Molecular Phylogenetics and Evolution 91 (2015) 56–67

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier .com/locate /ympev

Multilocus sequence data reveal dozens of putative cryptic speciesin a radiation of endemic Californian mygalomorph spiders(Araneae, Mygalomorphae, Nemesiidae) q

http://dx.doi.org/10.1016/j.ympev.2015.05.0161055-7903/� 2015 Elsevier Inc. All rights reserved.

q This paper has been recommended for acceptance by Marcos Perez-Losada.⇑ Corresponding author.

E-mail address: dhleavitt@gmail.com (D.H. Leavitt).

Dean H. Leavitt a,⇑, James Starrett a, Michael F. Westphal b, Marshal Hedin a

a Department of Biology, San Diego State University, San Diego, CA 92182, United Statesb Bureau of Land Management, Hollister Field Office, Hollister, CA 95023, United States

a r t i c l e i n f o a b s t r a c t

Article history:Received 6 March 2015Revised 11 May 2015Accepted 19 May 2015Available online 27 May 2015

Keywords:DNA barcodingCryptic speciesGenealogical concordanceSpecies treeSympatry

We use mitochondrial and multi-locus nuclear DNA sequence data to infer both species boundaries andspecies relationships within California nemesiid spiders. Higher-level phylogenetic data show that theCalifornia radiation is monophyletic and distantly related to European members of the genusBrachythele. As such, we consider all California nemesiid taxa to belong to the genus CalisogaChamberlin, 1937. Rather than find support for one or two taxa as previously hypothesized, genetic datareveal Calisoga to be a species-rich radiation of spiders, including perhaps dozens of species. This conclu-sion is supported by multiple mitochondrial barcoding analyses, and also independent analyses ofnuclear data that reveal general genealogical congruence. We discovered three instances of sympatry,and genetic data indicate reproductive isolation when in sympatry. An examination of female reproduc-tive morphology does not reveal species-specific characters, and observed male morphological differ-ences for a subset of putative species are subtle. Our coalescent species tree analysis of putativespecies lays the groundwork for future research on the taxonomy and biogeographic history of thisremarkable endemic radiation.

� 2015 Elsevier Inc. All rights reserved.

1. Introduction

Mygalomorph spiders, the radiation that includes tarantulasand trapdoors spiders, are ancient and diverse. Because of generaldispersal limitation and high levels of regional endemism, severalresearch groups from different continents have developed mygalo-morphs as models for studies of species delimitation, cryptic spe-ciation, and historical biogeography (e.g., Bond and Stockman,2008; Satler et al., 2013; Hedin et al., 2013; Hendrixson et al.,2013; Castalanelli et al., 2014; Hamilton et al., 2014; Opatovaand Arnedo, 2014a). Still, many features of mygalomorph biologymake them challenging to study. These spiders are largely subter-ranean and spend much of their lives out of sight; also, much oftheir diversity is obfuscated by conserved morphology across bothspecies and higher taxonomic groups. Raven (1985) used morphol-ogy to revise taxonomy and contributed greatly to our understand-ing of higher-level mygalomorph phylogenetic relationships.Recent phylogenetic studies using DNA sequence data have

recovered support for some previous morphology-based hypothe-ses, but have also demonstrated the limitations of morphologicaldata and revealed some surprising relationships (Hedin andBond, 2006; Bond et al., 2012; Starrett et al., 2013; Opatovaet al., 2013; Opatova and Arnedo, 2014a).

In addition to improving our understanding of deeper phyloge-netic relationships, DNA sequence data have played a crucial rolein uncovering and recognizing species-level diversity in mygalo-morphs. Recent genetic studies have revealed ancient but morpho-logically cryptic species diversity in various mygalomorph groups(Starrett and Hedin, 2007; Stockman and Bond, 2007; Hedin andCarlson, 2011; Graham et al., 2015). Thus far, most genetic studieshave relied principally on one or two fragments of mitochondrialDNA. Theory has long emphasized the shortcomings of relying ona single marker (Maddison, 1997), and empirical data increasinglydemonstrate that inferences based on mtDNA alone can be mis-leading (e.g., McGuire et al., 2007). For example, the mitochondrialgenome is particularly prone to asymmetric introgression acrossspecies boundaries (Schwenk et al., 2008), leading to both misun-derstanding species boundaries and gene tree/species tree incon-gruence if mtDNA is the sole locus employed (e.g., Harringtonand Near, 2012). The need for accompanying nuclear markers is

D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67 57

generally well appreciated, and while nuclear DNA data havebecome standard in many taxonomic groups, thus far only a fewrecent studies of mygalomorphs have included multiple nuclearloci (e.g., Satler et al., 2013; Hedin et al., 2013; Opatova andArnedo, 2014b).

The mygalomorph family Nemesiidae has a global distribution,with particularly high generic and species diversity in Australiaand South America (Raven, 1985; Goloboff, 1995; World SpiderCatalog, 2015). While recent phylogenetic data reveal uncertaintyin the composition of the family (Bond et al., 2012), NorthAmerican nemesiid diversity is known to be depauperate relativeto other continental regions. Mexico is home to Mexentypesa chia-pas (Raven, 1987) and at least one undescribed species (Hedin andBond, 2006). Other North American representatives include fivenominal species from California: Brachythele longitarsis Simon1891, B. anomala Schenkel 1950, Calisoga theveneti Simon 1891,Calisoga sacra Chamberlin 1937 and Calisoga centronethaChamberlin and Ivie, 1939 (World Spider Catalog, 2015).

In his unpublished dissertation, Bentzien (1976) foundCalifornia Brachythele to be morphologically distinct fromEuropean members of that genus and assigned all California spe-cies to the genus Calisoga. These spiders are robust, aggressive spi-ders that reside in subterranean burrows with ‘‘simple’’ openentrances, similar to certain tarantulas. Based on statistical analy-ses of morphology, Bentzien (1976) hypothesized that there areonly two species of Calisoga in California: the common, widespread‘‘C. longitarsis’’ and the smaller, rare C. theveneti. However, he alsoconsidered the possibility that spiders identified as C. thevenetiare merely ‘‘C. longitarsis’’ males that mature at an earlier age,and that all populations represented a single species. Despite thefact that these results were never published, many arachnologistshave come to use the combination ‘‘Calisoga longitarsis’’ to referto all Californian nemesiids (e.g., Ubick et al., 2005). Californianemesiids are one of the more commonly encountered mygalo-morphs in central and northern California, particularly during thefall when wandering males enter houses in the San Francisco Bayarea. That the binomial most commonly applied to this taxon –‘‘Calisoga longitarsis’’ – remains unpublished and technicallyunavailable (World Spider Catalog, 2015) underscores the needfor systematic attention. In this study we employ multi-locusDNA sequence data to examine the phylogenetic relationship ofCalifornia nemesiids relative to Old World Brachythele and to inferboth species boundaries and phylogenetic relationships within theCalifornia radiation. This robust phylogenetic hypothesis providesthe framework for future biogeographic studies and formal mor-phological taxonomic revision.

2. Methods

2.1. Taxon sampling and data collection

We tested the distinctiveness of Californian nemesiids relativeto the Palearctic Brachythele by performing molecular phylogeneticanalyses of available nemesiid taxa. Using published data andnewly sampled spiders, we compiled a data set including multipleoutgroup taxa representing the families Cyrtaucheniidae andCtenizidae as well as 21 nemesiids and allied taxa(Microstigmatidae and ‘‘Cyrtaucheniidae’’). This sample alsoincluded multiple representatives of both European Brachytheleand Californian spiders currently referred to as ‘‘Calisoga longitar-sis’’ (see Supplementary Material for specimen information).

