OUP user menu

New Genus and Species of Nectar-Feeding Bat in the Tribe Lonchophyllini (Phyllostomidae: Glossophaginae) from Northeastern Brazil

Renato Gregorin , Albert David Ditchfield
DOI: http://dx.doi.org/10.1644/BRB-229.1 403-414 First published online: 15 April 2005

Abstract

A new genus and species in the tribe Lonchophyllini, subfamily Glossophaginae (Chiroptera: Phyllostomidae) is described based on the analysis of 4 specimens collected in 3 different localities in a semiarid area of northeastern Brazil. The presence of a groove and filiform papillae on the lateral side of the tongue places this taxon as a member of the tribe Lonchophyllini. Fourteen cranial and dental characters distinguish the new genus from all other genera in the tribe (Lonchophylla, Lionycteris, and Platalina). The new genus is characterized by the presence of inward-facing molars on the palate; the loss of the cingulum in P3 and P4; the reduction or lack of the metastyle and preprotocrest in Ml, ectoflexus, preprotocrest, and postprotocrest in M2, and postprotocrest in M3; and loss of hypoflexid, metacristid, entocristid, and entoconid on m 1–3. A wider gap also occurs between lower incisors and canines than in other genera of Lonchophyllini. The palate in the new genus presents an unusual morphology. The palate of members of this genus is the longest among lonchophyllines and has 4 pits that permit the insertion of the main cusps of the lower last premolar and all 3 lower molars. Cladistic analysis suggests that the new genus is sister to Platalina.

Key words
  • Brazil
  • Chiroptera
  • Lonchophyllini
  • systematics
  • Xeronycteris vieirai

The subfamily Glossophaginae (Chiroptera: Phyllostomidae) includes bats that feed primarily on nectar, pollen, and petals, as well as on fruit and insects to some extent (Gardner 1977; Gimenez et al. 1996; Solmsen 1998). All representatives of this subfamily possess characteristics strictly related to nectar-feeding habits, including an elongated rostrum, reduction of dental structures, gaps between teeth, shortened pinnae and noseleaf, less distinct glenoid fossae, and a long and extensible tongue (Carstens et al. 2002; Freeman 1995; Gimenez et al. 1996; Greenbaum and Phillips 1974; Griffiths 1982; Howell and Hodgkin 1976). Genera that compose the Glossophaginae are a subject of some controversy. According to most authors, Glossophaginae is composed of 15 genera, usually subdivided into 3 distinct groups: Phyllonycteris and Erophylla (also known as phylonycterines); Anoura, Choeroniscus, Choeronycteris, Glossophaga, Hylonycteris, Leptonycteris, Lichonycteris, Monophyllus, Musonycteris, and Scleronycteris (glossopha-gines); and Lonchophylla, Platalina, and Lionycteris (lonchophyllines).

The relationships of the Lonchophylllatalinionycteris clade within Glossophaginae and with other Phyllostomidae are disputed. Griffiths (1982), followed by Koopman (1993, 1994), considered the clade distinct from typical glossophagines and created a new subfamily, Lonchophyllinae, despite some hyoidean and muscle characters shared between both. Griffiths (1982) considered the characters that unite Lonchophyllinae and Glossophaginae of independent origin, or homoplasies representing convergence for nectivory. Alternatively, some authors (e.g., Miller 1907; Wetterer et al. 2000) consider Lonchophylllatalinionycteris as a derived but typical group within Glossophaginae (= tribe Lonchophyllini). Wetterer et al. (2000) employed cladistic analyses, predominantly using morphological data but combining some molecular and karyotypic data, and their tree represents a good phylogenetic hypothesis for available morphological data. Nevertheless, Baker et al. (2003) employed molecular data and present a technically more robust hypothesis than that of Wetterer et al. (2000), which differs in some crucial aspects. However, the close evolutionary relationship among Lonchophylla, Platalina, and Lionycteris is the consensus among most authors (Gimenez et al. 1996; Griffiths 1982; Haiduk and Baker 1982; Koopman 1993), including both Wetterer et al. (2000) and Baker et al. (2003). This clade is supported by several synapomorphies, including hyoidean, muscle, tongue, and molecular characters. Regardless of the methodology and the number of synapomorphies that diagnose the group, the tribe Lonchophyllini is monophyletic. Among the various shared derived characters for the 3 taxa are the presence of a groove and a line of filiform papillae on the lateral side of the tongue (see Gimenez et al. 1996; Griffiths 1982). Enlarged and spatulate inner upper incisors with the outer upper incisors small and peglike also define the clade. Within the clade, each of the 3 genera also is clearly defined (Koopman 1994). Because we are employing morphological data, we will adopt the hypothesis and character descriptions of Wetterer et al. (2000).

The genus Lonchophylla includes 8 species (L. hesperia, L. concava, L. robusta, L. handleyi, L. thomasi, L. mordax, L. dekeyseri, and L. bokermanni) and is distributed from Nicaragua to southeastern Brazil. The other 2 genera are monotypic, with Platalina genovensium being restricted to open dry areas of coastal Peru, and Lionycteris spurrelli occurring from Panama to southeastern Brazil. Five species of Lonchophyllini are known to occur in Brazil, including 4 belonging to the genus Lonchophylla. The species L. dekeyseri and L. bokermanni initially were described from the open grasslands of central Brazil and the latter occurs in forested areas (Taddei et al. 1978, 1983), whereas L. mordax and L. thomasi are distributed broadly in forested and open areas. Lionycteris spurrelli, usually associated with wet forest (Handley 1976), has been found in dry and open regions in eastern Brazil (Gregorin and Mendes 1999; Trajano and Gimenez 1998).

In this paper we report the discovery of a new form of the tribe Lonchophyllini collected in northeastern Brazil. The characters exhibited by the specimens of this new form suggest that it clearly is not a species of Lonchophylla, Lionycteris, or Platalina, but a new genus closely related to Platalina. We define the new taxon with comparisons to the relevant taxa and propose a hypothesis of its relationships to the other members of the subfamily.

