New approaches to cataloguing and understanding ...biology-assets.anu.edu.au/hosted_sites/Scott/2014oliveretalajz.pdf · New approaches to cataloguing and understanding evolutionary

New approaches to cataloguing and understanding evolutionarydiversity a perspective from Australian herpetology

Paul OliverAB J Scott KeoghA and Craig MoritzA

AEvolution Ecology and Genetics Research School of Biology The Australian National UniversityCanberra ACT 0200 Australia

BCorresponding author Email pauloliveranueduau

Abstract Species are a fundamental unit for all fields of biology but conceptual and practical limitations have hamperedthe process of identifying and describing species in many organismal groups One outcome of these challenges is theaccumulation of genetically divergent lineages and morphologically distinctive populations that are lsquoknownrsquo but remainof uncertain taxonomic status and evolutionary significance These lineages are also currently not effectively incorporatedinto evolutionary studies or conservation planning andmanagement Here we suggest three ways to address this issue Firstthere is a need to develop improved frameworks to systematically capture taxonomically unrecognised lineage diversitySecond increased utilisation of metadata frameworks will allow better recording and dissemination of biodiversityinformation Finally emerging genomic and analytical techniques will provide powerful new tools to improve ouridentification and understanding of evolutionary lineages

Additional keywords biodiversity candidate species conservation cryptic species divergent lineage generalisedlineage concept genomics species delimitation taxonomy

Received 24 October 2014 accepted 14 January 2015 published online 19 February 2015

Introduction

Species are a fundamental unit of biology (Hull 1977)However species boundaries are also testable hypothesesTests of these hypotheses need to be set in context of the factsthat species formation occurs through diverse and interactingprocesses manifest in differing outcomes for phenotypic andgenetic divergence and it is often a protracted process (deQueiroz 1998 Fujita et al 2012) Conceptual advances andnew empirical data thus result in the continual refinement ofspecies limits and taxonomic arrangements More so thanmany other fields of biology the resulting changes haveimplications to a wide range of end users with highly variablelevels of interest in the systematic process itself On the onehand this creates a need to explain the rationale behindtaxonomic decisions to the wider community using thisframework On the other hand systematists need to continuallyevaluate their processes to ensure that they are effectivelycapturing evolutionary diversity and maximising rigor androbustness of species delineation for end users

In this paper we have three specific aims First we highlighta particular issue in systematics ndash the ongoing proliferationof genetically divergent lineages or lsquocandidate speciesrsquo in theliterature that remain almost invisible to other scientific workincluding conservation assessments management planningand evolutionary analyses This issue is especially notable forlow-dispersal and morphologically conservative taxa such asmany clades of reptiles on which we focus here Second we

attempt to provide a brief introduction to the philosophy andpractice of modern taxonomy ndash in part an explanation of whytaxonomists continue to make changes Finally we discussthree emerging approaches ndash a standardised provisionalnomenclature metadata frameworks and genomic techniques ndashthat may provide opportunities to more effectively documentunderstand and manage biodiversity in taxonomicallychallenging groups

Known unknowns the proliferation of unrecognisedlineage diversity

New species continue to be recognised at a high rate even incomparatively well known vertebrate groups For instance over100 new reptile species were recognised in 2013 (Uetz andHosek 2014) Some new species recognised primarily on thebasis of highly divergent phenotype cannot be confounded withany other recognised taxa (eg Hoskin and Couper 2013)However at least amongst vertebrates an increasing majorityof new species are part of lsquospecies complexesrsquo ndash groups of taxathat are difficult to elucidate and diagnose without detailedmultidimensional analysis (eg Donnellan et al 1993 Bickfordet al 2007)

Molecular genetic tools especially mitochondrial DNAsequences allow us to identify potentially divergent lineageswithin species complexes quickly and efficiently Howeverthis is not without error (for example due to mtDNAintrogression across species boundaries or incomplete lineage

Journal compilation CSIRO 2014 wwwpublishcsiroaujournalsajz

CSIRO PUBLISHING

Australian Journal of Zoology 2014 62 417ndash430 Reviewhttpdxdoiorg101071ZO14091

sorting) such that additional independent data sources arerequired to test the hypothesis that these lineages representdistinctive evolutionary lineages or species (see below) Insome cases morphological characters that allow diagnosis inthe field are then identified (eg Shea et al 2011 Oliver et al2014a) However other species appear to be genuinely crypticand show little to no evidence of consistent morphologicaldifferentiation (Hoskin 2007 Smith et al 2011 Pepper et al2011a) In both cases assembling appropriate datasets toresolve species boundaries in species complexes is timeconsuming dependent on appropriate sampling and oftenexpensive

This comparative ease of initial genetic screening on theone hand juxtaposed against the challenges of speciesdelineation on the other has resulted in a proliferation ofcandidate species and genetic lineages that are not capturedby current taxonomy (Fouquet et al 2007 Vieites et al2009 Oliver et al 2010) This evolutionary diversity thusalso remains almost entirely missing from conservationassessments management planning and evolutionary analysesFinding new ways to effectively incorporate this unnameddiversity into biological research in the short term and

ultimately resolving the taxonomy of these lineages in thelonger term are two of the biggest challenges facingsystematists in the coming decade

Australian reptiles as a case study

Australia has an exceptionally diverse reptile fauna with over950 recognised species (Wilson and Swan 2013) This state ofknowledge is the result of decades of mostly painstaking peer-reviewed lsquoalpharsquo taxonomy based largely on morphologicalcharacters This creates the perception that the job is nearly doneexcept for ongoing discoveries in remote andor highly endemicregions (Chapman 2009 Hoskin and Couper 2013) By contrastgenetic studies continue to reveal additional highly divergentlineages within species as currently recognised (Pepper et al2006 2011b Fujita et al 2010 Oliver et al 2010 2014b Marinet al 2013) These data indicate that significant evolutionarydiversity remains uncaptured by taxonomy (see Box 1)Conversely there are also some instances where molecularanalyses have shown that taxonomy based on highly plasticcharacters inflates estimates of species diversity (Keogh et al2005 Rabosky et al 2014a) (Fig 1) These deficiencies in our

Box 1 Australiarsquos increasing reptile diversity geckos

(a) Diplodactylus conspicillatus complex

(b) Heteronotia binoei complex

(c) Oedura marmorata complex

9+ species or lineages Oliver et al 2014a

Fig II Hyperdiverse species complexes

10 Candidate species Fujita et al 2010

10 major lineages Oliver et al 2014b

Geckos (including pygopods) are the secondmost diverse groupof Australia reptiles and comprise close to 200 species Howevernew species continue to be recognised at a high rate ~20 ofrecognised species have been described since 2000 (Macdonald2014) (Fig I) While some new taxa are highly distinct anddivergent novelties most have been partitioned from previouslyrecognised species such asCyrtodactylus tuberculatus sl (Sheaet al 2011) and Rhynoedura ornata sl (Pepper et al 2011a)However even with this recent growth in recognised diversitypublished data indicate that species diversity remainsunderestimated (contra Chapman 2009) Of particular noteare hyperdiverse species complexes that comprise up to 10 (oreven more) highly divergent lineages (Fig I) Many of theselineages are corroborated by multiple independent datasets andclearly representundescribedspeciesOngoingresearch indicatesthat similarly diverse species complexes exist in genera such asAmolosia Gehyra and Heteronotia Furthermore regardlessof the final taxonomic status of these lineages it is also clear thatour current taxonomy does not fully capture the evolutionarydiversity of Australian geckos especially in the MonsoonTropics

200

Fig I Australian gecko speciesdiversity through time

150

100

Num

ber

of s

peci

es

50

0

1800 1850 1900

Year1950 2000

Data and graph courtesy of S Macdonald (2014)

418 Australian Journal of Zoology P Oliver et al

systematic framework can conflate biodiversity mappingimpede or confound monitoring of landscape-scale changes andpotentially even result in the misappropriation of conservationresources (Siler et al 2014)

Recent Australian reptile descriptions include localisedspecies of conservation concern from often highly disturbedhabitats (Kay and Keogh 2012 Doughty and Oliver 2013McLean et al 2013 Melville et al 2014) (Fig 2) Shrinkingislands of high-altitude habitat across eastern Australia alsocontinue to reveal previously unrecognised endemic lineagesand taxa (Bell et al 2010 Hoskin and Couper 2014 Haineset al 2014) Finally and perhaps most strikingly recent workacross the Australian Monsoonal Tropics has revealed anexceptional number of apparently restricted lineages that stillawait proper taxonomic assessment (Fujita et al 2010Pepper et al 2011b Smith et al 2011 Oliver et al 20122014b Marin et al 2013 Rabosky et al 2014a) Manymammal taxa have recently disappeared from this region(Woinarski et al 2011) and this undocumented reptile biotamight also include species that are similarly vulnerable tolandscape change