To investigate species diversity within Californian nemesiids,we sampled spiders from throughout the known distribution incentral and northern California (following Bentzien, 1976). Oursampling effort also included original collections that extended

the known distribution both to the north and the southwest. Weattempted to collect 3–5 individuals per locality, although somelocalities yielded fewer individuals. In total, this study includesgenetic data for 197 California nemesiid specimens from 106 dis-tinct localities (Fig. 1). Whole spiders were vouchered and depos-ited in the San Diego State University Terrestrial ArthropodsCollection, and typically one leg was cryopreserved for subsequentgenetic work. DNA was extracted from ethanol-preserved tissueusing the Qiagen DNeasy kit.

2.2. Molecular protocols

For the higher-level phylogenetic analysis of Nemesiidae, wedownloaded published data from Genbank for 28S and elongationfactor 1 c (EF-1c), nuclear loci that are widely used in phylogeneticanalyses of mygalomorphs (Hedin and Bond, 2006; Ayoub et al.,2007; Bond et al., 2012; Opatova et al., 2013). We also includedadditional data from transcriptomes and supplementary PCR reac-tions (see Supplementary Material, which also includes 28S andEF-1c primers). All PCR products were purified with MilliporeMultiscreen PCR filter plates and Sanger sequenced in both direc-tions by Macrogen, USA.

Using a published ‘‘Calisoga longistarsis’’ mitochondrial genome(Masta and Boore, 2008), we modified universal primers to amplifya region of the cytochrome oxidase I gene (COI) comprising over1200 base pairs (bp). In addition to the primers used in amplifica-tion, we sometimes used internal sequencing primers(Supplementary Material). The COI fragment was targeted for allCalifornian samples included in the study. Additionally, an approx-imately 1000 bp mitochondrial region that contained partial 12S,val-tRNA and 16S regions was sequenced for a subset ofCalifornian specimens.

Based on preliminary analyses of mitochondrial data and geo-graphic distribution, we collected nuclear data for forty individu-als. We amplified the internal transcribed spacer (ITS) units 1and 2 using published primers (Ji et al., 2003). Additional nuclearloci were developed from genomic resources generated specificallyfor Californian samples using comparative transcriptomic data (seeThomson et al., 2010) derived from next-generation sequencing. Atranscriptome was characterized for a spider from Panoche Hills(Fresno Co.), and compared to an already published transcriptomefrom Jasper Ridge (San Mateo Co., SRX652491; Bond et al., 2014).Methodological details (including phasing of data) are similar tothose used in Derkarabetian and Hedin (2014) and are further out-lined in the Supplementary Material. From this effort we selectedseven exonic loci that amplified consistently across a geographictest panel. We performed PCR reactions with typical reagents,and primers are available in Supplementary Material. Purifiednuclear loci were Sanger sequenced using forward and reverse pri-mers by Macrogen USA.

2.3. Higher-level alignment and phylogenetic analyses

An initial alignment of EF-1c was performed in Sequencher andmanually adjusted after translating codons into amino acids. Toalign the ribosomal 28S fragment, we used the option Q-INS-i inMAFFT v7 (Katoh et al., 2005) as it considers RNA secondary struc-ture during alignment. The default settings include a gap-openingpenalty of 1.53 and the offset value equal to 0.0. We used the pro-gram GBLOCKS v0.91b (Castresana, 2000) to identify and excluderegions that could not be reliably aligned. Settings for this analysisallowed gap positions within the final blocks as well as smallerfinal blocks. EF-1c and 28S data were concatenated for phyloge-netic analyses. The program jModeltest (Posada, 2008) was usedto aid in selection of nucleotide models for these and subsequentphylogenetic analyses. We performed a maximum-likelihood

Fig. 1. Map of central and northern California showing collecting localities included in this study. Colors and clade names correspond to ten major biogeographic cladesrecovered in mitochondrial gene tree analyses. Localities where divergent mitochondrial and nuclear lineages were found in sympatry are indicated in white. Black x’sdesignate collection locations for CAS male specimens. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of thisarticle.)

58 D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67

analysis using the RAxML-HPC BlackBox 8.0.24 (Stamatakis et al.,2008) on the CIPRES Science Gateway (Miller et al., 2011). AGTRgamma model was used for each of four partitions: three forcorresponding codon positions within EF-1c and a fourth for the28S fragment. The sufficient number of bootstrap replicates wasautomatically detected by the analysis. We consider bootstrap val-ues >70 as evidence for high support (Hillis and Bull, 1993; Alfaroet al., 2003). For a Bayesian BEAST analysis, we used the same

partitioning scheme and nucleotide models (GTR + G), exceptHKY + G models were used for the three EF-1c partitions to achieveconvergence in certain estimated parameters. This analysisemployed a lognormal relaxed clock for each of the two generegions and a tree prior of birth–death with incomplete sampling.The BEAST analysis was run for 100 million generations savingevery 4000th. We used TRACER v.1.5 (Rambaut and Drummond,2007) to plot likelihood scores in order to assess convergence, to

D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67 59

identify the appropriate burn-in length, and to verify that all effec-tive sample size (ESS) values were well over 200.

2.4. Mitochondrial gene tree, species delimitation and DNA barcoding

Mitochondrial COI data were translated into amino acids to ver-ify an open reading frame. With no insertions or deletions, the COIalignment was trivial and performed in Sequencher. For the riboso-mal 12S/16S data, we performed analyses with MAFFT andGBLOCKS v0.91b with the same settings as outlined above. To inferthe mitochondrial gene tree, we concatenated the two fragments.Similar to higher-level phylogenetic analyses, likelihood andBayesian analyses contained four partitions: three for the COIcodon positions and a fourth for the ribosomal fragment. We per-formed maximum-likelihood analysis in RAxML with the modelGTRgamma for each partition and automatic bootstrapping. Forthe BEAST analysis, the use of HKY + G models for each of the fourpartitions was necessary to achieve convergence. A strict clock wasused for the COI fragment and a lognormal clock for the 12S frag-ment, with the 12S rate estimated relative to the COI rate of 1.0substitution per site. We used a birth–death and incomplete sam-pling tree prior.

We used three different approaches to DNA barcoding to iden-tify putative species limits. The first method was the traditionalbarcoding approach (Hebert et al., 2003a, 2003b) where geneticdistances are calculated and a threshold value that designates dis-tinct species status is assigned. The divergences among relevantsister clades recovered from the BEAST mtDNA gene tree were cal-culated using the uncorrected p-distance similar to recently pub-lished barcoding studies of mygalomorphs (Castalanelli et al.,2014; Hamilton et al., 2014). To be able to directly compare theresults of this system to the aforementioned studies, we identifiedputative species using the threshold values of both 9.5%(Castalanelli et al., 2014) and 5% (Hamilton et al., 2014). The 9.5%value was used to characterize species diversity of mygalomorphswithin the Pilbara region of Australia (the majority of whichbelonged to the family Nemesiidae), and the 5% cutoff was usedfor species discovery within the North American tarantula genusAphonopelma sampled primarily from the United States.

The third barcoding approach used was the Automatic BarcodeGap Discovery (ABGD) method (Puillandre et al., 2012), imple-mented on the online ABGD server (http://www.abi.snv.jussieu.fr/public/abgd/). This analysis uses the ordered, ranked genetic dis-tances from COI sequences to identify marked shifts from lowintraspecific distances to higher interspecific distances. Theseobserved transitions in genetic distances represent barcode gaps.Gap detection is an iterative process that assigns sampled individ-uals into increasingly partitioned groups. We used the sameparameters employed in the tarantula study (Hamilton et al.,2014) for comparative purposes, as follows: Pmin = 0.0001,Pmax = 0.200, Steps = 10, X = 1, Nb bins = 20. Hamilton et al.(2014) found that species delimitation with this approach appearsto be fairly robust to varying parameter settings.

