ORNITOLOGIA NEOTROPICAL 16: 347?359, 2005 ? The Neotropical Ornithological Society INDEPENDENT EVOLUTION OF TWO DARWINIAN MARSH- DWELLING OVENBIRDS (FURNARIIDAE: LIMNORNIS, LIMNOCTITES) Storrs L. Olson1, Martin Irestedt2, 3, Per G. P. Ericson2, & Jon Fjelds?4 1Division of Birds, National Museum of Natural History, Smithsonian Institution, P.O. Box 37012, Washington, DC 20013-7012, U.S.A. E-mail: olsons@si.edu 2Department of Vertebrate Zoology and Molecular Systematics Laboratory, Swedish Museum of Natural History, P.O. Box 50007, SE?104 05 Stockholm, Sweden. E-mail: martin.irestedt@nrm.se & per.ericson@nrm.se 3Department of Zoology, University of Stockholm, SE?106 91 Stockholm, Sweden. 4Vertebrate Department, Zoological Museum, University of Copenhagen, Universitetsparken 15, DK-2100 Copenhagen ?, Denmark. E-mail: jfjeldsaa@zmuc.ku.dk Resumen. ? Evoluci?n independiente de dos horneros de pantano (Furnariidae: Limnornis, Limnoc- tites). ? La Pajonalera Pico Curvo (Limnornis curvirostris) y la Pajonalera Pico Recto (Limnoctites rectirostris) son dos especies de hornero de pantano colectados por primera vez por Charles Darwin en Uruguay. Ambas tienen una distribuci?n limitada a Uruguay, el sur de Brasil y el norte de Argentina, ?rea en la cual ocupan h?bitat muy diferentes. Descritas originalmente como cong?neres debido a sus similitudes en plu- maje, las dos especies han sido consideradas parientes cercanas a pesar de diferencias estructurales obvias entre ambas. Analizamos secuencias de ADN de tres genes de estas dos especies y las comparamos con una amplia variedad de otras especies de Furnariidae y varios grupos externos. Limnoctites rectirostris perte- nece al grupo de especies tradicionalmente agrupadas en Cranoleuca, estando m?s cercanamente relacio- nada al Curuti? Ocr?ceo (C. sulphurifera) entre las especies muestreadas. Estos resultados est?n respaldados por vocalizaciones y nidificaci?n. Limnornis curvirostris forma un clado con el Junquero (Phleocryptes mela- nops); este clado tiene al Macuqui?o (Lochmias nematura) como grupo hermano. Una relaci?n cercana entre Limnornis y Phleocryptes es respaldada por el color azul de los huevos y la arquitectura del nido, la cual es apa- rentemente ?nica en estos dos g?neros. Abstract. ? The Curve-billed Reedhaunter (Limnornis curvirostris) and the Straight-billed Reedhaunter (Lim- noctites rectirostris) are marsh-dwelling ovenbirds that were first collected by Charles Darwin in Uruguay. Each has a limited distribution in southernmost Brazil, Uruguay, and northern Argentina, within which the birds occupy very distinct habitats. Originally described as congeners because of overall similarity of plum- age, the two species have been treated as close relatives through most of their history despite obvious structural differences. We analyzed DNA sequences from three different genes of these species, compar- ing them with a wide variety of other species of Furnariidae and several outgroup taxa. Limnoctites rectirostris belongs among the species traditionally placed in Cranioleuca, being most closely related to the marsh- dwelling Sulphur-throated Spinetail (C. sulphurifera) among the species we sampled. This is supported by vocalizations and nidification. Limnornis curvirostris forms a clade with the Wren-like Rushbird (Phleocryptes melanops), with the Sharp-tailed Streamcreeper (Lochmias nematura) as a rather distant sister-taxon. A close relationship between Limnornis and Phleocryptes is supported by the apparently unique nest architecture and blue-green egg color. Accepted 5 April 2005.347 Key words: Furnariidae, Limnoctites, Limnornis, molecular systematics, nidification, ovenbirds. OLSON ET AL. INTRODUCTION Charles Darwin was the first naturalist to col- lect two paludicolous species of ovenbirds that are now known to be of very limited dis- tribution in southeastern South America. These were obtained in June 1833 (Steinhe- imer 2004; not 1832 as per Vaurie 1980: 211) in what is now the province of Maldonado in the Republica Oriental del Uruguay. When John Gould (1839: 80?81, pll. 25?26) identi- fied and described the birds from the voyage of H. M. S. Beagle, he created a new genus, Limnornis, for these two species, calling the first L. rectirostris, for its very straight, pointed bill, and the second L. curvirostris, for its bill with a more typically curved tip. In English these species are now called the Straight- billed and the Curve-billed reedhaunters, respectively. Both species were said by Darwin (in Gould 1839:80?81) to live in the same habitat ?amongst the reeds on the borders of lakes? and that he was ?unable to point out any dif- ferences? in the habits of the two. This mis- taken impression probably colored much of the subsequent thinking about these birds. Apart from the supposedly shared habitat, the main similarity between the two reedhaunters is in general coloration and plumage pattern. This is doubtless what lead Gould (1839) to his original decision to place them both in the same genus and also why