A new species of Myotis (Chiroptera, Vespertilionidae) from Uruguay

The genus Myotis comprises a diverse group of vesper bats with worldwide distribution. Twenty-eight neotropical species are currently recognized. Based on a morphological approach, we describe a new species of Myotis from the Uruguayan Pampas grasslands, an ecoregion under high anthropogenic pressure with a largely unknown bat fauna. Qualitative and quantitative morphological analyses support the recognition of the new species and we present a set of external and cranial diagnostic characters by comparing them with other neotropical Myotis species. The new species reassembles Myotis riparius , but can be distinguished by a set of qualitative and quantitative morphological traits, including its clearly bicolored dorsal fur, tricolored ventral fur, a pelage on the dorsal surface of uropatagium, sagittal crest lower, braincase lower in lateral view and overall smaller size. DFA support the full distinction of Myotis sp. nov. from Uruguay compared to their more similar congeners and other South American Myotis . We present here the description of the new species from the Uruguayan Pampas, including a morphological diagnosis and comparisons with the other neotropical Myotis species.


Introduction
Myotis Kaup, 1829 comprises a diverse and worldwide distributed genus of vesper bats with ca. 140 species recognized (Moratelli et al. 2019a, b;MDD 2021). Despite the high diversity, species have reduced phenotypic variation and conspicuously conserved morphology compared to most of the other genera of vesper bats (Gunnell et al. 2012; Moratelli et al. 2019a). Therefore, Myotis is among the most difficult genera of bats to delimit species based on morphology only. However, several neotropical species whose evidence was based solely on morphology have been confirmed by subsequent molecular assessments, such as the cases of Myotis lavali Moratelli, Peracchi, Dias & Oliveira, 2011 from the South American dry Diagonal, and Myotis larensis LaVal, 1973from Venezuela (Moratelli et al. 2017Novaes et al. 2021). Even Myotis diminutus Moratelli & Wilson, 2011, whose description was based on one subadult individual, was confirmed as a distinct lineage based on molecular evidence (Platt et al. 2018).
New World Myotis form a monophyletic group divided into two clades with strong biogeographic associa-tions, one Nearctic and the other predominantly Neotropical (Stadelmann et al. 2007). Based on the congruence between molecular and morphological data, Moratelli et al. (2019a) recognized four groups of species. Among them, the lucifugus and vivesi groups are restricted to the Nearctic region; the albescens-group is predominantly neotropical, but with a few nearctic species; and the ruber-group is restricted to the Neotropics.
During a taxonomic revision of neotropical Myotis, we identified a sample of 15 specimens from the Uruguayan Pampa grasslands with a unique set of qualitative and quantitative characters that distinguish them from all South American congeners currently recognized. Based on the morphological discontinuity, we hypothesize that this sample comprises a new species, which we describe and compare with neotropical congeners.

Methods
This research is part of a critical review of neotropical Myotis, and material examined comprises more than 7,500 specimens covering all neotropical species currently recognized, including their primary types, and comparative series from different localities. The 15 specimens from the Uruguayan Pampa grasslands that represent the putative new species were collected in 1963, and are deposited at the American Museum of Natural History, USA. Due to the absence of fresh tissue samples or DNA sequences available in biological collections and DNA libraries, respectively, we restricted our analyses to qualitative and quantitative morphological comparisons.
Comparative analyses were based on 15 specimens of the putative new species of Myotis from Uruguay, and other 720 specimens from 15 neotropical Myotis species deposited in 19 scientific collections from Argentina, Brazil, Canada, France, Switzerland, and United States (Appendix 1). Species selected for comparisons include all species in the ruber-group (M. armiensis, M. elegans, M. keaysi, M. midastactus, M. pilosatibialis, M. riparius, M. ruber, M. simus), and those representing the albescens-group that occur in the Southern Cone (M. albescens, M. chiloensis, M. dinellii, M. izecksohni, M. lavali, M. levis, and M. nigricans).
The intra and interspecific morphological variation was evaluated from qualitative and quantitative analyses based on adults (classified as such based on closed epiphyses; see Brunet-Rossini and Wilkinson 2009). Qualitative traits employed here to characterize and distinguish species follow Moratelli et al. (2013) and Novaes et al. (2021). Capitalized color nomenclature follows Ridgway (1912).
Discriminant Function Analysis (DFA), performed in the R software (R Development Core Team 2008), with the Mass (Vanables and Ripley 2002) and Lattice (Sarkar 2008) packages, was used in order to characterize and discriminate samples. We selected a subset of the cranial dimensions (GLS, CIL, MAB, BCB, POB, IOB, BAC, BAM, MTL, M1-3, MAN, MAL) representing different axes of length and width of skull, rostrum, and mandible. Multivariate analyses require complete datasets, thus, the missing values (< 5% of total dataset) were estimated from the existing raw data using the Amelia II package (Honaker et al. 2011) implemented in R software. Measurements were transformed to natural logs and covariance matrices were computed considering all variables. DFA was run with 15 selected individuals of M. riparius, M. ruber, and Myotis sp. nov. from Uruguay, considering that these species have similar morphology and occurs in the Southern Cone.

