Emended diagnosis of Xeronycteris vieirai ( Mammalia : Chiroptera ) , with the first record of polyodontia for the genus

The Caatinga-endemic nectar-feeding bat, Xeronycteris vieirai Gregorin & Ditchfield, 2005 was described based on four specimens. Since then, only two additional specimens have been reported in the literature. Examination of a new specimen that closely agrees with the original description but presents two additional upper premolars, led us to review the type series of this taxon. Our analysis provided support to the recognition of the new specimen as a X. vieirai with supernumerary teeth and revealed new diagnostic characters that can help in the field identification of this species. Xeronycteris vieirai presents tricolored dorsal fur, entirely naked forearm, connection of the base of the spear of the noseleaf with the upper limit of the horseshoe with a marked ridge, and chin with simple dermal pads and a relatively slight cleft. Additional new characters described here include anterior zygomatic arches reduced and not extending laterally and upward with respect to upper toothrow, basioccipital pits deep and separated by a high and thin bone septum, and mandible with a pronounced ridge at anterior symphysis. We propose an emended diagnosis based on morphological characters and provide a rectification on the original notation related to the holotype.

The type series of X. vieirai is composed of four specimens, as follows: an adult female (MZUSP 14170) and a subadult male (MZUSP 14173) obtained in 1973 in the municipality of Cocorobó, state of Bahia; an adult female (MZUSP 14804) collected in 1978 in the municipality of Exu, state of Pernambuco (see below for the status of this specimen); and an adult male (MZUSP 29777), assigned as the holotype, collected in 1993 in the municipality of Soledade, state of Paraíba (GREGORIN & DITCHFIELD 2005).As revealed by original notations in tags, MZUSP 14170 and 14173 were first referred to as Lonchophylla bokermanni Sazima, Vizotto & Taddei, 1978.More recently, additional material of X. vieirai was reported in the literature, bringing a new record for the state of Pernambuco (ASTÚA & GUERRA 2008) and the first mention of the species for the states of Sergipe (ASTÚA & GUERRA 2008) and Minas Gerais (NOGUEIRA et al. 2007).
Upon examining the specimen from the region of Jaíba, state of Minas Gerais (NOGUEIRA et al. 2007), we noticed that it closely agrees with the original description of X. vieirai, differing from it by the presence of additional upper premolar.Aiming to consistently access the taxonomic identification of this specimen, we directly compare it with the type series of X. vieirai.Our analysis confirmed the first case of polyodontia for this taxon, and led us to propose an emended diagnosis that will further help to recognize this endemic bat even in the field.Because the original description lacks a detail account of external characters, our emended diagnosis is preceded by a revised description of those morphological traits, and by additional comments on craniodental features and measurements.

MATERIAL AND METHODS
We examined specimens (see below) housed at the following institutions: American Museum of Natural History (AMNH), New York; Coleção Adriano Lúcio Peracchi (ALP), Instituto de Biologia, Universidade Federal Rural do Rio de Janeiro; Museu Nacional, Universidade Federal do Rio de Janeiro (MN); and Museu de Zoologia da Universidade de São Paulo (MZUSP).Specimens from the type series of X. vieirai are de-