Vaurie (1980) rather vigorously defended this course. On the other hand, there are manifest differences between the two, so that they have most often been placed in separate monotypic genera (Limnor- nis and Limnoctites). Regardless, the two reedhaunters first collected by Darwin have long been regarded as each other?s closest rel- ative. This also appeared to receive support We decided to test this hypothesis by review- ing the information on morphology, ecology, and nidification, and by comparing this with new molecular evidence. SYNOPSIS OF MORPHOLOGY AND NOMENCLATURAL HISTORY Both reedhaunters are plain brownish or dull rufous above, with rufous tails, dull whitish undersides, and no adornment apart from a whitish superciliary stripe. Limnornis curviros- tris is a reasonably robust bird (26.7?33.3 g, mean 29.0 g, n = 10, according to USNM specimen data; mean of 21 eggs 24.6 x 17.9 mm, according to Narosky et al. 1983) with a longish, rounded tail and a stout, curved bill. It has very much the appearance of a drab version of one of the smaller species of Fur- narius. Limnoctites rectirostris is a much slighter bird (15.6?24.5 g, mean 19.2 g, n = 5, accord- ing to USNM specimen data [the heaviest was a very fat laying female]; mean of 3 eggs 20.3 X 15.3 mm, according to Ricci & Ricci 1984) with an extremely long, straight, slender bill, and a shorter, graduated tail with very pointed, usually worn, rectrices. The nomen- clatural history of the two species has revolved around whether to emphasize their similarities or their differences. Gould (1839) did not designate a type species for Limnornis and Gray (1840) subse- quently selected L. curvirostris as the genotype. Sclater (1889) overlooked Gray?s action and applied a new name, Limnophyes, to L. curviros- tris, reserving the name Limnornis for L. rec- tirostris. But Limnophyes was preoccupied by a genus of Diptera, so Oberholser (1899) pro- posed Thryolegus as a replacement. Hellmayr (1925) pointed out Gray?s (1840) type desig- nation, returned curvirostris to Limnornis, and348 from a phylogenetic analysis of nest structure (Zyskowski & Prum 1999) in which Limnornis and Limnoctites were said to form a group with the Wren-like Rushbird (Phleocryptes melanops). proposed the new generic name Limnoctites for L. rectirostris. Since then, most authors have maintained the two species in monotypic genera, usually placing them next to one EVOLUTION OF TWO MARSH-DWELLING OVENBIRDS another. The first exception appears to be Esteban (1949), who advocated removing Limnoctites rectirostris to the subfamily Synallaxinae near Certhiaxis, presumably leaving Limnornis curvi- rostris in the Philydorinae, where both were placed by Sclater (1890). Peters (1951) also dissociated the two, placing Limnornis immedi- ately after Furnarius, with 12 genera between it and Limnoctites, which was placed between Cer- thiaxis and Poecilurus/Cranioleuca. Although Meyer de Schauensee (1966) followed Peters? sequence almost exactly, one of his few depar- tures was to place Limnoctites immediately after Limnornis. Vaurie (1971, 1980) and Sibley & Monroe (1990) returned to the original Goul- dian nomenclature and combined both spe- cies in Limnornis. Other authors (e.g., Ridgely & Tudor 1994) have preferred to emphasize differences by recognizing two genera for the reed-haunters, although still maintaining their close association. Remsen (2003) maintained them as adjacent in his linear sequence but noted that differences in tail structure, nesting materials, and egg color called into question their proposed sister relationship. DISTRIBUTION AND HABITAT For nearly a century, Darwin?s original two specimens of L. rectirostris were the only ones known. Sanborn (1929) next obtained the species at another locality in Uruguay in 1926. Gradually, its range was extended from the provinces of Entre Rios and Buenos Aires, Argentina, to Rio Grande do Sul and Santa Catarina, in southernmost Brazil, and the spe- cies was also found more extensively in Uru- guay (Daguerre 1933, Pereyra 1938, Esteban 1949, Escalante 1956, Gerzenstein & Achaval 1967, Zorilla de San Martin 1963, Alda do Argentina, by W. H. Hudson (Sclater & Salvin 1868). Durnford (1877: 182) found the spe- cies common in the same province and was ?at a loss to understand how this bird could have escaped the observation of naturalists till Mr. Darwin?s visit to South America.? Addi- tional specimens were obtained in Uruguay (Sclater & Hudson 1888, Sclater 1890, Hell- mayr 1925), Sanborn?s (1929) assertion that his 1926 specimens from Uruguay were the first since Darwin being erroneous. By 1899, the range of the species was extended to Rio Grande do Sul, Brazil (Ihering 1899). Although Darwin stated (in Gould 1839) that the two reedhaunters occurred together and Vaurie (1980: 212) asserted that ?L. rec- tirostris shares the same habitat [as L. curviros- tris], but its requirements are less rigid,? this is not, in fact, the case. Ridgely & Tudor (1994: 61) note that the two species ?appear never to