RHINOCRYPTIDAE SYSTEMATICS

Maurício, Giovanni Nachtigall, Areta, Juan Ignacio, Bornschein, Marcos Ricardo & Reis, Roberto E., 2012, Morphology-based phylogenetic analysis and classification of the family Rhinocryptidae (Aves: Passeriformes), Zoological Journal of the Linnean Society 166 (2), pp. 377-432 : 378-381

HISTORY OF RHINOCRYPTIDAE SYSTEMATICS

Limits of the family

The large and membranous nasal operculae that hide the nares of the Rhinocryptidae and after which the family takes its name has intrigued ornithologists for a long time (e.g. Lesson, 1831 [1830]; Saint-Hilaire 1832) and has been used as a landmark for the diagnosis of the group. As a supra-generic entity, the Rhinocryptidae has its early taxonomic history rooted in contributions that appeared in the 1830s and 1840s. On the basis of the possession of the operculum covering the nostrils Ménétriés (1835) recognized the close relationship between several taxa presently allocated in four rhinocryptid genera ( Merulaxis , Eleoscytalopus , Eugralla , and Scytalopus ) and placed them in the genus Malacorhynchus , within the family ‘Myiotherinae’ (an entity equivalent to the old sense Formicariidae ). He also suggested a close relationship between his Malacorhynchus and the old sense Pteroptochos (including Scelorchilus ). Lafresnaye & d’Orbigny (1837) erected the family Rhinomyidae (i.e. the present-day Rhinocryptidae ) exclusively for the genera Rhinomya (= Rhinocrypta ) and Megalonyx (= Pteroptochos , then including Scelorchilus ) based on the possession of the diagnostic, prominent operculum covering the nostrils. When describing Merulaxis senilis (now Myornis ; Chapman, 1915) Lafresnaye (1840) used the presence of the nasal operculum to suggest a close relationship between Rhinomya , Megalonyx (= Pteroptochos ), and Merulaxis (including Malacorhynchus sensu Ménétriés, [1835] ), treating all within the family ‘Myotherinae’. Lesson’s (1841 [1842]) inclusive version of Megalonyx included several subgenera containing the actual Rhinocrypta , Pteroptochos , Scelorchilus , Eugralla , and Scytalopus , which were united among other features by the possession of a ‘naribus basalihus, membrane subtumenenli pilisque per mediam longitudinem tectâ’. Surprisingly, he did not include Merulaxis among them, a genus that he himself had previously described as having the nasal operculum ( Lesson, 1831 [1830]). A few years later, when describing Merulaxis orthonyx (now Acropternis ; Cabanis & Heine, 1859), Lafresnaye (1843) explicitly united Rhinomya , Megalonyx , and Merulaxis under the subfamily Rhinomyinae , then included in the old sense Formicariidae . Shortly afterwards, Lafresnaye (1844) justified his treatment of the Rhinomyidae at the subfamily level under Formicariidae [ sensu lato (s.l.)] and tried to demonstrate the cohesion between the three genera admitted by him in this group, reaffirming and somewhat expanding his earlier conclusions. As the three-genera treatment of Lafresnaye (1843, 1844) embraced diverse taxa representing nine out of 12 genera now included in Rhinocryptidae ( Rhinomya [= Rhinocrypta ], Megalonyx [= Pteroptochos and Scelorchilus ], and the enlarged Merulaxis [= Merulaxis , Eleoscytalopus , Eugralla , Scytalopus , Myornis and Acropternis ]), the limits and the concept of the group were already well established in the first half of the 19th century, independently of its treatment at the family or subfamily level.

Gould & Darwin (1841) also recognized the affinities between some diverse rhinocryptid genera. In addition to their inclusion of Eugralla , Scelorchilus , and Pteroptochos in an expanded genus Pteroptochos , they placed Scytalopus ( magellanicus ) just below P. paradoxus (= Eugralla paradoxa ) and regarded both genera as closely allied, remarking on their close relationship based on behaviuor. The position of Rhinomya (= Rhinocrypta ) was somewhat ambiguous: it was placed immediately above Pteroptochos and indicated to replace the Pacific Pteroptochos on the Atlantic side of the South American continent but was later said to be distantly allied to the furnariid Eremobius phoenicurus .

