bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. 1 Phylogenetic systematics and evolution of the spider infraorder Mygalomorphae using genomic 2 scale data 3 Vera Opatova1, Chris A. Hamilton2, Marshal Hedin3, Laura Montes de Oca4, Jiří Král5, 4 Jason E. Bond1 5 6 1 Department of Entomology and Nematology, University of California, Davis, CA 95616, USA 7 2 Department of Entomology, Plant Pathology & Nematology, University of Idaho, Moscow, ID 8 83844, USA 9 3 Department of Biology, San Diego State University, San Diego, CA, 92182–4614, USA 10 4 Departamento de Ecología y Biología Evolutiva, Instituto de Investigaciones Biológicas 11 Clemente Estable, Montevideo, Uruguay. 12 5 13 128 44, Czech Republic Department of Genetics and Microbiology, Faculty of Sciences, Charles University, Prague, 14 15 16 Corresponding authors: 17 Vera Opatova, 1282 Academic Surge, One Shields Avenue, Davis, CA 95616 18 Telephone: +1 530-754-5805, E-mail: vopatova@ucdavis.edu 19 20 Jason E. Bond, 1282 Academic Surge, One Shields Avenue, Davis, CA 95616 21 Telephone: +1 530-754-5805, E-mail: jbond@ucdavis.edu 22 23 24 Running Head: PHYLOGENY OF MYGALOMORPH SPIDERS bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 25 ABSTRACT 26 27 The Infraorder Mygalomorphae is one of the three main lineages of spiders comprising over 28 3,000 nominal species. This ancient group has a world-wide distribution that includes among its 29 ranks large and charismatic taxa such as tarantulas, trapdoor spiders, and highly venomous 30 funnel web spiders. Based on past molecular studies using Sanger-sequencing approaches, 31 numerous mygalomorph families (e.g., Hexathelidae, Ctenizidae, Cyrtaucheniidae, Dipluridae 32 and Nemesiidae) have been identified as non-monophyletic. However, these data were unable to 33 sufficiently resolve the higher-level (intra- and interfamilial) relationships such that the 34 necessary changes in classification could be made with confidence. Here we present a 35 comprehensive phylogenomic treatment of the spider infraorder Mygalomorphae. We employ 36 472 loci obtained through Anchored Hybrid Enrichment to reconstruct relationships among all 37 the mygalomorph spider families and estimate the timeframe of their diversification. We 38 sampled all currently recognized families, which has allowed us to assess their status, and as a 39 result, propose a new classification scheme. Our generic-level sampling has also provided an 40 evolutionary framework for revisiting questions regarding silk use in mygalomorph spiders. The 41 first such analysis for the group within a strict phylogenetic framework shows that a sheet web is 42 likely the plesiomorphic condition for mygalomorphs, as well as providing hints to the ancestral 43 foraging behavior for all spiders. Our divergence time estimates, concomitant with detailed 44 biogeographic analysis, suggest that both ancient continental-level vicariance and more recent 45 dispersal events have played an important role in shaping modern day distributional patterns. 46 Based on our results, we relimit the generic composition of the Ctenizidae, Cyrtaucheniidae, 47 Dipluridae and Nemesiidae. We also elevate five subfamilies to family rank: Anamidae (NEW 2 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 48 RANK), Euagridae (NEW RANK), Ischnothelidae (NEW RANK), Pycnothelidae (NEW 49 RANK), and Bemmeridae (NEW RANK). The three families Entypesidae (NEW FAMILY), 50 Microhexuridae (NEW FAMILY), and Stasimopidae (NEW FAMILY) are newly proposed. 51 Such a major rearrangement in classification, recognizing eight newly established family-level 52 rank taxa, is the largest the group has seen in over three decades since Raven’s (1985) taxonomic 53 treatment. 54 55 Keywords: Biogeography, Molecular clocks, Phylogenomics, Spider web foraging, Taxonomy 56 57 Spiders placed in the infraorder Mygalomorphae are a charismatic assemblage of taxa 58 that includes among its ranks the tarantulas, trapdoor spiders, and also some of the most 59 venomous species, like the Sydney funnel web spider and its close relatives (Hedin et al. 2018). 60 Today, the infraorder comprises 20 families, divided into 354 genera, with 3,006 species (World 61 Spider Catalog 2018; Table 1). The group is ancient, known to occur in the fossil record since 62 the Middle Triassic (Selden and Gall 1992), but with origins estimated further back into the 63 Carboniferous over 300 million years ago (Ayoub et al. 2007; Starrett et al. 2013; Garrison et al. 64 2016). Their ancient origins and concomitant sedentary nature provide a rich biogeographical 65 context for studying geological-scale continental vicariant events (Raven 1980; Hedin and Bond 66 2006; Opatova et al. 2013; Opatova and Arnedo 2014a). Not surprisingly for such an ancient 67 lineage, mygalomorph spider morphology is complicated; they have retained a number of 68 features considered primitive for spiders, such as a simple silk spinning system, two pairs of 69 book lungs and paraxial cheliceral arrangement, yet are also relatively homogenous when 70 compared to the overall morphological diversity observed among their araneomorph counterparts 3 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 71 (Hendrixson and Bond 2009). Although such striking homogeneity presents significant 72 challenges for morphology-based taxonomy, mygalomorphs have become a noteworthy system 73 for studying allopatric speciation and species crypsis (Bond et al. 2001; Bond and Stockman 74 2008; Satler et al. 2013; Hamilton et al. 2014; Opatova and Arnedo 2014b; Hedin et al. 2015; 75 Leavitt et al. 2015; Castalanelli et al. 2017; Rix et al. 2017b; Starrett et al. 2018). 76 For less than two decades, since the first application of molecular systematics to 77 questions regarding mygalomorph phylogeny (Hedin and Bond 2006), it has been generally 78 acknowledged that the current system of classification is replete with problems that include 79 numerous instances of family-level para and/or polyphyly. In short, much of today’s 80 classification scheme dates back to Raven (1985) Fig. 1a, which first applied cladistic thinking 81 towards evaluating family-level relationships on the basis of morphological character 82 argumentation. Post Raven (1985), a number of authors continued to explore morphological 83 characters as evidence for broad scale familial-level relationships and identity (Eskov and 84 Zonshtein 1990), further facilitated by the availability of computational phylogenetic inference 85 (Goloboff 1993; Goloboff 1995; Bond and Opell 2002; Fig. 1b). These morphological studies 86 were followed shortly thereafter by a number of early molecular treatments (Hedin and Bond 87 2006; Ayoub et al. 2007; Fig. 2a). The first major taxonomic changes to Raven’s classification 88 followed in 2012 (Bond et al. 2012; Fig. 2b) with the elevation of the subfamily Euctenizinae to 89 family status. A number of subsequent studies employing Sanger-sequencing based approaches 90 continued to make progress in exploring relationships between closely related taxa (Hamilton et 91 al. 2014; Opatova et al. 2016; Mora et al. 2017; Rix et al. 2017a; Ortiz et al. 2018; Starrett et al. 92 2018) and within families (Opatova et al. 2013; Opatova and Arnedo 2014a; Harrison et al. 93 2017; Rix et al. 2017c; Harvey et al. 2018; Lüddecke et al. 2018; Turner et al. 2018). Although 4 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 94 Sanger approaches have been fruitful in terms of advancing our understanding of some 95 relationships, it has been long recognized that ribosomal RNA genes (e.g., 28S and 18S), 96 elongation factors, and mitochondrial genes, are often of limited utility in the context of deep 97 scale divergences of mygalomorph spiders (Hedin and Bond 2006; Ayoub et al. 2007; Bond et 98 al. 2012). 99 Within the last several years, the application of phylogenomic approaches has helped to 100 transform spider systematics (Bond et al. 2014; Fernández et al. 2014; Garrison et al. 2016; 101 Hamilton et al. 2016b; Hedin et al. 2018; Kuntner et al. 2018; Wood et al. 2018; Hedin et al. 102 2019), by providing the data and impetus for making major changes in mygalomorph spider 103 classification. Specifically, targeted capture approaches using anchored hybrid enrichment 104 (AHE; Lemmon et al. 2012) and ultra-conserved elements (UCE; Faircloth et al. 2012) are now 105 being applied to questions at multiple hierarchical levels across the spider tree of life (Hamilton 106 et al. 2016a; Hamilton et al. 2016b; Maddison et al. 2017; Chamberland et al. 2018; Godwin et 107 al. 2018; Hedin et al. 2018) and other arachnids (Starrett et al. 2017; Sharma et al. 2018). 108 Implementation of genomic-based methods seemingly overcomes the perceived limitations of the 109 traditional Sanger loci, enabling mygalomorph systematists to formally address a number of 110 long-standing issues like the polyphyletic nature of ctenizids and hexathelids (Godwin et al. 111 2018; Hedin et al. 2018). In both cases, phylogenomic results stabilized the taxonomic position 112 of genera that were notoriously challenging to place and established new familial ranks 113 (Halonoproctidae (Godwin et al. 2018), Atracidae, Macrothelidae and Porrhothelidae (Hedin et 114 al. 2018)). Although these smaller scale studies have made important advances in our 115 understanding of mygalomorph phylogeny and classification, they have all thus far lacked the 5 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 116 broad scale phylogenetic context instrumental to understanding relationships spanning the entire 117 infraorder. 118 We present here the most comprehensive phylogenomic treatment of Mygalomorphae 119 conducted to date. Our analysis includes broad sampling of taxa from all currently recognized 120 families (88 genera, 111 species), allowing us to assess the monophyly of a number of key 121 families (e.g., Cyrtaucheniidae, Nemesiidae, Dipluridae, Ctenizidae) and address some long- 122 standing taxonomic problems. Our generic-level sampling also provides an evolutionary 123 framework for revisiting questions regarding silk use in non-araneomorph spiders. We explore 124 questions regarding the timing and biogeographical context of mygalomorph spider 125 diversification, and explicitly test a number of hypotheses related to the timing and origin of 126 major mygalomorph lineages, the use of silk in mygalomorph prey capture, and the monophyly 127 of the major mygalomorph higher-level groups and families. 128 129 MATERIALS & METHODS 130 131 We sampled a total of 113 taxa, 111 ingroup specimens representing 88 genera from all 132 20 currently recognized mygalomorph families (Table 1) and two representatives of the suborder 133 Mesothelae (Liphisthius and Vinathela, family Liphistiidae) as outgroups to root phylogenies. 134 Our sampling represents approximately 25% of the infraorder’s known generic diversity and 135 about 47% if the two most diverse mygalomorph families, Barychelidae and Theraphosidae, are 136 excluded from the calculation. Approximately 60% of AHE sequence data was newly generated 137 for this study, additional sequences from Hamilton et al. (2016b) and Godwin et al. (2018) were 138 included. Taxon sampling is summarized in Supplemental Table S1; all statistics related to 6 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 139 family and generic-level species diversity are extracted from the World Spider Catalog 140 (https://wsc.nmbe.ch/) accessed November 2018. 141 Whole genomic DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen), 142 following the manufacturer’s guidelines; to ensure the complete digestion of RNA, RNase A was 143 added to the mix after the tissue lysis step. Library preparation, enrichment, and sequencing were 144 performed at the Center for Anchored Phylogenomics at Florida State University 145 (http://anchoredphylogeny.com/) following the methods described in Lemmon et al. (2012), 146 Hamilton et al. (2016b) and Godwin et al. (2018). During the development of the spider AHE 147 loci, a mixture of spider genome sequencing and transcriptomes were utilized to build scaffolds 148 from which probes were designed in conserved anchor regions neighboring variable flanking 149 regions, creating a set of probes that include exons, introns, intergenic, and conserved regions of 150 the genome. Initial probe design was based on the core arthropod orthologs and the targeted loci 151 were then found in the spider reference genomic datasets (both spider genomes and 152 transcriptomes). Loci that did not contain both Aliatypus (Antrodiaetidae) and Aphonopelma 153 (Theraphosidae) were not considered for further usage. After sequencing, all reads were 154 assembled into contigs following Prum et al. (2015), but using references derived from the 155 Ixodes (a tick), Hypochilus (araneomorph), Aphonopelma, and Aliatypus (mygalomorphs) 156 sequences used for probe design. Homologous nucleotide sequence sets were produced by 157 grouping filtered consensus sequences across individuals by target locus. Orthologous groups 158 were then determined for each target locus by a) employing a pairwise-distance measure 159 computed for pairs of homologs; b) a neighbor-joining clustering algorithm to cluster the 160 consensus sequences into orthologous sets; c) the identified sequences are then grouped into a 161 cluster; d) steps A-C are repeated, treating each cluster as a sequence by using the average 7 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 162 distance between the cluster and other clusters/sequence when assessing distance. This approach 163 efficiently divides the homologs into two or more orthologous sets, allowing us to identify 164 duplications occurring before the ancestor of the taxonomic group being analyzed. For 165 duplications occurring within the taxonomic group under investigation, this approach tends to 166 produce a single cluster containing one sequence per taxon and a second cluster containing 167 sequences from those taxa that are descendants from the ancestor in which the duplication 168 occurred. 169 Following the targeting of the 585 loci within the Spider Probe Kit v1 (Hamilton et al. 170 2016b), 472 loci were recovered and assembled. All loci were aligned with MAFFT 7.402 171 (Katoh and Standley 2013) using the L-INS-i algorithm (--localprior and --maxiterate 1000 172 flags). The alignments were scored for accuracy in ALISCORE (Misof and Misof 2009) and 173 ambiguously aligned positions were removed with ALICUT (Kück 2009). Individual loci were 174 concatenated with FASconCAT (Kück and Longo 2014) yielding a supermatrix of 93,410 175 nucleotides for 111 ingroup and two outgroup taxa (hereafter “All_taxa”); all alignments were 176 visually inspected in Geneious 10.1.3 (Biomatters 2017) prior to concatenation. 177 Two additional matrices were constructed to assess the effect of taxon removal or 178 sequence data addition on tree topology and support values. First, terminals representing the 179 family Paratropididae were removed from the “All_taxa” (hereafter referred to as 180 “No_Paratropis”). The family was recovered as a “stand-alone” lineage with unresolved and 181 unstable placement (Hedin and Bond 2006; Bond et al. 2012), potentially causing conflict in 182 topology and lowering the support of deeper nodes. To test the effect of data addition on node 183 support and the unresolved placement of Stasimopidae, we combined our AHE loci with pre- 184 existing transcriptomic data (Garrison et al. 2016). To maximize the taxon sampling, we 8 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 185 extracted the mygalomorph terminals from “BCC-75” matrix (Garrison et al. 2016) and 186 combined them into the “DNAAA_matrix” with corresponding terminals represented in our 187 AHE dataset, in two instances, forming chimaeras at generic level (Supplemental Table S2). In 188 order to maintain a relevant phylogenetic framework, we also retained additional individuals 189 representing key lineages/clades that were only represented by AHE data. Terminals extracted 190 from the “BCC-75” matrix (Garrison et al. 2016) were parsed into individual loci using 191 BIOPYTHON (Hamelryck and Manderick 2003; Cock et al. 2009). Resulting 1675 192 transcriptomic loci were blasted (tblastn command) in Geneious (Biomatters 2017) against a 193 custom AHE locus database comprising nucleotide sequence data prior to ALICUT treatment. 194 Amino acid loci that received an e-value of < 1e-10; that is, the loci already represented in AHE 195 dataset, were excluded from the analyses. Loci that did not contain all 20 amino acids were 196 further removed from the dataset (per Stamatakis 2014). 197 Phylogenetic analyses were run on the Hopper Community Cluster at Auburn University. 198 For the DNA matrices, partition schemes under the GTR substitution model were evaluated and 199 selected based on AICc criterion in PartionFinder 2 (Lanfear et al. 2016) using the rcluster 200 algorithm (Lanfear et al. 2014) with RAxML implementation (Stamatakis 2014). Partitioning by 201 codon was not considered in the analyses because ambiguously aligned positions were removed 202 with ALICUT. Partition schemes and models of protein evolution for the amino acid data were 203 inferred independently using PartionFinder 2 with rcluster algorithm prior to their concatenation 204 with the nucleotide AHE sequences into the “DNAAA_matrix”. 205 Maximum likelihood analyses (ML) were conducted in RAxML v 8.2.9 (Stamatakis 206 2014). Both “All_taxa” and “No_Paratropis” datasets were analyzed using the 191-partition 207 scheme and an independent GTR+G model defined for each partition. The best ML tree for each 9 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 208 matrix was selected from 1000 iterations, each starting from an independently derived 209 parsimony-based tree. Bootstrap support was inferred from 50 replicates, determined as 210 sufficient by the automatic bootstopping criterion (- N autoMRE flag) for both datasets 211 (Pattengale et al. 2010). Additional ML analysis of “All_taxa” dataset was also conducted in IQ- 212 TREE (Nguyen et al. 2014), model selection for each partition was performed by with the 213 ModelFinder Plus; support values were inferred via SH-aLRT test (Anisimova et al. 2011) and 214 ultrafast bootstrapping (both 1000 replicates). 215 The “DNAAA_matrix” mixed dataset combining both nucleotide and amino acid data 216 was analyzed with 1367-partition scheme, applying an independent GTR+G model for each 217 AHE partition and a corresponding protein model inferred by PartionFinder to each 218 transcriptome partition. The best ML tree was selected from 1000 iterations and support was 219 assessed from 100 bootstrap replicates (autoMRE). 220 Statistical significance of alternative tree topologies; that is, different, or traditional, 221 family-level arrangements (Table 2) than those recovered in our analyses, was assessed with 222 Approximately Unbiased (AU) topology test (Shimodaira 2002) implemented in CONSEL 223 (Shimodaira and Hasegawa 2001). Alternative tree topologies were obtained from 1000 224 independent constrained searches conducted in RAxML implementing the same settings, 225 partition scheme, and nucleotide substitution model as described above. 226 Bayesian Inference (BI) analyses of the “All_taxa” dataset was conducted in ExaBayes v 227 1.4.1 (Aberer et al. 2014) implementing the same partition scheme and nucleotide substitution 228 model as in the ML analyses. Two independent runs of 4x107 generations with four coupled 229 chains each, starting from a parsimony tree with resampling every 1000 generations, were run 230 simultaneously. Standard deviation of split frequencies was monitored (<0.01), and the first 25% 10 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 231 were discarded as a burn-in for the analyses. An extended majority rule consensus tree was 232 obtained in the ExaBayes accompanying program consense (Aberer et al. 2014). 233 A species tree was inferred from 472 gene trees using ASTRAL v 4.11.2 (Mirarab and 234 Warnow 2015). Node support was assessed using ASTRAL’s local posterior probabilities. Single 235 gene trees were inferred from AHE nucleotide alignments implementing GTR+G model in 236 RAxML; the best ML trees were selected from 1000 independent iterations for each locus 237 individually. 238 Divergence times were estimated in the penalized likelihood environment (Sanderson 239 2002) using the treePL algorithm (Smith and O’Meara 2012) on the tree topology obtained in 240 RAxML, and calibrated with six fossils. The optimal settings for the thorough analysis was 241 determined with the prime option in treePL; smoothing value = 1, determined by random 242 subsample and replicate cross-validation method (RSRCV), was used for both dating analysis 243 and bootstrapping (see below). The calibrations (shown in, Supplemental Fig.7) are as follows: 244 1) Mygalomorphae: Due to the lack of relevant fossil calibration points for the Mygalomorphae – 245 Mesothelae split, we used the information available from phylogenomic studies based on 246 extensive outgroup fossil calibrations (Garrison et al. 2016; Fernández et al. 2018) in order to 247 constrain the age of the root node. Namely, we used the 95% confidence intervals estimated for 248 this split as the minimum (287 Ma) and maximum bounds (398 Ma); 2) Avicularioidea: The age 249 of the oldest mygalomorph fossil Rosamygale grauvogely (Selden and Gall 1992) from Gres-a- 250 Voltzia Formation, Vosges, France dated to Middle Triassic (Anisian) 242 Million years ago 251 (Ma) was assigned as the minimum bound for the Avicularioidea – Atypoidea split. The fossil 252 has been tentatively assigned to the family “Hexathelidae” (sensu Raven) but given the 253 polyphyly of the family (Hedin et al. 2018), we rather interpret Rosamygale as an Avicularioidea 11 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 254 crown-group. The maximum bound for the Avicularioidea – Atypoidea split was assigned the 255 age of 323 Ma, corresponding to the oldest age of the Bashkirian stage (Pennsylvanian, 256 Carboniferous) that yielded the oldest spider fossil Arthrolycosa sp. putatively assigned as 257 Mesothelae stem-group (Selden et al. 2014; Garwood et al. 2016). This age also overlaps with 258 the oldest boundary of the 95% confidence intervals for the Avicularioidea – Atypoidea split 259 estimated by recent phylogenomic studies (Garrison et al. 2016; Fernández et al. 2018); 3) 260 “Nemesioidina” clade: The age of Cretamygale chasei 125 Ma (Selden 2002) from Cretaceous 261 (Barremian, Hughes and McDougall 1990) currently classified as Nemesiidae was 262 conservatively assigned as the minimum bound for the split between Bemmeridae/Theraphosidae 263 and “Nemesioidina” clades. The maximum bound was assigned 242 Ma, corresponding to the 264 age of Rosamygale fossil. 4) Atypoidea: Atypidae – Antrodiaetidae split was assigned minimum 265 bound of 100 Ma representing Ambiortiphagus ponomarenkoi (Eskov and Zonshtein 1990) from 266 the Lower Cretaceous (Albian) proceeding from Dund-Ula Mountain Ridge in Bayan-Khongor 267 Province, Central Mongolia. The age of Rosamygale (242 Ma) was assigned as a maximum 268 bound for the split; and 5) Ummidia fossils (44 Ma) from Baltic amber (Wunderlich 2011) were 269 used to calibrate the minimum bound for the split between the North American and European 270 species of Ummidia. Maximum bound of 90 Ma corresponds to the oldest boundary of the 95% 271 confidence interval estimated for this particular split (Opatova et al. 2013). The confidence 272 intervals for node ages were assessed by dating 100 ML bootstrap phylograms inferred in 273 PAUP* (Swofford 2003). The trees were dated in TreePL using the same settings as described 274 above; TreeAnnotator v. 2.4.5 (Bouckaert et al. 2018) was used for calculating the age statistics 275 for the nodes. 12 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 276 Biogeographic analyses were conducted in Reconstruct Ancestral State in Phylogenies 277 (RASP v 3.1, Yu et al. 2015) with the dispersal-extinction-cladogenesis (DEC) model (Ree and 278 Smith 2008) implemented in C++ version of Lagrange (Smith 2010). Analyses were run with no 279 dispersal constraints, assigning equal probabilities to dispersal events among all defined areas, 280 and allowing two unit areas in the ancestral distributions. The terminal taxa represented in the 281 tree were assigned to six distribution ranges: (A) North America; (B) Europe; (C) Asia; (D) 282 South America; (E) Africa; (F) Australia plus New Zealand (Fig. 5b). The consensus tree 283 obtained in ExaBayes was used as an input for the analyses. 284 For ancestral state reconstruction analyses of foraging construct, each terminal taxon was 285 scored using six character states: (0) burrow with trapdoor(s); (1) brush sheet/funnel web; (2) 286 burrow with collar door; (3) purse web; (4) open burrow; (5) turret. Character codings were 287 scored from personal observations and/or the primary literature as documented in Supplemental 288 Table S3. Taxa that forage as cursorial hunters were scored as unknown or missing (Coddington 289 et al. 2018); polymorphic character states were assigned to terminals for instances where species 290 build multiple/combination of constructs (e.g., some Cyrtauchenius terminals build turrets as 291 juveniles but cover those turrets with trapdoors as adults, Opatova persn. obsv.), or closely 292 related species build more than one construct e.g., both open burrows and trapdoors (some 293 bemmerids and pycnothelids) and both turrets and collars (e.g., some Antrodiaetus). Ancestral 294 character state reconstructions were conducted using the R package corHMM (Beaulieu et al. 295 2013) on an ultrametric scaled tree with node.states=marginal; corHMM was preferred over 296 other methods given its capacity to accommodate missing and polymorphic characters. The R 297 package ape was used to scale the preferred tree topology (‘chronopl’) with an assigned lambda 298 value of 0.1. Character optimizations using equal (ER), symmetric (SYR), and all rates different 13 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 299 (ARD) models were explored for these categorical data using ‘rayDISC’ (corHMM). The 300 preferred model was chosen by comparing AICc values calculated using corHMM. 