Aneides, Baird, 1851

Frost, Darrel R., Grant, Taran, Faivovich, Julián, Bain, Raoul H., Haas, Alexander, De Sá, Célio F. B. Haddad Rafael O., Channing, Alan, Wilkinson, Mark, Donnellan, Stephen C., Raxworthy, Christopher J., Campbell, Jonathan A., Blotto, Boris L., Moler, Paul, Drewes, Robert C., Nussbaum, Ronald A., Lynch, John D., Green, David M. & Wheeler, And Ward C., 2006, The Amphibian Tree Of Life, Bulletin of the American Museum of Natural History 2006 (297), pp. 1-291 : 1-291

Aneides to the exclusion of eastern species, Of Bolitoglossinae we sampled 11 of

and others (e.g., Chippindale et al., 2004; 14 nominal genera: supergenus Batrachoseps

36 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297

pergenus Bolitoglossa ( Bolitoglossa rufes­ in addition to two cranial characters (prescens, Cryptotriton alvarezdeltoroi , Dendrotri­ ence of a frontosquamosal arch and fusion of ton rabbi , Ixalotriton niger , Lineatriton lineo­ the premaxillaries [reversed in Pleurodeles 1 lus, Nototriton abscondens , Oedipina unifor­ Tylototriton , and Chioglossa ]).

mis, Parvimolge townsendi , Pseudoeurycea Titus and Larson (1995) provided a phyconanti, and Thorius sp. ). logenetic tree on the basis of a study of mt SALAMANDRIDAE (18 GENERA, 73 SPECIES): rRNA and morphology data ( fig. 12 ). Scholz Salamandridae is found more­or­less (1995; fig. 13) obtained similar results on the throughout the Holarctic, with the bulk of its basis of morphology and courtship behavior. phylogenetic and species diversity in tem­ Zacj and Arntzen (1999) also reported on

perate Eurasia. Salamandrids are character­ phylogenetics of Triturus , showing (as did ized by strongly keratinized skin in adults Titus and Larson, 1995) that it is composed (except for the strongly aquatic Pachytriton ), of two groups: (1) Triturus vulgaris 1 Tri­

2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

37

written, but we chose taxa that should allow

the basic structure of salamandrid phylogeny

to be elucidated. To bracket this suggested

topology with appropriate taxonomic sam­

ples we chose Euproctus asper , Neurergus

crocatus, Notophthalmus viridescens , Pach­

ytriton brevipes , Paramesotriton sp. , Pleu­

rodeles waltl , Salamandra salamandra , Tar­

icha sp., Triturus cristatus , and Tylototriton

shanjing.

DICAMPTODONTIDAE (1 GENUS, 4 SPECIES):

The North American Dicamptodon is related

to Ambystomatidae (Larson and Dimmick,

1993; fig. 4 ) and, like them, some popula­

tions are neotenic (Nussbaum, 1976). Like

other salamandroid salamanders they have

internal fertilization and a suite of morpho­ Fig. 13. Consensus of salamandrid relation­ logical features associated with forming and ships suggested by Scholz (1995) based on a par­ collecting spermatophores. Dicamptodon difsimony analysis of 27 character transformations fers from Ambystomatidae in glandular feaof morphology and behavior. Triturus in Scholz’s tures of the cloaca and in attaining a large sense included what is now Lissotriton , Mesotri­

size, but is considered by most workers as ton, and Triturus (García­París et al., 2004b) .

the sister taxon of Ambystomatidae (e.g.,

Larson et al., 2003—fig. 6; Wiens et al., turus marmoratus species groups; (2) and 2005— fig. 7 ). We sampled both Dicampto­ Triturus cristatus group, but not addressing don aterrimus and D. tenebrosus . its polyphyly. Steinfartz et al. (2002) report­ AMBYSTOMATIDAE (1 GENUS, 31 SPECIES): ed on salamandrid phylogeny and substanti­ North American Ambystomatidae is a morated the polyphyly of Triturus and of Mer­ phologically compact family having internal tensiella. Subsequently (and appearing after fertilization via a spermatophore and the this analysis was completed), García­París et suite of morphological characters that supal. (2004b) partitioned the polyphyletic ‘‘ Tri­ port this attribute. Some populations exhibit turus ’’ into three genera ( Triturus , Lissotri­ neotenic aquatic adults. ton, and Mesotriton ), based on the sugges­ The last summary of phylogeny within the tions that (1) Triturus , sensu stricto ( Triturus group based on explicit evidence was precristatus 1 T. marmoratus species groups) is sented by Shaffer et al. (1991; see also Larmost closely related to Euproctus ; (2) Me­ son et al., 2003), who provided a cladogram sotriton ( Triturus alpestris ) is the sister taxon based on 32 morphological transformation of a group composed of Cynops , Parame­ series and 26 allozymic transformation sesotriton, and Pachytriton ); and (3) Lissotri­ ries. The basal dichotomy in this tree is beton ( Triturus vulgaris species group) is of tween Ambystoma gracile 1 A. maculatum uncertain relationship to the other compo­ 1 A. talpoideum on one hand, and all other nents, but does not form a monophyletic species of Ambystoma , on the other. We were group with either Mesotriton or Triturus . unable to obtain any of these three species, García­París et al. (2004a: 602) also sug­ but we did sample Ambystoma cingulatum , gested that ongoing molecular work (evi­ A. mexicanum and A. tigrinum . Ambystoma dence undisclosed), will show Euproctus to mexicanum and A. tigrinum are very closely be paraphyletic and that Triturus vittatus will related, and A. cingulatum is distantly related not be included within Triturus , the oldest to them. This is a weaker test of monophyly available name for this taxon being Omma­ than we would have liked because it does not

totriton Gray, 1850. include A. gracile , A. maculatum , or A. tal­ We could not address these final issues, poideum. Further, Larson et al. (2003) sugthese appearing well after the manuscript was gested that, in addition to A. gracile , A. ma­

38 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297

culatum, and A. talpoideum , A. jeffersonian­ the extent of character conflict within their um, A. laterale , A. macrodactylum , and A. data set was never adequately exposed. More opacum were likely to be outside of the taxa recently, Haas (2003; fig. 15) provided a disbracketed by our species, although the evi­ cussion of frog evolution, based primarily on dence for this was not presented. new larval characters. Haas did, however, exclude several of the adult characters included ANURA by Ford and Cannatella (1993) as insufficiently characterized or assayed. More re­ Frogs (32 families, ca. 372 genera, 5227 cently, important discussions of phylogeny species) constitute the vast majority (88%) of have been made in the context of DNA seliving species of amphibians and the bulk of quence studies (Roelants and Bossuyt, their genetic, physiological, ecological, and 2005— fig. 16 ; San Mauro et al., 2005—fig. morphological diversity. Despite numerous 17) that will be cited throughout our review. studies that point towards its deficiencies The monophyly of frogs ( Anura ) relative (e.g. Kluge and Farris, 1969; Lynch , 1973; to other living amphibians has not been gen­ Sokol, 1975, 1977; Duellman and Trueb, erally questioned 6 (although the universality 1986; Ruvinsky and Maxson, 1996; Maglia , of this taxon with respect to some fossil an­ 1998; Emerson et al., 2000; Maglia et al., tecedent taxa has (e.g., Griffiths, 1963; Ro­ 2001; Scheltinga et al., 2002; Haas, 2003; ĉek, 1989, 1990), and the number of mor­ Roelants and Bossuyt, 2005; San Mauro et phological characters corroborating this al., 2005; Van der Meijden et al., 2005), the monophyly is large—e.g., (1) reduction of current classification continues in many of its vertebrae to 9 or fewer; (2) atlas with a single parts to reflect sociological conservatism and centrum; (3) hindlimbs significantly longer the traditional preoccupation with groupings than forelimbs, including elongation of ankle by subjective impressions of overall similar­ bones; (4) fusions of radius and ulna and tibity; special pleading for characters consid­ ia and fibula; (5) fusion of caudal vertebral ered to be of transcendent importance; and segments into a urostyle; (6) fusion of hyobnotions of ‘‘primitive’’, ‘‘transitional’’, and ranchial elements into a hyoid plate; (7) pres­ ‘‘advanced’’ groups instead of evolutionary ence of keratinous jaw sheaths and keratopropinquity. Understanding of frog relation­ donts on larval mouthparts; (8) a single meships remains largely a tapestry of conflicting dian spiracle in the larva, a characteristic of opinion, isolated lines of evidence, unsub­ the Type III tadpole (consideration of this as stantiated assertion, and unresolved paraphy­ a synapomorphy being highly contingent on ly and polyphyly. Indeed, the current taxon­ the preferred overall cladogram); (9) skin omy of frogs is based on a relatively small with large subcutaneous lymph spaces; and sampling of species and in many cases the (10) two m. protractor lentis attached to lens, putative morphological characteristics of ma­ based on very narrow taxon sampling (Saintjor clades within Anura are overly­general­ Aubain, 1981; Ford and Cannatella, 1993).