Landscape-scale and community-level approaches tomodelling biodiversity also present an exciting new opportunityto improve predictions of the consequences of climate change andallocate management resources (Reside et al 2013) howeverthey tend to use current species-level taxonomies and associateddistributional data The ongoing discovery of lineage diversityindicates that current taxonomy may not adequately captureendemism across many major regions of Australia especially innorthern Australia (Fujita et al 2010 Pepper et al 2011b 2013Smith et al 2011Oliver et al 2012 2014b)where ongoingworkis re-emphasising known regions of endemism such as Cape

Range and Cape Melville (Doughty et al 2008 Hoskin andCouper 2014) and highlighting overlooked areas such as the MtIsa Inlier and the limestone ranges of the southern Kimberley(Fujita et al 2010 Oliver et al 2014c)

With few exceptions comparative macroevolutionary andmacroecological analyses typically use species names aslsquoprovidedrsquo by taxonomists however inadequate taxonomy canalso impede autecological and evolutionary studies Because oftheir high diversity and endemism Australian reptiles havefeatured prominently in evolutionary biology they include focalgroups for investigation of topics ranging from adaptivediversification (Rabosky et al 2007 Gruumlndler and Rabosky2014) community ecology and assembly (Pianka 1986 Powneyet al 2010) thermal ecology (Grigg and Buckley 2014)communication and sexual selection (Chen et al 2013) sociality(Chapple 2003) and parthenogenesis (Kearney and Shine 2004)An inaccurate phylogenetic and systematic framework canconfound such studies (Donnellan et al 1993 Bickford et al2007) Furthermore the same processes that make speciescomplexes difficult to resolve ndash phenotypic conservatismconvergence and plasticity occasional gene-flow betweenlineages and adaptive and non-adaptive diversification ndash alsoplay important roles in shaping biological diversity Detailedexamination of complexes of related taxa thus tends to provideframeworks for broader insight into how these processes shapeevolutionary diversification (eg Smith et al 2011 Singhal andMoritz 2014 Rabosky et al 2014a)

Species in concept and in practice the general lineageconcept and integrative taxonomy

Before addressing how best to deal with undescribedlineage diversity it is necessary to take a brief look at the

(a) (b) (c) (e) (f ) (g)(d)

Fig 1 Overestimation of species diversity in lineages with highly plastic morphology Figured specimens have been ascribed to four different specieson the basis of variation in dorsal colour pattern but in fact all represent just one species ndash Ctenotus helenae Reproduced with permission from Raboskyet al (2014a)

Cataloguing and understanding evolutionary diversity Australian Journal of Zoology 419

complex issue of how scientists recognise species Whilethere has been exhaustive discussion of this topic over manydecades there is now perhaps an increasing consensus onhow we should approach the delimitation of species at leastfor sexually reproducing vertebrates At the conceptual levelwe advocate the General Lineage Concept of species (deQueiroz 1998 Fujita et al 2012) This is deliberately broad butgenerally specifies that species are separately evolving lineages ndashdiscrete cohesive groups of organisms (metapopulations)that are on independent evolutionary trajectories from all othersuch groups (de Queiroz 1998) To a significant extent andsubsuming ideas such as the Biological and EvolutionarySpecies Concepts this framework circumvents thearguments that dominated the field for many years (de Queiroz2007)

In practice there is also widespread acknowledgementthat delineating species requires accumulating independentdatasets of genotypic morphological or ecological data to helpidentify these evolutionarily independent lineages (ie species)(Dayrat 2005 Fujita et al 2012) Here we follow the recentfashion to term this lsquointegrative taxonomyrsquo (Dayrat 2005Padial et al 2010 but see Yeates et al 2011) Leaving

aside the issue of terminology under this methodologicalframework no one source of evidence (morphology genesecology) has priority or is explicitly necessary but the largerthe number of independent data sources that support a distinctevolutionary trajectory the stronger the case for taxonomicrecognition (Miralles and Vences 2013)

In practice this approach is simply a formalisation of thelsquoreciprocal illuminationrsquo ethos and methodologies that manysystematists have followed for many decades (Sites andMarshall 2004) Indeed Australian herpetology since the1970s provides numerous examples of the incorporation ofmultiple sources of data into delimiting species boundaries indifficult groups including morphology chromosomesallozymes mtDNA and increasingly nuclear sequence data(King 1982 Hutchinson and Donnellan 1992 Donnellan et al2002 Oliver et al 2007 Sistrom et al 2013 Pepper et al2013) In many cases these workers specifically stated that theentities they were identifying were congruent across multipledifferent species delineation methodologies (Hutchinson andDonnellan 1992 Oliver et al 2007) in short they werebroadly following what are now frequently termed lsquointegratedrsquotaxonomic protocols

Fig 2 Australia reptileswith highly restricted ranges recently partitioned from species complexes Clockwise from top left the coastal plain ctenotus (Ctenotusora) and the cloudy stone gecko (Diplodactylus nebulosus) from the south-west sandplains the black pilbara gecko (Heteronotia atra) from two mesas in thePilbara region and the Barrier Ranges dragon (Ctenophorous mirritiyana) from the Barrier Range in far western New South Wales Photographs by BradMaryan and Steve Sass


Over the past few years a plethora of methods to enable sequencing of a large number of nuclear genes has emerged (Lemmon and Lemmon2013) (Fig III)

Whole genome approaches At one end of the spectrum is whole genome resequencing For some questions (eg mapping regions ofgenome divergence (eg Kawakami et al 2014) this is the ideal yet for taxa with modest size (gt1GB) genomes and which do not have ahigh-quality reference genome this approach remains cost-prohibitive for phylogeographic and species boundary scale questions Henceseveral methods have been developed to enable sequencing at subgenomic scale

Genome reduction ndash RADseq RADseq and its variants samples the genome at random using size-selected fragments generated by oneor more restriction enzymes (eg Peterson et al 2012) So long as a high proportion of restriction sites are conserved the method cangenerate sequence data for 1000s of loci though there is often considerable missing data and the loci recovered might include repetitivesequences as well as the unique orthologous genes This approach has proven effective at population to phylogeographic scales includingspecies delimitation (eg Leacheacute et al 2014) However at deeper divergences mutations at restriction sites erode the number oforthologous loci that can be captured

Genome reduction ndash Target capture Fragments from randomly sheared DNA containing the target sequences can be enriched viahybridisation against complementary probes (short regions of DNA used to target complementary regions) There are several strategies togenerate these

Approach 1 Where 10s to 100 loci will suffice probes can be generated from the target taxon by PCR amplifying the genes from one ortwo individuals (Pentildealba et al 2014 eg including loci previously studied via Sanger sequencing)

Approach 2 Synthetic probes can be designed against exon sequences for the clade in question which themselves were derived from denovo transciptome assembly (Bi et al 2012) This lsquocustom exon-capturersquo method has proven highly effective at reasonablephylogenetic distances such as across skinks of the Eugongylus group (Bragg et al unpubl data) and has the advantage ofproducing high-quality near-complete data matrices that can be readily connected across scales from populations to entireclades There is however a higher start-up cost and the approach is best suited where there is intent to focus on one cladeaddressing questions from population genetics through to macroevolution

Approach 3 An alternative to clade-specific exon capture is to use more generic probes ndash conserved low-copy loci identified bycomparing phylogenetically diverse genomes This includes two methods that are being used with increasing frequencyultraconserved elements (Faircloth et al 2012) and conserved exons (Lemmon and Lemmon 2012) Although the targetsthemselves are highly conserved the flanking regions also recovered to some extent by hybridisation have higher levelsof variation As a consequence these more generic approaches which typically yield high-quality data across 100s of lociare informative at the phylogeographic scale (Smith et al 2014 Brandley et al unpubl data) and so have promise forspecies delimitation as well as phylogenetic analyses

Box 2 The expanding toolbox for accessing genome-scale data

Genome (1ndash2 Gb)

Genome reduction

Random sample(eg RAD)1

1 Peterson et al (2012)2 Pentildealba et al (2014)

3 Lemmon and Lemmon (2012)

4 Faircloth et al (2012)

5 Bi et al (2012) Bragg et al in review

SCPP (PCRprobes)2

Loci gt1000 10ndash100 300ndash500 100sndash1000 gt1000

Completeness

Connectivity

Investment

low

low-mod

low

high

high

medium medium medium

high

high

high

high

high

moderate

moderate

Anchoredenrichment(conservedexons)3

Ultraconservedelements (noncoding)4

Custom exoncapture5

Whole genome sequencing(at low coverage)

Target enrichment

Fig III Schematic summary of some emerging methods for generating genomic and subgenomic datasets


Evolutionary distinctiveness ndash necessary and sufficientevidence

The shift away from species concept semantics is most welcomeHowever deciding what levels or patterns of variation providesufficient evidence of evolutionary distinctiveness still remainsan important challenge (Bond and Stockman 2008 Miralles andVences 2013) As increasingly large-scale genomic datasetsemerge (see Box 2) practical limitations on the resolution ofgenetic data for resolving species boundarieswill be a thing of thepast However the continuous and sometimes meanderingprocess of species formation (Rosenblum et al 2012) means thatthere is no simple threshold of genetic divergence for recognisingspecies Thus even with extensive genetic data and powerfulanalytical methods the question of howmuch genetic divergenceis enough will remain It follows that additional evidence ndash egfrom morphology behaviour or tests for introgression acrosscontact zone remains highly desirable

