comparative phylogeography and population genetics of ... · conservation genetics resources, 5,...
TRANSCRIPT
1
Comparative phylogeography and
population genetics of three Banded
Iron Formation endemics
Heidi M. Nistelberger
B. Biol. Sc. (Hons)
Supervisors: Prof. J. Dale Roberts
Dr Margaret Byrne
Dr David Coates
This thesis is presented for the degree of Doctor of Philosophy of The University of Western
Australia, School of Animal Biology 2014
2014
2
3
Thesis Declaration
This thesis contains published work and work prepared for publication, some of which has
been co-authored. The bibliographic details of the works and where they appear in the thesis
are set out below.
Chapter 2
Nistelberger HM, Byrne M, Coates DJ and Roberts JD (2014) Strong phylogeographic structure
in a millipede indicates Pleistocene vicariance between populations on banded iron formations
in semi-arid Australia. PLoS ONE, 9, e93038.
Nistelberger HM, Byrne M, Roberts JD and Coates DJ (2013) Isolation and characterisation of
11 microsatellite loci from the Western Australian Spirostreptid millipede Atelomastix
bamfordi. Conservation Genetics Resources, 5, 533-535. (Appendix D)
Chapter 3
Nistelberger HM, Byrne M, Coates DJ and Roberts JD (2014) Terrestrial islands in an ancient
landscape: Evidence of both historical isolation and connectivity of populations of Grevillea
georgeana, a banded iron formation endemic. Submitted: Journal of Biogeography.
Nistelberger HM, Coates DJ, Byrne M and Roberts, JD (2013) Isolation and characterisation of
11 microsatellite loci in the short-range endemic shrub Grevillea georgeana McGill
(Proteaceae). Conservation Genetics Resources, 6, 221-222. (Appendix D)
Chapter 4
Nistelberger HM, Byrne M, Coates DJ and Roberts JD (2014) Patterns of isolation and
connectivity in terrestrial island populations of Banksia arborea (Proteaceae). Submitted:
Biological Journal of the Linnean Society.
Nistelberger HM, Roberts JD, Coates DJ and Byrne M (2013) Isolation and characterisation of
14 microsatellite loci from a short-range endemic, Western Australian tree, Banksia arborea
(C.A. Gardner). Conservation Genetics Resources, 5, 1143-1145. (Appendix D)
4
All published work was primarily conducted by HMN (approx. 90%). JDR, MB and DC assisted
with project design, provided analysis advice and editing. One more paper, Chapter 5 -
Conservation of the Yilgarn BIFs, is in preparation for publication.
5
Summary
The Yilgarn Banded Iron Formations (BIFs) of south-western Australia are terrestrial islands
embedded within the largest remaining stand of Mediterranean woodland on Earth. These
ancient geological features are centres of high species richness, turnover and endemism -
characteristics attributed to the diversity of niche habitats provided by their topographic
complexity as well as historical factors, including an absence of glaciation since the Permian
(250 million years ago (Ma)). The BIFs to the south west of the Yilgarn province, along the
boundary of the South Western Australian Floristic Region, are particularly notable for their
high endemism and many species here occur across several formations, reminiscent of an
oceanic island archipelago. It is not known to what extent these disjunct BIF populations have
been historically connected by population expansion and contraction, or by long distance
dispersal events, or whether they have been isolated for long periods of time as found in taxa
endemic to granitic terrestrial islands. Regional BIF endemics provide an excellent opportunity
to assess population genetic connectivity, or lack thereof, using phylogeographic and
population genetic approaches.
I selected three taxa for phylogeographic and population genetic analysis that a) appeared to
show limited dispersal capability, b) occurred across five or more Yilgarn BIFs within the same
region and did not occur in the intervening landscape and c) were unrelated, so any
congruence in phylogeographic patterns between the taxa could not be ascribed to
phylogenetic relatedness. These taxa are:
6
1 ) Atelomastix bamfordi - a spirostreptid millipede that is restricted to pockets of moist
habitat provided by the shady gullies and areas of dense leaf litter, present on five of the
Yilgarn BIFs. Unlike its congeners found in the wetter south-western corner of the continent, A.
bamfordi was often difficult to find.
2 ) Grevillea georgeana - a proteaceous perennial shrub that is restricted to banded ironstone
and was sampled from seven Yilgarn BIFs for the purposes of this study. The showy, red
flowers of this attractive plant form a dominant food source for several species of honeyeater
pollinators. The species can be common where it occurs but on some BIFs persists only as very
small populations.
3 ) Banksia arborea - formerly known as Dryandra arborea and commonly known as the Yilgarn
tree, is a conspicuous element of the BIF flora and was sampled from eight Yilgarn BIFs. This
insect and bird pollinated species has a broader habitat requirement than either A. bamfordi
or G. georgeana and occurs anywhere from the exposed summits of outcrops to the shale and
sandy slopes lower down.
I used a combination of nuclear and cytoplasmic markers to investigate patterns of genetic
diversity and structure in these BIF endemics. The three species showed different patterns of
isolation and connectivity amongst the BIF populations, but all highlighted the Pleistocene as
an important period for lineage divergence in this landscape.
Atelomastix bamfordi: Strong phylogeographic structure indicated that the five BIF populations
have experienced long-term isolation following their separation during the Pleistocene. The
timing coincides with increasing aridity in this landscape, a process which likely marooned
millipedes in the remaining moist pockets of habitat provided by these topographical
formations, where they persist today.
7
Grevillea georgeana: The seven BIF populations of this species showed more complex
phylogeographic patterning. Four populations were behaving as isolated islands, with genetic
drift driving genetic differentiation, similar to patterns in A. bamfordi. In contrast three
populations showed evidence of more recent connectivity, possibly via inaccessible, extant
populations on low lying banded ironstone outcrops between sampled populations. Pollen
movement was restricted, although a pattern of isolation by distance across the landscape
suggests some level of connectivity between neighbouring BIFs. Again, lineage and haplotype
divergence occurred during the Pleistocene.
Banksia arborea: The eight BIF populations of this species showed complex phylogeographic
patterns with genetic divergence of some populations, reflecting isolation and genetic drift but
an absence of differentiation amongst others, suggesting more recent connectivity. Pollen
dispersal was limited and did not appear to occur between neighbouring BIF populations.
Divergence of the differentiated populations occurred during the Pleistocene.
Several common patterns of diversity and divergence emerged from the data that has allowed
identification of BIFs that may warrant conservation priority. Large, topographically complex
BIFs were found, in general, to act as reservoirs of genetic diversity, whilst small and/or
spatially isolated populations were often the most genetically divergent. These findings have
significant implications for the management of these BIFs and potentially, for other terrestrial
islands throughout the world.
8
9
Table of Contents
Summary 5
Table of Contents 9
List of Tables 13
List of Figures 14
Acknowledgements 15
Chapter 1: Introduction 19
1.1 Comparative phylogeography 19
1.1.1 Genetic markers for phylogeography and population genetic analysis 20
1.1.2 Molecular clocks 21
1.2 Genetic studies for conservation outcomes 22
1.3 Island biogeography and terrestrial islands 22
1.4 Phylogeography and population genetics of terrestrial island endemics – a global
perspective 23
1.5 Banded Iron Formations 24
1.6 The Yilgarn Banded Iron Formations 24
1.7 History of the Yilgarn 25
1.8 Yilgarn BIF endemics 26
1.9 Conservation of the Yilgarn BIFs 26
1.10 Study species and rationale 27
1.11 Aims and objectives 29
Chapter 2: Strong phylogeographic structure in a millipede indicates Pleistocene vicariance
between populations on banded iron formations in semi-arid Australia
2.1 Abstract 33
2.2 Introduction 34
2.3 Materials and Methods 39
2.3.1 Sampling and DNA extraction 39
2.3.2 Mitochondrial DNA amplification and sequencing 40
2.3.3 Mitochondrial DNA analysis 41
10
2.3.4 Nuclear microsatellite genotyping and data analysis 43
2.4 Results 45
2.4.1 Genetic variation 45
2.4.1.1 Mitochondrial DNA data 45
2.4.1.2 Nuclear microsatellite data 46
2.4.2 Phylogeographic structure 49
2.4.3 Population genetic structure 50
2.5 Discussion 53
2.5.1 Genetic differentiation of BIF populations 53
2.5.1.1 Spatial patterns 53
2.5.1.2 Temporal patterns 55
2.5.2 BIF population dynamics-historical and contemporary patterns 55
2.5.2.1 Historical patterns 55
2.5.2.2 Contemporary patterns 56
2.5.3 Conservation implications 57
2.5.4 Conclusion 58
Chapter 3: Terrestrial islands in an ancient landscape: Evidence of both historical isolation
and connectivity of populations of Grevillea georgeana, a Banded Iron Formation endemic
3.1 Abstract 61
3.2 Introduction 62
3.3 Materials and Methods 65
3.3.1 Sampling and DNA extraction 65
3.3.2 cpDNA sequencing 66
3.3.3 Fragment analysis 66
3.3.4 cpDNA data analysis 66
3.3.5 Microsatellite data analysis 68
3.4 Results 69
3.4.1 cpDNA diversity 69
3.4.2 Microsatellite data 70
3.5 Discussion 79
3.5.1 Genetic differentiation among populations 80
11
3.5.2 Genetic variation within populations 82
3.5.3 Conservation implications 83
Chapter 4: Patterns of genetic isolation and connectivity in terrestrial island populations of
Banksia arborea (Proteaceae)
4.1 Abstract 87
4.2 Introduction 88
4.3 Materials and Methods 91
4.3.1 Sampling and DNA extraction 91
4.3.2 cpDNA amplification 91
4.3.3 Microsatellite amplification 92
4.3.4 cpDNA analysis 92
4.3.5 Microsatellite analysis 93
4.4 Results 95
4.4.1 cpDNA variation 95
4.4.2 Microsatellite variation 97
4.5 Discussion 105
4.5.1 Genetic differentiation among BIF populations 105
4.5.2 Genetic variation and differentiation within a BIF population 108
4.5.3 Conservation implications 109
Chapter 5: Hotspots of genetic diversity identify terrestrial islands of conservation value in a
semi-arid Australian landscape
5.1 Abstract 113
5.2 Introduction 114
5.3 Materials and Methods 117
5.3.1 Study species and population sampling 117
5.3.1.1 Atelomastix bamfordi 118
5.3.1.2 Grevillea georgeana 120
5.3.1.3 Banksia arborea 120
5.3.2 Diversity and differentiation analysis 120
5.4 Results 121
5.4.1 Contribution to population diversity (CSR an CS) 121
12
5.4.1.1 Atelomastix bamfordi 121
5.4.1.2 Grevillea georgeana 121
5.4.1.3 Banksia arborea 122
5.4.2 Contribution to divergence (CDR and CD) 122
5.4.2.1 Atelomastix bamfordi 122
5.4.2.2 Grevillea georgeana 122
5.4.2.3 Banksia arborea 122
5.4.3 Contribution to overall diversity (CTR and CT) 123
5.4.3.1 Atelomastix bamfordi 123
5.4.3.2 Grevillea georgeana 123
5.4.3.3 Banksia arborea 123
5.5 Discussion 128
5.5.1 Large, topographically complex BIFs support populations with higher genetic
diversity 128
5.5.2 Small and spatially isolated BIFs harbour populations with greater genetic
divergence 130
5.5.3 Prioritisation of BIFs for conservation 130
Concluding remarks 135
References 139
Appendix
Appendix A 157
Appendix B 160
Appendix C 162
Appendix D 163
13
List of Tables
Chapter 2:
2.1 Atelomastix bamfordi mitochondrial and nuclear data diversity statistics 47
2.2 Population pairwise nucleotide divergence estimates of A. bamfordi 48
Chapter 3:
3.1 Grevillea georgeana chloroplast data diversity statistics 71
3.2 Population pairwise nucleotide divergence estimates of G. georgeana 71
3.3 Microsatellite genetic diversity estimates for nine G. georgeana populations 77
3.4 Population pairwise FST values for G. georgeana 78
Chapter 4:
4.1 Banksia arborea chloroplast data diversity statistics 99
4.2 Population pairwise nucleotide divergence estimates of Banksia arborea 99
4.3 Microsatellite genetic diversity estimates for 10 B. arborea populations 100
4.4 Population pairwise FST and DEST values for B. arborea 103
4.5 Population pairwise indirect gene flow estimates of B. arborea 104
Chapter 5:
5.1 Details of the 10 Yilgarn BIFs sampled for this study 119
5.2a Contributions of BIF populations to genetic diversity (cytoplasmic DNA data) 127
5.2b Contributions of BIF populations to genetic diversity (nuclear microsatellite data) 127
14
List of Figures
Chapter 1:
1.1 Map of the 10 Yilgarn BIFs sampled for this study 29
Chapter 2:
2.1 Map of the five BIF populations of Atelomastix bamfordi sampled, coloured according to
mtDNA genetic clade 49
2.2 Bayesian cladogram of A. bamfordi individuals based on mtDNA data 51
2.3 Bayesian population assignment of A. bamfordi individuals (microsatellite data) 52
2.4 Unrooted neighbour joining tree of Cavalli Sforza chord distances between the five BIF
populations of A. bamfordi 53
Chapter 3:
3.1 Map of Grevillea georgeana populations and cpDNA haplotypes identified 72
3.2 Bayesian tree of G. georgeana haplotypes based on cpDNA data 73
3.3 Maximum parsimony network for all G. georgeana haplotypes identified 74
3.4 Map of G. georgeana populations with pie charts showing proportion of assignment of
individuals to genetic clusters identified based on microsatellite data 75
3.5 Principal coordinates analysis of G. georgeana populations 79
Chapter 4:
4.1 Map of Banksia arborea populations and cpDNA haplotypes identified 96
4.2 Maximum parsimony network of all B. arborea haplotypes identified 97
4.3 Bayesian cladogram of B. arborea haplotypes based on cpDNA data 98
4.4 Principal coordinates analysis of B. arborea populations 102
Chapter 5:
5.1 Map of the 10 Yilgarn BIFs sampled for this study 118
5.2 Contributions of Atelomastix bamfordi populations to total genetic diversity 124
5.3 Contributions of Grevillea georgeana populations to total genetic diversity 125
5.4 Contributions of Banksia arborea populations to total genetic diversity 126
15
Acknowledgements
You know how in some theses the student acknowledges the supervisors, but you can tell it’s out of
duty and not because they really mean it? This is not one of those theses. The last four years have been
a wonderful experience for me and I have three awesome supervisors to thank for it.
Dale, I couldn’t have asked for a better primary supervisor. You never suffocated or hovered over me,
neither did I ever feel abandoned or without support. Even when you moved to Albany you were always
a phone call away…never too busy to discuss my ‘conceptual crises’ and run through them with me on
the whiteboard. I loved Friday student coffee sessions, chatting about life, science and even occasionally
our research. Your dedication towards me (and all of your students) went above and beyond what I
could ever have hoped for - I will always remember this as an amazing period of my life…and I have you
to thank for it.
Margaret, you have been my biggest critic and therefore biggest supporter for years now. You
continually push me to improve my understanding in all areas of science and your unwavering
commitment to me, despite being one of the busiest women on earth, leaves me grateful beyond
words. That you managed to return writing to me within the week despite all your other commitments
was always noticed and shows how seriously you take on the role of supervisor. Thankyou thankyou
thankyou.
Dave, you were always the person I turned to when I experienced occasional episodes of self-doubt.
That you could recognise when I was struggling and would make the effort to ring me up to make me
laugh, and make sure I was confident and on-track shows how much you cared…and went a long way
toward helping me through the really busy times. Whenever I need to be reminded of why I’m working
in science I go and park myself in your office. Your passion and enthusiasm for research are infectious
and always re-ignite my love of science. I look forward to many more discussions over a beer in the
future. Thankyou for all your support.
My family. Mum, Dad and Alan, you’ve always been my biggest supporters, encouraging me along the
way, getting as excited about my results as I was. Mum, I’ll never forget the day you pointed out that
although you weren’t sure what the term meant you thought I had spelled “heterozygote”
incorrectly…and you were right, of course. You have all inspired me throughout my whole life and
continue to do so. I wouldn’t be where I am today without your unwavering love and support. I’m the
luckiest girl in the world to be able to call you all my own.
Brad, you came into my life during what some describe as the most stressful times of one’s life. Yet
straight away you jumped right in, reassuring me when I panicked and broke down in tears because I
realised I only had 6 months left and didn’t want it all to end! You never complained about my lack of
time, money and occasionally, ability to talk of anything but my Ph D . You have been my rock this last
year and I couldn’t have finished in the semi-sane state I did without your loving influence and support.
Thanks Handsome.
Uni crew. Ahh I’ve had such fun here. This place is buzzing with interesting, passionate people, so many
of whom have made my time here. The zoology building feels like an extension of my home and there
are too many people to thank to list here. In particular I’m going to miss the Friday sessions at the pub,
summer drinks on Kath and Emile’s balcony. Renee, we’ve had some ridiculous adventures tapering to
the more sensible ‘swim sessions’ as I got ready to submit. Sam, Steve and Georgina it’s been a blast to
share an office with you all, always ready to discuss the finer horrors of ms word, or to solve the world’s
problems over a packet of stale ginger nuts. Danilo, we were the lone phylogeographers, thank
16
goodness we had each other to support, to bitch about analyses that took weeks on end (only to crash
days before completion due to power outages). Thanks for all your feedback and support.
DPaW ladies. My other home away from home. Bron and Shell you have always been so supportive of
me, both in and out of the lab. I love that we can laugh over the minor disasters… the smiley gels and
snotty DNA extracts and enjoy a good session of tip packing when there was some need of therapy time.
You made my time of crazy lab work bearable and even enjoyable, thankyou. All the other ladies down
the corridor, you all contributed to making my visits something I looked forward to and in particular
Kym, Sarzie, Binksie, Hevs, Donna and Tan L …you made me laugh till my tummy hurt. I’ll always
remember hoopla mornings with fondness.
Emily, thanks for all the amazing help in the field, your keen ‘milli’ eyes, endless enthusiasm and sense
of humour made our long days of fossicking a blast. I absolutely could not have done this project
without you.
My school gals. Even though it was a world removed from yours you still supported me these last 4
years. I’m so lucky to have friends that span the length of my life. I always look forward to our coffee
catch ups where I no doubt bore you senseless, gabbing on about my research. Thanks for listening, now
and always.
Outdoor crew. For introducing me to a world away from academia, thanks Bruno, Steve, Jerome, Robin
and all the UWA ODC crew for keeping me sane on the weekends and my hands nicely calloused for all
those climbing adventures.
Funding. I’d like to thank Cliffs Natural Resources for their generous contribution. This research would
not have been possible without their support. The school of Animal Biology (UWA), for annual funding
and the Holsworth Wildlife Research Grant, for making the microsatellite libraries possible. Finally,
Department of Parks and Wildlife for help with lab consumables and other unforeseen costs.
Specific chapters
Chapter 2: Thanks to Mark Harvey, Mike Rix (Western Australian Museum), Roy Teale (Biota) and Karl
Brennan (DPaW) for millipede discussions and specimens. Thanks to Janine Rix, Maxine Lovegrove and
Yvette Hitchen for extraction advice. Nicole Harry, Jeremy Sheperdson (Cliffs) and Scott Reiffer for
arranging site access, fuel and accommodation when required and Emily Ager for fieldwork.
Chapter 3: Thanks to Kym Ottewell for advice regarding pcr multiplexing…saved me many swears and
Neil Gibson (DPaW) for BIF chats and Emily Ager for fieldwork.
Chapter 4: Thanks to Marcel Cardillo for providing Banksia node date estimates and to Neil Gibson for all
the helpful BIF discussions and Emily Ager for fieldwork.
How else could I end but to say thanks to the Yilgarn BIFs, what an amazing place…
17
A tale of three species,
one landscape
18
19
Chapter 1
Introduction
1.1 Comparative phylogeography
Phylogeography ties together the disciplines of biogeography, cladistics and population
genetics. By definition, it is the study of historical phylogenetic components that result in the
spatial distributions of gene lineages (Avise, 2000). In comparative phylogeography, these
lineages are compared in order to infer common historical events that may have led to
unrelated taxa sharing similar distributions (Carstens et al., 2005). Comparative
phylogeography studies generally focus on two processes affecting species distributions and
division: vicariance and dispersal and their role in speciation (Cox & Moore, 2000; Raxworthy
et al., 2002; Yoder & Nowak, 2006; Gamble et al., 2008).
By studying the historical patterns present in a suite of co-occurring taxa, we can infer a great
deal about the important evolutionary processes acting in landscapes. For example, in a study
of five widespread, south-west Australian tree species, three were identified as showing
concordant patterns of lineage divergence, the timing of which suggested a generalised
response to climatic fluctuations during the Pleistocene (Byrne, 2007). Another study of two
plant and four animal species from the Pacific Northwest, U.S.A., was used to test three
hypotheses, one of vicariance and two of dispersal, that would explain disjunctions in the
mesic forest ecosystem of the region. Three amphibian lineages indicated a generalised
response to an ancient vicariance event, whilst the other three species showed recent
dispersal in a northern direction, suggesting that a combination of historical events had led to
the congruent, disjunct distributions of these species (Carstens et al., 2005). Comparative
studies may also be used to uncover regions within a landscape that commonly exhibit high
20
genetic diversity or distinctiveness, characteristics that may warrant conservation of these
areas (Moritz & Faith, 1998).
1.1.1 Genetic markers for phylogeographic and population genetic analysis
Mitochondrial DNA (mtDNA) and chloroplast DNA (cpDNA) sequences have been popular
markers for inferring phylogeographic history because of their beneficial characteristics
including a lack of recombination, effective haploidy (Harris & Ingram, 1991; Zhang & Hewitt,
1996; Avise, 2000) and a maternal mode of inheritance, allowing observation of lineages not
confounded by recombination during sexual reproduction (but see Hipkins et al., 1994;
Testolin & Cipriani, 1997; McCauley et al., 2007 for exceptions in plants, and Kondo et al.,
1990; Gyllensten et al., 1991 for exceptions in animals). Maternal inheritance of cpDNA is
particularly useful in population studies of plants as contrasting cpDNA polymorphism with
biparentally inherited marker polymorphism can allow evaluation of the relative influences of
seed and pollen on gene flow (McCauley, 1995). Several criticisms have been raised regarding
the exclusive use of organelle DNA markers in phylogeographic studies. Their asexual mode of
inheritance effectively means that these genomes can be thought of as a single locus (Zhang &
Hewitt, 2003) where all character states are linked (Avise, 2000), and therefore, potentially
subject to selective sweeps (Avise et al., 1987; Zhang & Hewitt, 1996; Gerber et al., 2001). The
majority of mutations in mtDNA have been observed in silent positions of protein-coding
genes and in the non-transcribed D-loop region. These positions are not expected to be linked
to organismal fitness and so are considered useful for phylogeography (Brown, 1983; Avise et
al., 1987), likewise mutations in the cpDNA typically occur in non-coding regions (Palmer et al.,
1988; Taberlet et al., 1991). Another potential problem with mtDNA and cpDNA genealogies is
that they allow only one, matrilineal perspective on evolutionary history, and, as the structure
and history of any one particular gene or DNA sequence is likely to differ from that of the
population or species, it is necessary to take a multi-locus approach to phylogeographic studies
(Hare, 2001; Zhang & Hewitt, 2003; Brito & Edwards, 2009). The use of mtDNA and cpDNA
21
markers needs to be combined with nuclear DNA (nrDNA) markers where possible.
Unfortunately, sequence markers commonly available for nuclear genes may lack sufficient
variation to be useful at the population level (Zhang & Hewitt, 2003). Other, more rapidly-
evolving nuclear markers such as microsatellites, can provide a useful alternative and have
been used in conjunction with traditional mtDNA and cpDNA markers in many studies (Bech et
al., 2009; Beatty & Provan, 2011; Keir et al., 2011; Sakaguchi et al., 2012). The high variability
of these markers and their co-dominant, Mendelian inheritance patterns, provides important
population level information on genetic diversity and structure that can be used to detect
more contemporary evolutionary processes influencing populations, such as recent
bottlenecks (Luikart et al., 1998) and contemporary gene flow (Sork et al., 1999). Using both
rapidly evolving markers and the slower mtDNA and cpDNA markers is advantageous in
situations where phylogenetic splits may be relatively recent, which may occur in population-
level studies. In these cases, a lack of variation at the cytoplasmic level may be compensated
for by the highly informative microsatellite markers. This is particularly important in plants,
where cpDNA variation may often be very low (Keir et al., 2011).
1.1.2 Molecular clocks
Molecular clocks have become an important component of phylogeographic analysis, allowing
for the timing of divergence of major clades and lineages at the intra and interspecific level to
be estimated. The molecular clock hypothesis is based on the finding that DNA sequences
accrue mutations at a relatively constant rate (Margoliash, 1963; Zuckerkandl & Pauling, 1965).
In some species, this rate can be determined by calibrating a phylogenetic tree with a known
divergence date from fossil or geological data (Weir & Schluter, 2008). When this information
is not available, a generalised mutation rate inferred from other, closely related species can be
used (Byrne, 2007; Edwards et al., 2007). Ideal phylogeographic markers possess a rate of
evolution fast enough to reveal adequate variation but not so fast as to incur the detrimental
effects of homoplasy and saturation (Goldman, 1998). This has commonly involved the use of
mtDNA sequence markers such as CO1 in animals (Hebert et al., 2003) and non-coding
22
intergenic cpDNA markers in plants (Shaw et al., 2007) because of their optimal rates of
mutation at intra-specific scales.
1.2 Genetic studies for conservation outcomes
Genetic analysis has become an increasingly important component of species’ management
plans (Hedrick, 2001; Allendorf et al., 2013). In the USA, agencies administering the
Endangered Species Act are now turning to genetic data to target populations for conservation
(Kelly, 2010). Preservation of genetic diversity is at the forefront of conservation decisions, as
higher diversity levels are thought to enable continual adaptation to changing environmental
conditions (Soulé, 1980). The degree of genetic divergence present within a species is also
important. Populations that have been isolated for extensive periods of time may show genetic
divergence that may result in future allopatric speciation if isolation is prolonged (Dobzhansky,
1970; Mayr, 1970). Conservation efforts will ideally encourage preservation of populations or
regions of both high genetic diversity, as well as genetic divergence (Faith, 1992; Petit et al.,
1998).
1.3 Island biogeography and terrestrial islands
Islands provide a unique context for studying the evolutionary processes that drive species
distribution. They are discrete and quantifiable entities that provide a natural laboratory from
which to examine processes of speciation, and patterns of species colonisation, persistence
and extinction (Whittaker, 1998). Terrestrial islands, or habitat islands, are discrete habitat
patches embedded in the surrounding landscape, and include anything from large and
conspicuous geological formations, such as mountain tops, to dung piles in a grassland
landscape (Whittaker, 1998). Large and temporally stable terrestrial islands such as granite
outcrops and inselbergs, may be characterised by high species endemism in comparison to the
surrounding environment, a result of the different habitats provided by these topographical
23
features (Burke, 2003; Jacobi et al., 2007; Gibson et al., 2010). Species endemic to these
terrestrial islands are of particular interest to evolutionary biologists, as they provide a model
for studying patterns of species dispersal and colonisation in a fragmented landscape.
1.4 Phylogeography and population genetics of terrestrial island endemics - a global
perspective
Sky islands are amongst the most well-studied terrestrial island systems (Knowles, 2000;
DeChaine & Martin, 2005; Smith & Farrell, 2005; Assefa et al., 2007; Bech et al., 2009). These
are mountains or mountain-tops surrounded by low-lying areas of starkly contrasting
environmental habitat. Flora and fauna may be restricted to these mountain top ‘islands’ or to
the habitat ‘islands’ between the mountain tops. Phylogeography and population genetic
studies have been used to show whether disjunct sky island populations have maintained
genetic connectivity, either via dispersal or repeated expansion dynamics during favourable
climatic periods (Floyd et al., 2005), or whether they have been isolated for long periods of
time leading to substantial genetic divergence (Masta, 2000). When examined from a
landscape context, studies can uncover common patterns relating to the history of a region,
for example, research from the northern hemisphere has highlighted the influence of
Quaternary glacial cycles on the geographic distributions of population genomes, with several
species showing rapid northward expansion of populations into newly inhabitable landscapes
during more moderate climates of the inter-glacial periods (Hewitt, 2000).
More recently, phylogeographic studies have been undertaken on the more low-lying granites,
inselbergs and other rock outcrops as terrestrial islands (Levy & Neal, 1999; Byrne & Hopper,
2008; Duputié et al., 2009; Levy et al., 2012; Tapper et al., 2014a, 2014b). These have
identified contrasting patterns of connectivity and isolation reflecting different evolutionary
responses of species to historical events.
24
1.5 Banded Iron Formations
Banded Iron Formations, or BIFs, are ancient geological features associated with Archaean
Cratons throughout the world that behave as terrestrial islands. Often prospective for resource
development (Jacobi & Carmo, 2008; Gibson et al., 2012), BIFs are surprisingly under-studied
in comparison to other terrestrial islands systems, a concerning fact given that flora surveys of
BIFs in south-eastern Brazil and south-western Australia have both identified high species
richness and endemism (Jacobi et al., 2007; Meissner & Caruso, 2008a; Gibson et al., 2010). To
date there has been just one population-level genetic study conducted on a BIF endemic;
genetic diversity and differentiation was examined across populations of Acacia
woodmaniorum, found on four BIF islands in Western Australia. The study identified sufficient
connectivity amongst BIF islands to counteract the potentially negative influences of genetic
drift and inbreeding on populations, an outcome which has significant implications for
management and future restoration of this species (Millar et al., 2013).
