Branchiosyllis, Haplosyllis, Opisthosyllis and Trypanosyllis (Annelida: Syllidae) from Brazil, with the Description of Two New Species

Brazilian specimens of Branchiosyllis cf. exilis, B. tamandarensis sp. n., Haplosyllis lattigae sp. n., H. loboi, Opisthosyllis brunnea and O. viridis are described and illustrated herein, from recently collected material; also, the distributions of Haplosyllis amphimedonicola and H. rosenalessoae are expanded to other localities in the states of Paraíba and Pernambuco. Branchiosyllis tamandarensis sp. n. was found associated with sponges and is characterized by having a flattened, ribbon-like body, with longitudinal line of mid-dorsal papillae, peristomium dorsally inconspicuous, branchiae with up to six lobes, branchiae and ungulae on all parapodia, and falcigers absent. Haplosyllis lattigae sp. n. is characterized by having two kinds of chaetae with different sizes and shapes per parapodium, papillate dorsum from midbody, and midbody dorsal cirri alternating in length. Additionally, we provide keys to the Brazilian species of Branchiosyllis, Haplosyllis and Opisthosyllis, comparative tables of the new species described herein and selected similar congeners, and the first record for Trypanosyllis zebra in the states of Espírito Santo and Paraíba.


Introduction
Syllidae Grube, 1850 is one of the most diverse families of polychaetes, with 74 valid genera and more than 700 species [1]. The family is represented by very abundant animals living epibiontically on various substrata, from hard to soft bottoms, and from intertidal to abyssal depths [1,2]. The subfamily Syllinae Grube, 1850, with about 260 species and 20 genera [1], is monophyletic [3]. According to the classification suggested by [3], the subfamily is characterized by reproducing schizogamically by scissiparity, with distinct types of stolons, corresponding to the internal monophyletic groups. Schyzogamy is also found in the Autolytinae Langerhans, 1879, but originating independently [3]. The results of [3] also suggests a close relationship between Branchiosyllis Ehlers, 1887 and Haplosyllis Langerhans, 1879, as both genera have species living in association with other organisms and have modified chaetae, which could be considered as a particular adaptation to symbiotic life styles [2,[4][5][6][7]. In contraposition, Opisthosyllis Langerhans, 1879 and Trypanosyllis Claparède, 1864 are more closely related to the remaining genera of Syllinae. A historical approach and a general introduction to the subfamily is given by [6].
Up to the present, 8 genera and 47 species of the subfamily were recorded from Brazil [8][9][10], however, most of these records came from unpublished thesis and ecological studies [9], and several of these species were not found in recent taxonomic studies [11][12][13][14], rendering the real number of occurrences in the country difficult to access.

Material and Methods
Field permit and deposit of specimens

Study area and sample processing
The material examined for the present study came from six independent studies (Fig 1; Table 1). The first, project 'Diversity of Polychaeta (Annelida) on sandstone reefs off northeastern Brazil, states of Paraíba and Pernambuco' ('BioPol-NE'), was conducted by the Laboratório de Poliquetologia (LaPol), Instituto de Biociências, Universidade de São Paulo (IB/USP), with collections from the intertidal zone to~1 m deep on sandstones, at neap tide, from reefs off the states of Paraíba and Pernambuco, northeastern Brazil. Algae, sponges, ascidians, mussel beds and similar substrates were scrapped from the rocks. In the laboratory, the material was examined alive under a stereomicroscope, polychaetes were sorted, relaxed in menthol solution, preserved in 4% formalin, and, a few weeks later, rinsed in fresh water and transferred to 70% ethanol.
The project 'HABITATS-Environmental Heterogeneity in the Campos Basin' ('HABI-TATS') sampled in the Campos Basin, State of Rio de Janeiro, in a survey coordinated by the Brazilian energy company PETROBRAS. Collections were made from 12 to 3301 m deep in 2008 and 2009. Material was fixed in 4% formalin and later rinsed in fresh water and transferred to 70% ethanol.
The project 'REVIZEE/Southern Score/Benthos' ('REVIZEE') investigated the fauna occurring in the Brazilian Economic Exclusive Zone. Material was collected from 60 to 800 m deep, with dredges, box corers and Van Veen grabs (see [16] for collection details).