Conversely while many species can be diagnosed by thecharacters that systematists have relied on in the past (especiallydiscrete morphological traits) there is no intrinsic reason whythis has to be the case To use morphology as an example in thelong history of thought and writing in relation to lsquosibling speciesrsquo(a precursor idea to cryptic species) it was emphasised that (1)morphologically indistinguishable species are just as real asmorphospecies and that (2) more such cases can be expected aslsquobiochemicalrsquo methods are used more extensively (Mayr 1970Dobzhansky 1970 Grant 1981) As predicted new multilocusdatasets are increasingly providing compelling evidence ofreduced gene flow across hybrid zones of lineages that are nearindistinguishable usingmorphology (Hoskin et al 2005 Singhaland Moritz 2014) While they are effectively indistinguishablelsquoin the fieldrsquo these more divergent lineages are clearly onseparate evolutionary trajectories and a compelling argumentcan be made that they should be recognised as species (Hoskin2007)

To make matters yet more complicated with the increasingapplication of multilocus and now genome-scale data it is clearthat what appear to be very well defined species boundariescan sometimes be porous ndash some genetically ecologically andmorphologically distinct species do nonetheless exchangegenes (Mallet 2005 Pinho and Hey 2010) Such introgressionhas been detected via multilocus studies in several Australianreptiles (Rabosky et al 2009 Haines et al 2014) and amphibians(Catullo and Keogh 2014) and has even occurred in humanevolution where regional populations of modern Homo sapiensshow evidence of introgression fromother differentiated hominidlineages ndash the Denisovians and Neanderthals (Reich et al 2011Yang et al 2012) This poses important questions about howfrequent gene flow is between lineages that otherwise showstrong evidence of evolutionary divergence (ie species) Itseems likely that only more comprehensive genome-scaledatasets (Box 2 and below) will allow us to better address thesequestions about what species boundaries represent and how theyfunction

Some directions for the next decade

The Australian herpetofauna includes a large number ofspecies with divergent lineages many of which have been

identified initially by mtDNA sequencing that remain ofuncertain taxonomic and evolutionary significance (Fujitaet al 2010 Oliver et al 2010 2014b Marin et al 2013) Atthe same time Australia is well served by a pre-existingtaxonomy and an extensive relevant infrastructure includingmajor museum collections several university- and museum-based research laboratories and the online Atlas of LivingAustralia This provides excellent opportunities to developstrategies to improve our capture and understanding ofunrecognised lineage diversity How Australian biologistsaddress these challenges may have important implications forscientists in other less developed megadiverse countries thatare also trying to deal with the challenges of documenting andunderstanding their biota (Mittermeier et al 1997)

Here we suggest three keys areas that could be targetedto more effectively incorporate biodiversity information intoconservation and biology and improve our understanding ofspecies complexes(1) Systematic adoption of strategies to facilitate the widespread

inclusion of significant undescribed evolutionary diversityinto biodiversity analyses

(2) Increasing utilisation of metadata frameworks to bettercapture genetic and morphological data to increase theefficiency of systematics research and improve linksbetween this and other areas of biology

(3) Increasing adoption of genomic technologies to refineour understanding of what species are and to improvedelineation of species boundaries in difficult groups

Formalising processes for the lsquorecognitionrsquoof lsquounrecognisedrsquo diversity

There do not appear to be any short cuts to delineating taxa inspecies complexes Based on current trends generating andprocessing the genomic and morphological data required toresolve the more difficult Australian reptile groups alone willprobably take the best part of the coming decade (or more)Divergent lineages are often identified many years beforethey are described (Oliver et al 2009 2014a) and wellsupported candidate species in many other groups remainunnamed (eg Oliver et al 2009 2010 Fujita et al 2010Rabosky et al 2014a) Furthermore even if some divergentlineages do not ultimately warrant recognition as speciesthey often still capture evolutionary diversity and reflectimportant landscape-level processes that generate shape andmaintain biodiversity (Moritz et al 2009 Carnaval et al2014) and a strong argument can be made for their inclusioninto biodiversity assessments and mapping (Moritz 2002Rissler et al 2006)

Against this background a compelling argument can be toformalise a process that will expedite the widespreadrecognition of candidate species and their component lineagesand facilitate their inclusion into biological analyses (Schindeland Miller 2010 Oliver and Lee 2010) This would improvethe accuracy of biodiversity data used in diversity modellingand aid allocation of conservation resources Of particular notehere is the emerging lineage diversity of reptiles from themonsoon tropics of northern Australia (Fujita et al 2010Smith et al 2011 Oliver et al 2012 2014b) This is where


lineage diversity information (independent of taxonomy) mayfeed in most usefully ndash for instance testing whether highdiversity is correlated with particular combinations of climaticand topographic characters and using this information toeffectively assign conservation resources and monitorlandscape-scale changes (Rissler et al 2006)

The Evolutionarily Significant Unit (ESU) concept wasproposed nearly two decades ago partly in response to theissues outlined above (Ryder 1986 Moritz 1994) In essencethe goal was to formalise the recognition of majorindependently evolving lineages within taxonomic species(ESUs) and of demographically independent units withinthese Yet there is debate around how to define ESUs andin most areas of the world (including Australia) this concepthas rarely been applied (beyond threatened speciesmanagement) to enable including taxonomically unrecognisedlineage diversity in regional conservation assessments andplanning Furthermore in light of ongoing conceptual andmethodological advances an argument can be made that manyof these independently evolving lineages ESUs sensu Moritz1994 should now be recognised as species

Herpetologists working on diverse faunas elsewhere havedeveloped alternative protocols for lsquonamingrsquo unrecogniseddiversity Perhaps the most relevant model is that proposed byVieites et al (2009) in their comprehensive assessment of theMadagascan frog fauna This centred around three categories(1) confirmed candidate species (CCS) ndash strong evidencefor evolutionary independence from multiple datasources(2) unconfirmed candidate species (UCS) ndash geneticdivergence other datasources as yet not assessed and(3) deeply divergent lineages (DDL) ndash genetic divergencebut no support from other datasources We believe this servesas a good starting point for the Australian fauna However wepropose a modification of this system that involves just twocategories and also attempts to capture the theme and purposeof ESUs(1) Candidate species (CS) ndash those lineages for which at

least two different independent datasets relevant toevolutionary divergence (mtDNA nDNA different aspectsof phenotype reduced gene flow at contact zones) suggest asignificant history of evolutionary distinctiveness

(2) Evolutionary Significant Unit (ESU) ndash those lineagesin which divergence levels (morphologicalmolecular)approach those seen between related species but are notjudged sufficient to warrant recognition as species Thisagain requires more than one data source to provide robustevidence (eg mtDNA+ nDNA divergence for ESUs sensuMoritz 1994)These categories are deliberately broad and generalised

We present a worked example of their application in Fig 3We propose just two categories (contra Vieites et al 2009) forseveral reasons The CS category is designed to be a proxy forcapturing species diversity as accurately as possible pendingformal descriptions while the ESU category is designed tocapture significant evolutionary diversity within recognised orcandidate species Thus they are somewhat complementaryand also hierarchical ndash for instance one CS may comprisemultiple ESUs Trying to apply a third category (DDL sensuVieites et al 2009) based on genetic lineage divergence

(especially mtDNA data) alone creates a danger of recognisinglineages that are of ephemeral significance or do not representevolutionary boundaries and including these in biodiversityanalyses There are also practical and logistical benefits toa ranking system that comprises just two categories (seebelow)

Different sometimes non-overlapping datasets are oftenused to delimit evolutionary diversity in different groupsTherefore it is essential that the evidence used to assess lineagediversity is explicitly stated and freely available (ie sourcesof data levels of genetic divergence and thoroughnessof sampling) This background information provides theframework needed for ongoing assessment of the accuracyand comparability of provisional taxonomies

Expanding metadata frameworks to expedite the captureof biodiversity information

Online databases of biological and climatic data have made amajor contribution to expediting systematic and evolutionaryresearch For example online repositories of genetic sequencedata (eg GenBank) now allow synthesis and leverage of millionsof dollars of public investment into generating genetic data (egPyron et al 2013) Similarly online databases of museummaterial with matching coordinates are also now available formany areas (including Australia) and are making a majorcontribution to facilitating biodiversity research and conservation(Graham et al 2004 Soberoacuten and Peterson 2004) It is timely tostart considering additional ways that metadata frameworks mayimprove the capture of lineage diversity and taxonomicinformation