Materials and Methods

Specimens examined.—We examined 77 specimens representing 11 genera and 14 species of glossophagines (Appendix I). For comparative purposes, we examined 25 specimens that included relevant genera of Glossophagini and a genus of Phyllonycterini. The remaining 52 specimens belong to Lonchophyllini and include all species occurring in Brazil and Platalina from Peru. In addition we have compiled data from the literature for species of Lonchophylla that occur outside Brazil (Appendix I). The specimens examined were all adults, which we defined as individuals with skull and digital symphysis fully calcified. We attempted to examine representative collections for all species involved, by sampling both sexes for as wide a portion of the range as possible.

Animal care and use followed standard procedure as recommended by the guidelines of the American Society of Mammalogists (http://www.mammalogy.org/committees/index.asp). Collecting permits for all animals collected in Brazil were issued by the Instituto Brasileiro de Meio Ambiente e dos Recursos Renováveis.

We employed both qualitative and quantitative characters. Weight was recorded from a freshly obtained specimen and is given in grams. All measurements obtained are given in millimeters. Measurement procedure was based on standard bat body and skull measurements such as those used by Kalko and Handley (1994) and Vizotto and Taddei (1973). Some measurements were designed specifically for the taxa at hand, although biometrie characters were assessed in order to define range of size but not to cover phylogeny or to diagnose taxa because overlap is significant. Measurement acronyms and procedures are given in Appendix II.

Character description.—Phylogenetic relationships were assessed by using morphological characters, including some of the characters that were important for understanding the relationships among genera of glossophagines as listed elsewhere (Wetterer et al. 2000). Qualitative characters include internal and external anatomy, with dental, skull, skin, and tongue characters obtained from fluireserved and dry (skin and skull) specimens. Additional observations were made with scanning electron microscopy, particularly for dentition, to verify character states assigned optically with a morphological tool of much higher resolution. Primary homologies for characters were proposed by considering the topographical criterion (e.g., de Pinna 1991), and Arabic numbers were employed to denote character states (Appendix II). Missing or noncomparable data were designated by a question mark (?). Estimates of phylogeny relationships for morphological characters were performed by using the Hennig86 computer program (Farris 1989). Additional analyses were performed with PAUP* software (Swofford 2001). Maximum parsimony was considered with the outgroup not previously defined (Farris 1982; Nixon and Carpenter 1993), and characters were coded both as unordered and ordered, with the former being used for character optimization. All characters were given equal weighting, and both ACCTRAN (ACCelerated TRANsformation) and DELTRAN (DELayed TRANsformation) optimization criteria were employed. ACCTRAN assumes that reversals are more likely than convergence, whereas DELTRAN optimization assumes convergences are more likely than reversals. A strict consensus tree was employed to study and discuss character differentiation and in diagnosing the genera of Lonchophyllini. Robustness and nodal support of all parsimony trees were evaluated with standard bootstrap heuristic analyses of 1,000 replicates (Felsenstein 1985). The starting trees were obtained via stepwise addition and the brancwapping algorithm was TBR (treisectioeconnection). No topological constraints were enforced.

Taxon sampling.—The choice of outgroup is a critical element in phylogenetic analysis (Griffiths 1982,1983; Smith and Hood 1984). Most authors agree that the tribe Lonchophyllini is monophyletic, defined by at least 2 lingual characters: presence of a lateral groove and a fringe of papillae along the side of the tongue. Morphological examination of the specimens of the new taxon demonstrated that it does possess these defining characters. Therefore, initially the taxon was considered a member of Lonchophyllini, although a more inclusive analysis could possibly falsify this assumption. Therefore, other glossophagine genera were used as outgroups for establishing the relationships of the new taxon to other lonchophyllines. The following genera were used as outgroups: Erophylla to represent the phyllonycterines; Glossophaga; Monophyllus; Anoura; Leptonycteris; and Choeronycteris to represent the clade Choeronycteris, Choeroniscus, and Musonycteris.

Results

Two maximum parsimony analyses were performed. First, 6 taxa were used as outgroups, including Erophylla, a genus considered by Wetterer et al. (2000) as a distinct subfamily; and 2nd, members of the Glossophaginae were used. This was done because Erophylla shows reduction or loss of a series of dental elements similar in pattern to that of the new taxon, although the shape of teeth in these taxa is clearly different, and the potential homology seems weak. The 1st analysis resulted in 11 equally parsimonious trees (length = 64, consistency index [CI] = 0.62, and retention index [RI] = 0.61). In all, the monophyly of Lonchophyllini was maintained, with variation restricted to different rearrangements within the outgroup taxa. The consensus tree had a length of 70, CI = 0.57, and RI = 0.50 (Fig. 1). Nine synapomorphies united all taxa, including characters 1 (state 1 [character states are defined in Appendix II]), 2 (state 1), 3 (state 1), 4 (state 1), 5 (state 1), 11 (state 1), and 27 (state 1). The analysis excluding Erophylla resulted in 3 equallarsimonious trees (length = 58, CI = 0.70, and RI = 0.65), with a consensus tree similar to that obtained in the 1st analysis. This tree also resulted in monophyly for Lonchophyllini. In addition, both analyses displayed branching sequence within Lonchophyllini, with Lionycteris more basal, and the new taxon and Platalina the more derived taxa. The characters that support both monophyly of the tribe and clades within Lonchophyllini are the same in each analysis.

Fig. 1

Consensus tree of the tribe Lonchophyllini resulting from 11 equally mosarsimonious cladograms. Numbers below branches represent bootstrap support.

As expected, the 1st analysis reflected that the reductions of dental structure in the new taxon and Platalina are homoplasic in Erophylla, and to a lesser degree in Choeronycteris, such as characters 16 (state 0), 20 (state 0), 21 (state 0), and 30 (state 1). An exhaustive search with parsimony as the optimality criterion and 6 outgroup taxa including Erophylla also yielded 11 trees. The strict consensus tree obtained was identical in topology to the tree obtained with the 1st Hennig86 search. The bootstrap 50% majoritule consensus tree was identical to the strict consensus tree obtained above. The monophyly of the lonchophyllines is upheld with a 63% bootstrap support value, whereas Lonchophylla was sister to the the new taxon-Platalina clade (55%). The placement of the new taxon as the sister taxon to Platalina is supported by a bootstrap value of 98% and 8 derived characters, of which the following are autapomorphic to both genera: no ribs (character 6, state 1), upper incisors parallel to saggital plane of skull (character 13, state 1), and entoconid on molars lacking (character 33).