301 All relevant data, DNA and amino acid sequence alignments, phylogenetic trees, and 302 miscellaneous data matrices, were deposited in the Dryad Data Repository 303 (doi:XXXXXXXXXXX) on XX XX 2019. 304 305 RESULTS 306 307 Concatenated analyses, all taxa.—The complete dataset “All_taxa” supermatrix 308 comprises 472 loci (93,410 nucleotides) for 113 terminals (Supplemental Table S1), whereas the 309 reduced “No_Paratropis” dataset comprises 111 terminals. The proportion of missing data in the 310 datasets is 14.7% and 14.5%, respectively. The combined “DNAAA_matrix” includes 30 311 terminals, with 14 individuals represented by both genomic (472 AHE loci) and transcriptomic 312 data (1519 loci proceeding from Garrison et al. (2016)). The total length of the 313 “DNAAA_matrix” is 305,860 characters; gaps and missing data represent 54.5% of the dataset. 314 For details see Supplemental Table S2. 315 Both ML (-ln 1728131.7475) and BI analyses performed on the “All_taxa” dataset 316 yielded similar tree topologies (Fig. 3; Supplemental Fig. 1, 2, 3). Overall, the resulting trees are 317 highly supported (bootstrap > 90, posterior probability (PP) > 0.95, SH-aLRT > 95), with the 318 exception of a few deeper nodes. The ML and BI analyses both recover two main clades 319 Atypoidea and Avicularioidea with strong support. The relationships within Atypoidea are fully 320 supported; the family Atypidae is recovered as a sister to Antrodiaetidae inclusive of Hexura 321 (Mecicobothriidae). 14 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 322 Four mygalomorph families belonging to Avicularioidea are recovered as para- or 323 polyphyletic (namely Dipluridae, Ctenizidae, Cyrtaucheniidae and Nemesiidae), prompting 324 extensive higher-level changes to mygalomorph classification (see below). Several lineages 325 formerly belonging to the families Dipluridae and Hexathelidae (see Hedin et al. (2018)) 326 successively branch off near the root node of Avicularioidea as follows: Ischnothelidae (NEW 327 RANK), Microhexuridae (NEW FAMILY), Hexathelidae, Euagridae (NEW RANK), 328 Porrhothelidae and Macrothelidae. The Bipectina clade (sensu Goloboff (1993)) inclusive of the 329 family Atracidae is recovered in our analysis. The family Paratropididae is recovered as sister to 330 all the remaining Bipectina. The remaining taxa are placed into a clade comprising Stasimopidae 331 and the “Venom Clade” (Atracidae + Actinopodidae) and two additional clades corresponding to 332 Domiothelina and Crassitarsae (sensu Bond et al. (2012), but excluding Stasimopus). The 333 monogeneric family Stasimopidae (NEW FAMILY, formerly Ctenizidae), is placed as sister to 334 “Venom Clade”, albeit with low support. The ML analysis performed in IQ-TREE resulted in 335 different placement of Stasimopidae and the “Venom Clade”. The “Venom Clade” is supported 336 as sister to all Bipectina minus Paratropididae, whereas the family Stasimopidae is recovered 337 with high SH-aLRT support as sister to Domiothelina. Domiothelina form a monophyletic group 338 with well resolved relationships in both ML and BI analyses. The family Halonoproctidae is 339 recovered as sister to all remaining lineages of Domiothelina. The family Migidae is recovered 340 as sister to Idiopidae, Ctenizidae (now comprising Cteniza and Cyrtocarenum) and Euctenizidae. 341 Crassitarsae (sensu Bond et al. (2012)), recovered with high SH-aLRT support in IQ-TREE ML 342 inference, is subdivided into two clades supported in all analyses. The first one comprises the 343 family Bemmeridae (NEW RANK, formerly Cyrtaucheniidae), recovered as sister to the 344 Barychelidae and Theraphosidae. The second clade, “Nemesioidina”, comprises the Nemesiidae 15 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 345 plus three lineages formerly belonging to the Nemesiidae (Pycnothelidae (NEW RANK), 346 Entypesidae (NEW FAMILY) and Anamidae (NEW FAMILY)), as well as Dipluridae, 347 Cyrtaucheniidae and Microstigmatidae. Nemesiidae is recovered as sister to all the remaining 348 families; Pycnothelidae is recovered as sister to the Dipluridae/Cyrtaucheniidae clade and the 349 Entypesidae /Microstigmatidae/Anamidae clade. 350 No_Paratropis analysis.—Maximum likelihood analyses of the “No_Paratropis” dataset 351 recovered a moderately different Bipectina topology contra the analyses on the full dataset 352 (Supplemental Fig. 4; ML best tree –ln = 1702659.2948). The discordance, however, mostly 353 affects nodes unsupported in both analyses. In the “No_Paratropis” analyses the clade 354 comprising Bemmeridae + Barychelidae + Theraphosidae is recovered with high support as 355 sister to all the remaining Bipectina families. The Stasimopidae + “Venom Clade” receive 356 moderate bootstrap support, but its sister relationship to the Domiothelina clade remains 357 unsupported, albeit with increased support when compared to the analyses of the full dataset. 358 DNAAA analysis.—Similar to the “No_Paratropis” dataset, analyses of “DNAAA” (ML 359 best tree –ln = 2543315.0197) resulted in different topological arrangements within the Bipectina 360 clade (Fig. 4). The Domiothelina clade is recovered as sister to all the remaining diversity. 361 Stasimopidae + “Venom Clade” (moderately supported in this analysis) is recovered with low 362 support as sister to Crassitarsae. The Crassitarsae clade is highly supported and inclusive of 363 Paratropis. 364 Topology tests.—The AU topology test (Table 2) rejected the monophyly of all 365 alternative arrangements (i.e., alternatives using existing compositions of genera) of the families 366 Ctenizidae (sensu Raven (1985)), Cyrtaucheniidae (sensu Bond et al. (2012)), Dipluridae, 16 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 367 Hexathelidae (sensu Raven (1985)) and Nemesiidae in favor of the results of the unconstrained 368 search. 369 Species tree analysis.—Coalescence-based analysis produced a species tree from the 370 individual 472 input gene trees obtained in RAxML. The resulting quartet-based supertree 371 estimated in ASTRAL (Supplemental Fig. 5) comprises 1,228,051,216 induced quartet trees 372 from the input gene trees, representing 65.2% of all quartets present in the species tree. ASTRAL 373 yielded a different topology than the concatenated approach (Supplemental Fig. 6). The analyses 374 recover both Atypoidea and Avicularioidea clades. Ischnothelidae (NEW RANK), 375 Microhexuridae (NEW FAMILY), Hexathelidae, Euagridae (NEW RANK) and Porrhothelidae, 376 a grade of lineages branching at the base of the Avicularioidea clade in the concatenated 377 analyses, form a highly supported clade in the species tree approach. The topology of the 378 Bipectina clade (unsupported in ASTRAL) is also somewhat differed from the concatenated 379 analyses. Bemmeridae/Barychelidae/Theraphosidae clade is recovered as sister to moderately 380 supported Domiothelina + Stasimopidae/“Venom Clade”, rendering the Crassitarsae 381 paraphyletic. 382 Molecular dating.—The resulting dated tree is depicted in Fig. 5; Supplemental Figures 383 7, 8. The age of the root (398 Ma) corresponds to the maximum age constraints assigned to this 384 split. The split between Atypoidea and Avicularioidea is dated at ~ 323 Ma. The early branching 385 events within Avicularioidea dated between 254 Ma [95% confidence interval (CI): 259 – 247] 386 and 174 Ma (179 – 162) (Ischnothelidae through Macrothelidae divergences). The first 387 diversification of Bipectina is dated to 166 Ma (171 – 155) (divergence of Paratropididae) and 388 continues at 158 Ma (163 – 147), with diversification into Domiothelina + 17 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 389 Stasimopidae/“Venom Clade” and Crassitarsae. The divergence time estimates for individual 390 families are reported in Table 3. 391 Historical biogeography.—The results of DEC analyses are summarized in Fig. 6. The 392 ancestral distributional area of the mygalomorph ancestor and the early branching Avicularioidea 393 lineages remain unresolved. The distribution of the Atypoidea ancestor is inferred to have 394 originated in the Northern hemisphere continents (North America and Asia); Bipectina likely 395 originated in Africa. The ancestral ranges of several Avicularioidea families are inferred to areas 396 spanning more than one continent, for example Africa and Australia (Migidae, Idiopidae, 397 Barychelidae, Microstigmatidae); Africa and South America (Cyrtaucheniidae, Pycnothelidae). 398 On the other hand, some Avicularioidea families experienced in situ diversification within a 399 single area. Stasimopidae, Bemmeridae, and Entypesidae likely evolved in Africa, Paratropididae 400 and Dipluridae in South America, Anamidae in Australia, Euctenizidae in North America and 401 Ctenizidae in Europe. DEC analyses also infers a large number of dispersal and vicariance events 402 across the mygalomorph phylogeny, however, most of the events receive low probabilities 403 (<0.7). 404 Evolution of foraging construct.—We scored foraging construct for 109 taxa. Taxa that 405 are cursorial hunters (three terminals), Paratropis and Microstigmata, were scored as 406 missing/inapplicable. The foraging construct of nemesiid Mexentypesa is unknown to us. 407 Because bemmerids and pycnothelids construct both trapdoors and open burrows they were 408 scored as polymorphic. Based on AICc values the equal rates model (ER) produced the preferred 409 reconstruction (AICc = 168.3399). Although the ER model was preferred, ancestral state 410 reconstructions for the other models were generally very similar. 18 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 411 The ancestral state reconstruction of mygalomorph foraging construct, summarized in 412 Fig. 7, shows a rather simple and concise pattern. Towards the root of the phylogeny there 413 appears to be an equal probability of an ancestor having either a funnel-and-sheet web or a 414 burrow with a trapdoor. Atypoids likely had a most recent common ancestor (MRCA) with a 415 funnel-and-sheet web, modified as a purse web in atypids or as a collar, turret, or trapdoor in 416 antrodiaetids. As discussed below, atypoids demonstrate the full scope of variability in burrow 417 construct modifications that can evolve from a sheet web ancestral form. Avicularioidea have a 418 MRCA that optimizes as having a sheet web. The MRCA for the Bipectina clade likely had 419 either a trapdoor or funnel-and-sheet web (equivocal). A burrow with a trapdoor more 420 definitively optimizes as the ancestral foraging construct for Crassitarsae with subsequent losses 421 of the trapdoor covering (optimized as the open burrow state) four times in nemesiids, 422 pycnothelids, theraphosids, and the euctenizid genus Apomastus; some pycnothelid and anamid 423 taxa independently lose the trapdoor and instead forage from burrows with distinct turret or 424 collar door constructs. Funnel-and-sheet webs are regained twice, once in atracids and again in 425 diplurids. 426 427 DISCUSSION 428 429 In light of the results reported herein and other past molecular based studies of 430 mygalomorph phylogeny, two points seem clear. First, morphological data alone are unsuitable 431 for resolving the evolutionary history of mygalomorphs. Second, if our aim is to have a 432 classification scheme that reflects evolutionary history, other characters are needed. The results 433 we discuss below allow us to assess the status of nearly all mygalomorph families and resolve 19 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 434 several long-standing questions in the group’s phylogeny and higher-level classification. As a 435 consequence, we propose extensive taxonomic changes that better reflect the origins and 436 evolutionary relationships of the taxa. We discuss in detail 1) the temporal and biogeographical 437 context of mygalomorph family-level diversification; 2) revisit longstanding questions regarding 438 silk use in non-araneomorph spiders; and 3) provide a detailed account regarding the taxonomic 439 implications of our newly formed phylogenetic hypothesis (Fig. 8). The Appendix to this paper 440 (Bond et al. 2019) outlines the formal taxonomic changes proposed herein and discusses in more 441 detail the systematics of each mygalomorph family. 442 443 Diversification and Biogeography of Mygalomorph Spiders 444 445 Mygalomorph spiders with nearly a cosmopolitan distribution, highly sedentary nature 446 (Raven 1985; Bond et al. 2001) and evolutionary history dating back to Carboniferous (Ayoub et 447 al. 2007; Starrett et al. 2013; Garrison et al. 2016), represent the type of group that has captivated 448 the attention of biogeographers for decades (Cracraft 1988; Wiley 1988; Morrone and Crisci 449 1995). By implementing divergence time estimation methods, many distribution patterns 450 observed in mygalomorphs could be linked to specific geologic events such as continental 451 breakups (Opatova et al. 2013; Opatova and Arnedo 2014a), Mediterranean tectonics (Opatova et 452 al. 2016; Mora et al. 2017) and opening of the Okinawa Trough (Su et al. 2016). However, recent 453 findings suggest that dispersal via rafting or could be a plausible mechanism for overcoming 454 both short (Hedin et al. 2013; Opatova and Arnedo 2014b) and long distances (Harrison et al. 455 2017). 20 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 456 The results of our biogeographic analyses (Fig. 6) place the ancestral area of distribution 457 of Atypoidea within the Northern hemisphere (across North America and Asia), whereas the 458 ancestor of Porrhothelidae, Macrothelidae and Bipectina most likely evolved in the Southern 459 hemisphere (across Africa and Australia). In our dating analyses, we recovered two repeated 460 patterns of divergences occurring among taxa with present day distributions confined mostly to 461 the Southern hemisphere that in dispersal-limited invertebrate groups have been repeatedly 462 linked to vicariant events related to the different phases of Gondwana breakup (Wood et al. 463 2012; Frazão et al. 2015; Kim and Farrell 2015; Xu et al. 2015; Andújar et al. 2016; Toussaint et 464 al. 2017). 465 The initial phase of Gondwana breakup (Fig 5), i.e., the formation of East and West 466 Gondwana, began approximately 165 Ma with South America and Africa drifting away from 467 India, Madagascar, Antarctica and Australia. The pattern concordant with the East – West 468 Gondwana breakup observed in our results involved splits within the families Euagridae, 469 Migidae, Idiopidae (putatively also Barychelidae), and divergence between Atracidae – 470 Actinopodidae and Anamidae – Entypesidae /Microstigmatidae clade (marked with stars on the 471 chronogram in Fig. 5) was dated between 101 – 88 Ma. The second pattern, concordant with 472 West Gondwana fragmentation into South America and Africa occurring between 132 – 112 Ma 473 (Will and Frimmel 2018) involved divergences within Theraphosidae, Pycnothelidae and 474 Cyrtaucheniidae (marked as triangles) was dated to 81 – 61 Ma. The divergence times estimated 475 in our analyses were considerably younger than the ages estimated for other sedentary taxa with 476 former Gondwanan distribution, such as velvet worms or bothriurid scorpions (Murienne et al. 477 2014; Sharma et al. 2018). Our inferred divergences also postdate the onset of continental 478 drifting in both instances by approximately 50 My, however, the initial phases of different 21 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 479 Gondwana breakup stages were often characterized by slow tectonic movement, or even 480 quiescence (Boger 2011; Will and Frimmel 2018; Young et al. 2018). In case of East and West 481 Gondwana, the drifting was occurring at an extremely slow rate for at least the first ten million 482 years (Young et al. 2018). A connection between South America and Australia also likely 483 persisted via Antarctica until late Eocene, approximately 35 Ma (Sanmartín and Ronquist 2004), 484 which links the divergence time estimates between Neotropical and Australasian groups of 485 Darwin’s stag beetles dated to 47 and 58 Ma, respectively, to Gondwana breakup (Kim and 486 Farrell 2015). 487 Conversely, the divergence time estimates inferred in our analyses could also be affected 488 by the limited fossil record for mygalomorphs, and our conservative approach towards node 489 calibration. For example, the Cretamygale chasei fossil, formerly interpreted as Nemesiidae, was 490 used to set a minimum bound for the whole “Nemesioidina” clade (see methods), rather than 491 arbitrarily assigning the fossil to the family Nemesiidae (sensu this study). Despite the apparent 492 advantages in the ability of handling genomic data, the dating analyses used in this study (Smith 493 and O’Meara 2012), does not allow to assign probability distribution priors to the calibration 494 points or infer confidence intervals for the divergences. Confidence intervals are often essential 495 for connecting the age estimates and specific geological events and while assessing them via 496 bootstrapping potentially informs our analyses, this approach likely does not capture the full span 497 of variation as accurately as when node calibration priors are taken into account. 498 Alternatively, the observed patterns could be explained by long-distance dispersal events 499 as in case of other poorly dispersing groups such as euedaphic beetles (Andújar et al. 2016) or 500 short-tailed whip-scorpion order Schizomida (Clouse et al. 2017). Although evidence suggests 501 that mygalomorph spiders are capable of trans-oceanic dispersal (Harrison et al. 2017), repeated 22 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 502 divergence patterns imply multiple long-distance dispersal events occurring independently 503 among unrelated taxa during the same timeframe. Continental-level vicariance thus provides a 504 more plausible explanation for the observed pattern. 505 The East – West Gondwana vicariant origin of the taxa is also supported by the results of 506 the ancestral range reconstruction analyses. The inferred ranges for all the taxon groups 507 displaying a distribution pattern concordant with the vicariance (Euagridae, Atracidae and 508 Actinopodidae, Migidae, Idiopidae, Barychelidae, Entypesidae and Anamidae) comprised former 509 Gondwana continents, although the ancestral range inferred for Idiopidae, Barychelidae and 510 Migidae (Africa and Australia) could potentially represent an artifact stemming from a 511 geographical bias in our sampling (i.e. lacking their South American representatives in our 512 analyses). Vicariance was also inferred with high probability in the divergences between 513 Allothele and Australian euagrids, Atrophothele and remaining barychelids. The West Gondwana 514 vicariant split between Africa and South America can likely be attributed to the structure within 515 Pycnothelidae (Pionothele – remaining pycnothelids) and South American and African 516 Cyrtaucheniids which also optimizes as an ancestral distribution spanning across both continents 517 in both groups. 518 Another repeated pattern observed across our tree includes taxa with currently disjunct 519 amphi-Atlantic distributions (marked as dots on the chronogram). Such patterns of divergence 520 were observed within Ummidia and Cyclocosmia (Halonoproctidae) and between Sphodros and 521 Atypus (Atypidae), and were dated to 44 Ma, 33 Ma (35 – 29) and 39 Ma (41 – 37), respectively. 522 This pattern has previously been connected with the breakup of Laurasia (Opatova et al. 2013; 523 Godwin et al. 2018) dating to the end of Paleocene, approximately 60 – 55 Ma (Sanmartín et al. 524 2001) and it is supported by the ancestral distributions of the taxa projected to the Northern 23 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 525 hemisphere continents in our analyses. Our divergences postdate both the geological event and 526 previous dating results (Opatova et al. 2013). The discrepancy between these time estimates and 527 the Laurasian breakup could be explained by either an artefact of the analyses (discussed above), 528 or alternatively by a persistent connection of North America and Europe via a North American 529 land bridge. This connection existed to some extent until 25 Ma and is hypothesized to have 530 played an important role in plant dispersal (Tiffney and Manchester 2001; Milne 2006), 531 suggesting also a suitable dispersal route for ballooning spiders like Atypus, Sphodros, Ummidia 532 (Opatova et al. 2013). The current distribution of Cyclocosmia (East Asia and South East of 533 North America), represents a well-known biogeographic pattern linked to continental vicariance 534 (Boufford and Spongberg 1983; Bolotov et al. 2016), which is further supported by the inferred 535 vicariant origin of the genus. Alternatively, similar distribution patterns could have resulted from 536 dispersal events between Palearctic and Nearctic regions via Beringia (Li et al. 2015; Bolotov et 537 al. 2016). However, the timing of the dispersal in other organisms postdates the results of our 538 analyses (Li et al. 2015; Maguilla et al. 2018). 539 Older divergences recovered in our analyses are more challenging to explain within a 540 biogeographic context. If we follow the same line of reasoning as explained above, and assume 541 that the actual ages of the splits are older than indicated by the results, the splits between the 542 Idiopidae and Ctenizidae/Euctenizidae clade, and Nemesiidae and the remaining “Nemesioidina” 543 diversity (marked as inverted triangles), could correspond to the disintegration of the 544 supercontinent Pangea occurring in Lower Jurassic (190 – 180 Ma) (Beutel et al. 2005; Will and 545 Frimmel 2018). Divergences proceeding the breakup of Pangea could be potentially linked to 546 paleoclimatic events. Since its formation around 520 – 510 Ma (Veevers 2004; Domeier and 547 Torsvik 2014; Will and Frimmel 2018), large parts of Gondwana experienced repeated 24 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 548 glaciations in the Silurian, Devonian, late Carboniferous and early Triassic. Additional 549 glaciations were caused by the orogenic processes prompted by the collision of Gondwana with 550 Laurussia and subsequent formation of Pangea ~320 Ma (Golonka 2007; Nance et al. 2014; 551 Matthews et al. 2016) in the Permian. Suitable habitats were thus repeatedly fragmented, likely 552 causing taxon distributional shifts and large-scale extinction events (Coccioni and Galeotti 1994; 553 Zachos et al. 2001; Chen and Benton 2012). The consequences of these events could be 554 potentially reflected in the mygalomorph phylogeny by the existence of many early-branching 555 monogeneric lineages such as Microhexuridae, Porrhothelidae, Macrothelidae, Paratropididae 556 and Stasimopidae, that do not seem to have any close relatives among extant taxa. Deep 557 divergences far preceding the breakup of Pangea have also been observed for a similarly ancient 558 and sedentary group – the velvet worms (Boyer et al. 2007; Murienne et al. 2014), indicating that 559 paleoclimatic events and habitat heterogeneity played an important role in the evolution of these 560 ancient groups. 561 We also detected potential long-distance dispersal events across the tree (denoted with 562 “*”). The Australasian genus Conothele (Halonoproctidae), shows divergence between 563 Australian and Vietnamese species dating back to approximately 46 Ma (47 – 44), which 564 postdates their vicariance during the northward drifting of India ~ 135 Ma (Will and Frimmel 565 2018). The genus exhibits some behavioral traits associated with ballooning (Main 1957), and it 566 is possible that aerial dispersal played a role in Conothele distribution. Other cases involve 567 Australian and African taxa. Poecilomigas (family Migidae) diverged from its Australia 568 inhabiting sister taxon Bertmainius approximately 66 Ma (72 – 57), and Kiama and 569 Microstigmata (both Microstigmatidae) split around 64 Ma (70 – 54). Both divergences are 570 likely too young to reflect the East – West Gondwana vicariance (inferred by the biogeographic 25 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 571 analyses) that started approximately 165 Ma, even when a possible bias in our dating analyses is 572 considered. Trans-oceanic dispersal has previously been reported from the family Migidae 573 (Harrison et al. 2017), which suggests that dispersal could plausibly explain the disjunct 574 distribution between Bertmainius and Poecilomigas. 575 Another unusual pattern occurs within the family Pycnothelidae, where the closest 576 relatives of the Australian genus Stanwellia have South American distributions. The split dates 577 back to approximately 28 Ma (30 - 26) which by far postdates the East – West Gondwana 578 vicariance but might actually reflect a connection between South America and Australia via 579 Antarctica that presumably persisted until 35 Ma (Sanmartín and Ronquist 2004). The same 580 vicariant scenario has been postulated to explain a similar distribution pattern recovered in Stag 581 beetles (Kim and Farrell 2015). The biogeographic analyses also inferred vicariance for the 582 Stanwellia – Acanthogonatus favoring South America – Australia connection over dispersal 583 event. 