ized, overly­interpreted, and reified through Haas (2003) suggested ( fig. 15) an addigenerations of literature reviews (e.g., Ford tional 20 synapomorphies from larval mor­ and Cannatella, 1993), of which this review phology: (1) paired venae caudalis lateralis is presumably guilty as well. This general short; (2) operculum fused to abdominal lack of detailed understanding of anuran re­ wall; (3) m. geniohyoideus origin from cerlationships has been exacerbated by the ex­ atobranchials I–II; (4) m. interhyoideus posplosive discovery of new species in the past terior absent; (5) larval jaw depressors orig­ 20 years. inate from palatoquadrate; (6) ramus maxil­ Currently, the most widely cited review of laris (cranial nerve V 2) medial to the m. lefrog phylogeny is Ford and Cannatella

(1993; fig. 14 ), which provided a narrative 6 Roček and Vesely (1989) suggested a diphyletic ordiscussion of the evidence for a novel view igin of Anura based on a hypothesized nonhomology

of frog phylogeny without providing all of between the rostral plate of pipoid larvae and the cornua trabeculae of other anuran larvae. The developmental

2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

39 vator manidbulae longus; (7) ramus hypobranchiale; (12) processus urobranchialmandibularis (cranial nerve V 3) anterior (dor­ is short, not reaching beyond the hypobransal) to the m. levator mandibulae longus; (8) chial plates; (13) commisura proximalis I ramus mandibularis (cranial nerve V 3) ante­ present; (14) commisura proximalis II prerior (dorsal) to the externus group; (9) car­ sent; (15) commisura proximalis III present;)

tilago labialis superior (suprarostral cartilage) (16) ceratohyal with diarthrotic articulation present; (10) two perilymphatic foramina; present, medial part broad; (17) cleft between (11) hypobrachial skeletal parts as planum hyal arch and branchial arch I closed; (18 40 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297 2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

41

ligamentum cornuquadratum present; (19) sacral vertebrae (Ford and Cannatella, 1993), ventral valvular velum present; (20) branchi­ both considered plesiomorphic within Anual food traps present. Haas also suggested ra 7. Ascaphus has an intromittant organ (apothat the following were synapomorphies not morphic) in males and a highly modified tormentioned as such by Ford and Cannatella rent­dwelling tadpole. The vertebrae are am­ (1993): (1) amplexus inguinal; (2) vertical phicoelous and ectochordal (Nicholls, 1916; pupil shape; (3) clavicle overlapping scapula Laurent, 1986), presumably plesiomorphic at anteriorly; and (4) cricoid cartilage as a this level of generality. Our sampled species closed ring. for this taxon is Ascaphus truei , one of the two closely­related species. ‘‘ PRIMITIVE ’’ FROGS LEIOPELMATIDAE (1 GENUS, 4 SPECIES): Isolated in New Zealand, Leiopelmatidae , like We first address the groups that are some­ Ascaphidae , is a generally very plesiomorphtimes referred to collectively as Archaeoba­ ic group of frogs. Nevertheless, it possesses trachia (Duellman, 1975) and traditionally are considered ‘‘primitive’’, even though the apomorphies, such as ventral inscription ribs, component taxa have their own apomorphies found nowhere else among frogs (Noble, and the preponderance of evidence suggests 1931; Laurent, 1986; Ford and Cannatella, strongly that they do not form a monophy­ 1993). Unlike Ascaphus , Leiopelma does not letic group (Roelants and Bossuyt, 2005; San have feeding larvae (Archey, 1922; Altig and Mauro et al., 2005). McDiarmid, 1999; Bell and Wassersug, ASCAPHIDAE (1 GENUS, 2 SPECIES): Ford and 2003). As in Ascaphidae , the vertebrae are Cannatella (1993) considered North Ameri­ amphicoelous and ectochordal with a persiscan Ascaphus ( Ascaphidae ) to be the sister tent notochord (Noble, 1924; Ritland, 1955) taxon of all other frogs ( fig. 14 ), although on and both vocal sacs and vocalization are abthe basis of allozyme study by Green et al. sent (Noble and Putnam, 1931). (1989) and, more recently, Roelants et al. Ford and Cannatella (1993) suggested that (2005; fig. 16 ) and San Mauro et al. (2005; Leiopelmatidae is the nearest relative of all fig. 17 ), on the basis of evidence from DNA other frogs (excluding Ascaphidae ) and listed sequences, suggested that Ascaphidae 1 five synapomorphies in support of this Leiopelmatidae forms a monophyletic group. grouping (their Leiopelmatanura): (1) elon­ Báez and Basso (1996) presented a phylo­ gate arms on the sternum; (2) loss of the asgenetic analysis designed to explore the re­ cending process of the palatoquadrate; (3) lationships of the fossil anurans Vieraella sphenethmoid ossifying in the anterior posi­ and Notobatrachus with the extant taxa As­ tion; (4) exit of the root of the facial nerve caphus, Leiopelma , Bombina , Alytes , and from the braincase through the facial fora­ Discoglossus . Despite their restricted taxon men, anterior to the auditory capsule, rather sampling, their results also support the than via the anterior acoustic foramen into monophyly of Ascaphus 1 Leiopelma , al­ the auditory capsule; (5) palatoquadrate arthough the authors considered their evidence weak for reasons of difficulty in evaluating 7 Ritland (1955) suggested the possibility that the m. characters. caudalipuboischiotibialis is not homologous with the Green and Cannatella (1993) did not find tail­wagging muscles of salamanders but instead, an ac­ a monophyletic Ascaphus 1 Leiopelma . As­ cessory coccygeal head of the m. semimembranosus. In that case, the character would be judged to be a synacaphus and Leiopelma share the presence of pomorphy of Ascaphus 1 Leiopelma , rather than a sym­ a m. caudalipuboischiotibialis and nine pre­ plesiomorphy shared by those taxa.

trees discovered by parsimony analysis of 151 character­transformation series (excluding his morphometric characters 12, 83, 116, and 117, as well as 102) of larval and adult morphology and reproductive mode (ci 5 0.31; ri 0.77). Taxonomy is updated to reflect subsequent publications.

