Phylogeny and taxonomic revision of Heteropachylinae (Opiliones: Laniatores: Gonyleptidae)zoj_

Zoological Journal of the Linnean Society, 2011, 163, 437 483. With 28 figures Phylogeny and taxonomic revision of Heteropachylinae (Opiliones: Laniatores: Gonyleptidae)zoj_706 437..483 AMANDA CRUZ MENDES* Museu Nacional Universidade Federal do Rio de Janeiro Departamento de Invertebrados, Quinta da Boa Vista s/n, São Cristóvão, Rio de Janeiro 20940-040, Brazil Received 19 May 2010; revised 25 September 2010; accepted for publication 28 September 2010 Heteropachylinae Kury, 1994 is currently considered an early lineage of Gonyleptidae Sundevall, 1833 and includes small harvestmen that occur in the Atlantic Forests, mainly in the Northeast Region of Brazil. The species of Heteropachylinae were spread in different subfamilies until the establishment of the group. As a result of that, until now there were no comparative works on the subfamily in the literature. In this study, the first systematic review for Heteropachylinae was carried out and the first phylogenetic hypothesis for their taxa is proposed, by means of a cladistic approach. A total of 102 morphological characters was coded for 21 terminal taxa: 11 species of Heteropachylinae, two species from the genus Mangaratiba Mello-Leitão, 1940 (considered as Heteropachylinae before this paper), seven other gonyleptids, and one cosmetid. The data were analysed using heuristic search algorithms, resulting in one most parsimonious tree (280 steps, consistency index = 0.47, retention index = 0.63). A parsimony analysis under implied weights was also performed, resulting in three trees. According to the results, Heteropachylinae is a monophyletic and well-supported group if Mangaratiba is excluded. This genus should be returned to where it was originally described Pachylinae. The subfamily distribution is restricted from Ceará to Espírito Santo Brazilian states. Pseudopucrolia Roewer, 1912 is considered a senior synonym of Melloa Roewer, 1930, Thaumatoleptes Roewer, 1930, and Tribunosoma Roewer, 1943. Two species of Heteropachylus Roewer, 1913, Heteropachylus ramphonotus sp. nov. and Heteropachylus gracilis sp. nov., from Bahia and Espírito Santo Brazilian states, are newly described. Heteropachylus is considered a senior synonym of Chavesincola Soares & Soares, 1946 and Aesotrinoma H. Soares, 1977. A new genus is also described for a new species from Bahia and Espírito Santo Brazilian states, Magnispina neptunus gen. et sp. nov. Lectotypes are designated for the following species: Pseudopucrolia incerta (Mello-Leitão, 1928) comb. nov., Pseudopucrolia discrepans (Roewer, 1943) comb.nov., Pseudopucrolia rugosa (Roewer, 1930) comb. nov., Heteropachylus inexpectabilis (Soares & Soares, 1946) comb. nov. and Heteropachylus spiniger Roewer, 1913.. doi: 10.1111/j.1096-3642.2011.00706.x ADDITIONAL KEYWORDS: Arachnida Atlantic rain forest Grassatores male polymorphism Neotropics Pachylinae. INTRODUCTION Heteropachylinae Kury, 1994 was proposed in a study of the early lineages of Gonyleptidae Sundevall, 1833 (Kury, 1994), in which two other subfamilies were also newly proposed (Cobaniinae Kury, 1994 and Metasarcinae Kury, 1994). The subfamily was erected *E-mail: amanda.mendes@gmail.com based on the synapomorphic free tergites with projected corners and more or less fused with scutum in males (Kury, 1994). Until the establishment of Heteropachylinae, its species were distributed in different subfamilies of Gonyleptidae and Phalangodidae Simon, 1879. This fact in part justifies the lack of any comparative study of the species of the subfamily in the literature (Mendes & Kury, 2003). Besides the allocation of its species to different taxonomic groups, the 437

438 A. C. MENDES predominance of monotypic genera is another problem associated with the classificatory model proposed by Roewer (e.g. 1923) and followed by other authors (e.g. Mello-Leitão, 1932; Soares & Soares, 1946, 1949, 1954). Most of the available descriptions, redescriptions, and illustrations in the literature are insufficient to recognize the species. The male genitalia have been illustrated only for two species (Soares, 1977; Mendes & Kury, 2003); thus, until now, there has been no characterization and inquiry on the taxonomic importance of this structure in the group. There are in the literature indications of relationship between some of its genera based on morphological resemblance (Roewer, 1930; Soares, 1968, 1977; Mendes & Kury, 2003), but none includes a detailed study comparing the species of the group. Since these species were grouped by Kury (1994) little work has been carried out on the taxonomy of the subfamily, so there were no comparative descriptions and diagnoses, or a key to the genera and species of Heteropachylinae until now. In the present work the monophyly of Heteropachylinae was tested and the first hypothesis of phylogeny proposed for its taxa using a cladistic paradigm, including a new classification of the subfamily congruent with the results obtained. Redescriptions and illustrations are also provided for all species and a new genus and three new species are proposed and described. SYSTEMATIC BACKGROUND Until this paper the distribution of Heteropachylinae was restricted to the remnant Atlantic Forest from Ceará to Paraná Brazilian states and had 11 species grouped in eight genera (Kury, 1994, 2003a; Mendes & Kury, 2003). In a paper on the arthropods collected in the Bavarian expedition to Brazil at the beginning of the 19 th century, Perty (1833) described the first species of the group, Eusarcus muticus, in a description of few features without illustrations. Koch (1839a) transferred this species to Gonyleptes Kirby, 1818 and gave a more detailed description with a coloured figure, but little information. Later, Roewer (1913a) created Heterogonyleptes for this species, placing it in Phalangodidae Tricommatinae Roewer, 1912. In those times, Laniatores without pseudonychium sensu Roewer (1923) were placed in Phalangodidae and the species with distitarsus I trimerous (as opposed to bimerous) were assigned to Tricommatinae [much later transferred to Gonyleptidae (Kury, 2003a)]. All the other species were described (and redescribed) under Gonyleptidae during the 20 th century by the prominent arachnologists of that time: Carl Friedrich Roewer, Cândido de Mello-Leitão, Benedicto Soares, and Helia Soares (Roewer, 1913a,b, 1930, 1943; Mello-Leitão, 1928; Soares & Soares, 1946, 1974; Soares, 1977). Following the Roewerian method, those species with four mesotergal areas were described under Pachylinae Sørensen, 