In Australia we have strong infrastructure in the Atlas ofLiving Australia which itself builds on previous federatedmuseumdatabases (OZCAMAVH) and the substantial efforts ofcurators in digitising and validating the specimen records Thisconsiderable investment of time and resources now provides aframework for accessing a vast number of museum records(approaching half a million for reptiles and amphibians) andlinking thesewith a diverse array of environmental variables suchas climate and geology Future options for incorporatingphylogenetic hypotheses and other data sources into this networkare now also being explored (eg Jolley-Rogers et al 2014)Efforts to lsquophylogenisersquo the Atlas are especially exciting as wewill then be able to properly represent and interrogate the tree oflife above and below recognised species as it exists acrossAustralia There are however abundant further opportunities tobuild extra data interfaces into this framework Here we focus onjust two areas that are often reported upon in systematic literaturebut aremissing fromcurrent databases (1) candidate species (CS)and ESUs (as discussed above) and (2) phenotypic data

The argument for recognising candidate taxa and significantlineage diversity is outlined above and elsewhere (Schindel andMiller 2010) The Atlas of Living Australia provides an excellentframework for attempting to do this in a systematic manner forlarge portions of the biota To enable collaborative researchand ultimately share the outcomes it will be possible to linkuser-provided records (pertaining to tissues andor voucheredspecimens) to phylogeny from intraspecific ESUs and CSs torecognised species and entire endemic clades When research


results are robust and ready to be shared publically there couldbe a link under the relevant species page(s) highlighting theexistence distribution and samples used to identify bothadditional candidate species and also divergent lineages Oncethe candidate species are formally recognised the original linksand data can be amended accordingly but the phylogeneticrepresentation would remain

Phenotypic analyses for taxonomic work (and often alsoevolutionary research) are usually centred upon measurement ofa fairly standard set of morphological characters (some that arecommon across many groups (lengths head proportions) andothers that are more group-specific) These data are gathered tocharacterise taxa and provide a framework to allow fieldworkersand others to readily identify specimens many of which lackmatching genetic samples However despite the large amountof time involved in gathering this data the majority ofmorphological information gathered in taxonomic revisionsremains difficult to access For instance over several decadesof work on the Western Australian reptile fauna Glenn Storr

(Australiarsquos most prolific Australian reptile taxonomist) wouldhave measured many thousands of specimens yet thesespecimen-level data (housed in notebooks in the WesternAustralian Museum) are inaccessible to most researchers

However these phenotypic data are extremely valuable Inthe first instance they provide the framework for further testingand refining previous taxonomic work More broadly biologistsare also increasingly linking morphological data with geneticand climatic datasets in order to understand patterns of micro-and macro-evolutionary change Some good recent examplesinclude using voucher material to detect temporal andaltitudinal patterns of body size variation in several groups(Gardner et al 2009 Leacheacute et al 2010)

Individual specimen records in the Atlas of Living Australiaprovide an opportunity to improve the efficiency and value ofphenotypic data capture While some morphological datacollected for revisionary work are difficult if not impossible tosynthesise across different studies (different authors often useslightly different techniques) there remains a subset of

Fig 3 Provisional taxonomy for a species complex of Australian lizards The Oedura marmorata complex in northern Australia includes numerousdivergent but often poorly sampled mitochondrial lineages (Oliver et al 2014b) Deep mitochondrial divergences numerous fixed allozyme differences andstrong morphological differentiation support the recognition of two candidate species in this clade lsquoGulfrsquo and lsquoNorthrsquo Both candidate lineages contain furtherdivergent mitochondrial clades that also show evidence of concordant allozyme differentiation and sometimes ecological and morphological divergenceLineages that are comparatively well sampled andor show evidence of divergence at multiple sources of information are provisionally recognised as ESUswhile poorly sampled lineages that show evidence of less divergence are not recognised at this stage Further sampling and additional data will lead to therefinement of this arrangement Photographs by Stewart Macdonald and Steve Richards


characters including snoutndashvent length and aspects of bodylength tail length and some head proportions that are probablyalready close to standardised across studies or many groups

Having these data publicly available may also serve toincrease the care and rigour with which they are collected andvetted

As with methods for generating the data there are diverse methods for employing multilocus evidence to test boundaries among candidatespecies (Fujita et al 2012) Moving beyond simplistic criteria like reciprocal monophyly which could be overly conservative across nucleargenes (Hudson and Coyne 2002) these new approaches apply explicit model-based tests for evolutionary independence that incorporate theamong-gene variance in coalescence times These approaches which typically contrast alternative hypotheses of one versus two lineages atproposed bifurcations promise to provide an objective statistical approach to species delineation Logically the process is to (1) proposecandidate lineages and membership of these from some mix of mtDNA phylogeography distinct phenotypes etc (2) propose relationshipsamong these (ie one or more assumed tree topologies) and (3) using independent evidence (nDNA genes) evaluate evidence for evolutionarydistinctiveness (ie time gt0 m ~0 see Fig IV) by sequentially collapsing nodes and comparing likelihood-AIC or Bayes Factors against a splitmodel

There are however some issues First incorrect allocation of individuals to candidate species at the outset can compromise these (and anyother) tests (Olave et al 2014) Given sufficient loci this can be overcome by clustering individuals a priori via ordination or a populationgenetic approach (eg STRUCTURE) Second under some conditions the outcome might be sensitive to assumed tree topology although atleast one method appears robust to error expected under realistic conditions (Zhang et al 2014) Third the models typically assume that eachcandidate species is randomly mating Where as is more usual there is isolation-by-distance sampling gaps could result in incorrect inferenceof multiple species though underlying models could be robust to isolation by distance or low migration (Nm ltlt 1) (Zhang et al 2011)

It is important to stress that the goal here is not to identify the smallest genotypic cluster (ie a semi-isolated component of ametapopulation) but rather truly independently evolving lineages In practice different inference methods available now can infer varyingnumbers of lineages especially in low-gene-flow taxa with fractal genealogical structure (Camargo et al 2012 Carstens et al 2013) Whetherthese different outcomes are a result of differences in model structure and assumptions or inadequacy of sampling (genes individuals Rittmeyerand Austin 2012) remains to be determined On the latter while it is true that relatively small numbers (10sndash100s) of genes are sufficient undermost conditions (Zhang et al 2011) using a large number (eg 1000s) could improve initial clustering and otherwise provide more robust tests(Leacheacute et al 2014)

Countering the ideal for objective coalescent-based procedures for inferring species from genomic data are the observations that (1)lineages that have split very recently could well merge again in the absence of strong prezygotic isolation (Rosenblum et al 2012) andconversely (2) there is increasing evidence that given strong divergent selection species can form despite modest levels of gene flow Theformer case would correspond to oversplitting and the latter to incorrect lumping Thus we and others (Padial et al 2010 Fujita et al 2012)continue to advocate for a pluralist approach in which congruence across multiple lines of evidence is desirable yet no one dimension(genomic phenotypic etc) is necessary

Box 3 Inferring species boundaries from genomic data

The structured coalescent model thatunderpins inference of species status frommultilocus data Divergence history of threespecies (AndashC) in relation to time (T) andpopulation size (N) with a mismatched genetree shown in red Delimitation methodsassume no gene flow between species (m)but are robust to low levels

T

NTopology

nDNA loci

Phenotypegenomicdivergenceintrogressionat contactzones etc

CandidateSpecies

ESUs

Reality check

Coalescentmodel tests

Predicted lineages

mtDNA phenotype

CBA

m

Fig IV Theoretical and hierarchical frameworks underpinning the assessments of evolutionary divergence and species boundaries from integrated taxonomic datasets


Ultimately it may be possible to work towards a systemwhereresearchers working on a particular group (with appropriateaccreditation) are able to directly enter measurement data intothe matching specimen record on the Atlas of Living Australiawith repatriation to individual museum databases These datacould also be routinely collected for newly vouchered specimensand the associated data uploaded as a matter of course If andwhere appropriate this information could be embargoed (similarto the current system on GenBank) While there are certainlyissues to work through increasing utilisation of systems suchas this would almost certainly provide a valuable framework forfuture taxonomic and evolutionary research

Entering the Genomic Age

Molecular systematics has been through a series of revolutions ndashfrom the introduction of allozyme electrophoresis andcytogenetics in the 1960sndash1970s RFLP analyses in the 1980sPCRgene amplification and Sanger sequencing (ormicrosatelliteloci) in the 1990s andmost recently next-generation sequencingof 100s to 1000s of loci (Hillis et al 1996 Lemmon andLemmon 2013) Interestingly this journey has taken us frommultilocus analysis and a view of the entire genome albeit atlow resolution to fixation on a few readily amplifiable genefragments and just now back to a genome perspectiveHerpetologists including those studying Australian specieshave always been early adopters of these new approaches (egKing 1979 Brown and Wright 1979 Gartside 1982 Lemmonand Lemmon 2012)