42 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297 2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

43 ticulates with the braincase via a pseudobasal respect to salamanders; the other three charthe process rather than a basal process. acters were likely polarized on the assump­ Characters 4 (facial nerve exit) and 5 (pal­ tion that Ascaphus is plesiomorphic and atoquadrate articulation) are polarized with sister taxon of remaining frogs, thereby pre­

44 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297

supposing the results, although this was not caudalipuboischiotibialis. In addition, stated. With respect to character 1 (the tri­ Abourachid and Green (1999) noted that alradiate sternum), the parsimony cost of this though Leiopelma and Ascaphus do hop, transformation on the overall tree is identical they swim with alternating sweeps of the if Ascaphidae and Leiopelmatidae are sister hind legs (the presumably plesiomorphic taxa and Bombinatoridae and Discoglossidae condition), unlike those in Bombianura, are sister taxa. The remaining characters, 2 which swim with coordinated thrusts of the and 3, were not discussed with respect to out­ hind limbs, a likely synapomorphy. groups or reversals in the remainder of Ford Bombinatoridae was considered (Ford and and Cannatella’s tree, implying that they are Cannatella, 1993) to have as synapomorphies unreversed and unique. (1) expanded flange of the quadratojugal, and With Ascaphus , Leiopelma shares the apo­ (2) presence of endochondral ossifications in morphy of columella not present (N.G. Ste­ the hyoid plate (both unreversed). We samphenson, 1951). Haas (2003) did not include pled four species of Bombina : B. bombina , Leiopelma in his analysis of exotrophic lar­ B. microdeladigitora , B. orientalis , and B. val morphology because of their endotrophy. variegata . The genus may be monophyletic, We included in our analysis Leiopelma ar­ but no rigorous phylogenetic study has been cheyi and L. hochstetteri , which bracket the performed so far, and paraphyly of Bombina phylogenetic diversity of Leiopelmatidae with respect to Barbourula remains an open (E.M. Stephenson et al., 1974), although it is question. We could not obtain tissues of Barnot sufficient to test hypotheses of the evo­ bourula so its phylogenetic position will relution of direct development (exoviviparity main questionable. Bombina has aquatic in this case; Thibaudeau and Altig, 1999) feeding tadpoles, but larvae of Barbourula within Leiopelma . are unknown and are suspected to be endo­ DISCOGLOSSIDAE 8 (2 GENERA, 12 SPECIES) trophic (Altig and McDiarmid, 1999). Dis­ AND BOMBINATORIDAE (2 GENERA, 10 SPECIES): coglossidae (sensu stricto) also has free­liv­ Ford and Cannatella (1993; fig. 14 ) suggest­ ing aquatic tadpoles (Boulenger, 1892 ed that Bombina 1 Barbourula forms the sis­ ‘‘1891’’; Altig and McDiarmid, 1999). ter taxon of all other frogs, exclusive of Leio­ Ford and Cannatella (1993; fig. 14 ) also pelmatidae and Ascaphidae , although recent posited a taxon, Discoglossanura, composed molecular evidence (Roelants and Bossuyt, of Discoglossidae (sensu stricto) and the re­ 2005; fig. 16 ) placed Bombinatoridae and maining frogs (exclusive of Ascaphidae , Discoglossidae in the familiar position of sis­ Leiopelmatidae , and Bombinatoridae ) which ter taxa. they suggested to be monophyletic on the ba­ Ford and Cannatella’s (1993) arrangement sis of two synapomorphies: (1) bicondylar ( fig. 14 ; i.e., paraphyly of Bombinatoridae 1 sacrococcygeal articulation; and (2) epister­ Discoglossidae ) required a partition of the num present. Monophyly of Discoglossidae traditionally recognized Discoglossidae (sen­ (sensu stricto) was supported by their possu lato) to place Bombina and Barbourula in session of (1) V­shaped parahyoid bones their own family, Bombinatoridae . In their (also in Pelodytes ) and (2) a narrow epipubic system, Bombinatoridae 1 its sister taxon cartilage plate. (all frogs excluding Leiopelmatidae and As­ Haas (2003; fig. 15) presented a cladogram caphidae) was named Bombianura. Bombi­ that is both deeply at variance with the reanura is corroborated by four synapomor­ lationships suggested by Ford and Cannatella phies: (1) fusion of the halves of the sphe­ (1993) and, at least with respect to this part nethmoid; (2) reduction to eight presacral of their cladogram, consistent with the movertebrae; (3) loss of the m. epipubicus (re­ lecular evidence presented by Roelants and gained in Xenopus ); and (4) loss of the m. Bossuyt (2005; fig. 16 ). Haas (2003) pre­

sented six morphological synapomorphies of

Discoglossidae 1 Bombinatoridae (as Dis­

8 Sanchíz (1998) and Dubois (2005) noted that the

coglossidae, sensu lato) and rejected Discog­2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

45

of the remaining members of Discoglossidae Ford and Cannatella (1993; fig. 14 ) sug­ 1 Bombinatoridae . Synapomorphies of gested this group, Mesobatrachia, to be Haas’ Discoglossidae are: (1) origin of m. monophyletic and composed of Pipoidea intermandibularis restricted to the medial ( Pipidae 1 Rhinophrynidae ) and Pelobatoface of the cartilago meckelii; (2) larval m. idea ( Pelobatidae [including Scaphiopodidae ] levator mandibulae externus present as two 1 Megophryidae 1 Pelodytidae ). They probundles (profundus and superficialis); (3) vided four synapomorphies for their Mesoposterior processes of pars alaris double; (4) batrachia: (1) closure of the frontoparietal cartilaginous roofing of the cavum cranii fontanelle by juxtaposition of the frontopapresent only as taenia traversalis; (5) verte­ rietal bones (not in Pelodytes or Spea ); (2) bral centra formation epichordal; and (6) pro­ partial closure of the hyoglossal sinus by the cessus urobranchialis absent. Synapomor­ ceratohyals; (3) absence of the taenia tecti phies suggested by Haas (2003; fig. 15) for medialis; and (4) absence of the taenia tecti Discoglossidae , excluding Alytes are (1) epi­ transversum.

dermal melanocytes forming an orthogonal Pugener et al. (2003) rejected Mesobatrapattern; (2) advertisement call inspiratory; chia and suggested three synapomorphies for and (3) pupil an inverted drop­shape (trian­ a clade composed of all frogs excluding pigular). Of Discoglossidae (sensu stricto), we poids. (This statement is based on Pugener sampled one species of Alytes (A. obstetri­ et al.’s, 2003, figure 12 ; they provided no cans) and two species of Discoglossus (D. comprehensive list of synapomorphies.) galganoi and D. pictus ). Discoglossidae and Haas (2003; fig. 15), in contrast, suggested Bombinatoridae show opisthocoelous and a number of characters that placed Pipoidea epichordal vertebrae according to Mookerjee as the sister taxon of all frogs except Asca­ (1931), Griffiths (1963), and Haas (2003). phidae (although he did not study Leiopel­ Kluge and Farris (1969: 23) suggested that ma). This is consistent with the molecular vertebral development in Discoglossus pictus studies of San Mauro et al. (2005; fig. 17 ). is perichordal, although Haas (2003) reported Haas’ characters also placed Pelobatoidea (as it as epichordal. represented by his exemplars) as a paraphy­ Roelants and Bossuyt (2005; fig. 16 ) and, letic series of Spea , ( Pelodytes, Heleophrywith denser taxon sampling, San Mauro et ne), and Pelobates 1 Megophrys 1 Leptoal. (2005; fig. 17 ) provided substantial brachium, ‘‘between’’ Discoglossidae (sensu amounts of DNA evidence suggesting lato) and Limnodynastes on a pectinate tree. strongly that Bombinatoridae 1 Discoglos­ This is inconsistent with the results of Roesidae forms a monophyletic group, thereby lants and Bossuyt (2005). Larval characters rejecting Discoglossanura, Leiopelmatanura, suggested by Haas (2003) to support the and Bombianura of Ford and Cannatella group of all frogs exclusive of Ascaphidae (1993). and Pipoidea are (1) m. mandibulolabialis present; (2) upper jaw cartilages powered by ‘‘ TRANSITIONAL ’’ FROGS jaw muscles; (3) larval m. levator mandibulae externus main portion inserts in upper The following few groups traditionally jaw cartilages; (4) insertion of the larval m. have been considered ‘‘transitional’’ from the levator mandibulae internus in relation to jaw primitive to advanced frogs, even though one articulation lateral; (5) m. levator mandibulae component taxon in particular, Pipidae , is longus superficialis and profundus in two highly apomorphic in several ways. The bundles; (6) processus anterolateralis of crismonophyly of this collection of families was ta parotica present; (7) processus muscularis supported by some authors (e.g., Ford and present; (8) distal end of cartilago meckeli Cannatella, 1993; García­París et al., 2003), with stout dorsal and ventral processes formbut recent morphological (e.g., Haas, 2003; ing a shallow articular fossa; and (9) liga­ Pugener et al., 2003) and DNA sequence ev­ mentum mandibulosuprarostrale present.