Emended diagnosis of Xeronycteris vieirai (Mammalia: Chiroptera),
with the first record of polyodontia for the genus Marcelo R.  Emended diagnosis.A medium-sized Lonchophyllinae (mean forearm 37 mm) with tricolored dorsal fur (cream-white in a basal narrow stripe, pale brown in the middle, and darker brown distally); naked forearm; base of spear of noseleaf, at internarinal region, inserted behind upper limit of horseshoe; lower border of horseshoe entirely fused to face, with no free edge associated with upper lip; lower lip with pair of simple (not scalloped) dermal pads strait along its extension; chin with relatively slight cleft present in midline; foot longer than calcar; thumb relatively short; rostrum slightly longer than braincase and conspicuously tapered anteriorly; zygomatic arch incomplete and reduced to thin projection that follows labial outline of the upper toothrow; medial palate deeply concave and bearing pits lingually adjacent to each molar and, occasionally, also to premolars, allowing insertion of main cusps at lingual side in lower dentition; posterior border of medial palate well-developed; posterior palate concave at midline, and with a broadly U-shaped posterior margin reaching posterior border of optic foramen; glenoid fossae poorly developed; basioccipital pits markedly concave and separated by a thin but high septum; ridge at anterior symphysis of mandible welldeveloped; upper premolars (P3 and P4) lacking cingulum; parastyle anteriorly very extended in M1 and M2; upper molar series extremely reduced, with oclusal plane progressively more lingually oriented from M1 to M3, and mesostyle, metastyle, preprotocrest, postprotocrest, and ectoflexus all missing or greatly reduced, resulting in the virtual absence of a trigon basin; lower dentition reduced with narrowness of premolars, and loss of hypoflexid, metacristid, entocristid, and entoconid.back); dorsal hairs slightly tricolored (cream-white in a basal narrow stripe, pale brown in middle, and darker brown distally), with external layer frosted at tips.Ears, noseleaf and uropatagium drab in most specimens, but darker in holotype (MZUSP 29777); wing membrane dark brown in all specimens.Craniodental features.The ALP 6824 departs from all specimens in the type series of X. vieirai (dental formula I2/2, C1/1, P2/3, M3/3 x 2 = 34) by presenting one additional upper premolar on each side of the arcade (total of 36 teeth; Fig. 1).This extra upper premolar (assumed to be that after the canine) is smaller in antero-posterior length than the two regular premolars (P3 and P4), but similar to these teeth in presenting small anterior and posterior accessory cusps.In specimens from the type series, P3 is separated from canine by a larger diastema than that between P3 and P4.In ALP 6824, however, these diastemata are similarly reduced (Fig. 1), which might be a consequence of the presence of the extra tooth.Aside from these differences, craniodental characters in ALP 6824 are generally in agreement with those in the type specimens of X. vieirai.The following characters were not included in the original description of this taxon, but are shared by all X. vieirai we were able to examine: zygomatic arch incomplete and reduced to thin anterior projection that follows labial outline of upper toothrow (broken in most specimens but present on left side of holotype and in ALP 6824); glenoid fossae poorly developed; basioccipital pits markedly concave and separated by thin but high septum; ridge at mandibular symphysis prominent in most specimens (reduced in MZUSP 14170), projected beyond level of lower incisors, and conferring an arched anteriormost profile to ramus in lateral view (Fig. 1); coronoid process broken in most specimens, including holotype, but based on ALP 6824 can be characterized as relatively low; angular process prominent forming posteriormost point of the mandible.As reported in the original description, the medial palate of X. vieirai is strongly arched and bearing pits lingually adjacent to each molar.Our analysis, however, showed that these pits can also occur adjacent to premolars.The MZUSP 14170 present pits adjacent to P2 and P3, but only in one side of the arcades.
Comparisons of individual measurements.The specimen of X. vieirai from Jaíba (ALP 6824) is an adult female preserved in alcohol; the skull has been removed and cleaned.The measurements are, in general, very close to those reported here for the two adult females in the type series of X. vieirai (Table I).Upper and lower toothrows, however, are slightly larger in ALP 6824.When compared to specimens in the type series of L. bokermanni, X. vieirai can be promptly differentiated by its shorter forearm, narrower braincase, and longer mandible (Table I).
Remarks.While analyzing the specimens from the type series of X. vieirai we noticed that the holotype (MZUSP 29777), previously reported as a male (see "Type series" section, page 405, and table 2, page 409, both in the original manuscript), is in fact an adult female.In the original tag, where this speci-men is assigned under the collection number 186020, it is also erroneously reported as a male.The only male available from the type series is, therefore, MZUSP 14173, which is a subadult.