The project 'BIOTA/FAPESP/Benthic Marine Biodiversity in the State of São Paulo' ('BIOTA') studied the fauna occurring off the northern coast of São Paulo, with collections  Sand and also algae, sponges, ascidians, mussel beds and similar substrates from the intertidal zone to~45 m deep, on rocky shores, assemblages of algae and soft bottoms. For that study, material was preserved in 4% formalin immediately after collection, and specimens were posteriorly sorted out by family and transferred to 70% ethanol [17]. The sixth project, 'Diversity of Polychaeta (Annelida) on rocky shores off the State of São Paulo, southeastern Brazil' ('BioPol-SP') was also conducted by the LaPol, following the same methodology as project BioPol-NE, except for all collections having been made intertidally.
Further analyses under stereo-and light microscopes were made from specimens preserved in ethanol, some of which were permanently mounted on slides in glycerin jelly. For examination under scanning electron microscope (SEM), specimens were dehydrated in a series of increasingly stronger ethanol solutions, then critical point dried, covered with 25 nm of gold, and examined and photographed under the SEM at Laboratório de Microscopia Eletrônica, Instituto de Biociências, Universidade de São Paulo (IB/USP). Line drawings were made from slide-mounted specimens, with the aid of a drawing tube. Measurements were taken under compound and stereo microscopes. Length of the specimens was measured from the tip of palps to the tip of pygidium, excluding anal cirri; width was measured at proventricular level, excluding parapodia.
Comparative material was examined from specimens lodged at the MNCN, the Australian Museum, Sydney, Australia (AM), the Zoological Museum Hamburg, Germany (ZMH), United States National History Museum, Smithsonian Institution, Washington DC, USA (USNM) and National Museum of Nature and Science, Tsukuba, Japan (NSMT). The number of access of all material analysed in the present study can be found at sections "material examined" and additional material examined"along the text.
The nomenclature used for Haplosyllis chaetal morphology follows that suggested by [18] and [19]. The chaetal characters used here are: • length of main fang: as long as, or longer than chaetal width; • upper side of main fang: with or without denticles; • mid-joining point between apical teeth and main fang: curved (usually short) or straight (at right angle with main fang, short or long); • proximal and distal apical teeth: similar in size or one of them larger.

Nomenclatural Acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix "http://zoobank.org/". The LSID for this publication is: urn:lsid:zoobank.org:pub:6353431B-6724-42BC-9F54-D06CD5FA0F28. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.

Taxonomic Account
Family Syllidae Grube, 1850 Subfamily Syllinae Grube, 1850 Genus Branchiosyllis Ehlers, 1887 Type species: Branchiosyllis oculata Ehlers, 1887. Diagnosis. Relatively medium-sized to large body, subcilindrical to dorso-ventrally or, more rarely, laterally flattened. Palps free or fused at bases. Prostomium with four eyes, occasionally also with two anterior eyespots, and three antennae. Peristomium with two pairs of peristomial cirri. Antennae, peristomial, dorsal and anal cirri articulated. Parapodial lobes sometimes with branchiae. Compound chaetae as falcigers with regular blades and with modified, claw-shaped blades, arranged in 90°with shaft ('ungulae'), at least on part of body. Simple chaetae not known, apparently altogether absent. Pharynx with single anterior tooth; proventricle usually of approximately same length as pharynx. Reproduction by acephalous stolon.
Remarks. Branchiosyllis exilis is a cosmopolitan species, which was reviewed recently by [28], who analysed morphologically a great number of specimens from several localities around the world and divided them in two morphological groups. The authors concluded that a molecular approach is necessary to determine if there is a single species, with slight differences among populations, or if there are several sibling species grouped together under the name Branchiosyllis exilis. Considering this, Brazilian specimens of Branchiosyllis cf. exilis fit the group of the holotype (and material from Australia and Cuba), with slender and elongate cirri (longer than body width, 30-40 articles), blades of falcigers with dorso-ventral gradation in length and with both teeth similar in size. About the colour pattern, they noticed this character to be highly variable and thus not useful for identification, at least in preserved specimens. The Brazilian material matches the description provided by [28], although with slight differences on dorsal pigmentation. Considering this variation, we treated the Brazilian specimens as Branchiosyllis cf. exilis.
