A data matrix with 46 taxa and 71 characters (Table 1) was analyzed with two programs using equal (Nona 2.0 for Windows (Goloboff 1998)) and implied character weighting (Piwe 3.0 for Windows (Goloboff 1997)). For Piwe, concavities from 1 to 6 were used. All characters were treated as unordered. Commands used to search for trees were h10000, h/1000, amb-, mult*500. Absolute and relative Bremer supports (Bremer 1994) were estimated in Piwe using the following command: h1000, find*, bsupport50, bsupport*.

In the cladistic analyses, representative species were included of seven of the eight theraphosid subfamilies recognized by Raven (1985), as well as species of all aviculariine genera and other available arboreal taxa with uncertain position, e.g., Poecilotheria Simon, 1885, Heteroscodra Pocock, 1899, and Stromatopelma Karsch, 1881, in order to test Aviculariinae monophyly. Melloina santuario Bertani, 2013 (Paratropididae) was used to root the cladogram. The choice of this genus was done based on two considerations: paratropidids were at the beginning of this study considered as the sister-group of Theraphosidae in morphological analysis (Raven 1985), and Melloina is the paratropidid genus which retains more plesiomorphic characters shared with the Theraphosidae (Raven 1985). However, recent molecular analyses indicate Barychelidae as the sister-group of Theraphosidae (Bond et al. 2012, Hamilton et al. 2016, Garrison et al. 2016, Wheeler et al. 2016), so future studies should aim to include more taxa based on molecular analyses. The ingroup specimens used in cladistics analysis are those used in the descriptions. Specimens used as outgroups are: Paratropididae, Glabropelmatinae: Melloina santuario Bertani, 2013, holotype male (MIZA 520) and paratype female (MNRJ 12965), both from Venezuela, Lara, Cueva El Santuario [9°49'N, 70°03'W], O. Villarreal col., 19 April 2000. Theraphosidae, Aviculariinae: Ephebopus murinus (Walckenaer, 1837), male (IBSP 9650) and female (IBSP 9658), Brazil, state of Pará, Tucuruí, U.H.E. Tucuruí, [4°20'S, 49°31'W], Equipe de Resgate de Fauna col., 1984; Ephebopus uatuman Lucas, Silva & Bertani, 1992, male holotype (IBSP 4939) and paratype female (IBSP 4940), Brazil, state of Amazonas, Presidente Figueiredo [2°02'S, 60°01'W], Uatuman River, Balbina Hydroelectric Power Station, M. Costa col., 19 February 1988; Heteroscodra maculata Pocock, 1899, male, Africa, pet trade (IBSP 9642); female, Guinea-Bissau, pet trade (IBSP 9644); Iridopelma hirsutum Pocock, 1901, male, Brazil, state of Paraíba, João Pessoa [7°07'S, 34°52'W], P. F. L. Duarte col., 30 November 1979 (IBSP 8078); female, Brazil, state of Alagoas, Murici, E. E. Murici, UFAL (09°14'1.73"S, 35°50'1.61"W), R. Bertani, D. R. M. Ortega and R. H. Nagahama col., 16 August 2006 (MNRJ 06252); Iridopelma katiae Bertani, 2012, holotype female, Brazil, state of Bahia, Mucugê, Parque Nacional da Chapada Diamantina (12°45'7.00"S, 41°30'2.18"W), 1399 m a.s.l., R. Bertani, C. S. Fukushima and R. H. Nagahama col., 18 February 2008 (MZUSP 36887) and paratype male, same data, (12°45'3.77"S, 41°30'4.01"W), 1377 m a.s.l., R. Bertani, C. S. Fukushima and R. H. Nagahama col., 17 February 2008 (MZUSP 36888); Iridopelma marcoi Bertani, 2012, holotype female, Brazil, state of Bahia, São Desidério (12°28'52"S, 45°09'10"W), 724 m a.s.l, M. A. Freitas col., October 2009, under tree bark, 1 m above the ground, area of carrasco vegetation (MZUSP 36891); Iridopelma oliveirai Bertani, 2012, holotype male, Brazil, state of Bahia, Central, Toca dos Pilões [11°08'S, 42°06'W], A. D. Brescovit col., July 2000 (IBSP 10100) and paratype female, at night in “Macambira” bromeliad (Bromelia laciniosa), same data (IBSP 8714); Iridopelma vanini Bertani, 2012, holotype female, Brazil, state of Piauí, Parnaíba [2°53'S, 41°41'W], 5 m a.s.l., R. Bertani col., November 1994 (IBSP Ref. 74.595) and paratype male, Brazil, state of Maranhão, Barreirinhas, Parque Nacional dos Lençóis Maranhenses [2°41'S, 42°55'W), 32 m a.s.l., Equipe Biota col., 12–18 October 2001 (IBSP 11328); Iridopelma zorodes (Mello-Leitão, 1926), male, Brazil, state of Bahia, Mata de São João, RPPN Camurujipe, in a leaf at night, R. Bertani, C. S. Fukushima and R. H. Nagahama col., 04 October 2007 (MNRJ 6254); female, Brazil, state of Sergipe, Areia Branca, Mosqueiro, I. Matos col., 01 November 1992, ref. 79901 (IBSP 11760); Pachistopelma bromelicola Bertani, 2012, holotype