1.6 The Yilgarn Banded Iron Formations
The Yilgarn BIFs are part of the Yilgarn Craton of Western Australia, one of the oldest regoliths
on Earth that covers an area of 657 000 km2 (Freeman, 2001; Chivas & Atlhopheng, 2010). The
BIFs are comprised of banded iron talus slopes and areas of weathered duricrust (Hocking et
al., 2007), and are associated with greenstones and granites that formed 2.6 to 3 billion years
ago (Myers, 1993). These island-like formations occur along the boundary of the Southwest
Australian Floristic Region (SWAFR) and extend inland to the arid zone for over 750 km (Gibson
et al., 2012). BIFs vary widely in size, shape and distance from one another. Those bordering
the SWAFR show higher endemism and species turnover than more arid BIFs (Gibson et al.,
2010, 2012), characteristics attributed to their positioning along the boundary of the wetter
south-western environment and the arid interior (Butcher et al., 2007). The landscape
25
surrounding the BIFs is a mosaic of alluvial clay soils dominated by eucalypt woodland,
interspersed with sandplains and occasional granite outcropping.
1.7 History of the Yilgarn
The Yilgarn region of Western Australia has been geologically stable, un-glaciated and above
sea level since the Permian (250 Ma) (Hopper, 2000; McLoughlin, 2001). Nonetheless there
have been substantial changes in climate and vegetation communities over time. At the
beginning of the Cenozoic (65 Ma) the region was warm, humid and dominated by temperate
rainforest (Martin, 2006). Over time, the landscape became increasingly arid, beginning
approximately 11 Ma in the mid-Miocene, as evidenced by the absence of regular flows in
paleo-drainage systems (Martin, 2006). By the early Pleistocene (approx. 2 Ma) the modern
climate was established but mesic-arid fluctuations continued throughout this period (Zheng et
al., 1998).
The complex topography of BIFs provide an abundance of niche habitats that are believed to
have allowed for the persistence of species that were progressively extirpated in the
increasingly arid environments surrounding these BIFs (Byrne et al., 2008). The current
biodiversity is thought to be a mixture of these relictual species and more recently derived,
arid-adapted species (Byrne et al., 2008). Today the region is considered semi-arid, with
rainfall occurring throughout the year but predominantly in winter months (200-300 mm
annually) (Gibson et al., 2012).
26
1.8 Yilgarn BIF endemics
Yilgarn BIF endemics can be broadly categorised into those that are localised, occurring on just
one BIF and those that are regional, occurring across multiple formations. Flora surveys of the
Yilgarn BIFs have identified many regionally endemic plant species with distributions restricted
to, or centred on BIFs (Gibson et al., 2007). Less is known about the fauna of the BIFs, but it is
expected that species with limited dispersal capabilities (short-range endemics), may exhibit
regional endemism, particularly relictual invertebrates such as millipedes, land snails and
troglofauna that persist in the mesic micro-habitats provided by BIFs (Harvey, 2002;
Bennelongia, 2008; Biota, 2009). It is not known whether individual BIF populations of regional
endemics are genetically isolated, or whether they show evidence of population connectivity.
In plant populations connectivity will be maintained by pollen or seed dispersal. In fauna,
connectivity must occur via migration, which in less mobile species may be facilitated by
animal vectors or flood events.
1.9 Conservation of the Yilgarn BIFs
The Yilgarn BIFs and their surrounds are unique in that until now, they have remained in a
relatively pristine condition in comparison to other BIFs throughout the world (Burke, 2003).
Mining for iron-ore and gold in the region has led to identification of the diversity and
endemism present on these formations (Gibson et al., 2012) and there has been recent
political interest as to whether the balance between conservation and resource development
in this landscape is being met (Environmental Protection Authority, 2013). The current speed
of development has hastened the importance of understanding the processes resulting in high
species endemism on the BIFs, in order to mitigate the potential impact on these restricted
species.
27
1.10 Study species and rationale
I investigated the patterns of genetic diversity and differentiation present in regional endemics
of the Yilgarn BIFs, in order to understand the evolutionary forces shaping distributions in this
landscape. To select my target taxa I worked with four major criteria:
(1) Taxa showed evidence of restricted dispersal capabilities. In plants this meant selecting species
with seed that showed no active dispersal mechanisms and were not wind-pollinated, as wind-
pollination generally maintains population connectivity (Loveless & Hamrick, 1984).
(2) Taxa were each found on more than three Yilgarn BIFs.
(3) Taxa were taxonomically resolved, with vouchered specimens from multiple locations.
(4) Taxa were unrelated. The more dissimilar the species selected, the greater the strength of the
study, as congruent patterns between taxonomically diverse species are more likely to suggest
a wide-spread response to an historical event (Zink, 1996).
Based on these criteria I selected three species, a millipede and two plants.
(a) Atelomastix bamfordi (Edward and Harvey) is a recently described Spirostreptid millipede,
originally recorded from five Yilgarn BIFs (Edward & Harvey, 2010) that is now known to occur
on six; the Die Hardy Range, Helena Aurora Range, Windarling Range, Mt Jackson,
Koolyanobbing Range and Marvel Loch. Windarling, Koolyanobbing and Mt Jackson are all
actively mined for iron-ore and for Windarling and Koolyanobbing, specimens were sourced
from the Western Australian Museum (WAM) where they had been stored in 100 % ethanol.
Specimens from Helena Aurora, Die Hardy and Mt Jackson were collected by hand. Samples
could not be collected from Marvel Loch due to mining operations. In contrast to its congeners
found around the wetter regions of the south-western Australian coastline, A. bamfordi was
28
often difficult to find. Sampling success was greatest in the cold winter months or following
rain events that occur sporadically throughout the year. The majority of millipedes were
associated with deep leaf litter at the base of rocky slopes and under rocks in shaded gullies.
Several were found behind bark at the base of large, shaded trees. At the Helena Aurora
Range, millipedes were found under logs and leaf litter on flat clay soils surrounding the BIF at
the base of drainage lines that were prone to flooding. To ensure the species was restricted to
BIF, other flood-prone sites in the intervening flats were also searched and, although other
millipedes from the genus Antichiropus were recovered, A. bamfordi was never found away
from a BIF.
(b) Grevillea georgeana (McGill) is a member of the family Proteaceae. This bird-pollinated,
perennial woody shrub has attractive pink to red flowers that provide a major food source for
a number of local honeyeater species (Olde & Marriott, 1995). It is an obligate seeder and was
noted to respond well to disturbance at one site. Strongly associated with the dark,
mineralised rocky outcrops of the BIF, G. georgeana is capable of tolerating extremes in heat
(Olde & Marriott, 1995). The species tended to be patchily distributed across most of the BIFs
but was common where it did occur. I sampled G. georgeana from seven Yilgarn BIFs; the
Helena Aurora Range, Die Hardy Range, Hunt Range, Mt Correll, Mt Manning, Mt Dimer and
Mt Finnerty.
(c) Banksia arborea (C. A. Gardner), formerly Dryandra arborea, is another member of the
Proteaceae but the genus is significantly removed from Grevillea, having diverged
approximately 77 Ma (Sauquet et al., 2009; Cardillo & Pratt, 2013). This long-lived tree forms a
conspicuous component of the BIF vegetation, producing an abundance of yellow flowers that
are visited by a range of birds and invertebrates. The woody fruits are serotinous, releasing
their seed following fire (Wrigley & Fagg, 1989). Banksia arborea tends to be common where it
occurs and is not as habitat specific as G. georgeana, with some populations occurring at the
base of BIFs and one population noted at the base of a granite outcrop (pers. obs. N. Gibson),
suggesting the species prefers the well-drained habitats of upland areas. I sampled B. arborea
29
from eight Yilgarn BIFs; the Koolyanobbing Range, Mt Jackson, Windarling Range, Die Hardy
Range, Mt Manning, Taipan Hill, Hunt Range and the Helena Aurora Range.
Figure 1.1 Map of the 10 Yilgarn BIFs in Western Australia sampled for Atelomastix bamfordi, Grevillea georgeana and Banksia arborea. Image adapted from Google Earth.
1.11 Aims and objectives
The research undertaken in this thesis had three primary objectives:
1) Determine the evolutionary processes shaping species distributions by assessing patterns in
phylogeography and population genetic structure. This is composed of two parts to be
addressed separately for each species:
i) Quantify genetic diversity within and between island populations and determine whether
diversity is geographically structured.
ii) If genetically divergent, to estimate the time of divergence of island populations.
30
2) Determine whether there are any congruent phylogeographic patterns between taxa that
could indicate shared responses to historical events. For example, has population divergence
coincided with a known period of increasing aridity?
3) Propose conservation strategies to ensure the long-term persistence of these BIF endemics
based on collective data.
31
Oh shining
millipede…Where art
thou?
32
Work in this chapter has been published as:
Nistelberger HM, Byrne M, Coates DJ and Roberts JD (2014) Strong phylogeographic structure
in a millipede indicates Pleistocene vicariance between populations on banded iron formations
in semi-arid Australia. PLoS ONE, 9, e93038.
Nistelberger HM, Byrne M, Roberts JD and Coates DJ (2013) Isolation and characterisation of
11 microsatellite loci from the Western Australian Spirostreptid millipede Atelomastix
bamfordi. Conservation Genetics Resources, 5, 533-535. (Appendix D)
33
Chapter 2
Strong phylogeographic structure in a millipede indicates Pleistocene
vicariance between populations on Banded Iron Formations in semi-arid
Australia.
2.1 ABSTRACT
The Yilgarn Banded Iron Formations of Western Australia are topographical features that
behave as terrestrial islands within the otherwise flat, semi-arid landscape. The formations are
characterised by a high number of endemic species, some of which are distributed across
multiple formations without inhabiting the intervening landscape. These species provide an
ideal context for phylogeographic analysis, to investigate patterns of genetic variation at both
spatial and temporal scales. We examined genetic variation in the spirostreptid millipede,
Atelomastix bamfordi, found on five of these Banded Iron Formations at two mitochondrial loci
and 11 microsatellite loci. Strong phylogeographic structuring indicated the five populations
became isolated during the Pleistocene, a period of intensifying aridity in this landscape, when
it appears populations have been restricted to pockets of moist habitat provided by the
formations. The pattern of reciprocal monophyly identified within the mtDNA and strong
differentiation within the nuclear microsatellite data, highlight the evolutionary significance of
these divergent populations and we suggest the degree of differentiation warrants designation
of each as a conservation unit.
34
2.2 INTRODUCTION
Islands provide a unique context for studying the evolutionary processes that drive species
distributions. They are natural laboratories, allowing examination of processes of speciation,
and patterns of colonisation, persistence, dispersal and extinction (Whittaker, 1998). The
genetic consequences of island distributions are well documented by phylogeographic and
population genetic studies that infer historical patterns of natural and sexual selection, drift
and their interactions with historical and ongoing dispersal (Raxworthy et al., 2002;
Beheregaray et al., 2004; Jones & Kennedy, 2008; Wesener et al., 2010). The study of
terrestrial or ‘habitat’ islands appears to offer a different challenge to oceanic islands, as the
matrix between islands can be more permeable than water barriers, resulting in potentially
complex spatial and temporal patterns of genetic structure (Masta, 2000; Robin et al., 2010).
The most conspicuous and well-studied terrestrial islands include mountain ranges that form
‘sky islands’, such as the tepui summits of South America, the Western Ghats of southern India
and the Madrean and Rocky Mountain sky island systems of central and northern America
(DeChaine & Martin, 2004, 2005; Smith & Farrell, 2005; Storfer et al., 2007; Robin et al., 2010;
Salerno et al., 2012). Here, ‘island species’ form disjunct populations on either mountain tops
or the lowland valleys separating mountains. Other, less conspicuous topographic habitat
islands are found throughout the world and include granite outcrops, inselbergs and Banded
Iron Formations (BIFs), many of which display high endemism. Although topographically less
dramatic, these provide equally compelling bases for phylogeographic study (Baskin & Baskin,
1988; Burke, 2003; Jacobi et al., 2007; Byrne & Hopper, 2008; Gibson et al., 2010).
Inferring the evolutionary history of terrestrial island endemics can be difficult given that
isolation is not always absolute, either spatially or temporally. Dispersal events can occur
across valleys and lowlands for mountain-top endemics or over mountains for valley floor
endemics, depending on the vagility of the organism and their environmental requirements
35
(Masta, 2000; Robin et al., 2010). Despite this, phylogeographic patterns, although potentially
complex, may reveal the degree of connectivity or isolation experienced by island populations.
Restricted gene flow between island populations over long periods of time will result in
substantial genetic divergence due to genetic drift and/or, local adaptation (Dobzhansky,
1970). Given sufficient time, populations may show patterns of reciprocal monophyly, where
all individuals can be traced to a common ancestor within that population (Avise et al., 1983;
Moritz, 1994). Recurrent migration and gene flow between islands on the other hand, has a
homogenising effect on genetic structure, unless there is strong selection favouring local
genotypes on each island. For example, populations of yellow bellied marmots, restricted to
the Great Basin mountain-tops of North America, were believed to have been marooned on
the rocky outcrops of mountain tops following habitat contraction during the Pleistocene
(Floyd et al., 2005). A pattern of isolation by distance and high genetic diversity showed that
occasional dispersal events across the arid scrublands of the Basin have been an important
process connecting mountain-top populations (Floyd et al., 2005). A contrasting pattern was
found in the tree Eucalyptus caesia, endemic to granite outcrop ‘islands’ of Western Australia.
This species has high levels of genetic divergence with little geographic pattern to the
distribution of chloroplast haplotypes, reflecting long-term isolation and persistence without
connectivity and emphasising the role of genetic drift in this landscape (Moran & Hopper,
1983).
The Yilgarn Banded Iron Formations (BIFs) of south-western Australia are ancient, island-like
rock outcrops that provide an ideal system to investigate island population dynamics. They
border the transitional rainfall zone, a biogeographic region separating the high rainfall forest
region from the arid interior and are characterised by high species endemism and richness
(Gibson et al., 2007, 2010). BIFs vary widely in height (highest point approx. 700 m above sea
36
level), shape and distance from one another and are comprised of banded iron talus slopes
and areas of weathered duricrust (Hocking et al., 2007). These topographically complex
formations occur in an otherwise flat landscape that is dominated by clay, silt and sand plains.
The region has been geologically stable, un-glaciated and above sea level since the Permian
(250 Ma) (Hopper, 2000; McLoughlin, 2001; Anand & Paine, 2002). Nonetheless, there have
been substantial changes in climate and vegetation communities over this time. At the
beginning of the Cenozoic (65 Ma), the region was warm, humid and dominated by temperate
rainforest (Martin, 2006). Over time, the landscape became increasingly arid, beginning
approximately 11 Ma in the mid-Miocene, shown by the absence of regular flows in paleo-
drainage systems (Martin, 2006). Another aridification event occurred 7 Ma in the late
Miocene with several more throughout the Pliocene, shown by changes in vegetation structure
in palynological deposits to more arid-adapted forms (Dodson & Macphail, 2004; Martin,
2006). By the early Pleistocene (approx. 2.6 Ma), the modern climate had more or less become
established but fluctuations throughout this period continued, associated with successive
glacial and interglacial periods, the latter with intensifying arid conditions (Martin, 2006).
Historically, the altitudinal relief of BIFs and their complex structure, including deep valleys and
shaded gullies as well as exposed outcrops and talus slopes (Hocking et al., 2007), would have
provided an abundance of niche habitats that allowed for the persistence of both mesic and
xeric-adapted species during changing climates (Hopper et al., 1996). Whilst this may account
for the high species diversity found on the BIFs today, the high endemism (Gibson et al., 2007,
2010; Yates et al., 2011) also suggests long periods of isolation, allowing for speciation and the
development and persistence of a unique flora and fauna. The Yilgarn BIF endemics consist of
species that are either localised (found on one BIF), or regional (located on several BIFs but not
in the intervening landscape) (Gibson et al., 2007). Regional endemics provide an opportunity
to study phylogeographic patterns, to determine what processes have influenced current
37
distributions, and historically, how diversity has evolved and accumulated. The floral
assemblages of the Yilgarn BIFs are well known, with many examples of regional endemics
(Gibson et al., 2007). Although less is known of the faunal assemblages, taxa that exhibit short-
range endemism consistent with limited dispersal capacity, such as millipedes, mygalomorph
spiders, land snails and troglofauna (Harvey, 2002; Bennelongia, 2008; Biota, 2009), may also
exhibit patterns of regional endemism, persisting in micro-habitats provided by the BIF.
We investigated phylogeographic pattern in the spirostreptid millipede Atelomastix bamfordi,
known to occur on five BIFs on the Yilgarn Plateau in Western Australia. Millipedes lack the
waxy cuticle that protects other arthropods from desiccation and are reliant on moist
environments for survival (Mesibov, 1994; Harvey, 2002). Atelomastix, from the family
Iulomorphidae, is a predominantly south-west Australian genus, with many species distributed
across the wetter parts of the southern and south-western coastline and mesic inland Ranges,
including the Stirling and Porongorup Ranges (Moir et al., 2009; Edward & Harvey, 2010).
Almost all species have restricted distributions and have been characterised as short-range
endemics (geographic range < 10000 km2) (Harvey, 2002). Atelomastix bamfordi is the most
easterly occurring member of this genus and inhabits moist habitats provided by BIFs, such as
the base of rocky slopes, under large rocks and under dense leaf litter. The life cycle is likely
annual with adult millipedes dying prior to, or during the harsh summer months (pers. comm.
M Harvey). The species has been reported as abundant on the large Koolyanobbing Range
(Biota, 2009) but appears scarcer on the drier BIFs to the north and east. Increased exploration
and mining of iron ore within the region (Gibson et al., 2007) has resulted in the need for a
delicate balance between mining activities and conservation of BIF endemic species.
Understanding the patterns and evolution of diversity in these sites may guide two key
conservation issues: i) the viability and conservation of highly endemic species ii) the
conservation of landscape features which maximise the chances of maintaining patterns of
38
dispersal, gene flow and local adaptation that have shaped the history of extant diversity and
continue to be major factors in the persistence of this diversity.
Although unconnected today, BIF populations of A. bamfordi may have repeatedly come into
contact through population expansion associated with historical climate change in this
landscape. If this were the case, we would expect periods of gene exchange between
neighbouring BIF populations to result in a pattern of isolation-by-distance, where proximate
populations are more genetically related, or, if gene exchange has been extensive, to result in
genetic homogeneity of populations (Kimura & Weiss, 1964a). Alternately, the current
distribution may be due to contraction of a formerly widespread ancestor, followed by long-
term isolation and persistence on BIF islands. If this were the case, we would expect to see
signatures of deep genetic divergence amongst BIF populations but no clear pattern of
isolation by distance. Haplotypes would be conserved and there may be evidence of reciprocal
monophyly, if sufficient time for lineage sorting has occurred (Avise, 1989). This is the first
phylogeographic study to examine patterns of population level divergence in a regional BIF
endemic species from this biodiverse region.
We evaluated the genetic variation present in A. bamfordi to investigate evolutionary
relationships and patterns of genetic diversity and structure across BIF populations. Our goals
were:
1. To determine patterns of genetic differentiation among the five known BIF populations
and to use those data to infer the impact of historical processes on the distribution of
millipedes across the landscape.
2. To assess genetic diversity and structure to determine how limited dispersal
capabilities affect structure within BIF millipede populations.
39
3. To use patterns of genetic differentiation, both within and among BIFs, to inform
conservation and management of this species.
2.3 MATERIALS AND METHODS
2.3.1 Sampling and DNA extraction
A. bamfordi had been recorded from four of the Yilgarn BIFs; Koolyanobbing (KOO), Windarling
(WIN), Mt Jackson (MTJ) and the Die Hardy Range (DH) (pers. comm. R. Teale) (Biota, 2009).
This study extended the known distribution to include the Helena Aurora (HA) Range. Four
more BIFs in the region (Mt Dimer, Mt Correll, Hunt Range and Mt Manning) were also
searched but no specimens were found. Specimens from the Windarling and Koolyanobbing
Ranges, preserved in 100% ethanol were available from the Western Australian Museum’s
Department of Terrestrial Zoology collection (Appendix Table A1). For the remaining BIFs (DH,
MTJ and HA) sampling was spread across the geographic extent of each as far as was practical
to avoid sampling closely related individuals. For some BIFs this was not possible, as millipedes
were either not found and/or there was no suitable habitat. The same search procedure was
applied to all BIFs, so we believe it is reasonable to assume that the specimens collected reflect
the genetic diversity contained within each site. The five BIFs vary in size, topographic
complexity and orientation. Koolyanobbing is large, approximately 12 km in length and
reaching heights of 500 m. This BIF is topographically complex and contains many sites with
habitat suitable for millipedes. Helena Aurora is also large (~ 12 km) and elevated (600-700 m
height), but is comprised of more rounded hills with gentle slopes. Millipedes here were
associated with drainage lines coming off the peaks. The Die Hardy Range is large and X-
shaped, a topology providing many shaded slopes that supported habitat suitable for
millipedes. The two smaller Ranges, Windarling and Mt Jackson, have a narrow east to west
orientation and are therefore subject to an abundance of northern sunlight. These smaller
Ranges did not possess as many water-gaining sites and millipedes were scarcer. Invertebrate
40
surveys conducted on the intervening ‘flats’ between the Yilgarn BIF have not recorded A.
bamfordi (pers. comm. R. Teale). We also searched any water-gaining or densely vegetated
sites between BIFs and although these recovered other millipedes from the genus
Antichiropus, A. bamfordi was not found. As such, to the best of our knowledge, A. bamfordi
populations are restricted to BIF habitat and are not presently connected by populations in
intervening habitat. The field study did not involve any endangered or protected species and
none of the field sites from which we sampled (i.e. excluding KOO and WIN) required access
permission, with the exception of Mt Jackson (permission required from Cliffs Natural
Resources) and Mt Manning (Regulation 4 permit required from the Western Australian
Department of Parks and Wildlife (DPaW)). A license to take fauna for scientific purposes was
obtained prior to sampling from DPaW. GPS coordinates of the sampling locations are
provided in the supporting information (Appendix Table A1).
In total, 77 individuals from five BIFs were sequenced at two mitochondrial regions (77 at 16S,
75 at CO1) and genotyped at 11 nuclear microsatellite loci. DNA was extracted from the legs or
from the crushed heads of juvenile specimens stored in 100 % ethanol using a Qiagen DNeasy
Kit or a salt-based DNA extraction protocol (Beveridge & Simmons, 2004). Elutions were
performed in 100 µl or 50 µl of TE buffer respectively. Legs and head tissue were homogenized
in the lysis buffer prior to addition of proteinase K using mini micropestles (Geneworks).
2.3.2 Mitochondrial DNA amplification and sequencing
We sequenced two partial regions of the mitochondria, the Cytochrome Oxidase Subunit 1
(CO1) and the large Ribosomal Subunit 16S. For amplification of CO1, we used the Folmer
primers LCO1490 (5’GGTCAACAAATCATAAAGATATTGG3’) and HCO2198
(5’TAAACTTCAGGGTGACCAAAAAATCA3’) (Folmer et al., 1994) and for 16S, the ‘universal’
41
primers 16Sar (5’CGCCTGTTTTTCAAAAACAT3’) and 16Sbr (5’CCGGTTTGAACTCAGATCATGT3’)
(Simon et al., 1994). We also amplified two nuclear gene regions, the Internal Transcribed
Spacer 1 (ITS-1) (Ji et al., 2003) and the 18S rRNA gene (Hillis & Dixon, 1991). ITS-1 sequenced
products showed evidence of site heterogeneity and were excluded from use, and 18S
sequences were monomorphic. Amplifications were performed in 25 µl reactions containing 1x
PCR buffer, 3 mM MgCl2, 200 µM dNTP’s, 0.4 µM of each forward and reverse primer, 0.25 µg
of Bovine Serum Albumin (Fisher Biotech), 0.3 U of Platinum Taq DNA Polymerase (Invitrogen)
and approximately 20 ng/µl DNA. The thermal cycler profile for both CO1 and 16S involved an
initial denaturation at 94°C for 3 min, followed by 35 cycles of 94°C for 30 sec, 46°C for 30 sec
and 72°C for 30 sec followed by a final extension at 72°C for 2 min. PCR products were purified
using an UltraClean PCR Cleanup kit (Mo-Bio technologies, Geneworks) and the cleaned
product was quantified using a Nanodrop (Nanodrop Technologies). Sequencing reactions
(1/8) were carried out in 10 µl volumes containing 1 µl of Big Dye Terminator, 1.5 µl of 5 x
sequencing buffer, 1 µl of 3.2 pmol primer and approximately 40 ng of purified PCR product.
Thermal cycling conditions for sequencing reactions involved an initial denaturation at 96°C for
2 min, followed by 25 cycles of 96°C for 10 sec, 46°C for 5 sec and 60°C for 4 min. Sequenced
products were cleaned prior to electrophoresis using a standard ethanol precipitation method
(Applied Biosystems). Products were sequenced on an Applied Biosystems 3730 capillary
sequencer (Murdoch University).
2.3.3 Mitochondrial DNA analysis
Sequence data were aligned using Clustal W in the program BioEdit [26] and finalised by eye
with all mutations checked against raw sequence data. We used AMOVA, implemented in
Arlequin v. 3.5 (Excoffier & Lischer, 2010) to determine the structuring of mitochondrial DNA
diversity amongst BIF populations using the number of pairwise base-pair differences and
tested for significance with 10000 permutations. Bayesian analysis was used to visualise
42
genetic relationships and to estimate coalescence times of the populations in BEAST v. 1.7.5
(Drummond et al., 2012). Data were partitioned into the CO1 and 16S datasets, enforcing
monophyly for all A. bamfordi samples, with a SRD06 codon-partitioned model of sequence
evolution applied to the CO1 dataset and an HKY+G model to the 16S partition. Due to an
absence of fossil data and geological calibration points, estimates of population coalescence
times were generated by applying the average arthropod mutation rate of 2.3% per million
years to the partitioned data set (Brower, 1994). We applied a strict clock to both partitions
and used a Coalescent-Constant Size tree prior for four independent runs of 40 million
generations, sampling every 1000 generations. Log files were combined in Log Combiner and
the first 25% discarded as burnin. Visual inspection of the traces and high ESS values (>200)
confirmed stationarity had been reached in Tracer (BEAST package). Tree files were combined
in LogCombiner and the final file annotated in TreeAnnotater (BEAST package).
To distinguish between incomplete lineage sorting and migration for the two populations that
were paraphyletic (see results), we performed nested model likelihood ratio tests of various
isolation versus migration models in ‘L mode’ in the program IMa2 (Hey, 2010). IMa2 uses
Markov chain Monte Carlo simulations of gene genealogies to estimate divergence times of
populations and effective population sizes of extant and ancestral populations (Hey & Nielsen,
2004). To assess whether migration or drift was responsible for the paraphyly observed in the
two populations we performed a run on the two populations in ‘M mode’ using the following
command line priors: q200, t10, m1, b100000, l1000000, hfg, hn6, ha0.96, hb0.9. Convergence
was assessed by monitoring ESS values (all >20 000), chain swap rates (>80%) and running the
same command line four times independently using different seed values to ensure
consistency of results. The output was then used to test the likelihood of 24 different
migration/isolation models outlined in Hey and Nielsen (Hey & Nielsen, 2007) using the
following command line: q200, t10, m1, r0, c2 b100000 l1000000. Likelihood ratio tests were
43
used to reject models that were significantly worse than the full model, and the remaining
models were compared using the AIC criterion (AIC = -2(log(P) + 2K)) to identify the model of
best fit, where K is the number of parameters in the statistical model and L is the maximised
value of the likelihood function (Akaike, 1974).
To assess the historical stability of BIF population sizes we performed tests of neutrality that
can also serve as a measure of demographic expansion or contraction. Estimates of Tajima’s D
were calculated in Arlequin and diversity statistics including haplotype diversity, nucleotide
diversity (Pi), and nucleotide divergence (Jukes and Cantor) between populations were
calculated in DnaSP v. 5.10 (Librado & Rozas, 2009).
To test for isolation by distance within BIF populations using the combined sequence data, we
conducted mantel tests in R using the Vegan package (R Core Team, 2013). Permutations were
performed 10000 times on matrices of the number of pairwise sequence differences and
geographic distance. The R statistic is based on Pearson’s product-moment correlation
coefficient. We plotted the number of pairwise differences on geographic distance and
established a linear relationship, meaning no transformations of the data were required
2.3.4 Nuclear microsatellite genotyping and data analysis
Genotyping of the 11 microsatellite markers was carried out according to Nistelberger et al.,
(2013a). We checked all pairwise combinations of loci for linkage disequilibrium by performing
exact tests in Genepop v.1.2 (Raymond & Rousset, 1995) and applying a sequential Bonferroni
correction to determine significance (Rice, 1989). There was no significant association among
loci following sequential Bonferroni correction. Traditional F statistics were of limited use in
our study due to the likelihood that our BIF ‘populations’ do not represent true panmictic
44
populations given the distances over which individuals were sampled and the limited dispersal
capabilities of A. bamfordi. This was confirmed, with all populations showing significant
deviation from Hardy-Weinberg equilibrium, with the exception of Windarling, where
individuals had been sampled in close proximity to one another. Rarefied allelic richness
provided a better estimator of genetic diversity, given the sampling, and was calculated in HP-
Rare v 1.0 (Kalinowski, 2005). The partitioning of genetic diversity amongst BIF populations
was assessed using an Analysis of Molecular Variance (AMOVA) implemented in Arlequin, with
significance determined with 10000 permutations. An analysis of genetic structure across the
five BIF populations was assessed using a Bayesian clustering algorithm implemented in TESS
ver. 2.3 (Chen et al., 2007). This program assumes proximate interactions between individuals,
with spatial information prescribed at the individual level. We ran TESS with no admixture,
owing to the limited dispersal capabilities of millipedes and performed 100 iterations of K
values between 1 and 7, with a burnin length of 20000, 100000 sweeps and a spatial
interaction parameter of 0.8. Lowest DIC values for each K value were plotted in order to
determine the optimal K value. The results were then used in the downstream program
CLUMPP (Jakobsson & Rosenberg, 2007), which aligns cluster assignment across the replicate
analyses and ensures the correct assignment of K. Structure output graphics were created
using DISTRUCT (Rosenberg, 2004).