occur together in the same marsh.? Belton (1984: 622) shows no overlap in range between the species in Rio Grande do Sul, Brazil. Olson has experience with L. curviros- tris in Argentina, and with both species through much of Uruguay, and has not yet visited a site where both species might be expected to occur in proximity. Limnoctites rectirostris typically occurs in marshes. Although a variety of plants may occur in such sites, particularly at the edges, the bird is found only where the spiny cara- guat?, Eryngium spp., dominates. In the orni- thological literature this plant has erroneously been referred to as a sedge (e.g., Ridgely & Tudor 1994, Remsen 2003: 226, but correctly as an ?apiaceous herb? on 261) or a grass (Gerzenstein & Achaval 1967, Vaurie 1980, Babarskas & Fraga 1998). It is actually a dicot that belongs in the carrot family (Apiaceae or Umbelliferae). Ricci & Ricci (1984: 205) cor-349 Ros?rio 1996, Babarskas & Fraga 1998). After Darwin?s original collection, Limnor- nis curvirostris was next collected from 1866 to 1868 at Conchitas, Buenos Aires Province, rectly describe the plants as growing in ?bro- meliad-like rosettes,? the leaves of which are beset with sharp spines that make the pursuit of birds in this habitat a decided challenge to OLSON ET AL. the ornithologist?s flesh and clothing. The species that have been mentioned in connec- tion with Limnoctites are Eryngium horridum (Belton 1984, Remsen 2003), E. pandanifolium (Gerzenstein & Achaval 1967, Babarskas & Fraga 1998, Remsen 2003), and E. eburneum (Ricci & Ricci 1984). The ranges of these plants are given as southern Brazil to NE Argentina, with the last two extending to Par- aguay (Cabrera 1965). That the birds are tied to the plant and not necessarily to marshy environments is shown by their occurrence at 200 to 250 m in rocky scrub forest where Eryngium pandanifo- lium occurs in 5 to 20 m wide patches along streams (Gerzenstein & Achaval 1967). That L. rectirostris occurs as high as 1100 m (Ridgely & Tudor 1994, BirdLife International 2000, Remsen 2003) is presumably based on the populations in Aparados da Serra National Park mentioned by Belton (1984: 621), though no substantive documentation of their occurring so high appears to exist. Although also a marsh bird, L. curvirostris is found in extensive reedbeds, especially of the giant sedge or pajonal (Scirpus giganteus), the grass known as espada?a (Zizaniopsis bona- riensis), and also cattails (Typha). Although its briefly stated range from Rio Grande do Sul to Buenos Aires is the same as that of L. rec- tirostris, within that area it is much the more abundant and widespread of the two species because of the greater extent of its habitat. In summary, although the distributions of the two reedhaunters are superficially similar, they are adapted to distinct habitats and are probably never syntopic. NIDIFICATION Nest structure in the Furnariidae is extremely basis of their supposedly building a domed nest with ?a small awning over the nest entrance? (p. 899). The nest of Limnornis curvi- rostris is described by von Ihering (1902) and Belton (1984); that of Limnoctites rectirostris by Daguerre (1933), Ricci & Ricci (1984) and Sick (1993); and those of both species in Vau- rie (1980) and Narosky et al. (1983). Included here are all the references cited by Zyskowski & Prum (1999), plus some others, but none mentions an awning over the entrance of the nest of Limnoctites. This error came about from misinterpretation (K. Zyskowski pers. com. to Olson, June 2004) of photographs of the nest of Phylloscartes ventralis in the article preceding the paper by Ricci & Ricci that were erroneously captioned as Limnoctites. Zyskowski (pers. com. ibid.) has since found and photographed a nest of L. rectirostris in Uruguay and confirmed that it does not have an awning. Remsen (2003) took his descrip- tions of the nests of both reedhaunters as having awnings from Zyskowski & Prum (1999). The first mention of an ?awning? being constructed by Limnornis curvirostris appears to be that in one of two nests of described by Narosky (in Vaurie 1980: 213). Later, Narosky et al. (1983: 36) confirmed that the nest of L. curvirostris ?possesses, like that of the junquero Ph[leocryptes] melanops, a 3 cm projection or eave above the mouth of the entrance? (our translation). On the other hand, it does seem that the basic nest structure of Limnornis and Phleo- cryptes is similar, best exemplified by Narosky?s descriptions in Vaurie (1980). Phleocryptes pre- sents a presumably more derived condition in covering its nest with mud. The vast majority of species of Furnari- idae have pure white eggs, although the eggs350 diverse and has been used to devise a phylog- eny of the family (Zyskowski & Prum 1999). In this phylogeny, Limnornis and Limnoctites were grouped with Phleocryptes melanops on the in a few species of Synallaxis may have a light bluish, greenish or yellowish cast (Sch?nwet- ter & Meise 1967: 12, Sick 1993: 428). In con- trast, the eggs of Limnornis