Branchiosyllis lorenae, B. salazari, B. maculata (Imajima, 1966) and B. thylacine San Martín, Hutchings & Aguado, 2008 also lack branchiae, have ungulae starting from midbody chaetigers and dark spots on cirri, similar to the pattern found in B. cf. exilis. The first two species are known from the Caribbean, while the remaining species are both from the Pacific [28]. Branchiosyllis lorenae differs from Branchiosyllis cf. exilis by having three ovate spots dorsally on each segment, except for the anteriormost chaetigers; peristomium dorsally reduced; falcigers with subdistal tooth shorter than distal tooth; and ungulae present only on posterior parapodia. Brazilian specimens of Branchiosyllis cf. exilis have a different colour patern, as a transverse dorsal stripe per chaetiger, sometimes with additionally diffused dark band dorsal and ventrally; peristomium dorsally conspicuous, although shorter than following chaetigers; and ungulae starting from midbody.
Branchiosyllis salazari was described by [29] in a paper which also analysed the type material of B. exilis. The authors noticed that the holotype of B. exilis does not have dorsal pigmentation, however, as already mentioned, this character is extremely variable, especially considering preserved material [28]. Branchiosyllis salazari and B. exilis are distinguished from each other mostly based on the morphology of anterior and midbody aciculae, and also on the size of the teeth of midbody falciger blades. According to description provided by [29], B. salazari has three aciculae per parapodium anteriorly, two of which straight, with acute tips, the remaining subdistally oblique, with acute tip; from midbody, each parapodium has two aciculae, one of which straight, with acute tip, the other subdistally oblique, with tip directed upwards ( Fig 3B, D in [29]); on the other hand, according to [29], B. exilis has two aciculae per parapodium anteriorly, one L-shaped and another subdistally oblique, and posterior parapodia with single acicula each, distally tapering and curved. Furthermore, the blades of falcigers on midbody parapodia are bidentate in B. exilis, with distal tooth ca. three times larger than the subdistal one (at least in ventralmost chaetae), while midbody falciger blades of B. salazari are unidentade [29]. Also, live specimens of B. salazari have different pattern of dorsal pigmentation when compared with live material of Brazilian Branchiosyllis cf. exilis (Leslie Harris, personal communication). According to an unpublished photo by Leslie Harris, B. salazari has anterior chaetigers with two brownish, dimmed, transverse bands per chaetiger, one discontinuous, mid-dorsal with lateral gaps, and another near posterior border of each chaetiger.
Brazilian specimens of Branchiosyllis cf. exilis differ from the description of the holotype provided [29] by having ventralmost falcigers subbidentate to unidentate and lacking L-shaped aciculae.
Branchiosyllis maculata differs from Branchiosyllis cf. exilis by having some completely black articles on dorsal cirri throughout, dorsum lacking pigmentation patterns, longer dorsal cirri throughout, and ungulae only present on posteriormost parapodia. On the other hand, Branchiosyllis cf. exilis does not have black articles on dorsal cirri, although some scattered black spots are common. Finally, B. thylacine can be differentiated from Brazilian specimens of Branchiosyllis cf. exilis by only having ungulae on posterior parapodia, all falcigers with unidentate blades and different aciculae, straight and with rounded tips.
Type locality. Red Sea (Indian Ocean). Distribution. Apparently circumtropical, also present in the warmest regions of the Mediterranean sea [28]. First record from off the northeastern Brazilian coast.
Remarks. Branchiosyllis tamandarensis sp. n. resembles B. oculata, B. pacifica and B. lamellifera by having flattened, ribbon-like body, branchiae present, ungulae on all parapodia, and falcigers absent (Table 3). Branchiosyllis oculata and B. lamellifera were described from Florida and Gulf of Mexico, respectively, whereas B. pacifica was described from Mexican Pacific. According to [21], B. oculata occurs in the Gulf of Mexico and Caribbean Sea, from Florida to Venezuela; B. pacifica is found in the eastern tropical Pacific Ocean, while B. lamellifera occurs in Bermuda, Gulf of Mexico, Venezuela and, probably, also Curaçao.