To further reconstruct the phylogenetic relationships of populations using microsatellite data,
we generated Cavalli-Sforza chord distances (DC) (Cavalli-Sforza & Edwards, 1967) between
populations in the program Populations 1.2.30 (Langella, 2000) and drew a neighbour-joining
tree of populations after 100 bootstrap replications in TreeView (Page, 1996). This geometric
distance measure does not make any biological assumptions about the data and can be more
useful for phylogenetic inference than estimates that use stepwise mutation models when
small population sizes are involved (Cavalli-Sforza & Edwards, 1967; Goldstein & Pollock,
45
1997). To assess isolation by distance within BIF populations using microsatellite data, we
correlated pairwise population matrices of genetic distance and geographic distance and
tested for significance using a Mantel test with 9999 permutations in the program IBDWS v
3.23 (Jensen et al., 2005).
2.4 RESULTS
2.4.1 Genetic variation
2.4.1.1 Mitochondrial DNA data
Of the two mitochondrial regions amplified, CO1 was the most variable, with 45 haplotypes
identified from the 75 individuals sequenced (two WIN samples did not amplify at CO1). The
variation was composed of 54 transitions, eight transversions and two multistate substitutions.
The 16S region yielded 24 haplotypes from the 77 individuals sequenced and included 21
transitions, one transversion and one single base pair indel. Combining the two regions
resulted in a total aligned sequence length of 1114 bp and 48 haplotypes. The most prevalent
haplotype (Hap 47 WIN) had a frequency of 23 %, followed by Hap 35 (MTJ) and Hap 39 (HA)
with frequencies of 8%, and Hap 33 (MTJ) and Hap 40 (HA) with 6%. The remaining haplotypes
were only found in one or two individuals. Measures of genetic diversity were consistently
lowest in the Windarling population and highest in Koolyanobbing. Haplotype diversity ranged
from 0.295 (WIN) to 0.986 (KOO), with an average of 0.794. Nucleotide diversity ranged from
0.06 % (WIN) to 8.5 % (KOO), with an average of 4%. No populations departed from the neutral
model of evolution except for Windarling (D = -1.775) (Table 2.1). Measures of diversity in
Windarling are likely to be biased due to restricted sampling. Millipedes had been previously
collected from two closely located sites on this BIF (<260 m apart), and although noted to be
rare along the length of the Range (pers. comm. R. Teale), further sampling was not possible
due to restricted access associated with mining activity.
46
There was a large and highly significant degree of genetic structuring of the mitochondrial
genetic diversity, with most variation (AMOVA) occurring among BIF populations (71.56%)
rather than within (28.4%). The greatest genetic divergence occurred between the Mt Jackson
and Helena Aurora populations (approx. 40 km apart) and the least between the two most
spatially isolated populations, Koolyanobbing and Die Hardy (approx. 100 km apart) (Table 2.2,
Fig.2.1).
2.4.1.2 Nuclear microsatellite data
Microsatellite genetic diversity was low, both within and amongst BIF populations. Rarefied
allelic richness ranged from 1.89 (HA) to 3.28 (KOO), with an average of 2.47. There was a
highly significant degree of structuring of the microsatellite diversity, with the majority of
variation (68.11%) maintained within populations, compared to (31.89%) among populations.
47
Table 2.1 Diversity statistics for the five Atelomastix bamfordi BIF populations for mitochondrial DNA data and nuclear microsatellite data (bold), standard
error in parentheses: N, number of individuals; # haps, number of haplotypes; HD, haplotype diversity; Pi, nucleotide diversity; D, Tajima’s D; n, average
number of individuals genotyped; AR^, rarefied allelic richness; * P value<0.05
Population # Haplotype Genbank accession #
Population abbrev. N haps HD Pi D numbers CO1 16S n AR^
Die Hardy Range DH 16 14 0.983 (0.002) 0.0050 (0.000) -0.569 19-32 KC689871-KC689883 KC689837-KC689841 13.7 2.62 (0.34)
Helena Aurora Range HA 14 8 0.890 (0.004) 0.0042 (0.000) -0.733 38-45 KC689889-KC689894 KC689845-KC689851 13.9 1.89 (0.28)
Koolyanobbing Range KOO 21 18 0.986 (0.001) 0.0085 (0.000) -0.316 1-18 KC689853-KC689870 KC689829-KC689836 21.7 3.28 (0.40)
Mt Jackson MTJ 11 5 0.818 (0.008) 0.0017 (0.000) -0.234 33-37 KC689884-KC689888 KC689842-KC689844 9.6 2.57 (0.36)
Windarling Range WIN 13 3 0.295 (0.012) 0.0006 (0.000) -1.775* 46-48 KC689895-KC689897 KC689852 14 1.97 (0.27)
48
Table 2.2 Pairwise population nucleotide divergence based on Jukes and Cantor of five Atelomastix bamfordi populations using mtDNA data. DH, Die Hardy; HA, Helena Aurora;
MTJ, Mt Jackson; Win, Windarling; KOO, Koolyanobbing.
DH HA MTJ WIN KOO
DH 0
HA 0.016 0
MTJ 0.017 0.022 0
WIN 0.013 0.018 0.019 0
KOO 0.013 0.014 0.018 0.017 0
49
Figure 2.1 Schematic representation of the five Banded Iron Formation populations, coloured according to the five genetic clades identified with mtDNA analysis. Black circles indicate sample sites. Study area shown in inset.
2.4.2 Phylogeographic structure
There was evidence of strong phylogeographic structure within A. bamfordi, with minimal
sharing of CO1 and 16S haplotypes amongst individuals within BIF populations and no sharing
of haplotypes between populations. Bayesian analysis identified five genetic clades in which
populations were reciprocally monophyletic, with the exception of Koolyanobbing and Helena
Aurora, for which some samples were paraphyletic (Fig. 2.2). This was due to a subset of six
individuals from the north of the Koolyanobbing Range being grouped with samples from
Helena Aurora (Fig.2.1). Support for the five genetic clades was high (posterior probabilities >
0.85) although the relationships between the clades was poorly supported. Dating estimates
50
showed coalescence during the Pleistocene, between 0.729 Ma [0.532-0.952] and 1.078 Ma
[0.792-1.433] (Fig.2.2).
Isolation and migration analysis in the program IMa2 was used to determine whether the
paraphyly between Koolyanobbing and Helena Aurora was due to incomplete lineage sorting
(isolation) or gene flow (migration). The three models with the highest AIC support all
indicated there was no historical migration between the two populations (Appendix Table A2).
Mantel tests of pairwise sequence differences against geographic distance indicated significant
isolation by distance within the Koolyanobbing (r=0.740 p=0.0001), Mt Jackson (r=0.225
p=0.0375) and Helena Aurora (r=0.399 p=0.0025) populations.
2.4.3 Population genetic structure
Bayesian cluster analysis identified four genetic clusters in the microsatellite data. The first
corresponded to the Die Hardy and Mt Jackson populations, and the remaining clusters to each
of the three other populations (HA, WIN, KOO) (Fig. 2.3). Further investigation of structure in
the first cluster revealed two more groupings corresponding to the separate BIF populations
(DH, MTJ).
51
Figure 2.2 Bayesian cladogram of 75 Atelomastix bamfordi individuals based on combined CO1 and 16S mtDNA data. Clades are designated separate colours as per Fig. 2.1. Numbers on terminals indicate haplotype number. The * denotes major nodes with high support (posterior probability > 0.85) and the numbered node indicates divergence estimate in millions of years, with confidence intervals in brackets. NB the A. nigrescens branch has been truncated (//) for viewing purposes.
52
There was some evidence of admixture in the Mt Jackson and Koolyanobbing populations (Fig.
2.3). Phylogenetic analysis of relationships between populations using Cavalli-Sforza chord
distances, showed distinct genetic structure amongst populations using microsatellite data,
but with some clustering of Koolyanobbing and Helena Aurora (97% bootstrap support) and
Die Hardy and Mt Jackson (62%) (Fig. 2.4). Two of the five populations, Helena Aurora (r=0.30
p=0.004) and Die Hardy (r=0.32 p=0.009), showed weak but significant patterns of isolation-by-
distance at microsatellite loci following Mantel tests.
Figure 2.3 Bayesian population assignment of Atelomastix bamfordi individuals based on nuclear microsatellite data generated in TESS for the optimal grouping of K=4. Each vertical line represents an individual and the colour displays the probability of belonging to each genetic cluster. DH, Die Hardy; MTJ, Mt Jackson; HA, Helena Aurora; WIN, Windarling; KOO, Koolyanobbing.
53
Figure 2.4 Unrooted neighbour-joining tree of Cavalli-Sforza chord distances between the five BIF populations of Atelomastix bamfordi. Distances are based on 11 microsatellite loci with bootstrap support shown at nodes. DH, Die Hardy; MTJ, Mt Jackson; HA, Helena Aurora; WIN, Windarling; KOO, Koolyanobbing.
2.5 DISCUSSION
2.5.1 Genetic differentiation of BIF populations
2.5.1.1 Spatial patterns
Phylogeographic analysis of A. bamfordi revealed strong genetic differentiation between BIF
populations, with mtDNA genetic lineages corresponding to each of the five BIF populations,
except for six individuals from Koolyanobbing that were paraphyletic with the Helena Aurora
Range due to incomplete lineage sorting. Microsatellite data were broadly consistent with this
result, identifying four genetic clusters, the first of which was further broken down into
individual populations.
The lack of correlation between genetic divergence and geographic distance, with the two
most spatially isolated populations showing the least genetic divergence, suggests there has
54
been little, to no connectivity of BIF populations following their separation. This result was
emphasized in the Bayesian analysis which showed high support for genetic clades but poor
support for the relationships between clades (Fig. 2.2). It is also consistent with estimates of
sequence neutrality (D) that indicated a lack of demographic expansion and contraction
events, with the exception of Windarling, where the result can be explained by close sample
proximity. This pattern of differentiation has been observed in other taxa with restricted
dispersal capabilities, where populations have become isolated as a response to climate
change and habitat loss. Like A. bamfordi, these studies emphasise the role of genetic drift in
driving population differentiation. For example, mitochondrial and nuclear sequence data
revealed extraordinarily strong phylogeographic structure in populations of trapdoor spiders
(Moggridgea tingle) in south-western Australia, often situated just kilometres apart.
Molecular dating suggested these populations became isolated in the mesic valleys and
uplands of the Porongorup and Stirling Ranges following the onset of aridity in the late
Miocene (Cooper et al., 2011). Like A. bamfordi, reciprocally monophyletic lineages across
three geographic locations highlighted the reproductive isolation of M. tingle populations, but
the pattern emphasized a more ancient division, with monophyly observed in both mtDNA and
the more slowly evolving nuclear genome (Cooper et al., 2011). Studies of genetic structure in
two frog species in south-western Australia, Geocrinia lutea and G. rosea, also described
patterns of extremely low dispersal and long-term isolation between habitat ‘islands’, with
divergence attributed to isolation via range changes resulting from climate change rather than
geographical barriers (Driscoll, 1998). Another striking example of genetic divergence across
terrestrial islands occurred in the sky islands of south-eastern Arizona - forested mountain
Ranges, separated by low-lying grassland and desert. Here, the jumping spider Habronattus
pugillis occupies woodland habitat at altitudes above 1400 m but is absent from the lowlands.
Using mtDNA, Masta (Masta, 2000) showed that like A. bamfordi, smaller populations of H.
pugillis formed monophyletic clades, whereas larger Ranges showed evidence of paraphyly.
Like A. bamfordi, this was ascribed to incomplete lineage sorting rather than resulting from
55
migration (Masta, 2000). The isolation of populations was attributed to climate-induced
contraction of woodland populations onto disjunct mountain Ranges.
2.5.1.2 Temporal patterns
Molecular dating with the arthropod clock rate has placed the separation of A. bamfordi clades
at 0.7 to 1.1 million years ago during the Pleistocene. The timing coincides with known
increases in aridity in the transitional rainfall zone (TRZ) due to climatic fluctuations associated
with Milankovitch cycles during this period (Bowler, 1982). In particular, there was a major arid
shift in southern Australia corresponding to the middle Pleistocene transition (1250 ka - 700
ka), a period of increased severity of glaciation, associated with greater ice volumes, lowered
temperatures and increased aridity (Clark et al., 2006). Although the TRZ remained un-
glaciated, this period of intense climatic instability is likely to have caused the fragmentation of
species distributions due to changes in suitable habitat within this region (Byrne, 2007).
Although the divergence estimates presented here must be treated with caution given the use
of a generalised arthropod mutation rate, the errors associated with this method are lessened
given we are working at shallow phylogenetic levels (Buckley et al., 2001; Pons et al., 2010).
2.5.2 BIF population dynamics – historical and contemporary patterns
2.5.2.1 Historical patterns
The process of genetic drift acting independently on each BIF population, is reflected in
different patterns of genetic divergence and diversity seen in the mtDNA data. The smaller
populations, Mt Jackson and Windarling, and the larger population Helena Aurora, which had
less suitable millipede habitat, showed lower genetic diversity and higher levels of genetic
divergence than the larger populations (KOO and DH). These three BIFs have probably
harboured historically small effective population sizes, owing to the limited availability of
56
moist sites, increasing their susceptibility to forces of genetic drift driving divergence (Wright,
1929). They may also be influenced by different selective forces associated with the more arid
conditions present on these BIFs. On the other hand, the two large BIF populations
characterised by a greater availability of suitable habitat (KOO and DH), retained genetic
diversity and had less genetic divergence, consistent with historically large population sizes
(Wright, 1929). The paraphyly of individuals from Koolyanobbing with those of Helena Aurora
further supports the likelihood of a large historical population size at Koolyanobbing, as larger
populations take longer to achieve reciprocal monophyly through lineage sorting. Over time
this signature is likely to erode, leaving the five BIFs reciprocally monophyletic.
2.5.2.2 Contemporary patterns
Rapidly evolving microsatellite markers can provide a more contemporary view of BIF
population dynamics. The information we gained from these markers was limited by a lack of
Hardy-Weinberg equilibrium within our BIF populations, yet we still identified patterns in
genetic diversity that appear to support the impact of genetic drift on populations: lower
genetic diversity in BIFs likely to have had small effective population sizes (WIN, MTJ and HA).
Signals of isolation by distance (IBD) were seen in some of the BIF populations (HA, DH), but
these were weak and likely to be heavily influenced by limited sample sizes. The small pairwise
distances between samples at Windarling explains the lack of IBD seen in this BIF, but we were
surprised to see evidence of IBD in the Helena Aurora and Mt Jackson populations. Here, we
expected limited gene flow to result in a pattern of ‘islands within islands’, where significant
isolation has resulted in genetically divergent populations within BIFs that do not correlate
with geographic distance. An increased number of samples from more locations would be
required to determine whether IBD is really occurring at these spatial scales.
57
2.5.3 Conservation implications
The evolutionary significance of geographically isolated genetic lineages has long been
recognised, and conservation biologists have attempted to formalise this by delineating
conservation units within a species, such as Evolutionary Significant Units (ESUs) and
Management Units (MUs) (Ryder, 1986). Conservation units have been defined at various
levels of genetic differentiation within species and are likely to experience limited gene flow or
even total reproductive isolation. For example, ESUs have been defined as a population or a
group of populations that warrant separate management for conservation because of high
genetic and ecological distinctiveness (Allendorf et al., 2013), with Moritz (1994) suggesting
that ESUs might be characterised by a pattern of reciprocal monophyly amongst populations.
Applying this criterion to our mtDNA data, results in the recognition of three BIF populations
(DH, MTJ, WIN) as ESUs. However, Paetkau (1999) and Crandall et al. (2000) highlight the
problems inherent in ignoring paraphyletic lineages, such as Koolyanobbing and Helena
Aurora, which are often ancestral and maintain high levels of diversity and divergence within
themselves. In our case, the paraphyly of Koolyanobbing with Helena Aurora is a result of a
slower progression towards monophyly due to ongoing lineage sorting, probably owing to
large population sizes. Given the clear evidence for restricted gene flow between these two
populations, both could readily be recognised as MUs. Therefore we consider that all five BIF
populations warrant designation as conservation units.
Although it is difficult to predict future climate change at such fine scales, broad-scale models
for this region over the 21st century indicate increasing temperature, decreasing rainfall and a
greater incidence of fire (Hughes, 2003). If these predictions are correct, populations are
unlikely to re-connect via suitable intervening habitat in the future, further consolidating their
isolation to the remaining moist pockets provided by BIF, where they will continue to be
58
influenced by forces of genetic drift and local selection. Given their dependence on moisture,
conservation efforts should focus on protecting areas of dense leaf litter, large outcrops and
shaded gullies on each BIF in order to preserve the five genetic lineages identified here. To
achieve this, preservation of the height and topographical complexity of BIFs is essential, in
order to maintain drainage lines and water-collecting sites. Conservation of these areas is also
likely to support populations of other relict fauna and flora that are adapted to moist
environments.
2.5.4 Conclusion
This is the first phylogeographic study conducted on a species inhabiting multiple Banded Iron
Formations in this landscape and provides evidence that the disjunct populations have
experienced long-term isolation following their separation during the Pleistocene. The
divergence of populations during this period coincides with increases in aridity and emphasizes
the role of climatic instability, along with topographic features in driving differentiation and,
ultimately speciation in this landscape. We consider the strong phylogeographic structure and
evidence of restricted gene flow between BIF populations warrant the designation of all five
populations as distinct conservation units.
59
Attractive,
interesting…what more
could a biologist ask for?
60
Work in this chapter has been submitted for publication and published as:
Nistelberger HM, Byrne M, Coates DJ and Roberts JD (2014) Terrestrial islands in an ancient
landscape: Evidence of both historical isolation and connectivity of populations of Grevillea
georgeana, a banded iron formation endemic. Submitted: Journal of Biogeography.
Nistelberger HM, Coates DJ, Byrne M and Roberts, J D (2013) Isolation and characterisation of
11 microsatellite loci in the short-range endemic shrub Grevillea georgeana McGill
(Proteaceae). Conservation Genetics Resources, 6, 221-222. (Appendix D)
61
Chapter 3
Terrestrial islands in an ancient landscape: Evidence of both historical
isolation and connectivity of populations of Grevillea georgeana, a
Banded Iron Formation endemic.
3.1 ABSTRACT
Plant species with island distributions might be expected to have restricted seed and pollen
dispersal with significant genetic divergence among populations. In this study we assessed
population genetic connectivity through patterns of genetic diversity and structure in seven
terrestrial island populations of Grevillea georgeana. Phylogeographic pattern and population
genetics of seven Yilgarn Banded Iron Formation (BIF) populations were examined using
sequence data from two chloroplast intergenic spacers and 10 nuclear microsatellite loci.
Phylogenetic relationships of 67 individuals were assessed using both Bayesian and statistical
parsimony methods, and genetic divergence of haplotypes was dated using two angiosperm
chloroplast substitution rates. Microsatellite variation was assessed in geographically clustered
groups of 24 individuals from each BIF, with the exception of Die Hardy Range where three
groups of 24 individuals were sampled to infer patterns within a large range. Pollen movement
was assessed using indirect methods based on the average frequency of private alleles at
microsatellite loci. The results show that different historical processes have shaped the
evolution of Grevillea georgeana populations. Some populations are behaving as isolated
islands, with genetic drift driving patterns of differentiation, whilst others show patterns of
more recent connectivity. Pollen movement, although generally restricted, appears to be an
important process connecting neighbouring populations that should be considered in the face
of future habitat disturbance.
62
3.2 INTRODUCTION
The evolutionary consequences of island distributions have long been of interest to biologists.
Species with disjunct, island populations may experience restricted gene flow that over time,
leads to genetic divergence as population specific processes, such as genetic drift, mutation,
bottlenecks and natural selection, alter genetic structure (Wilson & MacArthur, 1967;
Dobzhansky, 1970; Mayr, 1970; Whittaker, 1998). If genetic isolation is prolonged, populations
follow independent evolutionary trajectories that can result in reproductive barriers and
allopatric speciation (Dobzhansky, 1970; Mayr, 1970).
Oceanic and continental islands generally have well defined boundaries to dispersal, with
pronounced genetic isolation leading to patterns of deep genetic divergence (Beheregaray et
al., 2003; Juan et al., 2004; Measey et al., 2007; Zhai et al., 2012). In comparison, terrestrial or
habitat islands have less obvious barriers to dispersal and the degree of genetic isolation
among populations may be difficult to predict (Masta, 2000; Robin et al., 2010). In these
situations, evidence of long-term isolation or connectivity between habitat islands can be
inferred from phylogeographic and population genetic studies (Masta, 2000; Floyd et al., 2005;
Byrne & Hopper, 2008; Cooper et al., 2011).
The degree of connectivity or isolation of island populations is dependent on dispersal
mechanisms that in plants, includes movement of both seed and pollen (Epperson, 1993).
Distinguishing between these two modes of dispersal requires the use of both nuclear (nrDNA)
and chloroplast DNA (cpDNA) markers. CpDNA is generally maternally inherited in
angiosperms, reflecting patterns of seed dispersal (McCauley et al., 2007) and provides a good
indicator of historical relationships owing to its uni-parental inheritance, low mutation rate
and lack of recombination (Avise, 2000). Nuclear markers are bi-parentally inherited, providing
63
a means of estimating gene flow through both pollen and seed dispersal (Ennos, 1994). At the
population level, nrDNA sequence markers may lack sufficient variation to provide
phylogeographic information (Zhang & Hewitt, 2003), but rapidly mutating neutral markers,
such as microsatellites, can be used to provide a nuclear perspective on patterns of population
diversity and structure caused by more recent evolutionary processes (Zhang & Hewitt, 2003).
When used together, both cpDNA and nrDNA genetic markers allow for inference of historical
phylogeographic patterns relating to seed movement, as well as more recent population
processes, including pollen dispersal between island populations of plant species.
The Yilgarn Banded Iron Formations (BIFs) of Western Australia are topographical features
characterised by high species richness and endemism that behave as terrestrial islands in a
semi-arid landscape (Gibson et al., 2012). Several endemic species occur across multiple
formations without inhabiting the intervening landscape and these provide an excellent
system for phylogeographic and population genetic analysis to assess patterns of connectivity
among these terrestrial island formations (Gibson et al., 2007). Grevillea georgeana McGill is a
bird-pollinated perennial shrub (2 - 2.5 m height) that occurs across multiple BIFs, but not in
the surrounding sand and alluvial plains (McGillivray & Makinson, 1993; Gibson et al., 2007).
The geographic distance between the island populations ranges from approximately 7 km to
115 km and it is not known whether populations are connected via gene flow or if they are
reproductively isolated. Seeds in G.georgeana have no active mechanism of dispersal and the
species regenerates from seed following fire (Olde & Marriott, 1995). Given the lack of
dispersal mechanism we hypothesise little connectivity of BIF populations via seed. If disjunct
populations have been separated for long periods, this may result in population differentiation
of cpDNA that is not geographically structured, with haplotypes restricted to particular
populations (Lee, 2000). If there has been connectivity, either via movement of seed between
BIFs or via the connection of current populations through now extinct populations, we would
64
expect shallow genetic differentiation that is potentially geographically structured, with
sharing of haplotypes between adjacent BIF populations (Kimura & Weiss, 1964b).
Patterns in nuclear microsatellite data will reflect contemporary pollen movement across the
landscape as well as seed dispersal (Sork et al., 1999). Bird pollination can promote or diminish
genetic structure depending on the species and their foraging habits, leading to difficulties in
predicting pollen flow and nuclear genetic structure (Degen et al., 2004; Llorens, 2004; Byrne
et al., 2007). Depending on the movement of pollinators, we can expect one of three potential
outcomes. An absence of pollen flow between BIFs would lead to a pattern of genetic
differentiation that is not geographically structured, with genetic drift driving patterns of
random genetic structure. If pollen movement is restricted to neighbouring BIFs then a pattern
of isolation by distance (IBD) will develop (Wright, 1943), with less genetic divergence among
adjacent BIFs. Finally, if pollen movement is unrestricted across all BIF populations, we would
expect a lack of genetic structure. In evaluating the influence of historical and contemporary
processes on relationships among G. georgeana populations, we expect to see genetic
signatures of differentiation in cpDNA arising from limited seed dispersal, and some structure
amongst BIF populations in the nrDNA, depending on the level of pollen dispersal.
Understanding the patterns of genetic differentiation in terrestrial island species, such as those
endemic to BIF, not only provides valuable information on the evolution and consequences of
species inhabiting island distributions, but also informs conservation of these terrestrial island
habitats. Species with small populations and disjunct distributions may be at heightened risk of
extinction, either from demographic effects associated with the increased impact of
disturbance or habitat loss, or from reduced genetic diversity and inbreeding depression
(Frankham et al., 2002). Understanding levels of connectivity via gene flow will be critical in
assessing the impact of reduced genetic diversity and potential inbreeding depression.
65
Evaluating the extent of genetic relationships among populations or groups of populations is
also important, as high genetic differentiation could require the recognition of distinct units for
conservation (Allendorf et al 2013). This is particularly important in the case of G. georgeana
as many of the Yilgarn BIFs are prospective for resource development, with potential for
extensive habitat removal and disturbance.
We used a combination of non-coding cpDNA sequences and nuclear microsatellite markers
specifically developed for G. georgeana in order to test our hypotheses through (i) assessment
of historical connectivity of BIF populations via seed and pollen dispersal, and (ii) assessment
of patterns of genetic variation within BIF populations to infer population specific dynamics
and evolutionary history.
3.3 MATERIALS AND METHODS
3.3.1 Sampling and DNA extraction
Samples were collected from seven Yilgarn BIFs, and voucher specimens collected from each
population are lodged at the Western Australian Herbarium. The BIFs vary in size and
accessibility, so distances between individuals sampled varied considerably. Sampling for the
phylogeographic (cpDNA) component of the study (10 individuals) was spaced out along each
BIF where possible, in order to capture the greatest possible spatial haplotype diversity. To
determine population differentiation using microsatellite markers, a geographically clustered
group of 24 individuals was collected from a site where the species was common. For the large
Die Hardy Range, we collected from three sites (north, central and south) to allow assessment
of within BIF genetic structure. DNA was extracted from freeze-dried (Labconco) leaves using a
modified 2% CTAB method as outlined in Nistelberger et al. (2013).
66
3.3.2 cpDNA Sequencing
We screened six non-coding cpDNA regions known to detect phylogeographic structure in
Australian plants (Byrne & Hankinson, 2012) and selected the two most variable, trnQ-rps16
(primers trnQ, rps16) and psbD-trnT (primers psbD, trnT)(Shaw et al., 2007). PCR reactions (50
µl) contained 50 mM KCl, 20 mM Tris-HCl (pH 8.4), 0.2 mM each dNTP, 0.1 µM primers, 1.5 and
2 mM MgCl2 (psbD-trnT, trnQ-rps16 respectively) and 0.1 U of Taq polymerase (Invitrogen).
The amplification protocol included a hot start of 5 min at 80°C, followed by 30 cycles of 95°C
for 1 min, 50°C for 1 min and 65°C for 4 min, with a final extension of 65°C for 5 min. PCR
products were cleaned and sequenced by AGRF-Perth (Australian Genome Research Facility)
on an Applied Biosystems 3730x1 DNA Analyzer platform. DNA sequence data were aligned
and edited using ClustalW in BioEdit (Hall, 1999) and checked by eye. The two sequences were
combined to form a total aligned sequence length of 2777 base pairs.
3.3.3 Fragment Analysis
For microsatellite analysis, multiplexed PCR reactions were performed in a total volume of 5µl
as outlined in Nistelberger et al. (2013). Amplified products were run on a 3% agarose gel and
eleven primer pairs were selected based on clear, interpretable banding pattern. PCR products
were visualised on an Applied Biosystems 3730 capillary sequencer (Murdoch University) and
scored using GeneMapper v 4.0 (Applied Biosystems).
3.3.4 cpDNA data analysis
Descriptive statistics, including the number of haplotypes, haplotype diversity, nucleotide
diversity and population divergence (Jukes & Cantor, 1969), were calculated in DNAsp v5.10
(Librado & Rozas, 2009). Partitioning of chloroplast diversity was determined with an analysis
of molecular variance (AMOVA) using the number of pairwise base-pair differences and testing
67
for significance with 10,000 permutations in Arlequin v.3.5 (Excoffier & Lischer, 2010).
Historical stability of populations was assessed with tests of neutrality (Tajima’s D), that also
serve as indicators of population expansion and contraction in Arlequin, with significance
tested with 10,000 permutations (Tajima, 1989). These analyses were performed using data
with indels manually encoded as transitions. We inferred the evolutionary relationships of
haplotypes and populations using a Bayesian analysis implemented in MrBayes v3.1
(Huelsenbeck & Ronquist, 2001). Data were partitioned with indels coded according to
Simmons & Ochoterena (2000) as implemented in Seqstate v.1.4.1 (Muller, 2005). An HKY+I+G
nucleotide substitution model was used as determined by Jmodeltest v0.1.1 (Posada, 2008)
and the indel partition was set to nst=1. We ran the Markov Chain Monte Carlo method for
four million generations, sampling from the posterior distribution every 1000 generations. The
first 25% of trees were discarded as burnin. The time to most recent common ancestor
(TMRCA) of haplotypes was estimated using two nucleotide substitution rates commonly
reported for angiosperms, 1.2x10-9 (Graur & Li, 2000) and 2.0x10-9 substitutions per site per
year (Wolfe et al., 1987) in BEAST v.1.7 (Drummond et al., 2012). We used an uncorrelated,
log-normal clock with a coalescent-constant size tree prior to estimate divergence times. Two
independent runs of 10 million generations were conducted, sampling every 1000 generations.
