Molecular phylogeny and revision of the false violin spiders (Araneae: Drymusidae) of Africa

Facundo M Labarque; Abel Pérez-González; Charles E Griswold

doi:10.1093/zoolinnean/zlx088

Molecular phylogeny and revision of the false violin spiders (Araneae: Drymusidae) of Africa

Labarque, Facundo M;Pérez-González, Abel;Griswold, Charles E 2017-12-14 00:00:00 Abstract The family Drymusidae includes 16 cryptic spiders that build irregular webs in dark places. The family is distributed in South Africa, the Neotropical and Andean regions. Here, we use a molecular approach to infer the relationships of Drymusidae using three mitochondrial (COI, 16S, 12S) and three nuclear (H3, 28S, 18S) markers. Our preferred analyses support Drymusidae and its American and African clades, which emerge as sister groups. Our analyses suggest a Gondwanan distribution of Drymusidae and a Westward radiation of Izithunzi gen. nov. within South Africa, but both hypotheses remain to be thoroughly tested. We describe Izithunzi gen. nov. for the African species. All previous African species are redescribed and new combinations are proposed: Izithunzi capense (Simon) comb. nov., Izithunzi productum (Purcell) comb. nov. and Izithunzi silvicola (Purcell) comb. nov. Two new species are described: Izithunzi lina sp. nov. (known from both sexes) and Izithunzi zondii sp. nov. (known only from females). The male of I. productum (Purcell) comb. nov. is also described for the first time. We consider Loxosceles valida Lawrence, 1964, a junior synonym of I. capense (Simon) comb. nov. (new synonymy). We also provide a dichotomous key for Izithunzi gen. nov. species. Gondwana, target-genes, Scytodoidea, Synspermiata INTRODUCTION African Drymusidae Simon, 1893, are shy, cryptic spiders that build irregular, tangled webs beneath logs and stones and in other dark, damp places, such as rock overhangs and caves (Fig. 1; Larsen, 1994). Because of their retiring nature, most will never be encountered by people. Chance encounters with the huge species from Table Mountain may elicit surprise and concern due to their size and resemblance to the venomous violin spiders (Sicariidae Keyserling, 1880: Loxosceles Heineken & Lowe, 1832) (Purcell, 1904; Larsen, 1994; Binford & Wells, 2003). This resemblance has led to the common name ‘false violin spiders’ for the African Drymusidae (Dippenaar-Schoeman & Jocqué, 1997; Filmer, 2010). In spite of this morphological similarity, and close relationship to Sicariidae in the superfamily Scytodoidea, Drymusidae species show no evidence of being dangerous to humans (Binford & Wells, 2003). Figure 1. View largeDownload slide African Drymusidae habitat. Izithunzi capense comb. nov. (♀, A–C CASENT 9023625), Izithunzi aff. lina sp. nov. (♀, D), live specimens. A–C, Newlands Forest Preserve, specimens hanging beneath loose space webs inside a cave (photographs by Norman Larsen) (C specimen feeding on Theridion delicatum O. Pickard-Cambridge, 1904). D, Helderberg Nature Reserve, specimen on log (photograph by Gonzalo Giribet). Figure 1. View largeDownload slide African Drymusidae habitat. Izithunzi capense comb. nov. (♀, A–C CASENT 9023625), Izithunzi aff. lina sp. nov. (♀, D), live specimens. A–C, Newlands Forest Preserve, specimens hanging beneath loose space webs inside a cave (photographs by Norman Larsen) (C specimen feeding on Theridion delicatum O. Pickard-Cambridge, 1904). D, Helderberg Nature Reserve, specimen on log (photograph by Gonzalo Giribet). Drymusidae include the genus Drymusa Simon, 1891, with 16 extant species described so far, four of them from South Africa (World Spider Catalog, 2017). The other Drymusidae species are distributed across the Neotropical (11 species) and Andean regions (one species) (Morrone, 2004, 2014; World Spider Catalog, 2017). The first Drymusidae described for South Africa was Drymusa capensis Simon, 1893, from Table Mountain (Western Cape). The original description was brief, based on one immature holotype specimen, and lacks important morphological details, for example, those of genitalic organs. Purcell (1904) was the first to mention the female of D. capensis (although he did not describe it), along with the descriptions of two new species for South Africa: Drymusa producta Purcell, 1904 (originally known from female specimens) and Drymusa silvicola Purcell, 1904 (originally known from both sexes). He also provided a key to females (Purcell, 1904: 154). Recently, Lotz (2012) proposed the new combination, Drymusa valida (Lawrence, 1964), after he transferred this species (originally described in Loxosceles, both sexes known) from Sicariidae to Drymusidae. This species is also known from Table Mountain (i.e. sympatric with D. capensis). Drymusidae have long been associated with violin spiders (as Loxoscelidae or Sicariidae) and spitting spiders (Scytodidae Blackwall, 1864) (Simon, 1893; Gertsch, 1967; Alayón, 1981; Lehtinen, 1986). Cladistic analyses over the last two decades (Platnick et al., 1991; Ramírez, 2000; Wheeler et al., 2016) have suggested the sister group relationship of Drymusidae and Scytodidae and confirmed their distinctness from Loxosceles. A recent morphological cladistic analysis (Labarque & Ramírez, 2012) proposed that the Australasian Periegopidae Simon, 1893, are the sister group of Drymusidae. Nevertheless, the molecular cladistic results of the Assembling the Tree of Life (AToL): Phylogeny of Spiders Project (Wheeler et al., 2016) suggested that the traditional limits of Drymusidae are paraphyletic with respect to Periegopidae and that the African species placed in Drymusa are more closely related to Periegops Simon, 1893, from Australia and New Zealand than to other Drymusa from the Americas. Suggestions that Drymusidae may include more than one genus have been made by Lehtinen (1967) and Platnick et al. (1991), especially by the former, who reviewed Neotropical and South African species and considered them not congeneric. In this study, we describe Izithunzi gen. nov. for the African species formerly placed in Drymusa. The male of Izithunzi productum (Purcell, 1904) comb. nov. is described for the first time, and D. valida is considered a junior synonym of Izithunzi capense (Simon, 1893) comb. nov. (new syn.). We describe Izithunzi lina sp. nov. (male and female) and Izithunzi zondii sp. nov. (female) and redescribe and key all five species. Furthermore, in cooperation with SANSA (South African National Survey of Arachnida), we map and produce a *.kml file documenting all known records of Drymusidae in South Africa (see Supporting Information). With the aim to investigate the monophyly of Izithunzi gen. nov., examine the delimitation of contentious genera and determine the placement of Drymusidae among Scytodoidea, we generated a molecular data set for Drymusidae species (i.e. Neotropical, Andean and South African) including six markers: three nuclear genes (the ribosomal 28S rRNA and 18S rRNA and the protein-coding histone H3) and three mitochondrial genes [the ribosomal 16S rRNA and 12S rRNA and the protein-coding cytochrome c oxidase subunit I (COI)]. These sequences were included in a major data set, the spider AToL Synspermiata data subset (Wheeler et al., 2016), and analysed under maximum likelihood, Bayesian inference and parsimony, separately and concatenated. Our results highly support Drymusidae as monophyletic, the sister group relationship between the American and African clades and the monophyly of both of those clades. MATERIAL AND METHODS Molecular data The molecular data set included 108 species, six of which were sequenced in the present study and 102 were extracted from the spiders AToL project (Wheeler et al., 2016). Synspermiata is represented by 99 species in 68 genera (Supporting Information, Table S1). Outgroup terminals included Hypochilus pococki Platnick, 1987 (Hypochilidae Marx, 1888) and eight genera of Filistatidae Ausserer, 1867. Total DNA was extracted from one or two legs of freshly collected specimens that were preserved in 95–100% ethanol or propylene glycol (i.e. Drymusa armasi Alayón, 1981). Extraction, amplification and sequencing followed the protocols described by Polotow, Carmichael & Griswold (2015) and Wheeler et al. (2016). Partial fragments of the mitochondrial genes cytochrome c oxidase subunit I (COI, ranging in length from 321 to 1076 bp), 12S rRNA (12S, from 169 to 348 bp) and 16S rRNA (16S, from 323 to 453 bp) and the nuclear genes 28S rRNA (18S, from 312 to 1729 bp), 28S rRNA (28S, from 507 to 2236 bp) and Histone H3 (H3, from 295 to 318 bp) were amplified. Primer pairs and PCR annealing conditions for each locus are listed in Table 1. Table 1. Primer sequences, source and annealing temperatures used to generate data for this study Gene Direction Sequence Reference Annealing temperature COI Fwd: LC011490-oono 5′-CWA CAA AYC ATA RRG ATA TTG G-3′ Modified from Folmer et al. (1994) 50–54 °C Rev: C1-N-2191 5′-CCC CGT AAA ATT AAA ATA TAA ACT TC-3′ Simon et al. (1994) H3 Fwd: H3aF 5′-ATG GCT CGT ACC AAG CAG ACV GC-3′ Colgan et al. (1998) 52 °C Rev: H3aR 5′-ATA TCC TTRGGC ATR ATRGTG AC-3′ Colgan et al. (1998) 28S Fwd: 28S-0sc 5′-CGT GAA ACT GCT CAG AGG-3′ Modified from Hedin & Maddison (2001) 50–53 °C Rev: 28S-C 5′-GGT TCG ATT AGT CTT TCG CC-3′ Hedin & Maddison (2001) 18S Fwd: 18S-1F 5′-TAC CTG GTT GAT CCT GCC AGT AG-3′ Giribet et al. (1996) 50 °C Rev: 18S-5-9-intR 5′-ATT CCG WTA ACG ADC GAG-3′ Miller et al. (2010) 18S Fwd: 18S-5F 5′-GCG AAA GCA TTT GCC AAG AA-3′ Giribet et al. (1996) 50 °C Rev: 18S-9R 5′-GAT CCT TCC GCA GGT TCA CCT AC-3′ Giribet et al. (1996) Gene Direction Sequence Reference Annealing temperature COI Fwd: LC011490-oono 5′-CWA CAA AYC ATA RRG ATA TTG G-3′ Modified from Folmer et al. (1994) 50–54 °C Rev: C1-N-2191 5′-CCC CGT AAA ATT AAA ATA TAA ACT TC-3′ Simon et al. (1994) H3 Fwd: H3aF 5′-ATG GCT CGT ACC AAG CAG ACV GC-3′ Colgan et al. (1998) 52 °C Rev: H3aR 5′-ATA TCC TTRGGC ATR ATRGTG AC-3′ Colgan et al. (1998) 28S Fwd: 28S-0sc 5′-CGT GAA ACT GCT CAG AGG-3′ Modified from Hedin & Maddison (2001) 50–53 °C Rev: 28S-C 5′-GGT TCG ATT AGT CTT TCG CC-3′ Hedin & Maddison (2001) 18S Fwd: 18S-1F 5′-TAC CTG GTT GAT CCT GCC AGT AG-3′ Giribet et al. (1996) 50 °C Rev: 18S-5-9-intR 5′-ATT CCG WTA ACG ADC GAG-3′ Miller et al. (2010) 18S Fwd: 18S-5F 5′-GCG AAA GCA TTT GCC AAG AA-3′ Giribet et al. (1996) 50 °C Rev: 18S-9R 5′-GAT CCT TCC GCA GGT TCA CCT AC-3′ Giribet et al. (1996) COI, mitochondrial cytochrome c oxidase subunit I; 28S, nuclear 28S rDNA; 18S, nuclear 18S rDNA; H3, nuclear histone 3; Fwd, forward; Rev, reverse. View Large Table 1. Primer sequences, source and annealing temperatures used to generate data for this study Gene Direction Sequence Reference Annealing temperature COI Fwd: LC011490-oono 5′-CWA CAA AYC ATA RRG ATA TTG G-3′ Modified from Folmer et al. (1994) 50–54 °C Rev: C1-N-2191 5′-CCC CGT AAA ATT AAA ATA TAA ACT TC-3′ Simon et al. (1994) H3 Fwd: H3aF 5′-ATG GCT CGT ACC AAG CAG ACV GC-3′ Colgan et al. (1998) 52 °C Rev: H3aR 5′-ATA TCC TTRGGC ATR ATRGTG AC-3′ Colgan et al. (1998) 28S Fwd: 28S-0sc 5′-CGT GAA ACT GCT CAG AGG-3′ Modified from Hedin & Maddison (2001) 50–53 °C Rev: 28S-C 5′-GGT TCG ATT AGT CTT TCG CC-3′ Hedin & Maddison (2001) 18S Fwd: 18S-1F 5′-TAC CTG GTT GAT CCT GCC AGT AG-3′ Giribet et al. (1996) 50 °C Rev: 18S-5-9-intR 5′-ATT CCG WTA ACG ADC GAG-3′ Miller et al. (2010) 18S Fwd: 18S-5F 5′-GCG AAA GCA TTT GCC AAG AA-3′ Giribet et al. (1996) 50 °C Rev: 18S-9R 5′-GAT CCT TCC GCA GGT TCA CCT AC-3′ Giribet et al. (1996) Gene Direction Sequence Reference Annealing temperature COI Fwd: LC011490-oono 5′-CWA CAA AYC ATA RRG ATA TTG G-3′ Modified from Folmer et al. (1994) 50–54 °C Rev: C1-N-2191 5′-CCC CGT AAA ATT AAA ATA TAA ACT TC-3′ Simon et al. (1994) H3 Fwd: H3aF 5′-ATG GCT CGT ACC AAG CAG ACV GC-3′ Colgan et al. (1998) 52 °C Rev: H3aR 5′-ATA TCC TTRGGC ATR ATRGTG AC-3′ Colgan et al. (1998) 28S Fwd: 28S-0sc 5′-CGT GAA ACT GCT CAG AGG-3′ Modified from Hedin & Maddison (2001) 50–53 °C Rev: 28S-C 5′-GGT TCG ATT AGT CTT TCG CC-3′ Hedin & Maddison (2001) 18S Fwd: 18S-1F 5′-TAC CTG GTT GAT CCT GCC AGT AG-3′ Giribet et al. (1996) 50 °C Rev: 18S-5-9-intR 5′-ATT CCG WTA ACG ADC GAG-3′ Miller et al. (2010) 18S Fwd: 18S-5F 5′-GCG AAA GCA TTT GCC AAG AA-3′ Giribet et al. (1996) 50 °C Rev: 18S-9R 5′-GAT CCT TCC GCA GGT TCA CCT AC-3′ Giribet et al. (1996) COI, mitochondrial cytochrome c oxidase subunit I; 28S, nuclear 28S rDNA; 18S, nuclear 18S rDNA; H3, nuclear histone 3; Fwd, forward; Rev, reverse. View Large Sequence reconciliation, edition and chromatogram evaluation were performed using Geneious 7.1.5 (Biomatters; http://www.geneious.com/). Each sequence was checked for contamination using an NCBI BLAST search (http://ncbi.nlm.nih.gov/BLAST). Low-quality samples with ambiguous readings were removed from the matrix. In addition, sequences were managed in Bioedit ver. 7.0.5.2 (Hall, 1999). Taxonomic and sequence information for the study specimens are listed in Supporting Information (Table S1). Automatic multiple alignments were built using the E-INS-i very slow search strategy in MAFFT ver. 7 (Katoh & Standley, 2013), which incorporates affine gap costs. Several gap opening (GOP, ‘Gap opening penalty’) costs were investigated to assess their influence on the results. The following parameter costs were used (GOP): 1.00, 1.53 (default), 3.00. These values resulted in several alignments ranging from numerous short gaps (‘gappy’) to fewer and longer gaps in the sequences (‘compacted’). The minimum value of the rescaled incongruence length difference (RILD) was used as a criterion to select the optimal alignment for each gene fragment (Wheeler & Hayashi, 1998). The RILD measures the increase in homoplasy generated as a result of combining data sets; hence, selecting the alignment combination that minimizes the RILD value maximizes the congruence among partitions. To avoid overweighting contiguous gap positions, gaps were recoded as presence/absence characters using the simple code method by Simmons & Ochoterena (2000) as implemented in the program Gapcoder (Young & Healy, 2002). Phylogenetic analyses Maximum likelihood analysis was performed using the program RAxML ver. 7.3.0 (Stamatakis, 2006; Stamatakis, Hoover & Rougemont, 2008) and was run remotely at the CIPRES Science Gateway (Miller, Pfeiffer & Schwartz, 2010; https://www.phylo.org/). Ribosomal genes (12S, 16S, 18S and 28S) were each treated as a single partition. Three different partition schemes were implemented for the protein-coding genes (COI and H3), namely unpartitioned (each gene a single partition), 3rd codon (two partitions: 1st + 2nd codons, and 3rd codon) and codon specific (one partition for each codon position). The partition corresponding to the gaps scored as absence/presence characters was treated as binary. Each gene partition was assigned an unlinked general time-reversible model with gamma distributed among invariable sites (GTR + Γ + I). The best likelihood tree was obtained out of 100 random iterations, and support was assessed by conducting 1000 non-parametric bootstrap replicates for each analysis. Both independent gene trees and concatenated analyses were run under the different partition schemes. Bayesian analyses were performed using MrBayes 3.2.6 (Ronquist & Huelsenbeck, 2003), and the program jModeltest ver. 2.1.6 (Guindon & Gascuel, 2003; Darriba et al., 2012) was used to select the best-fitting model of evolution for each partition using Akaike’s information criterion (Akaike, 1973; Buckley et al., 2002). Both programs were run remotely at the CIPRES Science Gateway. Partition schemes were implemented as in the previous paragraph and evaluated using Bayes factors (BF) (Brown & Lemmon, 2007). Acceptance or rejection of each strategy was based on the following cut-off: BF ≥ 10 (strong evidence against the competing hypothesis), 10 < BF ≥ −10 (ambiguous, select least complex strategy) and BF ≤ −10 (strong evidence for the competing hypothesis). Recoded gaps were treated with a standard discrete model. The substitution estimates were allowed to vary independently between each partition. For phylogenetic analysis, two independent runs with four simultaneous Markov chain Monte Carlo chains (one cold and three heated), each with random starting trees, were conducted simultaneously, sampling every 1000 generations until the SD of the split frequencies of these two runs dropped below 0.01 (107 generations). The program Tracer ver. 1.5 (Rambaut & Drummund, 2009) was used to ensure that the Markov chains had reached stationarity by examining the effective sample size values (> 200) and to determine the correct number of generations to discard as a burn-in (first 10% of generations). Parsimony analyses were performed using TNT version 1.5 (Goloboff, Farris & Nixon, 2008a; Goloboff & Catalano, 2016). These data were analysed under both equal and implied weights (Goloboff, 1993) with concavity constants of the weighting function values of 6, 20, 50, 100, 500 and 1000, as suggested by Goloboff et al. (2008b). All tree searches were driven to independently hit 15 times the optimal scoring, using the default values of the ‘New Technologies’ search in TNT with sectorial searches, tree fusing and ratchet, followed by TBR branch swapping (string of commands hold 20000; xmult = hits 15 ratchet 10; bb = fillonly;). Support values were estimated by jackknifing (J) frequencies; each of the 1000 pseudoreplicates used three random additional sequences plus TBR, followed by TBR collapsing to calculate the consensus. Trees were edited with the program FigTree ver. 1.4.3 (Rambaut, 2006–2016) and WinClada. All analyses were rooted with the branch separating H. pococki and Filistatidae from the remaining taxa. The data sets can be obtained in TreeBase with the accession number 21309. Morphological data The examined material is deposited in the following institutions (abbreviation and curators in parentheses): American Museum of Natural History, New York, NY, USA (AMNH, L. Prendini); California Academy of Sciences, San Francisco, CA, USA (CAS, L. Esposito), Field Museum of Natural History, Chicago, IL, USA (FMNH, P. Sierwald); Muséum National d’Histoire Naturelle, Paris, France (MNHN, C. Rollard); National Collection of Arachnida, Plant Protection Research Institute, Pretoria, South Africa (NCP, A. Dippenaar-Schoeman, P. Marais and R. Lyle); National Museum, Bloemfontein (NMBA, L. Lotz); Natal Museum, Pietermaritzburg, Kwa-Zulu Natal, South Africa (NMSA, D. Herbert, A. Ndaba and C. Stoffels); and Iziko Museum of Capetown (formerly South African Museum), Cape Town, South Africa (SAMC, S. van Noort and D. Larsen). Morphological observations were made using Leica M165 C or M125 stereomicroscopes and Olympus BH-2 or Leica DM4000 M compound microscopes. Light micrographs were prepared using a Nikon DXM1200 digital camera mounted on a Nikon SMZ1500 or Leica MZ 16 stereomicroscope, a Leica DFC 500 digital camera mounted on a Leica M216 and M165 C stereomicroscope or a Leica DM 4000 compound microscope. Extended focal range images were composed with Leica Application Suite version 3.6.0. and Helicon Focus 3.10.3, 4.01 Pro and 4.62 Pro (Khmelik, Kozub & Glazunov, 2006). Female vulva illustrations were drawn in Adobe Illustrator (Adobe Systems) over digital images using a Wacom drawing table (Wacom Co. Ltd.). The male pedipalp was checked for haematodochal expansion by placing it in a vial of hot 92% lactic acid solution (Sigma–Aldrich, St Louis, MO, USA) that was placed into boiling water for 2–3 min, after which the pedipalp was transferred to distilled water (Ledford et al., 2011; Griswold, Audisio & Ledford, 2012). The internal anatomy of the pedipalpal bulb and course of the reservoir were examined by immersing the bulb in a solution of methyl salicylate (synthetic oil of wintergreen), from Mallunckrodt Baker Chemicals, Philipsburg, NJ, USA (Miller, Griswold & Yin, 2009). Preparations were carefully cleaned using fine brushes, a thin jet of alcohol from a thinned pipette or in an ultrasonic cleaner; some setae were removed to expose structures, especially those on the legs and pedipalp, spinnerets and chelicerae. For scanning electron microscopy (SEM), all preparations were dehydrated in a series of increasing concentrations of ethanol (80, 90, 95 and 100%) and critical point dried. After drying and brushing, they were mounted on adhesive copper tape (Electron Microscopy Sciences, EMS 77802) affixed to a stub and secured with a conductive paint of colloidal graphite on isopropyl alcohol base (EMS 12660). Prior to SEM examination under high vacuum with an FEI XL30 TMP or Leo 1450VP Scanning Electron Microscope, the structures were sputter coated with Au-Pd. Measurements were taken with a micrometric ocular and are given in millimetres. Female genitalia were digested after dissection with KOH, pancreatin (Álvarez-Padilla & Hormiga, 2008) or contact lens cleaner solution (ReNu) and observed in clove oil or lactic acid. Coloration patterns are described based on specimens preserved in 80% ethanol. Reciprocal species diagnoses were made between sister species when possible. Discussed morphological characters are used ad hoc and were not included in the phylogenetic analyses. Coordinates given in brackets were obtained with Google Earth and The World Coordinate Converter (http://twcc.free.fr/) and those written in italic should be considered broad approximations. Cape Town’s cave coordinates were obtained at geoview.info (http://za.geoview.info/), and information regarding caves’ names obtained at Cape Peninsula Speleological Society web page (http://cpss.caving.org.za/) and from Craven (2012). RESULTS A summary of results, statistics and conditions for each phylogenetic analysis is presented in Table 2. We analysed the gene fragments separately and in combination. Results from the different phylogenetic analyses are summarized in the form of a preferred working tree (Fig. 2; see Congruence and selection of working tree). We focus our discussion on the groups with higher support, hereby defined as those with a bootstrap proportions (BS) > 0.75, Bayesian posterior probability (PP) > 0.95 and jackknifing proportions > 0.75 (denoted with black in circles in Fig. 2 and Supporting Information, Fig. S1), and also on clades that were with low support but were consistently recovered in likelihood, Bayesian and parsimony analyses (grey circles in Fig. 2 and Supporting Information, Fig. S1). Table 2. Summary tree statistics and conditions for each analysis Analysis Optimality criterion, Software Conditions Statistics COI Parsimony, TNT v1.5 EW 24 trees, 6431 steps Histone 3 EW 8 trees, 1788 steps 12S rDNA + gaps EW 11 trees, 4838 steps 16S rDNA + gaps EW 56 trees, 5981 steps 18S rDNA + gaps EW 138 trees, 2783 steps 28S rDNA + gaps EW 10 trees, 8783 steps Six-genes concatenated + gaps EW 36 trees, 31532 steps IW k = 6 1 tree, 31637 steps, fit 1343.78383 IW k = 20 1 tree, 31573 steps, fit 720.10758 IW k = 50 1 tree, 31539 steps, fit 374.70066 IW k = 100 1 tree, 31534 steps, fit 210.02483 IW k = 500 1 tree, 31532 steps, fit 46.75595 IW k = 1000 1 tree, 31532 steps, fit 23.72344 Reduce six-genes concatenated + gaps 84 trees, 29 373 steps COI (unpartitioned) Likelihood, RAxML v7.3.0 1000 non-parametric bootstrap replicates −lnL 24507.664201 COI (3rd codon) −lnL 24018.961966 COI (codon specific) −lnL 23605.298821 Histone 3 (unpartitioned) −lnL 7172.273998 Histone 3 (3rd codon) −lnL 6923.293562 Histone 3 (codon specific) −lnL 6838.500314 12S rDNA + gaps −lnL 17083.095847 16S rDNA + gaps −lnL 22180.794164 18S rDNA + gaps −lnL 16446.281902 28S rDNA + gaps −lnL 40176.652458 Six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 132036.031678 Reduce six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 124 410.091077 COI (unpartitioned) Bayesian, Mr. Bayes v.3.2.6 50000000 generations, burn-in = 10% sdsf 0.006877 COI (3rd codon) sdsf 0.007477 COI (codon specific) sdsf 0.010772 Histone 3 (unpartitioned) sdsf 0.004276 Histone 3 (3rd codon) sdsf 0.003138 Histone 3 (codon specific) sdsf 0.005141 12S rDNA + gaps sdsf 0.005199 16S rDNA + gaps sdsf 0.003927 18S rDNA + gaps sdsf 0.007870 28S rDNA + gaps sdsf 0.023531 Six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps 30000000 generations, burn-in = 10% sdsf 0.030925 Reduce six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps sdsf 0.001968 Analysis Optimality criterion, Software Conditions Statistics COI Parsimony, TNT v1.5 EW 24 trees, 6431 steps Histone 3 EW 8 trees, 1788 steps 12S rDNA + gaps EW 11 trees, 4838 steps 16S rDNA + gaps EW 56 trees, 5981 steps 18S rDNA + gaps EW 138 trees, 2783 steps 28S rDNA + gaps EW 10 trees, 8783 steps Six-genes concatenated + gaps EW 36 trees, 31532 steps IW k = 6 1 tree, 31637 steps, fit 1343.78383 IW k = 20 1 tree, 31573 steps, fit 720.10758 IW k = 50 1 tree, 31539 steps, fit 374.70066 IW k = 100 1 tree, 31534 steps, fit 210.02483 IW k = 500 1 tree, 31532 steps, fit 46.75595 IW k = 1000 1 tree, 31532 steps, fit 23.72344 Reduce six-genes concatenated + gaps 84 trees, 29 373 steps COI (unpartitioned) Likelihood, RAxML v7.3.0 1000 non-parametric bootstrap replicates −lnL 24507.664201 COI (3rd codon) −lnL 24018.961966 COI (codon specific) −lnL 23605.298821 Histone 3 (unpartitioned) −lnL 7172.273998 Histone 3 (3rd codon) −lnL 6923.293562 Histone 3 (codon specific) −lnL 6838.500314 12S rDNA + gaps −lnL 17083.095847 16S rDNA + gaps −lnL 22180.794164 18S rDNA + gaps −lnL 16446.281902 28S rDNA + gaps −lnL 40176.652458 Six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 132036.031678 Reduce six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 124 410.091077 COI (unpartitioned) Bayesian, Mr. Bayes v.3.2.6 50000000 generations, burn-in = 10% sdsf 0.006877 COI (3rd codon) sdsf 0.007477 COI (codon specific) sdsf 0.010772 Histone 3 (unpartitioned) sdsf 0.004276 Histone 3 (3rd codon) sdsf 0.003138 Histone 3 (codon specific) sdsf 0.005141 12S rDNA + gaps sdsf 0.005199 16S rDNA + gaps sdsf 0.003927 18S rDNA + gaps sdsf 0.007870 28S rDNA + gaps sdsf 0.023531 Six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps 30000000 generations, burn-in = 10% sdsf 0.030925 Reduce six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps sdsf 0.001968 The preferred analyses for each optimality criteria are given in bold. COI, mitochondrial cytochrome c oxidase subunit I; EW, equal weights; Gaps, number of gap absence/presence characters added; IW, implied weights; k, constant of cavity. View Large Table 2. Summary tree statistics and conditions for each analysis Analysis Optimality criterion, Software Conditions Statistics COI Parsimony, TNT v1.5 EW 24 trees, 6431 steps Histone 3 EW 8 trees, 1788 steps 12S rDNA + gaps EW 11 trees, 4838 steps 16S rDNA + gaps EW 56 trees, 5981 steps 18S rDNA + gaps EW 138 trees, 2783 steps 28S rDNA + gaps EW 10 trees, 8783 steps Six-genes concatenated + gaps EW 36 trees, 31532 steps IW k = 6 1 tree, 31637 steps, fit 1343.78383 IW k = 20 1 tree, 31573 steps, fit 720.10758 IW k = 50 1 tree, 31539 steps, fit 374.70066 IW k = 100 1 tree, 31534 steps, fit 210.02483 IW k = 500 1 tree, 31532 steps, fit 46.75595 IW k = 1000 1 tree, 31532 steps, fit 23.72344 Reduce six-genes concatenated + gaps 84 trees, 29 373 steps COI (unpartitioned) Likelihood, RAxML v7.3.0 1000 non-parametric bootstrap replicates −lnL 24507.664201 COI (3rd codon) −lnL 24018.961966 COI (codon specific) −lnL 23605.298821 Histone 3 (unpartitioned) −lnL 7172.273998 Histone 3 (3rd codon) −lnL 6923.293562 Histone 3 (codon specific) −lnL 6838.500314 12S rDNA + gaps −lnL 17083.095847 16S rDNA + gaps −lnL 22180.794164 18S rDNA + gaps −lnL 16446.281902 28S rDNA + gaps −lnL 40176.652458 Six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 132036.031678 Reduce six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 124 410.091077 COI (unpartitioned) Bayesian, Mr. Bayes v.3.2.6 50000000 generations, burn-in = 10% sdsf 0.006877 COI (3rd codon) sdsf 0.007477 COI (codon specific) sdsf 0.010772 Histone 3 (unpartitioned) sdsf 0.004276 Histone 3 (3rd codon) sdsf 0.003138 Histone 3 (codon specific) sdsf 0.005141 12S rDNA + gaps sdsf 0.005199 16S rDNA + gaps sdsf 0.003927 18S rDNA + gaps sdsf 0.007870 28S rDNA + gaps sdsf 0.023531 Six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps 30000000 generations, burn-in = 10% sdsf 0.030925 Reduce six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps sdsf 0.001968 Analysis Optimality criterion, Software Conditions Statistics COI Parsimony, TNT v1.5 EW 24 trees, 6431 steps Histone 3 EW 8 trees, 1788 steps 12S rDNA + gaps EW 11 trees, 4838 steps 16S rDNA + gaps EW 56 trees, 5981 steps 18S rDNA + gaps EW 138 trees, 2783 steps 28S rDNA + gaps EW 10 trees, 8783 steps Six-genes concatenated + gaps EW 36 trees, 31532 steps IW k = 6 1 tree, 31637 