amnb 00189 Mp 1 File # 01TQ

PHYLOGENETIC RELATIONSHIPS OF MORMOOPID BATS (CHIROPTERA: MORMOOPIDAE) BASED ON MORPHOLOGICAL DATA

NANCY B. SIMMONS Associate Curator Division of Vertebrate Zoology (Mammalogy) American Museum of Natural History

TENLEY M. CONWAY Scientific Assistant Division of Vertebrate Zoology (Mammalogy) American Museum of Natural History

BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY Number 258, 97 pp., 12 figures, 4 tables Issued February 15, 2001 Price: $10.10 a copy

Copyright ᭧ American Museum of Natural History 2001 ISSN 0003-0090 2 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

CONTENTS Abstract ...... 3 Introduction ...... 3 Historical Background ...... 3 Mormoopid Species: A Synopsis ...... 8 Goals of the Present Study ...... 12 Materials and Methods ...... 12 Taxonomic Sampling, Outgroups, and Tree Rooting ...... 12 Sources of Data ...... 13 Definition of Characters and Ordering of Character States ...... 13 Polarity ...... 15 Completeness ...... 15 Methods of Phylogenetic Analysis ...... 16 Character Descriptions ...... 17 Skull ...... 17 Dentition ...... 24 Vomeronasal Complex and Brain ...... 29 Trachea and Hyoid Apparatus ...... 30 Tongue ...... 35 Face, Ears, and Vibrissae ...... 36 Pelage and Patagia ...... 42 Postcranial Skeleton ...... 45 Postcranial Myology ...... 58 Reproductive Tract ...... 65 Digestive Tract ...... 66 Results ...... 68 Discussion and Conclusions ...... 73 Intrafamilial Relationships of Mormoopids ...... 73 Classification and Taxonomic Diagnoses of Mormoopid Clades ...... 74 Monophyly and Diagnoses of Noctilionoid Families ...... 81 Interfamilial Relationships ...... 83 Directions for Future Research ...... 84 Acknowledgments ...... 85 References ...... 85 Appendix 1: Specimens Examined ...... 93 Appendix 2: Taxon-Character Matrix ...... 96 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 3

ABSTRACT Mormoopidae is a small family of Neotropical microchiropteran bats that includes two genera (Mormoops and Pteronotus) and ten species, two of which are known only from fossils. Mormoopidae is typically classified as a member of Noctilionoidea, a group that minimally includes two other Neotropical families (Phyllostomidae and Noctilionidae) and may also include Mystacinidae, a taxon endemic to New Zealand and Australia. Phylogenetic relation- ships of extant mormoopid species and one extinct taxon, Pteronotus pristinus, were investi- gated in a series of parsimony analyses of 209 morphological characters including features of the skull, dentition, vomeronasal organ complex and brain, trachea and hyoid apparatus, tongue, face, ears, pelage, patagia, postcranial skeleton, postcranial myology, reproductive tract, and digestive tract. Three extant phyllostomid species, two noctilionids, two mystacinids, and one emballonurid species were included as outgroups to test monophyly of Mormoopidae and to provide a context for determining the sister group of the family. Results of parsimony analyses under a variety of different assumption sets indicate that Mormoopidae is monophy- letic, and that Pteronotus and Mormoops are monophyletic sister taxa. Within the genus Pter- onotus, several clades were repeatedly recovered: (1) P. davyi ϩ P. gymnonotus (ϭ subgenus Pteronotus); (2) P. macleayi ϩ P. quadridens; (3) P. personatus ϩ P. macleayi ϩ P. quad- ridens (ϭ subgenus Chilonycteris); (4) P. parnellii ϩ P. pristinus ϩ P. personatus ϩ P. macleayi ϩ P. quadridens; and (5) P. parnellii ϩ P. pristinus (ϭ subgenus Phyllodia). These results support monophyly of all subgenera of Pteronotus previously recognized, and addi- tionally indicate that the subgenera Pteronotus and Chilonycteris are sister taxa. Comprehen- sive diagnoses for each species and clade of Mormoopidae are provided based on character optimizations and ancillary morphometric data from the literature. Results of our parsimony analyses also have implications for understanding higher-level phy- logeny of noctilionoid bats. Monophyly of each of the traditionally recognized families (i.e., Noctilionidae, Mystacinidae, Mormoopidae, and Phyllostomidae) was strongly supported. In con- gruence with recent analyses of mitochondrial gene-sequence data and DNA hybridization ex- periments, we found strong support for inclusion of Mystacinidae in Noctilionoidea. Using an emballonurid species to root the tree, we found the following interfamilial relationships of noc- tilionoids: (Noctilionidae (Mystacinidae (Phyllostomidae, Mormoopidae))). Lists of morpholog- ical synapomorphies of each of these groups are provided based on character optimizations.

INTRODUCTION tropical forest to semiarid and arid subtropical forest and scrubland (Handley, 1976; Em- Mormoopidae is a small family of Neo- mons, 1997; Reid, 1997; Smith, 1972). The tropical microchiropteran bats that currently most recent revision of Mormoopidae was includes two genera (Mormoops and Pteron- that of Smith (1972), which has formed the otus) and eight extant species; two additional basis for most modern treatments of the fam- species are known only from Quaternary fos- ily. sils (Smith, 1972; Silva-Taboada, 1974, 1979; The history of classification and nomen- Koopman, 1993, 1994). Commonly known as clature of mormoopids is complex and often mustached, ghost-faced, or naked-backed confusing. Smith (1972) provided a compre- bats, these taxa are characterized by the pres- hensive systematic literature review for the ence of flaplike outgrowths below the lower family, which we will not reproduce here. A lip and funnel-shaped ears. Mormoopids are few critical points from the pre-1970 litera- small to medium in size (e.g., forearm length ture, together with more recent taxonomic 35Ð66 mm) and are thought to be exclusively changes and phylogenetic inferences, are insectivorous (Koopman, 1984, 1994). Mem- summarized below. bers of the family presently range from the southwestern United States to southern Brazil, HISTORICAL BACKGROUND and also occur in the Greater Antilles, Lesser Antilles, Dutch West Indes, and Trinidad and CLASSIFICATION Tobago (Koopman, 1993, 1994). They live in The taxonomic history of Mormoopidae a wide variety of habitats ranging from humid began with Leach (1821a, 1821b), who de- 4 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 scribed Mormoops and Aello. Pteronotus and sification did not entirely match his phylo- Chilonycteris were subsequently described genetic tree (fig. 1), which depicted Pteron- by Gray (1838, 1839), and Lobostoma was otus personatus as more closely related to P. named by Gundlach (1840). Gray (1843) davyi and P. suapurensis than to P. macleayii named a sixth genus, Phyllodia, followed al- and P. fuliginosus. However, Smith’s (1972) most 60 years later by Gill’s (1901) propo- revision was completed in a precladistic con- sition of Dermonotus as a replacement name text, so this discrepancy was not widely rec- for Pteronotus. Various authors have pro- ognized. Subsequent subgeneric classifica- posed different synonomies for these taxa tions (e.g., Corbet and Hill, 1980; Herd, over the last 150 years, and misplaced ho- 1983; Koopman, 1994) retained Smith’s lotypes, overlooked names, and arguments (1972) subgeneric usage. over nomenclatural priority have served to Although Smith’s (1972) classification of further complicate matters (see review in Mormoopidae survived more-or-less intact Smith, 1972). For most of the 20th century, into the 1990s, several significant nomencla- classifications of Mormoopidae were based tural changes have occurred since that pub- on that of Miller (1907), who recognized lication. Silva-Taboada (1976) showed that three genera (Chilonycteris, Pteronotus, and Lobostoma quadridens Gundlach, 1840 is a Mormoops), which he placed in the subfam- senior synonym of Chilonycteris fuliginosus ily Chilonycterinae in the family Phyllostom- Gray, 1843, so Pteronotus quadridens is now idae. Dalquest and Werner (1954) recom- recognized as the correct name for this spe- mended elevation of the group to family rank cies. Similarly, Smith (1977) subsequently as Chilonycteridae, but this suggestion was demonstrated that Chilonycteris gymnonotus overlooked or ignored by most authors prior Wagner, 1843, is a senior synonym of Der- to Smith’s (1972) revision. Smith (1972) monotus suapurensis J. A. Allen, 1904, so agreed that this group deserved familial sta- Pteronotus gymnonotus is now recognized as tus, but argued that the correct name for this the correct name for this species (see Carter taxon is Mormoopidae. Most subsequent au- and Dolan [1978] for discussion of author- thors have followed this recommendation. A ship of gymnonotus). In addition, two fossil lone dissenter was Hall (1981), who contin- mormoopid species, Mormoops magna and ued to recognize Chilonycterinae as a sub- Pteronotus pristinus, were described from family of Phyllostomidae based on taxonom- Cuban cave deposits by Silva-Taboada ic history and his perception that the ‘‘grade (1974). Partial or complete synonomies for of differentiation’’ of this group was similar all currently recognized species and subspe- to that of ‘‘some other groups of mammals cies of mormoopids can be found in Smith currently treated as subfamilies.’’ These ar- (1972, 1977), Silva-Taboada (1976, 1979), guments did not convince the rest of the sys- Herd (1983), Adams (1989), Rodr«õguez-Dur- tematic community, and the group has been a«n and Kunz (1992), Koopman (1993), and listed under Mormoopidae in most classifi- Rezsutek and Cameron (1993). cations published since 1972 (e.g., Corbet and Hill, 1980; Koopman, 1984, 1993, 1994; PHYLOGENETIC RELATIONSHIPS McKenna and Bell, 1997; Simmons, 1998; Simmons and Geisler, 1998). The first explicit phylogenetic tree of Mor- Smith (1972) recognized two genera, three moopidae was that of Smith (1972; fig. 1). subgenera, and eight extant species within Although Smith later became well known for Mormoopidae. Mormoops was shown to in- his applications of cladistic methodology to clude two species, M. blainvillii and M. me- bat systematics (e.g., Smith, 1976, 1980; galophylla (Smith, 1972). Smith (1972) di- Smith and Madkour, 1980; Hood and Smith, vided Pteronotus into three subgenera: Phyl- 1982, 1983), his tree of mormoopid relation- lodia (including only Pteronotus parnellii), ships was developed in a precladistic context Chilonycteris (including Pteronotus macle- using methods of numerical taxonomy ayii, P. fuliginosus, and P. personatus), and (Smith, 1972). This phenetic tree assumed Pteronotus (including P. davyi and P. sua- monophyly of the family, a hypothesis that purensis). Interestingly, Smith’s (1972) clas- he supported based on morphological com- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 5

Fig. 1. Smith’s (1972) phylogenetic tree of mormoopid species, with the taxonomy changed to reflect modern usage (i.e., P. quadridens ϭ P. fuliginosus, P. gymnonotus ϭ P. suapurensis). This tree, which was developed in a pre-cladistic context using methods of numerical taxonomy, was derived from analysis of 22 skull measurements and 20 qualitative multistate characters of the dentition and external morphology. The horizontal and vertical spacing of various branches was intended to depict degrees of similarity and differences among taxa. Redrawn from Smith (1972: fig. 14).

parisons to members of other chiropteran group relationship between Mormoopidae groups. In large part due to the thoroughness and Noctilionidae was proposed based on of Smith’s (1972) study, this tree was very five inferred Robertsonian fusion events influential in shaping subsequent classifica- shared by these taxa (fusion of chromosomes tions of mormoopids. 21/14, 17/9, 13/8, 22/3, and 18/2; Patton and Karyotypic studies including mormoopid Baker, 1978). All of these results were con- species were completed in the late 1970s and gruent with Smith’s (1972) phylogenetic hy- early 1980s (e.g., Patton and Baker, 1978; pothesis. Baker and Bickham, 1980; Sites et al., 1981). Limited immunological studies conducted All species of Pteronotus and Mormoops around the same time by Honeycutt (1981) were found to have the same gross karyotype produced divergent results. Analysis of al- (2N ϭ 38, FN ϭ 60), and banding compar- bumin immunological distance data for a set isons indicated that they share a unique fis- of taxa including Mormoops megalophylla, sion of chromosome 6/7 that occurs in no Pteronotus parnellii, Noctilio leporinus, a other bat family (Patton and Baker, 1978; mixed sample of several phyllostomid spe- Baker and Bickham, 1980; Sites et al., 1981). cies, and a mixed sample of pteropodids All extant species of Pteronotus share an (used to root the tree) resulted in a tree in identical, unique G-band karyotype, thus which Mormoopidae was paraphyletic with supporting monophyly of the genus (Sites et respect to Phyllostomidae. However, poor al., 1981). Monophyly of Mormoops was sampling within Mormoopidae and inconsis- also supported by banding studies, and only tencies in the data led Honeycutt (1981) to a single difference in banding pattern was exercise caution in interpreting his results, found between the karyotypes of Mormoops which he regarded as inconclusive. and Pteronotus (Sites et al., 1981). A sister- In one of the first papers to explicitly ad- 6 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 dress the problem of combining disparate four data sets (fig. 2). Within Mormoopidae, data sets in phylogenetic studies, Arnold et the relationships depicted were completely al. (1982) reviewed available morphological, compatible with Smith’s (1972) tree (fig. 1). immunological, allozyme, and karyotypic However, the Arnold et al. (1982) tree did data relevant to mormoopid relationships. not include all mormoopid species (Pteron- Phylogenetic trees for each data set were de- otus gymnonotus was excluded), and rela- rived by hand by applying parsimony criteria tionships among species of Pteronotus were to the discrete data sets (morphology, kar- poorly resolved. yotypes, and allozymes) and distance meth- No new comprehensive phylogeny of ods to the immunological data (Arnold et el., Mormoopidae has been published since 1982). The karyotype tree presented by Ar- 1982. However, several studies have ad- nold et al. (1982: fig. 3) was based on data dressed the relationships of mormoopids to from Patton and Baker (1978) and Baker and other families. Pierson (1986) and Pierson et Bickham (1980), which supported monophy- al. (1986) proposed a phylogeny of bats ly of a Mormoopidae ϩ Noctilionidae clade, based on a study of transferrin immunologi- Mormoopidae, and Pteronotus.1 The immu- cal distance data. They found that Mormoop- nological tree was taken from Honeycutt idae, Noctilionidae, and Mystacinidae (1981), which indicated paraphyly of Mor- formed a clade, with Phyllostomidae as a moopidae with respect to Phyllostomidae, very close relative to this group. More than with Noctilionidae as the sister group to a decade later, Kirsch et al. (1998) found these taxa. The morphology tree, which was very strong support for monophyly of a Mys- based on 16 morphological characters de- tacinidae ϩ Noctilionidae ϩ Mormoopidae rived from Smith (1972), supported mono- ϩ Phyllostomidae clade in a study of DNA phyly of Mormoopidae, Mormoops, Pteron- hybridization data. Within this group, mono- otus, and a clade consisting of P. personatus, phyly of a clade comprising Mormoopidae ϩ P. quadridens, P. macleayii, and P. davyi Phyllostomidae was also very strongly sup- (Arnold et al., 1982: fig. 2). Mormoopidae, ported, although only one mormoopid (Pter- Noctilionidae, and Phyllostomidae together onotus parnellii) was included in the analy- formed a monophyletic clade Phyllostomo- sis. Relationships of this clade to Noctilion- idea (ϭ Noctilionoidea) in the morphology idae and Mystacinidae remained poorly re- tree, which was rooted using emballonurids solved (Kirsch et al., 1998). (Arnold et al., 1982: fig. 2). Finally, Arnold Simmons (1998) and Simmons and Geis- et al. (1982) presented a new data set derived ler (1998) found support for somewhat dif- from allozyme studies of 16 populations of ferent relationships in a series of analyses of mormoopids and 4 populations of noctilion- higher-level bat relationships based on mor- ids. A cladogram derived from these data phological data. Simmons (1998) and Sim- supported monophyly of Mormoops, Pter- mons and Geisler (1998) found strong sup- onotus, and a clade consisting of P. person- port for Noctilionoidea as traditionally rec- atus, P. quadridens, P. macleayii, and P. ognized (Mormoopidae ϩ Phyllostomidae ϩ davyi (Arnold et al., 1982: fig. 1). Noctilionidae). Within this group, weak sup- Arnold et al. (1982) concluded their paper port was found for a sister-group relationship by presenting a summary ‘‘composite clad- between Mormoopidae and Noctilionidae in ogram’’ that was constructed using principles both analyses. In each case, Mystacina was of taxonomic congruence and majority rule tentatively placed as the most basal branch consensus. The incompatible immunological of the sister clade to Noctilionoidea. How- data were discounted for a variety of meth- ever, support for this placement was weak, odological reasons, and all nodes in the com- and the authors noted that several other po- posite cladogram were ultimately found to be sitions for Mystacinidae were only slightly supported by between one and three of the less parsimonious given their data. DNA sequence data are as yet unavailable 1 Arnold et al. (1982) completed their study before for most mormoopids, but two studies have publication of Sites et al. (1981), so they were not aware addressed relationships of mormoopids in the of the karyotypic evidence for monophyly of Mormoops. context of projects designed to resolve the 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 7

