Two overlooked elusive crakes (Aves, Rallidae): first country record of Ocellated Crake Micropygia schomburgkii in Ecuador and Rufous-faced Crake Laterallus xenopterus in Peru

ABSTRACT We present two new distributional country records based on previously misidentified sound recordings: Ocellated Crake, Micropygia schomburgkii (Schomburgk, 1848), in Zamora Chinchipe (Ecuador) and Rufous-faced Crake, Laterallus xenopterus Conover, 1934, in Madre de Dios (Peru). We show supporting spectrograms of sound recordings and provide tools to justify the correct identification of the trills of both species. This range extension is especially relevant for the Vulnerable L. xenopterus, representing a new important breeding site in the Bahuaja Sonene National Park, Peru.

Many crake species (Aves, Rallidae) figure among the least known birds in the World, in part, because of their secretive habits in densely vegetated and flooded habitats [1]. One of the easiest ways to detect these species is through their trill, which is their most frequent, loudest, and longest vocalization. However, trills of many crake species of the genera Anurolimnas, Laterallus, and Micropygia are so similar that they are not easily identified. The Ocellated Crake, Micropygia schomburgkii (Schomburgk, 1848), is a poorly known bird species [1,2]. It has a vast range and is categorized as Least Concern for conservation [3], but it is only found at scattered localities in diverse habitats from dry to wet grasslands in Argentina, Costa Rica, Colombia, Venezuela, Guianas, Peru, Bolivia, and Paraguay [1,4,5]. Its main vocalization is a relatively pure and penetrating solo trill that sometimes finishes in a series of raspy drawn-out notes [1,6] (Figure 1). The Rufous-faced Crake, Laterallus xenopterus Conover, 1934, is considered Vulnerable and it is probably among the five most secretive Neotropical crake species [1,3]. It was discovered in 1933 in Paraguay [7] and nothing was known of it until 45 years later when Sick reported a specimen captured in Brazil [8]. During the last 35 years, a few studies have added scant new information about this crake [e.g. [9][10][11][12][13][14]. Its range remains poorly understood and it is certainly known from a few localities in Bolivia, Brazil, and Paraguay [11,12,14]. It inhabits moist campo limpo dense grasslands in the Cerrado region and shallowly flooded savannas dominated by tussock grasses in the Beni Savanna region [11,14]. Its main territorial vocalization is a duet trill that is relatively short, more gargled, and with clearly distinguishable notes in comparison to other Laterallus [6] (Figure 2).
Here we provide evidence of two new distributional country records based on previously misidentified crake sound-recordings: M. schomburgkii for Ecuador and L. xenopterus for Peru. We present supporting spectrograms of sound recordings and provide tools to justify the correct identification of the song of both species.
EAD reviewed all available Neotropical crake records with associated supporting evidence at eBird (Cornell Lab of Ornithology, www.ebird.org), iNaturalist (www.inatural ist.org), Macaulay Library (Cornell Lab of Ornithology, www.macaulaylibrary.org), and Xeno-canto (www.xenocanto.org) up to September 2019 [6]. Thus, the identification of all the media of interest archived at those repositories (audio, photos, and videos) was checked. Particularly, the identification of sound recordings was corroborated aurally and visually (with the help of spectrograms) using Raven Pro version 1.5 (Cornell Lab of Ornithology, www.birds.cornell.edu/raven). Spectrogram parameters were set at window type = Hann, size = 312 samples (6.5 ms), 3 dB filter bandwidth = 221 Hz; timegrid overlap = 50%, hop size = 156 samples (3.25 ms); frequency grid-DFT size = 512 samples, grid spacing = 93.8 Hz. Specifically, two misidentified soundrecordings representing first country records were found: M. schomburgkii for Ecuador ( Figure 1; XC251040/ ML246349), and L. xenopterus for Peru (Figure 2; ML44076). Spectrograms of these sound recordings were compared to other correctly identified recordings (Figures 1 and 2, Appendix A). Using Esri ArcMap 10.4 (www.esri.com), the new country records were plotted in a map together with previous records of each species that count with supporting documented media [data from 6] ( Figure 3). We follow the taxonomy of Remsen et al. [15]. Associated information and comments of the first country records of M. schomburgkii for Ecuador and L. xenopterus for Peru are provided below.