steps, fit 1343.78383 IW k = 20 1 tree, 31573 steps, fit 720.10758 IW k = 50 1 tree, 31539 steps, fit 374.70066 IW k = 100 1 tree, 31534 steps, fit 210.02483 IW k = 500 1 tree, 31532 steps, fit 46.75595 IW k = 1000 1 tree, 31532 steps, fit 23.72344 Reduce six-genes concatenated + gaps 84 trees, 29 373 steps COI (unpartitioned) Likelihood, RAxML v7.3.0 1000 non-parametric bootstrap replicates −lnL 24507.664201 COI (3rd codon) −lnL 24018.961966 COI (codon specific) −lnL 23605.298821 Histone 3 (unpartitioned) −lnL 7172.273998 Histone 3 (3rd codon) −lnL 6923.293562 Histone 3 (codon specific) −lnL 6838.500314 12S rDNA + gaps −lnL 17083.095847 16S rDNA + gaps −lnL 22180.794164 18S rDNA + gaps −lnL 16446.281902 28S rDNA + gaps −lnL 40176.652458 Six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 132036.031678 Reduce six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 124 410.091077 COI (unpartitioned) Bayesian, Mr. Bayes v.3.2.6 50000000 generations, burn-in = 10% sdsf 0.006877 COI (3rd codon) sdsf 0.007477 COI (codon specific) sdsf 0.010772 Histone 3 (unpartitioned) sdsf 0.004276 Histone 3 (3rd codon) sdsf 0.003138 Histone 3 (codon specific) sdsf 0.005141 12S rDNA + gaps sdsf 0.005199 16S rDNA + gaps sdsf 0.003927 18S rDNA + gaps sdsf 0.007870 28S rDNA + gaps sdsf 0.023531 Six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps 30000000 generations, burn-in = 10% sdsf 0.030925 Reduce six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps sdsf 0.001968 The preferred analyses for each optimality criteria are given in bold. COI, mitochondrial cytochrome c oxidase subunit I; EW, equal weights; Gaps, number of gap absence/presence characters added; IW, implied weights; k, constant of cavity. View Large Figure 2. View largeDownload slide Preferred tree obtained in the maximum likelihood analysis of the combined molecular alignments excluding ambiguously aligned regions, with the preferred model for each gene fragment and gaps scored as absence/presence. The information of the results of the analyses under Bayesian (first 10% of generations were removed as burn-in) and parsimony (equal weights) are summarized in the node support. Circles at nodes indicate support levels subdivided by analysis (L, maximum likelihood; B, Bayesian; P, parsimony: bottom; see inset). Black indicates bootstrap proportions > 0.75, posterior probabilities > 0.95 and jackknifing proportions > 0.75; grey indicates that the clade was recovered but with lower support than the previous values; white indicates that the clade was not recovered. Supported families and genera discussed in the text are highlighted with lateral bars and shaded boxes, respectively. Figure 2. View largeDownload slide Preferred tree obtained in the maximum likelihood analysis of the combined molecular alignments excluding ambiguously aligned regions, with the preferred model for each gene fragment and gaps scored as absence/presence. The information of the results of the analyses under Bayesian (first 10% of generations were removed as burn-in) and parsimony (equal weights) are summarized in the node support. Circles at nodes indicate support levels subdivided by analysis (L, maximum likelihood; B, Bayesian; P, parsimony: bottom; see inset). Black indicates bootstrap proportions > 0.75, posterior probabilities > 0.95 and jackknifing proportions > 0.75; grey indicates that the clade was recovered but with lower support than the previous values; white indicates that the clade was not recovered. Supported families and genera discussed in the text are highlighted with lateral bars and shaded boxes, respectively. New molecular data The molecular data set includes six species sequenced in the present study, yielding 18 sequences all available in GenBank (Supporting Information, Table S1): Drymusa serrana Goloboff & Ramírez, 1992 (H3 MF49776, COI MF497758), D. armasi (H3 MF564217, 28S MF506838, 18S MF506830), Izithunzi silvicola comb. nov. (Purcell, 1904) (H3 MF497760, COI MF497757, 28S MF506837, 18S MF506834), I. capense comb. nov. (H3 MF497759, COI MF497755, 28S MF506835, 18S MF506832), I. lina sp. nov. (COI MF497754, 28S MF506836, 18S MF506831) and I. productum comb. nov. (COI MF497756, 18S MF506833). Analysis of molecular data The alignment of the protein-coding genes COI and H3 was trivial because they show no evidence of indel mutations. For the 12S and 18S ribosomal genes, the alignment with the parameter value GOP = 1.53 minimized RILD when combined with the unambiguously aligned protein-coding genes, whereas for the 16S and 28S ribosomal genes, the preferred parameter value was GOP = 1 (Table 3). The total length of the concatenated aligned data matrix was 7630 characters, of which 6790 corresponded to nucleotides (1078 bp of COI, 320 bp of H3, 401 bp of 12S, 532 bp of 16S, 1795 bp of 18S, 2664 bp of 28S) and 840 to gaps coded as absence/presence (111 from 12S, 150 from 16S, 87 from 18S, 492 from 28S). Table 3. Summary of the results of the parsimony analysis of the alignments obtained with MAFFT under different parameter combinations and similarity values with the results of the combined data matrix (ribosomal gene + gaps + protein-coding genes) GOP T L Gaps ∑ Lc ILD MaxLc RILD 12S 1.00 4 4813 112 13033 13305 272 18603 0.0488330341 1.53 (default) 3 4838 111 13 058 13 322 264 18 635 0.0473372781 3.00 4 4864 85 13084 13349 265 18664 0.0474910394 16S 1.00 9 5981 150 14 201 14 444 243 20 667 0.0375811939 1.53 1 5998 126 14218 14468 250 20712 0.0384970742 3.00 3 6106 105 14326 14580 254 20882 0.0387431361 18S 1.00 15 2771 90 10991 11251 260 16292 0.0490473496 1.53 22 2783 87 11 003 11 259 256 16 317 0.048174633 3.00 25 2793 89 11013 11271 258 16338 0.0484507042 28S 1.00 5 8783 492 17 003 17 369 366 25 657 0.0422925815 1.53 2 8912 464 17132 17509 377 25835 0.043318396 3.00 11 9029 423 17249 17625 376 26090 0.0425291257 GOP T L Gaps ∑ Lc ILD MaxLc RILD 12S 1.00 4 4813 112 13033 13305 272 18603 0.0488330341 1.53 (default) 3 4838 111 13 058 13 322 264 18 635 0.0473372781 3.00 4 4864 85 13084 13349 265 18664 0.0474910394 16S 1.00 9 5981 150 14 201 14 444 243 20 667 0.0375811939 1.53 1 5998 126 14218 14468 250 20712 0.0384970742 3.00 3 6106 105 14326 14580 254 20882 0.0387431361 18S 1.00 15 2771 90 10991 11251 260 16292 0.0490473496 1.53 22 2783 87 11 003 11 259 256 16 317 0.048174633 3.00 25 2793 89 11013 11271 258 16338 0.0484507042 28S 1.00 5 8783 492 17 003 17 369 366 25 657 0.0422925815 1.53 2 8912 464 17132 17509 377 25835 0.043318396 3.00 11 9029 423 17249 17625 376 26090 0.0425291257 The preferred combinations are given in bold. 12S, mitochondrial 12S rDNA; 16S, 16S rDNA; 18S, nuclear 18S rDNA; 28S, 28S rDNA; GOP, gap opening cost; T, number of most parsimonious trees; L, tree length; gaps, number of gap absence/presence characters added; ∑, summary of the tree length of each partition; Lc, tree length of the combined data matrix; ILD, incongruence length difference; MaxLc, maximum tree length of the combined data matrix; RILD, rescaled incongruence length difference. Table 3. Summary of the results of the parsimony analysis of the alignments obtained with MAFFT under different parameter combinations and similarity values with the results of the combined data matrix (ribosomal gene + gaps + protein-coding genes) GOP T L Gaps ∑ Lc ILD MaxLc RILD 12S 1.00 4 4813 112 13033 13305 272 18603 0.0488330341 1.53 (default) 3 4838 111 13 058 13 322 264 18 635 0.0473372781 3.00 4 4864 85 13084 13349 265 18664 0.0474910394 16S 1.00 9 5981 150 14 201 14 444 243 20 667 0.0375811939 1.53 1 5998 126 14218 14468 250 20712 0.0384970742 3.00 3 6106 105 14326 14580 254 20882 0.0387431361 18S 1.00 15 2771 90 10991 11251 260 16292 0.0490473496 1.53 22 2783 87 11 003 11 259 256 16 317 0.048174633 3.00 25 2793 89 11013 11271 258 16338 0.0484507042 28S 1.00 5 8783 492 17 003 17 369 366 25 657 0.0422925815 1.53 2 8912 464 17132 17509 377 25835 0.043318396 3.00 11 9029 423 17249 17625 376 26090 0.0425291257 GOP T L Gaps ∑ Lc ILD MaxLc RILD 12S 1.00 4 4813 112 13033 13305 272 18603 0.0488330341 1.53 (default) 3 4838 111 13 058 13 322 264 18 635 0.0473372781 3.00 4 4864 85 13084 13349 265 18664 0.0474910394 16S 1.00 9 5981 150 14 201 14 444 243 20 667 0.0375811939 1.53 1 5998 126 14218 14468 250 20712 0.0384970742 3.00 3 6106 105 14326 14580 254 20882 0.0387431361 18S 1.00 15 2771 90 10991 11251 260 16292 0.0490473496 1.53 22 2783 87 11 003 11 259 256 16 317 0.048174633 3.00 25 2793 89 11013 11271 258 16338 0.0484507042 28S 1.00 5 8783 492 17 003 17 369 366 25 657 0.0422925815 1.53 2 8912 464 17132 17509 377 25835 0.043318396 3.00 11 9029 423 17249 17625 376 26090 0.0425291257 The preferred combinations are given in bold. 12S, mitochondrial 12S rDNA; 16S, 16S rDNA; 18S, nuclear 18S rDNA; 28S, 28S rDNA; GOP, gap opening cost; T, number of most parsimonious trees; L, tree length; gaps, number of gap absence/presence characters added; ∑, summary of the tree length of each partition; Lc, tree length of the combined data matrix; ILD, incongruence length difference; MaxLc, maximum tree length of the combined data matrix; RILD, rescaled incongruence length difference. The maximum likelihood analysis of the alternative partition schemes revealed better likelihoods for codon-specific partitions for both the COI and H3 genes (Table 2). Maximum likelihood analyses of the complete concatenated data matrix with a/p gaps (C-ML) yielded one single tree with a score of −lnL = −132036.032 (Table 2; Supporting Information, Fig. S1). Summary results of the evolutionary models selected by Akaike’s information criterion in JModeltest are presented in Table 4 and for BF partitioning strategies in Table 5. Comparison of the alternative partitioning strategies revealed a strong preference for the 3rd codon partition in COI and a strong preference for the unpartitioned partition in H3 (Table 5). Bayesian analysis of independent genes achieved convergence within 5 × 107 generations, while that of concatenated genes achieved convergence within 3 × 107 generations (i.e. due to the matrix size) (mostly all SD of split frequencies < 0.01; see Table 2). The Bayesian results of the complete concatenated matrix (C-BI) mostly differ from the maximum likelihood results in the basal nodes of the clade uniting Scytodoidea and among the Lost Tracheal clade members (clade A; Supporting Information, Fig. S1). Table 4. Evolutionary models used for Bayesian analysis as selected by AIC in Jmodeltest ver. 2.1.6 Partition Model selected AIC score Model used COI GTR + I + G 24120.0236 GTR + I + G H3 TVM + I + G 7035.7133 GTR + I + G 12S TIM2 + I + G 15799.2414 GTR + I + G 16S TVM + I + G 20417.1281 HKY + I + G 18S TIM2 + I + G 15592.0110 GTR + I + G 28S GTR + I + G 35977.6747 GTR + I + G Partition Model selected AIC score Model used COI GTR + I + G 24120.0236 GTR + I + G H3 TVM + I + G 7035.7133 GTR + I + G 12S TIM2 + I + G 15799.2414 GTR + I + G 16S TVM + I + G 20417.1281 HKY + I + G 18S TIM2 + I + G 15592.0110 GTR + I + G 28S GTR + I + G 35977.6747 GTR + I + G COI, mitochondrial cytochrome c oxidase subunit I; 12S, mitochondrial 12S rDNA; 16S, mitochondrial 16S rDNA; 18S, nuclear 18S rDNA; 28S, nuclear 28S rDNA; H3, nuclear histone 3; AIC, Akaike information criterion. Table 4. Evolutionary models used for Bayesian analysis as selected by AIC in Jmodeltest ver. 2.1.6 Partition Model selected AIC score Model used COI GTR + I + G 24120.0236 GTR + I + G H3 TVM + I + G 7035.7133 GTR + I + G 12S TIM2 + I + G 15799.2414 GTR + I + G 16S TVM + I + G 20417.1281 HKY + I + G 18S TIM2 + I + G 15592.0110 GTR + I + G 28S GTR + I + G 35977.6747 GTR + I + G Partition Model selected AIC score Model used COI GTR + I + G 24120.0236 GTR + I + G H3 TVM + I + G 7035.7133 GTR + I + G 12S TIM2 + I + G 15799.2414 GTR + I + G 16S TVM + I + G 20417.1281 HKY + I + G 18S TIM2 + I + G 15592.0110 GTR + I + G 28S GTR + I + G 35977.6747 GTR + I + G COI, mitochondrial cytochrome c oxidase subunit I; 12S, mitochondrial 12S rDNA; 16S, mitochondrial 16S rDNA; 18S, nuclear 18S rDNA; 28S, nuclear 28S rDNA; H3, nuclear histone 3; AIC, Akaike information criterion. Table 5. Results of Bayes factor hypothesis testing for partitioning strategies Partition −lnL harmonic mean (post burn-in) Bayes factor Support for rejection COI (codon specific) −23331.07 – – COI (3rd codon) −22758.04 −573.03 None COI (unpartitioned) −22894.01 −437.06 None H3 (codon specific) −6925.449 – – H3 (3rd codon) −6862.995 −62.454 None H3 (unpartitioned) −6679.128 −246.321 None Partition −lnL harmonic mean (post burn-in) Bayes factor Support for rejection COI (codon specific) −23331.07 – – COI (3rd codon) −22758.04 −573.03 None COI (unpartitioned) −22894.01 −437.06 None H3 (codon specific) −6925.449 – – H3 (3rd codon) −6862.995 −62.454 None H3 (unpartitioned) −6679.128 −246.321 None The preferred hypotheses are given in bold. COI, mitochondrial cytochrome c oxidase subunit I; H3, nuclear histone 3. Table 5. Results of Bayes factor hypothesis testing for partitioning strategies Partition −lnL harmonic mean (post burn-in) Bayes factor Support for rejection COI (codon specific) −23331.07 – – COI (3rd codon) −22758.04 −573.03 None COI (unpartitioned) −22894.01 −437.06 None H3 (codon specific) −6925.449 – – H3 (3rd codon) −6862.995 −62.454 None H3 (unpartitioned) −6679.128 −246.321 None Partition −lnL harmonic mean (post burn-in) Bayes factor Support for rejection COI (codon specific) −23331.07 – – COI (3rd codon) −22758.04 −573.03 None COI (unpartitioned) −22894.01 −437.06 None H3 (codon specific) −6925.449 – – H3 (3rd codon) −6862.995 −62.454 None H3 (unpartitioned) −6679.128 −246.321 None The preferred hypotheses are given in bold. COI, mitochondrial cytochrome c oxidase subunit I; H3, nuclear histone 3. In the parsimony analyses of the complete data matrix, the heuristic search strategies under equal weights (C-EW) produced 36 trees of length 31532. As expected, milder down-weighting against homoplasy (higher k values) resulted in topologies more similar to equally weighted trees. The k = 1000 and 500 were identical to each other and to one of the 36 equal weighted trees, but the topology of the k = 6, 20, 50 and 100 analyses was slightly different in the resolution of a few groups with low bootstrap values. Generally, the parsimony results of the concatenated molecular matrix mostly differ from the maximum likelihood analysis regarding nodes with low support, especially in clade A (Supporting Information, Fig. S1). Congruence and selection of working tree To explore congruence and sensitivity among all the trees produced by different optimality criteria, we summarized our results in the maximum likelihood tree obtained from the concatenated alignment matrix as ‘Navajo rugs’ (Supporting Information, Fig. S2; Giribet & Edgecomb, 2006). These analyses indicated that no individual gene reliably anticipates the results from the concatenated alignment. As expected, the incongruent nodes also showed low support (Supporting Information, Figs S1, S2; Rindal & Brower, 2011). Five terminals were very unstable across analyses, and few had very long branches within clade A: D. armasi, cf. Psiloderces sp. MA186 (Ochyroceratidae Fage, 1912), Theotima sp. MR15 (Ochyroceratidae), Stedocys pagodas Labarque, Grismado, Ramírez, Yan & Griswold, 2009 (Scytodidae) and Usofila sp. MR71 (Telemidae Fage, 1913). The removal of these five terminals from the data set and then running various phylogenetic algorithms (e.g. maximum likelihood, Bayesian, parsimony) produce an increment in the support values within clade A. Furthermore, these results also recovered Scytodoidea and the Lost Tracheal Clade as monophyletic, contrary to the results of the complete data set (compare Fig. 2 against Supporting Information, Fig. S1). Therefore, we present these analyses of 103 terminals as our preferred hypothesis, which are summarized in a maximum likelihood tree (Fig. 2). Our working tree reflects that of Wheeler et al. (2016: fig. 3), which mostly agrees with the taxonomic and transcriptomic congruences computed in that work. Notably, our preferred tree results from unconstrainted analyses, contrary to Wheeler et al. (2016: fig. 3; although their unconstrained and constrained maximum likelihood trees were very similar). Maximum likelihood analysis of the reduced, concatenated data matrix with a/p gaps (R-ML) yielded one single tree with a score of −lnL = −124410.091 (Fig. 2; Table 2). Bayesian analysis of the reduced, concatenated genes (R-BI) achieved convergence within 3 × 107 generations (SD of split frequencies < 0.01). In the parsimony analysis of the reduced, concatenated matrix, the heuristic search strategies under equal weights (R-EW) produced 84 trees of length 29373. DISCUSSION Synspermiata relationships and clade support Our preferred tree topology (Fig. 2) mostly resembles that of Wheeler et al. (2016: fig. 3), except for a few differences. The R-ML and R-EW analyses recovered Trogloraptoridae Griswold, Audosio & Ledford, 2012 and Caponiidae Simon, 1890, as sister groups, but both families emerged as sister to the other members of all Synspermiata instead of only Dysderoidea (Fig. 2; Wheeler et al., 2016: fig. 3). Furthermore, in our C-ML and C-EW analyses, Caponiidae emerged as sister to the rest of Synspermiata, while Trogloraptoridae were recovered as sister to Dysderoidea (Fig. 2; although in the R-BI and C-BI analyses Caponiidae go to an odd place, as sister to Filistatidae with PP = 0.93). Most analyses (except R-BI and C-BI) recovered Oonopidae Simon, 1890 and Dysderidae C. L. Koch, 1837, as sister groups (Fig. 2), contrary to Wheeler et al. (2016), which showed a close relationship between Orsolobidae Cooke, 1965 and Dysderidae. However, these relationships have low support in both cladistic analyses (Fig. 2; Wheeler et al., 2016: fig. 3). Orchestina Simon, 1882 was recovered in all our analyses, but it was not resolved as monophyletic in Wheeler et al. (2016). Our working tree also yielded higher average support values than that of Wheeler et al. (2016: fig. 3), highly supporting several clades only recovered (i.e. with low support) by the former. Oonopidae and the informal higher gamasomorphines (following Grismado, Ramírez & Izquierdo, 2014) were supported by most of our analyses, except by parsimony (i.e. R-EW and C-EW). The clade including Tetrablemmidae O. Pickard-Cambridge, 1873, Plectreuridae Simon, 1893, Diguetidae, F. O. Pickard-Cambridge, 1899, and Pacullidae Simon, 1893, was also supported by most of our analyses, except by the R-BI. All reduced data set analyses supported Ochyroceratidae and Scytodidae (Fig. 2). Synspermiata was supported by the R-ML and C-ML analyses (although Caponiidae were recovered as the sister of Filistatidae in the R-BI and C-BI analyses, see previous paragraph). The R-ML and R-BI analyses supported the genus Sicarius Walckenaer, 1847, and the group formed by Scytodoidea and the Lost Tracheal clade (Fig. 2: clade A). Scytodoidea were supported by the R-BI analysis. The clade including Plectreuridae, Diguetidae and Pacullidae was recovered by our results (i.e. R-ML and R-BI) and by Wheeler et al. (2016: fig. 3). Ochyroceratidae as member of Scytodoidea Ochyroceratidae and Scytodidae were recovered as sister groups, and each family is well supported by R-ML, R-BI and R-EW analyses (Fig. 2). Note that the low support values obtained for each family in Wheeler et al. (2016: fig. 3 – Ochyroceratidae BS = 55, Scytodidae BS = 39) were due to the use of inconsistent terminals in their analyses (i.e. cf. Psiloderces sp. and S. pagodas). Our hypothesis (Ochyroceratidae plus Scytodidae) agrees with two morphological characters including bipectinated proclaws on tarsi I–II (Lehtinen, 1986; Labarque & Ramírez, 2012; Pérez-González, Rubio & Ramírez, 2016: fig. 8G) and a distal hood on podotarsite (Labarque et al., 2017); characters also shared by Drymusidae and Periegopidae (Fig. 3A, B; Labarque & Ramírez, 2012; Labarque et al., 2017). However, Ochyroceratidae lack several of the features considered to be Scytodoidea morphological synapomorphies, including Non-homoplastic: (1) tracheae with fused third opisthosomal apodemes (Fig. 3C; Ramírez, 2000: ch. 32; Labarque & Ramírez, 2012: ch. 53); (2) an enlarged subchromosomal space (Michalik & Ramírez, 2014: ch. 11); Homoplastic: (3) the cheliceral promarginal lobe (Fig. 3D; Labarque & Ramírez, 2012: ch. 28); (4) female cheliceral stridulatory ridges and pedipalpal femoral thorns (Fig. 3E, F; Labarque & Ramírez, 2012: ch. 32); (5) a conical acrosomal vacuole (Michalik & Ramírez, 2014: ch. 8); (6) the ends of the acrosomal filament clearly after the axonemal basis (Michalik & Ramírez, 2014: ch. 12); and (7) the indented anterior pole of the nucleus (Michalik & Ramírez, 2014: ch. 14). Figure 3. View largeDownload slide Scytodoidea and Drymusidae plus Periegopidae putative synapomorphies. Izithunzi capense comb. nov. (♀, A–C CASENT 9023625), Izithunzi silvicola comb. nov. (♀, D–F CASENT-9048599), preserved specimens. A, B, right feet I (C, retrolateral; D, apical). C, tracheae, dorsal. D–E, left chelicerae detail [D, promargin; E, ectal (corner, shallow scales, detail)]. F, left pedipalp, prolateral, tarsal tip (*, curved setae; corner, femoral thorn, detail). Abbreviations: ant., anterior; pro., prolateral. Scale bar: C, E, 0.2 mm; A, B, 0.1 mm; D, F, 0.05 mm. Figure 3. View largeDownload slide Scytodoidea and Drymusidae plus Periegopidae putative synapomorphies. Izithunzi capense comb. nov. (♀, A–C CASENT 9023625), Izithunzi silvicola comb. nov. (♀, D–F CASENT-9048599), preserved specimens. A, B, right feet I (C, retrolateral; D, apical). C, tracheae, dorsal. D–E, left chelicerae detail [D, promargin; E, ectal (corner, shallow scales, detail)]. F, left pedipalp, prolateral, tarsal tip (*, curved setae; corner, femoral thorn, detail). Abbreviations: ant., anterior; pro., prolateral. Scale bar: C, E, 0.2 mm; A, B, 0.1 mm; D, F, 0.05 mm. Drymusidae monophyly Drymusidae and Periegopidae were recovered as sister groups in the R-ML analysis (Fig. 2) agreeing with Labarque & Ramírez (2012) but contrary to Wheeler et al. (2016: fig. 3), who recovered Drymusidae paraphyletic including Periegopidae in their preferred tree (although in several of their analyses Drymusidae was recovered monophyletic). The R-ML and R-BI analyses supported Drymusidae (Fig. 2), but the family was recovered polyphyletic in our complete data set analyses (Supporting Information, Fig. S1). The Antillean D. armasi emerged as sister to the Ochyrocera species in the C-ML and C-BI analyses, and it was related to S. pagodas and Usofila sp. in the C-EW analysis. Independent gene analyses recovered D. armasi with Ochyrocera Simon, 1892, (i.e. using nuclear ribosomal genes), with Scytodidae (i.e. in parsimony analyses) and forming a polytomy with several Synspermiata members (i.e. in Bayesian analyses). Despite these results, most analyses, except the C-BI analysis, showed low support for the polyphyly of Drymusidae (Supporting Information, Fig. S1). Drymusa armasi was finally discarded from the working data set due to its instability across the analyses. We also tested the COI and 28S sequences of Drymusa dinora Valerio, 1971, (EU817714 and EU817719, respectively; Binford et al., 2008), another Caribbean Drymusidae (analyses not shown). We found several indels in the COI sequence; therefore, this sequence was discarded. The maximum likelihood analysis of the 28S supported D. dinora as sister to D. armasi and both species were recovered within Ochyrocera with low support, mirroring our C-ML results (Supporting Information, Fig. S1). Binford et al. (2008) found that members of Scytodes Latreille, 1804, Usofila Keyserling, 1891, and Drymusa, especially D. dinora, presented very long branches in their analyses. We arrived at the same conclusions; hence, we discarded the 28S sequence of D. dinora, D. armasi, S. pagodas and Usofila sp. from our complete and reduced data sets. Despite the low support in our molecular results and in those of Wheeler et al. (2016), Drymusidae, that is, Neotropical, Andean and South African taxa, and Periegopidae are united by morphological synapomorphies, as suggested by a previous cladistic analysis (Labarque & Ramírez, 2012; although Caribbean Drymusidae were not tested in that analysis, the synapomorphies are confirmed here). Drymusidae plus Periegopidae presented (Labarque & Ramírez, 2012) the following morphological synapomorphies: Non-homoplastic: (1) venom outlet of cheliceral fang facing anteriorly (Fig. 3D; but see comments in Labarque & Ramírez, 2012: ch. 34); Homoplastic: (2) regain of cheliceral promarginal teeth (Fig. 3D; Labarque & Ramírez, 2012: ch. 24). From the present analysis, we suggest several additional putative synapomorphies for this clade including: (3) a prolateral tooth row on the superior retroclaws of pretarsi I–II (Fig. 3A); (4) the foot articulated, with the podotarsite subdivided by an additional articulation (Fig. 3B; Labarque et al., 2017: ch. 1, fig. 1I); (5) regain of cheliceral retromarginal teeth (Fig. 4A); and (6) dense brushes on metatarsi III–IV (Fig. 4B). Despite efforts to find morphological synapomorphies for Drymusidae, only a few homoplastic characters may be tentatively suggested so far: (1) regain of a second major ampullate gland spigot on the anterior lateral spinnerets (instead of a nubbin) (Fig. 4C; Labarque & Ramírez, 2012: ch. 70); (2) regain of long deferent ducts on the testis (Michalik & Ramírez, 2014: ch. 4); (3) regain of cleistospermia (Michalik & Ramírez, 2014: ch. 40). After close examination of American Drymusa and several other Synspermiata, we here suggest an additional potential synapomorphy for Drymusidae: (4) inner spermathecae of vulva connected to the uterus externus through a duct (Fig. 4D; except in D. serrana). Figure 4. View largeDownload slide Drymusidae plus Periegopidae putative synapomorphies. Izithunzi capense comb. nov. (♀, A CASENT 9023625), Periegops suterii (Urquhart, 1892) (♀, B AMNH), Izithunzi productum comb. nov. (♀, C CASENT 9043066), Izithunzi silvicola comb. nov. (♀, D CASENT 9021765), preserved specimens. A, left chelicerae detail, retromargin. B, left tarsus-metatarsus IV, ventral. C, right ALS field, detail (corner, spinneret field diagram). D, right spermathecae, detail. Abbreviations: MaAm, major ampullate; Pi, piriform; retro., retrolateral. Scale bar: B, 0.5 mm; A, 0.2 mm; C, 0.09 mm; D, 0.05 mm. Figure 4. View largeDownload slide Drymusidae plus Periegopidae putative synapomorphies. Izithunzi capense comb. nov. (♀, A CASENT 9023625), Periegops suterii (Urquhart, 1892) (♀, B AMNH), Izithunzi productum comb. nov. (♀, C CASENT 9043066), Izithunzi silvicola comb. nov. (♀, D CASENT 9021765), preserved specimens. A, left chelicerae detail, retromargin. B, left tarsus-metatarsus IV, ventral. C, right ALS field, detail (corner, spinneret field diagram). D, right spermathecae, detail. Abbreviations: MaAm, major ampullate; Pi, piriform; retro., retrolateral. Scale bar: B, 0.5 mm; A, 0.2 mm; C, 0.09 mm; D, 0.05 mm. Gondwanan Drymusidae and westward radiation of Izithunzi species in South Africa Most of our analyses show good support for the relationship between the Southern Neotropical, Andean and South African Drymusidae, using both the complete and reduced data sets (Fig. 2; Supporting Information, Fig. S1). This is consistent with a temperate Gondwanan distribution, meaning that these species are found on continental plates once forming Western Gondwana. As the complete separation of Africa and America occurred 95 Mya (Pitman et al., 1993), the potential Mesozoic age of Drymusidae requires testing. Radiation of Izithunzi species within South Africa may have followed a Westward (from East to West) pattern (compare Fig. 2 against Figs. 5, 6), as the more recent proposed sister taxon splits occur westward in Izithunzi distribution. Unfortunately, we lack sequences from I. zondii sp. nov., the easternmost species described for this genus (KwaZulu-Natal Province), which are necessary to fully test this hypothesis. Nevertheless, I. silvicola comb. nov. appears to have diverged early, are found to be sister to the remaining species of the genus and occur at the border between Eastern and Western Cape provinces (Figs 2, 5, 6). The next species to diverge was I. productum comb. nov., which emerges as sister to the western species (Figs 2, 5, 6). This species occurs in the middle of Western Cape Province. Finally, the last two species to diverge were I. capense comb. nov. and I. lina sp. nov. both occurring on the western extreme of Izithunzi distribution. (Figs 2, 5, 6). Figure 5. View largeDownload slide Distribution map of Izithunzi capense comb. nov., Izithunzi lina sp. nov., Izithunzi productum comb. nov. and Izithunzi silvicola comb. nov. in Western Cape and