2.5. Nuclear gene genealogical concordance

Species delimitation using a single mitochondrial marker can bemisled due to various processes, including both incomplete lineagesorting and introgression. Additionally, deep mitochondrial struc-ture is not always accompanied with correspondingly deep nucleardifferentiation (Ogden and Thorpe, 2002), and dispersal-limitedtaxa (e.g., mygalomorph spiders) in particular may show spuriousmtDNA breaks (Irwin, 2002). On the other hand, genealogical con-cordance among multiple sampled loci can provide convincing evi-dence for genetic boundaries (Avise and Ball, 1990; Kuo and Avise,2005), and thus multi-locus data can validate the presence of

genetically distinctive but morphologically cryptic lineages(Satler et al., 2013).

We evaluated nuclear data separately from mitochondrial evi-dence to provide an independent test of putative species identifiedfrom mtDNA barcoding. We performed a species tree analysiswhere each of the forty individuals with sampled nuclear datawas treated as a distinct OTU. Identified mtDNA barcode speciesare considered validated by the nuclear species tree analysis whenindividuals are recovered in a clade with other members of thatspecies (when multiple individuals sampled), with non-trivialbranch lengths separating a clade/individual from their respectivesister clade. Species tree inference was performed using *BEASTwith BEAST 1.8. Because initial analyses using more complex GTRmodels failed to converge, all eight phased nuclear loci were ana-lyzed with HKY + G models (Grummer et al., 2014) and a lognormalrelaxed clock. The ITS rate was set to 1.0 and the rates of other lociwere estimated relative to that of ITS. An inverse gamma shape wasused for the species population mean prior, and an exponentialshape was used for the Yule birth–death prior. Analyses were runfor 200 million generations with every 5000th generation saved.

2.6. Combined species tree analyses

We performed a combined species tree analysis of the nuclearand mitochondrial data similar to the *BEAST analysis of thenuclear-only data. OTUs for this analysis were mtDNA barcode spe-cies identified at the 9.5% threshold. In addition, in three instancesindividuals from divergent clades just under this threshold (�9%)were treated as distinct taxa (species Q was divided into Q1 andQ2, H into clades H and I, and F into F1 and F2). Mitochondrial bar-code species that lacked nuclear data were excluded in this analy-sis. As with other *BEAST analyses of these data, all loci wereanalyzed with an HKY + G model of sequence evolution. The COIfragment was partitioned by codon position, and a strict clockwas employed for this gene region alone. Other gene regions usedlognormal relaxed clocks whose rates were estimated relative tothat of COI. Other priors were the same as the nuclear-only analy-sis, with analyses run for 200 million generations with every5000th generation saved.

2.7. Morphological data

The morphology of adult male pedipalps and first legs (whichpossess mating spurs) is often used as a primary character for spe-cies delimitation in mygalomorph spiders, including nemesiids(e.g., Goloboff, 1995; Decae and Cardoso, 2005; Decae et al.,2014). To assess qualitative morphological divergence acrossCalifornian populations, adult males (N = 9) were borrowed fromthe California Academy of Sciences (the same males studied byBentzien, 1976), and augmented with males (N = 9) from ourown collections. Also, mature spermathecal organs for all adultfemales included in our molecular study were dissected and exam-ined (N = 103). Specimens were imaged using a Visionary DigitalBK Plus system (http://www.visionarydigital.com), including aCanon 5D digital camera, Infinity Optics Long DistanceMicroscope, P-51 camera controller, and FX2 lighting system.Individual images were combined into a composite image usingHelicon Focus V5.3, and then edited using Adobe Photoshop CS6.

3. Results

3.1. Higher-level phylogenetic placement

The EF-1c data included for phylogenetic analysis consists of924 aligned base pairs. The MAFFT alignment of the 28S fragment

60 D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67

resulted in 1964 aligned bp, and the GBLOCKS analysis reducedthis length to 1812 bp. All sampled taxa were represented by theEF-1c fragment; three individuals lacked the 28S fragment butwere closely related to taxa with complete data. Of the total2736 aligned characters, 665 were variable and 440 were parsi-mony informative.

Maximum-likelihood and Bayesian BEAST analyses recoveredsimilar topologies, differing only in the placement of the clade con-taining Hermacha sp. and the undescribed taxon from NgomeForest, South Africa. The topology recovered from the BEAST anal-ysis is reported in Fig. 2. The sampled Californian spiders are recov-ered with high support in a clade containing representatives of twoMediterranean genera, Iberesia and Ambylocarenum. In the RAxMLanalysis, the Ambylocarenum sp. sample from Spain is recoveredas sister to Californian spiders with a high support (BS = 94); thisrelationship is also recovered in the Bayesian BEAST analysis butwith lower support (PP = 0.66). The European Brachythele is weaklyrecovered as sister to Fufius sp., and this clade is recovered as sisterto the aforementioned Iberesia + Amblyocarenum + Californianclade, also with weak support. These genera are separated bybranch lengths that are long relative to crown Nemesiidae (Fig. 2).

3.2. Mitochondrial gene tree and DNA barcoding

The COI matrix included data for 194 spiders that we sampled(Supplementary Material), plus COI and ribosomal mtDNA from apublished mitochondrial genome (Masta and Boore, 2008). Thealigned length of the COI fragment was 1263 base pairs; 560 posi-tions were variable, of which 515 were parsimony informative. Themitochondrial ribosomal fragment was obtained for 55 spiders.The mtDNA ribosomal alignment after MAFFT and GBLOCKS

Fig. 2. Maximum clade credibility tree from the BEAST analysis of concatenated EF-1c anat nodes represent Bayesian posterior probabilities first, followed by ML bootstrap valu

analyses consisted of 986 bp; 519 characters were variable with433 of those being parsimony informative.

Likelihood and Bayesian analyses recovered similar gene treetopologies; the Bayesian BEAST topology is depicted in Fig. 3. Thepartitioned HKY + G model with the strict molecular clock asemployed in BEAST indicates that deepest corrected divergencesamong sampled California specimens are �22%. The mtDNA genetree is composed of divergent, geographically cohesive clades thatare allopatric or parapatric to neighboring clades with few excep-tions. We arbitrarily name ten major clades with reference to theirgeographical distribution (Fig. 1, Supplementary Material). Theseclades include: Bay Area (locations 1–29), Sacramento Valley(30–36), Northern Sierra (37–53), Calaveras (50, 54), Tuolumne(55–58), Coulterville (59–63), Central Sierra (60, 64–74), SouthernSierra (75–78), Monterey (1, 79–101), and Panoche (l02–106).

We found sympatry of divergent mitochondrial lineages atthree localities (Fig. 1). Bay Area and Monterey clades were foundat locality 1 (Coalinga Road), Northern Sierra and Calaveras cladesat locality 50 (west of Sheep Ranch), and Coulterville and CentralSierra clades at locality 61 (west of Coulterville). At almost all otherlocalities with multiple sampled individuals, COI sequences wereeither identical or minimally divergent. An exception is locality45 (east of Mount Aukum), where the two sampled COI sequenceswere just over 6% divergent.

Using the traditional barcoding approach with a 9.5% threshold,26 putative species are identified (Fig. 3, Supplementary Material).Most of the major geographic clades previously referenced arecomposed of multiple species at this cutoff, but Tuolumne,Calaveras and Coulterville clades each correspond to a single spe-cies (Fig. 3). Using the 5% threshold (similar to Hamilton et al.,2014), the number of species identified more than doubles to 59species (Fig. 3, Supplementary Material). Of the aforementioned

d 28S data. Clade of sampled Nemesiidae and allied taxa is indicated. Support valueses.

