Lepidocolaptes angustirostris ( Vieillot, 1818 )

BOLÍVAR-LEGUIZAMÓN, SERGIO & SILVEIRA, LUÍS FÁBIO, 2015, MORPHOLOGICAL VARIATION AND TAXONOMY OF LEPIDOCOLAPTES ANGUSTIROSTRIS (VIEILLOT, 1818) (PASSERIFORMES: DENDROCOLAPTIDAE), Papéis Avulsos de Zoologia (São Paulo) 55 (21), pp. 281-316 : 289-295

Lepidocolaptes angustirostris ( Vieillot, 1818)

Trepador Del Comun , Azara, 1802, Apuntamientos para la Historia Natural de los Páxaros del Paragüay y del Rio de la Plata, Tomo II, p. 279. Type no existent .

Dendrocopus angustirostris Vieillot, 1818 , Nouveau Dictionnaire D’historie Naturelle, Appliquée Aux Arts. XXVI, p. 116, locality: “ Paraguay ”, based in Azara (1802).

Dendrocolaptes bivittatus Lichtenstein, 1822 , Abhandlungen der physikalischen (-mathematischen) Klasse der Koeniglich-Preussischen Akademie der Wissenschaften, (1820-1821), p. 258-266, pl. 2, fig. 2, locality: “in province São Paulo ”. Spix, 1824, Avium species novae, quas Brasíliam, I, p. 87, locality: Piauí, Brazil. Lafresnaye & D’Orbigny, 1838, Magasin Zoologie, p. 8, locality: Corrientes.

Picolaptescoronatus Lesson, 1830, Traited’Ornithologie, Livr. 4, p. 314, locality: Piauí, Brazil.

Dendrocolaptes rufus Wied, 1830 , Beiträge zur Naturgeschichte von Brasilien, 3, p. 1130, locality: between provinces Minas and Bahia, Brazil.

Picolates bivittatus Lafresnaye, 1850 , Revue et Magasin de Zoologie pure et Appliquée, 2, p. 152, locality: São Paulo. Allen, 1876, Bulletin of the Essex Institute, 8, p. 80, locality: Santarém, Brazil.

Picolaptes angustirostris Lafresnaye, 1850 , Revue et Magasin de Zoologie pure et Appliquée, 2, p. 151, locality: “ Paraguay ”. Sclater, 1890, Catalogue of Birds in the British Museum, 15, p. 155, locality: “ Bolivia ”. Salvadori, 1897, Bollettino dei Musei di Zoologia e Anatomia Comparata della Reale Universitá di Torino, 12, N. 292, p. 21, locality: San Francisco, Caiza, province Tarija, Bolivia.

Lepidocolaptes atripes Hudson, 1870 , Proceedings of the Zoological Society of London, p. 113, locality: Concepción del Uruguay.

Picolaptes bivittatus bahiae Hellmayr, 1903 , Verhandlungen der kaiserlichkönigiglichen zoologisch-botanischen Gesellschaft in Wien, 53, p. 219, locality: Joazeiro, Bahia, Brazil.

Picolaptes angustirostris bivittatus Hellmayr, 1908 , Novitates Zoologicae, 15, p. 65, locality: “Goyaz”, Rio Araguaya, Rio Thesouras, Faz. Esperança (“Goyaz” = Goiás, Brazil).

Picolaptes angustirostris angustirostris Dabbene, 1910 , Anales del Museo Nacional de Buenos Aires, 18, p. 307, locality: Catamarca, Entre Ríos, Córdoba, Jujuy, Mendoza, Tucumán, Salta, Chaco; idem, l.c. 23, p. 318, 1912 – San Rafael, Uruguay.

Lepidocolaptes angustirostris certhiolus Todd, 1913 , Proceedings of the Biological Society of Washington, 26, p. 173, locality: Curiche, Rio Grande, Bolivia.

Lepidocolaptes angustirostris praedatus Cherrie, 1916 , Bulletin of the American Museum of Natural History, 35, p. 187, locality: Entre Ríos, Concepción del Uruguay (Probably Argentina / Uruguay limits).

Lepidocolaptes angustirostris hellmayri Naumburg, 1925 , Auk, 42, p. 421, locality: Chilón, Santa Cruz, Bolivia.

