Post on 24-May-2020
transcript
Information on this journal can be accessed at http://wileyonlinelibrary.com/journal/mec
The journal is covered by AGRICOLA, Chemical Abstracts, Current Awareness Biological Sciencesand Current Contents.
This journal is available at Wiley Online Library. Visit http://wileyonlinelibrary.com to searchthe articles and register for table of contents e-mail alerts.
409 The putative functional ecology and distribution of archaealcommunities in sponges, sediment and seawater in a coralreef environmentA. R. M. Polónia, D. F. R. Cleary, R. Freitas, N. J. de Voogd& N. C. M. Gomes
424 Long-term experimental warming alters communitycomposition of ascomycetes in Alaskan moist and dry arctictundraT. A. Semenova, L. N. Morgado, J. M. Welker, M. D. Walker, E. Smets & J. Geml
Molecular Adaptation438 Rapid acclimation of juvenile corals to CO2-mediated
acidification by upregulation of heat shock protein and Bcl-2 genesA. Moya, L. Huisman, S. Forêt, J.- P. Gattuso, D. C. Hayward, E. E. Ball & D. J. Miller
453 Fluorescent protein-mediated colour polymorphism in reefcorals: multicopy genes extend theadaptation/acclimatization potential to variable lightenvironmentsJ. R. Gittins, C. D’Angelo, F. Oswald, R. Edwards & J. Wiedenmann
Speciation and Hybridization466 Combined next-generation sequencing and morphology
reveal fine-scale speciation in Crocodile Skinks (Squamata: Scincidae: Tribolonotus)E. N. Rittmeyer & C. C. Austin
Phylogeography484 Phylogeography of var gene repertoires reveals fine-scale
geospatial clustering of Plasmodium falciparum populationsin a highly endemic areaS. K. Tessema, S. L. Monk, M. B. Schultz, L. Tavul, J. C. Reeder, P. M. Siba, I. Mueller & A. E. Barry
Kinship, Percentage and Behaviour498 Social structure varies with elevation in an Alpine ant
J. Purcell, L. Pellissier & M. Chapuisat
Ecological Interactions508 Stable coexistence of incompatible Wolbachia along a
narrow contact zone in mosquito field populationsC. M. Atyame, P. Labbé, F. Rousset, M. Beji, P. Makoundou, O. Duron, E. Dumas, N. Pasteur, A. Bouattour, P. Fort & M. Weill
MOLECULARECOLOGY VOLUME 24, NUMBER 2, JANUARY 2015
MO
LE
CU
LA
RE
CO
LO
GY
VO
LU
ME
24, NU
MB
ER
2, JAN
UA
RY
2015, pp.259–522
MOLECULARECOLOGY
VOLUME 24NUMBER 2JANUARY
2015
Published by
ISSN 0962-1083
NEWS AND VIEWSPerspective
259 Pushing the envelope in genetic analysis of species invasionS. A. Cushman
INVITED REVIEWS AND SYNTHESES263 Multicollinearity in spatial genetics: separating the wheat
from the chaff using commonality analysesJ. G. Prunier, M. Colyn, X. Legendre, K. F. Nimon & M. C. Flamand
ORIGINAL ARTICLESPopulation and Conservation Genetics
284 Human-aided and natural dispersal drive gene flow acrossthe range of an invasive mosquitoK. A. Medley, D. G. Jenkins & E. A. Hoffman
296 Life-stage differences in spatial genetic structure in anirruptive forest insect: implications for dispersal and spatialsynchronyP. M. A. James, B. Cooke, B. M. T. Brunet, L. M. Lumley, F. A. H. Sperling, M.- J. Fortin, V. S. Quinn & B. R. Sturtevant
310 Reconstructing the demographic history of orang-utans usingApproximate Bayesian ComputationA. Nater, M. P. Greminger, N. Arora, C. P. van Schaik, B. Goossens, I. Singleton, E. J. Verschoor, K. S. Warren & M. Krützen
328 Demographic inferences using short-read genomic data in anapproximate Bayesian computation framework: in silicoevaluation of power, biases and proof of concept in AtlanticwalrusA. B. A. Shafer, L. M. Gattepaille, R. E. A. Stewart & J. B. W. Wolf
346 High-stakes species delimitation in eyeless cave spiders(Cicurina, Dictynidae, Araneae) from central TexasM. Hedin
362 MHC variation reflects the bottleneck histories of NewZealand passerinesJ. T. Sutton, B. C. Robertson & I. G. Jamieson
374 Global invasion history of the tropical fire ant: a stowawayon the first global trade routesD. Gotzek, H. J. Axen, A. V. Suarez, S. H. Cahan & D. Shoemaker
Ecological Genomics389 Population genomics of natural and experimental
populations of guppies (Poecilia reticulata)B. A. Fraser, A. Künstner, D. N. Reznick, C. Dreyer & D. Weigel
mec_24_2_oc_Layout 1 13-01-2015 11:36 Page 1
High-stakes species delimitation in eyeless cave spiders(Cicurina, Dictynidae, Araneae) from central Texas
MARSHAL HEDIN
Department of Biology, San Diego State University, San Diego, CA 92182, USA
Abstract
A remarkable radiation of completely eyeless, cave-obligate spider species (Cicurina)has been described from limestone caves of Texas. This radiation includes over 50
described species, with a large number of hypothesized single-cave endemics, and four
species listed as US Federally Endangered. Because of this conservation importance,
species delimitation in the group is ‘high-stakes’– it is imperative that species hypoth-
eses are data rich, objective, and robust. This study focuses on a complex of four cave-
dwelling Cicurina distributed on the northwestern edge of Austin, Texas. Several of
the existing species hypotheses in this complex are weak, based on morphological
comparisons of small samples of adult female specimens; one species description (for
C. wartoni) is based on a single adult specimen. Species limits in this group were
newly assessed using morphological, mitochondrial and nuclear DNA sequence data
evidence, analysed using a variety of approaches. All data support a clear lineage sepa-
ration between C. buwata versus the C. travisae complex (including C. travisae,C. wartoni and C. reddelli). Observed congruence across multiple analyses indicate that
the C. travisae complex represents a single species, and the formal species synonymy
presented here has important conservation implications. The integrative framework
utilized in this study serves as a potential model for other Texas cave Cicurina, includ-ing US Federally Endangered species. More generally, this study illustrates how and
why taxon-focused conservation efforts must prioritize modern species delimitation
research (if the existing taxonomy is weak), before devoting precious downstream
resources to conservation efforts. The study also highlights the issue of taxonomic type
II error that diversity biologists increasingly face as species delimitation moves into
the genomics era.
Keywords: Bayesian phylogenetics & phylogeography, conservation biology, endangered
species, population structure, species delimitation, species synonymy
Received 24 September 2014; revision received 3 December 2014; accepted 3 December 2014
Introduction
Species delimitation is the process whereby systematists
or evolutionary biologists combine specimens, specimen
data and data analyses to delimit species taxa while
working under an explicit species concept. Species
delimitation is an original and fundamental goal of sys-
tematic biology (Wiens 2007; Camargo & Sites 2013).
For some researchers, this process has changed dramati-
cally over the past 10–20 years, these changes paralleling
a flood of DNA sequence data and the development of
new analytical delimitation tools for both molecules
and morphology (Fujita et al. 2012; Camargo & Sites
2013; Carstens et al. 2013). However, these changes have
not been universally adopted, as the majority of new
species are still delimited and described under a quali-
tative ‘distinct morphological groups’ sorting process
(Pante et al. 2015).
In principle, it could be claimed that all species
delimitations are ‘high-stakes’ hypotheses, following
familiar arguments that species are fundamental units
of evolutionary biology, ecology, etc. More realistically,
however, most initial species hypotheses have neverCorrespondence: Marshal Hedin, Fax: (619) 594 5676; E-mail:
mhedin@mail.sdsu.edu
© 2014 John Wiley & Sons Ltd
Molecular Ecology (2015) 24, 346–361 doi: 10.1111/mec.13036
been subsequently tested, and in fact, species hypothe-
ses do not even exist for a majority of Earth’s biodiver-
sity (e.g. Mora et al. 2011). That said, certain species
delimitations are particularly important for political,
economic, and/or applied reasons. Examples include
species complexes that include disease vectors, venom-
ous taxa, and taxa of conservation concern. With respect
to taxon-focused conservation efforts, it is fundamen-
tally important to rigorously and accurately delimit con-
servation units (at or below the species level) before
dedicating finite conservation resources (Paquin et al.
2008; Weisrock et al. 2010; Niemiller et al. 2012, 2013;
Malaney & Cook 2013; McCormack & Maley 2015).
Taxa restricted to cave habitats present special chal-
lenges for species delimitation. Complexes of cave-
dwelling populations/taxa are often distributed in strict
allopatry, such that cases of sympatry (and natural tests
of reproductive isolation) are rare or nonexistent. Simi-
lar selective pressures across spatially isolated cave hab-
itats promote morphological stasis or homoplasy
(Derkarabetian et al. 2010; Bendik et al. 2013; Derkarabe-
tian & Hedin 2014), and many studies have now shown
that hypothesized widespread cave species often consist
of multiple, morphologically similar or cryptic species
(Juan et al. 2010; Niemiller et al. 2012, 2013; Derkarabe-
tian & Hedin 2014). In contrast, constrained gene flow
among isolated cave habitats often leads to extreme
population genetic structuring (e.g. Hedin 1997; Chiari
et al. 2012). If only single gene data are available (e.g.
mtDNA only), population genetic structure can be diffi-
cult (or impossible) to distinguish from population
divergence and speciation. For example, the single-locus
generalized mixed Yule coalescent method (Pons et al.