male, Brazil, state of Bahia, Elísio Medrado, RPPN Jequitibá (12°52'3.20"S, 39°28'9.09"W), R. Bertani, C. S. Fukushima and R. H. Nagahama col., 07 October 2007, collected at night, found immature inside bromeliads, matured in captivity in May 2010 (MNRJ 06241) and paratype female, same data (MNRJ 06242); Pachistopelma rufonigrum Pocock, 1901, male and female, Brazil, state of Alagoas, Murici, Estação Ecológica de Murici (9°14'9.52"S, 35°48'0.25"W), 245 m a.s.l., R. Bertani, R. H. Nagahama and D. R. M. Ortega col., 11 August 2006 (MNRJ 6246 AL1100); Psalmopoeus cambridgei Pocock, 1895, male, Trinidad-Tobago, pet trade (IBSP 9653); Psalmopoeus sp., female, Venezuela, pet trade (IBSP 9655); Stromatopelma sp., male, Sierra Leone, pet trade (IBSP 9665); female, Africa, pet trade (IBSP 11136); Tapinauchenius violaceus (Mello-Leitão, 1930), holotype female of T. purpureus Schmidt, 1995 (SMF 38042) and paratype male of T. purpureus Schmidt, 1995, French Guiana (SMF 38046); Typhochlaena amma Bertani, 2012, holotype female, Brazil, state of Espírito Santo, Santa Teresa, Estação Ecológica de Santa Lúcia [19°58'S, 40°32'W], 672 m a.s.l, A. P. L. Giupponi col., February 2008 (MNRJ 06239) and paratype male, same locality, no further data (MNRJ 12926); Typhochlaena costae Bertani, 2012, holotype female, Brazil, state of Tocantins, Palmas, U.H.E. Luís Eduardo Magalhães [10°12'S, 48°24'W], 211 m a.s.l, during faunal rescuing, M. Costa and D. Cândido col., 12 January 2002 (IBSP unnumbered) and paratype male, Brazil, Tocantins, Lajeado (9°46'4.85"S, 48°21'6.69"W), 226 m a.s.l, G. Puorto, R. Martins and I. Knysak col. (pitfall trap), April 2002 (IBSP unnumbered); Typhochlaena curumim Bertani, 2012, holotype female, Brazil, state of Paraíba, Areia, Reserva Ecológica Estadual Mata do Pau-Ferro [6°58'S, 35°42'W], 500 m a.s.l., under tree bark, A. D. Brescovit, R. Bertani, A. B. Bonaldo and S. C. Dias col., September 1999 (IBSP 8701) and paratype female (IBSP 8354), same data; male, Brazil, state of Rio Grande do Norte, Baía Formosa, lying over a leaf in a bush at night, S. N. Migliore col., 19 April 2014, S90 (MNRJ 06915); Typhochlaena paschoali Bertani, 2012, holotype female and immature paratype, Brazil, state of Bahia, Camacan [15°24'S, 39°30'W], no further data (MNRJ 13723); paratypes 1 female, 1 subadult male, 9 immatures, Brazil, state of Bahia, Jussari [15°10'S, 39°29'W], no further data (MNRJ 12928 – R2981); Typhochlaena seladonia (C. L. Koch, 1841), 1 male, 1 female exuvium, Brazil, state of Bahia, Camaçari, Jacuipe [12°42'S, 38°07'W], T. Brazil ded., August 1980 (IBSP 4551); 1 female, Brazil, state of Bahia, Salvador, Alphaville [12°56'S, 38°21'W], G. G. Montingelli col., 11 December 2001 (IBSP 109718); Eumenophorinae: Pelinobius muticus Karsch, 1885, male (IBSP 8530), female (IBSP 9643), both from Kenya, born in captivity; Harpactirinae: Pterinochilus sp., male, Angola, Biula-Dala [11°10'S, 20°12'E] (IBSP 9647); female, Africa, pet trade (IBSP 8765); Ischnocolinae: Holothele rondoni (Lucas & Bücherl, 1972), holotype male and paratype female, Brazil, state of Amazonas, Iauaretê [0°36'N, 69°11'W] (IBSP 4090); Ornithoctoninae: Haplopelma longipes von Wirth & Striffler, 2005, male, Cambodia, Skuon [1°33'N, 104°55'E], A. Anderson col., 10 July 2003 (MZUSP 28761); Haplopelma minax (Thorell, 1897), female, Thailand, 1 mi east of Bangkok [13°43'N, 100°31'E] (IBSP 9645); Selenocosmiinae: Phlogiellus sp, male, Vietnam, R. Blauman col., April 1989 (MZUSP 28762); female (AMNH), pet trade; Poecilotheria sp., male, India, pet trade (IBSP 9660); Poecilotheria ornata Pocock, 1899, female, Sri Lanka, pet trade (IBSP 8767); Theraphosinae: Phrixotrichus vulpinus (Karsch, 1880), 3 males, Chile (IBSP 3817–A); 1 female, Chile, Osorno [40°34'S, 73°09'W] (IBSP 3817–B); Lasiodora sp., male, Brazil, state of Paraíba, João Pessoa [7°07'S, 34°52'W] (IBSP 11143); female, Brazil, state of Pernambuco, Jaboatão dos Guararapes [8°06'S, 35°00'W], Conjunto Murebeca (IBSP 10293); Subfamily Incertae sedis: Encyocratella olivacea Strand, 1907, holotype female, Tanzania, Amani (05°09'S, 38°36'E), Vosseler leg., November 1903 (ZMB 10484); male, Tanzania, mountains near Arusha (03°23'S, 36°40'E) wild caught, reared to maturity, 2004 (BMNH); female holotype and 2 females paratypes of Xenodendrophila gabrieli Gallon, 2003, northern Tanzania, mountains near Arusha (03°23'S, 36°40'E), Joe Beraducci leg., circa April 1999 (BMNH 2005.123).