1884 (Chavesincola inexpectabilis Soares & Soares, 1946; Heteropachylus peracchii Soares & Soares, 1974; Heteropachylus spiniger Roewer, 1913; Mangaratiba acanthoprocta (H. Soares, 1968); Mangaratiba angulispinosa (H. Soares, 1966); Mangaratiba monstrosa Mello-Leitão, 1940; Tribunosoma discrepans Roewer, 1943 and those with three mesotergal areas under Gonyleptinae Sundevall, 1833 (Aesotrinoma crassicalcanei H. Soares, 1977; Melloa incerta (Mello- Leitão, 1928); Thaumatoleptes rugosus Roewer, 1930). Except for H. peracchii, all the species were originally described or later transferred (M. incerta) to monotypic genera. The standard of species descriptions also followed the pattern implemented by Roewer and are poorly illustrated, most of them with just a drawing of the dorsum without the appendages. The fusion of the free tergites to the dorsal scutum in males, a great conspicuous characteristic and considered a synapomorphy of the group (Kury, 1994; this paper) was commented on by Roewer (1930: 405, 407, 1943: 25) in few of his descriptions/ redescriptions, and by Soares & Soares (1946) in the description of C. inexpectabilis. In the latter, the authors made a curious comment at the end of the genus diagnosis. They said that if this fact was constant in both sexes and in many exemplars, it should result in the formation of a new family for the species. In spite of recognizing the proximity of the genera Thaumatoleptes Roewer, 1930, Triaenosoma Roewer, 1913 (= Pseudopucrolia Roewer, 1912), and Melloa Roewer, 1930 based on that characteristic, Roewer (1930), continuing with the system created by himself, never put them in them same group, leaving this information lost and being rescued only more than six decades later (Kury, 1994). Studying the phylogeny of the early lineages of Gonyleptidae, Kury (1994) established Heteropachylinae for nine genera that were in different groups: Aesotrinoma H. Soares, 1977; Canopilio H. Soares, 1968; Chavesincola Soares & Soares, 1946; Heterogonyleptes Roewer, 1913 (= Pseudopucrolia = Triaenosoma = Pseudotriaenosoma Mello-Leitão, 1927); Heteropachylus Roewer, 1913; Mangaratiba Mello-Leitão, 1940; Melloa; Thaumatoleptes; and Tribunosoma Roewer, 1943. Later, Kury (2003a) added the genus Pseudopucrolia and proposed new synonymies totalling eight genera and 11 valid species (see Systematic section for the current classification). Mendes & Kury (2003) redescribed Thaumatoleptes rugosus, proposing a paradigm of description of the structures for the species of Heteropachylinae. In that

REVISION OF HETEROPACHYLINAE 439 paper, the proximity of the genera Melloa, Pseudopucrolia, Thaumatoleptes, and Tribunosoma was suggested based on the morphology of the free tergites and femur IV. MATERIAL AND METHODS MATERIAL EXAMINED For this study, 708 specimens were examined. The material of Heteropachylinae is listed before the description of each species in the Systematic results section. The outgroup material examined is listed in Table 1. The collections cited in this paper are listed below with their respective acronyms. COLLECTION ABBREVIATIONS HEMS, Collection Helia Eller Monteiro Soares. Formerly a private collection, now kept by MNRJ, Universidade Federal do Rio de Janeiro. Rio de Janeiro. Brazil; IBSP, Instituto Butantan. Laboratório de Artrópodes Peçonhentos, São Paulo, Brazil; MACN, Museo Argentino de Ciências Naturales Bernardino Rivadavia, Buenos Aires, Argentina; MNRJ, Museu Nacional, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil; MZSP, Museu de Zoologia da Universidade de São Paulo, São Paulo, Brazil; SMF, Naturmuseum Senckenberg Sektion Arachnologie, Frankfurt am Main, Germany; ZSMC, Zoologische Staatssammlung München, München, Germany. CLADISTIC ANALYSIS The ingroup consists of all the currently valid species of Heteropachylinae, plus three undescribed species. The outgroup taxa correspond to seven species representing six subfamilies of Gonyleptidae and one representative of Cosmetidae. Table 1 provides the list of species used as outgroups in the analysis, indicating the material used. A total of 102 characters was coded for 21 terminal taxa, with the data matrix (Appendix) constructed in MESQUITE (Maddison & Maddison, 2009). The characters are based almost exclusively on males because the females of Heteropachylinae, as in many other Gonyleptidae subfamilies, are highly homogeneous and lack complex structures in the free tergites, coxa, trochanter, and femur IV. These structures, besides those from male genitalia, compose 51% of the characters in the present analysis. The characters were treated as unordered and the polarization was performed a posteriori, resulting from the rooting of the diagrams obtained by the parsimony analysis (Nixon & Carpenter, 1993). The autapomorphies were included because of their great descriptive value and to evaluate morphological gaps that could justify arbitrary decisions of cuts of taxonomical categories. The coding for the terminals is shown in the Appendix. An exact analysis using implicit enumeration was performed by the program TNT (Goloboff, Farris & Nixon, 2008b). The branch-support was evaluated through absolute Bremer support (Bremer, 1994), relative Bremer support (RBS, relative fit difference; Goloboff & Farris, 2001), and bootstrap (Felsenstein, 1985), both performed by TNT. The Bremer support was calculated by searching for suboptimal trees, increasing the number of suboptimal trees in steps. This was carried out to make sure that the buffer would not be filled just by highly suboptimal trees, which would cause an overestimation of the support. The bootstrap analysis was performed with 5000 replicates with tree bisection-reconnection branch swapping. The distribution and optimization of the characters on the trees obtained were studied through the program WINCLADA (Nixon, 1999 2002). Table 1. List of vouchers of outgroups used in the cladistic analyses Family/subfamily Species Depository Cosmetidae Cynorta conspersa MNRJ 4560 Gonyleptidae/Metasarcinae Tschaidicancha weyrauchi CAS Gonyleptidae/Borguyiinae Borguyia bocaina MNRJ 8977 Gonyleptidae/Cobaniinae Cobania picea MNRJ 06791 Gonyleptidae/Caelopyginae Pristocnemis pustulatus MNRJ 05474 Gonyleptidae/Goniosomatinae Goniosoma varium MNRJ 6929 Gonyleptidae/Pachylinae Discocyrtus infelix MNRJ 5698 Gonyleptidae/Pachylinae Pachylus chilensis MNRJ Gonyleptidae/Pachylinae* Mangaratiba monstrosa MNRJ 17533 Gonyleptidae/Pachylinae* Mangaratiba acanthoprocta MZSP 7870, Holotype *Although Mangaratiba was initially part of the ingroup, it was included here because it was excluded from Heteropachylinae.