Members of B. oculata are dark, from brown to black, and the branchiae are dome shaped, different from B. tamandarensis sp. n., which has yellow to orange body and multi-lobed branchiae. Branchiosyllis pacifica has multi-lobed branchiae but differs from B. tamandarensis sp. n. by having shorter median and lateral antennae, with 8-9 and 10 articles each, respectively; shorter dorsal and peristomial cirri, with 16 and 10-12 articles each, respectively; shorter anterior and midbody dorsal cirri, with 14-17 and 13-17 articles each, respectively; longer posterior dorsal cirri, with 10-26 articles each and 2-4 ungulae per anterior parapodium [21]. On the other hand, B. tamandarensis sp. n. has longer appendages (i. e. antennae, peristomial, anterior and midbody dorsal cirri, Table 3) and has 4-6 ungulae on each anterior parapodium. Branchiosyllis lamellifera, according to the redescription by [21], is very similar to B. tamandarensis sp. n., in overall body shape. However, B. tamandarensis sp. n. has the median antenna inserted more posteriorly in relation to lateral antennae; prostomium slightly broader than long; appendages (i. e. antennae, peristomial and dorsal cirri throughout) more slender and elongated; midbody branchiae with up to six lobes; up to six ungulae on each anterior parapodium; pharynx through ca. 5 chaetigers and proventricle through 3.5-5 chaetigers. On the other hand, B. lamellifera has all antennae inserted nearly in line (Fig 4B, D in [21]); prostomium conspicuously broader than long; stouter appendages; midbody branchiae with up to three lobes; up to 3-4 ungulae on each anterior parapodium; and pharynx and proventricle both through 9 chaetigers.
Biology. Some specimens were found associated with the sponge Tedania ignis. However, our methodology does not allow to state that this species lives exclusively within T. ignis, instead of also being present in other different types of substrates, including other species of sponges, found close to T. ignis.
Some specimens have one mid-dorsal papilla per chaetiger, arranged in a longitudinal row at dorsal midline (Fig 10C and 10F), while in others such papillae are absent (or at least not visible). We considered this as an intraspecific variation, as it also occurs among members of B. pacifica (Fig 8A [21]), although this character was not mentioned in [21].
Etymology. This species is named after the city of Tamandaré (state of Pernambuco), where Praia dos Carneiros, the type locality, is located. Praia dos Carneiros is one of the most beautiful beaches in the world.
Distribution. Atlantic Ocean: Brazil (states of Paraíba and Pernambuco). Genus Haplosyllis Langerhans, 1887 Type species: Syllis spongicola Grube, 1855. Diagnosis. Body of variable size, usually up to 10 mm long, with numerous nearly cylindrical segments. Palps robust, fused at bases. Prostomium with four eyes and three antennae. Two pairs of peristomial cirri. Antennae, peristomial, anal and dorsal cirri throughout moniliform, the latter sometimes with a single small article on either side of posterior parapodia. Ventral cirri digitiform. Pharynx with anterior tooth, opening surrounded by crown of 10-12 soft papillae and ring of cilia, occasionally with trepan of small teeth. All chaetae simple, with bi-or unidentate tips, 1-12 per parapodium. Reproduction by acephalous stolon.
Remarks. A large revision of this genus was recently carried out, providing descriptions for all known species [19]; also, a complete comparative table with all valid species was recently provided by [10] Diagnosis. Species of Haplosyllis with midbody dorsal cirri alternating long (3-7 articles) and short (1-3 articles), long cirri not reaching half body width at corresponding chaetiger. Chaetae bidentate, often with eroded tips; main fang about same length or shorter than shaft width; mid-joining point straight and long; upper side of main fang with few, short denticles.
Remarks. This species was described from northern Paraíba. The distribution range is herein expanded to southern Pernambuco.
Type locality. Diagnosis. Haplosyllis with dorsal inclusions on midbody chaetigers. Chaetae with main fang shorter than shaft width; apical teeth close to each other, chaetae nearly unidentate at first glance; aciculae distally curved.
Remarks. This species was also described from Paraíba. The distribution range is herein expanded to southern Pernambuco.
Remarks. The most similar species to Haplosyllis lattigae sp. n. is H. djiboutiensis Gravier, 1900, which is recognized by the alternation in length of dorsal cirri on midbody, abundant dorsal granules on posterior body, anterior chaetae with very short main fang, with short denticles on upper side, and presence of two types of chaetae per parapodium from midbody chaetigers, short chaetae with short proximal tooth, large chaetae with mid-joining point long and straight, and apical teeth equal in size [31]. Material of H. djiboutiensis from Australia [32], Djubal (Qatar) and Abu Dhabi, including paratypes [31], was recently analysed; however, although slight morphological differences were noticed among populations, those are too slight characterize different species (Table 5).