Discussion

Although Platalina was sister taxon to the new taxon in all analyses, the recognition of a new genus for the species described herein requires justification. The generic concepts of the 3 previously described taxa (their diagnoses) indicate that they are defined by a small number of autapomorphies, but more commonly by an exclusive set of characters, including plesiomorphic and derived characters (see diagnoses of groups based on Koopman [1994]). The new taxon described here presents 14 apomorphic characters, making it the most derived taxon within lonchophyllines. These unique characters include dentition and skull morphological complexes. Also, to reinforce the decision to recognize a new genus, the set of characteristics that are of current use in diagnoses for the genera indicates that the new taxon shares some characters with all 3 genera, making it difficult to place it in any 1 genus. To include the new taxon in 1 of 3 previously described genera would require a complete redefinition of any genus that includes the new genus and species. However, if we consider only characters that have proved to be derived in this analysis, the new genus and species shares a considerable number of characters with Platalina. However, the species Xeronycteris vieirai is diagnosed by 14 autapomorphies, which would completely and radically change the current definition of Platalina if it were to be included therein. Therefore, taking into account the very distinct morphology of the new genus and species and wishing to avoid a drastic redefinition of all Lonchophyllini, we suggest allocating the species described to a new genus.

Systematics (following Wetterer et al. 2000)

Family Phyllostomidae Subfamily Glossophaginae Tribe Lonchophyllini.

Xeronycteris vieirai, new genus and species

Type series.—Holotype: MZUSP 29777 (Museu de Zoologia da Universidade de São Paulo; collector number AD 15). Specimen is an adult male collected by A. D. Ditchfield 11 July 1993 in Fazenda Espirito Santo, Município de Soledade, state of Paraíba, Brazil (07°05'S, 36°21 'W). Specimen is fluid preserved with skull extracted. The skull is intact with orbital and postorbital regions moderately clean, with perfect dentition but with the right coronoid process of the mandible broken. Liver tissue has been placed in the Museum of Vertebrate Zoology (MVZ) tissue collection, specimen number MVZ 186020, and at the University of São Paulo, Brazil, collector number AD15.

Paratypes: MZUSP 14170 (collector number ABZUSP expedition 73.0594)—adult female collected 1–4 May 1973, in the municipality of Cocorobó, state of Bahia, Brazil (09°53′S, 39°02′W). Specimen is clean with palatal and orbital region somewhat damaged. Mandible is broken with left coronoid process missing. MZUSP 14173 (collector number ABZUSP expedition 73.0594)—subadult male collected 1–4 May 1973, in the municipality of Cocorobó, state of Bahia, Brazil. Specimen is fluid preserved with a broken mandible. MZUSP 14804 (collector number ABINEP 3685)—adult female collected 12 January 1978, in the Serra da Gritadeira, 18 km soutouthwest of Exu, municipality of Exu, state of Pernambuco, Brazil (07°40′S, 39°47′W). Specimen is fluid preserved with tongue and skull extracted. Skull is clean but premaxillary bone and inner superior incisors are missing. Lower mandible is intact. This specimen was gold dusted for electron microscopic analysis. Specific localities are shown in Fig. 2.

Fig. 2

Map representing the upper portion of South America, showing distribution of Xeronycteris vieirai sp. nov. (localities 1–3) and Platalina (dotted area). Note wide gap (approximately 4,000 km) separating the 2 genera. 1, Soledade (type locality); 2, Serra da Gritadeira, Exu; 3, Cocorobó. Distribution of Platalina was compiled from Eisenberg and Redford (1999).

Diagnosis.—Skull with rostral length slightly larger than braincase (Fig. 3). Posterior edge of palate extends to optic foramen, with rounded hape margin. Upper premolars (P3 and P4) lack cingulum. Parastyle anteriorly very extended in Ml and M2. Upper molar series is extremely reduced, with meso-style, metastyle, preprotocrest, postprotocrest, and ectoflexus all missing or greatly reduced, resulting in the virtual absence of a trigon basin (Fig. 4). Lower dentition reduced with narrowness of premolars, and loss of hypoflexid, metacristid, entocristid, and entoconid (Fig. 5). Medial region of palate seems deeply concave because molar series is not perpendicular to palate but curves inward, with plane of occlusion being oriented lingually and at an angle to transversal plane (Fig. 3). Presence of 4 pits lingually positioned on each side of palate to permit the insertion of main cusps at lingual side in lower dentition.

Fig. 3

Outline of skull (left side) in A) Lionycteris, B) Lonchophylla, C) Platalina, and D) Xeronycteris vieirai. Note increase in the projection of upper incisors and premaxillary branch of palate from A to D, the greater rostral length in C and D, and obscured cusps of the 2 last molars due the inward position on the palate in X. vieirai sp. nov.

Description and comparisons.—Externally, Xeronycteris is similar to other glossophagine genera, with an elongate rostrum, reduced pinnae and noseleaf, reduced tail, and a long tongue with bristles. The tongue of Xeronycteris with its lateral groove, row of filiform papillae on lateral side, and reduced dorsal and lateral vallate papillae clearly place the genus within Lonchophyllini. (Tongue description and terminology is according to Gimenez et al. [1996].) Measurements of external morphology and skull of all relevant taxa are given in Tables 1 and 2.

View this table:
Table 1

Comparative morphometries for females for selected taxa in the tribe Lonchophyllini. Acronyms are defined in Appendix II. Upper line is mean ± SD;lower line indicates range with sample size in parentheses. Measurements are in millimeters.

