Revised generic placement of Brachypelma embrithes (Chamberlin & Ivie, 1936) and Brachypelma angustum Valerio, 1980, with definition of the taxonomic features for identification of female Sericopelma Ausserer, 1875 (Araneae, Theraphosidae)

Abstract The tarantula genus Sericopelma was originally defined based on male specimens, most notably lacking tibial spurs on leg I. Early female specimens were unrecognised as Sericopelma, and typically placed in Eurypelma – a dumping ground for problem specimens. The first females were only later recognised, but authors failed to adequately define female Sericopelma. Here, the holotypes of the southern-most alleged Brachypelma species, Brachypelma embrithes (Chamberlin & Ivie, 1936) and Brachypelma angustum Valerio, 1980 were examined, and finding both to possess defining characteristics of Sericopelma were transferred. The taxonomic attributes to define Sericopelma relative to Brachypelma and select other Neotropical genera are discussed, especially for females. As important diagnostic characters for Sericopelma, the single (unilobar) spermathecae swollen at the apex forming a P-shaped cross-section, metatarsus IV with trace scopula, femur IV with a dense retrolateral pad of plumose hair, plus other attributes. Some past confusion in these characters are clarified and Sericopelma relative to Brachypelma and Megaphobema mesomelas are discussed. Finally recommendations are given about these taxonomic changes for CITES regulations.


Introduction
was established for a male tarantula from an unspecified location in Panama without leg I tibial apophyses, named S. rubronitens Ausserer, 1875. Sericopelma was originally a subgenus of Eurypelma Koch, 1851, but later given full generic status (Simon 1892). Karsch (1880) also described an early male tarantula from Chiriquí Panama without leg I tibial apophyses as Theraphosa panamana Karsch, 1880. In revision, Simon (1892) synonymized T. panamana into Ausserer's Sericopelma rubronitens, emphasizing the lack of male tibial apophyses. He also considered that another male in the Paris collection from Chiriquí, Panama, might be the same. Like Karsch, he drew on similarities to the genus Theraphosa, where males of T. blondi (Latreille, 1804) also lack tibial apophyses, but distinguished the genera by several other features such as bulb shape, eye ratios, and cephalothorax dimensions. Soon after, Pickard-Cambridge (1897) described another species from four males also collected around Chiriquí province in Panama, which he named Sericopelma commune F.O.P.-Cambridge, 1897. He distinguished Sericopelma by femur IV "with a thick scopuliform pad on inner side", male tibia I without spurs, and emphasized the lack of scopulae on protarsus (metatarsus) of leg IV "with no thick scopulae on the inner side". Pocock (1901) again treated Sericopelma as congeneric with Theraphosa, but was not subsequently accepted.
Throughout the early twentieth century, only male Sericopelma were formally known and females remained unrecognised. Simon had described Eurypelma panamense Simon (1891) from a female with the vague locality of "Panama, Guatemala" emphasising conspicuous scopulae on femural leg IV, but failed to recognize it as Sericopelma (see Gabriel 2009). Schiapelli and Gerschman de Pikelin (1967) then evaluated both sexes of a Sericopelma sp. from "Rio Grande, Nicaragua" (? = Río Grande de Matagalpa) and illustrated the first female spermathecae. Next, Valerio (1980) described seven new Costa Rican species including three from both sexes, namely Sericopelma generala, S. immensum and S. silvicola, but only males for S. dota, S. ferrugineum, S. melanotarsum and S. upala. Following Schiapelli and Gerschman (1967), Sericopelma was characterized in Valerio (1980) by the "presence of a thick scopula on the inner side of femur IV, and by the absence of spurs on tibia I [of males], and by the absence of stridulatory setae on trochanter I, and [absence of] scopula on metatarsus IV". Smith (1991b) then re-described a syntype male of S. commune and illustrated the spermathecae of a female Sericopelma sp. in the BMNH collection. He suggested the latter was the un-described female of S. commune, and although stating "not a species description", it has been subsequently treated as such (i.e. World Spider Catalog 2015). We deduce that Smith (1991b) was referring to a female from Pozo Azul de Pirrís, Costa Rica, assigned by Valerio to S. immensum [see discussion]. Schmidt (1994) described the exuvia of a female as S. melanotarsum, illustrating the spermathecae, but did not give collection locality nor list any museum deposit. Most recently, Gabriel (2009) transferred the Panamanian Sericopelma panamense (Simon, 1891) from Eurypelma, illustrating the holotype spermathecae plus of another Panamanian Sericopelma sp. from Boquete, Chiriquí province, whilst Gabriel and Longhorn (2011) illustrated the spermathecae of a Sericopelma sp. from Bocas del Toro province, Panama. Finally Andre and Esche (2011) showed the spermathecae of S. melanotarsum alongside other morphological data, plus substantial ecological, behavioural, and captive breeding data. However, despite these studies, Sericopelma as a whole remains poorly defined.