With the recent coupling of various genome-reductionmethods (Box 2) and massively parallel sequencing platformsit is now feasible to readily obtain 100s to 1000s of loci fromspecies and clades that previously lacked genome resourcesThus we havemoved frombeing reliant on a few often relativelyconservative loci to having more data than most model-basedanalytical methods can handle (Box 3) In essence we are nowmore limited by computation and inference methods than bynumbers of loci In both theory andmost applications only a few10s of sufficiently informative loci are required to resolve speciesboundaries and estimate relationships (Zhang et al 2011Lemmon and Lemmon 2013 Brandley et al unpubl data) Yetanalysis of 1000s of independent loci then enable more preciseestimation of demographic and divergence histories The latterincludes testing for the extent of gene flow across borders ofcandidate lineages where they now meet (Singhal and Moritz2014)

We expect that over the next few years these lsquonext-genrsquoapproaches will become increasingly accessible to thesystematics community and will undoubtedly yield newinsights into processes of speciation including geneticexchange across porous species boundaries It is also true thatmoving beyond a small number (often lt5) of relativelyconserved nuclear loci is important to provide more robustdelineation of species through genetic methods (eg Leacheacuteet al 2014) and is crucial when this is the primary source ofevidence (eg phenotypically cryptic species) or whererelatively few samples can be obtained because of intrinsicrarity or access issues (eg Oliver et al 2012) That said thereremains much value in applying current methods (eg

integration across mtDNA phenotypes and a few nDNA loci)to identify obviously divergent species leaving the moreintensive genome-scale analyses to more difficult cases ordetailed studies where the goal is to infer divergence processesIn the former case it should also be recalled that allozymeelectrophoresis remains a very effective tool for speciesdelineation (eg Horner and Adams 2007 Oliver et al 20072010)

The bottom line is that a deluge of increasingly extensivegenome-scale data are coming and as a community systematistsneed to identify howbest to exploit this to improve our knowledgeof diversity of species their relationships and the evolutionaryprocesses bywhich they formDespitemuchpromise (Fujita et al2012) much remains to be done to test and improve on currentanalytical methods for delineating species from multilocus data(Camargo et al 2012 Carstens et al 2013 Miralles and Vences2013) (Box 3) But this will come and with suitably conservativeand rigorous application will enrich our understanding ofdiversity

Conclusions

Systematists aim to provide frameworks that effectively captureevolutionary biodiversity However the more we learn aboutspeciation and diversification processes the more we realise thatspecies delineation will never be a simple rule-driven matterAlong with many others we maintain that there is great value inmaintaining and in fact strengthening the nexus betweengenomic andphenotypic analyses of diversity Indeed to giveoneof these primacy over the other is not only intellectually flawedbut also misses great opportunities for understandingevolutionary processes across different scales (eg Hoskin et al2011 Rabosky et al 2014b)With this realisation comes the needto improve access to and visualisation of taxon characteristics(DNA sequences traits distributions) in the context of the tree oflife from the very tips (ESUs and CSs) through recognisedspecies to whole clades Reasonably well known and highlyendemic groups such as Australian reptiles provide a uniqueopportunity to develop approaches for dealing with thesechallenges and provide a road map for understanding thediversity of developing countries and more poorly knowntaxonomic groups

Acknowledgements

We thank theAtlas of LivingAustralia for providing data StewartMacdonaldfor compiling a species accumulation curve for Australian geckos and BradMaryan Stewart MaDonald Steve Sass and Stephen Richards for providingphotographsWealso thankDanRaboskyandElsevier Press for permission torepublish the image used in Fig 1 We all thank the Australian ResearchCouncil for ongoing support

References

Bell R C Parra J L TonioneM Hoskin CMacKenzie J BWilliamsS E andMoritz C (2010) Patterns of persistence and isolation indicateresilience to climate change in montane rainforest lizards MolecularEcology 19 2531ndash2544

Bi K Vanderpool D Singhal S Moritz C and Good J M (2012)Transcriptome-based exon capture enables highly cost-effectivecomparative genomic data collection at moderate evolutionary scalesBMC Genomics 13 403 doi1011861471-2164-13-403


Bickford D Lohman D J Sodhi N S Ng P K Meier R Winker KIngram K and Das I (2007) Cryptic species as a window on diversityand conservation Trends in Ecology amp Evolution 22 148ndash155doi101016jtree200611004

Bond J E andStockmanAK (2008)An integrativemethod for delimitingcohesion species finding the populationndashspecies interface in a group ofCalifornian trapdoor spiders with extreme genetic divergence andgeographic structuring Systematic Biology 57 628ndash646 doi10108010635150802302443

Brown W M and Wright J W (1979) Mitochondrial DNA analyses andthe origin and relative age of parthenogenetic lizards (genusCnemidophurus) Science 203 1247ndash1249 doi101126science424751

Camargo A Morando M Avila L J and Sites J W (2012) Speciesdelimitation with ABC and other coalescent-based methods a test ofaccuracy with simulations and an empirical example with lizards of theLiolaemus darwinii complex (Squamata Liolaemidae) Evolution 662834ndash2849 doi101111j1558-5646201201640x

Carnaval A C Waltari E Rodrigues M T Rosauer D VanDerWal JDamasceno R Prates I Strangas M Spanos Z Rivera D Pie MFirkowskiCR BornscheinMR Ribeiro L F andMoritz C (2014)Prediction of phylogeographic endemism in an environmentally complexbiome Proceedings of the Royal Society B Biological Sciences 28120141461 doi101098rspb20141461

Carstens B C Pelletier T A Reid N M and Satler J D (2013) How tofail at species delimitation Molecular Ecology 22 4369ndash4383doi101111mec12413

Catullo R and Keogh J S (2014) Aridification drove repeated episodesof diversification between Australian biomes evidence from a multi-locus phylogeny of Australian toadlets (Uperoleia Myobatrachidae)Molecular Phylogenetics and Evolution 79 106ndash117 doi101016jympev201406012

Chapman A D (2009) lsquoNumbers of Living Species in Australia and theWorldrsquo 2nd edn (Australian Biological Resources Study Canberra)

Chapple D G (2003) Ecology life-history and behavior in the Australianscincid genus Egernia with comments on the evolution of complexsociality in lizards Herpetological Monograph 17 145ndash180doi1016550733-1347(2003)017[0145ELABIT]20CO2

Chen I-P Symonds M R E Melville J and Stuart-Fox D (2013)Factors shaping the evolution of colour patterns in Australian agamidlizards (Agamidae) a comparative study Biological Journal of theLinnean Society 109 101ndash112 doi101111bij12030

Dayrat B (2005) Towards integrative taxonomy Biological Journal of theLinnean Society 85 407ndash415 doi101111j1095-8312200500503x

deQueirozK (1998) The general lineage concept of species species criteriaand the process of speciation a conceptual unification and terminologicalrecommendations In lsquoEndless Forms Species and Speciationrsquo (Eds D JHoward and S H Berlocher) pp 57ndash75 (Oxford University PressOxford)

de Queiroz K (2007) Species concepts and species delimitation SystematicBiology 56 879ndash886 doi10108010635150701701083

Dobzhansky T (1970) lsquoGenetics of the Evolutionary Processrsquo (ColumbiaUniversity Press New York)

Donnellan SAdamsMHutchinsonM andBaverstock PR (1993) Theidentification of cryptic species in the Australian herpetofauna a highresearch priority In lsquoHerpetology in Australia a Diverse Disciplinersquo(Eds D Lunney and D Ayres) pp 121ndash126 (Surrey Beatty Sydney)

Donnellan S C Hutchinson M N Dempsey P and Osborne W S(2002) Systematics of the Egernia whitii species group (LacertiliaScincidae) in south-eastern Australia Australian Journal of Zoology 50439ndash459 doi101071ZO01065

Doughty P and Oliver P M (2013) Systematics of Diplodactylus(Squamata Diplodactylidae) from south-western Australia redefinitionof D polyophthalmus and the description of two new species Recordsof the Western Australian Museum 28 44ndash65

Doughty P Oliver P M and Adams M (2008) Systematics of stonegeckos in the genus Diplodactylus (Reptilia Diplodactylidae) fromnorthwestern Australia with a description of a new species from theNorthwest Cape Western Australia Records of the Western AustralianMuseum 24 247ndash265

Faircloth B C McCormack J E Crawford N G Harvey M GBrumfield R T and Glenn T C (2012) Ultraconserved elementsanchor thousands of genetic markers spanning multiple evolutionarytimescales Systematic Biology 61 717ndash726 doi101093sysbiosys004

Fouquet A Gilles A VencesMMarty C BlancM andGemmell N J(2007) Underestimation of species richness in Neotropical frogsrevealed by mtDNA analyses PLoS ONE 2(10) e1109 doi101371journalpone0001109

Fujita M K McGuire J A Donnellan S C and Moritz C M (2010)Diversification at the aridndashmonsoonal interface Australia-widebiogeography of the Bynoersquos gecko (Heteronotia binoei Gekkonidae)Evolution 64 2293ndash2314

Fujita M K Leacheacute A D Burbrink F T McGuire J A and Moritz C(2012) Coalescent-based species delimitation in an integrativetaxonomy Trends in Ecology amp Evolution 27 480ndash488 doi101016jtree201204012