CharacterXeronycteris vieiraiLionycterisaLonchophylla mordaxL. bokermannibL. thomasi
gls27.00 (1)19.60 (1)23.00 ± 0.3925.67 ± 0.3020.31 ± 0.69
22.13–23.34 (8)25.40–26.30 (7)19.80–20.80 (2)
cil25.20 (1)18.00 (1)21.63 ± 0.3924.38 ± 0.3119.24 ± 0.79
20.93–22.03 (7)24.00–25.00 (7)18.68–19.80 (2)
Pi8.80 (1)12.36 ± 0.1014.20 ± 0.2110.91 ± 0.55
12.22–12.47 (7)13.80–14.30 (7)10.52–11.30 (2)
C–M38.52 ± 0.0217.50 ± 0.287.80 ± 0.108.18 ± 0.276.48 ± 0.02
8.50–8.54 (2)17.30–17.70 (2)7.67–7.94 (8)7.80–8.60 (7)6.46–6.50 (2)
I–C3.20 (1)2.89 ± 0.132.39 (1)
2.73–3.10 (7)
bb8.77 ± 0.358.20 ± 0.288.29 ± 0.239.42 ± 0.177.96 ± 0.33
8.75–8.80 (2)8.00–8.40 (2)7.88–8.70 (9)9.30–9.70 (7)7.72–8.20 (2)
M–M5.79 ± 0.085.05 ± 0.215.06 ± 0.185.42 ± 0.145.10 ± 0.14
5.73–5.85 (2)4.90–5.20 (2)4.78–5.26 (9)5.20–5.70 (7)5.00–5.20 (2)
ml19.66 (1)13.05 ± 0.2115.73 ± 0.7217.55 ± 0.1913.56 ± 0.19
12.90–13.2 (2)13.90–16.59 (9)17.40–18.00 (7)13.43–13.70 (2)
c–m39.14 ± 0.226.50 ± 0.288.01 ± 0.318.62 ± 0.276.72 ± 0.11
8.98–9.30 (2)6.30–6.70 (2)7.45–8.46 (8)8.20–8.90 (7)6.64–6.80 (2)
i–c2.57 ± 0.01.75 ± 0.121.47 (1)
2.57 (2)1.61–2.01 (8)
fal36.77 ± 1.9035.00 ± 1.2435.45 ± 1.2840.38 ± 0.48
35.42–38.12 (2)33.60–36.00 (3)34.00–37.73 (6)40.00–41.30 (7)
  • a Data calculated from Taddei et al. (1978).

  • b Data calculated from Sazima et al. (1978).

View this table:
Table 2

Comparative morphometrics for males of selected taxa in the tribe Lonchophyllini. Acronyms are defined in Appendix II. Upper line is mean ± SD; lower line indicates range with sample size in parentheses. Measurements are given in millimeters.

CharacterXeronycteris vieiraiPlatalinaLionycterisLonchophylla mordaxL. bokermanniaL. dekeyserib
gls25.50 (1)32.15 ± 0.2120.28 ± 0.4123.05 ± 0.6225.25 ± 0.0722.30 ± 0.42
32.00–32.30 (2)19.35–20.67 (8)21.87–23.98 (18)25.20–26.30 (2)22.00–22.60 (2)
cil24.20 (1)30.75 ± 0.3519.02 ± 0.3920.89 ± 0.8824.6 ± 0.9820.75 ± 0.49
30.50–31.00 (2)18.17–19.43 (8)19.80–22.50 (17)23.90–25.30 (2)20.40–21.10 (2)
pl16.68 (1)19.00 (1)10.23 ± 0.2412.39 ± 0.4614.05 ± 0.21
9.68–10.60 (8)11.66–13.10 (16)13.90–14.20 (2)
C–M311.00 (1)10.85 ± 0.216.45 ± 0.247.84 ± 0.258.10 ± 0.147.50 (1)
10.7–11.00 (2)6.00–6.80 (8)7.25–8.25 (17)8.00–8.20 (2)
I–C3.16 (1)3.60 (1)2.21 ± 0.112.87 ± 0.21
2.04–3.38 (8)2.56–3.2 (10)
bb8.64 (1)10.35 ± 0.918.17 ± 0.158.41 ± 0.259.35 ± 0.218.60 (1)
9.70–11.00 (2)8.00–8.38 (8)7.92–8.83 (17)9.20–9.50 (2)
M–M5.67 (1)5.60 (1)4.98 ± 0.205.50 ± 0.285.00 ± 0.28
4.62–5.28 (8)5.30–5.70 (2)4.80–5.20 (2)
ml22.58 (1)23.10 (1)13.65 ± 0.3715.79 ± 0.6317.50 ± 0.0715.05 ± 0.35
12.88–13.99 (8)14.39–16.75 (14)17.30–17.4 (2)14.80–15.30 (2)
c–m38.54 (1)11.30 (1)6.53 ± 0.198.14 ± 0.318.55 ± 0.217.85 ± 0.07
6.15–6.80 (8)7.49–8.65 (15)8.40–8.70 (2)7.80–7.90 (2)
i–c2.32 (1)2.30 (1)1.41 ± 0.091.71 ± 0.10
1.26–1.52 (8)1.58–1.86 (8)
fal35.70 (1)46.00 (1)35.44 ± 0.7635.36 ± 0.9939.50 ± 0.9137.50 ± 0.21
34.50–36.70 (9)34.00–36.90 (16)38.70–40.00 (2)37.40–37.70 (2)
  • a Data calculated from Sazima et al. (1978).

  • b Data calculated from Taddei et al. (1983).