Fig. 2. Arnold et al.’s (1982) ‘‘composite cladogram’’ of relationships among noctilionoid species. This phylogeny was derived from taxonomic-congruence comparisons of trees derived from four data sets: albumin immunology, chromosomes, allozymes, and morphology. Data sets supporting each clade are indicated on the tree. Redrawn from Arnold et al. (1982: fig. 5).

phylogenetic position of Mystacina relative support for mormoopid monophyly was es- to other bat families. Kennedy et al. (1999) sentially nonexistent. Kennedy et al. (1999) used mitochondrial cytochrome-b gene se- additionally found strong support for place- quences to investigate relationships among ment of Mystacina within Noctilionoidea, al- 34 species including Mormoops megalophyl- though the exact position of this taxon rela- la, Pteronotus parnellii, and P. davyi. The tive to Phyllostomidae, Noctilio, and mor- most surprising result of this study was the moopids varied under different analysis pa- finding that Mormoopidae is not monophy- rameters. letic; Noctilio was placed as the sister group Results of another study of mitochondrial of Pteronotus in all analyses, with Mormoops gene sequences produced somewhat different falling outside that clade (Kennedy et al., results. Van Den Bussche and Hoofer (2000) 1999). Support for this arrangement was very analyzed DNA sequence data from three ad- high in bootstrap and decay analyses of a jacent mitochondrial genes (12S rRNA, t- weighted data set (codon positions weighted RNAVal, and 16S rRNA) sampled in 11 noc- 5:16:1, transversions/transitions 2:1), while tilionoids and 17 outgroup species. Results 8 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 of unweighted parsimony analyses supported (1974), Wilkins (1983), Steadman et al. monophyly of Noctilionoidea including Mys- (1984), Eshelman and Morgan (1985), Mor- tacina, and monophyly of each of families gan and Woods (1986), and Morgan (1989, within this group. Mormoopidae and Phyl- 1991). Tertiary mormoopid fossils are as yet lostomidae were found to be sister taxa, unknown (Czaplewski, 1997; McKenna and echoing the results of Kirsch et al. (1998). Bell, 1997). Mystacinidae was placed as the sister group Mormoops blainvillii Leach, 1821 (Antil- to the Mormoopidae ϩ Phyllostomidae lean ghost-faced bat) is a medium-small spe- clade. Mormoopidae, represented by Mor- cies (forearm length ϭ 43Ð50 mm; condy- moops megalophylla and Pteronotus parnel- lobasal length ϭ 12Ð14 mm; 6Ð11 g) pres- lii, was found to be monophyletic in 71Ð88% ently restricted to the Greater Antillean is- of bootstrap replicates (higher values were lands of Cuba, Jamaica, Hispaniola, Puerto obtained using successive approximations Rico, and adjacent small islands (Silva-Ta- weighting). boada, 1979; Koopman, 1993, 1994; Lancas- ter and Kalko, 1996; Emmons, 1997). No MORMOOPID SPECIES:ASYNOPSIS subspecies are currently recognized (Smith, 1972; Hall, 1981; Koopman, 1994). Fossils The principal reference on morphology of M. blainvillii have been reported from and distribution of mormoopid species is caves in the Bahamas, Antigua and Barbuda Smith (1972), but this resource is over 25 in the northern Lesser Antilles, and La Gon- years old and it lacks any substantive con- ave off the west coast of Hispaniola (Koop- sideration of ecology or the fossil record. To man, 1951, 1955, 1989; Steadman et al., help fill these gaps, we provide below a brief 1984; Morgan and Woods, 1986; Morgan, synopsis of each mormoopid species includ- 1989), indicating that the Late Quaternary ing information on presently recognized sub- range of this species was more extensive than species, geographic range, fossil record, hab- it is at present. itat preferences, roosting habits, and dietary Like other mormoopids, M. blainvillii habits. Literature citations are provided to fa- roosts in hot, humid caves (i.e., those with cilitate access to the primary literature. For temperatures 26Ð40ЊC; Goodwin, 1970; Sil- identification purposes, readers should con- va-Taboada, 1979; Rodrõ«guez-Dura«n and sult keys to genera and extant species of Lewis, 1987; Rodr«õguez-Dura«n, 1995, 1998; mormoopids published in Smith (1972), Hall Lancaster and Kalko, 1996). Colonies may (1981), Herd (1983), Eisenberg (1989), and number over 40,000 individuals (Rodr«õguez- Rodr«õguez-Dura«n and Kunz (1992). Another Dura«n and Lewis, 1987). Moths are by far helpful resource is Koopman (1994), which the most common item in the diet, which also includes measurements and some diagnostic includes some beetles, flies, homopterans, features of each extant species. Useful sum- and hemipterans (Silva-Taboada, 1979; Rod- maries of morphological and ecological data r«õguez-Dura«n and Lewis, 1987; Rodr«õguez- for various species were provided by Bate- Dura«n et al., 1993). man and Vaughan (1974), Silva-Taboada Mormoops megalophylla Peters, 1864 (1976, 1979), Herd (1983), Adams (1989), (Peters’ ghost-faced bat; Ghost-faced bat; Eisenberg (1989), Rodrõ«guez-Dura«n and Leaf-chinned bat) is a medium-large species Kunz (1992), Rezsutek and Cameron (1993), (forearm length ϭ 49Ð61 mm; condylobasal Emmons (1997), and Reid (1997). Maps length ϭ 12Ð14 mm; 12Ð20 g) that is broadly showing the geographic ranges of extant distributed in the Caribbean and Central and populations can be found in Hall (1981), South America (Koopman, 1993, 1994; Em- Koopman (1982), Herd (1983), Adams mons, 1997; Reid, 1997). Four subspecies (1989), Eisenberg (1989), Rodr«õguez-Dura«n are currently recognized (Smith, 1972; Rez- and Kunz (1992), Rezsutek and Cameron sutek and Cameron, 1993; Koopman, 1994): (1993), Lancaster and Kalko (1996), and M. m. megalophylla (Baja California, south- Reid (1997). Information on Quaternary fos- ern Arizona, and southern Texas south to sils can be found in Koopman (1951, 1955, Honduras); M. m. tumidiceps (northern Co- 1958, 1989), Ray et al. (1963), Silva Taboada lombia, northern Venezuela, Margarita Island 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 9

and Trinidad); M. m. intermedia (Aruba, Cu- Pteronotus parnellii Gray, 1843 (Par- rac¸ao, and Bonaire islands), and M. m. car- nell’s mustached bat; Mustached bat) is a teri (coastal Ecuador and northwestern Peru). medium-sized species (forearm length ϭ 48Ð Pleistocene fossils of M. megalophylla have 66 mm; condylobasal length ϭ 16Ð22 mm; been found in Florida, Mexico, Cuba, His- 10Ð28 g) that is broadly distributed from the paniola, Jamaica, the Bahamas (Andros), Greater Antilles and tropical Mexico to Aruba, Curac¸ao, Margarita Island, Trinidad, northeastern Brazil, although it does not oc- Tobago, and southeastern Brazil (Koopman, cur west of the Andes in South America (Sil- 1958, 1989; Ray et al., 1963; Silva-Taboada, va-Taboada, 1979; Koopman, 1993, 1994; 1974; Wilkins, 1983; Eshelman and Morgan, Emmons, 1997; Reid, 1997). Nine subspe- 1985; Morgan and Woods, 1986; Morgan, cies are currently recognized (Smith, 1972; 1989, 1991; Arroyo-Cabrales and Alvarez, Linares and Ojasti, 1974; Herd, 1983; Koop- 1990; Arroyo-Cabrales, 1992; Czaplewski man, 1994): P. p. parnellii (ϭ boothi; Cuba and Cartelle, 1998). The records from Flor- and Jamaica); P. p. pusillus (Hispaniola); P. ida, the Bahamas, the Greater Antilles, and p. gonavensis (Gonave Island off the east eastern Brazil (which do not support extant coast of Hispaniola); P. p. portoricensis populations) indicate that the range of M. (Puerto Rico); P. p. mexicanus (Mexico from megalophylla was once more extensive than Sonora and Tamaulipas to Oaxaca and Ve- it is today. racruz); P. p. mesoamericanus (western coast Mormoops megalophylla has been cap- of Central America from Chiapas to western tured at sites from sea level to about 3000 m Panama; eastern coast of Central America in habitats ranging from lowland rainforest from Veracruz and Yucatan to Honduras); P. to semiarid and arid scrub forest (Webb and p. rubiginosus (Honduras to Panama, Trini- Baker, 1962; Smith, 1972; Bateman and dad and Tobago, southern Venezuela to Su- Vaughan, 1974; Albuja, 1982; Graham and rinam, eastern Peru, and northeastern Brazil); Barkley, 1984; Sanchez-Herrera et al., 1986; P. p. fuscus (northeastern Colombia and Bonaccorso et al., 1992; Rezsutek and Cam- northern Venezuela except for the Paraguana eron; 1993; Reid, 1997). This species varies peninsula); and P. p. paraguensis (Paraguana from uncommon to locally common depend- peninusla of northern Venezuela). P. parnel- ing on locality (Reid, 1997), typically ap- lii has recently been found on the island of pearing to be most abundant in hot lowland St. Vincent in the Lesser Antilles, but the areas in the northern parts of its range (Eas- subspecific status of this population has not terla, 1973) and moist forested areas in the yet been assessed (Vaughan and Hill, 1996). tropics (Handley, 1976). Favored roosting The fossil record of P. parnellii within its sites include caves and abandoned mine current range includes late Pleistocene fossils shafts, where these bats occur in colonies that from Puerto Rico and subfossil remains from may include as many as 500,000 individuals Mexico, Cuba, Jamaica, and eastern Brazil (Barbour and Davis, 1969; Bateman and (Martin, 1972; Silva-Taboada, 1979; Herd, Vaughan, 1974; Graham and Barkley, 1984; 1983; Koopman, 1989; Morgan, 1989; Cza- Bonaccorso et al., 1992; Rezsutek and Cam- plewski and Cartelle, 1998). Extinct popu- eron; 1993). The diet of M. megalophylla lations are known from a number of Carib- consists of moths, beetles, and flies; most bean islands including Isla de Pinos, New prey items are relatively large, with body Providence in the Bahamas, La Gonave, An- lengths of 5Ð6 mm (Easterla and Whitaker, tigua, Grand Cayman, and Tobago (Koop- 1972; Ceballos and Galindo, 1984). man, 1955, 1989; Silva-Taboada, 1979; Mormoops magna Silva-Taboada, 1974 Steadman et al., 1984; Eshelman and Mor- is known only from a pair of fossil humeri gan, 1985; Morgan and Woods, 1986; Mor- collected from cave in Cuba (Silva-Taboada, gan, 1989). These records indicate that the 1974, 1979). These elements resemble those Late Quaternary range of P. parnellii was of M. megalophylla but are larger than other somewhat more extensive than it is today. specimens from the same fossil layer that Pteronotus parnellii has been captured at have been referred to M. megalophylla (Sil- sites from sea level to about 3000 m in hab- va-Taboada, 1974, 1979). itats ranging from lowland rainforest to semi- 10 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 arid and arid scrub forest (Smith, 1972; a«n and Kunz, 1992; Koopman, 1994): P. q. Bowles et al., 1979; Silva-Taboada, 1979; quadridens (Cuba) and P. q. fuliginosus (ϭ Herd, 1983; Eisenberg, 1989; Reid, 1997). It inflata; Jamaica, Hispaniola, and Puerto is a relatively common bat throughout much Rico). The fossil record of P. quadridens in- of its range and is abundant at some locali- cludes late Pleistocene and Holocene mate- ties, particularly in moist forests (Smith, rial from several caves in Cuba (Silva-Ta- 1972; Handley, 1976; Silva-Taboada, 1979; boada, 1974, 1979; Woloszyn and Silva-Ta- Herd, 1983; Reid, 1997). Preferred roosting boada, 1977) and the Bahamas (Andros and sites include hot, humid caves and mines New Providence; Morgan, 1989). The latter (Goodwin, 1970; Bowles et al., 1979; Silva- populations are now extinct. Taboada, 1979; Vizotto et al., 1980; Herd, Pteronotus quadridens is one of the most 1983; Bonaccorso et al., 1992; Rodr«õguez- abundant bats in Cuba and Puerto Rico, Dura«n, 1998), but P. parnellii also occurs in where it roosts principally in hot, humid rainforest regions that have no such refugia caves (Silva-Taboada, 1979; Rodr«õguez-Dur- (Simmons and Voss, 1998). In these areas, a«n and Lewis, 1987; Rodr«õguez-Dura«n and this species may occupy cavelike hollows in Kunz, 1992; Rodr«õguez-Dura«n, 1995, 1998). large trees. Colony size may exceed 140,000 individuals The diet of P. parnellii consists principally (Rodr«õguez-Dura«n and Lewis, 1987). of moths, beetles, orthopterans, and flies, Pteronotus quadridens feeds on flying in- with relative proportions of each varying sects that are captured primarily in the forest among populations and seasons (Bateman understory (Rodrõ«guez-Dura«n and Kunz, and Vaughan, 1974; Howell and Burch, 1992). Dietary analyses based on stomach 1974; Silva-Taboada, 1979; Whitaker and contents and fecal samples indicate that a Findley, 1980; Herd, 1983). Other insects oc- wide variety of prey are taken by these bats, casionally taken include odonatans and hy- including beetles, moths, flies, orthopterans, menopterans (Silva-Taboada, 1979). Seeds homopterans, hymenopterans, dyctiopterans, were found in the feces of four individuals dermapterans, hemipterans, and heteropter- (Whitaker and Findley, 1980), but it is not ans (Silva-Taboada, 1979; Rodr«õguez-Dura«n known if fruit is intentionally ingested by and Lewis, 1987; Rodr«õguez-Dura«n et al., this species. P. parnellii is unique among 1993). Remains from as many as seven dif- mormoopids in using long CF (constant fre- ferent families of insects have been found in quency) echolocation calls and Doppler com- samples from a single individual P. quadri- pensation (Novick, 1963, 1965, 1977; Nov- dens, suggesting that these bats may be op- ick and Vaisnys, 1964; Schnitzler, 1970a, portunistic feeders (Silva-Taboada, 1979; 1970b, 1987; Fenton, 1980, 1982, 1984, Rodr«õguez-Dura«n and Kunz, 1992). Small 1994, 1995; Schnitzler and Henson, 1980; amounts of pollen have also been found in Simmons, 1980; Simmons and Stein, 1980; some stomach and fecal samples (Silva-Ta- Neuweiler, 1984, 1989, 1990; Neuweiler and boada, 1979; Rodr«õguez-Dura«n and Lewis, Fenton, 1988; Schnitzler and Kalko, 1998; 1987), but this is thought to be a result of Kalko and Schnitzler, 1998). All prey is ap- feeding of pollen-covered insects rather than parently captured on the wing and there is no intentional feeding directly on flower prod- evidence that feeding roosts are ever used ucts (Rodr«õguez-Dura«n and Kunz, 1992). (Bateman and Vaughan, 1974; Herd, 1983). Pteronotus macleayi Gray 1839 (Ma- Pteronotus quadridens Gundlach, 1840 cleay’s mustached bat) is another small-sized (Sooty mustached bat) is the smallest species species (forearm length ϭ 41Ð46 mm; con- of Pteronotus (forearm length ϭ 35Ð40 mm; dylobasal length ϭ 14Ð16 mm; 4Ð8 g) that condylobasal length ϭ 12Ð14 mm; 3Ð6g; is presently confined to Cuba and Jamaica Silva-Taboada, 1979; Koopman, 1994). The (Silva-Taboada, 1979; Koopman, 1993, geographic range of this species is presently 1994). Two subspecies are currently recog- restricted to several islands in the Greater nized (Smith, 1972; Koopman, 1994): P. m. Antilles (Koopman, 1993, 1994). Two sub- macleayi (Cuba) and P. m. griseus (Jamaica). species are currently recognized (Smith, Fossils of this species are known from the 1972; Silva-Taboada, 1979; Rodr«õguez-Dur- late Quaternary of the Bahamas (New Prov- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 11