It is not clear how the preceding authors influenced each other in their attempt to define and classify the rhinocryptids, as there are no clear references to previous treatments or recognition of credits in these works. It appears that some sort of tacit synergy and convergent thoughts finally gave shape to the closely knitted group that we now refer to as Rhinocryptidae . It is also worthy of note that the family name Rhinomyidae was based on Rhinomya Geoffroy Saint- Hilaire, 1832, a generic name replaced by Rhinocrypta G. R. Gray, 1841 due to the fact that Rhinomya Robineau-Desvoidy, 1830 , a genus of Diptera , had nomenclatural priority ( Gray, 1841; Peters, 1951; Bock, 1994). Thus, Wetmore (1926) introduced the term Rhinocryptidae to replace Rhinomyidae , understanding that the family designation should be based on the generic name (i.e. Rhinocrypta ) erected to substitute for Rhinomya , this latter being the first name used to form a family-group name having as basis a rhinocryptid taxon. Other names sometimes used to designate the family, such as Scytalopodidae , Megalonychidae , Hylactidae, and Pteroptochidae , the last-named being used for a long period until Wetmore’s (1926) proposal, must be conditionally suppressed in favour of Rhinocryptidae ( Bock, 1994) .

The next sound contribution to the systematics and characterization of the rhinocryptids was the pioneering anatomical study of Johannes Müller ( Müller, 1878 [1847]). This study on the syrinx and osteology of Passeriformes and groups that were not considered to be closely related was first published in German in 1847 and translated into English in 1878. Based on a peculiar conformation of the syrinx, Müller (1878 [1847]) conceived and delimited the ‘Tracheophone’ division, a higher-level entity equivalent to the present-day infraorder Furnariides . He also proposed its division into three families, ‘Scytalopidae’ (= Rhinocryptidae ), Anabatidae (= Furnariidae s.l.) and Myiotheridae (= Formicariidae s.l.). Müller (1878 [1847]) highlighted that the Scytalopidae was distinguished from the remaining Tracheophone based on the presence of four notches in the posterior margin of the sternum (two on each side), a condition previously described for Pteroptochos (= Pteroptochos + Scelorchilus ) by Eyton (1841). Accordingly, he placed in his Scytalopidae only the few taxa whose sternum was known to him, namely Scytalopus indigoticus (= Eleoscytalopus indigoticus ) and the old sense Pteroptochos .

Subsequently, although some authors did not recognize the group taxonomically (e.g. Cabanis, 1847; Bonaparte, 1850), others such as Chenu & Des Murs (1852; as Megalonychidae ) and Sclater (1858a, b; as Pteroptochidae ) did so. Although Sclater (1858a) regarded the rhinocryptids as a well-distinguished group, stating that ‘their most essential character (...) consists in the covered nostril (...), which occurs in every species’, he remotely admitted the possiblity of considering them as a subfamily of Formicariidae . Shortly afterwards, however, he strongly defended that they ‘form a division rather parallel than subordinate to the family Formicariidae’ ( Sclater, 1858b), citing the results of Müller (1878 [1847]) as the main basis for this treatment. Nevertheless, some doubts regarding the limits of the family remained as he later suspected that the genera Psilorhamphus and Rhamphocaenus (a polioptilid in the Passeri radiation) ‘might perhaps be more naturally placed as a distinct subfamily of Pteroptochidae’, although he left them in their traditional positions in the Formicariidae (s.l.) ( Sclater, 1858c).

Still in the 19th century, representatives of two new genera were described: Pteroptochus thoracicus , described in the subgenus Liosceles ( Sclater, 1864) but later elevated to the rank of genus ( Sclater, 1874), and Rhinocrypta fusca ( Sclater & Salvin, 1873) , later placed in the monotypic genus Teledromas ( Wetmore & Peters, 1922) . The modern conformation of morphotypes in the family is almost the same as that in the taxonomic revision of the family by Sclater (1874), whose grouping included representatives of 11 out of 12 currently recognized genera, even though he only recognized eight genera ( Scytalopus , Merulaxis , Rhinocrypta , Liosceles , Pteroptochus [= Scelorchilus ], Hylactes [= Pteroptochos ], Acropternis , and Triptorhinus [= Eugralla ]). Since then, the genera Myornis ( Chapman, 1915) , Teledromas ( Wetmore & Peters, 1922) , and Eleoscytalopus ( Maurício et al., 2008) were described and adopted for species allocated to the Rhinocryptidae since their descriptions, and the position of the two conflictive genera Psilorhamphus and Melanopareia in the family was debated. Wetmore (1926) transferred Melanopareia from the Formicariidae s.l. to the Rhinocryptidae , as it had a fournotched sternum, a feature considered to be exclusive of the Rhinocryptidae within the Passeriformes ( Heimerdinger & Ames, 1967; Feduccia & Olson, 1982). By virtue of its peculiar external morphology, Psilorhamphus was transferred from the Formicariidae s.l. to the Sylviidae , jumping to the suborder Passeri ( Peters, 1951) , only to be placed again in the suborder Tyranni by Sick (1954; see also Sick, 1960). Sick (1954) suspected that Psilorhamphus could be a Rhinocryptidae , having even illustrated a nasal operculum covering the nares. However, it was only after the syrinx and the sternum (with four well-marked notches) were examined that Psilorhamphus was firmly allocated within the Rhinocryptidae ( Plótnick, 1958) . Despite this, the placement of Psilorhamphus and Melanopareia in the family was questioned, suggesting that their phylogenetic affinities may lay outside the Rhinocryptidae ( Ridgely & Tudor, 1994; Krabbe & Schulenberg, 2003). Recent molecular studies support the inclusion of Psilorhamphus in the Rhinocryptidae , but strongly argue against the inclusion of Melanopareia ( Irestedt et al., 2002; Moyle et al., 2009; Ericson et al., 2010). In this way, the family Rhinocryptidae today comprises 12 genera ( Remsen et al., 2011).