The skull of Xeronycteris has a rostrum approximately equal to braincase in length. This is similar to Lonchophylla except that in Xeronycteris the skull is longer overall, with braincase merging with the rostrum without a distinct change of plane as in Lonchophylla. The skull of Lionycteris is smaller than that of the other genera, with rostrum clearly shorter than braincase, whereas Platalina has a largest skull with a rostrum clearly longer than braincase (Fig. 3). The premaxillary bone projects beyond the canines to a great degree in Xeronycteris and Platalina, producing a triangular projection, which differs from Lonchophylla and Lionycteris, where the anterior portion of the palate is a shallow arch. The palate of Lonchophylla, Lionycteris, and Xeronycteris is deeply concave in its medial portion, between the molars, a condition differing from the flat palate of Platalina. The posterior portion of palate is short in Lionycteris but long in Xeronycteris, reaching the distal border of optic foramen, whereas in the remaining genera the palate ends in the anterior border of optic foramen. The medial posterior margin of palate in Xeronycteris is rounded and distinctly haped, whereas in Platalina and Lonchophylla it is an inflected V, like a narrow lyre. In Lionycteris, the shape is square, resembling a flaottomed U. Four pits are present on each side of palate, lingually positioned to dentition. These pits permit the insertion of the main cusps (protoconid) of lower last premolar and 3 molars. This characteristic is odd to Xeronycteris. The most anterior of these pits are positioned on the anterior onalf of last upper premolar, the 2nd pit between last premolar and 1st molar, the 3rd and deepest pit is placed between 1st and 2nd molar, and the most posterior pit is between 2nd and 3rd molar.

The morphology and position of the upper incisors is similar in Xeronycteris, Lonchophylla, and Lionycteris. All 3 have similar inner upper incisors, with medial edge of incisor straight but labial margin flaring out into a broad spatulate shape. In Platalina, the inner upper incisors are broad and square with no flaring from base to tip. In addition, for the 3 genera (Xeronycteris, Lonchophylla, and Lionycteris), the inner and outer incisors are separated by a narrow gap. In Platalina, all incisors are in contact with one another (Fig. 4). However, the angle of insertion of upper incisors in Platalina and Xeronycteris is similar, with the upper incisors projecting forward edge on, almost as an extension of the rostrum, whereas in Lonchophylla and Lionycteris, the incisors are placed at an angle of approximately 45° to the plane of the palate, facing downward instead of anteriorly (Fig. 3). Upper canines are more flattened in the lingual surface in Xeronycteris and Platalina than in Lonchophylla and Lionycteris, with a weak cingulum (Fig. 4). The shape of premolars of Xeronycteris and Platalina tends to be lower, longer, and narrower when compared to Lonchophylla and Lionycteris. In Xeronycteris and Platalina, the paracone of P3 is short and the cingulum is absent, whereas in Lonchophylla and Lionycteris the paracone is prominent and the cingulum is distinctly developed. The shape of P4 is distinctive among the 4 genera. Xeronycteris presents the simplest morphology, with the tooth being similar to the P3. In Platalina, there is a weakly developed cingulum, and an accessory posterior cuspid (Fig. 4). In Lonchophylla, the paracone is high and the cingulum is well developed, with a small protocone projected lingually and connected by a narrow crest. In Lionycteris, the paracone is high and the cingulum also is well developed, but the protocone is absent; however, 2 mesostylar cuspids are present in the posterior border of the cingulum. The morphology of Ml and M2 is distinct in Xeronycteris, with extreme reduction of tooth structure with absence of mesostyle, metaconule, postprotocrest, metaconule, and ectoflexus (Fig. 4). Metastyle is absent or it is so reduced that its visualization is impossible. Also, the trigon basin is absent and protocone is connected to the tooth by a narrow crest in Xeronycteris. In Platalina, the dental structure is reduced but not to the same degree as in Xeronycteris. Metastyle is reduced in Platalina, and the protocrest, preprotocrest, and postprotocrest are all present but the trigon basin is shallow on the 2nd upper molar. In Lonchophylla and Lionycteris, the molars are broad and with a deep trigon basin. The metaconule on Ml is absent in Lonchophylla but present in Lionycteris. The M3 of Xeronycteris differs from that of other taxa only by the absence of an ectoflexus, postprotocrest, and trigon. Also, in Xeronycteris, the molar series differs from all other genera of Lonchophyllini in that the plane of occlusion of the molars is oriented lingually and at an angle to the plane defined by the palate. Consequently, when viewed laterally, the main cuspids of the molars of Xeronycteris are not visible, differing from the other taxa (Fig. 3).

Fig. 4

Upper toothrow (occlusal view, left side) in A) Lionycteris, B) Lonchophylla, C) Platalina, and D) Xeronycteris vieirai sp. nov. Note the trend in reduction of dentition from A to D, wide gap between incisors in D, reduction of cingulum on canine and premolars in C and D, and reduction of crests and trigon in C. Note the absence of ectoflexus, mesostyle, pre- and postprotocrest, and the virtual trigon in X. vieirai sp. nov. Ci, cingulum; Efx, ectoflexus; Me, metacone; Mecr, metacrest; Mest, mesostyle; Met, metaconule; Metst, metastyle; Pa, paracone; Pacr, paracrest; Past, parastyle; Popcr, postprotocrest; Pr, protocone; Prpcr, preprotocrest.

View this table:
Table 3

Character codes for states of morphological characters for 10 genera of Glossophaginae, including the newly described genus. Character states are described in Appendix II. Characters are grouped into sets of 5 for easier reading. Missing data are indicated by a question mark (?).

Characters
Taxon1–56–1011–1516–2021–2526–3031–34
Erophylla0000010000000020000000000100000000
Monophyllus0000000000000111001111010001010000
Choeronycteris0000000002000111001000010101010000
Leptonycteris0000001000010001201111??0?00000000
Anoura0000000000000011101111000001010000
Glossophaga0000001000000021000111000001011000
Lionycteris1111100010110021111110000011011000
Lonchophylla1111101001110111101111010011010000
Platalina1101111002101110101000010111000010
Xeronycterisgen. nov.1101112101121100201001111212100111