idence), which indicates that the range of P. Three subspecies are currently recognized macleayi once extended farther north than it (Smith, 1972; Adams, 1989; Koopman, does today (Morgan, 1989). 1994): P. d. fulvus (Sonora and Nuevo Leon Like other mormoopids, P. macleayi usu- to Honduras), P. d. davyi (Nicaragua to ally roosts in hot, humid caves (Goodwin, coastal Venezuela and Trinidad, and north in 1970; Smith, 1972, Silva-Taboada, 1979). the Lesser Antilles to Marie Galante [exclud- Analyses of stomach contents indicate that ing St. Lucia, St. Vincent, Barbados, and the this species preys only on flying insects, in- Grenadines]), and P. d. incae (northwestern cluding flies, beetles, orthopterans, and ho- Peru on both sides of the Andes). Although mopterans (Silva-Taboada, 1979). Unlike Vieira (1955), Mares et al. (1981), and Willig other mormoopids, P. macleayi apparently (1983, 1985) reported P. davyi from north- does not eat moths (Silva-Taboada, 1979). eastern Brazil, these identifications were ap- Pteronotus personatus Wagner, 1843 parently based on the erroneous opinion that (Wagner’s mustached bat; Mustached bat) is P. davyi is a senior synonym of P. gynon- a medium-sized species (forearm length ϭ ontus; extant populations of true P. davyi re- 40Ð49 mm; condylobasal length ϭ 13Ð16 main unknown south of the Amazon River mm; 5Ð11 g) that is broadly distributed in (Smith, 1977; Willig and Mares, 1989; mainland Central and South America Koopman, 1993; da Fonseca et al., 1996; (Bowles et al., 1979; Koopman, 1993, 1994; Czaplewski and Cartelle, 1998). However, Emmons, 1997; Reid, 1997). Two subspecies Pleistocene fossils of P. davyi have been are currently recognized (Smith, 1972; Bros- found in Bahia in eastern Brazil (Czaplewski set and Charles-Dominique, 1990; Koopman, and Cartelle, 1998), suggesting that the range 1994): P. p. psilotis (Sonora and Tamaulipas of this species was once much larger. Qua- to Honduras) and P. p. personatus (Nicara- ternary fossils of P. davyi have been found gua south to French Guiana and eastern Peru, in Tobago, where this species is now extinct from there east to northeastern Brazil). Qua- (Eshelman and Morgan, 1985). ternary fossils of P. personatus have been Pteronotus davyi has been captured at sites found in Tobago, where this species is now from sea level to about 2300 m in habitats extinct (Eshelman and Morgan, 1985). that range from rainforest to dry deciduous Pteronotus personatus has been captured forest (Smith, 1972; Koopman, 1978, 1982; at sites from sea level to about 1000 m in Adams, 1989; Reid, 1997). These bats may habitats that range from rainforest to dry de- be locally common in dry and semideciduous ciduous forest (Smith, 1972; Handley, 1976; lowland forest, but are less common in moist, Bowles et al., 1979; Brosset and Charles- evergreen forest (Handley, 1976; Reid, Dominique, 1990; Reid, 1997). This species 1997). Hot, humid caves are the preferred is relatively uncommon to locally common roosting site, and colonies may include as in Central America, and is less common many as 400,000Ð800,000 bats (Goodwin south of Honduras (Reid, 1997). The pre- and Greenhall, 1961; Villa-R., 1967; Bate- ferred roosting site is hot, humid caves, man and Vaughan, 1974; Birney et al., 1974; where colonies may exceed 15,000 individ- Bonaccorso et al., 1992). The diet of P. davyi uals (Dalquest and Hall, 1949; Smith, 1972; consists of moths, some flies, and occasion- Bateman and Vaughan, 1974; Bowles et al., ally earwigs (forficulids; Villa-R, 1967; 1979; Vizotto et al., 1980; Bonaccorso et al., Howell and Burch, 1974; Adams, 1989). 1992). Prey is captured on the wing and there is no Pteronotus davyi Gray, 1838 (Davy’s na- evidence that feeding roosts are ever used ked-backed bat; Lesser naked-backed bat; (Bateman and Vaughan, 1974). Little naked-backed bat; Trinidadian naked- Pteronotus gymnonotus Wagner, 1843 backed bat) is a medium-small species (fore- (Big naked-backed bat; Greater naked- arm length ϭ 40Ð50 mm; condylobasal backed bat) is a medium-large species (fore- length ϭ 13Ð16 mm; 5Ð11 g) that is broadly arm length ϭ 49Ð56 mm; condylobasal distributed in Central and South America and length ϭ 15Ð17 mm; 11Ð18 g) that ranges the Lesser Antilles (Jones, 1989; Koopman, from Veracruz to French Guiana and central 1993, 1994; Emmons, 1997; Reid, 1997). Brazil, although it does not occur west of the 12 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Andes in South America (Koopman, 1993, pristinus to extant taxa. We chose to focus 1994; Brosset et al., 1996; Emmons, 1997; on morphological data in this study because Reid, 1997). No subspecies are currently rec- of our past experience with such data sets, ognized (Smith, 1972; Koopman, 1994). and because of the promising patterns of Quaternary fossils of P. gymnonotus have character variation among mormoopids de- been found in Tobago, where this species is scribed by Smith (1972). Our objective was now extinct (Eshelman and Morgan, 1985). to produce a relatively well-resolved, well- Pteronotus gymnonotus has been captured supported phylogeny of Mormoopidae that at sites from sea level to about 3000 m in could provide both a reference point for fu- habitats that range from rainforest to dry ture phylogenetic analyses of molecular se- semideciduous forest (Smith, 1972; Reid, quence data and serve as a basis for our on- 1997). This species is typically uncommon going work on morphological evolution, his- but may be locally abundant, particularly in torical biogeography, and the evolution of dry, relatively open areas (Handley, 1966, echolocation call structure and foraging strat- 1976; Reid, 1997). P. gymnonotus apparently egies. prefers to roost in caves, often with other mormoopids, and colonies may number in MATERIALS AND METHODS the thousands (Vizotto et al., 1980). Little is TAXONOMIC SAMPLING,OUTGROUPS, AND known about the diet of P. gymnonotus, but TREE ROOTING available data suggest that these bats con- sume mainly orthopterans, beetles, and Our taxonomic sample included all extant moths (Howell and Burch; 1974; Whitaker species presently recognized within Mor- and Findley, 1980). moopidae (see above). Mormoops magna, Pteronotus pristinus Silva-Taboada, known only from fossil humeri, was judged 1974 is a medium-sized species (greatest too incomplete to be included in our study. length of skull ϭ 17Ð18 mm) known only Pteronotus pristinus, another fossil taxon, from fossils collected from caves in Cuba was included in some but not all analyses. A (Silva-Taboada, 1974, 1979). Available ma- list of specimens examined is provided in ap- terial includes several complete skulls plus pendix 1. assorted postcranial elements (Silva-Taboa- In gathering data for this study, we ini- da, 1974, 1979). Two mandibles from a fossil tially scored each currently recognized mor- site in central Florida were referred to ‘‘Pter- moopid subspecies as a separate taxon in or- onotus cf. P. pristinus’’ by Morgan (1991). der to facilitate preliminary tests of Smith’s These specimens may represent P. pristinus; (1972) hypotheses of species limits. Exami- Morgan’s (1991) identification was left as nation of the resulting data matrix revealed tentative only because he was unable to di- many gaps (particularly in soft-tissue char- rectly compare the Florida fossils with any acters) but no substantive within-species of the Cuban material. character variation in mormoopids. We there- fore pooled our observations by species for GOALS OF THE PRESENT STUDY the ingroup mormoopid taxa, which simul- taneously reduced redundancy and increased The present study was designed with four completeness within the data set. With re- explicit goals: (1) to test the monophyly of spect to species limits within Mormoopidae, the family Mormoopidae and provide a re- we concluded that available discrete charac- vised diagnosis of this group if monophyly ter data are entirely consistent with Smith’s can be demonstrated; (2) to identify the sister (1972) taxonomic conclusions. Further spe- group of Mormoopidae; (3) to test previous cies-level revisionary work on mormoopids hypotheses of interrelationships among mor- will require morphometric and/or biochemi- moopid species (e.g., those of Smith [1972] cal analyses far beyond the scope of the pre- and Arnold et al. [1982]) and provide revised sent study. diagnoses of all superspecific clades (e.g., Only one outgroup is necessary to root a genera and subgenera); and (4) to evaluate phylogenetic tree (Nixon and Carpenter, the affinities of the fossil species Pteronotus 1993), but at least two outgroups are usually 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 13

included in cladistic analyses to establish SOURCES OF DATA character polarities and to permit testing of Within the limits of specimen availability, ingroup monophyly (Maddison et al., 1984). we examined standard museum specimens Ideally, outgroups should comprise the near- (dried skins, skulls, skeletons, and fluid-pre- est sister-taxa to the ingroup because the served material) of all extant species includ- probability of homoplasy increases with time ed in our study (see appendix 1 for speci- since divergence from a common ancestor. mens examined). A single skull of the fossil As discussed in the Introduction, Mormoop- species Pteronotus pristinus was also includ- idae is widely regarded as belonging to the ed in our study (fig. 3). We initially obtained monophyletic superfamily Noctilionoidea, morphological information from the litera- where phylogenetic analyses have placed it ture (e.g., Smith, 1972; Hill and Daniel, either with Noctilionidae (e.g., Patton and 1985), but subsequently we scored most cra- Baker, 1978; Arnold et al., 1982; Simmons, nial, dental, skeletal, and external characters 1998; Simmons and Geisler, 1998; Kennedy directly from specimens. We did not collect et al., 1999) or as the sister group of Phyl- any new information on features of the vom- lostomidae (Van Valen, 1979; Novacek, eronasal complex and brain (characters 48Ð 1991; Kirsch et al., 1998; Van Den Bussche 52), hyoid apparatus (characters 54Ð71), hair and Hoofer, 2000). Several authors have con- microstructure (characters 108Ð110), post- vincingly argued that another family, Mys- cranial myology (characters 177Ð196), repro- tacinidae, is also closely related to these taxa ductive tract (characters 200Ð204), or diges- (Pierson, 1986; Pierson et al., 1986; Kirsch tive tract (characters 205Ð209). As described et al., 1998; Kennedy et al., 1999; Van Den below under the character descriptions, these Bussche and Hoofer, 2000). features were scored directly from literature Based on the studies cited above, we in- accounts. cluded representatives of three noctilionoid The following institutional abbreviations families in our study as putative outgroups: are used in the text, figures, and appendices: Phyllostomidae (represented by Macrotus AMNH, American Museum of Natural waterhousii, Macrotus californicus, and Ar- History, New York, New York, USA; tibeus jamaicensis), Noctilionidae (Noctilio MNHN, Muse«um National d’Histoire Natu- leporinus, Noctilio albiventris), and Mysta- relle, Paris, France; MVZ, Museum of Ver- cinidae (Mystacina tuberculata and Mysta- tebrate Zoology, University of California, cina robusta). Because relationships among Berkeley, California, USA; ROM, Royal these families remain uncertain, we addition- Ontario Museum, Toronto, Canada; USNM, ally included a member of the family Em- National Museum of Natural History, Smith- ballonuridae (Saccopteryx bilineata)asa sonian Institution, Washington, D.C., USA. more distal outgroup. Recent studies all agree that Emballonuridae falls outside the DEFINITION OF CHARACTERS AND ORDERING smallest clade containing the noctilionoid OF CHARACTER STATES taxa listed above (e.g., Pierson, 1986; Pier- Most of the characters employed in the son et al., 1986; Kirsch et al., 1998; Sim- present study reflect features that vary among mons, 1998; Simmons and Geisler, 1998; mormoopids and are thus potentially phylo- Van Den Bussche and Hoofer, 2000). Ac- genetically informative with respect to spe- cordingly, we rooted our trees using Saccop- cies-level relationships in the family. How- teryx bilineata, and allowed the relative po- ever, the goals of our study included testing sitions of the noctilionoid taxa (including mormoopid monophyly and identification of mystacinids) to vary. Although we could not the sister group of Mormoopidae, so we also score every species for all of the characters included characters representing putative in our data set, we chose to use species (rath- synapomorphies of Mormoopidae. Other er than genera or families) as outgroup OTUs characters relevant to relationships among to facilitate future combination of our data the outgroups were also included in our data with molecular data sets. set (even if they are invariant within Mor- 14 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Fig. 3. Dorsal (A), ventral (B), and lateral (C) views of the skull of Pteronotus pristinus (ROM 59132), a fossil species known only from Cuba. Scale bar ϭ 10 mm. 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 15

