Distribution extension of Lasiurus arequipae Málaga, Díaz, Arias & Medina, 2020 (Chiroptera, Vespertilionidae) and contributions to its natural history

The recently described bat Lasiurus arequipae Málaga, Díaz, Arias & Medina, 2020 is endemic to Peru . Lasiurus are­ quipae is currently known to occur at two localities in the department of Arequipa and one locality in the department of Tacna in southwestern Peru. Here, based on morphological and molecular data, we present a fourth locality that represents the southernmost record of this species and extends its known range 62 km south from the nearest locality in Tacna. Additionally, we provide information some aspects on the natural history and acoustic profile of L. arequipae .


Introduction
Vespertilionidae Gray, 1821, is the most taxonomically diverse bat family in the world with more than 520 species (MMD 2021). Approximately 110 vespertilionid species occur in the Americas (Solari and Martinez-Arias 2014; Moratelli et al. 2019a). In Peru, the family is the third most diverse family, after Phyllostomidae and Molossidae, with 25 species grouped in four genera: Eptesicus Rafinesque, 1820; Lasiurus Gray, 1831; Myo tis Kaup, 1829;andRhogeessa Allen, 1866 (Pacheco et al. 2021). The genus Lasiurus includes the red, yellow, and hoary fur bats and is easily distinguished from other genera by its wide palatal emargination (approximately a third of the distance across the canines) that separates the premaxillae and upper incisors, and by its hairy uropatagium on its dorsal surface (Gardner and Handley 2008). Recently, Baird et al. (2015Baird et al. ( , 2017Baird et al. ( , 2021 suggested the division of Lasiurus into three genera based on molecular data and external differences in fur colors: Lasiurus (red bats), Aeorestes (hoary bats), and Dasypterus (yellow bats). This taxonomic arrangement has been disputed by Ziegler et al. (2016), Novaes et al. (2018), and Teta (2019), citing a lack of support for this division in a monophyletic genus and stating that using subgeneric names preserves the nomenclatural stability and facilitates the searching in databases, especially when Aeorestes had never been used as a name for hoary bats. In accordance with these publications, we recognize Lasiurus to include the subgenera Aeorestes and Dasypterus.
Twenty species of Lasiurus are currently recognized, and these are mostly distributed throughout the Americas, from Canada to Argentina. Moreover, vagrant records are also known in northern Canada, Bermuda, Hawaii, the Orkney Islands, and Iceland (Allen 1939;Hall and Jones 1961;Koopman and Gudmundsson 1966;Gardner and Handley 2008).
In Peru, Lasiurus is represented by four of the eight species occurring in South America and is distributed from sea level to 3300 m a.s.l. (Graham 1983;Málaga et al. 2020;Pacheco et al. 2021): • Lasiurus blossevillii (Lesson & Garnot, 1826), is characterized by the gray tips of the hairs on the dorsal fur giving it a hoary or frosted aspect (Díaz et al. 2016) and occurs on both sides of the Andes in the Bosque Seco Ecuatorial, Bosque Tropical Amazónico, Bosque Tropical del Pacífico, Yungas, Páramo, Puna, and Desierto Costero ecoregions (Allen and Keays 1901;Thomas 1926;Ortiz de la Puente 1951;Graham and Barkley 1984;Solari et al. 2001;Quintana et al. 2009); • Lasiurus ega (Gervais, 1856), characterized by its olivaceous yellow coloration, has also been recorded on both sides of the Andes in the Bosque Tropical Amazónico, Bosque Tropical del Pacífico, and the Yungas ecoregions (Hall and Jones 1961;Pirlot 1968;Gardner 1976;Woodman et al. 1991;Ascorra et al. 1996);• Lasiurus villosissimus (E. Geoffroy St.