The genus Brachypelma Simon, 1891 was created for Mygale emilia White, 1856, originally listed from Panama. However, this location is erroneous, as the natural distribution of the type species and allies is South-western México (i.e. Smith 1994, Locht et al. 1999, Schmidt 2003. Brachypelma is currently said to range from México to Panama, though the southern-most species have not been revised until now. Pickard-Cambridge (1897) could not distinguish Brachypelma from Eurypelma, and considered the genera synonymous, describing other species such as B. smithi (F.O.P.-Cambridge, 1897), which has become a flagship for conservation efforts under the Convention on International Trade in Endangered Species (CITES). Eurypelma was partly dismembered by Pocock (1903), who recognised the importance of plumose hairs on leg I and palp to define Brachypelma, whilst Simon (1903) admitted Eurypelma was previously insufficiently characterised. Brachypelma was considered valid by Valerio (1980) who described three new species from Costa Rica, B. albopilosum Valerio, 1980, B. fossorium Valerio, 1980and B. angustum Valerio, 1980. Soon after, Smith (1986 formalised additional transfers from Eurypelma to Brachypelma. Valerio (1980) had previously also transferred the Costa Rican Eurypelma mesomelas O.P.-Cambridge, 1892 into Brachypelma and described the female. Smith (1986Smith ( , 1987 agreed, but not Schmidt (1991a/b), who further transferred it to Megaphobema despite objections by Smith (1991a/b). Schmidt (1993Schmidt ( , 2003 continued to list this as Megaphobema mesomelas, as does the current World Spider Catalog (World Spider Catalog 2015).
Here taxonomic placement of some Costa Rican and Panamanian species is re-evaluated. Petrunkevitch (1925) had previously recorded several alleged Eurypelma from Panama, listing some as species now placed in Brachypelma (namely emilia, sabulosum and vagans) since known only from México, Guatemala and Belize (Smith 1994, Locht et al. 1999. Chamberlin and Ivie (1936) went further and described a species from Barro Colorado Island [Panama] as Eurypelma embrithes, placing it in that genus without explanation. Already, the robustness of Eurypelma should have been suspicious, as many species had been placed there without justification. Petrunkevitch (1939) later considered Eurypelma as "genus incertum and invalidum", although was treated as valid by Roewer (1942). Raven (1985) went on to regard Eurypelma as a junior synonym of the arboreal Avicularia Lamarck, 1818. Consequently several species were transferred to Avicularia that clearly did not belong there. Schmidt (1993) instead transferred several former Eurypelma into Aphonopelma, leading to the new combination Aphonopelma embrithes (Chamberlin & Ivie, 1936) although gave no justification, nor apparently examined any relevant types. Smith (1994) relocated embrithes to Brachypelma after reviewing many historical specimens, but did not explain his placement of this Panamanian species, thereby becoming the southern-most representative of Brachypelma. However, B. embrithes has since been listed as such (e.g. World Spider Catalog 2015) and receives legal protection under CITES legislation. However, much taxonomic revision is necessary for this protected genus in the context of others genera such as Sericopelma, although Smith (1994), Locht et al. (1999) and West (2005) have each made valuable contributions. Here, type material of B. embrithes and B. angustum are re-examined and their taxonomic placement is reconsidered in a modern context.
Colour. Type specimen alcohol faded brown. Live freshly moulted specimens from type locality are an overall blackish with longer red hairs on the abdomen, with grayish hairs on the dorsal trochanter, coxae and edges of the carapace, and two converging stripes on patella in older specimens (Fig. 2). These colours fade to overall brown with subdued russet abdominal hairs after a few months and the first dry season (RG pers.obs.).
