journal of field ornithology · s. cinnamomea), morfo de color i (morfo de s. cinnamomea) y morfo...

J. Field Ornithol. 79(4):352–363, 2008 DOI: 10.1111/j.1557-9263.2008.00186.x

Entre Rıos Seedeater (Sporophila zelichi):a species that never was

Juan I. Areta

Grupo FALCO, Calle 67 Nro 1725 e/67 y 68, 1900, La Plata, Buenos Aires, Argentina

Received 2 March 2008; accepted 16 May 2008

ABSTRACT. Entre Rıos Seedeaters (Sporophila zelichi), also called Zelich’s Seedeaters, White-collaredSeedeaters, and Narosky’s Seedeaters, are one of the rarest birds in the Neotropics. However, doubts have beenraised about the validity of this species. Therefore, I evaluated the systematic status of Entre Rıos Seedeaters basedon analysis of previously unpublished vocal and habitat data. I tested four hypotheses regarding the systematicstatus of S. zelichi: Good Species Hypothesis (valid species), Hybridization Hypothesis (hybrid S. palustris ×S. cinnamomea), Color Morph Hypothesis I (morph of S. cinnamomea), and Color Morph Hypothesis II (morphof S. palustris). The songs and preferred habitat of S. zelichi are indistinguishable from those of Marsh Seedeaters(S. palustris), and the songs of both forms have exhibited similar changes from the early 1990s to 2003–2007. Incontrast, the songs and preferred habitat of Chestnut Seedeaters (S. cinnamomea) differ from those of S. zelichi.Therefore, the Good Species Hypothesis is rejected by vocalization and habitat data, the Hybrid Hypothesis isundermined by the absence of shared vocal characters and limited habitat overlap of the proposed parental formsS. cinnamomea/S. palustris, and Color Morph Hypothesis I is rejected by both song and habitat data. However,Color Morph Hypothesis II is supported by both song and habitat data. Thus, I propose that S. zelichi be considereda color morph of S. palustris.

SINOPSIS. El Capuchino Collar Blanco (Sporophila zelichi): una especie que nunca fueEl Capuchino de Collar Blanco (S. zelichi) es una de las aves mas raras del Neotropico. Su validez como especie

se ha puesto en duda repetidas veces. Por esto, evalue el status de S. zelichi basandome en el analisis de datos devocalizaciones y habitat previamente no publicados. Pongo a prueba explıcitamente cuatro hipotesis sobre el estatussistematico de S. zelichi: hipotesis de especie buena (especie valida), hipotesis de hibridacion (hıbrido S. palustris xS. cinnamomea), morfo de color I (morfo de S. cinnamomea) y morfo de color II (S. palustris). El Capuchino deCollar Blanco es indistinguible del Capuchino Pecho Blanco (S. palustris) tanto en voces como en habitat preferido,y las vocalizaciones de ambos cambiaron en tandem entre principios de los 1990’s y 2003–2007. El CapuchinoCastano (S. cinnamomea) difiere marcadamende de S. zelichi en vocalizaciones y habitat preferido. Por lo tanto, lahipotesis de especie buena es refutada por los datos de vocalizaciones y habitat, la hipotesis de hibridacion no esapoyada por la falta de caracteres vocales compartidos y la escasa superposicion de habitat entre sus supuestas formasparentales S. cinnamomea/S. palustris, la hipotesis de morfo de color I es refutada por los datos de vocalizaciones yhabitat. Sin embargo, la hipotesis de morfo de color II es fuertemente corroborada por los datos de vocalizacionesy habitat. Propongo tratar a S. zelichi como morfo de color de S. palustris.

Key words: color morph, preferred habitat, seedeater, Sporophila, systematics, vocalizations

The Capuchinos, or Caboclinhos, represent asubgroup of the genus Sporophila that includes11 species: Dark-throated Seedeater (S. ruficol-lis), Rufous-rumped Seedeater (S. hypochroma),Chestnut Seedeater (S. cinnamomea), MarshSeedeater (S. palustris), Entre Rıos Seedeater(S. zelichi), Tawny-bellied Seedeater (S. hy-poxantha), Ruddy-breasted Seedeater (S. min-uta), Black-and-tawny Seedeater (S. nigrorufa),

Current address: CICyTTP-CONICET, Materi& Espana, 3105, Diamante, Entre Rıos, Argentina.Email: [email protected]

Capped Seedeater (S. bouvreuil) and, althoughapparently not as closely related, Chestnut-bellied Seedeater (S. castaneiventris) and Black-bellied Seedeater (S. melanogaster) (Sick 1997,Lijtmaer et al. 2004). Based on limited infor-mation, there have been several systematic treat-ments of the Capuchinos in the ruficollis group(from S. ruficollis to S. hypoxantha in the abovedefined Capuchinos group) (Sclater 1871, Hell-mayr 1904, 1938, Meyer de Schauensee 1952,1966, Paynter and Storer 1970, Short 1969,1975, Narosky and Yzurieta 1987, Ridgely andTudor 1989, De la Pena 1989, Ouellet 1992,Pearman and Abadie 1995, Silva 1999, Remsenet al. 2006).

C©2008 The Author(s). Journal compilation C©2008 Association of Field Ornithologists

352

Journal of Field Ornithology

Vol. 79, No. 4 Seedeater Systematics 353

All species in the ruficollis group are long-distance migrants whose seasonal movementsseem to be related to being stem gleaner spe-cialists, vacating their breeding grounds as seedsmature and fall from the stalks to the ground(Silva 1999). Three species, Entre Rios Seedeater(S. zelichi), Marsh Seedeater (S. palustris), andChestnut Seedeater (S. cinnamomea), breedonly in eastern Argentina, southeastern Brazil,and eastern and western Uruguay, and migratethrough Paraguay to winter in the Cerrado andPantanal of Brazil (Silva 1999, J. Areta, pers.obs.). The distributional patterns of S. cinnamo-mea, S. palustris, and S. zelichi are similar, anddiffer from the more widespread S. hypochroma,S. hypoxantha, and S. ruficollis (Short 1975,Ridgely and Tudor 1989, Silva 1999, Azpiroz2003, Clay and Field 2003, Bencke 2004, GuyraParaguay 2004).

