Using the Integrative Approach to Update a Gap of One Century: Redescription and New Distribution Records of the South American Tarantulas Grammostola pulchra (Araneae: Mygalomorphae: Theraphosidae)

© 2023 Academia Sinica, Taiwan Open Access Using the Integrative Approach to Update a Gap of One Century: Redescription and New Distribution Records of the South American Tarantulas Grammostola pulchra (Araneae: Mygalomorphae: Theraphosidae) Renan S. Pittella1,* , Pedro G. Bassa2, Edison Zefa3, and Filipe M. Bianchi4,* 1Laboratório de Ecologia de Interações, Departamento de Ecologia, Universidade Federal do Rio Grande do Sul, Porto Alegre-RS, Brazil. *Correspondence: E-mail: [email protected] (Pittella) 2Laboratório de Comportamento e Ecologia de Formigas, Departamento de Ecologia, Zooloogia e Genética, Universidade Federal de Pelotas, Pelotas-RS, Brazil. E-mail: [email protected] (Bassa) 3Laboratório de Invertebrados, Departamento de Ecologia, Zoologia e Genética, Universidade Federal de Pelotas, Pelotas-RS, Brazil. E-mail: [email protected] (Zefa) 4Laboratório de Entomologia Sistemática, Departamento de Zoologia, Universidade Federal do Rio Grande do Sul, Porto Alegre-RS, Brazil. *Correspondence: E-mail: [email protected] (Bianchi) Received 8 June 2021 / Accepted 16 November 2022 / Published 6 April 2023 Communicated by I-Min Tso Taxonomic researchers have used multiple sources of evidence to support species hypotheses and delimitations. Grammostola Simon (Mygalomorphae: Theraphosidae) comprises 20 valid species endemic to South America, six occurring in Brazil. The classical morphological approach based mainly on genitalia may be misleading in recognizing species in this genus. Thus, we used morphology, geographical distribution, genetic distance, and phylogeny to support the redescription of Grammostola pulchra from southern Brazil, a species described a century ago. We also diagnosed and illustrated the species. Males have a developed apical keel at the apex of the embolus; for the first time, this type of structure has been reported in a species of Grammostola. The molecular analyses using the partial sequence of Cytochrome c oxidase subunit I showed 7% of genetic distance (p-distance) between G. pulchra and Grammostola anthracina. Distance and tree-based methods (ASAP and bPTP, respectively) assigned G. pulchra as a valid species. The gene-tree under Bayesian and Maximum-Likelihood recovered a similar topology, placing G. pulchra as closely related to Grammostola burzaquensis and G. anthracina. Morphological characters which could be important in the taxonomy of the genus are further discussed. Key words: Barcoding, COI, Integrative taxonomy, Neotropical region, Tarantula. Citation: Pittella RS, Bassa PG, Zefa E, Bianchi FM. 2023. Using the integrative approach to update a gap of one century: redescription and new distribution records of the South American Tarantulas Grammostola pulchra (Araneae: Mygalomorphae: Theraphosidae). Zool Stud 62:05. doi:10.6620/ZS.2023.62-05. BACKGROUND In a broad sense, taxonomy is a scientific discipline of biology that consists of varied activities, e.g., comparing taxa intending to infer relational hypotheses, approaching intraand interspecific variation, and constructing tools for accurate identification (Enghoff and Seberg 2006). Nevertheless, a core priority of taxonomic studies is to discover and delimitate taxa. Furthermore, since the species boundaries should be treated as a testable hypothesis, species concepts are built on various pieces of evidence and criteria to distinguish species lineages, and recognize taxonomic units (Yeates et al. 2011). Thus, taxonomic problems Zoological Studies 62: 5 (2023) doi:10.6620/ZS.2023.62-05 1 © 2023 Academia Sinica, Taiwan have been approached simultaneously by distinct