversification
a)
b)
Figure 3.3. Sampling locations for study species from (a) the Platystele-Dryadella and (b)
the Epidendrum-Brassavola clade pairs.
82
Chapter 3. Gene flow and diversification
a)
b)
Figure 3.4. Sampling locations for study species from the Scaphyglottis-Jacquiniella
clade pair. (b) is a detail of the boxed area in (a).
83
Chapter 3. Gene flow and diversification
AFLP genotyping
DNA from all genotyped samples was extracted using the Qiagen DNeasy Plant
Kit, following the manufacturer’s protocol. DNA extractions for about a third of samples
were carried out using the Plant Mini Kit, with individual extraction tubes; the rest were
carried out using the 96 Plant Kit, with 96-well plates. Except when there was too little
sample material available (for small species), approximately 20 mg silica-dried material
was used for each extraction. Flowers were used in preference to leaves when available.
Some leaves had algae, moss or other contaminants and were scraped clean or wiped with
ethanol before being used. Extractions were eluted twice with either 50 or 100 μl AE
elution buffer or eluted with 50 and then 25 μl AE elution buffer, depending on amount of
sample material available and ease of extracting DNA from each species. Each round of
extraction included one or two blanks to check for contamination between tubes/wells. In
addition, 10-20% of individuals from each population of each species were repeated, by
extracting and genotyping them twice independently, to allow later quantification of
genotyping error rates (Bonin et al. 2004).
Trials of potential selective primer combinations were carried out for each species
individually, testing genotyping quality of six or twelve primer combinations. Primers
were chosen to maximise number of peaks per sample, number of polymorphic peaks per
species, evenness of spread of peak sizes (even if this required choosing a primer that
produced fewer peaks) and profile repeatability.
84
Chapter 3. Gene flow and diversification
AFLP reactions were carried out following the methods of Vos et al. (1995)
except for the following modifications.
Only 300 ng DNA was used for species that did not successfully amplify using the
original protocol (Epidendrum laucheanum, Trisetella triglochin, Lepanthes elata,
Lepanthes ciliisepala and Platystele propinqua). DNA was dried in a vacuum oven at 4060°C and, once dry, resuspended in 5.5 μl distilled water.
Enzyme restriction used EcoRI and MseI (Promega and New England BioLabs,
respectively); ligation used T4 Ligase (Promega). DNA restriction and subsequent
adaptor ligation were carried out simultaneously for most species by incubating each
sample at 37°C for 2 hours with 5 U EcoRI, 1 U MseI, 1.1 μl 0.5M NaCl, 0.55 μg bovine
serum albumen (BSA), 1.1 μl 10X Ligase buffer (Promega), 1 U Ligase, 1 μl each of
MseI and EcoRI adaptor pairs (Applied Biosystems) and distilled water to make up the
volume to 5.5 μl (total reaction volume including sample = 11 μl). For some species,
however, the ligation step worked only if it was carried out separately. In this case,
samples were first incubated as above but in a reaction mixture with water replacing the
ligase and adaptor pairs. Ligation was then carried out by incubating 4 μl of the diluted
restriction product (11 units product diluted with 80 units TE buffer) for 2-3 hours at
room temperature with 1 U Ligase, 1 μl each of MseI and EcoRI adaptor pairs, 1.1 μl
Ligase buffer, 0.55 μg BSA and water to make up the volume to 4 μl (total reaction
volume including sample = 8 μl). For species where restriction and ligation were carried
out together, the reaction product was diluted before the preamplification PCR (11 units
of product diluted with 189 units TE buffer).
85
Chapter 3. Gene flow and diversification
All PCR primers came from the Applied Biosystems AFLP Regular Genome
Plant Mapping Kit and all PCR reactions used a Fermentas PCR mastermix. For the
preamplification PCR, 2 μl restriction-ligation product were mixed with 7.5 μl PCR
mastermix and 0.5 μl preamplification primers and amplified with the following program:
72°C for 2 min; cycles of 20 s at 94°C, 30 s at 56°C and 2.5 min at 72°C; 30 min at 60°C.
For some species, the 72°C step was shortened to 2 minutes, and the number of cycles
used varied between species. The preamplification PCR product was then diluted (5 units
product with 95 units TE buffer) for some species, but for most species the final peak
profile was stronger if the preamplification PCR product was used undiluted. For the
selective PCR, 1.5 μl preamplification product were mixed with 7.5 μl PCR mastermix
and 0.5 μl of each of two selective primers and amplified with the following program:
72°C for 2 min; 10 cycles of 20 s at 94°C, 30 s at 66°C and 2 min at 72°C, with a
decrease in annealing temperature of 1°C per cycle; 35 cycles of 20 s at 94°C, 30 s at
56°C and 2 min at 72°C; 30 min at 60°C. Three selective PCRs were carried out for each
sample: with NED (Yellow), JOE (Green) and FAM (Blue) labelled primer combinations.
Finally, 1.2 μl of each of the three selective PCR products for each sample were
combined with 10 μl formamide and 0.2 μl GeneScan-500 ROX size standard (Applied
Biosystems) in a single well for simultaneous genotyping on a capillary sequencer
(Applied Biosystems 3130xl Genetic Analyzer).
Details of the protocol variations used for each species, including the selective
primers used, are listed in Table 3.4. If possible, all samples for each species were run
together on a single plate for all AFLP reactions to eliminate variability from slight
variations in run temperatures etc. When this was not possible, samples were split
86
Chapter 3. Gene flow and diversification
between two plates, each containing at least one pair of repeated samples. When samples
of one species were split between plates, both plates were run on the same PCR machine
for all reactions.
87
88
ACT-CTG
ACT-CAA
ACA-CAT
ACT-CTG
ACT-CTG
ACT-CTG
ACT-CTA
ACA-CTT
ACA-CTA
ACA-CAT
ACA-CTT
ACT-CAA
ACT-CAA
ACT-CTG
T. triglochin
L. ciliisepala
L. elata
L. floripecten
P. propinqua
P. stenostachya
D. odontostele
S. jimenezii
S. fusiformis
J. aporophylla
J. teretifolia
E. exasperatum
E. laucheanum
E. vulgoamparoanum
B. nodosa
AGC-CAT
AGC-CTA
AGC-CTA
AGC-CTA
ACC-CTA
AGC-CAT
AGC-CAT
AAC-CAC
ACC-CAG
AAC-CAC
AGC-CTA
AAC-CAG
AAC-CTA
AAC-CTA
AAG-CTC
AAG-CTC
ACG-CAG
AAG-CTC
AAG-CTC
AAG-CTC
AAG-CTC
AAG-CTG
ACG-CAC
ACG-CAC
AAG-CTC
ACG-CAG
ACG-CTG
ACG-CTG
AAG-CAA
ACG-CAG
ACA-CAT
ACA-CAT
M. nidifica
M. rafaeliana
ACC-CTT
AGC-CAT
FAM (Blue) NED (Yellow) JOE (Green)
primers used primers used primers used
Species
extension time was used instead of the standard 2.5 min.
one reaction
one reaction
separate reactions
one reaction
one reaction
one reaction
one reaction
one reaction
separate reactions
separate reactions
separate reactions
separate reactions
separate reactions
separate reactions
separate reactions
one reaction
one reaction
Restriction/ligation
yes
no
no
no
yes
no
yes
no
no
no
no
no
no
no
no
yes
yes
25*
25
30
25*
25
25*
25
25*
30
30
30
30
30
30
30
25
25
Preamp. PCR
# preamp.
product diluted? PCR cycles
50
27
31
38
37
46
47
42
12
24
50
20
36
52
70
34
30
# samples
genotyped
10
5
9
10
4
5
6
7
6
4
9
7
7
11
15
4
2
# samples
repeated
Table 3.4. Details of AFLP method used for each study species. A star by the number of preamplification PCR cycles indicates that a 2 min
Chapter 3. Gene flow and diversification
AFLP scoring
AFLP scoring was carried out using GeneMapper v. 4.0 (Applied Biosystems) to
manually identify bins and AFLPScore v. 1.4a (available at
http://www.sheffield.ac.uk/molecol/software~/aflpscore.html; Whitlock et al. 2008) to
optimise scoring parameters and create a binary genotype table for each species.
First, all profiles with poor sizing or evidence of poor PCR amplification (few
peaks or peak strength decreasing rapidly with fragment size) were excluded from
analysis. Then, as many bins as possible were created in the 50-500 base pair (bp) range
for each species dataset. Bins had to be less than 1 bp wide and non-overlapping. In
addition, bins had to include at least one peak of height 100 relative frequency units
(RFU) or greater; peaks within a single bin had to be 0.3 bp apart or less and those from
different bins had to be at least 0.4 bp apart. These criteria were set to minimise the
chance of homoplasy due to including peaks from more than one locus in a single bin,
based on the fact that homologous peaks from repeated samples never differed in position
by more than 0.3 bp, except in a few cases involving unusually wide peaks. These criteria
are strict, but I preferred to exclude a few valid loci than to score non-homologous peaks
as bands from the same locus.
Once the maximal bin set was created, scoring parameters were chosen using a
version of AFLPScore that I modified, which removes loci with no peaks in any sample
after scoring and before calculating error rates. A range of locus and phenotype selection
thresholds were tested for each of the four scoring methods available in AFLPScore
89
Chapter 3. Gene flow and diversification
(filtered loci/absolute thresholds, unfiltered/absolute, filtered/relative and
unfiltered/relative) and thresholds were chosen that resulted in the most loci being
retained with an error rate of 5% or less. These optimised thresholds were then used to
generate a binary allele table using AFLPScore. For some species and primer
combinations, no scoring parameters resulted in error rates less than 5%. In these cases,
the scoring parameters giving the lowest error rate were chosen, as long as this error rate
was less than 10%. If an error rate lower than 10% could not be obtained for a particular
species/primer dataset, that dataset was excluded from further analysis. After scoring, all
loci were removed that had a band presence or absence in only one individual, as these
are likely to represent errors in the genotyping process. Details of scoring parameters used
and error rates for each species and primer combination are given in Appendix Table II.2.
Finalising AFLP datasets
For all species, the populations used as units for analysis were defined by
distance: all samples collected within 1.5 km of one another were treated as a single
population.
Before carrying out any analyses, the AFLP datasets for each species were
checked for outlier loci potentially under the influence of selection. Outlier loci are those
that have much higher or lower Fst values than expected from the overall distribution of
locus-specific Fst values, taking into account the heterozygosity of each locus. Loci with
higher Fst than expected are potential evidence of divergent selection, whereas loci with
lower Fst than expected are potential evidence of balancing selection (Beaumont and
Nichols 1996; Beaumont and Balding 2004). Outlier loci were identified using the
90
Chapter 3. Gene flow and diversification
software DfDist (available from http://www.rubic.rdg.ac.uk/~mab/stuff/; modified to
allow dominant data from Beaumont and Balding 2004), which calculates p-values
representing the likelihood of selection influencing each locus after simulating Fstheterozygosity distributions. Loci with p-values less than 0.005 for either balancing or
divergent selection were excluded from further analyses in order to calculate Fst without
bias from loci under selection. Over all species, only 26 loci were excluded.
Some final AFLP datasets included populations with only one or a few individuals
due to difficulties with collecting in the field or with AFLP genotyping. To deal with this
problem, all analyses were carried out both on the full dataset including all populations
and on a reduced dataset that excluded populations with less than three individuals. In the
case of Lepanthopsis floripecten, which had only one population with three or more
samples, the reduced dataset only excluded populations with one individual.
Analysing Fst patterns
Except where stated otherwise, all analyses were carried out using R v. 2.8.1 (R
development core team 2008).
Fst values were estimated for each species using Arlequin v. 3.5 (Excoffier and
Lischer 2010), as Φst values (Excoffier et al. 1992), for which the between- and withingroup variances are calculated using analysis of molecular variance (AMOVA) of genetic
distances between sample haplotypes. The significance of each Fst value was tested
through permutation of the original haplotype table. Negative values of Fst were replaced
with 0 for further analyses (Long 1986). It was also noted for each species whether any
91
Chapter 3. Gene flow and diversification
pair-wise Fst values between populations were greater than 0.2, which corresponds to a
migration rate (Nm) of 1 individual per generation using the formula Fst ≈ 1/(4Nm+1) and
indicates gene flow reduced enough to allow neutral divergence of populations (Wright
1931; Slatkin 1985). Because the geographic arrangement of population samples varied
between species, the relationship between Fst and geographic distance was investigated.
Matrices of pair-wise Fst values between all populations for each species were generated
in Arlequin based on pair-wise differences between haplotypes. Matrices of pair-wise
geographic distances were generated with the AFLPdat package in R (Ehrich 2006) based
on the latitude and longitude for each population. For each species, the significance of the
relationship between Fst and geographic distance was tested using a Mantel test with pvalue calculated from 1000 simulated permutations of the original matrices, using the
ade4 package in R (Dray et al. 2007). In addition, the average pair-wise Fst between
populations separated by 50 km or less (Fst<50) was calculated for each species to give a
measure of Fst calculated at a constant geographical scale.
Heritability was estimated for overall Fst by calculating the phylogenetic signal
using the lambda measure (Pagel 1999). Lambda varies from 0 to 1, where a value of 0
means a trait evolves independently of the phylogenetic tree (is not heritable along
lineages), and a value of 1 means trait values are entirely determined by the tree (are
completely heritable). The maximum likelihood value of lambda was calculated using the
CAIC package in R (Orme et al. 2008) and the matK species tree described earlier.
Likelihood ratio tests were used to test whether lambda was significantly different from 0
or 1, by computing the likelihood ratio between a model optimising lambda and a model
92
Chapter 3. Gene flow and diversification
fixing lambda to 0 or 1, respectively, and computing the p-value using a Chi squared
distribution (Freckleton et al. 2002).
The effect of gene flow on diversification was tested using a two-tailed Wilcoxon
signed rank test comparing mean overall Fst values between sister clades.
Testing the influences of species range size and ecology
Measures of Fst were tested for associations with species range size, using two
measures of range size. First, overall Fst values and whether any pair-wise Fst was greater
than 0.2 were compared between restricted and widespread species using two-tailed
Wilcoxon rank sum tests. “Restricted” and “widespread” were defined as described in the
introduction. Second, overall Fst was regressed against species range size quantified as the
number of TDWG Level 2 Regions (http://www.tdwg.org/standards/109/) for which
species occurrence was noted in the World Checklist of Orchids (Govaerts et al. 2007).
The effect of the interaction between range size and Fst on diversification rate was
tested using a two-tailed Wilcoxon signed rank test comparing mean Fst values of
widespread species between sister clades. Fst values from restricted species were not
compared between sister clades because only one restricted species from a species-poor
clade was collected.
A range of ecological variables were tested to see if they could explain variation
between species in Fst values: branch circumference, elevation range and habitat range.
The circumference of the branch on which each sample plant was growing was measured
in the field at the time of collection. Branch circumference was included because it is a
93
Chapter 3. Gene flow and diversification
proxy for many important aspects of epiphyte niche - smaller branches are associated
with shorter lifespans, lower water availability and higher light availability (Chase 1988;
Gravendeel et al. 2004). Species elevation ranges and habitat descriptions were taken
from the species descriptions in the MPCR. The mean and variance of branch
circumference for each species were regressed against overall Fst. The minimum,
maximum and range ( = maximum - minimum) of elevation for each species were also
regressed against overall Fst, as was the number of habitats occupied. Variance in branch
circumference, elevation range and number of habitats were included as proxies of
ecological specificity/niche breadth. All these variables were also compared between
differentiated (any pair-wise Fst > 0.2) and non-differentiated species using two-tailed
Wilcoxon rank sum tests. In both cases, Bonferroni correction was used to correct for
multiple tests. Because each measure of Fst was tested for association with 6 ecological
variables, only p-values smaller than 0.0083 were considered significant. The same
ecological variables were tested for associations with species range size measured as
number of regions using regressions, and with species range as widespread/restricted
using ANOVAs. Elevation and habitat ranges of study species are given in Table 3.2.
Branch circumference mean and variance for each species are given in Appendix Table
II.3.
Mean species range size, elevation range and number of habitats were compared
between sister clades using two-tailed Wilcoxon rank sum tests to investigate whether
differences in range size or ecology could explain differences between clades in
diversification. For these tests, range sizes for all species in all genera in each clade were
taken from the World Checklist of Orchidaceae (Govaerts et al. 2007) as described above.
94
Chapter 3. Gene flow and diversification
Species elevation and habitat ranges were compiled for all species native to Costa Rica
from all genera in each clade from the MPCR. Elevation and habitat data were compiled
only for species native to Costa Rica as this way all data came from a single source and
were comparable. Mean range sizes, elevation ranges and habitat ranges of study clades
are given in Appendix Table II.4.
Results
A total of 1341 plants were sampled from 26 consistently identifiable species in 5
pairs of sister genera; 647 of these samples from 17 species in the 5 sister clade pairs
were successfully genotyped.
Overall Fst is variable between species, ranging from zero, indicating no
differentiation between populations, to a maximum value of 0.358, indicating clear
differentiation between populations. Similarly, Fst between populations less than 50 km
apart (Fst<50) ranges from zero to 0.431, and the two measures are strongly correlated (p
< 0.005 for both datasets). The relationship between Fst and geographic distance is not
significant for any species, and only near significance for one (E. laucheanum, p =
0.0509; other p-values ranging from 0.11 to 1). Plots of Fst against geographic distance
for all species are given in Figures 3.5 and 3.6. Because there is no evidence for a
relationship between Fst and geographic distance, and overall Fst and Fst<50 are strongly
correlated, Fst<50 is not used in further analyses. Fst values for all species are given in
Table 3.5.
95
Chapter 3. Gene flow and diversification
Table 3.5. Fst values for all study species. Significance of overall Fst is indicated by stars *, p < 0.05; **, p < 0.005; ***, p < 0.0005. Datasets with a dash for Fst < 50 had no
populations separated by less than 50 km.
overall
Fst
any pair-wise
Fst > 0.2
Fst < 50
M. rafaeliana
T. triglochin
.36 ***
.35 ***
0
.063 **
yes
yes
no
no
0.43
0
0.063
L. ciliisepala
L. elata
L. floripecten - full dataset
L. floripecten - reduced dataset
0.027
.072 ***
0.14
0
no
no
yes
no
0.027
0.072
0.43
0
P. propinqua - full dataset
P. propinqua - reduced dataset
P. stenostachya
D. odontostele
.17 ***
.12 ***
.24 ***
.18 *
yes
no
yes
yes
0
0.16
0.037
S. jimenezii
S. fusiformis
J. aporophylla- full dataset
J. aporophylla - reduced dataset
J. teretifolia - full dataset
J. teretifolia - reduced dataset
0
.24 ***
0
0
0
0.013
no
yes
no
no
no
no
0.00084
0.14
0
0
0.0012
0.01
E. exasperatum
E. laucheanum
E. vulgoamparoanum
B. nodosa - full dataset
B. nodosa - reduced dataset
.13 **
0.045
0.019
.059 **
.054 **
no
no
no
no
no
0.14
0.038
0.027
0.057
0.053
Species
M. nidifica - full dataset
M. nidifica - reduced dataset
96
Chapter 3. Gene flow and diversification
Figure 3.5. Relationship between Fst and distance for all study species for the full dataset.
Data are presented by sister clade pair. Species from the species-rich clade in each case
are shown in black; species from the species-poor clade are shown in red. Restricted
species are shown with open symbols and widespread species with filled symbols. The
Epidendrum clade includes two restricted species: E. exasperatum is represented by open
circles, whereas E. vulgoamparoanum is represented by open triangles.
97
Chapter 3. Gene flow and diversification
Figure 3.6. Relationship between Fst and distance for all study species for the reduced
dataset. All plotting symbols follow the conventions described for Figure 3.5, but the
vertical scale here is smaller.
98
Chapter 3. Gene flow and diversification
Overall Fst is not heritable over the clades examined here. The maximum
likelihood value of lambda for overall Fst for both datasets is 6.61x10-5 and is not
significantly different from 0 (for both datasets, p value for difference from 0 = 1; p value
for difference from 1 < 0.001). Figure 3.7 shows how Fst maps on to the species tree (data
from the full dataset; the equivalent tree with Fst values from the reduced dataset is given
in Appendix Figure II.1).
There is no support for my hypothesis that gene flow is associated with
diversification rate. There is no significant difference in mean overall Fst between speciesrich and species-poor sister clades (full dataset, p = 0.313, reduced dataset, p = 0.125).
Instead, there is a suggestion of an inverse association between the level of gene
flow and species range size. Overall Fst is higher for widespread species than restricted
species, although the difference is only significant for the full dataset (full dataset, p =
0.0495, reduced dataset, p = 0.0951), and there is no significant relationship between
overall Fst and species range size measured as number of regions (full dataset, p = 0.862,
reduced dataset, p = 0.955). Furthermore, no difference exists between widespread and
restricted species in whether any pair-wise Fst is greater than 0.2 (full dataset, p = 0.158,
reduced dataset, p = 0.0734).
There is also a hint of an interaction between species range size and the level of
gene flow in determining diversification rates. In both datasets, the only species with
overall Fst over 0.2, indicating gene flow reduced enough to allow independent evolution
of populations, are widespread species from large clades (Masdevallia nidifica, Platystele
stenostachya and Scaphyglottis fusiformis). This association can be seen in Figure 3.7,
99
Chapter 3. Gene flow and diversification
which shows both range size and Fst values mapped on to the species tree. However, there
are an equal number of widespread species from large clades with low Fst and there is no
significant difference between the overall Fst of widespread species from small and large
sister clades (full dataset, p = 0.438, reduced dataset, p = 0.125).
Figure 3.7. Associations of Fst and species range size with species phylogeny (data from
the full dataset). Circles at branch tips are sized proportionally to the log of species range
size measured as number of regions and shaded according to Fst. Species names are
abbreviated by the first letter of genus and species.
100
Chapter 3. Gene flow and diversification
Some variation in Fst between species can be explained by species ecology.