2006; Fujisawa & Barraclough 2013), commonly applied
in modern species delimitation, is susceptible to over-
splitting species diversity in the face of strong popula-
tion structure (Lohse 2009; Keith & Hedin 2012; Satler
et al. 2013; Hamilton et al. 2014). Likewise, population
structure in the nuclear genome presents potential prob-
lems for newer multispecies coalescent methods (O’Me-
ara 2010; Camargo & Sites 2013). An example is
Bayesian phylogenetics and phylogeography (BPP;
Yang & Rannala 2010; Rannala & Yang 2013), a method
that assumes panmixia within species, but has been
suggested to potentially oversplit diversity in dispersal-
limited taxa (e.g. Niemiller et al. 2012; Barley et al. 2013;
Carstens & Satler 2013; McKay et al. 2013).
A remarkable radiation of completely eyeless, cave-
obligate spider species has been described from lime-
stone cave habitats of Texas. The subgenus Cicurella
(genus Cicurina) includes about 80 species, over 50 of
which are eyeless and restricted to Texas caves (Gertsch
1992; Paquin & Dup�err�e 2009); available evidence sug-
gests monophyly of this group (Paquin & Hedin 2004;
Paquin & Dup�err�e 2009). Species delimitation in eyeless
Cicurina is challenging for several reasons beyond the
generalities discussed above. Adult specimens are rare,
with an estimated ratio of immature/adult female/
adult male specimens of 100/10/1 (Paquin & Dup�err�e
2009). Most araneomorph spider species are described
based on a combination of both male and female adult
morphology, but male-based evidence is essentially
lacking in eyeless Cicurina. Female somatic morphology
is highly conserved among eyeless Cicurina species,
while female genitalic morphology often varies within
species, blurring the distinction between geographic
variation and species-level divergence (Paquin &
Dup�err�e 2009). Finally, access to Texas caves can be
challenging, leading to geographic sampling gaps that
can impact genealogical interpretations (Lohse 2009;
Niemiller et al. 2012). Despite these several difficulties,
species delimitation in the group is ‘high-stakes,’ with
over 35 blind Texas Cicurina species known only from
their respective type localities (Paquin & Dup�err�e 2009),
and four cave taxa listed as US Federally Endangered
Species (Longacre 2000). Because of this obvious conser-
vation and political importance, it is imperative that
species hypotheses in the group are robust.
This study focuses on a complex of four eyeless spe-
cies (Cicurina wartoni, C. reddelli, C. travisae and C. buw-
ata) restricted to caves in Travis and Williamson
Counties, Texas. Existing species hypotheses in this
complex are not data rich; for example, the taxonomic
description of C. wartoni is based on a single adult
female specimen, which remains the only adult speci-
men known for this taxon (Gertsch 1992; Paquin &
Dup�err�e 2009). Cicurina reddelli is likewise known only
from the type locality. Cicurina travisae and C. buwata
are known from females for a larger number of caves
(>7 caves per taxon), but adult males have never been
described for these species. Taking into account intra-
specific variation, female genital morphologies for
members of this complex are highly similar (Paquin &
Dup�err�e 2009; fig. 131), suggesting the possibility that
these four species actually comprise a single taxon. Spe-
cies delimitation in this complex is particularly impor-
tant because of the rarity and apparent endemicity of
contained taxa in a region impacted by development on
the northwestern edge of Austin. For example, the only
known location for C. wartoni is a small, shallow cave
with many threats (e.g. fire ants, pollution, trash dump-
ing, etc.), and this species has been a candidate for list-
ing as a US Federally Endangered Species (U.S. Fish &
Wildlife Service 2010).
The research summarized here seeks to clarify the
evolutionary independence of species in this complex,
with an emphasis on the distinctiveness of C. wartoni.
Because any single data source or analytical method is
© 2014 John Wiley & Sons Ltd
CAVE SPIDER SPECIES DELIMITATION 347
susceptible to error, most modern species delimitation
analyses now include both independent lines of evi-
dence (e.g. mtDNA data, nuclear data, morphology,
etc.) and comparisons of multiple analyses (Fujita et al.
2012; Carstens et al. 2013; Satler et al. 2013; Derkarabe-
tian & Hedin 2014). This integrative approach to the
species delimitation problem was used in this research.
Materials and methods
Specimen sampling
Specimens were available from 27 regional caves (Fig. 1,
Appendix 1), with one to three spiders sampled from
each cave. A priori specimen identification was based
on geographic location and/or genetic affiliation as fol-
lows: Cicurina buwata – sample including eyeless imma-
ture specimens from three caves (Buttercup Creek Cave,
Marigold Cave, and Testudo Tube) that are known loca-
tions for C. buwata (Paquin & Dup�err�e 2009). Eyeless
immature and/or adult specimens from eight additional
caves were identified as C. buwata based on consistent
placement into a C. buwata genetic clade (see Results);
C. wartoni – three eyeless immature specimens from the
type locality (Pickle Pit) were tentatively identified as
this species, as no other eyeless Cicurina have been
recorded from this cave; C. travisae – sample including
eyeless immature and/or adult specimens from five
caves (McDonald Cave, Amber Cave, Kretschmarr Dou-
ble Pit, North Root Cave, Tooth Cave = type locality)
that are known locations for C. travisae (Paquin &
Dup�err�e 2009); C. reddelli – an adult female and imma-
ture male specimen from Cotterell Cave, the type and
only known locality for C. reddelli; A priori unidentified –both immature and adult specimens from nine caves
geographically situated between Cotterell Cave and
McDonald Cave (Fig. 1, Appendix 1) are genetic mem-
bers of the clade of interest (see Results), but are from
Broken Arrow
Testudo
Marigold
ButtercupCreek
BabeLakeline
Weldon
NoRent
McNeilBat
AppleRiata
McDonald
Kretschmarr
Amber
TwoTrunks
Tooth
NorthRoot
Gallifer
Geode
Stovepipe
PicklePit
Spider
BeardRanch
KenButlerPit
JesterEstates
JestJohn
Cotterell
DiesRanchTreasure
Williamson County
Travis County
BB B
B
BB
B
B BB
BW
R
T
TT
T~ 5 km
T
(A) (B)
(C)
Fig. 1 (A) Map of Texas with counties,
Travis and Williamson counties high-
lighted. (B) Adult female Cicurina buwata
from Lakeline Cave (G2001). (C) Map
showing geographic distribution of
sampled cave populations – locations
approximate, to protect location anonym-
ity. Specimens allocated into five primary
a priori groups based on consideration of
morphology, geographic location and/or
genetic affiliation. Colours used to desig-
nate taxa (amber = C. buwata, blue =C. travisae, green = C. reddelli, red =C. wartoni, black boxes = ‘western undeter-
mined’, black circles = ‘eastern undeter-
mined’; see text for details).
© 2014 John Wiley & Sons Ltd
348 M. HEDIN
caves without previous records for adult Cicurina. Adult
females are available from some of these caves, but
because C. travisae, C. reddelli and C. wartoni have very
similar female epigynal morphologies (Paquin &
Dup�err�e 2009, fig. 131), these specimens were not iden-
tified to species a priori.
Morphology
Genitalic structures (female epigyna, male pedipalps)
for all adult specimens used in genetic analyses were
digitally imaged. Specimens were imaged using a
Visionary Digital BK Plus system (http://www.vision-
arydigital.com), including a Canon 5D digital camera,
Infinity Optics Long Distance Microscope, P-51 camera
controller and FX2 lighting system. Individual images
were combined into a composite image using ZERENE
STACKER v1.04 (http://www.zerenesystems.com/); this
composite image was then edited using Adobe Photo-
shop CS6. Female epigyna were dissected from speci-
mens using fine forceps, immersed for 2–5 min in
BioQuip specimen clearing fluid (http://www.bio-
quip.com) on a depression slide and then imaged in
this fluid on slides.
Mitochondrial analyses
Mitochondrial cytochrome oxidase I (COI) data were
gathered for 46 specimens from 26 caves (Appendix 1),
using PCR and genomic DNA extracted from leg tis-
sues. Amplified PCR products were purified using Mil-
lipore plates and Sanger-sequenced in both directions at
Macrogen USA. DNA sequences were edited using GENE-
IOUS PRO R7 software (http://www.geneious.com/) and
trimmed to exclude primer sequences. Original
sequences were supplemented with transcriptome-
derived COI data for C. vibora (see below) and pub-
lished data for other Texas cave- and surface-dwelling
Cicurina (Paquin & Hedin 2004).
A COI gene tree was reconstructed using maximum
likelihood, using RAXML searches (Stamatakis 2006, 2014)
as implemented in RAXMLGUI 1.31 (Silvestro & Michalak
2012). This analysis included a thorough bootstrap
analysis (1000 bootstrap replicates) followed by multiple
inferences (100) on an alignment with redundant haplo-
types collapsed, using a GTR_Γ model for separate
codon partitions. A mitochondrial gene tree was also
reconstructed using Bayesian inference in the BEAST
v1.7.2 package (Drummond et al. 2012). BEAST analyses
were conducted using an uncorrelated lognormal
relaxed clock model (Drummond et al. 2006), imple-
menting best-fit models of molecular evolution chosen
using default parameters in PARTITIONFINDER v1.1.1 (Lan-
fear et al. 2012). Two replicate MCMC chains were run
for 50 million generations, with sampling every 1000
generations, using the ‘auto optimize’ operators option,
and a Speciation: Yule Process tree prior. The consis-
tency of parameter estimates across replicate runs was
assessed using Tracer (Drummond et al. 2012), and
results of replicate runs were combined such that ESS
values exceeded 200 for all parameters. LOGCOMBINER
was used to combine separate tree files
(burnin = 20 000), with a reduced resample frequency
of 25 000. From this reduced tree sample, TREEANNOTATOR
was used to reconstruct a maximum clade credibility
(mcc) tree.