Gardner J Heinsohn R and Joseph L (2009) Shifting latitudinal clines inavian body size correlate with global warming in Australian passerinesProceedings of the Royal Society of London B Biological Sciences 2763845ndash3852

Gartside D F (1982) The Litoria ewingi complex (Anura Hylidae) insoutheastern Australia VI Geographic variation in transferrins offour taxa Australian Journal of Zoology 30 103ndash113 doi101071ZO9820103

Graham C H Ferrier S Huettman F Moritz C and Peterson A T(2004) New developments in museum-based informatics andapplications in biodiversity analysis Trends in Ecology amp Evolution 19497ndash503 doi101016jtree200407006

Grant V (1981) lsquoPlant Speciationrsquo 2nd edn (Columbia University PressNew York)

Grigg J W and Buckley L B (2014) Conservatism of lizard thermaltolerances and body temperature across evolutionary history andgeography Biology Letters 9 20121056

Gruumlndler M C and Rabosky D L (2014) Trophic divergence despitemorphological convergence in a continental radiation of snakesProceedings of the Royal Society of London B Biological Sciences 28120140413 doi101098rspb20140413

Haines M L Mousalli A Stuart-Fox D Clemann N and Melville J(2014) Phylogenetic evidence of historical mitochondrial introgressionand cryptic diversity in the genus Pseudomoia (Squmata Scincidae)Molecular Phylogenetics and Evolution 81 86ndash95 doi101016jympev201409006

Hillis D M Moritz C and Mable B K (1996) lsquoMolecular Systematicsrsquo(Sinauer Associates Sunderland MA)

Horner P and Adams M (2007) A molecular systematic assessmentof species boundaries in Australian Cryptoblepharus (ReptiliaSquamata Scincidae) ndash a case study for the combined use ofallozymes and morphology to explore cryptic biodiversity The BeagleRecords of the Museums and Art Galleries of the Northern Territory1ndash19

Hoskin C J (2007) Description biology and conservation of a new speciesof Australian tree frog (Amphibia Anura Hylidae Litoria) and anassessment of the remaining populations of Litoria genimaculataHorst 1883 systematic and conservation implications of an unusualspeciation event Biological Journal of the Linnean Society 91 549ndash563doi101111j1095-8312200700805x

Hoskin C J and Couper P J (2013) A spectacular new leaf-tailedgecko (CarphodactylidaeSaltuarius) from theMelvilleRange north-eastAustralia Zootaxa 3717 543ndash558 doi1011646zootaxa371746


Hoskin C J and Couper P J (2014) Two new skinks (ScincidaeGlaphyromorphus) from rainforest habitats in north-eastern AustraliaZootaxa 3869 001ndash016 doi1011646zootaxa386911

Hoskin C J Higgie M McDonald K R and Moritz C (2005)Reinforcement drives rapid allopatric speciationNature437 1353ndash1356doi101038nature04004

Hoskin C J Tonione M Higgie M MacKenzie J B Williams S EVanDerWal J and Moritz C (2011) Persistence in peripheral refugiapromotes phenotypic divergence and speciation in a rainforest frogAmerican Naturalist 178 561ndash578 doi101086662164

Hudson R R and Coyne J A (2002) Mathematical consequences of thegenealogical species concept Evolution 56 1557ndash1565 doi101111j0014-38202002tb01467x

Hull D L (1977) The ontological status of species as evolutionary unitsIn lsquoFoundational Problems in the Special Sciencesrsquo (Eds R Butts andJ Hintikka) pp 91ndash102 (D Reidel Publishing Company DordrechtHolland)

Hutchinson M N and Donnellan S C (1992) Taxonomy and geneticvariation in the Australian lizards of the genus Pseudemoia (ScincidaeLygosominae) Journal of Natural History 26 215ndash264 doi10108000222939200770091

Jolley-Rogers G Varghese T Harvey P dos Remedios N and MillerJ T (2014) PhyloJIVE integrating biodiversity data with the Treeof Life Bioinformatics 30 1308ndash1309 doi101093bioinformaticsbtu024

Kawakami T Smeds L BackstroumlmNHusbyAQvarnstroumlmAMugalC F Olason P and Ellegren H (2014) A high-density linkagemap enables a second-generation collared flycatcher genome assemblyand reveals the patterns of avian recombination rate variation andchromosomal evolution Molecular Ecology 23 4035ndash4058doi101111mec12810

Kay G and Keogh J S (2012) Molecular phylogeny and morphologicalrevision of the Ctenotus labillardieri (Reptilia Squamata Scincidae)species group and a new species of immediate conservation concern inthe southwestern Australian biodiversity hotspot Zootaxa 3390 1ndash18

Kearney M and Shine R (2004) Developmental success stability andplasticity in closely-related parthenogenetic and sexual lizards(Heteronotia Gekkonidae) Evolution 58 1560ndash1572 doi101111j0014-38202004tb01736x

Keogh J S Scott I AW and Hayes C (2005) Rapid and repeated originof insulargigantismanddwarfism inAustralian tiger snakesEvolution59226ndash233 doi101111j0014-38202005tb00909x

King M (1979) Karyotypic evolution in Gehyra (GekkonidaeReptilia)I The Gehyra variegatandashpunctata complex Australian Journal ofZoology 27 373ndash393 doi101071ZO9790373

King M (1982) Karyotypic evolution in Gehyra (Gekkonidae Reptilia)II A new species from the Alligator Rivers region in northern AustraliaAustralian Journal of Zoology 30 93ndash101 doi101071ZO9820093

Leacheacute A D Helmer D and Moritz C (2010) Phenotypic evolution inhigh elevation populations of western fence lizards (Sceloporusoccidentalis) in the Sierra Nevada Mountains Biological Journal ofthe Linnean Society 100 630ndash641 doi101111j1095-8312201001462x

Leacheacute A D Fujita M K Minin V and Bouckaert R (2014) Speciesdelimitation using genome-wide SNP data Systematic Biology 63534ndash542 doi101093sysbiosyu018

Lemmon A R and Lemmon E M (2012) High-throughput developmentof informative nuclear loci for shallow-scale phylogenetics andphylogeography Systematic Biology 61 745ndash761 doi101093sysbiosys051

LemmonEM andLemmonAR (2013)High-throughputgenomicdata insystematics and phylogenetics Annual Review of Ecology Evolution andSystematics 44 99ndash121 doi101146annurev-ecolsys-110512-135822

Macdonald S (2014) Australian Reptiles Online Database Available athttpwwwarodcomauarod (accessed 20 October 2014)

Mallet J (2005) Hybridization as an invasion of the genome Trends inEcology amp Evolution 20 229ndash237 doi101016jtree200502010

Marin J Donnellan S C Hedges S B Puillandre N Aplin K PDoughty P Hutchinson M N Couloux A and Vidal N (2013)Hidden species diversity of Australian burrowing snakes(Ramphotyphlops) Biological Journal of the Linnean Society 110427ndash441 doi101111bij12132

Mayr E (1970) lsquoPopulations Species and Evolutionrsquo (Harvard UniversityPress Cambridge MA)

McLeanCAMoussalliA Sass S andStuart-FoxD (2013) Taxonomicassessment of the Ctenophorus decresii complex (Reptilia Agamidae)reveals a new species of dragon lizard from western New South WalesRecords of the Australian Museum 65 51ndash63 doi103853j2201-43496520131600

Melville J E Smith K L Hobson R Hunjan S and Shoo L (2014)The role of integrative taxonomy in the conservation managementof cryptic species the taxonomic status of endangered earlessdragons (Agamidae Tympanocryptis) in the grasslands ofQueensland Australia PLoS ONE 9 e101847 doi101371journalpone0101847

Miralles A and Vences M (2013) New metrics for comparison oftaxonomies reveal striking discrepancies among species delimitationmethods in Madascincus lizards PLoS ONE 8 e68242 doi101371journalpone0068242

Mittermeier R A Robles-Gil P and Mittermeier C G (Eds) (1997)lsquoMegadiversity Earthrsquos Biologically Wealthiest Nationsrsquo (CEMEXAgrupaciaon Sierra Madre Mexico City)

Moritz C (1994) Defining lsquoEvolutionarily Significant Unitsrsquo forconservation Trends in Ecology amp Evolution 9 373ndash375 doi1010160169-5347(94)90057-4

Moritz C (2002) Strategies to protect biological diversity and theevolutionary processes that sustain it Systematic Biology 51 238ndash254doi10108010635150252899752

MoritzCHoskinC JMacKenzie J B PhillipsB L TonioneM SilvaN VanDerWal J Williams S E and Graham C H (2009)Identification and dynamics of a cryptic suture zone in tropical rainforestProceedings of the Royal Society of London B Biological Sciences 2761235ndash1244 doi101098rspb20081622

Olave M Solagrave E and Knowles L L (2014) Upstream analyses createproblems with DNAbased species delimitation Systematic Biology 63263ndash271 doi101093sysbiosyt106