Interfamilial relationships

Some early taxonomists considered the rhinocryptids to be related to oscine passerines such as wrens ( Troglodytidae ) and lyrebirds and allies (Menurae) (e.g. Kittlitz, 1830; Gray, 1841; Sclater, 1855; Cabanis & Heine, 1859), but most workers linked them exclusively to suboscine passerines such as antbirds, antthrushes, and allies, at that time collectively referred to as ‘Myiotherinae’ (or Myiotheridae), a paraphyletic assemblage containing the modern Thamnophilidae , Conopophagidae , Melanopareiidae , Formicariidae , and Grallariidae (see above). The hypothesis of a close relationship between rhinocryptids and the Menurae (families Menuridae and Atrichornithidae ) was later revived by Feduccia & Olson (1982) on the basis of the discovery of osteological similarities. However, the concept of relationship they adopted includes paraphyletic arrangements (a characteristic of the evolutionary school of systematics; see Wiley, 1981) with the authors suggesting that the two groups might be ‘basally related’, i.e. they would lay close to each other at the base of the passerine radiation, but also admitted that they might be related in the sense of ‘the strict redefinition of monophyly advocated by cladists’ ( Feduccia & Olson, 1982: 17). In any event, a massive body of morphological ( Bock & Clench, 1985; Clench, 1985; Raikow, 1985, 1987; Rich, McEvey & Baird, 1985) and molecular ( Ericson et al., 2002a; Barker et al., 2004; Chesser & Have, 2007; Hackett et al., 2008) data reject these hypotheses and show that the similarities between the two groups as described by Feduccia & Olson (1982) are the product of convergent evolution.

That the Rhinocryptidae was part of a group composed exclusively by the presently recognized families Thamnophilidae , Conopophagidae , Melanopareiidae , Rhinocryptidae , Grallariidae , Formicariidae , Furnariidae , Dendrocolaptidae , and Scleruridae (i.e. the infraorder Furnariides ) has been widely accepted since Müller (1878 [1847]), who showed that members of this assemblage are characterized by the possession of a tracheal syrinx, which is more complex than that of other members of the suborder Tyranni and is unique in the class Aves ( Ames, 1971; Raikow, 1987; Krabbe & Schulenberg, 2003). Within Furnariides , the ‘ground antbirds’ [ sensu Ames (1971) , i.e. an assemblage composed by the actual families Formicariidae , Grallariidae , and Conopophagidae , excluding Thamnophilidae ], the Furnariidae (including the actual Scleruridae), and Conopophagidae were alternately considered closely related to the Rhinocryptidae based on anatomical features, especially sternum and syrinx morphology ( Garrod, 1877a; Forbes, 1881; Ames, 1971), whereas on the basis of DNA–DNA hybridization data the latter two families were found to be sisters ( Sibley & Ahlquist, 1985, 1990). Even recent molecular phylogenetic analyses with fairly comprehensive taxon sampling were not congruent in defining the sister-taxon to the Rhinocryptidae , with possible alternatives being the Grallariidae ( Moyle et al., 2009) , the Formicariidae ( Irestedt et al., 2002; Rice, 2005), or a clade formed by Scleruridae, Dendrocolaptidae , Furnariidae , and Formicariidae , with Grallariidae basal to them ( Irestedt et al., 2002; Chesser, 2004). The clade Grallariidae + Rhinocryptidae was given superfamily rank as Grallarioidea, despite its moderate statistical support, and was found to be sister to a clade including Scleruridae, Dendrocolaptidae , Furnariidae , and Formicariidae with high statistical support ( Moyle et al., 2009).