techniques, using plural sources of evidence to elucidate species delimitation (Padial et al. 2010). Morphological characters have been used to classify spiders, mainly those related to the copulatory organs (Pérez-Miles et al. 1996; Bertani 2000). However, some groups, like mygalomorphs, present very similar somatic morphology and less diverse genital structures compared to the sister group Araneomorphae (Hedin and Bond 2006). Because of this, the traditional morphological approach may be insufficient to recognize species accurately, bringing up taxonomic incongruences and inconsistencies (Raven 1990). For example, within the Theraphosidae, the species of Grammostola Simon, 1892, are difficult to sort out and identify due to the copulatory structures being extremely homogeneous among them (Bücherl 1957; Montes de Oca et al. 2016). In these situations, molecular data can provide promising evidence to refine the understanding of Mygalomorphae delimitations and species recognition (Starret and Hedin 2007; Bond and Stockman 2008; Hendrixon et al. 2013; Hamilton et al. 2014). Furthermore, multiple kinds of evidence (e.g., molecular data, geographic distribution, behavioral aspects) may generate a more robust hypothesis of species when added to morphological characters (Montes de Oca et al. 2016). Grammostola is endemic to the subtropical region of South America (Ferretti et al. 2013). Currently, 20 valid species are distributed in Argentina, Bolivia, Brazil, Chile, Paraguay, and Uruguay (World Spider Catalog 2022). Although the identification of specimens and description of new taxa is challenging (Ferretti et al. 2011), studies related to taxonomy and natural history have been conducted in Argentina (Ferretti et al. 2011 2013) and Uruguay (Vol 2008; Montes de Oca et al. 2016). However, the specimens of Grammostola from Brazil have been scarcely explored since prominent research decades ago (Mello-Leitão 1921 1923; Bücherl 1951). The Brazilian species are distributed throughout the southern region of the country (Bücherl 1951), mainly in the state of Rio Grande do Sul, where six of them have been documented: Grammostola actaeon (Pocock, 1903), Grammostola anthracina (Koch, 1842), Grammostola iheringi (Keyserling, 1891), Grammostola pulchra Mello-Leitão, 1921, Grammostola pulchripes Simon, 1891, and Grammostola quirogai Montes de Oca, D’elía and Pérez-Miles, 2016 (Buckup et al. 2010; Malta-Borges et al. 2016). The descriptions of Grammostola species from Brazil made in the 19th and early 20th centuries are short and superficial compared to the current taxonomic descriptions for the genus. Furthermore, identifying taxa may be a tangled task for non-specialists since understanding species limits and the capacity to refine descriptions and diagnoses have changed throughout history (De Queiroz 2007). A way to improve the taxonomic scenario is updating the species descriptions under the scrutiny of the current methods, homology hypotheses, and terminologies. Grammostola pulchra was described a century ago (Mello-Leitão 1921). The original description is brief and vague, lacking essential features to the current taxonomy of the genus, such as the general aspect of spermathecae and morphology of palpal bulbs. Moreover, little additional information related to the taxonomic aspects of G. pulchra has been published. After a hundred years since the original description, we assessed type specimens and additional samples to update the taxonomic knowledge of G. pulchra. We used the somatic and genital morphology, geographical distribution, genetic distance, and phylogeny to build a robust redescription of G. pulchra. Diagnosis, illustrations of copulatory organs, distributional map, species delimitation based on a molecular marker, and phylogenetic discussion