curvirostris and 351 EVO LU TIO N O F TW O M A RSH -D W ELLIN G O V EN BIRD S TAB signations are given according to two alter Swedish Museum of Natural History; USN rsity of Copenhagen. Spec tochrome b Myoglobin G3PDH Cinc Furn Furn Upu Auto Loch Phily Thri Anu Asth Cory Cran Cran Cran Syna Lim Lim Phleo Geos Scler Drym Xiph Dend Sitta Cham Ptero Scyta Y590044d Y064279c Y065700a Y590048d Y065699a Y065702a Y065701a Y065709a Y065705a Y065710a Y065708a Y065707a Y590043d Y065715a Y065711a Y065712a Y065713b Y065714a Y065718a Y065717a Y065716a AY590054d AY064255c AY065756a AY590058d AY065755a AY065758a AY065757a AY065765a AY065761a AY065766a AY065764a AY065763a AY590053d AY065772a AY065768a AY065769 a AY065770b AY065771a AY065776a AY065774a AY065773a AY590065d AY590066d AY590078d AY590081d AY590076d AY590077d AY590072d AY590070d AY590073d AY590069d AY590068d AY590063d AY590080d AY590088d AY590093d AY590087d AY590092d AY590095d AY590096d AY590097 RefeLE 1. Specimen data and Genbank accession numbers for samples used in the study. Family and subfamily de native classification schemes. Acronyms: AMNH = American Museum of Natural History, New York; NRM = M = National Museum of Natural History, Smithsonian Institution; ZMUC = Zoological Museum of the Unive ies Classif. by Irestedt et al. (2002, in press) Classif. by Remsen (2003) Voucher no. Cy lodes fuscus arius cristatus arius leucopus certhia jelskii molus leucophthalmus mias nematura dor atricapillus padectes flammulatus mbius annumbi (*) enes cactorum phistera alaudina ioleuca albicapilla ioleuca pyrrhophia ioleuca sulphurifera llaxis ruficapilla nocites rectirostris nornis curvirostris cryptes melanops itta tenuirostris urus scansor ornis bridgesii ocolaptes major rocincla tyrannina somus griseicapillus aeza meruloides ptochos tarnii lopus spillmanni Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Sclerurinae Furnariidae: Sclerurinae Furnariidae: Dendrocolaptinae Furnariidae: Dendrocolaptinae Furnariidae: Dendrocolaptinae Furnariidae: Dendrocolaptinae Formicariidae Rhinocryptidae Rhinocryptidae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Furnariinae Furnariidae: Philydorinae Furnariidae: Philydorinae Furnariidae: Philydorinae Furnariidae: Philydorinae Furnariidae: Synallaxinae Furnariidae: Synallaxinae Furnariidae: Synallaxinae Furnariidae: Synallaxinae Furnariidae: Synallaxinae Furnariidae: Synallaxinae Furnariidae: Synallaxinae Furnariidae: Synallaxinae Furnariidae: Synallaxinae Furnariidae: Synallaxinae Furnariidae: Furnariinae Furnariidae: Philydorinae Dendrocolaptidae Dendrocolaptidae Dendrocolaptidae Dendrocolaptidae Formicariidae Rhinocryptidae Rhinocryptidae ZMUC S220 NRM 966772 ZMUC 125590 ZMUC S439 NRM 937251 ZMUC S2577 NRM 937334 ZMUC S428 NRM 966903 ZMUC S150 NRM 966910 ZMUC 124797 NRM 966821 USNM B17199 NRM 956643 USNM B14895 USNM B2735 USNM B2734 ZMUC S292 NRM 937258 NRM 966930 NRM 966847 NRM 976662 NRM 967031 ZMUC S2053 AMNH RTC467 ZMUC S540 A A A A A A A A A A A A A A A A A A A A A rences: aIrestedt et al. (2001), bIrestedt et al. (in press), cEricson et al. (2002), dFjelds? et al. (2005). OLSON ET AL. Phleocryptes melanops are a deep greenish-blue (Sch?nwetter & Meise 1967: 12; Vaurie 1980). We reviewed the egg color of at least 32 gen- era of Furnariidae, mainly based on the col- lections of the British Museum (Michael Walters, pers. com. to Olson, June 2004) and scattered references in more recent literature and found no exceptions to the preceding observations. Thus, the decidedly greenish blue eggs of Limnornis curvirostris and Phleoc- ryptes appear to be unique within the Furnari- idae, a similarity noticed at least as early as Pereyra (1938). In contrast, the nest and eggs of Limnoc- tites rectirostris have been likened to that of Cranioleuca. ?The [white] eggs and nest more closely resemble those of C. sulphurifera than those of L. curvirostris, whose eggs are bluish green and the nest [of L. curvirostris], although spherical with a lateral entrance, is more con- spicuous for being constructed at consider- able height (up to 1.6 m), and is without much differentiation between the external material and the lining? (translated from L?pez-Lan?s et al. 1999: 63). Within the Furnariidae the white eggs and less elaborate nest of L. rectirostris are proba- bly plesiomorphic states, or relatively so in the case of the nest, the apparently unique awninged nest and blue-green eggs of L. curvi- rostris and Phleocryptes are derived conditions that argue for a sister-group relationship for these two taxa. MOLECULAR SYSTEMATICS Materials and methods. Twenty species of oven- birds and four woodcreepers were selected for the molecular analysis. In addition to Lim- nornis and Limnoctites, we included representa- tives of all major clades of furnariids herein) and the woodcreepers (Irestedt et al. 2002, Chesser 2004, Fjelds? et al. 2005). Serv- ing as outgroups are three representatives of the proposed sister clade of Furnariidae (Irestedt et al., 2002, Chesser 2004): Pteropto- chos tarnii and Scytalopus