One  [19,20]. Because trills of these species were misidentified by T. Parker at Pampas del Heath, and Parker et al. [21] reported sound recordings of L. exilis from other localities with the possible presence of L. xenopterus, we also reviewed Parker sound recordings from localities reported by Parker et al. [21] contained in unarchived cassettes at the Macaulay Library (Cornell Lab of Ornithology), as well as all his crake sound-recordings listed in those cassettes (Santa Cruz and Beni departments, Bolivia; Madre de Dios and Loreto departments, Peru; Brasilia, Espirito Santo, Parana, and Rio Grande do Sul states, Brazil; mostly between years 1988 and 1990). However, these tapes contained no L. xenopterus recordings.
Micropygia schomburgkii utters a trill that rises in pitch and then maintains a fairly constant frequency, composed of a long series of relatively pure and loud pr pr pr notes [1,6]. Because these trills are delivered in solo (i.e., by one individual), the note rate heard is lower than in duet trills of other crakes (EAD pers. observation) (Figure 1). Nevertheless, the trills of this species are frequently confused with those of Anurolimnas viridis, and L. xenopterus (EAD pers. observation). However, the trills of A. viridis are (1) more strident, (2) much longer in duration, (3) descending in frequency from start to end, and (4) composed of glassy tinkling notes; whilst the trills of L. xenopterus are (1) slightly lower pitched, (2) shorter, and (3) with almost twice the number of notes per second [6,22]. Additionally, M. schomburgkii also gives a series of harsh progressively more drawn-out pjrrr or prrrxzzz "crying" notes that can be given sometimes immediately after solo trills (Figure 1) or sometimes as isolated calls [1,6]. The identification of XC251040/ML246349 as M. schomburgkii is justified because no other neotropical crake is known to give a solo trill followed by "crying" notes [1,6] (Figure 1).
Laterallus xenopterus gives a duet trill typical of other members of the genus [1,6]. Its trills resemble those of the Red-and-white Crake, L. leucopyrrhus (Vieillot, 1819), in being somewhat gargled, and with clearly distinguishable notes, but similar in pitch to those of the Rufous-sided Crake, L. melanophaius (Vieillot, 1819) [1,6]. Diagnostic features of the trill of L. xenopterus are (1) its relatively short duration, typically the shortest trills of the genus, and (2) the ending in one or more isolated final notes that follow the main body of the trill [6] (see Figure 2). In contrast, the trills of L. exilis are faster and harsher rattles that descend in pitch ( Figure 2) [1,6,23]. Duet trills of these species are two-parted, beginning with soft initial notes audible at close range that are then followed by the loud-trilled portion which is audible at longer distances [22]. The soft initial notes of Laterallus crakes are often diagnostic; L. exilis presents prí prí prí initial notes, while L. xenopterus typically utters two well-spaced wavy and pure whistled notes that ascend and descend in frequency ("preew . . . . . . preew") [see 22]. However, soft initial notes are not always given before the trills and can be difficult to record because they are only audible at close range [6,22]. The soft introductory notes are missing or not audible in ML44076, possibly because the duet trill is not in focus at the beginning, the highbackground noise of the recording may mask the soft initial notes, and/or the initial parts of the trills overlap with the loud song of an Ash-throated Crake, Mustelirallus albicollis (Vieillot, 1819) that figures prominently in the recording. Because of all the arguments brought above, and especially because no other neotropical crake utters final isolated notes at the ending of the duet trill, the identification of ML44076 is justified as L. xenopterus [6] (Figure 2).
These new records extend the known ranges of these species by 265 km NW for M. schomburgkii and by 335 km NW for L. xenopterus (measured from the nearest previous species record showed in Figure 3, records from [6]). This range extension is especially relevant for the Vulnerable L. xenopterus. Both records occurred in habitats known to be used by these species. Also, these new range extension records do not constitute extensions in the known elevational range of the species, neitherfor M. schomburgkii (new record at 1,100 m. a.s.l; known range from 4 to 1,707 m. a.s.l) nor for L. xenopterus (new record at 210 m. a.s.l; known range from 128 to 1,236 m. a.s.l.) -elevational data was gathered only from species records with associated supporting data such as photos, sound-recordings or museum specimens, and reliable bibliography [4,6].