Stationarity was assessed by viewing log files in TRACER and ensuring effective samples sizes
for all parameters exceeded 200. Tree files were combined using LOGCOMBINER and
annotated in TREEANNOTATER (available BEAST package) with the first 25% of trees discarded
as burnin. Haplotype relationships were also visualised with an unrooted haplotype network
based on statistical parsimony analysis, implemented in TCS v.1.21 (Clement et al., 2000).
To assess whether haplotypes show IBD across the landscape we conducted Mantel tests in R
v. 3 (Vegan package) on the number of pairwise sequence differences and geographic distance,
permuted 10,000 times (R Core Team, 2013).
68
3.3.5 Microsatellite data analysis
We tested for the presence of null alleles using a maximum likelihood method implemented in
ML-Null (Kalinowski & Taper, 2006). Tests for linkage disequilibrium and Hardy-Weinberg
equilibrium were conducted using FSTAT v.2.9.3 (Goudet, 2001). Measures of population
genetic diversity and structure, as well as an AMOVA, were calculated using Arlequin.
Differences in expected heterozygosity and the number of alleles per population were tested
with a Friedman ANOVA (for multiple paired comparisons), with samples paired by locus.
We used the Bayesian assignment approach of TESS v.2.3 to determine the number of distinct
genetic units (K) present in the dataset (Chen et al., 2007). We performed a total of 50,000
sweeps, with 20,000 burnin and a spatial interaction prior of 0.8, and both admixture and no-
admixture models. The optimal value of K was determined by plotting the lowest DIC values of
each iteration and convergence was assessed using CLUMPP (Jakobsson & Rosenberg, 2007).
We also visualised the population relationships with principal co-ordinates analysis (PCoA) of
pairwise population estimates of Nei’s unbiased genetic distance (Nei, 1978), implemented in
Genalex v.6 (Peakall & Smouse, 2006). IBD across populations was tested for significance
(10,000 permutations) with a Mantel test Implemented in R (Vegan package). IBD within a BIF
range was also tested among individuals from the three Die Hardy Range populations.
Indirect measures of gene flow were calculated to provide an indication of the degree of
population connectivity or isolation (Slatkin & Barton, 1989) based on nine populations (ie.
including all three DH populations). The effective number of immigrants per generation (Nm)
was calculated according to Slatkin (1985), as implemented in Genepop v.1.2 (Raymond &
69
Rousset, 1995). This estimate must always be treated cautiously as it also assumes that
populations are large and stable, and that gene frequencies are not influenced by mutation or
selection (Slatkin, 1985).
We tested for bottleneck events using a Wilcoxon signed rank test for excess of observed
heterozygotes using BOTTLENECK v1.2.02 (Piry et al., 1999). Heterozygote excess was tested
under the Infinite Allele Model, the Stepwise Mutation Model and the Two Phase Model. For
the Two Phase Model we used settings typical for microsatellite markers with the frequency of
single step mutations set to 90% and the variance of those mutations as 12% (Busch et al.,
2007).
3.4 RESULTS
3.4.1 cpDNA diversity
The total combined cpDNA sequence alignment (2777 bp) generated 20 haplotypes from 67
individuals. Variation was comprised of seven transitions, 15 transversions and nine variable
length indels. The Die Hardy and Mt Manning populations were significantly more diverse than
other populations (Table 3.1). Mt Correll showed the least genetic diversity with just one
haplotype identified. Two of the haplotypes, H11 and H12, were shared across three
populations (FIN, HUN, TH) but all other haplotypes were specific to individual populations
(Fig. 3.1). The majority of haplotype diversity occurred between populations, but a relatively
large amount (32.65%) was maintained within populations. Genetic divergence of populations
was high, with the exception of the three populations sharing haplotypes. The spatially
isolated Mt Correll showed the greatest average genetic divergence from all other populations
(0.25%) (Table 3.2). No population departed from the neutral model of evolution (Table 3.1). A
Mantel test showed no pattern of isolation by distance of haplotypes across the landscape
(R=0.034, P=0.37).
70
Bayesian analysis of haplotypes resulted in a tree with a mostly unresolved polytomy (Fig. 3.2).
There was some geographic clustering of related haplotypes within populations, eg. H3, H8,
H10 (DH), and H13, H14 (HA), but many haplotypes from the Die Hardy and Mt Manning
ranges were distributed throughout the tree, indicating differentiated haplotypes present
within these populations, a pattern confirmed in the parsimony network (Fig. 3.3).
Interestingly, the haplotype present in the geographically isolated population of Mt Correll was
most closely related to a subset of haplotypes at Mt Manning over 90 km away to the east (Fig.
3.2 and 3.3).
The TMRCA of haplotypes ranged from 0.99 [0.56-1.57] to 1.6 [0.96-2.64] million years ago
respectively, indicating divergence during the Mid Pleistocene. The divergence of haplotypes
within populations was estimated to be 0.08-0.14 to 0.989-1.6 million years ago (Fig. 3.2).
3.4.2 Microsatellite data
One microsatellite marker, Gg16, showed a high frequency of null alleles (40%) and was
discarded. Three other markers (Gg5, Gg18, Gg29) showed null alleles at frequencies between
10 and 16 % and initial analyses with removal of these markers did not significantly alter
results so all 10 loci were retained for analyses. The 10 loci were in Hardy-Weinberg
equilibrium and there was no evidence of linkage disequilibrium following Bonferroni
correction. Genetic diversity was low to moderate (Table 3.3) and a Friedman test indicated a
highly significant difference amongst expected heterozygosity values (Friedman stat = 36.4,
p=0.0001) but no significant difference in the average number of alleles per locus.
71
Table 3.1 Diversity statistics for the seven Grevillea georgeana BIF populations assayed for chloroplast DNA variation: N, number of individuals; # Haps, number of
haplotypes; HD, haplotype diversity (std dvn); Pi, nucleotide diversity (std dvn).
Population Pop abbrev.
N # Haps HD Pi Tajima’s D
Mt Correll COR 10 1 0 0 na
Die Hardy Range DH 10 9 0.978 (0.05) 0.00188 (0.00028) -0.573
Mt Finnerty FIN 10 2 0.356 (0.16) 0.00013 (0.00006) 0.015
Helena Aurora Range HA 9 2 0.222 (0.17) 0.00017 (0.00013) -1.362
Hunt Range HUN 10 3 0.378 (0.18) 0.00044 (0.00028) -1.388
Mt Manning MM 9 5 0.833 (0.10) 0.00215 (0.00030) 1.343
Taipan Hill TH 9 2 0.500 (0.13) 0.00019 (0.00005) 0.987
Table 3.2 Estimates of pairwise divergence % (JC) based on chloroplast DNA variation amongst populations of Grevillea georgeana. The distances represent an average among individuals from each population.
COR DH FIN HA HUN MM
DH 0.29 FIN 0.25 0.12
HA 0.29 0.15 0.11 HUN 0.24 0.13 0.02 0.12
MM 0.19 0.21 0.17 0.20 0.17 TH 0.25 0.13 0.02 0.12 0.03 0.17
72
Figure 3.1 Map of the seven BIF Grevillea georgeana populations with pie charts showing the frequency and occurence of each cpDNA haplotype (legend). Only two haplotypes were shared amongst populations (H11 and H12). Image modified from Google Earth.
73
Figure 3.2 Bayesian tree of 20 Grevillea georgeana haplotypes based on combined trnQ-rps16 and psbD-trnT cpDNA data. Numbers in red reflect divergence estimates for the conservative mutation rate in millions of years. Numbers in black are posterior probabilities. COR, Mt Correll; DH, Die Hardy; FIN, Mt Finnerty; HA, Helena Aurora; HUN, Hunt; MM, Mt Manning; TH, Taipan Hill.
74
Figure 3.3 Maximum parsimony network for all Grevillea georgeana haplotypes (numbered) based on combined cpDNA data. Node size is proportional to the frequency of haplotypes with the high frequency Hap 11 representing 34% of individuals. Black dots indicate mutational steps. Colours identify the haplotype according to the key in Fig. 3.1.
75
Figure 3.4 Map of the seven Grevillea georgeana BIF populations with pie charts indicating the proportion of assignment of sampled populations to each of the six genetic clusters (colours) identified in the nrDNA microsatellites using TESS.
Estimates of population differentiation using FST were moderate to high, with an average value
of 0.174. All pairwise combinations were significant following Bonferroni correction, with the
exception of Die Hardy south and Hunt Range (FST = 0.026), and Die Hardy south and Die Hardy
central (FST = 0.031). The spatially isolated population at Mt Correll showed the highest
average pairwise FST estimate (0.247), and Hunt Range the least (0.119) (Table 3.4). Bayesian
analysis with TESS resulted in the identification of six genetic clusters corresponding to each
population, with the exception of Die Hardy and Hunt Range that clustered together (Fig. 3.4).
In contrast, PCoA identified two major genetic clusters, the first corresponded to Mt Finnerty,
Taipan Hill and Helena Aurora, and the second to all other populations, with further
subdivision of Mt Manning and Mt Correll separate from Die Hardy and Hunt Range. The first
two axes explained 36 % and 22.5% of the variation present (Fig. 3.5). The Mantel test showed
76
weak but significant IBD across all populations (r=0.21, p=0.0001) and also within the Die
Hardy Range (r=0.24, p=0.0001).
Indirect estimates of gene flow indicated low levels of migration between populations
(average Nm=0.75), and within a BIF (average Nm=1.31) (Table 3.4). The greatest gene flow
occurred between the Die Hardy south and Hunt Range populations (Nm = 1.458), and the
least between two of the most spatially distant, Mt Correll and Mt Finnerty (Nm = 0.182). The
centrally located Helena Aurora population received the highest gene flow with an average Nm
of 0.991.
Four of the nine populations (TH, DHN, DHS and HUN) showed significant departure from
mutation-drift equilibrium under the infinite allele mutation model using the Wilcoxon test
(P=<0.05). None of the populations deviated from mutation-drift equilibrium under the
stepwise mutation or two phase mutation model.
77
Table 3.3 Microsatellite genetic diversity estimates for nine populations of Grevillea georgeana genotyped using 10 loci. N, number of individuals genotyped; #
alleles, average number of alleles per locus; HO, observed heterozygosity; HE, expected heterozygosity; FIS, inbreeding coefficient. Standard deviations in
parentheses.
Population Pop abb. Latitude Longitude N # alleles HO HE FIS
Mt Correll COR -30.53937449 118.9105704 24 2.600(0.84) 0.390(0.28) 0.395(0.16) 0.012
Die Hardy north DHN -29.86778772 119.4049258 24 3.100(0.99) 0.406(0.26) 0.481(0.26) 0.158
Die Hardy central DHC -29.89930478 119.3565458 24 3.444(1.88) 0.505(0.35) 0.431(0.25) -0.176
Die Hardy south DHS -29.94275564 119.3747744 23 3.778(1.72) 0.518(0.23) 0.541(0.16) 0.043
Mt Finnerty FIN -30.67567776 120.1341005 24 3.700(1.64) 0.407(0.26) 0.451(0.25) 0.1
Helena Aurora HA -30.35353016 119.7039402 24 4.100(2.77) 0.491(0.30) 0.457(0.21) -0.076
Hunt Range HUN -29.94269066 119.3748764 17 4.000(1.80) 0.609(0.29) 0.580(0.15) -0.045
Mt Manning MM -29.81221491 119.6053819 24 3.143(1.46) 0.505(0.29) 0.436(0.25) -0.164
Taipan hill TH -30.38923472 119.9079597 24 3.900 (1.66) 0.497(0.22) 0.530(0.22) 0.062
78
Table 3.4 Pairwise FST values based on 10 microsatellite loci (upper diagonal) and Nm estimates (Slatkin, 1985) (lower diagonal) *P < 0.005. COR, Mt Correll; DH,
Die Hardy north, central, south; FIN, Mt Finnerty; HA, Helena Aurora; HUN, Hunt; MM, Mt Manning; TH, Taipan Hill.
COR DHN DHC DHS FIN HA HUN MM TH
COR 0 0.284* 0.193* 0.232* 0.382* 0.287* 0.146* 0.169* 0.284*
DHN 0.197 0 0.108* 0.046* 0.189* 0.189* 0.091* 0.232* 0.108*
DHC 0.322 0.959 0 0.031 0.198* 0.203* 0.050* 0.115* 0.162*
DHS 0.431 0.792 1.752 0 0.239* 0.209* 0.026 0.163* 0.117*
FIN 0.182 0.438 0.635 0.635 0 0.144* 0.202* 0.330* 0.087*
HA 0.399 0.986 1.420 1.266 1.178 0 0.192* 0.261* 0.098*
HUN 0.360 0.697 0.830 1.458 0.613 1.300 0 0.130* 0.119*
MM 0.226 0.254 0.532 0.607 0.315 0.654 0.391 0 0.247*
TH 0.278 0.308 0.385 0.510 0.603 0.726 0.587 0.316 0
79
Figure 3.5 Principal coordinates analysis of all Grevillea georgeana populations based on Nei’s genetic distance, calculated across 10 microsatellite loci. Primary and secondary axes shown, with the percentage of total variation explained by each. COR, Mt Correll; DH, Die Hardy north, central, south; FIN, Mt Finnerty; HA, Helena Aurora; HUN, Hunt; MM, Mt Manning; TH, Taipan Hill.
3.5 DISCUSSION
Phylogeographic analysis of G. georgeana revealed patterns of both long-term isolation and
connectivity, indicating different historical processes shaping evolution of these terrestrial
island populations. Some are behaving as isolated islands, with genetic drift driving patterns of
differentiation without geographic structure, whilst others show patterns of more recent
connectivity. The patterns of isolation are consistent with expectations from island
biogeography, but evidence of some pollen dispersal demonstrates that populations on
terrestrial habitat islands may have opportunities for connectivity. The potential for some
pollen movement between neighbouring BIFs is an important process that should be taken
into consideration when making decisions regarding land use management in this region.
80
3.5.1 Genetic differentiation among BIF populations
Analysis of cpDNA variation revealed complex patterns of genetic structure amongst
populations, with some geographic structuring of haplotypes, such as those specific to Helena
Aurora and Mt Correll, and others shared among populations, such as H11 and H12 from Mt
Finnerty, Taipan Hill and Hunt Range. The high diversity and divergence of haplotypes at Die
Hardy and Mt Manning is in contrast to specific haplotypes in other populations, such as
Helena Aurora and Mt Correll. The high level of genetic divergence among many populations
suggests long-term isolation of these populations, allowing the independent processes of
mutation and genetic drift to shape genetic variation (Byrne & Hopper, 2008). In particular, the
unresolved relationship of Die Hardy and Mt Manning haplotypes may reflect long periods of
isolation and persistence of G. georgeana on these large, northern populations (Lee, 2000).
Divergence of haplotypes both within and between BIFs, suggests isolation since the Mid
Pleistocene. The Pleistocene was a period of intensifying aridity in this landscape that
impacted the distribution of many species, causing localised extinctions and range shifts
(Bowler, 1982; Dodson & Macphail, 2004; Byrne, 2007; Fujioka et al., 2009), and this is likely to
have led to increased isolation in populations of G. georgeana that may have been restricted
to BIF islands prior to the Pleistocene. Similar high differentiation arising from historical
isolation from the early-mid Pleistocene has been observed in a millipede endemic to the
Yilgarn BIFs (Nistelberger et al., 2014a), and in other rare and common species that are
endemic to granite outcrops in the south western Australian landscape that also act as
terrestrial islands (Byrne & Hopper, 2008; Tapper et al., 2014a, 2014b). This pattern has also
been observed in intra-specific studies from other parts of the world on populations endemic
to terrestrial islands, such as the sky islands of Southern India (Robin et al., 2010), the Afro-
Alpine sky islands of East Africa (Assefa et al., 2007) and the Madrean sky islands of south
81
western USA (Smith & Farrell, 2005), highlighting the global influence of Pleistocene climatic
oscillations on driving species biogeography (Hewitt, 1996, 2000).
The general pattern of population divergence is consistent with our hypothesis of historical
isolation of populations, but we were surprised to see some evidence of connectivity between
the three south-eastern populations. This pattern may be explained by the underlying geology
of this part of the landscape. Mt Finnerty, Taipan Hill and Hunt Range, although separate
regions of uplift in the Yilgarn craton, are all part of the large Yendilberin/Watt Hills
greenstone belt that extends for approx. 90 km in a north-west to south-east direction in the
southern part of this species’ distribution. The results suggest populations of G. georgeana
along these hills may have been connected relatively recently, or indeed may still be
connected to some extent across portions of the belt that were difficult to access. The
widespread nature of the two haplotypes present in these populations and their position in
the haplotype network suggested these may be ancestral (Crandall & Templeton, 1993).
However, they form part of the unresolved polytomy in the Bayesian tree, making their
position relevant to other haplotypes difficult to determine.
Significant differentiation among many populations was also evident in the nrDNA, along with
isolation by distance, suggesting historical pollen flow between neighbouring BIF populations
was not frequent enough to override forces of genetic drift (Slatkin, 1985; Lowe & Allendorf,
2010). Pollen flow between neighbouring BIF populations was observed in a study of the
insect-pollinated Acacia woodmaniorum, and this was attributed to the long-distance dispersal
of pollinators by prevailing winds (Millar et al., 2013), although the geographic distances
between these BIFs were substantially less than those in our study. Grevillea georgeana is
visited by a range of honeyeaters that probably facilitate the majority of pollination, although
some insect pollination may occur (McGillivray & Makinson, 1993). Foraging behaviours are
82
variable amongst honeyeaters, ranging from within-tree and nearest-neighbour flower
visitation to long-distance dispersal events, and can be highly influenced by other species of
birds occupying the area (Ford, 1981; Hopper & Moran, 1981; Ford et al., 2000; Southerton et
al., 2004). The genetic results identified here suggest inter-population pollinator movements
are restricted, as would be expected given the large distances and dramatic change in
vegetation composition between BIFs that are highly likely to act as barriers to dispersal for
pollinators. Higher estimates of gene flow for the centrally located Helena Aurora population
suggest it may provide a stepping stone for birds moving through the broader landscape.
By sampling three populations across the large Die Hardy Range we were able to assess the
presence of genetic structure across a BIF that supports a large but discontinuous population.
The differentiation and pattern of isolation by distance between the Die Hardy sub-
populations suggests pollen movement is also restricted within a BIF, allowing differentiation
to develop at localised scales (approximately 4 km between sub-populations).
3.5.2 Genetic variation within BIF populations
The slow mutation rate of cpDNA can make it difficult to detect genetic variation at the
population level in angiosperms (McCauley 1995), yet we found high cpDNA variation,
particularly within the Mt Manning and Die Hardy populations. This high diversity suggests
historically large effective population size of G. georgeana on these BIFs (Frankham et al.,
2002). As expected, smaller populations, such as Mt Correll and Mt Finnerty, showed less
genetic variation, most likely due to limited gene flow, population bottlenecks and genetic drift
(Hamrick & Godt, 1989; Barrett & Kohn, 1991; Gitzendanner & Soltis, 2000). The population at
Mt Correll, a small and isolated BIF, was identified by both data sets as the least genetically
diverse and the most divergent. The overall microsatellite diversity in G. georgeana was similar
83
to that identified using microsatellite data in other Grevillea species with naturally restricted
distributions, including G. caleyi, G. longifolia and G. macleayana, indicating that long-term
isolation on BIF islands has not necessarily driven lower diversity in this species compared to
its congeners (England et al., 2003; Llorens et al., 2004).
The identification of recent bottlenecks under the IAM model in four of the populations
suggests fluctuations in population size may be a common occurrence in G. georgeana, due to
stochastic influences of fire, competition or climatic fluctuations on populations. Many
grevilleas, including G. georgeana, are killed by fire, with re-establishment occurring from the
seed bank (Gill, 1981; Olde & Marriott, 1995; Burne et al., 2003). England et al (2002)
suggested that fluctuations in the size of the accumulated seed bank due to varying fire
intervals may be responsible for population bottlenecks in some of the naturally fragmented G.
macleayana populations. Fire is likely to have played a large role in generating population
structure in this semi-arid landscape and may have caused periodic bottleneck events.
3.5.3 Conservation implications
This study has highlighted the importance of understanding genetic patterns in species
endemic to terrestrial islands, given the possibility for complex genetic structure that has
implications for conservation. A key goal of conservation biology is to preserve a range of
genetic diversity to ensure the potential for future adaptation and prevent inbreeding
depression (Soulé, 1980), and also to protect divergent lineages that may represent cryptic
species (Hebert et al., 2004) or progress towards speciation (Faith, 1992). In the case of G.
georgeana, the protection of Die Hardy Range and Mt Manning would capture both large
amounts of cpDNA genetic variation and divergent haplotypes. Mt Correll may also be
considered important, given the data indicates long periods of genetic isolation, which is likely
84
to continue into the future given this population’s spatial isolation. Perhaps most importantly,
the microsatellite data shows that whilst pollen flow is generally restricted, it has occurred to
some extent between neighbouring populations. Maintenance of connectivity between these
neighbouring populations is likely to be an integral process contributing to the persistence of
these otherwise isolated BIF populations.
85
Majestic, grand…and
incredibly prickly
86
Work in this chapter has been submitted for publication and published as:
Nistelberger HM, Byrne M, Coates DJ and Roberts JD (2014) Patterns of isolation and
connectivity in terrestrial island populations of Banksia arborea (Proteaceae). Submitted
Biological Journal of the Linnean Society.
Nistelberger HM, Roberts JD, Coates DJ and Byrne M (2013) Isolation and characterisation of
14 microsatellite loci from a short-range endemic, Western Australian tree, Banksia arborea
(C.A. Gardner). Conservation Genetics Resources, 5, 1143-1145. (Appendix D)
87
Chapter 4
Patterns of genetic isolation and connectivity in terrestrial island
populations of Banksia arborea.
4.1 ABSTRACT
The Banded Iron Formations (BIFs) of south-western Australia are terrestrial islands
characterised by high species richness and endemism. Regional endemics occur across multiple
formations without inhabiting the intervening landscape matrix. We asked whether eight BIF
populations of the regional endemic Banksia arborea, showed evidence of genetic connectivity
or isolation. Connectivity was assessed using three chloroplast DNA sequence markers and 11
nuclear microsatellite loci. Phylogenetic relationships were assessed with statistical parsimony
and Bayesian methods. Dating estimates of haplotype divergence were estimated using the
time to most recent common ancestor of B. arborea and B. purdieana, as well as a
conservative angiosperm chloroplast DNA mutation rate. Population genetic diversity and
structure was assessed amongst and within populations by genotyping 24 geographically
clustered individuals from each BIF and three sub-populations within the Die Hardy Range BIF.
Indirect gene flow estimates were determined using a method based on the frequency of
private alleles. Banksia arborea showed a complex evolutionary history with patterns of
genetic divergence amongst some BIF populations, reflecting processes of long-term isolation
and genetic drift, and an absence of differentiation amongst others. There was little evidence
of pollen dispersal among BIFs and within the large Die Hardy Range.
88
4.2 INTRODUCTION
Terrestrial or habitat islands are discrete habitat patches that are embedded within the
surrounding landscape. Depending on an organism’s habitat requirements, they can range
from large and conspicuous geological formations, such as mountain tops, to dung piles in a
grassland landscape (Whittaker, 1998). Terrestrial islands generally harbour different species
from the surrounding landscape due to the different habitats they provide (Rosenzweig, 2002).
Banded iron formations (BIFs) are ancient, geological features associated with Archean cratons
throughout the world and form some of the oldest structures known on Earth (Klein, 2005).
The most well-studied BIF terrestrial islands are those found in south-eastern Brazil and south-
western Australia. These are characterised by high species richness and endemism as a result
of the diversity of habitats, including changes in elevation, substrate, soil accumulation,
temperature and moisture (Jacobi et al., 2007; Jacobi & Carmo, 2008; Markey & Dillon, 2008;
Meissner & Caruso, 2008a, 2008b; Gibson et al., 2010, 2012), and also of the relative stability
of these landscapes that remained un-glaciated during the Pleistocene, allowing for the
development and persistence of diversity over time (Haffer, 1969; Gibson et al., 2010).
The BIFs of south-western Australia are features of the Yilgarn craton, one of the oldest
geological shields on earth (Freeman, 2001). They occur along the boundary of the Southwest
Australian Floristic Region (SWAFR) and extend inland to the arid zone for over 750 km (Gibson
et al., 2012). Floristic analysis of BIFs bordering the SWAFR and adjacent semi-arid zone
identified higher species turnover and endemism than those further inland in the arid zone
(Gibson et al., 2010, 2012). BIF endemics may be restricted to just one formation, or may be
regional, occurring across several BIFs, displaying an island-archipelago distribution. These
regional island endemics provide an excellent opportunity to examine patterns of population
connectivity or isolation using a combined phylogeographic and population genetic approach.
Compared to other terrestrial island systems, such as the sky islands of northern America
89
(DeChaine & Martin, 2004, 2005) and southern India (Robin et al., 2010), and rocky inselberg
and serpentine soils (Barbará et al., 2007; Byrne & Hopper, 2008; Boisselier-Dubayle et al.,
2010; Moore et al., 2013; Tapper et al., 2014a, 2014b), there has been little population-level
analysis of the patterns of genetic diversity and structure present in regional BIF endemics
worldwide. Recent studies on south-western Australian BIF flora and fauna have shown
organisms endemic to BIFs have complex phylogeographic histories, with some populations
showing evidence of isolation since the Pleistocene and others showing more recent
connectivity. For example, the millipede Atelomastix bamfordi occurs on five BIFs, where it
inhabits mesic microhabitats provided by shaded gullies containing accumulated leaf litter.
Analysis of populations using both mtDNA sequencing and nuclear microsatellite markers
indicated vicariance of a formerly widespread ancestor during the Pleistocene with subsequent
isolation of populations, resulting in patterns of genetic differentiation that were not
geographically structured (Nistelberger et al., 2014a). This pattern of long-term isolation on
terrestrial islands is similar to those observed in plant species restricted to granite outcrops of
Western Australia (Byrne & Hopper, 2008; Tapper et al., 2014a, 2014b) and to invertebrates
restricted to sky islands (Masta, 2000). In contrast, the BIF plant endemic Grevillea georgeana
has mixed phylogeographic patterns, with some populations showing evidence of genetic
isolation in cpDNA, and others showing evidence of recent population connectivity via both
seed and pollen flow (Nistelberger et al., 2014b). The insect-pollinated Acacia woodmaniorum
showed even greater evidence of population connectivity amongst disjunct BIF formations, a
pattern attributed to effective long-distance pollen dispersal (Millar et al., 2013). These studies
show different responses of species to isolation on BIF islands and highlight the need for
research on a range of species in order to identify whether common patterns exist. Major BIF
formations in Brazil and Western Australia are prospective for mining, placing populations at
risk from habitat removal and disruption of gene flow events. Understanding patterns of
genetic connectivity or isolation is important as it provides a basis for assessing the impact of
90
population loss and reduced connectivity on genetic diversity, inbreeding in small isolated
populations and species persistence.
To further the understanding of factors influencing population connectivity in BIFs, we
investigated phylogeographic pattern in a locally common tree species. Banksia arborea (C.A.
Gardner: A.R. Mast & K.R. Thiele) is restricted to the south–western, species rich BIFs that
occur close to the SWAFR (Gibson et al., 2007) and forms a conspicuous component of the BIF
vegetation community. It is a long-lived tree, up to eight metres high, with yellow flowers that
appear throughout the year (Wrigley & Fagg, 1989). Although little is known of the species’
ecology, it is presumed that insects, and potentially birds, are responsible for pollination (Ladd
et al., 1996). The woody fruits are serotinous and contain seeds that have no active
mechanism of dispersal (Wrigley & Fagg, 1989). We used cpDNA sequence data to assess
historical patterns of population connectivity, and nuclear microsatellite data to assess
contemporary patterns reflecting both pollen and seed dispersal within and between BIF
populations. Given the lack of active seed dispersal mechanisms in B. arborea we hypothesised
limited population connectivity via seed, resulting in strong signals of genetic differentiation in
the cpDNA data. Contemporary patterns of connectivity may be dependent on the mobility of
pollinators. If pollinator movement is limited we would expect significant population
differentiation that lacks geographic structure as a result of genetic drift driving genetic
structure, or, if pollen movement occurs between neighbouring BIFs, a pattern of isolation by
distance would be predicted.
91
4.3 MATERIALS AND METHODS
4.3.1 Sampling and DNA extraction
Banksia arborea was collected from eight Yilgarn BIFs; Koolyanobbing Range (KOO), Mt
Jackson (MTJ), Windarling Range (WIN), Die Hardy Range (DH), Mt Manning (MM), Taipan Hill
(TH), Hunt Range (HUN) and the Helena Aurora Range (HA). Voucher specimens from each
population were lodged at the Western Australian Herbarium. The BIFs vary in size, shape,
height and distance from one another. All form part of the Yilgarn Province, a region under
active exploration and resource development. For the phylogeographic component of the
study, 10 individuals were sampled across the geographic extent of each BIF where possible, in
order to avoid sampling closely related individuals and to maximise the genetic diversity
captured. Population genetic structure amongst the different BIF populations was analysed
using 11 specifically designed microsatellite markers on a geographically clustered group of 24
individuals (Nistelberger et al., 2013c). For the Die Hardy Range, three sub-populations of 24
individuals were collected, in the north, central and south of the range to allow assessment of
patterns of structure within this large BIF. Adult leaves from each tree were collected in paper
envelopes and placed immediately on silica gel, then desiccated in a freeze-drier (Labconco)
for 48 hours. DNA was extracted using a modified 2 % CTAB method outlined in Nistelberger et
al. (2013).