The RAxML gene tree was used as input in a Bayes-
ian Poisson Tree Processes (bPTP) analysis, as imple-
mented on the bPTP server (http://species.h-its.org/
ptp/; Zhang et al. 2013). PTP is a single-locus species
delimitation method using only nucleotide substitution
information, implementing a model assuming gene tree
branch lengths generated by two independent Poisson
process classes (within- and among-species substitution
events). The bPTP analysis was run using 100 000
MCMC generations, with a thinning of 100 and burn-in
of 0.1. In addition, the relaxed clock BEAST mcc tree was
used as input in single- and multiple-threshold GMYC
analyses (http://species.h-its.org/gmyc/). Available
simulation studies suggest that PTP outperforms GMYC
(Pons et al. 2006; Fujisawa & Barraclough 2013) for sin-
gle-locus species delimitations (Zhang et al. 2013), with-
out requiring an ultrametric tree.
Nuclear marker development and analysis
Cicurina-specific nuclear phylogenetic markers were
developed from comparative transcriptome data (sum-
marized in Appendix S1, Supporting information). After
preliminary PCR primer testing and sequencing, eight
nuclear gene regions were chosen for comprehensive
specimen sampling (primers, PCR conditions and tran-
script annotations are provided in Appendix S2, Sup-
porting information). Collection of nuclear gene data
specifically targeted the four focal species from 26
regional caves (Appendix 1). This focused sampling fol-
lows from COI gene tree results, which indicate the
monophyly of this species complex (see below). Also,
although eyeless Cicurina occur in the karst landscapes
to the north and south of the focal region (e.g. C. vibora,
C. bandida), these taxa are both morphologically (Paquin
& Dup�err�e 2009) and genetically distinct (see Results)
from the focal species.
Nuclear PCR products were purified, Sanger-
sequenced and edited as summarized above for mito-
chondrial sequences. The C. travisae transcriptome
data were not used for downstream phylogenetic
analyses since heterozygosity could not be assessed
© 2014 John Wiley & Sons Ltd
CAVE SPIDER SPECIES DELIMITATION 349
(transcriptomes derived from two individuals; Appendix
S1). Only one of the nuclear matrices included indels –for the B2_F10F11 matrix, a three-amino-acid indel
restricted to C. buwata specimens was recoded as a two-
state nucleotide transition. Heterozygous nuclear
sequences were bioinformatically phased to alleles. SEQ-
PHASE (Flot 2010) was used to convert matrices for input
into PHASE 2.1.1 (Stephens et al. 2001; Stephens & Donnel-
ly 2003) using default settings (phase threshold = 90%,
100 iterations, thinning interval = 1, burn-in = 100). Gene
trees for individual nuclear genes were reconstructed
using RAxML searches (as above), applying a single
GTR_Γ model to each gene region. Gene trees included
data for focal taxa, rooted with orthologous sequences
from the C. vibora transcriptome (Appendix S1).
Multigenic nuclear genetic distances among individu-
als were calculated using POFAD v1.03 (Joly & Bruneau
2006). Uncorrected p-distances for each locus were cal-
culated in PAUP v4.0b10 (Swofford 2002), and because all
nuclear matrices are similar in aligned length and levels
of variation, individual matrices were standardized to
have the same weight (standardized weights option).
Analyses using nonstandardized distances gave very
similar results (results not shown). POFAD distances
were used to reconstruct a NeighborNet network in
SPLITSTREE4 (Huson & Bryant 2006). To eliminate the
potentially confounding influence of female-based pop-
ulation structure, mitochondrial data were not included
in POFAD analyses. Also, more distant outgroups (i.e.
C. vibora) were excluded from this analysis. This
‘nuclear + ingroup only’ data set was also used for
Bayesian clustering and BPP analyses summarized
below.
Bayesian genetic clustering analyses were conducted
using STRUCTURAMA 2.0 (Huelsenbeck et al. 2011) and
STRUCTURE (Pritchard et al. 2000, 2010). Structurama treats
the number of populations (K = number of distinct
genetic clusters) as a random variable following a Di-
richlet process prior (Pella & Masuda 2006; Huelsen-
beck & Andolfatto 2007; Huelsenbeck et al. 2011), while
STRUCTURE requires multiple analyses at different fixed K
values, and the use of ad hoc statistics to choose an
optimal K value. For both analyses, SNAP Map (Price &
Carbone 2005; Aylor et al. 2006) was used to convert
phased nuclear DNA sequences to numbered unique
alleles (haplotypes). Both Structurama and STRUCTURE
have been used to discover the number of distinct
genetic clusters in other species delimitation studies
(e.g. Leach�e & Fujita 2010; Weisrock et al. 2010; Carstens
& Satler 2013; Satler et al. 2013).
Structurama settings were as follows: model num-
pops=rv, admixture=no, concparmprior=gamma(0.1,10),
mcmc ngen=1 000, 000, samplefreq=100, printfreq=1000,calcmarginal likelihood burnin=1000. Runs with differ-
ent alpha values (shape and scale values of gamma dis-
tribution) were conducted as follows: (1,1) (1,5) (1,10)
(0.1,1) (0.1,5) (0.1,10). STRUCTURE runs were conducted
assuming between 1 and 6 genetic clusters (K = 1–6),with analyses for each K value replicated three times.
Analyses used an admixture model with a burn-in of
100 000 (one million MCMC steps after burn-in), and
allele frequencies were considered independent among
populations. Both the maximum value of the log proba-
bility of the data given K (L(K)) and DK (= rate of
change in log probability between successive K values,
Evanno et al. 2005) were used to identify an optimal K
value. Estimates from multiple replicates for multiple K
values were calculated in STRUCTURE HARVESTER (Earl &
vonHoldt 2012), and data were summarized using the
FullSearch algorithm of CLUMPP (Jakobsson & Rosenberg
2007) and visualized with DISTRUCT (Rosenberg 2004).
The Bayesian phylogenetics and phylogeography
method (BPP, Yang & Rannala 2010; Rannala & Yang
2013) was used with the multilocus nuclear data to cal-
culate the posterior probabilities of different species
delimitation models. The rjMCMC species delimitation
method was used with algorithm 1 (a value = 2, m value
= 1), ambiguous data columns were not removed (clean-
data =0), and the analysis was set for automatic fine-tune
adjustments. Two different combinations for population
size (theta, h) gamma priors were used, including a large
theta prior G(alpha = 1, beta = 10) and a more diffuse
(uninformative) theta prior G(alpha = 2, beta = 100). For
both analyses, a diffuse tau (s) gamma prior of G(alpha
= 2, beta = 1000) was used for the age of the root (tau0),
while the other divergence time parameters were
assigned the Dirichlet prior (Yang & Rannala 2010). Each
prior combination was run twice to check for conver-
gence and proper mixing. Analyses were run for 100 000
generations, sampling every five generations with 10 000
burnin. Species tree nodes with posterior probability val-
ues >0.95 were considered supported species delimita-
tions; values below 0.95 were considered as evidence for
collapsing a species tree node.
Bayesian phylogenetics and phylogeography analyses
require a priori specimen allocation and an input guide
tree – two alternatives were considered. First, a conser-
vative analysis was conducted in which only specimens
from known cave locations were included (see Fig. 1).
The guide tree used was as follows: (C. buwata sister to
(C. travisae, (C. reddelli + C. wartoni)), which is a topol-
ogy consistent with POFAD, STRUCTURE (K = 3) and
Structurama results (see below). Second, following
results of POFAD, STRUCTURE (K = 3) and Structurama,
and geography (Fig. 1), four ‘undetermined’ cave popu-
lations in the vicinity of C. travisae (Two Trunks Cave,
Gallifer Cave, Geode Cave, Stovepipe Cave) were
allocated to C. travisae, and five eastern ‘undetermined’
© 2014 John Wiley & Sons Ltd
350 M. HEDIN
populations (Spider Cave, Ken Butler Pit, Jester Estates
Cave, Jest John Cave, Beard Ranch Cave) were allocated
to C. reddelli. The same guide tree as above was used.
Results
Morphology
Although formal quantitative analyses were not con-
ducted because of small sample sizes, consideration of
patterns of qualitative variation provided important
information. The epigynal morphology of available
female specimens is consistent with that reported for
the focal species of interest (compare Fig. 2 to Paquin &
Dup�err�e 2009; fig. 131). However, defining ‘diagnostic’
epigynal features for different species in this complex is
very challenging, since epigynal variation within a spe-
cies (or cave population) sometimes exceeds variation
between hypothesized species (e.g. Tooth Cave speci-
mens G1958, G2014; Fig. 2; see also fig. 5b of Paquin
(K)
(L)
(M) (N)
(A) (B) (C)
(D)(E) (F)
(G) (H) (I)
(J)
(O)
Fig. 2 Epigynal (ventral view) and palpal
(left palp, ventral view) morphologies for
adult specimens used in this study. (A–E), Cicurina buwata females: (A) Broken
Arrow Cave, G1961; (B) Broken Arrow
Cave, G1959; (C) Babe Cave, G2011; (D)
Buttercup Creek Cave, G2007; (E) Apple
Riata, G1997; Specimen determinations
for C. buwata based primarily on phylo-
genetic placement in DNA analyses,
except for Buttercup Creek Cave (known
location for C. buwata). (F) C. reddelli
female, Cotterell Cave, G 1960 (type
locality for C. reddelli). (G–J) C. travisae
females: G) North Root Cave, G1970; (H)
Tooth Cave, G1958; (I) Tooth Cave,
G2014; (J) Kretschmarr Double Pit,
G1966; all known localities for C. travisae,
including type locality (Tooth Cave). (K)
‘eastern undetermined’ female, Jester
Estates Cave, G 1985. (L–O) adult males:
(L) Spider Cave, G1981; (M) Jester Estates
Cave, G 1986; (N) Stovepipe Cave,
G1977; (O) Amber Cave, G 1998; Amber
Cave is a known locality for C. travisae,
although adult males for this species
have never been described. Colours and
symbols used to designate populations as
in Fig. 1.