-Hilaire, 1806), characterized by its large forearm (>50 mm) and frosted predominantly gray and yellowish fur, occurs on the east side of the Andes in the Yungas ecoregions (Sanborn 1953;Gardner 1976); and • Lasiurus arequipae Málaga, Díaz, Arias & Medina, 2020, recently described, can be distinguished from the other three species by its medium-sized forearm (46.4-47.6 mm), dense fur, cinnamon to orangish dorsal coloration mixed with dark brown giving it a marbled appearance, blackish wing membrane, and reddish hairs covering the forearm and metacarpals. The separation of L. arequipae from its sister species has been confirmed by its position in a phylogeny based on the mitochondrial marker cytochrome b (Cyt-b) and in a high genetic divergence (<9.4%) (Málaga et al. 2020). Lasiurus arequipae was originally only known by three specimens, one male collected at the type locality in the coastal valley of Huatiapa at 725 m a.s.l. and two (one male and one female individual) from Chaucalla at 860 m a.s.l. in the department of Arequipa. Most recently, this species was also reported for Lomas de Sama Grande at 555 m a.s.l., department of Tacna, collected while roosting in an olive tree (Mamani-Contreras and Aragón-Alvarado 2021). Lasiurus arequipae is the only species of Lasiurus endemic to Peru, and the second species of Lasiurus recorded in the Desierto Costero ecoregion where only L. blossevillii had been known. Ugarte-Núñez (2020) developed an echolocation key to 20 species of bats from the Pacific coast and the western slopes of the Andes of southern Peru, including L. arequipae, but did not provide the location and the source of the acoustic records. The echolocation patterns described of L. arequipae show peak frequencies of ~35 kHz and a descending modular frequency at the beginning and at the end, with a middle portion close to the quasi-constant frequency, supporting the distinction of L. arequipae from other species of Lasiurus (Briones-Salas et al. 2013;Rodriguez-San Pedro and Simonetti 2013;Rodriguez-San Pedro et al. 2014Rivera et al. 2015).
Here, based on morphological and molecular data, we report new specimens of L. arequipae from the department of Tacna, Peru, extending the known distribution of the species further south from Lomas de Sama Grande and from its type locality. We also provide information on its natural history and echolocation calls.

Methods
We sampled bats in Sama Valley, department of Tacna, Peru from the wet local season in 2015 to the wet local season in 2016. This valley, in which the Sama River flows, is in the Desierto Costero ecoregion. Xeric vegetation grows in the cracks and dry riverbeds of this desert region (Jaén et al. 1963;Morris and Panty 1999). We established six sampling stations with seven mistnets (12 × 2.5 m) each. The nets were set perpendicular to the Sama River. The sampling was carried out for four consecutive nights per station during both wet and dry seasons. Sampling was performed at each station for 11 hours after sunset. The total sampling effort total was 336 net-nights. This research was authorized by the Servicio Nacional Forestal y de Fauna Silvestre (SERFOR) through R.D. N°246-2017-SERFOR/ DGGSPFFS and authorization code for scientific purposes AUT-IFS-2017-057.
Our two collected specimens (one male, one female) were identified as Lasiurus based on morphological characters described by Díaz et al. (2016) and then compared with specimens deposited in the Museo de Historia Natural de la Universidad Nacional Mayor de San Marcos (MUSM). The specimens went unidentified to species until after the description of L. arequipae (Málaga et al. 2020).
Our specimens are deposited at MUSM. Both specimens are preserved in ethanol. In the field, external measurements were taken using Fowler 54-100-444 Sylvac Ultra-Cal Mark IV digital calipers with an accuracy of 0.01 mm; these measurements included the head-body length (HB), tail length (T), hindfoot length (HF), ear length (E), and forearm length (FA). Both individuals were weighed using a 50 g Pesola spring scale to the nearest 0.1 g. The skull of the female specimen was removed and cleaned, and the following craniodental measurements were taken with the same calipers, also recorded to the nearest 0.01 mm: greatest length of the skull (GLS), condyle-basal length (CBL), zygomatic breadth (ZB), interorbital breadth (IB), postorbital constriction (POC), mastoid breadth (MAB), braincase breadth (BCB), palatal length (PL), maxillary toothrow length (C-M), upper canine breadth (C-C), upper molar breadth (M-M), lower toothrow length (c-m), and greatest length of mandible (GML). The craniodental measurements follow Barquez et al. (1999) and Málaga et al. (2020).