4.3.2 cpDNA amplification
Non-coding cpDNA regions that have been found useful for phylogeographical studies in
Australian plants (Byrne & Hankinson, 2012) were trialled in B. arborea, and the three regions
that produced the highest sequence quality and nucleotide diversity, the intergenic spacers
psbD-trnT and psbA-trnH and the D loop intron PetB, were selected for further analysis. PCR
reactions were carried out in 50 µl individual, non-multiplexed volumes containing 50 mM KCl,
92
20 mM Tris-HCl (pH8.4), 0.2 mM of each dNTP, 0.1 µM (psbD-trnT and psbA-trnH) or 0.08 µM
(PetB) primers, 1.5 mM (psbD-trnT and PetB) or 2.5 mM (psbA-trnH) MgCl2 and 0.1 U of Taq
polymerase (Invitrogen). The amplification protocol involved an initial hot start of 5 min at
80˚C, followed by 30 cycles of 95˚C for 1 min, 50˚C for 1 min and 65˚C for 4 min, with a final
extension step of 65˚C for 5 min. PCR products were cleaned and sequenced at the AGRF-Perth
(Australian Genome Research Facility) on an Applied Biosystems 3730 x 1 DNA Analyser
platform. Sequence data were aligned and edited in BioEdit (Hall, 1999) and finalised by eye.
The three regions were combined to form a total aligned sequence length of 1914 base pairs.
4.3.3 Microsatellilte amplification
PCR reactions (non-multiplexed) were performed in a total volume of 15 µl as outlined in
Nistelberger et al. (2013). Amplified products were run on 3% agarose gel with gel red nucleic
acid stain (Biotium) to assess amplification and polymorphism. Fourteen M13 labelled primer
pairs were initially used to amplify DNA; three of these were subsequently discarded, two
(Ba18 and Ba28) due to detection of high frequencies of null alleles, and the third (Ba04) due
to inconsistent banding patterns. PCR products were visualised on an Applied Biosystems 3730
capillary sequencer (Murdoch University) and scored using GeneMapper version 4.0 (Applied
Biosystems).
4.3.4 cpDNA analysis
Descriptive statistics, including number of haplotypes, nucleotide diversity and population
divergence (Jukes & Cantor, 1969) were calculated using DNAsp v5.10 (Librado & Rozas, 2009).
The structuring of cpDNA diversity amongst BIF populations was determined with an analysis
of molecular variance (AMOVA), based on the number of pairwise, base-pair differences, as
implemented in Arlequin v 3.5 (Excoffier & Lischer, 2010).
93
We used a Statistical Parsimony method to create an un-rooted haplotype network in the
program TCS (Clement et al., 2000) and estimated divergence times of the different haplotypes
using a Bayesian analysis in BEAST (Drummond et al., 2012). The best substitution model for
the dataset was identified as F81 based on AIC log likelihoods (Posada, 2008), so we
implemented the simplest substitution model available in the BEAST package (HKY) and
applied a Strict Clock model, owing to the low diversity of the cpDNA data, with a
coalescent:constant size tree prior. We used the time to most recent common ancestor
(TMRCA) of B. arborea and B. purdieana (Diels) A.R. Mast & K. R. Thiele, taken from Cardillo
and Pratt (2013) to calibrate the tree for molecular dating. For the root tree node we used a
normal prior (mean = 20.39, sd = 1) and initiated two independent MCMC runs of 10 million
generations from starting trees, sampling every 1000 generations. Stationarity was assessed by
inspecting log files for high ESS values in Tracer. Tree files were combined in LogCombiner and
annotated in TreeAnnotater discarding the first 25% of trees as burnin (available BEAST
package). As a further test we compared these dates to those obtained using the conservative
mutation rate commonly reported for angiosperms 1.2 x 10-9 (Graur & Li, 2000).
4.3.5 Microsatellite analysis
Tests for linkage disequilibrium and Hardy-Weinberg equilibrium along with estimates of FIS
were calculated using FSTAT v 2.9.3 (Goudet, 2001). Estimates of population genetic diversity
and structure parameters, including the average number of alleles per locus, HO and HE,
pairwise FST as well as AMOVA were calculated in Arlequin v.3.5 (Excoffier & Lischer, 2010).
Jost’s D statistic, based on the effective number of alleles rather than heterozygosity (Jost,
2008) was also used to measure population differentiation, calculated in GenAlEx v. 6.5
(Peakall & Smouse, 2006). Differences in population heterozygosity and the number of alleles
per population were tested with a Friedman ANOVA (for multiple paired comparisons) with
94
samples paired by locus. The presence and frequency of null alleles were estimated using a
maximum likelihood method in the program ML-null (Kalinowski & Taper, 2006).
Genetic structure amongst BIF populations was also assessed using Bayesian clustering
algorithms implemented in the programs STRUCTURE v.2.3.3 (Pritchard et al., 2000) and TESS
v.2.3.1 (Chen et al., 2007). STRUCTURE assigns individuals to the most likely number of groups
(K) without prior knowledge of spatial location by maximising within-K Hardy-Weinberg and
linkage equilibria (Pritchard & Wen, 2004). We ran STRUCTURE for 1 million MCMC repetitions
with a burnin length of 100,000. Simulations for proposed K values between 1 and 11 were
iterated 10 times and the posterior mean recorded. We used a run model that assumed
independence of allele frequencies and admixed ancestry. The output was run through
STRUCTURE HARVESTER (http://taylor0.biology.ucla.edu/structureHarvester/) that uses the
Evanno method (Evanno et al., 2005) to determine the optimal value of K. These results were
then used in the program CLUMPP (Jakobsson & Rosenberg, 2007) that aligns cluster
assignment across the replicate analyses and ensures the correct assignment of K. In contrast
to STRUCTURE, TESS assumes proximate interactions between individuals, with spatial
information prescribed at the individual level. TESS has been shown to perform better under
scenarios with nil to moderate admixture, whilst STRUCTURE is better at assigning K under
high levels of admixture (Chen et al., 2007). We tested for structure across the BIF populations
with TESS under both admixture and no-admixture models. For the admixture model,
simulations for K values of one to 11 were iterated 100 times each, with a total of 100,000
sweeps run and a burnin of 20,000. Under the no-admixture model we conducted the same
number of runs and iterations and set the spatial interaction parameter to 0.8. The optimal
value of K was determined by plotting the lowest DIC values of each K iteration and
convergence was assessed using the downstream program CLUMPP. The genetic relationships
among populations were further examined using principal coordinates analysis (PCoA) of
95
pairwise population estimates of Nei’s unbiased genetic distance (1978), implemented in
Genalex6.5 (Peakall & Smouse, 2006).
We tested for a pattern of isolation by distance (IBD) across all eight BIF populations and
across the three Die Hardy sub-populations by performing a Mantel test on matrices of genetic
distance (Nei, 1978) and log geographic distance in the program IBDWS. A Wilcoxon signed
rank test was used to test for an excess of observed heterozygotes that can indicate a
bottleneck event using the program BOTTLENECK v 1.2.02 (Piry et al., 1999). Heterozygote
excess was tested under the Infinite Allele Model, the Stepwise Mutation Model and the Two
Phase Model. For the Two Phase Model, we used settings typical for microsatellite markers
with the frequency of single step mutations set to 90% and the variance of those mutations as
12% (Busch et al., 2007). Indirect measures of gene flow (Nm) between all BIF populations and
the Die Hardy sub-populations were calculated using a method based on the frequency of
private alleles (Slatkin, 1985), in the program Genepop v 1.2 (Raymond & Rousset, 1995).
4.4 RESULTS
4.4.1 cpDNA variation
Amplification of the three cpDNA regions across 79 individuals (one individual from Helena
Aurora failed to amplify) revealed five haplotypes comprised of three transitions and three
transversions. One haplotype was specific to two individuals from the Die Hardy population
(H1), one haplotype was restricted to Taipan Hill (H4) and one to Koolyanobbing (H5). Two
were specific to multiple populations (MTJ, HA, HUN) (H2) and (WIN, DH, MM) (H3) (Table 4.1)
(Fig. 4.1). The majority of cpDNA variation (98%) occurred amongst populations, with only the
Die Hardy population possessing more than one haplotype.
96
The parsimony network showed some structure among haplotypes, with the two common
haplotypes and the haplotype restricted to Die Hardy being more closely related, and
divergence of two haplotypes present at Koolyanobbing and Taipan Hill (Fig. 4.2, Table 4.2).
The divergence times amongst haplotypes based on the calibrated BEAST tree ranged from 1.2
Ma (95% hpd levels 0.6-2.0 Ma) and 1.4 Ma (95% hpd levels 0.5-2.6 Ma) between haplotypes 2
and 3 and haplotypes 4 and 5 respectively, to 2.7 Ma (95% hpd levels 1.2-4.7 Ma) for the
divergence of haplotypes 4 and 5 from haplotypes 1, 2 and 3 (Fig. 4.3). These estimates were
older than those generated using the conservative, angiosperm mutation rate that indicated
divergence of 0.3 Ma (95% hpd levels 0.1-0.5 Ma) between haplotypes 2 and 3 and 0.74 (95%
hpd levels 0.3-1.2 Ma) for the divergence of haplotypes 4 and 5 from haplotypes 1, 2 and 3.
Figure 4.1 Map of the eight Banksia arborea BIF population locations with haplotypes represented by different colours. Inset shows study location. Image adapted from Google Earth.
97
Figure 4.2 Maximum parsimony haplotype network for Banksia arborea based on combined cpDNA data. Node size is proportional to the frequency of haplotypes with H1 representing 2 individuals and H2 29 individuals. The inferred ancestral haplotype is shown as a rectangle and black dots indicate un-sampled or extinct haplotypes. Colours identify sequence haplotypes as shown in Fig. 4.1. DH, Die Hardy; HA, Helena Aurora; HUN, Hunt; KOO, Koolyanobbing; MM, Mt Manning; MTJ, Mt Jackson; TH, Taipan Hill; WIN, Windarling.
4.4.2 Microsatellite variation
There was no evidence of linkage disequilibrium between any pairwise combinations of loci.
Null alleles were estimated to occur at frequencies between 10% and 20% in Barb 10, 16 and
36. To examine the influence of these three loci on the results we also analysed our data
without these markers, but this did not significantly alter the analysis outcomes and the data
for all 11 loci is presented. Genetic variation within populations was low, with an average of
4.18 alleles per locus across all populations. Observed and expected heterozygosity was also
low, ranging from 0.39 - 0.56 and 0.46 - 0.55 respectively. Seven out of 10 populations had
private alleles, with Die Hardy north showing significantly higher numbers than all other
populations (> 2SD from the mean) (Table 4.3). All populations were in Hardy-Weinberg
equilibrium following Bonferroni correction (Table 4.3).
98
Figure 4.3 Bayesian cladogram of five Banksia arborea haplotypes based on combined cpDNA data using B. densa and B. purdieana as outgroups. Numbers on the left-hand side of nodes indicate posterior probabilities. Numbers on the right-hand side of nodes indicate divergence estimates in millions of years with confidence intervals shown in brackets. Age estimates shown were based on the TMRCA of B. arborea and B. purdieana. NB. the branch to B. purdieana was truncated to improve tree appearance. DH, Die Hardy; HA, Helena Aurora; HUN, Hunt; KOO, Koolyanobbing; MM, Mt Manning; MTJ, Mt Jackson; TH, Taipan Hill; WIN, Windarling.
99
Table 4.1 Diversity statistics for chloroplast DNA data for eight Banksia arborea populations: N, number of individuals; # Haps, number of haplotypes; HD, haplotype diversity; Pi, nucleotide diversity. DH, Die Hardy; HA, Helena Aurora; HUN, Hunt; KOO, Koolyanobbing; MM, Mt Manning; MTJ, Mt Jackson; TH, Taipan Hill; WIN, Windarling.
Genbank numbers
Population N # HAPS HD PI (%) PetB PsbA PsbD
DH 10 2 0.356 0.019 KF668088 KF668091 KF668094 KF668095
HA 9 1 0 0 KF668089 KF668091 KF668094
HUN 10 1 0 0 KF668089 KF668091 KF668094
KOO 10 1 0 0 KF668090 KF668092 KF668095
MM 10 1 0 0 KF668088 KF668091 KF668094
MTJ 10 1 0 0 KF668089 KF668091 KF668094
TH 10 1 0 0 KF668090 KF668093 KF668095
WIN 10 1 0 0 KF668088 KF668091 KF668094
Table 4.2 Estimates of cpDNA divergence (JC) (%) between populations of Banksia arborea
DH HA HUN KOO MM MTJ TH
HA 0.063 HUN 0.063 0
KOO 0.304 0.262 0.262 MM 0.010 0.052 0.052 0.314
MTJ 0.063 0 0 0.262 0.052 TH 0.251 0.209 0.209 0.052 0.262 0.209
WIN 0.010 0.052 0.052 0.314 0 0.052 0.262
100
Table 4.3 Microsatellite genetic diversity estimates for 10 populations of Banksia arborea genotyped at 11 loci. N, number of individuals genotyped; # alleles,
average number of alleles per locus; HO, observed heterozygosity; HE, expected heterozygosity; FIS, inbreeding coefficient. Standard error in parentheses. *
Significantly higher than the mean number of private alleles.
Population pop abb.
N # alleles
# Private HO HE FIS alleles
Die Hardy north DHN 24 4.80(0.50) 7* 0.43(0.06) 0.50(0.05) 0.147
Die Hardy central DHC 23 4.10(0.53) 0 0.43(0.06) 0.53(0.04) 0.186
Die Hardy south DHS 24 4.60(0.58) 1 0.52(0.05) 0.55(0.05) 0.057
Helena Aurora HA 24 4.60(0.47) 2 0.52(0.07) 0.55(0.06) 0.063
Hunt Range HUN 11 3.46(0.46) 1 0.42(0.08) 0.51(0.06) 0.170
Koolyanobbing KOO 24 3.82(0.55) 3 0.43(0.05) 0.46(0.04) 0.068
Mt Jackson MTJ 24 4.46(0.44) 0 0.39(0.05) 0.52(0.05) 0.248
Mt Manning MM 24 3.46(0.36) 0 0.47(0.05) 0.54(0.04) 0.139
Taipan Hill TH 24 4.18(0.53) 3 0.44(0.07) 0.49(0.06) 0.115
Windarling WIN 20 4.27(0.61) 2 0.56(0.06) 0.51(0.05) -0.108
101
Analysis of molecular variance indicated significant partitioning of the overall genetic diversity
with 12.1% of the variation found amongst populations. All pairwise population FST values were
significant, with the exception of the values for Die Hardy central and Die Hardy south, Die
Hardy south and Die Hardy north, and Mt Manning and Hunt Range (Table 4.4). The average
FST across all BIF population pairs was 0.14 (excluding DHN and DHS). All pairwise population
estimates of DEST were significant (Table 4.4). The lack of significant differentiation obtained in
FST values between Mt Manning and Hunt Range was not reflected in DEST values or in a higher
Nm estimate between these two populations, and is likely an artefact of lower sample size
(n=11) taken from Hunt Range.
The two Bayesian analyses of genetic structure produced conflicting results. The program
STRUCTURE resolved two genetic clusters, both showing high levels of admixture. TESS analysis
showed K = 10 under a no-admixture model and K = 11 under an admixture model but none of
the runs from STRUCTURE or TESS indicated convergence following analysis in the downstream
program CLUMPP (all H’ values <0.8). This failure to detect genetic clusters, despite significant
population differentiation (FST and DEST) is most likely due to a lack of sufficient genetic
variation within the data (Rosenberg et al., 2001). Principal coordinates analysis showed
differentiation amongst all populations, but with some clustering of the Die Hardy sub-
populations and greatest divergence of the Windarling population. The first two principal
component (PC) axes explained 27% and 20% of the genetic variation among populations (Fig.
4.4).
There was no signal of IBD across all BIF populations (p = 0.88), although there was a weak, but
significant relationship across the three sub-populations sampled from the Die Hardy Range (r=
0.12, p=< 0.001) (appendix Fig. C1). Three populations (DHC, DHS and MM) showed significant
heterozygote excess, indicating potential bottleneck events using the Wilcoxon sign rank test
102
under the Infinite allele method. No bottlenecks were evident under the stepwise mutation
model or two-phase mutation model.
Estimates of gene flow indicated low pollen movement both within and between BIFs, with an
average Nm of 0.94 across all Die Hardy sub-populations and 0.78 between all BIF populations
(Table 4.5). The highest level of gene flow between BIF populations occurred between Die
Hardy north and Taipan Hill (Nm = 1.81), and the least between Hunt Range and Mt Manning
(Nm = 0.32) (Table 4.5). This latter result was interesting given Hunt Range and Mt Manning
were not significantly differentiated according to FST estimates. Die Hardy north showed the
highest average Nm (1.16) estimate across all populations (when excluding DHC and DHS) and
Mt Manning the least (0.06).
Figure 4.4 Principal coordinate analysis of Banksia arborea populations based on pairwise population comparisons of Nei’s unbiased genetic distance calculated across 11 microsatellite loci. Primary and secondary axes shown, with the percentage of total variation explained by each. DH, Die Hardy north, central, south; HA, Helena Aurora; HUN, Hunt; KOO, Koolyanobbing; MM, Mt Manning; MTJ, Mt Jackson; TH, Taipan Hill; WIN, Windarling.
103
Table 4.4 Banksia arborea population pairwise FST values (lower diagonal) and DEST values (upper diagonal).* P < 0.05. DH, Die Hardy north, central, south; HA,
Helena Aurora; HUN, Hunt; KOO, Koolyanobbing; MTJ, Mt Jackson; MM, Mt Manning; TH, Taipan Hill; WIN, Windarling.
DHN DHC DHS HA HUN KOO MTJ MM TH WIN
DHN 0 0.062* 0.036* 0.079* 0.042* 0.103* 0.184* 0.091* 0.039* 0.158*
DHC 0.043* 0 0.022* 0.062* 0.086* 0.119* 0.127* 0.063* 0.108* 0.088*
DHS 0.027 0.020 0 0.043* 0.024* 0.099* 0.112* 0.065* 0.081* 0.100*
HA 0.074* 0.126* 0.067* 0 0.051* 0.057* 0.036* 0.110* 0.055* 0.084*
HUN 0.101* 0.164* 0.054* 0.119* 0 0.056* 0.074* 0.063* 0.075* 0.178*
KOO 0.148* 0.193* 0.129* 0.120* 0.086* 0 0.113* 0.102* 0.126* 0.169*
MTJ 0.213* 0.210* 0.121* 0.148* 0.110* 0.211* 0 0.099* 0.107* 0.067*
MM 0.108* 0.141* 0.074* 0.144* 0.047 0.112* 0.158* 0 0.135* 0.132*
TH 0.126* 0.150* 0.072* 0.058* 0.079* 0.172* 0.123* 0.130* 0 0.139*
WIN 0.054* 0.060* 0.068* 0.088* 0.185* 0.138* 0.282* 0.110* 0.157* 0
104
Table 4.5 Pairwise indirect gene flow estimates (Nm) between populations of Banksia arborea based on the private allele method of Slatkin (1985). DH, Die Hardy north, central, south; HA, Helena Aurora; HUN, Hunt; KOO, Koolyanobbing; MTJ, Mt Jackson; MM, Mt Manning; TH, Taipan Hill; WIN, Windarling.
DHN DHC DHS HA HUN KOO MTJ MM TH WIN
DHN 0 DHC 1.42 0
DHS 1.65 1.15 0 HA 1.06 0.61 0.75 0
HUN 1.49 0.45 1.00 1.09 0 KOO 0.97 0.50 0.49 0.74 0.57 0
MTJ 1.13 1.26 0.90 1.08 0.92 1.06 0 MM 0.81 0.76 0.50 0.71 0.32 0.47 0.84 0
TH 1.81 0.66 0.83 1.30 0.69 0.90 0.71 0.56 0 WIN 0.84 0.78 0.83 0.95 0.52 0.56 1.02 0.45 1.06 0
105
4.5 DISCUSSION
Phylogeographic studies of species with island distributions can aid in our understanding of
factors influencing connectivity and isolation through their effect on genetic structure.
Phylogeography of B. arborea populations has shown contrasting patterns of both isolation
and connectivity amongst BIF populations reflecting, different historical processes operating at
different scales throughout this landscape. The northern populations showed evidence of
connectivity in the cpDNA, although appear to have limited connectivity through pollen
dispersal, whereas divergence is evident in the geographically isolated and demographically
small populations of Koolyanobbing and Taipan Hill respectively.
4.5.1 Genetic differentiation among BIF populations
Variation in the cpDNA showed genetic differentiation of some BIF populations, such as Taipan
Hill and Koolyanobbing that both possessed population specific haplotypes. In contrast the
most northern populations were genetically similar with shared haplotypes across several BIFs.
The sharing of haplotypes between Die Hardy, Windarling and Mt Manning, and between Mt
Jackson, Helena Aurora and Hunt Range, may be due to ongoing connectivity among these BIF
populations via dispersal of seed, or may indicate a more widespread historical distribution
that covered both BIFs and the intervening landscape. Banksia arborea has less restrictive
habitat requirements than some other endemic BIF plants, such as G. georgeana and A.
woodmaniorum, and although it is typically confined to BIF, it has also been identified on the
edge of BIF in sand habitat and one population has been sighted at the base of a granite
outcrop (pers. comm. N. Gibson). The landscape between the BIFs consists of alluvial soils
dominated by eucalypt woodland, interspersed with patches of sandplain generated by the
weathering of surrounding granite outcrops (Payne et al., 1998). Ancestral populations of B.
arborea may have occurred on these sandplains in the past, connecting the now disjunct
northern populations. The other alternative of long-distance seed dispersal connecting
106
populations is also plausible, as long-distance seed dispersal has been shown to occur in other
Banksia species. Studies by He et al. (2004, 2009) revealed instances of remarkably long-
distance seed dispersal in Banksia hookeriana (Meisn) and B. attenuata (R. Br.), species with
patchy distributions endemic to the Eneabba sand plains in south-west Australia. Those studies
showed seed dispersal over several kilometres, and attributed this to transportation of cones
by birds or via wind vortices that can occur following fire. The dispersal of B. arborea seed
across the large distances separating the northern Yilgarn BIFs may be a rare phenomenon, but
even rare events, over the course of millions of years, may be sufficient to maintain genetic
connectivity among populations. Unfortunately, due to the lack of variation within
populations, it was not possible to distinguish between the contraction of a widespread
ancestor to BIF or connectivity via seed dispersal. The greater genetic divergence of the
demographically smaller Taipan Hill population and the geographically isolated population at
Koolyanobbing, suggests either longer periods of isolation of these populations or the greater
influence of genetic drift on historically small effective population size in comparison to the
larger northern populations. Contrasting patterns of isolation and connectivity were also
observed in the Yilgarn BIF endemic Grevillea georgeana, where the majority of populations
exhibited population specific haplotypes reflecting long-term isolation with the exception of
three southern populations which demonstrated evidence of connectivity, sharing haplotypes
and showing little genetic differentiation (Nistelberger et al., 2014b).
Molecular dating showed the divergence of all haplotypes occurred during the Pleistocene.
This was a period of intensifying aridity, associated with climatic fluctuations in this landscape
that saw the transition of vegetation communities to dominance by more arid-adapted species
(Bowler, 1982; Crisp et al., 2004; Dodson & Macphail, 2004; Martin, 2006). Populations of B.
arborea may have already been restricted to BIF outcrops by this period, yet climatic
fluctuations would still have shaped population distributions in this landscape, driving localised
107
extinction as well as contraction and expansion of population boundaries. The broad influence
of these processes is evident in similar genetic patterns seen in other taxa endemic to the
Yilgarn BIFs, including the millipede Atelomastix bamfordi (Nistelberger et al., 2014a) and the
plant Grevillea georgeana (Nistelberger et al., 2014b), as well as in other more widely
distributed species from south-western Australia (Byrne et al., 2002, 2003; Byrne & Hines,
2004; Hopper & Gioia, 2004; Byrne, 2007).
Patterns of genetic structure in the microsatellite data did not reflect those in the cpDNA and
showed all populations, with the exception of Mt Manning and Hunt Range, to be significantly
differentiated but without a geographical pattern to the structure. Discordance between the
two marker data sets can occur given the different modes of inheritance and mutation rates of
cpDNA and nuclear DNA (Avise, 2000). The pattern of differentiation seen in the microsatellite
data is consistent with processes of population isolation and genetic drift driving
differentiation and the derived migration rates suggest very limited pollen movement between
populations (Mills & Allendorf, 1996; Lowe & Allendorf, 2010). This was in contrast to the
patterns observed in other BIF endemics Grevillea georgeana and Acacia woodmaniorum,
where isolation by distance amongst populations indicated some pollen flow between
adjacent BIFs (Millar et al., 2013; Nistelberger et al., 2014b). This contrast may reflect different
behaviours of pollinators. The bright, yellow flowers of B. arborea attract a range of insects
and birds (Pers. obs. H. Nistelberger). The dominant pollinators are not known, but insect
pollinators often exhibit restricted movement patterns (Heinrich & Raven, 1972; Loveless &
Hamrick, 1984), but many bird species can also show limited movement due to territorial and
nearest-neighbour foraging behaviours (Ford, 1981; Turner et al., 1982).
By sampling three sub-populations within the Die Hardy Range we were able to assess patterns
of genetic structure within a BIF. Sub-population differentiation was lower than that between
108
BIFs, but although there was a weak pattern of isolation by distance, migration estimates still
indicated limited pollen flow within a BIF. This is surprising, as habitat conditions within a BIF
would be expected to promote pollinator movement, compared to the differing habitat of the
surrounding landscape.
4.5.2 Genetic variation and differentiation within a BIF population
Both cpDNA and nrDNA markers used in this study showed low levels of genetic diversity
within populations. The microsatellite data in particular showed low allelic richness and
expected heterozygosity in comparison to other Banksia studies that have used microsatellites,
including two species from SWAFR, Banksia sphaerocarpa (R. Br.) (Llorens et al., 2012) and B.
hookeriana (Krauss et al., 2006). Both of these species have more widespread distributions
that typically maintain higher levels of genetic diversity owing to greater gene flow, less
genetic drift and fewer population bottlenecks (Hamrick & Godt, 1989). Low diversity detected
here in B. arborea suggests populations have been through bottlenecks, perhaps both
historically and more recently. Indeed, evidence for recent bottleneck events were detected in
some populations. Larger BIFs might be expected to maintain larger populations that are more
resistant to loss of diversity caused by genetic drift and population bottlenecks (Wright, 1929).
The populations on the large Die Hardy Range did show higher genetic diversity in terms of
cpDNA haplotypes and private alleles in the nuclear data in Die Hardy north, although there
was also evidence of bottlenecks in two populations. Persistence of low frequency alleles in
the Die Hardy north population suggests maintenance of large effective population size, as
bottleneck events and genetic drift remove novel, low-frequency alleles leaving dominant,
high-frequency alleles (Luikart et al., 1998).
109
4.5.3 Conservation implications
This study highlights that species-specific patterns of population genetic differentiation can
occur in regional BIF endemics. Genetic divergence of the small Taipan Hill and geographically
isolated Koolyanobbing populations in B. arborea is consistent with population genetic theory
and with results from other studies on terrestrial island habitats (Byrne & Hopper, 2008; Millar
et al., 2013; Nistelberger et al., 2014a, 2014b; Tapper et al., 2014a, 2014b). However, the
extent of apparent connectivity in the cpDNA among many of the populations is in contrast to
the general expectations of island biogeography, and in contrast to the level of genetic
structure and isolation evident in a previous study of the shrub Grevillea georgeana that also
occurs on these BIFs (Nistelberger et al., 2014b). These differences in connectivity may relate
to different seed dispersal mechanisms or the possibility that B. arborea has been connected
through populations in the intervening matrix in the past.
Understanding these patterns of genetic connectivity and isolation is important if managers
need to identify target regions or particular BIFs as priorities for conservation or reservation.
Those areas with habitats that facilitate maintenance of high genetic diversity, and/or Ranges
where isolation has driven genetic divergence, would have high priority for conservation.
While other factors such as species richness and endemism are often considered critical we
suggest that these zones of high genetic diversity should also be taken into account.
Knowledge of population dynamics is also important in order to assess the impact of
population loss and reduced connectivity on genetic diversity loss and potential inbreeding in
small isolated populations. In this study, the Die Hardy Range emerged as a region of higher
genetic diversity. This large BIF is topographically complex, comprised of a mixture of sheer
cliff faces, gentle slopes and shaded valleys. This has resulted in localised groups of individuals,
a feature that may allow for the retention of diversity over time, with some sub-populations
influenced by bottlenecks, such as Die Hardy central and Die Hardy south, whilst others, like
110
Die Hardy north remaining stable. Populations that are highly divergent, such as those at
Koolyanobbing and Taipan Hill may also be considered significant for conservation purposes
given the potential for prolonged isolation of divergent lineages to result in reproductive
isolation and allopatric speciation.
111
Striking the balance:
Conservation and the
Yilgarn BIFs
112
113
Chapter 5
Hotspots of genetic diversity identify terrestrial islands of conservation
value in a semi-arid Australian landscape.
5.1 ABSTRACT
The Yilgarn Banded Iron Formations (BIFs) of Western Australia are terrestrial islands
characterised by high species richness and endemism that are subject to increasing land-use
pressures. BIFs vary in size, complexity and isolation, and the distribution of genetic diversity
and divergence in species endemic to these BIFs may also vary accordingly. We used genetic
data from three regional BIF endemics, which occur across multiple ‘islands’ but not in the
intervening landscape, to determine whether patterns of high genetic diversity and/or
divergence were similar across all three taxa. The relative contribution of BIF populations to
total species genetic diversity was determined using measures of allelic richness and gene
diversity. Large BIFs, including the Die Hardy Range and Koolyanobbing, consistently showed
higher contributions to within population diversity, suggesting a role of these BIFs as genetic
reservoirs in this landscape. Smaller or more spatially isolated BIFs, such as Windarling and
Koolyanobbing, showed high contributions to population divergence. Four BIFs,
Koolyanobbing, Die Hardy Range, Taipan Hill and Windarling Range emerged as major
contributors to overall landscape genetic diversity in these species and warrant consideration
as priorities for conservation.