Lepidocolaptes angustirostris bahiae Cory & Hellmayr, 1925 , Catalogue of Birds of the Americas and The Adjacent Islands in Field Museum of Natural History, vol. IV. p. 339.

Lepidocolaptes angustirostris bivittatus Wetmore, 1926 , US National Museum Bolletin, 133, p. 235.

Lepidocolaptes angustirostris coronatus Wetmore, 1926 , US National Museum Bolletin, 133, p. 235.

Lepidocolaptes angustirostris immaculatus Carriker, 1935 , Proc. Acad. Nat. Sciences Philad. 87, p. 328, locality: Bolivia, Beni, Charatona.

Lepidocolaptes angustirostris chacoensis Laubmann, 1935 , Verh. Ornith. Gesel. Bayern 20(4), p. 336, locality: Argentina, Formosa, San José and Lapango.

Lepidocolaptes angustirostris dabbenei Esteban, 1948 , Acta Zool. Lilloana 5, p. 384, Argentina, Tucumán, Los Goméz.

Lepidocolaptes angustirostris griseiceps Mees, 1974 , Zool. Mededelingen Rijksmus. Nat. Hist. Leiden, 48, p. 57, locality: Sipaliwini, Suriname.

Distribution: From extreme northeastern Brazil to southern-central regions at Argentina, including isolated populations at the Sipaliwini savanna and the Amazonian regions of east Amapá and Pará. Inhabits mainly in the open/dry lands of Caatinga, Cerrado, and El Gran Chaco ecoregions, and adjacent regions (Pantanal, Chiquitano Dry Forest, Espinal, Paraná Flooded Savanna, Southern Cone Mesopotamian Savanna, Alto Paraná Atlantic Forests ecoregions).

Diagnosis: Unlike other Lepidocolaptes from South America, the white superciliaries are broad, the bill is more pinkish than the other Lepidocolaptes , head and neck blackish with numerous pale fulvous shaft-spots, darker than others species.

Description: A woodcreeper of medium size (19-22 cm). A long, slim and moderately decurved pale grey to pinkish-horn bill; base of upper mandible with dusky sides. Brown to Brown-olivaceous above; superciliaries broad and white; head and neck blackish with numerous pale fulvous shaft-spots; tail ferruginous; underparts from unstreaked cinnamon-ochraceous to dark-brown streaked, from the north to south of the distribution; legs and feets grey to dark grey. No sexual dimorphism in the plumage. Despite of high intergradation and geographical variation, two main morphs can be found: an unstreaked group found from the northeastern to southern Brazil (“bivitattus” group), including isolated populations in the north Amazonas and in the savannas of Sipaliwini; and a streaked morph, found from the southern Brazil to north-central Argentina, including populations from the eastern Bolivia ( “angustirostris ” group).

Intraspecific variation: The populations of L. angustirostris complex are highly polymorphic. The phenotypic variation has allowed the classification of this taxon in the past into several subspecies by many authors. However, the undefined geographical boundaries and the intermediate stages of morphological and plumage characters do not support this subspecific division. The plumage patterns are the most variable traits in the species and probably this is the most variable woodcreeper. The size varies, but this morphological variation was not unidirectional. The largest populations can be found in the northern and southern regions (Caatinga and southern Chacoan ecoregions), while smaller individuals were identified between the southern Cerrado -Pantanal to the northern Chacoan ecoregions (see Fig. 9). The size decreases gradually from the northeastern to southern Brazil, where we found the smaller individuals.

Delimitation of the species

The main objective of this work was to review the taxonomic status of Narrow-billed Woodcreeper based on plumage and morphometric characters. Our results showed any significant differences among the named populations, and intergradation and a possible latitudinal geographic variation were found. Statistical analyzes were also inconclusive; PCA tests found significant differences among groups studied, but is not congruent with the level of intergradation observed. So, how to explain the high colour-polymorphism found in the populations of Lepidocolaptes angustirostris through South American open/dry lands? Is there any convenience in splitting contiguous colour-polymorphic populations from a unique recognized species into several valid taxonomic units?