the idence (Roelants and Bossuyt, 2005; San García­París et al. (2004b; fig. 18 ) pre­ Mauro et al., 2005) does not support its sented mtDNA sequence evidence for monophyly. monophyly of Mesobatrachia although their

46 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297

outgroup sampling (which was limited to As­ changing places, with San Mauro et al.’s caphus truei , A. montanus , Discoglossus gal­ (2005) placement of Pipoidea agreeing with ganoi, and Rana iberica ) provided only a that of Haas (2003).

minimal test of this proposition. Even more PIPOIDEA: Pipoidea ( Pipidae 1 Rhinorecently, on the basis of more DNA sequence phrynidae) is clearly well corroborated as evidence and better sampling, Roelants and monophyletic but not clearly resolved with Bossuyt (2005; fig. 16 ) and San Mauro et al. respect to its rather dense fossil record. Ford (2005; fig. 17 ) found ‘‘Mesobatrachia’’ to and Cannatella (1993; fig. 14 ) considered Pihave its elements in a paraphyletic series poidea to be supported by five morphological with respect to Neobatrachia. Roelants and synapomorphies: (1) lack of mentomeckelian Bossuyt (2005) found ( Ascaphidae 1 Leio­ bones; (2) absence of lateral alae of the parpelmatidae) 1 (Discoglossoidea 1 (Pipoidea asphenoid; (3) fusion of the frontoparietals 1 (Pelobatoidea 1 Neobatrachia))) and San into an azygous element; (4) greatly enlarged Mauro et al. (2005) found Ascaphidae 1 otic capsule; and (5) tadpole with paired spi­ Leiopelmatidae as the sister taxon of Pipo­ racles and lacking keratinized jaw sheaths idea 1 (Discoglossoidea 1 ( Pelobatidae 1 and keratodonts (Type I tadpole). Haas

Neobatrachia)). In other words, their substan­ (2003) added a substantial number of larval tial difference is in Discoglossoidea (5 Bom­ characters: (1) eye position lateral; (2) operbinatoridae 1 Discoglossidae ) and Pipoidea cular canal and spiracles paired; (3) insertion

2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

47

of m. levator arcuum branchialium reaching phy of Rhinophrynidae at this level of genmedially and extending on proximal parts of erality.

ceratobranchial IV; (4) m. constrictor bran­ PIPIDAE (5 GENERA, 30 SPECIES): South chialis I absent; (5) m. levator mandibulae American and African Pipidae is a highly internus shifted anteriorly; (6) m. levator apomorphic group of bizarre, highly aquatic mandibulae longus originates exclusively species. Ford and Cannatella (1993) provided from arcus subocularis; (7) posterolateral 11 characters in support of its monophyly: projections of the crista parotica with expan­ (1) lack of a quadratojugal; (2) presence of sive flat chondrifications; (8) arcus subocu­ an epipubic cartilage; (3) unpaired epipubic laris with a distinct processus lateralis pos­ muscle; (4) free ribs in larvae; (5) fused arterior projecting laterally from the posterior ticulation between the coccyx and the sapalatoquadrate; (9) articulation of cartilago crum; (6) short, stocky scapula; (7) elongate labialis superior with cornu trabeculae fused septomaxillary bones; (8) ossified pubis; (9) into rostral plate; and (10) forelimb erupts a single median palatal opening of the eusout of limb pouch, outside of peribranchial tachian tube; (10) lateral line organs in the space. In addition, recent DNA sequence adults; and (11) loss of tongue. Báez and data (Roelants and Bossuyt, 2005; fig. 16 ) Trueb (1997) added to this list (fossil taxa strongly support a monophyletic group of pruned by us for purposes of this discussion): Rhinophrynidae 1 Pipidae . (1) the possession of an optic foramen with R(1, 1): a complete bony margin formed by the sphe­ HINOPHRYNIDAE GENUS SPECIES

Tropical North American and Central Amer­ nethmoid; (2) anterior ramus of the pterygoid ican Rhinophrynus dorsalis is a burrowing arises near the anteromedial corner of the frog with a number of apomorphies with re­

otic capsule; (3) parasphenoid fused at least spect to its nearest living relative, Pipidae :

partially with the overlying braincase; (4) vomer without an anterior process if the bone (1) division of the distal condyle of the femur

is present; (5) mandible bears a broad­based, into lateral and medial condyles; (2) modi­

bladelike coronoid process along its posterofication of the prehallux and distal phalanx

medial margin; (6) sternal end of the coraof the first digit into a spade for digging; (3)

coid not widely expanded; (7) anterior ramus tibiale and fibulare short and stocky, with

of pterygoid dorsal with respect to the maxdistal ends fused; and (4) an elongate atlantal

illa; and (8) premaxillary alary processes exneural arch. In addition to the previous char­

panded dorsolaterally. Haas (2003) provided acters provided by Ford and Cannatella

11 additional larval characters: (1) origin of (1993; fig. 14 ), Haas (2003; fig. 15) provided

the m. subarcualis rectus II–IV placed far lat­ (1) larval m. geniohyoideus absent; (2) larval erally; (2) anterior insertion of m. subarcualis m. levator mandibulae externus present in

rectus II–IV on ceratohyal III; (3) commistwo bundles (profundus and superficialis); surae craniobranchiales present; (4) arcus su­ (3) ramus mandibularis (cranial nerve V 3) bocularis round in cross section; (5) one periposterior (ventral) to m. levator mandibulae lymphatic foramen; (6) vertebral centra forexternus group; (4) endolymphatic spaces mation epichordal; (7) processus urobranextend into more than half of the vertebral chialis absent; and (8) ventral valvular velum canal (presacral vertebrae 4 or beyond); (5) absent, as well as these additional characters branchial food traps divided crescentrically; of the adult: (9) advertisement call without (6) cricoid ring with dorsal gap; and (7) uro­ airflow; (10) pupil shape round; and (11) branchial process very long. Available DNA pectoral girdle pseudofirmisternal.

sequence data (e.g., Roelants and Bossuyt, On the basis of morphology, Cannatella 2005) also suggest strongly that Rhinophry­ and Trueb (1988; fig. 19A) considered the nus is the sister taxon of Pipidae . We sam­ generic relationships to be Xenopus 1 (Silpled the single species in this taxon, Rhino­ urana 1 (( Hymenochirus 1 Pseudhymenophrynus dorsalis . Báez and Trueb (1997) not­ chirus) 1 Pipa )). Subsequently, de Sá and, ed that Rhinophrynus also has amphicoelous Hillis (1990; fig. 19B), on the basis of a comectochordal vertebrae, as in Ascaphidae and bined analysis of morphology and mtDNA Leiopelmatidae , which may be a synapomor­ proposed the arrangement Hymenochirus

48 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297

postzygopophyses bear sulci and ridges, with

the prezygopophyses covering the lateral

margin of the postzygopophysis; and (4) an­

terior process of the pterygoid laminae. They

also suggested the following synapomorphies

for Pipinae ( Pipa 1 Hymenochirus ) (fossil

taxa pruned for purposes of this discussion):