The most recently discovered species in thegenus is the Entre Rıos Seedeater (S. zelichi),described as a new species on the basis of twomales from the Entre Rıos province, Argentina,and collected by Mateo Zelich (Narosky 1977).Based on plumage features (white collar, ru-fous body, and gray cap in S. zelichi; whitethroat, rufous belly, gray back, nape, and cap inS. palustris; and chestnut body and gray cap inS. cinnamomea), doubts were raised about thevalidity of this species even in its formal descrip-tion, where it was considered to be “a hybrid ofS. palustris and S. cinnamomea” (E. Eisenmann,in Narosky 1977: 347). Later, raising additionaldoubts about its identity, investigators suggestedthat S. zelichi was a color morph of eitherS. cinnamomea (Ridgely and Tudor 1989) orS. palustris (Pearman and Abadie 1995).

Entre Rıos Seedeaters, also called Zelich’sSeedeaters, White-collared Seedeaters, andNarosky’s Seedeaters, are one of the rarest birdsin the Neotropics, and have only recently beenreported outside of Argentina, in Uruguay,Paraguay, and Brazil (Azpiroz 2003, Clay andField 2003, Bencke 2004). With a decliningpopulation estimated at 50–250 individuals inits restricted range, Entre Rıos Seedeaters areamong the few critically endangered species inSouth America (BirdLife 2000, 2005).

A conservation priority for the Entre RıosSeedeater is to determine its systematic position(BirdLife 2005). Progress on this issue has beenthwarted by their limited numbers (BirdLife2005, J. Areta, pers. obs.) and an apparent

absence of genetic differentiation among theCapuchinos (Lijtmaer et al. 2004). Thus, toelucidate the taxonomic status of Entre RıosSeedeaters, comparative information about vo-calizations, natural history, and habitat use isneeded.

The vocalizations of Entre Rıos Seedeatershave not been studied and their natural historyand habitat preferences remain poorly docu-mented (Pearman and Abadie 1995, Azpiroz2003, Bencke 2004). My objective was to exam-ine and compare the vocalizations and habitatuse of Entre Rıos Seedeaters with those of tworelated forms, Marsh Seedeaters and ChestnutSeedeaters, to shed light on their taxonomicstatus. I also use data concerning distributions,relative abundance, plumage, continuity, anddistributional ranges of other Sporophila speciesto elucidate the taxonomic position of the EntreRios Seedeater.

METHODS

I surveyed an area encompassing most of thebreeding ranges of S. zelichi, S. palustris, andS. cinnamomea, including the Mesopotamiangrasslands of Entre Rıos, Corrientes and SantaFe provinces in Argentina, and the departmentsof Rocha and Paysandu in Uruguay. I searchedfor Capuchinos during four breeding seasonsfrom October 2003 to March 2007. I spent69 d in the field at 25 locations, with 33 d at 11locations in Entre Rıos, 28 d at 7 locations inCorrientes, 3 d at 2 locations in Santa Fe, 3 d at2 locations in Paysandu, and 3 d at 3 locationsin Rocha (Fig. 1).

To characterize habitat use, I assigned eachterritorial bird to one of three broad habitatcategories: (1) marsh or wet grassland habitat,(2) undulating grassland, and (3) Caraguatal-Cardal. Marsh or wet grassland habitat includedany grass- or sedge-dominated plant communityregularly flooded during the rainy season orwith permanent water. Plants characterizing thishabitat included diverse grasses (e.g., Paspalumand Panicum), reeds (Typha), and rushes (Scir-pus). Undulating grassland included dry grass-lands located on the ridges, valleys, and slopesof the Mesopotamian grasslands of Entre Rıos,Corrientes, and close to the Uruguay River inPaysandu. Typical features of this habitat includerocky outcrops, sandy soils, and diverse grasses(e.g., Paspalum, Bromus, and Piptochaetium),

354 J. I. Areta J. Field Ornithol.

Fig. 1. Location of sites in Argentina (1–20) and Uruguay (21–25) visited during my study.1-Ibicuy (59◦09′W, 33◦44′S), 2-Arroyo Nancay (58◦44′W, 33◦23′S), 3-Ea. La Marita (58◦35′W, 33◦20′S),4-Gualeguaychu (58◦30′W, 33◦00′S), 5-Las Piedras (58◦33′W, 32◦53′S), 6-Larroque (59◦00′W, 33◦02′S),7-Urdinarrain (58◦53′W, 32◦41′S), 8-Pto. Liebig and Arroyo Caraballo (58◦11′W, 32◦09′S), 9-PN El Palmar(58◦18′W, 31◦55′S), 10-Arroyo Baru (58◦27′W, 31◦52′S), 11-San Salvador (58◦30′W, 31◦37′S). Santa FeProvince. 12-Sauce Viejo (60◦50′W, 31◦46′S), 13-Campo del Medio (60◦08′W, 31◦08′S). Corrientes Province.14-Mercedes (58◦05′W, 29◦10′S), 15-Ea. Rincon del Socorro (57◦10′W, 28◦32′S), 16-Cnia. Pellegrini(57◦10′W, 28◦35′S), 17-Camba Trapo (56◦51′W, 28◦27′S), 18-Cuenca del Rıo Aguapey (56◦56′W, 28◦36′S),19-PN Mburucuya (58◦05′W, 28◦00′S), 20-Rincon Sta. Marıa (56◦35′W, 27◦30′S). 21- Banados de la IndiaMuerta (53◦50′W, 33◦45′S), 22- Cebollatı (53◦38′W, 33◦15′S), 23- Laguna Negra (53◦40′W, 34◦00′S).Paysandu Department, 24- Lorenzo Geyres-Quebracho (57◦55′W, 32◦04′S), and 25- Queguayar (57◦50′W,32◦00′S).

and the occasional presence of yatay palms(Butia yatay). Caraguatal-Cardal is a dry grass-land with large stands of saw-grass, known ascaraguata (Eryngium horridum), cardos (natu-ralized thistles, Cardus spp.), isolated bushes ofChilca (Bacharis spp.), and occasional espinillo(Acacia caven) and nandubay (Prosopis affinis)trees. These habitats are widely interspersed, and

habitat types may coexist as a mosaic of differenthabitats in a single locality or in a smallergeographic area. Although the broad habitatcategories used may mask subtle differences thatcould indicate habitat segregation on a finergrain, they were designed to apply over a widearea where more detailed descriptions of habitatswould have precluded the discovery of any


habitat-use pattern because features unique toeach site would have masked overall similaritiesbetween sites.