Eastern Cape provinces (South Africa). Figure 5. View largeDownload slide Distribution map of Izithunzi capense comb. nov., Izithunzi lina sp. nov., Izithunzi productum comb. nov. and Izithunzi silvicola comb. nov. in Western Cape and Eastern Cape provinces (South Africa). Figure 6. View largeDownload slide Distribution map of Izithunzi zondii sp. nov. in KwaZulu-Natal province (South Africa). Figure 6. View largeDownload slide Distribution map of Izithunzi zondii sp. nov. in KwaZulu-Natal province (South Africa). CONCLUSIONS Further analyses such as phylogenomic, that is, next-generation sequencing, or total evidence approaches, that is, including morphological and molecular characters, could increase the Synspermiata tree resolution and help to resolve and support the position of Ochyroceratidae within Scytodoidea, the clade Periegopidae plus Drymusidae and the relationships between the Caribbean and South American Drymusidae. In addition, to improve Scytodoidea and Drymusidae resolution, it is very important to increase the sampling of Drymusidae representatives, especially those species of the Neotropical region. Regarding the biogeography of Izithunzi species, there is no doubt that obtaining new molecular data for several specimens of the described species (and potentially new species) will help to better understand the distributional patterns of this genus in South Africa. TAXONOMIC ACCOUNT Drymusidae Simon, 1893 Drymusinae Simon, 1893: 264, 276; Petrunkevitch, 1928: 42, 19; Petrunkevitch, 1939: 158; Gerhardt & Kästner, 1938: 596; Roewer, 1942: 321; Bonnet, 1956: 1610; Lehtinen, 1967: 301. Drymusidae—Lehtinen, 1986: 156 (new rank); Dippenaar-Schoeman & Jocqué, 1997: 152; Jocqué & Dippenaar-Schoeman, 2006: 118; Filmer, 1991: 80; Ramírez et al., 1999; Filmer 2010: 71; Grismado et al., 2014: 80; World Spider Catalog, 2017. Taxonomic history: The subfamily Drymusinae was established by Simon (1893) in the Deuxième édition of his Histoire Naturelle des Araignées and was monogeneric including the type genus Drymusa. Drymusinae were placed within Sicariidae together with Plectreurinae, Scytodinae, Periegopinae, Loxoscelinae and the s.s. Sicariinae (Simon, 1893); this arrangement was followed by Petrunkevitch (1928, 1939), Gerhardt & Kästner (1938), Roewer (1942) and Bonnet (1956). However, some authors treated Drymusa species as members of Sicariidae (Purcell, 1904; Petrunkevitch, 1911; Bryant, 1948) or Scytodidae (Bristowe, 1938; Lehtinen, 1967, who placed them there tentatively). Gertsch (1967) proposed a new concept of Scytodidae containing three genera: Loxosceles, Scytodes and Drymusa; this arrangement was followed by Valerio (1971, 1974) and Brignoli (1976, 1983). Alayón (1981) revalidated Loxoscelidae (= Sicariidae), including Loxosceles and Drymusa, and ignored Drymusinae; this opinion was followed only by himself (Alayón, 1987). Lehtinen (1986) elevated Drymusinae for the first time to family rank: Drymusidae. Unfortunately, neither of the previous hypotheses (Alayón, 1981; Lehtinen, 1986) was based on any supporting morphological characters, that is, autapomorphies for Drymusa, a critical deficiency as indicated by Goloboff & Ramírez (1992). Finally, Platnick (1989) recognized Drymusidae in his catalogue following Lehtinen (1986), and this family rank has been generally accepted (Coddington & Levi, 1991; Filmer, 1991, 2010; Platnick et al., 1991; Goloboff & Ramírez, 1992; Platnick, 1993; Dippenaar-Schoeman & Jocqué, 1997; Ramírez et al., 1999; Alayón, 2000; Ramírez, 2000; Brescovit, Bonaldo & Rheims, 2004; Coddington et al., 2004; Coddington, 2005; Bonaldo, Rheims & Brescovit, 2006; Jocqué & Dippenaar-Schoeman, 2006; Labarque & Ramírez, 2007a, b; Binford et al., 2008; Rheims, Brescovit & Bonaldo, 2008; Labarque & Ramírez, 2012; Lotz, 2012; Agnarsson, Coddington & Kuntner, 2013; Grismado et al., 2014; Michalik & Ramírez, 2014; Wheeler et al., 2016; Griswold & Ramírez, 2017; Labarque et al., 2017; World Spider Catalog, 2017). Monophyly: Most of our preferred analyses (i.e. R-ML and R-BI) highly supported Drymusidae (Fig. 2), even though the family was recovered polyphyletic using the complete dataset (Supporting Information, Fig. S1). The complete-data set analyses showed that the Caribbean Drymusa sequences had very long branches and were very unstable across analyses; hence, we discarded these from the reduced data set (Fig. 2). In addition, our preferred cladistic results showed a consistently and highly supported sister relationship between the South African Izithunzi gen. nov. and the Southern South American Drymusa (Fig. 2). The following homoplastic morphological synapomorphies have been suggested for Drymusidae: (1) regain of a second major ampullate gland spigot on the anterior lateral spinnerets (instead of a nubbin) (Fig. 4C; Labarque & Ramírez, 2012); (2) regain of long testicular deferent ducts (Michalik & Ramírez, 2014); (3) regain of cleistospermia (Michalik & Ramírez, 2014). Herein, we propose another putative synapomorphy: (4) inner spermathecae of vulva connected to the uterus externus through a duct (Fig. 4D; except in D. serrana). Diagnosis: Synspermiata (although with a regain of cleistospermia in Drymusa rengan Labarque & Ramírez, 2007, see Michalik & Ramírez, 2014), Scytodoidea with six eyes in three separated dyads (no anterior median eyes), intercheliceral articulation sclerotized (known as ‘chelicerae fused’, see Ramírez, 2014), chelicera with lamina and promarginal lobe, female cheliceral stridulatory ridges and pedipalpal femoral thorns, male copulatory bulb presenting one non-expandable sclerite, podotarsite with distal dorsal hood and a third area of articulation, females with haplogyne condition (i.e. cul-de-sac type, Austad, 1984), one posterior spiracle near the spinnerets and tracheae with fused third opisthosomal apodemes (Labarque & Ramírez, 2012; Michalik & Ramírez, 2014; Labarque et al., 2017). Drymusids differ from most Scytodoidea in having two pairs of spermathecae on the vulva with the inner spermathecae connected to the uterus externus through a duct (except Stedocys Ono, 1995; Labarque et al., 2009; Labarque & Ramírez, 2012); differ from Periegopidae in having a second major ampullate gland spigot on the anterior lateral spinnerets, a dorsal cuticular roof for the pretarsal depressor tendon (Labarque & Ramírez, 2007a; Labarque & Ramírez, 2012; Labarque et al., 2017), and by hanging beneath loose space webs (see Labarque & Ramírez, 2012); differ from Sicariidae and Scytodidae in having a closed podotarsite with one subdivision (Labarque et al., 2017); differ also from Sicariidae in having three pretarsal claws, bipectinated proclaws on tarsi I–II, a median field of spicules on the posterior median spinnerets (Platnick et al., 1991; Labarque & Ramírez, 2012), and by venom that lacks sphingomyelinase D proteins (i.e. toxin causative agent in ‘loxoscelism’ lesion formation; Binford & Wells, 2003); differ also from Scytodidae in having the cephalic and thoracic areas of the carapace with similar height (i.e. not forming a dome), the cheliceral stridulatory file composed by shallow multiple scales, the venom outlet facing anteriorly, in lacking apical blunt macrosetae on the female pedipalpal tarsus (Labarque & Ramírez, 2012), and lacking the mechanism for spitting a glue-like substance (Kovoor & Zylberberg, 1972; Foelix, 2011) that is typical of scytodids (‘spitting spiders’); differ from Ochyroceratidae in most of the previous characters, except that Drymusidae are like Ochyroceratidae in having three pretarsal claws, bipectinated proclaws on tarsi I–II (Lehtinen, 1986; Pérez-González et al., 2016), a distal hood on podotarsite (Labarque et al., 2017) and a cheliceral lamina. Genera: Drymusa Simon, 1891, Izithunzi gen. nov. Distribution: From Neotropical and Andean regions (Morrone, 2004, 2014) to South Africa. Natural history: Drymusidae species hang beneath loose space webs hidden in wall crevices or below leaf litter. Females may build small, spherical, wrinkled egg sacs, which they carry with their chelicerae, or build huge, irregular egg sacs, which are covered with detritus and attached to the webs (Valerio, 1974; Alayón, 1981; Jocqué & Dippenaar-Schoeman, 2006: 118–119; Labarque & Ramírez, 2007b; Labarque & Ramírez, 2012). Genus Izithunzi gen. nov. urn:lsid:zoobank.org:act:250052B4-E682-4C4C-A0B5-463F34C3FBC1 Type species: Drymusa capensis Simon, 1893. Species included: Izithunzi capense (Simon, 1893) comb. nov., I. lina sp. nov., I. productum (Purcell, 1904) comb. nov., I. silvicola (Purcell, 1904) comb. nov. and I. zondii sp. nov. Etymology: The generic name means shadows (Izithunzi) in Xhosa, a South African Nguni language of Bantu People. It refers to the retiring nature and cryptic environments (i.e. hidden in caves’ crevices or under dense vegetation) where the members of this genus live. The name is neuter in gender. Remarks:Lehtinen (1967) and Platnick et al. (1991) predicted the non-monotypy of Drymusidae. Lehtinen (1967: 301) compared the Neotropical Drymusa nubila Simon, 1892 with the South African D. capensis (= I. capense comb. nov.) and considered that they were not congeneric (although he took no formal action). Monophyly: Putative synapomorphies include (1) cheliceral promargin with ‘rubble teeth’ that are massive, blunt and boulder shaped (Fig. 3D); (2) female with two sclerotized plates on the vulva, anterior and posterior to the uterus externus (Fig. 7A, B); and (3) male with anterior stridulatory ridge plate on opisthosoma (Fig. 7C, D). Figure 7. View largeDownload slide Izithunzi gen. nov. putative synapomorphies. Izithunzi capense comb. nov. (♀, A CASENT 9023625), Izithunzi lina sp. nov. (♀, B NMBA; ♂, C NMBA), Izithunzi silvicola comb. nov. (♂, D CASENT-9048600), preserved specimens. A, B, female internal genitalia (A, lateral; B, dorsal, right detail). C, D, male opisthosoma, stridulatory ridge plate (C, frontal; D, detail). Abbreviations: ext., externus; recep., receptacle; sp., spermathecae. Scale bar: C, 0.5 mm; D, 0.15 mm; A, B, 0.1 mm; F, 0.05 mm. Figure 7. View largeDownload slide Izithunzi gen. nov. putative synapomorphies. Izithunzi capense comb. nov. (♀, A CASENT 9023625), Izithunzi lina sp. nov. (♀, B NMBA; ♂, C NMBA), Izithunzi silvicola comb. nov. (♂, D CASENT-9048600), preserved specimens. A, B, female internal genitalia (A, lateral; B, dorsal, right detail). C, D, male opisthosoma, stridulatory ridge plate (C, frontal; D, detail). Abbreviations: ext., externus; recep., receptacle; sp., spermathecae. Scale bar: C, 0.5 mm; D, 0.15 mm; A, B, 0.1 mm; F, 0.05 mm. Diagnosis: Izithunzi gen. nov. can be distinguished from Drymusa by the boulder-shaped cheliceral promarginal rubble teeth, the female genitalia having two sclerotized plates anterior and posterior to the uterus externus and the male opisthosoma having an anterior stridulatory ridge plate. Description: Female total length between 5.65 and 15.03 and male total length between 4.95 and 12.27. Carapace reddish with distinguished dun pattern forming two central large patches making a backward-pointing V-shaped mark, each anteriorly prolonged in three longitudinal lines surrounding the eyes and laterally extending in three wavy lines; clypeus with cross-linked pattern. Cephalic area slightly more than one-half width of thoracic area. Chelicerae promargin with two massive and blunt teeth, the rubble teeth, usually dark under the microscope, a triangular lamina contiguous with the paturon margin with a row of several macrosetae against it and a fleshy lobe apically blunt with a basal row of several long and filiform setae, the bracket setae (Figs 3D, 8A). Cheliceral retromargin with two rubble teeth, rarely three (see I. silvicola comb. nov.), forming a row perpendicular to the fang furrow, apical tooth large and proximal small (Fig. 4A). Chelicerae ectal margin stridulatory ridge formed by multiple shallow scales (Fig. 3E). Fang venom gland outlet distal, facing anteriorly (Fig. 3D). Cheliceral bases articulated with a small, sclerotized, posterior intercheliceral sclerite (Fig. 8B). Venom gland extending into carapace (Fig. 8C). Endites longer than wide, bending prolaterally and converging in front of the labium, fleshy apical profiles almost touching each other (Fig. 8D). Labium trapezoidal, longer than wide, narrowed close to the acute apical margin and partially separated from sternum by a membranous suture (Figs 8, 9). Maxillary gland pore field clumped (Fig. 8). Labral tongue apically concave, with dorsal setae (Fig. 8). Pedipalpal claw reduced to a nubbin, distal tarsus with four prolateral macrosetae curved distally, and basal femur with prolateral thorn that might be distally acute or blunt (Fig. 3F). Sternum oval, longer than wide, bordered (Fig. 9). Precoxal triangles fused to sternum (Fig. 9). Legs tan dappled with dun. Proclaws I–II clearly bipectinate with longer teeth on the prolateral row; retroclaws I–II with prolateral tooth row (Fig. 3A). Foot articulated, closed podotarsite with one subdivision, distal hood covering the base of the superior claws, with distal-ventral frictional setae (Fig. 3A, B). Tarsal organ exposed, ovoid and with three receptor-cell dendrite terminals (sensilla) (Fig. 9). Metatarsal trichobothria opening distal margin entire; proximal and distal plates of trichobothria smooth, not well differentiated, distal plate contiguous with the surrounding cuticle; sculpture on basal expansion of trichobothrial setae smooth (Fig. 9). Metatarsi III–IV with dense brush of apical ventral setae (Fig. 4B). Pedicel lorum triangular, not transversely divided, narrowed posteriorly to a pointed end, without slit sensilla stripes (Fig. 9). Pedicel sternites and pleurites separated. Opisthosoma colour overall dark brown with or without chevrons. Epigastrium posterior border with two small, lateral, sclerotized grooves, which seemingly provide a guide for the copulatory bulb apex (see at end of paragraph) to enter into the vulva. Epigastrium and postepigastrium on the external genital area may be heavily sclerotized, forming the epigastrium plate and the postepigastrium plate, respectively (see I. silvicola comb. nov.). Postepigastric foveae absent from abdominal venter (present in many Scytodes and some Neotropical Drymusa). Vulva with an anterior pair of spermathecae (inner), each connected to the uterus externus through a duct (Figs 4D, 7, 10A–C), an anterior sclerotized plate with a second pair of spermathecae (outer) and a posterior sclerotized plate with two dorsal receptacula (Figs 7, 10A–C). Inner spermatheca larger than outer, its duct may open far from (separated, Figs 10A, B) or directly next to the latter (clustered, Figs 4D, 10C). Outer spermathecae and dorsal receptaculum may provide opisthosomal muscle insertions (Fig. 10C). Spermathecae and dorsal receptaculum with ductless glands well spaced or in patches of two to many glands (Figs 4D, 7, 10A–C). Sclerotized plates may be separated from each other surrounding the uterus externus (Figs 7A, 10A), pressed together anteriorly squeezing the uterus externus (Figs 7B, 10B) or completely integrated with the uterus externus forming a unit, the integrated plate (Figs 10). Integrated plate may present a conspicuous or diffuse anterior longitudinal middle ridge (Fig. 10C) forming a sort of crest, which extends across the plate. Uterus externus may or may not extend beyond the vulval plate anterior borders (compare I. capense comb. nov. against I. zondii sp. nov.). Tracheal spiracle wide, separated from spinnerets (Fig. 11). Third opisthosomal entapophyses fused, forming a long median trachea (Fig. 3C). Colulus well defined, ovoid and posteriorly narrowed (Fig. 11). Anterior lateral spinnerets (ALS) with three articles (Fig. 11). ALS with two major ampullate gland spigots (MaAm) and a separated field of several piriform (Pi) gland spigots (Fig. 11). Posterior median spinnerets (PMS) tetrahedral with basal straight and plumose setae on anteromedian surface, a single aciniform gland spigot (Ac), and a field of spicules on mesal surface Fig. 11. Posterior lateral spinnerets (PLS) conical, with several aciniform gland spigots (Fig. 11). Male markings as in female. Opisthosoma with an anterior sclerotized plate with stridulatory ridge, dorsally facing posterior prosoma (Fig. 7). Epiandrous spigots arising in several bunches from isolated pits (Fig. 12). Male with spinnerets as in female but differs in having the PLS with one aciniform gland spigot instead of several spigots (Fig. 11). Male pedipalp with prolateral femoral thorns as in females (Fig. 13). Patella-tibia joint dicondylic, ventral condyle heavily sclerotized, projecting prolaterally, arthrodial membrane broad prolaterally, occupying one third of patella (Figs 12, 13); these modifications on the articulation seem to allow a prolateral movement to the tibia. Femora narrow (L ≤ 3.5× W) to elongated (L > 4× W). Tibiae swollen (L < 2× W) to thin (L > 2.5× W) (compare I. capense comb. nov. against I. silvicola comb. nov.). Cymbium as long as wide to swollen (L ≤ 1.5× W), apically blunt, and as long as or 1.5 times longer than the copulatory bulb base (compare I. capense comb. nov. against I. silvicola comb. nov.). Copulatory bulb presenting one non-expandable piriform sclerite (subtegulum, tegulum and embolus fused). Base (subtegulum + tegulum) showing the sperm duct. Apex (embolus) as long as to more than three times longer than the base, laminated, slightly curved and implanted prolaterally (compare I. productum comb. nov. against I. silvicola comb. nov.). Figure 8. View largeDownload slide Izithunzi gen. nov. chelicerae and mouthparts. Izithunzi zondii sp. nov. (♀, A, B, F, NMSA), Izithunzi silvicola comb. nov. (♀, D, CASENT 9021766, ♀, E CASENT 9048599), Izithunzi capense comb. nov. (♂, C CASENT 9026022), preserved specimens. A–C, Chelicerae (A, anterior; B, posterior; C, venom gland; anterior). D, labium and endites (ventral). E, left maxillary gland (detail). F, labrum and endites (dorsal). Abbreviations: scl., sclerite. Scale bar: A, B, D, 0.5 mm; F, 0.25 mm; C, 0.2 mm; E, 0.02 mm. Figure 8. View largeDownload slide Izithunzi gen. nov. chelicerae and mouthparts. Izithunzi zondii sp. nov. (♀, A, B, F, NMSA), Izithunzi silvicola comb. nov. (♀, D, CASENT 9021766, ♀, E CASENT 9048599), Izithunzi capense comb. nov. (♂, C CASENT 9026022), preserved specimens. A–C, Chelicerae (A, anterior; B, posterior; C, venom gland; anterior). D, labium and endites (ventral). E, left maxillary gland (detail). F, labrum and endites (dorsal). Abbreviations: scl., sclerite. Scale bar: A, B, D, 0.5 mm; F, 0.25 mm; C, 0.2 mm; E, 0.02 mm. Figure 9. View largeDownload slide Izithunzi gen. nov., female prosoma, legs and pedicel. Izithunzi silvicola comb. nov. (A CASENT 9048599), Izithunzi capense comb. nov. (♀, B, C CASENT 9023625), Izithunzi productum comb. nov. (♀, D CASENT 9043066), preserved specimens. A, sternum. B, tarsal organ, I left. C, metatarsal trichobothria, I right. D, pedicel, dorsal. Scale bar: A, 1 mm; D, 0.09 mm; C, 0.02 mm; B, 0.009 mm. Figure 9. View largeDownload slide Izithunzi gen. nov., female prosoma, legs and pedicel. Izithunzi silvicola comb. nov. (A CASENT 9048599), Izithunzi capense comb. nov. (♀, B, C CASENT 9023625), Izithunzi productum comb. nov. (♀, D CASENT 9043066), preserved specimens. A, sternum. B, tarsal organ, I left. C, metatarsal trichobothria, I right. D, pedicel, dorsal. Scale bar: A, 1 mm; D, 0.09 mm; C, 0.02 mm; B, 0.009 mm. Figure 10. View largeDownload slide Izithunzi gen. nov., female vulva types, dorsal. Izithunzi capense comb. nov. (♀, A CASENT-9021768), Izithunzi lina sp. nov. (♀, B NMBA), Izithunzi silvicola comb. nov. (♀, C CASENT-9021765). A, separated plates surrounding the uterus externus. B, pressed plates anteriorly squeezing the uterus externus. C, integrated plate forming a unit with the uterus externus. Abbreviations: ant., anterior; ext., externus; int., integrated; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A–C, 0.2 mm. Figure 10. View largeDownload slide Izithunzi gen. nov., female vulva types, dorsal. Izithunzi capense comb. nov. (♀, A CASENT-9021768), Izithunzi lina sp. nov. (♀, B NMBA), Izithunzi silvicola comb. nov. (♀, C CASENT-9021765). A, separated plates surrounding the uterus externus. B, pressed plates anteriorly squeezing the uterus externus. C, integrated plate forming a unit with the uterus externus. Abbreviations: ant., anterior; ext., externus; int., integrated; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A–C, 0.2 mm. Figure 11. View largeDownload slide Izithunzi gen. nov., posterior respiratory system and spinnerets. Izithunzi lina sp. nov. (♀, A NMBA), Izithunzi zondii sp. nov. (♀, B NMSA), Izithunzi productum comb. nov. (♀, C CASENT 9043066), Izithunzi silvicola comb. nov. (♀, D, E CASENT 9048599, ♂, F CASENT 9048600), preserved specimens. A, opisthosoma, tracheae spiracle detail. B, spinnerets field, posterior. C, right ALS field. D, left PMS field, inner surface. E, F, left PMS-PLS field (C–F corners, spinnerets field diagram). Abbreviations: Ac, aciniform; ALS, anterior lateral spinnerets; MaAm, major ampullate; Pi, piriform; PMS, posterior median spinnerets; PLS, posterior lateral spinnerets. Scale bar: A, 0.5 mm; B, 0.2 mm; D, 0.09 mm; E, 0.03 mm; C, F, 0.02 mm. Figure 11. View largeDownload slide Izithunzi gen. nov., posterior respiratory system and spinnerets. Izithunzi lina sp. nov. (♀, A NMBA), Izithunzi zondii sp. nov. (♀, B NMSA), Izithunzi productum comb. nov. (♀, C CASENT 9043066), Izithunzi silvicola comb. nov. (♀, D, E CASENT 9048599, ♂, F CASENT 9048600), preserved specimens. A, opisthosoma, tracheae spiracle detail. B, spinnerets field, posterior. C, right ALS field. D, left PMS field, inner surface. E, F, left PMS-PLS field (C–F corners, spinnerets field diagram). Abbreviations: Ac, aciniform; ALS, anterior lateral spinnerets; MaAm, major ampullate; Pi, piriform; PMS, posterior median spinnerets; PLS, posterior lateral spinnerets. Scale bar: A, 0.5 mm; B, 0.2 mm; D, 0.09 mm; E, 0.03 mm; C, F, 0.02 mm. Figure 12. View largeDownload slide Izithunzi lina sp. nov. (♂, A, B NMBA), preserved specimen. A, epiandrous spigots field (note several bunches arising from isolated pits). B, male right pedipalp, dicondylic patella-tibia joint, prolateral. Scale bar: A, 0.2 mm; B, 0.1 mm. Figure 12. View largeDownload slide Izithunzi lina sp. nov. (♂, A, B NMBA), preserved specimen. A, epiandrous spigots field (note several bunches arising from isolated pits). B, male right pedipalp, dicondylic patella-tibia joint, prolateral. Scale bar: A, 0.2 mm; B, 0.1 mm. Figure 13. View largeDownload slide Izithunzi capense comb. nov., copulatory organs (♀, A, C CASENT-9023625; ♀, B CASENT-9021768; ♂, D–F CASENT-9021768). A, external female genitalia (ventral). B, vulva (dorsal). C, left dorsal receptaculum, detail. D–F, left pedipalp (D, prolateral; E, retrolateral; F, apical). Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, D–F, 0.5 mm; B, 0.2 mm; C, 0.05 mm. Figure 13. View largeDownload slide Izithunzi capense comb. nov., copulatory organs (♀, A, C CASENT-9023625; ♀, B CASENT-9021768; ♂, D–F CASENT-9021768). A, external female genitalia (ventral). B, vulva (dorsal). C, left dorsal receptaculum, detail. D–F, left pedipalp (D, prolateral; E, retrolateral; F, apical). Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, D–F, 0.5 mm; B, 0.2 mm; C, 0.05 mm. Izithunzi capense (Simon, 1893) comb. nov. (Figs 1a–c, 3a–c, 4a, 5, 7A, 8C, 9B, C, 10A, 13A–C, 14, 15) Drymusa capensis Simon, 1893: 278; Purcell, 1904: 154; Filmer, 1991: 80; Filmer, 2010: 71; Labarque & Ramírez, 2012: 3, figs 2B, 3C–D, 6C–D, 9B, 10B, 11B, 12B, 13B, 14B, 15B, 16B, 17B, 18B, 19A, 23A–F, 29C, D, 31B [one immature holotype (MNHN AR-1326) from South Africa, Western Cape, examined.] Figure 14. View largeDownload slide Izithunzi capense comb. nov. (♂, A, B CASENT-9026022), Izithunzi lina sp. nov. (C, D CASENT-9043226), male chelicerae. A, B, right, chelicerae, promargin (flattened teeth detail). C, both chelicera (frontal). D, left chelicerae (medial). Abbreviations: pro., prolateral; retro., retrolateral. Scale bar: C, D, 0.2 mm; A, 0.1 mm; B, 0.03 mm. Figure 14. View largeDownload slide Izithunzi capense comb. nov. (♂, A, B CASENT-9026022), Izithunzi lina sp. nov. (C, D CASENT-9043226), male chelicerae. A, B, right, chelicerae, promargin (flattened teeth detail). C, both chelicera (frontal). D, left chelicerae (medial). Abbreviations: pro., prolateral; retro., retrolateral. Scale bar: C, D, 0.2 mm; A, 0.1 mm; B, 0.03 mm. Figure 15. View largeDownload slide Izithunzi capense comb. nov., preserved specimens (♀, A, B CASENT-9023625; ♂, C, D CASENT-9021768; immature, E, F MNHN AR-1326 holotype). A–D, habitus (A, C, E, dorsal; B, D, F, ventral). Scale bar: A–D, 2 mm; E, F, 0.5 mm. Figure 15. View largeDownload slide Izithunzi capense comb. nov., preserved specimens (♀, A, B CASENT-9023625; ♂, C, D CASENT-9021768; immature, E, F MNHN AR-1326 holotype). A–D, habitus (A, C, E, dorsal; B, D, F, ventral). Scale bar: A–D, 2 mm; E, F, 0.5 mm. Loxosceles valida Lawrence, 1964: 61, figs 3, 4; Newlands, 1975: 146, fig. 7; 1986: 80, figs 55–57; Brignoli, 1976: 147. [one ♂ holotype (SAM-ENW B010012) from South Africa, Western Cape Province, Cape Town, Kalk Bay, Echo Valley, Echo Halt Cave, [−34.116667, 18.433333], Apr.1954, J. Grindley col., not examined.] syn. nov. Drymusa valida – Lotz, 2012: 37, fig. 18A, B. Remarks: Whereas we have not examined the type specimens of other Izithunzi species and L. valida, we believe that species attribution is unproblematic. Purcell (1904: 154) reported that the type of D. capensis is very immature and that the adults of this species are much larger than the other species from South Africa. Here, we reconfirm his assessment: the type specimen is very immature (Fig. 13E, F), and the total length of females (14.30) and males (12.80) that we identify as this species fits Purcell’s suggestion of the large size of this species. Purcell (1904: 154) also provided a key to females of the African species using characters that match those that we use here to diagnose I. capense comb. nov. Regarding L. valida, we think the evidence for synonymy with I. capense comb. nov. is convincing. Lawrence (1964: 61) described a row of several macrosetae against the triangular lamina on the chelicerae promargin of L. valida, a character absent in Loxosceles but present in Drymusidae. Brignoli (1976: 147) suggested that L. valida might be a D. capensis based on their similar geographical distribution. The presence of three pretarsal claws in L. valida was brought to Norman Larsen’s attention by Vince Roth during a visit to the South African Museum in Cape Town (Larsen, 1994). This feature contrasts to the two claws of Loxosceles and other Sicariidae. Finally, Lotz (2012: 37) transferred L. valida to Drymusa (and to Drymusidae) based on the presence of an articulated podotarsite (‘long and clear onicium’) and three pretarsal claws. The opisthosomal coloration pattern and male pedipalpal configuration of L. valida resemble those of I. capense comb. nov. (compare Figs 13, 15 against Lawrence, 1964: figs 3, 4). The examination of a female topotype (i.e. South Africa, Western Cape Province, Cape Town, Kalk Bay, Echo Valley, Echo Halt Cave [−34.116667, 18.433333]), further corroborates our new synonymy. Key to species of Izithunzi gen. nov. 1 Females 2 – Males … 6 2(1) Small, prosoma length less than 4.0 mm … 3 – Large, prosoma length greater than 5.0 mm … 4 3(2) Spermathecae clustered … 5 – Spermathecae separated … I. zondii sp. nov. 4(2) Vulval plates separated; posterior plate nearly straight … I. capense comb. nov. – Vulval plates pressed together anteriorly; posterior plate curved anteriorly … I. lina sp. nov. 5(3) Epigastrium posterior border with dark, small, thick setae; epigastrium not forming an elongated plate … I. productum comb. nov. – Epigastrium posterior border lacking those setae; epigastrium forming a plate, heavily sclerotized and posteriorly elongated to the median point of the opisthosoma … I. silvicola comb. nov. 6(1) Pedipalp short, femur length less than 3.5 times its width; tibia length less than 2 times its width … 7 – Pedipalp elongated, femur length more than four times its width; tibia length more than 2.5 times its width … I. silvicola comb. nov. 7(6) Copulatory bulb base and apex transition smooth; apex distally dark … 8 – Copulatory bulb base and apex transition indented; apex mostly dark … I. lina sp. nov. 8(7) Copulatory bulb apex 1½ times longer than its base, with an acute and slightly curved tip … I. capense comb. nov. – Copulatory bulb apex no longer than its base, with laminar truncated tip … I. productum comb. nov. Material examined: ♀ from South Africa, Western