(1) wedge­shaped skull; (2) vertebrae with

parasagittal spinous processes; (3) anterior

position of the posterior margin of the par­

asphenoid; (4) possession of short coracoids

broadly expanded at their sternal ends. In ad­

dition, they noted other characters of more

ambiguous placement that optimize on this

stem in this topology. Recent DNA sequence Fig. 19. Trees of intergeneric relationships data (Roelants and Bossuyt, 2005; figs. 16 , within Pipidae : A, Analysis of Cannatella and 19D), however, suggest a topology of Pipa Trueb (1988) based on 94 character transforma­ 1 ( Hymenochirus 1 ( Xenopus 1 Silurana )). tions of morphology and 7 ingroup taxa (4 species We sampled three species of Dactylethriof Pipa collapsed and Pseudhymenochirus consid­ nae (Africa): Silurana tropicalis , Xenopus ered a synonym of Hymenochirus in our figure for laevis , and X. gilli . From Pipinae (South clarity of discussion). Monophyly of Pipidae was America and Africa) we sampled Hymenoassumed as well as the sister­taxon relationship of chirus boettgeri , Pipa pipa , and P. carvalhoi . Rhinophrynidae , with pelobatoids accepted as the According to the cladogram provided by second taxonomic outgroup (no tree statistics provided). B, Analysis of de Sá and Hillis (1990) Trueb and Cannatella (1986), inclusion of eibased on 1.486kb of sequence from nuclear 18S ther Pipa parva or P. myersi would have and 28S rDNA and the morphological data from bracketed the phylogenetic diversity of Pipa Cannatella and Trueb (1988) . Sequences were somewhat better, although our sampling was aligned manually and analyzed under equally adequate to test pipine (weakly), dactylethweighted parsimony; gaps were not treated as ev­ rine, and pipid monophyly, and the placeidence. The tree was rooted on Spea (tree length ment of Pipidae among other frogs. counting only informative characters 5 81, ci 5 PELOBATOIDEA: Pelobatoidea (Megophryi­ 0.74). C, Parsimony tree of Báez and Pugener (2003) based on 49 characters of adult morphol­ dae, Pelobatidae , Pelodytidae , and Scaphioogy, outgroups and fossils pruned for graphic pur­ podidae) has also been the source of considposes (the effect of this pruning on the number of erable controversy. Haas (2003; fig. 15) did characters being relevant is not known). The tree not recover the group as monophyletic (see was rooted on Rhinophrynus , Discoglossus , and the earlier discussion under Mesobatrachia), Ascaphus . (The length of original tree 5 93, ci 5 although Ford and Cannatella (1993) sug­ 0.677.) D, Relevant section of tree from Roelants gested that synapomorphies include the pres­ and Bossuyt (2005). See figure 16 for information ence of a palatine process of the maxilla and on alignment and analysis. ossification of the sternum into a bony style.

Gao and Wang (2001) found Pelobatoidea to ( Xenopus 1 Silurana ), and this was further be more closely related to Discoglossidae on corroborated by Báez and Trueb (1997) and the basis of a limited analysis of fossil taxa. Báez and Pugener (2003; who found [Hy­ But, García­París et al. (2003; fig. 18 ) sugmenochirus 1 Pipa ] 1 [ Xenopus 1 Silur­ gested that Pelobatoidea is the sister taxon of ana ]; fig. 19C), and suggested the following Pipoidea on the basis of a maximum­likelisynapomorphies for Dactylethrinae ( Xenopus hood analysis of mtDNA evidence, although 1 Silurana ; fossil taxa pruned for this dis­ their outgroup structure was insufficient to

cussion): (1) scapula extremely reduced; (2) provide a strong test of this proposition. margins of olfactory foramina cartilaginous; (This position was effectively rejected by re­ (3) articular surfaces of the vertebral pre­ and cent molecular evidence [Roelants and Bos­

2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

49

suyt, 2005; San Mauro et al., 2005; figs. 16 , calcaneum (also found in some Centroleni­ 17].) dae; Sanchíz and de la Riva, 1993) and

Maglia (1998) also provided an analysis of placed them in their Pelobatoidea as did Pelobatoidea, but because she constrained García­París et al. (2003; fig. 18 ). Haas the monophyly of this group we are not sure (2003), however, recovered Pelodytes in a how to interpret the distribution of her mor­ polytomy with Heleophryne, Neobatrachia phological evidence. Pugener et al. (2003) and Megophrys 1 Pelobates 1 Leptobrachprovided a cladogram based on morphology ium. We sampled Pelodytes punctatus as our in which Pelobatoidea was recovered as exemplar of Pelodytidae . Larvae in pelodymonophyletic (and imbedded within Neoba­ tids are also typical free­living exotrophs trachia), but the underlying data were not (Altig and McDiarmid, 1999). provided. Roelants and Bossuyt (2005; fig. MEGOPHRYIDAE (11 GENERA, 129 SPECIES): 16) suggested on the basis of DNA evidence Ford and Cannatella (1993; fig. 14 ) diagthat Pelobatoidea is the sister taxon of Neo­ nosed Megophryidae as having (1) a combatrachia, a result that is consistent with the plete or nearly complete absence of ceratoolder view of Savage (1973; cf. Noble, hyals in adults; (2) intervertebral cartilages 1931). Dubois (2005) most recently treated with an ossified center; and (3) paddleall pelobatoids as a single family composed shaped tongue. Haas (2003; fig. 15) recovof four subfamilies, but this was merely a ered a group consisting of the megophryids change in Linnaean rank without a concom­ ( Leptobrachium and Megophrys being his itant change in understanding phylogenetic exemplars) and Pelobates but did not resolve history. the megophryids sensu stricto. Evidence for

PELOBATIDAE (1 GENUS, 4 SPECIES) AND this megophryid 1 Pelobates clade is: (1) SCAPHIOPODIDAE (2 GENERA, 7 SPECIES): Ford distal anterior labial ridge and keratodont­ and Cannatella (1993; fig. 14 ) diagnosed Pe­ bearing row very short and median; (2) vena lobatidae (including Scaphiopodidae in their caudalis dorsalis present; (3) anterior insersense) on the basis of (1) fusion of the joint tion of the m. subarcualis rectus II–IV on between the sacrum and urostyle; (2) exos­ ceratobranchial III; (4) m. mandibulolabialis tosed frontoparietals; and (3) presence of a superior present; (5) adrostral cartilage very metatarsal spade supported by a well­ossified large and elongate; and (6) cricoid ring with prehallux. As noted earlier, Haas (2003; fig. a dorsal gap. 15) did not recover Pelobatidae (sensu lato) Dubois (1980) and Dubois and Ohler as monophyletic, instead placing Spea phy­ (1998) suggested that megophryids form two logenetically far from Pelobatidae , more dis­ subfamilies based on whether the larvae have tant than Heleophryne . More recently, funnel­shaped oral discs (Megophryinae), an García­París et al. (2003; fig. 18 ) provided apomorphy, or nonmodified oral discs (Lepmolecular data suggesting that Pelobatidae tobrachiinae), a plesiomorphy. Megophryi­ and Scaphiopodidae are not each other’s nae includes Atympanophrys, Brachytarsoclosest relatives. These results were aug­ phrys , Megophrys , Ophryophryne , and Xenmented by the DNA sequence studies of ophrys. Their Leptobrachiinae includes Lep­ Roelants and Bossuyt (2005) and San Mauro tobrachella, Leptolalax , Leptobrachium , et al. (2005), both of which supported Sca­ Oreolalax , Scutiger , and Vibrissaphora. Dephiopodidae as the sister taxon of Pelodyti­ lorme and Dubois (2001) presented a condae 1 ( Pelobatidae 1 Megophryidae ) (figs. sensus tree ( fig. 20 ) based on 54 transfor­ 16, 17). All species have typical exotrophic mation series of morphology (not including aquatic larvae (Altig and McDiarmid, 1999). Vibrissaphora). This tree suggests that Me­ We sampled Spea hammondii , Scaphiopus gophryinae ( Megophrys montana being their couchii , and S. holbrooki from Scaphiopod­ exemplar) is deeply imbedded within a paridae, and Pelobates fuscus and P. cultripes aphyletic Leptobrachiinae (the remaining from Pelobatidae . megophryid exemplars being of this nominal

a PELODYTIDAE (1 GENUS, 3 SPECIES): Ford subfamily); that Scutiger is composed of and Cannatella (1993; fig. 14 ) diagnosed Pe­ paraphyletic subgenus Scutiger and a monolodytidae as having a fused astragalus and phyletic subgenus Aelurophryne ), and that

50 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297

Oreolalax is composed of a paraphyletic sub­ were Brachytarsophrys feae , Leptobrachium genus Oreolalax and a monotypic subgenus chapaense , L. hasselti , Leptolalax bourreti , Aelurolalax . Megophrys nasuta , Ophryophryne hansi , O.