moopidae) so as to contribute structure to the multistate characters described progressive basal part of the tree. gradations in size, shape, degree of devel- Our methods of character definition and opment, meristic counts, or color patterns, description followed Simmons (1993, 1998) and these were treated as ordered characters and Simmons and Geisler (1998). Unlike in many of our analyses. The proscribed se- those studies, however, we chose to include quences of character-state transformation putative autapomorphies (derived character (and our justification for expecting changes states observed in only one terminal taxon) to be restricted to these sequences) are dis- in our data set. In this we follow Omland cussed below under the relevant character (1997) and Livezey (1998), who have argued descriptions. that such characters contribute to estimates of overall divergence. Formal scoring of such POLARITY traits additionally facilitates development of Patterns of distribution of character states taxonomic diagnoses for terminal taxa as among multiple outgroups are often used to well as future integration of data from spe- provide preliminary assessments of character cies not included in the original study. The polarity prior to phylogenetic analyses (e.g., few instances of within-species polymor- Simmons and Geisler, 1998). Although the phism detected were treated by erecting al- present study included multiple putative out- ternative character states to describe poly- groups, only one taxon (Saccopteryx bilinea- morphic conditions. ta) was used to root the tree. Suggesting that Two categories of missing data are com- character states seen in this species necessar- monly recognized in phylogenetic analyses: ily represent the primitive condition would (1) missing data that result from lack of spec- be misleading, particularly since Emballon- imens or appropriate observational data (e.g., uridae is highly derived in some respects and soft-tissue characters that cannot be scored in the family has been distinct since at least the fossil taxa or rare extant species), and (2) Middle Eocene (Simmons and Geisler, missing data that result from evolutionary 1998). Accordingly, we do not provide a modifications that may render characters in- priori assessments of character polarity be- applicable under some conditions (e.g., pres- low. A posteriori conclusions concerning po- ence/absence of a hypocone on M3 cannot larity of selected characters are discussed in be scored in taxa that lack M3). We followed the context of taxonomic diagnoses in the convention in scoring the former type of ‘‘Discussion and Conclusions’’ section be- missing data as ‘‘?’’ and the latter as ‘‘–’’ in low. our data matrix. Our final data set consisted of a total of COMPLETENESS 209 characters: 27 characters of the skull, 20 dental characters, 5 characters of the vom- Data completeness is a problem that per- eronasal complex and brain, 19 characters of petually plagues phylogenetic studies that in- the trachea and hyoid apparatus, 5 characters clude fossils and/or difficult-to-study soft-tis- of the tongue, 26 characters describing fea- sue characters. Completeness of a taxon may tures of the face, ears, and vibrissae, 14 char- be defined as the percentage of characters for acters of the pelage and patagia, 60 charac- which it can be scored in a given analysis ters of the postcranial skeleton, 21 characters (Simmons, 1993; Simmons and Geisler, of postcranial myology, 7 features of the re- 1998). In the context of morphological data productive tract, and 5 characters of the di- matrices, fossil taxa are typically much less gestive system. Character-state data were re- complete than extant taxa, and rare taxa corded in a taxon-character matrix using (which are often omitted from studies that MacClade version 3.0 (Maddison and Mad- require destructive sampling) are typically dison, 1992) and were imported into other less complete than more common taxa. Data software for analysis. While most of our gaps resulting from these patterns may re- characters were binary (155 ϭ 74%), our duce both phylogenetic resolution and per- data set included a number of multistate ceived level of clade support in phylogenetic characters (54 ϭ 26%). A total of 34 of these analyses. However, the effects of these prob- 16 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

TABLE 1 Statistics on Character Codings for Mormoopid Data Set

lems are frequently overestimated, and inclu- shortest trees and to calculate tree statistics. sion of fossil taxa and soft-tissue characters Near-most-parsimonious trees (one to five may have significant positive effects on the steps longer) were identified in subsequent outcome of an analysis (see discussion in branch-and-bound analyses, and a decay Simmons and Geisler [1998] and references analysis was performed following the meth- cited therein). We therefore followed many ods of Bremer (1988). A branch-and-bound recent authors in concluding that the poten- bootstrap analysis with 10,000 replicates and tial benefits of including fossils and soft-tis- a character jackknife analysis (50% character sue characters in phylogenetic studies far deletion, branch-and-bound, 10,000 repli- outweigh the analytical problems that they cates) were additionally used to evaluate rel- may introduce (Gauthier et al., 1988; Dono- ative support for various groupings. ghue et al., 1989; Novacek, 1992, 1994; Sim- Two analyses were conducted including mons, 1993; Simmons and Geisler, 1998). different subsets of taxa: one including only Simmons and Geisler (1998) argued that cal- extant species (excluding Pteronotus pristin- culating the percent completeness of each us), and a second analysis including all taxa. taxon in a given data set can provide a useful To test the effects of ordering multistate char- basis for a posteriori consideration of the ef- acters on the resulting tree topology, we also fects of completeness on the analysis. These conducted a separate analysis of the com- data for our study are presented in table 1. plete data set with all characters unordered. Additional analyses of our complete data set (with characters ordered) were also per- METHODS OF PHYLOGENETIC ANALYSIS formed to find the shortest trees compatible Phylogenetic analyses were conducted us- with the phylogenies previously published by ing the parsimony algorithm implemented in Smith (1972), Arnold et al. (1982), and Ken- PAUP* version 4.0b3a (Swofford, 2000). nedy et al. (1999). Results and comparisons The branch-and-bound option of PAUP was of these analyses are presented under ‘‘Re- used to identify all equally parsimonious sults’’ below. 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 17

Character optimizations were calculated tures are not necessarily linked in bats. Some using both the ACTRAN (accelerated trans- taxa that have freely-movable premaxilla formation optimization) and DELTRAN (de- also have well-developed palatal processes layed transformation optimization) as imple- that are fused across the midline (e.g., nyc- mented in PAUP* version 4.0b3a (Swofford, terids). Accordingly, we treated these as sep- 2000). Discussions of the similarities and dif- arate characters in our analysis. ferences between these methods can be Character 3: Pair of incisive foramina found in Simmons (1993) and Simmons and present (0); or incisive foramina absent (1). Geisler (1998). MacClade version 3.0 (Mad- A pair of incisive foramina, bordered ante- dison and Maddison, 1992) was used to vi- riorly, medially, and laterally by the premax- sualize the results of character mapping. illae and posteriorly by the maxillae, is found in the anterior palate of most taxa including CHARACTER DESCRIPTIONS all species of Pteronotus and Mormoops. Among the outgroups, a similar condition is SKULL seen in Artibeus jamaicensis, Macrotus wa- Character 1: Premaxilla fused to maxilla terhousii, Macrotus californicus, Mystacina (0); or premaxilla articulates with maxilla tuberculata, and Mystacina robusta. In con- via ligaments, premaxilla freely movable (1). trast, the incisive foramina are absent in both The nasal branch of the premaxilla is defined species of Noctilio, which lack any perfora- as that portion of the premaxilla that lies on tions in the anterior palate. This character the face adjacent to the narial opening. The cannot be scored in Saccopteryx bilineata, nasal branch of the premaxilla is fused to the which lacks a premaxillary contribution to maxilla in all mormoopids and most of the the anterior palate (see character 2). outgroups (Artibeus jamaicensis, Macrotus Character 4: Hard palate extends poste- waterhousii, Macrotus californicus, Mysta- riorly into interorbital region (0); or termi- cina tuberculata, Mystacina robusta, Nocti- nates at the level of the zygomatic roots (1). lio leporinus, Noctilio albiventris). In con- The hard palate, which forms a bony sepa- trast, the premaxilla in Saccopteryx bilineata ration between the oral and nasal passages, articulates with the maxilla via ligaments. terminates posteriorly at the mesopterygoid This arrangement allows the premaxillae to fossae. The position of the posterior edge of move freely with respect to the rest of the the hard palate varies independently of the facial skeleton. posterior extent of the molar toothrow, and Character 2: Palatal branch of premax- is apparently linked to the structure of the illa well-developed, right and left premaxil- nasal passages and soft tissues of the pha- lae fused across midline of palate (0); or pal- ryngeal region. The hard palate extends pos- atal branch absent, right and left premaxilla teriorly into the interorbital region in mor- not in contact on palate (1). The palatal moopids and most of the outgroups (Artibeus branches of the premaxillae are defined as jamaicensis, Macrotus waterhousii, Macro- those portion of the premaxillae that contrib- tus californicus, Mystacina tuberculata, Mys- ute to the anterior palate. The palatal branch- tacina robusta, Noctilio leporinus, Noctilio es of the premaxilla are well-developed and albiventris). In contrast, the hard palate ter- the right and left premaxillae are fused across minates at the level of the zygomatic roots the midline of the palate in all mormoopids in Saccopteryx bilineata. and most of the outgroups (Artibeus jamai- Character 5: Rostrum not upturned, angle censis, Macrotus waterhousii, Macrotus cal- between long axis of anterior half of zygo- ifornicus, Mystacina tuberculata, Mystacina matic arch and occlusal surface of molar robusta, Noctilio leporinus, Noctilio albiven- toothrow less than 10Њ (0); or rostrum slight- tris). In contrast, the palatal branches of the ly upturned, angle 13Ð26Њ (1); or rostrum premaxilla are absent in Saccopteryx bilinea- strongly upturned, angle 28Ð34Њ (2). The ta. As a result, the right and left premaxillae long axis of the rostrum lies roughly parallel do not contact one another in the palatal re- to the long axis of the zygomatic arch in gion. Although this character has the same many bats, resulting in a skull morphology distribution of states as character 1, these fea- that lacks any obvious ‘‘upturning’’ of the 18 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 rostrum relative to the braincase. Measuring slightly upturned rostrum, but lack concave the degree of upturning of the rostrum is nasals. Accordingly, we chose to treat these complicated by numerous factors, including as separate characters in our analysis. modifications of the zygomatic arch (e.g., Character 7: Nasal foramina small or ab- dorsoventral curvature) and modifications of sent (0); or pair of large foramina present in the rostrum (e.g., foreshortening and dorsal posterior nasal region below forehead (1). curvature). In noctilionoids, the anterior half The nasal bones of most bats are imperforate of the zygomatic arch usually appears or include only a few small foramina. This straight in lateral view, and the occlusal sur- condition is seen in all mormoopids and most face of the first and second molar teeth (M1 of the outgroups (Artibeus jamaicensis, Ma- and M2) describe a plane with which zygo- crotus waterhousii, Macrotus californicus, matic arch orientation may be compared. The Mystacina tuberculata, Mystacina robusta, occlusal plane of molar teeth lies at an angle Saccopteryx bilineata). In contrast, both spe- of 13Ð26Њ to the long axis of the anterior half cies of Noctilio are characterized by presence of the zygomatic arch in all species of Pter- of a pair of large foramina in the posterior onotus, giving these bats the appearance of portion of the nasals just below the forehead. having a slightly upturned rostrum (fig. 4). Character 8: Maximum rostral breadth A more extreme condition is seen in both less than or equal to length of maxillary species of Mormoops, which have a strongly- toothrow (0); or greater than length of max- upturned rostrum characterized by an angle illary toothrow (1). The shape of the rostrum of 28Ð34Њ between the zygomatic arch and in dorsal view can be described by compar- molar occlusal plane (fig. 4). Among the out- ing maximum breadth of the rostrum (which groups, a slightly upturned rostrum (mea- usually occurs at the level of the last pre- sured using the criteria described above) is molar or first molar) with the length of the seen in Artibeus jamaicensis, Macrotus wa- maxillary toothrow. Pteronotus parnellii, P. terhousii, Macrotus californicus, Mystacina macleayii, P. quadridens, P. personatus, P. tuberculata, Noctilio leporinus, and Noctilio pristinus, and both species of Mormoops albiventris. Both species of Mystacina and have a rostral breadth that is less than or Saccopteryx bilineata are characterized by a equal to the length of the maxillary toothrow. different condition in which the rostrum is In contrast, Pteronotus davyi and P. gymnon- not upturned, and the occlusal plane of molar otus have a rostral breadth that is greater than teeth lies at an angle of less than 10Њ to the the length of the maxillary toothrow. Similar long axis of the anterior half of the zygo- variation is seen among the outgroups. Sac- matic arch. Transformations in this character copteryx bilineata, both species of Mystaci- were ordered 0 ↔ 1 ↔ 2toreflect the logical na, and both species of Macrotus have a ros- sequence of transformation from an unflexed tral breadth that is less than or equal to the rostrum (‘‘0’’) to a strongly upturned rostrum length of the maxillary toothrow, while Ar- (‘‘2’’). tibeus jamaicensis and both species of Noc- Character 6: Nasals flat or convex (0); or tilio have a rostral breadth that is greater than nasals concave (1). The nasal bones in most the length of the maxillary toothrow. bats are either flat or convex upward when Character 9: Rostrum length less than seen in lateral view. In contrast, all species one-half of total length of skull (0); or equal of Pteronotus and Mormoops have nasals to or greater than one-half of length of skull that are concave upward (fig. 4). The out- (1). The relative length of the rostrum in bats groups all have nasals that are flat or convex, varies independently of rostral breadth. We suggesting that concave nasals represent a defined rostral length as the distance between derived condition. This character appears to the anteriormost point on the skull (exclud- be correlated with presence of an upturned ing teeth) and the interorbital constriction rostrum in mormoopids (see character 1), but (the point of minimum skull breadth in the these features are decoupled in the out- orbital region). Skull length was defined as groups. Artibeus jamaicensis, Macrotus wa- the greatest length of the skull excluding terhousii, Macrotus californicus, Noctilio le- teeth. Pteronotus parnellii, P. pristinus, P. porinus, and Noctilio albiventris each have a personatus, P. davyi, P. gymnonotus, and 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 19

Fig. 4. Lateral views of the skulls of (A) Mormoops blainvillii (AMNH 45197), (B) Pteronotus davyi (AMNH 204965), (C) Mormoops megalophylla (AMNH 27300), (D) Pteronotus gymnonotus (AMNH 32092), (E) Pteronotus parnellii (AMNH 129749), (F) Pteronotus macleayii (AMNH 60899), (G) Pteronotus personatus (AMNH 32136), and (H) Pteronotus quadridens (AMNH 39357). Note the differences in degree of upturn of the rostrum relative to the braincase, and in the position of the infraorbital foramen and anterior rim of the orbit relative to the toothrow. Scale bars ϭ 5 mm. 20 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Fig. 5. Dorsal views of the skulls of (A) Mormoops blainvillii (AMNH 45197), (B) Mormoops megalophylla (AMNH 27300), (C) Pteronotus davyi (AMNH 204965), (D) Pteronotus gymnonotus (AMNH 32092), (E) Pteronotus parnellii (AMNH 129749), (F) Pteronotus personatus (AMNH 32136), (G) Pteronotus macleayii (AMNH 60899), and (H) Pteronotus quadridens (AMNH 39357). Note the differences in relative proportions of the skull (e.g., breadth of the rostrum, zygomatic arches, and mastoid region compared with skull length). Scale bars ϭ 5 mm. both species of Mormoops are characterized Character 10: Infraorbital foramen not by having a rostral length that is less than enlarged, diameter less than one-eighth the one-half the length of the skull (fig. 5). In height of the rostrum (0); or enlarged, di- contrast, the rostrum is noticeably longer in ameter one-quarter to one-half the height of Pteronotus quadridens and P. macleayi, the rostrum (1). The infraorbital foramen, where rostral length (as defined above) is which is located in the maxilla, marks the equal to or greater than one-half of the total anterior opening of the infraorbital canal. length of the skull (fig. 5). All of the out- The size of the infraorbital foramen can be groups have a rostrum length that is less than evaluated by comparing diameter of the fo- one-half of the length of the skull. ramen to the height of the rostrum above M1. 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 21