Remarks. Originally, this species was described by "Carapace is decidedly longer than wide. Median depression transverse; deep" and "barely a trace of scopula on metatarsus IV". Our examination confirmed these features, but lead us to conclude identification as Sericopelma as defined here, including presence of an apically swollen unilobar spermathecae ( Fig. 1, see also Figs 3-5, 7-9, contrast 13-16). The type locality of Barro Colorado Island is the site of a Smithsonian Institute field-centre; hence there is a large series of specimens from type locality assignable to Sericopelma embrithes (Fig. 2) in the MCZ, MIUP and PMY (supplementary material). It is possible that S. embrithes (Chamberlin and Ivie 1936) is a junior synonym of another Sericopelma sp. such as S. commune Pickard-Cambridge, 1897 or S. panamanum (Karsch, 1880). Unfortunately, the mature male of S. embrithes remains unknown. However, geographic considerations can be vital to make confident decisions about both generic and species identities as many tarantulas have narrow distributions, and we contend these older named Panamanian species were collected in distant western Panama, namely 'Chiriquí', likely the cool highlands near Volcán Baru and Boquete (Prov. de Chiriquí) where Europeans would acclimatize (rather than the small modern village of Chiriquí, Prov. de Chiriquí). Conversely, S. embrithes from Barro Colorado Island (Prov. de Panamá) is within the central Canal Zone, a distance of over 300 km from 'Chiriquí' (Specifically ca. 320 km from Panama City to Boquete).
Anterior row procurved, posterior row recurved. Eyes, ALE > PLE, PLE > AME, AME > PME. Clypeus, 0.5, clypeal fringe long. Fovea, deep transverse. Maxillae, with 80-100 cuspules, covering approximately 60% of proximal edge. Labium, length 2.9, width 3.7, with 21 labial cuspules (a bald area in the centre of the labium lacks sockets for cuspules and may indicate previous damage, this cannot be confirmed until further specimens are examined) most separated by less than 0.5-1 times the width of a single cuspule. Labio-sternal mounds separate. Sternum damaged, narrow, length 10.2 (approx), width 8.4 with three pairs of sigilla. Femur IV with a dense pad of plumose hair on retro-lateral surface, pro-lateral surfaces of trochanter/femur of anterior legs lacking stridulatory setae. Tarsi I-IV densely scopulate, tarsus IV with spines along central axis. Metatarsal scopulae, I 84%, II 78%, III 35%, of the length of the segment, IV lacking scopulae. Lengths of leg and palpal segments see Table 1. Spination: femurs I, II, , v 3-5-10 (6 apical), IV d 6-5-4, v 8-11-16 (6 apical). Posterior lateral spinnerets with three segments, basal 3.9, medial 3.2, digitiform apical 5. Remarks. The holotype is now fragmented (Figs 6-11) and right legs II and III both appear to have been lost in life as coxal stumps are blackened indicating wound healing. Accession data from UCR and jar labels specify the holotype was collected on 01-Oct.-1974 by Edgar Vargas, but this information was not given by Valerio 1980. In the holotype jar of S. angustum a label "iqual a Sericopelma upala (?) CEV 13 julio 83" (Fig. 6) shows Valerio himself (= CEV) had doubts about placement in Brachypelma, also considering it conspecific to the male he described as S. upala. The type localities are close, less than 50 km apart in Alajuela with similar ecotypes of lowland tropical forest, now largely fragmented to sugarcane plantation and cattle pasture (S. Longhorn pers. obs). However, until further specimens of Sericopelma upala and/or S. angustum are examined, we are not prepared to place them into synonymy at this time. We suspect Valerio (1980) lacked sufficient access to Brachypelma material to make a more informed decision about the genus, failing to recognise defining characteristics (as outlined below). Remarks. Smith (1991b) refers to three of four male syntypes from Chiriquí as S. commune, specifically BMNH 1898-12-24-19-21 (i.e. accessioned 24 th -Dec-1898, coded '19-21'), then described a female, saying "Female BMNH 98-12-24-22. Assigned to the species by Valerio". The only female BMNH specimen with this accession has the oldest label "Museo Nacional de Costa Rica, Pozo Azul de Pirrís, José C. Zeldón", naming a collector from the 1890s. A later label "Sericopelma immensa n. sp. Det. C. E. Valerio, Jan 10, 1979" matches his paper (Valerio 1980) referring to a BMNH specimen from this same locality as S. immensum. However, the species on the Valerio label has been physically scored out, but likely reads immensa. Another pen-written label says "Sericopelma commune F.O Pick-Cambr." (in handwriting of curator Doug Clark, died 1972), apparently present when both Valerio and Smith examined the specimen. We suspect this label misled Smith (1991b) to reconsider the specimen as the un-described female S. commune, even though collected at a Costa Rican locality (Parrita Cantón, Puntarenas), approx. 