In the cladistic analyses we used the characters proposed by West et al. (2008), Bertani (2012), and 15 new characters. Forty characters (0–3, 5–18, 20–22, 31–33, 41, 43, 47–48, 52–57, 60–61, 64–66, and 69–70) are the same used by West et al. (2008) or were slightly modified here. Ten characters (4, 19, 23, 26, 35, 37–38, 59, 67 and 68) are strictly the same used by Bertani (2012). Characters 24, 44–46 and 50 were adapted from that work. The newly introduced characters are 25, 27–30, 34, 36, 39–40, 42, 49, 51, 58, 62–63. In characters marked with an asterisk (*, viz 7–8, 14, 20, 31, 60 and 65) fit and steps are not given as the character is not significant to the analyses but were kept to indicate autapomorphies for terminal taxa.

Character coding is always challenging, especially regarding problems concerning “part” coding (e.g., presence of tibial apophysis) versus “character-variable” coding (e.g., development of tibial apophysis). We decided to code “part” characters separately from its “character-variable” characters to preserve transformational independence between them, assuring that primary homology statements of both characters were included in the analyses (Lee and Bryant 1999). Fusing coding results in a loss of phylogenetic information and also in the determination of taxa synapomorphies by others characters (Lee and Bryant 1999). The characters that were splited are: 8/9, 11/12, 24/25, 31/32, 44/45, 56/57/58, 61/62, and 61/63.

Theraphosids are notorious for their morphological homogeneity (Bertani 2001). Due to this, it is extremely difficult to find characters suitable for taxonomy and cladistics. In order to find new characters, we have tried many approaches. We took measurements of leg and palp articles such as tibia, metatarsi and femora, and established ratios between them and ratios between one of these articles and the carapace. We also calculated the ratio between the length of tibial apophysis and tibia. Unfortunately, all results were not significant enough to be used in the analyses. Measurements of male palpal bulb, such as the ratio between the length of embolus and carapace, showed the same problems. Besides this, the morphology of the Aviculariinae male palpal bulb is a problem per se. It is very difficult to position the specimen in order to make reliable and replicable embolus measurements.

Another feature we have tried to use is the length of type II urticating setae. Some species, such as A. hirschii, A. minatrix, A. taunayi, Y. sooretama comb. n. and Y. gamba comb. n., apparently have unusual urticating setae length. However, we could not confirm this suspicion since we have just one or two specimens of those species. Thus, it was impossible to determine the intraspecific length variation and establish a reliable length range for each species. Beside setae length, another urticating setae feature that could be used is the distribution of barbs along setae in females (Figs 16–17). Usually, females present barbs only in a short area near hair stalk (Stradling 1978). In contrast, morphotype 5 of A. avicularia presents barbs along 1/3 of the seta length (Fig. 17). However, since we have only three immature females, it is not possible to state if that is a typical feature of the morphotype population.

Thus, after trying different attempts and approaches, we consider the following characters suitable to be used:

0. Anterior row of eyes: (0) procurved (Fig. 13), (1) straight (Fig. 12).

1. Clypeus: (0) absent or narrow; (1) present, wide.

West et al. (2008) and Bertani (2012) used three states for this character: “absent”, “narrow” and “wide”. Herein, we consider that the states “absent” and “narrow” should be merged due to difficulty in establishing limits between them. The state “wide”, on the other hand, is easily distinguishable.