The delimitation of species and subspecies is a topic highly debated, and a number of theoretical and methodological approaches have been proposed ( Amadon, 1949; Zink, 2006; Alström et al., 2008; Cicero, 2010; Marantz & Patten, 2010; Tobias et al., 2010; Zapata & Jiménez, 2012; Camargo & Sites, 2013). However, there is still no consensus about concepts and methods applied to different taxa ( Agapow et al., 2004; Queiroz, 2007; Gill, 2014; Sangster, 2014). In this study, using mainly the diagnosability criteria the taxonomic validity of subspecies of L. angustirostris proposed in the literature is rejected due to poor definition of plumage, morphological, and geographical boundaries among them. The Narrow-billed Woodcreeper is composed by polymorphic populations intergrading from the northeast of Brazil to south of Chacoan ecoregions (Dry and Humid Chaco), through the open/dry lands of Caatinga, Cerrado, and the Gran Chaco. So, division of this taxon could not be applied. Cardoso et al. (2003) state that species like as L. angustirostris inhabiting climatic gradients show morphological variation in a clinal trend. This variation could be difficult in taxonomy, as no clear morphological entities can be delineated. It is common for taxonomists to split the clines into distinct subspecies, but in this case this is purely arbitrary and artificial.

Several authors have criticized the concept and application of subspecies rank in taxonomic works ( Zink, 2004; Alström et al., 2008, among others). For Fitzpatrick (2010), subspecies, unlike species are human constructs, while Zink (2004) concluded that the use of subspecific rank in taxonomy is not useful and adds uncertainty to biological classifications. The absence of defined criteria in the delimitation of these biological groups is one of obstacles when are implemented. At the other hand, some works have shown the convenience of subspecific status in study of certain taxa ( Patten & Unitt, 2002; Cicero, 2010; Patten, 2010). For example, Haig & Winker (2010) states that the study of subspecies address the geographic component of variation and differentiation, if the criteria used to define the subespecific groups each case were made explicit.

A plausible hypothesis to explain the high phenotypic diversity in L. angustirostris is an existence of an incipient speciation in this taxon. In an incipient speciation, two or more populations from one species are being splitted into two new ones, but are still capable of interbreeding. These ‘early stages’ of divergence among populations of the same species has been analyzed in some groups of birds ( Dendroica coronata, Brelsford & Irwin, 2009 ) and other taxa (the Arctic Charr Salvelinus alpinus, Adams & Huntingford, 2004 ). Indicators of incipient speciation can be the presence of low genetic divergence among populations and morphological polymorphism ( Price, 2008). For instance, in an analyses testing adaptive radiation of the capuchino seedeaters (11 species from genus Sporophila ), Campagna et al. (2011) found high phenotypic variation in vocalizations and coloration, but extremely low levels of neutral genetic differentiation, proposing a recent speciation (middle Pleistocene) occurring in the group.

Some molecular evidence could support this hypothesis in the Narrow-billed Woodcreeper; in a phylogenetic analysis of genus Lepidocolaptes, Arbelaéz-Cortés et al. (2012) found low genetic differences between both COI and cyt b haplotypes for L. angustirostris samples from Brazil, Bolivia and Argentina. Here, 10 individuals from six localities were analyzed (COI haplotype). According to these authors, four subspecies (L. a. bahiae, L. a. hellmayri or L. a. certhiolus, and L. a. praedatus) could be sampled, and their observed genetic variation was low, suggesting “a recent range expansion”. The low genetic differentiation among populations analyzed and the high color polymorphism in the individuals sampled in this study may support the hypothesis of an eventual early speciation occurring in the populations of the Narrow-billed Woodcreeper.

