A new species of Long-eared Brown Bat of the genus Histiotus (Chiroptera) and the revalidation of Histiotus colombiae

The South American bats of the genus Histiotus comprise between four and eight species, but their taxonomy has been controversial and the limits between species and their distribution are not well understood. In Colombia, Histiotus humboldti and H . montanus colombiae have been recorded, but undescribed species has been suggested. We evaluated the species richness and distribution of Colombian Histiotus using morphological, molecular, and acoustic traits. Our results evidence three species in Colombia, the two previously recorded taxa and a new species from the Cordillera Central of Colombia and northern Ecuador that we describe here. We also revalidated H . colombiae as a full species. H. humboldti is widely distributed in the Colombian and Ecuadorean Andes and can be sympatric with the other two species. H. colombiae is restricted to the Colombian Cordillera Oriental. Finally, we highlight the potential hidden diversity within Histiotus in the Peruvian and Bolivian Andes, the need to resolve the evolutionary relationships of the genus, and its implications to the understanding of the processes that have structured the Andean mammal fauna.


INTRODUCTION
demic to South America, occurring along the Andean chains in a variety of mountainous ecosystems, from Venezuela and Colombia to the most southern continental lands of Argentina and Chile, the coastal Atlantic Forest of eastern Brazil, and the semiarid regions of Argentina and Brazil (Handley and Gardner 2008, Feijó et al. 2015, Moratelli et al. 2019. Histiotus are easily differentiated from other South American Vespertilionids by their very long pinnae and large tympanic bullae (Handley and Gardner 2008). Recently, molecular phylogenetic analyses found that Histiotus was nested within the clade of the New World Eptesicus (Hoofer and Van Den Bussche 2003, Roehrs et ships among worldwide Eptesicus remain unresolved and no taxonomic decisions have been taken. Despite that, based on the unique morphology within the South Ameri-al. 2010, Amador et al. 2018. However, the relation can vespertilionid bats, Histiotus has been recognized as a valid genus in recent taxonomic treatments (Simmons 2005, Handley and Gardner 2008, Burgin et al. 2018, Moratelli et al. 2019).
The Big-eared Brown Bats of the genus Histiotus are en versial and the species limits unclear, due to the similar interspecific skull morphology, few specimens available in ed taxonomic and geographic representativeness (Feijó et ample, Simmons (2005) recognized seven species: H.
The alpha taxonomy of Histiotus has been also contro Contributions on the taxonomy and distribution of Histiotus are few and mostly dedicated to southern South American forms. For instance, Feijó et al. (2015) atelli et al. (2019): H. montanus montanus, H. m. colombiae, H. m. inambarus, H. diaphanopterus, H. humboldti, H. ed specimens from two independent undescribed taxa, both from Peru. One of them is based on one specimen from Cajamarca housed at the American Museum of Natural History (AMNH 268090) that could not be assigned to any Histiotus species known to date. The second has been mentioned in the literature as Histiotus sp. by Giménez et al. (2019) based on two specimens (AMNH 278521, 278524) from Piura (Supplementary material, Appendix rial: total length (TL), tail length (TaL), hindfoot length (HFL), ear length (Ear), forearm length (FA), greatest length of skull (GLS), condylo-incisive length (CI), zygolaephotis, H. macrotus, and H. velatus. We also includ 1). We recorded the sex, the locality, and five external measurements from the specimen's labels (in mm). Thirteen craniodental and mandibular measurements were taken from specimens with calipers accurate to the nearest 0.01 mm. The eighteen standard measurements, including external, craniodental and mandibular, are based on Handley (1996), Moratelli et al. (2013), andFeijó et al. (2015). These measurements are described in appendix 2 of the supplementary mate matic breadth (ZB), postorbital breadth (PO), braincase breadth (BCB), mastoid breadth (MB), braincase height (BCH), length of the maxillary tooth row (MTR), palatal length (PAL), post-palatal length (PPAL), breadth across upper molars (M-M), breadth across the canines (C-C), mandibular tooth row (Dent-L).