440 A. C. MENDES Besides the equal-weighted analysis, I also performed an analysis under implied weights (Goloboff, 1993) via TNT, which attributes higher values to characters which better fit (fi) the trees. There may not be an optimal value for k, but for data where the most homoplastic characters rarely exceed ten extra steps, k = 5 16 can be considered adequate (Goloboff et al., 2008a). I performed analyses with k = 1 6, 15, and 20 to observe the effect of the constant variation over the topology of the trees. These results are summarized as space plots ( Navajo rugs ). Other authors have also estimated the stability of results under different concavity values (k) in phylogenies using morphological data (Prendini, 2003; de Bivort & Giribet, 2004). The evaluation of the stability of results under different parameters, such as different concavity values (k), is comparable to a sensitivity analysis (Wheeler, 1995), as used in molecular analyses. The analyses under implied weights were also performed using implicit enumeration. MEASUREMENTS AND TERMINOLOGY Measurements are given in millimetres. Whenever possible, they are given for the holotype, followed by medians of other exemplars (males and females) with minor and maximum values in parentheses. Abbreviations: CL, carapace length; MCW, maximum carapace width; ASL, abdominal scutum length; MASW, maximum abdominal scutum width; FL, femur IV length. The descriptions and illustrations were created following the pattern of Mendes & Kury (2003). The spines of the tibia and tarsus of pedipalps are coded as I for long spine and i for short spine as is usual in the harvestman literature, but I used brackets to indicate spines that come from the same socket and a plus signal to indicate a gap between the spines. GEOGRAPHICAL DISTRIBUTION The primary sources for the geographical distribution of the species were the labels of the analysed material and literature records. Some doubtful literature records were not included, but they were mentioned in the notes after description of each species. The given distributions for the species have reference to the Global Ecoregions adopted by the World Wild Fund for Nature (Olson & Dinerstein, 2002) CLADISTIC ANALYSIS CHARACTERS A list of the characters used in this analysis is given below with their respective length (L), consistency (CI), and retention (RI) indexes (relative to the analysis with equal weights, Fig. 1), and also comments when relevant. As females do not bear well-developed armatures on free tergites and leg IV, characters relative to those structures are based only on males (characters 24 39 and 62 90). MORPHOLOGY OF DORSUM AND VENTER 1. Carapace, lateral of anterior margin, ornamentation: (0) absent; (1) present. L = 3; CI = 0.33; RI = 0.60. 2. Carapace, lateral of anterior margin, kind of ornamentation: (0) small granules (Fig. 5A); (1) tubercles (Fig. 19A). L = 2; CI = 0.50; RI = 0.80. In this character, state (0) is the plesiomorphic state, and the appearance of tubercles, state (1), was considered as a synapomorphy of Heteropachylus and an independent autapomorphy of Magnispina gen. nov. in delayed transformation (DELTRAN) optimization. 3. Carapace, frontal swelling, granules: (0) absent; (1) present (Fig. 5A). L = 2; CI = 0.50; RI = 0.75. 4. Eye mound armature (tubercles or spines): (0) absent; (1) present (Fig. 3C). L = 3; CI = 0.33; RI = 0. 5. Eye mound, number of armature/ ornamentation: (0) two (Fig. 3C); (1) one. L = 3; CI = 0.33; RI = 0. The unpaired median armature, state (1), just occurs in the outgroup independently in Pachylus chilensis (Gray, 1833), Cobania picea (Bertkau, 1880), and Borguyia bocaina Yamaguti & Pinto-da-Rocha, 2009, in DELTRAN optimization. However, from an accelerated transformation (ACCTRAN) point of view it could be considered independently as an autapomorphy of Borguyia bocaina and a synapomorphy of (P. chilensis + C. picea) + Heteropachylinae, with a reversion to state (0) in Heteropachylinae. 6. Eye mound, kind of pair of ornamentation: (0) granules (Fig. 5C); (1) spines (Fig. 3C). L = 4; CI = 0.25; RI = 0.25. 7. Eye mound, shape: (0) saddle-shaped; (1) convex (Fig. 3C). L = 3; CI = 0.33; RI = 0.33. 8. Ozopore, integumentar dome: (0) absent; (1) present (Fig. 3A, E). L = 1; CI = 1; RI = 1. This character is modified from morphological character I from Hara & Gnaspini (2003: 268), just considering absence/presence, dismissing the position of the dome (present in the original character) as it should constitute another character not relevant to this analysis. 9. Ozopore, integumentar dome, V-shaped cut: (0) absent; (1) present (Fig. 3E). L = 3; CI = 0.33; RI = 0.60. This character corresponds to morphological character II from Hara & Gnaspini (2003: 268).

REVISION OF HETEROPACHYLINAE 441 Cynorta conspersa Cosmetidae Tschaidicancha acanthoma Metasarcinae Borguyia bocaina Borguyiinae Pristocnemis pustulatus Caelopyginae 7/38 Discocyrtus infelix Pachylinae 88 2/20 Goniosoma varium 2/20 3/20 Goniosomatinae 11/45 Mangaratiba acanthoprocta Pachylinae 99 Mangaratiba monstrosa 4/20 1 4 10 k 2 5 15 3 6 20 2/20 4/20 Pachylus chilensis Pachylinae Cobania picea Cobaniinae Magnispina neptunus 1/14 6/25 Pseudopucrolia incerta 58 2/26 P. discrepans 59 1/24 P. mutica 12/41 P. rugosa 96 1/14 1/14 Heteropachylus ramphonotus 3/23 H. inexpectabilis 54 H. crassicalcanei 1/14 5/28 89 H. gracilis H. peracchii H. spiniger Heteropachylinae Gonyleptidae Figure 1. The single most parsimonious tree (280 steps, consistency index = 0.47, retention index = 0.66) resulting from analysis under equal weights. Clade support is indicated above (absolute/relative Bremer support values) and below (bootstrap proportions) branches. Sensivity plots ( Navajo rugs ) indicate the recovery of the nodes in the analysis under implied weights with different values of k (black squares indicate monophyly; white squares indicate non-monophyly). 10. Ozopore, posterior opening: (0) absent (Fig. 3E); (1) present. L = 3; CI = 0.33; RI = 0.66. This character is modified from the morphological character III from Hara & Gnaspini (2003: 268), just considering absence/presence, dismissing the size of the opening (present in the original character) as once more it should constitute another character not relevant to this analysis. 11. High density of dark granules covering dorsal surface of the body: (0) absent; (1) present. L = 2; CI = 0.5; RI = 0. 