Intrafamilial relationships

Besides the simple grouping of species into distinct genera as reviewed above, a significant early attempt to organize the family Rhinocryptidae internally was that of Lafresnaye (1851). He divided his expanded genus Merulaxis ( sensu Lafresnaye, 1843, 1844 ) into four sections, the linear sequencing of which apparently being intended to reflect relationships: (1) ‘Galeati’ (rostri culmine in spatio frontali, horizontali limbo cirumdato, modern Eugralla and Acropternis ); (2) ‘Rhinolophi’ (rostri culmine basi elevato, compresso, plumulis frontalibus longiusculis erectis, modern Merulaxis sensu stricto); (3) ‘Albiventris’ (gutture, collo ântico pectoreque albis, rostro simplici, modern Eleoscytalopus ); and (4) ‘Concolores’ (aut ferê concolores, collo ântico pectoreque non albis sed dorso fere concolore, modern Scytalopus and Myornis ). Within this simple scheme, two aspects involve important systematic hypotheses. The ‘Galeati’ section groups two diverse genera that resemble each other only in the variably elevated base of the bill, an arrangement that was implemented a few years earlier by Cabanis (1847), who united Eugralla and Acropternis under his new genus Triptorhinus . The other important aspect is the recognition of the ‘Albiventris’ section for the white-bellied taxa, dissociating them from the generally grey ‘Concolores’, thus agreeing with the recently proposed genus Eleoscytalopus .

In his taxonomic revision of the family, Sclater (1874) recognized some relationships between the eight genera admitted by him in the Rhinocryptidae . He considered Scytalopus (then including Myornis and Eleoscytalopus ) and Merulaxis as allied to each other and regarded Rhinocrypta (then including Teledromas ) and Liosceles as allied to Scelorchilus (treated as Pteroptochus ), with the present-day Pteroptochos ( Hylactes at that time) being regarded as ‘a strong form’ of Scelorchilus . Although Sclater regarded Eugralla (as Triptorhinus ) as being most similar to Scytalopus , he maintained it next to Acropternis based on the similarly shaped bill.

Relationships within the Rhinocryptidae were first studied using explicit methods by Sibley & Ahlquist (1985, 1990), whose classical DNA–DNA hybridization study recovered Scytalopus + Liosceles as sister to Pteroptochos . A study on the molecular phylogenetic relationships of Pteroptochos found it to be monophyletic and sister to Scelorchilus , even though only Scytalopus and Rhinocrypta were sampled as additional members of the family ( Chesser, 1999). In a study investigating the molecular phylogenetic relationships of the Furnariides based on extensive sampling the included Rhinocryptidae formed a monophyletic clade, with Scytalopus + Pteroptochos as sister to Rhinocrypta ( Irestedt et al., 2002) . Melanopareia (considered a member of the family at that time) was highly divergent and grouped either with the Conopophagidae or with the Thamnophilidae , or was recovered as being basal to the whole infraorder ( Irestedt et al., 2002). This result prompted the creation of the new family Melanopareiidae to host Melanopareia together with Teledromas , a relationship based on reported behavioural and morphological similarities (despite Teledromas not being sampled for molecular characters) ( Irestedt et al., 2002). The same rhinocryptid genera were sampled for a wider molecular phylogenetic analysis including members of all the Tyranni , but in this case a clade Pteroptochos + Rhinocrypta was sister to Scytalopus ( Chesser, 2004) . In turn, Melanopareia was part of a polytomy together with Conopophagidae + Thamnophilidae , which was basal to the clade including Scleruridae, Dendrocolaptidae , Furnariidae , Formicariidae , Grallariidae , and Rhinocryptidae ( Chesser, 2004) . Some phylogenetic studies of other groups have also included members of the Rhinocryptidae . For example, Scytalopus was sister to a clade Rhinocrypta + Liosceles in a study on the molecular phylogenetic relationships of the Grallariidae ( Rice, 2005) , and Rhinocrypta was grouped with Pteroptochos in a study focused on the relationhips of some species of Furnariidae ( Claramunt & Rinderknecht, 2005) , in the single morphological phylogenetic study including members of the Rhinocryptidae .