Morphological analyses
We studied 103 specimens belonging to eight proposed taxa of Histiotus according to Simmons (2005) and Mor To evaluate morphometric variation within Histiotus from Colombia, we performed a Principal Component Analysis with the thirteen cranial measurements using the statistical package PAST version 4 (Hammer et al. 2001). We used the covariance matrix to preserve the information about the relative scale among variables and all data were transformed log 10 . Due to the completeness of our dataset, only 60 adult specimens (Appendix 1 in supplementary material) with all the variables representing eight taxa (H. humboldti, H. laephotis, H. velatus, H. macrotus, H. montanus montanus, and H. montanus colombiae) were included in our morphometric analyses. For H. m. colombiae we included samples from both the cordilleras Oriental and Central of Colombia and from northern Ecuador. To find discrete characters supporting morphometric groups we also explored the diagnostic external and craniodental traits available in the literature for all studied specimens.

Molecular methods
boldti from the Cordillera Oriental of Colombia, using standard phenol-chloroform methods (Sambrook et al. 1989) on muscle tissues preserved in 96 % ethanol. Unfortunately, we could not access to tissues of specimens identified as H. montanus colombiae from the Cordillera Central. Preservation in formalin of ICN specimens refrain us from successful DNA amplification. Amplification of cytochrome-b (Cyt b) was performed using primers glo7L and glo6H (Hoffmann and Baker 2001). A partial sequence between 900 and 1100 base pairs of all individuals was ampli lecular SSiGMol at the Universidad Nacional de Colombia, Colombia, Bogotá. We submitted genetic sequences to Gen-ed and sequenced following Porter et al. (2007), although we increased the annealing temperature from 45-48 °C to 50.5 °C. We carried out purification and sequencing of both strands with the amplification primers on an ABI 3500 sequencer (Applied Biosystems, Waltham, MA, USA) at the Servicio de Secuenciación y Análisis Mo Bank under accession numbers presented in Table 1 in the We extracted whole genomic DNA from two individuals of H. montanus colombiae and four individuals of H. hum fi supplementary material. Sequences were manually checked and aligned using BioEdit 7.2.6 software (Hall 1999 Handley, 1960, to be used in phylogenetic analyses and genetic distance comparisons (accession numbers available in Table 1 in supplementary material). The complete data set consisted of 26 sequences. We inferred a phylogenetic tree using Maximum Likelihood analyses in ware (Ronquist and Huelsenbeck 2003), running 10 × 10 6 generations with one cold and three incrementally heated Markov chains, random starting trees for each chain, and trees sampled every 100 000 generations. We discarded 20 % of the resulting trees as burn-in, and 85 % of the trees were used for generating a 50 % majority-rule consensus. Statistical support for resulting phylogenies was measured in bootstrap support values (of 500 iterations) and Bayes-b gene (1140 bp) sequences from GenBank represent macrotus, and H. sp. from Peru), and the out RAxML 1.5 beta software (Stamatakis 2014), and Bayesian Inference analyses using MrBayes v.3.1 soft ian posterior probabilities. Genetic distance values were estimated using the p-distance method.
We retrieved two additional sequences from BoldSys Acoustic characterization tector (Anabat Swift, Titley Scientific) with a sampling rate of 500 kHz. The bat detector was placed at a height of two meters above ground level in an open space, close to a roost in an abandoned rural house roof (Cundinamarca: Guachetá, Gachetá Alto: 5°27'39.47"N; 73°39'44.57"W; 2893 m). We collected four voucher specimens to confirm the species identification and deposited them at the Colección de Mamíferos Alberto Cadena, Instituto de Ciencias Naturales -ICN collections (ICN 24845-24848). We analyzed only the recordings with at least five pulses of good signal to noise ratio. Sequences of echolocation calls were displayed simultaneously as spectrograms and oscillograms using the software Raven Pro 1.6.1 (Center for Conservation Bioacoustics 2019). Spectrograms were made of consecutive fast Fourier transformations with an 85 % overlap and visualized on a Hamming type window. For each pulse, we manually measured the following parameters: 1) duration (time between start and end of a call, measured in milliseconds); 2) peak frequency (frequency in kHertz corresponding with maximal intensity in the power spectrum), 3) start frequency, 4) end frequency, and 5) interpulse interval (measured from the beginning of a call to the start of next call). We measured these variables to compare them with the data from H. montanus montanus from Chile reported by Ossa et al. (2015). All the values of acoustic parameters are given as mean ± 1 SD, since the data sets were not normally distributed, we used nonparametric statistics (Kolmogorov-Smirnov test). We carried out the tests using the software PAST version 4 (Hammer et al. 2001).
We recorded echolocation calls of H. montanus colombiae from the Cordillera Oriental using an ultrasound bat-de

Nomenclatural statement
This article has been registered in the Official Registry of Zoological Nomenclature (Zoobank) as 3AEB5C6F-509B-4BB3-AB7E-E5C81A946571. In this way, we comply with the requirements of the International Code of Zoological Nomenclature (Article 8.5.3).