12. Groove between mesotergal areas III and IV: (0) complete, four areas (Fig. 15A, B); (1) absent, three areas (Fig. 13A); (2) incomplete (Figs 3A, 11). L = 6; CI = 0.33; RI = 0.50. The number of mesotergal areas was used by Roewer to define groups (e.g. Roewer, 1923), and it is still used to differentiate Gonyleptinae from Pachylinae, as none of both subfamilies currently has a reasonable diagnosis. This character varies a lot inside Gonyleptidae. In this analysis the condition of state (0), four defined areas, was the plesiomorphic state. The fusion of areas III and IV can be presumed as long as the species of Pseudopucrolia shows a vestige of this fusion in area III, state (2) (although some specimens do not show the vestige). The fusion of the original areas III and IV in Heteropachylinae occurs independently in two more species, Heteropachylus ramphonotus sp. nov. and Heteropachylus crassicalcanei. 13. Mesotergal area I, longitudinal division: (0) present, forming right and left halves (Fig. 3A); (1) absent, entire (Fig. 13A). L = 2; CI = 0.50; RI = 0. 14. Mesotergal area II, pair of paramedian granules, size in relation to the other granules of the area: (0) same size (sometimes absent) (Fig. 11); (1) larger (Fig. 17A). L = 7; C = 0.14; RI = 0.14. The presence of a pair of granules on the areas can be a subjective character, as when the area is entirely granulous the absence or presence could

442 A. C. MENDES be misinterpreted. Additionally, the granules might not be seen if very small. I just considered the pair of paramedian granules as present if it was the only ornamentation of the mesotergal area or if visibly larger, in the case of ocurring with others. The phylogenetic signal of this character was very weak, showing that it is not an adequate character on which to rely. In both optimizations state (1) was synapomorphic for Gonyleptidae with a reversion to state (0) in the most inclusive clade including Mangaratiba and Heteropachylinae. In the DELTRAN optimization, state (1) evolved independently in the species of Heteropachylinae in which it occurs. In the ACCTRAN optimization, state (1) occurred independently in H. spiniger, H. crassicalcanei, and Magnispina gen. nov. + Pseudopucrolia. 15. Mesotergal area II surface: (0) smooth (Fig. 5A); (1) granulous (Fig. 19A). L = 6; CI = 0.16; RI = 0.37. 16. Groove between mesotergal areas II and III/ III + IV: (0) complete (Figs 3A, B, 15A, B); (1) incomplete (Fig. 13A). Uninformative. 17. Mesotergal area III, number of armature: (0) pair of tubercles or spines (Fig. 3A, B); (1) one tubercle or spine (Figs 13A, 15A, B). L = 2; CI = 0.50; RI = 0.66. 18. Mesotergal area III, aspect of surface: (0) smooth (Fig. 5A, B); (1) granulous (Fig. 19A, B). L = 4; CI = 0.25; RI = 0.40. 19. Mesotergal area III, level in relation to the other areas: (0) same (Fig. 3A, E); (1) elevated (Fig. 13A, D). Uninformative. 20. Mesotergal area IV, longitudinal division: (0) present, forming right and left halves, divided (Fig. 15A, B); (1) absent, entire (Fig. 19A, B). L = 4; CI = 0.25; RI = 0. 21. Mesotergal grooves, depth: (0) deep (Fig. 3B, E); (1) shallow (Fig. 5B, E) (superficial). L = 3; CI = 0.33; RI = 0. 22. Dorsal scutum, surface of lateral area: (0) smooth; (1) with row of tubercles (Fig. 3A, E). L = 1; CI = 1; RI = 1. 23. Dorsal scutum, shape: (0) more or less rounded, as in most cosmetids (e.g. Cynorta, Metavononoides); (1) rectangular; (2) pyriform (Fig. 3A). L = 2. Uninformative. 24. Free tergites in males, degree of fusion to the dorsal scutum: (0) without fusion; (1) free tergites I and II fused to the dorsal scutum (Fig. 3A, E); (2) free tergites I, II, and III fused to the dorsal scutum (Fig. 3A, D). L = 2; CI = 1; RI = 1. This character, along with 25, 30, and 38, is equivalent to character 13 of Kury (1994). In that character the author put information on the lateral armature and fusion of the free tergites together. I chose to treat these as independent characters. 25. Free tergite I, corners (lateral margin), armature: (0) absent (Fig. 5A); (1) present (Fig. 23A). L = 3; CI = 0.33; RI = 0.66. 26. Free tergite II, median armature: (0) absent (or much reduced) (Fig. 11); (1) present (Fig. 3A). L = 3; CI = 0.33; RI = 0.75. 27. Free tergite II, median apophysis, shape: (0) conic (Fig. 3A, B, E, F); (1) square-like. Uninformative. 28. Free tergite II, median apophysis, length in relation to the length of free tergite II: (0) comparable size (Fig. 7A, B); (1) at least double the size (Figs 3A, B, 5A, B). L = 2; CI = 0.50; RI = 0. State (1) could be considered independent autapomorphies of Magnispina gen. nov. and Pseudopucrolia incerta in DELTRAN, or a synapomorphy of Magnispina gen. nov. + Pseudopucrolia in ACCTRAN. 29. Free tergite II, shape: (0) triangular (Fig. 15A); (1) substraight (Fig. 13A). L = 3; CI = 0.33; RI = 0.50. 30. Free tergite II, corners (lateral margin), armature: (0) absent (Fig. 5A); (1) present (Fig. 3A). L = 4; CI = 0.25; RI = 0.66. 31. Free tergite II, lateral armature, kind of armature: (0) tubercle (Fig. 13A); (1) apophysis (Fig. 3A). L = 2; CI = 0.50; RI = 0.66. 32. Free tergite II, apophysis of lateral armature, aspect: (0) spiniform, powerful, thick, with acuminate apex (Fig. 3A); (1) blunt, thin, not acuminate. L = 1; CI = 1; RI = 1. 33. Free tergite III, aspect of posterior margin: (0) substraight (Fig. 15A); (1) slightly rounded (convex) (Fig. 7A); (2) distinctly convex (Fig. 5A). L = 3; CI = 0.66; RI = 0.50. 34. Free tergite III, longitudinal keel: (0) absent (Fig. 17B); (1) present (Figs 9F, 15E). L = 2; CI = 0.50; RI = 0.75. 35. Median armature of free tergite III: (0) absent (Fig. 9B, F); (1) present (Fig. 3B, E, F). L = 6; CI = 0.16; RI = 0.16. 36. Free tergite III, number of median armature: (0) one (Fig. 3A, B, E, F); (1) two. L = 1. Uninformative. 37. Free tergite III, median apophysis (tubercle), shape: (0) with broad base, blunt (Fig. 3A, B, E, F); (1) with narrow base, spiniform. Uninformative. 38. Free tergite III, corners (lateral margin), armature: (0) absent (Fig. 5F); (1) present (Figs 21A, 23A). L = 3; CI = 0.33; R = 0.33. 39. Free tergite III, aspect of surface: (0) smooth (Fig. 5A, F); (1) with a row of granules (Fig. 3A, B,