As for the inner jaw, lower outer incisors are separated from the canines by a gap in Platalina, Lonchophylla, and Lionycteris (0.1–0.3 mm), whereas in Xeronycteris this gap is somewhat larger (0.6–0.82 mm) (Fig. 5). The lingual cingulum of lower canines is evident in Platalina, Lonchophylla, and Lionycteris but extremely reduced in Xeronycteris. The posterior cingular crest of the lower canines is reduced in Xeronycteris and Lionycteris, whereas it is developed in Platalina and Lonchophylla. The paraconid in p2–1 is low in Xeronycteris and Platalina, whereas it is high in Lonchophylla and Lionycteris (Fig. 3). The p3 and p4 in Xeronycteris and Platalina are slightly inclined inward in relation to the molars, and they are narrow and bladelike in occlusal view, with the absence of lingual structures. In Lonchophylla, the cingulum in p3 and p4 is weakly developed, whereas in Lionycteris, the lingual region is well developed with the presence of metaconid. The entoconid is present in the molars of Lonchophylla and Lionycteris, but it is lacking in Xeronycteris and Platalina. The hypoflexid, metacristid, and entocristid of m1–3 are present in all Lonchophyllini genera save for Xeronycteris. Likewise, the talonid basin is very reduced in Xeronycteris and well defined in the remaining Lonchophyllini taxa (Fig. 5). To conclude, in the tribe Lonchophyllini, in both lower and upper premolars and molars, the primitive molar pattern of phyllostomids with the presence of deep and welefined fossae demonstrated by Lionycteris and Lonchophylla are considered derived and resulting from character reversion (Wetterer et al. 2000). In Platalina and even more so in Xeronycteris, there is a tendency for the compression of molars into a narrow blade with the reduction and loss of structures in the lingual border, a plesiomorphic state in Lonchophyllini. However, it is important to note that derived or primitive states of characters would be assigned throughout global parsimony analysis and in our analysis the characters are employed strictly to access Lonchophyllini phylogeny.

Fig. 5

Lower toothrow (occlusal view, right side) in A) Lionycteris, B) Lonchophylla, C) Platalina, and D) Xeronycteris vieirai sp. nov. Note trend in narrowness of premolars and reduction of crests and cusps on molars from A to D, and wide gap between outer incisor and canine, narrowness of premolars, and lack of hypoflexid, metacristid, entocristid, and entoconid in D. Cc, cingular crest; En, entoconid; Ec, entocristid; Hy, hypoconid; Hyf, hypoflexid; Me, metaconid; Mec, metacristid; Pa, paraconid; Poc, postcristid; Pr, protoconid. Question mark (?) indicates that some structures are lacking on molars of X. vieirai.

Geographical distribution.—Currently known only from semiarid localities in the interior of northeastern Brazil, inland and quite distant from the belt of humid coastal climate on the Atlantic seaboard (Fig. 1). The area is covered by Caatinga, a scrubland rich in cactii and xerophyllic thorn brush (Mares et al. 1981; Sick et al. 1987).

Etymology.—Xeronycteris is derived from a combination of Greek roots, Xeros, dry, and Nykteris, bat, in reference to the species living in the Caatinga, a semiarid habitat in northeastern Brazil. The epithet vieirai is in honor of Dr. Carlos O. C. Vieira, whose monograph became a major reference for Brazilian bats (Vieira 1942).

Concluding statements.—The close relationship of the species P. genovensium and X. vierai and the fact that both occur in arid environments makes it probable that their hypothetical ancestor also inhabited the arid regions of South America in the past. Other mammal groups presently restricted to the Andes had a wider distribution in South America in the past, occurring in eastern Brazil when that region was more arid during climatic changes (e.g., camelids—Cartelle 1999; and the freailed bat MormopterusCzaplewski 1997; Legendre 1985). In addition, the lingual characters and particularly the extreme reduction of dental morphology, especially conspicuous in Xeronycteris, indicate that both genera may have a predominantly liquid diet (i.e., nectar), as opposed to the other glossophagines that in general have a more opportunistic diet of insects, fruit, and flower parts, in addition to nectar (Gardner 1977; von Helversen and Winter 2003). Unfortunately, we have few ecological data for Platalina and none for Xeronycteris, making it extremely desirable to study the ecology of these taxa in the field to test the evolutionary hypothesis proposed here.

Xeronycteris vieirai is the 2nd Brazilian bat endemic to the open and dry biomes (Cerrado and Caatinga) in Brazil, the other endemic being Lonchophylla dekeyseri. This observation has conservation implications because the original vegetation of the semiarid regions of northeastern Brazil has been severely affected by establishment of plantations and logging for charcoal (Coimbrilho and Câmara 1996). For species with restricted geographical distribution and highly specific diet requirements, such a loss of habitat might present a serious threat. Thus, the endemic character of X. vieirai and its completely unknown ecological needs makes it potentially one of most threatened species of mammals in Brazil. Further studies in the near future are needed to obtain crucial biological data for this new taxon.

Acknowledgments

We are grateful to the following museum curators for kindly loaning us the relevant material and allowing us to analyze it, namely M. de Vivo (MZUSP), J. L. Patton (MVZ), and V. A. Taddei (DZSJRP). We thank J. L. Patton for suggestions and criticism on earlier versions of the manuscript. This study was supported by Fundação de Ampara a Pesquisa do Estado de São Paulo (FAPESP), Processes 98/05075 (Programa Biota), 01/07067–6 (to RG), and 99/02403–6 (to ADD).

Appendix I

Specimens examined.—Museums and acronyms follow Hafner et al. (1997): Museu de Zoologia da Universidade de São Paulo, Sao Paulo, Brazil (MZUSP), in which the type specimens examined are housed; Laboratório de Chiroptera, Departamento de Zoologia e Botânica da Universidade Estadual Paulista, São José do Rio Preto (DZSJRP); Museum of Vertebrate Zoology, University of California, Berkeley, California (MVZ).

Anoura geoffroyi.—Brazil: Ceará, Floresta Nacional do Araripe (MZUSP 16974, 16980); Ceará, Nova Olinda (MZUSP 17014, 17015).

Choeronyscus minor.—Brazil: Amazonas, Rio Juruá (MZUSP 1207); Mato Grosso, Alta Floresta (Rio Cristalino—MZUSP 22712); Pará, Lago Leonardo, Rio Trombetas (MZUSP 212919); Pará, Rio Trombetas, between Lago do Arrozal and Cachoeira Porteira (MZUSP 13178).

Glossophaga soricina.—Brazil: Ceará, Floresta Nacional do Araripe (MZUSP 18005, 18013, 18017); Ceará, Nova Olinda (MZUSP 18022); Ceará, Crato (MZUSP 18168, 18173).