Because individuals were not marked, I in-cluded only data from territorial males thatwere spatially or temporally segregated to avoidpseudoreplication. For areas visited more thanonce, I included only one record per territory.I considered each contact to be an indepen-dent sample for evaluating habitat preference,with habitat preference evaluated by comparingfrequency of occurrence of each taxon in eachhabitat. Preferred habitat was defined as thehabitat where I found at least 65% of theterritorial males. Habitat overlap was calculatedas the proportion of encounters where two ormore species shared a habitat.

I recorded vocalizations using a directionalmicrophone (ME-67, Sennheiser, Old Lyme,CT) protected with a shock-mount, windshield,and windjammer (Rycote, Gloucestershire, UK)and a tape recorder (PMD-222, Marantz,Aurora, Illinois). Spectrograms were preparedusing Syrinx 2.1 (J. Burt, www.syrinxpc.com).Additional recordings were provided by otherinvestigators (see Appendix 1). All my recordingswill be archived at the Macaulay Library ofNatural Sounds (Cornell Lab of Ornithology,Ithaca, New York) and those by M. Pearman aredeposited at the British Library Sound Archive(Kensington, London, UK).

The songs of Capuchinos include many dif-ferent, nonrepetitive, and morphologically com-plicated notes. To examine differences and simi-larities among different taxa, I first characterizednotes that, based on shape (including durationand frequency distribution) and relative positionin the songs, could be identified unambiguouslydespite variation among individuals. I then com-

Table 1. Four alternative hypotheses and predictions derived from them that allow for an evaluation of thesystematic status of the Entre Rios Seedeater (Sporophila zelichi).

Predictions regarding S. zelichi

Hypothesis Vocalizations Preferred habitat

1 Good species hypothesis Unique (−) Unique (−)2 Hybridization hypothesis Intermediate between S. palustris Shared by both S. palustris

and S. cinnamomea (−) or identical to and S. cinnamomea (±, −)S. palustris or S. cinnamomea (±, −)

3 Color morph hypothesis I Same as S. cinnamomea (−) Same as S. cinnamomea (−)4 Color morph hypothesis II Same as S. palustris (+) Same as S. palustris (+)For each prediction, support is indicated by (+), rejection by (−), and inconclusive evidence by (±).

pared the frequency of occurrence of variousnotes for individuals within and among species.

I estimated the relative abundance of S. zelichiby calculating a relative abundance ratio. I onlyused data from my own surveys to ensure thatrelative abundance was not biased by observeror sampling effort. The relative abundance ratio(RAR) was calculated as the number of malesof S. zelichi/number of males of S. palustris orS. cinnamomea. I used data from the literatureand from my surveys to describe patterns ofdistribution, ranges, abundance, and tempo-ral continuity of S. zelichi, S. palustris, andS. cinnamomea.

Systematic hypotheses. Four hypotheseshave been proposed regarding the status of S.zelichi (Table 1). To avoid unconscious pref-erences, I used the method of multiple work-ing hypotheses (Chamberlin 1965). I deducedpredictions for each hypothesis based on theRecognition Species Concept (Paterson 1985),whereby any bird species must have a pre-ferred or normal habitat where a specific materecognition system (SMRS) ensures successfulreproductive encounters between members of aspecies (Paterson 1985, Vrba 1995). Habitat-dependent selection acting on the SMRS is thusthe main force thought to influence speciation,and two closely related valid species should notshare the same preferred habitat. I consideredvocalizations to be a key element in species-specific recognition (i.e., a crucial part of theSMRS; see Slabekoorn and Smith 2002 forreview of habitat-dependent divergence in vocal-izations and speciation). I used the RecognitionConcept of Species to evaluate the specific statusof the Capuchinos (Paterson 1985, Haffer 1997)because this concept clearly delimits species andnarrows the meaning of the word species to a


restricted biological phenomenon, permits thetesting of predictions, and makes the results ofmy study easily interpretable by researchers en-dorsing other species concepts. Although manySporophila species coexist and share habitats, thespecies of Capuchinos that I studied differ insongs and in preferred habitat, suggesting thatboth habitat preference and vocalizations havediverged together in the Capuchinos (Areta,unpubl. data). Therefore, although vocalizationsalone might serve to assess specific status ofpopulations, a strong test of specific identityshould test vocalizations and distinctive habitatas key features of any valid species.

The four hypotheses that I tested were:

(1) Good Species Hypothesis (Narosky 1977).According to this hypothesis, S. zelichi isa valid species. If so, I predicted that theywould have vocalizations and a preferredhabitat different from those of closelyrelated species.

(2) Hybridization Hypothesis (E. Eisenmann,in Narosky 1977). The Hybridization Hy-pothesis considers S. zelichi to be a hybridof S. cinnamomea and S. palustris. If true,I predicted that the songs of S. zelichiwould either be intermediate between S.palustris and S. cinnamomea (assumingsongs are genetically determined) or iden-tical to the paternal form (S. palustris orS. cinnamomea, if learned from the father).This hypothesis predicts that the habitat ofS. zelichi would be shared by both parentalforms (S. palustris and S. cinnamomea).

(3) Color Morph Hypothesis I (Ridgely andTudor 1989, Sibley and Monroe 1990).This hypothesis considers S. zelichi tobe a color morph of S. cinnamomea.If true, I predicted they would havesongs and preferred habitats like those ofS. cinnamomea.

(4) Color Morph Hypothesis II (Pearman andAbadie 1995). This hypothesis considersS. zelichi to be a color morph of S. palustris.If true, S. zelichi should have songs andpreferred habitats like those of S. palustris.