REVISION OF HETEROPACHYLINAE 443 F); (2) with an edging row of tubercles. L = 5; CI = 0.40; RI = 0.57. 40. Dorsal anal operculum, armature: (0) absent (Fig. 15D, E); (1) present (Fig. 5E, F). L = 4; CI = 0.25; RI = 0.57. 41. Dorsal anal operculum, shape of median paired armature (when present): (0) spines; (1) tubercles. Uninformative. 42. Dorsal anal operculum, position of median apophysis: (0) coming from the posterior rim of the operculum (Fig. 7E); (1) in a central position (Fig. 23D). L = 1; CI = 1; RI = 1. 43. Ventral anal operculum, aspect of surface: (0) smooth (Fig. 5F); (1) with saliencies that look like the result of the lateral merging of two or more tubercles (Fig. 3F); (2) with lateral large tubercles; (3) row of normal tubercles/granules (Fig. 7F). L = 7; CI = 0.42; RI = 0. 44. Sternite V, size in relation to the remaining sternites in lateral view: (0) similar size; (1) double the size (Fig. 3E, F). L = 1; CI = 1; RI = 1. APPENDAGES 45. Cheliceral bulla, aspect of lateral surface: (0) granulous; (1) smooth. L = 1; CI = 1; RI = 1. 46. Chelicerae, sexual dimorphism (chelicerae of males hypertelic): (0) absent (Fig. 3A, B); (1) present. L = 1; CI = 1; RI = 1. 47. Pedipalpal femur, subapical mesal spine: (0) conspicuous (Fig. 3A, B); (1) very small; (2) absent. L = 5; CI = 0.40; RI = 0.25. 48. Pedipalpal tibia, ectal spines: (0) [Ii] distal (Fig. 19E); (1) i + [Ii] (Figs 3E, 15D); (2) I + [Ii]; (3) Ii[Ii]. L = 5; CI = 0.60; RI = 0.60. 49. Pedipalpal tibia, mesal spines: (0) Ii[Ii]; (1) IiIi (Fig. 13A); (2) I Ii (Fig. 15A). L = 1; CI = 1; RI = 1. 50. Pedipalpal tarsus, mesal spines: (0) IiIi (Fig. 5A); (1) i[ii]; (2) II; (3) IIi distal. L = 5; CI = 0.60; RI = 0.66. 51. Tarsus I, number of tarsomeres: (0) six; (1) five; (2) nine; (3) eight. L = 6; CI = 0.50; RI = 0.25. 52. Femur I, ventral surface aspect: (0) with row of granules; (1) with row of tubercles; (2) smooth. L = 4; CI = 0.50; RI = 0.66. 53. Calcaneus I, length in relation to metatarsus I: (0) about half; (1) smaller than half; (2) much larger than half. L = 4; CI = 0.50; RI = 0.50. 54. Calcaneus I, colour in relation to the rest of the podomere: (0) lighter; (1) same colour or subtle difference; (2) darker. L = 2; CI = 0.50; RI = 0.83. 55. Calcaneus I, thickness: (0) not enlarged; (1) enlarged. L = 2; CI = 0.50; RI = 0.88. 56. Coxa II, dorsal apophysis aspect: (0) uniramous; (1) bifurcated (Fig. 5E). L = 2; CI = 0.50; RI = 0.87. 57. Femur III, ventral surface, row of conspicuous ornamentation: (0) absent; (1) present. L = 1; CI = 1; RI = 1. 58. Femur III, ventral surface, size of granules: (0) equal sized; (1) increasing size progressively towards the apex. L = 4; CI = 0.25; RI = 0. 59. Trochanter III, retrolateral margin, distal (apical) armature: (0) absent; (1) present. L = 1. Uninformative. 60. Trochanter III, retrolateral margin, apical armature shape: (0) granule/small tubercle; (1) spiniform apophysis. L = 1; CI = 1; RI = 1. 61. Calcaneus III, length in relation to the total length of metatarsus III: (0) about one third; (1) smaller than one third; (2) much larger than one third. L = 4; CI = 0.50; RI = 0.33. 62. Trochanter IV, prolateral portion, sub-basal rounded tubercle: (0) absent; (1) present. L = 3; CI = 0.33; RI = 0. 63. Sclerite on the articular membrane between coxa and trochanter IV: (0) square (Fig. 3A); (1) rectangular. L = 1; CI = 1; RI = 1. 64. Coxa IV, prolateral margin, distal armature: (0) absent; (1) present (Fig. 3A). L = 1. Uninformative. 65. Coxa IV, prolateral margin, distal armature, shape: (0) small tubercle; (1) apophysis. L = 3; CI = 0.33; RI = 0. 66. Coxa IV, prolateral margin, apex of distal apophysis: (0) blunt (Fig. 3A); (1) spiniform. L = 3; CI = 0.33; RI = 0.50. 67. Coxa IV, prolateral margin, aspect of distal apophysis: (0) simple; (1) bifid. L = 3; CI = 0.33; RI = 0.50. 68. Coxa IV, prolateral margin, direction of distal apophysis in dorsal view: (0) posteriorly directed; (1) transverse (Fig. 7A). L = 1; CI = 1; RI = 1. 69. Coxa IV, retrolateral margin, apical armature: (0) absent; (1) present. L = 3; CI = 0.33; RI = 0.66. 70. Coxa IV, retrolateral margin, apical armature, shape: (0) spiniform apophysis; (1) small tubercle. L = 1; CI = 1; RI = 1. 71. Trochanter IV, prolateral margin, distal armature: (0) absent; (1) blunt curve apophysis (Figs 3A, 4A); (2) spiniform curve apophysis, dorsally directed. L = 2; CI = 1; RI = 1. 72. Trochanter IV, retrolateral margin, distal armature: (0) absent; (1) present (Figs 3A, 4A). L = 2; CI = 0.50; RI = 0.80. 73. Trochanter IV, ventral surface, retrolateral spiniform apophysis: (0) absent; (1) present. L = 1; CI = 1; RI = 1. 74. Trochanter IV, shape: (0) cylindrical, longer than wide; (1) wider than long (Figs 3A, 4A). L = 1; CI = 1; RI = 1.

444 A. C. MENDES 75. Femur IV, aspect in dorsal view (curvature): (0) straight or slightly curved (Fig. 8A); (1) distinctly curved (Fig. 4A). L = 5; CI = 0.20; 0.42. 76. Femur IV, aspect in lateral view (curvature): (0) straight (Fig. 6C, D); (1) two curvatures ( Sshaped ) (Fig. 4C, D); (2) one curvature. L = 3; CI = 0.66; RI = 0.88. 77. Femur IV, structure of ornamentation: (0) small granules randomly distributed; (1) conspicuous apophyses/tubercles organized into rows (Fig. 4). L = 2; CI = 0.50; RI = 0.80. 78. Femur IV, dorsobasal apophysis: (0) absent (Fig. 6A, C, D); (1) present (Fig. 3A, C, D). L = 3; CI = 0.33; RI = 0.71. 79. Femur IV, dorsobasal apophysis, shape: (0) conical (Fig. 20C, D); (1) rectangular (parallel sides) (Figs 4C, D, 14C, D); (2) anvil-shaped (Figs 22C, D, 24C, D); (3) rounded tubercle. L = 3; CI = 1; RI = 1. 80. Femur IV, main dorsal row of tubercles structure: (0) with rounded tubercles equal-sized (Fig. 24C, D); (1) with larger tubercles immediately after the dorsobasal apophysis (Fig. 18C). L = 1; CI = 1; RI = 1. 81. Femur IV, main dorsal row of tubercles, conspicuous medial apophysis: (0) absent (Fig. 10C, D); (1) present (Fig. 8C, D). L = 2; CI = 0.50; RI = 0. 82. Femur IV, main dorsal row of tubercles, one conspicuous apical tubercle, rounded with acuminate apex: (0) absent (Fig. 13E); (1) present (Fig. 7D). L = 3; CI = 0.33; RI = 0.50. 83. Femur IV, dorsal distal tubercles, distribution: (0) forming a circle; (1) random, not forming a circle. L = 1; CI = 1; RI = 1. 84. Femur IV, rows of tubercles lateral to the main dorsal row: (0) present; (1) absent. L = 2; CI = 0.50; RI = 0. 85. Femur IV, retrolateral margin, structure of ornamentation: (0) rounded tubercles equal-sized (Fig. 6D); (1) two subdistal, large, rounded tubercles (Fig. 24D); (2) three/four last tubercles larger than the remaining (Figs 14A, 20A). L = 4; CI = 0.50; RI = 0.33. 86. Femur IV, prolateral margin, structure of ornamentation: (0) row of rounded tubercles (in distal half) larger in the middle of the row (Fig. 4A, B); (1) row of rounded tubercles increasing progressively towards the apex of the femur (Figs 6C, 12C); (2) rounded tubercles equal-sized. L = 2; CI = 1; RI = 1. 