Erophylla zekorni.—Dominican Republic: District Nacional, Paseo de los Indios (MVZ 166156, 166157).

Leptonycteris curasoae.—Mexico: Baja California, Santa Rosalia (MVZ 110164, 110167, 110176).

Lichonycteris obscura.—Brazil: Amapá, Serra do Navio (MZUSP 18925); Espfrito Santo (MZUSP 2253); Pará, Uruá, Itaiatuba (MZUSP 12632).

Lionycteris spurrelli.—Brazil: Bahia, Itaetê, Caverna Poço Encan-tado (MZUSP 29069); Minas Gérais, Itacarambi, Caverna Olhos d'Água (MZUSP 28952–28958).

Lonchophylla mordax.—Brazil: Bahia, Barra (MZUSP 2659–2661); Bahia, Cocorobó (MZUSP 14171, 14174); Bahia, Itaetê, Caverna Poço Encantado (MZUSP 29132); Ceará, Nova Olinda (MZUSP 18176–18183); Goiás, Mambaí (MZUSP 13662); Pernambuco, Exu (MZUSP 18212, 18213, 18216, 18218, 18220, 18224); Pernambuco, Serrote das Lajes, 17 km S from Exu (MZUSP 18191, 18196, 18199, 18200, 18205, 18210).

Lonchophylla bokermanni.—Brazil: Minas Gerais, Serra do Cipó, Jaboticatubas (DZSJRP 10342, 10347 [holotype], 10408, 11410, 11411).

Lonchophylla dekeyseri.—Brazil: Distrito Federal, Parque Nacional de Brasilia (DSZJRP 10099 [holotype]).

Lonchophylla thomasi.—Brazil: Mato Grosso, Alta Floresta (MZUSP 28279); Pará, Beiérn (MZUSP 3335); Pará, Rodovia Belérasília, km 97 (MZUSP 14100); Rondônia, Pedra Branca (MZUSP 22891).

Monophyllus plethodon.—Lesser Antilles (formerly a possession of the United Kingdom): Montserrat (MVZ 166189, 166190, 166193).

Platalina genovensium.—Peru: Depto. Piura, Talara (MVZ 135547, 135550).

Xeronycteris vieirai, new genus and species.—Brazil: Bahia, Cocorobó (MZUSP 14170, 14173 [paratypes]); Paraiba, Soledade (Fazenda Espfrito Santo—MZUSP 29777 [holotype]); Pernambuco, Serrote da Gritadeira, 18 km SSW Exu (MZUSP 14804 [paratype]).

Appendix II

Definition of 10 biométric variables and 34 descriptions of qualitative characters used in analyzing genera of Glossophaginae. Codes are given for various states of qualitative characters.

Morphometric variables and acronyms.—Forearm length (fal): Distance from elbow to upper wrist joint including carpal elements. Total length of skull (gls): Distance from posterior edge of occipital to anterior edge of incisors. Condylncisive length (cil): Distance between the most posterior point of occipital condyles to tip of incisors. Braincase breadth (bb): Greatest breadth of globular part of braincase. Palatal length (pl): Distance between medial edge of palate to incisors. Upper caninolar length (3): The length of maxillary toothrow from anterior edge of the canine cingulum to posterior edge of last molar. Maxillary caninncisive length (): Measured between posterior edge of cingulum of canines to labial tip of upper incisors. Maxillary breadth (): Measured across labial edge of most external molars. Mandibular length (ml): The distance between coronoid process to labial tip of lower incisors. Lower caninolar length (3): The length of maxillary toothrow from anterior edge of lower canine cingulum to posterior edge of last molar. Lower caninncisive length (): Measured between posterior edge of cingulum of canines to labial tip of lower incisors.

Facial and tongue morphology.—1. Lingual sulcus: absent (0); present (1). A lateral lingual sulcus is exclusive for all Lonchophyllini taxa, including Xeronycteris. See detailed description in Gimenez et al. (1996) and Wetterer et al. (2000). 2. Hairlike papillae ventrally to the sulcus: absent (0); present (1). A single and posterior line of hairlike papillae running ventrally to the lingual sulcus is present exclusively in Lonchophyllini (see Wetterer et al. 2000) including Xeronycteris. 3. Basketlike: absent (0); present (1). Basketlike papillae (definition in Wetterer et al. 2000) are present only in Lonchophylla and Lionycteris. 4. Central rib in midsaggital line of spear: absent (0); present (1). Wetterer et al. (2000) provided the distribution of the central rib in the spear of the phyllostomids. We agree only partially with them, because an analysis of a large sample of Lonchophylla revealed that L. mordax has no ribs, suggesting that this is a polymorphic character for the genus. Additionally, Lionycteris and Xeronycteris have no ribs. We followed Wetterer et al. (2000) regarding the description of noseleaf morphology for Platalina. 5. Number of interramal vibrissae: 2 (0); 3 (1). Three interramal vibrissae are present exclusively in Xeronycteris and Lonchophyllini. 6. Line of papillae or fleshy ridge in internarial region: absent (0); present (1). Presence of internarial papillae or fleshy ridge was recorded in Lonchophylla, Lionycteris, and Glossophaginae, but not in Xeronycteris and Erophylla.

Skull.—7. Premaxillary: short (0); long (1); very projected (2). A short premaxillary bone was recorded for most outgroup taxa and Lionycteris. Lonchophylla, Platalina, Leptonycteris, and Glossophaga showed the bone anteriorly extended, and Xeronycteris possesses the bone very projected, much exceeding the anterior side of canines, in a dorsal view. 8. Palatal length: short (0); long (1). Palate defined as short when its posteromedial edge reaches the anterior region of optic foramina as recorded for most taxa. Xeronycteris presents the border of palate posteriorly exceeding the optic foramina. 9. Oval foramen: small (1); large (2). A huge oval foramen was recorded exclusively in Lionycteris. 10. Rostrum length: shorter than braincase (0); equal (1); larger than braincase (2). Most taxa analyzed showed the rostrum shorter than braincase. Platalina and Choeronycteris presented the rostrum clearly longer than the braincase, and Lonchophylla and Xeronycteris have a similar size in both rostrum and braincase. 11. Zygomatic arch: complete (0); incomplete (1). Incomplete zygomatic arch is characteristic of Lonchophyllini and a few genera of nectaeeding bats, such as Phyllonycteris, Lichonycteris, and Choeroniscus (see Wetterer et al. [2000] for distribution of this character in Phyllostomidae).