250 km from the Chiriquí type site. However Smith only records the distribution (indicating both sexes) from Chiriquí, Panama. Further confusion occurs with another mature male in BMNH with an old pencil-written label "Panama", then two pen labels in Clark's handwriting, "S. commune PDA Costa Rica BMNH 1898-12-24-22" and "Sericopelma commune det. Clark 1960". We suspect these latter labels were an attempt by Clark to wrongly allocate this "Panama" male to both the Pozo Azul de Pirrís accession, and as a 'missing' fourth male syntype of S. commune. Clark perhaps did not realise that fourth male is in the Pickard-Cambridge collection at OUMNH, where a male labelled 'syntype' had the unequivocal label "Sericopelma communis Fopc Chiriqui -Champion". In a BMNH accessions book, 1898-12-24-22 corresponds to "Sericopelma sp? Pozo Azul de Pirrís (Costa Rica). Pres. by F.D. Godman, Esq., Costa Rica Mus, F.O.P.-Cambridge". However, although F.O. Pickard-Cambridge apparently recognised it as a possible female Sericopelma sp, the lack of accounts before Valerio (1980) indicate it was ignored, perhaps due to uncertainty about matching it with known males. We consider this female to be the same listed by both Valerio (1980) and Smith (1991b) and suggest its unsecure designation as the first described female of S. commune be suspended, instead to favour topotypic specimens from Chiriquí, such as the region of Volcán where G. Champion likely collected the four male syntypes. Distribution. Only known from type locality, Chiriquí = Chiriquí, Provincia de Chiriquí, República de Panamá. (Karsch, 1880) Remarks. Simon (1892) makes no clear justification why Karsh's T. panamana from Chiriquí should be synonymous with S. rubronitens, only referring to similarities in eye pattern and absence of tibial spurs in Karsh's description against another nontype male specimen in the Paris collection, which he had assigned as S. rubronitens. Our re-examination of the type specimen confirmed its designation as a Sericopelma sp., but not its synonymy with S. rubronitens, which is here reversed.

Geographic distribution, and generic limits
We believe it is important to re-clarify the characteristics of Brachypelma in this context. The type of Brachypelma is B. emilia, originally suggested in the paper's title to be from Panama (White 1856). A later male from México: Ventanas, Prov. de Durango (leg. Forrier) was described by Simon (1891) as generic type. Smith (1994) incorrectly says "Simon lists his specimen as coming from Panama" (p.166). We suspect this stems from a mis-listing by F.O.P-Cambridge (1897) of "PANAMA (coll. Simon: Male)" where locality was confused with the original type. While the original specimen appears lost (Smith 1994) or 'non-existent' (Pickard-Cambridge 1897), an excellent illustration in White's paper allows identification, showing an adult male with tibial spurs. The route of the collector (Berthold Seemann) is well known (Seemann 1852), joining his ship in Panama and voyaging north along the Pacific, docking in México both at San Blas (Estado Nayarit) and Mazatlán (Sinaloa). The original type taken to the BMNH was most likely collected during the second inland foray in 1849/50 to Ciudad de Durango (modern Victoria de Durango, Durango) and Tepic (Nayarit). However, another male deposited in MNHN was used as generic type of Brachypel-ma, from Ventanas (leg. Forrer). This is likely modern Villa Corona, Estado Durango (DMS = 23°52'51"N, 105°46'19"W) (Selander and Vaurie 1962), but concurs both with the route of Seemann (within 15 km from Mazatlán to Ciudad de Durango) and with modern understanding of the species distribution across Sinaloa, Durango and Nayarit (Locht et al. 1999). Pickard-Cambridge (1897) mentions two males by Mr. Forrer from Ciudad [modern Victoria de Durango] plus Simon's male from Ventanas, but none specifically as neotype. One adult male which Smith (1994) refers to as neotype was accessioned in NHM as BMNH 98-12-24-32 where it is labelled 'leg. Forrer' plus 'Ciudad'. The second adult male is in the Pickard-Cambridge collection at OUNMH (Jar 65), with the same collection details of 'Ciudad. Mex, Forrer', plus labelled 'paratype A.M. Smith'. However, we argue preference could