2. Fovea, closure: (0) slit like; (1) pit like.

3. Labial cuspules, number: (0) 30–300; (1) 0–20; (2) 350–450.

4. Sternum, shape: (0) longer than wide, not truncated behind (Fig. 9); (1) as long as wide, truncated behind (Fig. 7).

5. Sigilla, posterior pair, position: (0) marginal, less than 1.5 diameters from margin; (1) close to the center, more than 2 diameters from margin.

6. Setae on metatarsi and tibia I–IV, length, males: (0) same length as other articles (Fig. 299); (1) longer setae laterally projected, forming a brush (Fig. 301).

7. Scopulae on metatarsi IV, division: (0) divided by setae or spiniform setae; (1) not divided.*

8. Tarsal scopulae, occurrence: (0) no true scopula; (1) true scopula.*

9. Tarsal scopulae, setae development: (0) scopula composed of sparse setae; (1) dense scopula that does not extend much laterally; (2) scopula very extensive laterally, giving the tarsi and metatarsi I and II a spatulate appearance (Figs 299–301).

10. Tarsi IV, division, males: (0) cracked; (1) integral.

11. Leg spines, occurrence: (0) present; (1) absent.

12. Leg spines, distribution: (0) in whole tibiae and/or metatarsi; (1) only in ventral apical tibiae and/or metatarsi.

13. Palpal femora, scopula on retrolateral face, occurrence: (0) absent, (1) present.

14. Femora IV, scopulae on retrolateral face, occurrence: (0) absent; (1) present.*

15. Chelicerae, scopulae on retrolateral face, occurrence: (0) absent, (1) present.

16. Maxillae, spiniform setae on lower prolateral face: (0) absent, (1) present.

17. Stridulatory bristles form maxillae lyra: (0) absent, (1) present (Fig. 11).

18. Stridulatory bristles on coxae I, occurrence: (0) absent, (1) present.

19. Posterior lateral spinnerets, distal article, shape: (0) digitiform (Fig. 10); (1) domed (Fig. 8).

20. Patellae and tibia, stripes, color: (0) same color of the rest of the segment, (1) white.*

21. Leg rings on distal femora, tibiae and metatarsi, coloration: (0) same color of the rest of the segment, (1) white (Fig. 299), (2), yellow or orange (Fig. 300).

Avicularia avicularia and A. variegata stat. n. have morphotypes with different colors of leg rings. Due to this intraespecific variability, these two taxa were coded as “?” in the matrix.

22. Tibiae, metatarsi and tarsi, dorsal coloration: (0) homogeneous color, (1) with black marking (Fig. 5).

23. Color pattern, ontogenetic change: (0) pattern remains practically the same during ontogeny; (1) pattern presents drastic changes during ontogeny.

Most outgroup species have only slight variations in the general color and abdominal pattern during their ontogeny. We consider a drastic change when legs, carapace, and abdomen change their color during ontogeny. Most aviculariines have abdomen dorsum pattern with stripes in their early life stages, losing it when reaching maturity, which is, herein, considered as state 1. Other species, such as Ephebopus murinus, despite not having heterogeneous abdominal pattern during all ontogeny, have some conspicuous changes: immatures have light brown legs with black tarsi and black carapace, while adults have black legs with very conspicuous white stripes and light brown carapace. This condition is also considered as a drastic change and codified as “1”.

24. Dorsal abdominal pattern in immatures (see figures in Bertani 2012): (0) homogeneous; (1) herringbone; (2) with a zigzag central longitudinal dark stripe over a clear spot, which is marginated in black and connects to five narrow transversal black stripes; (3) two median dorso-lateral spots; (4) central longitudinal black stripe with 5–6 lateral stripes, connecting or not with the central stripe (Fig. 37); (5) leaf pattern: black with a large central clear area having a longitudinal dark stripe, area closer to longitudinal dark stripe usually reddish; (6) central longitudinal reddish stripe inside a dark area with zigzag borders connected to transverse dark stripes (Fig. 295); (7) longitudinal central stripe of a different color of remaining abdomen (Fig. 184).

Character modified from Bertani (2012). A new state was added (state 7), which was applied here only to A. hirschii. State 4 is present in most Avicularia species. This abdominal stripped pattern is also spread over the subfamily. It can also be found in Caribena gen. n. species, in both Pachistopelma species, in T. curumim, T. costae and T. seladonia as well as Tapinauchenius sp. and Psalmopoeus sp.

25. Abdominal pattern, immatures, central longitudinal stripe, connection with lateral stripes: (0) connected with all lateral stripes (Fig. 167); (1) connected only with first and second lateral pair of stripes (Fig. 101); (2) disconnected from lateral stripes (Fig. 44).