584 585 The Role of Silk in Prey Capture by Non-Araneomorph Spiders Revisited 586 587 As noted by Coyle (1986), progress in our understanding of the role silk plays in non- 588 araneomorph spiders lags far behind what is known about araneomorphs; this was an accurate 589 statement in 1986, and largely remains true today. As we discuss below, a prevailing perception 590 considered by us misleading, if not erroneous, is that mygalomorph spiders do not use silk 591 directly in foraging for prey; that is, it has been perceived that they do not spin webs defined in 592 either a restricted or broad sense. For the purposes of our discussion we consider two 593 characteristics as fundamental to ‘qualifying’ silk as playing a role in prey capture: 1) silk is 26 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 594 employed by the spider in a sensory capacity to transmit information and extend prey sensing 595 space outward to localize prey; and/or 2) silks can also be used by the spider to slow, impede, or 596 restrain prey items. These two criteria are not mutually exclusive, particularly in mygalomorphs 597 where visual clues are not known to play a significant role in prey capture; simply put silk is 598 employed to increase predator speed (fast predator and/or slower prey (Coddington et al. 2019)). 599 Note that Vollrath and Selden (2007) had a much broader definition of a web that included any 600 spider structure made of silk. Because the question we address here is specifically related to the 601 use of silk in prey capture, our working definition is narrower than theirs. It is our hypothesis, 602 forwarded herein, that mygalomorphs, and consequently all spiders, plesiomorphically used silk 603 to both localize and impede prey for capture. Whereas not all mygalomorphs use silken 604 constructs to impede prey (criterion 2), nearly all employ it in some fashion to transmit 605 vibrations to facilitate prey localization (criterion 1). 606 As documented in Figure 7, the MRCA for all mygalomorphs shows an unambiguous 607 optimization for a funnel-and-sheet web versus trapdoor. Although the optimization of this 608 character is equivocal at the root node (the MRCA of liphistiids and mygalomorphs), we would 609 contend that the likely character state at this node, if araneomorphs and uraraneids (e.g., 610 Chimerarachne yingi, (Wang et al. 2018)) were to be included in the analysis, would likely 611 resolve as a funnel-and-sheet web. As noted by Vollrath and Selden (2007), there is a clear link 612 between body structure and behavior in extant spiders. As such, the proportionally long 613 spinnerets of spider sister taxon Chimerarachne suggest that they were funnel-and-sheet web 614 builders (mygalomorph sheet web builders have proportionally longer spinnerets than their 615 counterparts – see Ischnothelidae). Eskov and Selden (2005) hypothesized that Permarachne (at 616 the time considered an extinct mesothele – now placed in the Uraraneida) constructed a funnel- 27 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 617 and-sheet web but likely evolved from a protypical mesothele ancestor that constructed a 618 trapdoor burrow with radiating lines. Alternatively, the uraraneid placement as the sister group to 619 all spiders (Wang et al. 2018) likely requires the optimization that funnel-and-sheet webs are the 620 plesiomorphic condition with trapdoor burrows independently derived in Liphistiidae – a lineage 621 that is morphologically unique among spiders (see below). Funnel-and-sheet webs meet both 622 criteria outlined above (Coyle 1986); that is, they transmit and extend prey localization from a 623 tubular retreat (typically) and serve to entangle prey. Although silk may not have initially 624 evolved for this function in the group’s common ancestor (that is, the common ancestor of 625 Uraraneida + Araneae), these results strongly suggest that early silk use likely played a role in 626 prey capture. Some authors have hypothesized that silk was initially used in reproduction (e.g., 627 protecting eggsacs, sperm webs, etc. (Shultz 1987) or played a sensory role (Bond 1994)). 628 However, these early functions may not be mutually exclusive; silken threads extending from an 629 eggsac could have allowed a spider to care for her young while still able to sense prey (or 630 predators) from a distance. 631 Irrespective of the uraraneid sheet web hypothesis originally proposed by Eskov and 632 Selden (2005), the funnel-and-sheet web optimization at the mygalomorph node inferred here 633 may be considered by some to be surprising given that the sister group to all other spiders 634 (Opisthothelae), Mesothelae, build burrows covered by trapdoors. As we discuss in more detail 635 below, a trapdoor covered burrow may have evolved from a sheet web (certainly the case among 636 more derived mygalomorph taxa). The Mygalomorphae MRCA with high probability spun a 637 sheet web from which trapdoor burrows were derived independently twice across the tree (once 638 in the Atypoidea and a second time in the Bipectina). Nevertheless, we consider trapdoor burrow 639 entrances to play a role in prey localization (discussed in more detail below), further enhanced in 28 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 640 some liphistiids and other trapdoor spiders by extending the prey sensing surface of their burrow 641 opening via silken trip lines or leaves, and other organic debris. 642 The evolutionary plasticity of the ancestral funnel-and-sheet web is best illustrated in the 643 Atypoidea. The antrodiaetid clade includes Antrodiaetus (plus Atypoides), Aliatypus, and Hexura 644 – all of which construct collared burrows, turreted burrows, trapdoors, and sheet webs, 645 respectively. As demonstrated in a number of experiments conducted by Coyle (1986), the 646 trapdoor spider Aliatypus likely suffers from reduced foraging efficiency relative to its more 647 open burrow (collared door) and sheet web related genera. The sister group, Atypidae, builds 648 purse webs, which are silken tubes that typically extend vertically up trees or horizontally along 649 a substrate. As discussed by Coyle (1986), these spiders respond rapidly to vibrations along the 650 tube surface; atypid purse webs definitively extend the spider’s prey sensory environment. An 651 optimization of web constructs across the Atypoidea conducted by Hedin et al. (2019) recovered 652 a similar pattern. 653 The Avicularioidea clade unequivocally has an MRCA that forages from a funnel-and- 654 sheet web. As discussed in the classification section below, diplurids are polyphyletic, with most 655 genera forming a grade of lineages that are sister to the remaining mygalomorphs. The 656 characteristics of the ischnothelid, euagrid, etc. and hexathelid funnel-and-sheet webs are well 657 documented for their abilities to aid in sensing/localizing prey, and in many taxa also impeding 658 and/or entangling prey items. These silken constructs play a significant role in prey capture and 659 are by any definition a “foraging web.” Funnel-and-sheet webs are independently derived twice 660 in atracids (formerly hexathelids) and much further up the tree in the clade that includes “true” 661 diplurids. The atracid funnel-and-sheet webs are highly modified from silken tubes (Gray 2010) 662 and likely not homologous to the ancestral funnel-and-sheet web. 29 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 663 The Bipectina clade resolves as having an MRCA with a trapdoor. This clade includes 664 typical trapdoor spiders like idiopids, migids, euctenizids, and nemesiids. As already discussed 665 above, and in a number of papers (Coyle 1986; Coyle et al. 1992; Bond and Coyle 1995), 666 trapdoor spiders employ silk in prey detection at the burrow entrance and in the door directly. 667 Moreover, many trapdoor spider species included in this clade add additional silk lines, plant 668 material, and tabs to the burrow entrance further emphasizing the role that the silken lined 669 burrow entrance plays in detecting and localizing prey. Although trapdoor covered burrows do 670 not seemingly serve to entangle or impede prey (criterion 2, above), aspects of the burrow do 671 serve to enhance the sensory capacity of the spider. Trapdoors are independently lost in a number 672 of Bipectina families in favor of other burrow constructs and foraging methods – Atracidae: 673 funnel-and-sheet web; Theraphosidae: open burrows (though trapdoors have reappeared ≥3 674 times; Hamilton persn. obsv.); Pycnothelidae: open burrows, turrets; Dipluridae: funnel-and- 675 sheet webs; Anamidae: open burrows and collar doors; Microstigmatidae: cursorial hunters, 676 scored as inapplicable (sensu paratropidids). Additionally, species of one euctenizid genus, 677 Apomastus, forages from an open burrow. In a number of these taxa, extensive field observations 678 demonstrate that vibrations at the silken burrow entrance can be used to coax the spider to the 679 surface, illustrating the functional role that silk plays in prey detection for these taxa. 680 As mentioned above, Coyle (1986) demonstrated that trapdoors appear to reduce foraging 681 efficiency in some taxa. As such it begs the question as to why some mygalomorph spiders have 682 abandoned funnel-and-sheet webs in favor of a foraging construct that appears to be less 683 effective at detecting prey and clearly lacks the ability to impede or entangle prey items. 684 Although this remains an open question, trapdoor spiders likely required additional protection 685 from predators (e.g., parasitoid insects like pompilid wasps), desiccation, or even flooding. The 30 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 686 reversal to an open burrow, often with an expanded collar for prey sensing (e.g., Apomastus 687 burrows, (Bond 2004)), suggests that once environmental conditions or predator selective 688 pressures change, the trapdoor can be easily lost. 689 It is our supposition that most mygalomorphs build foraging webs; that is, the majority of 690 species employ silk either in a funnel-and-sheet web or at a burrow entrance to detect and 691 localize prey and that a large number of taxa use silk to impede or entangle prey (the latter 692 considered to be the plesiomorphic condition). These points are important to make particularly in 693 light of a recent study by Fernández et al. (2018) that scored most mygalomorph taxa as having a 694 non-foraging burrow. Based on the results and data discussed here, not surprisingly, we would 695 question any analysis that hypothesizes mygalomorphs (and spiders as whole) to have a non-web 696 foraging MRCA. Such an interpretation likely stems from limited taxon sampling and major 697 differences in character state coding (Coddington et al. 2019). The latter in particular includes an 698 interpretation that a spider foraging from a burrow is not employing silk in the process of prey 699 capture, which, in our opinion contradicts the information available in the literature and direct 700 observations in the field. 701 702 Mygalomorphae Systematics and Higher-level Classification 703 704 The criterion of diagnosability, dictated by the ICZN, that all nominal taxa be 705 diagnosable (morphologically or otherwise) presents some problems for higher level 706 classification schemes based on these genomic data. Although one may recognize a molecularly 707 defined lineage as a family rank taxon (with respect to other previously defined lineages at that 708 equivalent rank), it is unreasonable to necessarily expect that some set of morphological features 31 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 709 will reveal themselves as diagnostic for this newly identified lineage/taxon. A process of 710 reciprocal illumination may lead to the discovery of such diagnostic characteristics, but that there 711 are no guarantees. One might infer (based on character homoplasy) that the overriding notion of 712 taxic homology, that diagnostic characters are used to unite groups (Panchen 1992), was an 713 implicit driving force in previous attempts delineate mygalomorph families and other higher- 714 level taxa (sensu Raven (1985)); an approach that at times seemingly interpreted diagnosability 715 as synapomorphy. A diagnostic character(s) for a family does not necessarily infer 716 synapomorphy any more than characters in a taxonomic key. A lineage that is hierarchically 717 located at the same level among other family rank taxa should potentially be viewed as a family 718 rank taxon regardless of whether or not it is morphologically diagnosable. 719 Taxonomic changes referenced herein are formally documented and new family-level 720 taxa diagnosed in Appendix 1 (authorship to be formally attributed as Bond, Opatova, and Hedin 721 2019). We acknowledge that the taxonomic changes we propose are unlikely to go without 722 criticism and will most likely require future revision and refinement. The four guiding principles 723 we used attempted to realign family-level limits as clearly as possible but remain conservative in 724 order to minimize errors: 1) revise higher-level mygalomorph classification and familial-level 725 taxonomy such that it better reflects the evolutionary relationships indicated by data available 726 (i.e., the sum of collected to data, not just this analysis); 2) further refine the classification, where 727 possible, based on previous morphological/molecular hypotheses in which subfamilial taxa are 728 clearly defined and supported by unambiguous synapomorphies; in some instances we 729 transferred unsampled genera to newly defined families in cases where their placement in said 730 group appears to be well supported in other studies; 3) minimize changes in family-level 731 classification to the extent possible that maintains a consistent phylogenetic level across the tree 32 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 732 for the family-level rank; and 4) instances where generic-level taxon sampling and criterion 2 733 above preclude an unambiguous family-level assignment of a genus (for relimited families), we 734 retain the original family assignment but denote its placement as incertae sedis (operationally for 735 the purposes of this study). We discuss below the restructuring of the major mygalomorph 736 groups and their constituent families with a particular emphasis on longstanding problems in the 737 families Dipluridae, Cyrtaucheniidae, and Nemesiidae. 738 739 Higher Level Groupings: Atypoidea – Avicularioidea – Bipectina – Domiothelina 740 Crassitarsae.—The interpretation of the significance of similarities in spinneret morphology, 741 constituted one of the center pieces in the debate revolving around the higher-level classification 742 of the Mygalomorphae (Platnick and Gertsch 1976; Raven 1985; Eskov and Zonshtein 1990; 743 Goloboff 1993) and taxon composition of some families (Raven 1985). Traditionally, 744 mygalomorph spiders were divided into two families: Atypidae and Aviculariidae 745 (=Theraphosidae; Simon 1864), later established as superfamilies Atypoidea and Ctenizoidea 746 (Chamberlin and Ivie 1946). Atypoidea, a sister group to all the remaining mygalomorph 747 families, consisted of Atypidae, Antrodiaetidae, and Mecicobothriidae. However, the external 748 similarities of mecicobothriids and diplurids motivated some authors to reject the placement of 749 Mecicobothriidae within Atypoidea (Platnick and Gertsch 1976), and subsequently even the 750 concept of Atypoidina (Atypidae + Antrodiaetidae; Raven 1985) being a sister group to all 751 remaining mygalomorphs was abandoned in favor of an alternative arrangement (Fig. 1a; Raven 752 1985). Eskov and Zonshtein (1990) disagreed with some of Raven’s character assessments and 753 alternatively supported Atypoidea (sensu Chamberlin and Ivie (1946)), arguing that the shared 754 traits in mecicobothriids and the diplurids, for example the elongation of spinnerets, constitute an 33 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 755 ecological convergence related to a construction of a funnel-and-sheet web (Coyle 1986). 756 However, their alternative hypothesis was rejected in what was the first computer-based cladistic 757 treatment of mygalomorph relationships (Goloboff 1993), on the basis of different interpretation 758 of Atypoidea unifying characters presented by Eskov and Zonshtein (1990). 759 Contrary to the prevailing morphology-based hypothesis (Raven 1985; Goloboff 1993), 760 the first studies using molecular data strongly supported the classic Atypoidea hypothesis (Hedin 761 and Bond 2006). The root node bifurcation of Mygalomorphae into Atypoidea and 762 Avicularioidea was further established after adding additional molecular data (Ayoub et al. 2007; 763 Bond et al. 2012). Although with limited taxon sampling, the same arrangement has been 764 consistently recovered in subsequent studies employing transcriptome data (Bond et al. 2014; 765 Garrison et al. 2016) and is also supported in our analyses. A recent study by Hedin et al. (2019); 766 published while this study was in review, provided a much needed overhaul of the Atypoidea, 767 recognizing two additional families (Hexurellidae and Megahexuridae). 768 The Avicularioidea clade was recovered in all the analyses with high support and a 769 degree of internal resolution never achieved in past comprehensive mygalomorph phylogenetic 770 studies. Similarly, to the results obtained in the past (Goloboff 1993; Hedin and Bond 2006; 771 Ayoub et al. 2007; Bond et al. 2012), the concatenated analyses recovered a grade of early 772 branching Avicularioidea lineages representing the so-called “non-diplurine diplurids” (Goloboff 773 1993), the Hexathelidae, Porrhothelidae and Macrothelidae. Alternatively, the species tree 774 analysis performed in ASTRAL recovered these lineages, excluding Macrothelidae, in a single 775 highly supported clade. We statistically explored alternative tree topologies (Table 2) to: 1) 776 assess the monophyly of the family Dipluridae (sensu Raven (1985)), 2) test the species tree 777 arrangement and 3) further test the monophyly of the family Hexathelidae (sensu Raven (1985)) 34 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 778 within a wider outgroup sampling framework than in Hedin et al. (2018). The rejection of the 779 alternative topologies by the AU topology tests, the reciprocal monophyly of the “non-diplurine 780 diplurid” lineages correlated with a unique combination of morphological characters present in 781 each taxon, provides strong support for taxonomic and nomenclatural changes proposed here and 782 by other authors (Hedin et al. 2018). 783 As in the previous analyses (Hedin and Bond 2006; Ayoub et al. 2007; Bond et al. 2012), 784 all Avicularioidea families, with exception of the grade of “diplurid” and former “hexathelid” 785 lineages, were recovered in a single clade, close in taxon composition to the original delineation 786 of Bipectina (sensu Goloboff (1993)). We recovered Paratropididae and Actinopodidae (removed 787 from Bipectina by Bond et al. (2012)) alongside the Atracidae as part of that clade. The 788 Bipectina including these three families represent the most inclusive monophyletic arrangement 789 that is consistently recovered in all the analyses. For this reason, we choose to relimit the 790 Bipectina clade (sensu Bond et al. (2012)) to include Actinopodidae, Atracidae, Stasimopidae, 791 and Paratropididae alongside the Crassitarsae and Domiothelina. 792 A monophyletic group roughly consistent with Domiothelina (sensu Bond et al. (2012)) 793 was recovered in all of our analyses. However, given the recent taxonomic relimitation of the 794 family Ctenizidae (Godwin et al. 2018) and further taxonomic changes made in this paper, we 795 formally recircumscribe the Domiothelina to include the following families: Halonoproctidae, 796 Migidae, Idiopidae, Ctenizidae and Euctenizidae. We also relimit the family Ctenizidae to 797 include only the genera Cteniza and Cyrtocarenum and establish a new family Stasimopidae 798 (NEW FAMILY) to accommodate Stasimopus herein removed from Ctenizidae. 799 800 Albeit generally accepted as a monophyletic group (Hedin and Bond 2006; Bond et al. 2012), Crassitarsae (Raven 1985; Goloboff 1993) is an unstable clade in our analyses. The group 35 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 801 was recovered in the concatenated analyses (supported in IQ-TREE ML inference and 802 genomic/transcriptomic dataset), whereas other analyses supported an alternative placement for 803 the Theraphosoidina clade (sensu Bond et al. (2012)), namely as sister to Domiothelina and the 804 Stasimopidae/Atracinae/Actinopodidae clade (species tree analysis) or as the sister group to all 805 Bipectina lineages (No_Paratropis dataset). Despite the uncertain status of Crassitarsae, both 806 Theraphosoidina and the clade comprising Cyrtaucheniidae, Dipluridae, Microstigmatidae and 807 former “Nemesiidae” lineages (the “Nemesioidina”, see below), were highly supported and 808 mostly resolved at internal levels in all analyses. With the results at hand, we can now address 809 the long-standing taxonomic issues involving the non-monophyly of the families Dipluridae, 810 Cyrtaucheniidae and Nemesiidae. As a result, we recircumscribe the families Cyrtaucheniidae, 811 Dipluridae and Nemesiidae and establish the new families Anamidae (NEW RANK), 812 Entypesidae (NEW FAMILY), Bemmeridae (NEW RANK), and Pycnothelidae (NEW RANK). 813 “Dipluridae”.—The monophyly of the diplurids was first questioned by Goloboff (1993) 814 relatively soon after the last relimitation of the family (Raven 1985), which without a formal 815 change being made remains still reflected in the current classification scheme. Diplurids in their 816 original composition comprised a variety of, now in retrospect, very distantly related lineages 817 possessing to some extent widely spaced “lower” spinnerets and elongated “upper” spinnerets 818 (Simon 1889); which typifies the pitfalls of establishing groups of the basis of homoplastic, 819 convergently evolved characteristics. Mecicobothriidae (Gertsch 1979) and Hexathelidae (Raven 820 1980) were subsequently removed from Dipluridae and established as independent families and a 821 large number of taxa was then transferred to the family Nemesiidae (Raven 1985). Despite that, 822 the elongation of posterior lateral spinnerets and the wide separation of posterior median 823 spinnerets remained among characters uniting the family Dipluridae (Raven 1985). Goloboff 36 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 824 (1993) debated the character state scoring of the spinneret elongation in diplurids, hexathelids, 825 and mecicobothriids, alluding that the character definition was ambiguous, even within 826 Dipluridae. Goloboff (1993) recovered the Dipluridae (sensu Raven (1985)) as four independent 827 lineages which portended the future relimitation of diplurids to the subfamily Diplurinae (sensu 828 Raven (1985)). The non-monophyly of the diplurids was also cautiously echoed by Coyle (1995) 829 who, while conducting a taxonomic revision of the subfamily Ischnothelinae, did not find any 830 strong synapomorphies establishing a sister relationship between ischnothelines and any other 831 diplurid taxa (Coyle 1995). Diplurids have been recovered both as monophyletic (Ayoub et al. 832 2007; Bond et al. 2012) and paraphyletic (Hedin and Bond 2006; Wheeler et al. 2017; Hedin et 833 al. 2018). However, with exception of Hedin et al. (2018), the taxon sampling was restricted to 834 members of the subfamily Euagrinae and the seemingly contradictory outcome was likely a 835 result of a lack of strong phylogenetic signal in the sequenced markers. In our analyses, we 836 sampled representatives from three diplurid subfamilies (Diplurinae, Ischnothelinae, Euagrinae) 837 plus Microhexura, a genus with uncertain placement (Raven 1985); Masteriinae was not 838 included in this study. Our results are congruent with Goloboff’s (1993) hypothesis that 839 Dipluridae constitutes an assemblage of distantly related lineages that do not share a common 840 ancestor. The subfamilies Ischnothelinae, Euagrinae + Microhexura were recovered as three 841 independent lineages at the base of the Avicularioidea clade (“non-diplurine diplurids” (Goloboff 842 1993; Hedin et al. 2018)), whereas Diplurinae taxa were placed as sister to cyrtaucheniids within 843 proximity to nemesiids (Supplemental figure Fig. 9). The presumed synapomorphy of diplurids 844 (sensu Raven (1985)), elongated and widely spaced spinnerets, could thus be an example of a 845 retention of plesiomorphic condition (see discussion on silk use) of the “non-diplurine diplurids” 846 that evolved convergently in Diplurinae – the genomic architecture likely already in place (see 37 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 847 Uraraneida: Chimerarachne yingi (Wang et al. 2018)). As already pointed out by Coyle (1971) 848 and Eskov and Zonshtein (1990), the traits related to the morphology of the spinnerets are likely 849 correlated to the type of foraging structure produced by the spiders. “Diplurids” generally build 850 conspicuous three-dimensional sheet-like webs with a funnel-like retreat (Coyle 1988; Coyle and 851 Ketner 1990; Coyle 1995; Eberhard and Hazzi 2013; Passanha and Brescovit 2018), a trait 852 present in many other mygalomorph taxa that have little to no evolutionary proximity to 853 diplurids (e.g., Mecicobothriidae (Gertsch 1979), Macrothelidae (Snazell and Allison 1989; 854 Shimojana and Haupt 1998), Atracidae (Gray 2010), and is also paralleled in araneomorph 855 funnel-web spider family Agelenidae (Nentwig 1983). With the results of our phylogenetic 856 analyses, it seems safe to assume that spinneret morphology is a plastic trait that has little to no 857 informative value for higher-level classification in mygalomorph spiders. This observation is 858 further supported by the results of the topology test rejecting the monophyly of Dipluridae (sensu 859 Raven (1985)) as well as the monophyly of a clade consisting of the “non-diplurine diplurids” 860 (Goloboff 1993), Hexathelidae and Porrhothelidae. We therefore substantially relimit the 861 boundaries of the family Dipluridae (new circumscription). We elevate the subfamily 862 Ischnothelidae (NEW RANK) and Euagridae (NEW RANK) to family-level and establish the 863 new family Microhexuridae (NEW FAMILY). 864 “Nemesiidae”.