Female Capuchinos in the ruficollis group areindistinguishable by plumage (visual spectrum)and their vocal repertoire is limited to calls andsoft chatters (J. Areta, pers. obs.). Thus, I usedmale songs and habitat use to test all hypotheses.Distribution patterns, extent of occurrence, rel-

ative abundance, and temporal continuity datawere also used to assess the validity of thesehypotheses.

RESULTS

Vocalizations. I obtained recordings ofthree S. zelichi males, 34 S. palustris males,and 24 S. cinnamomea males (Appendix 1). Iidentified 15 notes in the song repertoires ofmale S. palustris (Fig. 2, Table 2) and all butone of those notes were also used by maleS. zelichi (Fig. 2, Table 2). However, none ofthese notes were identified in the songs of maleS. cinnamomea (Table 2). In contrast, noneof 11 notes identified in the songs of maleS. cinnamomea (Fig. 3) were used by male S.palustris or S. zelichi. The last segments of thefull song of S. cinnamomea were seldom recordedand could not be characterized.

The songs of S. zelichi and S. palustris sharedistinctive notes not present in the songs ofother Capuchinos (Areta, unpubl. data). Thetwo most frequently notes used by S. zelichi andS. palustris were D and O. The sound of theuniquely inflected note D is readily identified inthe field as belonging to either of the two forms(Figs. 2A and D). The “V-shaped call” (noteO) is a short, high-pitched, pure-toned whistle,lasting almost one-third second, that descendsand then ascends to half of its initial frequencyand ends in a small “nail” that bends downward(Figs. 2A and 2E). Among the Capuchinos, onlyS. zelichi and S. palustris have these characteristicnotes (J. Areta, pers. obs.). The similarity in finestructure and tempo of these last two notes isstriking, being even audibly distinctive amongall the Capuchinos in the ruficollis group. The Fand K notes of S. cinnamomea can be consideredhomologous to notes D and O of S. palustrisand S. zelichi based on position in the song, anddiffer dramatically from them (cf. Figs. 2A vs.3A, 2D vs. 3B, and 2E vs. 3C).

I did not detect any geographical variation inany of the three species, with the same note typesidentified at all localities for each species. How-ever, when recordings of S. zelichi, S. palustris,and S. cinnamomea made in 1991–1993 and2003–2007 are compared (Appendix 1, Figs. 2and 3), morphological changes of apparentlyhomologue notes can be tracked through time.The recordings of S. palustris and S. zelichi fromthe early 1990s are more similar to each other


Fig. 2. Representative spectrograms comparing vocalizations of S. zelichi (N = 3) and S. palustris (N =34). Numbers in parentheses correspond with data of individuals in Appendix 1, locality name and localitynumber as in Figure 1 are also included here to facilitate comparison of similitude among localities. (A) Fullsong and call of S. palustris in 2003–2007 depicting delimitation of note types; song (10, Ea. Rincon delSocorro-15) and call (5, Ea. Rincon del Socorro-15), (B) Initiation and middle portion of songs in 1991–1993of S. palustris (above, AJ30, Mercedes-14) and S. zelichi (below, MP3, Gualeguaychu-4), (C) Middle and finalportion of songs in 1991–1993 of S. palustris (above, MP32, Gualeguaychu-4) and S. zelichi (below, MP3,Gualeguaychu-4), (D) Full songs in 2003–2007 of S. palustris (above, 10, Rincon del Socorro-15) and S.zelichi (below, 2, Ea. La Marita-3), (E) Calls in 2003–2007 of S. palustris (above, 5, Rincon del Socorro-15)and S. zelichi (below, 2, Rincon del Socorro-15) call. Inferred homologies are described by the use of anapostrophe (’). Notes 1, 2, 3, and notes with an apostrophe were only found in recordings made in the early1990s, and not in my recordings from 2003 to 2007.


Table 2. Note types in the songs of S. palustris and the percentage of individual male S. palustris (N = 34)and S. zelichi (N = 3) whose songs included those notes.

Species a b c d e f g h i j k l m n o

S. palustris 21% 59% 85% 85% 74% 74% 74% 68% 62% 50% 85% 32% 44% 41% 71%S. zelichi 33% 33% 100% 100% 67% 67% 67% 67% 33% 33% 0% 33% 33% 67% 33%No male S. cinnamomea (N = 24) had songs that included any of these notes. Letters correspond to lettersin Figure 2.

than they are to recordings from 2003 to 2007and vice versa (cf. Figs. 2B and 2C vs. 2D).The D note was almost vertical in the early1990s (labeled d’, Fig. 2), but had a moderateslope in the 2003–2007 recordings (labeled d,Fig. 2). A similar change in pattern is foundin note F of S. cinnamomea (Fig. 3), wherenotes identified from recordings made duringthe early 1990s differ in shape from those from2003 to 2007. The F note starts with a cleardownward inflexion in recordings from the early1990s (labeled f’) that is lacking in any birdrecorded in 2003–2007 (labeled f ) (cf. Fig. 3Babove and below). Moreover, three notes used byboth S. zelichi and S. palustris in the early 1990s(notes 1, 2, and 3 in Fig. 2C) were not identifiedin recordings from 2003 to 2007 (Figs. 2A and2D, Appendix 1). This suggests that the songs ofthe Capuchinos changed from the early 1990sto 2003–2007. More important, however, is thatthe songs of S. zelichi and S. palustris exhibitedsimilar changes, whereas those of S. cinnamomea,although diverging in time, did not.

Habitat. S. zelichi (12 territories) andS. palustris (46 territories) were always found inmarsh or wet grassland habitat. S. cinnamomea(34 territories) were found in undulating grass-land (N = 25, or 73.5%), marsh or wet grassland(N = 5, or 14.7%), and in cardal-caraguatal(N = 4, or 11.7%). The only habitat sharedby the three species was marsh or wet grassland(see Appendix 2 for sampling localities). Overlapbetween S. cinnamomea and S. palustris occurredonly in wet grassland where 46 S. palustris andfive S. cinnamomea were found, whereas neitherS. palustris nor S. zelichi were ever found inthe typical S. cinnamomea habitat of undulatinggrasslands. Thus, although differing in preferredhabitat, there is some overlap between S. palustrisand S. cinnamomea, and this overlap occurs intypical S. zelichi habitat.