Cape Province, Cape Town, Table Mountain National Park, Newlands Forest Preserve, −33.973999, 18.444133, 25 February 2006, elev. 145 m, J. Miller, H. Wood, N. Larsen cols., preparation codes FML-00435-00444 [♀], deposited in CAS (CASENT9023625); ♂, ♀ and three immatures, same data, preparation codes FML-01010, FML-01108 and FML-01118 [♂], and FML-01107 [♀], CAS (CASENT9026022); 4♂, 3♀ and four immatures, same locality, 4 October 2001, N. Larsen, K. Muller, S. Prinsloo, D. Ubick, S. Ubick cols., CAS (CASENT9048605); ♀ and two immatures, same data, 4 November 1925, Lang col., AMNH; ♀ and one immature, same locality, 18 December 1996, elev. 150 m, indigenous forest at night, P. Sierwald col., FMNH; three immatures, same data, FMNH; one immature, same data, elev. 120 m, during the day, FMNH; two immatures, same data, FMNH; two immatures, same data, FMNH; three immatures, same data, FMNH; seven immatures, same data, C. Griswold col., CAS (CASENT9053375); one immature, same locality, 15 January 2009, pine plantation decayed log, site 2, C. Uys col., NCP (2010/1949); two immatures, same data, 23 May 2008, afrotemperate forest decayed log, NCP (2010/1922); ♂, same locality, March 1993, S. Muller col., NCP; ♀, same data, NCP; ♀, same locality, −33.977317, 18.439783, 4 October 2011, elev. 195 m, L. Almeida, C. Griswold, T. Meikle, N. Larsen cols., CAS (CASENT9042516); 2♀ and two immatures, same data, CAS (CASENT9043287); ♀, same data, CAS (CASENT9043286); ♀, same data, CAS (CASENT9043173); ♀ and one immature, same data, CAS (CASENT9043171); 5♀ and nine immatures, same locality, Fernwood Gully, −33.966667, 18.450000, 18 December 1996, elev. 120–150 m, indigenous forest, C. E. Griswold col., preparation code FML-01151 [♀], CAS (CASENT9048602); ♀, same data, CAS (CASENT9048603); 2♂ and 2♀, same locality, Kirstenbosch Botanic Garden, Skeleton Gorge Forest, −33.983333, 18.433333, 7 January 1985, elev. 700 ft, webs beneath logs, C. Griswold, T. Meikle Griswold cols., preparation codes APG-00055 [♀], FML-00546-00547 [♂] and FML-00705 [♀], CAS (CASENT9021768); 2♀ and three immatures, same data, NMSA; five immatures, same data, AMNH; 3♂ and one immature, same data, moulted in captivity, NMSA; ♂ and 2♀, same data, C. Griswold col., NMSA; ♀, same locality, October 1985, 800 ft, C. Griswold col., CAS (CASENT9021767); 3♀ and three immatures, same locality, 26–29 October 1985, elev. 700–1000 ft, C. Griswold, J. Doyen, T. Meikle Griswold cols. NMSA; 4♀ and nine immatures, same data, NMSA; four immatures, same locality, 5 February 2009, afrotemperate forest decayed log, site 5, C. Uys col., NCP (2010/1950); one immature, same locality, Nursery Ravine, Wynberg Caves, [−33.986430, 18.403772], 13 February 1991, cave entrance, V. D. Roth, B. Roth cols., CAS (CASENT9048604); one immature, same locality, Orange Kloof, [−33.998611, 18.392778], 28 January 2009, afrotemperate forest, sugar-baited ant trap, C. Uys col., NCP (2010/1952); ♀, same city, Kalk Bay, Echo Valley, Echo Halt Cave [−34.116667, 18.433333], December 1987, A. le. Roy col., NCP. Further material examined byLotz (2012): ♀ from South Africa, Western Cape Province, Cape Town, Kalk Bay, Echo Valley, Devil’s Pit, [−34.117942, 18.436667], June 1954, J. Grindley col., SAM (B010015); one immature, same locality, Tartarus Cave, [−34.113647, 18.441556], July 1961, J. Grindley col., SAM (B10013); ♀, same city, Table Mountain National Park, Nursery Ravine, Wynberg Caves, Powder Room, [−33.986430, 18.403772], March 1956, South African Speleological Association, SAM (B10018); ♀, same data, February 1956, SAM (B10016); one immature, same data, March 1931, R. Lawrence, SAM (B7892); one immature, same locality, Giants Workshop, [−33.989180, 18.407553], July 1956, J. Grindley col., SAM (B10017); ♀, same locality, Bats Cave, [−33.988569, 18.406797], September 1960, J. Grindley col., SAM (B10014). Diagnosis: Females of I. capense comb. nov. resemble those of I. lina sp. nov. by the epigastrium protruded ventro-anteriorly, the inner and outer spermathecae separated and the uterus externus exceeding beyond the vulval plate anterior borders (Figs 7, 10, 13, 16), but it can be distinguished by the epigastrium covered with long and thin setae, the postepigastrium with an anterior lip covering the epigastric furrow and the vulval plates separated (Figs 7, 10, 13), whereas I. lina sp. nov. presents long, thick and dark setae on the epigastrium, lacks the anterior lip on the postepigastrium and has pressed vulval plates (Figs 7, 10, 16). Males of I. capense comb. nov. resemble those of I. lina sp. nov. by having the cheliceral fang promarginally and distally excavated, and the pro- and retromarginal cheliceral teeth close to the base of the fang extremely modified, enlarged and flattened, which fit in the fang’s excavation (Fig. 14); but it can be distinguished by presenting a smooth transition between the base and apex of the copulatory bulb, and the apex 1½ times longer than the base with an acute and slightly curved tip (Fig. 13), whereas I. lina sp. nov. has an indented transition, and the apex two times longer than the base, heavily sclerotized (dark) and a broad tip (Fig. 16). Figure 16. View largeDownload slide Izithunzi lina sp. nov., copulatory organs (♀, A–C NMBA; ♂, D–F NMBA). A, external female genitalia (ventral). B, C, vulva (B, dorsal; C, lateral). D–F, right pedipalp [D, prolateral; E, retrolateral; F, apical (corner, tip detail, retrolateral)]. Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, F, 2.5 mm; D, E, 0.5 mm; B, 0.2 mm; C, 0.1 mm. Figure 16. View largeDownload slide Izithunzi lina sp. nov., copulatory organs (♀, A–C NMBA; ♂, D–F NMBA). A, external female genitalia (ventral). B, C, vulva (B, dorsal; C, lateral). D–F, right pedipalp [D, prolateral; E, retrolateral; F, apical (corner, tip detail, retrolateral)]. Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, F, 2.5 mm; D, E, 0.5 mm; B, 0.2 mm; C, 0.1 mm. Redescription female (Table Mountain National Park, Newlands Forest Preserve: Images CASENT 9023625; Measurements CASENT 9048605): Total length 15.03. Prosoma: length 5.7, width 3.96, height 2.77. Sternum: length 2.93, width 2.2. Leg measurements: femur: I: 15.15, II: 13.53, III: 11.02, IV: 13.78; patella: I: 1.71, II: 1.68, III: 1.59, IV: 1.72; tibia: I 14.78, II: 12.4, III: 9.2, IV: 12.27; metatarsus: I: 15.15, II: 13.4, III: 10.5, IV: 13.4; tarsus: I: 2.14, II: 2.02, III: 2.09, IV: 2.89; podotarsite: I: 0.22, II: 0.24, III: 0.28, IV: 0.22. Total: I: 49.15, II: 43.27, III: 34.68, IV: 44.27. Leg formula: 1423. Opisthosoma: length 8.72, width 4, height 4.65. Thoracic area lateral margins and central V-shaped pattern darkish, forming a continuum (Fig. 15). Chelicerae promargin with five bracket setae, and a row of seven to eight macrosetae against the triangular lamina. Labium dun, reddish at narrow area and white at apex (Figs 8, 15). Pedipalpal prolateral femoral thorn distally acute. Sternum dun (Figs 8, 15). Femora and tibiae dun, patellae, metatarsi and tarsi tan (Fig. 15). Opisthosoma colour overall dark brown forming thick chevrons extending anteriorly (Fig. 15). Chevrons well-spaced anteriorly, clustered posteriorly, the first two forming a continuum (Fig. 15). Anterior vulval plate slightly sclerotized, and posterior plate nearly rectangular and slightly curved anteriorly (Figs 7, 10, 13). Both spermathecae oval (Figs 7, 10, 13). Redescription male (Table Mountain National Park, Newlands Forest Preserve: Habitus CASENT 9021768; SEM CASENT 9026022; Measurements CASENT 9048605): Total length 12.27. Prosoma: length 5.6, width 4.12, height 2.8. Sternum: length 2.67, width 2.02. Leg measurements: femur: I: 17.1, II: 16.5, III: 13.02, IV: 14.9; patella: I: 1.74, II: 1.78, III: 1.68, IV: 1.7; tibia: I 16.5, II: 14.65, III: 10.4, IV: 12.77; metatarsus: I: 18.0, II: 16.1, III: 12.27, IV: 14.9; tarsus: I: 2.37, II: 2.37, III: 2.35, IV: 3.14; podotarsite: I: 0.27, II: 0.21, III: 0.32, IV: 0.33; total: I: 55.98, II: 51.61, III: 40.04, IV: 47.74. Leg formula: 1243. Opisthosoma: length 6.3, width 3.76, height 3.76. Male pedipalp: femur: 2.06, patella: 0.75, tibia: 1.5, tarsus: 0.59. Coloration as female (Fig. 15). Chelicerae promargin also with five bracket setae (Fig. 14). Epiandrous spigots arising in seven bunches from isolated pits. Pedipalpal prolateral femoral thorn also distally acute but longer than in females (Fig. 13); femora narrow (L ≤ 3.5× W); tibiae swollen, longer than width (L < 2× W) (Fig. 13). Copulatory bulb apex elongated and acute distally. In addition, the copulatory bulb apex looks slightly curved in lateral view (both pro- and prolateral) and straight in apical view (i.e. apical view of the cymbium) (Fig. 13). Distribution: Western Cape Province, South Africa, from Table Mountain National Park to Kalk Bay and surroundings (Fig. 5; Supporting Information, Fig. S1). Natural history: According to Larsen (1994), I. capense comb. nov. specimens are found under exfoliated bark or in crevices between boulders, always in cool shaded areas and hanging beneath loose space webs (Fig. 1A–C). Izithunzi capense comb. nov. individuals are sensitive to the light, and they quickly retreat to the darkness after a minute of exposure with an unfiltered electric torch (flashlight) (Larsen, 1994). Our own observations match these. Izithunzi lina sp. nov. urn:lsid:zoobank.org:act:269AB044-0E94-4F7A- 9A18-CEEA44D74686 (Figs 5, 7, 10–12, 14, 16, 17) Type material: Holotype: ♀ from South Africa, Western Cape Province, Overberg DC, Fernkloof Nature Reserve, 3.98 km 90° E Hermanus, −34.412683, 19.288350, 13 October 2011, elev. 14 m, general collecting at wet, mossy cliff and caves, L. Almeida, C. Griswold cols., preparation codes FML-01103 and FML-01351-01352 [♀], deposited in NMBA. Paratypes: ♂, same data as the type, preparation codes FML-01152 and FML-01353 [♂], NMBA; 2♀, same data, CAS (CASENT9043225); ♂ and three immatures, same locality, 2.81 km 43° NE Hermanus, −34.394617, 19.266117, 12 October 2011, elev. 97 m, general collecting at night in riparian vegetation, L. Almeida, C. Griswold, T. Meikle, V. Hamilton-Attwell cols., CAS (CASENT9043181). Figure 17. View largeDownload slide Izithunzi lina sp. nov., preserved specimens (♀, A–C NMBA; ♂, D–F NMBA). A–F, habitus (A, D, dorsal; B, E, lateral; C, F, ventral). Scale bar: A–F, 2 mm. Figure 17. View largeDownload slide Izithunzi lina sp. nov., preserved specimens (♀, A–C NMBA; ♂, D–F NMBA). A–F, habitus (A, D, dorsal; B, E, lateral; C, F, ventral). Scale bar: A–F, 2 mm. Further material examined: One immature, same city, Kogelberg Nature Reserve, [−34.292400, 18.919390], 5 July 2007, UGB Forest, ‘banc hand’, technician students cols., NCP (2008/4425). Etymology: The specific name is a name in apposition to honour our friend and fellow arachnologist Lina Maria Almeida-Silva, who collected part of the type series. Diagnosis: Females of I. lina sp. nov. resemble those of I. capense comb. nov. by the epigastrium protruded, the inner and outer spermathecae separated and the uterus externus extending beyond the vulval plates anterior borders (Figs 7, 10, 13, 16), but it can be distinguished by the epigastrium densely covered with long, thick and dark setae, and the vulval plates pressed together anteriorly squeezing the uterus externus (Figs 7, 10, 16), whereas I. capense comb. nov. has thin setae on the epigastrium and separated vulval plates (Figs 7, 10, 13). Males of I. lina sp. nov. resemble those of I. capense comb. nov. by having the cheliceral fang promarginally and distally excavated, and the pro- and retromarginal cheliceral teeth close to the base of the fang extremely modified, enlarged and flattened, which fit in the fang’s excavation (Fig. 14), but it can be distinguished by having an indented transition between the base and apex of the copulatory bulb, and the apex two times longer than the base, heavily sclerotized, and with a broad tip (Fig. 16), whereas I. capense comb. nov. presents a smooth transition between the base and apex, and the apex 1½ times longer than the base with an acute and slightly curved tip (Fig. 13). Description female (Fernkloof Nature Reserve: Habitus and measurements NMBA): Total length 12.4. Prosoma: length 5.2, width 3.36, height 3.08. Sternum: length 2.6, width 1.79. Leg measurements: femur: I: 13.65, II: 11.7, III: 9.4, IV: 12.02; patella: I: 1.4, II: 1.48, III: 1.48, IV: 1.41; tibia: I: 12.65, II: 10.3, III: 7.76, IV: 10.3; metatarsus: I: 13.15, II: 10.9, III: 8.64, IV: 10.9; tarsus: I: 1.88, II: 1.76, III: 1.76, IV: 2.37; podotarsite: I: 0.21, II: 0.22, III: 0.19, IV: 0.25; total: I: 42.94, II: 36.36, III: 29.23, IV: 37.25. Leg formula: 1423. Opisthosoma: Length 7.6, width 3.46, height 3.02. Thoracic area lateral margins and central V-shaped pattern darkish, forming a continuum (Fig. 17). Chelicerae promargin with five bracket setae and a row of seven macrosetae against the triangular lamina. Labium dun, reddish at narrow area and white at apex (Fig. 17). Sternum dun (Fig. 17). Femora and tibiae dun, patellae, metatarsi and tarsi tan (Fig. 17). Opisthosoma colour overall dark brown forming thick chevrons extending anteriorly (Fig. 17). Chevrons well-spaced anteriorly, clustered posteriorly, the first two forming a continuum (Fig. 17). Anterior plate heavily sclerotized, and both vulval plates curved anteriorly (Fig. 16). Both spermathecae oval (Fig. 16). Description male (Fernkloof Nature Reserve: Habitus and measurements NMBA): Total length 8.56. Prosoma: length 4.2, width 3.14, height 2.27. Sternum: length 2.22, width 1.59. Leg measurements: femur: I: 12.52, II: 11.2, III: 8.96, IV: 11.0; patella: I: 1.23, II: 1.16, III: 1.16, IV: 1.18; tibia: I: 11.2, II: 9.44, III: 7.52, IV: 9.5; metatarsus: I: 12.0, II: 9.1, III: 8.32, IV: 10.5; tarsus: I: 1.84, II: 1.73, III: lost, IV: 2.22; podotarsite: I: 0.16, II: 0.23, III: lost, IV: 0.16; total: I: 38.9, II: 32.86, III: 25.96 (without tarsus-podotarsite), IV: 34.56. Leg formula: 1423. Opisthosoma: length 4.2, width 2.32, height 2.06. Male pedipalp: femur: 1.1, patella: 0.37, tibia: 0.89, tarsus: 0.44. Coloration as in female (Fig. 17). Chelicerae promargin with four bracket setae, and a row of six to seven macrosetae against the triangular lamina (Fig. 15). Epiandrous spigots arising in five bunches from isolated pits (Fig. 17). Pedipalpal prolateral femoral thorn distally acute (Fig. 16); femora narrow (L ≤ 3.5× W); tibiae swollen, longer than width (L < 2× W) (Fig. 16). Copulatory bulb apex elongated and broad distally (Fig. 16). In addition, the copulatory bulb apex appears straight in lateral view (both pro- and prolateral) and slightly curved in apical view (i.e. of the cymbium) (Fig. 16). Distribution: Western Cape Province, South Africa, Fernkloof Nature Reserve, from Fernkloof east to Hermanus (Fig. 5; Supporting Information, Fig. S1). Izithunzi productum (Purcell, 1904) comb. nov. (Figs 5, 9, 11, 18–20) Drymusa producta Purcell, 1904: 153, pl. 11, fig. 26 [♀] [3♀ syntypes (SAMC 7905) from South Africa, Western Cape Province, Swellendam, mountainside forest, [−34.025708, 20.438125], August 1900, W. Purcell col., not examined]. Figure 18. View largeDownload slide Izithunzi productum comb. nov., copulatory organs (♀, A–C CASENT-9043066; ♂, D–F CASENT-9043066). A, external female genitalia (ventral). B, C, vulva (B, dorsal; C, ventral). D–F, left pedipalp (D, prolateral; E, retrolateral; F, apical). Abbreviations: ant., anterior; ext., externus; int., integrated; recep., receptacle; sp., spermathecae. Scale bar: A, 0.25 mm; D–F, 0.2 mm; C, 0.1 mm; B, 0.09 mm. Figure 18. View largeDownload slide Izithunzi productum comb. nov., copulatory organs (♀, A–C CASENT-9043066; ♂, D–F CASENT-9043066). A, external female genitalia (ventral). B, C, vulva (B, dorsal; C, ventral). D–F, left pedipalp (D, prolateral; E, retrolateral; F, apical). Abbreviations: ant., anterior; ext., externus; int., integrated; recep., receptacle; sp., spermathecae. Scale bar: A, 0.25 mm; D–F, 0.2 mm; C, 0.1 mm; B, 0.09 mm. Remarks: Whereas we have not examined the syntypes of this species, the key, description and illustrations provided by Purcell (1904: 152–154) leave no doubt as to the species identity. Unfortunately, no topotypes were available either. Material examined: ♀ and ♂, from South Africa, Western Cape Province, South Cape DC, Grootvadersbosch Nature Reserve, 14.97 km 316° NE Heidelberg, −33.983733, 20.831967, 17 October 2011, elev. 338 m, indigenous forest, general collecting, L. Almeida, C. Griswold, T. Meikle cols., preparation codes FML-01090, FML-01101-01102 and FML-01109-01115 [♀] and FML-01091-01092 and FML-01116 [♂], deposited in CAS (CASENT9043066); ♂ and ♀, same data, CAS (CASENT9043050); 2♂, 2♀ and one immature, same data, CAS (CASENT9043067); ♀ and ♂, same data, CAS (CASENT9053373); ♂, same locality, Bushbuck Trail (Fonteintjiesbos), −33.985000, 20.808333, 26 January–February 2004, elev. 370 m, young afromontane forest, flight intercept trap, N. Solodovnikov et al. cols., FMNH (2004/011); 2♀, same locality, 20 km WNW Heidelberg, −33.999999, 20.783333, 8–10 November 1985, elev. 1600 ft, indigenous forest, C. Griswold, J. Doyen, T. Meikle Griswold cols., NMSA; 2♂, 4♀ and three immatures, same data, NMSA. Diagnosis: Females of I. productum comb. nov. resemble those of I. silvicola comb. nov. by the inner and outer spermathecae clustered, and the integrated plate (Figs 4D, 18, 19, 21), but can be distinguished by the epigastrium posterior border covered with small, dark, thick setae, the lateral sclerotized grooves almost touching each other, the dorsal receptaculum lacking muscle insertions and completely integrated to the (posterior) plate and the anterior ridge conspicuous (Figs 18, 19), whereas I. silvicola comb. nov. lacks the thick setae on the epigastrium, has separated lateral grooves, partially integrated dorsal receptaculum that present muscle insertions and diffused anterior ridge (Figs 4D, 10, 21). Males of I. productum comb. nov. can be distinguished from other Izithunzi by the apex of the copulatory bulb as long as the base (1:1) with a laminar truncated tip (Fig. 18). Figure 19. View largeDownload slide Izithunzi productum comb. nov., female vulva (♀, A–C CASENT-9043066). A, dorsal. B, lateral, right. Abbreviations: ant., anterior; ext., externus; int., integrated; recep., receptacle; sclero., sclerotization; sp., spermathecae. Scale bar: B, 0.05 mm. Figure 19. View largeDownload slide Izithunzi productum comb. nov., female vulva (♀, A–C CASENT-9043066). A, dorsal. B, lateral, right. Abbreviations: ant., anterior; ext., externus; int., integrated; recep., receptacle; sclero., sclerotization; sp., spermathecae. Scale bar: B, 0.05 mm. Redescription female (Grootvadersbosch Nature Reserve: Habitus, SEM CASENT 9043066; Measurements CASENT 9053373): Total length 5.65. Prosoma: length 2.75, width 1.82, height 1.81. Sternum: length 1.42, width 1.03. Leg measurements: femur: I: 4.85, II: 4.08, III: 3.39, IV: 4.68; patella: I: 0.72, II: 0.8, III: 0.68, IV: 0.73; tibia: I: 5.0, II: 3.88, III: 2.9, IV: 4.36; metatarsus: I: 4.28, II: 3.56, III: 3.0, IV: 4.12; tarsus: I: 1.14, II: 1.05, III: 1.0, IV: 1.26; podotarsite: I: 0.09, II: 0.09, III: 0.11, IV: 0.08; total: I: 16.08, II: 13.46, III: 11.08, IV: 15.23. Leg formula: 1243. Opisthosoma: length 4.0, width 2.7, height 2.57. Chelicerae promargin with four bracket setae, and a row of six macrosetae against the triangular lamina. Labium dun, reddish at narrow area and white at apex (Fig. 20). Pedipalpal prolateral femoral thorn thick, distally blunt. Sternum dun (Fig. 20). Femora and tibiae dun, patellae, metatarsi and tarsi tan (Fig. 20). Opisthosoma colour overall dark brown, smooth, without chevrons (Fig. 20). Inner spermathecae oval, basally sclerotized (Figs 18, 19). Outer spermathecae tetrahedral (Figs 18, 19). Inner spermathecae duct opening to the integrated plate anteriorly, next to the outer spermathecae (Figs 18, 19). Figure 20. View largeDownload slide Izithunzi productum comb. nov., preserved specimens (♀, A–C CASENT-9043066; ♂, D–F CASENT-9043066). A–F, habitus (A–D, dorsal; B–E, lateral; C–F, ventral). Scale bar: A–F, 1 mm. Figure 20. View largeDownload slide Izithunzi productum comb. nov., preserved specimens (♀, A–C CASENT-9043066; ♂, D–F CASENT-9043066). A–F, habitus (A–D, dorsal; B–E, lateral; C–F, ventral). Scale bar: A–F, 1 mm. Description male (Grootvadersbosch Nature Reserve: Habitus, SEM CASENT 9043066; Measurements CASENT 9053373): (Note: Here, we provide the first description of the male of this species.) Total length 4.95. Prosoma: length 2.26, width 1.7, height 1.2. Sternum: length 1.14, width 0.9. Leg measurements: femur: I: 5.05, II: 4.32, III: 3.44, IV: 4.65; patella: I: 0.6, II: 0.62, III: 0.59, IV: 0.62; tibia: I 5.3, II: 4.12, III: 3.11, IV: 4.45; metatarsus: I: 4.75, II: 3.8, III: 3.11, IV: 4.3; tarsus: I: 1.2, II: 1.15, III: 1.04, IV: 1.32; podotarsite: I: 0.07, II: 0.09, III: 0.1, IV: 0.11; total: I: 16.97, II: 14.1, III: 11.39, IV: 15.45. Leg formula: 1243. Opisthosoma: length 2.52, width 1.46, height 1.41. Male pedipalp: femur: 0.74, patella: 0.22, tibia: 0.57, tarsus: 0.27. Coloration lighter than in female, V-shaped pattern reduced (Fig. 20). Chelicerae promargin also with four bracket setae. Epiandrous spigots arising in four bunches from isolated pits. Pedipalpal prolateral femoral thorn also thick and distally blunt as in females (Fig. 18); femora narrow (L ≤ 3.5× W); tibiae swollen, longer than width (L < 2× W) (Fig. 18). Copulatory bulb apex short and truncated distally (Fig. 18). In addition, the copulatory bulb apex appears slightly curved in lateral (both pro- and prolateral) and apical views (Fig. 18). Distribution: Western Cape Province, South Africa, Langeberg Mountains from Swellendam to Grootvadersbosch and surroundings (Fig. 5; Supporting Information, Fig. S1). Izithunzi silvicola (Purcell, 1904) comb. nov. (Figs 3D–F, 4D, 5, 7, 8–11, 21, 22) Drymusa silvicola Purcell, 1904: 152, pl. 11, figs 24, 25 [♀] [♂, 2♀ and three immature syntypes (SAMC 871) from South Africa, Western Cape Province, Knysna, [−34.035086, 23.046469], March 1896, W. Purcell col., not examined]. Figure 21. View largeDownload slide Izithunzi silvicola comb. nov., copulatory organs (♀, A CASENT-9021766; ♀, B–C CASENT-9021765; ♂, D–F CASENT-9021766). A, external female genitalia (ventral). B, C, vulva (B, dorsal; C, ventral). D, E, left pedipalp (D, prolateral; E, retrolateral). F, bulb, detail (F, ventral). Abbreviations: ant., anterior; epi., epigastrium; int., integrated; pepi., postepigastrium; recep., receptacle; sp., spermathecae. Scale bar: A, D, E, 0.5 mm; B, C, 0.2 mm. Figure 21. View largeDownload slide Izithunzi silvicola comb. nov., copulatory organs (♀, A CASENT-9021766; ♀, B–C CASENT-9021765; ♂, D–F CASENT-9021766). A, external female genitalia (ventral). B, C, vulva (B, dorsal; C, ventral). D, E, left pedipalp (D, prolateral; E, retrolateral). F, bulb, detail (F, ventral). Abbreviations: ant., anterior; epi., epigastrium; int., integrated; pepi., postepigastrium; recep., receptacle; sp., spermathecae. Scale bar: A, D, E, 0.5 mm; B, C, 0.2 mm. Remarks: Whereas we have not examined the syntypes of this species, the key, description and illustrations provided by Purcell (1904: 152–154) and the examination of topotypes leave no doubt as to the species identity. Material examined: 4♀ and eight immatures from South Africa, Western Cape Province, Knysna, [−34.033333, 23.033333], October 1946, Malkin col., preparation codes APG-00002 [♀] and APG-00063 [♀], CAS (CASENT9021765); two immatures, same locality, Industrial Area, [−34.044894, 23.077147]), 3 February 1991, collected in big tree, V. Roth col., CAS (CASENT9048598); 2♂ and 2♀, same province, Harkerville State Forest, 19 km E Knysna, −34.05, 23.233333, 11–13 December 1996, elev. 240 m, C. Griswold col., indigenous forest, preparation codes FML-00548-00549 [♂] and FML-00558 [♀], deposited in CAS (CASENT9021766); 2♂, 3♀ and 11 immatures, same data, CAS (CASENT9048601); ♂, same data, preparation codes FML-01008-01009 and FML-01117, CAS (CASENT9048600); ♀, same data, CAS (CASENT9053369); ♀, same data, CAS (CASENT9053371); ♀, same data, CAS (CASENT9053372); ♂, same data, CAS (CASENT9053367); ♂, same data, CAS (CASENT9053370); one immature, same data, CAS (CASENT9053368); one immature, same data, CAS (CASENT9053374); ten immature, same data, CAS (CASENT9048601); 2♀ and two immatures, same data, CAS (CASENT9021766); ♀ and two immatures, same data, forest, hand collecting, night, P. Sierwald col., FMNH; 5♀, same province, Kranshoek, 20 km E Knysna, −34.0833, 23.2333, 13 December 1996, elev. 180 m, indigenous forest, C. Griswold col., CAS (CASENT9021764); ♀, same data, forest, hand collecting, night, P. Sierwald col., FMNH; ♀, same province, Diepwalle Forest Station, 22 km NE Knysna, −33.948599, 23.157369, 10–13 January 1985, elev. 1800 ft., indigenous forest in holes in tree trucks, C. Griswold, T. Meikle Griswold cols., preparation codes FML-01002-01007 and FML-01011-01012, CAS (CASENT9048599); ♂, 5♀ and three immatures, same data, 11–13 November 1985, indigenous forest, C. Griswold, J. Doyen, T. Meikle Griswold cols., NMSA; ♀, same data, NMSA; 4♀ and two immatures, same data, 10–13 November 1985, in holes tree trunks in indigenous forest, C. Griswold, T. Meikle Griswold cols., NMSA; ♀ and two immatures, same data, 11–13 December 1996, elev. 540 m, C. Griswold col., CAS (CASENT9021761); three immatures, Eastern Cape Province, Tsitsikamma National Park, 78 km E Knysna, −34.023483, 23.890333, 17–18 February 2006, elev. 15 m, coastal forest, J. Miller, H. Wood cols., CAS (CASENT9023643). Diagnosis: Females of I. silvicola comb. nov. resemble those of I. productum comb. nov. by the inner and outer spermathecae clustered, and the integrated plate (Figs 10, 18, 19, 21), but can be distinguished by the epigastrium and postepigastrium plates, the dorsal receptaculum with muscle insertions and partially integrated to the (posterior) plate and the anterior ridge diffuse (Figs 4D, 21), while I. productum comb. nov. lacks both epigastrium and postepigastrium plates, has an integrated dorsal receptaculum without muscle insertions and conspicuous anterior ridge (Figs 18, 19). Males of I. silvicola comb. nov. can be distinguished from other Izithunzi by the pedipalp elongated, the ventral condyle of the patella-tibiae joint not projecting prolaterally, cymbium swollen (L ≤ 1.5× W) and 1.5 times longer than the base of the copulatory bulb, and the apex more than three times longer than the base (Fig. 21). Redescription female (Harkerville State Forest: Habitus CASENT 9021766; SEM CASENT 9043066; Measurements CASENT 9053369): Total length 8.8. Prosoma: length 3.49, width 2.67, height 2.18. Sternum: length 1.76, width 1.45. Leg measurements: femur: I: 9.8, II: 8.0, III: 6.74, IV: 8.72; patella: I: 0.82, II: 0.93, III: 0.92, IV: 0.97; tibia: I 9.44, II: 7.12, III: 5.85, IV: 8.4; metatarsus: I: 9.36, II: 7.6, III: 5.61, IV: 8.24; tarsus: I: 1.54, II: 1.48, III: 1.46, IV: 1.92; podotarsite: I: 0.12, II: 0.15, III: 0.14, IV: 0.14; total: I: 31.08, II: 25.28, III: 20.72, IV: 28.39. Leg formula: 1423. Opisthosoma: length 5.25, width 2.42, height 2.95. Chelicerae promargin with six bracket setae (Fig. 3D), and a row of seven macrosetae against the triangular lamina. Cheliceral retromargin with an extra small tooth, next to the apical tooth on the cheliceral furrow. Labium dun, reddish at narrow area and white at apex (Figs 8, 22). Pedipalpal prolateral femoral thorn thick, distally blunt (Fig. 3F). Sternum dun (Fig. 22). Femora and tibiae dun, patellae, metatarsi and tarsi tan (Fig. 22). Opisthosoma colour overall dark brown, smooth, without chevrons (Fig. 22). Epigastrium plate extending posteriorly beyond middle of opisthosoma, covered with thin setae (Fig. 21). Postepigastrium plate triangular, swollen, just below the epigastric furrow (Fig. 21). Lateral sclerotized grooves conspicuous, well defined, associated with the postepigastrium plate (Fig. 21). Inner spermathecae oval. Outer spermathecae tetrahedral (Fig. 21). Inner spermathecae duct opening to the integrated plate anteriorly, next to the outer spermathecae (Fig. 21). Figure 22. View largeDownload slide Izithunzi silvicola comb. nov., preserved specimens (♀, A–C CASENT-9021766; ♂, D–F CASENT-9021766). A–F, habitus (A–D, dorsal; B–E, lateral; C–F ventral). Scale bar: A–F, 1 mm (B, E, images mirrored). Figure 22. View largeDownload slide Izithunzi silvicola comb. nov., preserved specimens (♀, A–C CASENT-9021766; ♂, D–F CASENT-9021766). A–F, habitus (A–D, dorsal; B–E, lateral; C–F ventral). Scale bar: A–F, 1 mm (B, E, images mirrored). Redescription male (Harkerville State Forest: Habitus CASENT 9021766; SEM and leg measurements CASENT 9048600; Body measurements CASENT 9053367): Total length 5.93. Prosoma: length 2.73, width 1.98, height 1.39. Sternum: length 1.48, width 1.2. Leg measurements: femur: I: 11.7, II: 9.7, III: 7.6, IV: 9.8; patella: I: 0.86, II: 0.89, III: 0.83, IV: 0.92; tibia: I 11.5, II: 9.12, III: 6.92, IV: 9.3; metatarsus: I: 12.65, II: 9.9, III: 7.6, IV: 9.9; tarsus: I: 1.67, II: 1.6, III: 1.59, IV: 2.2; podotarsite: I: 0.14, II: 0.12, III: 0.19, IV: 0.16; total: I: 38.52, II: 31.33, III: 24.73, IV: 32.28. Leg formula: 1423. Opisthosoma: length 3.3, width 1.43, height 1.31. Male pedipalp: femur: 2.02, patella: 0.69, tibia: 1.04, tarsus: 0.47. Coloration lighter than in female, thoracic area pattern reduced, V-shaped mark narrower (Fig. 21D–F). Epiandrous spigots arising in three bunches from isolated pits. Pedipalpal prolateral femoral thorn distally acute and longer than in females; femora elongated (L > 4× W); tibiae thin (L > 2.5× W) (Fig. 21). Copulatory bulb apex elongated (Fig. 21). In addition, the copulatory bulb apex appears slightly curved in lateral (both pro- and prolateral) and apical views (Fig. 21). Distribution: Western and Eastern Cape Provinces, South Africa, in Knysna-Amatola montane forests (Fig. 5; Supporting Information, Fig. S1). Izithunzi zondii sp. nov. urn:lsid:zoobank.org:act:423DC628-2388-4C15-B69E-B020B1940BC6 (Figs 6, 8, 11, 23, 24) Type material: Holotype: ♀ from South Africa, KwaZulu-Natal Province, Indlovu DC, Natal Karkloof, 50 km NNW Pietermaritzburg, −29.433333, 30.316667, 20 October 1985, elev. 4600 ft, forest, C. Griswold, J. Doyen, T. Meikle Griswold cols., preparation codes FML-01228-1231, 1247, 1354 [♀], deposited in NMSA. Figure 23. View largeDownload slide Izithunzi zondii sp. nov., copulatory organs (♀, NMSA). A, external female genitalia (ventral). B–D, vulva (B, dorsal; C, ventral; D, anterior). Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, 0.5 mm; D, 0.1 mm. Figure 23. View largeDownload slide Izithunzi zondii sp. nov., copulatory organs (♀, NMSA). A, external female genitalia (ventral). B–D, vulva (B, dorsal; C, ventral; D, anterior). Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, 0.5 mm; D, 0.1 mm. Figure 24. View largeDownload slide Izithunzi zondii sp. nov., preserved specimen (♀, NMSA). A, B, habitus (A, dorsal; B, ventral). Scale bar: A, B, 2 mm. Figure 24. View largeDownload slide Izithunzi zondii sp. nov., preserved specimen (♀, NMSA). A, B, habitus (A, dorsal; B, ventral). Scale bar: A, B, 2 mm. Further material examined: 4♀, same province as the holotype, North Uthungulu, iSimangaliso Wetland Park (label St. Lucia National Park), Fanieseiland, camp, [−28.112461, 32.431443], 24 January 1991, V. D. Roth, B. Roth cols., CAS (CASENT9021763). Etymology: The specific name is a patronym in honour of Mathew Sibusiso Zondi, who taught author Griswold to say ‘hello’, ‘thank you’ and ‘we are collecting spiders’ in Zulu. Diagnosis: Females of I. zondii sp. nov. resemble those of I. lina sp. nov. by the inner and outer spermathecae separated, the vulval plates pressed together anteriorly, and the anterior plate heavily sclerotized (Figs 10, 16, 23), but can be distinguished by the anterior plate strongly curved anteriorly (inverted V shaped), the posterior plate straight and the dorsal receptaculum partially integrated to the posterior plate (Fig. 23), while I. lina sp. nov. has anteriorly curved vulval plates and integrated dorsal receptaculum (Figs 10, 16). Description female (Natal Karkloof: habitus and measurements NMSA): Total length 9.04. Prosoma: length 3.84, width 2.77, height 2.2. Sternum: length 1.84, width 1.49. Leg measurements: femur: I: 10.8, II: 9.0, III: 7.52, IV: 9.4; patella: I: 1.0, II: 1.03, III: 0.99, IV: 1.02; tibia: I: 10.0, II: 7.92, III: 6.61, IV: 8.64; metatarsus: I: 9.7, II: 7.92, III: 6.74, IV: 8.64; tarsus: I: 2.02, II: 1.94, III: 1.92, IV: 2.65; podotarsite: I: 0.12, II: 0.13, III: 0.15, IV: 0.16; total: I: 33.64, II: 27.94, III: 23.93, IV: 30.51. Leg formula: 1423. Opisthosoma: length 5.75, width 3.65, height 3.84. Thoracic area lateral margins and central V-shaped pattern darkish dun, forming a continuum (Fig. 24). Chelicerae promargin with three bracket setae, and a row of seven macrosetae against the triangular lamina (Fig. 8). Labium dun, reddish at narrow area and white at apex (Fig. 24). Sternum dun (Fig. 24). Femora and tibiae dun, patellae, metatarsi and tarsi tan. Opisthosoma colour overall dark brown forming thick chevrons extending anteriorly (Fig. 24). Chevrons well-spaced anteriorly, clustered posteriorly, the first two forming a continuum (Fig. 24). Outer spermathecae minute (Fig. 23). Inner spermathecae oval (Fig. 24). Distribution: KwaZulu-Natal Province, South Africa, from Natal midlands at Karkloof to coast at iSimangaliso Wetland Park and surroundings (Fig. 6; Supporting Information, Fig. S1). Supporting Information Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Figure S1. Tree obtained in the maximum likelihood analysis of the combined molecular alignments with the preferred model for each gene fragment and gaps scored as absence/presence. The information of the results of the analyses under Bayesian (first 10% of generations were removed as burn-in) and parsimony (equal weights) are summarized in the node support. Circles at nodes indicate support levels subdivided by analysis (L, maximum likelihood; B, Bayesian; P, parsimony: bottom; see inset). Black indicates bootstrap proportions > 0.75, posterior probabilities > 0.95 and jackknifing proportions > 0.75; gray indicates that the clade was recovered but with lower support than the previous values; white indicates that the clade was not recovered. Supported families and genera discussed in the text are highlighted with lateral bars and shaded boxes, respectively. Figure S2. Summary phylogeny of Synspermiata with congruence among data partitions and optimality criteria. Topology and clade composition are based on Bayesian analysis of total evidence (see Supporting Information, Fig. S1). Boxes at nodes indicate each partition (morphology, combined DNA, individual genes or total evidence) subdivided by analysis method (B, Bayesian; L, maximum likelihood; P, parsimony; see inset). Black squares indicate that the clade was recovered (regardless of support) in the given analysis, white squares indicate that the clade was not recovered, and a diagonal line indicates that the clade was not tested (missing sequence data). Terminals with ‘*’ are missing sequence data. Table S1. List of the specimens used for the molecular phylogenetic analysis in the study, localities and vouchers codes. ACKNOWLEDGEMENTS We especially thank Ansie Dippenaar-Schoeman, Lauren Esposito, Peter Croeser, Gustavo Hormiga, Dawn Larsen, Leon Lotz, Robin Lyle, Petro Marais, Lorenzo Prendini, Martín Ramírez, Christine Rollard and Petra Sierwald for loaning the studied specimens. We also thank Norman Larsen and Gonzalo Giribet for providing images of live specimens. Several friends and colleagues helped us in field trips and museum visits: we are grateful to Lina Almeida, Peter Croeser, John Doyen, Vic Hamilton-Attwell, Charles Haddad, Norman Larsen, Teresa Meikle, Audrey Ndaba, Petra Sierwald, Esther van der Westhuizen and Hannah Wood. We further thank the authorities at the South African National Survey of Arachnida (SANSA, Western Cape, Eastern Cape) and Ansie Dippenaar-Schoeman, Charles Haddad and Mikhail (Mike) B. Mostovski for collecting permits and assistance in the field. We also thank the Willi Hennig Society for providing the programme TNT for free. F.M.L. wishes to thank a Postdoctoral Fellowship from the Schlinger Chair of Arachnology at California Academy of Sciences and Postdoctoral from the São Paulo Research Foundation, FAPESP (grant number 2014/23369-2). C.E.G. thanks the Schlinger Foundation and California Academy of Sciences Exline-Frizzell Fund for Arachnology, and the National Science Foundation (USA) for ‘Assembling the Tree of Life: Phylogeny of Spiders’ (grant number EAR-0228699; Wheeler, Sierwald, Prendini, Hormiga and Coddington P.I.s). A.P.G. is member of the Scientific Researcher Career of CONICET, and the results of this publication were funded, in part, by the Argentinean ‘Agencia Nacional de Promoción Científica y Tecnológica’ (ANPCyT) (grant number PICT 2015-0283; Martín Ramírez and Andrés A. Ojanguren-Affilastro P.I.s). [Version of Record, published online 18 December 2017; http://zoobank.org/ urn:lsid:zoobank.org:pub:5A38A09D- 3C0C-43DB-B355-4952C4BB4B0D] References Agnarsson I , Coddington JA , Kuntner M . 2013 . Systematics: progress in the study of spider diversity and evolution . In: Penney D , ed. Spider research in the 21st century: trends and perspectives . Rochdale : Siri Scientific Press , 58 – 111 . Akaike H . 1973 . Information theory and an extension of the maximum likelihood principle . In: Petrov BN , Csaki F , eds. 2nd International Symposium on Information Theory, Tsahkadsor, Armenia, USSR , 2–8 September 1971 . Budapest : Akademiai Kiado , 267 – 281 . Alayón GG . 1981 . El género Drymusa (Araneae: Loxoscelidae) en Cuba . Poeyana 219 : 1 – 19 . Alayón GG . 1987 . Descripción del macho de Drymusa armasi Alayón (Arachnida: Araneae: Loxoscelidae) . Poeyana 351 : 1 – 4 . Alayón GG . 2000 . Las arañas endémicas de Cuba (Arachnida: Araneae) . Revista Ibérica de Aracnología 2 : 1 – 48 . Álvarez-Padilla F , Hormiga G . 2008 . A protocol for digesting internal soft tissues and mounting spiders for scanning electron microscopy . Journal of Arachnology 35 : 538 – 542 . Google Scholar CrossRef Search ADS Ausserer A . 1867 . Die Arachniden Tirols nach ihrer horizontalen und verticalen Verbreitung; I . Verhandlungen der Kaiserlich-Königlichen Zoologisch-Botanischen Gesellschaft in Wien 17 : 137 – 170 . Austad SN . 1984 . Evolution of sperm priority patterns in spiders . In: Smith RL , ed. Sperm competition and the evolution of animal mating systems . New York : Academic Press , 223 – 249 . Google Scholar CrossRef Search ADS Blackwall J . 1864 . A history of the spiders of Great Britain and Ireland, Vol. 2 . London : Ray Society , 175 – 384 . Binford GJ , Callahan MS , Bodner MR , Rynerson MR , Núñez PB , Ellison CE , Duncan RP . 2008 . Phylogenetic relationships of Loxosceles and Sicarius spiders are consistent with Western Gondwanan vicariance . Molecular Phylogenetics and Evolution 49 : 538 – 553 . Google Scholar CrossRef Search ADS Binford GJ , Wells MA . 2003 . The phylogenetic distribution of sphingomyelinase D activity in venoms of Haplogyne spiders . Comparative Biochemistry and Physiology B: Biochemistry & Molecular Biology 135 : 25 – 33 . Google Scholar CrossRef Search ADS Bonaldo AB , Rheims CA , Brescovit AD . 2006 . Four new species of Drymusa Simon (Araneae, Drymusidae) from Brazilian Oriental Amazonia . Revista Brasileira de Zoologia, Curitiba 23 : 455 – 359 . Google Scholar CrossRef Search ADS Bonnet P . 1956 . Bibliographia araneorum . Toulouse 2 : 919 – 1926 . Brescovit AD , Bonaldo AB , Rheims CA . 2004 . A new species of Drymusa Simon, 1891 (Araneae, Drymusidae) from Brazil . Zootaxa 697 : 1 – 5 . Brignoli PM . 1976 . Beiträge zur Kenntnis der Scytodidae (Araneae) . Revue Suisse de Zoologie 83 : 125 – 191 . Brignoli PM . 1983 . A catalogue of the Araneae described between 1940 and 1981 . Manchester : Manchester University Press , 1 – 755 . Bristowe WS . 1938 . The classification of spiders . Proceedings of the Zoological Society of London (B) 108 : 285 – 322 . Google Scholar CrossRef Search ADS Brown JM , Lemmon AR . 2007 . The importance of data partitioning and the utility of Bayes factors in Bayesian phylogenetics . Systematic Biology 56 : 643 – 655 . Google Scholar CrossRef Search ADS Bryant EB . 1948 . The spiders of Hispaniola . Bulletin of the Museum of Comparative Zoology at Harvard College 100 : 329 – 447 . Buckley TR , Arensburger P , Simon C , Chambers GK . 2002 . Combined data, Bayesian phylogenetics, and the origin of the New Zealand cicada genera . Systematic Biology 51 : 4 – 18 . Google Scholar CrossRef Search ADS Coddington JA . 2005 . Phylogeny and classification of spiders . In: Ubick D , Paquin P , Cushing PE , Roth V , eds. Spiders of North America: an identification manual . American Arachnological Society , 18 – 24 . Coddington JA , Giribet G , Harvey MS , Prendini L , Walter DE . 2004 . Arachnida . In: Cracraft J , Donoghue MJ , eds. Assembling the tree of life . New York : Oxford University Press , 296 – 318 . Coddington JA , Levi HW . 1991 . Systematics and evolution of spiders (Araneae) . Annual Review of Ecology and Systematics 22 : 565 – 592 . Google Scholar CrossRef Search ADS Colgan DJ , McLauchlan A , Wilson GDF , Livingston SP , Edgecombe GD , Macaranas J , Cassis G , Gray MR , 1998 . Histone H3 and U2 snRNA DNA sequences and arthropod molecular evolution . Australian Journal of Zoology . 46 , 419 – 437 . Google Scholar CrossRef Search ADS Craven SA . 2012 . A history of cave explorations on the Cape Peninsula mountain chain of South Africa . In: John GUNN and David LOWE, eds Cave and karst science, Vol. 39, No. 3 . Buxton : Transactions of the British Cave Research Association , 101 – 108 . Darriba D , Taboada GL , Doallo R , Posada D . 2012 . jModelTest 2: more models, new heuristics and parallel computing . Nature Methods 9 : 772 . Google Scholar CrossRef Search ADS Dippenaar-Schoeman AS , Jocqué R . 1997 . African spiders: an identification manual. Plant Protection Research Institute Handbook, No. 9 . Pretoria : Agricultural Research Council, South Africa , 1 – 392 . Fage L . 1912 . Etudes sur les araignées cavernicoles. I. Revision des Ochyroceratidae (n. fam.). In: Biospelogica, XXV . Archives de Zoologie Expérimentale et Générale 5 : 97 – 162 . Fage L . 1913 . Etudes sur les Araignées cavernicoles. II. Revision des Leptonetidae. In: Biospelogica, XXIX . Archives de Zoologie Expérimentale et Générale 10 : 479 – 576 . Filmer MR . 1991 . Southern African spiders: an identification guide . Cape Town : Struik Nature , 1 – 128 . Filmer MR . 2010 . Filmer’s spiders: an identification guide for Southern Africa, 2nd edn . Cape Town : Struik Nature , 1 – 128 . Foelix RF . 2011 . Biology of spiders, 3rd edn . New York : Oxford University Press . Folmer O , Black M , Hoeh W , Lutz R , Vrijenhoek R . 1994 . DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates . Molecular Marine Biology and Biotechnology . 3 : 294 – 299 . Gerhardt U , Kästner A . 1938 . Araneae = Echte Spinnen = Webspinnen . In: Kükenthal W , Krumbach T , eds. Handbuch der zoologie, Vol. 3 . Berlin : DeGruyter , 497 – 656 , figs 627–854. Gertsch WJ . 1967 . The spider genus Loxosceles in South America (Araneae, Scytodidae) . Bulletin of the American Museum of Natural History 136 : 117 – 174 . Giribet G , Carranza S , Baguñà J , Riutort M , Ribera C . 1996 . First molecular evidence for the existence of a Tardigrada + Arthropoda clade . Molecular Biology and Evolution . 13 , 76 – 84 . Google Scholar CrossRef Search ADS Giribet G , Edgecomb G . 2006 . Conflict between datasets and phylogeny of centipedes: an analysis based on seven genes and morphology . Proceedings of the Royal Society of London B: Biological Sciences 273 : 531 – 538 . Google Scholar CrossRef Search ADS Goloboff PA . 1993 . Estimating character weights during tree search . Cladistics 9 : 83 – 91 . Google Scholar CrossRef Search ADS Goloboff PA , Carpenter JM , Arias JS , Miranda Esquivel DR . 2008b . Weighting against homoplasy improves phylogenetic analysis of morphological data sets . Cladistics 24 : 1 – 16 . Google Scholar CrossRef Search ADS Goloboff PA , Catalano SC . 2016 . TNT version 1.5, including a full implementation of phylogenetic morphometric . Cladistics 32 : 221 – 238 . Google Scholar CrossRef Search ADS Goloboff PA , Farris JS , Nixon KC . 2008a . TNT, a free program for phylogenetic analysis . Cladistics 24 : 774 – 786 . Google Scholar CrossRef Search ADS Goloboff PA , Ramírez MJ . 1992 . A new species of Drymusa (Araneae: Scytodidae) from Argentina . Journal of the New York Entomological Society 99 : 691 – 695 . Grismado CJ , Ramírez MJ , Izquierdo MA . 2014 . Araneae: taxonomía, diversidad, y clave de identificación de familias de la Argentina . In: Roig-Juñent S , Claps LE , Morrone JJ , eds. Biodiversidad de Artrópodos Argentinos vol. 3. INSUE – UNT . Argentina : San Miguel de Tucumán , 55 – 93 . Griswold CE , Audisio T , Ledford JM . 2012 . An extraordinary new family of spiders from caves in the Pacific Northwest (Araneae, Trogloraptoridae, new family) . ZooKeys 215 : 77 – 102 . Google Scholar CrossRef Search ADS Griswold CE , Ramírez MJ . 2017 . Phylogeny of spiders . In: Ubick D , Paquin P , Cushing PE , eds. Spiders of North America: an identification manual, 2nd edn . American Arachnological Society , 17 – 29 . Guindon S , Gascuel O . 2003 . A simple, fast and accurate method to estimate large phylogenies by maximum-likelihood . Systematic Biology 52 : 696 – 704 . Google Scholar CrossRef Search ADS Hall TA . 1999 . BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT . Nucleic Acids Symposium Series 41 : 95 – 98 . Hedin MC , Maddison WP . 2001 . A combined molecular approach to phylogeny of the jumping spider subfamily Dendryphantinae (Araneae: Salticidae) . Molecular Phylogenetics and Evolution 18 , 386 – 403 . doi: 10.1006/mpev.2000.0883 Google Scholar CrossRef Search ADS Jocqué R , Dippenaar-Schoeman SA . 2006 . Spider families of the world . Tervuren : Royal Museum for Central Africa , 1 – 336 . Katoh K , Standley DM . 2013 . MAFFT multiple sequence alignment software version 7: improvements in performance and usability . Molecular Biology and Evolution 30 : 772 – 780 . Google Scholar CrossRef Search ADS Keyserling E . 1880 . Die Spinnen Amerikas, I. Laterigradae, Vol. 1 . Nürnberg : Verlag von Bauer & Raspe , 1 – 283 . Keyserling E . 1891 . Die Spinnen Amerikas. Brasilianische Spinnen, Vol. 3 . Nürnberg : Verlag von Bauer & Raspe , 1 – 278 . Khmelik VV , Kozub D , Glazunov A . 2006 . Helicon Focus, Versions 3.10.3 and 4.01 Pro . Available at http://helicon.com.ua/heliconfocus/ (accessed 21 August 2016 ). Kovoor J , Zylberberg L . 1972 . Histologie et infrastructure de la glande chélicérienne de Scytodes delicatula Sim. (Araneae, Scytodidae) . Annales des Sciencies Naturelles, Zoologie, Paris 14 : 333 – 388 . Labarque FM , Grismado CJ , Ramírez MJ , Yan HM , Griswold CE . 2009 . The southeast Asian genus Stedocys Ono, 1995 (Araneae: Scytodidae): first descriptions of female genitalia and a new species from China . Zootaxa 2297 : 1 – 14 . Labarque FM , Ramírez MJ . 2007a . A new species of Drymusa Simon, 1891 (Araneae: Drymusidae) from Chile . Studies on Neotropical Fauna and Environment 42 : 241 – 246 . Google Scholar CrossRef Search ADS Labarque FM , Ramírez MJ . 2007b . Description of the female of Drymusa serrana Goloboff & Ramírez, 1991 (Araneae: Drymusidae) with notes on its biology . Zootaxa 1580 : 27 – 33 . Labarque FM , Ramírez MJ . 2012 . The placement of the spider genus Periegops and the phylogeny of Scytodoidea (Araneae: Araneomorphae) . Zootaxa 3312 : 1 – 44 . Labarque FM , Wolff JO , Michalik P , Griswold CE , Ramírez MJ . 2017 . The evolution and function of spider feet (Araneae, Arachnida): multiple acquisitions of distal articulations . Zoological Journal of the Linnean Society 1 – 34 . Larsen N . 1994 . Net-casters, spitters and invalida spiders of the Cape . In: Dehning C , ed. The Spider Club News 9 . Johannesburg : The Spider Club of Southern Africa , 3 – 5 . Latreille PA . 1804 . Tableau methodique des Insectes . Nouveau Dictionnaire d’Histoire Naturelle, Paris 24 : 129 – 295 . Lawrence RF . 1964 . New cavernicolous spiders from South Africa . Annals of the South African Museum 48 : 57 – 75 . Ledford JM , Paquin P , Cokendolpher J , Campbell J , Griswold CE . 2011 . Systematics of the spider genus Neoleptoneta Brignoli, 1972 (Araneae: Leptonetidae) with a discussion of the morphology and relationships for the North American Leptonetidae . Invertebrate Systematics 25 : 334 – 388 . Google Scholar CrossRef Search ADS Lehtinen PT . 1967 . Classification of the cribellate spiders and some allied families, with notes on the evolution of the suborder Araneomorpha . Annales Zoologici Fennici 4 : 199 – 468 . Lehtinen PT . 1986 . Evolution of the Scytodoidea . In: Proceedings of the Ninth International Congress of Arachnology, Panama, 1983 . (Smithsonian Institution Press, Washington, DC, 1986). pp. 149 – 157 . (University of Panama, Smithsonian Tropical Research Institute, and Centre International de Documentation Arachnologique Panama City, Panama 1–8 August 1983). Lotz LN . 2012 . Present status of Sicariidae (Arachnida: Araneae) in the Afrotropical region . Zootaxa 3522 : 1 – 41 . Lowe RT . 1832 . Descriptions of two species of Araneidae, natives of Madeira . The Zoological Journal 5 : 320 – 323 . Marx G . 1888 . On a new and interesting spider . Entomologica Americana 4 : 160 – 162 . Michalik P , Ramírez MJ . 2014 . Evolutionary morphology of the male reproductive system, spermatozoa and seminal fluid of spiders (Araneae, Arachnida) – current knowledge and future directions . Arthropod Structure & Development 43 : 291 – 322 . Google Scholar CrossRef Search ADS Miller JA , Griswold CE , Yin CM . 2009 . The symphytognathoid spiders of the Gaoligongshan, Yunnan, China (Araneae: Araneoidea): systematics and diversity of micro-orbweavers . ZooKeys 11 : 9 – 195 . Miller MA , Pfeiffer W , Schwartz T . 2010 . Creating the CIPRES Science Gateway for inference of large phylogenetic trees . In: Proceedings of the Gateway Computing Environments Workshop (GCE) , 14 November 2010 , New Orleans, LA : IEEE (Institute of Electrical and Electronics Engineers) , 1 – 8 . Morrone JJ . 2004 . La zona de transición Sudamericana: caracterización y relevancia evolutiva . Acta Entomológica Chilena 28 : 41 – 50 . Morrone JJ . 2014 . Biogeographical regionalisation of the Neotropical region . Zootaxa 3782 : 1 – 110 . Google Scholar CrossRef Search ADS Newlands G . 1975 . A revision of the spider genus Loxosceles Heinecken & Lowe, 1835 (Araneae: Scytodidae) in southern Africa with notes on the natural history and morphology . Journal of the Entomological Society of South Africa 38 : 141 – 154 . Pérez-González A , Rubio GD , Ramírez MJ . 2016 . Insights on vulval morphology in Ochyroceratinae with a rediagnosis of the subfamily and description of the first Argentinean species (Araneae: Synspermiata: Ochyroceratidae) . Zoologischer Anzeiger 260 : 33 – 44 . Google Scholar CrossRef Search ADS Petrunkevitch A . 1911 . A synonymic index-catalogue of spiders of North, Central and South America with all adjacent islands, Greenland, Bermuda, West Indies, Terra del Fuego, Galapagos, etc . Bulletin of the American Museum of Natural History 29 : 1 – 791 . Petrunkevitch A . 1928 . Systema Aranearum . Transactions of the Connecticut Academy of Arts and Sciences 29 : 1 – 270 . Petrunkevitch A . 1939 . Catalogue of American spiders. Part one . Transactions of the Connecticut Academy of Arts and Sciences 33 : 133 – 338 . Pitman WC III , Cande S , LaBrecque J , Pindell J . 1993 . Fragmentation of Gondwana: the separation of Africa from South America . In: Goldblatt P , ed. Biological relationships between Africa and South America . New Haven : Yale University Press , 19 – 32 . Google Scholar CrossRef Search ADS Platnick NI . 1987 . Hypochilus pococki. In: Forster RR, Platnick NI,Gray MR. (1987). A review of the spider superfamilies Hypochiloidea and Austrochiloidea (Araneae, Araneomorphae) . Bulletin of the American Museum of Natural History 185 : 1 – 116 . Platnick NI . 1989 . Advances in spider taxonomy 1981–1987: a supplement to Brignoli’s catalogue of the Araneae described between 1940 and 1981 . Manchester : Manchester University Press , 1 – 673 . Platnick NI . 1993 . Advances in spider taxonomy 1988–1991, with synonymies and transfers 1940–1980 . New York : The New York Entomological Society , 1 – 846 . Platnick NI , Coddington JA , Forster RR , Griswold CE . 1991 . Spinneret morphology and the phylogeny of haplogyne spiders (Araneae, Araneomorphae) . American Museum Novitates 3016 : 1 – 73 . Polotow D , Carmichael A , Griswold C . 2015 . Total evidence analysis of the phylogenetic relationships of Lycosoidea spiders (Araneae, Entelegynae) . Invertebrate Systematics 29 : 124 – 163 . Google Scholar CrossRef Search ADS Purcell WF . 1904 . Descriptions of new genera and species of South African spiders . Transactions of the South African Philosophical Society 15 : 115 – 173 . Google Scholar CrossRef Search ADS Rambaut A . 2006–2016 . Tree Figure Drawing Tool Version 1.4 . Institute of Evolutionary Biology, University of Edinburgh . Available at: http://tree.bio.ed.ac.uk/ Rambaut A , Drummund AJ . 2009 . Tracer Version 1.5: MCMC trace analysis tool . Institute of Evolutionary Biology, University of Edinburgh . Available at: http://tree.bio.ed.ac.uk/software/tracer Ramírez MJ . 2000 . Respiratory system morphology and the phylogeny of haplogyne spiders (Araneae, Araneomorphae) . Journal of Arachnology 28 : 149 – 157 . Google Scholar CrossRef Search ADS Ramírez MJ . 2014 . The morphology and phylogeny of dionychan spiders (Araneae: Araneomorphae) . Bulletin of the American Museum of Natural History 390 : 1 – 374 . Google Scholar CrossRef Search ADS Ramírez MJ , Crespo FA , Iglesias MS , Valverde AC . 1999 . El ABC en la determinación de artrópodos. Claves para especímenes presentes en la Argentina I . Buenos Aires : Editorial CCC Educando , 1 – 107 . Rheims CA , Brescovit AD , Bonaldo AB . 2008 . Notes on the Amazonian species of the genus Drymusa Simon (Araneae, Drymusidae) . Journal of Arachnology 36 : 164 – 166 . Google Scholar CrossRef Search ADS Rindal E , Brower AVZ . 2011 . Do model-based phylogenetic analyses perform better than parsimony? A test with empirical data . Cladistics 27 : 331 – 334 . Google Scholar CrossRef Search ADS Roewer CF . 1942 . Katalog der Araneae von 1758 bis 1940 . Bremen : Natura/IRSN , 1 – 1040 . Ronquist F , Huelsenbeck JP . 2003 . MrBayes 3: Bayesian phylogenetic inference under mixed models . Bioinformatics 19 : 1572 – 1574 . Google Scholar CrossRef Search ADS Simmons MP , Ochoterena H . 2000 . Gaps as characters in sequence-based phylogenetic analyses . Systematic Biology 49 : 369 – 381 . Google Scholar CrossRef Search ADS Simon E . 1891 . On the spiders of the island of St. Vincent. Part 1 . Proceedings of the Zoological Society of London 1891 : 549 – 575 . Simon E . 1893 . Histoire naturelle das araignées . Paris : LIBRAIRIE ENCYCLOPÉDIQUE DE RORET , 257 – 488 . Simon C , Frati F , Beckenbach A , Crespi B , Liu H , Flook P . 1994 . Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers . Annals of the Entomological Society of America 87 : 651 – 701 . doi: 10.1093/aesa/87.6.651 Google Scholar CrossRef Search ADS Stamatakis A . 2006 . RAxML-VI-HPC: maximum likelihood based phylogenetic analyses with thousands of taxa and mixed models . Bioinformatics 22 : 2688 – 2690 . Google Scholar CrossRef Search ADS Stamatakis A , Hoover P , Rougemont J . 2008 . A fast bootstrapping algorithm for the RAxML web-servers . Systematic Biology 57 : 758 – 771 . Google Scholar CrossRef Search ADS Valerio CE . 1971 . The spider genus Drymusa in the New World (Araneae: Scytodidae) . Florida Entomologist 54 : 193 – 200 . Google Scholar CrossRef Search ADS Valerio CE . 1974 . Prey capture by Drymusa dinora (Aranea, Scytodidae) . Psyche 81 : 248 – 287 . Wheeler WC , Coddington JA , Crowley LM , Dimitrov D , Goloboff PA , Griswold CE , Hormiga G , Prendini L , Ramírez MJ , Sierwald P , Almeida-Silva L , Álvarez-Padilla F , Arnedo MA , Benavides Silva LR , Benjamin SP , Bond JE , Grismado CJ , Hasan E , Hedin M , Izquierdo MA , Labarque FM , Ledford J , Lopardo L , Maddison WP , Miller JA , Piacentini LN , Platnick NI , Polotow D , Silva-Dávila D , Scharff N , Szüts T , Ubick D , Vink CJ , Wood HM , Zhang J . 2016 . The spider tree of life: phylogeny of Araneae based on target-gene analyses from an extensive taxon sampling . Cladistics 1 – 43 . Wheeler WC , Hayashi CY . 1998 . The phylogeny of the extant chelicerate orders . Cladistics 14 : 173 – 192 . Google Scholar CrossRef Search ADS World Spider Catalog . 2017 . World Spider Catalog, version 18.0 . Natural History Museum Bern . Available at: http://wsc.nmbe.ch (accessed on 10 March 2017 ). Young ND , Healy J . 2002 . GapCoder . Available at: http://www.trinity.edu/nyoung/GapCoder/Download.html © 2017 The Linnean Society of London, Zoological Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Zoological Journal of the Linnean Society Oxford University Press http://www.deepdyve.com/lp/oxford-university-press/molecular-phylogeny-and-revision-of-the-false-violin-spiders-araneae-bxaJNsCq9k

Loading next page...