© 2014 John Wiley & Sons Ltd
CAVE SPIDER SPECIES DELIMITATION 351
et al. 2008). Four adult male specimens were also avail-
able for study and imaged (Fig. 2). Because male speci-
mens have never been described for any of the focal
species (Paquin & Dup�err�e 2009), there is no basis for
formal species-level comparison. It is notable, however,
that all male specimens have essentially identical palp
morphologies.
Mitochondrial analyses
New mitochondrial sequences have been submitted to
GenBank (Appendix 1); GenBank numbers for previ-
ously published sequences are provided in Fig. 3.
Maximum-likelihood and Bayesian analyses (Fig. 3) of
mitochondrial sequences result in generally similar
gene tree topologies, with two minor differences.
First, the C. bandida/C. puentecilla clade (outside the
focal group) differs in position, but this position is
weakly supported in both analyses. Second, RAxML
analyses suggest a different root placement within
C. buwata, with eastern populations (No Rent, Apple
Riata, Weldon, McNeil Bat; Fig. 3) forming a grade
leading to western populations. Both Bayesian and
maximum-likelihood trees recover a well-supported
clade (bootstrap proportion values >70, posterior prob-
ability values >0.95) that includes the focal taxa, and
a primary separation between C. buwata versus
C. reddelli/C. travisae/C. wartoni (hereafter called the
C. travisae complex). The hypothesis of Gertsch (1992)
that C. buwata occurs in the southern Cotterell and
Gallifer Caves (see fig. 131 of Paquin & Dup�err�e
2009) is not supported by mitochondrial or nuclear
evidence (see below).
Mitochondrial species delimitation results (bPTP
and GMYC) are summarized on Fig. 3. Outside the
focal group, there is a high degree of congruence
between bPTP groups and morphologically described
species. Within the focal group, C. buwata forms a
single bPTP group, and the three members of the
C. travisae complex together form a single bPTP
group. This result is consistent with the hypothesis
that members of the C. travisae complex represent a
single species, rather than three distinct species. The
single-threshold GMYC model also recovers C. buwata
and members of the C. travisae complex as separate
individual groups, but seems overly conservative in
that other single GMYC clusters (outside the focal
group) include described species that are clearly mor-
phologically distinct and occupy highly disjunct geo-
graphic distributions (e.g. C. vibora, C. troglobia,
C. hoodensis). The multiple-threshold GMYC model
appears biologically unrealistic, fragmenting most
described species into several individual clusters
(Fig. 3).
Nuclear analyses
Data were collected and phased for eight nuclear genes;
unphased sequences have been submitted to GenBank
(Appendix 1), and alignments of phased data are avail-
able on Dryad (doi:10.5061/dryad.qc3s0). All nuclear
sequences correspond to exons, and PCR-amplified San-
ger data match transcriptome data. Final matrices
include very little missing data (8 nuclear gene matrices
X 34 individuals per matrix – 7 total missing
sequences), and no single specimen is missing data for
more than one nuclear gene (Appendix 1). Gene regions
are similar in aligned length, and all include phyloge-
netic information (Table 1).
Each nuclear gene tree is topologically unique
(Appendix S3, Supporting information), but general pat-
terns are apparent. Members of the C. travisae complex
are recovered as a clade separate from a C. buwata clade
in seven of eight gene trees, and these clades are
strongly supported (maximum-likelihood bootstrap
>70). Within the C. travisae complex, genetic relation-
ships vary from gene to gene, and sequences from indi-
vidual hypothesized species do not cluster together –instead, sequences from different hypothesized species
within the C. travisae complex are intermixed on nuclear
gene trees. Sequences from C. wartoni specimens never
form an exclusive group on nuclear gene trees, but
instead are always intermixed with other members of
the C. travisae complex (Appendix S3).
The POFAD network shows an obvious division
between C. buwata and members of the C. travisae com-
plex (Fig. 4A). Within the C. travisae complex, C. wartoni
specimens are not grouped together on the network.
Specimen (and population) placement on the network
coincides roughly with geographic position (western
populations basal, eastern populations derived; Fig. 4B,
C). A group of populations including C. reddelli from
Cotterell Cave, five eastern ‘undetermined’ cave popu-
lations (Beard Ranch Cave, Jest John Cave, Jester Estates
Cave, Ken Butler Pit, Spider Cave) and two specimens
of C. wartoni cluster together – this genetic association
mirrors that recovered by STRUCTURE K = 3 and Structu-
rama (see below).
STRUCTURE results suggest two genetic partitions, with
K = 2 including the largest DK (461.86) as estimated
using the Evanno method (Fig. 5A). The two genetic
clusters correspond to C. buwata and the C. travisae
complex, consistent with the hypothesis that members
of the C. travisae complex represent a single species.
Although K = 2 (mean LnP(K) = �1165.2) is optimal
under the Evanno method, other K values were consid-
ered, allowing for the possibility that the Evanno
method is underestimating species diversity in this
complex. In particular, a K = 3 (mean LnP
© 2014 John Wiley & Sons Ltd
352 M. HEDIN
AY633093_Cpampa
AY633098_CbrunsiAY633012_013_Cloftini
AY633011_Cbullis
AY633008_Cbullis
AY633006_CbullisAY633007_CbullisAY633010_Cbullis
AY633009_Cbullis
AY633057_CmadlaAY633058_CmadlaAY633059_Cmadla
AY633056_CmadlaAY633061_Cmadla
AY633066_CmadlaAY633063_Cmadla
AY633071_CmadlaAY633062_Cmadla
AY633073_Cmadla
AY633055_CvesperaAY633068_Cmadla_AY633054_Cvespera
AY633069_Cmadla
AY633064_Cmadla
AY633072_CmadlaAY633052_070_Cmadla
AY633089_Cvibora_TemplesofThorAY633090_Cvibora_TemplesofThor
Cvibora_TemplesofThor_TranscriptomeAY633015_Ctroglobia
AY633014_Cmixmaster
AY633025_Choodensis
AY633024_ChoodensisAY633016_017_Ccaliga
AY633021_Choodensis_018_CcaligaAY633026_Choodensis
AY633020_023_ChoodensisAY633019_Choodensis
AY633082_083_Cbandida
AY633077_Cpuentecilla
AY633076_CpuentecillaAY633075_Cpuentecilla
AY633081_CpuentecillaAY633078_079_Cpuentecilla
AY633099_Creddelli_CotterellCave
AY633105_CplacidaAY633104_Cpallida
AY633030_Cvarians
8.0
Ctravisae_NorthRootCave_G1970
Creddelli_CotterellCave_G1960_G1989
StovepipeCave_G1977Cwartoni_PicklePit_G1979_G1980
Ctravisae_McDonaldCave_G1965Ctravisae_McDonaldCave_G1964GeodeCave_G1976
Ctravisae_ToothCave_G1958_G1972_G2014
KenButlerPit_G2009BeardRanchCave_G1984
SpiderCave_G1983SpiderCave_G1981SpiderCave_G1982JestJohnCave_G1988JesterEstatesCave_G1987JesterEstatesCave_G1985
Ctravisae_KretschmarrDPit_G1966Ctravisae_AmberCave_G1998_G1999
TwoTrunksCave_G1967_G1968
AY633084_085_Cbuwata_TestudoTube
Cbuwata_NoRentCave_G1994Cbuwata_AppleRiataTrace_G1997
Cbuwata_NoRentCave_G1995
Cbuwata_BrokenArrowCave_G1961Cbuwata_MarigoldCave_G2005_G2006
Cbuwata_NoRentCave_G1957_G1992
Cbuwata_McNeilBatCave_G1996
Cbuwata_LakelineCave_G2001Cbuwata_LakelineCave_G2000
Cbuwata_BabeCave_G2010_G2011Cbuwata_ButtercupCreekCave_G2007Cbuwata_DiesRanchTreasureCave_G2002
0.93/87
0.93/87
0.96 /79
0.99/99
0.84 /83
0.99/100
0.80/710.99/97
0.99/98
0.99 /93
0.99/90
0.98/82
0.99/100
0.99/98
RAxML
GMYC_Single
C. buwata
C. travisae Complex
bPTP
1.01.01.0
0.94
1.0
0.80
0.81
1.0
0.91
0.83
0.95
1.01.0
1.0
1.0
GMYC_Multiple
Fig. 3 Bayesian inference mitochondrial gene tree with posterior probability and RAxML bootstrap values (for primary lineages),
plus bPTP and GMYC results. Arrow denotes alternative RAxML phylogenetic placement for Cicurina puentecilla/C. bandida clade.
Redundant haplotypes not shown include CbuwataTestudoTube = CbuwataBroken ArrowCave_G1959_G1962 + CbuwataMarigold-
Cave_G2004, CbuwataWeldonCave_G1992 = CbuwataNoRentCave_G1994, TwoTrunksCave_G1967_G1968 = GalliferCave_G1975. Col-
ours and symbols used to designate focal populations as in Fig. 1.
© 2014 John Wiley & Sons Ltd
CAVE SPIDER SPECIES DELIMITATION 353
(K) = �1046.4) hypothesis was considered – this genetic
clustering implies three separate genetic groups corre-
sponding to C. buwata, C. travisae (plus western unde-
termined) and C. reddelli + plus eastern undetermined +C. wartoni (Fig. 5B). Structurama analyses, using a no
admixture model with multiple prior alpha values, also
support this same K = 3 hypothesis (Table 2). Analyses
under certain alpha values include a low posterior
probability for K = 4 (e.g. alpha 1,1, pp = 0.21; Table 2),
but the fourth genetic cluster does not consistently
include the same set of individuals.