Relative abundance and continuity. Ifound a skewed relative abundance ratio, with

one S. zelichi for every 15 S. cinnamomea andone S. zelichi for every 20 S. palustris. Althoughdiffering in abundance, all three species can befound continuously in nature.

Hypothesis testing. The hypothesis thatS. zelichi is a valid species is rejected by bothsong and habitat data, with no species-specificnotes in its songs and S. zelichi and S. palustrisoccupying similar habitat. Moreover, althoughthe distributional pattern of the three forms isconsistent with all the hypotheses, no S. zelichiare known to occur outside the ranges of S. cin-namomea and S. palustris, suggesting a causallink between the existence of these three formsand supporting the Hybridization Hypothesisand both Color Morph Hypotheses over theGood Species Hypothesis.

The hypothesis that S. zelichi is a hybrid ofS. palustris × S. cinnamomea is partially rejectedby both song and habitat data. Because thereis no sign of intermediacy in the vocalizationsof S. zelichi and its putative parental formsS. palustris and S. cinnamomea, the mixed-voiceprediction is rejected. However, if songs wereinnate or learned exclusively from male parents,similarities between songs of S. palustris andS. zelichi would be expected even if the latteris a hybrid. If S. zelichi is of hybrid origin andsongs are learned from the father, then evidencesuggests that S. palustris is the paternal form.Habitat overlap between the proposed parentalforms S. cinnamomea and S. palustris is limited,but sufficient to allow hybridization. However,no S. zelichi were found in the typical grasslandhabitat of S. cinnamomea, suggesting that maleS. cinnamomea might not be important in theorigin of S. zelichi. None of the predictionsof this hypothesis was fully supported by myresults. Importantly, S. zelichi occurs regularly(although at low densities) in the wild both inspace and time, unlike hybrids reported in thegenus Sporophila, a fact that undermines the Hy-brid Hypothesis. Continuity in time and wide


Fig. 3. Representative spectrograms of vocalizationsof S. cinnamomea (N = 24). Numbers in parenthesescorrespond with data of individuals in Appendix 1,locality name and locality number as in Figure 1are also included here to facilitate comparison ofsimilitude among localities. (A) Initiation and middleportion of song and call of S. cinnamomea in 2003–2007 depicting delimitation of note types; song (5,Rincon del Socorro-15) and call (5, Rincon delSocorro-15), (B) Initiation and middle portion ofsong in 1991–1993 (above, AJ21, Mercedes-14), andin 2003–2007 (below, 5, Rincon del Socorro-15), (C)Call in 2003–2007 (14, Lorenzo Geyres-24). Inferredhomologies are described by the use of an apostrophe(’), interrogation mark denotes doubt regarding thehomology.

geographic span are two other characteristics ofS. zelichi arguing against its hybrid origin.

The hypothesis that S. zelichi is a color morphof S. cinnamomea is rejected based on both songand habitat data. The songs of S. zelichi andS. cinnamomea differ in note structure, pace, andduration (Figs. 2 and 3, Table 2). In addition,there is little habitat overlap between S. zelichiand S. cinnamomea, and no S. zelichi were foundin the preferred habitat of S. cinnamomea.

The hypothesis that S. zelichi is a color morphof S. palustris is supported by both song andhabitat data. The note repertoire of S. zelichi wassimilar to that of S. palustris, and their songs havechanged in a similar manner over time (Fig. 2).Both S. zelichi and S. palustris were always foundin marshes or wet grasslands and the presence ofS. palustris predicts the presence of S. zelichibetter than any other species.

DISCUSSION

My results indicate that S. zelichi is not adistinct species, and support the hypothesis thatit is a syntopic color morph of S. palustris. Thesongs and habitats of S. zelichi were indistin-guishable from those of S. palustris, and differentfrom those of S. cinnamomea. Moreover, birdtrappers have reported the use of S. palustrisor S. zelichi individuals to attract each otherto traps, showing intense territorial responsebetween these two forms (R. Tato and T. Juani,pers. comm.). Playback experiments (althoughnot quantified) indicate that S. zelichi respondpositively to S. palustris and vice versa (Pearmanand Abadie 1995, unpubl. data). My findingsagree with those of Pearman and Abadie (1995)who found the vocalizations of S. palustris andS. zelichi to be identical. In contrast, Zelich(in Narosky 1977:346) mentioned a friendwho could presumably separate S. zelichi fromS. palustris in the field by their vocalizations.Previous reports indicate that both species werefound in marshes, as is shown in this study(Barrows 1883, Narosky 1977, Pearman andAbadie 1995, Azpiroz 2003). These findings areparalleled by the existence of an undescribedblack-collared and rusty-backed seedeater thatis indistinguishable from S. ruficollis in vocaliza-tions and habitat (Areta, unpubl data).

No valid seedeater species in the genusSporophila is rare over its entire range (J. Areta,pers. obs.), with even less abundant species beinglocally common when conditions are suitable(Ridgely and Tudor 1989, Sick 1997, Stiles2004). Population size estimates for S. palustris(1000–2500 individuals) and S. zelichi (50–250individuals) are reflected by the low relativeabundances of both species (BirdLife 2005).Pearman and Abadie (1995) found a relativeabundance ratio of one S. zelichi for every 14S. palustris similar to the one reported here.The extreme rarity of S. zelichi is not expected


in a valid seedeater species. Although hybridshave been reported in the genus Sporophila,none has been found to occur extensively andcontinuously in nature (Sick 1963, 1997) and,unlike many hybrids, S. zelichi has been reportedbreeding successfully in the wild (Pearman andAbadie 1995, M. Zelich, pers. comm.). How-ever, I cannot rule out the possibility thatS. zelichi is a hybrid between S. palustris and an-other Capuchino (most likely S. cinnamomea).Data on habitat are asymmetrical because someindividuals of S. cinnamomea were found in thepreferred habitat of S. zelichi and S. palustris(marsh or wet grassland), but no S. zelichi orS. palustris were found in the habitat typical ofS. cinnamomea (undulating grasslands).