spillmanni (family Rhi- nocryptidae) and Chamaeza meruloides (family Formicariidae). Sample identifications and GenBank accession numbers are given in Table 1. We sequenced the complete myoglobin intron 2 (along with 13 bp and 10 bp of the flanking regions of exons 2 and 3, respec- tively), the complete glyceraldehydes-3-phos- phodehydrogenase (G3PDH) intron 11 (along with 36 bp and 18 bp of exons 11 and 12, respectively), and 999 bp from the cyto- chrome b gene (see Ericson et al. 2002, Irestedt et al. 2002, and Fjelds? et al. 2003, 2004, for primer sequences and procedures). Positions where the nucleotide could not be determined with certainty were coded with the appropriate IUPAC code. Due to the low number of insertions in the introns, the combined sequences could easily be aligned by eye. All gaps in the myoglobin and the G3PDH sequences were treated as miss- ing data in the analyses. No insertions, dele- tions, stop or nonsense codons were observed in any of the cytochrome b sequences. ModelTest 3.06 (Posada & Crandall 1998) in conjunction with PAUP* (Swofford 1998) was used to evaluate the fit of the data to dif- ferent models for nucleotide substitutions. The GTR+I+? model has the best fit for the combined data set and for the cytochrome b partition, while GTR+G was selected for both the myoglobin intron 2 and the G3PDH intron 11 partitions. These models were used in the analyses of the individual genes, as352 identified by Fjelds? et al. (2005), as well as the genera Sclerurus and Geositta, which form the sister group to a clade consisting of all the other ovenbirds (called ?core ovenbirds? well in the analysis of the combined data set. The posterior probabilities of trees and parameters in the substitution models were approximated with Markov chain EVOLUTION OF TWO MARSH-DWELLING OVENBIRDS Monte Carlo and Metropolis coupling using the program MrBayes (Huelsenbeck & Ron- quist 2001, Ronquist & Huelsenbeck 2003). We ran two analyses of 500,000 genera- tions for each gene with trees sampled every 100 generation. The parameter estimates from the two separate MCMC runs for each data set were compared and found to be very simi- lar, thus allowing an inference from the con- catenated output. Posterior probabilities for the individual genes were based on a total of 9000 trees saved after discarding the trees saved during the ?burnin phase? (as estimated graphically) in each analysis. The analysis of the combined data set was conducted in the same manner as for the individual genes except that the number of generations in each run was two millions. The 50% majority-rule consensus trees were identical in the runs and the posterior probabilities are based on a total of 390,000 saved trees. Results. The concatenated sequences became 2164 basepairs long after alignment. Within the ingroup (all ovenbirds and woodcreepers) the lengths of the myoglobin sequences range from 677 in Philydor to 701 bp in Geositta, and the lengths of the G3PDH sequences range from 349 in Dendrocincla to 401 bp in Xiphoco- laptes. The myoglobin intron is the least varia- myoglobin intron 2 and G3PDH intron 11 sequences requires postulation of a few inser- tions and deletion events (indels) among the core ovenbirds, most of which involve only a single basepair (singletons) and/or are found only in a single taxon (autapomorphic). The phylogenetic trees obtained from the Bayesian analyses of the individual genetic markers, as well as of the combined data set, are generally similar (Figs 1A-C and 2). The analyses also agree well on the systematic positions of Limnornis and Limnoctites, which clearly are not sister taxa. Most major group- ings of ingroup taxa are recovered by all data partitions and receive generally high supports (Table 3). Monophyly of the ingroup is strongly corroborated, as is the basal position of Geositta and Sclerurus relative to the core ovenbirds and woodcreepers, which in turn are recovered as sister groups. All data parti- tions except cytochrome b also support a basal position among the core ovenbirds of a clade consisting of Automolus, Thripadectes and Philydor. Although the cytochrome b data set also recognizes monophyly of this clade, it leaves the group unresolved in relation to the other core ovenbirds. The remaining core ovenbirds are divided into two clades that receive strong support in the analyses of most data partitions. Limnornis groups with TABLE 2. Descriptive statistics for the observed pairwise, uncorrected sequence divergencies (p-distances) between selected groups of taxa. Larger distances suggest higher rates of nucleotide substitutions. Cytochrome b Myoglobin intron 2 G3PDH intron 11 Mean Min. Max. Mean Min. Max. Mean Min. Max. Within core ovenbirds Core ovenbirds vs woodcreepers Core ovenbirds vs Sclerurus/Geositta-clade Core ovenbirds vs Outgroups 12.63 14.53 15.57 17.7 2.4 12.21 13.51 15.84 15.62 17.12 18.12 20.32 2.3 4.17 3.77 7.09 0.29 2.96 2.8 5.67 3.62 5.16 4.92 8.6 5.16 6.84 7.71 9.31 0.76 4.59 4.53 6.68 7.73 8.83 9.63 11.54353 ble among