Haplosyllis lattigae sp. n. differs from H. djiboutiensis by having two types of chaetae of different sizes also on anterior parapodia, long chaetae with apical teeth of similar size or distal tooth slightly thinner, mid-joining point straight and relatively long, main fang proportionally longer when compared with that of H. djiboutensis; and short chaetae with apical teeth forward directed, distal tooth larger than proximal one, with relatively wide, nearly rounded space between teeth, and mid-joining point straight and relatively short. Additionally, H. lattigae sp. n. differs from H. djiboutiensis from Djibuti [19] by having shorter appendages anteriorly (Table 5).
Haplosyllis lattigae sp. n. differs from H. amphimedonicola and H. rosenalessoae mainly by having two types of chaetae per parapodium along the body. Furthermore, in contraposition to H. lattigae sp. n., H. amphimedonicola has a stouter body; papillae absent on dorsum; shorter antennae and midbody cirri; chaetae with distal teeth similar in size to each other; and aciculae with different morphology. Compared to H. lattigae sp. n., H. rosenalessoae is a longer and stouter species; with shorter and stouter antennae; chaetae on mid-and posterior body with smooth main fang; distal teeth inconspicuous from midbody; and aciculae also with different morphology.
Remarks. Haplosyllis loboi has a peculiar chaetal morphology, similar to Trypanoseta sp. from Japan [34], with denticles from the base of the proximal tooth to the upper side of the main fang. The genus Trypanoseta was synonymized with Haplosyllis by [19]. The specimen from Japan differ from those from H. loboi by having up to 7 chaetae per parapodium, each with two groups of spines, the first in the mid-joining point and the second in the surface of the main fang, with one longer spine in between groups.
Brazilian specimens of H. loboi differ slightly from the specimens from the type locality (La Plata, Argentina). Although the original description states that there are 2 chaetae per parapodium throughout, we observed some chaetigers with 3 chaetae on some anteriormost parapodia in the paratype examined at the MNCN. In general, dorsal cirri from chaetiger 2 onwards are longer in Brazilian specimens; also, Brazilian specimens have a larger number of musclecells rows in proventricle than Argentinean ones (up to 51 x~36). In addition, chaetal apical teeth seem to be sharper in Argentinean specimens (Fig 4E-4H in [33]) and the denticulation between the base of apical teeth and the surface of main fang is less conspicuous in Brazilian specimens.
Biology. Haplosyllis loboi was described from coarse sandy sediments, from La Plata, Argentina. The original description suggests an association with sponges, since some specimens were found with several spicules either buried into or attached to the body surface [33]; this was also found in the specimens herein analysed, corroborating the idea of such association.
Type locality. La Plata, Argentina (Atlantic Ocean). Distribution. Atlantic Ocean: from northern Rio de Janeiro, Brazil, to La Plata, Argentina.  dorsal cirri throughout distinctly articulated; ventral cirri digitiform to ovate. Compound chaetae as falcigers only. Pharynx with opening surrounded by crown of papillae, pharyngeal tooth located away from anterior border, usually on posterior half of pharynx. Reproduction by means of stolons [32].
Remarks. The genus was first recorded for the Brazilian coast by [35], which reported O. corallicola Hartmann-Schröder, 1965 associated with the bryozoan Schizoporella unicornis (Johnston in Wood, 1844) off the state of São Paulo. Later on, [12] and [8] recorded O. brunnea from off the State of São Paulo, and [15] and [8] found O. viridis Langerhans, 1879 among material from off the states of Espírito Santo and Rio de Janeiro, respectively. The present paper is the first formal record for Opisthosyllis for the northeastern Brazilian coast.
Remarks. Brazilian specimens of O. brunnea are similar to those from the Caribbean (Venezuela and Cuba), Iberian Peninsula [2], Australia [53] and Japan, although with slight differences among specimens from different localities. Brazilian and Australian specimens have up to four aciculae per parapodium anteriorly, while Caribbean and Iberian specimens have only two. Brazilian, Caribbean and Iberian specimens have falciger blades with subdistal tooth as a spine (subbidentate), and ca. 60 rows of muscle cells in the proventricle, while Australian specimens have bidentate blades of anterior body falcigers, with distal tooth larger than subdistal one, and ca. 30 rows of muscle cells in the proventricle. These discrepancies indicate that O. brunnea, as currently recognized, may be a complex of sibling species.