Within megophryines, Xie and Wang microstoma , Xenophrys major (formerly X. (2000) noted conflict between isozyme and lateralis ). We were unable to obtain samples karyological data regarding the monophyly of Atympanophrys , Leptobrachella, Oreolaof Brachytarsophrys , and also noted that lax, Scutiger , and Vibrissaphora, so, al­ Atympanophrys is only dubiously diagnos­ though we are confident that our sampling able from Megophrys or Xenophrys . They will allow phylogenetic generalizations to be also suggested that Xenophrys may not be made regarding the family, most of the probdiagnosable from Megophrys . lems within the group (e.g., the questionable

Lathrop (1997) suggested that, among monophyly of Leptobrachium , Leptolalax , nominal leptobrachiines, Leptolalax has no Megophrys , Scutiger , and Xenophrys ) will identified apomorphies. Xie and Wang remain unanswered. (2000) noted that Oreolalax is diagnosable from Scutiger on the basis of unique maxil­ ‘‘ ADVANCED ’’ FROGS—NEOBATRACHIA lary teeth and that Vibrissaphora has apo­ Neobatrachia9 includes about 96% of exmorphies (e.g., keratinized spines along the tant frogs and is a poorly understood array lips of adults), although the effect of recog­ of apparently likely paraphyletic groups with nizing Oreolalax and Vibrissaphora on the apomorphic satellites. So, at this juncture in monophyly of Scutiger has not been evalu­ our discussion the quantity of evidence sugated. Similarly, the monophyly of Leptobrachium is undocumented. 9 There is controversy regarding the appropriate name

2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

51 gested by authors to support major groups, prephylogenetic systematics, Neobatrachia and the quality of published taxonomic rea­ also has within it its own nominally ‘‘primsoning drops significantly to the realm of itive’’ groups aggregated on plesiomorphy grouping by overall similarity and special (e.g., Leptodactylidae ), as well as its own pleading for particularly favored characters. nominally ‘‘transitional’’ and ‘‘advanced’’,

, Like the larger­scale Archeobatrachia (prim­ groups (e.g., Ranidae and Rhacophoridae itive frogs), Mesobatrachia (transitional Arthroleptidae and Hyperoliidae ). Further frogs), and Neobatrachia (advanced frogs) of the unwillingness of the systematics com­

52 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297

munity to change taxonomies in the face of lossidae, and, presumably Rheobatrachidae evidence is best illustrated here. For exam­ as well. Darst and Cannatella (2004; fig. 22) ple, Brachycephalidae was shown to be im­ redelimited Hyloidea as the descendants of bedded within the leptodactylid taxon the most recent common ancestor of Eleuth­ Eleutherodactylinae, but the synonymy was erodactylini, Bufonidae , Centrolenidae, Hynot made by Darst and Canntella (2004), and linae, Phyllomedusinae , Pelodryadinae , and Leptopelinae was shown to be more closely Ceratophryinae , thereby excluding Heleorelated to Astylosternidae than to hyperoliine phrynidae, Limnodynastidae, Myobatrachihyperoliids by Vences et al. (2003c), but was dae, Rheobatrachidae , and Sooglossidae (and retained by those authors in an explicitly par­ by implication, presumably Nasikabatrachiaphyletic Hyperoliidae . dae) from Hyloidea 10. For this discussion, we Ford and Cannatella (1993; fig. 14 ) sug­ retain the older, more familiar definition of gested five characters in support of the Hyloidea as all neobatrachians excluding the monophyly of Neobatrachia: (1) (neo)palatine ranoids. bone present; (2) fusion of the third distal HELEOPHRYNIDAE (1 GENUS, 6 SPECIES): carpal to other carpals; (3) complete separa­ South African Heleophryne was considered tion of the m. sartorius from the m. semiten­ by Ford and Cannatella (1993; fig. 14 ) to be dinosus; (4) presence of an accessory head a member of Neobatrachia and Hyloidea. of the m. adductor longus; and (5) absence The synapomorphy of Heleophrynidae sugof the parahyoid bone. In addition, Haas gested by these authors includes only ab­ (2003; fig. 15) presented the following larval sence of keratinous jaw sheaths in exotrophic characters (but see Heleophrynidae ): (1) up­ free­living larvae. Haas (2003; fig. 15), in per lip papillation with broad diastema; (2) contrast, placed Heleophrynidae outside cartilage of the cavum cranii forms tectum Neobatrachia in a pectinate relationship parietale; (3) secretory ridges present; and among ‘‘pelobatoids’’ or as the sister taxon (4) pupil horizontally elliptical. The character of central importance historically to the 10 Because Darst and Cannatella (2004) used Limnorecognition of this taxon is the (neo)palatine dynastes and Heleophryne as outgroups to root the remainder of the tree, it was not possible for them to have bone, a character not without its own contro­ obtained a tree in which traditional Hyloidea is monoversy. phyletic so their statement (p. 46) that ‘‘the placement ‘‘ HYLOIDEA ’’: The worldwide Hyloidea, of some basal neobatrachian clades ( Heleophrynidae , for which no morphological synapomorphy Myobatrachidae , and Sooglossidae ) remains uncertain’’ is actually an assumption of their phylogenetic analysis. has been suggested, was long aggregated on Uncited by Darst and Cannatella (2004), Biju and Bosthe basis of its being ‘‘primitive’’ with re­ suyt (2003) differentiated between ‘‘Hyloidea’’ sensu spect to the ‘‘more advanced’’ Ranoidea , al­ lato (the traditional view of Hyloidea) and Hyloidea senthough molecular evidence under certain an­ su stricto, which they considered to be monophyletic and alytical methods and assumptions supports which, like the concept of Darst and Cannatella (2004), excluded Myobatrachidae , Limnodynastidae, Heleoits monophyly (Ruvinsky and Maxson, 1996; phrynidae, Sooglossidae , and Nasikabatrachidae . Anoth­ Feller and Hedges, 1998). Hyloidea is de­ er issue is that Ford and Cannatella (1993) and Cannafined by the plesiomorphic (at least within tella and Hillis (2004) defined the name Hylidae to apply Neobatrachia) possession of arciferal pecto­ cladographically to the hypothetical ancestor of Hemiphractinae , Hylinae , Pseudinae (now part of Hylinae ), ral girdles (coracoids not fused) and simple and Pelodryadinae , and all of its descendants. However, procoelous vertebrae, although descriptions Darst and Cannatella (2004) implied that their Hylidae of both characters have been highly reified was redefined to exclude Hemiphractinae . This redefithrough repetition and idealization. More re­ nition would be necessary to keep content and diagnosis cently, Biju and Bossuyt (2001: fig. 21 ) sug­ as stable as possible with respect to the traditional use of the term ‘‘Hylidae’’, because without this kind of regested on the basis of a DNA sequence anal­ definition in a system that aspires to precision, the preysis that Hyloidea, as traditionally viewed, is tense of precision is lost. For example, the cladographic paraphyletic with respect to Ranoidea , but definition of Hylidae by Ford and Cannatella (1993) and within ‘‘Hyloidea’’ is a monophyletic group Cannatella and Hillis (2004) applied to the cladogram of Darst and Cannatella (2004) would require that the

largely coextensive with ‘‘Hyloidea’’, but ex­ following be included within Hylidae : Brachycephalicluding Heleophrynidae , Limnodynastidae , dae, Leptodactylidae , Bufonidae , Centrolenidae , Den­ Myobatrachidae , Nasikabatrachidae , Soog­ drobatidae, and, likely, Rhinodermatidae .