87. Femur IV, retrolateral margin, apical apophysis/ tubercle: (0) absent; (1) present (Fig. 3D). L = 1; CI = 1; RI = 1. 88. Femur IV, retrolateral margin, apical apophysis/ tubercle, shape: (0) apophysis dorsally curved (Fig. 5D); (1) completely dorsally pointed (Fig. 23E); (2) tubercle (much reduced); (3) apophysis ventrally curved. L = 7; CI = 0.42; RI = 0.42. 89. Femur IV, prolateral margin, apical armature: (0) absent; (1) present (Fig. 3D). L = 3; CI = 0.33; RI = 0.50. 90. Femur IV, prolateral margin, apical armature, shape: (0) rounded tubercle with acuminate apex (Fig. 23E); (1) conical apophysis (Fig. 3D). L = 2; CI = 0.50; RI = 0. MALE GENITALIA 91. Glans, dorsal digitiform process: (0) present (Fig. 12); (1) absent (Fig. 25A D, G L). L = 4; CI = 0.25; RI = 0.66. 92. Glans, ventral process: (0) absent (Figs 12, 25); (1) present. L = 2; CI = 0.50; RI = 0.83. 93. Stylus, distal portion, setae: (0) present (Figs 12C H, 25); (1) absent (Fig. 12A, B). L = 4; CI = 0.25; RI = 0.62. 94. Ventral plate, shape: (0) pyriform; (1) subrectangular, slightly narrower at basal portion (Figs 12, 25); (2) subrectangular, slightly narrower at distal portion. L = 3; CI = 0.66; RI = 0.87. 95. Ventral plate, aspect of lateral portions of dorsal surface (dorsolateral surface): (0) not excavated; (1) excavated (Figs 12, 25). L = 1; CI = 1; RI = 1. 96. Ventral plate, ventral surface, blunt rounded small setae: (0) absent; (1) present (Fig. 12C, E). L = 2; CI = 0.50; RI = 0.87. 97. Ventral plate, aspect of distal border: (0) straight; (1) slightly concave (Fig. 12); (2) distinctly concave (Fig. 25A, B, E, F, I L). L = 6; CI = 0.33; RI = 0.50. 98. Ventral plate, lateral lobes: (0) absent (Figs 12, 25); (1) present. L = 2; CI = 0.50; RI = 0.80. 99. Ventral plate, dorsal surface, setae configuration: (0) 4 + 3 (seven pairs of setae); (1) 3 + 1 + 2or 3 + 3 (six pairs of setae); (2) 4 + 1 or 3 + 1 + 1 (five pairs of setae) (Figs 12, 25); (3) 3 + 1 + 3; (4) 3 + 1 + 4. L = 4; CI = 0.75; RI = 0.85. 100. Distal group of dorsal setae, position of the most basal in relation to the three most distal setae: (0) not close (Fig. 12A, B); (1) near (Fig. 12C H). L = 4; CI = 0.25; RI = 0.25. 101. Distal group of dorsal setae, most basal setae (fourth pair of setae from top to base), size in relation to the others: (0) same size (Fig. 12C H); (1) larger (Figs 12A, B, 25E, F). L = 3; CI = 0.33; RI = 0. 102. Ventral plate, distal setae, aspect of apical portion: (0) straight (Figs 12, 25); (1) helicoidal. L = 2; CI = 0.50; RI = 0.

REVISION OF HETEROPACHYLINAE 445 Table 2. Apomorphy list for the preferred hypothesis (Figs 1, 2) of the cladistic analysis of Heteropachylinae (analysis with equal weights, only clades included in the subfamily, outgroup omitted) Node Apomorphies Heteropachylinae 9(1), 10(0), 24(1), 30(1), 50(0), 55(1), 56(1), 66(0), 67(0), 71(1), 89(1), 93(0)A, 94(1)A, 95(1), 96(1), 99(2) Magnispina gen. nov. + Pseudopucrolia 12(2), 15(0), 28(1)A, 40(1), 85(0)A Pseudopucrolia 6(0), 18(0), 34(1), 75(0), 76(0), 86(1), 100(1) Pseudopucrolia discrepans + (Pseudopucrolia 28(0)A, 29(0), 38(1)A, 81(1)A, 82(1) mutica + Pseudopucrolia rugosa) Pseudopucrolia mutica + Pseudopucrolia rugosa 25(1), 88(2) Heteropachylus 2(1), 17(1), 31(0), 53(0), 54(0), 78(1)D, 88(1)A, 91(1), 97(2)A Heteropachylus crassicalcanei + Heteropachylus 15(0), 18(0), 31(1), 33(0), 47(1), 49(2), 61(0), 80(1), 85(0) inexpectabilis Heteropachylus gracilis (Heteropachylus 25(1) peracchii + Heteropachylus spiniger) H. peracchii + H. spiniger 17(0), 35(1), 38(1), 40(1), 42(1)D, 52(1), 53(1), 75(0), 79(2), 83(0), 85(1)A, 100(1) (H. crassicalcanei + H. inexpectabilis) + (H. gracilis 6(0), 39(0), 48(0), 79(0) (H. peracchii + H. spiniger) Nonhomoplastic synapomorphies are in bold type. A, accelerated transformation (ACCTRAN) optimization; D, delayed transformation (DELTRAN) optimization; R, reversion. RESULTS The analysis under equal weights resulted in one most parsimonious tree, with 280 steps, CI = 0.47, RI = 0.66 (270 steps, CI = 0.45, RI = 0.63 removing uninformative characters). The obtained hypothesis (Fig. 1) shows that Heteropachylinae can only be considered monophyletic if Mangaratiba is excluded from the group. A list with the apomorphies and optimizations for this tree is provided in Table 2, and these are also represented in Figure 2. Heteropachylinae after exclusion of Mangaratiba is recovered as monophyletic, highly supported by bootstrap value (96%) and with the highest Bremer support value of the analysis (12). The subfamily was supported by 14 unambiguous synapomorphies, amongst them: the fusion of free tergites I and II to the dorsal scutum in males (character 24, state 1); the presence of lateral armature on the corners of free tergite II (character 30, state 1); bifurcate apophysis on the dorsum of coxa II (character 56, state 1); distal blunt curve apophysis on the prolateral margin of trochanter IV (character 71, state 1); ventral plate of male genitalia subrectangular, slightly narrower at basal portion, excavated on the dorsolateral surface and with blunt, small, rounded setae on the ventral surface (respectively characters 94 96, state 1). This group includes two clades, one composed of Magnispina gen. nov. + Pseudopucrolia, and the other of Heteropachylus. The clade (Magnispina gen. nov. + Pseudopucrolia) is supported by three synapomorphies in both optimizations (ACCTRAN and DELTRAN) as the incomplete groove between mesotergal areas III and IV (character 12, state 2), and it is weakly supported by bootstrap (< 10%) and Bremer supports (1). Pseudopucrolia is supported by seven synapomorphies such as the presence of a longitudinal keel in free tergite III (character 34, state 1); femur IV only slightly curved in dorsal view (character 75, state 0) and straight in lateral view (character 76, state 0) with a prolateral row of tubercles increasing progressively in size towards the apex (character 86, state 1), the last one an exclusive synapomorphy. It obtained a bootstrap support of 58% and a Bremer support of 6. Heteropachylus is supported by six unambiguous synapomorphies such as the unpaired armature of mesotergal area III, which is an exclusive synapomorphy (character 17, state 1); calcaneus I lighter than and with size about half of metatarsus I (respectively character 54, state 0 and character 53, state 0); dorsal digitiform process in glans of male genitalia absent (character 91, state 1). This clade is weakly supported by bootstrap and Bremer supports (35% and 1, respectively). The analysis under implied weights obtained three trees with different topologies. Table 3 lists length, consistency and retention indexes, and fit from the trees obtained. The recovery of the clades obtained with implied weighting using different concavities is