—The cosmopolitan family Nemesiidae belongs among the most speciose 865 mygalomorph groups. As currently defined, it comprises over 400 nominal species that are 866 placed in 45 genera. The family was diagnosed by Raven (1985) as those taxa possessing two 867 rows of teeth on the paired tarsal claws and having scopulate tarsi (Raven 1985); a set of 868 characters found in other groups and thus likely homoplasious, and later reinterpreted as 869 plesiomorphic (Goloboff 1993). Not surprisingly, the monophyly of the family Nemesiidae has 38 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 870 never been recovered in any broad scale molecular phylogenetic analyses of mygalomorphs 871 (Hedin and Bond 2006; Ayoub et al. 2007; Bond et al. 2012). Moreover, even for analyses based 872 solely on morphological characters, the family was seemingly split into five different lineages 873 (Goloboff 1993). Paralleling the non-monophyletic nature of the family, the subfamilies 874 established by Raven (1985) were also typically recovered as para- or polyphyletic (Goloboff 875 1993; Goloboff 1995; Hedin and Bond 2006; Bond et al. 2012). For this current study, we 876 sampled five out of six nemesiid subfamilies (Raven 1985) and about 40% of the generic 877 diversity to more thoroughly assess the composition of the family, particularly with an emphasis 878 on resolving the position of the family Microstigmatidae and several difficult to place 879 “cyrtaucheniid” genera (Goloboff 1993; Hedin and Bond 2006; Ayoub et al. 2007; Bond et al. 880 2012; Wheeler et al. 2017). Like all other studies conducted to date (Goloboff 1993; Hedin and 881 Bond 2006; Ayoub et al. 2007; Bond et al. 2012; Garrison et al. 2016; Wheeler et al. 2017), we 882 did not recover the Nemesiidae as monophyletic (Supplemental Fig. 9). The South African genus 883 Spiroctenus was placed within Bemmeridae, whereas the remaining nemesiids were parceled 884 among five reciprocally monophyletic lineages forming a clade with the cyrtaucheniids, 885 diplurids, and microstigmatids (the “Nemesioidina” clade). Both microstigmatids and some 886 “cyrtaucheniids” have been recovered as lineages embedded within Nemesiidae in the past 887 (Goloboff 1993; Hedin and Bond 2006; Ayoub et al. 2007; Bond et al. 2012; Wheeler et al. 888 2017). Goloboff (1993), suggested that the some of the microstigmatid synapomorphies may be 889 of neotenic nature (as proposed previously by Raven (1985) and Griswold (1985)), 890 foreshadowing the eventual synonymy of microstigmatids with the nemesiids, or alternatively, 891 splintering Nemesiidae into multiple families. Similar alternatives were also proposed for 892 Cyrtaucheniidae (sensu Raven (1985)) based on the position of Acontius, “Cyrtauchenius”, and 39 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 893 Fufius (Hedin and Bond 2006; Bond et al. 2012). Unfortunately, we cannot assess the position of 894 Acontius and Fufius in this study. However, the sample representing “Cyrtauchenius” in Bond et 895 al. (2012), in fact, corresponds to Amblyocarenum, later removed from its synonymy (Decae and 896 Bosmans 2014). Although the phylogenetic placement of both genera could be conservatively 897 interpreted as “embedded within Nemesiidae sensu lato, the placement of Amblyocarenum along 898 with other genera likely has more severe taxonomic implications. Amblyocarenum is placed as a 899 sole “cyrtaucheniid” within a monophyletic clade comprising the genus Nemesia, the type genus 900 of the family. The affinities between Amblyocarenum and other nemesiids were also noted when 901 the genus was removed from synonymy with Cyrtauchenius (Decae and Bosmans 2014). On the 902 other hand, Cyrtauchenius forms a well-supported clade with two other cyrtaucheniid genera and 903 is placed as sister to another monophyletic and morphologically distinguishable clade, the 904 Dipluridae (sensu this study). The Dipluridae – Cyrtaucheniidae relationship and their placement 905 within Nemesiidae was previously recovered in the pilot assessment of mygalomorph phylogeny 906 with AHE data (Hamilton et al. 2016b), however, the sampling limitations of the study and the 907 interpretation of Fufius and “Linothele” (later identified as Diplura) as taxonomically misplaced 908 nemesiids alluded to a different conclusion than presented in this paper. The situation is similar 909 for Microstigmatidae, presumably another “candidate” nemesiid lineage (Hedin and Bond 2006; 910 Ayoub et al. 2007; Bond et al. 2012). The genus Microstigmata has been recovered in proximity 911 to Ixamatus and another Australian nemesiid genus Kiama (Bond et al. 2012), which is 912 congruent with our results. Kiama presents a unique combination of morphological features that, 913 to some extent, overlaps with diagnostic characters for a handful of families, and as a result, the 914 placement of the genus vacillated among Dipluridae (Main and Mascord 1969), Cyrtaucheniidae 915 (Raven 1985) and Nemesiidae (Bond et al. 2012). Both Kiama and Microstigmata possess a 40 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 916 uniquely modified cuticle and share a similar tarsal organ morphology (Raven 1985; Bond and 917 Opell 2002), however these similarities were dismissed as convergent or plesiomorphic by others 918 (Raven 1985; Goloboff 1993). The homology of these characters consequently needs to be 919 reassessed in a framework using other lines of evidence other than just morphology (Bond et al. 920 2012; Hedin et al. 2018). 921 The dilemma of whether to dissolve the Nemesiidae, or redefine the family boundaries to 922 include microstigmatids and cyrtaucheniids has been repeatedly discussed in the literature 923 (Goloboff 1993; Hedin and Bond 2006; Bond et al. 2012), with tendencies gravitating towards 924 the later solution (e.g. Bond et al. 2012; Fig. 1C in Hedin et al. (2018)). However, there were 925 several factors at play reinforcing the argument for the “inclusive” Nemesiidae. First and 926 foremost, the limited taxon sampling in the previous analyses (Hedin and Bond 2006; Ayoub et 927 al. 2007; Bond et al. 2012; Hamilton et al. 2016b) partially skewed the interpretation of the 928 results. The polyphyly of the family Cyrtaucheniidae (sensu Raven (1985)), misplacement of 929 Cyrtauchenius (see above) and lack of adequate representation of the diplurines (sensu Raven 930 (1985)) presented major hurdles for accurately interpreting nemesiid relationships. Recovering 931 Microstigmata or a few genera of an apparently polyphyletic family embedded within an 932 enlarged “nemesiid” construct understandably supported the notion that a broad synonymization 933 (lumping rather than splitting) might be the more stable and thus preferable taxonomic solution. 934 Adding to the overall ambiguity of the results, large sections of the recovered trees often lacked 935 support (Hedin and Bond 2006; Ayoub et al. 2007; Bond et al. 2012; Wheeler et al. 2017), 936 adding uncertainty to the monophyly of the groups in question. We would suggest that our 937 improved sampling and high support values brings long-needed insight and resolution to the 938 “nemesiid” problem, which is essential for ultimately answering the question of whether or not 41 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 939 the Nemesiidae should be fragmented into independent families or not. The family Nemesiidae 940 (sensu Raven (1985)) has lacked a well-supported synapomorphy, a widely recognized fact 941 (Goloboff 1993; Goloboff 1995; Bond et al. 2012; Harvey et al. 2018), since the family was first 942 delimited (Raven 1985). On the other hand, some groups (e.g., subfamilies) currently recognized 943 within the family, do possess diagnostic synapomorphic characters (Harvey et al. 2018), which is 944 also the case for the diplurids and cyrtaucheniids (sensu this study) that were recovered as 945 mutually exclusive lineages within the larger more inclusive “Nemesioidina” clade. We thus 946 argue that the original concept of the family (sensu Raven (1985)) must be abandoned in favor of 947 establishing multiple monophyletic families that can also be ultimately clearly defined 948 morphologically. Another fact in favor of such an arrangement is the relatively deep divergences 949 separating the lineages (Kuntner et al. 2018). Major lineages across the entire clade diversified 950 into their respective individual groups between 115 – 79 Ma, roughly congruent with dated 951 divergences among other families (e.g., Theraphosidae and Barychelidae split around 107 Ma 952 and Actinopodidae diverged from Atracidae approximately 98 Ma). In agreement with the results 953 of our analyses, we therefore relimit the family Nemesiidae (new circumscription) as outlined in 954 the Appendix 1. 955 956 CONCLUSIONS AND FUTURE DIRECTIONS 957 958 It is our hope that this study represents a significant leap forward in stabilizing 959 mygalomorph spider classification via a robust, well-supported phylogenetic framework. 960 Extensive taxon sampling combined with enhanced phylogenetic resolution using genomic-scale 961 data facilitates the proposal of an alternative classification scheme that more accurately reflects 42 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 962 the evolutionary relationships of the infraorder. Consequently, these results brought to their 963 logical fruition require extensive changes to existing familial composition as well as the 964 establishment of a number of new family-rank taxa. Understandably, such changes require 965 considerable downstream scrutiny and criticism. This newly derived framework is a hypothesis 966 that begs further testing through additional data and increased taxon sampling. 967 One observation in particular from this study that seems clear is that morphological data 968 are not uniformly suitable for reconstructing mygalomorph higher-level relationships, and to a 969 slightly lesser degree for delimiting family rank taxa (e.g., subfamily relationships seem to be 970 relatively stable as evidenced by the number of subfamilies elevated to family rank). Raven’s 971 (1985) results were in no small part afflicted by the rampant homoplasy that seemingly plagues 972 this group of spiders. Given the data and lack of computational resources available when the 973 analyses were conducted, the results Raven achieved represented the strongest prevailing 974 hypothesis at that time and provided a framework upon which to test new hypotheses. In 1985 975 and shortly thereafter (Eskov and Zonshtein 1990) disagreement regarding morphological 976 evidence for various clades and familial-level relationships and identity were in retrospect 977 matters of subjective opinion and consequently difficult to test (likely explaining why some 978 seemingly correct hypotheses, e.g., Atypoidea, were ignored). Alternatively though, it is 979 important to recognize that a relatively large percentage of the new family rank taxa were once 980 morphologically defined subfamilies – a number of which predate Raven (1985) – and thus 981 relatively correct. 982 Issues regarding spider classification notwithstanding, phylogenomics provides an 983 exciting new framework for considering broad scale questions related to spider evolution. Our 984 biogeographic and dating analyses suggest that ancient vicariance and dispersal shaped the 43 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 985 currently observed biogeographic patterns. Much of the deep phylogenetic structure across the 986 infraorder may be attributable to climatic heterogeneity and oscillation prior to the breakup of 987 Gondwana/Pangea. Such deep phylogenetic and ancient divergence, coupled with low vagility, 988 makes these spiders ideal candidates to test biogeographic hypotheses. 989 Furthermore, the recently held notion that most mygalomorph spiders capture prey from 990 non-foraging burrows (Fernández et al. 2018) has critical implications for how we might 991 understand and interpret silk across spiders as a whole. The reformulated hypothesis posits that 992 spiders likely foraged primitively from a silken funnel-and-sheet web that served classically-held 993 web functions like increasing predator speed and enhancing prey capture success. How 994 mygalomorph spiders build their webs and retreats, employ silk in prey capture, and the nature of 995 their spinning apparatus remains a rich area of study that has received little attention since 996 Coyle’s (1986) treatment. 997 Considerable work remains in achieving final resolution of mygalomorph relationships 998 and stable fully-resolved familial composition. Despite our best efforts, which included a 999 combination of transcriptomic and AHE data, we are still unable to resolve several of the deepest 1000 nodes; Paratropididae remains a difficult taxon to place as do the newly formed Stasimopidae. 1001 Despite our best efforts and major relimitation, many nemesiid, diplurid, and cyrtaucheniid 1002 genera remain in an uncertain taxonomic state. Although we could speculate about placement of 1003 many taxa (e.g., Aporoptychinae cyrtaucheniids, Masteriinae diplurids, as well as numerous 1004 South American nemesiids), such questions will have to be left to future studies. At the very 1005 least, it is our hope that this study facilitates further interest and provides a solid foundation for 1006 future work in this intriguing, and ancient group of spiders. 1007 44 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1008 ACKNOWLEDGEMENTS 1009 1010 We thank M. Arnedo, R. Booysen, A. Decae, M. Forman, N. Garrison, R. Godwin, J. Grismer, 1011 C. Haddad, S. Huber, V. Hula, P. Just, J. Malumbres-Olarte, T. Mazuch, JA. Neethling, J. 1012 Rodriguez, F. Šťáhlavský, J. Starrett, C. Stephen, and D. Weinmann for help with field collecting 1013 and providing additional samples. We are grateful to M. Harvey, M. Kuntner, C. Griswold, S. 1014 Zonshtein, B. Wiegmann, J. Ledford, and an anonymous reviewer for comments on earlier 1015 versions of the manuscript. The computational analyses were completed in part with resources 1016 provided by the Auburn University Hopper supercomputing cluster. This project was funded in 1017 part by National Science Foundation grants DEB-0841610 to JEB, Doctoral Dissertation 1018 Improvement Grant DEB-1311494 to CAH, Czech Science Foundation grant 16-10298S to JK, 1019 Auburn University Department of Biological Sciences and College of Sciences and Mathematics 1020 (JEB), an Auburn University Cellular and Molecular Biosciences Peaks of Excellence Research 1021 Fellowship (CAH), and Evert and Marion Schlinger Foundation, University of California Davis 1022 (JEB). 1023 45 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1024 LITERATURE CITED 1025 1026 Aberer AJ, Kobert K, Stamatakis A. 2014. ExaBayes: massively parallel Bayesian tree inference 1027 for the whole-genome era. Mol Biol Evol, 31:2553-2556. 1028 1029 Andújar C, Faille A, Pérez-González S, Zaballos JP, Vogler AP, Ribera I. 2016. Gondwanian 1030 relicts and oceanic dispersal in a cosmopolitan radiation of euedaphic ground beetles. 1031 Mol.Phylogenet.Evol., 99:235-246. 1032 1033 Anisimova M, Gil M, Dufayard J-F, Dessimoz C, Gascuel O. 2011. Survey of branch support 1034 methods demonstrates accuracy, power, and robustness of fast likelihood-based approximation 1035 schemes. Syst Biol, 60:685-699. 1036 1037 Ayoub NA, Garb JE, Hedin M, Hayashi CY. 2007. Utility of the nuclear protein-coding gene, 1038 elongation factor-1 gamma (EF-1g), for spider systematics, emphasizing family level 1039 relationships of tarantulas and their kin (Araneae: Mygalomorphae). Mol.Phylogenet.Evol., 1040 42:394-409. 1041 1042 Beaulieu JM, O’Meara BC, Donoghue MJ. 2013. Identifying hidden rate changes in the 1043 evolution of a binary morphological character: the evolution of plant habit in campanulid 1044 angiosperms. Syst Biol, 62:725-737. 1045 46 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1046 Beutel E, Nomade S, Fronabarger A, Renne P. 2005. Pangea's complex breakup: A new rapidly 1047 changing stress field model. Earth and Planetary Science Letters, 236:471-485. 1048 1049 Boger SD. 2011. Antarctica—before and after Gondwana. Gondwana Research, 19:335-371. 1050 1051 Bolotov IN, Vikhrev IV, Bespalaya YV, Gofarov MY, Kondakov AV, Konopleva ES, Bolotov 1052 NN, Lyubas AA. 2016. Multi-locus fossil-calibrated phylogeny, biogeography and a subgeneric 1053 revision of the Margaritiferidae (Mollusca: Bivalvia: Unionoida). Mol.Phylogenet.Evol., 1054 103:104-121. 1055 1056 Bond J, Opell BD. 2002. Phylogeny and taxonomy of thegenera of south-western North 1057 American Euctenizinae trapdoor spidersand their relatives (Araneae: Mygalomorphae, 1058 Cyrtaucheniidae). Zool. J. Linn. Soc. , 136:487-534. 1059 1060 Bond JE. 1994. Seta-spigot homology and silk production in first instar Antrodiaetus unicolor 1061 spiderlings (Araneae: Antrodiaetidae). J. Arachnol.:19-22. 1062 1063 Bond JE. 2004. Systematics of the Californian euctenizine spider genus Apomastus (Araneae : 1064 Mygalomorphae : Cyrtaucheniidae): the relationship between molecular and morphological 1065 taxonomy. Invertebr Systemat, 18:361-376. 1066 47 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1067 Bond JE. 2017. Euctenizidae. In Ubick, D., Paquin, P. Cushing, P., Roth, V. (eds.) Spiders of 1068 North American identification manual, 2nd Edition, 55-57. New Hampshire, American 1069 Arachnological Society. 1070 1071 Bond JE, Coyle FA. 1995. Observations on the natural history of an Ummidia trapdoor spider 1072 from Costa Rica (Araneae, Ctenizidae). J. Arachnol., 23:157-164. 1073 1074 Bond JE, Garrison NL, Hamilton CA, Godwin RL, Hedin M, Agnarsson I. 2014. Phylogenomics 1075 resolves a spider backbone phylogeny and rejects a prevailing paradigm for orb web evolution. 1076 Current Biology, 24:1765-1771. 1077 1078 Bond JE, Hedin M. 2006. A total evidence assessment of the phylogeny of North American 1079 euctenizine trapdoor spiders (Araneae, Mygalomorphae, Cyrtaucheniidae) using Bayesian 1080 inference. Mol.Phylogenet.Evol., 41:70-85. 1081 1082 Bond JE, Hedin MC, Ramirez MG, Opell BD. 2001. Deep molecular divergence in the absence 1083 of morphological and ecological change in the Californian coastal dune endemic trapdoor spider 1084 Aptostichus simus. Mol Ecol, 10:899-910. 1085 1086 Bond JE, Hendrixson BE, Hamilton CA, Hedin M. 2012. A Reconsideration of the Classification 1087 of the Spider Infraorder Mygalomorphae (Arachnida: Araneae) Based on Three Nuclear Genes 1088 and Morphology. PLoS ONE, 7:e38753. 1089 48 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1090 Bond JE, Stockman AK. 2008. An integrative method for delimiting cohesion species: Finding 1091 the population-species interface in a group of Californian trapdoor spiders with extreme genetic 1092 divergence and geographic structuring. Syst Biol, 57:628 - 646. 1093 1094 Bouckaert R, Vaughan TG, Barido-Sottani J, Duchene S, Fourment M, Gavryushkina A, Heled J, 1095 Jones G, Kuhnert D, De Maio N. 2018. BEAST 2.5: An Advanced Software Platform for 1096 Bayesian Evolutionary Analysis. BioRxiv:474296. 1097 1098 Boufford DE, Spongberg SA. 1983. Eastern Asian-eastern North American phytogeographical 1099 relationships-a history from the time of Linnaeus to the twentieth century. Ann Mo Bot Gard, 1100 70:423-439. 1101 1102 Boyer SL, Clouse RM, Benavides LR, Sharma P, Schwendinger PJ, Karunarathna I, Giribet G. 1103 2007. Biogeography of the world: a case study from cyphophthalmid Opiliones, a globally 1104 distributed group of arachnids. J Biogeography, 34:2070-2085. 1105 1106 Castalanelli MA, Huey JA, Hillyer MJ, Harvey MS. 2017. Molecular and morphological 1107 evidence for a new genus of small trapdoor spiders from arid Western Australia (Araneae: 1108 Mygalomorphae: Nemesiidae: Anaminae). Invertebr Systemat, 31:492-505. 1109 1110 Chamberland L, McHugh A, Kechejian S, Binford GJ, Bond JE, Coddington J, Dolman G, 1111 Hamilton CA, Harvey MS, Kuntner M. 2018. From Gondwana to GAAR landia: Evolutionary 1112 history and biogeography of ogre-faced spiders (Deinopis). J Biogeography, 45:2442-2457. 49 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1113 1114 Chamberlin RV. 1917. New spiders of the family Aviculariidae. Bull. Mus. Comp. Zool., 61:25- 1115 75 1116 1117 Chamberlin RV, Ivie W. 1946. On several new American spiders. University of Utah. 1118 1119 Chen Z-Q, Benton MJ. 2012. The timing and pattern of biotic recovery following the end- 1120 Permian mass extinction. Nature Geoscience, 5:375. 1121 1122 Clouse RM, Branstetter MG, Buenavente P, Crowley LM, Czekanski-Moir J, General DEM, 1123 Giribet G, Harvey MS, Janies DA, Mohagan AB. 2017. First global molecular phylogeny and 1124 biogeographical analysis of two arachnid orders (Schizomida and Uropygi) supports a tropical 1125 Pangean origin and mid-Cretaceous diversification. J Biogeography, 44:2660-2672. 1126 1127 Coccioni R, Galeotti S. 1994. KT boundary extinction: Geologically instantaneous or gradual 1128 event? Evidence from deep-sea benthic foraminifera. Geology, 22:779-782. 1129 1130 Cock PJ, Antao T, Chang JT, Chapman BA, Cox CJ, Dalke A, Friedberg I, Hamelryck T, Kauff 1131 F, Wilczynski B. 2009. Biopython: freely available Python tools for computational molecular 1132 biology and bioinformatics. Bioinformatics, 25:1422-1423. 1133 1134 Coddington JA, Agnarsson I, Hamilton C, Bond JE. 2018. Spiders did not repeatedly gain, but 1135 repeatedly lost, foraging webs. PeerJ Preprints, 6:e27341v27341. 50 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1136 1137 Coyle F. 1986. The role of silk in prey capture by nonaraneomorph spiders. 1138 1139 Coyle FA. 1971. Systematics and natural history of the mygalomorph spider genus Antrodiaetus 1140 and related genera (Araneae: Antrodiaetidae). Harvard Univ Mus Compar Zool Bull. 1141 1142 Coyle FA. 1981. The mygalomorph spider genus Microhexura (Araneae, Dipluridae). Bull. Am. 1143 Mus. Nat. Hist., 170:64-75. 1144 1145 Coyle FA. 1986. Chilehexops, a new funnelweb mygalomorph spider genus from Chile 1146 (Araneae, Dipluridae). American Museum Novitates; no. 2860. 1147 1148 Coyle FA. 1988. A revision of the American funnel web mygalomorph spider genus Euagrus 1149 (Araneae, Dipluridae). Bulletin of the American Museum of Natural History., 187:203-292. 1150 1151 Coyle FA. 1995. A revision of the funnelweb mygalomorph spider subfamily Ischnothelinae 1152 (Araneae, Dipluridae). Bulletin of the AMNH; no. 226. 1153 1154 Coyle FA, Dellinger RE, Bennet R. 1992. Retreat architecture and construction behavior of an 1155 East African idiopine trapdoor spider (Araneae, Idiopidae). Bulletin of the of the British 1156 Arachnological Society, 9:99-104. 1157 51 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1158 Coyle FA, Ketner ND. 1990. Observations on the prey and prey capture behaviour of the 1159 funnelweb mygalomorph spider genus Ischnothele (Araneae, Dipluridae). Bull Br arachnol Soc, 1160 8:97-104. 1161 1162 Cracraft J. 1988. Deep-history biogeography: retrieving the historical pattern of evolving 1163 continental biotas. Syst Biol, 37:221-236. 1164 1165 Decae AE. 1996. Systematics of the trapdoor spider genus Cyrtocarenum Ausserer, 1871 1166 (Araneae, Ctenizidae). Bulletin of British arachnological Society, 10:161-170. 1167 1168 Decae AE, Bosmans R. 2014. Synonymy of the Trapdoor Spider Genera Cyrtauchenius Thorell, 1169 1869 and Amblyocarenum Simon, 1892 Reconsidered (Araneae, Mygalomorphae, 1170 Cyrtaucheniidae). Arachnology, 16:182-192. 1171 1172 Dippenaar-Schoeman AS. 2002. Baboon and trapdoor spiders of Southern Africa: an 1173 identification manual. Agricultural Research Council Pretoria. 1174 1175 Domeier M, Torsvik TH. 2014. Plate tectonics in the late Paleozoic. Geoscience Frontiers, 1176 5:303-350. 1177 1178 Eberhard WG, Hazzi NA. 2013. Web construction of Linothele macrothelifera (Araneae: 1179 Dipluridae). The Journal of Arachnology, 41:70-75. 1180 52 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1181 Engelbrecht I, Prendini L. 2012. Cryptic diversity of South African trapdoor spiders: Three new 1182 species of Stasimopus Simon, 1892 (Mygalomorphae, Ctenizidae), and redescription of 1183 Stasimopus robertsi Hewitt, 1910. American Museum Novitates:1-42. 1184 1185 Eskov K, Zonshtein S. 1990. First Mesozoic mygalomorph spiders for the Lower Cretaceous of 1186 Siberia and Mongolia, with notes on the infraorder Mygalomorphae (Chelicerata: Araneae). 1187 Neue Jahrbuch für Geologie und Palaeontologie Abhandlungen, 178:325–368. 1188 1189 Eskov KY, Selden PA. 2005. First record of spiders from the Permian period (Araneae: 1190 Mesothelae). Bull Br arachnol Soc, 13:111-116. 1191 1192 Faircloth BC, McCormack JE, Crawford NG, Harvey MG, Brumfield RT, Glenn TC. 2012. 1193 Ultraconserved elements anchor thousands of genetic markers spanning multiple evolutionary 1194 timescales. Syst Biol, 61:717-726. 1195 1196 Fernández R, Hormiga G, Giribet G. 2014. Phylogenomic analysis of spiders reveals 1197 nonmonophyly of orb weavers. Current Biology, 24:1772-1777. 1198 1199 Fernández R, Kallal RJ, Dimitrov D, Ballesteros JA, Arnedo MA, Giribet G, Hormiga G. 2018. 1200 Phylogenomics, Diversification Dynamics, and Comparative Transcriptomics across the Spider 1201 Tree of Life. Current Biology, 28:1489-1497. e1485. 1202 53 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1203 Forster R. 1968. The spiders of New Zealand. Part II. Ctenizidae, Dipluridae and Migidae. Otago 1204 Mus. Bull. 1205 1206 Frazão A, da Silva HR, de Moraes Russo CA. 2015. The Gondwana breakup and the history of 1207 the Atlantic and Indian oceans unveils two new clades for early neobatrachian diversification. 1208 PloS one, 10:e0143926. 1209 1210 Gabriel R, Longhorn SJ. 2015. Revised generic placement of Brachypelma embrithes 1211 (Chamberlin & Ivie, 1936) and Brachypelma angustum Valerio, 1980, with definition of the 1212 taxonomic features for identification of female Sericopelma Ausserer, 1875 (Araneae, 1213 Theraphosidae). ZooKeys:75. 