DNA was extracted from a fragment of muscular tissue preserved in alcohol 96° GL of the specimen MUSM 52933 using the GeneON "Vivantis" kit (GF-TD-100). A fragment of the cytochrome b gene (Cyt-b) was amplified using primers glo7L (5′-CAY-CGTTGTATTTCAACTRTAAGAAC-3′) and glo6H (5′-CGGTGTAATGRATATACTACATRG-3′). We followed the thermal profile described in Hoffmann and Baker (2001) for PCR, and the amplification product was sent for sequencing only with the forward primer at Macrogen, Inc (Seoul, South Korea). We obtained 778 bp of Cyt-b and deposited the sequence at GenBank (accession code OK030950).
A phylogenetic analysis was performed using the sequence obtained in this study and 34 sequences were retrieved from GeneBank, which had previously been used by Baird et al. (2015) and Málaga et al. (2020). These sequences belong to the 16 species of Lasiurus and are listed in the Appendix, Table A1. Sequences of Tadarida brasiliensis (I. Geoffroy St.-Hilaire, 1824) and Myotis velifer (J.A. Allen 1890) were included as outgroups. The sequences were aligned in Geneious v. 7.12.3 using the ClustalW algorithm (Thompson et al. 1997). The final alignment had 1040 bp. The phylogenetic analyzes were performed using the maximum likelihood (ML) and Bayesian inference (BI) methods. For the ML analysis we used RAxML v. 7.2.8 (Stamatakis 2014) using GTRGAMMAI as the nucleotide substitution model and 1,000 pseudo-replications for bootstrapping and nodal support. For the BI analysis, the best nucleotide substitution model was estimated using jModelTest (Darriba et al. 2012) using three schemes of substitution (= 24 models) and the Akaike Information Criterion (AIC). The best model was HKI+I+G, which was implemented in MrBayes v. 3.2.7 (Ronquist et al. 2012) using four independent MCMC runs, 10 million generation sampled every 1,000 generation to estimate the posterior probabilities as nodal support. We considered 25% of the samples as burn-in. Convergences of the Bayesian analysis were checked on Tracer v. 1.6 and followed the recommendations described in the MrBayes manual: standard deviation less than 0.01 (it was 0.008), the potential scale reduction factor (PSRF) near to 1.0 (it was 0.99-1.0), and the effective samples size (ESS) more than 200 (it was 6000). Finally, genetic distance was estimated with Kimura 2-parameters (Kimura 1980) in MEGA v. 7.0.14 (Kumar et al. 2016). The phylogenetic tree was visualized in FigTree v. 1.3.4 and edited in Inkscape v. 0.9.
The lack of information about the acoustic records for this species, apart from the information provided by Ugarte-Núñez (2020), motivated us to provide complementary data. In Chaupo (Quechualla district, Arequipa), an individual with similar external characteristics (medium FA = 47.1 mm and the orange coloration in dorsal fur) to L. arequipae was captured in a mist-net; it was released following the methodologies of Kunz andParsons (2009) andEstrada-Villegas et al. (2013), in order to perform acoustic recordings after the bat started to fly. Additionally, in Velinga (Quechualla district) and Sogay (Yarabamba district), bat acoustic recordings were made without capture. The recordings were made using two devices: Echo Meter 3 (Wildlife Acoustics) and Anabat Walkabout (Titley Scientific); both with a full-spectrum 16-bit resolution and a sampling quotient of 384 kHz set at 10 sec.