The infraorbital foramen is not enlarged (i.e., the eye, which typically occupies the anter- foramen diameter less than one-eighth the iormost portion of the orbit, and the length height of the rostrum) in all extant species of of the infraorbital canal, which begins in the Pteronotus and Mormoops (fig. 4). In con- anterior orbit and terminates on the face at trast, the infraorbital canal is enlarged and its the infraorbital foramen (see character 11). diameter is approximately one-quarter to The anterior rim of the orbit terminates one-third the height of the rostrum in Pter- above (dorsal to) the posterior half of M2 in onotus pristinus (fig. 3) Among the out- all species of Pteronotus (fig. 4). In contrast, groups, the infraorbital foramen is not en- the anterior orbital rim terminates above M1 larged in Artibeus jamaicensis, Macrotus wa- in both species of Mormoops (fig. 4). Among terhousii, Macrotus californicus, Noctilio le- the outgroups, the anterior rim of the orbit porinus, Noctilio albiventris, and Saccopteryx terminates above P4 in Saccopteryx bilinea- bilineata, but this foramen is enlarged (one- ta, above M1 in both species of Noctilio, half the height of the rostrum) in both species above the anterior half of M2 in Artibeus ja- of Mystacina. maicensis and both species of Macrotus, and Character 11: Infraorbital foramen locat- above the posterior half of M2 in both spe- ed above posterior half of P4 (0); or above cies of Mystacina. Transformations in this anterior half of M1 (1); or above posterior character were ordered 0 ↔ 1 ↔ 2 ↔ 3to half of M1 (2); or anterior half of M2 (3). reflect our assumption that changes in the po- The infraorbital foramen in Pteronotus per- sition of the infraorbital foramen relative to sonatus is located directly above (dorsal to) the toothrow occur in a progressive fashion. the anterior half of M2, roughly above the Character 13: Postorbital process present parastyle (fig. 4). In contrast, the infraorbital (0); or absent (1). The postorbital process is foramen is located above the posterior half a laterally projecting process of the frontal of M1 (above the metacone) in P. davyi and that forms part of the posterodorsal rim of P. gymnonotus (fig. 4). The infraorbital fo- the orbit. A postorbital process is absent in ramen is located above the anterior half of all mormoopids (fig. 5) and most of the out- M1 (above the paracone) in P. parnellii, P. groups. In contrast, a well-developed, elon- macleayi, P. quadridens, and P. pristinus gate postorbital process is present in Saccop- (figs. 3, 4). Both species of Mormoops ex- teryx bilineata. hibit a fourth condition in which the infra- Character 14: Zygomatic breadth greater orbital foramen occurs above the posterior than mastoid breadth (0); or less than mas- half of P4 (fig. 4). Among the outgroups, toid breadth (1). The maximum breadth of both species of Mystacina have an infraor- the skull in bats may occur in either the zy- bital foramen that opens above the posterior gomatic region anterior to the braincase or in half of M1. Both species of Noctilio and both the mastoid region at the posterior end of the species of Macrotus have the foramen above braincase. Zygomatic breadth (the maximum the anterior half of M1, and the infraorbital breadth of the skull measured across the zy- foramen opens above the posterior half of P4 gomatic arches) is greater than mastoid in Artibeus jamaicensis and Saccopteryx bil- breadth (maximum breadth of the skull mea- ineata. Transformations in this character sured across the mastoid region) in Pteron- were ordered 0 ↔ 1 ↔ 2 ↔ 3toreflect our otus parnellii, P. pristinus, P. davyi, P gym- assumption that changes in the position of nonotus, and both species of Mormoops (figs. the infraorbital foramen relative to the tooth- 3, 5). In contrast, mastoid breadth is greater row occur in a progressive fashion. than zygomatic breadth in Pteronotus per- Character 12: Anterior rim of orbit ter- sonatus, P. quadridens, and P. macleayii. minates above P4 (0); or above M1 (1); or Zygomatic breadth is greater than mastoid above anterior half of M2 (2); or above pos- breadth in all of the outgroups. terior half of M2 (3). The anterior terminus Character 15: Parietals not inflated (0); of the orbit, which in most bats is marked by or inflated in area encasing cerebellum (1). a sharply demarcated rim, may occur in dif- In most bats the parietal bones, which cover ferent positions relative to the toothrow. This the brain in the region of the cerebellum, are variation affects both the relative position of not notably inflated with respect to the bones 22 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 of the anterior braincase (i.e., the frontals). imum diameter Ͼ25% of the external width In contrast, the parietal bones are markedly of the first half turn of the cochlea. A similar inflated over the cerebellum in Mormoops condition is seen in most of the outgroups. (fig. 5). All species of Pteronotus lack in- In contrast, the fenestra cochleae is smaller flated parietals (fig. 5), as do all of the out- (maximum diameter Ͻ20% of the external groups. width of the first half turn) in Saccopteryx Character 16: Epitympanic recess shal- bilineata. Data were not available for Pter- low and broad (0); or deep, constricted in onotus pristinus, which was scored ‘‘?’’ for area (1). The epitympanic recess in bats lies this character. in the dorsolateral roof of the middle-ear Character 19: Pars cochlearis of petrosal space, where it houses the incus and the joint sutured to basisphenoid (0); or loosely at- between the malleus and incus (Simmons tached to basisphenoid via ligaments and/or and Geisler, 1998). The epitympanic recess thin splints of bone (1). The pars cochlearis is deep and constricted in area in all species of the petrosal (which in microchiropterans of Pteronotus and Mormoops. Among the consists principally of the bone enclosing the outgroups, a similar condition is seen in Sac- enlarged cochlea) is loosely attached to the copteryx bilineata, both species of Noctilio, basisphenoid via ligaments and/or thin and both species of Mystacina. In contrast, splints of bone in all species of Mormoopids the epitympanic recess is shallow and broad and Pteronotus. A similar condition is seen in Artibeus jamaicensis and both species of in most of the outgroups. In contrast, the pars Macrotus. cochlearis of the petrosal is sutured to the Character 17: Fossa for m. stapedius in- basisphenoid in Saccopteryx bilineata. Data distinct (0); or deep, constricted in area, were not available for Pteronotus pristinus, forms a crescent-shaped fissure (1). The fos- which was scored ‘‘?’’ for this character. sa for the origin of m. stapedius is located in Character 20: Cochlea phanerocochlear the posterior roof of the middle-ear space (0); or cryptocochlear (1). Novacek (1985b, (Simmons and Geisler, 1998). When well de- 1991) described variation in cochlear struc- fined, it occupies a cavity dorsal to the crista ture among bats and recognized two distinct parotica (Simmons and Geisler, 1998). The patterns of petrosal ossification and adult co- fossa for m. stapedius is indistinct in all spe- chlear morphology. The ‘‘phanerocochlear’’ cies of Pteronotus and Mormoops. Among state occurs when the petrosal wall is thin the outgroups, a similar condition is seen in and poorly ossified, resulting in a condition both species of Mystacina. In contrast, the where the cochlear labyrinth is clearly visible fossa for the origin of m. stapedius is deep externally in the adult (Novacek, 1985b, and constricted in area in Saccopteryx bili- 1991). In contrast, a ‘‘cryptocochlear’’ con- neata, both species of Noctilio, Artibeus ja- dition occurs when strong petrosal ossifica- maicensis, and both species of Macrotus. In tion produces a thicker encasement of bone these taxa, the fossa for m. stapedius forms around the cochlea, hiding the cochlear lab- a crescent-shaped fissure. yrinth from external view in adults (Nova- Character 18: Fenestra cochleae small or cek, 1985b, 1991). Pteronotus parnellii ex- of moderate size, maximum diameter Ͻ20% hibits the cryptocochlear condition. In con- of the external width of the first half turn of trast, all other extant species of Pteronotus the cochlea (0); or enlarged, maximum di- and both species of Mormoops have a pha- ameter Ͼ25% of the external width of the nerocochlear cochlea. Among the outgroups, first half turn of the cochlea (1). The fenestra the phanerocochlear condition occurs in Sac- cochleae (ϭ fenestra rotundum) is a mem- copteryx bilineata, Artibeus jamaicensis, brane-covered opening in the tympanic wall both species of Macrotus, and both species of the petrosal. It faces posteriorly or pos- of Mystacina. Both species of Noctilio have terolaterally, and separates the scala tympani a cryptocochlear cochlea. Data were not (at the base of the cochlear labyrinth) from available for Pteronotus pristinus, which was the middle-ear cavity. The fenestra cochleae scored ‘‘?’’ for this character. appears relative large in all species of Pter- Character 21: Cochlea greatly enlarged onotus and Mormoops, where it has a max- (0); or moderately enlarged (1). Numerous 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 23 authors have discussed the degree of enlarge- study. We found that all species of Pteron- ment of the cochlea in microchiropteran bats, otus and Mormoops have a greatly enlarged a feature that appears to be correlated with cochlea. Among the outgroups, a similar echolocation and foraging habits (Novacek, condition is seen in Saccopteryx bilineata, 1985a, 1987, 1991; Habersetzer and Storch, both species of Noctilio, Mystacina tuber- 1992; Simmons and Geisler, 1998). Haber- culata, and Artibeus jamaicensis. In contrast, setzer and Storch (1992) plotted cochlear Mystacina robusta and both species of Ma- width (measured from the end of the first half crotus have a moderately enlarged cochlea. turn of the cochlea to the end of the second Character 22: Ectotympanic bulla ex- half turn) versus basicranial width, and found tends medially across two-thirds or more of a small zone of overlap between nonechol- cochlea (0); or extends across one-half of co- ocating megachiropterans and echolocating chlea (1); or extends across no more than microchiropterans. Simmons and Geisler one-third of cochlea (2). The auditory bulla (1998) added additional data points to this of bats is poorly developed, consisting of a plot, and formalized these data in phyloge- single ectotympanic element that typically netic character which they used in assess- covers only part of the ventral exposure of ment of interfamilial relationships. Based on the cochlea. The ectotympanic bulla extends their data plot (Simmons and Geisler, 1998: medially across two-thirds or more of the co- fig. 29), they recognized three character chlea in both species of Mormoops and all states: (A) cochlea not enlarged (taxa that fall extant species of Pteronotus, with the excep- below the zone of overlap between Mega- tion of P. parnellii. In the latter species, the chiroptera and Microchiroptera); (B) cochlea ectotympanic bulla extends across no more moderately enlarged (taxa that fall in the than one-third of the cochlea. Among the zone of overlap between Megachiroptera and outgroups, the ectotympanic bulla extends Microchiroptera); and (C) cochlea greatly en- medially across two-thirds or more of the co- larged (taxa that fall above the zone of over- chlea in Artibeus jamaicensis, both species of lap between Megachiroptera and Microchi- Macrotus, and both species of Mystacina. In roptera). Extant bats in the ‘‘cochlea not en- contrast, the ectotympanic bulla extends larged’’ group do not use sophisticated echo- across one-half of the cochlea in Saccopteryx location. Some bats with a ‘‘moderately bilineata and both species of Noctilio. Trans- enlarged’’ cochlea do not use sophisticated formations in this character were ordered 0 echolocation (i.e., the megachiropterans), but ↔ 1 ↔ 2toreflect our hypothesis that others do (i.e., the microchiropterans). How- changes in relative coverage of the cochlea ever, none of the latter group are typical ae- by the ectotympanic bulla occur in a pro- rial insectivores (Simmons and Geisler, gressive fashion. Data were not available for 1998). Although some may combine aerial Pteronotus pristinus, which was scored ‘‘?’’ hawking with other foraging strategies, most for this character. apparently rely on passive acoustic cues and/ Character 23: Tympanic annulus in- or vision to detect invertebrate or small ver- clined, lies at angle of 35Ð55Њ relative to the tebrate prey that are captured by gleaning or basicranial plane (0); or semivertical, lies at landing on the prey, or they feed on fruit or angle of 75Ð90Њ relative to basicranial plane nectar that is detected by vision or olfaction (1). The tympanic annulus is that part of the (Simmons and Geisler [1998] and references ectotympanic that forms the ring that sup- cited therein). In contrast, all bats with the ports the tympanic membrane. In bats the an- ‘‘cochlea greatly enlarged’’ use sophisticated nulus is typically composed of relatively laryngeal echolocation, and most are expert dense bone (as opposed to the thinner medial aerial hawkers that obtain most or all of their portion of the ectotympanic, which forms the food in this manner (Simmons and Geisler bulla), and it lies in a plane that defines the [1998] and references cited therein). orientation of the tympanic membrane. The In the present study, we scored taxa using tympanic annulus is semivertical (plane of Simmons and Geisler’s (1998: fig. 29) meth- annulus lies at angle of 35Ð55Њ relative to the ods and data, adding new observations only basicranial plane) in all extant species of for the species not explicitly scored in that Pteronotus and Mormoops. A similar condi- 24 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 tion is found among the outgroups in Arti- not available for Pteronotus pristinus, which beus jamaicensis and both species of Macro- was scored ‘‘?’’ for this character. tus. In contrast, the tympanic annulus is in- Character 26: Angular process projects clined at an angle of 35Ð55Њ relative to the above the level of the occlusal plane (0); or basicranial plane in Saccopteryx bilineata, at or below the level of the occlusal plane both species of Noctilio, and both species of (1). The angular process in bats extends from Mystacina. Data were not available for Pter- the posteroventral ‘‘corner’’ of the lower jaw. onotus pristinus, which was scored ‘‘?’’ for This process projects above the level of the this character. occlusal plane of the molar toothrow in all Character 24: Basisphenoid with wide, species of Pteronotus and Mormoops. A sim- shallow longitudinal furrows (0); or with ilar condition is seen in Saccopteryx bilinea- narrow, deep furrows (1); or with deep pits ta and both species of Noctilio. In contrast, (2). In most noctilionoids the basisphenoid is the angular process lies either at or entirely trapezoid-shaped, wider anteriorly than pos- below the level of the occlusal plane in both teriorly, and has two parallel furrows along species of Mystacina, Artibeus jamaicensis, its entire length. In part owing to the shape and both species of Macrotus. Data were not of the basisphenoid, the furrows are less well available for Pteronotus pristinus, which was defined anteriorly than posteriorly. Pterono- scored ‘‘?’’ for this character. tus macleayii, P. quadridens, P. personatus, Character 27: Coronoid process of lower P. davyi, P. gymnonotus, and both species of jaw does not extend dorsally above level of Mormoops have basisphenoid furrows that condyloid process (0); or coronoid process are wide and shallow, covering the entire one and one-half times the height of the con- width of the basisphenoid. In contrast, Pter- dyloid process (1); or two times the height onotus parnellii and P. pristinus have nar- of the condyloid process (2). The coronoid row, deep furrows down the center of the ba- process of the lower jaw does not extend dor- sally above the level of the condyloid process sisphenoid. Both species of Macrotus have in any species of Pteronotus and Mormoops. similar narrow and deep furrows, while Ar- A similar condition is seen in Saccopteryx tibeus jamaicensis, both species of Noctilio, bilineata and both species of Noctilio. In and both species of Mystacina have furrows contrast, the coronoid process extends above that are wide and shallow. Saccopteryx bili- the level of the condyloid process in both neata exhibits an alternative condition in species of Mystacina, Artibeus jamaicensis, which the basisphenoid is dominated by a and both species of Macrotus. The relative pair of large, deep pits. height of these processes (as measured from Character 25: Basioccipital not constrict- the lower border of the body of the mandi- ed (0); or constricted (1). In bats the basi- ble) varies considerably. The coronoid pro- occipital forms the posterior floor of the cess is one and one-half times the height of braincase, typically extending anteriorly be- the condyloid process in both species of Mys- tween the cochlea. In Pteronotus parnellii, tacina and Artibeus jamaicensis, while it is the basioccipital is markedly constricted an- two times the height of the condyloid process teriorly so that the width of the anterior por- in both species of Macrotus. Data were not tion of this element is approximately one- available for Pteronotus pristinus, which was third the width of the foramen magnum. In scored ‘‘?’’ for this character. contrast, the basioccipital is not so constrict- ed in Mormoops and the remaining species DENTITION of Pteronotus. In these forms, the width of the anterior basioccipital is approximately Character 28: Two pairs upper incisors one-half the width of the foramen magnum. (I1 and I2) present (0); or only one pair of The basioccipital is similarly unconstricted in upper incisors (I1) present, I2 absent (1). both species of Noctilio, both species of Mys- Most bats have two pairs of upper incisor tacina, both species of Macrotus, and Arti- teeth. The inner incisors (which lie adjacent beus jamaicensis. The basioccipital is con- to each other and the midsagittal plane) are stricted in Saccopteryx bilineata. Data were typically interpreted as homologous to I1 in 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 25