1214 1215 Gallon RC. 2002. Revision of the African genera Pterinochilus and Eucratoscelus (Araneae, 1216 Theraphosidae, Harpactirinae) with description of two new genera. Bull Br arachnol Soc, 1217 12:201-232. 1218 1219 Garrison NL, Rodriguez J, Agnarsson I, Coddington JA, Griswold CE, Hamilton CA, Hedin M, 1220 Kocot KM, Ledford JM, Bond JE. 2016. Spider phylogenomics: untangling the Spider Tree of 1221 Life. PeerJ, 4:e1719. 1222 1223 Garwood RJ, Dunlop JA, Selden PA, Spencer AR, Atwood RC, Vo NT, Drakopoulos M. 2016. 1224 Almost a spider: a 305-million-year-old fossil arachnid and spider origins. Proc. R. Soc. B, 1225 283:20160125. 54 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1226 1227 Gertsch WJ. 1979. A revision of the spider family Mecicobothriidae (Araneae, Mygalomorphae). 1228 American Museum novitates; no. 2687. 1229 1230 Gertsch WJ, Platnick NI. 1980. A revision of the American spiders of the family Atypidae 1231 (Araneae, Mygalomorphae). American Museum novitates; no. 2704. 1232 1233 Godwin RL, Opatova V, Garrison NL, Hamilton CA, Bond JE. 2018. Phylogeny of a 1234 cosmopolitan family of morphologically conserved trapdoor spiders (Mygalomorphae, 1235 Ctenizidae) using Anchored Hybrid Enrichment, with a description of the family, 1236 Halonoproctidae Pocock 1901. Mol.Phylogenet.Evol., 126:303-313. 1237 1238 Goloboff PA. 1993. A reanalysis of mygalomorph spider families (Araneae). American Museum 1239 Novitates, 3056:1–32. 1240 1241 Goloboff PA. 1995. A revision of the South American spiders of the family Nemesiidae 1242 (Araneae, Mygalomorphae): Part I: Species from Peru, Chile, Argentina, and Uruguay. Bull. 1243 Am. Mus. Nat. Hist.:1-189. 1244 1245 Golonka J. 2007. Late Triassic and Early Jurassic palaeogeography of the world. 1246 Palaeogeography, Palaeoclimatology, Palaeoecology, 244:297-307. 1247 55 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1248 Graham MR, Hendrixson BE, Hamilton CA, Bond JE. 2015. Miocene extensional tectonics 1249 explain ancient patterns of diversification among turret-building tarantulas (Aphonopelma 1250 mojave group) in the Mojave and Sonoran deserts. J Biogeography, 42:1052-1065. 1251 1252 Gray M. 2010. A Revision of the Australian Funnel-web Spiders (Hexathelidae: Atracinae). 1253 Records of the Australian Museum, 62:285-392. 1254 1255 Griswold CE. 1985. A revision of the African spiders of the family Microstigmatidae (Araneae: 1256 Mygalomorphae). Annals of the Natal Museum, 27:1-37. 1257 1258 Griswold CE, Ledford J. 2001. A Monograph of the Migid Trap Door Spiders of Madagascar: 1259 And a Review of the World Genera (Araneae, Mygalomorphae, Migidae). California Academy 1260 of Sciences. 1261 1262 Guadanucci JPL. 2005. Tarsal scopula significance in ischnocolinae phylogenetics (Araneae, 1263 Mygalomorphae, Theraphosidae). J. Arachnol.:456-467. 1264 1265 Guadanucci JPL. 2011. Cladistic analysis and biogeography of the genus Oligoxystre Vellard 1266 1924 (Araneae: Mygalomorphae: Theraphosidae). J. Arachnol., 39:320-326. 1267 1268 Guadanucci JPL. 2014. Theraphosidae phylogeny: relationships of the ‘Ischnocolinae’genera 1269 (Araneae, Mygalomorphae). Zool.Scr., 43:508-518. 1270 56 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1271 Guadanucci JPL, Wendt I. 2014. Revision of the spider genus Ischnocolus Ausserer, 1871 1272 (Mygalomorphae: Theraphosidae: Ischnocolinae). J. Nat. Hist. , 48:387-402. 1273 1274 Hamelryck T, Manderick B. 2003. PDB file parser and structure class implemented in Python. 1275 Bioinformatics, 19:2308-2310. 1276 1277 Hamilton CA, Formanowicz DR, Bond JE. 2011. Species Delimitation and Phylogeography of 1278 Aphonopelma hentzi (Araneae, Mygalomorphae, Theraphosidae): Cryptic Diversity in North 1279 American Tarantulas. PLoS ONE, 6:e26207. 1280 1281 Hamilton CA, Hendrixson BE, Bond JE. 2016a. Taxonomic revision of the tarantula genus 1282 Aphonopelma Pocock, 1901 (Araneae, Mygalomorphae, Theraphosidae) within the United 1283 States. ZooKeys:1. 1284 1285 Hamilton CA, Hendrixson BE, Brewer MS, Bond JE. 2014. An evaluation of sampling effects on 1286 multiple DNA barcoding methods leads to an integrative approach for delimiting species: a case 1287 study of the North American tarantula genus Aphonopelma (Araneae, Mygalomorphae, 1288 Theraphosidae). Mol Phylogenet Evol, 71:79-93. 1289 1290 Hamilton CA, Lemmon AR, Lemmon EM, Bond JE. 2016b. Expanding anchored hybrid 1291 enrichment to resolve both deep and shallow relationships within the spider tree of life. BMC 1292 Evolutionary Biology, 16:212. 1293 57 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1294 Harrison SE, Harvey MS, Cooper SJ, Austin AD, Rix MG. 2017. Across the Indian Ocean: A 1295 remarkable example of trans-oceanic dispersal in an austral mygalomorph spider. PloS one, 1296 12:e0180139. 1297 1298 Harvey MS, Hillyer MJ, Main BY, Moulds TA, Raven RJ, Rix MG, Vink CJ, Huey JA. 2018. 1299 Phylogenetic relationships of the Australasian open-holed trapdoor spiders (Araneae: 1300 Mygalomorphae: Nemesiidae: Anaminae): multi-locus molecular analyses resolve the generic 1301 classification of a highly diverse fauna. Zool. J. Linn. Soc., 184:407-452 1302 1303 Hedin M, Bond JE. 2006. Molecular phylogenetics of the spider infraorder Mygalomorphae 1304 using nuclear rRNA genes (18S and 28S): Conflict and agreement with the current system of 1305 classification. Mol.Phylogenet.Evol., 41:454-471. 1306 1307 Hedin M, Carlson D, Coyle F. 2015. Sky island diversification meets the multispecies 1308 coalescent–divergence in the spruce-fir moss spider (Microhexura montivaga, Araneae, 1309 Mygalomorphae) on the highest peaks of southern Appalachia. Mol Ecol, 24:3467-3484. 1310 1311 Hedin M, Derkarabetian S, Alfaro A, Ramírez MJ, Bond JE. 2019. Phylogenomic analysis and 1312 revised classification of atypoid mygalomorph spiders (Araneae, Mygalomorphae), with notes on 1313 arachnid ultraconserved element loci. PeerJ 7:e6864 http://doi.org/10.7717/peerj.6864. 1314 58 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1315 Hedin M, Derkarabetian S, Ramírez MJ, Vink C, Bond JE. 2018. Phylogenomic reclassification 1316 of the world’s most venomous spiders (Mygalomorphae, Atracinae), with implications for 1317 venom evolution. Scientific reports, 8:1636. 1318 1319 Hedin M, Starrett J, Hayashi C. 2013. Crossing the uncrossable: novel trans-valley 1320 biogeographic patterns revealed in the genetic history of low-dispersal mygalomorph spiders 1321 (Antrodiaetidae, Antrodiaetus) from California. Mol Ecol, 22:508-526. 1322 1323 Hendrixson BE, Bond JE. 2009. Evaluating the efficacy of continuous quantitative characters for 1324 reconstructing the phylogeny of a morphologically homogeneous spider taxon (Araneae, 1325 Mygalomorphae, Antrodiaetidae, Antrodiaetus). Mol.Phylogenet.Evol., 53:300-313. 1326 1327 Hendrixson BE, DeRussy BM, Hamilton CA, Bond JE. 2013. An exploration of species 1328 boundaries in turret-building tarantulas of the Mojave Desert (Araneae, Mygalomorphae, 1329 Theraphosidae, Aphonopelma). Mol Phylogenet Evol, 66:327-340. 1330 1331 Hewitt J. 1915. VI. Notes on several four-lunged spiders in the coliection of the Durban Museum 1332 with descriptions of two new forms. Durban Museum Novitates, 1:125-133. 1333 1334 Hewitt J. 1916. Descriptions of new South African spiders. Annals of the Transvaal Museum, 1335 5:180-213. 1336 59 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1337 Hughes N, McDougall A. 1990. New Wealden correlation for the Wessex basin. Proceedings of 1338 the Geologists' Association, 101:85-90. 1339 1340 Indicatti RP, Lucas SM. 2005. Description of a new genus of Nemesiidae (Araneae, 1341 Mygalomorphae) from the Brazilian Cerrado. Zootaxa, 1088:11-16. 1342 1343 Isbister GK, Gray MR, Balit CR, Raven RJ, Stokes BJ, Porges K, Tankel AS, Turner E, White J, 1344 Fisher MM. 2005. Funnel-web spider bite: a systematic review of recorded clinical cases. The 1345 Medical Journal of Australia, 182:407-411. 1346 1347 Isbister GK, Sellors KV, Beckmann U, Chiew AL, Downes MA, Berling I. 2015. 1348 Catecholamine-induced cardiomyopathy resulting from life-threatening funnel-web spider 1349 envenoming. Med J Aust, 203:302-304. 1350 1351 Katoh K, Standley DM. 2013. MAFFT multiple sequence alignment software version 7: 1352 improvements in performance and usability. Mol Biol Evol, 30:772-780. 1353 1354 Kim SI, Farrell BD. 2015. Phylogeny of world stag beetles (Coleoptera: Lucanidae) reveals a 1355 Gondwanan origin of Darwin’s stag beetle. Mol.Phylogenet.Evol., 86:35-48. 1356 1357 Král J, Kořínková T, Krkavcová L, Musilová J, Forman M, Herrera IMÁ, Haddad CR, Vítková 1358 M, Henriques S, Vargas JGP. 2013. Evolution of karyotype, sex chromosomes, and meiosis in 1359 mygalomorph spiders (Araneae: Mygalomorphae). Biol. J. Linn. Soc. , 109:377-408. 60 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1360 1361 Kück P. 2009. ALICUT: a Perlscript which cuts ALISCORE identified RSS. Department of 1362 Bioinformatics, Zoologisches Forschungsmuseum A. Koenig (ZFMK), Bonn, Germany, version, 1363 2. 1364 1365 Kück P, Longo GC. 2014. FASconCAT-G: extensive functions for multiple sequence alignment 1366 preparations concerning phylogenetic studies. Frontiers in Zoology, 11:81. 1367 1368 Kuntner M, Hamilton CA, Cheng R-C, Gregoric M, Lupse N, Lokovsek T, Lemmon E, Lemmon 1369 A, Agnarsson I, Coddington JA. 2018. Golden Orbweavers Ignore Biological Rules: 1370 Phylogenomic and Comparative Analyses Unravel a Complex Evolution of Sexual Size 1371 Dimorphism. bioRxiv:368233. 1372 1373 Lanfear R, Calcott B, Kainer D, Mayer C, Stamatakis A. 2014. Selecting optimal partitioning 1374 schemes for phylogenomic datasets. BMC evolutionary biology, 14:82. 1375 1376 Lanfear R, Frandsen PB, Wright AM, Senfeld T, Calcott B. 2016. PartitionFinder 2: new 1377 methods for selecting partitioned models of evolution for molecular and morphological 1378 phylogenetic analyses. Mol Biol Evol, 34:772-773. 1379 1380 Leavitt DH, Starrett J, Westphal MF, Hedin M. 2015. Multilocus sequence data reveal dozens of 1381 putative cryptic species in a radiation of endemic Californian mygalomorph spiders (Araneae, 1382 Mygalomorphae, Nemesiidae). Mol.Phylogenet.Evol., 91:56-67. 61 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1383 1384 Lemmon AR, Emme SA, Lemmon EM. 2012. Anchored Hybrid Enrichment for Massively 1385 High-Throughput Phylogenomics. Syst Biol, 61:727-744. 1386 1387 Li J-T, Wang J-S, Nian H-H, Litvinchuk SN, Wang J, Li Y, Rao D-Q, Klaus S. 2015. 1388 Amphibians crossing the bering land bridge: evidence from holarctic treefrogs (Hyla, Hylidae, 1389 Anura). Mol.Phylogenet.Evol., 87:80-90. 1390 1391 Lüddecke T, Krehenwinkel H, Canning G, Glaw F, Longhorn SJ, Taenzler R, Wendt I, Vences 1392 M. 2018. Discovering the silk road: Nuclear and mitochondrial sequence data resolve the 1393 phylogenetic relationships among theraphosid spider subfamilies. Mol.Phylogenet.Evol., 119:63- 1394 70. 1395 1396 Maddison WP, Evans SC, Hamilton CA, Bond JE, Lemmon AR, Lemmon EM. 2017. A 1397 genome-wide phylogeny of jumping spiders (Araneae, Salticidae), using anchored hybrid 1398 enrichment. ZooKeys:89. 1399 1400 Maguilla E, Escudero M, Luceño M. 2018. Vicariance versus dispersal across Beringian land 1401 bridges to explain circumpolar distribution: A case study in plants with high dispersal potential. J 1402 Biogeography, 45:771-783. 1403 1404 Main B. 1957. Occurrence of the trap-door spider Conothele malayana (Doleschall) in Australia 1405 (Mygalomorphae: Ctenizidae). Western Australian Naturalist, 5:209-216. 62 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1406 1407 Main B. 1993. Biogeographic significance of the Nullarbor cave mygalomorph spider 1408 Troglodiplura and its taxonomic affinities. Journal of the Royal Society of Western Australia, 1409 76:77-85. 1410 1411 Main BY. 1969. A blind mygalomorph spider from a Nullarbor plain cave. Roy Soc West 1412 Australia J. 52:9-11 1413 1414 Main BY, Mascord R. 1969. A new genus of diplurid spider (Araneae: Mygalomorphae) from 1415 New South Wales. Journal of the Entomological Society of Australia (NSW), 6:24. 1416 1417 Matthews KJ, Maloney KT, Zahirovic S, Williams SE, Seton M, Mueller RD. 2016. Global plate 1418 boundary evolution and kinematics since the late Paleozoic. Global and Planetary Change, 1419 146:226-250. 1420 1421 Milne R. 2006. Northern hemisphere plant disjunctions: a window on Tertiary land bridges and 1422 climate change? Annals of Botany, 98:465-472. 1423 1424 Mirarab S, Warnow T. 2015. ASTRAL-II: coalescent-based species tree estimation with many 1425 hundreds of taxa and thousands of genes. Bioinformatics, 31:i44-i52. 1426 1427 Misof B, Misof K. 2009. A Monte Carlo approach successfully identifies randomness in multiple 1428 sequence alignments: a more objective means of data exclusion. Syst Biol, 58:21-34. 63 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1429 1430 Mittermeier RA, Turner WR, Larsen FW, Brooks TM, Gascon C. 2011. Global biodiversity 1431 conservation: the critical role of hotspots. Biodiversity hotspots, Springer, p. 3-22. 1432 1433 Montes de Oca L, D'Elía G, Pérez-Miles F. 2016. An integrative approach for species 1434 delimitation in the spider genus Grammostola (Theraphosidae, Mygalomorphae). Zool.Scr., 1435 45:322-333. 1436 1437 Mora E, Paspati A, Decae AE, Arnedo MA. 2017. Rafting spiders or drifting islands? Origins 1438 and diversification of the endemic trap-door spiders from the Balearic Islands, Western 1439 Mediterranean. J Biogeography, 44:924-936. 1440 1441 Morrone JJ, Crisci JV. 1995. Historical biogeography: introduction to methods. Annu Rev Ecol 1442 Evol Systemat, 26:373-401. 1443 1444 Murienne J, Daniels SR, Buckley TR, Mayer G, Giribet G. 2014. A living fossil tale of Pangaean 1445 biogeography. Proceedings of the Royal Society of London B: Biological Sciences, 1446 281:20132648. 1447 1448 Myers N, Mittermeier RA, Mittermeier CC, da Fonseca GA, Kent J. 2000. Biodiversity hotspots 1449 for conservation priorities. Nature, 403:853-858. 1450 64 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1451 Nance RD, Murphy JB, Santosh M. 2014. The supercontinent cycle: a retrospective essay. 1452 Gondwana Research, 25:4-29. 1453 1454 Nentwig W. 1983. The prey of web-building spiders compared with feeding experiments 1455 (Araneae: Araneidae, Linyphiidae, Pholcidae, Agelenidae). Oecologia, 56:132-139. 1456 1457 Nguyen LT, Schmidt HA, von Haeseler A, Minh BQ. 2014. IQ-TREE: a fast and effective 1458 stochastic algorithm for estimating maximum-likelihood phylogenies. Mol Biol Evol, 32:268- 1459 274. 1460 1461 Opatova V, Arnedo MA. 2014a. From Gondwana to Europe: inferring the origins of 1462 Mediterranean Macrothele spiders (Araneae: Hexathelidae) and the limits of the family 1463 Hexathelidae. Invertebr Systemat, 28:361-374. 1464 1465 Opatova V, Arnedo MA. 2014b. Spiders on a Hot Volcanic Roof: Colonisation Pathways and 1466 Phylogeography of the Canary Islands Endemic Trap-Door Spider Titanidiops canariensis 1467 (Araneae, Idiopidae). PLoS ONE, 9:e115078. doi:115010.111371/journal.pone.0115078. 1468 1469 Opatova V, Bond JE, Arnedo MA. 2013. Ancient origins of the Mediterranean trap-door spiders 1470 of the family Ctenizidae (Araneae, Mygalomorphae). Mol Phylogenet Evol, 69:1135-1145. 1471 65 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1472 Opatova V, Bond JE, Arnedo MA. 2016. Uncovering the role of the Western Mediterranean 1473 tectonics in shaping the diversity and distribution of the trap-door spider genus Ummidia 1474 (Araneae, Ctenizidae). J Biogeography, 43:1955-1966. 1475 1476 Ortiz D, Francke OF, Bond JE. 2018. A tangle of forms and phylogeny: Extensive morphological 1477 homoplasy and molecular clock heterogeneity in Bonnetina and related tarantulas. 1478 Mol.Phylogenet.Evol., 127:55-73. 1479 1480 Ott R, Höfer H. 2003. Envia garciai, a new genus and species of mygalomorph spiders (Araneae, 1481 Microstigmatidae) from Brazilian Amazonia. Iheringia. Série Zoologia, 93:373-379. 1482 1483 Panchen AL. 1992. Classification, evolution, and the nature of biology. Cambridge University 1484 Press. 1485 1486 Passanha V, Brescovit AD. 2018. On the Neotropical spider Subfamily Masteriinae (Araneae, 1487 Dipluridae). Zootaxa, 4463:1-73. 1488 1489 Pattengale ND, Alipour M, Bininda-Emonds ORP, Moret BME, Stamatakis A. 2010. How many 1490 bootstrap replicates are necessary? Journal of Computational Biology, 17:337-354. 1491 1492 Pérez-Miles F. 1994. Tarsal scopula division in Theraphosinae (Araneae, Theraphosidae): its 1493 systematic significance. J. Arachnol.:46-53. 1494 66 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1495 Pérez-Miles F. 2002. The occurrence of abdominal urticating hairs during development in 1496 Theraphosinae (Araneae, Theraphosidae): phylogenetic implications. J. Arachnol., 30:316-320. 1497 1498 Pérez-Miles F, Lucas S, da Silva Jr P, Bertani R. 1996. 1499 Prum RO, Berv JS, Dornburg A, Field DJ, Townsend JP, Lemmon EM, Lemmon AR. 2015. A 1500 comprehensive phylogeny of birds (Aves) using targeted next-generation DNA sequencing. 1501 Nature, 526:569. 1502 1503 Raven R. 1991. A revision of the mygalomorph spider family Dipluridae in New Caledonia 1504 (Araneae). Mémoires du Muséum National d’Histoire Naturelle, Paris (A), 149:87-117. 1505 1506 Raven RJ, Schwendinger P. 1995. Three new mygalomorph spider genera from Thailand and 1507 China (Araneae). Memoirs-Queenslad Museum, 38:623-642. 1508 1509 Raven RJ. 1980. The evolution and biogeography of the mygalomorph spider family 1510 Hexathelidae (Araneae, Chelicerata). J. Arachnol., 8:251-266. 1511 1512 Raven RJ. 1981. A Review of the Australian Genera of the Mygalomorph Spider Subfamily 1513 Diplurinae (Dipluridae: Chelicerata). Aust. J. Zool. , 29:321-363. 1514 1515 Raven RJ. 1985. The spider infraorder Mygalomorphae (Araneae): cladistics and systematics. 1516 Bull. Amer. Mus. Nat. Hist., 182:1-180. 1517 67 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1518 Raven RJ. 1985. The spider infraorder Mygalomorphae (Araneae): cladistics and systematics. 1519 Bull. Amer. Mus. Nat. Hist., 182:1-180. 1520 1521 Raven RJ. 1994. Mygalomorph spiders of the Barychelidae in Australia and the western Pacific. 1522 Mem. Queensl. Mus., 35:291-706. 1523 1524 Ree RH, Smith SA. 2008. Maximum Likelihood Inference of Geographic Range Evolution by 1525 Dispersal, Local Extinction, and Cladogenesis. Syst Biol, 57:4 - 14. 1526 1527 Ríos Tamayo D. 2017. Re-description of Trichopelma cubanum (Theraphosidae: Ischnocolinae) 1528 and comments about the familial placement of Trichopelma. Mun. Ent. Zool., 12:194-198. 1529 1530 Ríos-Tamayo D, Goloboff P. 2017. Vilchura calderoni, a new genus and species of Euagrinae 1531 (Araneae: Mygalomorphae: Dipluridae) from Chile. Arachnology, 17:183-187. 1532 1533 Rix MG, Bain K, Main BY, Raven RJ, Austin AD, Cooper SJ, Harvey MS. 2017a. Systematics 1534 of the spiny trapdoor spiders of the genus Cataxia (Mygalomorphae: Idiopidae) from south- 1535 western Australia: documenting a threatened fauna in a sky-island landscape. The Journal of 1536 Arachnology, 45:395-423. 1537 1538 Rix MG, Cooper SJ, Meusemann K, Klopfstein S, Harrison SE, Harvey MS, Austin AD. 2017b. 1539 Post-Eocene climate change across continental Australia and the diversification of Australasian 1540 spiny trapdoor spiders (Idiopidae: Arbanitinae). Mol.Phylogenet.Evol., 109:302-320. 68 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1541 1542 Rix MG, Raven RJ, Main BY, Harrison SE, Austin AD, Cooper SJ, Harvey MS. 2017c. The 1543 Australasian spiny trapdoor spiders of the family Idiopidae (Mygalomorphae: Arbanitinae): a 1544 relimitation and revision at the generic level. Invertebr Systemat, 31:566-634. 1545 1546 Sanderson MJ. 2002. Estimating absolute rates of molecular evolution and divergence times: a 1547 penalized likelihood approach. Mol Biol Evol, 19:101-109. 1548 1549 Sanmartín I, Enghoff H, Ronquist F. 2001. Patterns of animal dispersal, vicariance and 1550 diversification in the Holarctic. Biol. J. Linn. Soc., 73:345-390. 1551 1552 Sanmartín I, Ronquist F. 2004. Southern hemisphere biogeography inferred by event-based 1553 models: plant versus animal patterns. Syst Biol, 53:216-243. 1554 1555 Satler JD, Carstens BC, Hedin M. 2013. Multilocus Species Delimitation in a Complex of 1556 Morphologically Conserved Trapdoor Spiders (Mygalomorphae, Antrodiaetidae, Aliatypus). Syst 1557 Biol, 62:805-823. 1558 1559 Schmidt G. 2002. Gehören Brachionopus Pocock, 1897 und Harpactirella Purcell, 1902 zu den 1560 Theraphosiden. Arthropoda, 10:12-17. 1561 1562 Selden PA. 2002. First British Mesozoic Spider, From Cretaceous Amber Of The Isle Of Wight, 1563 Southern England. Palaeontology, 45:973-983. 69 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1564 1565 Selden PA, Gall JC. 1992. A Triassic mygalomorph spider from the northern Vosges, France. 1566 Palaeontology, 35:211–235. 1567 1568 Selden PA, Shcherbakov DE, Dunlop JA, Eskov KY. 2014. Arachnids from the Carboniferous of 1569 Russia and Ukraine, and the Permian of Kazakhstan. Paläontologische Zeitschrift, 88:297-307. 1570 1571 Sharma PP, Baker CM, Cosgrove JG, Johnson JE, Oberski JT, Raven RJ, Harvey MS, Boyer SL, 1572 Giribet G. 2018. A revised dated phylogeny of scorpions: Phylogenomic support for ancient 1573 divergence of the temperate Gondwanan family Bothriuridae. Mol.Phylogenet.Evol., 122:37-45. 1574 1575 Shimodaira H. 2002. An approximately unbiased test of phylogenetic tree selection. Syst Biol, 1576 51:492-508. 1577 1578 Shimodaira H, Hasegawa M. 2001. CONSEL: for assessing the confidence of phylogenetic tree 1579 selection. Bioinformatics, 17:1246 - 1247. 1580 1581 Shimojana M, Haupt J. 1998. Taxonomy and natural history of the funnel-web spider genus 1582 Macrothele (Araneae: Hexathelidae: Macrothelinae) in the Ryukyu Islands (Japan) and Taiwan. 1583 Species Diversity, 3:1-15. 1584 1585 Shultz JW. 1987. The origin of the spinning apparatus in spiders. Biological Reviews, 62:89-113. 1586 70 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1587 Simon E. 1864. Histoire naturelle des Araignées (Aranéides). Paris, Roret edit. 1588 1589 Simon E. 1889. Voyage de M. E. Simon au Venezuela (Decembre 1887–Avril 1888) 4e 1590 Mémoire. . Annales de la Société Entomologique de France, Paris, p. 169-220. 1591 1592 Smith SA. 2010. Lagrange C++ Manual. 1593 1594 Smith SA, O’Meara BC. 2012. treePL: divergence time estimation using penalized likelihood for 1595 large phylogenies. Bioinformatics, 28:2689-2690. 1596 1597 Snazell RG, Allison R. 1989. The genus Macrothele Ausserer (Araneae; Hexathelidae) in 1598 Europe. Bull Br arachnol Soc, 8:65-72. 1599 1600 Stamatakis A. 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of 1601 large phylogenies. Bioinformatics, 30:1312-1313. 1602 1603 Starrett J, Derkarabetian S, Hedin M, Bryson Jr RW, McCormack JE, Faircloth BC. 2017. High 1604 phylogenetic utility of an ultraconserved element probe set designed for Arachnida. Mol. Ecol. 1605 Res., 17:812-823. 1606 1607 Starrett J, Hayashi CY, Derkarabetian S, Hedin M. 2018. Cryptic elevational zonation in 1608 trapdoor spiders (Araneae, Antrodiaetidae, Aliatypus janus complex) from the California 1609 southern Sierra Nevada. Mol.Phylogenet.Evol., 118:403-413. 71 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1610 1611 Starrett J, Hedin M, Ayoub N, Hayashi CY. 2013. Hemocyanin gene family evolution in spiders 1612 (Araneae), with implications for phylogenetic relationships and divergence times in the 1613 infraorder Mygalomorphae. Gene, 524:175-186. 1614 1615 Su YC, Brown RM, Chang YH, Lin CP, Tso IM. 2016. Did a Miocene–Pliocene island isolation 1616 sequence structure diversification of funnel web spiders in the Taiwan-Ryukyu Archipelago? J 1617 Biogeography, 43:991-1003. 1618 1619 Swofford D. 2003. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods), 1620 version 4.0b10. 1621 1622 Tiffney, Manchester. 2001. The use of geological and paleontological evidence in evaluating 1623 plant phylogeographic hypotheses in the Northern Hemisphere tertiary. International Journal of 1624 Plant Sciences, 162:S3-S17. 1625 1626 Toussaint EF, Bloom D, Short AE. 2017. Cretaceous West Gondwana vicariance shaped giant 1627 water scavenger beetle biogeography. J Biogeography, 44:1952-1965. 1628 1629 Turner SP, Longhorn SJ, Hamilton CA, Gabriel R, Pérez-Miles F, Vogler AP. 2018. Re- 1630 evaluating conservation priorities of New World tarantulas (Araneae: Theraphosidae) in a 1631 molecular framework indicates non-monophyly of the genera, Aphonopelma and Brachypelma. 1632 Systematics and Biodiversity, 16:89-107. 72 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1633 1634 Veevers J. 2004. Gondwanaland from 650–500 Ma assembly through 320 Ma merger in Pangea 1635 to 185–100 Ma breakup: supercontinental tectonics via stratigraphy and radiometric dating. 1636 Earth-Science Reviews, 68:1-132. 1637 1638 Vollrath F, Selden P. 2007. The Role of Behavior in the Evolution of Spiders, Silks, and Webs. 1639 Annu Rev Ecol Evol Systemat, 38:819-846. 1640 1641 Wang B, Dunlop JA, Selden PA, Garwood RJ, Shear WA, Müller P, Lei X. 2018. Cretaceous 1642 arachnid Chimerarachne yingi gen. et sp. nov. illuminates spider origins. Nature ecology & 1643 evolution, 2:614. 1644 1645 Wheeler WC, Coddington JA, Crowley LM, Dimitrov D, Goloboff PA, Griswold CE, Hormiga 1646 G, Prendini L, Ramírez MJ, Sierwald P. 2017. The spider tree of life: phylogeny of Araneae 1647 based on target-gene analyses from an extensive taxon sampling. Cladistics, 33:574-616. 1648 1649 Wiley EO. 1988. Vicariance biogeography. Annu Rev Ecol Evol Systemat, 19:513-542. 1650 1651 Will TM, Frimmel HE. 2018. Where does a continent prefer to break up? Some lessons from the 1652 South Atlantic margins. Gondwana Research, 53:9-19. 1653 73 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1654 Wood HM, González VL, Lloyd M, Coddington J, Scharff N. 2018. Next-generation museum 1655 genomics: Phylogenetic relationships among palpimanoid spiders using sequence capture 1656 techniques (Araneae: Palpimanoidea). Mol.Phylogenet.Evol., 127:907-918. 