Fig. 3. Maximum clade credibility tree from the mtDNA BEAST analysis inferred from concatenated COI and ribosomal fragments. Nodal support is shown, with Bayesianposterior probabilities first and ML bootstraps second. For simplicity, individual haplotypes have been collapsed into clades corresponding to putative species at the 5%threshold. Clades are also labeled according to recovered species at the 9.5% threshold. The sampled localities comprising each clade are listed. Major geographic clades arealso labeled (see Fig. 1).

Table 1Nuclear loci summary information.

Primer combo Annotation # Seq. Length bp # Inform. # Uninform

B1_B2 Orthologous to Ixodes ISCW003568 (conserved hypothetical protein) 40 945 44 3D1_D3 Orthologous to Ixodes ISCW012830 (60S ribosomal protein L3) 38 924 87 10D6_D7 Orthologous to Ixodes ISCW013467 (conserved hypothetical protein) 39 876 38 9E10_E11 Orthologous to Ixodes ISCW019432 (ccaat enhanced binding protein) 40 603 25 10F9_F10 Orthologous to Ixodes ISCW020841 (zinc finger protein) 37 578 56 14G9_G10 Orthologous to Ixodes ISCW024803 (60S ribosomal protein L23) 40 496 36 7H1_H2 Orthologous to Ixodes ISCW024894 (acetylcholinesterase) 40 585 41 7CAS18sF1 CAS5p8sB1d ITS ribosomal 34 992 69 16

Notes: # Seq. = the number of individuals sequenced, # Inform. = the number of parsimony informative sites, # Uninform = the number of variable but parsimony unin-formative sites.

D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67 61

geographic clades, only Coulterville corresponds to a single speciesat this cutoff. The Automatic Barcode Gap Discovery (ABGD)method results in a much higher number of putative species. Aminimum of 95 species are inferred with this approach(Supplementary Material), with most of these restricted to a singlesampled geographic location.

3.3. Species validation using nuclear-only data

Gene annotation and patterns of nuclear gene variation aresummarized in Table 1. The eight nuclear loci totaled 5999 bp with393 parsimony informative sites. The nuclear-only *BEAST analysiswith forty individuals treated as OTUs recovers a topology that dis-plays a great deal of topological concordance with the mtDNA genetree, both in terms of species membership as well as major clademembership (Fig. 4). Using the 9.5% threshold, 23 of 26 candidatespecies were sampled for nuclear data. Of those 23, ten species hadnuclear data collected for multiple individuals. In nine out of ten

instances individuals belonging to a single species were recoveredtogether in a clade. The exception was species H, as one member ofH was not recovered with the other two (Fig. 4). However, thesespiders belong to two divergent mitochondrial clades that are nearthe cutoff threshold (9.1%).

At the three localities with sympatric divergent mitochondrialclades, a specimen from each clade was sampled for nuclear data.In these three instances all eight nuclear loci were informative inregards to distinguishing species membership (Fig. 4). In each casenot only did the combined nuclear data substantiate the mtDNAclade membership, but also each of the nuclear loci individuallysupported the mitochondrial species membership.

In every case the species identified at both the 9.5% and 5% cut-offs are recovered in the same major clades with the nuclear-onlyspecies tree as they are with the mitochondrial gene tree.Relationships among major clades from the nuclear species treeand the mtDNA gene tree show a large degree of congruence, butthere are some differences. However, in none of these cases is a

Fig. 4. Maximum clade credibility tree from the *BEAST analysis of nuclear data with Bayesian posterior probabilities at each node. Individuals are labeled according toputative species at the 9.5% threshold, major geographic clade and voucher number. Sympatric specimens are indicated by a square (locality 50), a circle (locality 61) and astar (locality 1). Inset is adult female from Hare Canyon Road (Monterey County).

62 D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67

strongly supported node (0.95 PP or above) from the nuclear spe-cies tree in conflict with a strongly supported node in the mito-chondrial gene tree.

3.4. Combined data species tree

The species tree from the combined nuclear and mitochondrialdata is depicted in Fig. 5. Nodes that are in common with thenuclear-only species tree either maintain high support values orgenerally received increased support, indicating that the mito-chondrial gene tree has not been misled by introgression and canappropriately be analyzed with the nuclear loci in a combined spe-cies tree analysis. In the recovered phylogeny, the three northern-most clades (Sacramento Valley, North Sierra and Bay Area) arerecovered as sister to southerly-distributed clades. Among thesesouthern clades, the Central Sierra is sister to and deeply divergentfrom a clade containing the Tuolumne, Coulterville, Calaveras,Southern Sierra, Monterey and Panoche clades. Relationshipsamong these clades are not well supported except for the sisterrelationship between the Monterey and Panoche clades, bothdistributed west of the Central Valley.

3.5. Morphological variation

Our sample of adult males is geographically biased, as males areonly available for five of the ten primary geographic clades

(Supplementary Material) and an obviously smaller percentage ofputative molecular species. This sample includes the CAS speci-mens (without DNA data), most of which could be confidentlyplaced into larger geographic clades based on geographic location(Supplementary Material, Fig. 1). Comparison of palpal and leg Iclasper morphology among individuals from primary cladesindicates limited morphological divergence, with only very subtledifferences in spination, bulb shape, and relative length of legelements (Fig. 6). Obviously, an important task will now be targetedgeographic sampling of adult males, informed by the barcodingresults. Our sample of females with mature spermathecal organsis comprehensive, including all primary geographic clades, and all26 conservative barcode species (Supplementary Material).However, female variation showed no obvious trends relative toclade or species membership, with morphological asymmetry oftenapparent within individuals, and variation among individuals froma single location often greater than divergence among individualsfrom different primary clades (Supplementary Material).

4. Discussion

4.1. Taxonomic validity and composition of Calisoga

The combined results of our higher-level phylogenetic studyand our population-level sampling confirm that California neme-siid diversity consists of a single radiation that is divergent from

Fig. 5. Maximum clade credibility tree from the *BEAST analysis of the combined mitochondrial and nDNA data with Bayesian posterior probabilities at each node. Putativespecies were designated based on the 9.5% threshold, and the comprising localities are listed. Major geographic clades are also labeled. Sympatric localities are indicated by asquare (locality 50), a circle (locality 61) and a star (locality 1). Scale bar represents species tree branch lengths estimated using a rate of 1 substitution/site for the COIfragment.

D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67 63

(and not sister to) European Brachythele and should instead beplaced in the distinct genus Calisoga Chamberlin 1937. This conclu-sion was also reached by Bentzien (1976) in his unpublished dis-sertation based on observed morphological variation. The longbranches separating European Brachythele from Calisoga indicateancient divergences.

While the five named taxa from California were originally allo-cated to three different families, Raven (1985) transferred them allto Nemesiidae. He further recognized that they belonged in thegenus Calisoga rather than Brachythele, but he neglected to transferBrachythele longitarsis Simon 1891 and B. anomala Schenkel 1950(Ubick and Ledford, 2005). Thus, we recommend taxonomicchanges to make the two following names available.

Calisoga longitarsis (Simon 1891) comb. n.Brachythele longitarsis Simon 1891.Type locality: ‘‘California meridionalis’’.Remarks: The type of longitarsis was reported as being collected

in ‘‘southern California’’. The type specimen is a mature male and islocated in the M.N.H.N. in Paris. Bentzien (1976) examined thespecimen and provided measurements, a description, and partialillustrations. In the ‘‘1890’’ paper describing a female specimenof longitarsis (published after the 1891 description of the type),Simon lists a number of other regions where longitarsis specimensoccur; some of these are clearly in error (e.g., Idaho and Texas).While the specific geographic origin of the type of longitarsis is cur-rently unknown, it can be reasonably assumed that this species isrepresented in our comprehensive sampling (or is at least closely

related to sampled spiders). Due to their morphological similarityto the type, the majority of the populations sampled in this study(including the those in the higher-level analysis) have historicallybeen considered longitarsis.