have been given to the generic type of Simon from Ventanas. Simon (1891) also referred to a female specimen, though Pickard-Cambridge (1897) stated the female is unknown. However Smith (1994) gives a comprehensive description of both sexes, using a later Figure 12. Geographic distribution of the genus Sericopelma from published records (including this study), where complete black-centred shapes are for specimens examined during this study, whilst gray shapes [outlined in black] are further specimens listed by Valerio (1980), accordingly data for S. immensum has black shapes (for the holotype, allotype and further female from Pozo Azul de Pirrís examined here), and gray shapes for further sites of Valerio. S. rubronitens and S. panamense are of unspecific location, but canal-zone seems likely. female BMNH 1962-2-28-1, and as a result the taxonomic identity of this species is clear. Simon (1891) originally emphasized several characters for Brachypelma, including presence of distinct scopula on the metatarsus, and femur IV without inner scopula (i.e. no dense pad of plumose hairs), instead long and simple hairs ("metatarsus paris scopula crassa medium articulum fere attingente munitus, femora postica haud scopulata intus longe et simpliciter pilosa", Simon 1890). The genus is also characterised by plumose hairs on the prolateral face of leg I trochanter/femur and retrolateral face of the palp (Pocock 1903). These features have been supported by subsequent authors as diagnostic for Brachypelma (e.g. Smith 1994), such as both sexes without a plumose pad on leg IV femur, the metatarsus IV distally one-third to one-fifth scopulate, and no tarsal division by stiffened setae, along with male palpal bulb distally wide and flattened (spoon-shaped), two unequal spurs on male tibia of leg I, females with a simple undivided/fused spermathecae (Figs 13-14) which we further clarify have a flat crosssection. Despite some earlier confusion about the types, the type species B. emilia is well defined, and the genus is easily separated from Sericopelma. The geographic range of Brachypelma is securely centred in south-western Mexico, now with B. albopilosum and B. fossorium at its southern-most limit in Northern Costa Rica. Due to the generic transfers here of S. angustum and S. embrithes (and comments below on other specimens), there are now no reliable records of the genus Brachypelma in Panama. The transfers proposed here verify that the Brachypelma as currently defined ranges from Mexico to north Costa Rica, and is not native in Panama or further south.
Geographic distribution of Sericopelma. From examination of specimens (see methods and supplement), combined with data we consider reliable in Schiapelli and Gerschman (1967) and Valerio (1980), we consider that Sericopelma ranges from Nicaragua to Panama (Fig. 2), with the northern-most report from Nicaragua. This was confirmed by examination of a single male specimen from Matagalpa, Nicaragua held in MCZ.
We regard the inclusion of 'Guatemala' in the original type locality of S. panamense from 'Panama, Guatemala' as an error, and suggest that 'Guatemala' instead refers to the locality for a second specimen (actually from another genus, and seemingly not of a taxon from Panama) which we found in the same jar from the Paris collection.
Note: The extralimital Brazilian Sericopelma fallax Mello-Leitão, 1923 is considered misplaced (see Gabriel and Longhorn 2011). * Originally listed as Panama and Guatemala, though the latter is unlikely. ** Originally simply listed as Panama.

Discussion
Prior to Valerio (1980) the diagnostic features for Sericopelma were poorly known, with males primarily recognised by the palpal bulb shape and absence of tibial apophyses (Ausserer 1875, Karsch 1880, Simon 1891, while females were unrecognized until Schiapelli and Gerschman (1967). Over-reliance on the lack of male tibial apophyses led many museum specimens to be mislabelled and misplaced. In actuality, Simon (1891) had described the first female Sericopelma as Eurypelma panamense, but unrecognized until Gabriel (2009) rediscovered it as a former Eurypelma, a genus that Raven (1985) had described as a taxonomic "dumping ground". We now confirm that Chamberlin and Ivie (1936) misplaced another female into Eurypelma, here transferred to Sericopelma embrithes (Chamberlin & Ivie, 1936). As the female characteristics of Sericopelma have long been uncertain, the female description by Smith (1991b) was valuable to resolve uncertainty about spermathecae characteristics. Schiapelli and Gerschman (1967) illustrated the