To define the main acoustic parameters of the searching calls obtained, the echolocation analysis was made using Raven Pro v. 1.5. (Center for Conservation Bioacoustics, Cornell Lab 2014). Parameters like peak frequency (PF), minimum frequency (LF), maximum frequency (HF), bandwidth (BW) were given in kilohertz (kHz), while call duration (CD) and inter-pulse (IP) were given in milliseconds (ms).
We developed a distribution map for the four species of Lasiurus distributed in Peru with ArcMap v. 10.3 using the material we examined at MUSM, as well as bat records of Peruvian specimens available in the scien-

Results
Morphological and molecular analyses confirm that the Lasiurus specimens collected in the Sama Valley (Tacna, Peru) are L. arequipae. Moreover, based on bioacoustic monitoring and the parameters of the echolocation calls, the specimens recorded at Chaupo, Velinga, and Sogay corresponds also to L. arequipae. The locality, north of Tomasiri, is in a 1 km wide alluvial plain having a narrow riparian forest and a broad expanse of agricultural crops. One side of the river is restricted by a hillside about 50 m high. The individuals we captured were a non-lactating female and a male with descended testes.
Identification. The two individuals, MUSM 52933 (♀) (Fig. 1) and MUSM 52934 (♂), fit the description of L. arequipae (see Málaga et al. 2020). This is supported by a dense dorsal fur that varies from cinnamon-brown to orange. The dorsal hairs are tricolored (black basal band, yellowish mid-band, and dark brown tips), giving a marbled appearance to the fur. The ventral fur hairs are bicolored (black base and yellowish tips). Wing membranes are blackish with red hairs on both sides of the forearms and metacarpals. Cranio-dental measurements were also within the range provided by Málaga et al. (2020) (Table 1). The anterior border of the nasal bones is V-shaped. The shape of the ocular orbit is oval in dorsal view; the post palatal process is poorly developed; the mesopterygoid fossa is V-shaped; and the tympanic ring is circular in shape (Fig. 2). External measurements are given in Table 1.
The phylogenetic reconstruction with ML and BI analyses found that our Tacna specimen is sister to the only published sequence of L. arequipae (MUSA 21958). Moreover, the 778 bp of Cyt-b of these specimens from Tacna and Arequipa are 99.23% similar, while their K2p intraspecific genetic distance is less than 0.01% (Fig. 3).
Acoustic recordings of a Chaupo individual caught, then released, show that the echolocation call had a downward modulate frequency harmonic (FM-d) and    of 75.56 kHz ± 11.78, and a frequency peak of 35.39 kHz. The average length of the call was 3.97 ms ± 1.23 with an interval of 91.31 ms ± 46.69, and a bandwidth of 56.29 kHz ± 12.78 (Fig. 4). These calls are typical of vespertilionids and are in the parameters for Lasiurus (Schnitzler and Kalko 2001), which has a marked difference among the acoustic data of L. varius Poeppig, 1835, L. cinereus (Palisot de Beauvois, 1796), and L. blossevillii (Briones-Salas et al. 2013;Rodríguez-San Pedro and Simonetti 2013;Rodríguez-San Pedro et al. 2014Rivera et al. 2015) ( Table 2). The map (Fig. 5) displays the current distribution of the Lasiurus species in the Peruvian territory: Lasiurus arequipae, Lasiurus blossevillii, Lasiurus ega, and Lasi urus villosissimus.