are provided. MATERIALS AND METHODS Morphological analyses and species redescription Distinct sources of data were used to assess multiple lines of support for the redescription. We considered the accumulation of evidence from each method to recognize a more robust hypothesis for the species. Here, we assumed the unified species concept, understanding species as a lineage evolving separately from other lineages (De Queiroz 2007). General character descriptions followed Montes de Oca et al. (2016) with some modifications— the spine classification method was made according to Petrunkevitch (1925), palpal bulb structures were classified according to Bertani (2000) and the classification of urticating hairs according to Cooke et al. (1972). All measurements are given in millimeters and were made with the Discovery V20 - Zeiss® stereomicroscope equipped with the Axiohome system. The pedipalp and leg measurements were taken from the dorsal aspect of the left side of the specimen. Photographs were taken with the stereomicroscope equipped with camera Leica® for microscopy, except photos of the live specimens and habitat that were made with Nikon P600 digital camera. Abbreviations: A = apical keel, ALE = anterior lateral eyes, AME = anterior median eyes, D = dorsal, P = prolateral, PI = prolateral inferior keel, PLE = posterior lateral eyes, page 2 of 15Zoological Studies 62: 5 (2023) © 2023 Academia Sinica, Taiwan PLS = posterior lateral spinnerets, PME = posterior median eyes, PMS = posterior median spinnerets, PS = prolateral superior keel, PV = proventral, R = retrolateral, RV = retroventral, V = ventral. The examined specimens belong to the following collections (the acronyms following Evenhuis 2021): MCNZ = Brazil, Rio Grande do Sul, Porto Alegre, Museu de Ciências Naturais da Fundação Zoo-Botânica do Rio Grande do Sul; MCTP = Brazil, Rio Grande do Sul, Porto Alegre, Pontifícia Universidade, Museu de Ciências; MZSP = Brazil, São Paulo, São Paulo, Museu de Zoologia da Universidade de São Paulo. Therefore, the identifications were double-checked using the original descriptions. The individuals studied were deposited into arachnological collections at the Museu de Ciências Naturais da Fundação Zoo-Botânica do Rio Grande do Sul (MCNZ) and Museu de Ciências e Tecnologia da Pontifícia Universidade Católica do Rio Grande do Sul (MCTP) and preserved in ethanol 80%. The species distribution map was made using the Arcgis 10.8.1 program, and the geographic coordinates were obtained using Google Earth (Lat/Long - WGS84). DNA extraction, amplification, and sequencing The DNA samples were obtained from ~80% ethanol-preserved species. The specimens had the metatarsus and tarsus removed containing muscle tissues from the third or fourth legs. The genomic DNAs were extracted using DNeasy Blood and Tissue kit (Qiagen, Valencia, CA, U.S.A.) according to the manufacturer’s instructions, but eluting to a final volume of 80 μL. The total genomic DNAs were stored at -20°C before amplification. The pair of primers used was designed by Folmer et al. (1994), targeting the “Folmer region” of the mitochondrial gene COI, LCOI 1490 (5'-GGTCAACAAATCATAAAGATATTGG-3') and HCOI 2198 (5'-TAAACTTCAGGGTGACCAA AAAATCA-3'). The PCR conditions were modified from Petersen et al. (2007), increasing the annealing temperature to 50°C/40 s, once amplified products using the recommended 47°C/1 m yielded non-specific bands when electrophoresed in the agarose gel. The PCR products were purified using Exonuclease I and shrimp alkaline phosphatase (Affymetrix, Inc. USB Products, Cleveland, OH, U.S.A.). Macrogen, Inc. (Seoul, South Korea) sequenced both DNA strands for all PCR products. Sequence chromatograms were visually inspected, verified, and manually edited using the Staden package (Staden et al. 