Overall Fst shows no significant correlation with circumference of branches on which
sampled plants occurred, species elevation range or number of habitats per species (pvalues ranging from 0.0838 to 0.913). However, for the full dataset, whether any pairwise Fst is greater than 0.2 is significantly associated with both decreased elevation range
(p = 0.0021; Figure 3.8a) and decreased number of habitats (p = 0.0043; Figure 3.8b).
Figure 3.8. Relationships between species elevation range and number of habitats and
whether any pair-wise Fst value is over 0.2, for the full dataset.
There is no association between species range size and ecology - neither can
explain the other’s association with Fst. Among the study species, branch circumference,
elevation range and number of habitats do not differ between widespread and restricted
species (p-values between 0.263 and 0.971), nor do they correlate with range size
measured as number of regions (p-values between 0.264 and 0.842). In the wider dataset
of range size and ecology data compiled from the World Checklist of Orchidaceae and the
101
Chapter 3. Gene flow and diversification
MPCR, range size measured as number of regions is positively associated with both
elevation range (p = 0.038; Figure 3.9a) and with number of habitats occupied (p = 0.029;
Figure 3.9b), but in both cases the correlation is weak (elevation range, r2 = 0.0091;
number of habitats, r2 = 0.01).
Figure 3.9. Associations between species range size and elevation range and number of
habitats over all species native to Costa Rica from the study clades. Shading of circles is
proportional to the number of species with each combination of trait values.
Neither range size nor species ecology shows any association with clade
diversification. Mean range size does not differ between large and small sister clades
(p = 0.1); neither do mean elevation range (p = 0.625) or the mean number of habitats
occupied (p = 0.125).
102
Chapter 3. Gene flow and diversification
Discussion
Findings
There is no support for the hypothesis that the level of gene flow directly controls
rates of diversification. Across the species studied here, the level of gene flow is not
heritable and does not differ between species-poor and species-rich sister clades.
However, there is weak evidence for a relationship between the level of gene flow,
species range size and diversification rates: gene flow low enough to allow high
population differentiation is observed only in widespread species in large clades. Some
variation among species in the level of gene flow can be explained by species ecology,
but this does not contribute to the associations between gene flow and range size or gene
flow and diversification.
The most conservative interpretation of these results is that the level of gene flow
within species does not affect clade diversification rates. Even though gene flow is a
major factor limiting population divergence or speciation, there are a number of situations
in which it would not be expected to affect diversification rates. First, speciation rates
may be limited more by other steps in the speciation process. For these orchids, for
instance, speciation may be most limited by the rate at which species colonise new
regions that are separated by barriers from the ancestral range or by the diversity of
potential pollinators available to drive divergent selection and local adaptation. Second,
variation in diversification rates may be driven mainly by extinction rather than
speciation. This is especially likely if new species tend to have small ranges or population
103
Chapter 3. Gene flow and diversification
sizes, for example as predicted by point mutation (Rosindell et al. 2010) or peripatric
(Mayr 1982) models of speciation.
There is also some indication of a more complex scenario, where the level of gene
flow affects rates of speciation but changes through the lifetime of a species, making it
difficult to detect a link. Unlike organismal traits such as body size, the level of gene flow
within species is an emergent trait at the species level (Jablonski 2008) and likely to be
changed by the speciation process in a manner analogous to species range size. This is
because speciation is likely to divide the parent species range where population
differentiation is highest, so that most speciation events result in daughter species with
decreased average population differentiation as well as decreased range size compared to
the parent. As a result, younger species should tend to have both smaller ranges and less
differentiation/greater gene flow between populations, although the relationship between
species age, range size and gene flow will depend on how quickly ranges expand and
differentiation develops after speciation (Pigot et al. in press). In this case, clades with a
tendency for less gene flow might speciate more rapidly, but new species would always
show reduced population differentiation, weakening any observed relationship between
species gene flow and clade diversity. Two results are consistent with this scenario: the
lack of heritability for the level of gene flow and the finding that well-differentiated
populations only occur in widespread species from large clades, assuming that
widespread species tend to be older than restricted species. However, if this scenario were
the main explanation for variation within clades in level of gene flow, then widespread
species with low Fst would be expected only if they were young species that had recently
104
Chapter 3. Gene flow and diversification
expanded their range. To my knowledge, the relationship between species age and
population differentiation has not been tested but would be worth exploring.
It is interesting that no significant relationship between genetic differentiation and
geographic distance was found for any species, as such isolation by distance relationships
are expected unless dispersal ability is exceptionally high or low (Peterson and Denno
1998) or gene flow is not at equilibrium with genetic drift (Hutchison and Templeton
1999). Orchids are known to have a high propensity for long-distance dispersal as a result
of their abundant, tiny seeds (Arditti and Ghani 2000) and so a lack of isolation by
distance in orchid species should be associated with low population differentiation
(Peterson and Denno 1998) and the species with high values of Fst here are anomalous.
This suggests that population genetics in these highly differentiated species are dominated
by genetic drift within populations (Hutchison and Templeton 1999). One way this could
happen is through founder-effect drift (Mayr 1963), in the context of the “everything is
everywhere” hypothesis for microorganisms (Baas-Becking 1934; Finlay 2002).
According to this hypothesis, the distributions of microorganisms are limited not by
colonisation ability, but by habitat requirements, because of their great abundance and
long-distance dispersal ability (Finlay 2002). As a result, new populations can be founded
by a mix of colonists from throughout the original species range. If dispersal rates are
high, the entire species range will stay homogeneous, as in any model of gene flow.
However, if dispersal rates are low, populations should be randomly differentiated as a
result of receiving sets of colonists from different parts of the species range and there
should be no pattern of isolation by distance (Fontaneto et al. 2008). In theory, orchids
might also follow this model, as their seeds are on the same scale as microorganisms
105
Chapter 3. Gene flow and diversification
covered by the “everything is everywhere” hypothesis (< 2 mm), and single seed pods can
contain millions of seeds (Arditti and Ghani 2000). The data presented here are not
sufficient to test this possibility, but it would be worthwhile to investigate it further by
testing for isolation by distance using data from more individuals and more populations
throughout the full extent of an orchid species range.
Study limitations
An important limitation of this study is the use of current diversities of clades to
quantify diversification rates, as it is possible that differences in sister clade diversity are
instead the result of differences in diversity limits (Rabosky 2009a). Evidence is building
in the field of diversity patterns that much variation in clade richness (at least at the
family scale) is not the result of variation in diversification rates, but in limits to the
number of related species that can coexist in a region (Appendix I; Rabosky 2009a;
Vamosi and Vamosi 2010). This possibility cannot be ruled out here, but further studies
using younger sister clades or estimating diversification rates over an entire phylogenetic
tree (and thus able to test explicitly for diversity-dependent diversification) would avoid
this problem.
Another possible reason for the lack of evidence in favour of the original
hypothesis is the small number of species used. In particular, the ability of this study to
explore the interaction of range size with gene flow in affecting diversification was
limited because only one restricted species from a species-poor sister clade was sampled.
It is also possible that the species used here are not adequate representatives of their
clades and that a different or larger set of species would have shown a different pattern.
106
Chapter 3. Gene flow and diversification
Even so, the data here are a clear indication that the level of gene flow is not consistent
within clades.
Largely because most study species have patchy distributions, sampling within
species was not optimal. Some species could only be found in a restricted region,
precluding between-species comparisons of long-distance patterns of differentiation, and
few species had consistently large populations, leading to generally low sample sizes. In
addition, difficulties with AFLP genotyping reduced sample sizes for some species even
further. However, Fst values were tested for significance and only one Fst value over 0.05
(indicating at least moderate population differentiation; Freeland 2005) was found to be
not significant (Lepanthopsis floripecten, full dataset), even though there were a number
of relatively poorly sampled species with high Fst. This suggests that despite sub-optimal
sampling, the Fst values reported here reflect real patterns of differentiation in the study
species.
Additional ecological data would also be helpful for better understanding
diversification patterns in these orchids. Data regarding pollinators and mycorrhizal fungi,
neither of which is well known, would be particularly useful, as both could be major
drivers of orchid population structure. The pollen dispersal potential of different insects
known to pollinate orchids, for example, gnats versus hawkmoths, varies greatly.
Additionally, all orchid seeds require a mycorrhizal fungus partner in order to germinate
(Benzing 1987; Rasmussen 1995), and so orchid distributions should be limited by their
mycorrhizal specificity and the distributions of their mycorrhizal partners (Swarts et al.
2010). However, even anecdotal identification of pollinators is unavailable for most
tropical orchid species and even less is known about their mycorrhizal associations.
107
Chapter 3. Gene flow and diversification
Among the species studied here, only broad generalisations are known regarding
pollinators: Brassavola species are pollinated by sphingid moths; Lepanthes probably by
fungus gnats although pleurothallid orchids in general (including Lepanthes,
Lepanthopsis, Masdevallia, Trisetella, Platystele and Dryadella) tend to be fly-pollinated;
and Epidendrum species tend to be pollinated by Lepidoptera although there is a lot of
variability in the group (Pijl and Dodson 1966; Cingel 2001). Mycorrhizae of most
tropical orchids have not been studied at all and the few data that exist are not enough to
make any conclusions regarding mycorrhizal specificity. For the study genera, the only
indications in this direction are that multiple Epidendrum species tend to associate with
the same fungal genus, Epulorhiza (Zettler et al. 1998; Nogueira et al. 2005; Zettler et al.
2007; Pereira et al. 2009); and that in Masdevallia, for which mycorrhizal associations
were investigated using roots sampled from the same plants used in this study, the
widespread species M. nidifica appears to be more specific in mycorrhizal associations
than the narrow endemic M. rafaeliana (Renshaw 2010).
Ideas for future work
The difficulties encountered in this study would be most easily avoided by
carrying out a similar study using a better-known taxon or using data drawn from the
published literature. For better-studied taxa, such as temperate plants or birds, much more
detailed ecological data would be available to provide context and more previous
collection data would be available to help guide sampling. With a meta-analysis of
published data, it would be easier to collect data for many species. That said, there are
some obvious patterns in the orchid data that would be important to compare to patterns
108
Chapter 3. Gene flow and diversification
from other taxa. Tropical orchids have many unusual characteristics, including their tiny
seeds, complex relationships with pollinators and mycorrhizae and incredibly patchy
distributions, but it is unclear whether this means that they should also be unique in the
context of diversification.
Conclusion
The results presented here are inconclusive. It appears that the average level of
gene flow of a clade does not directly affect its diversification rate, but there is weak
support for a more complex scenario in which gene flow within species affects lineage
rates of diversification but changes through species’ lifetimes. Furthermore, no effect of
species ecology on the level of gene flow was found, but no data were available to test the
effects of pollinators and mycorrhizae, which should have strong effects on gene flow.
More data are needed to resolve these results. In addition, the lack of a strong effect of
gene flow on diversification rates does not preclude the existence of an effect of genetic
drift or local adaptation on diversification. I test this possibility in chapter four.
109
Chapter 4. Genetic drift, local adaptation, and diversification
Chapter 4. The effects of genetic drift and local adaptation on
clade diversification rates in Costa Rican orchids
Introduction
Just as gene flow is expected to affect speciation rates because of its role in
limiting population divergence, genetic drift and local adaptation could be expected to
affect speciation rates because of their role in generating population divergence (Grant
1981). In chapter three I showed that, for the set of orchid clades studied, there is no
direct link between the level of gene flow within species and the diversification rates of
clades. However, this does not rule out the possibility that levels of genetic drift or local
adaptation within species are traits that affect clade diversification.
Genetic drift, though once given a key role in models of adaptation (Wright 1931,
1982) and speciation (primarily in a set of related models involving small founder
populations, summarised in Carson and Templeton 1984) is now generally thought to
have only a minor role relative to selection (Sobel et al. 2010). Selection and gene flow
are thought to drive most speciation events, as they are expected to act much more
effectively than drift (Coyne et al. 1997, 2000; Gavrilets 2003). Furthermore, there is no
empirical support for a strong general role of drift in speciation (Rice and Hostert 1993;
Coyne et al. 1997, 2000). However, there are special cases in which theoretical models
suggest that genetic drift may have an important role in driving divergence (Gavrilets
2003): when it acts in combination with selection, and in taxa with characteristics
producing small effective population sizes (including low population density, patchy
110
Chapter 4. Genetic drift, local adaptation, and diversification
distributions and ephemeral populations). Epiphytic orchids may be one of the
exceptional taxa strongly affected by genetic drift. This is because small population sizes
and patchy distributions are common traits of epiphytic orchids, as well as high variance
in reproductive success, which also reduces effective population size (Tremblay 1997;
Tremblay et al. 2005). Furthermore, there is evidence that drift drives floral colour
variation in some orchid species (Tremblay et al. 2005).
Selection is central to most models of speciation, as it is believed to be the most
important process driving population differentiation and divergence (Schluter 2000, 2009;
Sobel et al. 2010). Unlike gene flow and drift, its role in driving diversification rates has
been tested in a number of studies. Barraclough (1995) found evidence for a positive link
between the strength of sexual selection (indicated by the magnitude of colour differences
between sexes) and diversification rate of birds. Seddon et al. (2008) confirmed this result
for birds both for song and colour and Stuart-Fox and Owens (2003) found similar results
for agamid lizards. However, all of these studies focused on the special case of sexual
selection, and used indirect surrogates of the likely strength of selection, rather than direct
measures. There is a lack of tests of the link between the degree of local adaptation and
diversification rates and of tests that use direct measures of selection derived from
population-level data.
Here I estimate the level of genetic drift by measuring the amount of neutral
genetic diversity within populations using AFLP data, and I estimate the level of local
adaptation by measuring the level of divergence between populations in leaf shape.
Average genetic diversity within populations should be lower in species that have
experienced increased genetic drift (Hedrick 2000). The level of morphological
111
Chapter 4. Genetic drift, local adaptation, and diversification
divergence between populations, when compared to divergence in neutral genetic
markers, is a direct measure of divergent selection (Merilä and Crnokrak 2001). I measure
divergence in leaf shape using geometric morphometrics and the Pst metric. Leaf shape is
used because it is an ecologically relevant trait associated with plant water, temperature
and light relations (Takenaka 1994; Royer et al. 2005; Wang 2007) and easily measured
in the field. Pst is a phenotype-based analogue to Fst that measures the fraction of variation
in phenotype that lies between populations out of the total amount of variation in the
sampled individuals. As with Fst, a Pst value of 0 indicates no differentiation between
populations, and the maximum Pst value of 1 indicates clear differentiation. Pst is also
similar to Qst, a metric of divergence in genes for quantitative traits (Spitze 1993), except
that Pst is influenced by environmental effects (Leinonen et al. 2006). When Qst is greater
than Fst it indicates that morphological divergence between populations is greater than
that expected from drift and gene flow alone, making the difference between Qst and Fst a
measure of local adaptation (Merilä and Crnokrak 2001), and the difference between Pst
and Fst an estimate of local adaptation. I also analyse overall Pst independently as a
measure of total phenotypic divergence, which should be the result of the combined
effects of gene flow, drift and local adaptation on morphological variation.
I hypothesise that increased genetic drift and local adaptation drive greater
population divergence within species and, thus, higher speciation rates. I use the same
approach as in chapter three to test this hypothesis. First, I test whether the levels of
genetic drift and local adaptation are heritable across lineages. Then I test for associations
between the level of each process and diversification rates of clades, taking into account
possible confounding influences of species range size and ecology.
112
Chapter 4. Genetic drift, local adaptation, and diversification
Methods
Study species and genetic data were the same as those used in chapter three. As in
chapter three, analyses (with the exception of estimating Pst) were carried out both on the
full dataset and a reduced dataset excluding populations with less than three samples (less
than two samples for Lepanthopsis floripecten). Except where stated otherwise, all
analyses were carried out using R v. 2.8.1 (R development core team 2008).
Genetic diversity within populations was estimated using Nei’s gene diversity, a
simple measure of variation based on allele frequencies (Nei 1987). Nei’s gene diversity
was calculated as
D
n
1  freq(allele1 )2  freq(allele0 )2 ,
n 1
(Eq. 4.1)
where n is the number of samples (Nei 1987), using the diversity function in the
AFLPdat package for R (Ehrich 2006). Gene diversity was only calculated for
populations containing five or more samples to reduce noise from small sample sizes. I
checked for a sample size bias using linear mixed effects models of genetic diversity
versus sample size with species as a random effect, compared to null models lacking a
sample size parameter. Because there was no evidence for a sample size bias (full dataset,
p = 0.12; reduced dataset, p = 0.22), average gene diversity for each species was
calculated simply as the mean over all populations.
To calculate Pst, leaf shapes were first recorded by photographing leaves from
each sample at the time of collection. Leaf outlines were digitised by hand using the
113
Chapter 4. Genetic drift, local adaptation, and diversification
Curves tool in TpsDig (available from the SUNY Stony Brook morphometrics webpage,
http://life.bio.sunysb.edu/morph/index.html; Rohlf 2008). Full outlines were digitised for
species with asymmetric leaves (Jacquiniella aporophylla) or leaves that are long, thin
and rarely straight (Brassavola nodosa, Dryadella odontostele and Trisetella triglochin).
Otherwise only the left-hand side of the leaf outline was digitised. Whenever possible,
two leaves per sample were digitised, and when a sample contained more than two leaves,
the longest leaves were chosen. Examples of digitised leaves are shown in Figure 4.1.
Populations with only one sample were excluded from analysis, except in the case of
Lepanthopsis floripecten. The digitised leaf shapes were converted into scores for
independent shape traits using eigenshape analysis (Lohmann 1983; MacLeod 1999) in
Mathematica v. 7 (Wolfram Research), with the morpho-tools notebooks kindly provided
by Jonathan Krieger (Krieger and MacLeod pers. comm.). Full leaf outlines were
analysed using the extended eigenshape method, with the leaf tip marked as a landmark;
half outlines were analysed as simple curves, with no landmarks, using the standard
eigenshape method. A different set of shape traits was generated for each species. After
shape scores were generated, the scores from multiple leaves from the same sample were
averaged in order to reduce noise from within-sample variability. Between and within
population variance components for each shape trait for each species were then calculated
from a linear mixed effects model with the shape trait as response variable and population
as a random effect. Pst for each shape trait was calculated as
Pst 
vb
,
v b  2v w h 2
(Eq. 4.2)
114
Chapter 4. Genetic drift, local adaptation, and diversification
where vb is the variance between populations, vw the variance within populations and h 2
the heritability of each trait (Leinonen et al. 2006). Because no data are available
regarding the heritability of leaf shape in orchids, I conservatively assumed a heritability
of 0.5 for all shape traits, meaning that half of the observed morphological variation
results from environmental or nonadditive genetic effects (Leinonen et al. 2006). Overall
Pst for each species was calculated as the mean Pst over all shape traits. The difference
between Pst and Fst was calculated by subtracting overall Fst for each species from overall
Pst.
Figure 4.1. Examples of digitised leaf outlines. Red circles mark points whose
coordinates were recorded. a) shows a Lepanthes ciliisepala sample; because leaves are
symmetrical, only one side of each leaf was digitised. b) shows a sample of Jacquiniella
aporophylla; because leaves are asymmetrical, whole leaf outlines were digitised.
115
Chapter 4. Genetic drift, local adaptation, and diversification
Phylogenetic heritability was estimated for gene diversity, overall Pst and the
difference between Pst and Fst as described in chapter three. The effects of genetic drift
and local adaptation on diversification were tested using two-tailed Wilcoxon signed rank
tests comparing mean trait values between sister clades. The mean gene diversity, overall
Pst and difference between Pst and Fst were calculated for each clade over all sampled
species.
Gene diversity, overall Pst and the difference between Pst and Fst were tested for
association with species range size and ecology in the same way that measures of Fst were
tested in chapter three. Association between all three measures and species range
measured as number of regions was tested using linear regression and the three measures
were compared between restricted and widespread species using two-tailed Wilcoxon
rank-sum tests. In addition, all three measures were tested for association with mean
branch circumference, variance in branch circumference, minimum elevation, maximum
elevation, elevation range and number of habitats using linear regression. For the tests
involving ecological variables, Bonferroni correction was used to correct for multiple
tests by considering only p-values less than 0.0083 to be significant.
Results
Gene diversity within populations does not vary greatly between species. It ranges
from 0.12 in Trisetella triglochin to 0.24 in Jacquiniella aporophylla (full dataset).
Overall Pst values vary from 0.015 in Trisetella triglochin to 0.106 in Lepanthopsis
floripecten. The difference between Pst and Fst ranges from -0.33 to 0.11. Overall Pst is
116
Chapter 4. Genetic drift, local adaptation, and diversification
greater than overall Fst, indicating local adaptation, for about one third of the study
species, in all cases species with very low Fst values (< 0.03). Gene diversity, overall Pst
and difference between Pst and Fst are listed for all species in Table 4.1.
The maximum likelihood value of lambda for gene diversity is 6.61x10-5 (p-value
for difference from 0 = 1; p-value for difference from 1 < 0.005). The maximum
likelihood value of lambda for overall Pst is 0.94 (p-value for difference from 0 = 0.15;
p-value for difference from 1 = 0.39) and this result is robust to variation in branch
lengths: the maximum likelihood value of lambda calculated with equal branch lengths is
0.27 (p-value for difference from 0 = 0.65; p-value for difference from 1 = 0.36). Unlike
lambda for overall Pst, the maximum likelihood value of lambda for the difference
between Pst and Fst is 6.61x10-5 (p-value for difference from 0 = 1; p-value for difference
from 1 < 0.005). The association between all three traits with species phylogeny is shown
in Figure 4.2 for the full dataset and Figure 4.3 for the reduced dataset.
Neither gene diversity nor overall Pst shows a significant association with
diversification rate (gene diversity, full dataset, p = 0.813, reduced dataset, p = 0.625;
overall Pst, p = 0.813). The relationship between the difference between Pst and Fst and
diversification approaches significance, but only for the reduced dataset, and in the
opposite direction as expected: increased diversification is associated with smaller values
of (Pst - Fst) (full dataset, p = 0.313, reduced dataset, p = 0.0625).