Although my results do not eliminate the pos-sibility of a hybrid origin for S. zelichi, this wouldrequire that an S. palustris male (explaining thesimilar songs) pair with an S. cinnamomea female(possibly explaining the intermediate plumage).This scenario, although plausible, would alsorequire that (1) female S. cinnamomea possessgenetic information from the color pattern ofits father and that these genes interact withthose of a male S. palustris to generate theplumage of S. zelichi, and (2) S. zelichi malesoccur only in the habitats occupied by thepaternal species and not the maternal species.The hypothesis that S. zelichi is a morph ofS. palustris requires no such assumptions, andexplains the similarities in songs and habitatuse. In support of the Color Morph Hypothesis,Stiles (2004) proposed that Tumaco Seedeaters(S. insulata) are a color morph of S. telascorather than hybrids between Ruddy-breasted(S. minuta) and Chestnut-throated (S. telasco)Seedeaters as previously suggested (Ridgely andTudor 1989). Analysis of mitDNA showed thatS. insulata is more closely related to S. telascothan to S. minuta (de las Casas 2004), eithersupporting the hypothesis that S. insulata is avariant of S. telasco or suggesting hybridizationbetween S. telasco and another species (mostlikely S. minuta).

Sick (1997) showed that Oryzoborus ( =Sporophila sensu Lijtmaer et al. 2004) seed-finches can learn songs, whereas Rusty-collaredSeedeaters (S. collaris) are known to imitatemany bird species (Moschione 1989). Therapid and concordant changes in the songs ofS. palustris and S. zelichi suggest a commonmechanism of restriction in vocal learning of

both forms, and the changes in S. cinnamomeaalso suggest vocal learning in the Capuchinos.It is not known whether Capuchinos have anyinnate pattern selectivity (template) that couldprevent juveniles from learning heterospecificsongs (see Hultsch and Todt 2004). Capuchinoslikely learn songs from their fathers or fromneighbors during their first (or subsequent)breeding seasons (Grant 1999, Kroodsma 2004,Liu and Kroodsma 2006). Further elucidationof learning mechanisms in the Capuchinos willprovide insight into their systematics.

In a recent mitDNA study, Lijtmaer et al.(2004) were unable to identify any diagnosableunit in the ruficollis group and instead found anarray of forms/species as part of a single poly-tomy. However, vocalization and habitat datasuggest a closer relationship (i.e., conspecificity)between S. zelichi and S. palustris than betweeneither of these forms and S. cinnamomea. Thesedifferences also occur in several other species inthe Capuchinos of the ruficollis group (Areta,unpubl. data). Discrepancy between molecularand field data could reflect (1) recent diversifi-cation not yet evidenced in the studied genes,(2) poor molecular resolution due to method-ological inadequacies, or (3) the existence ofa polymorphic species with varying degrees ofisolation among different forms. This thirdprovocative explanation for the discordance be-tween the molecular and the field perspectivessuggests that female Capuchinos in the ruficollisgroup mate freely with all male plumage forms,hence erasing any genetic distinctiveness, andthat diagnostic features of the vocalizations andhabitat preferences of males are cultural traitssuperimposed on the shared genetic basis of apolymorphic species. At least four critical issuesconcerning the systematics of the Capuchinosremain to be resolved: (1) how, when, and fromwhom do males learn their songs, (2) what arethe female preferences for males with differentplumage and songs, (3) what are the genetic anddevelopmental bases of plumage patterns, and(4) assess the degree of genetic relationship anddistinctiveness of the different known forms.

Taxonomic recommendation. Barrows(1883) described Spermophila palustris basedon material collected at Concepcion delUruguay (Entre Rıos Argentina), and thespecies was subsequently transferred to thegenus Sporophila (Hellmayr 1938). My resultsinvalidate S. zelichi as a valid species under


the Recognition Concept (however, the resultswould be the same under the Biological SpeciesConcept and probably the Phylogenetic SpeciesConcept as well; see Eldredge 1995, Haffer1997). One possible course of action wouldbe to acknowledge that the name S. palustris(Barrows 1883) has priority over S. zelichi(Narosky 1977). Because of similarities in songsand habitat use, I recommend that S. zelichibe treated as a synonym of S. palustris. Someprevious taxonomic proposals might havebearings on the taxonomic validity of the nameS. palustris and, therefore, on the suggestedname for S. zelichi. Short (1975) suggested thatS. palustris was a color morph of S. hypoxantha.S. hypoxantha (Cabanis 1851) has priority overS. palustris (Barrows 1883) and, therefore, ifShort (1975) was correct, then S. hypoxanthawould be the appropriate name for bothS. palustris and S. zelichi. However, my resultsdo not support Short’s (1975) proposal becausethe songs and habitats of S. hypoxantha andS. palustris differ (Areta, unpubl. data). A morecautious approach would be to acknowledgethat S. zelichi is not a valid species, but to deferfrom considering it as a synonym of S. palustrisuntil there is irrefutable evidence that it is not ahybrid of S. palustris × S. cinnamomea.

Although S. cinnamomea has been suggestedto be a color morph of S. hypochroma (Short1975, Ridgely and Tudor 1989, Sibley and Mon-roe 1990), this hypothesis has no bearing on thedecision regarding the future use of the nameS. zelichi. Regardless, the binomial Sporophilazelichi should not be used to designate anyanimal taxon.

Conservation implications. Conservationof S. zelichi will require protection of habitatoccupied by the endangered S. palustris ratherthan exclusive protection measures for S. zelichi.The main threat to these forms is afforestationwith Pinus spp. and Eucalyptus spp. (Pearmanand Abadie 1995, BirdLife 2005, J. Areta, pers.obs.) that not only eliminates adequate grasslandhabitats, but also decreases the chances of mi-grating birds of finding their naturally variableand spatially disperse food resources. Althoughovergrazing by cattle is also detrimental forS. palustris/zelichi and other Capuchinos, mod-erate grazing seems to be tolerated (J. Areta, pers.obs.) and might provide a suitable conservationalternative to total replacement of grasslandhabitat by monocultures and wood plantations.

Due to their scarcity, S. zelichi are highlyprized by bird-keepers. However, the chancesof capturing a male S. zelichi are small somost are captured incidentally by bird-keepersattempting to trap S. palustris. More commonCapuchinos are trapped in greater numbers (J.Areta, pers. obs.). The effects of trapping for thecage-bird market on populations are unknown,but are probably high.