114
5.2 INTRODUCTION
Global human population growth and associated changes in land-use, are placing increasing
pressure on remaining natural environments, driving the need for informed decision-making to
determine what should receive conservation priority. Studies of genetic diversity complement
assessment of species richness and endemism, and provide a means of including an
evolutionary perspective in identification of populations, species or regions that may have
higher priority for conservation (Allendorf et al., 2013).
There are several approaches based on the identification of high genetic diversity and
divergence that have been designed to assist in determining priorities for conservation.
Preservation of genetic diversity is important for maximising persistence and capacity to
respond to changing environmental conditions (Frankel & Soule, 1981; Shaffer, 1981), whilst
maximising genetic divergence, or ‘distinctiveness’ (Faith, 1992; Moritz & Faith, 1998; Petit et
al., 1998) is important as divergent lineages can indicate historical isolation, and may result in
allopatric speciation with prolonged isolation (Ehrlich, 1988; Crozier, 1992; Faith, 1992). The
Phylogenetic Diversity (PD) method of Faith (1992) uses cladistic relations among taxa to
determine the subset that displays the maximum underlying feature (phylogenetic) diversity
and is calculated by summing the tree branch lengths in the minimum spanning path between
taxa (Faith, 1992). This measure has been implemented in a number of studies, typically at
higher systematic levels, that have identified regions of evolutionary potential in the Cape flora
of South Africa (Forest et al., 2007), mammal speciation in California (Davis et al., 2008) and
phylogenetic diversity of bird genera in north West Africa (Rodrigues & Gaston, 2002). Another
approach is to identify Evolutionarily Significant Units (ESUs), groups that show either genetic
or ecological distinctiveness, as a means to identify lineages that are a priority for conservation
effort (Ryder, 1986; Allendorf et al., 2013). Moritz (1994) provided guidelines to enable clear
identification of ESUs based on genetic data, such as evidence of reciprocal monophyly in the
115
faster evolving mitochondrial DNA genome and significant divergence of allele frequencies in
the nuclear genome (Moritz, 1994).
These methods provide effective means of identifying evolutionary units and geographical
areas that are a focus for conservation, but a downfall is their reliance on resolved
phylogenetic trees (Faith, 1992). When dealing with population-level data, particularly
populations that are heavily influenced by genetic drift, obtaining well-resolved trees can be
problematic (Purvis & Garland, 1993; Kim et al., 1998; Lee, 2000). These phylogenies may
contain unresolved branches (polytomy), hampering efforts to identify monophyly or to
calculate the minimum spanning path between a subset of populations (Swenson, 2009).
An alternative method developed by Petit et al (1998), allows for the relative contribution of
populations to total species diversity and divergence to be determined when relationships
amongst populations are unresolved. The method involves calculating total genetic diversity -
in the form of either allelic richness or gene diversity, both of which can be broken down into
their two components: variation within populations and divergence among populations. This
method is useful as estimates of allelic richness and gene diversity are widely used in
conservation genetic studies (Petit et al., 1998), and allelic richness in particular has been
widely argued to provide one of the best measures of genetic diversity in the context of
conservation genetics (Schoen & Brown, 1993; Bataillon et al., 1996; El Mousadik & Petit,
1996) as it is a strong indicator of past demographic changes within populations (Petit et al.,
1998). A drawback of allelic richness is its reliance on sample size, and this problem has been
dealt with in the software CONTRIB (Petit et al. 1998) by using rarefaction, a statistical
technique that compensates for variation in sample size (Petit et al., 1998; Kalinowski, 2005).
The CONTRIB method has been used to estimate the relative contribution of populations to
116
total species diversity in tree species and horse and sheep breeds (Petit et al., 1998; Comps et
al., 2001) (Solis et al., 2005; Tapio et al., 2005).
Population-level conservation genetic studies are often designed around single species that
are likely to be heavily impacted by some form of disturbance; either because they are
charismatic, rare and/or have restricted distributions (Soulé & Mills, 1992; Ueno et al., 2005;
Marshall et al., 2008; Butcher et al., 2009; Allendorf et al., 2013; Millar et al., 2013). However,
the same data can be used in a multi-species approach to identify specific areas or habitats
where protection may result in more effective net conservation outcomes at a landscape level
(Mace et al., 2003). This approach may be particularly useful in analysis of sympatric species
that exhibit disjunct distributions, such as found in terrestrial island systems, to identify
landscape genetic hotspots that require conservation priority, rather than relying on
population analyses of single species.
The Yilgarn Banded Iron Formations (BIFs) of Western Australia are ancient topographical
features that behave as terrestrial islands in a flat, semi-arid landscape. They are centres of
species richness and endemism and are therefore of high conservation value (Gibson et al.,
2007, 2010, 2012). The southern Yilgarn BIFs that border the South West Australian Floristic
Region are in a relatively pristine state in comparison to other BIFs throughout the world
(Burke, 2003); although many are prospective for resource development (Gibson et al., 2012)
and are subject to pastoral activities, and there is a need to find a balance between land-use
activities and conservation in this region (Environmental Protection Authority, 2013). The BIFs
vary in size, shape and degree of isolation, and there is a high turnover in plant species
diversity both within and between BIFs (Gibson et al., 2010). There is also variation in the
number of endemic species present on each BIF (Gibson et al., 2010), but it is not known
whether these differences in species richness and endemism amongst BIFs are also reflected in
117
differences in genetic diversity within and divergence between populations in plant and animal
species found on the formations. Regional endemic species provide an excellent opportunity
to investigate whether common patterns of genetic diversity and structure exist within the
landscape, as this might enable the identification of BIFs that can be considered a higher
priority for conservation.
Recent studies (Nistelberger et al., 2014; Chapter 3; Chapter 4) have evaluated the
phylogeography and population genetics of three regional endemics found on the most
southerly Yilgarn BIFs in south-western Australia. Here, we use these data to answer three key
questions. First, is genetic diversity across species partitioned equally amongst BIF islands or
are there particular BIFs with higher diversity? Second, are there any BIFs which exhibit higher
levels of genetic divergence than others? Finally, are there BIFs that encompass both of these
attributes that may be prioritised for conservation? The results will not only aid in our
understanding of the evolutionary history of the Yilgarn but will contribute to informed
decisions that consider the underlying genetic diversity that is important for persistence of
species endemic to this region.
5.3 MATERIALS AND METHODS
5.3.1 Study species and population sampling
A total of 10 south-western Yilgarn BIFs were included in this study (Fig. 5.1). We used
cytoplasmic and nuclear DNA data from three species endemic to the southern Yilgarn BIFs,
Atelomastix bamfordi, Grevillea georgeana and Banksia arborea (Table 5.1), to assess the
relative contribution of BIF populations to genetic diversity and divergence.
118
Figure 5.1 Map of the 10 Yilgarn BIFs sampled for Atelomastix bamfordi, Grevillea georgeana and Banksia arborea. Image adapted from Google Earth.
5.3.1.1 Atelomastix bamfordi
The spirostreptid millipede, A. bamfordi, is found on five of the Yilgarn BIFs; the Die Hardy,
Windarling, Helena Aurora and Koolyanobbing Ranges, and Mt Jackson. It occurs in mesic
habitat provided by shaded gullies and areas of dense leaf litter on BIFs but has not been
found on the flats between BIFs. The analysis used data from two sequenced regions of the
mtDNA genome (16S and CO1) and genotype data from 11, specifically developed, nuclear
microsatellite markers as outlined in Nistelberger et al. (2013, 2014). Millipedes were acquired
by searching in litter, under rocks etc. or tissue samples were taken from specimens lodged at
the Western Australian Museum. All available specimens were sequenced and genotyped as
previously described in Nistelberger et al. (2014).
119
Table 5.1. Details of the 10 Yilgarn banded iron formations (BIFs) sampled for this study and the presence (+) of the three species on each.
BIF Pop
code Latitude Longitude Atelomastix bamfordi Grevillea georgeana Banksia arborea
Die Hardy Range DH -29.89575 119.36179 + + +
Helena Aurora Range HA -30.34780 119.70820 + + +
Hunt Range HUN -30.20180 119.86395
+ +
Koolyanobbing Range KOO -30.84039 119.54244 +
+
Mt Correll COR -30.53926 118.90991
+
Mt Finnerty FIN -30.67577 120.13417
+
Mt Jackson MTJ -30.25508 119.27631 +
+
Mt Manning MM -29.99582 119.64289
+ +
Taipan Hill TH -30.39248 119.91232
+ +
Windarling Range WIN -30.01100 119.31392 + +
120
5.3.1.2 Grevillea georgeana
The bird-pollinated, proteaceous shrub, G. georgeana, was sampled from seven Yilgarn BIFs
the Helena Aurora, Die Hardy Range and Hunt Ranges, Mt Correll, Mt Manning, Mt Dimer and
Mt Finnerty. Analysis was based on sequence data from two intergenic chloroplast DNA
spacers (trnQ-rps16 and psbD-trnT) and 10 specifically developed nuclear microsatellite loci as
previously described in Nistelberger et al. (2013b; Chapter 3). Samples for the sequence data
were taken from 10 plants spaced out across the BIF as far as practicably possible, whereas the
samples for microsatellite genotyping were taken from a geographically clustered subset of 24
individuals from each BIF. The previous study (Chapter 3) analysed three subpopulations from
the Die Hardy Range, but only one subpopulation (northern) was used in this analysis to avoid
bias when estimating the contribution of a BIF population to overall genetic diversity.
5.3.1.3 Banksia arborea
Banksia arborea, a proteaceous, perennial tree that is likely pollinated by both birds and
insects, was sampled from eight Yilgarn BIFs, the Koolyanobbing, Mt Jackson, Windarling, Die
Hardy, Hunt and Helena Aurora Ranges, Mt Manning and Taipan Hill. Analysis was based on
sequence data from three chloroplast DNA regions (psbD-trnT, psbA-trnH, PetB) and genotype
data from 11 specifically developed nuclear microsatellite loci as previously described in
Nistelberger et al. (2013c; Chapter 4). Sampling was the same as for G. georgeana.
5.3.2 Diversity and differentiation analysis
The contributions of populations to total species allelic richness (CTR) and gene diversity (CT)
were calculated for both cytoplasmic DNA (mitochondrial (mtDNA) and chloroplast (cpDNA))
and nuclear DNA (nrDNA) datasets using the software CONTRIB (available at
http://www.pierroton.inra.fr/genetics/labo/Software/Contrib/). Total allelic richness and total
121
gene diversity measures are comprised of two components; within population diversity (allelic
richness = CSR, gene diversity = CS) and between population differentiation (allelic richness =
CDR, gene diversity = CD) respectively. The joint summation of the diversity (CSR, CS) and the
differentiation (CDR, CD) components determines if the overall contribution of a particular
population to diversity (allelic richness = CRT, gene diversity = CT) within the species is positive
or negative. These measures allow for the comparative assessment of the relative
contributions of populations to total genetic diversity. Populations that contributed the most
to each of these three measures (within population diversity, among population divergence
and total genetic diversity) in each species were identified (Fig. 5.2, 5.3, 5.4). Comparisons
across BIFs for all three species were then used to determine those BIFs that harbour
populations that contribute the most to total genetic diversity (i.e. diversity and
differentiation) of these species (Table 5.2).
5.4 RESULTS
5.4.1 Contribution to population diversity (CSR and CS)
5.4.1.1 Atelomastix bamfordi
The large Koolyanobbing BIF population made the greatest contribution to genetic diversity
across both chloroplast and nuclear data sets, based on both allelic richness and gene diversity
estimates (Fig. 5.2).
5.4.1.2 Grevillea georgeana
The large Die Hardy BIF population made the greatest contribution to genetic diversity based
on both allelic richness and gene diversity estimates in the cpDNA data and the Taipan Hill
population showed greatest diversity in the nrDNA data (Fig. 5.3).
122
5.4.1.3 Banksia arborea
Again, Die Hardy showed high genetic diversity in the cpDNA data based on both allelic
richness and gene diversity estimates but there were contrasting patterns in the nrDNA, with
the Mt Manning population contributing the most to gene diversity and the Mt Jackson
population contributing the most to allelic richness (Fig. 5.4).
5.4.2 Contribution to divergence (CDR and CD)
5.4.2.1 Atelomastix bamfordi
The small Windarling BIF population was the most genetically divergent from all other
populations in both chloroplast and nuclear data sets and across both diversity measures (Fig.
5.2).
5.4.2.2 Grevillea georgeana
Mt Correll, the small and spatially isolated BIF at the western boundary of the species’
distribution, was the most genetically divergent population based on both allelic richness and
gene diversity estimates, with the exception of the Mt Manning population in the nrDNA data
that showed more divergence using allelic richness (Fig. 5.3).
5.4.2.3 Banksia arborea
The cpDNA data showed both the Koolyanobbing and Taipan Hill populations to be the most
divergent in this species. Results from the nrDNA showed the population at the small
Windarling BIF to be the most divergent (Fig. 5.4).
123
5.4.3 Contribution to total genetic diversity (CTR and CT)
Total genetic diversity is the sum of a population’s contribution to within population diversity
(CSR, CS) and between population divergence (CDR, CD).
5.4.3.1 Atelomastix bamfordi
Contribution to both population diversity and divergence varied across the mitochondrial and
nuclear datasets and two estimates (gene diversity and allelic richness). In the mtDNA data,
the Die Hardy Range population was a high overall contributor using allelic richness as a
measure of diversity. In the gene diversity estimate all populations were equal overall
contributors owing to the population specificity of all haplotypes in this species. Populations at
Koolyanobbing and Windarling showed highest contributions in the nrDNA dataset using allelic
richness and gene diversity measures respectively (Fig. 5.2) (Table 5.2).
5.4.3.2 Grevillea georgeana
The Taipan Hill population showed greatest contribution to overall diversity in the nrDNA data
and the Die Hardy population in the cpDNA data, but populations at Mt Correll, Helena Aurora
and Mt Manning also made a high contribution based on the cpDNA gene diversity measure
(Fig. 5.3) (Table 5.2).
5.4.3.3 Banksia arborea
Both Koolyanobbing and Taipan Hill populations showed high overall contribution in the
cpDNA data but the nrDNA data showed the Windarling population contributes the most to
total genetic diversity based on allelic richness and gene diversity, while the Mt Manning
population also made a high contribution based on gene diversity (Fig. 5.4) (Table 5.2).
124
Figure 5.2 a-d Contribution of Atelomastix bamfordi populations to total diversity in terms of allelic richness (CTR) (a,b) and gene diversity (CT)(c,d) at both
mitochondrial (mtDNA)(a,c) and nuclear microsatellite (nrDNA)(b,d) datasets. NB. In the gene diversity mtDNA estimate all populations were equal overall
contributors owing to the population specificity of all haplotypes in this species. DH, Die Hardy; HA, Helena Aurora; KOO, Koolyanobbing; MTJ, Mt Jackson;
WIN, Windarling.
125
Figure 5.3 a-d Contribution of Grevillea georgeana populations to total diversity in terms of allelic richness (CTR) (a,b) and gene diversity (CT)(c,d) at both
chloroplast (cpDNA)(a,c) and nuclear microsatellite (nrDNA)(b,d) datasets. COR, Mt Correll; DH, Die Hardy; FIN, Mt Finnerty; HA, Helena Aurora; HUN, Hunt;
MM, Mt Manning; TH, Taipan Hill.
126
Figure 5.4 a-d Contribution of Banksia arborea populations to total diversity in terms of allelic richness (CTR) (a,b) and gene diversity (CT)(c,d) at both
chloroplast (cpDNA)(a,c) and nuclear microsatellite (nrDNA)(b,d) datasets. DH, Die Hardy; HA, Helena Aurora; HUN, Hunt; KOO, Koolyanobbing; MM, Mt
Manning; MTJ, Mt Jackson; TH, Taipan Hill; WIN, Windarling.
127
Table 5.2a Contributions (%) of populations of Atelomastix bamfordi, Grevillea georgeana and Banksia arborea to total genetic diversity, comprised of within population diversity and between population divergence using cytoplasmic DNA data. COR, Mt Correll; DH, Die Hardy; FIN, Mt Finnerty; HA, Helena Aurora; HUN, Hunt; KOO, Koolyanobbing; MM, Mt Manning; MTJ, Mt Jackson; TH, Taipan Hill; WIN, Windarling.
cytoplasmic DNA Atelomastix bamfordi Grevillea georgeana Banksia arborea contribution to total genetic diversity
Gene Diversity (CT)
Allelic Richness (CTR)
Gene Diversity (CT)
Allelic Richness (CTR)
Gene Diversity (CT)
Allelic Richness (CTR)
>30% - DH - DH KOO, TH KOO, TH
21-30% - KOO COR, DH, HA, MM HA, MM - -
11-20% - HA, MTJ - COR - DH
0-10% - WIN FIN, HUN, TH FIN, HUN, TH DH, HA, HUN, MM, MTJ, WIN HA, HUN, MM, MTJ, WIN
Table 5.2b. Contributions (%) of populations of Atelomastix bamfordi, Grevillea georgeana and Banksia arborea to total genetic diversity, comprised of within population diversity and between population divergence using microsatellite DNA data. COR, Mt Correll; DH, Die Hardy; FIN, Mt Finnerty; HA, Helena Aurora; HUN, Hunt; KOO, Koolyanobbing; MM, Mt Manning; MTJ, Mt Jackson; TH, Taipan Hill; WIN, Windarling.
microsatellite DNA Atelomastix bamfordi Grevillea georgeana Banksia arborea contribution to total genetic diversity
Gene Diversity (CT)
Allelic Richness (CTR)
Gene diversity (CT)
Allelic Richness (CTR)
Gene diversity (CT)
Allelic Richness (CTR)
>30% WIN KOO - TH - WIN
21-30% DH WIN TH - MM, WIN -
11-20% HA DH, MTJ COR, DH, FIN, HA, HUN DH, FIN, HUN MTJ MM, MTJ, TH
0-10% KOO, MTJ HA MM COR, HA, MM DH, HA, HUN, KOO, TH DH, HA, HUN, KOO
128
5.5 DISCUSSION
Genetic diversity and divergence are important for ongoing persistence and evolutionary
potential (Frankel & Soule, 1981; Faith, 1992). Our analysis of diversity in three endemic BIF
species revealed patterns that were largely congruent and consistent with expectations based
on size and isolation of BIF populations, allowing for the identification of key BIFs that can be
prioritised for conservation. Although the three species used in this study are not sympatric
across all 10 BIFs, the relative contributions of each BIF population to diversity and divergence
within each species showed similar trends with large, topographically complex BIFs supporting
populations with higher genetic diversity and small and/or isolated BIFs harbouring
populations with greater divergence.
5.5.1 Large, topographically complex BIFs support populations with higher genetic diversity
The large and topographically complex BIFs at Die Hardy Range and Koolyanobbing were
consistent in harbouring populations with high genetic variation in all three species. Genetic
theory and practical examples have reinforced the importance of large effective population
size in maintaining diversity by conferring resilience to bottlenecks, genetic drift, and
inbreeding (Wright, 1929; Ellstrand & Elam, 1993; Frankham, 1996; Allendorf et al., 2013).
Consequently, large populations, such as those at Die Hardy and Koolyanobbing, may play an
important role as genetic reservoirs within this landscape. High genetic diversity has also been
connected with high species diversity (Vellend & Geber, 2005). Whilst some surveys of species
diversity have been carried out in this landscape (Gibson et al., 2010), not all the BIFs we
evaluated have been assessed for species diversity and those that have were not surveyed
equally, so we were unable to correlate BIF species richness with genetic diversity in the
species we investigated.
129
We did not observe a high contribution of the populations from the large Helena Aurora Range
to genetic diversity, which is surprising since it is characterised by a high number of localised
and regional plant endemics making it of conservation significance in that context (Gibson et
al., 1997, 2007, 2010). However, large BIFs may not always support large populations of
particular species, and a similar study of two common plant species restricted to granite
outcrop islands in this region showed no relationship between outcrop area and population
diversity (Tapper et al. 2014a,b). Although large (approx. 12 km long) and elevated (600-700 m
height), the topology of the Helena Aurora Range with its gentle, rounded hills, creates a
different environment to the other large BIFs. Individuals of A. bamfordi were difficult to find
here in comparison to the Die Hardy and Koolyanobbing Ranges (Biota, 2009), and G.
georgeana was also less common in comparison to the diverse Die Hardy Range (it does not
occur on Koolyanobbing). Personal observations of the three species and their distributions,
suggest that it is not only size but complexity of habitats that are important for ensuring
persistence of large populations and associated genetic diversity in these species. Habitat
complexity has often been shown to be associated with species diversity (Kohn & Leviten,
1976; Cowling & Lombard, 2002), and has been associated with genetic diversity (Nistelberger
et al., 2014a) and divergence (Byrne et al., 2001; Byrne & Hopper, 2008) in some species.
Quantifying topographic and habitat complexity of BIFs is currently a difficult and expensive
exercise, yet increasingly this is being seen as an important factor in identifying areas that may
support higher biological diversity (Ashcroft et al., 2012; Schut et al., 2014). Recently
developed techniques based on the use of fine-scale topo-climatic grids, should enable more
precise quantification of these traits in the future (Ashcroft et al., 2012) and this would allow
the correlation between genetic diversity and habitat complexity to be investigated.
130
5.5.2 Small and spatially isolated BIFs harbour populations with greater genetic divergence
Populations on the small Windarling Range and at Mt Correll, along with those on the spatially
isolated (but diverse) Koolyanobbing Range, frequently made the greatest contributions to
genetic divergence within the species. In comparison to large populations, small populations
suffer increased susceptibility to population bottlenecks and genetic drift, generally resulting
in genetic divergence (Ellstrand & Elam, 1993). Peripheral or spatially isolated populations may
also show genetic divergence as a result of genetic isolation, genetic drift and natural selection
(Lesica & Allendorf, 1995). Peripheral populations can also show reduced genetic diversity as a
result of these processes (Lesica & Allendorf, 1995) yet the spatially isolated Koolyanobbing
Range had populations that contributed to divergence, as expected, but also maintained high
genetic diversity. This high genetic diversity is likely due to the large and complex nature of this
Range that would facilitate maintenance of larger population size over time. However, the
spatial isolation of this BIF has limited gene flow and driven divergence of populations, as
expected from genetic theory (Lesica & Allendorf, 1995).
5.5.3 Prioritisation of BIFs for conservation
While the comparisons across the three species show the pattern of distribution of genetic
diversity is not fully consistent, there is sufficient similarity to enable an assessment of overall
contribution to genetic diversity. From a landscape perspective, evaluation of genetic diversity
in populations of three species endemic to the BIFs in this region suggests that conservation of
the large Die Hardy and Koolyanobbing Ranges, along with the smaller Taipan Hill and
Windarling Range would capture maximum genetic diversity and distinctiveness of these three
species in this region, with populations in these four Ranges frequently contributing more than
30% of the total genetic diversity within a species. Thus, these four BIFs would rate higher in
terms of priority for conservation when evaluating competing land use practices. Identifying
regions of genetic diversity and distinctiveness can complement species-based approaches to
131
conservation through inclusion of an evolutionary perspective (Mace et al., 2003). This analysis
has revealed important patterns of genetic diversity present in this landscape and has
highlighted the characteristics of BIFs that might be expected to harbour high genetic diversity
(large and complex) and high genetic divergence (small and/or isolated). These patterns are
consistent with genetic theory and are likely to apply to other island landscapes.
There is always a degree of caution required when using this type of analysis as a basis for
prioritising populations and geographical areas for conservation (Petit et al., 1998). The use of
this approach is based on three major assumptions; the first is that the diversity observed at
neutral loci reflects functional diversity, the second is that diversity at the selected loci reflects
genome-wide diversity, and the third is that patterns in these three taxa reflect genetic
patterns in other taxa present in these locations. These issues have been discussed at length
by Petit et al. (1998) but despite the drawbacks, assessment of patterns of neutral genetic
diversity still provides one of the most accessible methods available to biologists for identifying
populations of conservation significance (Petit et al., 1998; Reed & Frankham, 2003). The data
obtained here conforms to theoretical expectations of the relationship between diversity and
size, and isolation and divergence; and identification of concordant patterns across different
species provides a basis for generalisations about patterns of genetic diversity in this
landscape. This information can be combined with evaluation of endemism and species
richness to provide a more comprehensive basis for decisions aiming to find a balance
between competing land uses in this region.
132
133
Where to from here…?
134
135
Concluding Remarks
“The one process now going on that will take millions of years to correct is
the loss of genetic and species diversity by the destruction of natural
habitats. This is the folly our descendants are least likely to forgive us.”
E. O. Wilson
Our natural landscapes are facing unprecedented levels of degradation, with approximately
two thirds of the planet’s ecosystems already considered degraded (Nellemann & Corcoran,
2010). As a consequence, there is increasing pressure for areas of conservation priority to be
identified (Riggs, 1990). The southern BIFs of the Yilgarn, with their high species richness and
endemism (Gibson et al., 2010, 2012) and presence of aboriginal sites (Department of
Aboriginal Affairs), are of great natural and indigenous value. There has been a long history of
prospecting, resource development, and pastoralism in this landscape, with gold mining
commencing in the 19th century and iron-ore production at the Koolyanobbing Range and
pastoralism commencing in the 1960’s (Lord & Trendall, 1976). Conservation in this region has
focussed on protecting threatened flora, referred to as Declared Rare Flora (DRF) under the
Wildlife Conservation Act 1953, and Priority Flora (Smith, 2013), with resource companies
required to satisfy a range of conditions that afford these species protection and continued
preservation (Department of Mines and Petroleum, 2006). This requirement has resulted in
the protection of threatened flora, such as Tetratheca paynterae and Ricinocarpos brevis
endemic to Windarling Range (Butcher et al., 2009; Environmental Protection Authority, 2012)
and a number of Priority Flora (Smith, 2013). The establishment of conservation parks and
reserves in the area, such as the Mt Manning Nature Reserve, has also contributed to
136
conservation efforts; however, these areas are not precluded from development activities
(Environmental Protection Authority, 2007).
While the protection of DRF and Priority Flora is an important aspect of conservation in this
landscape, to ensure the preservation of landscape ecosystem functioning, we need to
understand the wider evolutionary dynamics that drive patterns of genetic diversity and
structure in more common species. These are the species that form the backbone of the
unique BIF flora communities and those that may be required for restoration practices in the
future.
As discussed in Chapter 5, the best way to ensure the persistence and evolutionary potential of
species is to preserve the maximum amount of genetic diversity and genetic ‘distinctiveness’.
The common patterns of genetic diversity in the three taxa studied for my thesis have
highlighted the importance of maintaining large effective population sizes and topographic
complexity to enable preservation of maximum genetic diversity in BIF endemics. We have also
seen the importance of small or spatially isolated BIF populations in contributing towards
genetic divergence or ‘distinctiveness’ (Chapter 5). Ideally, these large and isolated BIFs would
receive conservation protection, yet the BIFs identified as high conservation value in this study
are also of high resource value, posing immense challenges for conservation (Gibson et al.,
2012).
Mining in the Yilgarn landscape often involves the partial removal of a BIF and may be
structured around requirements to protect DRF, such as T. paynterae. I suggest that the
conservation of disturbed BIFs should also focus on leaving intact portions of BIF that capture
maximum habitat complexity. Many of the endemic species found on these BIFs are reliant on
137
the multitude of habitats provided by topographic complexity including tors, gullies and
shaded sites. For example, the millipede Atelomastix bamfordi requires the moist pockets of
habitat provided by the shaded gullies and leaf litter that accumulates at the base of rocky BIF
slopes (Nistelberger et al., 2014a). Preservation of these habitats is likely to protect other
groups of short-range endemics that also depend upon these moist environments (e.g.
isopods, snails, mygalomorph spiders)(Harvey, 2002).
Post-disturbance
Prevention of habitat degradation in the first instance is better than attempting to restore
degraded environments (Menz et al 2013). Yet the reality remains that large tracts of land
worldwide currently require rehabilitation as a result of human disturbance (Merritt and Dixon
2011). Understanding how to restore the Yilgarn landscape post-disturbance, is another crucial
component to BIF biodiversity conservation. Restoration attempts in dryland regions, such as
Western Australia’s Pilbara region, have been problematic, with less than 15% of species
biodiversity being returned and, in some cases, failure to restore even the most common of
native plant species (Merritt & Dixon, 2011; Environmental Protection Authority, 2013). There
is a need for research and innovation into restoration ecology practices in these semi-arid
landscapes, a challenge that will require a bottom-up approach to rebuilding the landscape.
This will require getting the foundations of the rehabilitation site as close to its original state as
possible, in terms of substrate, aspect and drainage, before species are returned to the site.
Essentially, the foundation requires the return of topographic complexity. Undoubtedly, this is
a challenging task that requires financial investment and a period of trial and error, but it also
presents an exciting opportunity to think more creatively about restoration practices. Rather
than asking how can we improve the survival rate of replanted species, a current problem in
dryland rehabilitation (Merritt & Dixon, 2011), what about asking; how can we return large,
complex physical structure back to these restored hills? How can we re-create deep rock
138
fissures and cracks in which species such as the BIF specialist Tetratheca species can re-
germinate? And how can we ensure there are pockets of mesic habitat available for moisture-
dependent species such as A. bamfordi; habitat that receives higher shade and is covered in
dense leaf litter, or an appropriate analogue? These questions require a creative range of
engineering, biological and geological solutions that, whilst initially expensive, should result in
a higher rehabilitation success rate in terms of biodiversity return for these landscapes.