The only significant association between a species genetic trait and an ecological
trait is between gene diversity and elevation range for the reduced dataset (p = 0.0081,
r2 = 0.34; Figure 4.4). Otherwise, gene diversity, overall Pst and the difference between Pst
117
Chapter 4. Genetic drift, local adaptation, and diversification
and Fst are not associated with any measure of species range size (p-values between 0.07
and 0.536) or species ecology (p-values between 0.0166 and 0.995).
Table 4.1. Measures of levels of genetic drift, local adaptation and overall phenotypic
divergence for all study species. Species are separated according to sister clade pair.
Gene
diversity
Overall Pst
Pst - Fst
M. nidifica - full dataset
M. nidifica - reduced dataset
M. rafaeliana
T. triglochin
0.221
0.196
0.23
0.124
0.024
0.024
0.018
0.015
-0.333
-0.326
0.018
-0.048
L. ciliisepala
L. elata
L. floripecten - full dataset
L. floripecten - reduced dataset
0.15
0.126
0.219
0.236
0.051
0.031
0.106
0.106
0.024
-0.041
-0.035
0.106
P. propinqua - full dataset
P. propinqua - reduced dataset
P. stenostachya
D. odontostele
0.23
0.233
0.173
0.235
0.03
0.03
0.031
0.08
-0.144
-0.09
-0.211
-0.102
S. fusiformis
S. jimenezii
J. aporophylla - full dataset
J. aporophylla - reduced dataset
J. teretifolia - full dataset
J. teretifolia - reduced dataset
0.149
0.178
0.239
0.229
0.175
0.172
0.031
0.02
0.016
0.016
0.023
0.023
-0.206
0.02
0.016
0.016
0.023
0.01
E. exasperatum
E. laucheanum
E. vulgoamparoanum
B. nodosa - full dataset
B. nodosa - reduced dataset
0.177
0.217
0.229
0.176
0.176
0.035
0.042
0.059
0.038
0.038
-0.094
-0.003
0.04
-0.02
-0.016
Species
118
119
of genus and species.
dataset. Trait values for species are represented by colour of circles at branch tips. Species names are abbreviated by the first letters
Figure 4.2. Associations of (a) gene diversity, (b) overall Pst and (c) difference between Pst and Fst with species phylogeny for the full
120
reduced dataset. Traits are represented and species names abbreviated as in Figure 4.2.
Figure 4.3. Associations of (a) gene diversity, (b) overall Pst and (c) difference between Pst and Fst with species phylogeny for the
Chapter 4. Genetic drift, local adaptation, and diversification
Figure 4.4. Relationship between gene diversity and species elevation range for the
reduced dataset.
Discussion
Main findings
There is no evidence for the heritability of either the level of genetic drift or the
level of local adaptation within species. Following from this finding, it is unsurprising
that neither trait is associated with clade diversification rates. Both appear to be too labile
to be considered as clade traits, at least at the taxonomic level investigated here.
121
Chapter 4. Genetic drift, local adaptation, and diversification
In contrast, it is interesting that overall Pst has such a high heritability value,
although there is insufficient power to say that this value is significantly different from 0.
This may mean that the total level of phenotypic divergence within species is a heritable
clade trait, perhaps as a result of heritability in the variety of environments used by
species or in species responses to selection. It would be worthwhile to test this possibility
further with data from more species. However, like measures of drift and local adaptation,
overall Pst does not show any association with clade diversification rates.
The simplest conclusion from these results is that there is no relationship between
levels of local adaptation or genetic drift within species and diversification rates of higher
clades. There are also a number of alternative possibilities.
First, population genetic characteristics of species may be meaningful predictors
of diversification only at higher taxonomic levels, such as families, where differences
between clades outweigh the variability within them or where bigger differences exist
between clades. In this group of study clades, for instance, the measure of genetic drift
did not vary greatly either within or between clades. This possibility could be tested by
repeating this study with higher-level sister clades and data for more species.
Second, as discussed in chapter three, population genetic characteristics of species
may correlate with diversification rate, but with the signal confounded by the tendency
for the speciation process itself to change species population genetics. This does not seem
likely for genetic drift, as it is measured within populations and will not be affected by
populations being split between new daughter species. However, as local adaptation is a
relative measure quantified by comparing populations, it should generally be decreased
122
Chapter 4. Genetic drift, local adaptation, and diversification
by speciation in the same way as neutral genetic differentiation. If this were the case, an
association between species range size and degree of local adaptation, as was found for
gene flow in chapter three, would be expected. Alhough this association was not found,
such a complex scenario cannot yet be ruled out and would be worth exploring by
studying further the relationships between degree of local adaptation and species range
size and age.
Finally, it is possible that all lineages do not contribute equally to clade
diversification (i.e. there is imbalance even within the clades studied here) and that
lineages that contribute more to diversification, through higher rates of speciation, differ
in their population genetics from those that contribute less to diversification. This
possibility could be addressed by measuring the association between population genetics
and diversification rates using data for all species within a single clade.
Study limitations
The dataset used here is not perfect, and its imperfections illustrate the difficulties
inherent in comparative population genetic studies. Sampling was the greatest problem. In
this study there were four levels at which extensive sampling was important: number of
sister clade pairs included, number of species included in each clade, number of
populations included for each species and number of plants sampled from each
population. Sampling at the level of individuals and populations is equally important to
sampling at the level of species and clades, in order to have confidence in both the
population genetic data being compared and the outcome of the species comparisons
themselves. Optimally, many more species and sister clade pairs would be included than
123
Chapter 4. Genetic drift, local adaptation, and diversification
were included here; at least five populations would be sampled for each species,
extending across each species range; and genetic data would be available for ten
individuals for each population. In reality, however, with limited resources for fieldwork
and labwork, tradeoffs must be accepted. In further work, this problem would most easily
be resolved with collaborative efforts such as the IntraBioDiv consortium (Gugerli et al.
2008) or with databases of population genetic data. Nevertheless, this study had sufficient
sampling for a first test of the relationship between population genetics and
diversification and could have found evidence for a strong direct association if it existed.
Another difficulty in this study was choosing appropriate metrics to compare
between species and clades. Many metrics exist for measuring levels of genetic drift and
local adaptation (Freeland 2005), but it is impossible to distil all complexities of either
process into any single measure. The measures of genetic drift and local adaptation used
here were chosen as the simplest available measures that also make the fewest
assumptions in order to summarise the overall level of each process. It is unlikely that any
other measures of drift or local adaptation would have given qualitatively different
results.
Although it is unlikely that a different measure of local adaptation based on leaf
shape would have given a qualitatively different result, it is possible that leaf shape is not
representative of ecological divergence in these species and that calculating Pst for a
different trait would have changed the results. It would have been particularly worthwhile
to calculate Pst for floral traits. Orchid species are highly variable and often specialised in
pollination traits, and often reproductively isolated from close relatives by pollination
barriers alone (Schiestl and Schluter 2009). For these reasons, pollinator interactions are
124
Chapter 4. Genetic drift, local adaptation, and diversification
thought to be important in driving orchid speciation (Cozzolino and Widmer 2005;
Schiestl 2005), and variation in floral traits, which mediate pollinator interactions, might
be expected to correlate with speciation rates. In this study, analysing floral traits was not
possible because few flowering individuals were found in the field. However, future
studies using cultivated plants or species that flower more frequently in the field might be
able to address this.
Conclusion
Notwithstanding the obstacles discussed above, this study shows that comparative
population genetics can be used to address macroevolutionary questions. Comparative
studies so far have focused mainly on exploring the range of natural variation in
population genetics and testing how much of this variation can be explained by species
traits (Loveless and Hamrick 1984; Hamrick and Godt 1996; Morjan and Rieseberg
2004). As population genetic data become easier to generate with the advance of
molecular technology, and as population genetic data accumulate, potential applications
for such data increase. Rather than exploring only factors that control population genetics,
it is now possible to study the extent to which population genetics control other processes.
In this study, no association was found between population genetics and diversification,
but considering the potential complexities of the relationship between the two and the
central role of population genetics in speciation theory, this is a research area worth
exploring further.
125
Chapter 5. Conclusion
Chapter 5. Conclusion
In this thesis, I aimed to test a proposed framework in which the spatial scale of
speciation, set by population genetic characteristics of taxa, interacts with region area to
determine diversification rates.
Summary of results
In chapter two, I surveyed speciation events in angiosperms, bats, birds,
carnivorous mammals, ferns, lizards, Macrolepidoptera and snails on isolated oceanic
islands to measure taxonomic variation in the spatial scale of speciation. Using the
minimum island area with evidence of in situ speciation as an estimate of the spatial scale
of speciation, I found great variation between taxa: at one extreme, I found evidence for
snail speciation even on Nihoa, which is only 0.8 km2; at the other extreme, I found no
evidence for carnivore speciation on any island smaller than Madagascar (587,713 km2). I
also found that the probability of speciation within an island increases with island area for
all taxa except ferns, supporting a strong role of area in limiting speciation and
diversification. Finally, using data from a survey of the population genetic literature, I
found that minimum island areas for speciation are strongly correlated across taxa with
their average level of gene flow, supporting a link between the population genetic
characteristics of taxa and the spatial scale of speciation.
In chapters three and four, I generated population genetic data for orchid species
from sister clades differing in species richness to test the link between population genetic
126
Chapter 5. Conclusion
characteristics of taxa and their rates of diversification. I investigated the effects of three
population genetic processes: gene flow, quantified using Fst, a measure of betweenpopulation differentiation in neutral genetic markers; genetic drift, quantified using gene
diversity, a measure of genetic diversity within populations; and local adaptation,
quantified using the difference between Fst and Pst, a measure of between-population
differentiation in phenotypic traits. I found no evidence for a direct link between any of
these three processes and rates of diversification. However, I did find some support for a
link between gene flow and diversification rate mediated by species range size - all
species with great population differentiation were widespread species from the more
species-rich sister clade.
These results provide partial support for the framework I originally proposed.
Although the spatial scale of speciation is clearly linked to both gene flow and the
probability of speciation within a region, there is no support for a link between the
diversification of taxa and their population genetic characteristics (which I expected
would affect diversification via their effect on the spatial scale of speciation). This may
mean that the spatial scale of speciation is important in the context of individual
speciation events, but does not affect overall diversification rates of clades. However, the
results of the orchid study are not conclusive - it is possible that different results would
come from choosing sister clades at a different taxonomic scale or from including data for
more species. Thus, the relationship between the spatial scale of speciation and
diversification rates of taxa requires further study.
127
Chapter 5. Conclusion
Directions for future work
Following on the results presented here, the most interesting question to address
next is whether a direct relationship exists between the spatial scale of speciation and
diversification rates. This could be tested by comparing minimum areas for speciation
across sister groups or across clades with known phylogenies, while also accounting for
clade range size or the amount of area available to each clade. If a strong relationship
exists, groups with smaller minimum areas for speciation should contain more species.
More broadly, there is great potential for improving our understanding of
diversification and diversity patterns by considering the spatial scale of speciation and
population genetic characteristics of taxa in macroevolutionary studies. Comparative
population genetics is a challenging approach, due to the difficulty of assembling
comparable and high quality data across a suite of species. Nevertheless, as I have
demonstrated throughout this thesis, it can be used effectively for investigating
macroevolutionary questions, and it could be applied to a range of macroevolutionary
topics. Measuring the spatial scale of speciation of taxa is also difficult, unless wellbounded regions (such as the oceanic islands I used in chapter two) can be identified in
which speciation events can be inferred. However, with sufficient environmental and
biogeographical information, it should also be possible to make such inferences in
continental settings. This would expand the range of taxa for which the spatial scale of
speciation could be measured and make it possible to use this measure in a broad range of
macroevolutionary studies.
128
Chapter 5. Conclusion
Finally, this work demonstrates the value of integrating the effects of both
environmental and taxon characters in studies of diversification. In chapter two, I found
that the probability of speciation within an island was best modelled taking into account
both taxon identity and island environmental characteristics. This result, combined with
many other studies that have found evidence for an interaction between environmental
and taxon characteristics (reviewed in chapter one), suggests that studies of
diversification should either explicitly control for regional effects (as I did in chapters
three and four by limiting study species to natives of Costa Rica and sampling
populations within Costa Rica only) or include environmental characteristics of regions in
analyses.
General conclusions
The variety and complexity of diversity patterns on Earth are fascinating, yet
challenging to untangle. Centuries of thought and research have identified many possible
factors and mechanisms that could contribute to structuring species richness, but as yet
there is no consensus as to which are the most important or most general in their effects.
The framework proposed and explored in this thesis, which links area, population genetic
characteristics of clades and the spatial scale of speciation, is surely insufficient to explain
the full breadth of existing diversity patterns. Nevertheless, as it is applicable to all taxa
and integrates both environmental and taxon characteristics, it has potential as a backbone
for theories of diversification and diversity patterns.
129
Bibliography
Bibliography
Ackerman, J. D., J. C. Trejo-Torres and Y. Crespo-Chuy. 2007. Orchids of the West
Indies: predictability of diversity and endemism. Journal of Biogeography 34: 779786.
Arditti, J. and A. K. A. Ghani. 2000. Tansley review No. 110 - Numerical and physical
properties of orchid seeds and their biological implications. New Phytologist 145:
367-421.
Australian Biological Resources Study, Canberra. 2008. "Flora of Australia online:
oceanic islands excluding Norfolk and Lord Howe Islands." from
http://www.environment.gov.au/biodiversity/abrs/onlineresources/flora/50/index.html
Baas-Becking, L. G. M. 1934. Geobiologie of inleiding tot de milieukunde. W. P. van
Stockum and Zoon, The Hague, Netherlands.
Barluenga, M., K. Stolting, W. Salzburger, M. Muschick and A. Meyer. 2006. Sympatric
speciation in Nicaraguan crater lake cichlid fish. Nature 439: 719-723.
Barraclough, T. G., P. H. Harvey and S. Nee. 1995. Sexual selection and taxonomic
diversity in passerine birds. Proceedings of the Royal Society B: Biological Sciences
259: 211-215.
Barraclough, T. G. and S. Nee. 2001. Phylogenetics and speciation. Trends in Ecology &
Evolution 16: 391-399.
Barraclough, T. G. and A. P. Vogler. 2000. Detecting the geographical pattern of
speciation from species-level phylogenies. American Naturalist 155: 419-434.
Barraclough, T. G., A. P. Vogler and P. H. Harvey. 1998. Revealing the factors that
promote speciation. Philosophical Transactions of the Royal Society B: Biological
Sciences 353: 241-249.
130
Bibliography
Barton, N. H. 2001. The evolutionary consequences of gene flow and local adaptation:
future approaches. Pages 329-340 in J. Clobert, E. Danchin, A. A. Dhondt and J. D.
Nichols, eds. Dispersal. Oxford University Press, Oxford.
Bates, D. 2007. lme4: linear mixed-effects models using S4 classes. R package version
0.99875-9.
Bauer, A. M. 1988. Reptiles and the biogeographic interpretation of New Caledonia.
Tuatara 30: 39-50.
Beaumont, M. A. and D. J. Balding. 2004. Identifying adaptive genetic divergence among
populations from genome scans. Molecular Ecology 13: 969-980.
Beaumont, M. A. and R. A. Nichols. 1996. Evaluating loci for use in the genetic analysis
of population structure. Proceedings of the Royal Society B: Biological Sciences 263:
1619-1626.
Belliure, J., G. Sorci, A. P. Moller and J. Clobert. 2000. Dispersal distances predict
subspecies richness in birds. Journal of Evolutionary Biology 13: 480-487.
Benzing, D. H. 1987. Vascular epiphytism: taxonomic participation and adaptive
diversity. Annals of the Missouri Botanical Garden 74: 183-204.
Berlocher, S. H. 1998. Can sympatric speciation via host or habitat shift be proven from
phylogenetic and biogeographic evidence? Pages 99-113 in D. J. Howard and S. H.
Berlocher, eds. Endless forms - species and speciation. Oxford University Press, New
York.
Bohonak, A. J. 1999. Dispersal, gene flow, and population structure. Quarterly Review of
Biology 74: 21-45.
Bonin, A., E. Bellemain, P. Bronken Eidesen, F. Pompanon, C. Brochmann and P.
Taberlet. 2004. How to track and assess genotyping errors in population genetics
studies. Molecular Ecology 13: 3261-3273.
131
Bibliography
Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: a
practical information-theoretic approach. Springer-Verlag, New York.
Butlin, R. K., J. Galindo and J. W. Grahame. 2008. Sympatric, parapatric or allopatric:
the most important way to classify speciation? Philosophical Transactions of the
Royal Society B: Biological Sciences 363: 2997-3007.
Cardillo, M., J. S. Huxtable and L. Bromham. 2003. Geographic range size, life history
and rates of diversification in Australian mammals. Journal of Evolutionary Biology
16: 282-288.
Cardillo, M., C. D. L. Orme and I. P. F. Owens. 2005. Testing for latitudinal bias in
diversification rates: An example using New World birds. Ecology 86: 2278-2287.
Cardoso, G. C. and P. G. Mota. 2008. Speciational evolution of coloration in the genus
Carduelis. Evolution 62: 753-762.
Carlquist, S. J. 1974. Island biology. Columbia University Press, New York.
Carson, H. L. and A. R. Templeton. 1984. Genetic revolutions in relation to speciation
phenomena: The founding of new populations. Annual Review of Ecology and
Systematics 15: 97-131.
Castella, V., M. Ruedi, L. Excoffier, C. Ibanez, R. Arlettaz and J. Hausser. 2000. Is the
Gibraltar Strait a barrier to gene flow for the bat Myotis myotis (Chiroptera :
Vespertilionidae)? Molecular Ecology 9: 1761-1772.
Chan, K. M. A. and B. R. Moore. 2002. Whole-tree methods for detecting differential
diversification rates. Systematic Biology 51: 855-865.
Chase, M. W. 1988. Obligate twig epiphytes: a distinct subset of Neotropical
orchidaceous epiphytes. Selbyana 10: 24-30.
Chase, M. W. and H. H. Hills. 1991. Silica-gel - an ideal material for field preservation of
leaf samples for DNA studies. Taxon 40: 215-220.
132
Bibliography
Chiba, S. 1999. Accelerated evolution of land snails Mandarina in the oceanic Bonin
Islands: evidence from mitochondrial DNA sequences. Evolution 53: 460-471.
Cingel, N. A. v. d. 2001. An atlas of orchid pollination: America, Africa, Asia and
Australia. A.A. Balkema Publishers, Rotterdam.
Coyne, J. A., N. H. Barton and M. Turelli. 1997. Perspective: a critique of Sewall
Wright's shifting balance theory of evolution. Evolution 51: 643-671.
Coyne, J. A., N. H. Barton and M. Turelli. 2000. Is Wright's shifting balance process
important in evolution? Evolution 54: 306-317.
Coyne, J. A. and H. A. Orr. 2004. Speciation. Sinauer Associates, Inc., Sunderland,
Massachusetts.
Coyne, J. A. and T. D. Price. 2000. Little evidence for sympatric speciation in island
birds. Evolution 54: 2166-2171.
Cozzolino, S. and A. Widmer. 2005. Orchid diversity: an evolutionary consequence of
deception? Trends in Ecology & Evolution 20: 487-494.
Cracraft, J. 1982. A non-equilibrium theory for the rate-control of speciation and
extinction and the origin of macroevolutionary patterns. Systematic Zoology 31: 348365.
Cracraft, J. 1985. Biological diversification and its causes. Annals of the Missouri
Botanical Garden 72: 794-822.
Dahl, A. 2004. "United Nations Environment Program island directory." from
http://islands.unep.ch/isldir.htm
Darwin, C. 1862. The various contrivances by which orchids are fertilized by insects.
John Murray, London.
133
Bibliography
Davies, T. J., V. Savolainen, M. W. Chase, J. Moat and T. G. Barraclough. 2004.
Environmental energy and evolutionary rates in flowering plants. Proceedings of the
Royal Society B: Biological Sciences 271: 2195-2200.
Dial, K. P. and J. M. Marzluff. 1989. Nonrandom diversification within taxonomic
assemblages. Systematic Zoology 38: 26-37.
Diamond, J. M. 1977. Continental and insular speciation in Pacific land birds. Systematic
Zoology 26: 263-268.
Diamond, J. M. 2005. Collapse : how societies choose to fail or succeed. Viking, New
York.
Doebeli, M. and U. Dieckmann. 2003. Speciation along environmental gradients. Nature
421: 259-264.
Dray, S., A. B. Dufour and D. Chessel. 2007. The ade4 package-II: Two-table and Ktable methods. R News 7: 47-52.
Dressler, R. L. 2005. How many orchid species. Selbyana 26: 155-158.
Eastwood, A., Q. C. B. Cronk, J. C. Vogel, A. Hemp and M. Gibby. 2004. Comparison of
molecular and morphological data on St Helena: Elaphoglossum. Plant Systematics
and Evolution 245: 93-106.
Ehrich, D. 2006. AFLPDAT: a collection of R functions for convenient handling of AFLP
data. Molecular Ecology Notes 6: 603-604.
Emerson, B. C. and P. Oromi. 2005. Diversification of the forest beetle genus Tarphius
on the Canary Islands, and the evolutionary origins of island endemics. Evolution 59:
586-598.
Endler, J. A. 1977. Geographic variation, speciation, and clines. Princeton University
Press, Princeton.
134
Bibliography
Eriksson, O. and B. Bremer. 1991. Fruit characteristics, life forms, and species richness in
the plant family Rubiaceae. American Naturalist 138: 751-761.
Excoffier, L. 2001. Analysis of population subdivision. Pages 271-307 in D. J. Balding,
M. Bishop and C. Cannings, eds. Handbook of statistical genetics. John Wiley &
Sons, Chichester.
Excoffier, L. and H. E. L. Lischer. 2010. Arlequin suite ver 3.5: a new series of programs
to perform population genetics analyses under Linux and Windows. Molecular
Ecology Resources 10: 564-567.