In species with abdominal pattern corresponding to state 4 of character 24, the central longitudinal black stripe can connect or not with the lateral stripes. State 0 is found in many Aviculariinae species such as A. purpurea and A. minatrix as well as in C. versicolor comb. n., T. costae, T. seladonia, P. bromelicola, Tapinauchenius sp. and Psalmopoeus sp. State 1 is found only in two Avicularia species from the early stages: A. merianae sp. n. and A. taunayi. Central stripe disconnected from lateral stripes (state 2) can be found in A. rufa, A. juruensis, A. variegata stat. n. and A. avicularia, as well as in C. laeta comb. n., T. curumim and in P. rufonigrum.

26. Body coloration, juveniles: (0) brownish or grayish (Fig. 44); (1) metallic green or blue (Fig. 239).

27. Tarsi, coloration, juveniles: (0) same color of other articles (Fig. 167); (1) black (Fig. 44).

Black tarsi contrasting with other lighter articles in legs and palpi are present in juveniles of some Tapinauchenius and Ephebopus species. It also occurs in all Avicularia species with known immature stages (except A. purpurea) as well as in I. katiae and T. paschoali. Since Psalmopoeus sp. has black tarsi contrasting with clear metatarsi, it was codified as state 1 despite most articles having a dark color. Typhochlaena curumim has black tarsi and metatarsi; but as the other articles are lighter and there is a strong contrast between them, it is considered as having state 1.

28. Dorsal abdominal pattern, single dorsal stripe, male, occurrence: (0) absent (Fig. 43); (1) present (Fig. 297).

Males of Y. sooretama comb. n., Y. gamba comb. n., A. lynnae sp. n., and I. hirsutum have a single dorsal central stripe on abdomen dorsum. This pattern is very distinct from those found in females of these same species and also very distinct from patterns commonly found in other aviculariine males. Patterns other than dorsal central stripe were all codified as state 0. Avicularia hirschii is coded as “?” since the holotype, the single known male, has the abdomen in poor condition. The description also does not allow us to recognize its abdominal pattern.

29. Abdominal setae, distribution, females: (0) homogeneous (Fig. 304); (1) heterogeneous, with long guard-setae grouped on lateral and dorso anterior areas (Figs 302–303).

Presence of long guard-setae grouped on lateral and dorsal anterior areas of the abdomen is a characteristic found only in females of A. avicularia, A. rufa, A. juruensis and A. variegata stat. n. These longer guard-setae have different coloration from the shorter body setae found on remaining abdominal areas. In males, the distribution pattern is homogeneous, with long guard-setae spread all over the dorsal abdomen.

This character is coded “?” in A. juruensis since each morphotype has a different type of abdominal setae pattern.

30. Legs and palp, long guard-setae, coloration: (0) setae with homogeneous coloration or gradually lightening along its length (Fig. 299); (1) setae with darker base and contrasting whitish apex (grizzled setae) (Fig. 300).

Theraphosid spiders have two main covering setae types in legs and pedipalps; the short setae that densely covers most of the article area, providing a velvety aspect, is known as short body setae (Foelix 2010). The other type is the very long tactile setae, which presents insertion into a socket, allowing them to move, and is more spaced out among each other; this is known as guard-setae (Foelix 2010). These long setae are normally homogeneously dark or present a gradual lightening toward apex. But, in some species, these setae are dark and present a contrasting whitish apex.

All examined outgroup taxa have setae with homogeneous coloration along its length. In the ingroup, some species such as A. avicularia, A. rufa, A. juruensis and A. variegata stat. n. have the long setae with a contrasting whitish apex, giving a grizzled body aspect (Fig. 300). However, this character is coded “?” in A. juruensis and A. avicularia since some morphotypes (morphotypes 1 and 3 for A. juruensis and morphotypes 1, 2, 4, and 5 for A. avicularia) have homogeneous coloration along most of the guard-setae.

31. Spermathecae, occurrence: (0) present; (1) absent.*

32. Spermathecae, number: (0) two, completely separated (Fig. 14), (1), two, fused at base; (2) one, totally fused.

33. Spermathecae, walls, shape: (0) without projections or lobes (Fig. 21); (1) with projections or lobes (Fig. 91).

Of the several specimens of A. rufa, A. juruensis, A. avicularia, and A. variegata stat. n. examined, a few specimens have one or more discrete lobes in the spermathecae. We consider them an extreme morphological variation, as the typical spermathecae of those species lack any type of lobes.

Caribena laeta comb. n. spermathecae have a slight intumescence in their apex (Figs 227–229). It is not a well-defined lobe as in other taxa, such as Y. sooretama comb. n. (Fig. 257) or in A. taunayi (Figs 91–92). Due to this, we considered C. laeta comb. n. as having state 0. State 1 is widespread among theraphosid genera.