Bayesian phylogenetics and phylogeography analyses
provide strong support for a node separating C. buwata
versus the C. travisae complex, but fail to support fur-
ther subdivisions within the C. travisae complex (Fig. 6).
Again, these results are consistent with the hypothesis
that members of the C. travisae complex represent a sin-
gle species.
Discussion
The Cicurina travisae species complex
The approach to species delimitation used here relied
upon multiple lines of independent evidence, included
analyses with different analytical assumptions and inte-
grated results via observed congruence. When congru-
ence was not observed (e.g. K = 2 versus K = 3, K = 2
STRUCTURE versus STRUCTURAMA), additional analyses were
conducted to explore this incongruence. Taken together,
multiple lines of evidence support a clear distinction
between northern C. buwata and a southern C. travisae
complex that currently includes three described species.
As shown here and elsewhere (Paquin & Dup�err�e
2009), different members of the C. travisae complex have
extremely similar female epigynal morphologies
(Fig. 2). Male specimens are not available from all cave
populations in the complex, but those available
(spanning much of the range of the C. travisae complex)
have essentially indistinguishable pedipalps (Fig. 2).
Although mitochondrial data indicate high levels of
female-based genetic structuring (e.g. exclusive mito-
chondrial clades that correspond to single-cave popula-
tions), single-locus species delimitation analyses (bPTP,
single-threshold GMYC) recover members of this com-
plex as a single lineage. This result provides a general
illustration as to why simple phylogenetic patterns
observed on single gene trees (e.g. genealogical exclu-
sivity for single caves) should not be overinterpreted as
evidence for species status.
The multigenic nuclear perspective shows that
sequences from C. wartoni, C. travisae and C. reddelli are
intermixed on nuclear gene trees. There is some signal
for east to west geographic structuring in the nuclear-
only POFAD network, mirrored in K = 3 STRUCTURE and
STRUCTURAMA analyses, as might be expected in a system
where populations are restricted to caves. However, this
geographic structure is not supported as species-level
divergence in BPP validation analyses, which instead
support the C. travisae complex as a single genetic line-
age. This result is consistent with the K = 2 STRUCTURE
results, POFAD analyses and the above summarized
morphological and mitochondrial results.
It is important for researchers to complete taxonomic
actions implied by robust results, be this species synon-
ymy or new species description (e.g., Fujita & Leach�e
2010; Satler et al. 2013; Derkarabetian & Hedin 2014). To
this end, the species C. wartoni, C. travisae and C. reddel-
li are formally synonymized below. The species concept
used here corresponds to a general lineage concept (de
Queiroz 2007), with species viewed as separately evolv-
ing metapopulation lineages, thus allowing for some
internal genetic structuring (i.e. connected subpopula-
tions). Multiple lines of congruent evidence were used
to operationally recognize this metapopulation lineage,
with both population genetic and multispecies coales-
cent evidence from the nuclear genome emphasized.
Family DICTYNIDAE O. Pickard-Cambridge 1871
[urn:lsid:nmbe.ch:spiderfam:0042]
Genus Cicurina Menge 1871 [urn:lsid:nmbe.ch:spider-
gen:01972]
Subgenus Cicurella Chamberlin & Ivie 1940
Cicurina (Cicurella) travisae Gertsch 1992 [urn:lsid:
nmbe.ch:spidersp:022224]; FIGS 1-6
Cicurina travisae Gertsch 1992: 101, figs 63–70.Cicurina travisae Paquin et al. 2008: 147, fig. 5b.
Cicurina travisae Paquin & Dup�err�e 2009: 47, figs 106-
107, fig. 131.
Cicurina reddelli Gertsch 1992: 105, figs 77–78; new
synonymy.
Cicurina reddelli Paquin & Dup�err�e 2009: 40, figs 86–87, 131.
Table 1 Nuclear gene data summary
Primer
Names
Aligned
length
No.
Ingroup
Sequences
Parsimony
Informative
Sites
Nucleotide
Diversity
B1_B11B12 662 47 33 0.018
B1_C9C10 706 39 11 0.006
B1_E1E2 667 45 12 0.007
B1_G5G6 827 35 14 0.008
B1_G11G12 840 44 14 0.005
B2_D2D3 560 48 20 0.012
B2_F10F11 677 47 25 0.011
B2_H6H7 647 53 22 0.007
Diversity statistics calculated using MEGA 6.06 (Tamura et al.
2013), ingroup data only.
© 2014 John Wiley & Sons Ltd
354 M. HEDIN
Cicurina wartoni Gertsch 1992: 101, figs. 75–76; new
synonymy.
Cicurina wartoni Paquin & Dup�err�e 2009:55, figs. 122–123, 131.
Taxonomic type II error and the ‘burden of proof’
Like all species hypotheses, the synonymy hypothesis
proposed above remains falsifiable. It is possible that
the combination of species criterion, data analyses and
data type utilized has failed to detect a species-level dif-
ference that actually exists, resulting in a so-called taxo-
nomic type II error (Padial et al. 2010; Miralles &
Vences 2013; McCormack & Maley 2015). For example,
until diversity biologists employ complete genomic data
for large, theoretically sufficient samples, it will always
be possible to speculate about the proverbial data nee-
dle (or needles) in the haystack. Although such argu-
ments are frankly difficult to counter, several are made
here for the case of Cicurina travisae.
First, there are clear indications that many genetic
species delimitation approaches are inflationist in natu-
rally fragmented systems – if population structure
exists, there is an apparent analytical bias towards
C_travisae Complex
0.1
CTRAVISAE_TOOTHCAVE_G1958
STOVEPIPECAVE_G1977
GALLIFERCAVE_G1975
CTRAVISAE_MCDONALDCAVE_G1964
CTRAVISAE_MCDONALDCAVE_G1965
CTRAVISAE_AMBERCAVE_G1998
GEODECAVE_G1976CWARTONI_PICKLEPIT_G1979
CREDDELLI_COTTERELLCAVE_G1960
CREDDELLI_COTTERELLCAVE_G1989
JESTJOHNCAVE_G1988
KENBUTLERPIT_G2009SPIDERCAVE_G1982
SPIDERCAVE_G1981
JESTERESTATESCAVE_G1985
BEARDRANCHCAVE_G1984
CWARTONI_PICKLEPIT_G1978CWARTONI_PICKLEPIT_G1980
CTRAVISAE_NORTHROOTCAVE_G1970TWOTRUNKSCAVE_G1968
CTRAVISAE_TOOTHCAVE_G2014
CTRAVISAE_KRETSCHMARRDOUBLEPIT_G1966 JESTERESTATES
CAVE_G1986
C_buwata(A)
(B)
(C)McDonald
Kretschmarr
Amber
TwoTrunks
Tooth
NorthRoot
Gallifer
Geode
Stovepipe
PicklePit
Spider
BeardRanch
KenButlerPit
JesterEstates
JestJohn
Cotterell
W
R
T
TT
TT
Fig. 4 NeighborNet network recon-
structed using standardized POFAD
nuclear distances. (A) Entire network,
showing primary division between Cic-
urina buwata and the C. travisae complex.
(B) POFAD network for C. travisae com-
plex. Colours and symbols as in Fig. 1.
(C) Map inset, colours and symbols used
to designate populations as in Fig. 1.
© 2014 John Wiley & Sons Ltd
CAVE SPIDER SPECIES DELIMITATION 355
oversplitting (type I error; e.g. Hey 2009; Camargo &
Sites 2013). For example, both single-threshold GMYC
(Keith & Hedin 2012; Satler et al. 2013; Hamilton et al.
2014) and BPP (Barley et al. 2013; Carstens & Satler
2013; McKay et al. 2013; Miralles & Vences 2013; Satler
et al. 2013) have been suggested to oversplit in naturally
fragmented systems. These methods, however, support
a single species hypothesis for C. travisae. Second, for
nuclear analyses that support a more finely subdivided
taxonomy (e.g. STRUCTURE K = 3, Structurama), the smal-
ler units correspond to arrays of multiple adjacent pop-
ulations (e.g. eastern versus western cave populations,
Figs 4 and 5). None of the nuclear analyses conducted
support single cave populations as species. Third,
although genomic-scale data were not technically used
here (e.g. genome-wide SNP data, Leach�e et al. 2014),
considerable resources were devoted towards the devel-
opment of Cicurina-specific ‘rapidly evolving’ nuclear
markers. If speciation in Cicurina proceeds via an ‘isola-
tion plus drift’ model as expected, then nucleotide dif-
ferences observed in a subset of the nuclear genome
should reflect overall genomic divergence (Feder et al.
2012). If Cicurina speciation instead corresponds to a
‘genomic islands of divergence’ model, then speciation
would be potentially missed if selectively important loci
were not sampled, but this model of speciation is con-
sidered unlikely here. Finally, populations from Pickle
Pit (type for ‘C. wartoni’) and Cotterell Cave (type for
‘C. reddelli’) correspond to ‘evolutionary significant
units’, if monophyly for mitochondrial alleles is used as
the criteria for such units (Moritz 1994). As such, these
cave populations retain conservation importance, just
not as distinct species.
Genomics-scale species delimitation for all taxa on
Earth seems unrealistic and prohibitively expensive,
particularly given the enormity of the current task faced
by diversity biologists (i.e. most existing species
hypotheses never tested, millions of species remain
undescribed). However, for ‘high-stakes’ research on
conservation-relevant taxa (e.g., hypothesized single site
endemics), we might expect such data sets to become
the norm, or even the expectation (McCormack & Ma-
ley 2015). Such data sets, however, do not imply an
easy resolution to the species delimitation problem, as
both the genomic architecture of speciation and popula-
tion subdivision continue to present analytical chal-
lenges for most currently available methods.