Discussion
Our records of Lasiurus arequipae from Sama Valley in Tacna extends the known geographic distribution of this species about 292 km south of its type locality and 62 km south from its occurrence at Lomas de Sama Grande, department of Tacna. The Sama Valley, belonging to the Desierto Costero ecoregion, has typical features of Peruvian coastal valleys, with moderate slopes (greater than 60°), warm temperatures (15-24 °C), and little, irregular precipitation (Málaga et al. 2020;Portugal-Zegarra et al. 2020). These climatic and geographic characteristics are also present in northern coastal Chile, such as in the Lluta and Azapa valleys (less than 100 km from Sama Valley), where other species like Platalina genovensium Thomas, 1928, Promops davisoni Thomas, 1921, and Nyctino mops aurispinosus (Peale, 1848, were recently confirmed to occur (Ossa et al. , 2018Rodríguez-San Pedro et al. 2022). Therefore, L. arequipae may occur to this area or even further south. Moreover, discovery of L. arequipae in the Desierto Costero partially fills a known geographic gap for the genus, which extends from the Atacama Desert in northern Chile (where only L. villo sissimus is known) to the Central Peruvian desert (where only L. blossevillii is known); no other Lasiurus species have been recorded before. It is unknown if L. arequipae   (2015). may occur in sympatry with L. blossevillii, L. villosis simus, or both, and if the more northern L. villosissimus might also be found in the southern Peruvian desert. Lasiurus arequipae has been described as occurring at 555-860 m a.s.l. (Málaga et al. 2020; Mamani-Contreras and Aragón-Alvarado 2021), although Málaga et al. (2020) mentioned the possibility that it may occur at 2760 m a.s.l. based on an individual collected in Sogay (Arequipa) but released. Our records from Sama valley, at 464 m a.s.l., are the lowest known occurrences of the species, while our echolocation surveys confirm the highest record of L. arequipae, at 2730 m a.s.l. in Yarabamba.
Our new acoustic data of L. arequipae allowed us to differentiate this species from L. blossevillii. Thus, L. arequipae has a lower peak frequency (35.39 kHz ± 3.44) and a distinctive major bandwidth, as well as a significative lower minimum frequency. The length of the call of L. arequipae is shorter than the other Lasiurus species. Therefore, the acoustic parameters of L. areq uipae do not overlap or approach those of the other species within their geographic distribution. Schnitzler and Kalko (2001) pointed out that bats with longer signals of low frequency would be adapted for large prey detection at long distances, which would apply for L. arequipae.
Among the Lasiurus species in Peru (Fig. 5), L. blos sevillii has a wider distribution and a preference for lowland forests, highland forests, the Puna, the Yungas, and even the Pacific Coastal Desert (Allen and Keays 1901;Thomas 1926; Ortiz de la Puente 1951; Graham and Barkley 1984;Solari et al. 2001;Quintana et al. 2009). The distribution of L. ega is restricted to ecosystems of the Bosque Tropical del Pacífico, Bosque Tropical Amazónico, and the Yungas (Hall and Jones 1961;Pirlot 1968;Gardner 1976;Woodman et al. 1991;Ascorra et al. 1996), while L. villosissimus has been recorded in the Yungas (Sanborn 1953;Gardner 1976). The distribution of insectivorous bat species like Lasiurus corresponds to temperature and altitude because of ecophysiological tolerances and thermoregulation (Pérez 2011). In this context, the three species converge in similar environments to the eastern Andean slopes and the lowland forests of Argentina, Bolivia, Brazil, Peru, and Uruguay, with L. blossevillii most adapted to higher altitudes and extreme environmental conditions.
Finally, we suggest that additional fieldwork must be conducted to the north and the south of the currently known distribution of L. arequipae, where this species could be facing threats from mining and urban expansion, similar to other insectivores such as Myotis ataca mensis (Lataste, 1892), Amorphochilus schnablii Peters, 1877, and Promops davisoni (Aragón-Alvarado and Aguirre 2014; Flores-Quispe et al. 2015). mens at Museum of Vertebrate Zoology, University of California, Berkeley, Dan Vivas (Laboratorio de Biología Molecular, Facultad de Ciencias Biológicas, Universidad Nacional Mayor de San Marcos) for his assessment of our molecular data and analysis, and Peter Steinour for his invaluable support, time, and patience to improve the language of this manuscript. Yazmín Ticona (Departamento de Mastozoología, Museo de Historia Natural, Universidad Nacional Mayor de San Marcos) and Lisseth Montes de Oca (Programa de Conservación de Murciélagos del Perú -Sede Tacna) provided invaluable support during the fieldwork.