139
REFERENCES
Akaike H. (1974) A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19, 716–723.
Allendorf F.W., Luikart G., & Aitken S.N. (2013) Conservation and the Genetics of Populations. Wiley-Blackwell, West Sussex, UK.
Anand R. & Paine M. (2002) Regolith geology of the Yilgarn Craton, Western Australia: implications for exploration. Australian Journal of Earth Sciences, 49, 3–162.
Ashcroft M., Gollan J., Warton D., & Ramp D. (2012) A novel approach to quantify and locate potential microrefugia using topoclimate, climate stability, and isolation from the matrix. Global Change Biology, 18, 1866–1879.
Assefa A., Ehrich D., Taberlet P., Nemomissa S., & Brochmann C. (2007) Pleistocene colonization of afro-alpine “sky islands” by the arctic-alpine Arabis alpina. Heredity, 99, 133–42.
Avise J.C. (1989) Gene trees and organismal histories: A phylogenetic approach to population biology. Evolution, 43, 1192–1208.
Avise J.C. (2000) Phylogeography: The history and formation of species. Harvard University Press, Cambridge, Massachusetts.
Avise J.C., Arnold J., Ball R.M., Bermingham E., Lamb T., Neigel J.E., Reeb C.A., & Saunders N.C. (1987) Intraspecific phylogeography: The mitochondrial DNA bridge between population genetics and systematics. Annual Review of Ecology and Systematics, 18, 489–522.
Avise J.C., Shapira J.F., Daniel S.E., Aquador C.F., & Lansman R.A. (1983) Mitochondrial DNA differentiation during the speciation process in Peromyscus. Molecular Biology and Evolution, 1, 38–56.
Barbará T., Martinelli G., Fay M.F., Mayo S.J., & Lexer C. (2007) Population differentiation and species cohesion in two closely related plants adapted to neotropical high-altitude “inselbergs”, Alcantarea imperialis and Alcantarea geniculata (Bromeliaceae). Molecular Ecology, 16, 1981–92.
Barrett S.C.H. & Kohn J.R. (1991) Genetic and evolutionary consequences of small population size in plants: implications for conservation. Genetics and conservation of rare plants (ed. by D.A. Falk and K.E. Holsinger), pp. 3–30. Oxford University Press, New York, USA.
Baskin J.M. & Baskin C.C. (1988) Endemism in rock outcrop plant communities of unglaciated eastern United States: an evaluation of the roles of the edaphic, genetic and light factors. Journal of Biogeography, 15, 829–840.
Bataillon T.M., David J.L., & Schoen D.J. (1996) Neutral genetic markers and conservation genetics: simulated germplasm collections. Genetics, 144, 409–417.
140
Beatty G. & Provan J. (2011) Comparative phylogeography of two related plant species with overlapping ranges in Europe, and the potential effects of climate change on their intraspecific genetic diversity. BMC Evolutionary Biology, 11, 29.
Bech N., Boissier J., Drovetski S., & Novoa C. (2009) Population genetic structure of rock ptarmigan in the “sky islands” of French Pyrenees: implications for conservation. Animal Conservation, 12, 138–146.
Beheregaray L.B., Ciofi C., Caccone A., Gibbs J.P., & Powell J.R. (2003) Genetic divergence, phylogeography and conservation units of giant tortoises from Santa Cruz and Pinzon, Galapagos islands. Conservation Genetics, 4, 31–46.
Beheregaray L.B., Gibbs J.P., Havill N., Fritts T.H., Powell J.R., & Caccone A. (2004) Giant tortoises are not so slow: Rapid diversification and biogeographic consensus in the Galapagos. Proceedings of the National Academy of Sciences of, 101, 6514–6519.
Bennelongia (2008) Troglofauna Survey of Mt Jackson Range, Western Australia - Report 2008/50. Bennelongia Environmental Consultants, December 2005. .
Beveridge M. & Simmons L.W. (2004) Microsatellite loci for Dawson’s burrowing bee (Amegilla dawsoni) and their cross-utility in other Amegilla species. Molecular Ecology Notes, 4, 379–381.
Biota (2009) Mt Jackson J1 Deposit - Summary of Atelomastix sampling, August 2009. Biota Environmental Sciences, Letter, 9 November 2009. .
Boisselier-Dubayle M.-C., Leblois R., Samadi S., Lambourdiere J., & Sarthou C. (2010) Genetic structure of the xerophilous bromeliad Pitcairnia geyskesii on inselbers in French Guiana- a test of the forest refuge hypothesis. Ecography, 33, 175–194.
Bowler J.M. (1982) Aridity in the late Tertiary and Quaternary of Australia . Evolution of the flora and fauna of arid Australia (ed. by W.R. Barker and P.J.M. Greenslade), pp. 35–45. Peacock Publications, Frewville.
Brito P.H. & Edwards S. V (2009) Multilocus phylogeography and phylogenetics using sequence-based markers. Genetica, 135, 439–455.
Brower A. (1994) Rapid morphological radiation and convergence among races of the butterfly Heliconius erato inferred from patterns of mitochondrial DNA evolution. Proceedings of the National Academy of Sciences of the United States of America, 91, 6491–6495.
Brown W.M. (1983) Evolution of animal mitochondrial DNA. Evolution of Genes and Proteins (ed. by M. Nei and K. Koehn), pp. 62–88. Sinauer, Sunderland, MA.
Buckley T., Simon C., & Chambers G. (2001) Exploring among-site rate variation models in maximum likelihood framework using empirical data. Effects of model assumptions on estimates of topology branch lengths and bootstrap support. Systematic Biology, 50, 67–86.
Burke A. (2003) Inselbergs in a changing world-global trends. Diversity & Distributions, 9, 375–383.
141
Burne H.M., Yates C.J., & Ladd P.G. (2003) Comparative population structure and reproductive biology of the critically endangered shrub Grevillea althoferorum and two closely related more common congeners. Biological Conservation, 114, 53–65.
Busch J., Waser P., & Dewoody J. (2007) Recent demographic bottlenecks are not accompanied by a genetic signature in banner-tailed kangaroo rats (Dipodomys spectabilis). Molecular Ecology, 16, 2450–2462.
Butcher P.A., McNee S.A., & Krauss S.L. (2009) Genetic impacts of habitat loss on the rare ironstone endemic Tetratheca paynterae subsp. paynterae. Conservation Genetics, 10, 1735–1746.
Butcher R., Byrne M., & Crayn D.M. (2007) Evidence for convergent evolution among phylogenetically distant rare species of Tetratheca (Elaeocarpaceae, formerly Tremandraceae) from Western Australia. Australian Systematic Botany, 20, 126–138.
Byrne M. (2007) Phylogeography provides an evolutionary context for the conservation of a diverse and ancient flora. Australian Journal of Botany, 55, 316–325.
Byrne M., Elliott C.P., Yates C., & Coates D.J. (2007) Extensive pollen dispersal in a bird-pollinated shrub, Calothamnus quadrifidus, in a fragmented landscape. Molecular ecology, 16, 1303–14.
Byrne M. & Hankinson M. (2012) Testing the variability of chloroplast sequences for plant phylogeography. Australian Journal of Botany, 60, 569–574.
Byrne M. & Hines B. (2004) Phylogeographical analysis of cpDNA variation in Eucalyptus loxophleba (Myrtaceae). Australian Journal of Botany, 52, 459–470.
Byrne M. & Hopper S. (2008) Granite outcrops as ancient islands in old landscapes: evidence from the phylogeography and population genetics of Eucalyptus caesia (Myrtaceae) in Western Australia. Biological Journal of the Linnean Society, 93, 177–188.
Byrne M., MacDonald B., Broadhurst L., & Brand J. (2003) Regional genetic differentiation in Western Australian sandalwood (Santalum spicatum) as revealed by nuclear RFLP analysis. Theoretical and Applied Genetics, 107, 1208–1214.
Byrne M., MacDonald B., & Coates D.J. (2002) Phylogeographical patterns in chloroplast DNA variation within the Acacia acuminata (Leguminosae: Mimosoideae) complex in Western Australia. Journal of Evolutionary Biology, 15, 576–587.
Byrne M., Tischler G., Macdonald B., Coates D.J., & McComb J.A. (2001) Phylogenetic relationships between two rare acacia and their common, widespread relatives in south-western Australia. Conservation Genetics, 2, 157–166.
Byrne M., Yeates D.K., Joseph L., Kearney M., Bowler J., Williams M.A.J., Cooper S., Donnellan S.C., Keogh J.S., Leys R., Melville J., Murphy D.J., Porch N., & Wyrwoll K.H. (2008) Birth of a biome: insights into the assembly and maintenance of the Australian arid zone biota. Molecular Ecology, 17, 4398–4417.
Cardillo M. & Pratt R. (2013) Evolution of a hotspot genus: geographic variation in speciation and extinction rates in Banksia (Proteaceae). BMC evolutionary biology, 13, 155.
142
Carstens B.C., Brunsfeld S.J., Demboski J.R., Good J.M., & Sullivan J. (2005) Investigating the evolutionary history of the pacific northwest mesic forest ecosystem: Hypothesis testing within a comparative phylogeographic framework. Evolution, 59, 1639–1652.
Cavalli-Sforza L.L. & Edwards A.W.F. (1967) Phylogenetic analysis: models and estimation procedures. The American Journal of Human Genetics, 19, 233–257.
Chen C., Durand E., Forbes F., & Francois O. (2007) Bayesian clustering algorithms ascertaining spatial population structure: a new computer program and a comparison study. Molecular Ecology Notes, 7, 747–756.
Chivas A.R. & Atlhopheng J.R. (2010) Oxygen-isotope dating the Yilgarn regolith. Australian Landscapes (ed. by P. Bishop and B. Pillans), pp. 309–320. The Geological Society of London, London.
Clark P.U., Archer D., Pollard D., Blum J.D., Rial J.A., Brovkin V., Mix A.C., Pisias N.G., & Roy M. (2006) The middle Pleistocene transition: characteristics, mechanisms, and implications for long-term changes in atmospheric pCO2. Quaternary Science Reviews, 25, 3150–3184.
Clement M., Posada D., & Crandall K.A. (2000) TCS: a computer program to estimate gene genealogies. Molecular Ecology, 9, 1657–1660.
Comps B., Gömöry D., Letouzey J., Thiébaut B., & Petit R.J. (2001) Diverging trends between heterozygosity and allelic richness during postglacial colonization in the European beech. Genetics, 157, 389–97.
Cooper S.J.B., Harvey M.S., Saint K.M., & Main B.Y. (2011) Deep phylogeographic structuring of populations of the trapdoor spider Morggridgea tingle (Migidae) from southwestern Australia: evidence for long-term refugia within refugia. Molecular Ecology, 20, 3219–3236.
Cowling R.M. & Lombard A.T. (2002) Heterogeneity, speciation/extinction history and climate: explaining regional plant diversity patterns in the Cape Floristic Region. Diversity & Distributions, 8, 163–179.
Cox C.B. & Moore P.D. (2000) Biogeography: An ecological and evolutionary approach. Blackwell Science Ltd, Oxford.
Crandall K.A., Bininda-Emonds O.R.P., Mace G.M., & Wayne R.K. (2000) Considering evolutionary processes in conservation biology. Trends in Ecology & Evolution, 15, 290–295.
Crandall K.A. & Templeton A.R. (1993) Empirical tests of some predictions from coalescent theory with applications to intraspecific phylogeny reconstruction. Genetics, 134, 959–69.
Crisp M., Cook L., & Steane D. (2004) Radiation of the Australian flora: what can comparisons of molecular phylogenies across multiple taxa tell us about the evolution of diversity in present–day communities? Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences, 359, 1551–1571.
Crozier R.H. (1992) Genetic diversity and the agony of choice. Biological Conservation, 61, 11–15.
143
Davis E.B., Koo M.S., Conroy C., Patton J.L., & Moritz C. (2008) The California Hotspots Project: identifying regions of rapid diversification of mammals. Molecular ecology, 17, 120–38.
DeChaine E.G. & Martin A.P. (2004) Historic cycles of fragmentation and expansion in Parnassius smintheus (Papilionidae) inferred using mithochondrial DNA. Evolution, 58, 113–127.
DeChaine E.G. & Martin A.P. (2005) Marked genetic divergence among sky island populations of Sedum lanceolatum (Crassulaceae) in the Rocky Mountains. American Journal of Botany, 92, 477–486.
Degen B., Bandou E., & Caron H. (2004) Limited pollen dispersal and biparental inbreeding in Symphonia globulifera in French Guiana. Heredity, 93, 585–91.
Department of Mines and Petroleum (2006) Management of Declared Rare Flora for onshore petroleum and mineral activities. .
Dobzhansky T.H. (1970) Genetics and the evolutionary process. Columbia University Press, New York.
Dodson J.R. & Macphail M.K. (2004) Palynological evidence for aridity events and vegetation change during the Middle Pliocene, a warm period in Southwestern Australia. Global and Planetary Change, 41, 285–307.
Driscoll D.A. (1998) Genetic structure of the frogs Geocrinia lutea and Geocrinia rosea reflects extreme population divergence and range changes, not dispersal barriers. Evolution, 52, 1147–1157.
Drummond A.J., Suchard M.A., Xie D., & Rambaut A. (2012) Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution, 29, 1969–1973.
Duputié A., Delêtre M., De Granville J.J., & McKey D. (2009) Population genetics of Manihot esculenta ssp. flabellifolia gives insight into past distribution of xeric vegetation in a postulated forest refugium area in northern Amazonia. Molecular ecology, 18, 2897–907.
Edward K. & Harvey M.S. (2010) A review of the Australian millipede genus Atelomastix (Diplopoda: Spirostreptida: Iulomorphidae). Zootaxa, 2371, 1–63.
Edwards D.L., Roberts D.J., & Keogh S.J. (2007) Impact of Plio-Pleistocene arid cycling on the population history of a southwestern Australian frog. Molecular Ecology, 16, 2782–2796.
Ehrlich P.R. (1988) The loss of diversity: causes and consequences. Biodiversity (ed. by E.O. Wilson), pp. 21–27. National Academy Press, Washington, D.C.
Ellstrand N.C. & Elam D.R. (1993) Population genetic consequences of small population size: implications for plant conservation. Annual Review of Ecology and Systematics, 24, 217–242.
England P.R., Whelan R.J., & Ayre D.J. (2003) Effects of seed bank disturbance on the fine-scale genetic structure of populations of the rare shrub Grevillea macleayana. Heredity, 91, 475–480.
144
Ennos R.A. (1994) Estimating the relative rates of pollen and seed migration among plant populations. Heredity, 72, 250–259.
Environmental Protection Authority (2007) Advice on areas of the highest conservation value in the proposed extensions to Mount Manning Nature Reserve. .
Environmental Protection Authority (2012) Yilgarn Operations - Windarling Range W4 East Deposit. .
Environmental Protection Authority (2013) Environmental Protection Authority 2012-2013 Annual Report. .
Evanno G., Regnaut S., & Goudet J. (2005) Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology, 14, 2611–2620.
Excoffier L. & Lischer H.E.L. (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.
Faith D.P. (1992) Conservation evaluation and phylogenetic diversity. Biological Conservation, 61, 1–10.
Floyd C.H., Van Vuren D.H., & May B. (2005) Marmots on Great Basin mountaintops: using genetics to test a biogeographic paradigm. Ecology, 86, 2145–2153.
Folmer O., Black M., Hoeh W., Lutz R., & Vrijenhoek R. (1994) DNA primers for amplification of mitochondrial Cytochrome C Oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology, 3, 294–299.
Ford H.A. (1981) Territorial behaviour in an Australian nectar-feeding bird. Australian Journal of Ecology, 6, 131–134.
Ford H.A., Geering D., & Ley A. (2000) Radio-tracking trials with regent honeyeaters Xanthomyza phrygia and other honeyeaters. Corella, 24, 25–29.
Forest F., Grenyer R., Rouget M., Davies T.J., Cowling R.M., Faith D.P., Balmford A., Manning J.C., Procheş S., van der Bank M., Reeves G., Hedderson T.A.J., & Savolainen V. (2007) Preserving the evolutionary potential of floras in biodiversity hotspots. Nature, 445, 757–60.
Frankel O.H. & Soule M.E. (1981) Conservation and evolution. Cambridge University Press, Cambridge, England.
Frankham R. (1996) Relationship of genetic variation to population size in wildlife. Conservation Biology, 10, 1500–1508.
Frankham R., Ballou J.D., & Briscoe D.A. (2002) Introduction to conservation genetics. Cambridge University Press, Cambridge, UK.
Freeman M.J. (2001) Avon River paleodrainage system, Western Australia: geomorphological evolution and environmental issues related to geology. Gondwana to Greenhouse: Australian Environmental Geoscience (ed. by V.A. Gostin), pp. 37–47. Geological Society of Australia. Special Publication, 21.,
145
Fujioka T., Chappell J., Fifield L.K., & Rhodes E.J. (2009) Australian desert dune fields initiated with Pliocene-Pleistocene global climatic shift. Geology, 37, 51–54.
Gamble T., Bauer A.M., Greenbaum E., & Jackman T.R. (2008) Evidence for Gondwanan vicariance in an ancient clade of gecko lizards. Journal of Biogeography, 35, 88–104.
Gerber A.S., Loggins R., Kumar S., & Dowling T.E. (2001) Does nonneutral evolution shape observed patterns of DNA variation in animal mitochondrial genomes? Annual Review of Genetics, 35, 539.
Gibson N., Coates D.J., & Thiele K.R. (2007) Taxonomic research and the conservation status of flora in the Yilgarn Banded Iron Formation ranges. Nuytsia, 17, 1–12.
Gibson N., Lyons M.N., & Lepschi B.J. (1997) Flora and vegetation of the eastern goldfields ranges, Part I: Helena and Aurora Range. CALMScience, 2, 231–246.
Gibson N., Meissner R., Markey A.S., & Thompson W.A. (2012) Patterns of plant diversity in ironstone ranges in arid south western Australia. Journal of Arid Environments, 77, 25–31.
Gibson N., Yates C., & Dillon R. (2010) Plant communities of the ironstone ranges of south western Australia: hotspots for plant diversity and mineral deposits. Biodiversity and Conservation, 19, 3951–3962.
Gill A.M. (1981) Adaptive responses of Australian vascular plant species to fire. Fire and the Australian Biota pp. 243–272.
Gitzendanner M.A. & Soltis P.S. (2000) Patterns of genetic variation in rare and widespread plant congeners. American Journal of Botany, 87, 783–792.
Goldstein D.B. & Pollock D.D. (1997) Launching microsatellites. Journal of Heredity, 88, 3235–3242.
Goudet J. (2001) FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9.3). .
Graur D. & Li W.H. (2000) Fundamentals of Molecular Evolution. Sinauer Associates Inc., Sunderland, Massachusetts.
Gyllensten U., Wharton D., Josefsson A., & Wilson A.C. (1991) Paternal inheritance of mitochondrial DNA in mice. Nature, 352, 255–257.
Haffer J. (1969) Speciation in Amazonian forest birds. Science, 165, 131–137.
Hall T.A. (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symposium, 41, 95–98.
Hamrick J.L. & Godt M.J.W. (1989) Allozyme diversity in plant species. Plant population genetics, breeding and genetic resources (ed. by A.H.D. Brown, M.T. Clegg, A.L. Kahler, and B.S. Weir), pp. 43–63. Sinauer, Sunderland, MA.
Hare M.P. (2001) Prospects for nuclear gene phylogeography. Trends in Ecology & Evolution, 16, 700–706.
146
Harris S.A. & Ingram R. (1991) Chloroplast DNA and biosystematics: The effects of intraspecific diversity and plastid transmission. Taxon, 40, 393–412.
Harvey M.S. (2002) Short-range endemism in the Australian fauna: some examples from non-marine environments. Invertebrate Systematics, 16, 555–570.
He T., Krauss S.L., Lamont B.B., Miller B.P., & Enright N.J. (2004) Long-distance seed dispersal in a metapopulation of Banksia hookeriana inferred from a population allocation analysis of amplified fragment length polymorphism data. Molecular ecology, 13, 1099–109.
He T., Lamont B.B., Krauss S.L., Enright N.J., & Miller B.P. (2009) Long-distance dispersal of seeds in the fire-tolerant shrub Banksia attenuata. Ecography, 32, 571–580.
Hebert P.D.N., Penton E.H., Burns J.M., Janzen D.H., & Hallwachs W. (2004) Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator. Proceedings of the National Academy of Sciences of the United States of America, 101, 14812–7.
Hebert P.D.N., Ratnasingham S., & DeWaard J.R. (2003) Barcoding animal life: cytochrome c oxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society London B (suppl.), 270, S96–9.
Hedrick P.W. (2001) Conservation genetics: where are we now? Trends in Ecology & Evolution, 16, 629–636.
Heinrich B. & Raven P.H. (1972) Energetics and pollination ecology. Science, 176, 597–602.
Hewitt G. (2000) The genetic legacy of the Quaternary ice ages. Nature, 405, 907.
Hewitt G.M. (1996) Some genetic consequences of ice ages, and their role in divergence and speciation. Biological journal of the Linnean Society, 58, 247–267.
Hey J. (2010) Isolation with migration models for more than two populations. Molecular Biology and Evolution, 27, 905–920.
Hey J. & Nielsen R. (2004) Multilocus methods for estimating population sizes, migration rates and divergence time, with applications to the divergence of Drosophila psuedoobscura and D. persimilis. Genetics, 167, 747–760.
Hey J. & Nielsen R. (2007) Integration within the Felsenstein equation for improved Markov chain Monte Carlo methods in population genetics. Proceedings of the National Academy of Sciences of the United States of America, 104, 2785–2790.
Hillis D.M. & Dixon M.T. (1991) Ribosomal DNA : molecular evolution and phylogenetic inference. The Quarterly Review of Biology, 66, 411–453.
Hipkins V.D., Krutovskii K. V, & Strauss S.H. (1994) Organelle genomes in conifers: structure, evolution and diversity. Forest Genetics, 1, 179–189.
Hocking R.M., Langford R.L., Thorne A.M., Sanders A.J., Morris P.A., Strong C.A., & Gozzard J.R. (2007) A classification system for regolith in Western Australia. Western Australia Geological Survey, 2007/8, 19.
147
Hopper S.D. (2000) Floristics of Australian granitoid inselberg vegetation. Inselbergs (ed. by S. Porembeski and W. Barthlott), pp. 391–407. Springer-Verlag, Berlin.
Hopper S.D. & Gioia P. (2004) The southwest Australian floristic region: Evolution and conservation of a global hot spot of biodiversity. Annual Review of Ecology, Evolution, and Systematics, 35, 623–650.
Hopper S.D., Harvey M.S., Chappill J.A., Main A.R., & Main B.Y. (1996) The Western Australian biota as Gondwanan heritage - a review. Gondwanan Heritage: Past, present and future of the Western Australian Biota (ed. by S.D. Hopper, J.A. Chappill, M.S. Harvey, and A.S. George), pp. 1–46. Surrey Beatty & Sons Pty limited, Chipping Norton, NSW.
Hopper S.D. & Moran G.F. (1981) Bird pollination and the mating system of Eucalyptus stoatei. Australian Journal of Botany, 29, 625–638.
Hoskin C.J., Higgie M., McDonald K.R., & Moritz C. (2005) Reinforcement drives rapid allopatric speciation. Nature, 437, 1353–6.
Huelsenbeck J.P. & Ronquist F. (2001) MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics, 17, 754–755.
Hughes L. (2003) Climate change and Australia: Trends, projections and impacts. Austral Ecology, 28, 423–443.
Jacobi C.M. & Carmo F.F. (2008) The contribution of ironstone outcrops to plant diversity in the Iron Quadrangle, a threatened Brazilian landscape. Ambio, 37, 324–326.
Jacobi C.M., Carmo F.F., Vincent R.C., & Stehmann J.R. (2007) Plant communities on ironstone outcrops: a diverse and endangered Brazilian ecosystem. Biodiversity and Conservation, 16, 2185–2200.
Jakobsson M. & Rosenberg N.A. (2007) CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics, 23, 1801–1806.
Jensen J.L., Bohonak A.J., & Kelley S.T. (2005) Isolation by distance, web service. BMC Genetics, 6, .
Ji Y.-J., Zhang D.-X., & He L.-J. (2003) Evolutionary conservation and versatility of a new set of primers for amplifying the ribosomal internal transcribed spacer regions in insects and other invertebrates. 3, 581–585.
Jones A.W. & Kennedy R.S. (2008) Evolution in a tropical archipelago: comparative phylogeography of Philippine fauna and flora reveals complex patterns of colonization and diversification. Biological journal of the Linnean Society, 95, 620–639.
Jost L. (2008) GST and its relatives do not measure differentiation. Molcecular Ecology, 17, 4015–4026.
Juan A., Crespo M.B., Cowan R.S., Lexer C., & Fay M.F. (2004) Patterns of variability and gene flow in Medicago citrina, an endangered endemic of islands in the western Mediterranean, as revealed by amplified fragment length polymorphism (AFLP). Molecular Ecology, 13, 2679–2690.
148
Jukes T.H. & Cantor C.R. (1969) Evolution of protein molecules. Mammalian Protein Metabolism (ed. by H.N. Munro), pp. 21–132. Academic Press, New York.
Kalinowski S.T. (2005) HP-RARE 1.0: a computer program for performing rarefaction on measures of allelic richness. Molecular Ecology Notes, 5, 187–189.
Kalinowski S.T. & Taper M.L. (2006) Maximum likelihood estimation of the frequency of null alleles at microsatellite loci. Conservation Genetics, 7, 991–995.
Keir K.R., Bemmels J.B., & Aitken S.N. (2011) Low genetic diversity, moderate local adaptation, and phylogeographic insights in Cornus nuttallii (Cornaceae). American Journal of Botany, 98, 1327–1336.
Kelly R.P. (2010) The use of population genetics in Endangered Species Act listing decisions. Ecology Law Quarterly, 37, 1107–1158.
Kim I., Phillips C.J., Monjeau J.A., Birney E.C., Noack K., Pumo D.E., Sikes R.S., & Dole J.A. (1998) Habitat islands, genetic diversity, and gene flow in a Patagonian rodent. Molecular Ecology, 7, 667–678.
Kimura M. & Weiss G.H. (1964a) The stepping stone model of population structure and the decrease of genetic correlation with distance. Genetics, 49, 561–576.
Kimura M. & Weiss G.H. (1964b) The stepping stone model of population structure and the decrease of genetic correlation with distance. Genetics, 49, 561–576.
Klein C. (2005) Some Precambrian banded iron-formations (BIFs) from around the world: Their age, geologic setting, mineralogy, metamorphism, geochemistry, and origins. American Mineralogist, 90, 1473–1499.
Knowles L.L. (2000) Tests of Pleistocene speciation in montane grasshoppers (genus Melanoplus) from the sky islands of western North America. Evolution, 54, 1337–48.
Kohn A.J. & Leviten P.J. (1976) Effect of habitat complexity on population density and species richness in tropical intertidal predatory gastropod assemblages. Oecologia, 25, 199–210.
Kondo R., Satta Y., & Matsuura E.T. (1990) Incomplete maternal transmission of mitochondrial DNA in Drosophila. Genetics, 126, 657–663.
Krauss S.L., He T., Lamont B.B., Miller B.P., & Enright N.J. (2006) Late Quarternary climate change and spatial genetic structure in the shrub Banksia hookeriana. Molecular Ecology, 15, 1125–1137.
Ladd P.G., Alkema A.J., & Thomson G.J. (1996) Pollen Presenter Morphology and Anatomy in Banksia and Dryandra. Australian Journal of Botany, 44, 447–471.
Langella O. (2000) Populations 1.2.19: population genetic software individuals or populations distances, phylogenetic trees. .
Lee C.E. (2000) Global phylogeography of a cryptic copepod species complex and reproductive isolation between genetically proximate “populations”. Evolution, 54, 2014–27.
149
Lesica P. & Allendorf F.W. (1995) When are peripheral populations valuable for conservation? Conservation Biology, 9, 753–760.
Levy E., Kennington J.W., Tomkins J.L., & LeBas N.R. (2012) Phylogeography and population genetic structure of the Ornate Dragon Lizard, Ctenophorus ornatus. PLoS ONE, 7, e46351.
Levy F. & Neal C. (1999) Spatial and temporal genetic structure in chloroplast and allozyme markers in Phacelia dubia implicate genetic drift. Heredity, 82, 422–31.
Librado P. & Rozas J. (2009) DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics, 25, 1451–1452.
Llorens T.M. (2004) Conservation genetics and ecology of two rare Grevillea species. PhD Thesis. University of Wollonong, New South Wales, Australia,
Llorens T.M., Ayre D.J., & Whelan R.J. (2004) Evidence for ancient genetic subdivision among recently fragmented populations of the endangered shrub Grevillea caleyi (Proteaceae). Heredity, 92, 519–526.
Llorens T.M., Byrne M., Yates C.J., Nistelberger H.M., & Coates D.J. (2012) Evaluating the influence of different aspects of habitat fragmentation on mating patterns and pollen dispersal in the bird-pollinated Banksia sphaerocarpa var. caesia. Molecular ecology, 21, 314–28.
Lord J.H. & Trendall A.F. (1976) Iron-ore deposits of Western Australia- Geology and Development: Minerals. Circum-Pacific Energy and Mineral Resources pp. 410–417.
Loveless M.D. & Hamrick J.L. (1984) Ecological determinants of genetic structure in plant populations. Annual Review of Ecology and Systematics, 15, 65–95.
Lowe W.H. & Allendorf F.W. (2010) What can genetics tell us about population connectivity? Molecular Ecology, 19, 3038–3051.