ACKNOWLEDGMENTS

I enormously and especially thank J. Noriega for hisfriendship, patience, and continuous support in logisticaland moral matters during the development of my Ph.D.research. M. Zelich’s enthusiasm and ideas are largelyresponsible for the success of this research project. I alsothank Powerdance, Lui, Baiların, Gringuita, Sol, Viole,Enrique, Joaco, Guada y mi Chinita for their help inthe field and for the good rains that we shared. I thankM. Pearman and K. Cockle for comments that greatlyimproved the quality of an early version of the manuscript,R. Restall for hours of productive discussion on seedeaters,Z. Z. Joroscho for producing the map, and K. Roesler, D.Monteleone, A. Jaramillo, C. Fraga, B. Lopez Lanus, T.Narosky, and M. Pearman for sharing their recordingsand knowledge of the Capuchinos. K. Cockle, one oftwo anonymous reviewers, and the editor, G. Ritchison,made very useful comments that greatly improved themanuscript in its final stages. Finalmente le quiero dar lasgracias al Peti, a Javiercho y a Josecho, por compartir lavida sin importar la distancia, be my blood.

LITERATURE CITED

AZPIROZ, A. 2003. Primeros registros del capuchino decollar (Sporophila zelichi) en Uruguay. OrnitologiaNeotropical 14: 117–119.

BARROWS, W. B. 1883. Birds of the lower Uruguay.Bulletin of the Nuttall Ornithological Club 8: 82–94.

BENCKE, G. A. 2004. O caboclinho Sporophila zelichiobservado no Rio Grande do Sul, Brasil. Ararajuba12: 88–89.

BIRDLIFE INTERNATIONAL. 2000. Threatened birds ofthe world. Lynx Edicions and Birdlife Internacional,Barcelona, Spain, and Cambridge, UK.

BIRDLIFE INTERNATIONAL. 2005. Species factsheets:Sporophila zelichi, Sporophila palustris. Available at://www.birdlife.org (18 March 2006).

CABANIS, J. 1851. Museum Heineanum. Verzeichnissder ornithologischen Sammlung des Ober-amtmannFerdinand Heine auf Gut St. Burchard vor Halber-stadt mit kritischen Anmerkungen und Beschreibungder neuen Arten, systematisch bearbeitet. I Singvogel:1–233.

CHAMBERLIN, T. C. 1965. The method of multipleworking hypotheses. Science 148: 754–759.

CLAY, R. P., AND B. FIELD. 2003. First records of Narosky’sSeedeater Sporophila zelichi in Paraguay. Cotinga 23:84–85.


DE LA PENA, M. 1989. Guıa de Aves Argentinas. TomoVI. Editorial Literature of Latin America. (L.O.L.A.).Buenos Aires, Argentina.

DE LAS CASAS, J. C. 2004. Evaluacion del estadotaxonomico del Semillero de Tumaco Sporophila in-sulata (Fringillidae: Emberizinae) utilizando metodosmorfologicos y geneticos. Tesis de pregrado. Univer-sidad Nacional de Colombia, Bogota, Colombia.

ELDREDGE, N. 1995. Species, selection, and Paterson’sconcept of the specific-mate recognition system. In:Speciation and the recognition concept: theory andapplication (D. M. Lambert and H. G. Spencer,eds.), pp. 464–477. Johns Hopkins University Press,Baltimore, MD.

GRANT, P. R. 1999. Ecology and evolution of Dar-win’s finches. Princeton University Press, Princeton,NJ.

GUYRA-PARAGUAY. 2004. Lista comentada de las avesdel Paraguay. Asociacion Guyra Paraguay, Asuncion,Paraguay.

HAFFER, J. 1997. Species concepts and species limits inornithology. In: Handbook of the birds of the world,vol. 4. Sandgrouse to cuckoos (J. del Hoyo, A. Elliott,and J. SARGATAL, eds.), pp. 11–24. Lynx Edicions,Barcelona, Spain.

HELLMAYR, K. E. 1904. Uber neue und wenig bekan-nte Fringilliden Brasiliens, nebst Bemerkungen ubernotwendige Anderungen in der Nomenklatur einigerArten. Verhandl Zoologische und Botanische Wien54: 516–537.

HELLMAYR, K. E. 1938. Catalogue of birds of the Americasand adjacent islands. Part XI. Field Museum ofNatural History Publications 430, Zoology Series,Vol XIII, Chicago, IL.

HULTSCH, H., AND D. TODT. 2004. Learning to sing. In:Nature′s music: the science of birdsong (P. Marler andH. Slabekoorn, eds.), pp. 80–107. Academic Press,San Diego, CA.

KROODSMA, D. 2004. The diversity and plasticity ofbirdsong. In: Nature′s music: the science of birdsong(P. Marler and H. Slabekoorn, eds.), pp. 108–131.Elsevier Academic Press, San Diego, CA.

LIJTMAER, D., N. M. M. SHARPE, P. L. TUBARO, AND S.C. LOUGHEED. 2004. Molecular phylogenetics anddiversification of the genus Sporophila (Aves: Passeri-formes). Molecular Phylogenetics and Evolution 33:562–579.

LIU, W., AND D. E. KROODSMA. 2006. Song learning byChipping Sparrows: when, where, and from whom.Condor 108: 509–517.

MEYER DE SCHAUENSEE, R. 1952. A review of the genusSporophila. Proceedings of the Academy of NaturalSciences of Philadelphia 54: 153–198.

MEYER DE SCHAUENSEE, R. 1966. The species of birdsof South America with their distribution. LivingstonPublishing Company, Narberth, PA.

MOSCHIONE, F. N. 1989. El gran imitador. Observacionessobre la conducta vocal del Domino (Sporophilacollaris). Garganchillo 9: 10–13.

NAROSKY, S. 1977. Una nueva especie del generoSporophila. Hornero 11: 345–348.

NAROSKY, T., AND D. YZURIETA. 1987. Guıa para laidentificacion de las aves de Argentina y Uruguay.Asociacion Ornitologica del Plata/Vazquez Mazzini.Buenos Aires, Argentina.