Only very recently have molecular phylogenetic studies included comprehensive taxon sampling of the family. In a study aimed at investigating the relationships of some genera, Maurício et al. (2008) sampled nine genera, recovering the topology (((( Myornis ( Scytalopus + Eugralla )) ( Eleoscytalopus + Merulaxis )) ( Rhinocrypta + Psilorhamphus )) ( Scelorchilus + Pteroptochos )). In their study designed to elucidate relationships within the infraorder Furnariides, Moyle et al. (2009) also included nine genera and recovered a monophyletic Rhinocryptidae that was divided into two large clades, the subfamilies Scytalopodinae ( Myornis ( Eugralla + Scytalopus )) and Rhinocryptinae (( Pteroptochos + Scelorchilus ) ( Liosceles ( Acropternis ( Rhinocrypta + Teledromas )))). Ericson et al. (2010) also recovered a monophyletic Rhinocryptidae , in the only study to date that included all genera of the family. They found two major clades, one partially coinciding with the Rhinocryptinae of Moyle et al. (2009), with the topology (( Teledromas ( Acropternis + Rhinocrypta )) ( Liosceles + Psilorhamphus )), and another containing the Scytalopodinae of the latter authors plus the remaining genera, with the topology ((( Myornis ( Eugralla + Scytalopus )) ( Eleoscytalopus + Merulaxis )) ( Pteroptochos + Scelorchilus )).

At the species level, several studies focusing on particular groups of both Andean ( Arctander & Fjeldså, 1994; Cuervo et al., 2005; Krabbe et al., 2005; Krabbe & Cadena, 2010) and Brazilian ( Bornschein et al., 2007; Mata et al., 2009) members of the genus Scytalopus have been published, all of these including results of molecular phylogenetic analyses. Of these contributions, that of Mata et al. (2009) was the most taxonomically comprehensive and significant as it investigated the relationships of all Brazilian taxa, a diversified assemblage frequently referred to as the S. speluncae group, whose monophyly was recovered with some confidence.

Additional hypotheses on the relationships within the Rhinocryptidae have been proposed albeit without any formal phylogenetic analysis. Krabbe & Schulenberg (1997, 2003) proposed, based on vocalizations, plumage and body proportions, that Merulaxis , Myornis , Eugralla , and Scytalopus (then including Eleoscytalopus ) formed a more closely related group within the family, with the former two being hypothesized to be sister groups. Several hypotheses of relationship at the species level have been proposed on the basis of vocal and plumage characters (especially in Scytalopus ) ( Zimmer, 1939; Vielliard, 1990; Whitney, 1994; Krabbe & Schulenberg, 1997, 2003; Bornschein, Reinert & Pichorim, 1998; Maurício, 2005; Bornschein et al., 2007), some of which have been corroborated by molecular analyses.

Higher-level classification adopted in the present study

We follow the classification of Moyle et al. (2009) and Tello et al. (2009), which divide the order Passeriformes into three suborders: Acanthisitti (a New Zealand relict), Passeri (the Oscine passerines, worldwide distribution), and Tyranni (the Suboscine passerines, mostly New World but with few Old World representatives). Ericson et al. (2002a, b) and Ericson, Irestedt & Johansson (2003) also delimited these same higher-level groups, but using different names and taxonomic subordinations. Within Tyranni , we recognize three infraorders: Eurylaimides (Old World suboscines), Tyrannides , and Furnariides (New World suboscines) ( Moyle et al., 2009; Tello et al., 2009). The treatment of the Furnariides at the level of infraorder is in agreement with Cracraft (1981) and Raikow (1987), and contrasts with the traditional status of superfamily or suborder given to the group [see Sibley & Ahlquist (1990) for a revision of the topic]. The Tyrannides , sister-taxon to the Furnariides , are also given infraordinal level in the present study ( Johansson et al., 2001; Ericson et al., 2002b; Barker et al., 2004; Chesser, 2004; Irestedt et al., 2009; Tello et al., 2009). Following Moyle et al. (2009), we recognize nine family-level entities in the infraorder Furnariides , namely Thamnophilidae , Conopophagidae , Melanopareiidae , Grallariidae , Rhinocryptidae , Formicariidae , Scleruridae, Dendrocolaptidae , and Furnariidae .

Aims of the present study

This study aims to derive a phylogenetic hypotheses for the family Rhinocryptidae based on skeletal and syringeal features. Implicit in this main objective are the following empirical questions: (1) What is the potential of variation in internal morphological characters to help elucidate phylogenetic relationships in the Rhinocryptidae ? (2) Is the Rhinocryptidae a monophyletic group? (3) Do the disputed genera Psilorhamphus and Melanopareia belong to this family? (4) Are the genera Merulaxis and Eleoscytalopus sister taxa and closley related to Scytalopus , Myornis , and Eugralla as molecular phylogenetic studies proposed? (5) Is the highly speciose genus Scytalopus monophyletic?