the three genetic markers studied herein (Table 2). The observed substitution rate is larger in the mitochondrial cytochrome b gene, in accordance with previous studies (Irestedt et al. 2004). The alignment of the Phleocryptes with 100% posterior probabilities in the analyses of both myoglobin and the genes combined. The analyses of cytochrome b and G3PDH also suggest a Limnornis-Phleo- cryptes clade, albeit with weaker support. OLSON ET AL. Cinclodes, Lochmias, Upucerthia and the two Asthenes, Anumbius and Coryphistera. Limnoctites FIG. 1. Majority-rule consensus trees obtained from Bayesian analyses of three genetic markers: A. cyto- chrome b, B. myoglobin intron 2, C. glyceraldehydes-3-phosphodehydrogenase (G3PDH) intron 11. Pos- terior probabilities are indicated at the nodes.354 species of Furnarius are the other members of this larger group of core ovenbirds. The other group consists of the three species of Cranioleuca, Limnoctites, Synallaxis, falls well within the Cranioleuca clade, but it groups with different species of Cranio- leuca depending on which genetic marker is studied. EVOLUTION OF TWO MARSH-DWELLING OVENBIRDS DISCUSSION and within the sample we tested was closest to FIG. 2. Maximum-likelihood tree calculated from the combined data set (cytochrome b, myoglobin intron 2 and glyceraldehydes-3-phosphodehydrogenase intron 11). Posterior probabilities from the Bayesian analysis are indicated at the nodes.355 DNA sequence data unambiguously show that Limnornis curvirostris and Limnoctites rectiros- tris are not particularly closely related. Limnoc- tites falls out among the species of Cranioleuca the Sulphur-throated Spinetail C. sulphurifera. This is another marsh-inhabiting species with a distribution similar to the two reedhaunters, except that it also occurs farther inland and to the south in Argentina (Remsen 2003). As OLSON ET AL. recounted above, Esteban (1949) advocated putting Limnoctites in the Synallaxinae. Gerzenstein & Acheval (1967), followed by Ricci & Ricci (1984), noted its general similar- ity in appearance to Cranioleuca sulphurifera. L?pez-Lan?s et al. (1999) concurred and also considered that the vocalizations, nest, and eggs of Limnoctites were more similar to C. sul- phurifera than to Limnornis curvirostris. In the straight bill and in tail and rectrix shape, Limnoctites agrees better with Cranio- leuca than with Limnornis. The juvenal plumage of C. sulphurifera, which lacks the breast streaks and yellow throat of the adult, is like that of Limnoctites in every respect except for its rufous and black pattern in the wing. Similarities extend to vocalizations as well. ?L. rectirostris gives a hissing trill Ti-ti-ti-ti- titititritriiiii, accelerating at the end. This is maintained at a stable frequency of between 4 and 6.5 kHz, lasting about 2.5 and 3.5 s, com- posed of 14 to 18 elements, varying up to 22 ... . The territorial advertising vocalization of sulphurifera, which gives a vocalization of simi- lar length (between 2.5 and 3.5 s) but with a wider frequency range (between 1 and 6.5 kHz) and a greater and more complex quan- tity of elements? (translated from L?pez- Lan?s et al. 1999: 62). In our molecular phylogeny Limnornis is well separated from Limnoctites and its closest relative among the taxa we sampled is the Wren-like Rushbird Phleocryptes melanops, with the Sharp-tailed Streamcreeper Lochmias ne- matura as the closest outlying sister group. There is nothing in the external morphology of these three genera that suggests a particu- larly close relationship. An obligate inhabitant of streams in dense forests (Remsen 2003 ), Lochmias differs strikingly from the other two in its habits. Phleocryptes, like Limnornis, inhabits marshy reedbeds and the two may occur together in the same marsh, although Phleocryptes is much more widely distributed, from Pacific Ecua- dor to Chile and east across parts of Bolivia, TABLE 3. Posterior probabilities for selected nodes obtained in analyses of different data partitions. Combined data set Cytochrome b Myoglobin intron 2 G3PDH 11 Monophyly of Furnariidae (ovenbirds and woodcreepers) Basal position of a Sclerurus/Geositta-clade Monophyly of core-ovenbirds Basal position of Automolus, Thripadectes and Philydor among core-ovenbirds Monophyly of a Cranioleuca spp. and Limnoc- tites clade Sister group relationship between Limnoctites and Cranioleuca sulphurifera Sister group relationship of Limnornis and Phleocryptes 100% 100% 100% 100% 100% 100% 100% 100% 98% 100% No support, but no contradition 99% 100% 76% 100% 100% 100% 92% 100% 100% 100% 100% 90% 100% 93% 100% Alternative tree topology suggested 61%356 this species is well differentiated from that of L. curvirostris, whose song sounds rough and faltering. On the other hand, it could be con- founded by an untrained ear with that of C. Paraguay, most of Argentina, Uruguay and southern Brazil (Remsen 2003). It is a smaller bird with a much more variegated plumage than Limnornis. Apart from DNA sequences, EVOLUTION OF TWO MARSH-DWELLING