first spermatheca of a probable Sericopelma from Nicaragua (Fig. 3) [Nb. specimen not seen]. Their relatively poor illustration shows possible indentations or notches on the apex, which appears atypical of the genus. However, we confirm that Sericopelma indeed exists in that region from another examined male Sericopelma sp. in MCZ with the label "Matagalpa, Nicaragua". Valerio (1980) described seven species from Costa Rica, only illustrating the spermathecae of both S. immensum and S. silvicola as simple domes, and neither shows any such notches. Neither do spermathecae of Smith (1991b) nor Schmidt (1994) show any such notches. Perez-Miles et al. (1996) reproduced the Schiapelli and Gerschman (1967) illustration, stating female Sericopelma have "a single spermathecae receptaculum with a median notch", plus key "19. Female with notched spermathecae". Schmidt (2003) also referred to the Sericopelma spermathecae as "Einteilige flache" (i.e. single flat) using the same illustration, not mentioning any apical notches or indentations. We regard the 'notched spermathecae' of Schiapelli and Gerschman (1967) as misleading, and its use to define female Sericopelma as erroneous. We find that mature female Sericopelma spermathecae lack any distinct median notch (Fig. 4) and furthermore, are distinctly swollen on the apex producing a diagnostic P-shape when viewed in profile (Fig. 5), which is also diagnostic for most immature Sericopelma females. We suggest this apical swelling probably expands with age (i.e. ontogenetic modification). Although the holotype spermathecae of S. angustum does have a slight medial concaved indentation, we consider this unique. It also shows the diagnostic swollen apex with P-shaped profile diagnostic for Sericopelma. The swollen apex is not found in the other Neotropical theraphosid genera where females have a single unilobar spermathecae, instead flattened or apically narrowed cross-section, such as Brachypelma Simon, 1890, Megaphobema Pocock, 1901and Theraphosa Thorell, 1870. Female Sericopelma can be distinguished from Eupalaestrus Pocock, 1901, Vitalius Lucas, Silva & Bertani, 1993, Nhandu Lucas, 1983, Pamphobeteus Pocock, 1901 and Xenesthis Simon, 1891 by the unilobar spermathecae lacking two separated apical projections (Bertani 2001), and from Mygalarachnae Ausserer, 1871, by the unilobar structure lacking a broad median notch (Gabriel and Longhorn 2011).
Along with spermathecae attributes, Sericopelma can now be defined by; Carapace longer than wide (Ausserer 1875, Karsch 1880, Simon 1892, Pickard-Cambridge 1897, Schiapelli and Gerschman 1967, deep transverse fovea (Ausserer 1875, Karsch 1880, Pickard-Cambridge 1897) and distinct radiating sulci (Ausserer 1875). We confirm these attributes as useful for both sexes, although carapace is more rounded in mature males than females. Another useful diagnostic is few/weak metatarsal scopulae on distal leg IV forming two distinct pads, elsewhere defined as "barely a trace of scopula on metatarsus IV" (Chamberlin and Ivie 1936), "not scopulate, or very slightly so at the apex" (Pickard-Cambridge 1897), or absent (Ausserer 1875, Simon 1892, Valerio 1980, Schiapelli and Gerschman 1967. Here we confirm that almost every examined specimen of Sericopelma actually does have trace of scopulae on the distal leg IV metatarsus, most forming two small distinct pads when viewed ventrally (Fig. 17, in most extensive form). Such 'trace scopulae' are typically present on in both mature sexes, but in some specimens are distinct while in others greatly reduced. The fresh specimens that lacked trace scopulae were smaller juveniles, suggesting the feature may become

18
more conspicuous through development. Trace scopulae were absent on some larger specimens, but only when eroded through wear or damage. Our examination of S. angustum confirmed trace scopulae on leg IV metatarsus as with other Sericopelma, unlike the one-third to one-fifth scopulae present in Brachypelma. From a large array of specimens (see Supplement), female Sericopelma may be robustly defined by: Spermathecae single (unilobar), swollen at the apex to form a P-shaped cross-section, femur IV with a dense retrolateral pad of plumose hair, trochanter/femur of leg I lacking stridulatory setae, carapace longer than wide, deep transverse fovea and distinct radiating sulci, ventral metatarsus IV with a divided and reduced trace of scopulate hairs at the distal end. Apart from spermathecae attributes, these remaining features also define mature males along with the absence of tibial spurs and characteristic embolus shape.