Morphological traits
Based on exclusive combinations of discrete characters, for Colombia, we identified three morphological groups of Histiotus including H. humboldti, H. m. colombiae from the Cordillera Oriental and H. m. colombiae from the Cordillera Central and northern Ecuador; the latter was considered as a new species described below (see comparisons below and Table 2 in supplementary material). In addition, there is morphological evidence to differentiate the nominal H. montanus from southern South America from the three morphological forms found in Colombia (see description and comparisons below and Table 2 of the supplementary material).
The PCA analysis of the morphometric data including specimens from all South America shows that the first two principal components (PC) accounted for 78.8 % of the total variance (PC1 = 70.9 %, PC2 = 7.9 %). The length of the maxillary toothrow (MTR), the palatal length (PAL), the breadth across upper molars (M-M), and breadth across canines (C-C) were related with PC1 and the postorbital breadth (PO) with PC2.
The projection of the first two components of the PCA ( Fig. 1)

Acoustic results
We obtained a total of seven wave files from H. m. colombiae from individuals flying in the surroundings of their roost.

Taxonomic remarks
We recognize the bats previously identified as H. montanus colombiae from the Cordillera Central of Colombia and northern Ecuador as a new species based on a unique combination of morphological traits (    Distribution: Histiotus cadenai sp. n. is currently known to occur in 10 localities (Appendix 1 in Supplementary material) along the northern Andes in a latitudinal range from 01°18'S to 05°04'N. In Colombia, the species has been recorded in six localities in the Cordillera Central (departments of Caldas, Quindío, Risaralda, Tolima, and Valle del Cauca). In Ecuador, the species has been recorded in four localities of the northern portion of the Ecuadorean Andes (provinces of Napo, Pichincha, and Tungurahua). The known altitudinal distribution encompasses a gradient between 2550 to 4000 m a.s.l, including highland ecosystems such as Paramo, Andean montane forest, and anthropized areas. The distance between the extreme localities is approximately 797 km.
Etymology. The species epithet honors the Colombian pioneering mammalogist Alberto Cadena-García. He trained several generations of researchers currently working around the world on topics such as biology and conservation of mammals, continuing his legacy for new generations of biologists.
Diagnosis. This species is distinguished from all other species of the genus Histiotus by the following combination of characters: a medium-sized bat (total length: 102-116 mm; forearm length: 45-50 mm; greatest length of skull: 17.12-18.13 mm), with long and silky dorsal fur light brown, ventral coloration is yellowish-brown. Ears are very long (length: 31-36 mm) and not connected by a band. Ears and patagium dark brown, darker than fur coloration. Pinnae are triangular with rounded points and a notch on the outer edge near the tip. The tragus is ensiform, with a sword-like shape with parallel edges and an acute tip. Calcar is well developed and keeled without lappet. The skull is robust, has a globular braincase with a continuous slope in lateral view, but the posterior border is rounded. Sagittal crest present but form a triangular plate of bone at the intersection with the nuchal crests. Nuchal crests are well developed. Paraoccipital process well developed wide and blunt. The anterior region of the braincase has the parietals straight and highly convergent, forming an abrupt angle with the frontal borders in a conspicuous waist shape towards the postorbital constriction. The supraorbital region is swollen without marked postorbital ridges. The upper toothrow is straight. The position of I2 is lateral to I1. Upper premolar without an anterior projection. M1 has a well-developed but blunt protocone. The angular process has a laminar shape, and it is longer than the condylar process, projected outward, and outside of the condylar process plane. The anterior edge of the coronoid process is straight and forms a 90° angle with the dentary.
Description. Histiotus cadenai is a medium-sized bat (total length: 102-116 mm; forearm length: 45-50 mm; greatest length of skull: 17.12-18.13 mm). The dorsal fur is long, silky, and light brown with a brown base (3/4), and light brown tips (1/4). Ventral coloration is yellowishbrown with a dark brown base (2/3) and creamy yellowish tips (1/3). The ears are very long (31-36 mm) and not connected by a band. Pinnae are deltate (equilateral triangle form) with rounded points and a notch on the outer edge near the tip. The tragus is ensiform with parallel edges and an acute tip, the edge is crenulated, but it appears smooth without magnification. Ears and patagia are dark brown, darker than fur coloration. Wings are attached to inferior extremities on the metatarsal. Calcar is well developed and keeled at 2/3 basal of its length. The tail protrudes beyond the uropatagium (ca. 5 mm).