OVENBIRDS the best evidence for a relationship between these two genera comes from the nests and eggs, as outlined above. CONCLUSIONS The available molecular, morphological, and behavioral data all indicate that Limnoctites rec- tirostris belongs among the species currently included in the genus Cranioleuca, within which it appears to be most closely related to another marsh-dwelling species, C. sulphurifera. Thus it appears possible that Limnoctites rec- tirostris may be a large, paedomorphic (in plumage) derivative of C. sulphurifera that moved out of reedbeds and became adapted to marshes of Eryngium, where its long, straight bill is possibly an adaptation for extracting prey from the spiny rosettes of that plant, as suggested by Ricci & Ricci (1984). It would be premature, however, to make any taxonomic or nomenclatural recommen- dations until the systematics of the entire genus Cranioleuca has been undertaken. Zyskowski & Prum (1999) indicate that there are two distinct groups within Cranioleuca based on nest structure, which they designate as the albiceps group and the pyrrhophia group. Limnoctites and C. sulphurifera belong to the pyrrhophia group. The type species of Cranio- leuca is C. albiceps and there does not seem to be any previously recognized generic name available for the pyrrhophia group (Hellmayr 1925). Thus, if this group were to be sepa- rated generically from Cranioleuca, the 8 spe- cies now in it would presumably have to take the name Limnoctites, which would certainly be an ironic turn of events. Another possibility is that L. rectirostris and C. sulphurifera may be sufficiently distinct as to be separated in Lim- noctites, so that the rest of the pyrrhophia group hypothesis is corroborated by the similar nest structure, egg coloration, and to some extent by the similarity in microhabitat. These two taxa otherwise appear to be sufficiently dis- tinct from one another morphologically and molecularly to justify the recognition of sepa- rate monotypic genera. ACKNOWLEDGMENTS For assistance in the field in Argentina and Uruguay, we are greatly indebted to Joaquin Aldabe, J. Phillip Angle, Luis Chiappe, Miguel Clara, Santiago Claramunt, Christina Geb- hard, and Christopher Milensky. Jorge Cravino, Director, Departamento de Fauna, Montevideo was instrumental in providing permits and critical information as well as arranging for accommodations for fieldwork on numerous estancias in Uruguay, to whose generous owners we are most grateful. George and Janet Winter provided a sumptu- ous base of operations for fieldwork in Uru- guay in 2002. J. V. Remsen, Michael Walters, and Kristof Zyskowski provided much useful information and references. Field work in Argentina and Uruguay was supported by the Virginia Y. Hendry Fund and the Alexander Wetmore Endowment Fund, Smithsonian Institution, respectively. Samples owned by the Swedish Museum of Natural History were collected in Paraguay in collaboration with the Museo Nacional de Historia Natural del Para- guay, San Lorenzo. The Swedish Research Council (grant no. 621-2001-2773 to P.E.) funded the laboratory work. We thank J. V. Remsen and Frank Steinheimer for many use- ful comments on the manuscript. REFERENCES357 would require a new name. Our molecular evidence indicates that the previous intimation of a relationship between Limnornis and Phleocryptes is correct. This Alda do Ros?rio, L. 1996. As aves em Santa Catar- ina: distribui??o geogr?fica e meio ambiente. Funda??o do Meio Ambiente-FATMA, Flori- an?polis, Brazil. OLSON ET AL. Babarskas, M, & R. Fraga. 1998. Actualizando la distribuci?n de la Pajonalera Pico Recto Limnoc- tites rectirostris en la provincia de Entre R?os, Argentina. Cotinga 10: 79?81. Belton, W. 1984. Birds of Rio Grande do Sul, Bra- zil. Part 1. Rheidae through Furnariidae. Bull. Am. Mus. Nat. Hist. 178: 369?636. BirdLife International. 2000. Threatened birds of the world. Lynx Edicions, Barcelona. Cabrera, A. L. 1965. Flora de la provincia de Bue- nos Aires. Parte IV: Oxalid?ceas a Umbel?feras. Coleccion Cientifica del Instituto Nacional de Tecnolog?a Agropecuaria, Buenos Aires, Argentina. Chesser, R. T. 2004. Molecular systematics of New World Suboscine birds. Mol. Phyl. Evol. 32: 11?24. Daguerre, J. B. 1933. Dos aves nuevas para la fauna Argentina. Hornero 5: 213?215. Durnford, H. 1877. Notes on the birds of the Province of Buenos Aires. Ibis ser. 4, 1: 166? 203. Ericson, P. G. P., L. Christidis, M. Irestedt, & J. A. Norman. 2002. Systematic affinities of the lyre- birds (Passeriformes: Menura), with a novel classification of the major groups of passerine birds. Mol. Phyl. Evol. 25: 53?62. Escalante Rossi, R. 1956. Nuevo hallazgo de la pajera de pico recto en El Uruguay. Hornero 10: 164?166. Esteban, J. G. 1949. Una nueva ave para Brasil ?Limnoctites rectirostris? (Gould). Acta Zool. Lilloana 8: 147?150. Fjelds?, J., M. Irestedt, & P. G. P. Ericson. 2005. Molecular data reveal some major adaptational shifts in the early evolution of the most diverse avian family, the Furnariidae. J. Ornithol. 