One of the most debated points in taxonomy is the dichotomy “splitting vs lumping” of taxa with intraspecific variation and uncertain taxonomic position; the need to describe all the diversity present in biological groups faces the absence of strong and verifiable evidence about the limits of these divisions. In cases like as L. angustirostris the situation is similar: the phenotypic variation exists, but evidence was not found that allow the delimitation of the populations identified as diagnosable units. Additionally, geographic barriers (frequently used in ornithology to delimit subspecific lineages) in the distribution of Narrow-billed Woodcreeper are absent or seem not to affect the gene flow among the populations. For these reasons, contrary to the cited literature, the subspecific rank was dismissed and only one valid taxonomic unit is proposed here. Literature about L. angustirostris used the subspecific rank to describe the differences in the ventral/dorsal plumage (see Marantz et al., 2003; Ridgely & Tudor, 2009). Undertail coverts coloration, plumage pattern of crown, and tone of color in dorsal and ventral sides has been used to sort this taxon in at least eight subspecies recognized. Despite finding these variations, their distribution along the sampled individuals cannot identify clear patterns of differentiation among populations. A possible cause is the scarce comparative analysis of additional samples from intermediate regions of the geographical distribution of the taxon, in a more comprehensive analysis of the total variation before describing new taxa. Early descriptions were based on very restricted samples, putting aside other samples from the same location or adjacent areas. In this study, more than one ventral plumage pattern was identified in the same localities (e.g., localities from the central Brazil [Goiás], and in the boundaries of southern Brazil and Paraguay), rejecting a possible delimitation among populations. In a recent study about morphological variation in Schistochlamys ruficapillus ( Lopes & Gonzaga, 2014) , a similar conclusion was proposed; the recognition of the three subspecies in the Cinnamon Tanager were based individuals from distant populations (scarce sampling), and intermediate stages were not included in these descriptions.

Plumage patterns and polymorphism in L. angustirostris

The plumage patterns in L. angustirostris complex are highly diverse, showing three dorsal and five ventral patterns along its distribution, covering the eight subspecies currently recognized. Intergradation and intermediate colorations were found among our vast sample in this study. Despite of high divergence between the populations from the northeast Brazil and the northern Argentina, defined boundaries among plumage patterns were not possible to identify. Given the existence of a high color polymorphism in the taxon, we reviewed some causes that may influence this plumage variation.

The color polymorphism is defined as the presence of two or more distinct, genetically determined color morphs within a single interbreeding population. This different color exemplifies extreme morphological diversity within populations ( Huxley, 1955; Gray & McKinnon, 2007). Has been theorized that, in color-polymorphic species with large geographic ranges (similar to L. angustirostris ), there is probability to occur parapatric speciation at the ends of a ratio cline in morph frequencies ( Endler, 1977). The color polymorphism has been associated with differences in groups of correlated traits (behavior, life history, morphology, physiology etc.) due to correlational or epistatic selection or shared developmental pathways ( Forsman et al., 2008). Other situations that could predispose color polymorphism occurs when populations come into secondary contact after diverging in coloration allopatrically or when color forms are under disruptive selection associated with different microhabitats ( Endler, 1977; Roulin, 2004). Also, in some circumstances, the color polymorphism may represent incomplete speciation ( Gray & McKinnon, 2007; Hugall & Stuart-Fox, 2012).

Galeotti et al. (2003) developed a study about color-polymorphism in birds in order to analyze some biological mechanisms proposed to explain the maintenance of polymorphism. The authors tested three forms of selection: apostatic, disruptive, and sexual selection, plus a no selection model. In addition to establishing that the polymorphism is a relatively rare phenomenon (only 3.5% of bird species show polymorphism), one of the conclusions proposed was that the color polymorphism in birds is not a nonadaptive consequence of selection on other adaptive traits, but a trait that evolved probably by disruptive selection. In this disruptive selection hypothesis, the patterns of variation in light conditions may be the most important selective mechanism maintaining color polymorphism in birds. Hugall & Stuart-Fox (2012) developed a study about speciation in color-polymorphic birds using genetic data from five families of non-passerine taxa. The authors concluded that the color polymorphism tends to be associated with diverse ecological conditions or relatively recent speciation, being this statement applied to the passerines taxa. For Roulin (2004), the presence of color morphs in a species have an adaptive value, namely, the different attributes of morphs could be correlated to environmental variations. This hypothesis predicts covariation with life history, behavioral, morphological and physiological traits.

For the Narrow-billed Woodcreeper the color polymorphism precludes the splitting of this taxon into other valid taxonomic units due to the high level of intergradation and lack of defined boundaries among populations. Second, environmental factors were correlated to the ventral plumage in the species. Based in the GLM analyses, temperature seasonality (BIOCLIM 4) seems to explain the ventral plumage variation using the two main states identified in the groups: unstreaked/Streaked patterns. Additionally, the Gloger’s rule (darker plumages are more expected in more humid environments, and lighter plumages at dry regions) fit with the variation observed. Third, if the statements of Arbelaéz-Cortés et al. (2012) are considered and added to the findings of this work, the color polymorphism could be a result of a recent speciation of populations in the open/dry lands of Caatinga, Cerrado and Chaco ecoregions.