Excoffier, L., P. E. Smouse and J. M. Quattro. 1992. Analysis of molecular variance
inferred from metric distances among DNA haplotypes: application to human
mitochondrial DNA restriction data. Genetics 131: 479-91.
Farris, J. S. 1976. Expected asymmetry of phylogenetic trees. Systematic Zoology 25:
196-198.
Felsenstein, J. 1985. Phylogenies and the comparative method. American Naturalist 125:
1-15.
Finlay, B. J. 2002. Global dispersal of free-living microbial eukaryote species. Science
296: 1061-1063.
Fontaneto, D., T. G. Barraclough, K. Chen, C. Ricci and E. A. Herniou. 2008. Molecular
evidence for broad-scale distributions in bdelloid rotifers: everything is not
everywhere but most things are very widespread. Molecular Ecology 17: 3136-3146.
Frankham, R., J. D. Ballou and D. A. Briscoe. 2010. Introduction to conservation
genetics. Cambridge University Press, Cambridge, UK.
Freckleton, R. P., P. H. Harvey and M. Pagel. 2002. Phylogenetic analysis and
comparative data: a test and review of evidence. American Naturalist 160: 712-26.
Freeland, J. R. 2005. Molecular Ecology. John Wiley & Sons, Ltd, Chichester.
135
Bibliography
Fu, Y. X. 1997. Statistical tests of neutrality of mutations against population growth,
hitchhiking and background selection. Genetics 147: 915-925.
Fuller, R. C. 2008. Genetic incompatibilities in killifish and the role of environment.
Evolution 62: 3056-3068.
Funk, D. J., K. E. Filchak and J. L. Feder. 2002. Herbivorous insects: model systems for
the comparative study of speciation ecology. Genetica 116: 251-267.
Funk, D. J., P. Nosil and W. J. Etges. 2006. Ecological divergence exhibits consistently
positive associations with reproductive isolation across disparate taxa. Proceedings of
the National Academy of Sciences of the USA 103: 3209-3213.
Futuyma, D. 1986. Evolutionary Biology. Sinauer, Sunderland.
Gaggioti, O. E. and I. Hanski. 2004. Mechanisms of population extinction. Pages 337-366
in I. Hanski and O. E. Gaggioti, eds. Ecology, genetics, and evolution of
metapopulations. Elsevier Academic Press, Amsterdam.
Gaston, K. J. and T. M. Blackburn. 1997. Age, area and avian diversification. Biological
Journal of the Linnean Society 62: 239-253.
Gavrilets, S. 2003. Perspective: Models of speciation: What have we learned in 40 years?
Evolution 57: 2197-2215.
Gavrilets, S. 2004. Fitness landscapes and the origin of species. Princeton University
Press, Princeton.
Gavrilets, S. and J. B. Losos. 2009. Adaptive radiation: contrasting theory with data.
Science 323: 732-737.
Gavrilets, S. and A. Vose. 2005. Dynamic patterns of adaptive radiation. Proceedings of
the National Academy of Sciences of the USA 102: 18040-18045.
Gentry, A. H. and C. H. Dodson. 1987. Diversity and biogeography of Neotropical
vascular epiphytes. Annals of the Missouri Botanical Garden 74: 205-233.
136
Bibliography
Gillespie, R. 2004. Community assembly through adaptive radiation in Hawaiian spiders.
Science 303: 356-359.
Gillespie, R. G. and B. G. Baldwin. 2009. Island biogeography of remote archipelagoes:
interplay between ecological and evolutionary processes. in J. B. Losos and R. E.
Ricklefs, eds. The theory of island biogeography at 40: impacts and prospects.
Princeton University Press, Princeton.
Givnish, T. J., K. C. Millam, A. R. Mast, T. B. Paterson, T. J. Theim, A. L. Hipp, J. M.
Henss, J. F. Smith, K. R. Wood and K. J. Sytsma. 2009. Origin, adaptive radiation and
diversification of the Hawaiian lobeliads (Asterales: Campanulaceae). Proceedings of
the Royal Society B: Biological Sciences 276: 407-416.
Govaerts, R., J. Pfahl, M. A. Campacci, D. Holland Baptista, H. Tigges, J. Shaw, P.
Cribb, A. George, K. Kreuz and J. Wood. 2007. World Checklist of Orchidaceae. The
Board of Trustees of the Royal Botanic Gardens, Kew, Published on the Internet;
http://www.kew.org./wcsp/.
Govaerts, R., J. Pfahl, M. A. Campacci, D. Holland Baptista, H. Tigges, J. Shaw, P.
Cribb, A. George, K. Kreuz and J. Wood. 2008. World Checklist of Orchidaceae. The
Board of Trustees of the Royal Botanic Gardens, Kew, Published on the Internet;
http://www.kew.org./wcsp/.
Govaerts, R., J. Pfahl, M. A. Campacci, D. Holland Baptista, H. Tigges, J. Shaw, P.
Cribb, A. George, K. Kreuz and J. Wood. 2010. World Checklist of Orchidaceae. The
Board of Trustees of the Royal Botanic Gardens, Kew, Published on the Internet;
http://www.kew.org./wcsp/.
Govindaraju, D. R. 1988. Relationship between dispersal ability and levels of gene flow
in plants. Oikos 52: 31-35.
Grant, P. R. and B. R. Grant. 2009. Sympatric speciation, immigration, and hybridization
in island birds. in J. B. Losos and R. E. Ricklefs, eds. The theory of island
biogeography at 40: impacts and prospects. Princeton University Press, Princeton.
137
Bibliography
Grant, V. 1981. Plant Speciation. Columbia University Press, New York.
Gravendeel, B., A. Smithson, F. J. W. Slik and A. Schuiteman. 2004. Epiphytism and
pollinator specialization: drivers for orchid diversity? Philosophical Transactions of
the Royal Society B: Biological Sciences 359: 1523-1535.
Gugerli, F., T. Englisch, H. Niklfeld, A. Tribsch, Z. Mirek, M. Ronikier, N. E.
Zimmermann, R. Holderegger, P. Taberlet and C. IntraBioDiv. 2008. Relationships
among levels of biodiversity and the relevance of intraspecific diversity in
conservation - a project synopsis. Perspectives in Plant Ecology Evolution and
Systematics 10: 259-281.
Hammel, B. E., M. H. Grayum, C. Herrera and N. Zamora, Eds. 2003. Manual de Plantas
de Costa Rica Volumen III. Missouri Botanical Garden Press, St. Louis.
Hamrick, J. L. and M. J. W. Godt. 1996. Effects of life history traits on genetic diversity
in plant species. Philosophical Transactions of the Royal Society B: Biological
Sciences 351: 1291-1298.
Harvey, P. H., R. M. May and S. Nee. 1994. Phylogenies without fossils. Evolution 48:
523-529.
Hedrick, P. W. 2000. Genetics of populations. Jones and Bartlett Publishers, Sudbury,
Massachusetts.
Herrera, C. M. 1989. Seed dispersal by animals - a role in angiosperm diversification.
American Naturalist 133: 309-322.
Hodges, S. A. and M. L. Arnold. 1995. Spurring plant diversification: Are floral nectar
spurs a key innovation? Proceedings of the Royal Society B: Biological Sciences 262:
343-348.
Holt, R. D. and Gomulkiewicz. 2004. Conservation implications of niche conservatism
and evolution in heterogeneous environments. Pages 244-264 in R. Ferrière, U.
138
Bibliography
Dieckmann and D. Couvet, eds. Evolutionary conservation biology. Cambridge
University Press, Cambridge.
Hutchison, D. W. and A. R. Templeton. 1999. Correlation of pair-wise genetic and
geographic distance measures: Inferring the relative influences of gene flow and drift
on the distribution of genetic variability. Evolution 53: 1898-1914.
Isaac, N. J. B., P. M. Agapow, P. H. Harvey and A. Purvis. 2003. Phylogenetically nested
comparisons for testing correlates of species richness: A simulation study of
continuous variables. Evolution 57: 18-26.
Jablonski, D. 1986. Larval ecology and macroevolution in marine invertebrates. Bulletin
of Marine Science 39: 565-587.
Jablonski, D. 2008. Species selection: theory and data. Annual Review of Ecology,
Evolution, and Systematics 39: 501-524.
Jobson, R. W. and V. A. Albert. 2002. Molecular rates parallel diversification contrasts
between carnivorous plant sister lineages. Cladistics 18: 127-136.
Katoh, K., K. Misawa, K. Kuma and T. Miyata. 2002. MAFFT: a novel method for rapid
multiple sequence alignment based on fast Fourier transform. Nucleic Acids Research
30: 3059-66.
Kay, K. M. and R. D. Sargent. 2009. The role of animal pollination in plant speciation:
Integrating ecology, geography, and genetics. Annual Review of Ecology Evolution
and Systematics 40: 637-656.
Kinlan, B. P. and S. D. Gaines. 2003. Propagule dispersal in marine and terrestrial
environments: a community perspective. Ecology 84: 2007-2020.
Kisel, Y., L. M. McInnes, N. H. Toomey and C. D. L. Orme. in review. How
diversification rates and diversity limits combine to create large-scale species-area
relationships. Philosophical Transactions of the Royal Society B: Biological Sciences.
139
Bibliography
Kodandaramaiah, U. and N. Wahlberg. 2007. Out-of-Africa origin and dispersal-mediated
diversification of the butterfly genus Junonia (Nymphalidae : Nymphalinae). Journal
of Evolutionary Biology 20: 2181-2191.
Lande, R. 1988. Genetics and demography in biological conservation. Science 241: 14551460.
Leinonen, T., J. M. Cano, H. Makinen and J. Merila. 2006. Contrasting patterns of body
shape and neutral genetic divergence in marine and lake populations of threespine
sticklebacks. Journal of Evolutionary Biology 19: 1803-1812.
Lepage, D. 2008. "Avibase - the world bird database." from http://www.bsceoc.org/avibase/avibase.jsp
Lohmann, G. P. 1983. Eigenshape analysis of micro-fossils - a general morphometric
procedure for describing changes in shape. Journal of the International Association for
Mathematical Geology 15: 659-672.
Long, J. C. 1986. The allelic correlation structure of Gainj-speaking and Kalam-speaking
people. 1. The estimation and interpretation of Wright's F-Statistics. Genetics 112:
629-647.
Losos, J. B. and C. E. Parent. 2009. The speciation-area relationship. in J. B. Losos and
R. E. Ricklefs, eds. The theory of island biogeography at 40: impacts and prospects.
Princeton University Press, Princeton.
Losos, J. B. and D. Schluter. 2000. Analysis of an evolutionary species-area relationship.
Nature 408: 847-850.
Loveless, M. D. and J. L. Hamrick. 1984. Ecological determinants of genetic structure in
plant populations. Annual Review of Ecology and Systematics 15: 65-95.
MacArthur, R. H. and E. O. Wilson. 1967. The theory of island biogeography. Princeton
University Press, Princeton.
140
Bibliography
MacLeod, N. 1999. Generalizing and extending the eigenshape method of shape space
visualization and analysis. Paleobiology 25: 107-138.
Marzluff, J. M. and K. P. Dial. 1991. Life-history correlates of taxonomic diversity.
Ecology 72: 428-439.
Mayhew, P. J. 2007. Why are there so many insect species? Perspectives from fossils and
phylogenies. Biological Reviews 82: 425-454.
Mayr, E. 1942. Systematics and the origin of species from the viewpoint of a zoologist.
Columbia University Press, New York.
Mayr, E. 1963. Animal species and evolution. Belknap Press of Harvard University Press,
Cambridge.
Mayr, E. 1965. Avifauna: turnover on islands. Science 150: 1587-1588.
Mayr, E. 1982. Processes of speciation in animals. Pages 1-19 in A. R. I. Liss, ed.
Mechanisms of Speciation. Alan R. Liss Inc., New York.
McPeek, M. A. 2008. The ecological dynamics of clade diversification and community
assembly. American Naturalist 172: E270-E284.
Merilä, J. and P. Crnokrak. 2001. Comparison of genetic differentiation at marker loci
and quantitative traits. Journal of Evolutionary Biology 14: 892-903.
Mitter, C., B. Farrell and B. Wiegmann. 1988. The phylogenetic study of adaptive zones has phytophagy promoted insect diversification. American Naturalist 132: 107-128.
Morjan, C. L. and L. H. Rieseberg. 2004. How species evolve collectively: implications
of gene flow and selection for the spread of advantageous alleles. Molecular Ecology
13: 1341-1356.
Nee, S. 2006. Birth-death models in macroevolution. Annual Review of Ecology
Evolution and Systematics 37: 1-17.
141
Bibliography
Nee, S., E. C. Holmes, R. M. May and P. H. Harvey. 1994. Extinction rates can be
estimated from molecular phylogenies. Philosophical Transactions of the Royal
Society B: Biological Sciences 344: 77-82.
Nee, S., A. O. Mooers and P. H. Harvey. 1992. Tempo and mode of evolution revealed
from molecular phylogenies. Proceedings of the National Academy of Sciences of the
USA 89: 8322-8326.
Nei, M. 1973. Analysis of gene diversity in subdivided populations. Proceedings of the
National Academy of Sciences of the USA 70: 3321-3.
Nei, M. 1987. Molecular evolutionary genetics. Columbia University Press, New York.
Nogueira, R. E., O. L. Pereira, M. C. M. Kasuya, M. C. d. S. Lanna and M. P. Mendonça.
2005. Mycorrhizal fungi associated to orchids growing in "campos rupestres" in
"Quadrilatero Ferrifero" region, Minas Gerais State, Brazil. Acta Botanica Brasilica
19: 417-424.
Nosil, P. 2009. Adaptive population divergence in cryptic color-pattern following a
reduction in gene flow. Evolution 63: 1902-12.
Orme, C. D. L., R. P. Freckleton and G. H. Thomas. 2008. CAIC: Comparative Analyses
using Independent Contrasts.
Otto, S. P. and J. Whitton. 2000. Polyploid incidence and evolution. Annual Review of
Genetics 34: 401-437.
Owens, I., P. Bennett and P. Harvey. 1999. Species richness among birds: body size, life
history, sexual selection or ecology? Proceedings of the Royal Society B: Biological
Sciences 266: 933-939.
Pagel, M. 1999. Inferring the historical patterns of biological evolution. Nature 401: 877884.
142
Bibliography
Parent, C. E. and B. J. Crespi. 2006. Sequential colonization and diversification of
Galapagos endemic land snail genus Bulimulus (Gastropoda, Stylommatophora).
Evolution 60: 2311-2328.
Paulay, G. 1985. Adaptive radiation on an isolated oceanic island - the Cryptorhynchinae
(Curculionidae) of Rapa revisited. Biological Journal of the Linnean Society 26: 95187.
Paulay, G. 1994. Biodiversity on oceanic islands - its origin and extinction. American
Zoologist 34: 134-144.
Paulay, G. and C. Meyer. 2006. Dispersal and divergence across the greatest ocean
region: do larvae matter? Integrative and Comparative Biology 46: 269-281.
Pereira, M. C., O. L. Pereira, M. D. Costa, R. B. Rocha and M. C. M. Kasuya. 2009.
Diversity of mycorrhizal fungi Epulorhiza spp. isolated from Epidendrum secundum
(Orchidaceae). Revista Brasileira De Ciencia Do Solo 33: 1187-1197.
Peterson, M. A. and R. F. Denno. 1998. The influence of dispersal and diet breadth on
patterns of genetic isolation by distance in phytophagous insects. American Naturalist
152: 428-446.
Phillimore, A. B., R. P. Freckleton, C. D. L. Orme and I. P. F. Owens. 2006. Ecology
predicts large-scale patterns of phylogenetic diversification in birds. American
Naturalist 168: 220-229.
Phillimore, A. B., C. D. L. Orme, R. G. Davies, J. D. Hadfield, W. J. Reed, K. J. Gaston,
R. P. Freckleton and I. P. F. Owens. 2007. Biogeographical basis of recent phenotypic
divergence among birds: a global study of subspecies richness. Evolution 61: 942957.
Phillimore, A. B., C. D. L. Orme, G. H. Thomas, T. M. Blackburn, P. M. Bennett, K. J.
Gaston and I. P. F. Owens. 2008. Sympatric speciation in birds is rare: insights from
range data and simulations. American Naturalist 171: 646-657.
143
Bibliography
Pickle, J. and J. Kirtley. 2008. Analyzing Digital Images 2008. Museum of Science,
Boston, Boston.
Pigot, A. L., A. B. Phillimore, I. P. F. Owens and C. D. L. Orme. in press. The shape and
temporal dynamics of phylogenetic trees arising from geographic speciation.
Systematic Biology.
Pijl, L. v. d. and C. H. Dodson. 1966. Orchid flowers: their pollination and evolution.
Fairchild Tropical Garden and University of Miami Press, Coral Gables.
Price, J. P. and W. L. Wagner. 2004. Speciation in Hawaiian angiosperm lineages: cause,
consequence, and mode. Evolution 58: 2185-2200.
Pridgeon, A. M., R. Solano and M. W. Chase. 2001. Phylogenetic relationships in
Pleurothallidinae (Orchidaceae): Combined evidence from nuclear and plastid DNA
sequences. American Journal of Botany 88: 2286-2308.
Purvis, A. 2008. Phylogenetic approaches to the study of extinction. Annual Review of
Ecology Evolution and Systematics 39: 301-319.
Quental, T. B. and C. R. Marshall. 2010. Diversity dynamics: molecular phylogenies need
the fossil record. Trends in Ecology & Evolution 25: 434-441.
R development core team. 2007. R: a language and environment for statistical computing.
R Foundation for Statistical Computing, Vienna.
R development core team. 2008. R: a language and environment for statistical computing.
R Foundation for Statistical Computing, Vienna.
Rabosky, D. L. 2009a. Ecological limits and diversification rate: alternative paradigms to
explain the variation in species richness among clades and regions. Ecology Letters
12: 735-743.
Rabosky, D. L. 2009b. Ecological limits on clade diversification in higher taxa. American
Naturalist 173: 662-674.
144
Bibliography
Rabosky, D. L. 2010. Extinction rates should not be estimated from molecular
phylogenies. Evolution 64: 1816-1824.
Ramos-Onsins, S. E. and J. Rozas. 2002. Statistical properties of new neutrality tests
against population growth. Molecular Biology and Evolution 19: 2092-100.
Ranker, T., S. Floyd and P. Trapp. 1994. Multiple colonizations of Asplenium adiantumnigrum onto the Hawaiian archipelago. Evolution 48: 1364-1370.
Rasmussen, H. N. 1995. Terrestrial orchids from seed to mycotrophic plant. Cambridge
University Press, Cambridge.
Raup, D. M., S. J. Gould, T. J. M. Schopf and D. S. Simberloff. 1973. Stochastic models
of phylogeny and evolution of diversity. Journal of Geology 81: 525-542.
Renshaw, P. B. 2010. Are patterns of mycorrhizal specificity conserved across the orchid
genus Masdevallia? Imperial College, London. MSc thesis.
Rice, W. R. and E. E. Hostert. 1993. Laboratory experiments on speciation - what have
we learned in 40 years. Evolution 47: 1637-1653.
Ricklefs, R. E. 2003. Global diversification rates of passerine birds. Proceedings of the
Royal Society B: Biological Sciences 270: 2285-2291.
Ricklefs, R. E. 2007a. Estimating diversification rates from phylogenetic information.
Trends in Ecology & Evolution 22: 601-610.
Ricklefs, R. E. 2007b. History and diversity: explorations at the intersection of ecology
and evolution. American Naturalist 170: S56-S70.
Rohlf, F. J. 2008. tpsDig, digitize landmarks and outlines, version 2.12, Department of
Ecology and Evolution, State University of New York at Stony Brook.
Rosenzweig, M. L. 1995. Species diversity in space and time. Cambridge University
Press, Cambridge.
145
Bibliography
Rosindell, J., S. J. Cornell, S. P. Hubbell and R. S. Etienne. 2010. Protracted speciation
revitalizes the neutral theory of biodiversity. Ecology Letters 13: 716-727.
Royer, D. L., P. Wilf, D. A. Janesko, E. A. Kowalski and D. L. Dilcher. 2005.
Correlations of climate and plant ecology to leaf size and shape: Potential proxies for
the fossil record. American Journal of Botany 92: 1141-1151.
Ryan, P. G., P. Bloomer, C. L. Moloney, T. J. Grant and W. Delport. 2007. Ecological
speciation in South Atlantic island finches. Science 315: 1420-1423.
Sanderson, M. J. and M. J. Donoghue. 1996. Reconstructing shifts in diversification rates
on phylogenetic trees. Trends in Ecology & Evolution 11: 15-20.
Savolainen, V., M. C. Anstett, C. Lexer, I. Hutton, J. J. Clarkson, M. V. Norup, M. P.
Powell, D. Springate, N. Salamin and W. J. Baker. 2006. Sympatric speciation in
palms on an oceanic island. Nature 441: 210-213.
Schiestl, F. P. 2005. On the success of a swindle: pollination by deception in orchids.
Naturwissenschaften 92: 255-264.
Schiestl, F. P. and P. M. Schluter. 2009. Floral isolation, specialized pollination, and
pollinator behavior in orchids. Annual Review of Entomology 54: 425-446.
Schliewen, U. K., D. Tautz and S. Paabo. 1994. Sympatric speciation suggested by
monophyly of crater lake cichlids. Nature 368: 629-632.
Schluter, D. 2000. The ecology of adaptive radiation. Oxford University Press, Oxford.
Schluter, D. 2001. Ecology and the origin of species. Trends in Ecology & Evolution 16:
372-380.
Schluter, D. 2009. Evidence for ecological speciation and its alternative. Science 323:
737-741.
146
Bibliography
Seddon, N., R. M. Merrill and J. A. Tobias. 2008. Sexually selected traits predict patterns
of species richness in a diverse clade of suboscine birds. American Naturalist 171:
620-631.
Seehausen, O. 2006. African cichlid fish: a model system in adaptive radiation research.