Results
The sample from Uruguay that represents the putative new species (hereafter refered as Myotis sp. nov. from Uruguay) can be distinguished from all South American congeners, except M. riparius and M. ruber, by the following set of traits: small to medium sized species .3 mm; GLS 13.4-14.6 mm); comparatively long wooly dorsal fur subtly bicolored; skull with very low sagittal crest and braincase roof not inclined forward. Although M. ruber has a overall larger size than Myotis sp. nov from Uruguay (Table 2), the smallest specimens of M. ruber may resemble the largest specimens of Myotis sp. nov. Table 1. External and craniodental measurements taken from Myotis specimens.
In the DFA, the 1st discriminant function, more related with the skull size, represented 93% of the total variation, whereas the 2nd discriminant function, more related with the skull shape, represented 7% (Fig. 3). Along the 1 st axis, M. riparius and M. sp. nov. overlapped, indicating similar skull sizes, and both do not overlap with M. ruber (Fig. 3). Myotis sp. nov. from Uruguay and M. riparius were retrieved with full overlap along the first axis (related to skull size), but there was a partial segregation of these taxa along the second axis (more related to skull shape). This analysis relied mainly on measurements related to skull and mandibular length, and braincase breadth (Figure 3). Linear measurements reveal that Myotis sp. nov. is slightly smaller than M. riparius in relation to both skull and external measurements and consistently smaller than M. ruber ( Table 2).
The set of qualitative morphological characters and the partial distinction of the skull shape evidenced by the

Distribution and habitat.
Known from only two localities in the provinces of Artigas and Tacuarembó, Uruguay (Fig. 6). The records come from the Uruguayan Pampas, between 30 and 240 m in altitude. The region is characterized by grassland plains with shrubby vegetation patches typical of subtropical landscapes. However, the margins of rivers can present denser riparian forests, with arborescent ferns, orchids, lianas and several trees with more than 10 m in height (Sganga 1994;Panario et al. 2011). According to information available on museum labels, most individuals were captured with mist-nets near watercourses.