34. Spermathecae, midwidth: (0) as wide as its base, or midwidth more slender than the base width but wider than apex width (Fig. 21); (1) midwidth expanded, about 1.5 times its basal and apical portion widths, or wider (Fig. 54).

We consider the spermathecae with midwidth expanded when this region is 1.5 times or more wider than both apex and base widths. Spermathecae of A. variegata stat. n. (Fig. 54) and A. juruensis (Fig. 106) clearly have state 1. In other aviculariine species, it is also possible to recognize this pattern despite their unusual spermathecae morphology, as in spermathecae with lobes in A. taunyai (Fig. 91). Spermathecae midwidth of Y. sooretama comb. n., Y. gamba comb. n. and A. merianae sp. n. are slightly wider than their bases’ width, but not enough to consider them as state 1.

In other taxa such as Haplopelma sp., Poecilotheria sp., Lasiodora sp., P. vulpinus and A. rickwesti comb. n., spermathecae (Fig. 284) are very modified or completely fused. In these cases we consider the character inapplicable.

35. Spermathecae, shape: (0) non-spiraled or not-twisted (Fig. 21); (1) twisted or spiraled (Fig. 183).

36. Spermathecae, weakly-sclerotized area, size: (0) weakly-developed or shorter than half the length of the well-sclerotized area (Fig. 21); (1) twice the length or the same length of the well-sclerotized area (Fig. 154); (2) spermatheca virtually non-sclerotized (Figs 257–259).

In some species, the spermathecae have a weakly-sclerotized basal portion, lighter and softer than the more sclerotized distal area, and without large visible pores. As this weakly-sclerotized area is barely present in most outgroup species, we consider them as having state 0. State 1 is found only in A. purpurea (Fig. 154) and A. merianae sp. n. (Fig. 196). All three species of Ybyrapora gen. n. have spermathecae with a very small sclerotization degree, even in adult females (Figs 257–259), and they were coded as state 2. Melloina santuario and T. seladonia have spermathecae virtually non-sclerotized, very similar to Ybyrapora gen. n. spermathecae. However, we know a single female of each species, both very small in size, in which the non-sclerotized spermathecae condition may be due to a not fully mature life stage. Thus, we decided to code both species in matrix as “?”.

37. Spermathecae, curvature: (0) straight or almost so, curved inward (Fig. 14); (1) with an accentuated outwards curvature medially (Fig. 21).

38. Cymbium, process in retrolateral lobe, occurrence: (0) absent or very weakly-developed (Figs 162, 305); (1) present, well-developed (Figs 306–307).

The presence of a process on the cymbium was considered by Raven (1985) a synapomorphy of Aviculariinae. However, more recent analyses showed that the structure is lacking in many Aviculariinae genera (West et al. 2008, Bertani 2012). Despite this, the character is informative and was studied in detail, herein. The process is a projection on the apical area of cymbium retrolateral lobes (Figs 306–307) and shows a considerable level of interspecific variation in size, shape, and associated setae. If on one hand this variation provides more information for cladistic analysis, on the other hand sometimes is difficult to establish limits for the states. We consider the process well-developed only when there is a clear and well-defined projection in the cymbium lobes (Figs 306–307). If there is a small bump without clear delimitation on cymbium retrolateral area, it was considered a very weakly-developed process, or state 0 (Figs 162, 305). We consider that Pachistopelma spp., Iridopelma spp., Caribena gen. n. spp., Y. diversipes comb. n., A. rickwesti comb. n., A. avicularia, A. rufa, A. juruensis, A. variegata stat. n. and A. taunayi have well-developed process (state 1).

39. Cymbium, setae covering the retrolateral process, thickness: (0) thin (Fig. 306); (1) thick (Fig. 307).

Thin setae covering the retrolateral process are found both in some Avicularia species as A. hirschii, A. lynnae sp. n. and A. caei sp. n., as well as in species of Caribena gen. n. and in Pachistopelma species.

40. Cymbium, process in retrolateral lobe, shape: (0) rounded (Fig. 307); (1) sharp (Fig. 306).

The apex of the process on retrolateral lobe is usually rounded, except in both species of Caribena gen. n., in which the process apex is sharp.

41. Cymbium, prolateral lobe, shape: (0) rounded; (1) subtriangular.

42. Tegulum, format: (0) globous (Fig. 29); (1) piriform (Fig. 320).

We detected two distinct tegula shapes. The globous tegulum is short, narrowing abruptly and giving origin to a slender embolus since its beginning. The pyriform tegulum is longer and tapers to form the embolus.

All Aviculariinae possess globous tegulum, except T. costae.