other mammals; the outer incisors are lacks a lingual cingulum on the upper inci- thought to represent I2 (Thomas, 1908; An- sors, but the remaining outgroups have a lin- dersen, 1912; Slaughter, 1970). All members gual cingulum on the base of each upper in- of Pteronotus and Mormoops have two pairs cisor. Data were not available for Pteronotus of upper incisors. Among the outgroups, both pristinus, which was scored ‘‘?’’ for this species of Noctilio, both species of Macro- character. tus, and Artibeus jamaicensis also have two Character 32: Diastema absent between pairs of upper incisors. In contrast, both spe- outer upper incisor (I2) and canine (0); or cies of Mystacina have only one pair of low- present (1). There is no diastema between the er incisors, which correspond in shape and outer upper incisor (I2) and canine tooth (C) position to I1 of other bats; I2 appears to be in Pteronotus parnellii. In contrast, a distinct absent in this genus. Saccopteryx bilineata diastema is present between these teeth in also has only a single pair of upper incisors, Pteronotus macleayii, P. quadridens, P. per- which we interpret as homologous to I1 sonatus, P. davyi, P. gymnonotus, P. pristin- based on comparisons with other emballon- us, and both species of Mormoops. Among urids that retain two pairs of upper incisors the outgroups, Artibeus jamaicensis and both (e.g., Emballonura). species of Macrotus and lack a diastema be- Character 29: Inner upper incisors (I1) tween I2 and C, while such a diastema is taper to blunt point (0); or bilobed (1). All present in both species of Noctilio. Saccop- extant species of Pteronotus and Mormoops teryx bilineata and both species of Mystacina have bilobed inner upper incisors (I1). have a diastema present anterior to the ca- Among the outgroups, both species of Ma- nine, but this taxon lacks I2 so the condition crotus also have similar bilobed inner upper is not comparable to that seen in other taxa. incisors. In contrast, Saccopteryx bilineata, Accordingly, we scored these species ‘‘–’’ both species of Noctilio, both species of Mys- for this character. tacina, and Artibeus jamaicensis exhibit a Character 33: Three pairs of lower inci- different condition in which each inner upper sors (i1, i2, i3) present (0); or two pairs (i1 incisor tapers to a blunt point. Data were not and i2) present (0); or only one pair (i1) pre- available for Pteronotus pristinus, which was sent (1). The maximum number of lower in- scored ‘‘?’’ for this character. cisors found in any bat is three pairs. All Character 30: Height of outer upper in- members of Pteronotus and Mormoops have cisors (I2) less than one-half that of inner two pairs of lower incisors (fig. 6), which we incisors (I1) (0); or I2 and I1 subequal in presume to be homologous to i1 and i2 fol- height (1). The height of the outer upper in- lowing Miller (1907). Artibeus jamaicensis cisors (I2) is less than one-half that of the and both species of Macrotus also have two inner incisors (I1) in all species of Pterono- pairs of lower incisors, again presumed to tus and Mormoops (fig. 4). This condition is represent i1 and i2. In contrast, Saccopteryx also seen in both species of Noctilio and both bilineata retains three pairs of lower incisors. species of Macrotus. In contrast, I2 and I1 Both species of Noctilio and both species of are subequal in height in Artibeus jamaicen- Mystacina have only one pair of lower inci- sis. Taxa lacking I2 (Saccopteryx bilineata sors that seem to correspond in shape and and both species of Mystacina; see character position to i1 of other bats; i2 and i3 appear 28) could not be evaluated for this character, to be absent in these forms (Miller, 1907). and so were scored ‘‘–’’ in our data set. Data Transformations in this character were or- were not available for Pteronotus pristinus, dered 0 ↔ 1 ↔ 2toreflect our hypothesis which was scored ‘‘?’’ for this character. that changes in the number of incisor teeth Character 31: Lingual cingulae present occurred in a progressive fashion. on upper incisors (I1 and I2) (0); or absent Although the number of upper and lower (1). Each upper incisor has a distinct lingual incisors is correlated in many species (e.g., cingulum at its base in all extant species of mormoopids), this is not true in all taxa. For Pteronotus. In contrast, lingual cingulae are example, Saccopteryx bilineata has three absent from these teeth in both species of pairs of lower incisors but only one pair of Mormoops. Saccopteryx bilineata similarly upper incisors; both species of Noctilio have 26 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Fig. 6. Occlusal views of the left lower toothrow of (A) Mormoops blainvillii (AMNH 45197), (B) Pteronotus quadridens (AMNH 39357), and (C) Pteronotus davyi (AMNH 204965). Note the relative proportions of the premolars (p3 reduced in Pteronotus species) and the morphology of the talonid basin (nyctalodont in Mormoops blainvillii and Pteronotus quadridens, myotodont in Pteronotus davyi). two pairs of upper incisors but only one pair tris similarly have a trilobed i1. In contrast, of lower incisors. The number of upper ver- i1 is bilobed in Noctilio leporinus, both spe- sus lower incisors therefore appears to be de- cies of Macrotus, and Artibeus jamaicensis. coupled among the taxa in our study. Ac- Data were not available for Pteronotus pris- cordingly, we have treated these as separate tinus, which was scored ‘‘?’’ for this char- characters. acter. Character 34: Inner lower incisors (i1) Character 35: Second lower incisors (i2) trilobed (0); or bilobed (1). All extant spe- trilobed (0); or bilobed (1). The second low- cies of Pteronotus and Mormoops have tri- er incisors (i2) are trilobed in Pteronotus ma- lobed inner lower incisors (i1; fig. 6). Among cleayii, P. quadridens, P. personatus, P. dav- the outgroups, Saccopteryx bilineata, both yi, P. gymnonotus, and both species of Mor- species of Mystacina, and Noctilio albiven- moops (fig. 6). In contrast, Pteronotus par- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 27

nellii has bilobed outer lower incisors. diastema between P3 and P4 in any species Among the outgroups, Saccopteryx bilineata of Pteronotus (fig. 4). In contrast, a distinct has trilobed outer lower incisors, while these diastema is present between these teeth in teeth are bilobed in and both species of Ma- both species of Mormoops (fig. 4). Among crotus and Artibeus jamaicensis. Both spe- the outgroups, a diastema between P3 and P4 cies of Noctilio and both species of Mysta- is lacking in Artibeus jamaicensis and both cina cannot be scored for this character be- species Mystacina, while such a diastema is cause they lack i2 (see character 34 above); present in both species Macrotus. P3 is ab- these taxa were scored ‘‘–’’ for this character sent in Saccopteryx bilineata and both spe- in our analysis. Data were not available for cies of Noctilio, so these taxa were scored Pteronotus pristinus, which was scored ‘‘?’’ ‘‘–’’ for this character. for this character. Character 39: Crown height of P3 sube- Character 36: Upper first premolar (P1) qual to that of P4 (0); or height of P3 less present (0); or absent (1). Much has been than one-half height of P4 (1). The relative written regarding the homologies of the pre- crown height of P3 and P4 varies among molar cheekteeth in bats (e.g., Miller, 1907; bats. P4 is typically a large tooth with a Thomas, 1908; Andersen, 1912; Slaughter, crown height equal to or greater than that of 1970; Handley, 1959; Simmons and Han- the molars; P3 may be of similar size, or re- dley, 1998). No bat has more than three up- duced relative to the other cheek teeth. The per premolars, and opinions have varied with crown height of P3 is reduced to less than respect to the identity of the missing tooth. one-half the height of P4 in all extant species We follow Thomas (1908), Andersen (1912), of Pteronotus and Mormoops (fig. 4). In con- Handley (1959), and Simmons and Handley trast, crown height of P3 is subequal to that (1998) in recognizing the premolars in bats of P4 in both species of Mystacina, both spe- as P1, P3, and P4; the missing tooth is pre- cies of Macrotus, and Artibeus jamaicensis. sumed to be P2. Changes in the composition Both species of Noctilio and Saccopteryx bil- of the premolar dentition were treated as bi- ineata cannot be scored for this character be- nary characters (rather than as a single or- cause they lack P3 (see character 38 above); dered multistate character, as we did with the these taxa were scored ‘‘–’’ for this character incisor dentition) because the sequence of in our analysis. Data were not available for loss or gain of teeth at the P1/p1 and P3/p3 Pteronotus pristinus, which was scored ‘‘?’’ loci are not clear. for this character. Two upper premolars are present in all Character 40: First lower premolar (p1) species of Pteronotus and Mormoops (fig. 4). present (0); or absent (1). No bat has more These apparently correspond to P3 and P4; than three lower premolars. As with the up- P1 is absent in these taxa (Miller, 1907). P1 per dentition, there has been considerable de- is similarly absent in both species of Nocti- bate regarding the identity of these teeth. lio, both species of Mystacina, Artibeus ja- Many authors have identified the three bat maicensis, and both species of Macrotus premolars as p2, p3, and p4 (e.g., Miller, (Miller, 1907). In contrast, P1 is present in 1907; Slaughter, 1970), but the anteriormost Saccopteryx bilineata (Simmons and Han- tooth is more likely homologous with p1 of dley, 1998). other mammals (Thomas, 1908; Handley, Character 37: Upper third premolar (P3) 1959). Three lower premolars (which we pre- present (0); or absent (1). The upper third sume to be p1, p3, and p4) are present in all premolar (P3) is present in all species of species of Pteronotus and Mormoops (fig. 6). Pteronotus and Mormoops (fig. 4). This tooth Saccopteryx bilineata and both species of is also present in both species of Mystacina, Macrotus also retain p1. In contrast, p1 is both species of Macrotus, and Artibeus ja- absent in both species of Noctilio, both spe- maicensis. In contrast, P3 is absent in Sac- cies of Mystacina, and Artibeus jamaicensis. copteryx bilineata and both species of Noc- Character 41: Third lower premolar (p3) tilio. present (0); or absent (1). The third lower Character 38: Diastema absent between premolar is present in all species of Pteron- P3 and P4 (0); or present (1). There is no otus and Mormoops (fig. 6). Similarly, p3 is 28 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 present in both species of Noctilio, both spe- m3) are present in all species of Pteronotus cies of Mystacina, both species of Macrotus, and Mormoops (figs. 4, 6). These teeth are and Artibeus jamaicensis. In contrast, p3 is also present in most of the outgroups. In con- absent in Saccopteryx bilineata (Simmons trast, M3 and m3 are both absent in Artibeus and Handley, 1998). jamaicensis. Unlike the situation with the in- Character 42: Lower p3 with one root cisors and premolars, where evolution in the (0); or with two roots (1). The lower third upper and lower dentitions appears decou- premolar p3 is a single-rooted tooth in all pled, changes in the upper and lower molar species of Pteronotus. In contrast, this tooth dentition are correlated among our taxa. Ac- has two roots in both species of Mormoops. cordingly, we treated absence of M3/m3 as Among the outgroups, p3 is single-rooted in a single character. both species of Noctilio, both species of Mys- Character 46: Lower first and second mo- tacina robusta, and Artibeus jamaicensis, but lars (m1 and m2) myotodont (0); or nyctal- this tooth has two roots in both species of odont (1); or modified for fruit feeding, cusps Macrotus. This character was scored ‘‘–’’ in and cristids not distinct (2). As discussed by Saccopteryx bilineata, which lacks p3 (see Simmons and Geisler (1998), morphology of character 41). the talonid of the lower first and second mo- Character 43: Crown length of lower p3 lars (m1 and m2) varies among bats. Most and p4 subequal (0); or p3 reduced, crown extant species have lower molar teeth in length of p3 less than or equal to one-half which the hypoconulid lies adjacent to the the crown length of p4 (1). Crown length of entoconid on the lingual edge of the tooth. the lower p3 and p4 are approximately sub- Menu and Sige« (1971) distinguished two equal in both species of Mormoops. In con- forms of the talonid that occur among these trast, p3 is reduced to a peglike tooth whose taxa: the postcristid (which originates from crown length is less than or equal to one-half hypoconid) may either connect with the hy- the crown length of p4 in all species of Pter- poconulid (a condition described as ‘‘nyctal- onotus (fig. 6). Among the outgroups, p3 and odonty’’) or bypass the hypoconulid to con- p4 are subequal in both species of Mystacina, nect with the entoconid (‘‘myotodonty’’). Artibeus jamaicensis, and both species of Pteronotus parnellii, P. davyi, P. gymnono- Macrotus, while p3 is reduced (crown length tus, and P. pristinus have myotodont lower approximately one-half the length of p4) in molars (fig. 6). In contrast, P. macleayii, P. both species Noctilio. This character was quadridens, P. personatus, and both species scored ‘‘–’’ in Saccopteryx bilineata, which of Mormoops have nyctalodont lower molars lacks p3 (see character 41). (fig. 6). Among the outgroups, Saccopteryx Character 44: No cuspule present on pre- bilineata is nyctalodont, but both species of metacrista on M1 and M2 (0); or cuspule Mystacina, both species of Noctilio, and both present (1). The premetacrista is a crestlike species of Macrotus have myotodont lower ridge that connects the mesostyle and meta- molars. Artibeus jamaicensis has molar teeth cone on the upper molars A small cuspule is that have been modified for frugivory in such present on the labial premetacrista in Pter- a way that the cusps and cristids are no lon- onotus personatus. In contrast, no cuspule is ger distinct. present in this position in Mormoops or any Character 47: Cristid obliqua on lower other species of Pteronotus. A cuspule is first and second molars (m1 and m2) extends similarly lacking on the premetacrista in Sac- from hypoconid to middle of posterior wall copteryx bilineata, both species of Noctilio, of protocristid (0); or extends from hypocon- both species of Mystacina, and both species id to metaconid (1). As noted by Czaplewski of Macrotus. Artibeus cannot be scored for (1997), the cristid obliqua on the lower first this character due to modification of the mo- and second molars (m1 and m2) extends lar teeth for frugivory, so this taxon is scored from hypoconid to middle of posterior wall ‘‘–’’ for this character. of protocristid in most bats. This condition Character 45: Upper and lower third mo- occurs in all species of Pteronotus and Mor- lars (M3 and m3) present (0); or absent (1). moops (fig. 6) and also characterizes Saccop- The upper and lower third molars (M3 and teryx bilineata, both species of Mystacina, 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 29