1657 1658 Wood HM, Matzke NJ, Gillespie RG, Griswold CE. 2012. Treating fossils as terminal taxa in 1659 divergence time estimation reveals ancient vicariance patterns in the palpimanoid spiders. Syst 1660 Biol, 62:264-284. 1661 1662 World Spider Catalog. 2018. World Spider Catalog (2018). World Spider Catalog. Version 19.5. 1663 Natural History Museum Bern, online at http://wsc.nmbe.ch/, accessed on {16 November 2018}. 1664 doi: 10.24436/2 1665 1666 Wunderlich Jh. 2011. Some fossil spiders (Araneae) in Eocene European ambers. In: Wunderlich 1667 J editor. Beiträge zur Araneologie, p. 472–538. 1668 1669 Xu X, Liu F, Cheng R-C, Chen J, Xu X, Zhang Z, Ono H, Pham DS, Norma-Rashid Y, Arnedo 1670 MA. 2015. Extant primitively segmented spiders have recently diversified from an ancient 1671 lineage. Proc. R. Soc. B, 282:20142486. 1672 1673 Young A, Flament N, Maloney K, Williams S, Matthews K, Zahirovic S, Müller RD. 2018. 1674 Global kinematics of tectonic plates and subduction zones since the late Paleozoic Era. 1675 Geoscience Frontiers. https://doi.org/10.1016/j.gsf.2018.05.011 1676 74 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1677 Yu Y, Harris AJ, Blair C, He X. 2015. RASP (Reconstruct Ancestral State in Phylogenies): a 1678 tool for historical biogeography. Mol.Phylogenet.Evol., 87:46-49. 1679 1680 Zachos J, Pagani M, Sloan L, Thomas E, Billups K. 2001. Trends, Rhythms, and Aberrations in 1681 Global Climate 65 Ma to Present. Science, 292:686-693. 1682 1683 1684 75 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1685 Figure 1. Phylogenetic relationships of mygalomorph families reconstructed from morphological 1686 characters: (a) Raven’s classification scheme (Raven 1985), (b) results of the first cladistic 1687 analyses performed by Goloboff (1993). Poly- and paraphyletic lineages are indicated by grey 1688 boxes. 1689 1690 Figure 2. Phylogenetic relationships of mygalomorph families: (a) inferred from ribosomal data 1691 (18S, 28S) in the first molecular analyses of the infraorder (Hedin and Bond 2006), (b) inferred 1692 in the “total evidence” approach combining three loci (18S, 28S, EF-1g) and morphological data 1693 (Bond et al. 2012). Poly- and paraphyletic lineages are indicated by grey boxes. 1694 1695 Figure 3. Phylogenetic tree summarizing results from concatenated and species tree approaches. 1696 Topology obtained in the Maximum Likelihood analyses. Boxes on nodes denote support values 1697 obtained in each approach (left to right): RAxML bootstrap support, ExaBayes Bayesian 1698 posterior probabilities (PP), ASTRAL support values, IQ-Tree SH-aLRT support values. Color 1699 coding of boxes corresponds to distinct support level categories depicted in bottom left corner. 1700 One filled box indicates full support in all analyses; white box=topology not recovered in species 1701 tree analysis. 1702 1703 Figure 4. Phylogenetic tree inferred from combined AHE loci and transcriptomic data 1704 (“DNAAA” matrix). Tree topology obtained in maximum likelihood (ML) analyses conducted in 1705 RAxML. Asterisk marks placement of Paratropis within Crassitarsae. Boxes near the branches 1706 denote bootstrap support values; colors correspond to distinct support level categories depicted 1707 on left. 76 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1708 1709 Figure 5. Divergence time estimates inferred by treePL on a topology obtained in RAxML. The 1710 x-axis is time in million years. Geologic time scale abbreviations: N: Neogene, Pg: Paleogene, 1711 K: Cretaceous, J: Jurassic, Tr: Triassic, P: Permian, C: Carboniferous, D: Devonian. Light green 1712 block marks time frame of Gondwana breakup. Symbols on the tree nodes denote vicariant or 1713 dispersal events hypothesized for the divergences (left to right): inverted triangle=Pangea 1714 breakup, star=East-West Gondwana breakup, triangle=West Gondwana breakup, dot=Laurasia 1715 breakup, asterisk=long distance dispersal. Left corner: Map showing the position of the 1716 continents (a) prior to Gondwana breakup (adapted from (Will and Frimmel 2018), thick black 1717 lines mark the zones of continental drifting, times indicate the initiation of drifting in each zone; 1718 (b) present day. Terminal tree taxa are color coded according to geographic areas of their 1719 sampling locations depicted in the map as follows: yellow=South America, orange=Africa, 1720 purple=Asia, red=Australia, light blue=North America, dark blue=Europe, lime 1721 green=Antarctica (no taxa assigned). 1722 1723 Figure 6. Ancestral areas of distribution inferred with DEC analyses implemented in RASP. Tree 1724 topology obtained in ExaBayes. Terminal tree taxa are color coded according to the 1725 biogeographic regions of their sampling locations depicted in the map (bottom left) as follows: 1726 A=North America, B=Europe, C=Asia, D=South America, E=Africa, F=Australia. Color coding 1727 of the inferred ancestral distributions corresponds to the assigned biogeographic regions, or 1728 combination of thereof depicted in the combined area legend on the left. Biogeographic events 1729 marked on the nodes: Di=dispersal, V=vicariance, Ex=extinction; black letters: events inferred 1730 with probability > 0.7. 77 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1731 1732 Figure 7. Ancestral state reconstruction of mygalomorph foraging constructs. Preferred ancestral 1733 state reconstruction of foraging type using an equal rates model in the R-package corHMM with 1734 cursorial hunting taxa treated as inapplicable (?); tree was modified as ultrametric; AICc = 1735 173.9083. Terminal tree taxa are color coded according the type of spinning structure depicted in 1736 the legend panel of the left. 1737 1738 Figure 8. Cladogram summarizing the phylogenetic relationships of mygalomorph families 1739 recovered in this paper. The topology corresponds to the arrangement obtained in the 1740 concatenated analyses (RAxML, ExaBayes, IQ-TREE). The crossed boxes on the branches 1741 represent unsupported relationships: left ML bootstrap support < 70%, right Bayesian inference 1742 Posterior Probability< 0.95, SH-aLRT < 95). Filled box=node supported only in Bayesian 1743 inference, otherwise relationships supported in both analyses. 78 PHYLOGENY OF MYGALOMORPH SPIDERS 1744 1745 Table 1. Taxon counts obtained from the World Spider Catalog (2018), undescribed and undetermined genera marked with *. 1746 Taxon # genera # species # genera sampled in this study Actinopodidae 3 69 1 Antrodiaetidae 2 35 2 Atracidae 3 35 2 Atypidae 3 54 3 Barychelidae 42 295 3 Ctenizidae 3 54 3 Cyrtaucheniidae 11 109 4 + 1* Dipluridae 26 199 11 Euctenizidae 7 76 7 + 1* Hexathelidae 7 45 1 Halonoproctidae 6 84 5 Idiopidae 22 374 9 79 OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND Macrothelidae 1 29 1 Mecicobothriidae 4 9 1 Microstigmatidae 7 17 1 Migidae 11 97 4 Nemesiidae 45 420 18 Paratropididae 4 11 1 Porrhothelidae 1 5 1 Theraphosidae 146 989 7 + 1* Mygalomorphae 354 3006 88 1747 1748 80 PHYLOGENY OF MYGALOMORPH SPIDERS 1749 Table 2. Maximum likelihood topology test results: obs, the observed log-likelihood difference to the best unconstraint topology (Fig. 1750 3), AU, p-value of the approximately unbiased test conducted in CONSEL; np, bootstrap probability calculated from the multiscale 1751 bootstrap; significant results (P<0.05) * 1752 1753 Family/Grouping Constraint hypothesis Included taxa obs AU np Ctenizidae Ctenizidae_h1 Conothele, Cteniza, Cyclocosmia, Cyrtocarenum, Hebestatis, 2049.1 2e-05* 3e-05 874.9 3e-44* 4e-16 Latouchia, Stasimopus, Ummidia Ctenizidae_h2 Conothele, Cteniza, Cyclocosmia, Cyrtocarenum, Hebestatis, Latouchia, Ummidia Cyrtaucheniidae Ctenizidae_h3 Cteniza, Cyrtocarenum, Stasimopus 1906.6 6e-05* 1e-06 Cyrtaucheniidae_h1 Amblyocarenum, Ancylotrypa, Cyrtaucheniidae sp., Cyrtauchenius, 15966.1 3e-40* 6e-15 Homostola Dipluridae Cyrtaucheniidae_h2 Amblyocarenum, Ancylotrypa, Cyrtauchenius 3487.7 6e-05* 2e-05 Dipluridae_h1 Allothele, Australothele, Cethegus, Diplura, Euagrus, Harmonicon, 7064.5 2e-07* 2e-07 156.8 0.007* 0.006 Ischnothele, Linothele, Microhexura, Namirea, Thelechoris Dipluridae_h2 Allothele, Australothele, Cethegus, Euagrus, Ischnothele, Microhexura, Namirea, Thelechoris 81 OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND Dipluridae_h3 Bymainiella, Allothele, Australothele, Cethegus, Euagrus, Ischnothele, 157.0 0.006* 0.006 157.1 0.007* 0.006 Microhexura, Namirea, Thelechoris Dipluridae_h4_astral Porrhothele, Bymainiella, Allothele, Australothele, Cethegus, Euagrus, Ischnothele, Microhexura, Namirea, Thelechoris Hexathelidae Nemesiidae Hexathelidae_h1 Atrax, Bymainiella, Hadronyche, Macrothele, Porrhothele 4719.6 1e-58* 1e-17 Hexathelidae_h2 Bymainiella, Macrothele, Porrhothele 1281.5 8e-43* 6e-16 Hexathelidae_h3 Macrothele, Porrhothele 523.1 1e-40* 6e-15 Nemesiidae_h1 Acanthogonatus, Aname, Bayana, Calisoga, Entypesa, Iberesia, 18386.8 3e-73* 6e-20 Ixamatus, Kiama, Kwonkan, Mexentypesa, Namea, Nemesia, Pionothele, Proshermacha, Pseudoteyl, Spiroctenus, Stanwellia, Stenoterommata, Teyl 1754 1755 82 PHYLOGENY OF MYGALOMORPH SPIDERS 1756 Table 3. Divergence time estimates obtained by treePL algorithm (Smith and O’Meara 2012) representing taxa origins and the 1757 respective timeframe of their diversification inferred for tree topology depicted in Fig. 5. Confidence intervals (CI) for the highest 1758 posterior density (HDP) were inferred from 100 bootstrap replicates. Taxon origin Age HDP 95% CI Taxon diversification HDP 95% CI (Ma) (Ma) (Ma) Mygalomorphae/Liphistiomorphae 398 397-398 Antrodiaetidae + Mecicobothriidae 83 – 38 Atypoidea/Avicularoidea 323 323 Atypidae 123 – 10 Ischnothelidae 254 259 – 247 Ischnothelidae 125 – 118 Microhexuridae 225 232 – 216 Euagridae 174 – 41 Hexathelidae/ Euagridae 202 209 – 192 Macrothelidae 42 – 31 Porrhothelidae 187 194 – 176 Paratropididae 19 – 16 Macrothelidae/Bipectina 174 179 – 162 Atracidae 12 – 11 Paratropididae 166 171 – 155 Actinopodidae 50 – 36 Crassitarsae 158 163 – 147 Stasimopidae 20 – 12 Stasimopidae 146 152 – 135 Halonoproctidae 131– 4 Domiothelina 155 161– 145 Migidae 95 – 57 83 OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND Atypidae/ Antrodiaetidae + 172 176 – 167 Idiopidae 105 – 18 Halonoproctidae 143 148 – 133 Ctenizidae 83 – 64 Migidae 140 145 – 129 Euctenizidae 89 – 31 Idiopidae 133 138 – 123 Bemmeridae 49 – 9 Bemmeridae 134 139 – 126 Barychelidae 94 – 27 Ctenizidae/ Euctenizidae 121 127 – 109 Theraphosidae 95 – 9 Nemesiidae 115 120 – 107 Nemesiidae 106 – 6 Barychelidae/ Theraphosidae 107 118 – 103 Pycnothelidae 84 – 26 Atracidae/ Actinopodidae 98 102 – 89 Dipluridae 57 – 37 Pycnothelidae 102 109 – 96 Cyrtaucheniidae 70 – 1 Dipluridae/ Cyrtaucheniidae 90 98 – 85 Entypesidae 28 – 21 Anaminidae 88 92 – 81 Microstigmatidae 70 – 54 Entypesidae/Microstigmatidae 79 84 – 70 Anaminidae 47 – 5 Mecicobothriidae 1759 84 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. Migoidea Theraphosoidea Theraphosoidina Ctenizoidina Domiothelina Crassitarsae Rastelloidina Quadrithelina Migoidea ae atid Mic ros tigm dae esii rap Par The Nem ae elid elid ych ych Bar Bar ae hos idae atro pidi dae ae niid che idae ino pod ida Idio e pid ae Cte nizi dae Act Mig luri dae TUBERCULOTAE Dip (b) Ant rod iae ti Aty pida dae e Mec icob oth rida Hex e ath elid ae FORNICEPHALAE tau Atypoidina Mecicobothrioidina Cyr (a) Ant rod Aty iaetid pida ae e Mig idae Act ino Cte podida nizi e Idio dae pid Cyr a e tau c Nem heniid ae esii da Bar ych e T h e elidae rap Par hosida atro e pid Dip l u r i idae Hex dae ath Mic elidae ros Mec tigmat ida icob e oth rida e PHYLOGENY OF MYGALOMORPH SPIDERS Theraphosoidea Domiothelina Theraphosoidina Rastelloidina Crassitarsae (redefined) Atypoidina Bipectina 1760 1761 Figure 1. Phylogenetic relationships of mygalomorph families reconstructed from morphological 1762 characters: (a) Raven’s classification scheme (Raven 1985), (b) results of the first cladistic 1763 analyses performed by Goloboff (1993). Poly- and paraphyletic lineages are indicated by grey 1764 boxes. 85 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. dae nizi luri Dip Par Cte dae elid ath Dip Hex dae luri ae atro pidi dae Hex ath elid ae Idio pid ae e tida iae oth rod Ant icob Mec Aty pida e rida e (a) Mig idae Bar ych elid ae The rap Cyr hosida tau che e niid Nem ae e C y r siidae tau c Euc ten heniid Act izinae ae = ino pod i Cte nizi dae Mic dae ros tigm atid Cyr ae tau che ni Nem esii idae dae OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND “Nemesiids” ?? ae elid rap The ros ae niid ae che ych Bar tau Cyr izid ae pid Idio ten Euc idae Mig dae Cte nizi ae e elid ida ath Hex pod ino Act atro Par pidi luri Dip dae ae dae e elid tida ath Hex iae rod Ant oth icob Mec Aty pida e rida e (b) hos idae tigm atid Cyr ae tau che n iida Nem e esii dae Bipectina Mic Atypoidea Euctenizoidina “Homostolines” Theraphosoidina Crassitarsae Domiothelina Atypoidea Bipectina Avicularioidea 1765 1766 Figure 2. Phylogenetic relationships of mygalomorph families: (a) inferred from ribosomal data 1767 (18S, 28S) in the first molecular analyses of the infraorder (Hedin and Bond 2006), (b) inferred 1768 in the “total evidence” approach combining three loci (18S, 28S, EF-1g) and morphological data 1769 (Bond et al. 2012). Poly- and paraphyletic lineages are indicated by grey boxes. 86 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1770 87 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1771 Figure 3. Phylogenetic tree summarizing results from concatenated and species tree approaches. 1772 Topology obtained in the Maximum Likelihood analyses. Boxes on nodes denote support values 1773 obtained in each approach (left to right): RAxML bootstrap support, ExaBayes Bayesian 1774 posterior probabilities (PP), ASTRAL support values, IQ-Tree SH-aLRT support values. Color 1775 coding of the boxes corresponds to distinct support level categories depicted in bottom left 1776 corner. One filled box indicates full support in all analyses; white box=topology not recovered in 1777 species tree analysis. 1778 88 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1779 1780 1781 Figure 4. Phylogenetic tree of inferred combined AHE loci and transcriptomic data (“DNAAA” 1782 matrix). Tree topology obtained in maximum likelihood (ML) analyses conducted in RAxML. 1783 Asterisk marks placement of Paratropis within Crassitarsae. Boxes near the branches denote 1784 bootstrap support values; colors correspond to distinct support level categories depicted on left. 89 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1785 90 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1786 Figure 5. Divergence time estimates inferred by treePL on a topology obtained in RAxML. The 1787 x-axis is time in million years. Geologic time scale abbreviations: N: Neogene, Pg: Paleogene, 1788 K: Cretaceous, J: Jurassic, Tr: Triassic, P: Permian, C: Carboniferous, D: Devonian. Light green 1789 block marks time frame of Gondwana breakup. Symbols on the tree nodes denote vicariant or 1790 dispersal events hypothesized for the divergences (left to right): inverted triangle=Pangea 1791 breakup, star=East-West Gondwana breakup, triangle=West Gondwana breakup, dot=Laurasia 1792 breakup, asterisk=long distance dispersal. Left corner: Map showing the position of the 1793 continents (a) prior to Gondwana breakup (adapted from (Will and Frimmel 2018), thick black 1794 lines mark the zones of continental drifting, times indicate the initiation of drifting in each zone; 1795 (b) present day. Terminal tree taxa are color coded according to geographic areas of their 1796 sampling locations depicted in the map as follows: yellow=South America, orange=Africa, 1797 purple=Asia, red=Australia, light blue=North America, dark blue=Europe, lime 1798 green=Antarctica (no taxa assigned). 1799 91 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1800 92 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1801 Figure 6. Ancestral areas of distribution inferred for mygalomorph spiders with DEC analyses 1802 implemented in RASP. Tree topology obtained in ExaBayes. Terminal tree taxa are color coded 1803 according to the biogeographic regions of their sampling locations depicted in the map (bottom 1804 left) as follows: A=North America, B=Europe, C=Asia, D=South America, E=Africa, 1805 F=Australia. Color coding of the inferred ancestral distributions corresponds to the assigned 1806 biogeographic regions, or combination of thereof depicted in the combined area legend on the 1807 left. Biogeographic events marked on the nodes: Di=dispersal, V=vicariance, Ex=extinction; 1808 black letters: events inferred with probability > 0.7. 93 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1809 94 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1810 Figure 7. Ancestral state reconstruction of mygalomorph foraging constructs. Preferred ancestral 1811 state reconstruction of foraging type using an equal rates model in the R-package corHMM with 1812 cursorial hunting taxa treated as inapplicable (?); tree was modified as ultrametric; AICc = 1813 173.9083. Terminal tree taxa are color coded according the type of spinning structure depicted in 1814 the legend panel of the left. 1815 95 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1816 1817 1818 Figure 8. Cladogram summarizing the phylogenetic relationships of mygalomorph families 1819 recovered in this paper. The topology corresponds to the arrangement obtained in the 1820 concatenated analyses (RAxML, ExaBayes, IQ-TREE). The crossed boxes on the branches 1821 represent unsupported relationships: left ML bootstrap support < 70%, right Bayesian inference 1822 Posterior Probability< 0.95, SH-aLRT < 95). Filled box=node supported only in Bayesian 1823 inference, otherwise relationships supported in both analyses. 1824 96 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1825 APPENDIX 1: Revised Taxonomy for some of the families of the Spider Infraorder 1826 Mygalomorphae (Araneae) 1827 1828 SUPPLEMENTAL FILES: 1829 1830 Data available from the Dryad Digital Repository: http://dx.doi.org/10.5061/dryad.[NNNN] 1831 1832 Tables and Figures 1833 1834 Table S1: Specimen locality data 1835 1836 Table S2: Specimens used in the combined analyses of genomic and transcriptomic data 1837 (DNAAA_matrix). AHE dataset: specimens sequenced in this study, AA dataset: data 1838 proceeding from Garrison et al. (2016) “BCC-75” matrix. 1839 1840 Table S3: Character coding for web type ancestral state reconstruction 1841 1842 Supplemental Figure 1. Tree topology obtained in maximum likelihood (ML) analyses conducted 1843 in RAxML. Values on the nodes correspond to bootstrap support. 1844 1845 Supplemental Figure 2. Tree topology obtained in Bayesian inference (BI) analyses conducted in 1846 ExaBayes. Values on the nodes correspond to the Bayesian posterior probability (PP). 1847 97 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1848 Supplemental Figure 3. Tree topology obtained in maximum likelihood (ML) analyses conducted 1849 in IQ-TREE. Values on the nodes correspond to SH-aLRT support. 1850 1851 Supplemental Figure 4. Phylogenetic tree inferred after removal of Paratropididae terminals 1852 (“No_Paratropis” dataset). Tree topology obtained in maximum likelihood (ML) analyses 1853 conducted in RAxML. Boxes near the branches denote full bootstrap support values; numbers 1854 report bootstrap support values obtained in this analysis (black) and the analysis of the full 1855 “All_taxa” dataset (red). Asterisks mark differences in topology obtained from the “All_taxa” 1856 dataset. 1857 1858 Supplemental Figure 5. Tree topology obtained in species tree analysis conducted in ASTRAL. 1859 Values on the nodes correspond to the ASTRAL support values. 1860 1861 Supplemental Figure 6. Comparison of resulting tree topologies obtained by concatenated 1862 analyses conducted in RAxML and ExaBayes (left) and species tree approach conducted in 1863 ASTRAL (right). Asterisks mark differences in topologies. Doted grey line denote array of 1864 lineages forming a highly supported clade in species tree analysis. Boxes on nodes denote 1865 support values obtained in each approach, color coded according to distinct support level 1866 categories depicted in bottom left corner in following order (left to right): RAxML bootstrap 1867 support, ExaBayes Bayesian posterior probabilities, ASTRAL support values. 1868 1869 98 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1870 Supplemental Figure 7. Divergence time estimates of mygalomorph spiders inferred by treePL 1871 on a topology obtained in RAxML, values near the nodes correspond to the inferred timing of the 1872 divergences. Numbered nodes represent placement of the calibrations and their minimum and 1873 maximum ages (in brackets), 1) Mygalomorphae (Garrison et al. 2016; Fernández et al. 2018), 2) 1874 Avicularoidea: Rosamygale grauvogely (Selden and Gall 1992), 3) “Nemesioidea”: Cretamygale 1875 chasei (Selden 2002), 4) Atypoidea: Ambiortiphagus ponomarenkoi (Eskov and Zonshtein 1876 1990), 5) Ummidia (Wunderlich 2011). The x-axis is time in million years. 1877 1878 Supplemental Figure 8. Divergence time estimates of mygalomorph spiders inferred by treePL 1879 on a topology obtained in RAxML, values near the nodes correspond to the confidence intervals 1880 (CI) for the highest posterior density (HDP) inferred from 100 bootstrap replicates. The x-axis is 1881 time in million years. 1882 1883 Supplemental Figure 9. Phylogenetic tree of Mygalomorphae summarizing results from 1884 concatenated and species tree approaches within a framework of a current classification scheme 1885 (World Spider Catalog 2018). Topology obtained in the Maximum Likelihood analyses. Boxes 1886 on nodes denote support values obtained in each approach (left to right): RAxML bootstrap 1887 support, ExaBayes Bayesian posterior probabilities (PP), ASTRAL support values, IQ-Tree SH- 1888 aLRT support values. Colour of the boxes corresponds to distinct support level categories 1889 depicted in bottom left corner. One filled box indicates full support in all analyses; white 1890 box=topology not recovered in species tree analysis. 1891 1892 Data Files 1893 99 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1894 ASTRAL_all_taxa.tre 1895 ASTRAL_all_taxa.trees 1896 ExaBayes_all_taxa.tre 1897 RASP_ingroup.tre_dis.csv 1898 RAxML_DNAAA.nex 1899 RAxML_DNAAA.tre 1900 RAxML_DNAAA_partitions.txt 1901 RAxML_ExaBayes_all_taxa.nex 1902 RAxML_ExaBayes_all_taxa_partitions.txt 1903 RAxML_No_Paratropis.nex 1904 RAxML_No_Paratropis.tre 1905 RAxML_No_Paratropis_partitions.txt 1906 RAxML_all_taxa.tre 1907 silkUseAncStateRecon.txt 1908 treePL_dated.tre 100 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1909 Taxonomic Appendix: Revised Taxonomy for some of the families of the Spider Infraorder 1910 Mygalomorphae (Araneae) 1911 1912 Jason E. Bond1, Vera Opatova1, and Marshal Hedin2 1913 1914 1 Department of Entomology and Nematology, University of California, Davis, CA 95616, USA 1915 2 Department of Biology, San Diego State University, San Diego, CA, 92182–4614, USA 1916 1917 Families remaining unchanged in their composition 1918 1919 Actinopodidae – Atracidae (The Venom Clade) 1920 Remarks. The family Atracidae comprises three genera endemic to eastern Australia. Formerly 1921 designated as the “hexathelid” subfamily Atracinae (Gray 2010), the group was recently elevated 1922 to family level by Hedin et al. (2018) and supported by the topology test performed on our 1923 dataset. Contrary to the hypothesis based on morphological characters (Raven 1985; Goloboff 1924 1993), Atracidae has consistently been recovered as sister to Actinopodidae (Hedin and Bond 1925 2006; Ayoub et al. 2007; Bond et al. 2012; Hamilton et al. 2016b; Hedin et al. 2018), further 1926 corroborated by our results. Actinopodidae comprises three genera with a disjunct South 1927 American – Australian distribution and on the basis of morphological characters (eyes widely 1928 spread across the carapace), was placed as sister to Migidae (Raven 1985; Goloboff 1993). 1929 However, this relationship has never been supported in any molecular phylogenetic study. 1930 Despite the differences in the morphology of the families, the phylogenetic proximity of 101 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1931 Atracidae/Actinopodidae is supported not only by the overwhelming amount of evidence 1932 proceeding from molecular data, but also by similarities in venom composition (for details see 1933 Hedin et al. (2018). Certain members of both families are medically important; the bites of the 1934 Sydney funnel-web spider, Atrax robustus, cause lethal or life-threatening conditions, and other 1935 taxa can also be dangerous to humans (Isbister et al. 2005, 2015). The Atracidae/Actinopodidae 1936 clade was recovered within the Bipectina, which contradicts the results of a number of previous 1937 analyses (Goloboff 1993; Hedin and Bond 2006; Bond et al. 2012). Unfortunately, the exact 1938 relationships of the Atracidae/Actinopodidae with other taxa remain uncertain. The group was 1939 consistently placed in all analyses as sister to the Stasimopidae, although with varying degrees of 1940 support (see below). In our analyses, both families were represented by two out of three genera; 1941 Atrax and Hadronyche for Atracidae, and Actinopus and Missulena for Actinopodidae, however, 1942 the monophyly of both families, based on complete generic sampling, appears to be strong 1943 (Hedin et al. 2018). 1944 1945 Antrodiaetidae – Megahexuridae – Atypidae – Mecicobothriidae - Hexurellidae 1946 Remarks. Atypoid spider classification has recently undergone major revision with the 1947 designation of new families and phylogenetic structure; see Hedin et al. (2019). 