Etymology.
We name Myotis pampa after the Pampas ecoregion, the main habitat occurrence for the species. Pampa is a Quechua word for "plain", and in Latin languages it is used in the masculine gender, which agrees with the generic name Myotis, also masculine (ICZN 1958;Pritchard 1994).
Diagnosis. The tricolored ventral fur, with a Drab basal band, Smoke Gray middle band, and yellowish (generally Naples Yellow) tip is unique among South American Myotis. In addition, the following set of traits distinguishes Myotis pampa from all its South American congeners: small to medium sized species (FA 33.8-36.3 mm; sagittal crest absent or very low; elongated skull (GLS 13.4-14.4 mm); braincase not inflated and low in profile; braincase roof formed by the parietal bone strongly straight and not inclined forward; frontal bone steeply sloped towards the rostrum; posterior region of the braincase rounded and projected beyond the limit of the occipital condyles; mastoid processes narrower and practically not visible in dorsal view; dorsal fur notably long (7-9 mm), woolly and bicolored, with tips Dresden Brown or paler Ochraceous-Tawny, and bases brownish-gray (giving a fuliginous aspect); ventral fur moderately long (6-7 mm); legs and dorsal surface of the uropatagium covered by scattered fur that extends up to the knees; plagiopatagium inserted into the foot by a broad band of membrane.
Morphological description and comparisons. Myotis pampa is a small to medium species of Myotis .3 mm; other measurements in Table 2) and the fur texture (wooly) and cranial morphology (broad skull with sagittal and lambdoidal crests presents) resembles species allocated to the ruber-group (q.v., Moratelli et al. 2013Moratelli et al. , 2019a. Ears are Mummy Brown and comparatively medium sized (EL 11-14 mm), reaching the portion of the rostrum between the eyes and nostrils when extended forward. Tragus is long and slender, with a wide base and a narrower spear-shaped terminal half, almost straight anterior edge, and rounded tip. Membranes are Mummy Brown; the dorsal surface of tibia, and uropatagium has scattered hairs extending to the level of the knees. The uropatagium lacks the fringe of hairs along the trailing edge. The plagiopatagium is attached to the foot at the level of the toes by a broad band of membrane.
Dorsal and ventral fur are wooly and long (LDF 7-9 mm, LVF 6-7 mm). The dorsal pelage is bicolored with brownish-gray bases (3/4 of the total fur length) and reddish tips (1/4 of the total fur length), ranging from Dresden Brown to paler Ochraceous-Tawny. The dorsal fur has a coloration that looks like the bat is partially covered by soot and with the tip of the hairs burned. The ventral fur is tricolored, with Drab bases (1/4 of the total fur length) progressively turning Smoke Gray in the middle band (2/4 length) and light-yellow tips (1/4 [ranging from Naples Yellow to Cream-Buff]). The tricolored ventral fur is unique among all South American Myotis species (Fig. 2). However, this pattern may be less evident in some lighter specimens, where the coloration of the middle band may be confused with the tip band of the hairs.
Dental formula is I 2/3, C 1/1, PM 3/3, M 3/3 = 38, typical of all New World species of Myotis. Skull delicate and medium-sized in length (GLS 13.1-14.1 mm), resembling Myotis of the ruber-group species. The 2nd upper premolar (P3) is in the toothrow, not displaced to the lingual side and smaller than first upper premolar (P2). The 1st lower molar (m1) is myotodont, with the postcristid connecting hypoconid and entoconid. Braincase delicate and elongated; sagittal crest and lambdoidal crests are absent or very low; the occipital region is rounded and projects beyond the posterior limits of the occipital condyles; mastoid processes weakly developed. Frontal bone slightly sloping; rostrum comparatively short and narrow.
These specimens of M. pampa were originally identified as M. riparius and this identification was maintained by LaVal (1973) and López- González et al. (2001). In fact, M. pampa resembles M. riparius, but it can be distinguished by its smaller skull (GLS < 14.4 in pampa and generally > 14.5 mm in riparius), braincase lower in lateral view, sagittal crest lower (Fig. 1), longer and bicolored dorsal fur with grayish base, tricolored ventral fur (Fig. 2), and presence of scattered hairs along the leg and dorsal surface of the uropatagium. Myotis riparius have a higher braincase in lateral view, sagittal crest well-developed, shorter and unicolored dorsal fur, bicolored ventral fur and absence of scattered hairs along the leg and dorsal surface of the uropatagium.
Myotis pampa can also be confused with M. ruber, but it can be easly distinguished by its smaller size (both external and cranial; FA > 37.5 mm and GLS > 14.0 mm in M. ruber), skull more delicate, braincase lower in lateral view, mastoid processes less developed, dorsal fur longer and clearly bicolored with grayish base (unicolored in ruber), tricolored ventral fur (bicolored in ruber), and presence of scattered hairs along the leg and dorsal surface of the uropatagium (absent in ruber).
Myotis pampa differs from M. armiensis, M. keaysi, and M. pilosatibialis by its external smaller size (FA > 36 mm in M. armiensis and M. keaysi), more delicate skull, braincase lower in lateral view, sagittal crest lower, mastoid processes less developed, dorsal fur longer and clearly bicolored with grayish base, tricolored ventral fur, and pelage on the dorsal surface of the uropatagium composed of scattered hairs not extending beyond the knees. It differs from M. elegans by larger size in general (both external and cranial; FA < 34.5 mm, GLS < 13.0 mm, and BCB < 6.3 mm in elegans), skull more robust in general, parietal bone not inclined forward, rostrum longer, dorsal fur clearly bicolored with grayish base, ventral fur tricolored. Myotis pampa can be easily distinguished from M. simus and M. midastactus by its smaller size (generally FA > 36 mm and GLS > 13.5 mm in M. simus and M. midastactus), more delicate and narrower skull (BCB < 6.8 in pampa and > 6.8 in simus and midastactus), dorsal fur longer and clearly bicolored, ventral fur tricolored, dorsal surface of the uropatagium covered by scattered hairs that extend up to the knees; plagiopatagium inserted into the foot by a broad band of membrane.
In comparison to species from the albescens-group, M. pampa can be distinguished from M. albescens, M. dinellii and M. levis by the absence of a fringe of hairs on the posterior margin of the uropatagium, reddish dorsal fur bicolored with grayish base, and braincase less inflated. In addition, M. pampa differs from M. dinellii and M. levis by its generally smaller size and comparatively shorter ears. It differs from M. izecksohni and M. nigricans by the braincase less inflated and reddish dorsal fur clearly bicolored with grayish base. Myotis pampa can be distinguished from M. chiloensis from its smaller size in general, skull lower in lateral view, braincase less inflated, dorsal fur shorter and bicolored, with grayish bases and reddish tips. It differs from M. lavali by its second upper premolar (P3) smaller than first upper premolar (P2), and by its reddish fur with grayish bases (being blackish bases in M. lavali). In addition, M. pampa can be distinguished of all species from the albescens-group by its wooly fur (silky in albescens-group species), ventral fur tricolored, and fur composed of scattered hairs on dorsal surface of the uropatagium.