Publisher: Oxford University Press
ISSN: 0024-4082
eISSN: 1096-3642
DOI: 10.1093/zoolinnean/zlx088
Publisher site: See Article on Publisher Site

Abstract

Abstract The family Drymusidae includes 16 cryptic spiders that build irregular webs in dark places. The family is distributed in South Africa, the Neotropical and Andean regions. Here, we use a molecular approach to infer the relationships of Drymusidae using three mitochondrial (COI, 16S, 12S) and three nuclear (H3, 28S, 18S) markers. Our preferred analyses support Drymusidae and its American and African clades, which emerge as sister groups. Our analyses suggest a Gondwanan distribution of Drymusidae and a Westward radiation of Izithunzi gen. nov. within South Africa, but both hypotheses remain to be thoroughly tested. We describe Izithunzi gen. nov. for the African species. All previous African species are redescribed and new combinations are proposed: Izithunzi capense (Simon) comb. nov., Izithunzi productum (Purcell) comb. nov. and Izithunzi silvicola (Purcell) comb. nov. Two new species are described: Izithunzi lina sp. nov. (known from both sexes) and Izithunzi zondii sp. nov. (known only from females). The male of I. productum (Purcell) comb. nov. is also described for the first time. We consider Loxosceles valida Lawrence, 1964, a junior synonym of I. capense (Simon) comb. nov. (new synonymy). We also provide a dichotomous key for Izithunzi gen. nov. species. Gondwana, target-genes, Scytodoidea, Synspermiata INTRODUCTION African Drymusidae Simon, 1893, are shy, cryptic spiders that build irregular, tangled webs beneath logs and stones and in other dark, damp places, such as rock overhangs and caves (Fig. 1; Larsen, 1994). Because of their retiring nature, most will never be encountered by people. Chance encounters with the huge species from Table Mountain may elicit surprise and concern due to their size and resemblance to the venomous violin spiders (Sicariidae Keyserling, 1880: Loxosceles Heineken & Lowe, 1832) (Purcell, 1904; Larsen, 1994; Binford & Wells, 2003). This resemblance has led to the common name ‘false violin spiders’ for the African Drymusidae (Dippenaar-Schoeman & Jocqué, 1997; Filmer, 2010). In spite of this morphological similarity, and close relationship to Sicariidae in the superfamily Scytodoidea, Drymusidae species show no evidence of being dangerous to humans (Binford & Wells, 2003). Figure 1. View largeDownload slide African Drymusidae habitat. Izithunzi capense comb. nov. (♀, A–C CASENT 9023625), Izithunzi aff. lina sp. nov. (♀, D), live specimens. A–C, Newlands Forest Preserve, specimens hanging beneath loose space webs inside a cave (photographs by Norman Larsen) (C specimen feeding on Theridion delicatum O. Pickard-Cambridge, 1904). D, Helderberg Nature Reserve, specimen on log (photograph by Gonzalo Giribet). Figure 1. View largeDownload slide African Drymusidae habitat. Izithunzi capense comb. nov. (♀, A–C CASENT 9023625), Izithunzi aff. lina sp. nov. (♀, D), live specimens. A–C, Newlands Forest Preserve, specimens hanging beneath loose space webs inside a cave (photographs by Norman Larsen) (C specimen feeding on Theridion delicatum O. Pickard-Cambridge, 1904). D, Helderberg Nature Reserve, specimen on log (photograph by Gonzalo Giribet). Drymusidae include the genus Drymusa Simon, 1891, with 16 extant species described so far, four of them from South Africa (World Spider Catalog, 2017). The other Drymusidae species are distributed across the Neotropical (11 species) and Andean regions (one species) (Morrone, 2004, 2014; World Spider Catalog, 2017). The first Drymusidae described for South Africa was Drymusa capensis Simon, 1893, from Table Mountain (Western Cape). The original description was brief, based on one immature holotype specimen, and lacks important morphological details, for example, those of genitalic organs. Purcell (1904) was the first to mention the female of D. capensis (although he did not describe it), along with the descriptions of two new species for South Africa: Drymusa producta Purcell, 1904 (originally known from female specimens) and Drymusa silvicola Purcell, 1904 (originally known from both sexes). He also provided a key to females (Purcell, 1904: 154). Recently, Lotz (2012) proposed the new combination, Drymusa valida (Lawrence, 1964), after he transferred this species (originally described in Loxosceles, both sexes known) from Sicariidae to Drymusidae. This species is also known from Table Mountain (i.e. sympatric with D. capensis). Drymusidae have long been associated with violin spiders (as Loxoscelidae or Sicariidae) and spitting spiders (Scytodidae Blackwall, 1864) (Simon, 1893; Gertsch, 1967; Alayón, 1981; Lehtinen, 1986). Cladistic analyses over the last two decades (Platnick et al., 1991; Ramírez, 2000; Wheeler et al., 2016) have suggested the sister group relationship of Drymusidae and Scytodidae and confirmed their distinctness from Loxosceles. A recent morphological cladistic analysis (Labarque & Ramírez, 2012) proposed that the Australasian Periegopidae Simon, 1893, are the sister group of Drymusidae. Nevertheless, the molecular cladistic results of the Assembling the Tree of Life (AToL): Phylogeny of Spiders Project (Wheeler et al., 2016) suggested that the traditional limits of Drymusidae are paraphyletic with respect to Periegopidae and that the African species placed in Drymusa are more closely related to Periegops Simon, 1893, from Australia and New Zealand than to other Drymusa from the Americas. Suggestions that Drymusidae may include more than one genus have been made by Lehtinen (1967) and Platnick et al. (1991), especially by the former, who reviewed Neotropical and South African species and considered them not congeneric. In this study, we describe Izithunzi gen. nov. for the African species formerly placed in Drymusa. The male of Izithunzi productum (Purcell, 1904) comb. nov. is described for the first time, and D. valida is considered a junior synonym of Izithunzi capense (Simon, 1893) comb. nov. (new syn.). We describe Izithunzi lina sp. nov. (male and female) and Izithunzi zondii sp. nov. (female) and redescribe and key all five species. Furthermore, in cooperation with SANSA (South African National Survey of Arachnida), we map and produce a *.kml file documenting all known records of Drymusidae in South Africa (see Supporting Information). With the aim to investigate the monophyly of Izithunzi gen. nov., examine the delimitation of contentious genera and determine the placement of Drymusidae among Scytodoidea, we generated a molecular data set for Drymusidae species (i.e. Neotropical, Andean and South African) including six markers: three nuclear genes (the ribosomal 28S rRNA and 18S rRNA and the protein-coding histone H3) and three mitochondrial genes [the ribosomal 16S rRNA and 12S rRNA and the protein-coding cytochrome c oxidase subunit I (COI)]. These sequences were included in a major data set, the spider AToL Synspermiata data subset (Wheeler et al., 2016), and analysed under maximum likelihood, Bayesian inference and parsimony, separately and concatenated. Our results highly support Drymusidae as monophyletic, the sister group relationship between the American and African clades and the monophyly of both of those clades. MATERIAL AND METHODS Molecular data The molecular data set included 108 species, six of which were sequenced in the present study and 102 were extracted from the spiders AToL project (Wheeler et al., 2016). Synspermiata is represented by 99 species in 68 genera (Supporting Information, Table S1). Outgroup terminals included Hypochilus pococki Platnick, 1987 (Hypochilidae Marx, 1888) and eight genera of Filistatidae Ausserer, 1867. Total DNA was extracted from one or two legs of freshly collected specimens that were preserved in 95–100% ethanol or propylene glycol (i.e. Drymusa armasi Alayón, 1981). Extraction, amplification and sequencing followed the protocols described by Polotow, Carmichael & Griswold (2015) and Wheeler et al. (2016). Partial fragments of the mitochondrial genes cytochrome c oxidase subunit I (COI, ranging in length from 321 to 1076 bp), 12S rRNA (12S, from 169 to 348 bp) and 16S rRNA (16S, from 323 to 453 bp) and the nuclear genes 28S rRNA (18S, from 312 to 1729 bp), 28S rRNA (28S, from 507 to 2236 bp) and Histone H3 (H3, from 295 to 318 bp) were amplified. Primer pairs and PCR annealing conditions for each locus are listed in Table 1. Table 1. Primer sequences, source and annealing temperatures used to generate data for this study Gene Direction Sequence Reference Annealing temperature COI Fwd: LC011490-oono 5′-CWA CAA AYC ATA RRG ATA TTG G-3′ Modified from Folmer et al. (1994) 50–54 °C Rev: C1-N-2191 5′-CCC CGT AAA ATT AAA ATA TAA ACT TC-3′ Simon et al. (1994) H3 Fwd: H3aF 5′-ATG GCT CGT ACC AAG CAG ACV GC-3′ Colgan et al. (1998) 52 °C Rev: H3aR 5′-ATA TCC TTRGGC ATR ATRGTG AC-3′ Colgan et al. (1998) 28S Fwd: 28S-0sc 5′-CGT GAA ACT GCT CAG AGG-3′ Modified from Hedin & Maddison (2001) 50–53 °C Rev: 28S-C 5′-GGT TCG ATT AGT CTT TCG CC-3′ Hedin & Maddison (2001) 18S Fwd: 18S-1F 5′-TAC CTG GTT GAT CCT GCC AGT AG-3′ Giribet et al. (1996) 50 °C Rev: 18S-5-9-intR 5′-ATT CCG WTA ACG ADC GAG-3′ Miller et al. (2010) 18S Fwd: 18S-5F 5′-GCG AAA GCA TTT GCC AAG AA-3′ Giribet et al. (1996) 50 °C Rev: 18S-9R 5′-GAT CCT TCC GCA GGT TCA CCT AC-3′ Giribet et al. (1996) Gene Direction Sequence Reference Annealing temperature COI Fwd: LC011490-oono 5′-CWA CAA AYC ATA RRG ATA TTG G-3′ Modified from Folmer et al. (1994) 50–54 °C Rev: C1-N-2191 5′-CCC CGT AAA ATT AAA ATA TAA ACT TC-3′ Simon et al. (1994) H3 Fwd: H3aF 5′-ATG GCT CGT ACC AAG CAG ACV GC-3′ Colgan et al. (1998) 52 °C Rev: H3aR 5′-ATA TCC TTRGGC ATR ATRGTG AC-3′ Colgan et al. (1998) 28S Fwd: 28S-0sc 5′-CGT GAA ACT GCT CAG AGG-3′ Modified from Hedin & Maddison (2001) 50–53 °C Rev: 28S-C 5′-GGT TCG ATT AGT CTT TCG CC-3′ Hedin & Maddison (2001) 18S Fwd: 18S-1F 5′-TAC CTG GTT GAT CCT GCC AGT AG-3′ Giribet et al. (1996) 50 °C Rev: 18S-5-9-intR 5′-ATT CCG WTA ACG ADC GAG-3′ Miller et al. (2010) 18S Fwd: 18S-5F 5′-GCG AAA GCA TTT GCC AAG AA-3′ Giribet et al. (1996) 50 °C Rev: 18S-9R 5′-GAT CCT TCC GCA GGT TCA CCT AC-3′ Giribet et al. (1996) COI, mitochondrial cytochrome c oxidase subunit I; 28S, nuclear 28S rDNA; 18S, nuclear 18S rDNA; H3, nuclear histone 3; Fwd, forward; Rev, reverse. View Large Table 1. Primer sequences, source and annealing temperatures used to generate data for this study Gene Direction Sequence Reference Annealing temperature COI Fwd: LC011490-oono 5′-CWA CAA AYC ATA RRG ATA TTG G-3′ Modified from Folmer et al. (1994) 50–54 °C Rev: C1-N-2191 5′-CCC CGT AAA ATT AAA ATA TAA ACT TC-3′ Simon et al. (1994) H3 Fwd: H3aF 5′-ATG GCT CGT ACC AAG CAG ACV GC-3′ Colgan et al. (1998) 52 °C Rev: H3aR 5′-ATA TCC TTRGGC ATR ATRGTG AC-3′ Colgan et al. (1998) 28S Fwd: 28S-0sc 5′-CGT GAA ACT GCT CAG AGG-3′ Modified from Hedin & Maddison (2001) 50–53 °C Rev: 28S-C 5′-GGT TCG ATT AGT CTT TCG CC-3′ Hedin & Maddison (2001) 18S Fwd: 18S-1F 5′-TAC CTG GTT GAT CCT GCC AGT AG-3′ Giribet et al. (1996) 50 °C Rev: 18S-5-9-intR 5′-ATT CCG WTA ACG ADC GAG-3′ Miller et al. (2010) 18S Fwd: 18S-5F 5′-GCG AAA GCA TTT GCC AAG AA-3′ Giribet et al. (1996) 50 °C Rev: 18S-9R 5′-GAT CCT TCC GCA GGT TCA CCT AC-3′ Giribet et al. (1996) Gene Direction Sequence Reference Annealing temperature COI Fwd: LC011490-oono 5′-CWA CAA AYC ATA RRG ATA TTG G-3′ Modified from Folmer et al. (1994) 50–54 °C Rev: C1-N-2191 5′-CCC CGT AAA ATT AAA ATA TAA ACT TC-3′ Simon et al. (1994) H3 Fwd: H3aF 5′-ATG GCT CGT ACC AAG CAG ACV GC-3′ Colgan et al. (1998) 52 °C Rev: H3aR 5′-ATA TCC TTRGGC ATR ATRGTG AC-3′ Colgan et al. (1998) 28S Fwd: 28S-0sc 5′-CGT GAA ACT GCT CAG AGG-3′ Modified from Hedin & Maddison (2001) 50–53 °C Rev: 28S-C 5′-GGT TCG ATT AGT CTT TCG CC-3′ Hedin & Maddison (2001) 18S Fwd: 18S-1F 5′-TAC CTG GTT GAT CCT GCC AGT AG-3′ Giribet et al. (1996) 50 °C Rev: 18S-5-9-intR 5′-ATT CCG WTA ACG ADC GAG-3′ Miller et al. (2010) 18S Fwd: 18S-5F 5′-GCG AAA GCA TTT GCC AAG AA-3′ Giribet et al. (1996) 50 °C Rev: 18S-9R 5′-GAT CCT TCC GCA GGT TCA CCT AC-3′ Giribet et al. (1996) COI, mitochondrial cytochrome c oxidase subunit I; 28S, nuclear 28S rDNA; 18S, nuclear 18S rDNA; H3, nuclear histone 3; Fwd, forward; Rev, reverse. View Large Sequence reconciliation, edition and chromatogram evaluation were performed using Geneious 7.1.5 (Biomatters; http://www.geneious.com/). Each sequence was checked for contamination using an NCBI BLAST search (http://ncbi.nlm.nih.gov/BLAST). Low-quality samples with ambiguous readings were removed from the matrix. In addition, sequences were managed in Bioedit ver. 7.0.5.2 (Hall, 1999). Taxonomic and sequence information for the study specimens are listed in Supporting Information (Table S1). Automatic multiple alignments were built using the E-INS-i very slow search strategy in MAFFT ver. 7 (Katoh & Standley, 2013), which incorporates affine gap costs. Several gap opening (GOP, ‘Gap opening penalty’) costs were investigated to assess their influence on the results. The following parameter costs were used (GOP): 1.00, 1.53 (default), 3.00. These values resulted in several alignments ranging from numerous short gaps (‘gappy’) to fewer and longer gaps in the sequences (‘compacted’). The minimum value of the rescaled incongruence length difference (RILD) was used as a criterion to select the optimal alignment for each gene fragment (Wheeler & Hayashi, 1998). The RILD measures the increase in homoplasy generated as a result of combining data sets; hence, selecting the alignment combination that minimizes the RILD value maximizes the congruence among partitions. To avoid overweighting contiguous gap positions, gaps were recoded as presence/absence characters using the simple code method by Simmons & Ochoterena (2000) as implemented in the program Gapcoder (Young & Healy, 2002). Phylogenetic analyses Maximum likelihood analysis was performed using the program RAxML ver. 7.3.0 (Stamatakis, 2006; Stamatakis, Hoover & Rougemont, 2008) and was run remotely at the CIPRES Science Gateway (Miller, Pfeiffer & Schwartz, 2010; https://www.phylo.org/). Ribosomal genes (12S, 16S, 18S and 28S) were each treated as a single partition. Three different partition schemes were implemented for the protein-coding genes (COI and H3), namely unpartitioned (each gene a single partition), 3rd codon (two partitions: 1st + 2nd codons, and 3rd codon) and codon specific (one partition for each codon position). The partition corresponding to the gaps scored as absence/presence characters was treated as binary. Each gene partition was assigned an unlinked general time-reversible model with gamma distributed among invariable sites (GTR + Γ + I). The best likelihood tree was obtained out of 100 random iterations, and support was assessed by conducting 1000 non-parametric bootstrap replicates for each analysis. Both independent gene trees and concatenated analyses were run under the different partition schemes. Bayesian analyses were performed using MrBayes 3.2.6 (Ronquist & Huelsenbeck, 2003), and the program jModeltest ver. 2.1.6 (Guindon & Gascuel, 2003; Darriba et al., 2012) was used to select the best-fitting model of evolution for each partition using Akaike’s information criterion (Akaike, 1973; Buckley et al., 2002). Both programs were run remotely at the CIPRES Science Gateway. Partition schemes were implemented as in the previous paragraph and evaluated using Bayes factors (BF) (Brown & Lemmon, 2007). Acceptance or rejection of each strategy was based on the following cut-off: BF ≥ 10 (strong evidence against the competing hypothesis), 10 < BF ≥ −10 (ambiguous, select least complex strategy) and BF ≤ −10 (strong evidence for the competing hypothesis). Recoded gaps were treated with a standard discrete model. The substitution estimates were allowed to vary independently between each partition. For phylogenetic analysis, two independent runs with four simultaneous Markov chain Monte Carlo chains (one cold and three heated), each with random starting trees, were conducted simultaneously, sampling every 1000 generations until the SD of the split frequencies of these two runs dropped below 0.01 (107 generations). The program Tracer ver. 1.5 (Rambaut & Drummund, 2009) was used to ensure that the Markov chains had reached stationarity by examining the effective sample size values (> 200) and to determine the correct number of generations to discard as a burn-in (first 10% of generations). Parsimony analyses were performed using TNT version 1.5 (Goloboff, Farris & Nixon, 2008a; Goloboff & Catalano, 2016). These data were analysed under both equal and implied weights (Goloboff, 1993) with concavity constants of the weighting function values of 6, 20, 50, 100, 500 and 1000, as suggested by Goloboff et al. (2008b). All tree searches were driven to independently hit 15 times the optimal scoring, using the default values of the ‘New Technologies’ search in TNT with sectorial searches, tree fusing and ratchet, followed by TBR branch swapping (string of commands hold 20000; xmult = hits 15 ratchet 10; bb = fillonly;). Support values were estimated by jackknifing (J) frequencies; each of the 1000 pseudoreplicates used three random additional sequences plus TBR, followed by TBR collapsing to calculate the consensus. Trees were edited with the program FigTree ver. 1.4.3 (Rambaut, 2006–2016) and WinClada. All analyses were rooted with the branch separating H. pococki and Filistatidae from the remaining taxa. The data sets can be obtained in TreeBase with the accession number 21309. Morphological data The examined material is deposited in the following institutions (abbreviation and curators in parentheses): American Museum of Natural History, New York, NY, USA (AMNH, L. Prendini); California Academy of Sciences, San Francisco, CA, USA (CAS, L. Esposito), Field Museum of Natural History, Chicago, IL, USA (FMNH, P. Sierwald); Muséum National d’Histoire Naturelle, Paris, France (MNHN, C. Rollard); National Collection of Arachnida, Plant Protection Research Institute, Pretoria, South Africa (NCP, A. Dippenaar-Schoeman, P. Marais and R. Lyle); National Museum, Bloemfontein (NMBA, L. Lotz); Natal Museum, Pietermaritzburg, Kwa-Zulu Natal, South Africa (NMSA, D. Herbert, A. Ndaba and C. Stoffels); and Iziko Museum of Capetown (formerly South African Museum), Cape Town, South Africa (SAMC, S. van Noort and D. Larsen). Morphological observations were made using Leica M165 C or M125 stereomicroscopes and Olympus BH-2 or Leica DM4000 M compound microscopes. Light micrographs were prepared using a Nikon DXM1200 digital camera mounted on a Nikon SMZ1500 or Leica MZ 16 stereomicroscope, a Leica DFC 500 digital camera mounted on a Leica M216 and M165 C stereomicroscope or a Leica DM 4000 compound microscope. Extended focal range images were composed with Leica Application Suite version 3.6.0. and Helicon Focus 3.10.3, 4.01 Pro and 4.62 Pro (Khmelik, Kozub & Glazunov, 2006). Female vulva illustrations were drawn in Adobe Illustrator (Adobe Systems) over digital images using a Wacom drawing table (Wacom Co. Ltd.). The male pedipalp was checked for haematodochal expansion by placing it in a vial of hot 92% lactic acid solution (Sigma–Aldrich, St Louis, MO, USA) that was placed into boiling water for 2–3 min, after which the pedipalp was transferred to distilled water (Ledford et al., 2011; Griswold, Audisio & Ledford, 2012). The internal anatomy of the pedipalpal bulb and course of the reservoir were examined by immersing the bulb in a solution of methyl salicylate (synthetic oil of wintergreen), from Mallunckrodt Baker Chemicals, Philipsburg, NJ, USA (Miller, Griswold & Yin, 2009). Preparations were carefully cleaned using fine brushes, a thin jet of alcohol from a thinned pipette or in an ultrasonic cleaner; some setae were removed to expose structures, especially those on the legs and pedipalp, spinnerets and chelicerae. For scanning electron microscopy (SEM), all preparations were dehydrated in a series of increasing concentrations of ethanol (80, 90, 95 and 100%) and critical point dried. After drying and brushing, they were mounted on adhesive copper tape (Electron Microscopy Sciences, EMS 77802) affixed to a stub and secured with a conductive paint of colloidal graphite on isopropyl alcohol base (EMS 12660). Prior to SEM examination under high vacuum with an FEI XL30 TMP or Leo 1450VP Scanning Electron Microscope, the structures were sputter coated with Au-Pd. Measurements were taken with a micrometric ocular and are given in millimetres. Female genitalia were digested after dissection with KOH, pancreatin (Álvarez-Padilla & Hormiga, 2008) or contact lens cleaner solution (ReNu) and observed in clove oil or lactic acid. Coloration patterns are described based on specimens preserved in 80% ethanol. Reciprocal species diagnoses were made between sister species when possible. Discussed morphological characters are used ad hoc and were not included in the phylogenetic analyses. Coordinates given in brackets were obtained with Google Earth and The World Coordinate Converter (http://twcc.free.fr/) and those written in italic should be considered broad approximations. Cape Town’s cave coordinates were obtained at geoview.info (http://za.geoview.info/), and information regarding caves’ names obtained at Cape Peninsula Speleological Society web page (http://cpss.caving.org.za/) and from Craven (2012). RESULTS A summary of results, statistics and conditions for each phylogenetic analysis is presented in Table 2. We analysed the gene fragments separately and in combination. Results from the different phylogenetic analyses are summarized in the form of a preferred working tree (Fig. 2; see Congruence and selection of working tree). We focus our discussion on the groups with higher support, hereby defined as those with a bootstrap proportions (BS) > 0.75, Bayesian posterior probability (PP) > 0.95 and jackknifing proportions > 0.75 (denoted with black in circles in Fig. 2 and Supporting Information, Fig. S1), and also on clades that were with low support but were consistently recovered in likelihood, Bayesian and parsimony analyses (grey circles in Fig. 2 and Supporting Information, Fig. S1). Table 2. Summary tree statistics and conditions for each analysis Analysis Optimality criterion, Software Conditions Statistics COI Parsimony, TNT v1.5 EW 24 trees, 6431 steps Histone 3 EW 8 trees, 1788 steps 12S rDNA + gaps EW 11 trees, 4838 steps 16S rDNA + gaps EW 56 trees, 5981 steps 18S rDNA + gaps EW 138 trees, 2783 steps 28S rDNA + gaps EW 10 trees, 8783 steps Six-genes concatenated + gaps EW 36 trees, 31532 steps IW k = 6 1 tree, 31637 steps, fit 1343.78383 IW k = 20 1 tree, 31573 steps, fit 720.10758 IW k = 50 1 tree, 31539 steps, fit 374.70066 IW k = 100 1 tree, 31534 steps, fit 210.02483 IW k = 500 1 tree, 31532 steps, fit 46.75595 IW k = 1000 1 tree, 31532 steps, fit 23.72344 Reduce six-genes concatenated + gaps 84 trees, 29 373 steps COI (unpartitioned) Likelihood, RAxML v7.3.0 1000 non-parametric bootstrap replicates −lnL 24507.664201 COI (3rd codon) −lnL 24018.961966 COI (codon specific) −lnL 23605.298821 Histone 3 (unpartitioned) −lnL 7172.273998 Histone 3 (3rd codon) −lnL 6923.293562 Histone 3 (codon specific) −lnL 6838.500314 12S rDNA + gaps −lnL 17083.095847 16S rDNA + gaps −lnL 22180.794164 18S rDNA + gaps −lnL 16446.281902 28S rDNA + gaps −lnL 40176.652458 Six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 132036.031678 Reduce six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 124 410.091077 COI (unpartitioned) Bayesian, Mr. Bayes v.3.2.6 50000000 generations, burn-in = 10% sdsf 0.006877 COI (3rd codon) sdsf 0.007477 COI (codon specific) sdsf 0.010772 Histone 3 (unpartitioned) sdsf 0.004276 Histone 3 (3rd codon) sdsf 0.003138 Histone 3 (codon specific) sdsf 0.005141 12S rDNA + gaps sdsf 0.005199 16S rDNA + gaps sdsf 0.003927 18S rDNA + gaps sdsf 0.007870 28S rDNA + gaps sdsf 0.023531 Six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps 30000000 generations, burn-in = 10% sdsf 0.030925 Reduce six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps sdsf 0.001968 Analysis Optimality criterion, Software Conditions Statistics COI Parsimony, TNT v1.5 EW 24 trees, 6431 steps Histone 3 EW 8 trees, 1788 steps 12S rDNA + gaps EW 11 trees, 4838 steps 16S rDNA + gaps EW 56 trees, 5981 steps 18S rDNA + gaps EW 138 trees, 2783 steps 28S rDNA + gaps EW 10 trees, 8783 steps Six-genes concatenated + gaps EW 36 trees, 31532 steps IW k = 6 1 tree, 31637 steps, fit 1343.78383 IW k = 20 1 tree, 31573 steps, fit 720.10758 IW k = 50 1 tree, 31539 steps, fit 374.70066 IW k = 100 1 tree, 31534 steps, fit 210.02483 IW k = 500 1 tree, 31532 steps, fit 46.75595 IW k = 1000 1 tree, 31532 steps, fit 23.72344 Reduce six-genes concatenated + gaps 84 trees, 29 373 steps COI (unpartitioned) Likelihood, RAxML v7.3.0 1000 non-parametric bootstrap replicates −lnL 24507.664201 COI (3rd codon) −lnL 24018.961966 COI (codon specific) −lnL 23605.298821 Histone 3 (unpartitioned) −lnL 7172.273998 Histone 3 (3rd codon) −lnL 6923.293562 Histone 3 (codon specific) −lnL 6838.500314 12S rDNA + gaps −lnL 17083.095847 16S rDNA + gaps −lnL 22180.794164 18S rDNA + gaps −lnL 16446.281902 28S rDNA + gaps −lnL 40176.652458 Six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 132036.031678 Reduce six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 124 410.091077 COI (unpartitioned) Bayesian, Mr. Bayes v.3.2.6 50000000 generations, burn-in = 10% sdsf 0.006877 COI (3rd codon) sdsf 0.007477 COI (codon specific) sdsf 0.010772 Histone 3 (unpartitioned) sdsf 0.004276 Histone 3 (3rd codon) sdsf 0.003138 Histone 3 (codon specific) sdsf 0.005141 12S rDNA + gaps sdsf 0.005199 16S rDNA + gaps sdsf 0.003927 18S rDNA + gaps sdsf 0.007870 28S rDNA + gaps sdsf 0.023531 Six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps 30000000 generations, burn-in = 10% sdsf 0.030925 Reduce six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps sdsf 0.001968 The preferred analyses for each optimality criteria are given in bold. COI, mitochondrial cytochrome c oxidase subunit I; EW, equal weights; Gaps, number of gap absence/presence characters added; IW, implied weights; k, constant of cavity. View Large Table 2. Summary tree statistics and conditions for each analysis Analysis Optimality criterion, Software Conditions Statistics COI Parsimony, TNT v1.5 EW 24 trees, 6431 steps Histone 3 EW 8 trees, 1788 steps 12S rDNA + gaps EW 11 trees, 4838 steps 16S rDNA + gaps EW 56 trees, 5981 steps 18S rDNA + gaps EW 138 trees, 2783 steps 28S rDNA + gaps EW 10 trees, 8783 steps Six-genes concatenated + gaps EW 36 trees, 31532 steps IW k = 6 1 tree, 31637 steps, fit 1343.78383 IW k = 20 1 tree, 31573 steps, fit 720.10758 IW k = 50 1 tree, 31539 steps, fit 374.70066 IW k = 100 1 tree, 31534 steps, fit 210.02483 IW k = 500 1 tree, 31532 steps, fit 46.75595 IW k = 1000 1 tree, 31532 steps, fit 23.72344 Reduce six-genes concatenated + gaps 84 trees, 29 373 steps COI (unpartitioned) Likelihood, RAxML v7.3.0 1000 non-parametric bootstrap replicates −lnL 24507.664201 COI (3rd codon) −lnL 24018.961966 COI (codon specific) −lnL 23605.298821 Histone 3 (unpartitioned) −lnL 7172.273998 Histone 3 (3rd codon) −lnL 6923.293562 Histone 3 (codon specific) −lnL 6838.500314 12S rDNA + gaps −lnL 17083.095847 16S rDNA + gaps −lnL 22180.794164 18S rDNA + gaps −lnL 16446.281902 28S rDNA + gaps −lnL 40176.652458 Six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 132036.031678 Reduce six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 124 410.091077 COI (unpartitioned) Bayesian, Mr. Bayes v.3.2.6 50000000 generations, burn-in = 10% sdsf 0.006877 COI (3rd codon) sdsf 0.007477 COI (codon specific) sdsf 0.010772 Histone 3 (unpartitioned) sdsf 0.004276 Histone 3 (3rd codon) sdsf 0.003138 Histone 3 (codon specific) sdsf 0.005141 12S rDNA + gaps sdsf 0.005199 16S rDNA + gaps sdsf 0.003927 18S rDNA + gaps sdsf 0.007870 28S rDNA + gaps sdsf 0.023531 Six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps 30000000 generations, burn-in = 10% sdsf 0.030925 Reduce six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps sdsf 0.001968 Analysis Optimality criterion, Software Conditions Statistics COI Parsimony, TNT v1.5 EW 24 trees, 6431 steps Histone 3 EW 8 trees, 1788 steps 12S rDNA + gaps EW 11 trees, 4838 steps 16S rDNA + gaps EW 56 trees, 5981 steps 18S rDNA + gaps EW 138 trees, 2783 steps 28S rDNA + gaps EW 10 trees, 8783 steps Six-genes concatenated + gaps EW 36 trees, 31532 steps IW k = 6 1 tree, 31637 steps, fit 1343.78383 IW k = 20 1 tree, 31573 steps, fit 720.10758 IW k = 50 1 tree, 31539 steps, fit 374.70066 IW k = 100 1 tree, 31534 steps, fit 210.02483 IW k = 500 1 tree, 31532 steps, fit 46.75595 IW k = 1000 1 tree, 31532 steps, fit 23.72344 Reduce six-genes concatenated + gaps 84 trees, 29 373 steps COI (unpartitioned) Likelihood, RAxML v7.3.0 1000 non-parametric bootstrap replicates −lnL 24507.664201 COI (3rd codon) −lnL 24018.961966 COI (codon specific) −lnL 23605.298821 Histone 3 (unpartitioned) −lnL 7172.273998 Histone 3 (3rd codon) −lnL 6923.293562 Histone 3 (codon specific) −lnL 6838.500314 12S rDNA + gaps −lnL 17083.095847 16S rDNA + gaps −lnL 22180.794164 18S rDNA + gaps −lnL 16446.281902 28S rDNA + gaps −lnL 40176.652458 Six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 132036.031678 Reduce six-gene concatenated + COI–H3 (codon specific) + gaps −lnL 124 410.091077 COI (unpartitioned) Bayesian, Mr. Bayes v.3.2.6 50000000 generations, burn-in = 10% sdsf 0.006877 COI (3rd codon) sdsf 0.007477 COI (codon specific) sdsf 0.010772 Histone 3 (unpartitioned) sdsf 0.004276 Histone 3 (3rd codon) sdsf 0.003138 Histone 3 (codon specific) sdsf 0.005141 12S rDNA + gaps sdsf 0.005199 16S rDNA + gaps sdsf 0.003927 18S rDNA + gaps sdsf 0.007870 28S rDNA + gaps sdsf 0.023531 Six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps 30000000 generations, burn-in = 10% sdsf 0.030925 Reduce six-gene concatenated + COI (3rd codon) + H3 (unpartitioned) + gaps sdsf 0.001968 The preferred analyses for each optimality criteria are given in bold. COI, mitochondrial cytochrome c oxidase subunit I; EW, equal weights; Gaps, number of gap absence/presence characters added; IW, implied weights; k, constant of cavity. View Large Figure 2. View largeDownload slide Preferred tree obtained in the maximum likelihood analysis of the combined molecular alignments excluding ambiguously aligned regions, with the preferred model for each gene fragment and gaps scored as absence/presence. The information of the results of the analyses under Bayesian (first 10% of generations were removed as burn-in) and parsimony (equal weights) are summarized in the node support. Circles at nodes indicate support levels subdivided by analysis (L, maximum likelihood; B, Bayesian; P, parsimony: bottom; see inset). Black indicates bootstrap proportions > 0.75, posterior probabilities > 0.95 and jackknifing proportions > 0.75; grey indicates that the clade was recovered but with lower support than the previous values; white indicates that the clade was not recovered. Supported families and genera discussed in the text are highlighted with lateral bars and shaded boxes, respectively. Figure 2. View largeDownload slide Preferred tree obtained in the maximum likelihood analysis of the combined molecular alignments excluding ambiguously aligned regions, with the preferred model for each gene fragment and gaps scored as absence/presence. The information of the results of the analyses under Bayesian (first 10% of generations were removed as burn-in) and parsimony (equal weights) are summarized in the node support. Circles at nodes indicate support levels subdivided by analysis (L, maximum likelihood; B, Bayesian; P, parsimony: bottom; see inset). Black indicates bootstrap proportions > 0.75, posterior probabilities > 0.95 and jackknifing proportions > 0.75; grey indicates that the clade was recovered but with lower support than the previous values; white