Upper dentition.—12. Gap between inner and outer incisors: absent (0); present (1). A noticeable gap between incisors was recorded exclusively in Xeronycteris, whereas Lonchophylla and Lionycteris present a narrow space between the teeth. In contrast, the remaining taxa have the incisors fully in contact or very close together. 13. Position of incisors: facing downward (0); projecting forward (1). The angle of insertion of the incisors in Platalina and Xeronycteris is similar in projecting forward edge on, whereas in Lonchophylla, Lionycteris, and all taxa of the outgroup the incisors are placed facing downward instead of ahead. 14. Parastyle on P3: weak (0); high (1). High and pointed parastyle is present only in Xeronycteris, Platalina, Leptonycteris, Choeronycteris, and Monophyllus. 15. Cingulum on P3: absent (0); weak (1); developed (2). The cingular shelf in taxa varied from absent (Xeronycteris and Leptonycteris) to notably developed, such as in Lionycteris and Glossophaga. Most of the taxa analyzed showed moderated cingulum. 16. Paracone on P3: low (0); high (1). Paracone is reduced (low) in Xeronycteris, Platalina, and Erophylla. However, in Erophylla, the structure is very low and rounded, almost vestigial, whereas in the 2 other taxa, the paracone is a distinct and point cusped. 17. Anterolingual cingulum on P4: developed (0); reduced (1); absent (2). Xeronycteris and Leptonycteris apparently have no cingulum in their premolars. In Platalina, Lionycteris, and Anoura the structure is weak and in the remaining taxa the cingulum is evident. 18. Accessory cusp on P4: absent (0); present (1). A posterior accessory cusp on P4 was recorded exclusively in Lionycteris. 19. Thickness of premolars: thick (0); narrow (1). Laterally compressed premolars are present in most of taxa, either with or without cingulum and accessory cusp. However, Erophylla and Glossophaga possess thick premolars. 20. Mesostyle on Ml and M2: absent (0); present (1). Mesostyle is lacking in Xeronycteris, Platalina, Choeronycteris, and Erophylla. 21. Parastyle on Ml and M2: present (0); absent or extremely reduced (1). Only Choeronycteris and Platalina have the parastyle absent or so reduced that it is virtually absent. 22. Metastyle on Ml: reduced (0); developed (1). The structure is reduced or almost absent in Platalina, Lionycteris, Lonchophylla, Erophylla, and Choeronycteris. 23. Postprotocrest on Ml, M2, and M3: present (0); absent or extremely reduced (1). Postprotocrest is lacking or so reduced that is impossible to record in Xeronycteris. The fact that postprotocrest is lacking on molars in Xeronycteris results in the total absence of fossae (trigon), and thus the protocone is connected with the labial side of the teeth exclusively by the preprotocrest. This is one of the particularly conspicuous characters diagnosing the taxon. Because Leptonycteris has no M3, characters related to this tooth were considered not comparable. 24. Metaconule on M2: present (0); absent (1). Metaconule is lacking in Lonchophyllini (excepting Lionycteris), Monophyllus, and Choeronycteris. 25. Insertion of M2 and M3 on the palate: parallel (0); inward (1). The last 2 molars in Xeronycteris are placed intrinsically inward to the posterior palate, which indicates that the occlusion occurs with the mandible against the palate and not parallel to it. 26. Position of M3 on the palate: below the zygomatic arch (0); slightly ahead of the zygomatic arch (1); quite distant from the zygomatic arch (2). The combination of reduction of dentary structures and elongation of palate in Xeronycteris resulted in an exclusive disposition of tooth anteriorly quite distant from the zygomatic arch. The most frequent position of the last molar (M3) is below the zygomatic arch, such as in Lonchophylla, Lionycteris, Glossophaga, Anoura, and Monophyllus, or immediately anterior to the arch, as in Platalina, Erophylla, and Choeronycteris.

Lower dentition.—27. Number of lobes on incisors cingulum: 1 (0); 3 (1). Large and trifid incisors are present exclusively in Lonchophyllini according to Phillips (1971). 28. Gap between outer incisor and the canine: absent (0); reduced (1); noticeable (2). Most of the taxa studied presented a narrow gap between incisors and canine. In Erophylla and Leptonycteris, this gap is lacking, whereas in Xeronycteris, it is large. This noticeable gap is a result of the presence of a delicate canine and the projection of the anterior region of dentary in Xeronycteris, another obvious diagnostic character for this taxon. 29. Cingulum on canine: developed (0); absent (1). Cingulum on canine is lacking in Xeronycteris, contrasting with the developed cingulum in all studied taxa. 30. Paraconid on p2, p3, and p4: low (0); high (1). Paraconid is high in most of taxa and very low in Xeronycteris, Platalina, Leptonycteris, and Erophylla. 31. Cingular cristid on p4: absent (0); present (1). A cingular cristid on the buccal side of the teeth was recorded in Lionycteris and Glossophaga. 32. Metacristid on ml, m2, and m3: present (0); absent (1). The metacristid is present in most of taxa except Xeronycteris, reflecting the trends in dentary reduction in the genera. Erophylla also presents a trend in reduction of cusps and crest, frequently related to size and not lost. 33. Entoconid on ml, m2, and m3: present (0); absent (1). Entoconid cusp also is lacking in Xeronycteris and Platalina, and noticeably reduced or absent in Erophylla. 34. Entocristid on molars: present (0); absent (1). The entocristid is lacking exclusively in Xeronycteris. As a result, the postcristid, a small accessory crest on the buccal side of the molars, is also lacking in Xeronycteris. This extreme dentary reduction due the loss of structures also resulted in the absence of trigonid on molars in Xeronycteris.

Footnotes

  • Associate Editor was Robert D. Bradley.

Literature Cited

View Abstract