Proceedings of the Royal Society B: Biological Sciences 273: 1987-1998.
Sequeira, A., A. Lanteri, R. Albelo, S. Bhattacharya and M. Sijapati. 2008. Colonization
history, ecological shifts and diversification in the evolution of endemic Galapagos
weevils. Molecular Ecology 17: 1089-1107.
Simpson, G. G. 1953. The major features of evolution. Columbia University Press, New
York.
Slatkin, M. 1973. Gene flow and selection in a cline. Genetics 75: 733-756.
Slatkin, M. 1985. Gene flow in natural populations. Annual Review of Ecology and
Systematics 16: 393-430.
Smith, A. B. 2007. Marine diversity through the Phanerozoic: problems and prospects.
Journal of the Geological Society 164: 731-745.
Smith, P. L. 1988. Paleoscene .11. Paleobiogeography and plate-tectonics. Geoscience
Canada 15: 261-279.
Sobel, J. M., G. F. Chen, L. R. Watt and D. W. Schemske. 2010. The biology of
speciation. Evolution 64: 295-315.
Spitze, K. 1993. Population structure in Daphnia obtusa: Quantitative genetic and
allozymic variation. Genetics 135: 367-374.
Steadman, D. W. 2006. Extinction & biogeography of tropical Pacific birds. University of
Chicago Press, Chicago.
Stebbins, G. L. 1950. Variation and evolution in plants. Columbia University Press, New
York.
147
Bibliography
Stuart-Fox, D. and I. P. F. Owens. 2003. Species richness in agamid lizards: chance, body
size, sexual selection or ecology? Journal of Evolutionary Biology 16: 659-669.
Stuessy, T. F., D. J. Crawford and C. Marticorena. 1990. Patterns of phylogeny in the
endemic vascular flora of the Juan Fernandez Islands, Chile. Systematic Botany 15:
338-346.
Stuessy, T. F., G. Jakubowsky, R. S. Gomez, M. Pfosser, P. M. Schluter, T. Fer, B. Y.
Sun and H. Kato. 2006. Anagenetic evolution in island plants. Journal of
Biogeography 33: 1259-1265.
Swarts, N. D., E. A. Sinclair, A. Francis and K. W. Dixon. 2010. Ecological
specialization in mycorrhizal symbiosis leads to rarity in an endangered orchid.
Molecular Ecology 19: 3226-3242.
Takenaka, A. 1994. Effects of leaf blade narrowness and petiole length on the light
capture efficiency of a shoot. Ecological Research 9: 109-114.
Thorpe, W. H. 1945. The evolutionary significance of habitat selection. Journal of
Animal Ecology 14: 67-70.
Thrall, P. H., J. J. Burdon and B. R. Murray. 2000. The metapopulation paradigm: a
fragmented view of conservation biology. Pages 75-95 in A. G. Young and G. M.
Clarke, eds. Genetics, demography and viability of fragmented populations.
Cambridge University Press, Cambridge.
Tiffney, B. H. and S. J. Mazer. 1995. Angiosperm growth habit, dispersal and
diversification reconsidered. Evolutionary Ecology 9: 93-117.
Tilman, D. and P. M. Kareiva. 1997. Spatial ecology: the role of space in population
dynamics and interspecific interactions. Princeton University Press, Princeton.
Tremblay, R. L. 1997. Distribution and dispersion patterns of individuals in nine species
of Lepanthes (Orchidaceae). Biotropica 29: 38-45.
148
Bibliography
Tremblay, R. L., J. D. Ackerman, J. K. Zimmerman and R. N. Calvo. 2005. Variation in
sexual reproduction in orchids and its evolutionary consequences: a spasmodic
journey to diversification. Biological Journal of the Linnean Society 84: 1-54.
Valentine, J. W. and E. M. Moores. 1970. Plate-tectonic regulation of faunal diversity and
sea level - a model. Nature 228: 657-659.
Vamosi, J. C. and S. M. Vamosi. 2010. Key innovations within a geographical context in
flowering plants: towards resolving Darwin's abominable mystery. Ecology Letters
13: 1270-1279.
van den Berg, C., W. E. Higgins, R. L. Dressler, W. M. Whitten, M. A. Soto-Arenas and
M. W. Chase. 2009. A phylogenetic study of Laeliinae (Orchidaceae) based on
combined nuclear and plastid DNA sequences. Annals of Botany 104: 417-30.
Vekemans, X. and O. J. Hardy. 2004. New insights from fine-scale spatial genetic
structure analyses in plant populations. Molecular Ecology 13: 921-935.
Veness, C. 2008. "Geodesic distance between two latitude/longitude points using
Vincenty ellipsoid formula in JavaScript." from http://www.movabletype.co.uk/scripts/latlong-vincenty.html.
von Hagen, K. B. and J. W. Kadereit. 2003. The diversification of Halenia
(Gentianaceae): Ecological opportunity versus key innovation. Evolution 57: 25072518.
Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. v. d. Lee, M. Hornes, A. Frijters, J. Pot, J.
Peleman, M. Kuiper and M. Zabeau. 1995. AFLP: a new technique for DNA
fingerprinting. Nucleic Acids Research 23: 4407-4414.
Vrba, E. S. 1984. Evolutionary pattern and process in the sister-group AlcelaphiniAepycerotini (Mammalia: Bovidae). Pages 62-79 in N. Eldredge and S. M. Stanley,
eds. Living Fossils. Springer-Verlag, New York.
149
Bibliography
Wagner, W. H. 1969. The role and taxonomic treatment of hybrids. Bioscience 19: 785795.
Wang, Y. T. 2007. Average daily temperature and reversed day/night temperature
regulate vegetative and reproductive responses of a Doritis pulcherrima Lindley
hybrid. Hortscience 42: 68-70.
Webb, T. J. and K. J. Gaston. 2003. On the heritability of geographic range sizes.
American Naturalist 161: 553-566.
Whitlock, R., H. Hipperson, M. Mannarelli, R. K. Butlin and T. Burke. 2008. An
objective, rapid and reproducible method for scoring AFLP peak-height data that
minimizes genotyping error. Molecular Ecology Resources 8: 725-735.
Whittaker, R. J. and J. M. Fernandez-Palacios. 2007. Island biogeography. Oxford
University Press, Oxford.
Whittaker, R. J., K. A. Triantis and R. J. Ladle. 2008. A general dynamic theory of
oceanic island biogeography. Journal of Biogeography 35: 977-994.
Whittaker, R. J., K. A. Triantis and R. J. Ladle. 2009. A general dynamic theory of
oceanic island biogeography: extending the MacArthur-Wilson theory to
accommodate the rise and fall of volcanic islands. in J. B. Losos and R. E. Ricklefs,
eds. The theory of island biogeography at 40: impacts and prospects. Princeton
University Press, Princeton.
Wright, S. 1931. Evolution in Mendelian populations. Genetics 16: 97-159.
Wright, S. 1982. The shifting balance theory and macroevolution. Annual Review of
Genetics 16: 1-19.
Wu, C. I. and C. T. Ting. 2004. Genes and speciation. Nature Reviews Genetics 5: 114122.
150
Bibliography
Zera, A. J. 1981. Genetic structure of 2 species of waterstriders (Gerridae, Hemiptera)
with differing degrees of winglessness. Evolution 35: 218-225.
Zettler, L. W., S. B. Poulter, K. I. McDonald and S. L. Stewart. 2007. Conservationdriven propagation of an epiphytic orchid (Epidendrum nocturnum) with a
mycorrhizal fungus. Hortscience 42: 135-139.
Zettler, L. W., T. Wilson Delaney and J. A. Sunley. 1998. Seed propagation of the
epiphytic green-fly orchid, Epidendrum conopseum R. Brown, using its endophytic
fungus. Selbyana 19: 249-253.
Zwickl, D. J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large
biological sequence datasets under the maximum likelihood criterion. The University
of Texas at Austin. Ph.D. thesis.
151
Appendix I.
Appendix I.
How diversification rates and diversity limits combine to create large-scale speciesarea relationships
Yael Kisel1*, Lynsey McInnes 1,2*, Nicola H Toomey1, C David L Orme1,2
1
Division of Biology, Imperial College London, Silwood Park, Ascot, Berkshire, SL5 7PY, UK
2
Grantham Institute for Climate Change, Imperial College, London SW7 2AZ, UK
* Authors for correspondence: [email protected] and [email protected]; these
authors contributed equally to this work.
Abstract
Species-area relationships have mostly been treated from an ecological perspective,
focusing on immigration, local extinction, and resource-based limits to species
coexistence. However, a full understanding across large regions is impossible without
also considering speciation and global extinction. Rates of both speciation and extinction
are known to be strongly affected by area and thus should contribute to spatial patterns of
diversity. Here, we explore how variation in diversification rates and ecologicallymediated diversity limits among regions of different sizes can result in the formation of
species-area relationships. We explain how this area-related variation in diversification
can be caused by either the direct effects of area or the effects of factors that are highly
correlated with area, such as habitat diversity and population size. We also review

This appendix is an invited contribution for a special issue of the Philosophical Transactions of the Royal
Society B: Biological Sciences, with the theme of “Global Biodiversity of Mammals”, to be published in
2011.
152
Appendix I.
environmental, clade-specific, and historical factors that affect diversification and
diversity limits but are not highly correlated with region area, and thus are likely to cause
scatter in observed species-area relationships. We present new analyses using data on the
distributions, ages and traits of mammalian species to illustrate these mechanisms; in
doing so we provide an integrated perspective on the evolutionary processes shaping
species-area relationships.
Introduction
The species-area relationship (SAR), which describes an increase in the number of
species as region size increases, is a nearly ubiquitous pattern of biodiversity. SARs exist
at a wide range of spatial scales, from local to global, and in a wide range of taxa,
including mammals (Pagel et al. 1991). In the ecological literature, SARs have been
explained by considering the factors that limit species from immigrating into,
establishing, and persisting in a region (Arrhenius 1921; MacArthur & Wilson 1967;
Preston 1960). However, at large geographic scales, in situ diversification contributes
significantly to generating diversity, and so a full understanding of the generation of
species-area relationships at such scales is impossible without also considering the
macroevolutionary processes of speciation and extinction (Losos & Schluter 2000;
Rosenzweig 1995; 1998).
Here, we explore the evolutionary underpinnings of large-scale SARs, outlining the roles
of area itself, environmental variation, clade traits and historical contingency. We adopt a
model of clade diversity in which clade diversification within regions is diversitydependent and SARs are created by the scaling of both diversity limits and diversification
153
Appendix I.
rates with area. We support this discussion with new analyses using mammals as they are
a well-known, diverse, and globally-distributed group with a wide variety of life
histories, occupying a wide range of habitats and with robust data for many key traits
(Jones et al. 2009).
SARs have traditionally been treated as the outcome of differences between regions in
the balance between immigration and local extinction (MacArthur & Wilson 1967) and in
the number of species that can coexist (Arrhenius 1921; Preston 1960). However, it was
later recognized that SARs may not be controlled by the same processes at all spatial
scales (Palmer & White 1994; Rosenzweig 1995). At the smallest scales, SARs result
from more complete sampling of the local biota as the area sampled increases, and as
such they are sampling rather than biological phenomena. At larger scales (sampling all
of the local biota), classical ecological explanations apply, with SARs emerging as a
result of more species being able to immigrate into and persist in larger areas. Finally, at
the largest scales, differences between regions in rates of speciation and extinction should
be the main factor generating SARs (Losos & Schluter 2000; Rosenzweig 1995, 1998).
Here, we focus on SARs at the largest geographic scale. For mammals, this large-scale
phase is likely to occur only when considering quite large regions: in Kisel &
Barraclough’s (2010) study of the spatial scale of speciation, the two mammal groups
represented (bats and carnivores) both required a region larger than 400,000 - 500,000
km2 for any in situ speciation to occur.
We use a framework of diversity-dependent cladogenesis (Section 1) to explore how the
area (Section 2) and environment (Section 3) of regions affect diversification and
154
Appendix I.
diversity limits in the generation of SARs. We also examine the role of clade traits
(Section 4) and temporal patterns of diversification (Section 5) in modulating the shape
of SARs. See Table 1 for a summary of the factors addressed.
Table 1. Summary of factors affecting diversification rates and diversity limits.
Type of factor
Factor
Effects on Speciation
Rate
Effects on Extinction
Rate
Effects on Diversity
Limits
Area
↑ potential for
geographic isolation of
separated populations
↓ by ↑ survival of
refuge populations
Population Size
↑ rate of appearance of
beneficial mutations,
standing genetic
variation, persistence of
incipient species
↓ by buffering
populations from
demographic
stochasticity,
environmental
disasters, habitat loss
Habitat Diversity
↑ population divergence
through local adaptation
Fragmentation
/Topographic
Diversity
↑ isolation of
populations; however
past a certain point, will
↓ speciation by ↓
population size
Environmental
factors not
strongly
correlated with
area
Energy
availability
↑ rate of molecular
evolution, rate of coevolutionary dynamics,
size of populations
supported
Clade traits
Life history traits
faster life cycle ↑
speciation by increasing
mutation rate
faster life cycle ↓
extinction by ↑
resilience to
disturbance
Range size
larger range sizes ↑
speciation by ↑ potential
for isolation of
populations
larger range sizes ↓
extinction by ↑
survival of refuge
populations
smaller range sizes ↑
diversity limit by allowing
more species to pack into
same area
Niche breadth
narrower niche breadths
associated with ↑
speciation
narrower niche
breadths associated
with ↑ extinction
narrower niche breadths ↑
diversity limit by allowing
finer subdivision of niche
space
Dispersal
↓ by reducing potential
isolation of populations,
but can also ↑ speciation
rate by ↑ rate at which
species colonize new
regions
↓ by ↑ resilience to
disturbance
↓ if high dispersal ability
is associated with large,
nonoverlapping species
ranges
Environmental
factors strongly
correlated with
area
↑ number of species with
viable populations
supported
↑ niche space available
if fragmentation
results in too small
patches of area or
habitat, will ↑
extinction rate
↑ by allowing more
ecologically equivalent
species to be supported
↑ by facilitating
specialization to narrow
niches
155
Appendix I.
Methods
We used the geographic distributions of 4650 terrestrial mammal species within
PanTHERIA (Jones et al. 2009) to explore the scaling of species richness with area. The
choice of appropriate regions at a global scale is not obvious, so we have taken two
approaches to identifying provinces. First, we used botanical sampling regions based on
geopolitical units (Taxonomic Database Working Group (TDWG), Brummit 2001) to
subdivide continental landmasses, although we further separated disjunct sub-regions,
such as islands. Second, we identified species presence in equal-area grid cells at a
resolution (96.5 km) comparable to a 1° grid. We then used complete linkage hierarchical
clustering on the Jaccard distance (Linder et al. 2005; but see Kreft & Jetz in press)
between grid cells to identify approximate mammalian biotic regions. Both methods are
hierarchically nested between levels but regions within the same level are not nested. The
fineness of subdivision can also be varied: the TDWG standard defines four levels,
ranging roughly from different biomes at the coarsest scale (level 1) to subdivisions
within countries at the finest scale (level 4); the hierarchical clustering can be cut at
different “heights” to give different numbers of regions and we have used 50, 100, 150
and 200 regions (mapped in Fig S1). The two region types differ in ways that are likely to
affect the outcome: for example, political boundaries are likely to more finely partition
large biotically homogenous regions in the temperate zone and agglomerate smaller
biotically heterogeneous tropical regions. We used both methods and the variety of scales
to assess the robustness of our conclusions to the details of sampling. Separating
discontinuous parts of detailed polygons of TDWG regions, in combination with the
imprecision in global species distribution maps, led to a large number of tiny islands and
boundary regions with implausible biotas. We therefore removed all regions at the
156
Appendix I.
coarsest TDWG scale that did not contain at least one species endemic to that region,
reducing 3974 candidate regions to 117. All nested subdivisions of these 117 regions at
the finer TDWG scales were retained.
The areas of both geopolitical and clustered regions were calculated using an equal-area
projection of the land within each region (Fig S2). We recorded both the total and
endemic mammalian species richness for each region and fitted SARs at each scale of
subdivision using linear models on log-log axes to estimate the slope. We modeled
species richness (S) as a power of area (A) as S = cAz (Arrhenius 1921; Rosenzweig
1995): although there has been considerable debate about the shape of SARs (Lomolino
2000; Scheiner 2003), our results should be general to alternative functions. For all
further analyses, we used the most finely divided regions and compared results using
TDWG Level 4 and 200 biotic regions. We also explored the differences between slopes
of SARs arising from species endemic to a region versus those occurring in more than
one region, and the variation in slopes of SARs within mammalian orders.
To investigate the additional explanatory power of habitat diversity and environmental
variables, we used two variables to capture different elements of habitat diversity: the
diversity of land cover classes (GLCC v2.0, http://edc2.usgs.gov/glcc/glcc.php),
calculated as the inverse of Simpson’s diversity index (1-D) on the relative areas of the
classes within each region; and the log range in elevation (GTOPO30,
http://eros.usgs.gov/) within each region. We considered two environmental variables
within regions: the mean annual temperature (www.worldclim.org) and the mean
normalized difference vegetation index (NDVI, Los, pers comm., updated versions of
157
Appendix I.
Los et al. 2000). We fitted multiple regressions with log area and each of these four
variables in turn as predictors of log species richness. For each variable, we tested
whether it showed a significant interaction with area as well as its significance as a main
effect. All covariates were mean centred and standardized to facilitate the interpretation
and comparison of these models (Schielzeth in press).
An approximate measure of habitat breadth for mammalian species was found by
counting the number of GLCC habitat cover classes across all the 96.5km cells
intersecting each species' range. This number correlates strongly with the species'
geographic range (Kendall's tau = 0.61) and we therefore also estimated a number of
major habitats by counting only those habitats with a proportional contribution of at least
0.142. This cutoff was selected because it minimises the observed correlation between the
resulting number of major habitats and the species' range size (Kendall's tau = -0.0002).
We then calculated the Kendall's correlation between family species richness and both the
number of habitats and number of major habitats.
In order to explore the effects of area on the temporal patterns of recent diversification
within mammals, we identified two sets of monophyletic clades from the mammal
supertree (Bininda-Emonds et al. 2007; 2008), excluding monotypic clades. One set had
crown ages younger than 20 MY (421 clades), the other had crown ages younger than 10
MY (616 clades) and was nested inside the older set. We recorded each clade’s species
richness, stem-group age and present-day area (either the total area of all TDWG level 4
provinces or of all biotic regions (finest scale) in which the component species occurred).
We then fitted a suite of six models of diversification rate across each set (Rabosky
158
Appendix I.
2009b; Phillimore 2010). The most complex model is an extension of those outlined in
Rabosky (2009b) and Phillimore (2010) and fits an exponential decline in diversification
with rate z over clade age (t) from an initial diversification rate (), but where log
present-day area (A) contributes to both initial  (scaling by c) and the rate of decline
(scaling by p); the overall diversification rate is always scaled by the relative extinction
rate ():

ri   c log Ai e (z  p log A i )t i 1   
We also fitted five simplifications of this model by fixing sets of parameters at zero: a
constant diversification rate across clades (c, z and p fixed), a constant diversification rate
scaled by individual clade area (z and p fixed), an exponential decline in rate within
clades (c and p fixed), an exponential decline from an initial  scaled by area (p fixed)
and an exponential decline at a rate z scaled by area (c fixed). We optimized parameter
estimates for the free variables in each model by maximizing the sum of log likelihoods
of the observed species richness (n) across clades given clade age and the model
estimates (following Bokma 2003; Ricklefs 2009; Phillimore 2010). The models were not
nested and we therefore used AIC to assess relative model support. As the two methods
to define regions gave qualitatively similar results, we report only the TDWG analysis
here (see ESM for biotic region results).
Section 1: A verbal model for clade diversification in space
Diversity-dependent models of diversification have two main features: a growth phase,
where the clade in question diversifies until it reaches an external limit; and an
equilibrium phase, where species identity turns over but clade size fluctuates about that
159
Appendix I.
limit (Sepkoski 1978, Alroy 1998). The precise shape of diversity-dependent
diversification has been debated (Nee et al. 1992; Rabosky 2009a), but the exact shape of
the diversification trajectory should not change the broad-scale implications of the
existence of diversity-dependent diversification. There is taxonomic, phylogenetic, and
paleontological evidence to support the existence of diversity-dependent diversification
in many cases, described variously as “ecological limits on diversity,” “diversification
slowdowns,” and “diversity equilibria” (Sepkoski 1976; Rabosky 2009a; Alroy 1998,
Nee et al. 1992; Rabosky & Lovette 2008, Vamosi & Vamosi in press).
A variety of processes could generate diversity-dependent diversification. Perhaps the
most commonly referenced is a model of ecological limits wherein, as available niches
are filled, speciation declines and new species are only added to a region following
extinctions and release of sufficient niche space (McKinney 1998; Rabosky 2009a). Such
a mechanism would provide a link between the ecological processes typically associated
with SARs and the evolutionary processes being proposed here. Alternatively, reduction
of both population and range sizes as diversity increases could lead to decreased rates of
speciation and increased rates of extinction and thus a diversification slowdown
conceivably divorced from any niche-based mechanism (Pigot et al. in press; Rosenzweig
1975).
Within our diversity-dependent framework, there are only three features of a clade’s
diversification curve that can vary: the speed at which a region initially accumulates
species (Fig. 1a), the diversity limit (or equilibrium species richness, Fig. 1b), and the age
at which diversification begins (Fig. 1c) (see also Rabosky 2009a). Before equilibrium is
160
Appendix I.
reached, the richness of clades depends only on their age and their rate of diversification.
In contrast, clade sizes at equilibrium depend on their diversity limits, which are
controlled by the interaction of external factors with clade traits (Mallet submitted, and
see below). SARs will emerge from this model whenever diversification rates and/or
diversity limits are higher in larger regions (Fig. 2). When a clade inhabits multiple
separate regions of different areas, the species richness of that clade will be higher in the
larger regions, creating a SAR.