Table 2 Structurama results
K Alpha 1,1 Alpha 1,5 Alpha 1,10 Alpha 0.1,1 Alpha 0.1,5 Alpha 0.1,10
Marginal likelihood �516.32 �517.41 �518.98 �517.31 �517.31 �516.67
Model Prob (# pops) 1 0 0 0 0 0 0
2 0 0 0 0 0 0
3 0.76 0.88 0.91 0.81 0.91 0.94
4 0.21 0.12 0.09 0.18 0.09 0.06
5 0.02 0.01 0 0.01 0 0
Appl
eRia
taBa
beCa
ve
Brok
enAr
row
Cave
Butte
rcup
Cree
kCav
eLa
kelin
eCav
eM
arig
oldC
ave
McN
eilB
atCa
ve
NoRen
tCav
e
Test
udoT
ube
Wel
donC
ave
Bear
dRan
chCa
veCo
ttere
llCav
e
Jest
John
Cave
Jest
erEs
tate
sCav
eKe
nBut
lerP
itSp
ider
Cave
Ambe
rCav
e
Gallif
erCa
veGeo
deCa
ve
Kret
schm
arrD
oubl
ePit
McD
onal
dCav
e
North
Root
Cave
Stov
epip
eCav
eTo
othC
ave
TwoT
runk
sCav
ePi
ckle
Pit
Appl
eRia
taBa
beCa
ve
Brok
enAr
row
Cave
Butte
rcup
Cree
kCav
eLa
kelin
eCav
eM
arig
oldC
ave
McN
eilB
atCa
ve
NoRen
tCav
e
Test
udoT
ube
Wel
donC
ave
Bear
dRan
chCa
veCo
ttere
llCav
e
Jest
John
Cave
Jest
erEs
tate
sCav
eKe
nBut
lerP
itSp
ider
Cave
Ambe
rCav
e
Gallif
erCa
veGeo
deCa
ve
Kret
schm
arrD
oubl
ePit
McD
onal
dCav
e
North
Root
Cave
Stov
epip
eCav
eTo
othC
ave
TwoT
runk
sCav
ePi
ckle
Pit
K = 2, mean LnP(K) = –1165.2
C buwata C travisae complex
C buwata
C. wartoni, C reddelli + “EASTERN undetermined”
C wartoni
G19
80
K = 3, mean LnP(K) = –1046.4
(A)
(B)
Fig. 5 STRUCTURE graphics (resulting from DISTRUCT) for K = 2
(A) and K = 3 (B). Each column represents a specimen,
grouped by cave population of origin. Different colours repre-
sent different genetic clusters (K); estimated membership coef-
ficients are proportional to bar colour height.
© 2014 John Wiley & Sons Ltd
356 M. HEDIN
The importance or robust species hypotheses inconservation biology
This study has relevance to systematic studies of other
Texas cave Cicurina, including species currently listed
as US Federally Endangered (Longacre 2000; U.S. Fish
& Wildlife Service 2008). First, the mitochondrial delim-
itation results presented suggest the possible synonymy
of certain taxa (e.g. C. madla and C. vespera, C. mixmas-
ter/C. caliga/C. hoodensis, etc.; Fig. 3). This issue of syn-
onymy, and the idea that many Texas cave Cicurina
may not be as highly endemic as indicated by the cur-
rent taxonomy, has been mentioned elsewhere multiple
times (Paquin & Hedin 2004; Paquin et al. 2008). Indeed,
Paquin & Dup�err�e (2009) called species synonymy a
‘generalized problem’ in Texas cave Cicurina. The revi-
sionary work of Ledford et al. (2012) on Texas cave lep-
tonetid spiders suggests a similar pattern, with a new
integrative taxonomy indicating fewer, more wide-
spread species than a prior taxonomy, suggesting a
higher number of more narrowly endemic species.
More generally, the integrative framework applied here
(or something akin to this framework, e.g. Satler et al.
2013; Derkarabetian & Hedin 2014; etc.) should be
applied elsewhere in Texas cave Cicurina. Of the four
federally listed taxa (C. baronia, C. madla, C. venii, C. ves-
pera), three species are single site endemics, two species
hypotheses are based on single female specimens
(Paquin & Dup�err�e 2009), and only one species has
been studied using molecular evidence (mitochondrial
only, Paquin & Hedin 2004). A major hurdle in the
study of cave Cicurina species limits has been specimen
rarity and availability – new sequencing technologies
potentially allow for the collection of massive quantities
of DNA sequence data from museum specimens (Bi
et al. 2013), perhaps overcoming this hurdle.
The Draft Recovery Plan for listed Cicurina taxa (U.S.
Fish & Wildlife Service 2008) does not include ‘conduct
integrative taxonomy’ as a suggested research need.
Although the existing species delimitations may be
accurate, hypotheses with such obvious conservation,
political and economic importance need to be more data
rich [arguments echoed by Paquin & Dup�err�e (2009)],
and conducting modern integrative taxonomy must be
a fundamental and primary research need. If the con-
servation action is taxon focused, then this same argu-
ment applies to all taxa where the existing taxonomy is
arguably weak, whether these taxa are new candidates
for conservation action, or taxa that are already pro-
tected. Taxon-focused conservation biology is expensive
(e.g. McCarthy et al. 2012), financial resources are lim-
ited, and downstream resource allocations flow from
the most fundamental question – is the taxonomic unit
evolutionarily distinct? This question needs to be
answered first.
buwata
1.0
travisae
wartoni
reddelli
0.37/0.35
0.53/0.53
0.81/ 0.82 0.50/
0.51
(A) (B)
(C) (D)
Theta prior (2, 100)
Theta prior (2, 100)
Theta prior (1, 10)
Theta prior (1, 10)
buwata
1.0
travisae
wartoni
reddelli
buwata
1.0
travisae
wartoni
reddelli
buwata
1.0
travisae
wartoni
reddelli
Fig. 6 Summary of Bayesian phylogenet-
ics and phylogeography analyses. (A and
B) Undetermined cave populations not
included in analysis, guide tree = Cicuri-
na buwata sister to (C. travisae, (C. reddelli
+ C. wartoni)) – A) G (alpha = 2, beta =100) theta prior, (B) G (alpha = 1, beta =10) theta prior. (C and D) Western ‘unde-
termined’ cave populations allocated to
C. travisae, eastern ‘undetermined’ popu-
lations (Spider Cave, Ken Butler Pit, Jes-
ter Estates Cave, Jest John Cave, Beard
Ranch Cave) allocated to C. reddelli.
Guide tree = C. buwata sister to (C. travi-
sae, (C. reddelli + C. wartoni)) – (C) G
(alpha = 2, beta = 100) theta prior, (D) G
(alpha = 1, beta = 10) theta prior.
© 2014 John Wiley & Sons Ltd
CAVE SPIDER SPECIES DELIMITATION 357
Acknowledgements
This research was funded by the US Fish & Wildlife Service,
contract #F13PX00770. Cyndee Watson provided support for
this project from inception to completion and deserves special
thanks. Specimens were obtained from Mark Sanders, Todd
Bayless, P. Fushille and Jet Larsen. Kristen Emata provided
expert laboratory assistance, with help from David Zezoff.
James Starrett extracted RNA, while Shahan Derkarabetian and
Dave Carlson assisted in the transcriptome assembly process.
Members of the Hedin laboratory, Jordan Satler, Bryan Car-
stens and two anonymous reviewers provided comments that
helped to improve the manuscript.
References
Aylor DL, Price EW, Carbone I (2006) SNAP: combine and map
modules for multilocus population genetic analysis. Bioinfor-
matics, 22, 1399–1401.Barley AJ, White J, Diesmos AC, Brown RM (2013) The chal-
lenge of species delimitation at the extremes: Diversification
without morphological change in Philippine sun skinks. Evo-
lution, 67, 3556–3572.Bendik NF, Meik JM, Gluesenkampo AG, Roelke CE, Chippin-
dale PT (2013) Biogeography, phylogeny, and morphological
evolution of central Texas cave and spring salamanders.
BMC Evolutionary Biology, 13, 201.
Bi K, Linderoth T, Vanderpool D, Good JM, Nielsen R, Moritz
C (2013) Unlocking the vault: next-generation museum pop-
ulation genomics. Molecular Ecology, 22, 6018–6032.Camargo A, Sites J Jr (2013) Species delimitation: a decade after
the renaissance. In: The Species Problem – Ongoing Issues (ed.
Pavlinov I), pp. 225–247. InTech, New York, New York.
Carstens BC, Satler JD (2013) The carnivorous plant described
as Sarracenia alata contains two cryptic species. Biological Jour-
nal of the Linnean Society, 4, 737–746.Carstens BC, Pelletier TA, Reid NM, Satler JD (2013) How to
fail at species delimitation. Molecular Ecology, 22, 4369–4383.Chiari Y, van der Meijden A, Mucedda M, Lourenco JM, Ho-
chkirch A, Veith M (2012) Phylogeography of Sardinian cave
salamanders (genus Hydromantes) is mainly determined by
geomorphology. PLoS ONE, 7, e32332.
Derkarabetian S, Hedin M (2014) Integrative taxonomy and
species delimitation in harvestmen: a revision of the western
North American genus Sclerobunus (Opiliones: Laniatores:
Travunioidea). PLoS ONE, 9, e104982.
Derkarabetian S, Steinmann DB, Hedin M (2010) Repeated and
time-correlated morphological convergence in cave-dwelling
harvestmen (Opiliones, Laniatores) from montane western
North America. PLoS ONE, 5, e10388.
Drummond AJ, Ho SYW, Phillips MJ, Rambaut A (2006)
Relaxed phylogenetics and dating with confidence. PLoS
Biology, 4, e88. doi:10.1371/journal.pbio.0040088.