Oliver P M and Lee M S Y (2010) The botanical and zoological codesimpede biodiversity research by discouraging publication of unnamednew species Taxon 59 1201ndash1205

Oliver PM Hugall A H AdamsM A Cooper S J B andHutchinsonM N (2007) Genetic elucidation of cryptic and ancient diversity in agroup of Australian diplodactyline geckos the Diplodactylus vittatuscomplex Molecular Phylogenetics and Evolution 44 77ndash88doi101016jympev200702002

Oliver P Doughty P Hutchinson M N Lee M S Y and Adams A(2009) The taxonomic impediment in vertebrates DNA sequenceallozyme and chromosomal data double estimates of species diversity ina lineage of Australian lizards (Diplodactylus Gekkota) Proceedings ofthe Royal Society of London B Biological Sciences 276 2001ndash2007doi101098rspb20081881

Oliver P M Adams M and Doughty P (2010) Extreme underestimationof evolutionary diversity within a nominal Australian gecko species(Crenadactylus ocellatus) BMC Evolutionary Biology 10 386doi1011861471-2148-10-386

Oliver P M Doughty P and Palmer R (2012) Hidden biodiversity inrare northern Australian vertebrates the case of the clawless geckos


(CrenadactylusDiplodactylidae) of theKimberleyWildlifeResearch39429ndash435 doi101071WR12024

Oliver P M Couper P and Pepper M (2014a) Independent transitionsbetween monsoonal and arid biomes revealed by systematic revison of acomplex of Australian geckos (Diplodactylus Diplodactylidae) PLoSONE 9 e111895 doi101371journalpone0111895

Oliver P M Smith K L Laver R L Doughty P and Adams M(2014b) Contrasting patterns of persistence and diversification in vicarsof a widespread Australian lizard lineage (the Oedura marmoratacomplex) Journal of Biogeography 41 2068ndash2079 doi101111jbi12364

Oliver P M Laver R L Melville J M and Doughty P (2014c) Anew species of Oedura from the limestone ranges of the southernKimberley Western Australia Zootaxa 3873 49ndash61 doi1011646zootaxa387314

Padial J M Miralles A De la Riva I and Vences M (2010) Theintegrative future of taxonomy Frontiers in Zoology 7 16 doi1011861742-9994-7-16

Pentildealba J V Smith L L TonioneMA Sass C Hykin SM SkipwithP L McGuire J A Bowie R C K and Moritz C (2014) Sequencecapture using PCR-generated probes (SCPP) a cost-effective methodof targeted high-throughput sequencing for non-model organismsMolecular Ecology Resources 14 1000ndash1010

Pepper M Doughty P and Keogh J S (2006) Molecular phylogenyand phylogeography of the Australian Diplodactylus stenodactylus(Gekkota Reptilia) species-group based on mitochondrial and nucleargenes reveals an ancient split between Pilbara and non-PilbaraD stenodactylus Molecular Phylogenetics and Evolution 41 539ndash555doi101016jympev200605028

Pepper M Doughty P Hutchinson M N and Keogh J S (2011a)Ancient drainages divide cryptic species in Australiarsquos arid zonemorphological and multi-gene evidence for four new species of beakedgeckos (Rhynchoedura) Molecular Phylogenetics and Evolution 61810ndash822 doi101016jympev201108012

PepperM FujitaMKMoritz C andKeogh J S (2011b) Palaeoclimatechange drove diversification among isolated mountain refugia in theAustralian arid zone Molecular Ecology 20 1529ndash1545 doi101111j1365-294X201105036x

PepperMDoughtyP andKeogh J S (2013)Geodiversity and endemismin the iconic Australian Pilbara region a review of landscape evolutionand biotic response in an ancient refugium Journal of Biogeography 401225ndash1239 doi101111jbi12080

Peterson B K Weber J N Kay E H Fisher H S and Hoekstra H E(2012) Double digest RADseq an inexpensive method for de novoSNP discovery and genotyping in model and non-model species PLoSONE 7 e37135 doi101371journalpone0037135

Pianka E R (1986) lsquoEcology and Natural History of Desert Lizardsrsquo(Princeton University Press Princeton NJ)

Pinho C and Hey J (2010) Divergence with gene flow models and dataAnnual Review of Ecology Evolution and Systematics 41 215ndash230doi101146annurev-ecolsys-102209-144644

Powney G D Grenyer R Orme C D L Owens I P F and Meiri S(2010) Hot dry and different Australian lizard richness is unlike that ofmammals amphibians and birds Global Ecology and Biogeography 19386ndash396 doi101111j1466-8238200900521x

Pyron R A Burbrink F T and Wiens J J (2013) A phylogenyand updated classification of Squamata including 4161 species of lizardsand snakes BMC Evolutionary Biology 13 93 doi1011861471-2148-13-93

Rabosky D L Donnellan S C Talaba A L and Lovette I J (2007)Exceptional among-lineage variation in diversification rates during theradiation of Australiarsquos largest vertebrate clade Proceedings of the RoyalSociety of London B Biological Sciences 274 2915ndash2923 doi101098rspb20070924

Rabosky D L Talaba A L Donnellan S C and Lovette I J (2009)Molecular evidence for hybridization between two Australian desertskinks Ctenotus leonhardii and Ctenotus quattuordecimlineatus(Scincidae Squamata) Molecular Phylogenetics and Evolution 53368ndash377 doi101016jympev200906020

Rabosky D L Hutchinson M N Donnellan S C Talaba A L andLovette I J (2014a) Phylogenetic disassemblyof species boundaries in awidespread group of Australian skinks (Scincidae Ctenotus)MolecularPhylogenetics and Evolution 77 71ndash82 doi101016jympev201403026

Rabosky D L Donnellan S C Grundler M and Lovette I J (2014b)Analysis and visualization of complex macroevolutionary dynamicsan example from Australian scincid lizards Systematic Biology 63610ndash627 doi101093sysbiosyu025

Reich D PattersonN KircherM Delfin F NandineniM R Pugach IKo A M Ko Y Jinam T A Phipps M E Saitou N Wollstein AKayser M Pa S and Stoneking M (2011) Denisova admixture andthe first modern human dispersals into Southeast Asia and OceaniaAmerican Journal of Human Genetics 89 516ndash528 doi101016jajhg201109005

ResideAEVanDerWal J PhillipsBL ShooLRosauerDAndersonB JWelbergen J AMoritz C Ferrier S andHarwood TD (2013)Climate change refugia for terrestrial biodiversity defining areas thatpromote species persistence and ecosystem resilience in the face of globalclimate change National Climate Change Adaptation Research FacilityCanberra

Rissler L J Hijmans R J Graham C H Moritz C and Wake D B(2006) Phylogeographic lineages and species comparisons inconservation analyses a case study of California herpetofaunaAmericanNaturalist 167 655ndash666 doi101086503332

Rittmeyer E N and Austin C C (2012) The effects of sampling ondelimiting species from multilocus data Molecular Phylogenetics andEvolution 65 451ndash463 doi101016jympev201206031

Rosenblum E B Sarver B A J Brown J W Des Roches S HardwickK M Hether T D Eastman J M Pennell M W and Harmon L J(2012) Goldilocks meets Santa Rosalia an ephemeral speciation modelexplains patterns of diversification across time scales EvolutionaryBiology 39 255ndash261 doi101007s11692-012-9171-x

Ryder O A (1986) Species conservation and systematics the dilemma ofsubspecies Trends in Ecology amp Evolution 1 9ndash10 doi1010160169-5347(86)90059-5

Schindel D and Miller S E (2010) Provisional nomenclaturethe on-ramp to taxonomic names In lsquoSystema Naturae 250 TheLinnaean Arkrsquo (Ed A Polaszek) pp 109ndash115 (CRC Press BocaRaton FL)

SheaG Couper PWorthingtonWilmer J andAmeyA (2011) Revisionof the genus Cyrtodactylus Gray 1827 (Squamata Gekkonidae) inAustralia Zootaxa 3146 1ndash63

SilerCDOaks JRCobbKOtaH andBrownRM (2014)Criticallyendangered island endemic or peripheral population of a widespreadspecies Conservation genetics of Kikuchirsquos gecko and the globalchallenge of protecting peripheral oceanic island endemic vertebratesDiversity amp Distributions 20 756ndash772 doi101111ddi12169

Singhal S and Moritz C (2014) Reproductive isolation betweenphylogeographic lineages scales with divergence Proceedings of theRoyal Society B Biological Sciences 280 20132246

Sistrom M Donnellan S C and Hutchinson M N (2013)Delimiting species in recent radiations with low levels of morphologicaldivergence a case study in Australian Gehyra geckos MolecularPhylogenetics and Evolution 68 135ndash143 doi101016jympev201303007

Sites J W and Marshall J C (2004) Operational criteria for delimitingspecies Annual Review of Ecology Evolution and Systematics 35199ndash227


Smith B T Harvey M G Faircloth B C Glenn T C and BrumfieldR T (2014) Target capture and massively parallel sequencing ofultraconserved elements (UCEs) for comparative studies at shallowevolutionary time scales Systematic Biology 63 83ndash95 doi101093sysbiosyt061