The dense retrolateral pad of plumose hair on femur IV is another useful character to separate Sericopelma from Brachypelma. We clarify the term 'femoral scopula/e' in Sericopelma as a broad pad of plumose hairs. Valerio (1980) defined Brachypelma with "Scopula in femur IV inconspicuous or absent", as did subsequent authors (Smith 1994, Schmidt 2003). Yet Valerio (1980) had previously confirmed that femur IV of B. angustum does indeed have a modified patch of hairs, by "Femur IV con cojinete medial" (p. 270), and elsewhere confirmed Sericopelma indeed posses such. Our examination of the S. angustum holotype (Figs 6-11) showed a broad pad of plumose hair on retrolateral femur IV (Fig. 10) as in other Sericopelma spp., but not Brachypelma. Schmidt and Krause (1994) reported that Brachypelma klaasi is exceptional with a "thin pad of plumose hairs on femur IV", used to support a new genus Brachypelmides, since rejected. They gave no indication of which sex was examined nor where femoral hairs were found. We therefore also examined mature B. klaasi specimens of both sexes and found no distinct pad on retrolateral femur IV, just a few sporadic fine-hairs slightly plumose basally, near the distal femur. We suggest these conform to the diagnostic 'short weak-feathered hairs (= kurze schwachgefiederte Haare) of Schmidt and Krause (1994), but do not form any distinctive pad as in Sericopelma (as S. angustum and S. embrithes). Instead in B. klaasi, these modified hairs are interspersed among more numerous long-fine hairs and thicker bristle-like hairs. Further, there is a baldline forming a longitudinal strip along the axis in B. klaasi, observable in both fresh and alcohol preserved specimens, contrasting with the dense pad of plumose hairs in Sericopelma. Modified hairs of B. klaasi hind-femurs were difficult to distinguish on alcohol-preserved specimens, so we also examined dried exuvia as Schmidt and Krause (1994), where fine-basally plumose hairs were more easily detected. Other examined Brachypelma spp. only showed fine hairs and bristle like hairs on femur IV, as reported for B. albiceps by Locht et al. (1999).
With a more robust definition of Sericopelma (including female characteristics), we can be increasingly certain about generic boundaries. Valerio (1980) defined Sericopelma by "the presence of a thick scopula in the inner side of femur IV, the absence of spurs on tibia I, [absence of] stridulatory setae on trochanter I, and [absence of] scopula on metatarsus IV". Also "One spermathecae, semicircular, sometimes with lateral extensions, covered with fine spinules", or as "Receptaculum seminis opens on dorsal side of apical region, communicating with distal tip of bulb by and open groove." This may be alluding to the apically swollen P-shaped cross-section that we consider diagnostic for Sericopelma. Valerio appears to have been misled by the central depression he characterised as "Spermathecae with a shallow notch in anterior edge (Fig. 7 [his figure 19])", leading him to recognise similarity with B. albopilosum, and misdiagnosing them both as Brachypelma by shared "Spermathecae with a conspicuous depression on the anterior edge". Our examination of S. angustum showed the spermathecae indeed possesses a slight medial indentation, but less defined than Valerio suggested, and we further recognise the apical swelling with a P-shape cross-section (Figs 8, 9) as diagnostic of Sericopelma. Spermathecae of other genera like Brachypelma (Figs 13-16) are flat throughout in cross-section. Further, S. angustum does not have any plumose hairs on the prolateral trochanter or femur of leg I (or II), nor the retrolateral palpal trochanter (i.e. Smith 1994, Schmidt 2003, but does have a distinctive pad of plumose hairs on femur IV (Fig. 10), together confirming it as Sericopelma, representing a unique species due in part to distinctive spines on tarsus IV (Fig. 11).
During this study, we found many historical museum specimens with mistaken identities, most importantly several wrongly reported as Panamanian Brachypelma. Petrunkevitch (1925) listed Sericopelma commune, 1 male and 1 female from the Canal zone. Sericopelma rubronitens from 2 females from Culebra (probably Pacific Canal Zone, 'Gaillard Cut'), and 2 females from Bocas del Toro. As discussed above, S. commune was described from males collected in distant Chiriquí, hence the identity of his Canal Zone species is dubious. Petrunkevitch did not compare his specimens to the earlier male types (nor could he with females), so his determination of various females as S. rubronitens cannot be regarded as reliable descriptions. Our confidence in Petrunkevitch determinations is greatly reduced as he also misidentified other geographically diverse specimens as S. rubronitens, all from outside the geographic range of the genus Sericopelma, such as from México, Haiti, and Ecuador (see supplement for re-evaluation), probably as all were similarly coloured with dark bodies and reddish abdominal hairs. He also inconsistently referred to specimens from Barro Colorado Island as either S. rubronitens or S. commune (see supplement), despite being the type locality for S. embrithes. Petrunkevitch (1925) Petrunkevitch (1925) merely repeated the erroneous location from the original description. Interestingly, some male Sericopelma from Chiriquí do superficially resemble B. emilia by light pinkish lower legs and carapace, plus black triangle on the carapace, perhaps leading to early confusion. On re-examination of the Petrunkevitch specimens in PMY, his alleged B. sabulosum was a Sericopelma sp, as likely are the 4 immatures of alleged B. vagans. The immatures are pre-dispersal nymphs, with the wrong proportions for B. vagans -where nymphs are almost one fifth of this size. In B. vagans, the legs remain proportionally shorter even when older post-dispersal 'spiderlings' of equivalent size (Fig. 18). The most likely genus for these large nymphs is Sericopelma. The alleged female B. vagans was not located, but we also expect to be a misidentified Sericopelma, which can be similarly coloured and often confused by non-specialists. Distribution of B. sabulosum and B. vagans from Panama should be regarded as mistaken, B. sabulosum is only validly recorded from Guatemala, whilst B. vagans is recorded from México, Belize and Guatemala.