Calisoga anomala (Schenkel 1950) comb. n.Brachythele anomala Schenkel 1950.Type locality: Wald um Guerneville und Monte Rio am Russian

River.Remarks: The type specimen of C. anomala is an immature ani-

mal in poor condition (Bentzien, 1976). However, the type localityis perhaps the most specific of named Calisoga: forest aroundGuerneville and Monte Rio along the Russian River (SonomaCounty, CA).

In addition to the two taxa above, the following species namesare also available for the genus Calisoga: C. sacra Chamberlin, 1937from Sacramento, CA, the type of genus; C. theveneti (Simon, 1891)from Mariposa, CA; and C. centronetha (Chamberlin and Ivie, 1939),from Mayfield (Palo Alto), CA. Thus, five names are available forCalifornia nemesiid diversity. In his unpublished dissertation,Bentzien (1976) considered Calisoga longitarsis, C. sacra, C. cen-tronetha, and C. anomala as synonyms.

With at least 26 molecular species and five available names,taxonomy within Calisoga clearly needs further attention. Thesmall body size and the geographic distribution of ‘‘species C’’ indi-cate that it should tentatively be considered C. theveneti. The sam-pled population that is closest to the type locality of C. sacra(Sacramento) is Deer Valley Rd. and belongs to ‘‘species Q1’’. The

noSier_Loc44 _GMY187

Bay_Loc21_MY1514

CentSier_Loc65_MY4603

Coulter_Loc60 _MY 4658

Fig. 6. Digital images of palpal and leg I clasper morphology from male spiders representing four primary geographic clades. Spiders are labeled with primary clade, localityand voucher number. All views prolateral, scale bars = 1 mm. Arrows point to minor differences in spination and relative leg length observed in the sample.

64 D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67

type locality of C. anomala falls within the distribution of ‘‘speciesS’’, and the type of C. centronetha falls within that of ‘‘species U’’.Unfortunately, the type locality of C. longitarsis is currentlyunknown; should its origin be determined, this older name wouldhave priority if synonymous with one of the three latter namedtaxa.

4.2. Molecular species diversity

Given the results of other molecular studies of California myga-lomorph spiders (e.g., Bond and Stockman, 2008; Hedin et al.,2013; Satler et al., 2013), it is not surprising that themorphology-based ‘‘two species’’ hypothesis of species diversitywithin Calisoga is an underestimate. Using the conservative 9.5%

cutoff used in the Pilbara study (Castalanelli et al., 2014), we finda similar number of molecular species per sampled spider: about1 species per 7 spiders sampled (7.5 in this study, 7 [1134 spi-ders/161 lineages] in Castalanelli et al., 2014). Even within themore conspicuous genus Aphonopelma, mitochondrial barcodingidentified previously unknown species diversity with perhaps ninemorphologically cryptic species recovered (Hamilton et al., 2014).Of course, the number of ‘‘real species’’ represented in this studyremains unknown. How much divergence is needed to correctlyidentify distinct species is a looming question in the DNA barcod-ing literature (Hamilton et al., 2014) and is likely taxon specific. Inthis study system, the 9.5% threshold does not appear to result inan overestimate of putative species, as the nuclear data supportgenetic distinctiveness at this cutoff. This divergence threshold

D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67 65

may in fact be conservative, and in a few instances thenuclear-only topology led us to treat clades just under the thresh-old (�9%) as distinct OTUs.

On the other hand, the approaches of species delimitation usingmethods such as ABGD and GMYC (Pons et al., 2006), which do notuse an arbitrary threshold, have been shown to return unrealisticnumbers of species in other mygalomorph studies (e.g., Hamiltonet al., 2014). In a study of the California trapdoor spider speciesAliatypus thompsoni, bGMYC identifies 61–79 species at a 95% con-fidence interval with mitochondrial data, whereas the authors pre-fer a conservative three species hypothesis based on analyses ofmultilocus variation (Satler et al., 2013). In the current study,ABGD infers a minimum of 95 species from 106 sampled locations– almost one species per locality. The overestimation of speciesusing these methods is due to structured genetic patterns emerg-ing from the viscous nature of mygalomorph populations (e.g.,Hedin et al., in press). The biology of burrow-dwelling mygalo-morphs may be particularly prone to generating extreme mito-chondrial structuring. In species not characterized by balloondispersal, spiderlings generally establish burrows in favorable sub-strate near the maternal burrow. Female mygalomorphs usuallyremain in the same burrow throughout their entire lives. Males,on the other hand, are known to disperse from their burrow oncethey reach reproductive maturity as they seek opportunities tomate with receptive females. Female philopatry and male-baseddispersal could create situations where mitochondrial data overes-timate isolation and the number of distinct lineages.

The boundary between intraspecific variation and interspecificvariation appears to be better defined in Aphonopelma tarantulas(Hamilton et al., 2014) than in Calisoga. Some tarantula cladesshow patterns of shallow intraspecific mtDNA divergence, indica-tive of either recent population expansion or a nonviscous geneticstructure (Hamilton et al., 2011). This includes a recent phylogeo-graphic analysis of tarantulas from central California (Wilson et al.,2013) where the sampled area overlaps a great deal with the areasampled in this study. Because of some ecological similarities totarantulas, we expected Calisoga to share similar phylogeographicpatterns. However, the extreme level of mitochondrial structuringobserved within Calisoga is more similar to that seen in trapdoorspiders (Satler et al., 2013; Stockman and Bond, 2007; Cooperet al., 2011) and turret spiders (Starrett and Hedin, 2007; Hedinet al., 2013). In these systems, mtDNA clades are narrowly dis-tributed, often to a single sampled locality, and genetic divergenceto neighboring clades can be surprisingly deep. Identifying theboundary between intraspecific phylogeographic structure versusspecies boundaries is challenging in low-dispersing, spatially vis-cous organisms (Bond and Stockman, 2008; Keith and Hedin,2012; Schuchert, 2014). In an example from California, mitochon-drial lineages of the slender salamander Batrachoseps attenuatusare narrowly distributed, leading to the suggestion that this taxoncomprises 39 extant lineages (Highton, 2014), whereas the initialresearchers that generated the data (Martínez-Solano et al.,2007) consider this taxon to represent a single species with deepphylogeographic structure.

Because species boundaries based on mitochondrial data can bemisled by processes like introgression, incomplete lineage sortingand/or phylogeographic structure unaccompanied by nuclear dif-ferentiation, the need for nuclear data is well appreciated. Giventhe biology of these spiders, the possibility for discordance betweenthe mitochondrial genome and the rest of the genome seemed con-siderable. But rather than reveal potential pitfalls of amitochondrial-only analysis, the nuclear-only species tree analysisprovided strong validation for the mtDNA-identified species-leveldiversity. Our nuclear-only approach to validating mitochondrialspecies boundaries took a qualitative approach. We tested whetherindividuals belonging to the same mtDNA-identified species were

recovered in a clade using nuclear-only data, and we evaluatedwhether the branch lengths separating different species or individ-uals were consistent with distinct species status. We did not per-form statistical evaluations of branch lengths, but ratherconsidered proportional length relative to overall tree length.Methods have been developed to statistically test species bound-aries using species trees. For example, Bayes factor delimitation(BFD) (Grummer et al., 2014) evaluates likelihood scores of differ-ent species models that either split or group individuals (or reas-signs them), and a related approach has been developed with theprogram DISSECT (Jones et al., 2015) that allows more flexibilityin testing split or lumped arrangements. The use of DISSECTappears to have the potential to over-split species with phylogeo-graphic structure (Jones et al., 2015), however, which is a concernthat likely extends to methods like BFD as well. Where the cutoffbetween interspecific structure and species boundaries lies in lowdispersing organisms will continue to be an interesting questionas relative information content increases with genomic datasets.