The skull is robust, has a globular braincase with a continuous slope from the parietals to the rostrum making it look flattened in lateral view, but the posterior border is rounded. Sagittal crest present, from the middle of the sagittal suture to the interparietal, where it forms a triangular plate of bone at the intersection of the nuchal crests. Nuchal crests are well developed along the suture occipitoparietalis and between the supraoccipital and the mastoid exposure of the petrosal. Paraoccipital process well developed, wide and blunt. At the basicranium, the basisphenoid has a flattened surface, the basioccipital has a wide longitudinal ridge with two deep fosses on each side, the lateral edge of these structures forms a laminar process curved up towards the bullae reaching the posterior basicochlear commissure. Bullae are well developed. Mastoid region bulges outward from the braincase. Zygomatic arches are straight on the squamosal ramus and convergent anteriorly towards the maxillary ramus. Glenoid fossa has a rectangular shape where the width is greater than the length, with a recurved posterior edge, and anterior edge well developed. Postorbital processes of the zygomatic arches are triangular, and their base breadth is greater than their height. In dorsal view, the anterior region of the braincase has straight parietals highly convergent towards the postorbital constriction, forming an abrupt angle with the frontal borders in a conspicuous waist shape. The rostrum is shorter than the braincase (1/3 of the total length of the skull). The supraorbital region is swollen without marked postorbital ridges. The external nasal fossa has a "U" shape with a conspicuous emargination of the nasal bones. Premaxillae are robust and continuous regarding the braincase and rostrum profile. Palatine processes of the maxillae of the hard palate are visible from the dorsal view. In the ventral view, the hard palate is deeply convex, and the posterior edge of the hard palatal has a spine forming an "M" shape. The hamular processes are thickened but straight. The dental formula is I 2/3 C 1/1 P 1/2 M 3/3, with a total of 32 teeth. The upper tooth row is straight. The inner upper incisors have convergent tips, each tooth has two asymmetrical cuspids. The position of I2 is beside or lateral to I1. The height of I2 reaches 2/3 of I1 height. Each upper incisor has a conspicuous cingulum around the tooth crown. Upper canine well developed, simple without accessorial cusps, but with a cingulum rounded the crown base. The upper premolar has a continuous cingulum around the crown base, without an anterior projection. Upper molars dilambdodont. The M1 has a well-developed protocone but blunt, the paracone is shorter than the metacone, reaching 2/3 of the size of the metacone. No stylar cusps are present. The M2 has a similar pattern and size as M1. The M3 has an "N" form without metastyle. Mandibular ramus is straight with an upward inflection below the coronoid process. The angular process has a laminar form, it is longer than the condylar process, projected outward, outside of the condylar process plane. The coronoid process is taller than wide with a rounded tip curved forward. The anterior edge of the process is straight and forms a 90° angle with the dentary. Lower incisors are trilobed and crowded one after the other.   Cranially, H. cadenai is easily distinguished from H. humboldti because the braincase and the rostrum of H. cadenai form a continuous slope in lateral view, whereas in H. humboldti the braincase and the rostrum form an angle (Figs. 4 and 5). H. cadenai has a developed sagittal crest that forms a triangular plate of bone at the intersection with the nuchal crests (not developed sagittal crest and the braincase is rounded without a bone plate in H. humboldti). The paraoccipital processes are well developed in H. cadenai, but wide and blunt (tubular in form and recurved downward) in H. humboldti (Figs. 4 and 5). Furthermore, H. cadenai can be distinguished from H. colombiae by the upper premolar without an anterior projection in the former and the upper premolar with a well-developed anterior projection in the latter. The supraorbital region of H. cadenai is swollen without marked postorbital ridges whereas H. colombiae has marked postorbital ridges. In H. cadenai the I2 is beside or lateral to I1, whereas in H. humboldti and H. colombiae the I2 is behind the I1. Finally, the angular process in H. cadenai has a laminar form and is projected outside the condylar process plane, whereas in H. humboldti the angular process has a tubular form and is projected upward in the same plane of the condylar process, and in H. colombiae has a laminar form, projected upward and outside, but in the same plane of the condylar process (Figs. 4 and 5).    and has a wider interorbital zone than other species such as H. laephotis and H. alienus. The upper premolar of H. colombiae has a well-developed anterior projection (see Thomas 1916). The supraorbital region of H. colombiae has marked postorbital ridges. The I2 is behind the I1. The angular process has a laminar form, projected upward and outside, but in the same plane of the condylar process (see Table 2 in the supplementary material for morphological comparison with additional Histiotus species).