Figure 1. Variation in patterns of clade diversification from A) initial rate of
diversification, B) equilibrium diversity, C) clade age and D) reinforcing (solid grey) and
opposing (dashed grey) combinations of rate and equilibrium diversity. Sampling clade
diversity at the time specified by the vertical line demonstrates the variation possible.
161
Appendix I.
Figure 2. The development of a species area relationship (SAR) across three regions (X,
Y, Z), in which both initial rate of diversification and equilibrium diversity increase with
area (A). The resulting SAR across regions (B) exhibits power law scaling both before
(dashed line) and after (solid line) the regions have reached equilibrium diversity. It is
important to discriminate between the clade diversification curves (A) and SARs (B);
each region will follow a particular diversification trajectory but contributes a single
point to the SAR.
162
Appendix I.
Globally, mammalian species richness shows strong scaling with area between nonnested provinces for both TDWG and clustered regions at all four scales (Fig. 3). These
are well described by power laws but there are differences between the two region types
(Fig 3a): clustered regions show consistent slopes across changing scales (0.41 - 0.43),
whereas TDWG regions show a decline in slope from 0.47 to 0.24 with increasing
subdivision. These slopes lie within the range of 65 previously reported slopes from
mammal power law SARs (Fig 3b; Drakare et al. 2006), but the higher values fall toward
the top of the reported range (92% quantile). The changes in slope between TDWG scales
is accompanied by higher intercepts (Table S1, Figures S3 and S4) and is primarily
driven by small political units, such as the Vatican City and Likoma, within species-rich
areas (Fig 3c); these outliers are not found in small regions based on mammalian biotas
(Fig 3d). In all cases, endemic species also show significant scaling with area but with
reduced slopes compared to total and non-endemic species richness (Fig 3b-c, Table S1,
Figures S3 and S4).
163
Appendix I.
Figure 3. A) Slopes and their standard errors of species area relationships (SARs) for
4560 terrestrial mammals at four different scales across geopolitical regions (T 1-4) and
biotic regions (C 1-4). B) Distribution of power law exponents from mammalian SARs
showing the range of non-nested region sizes considered (grey lines – data from Drakare
et al. 2006; black lines – values from panel above). Scatterplots show the distribution and
least squares fit of SARs for T4 (C) and C4 (D) for total (black) and endemic species
richness (grey). See also Table S1 and Figures S3 and S4.
164
Appendix I.
We also tested how well area explains variation in diversification rate across sets of
mammalian clades. For both sets of clades (crown group age < 20 MY and < 10MY,
Tables 2, S2), an exponential decline in diversification rate is best supported,
demonstrating apparent limits to diversity. For clades younger than 20 MY, the most
complex model was best supported, with clades occupying larger areas having increased
initial diversification rates and decreased rate of decline. For clades < 10 MY, a simpler
model, with area affecting only the rate of decline, could not be rejected. These results
suggest that for mammals the decline in diversification rate as a region fills is more
strongly affected by available area than the initial rate. Nevertheless, support for an effect
of available area on initial rate was still found for both clade sets and the similar
likelihoods for the younger clades may simply reflect individual clade differences within
the set tested (see also Cardillo et al. 2005; Linder 2008).
165
Appendix I.
Table 2. Summary of diversification models fitted to mammalian clades with crown age
younger than (a) 20 and (b) 10 million years before present using the TDWG Level 4,. Six
models of diversification are fitted representing: constant rate (1), constant rate scaled by
region area (2), exponential decline (3), exponential decline region area scaling initial
rate (4), rate of decline (5) or both (6). In each case, the maximum likelihood estimate of
the model is reported for each free parameter within the bounds shown. Dashed
parameter estimates were fixed at zero. The overall best-fit model for each period is
shown in bold. Results for biotic regions are presented in Table S2.
Lambda
c
[-1,1]
[-0.2,0.2]
a) 20 MY
1
0.340
--2 -0.300
0.040
3
0.790
--4 -0.300
0.040
5
0.474
--6 -0.260
0.040
b) TDWG Level 4, 10 MY
1
0.265
--2 -0.223
0.030
3
0.530
--4 -0.193
0.031
--5
0.377
6 -0.064
0.023
z
[-0.2,0.2]
p
[0.2,0.2]
Epsilon
[0.5,0.999]
-----0.030
-0.030
-0.138
-0.100
--------0.007
0.004
0.990
0.990
0.990
0.610
0.814
0.610
-----0.043
-0.043
-0.232
-0.120
--------0.012
0.005
0.999
0.990
0.999
0.520
0.711
0.500
ΔAICc
Likelihoo
d
222.4
136.8
187.2
53.1
19.1
0.0
-1410.0
-1366.2
-1391.4
-1323.3
-1306.3
-1295.8
164.2
93.4
110.6
30.0
0.0
1.16
-1578.3
-1541.9
-1550.5
-1509.1
-1494.1
-1493.7
166
Appendix I.
Section 2: Generating SARs in an evolutionary framework
In explanations of SARs, area is frequently viewed as a proxy or summary variable
(Hubbell 2001) acting only indirectly via other variables, such as population size and
habitat diversity that are highly correlated with area (MacArthur & Wilson 1967). The
individual effects of area and such correlated factors are difficult to separate in practice
(Triantis et al. 2003; Kallimanis et al. 2008), and their relative importance is likely to
vary depending on the taxon concerned (Rosenzweig 1995; Ricklefs & Lovette 1999).
However, we believe that area could conceivably have some direct effects, and we
discuss these first.
Direct effects of area
We can see only two ways that area could control diversity directly (ie. without invoking
increased population sizes or habitat variety). Firstly, extinction rates should be lower in
larger regions, in which refuge populations are more likely to survive after any
catastrophic disturbance affecting only part of the region (Wiley & Wunderle 1994).
Secondly, if populations are patchily distributed, speciation rates should be higher in
larger areas (Losos & Schluter 2000), where distances between populations can be larger
and barriers that can cause vicariant speciation are likely to be larger and more numerous
(Rosenzweig 1995). It could be argued that the effect of barriers is really an indirect
effect of area via fragmentation, and we discuss this point further below. Greater
geographic isolation between populations will lead to higher speciation rates if: 1) there
is sufficient selection pressure and/or genetic drift to drive population divergence through
to reproductive isolation (although there is no evidence for speciation via genetic drift on
its own: Coyne & Orr 2004); 2) gene flow is the main force preventing population
167
Appendix I.
divergence and speciation (Slatkin 1987); and 3) the regions considered are large enough
for populations to be sufficiently isolated to permit speciation. The definition of ‘large
enough’ will depend on the dispersal ability of the organism and the strength of selection
relative to gene flow, as poorer dispersers will attain sufficient isolation in smaller
regions (Kisel & Barraclough 2010), as will species whose populations experience
stronger divergent selection (Slatkin 1973; Slatkin 1985).
Effects of area via population size
Because larger regions are able to support greater total numbers of individuals (Brown
1995), and thus are also likely to have species with larger population sizes, the effects of
population size on diversification can contribute to the generation of SARs. In fact, many
of the effects of population size that we describe below have previously been described as
direct effects of area itself (MacArthur & Wilson 1967, Ricklefs & Lovette 1999). It is
well established that larger populations are less likely to go extinct, as they are more
buffered from the effects of demographic stochasticity, environmental disasters, and
habitat loss (Lande 1993; Rosenzweig 1995). Additionally, there are three ways that
larger population size may drive higher speciation rates. First, new beneficial mutations
will arise faster in larger populations (Willi et al. 2006), allowing faster divergence
between separated populations if mutation limits speciation (Schluter 2009). Second,
larger populations hold more standing genetic variation (Frankham 1996; Leimu et al.
2006) for selection to work on (Schluter & Conte 2009; Weber 1990). Third, newly
isolated populations resulting from the break-up of larger populations will also be larger,
and therefore more likely to survive long enough to diverge into new species (Chown &
Gaston 2000). In addition to effects on rates of diversification, the total abundance of
168
Appendix I.
individuals supported by a region places a hard limit on the number of species that the
region can hold. If we assume that all species are ecologically identical and so have the
same minimum viable population size (Gilpin & Soulé 1986, Hubbell 2001), then larger
regions will be able to support more species at sustainable equilibrium population sizes.
Effects of area via habitat diversity and fragmentation
Some authors have suggested that SARs are only a proxy for the scaling of species
richness with habitat diversity (MacArthur & Wilson 1967; Baldi 2008; Triantis et al.
2003; Losos & Parent 2010), and indeed habitat diversity and area are typically very
highly correlated. Along steep environmental gradients, and in heterogeneous habitats,
populations can more easily become specialised to different habitats, making ecological
speciation more likely and perhaps more rapid (Schluter 2009). Regions with high habitat
diversity also have a higher number of possible distinct niches or niche combinations
(Hutchinson & MacArthur 1959), thus increasing the number of species that can coexist
at equilibrium.
High levels of regional fragmentation can also elevate diversification rate and diversity
limits, by providing a textured landscape with subunits that are physically isolated from
one another but environmentally equivalent. Barrier formation can occur through many
processes, including river formation, mountain building, sea-level fluctuations, volcanic
uplift, and habitat fragmentation, and is more likely in larger regions. Barriers elevate
diversification rate by separating previously interacting populations, which are then more
likely to evolve reproductive isolation (Rosenzweig 1995). In addition, fragmentation can
boost equilibrium diversity, as ecologically equivalent species can be maintained in
169
Appendix I.
separated sub-regions (Shmida & Wilson 1985; Orme et al. in prep). For example,
Esselstyn et al. (2009) suggest that tree shrew diversity in the Phillippines has arisen
predominantly via speciation in allopatry on newly formed islands, with limited apparent
morphological or ecological differentiation. One particularly important measure of
regional fragmentation is topographic complexity, as environmental turnover along
altitude gradients is a barrier to many species’ ranges (McInnes et al. 2009). The richness
of uniquely adapted, restricted-range endemics found along altitudinal transects in
tropical mountains is perhaps the classic example of such fine-scale spatial partitioning
(Janzen 1967; Rahbek & Graves 2001).
The effects of fragmentation on species richness will show a complex relationship to the
total summed area of the subunits. While greater fragmentation of a region may permit
more species to exist within the same total area, it may also push the area of the
component fragments below a size which can maintain viable populations (Gilpin &
Soule 1986; Maurer & Nott 1998) or generate endemics (Losos & Schluter 2000; Kisel &
Barraclough 2010). Thus, plots of species richness against total area occupied may not
yield significant relationships unless the degree of fragmentation is also considered and
total area is scaled appropriately (see Orme et al. in prep). In addition, the dispersal
ability of a clade in combination with the geographic structure of the fragments will
influence the number of fragments that can be occupied. Finally, the effect of barriers
will depend on the average range sizes of species in a region: if the average range size is
small, barriers need not be large or bisect an entire region to cause speciation
(Rosenzweig 1975).
170
Appendix I.
Attesting to the importance of environmental features in the generation of SARs,
increased elevational range is associated with higher diversity in both geopolitical and
biotic regions; habitat diversity also drives higher diversity, but only in geopolitical
regions (Table 3, Figure S4). This arises from differences between the clustering
methods: areas with similar habitat are likely to be biotically homogenous and therefore
form a single biotic region, whereas political boundaries are more likely to cut across
such regions. As a result, Simpson’s index (1-D) of habitat diversity is low in biotic
clusters and scales extremely weakly with region area (intercept: 0.227, se = 0.042,
t=3.83; slope: 0.018, se = 0.014, t = 1.27; df = 148) whereas in TDWG regions it is
higher and scales strongly with area (intercept: 0.356, se = 0.025, t=14.39; slope: 0.055,
se = 0.005, t = 10.27; df = 578). In all these models, the high relative magnitude of the
standardized parameter estimate for area also implies it is not simply acting as a proxy for
either variable.
171
Appendix I.
Table 3. Multivariate regressions of SARs including a) habitat diversity, b) log range in
elevation, c) mean annual temperature and d) mean NDVI. The models are fitted to log
10 species richness within both geopolitical and biotic regions and the explanatory
covariates in all models are centred and standardized to facilitate model comparison (* p
< 0.05, ** p < 0.01, *** p < 0.001). The number of regions with available data is shown for
each model.
a)
b)
c)
d)
n
Intercept
Habitat diversity
log Area
Interaction
n
Intercept
log Elevation range
log Area
Interaction
n
Intercept
NDVI
log Area
Interaction
n
Intercept
Temperature
log Area
Interaction
a) Geopolitical regions
Estimate
SE
580
1.7419
0.0186
0.0234
0.0194
0.3407
0.0203
0.0410
0.0152
578
1.6810
0.0202
0.0533
0.0297
0.3956
0.0237
0.1114
0.0162
477
1.8201
0.0160
0.0142
0.0174
0.3075
0.0218
0.1007
0.0218
525
1.7463
0.0158
0.0577
0.0168
0.4110
0.0201
0.0979
0.0179
***
***
**
***
.
***
***
***
***
***
***
***
***
***
b) Biotic regions
Estimate
SE
150
1.2670 0.0342
-0.0649 0.0340
0.7608 0.0352
-0.0340 0.0325
200
1.0211 0.0391
0.2383 0.0573
0.5869 0.0478
0.1923 0.0343
130
1.0802 0.0496
0.1947 0.0521
0.8997 0.0536
-0.0442 0.0550
196
1.1339 0.0343
0.2750 0.0407
0.8411 0.0341
-0.0545 0.0432
***
.
***
***
***
***
***
***
***
***
***
***
***
172
Appendix I.
Section 3: Abiotic factors modulating the species-area relationship
Some abiotic factors, such as energy availability, do not correlate closely with area but
may still affect diversification rates or diversity limits of different regions, leading to
departures from SARs that depend on a region’s prevailing environmental conditions.
Energy availability is one of the key variables thought to contribute to large-scale spatial
patterns of diversity, and has mainly been discussed for its part in generating latitudinal
differences in diversity (reviewed in Willig et al. 2003; Mittelbach et al. 2007). On
average, energy availability (either ambient, e.g. temperature, or productive, e.g. plant
biomass) explains 60% of the variation in broad-scale richness across a range of plant and
animal groups (Hawkins et al. 2003). This variation should lead to consistent differences
between SARs of high- and low-energy regions.
As expected, increases in both mean annual temperature and mean NDVI act to
significantly elevate both overall mammal diversity and slopes of mammalian SARs
(Table 3, Figure S5). Again though, as in analyses including habitat and topographical
diversity, the relative magnitudes of standardized regression coefficients show that area is
the main driver of diversity within regions.
We expect energy to affect SARs through both diversification rates and diversity limits.
First, it could affect speciation rates through faster rates of molecular evolution, with
increased metabolic rates in higher-energy regions leading to both shorter generation
times and higher mutation rates (Rohde 1992). There has been mixed evidence for this
molecular rate hypothesis, with particularly weak support in endotherms (Cardillo et al.
173
Appendix I.
2005) and no support in angiosperms (although a direct effect of energy on species
richness is supported: Davies et al. 2004). However, Gillman et al. (2009) recently
presented evidence for higher rates of microevolution in tropical mammals and explained
this as an indirect consequence of more rapid co-evolution with other tropical ectotherms
(see also Fischer 1960; Schemske 2002). Energy is also expected to increase
diversification rates through effects on population dynamics, as aseasonal and elevated
productive energy can support larger populations, resulting in increased speciation and
reduced extinction, as described above. Such an aseasonal and high-energy environment
will also increase the equilibrium diversity limit by increasing resource availability,
facilitating specialisation to very narrow niches, and thus increasing the number of
distinct niches available (Janzen 1967). Conversely, seasonal habitats in temperate
regions may select for more motile, generalist species. These traits should decrease both
speciation rate and the number of species that can be supported in a region (Dynesius &
Jansson 2000; Sheldon 1996). Although not attempted here, incorporating ecological
covariates into our diversification models could lend insight into the effects of, for
example, energy availability on the diversification trajectory of clades in different regions
(Vamosi & Vamosi in press).
Section 4: Clade traits modulating the species-area relationship
So far our framework has considered species richness within a region as an outcome of
solely environmental and geographic influences, taking a neutral view of the organisms
themselves (MacArthur & Wilson 1967). However, there is abundant research (reviewed
in Coyne & Orr 2004) indicating that species traits affect clade diversity. Any clade traits
that affect diversity will give rise to clade-specific SARs, and create scatter around SARs
174
Appendix I.
that aggregate species richness across multiple clades. The effects of clade traits on SARs
are reflected in the clear differences between mammalian orders in the scaling of species
richness with area: order-specific slopes vary between -1.71 to 0.59 with medians of 0.16
for clustered regions and 0.11 for geopolitical regions (Table S3; because regions are not
nested, negative slopes arise simply where orders have high diversity in small regions).
According to our general model, clade traits can modulate SARs by modifying the net
rate of diversification (Fig 1a) and/or the diversity limit (Fig 1b). It is not straightforward
to assign traits to one of these mechanisms. Firstly, data are lacking: studies analysing
differences between clades in diversification (reviewed in: Jablonski 2008b; Rabosky &
McCune 2010) have not discriminated between effects on diversification rate and effects
on diversity limits (but see Vamosi & Vamosi in press), and studies of diversification
slowdowns in phylogenies (e.g. Phillimore & Price 2008) have not investigated the
influence of species’ traits. Secondly, individual traits are unlikely to act solely through
modification of either diversification rates or diversity limits (Mallet submitted). Finally,
many clade traits are strongly correlated (for example, geographic range size, dispersal
distance and body size: Jablonski 2008b; Jones et al. 2009) and so any traits acting
through one mechanism are likely to be associated with traits acting through the other.
Below, we discuss traits expected to influence SARs, with particular emphasis on those
that affect species’ use of space.
While most traits are likely to influence both diversification rates and diversity limits,
life history traits are perhaps the only class of traits expected to influence only
diversification rate. Typically, r selected species exhibit higher net rates of diversification
175
Appendix I.
than K selected species, and several mechanisms have been proposed to explain this
(Mayhew 2007, Marzluff & Dial 1991). Short generation times are associated with high
rates of population increase and the ability to rapidly exploit favourable conditions
(Mayhew 2007), conferring resilience to disturbance and leading to lower rates of
extinction. They are also associated with increased rates of evolution due to shorter
nucleotide generation times (Martin & Palumbi 1993; Mittelbach et al. 2007), and higher
metabolic rates (Martin & Palumbi 1993), both leading to higher rates of speciation. In
addition, the larger population sizes associated with r selection should increase speciation
rates and decrease extinction rates, as discussed in Section 2.
Clade traits that determine how space is occupied within a region also affect both the
generation and maintenance of SARs. Larger species ranges are associated with lower
clade diversity limits as well as reduced rates of extinction (e.g. Jablonski 2008a; Payne
& Finnegan 2007), and increased rates of speciation (Phillimore et al. 2006, but see
Jablonski & Roy 2003). Regarding diversity limits, there is evidence from both mammals
(Orme et al. in prep) and birds (Phillimore et al. 2008) that increasing species’ range
overlap is a stronger predictor of increased species richness than decreased median range
size.
Similarly to species’ range size, several aspects of narrow niche breadth, such as
ecological specialisation, high host specificity and narrow environmental tolerances, have
been associated with increased diversity limits as well as increased rates of extinction and
speciation (Jablonski 2008b). Increased clade diversity is also associated with greater
niche overlap rather than decreased niche breadth.
176
Appendix I.
Ricklefs (2009) has shown that South American bird families of varying species richness
do not differ in the average number of habitats occupied by species, suggesting that niche
overlap increases. We find the same in mammals, using simple measures of the number
of habitats used by species. There is no significant correlation between the richness of
mammalian families and either the average total number of habitats occupied
(tau = -0.076, p=0.21) or the average number of major habitats occupied (tau = -0.043,
p=0.49) nor is there a decrease in mean species range size with increasing family
richness (tau = -0.001, p=0.99).
Finally, increased dispersal ability has been found to reduce speciation and extinction
rates in some cases (Xiang et al. 2004), while in others it has been shown to increase
diversification rate (Phillimore et al. 2006; Phillimore & Price 2009). With respect to
diversity limits, high dispersal ability may lead to low equilibrium diversity within a
region if it leads to clades consisting of few species with large ranges. At the other
extreme, strong philopatry, where individuals retain or return to natal locations, might
both increase rates of diversification by accelerating rates of genetic differentiation
(Peterson 2008) and increase equilibrium diversity by impeding range expansion and
boosting the number of equivalent species that can persist in a region (Shmida & Wilson
1985; Seehausen 2006). Alternatively, high dispersal ability can increase the rate at
which new regions are occupied, increasing clade richness through occupation of
multiple regions. Such long-distance dispersal may significantly distort SARs if newly
colonised regions harbour clades with higher diversity due to competitive release.
Section 5: Historical and temporal effects on the species area relationship
177
Appendix I.
SARs will be clearest when clades have reached equilibrium throughout their ranges, but
this requires that they have had enough time to diversify to their limit in each region that
they occupy. Thus, in parts of the world where the current habitat has only recently
become available, current diversity is likely to be lower than expected (e.g. a recentlyformed island, Esselstyn et al. 2009, or a recently deglaciated region, Pielou 1979) and
may be biased toward large-ranged generalists (Dynesius & Jansson 2000). In contrast, a
comparison of mammalian sister taxon pairs with disjunct distributions across two realms
indicated that sisters remaining in the realm unambiguously reconstructed as ancestral
(DIVA: Ronquist 1997) are significantly less species rich (12 out of 41, binomial p =
0.004 Table S4) than sisters that dispersed. This suggests a diversification burst in newly
colonized regions, driven by competitive release. Finally, if a region is subject to frequent
extrinsic perturbations (such as an archipelago subject to repeated sea-level changes),
fluctuating extinction rates make it unlikely that equilibrium diversity will ever be
reached or maintained (Whittaker et al. 2008; Esselstyn et al. 2009). Indeed, explanations
for high tropical diversity, such as the time-for-speciation effect (Stephens & Wiens
2003) and reduced extinction due to long-term climatic stability (Fischer 1960), are
compatible with tropical regions being able to more closely approach diversity limits.