The observed concordance of nuclear and mitochondrial lineageboundaries provides convincing verification that Calisoga containsnot one or two species but likely over two dozen. Particularly com-pelling evidence that the lineages recovered in this study represent‘‘good species’’ is the discovery of three instances of sympatry, eachinvolving different species pairs (Figs. 1, 4 and 5). In all threeinstances, reproductive isolation appears to be quite high with nodetected evidence for gene flow among the sampled nuclear loci.Reproductive isolation in sympatry is considered the most con-vincing evidence of distinct species boundaries (Mayr, 1942), andhas been observed in morphologically cryptic diatoms (Amatoet al., 2007), lichens (Leavitt et al., 2011), earthworms (Donnellyet al., 2013) and lizards (Barley et al., 2013). Certainly the extentof sympatry among Calisoga species is greater than that shownhere since we sampled relatively few spiders per locality.Combining future fine-scale geographic sampling with larger sam-ple sizes, to discover and study additional cases of sympatry, willconstitute one of the strongest tests of species limits in Calisoga.

4.3. Phylogenetic relationships and taxonomy

Many of the phylogenetic relationships recovered from thecombined data species tree analysis are well supported, includingthose deepest in the tree (Fig. 5). Additional data is needed to solid-ify relationships among lineages within the Northern Sierra cladeas well as those among the geographically restricted clades inthe central Sierran region, however. As seen in the mtDNA genetree and nuclear-only species tree, the combined phylogeny showsthat sympatrically occurring species are distantly related, with twoof the three instances spanning the deepest divergence withinCalisoga (Fig. 5).

Among the Australian mygalomorphs from the Pilbara, mostsampled males were morphologically distinct at the 9.5% threshold(Castalanelli et al., 2014). The exception was the Aname mellosacomplex, where Castalanelli et al. recovered ten lineages at thatthreshold that were not obviously morphologically distinguishable.To what extent are the molecularly identified species of Calisogamorphologically identifiable? Surveying the morphology of femalegenitalic characters, we observed no obvious species-specific orclade specific morphology, even though variation among femaleswas clearly evident (Supplementary Material). An inspection ofmale morphology reveals limited morphological divergence, withonly subtle differentiation at the 9.5% cutoff (Fig. 6). Much moreextensive sampling of male specimens is needed to establishwhether these minor differences are taxonomically useful.

Bentzien (1976) considered most of the Californian populationsto belong to a single widespread species, but he did observe mor-phological variation among these populations, including differing

66 D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67

average body sizes. While our sample sizes are small, we alsoobserved some size variation among different populations.Additional sampling is needed to determine whether differentlysized populations specifically mentioned by Bentzien (1976) repre-sent distinct species. For example, the males he observed at IndianValley, Sierra Co. were among the smallest in his study. TheCalisoga specimens we found from the northern Sierra region wereoften very large, however, suggesting the Indian Valley spidersmay be a distinct, unsampled species. Additionally, Bentzien foundboth small (<7 mm) and large mature males (>10 mm) at HoplandField Station, Mendocino Co., which suggests the possibility ofsympatric species. Analyses of morphology have helped shed lighton the taxonomic status of the Nemesiidae genera Ambylocarenum(Decae et al., 2014) and Iberesia (Decae and Cardoso, 2005); amolecular perspective will undoubtedly provide additional clarifi-cation by revealing genetically unique lineages in these genera asthey have in Calisoga.

4.4. Conclusions

The results of our study reveal that the California genus Calisogais distinct from the European genus Brachythele, and that Calisogarepresents a species rich radiation, rather than just one or two spe-cies. A correct understanding of species boundaries is requisite forunderstanding evolutionary history as well as ecological relation-ships (e.g., Pante et al., 2015). Additionally, the genetic data pro-vides an important perspective in conservation planning (e.g.,Hedin, 2015) and habitat preservation (Moritz and Faith, 1998),and the inclusion of endemic arthropods has proven to be of highutility (Redak, 2000; Bond et al., 2006; Arnedo and Ferrández,2007; Harvey et al., 2011). Species discovery in organisms thatare narrowly distributed is an iterative process (Hedin et al.,2013), and further geographic sampling of Calisoga is needed totest hypotheses derived from this study and to reveal additionaldiversity. While much remains to be done regarding systematicwork in this genus, this study sets the stage for investigating thetemporal and spatial aspects of diversification of Calisoga in thegeologically dynamic region of central and northern California.Other recent biogeographic studies of California mygalomorphshave revealed both shared patterns with other organisms as wellas genuine surprises (Hedin et al., 2013).

Funding

This work was supported by a Theodore Roosevelt MemorialGrant from the American Museum of Natural History awarded toDean Leavitt and funding from the Bureau of Land Management.

Acknowledgments

For help in specimen collection we would like to thank BobThomson, Levi Gray, Peter Scott, Anna Bennett, Eileen Bennett,Chris Hamilton, Jordan Satler, Erika Garcia, Robin Hedin, BobKeith, Steven Thomas, Ryan Fitch, Dave Carlson, Axel Schönhofer,and Lars Hedin. Jason Bond provided assembled transcriptomesfor the Jasper Ridge Calisoga, and Pionothele. Charles Griswoldand Darrell Ubick facilitated the specimen loan from theCalifornia Academy of Sciences. Cheryl Hayashi provided lab spaceand other resources for RNA extraction work.

Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.ympev.2015.05.016.

References

Alfaro, M.E., Zoller, S., Lutzoni, F., 2003. Bayes or bootstrap? A simulation studycomparing the performance of Bayesian Markov chain Monte Carlo samplingand bootstrapping in assessing phylogenetic confidence. Mol. Biol. Evol. 20,255–266.

Amato, A., Kooistra, W.H., Levialdi Ghiron, J.H., Mann, D.G., Pröschold, T., Montresor,M., 2007. Reproductive isolation among sympatric cryptic species in marinediatoms. Protist 158, 193–207.

Arnedo, M.A., Ferrández, M.A., 2007. Mitochondrial markers reveal deep populationsubdivision in the European protected spider Macrothele calpeiana (Walckenaer,1805) (Araneae, Hexathelidae). Conserv. Genet. 8, 1147–1162.

Avise, J.C., Ball, R.M., 1990. Principles of genealogical concordance in speciesconcepts and biological taxonomy. Oxford Surv. Evol. Biol. 7, 45–67.

Ayoub, N.A., Garb, J.E., Hedin, M., Hayashi, C.Y., 2007. Utility of the nuclear protein-coding gene, elongation factor-1 gamma (EF-1c) for spider systematics,emphasizing family level relationships of tarantulas and their kin (Aranaea:Mygalomorphae). Mol. Phylogenet. Evol. 42, 394–409.

Barley, A.J., White, J., Diesmos, A.C., Brown, R.M., 2013. The challenge of speciesdelimitation at the extremes: diversification without morphological change inPhilippine sun skinks. Evolution 67, 3556–3572.

Bentzien, M.M., 1976. The Biosystematics of the Spider Genus Brachythele Ausserer(Araneidae: Dipluridae). PhD Dissertation, UC Berkeley, CA.

Bond, J.E., Stockman, A.K., 2008. An integrative method for delimiting cohesionspecies: finding the population-species interface in a group of Californiantrapdoor spiders with extreme genetic divergence and geographic structuring.Syst. Biol. 57, 628–646.

Bond, J.E., Beamer, D.A., Lamb, T., Hedin, M., 2006. Combining genetic and geospatialanalyses to infer population extinction in mygalomorph spiders endemic to theLos Angeles region. Anim. Conserv. 9, 145–157.