Drummond AJ, Suchard MA, Xie D, Rambaut A (2012) Bayes-
ian phylogenetics with BEAUti and the BEAST 1.7. Molecular
Biology & Evolution, 29, 1969–1973.Earl DA, vonHoldt BM (2012) STRUCTURE HARVESTER: a website
and program for visualizing STRUCTURE output and imple-
menting the Evanno method. Conservation Genetic Resources,
4, 359–361.
Evanno G, Regnaut S, Goudet J (2005) Detecting the number of
clusters of individuals using the software STRUCTURE: a simu-
lation study. Molecular Ecology, 14, 2611–2620.Feder JL, Egan SP, Nosil P (2012) The genomic of speciation-
with-gene-flow. Trends in Genetics, 28, 342–349.Flot JF (2010) SEQPHASE: a web tool for interconverting PHASE
input/output files and FASTA sequence alignments. Molecu-
lar Ecology Resources, 10, 162–166.Fujisawa T, Barraclough TG (2013) Delimiting species using
single-locus data and the Generalized Mixed Yule Coalescent
approach: a revised method and evaluation on simulated
data sets. Systematic Biology, 62, 707–724.Fujita MK, Leach�e AD (2010) A coalescent perspective on
delimiting and naming species: a reply to Bauer et al. Pro-
ceedings of the Royal Society B: Biological Sciences, 278, 493–495.
Fujita MK, Leach�e AD, Burbrink FT, McGuire JA, Moritz C
(2012) Coalescent-based species delimitation in an integrative
taxonomy. Trends in Ecology & Evolution, 27, 480–488.Gertsch WJ (1992) Distribution patterns and speciation in
North American cave spiders with a list of the troglobites
and revision of the cicurinas of the subgenus Cicurella.. Texas
Memorial Museum Speleological Monographs, 3. Studies on the
Endogean fauna of North America, 2, 75–122.Hamilton CA, Hendrixson BE, Brewer MS, Bond JE (2014) An
evaluation of sampling effects on multiple DNA barcoding
methods leads to an integrative approach for delimiting spe-
cies: a case study of the North American tarantula genus
Aphonopelma (Araneae, Mygalomorphae, Theraphosidae).
Molecular Phylogenetics and Evolution, 71, 79–93.Hedin M (1997) Molecular phylogenetics at the population/
species interface in cave spiders of the southern Appala-
chians (Araneae: Nesticidae: Nesticus). Molecular Biology and
Evolution, 14, 309–324.Hey J (2009) On the arbitrary identification of real species. In:
Speciation and Patterns of Biodiversity (eds Butlin RK, Bridle
JR, Schlueter D), pp. 15–28. Cambridge University Press,
New York, New York.
Huelsenbeck JP, Andolfatto P (2007) Inference of population
structure under a Dirichlet process model. Genetics, 175,
1787–1802.Huelsenbeck JP, Huelsenbeck ET, Andolfatto P (2011) Structu-
rama: Bayesian inference of population structure. Evolution-
ary Bioinformatics Online, 7, 55–59.Huson DH, Bryant D (2006) Application of phylogenetic net-
works in evolutionary studies. Molecular Biology and Evolu-
tion, 23, 254–267.Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster matching
and permutation program for dealing with label switching
and multimodality in analysis of population structure. Bioin-
formatics, 23, 1801–1806.Joly S, Bruneau A (2006) Incorporating allelic variation for
reconstructing the evolutionary history of organisms from
multiple genes: an example from Rosa in North America.
Systematic Biology, 55, 623–636.Juan C, Guzik MT, Jaume D, Cooper SJB (2010) Evolution in
caves: Darwin’s ‘wrecks of ancient life’ in the molecular era.
Molecular Ecology, 19, 3865–3880.Keith RM, Hedin M (2012) Extreme mitochondrial population
subdivision in southern Appalachian paleoendemic spiders
© 2014 John Wiley & Sons Ltd
358 M. HEDIN
(Araneae, Hypochilidae, Hypochilus), with implications for
species delimitation. Journal of Arachnology, 40, 167–181.Lanfear R, Calcott B, Ho SYW, Guindon S (2012) PARTITIONFIND-
ER: combined selection of partitioning schemes and substitu-
tion models for phylogenetic analyses. Molecular Biology and
Evolution, 29, 1695–1701.Leach�e AD, Fujita MK (2010) Bayesian species delimitation
in West African forest geckos (Hemidactylus fasciatus). Pro-
ceedings of the Royal Society B: Biological Sciences, 277, 3071–3077.
Leach�e AD, Fujita MK, Minin VN, Bouckaert RR (2014) Species
delimitation using genome-wide SNP data. Systematic Biol-
ogy, 63, 534–542.Ledford J, Paquin P, Cokendolpher J, Campbell J, Griswold C
(2012) Systematics, conservation and morphology of the spi-
der genus Tayshaneta (Araneae, Leptonetidae) in central
Texas caves. ZooKeys, 167, 1–102.Lohse K (2009) Can mtDNA barcodes be used to delimit spe-
cies? A response to Pons et al. (2006) Systematic Biology, 58,
439–442.Longacre C (2000) Department of the Interior, Fish and Wild-
life Service, 50 CFR Part 17, RIN 1018-AF33. Final rule to list
nine Bexar County, Texas invertebrate species as endan-
gered. Federal Register, 65, 81419–81433.Malaney JL, Cook JA (2013) Using biogeographical history to
inform conservation: the case of Preble’s meadow jumping
mouse. Molecular Ecology, 22, 6000–6017.McCarthy DP, Donald PF, Scharlemann JP et al. (2012) Finan-
cial costs of meeting global biodiversity conservation targets:
current spending and unmet needs. Science, 338, 946–949.McCormack JE, Maley JM (2015) Interpreting negative results
with conservation implications: another look at distinctness
of coastal California Gnatcatchers. The Auk, 132. doi:
10.1642/AUK-14-184.1.
McKay BD, Mays HL Jr, Wu Y et al. (2013) An empirical com-
parison of character-based and coalescent-based approaches
to species delimitation in a young avian complex. Molecular
Ecology, 22, 4943–4957.Miralles A, Vences M (2013) New metrics for comparison of
taxonomies reveal striking discrepancies among species
delimitation methods in Madascincus lizards. PLoS ONE, 8,
e68242.
Mora C, Tittensor DP, Adl S, Simpson AGB, Worm B (2011)
How many species are there on Earth and in the ocean?
PLoS Biology, 9, e1001127. doi:10.1371/journal.pbio.1001127.
Moritz C (1994) Defining “evolutionarily significant units” for
conservation. Trends in Ecology & Evolution, 9, 373–375.Niemiller ML, Near TJ, Fitzpatrick BM (2012) Delimiting spe-
cies using multilocus data: diagnosing cryptic diversity in
the southern cavefish, Typhlichthys subterraneus (Teleostei:
Amblyopsidae). Evolution, 66, 846–866.Niemiller ML, Graening GO, Fenolio DB et al. (2013) Doomed
before they are described? The need for conservation assess-
ments of cryptic species complexes using an amblyopsid
cavefish (Amblyopsidae, Typhlichthys) as a case study. Biodi-
versity & Conservation, 22, 1799–1820.O’Meara BC (2010) New heuristic methods for joint species
delimitation and species tree inference. Systematic Biology, 59,
59–73.Padial JM, Miralles A, De la Riva I, Vences M (2010) The inte-
grative future of taxonomy. Frontiers in Zoology, 7, 16.
Pante E, Schoelinck C, Puillandre N (2015) From integrative
taxonomy to species description: one step beyond. Systematic
Biology, 64, 152–160.Paquin P, Dup�err�e N (2009) A first step towards the revision
of Cicurina: redescription of type specimens of 60 troglobitic
species of the subgenus Cicurella (Araneae: Dictynidae), and
a first visual assessment of their distribution. Zootaxa, 2002,
1–67.Paquin P, Hedin M (2004) The power and perils of “molecular
taxonomy”: a case study of eyeless and endangered Cicurina
(Araneae: Dictynidae) from Texas caves. Molecular Ecology,
13, 3239–3255.Paquin P, Duperre N, Cokendolpher JC, White K, Hedin M
(2008) The fundamental importance of taxonomy in conser-
vation biology: the case of the eyeless Cicurina bandida (Ara-
neae: Dictynidae) of central Texas, including new synonyms
and the description of the male of the species. Invertebrate
Systematics, 22, 1–11.Pella J, Masuda M (2006) The Gibbs and split-merge sampler
for population mixture analysis from genetic data with
incomplete baselines. Canadian Journal Fish Aquatic Sciences,
63, 576–596.Pons J, Barraclough TG, Gomez-Zurita J et al. (2006) Sequence-
based species delimitation for DNA taxonomy of unde-
scribed insects. Systematic Biology, 55, 595–609.Price EW, Carbone I (2005) SNAP. Workbench management
tool for evolutionary population genetic analysis. Bioinformat-
ics Applications Note, 21, 402–404.Pritchard JK, Stephens M, Donnelly P (2000) Inference of popu-
lation structure using multilocus genotype data. Genetics,
155, 945–959.Pritchard JK, Wen X, Falush D (2010) Documentation for Struc-
ture software: Version 2.3. Available from http://
pritch.bsd.uchicago.edu/structure.html. 1-38.
de Queiroz K (2007) Species concepts and species delimitation.
Systematic Biology, 56, 879–886.Rannala B, Yang Z (2013) Improved reversible jump algo-
rithms for Bayesian species delimitation. Genetics, 194, 245–253.
Rosenberg NA (2004) DISTRUCT: a program for the graphical dis-
play of population structure. Molecular Ecology Notes, 4, 137–138.