Finally, another allied Costa Rican species with long ambiguous placement is Megaphobema mesomelas (O.P.-Cambridge, 1892), again originally placed in the poorly defined Eurypelma. Valerio (1980) described the first female and transferred it to Brachypelma before Schmidt (1991a/b) transferred to Megaphobema. Smith (1991b) also reevaluated the species, drawing tarsus IV with twin central lines of modified setae (his figure 6), which Valerio had recognised as "Cojinete del tarso IV dividido por varias filas de espinas". Against this, we considered S. angustum where scopulae are interspersed by thickened spines (Fig. 11), which we consider species specific -as not observed in other Sericopelma, nor mature specimens of other candidate genera. However, our inspection of various recent (both sexes) and historical specimens of M. mesomelas (including the male holotype and another male from same collector in the O.P.-Cambridge collection), each revealed only few long soft hairs on tarsus IV, not thickened spines. Our re-examination of M. mesomelas lead us to agree it does not belong in Brachypelma, nor Sericopelma, but neither do we agree with placement in Megaphobema (Gabriel and Longhorn, in prep). Female Sericopelma can be distinguished from Megaphobema by the form of the spermathecae, in the latter by greater ventral surface sculpturation with striated grooves more evenly spaced and extending to lateral edges, or a more cerebriform pattern, plus flatter cross-section (Fig. 15). Mature males of Sericopelma lack tibial apophyses (as do some other genera), but are present in Megaphobema (and other genera). Both sexes of Megaphobema also can be distinguished from Sericopelma by more extensive scopulae on metatarsus IV. For M. mesomelas the sternum is especially narrow and elongate, which Smith (1987) says "over twice as long as wide". We agree, observing the M. mesomelas sternum is more extremely narrowed than S. angustum. The narrowed form in both conflicts with Brachypelma, defined by a broad sternum (i.e. Simon 1891, "Sternum aeque longum ae latum"). S. angustum was diagnosed by Valerio (1980) by "Carapace longer than 18.0 mm" or "Carapace very narrow (1.6 times longer than broad)", and his specific epithet 'angust' (= narrow) refers to both the narrow cephalothorax and sternum. We suggest the narrowed sternum can be indicative of close evolutionary affinities of M. mesomelas with Sericopelma, particularly S. angustum.

Consequences for conservation, including CITES
Currently, all Brachypelma species are protected by international commercial trade regulation (CITES, Appendix II). Transfer of S. embrithes and S. angustum into Sericopelma means that consequently these species may now only be protected by national wildlife laws. However, there does not appear to be a current need to regulate trade in S. embrithes and S. angustum, so we assert both species should indeed be removed from CITES listing. As with most theraphosids, the major threat appears to be habitat destruction. For S. angustum, much of its probable habitat in northern Costa Rica has already been disrupted by human activity, often for sugar cane plantations. However, its conservation status within Costa Rica must be urgently evaluated. For S. embrithes, much of its original range was likely destroyed during the damming of the Chagres River for the Panama canal, isolating Barro Colorado Island. A more deserving candidate for CITES regulation is Megaphobema mesomelas; a large brightly coloured species which has regularly been targeted by illegal collection for commercial gain, and traded internationally. We also point out there remains need for continued regulation of all Brachypelma sp. traded as exotic pets, including those in the pet-markets still exchanged under the former name 'Brachypelma angustum', which would retain their CITES protected status under the aegis of Brachypelma sp.