Environmental correlations and biogeographical considerations

The Narrow-billed Woodcreeper is a widely distributed species that inhabits open/dry lands of South America, and its populations are subject to diverse environmental factors, which can have had an effect on the evolution of genetic/phenotypic traits in the species. The study of these variations in the populations of widespread taxa has allowed proposing ‘ecogeographic’ theories describing the correlation between morphological variation and environmental variation. Among them, the rules of Bergmann’s ( Bergmann, 1847), Allen’s ( Allen, 1877), Gloger’s ( Gloger, 1833), and the ‘Neo-Bergmannian’ rules (a re-interpretation by James, 1970) explain the variation in the phenotypes of the populations. The ecogeographic rule most tested in analyses of geographical variation is the Bergmann’s rule, that states in its original version that warm blooded vertebrate species from cooler climates tend to be larger than congeners from warmer climates (see Meiri & Dayan, 2003). Subsequently, the modification of James (1970) proposed that the intraspecific variation in size is related to a combination of climatic variables that includes temperature and humidity, i.e., the small size is associated with hot and humid conditions, larger size with cooler or drier conditions. The other two ecogeographic rules are less used, but are of equal importance when testing variation. In the Allen’s rule, the individuals in hot climates should have longer appendages relative to body core size in order to dissipate heat more efficiently. While that in the Gloger’s rule, defined as the expectation that plumages of birds are darker in more humid environments. It is important note that other selection forces (dense vegetation, interspecific competition, diet, among others factors) might also operate for slight variations in size at the population or species level (see Hamilton, 1961).

In works on geographical variation in Aves, several hypotheses have been proposed and tested with empirical methodologies. In an early review of the adaptive significances of the intra-specific variation in continental birds, ( Hamilton, 1961) concluded that the variation in wing length and body size in birds exist, and is correlated to gradient factors of the environment. Also, and maybe important for L. angustirostris , the clinal variation of each of these morphological traits could not be concordant. In the same way, James (1983) stated that phenotypic variation in the Red-winged Blackbird contains an important environmental component. In populations of birds, regional trends of size variation change gradually in a way that may reflect topographic features ( James, 1970). In a review by Meiri & Dayan (2003), the authors found that the presence of the Bergmann’s rule is more common in sedentary than in migratory species, and concluded that this ecogeographic rule is “a valid ecological generalization for birds and mammals at the class, order, and family levels”. In a study about Red-winged Blackbirds, Power (1969) concluded that size increases in arid regions may facilitate conservation of metabolic water and size decreases in humid regions may facilitate heat dissipation. In populations of Turdus migratorius , the morphological variation and plumage is concordant with the predicted by the Bergmann’s and Gloger’s rules, but the relationships among the explanatory variables were not well elucidated ( Aldrich & James, 1991).

In this work, the GLM analyses were conducted to establish a correlation between the geographical variation found in the Narrow-billed Woodcreeper and the climatic information gathered from each geographic record sampled. In these analyses, the climatic factor most explanatory of the geographical variation was the temperature seasonality (BIOCLIM 4) and the Precipitation of Warmest Quarter (BIOCLIM 18). In all models identified, these three variables appear to be the most correlated to the traits analyzed (PC1, PC2, size, and ventral plumage pattern).

For PC 1 (bill length, exposed culmen, and total culmen), a positive correlation was found with the latitude and the temperature seasonality. Namely, the size of components from PC1 increases as the temperature seasonality and latitude increases to south (from the Equator line to south). For PC 2, the bill width is positively correlated to the increase to south of latitude. The same type of positive correlation is recovered in models with temperature seasonality as explanatory variable. With the increase of this climatic variable, the bill width increases too. The correlation between the size and the climatic variables is not clear. The variation of size has two variation tendencies, a decreasing cline from the northeast to the central Brazil, and an increasing cline from this last zone to the southernmost region in central Argentina. The only climatic variable fitting in the size variation was the precipitation of warmest quarter. It is possible that the low levels of precipitation in the Caatinga and Chaco ecoregions (extreme regions of the distribution), additional to the high temperature in certain time of year can influence the size of individuals .