1948 1949 Barychelidae 1950 Remarks. Barychelidae is a highly diverse family (42 genera, 495 species) with a Gondwanan 1951 distribution spanning continental islands and also remote islands of volcanic origin (Raven 1994) 1952 indicating the ability to disperse. Like Theraphosidae, Barychelidae possess well developed 1953 scopulae and dense claw tufts, but differs in the number of cuspules, clavate trichobothria, and 102 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 1954 the shape of the maxillary anterior lobe (Raven 1985; Dippenaar-Schoeman 2002). Due to the 1955 potential ambiguity of these characters (Raven 1985; Goloboff 1993; Bond et al. 2012) some 1956 authors have considered Barychelidae, at least in part, a paraphyletic lineage of Theraphosidae 1957 (Goloboff 1993). Stemming from differences in character state interpretations, taxa tend to 1958 vacillate between the two families, with proposed membership by some authors in one family 1959 (Raven 1985; Guadanucci 2014) to be later rejected by others (Gallon 2002; Schmidt 2002; Ríos 1960 Tamayo 2017). One example of this problem, the family placement of Brachionopus (Raven 1961 1985; Gallon 2002; Schmidt 2002) was not successfully resolved until the implementation of a 1962 molecular approach (Lüddecke et al. 2018). With our limited sampling, we recovered the 1963 reciprocal monophyly of both Barychelidae and Theraphosidae. However, a thorough sampling 1964 of both families will be needed to stabilize the family boundaries. Barychelidae (as currently 1965 defined) is composed of two common and divergent “phenotypes” (a “naked” group that is much 1966 less setose and resemble most mygalomorph groups, and a “hairy” group that resemble small 1967 theraphosids. These two “phenotypes” have been the major drivers of taxa historically bouncing 1968 back and forth between the two families. 1969 1970 Euctenizidae 1971 Remarks. The group was originally designated as a subfamily of Cyrtaucheniidae (Raven 1985), 1972 but with increasing amount of evidence for a non-monophyletic Cyrtaucheniidae (Bond and 1973 Opell 2002; Bond and Hedin 2006), Euctenizidae were elevated to family level (Bond et al. 1974 2012). The family comprises 76 species within seven genera across a North American 1975 distribution (Bond 2017). We sampled all described genera and recovered the family as 1976 monophyletic, as well as the sister group to Ctenizidae (sensu this study) and Idiopidae. 103 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 1977 1978 Halonoproctidae 1979 Remarks. The family currently comprises six genera and has a nearly cosmopolitan distribution. 1980 It was recently removed from Ctenizidae (Godwin et al. 2018), whose monophyly was doubted 1981 both on morphological (Raven 1985) and molecular grounds (Hedin and Bond 2006; Bond et al. 1982 2012; Opatova et al. 2013). Similar to Godwin et al. (2018), we recovered Halonoproctidae as 1983 monophyletic in all analyses. Furthermore, the topology test rejected the monophyly of the 1984 family Ctenizidae (sensu Raven (1985)) as well as the monophyly of Halonoproctidae + 1985 Ctenizidae (sensu this study), verifying the status of the family within the context of a wider 1986 taxonomic framework. It should be noted that we did not include the genus Bothriocyrtum in our 1987 analyses, but given its stable position as sister to Hebestatis (Hedin and Bond 2006; Bond et al. 1988 2012; Opatova et al. 2013; Godwin et al. 2018), we believe that it had little to no effect on the 1989 resulting topology and support values. 1990 1991 Hexathelidae – Porrhothelidae – Macrothelidae 1992 Remarks. The monophyly of the family Hexathelidae in its original composition (sensu Raven 1993 (1985)) has been challenged on molecular grounds on numerous occasions (Hedin and Bond 1994 2006; Ayoub et al. 2007; Bond et al. 2012; Opatova and Arnedo 2014a; Hamilton et al. 2016b). 1995 Formal taxonomic changes were put on hold until the implementation of UCEs alongside 1996 complete generic sampling which resulted in parceling of the family Hexathelidae (sensu Raven 1997 (1985)) into four distinct families (Hedin et al. 2018). We sampled all the families in our 1998 analyses and similar to the results of Hedin et al. (2018), we recovered the families Hexathelidae, 1999 Porrhothelidae and Macrothelidae among the early branching lineages near the Avicularioidea 104 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2000 root; though Atracidae, was recovered within the Bipectina clade sister to Actinopodidae. The 2001 family Hexathelidae now comprises nine genera with a disjunct distribution (Australia, New 2002 Zealand, South America), of which we only sampled the Australian Bymainiella. It was 2003 recovered as sister to the family Euagridae with high support, which was previously indicated by 2004 EF-1g data (Ayoub et al. 2007), albeit the support for this relationship was low. Both 2005 Porrhothelidae and Macrothelidae were recovered as stand-alone lineages placed near the 2006 Avicularioidea root node. Both lineages are monogeneric and were previously considered sister 2007 taxa (Raven 1985), but this relationship was statistically rejected in the past (Opatova and 2008 Arnedo 2014a) and their unique position appears stable even in light of considerable genomic 2009 data ((Hedin et al. 2018); this study). Given the wide taxonomic framework of our analyses, we 2010 tested the monophyly of the family Hexathelidae, in its original composition (sensu Raven 2011 (1985)), as well the alternative arrangement recovered in the species tree analysis (Table 2); both 2012 were statistically rejected, confirming Hedin et al. (2018). 2013 Note. Because the new family Porrhothelidae was electronically published but not registered in 2014 ZooBank, the online registration system for the ICZN it is redscribed here. 2015 2016 Family Porrhothelidae (NEW FAMILY) Hedin and Bond 2017 2018 Type genus: Porrhothele Simon, 1892 [urn:lsid:nmbe.ch:spidergen:00021] (type species 2019 Porrhothele antipodiana (Walkenaer, 1837). 2020 2021 Diagnosis: As a consequence of its monotypy, characters used to diagnose the family 2022 Porrhothelidae are those characters attributed to the type genus. Porrhothele was thoroughly 105 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2023 diagnosed and described by Forster (1968) with additions by Raven (1980). Members of this 2024 family can be diagnosed on the basis of the following unique combination of characters: 1) small 2025 posterior sternal sigilla; 2) single row of promarginal cheliceral teeth; 3) male tibia swollen with 2026 dense pattern of strong promarginal spines (illustrated in Forster (1968) e.g., p. 170, figs. 543– 2027 548). This monotypic family is found in New Zealand. 2028 2029 Idiopidae 2030 Remarks. Idiopids are another family with predominantly Gondwanan distribution (South 2031 America, Asia, Africa, Australia and New Zealand), though some species are also known from 2032 the Middle East and Central America. The family is well characterized by male palpal 2033 synapomorphies (Raven 1985) and its monophyly has never been called into question by 2034 previous studies (Goloboff 1993; Hedin and Bond 2006; Ayoub et al. 2007; Bond et al. 2012). 2035 Therefore, it comes as no surprise that the family was recovered as monophyletic in our analyses. 2036 Idiopids are among the most diverse of the mygalomorphs; the family comprises 22 genera and 2037 three subfamilies: Arbanitinae, Genysinae and Idiopinae (Raven 1985; Rix et al. 2017c), of 2038 which we did not sample Genysinae. The position of Idiopidae within the Domiothelina clade 2039 was consistently recovered by both morphological (Goloboff 1993) and molecular approaches 2040 (Bond et al. 2012), but given the incomplete sampling of the family Ctenizidae (sensu Raven 2041 (1985)), the position within the clade varied. In our analyses, the family was placed as sister to 2042 Ctenizidae and Euctenizidae with high support. 2043 2044 Migidae 106 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2045 Remarks. The family Migidae currently comprises 11 genera with a classical Gondwanan 2046 distribution (Australia, New Zealand, Africa, Madagascar and South America). The family is 2047 well defined morphologically as taxa having longitudinal keels on cheliceral fangs and lack a 2048 rastellum (Raven 1985; Griswold and Ledford 2001). Similar to previous studies (Hedin and 2049 Bond 2006; Bond et al. 2012), our taxon sampling is rather limited. We only sampled four 2050 genera, lacking a neotropical representative; therefore, we cannot assess the monophyly of the 2051 group. The family has been recovered as sister to either Stasimopus (Hedin and Bond 2006) or 2052 Halonoproctidae (Bond et al. 2012), as well as in close proximity to Idiopidae and Euctenizidae 2053 (Ayoub et al. 2007; Bond et al. 2012), as we report here. In the concatenated analyses, the family 2054 was recovered with high support as sister to all the remaining Domiothelina taxa, except for 2055 Halonoproctidae. Despite the congruent topology, this relationship was poorly supported in the 2056 species tree analysis. 2057 2058 Paratropididae 2059 Remarks. Paratropididae is a poorly known, and highly distinct family comprising four genera 2060 known from Central and South America. Its phylogenetic placement has been regarded as 2061 “difficult” (Bond et al. 2012), given the conflict between the morphology-based hypotheses and 2062 the results of molecular analyses. The family was originally placed within a clade 2063 “Theraphosoidina”, alongside the Theraphosidae and Barychelidae (Raven 1985; Goloboff 2064 1993), however this relationship has never been recovered in any molecular analyses (Hedin and 2065 Bond 2006; Bond et al. 2012; Hamilton et al. 2016b) and, in fact, a close relationship between 2066 Paratropididae and any other mygalomorph taxon has never been supported. Instead, the 2067 paratropidids were placed among the grade of early branching Avicularioidea (Hedin and Bond 107 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2068 2006; Bond et al. 2012). In our analyses, the paratropidids represent a stand-alone lineage 2069 diverging at the root of Bipectina, however, similar to previous studies (Hedin and Bond 2006; 2070 Bond et al. 2012), we only sampled representatives of the genus Paratropis. This not only 2071 precludes the evaluation of the family’s monophyly, but may also be a mitigating factor in 2072 placing it on the tree. Moreover, the stand-alone status of the lineage may also have an impact on 2073 the inferred tree topology as a whole, a fact demonstrated by the improved support values at 2074 particular nodes and the topology changes in the unsupported nodes after paratropidids were 2075 removed from the dataset (Supplemental Fig. 4). Our exploratory analysis combining AHE and 2076 transcriptomic data recovered paratropidids with high support as sister to the “Nemesioidina” 2077 clade (Fig. 5). However, the sampling in this analysis was limited both in terms of taxa and data 2078 occupancy. Despite the equivocal nature of the placement of Paratropis within the Bipectina, it 2079 seems clear that there is no support for its placement within Theraphosoidina (sensu Raven 2080 (1985)). 2081 2082 Theraphosidae 2083 Remarks. The family Theraphosidae is the most diverse family of mygalomorphs with a nearly 2084 cosmopolitan distribution. Currently comprising 144 genera and over 970 species, the family is 2085 well-defined (aside from Barychelidae) by dense claw tufts and leg scopulae (Raven 1985). The 2086 monophyly is generally undoubted (Hedin and Bond 2006; Bond et al. 2012; Lüddecke et al. 2087 2018; Turner et al. 2018), although paraphyly with respect to Barychelidae has occurred due to 2088 taxon misplacement (Goloboff 1993; Bond et al. 2012). Until recent molecular phylogenetics 2089 work (Lüddecke et al. 2018; Turner et al. 2018), attempts to reconstruct the internal relationships 2090 of Theraphosidae mostly relied on the analysis of morphological characters (Pérez-Miles 1994; 108 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2091 Pérez-Miles et al. 1996; Pérez-Miles 2002; Guadanucci 2005, 2011; Guadanucci 2014). 2092 Molecular approaches have become more common to resolve taxonomic issues at shallow levels, 2093 such as species delimitation or uncovering species-specific phylogeographic patterns (Hamilton 2094 et al. 2011; Hendrixson et al. 2013; Hamilton et al. 2014; Graham et al. 2015; Montes de Oca et 2095 al. 2016; Ortiz et al. 2018). In our analyses, we sampled eight genera from North and South 2096 America, Africa, and Asia, recovering the family as monophyletic in all the analyses. The 2097 subfamily Theraphosinae, here represented by the tribe Theraphosini (Aphonopelma, 2098 Cyrthopholis, Sericopelma and Stichoplastoris) (Turner et al. 2018), was also recovered with 2099 high support. On the other hand, our results substantiate the para- or polyphyletic status of the 2100 subfamily Ischnocolinae, already debated on the basis of morphological characters (Raven 1985; 2101 Guadanucci 2005; Guadanucci 2014; Guadanucci and Wendt 2014) and karyotype structure 2102 (Král et al. 2013). The genus Catumiri (Ischnocolinae) was recovered as sister to the rest of the 2103 theraphosid lineages, whereas Ischnocolus, regarded as the most “basal group” of the family 2104 (Raven 1985), appears to be related to Ornitoctoninae and Harpactirinae (here represented by 2105 Brachionopus). The non-monophyletic status of higher-level taxonomic units, and apparent 2106 polyphyly of some wide spread genera arises from high levels of homoplasy in the 2107 morphological characters traditionally used in Theraphosidae taxonomy (Guadanucci and Wendt 2108 2014; Gabriel and Longhorn 2015; Hamilton et al. 2016a; Turner et al. 2018). Molecular 2109 approaches and more thorough taxonomic sampling will be essential to establishing an accurate 2110 classification scheme reflecting the evolutionary relationships within the family. 2111 2112 109 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2113 *denotes genera included in analyses in the accompanying paper (Opatova et al. 2019); @ 2114 denotes that the taxon was definitively placed by another study supporting the designated 2115 placement here 2116 2117 Relimitation of the family Dipluridae Simon, 1889 2118 2119 Family Dipluridae Simon, 1889 (new circumscription) 2120 In our analyses, the family Dipluridae is represented by three South American genera Diplura, 2121 Linothele and Harmonicon belonging to the diplurid subfamily Diplurinae (Raven 1985). We 2122 recovered the Dipluridae as sister to Cyrtaucheniidae placed within the same clade alongside 2123 Nemesiidae. These results are roughly congruent with Goloboff (Goloboff 1993), who analyzed 2124 four genera of diplurids (sensu Raven (1985)) and recovered Diplura as the sister to the rest of 2125 the Bipectina clade. As in Raven (1985), Goloboff (1993) also noted shared characters in 2126 Diplura and the Crassitarsae, and hypothesized that the diplurids would be restricted to the 2127 subfamily Diplurinae in the future. Unfortunately, we did not conduct an exhaustive sampling of 2128 the “diplurid” genera and acknowledge that even after removing the Euagridae, Ischnothelidae 2129 and Microhexuridae the family likely remains non-monophyletic. Given the rough congruence 2130 between Raven’s diplurid subfamilies (Raven 1985) and the lineages removed from the 2131 Dipluridae in this study, it is possible that the distinct subfamily Masteriinae (Raven 1985; 2132 Passanha and Brescovit 2018) will be elevated to family status as well. 2133 2134 Type genus: Diplura C. L. Koch, 1850 (type species Mygale macrura Koch, 1841) 2135 110 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2136 Remarks. The family Dipluridae is relimited here to include members of the subfamilies 2137 Diplurinae Simon, 1889 and Masteriinae Simon, 1889. Of the Diplurinae genera, Trechona C. L. 2138 Koch, 1850 and Troglodiplura Main, 1969 were not included in our analysis but are nevertheless 2139 retained, at least until such time that they are included in a molecular study; Trechona’s inclusion 2140 appears justified based on stridulatory and palpal endite characters (Raven 1985), whereas Main 2141 (1993) justifies the placement of Troglodiplura in Diplurinae contra Raven (1985). The relimited 2142 family herein conservatively also includes Masteriinae which we predict is likely to be 2143 recognized as a standalone family, closely related to Microhexuridae at some point in the future; 2144 the subfamily was recently reviewed by Passanha and Brescovit (2018), however, the authors 2145 provide no insight regarding affinities to other mygalomorph taxa. The relimited Dipluridae will 2146 require a significantly revised diagnosis at some point in the future. 2147 2148 List of included subfamilies and genera 2149 2150 Diplurinae Simon, 1889 2151 *Diplura C. L. Koch, 1850 2152 *Harmonicon F. O. Pickard-Cambridge, 1896 2153 *Linothele Karsch, 1879 2154 Trechona C. L. Koch, 1850 2155 Troglodiplura Main, 1969 2156 2157 Masteriinae Simon, 1889 2158 Masteria C. L. Koch, 1873 111 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2159 Siremata Passanha and Brescovit, 2018 2160 Striamea Raven, 1981 2161 2162 Family Ischnothelidae F.O.P Cambridge, 1897 NEW RANK 2163 The family corresponds to the former diplurid subfamily Ischnothelinae, defined by six 2164 morphological synapomorphies (Raven 1985; Coyle 1995). In our analyses, we sampled the 2165 genera Ischnothele (South America) and Thelechoris (Africa) and recovered these together in all 2166 analyses. In previous analyses based on morphological characters, Ischnothele was placed as a 2167 sister group to Bipectina (Goloboff 1993), but our results suggest that ischnothelids belong 2168 among the first groups that diverged within the Avicularioidea clade. Given the morphological 2169 cohesiveness of the group, and the fact that it was thoroughly revised in the past (Coyle 1995), 2170 we retain all the genera of the subfamily Ischnothelinae (Raven 1985; Coyle 1995) despite our 2171 incomplete sampling. The family Ischnothelidae comprises five genera (Andethele, Indothele, 2172 Ischnothele, Latrothele, Thelechoris) with a disjunct distribution spanning across Central and 2173 South America, the Antilles, Africa, Madagascar, India and Sri Lanka. All genera build 2174 conspicuous three-dimensional webs and often exhibit subsocial behavioral (Coyle 1995). 2175 2176 Type genus: Ischnothele Ausserer, 1875 (type species Ischnothele caudata Ausserer, 1875) 2177 2178 Remarks. Our sampling of ischonothelines comprised two genera that span the taxonomic 2179 breadth of the subfamily (Coyle 1995, Figures 21, 22) but omit three others. As discussed by 2180 Coyle (1995) in his revision of the subfamily, the group is supported by a number of unique 2181 morphological characters that taken in combination can be used for diagnosis. Distinguishing 112 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2182 characteristics for this newly elevated family are: 1) an elongate cymbial apophysis; 2) 2183 collariform trichobothrial bases; and 3) fused silk spigots. 2184 2185 List of included genera 2186 2187 *Ischnothele Ausserer, 1875 2188 Andethele Coyle, 1995 2189 Indothele Coyle, 1995 2190 Lathrothele Benoit, 1965 2191 *Thelechoris Karsch, 1881 2192 2193 Family Euagridae Raven, 1979 NEW RANK 2194 The family roughly corresponds to the former subfamily Euagrinae (Raven 1985) that currently 2195 comprises 11 genera with distribution spanning across Australia, Asia, Africa, and South and 2196 North America. Euagrinae differ from other “diplurids” by a combined lack of cuspules and 2197 broad and short serrula. In our analyses, we only sampled five Euagrinae genera: Allothele, 2198 Australothele, Cethegus, Euagrus and Namirea, which were all recovered forming a 2199 monophyletic group towards the root node of Avicularioidea. Similar results were obtained in 2200 previous analyses (Ayoub et al. 2007; Bond et al. 2012), though with limited taxon sampling. 2201 The internal relationships of these taxa resemble their morphological affinities (Raven 1985). 2202 The “austral genera” possess a crescent of hirsute cuticle at the base of the posterior median 2203 spinnerets (“australothelinae crescent”) and were recovered as a monophyletic group sister to 2204 Euagrus that lacks this character (Raven 1985; Coyle 1988). The lack of a strong synapomorphy 113 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2205 uniting the “austral” genera and Euagrus + Phyxioschema (Raven 1985; Coyle 1988) alongside 2206 deep divergence dating back to nearly 169 Ma (174 – 161), would provide the support for the 2207 division of Euagrinae in two different families. However, the decision would be premature given 2208 our taxon sampling, which represents less than half of the described generic diversity of 2209 Diplurinae, and in particular, lacks representatives of the subfamily Masteriinae (Raven 1985) – 2210 a group that overlaps in some characters with Euagrinae. We therefore prefer to elevate the 2211 subfamily Euagrinae to family rank, retaining its current generic composition – i.e., the taxa 2212 sampled in this paper along with the genera Caledothele, Carrai, Chilehexops, Leptothele, 2213 Phyxioschema, Stenygrocercus and Vilchura. 2214 2215 Type genus: Euagrus Ausserer, 1875 (type species Euagrus mexicanus Ausserer, 1875) 2216 2217 Remarks and revised diagnosis. Unfortunately, our sampling of Euagrinae diplurids is 2218 depauperate (and should include other Euagrus species, see below) and thus our placement of the 2219 remaining diplurids in the subfamily Euagrinae follows authors subsequent to Raven’s (1985) 2220 establishment of the family (i.e., inclusion in the subfamily is maintained per Raven (1991), 2221 Coyle (1986), Raven and Schwendinger (1995), and Ríos-Tamayo and Goloboff (2017)). Note 2222 that Euagrus, was considered by Coyle (1988) to be largely central/south American with the two 2223 species, Euagrus atropurpureus (S. Africa) and Euagrus formosanus (Taiwan) as misplaced. It is 2224 important to recognize that some of these taxa may be allocated to other formerly “diplurid” 2225 families at some point in the future (e.g., Chilehexops, Caledothele, Vilchura). Although 2226 morphological evidence (sensu Coyle (1988)) seems strong that Phyxioschema is the sister group 2227 to Euagrus, the morphological cladistic analysis conducted by Ríos-Tamayo and Goloboff 114 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2228 (2017) provides little encouragement that the subfamily is monophyletic and should indicate that 2229 any morphological diagnosis of this remaining assemblage of taxa is likely without substance. 2230 Following Raven (1985), the family can be diagnosed as those taxa having a distinct spur on the 2231 second male tibia and a centrally raised tarsal organ (also following from Ríos-Tamayo and 2232 Goloboff (2017)); Euagrinae taxa differ from australothelines by lacking an australotheline 2233 crescent (see details in the diagnosis below). 2234 2235 List of included Euagrinae genera 2236 2237 *Euagrus Ausserer, 1875 2238 Caledothele Raven, 1991 2239 Chilehexops Coyle, 1986 2240 Leptothele Raven and Schwendinger, 1995 2241 Phyxioschema Simon, 1889 2242 Vilchura Ríos-Tamayo and Goloboff, 2017 2243 2244 Australothelinae NEW SUBFAMILY Bond, Opatova, and Hedin 2245 2246 Type genus: Australothele Raven, 1984 (type species Australothele maculata Raven, 1984) 2247 2248 Diagnosis: Per Raven (1985), australothelines can be diagnosed from other euagrids as those 2249 taxa having an “australotheline crescent” on the posterior median spinnerets – which comprises a 2250 “crescent of hirsute cuticle isolated by pallid glabrous anterior to bases” and lacking cymbial 115 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2251 spines (the latter character is shared by numerous other mygalomorph taxa (mentioned by Ríos- 2252 Tamayo and Goloboff (2017)). As noted by Raven (1985), a similar such crescent is also found 2253 in the putatively true diplurid genus Masteria. The subfamily is also supported as monophyletic 2254 in the analysis conducted by Ríos-Tamayo and Goloboff (2017) and Hedin et al. (2018). 2255 2256 List of genera included 2257 2258 *Australothele Raven, 1984 2259 *Allothele Tucker, 1920 2260 Carrai Raven, 1984 2261 *Cethegus Thorell, 1881 2262 *Namirea Raven, 1984 2263 Stenygrocercus Simon, 1892 2264 2265 Microhexuridae NEW FAMILY Bond, Opatova, and Hedin 2266 Family Microhexuridae (NEW FAMILY).—Microhexuridae is a monogeneric family comprising 2267 two nominal species endemic to North America. Both species are very small (> 6 mm) and have 2268 allopatric distributions restricted to the high-elevation peaks in the Appalachian Mountains and 2269 Pacific Northwest Cascade Range (Coyle 1981; Hedin et al. 2015). Despite being recognized as a 2270 diplurid (sensu Raven (1985)), Microhexura exhibits a combination of characters not present in 2271 other diplurid genera. Most notably Microhexura possess a longitudinal fovea and spinnerets 2272 somewhat shorter than in other diplurids, which made the genus difficult to place (Coyle 1981; 2273 Raven 1985). The unique position of the genus noted in the previous studies (Coyle 1981; Raven 116 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2274 1985) is further supported by our results that recovered the genus Microhexura as a stand-alone 2275 lineage among other “non-diplurine diplurids” at the base of the Avicularioidea clade. The 2276 topology test results reject both its placement within the Dipluridae (sensu Raven (1985)), but 2277 also within the “non-diplurine diplurids”/Hexathelidae/Porrhothelidae clade recovered in the 2278 species tree analyses. 2279 2280 Type genus. Microhexura Crosby and Bishop, 1925 (type species Microhexura montivaga 2281 Crosby and Bishop, 1925) 2282 2283 Diagnosis: Because the family is monogeneric, characters used to diagnose Microhexuridae are 2284 those that can be attributed to the genus by Coyle (1981). Among the smallest mygalomorph 2285 spiders known, the family is diagnosed on the basis of the following unique combination of 2286 characters: 1) longitudinal fovea; 2) four spinnerets – posterior lateral spinnerets long (nearly the 2287 length of the abdomen); 3) lacking abdominal tergites; and 4) male pedipalp lacking conductor. 