and both species of Macrotus among the out- vomeronasal cartilage that is curved in cross groups. In contrast, the cristid obliqua ex- section. In contrast, Pteronotus parnellii and tends from hypoconid to the metaconid on P. personatus have bar-shaped vomeronasal all of the lower molars in both species of cartilage. Noctilio leporinus also has bar- Noctilio. shaped cartilage, while Macrotus californi- As noted under character 46, Artibeus ja- cus and Artibeus jamaicensis have a curved maicensis has molar teeth that have been vomeronasal cartilage. Mystacina tubercula- modified for frugivory in such a way that the ta lacks the vomeronasal cartilage entirely cusps and cristids are no longer distinct. To (Wible and Bhatnagar, 1996). Data were not avoid scoring this condition twice in our available for Pteronotus macleayi, P. quad- analysis, Artibeus jamaicensis was scored ‘‘Ð’’ ridens, P. davyi, P. gymnonotus, P. pristinus, for the present character. Mormoops blainvillii, Saccopteryx bilineata, Noctilio albiventris, Mystacina robusta, and VOMERONASAL COMPLEX AND BRAIN Macrotus waterhousii, which were scored ‘‘?’’ for this character. Character 48: Vomeronasal epithelial Character 50: Nasopalatine duct absent tube absent (0); or rudimentary, lacking a (0); or present (1). The nasopalatine duct neuroepithelial medial lining (1); or well-de- connects the nasal cavity to the oral cavity veloped, neuroepithelial medial lining pre- through the incisive foramen (Wible and sent (2). Wible and Bhatnagar (1996) inves- Bhatnagar, 1996). This duct is absent in Pter- tigated morphology of the vomeronasal sys- onotus parnellii, but is present in P. person- tem in 51 species of bats including seven of atus and Mormoops megalophylla. Among the taxa in our study. They found that the the outgroups, the nasopalatine duct is pre- vomeronasal epithelial tube is well devel- sent in Artibeus jamaicensis but is absent in oped, with a neuroepithelial medial lining, in Macrotus californicus, Noctilio leporinus, Pteronotus parnellii. The vomeronasal tube and Mystacina tuberculata. Data were not is rudimentary (lacking a neuroepithelial lin- available for Pteronotus macleayi, P. quad- ing) in Mormoops megalophylla, and the ridens, P. davyi, P. gymnonotus, P. pristinus, tube is entirely absent in Pteronotus person- Mormoops blainvillii, Saccopteryx bilineata, atus. Among the outgroups, Noctilio lepori- Noctilio albiventris, Mystacina robusta, and nus and Mystacina tuberculata lack a vom- Macrotus waterhousii, which were scored eronasal epithelial tube, while this structure ‘‘?’’ for this character. is well developed in Macrotus californicus Character 51: Accessory olfactory bulb and Artibeus jamaicensis (Wible and Bhat- absent (0); or present (1). The accessory ol- nagar, 1996). This character was ordered 0 factory bulb (AOB) is an anatomically dis- ↔ 1 ↔ 2 in our analyses to reflect the logical tinct portion of the forebrain that receives sequence through which a well-developed sensory information from the vomeronasal vomeronasal epithelial tube (‘‘2’’) is thought nerve (Wible and Bhatnagar, 1996). Al- to be progressively lost (‘‘0’’), or vice versa. though presence of an AOB is correlated Data were not available for Pteronotus ma- with degree of development of the vomero- cleayi, P. quadridens, P. davyi, P. gymnon- nasal epithelial tube (see character 14), we otus, P. pristinus, Mormoops blainvillii, Sac- followed Wible and Bhatnagar (1996) in copteryx bilineata, Noctilio albiventris, Mys- treating these as separate characters pending tacina robusta, and Macrotus waterhousii, further research on the connections of these which were scored ‘‘?’’ for this character. structures. An AOB is present in Pteronotus Character 49: Vomeronasal cartilage parnellii, but is absent in P. personatus and curved in cross section (‘‘J’’, ‘‘C’’, ‘‘U’’,or Mormoops megalophylla (Wible and Bhat- ‘‘O’’ shaped) (0); or bar-shaped in cross nagar, 1996). Among the outgroups, an AOB section (1); or cartilage absent (2). The is present in Artibeus jamaicensis but is ab- vomeronasal cartilage lies ventrolateral to the sent in Noctilio leporinus. Data were not anterior nasal septum and supports the vom- available for Pteronotus macleayi, P. quad- eronasal epithelial tube (Wible and Bhatna- ridens, P. davyi, P. gymnonotus, P. pristinus, gar, 1996). Mormoops megalophylla has a Mormoops blainvillii, Saccopteryx bilineata, 30 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Noctilio albiventris, Mystacina robusta, both extensively by Sprague (1943), Griffiths species of Macrotus, and both species of (1982, 1983, 1994), Griffiths and Smith Mystacina, which were scored ‘‘?’’ for this (1991) and Griffiths et al. (1992). We follow character. Griffiths’ terminology for osteological and Character 52: Inferior colliculi fully ex- myological features, which are identical to posed on dorsal surface of brain, not covered those used by Sprague (1943) with two ex- by cerebellar vermis or cerebral hemispheres ceptions: (1) the element termed the hypoh- (0); or inferior colliculi partially exposed, yal by Sprague (1943) is identified as the cer- medial longitudinal fissure and medial por- atohyal by Griffiths (1982, 1983, 1994), and tions of colliculi covered by rostral extension (2) the element termed the ceratohyal by of cerebellar vermis (1); or inferior colliculi Sprague (1943) is identified as the epihyal by not exposed, completely covered by cerebel- Griffiths (1982, 1983, 1994). lar vermis and cerebral hemispheres (2). As Character 53: Tracheal rings subequal in discussed and illustrated by Wetterer et al. diameter throughout length of trachea (0); or (2000), there is significant variation among first 5Ð8 rings enlarged to form tracheal ex- noctilionoids in the degree of exposure of the pansion posterior to larynx (1). In most inferior colliculi on the dorsal surface of the mammals, including most bats, the cartilag- brain. The inferior colliculi are partially ex- inous rings that support the trachea are of posed on the dorsal surface of the brain in more-or-less uniform diameter throughout Pteronotus parnellii, P. gymnonotus, and the length of the postlaryngeal trachea (Grif- Mormoops megalophylla (Schneider, 1957, fiths, 1982, 1983, 1994; Griffiths and Smith, 1972; Henson, 1970; Stephan, 1977). In 1991; Griffiths et al., 1992; Simmons, 1998; these forms, a rostral extension of the cere- Simmons and Geisler, 1998). However, some bellar vermis covers the medial portion of bats have ‘‘tracheal expansions’’ posterior to each colliculus and the longitudinal fissure the larynx (Griffiths, 1983, 1994; Griffiths between them, but the lateral portion of each and Smith, 1991; Griffiths et al., 1992; Sim- inferior colliculus is clearly visible along the mons, 1998; Simmons and Geisler, 1998). edges of the cerebellar vermis. Among the These expansions consist of balloonlike outgroups, a similar condition is seen in Noc- spaces supported by tracheal rings that have tilio leporinus (Baron et al., 1996). In con- trast, the inferior colliculi are entirely visible a diameter greater than that of the ‘‘normal’’ (no portion is covered by the cerebellum) in tracheal rings composing the remainder of Macrotus californicus (McDaniel, 1976). A the trachea (Griffiths, 1983, 1994; Griffiths third condition is seen in Artibeus jamaicen- and Smith, 1991; Griffiths et al., 1992). All sis, in which the inferior colliculi are com- extant species of Pteronotus and Mormoops pletely covered by the cerebellar vermis and have a tracheal expansion that involves the cerebral hemispheres (McDaniel, 1976; first 5Ð8 tracheal rings (fig. 7). No tracheal Bhatnagar and Kallen, 1974). No portion of expansion is seen in any of the outgroups the inferior colliculi are visible on the brain (fig. 7). Data were not available for Mysta- surface in dorsal view. Transformations in cina tuberculata and Pteronotus pristinus, this character were ordered 0 ↔ 1 ↔ 2to which were scored ‘‘?’’ for this character. reflect the hypothesis that covering or uncov- Character 54: Basihyal with entoglossal ering of the inferior colliculi by the cerebel- process (0); or with entoglossal process (1). lum occurs in a progressive fashion. Data The basihyal of many bats includes a medial were not available for Pteronotus macleayi, entoglossal process that provides an attach- P. quadridens, P. personatus, P. davyi, P. ment site for several hyoid muscles (Griffiths pristinus, Mormoops blainvillii, Saccopteryx 1982, 1983, 1994; Griffiths and Smith, 1991; bilineata, Noctilio albiventris, Mystacina ro- Griffiths et al. 1992). An entoglossal process busta, and Macrotus waterhousii, which is present in Pteronotus personatus and P. were scored ‘‘?’’ for this character. parnellii (Griffiths, 1982, 1983). Similarly, an entoglossal process is present on the ba- TRACHEA AND HYOID APPARATUS sihyal in Noctilio leporinus, Mystacina ro- The bones, cartilages, and muscles of the busta, Macrotus waterhousii, and Artibeus hyoid apparatus of bats have been discussed jamaicensis (Sprague, 1943; Griffiths, 1982, 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 31

Fig. 7. Ventral views of the larynx and anterior trachea of (A) Macrotus waterhousii and (B) Pter- onotus davyi (redrawn from Griffiths, 1978: figs. 2, 4). Note the expanded tracheal rings and division of m. cricothyroideus into two muscles in P. davyi. Scale bars ϭ 1 mm.

personal commun.). In contrast, the basihyal robusta, and Macrotus waterhousii have a element lacks an entoglossal process in Sac- ceratohyal that is roughly half the length of copteryx bilineata (Griffiths and Smith, the epihyal (Griffiths, 1982, personal com- 1991). Data were not available for Pterono- mun.). The ceratohyal is reduced to a tiny tus macleayii, P. quadridens, P. gymnonotus, cartilaginous element in Noctilio leporinus P. davyi, P. pristinus, Mormoops megalo- (Sprague, 1943). Transformations in this phylla, M. blainvillii, Noctilio albiventris, character were ordered 0 ↔ 1 ↔ 2toreflect Mystacina tuberculata, and Macrotus cali- our hypothesis that reduction or lengthening fornicus, which were scored ‘‘?’’ for this of the ceratohyal relative to the epihyal oc- character. curs in a progressive fashion. Data were not Character 55: Ceratohyal unreduced, ap- available for Pteronotus macleayii, P. quad- proximately equal in length to epihyal (0); ridens, P. gymnonotus, P. davyi, P. pristinus, or ceratohyal reduced to half the length of Mormoops megalophylla, M. blainvillii, Noc- epihyal (1); or ceratohyal reduced to tiny tilio albiventris, Mystacina tuberculata, Ma- cartilaginous element (2). In Pteronotus per- crotus californicus, and Artibeus jamaicen- sonatus, the ceratohyal and epihyal (sensu sis, which were scored ‘‘?’’ for this character. Griffiths) are rodlike elements that are ap- Character 56: M. mylohyoideus undivid- proximately the same length (Griffiths, 1982, ed sheet of muscle with no aponeurosis (0); 1983). In contrast, the ceratohyal is reduced or undivided but with anterior aponeurosis in length to a short, stubby element that is (1); or with a pronounced break, clearly di- roughly half the length of epihyal in Pter- vided into distinct anterior and posterior onotus parnellii (Griffiths, 1983). Among the muscular parts by a fleshy aponeurosis (2). outgroups, Saccopteryx bilineata, Mystacina M. mylohyoideus in bats originates from the 32 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 medial surface of the mandible and inserts with a tendon is present in Mystacina robusta into the elements of the hyoid and/or other (Griffiths, personal commun.). Data were not muscles of the hyoid region (Griffiths, 1982, available for Pteronotus macleayii, P. quad- 1983, 1994; Griffiths and Smith, 1991; Grif- ridens, P. personatus, P. gymnonotus, P. fiths et al., 1992). M. mylohyoideus consists davyi, P. pristinus, Mormoops megalophylla, of a undivided sheet of muscle with an apo- M. blainvillii, Noctilio albiventris, Mystacina neurotic anterior portion in Pteronotus par- tuberculata, and Macrotus californicus, nellii (Sprague, 1943). Among the outgroups, which were scored ‘‘?’’ for this character. Saccopteryx bilineata, Noctilio leporinus, Character 59: M. stylohyoideus present, and Mystacina robusta have an undivided m. passes superficial to digastric muscles (0); or mylohyoideus that lacks the anterior aponeu- superficial slip absent (1). M. stylohyoideus rosis seen in Pteronotus (Sprague, 1943; in bats originates from the medial stylohyal Griffiths, personal commun.; Griffiths and and inserts into the anterior cornu of the hy- Smith, 1991). In contrast, m. mylohyoideus oid (Sprague, 1943; Griffiths, 1982, 1983, exhibits a pronounced break and is clearly 1994; Griffiths and Smith, 1991; Griffiths et divided into distinct anterior and posterior al., 1992). M. stylohyoideus is absent (or is parts by a fleshy aponeurosis in Macrotus reduced to a few fibers fused with m. stylo- waterhousii and Artibeus jamaicensis (Grif- pharyngeus, deep to the digastric muscles) in fiths, 1982). Data were not available for Pter- Pteronotus parnellii, Artibeus jamaicensis, onotus macleayii, P. quadridens, P. person- Macrotus waterhousii, Mystacina robusta, atus, P. gymnonotus, P. davyi, P. pristinus, and Noctilio leporinus (Sprague, 1943; Grif- Mormoops megalophylla, M. blainvillii, Noc- fiths, 1982, 1983, personal commun.). In tilio albiventris, Mystacina tuberculata, and contrast, m. stylohyoideus is present as a dis- Macrotus californicus, which were scored tinct muscle that runs superficial to the di- ‘‘?’’ for this character. gastric muscles in Saccopteryx bilineata Character 57: M. mylohyoideus inserts on (Griffiths and Smith, 1991). Data were not basihyal, basihyal raphe, and thyrohyal (0); available for Pteronotus macleayii, P. quad- or on basihyal and basihyal raphe only (1). ridens, P. personatus, P. gymnonotus, P. M. mylohyoideus runs ventral to the midline davyi, P. pristinus, Mormoops megalophylla, strap muscles and inserts on the basihyal and M. blainvillii, Noctilio albiventris, Mystacina basihyal raphe in Pteronotus parnellii, Arti- tuberculata, and Macrotus californicus, beus jamaicensis, Macrotus waterhousii, which were scored ‘‘?’’ for this character. Mystacina robusta, and Noctilio leporinus Character 60: M. geniohyoideus origi- (Sprague, 1943; Griffiths, 1982, personal nates by fleshy fibers directly from the bone commun.). In contrast, m. mylohyoideus in- of the mandible (0); or originates by long serts on the basihyal raphe, basihyal, and also tendon from the mandible (1). M. geniohy- onto the thyrohyal in Saccopteryx bilineata oideus originates from the inner surface of (Griffiths and Smith, 1991). Data were not the anterior mandible and inserts into the available for Pteronotus macleayii, P. quad- tongue (Griffiths, 1982, 1983, 1994; Griffiths ridens, P. personatus, P. gymnonotus, P. and Smith, 1991; Griffiths et al., 1992). In davyi, P. pristinus, Mormoops megalophylla, Pteronotus parnellii, m. geniohyoideus orig- M. blainvillii, Noctilio albiventris, Mystacina inates by means of a long tendon from the tuberculata, and Macrotus californicus, mandible (Griffiths, personal commun.). A which were scored ‘‘?’’ for this character. similar condition is seen in Saccopteryx bil- Character 58: M. mandibulo-hyoideus ineata (Griffiths and Smith, 1991). In con- absent (0); or present (1). M. mandibulo-hy- trast, m. geniohyoideus originates by fleshy oideus, which represents the medial portion fibers directly from the bone of the mandible of the anterior digastric muscle, is absent in in Artibeus jamaicensis, Macrotus waterhou- Pteronotus parnellii, Artibeus jamaicensis, sii, Mystacina robusta, and Noctilio lepori- Macrotus waterhousii, Noctilio leporinus, nus (Sprague, 1943; Griffiths, 1982, personal and Saccopteryx bilineata (Sprague, 1943; commun.). Data were not available for Pter- Griffiths, 1982, Griffiths and Smith, 1991). onotus macleayii, P. quadridens, P. person- In contrast, a small m. mandibulo-hyoideus atus, P. gymnonotus, P. davyi, P. pristinus, 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 33