Bond, J.E., Hendrixson, B.E., Hamilton, C.A., Hedin, M., 2012. A reconsideration of theclassification of the spider infraorder Mygalomorphae based on three nucleargenes and morphology (Arachnida: Araneae). PLoS One 7, e38753.

Bond, J.E., Garrison, N.L., Hamilton, C.A., Godwin, R.L., Hedin, M., Agnarsson, I., 2014.Phylogenomics resolves a spider backbone phylogeny and rejects a prevailingparadigm for orb web evolution. Curr. Biol. 24, 1765–1771.

Castalanelli, M.A., Teale, R., Rix, M.G., Kennington, W.J., Harvey, M.S., 2014.Barcoding of mygalomorph spiders (Araneae: Mygalomorphae) in the Pilbarabioregion of Western Australia reveals a highly diverse biota. Invertebr. Syst. 28,375–385.

Castresana, J., 2000. Selection of conserved blocks from multiple alignments fortheir use in phylogenetic analysis. Mol. Phylogenet. Evol. 17, 540–552.

Chamberlin, R.V., Ivie, W., 1939. New tarantulas from the southwestern states. Bull.Univ. Utah 29, 1–17.

Cooper, S.J., Harvey, M.S., Saint, K.M., Main, B.Y., 2011. Deep phylogeographicstructuring of populations of the trapdoor spider Moggridgea tingle (Migidae)from southwestern Australia: evidence for long-term refugia within refugia.Mol. Ecol. 20, 3219–3236.

Decae, A., Cardoso, P., 2005. Iberesia, a new genus of trapdoor spiders (Araneae,Nemesiidae) from Portugal and Spain. Rev. Ibérica Aracnol. 12, 3–11.

Decae, A., Colombo, M., Manunza, B., 2014. Species diversity in the supposedlymonotypic genus Amblyocarenum Simon, 1892, with the description of a newspecies from Sardinia (Araneae, Mygalomorphae, Cyrtaucheniidae).Arachnology 16, 228–240.

Derkarabetian, S., Hedin, M., 2014. Integrative taxonomy and species delimitation inharvestmen: A revision of the western North American genus Sclerobunus(Opiliones: Laniatores: Travunioidea). PLoS One 9, e104982.

Donnelly, R.K., Harper, G.L., Morgan, A.J., Orozco-Terwengel, P., Pinto-Juma, G.A.,Bruford, M.W., 2013. Nuclear DNA recapitulates the cryptic mitochondriallineages of Lumbricus rubellus and suggests the existence of cryptic species in anecotoxological soil sentinel. Biol. J. Linn. Soc. 110, 780–795.

Goloboff, P.A., 1995. A revision of the South American spiders of the familyNemesiidae (Araneae, Mygalomorphae). Part I: species from Peru, Chile,Argentina, and Uruguay. Bull. Am. Mus. Nat. Hist. 224, 1–189.

Graham, M.R., Hendrixson, B.E., Hamilton, C.A., Bond, J.E., 2015. Mioceneextensional tectonics explain ancient patterns of diversification among turret-building tarantulas (Aphonopelma mojave group) in the Mojave and Sonorandeserts. J. Biogeogr. 42, 1052–1065.

Grummer, J.A., Bryson Jr., R.W., Reeder, T.W., 2014. Species delimitation using Bayesfactors: simulations and application to the Sceloporus scalaris species group(Squamata: Phrynosomatidae). Syst. Biol. 63, 119–133.

Hamilton, C.A., Formanowicz, D.R., Bond, J.E., 2011. Species delimitation andphylogeography of Aphonopelma hentzi (Araneae, Mygalomorphae,Theraphosidae): cryptic diversity in North American tarantulas. PLoS One 6,e26207.

Hamilton, C.A., Hendrixson, B.E., Brewer, M.S., Bond, J.E., 2014. An evaluation ofsampling effects on multiple DNA barcoding methods leads to an integrativeapproach for delimiting species: a case study of the North American tarantulagenus Aphonopelma (Araneae, Mygalomorphae, Theraphosidae). Mol.Phylogenet. Evol. 71, 79–93.

Harrington, R.C., Near, T.J., 2012. Phylogenetic and coalescent strategies ofspecies delimitation in snubnose darters (Percidae: Etheostoma). Syst. Biol.61, 63–79.

Harvey, M.S., Rix, M.G., Framenau, V.W., Hamilton, Z.R., Johnson, M.S., Teale, R.J.,Humphreys, G., Humphreys, W.F., 2011. Protecting the innocent: studying

D.H. Leavitt et al. / Molecular Phylogenetics and Evolution 91 (2015) 56–67 67

short-range endemic taxa enhances conservation outcomes. Invertebr. Syst. 25,1–10.

Hebert, P., Cywinska, A., Ball, S., DeWaard, J., 2003a. Biological identificationsthrough DNA barcodes. Philos. Trans. R. Soc. Lond. B 270, 313–321.

Hebert, P., Ratnasingham, S., de Waard, J.R., 2003b. Barcoding animal life:cytochrome c oxidase subunit 1 divergences among closely related species.Philos. Trans. R. Soc. B 270, S96–S99.

Hedin, M., 2015. High-stakes species delimitation in eyeless cave spiders (Cicurina,Dictynidae, Araneae) from central Texas. Mol. Ecol. 24, 346–361.

Hedin, M., Bond, J.E., 2006. Molecular phylogenetics of the spider infraorderMygalomorphae using nuclear rRNA genes (18S and 28S): conflict andagreement with the current system of classification. Mol. Phylogenet. Evol.41, 454–471.

Hedin, M., Carlson, D., 2011. A new trapdoor spider species from the southern CoastRanges of California (Mygalomorphae, Antrodiaetidae, Aliatypus coylei, nov sp),including consideration of mitochondrial phylogeographic structuring. Zootaxa2963, 55–68.

Hedin, M., Starrett, J., Hayashi, C., 2013. Crossing the uncrossable: novel trans-valleybiogeographic patterns revealed in the genetic history of low dispersalmygalomorph spiders (Antrodiaetidae, Antrodiaetus) from California. Mol.Ecol. 22, 508–526.

Hedin, M., Carlson, D., Coyle, F., in press. Sky island diversification meets themultispecies coalescent: divergence in the spruce-fir moss spider (Microhexuramontivaga) on the highest peaks of southern Appalachia. Mol. Ecol.

Hendrixson, B.E., DeRussy, B.M., Hamilton, C.A., Bond, J.E., 2013. An exploration ofspecies boundaries in turret-building tarantulas of the Mojave Desert (Araneae,Mygalomorphae, Theraphosidae, Aphonopelma). Mol. Phylogenet. Evol. 66, 327–340.

Highton, R., 2014. Detecting cryptic species in phylogeographic studies: Speciationin the California Slender Salamander, Batrachoseps attenuatus. Mol. Phylogenet.Evol. 71, 127–141.

Hillis, D.M., Bull, J.J., 1993. An empirical test of bootstrapping as a method forassessing confidence in phylogenetic analysis. Syst. Biol. 42, 182–192.

Irwin, D.E., 2002. Phylogeographic breaks without geographic barriers to gene flow.Evolution 56, 2383–2394.

Ji, Y.J., Zhang, D.X., He, L.J., 2003. Evolutionary conservation and versatility of a newset of primers for amplifying the ribosomal internal transcribed spacer regionsin insects and other invertebrates. Mol. Ecol. Notes 3, 581–585.

Jones, G., Aydin, Z., Oxelman, B., 2015. DISSECT: an assignment-free Bayesiandiscovery method for species delimitation under the multispecies coalescent.Bioinformatics 31, 991–998.

Katoh, K., Kuma, K.-I., Toh, H., Miyata, T., 2005. MAFFT version 5: improvement inaccuracy of multiple sequence alignment. Nucleic Acids Res. 33, 511–518.