indicates that the clade was not recovered. Supported families and genera discussed in the text are highlighted with lateral bars and shaded boxes, respectively. New molecular data The molecular data set includes six species sequenced in the present study, yielding 18 sequences all available in GenBank (Supporting Information, Table S1): Drymusa serrana Goloboff & Ramírez, 1992 (H3 MF49776, COI MF497758), D. armasi (H3 MF564217, 28S MF506838, 18S MF506830), Izithunzi silvicola comb. nov. (Purcell, 1904) (H3 MF497760, COI MF497757, 28S MF506837, 18S MF506834), I. capense comb. nov. (H3 MF497759, COI MF497755, 28S MF506835, 18S MF506832), I. lina sp. nov. (COI MF497754, 28S MF506836, 18S MF506831) and I. productum comb. nov. (COI MF497756, 18S MF506833). Analysis of molecular data The alignment of the protein-coding genes COI and H3 was trivial because they show no evidence of indel mutations. For the 12S and 18S ribosomal genes, the alignment with the parameter value GOP = 1.53 minimized RILD when combined with the unambiguously aligned protein-coding genes, whereas for the 16S and 28S ribosomal genes, the preferred parameter value was GOP = 1 (Table 3). The total length of the concatenated aligned data matrix was 7630 characters, of which 6790 corresponded to nucleotides (1078 bp of COI, 320 bp of H3, 401 bp of 12S, 532 bp of 16S, 1795 bp of 18S, 2664 bp of 28S) and 840 to gaps coded as absence/presence (111 from 12S, 150 from 16S, 87 from 18S, 492 from 28S). Table 3. Summary of the results of the parsimony analysis of the alignments obtained with MAFFT under different parameter combinations and similarity values with the results of the combined data matrix (ribosomal gene + gaps + protein-coding genes) GOP T L Gaps ∑ Lc ILD MaxLc RILD 12S 1.00 4 4813 112 13033 13305 272 18603 0.0488330341 1.53 (default) 3 4838 111 13 058 13 322 264 18 635 0.0473372781 3.00 4 4864 85 13084 13349 265 18664 0.0474910394 16S 1.00 9 5981 150 14 201 14 444 243 20 667 0.0375811939 1.53 1 5998 126 14218 14468 250 20712 0.0384970742 3.00 3 6106 105 14326 14580 254 20882 0.0387431361 18S 1.00 15 2771 90 10991 11251 260 16292 0.0490473496 1.53 22 2783 87 11 003 11 259 256 16 317 0.048174633 3.00 25 2793 89 11013 11271 258 16338 0.0484507042 28S 1.00 5 8783 492 17 003 17 369 366 25 657 0.0422925815 1.53 2 8912 464 17132 17509 377 25835 0.043318396 3.00 11 9029 423 17249 17625 376 26090 0.0425291257 GOP T L Gaps ∑ Lc ILD MaxLc RILD 12S 1.00 4 4813 112 13033 13305 272 18603 0.0488330341 1.53 (default) 3 4838 111 13 058 13 322 264 18 635 0.0473372781 3.00 4 4864 85 13084 13349 265 18664 0.0474910394 16S 1.00 9 5981 150 14 201 14 444 243 20 667 0.0375811939 1.53 1 5998 126 14218 14468 250 20712 0.0384970742 3.00 3 6106 105 14326 14580 254 20882 0.0387431361 18S 1.00 15 2771 90 10991 11251 260 16292 0.0490473496 1.53 22 2783 87 11 003 11 259 256 16 317 0.048174633 3.00 25 2793 89 11013 11271 258 16338 0.0484507042 28S 1.00 5 8783 492 17 003 17 369 366 25 657 0.0422925815 1.53 2 8912 464 17132 17509 377 25835 0.043318396 3.00 11 9029 423 17249 17625 376 26090 0.0425291257 The preferred combinations are given in bold. 12S, mitochondrial 12S rDNA; 16S, 16S rDNA; 18S, nuclear 18S rDNA; 28S, 28S rDNA; GOP, gap opening cost; T, number of most parsimonious trees; L, tree length; gaps, number of gap absence/presence characters added; ∑, summary of the tree length of each partition; Lc, tree length of the combined data matrix; ILD, incongruence length difference; MaxLc, maximum tree length of the combined data matrix; RILD, rescaled incongruence length difference. Table 3. Summary of the results of the parsimony analysis of the alignments obtained with MAFFT under different parameter combinations and similarity values with the results of the combined data matrix (ribosomal gene + gaps + protein-coding genes) GOP T L Gaps ∑ Lc ILD MaxLc RILD 12S 1.00 4 4813 112 13033 13305 272 18603 0.0488330341 1.53 (default) 3 4838 111 13 058 13 322 264 18 635 0.0473372781 3.00 4 4864 85 13084 13349 265 18664 0.0474910394 16S 1.00 9 5981 150 14 201 14 444 243 20 667 0.0375811939 1.53 1 5998 126 14218 14468 250 20712 0.0384970742 3.00 3 6106 105 14326 14580 254 20882 0.0387431361 18S 1.00 15 2771 90 10991 11251 260 16292 0.0490473496 1.53 22 2783 87 11 003 11 259 256 16 317 0.048174633 3.00 25 2793 89 11013 11271 258 16338 0.0484507042 28S 1.00 5 8783 492 17 003 17 369 366 25 657 0.0422925815 1.53 2 8912 464 17132 17509 377 25835 0.043318396 3.00 11 9029 423 17249 17625 376 26090 0.0425291257 GOP T L Gaps ∑ Lc ILD MaxLc RILD 12S 1.00 4 4813 112 13033 13305 272 18603 0.0488330341 1.53 (default) 3 4838 111 13 058 13 322 264 18 635 0.0473372781 3.00 4 4864 85 13084 13349 265 18664 0.0474910394 16S 1.00 9 5981 150 14 201 14 444 243 20 667 0.0375811939 1.53 1 5998 126 14218 14468 250 20712 0.0384970742 3.00 3 6106 105 14326 14580 254 20882 0.0387431361 18S 1.00 15 2771 90 10991 11251 260 16292 0.0490473496 1.53 22 2783 87 11 003 11 259 256 16 317 0.048174633 3.00 25 2793 89 11013 11271 258 16338 0.0484507042 28S 1.00 5 8783 492 17 003 17 369 366 25 657 0.0422925815 1.53 2 8912 464 17132 17509 377 25835 0.043318396 3.00 11 9029 423 17249 17625 376 26090 0.0425291257 The preferred combinations are given in bold. 12S, mitochondrial 12S rDNA; 16S, 16S rDNA; 18S, nuclear 18S rDNA; 28S, 28S rDNA; GOP, gap opening cost; T, number of most parsimonious trees; L, tree length; gaps, number of gap absence/presence characters added; ∑, summary of the tree length of each partition; Lc, tree length of the combined data matrix; ILD, incongruence length difference; MaxLc, maximum tree length of the combined data matrix; RILD, rescaled incongruence length difference. The maximum likelihood analysis of the alternative partition schemes revealed better likelihoods for codon-specific partitions for both the COI and H3 genes (Table 2). Maximum likelihood analyses of the complete concatenated data matrix with a/p gaps (C-ML) yielded one single tree with a score of −lnL = −132036.032 (Table 2; Supporting Information, Fig. S1). Summary results of the evolutionary models selected by Akaike’s information criterion in JModeltest are presented in Table 4 and for BF partitioning strategies in Table 5. Comparison of the alternative partitioning strategies revealed a strong preference for the 3rd codon partition in COI and a strong preference for the unpartitioned partition in H3 (Table 5). Bayesian analysis of independent genes achieved convergence within 5 × 107 generations, while that of concatenated genes achieved convergence within 3 × 107 generations (i.e. due to the matrix size) (mostly all SD of split frequencies < 0.01; see Table 2). The Bayesian results of the complete concatenated matrix (C-BI) mostly differ from the maximum likelihood results in the basal nodes of the clade uniting Scytodoidea and among the Lost Tracheal clade members (clade A; Supporting Information, Fig. S1). Table 4. Evolutionary models used for Bayesian analysis as selected by AIC in Jmodeltest ver. 2.1.6 Partition Model selected AIC score Model used COI GTR + I + G 24120.0236 GTR + I + G H3 TVM + I + G 7035.7133 GTR + I + G 12S TIM2 + I + G 15799.2414 GTR + I + G 16S TVM + I + G 20417.1281 HKY + I + G 18S TIM2 + I + G 15592.0110 GTR + I + G 28S GTR + I + G 35977.6747 GTR + I + G Partition Model selected AIC score Model used COI GTR + I + G 24120.0236 GTR + I + G H3 TVM + I + G 7035.7133 GTR + I + G 12S TIM2 + I + G 15799.2414 GTR + I + G 16S TVM + I + G 20417.1281 HKY + I + G 18S TIM2 + I + G 15592.0110 GTR + I + G 28S GTR + I + G 35977.6747 GTR + I + G COI, mitochondrial cytochrome c oxidase subunit I; 12S, mitochondrial 12S rDNA; 16S, mitochondrial 16S rDNA; 18S, nuclear 18S rDNA; 28S, nuclear 28S rDNA; H3, nuclear histone 3; AIC, Akaike information criterion. Table 4. Evolutionary models used for Bayesian analysis as selected by AIC in Jmodeltest ver. 2.1.6 Partition Model selected AIC score Model used COI GTR + I + G 24120.0236 GTR + I + G H3 TVM + I + G 7035.7133 GTR + I + G 12S TIM2 + I + G 15799.2414 GTR + I + G 16S TVM + I + G 20417.1281 HKY + I + G 18S TIM2 + I + G 15592.0110 GTR + I + G 28S GTR + I + G 35977.6747 GTR + I + G Partition Model selected AIC score Model used COI GTR + I + G 24120.0236 GTR + I + G H3 TVM + I + G 7035.7133 GTR + I + G 12S TIM2 + I + G 15799.2414 GTR + I + G 16S TVM + I + G 20417.1281 HKY + I + G 18S TIM2 + I + G 15592.0110 GTR + I + G 28S GTR + I + G 35977.6747 GTR + I + G COI, mitochondrial cytochrome c oxidase subunit I; 12S, mitochondrial 12S rDNA; 16S, mitochondrial 16S rDNA; 18S, nuclear 18S rDNA; 28S, nuclear 28S rDNA; H3, nuclear histone 3; AIC, Akaike information criterion. Table 5. Results of Bayes factor hypothesis testing for partitioning strategies Partition −lnL harmonic mean (post burn-in) Bayes factor Support for rejection COI (codon specific) −23331.07 – – COI (3rd codon) −22758.04 −573.03 None COI (unpartitioned) −22894.01 −437.06 None H3 (codon specific) −6925.449 – – H3 (3rd codon) −6862.995 −62.454 None H3 (unpartitioned) −6679.128 −246.321 None Partition −lnL harmonic mean (post burn-in) Bayes factor Support for rejection COI (codon specific) −23331.07 – – COI (3rd codon) −22758.04 −573.03 None COI (unpartitioned) −22894.01 −437.06 None H3 (codon specific) −6925.449 – – H3 (3rd codon) −6862.995 −62.454 None H3 (unpartitioned) −6679.128 −246.321 None The preferred hypotheses are given in bold. COI, mitochondrial cytochrome c oxidase subunit I; H3, nuclear histone 3. Table 5. Results of Bayes factor hypothesis testing for partitioning strategies Partition −lnL harmonic mean (post burn-in) Bayes factor Support for rejection COI (codon specific) −23331.07 – – COI (3rd codon) −22758.04 −573.03 None COI (unpartitioned) −22894.01 −437.06 None H3 (codon specific) −6925.449 – – H3 (3rd codon) −6862.995 −62.454 None H3 (unpartitioned) −6679.128 −246.321 None Partition −lnL harmonic mean (post burn-in) Bayes factor Support for rejection COI (codon specific) −23331.07 – – COI (3rd codon) −22758.04 −573.03 None COI (unpartitioned) −22894.01 −437.06 None H3 (codon specific) −6925.449 – – H3 (3rd codon) −6862.995 −62.454 None H3 (unpartitioned) −6679.128 −246.321 None The preferred hypotheses are given in bold. COI, mitochondrial cytochrome c oxidase subunit I; H3, nuclear histone 3. In the parsimony analyses of the complete data matrix, the heuristic search strategies under equal weights (C-EW) produced 36 trees of length 31532. As expected, milder down-weighting against homoplasy (higher k values) resulted in topologies more similar to equally weighted trees. The k = 1000 and 500 were identical to each other and to one of the 36 equal weighted trees, but the topology of the k = 6, 20, 50 and 100 analyses was slightly different in the resolution of a few groups with low bootstrap values. Generally, the parsimony results of the concatenated molecular matrix mostly differ from the maximum likelihood analysis regarding nodes with low support, especially in clade A (Supporting Information, Fig. S1). Congruence and selection of working tree To explore congruence and sensitivity among all the trees produced by different optimality criteria, we summarized our results in the maximum likelihood tree obtained from the concatenated alignment matrix as ‘Navajo rugs’ (Supporting Information, Fig. S2; Giribet & Edgecomb, 2006). These analyses indicated that no individual gene reliably anticipates the results from the concatenated alignment. As expected, the incongruent nodes also showed low support (Supporting Information, Figs S1, S2; Rindal & Brower, 2011). Five terminals were very unstable across analyses, and few had very long branches within clade A: D. armasi, cf. Psiloderces sp. MA186 (Ochyroceratidae Fage, 1912), Theotima sp. MR15 (Ochyroceratidae), Stedocys pagodas Labarque, Grismado, Ramírez, Yan & Griswold, 2009 (Scytodidae) and Usofila sp. MR71 (Telemidae Fage, 1913). The removal of these five terminals from the data set and then running various phylogenetic algorithms (e.g. maximum likelihood, Bayesian, parsimony) produce an increment in the support values within clade A. Furthermore, these results also recovered Scytodoidea and the Lost Tracheal Clade as monophyletic, contrary to the results of the complete data set (compare Fig. 2 against Supporting Information, Fig. S1). Therefore, we present these analyses of 103 terminals as our preferred hypothesis, which are summarized in a maximum likelihood tree (Fig. 2). Our working tree reflects that of Wheeler et al. (2016: fig. 3), which mostly agrees with the taxonomic and transcriptomic congruences computed in that work. Notably, our preferred tree results from unconstrainted analyses, contrary to Wheeler et al. (2016: fig. 3; although their unconstrained and constrained maximum likelihood trees were very similar). Maximum likelihood analysis of the reduced, concatenated data matrix with a/p gaps (R-ML) yielded one single tree with a score of −lnL = −124410.091 (Fig. 2; Table 2). Bayesian analysis of the reduced, concatenated genes (R-BI) achieved convergence within 3 × 107 generations (SD of split frequencies < 0.01). In the parsimony analysis of the reduced, concatenated matrix, the heuristic search strategies under equal weights (R-EW) produced 84 trees of length 29373. DISCUSSION Synspermiata relationships and clade support Our preferred tree topology (Fig. 2) mostly resembles that of Wheeler et al. (2016: fig. 3), except for a few differences. The R-ML and R-EW analyses recovered Trogloraptoridae Griswold, Audosio & Ledford, 2012 and Caponiidae Simon, 1890, as sister groups, but both families emerged as sister to the other members of all Synspermiata instead of only Dysderoidea (Fig. 2; Wheeler et al., 2016: fig. 3). Furthermore, in our C-ML and C-EW analyses, Caponiidae emerged as sister to the rest of Synspermiata, while Trogloraptoridae were recovered as sister to Dysderoidea (Fig. 2; although in the R-BI and C-BI analyses Caponiidae go to an odd place, as sister to Filistatidae with PP = 0.93). Most analyses (except R-BI and C-BI) recovered Oonopidae Simon, 1890 and Dysderidae C. L. Koch, 1837, as sister groups (Fig. 2), contrary to Wheeler et al. (2016), which showed a close relationship between Orsolobidae Cooke, 1965 and Dysderidae. However, these relationships have low support in both cladistic analyses (Fig. 2; Wheeler et al., 2016: fig. 3). Orchestina Simon, 1882 was recovered in all our analyses, but it was not resolved as monophyletic in Wheeler et al. (2016). Our working tree also yielded higher average support values than that of Wheeler et al. (2016: fig. 3), highly supporting several clades only recovered (i.e. with low support) by the former. Oonopidae and the informal higher gamasomorphines (following Grismado, Ramírez & Izquierdo, 2014) were supported by most of our analyses, except by parsimony (i.e. R-EW and C-EW). The clade including Tetrablemmidae O. Pickard-Cambridge, 1873, Plectreuridae Simon, 1893, Diguetidae, F. O. Pickard-Cambridge, 1899, and Pacullidae Simon, 1893, was also supported by most of our analyses, except by the R-BI. All reduced data set analyses supported Ochyroceratidae and Scytodidae (Fig. 2). Synspermiata was supported by the R-ML and C-ML analyses (although Caponiidae were recovered as the sister of Filistatidae in the R-BI and C-BI analyses, see previous paragraph). The R-ML and R-BI analyses supported the genus Sicarius Walckenaer, 1847, and the group formed by Scytodoidea and the Lost Tracheal clade (Fig. 2: clade A). Scytodoidea were supported by the R-BI analysis. The clade including Plectreuridae, Diguetidae and Pacullidae was recovered by our results (i.e. R-ML and R-BI) and by Wheeler et al. (2016: fig. 3). Ochyroceratidae as member of Scytodoidea Ochyroceratidae and Scytodidae were recovered as sister groups, and each family is well supported by R-ML, R-BI and R-EW analyses (Fig. 2). Note that the low support values obtained for each family in Wheeler et al. (2016: fig. 3 – Ochyroceratidae BS = 55, Scytodidae BS = 39) were due to the use of inconsistent terminals in their analyses (i.e. cf. Psiloderces sp. and S. pagodas). Our hypothesis (Ochyroceratidae plus Scytodidae) agrees with two morphological characters including bipectinated proclaws on tarsi I–II (Lehtinen, 1986; Labarque & Ramírez, 2012; Pérez-González, Rubio & Ramírez, 2016: fig. 8G) and a distal hood on podotarsite (Labarque et al., 2017); characters also shared by Drymusidae and Periegopidae (Fig. 3A, B; Labarque & Ramírez, 2012; Labarque et al., 2017). However, Ochyroceratidae lack several of the features considered to be Scytodoidea morphological synapomorphies, including Non-homoplastic: (1) tracheae with fused third opisthosomal apodemes (Fig. 3C; Ramírez, 2000: ch. 32; Labarque & Ramírez, 2012: ch. 53); (2) an enlarged subchromosomal space (Michalik & Ramírez, 2014: ch. 11); Homoplastic: (3) the cheliceral promarginal lobe (Fig. 3D; Labarque & Ramírez, 2012: ch. 28); (4) female cheliceral stridulatory ridges and pedipalpal femoral thorns (Fig. 3E, F; Labarque & Ramírez, 2012: ch. 32); (5) a conical acrosomal vacuole (Michalik & Ramírez, 2014: ch. 8); (6) the ends of the acrosomal filament clearly after the axonemal basis (Michalik & Ramírez, 2014: ch. 12); and (7) the indented anterior pole of the nucleus (Michalik & Ramírez, 2014: ch. 14). Figure 3. View largeDownload slide Scytodoidea and Drymusidae plus Periegopidae putative synapomorphies. Izithunzi capense comb. nov. (♀, A–C CASENT 9023625), Izithunzi silvicola comb. nov. (♀, D–F CASENT-9048599), preserved specimens. A, B, right feet I (C, retrolateral; D, apical). C, tracheae, dorsal. D–E, left chelicerae detail [D, promargin; E, ectal (corner, shallow scales, detail)]. F, left pedipalp, prolateral, tarsal tip (*, curved setae; corner, femoral thorn, detail). Abbreviations: ant., anterior; pro., prolateral. Scale bar: C, E, 0.2 mm; A, B, 0.1 mm; D, F, 0.05 mm. Figure 3. View largeDownload slide Scytodoidea and Drymusidae plus Periegopidae putative synapomorphies. Izithunzi capense comb. nov. (♀, A–C CASENT 9023625), Izithunzi silvicola comb. nov. (♀, D–F CASENT-9048599), preserved specimens. A, B, right feet I (C, retrolateral; D, apical). C, tracheae, dorsal. D–E, left chelicerae detail [D, promargin; E, ectal (corner, shallow scales, detail)]. F, left pedipalp, prolateral, tarsal tip (*, curved setae; corner, femoral thorn, detail). Abbreviations: ant., anterior; pro., prolateral. Scale bar: C, E, 0.2 mm; A, B, 0.1 mm; D, F, 0.05 mm. Drymusidae monophyly Drymusidae and Periegopidae were recovered as sister groups in the R-ML analysis (Fig. 2) agreeing with Labarque & Ramírez (2012) but contrary to Wheeler et al. (2016: fig. 3), who recovered Drymusidae paraphyletic including Periegopidae in their preferred tree (although in several of their analyses Drymusidae was recovered monophyletic). The R-ML and R-BI analyses supported Drymusidae (Fig. 2), but the family was recovered polyphyletic in our complete data set analyses (Supporting Information, Fig. S1). The Antillean D. armasi emerged as sister to the Ochyrocera species in the C-ML and C-BI analyses, and it was related to S. pagodas and Usofila sp. in the C-EW analysis. Independent gene analyses recovered D. armasi with Ochyrocera Simon, 1892, (i.e. using nuclear ribosomal genes), with Scytodidae (i.e. in parsimony analyses) and forming a polytomy with several Synspermiata members (i.e. in Bayesian analyses). Despite these results, most analyses, except the C-BI analysis, showed low support for the polyphyly of Drymusidae (Supporting Information, Fig. S1). Drymusa armasi was finally discarded from the working data set due to its instability across the analyses. We also tested the COI and 28S sequences of Drymusa dinora Valerio, 1971, (EU817714 and EU817719, respectively; Binford et al., 2008), another Caribbean Drymusidae (analyses not shown). We found several indels in the COI sequence; therefore, this sequence was discarded. The maximum likelihood analysis of the 28S supported D. dinora as sister to D. armasi and both species were recovered within Ochyrocera with low support, mirroring our C-ML results (Supporting Information, Fig. S1). Binford et al. (2008) found that members of Scytodes Latreille, 1804, Usofila Keyserling, 1891, and Drymusa, especially D. dinora, presented very long branches in their analyses. We arrived at the same conclusions; hence, we discarded the 28S sequence of D. dinora, D. armasi, S. pagodas and Usofila sp. from our complete and reduced data sets. Despite the low support in our molecular results and in those of Wheeler et al. (2016), Drymusidae, that is, Neotropical, Andean and South African taxa, and Periegopidae are united by morphological synapomorphies, as suggested by a previous cladistic analysis (Labarque & Ramírez, 2012; although Caribbean Drymusidae were not tested in that analysis, the synapomorphies are confirmed here). Drymusidae plus Periegopidae presented (Labarque & Ramírez, 2012) the following morphological synapomorphies: Non-homoplastic: (1) venom outlet of cheliceral fang facing anteriorly (Fig. 3D; but see comments in Labarque & Ramírez, 2012: ch. 34); Homoplastic: (2) regain of cheliceral promarginal teeth (Fig. 3D; Labarque & Ramírez, 2012: ch. 24). From the present analysis, we suggest several additional putative synapomorphies for this clade including: (3) a prolateral tooth row on the superior retroclaws of pretarsi I–II (Fig. 3A); (4) the foot articulated, with the podotarsite subdivided by an additional articulation (Fig. 3B; Labarque et al., 2017: ch. 1, fig. 1I); (5) regain of cheliceral retromarginal teeth (Fig. 4A); and (6) dense brushes on metatarsi III–IV (Fig. 4B). Despite efforts to find morphological synapomorphies for Drymusidae, only a few homoplastic characters may be tentatively suggested so far: (1) regain of a second major ampullate gland spigot on the anterior lateral spinnerets (instead of a nubbin) (Fig. 4C; Labarque & Ramírez, 2012: ch. 70); (2) regain of long deferent ducts on the testis (Michalik & Ramírez, 2014: ch. 4); (3) regain of cleistospermia (Michalik & Ramírez, 2014: ch. 40). After close examination of American Drymusa and several other Synspermiata, we here suggest an additional potential synapomorphy for Drymusidae: (4) inner spermathecae of vulva connected to the uterus externus through a duct (Fig. 4D; except in D. serrana). Figure 4. View largeDownload slide Drymusidae plus Periegopidae putative synapomorphies. Izithunzi capense comb. nov. (♀, A CASENT 9023625), Periegops suterii (Urquhart, 1892) (♀, B AMNH), Izithunzi productum comb. nov. (♀, C CASENT 9043066), Izithunzi silvicola comb. nov. (♀, D CASENT 9021765), preserved specimens. A, left chelicerae detail, retromargin. B, left tarsus-metatarsus IV, ventral. C, right ALS field, detail (corner, spinneret field diagram). D, right spermathecae, detail. Abbreviations: MaAm, major ampullate; Pi, piriform; retro., retrolateral. Scale bar: B, 0.5 mm; A, 0.2 mm; C, 0.09 mm; D, 0.05 mm. Figure 4. View largeDownload slide Drymusidae plus Periegopidae putative synapomorphies. Izithunzi capense comb. nov. (♀, A CASENT 9023625), Periegops suterii (Urquhart, 1892) (♀, B AMNH), Izithunzi productum comb. nov. (♀, C CASENT 9043066), Izithunzi silvicola comb. nov. (♀, D CASENT 9021765), preserved specimens. A, left chelicerae detail, retromargin. B, left tarsus-metatarsus IV, ventral. C, right ALS field, detail (corner, spinneret field diagram). D, right spermathecae, detail. Abbreviations: MaAm, major ampullate; Pi, piriform; retro., retrolateral. Scale bar: B, 0.5 mm; A, 0.2 mm; C, 0.09 mm; D, 0.05 mm. Gondwanan Drymusidae and westward radiation of Izithunzi species in South Africa Most of our analyses show good support for the relationship between the Southern Neotropical, Andean and South African Drymusidae, using both the complete and reduced data sets (Fig. 2; Supporting Information, Fig. S1). This is consistent with a temperate Gondwanan distribution, meaning that these species are found on continental plates once forming Western Gondwana. As the complete separation of Africa and America occurred 95 Mya (Pitman et al., 1993), the potential Mesozoic age of Drymusidae requires testing. Radiation of Izithunzi species within South Africa may have followed a Westward (from East to West) pattern (compare Fig. 2 against Figs. 5, 6), as the more recent proposed sister taxon splits occur westward in Izithunzi distribution. Unfortunately, we lack sequences from I. zondii sp. nov., the easternmost species described for this genus (KwaZulu-Natal Province), which are necessary to fully test this hypothesis. Nevertheless, I. silvicola comb. nov. appears to have diverged early, are found to be sister to the remaining species of the genus and occur at the border between Eastern and Western Cape provinces (Figs 2, 5, 6). The next species to diverge was I. productum comb. nov., which emerges as sister to the western species (Figs 2, 5, 6). This species occurs in the middle of Western Cape Province. Finally, the last two species to diverge were I. capense comb. nov. and I. lina sp. nov. both occurring on the western extreme of Izithunzi distribution. (Figs 2, 5, 6). Figure 5. View largeDownload slide Distribution map of Izithunzi capense comb. nov., Izithunzi lina sp. nov., Izithunzi productum comb. nov. and Izithunzi silvicola comb. nov. in Western Cape and Eastern Cape provinces (South Africa). Figure 5. View largeDownload slide Distribution map of Izithunzi capense comb. nov., Izithunzi lina sp. nov., Izithunzi productum comb. nov. and Izithunzi silvicola comb. nov. in Western Cape and Eastern Cape provinces (South Africa). Figure 6. View largeDownload slide Distribution map of Izithunzi zondii sp. nov. in KwaZulu-Natal province (South Africa). Figure 6. View largeDownload slide Distribution map of Izithunzi zondii sp. nov. in KwaZulu-Natal province (South Africa). CONCLUSIONS Further analyses such as phylogenomic, that is, next-generation sequencing, or total evidence approaches, that is, including morphological and molecular characters, could increase the Synspermiata tree resolution and help to resolve and support the position of Ochyroceratidae within Scytodoidea, the clade Periegopidae plus Drymusidae and the relationships between the Caribbean and South American Drymusidae. In addition, to improve Scytodoidea and Drymusidae resolution, it is very important to increase the sampling of Drymusidae representatives, especially those species of the Neotropical region. Regarding the biogeography of Izithunzi species, there is no doubt that obtaining new molecular data for several specimens of the described species (and potentially new species) will help to better understand the distributional patterns of this genus in South Africa. TAXONOMIC ACCOUNT Drymusidae Simon, 1893 Drymusinae Simon, 1893: 264, 276; Petrunkevitch, 1928: 42, 19; Petrunkevitch, 1939: 158; Gerhardt & Kästner, 1938: 596; Roewer, 1942: 321; Bonnet, 1956: 1610; Lehtinen, 1967: 301. Drymusidae—Lehtinen, 1986: 156 (new rank); Dippenaar-Schoeman & Jocqué, 1997: 152; Jocqué & Dippenaar-Schoeman, 2006: 118; Filmer, 1991: 80; Ramírez et al., 1999; Filmer 2010: 71; Grismado et al., 2014: 80; World Spider Catalog, 2017. Taxonomic history: The subfamily Drymusinae was established by Simon (1893) in the Deuxième édition of his Histoire Naturelle des Araignées and was monogeneric including the type genus Drymusa. Drymusinae were placed within Sicariidae together with Plectreurinae, Scytodinae, Periegopinae, Loxoscelinae and the s.s. Sicariinae (Simon, 1893); this arrangement was followed by Petrunkevitch (1928, 1939), Gerhardt & Kästner (1938), Roewer (1942) and Bonnet (1956). However, some authors treated Drymusa species as members of Sicariidae (Purcell, 1904; Petrunkevitch, 1911; Bryant, 1948) or Scytodidae (Bristowe, 1938; Lehtinen, 1967, who placed them there tentatively). Gertsch (1967) proposed a new concept of Scytodidae containing three genera: Loxosceles, Scytodes and Drymusa; this arrangement was followed by Valerio (1971, 1974) and Brignoli (1976, 1983). Alayón (1981) revalidated Loxoscelidae (= Sicariidae), including Loxosceles and Drymusa, and ignored Drymusinae; this opinion was followed only by himself (Alayón, 1987). Lehtinen (1986) elevated Drymusinae for the first time to family rank: Drymusidae. Unfortunately, neither of the previous hypotheses (Alayón, 1981; Lehtinen, 1986) was based on any supporting morphological characters, that is, autapomorphies for Drymusa, a critical deficiency as indicated by Goloboff & Ramírez (1992). Finally, Platnick (1989) recognized Drymusidae in his catalogue following Lehtinen (1986), and this family rank has been generally accepted (Coddington & Levi, 1991; Filmer, 1991, 2010; Platnick et al., 1991; Goloboff & Ramírez, 1992; Platnick, 1993; Dippenaar-Schoeman & Jocqué, 1997; Ramírez et al., 1999; Alayón, 2000; Ramírez, 2000; Brescovit, Bonaldo & Rheims, 2004; Coddington et al., 2004; Coddington, 2005; Bonaldo, Rheims & Brescovit, 2006; Jocqué & Dippenaar-Schoeman, 2006; Labarque & Ramírez, 2007a, b; Binford et al., 2008; Rheims, Brescovit & Bonaldo, 2008; Labarque & Ramírez, 2012; Lotz, 2012; Agnarsson, Coddington & Kuntner, 2013; Grismado et al., 2014; Michalik & Ramírez, 2014; Wheeler et al., 2016; Griswold & Ramírez, 2017; Labarque et al., 2017; World Spider Catalog, 2017). Monophyly: Most of our preferred analyses (i.e. R-ML and R-BI) highly supported Drymusidae (Fig. 2), even though the family was recovered polyphyletic using the complete dataset (Supporting Information, Fig. S1). The complete-data set analyses showed that the Caribbean Drymusa sequences had very long branches and were very unstable across analyses; hence, we discarded these from the reduced data set (Fig. 2). In addition, our preferred cladistic results showed a consistently and highly supported sister relationship between the South African Izithunzi gen. nov. and the Southern South American Drymusa (Fig. 2). The following homoplastic morphological synapomorphies have been suggested for Drymusidae: (1) regain of a second major ampullate gland spigot on the anterior lateral spinnerets (instead of a nubbin) (Fig. 4C; Labarque & Ramírez, 2012); (2) regain of long testicular deferent ducts (Michalik & Ramírez, 2014); (3) regain of cleistospermia (Michalik & Ramírez, 2014). Herein, we propose another putative synapomorphy: (4) inner spermathecae of vulva connected to the uterus externus through a duct (Fig. 4D; except in D. serrana). Diagnosis: Synspermiata (although with a regain of cleistospermia in Drymusa rengan Labarque & Ramírez, 2007, see Michalik & Ramírez, 2014), Scytodoidea with six eyes in three separated dyads (no anterior median eyes), intercheliceral articulation sclerotized (known as ‘chelicerae fused’, see Ramírez, 2014), chelicera with lamina and promarginal lobe, female cheliceral stridulatory ridges and pedipalpal femoral thorns, male copulatory bulb presenting one non-expandable sclerite, podotarsite with distal dorsal hood and a third area of articulation, females with haplogyne condition (i.e. cul-de-sac type, Austad, 1984), one posterior spiracle near the spinnerets and tracheae with fused third opisthosomal apodemes (Labarque & Ramírez, 2012; Michalik & Ramírez, 2014; Labarque et al., 2017). Drymusids differ from most Scytodoidea in having two pairs of spermathecae on the vulva with the inner spermathecae connected to the uterus externus through a duct (except Stedocys Ono, 1995; Labarque et al., 2009; Labarque & Ramírez, 2012); differ from Periegopidae in having a second major ampullate gland spigot on the anterior lateral spinnerets, a dorsal cuticular roof for the pretarsal depressor tendon (Labarque & Ramírez, 2007a; Labarque & Ramírez, 2012; Labarque et al., 2017), and by hanging beneath loose space webs (see Labarque & Ramírez, 2012); differ from Sicariidae and Scytodidae in having a closed podotarsite with one subdivision (Labarque et al., 2017); differ also from Sicariidae in having three pretarsal claws, bipectinated proclaws on tarsi I–II, a