446 A. C. MENDES 1 4 14 15 18 23 35 93 99 1 1 1 1 1 1 1 1 1 13 36 39 40 43 48 51 53 61 64 1 1 0 1 0 3 2,3 2 2 0 Tschaidicancha acanthoma 5 9 12 26 29 93 Borguyia bocaina 1 1 2 1 0 0 6 7 12 15 43 47 50 51 61 62 99102 Pristocnemis pustulatus 3 7 8 10 22 23 45 46 50 63 65 67 69 92 0 0 1 0 0 2 2 3 0 1 0 1 20 48 56 1 1 1 1 1 2 1 0 3 0 1 1 1 1 Discocyrtus infelix 1 89 97 0 3 1 35 66 72 91 97 98 7 9 12 33 39 43 50 5152 58 6162 7188 0 1 0 0 1 1 1 2 1 3 67 72 88 99 0 1 1 0 0 0 1 2 2 0 0 1 2 3 0 0 0 2 3 14 48 52 57 70 77 87 0 1 0 1 0 1 1 5 51 59 62 78 79 82 88102 1 1 0 1 1 3 1 2 1 Pachylus chilensis 5 11 20 43 47 52 69 84101 Goniosoma varium 6 25 30 32 49 51 60 66 73 75 76 77 89 94 2 10 35 37 38 39 40 41 0 1 1 1 0 1 1 0 1 1 2 0 0 0 69 74 75 76 1 1 0 1 1 2 1 1 1 0 1 1 1 43 Mangaratiba acanthoprocta 0 1 1 1 2 1 0 Cobania picea M. monstrosa 26 44 68 85 86 91 92 94 97 98 1 1 1 2 0 0 0 0 1 0 9 10 24 30 50 55 56 66 67 71 89 93 94 95 96 99 1 0 1 1 0 1 1 0 0 1 1 0 1 1 1 2 Heteropachylinae A 12 15 28 40 85 2 0 1 1 0 2 14 35 43 50 58 2 17 31 53 54 78 88 91 97 1 1 0 0 0 1 1 1 1 1 1 1 6 18 34 75 76 86 100 0 0 2 1 0 12 13 16 19 51 1 1 0 0 1 1 6 39 48 79 0 0 78 90 93 101 1 1 1 1 21 30 1 0 1 1 0 1 1 0 28 29 0 0 15 0 Magnispina neptunus 33 39 43 Pseudopucrolia incerta 2 0 0 14 53 65 P. discrepans 38 81 82 1 0 0 14 54 81 1 1 1 25 88 P.mutica 1 0 0 11 15 18 26 58 1 2 P. rugosa 1 1 1 0 0 Heteropachylus ramphonotus 1 4 20 24 27 29 30 34 55 58 75 84 88 90 97100 18 31 33 47 49 61 80 85 H. inexpectabilis 0 0 0 2 1 0 0 1 0 0 0 1 0 1 1 1 12 14 21 35 65 82 91 96101 0 1 0 1 2 0 1 0 H. crassicalcanei 1 1 1 1 0 1 0 0 1 4 88 97 H. gracilis 0 0 1 47 17 35 38 40 42 52 53 75 79 83 85100 H. peracchii 1 14 15 20 48 0 1 1 1 1 1 1 0 2 0 1 1 H. spiniger 25 1 Figure 2. The single most parsimonious tree resulting from analysis under equal weights with apomorphies plotted on branches. Black squares indicate nonhomoplastic apomorphies; empty circles indicate homoplastic apomorphies and reversals. (See Table 2 for details.) A, cladogram including outgroup represented. B, relationships amongst species of Heteropachylinae. 1 0 0 1 B illustrated in Figure 1. The trees obtained with all the concavities recovered the monophyly of Heteropachylinae, Pseudopucrolia, and Heteropachylus sensu this paper. The clade Magnispina gen. nov. + Pseudopucrolia was not recovered with any of the concavities; instead, Magnispina grouped with Heteropachylus. In the analyses with k = 1 6, Pseudopucrolia rugosa is sister of Pseudopucrolia discrepans instead of Pseudopucrolia mutica, obtained with k = 10, 15, 20, and equal weights. The internal relationship within Heteropachylus obtained in the analyses under implied weights differs from that found by the analysis with characters equally weighted in the positioning of H. ramphonotus and Heteropachylus gracilis. SYSTEMATIC ACCOUNTS The former and the proposed classifications of Heteropachylinae based upon the cladistic analysis are given in Table 4.

REVISION OF HETEROPACHYLINAE 447 Table 3. Number of most parsimonious trees (MPTs), length, consistency index (CI), retention index (RI), and fit (fi) from trees found in the analyses of Heteropachylinae Analysis MPTs Length CI/RI Fit Summary of topology EW 1 280 0.47/0.66 Fig. 1 IW k = 1 1 285 0.46/0.62 43.33 Differs from EW: Goniosoma varium is sister of Discocyrtus infelix and do not form a clade with Mangaratiba; Magnispina groups with Heteropachylus; Pseudopucrolia rugosa is sister of Pseudopucrolia discrepans; Heteropachylus ramphonotus and Heteropachylus gracilis IW k = 2 1 282 0.46/0.63 32.81 Differs from IW k = 1: G. varium, D. infelix and Mangaratiba form a clade, as in EW. IW k = 3 1 282 0.46/0.63 26.59 Same tree as IW k = 2. IW k = 4 1 282 0.46/0.63 22.42 Same tree as IW k = 2. IW k = 5 1 282 0.46/0.63 19.41 Same tree as IW k = 2. IW k = 6 1 282 0.46/0.63 17.13 Same tree as IW k = 2. IW k = 10 1 281 0.46/0.63 11.68 Differs from IW k = 2: P. rugosa is sister of P. mutica, asin EW. IW k = 15 1 281 0.46/0.63 8.37 Same tree as IW k = 10. IW k = 20 1 281 0.46/0.63 6.53 Same tree as IW k = 10. A short summary of the main differences between the topology of the trees is also included. EW; equal weighting; IW, implied weighting. Table 4. Former and proposed classification for Heteropachylinae Former classification Proposed classification Heteropachylinae Kury, 1994 Heteropachylinae Kury, 1994 Tribunosoma Roewer, 1930 Magnispina gen. nov. T. discrepans Roewer, 1943 Magnispina neptunus sp. nov. Melloa Roewer, 1930 Pseudopucrolia Roewer, 1912 M. incerta (Mello-Leitão, 1928) Pseudopucrolia discrepans (Roewer, 1943) comb. nov. Pseudopucrolia Roewer, 1930 P. incerta (Mello-Leitão, 1928) comb. nov. P. mutica (Perty, 1833) *P. mutica (Perty, 1833) Thaumatoleptes Roewer, 1930 P. rugosa (Roewer, 1930) comb. nov. T. rugosus Roewer, 1930 Heteropachylus Roewer, 1913a Heteropachylus Roewer, 1913a Heteropachylus ramphonotus sp. nov. H. peracchii Soares & Soares, 1974 H. peracchii Soares & Soares, 1974 *H. spiniger Roewer, 1913a *H. spiniger Roewer, 1913a Aesotrinoma H. Soares, 1977 Heteropachylus gracilis sp. nov. A. crassicalcanei H. Soares, 1977 H. crassicalcanei (H. Soares, 1977) comb. nov. Chavesincola Soares & Soares, 1946 H. inexpectabilis (Soares & Soares, 1946) comb. nov. C. inexpectabilis Soares & Soares, 1946 Pachylinae Sørensen, 1884 Mangaratiba Mello-Leitão, 1940 Mangaratiba Mello-Leitão, 1940 M. acanthoprocta (H. Soares, 1968) M. acanthoprocta (H. Soares, 1968) M. angulispinosa (H. Soares, 1966) M. angulispinosa (H. Soares, 1966) *M. monstrosa Mello-Leitão, 1940 *M. monstrosa Mello-Leitão, 1940 The type species of the nonmonotypic genera are indicated with an asterisk (*). HETEROPACHYLINAE KURY, 1994 Gonyleptinae [part]: Roewer, 1913a: 167, 1923: 463; Mello-Leitão, 1932: 231; Soares & Soares, 1949: 151. Pachylinae [part]: Roewer, 1913a: 10, 1923: 396; Mello-Leitão, 1932: 131; Soares & Soares, 1954: 225. Heteropachylinae Kury, 1994: 350; 2003a: 142; 2007: 168. Kury & Pinto-da-Rocha, 2007: 198; Sigrist & Carvalho, 2008: 35.