Luikart G., Allendorf F.W., Cornuet J., & Sherwin W.B. (1998) Distortion of allele frequency distributions provides a test for recent population bottlenecks. The Journal of Heredity, 89, 238–247.
Mace G.M., Gittleman J.L., & Purvis A. (2003) Preserving the tree of life. Science, 300, 1707–9.
Margoliash E. (1963) Primary structure and evolution of cytochrome c. Proceedings of the National Academy of Sciences of the United States of America, 50, 672–679.
Markey A.S. & Dillon S.J. (2008) Flora and vegetation of the banded ironstone formation of the Yilgarn Craton: the central Tallering land system. Conservation Science Western Australia, 7, 121–149.
Marshall J.C., Kingsbury B.A., & Minchella D.J. (2008) Microsatellite variation, population structure, and bottlenecks in the threatened copperbelly water snake. Conservation Genetics, 10, 465–476.
Martin H.A. (2006) Cenozoic climatic change and the development of the arid vegetation in Australia. Journal of Arid Environments, 66, 533–563.
150
Masta S.E. (2000) Phylogeography of the jumping spider Habronattus pugillis (Araneae: Salticidae): Recent vicariance of sky island populations? Evolution, 54, 1699–1711.
Mayr E. (1970) Populations, Species and Evolution. Belknap Press, Cambridge, Massachusetts.
McCauley D.E. (1995) The use of chloroplast DNA polymorphism in studies of gene flow in plants. Trends in ecology & evolution, 10, 198–202.
McCauley D.E., Sundby A.K., Bailey M.F., & Welch M.E. (2007) Inheritance of chloroplast DNA is not strictly maternal in Silene vulgaris (Caryophyllaceae): evidence from experimental crosses and natural populations. American Journal of Botany, 94, 1333–1337.
McGillivray D. & Makinson R.O. (1993) Grevillea, Proteaceae: a taxonomic revision. Melbourne University Press at the Miegunyah Press, Carlton, Victoria.
McLoughlin S. (2001) The breakup history of Gondwana and its impact on pre-Cenozoic floristic provincialism. Australian Journal of Botany, 49, 271–300.
Measey G.J., Vences M., Drewes R.C., Chiari Y., Melo M., & Bourles B. (2007) Freshwater paths across the ocean: molecular phylogeny of the frog Ptychadena newtoni gives insights into amphibian colonization of oceanic islands. Journal of Biogeography, 34, 7–20.
Meissner R. & Caruso Y. (2008a) Flora and vegetation of the banded ironstone formation of the Yilgarn Craton: Mt Gibson. Conservation Science Western Australia, 7, 105–120.
Meissner R. & Caruso Y. (2008b) Flora and vegetation of the banded ironstone formations of the Yilgarn Craton: Koolanooka and Perenjori Hills. Conservation Science Western Australia, 7, 73–88.
Merritt D.J. & Dixon K.W. (2011) Restoration seed banks — a matter of scale. Science, 332, 21–22.
Mesibov R. (1994) Faunal breaks in Tasmania and their significance for invertbrate conservation. Memoirs of the Queensland Museum, 36, 133–136.
Millar M.A., Coates D.J., & Byrne M. (2013) Genetic connectivity and diversity in inselberg populations of Acacia woodmaniorum, a rare endemic of the Yilgarn Craton banded iron formations. Heredity, 111, 437–44.
Mills L.S. & Allendorf F.W. (1996) The one-migrant-per-generation rule in conservation and management. Conservation Biology, 10, 1509–1518.
Moir M.L., Brennan K.E.C., & Harvey M.S. (2009) Diversity, endemism and species turnover of millipedes within the south-western Australia global biodiversity hotspot. Journal of Biogeography, 36, 1958–1971.
Moore A.J., Merges D., & Kadereit J.W. (2013) The origin of the serpentine endemic Minuartia laricifolia subsp. ophiolitica by vicariance and competitive exclusion. Molecular ecology, 22, 2218–31.
Moran G.F. & Hopper S.D. (1983) Genetic diversity and the insular population structure of the rare granite rock species, Eucalyptus caesia Benth. Australian Journal of Botany, 31, 161–172.
151
Moritz C. (1994) Defining “Evolutionary Significant Units” for conservation. Trends in Ecology & Evolution, 9, 373–375.
Moritz C. & Faith D.P. (1998) Comparative phylogeography and the identification of genetically divergent areas for conservation. Molecular Ecology, 7, 419–429.
El Mousadik A. & Petit R.J. (1996) High level of genetic differentiation for allelic richness among populations of the argan tree [Argania spinosa (L.) Skeels] endemic to Morocco. Theoretical and Applied Genetics, 92, 832–839.
Muller K. (2005) SeqState-primer design and sequence statistics for phylogenetic DNA data sets. Applied Bioinformatics, 4, 65–69.
Myers J.S. (1993) Precambrian history of the West Australian Craton and adjacent orogens. Annual Review of Earth and Planetary Sciences, 21, 453–485.
Nei M. (1978) Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics, 89, 583–590.
Nellemann C. & Corcoran E. (eds) (2010) Dead planet, living planet - Biodiversity and ecosystem restoration for sustainable development. A rapid response assessment. United Nations Environment Programme, GRID-Arendal.
Nistelberger H.M., Byrne M., Coates D.J., & Roberts J.D. (2014a) Strong phylogeographic structure in a millipede indicates Pleistocene vicariance between populations on banded iron formations in semi-arid Australia. PLoS ONE, 9, e93038.
Nistelberger H.M., Byrne M., Coates D.J., & Roberts J.D. (2014b) Terrestrial islands in an ancient landscape: Evidence of both historical isolation and connectivity of populations of Grevillea georgeana, a banded iron formation endemic. Journal of Biogeography, in review, .
Nistelberger H.M., Byrne M., Roberts J.D., & Coates D.J. (2013a) Isolation and characterisation of 11 microsatellite loci from the Western Australian Spirostreptid millipede, Atelomastix bamfordi. Conservation Genetics Resources, 5, 533–535.
Nistelberger H.M., Coates D.J., Byrne M., & Roberts J.D. (2013b) Isolation and characterisation of 11 microsatellite loci in the short-range endemic shrub Grevillea georgeana McGill (Proteaceae). Conservation Genetic Resources, 6, 221–222.
Nistelberger H.M., Roberts J.D., Coates D.J., & Byrne M. (2013c) Isolation and characterisation of 14 microsatellite loci from a short-range endemic, Western Australian tree, Banksia arborea (C.A. Gardner). Conservation Genetics Resources, 5, 1143–1145.
Olde P. & Marriott N. (1995) The Grevillea Book. Kangaroo Press Ltd, Kenthurst, NSW.
Paetkau D. (1999) Using genetics to identify intraspecific conservation units: a critique of current methods. Conservation Biology, 13, 1507–1509.
Page R.D.M. (1996) TREEVIEW: an application to display phylogenetic trees on personal computers. Computer Applications in the Biosciences, 12, 357–358.
152
Palmer J.D., Jansen R.K., Michaels H.J., Chase M.W., & Manhart J.R. (1988) Chloroplast DNA variation and plant phylogeny. Annals of the Missouri Botanical Garden, 75, 1180–1206.
Payne A.L., Vreeswyk A.M.E., Pringle H.J.R., Leighton K.A., & Henning P. (1998) An inventory and vegetation condition survey of the Sandstone-Yalgoo-Payne Finds area Western Australia. Agriculture Western Australia, South Perth.
Peakall R. & Smouse P.E. (2006) GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes, 6, 288–295.
Petit R.J., Mousadik A.E., & Pons O. (1998) Identifying populations for conservation on the basis of genetic markers. Conservation Biology, 12, 844–855.
Piry S., Luikart G., & Cornuet J. (1999) Bottleneck: a computer program for detecting recent reductions in the effective population size using allele frequency data. Journal of Heredity, 90, 502–503.
Pons J., Ribera I., Bertranpetit J., & Balke M. (2010) Nucleotide substitution rates for the full set of mitochondrial protein-coding genes in Coleoptera. Molecular Phylogenetics and Evolution, 56, 796–807.
Posada D. (2008) jModelTest: Phylogenetic Model Averaging. Molecular Biology and Evolution, 25, 1253–1256.
Pritchard J.K., Stephens M., & Donnelly P. (2000) Inference of population structure using multilocus genotype data. Genetics, 155, 945–959.
Pritchard J.K. & Wen W. (2004) Documentation for STRUCTURE software: version 2. Available from http://pritch.bsd.uchicago.edu. .
Purvis A. & Garland T.J. (1993) Polytomies in comparative analyses of continuous characters. Systematic Biology, 42, 569.
R Core Team (2013) R: A language and environment for statistical computing. .
Raxworthy C.J., Forstner M.R.J., & Nussbaum R.A. (2002) Chameleon radiation by oceanic dispersal. Nature, 415, 784–787.
Raymond M. & Rousset F. (1995) GENEPOP (version 1.2): population genetics software for exact tests and ecuminism. Heredity, 86, 248–249.
Reed D.H. & Frankham R. (2003) Correlation between Fitness and Genetic Diversity. Conservation Biology, 17, 230–237.
Rice W.R. (1989) Analyzing tables of statistical tests. Evolution, 43, 223–225.
Riggs L.A. (1990) Conserving genetic resources on-site in forest ecosystems. Forest Ecology and Management, 35, 45–68.
Robin V. V, Sinha A., & Ramakrishnan U. (2010) Ancient geographical gaps and paleo-climate shape the phylogeography of an endemic bird in the sky islands of southern India. PLoS ONE, 5, 1–13.
153
Rodrigues A.S.L. & Gaston K.J. (2002) Maximising phylogenetic diversity in the selection of networks of conservation areas. Biological Conservation, 105, 103–111.
Rosenberg N.A. (2004) Distruct: a program for the graphical display of population structure. Molecular Ecology Notes, 4, 137–138.
Rosenberg N.A., Burke T., Elo K., Feldman M.W., Freidlin P.J., Groenen M.A., Hillel J., Mäki-Tanila A., Tixier-Boichard M., Vignal A., Wimmers K., & Weigend S. (2001) Empirical evaluation of genetic clustering methods using multilocus genotypes from 20 chicken breeds. Genetics, 159, 699–713.
Rosenzweig M.L. (2002) Species diversity in space and time. Cambridge University Press, Cambridge, U.K.
Ryder O.A. (1986) Species conservation and systematics: the dilemma of subspecies. Trends in Ecology & Evolution, 1, 9–10.
Sakaguchi S., Qiu Y.-X., Liu Y.-H., Qi X.-S., Kim S.-H., Han J., Takeuchi Y., Worth J.R.P., Yamasaki M., Sakurai S., & Isagi Y. (2012) Climate oscillation during the Quaternary associated with landscape heterogeneity promoted allopatric lineage divergence of a temperate tree Kalopanax septemlobus (Araliaceae) in East Asia. Molecular ecology, 21, 3823–38.
Salerno P.E., Santiago R.R., Senaris J.C., Rojas-Runjaic F.J.M., Noonan B.P., & Cannatella D.C. (2012) Ancient Tepui summits harbor young rather than old lineages of endemic frogs. Evolution, 66, 3000–3013.
Sauquet H., Weston P.H., Anderson C.L., Barker N.P., Cantrill D.J., Mast A.R., & Savolainen V. (2009) Contrasted patterns of hyperdiversification in Mediterranean hotspots. Proceedings of the National Academy of Sciences of the United States of America, 106, 221–5.
Schoen D.J. & Brown A.H.D. (1993) Conservation of allelic richness in wild crop relatives is aided by assessment of genetic markers. Proceedings of the National Academy of Sciences of the United States of America, 90, 10623–10627.
Schut A.G.T., Wardell-Johnson G.W., Yates C.J., Keppel G., Baran I., Franklin S.E., Hopper S.D., Van Niel K.P., Mucina L., & Byrne M. (2014) Rapid characterisation of vegetation structure to predict refugia and climate change impacts across a global biodiversity hotspot. PLoS ONE, 9, e82778.
Shaffer M.L. (1981) Minimum population sizes for species conservation. BioScience, 31, 131–134.
Shaw J., Lickey E.B., Schilling E.E., & Small R.L. (2007) Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: the tortoise and the hare III. American Journal of Botany, 94, 275–288.
Simmons M.P. & Ochoterena H. (2000) Gaps as characters in sequence-based phylogenetic analyses. Systematic Biology, 49, 369–381.
Simon C., Frati F., Beckenbach A., Crespi B.J., Liu H., & Flook P. (1994) Evolution, weighting and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved
154
polymerase chain reaction primers. Annals of the Entomological Society of America, 87, 652–701.
Slatkin M. (1985) Rare alleles as indicators of gene flow. Evolution, 39, 53–65.
Slatkin M. & Barton N.H. (1989) A comparison of three indirect methods for estimating average levels of gene flow. Evolution, 43, 1349–1368.
Smith C. & Farrell B. (2005) Phylogeography of the longhorn cactus beetle Moneilema appressum LeConte (Coleoptera : Cerambycidae): was the differentiation of the Madrean sky islands driven by Pleistocene climate changes? Molecular Ecology, 14, 3049–3065.
Smith M. (2013) Threatened and Priority flora list for Western Australia. Department of Environment and Conservation, Kensington, Western Australia.
Solis A., Jugo B.M., Mériaux J.C., Iriondo M., Mazón L.I., Aguirre A.I., Vicario A., & Estomba A. (2005) Genetic diversity within and among four South European native horse breeds based on microsatellite DNA analysis: implications for conservation. The Journal of Heredity, 96, 670–8.
Sork V.L., Nason J.D., Campbell D.R., & Fernandez J.F. (1999) Landscape approaches to historical and contemporary gene flow in plants. Trends in Ecology & Evolution, 14, 219–224.
Soulé M.E. (1980) Thresholds for survival: maintaining fitness and evolutionary potential. Conservation Biology: An evolutionary-ecological perspective (ed. by M.E. Soule and B.A. Wilcox), pp. 345. Sinauer, Sunderland, Massachussets.
Soulé M.E. & Mills L.S. (1992) Conservation genetics and conservation biology: a troubled marriage. Species and Ecosystem Conservation (ed. by O.T. Sandlund, K. Hindar, and A.H.D. Brown), pp. 55–65. Scandinavian University Press, Oslo.
Southerton S.G., Birt P., Porter J., & Ford H.A. (2004) Review of gene movement by bats and birds and its potential significance for eucalypt plantation forestry. Australian Forestry, 67, 44–53.
Storfer A., Murphy M.A., Evans J.S., Goldberg C.S., Robinson S., Spear S.F., Dezzani R., Delmelle E., Vierling L., & Waits L.P. (2007) Putting the “landscape” in landscape genetics. Heredity, 98, 128–142.
Swenson N.G. (2009) Phylogenetic resolution and quantifying the phylogenetic diversity and dispersion of communities. PLoS ONE, 4, e4390.
Taberlet P., Gielly L., Pautou G., & Bouvet J. (1991) Universal primers for amplification of three non-coding regions of choloroplast DNA. Plant Molecular Biology, 17, 1105–1109.
Tajima F. (1989) Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics, 123, 585–595.
Tapio M., Tapio I., Grislis Z., Holm L.-E., Jeppsson S., Kantanen J., Miceikiene I., Olsaker I., Viinalass H., & Eythorsdottir E. (2005) Native breeds demonstrate high contributions to the molecular variation in northern European sheep. Molecular ecology, 14, 3951–63.
155
Tapper S.-L., Byrne M., Yates C.J., Keppel G., Hopper S.D., Van Niel K., Schut A.G.T., Mucina L., & Wardell-Johnson G.W. (2014a) Isolated with persistence or dynamically connected? Genetic patterns in a common granite outcrop endemic. Diversity & Distributions, DOI:10.1111/ddi.12185.
Tapper S.-L., Byrne M., Yates C.J., Keppel G., Hopper S.D., Van Niel K., Schut A.G.T., Mucina L., & Wardell-Johnson G.W. (2014b) Prolonged isolation and persistence of a common endemic on granite outcrops in both mesic and semi-arid environments. Journal of Biogeography, DOI:10.1111/jbi.12343.
Testolin R. & Cipriani G. (1997) Paternal inheritance of chloroplast DNA and maternal inheritance of mitochondrial DNA in the genus Actinidia. Theoretical and Applied Genetics, 94, 897–903.
Turner M.E., Stephens J.C., & Anderson W.W. (1982) Homozygosity and patch structure in plant populations as a result of nearest neighbour pollination. Proceedings of the National Academy of Sciences of the United States of America, 79, 203–207.
Ueno S., Setsuko S., Kawahara T., & Yoshimaru H. (2005) Genetic diversity and differentiation of the endangered Japanese endemic tree Magnolia stellata using nuclear and chloroplast microsatellite markers. Conservation Genetics, 6, 563–574.
Vellend M. & Geber M.A. (2005) Connections between species diversity and genetic diversity. Ecology Letters, 8, 767–781.
Weir J.T. & Schluter D. (2008) Calibrating the avian molecular clock. Molecular Ecology, 17, 2321–2328.
Wesener T., Raupach M.J., & Sierwald P. (2010) The origins of the giant pill-millipedes from Madagascar (Diplopoda: Sphaerotheriida: Arthrosphaeridae). Molecular Phylogenetics and Evolution, 57, 1184–1193.
Whittaker R.J. (1998) Island biogeography: ecology, evolution and conservation. Oxford University Press, New York.
Wilson E.O. & MacArthur R. (1967) The Theory of Island Biogeography. Princeton University Press, Princeton, NJ.
Wolfe K.H., Li W.H., & Sharp P.M. (1987) Rates of nucleotide substitution vary greatly among plant mitochondrial, chloroplast, and nuclear DNAs. Proceedings of the National Academy of Sciences of the United States of America, 84, 9054–9058.
Wright S. (1929) The evolution of dominance. The American Naturalist, 63, 556–561.
Wright S. (1943) Isolation by distance. Genetics, 28, 114–138.
Wrigley J.W. & Fagg M. (1989) Banksias, Waratahs and Grevilleas and all other plants in the Australian Proteaceae family. Collins Publishers Australia, Sydney, NSW.
Yates C., Gibson N., Pettit N.E., Dillon R., & Palmer R. (2011) The ecological relationships and demography of restricted ironstone endemic plant species: implications for conservation. Australian Journal of Botany, 59, 692–700.
156
Yoder A.D. & Nowak M.D. (2006) Has vicariance or dispersal been the predominant biogeographic force in Madagascar? Only time will tell. Annual Review of Ecology and Systematics, 37, 405–431.
Zhai S.-N., Comes H.P., Nakamura K., Yan H.-F., & Qiu Y.-X. (2012) Late Pleistocene lineage divergence among populations of Neolitsea sericea (Lauraceae) across a deep sea-barrier in the Ryukyu islands. Journal of Biogeography, 39, 1347–1360.
Zhang D. & Hewitt G.M. (1996) Nuclear integrations: challenges for mitochondrial DNA markers. Trends in Ecology & Evolution, 11, 247–251.
Zhang D. & Hewitt G.M. (2003) Nuclear DNA analyses in genetic studies of populations: practice, problems and prospects. Molecular Ecology, 12, 563–584.
Zheng H., Wyrwoll K.-H., Li Z., & McA Powell C. (1998) Onset of aridity in southern Western Australia—a preliminary palaeomagnetic appraisal. Global and Planetary Change, 18, 175–187.
Zink R.M. (1996) Comparative phylogeography in north American birds. Evolution, 50, 308–317.
Zuckerkandl E. & Pauling L. (1965) Molecules as documents of evolutionary history. Journal of Theoretical Biology, 8, 357–366.
157
APPENDIX A
Table A1 Specimen information, including spatial coordinates from UTM zone 50J and haplotype number based on combined CO1 and 16S mtDNA data. Samples from Windarling and Kooylanobbing (bold) are from specimens lodged at the Western Australian Museum, registration numbers shown (WAM_T).
Sample Population Easting Northing Haplotype no.
DH23 Die Hardy 729291 6685360 Hap21
DH2 Die Hardy 725805 6684839 Hap19
DH32 Die Hardy 727525 6687269 Hap24
DH33 Die Hardy 727525 6687269 Hap25
DH36 Die Hardy 727458 6687278 Hap26
DH37 Die Hardy 727458 6687278 Hap27
DH40 Die Hardy 727411 6687282 Hap28
DH41 Die Hardy 727347 6687287 Hap29
DH43 Die Hardy 727257 6687301 Hap24
DH46 Die Hardy 728060 6690423 Hap30
DH47 Die Hardy 728411 6690546 Hap31
DH48 Die Hardy 728411 6690546 Hap32
DH6 Die Hardy 729215 6685314 Hap20
DH7 Die Hardy 729215 6685314 Hap21
DH9 Die Hardy 729215 6685314 Hap23
HA36 Helena Aurora 760332 6639565 Hap42
HA37 Helena Aurora 760332 6639565 Hap43
HA38 Helena Aurora 760332 6639565 Hap44
HA41 Helena Aurora 760320 6639552 Hap43
HA42 Helena Aurora 760270 6639470 Hap45
HNAF11 Helena Aurora 752715 6637738 Hap39
HNAF13 Helena Aurora 754411 6633431 Hap40
HNAF18 Helena Aurora 754419 6633434 Hap40
HNAF19 Helena Aurora 754419 6633434 Hap41
HNAF24 Helena Aurora 754414 6633429 Hap40
HNAF2 Helena Aurora 752978 6637869 Hap38
HNAF3 Helena Aurora 752978 6637869 Hap39
HNAF4 Helena Aurora 752978 6637869 Hap39
HNAF5 Helena Aurora 752978 6637869 Hap39
MTJ39 Mt Jackson 718999 6650761 Hap35
MTJ41 Mt Jackson 719010 6650756 Hap35
MTJ42 Mt Jackson 718955 6650778 Hap33
MTJ43 Mt Jackson 718955 6650778 Hap36
MTJ44 Mt Jackson 718955 6650778 Hap33
MTJ52 Mt Jackson 719998 6650682 Hap35
MTJ53 Mt Jackson 720019 6650557 Hap34
MTJ54 Mt Jackson 720019 6650557 Hap37
MTJ6 Mt Jackson 718973 6650726 Hap33
158
MTJ7 Mt Jackson 718973 6650726 Hap34
WAM_T98784 Windarling 723016 6677816 Hap46
WAM_T98787 Windarling 723016 6677816 Hap47
WAM_T98791 Windarling 723016 6677816 Hap47
WAM_T98792 Windarling 723016 6677816 Hap47
WAM_T98793 Windarling 723016 6677816 Hap47
WAM_T98794 Windarling 723261 6677891 Hap47
WAM_T98796 Windarling 723261 6677891 na
WAM_T98798 Windarling 723261 6677891 Hap47
WAM_T98799 Windarling 723261 6677891 Hap47
WAM_T98800 Windarling 723261 6677891 Hap47
WAM_T98801 Windarling 723178 6677740 Hap47
WAM_T98803 Windarling 723178 6677740 Hap47
WAM_T98807 Windarling 723178 6677740 Hap48
WAM_T98810 Windarling 723178 6677740 Hap47
WAM_T98805 Koolyanobbing 748961 6581863 Hap1
WAM_T98806 Koolyanobbing 748968 6581893 Hap2
WAM_T98808 Koolyanobbing 748968 6581893 Hap3
WAM_T98814 Koolyanobbing 749410 6581553 na
WAM_T98816 Koolyanobbing 749410 6581553 Hap3
WAM_T98819 Koolyanobbing 749409 6581486 Hap4
WAM_T98821 Koolyanobbing 749409 6581486 Hap5
WAM_T98822 Koolyanobbing 749409 6581486 Hap6
WAM_T98824 Koolyanobbing 749741 6581383 Hap7
WAM_T98825 Koolyanobbing 749741 6581383 Hap8
WAM_T98831 Koolyanobbing 749884 6581470 Hap9
WAM_T98832 Koolyanobbing 749884 6581470 Hap8
WAM_T98834 Koolyanobbing 751727 6580142 Hap10
WAM_T98836 Koolyanobbing 751727 6580142 Hap11
WAM_T98838 Koolyanobbing 750267 6579600 Hap12
WAM_T98839 Koolyanobbing 750267 6579600 Hap12
WAM_T98842 Koolyanobbing 746153 6583769 Hap13
WAM_T98843 Koolyanobbing 744997 6584012 Hap14
WAM_T98844 Koolyanobbing 744997 6584012 Hap15
WAM_T98845 Koolyanobbing 744997 6584012 Hap16
WAM_T98848 Koolyanobbing 743158 6585319 Hap17
WAM_T98849 Koolyanobbing 743158 6585319 Hap18
159
Table A2 The three, best-fit models determined in IMa2 for tests of migration versus isolation between the paraphyletic populations Koolyanobbing and Helena
Aurora. Results are based on mtDNA variation.
Model log(p) terms df 2LLR AIC Delta AIC
Population size KOO equal to population size HA, migration rates zero -6.091 2 3 1.44 16.182 0
Migration rates zero -5.371 3 2 0 16.742 0.56
Population size KOO equal to population size 2 (ancestral population size), migration rates zero -6.708 2 3 2.675 17.416 1.234
160
APPENDIX B
Table B1 Specimen information, including spatial coordinates from UTM zone 50J and 51J (Mt Correll) and haplotype number based on combined trnQ-rps16 and psbD-trnT cpDNA data.
Sample Population Easting Northing Haplotype no.
corpop1 Mt Correll 683219 6619906 1
corpop2 Mt Correll 683219 6619906 1
corpop9 Mt Correll 683219 6619906 1
corpop13 Mt Correll 683243 6619901 1
corpop14 Mt Correll 683268 6619889 1
corpop18 Mt Correll 683282 6619892 1
corpop20 Mt Correll 683282 6619892 1
corpop22 Mt Correll 683282 6619892 1
corpop23 Mt Correll 683282 6619892 1
corpop25 Mt Correll 683300 6619872 1
MGs1 Die Hardy 731211 6683058 2
DHs1 Die Hardy 728545 6692270 3
DH2 Die Hardy 731078 6690882 4
dhs3 Die Hardy 727987 6687485 5
DHs4 Die Hardy 725814 6684856 6
dhs5 Die Hardy 727132 6685587 7
dhs6 Die Hardy 728074 6690048 8
dhs7 Die Hardy 731255 6692878 9
dhs9 Die Hardy 729291 6685360 10
dhs10 Die Hardy 728574 6680010 10
finpop1 Mt Finnerty 225443 6602805 11
finpop2 Mt Finnerty 225443 6602805 12
finpop3 Mt Finnerty 225443 6602805 12
finpop8 Mt Finnerty 225441 6602817 11
finpop10 Mt Finnerty 225441 6602817 11
finpop15 Mt Finnerty 225434 6602825 11
finpop20 Mt Finnerty 225420 6602843 11
finpop22 Mt Finnerty 225007 6602865 11
finpop24 Mt Finnerty 225407 6602865 11
finpop25 Mt Finnerty 225407 6602865 11
HAs2 Helena Aurora 759972 6638979 13
HAs3 Helena Aurora 756802 6638620 14
HAs4 Helena Aurora 760164 6639159 13
hapop8 Helena Aurora 759985 6638854 13
hapop14 Helena Aurora 759949 6638837 13
hapop17 Helena Aurora 759928 6638876 13
hapop22 Helena Aurora 759905 6638954 13
161
hapop25 Helena Aurora 759947 6638980 13
HUs1 Hunt Range 775758 6655357 11
hupop1 Hunt Range 775718 6655385 11
HUs2 Hunt Range 778002 6645231 11
hupop2 Hunt Range 775718 6655385 11
hupop4 Hunt Range 775724 6655349 11
hupop6 Hunt Range 775586 6655067 11
hupop7 Hunt Range 775585 6655791 12
hupop8 Hunt Range 775585 6655091 11
hupop13 Hunt Range 775580 6655114 11
hupop16 Hunt Range 775751 6655273 15
mmpop1 Mt Manning 751789 6699172 16
MM2 Mt Manning 753012 6678501 17
mms3 Mt Manning 753299 6678801 17
mms5 Mt Manning 753299 6678801 16
mms6 Mt Manning 754957 6678737 16
MMs7 Mt Manning 754957 6678737 18
MMs8 Mt Manning 754731 6680426 19
MMs9 Mt Manning 754526 6679481 17
mmpop17 Mt Manning 751811 6699176 20
ths1 Taipan Hill 779913 6634137 12
thpop1 Taipan Hill 779424 6634494 12
TH2 Taipan Hill 779834 6634123 12
ths3 Taipan Hill 779566 6634455 11
thpop3 Taipan Hill 779424 6634494 11
ths5 Taipan Hill 779278 6634562 11
ths6 Taipan Hill 779144 6634628 11
ths8 Taipan Hill 778490 6634412 11
ths9 Taipan Hill 778520 6634997 11
ths10 Taipan Hill 778618 6635295 11
162
APPENDIX C
Table C1 Diversity statistics for the eight Banksia arborea BIF populations for chloroplast DNA data: N, number of individuals; # Haps, number of haplotypes; HD, haplotype diversity; Pi, nucleotide diversity.
Pop N # Haps HD PI (%) Genbank
petB psbA psbD
DH 10 2 0.356 0.019 KF668088 KF668091 KF668094 KF668095
HA 9 1 0 0 KF668089 KF668091 KF668094 HUN 10 1 0 0 KF668089 KF668091 KF668094 KOO 10 1 0 0 KF668090 KF668092 KF668095 MM 10 1 0 0 KF668088 KF668091 KF668094 MTJ 10 1 0 0 KF668089 KF668091 KF668094 TH 10 1 0 0 KF668090 KF668093 KF668095 WIN 10 1 0 0 KF668088 KF668091 KF668094
Figure C1 Mantel correlograms of pairwise genetic distance (Nei, 1978) against log of geographic distance
for all individuals from Die Hardy north, central and south sub-populations.
Appendix D removed due to copyright issues.