Morphology and vocalization support specific status of the Chestnut-headed Chachalaca, Ortalis motmot ruficeps (Wagler, 1830) (Aves; Galliformes; Cracidae)

The Little Chachalaca (Ortalis motmot) is a widely distributed species in the Amazon basin, typically found in riverine habitats. There are two disjunct populations: the northern O. m. motmot and the southern O. m. ruficeps (known as Chestnutheaded Chachalaca). Here we performed a vocal, morphological and plumage comparison between these two taxa. Birds present differences in tail coloration but are otherwise undiagnosable in terms of plumage. Ortalis m. ruficeps is, however, markedly smaller and lighter than O. m. motmot, with no overlap in size or weight. We also found vocal differences between O. m. motmot and O. m. ruficeps, no overlap in geographic distributions and no signs of hybridization across its range. Based on the available data, Ortalis m. ruficeps thus must be considered a valid species, endemic to Brazil.


INTRODUCTION
Despite its large size, conspicuousness, ecologic, social and economic importance, the cracids comprise a group of Galliformes of largely neglected taxonomy. The definition of many taxa within this family is presently supported by early 19 th century analysis of few specimens and poorly analyzed populations. Despite the relatively large amount of available material in museums, just a handful of studies were performed based on good series of specimens thoroughly covering the species distribution. Unsurprisingly, recent studies using different sets of analytic tools (depending on the available material) have been demonstrating the need of revision of polytypic taxa, in order to discriminate true evolutionary entities from taxonomic rubbish, with the consequent uncovering of valid taxa of immediate conservation concern (e.g., Grau et al., 2003;Silveira et al., 2004;Costa et al., 2017;Evangelista-Vargas et al., 2017;Silveira et al., 2017;Evangelista-Vargas & Silveira, 2018). For example, the recognition of O. remota as a critically endangered valid species  promptly ignited a series of field studies directed to the conservation of the few remaining individuals.
The subordination of O. ruficeps to O. motmot was first proposed by Todd (1932), who analyzed a series of specimens from the lower Amazon. This author mentioned the smaller size and differences in head and tail coloration of O. ruficeps but suggested that it only represented a race of O. motmot.  agreed to this treatment and all subsequent authors followed suit, without presenting additional analyses Pinto, 1964;Vaurie, 1965Vaurie, , 1968Delacour & Amadon, 1973;Pinto, 1978;del Hoyo et al., 1994;. These studies maintained a similar description and diversely pointed out to various diagnostic features of O. ruficeps in comparison with O. motmot: the smaller size Pinto, 1964;Delacour & Amadon, 1973;; a lighter and/or brighter coloration of the crown in O. ruficeps Pinto, 1964); a redder coloration of the crown ; and a darker overall coloration (Pinto, 1964;Delacour & Amadon, 1973;. Differences in the color of the external rectrices were also raised , although Vaurie (1965) argued that this was a more subtle character. Sick (1985Sick ( , 1997 was the only author who consistently kept the species rank for O. ruficeps. Recent studies have shown that a closer examination of large cracids series can reveal several valid species within the subspecies of polytypic complexes (e.g., Ortalis guttata), in some instances with dramatic implications for conservation . Therefore, in the present study, we conducted a thorough examination of the available museum skins of O. m. motmot and O. m. ruficeps, together with high quality photographs and vocalizations available in online repositories, in order to provide a taxonomic revision of these two supposed subspecies, with updated diagnoses and geographic distribution. We also obtained photographic records from the WikiAves platform (www.wikiaves.com.br) to comple-ment the distribution map (31 localities). Finally, we gathered 40 voice recordings from Xeno-Canto (www. xeno-canto.org) and WikiAves, from which 18 were suitable for spectrographic analyses.