448 A. C. MENDES Distribution (Figs 27, 28). BRAZIL. From Ceará to north of Espírito Santo, restricted to tropical moist broadleaf Atlantic Forest. Type genus: Heteropachylus by original designation. Diagnosis: Gonyleptid with scutum length about 5 mm, anterior border of carapace usually with tubercles or granules. Eye mound wider than long, convex, not located directly on anterior margin of carapace, bearing pair of tubercles or granules. Area III with variable armature, mostly with pair of paramedian small tubercles. Males with at least free tergites I and II fused to the dorsal scutum (Fig. 3A), free tergite II bearing median apophysis (reduced or absent only in P. rugosa) and armature on the corners (Fig. 3A). Ozopore with V-shaped cut at the posterior rim of the anterior opening slit (Fig. 3E). Sternite V double the size of the others in lateral view (Fig. 3F). Basichelicerite smooth, with ventral granule, bulla well marked; cheliceral hand smooth, with bristles mainly on distal portion. Calcaneus I slightly thicker than astragalus I. Coxae I II with a pair of dorsal apophyses each, those of coxa II bifurcate (Fig. 5E). Male coxa IV (Fig. 3A, E) hypertelic, bearing apical apophysis on the prolateral margin [small spine in females (Fig. 3B)]. Male trochanter IV wider than long, with apical curve apophysis on the prolateral margin and acuminate on the retrolateral margin (Fig. 3A) [trochanter narrow in females with apical prolateral spine (Fig. 11)]. Femur IV (Fig. 4) short, curved in dorsal view, bearing rows of tubercles and apical apophyses in males. Male genitalia (Figs 12, 25). Ventral plate of male genitalia subrectangular, slightly narrower at basal portion, distal border concave, without lateral lobes, excavated on the dorsolateral surface seeming to be divided into ventral and dorsal layers and ventrally covered with minute setae. Blunt small rounded ventral setae and five pairs of dorsal spiniform setae. Glans without ventral process, in some cases bearing a dorsal digitiform process (Fig. 12D, H). Stylus with subapical small setae. MAGNISPINA GEN. NOV. Distribution: BRAZIL. Espírito Santo and Bahia. WWF Ecoregion NT0103, Bahia coastal forests. Type species: Magnispina neptunus gen. et sp. nov., here designated. Etymology: Genus name comes from the Latin magnus (large) + spina (thorn) in reference to the large median apophysis of free tergite II. Gender feminine. Diagnosis: Three mesotergal areas; mesotergal area III with a pair of tubercles (Fig. 3A). Median apophysis of free tergite II robust and spiniform (Fig. 3A, B, E, F), at least twice the tergite length, tergite bearing a stout spiniform apophysis on each corner; free tergite III without a longitudinal keel, with posterior border convex and unarmed corners (Fig. 3A). Dorsal anal operculum with rounded tubercles (Fig. 3F). Ventral anal operculum with pair of tubercles irregularly shaped (Fig. 3F). Calcaneus I same colour as the astragalus and less than half metatarsus I length. Femur IV distinctly curved in dorsal view and with two curvatures ( S-shaped ) in lateral view, bearing a large dorsobasal apophysis and a row of elongate, rounded tubercles on the prolateral margin, larger in the middle of the row (Fig. 4). Male genitalia. Most basal seta of the distal group of setae larger than the others; dorsal digitiform process of glans present (Fig. 12A, B). MAGNISPINA NEPTUNUS SP. NOV. (FIGS 3, 4, 12A, B, 27). Distribution (Fig. 27): BRAZIL, Espírito Santo (São Matheus) and Bahia (Ilhéus, Jussari and Una). WWF Ecoregion NT0103, Bahia coastal forests. Type material: Male (m) holotype (IBSP1052), BRAZIL, Bahia, Una, Reserva Biológica de Una, 13 16.iv.1998, Brescovit, Bertani & Pinto-da-Rocha. Paratypes. m (IBSP1052); m (IBSP1051); 1m and 4 females (f) (IBSP1068), same data as the holotype; m (IBSP1089) from Ilhéus, CEPLAC, 10 12.iv.1998, Brescovit A, Bertani R & Pinto-da-Rocha R; m (IBSP342); 2m and 2f (IBSP378); 3m (IBSP 383); m and f (IBSP 439); 5m and 2f (IBSP 446); 2m (IBSP 447); 2m and 3f (IBSP 464) from Espírito Santo, São Matheus, Reserva Florestal Vale do Rio Doce, i.1998, A.D. Brescovit et al. Etymology: Species name comes from the Latin Neptunus, the Roman god of the sea and of other waters, who possessed an enchanted trident. It is in reference to the aspect of free tergite II, which bears two lateral and a median large hooked apophyses. The name is used as a noun in apposition. Diagnosis: Same as the genus. Measurements: Male holotype: CL: 2.6; MCW: 3.3; ASL: 4.9; MASW: 7.1. FL: 3.8. Males (N = 4): CL: 2.3 (2.1 2.6); MCW: 3.1 (2.9 3.3); ASL: 4.3 (3.6 4.9); MASW: 6.5 (5.5 7.1). FL: 4.1 (3.8 4.4). Females