The seasonality of the temperature was the most correlated variable to the ventral plumage variation. In L. angustirostris the darker and streaked plumages were found in the southern Cerrado and Humid Chaco ecoregions, while the unstreaked patterns (cinnamon-ochraceous, pale yellow, Greyish-white pattern) inhabits northern Cerrado and Caatinga. Here, the Gloger’ rule could be applied, where the darker plumage are found in more humid environments (Humid Chaco, southern Cerrado ), and the lighter inhabits dry regions (Caatinga) .

Overall, the variation in the Narrow-billed Woodcreeper appears to follow the ecogeographic rules of Bergmann and the Gloger. First, individuals with larger bills are present at higher latitudes to south (El Gran Chaco ecoregions), while groups with small- er bills can be found near to Equator, and populations with darker plumages can be found in the south of the distribution with an high level of humidity, and the ligther/unstreaked populations at the central and north of Brazil (dry regions). Also, the Allen’s rule (individuals in hot climates should have longer appendages relative to body core size than individuals in cold environments) could be applied to the populations of L. angustirostris ; the tarsus-metatarsus length increases with the increase of latitude to south. However, this correlation is not clear (see Fig. 8).

For Werneck (2011) and Werneck et al. (2011), the biogeographical patterns of the open/dry lands in South America are the result of a correlation of geological processes that occurred during the Tertiary and Quaternary ages. Periodical glaciations and interglaciations affected the geographical extension of forest and savanna biomes. In the most accepted hypothesis, glaciation periods were characterized by cool and dry climates, with a reduction of Amazonian and Atlantic forests and the increase of open, drier biomes. During the interglacial periods (wet and hot climates), the savanna area was reduced and isolated refugia emerged. In this scenario, the geographical range of the savanna species was reduced to these isolated refugia (a similar situation to the Amazonian refugia proposed by Haffer, 1969), with a genetic and morphological differentiation among their populations. Campagna et al. (2011) proposed that the fluctuation in predominance of rainforest over open habitats and vice versa and the interdigitation of these two biomes could have contributed to isolating small populations in islands of suitable habitat, or a “grassland refugia”.

Using the grassland refugia hypothesis, the partial molecular data from Arbelaéz-Cortés et al., 2012 (low genetic divergence among the sampled individuals of the Narrow-billed Woodcreeper) combined with our results point to a plausible scenario which propose that the high variation in plumage and low genetic divergence among the populations of L. angustirostris could be influenced indirectly by the climatic variations during the Pleistocene. However, additional biogeographical analyses should be performed to confirm this hypothesis.

Isolated populations of Lepidocolaptes angustirostris were found in the Guianan Savanna, and the Uatuma-Trombetas Moist Forests/Tapajós-Xingú Moist Forests ecoregions. These populations were attribute to the griseiceps subspecies ( Mees, 1974; Marantz et al., 2003). Individuals from these groups show a dorsal plumage Strong brown and a greyish-white and pale yellow colorations in the ventral plumage, characteristic of the “bivittatus ” group. Morphometric data show no divergence with the other populations from Caatinga and central Cerrado. Lepidocolaptes angustirostis griseiceps is isolated, but their phenotypic characters show no signs of divergence from other populations. Mees (1974) proposes as diagnostic characters of griseiceps a brownish grey crown, lighter that that found in the adjacent groups (Caatinga-Cerrado populations), a distinctiveness not supported by our results.

The presence of an isolated population of L. angustirostris in Amazonia can be also explained by the Grassland refugia hypothesis ( Campagna et al., 2011, but see also Haffer, 1969). With the climatic variations of Pleistocene and Holocene, the forest regions of South America reduced its extension (glacial periods), and the grasslands regions were predominant (open areas), including the northern savannas of Los Llanos and Guianan savanna (see Werneck et al., 2011). In these glacial periods, populations of the Narrow-billed were able to expand through all these open areas, and, posteriorly, in the interglacial periods, forests were recovering their extension in these warmer and humid periods, and the northernmost populations were isolated from the other continuous populations. Another isolated population, with a few specimens collected, inhabits localities around the mouth of Tapajós river (e.g., Boca do Rio Tapajós, MZUSP 14675), south of Amazonas, and to the north of this river at Monte Alegre (MPEG 4732 and 54361), as a testimony of this old corridor of open vegetation.