Smith K Harmon L J Shoo L P and Melville J E (2011) Evidenceof constrained phenotypic evolution in a cryptic species complex ofagamid lizards Evolution 65 976ndash992 doi101111j1558-5646201001211x

Soberoacuten J and Peterson T (2004) Biodiversity informatics managing andapplying primary biodiversity data Philosophical Transactions of theRoyal Society of London B Biological Sciences 359 689ndash698doi101098rstb20031439

Uetz P andHosek J (2014) The reptile database Available at httpwwwreptile-databaseorg (accessed 10 October 2014)

Vieites D R Wollenberg K C Andreone F Koumlhler J Glaw F andVences M (2009) Vast underestimation of Madagascarrsquos biodiversityevidenced by an integrative amphibian inventory Proceedings of theNational Academy of Sciences of the United States of America 1068267ndash8272 doi101073pnas0810821106

Wilson S andSwanG (2013) lsquoACompleteGuide toReptiles ofAustraliarsquo4th edn (New Holland Publishers Sydney)

Woinarski J C Legge S Fitzsimons J A Traill B J Burbidge A AFisher A Firth R S C Gordon I J Griffiths A D Johnson C NMcKenzie N L Palmer C Radford I Rankmore B Ritchie E GWard S and Ziembicki M (2011) The disappearing mammal fauna ofnorthern Australia context cause and response Conservation Letters 4192ndash201 doi101111j1755-263X201100164x

Yang M A Malaspinas A S Durand E Y and Slatkin M (2012)Ancient structure in Africa unlikely to explain neanderthal and non-African genetic similarity Molecular Biology and Evolution 292987ndash2995 doi101093molbevmss117

YeatesDK SeagoANelsonLCameronSL JosephL andTruemanJ W H (2011) Integrative taxonomy or iterative taxonomy SystematicEntomology 36 209ndash217 doi101111j1365-3113201000558x

Zhang C Zhang D Zhu T andYang Z (2011) Evaluation of a Bayesiancoalescent method of species delimitation Systematic Biology 60747ndash761 doi101093sysbiosyr071

Zhang C Rannala B and Yang Z (2014) Bayesian species delimitationcan be robust to guide-tree inference errors Systematic Biology 63993ndash1004 doi101093sysbiosyu052

Handling Editor Paul Cooper


wwwpublishcsiroaujournalsajz

sorting) such that additional independent data sources arerequired to test the hypothesis that these lineages representdistinctive evolutionary lineages or species (see below) Insome cases morphological characters that allow diagnosis inthe field are then identified (eg Shea et al 2011 Oliver et al2014a) However other species appear to be genuinely crypticand show little to no evidence of consistent morphologicaldifferentiation (Hoskin 2007 Smith et al 2011 Pepper et al2011a) In both cases assembling appropriate datasets toresolve species boundaries in species complexes is timeconsuming dependent on appropriate sampling and oftenexpensive

This comparative ease of initial genetic screening on theone hand juxtaposed against the challenges of speciesdelineation on the other has resulted in a proliferation ofcandidate species and genetic lineages that are not capturedby current taxonomy (Fouquet et al 2007 Vieites et al2009 Oliver et al 2010) This evolutionary diversity thusalso remains almost entirely missing from conservationassessments management planning and evolutionary analysesFinding new ways to effectively incorporate this unnameddiversity into biological research in the short term and

ultimately resolving the taxonomy of these lineages in thelonger term are two of the biggest challenges facingsystematists in the coming decade

Australian reptiles as a case study

Australia has an exceptionally diverse reptile fauna with over950 recognised species (Wilson and Swan 2013) This state ofknowledge is the result of decades of mostly painstaking peer-reviewed lsquoalpharsquo taxonomy based largely on morphologicalcharacters This creates the perception that the job is nearly doneexcept for ongoing discoveries in remote andor highly endemicregions (Chapman 2009 Hoskin and Couper 2013) By contrastgenetic studies continue to reveal additional highly divergentlineages within species as currently recognised (Pepper et al2006 2011b Fujita et al 2010 Oliver et al 2010 2014b Marinet al 2013) These data indicate that significant evolutionarydiversity remains uncaptured by taxonomy (see Box 1)Conversely there are also some instances where molecularanalyses have shown that taxonomy based on highly plasticcharacters inflates estimates of species diversity (Keogh et al2005 Rabosky et al 2014a) (Fig 1) These deficiencies in our

Box 1 Australiarsquos increasing reptile diversity geckos

(a) Diplodactylus conspicillatus complex

(b) Heteronotia binoei complex

(c) Oedura marmorata complex

9+ species or lineages Oliver et al 2014a

Fig II Hyperdiverse species complexes

10 Candidate species Fujita et al 2010

10 major lineages Oliver et al 2014b

Geckos (including pygopods) are the secondmost diverse groupof Australia reptiles and comprise close to 200 species Howevernew species continue to be recognised at a high rate ~20 ofrecognised species have been described since 2000 (Macdonald2014) (Fig I) While some new taxa are highly distinct anddivergent novelties most have been partitioned from previouslyrecognised species such asCyrtodactylus tuberculatus sl (Sheaet al 2011) and Rhynoedura ornata sl (Pepper et al 2011a)However even with this recent growth in recognised diversitypublished data indicate that species diversity remainsunderestimated (contra Chapman 2009) Of particular noteare hyperdiverse species complexes that comprise up to 10 (oreven more) highly divergent lineages (Fig I) Many of theselineages are corroborated by multiple independent datasets andclearly representundescribedspeciesOngoingresearch indicatesthat similarly diverse species complexes exist in genera such asAmolosia Gehyra and Heteronotia Furthermore regardlessof the final taxonomic status of these lineages it is also clear thatour current taxonomy does not fully capture the evolutionarydiversity of Australian geckos especially in the MonsoonTropics

200

Fig I Australian gecko speciesdiversity through time

150

100

Num

ber

of s

peci

es

50

0

1800 1850 1900

Year1950 2000

Data and graph courtesy of S Macdonald (2014)









(a) (b) (c) (e) (f ) (g)(d)

















Genome (1ndash2 Gb)

Genome reduction






SCPP (PCRprobes)2


Completeness

Connectivity

Investment

low

low-mod

low

high

high


high

high

high

high

high

moderate

moderate



Custom exoncapture5


Target enrichment














































T

NTopology

nDNA loci


CandidateSpecies

ESUs

Reality check


Predicted lineages

mtDNA phenotype

CBA

m










Conclusions


Acknowledgements


References






























































































































(a) (b) (c) (e) (f ) (g)(d)

















Genome (1ndash2 Gb)

Genome reduction






SCPP (PCRprobes)2


Completeness

Connectivity

Investment

low

low-mod

low

high

high


high

high

high

high

high

moderate

moderate



Custom exoncapture5


Target enrichment














































T

NTopology

nDNA loci


CandidateSpecies

ESUs

Reality check


Predicted lineages

mtDNA phenotype

CBA

m










Conclusions


Acknowledgements


References





































































































































Genome (1ndash2 Gb)

Genome reduction






SCPP (PCRprobes)2


Completeness

Connectivity

Investment

low

low-mod

low

high

high


high

high

high

high

high

moderate

moderate



Custom exoncapture5


Target enrichment














































T

NTopology

nDNA loci


CandidateSpecies

ESUs

Reality check


Predicted lineages

mtDNA phenotype

CBA

m










Conclusions


Acknowledgements


References































































































































Genome (1ndash2 Gb)

Genome reduction






SCPP (PCRprobes)2


Completeness

Connectivity

Investment

low

low-mod

low

high

high


high

high

high

high

high

moderate

moderate



Custom exoncapture5


Target enrichment














































T

NTopology

nDNA loci


CandidateSpecies

ESUs

Reality check


Predicted lineages

mtDNA phenotype

CBA

m










Conclusions


Acknowledgements


References


































































































































































T

NTopology

nDNA loci


CandidateSpecies

ESUs

Reality check


Predicted lineages

mtDNA phenotype

CBA

m










Conclusions


Acknowledgements


References



















































































































































T

NTopology

nDNA loci


CandidateSpecies

ESUs

Reality check


Predicted lineages

mtDNA phenotype

CBA

m










Conclusions


Acknowledgements


References






































































































































T

NTopology

nDNA loci


CandidateSpecies

ESUs

Reality check


Predicted lineages

mtDNA phenotype

CBA

m










Conclusions


Acknowledgements


References































































































































T

NTopology

nDNA loci


CandidateSpecies

ESUs

Reality check


Predicted lineages

mtDNA phenotype

CBA

m










Conclusions


Acknowledgements


References






























































































































Conclusions


Acknowledgements


References























































































































































































































































































































































































Documents

New approaches to cataloguing and understanding ...biology-assets.anu.edu.au/hosted_sites/Scott/2014oliveretalajz.pdf · New approaches to cataloguing and understanding evolutionary