Diversity may also transiently over- or under-shoot the diversity limit of a region if
speciation or extinction occurs very rapidly, or if perturbations occur that suddenly alter
clade diversity limits (Gavrilets & Vose 2005). Alternatively, non-ecological modes of
speciation (e.g. via sexual selection or polyploidy), may produce transient species that are
unable to persist in the long-term given the niche space available, and thus are committed
to eventual extinction (McPeek 2008; Rosenzweig 1995, Chesson 2000). This may also
178
Appendix I.
apply to ecologically equivalent species formed in allopatry, if the barriers separating
them are themselves transient. Transient dynamics are now thought to be crucial in
predicting biodiversity responses to current global change (recently reviewed in Jackson
& Sax 2010); though the changes will likely not be as immediately apparent as for
ecological processes such as community assembly, evolutionary clade dynamics will
certainly be affected as well (Rosenzweig 2001).
Conclusions
We have presented a framework, based on a diversity-dependent model of clade
diversification, for understanding how evolutionary processes contribute to the creation
of large-scale SARs. This framework is supported by analyses on mammals using data
from the PanTHERIA database (Jones et al. 2009). SARs themselves result from direct
and positive effects of area on diversification rates and diversity limits, as well as indirect
effects of area through population size, habitat diversity, and habitat fragmentation. We
found that these effects are apparent in the histories of mammal diversification – clades
occupying larger areas had higher initial diversification rates and lower rates of decline in
diversification. We also confirmed that habitat and topographical diversity are significant
predictors of regional diversity in mammals, but found that neither is a proxy for area the most predictive models of diversity always include area as well. Environmental
factors and clade traits that are not tightly correlated with area also cause systematic
differences in SARs between clades or regions, and cause scatter around any general
SAR generated without accounting for them. We tested the influence of energy
availability on mammal diversity and showed that high energy availability significantly
increases the slopes and intercepts of SARs. In addition, mammal orders vary greatly in
179
Appendix I.
the slopes of their SARs. Finally, we provide evidence that historical contingencies
impact SARs, demonstrating that mammal clades able to colonize new, competitor-free
regions are more diverse than their stay-at-home sisters.
Schoener (1976) referred to the species-area relationship as the phenomenon closest to
attaining rule status in ecology, and SARs are indeed one of the most general diversity
patterns, existing for a wide range of organisms across a range of spatial scales. However,
we argue here that in addition to the processes most discussed in the ecological literature
– immigration, local extinction and species coexistence - SARs are also influenced by
macroevolutionary processes, in particular speciation and global extinction. None of
these processes operates in isolation, and every SAR is the result of interplay between
both ecological and evolutionary processes. Diversity limits, for instance, must ultimately
result from ecological limits on the number of species that can coexist in a region, though
the speed at which they are reached may depend on evolutionary processes. We suggest
that a full understanding of species-area relationships will require integrating both
ecological and evolutionary perspectives on the processes that generate and constrain
diversity.
180
Appendix I.
Acknowledgements
We thank Kate Jones for the invitation to contribute to this special issue and Tim
Barraclough, Natalie Cooper, Susanne Fritz, Alex Pigot and James Rosindell for
comments on previous versions of the manuscript. Special thanks are extended to Ally
Phillimore for insightful discussion and comments, and R code for implementing the
diversification models. YK was supported by a U.S. National Science Foundation
Graduate Research Fellowship and a Deputy Rector’s Award from Imperial College
London, LM by a Grantham Institute studentship, and CDLO by an RCUK fellowship.
181
Appendix I.
References
Alroy, J. 1998 Equilibrial diversity dynamics in North American mammals. In
Biodiversity Dynamics: Turnover of Populations, Taxa and Communities (ed. M. L.
McKinney & J. A. Drake). New York: Columbia University Press.
Arrhenius, O. 1921 Species and area. Journal of Ecology 9, 95-99.
Baldi, A. 2008 Habitat heterogeneity overrides the species-area relationship. Journal of
Biogeography 35, 675-681.
Bininda-Emonds, O. R. P., Cardillo, M., Jones, K. E., MacPhee, R. D. E., Beck, R. M. D.,
Grenyer, R., Price, S. A., Vos, R. A., Gittleman, J. L. & Purvis, A. 2007 The delayed
rise of present-day mammals. Nature 446, 507-512.
Bininda-Emonds, O. R. P., Cardillo, M., Jones, K. E., MacPhee, R. D. E., Beck, R. M. D.,
Grenyer, R., Price, S. A., Vos, R. A., Gittleman, J. L. & Purvis, A. 2008 The delayed
rise of present-day mammals (vol 446, pg 507, 2007). Nature 456, 274-274.
Bokma, F. 2003 Testing for equal rates of cladogenesis in diverse taxa. Evolution 57,
2469-2474.
Brown, J. H. 1995 Macroecology. Chicago: University of Chicago Press.
Brummit, R. K. 2001 World Geographical Scheme for Recording Plant Distributions.
Pittsburgh, Pennsylvania: Hunt Institute for Botanical Documentation, CarnegieMellon University (for the International Working Group on Taxonomic Databases for
Plant Sciences).
Cardillo, M., Orme, C. D. L. & Owens, I. P. F. 2005 Testing for latitudinal bias in
diversification rates: An example using New World birds. Ecology 86, 2278-2287.
Chesson, P. 2000 Mechanisms of maintenance of species diversity. Annual Review of
Ecology and Systematics 31, 343-366.
Chown, S. L. & Gaston, K. J. 2000 Areas, cradles and museums: the latitudinal gradient
in species richness. Trends in Ecology & Evolution 15, 311-315.
Coyne, J. A. & Orr, H. A. 2004 Speciation. Sunderland, Massachusetts: Sinauer
Associates, Inc.
Davies, T. J., Savolainen, V., Chase, M. W., Moat, J. & Barraclough, T. G. 2004
Environmental energy and evolutionary rates in flowering plants. Proceedings of the
Royal Society of London Series B-Biological Sciences 271, 2195-2200.
182
Appendix I.
Drakare, S., Lennon, J. J. & Hillebrand, H. 2006 The imprint of the geographical,
evolutionary and ecological context on species–area relationships. Ecology Letters 9,
215-227.
Dynesius, M. & Jansson, R. 2000 Evolutionary consequences of changes in species'
geographical distributions driven by Milankovitch climate oscillations. Proceedings
of the National Academy of Sciences of the United States of America 97, 9115-9120.
Esselstyn, J. A., Timm, R. M. & Brown, R. M. 2009 Do geological or climatic processes
drive speciation in dynamic archipelagos? The tempo and mode of diversification in
southeast Asian shrews. Evolution 63, 2595-2610.
Fischer, A. G. 1960 Latitudinal variations in organic diversity. Evolution 14, 64-81.
Frankham, R. 1996 Relationship of genetic variation to population size in wildlife.
Conservation Biology 10, 1500-1508.
Gavrilets, S. & Vose, A. 2005 Dynamic patterns of adaptive radiation. Proceedings of the
National Academy of Sciences of the United States of America 102, 18040-18045.
Gillman, L. N., Keeling, D. J., Ross, H. A. & Wright, S. D. 2009 Latitude, elevation and
the tempo of molecular evolution in mammals. Proceedings of the Royal Society BBiological Sciences 276, 3353-3359.
Gilpin, M. E. & Soulé, M. E. 1986 Minimum viable populations: processes of species
extinction. In Conservation biology: the science of scarcity and diversity (ed. M. E.
Soule), pp. 19-34. Sunderland, Massachusetts: Sinauer Associate, Inc.
Hawkins, B. A., Field, R., Cornell, H. V., Currie, D. J., Guegan, J. F., Kaufman, D. M.,
Kerr, J. T., Mittelbach, G. G., Oberdorff, T., O'Brien, E. M., Porter, E. E. & Turner, J.
R. G. 2003 Energy, water, and broad-scale geographic patterns of species richness.
Ecology 84, 3105-3117.
Hubbell, S. P. 2001 The Unified Neutral Theory of Biodiversity and Biogeography.
Princeton: Princeton University Press.
Hutchinson, G. E. & MacArthur, R. H. 1959 A theoretical ecological model of size
distributions among species of animals. The American Naturalist 93, 117.
Jablonski, D. 2008a Extinction and the spatial dynamics of biodiversity. Proceedings of
the National Academy of Sciences of the United States of America 105, 11528-11535.
183
Appendix I.
Jablonski, D. 2008b Species selection: theory and data. Annual Review of Ecology
Evolution and Systematics 39, 501-524.
Jablonski, D. & Roy, K. 2003 Geographical range and speciation in fossil and living
molluscs. Proceedings of the Royal Society B-Biological Sciences 270, 401-406.
Jackson, S. T. & Sax, D. F. 2010 Balancing biodiversity in a changing environment:
extinction debt, immigration credit and species turnover. Trends in Ecology &
Evolution 25, 153-160.
Janzen, D. 1967 Why mountain passes are higher in the tropics. American Naturalist 101,
233-249
Jones, K. E., Bielby, J., Cardillo, M., Fritz, S. A., O'Dell, J., Orme, C. D. L., Safi, K.,
Sechrest, W., Boakes, E. H., Carbone, C., Connolly, C., Cutts, M. J., Foster, J. K.,
Grenyer, R., Habib, M., Plaster, C. A., Price, S. A., Rigby, E. A., Rist, J., Teacher, A.,
Bininda-Emonds, O. R. P., Gittleman, J. L., Mace, G. M., Purvis, A. & Michener, W.
K. 2009 PanTHERIA: a species-level database of life history, ecology, and geography
of extant and recently extinct mammals. Ecology 90, 2648-2648.
Kallimanis, A. S., Mazaris, A. D., Tzanopoulos, J., Halley, J. M., Pantis, J. D. &
Sgardelis, S. P. 2008 How does habitat diversity affect the species-area relationship?
Global Ecology and Biogeography 17, 532-538.
Kisel, Y. & Barraclough, T. G. 2010 Speciation has a spatial scale that depends on levels
of gene flow. American Naturalist 175, 316-334.
Kreft, H. & Jetz, W. A framework for delineating biogeographical regions based on
species distributions. Journal of Biogeography, in press.
Lande, R. 1993 Risks of population extinction from demographic and environmental
stochasticity and random catastrophes. American Naturalist 142, 911-927.
Leimu, R., Mutikainen, P., Koricheva, J. & Fischer, M. 2006 How general are positive
relationships between plant population size, fitness and genetic variation? Journal of
Ecology 94, 942-952.
Linder, H. P. 2008 Plant species radiations: where, when, why? Philosophical
Transactions of the Royal Society B-Biological Sciences 363, 3097-3105.
184
Appendix I.
Linder, H. P., Lovett, J. C., Mutke, J., Barthlott, W., Jürgens, N., Rebelo, T. & Küper, W.
2005 A numerical re-evaluation of the sub-Saharan phytochoria of mainland Africa.
Biologiske Skrifter 55, 229-252.
Lomolino, M. V. 2000 Ecology's most general, yet protean pattern: the species-area
relationship. Journal of Biogeography 27, 17-26.
Los, S. O., Collatz, G. J., Sellers, P. J., Malmström, C. M., Pollack, N. H., DeFries, R. S.,
Bounoua, L., Parris, M. T., Tucker, C. J. & Dazlich, D. A. 2000 A global 9-year
biophysical land-surface data set from NOAA AVHRR data. Journal of
Hydrometerology 1, 183-199.
Losos, J. B. & Parent, C. E. 2010 The speciation-area relationship. In The Theory of
Island Biogeography Revisited (ed. J. B. Losos & R. E. Ricklefs), pp. 415-438.
Oxford: Princeton University Press.
Losos, J. B. & Schluter, D. 2000 Analysis of an evolutionary species-area relationship.
Nature 408, 847-850.
MacArthur, R. H. & Wilson, E. O. 1967 The Theory of Island Biogeography. Princeton:
Princeton University Press.
Mallet, J. The 'struggle for existence:' why the mismatch of theory in ecology and
evolution? . American Naturalist, submitted.
Martin, A. P. & Palumbi, S. R. 1993 Body size, metabolic rate, generation time and the
molecular clock. Proceedings of the National Academy of Sciences of the United
States of America 90, 4087-4091.
Marzluff, J. M. & Dial, K. P. 1991 Life history correlates of taxonomic diversity. Ecology
72, 428-439.
Maurer, B. A. & Nott, M. P. 1998 Geographic range fragmentation and the evolution of
biological diversity. In Biodiversity Dynamics: Turnover of Populations, Taxa and
Communities (ed. M. L. McKinney & J. A. Drake). New York: Columbia University
Press.
Mayhew, P. J. 2007 Why are there so many insect species? Perspectives from fossils and
phylogenies. Biological Reviews 82, 425-454.
185
Appendix I.
McInnes, L., Purvis, A. & Orme, C. D. L. 2009 Where do species' geographic ranges stop
and why? Landscape impermeability and the Afrotropical avifauna. Proceedings of
the Royal Society B-Biological Sciences 276, 3063-3070.
McKinney, M. L. 1998 Biodiversity dynamics: niche preemption and saturation in
diversity equilibria. In Biodiversity Dynamics: Turnover of Populations, Taxa and
Communities (ed. M. L. McKinney & J. A. Drake). New York: Columbia University
Press.
McPeek, M. A. 2008 The ecological dynamics of clade diversification and community
assembly. American Naturalist 172, E270-E284.
Mittelbach, G. G., Schemske, D. W., Cornell, H. V., Allen, A. P., Brown, J. M., Bush, M.
B., Harrison, S. P., Hurlbert, A. H., Knowlton, N., Lessios, H. A., McCain, C. M.,
McCune, A. R., McDade, L. A., McPeek, M. A., Near, T. J., Price, T. D., Ricklefs, R.
E., Roy, K., Sax, D. F., Schluter, D., Sobel, J. M. & Turelli, M. 2007 Evolution and
the latitudinal diversity gradient: speciation, extinction and biogeography. Ecology
Letters 10, 315-331.
Nee, S., Mooers, A. & Harvey, P. 1992 Tempo and mode of evolution revealed from
molecular phylogenies. Proceedings of the National Academy of Sciences of the
United States of America 89, 8322 - 8326.
Pagel, M. D., May, R. M. & Collie, A. R. 1991 Ecological aspects of the geographical
distribution and diversity of mammalian species. American Naturalist 137, 791-815.
Palmer, M. W. & White, P. S. 1994 Scale dependence and the species-area relationship.
American Naturalist 144, 717-740.
Payne, J. L. & Finnegan, S. 2007 The effect of geographic range on extinction risk during
background and mass extinction. Proceedings of the National Academy of Sciences of
the United States of America 104, 10506-10511.
Peterson, A. T. 2008 Philopatry and genetic differentiation in the Aphelocoma jays
(Corvidae). Biological Journal of the Linnean Society 47, 249-260.
Phillimore, A. B. 2010 Subspecies origination and extinction in birds. Ornithological
Monographs 67, 42-53.
186
Appendix I.
Phillimore, A. B., Freckleton, R. P., Orme, C. D. L. & Owens, I. P. F. 2006 Ecology
predicts large-scale patterns of phylogenetic diversification in birds. American
Naturalist 168, 220-229.
Phillimore, A. B., Orme, C. D. L., Thomas, G. H., Blackburn, T. M., Bennett, P. M.,
Gaston, K. J. & Owens, I. P. F. 2008 Sympatric speciation in birds is rare: Insights
from range data and simulations. American Naturalist 171, 646-657.
Phillimore, A. B. & Price, T. D. 2008 Density-dependent cladogenesis in birds. Plos
Biology 6, 483-489.
Phillimore, A. B. & Price, T. D. 2009 Ecological influences on the temporal pattern of
speciation. In Speciation and Patterns of Diversity (ed. R. Butlin, J. R. Bridle & D.
Schluter). Cambridge: Cambridge University Press.
Pielou, E. C. 1979 Biogeography. New York: Wiley.
Pigot, A. L., Phillimore, A. B., Owens, I. P. F. & Orme, C. D. L. The shape and temporal
dynamics of phylogenetic trees arising from geographic speciation. Systematic
Biology, in press.
Preston, F. W. 1960 Time and space and the variation of species. Ecology 41, 612-627.
Rabosky, D. L. 2009a Ecological limits and diversification rate: alternative paradigms to
explain the variation in species richness among clades and regions. Ecology Letters
12, 735-743.
Rabosky, D. L. 2009b Ecological limits on clade diversification in higher taxa. American
Naturalist 173, 662-674.
Rabosky, D. L. & Lovette, I. J. 2008 Density-dependent diversification in North
American wood warblers. Proceedings of the Royal Society B-Biological Sciences
275, 2363-2371.
Rabosky, D. L. & McCune, A. R. 2010 Reinventing species selection with molecular
phylogenies. Trends in Ecology & Evolution 25, 68-74.
Rahbek, C. & Graves, G. R. 2001 Multiscale assessment of patterns of avian species
richness. Proceedings of the National Academy of Sciences of the United States of
America 98, 4534-4539.
187
Appendix I.
Ricklefs, R. E. 2009 Speciation, extinction and diversity. In Speciation and Patterns of
Diversity (ed. R. Butlin, J. R. Bridle & D. Schluter). Cambridge: Cambridge
University Press.
Ricklefs, R. E. & Lovette, I. J. 1999 The roles of island area per se and habitat diversity in
the species-area relationships of four Lesser Antillean faunal groups. Journal of
Animal Ecology 68, 1142-1160.
Rohde, K. 1992 Latitudinal gradients in species diversity - the search for the primary
cause. Oikos 65, 514-527.
Ronquist, F. 1997 Dispersal-vicariance analysis: A new approach to the quantification of
historical biogeography. Systematic Biology 46, 195-203.
Rosenzweig, M. L. 1975 On continental steady states of species diversity. In The Ecology
and Evolution of Communities (ed. M. Cody & J. Diamond), pp. 121-140. Cambridge,
MA. : Harvard University Press.
Rosenzweig, M. L. 1995 Species Diversity in Space and Time. Cambridge: Cambridge
University Press.
Rosenzweig, M. L. 1998 Preston's ergodic conjecture: the accumulation of species in
space and time. In Biodiversity Dynamics: Turnover of Populations, Taxa and
Communities (ed. M. L. McKinney & J. A. Drake). New York: Columbia University
Press.
Rosenzweig, M. L. 2001 Loss of speciation rate will impoverish future diversity.
Proceedings of the National Academy of Sciences of the United States of America 98,
5404-5410.
Scheiner, S. M. 2003 Six types of species-area curves. Global Ecology and Biogeography
12, 441-447.
Schemske, D. W. 2002 Tropical diversity: patterns and processes. In Ecological and
evolutionary perspectives on the origins of tropical diversity: key papers and
commentaries (ed. R. Chazdon & T. Whitmore), pp. 163-173. Chicago: University of
Chicago Press.
Schielzeth, H. Simple means to improve the interpretability of regression coefficients.
Methods in Ecology and Evolution, in press.
188
Appendix I.
Schluter, D. 2009 Evidence for ecological speciation and Its alternative. Science 323,
737-741.
Schluter, D. & Conte, G. L. 2009 Genetics and ecological speciation. Proceedings of the
National Academy of Sciences of the United States of America 106, 9955-9962.
Schoener, T. W. 1976 The species-area relationship within archipelagoes: models and
evidence from island birds. Proceedings of XVI International Ornithological
Congress 6, 629-642.
Seehausen, O. 2006 African cichlid fish: a model system in adaptive radiation research.
Proceedings of the Royal Society B: Biological Sciences 273, 1987-1998.
Sepkoski, J. J. 1978 Kinetic model of Phanerozoic taxonomic diversity, 1. Analysis of
marine orders. . Paleobiology 4, 223-251.
Sepkoski, J. J., Jr. 1976 Species diversity in the Phanerozoic: species-area effects.
Paleobiology 2, 298-303.
Sheldon, P. R. 1996 Plus ca change - A model for stasis and evolution in different
environments. Palaeogeography Palaeoclimatology Palaeoecology 127, 209-227.
Shmida, A. & Wilson, M. V. 1985 Biological determinants of species diversity. Journal
of Biogeography 12, 1-20.
Slatkin, M. 1973 Gene flow and selection in a cline. Genetics 75, 733-756.
Slatkin, M. 1985 Gene flow in natural populations. Annual Review of Ecology and
Systematics 16, 393-430.
Slatkin, M. 1987 Gene flow and the geographic structure of natural populations. Science
236, 787-792.
Stephens, P. R. & Wiens, J. J. 2003 Explaining species richness from continents to
communities: The time-for-speciation effect in emydid turtles. American Naturalist
161, 112-128.
Triantis, K. A., Mylonas, M., Lika, K. & Vardinoyannis, K. 2003 A model for the
species-area-habitat relationship. Journal of Biogeography 30, 19-27.
Vamosi, J. C. & Vamosi, S. M. Key innovations within a geographical context in
flowering plants: towards resolving Darwin’s abominable mystery. Ecology Letters,
in press.
189
Appendix I.
Weber, K. E. 1990 Increased selection response in larger populations - 1. selection for
wing-tip height in Drosophila melanogaster at 3 population sizes. Genetics 125, 579584.
Whittaker, R. J., Triantis, K. A. & Ladle, R. J. 2008 A general dynamic theory of oceanic
island biogeography. Journal of Biogeography 35, 977-994.
Wiley, J. W. & Wunderle, J. M. 1994 The effects of hurricanes on birds, with special
reference to Caribbean islands. Bird Conservation International 3, 319-349.
Willi, Y., Van Buskirk, J. & Hoffmann, A. A. 2006 Limits to the adaptive potential of
small populations. Annual Review of Ecology Evolution and Systematics 37, 433-458.
Willig, M., Kaufmann, D. & Stevens, R. 2003 Latitudinal gradients of biodiversity:
pattern, process, scale, and synthesis. Annual Review of Ecology and Systematics 34,
273 - 309.