43. Subtegulum, length: (0) small; (1) large, extending down the bulb for half of the tegulum.

44. Palpal bulb, prominence on tegulum (frontal view), occurrence: (0) absent (Fig. 312); (1) present (Figs 313–315).

The tegula of some species have a groove in its prolateral side, which forms a prominence near it (Bertani 2012). This prominence can be better seen in frontal view (Figs 313–315). Presence of a prominence is shared by all Avicularia species except A. minatrix and also in species of Holothele, Pelinobius, Phlogiellus, Ephebopus, and Psalmopoeus. In other aviculariine species this groove is lacking and the tegulum prolateral area is rounded. This state (0) is found in species of Iridopelma, Pachistopelma, Typhochlaena (except T. seladonia), Ybyrapora gen. n. (except in Y. sooretama comb. n.), C. laeta comb. n. and A. rickwesti comb. n. Encyocratela olivacea is coded as “?” since this character was included after the specimen examination and we could not re-examine it in the proper position.

45. Palpal bulb, prominence on tegulum, development: (0) weakly-developed (Fig. 313); (1) developed (Fig. 314); (2) well-developed (Fig. 315).

We found that prominence on tegulum can have different development degrees. There are species in which prominence is well-developed (state 2), with a deep groove, as seen in A. variegata stat. n. and in A. juruensis (Fig. 315). When there is a distinct groove and it is clearly possible to determine a tegulum prominence, though it is not well-developed, it was coded as state 1. This state is found in most species analyzed (Fig. 314). However, if the groove is discrete, hard to see, and consequently the prominence is weakly-developed (Fig. 313), it was coded as state 0, as in H. rondoni, P. muticus, T. seladonia and A. caei sp. n.

46. Embolus, length, retrolateral view: (0) 1.5 to 2.5 times the tegulum’s length; (1) smaller than tegulum’s length (Fig. 320); (2) 3.0 to 3.5 times the tegulum’s length (Fig. 323); (3) more than 4 times the tegulum’s length (Fig. 277).

Most Aviculariinae have emboli from 3 to 3.5 times the tegulum’s length (state 2), including all Avicularia species, except A. lynnae sp. n. and A. caei sp. n., which have emboli described as very long (state 3).

47. Embolus, distal width: (0) thin, shorter than 1/5 tegulum’s length (Fig. 323); (1) thick, more than 1/3 tegulum’s length (Fig. 320).

48 Embolus, shape: (0) not flattened (Fig. 323); (1) slightly flattened (Fig. 320); (2) very flattened (Fig. 286).

49. Embolus, tip: (0) tapers (Figs 318–319); (1) narrows abruptly (Figs 316–317).

In some Aviculariinae taxa such as A. minatrix, A. taunayi, A. rufa, A. juruensis, A. avicularia, A. variegata stat. n. and Y. sooretama comb. n., the embolus tip narrows abruptly near 1/5 distal portion (better visualized in dorsal view). The remaining taxa show a gradual embolus tapering.

50. Embolus, proximal part in frontal view, shape: (0) straight; (1) slightly curved (Fig. 233); (2) very curved (Fig. 30).

Most outgroups (except Haplopelma sp. and P. muticus) and some Aviculariinae, such as Psalmopoeus sp., Tapinauchenius sp., Ephebopus spp., Stromatopelma sp. and Heteroscodra sp., have straight proximal portion of embolus in frontal view. All the other Aviculariinae taxa have some curvature degree in proximal portion of embolus. Caribena laeta comb. n. (Fig. 233), Y. gamba comb. n., T. costae, T. curumim and A. rickwesti comb. n. have slightly curved proximal embolus. The other Aviculariinae taxa have very curved proximal embolus (state 2) (Fig. 30).

51. Embolus, curved, retrolateral view, angle between tegulum’s margin and embolus medial portion: (0) very acute angle (Fig. 321) (1) acute angle (Fig. 323) (2) right or obtuse angle (Fig. 322).

We found that the angle between tegulum’s margin and embolus medial portion in retrolateral view can be informative for cladistic analysis. In Haplopelma sp., Pachistopelma spp. and Y. diversipes comb. n., the angle formed is very acute (state 0) (Fig. 321). In other taxa, such as T. costae, T. curumim, Iridopelma spp., Ybyrapora gen. n. (except Y. diversipes comb. n.), A. rickwesti comb. n., C. laeta comb. n. and P. muticus sp., an obtuse angle is formed between embolus medial portion and tegulum’s margin (state 2) (Fig. 322). In the remaining aviculariine species, including all Avicularia, an acute angle is formed between these two structures (state 1) (Fig. 323).