Distribution of Histiotus in Colombia
Histiotus cadenai is distributed from the Cordillera Central in Colombia to the north of Ecuador, over 2550 m (Fig. 6a). The distribution of H. colombiae encompasses the high plains of the Cordillera Oriental in the departments of Boyacá and Cundinamarca in an elevational range between 2600 and 3100 m (Fig. 6a). Finally, H. humboldti has the widest distribution covering the three Cordilleras of the Colombian Andes and the north of Ecuador in an elevational range between 1700 and 3000 m (Fig. 6b). The inclusion of a taxonomic framework based on different sources of evidence is needed to clarify the diversity of the genus and will facilitate species delimitation. Unfortunately, taxonomic assessments are challenging because acoustic or genetic data are not available for most Histiotus species, and there are few specimens in natural history collections. For instance, the northern South American forms had not been studied since Thomas (1916) and Handley (1996) (Loureiro et al. 2018). This is particularly important in areas like the northern Andes that are considered a highly threatened biodiversity hotspot (Myers et al. 2000).

AUTHOR´S CONTRIBUTIONS
MRP collected the holotype and one paratype, DMMM processed DNA samples and conducted the Maximum-likelihood analyses, MRP, DMMM, and HERC examined collection specimens, describe the species and did the morphological and morphometric analyses. DMM analyzed the acoustic data. CACA Conducted Bayesian analyses. All authors contributed to the theoretical and conceptual framework, wrote the paper, prepared figures, and tables and reviewed drafts of the paper. Tragus eccentric with the principal axe in one side of the middle point of the structure.
Tragus ensiform as a sword form with parallel edges and acute tip.
Tragus ensiform as a sword form with parallel edges and acute tip.
Tragus ensiform as a sword form with parallel edges and acute tip.
Braincase and rostrum form an obtuse angle.
Braincase and rostrum form a continuous slope.
Braincase and rostrum form a continuous slope.
Braincase and rostrum form an angle.
Sagittal crest developed a triangular plate of bone at the intersection with the nuchal crests.
Sagittal crest present, a triangular plate of bone at the intersection with the nuchal crests.
Sagittal crest present, a triangular plate of bone at the intersection with the nuchal crests.
Supraorbital region swollen with marked postorbital ridges.
Supraorbital region swollen with marked postorbital ridges.
The paraoccipital process is tubular in form and recurved downward.
The paraoccipital process is well developed, wide and blunt.
The paraoccipital process is well developed.
Paraoccipital process very well developed.
Upper premolar without an anterior projection.
Upper premolar without an anterior projection.
Upper premolar with a well-developed anterior projection.
Upper premolar without an anterior projection.
The position of I2 is behind to I1.
The position of I2 is lateral to I1.
The position of I2 is behind to I1.
The position of I2 is behind to I1.
The angular process has a tubular form and projected upward in the same plane of the condylar process.
The angular process has a laminar form and projected outward, outside of the condylar process plane.
The angular process has a laminar form, and projected upward and outside, but in the same plane of the condylar process.
The angular process has a laminar form and projected upward in the same plane of the condylar process. 9 ----