OUELLET, H. 1992. Speciation, zoogeography and taxo-nomic problems in the Neotropical genus Sporophila(Aves: Emberizinae). Bulletin of the British Ornitho-logical Club Centenary Supplement 112a: 225–235.

PATERSON, H. E. H. 1985. The recognition conceptof species. In: Species and speciation (E. S. Vrba,ed.), pp. 21–29. Transvaal Museum Monographs 4,Pretoria, South Africa.

PAYNTER, R. A., JR., AND R. W. STORER. 1970. Checklistof the birds of the world, vol. XIII. CambridgeMuseum of Comparative Zoology, Cambridge, MA.

PEARMAN, M., AND E. I. ABADIE. 1995. Mesopotamiagrasslands and wetlands survey, 1991–1993: con-servation of threatened birds and habitat in north-east Argentina. Final Report. Available on-line at:http://www.grupofalco.com.ar (16 May 2008).

REMSEN, J. V., JR., A. JARAMILLO, M. NORES, J. F.PACHECO, M. B. ROBBINS, T. S. SCHULENBERG, F.G. STILES, J. M. C. DA SILVA, D. F. STOTZ, AND K. J.ZIMMER. 2006. Version 20–3-2006. A classificationof the bird species of South America. AmericanOrnithologists’ Union. Available at: http://www.museum.lsu.edu/∼Remsen/SACCBaseline.html (20March 2006).

RIDGELY, R. S., AND G. TUDOR. 1989. The birds of SouthAmerica, vol. 1. University of Texas Press, Austin,TX.

SCLATER, P. L. 1871. A revision of the species of thefringilline genus Spermophila. Ibis 1: 1–23.

SHORT, L. L. 1969. Relationships among some SouthAmerican seedeaters (Sporophila), with a record of S.hypochroma for Argentina. Wilson Bulletin 81: 216–219.

SHORT, L. L. 1975. A zoogeographic analysis of the SouthAmerican Chaco avifauna. Bulletin of the AmericanMuseum of Natural History 154: 165–352.

SIBLEY, C. G., AND B. L. MONROE, JR. 1990. Distribu-tion and taxonomy of the birds of the world. YaleUniversity Press, New Haven, CT.

SICK, H. 1963. Hybridization in certain Brazilian fringill-idae (Sporophila and Oryzoborus). Proceedings ofthe XIIIth International Ornithological Congress,Ithaca: 161–170.

SICK, H. 1997. Ornitologıa Brasileira. Ed. Nova Fronteira,Rıo de Janeiro, Brasil.

SILVA, J. M. C. 1999. Seasonal movements and conserva-tion of seedeaters of the genus Sporophila in SouthAmerica. Studies in Avian Biology 19: 272–280.

SLABEKOORN, H., AND T. B. SMITH. 2002. Bird song,ecology and speciation. Philosophical Transactionsof the Royal Society of London B 357: 493–503.

STILES, G. 2004. The Tumaco Seedeater (Sporophilainsulata, Emberizidae): a species that never was?Ornitologia Neotropical 15: 17–30.

VRBA, E. 1995. Species as habitat-specific complex sys-tems. In: Speciation and the recognition concept:theory and application (D. M. Lambert and H. G.Spencer, eds.), pp. 3–44. Johns Hopkins UniversityPress, Baltimore, MD.

APPENDIX 1

Recording localities, dates, and source ofrecordings used in this paper. Data of localities


in Figure 1. Numbers between square brack-ets indicate number of individuals per localityand identification number of individual birdsassigned in my database is between paren-theses. All recordings by the author exceptthose noted by AJ = Alvaro Jaramillo, DM =Diego Monteleone, MP = Mark Pearman, andRF = Rosendo Fraga.

S. palustris [34]. ARGENTINA. 1991–1993:Mercedes [1] (AJ30), Gualeguaychu [2] (MP32,MP34), and Banco Caraballo–Entre Rıos [1](MP33). 2003–2007: Ea. La Marita [2] (1,2),Cnia. Pellegrini [2] (3,4), Ea. Rincon delSocorro [11] (5–14, DM31), Ea. Santa Isabel–Corrientes [1] (RF28), and Banado Santa Rosa–Corrientes [1] (RF29). URUGUAY. 2003–2007: Banados de la India Muerta [6] (15–20),and Cebollati [7] (21–27)

S. zelichi [3]. ARGENTINA. 1992:Gualeguaychu [1] (MP3). 2003–2007:Gualeguaychu [1] (1), and Ea. El Socorro [1](2).

S. cinnamomea [24]. ARGENTINA. 1992–1993: Mercedes [2] (AJ21, AJ22), PN El Pal-mar [1] (MP23), and Caza Pava – Corrientes

[1] (MP24). 2003–2007: Ea. El Socorro [9](1–5, 7–10), Mercedes [1] (1), Cnia. Pelle-grini [2] (11,12), and Gualeguaychu [1] (13).URUGUAY. 2003–2007: Lorenzo Geyres [5](14–18), San Javier – Rio Negro [1] (RF19),and Villa Soriano – Rio Negro [1] (RF20).

APPENDIX 2

Locations where I obtained informationabout habitat use (data for locations in Fig. 1).Numbers between square brackets indicatenumber of individuals per location.

S. palustris [46]. ARGENTINA. Cnia. Pelle-grini [7], Ea. El Socorro [15], Gualeguaychu [1],and Ea. La Marita [3]. URUGUAY. Cebollati[11] and Banados de la India Muerta [9].

S. zelichi [12]. ARGENTINA. Camba Trapo[1], Cnia. Pellegrini [2], Ea. El Socorro [1],Gualeguaychu [7], and Ibicuy [1].

S. cinnamomea [34]. ARGENTINA. Cnia.Pellegrini [7], Ea. El Socorro [9], Mercedes[1], PN El Palmar [8], Arroyo Baru [1], andGualeguaychu [1]. URUGUAY. Lorenzo Geyres[7] and Queguayar [2].

journal of field ornithology · s. cinnamomea), morfo de color i (morfo de s. cinnamomea) y morfo...

Documents