2000). Sequences were verified using BLAST (http://blast.ncbi.nlm.nih. gov/Blast.cgi), confirming the high similarity of our submitted sequences to Grammostola species. Alignment and Analyses Additional COI sequences of Grammostola were obtained from GenBank, of which the access numbers are provided in table 1. The sample contained 29 terminal taxa, representing eight species of Grammostola, two of Aphonopelma Pocock, and one of Brachypelma Simon, besides our target species. Alignments of the sequences were performed using Mafft 7 (Katoh et al. 2017), online version (http://mafft. cbrc.jp/alignment/server/index.html), applying the strategy “Auto”. The appropriate substitution model was chosen using jModelTest v2.1.6 (Guindon and Gascuel 2003; Darriba et al. 2012) via the Bayesian information criterion (BIC) as suggested by Luo et al. (2010), also implemented in jModeltest 2.1.5. The GTR+I+G was selected for the COI matrix. Two probabilistic methods were used to build the COI gene tree, Maximum-Likelihood (ML) and Table 1. Species and individual GenBank access number of partial sequences of Cytochrome c oxidase subunit I of Theraphosidae species. G. pulchra* sensu Montes de Oca et al. 2016 Taxon Access numbers Aphonopelma hentzi JF803357.1 Aphonopelma moderatum JF803409.1 Brachypelma verdezi KT995351.1 Grammostola andreleetzi KT965218.1 Grammostola anthracina KT965238.1 Grammostola anthracina KT965250.1 Grammostola anthracina KT965255.1 Grammostola anthracina KT965263.1 Grammostola anthracina KT965267.1 Grammostola burzaquensis KT965249.1 Grammostola porteri KT022080.1 Grammostola pulchra * KT965207.1 Grammostola pulchra * KT965210.1 Grammostola pulchra * KT965211.1 Grammostola pulchra * KT965220.1 Grammostola pulchripes MG273517.1 Grammostola quirogai KT965256.1 Grammostola quirogai KT965268.1 Grammostola quirogai KT965274.1 Grammostola quirogai KT965275.1 Grammostola quirogai KT965276.1 Grammostola rosea KT022079.1 Grammostola rosea KT022081.1 Grammostola rosea KT022082.1 Grammostola rosea KT965257.1 Grammostola rosea KT965262.1 Grammostola pulchra OM670234 Grammostola pulchra OM670235 Grammostola pulchra OM670236 page 3 of 15Zoological Studies 62: 5 (2023) © 2023 Academia Sinica, Taiwan Bayesian inference (BI). These analyses allow two simultaneous inferences: the reciprocal monophyly of the sampled taxa; and a phylogenetic hypothesis for the relationship of the species. The maximum-likelihood analysis was performed on the matrix using RAxMLHPC2 at CIPRES Science Gateway (Miller et al. 2011) (www.phylo.org/portal2/), applying the substitution model GTR-CAT. Nodal support was assessed with automatic Stop Bootstrapping Automatically with Majority Rule Criterion (autoMRE). Bayesian inference of the matrix was performed in the multithreading version of the program MrBayes 3.2.0 (Ronquist and Huelsenbeck 2003), setting nst = 6 rates = invgamma for the marker; 5 million of generation (nruns = 2 nchains = 4) with trees sampled every 1000 generations. Tracer v.1.6.0 (Rambaut et al. 2014) was used to inspect the convergence to the stationary distribution of the chains. The first 25% of the generations were discarded as “burn-in”, and then the chains were combined. The combined ESS values for each parameter were higher than 200. The posterior probability (PP) was estimated for the remaining generations. Phylogenetic trees were visualized and edited using FigTree v1.4.0 (Rambaut et al. 2014) (http://tree.bio.ed.ac.uk/software/figtree/). Brachypelma verdezi Schmidt was used to root the trees in both analyses. For molecular species delimitation, two approaches were used: a distance-based approach called “assemble species by automatic partitioning” (ASAP) (Puillandre et al. 2021) available on the ASAP web (https://bioinfo.mnhn.fr/abi/public/asap/), setting