Remarks. Opisthosyllis leslieharrisae and O. convexa Lee & Rho, 1994 also have the dorsal surface covered by numerous papillae. Brazilian specimens of O. viridis can be differentiated from O. leslieharrisae by having rounded papillae, shorter dorsal cirri, distinctly shorter spines on the edge of the falciger blades, and pharyngeal tooth located more posteriorly in the pharynx, close to the proventricle, while O. leslieharrisae has triangular papillae on dorsum and the pharyngeal tooth located at 3/4 pharyngeal length.
Opisthosyllis convexa has longer appendages than the Brazilian specimens of O. viridis. Furthermore, in O. convexa, the subdistal tooth of falcigers is more separated from the distal one than in O. viridis, and the pharynx occupies 5-6 chaetigers, while the proventricle extends for 10-11 chaetigers [49].
Opisthosyllis papillosa Hartmann-Schröder, 1960 and O. australis Augener, 1913, were recently recognized as junior-synonymes of O. viridis [53]. The authors [53] noticed that O. viridis, as currently considered, has high intraspecific variability, suggesting that this could be another case of a complex of sibling species. Brazilian specimens of O. viridis also present some differences from specimens from others localities, with shorter appendages than the described for Australian [53] and Japanese material; bifid dorsal simple chaetae and bidentate ventral simple chaetae; and proventricle with ca. 52 rows of muscle cells. Also different from the Brazilian material, the specimens analysed from Japan have an occipital flap. On the other hand, Australian specimens have bidentate dorsal simple chaetae and ventral simple chaetae with acute tip (Fig 27G in [53]); and proventricle with ca. 32 rows of muscle cells. Korean specimens of O. viridis have dorsal simple chaetae similar to those of the Brazilian specimens (bifid) and ventral simple chaetae similar to those of the Australian specimens (with acute tip) [49]. Papillae are absent on the prostomium and palps of the Brazilian and the Australian specimens ( Fig  27 in [53]), but present in material from Cape Verde [55]. Because of this great variability observed in the material of O. viridis from different localities, we agree with [53] in that a molecular approach is needed to clarify the diversity of species currently under this name.
Type locality. Madeira Island, Portugal (Atlantic Ocean). Distribution. Pacific Ocean: South Korea, Japan, Australia (Queensland and New South Wales). Indian Ocean: Australia (Western Australia). Atlantic Ocean: Portugal (Madeira Island) and Brazil (Pernambuco, Espírito Santo and São Paulo). First record from off the northeastern Brazilian coast.
Diagnosis. Medium to large sized syllines, up to 13 cm long, with flattened, ribbon-like body. Palps ovate, free from each other. Prostomium with 3 antennae, 2 pairs of eyes, sometimes with pair of anterior eyespots. Antennae, peristomial and dorsal cirri throughout articulated. Peristomium usually dorsally reduced, with two pairs of peristomial cirri. Compound chaetae as falcigers only, sometimes secondarily simple due to fusion of shafts and blades; dorsal and/or ventral simple chaetae sometimes present on posteriormost parapodia. Pharynx with anterior trepan; central, larger tooth sometimes also present. Reproduction by means of Tetraglene stolons [32].
Remarks. Three species of Trypanosyllis were recorded from Brazil: T. aurantiacus Nogueira & Fukuda, 2008, originally described from the state of São Paulo; T. parvidentata Perkins, 1981 found in the Guanabara Bay, state of Rio de Janeiro [59]; and T. zebra, recorded from the states of Pernambuco and Bahia (northeastern Brazilian coast) [22], and São Paulo [13,27,35,60] and Rio de Janeiro [12,61]  are also grateful to Drs Stephen Keable (AM), Javier Sánchez Almazán (MNCN), Aline Benetti (MZUSP), Geoff Keel and Karen Osborn (USNM), Hironori Komatsu (NSMT) and Angelika Brandt (ZMH), and to the staff of all these institutions, for providing access to the comparative material used for the present study.