DISCUSSION
The maximum number of teeth recorded to date for a phyllostomid bat is 34 (KOOPMAN 1994).Therefore, the specimen from Jaíba departs from all described dental formulae in the family.Here we recognize the condition seen in this bat as a congenital dental abnormality.We based our assessment on a morphological analysis showing congruence in diagnostic characters, and on the historical records of dental abnormalities in nectar-feeding bats (PHILLIPS 1971, LANZA et al. 2008).PHILLIPS (1971) assessed the prevalence of supernumerary teeth among glossophagines and lonchophyllines, and found extra first upper premolars in specimens of Glossophaga soricina (Pallas, 1766) and Leptonycteris sanborni Hoffmeister, 1957, but not in Anoura geoffroyi Gray, 1838.The first two species are similar to X. vieirai in presenting two upper premolars as their normal condition, while A. geoffroyi usually presents three upper premolars.We follow PHILLIPS (1971) in recognizing the extra premolars of ALP 6824 as resulting from a genetically determined double initiation in the process of formation of the first upper normal premolars (P3).Such origin is indicated by the presence of two roots in the extra premolars, and by a similarity in size and shape between these teeth and the normal first premolar (PHILLIPS 1971).A recent survey revealed that polyodontia (presence of extra teeth) is rare among bats, occurring in only 8% of the more than 160 described phyllostomid species (LANZA et al. 2008).
Among the several new characters described here for X. vieirai, some are of particular interest, due to their diagnostic value.The tricolored banding pattern in the upper back fur of X. vieirai seems to be unique among lonchophyllines (WETTERER et al. 2000), and the general brown body pelage is similar to that seen in L. peracchii, but clearly distinct from the paler, ventrally beige L. mordax, and the grayish L. bokermanni.The entirely naked forearm of Xeronycteris is similar to that of L. mordax and Platalina, but it can be used to differentiate this bat from L. bokermanni and L. peracchii, in which the forearm is hairy on the proximal half.Additional external features that may help to discriminate Xeronycteris from L. bokermanni include: the connection of the base of the spear of the noseleaf with the upper limit of the horseshoe, which is smooth in the latter species (and in L. mordax), while it is marked by a ridge in X. vieirai; the shape of dermal chin pads, which are simple in X. vieirai vs. scalloped in L. bokermanni (apparently extensive to whole genus - WETTERER et al. 2000); the presence of a relatively slight cleft on the chin (just after the dermal pads), which is absent in L. bokermanni (again this latter condition seems to be extensive to whole genus -WETTERER et al. 2000); and the relative length of the thumb (shorter in X. vieirai ) and noseleaf spear (shorter in L. bokermanni).When cleaned skulls and mandibles are available, the following additional characters are noteworthy: anterior zygomatic arches reduced and not extending laterally and upward with respect to upper toothrow, as in other lonchophyllines; basioccipital pits deeper and separated by higher and thinner bone septum than in L. bokermanni, L. mordax, and L. peracchii; and ridge at anterior symphysis more pronounced than in other lonchophyllines, extending well beyond the lower incisors.
We have emphasized here comparisons with L. bokermanni because in eastern Brazil this species is the most likely to be confused with X. vieirai.As revealed by original notations in specimens tags, the paratypes of X. vieirai MZUSP 14170 and 14173 were first referred to as L. bokermanni, which might explain the erroneous record of this latter species from Cocorobó, Bahia -GRIFFITHS & GARDNER (2008) based this record on the MZUSP 14170 -.As currently accepted, L. bokermanni is restricted to the state of Minas Gerais, southeastern Brazil (DIAS et al. 2013), and its type series came from an area of Cerrado less than 500 km from the meridional limit in the distribution of X. vieirai (NOGUEIRA et al. 2007).These two endemic species are listed as threatened in regional lists (e.g., PAGLIA et al. 2009), but according to the last IUCN assessment they are still datadeficient, requiring more information for a reliable determination of their conservation status (SAMPAIO et al. 2008, DITCHFIELD & GREGORIN 2008).In either case, these species should be focus of ecological and natural history investigations, ideally funded on a clear recognition of their taxonomic identity, particularly under field conditions, as supported by data presented here.
*Measurements from original tag.