Simple Distance (p-distances); and a tree-based approach, the Bayesian version of the Poisson Tree Processes model approach (bPTP) (Zhang et al. 2013) using Exelixis Lab’s web server (bPTP – http://species.h-its.org/ptp/) setting unrooted, 200000 MCMC generations, burn-in of 0.2. RESULTS (The taxonomy is based on the results of the integrative approach. See below) TAXONOMY Theraphosidae Thorell, 1869 Genus Grammostola Simon, 1892 Grammostola pulchra Mello-Leitão, 1921 (Figs. 1–21; Tables 2–3) Grammostola pulchra Mello-Leitão, 1921: 298. Grammostola pulchra Mello-Leitão, 1923: 198, figs. 66–68. Grammostola pulchripes pulchra Bücherl, 1951: 118, pl. IV (reduced to subspecies). Grammostola pulchripes pulchra Bücherl, 1957: 396, fig. 57. Grammostola pulchra Schmidt, 1986: 52, fig. 70. Grammostola pulchra Schmidt, 1993: 90, figs. 214, 221. Grammostola pulchra Schmidt, 1997: 16, figs. 55, 62. Grammostola pulchra Peters, 2000: 137, figs. 430–432. Grammostola pulchra Peters, 2003: 195, figs. 785–786, 790. Grammostola pulchra Schmidt, 2003: 167, figs. 382–383. Grammostola pulchra Montes de Oca, D’Elía & Pérez-Miles, 2016: 328, fig. 4E. Material examined Type material: BRAZIL: Rio Grande do Sul: Uruguaiana, IX.1914, 1 male and 1 female, E. Garbe leg. (MZSP 122) (examined). Additional material: 1 male, BRAZIL: Rio Grande do Sul: Capão do Leão, Campus Universitário da Universidade Federal de Pelotas [31°48'03.6"S, 52°25'18.7"W], 17.XI.2018, R. S. Pittella & P. G. Bassa leg. (MCTP 41831); 1 female, 20.XI.2018, R. S. Pittella leg. (MCTP 41833); 1 male, 01.XI.2018, R. S. Pittella leg. (MCN ARA-56829); 1 juvenile male, 20.XI.2018, R. S. Pittella leg. (MCTP 41832) matured in captivity; 1 female, 20.XI.2018, R. S. Pittella leg. (MCTP 43834); 1 male, 16.X.2018, R. S. Pittella & P. G. Bassa leg. (MCN ARA-56826); BRAZIL: Rio Grande do Sul: Pelotas, Road to Laranjal beach near Condomínio Veredas [31°45'27.5"S, 52°15'02.7"W], 1 male, 11.X.2018, R. S. Pittella leg. (MCN ARA-56827); Pelotas, Bairro Fragata [31°44'49.5"S, 52°22'51.7"W], 1 female, 20.X.2018, R. S. Pittella leg. (MCN ARA56828); BRAZIL: Rio Grande do Sul: São Borja, Reserva Biológica São Donato, 1 male, 10.X.2012, M. Machado leg. (MCTP 36905). Diagnosis: Grammostola pulchra differs from other Grammostola species by the following combination of characteristics: brownish black coloration; body thickly covered by long hairs of the same color but with yellowish or greyish tips, more abundant in the ventral region (Figs. 1–3, 17, 20–21); short broad-based spiniform setae on the prolateral coxal faces of legs I–IV (Fig. 5); male’s tibial apophysis, with two branches originating from the same base; primary branch smaller, straight, with a group of subapical macrosetae and presence of a long black spine with acuminate apex on the inner side, shorter than the branch; secondary branch larger with slight distal curvature and presence of a short broad base apical conical process; and 1–2 retrolateral tibial spines (Figs. 6–8). Male’s palpal tarsi with small tuft of erected and rigid spiniform setae in the apical region (Fig. 4). Piriform bulb with inferior prolateral keel (PI) and superior prolateral keel (PS) developed, short embolus with developed apical keel (A), slightly curved, folded in the middle portion and slightly curved in the apical page 4 of 15Zoological Studies 62: 5 (2023) © 2023 Academia Sinica, Taiwan portion (Figs. 9–11). Females have spermathecae with two short and straight seminal receptacles with rounded apex (Fig. 12). Grammostola pulchra resembles G. quirogai by the coloration but differs from them by their smaller size, the slender palpal bulb with developed apical keel and the morphology and spine combination of the tibial apophysis (illustrated by Montes de Oca et al. 2016); also could resemble