Keith, R., Hedin, M., 2012. Extreme mitochondrial population subdivision insouthern Appalachian paleoendemic spiders (Araneae: Hypochilidae:Hypochilus), with implications for species delimitation. J. Arachnol. 40, 167–181.

Kuo, C.H., Avise, J.C., 2005. Phylogeographic breaks in low-dispersal species: theemergence of concordance across gene trees. Genetica 124, 179–186.

Leavitt, S.D., Fankhauser, J.D., Leavitt, D.H., Porter, L.D., Johnson, L.A., St Clair, L.L.,2011. Complex patterns of speciation in cosmopolitan ‘‘rock posy’’ lichens–Discovering and delimiting cryptic fungal species in the lichen-formingRhizoplaca melanophthalma species-complex (Lecanoraceae, Ascomycota). Mol.Phylogenet. Evol. 59, 587–602.

Maddison, W.P., 1997. Gene trees in species trees. Syst. Biol. 46, 523–536.Martínez-Solano, I., Jockusch, E.L., Wake, D.B., 2007. Extreme population

subdivision throughout a continuous range: phylogeography of Batrachosepsattenuatus (Caudata: Plethodontidae) in western North America. Mol. Ecol. 16,4335–4355.

Masta, S.E., Boore, J.L., 2008. Parallel evolution of truncated transfer RNA genes inarachnid mitochondrial genomes. Mol. Phylogenet. Evol. 25, 949–959.

Mayr, E., 1942. Systematics and the Origin of Species. Columbia University Press,New York, New York.

McGuire, J.A., Linkem, C.W., Koo, M.S., Hutchison, D.W., Lappin, A.K., Orange, D.I.,Lemos-Espinal, J.A., Riddle, B.R., Jaeger, J.R., 2007. Mitochondrial introgressionand incomplete lineage sorting through space and time: phylogenetics ofcrotaphytid lizards. Evolution 61, 2879–2897.

Miller, M.A., Holder, M.T., Vos, R., Midford, P.R., Liebowitz, T., Chan, L., Hoover, P.,Warnow, T., 2011. The CIPRES Portals. CIPRES. <http://www.phylo.org/sub_sections/portal>.

Moritz, C., Faith, D.P., 1998. Comparative phylogeography and the identification ofgenetically divergent areas for conservation. Mol. Ecol. 7, 419–429.

Ogden, R., Thorpe, R.S., 2002. Molecular evidence for ecological speciation intropical habitats. Proc. Natl. Acad. Sci. 99, 13612–13615.

Opatova, V., Arnedo, M.A., 2014a. From Gondwana to Europe: inferring the originsof Mediterranean Macrothele spiders (Araneae: Hexathelidae) and the limits ofthe family Hexathelidae. Invertebr. Syst. 28, 361–374.

Opatova, V., Arnedo, M.A., 2014b. Spiders on a hot volcanic roof: colonisationpathways and phylogeography of the Canary Islands endemic trap-door spiderTitanidiops canariensis (Araneae, Idiopidae). PLoS One 9, e115078.

Opatova, V., Bond, J.E., Arnedo, M.A., 2013. Ancient origins of the Mediterraneantrap-door spiders of the family Ctenizidae (Araneae, Mygalomorphae). Mol.Phylogenet. Evol. 69, 1135–1145.

Pante, E., Puillandre, N., Viricel, A., Arnaud-Haond, S., Aurelle, D., Castelin, M.,Chenuil, A., Destombe, C., Forcioli, D., Valero, M., Viard, F., Samadi, S., 2015.Species are hypotheses: avoid connectivity assessments based on pillars ofsand. Mol. Ecol. 24, 525–544.

Pons, J., Barraclough, T.G., Gomez-Zurita, J., Cardoso, A., Duran, D.P., Hazell, S.,Kamoun, S., Sumlin, W.D., Vogler, A.P., 2006. Sequence-based speciesdelimitation for the DNA taxonomy of undescribed insects. Syst. Biol. 55,595–609.

Posada, D., 2008. JModelTest: phylogenetic model averaging. Mol. Biol. Evol. 25,1253–1256.

Puillandre, N., Lambert, A., Brouillet, S., Achaz, G., 2012. ABGD, Automatic BarcodeGap Discovery for primary species delimitation. Mol. Ecol. 21, 1864–1877.

Rambaut, A., Drummond, A.J., 2007. Tracer v1.5. <http://beast.bio.ed.ac.uk/Tracer>Raven, R.J., 1985. The spider infraorder Mygalomorphae (Araneae): cladistics and

systematics. Bull. Am. Mus. Nat. Hist. 182, 1–180.Raven, R.J., 1987. A new mygalomorph spider genus from Mexico (Nemesiinae,

Nemesiidae, Arachnida). J. Arachnol. 14, 357–362.Redak, R.A., 2000. Arthropods and multispecies habitat conservation plans: are we

missing something? Environ. Manage. 26, 97–107.Satler, J.D., Carstens, B.C., Hedin, M., 2013. Multilocus species delimitation in a

complex of morphologically conserved trapdoor spiders (Mygalomorphae,Antrodiaetidae, Aliatypus). Syst. Biol. 62, 805–823.

Schuchert, P., 2014. High genetic diversity in the hydroid Plumularia setacea: amultitude of cryptic species or extensive population subdivision? Mol.Phylogenet. Evol. 76, 1–9.

Schwenk, K., Brede, N., Streit, B., 2008. Introduction. Extent, processes andevolutionary impact of interspecific hybridization in animals. Philos. Trans. R.Soc. B 363, 2805–2811.

Simon, E., 1891. Etudes arachnologiques. 23e Mémoire. XXXVIII. Descriptionsd’espèces et de genres nouveaux de la famille des Aviculariidae. Ann. Soc.Entomol. France 60, 300–312.

Stamatakis, A., Hoover, P., Rougemont, J., 2008. A rapid bootstrap algorithm for theRAxML web-servers. Syst. Biol. 75, 758–771.

Starrett, J., Hedin, H., 2007. Multilocus genealogies reveal multiple cryptic speciesand biogeographic complexity in the California turret spider Antrodiaetus riversi(Mygalomorphae, Antrodieatidae). Mol. Ecol. 16, 583–604.

Starrett, J., Hedin, M., Ayoub, N., Hayashi, C.Y., 2013. Hemocyanin gene familyevolution in spiders (Araneae), with implications for phylogenetic relationshipsand divergence times in the infraorder Mygalomorphae. Gene 524, 175–186.

Stockman, A.K., Bond, J.E., 2007. Delimiting cohesion species: extreme populationstructuring and the role of ecological interchangeability. Mol. Ecol. 16, 3374–3392.

Thomson, R.C., Wang, I.J., Johnson, J.R., 2010. Genome-enabled development of DNAmarkers for ecology, evolution and conservation. Mol. Ecol. 19, 2184–2195.

Ubick, D., Ledford, J.M., 2005. Nemesiidae. In: Ubik, D., Paquin, P., Cushing, P.E., Roth,V. (Eds.), Spiders of North America: An Identification Manual. AmericanArachnological Society, pp. 52–53.

Ubick, D., Paquin, P., Cushing, P.E., Roth, V. (Eds.), 2005. Spiders of North America:An Identification Manual. American Arachnological Society, 377 pages.

Wilson, J.S., Gunnell, C.F., Wahl, D.B., Pitts, J.P., 2013. Testing the species limits of thetarantulas (Araneae: Theraphosidae) endemic to California’s Southern CoastRanges, USA. Insect Conserv. Divers. 6, 365–371.

World Spider Catalog, 2015. World Spider Catalog. Natural History Museum Bern.<http://wsc.nmbe.ch>, version 16 (accessed 02.02.15).