median field of spicules on the posterior median spinnerets (Platnick et al., 1991; Labarque & Ramírez, 2012), and by venom that lacks sphingomyelinase D proteins (i.e. toxin causative agent in ‘loxoscelism’ lesion formation; Binford & Wells, 2003); differ also from Scytodidae in having the cephalic and thoracic areas of the carapace with similar height (i.e. not forming a dome), the cheliceral stridulatory file composed by shallow multiple scales, the venom outlet facing anteriorly, in lacking apical blunt macrosetae on the female pedipalpal tarsus (Labarque & Ramírez, 2012), and lacking the mechanism for spitting a glue-like substance (Kovoor & Zylberberg, 1972; Foelix, 2011) that is typical of scytodids (‘spitting spiders’); differ from Ochyroceratidae in most of the previous characters, except that Drymusidae are like Ochyroceratidae in having three pretarsal claws, bipectinated proclaws on tarsi I–II (Lehtinen, 1986; Pérez-González et al., 2016), a distal hood on podotarsite (Labarque et al., 2017) and a cheliceral lamina. Genera: Drymusa Simon, 1891, Izithunzi gen. nov. Distribution: From Neotropical and Andean regions (Morrone, 2004, 2014) to South Africa. Natural history: Drymusidae species hang beneath loose space webs hidden in wall crevices or below leaf litter. Females may build small, spherical, wrinkled egg sacs, which they carry with their chelicerae, or build huge, irregular egg sacs, which are covered with detritus and attached to the webs (Valerio, 1974; Alayón, 1981; Jocqué & Dippenaar-Schoeman, 2006: 118–119; Labarque & Ramírez, 2007b; Labarque & Ramírez, 2012). Genus Izithunzi gen. nov. urn:lsid:zoobank.org:act:250052B4-E682-4C4C-A0B5-463F34C3FBC1 Type species: Drymusa capensis Simon, 1893. Species included: Izithunzi capense (Simon, 1893) comb. nov., I. lina sp. nov., I. productum (Purcell, 1904) comb. nov., I. silvicola (Purcell, 1904) comb. nov. and I. zondii sp. nov. Etymology: The generic name means shadows (Izithunzi) in Xhosa, a South African Nguni language of Bantu People. It refers to the retiring nature and cryptic environments (i.e. hidden in caves’ crevices or under dense vegetation) where the members of this genus live. The name is neuter in gender. Remarks:Lehtinen (1967) and Platnick et al. (1991) predicted the non-monotypy of Drymusidae. Lehtinen (1967: 301) compared the Neotropical Drymusa nubila Simon, 1892 with the South African D. capensis (= I. capense comb. nov.) and considered that they were not congeneric (although he took no formal action). Monophyly: Putative synapomorphies include (1) cheliceral promargin with ‘rubble teeth’ that are massive, blunt and boulder shaped (Fig. 3D); (2) female with two sclerotized plates on the vulva, anterior and posterior to the uterus externus (Fig. 7A, B); and (3) male with anterior stridulatory ridge plate on opisthosoma (Fig. 7C, D). Figure 7. View largeDownload slide Izithunzi gen. nov. putative synapomorphies. Izithunzi capense comb. nov. (♀, A CASENT 9023625), Izithunzi lina sp. nov. (♀, B NMBA; ♂, C NMBA), Izithunzi silvicola comb. nov. (♂, D CASENT-9048600), preserved specimens. A, B, female internal genitalia (A, lateral; B, dorsal, right detail). C, D, male opisthosoma, stridulatory ridge plate (C, frontal; D, detail). Abbreviations: ext., externus; recep., receptacle; sp., spermathecae. Scale bar: C, 0.5 mm; D, 0.15 mm; A, B, 0.1 mm; F, 0.05 mm. Figure 7. View largeDownload slide Izithunzi gen. nov. putative synapomorphies. Izithunzi capense comb. nov. (♀, A CASENT 9023625), Izithunzi lina sp. nov. (♀, B NMBA; ♂, C NMBA), Izithunzi silvicola comb. nov. (♂, D CASENT-9048600), preserved specimens. A, B, female internal genitalia (A, lateral; B, dorsal, right detail). C, D, male opisthosoma, stridulatory ridge plate (C, frontal; D, detail). Abbreviations: ext., externus; recep., receptacle; sp., spermathecae. Scale bar: C, 0.5 mm; D, 0.15 mm; A, B, 0.1 mm; F, 0.05 mm. Diagnosis: Izithunzi gen. nov. can be distinguished from Drymusa by the boulder-shaped cheliceral promarginal rubble teeth, the female genitalia having two sclerotized plates anterior and posterior to the uterus externus and the male opisthosoma having an anterior stridulatory ridge plate. Description: Female total length between 5.65 and 15.03 and male total length between 4.95 and 12.27. Carapace reddish with distinguished dun pattern forming two central large patches making a backward-pointing V-shaped mark, each anteriorly prolonged in three longitudinal lines surrounding the eyes and laterally extending in three wavy lines; clypeus with cross-linked pattern. Cephalic area slightly more than one-half width of thoracic area. Chelicerae promargin with two massive and blunt teeth, the rubble teeth, usually dark under the microscope, a triangular lamina contiguous with the paturon margin with a row of several macrosetae against it and a fleshy lobe apically blunt with a basal row of several long and filiform setae, the bracket setae (Figs 3D, 8A). Cheliceral retromargin with two rubble teeth, rarely three (see I. silvicola comb. nov.), forming a row perpendicular to the fang furrow, apical tooth large and proximal small (Fig. 4A). Chelicerae ectal margin stridulatory ridge formed by multiple shallow scales (Fig. 3E). Fang venom gland outlet distal, facing anteriorly (Fig. 3D). Cheliceral bases articulated with a small, sclerotized, posterior intercheliceral sclerite (Fig. 8B). Venom gland extending into carapace (Fig. 8C). Endites longer than wide, bending prolaterally and converging in front of the labium, fleshy apical profiles almost touching each other (Fig. 8D). Labium trapezoidal, longer than wide, narrowed close to the acute apical margin and partially separated from sternum by a membranous suture (Figs 8, 9). Maxillary gland pore field clumped (Fig. 8). Labral tongue apically concave, with dorsal setae (Fig. 8). Pedipalpal claw reduced to a nubbin, distal tarsus with four prolateral macrosetae curved distally, and basal femur with prolateral thorn that might be distally acute or blunt (Fig. 3F). Sternum oval, longer than wide, bordered (Fig. 9). Precoxal triangles fused to sternum (Fig. 9). Legs tan dappled with dun. Proclaws I–II clearly bipectinate with longer teeth on the prolateral row; retroclaws I–II with prolateral tooth row (Fig. 3A). Foot articulated, closed podotarsite with one subdivision, distal hood covering the base of the superior claws, with distal-ventral frictional setae (Fig. 3A, B). Tarsal organ exposed, ovoid and with three receptor-cell dendrite terminals (sensilla) (Fig. 9). Metatarsal trichobothria opening distal margin entire; proximal and distal plates of trichobothria smooth, not well differentiated, distal plate contiguous with the surrounding cuticle; sculpture on basal expansion of trichobothrial setae smooth (Fig. 9). Metatarsi III–IV with dense brush of apical ventral setae (Fig. 4B). Pedicel lorum triangular, not transversely divided, narrowed posteriorly to a pointed end, without slit sensilla stripes (Fig. 9). Pedicel sternites and pleurites separated. Opisthosoma colour overall dark brown with or without chevrons. Epigastrium posterior border with two small, lateral, sclerotized grooves, which seemingly provide a guide for the copulatory bulb apex (see at end of paragraph) to enter into the vulva. Epigastrium and postepigastrium on the external genital area may be heavily sclerotized, forming the epigastrium plate and the postepigastrium plate, respectively (see I. silvicola comb. nov.). Postepigastric foveae absent from abdominal venter (present in many Scytodes and some Neotropical Drymusa). Vulva with an anterior pair of spermathecae (inner), each connected to the uterus externus through a duct (Figs 4D, 7, 10A–C), an anterior sclerotized plate with a second pair of spermathecae (outer) and a posterior sclerotized plate with two dorsal receptacula (Figs 7, 10A–C). Inner spermatheca larger than outer, its duct may open far from (separated, Figs 10A, B) or directly next to the latter (clustered, Figs 4D, 10C). Outer spermathecae and dorsal receptaculum may provide opisthosomal muscle insertions (Fig. 10C). Spermathecae and dorsal receptaculum with ductless glands well spaced or in patches of two to many glands (Figs 4D, 7, 10A–C). Sclerotized plates may be separated from each other surrounding the uterus externus (Figs 7A, 10A), pressed together anteriorly squeezing the uterus externus (Figs 7B, 10B) or completely integrated with the uterus externus forming a unit, the integrated plate (Figs 10). Integrated plate may present a conspicuous or diffuse anterior longitudinal middle ridge (Fig. 10C) forming a sort of crest, which extends across the plate. Uterus externus may or may not extend beyond the vulval plate anterior borders (compare I. capense comb. nov. against I. zondii sp. nov.). Tracheal spiracle wide, separated from spinnerets (Fig. 11). Third opisthosomal entapophyses fused, forming a long median trachea (Fig. 3C). Colulus well defined, ovoid and posteriorly narrowed (Fig. 11). Anterior lateral spinnerets (ALS) with three articles (Fig. 11). ALS with two major ampullate gland spigots (MaAm) and a separated field of several piriform (Pi) gland spigots (Fig. 11). Posterior median spinnerets (PMS) tetrahedral with basal straight and plumose setae on anteromedian surface, a single aciniform gland spigot (Ac), and a field of spicules on mesal surface Fig. 11. Posterior lateral spinnerets (PLS) conical, with several aciniform gland spigots (Fig. 11). Male markings as in female. Opisthosoma with an anterior sclerotized plate with stridulatory ridge, dorsally facing posterior prosoma (Fig. 7). Epiandrous spigots arising in several bunches from isolated pits (Fig. 12). Male with spinnerets as in female but differs in having the PLS with one aciniform gland spigot instead of several spigots (Fig. 11). Male pedipalp with prolateral femoral thorns as in females (Fig. 13). Patella-tibia joint dicondylic, ventral condyle heavily sclerotized, projecting prolaterally, arthrodial membrane broad prolaterally, occupying one third of patella (Figs 12, 13); these modifications on the articulation seem to allow a prolateral movement to the tibia. Femora narrow (L ≤ 3.5× W) to elongated (L > 4× W). Tibiae swollen (L < 2× W) to thin (L > 2.5× W) (compare I. capense comb. nov. against I. silvicola comb. nov.). Cymbium as long as wide to swollen (L ≤ 1.5× W), apically blunt, and as long as or 1.5 times longer than the copulatory bulb base (compare I. capense comb. nov. against I. silvicola comb. nov.). Copulatory bulb presenting one non-expandable piriform sclerite (subtegulum, tegulum and embolus fused). Base (subtegulum + tegulum) showing the sperm duct. Apex (embolus) as long as to more than three times longer than the base, laminated, slightly curved and implanted prolaterally (compare I. productum comb. nov. against I. silvicola comb. nov.). Figure 8. View largeDownload slide Izithunzi gen. nov. chelicerae and mouthparts. Izithunzi zondii sp. nov. (♀, A, B, F, NMSA), Izithunzi silvicola comb. nov. (♀, D, CASENT 9021766, ♀, E CASENT 9048599), Izithunzi capense comb. nov. (♂, C CASENT 9026022), preserved specimens. A–C, Chelicerae (A, anterior; B, posterior; C, venom gland; anterior). D, labium and endites (ventral). E, left maxillary gland (detail). F, labrum and endites (dorsal). Abbreviations: scl., sclerite. Scale bar: A, B, D, 0.5 mm; F, 0.25 mm; C, 0.2 mm; E, 0.02 mm. Figure 8. View largeDownload slide Izithunzi gen. nov. chelicerae and mouthparts. Izithunzi zondii sp. nov. (♀, A, B, F, NMSA), Izithunzi silvicola comb. nov. (♀, D, CASENT 9021766, ♀, E CASENT 9048599), Izithunzi capense comb. nov. (♂, C CASENT 9026022), preserved specimens. A–C, Chelicerae (A, anterior; B, posterior; C, venom gland; anterior). D, labium and endites (ventral). E, left maxillary gland (detail). F, labrum and endites (dorsal). Abbreviations: scl., sclerite. Scale bar: A, B, D, 0.5 mm; F, 0.25 mm; C, 0.2 mm; E, 0.02 mm. Figure 9. View largeDownload slide Izithunzi gen. nov., female prosoma, legs and pedicel. Izithunzi silvicola comb. nov. (A CASENT 9048599), Izithunzi capense comb. nov. (♀, B, C CASENT 9023625), Izithunzi productum comb. nov. (♀, D CASENT 9043066), preserved specimens. A, sternum. B, tarsal organ, I left. C, metatarsal trichobothria, I right. D, pedicel, dorsal. Scale bar: A, 1 mm; D, 0.09 mm; C, 0.02 mm; B, 0.009 mm. Figure 9. View largeDownload slide Izithunzi gen. nov., female prosoma, legs and pedicel. Izithunzi silvicola comb. nov. (A CASENT 9048599), Izithunzi capense comb. nov. (♀, B, C CASENT 9023625), Izithunzi productum comb. nov. (♀, D CASENT 9043066), preserved specimens. A, sternum. B, tarsal organ, I left. C, metatarsal trichobothria, I right. D, pedicel, dorsal. Scale bar: A, 1 mm; D, 0.09 mm; C, 0.02 mm; B, 0.009 mm. Figure 10. View largeDownload slide Izithunzi gen. nov., female vulva types, dorsal. Izithunzi capense comb. nov. (♀, A CASENT-9021768), Izithunzi lina sp. nov. (♀, B NMBA), Izithunzi silvicola comb. nov. (♀, C CASENT-9021765). A, separated plates surrounding the uterus externus. B, pressed plates anteriorly squeezing the uterus externus. C, integrated plate forming a unit with the uterus externus. Abbreviations: ant., anterior; ext., externus; int., integrated; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A–C, 0.2 mm. Figure 10. View largeDownload slide Izithunzi gen. nov., female vulva types, dorsal. Izithunzi capense comb. nov. (♀, A CASENT-9021768), Izithunzi lina sp. nov. (♀, B NMBA), Izithunzi silvicola comb. nov. (♀, C CASENT-9021765). A, separated plates surrounding the uterus externus. B, pressed plates anteriorly squeezing the uterus externus. C, integrated plate forming a unit with the uterus externus. Abbreviations: ant., anterior; ext., externus; int., integrated; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A–C, 0.2 mm. Figure 11. View largeDownload slide Izithunzi gen. nov., posterior respiratory system and spinnerets. Izithunzi lina sp. nov. (♀, A NMBA), Izithunzi zondii sp. nov. (♀, B NMSA), Izithunzi productum comb. nov. (♀, C CASENT 9043066), Izithunzi silvicola comb. nov. (♀, D, E CASENT 9048599, ♂, F CASENT 9048600), preserved specimens. A, opisthosoma, tracheae spiracle detail. B, spinnerets field, posterior. C, right ALS field. D, left PMS field, inner surface. E, F, left PMS-PLS field (C–F corners, spinnerets field diagram). Abbreviations: Ac, aciniform; ALS, anterior lateral spinnerets; MaAm, major ampullate; Pi, piriform; PMS, posterior median spinnerets; PLS, posterior lateral spinnerets. Scale bar: A, 0.5 mm; B, 0.2 mm; D, 0.09 mm; E, 0.03 mm; C, F, 0.02 mm. Figure 11. View largeDownload slide Izithunzi gen. nov., posterior respiratory system and spinnerets. Izithunzi lina sp. nov. (♀, A NMBA), Izithunzi zondii sp. nov. (♀, B NMSA), Izithunzi productum comb. nov. (♀, C CASENT 9043066), Izithunzi silvicola comb. nov. (♀, D, E CASENT 9048599, ♂, F CASENT 9048600), preserved specimens. A, opisthosoma, tracheae spiracle detail. B, spinnerets field, posterior. C, right ALS field. D, left PMS field, inner surface. E, F, left PMS-PLS field (C–F corners, spinnerets field diagram). Abbreviations: Ac, aciniform; ALS, anterior lateral spinnerets; MaAm, major ampullate; Pi, piriform; PMS, posterior median spinnerets; PLS, posterior lateral spinnerets. Scale bar: A, 0.5 mm; B, 0.2 mm; D, 0.09 mm; E, 0.03 mm; C, F, 0.02 mm. Figure 12. View largeDownload slide Izithunzi lina sp. nov. (♂, A, B NMBA), preserved specimen. A, epiandrous spigots field (note several bunches arising from isolated pits). B, male right pedipalp, dicondylic patella-tibia joint, prolateral. Scale bar: A, 0.2 mm; B, 0.1 mm. Figure 12. View largeDownload slide Izithunzi lina sp. nov. (♂, A, B NMBA), preserved specimen. A, epiandrous spigots field (note several bunches arising from isolated pits). B, male right pedipalp, dicondylic patella-tibia joint, prolateral. Scale bar: A, 0.2 mm; B, 0.1 mm. Figure 13. View largeDownload slide Izithunzi capense comb. nov., copulatory organs (♀, A, C CASENT-9023625; ♀, B CASENT-9021768; ♂, D–F CASENT-9021768). A, external female genitalia (ventral). B, vulva (dorsal). C, left dorsal receptaculum, detail. D–F, left pedipalp (D, prolateral; E, retrolateral; F, apical). Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, D–F, 0.5 mm; B, 0.2 mm; C, 0.05 mm. Figure 13. View largeDownload slide Izithunzi capense comb. nov., copulatory organs (♀, A, C CASENT-9023625; ♀, B CASENT-9021768; ♂, D–F CASENT-9021768). A, external female genitalia (ventral). B, vulva (dorsal). C, left dorsal receptaculum, detail. D–F, left pedipalp (D, prolateral; E, retrolateral; F, apical). Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, D–F, 0.5 mm; B, 0.2 mm; C, 0.05 mm. Izithunzi capense (Simon, 1893) comb. nov. (Figs 1a–c, 3a–c, 4a, 5, 7A, 8C, 9B, C, 10A, 13A–C, 14, 15) Drymusa capensis Simon, 1893: 278; Purcell, 1904: 154; Filmer, 1991: 80; Filmer, 2010: 71; Labarque & Ramírez, 2012: 3, figs 2B, 3C–D, 6C–D, 9B, 10B, 11B, 12B, 13B, 14B, 15B, 16B, 17B, 18B, 19A, 23A–F, 29C, D, 31B [one immature holotype (MNHN AR-1326) from South Africa, Western Cape, examined.] Figure 14. View largeDownload slide Izithunzi capense comb. nov. (♂, A, B CASENT-9026022), Izithunzi lina sp. nov. (C, D CASENT-9043226), male chelicerae. A, B, right, chelicerae, promargin (flattened teeth detail). C, both chelicera (frontal). D, left chelicerae (medial). Abbreviations: pro., prolateral; retro., retrolateral. Scale bar: C, D, 0.2 mm; A, 0.1 mm; B, 0.03 mm. Figure 14. View largeDownload slide Izithunzi capense comb. nov. (♂, A, B CASENT-9026022), Izithunzi lina sp. nov. (C, D CASENT-9043226), male chelicerae. A, B, right, chelicerae, promargin (flattened teeth detail). C, both chelicera (frontal). D, left chelicerae (medial). Abbreviations: pro., prolateral; retro., retrolateral. Scale bar: C, D, 0.2 mm; A, 0.1 mm; B, 0.03 mm. Figure 15. View largeDownload slide Izithunzi capense comb. nov., preserved specimens (♀, A, B CASENT-9023625; ♂, C, D CASENT-9021768; immature, E, F MNHN AR-1326 holotype). A–D, habitus (A, C, E, dorsal; B, D, F, ventral). Scale bar: A–D, 2 mm; E, F, 0.5 mm. Figure 15. View largeDownload slide Izithunzi capense comb. nov., preserved specimens (♀, A, B CASENT-9023625; ♂, C, D CASENT-9021768; immature, E, F MNHN AR-1326 holotype). A–D, habitus (A, C, E, dorsal; B, D, F, ventral). Scale bar: A–D, 2 mm; E, F, 0.5 mm. Loxosceles valida Lawrence, 1964: 61, figs 3, 4; Newlands, 1975: 146, fig. 7; 1986: 80, figs 55–57; Brignoli, 1976: 147. [one ♂ holotype (SAM-ENW B010012) from South Africa, Western Cape Province, Cape Town, Kalk Bay, Echo Valley, Echo Halt Cave, [−34.116667, 18.433333], Apr.1954, J. Grindley col., not examined.] syn. nov. Drymusa valida – Lotz, 2012: 37, fig. 18A, B. Remarks: Whereas we have not examined the type specimens of other Izithunzi species and L. valida, we believe that species attribution is unproblematic. Purcell (1904: 154) reported that the type of D. capensis is very immature and that the adults of this species are much larger than the other species from South Africa. Here, we reconfirm his assessment: the type specimen is very immature (Fig. 13E, F), and the total length of females (14.30) and males (12.80) that we identify as this species fits Purcell’s suggestion of the large size of this species. Purcell (1904: 154) also provided a key to females of the African species using characters that match those that we use here to diagnose I. capense comb. nov. Regarding L. valida, we think the evidence for synonymy with I. capense comb. nov. is convincing. Lawrence (1964: 61) described a row of several macrosetae against the triangular lamina on the chelicerae promargin of L. valida, a character absent in Loxosceles but present in Drymusidae. Brignoli (1976: 147) suggested that L. valida might be a D. capensis based on their similar geographical distribution. The presence of three pretarsal claws in L. valida was brought to Norman Larsen’s attention by Vince Roth during a visit to the South African Museum in Cape Town (Larsen, 1994). This feature contrasts to the two claws of Loxosceles and other Sicariidae. Finally, Lotz (2012: 37) transferred L. valida to Drymusa (and to Drymusidae) based on the presence of an articulated podotarsite (‘long and clear onicium’) and three pretarsal claws. The opisthosomal coloration pattern and male pedipalpal configuration of L. valida resemble those of I. capense comb. nov. (compare Figs 13, 15 against Lawrence, 1964: figs 3, 4). The examination of a female topotype (i.e. South Africa, Western Cape Province, Cape Town, Kalk Bay, Echo Valley, Echo Halt Cave [−34.116667, 18.433333]), further corroborates our new synonymy. Key to species of Izithunzi gen. nov. 1 Females 2 – Males … 6 2(1) Small, prosoma length less than 4.0 mm … 3 – Large, prosoma length greater than 5.0 mm … 4 3(2) Spermathecae clustered … 5 – Spermathecae separated … I. zondii sp. nov. 4(2) Vulval plates separated; posterior plate nearly straight … I. capense comb. nov. – Vulval plates pressed together anteriorly; posterior plate curved anteriorly … I. lina sp. nov. 5(3) Epigastrium posterior border with dark, small, thick setae; epigastrium not forming an elongated plate … I. productum comb. nov. – Epigastrium posterior border lacking those setae; epigastrium forming a plate, heavily sclerotized and posteriorly elongated to the median point of the opisthosoma … I. silvicola comb. nov. 6(1) Pedipalp short, femur length less than 3.5 times its width; tibia length less than 2 times its width … 7 – Pedipalp elongated, femur length more than four times its width; tibia length more than 2.5 times its width … I. silvicola comb. nov. 7(6) Copulatory bulb base and apex transition smooth; apex distally dark … 8 – Copulatory bulb base and apex transition indented; apex mostly dark … I. lina sp. nov. 8(7) Copulatory bulb apex 1½ times longer than its base, with an acute and slightly curved tip … I. capense comb. nov. – Copulatory bulb apex no longer than its base, with laminar truncated tip … I. productum comb. nov. Material examined: ♀ from South Africa, Western Cape Province, Cape Town, Table Mountain National Park, Newlands Forest Preserve, −33.973999, 18.444133, 25 February 2006, elev. 145 m, J. Miller, H. Wood, N. Larsen cols., preparation codes FML-00435-00444 [♀], deposited in CAS (CASENT9023625); ♂, ♀ and three immatures, same data, preparation codes FML-01010, FML-01108 and FML-01118 [♂], and FML-01107 [♀], CAS (CASENT9026022); 4♂, 3♀ and four immatures, same locality, 4 October 2001, N. Larsen, K. Muller, S. Prinsloo, D. Ubick, S. Ubick cols., CAS (CASENT9048605); ♀ and two immatures, same data, 4 November 1925, Lang col., AMNH; ♀ and one immature, same locality, 18 December 1996, elev. 150 m, indigenous forest at night, P. Sierwald col., FMNH; three immatures, same data, FMNH; one immature, same data, elev. 120 m, during the day, FMNH; two immatures, same data, FMNH; two immatures, same data, FMNH; three immatures, same data, FMNH; seven immatures, same data, C. Griswold col., CAS (CASENT9053375); one immature, same locality, 15 January 2009, pine plantation decayed log, site 2, C. Uys col., NCP (2010/1949); two immatures, same data, 23 May 2008, afrotemperate forest decayed log, NCP (2010/1922); ♂, same locality, March 1993, S. Muller col., NCP; ♀, same data, NCP; ♀, same locality, −33.977317, 18.439783, 4 October 2011, elev. 195 m, L. Almeida, C. Griswold, T. Meikle, N. Larsen cols., CAS (CASENT9042516); 2♀ and two immatures, same data, CAS (CASENT9043287); ♀, same data, CAS (CASENT9043286); ♀, same data, CAS (CASENT9043173); ♀ and one immature, same data, CAS (CASENT9043171); 5♀ and nine immatures, same locality, Fernwood Gully, −33.966667, 18.450000, 18 December 1996, elev. 120–150 m, indigenous forest, C. E. Griswold col., preparation code FML-01151 [♀], CAS (CASENT9048602); ♀, same data, CAS (CASENT9048603); 2♂ and 2♀, same locality, Kirstenbosch Botanic Garden, Skeleton Gorge Forest, −33.983333, 18.433333, 7 January 1985, elev. 700 ft, webs beneath logs, C. Griswold, T. Meikle Griswold cols., preparation codes APG-00055 [♀], FML-00546-00547 [♂] and FML-00705 [♀], CAS (CASENT9021768); 2♀ and three immatures, same data, NMSA; five immatures, same data, AMNH; 3♂ and one immature, same data, moulted in captivity, NMSA; ♂ and 2♀, same data, C. Griswold col., NMSA; ♀, same locality, October 1985, 800 ft, C. Griswold col., CAS (CASENT9021767); 3♀ and three immatures, same locality, 26–29 October 1985, elev. 700–1000 ft, C. Griswold, J. Doyen, T. Meikle Griswold cols. NMSA; 4♀ and nine immatures, same data, NMSA; four immatures, same locality, 5 February 2009, afrotemperate forest decayed log, site 5, C. Uys col., NCP (2010/1950); one immature, same locality, Nursery Ravine, Wynberg Caves, [−33.986430, 18.403772], 13 February 1991, cave entrance, V. D. Roth, B. Roth cols., CAS (CASENT9048604); one immature, same locality, Orange Kloof, [−33.998611, 18.392778], 28 January 2009, afrotemperate forest, sugar-baited ant trap, C. Uys col., NCP (2010/1952); ♀, same city, Kalk Bay, Echo Valley, Echo Halt Cave [−34.116667, 18.433333], December 1987, A. le. Roy col., NCP. Further material examined byLotz (2012): ♀ from South Africa, Western Cape Province, Cape Town, Kalk Bay, Echo Valley, Devil’s Pit, [−34.117942, 18.436667], June 1954, J. Grindley col., SAM (B010015); one immature, same locality, Tartarus Cave, [−34.113647, 18.441556], July 1961, J. Grindley col., SAM (B10013); ♀, same city, Table Mountain National Park, Nursery Ravine, Wynberg Caves, Powder Room, [−33.986430, 18.403772], March 1956, South African Speleological Association, SAM (B10018); ♀, same data, February 1956, SAM (B10016); one immature, same data, March 1931, R. Lawrence, SAM (B7892); one immature, same locality, Giants Workshop, [−33.989180, 18.407553], July 1956, J. Grindley col., SAM (B10017); ♀, same locality, Bats Cave, [−33.988569, 18.406797], September 1960, J. Grindley col., SAM (B10014). Diagnosis: Females of I. capense comb. nov. resemble those of I. lina sp. nov. by the epigastrium protruded ventro-anteriorly, the inner and outer spermathecae separated and the uterus externus exceeding beyond the vulval plate anterior borders (Figs 7, 10, 13, 16), but it can be distinguished by the epigastrium covered with long and thin setae, the postepigastrium with an anterior lip covering the epigastric furrow and the vulval plates separated (Figs 7, 10, 13), whereas I. lina sp. nov. presents long, thick and dark setae on the epigastrium, lacks the anterior lip on the postepigastrium and has pressed vulval plates (Figs 7, 10, 16). Males of I. capense comb. nov. resemble those of I. lina sp. nov. by having the cheliceral fang promarginally and distally excavated, and the pro- and retromarginal cheliceral teeth close to the base of the fang extremely modified, enlarged and flattened, which fit in the fang’s excavation (Fig. 14); but it can be distinguished by presenting a smooth transition between the base and apex of the copulatory bulb, and the apex 1½ times longer than the base with an acute and slightly curved tip (Fig. 13), whereas I. lina sp. nov. has an indented transition, and the apex two times longer than the base, heavily sclerotized (dark) and a broad tip (Fig. 16). Figure 16. View largeDownload slide Izithunzi lina sp. nov., copulatory organs (♀, A–C NMBA; ♂, D–F NMBA). A, external female genitalia (ventral). B, C, vulva (B, dorsal; C, lateral). D–F, right pedipalp [D, prolateral; E, retrolateral; F, apical (corner, tip detail, retrolateral)]. Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, F, 2.5 mm; D, E, 0.5 mm; B, 0.2 mm; C, 0.1 mm. Figure 16. View largeDownload slide Izithunzi lina sp. nov., copulatory organs (♀, A–C NMBA; ♂, D–F NMBA). A, external female genitalia (ventral). B, C, vulva (B, dorsal; C, lateral). D–F, right pedipalp [D, prolateral; E, retrolateral; F, apical (corner, tip detail, retrolateral)]. Abbreviations: ant., anterior; ext., externus; post., posterior; recep., receptacle; sp., spermathecae. Scale bar: A, F, 2.5 mm; D, E, 0.5 mm; B, 0.2 mm; C, 0.1 mm. Redescription female (Table Mountain National Park, Newlands Forest Preserve: Images CASENT 9023625; Measurements CASENT 9048605): Total length 15.03. Prosoma: length 5.7, width 3.96, height 2.77. Sternum: length 2.93, width 2.2. Leg measurements: femur: I: 15.15, II: 13.53, III: 11.02, IV: 13.78; patella: I: 1.71, II: 1.68, III: 1.59, IV: 1.72; tibia: I 14.78, II: 12.4, III: 9.2, IV: 12.27; metatarsus: I: 15.15, II: 13.4, III: 10.5, IV: 13.4; tarsus: I: 2.14, II: 2.02, III: 2.09, IV: 2.89; podotarsite: I: 0.22, II: 0.24, III: 0.28, IV: 0.22. Total: I: 49.15, II: 43.27, III: 34.68, IV: 44.27. Leg formula: 1423. Opisthosoma: length 8.72, width 4, height 4.65. Thoracic area lateral margins and central V-shaped pattern darkish, forming a continuum (Fig. 15). Chelicerae promargin with five bracket setae, and a row of seven to eight macrosetae against the triangular lamina. Labium dun, reddish at narrow area and white at apex (Figs 8, 15). Pedipalpal prolateral femoral thorn distally acute. Sternum dun (Figs 8, 15). Femora and tibiae dun, patellae, metatarsi and tarsi tan (Fig. 15). Opisthosoma colour overall dark brown forming thick chevrons extending anteriorly (Fig. 15). Chevrons well-spaced anteriorly, clustered posteriorly, the first two forming a continuum (Fig. 15). Anterior vulval plate slightly sclerotized, and posterior plate nearly rectangular and slightly curved anteriorly (Figs 7, 10, 13). Both spermathecae oval (Figs 7, 10, 13). Redescription male (Table Mountain National Park, Newlands Forest Preserve: Habitus CASENT 9021768; SEM CASENT 9026022; Measurements CASENT 9048605): Total length 12.27. Prosoma: length 5.6, width 4.12, height 2.8. Sternum: length 2.67, width 2.02. Leg measurements: femur: I: 17.1, II: 16.5, III: 13.02, IV: 14.9; patella: I: 1.74, II: 1.78, III: 1.68, IV: 1.7; tibia: I 16.5, II: 14.65, III: 10.4, IV: 12.77; metatarsus: I: 18.0, II: 16.1, III: 12.27, IV: 14.9; tarsus: I: 2.37, II: 2.37, III: 2.35, IV: 3.14; podotarsite: I: 0.27, II: 0.21, III: 0.32, IV: 0.33; total: I: 55.98, II: 51.61, III: 40.04, IV: 47.74. Leg formula: 1243. Opisthosoma: length 6.3, width 3.76, height 3.76. Male pedipalp: femur: 2.06, patella: 0.75, tibia: 1.5, tarsus: 0.59. Coloration as female (Fig. 15). Chelicerae promargin also with five bracket setae (Fig. 14). Epiandrous spigots arising in seven bunches from isolated pits. Pedipalpal prolateral femoral thorn also distally acute but longer than in females (Fig. 13); femora narrow (L ≤ 3.5× W); tibiae swollen, longer than width (L < 2× W) (Fig. 13). Copulatory bulb apex elongated and acute distally. In addition, the copulatory bulb apex looks slightly curved in lateral view (both pro- and prolateral) and straight in apical view (i.e. apical view of the cymbium) (Fig. 13). Distribution: Western Cape Province, South Africa, from Table Mountain National Park to Kalk Bay and surroundings (Fig. 5; Supporting Information, Fig. S1). Natural history: According to Larsen (1994), I. capense comb. nov. specimens are found under exfoliated bark or in crevices between boulders, always in cool shaded areas and hanging beneath loose space webs (Fig. 1A–C). Izithunzi capense comb. nov. individuals are sensitive to the light, and they quickly retreat to the darkness after a minute of exposure with an unfiltered electric torch (flashlight) (Larsen, 1994). Our own observations match these. Izithunzi lina sp. nov. urn:lsid:zoobank.org:act:269AB044-0E94-4F7A- 9A18-CEEA44D74686 (Figs 5, 7, 10–12, 14, 16, 17) Type material: Holotype: ♀ from South Africa, Western Cape Province, Overberg DC, Fernkloof Nature Reserve, 3.98 km 90° E Hermanus, −34.412683, 19.288350, 13 October 2011, elev. 14 m, general collecting at wet, mossy cl

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Molecular phylogeny and revision of the false violin spiders (Araneae: Drymusidae) of Africa

Molecular phylogeny and revision of the false violin spiders (Araneae: Drymusidae) of Africa