G. burzaquensis in the general appearance and size, but differs due to the presence of short broad-based spiniform setae on prolateral coxal faces of legs I–IV, the morphology and spine combination of the tibial apophysis and the shape of palpal bulb (illustrated by Ferretti et al. 2011 2016). Additionally, it can be distinguished from G. anthracina by its small size, morphology and spine combination of the tibial apophysis and by the absence of reddish hairs on the ventral face of all legs (illustrated by Montes de Oca et al. 2016). Redescription: Male (MCTP 41831): Total body length, excluding chelicerae and spinnerets: 38.36. Carapace length: 15.39, width: 14.18, with slightly elevated cephalic area and shallow thoracic striae (Fig. 2). Chelicerae length: 6.12, width: 4.54. Abdomen length: 21.86, width: 14.98. Anterior eyes row recurve, posterior procurve. Eyes size and interdistances: PME: 0.32; PLE: 0.49; AME: 0.40; ALE: 0.49; PME–PME: 1.26; PME–PLE: 0.13; AME–AME: 0.49; AME–ALE: 0.23; ALE–PLE: 0.25. Eye tubercle length 1.98, width 2.41; clypeus: 0.13. Fovea shape: transverse and straight (Fig. 2); width: 1.63. Labium length: 2.45, width: 3.27, with 125 cuspules (Fig. 3). Maxillae length 5.13, width: 3.15, with 232 cuspules arranged in a triangular group with the base at the proximal edge (Fig. 3). Sternum length 7.04, width 7.22; posterior angle does not separate the coxae IV (Fig. 3). Sigillae: anterior: 1 pair not much evident, medians: 1 pair; posterior: 1 pair; all ellipsoid, submarginal (Fig. 3). Chelicera with 8 promarginal and 4 retromarginal teeth. Setae - stridulatory: present on the retrolateral face of the palpal maxilla and on the opposite prolateral face of leg I, Fig. 1. Maximum likelihood tree based on the partial sequence of Cytochrome c oxidase subunit I of Grammostola from South America and related Theraphosidae genus. Capital letters above nodes refer to lineages discussed in the text. Numbers close to nodes are the Bayesian posterior probabilities (PPs)/maximum likelihood bootstrap support (ML), respectively. Only nodal support above PP = 0.5 or ML = 50 is displayed (‘*’ indicates lower support values). Lineage assignments of distance-base (ASAP1 and ASAP2 means first and second best results of ASAP, respectively) and tree-based (bPTP) methods. Habitus photos of male and female of Grammostola pulchra. G. pulchra* means sensu Montes de Oca et al. 2016. page 5 of 15Zoological Studies 62: 5 (2023) © 2023 Academia Sinica, Taiwan claviform, located both above and below the suture, the upper ones occupy the apical area and slightly longer and numerous, the lower ones shorter and less numerous occupying a basal area with interspersed spiniform setae; palpal tarsi setae: small tuft of erect and rigid spiniform setae in the tarsal apical region (Fig. 4); coxal setae: broad-base short spiniform on the prolateral face of the legs I–IV (Fig. 5). Tarsi I–IV densely scopulate, scopula entire; metatarsi: I completely scopulate, II scopula on apical two thirds, III scopula on apical third, IV minimally scopulate on apical third. Tibia I with paired distal proventral apophyses, composed of two branches originating from the same base; secondary branch larger with slight distal curvature and presence of a short broad-based apical conical process; primary branch smaller, straight, with a group of subapical macrosetae and presence of a long black spine with an acuminate apex on the inner side, shorter than the branch; and 1–2 retrolateral tibial spines (Figs. 6–8). Flexion of metatarsus I retrolateral in relation to tibial apophysis. Palpal organ pyriform with prolateral keels present, PI and PS developed, short embolus with developed apical keel (A), slightly curved, folded in the middle portion and slightly curved in the apical portion (Figs. 9–11). Length of leg and palpal segments in table 2; Legs formulla: IV–I–II–III. Spination: Femora: palp 0; I 1P; II 1–0–1 P; III 1R; IV 1R. Patella: palp and I– IV 0. Tibia: palp 1 P, 1–0–1 PV; I 1–0–1 V, 1 PV, 2 R; II 1–0–1 P, 1–0–3–0–1 V; III 2–0–1 P, 1–0–1 R, 1–0–1 V, 1–0–1–0–2 PV, 1 RV; IV 1 P, 1–0–1 R, 1–0–2 V, 1 PV, 1 RV. Metatarsus: I 1 R, 1 PV, 2 RV; II 2–1 RV; III 1–1–1 P, 1–1–1 R, 1–1–1–1 PV, 1–1–1–2 V, 1–1 RV; IV 1–1–1 P, 1–1–1–1 R, 2–1–1–1 PV, 1–1–2 RV. Tarsus: palp and I–IV 0. Color (in vivo): brownish black body with long hairs of the same color with yellowish or greyish tips distributed over the cephalothorax, abdomen and legs, more abundant in the ventral region. Sternum, labium and coxae velvety blackish. Coxae of palp and chelicerae with reddish bristles. Presence of two slightly marked vertical striae on the patellae (Figs. 1–3, 20). In alcohol, the specimens have a dark brown color with yellowish bristles scattered throughout the body. Type III–IV urticating hairs present, gathered in the dorsal region of the abdomen, where they form a silver spot with semicircular shape. Spinnerets: PMS length: 1.97, monoarticulated; PLS length of articles: basal: 2.75, medial: 2.44, apical: 2.93, total: 8.12, triarticulated with apical segment digitiform. Variation (range (mean ± standard deviation)): Four adult males: total length 31.71–38.36 (34.93 ± 2.68); cephalothorax length 15–16.93 (15.66 ± 0.74), width 13–16.24 (14.19 ± 1.57); legs: I 46.05–51.72 (48.92 ± 2.43), II 44.25–49.24 (46.22 ± 1.91), III 42.51– 45.28 (44.15) ± 1.14), IV 53.42–56.67 (55.26 ± 1.17), palp 22.82–28.55 (24.61 ± 2.29). Female (MCTP 41833): Total body length, excluding chelicerae and spinnerets: 50.08. Carapace length 17.61, width 16.86, with slightly elevated cephalic area and shallow thoracic striae (Fig. 2). Chelicerae length 8.63, width 5.63. Abdomen length 29.96, width 25.79. Anterior eyes row recurve, posterior procurve. Eyes size and interdistances: PME: 0.29; PLE: 0.53; AME: 0.40; ALE: 0.57; PME–PME: 1.49; PME–PLE: 0.29; AME–AME: 0.70; AME–ALE: 0.47; ALE–PLE: 0.47. Eye tubercle length 1.87, width 2.82; clypeus: 0.18. Fovea shape: transverse and straight (Fig. 2); width 2.29. Labium length 2.46, width 4.41, with 210 cuspules. Maxillae length 6.44, width 4.44, with 233 cuspules arranged in a triangular group with the base at the proximal edge. Sternum length 8.28, width 8.29, posterior angle does not separate the coxae IV (Fig. 3). Sigillae: anterior: 1 pair not much evident, medians: 1 pair, posterior: 1 pair, all ellipsoid, submarginal (Fig. 3). Chelicera with 8 promarginal and 5 retromarginal teeth. Setae - stridulatory and coxal setae: as in male. Tarsi I–IV densely scopulate, scopula entire; metatarsi: I completely scopulate, II scopula on apical two thirds, III scopula on apical third, IV minimally scopulate on apical third. Length of leg and palpal segments in table 3; Legs formula: IV–I–II–III. Spination: Femora: palp 0; I 1 P; II 1 P; III 1 P; IV 1 R. Patella: palp and I–IV 0. Tibia: palp 1P, 1 R, 1 V, 1–0–2–1 PV, 1–0–2 RV; I 1 R, 1 PV, 1–0–1 RV; II 1 PV, 1–0–1–0–1 RV; III 1–1 P, 1–1–1 R, 2 V, IR 1 R, 1 V, 1 PV, 1 RV. Metatarsus: I 1–0–3 V, 1 PV, 2 RV; II 1 P, 3 V, 1 PV, 2–1 RV; III 1–1–1 P, 1–1 R, 1–1–0–5 V, 1–0–2–1–1 PV, 1–1 RV; IR 1–0–1 P, 1–2–1–1–2 R, 1–0–1–0–1 V, 1–0–1–0–2 PV, 2–2 RV. Tarsus: palp and I–IV 0. Color: as in male (Figs. 1, 17, 21). Type III-IV urticating hairs present, gathered in the dorsal region of the abdomen, where they form a silver spot and semicircular shape. Spinnerets: PMS length: 2.09; monoarticulated; PLS length of articles: basal: 3.72, medial: 2.58, apical: 2.91, triarticulated with apical segment digitiform. Spermathecae with two short and straight seminal receptacles with rounded apex (Fig. 12). Distribution and