2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

53

of Pelobates , Leptobrachium , and Mego­ We sampled one species each of the nomphrys. Heleophryne is included at this level inal sooglossid genera ( Nesomantis thomasin Haas’ analysis by having (1) m. tympan­ seti and Sooglossus sechellensis ). Of Nasiopharyngeus present; (2) m. interhyoideus kabatrachidae we sampled Nasikabatrachus posterior present; (3) m. diaphragmatoprae­ sahyadrensis as well as sequences attributed cordialis present; (4) m. constrictor bran­ by Dutta et al. (2004) only to an unnamed chialis I present; and (5) interbranchial sep­ species of Nasikabatachidae, also from the tum IV musculature with the lateral fibers of Western Ghats. Although Dutta et al. did not the m. subarcualis rectus II–IV invading the name their species as new, they explicitly septum, and lacking the characters listed by treated it as distinct from N. sahyadrensis Haas for Neobatrachia. In addition, the ver­ (Dutta et al., 2004: 214), and we therefore tical pupil and ectochordal vertebrae tie he­ follow their usage. (Our statement that Naleophrynids to myobatrachines, and non­neo­ sikabatrachidae contains two species rests on batrachians. Recent DNA sequence evidence this assertion, although any clear diagnosis (San Mauro et al., 2005; fig. 17 ) strongly of the second has yet to be cogently providsupports Heleophrynidae as the sister taxon ed.) All species of Sooglossidae that are of all other neobatrachians (although Biju known are endotrophic according to Thibau­ and Bossuyt, 2003, also on the basis of mo­ deau and Altig (1999). Sooglossus sechellenlecular evidence as well had suggested that sis has free tadpoles that are carried on the Heleophrynidae is the sister taxon of Lim­ back of the mother. The tadpoles are likely nodynastidae 1 Myobatrachidae ). endotrophic, but this is not definitely known We sampled Heleophryne purcelli and H. (R.A. Nussbaum, personal obs.). Dutta et al. regis . These species are likely close relatives (2004) reported exotrophic tadpoles occur­ (Boycott, 1982) so broader sampling (to have ring in fast­flowing streams for their unincluded H. rosei , whose isolation on Table named species of Nasikabatrachidae .

Mountain near Cape Town suggests a likely LIMNODYNASTIDAE (8 GENERA, 50 SPECIES), distant relationship to the other species) MYOBATRACHIDAE (11 GENERA, 71 SPECIES), would have been preferable. AND RHEOBATRACHIDAE (1 GENUS, 2 SPECIES): SOOGLOSSIDAE (2 GENERA, 4 SPECIES) AND Different authors consider this taxonomic NASIKABATRACHIDAE (1 GENUS, 2 SPECIES): cluster to be one family ( Myobatrachidae , Sooglossidae is a putative Gondwanan relict sensu lato) with two or three subfamilies (Savage, 1973) on the Seychelles, possibly (Heyer and Liem, 1976); to be two families, related to myobatrachids as evidenced by Limnodynastidae and Myobatrachidae (Zug sharing with that taxon the plesiomorphy of et al., 2001; Davies, 2003a, 2003b); or to be ectochordal vertebrae (J.D. Lynch , 1973), al­ three families, Limnodynastidae, Myobathough Bogart and Tandy (1981) suggested a trachidae, and Rheobatrachidae (Laurent, relationship with the arthroleptines (a ranoid 1986). Because Rheobatrachidae (Rheobagroup) . In fact, the group is plesiomorphic in trachus; Laurent, 1986) was only tentatively many characters, being arciferal (although associated with Myobatrachidae by Ford and having a bony sternum; see Kaplan, 2004, Cannatella (1993), we retain its familial stafor discussion of the various meanings of tus for clarity of discussion.

‘‘arcifery’’) and all statements as to its rela­ Limnodynastidae , Myobatrachidae , and tionships, based on morphology, have been Rheobatrachidae are primarily united on the highly conjectural. Biju and Bossuyt (2003; basis of their geographic propinquity on Ausfig. 21) suggested on the basis of DNA se­ tralia and New Guinea (Tyler, 1979; Ford quence evidence that Sooglossidae is the sis­ and Cannatella, 1993). And, only one line of ter taxon of the recently discovered Nasika­ evidence, that of spermatozoal morphology, batrachus, found in the Western Ghats of has ever suggested that these taxa taken to­ South India. Nasikabatrachus has so far had gether are monophyletic (Kwon and Lee, little of its morphology documented. They 1995). Heyer and Liem (1976) provided a).

also found Sooglossidae 1 Nasikabatrachi­ character analysis that assumed familial and dae to form the sister taxon of all other neo­ generic monophyly, but this was criticized batrachians. methodologically (Farris et al., 1982a 54 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297 2006 FROST ET AL.: AMPHIBIAN TREE OF LIFE

55

Rheobatrachinae ( Rheobatrachus ) is of un­ horizontal ones); (4) broad alary process certain position, although Farris et al. (Griffiths, 1959a), which they found in (1982a) in their reanalysis of Heyer and Myobatrachidae and Rheobatrachus (as well Liem’s (1976) data, considered it to be part as in Adenomera , Physalaemus [in the sense of Limnodynastinae. Ford and Cannatella of including Engystomops and Eupemphix ], (1993) subsequently argued that Rheobatra­ and Pseudopaludicola ); and (5) divided chinae is more closely related to Myoba­ sphenethmoid. trachidae than to Limnodynastidae , although Ford and Cannatella (1993) reported at this suggestion, like the first, rests on highly least one synapomorphy for Limnodynasticontingent phylogenetic evidence. Moreover, dae: a connection between the m. interman­ Myobatrachidae may be related to Sooglos­ dibularis and m. submentalis (also found in sidae (J.D. Lynch , 1973) and Limnodynasti­ Leptodactylinae and Eleutherodactylinae acnae to Heleophrynidae (J.D. Lynch , 1973; cording to Burton, 1998b). Rheobatrachus Ruvinsky and Maxson, 1996 ), although these was diagnosed by having gastric brooding of views are largely conjectural inasmuch as the larvae—an unusual reproductive mode, to character evidence of J.D. Lynch (1973) was say the least. It is tragic that the two species presented in scenario form. are likely now extinct (Couper, 1992).

Ford and Cannatella (1993) suggested, on Read et al. (2001) provided a phylogenetic the basis of discussion of characters present­ study of myobatrachine frogs (fig. 23) based ed by Heyer and Liem (1976), that Myoba­ on mtDNA sequence data that assumed trachidae (Myobatrachinae in their sense and monophyly of the group and used only Limpresumably including Rheobatrachus ) has nodynastes to root the myobatrachine tree. four morphological synapomorphies: (1) The evolutionary propinquity of Limnodynpresence of notochordal (ectochordal) verte­ astes ( Limnodynastidae ) and Myobatrachus brae with intervertebral discs; (2) m. petro­ ( Myobatrachidae ) was supported on the basis hyoideus anterior inserting on the ventral of DNA sequence evidence by Biju and Bosface of the hyoid; and, possibly, (3) reduction suyt (2003). of the vomers and concomitant reduction of We were able to sample at least one spevomerine teeth (J.D. Lynch , 1971). cies for most of the genera of the three nom­

Ford and Cannatella (1993) suggested sev­ inal families. For Limnodynastidae we sameral synapomorphies of Myobatrachidae and pled at least one species for all nominal gen­ Sooglossidae to the exclusion of Limnodyn­ era: Adelotus brevis , Heleioporus australiaastidae : (1) incomplete cricoid cartilage ring; cus, Lechriodus fletcheri , Limnodynastes (2) semitendinosus tendon inserting dorsal to depressus , L. dumerilii , L. lignarius , L. orthe m. gracilis (in myobatrachines excluding natus, L. peronii , L. salmini , Mixophyes car­ Taudactylus and Rheobatrachus , which have binensis, Neobatrachus sudelli , N. pictus , a ventral trajectory of the tendon; (3) hori­ Notaden melanoscaphus , Philoria sphagnizontal pupil (except in Uperoleia ; (also ver­ cola. Recent authors (e.g., Cogger et al., tical in Rheobatrachus ; limnodynastines 1983) have considered Kyarranus to be a primitively have a vertical pupil according to synonym of Philoria , and we follow this. Heyer and Liem, 1976, although several have J.D. Lynch (1971) provided morphological

to correspond with changes made after the paper was published. Use of Euhyas (instead of Eleutherodactylus ) is our modification to illuminate discussion.

56 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 297

characters that are evidence of monophyly of

Kyarranus 1 Philoria (e.g., presence of stub­

by fingers and concealed tympana as well as

direct development—Littlejohn, 1963; De

Bavay, 1993; Thibaudeau and Altig, 1999).