2288 2289 List of included genera 2290 2291 *Microhexura Crosby and Bishop, 1925 2292 2293 Relimitation of the family Ctenizidae Thorell, 1887 2294 2295 Ctenizidae Thorell, 1887 (new circumscription) 117 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2296 The generic composition of the family Ctenizidae was dramatically altered by Raven’s (1985) 2297 morphology-based cladistic analysis, but it became apparent that the family most likely 2298 comprised an assemblage of diverse lineages that do not share a common ancestor (Hedin and 2299 Bond 2006). This observation was repeatedly confirmed by subsequent molecular-based studies 2300 (Bond et al. 2012; Opatova et al. 2013; Wheeler et al. 2017). However, a lack of perceived 2301 phylogenetic signal, incomplete sampling, or a combination thereof precluded previous authors 2302 from making taxonomic changes that would rectify the problem. The polyphyletic nature of the 2303 family was partially resolved by Godwin et al. (2018) when six genera were transferred to the 2304 family Halonoproctidae, circumscribing Ctenizidae to Cteniza, Cyrtocarenum and Stasimopus. 2305 Godwin et al. (2018) recognized that Stasimopus did not share a common ancestor with the 2306 remaining genera, but incomplete outgroup sampling limited their ability to definitively address 2307 the phylogenetic placement of the taxon. In our analyses, we recovered Cteniza and 2308 Cyrtocarenum as sister taxa, which agrees with previous findings (Opatova et al. 2013; Godwin 2309 et al. 2018). Cteniza and Cyrtocarenum were placed as sister to the family Euctenizidae, whereas 2310 Stasimopus was recovered outside of Domiothelina in a weakly supported clade as the sister 2311 lineage to Atracidae and Actinopodidae. Given the distant phylogenetic position of Stasimopus, 2312 with regards to the remaining ctenizids, and considering that the topology test rejected the 2313 monophyly of the three genera, we remove Stasimopus from the family Ctenizidae and designate 2314 it to the family Stasimopidae (NEW FAMILY). Ctenizidae is circumscribed to contain only the 2315 genera Cteniza and Cyrtocarenum. The family Ctenizidae thus now constitutes a much smaller 2316 family comprising six species with a predominantly Mediterranean distribution. 2317 2318 Type genus: Cteniza Latreille, 1829 (type species Cteniza sauvagesi (Rossi, 1788)) 118 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2319 2320 Remarks and revised diagnosis. As discussed by others (e.g., Decae 1996; Godwin et al. 2018), 2321 Cteniza and Cyrtocarenum are virtually indistinguishable from each other and are very similar to 2322 other taxa like halonoproctids and stasimopids, save some questionable eye group characteristics. 2323 As outlined by Godwin et al. (2018), Ctenizidae, as delimited herein, can be distinguished from 2324 other similar taxa by lacking curved spines on the anterior legs, having a more arched and 2325 rounded caput and flatter ventral posterior carapace, and having a more oblong abdomen (rather 2326 than ovoid); legs tend to be less stocky. Ctenizids differ from their sister group, Euctenizidae, by 2327 lacking scopulae. 2328 2329 List of included genera 2330 2331 *Cteniza Latreille, 1829 2332 *Cyrtocarenum Ausserer, 1871 2333 2334 2335 Stasimopidae NEW FAMILY Bond, Opatova, and Hedin 2336 Although the placement of Stasimopus within the family Ctenizidae (Raven 1985) was 2337 repeatedly called into question based on earlier molecular studies (Hedin and Bond 2006; Bond 2338 et al. 2012; Opatova et al. 2013; Godwin et al. 2018), establishing its exact phylogenetic position 2339 has been difficult. In previous phylogenetic studies of the entire infraorder, Stasimopus was 2340 recovered as sister to Migidae (Hedin and Bond 2006), or as the sister group to the entire 2341 Domiothelina, albeit with low support (Bond et al. 2012). Importantly, when morphological 119 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2342 characters were added to the analyses, Stasimopus was recovered within the Ctenizidae (Bond et 2343 al. 2012), confirming the homoplasious nature of some morphological characters. Our results 2344 place Stasimopus as a stand-alone lineage either forming a clade with Atracidae and 2345 Actinopodidae or as sister to Domiothelina. The sister relationship with the “Venom Clade” was 2346 also recovered by Hamilton et al. (2016b), though in a much reduced AHE dataset, and similarly 2347 to our results, the clade was poorly supported. The support values for this clade increased with 2348 the removal of the paratropidid terminals (Supplemental Fig. 4) and with addition of 2349 transcriptomic data (Fig. 4), suggesting that adding more data to the matrix could be instrumental 2350 for stabilizing Stasimopidae placement. Despite the partial positional uncertainty of Stasimopus, 2351 the topology test rejects its inclusion within Ctenizidae. Considering the results of our analyses, 2352 alongside the genus’ restricted distribution to southern Africa – a biodiversity hotspot with high 2353 levels of endemicity (Myers et al. 2000; Mittermeier et al. 2011), it seems that there is sufficient 2354 evidence to support the hypothesis that Stasimopus represents a unique lineage within the 2355 Mygalomorphae and deserves family-level status. 2356 2357 Type genus: Stasimopus Simon, 1892 (type species Stasimopus caffrus (C. L. Koch, 1842) 2358 2359 Diagnosis: At this time, the family Stasimopidae is monogeneric, thus characteristics used to 2360 diagnose the genus are the distinguishing characters for the family. Per Raven (1985) and 2361 summarized by Engelbrecht and Prendini (2012), stasimopids can be distinguished from other 2362 ctenizid-like taxa on the basis of the following combination of unique characters: 1) lacking a 2363 saddle-like depression on tibia III; 2) having a very wide eye group (twice as wide as it is long); 2364 and 3) palpal endites with a modified anterior lobe (described as “produced”). Based on the 120 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2365 breadth of morphological diversity contained within the genus, which includes somewhat bizarre 2366 carapace modifications (Engelbrecht and Prendini 2012), it is possible that Stasimopus could be 2367 split in to multiple genera at some point in the future. 2368 2369 List of included genera 2370 2371 *Stasimopus Simon, 1892 2372 2373 2374 Relimitation of the family Cyrtaucheniidae Simon, 1889 2375 2376 Cyrtaucheniidae Simon, 1889 (new circumscription) 2377 Following the results of our phylogenetic analyses, we transfer the genus Homostola to the 2378 newly established family Bemmeridae (NEW RANK, see below). We also formally transfer the 2379 Mediterranean genus Amblyocarenum to the family Nemesiidae. The family Cyrtaucheniidae is 2380 thus relimited to nine genera (three of them monotypic) distributed across Africa, Asia and South 2381 America. In this study, we analyzed two African genera: Ancylotrypa, Cyrtauchenius and one 2382 South American lineage that is morphologically proximate to Acontius. As a result, 2383 Cyrtaucheniidae were recovered as monophyletic in all analyses and placed as sister to 2384 Dipluridae. At present, we retain the genera Acontius, Anemesia, Angka, Bolostromoides, 2385 Bolostromus, Fufius, and Rhytidicolus (not included in this study), within Cyrtaucheniidae. We 2386 acknowledge that the family still likely remains poly- or paraphyletic, because Fufius is known 2387 to have affinities with the Nemesiidae (Bond et al. 2012), however, we argue that even the partial 121 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2388 Cyrtaucheniidae taxon pruning performed in this study is preferable to retaining the high level of 2389 polyphyly. This is particularly true when obtaining fresh samples may be exceptionally difficult, 2390 especially for the African and South American representatives of the genera Acontius and 2391 Bolostromus. 2392 2393 2394 Type genus: Cyrtauchenius Thorell, 1869 (type species Cyrtauchenius terricola (Lucas, 1846)) 2395 2396 Remarks. Long recognized as a polyphyletic assemblage of genera, we herein relimit 2397 Cyrtaucheniidae to the Aporoptychinae minus Kiama and Angka based on Raven’s (1985) 2398 original composition of the subfamily. Given the diversity of the taxa left remaining in the 2399 family, it is unlikely that a revised diagnosis will have any lasting value (if even possible to 2400 formulate, given the diverse array of taxa remaining) and as such we decline to provide one. We 2401 anticipate significant changes to the composition of the family as additional aporoptychines are 2402 added to the tree. We justify our transfer of Angka to Microstigmatidae based on the original 2403 description of the genus by Raven and Schwendinger (1995), where they considered it as a clear 2404 sister taxon to Kiama (also transferred, by us below, to the family Microstigmatidae); the transfer 2405 of Amblyocarenum to Nemesiidae is justified on the basis of its phylogenetic position. 2406 2407 List of included genera 2408 2409 *Cyrtauchenius Thorell, 1869 2410 *Ancylotrypa Simon, 1889 122 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2411 2412 Incertae sedis 2413 2414 Acontius Karsch, 1879 2415 Anemesia Pocock, 1895 2416 Bolostromoides Schiapelli and Gerschman, 1945 2417 Bolostromus Ausserer, 1875 2418 Fufius Simon, 1888 2419 Rhytidicolus Simon, 1889 2420 2421 Transferred to other families (not included in the relimitation of Cyrtaucheniidae) 2422 2423 *Amblyocarenum Simon, 1892 here transferred to Nemesiidae justified on the basis of 2424 phylogenetic placement. 2425 2426 @ 2427 putative sister relationship to Kiama. Angka Raven and Schwendinger, 1995 here transferred to Microstigmatidae on the basis of its 2428 2429 *Homostola Simon, 1892 here transferred to Bemmeridae justified on the basis of phylogenetic 2430 placement. 2431 2432 Relimitation of the family Nemesiidae Simon, 1889 2433 123 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2434 Nemesiidae Simon, 1889 (new circumscription) 2435 2436 The family Nemesiidae is a cosmopolitan group. Our sampling comprised five genera, including 2437 the type genus Nemesia, spanning species sampled from three continents (Africa, Europe and 2438 North America). These genera form a single lineage and are placed in a clade that is sister to all 2439 the former “nemesiid” lineages plus Cyrtaucheniidae, Dipluridae and Microstigmatidae. In 2440 agreement with our results, we transfer the genus Spiroctenus to the newly established family 2441 Bemmeridae and transfer Amblyocarenum here from Cyrtaucheniidae (see discussion on 2442 Cyrtaucheniidae for justification). We remove the genera Pionothele, Bayana, Stenoterommata, 2443 Stanwellia, Acanthogonatus, Pycnothele and Entypesa and transfer them into newly erected 2444 families Pycnothelidae (NEW RANK) and Entypesidae (NEW FAMILY). We further remove 2445 the taxa included in the Anamini tribe (Harvey et al. 2018) and elevate it to the rank of family: 2446 Anamidae (NEW RANK). We also transfer the genera Kiama, Angka, and Xamiatus to the 2447 family Microstigmatidae (see details below). We acknowledge that many taxa, particularly those 2448 inhabiting South America and Asia, remain severely under-sampled in our study and therefore 2449 we cannot speak to their definitive placement at this time; as such they remain for time being 2450 assigned to Nemesiidae incertae sedis. 2451 2452 Type genus: Nemesia Audouin, 1826 (type species Nemesia cellicola Audouin, 1826) 2453 2454 Remarks. To a greater extent than Cyrtaucheniidae (above), what remains of the family 2455 Nemesiidae is a likely para-polyphyletic assemblage of genera. As such we consider the 2456 placement of a large number of genera in the family to be uncertain at best; many of which may 124 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2457 prove to be pycnothelids. Consequently, much work remains, particularly if we are able to place 2458 many of the South American genera. 2459 2460 List of included genera 2461 2462 *Nemesia Audouin, 1826 2463 *Amblyocarenum Simon, 1892 2464 *Calisoga Chamberlin, 1937 2465 *Iberesia Decae and Cardoso, 2006 2466 *Mexentypesa Raven, 1987 2467 2468 Incertae sedis 2469 2470 Atmetochilus Simon, 1887 2471 Brachythele Ausserer, 1871 2472 Chaco Tullgren, 1905 2473 Chilelopsis Goloboff, 1995 – most likely Pycnothelidae 2474 Damarchilus Siliwal, Molur, and Raven, 2015 2475 Damarchus Thorell, 1891 2476 Diplothelopsis Tullgren, 1905 2477 Flamencopsis Goloboff, 1995 2478 Gravelyia Mirza and Mondal, 2018 2479 Hermacha Simon, 1902 – likely to be transferred to Entypesidae 125 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2480 Hermachura Mello-Leitāo, 1923 2481 Lepthercus Purcell, 1902 – likely Entypesidae 2482 Longistylus Indicatti and Lucas, 2005; removed from Anamidae contra Indicatti and Lucas 2483 (2005) 2484 Lycinus Thorell, 1894 2485 Neostothis Vellard, 1925 2486 Prorachias Mello-Leitão, 1924 2487 Psalistopoides Mello-Leitão, 1934 2488 Pselligmus Simon, 1892 2489 Rachias Simon, 1892; likely to be transferred to Pycnothelidae per Goloboff (1995) 2490 Raveniola Zonshtein, 1987 2491 Sinopesa Raven and Schwendinger, 1995 2492 2493 Transferred to other families (not included in the relimitation of Nemesiidae) 2494 2495 Xamiatus Raven, 1981; here transferred to Microstigmatidae on the basis of its putative 2496 sister/close relationship with Ixamatus and Kiama (Raven 1981,1985) 2497 Ixamatus Simon, 1887 here transferred to Microstigmatidae on the basis of its putative 2498 phylogenetic position per Harvey et al. (2018). 2499 2500 Anamidae Simon, 1889 NEW RANK 2501 The family Anamidae corresponds to the Australian endemic tribe Anamini (Raven 1985; 2502 Harvey et al. 2018) formerly placed within the “nemesiid” subfamily Anaminae (Raven 1985). 126 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2503 The family comprises genera Aname, Chenistonia, Hesperonatalius, Kwonkan, Namea, 2504 Proshermacha, Teyl, Teyloides and Swolnpes, of which we analyzed six (including Pseudoteyl 2505 currently synonymized with Teyl). The family formed a monophyletic lineage in all analyses and 2506 was recovered as sister to Entypesidae/Microstigmatidae clade. The phylogenetic relationships of 2507 the group were recently evaluated, employing both extensive taxonomic sampling and multilocus 2508 data (Harvey et al. 2018). Harvey et al. (2018) performed a complete generic sampling of the 2509 Australian Anamini and recovered the group as a well-supported clade, also including genera not 2510 sampled in our analyses. Following the results of our study and Harvey et al. (2018), we 2511 therefore transfer all the genera listed above into the family Anamidae. Morphology and 2512 biogeography do not support the inclusion of Longistylus Indicatti and Lucas, 2005 in the 2513 Anamidae contra the tentative hypothesis proposed by Indicatti and Lucas (2005). 2514 2515 Type genus: Aname L. Koch, 1873 (type species Aname pallida L. Koch, 1873) 2516 2517 Remarks. A detailed diagnosis for this new family rank taxon is provided by Harvey et al. 2518 (2018). 2519 2520 List of included genera 2521 2522 *Aname L. Koch, 1873 2523 *Proshermacha Simon, 1908 2524 @ 2525 *Kwonkan Main, 1983 (=*Yilgarnia Main, 1986) Hesperonatalius Castalanelli et al. 2017 per Harvey et al. (2018) 127 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2526 *Namea Raven, 1984 2527 @ Chenistonia Hogg 1901 per Harvey et al. (2018) 2528 @ Swolnpes Main and Framenau, 2009 per Harvey et al. (2018) 2529 *Teyl Main, 1975 2530 @ 2531 to place elsewhere) Harvey et al. (2018) Teyloides Main, 1985 but check Bond et al. (2012) and Hedin and Bond (2006) – may be able 2532 2533 Bemmeridae Simon, 1903 NEW RANK 2534 The family Bemmeridae comprises 35 species classified in the genera Homostola and 2535 Spiroctenus, both endemic to South Africa. The group was repeatedly recovered as sister to 2536 Theraphosoidina (Hedin and Bond 2006; Bond et al. 2012), which is further substantiated by our 2537 analyses. Previous analyses also suggested the inclusion of the genus Ancylotrypa (Hedin and 2538 Bond 2006), an assumption that was unfortunately based on sample misidentification. The genus 2539 Homostola is being transferred to this family from Cyrtaucheniidae, the genus Spiroctenus from 2540 Nemesiidae. Both genera comprise medium-size spiders, constructing underground burrows with 2541 or without a trapdoor. 2542 2543 Type genus: Spiroctenus Simon, 1892 (type species Spiroctenus personatus, Simon, 1888) 2544 2545 Remarks and Diagnosis. The family name is derived from the available family-level taxon 2546 Bemmereae Simon, 1903; Bemmeris Simon, 1903 is a junior synonym of Spiroctenus. The 2547 identity and consequently family-level placement of Spiroctenus and Homostola is contentious 2548 and likely to remain so. Raven (1985) considered the two genera unrelated and placed in separate 128 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2549 families (Nemesiidae and Cyrtaucheniidae respectively), attributing Hewitt’s (1915) 2550 hypothesized synonym of the two to a misidentification. To date, it would seem that a male 2551 Homostola specimen has never been described in detail, further complicating the situation (but 2552 see Hewitt (1916)). Based on careful specimen identification and locality data, we are confident 2553 that our sample comprises both genera. The Spiroctenus exemplar represents an identical male 2554 and female sequence of Spiroctenus flavopunctatus (Purcell, 1903); Homostola terminals are 2555 likely H. pardalina affin. (Hewitt, 1913) and H. vulpecula Simon, 1892. Based on our 2556 examination of specimens, Bemmeridae females can be diagnosed on the basis of the following 2557 unique combination of characters: 1) eyes on a low tubercle; 2) metatarsus and tarsus with 2558 moderate to thick scopula; 3) preening combs on leg III; 4) labium with relatively dense patch of 2559 cuspules; and 5) a rastellum comprising a row of thickened, blunt spines not borne on an 2560 apophysis. Male Spiroctenus species have a very distinct mating clasper on leg I comprising a 2561 number of large spines and an apophysis at the tibia/metatarsus junction; we would predict a 2562 similar morphology for Homostola males. 2563 2564 List of included genera 2565 2566 *Homostola Simon, 1892 2567 *Spiroctenus Simon, 1889 2568 2569 Pycnothelidae Chamberlin, 1917 NEW RANK 2570 Family Pycnothelidae (NEW RANK).—The family Pycnothelidae is newly established here in 2571 order to accommodate taxa formerly belonging to a diverse mix of “nemesiid” subfamilies 129 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2572 (Raven 1985). Some of the close relationships are not entirely novel. For example the sister 2573 relationship between Stanwellia and Acanthogonatus was previously recovered by Hedin and 2574 Bond (2006). We recovered the African genus Pionothele, formerly placed within the subfamily 2575 Bemerinnae, as sister to Spiroctenus, and sister to all pycnothelid diversity. The family 2576 distribution spans across Africa, Australia, South America and New Zealand, therefore it could 2577 be expected that more genera will be eventually transferred here following a thorough 2578 assessment, particularly of African and South American lineages. 2579 2580 Type genus: Pycnothele Chamberlin, 1917 (type species Pycnothele perdita Chamberlin, 1917) 2581 2582 Remarks and Diagnosis: The genus Pycnothele was not included in our analysis but 2583 hypothesized morphological affinities (Goloboff 1995), have attributed it to be closely aligned 2584 with Stanwellia, Acanthogonatus, and Stenoterommata. Harvey et al. (2018) stated: “If a close 2585 relationship between Stanwellia, Acanthogonatus, Stenoterommata and Pycnothele is 2586 maintained, as proposed by Goloboff (1995), the subfamily-group name Pycnothelinae 2587 Chamberlin, 1917 is available for this clade (Chamberlin 1917).” In the interests of taxonomic 2588 stability, we believe the conservative decision is to transfer Pycnothele from Nemesiidae and 2589 include it in our newly defined clade (justified on the basis of morphological similarity). The 2590 alternative would be to propose a new family rank name (e.g., “Stanwellidae”) which will be 2591 inevitably deprecated at a later time. At this time, the family Pycnothelidae can be diagnosed on 2592 the basis of the following unique combination of characters: 1) rastellum absent; 2) aspinose 2593 tarsi; 3) preening combs absent, although reduced in some Stanwellia and Acanthogonatus; 4) 2594 labium lacks cuspules; 5) tibial spur on male leg I absent; and 6) most taxa have a reduced or 130 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2595 absent inferior tarsal claw. We anticipate that this diagnosis might change as other nemesiids 2596 (noted as incertae sedis above) are added to the tree. 2597 2598 List of included genera 2599 2600 @ 2601 *Stanwellia Rainbow and Pulleine, 1918 2602 *Acanthogonatus Karsch, 1880 (note: most likely polyphyletic) 2603 *Bayana Pérez-Miles, Costa and Montes de Oca, 2014 2604 *Pionothele Purcell, 1902 2605 *Stenoterommata Holmberg, 1881 Pycnothele Chamberlin, 1917 2606 2607 Entypesidae NEW FAMILY Bond, Opatova, and Hedin 2608 2609 We establish the family Entypesidae to accommodate the genus Entypesa, which currently 2610 comprises 7 species known from Madagascar and South Africa. The genus was originally placed 2611 within the nemesiid subfamily Anaminae (Raven 1985), however this taxon has been shown to 2612 be polyphyletic (Bond et al. 2012). Despite the rather undesirable monogeneric state of the 2613 family, we believe that establishing Entypesidae as a family is justified by both its reciprocal 2614 monophyly in relation to its sister taxon Microstigmatidae and the deep divergence between 2615 them, dating back to 79 Ma. We believe the monogeneric nature of the family is only temporary 2616 and other genera (particularly fellow South African genera Lepthercus and Hermacha) will 2617 eventually be transferred here following a thorough assessment of Nemesiidae phylogeny. 131 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2618 2619 Type genus: Entypesa Simon, 1902 (type species Entypesa nebulosa Simon, 1902) 2620 2621 Diagnosis: Because the family is monogeneric, characters used to distinguish it are those that 2622 can be attributed to diagnosing the genus. Based on examination of specimens, Raven’s (1985) 2623 description and Dippenaar-Shoeman (2002), Entypesidae may be diagnosed on the basis of the 2624 following unique combination of characters: 1) narrow straight thoracic groove; 2) distinctly 2625 raised eye tubercle; 3) labium lacks cuspules; 4) weak rastellum comprising numerous short, 2626 stout bristles; 5) dense scopula on tarsi I-III; 6) leg III preening comb, present; and 7) long, 2627 digitiform spinnerets. We anticipate the inclusion of Hermacha and Lepthercus in this family at 2628 some point in the future. 2629 2630 Included genus 2631 2632 * Entypesa Simon, 1902 2633 2634 Relimitation of the family Microstigmatidae Roewer, 1942 2635 2636 Family Microstigmatidae Roewer, 1942 (new circumscription) 2637 The family Microstigmatidae is well defined by morphological characters, such as small, round 2638 booklung openings and pustulose cuticle (Raven 1985; Hedin and Bond 2006). The monophyly 2639 of the family has never been assessed within a robust molecular framework because all previous 2640 studies exclusively included the South African genus Microstigmata (Hedin and Bond 2006; 132 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. PHYLOGENY OF MYGALOMORPH SPIDERS 2641 Ayoub et al. 2007; Bond et al. 2012; Wheeler et al. 2017); but see (Goloboff 1993). Following 2642 the results of our analyses we relimit the family Microstigmatidae to include the genus Kiama. 2643 The morphological similarities in tarsal organ and cuticle structure linking Kiama to 2644 microstigmatids are well documented (Raven 1985; Bond and Opell 2002). Similar 2645 characteristics are also found in the genera Ixamatus and Angka (Bond et al. 2012) suggesting 2646 their close relationships to the microstigmatids. 2647 2648 Type genus: Microstigmata Strand, 1932 (type species Microstigmata geophila (Hewitt, 1916)) 2649 2650 Remarks: The family Microstigmatidae was traditionally diagnosed as taxa have possessing, 2651 circular booklung openings. With Goloboff’s (1995) inclusion of Pseudonemesia and 2652 Spelocteniza in the family, as well as Envia (sensu Ott and Höfer (2003)), that feature now only 2653 serves to diagnose a subset of these relatively unique genera. Although a scaly or pustulose 2654 cuticle may be a diagnostic character uniting all microstigmatids (Bond and Opell 2002), future 2655 SEM study of all genera will be necessary to confirm. As noted above, the transfer of Angka to 2656 the Microstigmatidae is based on its affinities with Kiama whose phylogenetic position clearly 2657 aligns here. 2658 2659 List of included genera 2660 2661 *Microstigmata Strand, 1932 2662 @ 2663 Envia Ott and Höfer, 2003 Angka Raven and Schwendinger, 1995 133 bioRxiv preprint doi: https://doi.org/10.1101/531756; this version posted May 2, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license. OPATOVA V, HAMILTON CA, HEDIN MH, MONTES DE OCA L, KRÁL J & J BOND 2664 @ 2665 genus has a pustulose cuticle, a potentially diagnostic character of the family. 2666 *Kiama Main and Mascord, 1969 2667 Micromygale Platnick and Forster, 1982 2668 Ministigmata Raven and Platnick, 1981 2669 Pseudonemesia Caporiacco, 1955 2670 Spelocteniza Gertsch, 1982 2671 @ 2672 and its putative sister group relationship with Kiama; as noted by Raven (1985), the genus has a 2673 pustulose cuticle, a potentially diagnostic character of the family. 2674 Xenonemesia Goloboff, 1989 Ixamatus Simon, 1887 transferred here per Harvey et al. 2018; as noted by Raven (1985), the Xamiatus Raven, 1981 transferred here on the basis of morphological similarity with Ixamatus 134