natural history: G. pulchra is known for the western region (between the municipalities of Uruguaiana, São Borja and Maçambara) and for the southern region (which includes the municipalities of Capão do Leão and Pelotas) of the Pampa biome in the state of Rio Grande do Sul (Fig. 13), where occurs in rocky environments or near humid areas locally known as “banhados”. In the municipality of Pelotas, the individuals were found in burrows dug in the ground in open field areas with predominant herbaceous vegetation located between areas with anthropic modifications, such as the presence of a grove page 6 of 15Zoological Studies 62: 5 (2023) © 2023 Academia Sinica, Taiwan Figs. 2–8. Morphological characters of Grammostola pulchra. Male (MTCP 41831), (2) carapace and chelicerae; (3) sternum, maxillae, labium and coxae; (4) palpal tarsi, prolateral view, arrow indicate the small tuft of spiniform setae; (5) prolateral face of coxa I; (6–8) Right leg I tibial apophysis, (6) ventral view, (7) prolateral view, (8) small branch with spine on the inner side and group of subapical macrosetae. Scale bars: 2–3 = 5 mm; 4–8 = 1 mm. page 7 of 15Zoological Studies 62: 5 (2023) © 2023 Academia Sinica, Taiwan of Eucalyptus spp. and a native wetland region (Figs. 14–17). The burrows have circular entries and depths ranging from 50 to 80 cm (Fig. 16). Besides that, spiders were observed in remarkable agglomeration, with many shelters in a few square meters (Fig. 15). Additionally, G. pulchra occurs in sympatric distribution with other theraphosid genera, Eupalaestrus Pocock, 1901 and Catumiri Guadanucci, 2004 in their southern habitats, being also reasonable consider that it shares its habitat with G. quirogai by their western distribution in Rio Grande do Sul state. Remarks and affinities: Ferretti et al. (2013) mention a small bristle of erected and stiff setae on the palpal tarsi in males of Grammostola diminuta Figs. 9–12. Copulatory organs of Grammostola pulchra. (9–11): Male (MTCP 41831), (9) left palpal bulb retrolateral view, (10) left palpal bulb prolateral view, (11) embolus apex detail, arrow indicates the developed apical keel. (12) Female (MTCP 41833), spermathecae ventral view. Scale bars: 9, 10, 12 = 1 mm; 11 = 0.1 mm. page 8 of 15 Zoological Studies 62: 5 (2023) © 2023 Academia Sinica, Taiwan Ferretti, Pompozii, González & Pérez-Miles, 2013 as a diagnostic feature compared to all other species of the genus. However, a very similar structure was found in G. pulchra (Fig. 4), showing that this set of setae may not be unique to G. diminuta males. Thus, the analysis of this type of structure in other species may be relevant to understand the evolution of morphological characters and propose homologies between species of the genus. Molecular analyses The COI sequence was obtained for a total of three specimens of Grammostola pulchra, generating 650 bp sequences (see Material examined section: MCTP 41832, MCTP 41833, and MCN ARA-56826; Figs. S1–S3 for voucher pictures). The other specimens had low-quality DNA extraction, probably due to the Table 2. Length (mm) of legs and palpal segments of the male (MCTP 41831) of Grammostola pulchra I II III IV Palp Femur 13,67 12,90 11,61 13,12 8,84 Patella 7,53 7,11 6,26 7,09 4,78 Tibia 10,58 9,44 8,11 11,66 6,90 Metatarsus 8,00 7,98 9,87 13,20 - Tarsus 6,27 6,82 6,66 8,35 2,76 Total 46,05 44,25 42,51 53,42 23,28 Table 3. Length (mm) of legs and palpal segments of the female (MCTP 41833) of Grammostola pulchra I II III IV Palp Femur 14,44 12,37 10,86 13,88 10,07 Patella 7,86 7,17 6,45 7,24 5,66 Tibia 10,01 8,28 7,41 10,05 6,26 Metatarsus 8,05 8,09 9,03 13,04 - Tarsus 5,94 5,73 5,78 7,06 6,78 Total 46,39 41,64 39,53 51,27 28,77 Fig. 13. Distribution map of Grammostola pulchra. N page 9 of 15Zoological Studies 62: 5 (2023)