146: 1?13. Fjelds?, J., D. Zuccon, M. Irestedt, U. S. Johansson, & P. G. P. Ericson. 2003. Sapayoa aenigma: a New World representative of ?Old World suboscines?. Proc. R. Soc. Lond. B. (Suppl.) 270: 238?241. Gerzenstein, E., & F. Achaval. 1967. Nuevos datos Fitzroy, R. N., during the years 1832 to 1836. Signature pp. 57?96 issued Nov. 1839. Smith Elder & Co., London. . Gray, G. R. 1840. A list of the genera of birds, with an indication of the typical species of each genus. Richard & John E. Taylor, London, U. K. Hellmayr, C. E. 1925. Catalogue of birds of the Americas and the adjacent islands. Field Mus. Nat. Hist. Zool. Ser. 13, pt. 4: 1?390. Huelsenbeck, J. P. & F. Ronquist. 2001. MrBAYES: Bayesian inference of phylogenetic trees. Bioin- formatics 17: 754?755. Ihering, H. von. 1899. As aves do Estado do Rio Grande do Sul. Pp. 113?154 in De Azambuja, G. A. (ed.). Annuario do Estado do Rio Grande do Sul para o anno de 1900. Gundlach & Krahe, Porto Alegre, Brazil. Ihering, H. von. 1902. Descrip??o de novos ninhos e ovos. Rev. Museu Paulista 5: 291?303. Irestedt, M., J. Fjelds?, J. A. A. Nylander & P. G. P Ericson. 2004. Phylogenetic relationships of typical antbirds (Thamnophilidae) and test of incongruence based on Bayes factors. BMC Evol. Biol. 4: 23. Irestedt, M., J. Fjelds?, U. S. Johansson, & P. G. P. Ericson. 2002. Systematic relationships and biogeography of the tracheophone suboscines (Aves: Passeriformes). Mol. Phylogenet. Evol. 23: 499?512. L?pez-Lan?s, B., A. G. Di Giacomo, & M. Babars- kas. 1999. Estudios sobre ecologia y compor- tamiento de la Pajonalera Pico Recto Limnoctites rectirostris en la Reserva Otamendi, Buenos Aires, Argentina. Cotinga 12: 61?63. Meyer de Schauensee, R. 1966. The species of birds of South America and their distribution. Livingston Publishing Company, Narberth, Pennsylvania. Narosky, S., R. Fraga, & M. de la Pe?a. 1983. Nidi- ficacion de las aves argentinas (Dendrocolap- tidae y Furnariidae). Asociaci?n Ornitol?gica del Plata, Buenos Aires, Argentina. Oberholser, H. C. 1899. Some untenable names in358 sobre Limnornis rectirostris. Hornero 10: 307? 314. Gould, J. 1839. Birds. Part 3 [1838?1841] in Dar- win, C. (ed.). The zoology of the voyage of H. M. S. Beagle, under the command of Captain ornithology. Proc. Acad. Nat. Sci. Phila. 1899: 201?216. Pereyra, J. A. 1938. Aves de la zona ribere?a nor- deste de la Provincia de Buenos Aires. Mem. Jardin Zo?l. La Plata 9: 1?304. EVOLUTION OF TWO MARSH-DWELLING OVENBIRDS Peters, J. L. 1951. Check-list of birds of the world. Vollume 7. Museum of Comparative Zoology, Cambridge, Massachusetts. Posada, D., & K. A. Crandall. 1998. Modeltest: test- ing the model of DNA substitution. Bioinfor- matics 14: 817?818. Remsen, J. V. 2003. Family Furnariidae (Oven- birds). Pp. 162?357 in del Hoyo, J., A. Elliot, & D. A. Christie (eds.). Handbook of the birds of the world. Volume 8: Broadbills to tapaculos. Lynx Edicions, Barcelona, Spain. Ricci, J. J., & F. Ricci. 1984. Nidificaci?n de la pajonalera de pico recto (Limnornis rectirostris) en Benavidez, Buenos Aires, Argentina. Hornero 12: 205?208. Ridgely, R. S., & G. Tudor. 1994. The birds of South America. Volume 2. The suboscine pas- serines. Univ. of Texas Press, Austin. Ronquist, F., & J. P. Huelsenbeck. 2003. MrBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572?1574. Sanborn, C. C. 1929. Some Uruguay records. Auk 46: 251. Sch?nwetter, M., & W. Meise. 1967. Handbuch der Oologie. Lieferung 14, Akademie-Verlag, Ber- lin. Sclater, P. L. 1889. On some new species and gen- era of birds of the family Dendrocolaptidae. Proc. Zool. Soc. Lond. 1889: 32?34. Sclater, P. L. 1890. Catalogue of the birds in the British Museum. Volume 15. British Museum, London, U. K.. Sclater, P. L., & W. H. Hudson. 1888. Argentine Ornithology. Volume 1. R. H. Porter, London, U. K. Sclater, P. L., & O. Salvin. 1868. List of birds col- lected at Conchitas, Argentine Republica, by Mr. William H. Hudson. Proc. Zool. Soc. Lond. 1868: 137?146. Sibley, C. G., & B. L. Monroe, Jr. 1990. Distribution and taxonomy of birds of the world. Yale Univ. Press, New Haven, Connecticut. Sick, H. 1993. Birds in Brazil. Princeton Univ. Press, Princeton, New Jersey. Steinheimer, F. D. 2004. Charles Darwin?s bird col- lection and ornithological knowledge during the voyage of H .M. S. ?Beagle?, 1831?1836. J. Ornithol. 145: 300?320. Swofford, D. L. 1998. Paup* 4.0. Phylogenetic anal- ysis using parsimony (* and other methods). Sinauer Association, Sunderland, Massachu- setts. Vaurie, C. 1971. Classification of the ovenbirds (Furnariidae). H. F. & G Witherby Ltd., Lon- don, U. K.. Vaurie, C. 1980. Taxonomy and geographical distri- bution of the Furnariidae (Aves, Passeri- formes). Bull. Am. Mus. Nat. Hist. 166: 1?357. Zorilla de San Mart?n, J. C. 1963. Notes on the rare furnariid Limnoctites rectirostris of Uruguay. Con- dor 65: 531?533. Zyskowski, K., & R. O. Prum. 1999. Phylogenetic analysis of the nest architecture of Neotropical ovenbirds (Furnariidae). Auk 116: 891?911.359