We
We compared the birds in terms of plumage coloration, morphology and voice. For plumage coloration we analyzed the color of the crown, forehead, throat, chest, abdomen, flanks, crissum, mantle, wing coverts, primaries, tail coverts, external and central rectrices. For morphology we measured the exposed culmen, bill width and depth, and tarsus (nearest 0.01 mm) as well as the wing and tail length (nearest 0.1 cm). Finally, for the vocal analysis we measured the total syllable duration and the duration and spacing of each note, using the Software Raven Pro 1.5 (Bioacoustics Research Program, 2014).
Statistical analyses were conducted in R version 3.4.3 (R Core Team, 2017) for morphological and voice characters. We first tested separately whether each of the measurements and vocal characters were explained by taxa, sex and the interaction between taxa and sex (only taxa in the case of the vocal characters). Model selection was performed via stepwise backwards selection, dropping non-significant terms with the critical p-value corrected for multiple testing (Bonferroni correction = 0.05/number of tests) in each step. Next, we conducted a Principal Component Analysis on the standardized measurement values (mean centered at 0, standard error at 1) of the group of morphometry and voice characters. Missing values were replaced by zeros after the standardization (thus replaced by the mean value). Principal components 1 and 2 explained 81% of the variation for morphometrics and 62% for voice characters and were retained for subsequent analyses. Lastly, we conducted two analyses: a multivariate analysis of variance (MANOVA) with both PC1 and PC2 as response variables and two separate regressions for PC1 and PC2 separately (similar to the first analysis), once more with the critical p-value corrected for multiple testing.

RESULTS AND DISCUSSION
As previously reported in the literature (Wagler, 1832;Ogilvie-Grant, 1893Vaurie, 1965;, O. m. motmot and O. m. ruficeps differ from each other in the coloration of the rectrices, but are identical in other plumage characters, with little intraspecific variation. The border of the external rectrices are tinged with olive dark brown in O. m. ruficeps, while they are reddish brown in O. m. motmot (Fig. 1).
We did not observe any of the other differences in plumage reported in earlier studies, even after analyzing a large number of specimens. We did observe a lighter crown color in a few specimens of O. m. ruficeps (MZUSP 17054, 22045, 22046, 22818, 22819, 46256), as previously reported (Todd, 1932;Pinto, 1964), but in all other specimens the coloration was identical to O. m. motmot. Moreover, we did not observe differences in darkness or redness in other plumage characters (contra Pinto, 1964;Delacour & Amadon, 1973;. On the other hand, and confirming what was reported in the literature since its original description (Wagler, 1830;Todd, 1932;Pinto, 1964;Vaurie, 1965;Delacour & Amadon, 1973;, O. m. ruficeps is smaller than O. m. motmot in almost every character analyzed and the differences are maintained even when sexes differ in size. Furthermore, there is no overlap in size range between the two taxa. When compared to O. m. motmot, O. m. ruficeps has a smaller culmen (10%; percentage difference between brackets), bill width (13%), tarsus (16%), wing (12%), and tail length (16%; see Appendix 1). Ortalis m. ruficeps is also significantly smaller when using the principal components PC1 and PC2, with males larger than females (Fig. 2, MANOVA: Taxon F₂ , ₇₀ = 197.85, p < 0.01, Sex F₂ , ₆₉ = 6.46, p < 0.01; see also Table 1 and Appendix 2). To complement our morphological analyses, we obtained the fresh mass from the label of ten specimens of O. m. motmot (from Venezuela and Brazil deposited at MZUSP and EBRG) and a single record from the literature for O. m. ruficeps (Graves & Zusi, 1990). Ortalis m. motmot is heavier than O. m. ruficeps, weighing on average 476 g (ranging from 431 to 520 g, including males and females), while the sole adult male of O. m. ruficeps (Graves & Zusi, 1990) (Fig. 3, MANOVA: Taxon F₂ , ₄₈ = 12.61, p < 0.01), although the duration of each note varied among the two taxa (Appendix 3 and 4).  (Fig. 4).
Although occurring mostly in riverine habitats, there is no sign of contact between northern and southern populations. Thus, these birds comprise another classic case in Amazonian biogeography of taxa separated by the course of the main rivers (Cracraft, 1985). A few specimens of O. m. ruficeps allegedly share the same localities of O. m. motmot specimens (Lago Cuipeua and Pinhel, both in Pará). These individuals were also noticed by Vaurie (1965), who regarded the localities as erroneous,  Ellipses are 95% confidence ellipse (b) Loadings plot of the two first two first components obtained from the Principal Component Analysis of the vocal characters. Tomotani, B.M. et al.: Taxonomy of Ortalis motmot ruficeps Pap. Avulsos Zool., 2020; v.60: e20206012 4/12 which was later supported by Pinto (1978). These localities were also scrutinized by Wiley (2010), who raised a considerable suspicion of carelessness in labeling by the original collector A. Olalla. Thus, the correct provenance of these specimens is considered doubtful. Because of this and the lack of additional material indicating syntopy, we excluded these problematic localities from the map (Fig. 4).

APPENDIX 3
Outcomes of the principal component analysis for song.