For Rheobatrachidae , we obtained Rheo­

batrachus silus. And, for Myobatrachidae ,

we obtained at least single representatives of

all nominal genera: Arenophryne rotunda ,

Assa darlingtoni , Crinia nimbus , C. signi­

fera, Geocrinia victoriana , Metacrinia ni­

chollsi, Myobatrachus gouldii , Paracrinia

haswelli, Pseudophryne bibroni , P. coriacea ,

Spicospina flammocaerulea , Taudactylus

acutirostris , and Uperoleia laevigata . With

exceptions, this taxon selection will not al­

low us to comment on generic monophyly,

but it will identify major monophyletic

groups and questions that will guide future

research. All rheobatrachids and most myob­

atrachids have endotrophic larvae and vari­

ous degrees of direct development (Thibau­

deau and Altig, 1999).

‘‘ LEPTODACTYLIDAE ’’ (57 GENERA, 1243

SPECIES): ‘‘Leptodactylidae’’ holds the same

position in the Americas as Myobatrachidae

(sensu lato, as containing Limnodynastidae

and Rheobatrachidae ) does in Australia —a

likely nonmonophyletic hodgepodge ‘‘prim­

itive’’ holochordal or rarely stegochordal, ar­

ciferal, and procoelous neobatrachian group

united by geography and not synapomorphy.

‘‘Leptodactylidae’’ is currently divided into

five subfamilies, some of which are not

clearly monophyletic (or consistently diag­

nosable) and some of which may be poly­

Fig. 23. Parsimony tree of Crinia , Geocrinia , phyletic (Ruvinsky and Maxson, 1996; Haas, and allied myobatrachids, of Read et al. (2001). 2003; Darst and Cannatella, 2004; Faivovich Data were of mtDNA: approximately 621 bp (266 et al., 2005; San Mauro et al., 2005; figs. 17 , variable) from the 12S rRNA region and 677 bp 22, 24). (383 variable) of ND2. Sequence alignment of 12S and ND2 were done under ClustalX (Thomp­ J.D. Lynch (1971, 1973) considered lepson et al., 1997) with gap opening and extension todactylids to be divided into four subfamicosts set at 50, and transversion: transition cost lies, on the basis of both synapomorphy and ratio set at 2. Ambiguously alignable regions were symplesiomorphy: (1) Ceratophryinae (for excluded. In analysis, transversion:transition costs Ceratophrys and Lepidobatrachus ); (2) Elowere set at 2. It was not stated whether gaps were siinae (5 Hylodinae of other authors; for treated as evidence but we infer that gaps were Crossodactylus , Hylodes , and Megaelosia ); treated as missing data. Branches marked with an (3) Leptodactylinae (for Barycholos, Edalorasterisk were collapsed in the original publication hina, Hydrolaetare , Leptodactylus [including because of low bootstrap support. Adenomera ], Limnomedusa , Lithodytes , Par­

Figures

Fig. 12. Salamandrid relationships suggested by Titus and Larson (1995) based on a parsimony analysis of 44 morphological character transformations and 431 informative sites of ca. 1.8 kb of the 12S and 16S mt rRNA and tRNAVal fragments of mtDNA. Sequence alignment was done using MALIGN (W.C. Wheeler and Gladstein, 1992) with equal weighting of transversions and transitions and a gap penalty cost of 6. Sequence data and morphology in parsimony analysis had equal costs and gaps were treated as evidence. The tree was rooted on Eurycea 1 Phaeognathus; tree length 5 2,081. Generic names are updated to reflect the naming of Lyciasalamandra (Veith and Steinfartz, 2004) and the partition of Triturus into Mesotriton, Lissotriton (not studied by Titus and Larson, 1995), and Triturus (García­París et al., 2004b).

Fig. 4. Relationships of salamanders suggested by Larson and Dimmick (1993). Families are noted on right. Typhlonectes and Xenopus were employed as outgroups. Consensus of 40 equally­parsimonious trees (length 5 460, ci 5 0.59). Data are 32 morphological and 177 molecular (nu rDNA) character transformations (from Larson, 1991). The method of DNA alignment was not specified. Gaps were excluded as evidence.

Fig. 7. Relationships of salamanders suggested by Wiens et al. (2005). Families are noted on right. Results reflect a parsimony analysis of 326 character transformations of morphology (221 parsimonyinformative), and DNA sequences from nu rRNA (212 bp from Larson, 1991; 147 parsimony­informative) and RAG­1 (1,530 bp; 624 parsimony­informative). Sequence alignment was made using Sequencher (Gene Codes Corp.). Morphological characters identified as paedomorphic were treated as unknown for adult morphology and in some cases hypothetical terminals were related­species chimaeras of composite molecular and morphological data. Molecular transformations were weighted equally in analysis. Inferred insertion­deletion events were coded as binary characters separate from the nucleotide sequence characters and indel­required gaps within sequences were coded as missing. The tree was rooted on Gymnophiona 1 Anura.

Fig. 16. Tree of amphibians provided by Roelants and Bossuyt (2005). This tree reflects a maximumlikelihood analysis of 3,963 aligned positions (2,022 variable and 1,788 parsimony­informative) of three protein­coding nuDNA genes (ca. 555 bp of RAG­1, ca. 675 bp of CXCR­4, ca. 1280 bp of NCX­1) and ca. 1940 bp of the mitochondrial genome (part of 16S and tRNAMet, and all of tRNALeu, tRNAIle, ND­1, and tRNAGln). Alignment was done initially using ClustalX (Thompson et al., 1997; presumably applying default cost functions) followed by a probabilistic method implemented in the program ProAlign (Löytynoja and Milinkovitch, 2003) and, in the case of 16S and tRNA seqments, subsequently modified manually, guided by models of secondary structure for Xenopus. Gaps were treated as missing data and ambiguously aligned sequences were excluded. The model of evolution assumed was GTR 1 G 1 I.

Fig. 14. Narrative tree of relevant anuran taxa by Ford and Cannatella (1993). A branch subtending Hylidae 1 Pseudidae in the original figure is collapsed per errata distributed with reprint. An asterisk was used by these authors to denote a metataxon, and quotation marks to denote nonmonophyly.

Fig. 17. Tree of amphibians provided by San Mauro et al. (2005). This tree reflects a maximumlikelihood analysis of 1,368 bp of the nuclear protein­coding gene RAG­1, assuming the GTR 1 G 1 I substitution model (as suggested by ModelTest v. 3.6; Posada and Crandall, 1998). Sequence alignment was made manually with only one gap excluded from analysis.

Fig. 18. Tree of Pelobatoidea and outgroups of García­París et al. (2003) based on 1,000 bp of two mitochondrial genes: cytochrome c and 16S rRNA. The sequences were aligned using Clustal X (Thompson et al., 1997) using default costs then manually modified based on published secondary­structure models of the 16S gene. Gaps were treated as missing data and data were analyzed under the assumption of the GTR 1 G substitution model, as suggested by ModelTest 3.06 (Posada and Crandall, 1998). The tree was rooted on Ascaphus montanus 1 A. truei. Quotation marks denote nonmonophyly.

Fig. 20. Consensus of 12 equally parsimonious trees of selected members of Megophryidae of Delorme and Dubois (2001), rooted on Scaphiopus and Pelodytes. Underlying data were 54 transformation of morphology, rooted on Pelodytes and Scaphiopus (ci 5 0.581; ri 5 0.713). Although the tree and a list of the underlying character transformation were provided, no association was made between the character transformations and taxa or particular branches on the tree, rendering the analysis practically unrepeatable. Nominal subfamilies are noted on the right.

Fig. 21. Bayesian tree of anuran exemplars of Biju and Bossuyt (2003), with particular reference to Neobatrachia. Underlying data are two mtDNA fragments, covering part of 12S rRNA, complete t­ RNAVal, and part of 16S rRNA. In addition, one fragment of the nuclear genome: exon 1 of rhodopsin, single exon of RAG­1, and exon 2 of CXCR­4, for a total of 2,325 bp of sequence. Alignment was made using Clustal X (Thompson et al., 1997), alignment costs not disclosed, with ambiguous sections excluded and gaps excluded as evidence. Model of nucleotide substitution assumed for analysis was GTR 1 G1 I.