Discussion
The specimens used for the description of M. pampa were collected in 1963 during two expeditions in the Uruguayan Pampas, remaining identified as M. riparius for more than half a century. This highlights the importance of taxonomic studies based on the review of specimens in museums for the knowledge of the species diversity on Earth.
Myotis pampa occurs near the boundaries of Uruguay with Brazil (Rio Grande do Sul) and Argentina (Entre Ríos) and although individuals from grasslands of these two countries have not been examined, it is likely that the occurrence of M. pampa will extend to outside the political boundaries of Uruguay. We believe that M. pampa is endemic to the Pampa ecoregion, inasmuch as our taxonomic review of Myotis-which comprises more than 7,500 specimens examined-did not find evidence of the species in other South American ecoregions. If our hypothesis is correct, it is possible that M. riparius (sensu Novaes et al. 2017), does not occur in the Pampa grassland plain.
The localities where M. pampa occur are within the Pampas biome, an ecoregion with high biological diversity, with more than 3,000 species of plants, ca. 90 species of amphibians, 160 species of reptiles, 100 species of mammals, and more than 100 species of birds, including a very high rate of endemism (Pillar et al. 2016). However, it remains as one of the least studied ecoregions in relation to the bat fauna (Bernard et al. 2011). Currently, the Pampas has suffered from the intensification of agricultural production, such that at least 30% of the entire biome has been converted into pastures and agricultural areas and there are very few protected areas (Overbeck et al. 2007;Pillar et al. 2016;Achka 2017;Botto-Nuñez et al. 2019;Souza et al. 2020). Although it is premature to determine the conservation status of M. pampa, this possibly endemic species should be monitored due to the anthropogenic pressure.
The present taxonomic revision indicates the unequivocal occurrence of three Myotis species for Uruguay: M. albescens, M. levis, and M. pampa. All specimens of M. riparius from Uruguay analyzed in the present study were re-identified as M. pampa and, therefore, there is no evidence of the occurrence of this species in the Uruguayan Pampas. Recent literature points to the occurrence of M. nigricans in Uruguay (Achaval et al. 2007;González and Martínez-Lanfranco 2010;Botto-Nuñez et al. 2019). Myotis nigricans has been considered a species complex with a taxonomic puzzle far from being solved (e.g., Larsen et al. 2012;Moratelli et al. 2011Moratelli et al. , 2013Moratelli et al. , 2017Moratelli et al. , 2019a. In the present review, we did not identify any specimen that fit morphologically into what is known as the M. nigricans complex from southern Brazil and Southern Cone (sensu Moratelli et al. 2011). Likewise, we found no evidence of the occurrence of M. ruber in Uruguay and previous records (e.g., Acosta y Lara 1950) are due to misidentified specimens of M. levis, as also suggested by Botto-Nuñez et al. (2019). However, we do not rule out the possibility of the occurrence of other Myotis species in Uruguay, considering that (i) we did not have the opportunity to examine specimens deposited in Uruguayan collections (e.g., Museo Nacional de Historia Natural del Uruguay), and (ii) a large part of the territory does not have satisfactory samples for bats (González and Lessa 2014).