The M. schomburgkii new record was made by NK in a semi-natural savanna with introduced cattle grass, adjacent to a patch of recently burned grassland at the road that connects La Chonta and El Chorro villages, Zamora Chinchipe, Ecuador. A. viridis and L. exilis also occur in the area (EAD pers. observation). Indeed, the trills of A. viridis shown in this study and the new M. schomburgkii record were obtained at less than 700 m from each other (see Appendix A). However, as shown for other crakes, ecological segregation between these species is apparent to some extent [e.g. 23,24]. Even though M. schomburgkii and A. viridis share fern scrub with some Melastome shrubs and Curatella americana trees, M. schomburgkii apparently prefers more open and drier areas, while A. viridis prefers more humid areas with taller vegetation, and L. exilis uses 0.5-1.5 m tall open grasslands in the area (pers. observation). These three crake species utter trills and sometimes respond to interspecific playback trials among them [EAD pers. observation; also see [22][23][24]. As these species share general habitats to some extent, occur in close proximity, have similar vocalizations, and are difficult to see, M. schomburgkii could have remained overlooked in the area. Nevertheless, M. schomburgkii is considered to present some vagrant or occasional movements, especially associated with fires [1]. However, after being informed about the record here reported, Jonas Nilsson (in litt.) detected at least six M. schomburgkii territories at the same site during 12-14 November 2019 and 19-22 January 2020. These new records of the species 15 years after the first sound recording, might be the result of fortuitous vagrant events, but we believe they indicate the continued presence of the species in the area.
The Beni savannas, also known as the Moxos plains or Llanos de Mojos, lie mostly within Bolivia. However, small areas extend along the Iténez (Guaporé) River into the State of Rondônia in Brazil and in the Pampas del Heath of the Madre de Dios Department in Peru [25]. The presence of L. xenopterus was documented for El Beni Department of Bolivia in 1996 [11]. Our new record of L. xenopterus for Peru fits nicely in this biogeographic scenario because the Beni Savannas and the Pampas del Heath belong to the same biome complex. This also represents the first known sound recording of L. xenopterus, followed by those from Paraguay in 1997 by Juan Mazar Barnett (XC15482 to XC15485). Also, this new record would be the fourth known locality for L. xenopterus [7,8,26,27]. L. xenopterus is a very secretive species that has recently become better known [e.g. 13,14]. Lack of surveys at Pampas del Heath and the trill similarity between L. xenopterus and L. melanophaius [1,6], also present in the Beni savannas [11,28], may explain the lack of records from this area. No seasonal movements are known in L. xenopterus [1], suggesting that the Pampas del Heath is a new important breeding site for this species. Fortunately, all of the Pampas del Heath are protected in Peru by the Bahuaja Sonene National Park [25].
Finally, these new records based on sound-recorded vocalizations demonstrate the importance of natural history audio archives and highlight the need for more sound recordings of bird vocalizations. Crakes are more often heard than seen, and this seems especially true for the elusive M. schomburgkii and L. xenopterus. Visual detection of these species would have provided unmistakable clues to species identification; however, the two recordings discussed here were made without seeing the birds. Without these sound recordings, these new country records would have been lost. Continued economic and logistic support for sound-recording field trips at poorly surveyed remote areas is key to a complete understanding of the diverse Neotropical avifauna.

Acknowledgments
We are very grateful to Ted Parker, whose titanic survey efforts and documented records continue to enlighten us and to contribute to bird conservation. We thank Fabricio C. Gorleri, Jefferson García Loor, and all the rangers of Aguará Ñú 2016 for their companionship and help in field trips for sound-recording. Fabricio C. Gorleri also helped with elevational data. We specially thank Myriam Velázquez and Elisa Bonaccorso for their help with study permits and field trip logistical requirements in Paraguay and Ecuador, respectively. We thank Greg Budney, who gave us valuable comments about Ted Parker soundrecordings. We appreciate all the assistance provided by Matthew Medler, from the Macaulay Library (Cornell Lab of Ornithology). Our research was funded by the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), and the Neotropical Bird Club through a Conservation Award. Neither had input into the contents of the manuscript nor required their approval of the manuscript before submission or publication. All the study permits needed for the development of our field studies were granted by the pertinent governmental institutions of Paraguay and Ecuador.

Disclosure statement
No potential conflict of interest was reported by the authors.