Xiang, Q. Y. J., Zhang, W. H., Ricklefs, R. E., Qian, H., Chen, Z. D., Wen, J. & Li, J. H.
2004 Regional differences in rates of plant speciation and molecular evolution: A
comparison between eastern Asia and eastern North America. Evolution 58, 21752184.
190
Appendix I.
Supplementary information
Table S1. Parameters of linear regression models of log species richness as a function of
log region area for two region types at four scales and for all, endemic and non-endemic
species within each region.
all
endemic
nonendemic
int
se
p
slope
se
p
int
se
p
slope
se
p
int
se
p
slope
se
p
Clustering
TDWG
50
100
150
200
L1
L2
-0.855 -0.653 -0.617 -0.488 -0.591 -0.273
0.15 0.108 0.086 0.081 0.089 0.083
***
0.425
0.036
***
***
0.427
0.028
***
***
0.429
0.02
***
***
0.412
0.019
***
***
0.474
0.024
***
**
0.4
0.019
***
L3
0.197
0.072
**
0.329
0.015
***
-0.818 -0.915 -0.709 -0.555 -1.009 -0.762 -0.388
0.309 0.168 0.148
0.15
0.09 0.089
0.09
*
***
0.365
0.062
***
0.397
0.037
***
***
0.326
0.030
***
***
0.26
0.030
***
***
0.39
0.024
***
***
0.306
0.021
***
-0.849 -0.609 -0.610 -0.477 -0.454 -0.166
0.150 0.112 0.090 0.086 0.117
0.1
***
0.406
0.034
***
***
0.402
0.028
***
***
0.416
0.021
***
***
0.402
0.019
***
***
0.421
0.029
***
.
0.368
0.022
***
***
0.168
0.019
***
0.37
0.08
***
0.293
0.017
***
L4
0.721
0.059
***
0.237
0.013
***
-0.329
0.084
***
0.142
0.018
***
0.898
0.06
***
0.201
0.013
***
191
192
A. 20MY
1
2
3
4
5
6
B. 10MY
1
2
3
4
5
6
--0.040
--0.040
--0.040
--0.031
--0.041
--0.031
0.265
-0.251
0.530
-0.333
0.377
-0.197
c
[-0.2,0.2]
0.340
-0.320
0.790
-0.330
0.478
-0.260
Lambda
[-1,1]
-----0.043
-0.046
-0.233
-0.115
-----0.030
-0.030
-0.200
-0.100
--------0.011
0.004
--------0.011
0.004
0.999
0.990
0.999
0.581
0.712
0.500
0.990
0.990
0.990
0.560
0.908
0.650
z
p
Epsilon
[-0.2,0.2] [-0.2,0.2] [0.5,0.999]
at zero. The overall best-fit model for each period is shown in bold.
-160.5
-86.1
-106.9
-24.4
0.6
0.0
196.3
119.1
161.2
40.9
11.5
0.0
ΔAICc
-1578.3
-1540.1
-1550.5
-1508.2
-1495.7
-1495.0
-1412.7
-1373.1
-1394.2
-1333.0
-1318.3
-1311.5
Likelihood
maximum likelihood estimate of the model is reported for each free parameter within the bounds shown. Dashed parameter estimates were fixed
region area (2), exponential decline (3), exponential decline region area scaling initial rate (4), rate of decline (5) or both (6). In each case, the
present using the biotic regions (200) classification. Six models of diversification are fitted representing: constant rate (1), constant rate scaled by
Table S2 - Summary of diversification models fitted to mammalian clades with crown age younger than (a) 20 and (b) 10 million years before
Appendix I.
Table S3. Slopes of SARs for mammalian orders for both geopolitical and biotic regions.
Afrosoricida
Artiodactyla
Carnivora
Chiroptera
Cingulata
Dasyuromorphia
Dermoptera
Didelphimorphia
Diprotodontia
Erinaceomorpha
Hyracoidea
Lagomorpha
Macroscelidea
Microbiotheria
Monotremata
Notoryctemorphia
Paucituberculata
Peramelemorphia
Perissodactyla
Pholidota
Pilosa
Primates
Proboscidea
Rodentia
Scandentia
Soricomorpha
Tubulidentata
a) TDWG
Slope
SE
N
-0.015
0.112
0.106
0.131
0.183
0.311
0.000
0.127
0.302
0.035
-0.006
0.079
0.053
0.000
0.001
0.623
0.099
0.174
0.022
0.037
0.212
0.120
0.075
0.209
0.036
0.111
0.000
0.077
0.013
0.009
0.014
0.040
0.047
0.000
0.031
0.058
0.010
0.021
0.010
0.028
0.000
0.049
0.546
0.203
0.035
0.020
0.011
0.032
0.023
0.018
0.011
0.023
0.013
0.000
37
482
498
557
129
11
16
160
26
199
72
416
39
5
12
3
5
17
142
132
70
230
73
547
63
406
53
b) Clustering
Slope
-0.733
0.304
0.304
0.285
0.111
0.372
0.000
0.091
0.334
0.115
0.130
0.175
-0.685
0.000
0.096
0.396
0.388
0.127
0.240
0.079
0.041
0.217
0.182
0.403
0.038
0.274
0.000
SE
N
0.382
0.033
0.027
0.021
0.068
0.057
0.000
0.066
0.046
0.071
0.067
0.042
0.681
0.000
0.069
0.268
0.067
0.054
0.074
0.030
0.042
0.043
0.079
0.025
0.040
0.033
0.000
5
79
92
167
18
15
7
19
31
30
9
57
5
3
12
4
4
18
30
23
9
51
11
132
24
65
6
193
Appendix I.
Table S4. Sister clade pairs identified using DIVA that show unambiguous reconstructed
ancestral ranges and where one sister clade retains that ancestral range and the other
occupies a differing new range. Italicized rows show clades with higher species richness
in the new range.
Family
Bovidae
Bovidae
Bovidae
Canidae
Canidae
Canidae
Cercopethicidae
Cercopethicidae
Cercopethicidae
Cervidae
Cervidae
Emballonuridae
Emballonuridae
Equidae
Erinaceidae
Heteromyidae
Leporidae
Loridae
Manidae
Molossidae
Molossidae
Molossidae
Muridae
Muridae
Mustelidae
Mustelidae
Mustelidae
Mustelidae
Myoxidae
Ochotonidae
Pteropodidae
Pteropodidae
Rhinolophidae
Rhinolophidae
Rhinolophidae
Sciuridae
Sciuridae
Tapiridae
Vespertilionidae
Viverridae
Viverridae
Mammal
Tree Node
-1008
-940
-973
-1217
-1225
-1226
-780
-790
-801
-1029
-1023
-1429
-1417
-1116
-1788
-891
-878
-909
-1328
-1513
-1524
-1538
-68
-881
-1138
-1159
-1164
-1166
-670
-783
-1342
-1357
-1474
-1476
-1491
-503
-617
-1111
-1631
-1310
-1312
Ancestral
Area
IndoMalay
Afrotropics
Afrotropics
Palearctic
Palearctic
Palearctic
Afrotropics
IndoMalay
Afrotropics
Palearctic
Palearctic
Australasia
Australasia
Palearctic
Palearctic
Nearctic
Afrotropics
Afrotropics
Afrotropics
Afrotropics
Neotropics
Neotropics
IndoMalay
Afrotropics
Palearctic
IndoMalay
IndoMalay
IndoMalay
Palearctic
Palearctic
Australasia
Australasia
IndoMalay
Australasia
Afrotropics
Nearctic
IndoMalay
IndoMalay
IndoMalay
Afrotropics
Afrotropics
New Area
Nearctic, Palearctic
Palearctic
Palearctic
Neotropics
Afrotropics
Nearctic
IndoMalay
Australasia
IndoMalay
Neotropics
IndoMalay
Neotropics
Afrotropics
Afrotropics
IndoMalay
Neotropics
Nearctic
IndoMalay
IndoMalay
Neotropics
Australasia, IndoMalay
Afrotropics
Palearctic
IndoMalay
Neotropics
Neotropics
Afrotropics
Neotropics
Afrotropics
Nearctic
IndoMalay
Afrotropics
Australasia
Afrotropics
Australasia
Palearctic
Nearctic
Neotropics
Palearctic
IndoMalay
IndoMalay
# Spp.
Ancestral Area
1
12
4
1
9
9
8
2
9
1
5
7
7
2
11
4
6
3
3
2
15
15
2
11
3
2
1
4
2
2
14
60
3
1
2
2
4
1
5
1
1
# Spp.
New Area
2
22
3
9
1
2
11
4
23
14
18
18
2
3
6
31
15
3
4
3
4
22
20
4
2
3
2
8
14
2
21
28
14
2
5
11
2
3
7
2
5
194
Appendix I.
Figure S1. Geographic distributions of biotic clusters defined by between-cell Jaccard
distances (a – 50, b – 100, c –150 and d – 200 clusters).
195
Appendix I.
Figure S2. Size distributions of geopolitical (TDWG) and clustered biotic regions used to
measure species-area relationships.
196
Appendix I.
Figure S3. Species-area relationships for geopolitical regions at four spatial scales from
TDWG Level 1 (A) to TDWG Level 4 (D) for all (black circles and line, SA), widespread
(grey crosses and dashed grey line, SW) and endemic (grey dots and grey solid line, SE)
species in each region. See also Table S1.
197
Appendix I.
Figure S4. Species-area relationships for clustered biotic regions at four spatial scales
from 50 (A) to 200 (D) clusters for all (black circles and line, SA), widespread (grey
crosses and dashed grey line, SW) and endemic (grey dots and grey solid line, SE) species
in each region. See also Table S1.
198
Appendix I.
Figure S5. Prediction surfaces from models of log species richness. Four variables
(Simpson’s diversity index in habitat diversity, log elevational range, mean NDVI and
mean temperature) with regions were fitted in turn as a covariate with regional area with
the model including the interaction between each pair. The coloured surface shows the
predicted diversity (white – high, red – low) and the relative size of the points show the
observed diversity. Model coefficients were estimated using scaled and centred covariates
(bottom and left axes) but these plots also show the variables on their original scale (top
and right axes).
199
Appendix II.
Appendix II. Supplementary figures and tables for chapter
three
Appendix Table II.1. Sampling locations for all genotyped species. ..............................201
Appendix Table II.2. Scoring analysis details for AFLP datasets. ..................................207
Appendix Table II.3. Branch circumference mean and variance for all study species. ..208
Appendix Table II.4. Ecological characteristics of study clades. ...................................208
Appendix Figure II.1. Associations of Fst and species range size with species phylogeny
(data from the reduced dataset). ......................................................................................209
200
201
L. ciliisepala
T. triglochin
M. rafaeliana
La Paz
M. nidifica
Cerro Caraigres
forest near Navarro
Cerro Caraigres
road near CATIE station
Cerro de la Muerte pasture
road to Taus
Bosque de Paz reserve potrero
forest near Orosi
Bosque de Paz reserve
Location
Species
Table II.1. Sampling locations for all genotyped species.
9.717 N /
84.12 W
9.72 N /
84.12 W
9.757 N /
83.898 W
9.643 N /
83.846 W
9.55 N /
83.68 W
10.165 N /
84.549 W
10.208 N /
84.324 W
9.770 N /
94.5492 W
10.192 N /
84.321 W
9.783 N /
83.764 W
Latitude /
Longitude
montane rain forest
montane rain forest
2100
2160
montane rain forest
pasture in matrix of cloud
forest
cloud/oak forest
roadside, between rain forest
and pastures
pasture
roadside in humid forest
cloud forest
roadside among pastures
2160
2830
2770
1250
1560
1520
1670
1160
approximate
elevation (m) Habitat
27
20
50
16
14
1
5
26
1
1
# samples
analysed
202
Cachi finca
L. elata
P. propinqua
L. floripecten
Freddy's finca
L. ciliisepala
forest near Navarro
Bosque de Paz reserve
Taus
road to Taus
road to P. N. Tapantí
forest near Orosi
cafetal near Tapantí
km 43 marker on Interamerican
Highway
forest near Navarro
Location
Species
10.208 N /
84.324 W
9.757 N /
83.898 W
9.770 N /
83.829 W
9.770 N /
83.905 W
9.768 N /
83.799 W
9.785 N /
83.763 W
9.781 N /
83.720 W
9.812 N /
83.777 W
9.757 N /
83.978 W
9.757 N /
83.898 W
9.808 N /
83.995 W
Latitude /
Longitude
2100
1670
760
1250
1210
1520
1165
1900-2100
1930
1830
1960
montane rain forest
cloud forest
roadside, between humid forest
and pastures
roadside, between rain forest
and pastures
pastures in matrix of rain
forest
coffee plantation in matrix of
humid forest
roadside in humid forest
fragment of cloud forest by
roadside
montane rain forest
cloud forest
pasture in matrix of cloud
forest remnants
approximate
elevation (m) Habitat
9
14
1
14
2
2
1
13
20
19
9
# samples
analysed
203
Cachi finca
P. stenostachya
S. fusiformis
D. odontostele
Bosque de Paz reserve potrero
P. propinqua
P. N. Rincon de la Vieja, trail to main
crater
cafetal near Tapantí
road to P. N. Tenorio
La Esperanza
Shadehouse in Perez Zeledon (thought
to come from humid forest near
Quizzará)
Rara Avis Reserve
Horquetas
Taus
road to Taus
Location
Species
9.814 N /
83.655 W
10.751 N /
84.995 W
10.788 N /
85.349 W
9.763 N /
83.808 W
10.281 N /
84.043 W
10.340 N /
84.001 W
9.383 N /
83.6 W
9.819 N /
83.788 W
9.785 N /
83.763 W
9.7817 N /
83.718 W
10.192 N /
84.321 W
Latitude /
Longitude
coffee plantation in matrix of
humid forest
1260
1450
roadside, between pastures in
matrix of humid forest
dry forest
humid forest at edge of pasture
rain forest
humid forest
pasture bordering rain forest
roadside, between rain forest
and pastures
pasture among fragments of
rain forest
secondary cloud forest
pasture among fragments of
humid forest
unknown
750
710
1000
217
750
1250
1510-1650
1560
approximate
elevation (m) Habitat
6
17
8
11
3
6
3
4
9
37
1
# samples
analysed
204
J. teretifolia
J. aporophylla
road to P. N. Tapantí
S. jimenezii
San Ramon
road to Taus
road to P. N. Tapantí
Perez Zeledon finca
La Esperanza
Perez Zeledon pasture
Taus
San Ramon
road to Taus
Taus
road to Taus
Location
Species
9.814 N /
83.655 W
9.394 N /
83.594 W
9.770 N /
83.798 W
9.783 N /
83.767 W
10.130 N /
84.503 W
9.78 N /
83.76 W
10.130 N /
84.503 W
9.78 N /
83.72 W
9.383 N /
83.6 W
9.770 N /
83.796 W
9.785 N /
83.763 W
9.782 N /
83.718 W
Latitude /
Longitude
1130
1250
1250
1220
750
1000
750
1130
1240
750
1250
1260
roadside, between pastures and
coffee plantations
roadside, between rain forest
and pastures
roadside among pastures
humid forest
humid forest at edge of pasture
pastures in matrix of rain
forest
pasture
roadside, between rain forest
and pastures
roadside among pastures
roadside, between rain forest
and pastures
roadside, between rain forest
and pastures
pasture among fragments of
rain forest
approximate
elevation (m) Habitat
11
18
13
2
2
1
7
1
28
6
8
33
# samples
analysed
205
San Cristobal
Bosque de Paz reserve potrero
Freddy's finca
Cachi finca
Bosque de Paz reserve
San Cristobal
Playa Piñuela
P. N. Manuel Antonio
Playa Dominicalito
E.
vulgoamparoanum
Playa Herradura
E. laucheanum
Cerro Caraigres
E. exasperatum
road to P. N. Tapantí
Location
Species
9.226 N /
83.841 W
9.647 N /
84.656 W
9.382 N /
84.145 W
9.098 N /
83.686 W
10.208 N /
84.325 W
9.818 N /
83.782 W
9.808 N /
83.995 W
10.192 N /
84.321 W
9.779 N /
83.992 W
9.718 N /
84.128 W
9.768 N /
83.799 W
9.778 N /
83.992 W
Latitude /
Longitude
2
coastal forest bordering beach
coastal forest bordering beach
coastal forest bordering beach
2
2
coastal forest bordering beach
fragment of humid forest by
roadside
pasture in matrix of cloud
forest remnants
pasture
cloud forest
cloud forest
roadside, between humid forest
and pastures
roadside, between humid forest
and pastures
montane rain forest
2
1870
1560
1970
1630
1720-1740
1890
1210
2160
approximate
elevation (m) Habitat
16
9
3
10
12
4
5
3
7
5
18
4
# samples
analysed
206
9.604 N /
82.610 W
9.748 N /
82.813 W
10.834 N /
85.613 W
9.382 N /
84.145 W
9.193 N /
83.779 W
9.098 N /
83.686 W
10.806 N /
85.641 W
(thought to come from Gandoca Beach)
Road to Playa Naranjo, P. N. Santa
Rosa
Playa Piñuela
Playa Hermosa
P. N. Manuel Antonio
La Casona, P. N. Santa Rosa
P. N. Cahuita
Garden in Bri Bri
10.935 N /
85.733 W
Bahia Hachal, P. N. Santa Rosa
B. nodosa
Latitude /
Longitude
Location
Species
190
2
2
2
290
1
1
2
dry forest near coast
coastal forest bordering beach
coastal forest bordering beach
coastal forest bordering beach
dry forest near coast
coastal forest bordering beach
coastal forest bordering beach
mangroves bordering beach
approximate
elevation (m) Habitat
3
3
7
10
6
2
9
10
# samples
analysed
Appendix II.
Appendix Table II.2. Scoring analysis details for AFLP datasets. Datasets that had
unacceptably high error rates were not used and are not listed here. unf. = unfiltered;
fil. = filtered; abs. = absolute scoring threshold; rel. = relative scoring threshold.
species
M. nidifica
M. rafaeliana
T. triglochin
L. ciliisepala
L. elata
L. floripecten
P. propinqua
P. stenostachya
D. odontostele
S. jimenezii
S. fusiformis
J. aporophylla
J. teretifolia
E. exasperatum
E. laucheanum
E. vulgoamparoanum
B. nodosa
Y
Primer scoring
colour method
G
B
Y
G
B
Y
B
Y
G
B
G
B
Y
G
B
Y
G
B
B
G
B
Y
G
B
Y
G
B
Y
G
B
Y
G
B
B
Y
Y
G
B
Y
B
unf. abs.
fil. abs.
unf. abs.
fil. rel.
fil. abs.
unf. abs.
fil. abs.
unf. abs.
fil. abs.
fil. abs.
unf. abs.
unf. abs.
unf. abs.
fil. abs.
unf. abs.
fil. abs.
fil. abs.
fil. abs.
fil. abs.
fil. rel.
fil. rel.
unf. abs.
fil. rel.
fil. abs.
fil. rel.
fil. rel.
fil. rel.
fil. abs.
unf. abs.
fil. rel.
fil. abs.
unf. rel.
unf. rel.
unf. rel.
fil. abs.
unf. rel.
fil. rel.
fil. abs.
fil. abs.
unf. rel.
unf. abs.
locus
phenotype mismatch
threshold threshold error rate
130
50
60
160
115
75
50
50
120
30
30
95
30
80
120
150
30
70
30
100
175
120
40
40
40
125
90
40
70
155
160
100
65
50
30
28
90
30
50
35
35
50
50
190
30
80
75
50
160
190
75
125
30
125
85
105
100
30
50
30
64
26
60
36
45
30
30
10
35
70
44
54
90
50
50
50
70
10
120
150
60
155
6.45
6.02
6.94
5.13
4.55 23
5
4.87
1.19
4.98
5.08
4.91 28
4.55
4.91
4.76
4.76
4.76
4.76
4.88
4.91
3.92
4.6
4.55
3.37
4.24
4.92
4.86
4.64
3.65
5.16
5.95
4.81
6.329
5.977
5 13
4.86
5.88 6
4.8
4.69
5
7.82 27
5.71
# loci
scored
29
26
8
17
11
25
7
13
18
1
10
11
6
3
28
26
51
5
9
9
49
60
31
23
18
41
31
11
11
17
32
13
36
13
23
14
207
Appendix II.
Appendix Table II.3. Branch circumference mean and variance for all study species.
Dryadella odontostele was only found growing on a branch once, and so no variance is
given.
Mean branch
Variance in
circumference
branch
(cm)
circumference
Species
M. nidifica
M. rafaeliana
T. triglochin
23.7 293.8
66.3 2717.2
86.2 2517.7
L. ciliisepala
L. elata
L. floripecten
31 1133.3
37 1096.9
18 139
P. propinqua
P. stenostachya
D. odontostele
15.6 612.1
129.2 1804.8
6-
S. fusiformis
S. jimenezii
J. aporophylla
J. teretifolia
55.4 1066.2
47.4 3948.7
13.6 195.3
31.3 231.8
E. exasperatum
E. laucheanum
E. vulgoamparoanum
B. nodosa
36.2 363.9
27.4 309.9
59.7 1893.9
64.5 991
Appendix Table II.4. Ecological characteristics of study clades. Clade values are means
of species values. Species range size is measured as number of regions.
Clade
Mean
elevation
range (m)
Mean number
of habitats
Mean
range size
Masdevallia
Trisetella
659.7 1.9
1300 2.3
1.3
1.7
Lepanthes
Lepanthopsis
437.3 1.6
33.3 1
1.2
2
Platystele
Dryadella
748.6 1.9
483.3 1.7
2.3
1.9
Scaphyglottis
Jacquiniella
960 2.3
750 2
4.8
6.5
Epidendrum
Brassavola
625.5 1.8
250 2.2
2.1
3.2
208
Appendix II.
Appendix Figure II.5. Associations of Fst and species range size with species phylogeny
(data from the reduced dataset). Circles at branch tips are sized proportionally to the log
of species range size measured as number of regions, and shaded according to Fst. Species
names are abbreviated by the first letter of genus and species.
209