52. Bulb, prolateral inferior keel on embolus: (0) absent; (1) present.

53. Bulb, prolateral superior keel on embolus: (0) absent;(1) present.

54. Bulb, apical keel on embolus: (0) absent;(1) present.

55. Bulb, retrolateral keel on embolus: (0) absent; (1) present.

56. Tibial apophysis on leg I, occurrence: (0) present (Figs 310–311); (1) absent or weakly-developed (Figs 308–309).

Specimens of Y. sooretama comb. n., Y. gamba comb. n., Typhochlaena spp., P. muticus, Phlogiellus sp., Poecilotheria sp., E. olivacea, Heteroscodra sp., and Stromatopelma sp. do not present any type of apophysis or tibial modification. Some aviculariine species have a discrete elevation on prolateral tibia that can be covered or not by a cluster of setae but clearly does not form a branch. This discrete elevation on apical prolateral tibia is found in Y. diversipes comb. n., A. hirschii, A. minatrix, A. lynnae sp. n., and in A. caei sp. n., and was coded as state 0 since it cannot be considered as a true apophysis.

57. Tibial apophysis on leg I, shape: (0) 2 branches; (1) 1 branch with a megaspine; (2) 1 branch with setae (Figs 310–311).

Males of many barychelid and theraphosid species have tibial apophysis composed of two branches on leg I (Raven 1985) (state 0). This type of apophysis should have been present on the ancestral Theraphosoidina, suffering modifications and losses in all derivate groups (Raven 1985). This condition is present in Theraphosinae(represented here in the analysis by Lasiodora sp. and P. vulpinus), in Ischnocolinae (here represented by H. rondoni) and in Paratropididae (represented by Melloina santuario). The presence of tibial apophysis composed of single branch with a megaspine (state 1) is found exclusively in Harpactirinae (Pterinochilus sp. in our analysis) (Gallon 2003).

The presence of a tibial apophysis with a single branch bearing spiniform setae (state 2, figs 310–311) is a condition widely distributed in Aviculariinae. All Pachistopelma spp., Iridopelma spp., Caribena gen. n. spp. as well as A. rickwesti comb. n. and most Avicularia species have this condition. This condition also appears in outgroup taxon Haplopelma minax (Ornithoctoninae).

58. Tibial apophysis on leg I, one branch with setae, spiniform setae on a branch, branch development; (0) weakly-developed branch (Fig. 310); (1) well-developed branch (Fig. 311).

We detected different degrees of branch development in species that have a single and well-defined branch covered with spiniform setae (state 1 for character 57). The branch can be weakly-developed, not projecting far from tibia longitudinal axis. This is the condition of Pachistopelma spp., Iridopelma spp. and Caribena gen. n. A well-developed branch projected far from tibia longitudinal axis is present in most Avicularia species (A. avicularia, A. rufa, A. juruensis, A. variegata stat. n., A. taunayi, A. purpurea and A. merianae sp. n.), in A. rickwesti comb. n., and also in Haplopelma sp.

59. Tibial apophysis in leg II, occurrence: (0) absent; (1) present.

60. Type I urticating setae, occurrence: (0) absent; (1) present.*

61. Type II urticating setae, any life stage, occurrence: (0) absent; (1) present.

Despite females of Pachistopelma spp. lacking type II urticating setae, immature stages and males of both Pachistopelma species do have the type II urticating setae. As we do not know immature stages nor males of Iridopelma marcoi, this species was coded as “?”.

62. Type II urticating setae, adult female, occurrence: (0) present; (1) absent.

Mature females of Pachistopelma spp. and of I. marcoi lack type II urticating setae.

63. Type II urticating setae, length, females: (0) up to 0.90mm (Figs 16–17); (1) more than 1 mm (Fig. 243).

According to Cooke et al. (1972), urticating setae type II are approximately 0.5–1.5 mm in length. Bertani et al. (2003) observed that C. versicolor comb. n. has type II setae that is more slender and longer than the Avicularia species—up to 1.37 mm in females. In fact, this species as well as C. laeta comb. n. has longer and slender setae when compared with Avicularia spp., what was confirmed by comparison of scanning microscopy images of setae of both genera (Figs 17–18).

When urticating setae are analyzed, the gender of specimens examined should be considered because males have setae significantly longer than females (Stradling 1978, Bertani and Guadanucci 2013).

64. Type III urticating setae, occurrence: (0) absent; (1) present.

65. Type IV urticating setae, occurrence: (0) absent; (1) present.*

66. Type V urticating setae, occurrence: (0) absent; (1) present.

67. Legs, ratio between length of leg IV and I, males: (0) leg IV more than 10% longer than leg I; (1) leg IV roughly the same length as leg I; (2) leg IV more than 10% shorter than leg I.

68. Legs, ratio between length of leg IV and I, females: (0) leg IV more than 10% longer than leg I; (1) leg IV roughly the same length as leg I; (2) leg IV more than 10% shorter than leg I.

69. Habits, females: (0) retreat within surface layers of soil; (1) arboreal; (2) opportunistic; (3) fossorial.

70. Arboreal retreat made by adults: (0) built on tree trunk or on palm tree leaf base; (1) built in leaves, normally with two or more leaves connected by silk; (2) built under loosened tree bark; (3) bromelicolous.