Satler JD, Carstens BC, Hedin M (2013) Multilocus species
delimitation in a complex of morphologically conserved trap-
door spiders (Mygalomorphae, Antrodiaetidae, Aliatypus).
Systematic Biology, 62, 805–823.Silvestro D, Michalak I (2012) RAXMLGUI: a graphical front-end
for RAxML. Organisms Diversity & Evolution, 12, 335–337.
Stamatakis A (2006) RAxML-VI-HPC: maximum likelihood-
based phylogenetic analyses with thousands of taxa and
mixed models. Bioinformatics, 22, 2688–2690.Stamatakis A (2014) RAXML version 8: a tool for phylogenetic
analysis and post-analysis of large phylogenies. Bioinformat-
ics, 30, 1312–1313.Stephens M, Donnelly P (2003) A comparison of Bayesian
methods for haplotype reconstruction from population geno-
type data. American Journal of Human Genetics, 73, 1162–1169.Stephens M, Smith N, Donnelly P (2001) A new statistical
method for haplotype reconstruction from population data.
American Journal of Human Genetics, 68, 978–989.
© 2014 John Wiley & Sons Ltd
CAVE SPIDER SPECIES DELIMITATION 359
Swofford DL (2002) PAUP*: Phylogenetic Analysis Using Parsi-
mony (*and Other Methods). Version 4. Sinauer Associates,
Sunderland, Massachusetts.
Tamura K, Stecher G, Peterson D, Filipski A, Kumar S (2013)
MEGA6: molecular evolutionary genetics analysis version 6.0.
Molecular Biology and Evolution, 30, 2725–2729.U.S. Fish and Wildlife Service (2008) Bexar County Karst Inverte-
brates Draft Recovery Plan. U.S. Fish and Wildlife Service,
Albuquerque, New Mexico.
U.S. Fish and Wildlife Service (2010) Department of the Inte-
rior, Fish and Wildlife Service, 50 CFR Part 17, Docket No.
FWS–R9–ES–2010–0065; MO–9221050083–B2. Endangered
and threatened wildlife and plants’ review of native species
that are candidates for listing as endangered or threatened;
annual notice of findings on resubmitted petitions; annual
description of progress on listing actions. Federal Register, 75,
69222–69294.Weisrock DW, Rasoloarison RM, Fiorentino I et al. (2010)
Delimiting species without nuclear monophyly in Madagas-
car’s mouse lemurs. PLoS ONE, 5, e9883. doi:10.1371/jour-
nal.pone.0009883.
Wiens JJ (2007) Species delimitation: new approaches for dis-
covering diversity. Systematic Biology, 56, 875–878.Yang Z, Rannala B (2010) Bayesian species delimitation using
multilocus sequence data. Proceedings of the National Academy
of Sciences, USA, 107, 9264–9269.Zhang J, Kapli P, Pavlidis P, Stamatakis A (2013) A general
species delimitation method with applications to phyloge-
netic placements. Bioinformatics, 29, 2869–2876.
M.H. implemented the study design, directed data col-
lection, conducted analyses and wrote the manuscript.
Data accessibility
DNA sequences: Genbank accessions KP221938-
KP222248.
DNA sequence alignments, RAxML and BEAST.tre
files, morphology JPG files: Dryad doi:10.5061/
dryad.qc3s0.
Illumina data submitted to NCBI Short Read Archive
– C. vibora (SRX652499), C. travisae (SRX761208).
PCR conditions and nuclear gene trees uploaded as
online Supporting Information.
Supporting information
Additional supporting information may be found in the online ver-
sion of this article.
Appendix S1 Development of nuclear markers.
Appendix S2 Nuclear primers, PCR conditions, transcript
annotations.
Appendix S3 RAxML maximum likelihood nuclear gene trees.
© 2014 John Wiley & Sons Ltd
360 M. HEDIN
Appendix
1Vouch
ernumber,sp
eciesiden
tity
(original
taxonomy),sex,locationan
dGen
Ban
kinform
ation(unphased
sequen
ces)
LAB#
SPECIES
SEX
LOCALITY
COI
B1_G5G
6B1_B11B12
B2_D2D
3B1_C9C
10B2_F10F11
B2_H6H
7B1_E1E
2B1_G11G12
G1957
Cbuw
INoRen
tCav
eKP221938-40
KP221984
KP222016
KP222050
KP222084
KP222115
KP222148
KP222182
KP222216
G1992
Cbuw
IWeldonCav
eKP221941
KP221985
KP222017
KP222051
KP222085
KP222116
KP222149
KP222183
KP222217
G1996
Cbuw
IMcN
eilBat
Cav
eKP221942
KP221986
KP222018
KP222052
KP222086
KP222117
KP222150
KP222184
KP222218
G1997
Cbuw
FApple
Riata
Trace
KP221943
KP221987
KP222019
KP222053
KP222087
KP222118
KP222151
KP222185
KP222219
G2000
Cbuw
ILak
elineCav
eKP221944
KP221988
KP222020
KP222054
KP222088
KP222119
KP222152
KP222186
KP222220
G2001
Cbuw
ILak
elineCav
eKP221945
KP221989
KP222021
KP222055
KP222089
KP222120
KP222153
KP222187
KP222221
G1959
Cbuw
FBroken
Arrow
Cav
eKP221946-48
KP221990
KP222022
KP222056
KP222090
—KP222154
KP222188
KP222222
G1963
Cbuw
ITestudoTube
—KP221991
KP222023
KP222057
KP222091
KP222121
KP222155
KP222189
KP222223
G2004
Cbuw
IMarigold
Cav
eKP221949-51
KP221992
KP222024
KP222058
KP222092
KP222122
KP222156
KP222190
KP222224
G2007
Cbuw
FButtercu
pCreek
Cav
eKP221952
KP221993
KP222025
KP222059
KP222093
KP222123
KP222157
KP222191
KP222225
G2011
Cbuw
FBab
eCav
eKP221953,54
KP221994
KP222026
KP222060
KP222094
KP222124
KP222158
KP222192
KP222226
G2002
Cbuw
IDiesRan
chTreasure
Cav
eKP221955
——
——
——
——
G1964
Ctrav
IMcD
onaldCav
eKP221956
KP221995
KP222027
KP222061
KP222095
KP222125
KP222159
KP222193
KP222227
G1965
Ctrav
IMcD
onaldCav
eKP221957
KP221996
KP222028
KP222062
KP222096
KP222126
KP222160
KP222194
KP222228
G1966
Ctrav
FKretsch
marrDouble
Pit
KP221958
KP221997
KP222029
KP222063
KP222097
KP222127
KP222161
KP222195
KP2222229
G1968
APU
ITwoTrunksCav
eKP221959-61
KP221998
KP222030
KP222064
KP222098
KP222128
KP222162
KP222196
KP222230
G1970
Ctrav
FNorthRootCav
eKP221962
KP221999
KP222031
KP222065
KP222099
KP222129
KP222163
KP222197
KP222231
G1958
Ctrav
FTooth
Cav
eKP221963,64
KP222000
KP222032
KP222066
KP222100
KP222130
KP222164
KP222198
KP222232
G2014
Ctrav
FTooth
Cav
eKP221965
—KP222033
KP222067
KP222101
KP222131
KP222165
KP222199
KP222233
G1975
APU
IGalliferCav
eKP221966
KP222001
KP222034
KP222068
KP222102
KP222132
KP222166
KP222200
KP222234
G1976
APU
IGeo
deCav
eKP221967
KP222002
KP222035
KP222069
KP222103
KP222133
KP222167
KP222201
KP222235
G1998
Ctrav
MAmber
Cav
eKP221968,69
KP222003
KP222036
KP222070
KP222104
KP222134
KP222168
KP222202
KP222236
G1977
APU
MStovep
ipeCav
eKP221970
KP222004
KP222037
KP222071
KP222105
KP222135
KP222169
KP222203
KP222237
G1978
Cwar
IPickle
Pit
—KP222005
KP222038
KP222072
KP222106
KP222136
KP222170
KP222204
KP222238
G1979
Cwar
IPickle
Pit
KP221971
KP222006
KP222039
KP222073
—KP222137
KP222171
KP222205
KP222239
G1980
Cwar
IPickle
Pit
KP221972
KP222007
KP222040
KP222074
—KP222138
KP222172
KP222206
KP222240
G1981
APU
MSpider
Cav
eKP221973
KP222008
KP222041
KP222075
KP222107
KP222139
KP222173
KP222207
KP222241
G1982
APU
ISpider
Cav
eKP221974,75
KP222009
KP222042
KP222076
KP222108
KP222140
KP222174
KP222208
KP222242
G1984
APU
IBeard
Ran
chCav
eKP221976
—KP222043
KP222077
KP222109
KP222141
KP222175
KP222209
KP222243
G2009
APU
IKen
tButler
Pit
KP221977
KP222010
KP222044
KP222078
KP222110
KP222142
KP222176
KP222210
KP222244
G1985
APU
FJester
Estates
Cav
eKP221978
KP222011
KP222045
KP222079
KP222111
KP222143
KP222177
KP222211
KP222245
G1986
APU
MJester
Estates
Cav
eKP221979,80
KP222012
KP222046
KP222080
KP222112
KP222144
KP222178
KP222212
—G1988
APU
IJest
JohnCav
eKP221981
KP222013
KP222047
KP222081
KP222113
KP222145
KP222179
KP222213
KP222246
G1960
Cred
FCotterellCav
eKP221982
KP222014
KP222048
KP222082
KP222114
KP222146
KP222180
KP222214
KP222247
G1989
Cred
ICotterellCav
eKP221983
KP222015
KP222049
KP222083
—KP222147
KP222181
KP222215
KP222248
APU,apriori
uniden
tified
;I,im
mature.
© 2014 John Wiley & Sons Ltd
CAVE SPIDER SPECIES DELIMITATION 361