Mormoops megalophylla, M. blainvillii, Noc- tions in this character were ordered 0 ↔ 1 tilio albiventris, Mystacina tuberculata, and ↔ 2toreflect our hypothesis that changes in Macrotus californicus, which were scored the lateral extent of the origin of m. sterno- ‘‘?’’ for this character. hyoideus occur in a progressive fashion. Data Character 61: Paired m. geniohyoideus were not available for Pteronotus macleayii, muscles partially or completely fused across P. quadridens, P. personatus, P. gymnono- midline (0); or muscles not fused (1). M. ge- tus, P. pristinus, Mormoops blainvillii, Noc- niohyoideus is partially or completely fused tilio albiventris, Mystacina tuberculata, and with its counterpart across the midline in Macrotus californicus, which were scored Pteronotus parnellii (Sprague, 1943). ‘‘?’’ for this character. Among the outgroups, a similar condition is Character 63: Midline hyoid strap mus- seen in Artibeus jamaicensis (Sprague, culature with m. geniohyoideus and m. hyo- 1943). In contrast, the right and left m. ge- glossus directly attached to basihyal via niohyoideus are not fused in Noctilio lepor- fleshy fibers (0); or muscles attached indi- inus (Sprague, 1943). Data were not avail- rectly to basihyal via basihyal tendon (1). able for Pteronotus macleayii, P. quadridens, The midline hyoid strap musculature of bats P. personatus, P. gymnonotus, P. davyi, P. includes m. geniohyoideus and m. hyoglos- pristinus, Mormoops megalophylla, M. sus (Sprague, 1943; Griffiths, 1982, 1983, blainvillii, Saccopteryx bilineata Noctilio al- 1994; Griffiths and Smith, 1991; Griffiths et biventris, Mystacina tuberculata, M. robusta, al., 1992). In many bats, these muscles are Macrotus waterhousii, and M. californicus, attached indirectly to the basihyal via the ba- which were scored ‘‘?’’ for this character. sihyal tendon, resulting in what has been de- Character 62: M. sternohyoideus origin scribed as a ‘‘free-floating’’ strap muscle includes entire anterodorsal surface of ma- condition (Griffiths, 1982, 1983, 1994; Grif- nubrium and medial tip of clavicle (0); or fiths and Smith, 1991; Griffiths et al., 1992). origin includes entire anterodorsal surface of This condition is seen in Pteronotus parnel- manubrium but does not include clavicle (1); lii, Artibeus jamaicensis, Macrotus water- or origin restricted to medialmost surface housii, Mystacina robusta, and Noctilio le- manubrium in vicinity of anterior process porinus (Sprague, 1943; Griffiths, 1978, (2). M. sternohyoideus in bats extends be- 1982, personal commun.). In contrast, m. ge- tween the sternum and the hyoid region niohyoideus and m. hyoglossus attach di- (Sprague, 1943; Griffiths, 1978; 1982, 1983, rectly to the basihyal via fleshy fibers in Sac- 1994; Griffiths and Smith, 1991; Griffiths et copteryx bilineata (Griffiths and Smith, al., 1992). The origin of this muscle includes 1991). Data were not available for Pterono- the entire anterodorsal surface of the manu- tus macleayii, P. quadridens, P. personatus, brium and the medial tip of the clavicle in P. gymnonotus, P. davyi, P. pristinus, Mor- Pteronotus parnellii (Sprague, 1943). In con- moops megalophylla, M. blainvillii, Noctilio trast, the origin of m. sternohyoideus is re- albiventris, Mystacina tuberculata, and Ma- stricted to the anterodorsal surface of the ma- crotus californicus, which were scored ’’?’’ nubrium and the clavicular origin is lacking for this character. in P. davyi and M. megalophylla (Griffiths, Character 64: M. styloglossus with one 1978). Among the outgroups, the origin m. belly (0); or with two bellies separated by sternohyoideus includes the entire anterodor- lateral portion of m. hyoglossus (1). M. sty- sal surface of the manubrium and the medial loglossus in bats originates from the stylo- tip of the clavicle in Macrotus waterhousii, hyoid process and/or the adjacent surface of Artibeus jamaicensis, Noctilio leporinus, and the skull and inserts into the tongue Saccopteryx bilineata (Sprague, 1943; Grif- (Sprague, 1943; Griffiths, 1982, 1983, 1994; fiths, 1982; Griffiths and Smith, 1991). Mys- Griffiths and Smith, 1991; Griffiths et al., tacina robusta exhibits a different condition 1992). In Pteronotus parnellii this muscle in which the origin of this muscle is restrict- has two bellies that are separated by the lat- ed to the medialmost surface of the manu- eral portion of m. hyoglossus (Griffiths, brium in the vicinity of the anterior process 1983). Among the outgroups, m. styloglossus (Griffiths, personal commun.). Transforma- has two bellies in Mystacina robusta (Grif- 34 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

fiths, personal commun.), but only one belly moops megalophylla, M. blainvillii, Noctilio in Saccopteryx bilineata, Noctilio leporinus, albiventris, Mystacina tuberculata, and Ma- Macrotus waterhousii, and Artibeus jamai- crotus californicus, which were scored ‘‘?’’ censis (Sprague, 1943; Griffiths, 1982; Grif- for this character. fiths and Smith, 1991). Data were not avail- Character 67: M. sternothyroideus origi- able for Pteronotus macleayii, P. quadridens, nates from clavicle (0); or from manubrium P. personatus, P. gymnonotus, P. davyi, P. of sternum (1). M. sternothyroideus in bats pristinus, Mormoops megalophylla, M. originates from the sternal region and inserts blainvillii, Noctilio albiventris, Mystacina onto the thyroid cartilage (Sprague, 1943; tuberculata, and Macrotus californicus, Griffiths, 1982, 1983, 1994; Griffiths and which were scored ‘‘?’’ for this character. Smith, 1991; Griffiths et al., 1992). M. ster- Character 65: Origin of m. ceratohyoi- nothyroideus originates from the medial tip deus includes ceratohyal (0); or origin does of the clavicle (not from any part of the ster- not include ceratohyal (1). M. ceratohyoi- num) in Pteronotus parnellii (Sprague, 1943; deus in bats originates from the anterior cor- Griffiths, 1978). A similar condition is seen nu of the hyoid and inserts on the thyrohyal in Artibeus jamaicensis, Macrotus waterhou- (Sprague, 1943; Griffiths, 1982, 1983, 1994; sii, Noctilio leporinus, and Saccopteryx bili- Griffiths and Smith, 1991; Griffiths et al., neata (Sprague, 1943; Griffiths, 1982; Grif- 1992). The origin of m. ceratohyoideus in- fiths and Smith, 1991). In contrast, m. ster- cludes the epihyal element in all of the taxa nothyroideus originates from the lateral ma- in our study (Sprague, 1943; Griffiths, 1982, nubrium (not the clavicle) in Mystacina personal commun.; Griffiths and Smith, robusta (Griffiths, personal commun.). Data 1991). In Pteronotus parnellii, an adjacent were not available for Pteronotus macleayii, portion of this muscle originates from the P. quadridens, P. personatus, P. gymnono- ceratohyal (Griffiths, 1978). Among the out- tus, P. davyi, P. pristinus, Mormoops megal- groups, a similar origin occurs in Saccopte- ophylla, M. blainvillii, Noctilio albiventris, ryx bilineata, Noctilio leporinus, and Arti- Mystacina tuberculata, and Macrotus cali- beus jamaicensis (Sprague, 1943; Griffiths, fornicus, which were scored ‘‘?’’ for this 1982; Griffiths and Smith, 1991). In contrast, character. m. ceratohyoideus does not have a ceratohyal Character 68: M. omohyoideus originates origin in Macrotus waterhousii and Mysta- from clavicle (0); or originates from scapula cina robusta (Griffiths, 1982, personal com- mun.). Data were not available for Pterono- (1); or muscle absent (2) M. omohyoideus tus macleayii, P. quadridens, P. personatus, connects the shoulder girdle with the hyoid P. gymnonotus, P. davyi, P. pristinus, Mor- apparatus. In bats, the origin of this muscle moops megalophylla, M. blainvillii, Noctilio may be from either the scapula or the clavicle albiventris, Mystacina tuberculata, and Ma- (Sprague, 1943; Griffiths, 1982, 1983, 1994; crotus californicus, which were scored ‘‘?’’ Griffiths and Smith, 1991; Griffiths et al., for this character. 1992). M. omohyoideus is present and orig- Character 66: Origin of m. ceratohyoi- inates from the scapula in Pteronotus par- deus does not include stylohyal (0); or origin nellii (Sprague, 1943). Among the outgroups, does include stylohyal (1). M ceratohyoideus m. omohyoideus similarly originates from lacks a stylohyal origin in Pteronotus par- the scapula in Noctilio leporinus and Arti- nellii, Saccopteryx bilineata, Noctilio lepor- beus jamaicensis (Sprague, 1943). In con- inus, Mystacina robusta, and Artibeus ja- trast, this muscle originates from the clavicle maicensis (Sprague, 1943; Griffiths, 1978, in Saccopteryx bilineata (Griffiths and 1982, personal commun.; Griffiths and Smith, 1991). M. omohyoideus is absent in Smith, 1991). In contrast, the origin of m. Mystacina robusta (Griffiths, personal com- ceratohyoideus does include the stylohyal el- mun.). Data were not available for Pterono- ement in Macrotus waterhousii (Griffiths, tus macleayii, P. quadridens, P. personatus, 1982). Data were not available for Pterono- P. gymnonotus, P. davyi, P. pristinus, Mor- tus macleayii, P. quadridens, P. personatus, moops megalophylla, M. blainvillii, Noctilio P. gymnonotus, P. davyi, P. pristinus, Mor- albiventris, Mystacina tuberculata, Macrotus 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 35

californicus, and M. waterhousii, which were Pteronotus parnellii, P. davyi, and Mor- scored ‘‘?’’ for this character. moops megalophylla (fig. 7). In these bats, Character 69: M. sphincter colli profun- cricothyroideus anterior originates from the dus present (0); or absent (1). M. sphincter wide, calcified anterior portion of the carti- colli profundus in bats originates from the lage and inserts on the ventrolateral thyroid fascia of the mylohyoid region, forms a var- cartilage; m. cricothyroideus posterior origi- iable number of slips, and inserts on the inner nates from the soft, uncalcified lateral walls surface of the skin of the cervical region of the cricoid cartilage and inserts on the me- (Sprague, 1943; Griffiths, 1982). M. sphinc- dial surface of the thyroid cornu (Griffiths, ter colli profundus is absent in Pteronotus 1978). Both muscles are innervated my parnellii (Sprague, 1943). In contrast, this branches of n. laryngeus cranialis, the nerve muscle is present in Macrotus waterhousii, which innervates the single m. cricothyroi- Artibeus jamaicensis, and Noctilio lepori- deus found in other mammals (Griffiths, nus(Sprague, 1943; Griffiths, 1982). Data 1978). In contrast to the mormoopid condi- were not available for Pteronotus macleayii, tion, Saccopteryx bilineata, Macrotus water- P. quadridens, P. personatus, P. gymnono- housii, and Artibeus jamaicensis have a m. tus, P. davyi, P. pristinus, Mormoops megal- cricothyroideus that consists of a single, un- ophylla, M. blainvillii, Noctilio albiventris, divided sheet of muscle on each side of the Mystacina tuberculata, M. robusta, Saccop- larynx (fig. 7; Griffiths, 1978; Griffiths and teryx bilineata, and Macrotus californicus, Smith, 1991). Data were not available for which were scored ‘‘?’’ for this character. Pteronotus macleayii, P. quadridens, P. per- Character 70: M. cricopharyngeus con- sonatus, P. gymnonotus, P. pristinus, M. sists of a single large slip (0); or three or blainvillii, Noctilio albiventris, Mystacina more slips (1). M. cricopharyngeus in bats tuberculata, M. robusta, and Macrotus cali- originates from the cricoid cartilage and in- fornicus, which were scored ‘‘?’’ for this serts into the dorsal wall of the pharynx character. (Griffiths, 1978; Griffiths and Smith, 1991). M. cricopharyngeus consists of a single large slip in Pteronotus parnellii, P. davyi, and TONGUE Mormoops megalophylla (Griffiths, 1978). A similar condition is seen among the out- Character 72: Horny papillae of different groups in Noctilio leporinus and Saccopteryx sizes, largest near midline (0); or all horny bilineata, but m. cricopharyngeus consists of papillae of uniform size (1). Terminology for three or more slips in Macrotus waterhousii describing structure and distribution of and Artibeus jamaicensis (Sprague, 1943; tongue papillae is complex and often confus- Griffiths, 1978, 1982; Griffiths and Smith, ing. We follow Park and Hall (1951), Green- 1991). Data were not available for Pterono- baum and Phillips (1974), Griffiths (1982), tus macleayii, P. quadridens, P. personatus, Gimenez (1993), Gimenez et al. (1996), and P. gymnonotus, P. pristinus, Mormoops me- Wetterer et al. (2000) in recognizing groups galophylla, M. blainvillii, Noctilio albiven- of presumably homologous tongue papillae tris, Mystacina tuberculata, M. robusta, and based on morphology of individual papillae Macrotus californicus, which were scored and position of papillae on the tongue. ‘‘?’’ for this character. ‘‘Horny papillae’’ describes a cluster of en- Character 71: M. cricothyroideus consists larged, bifid papillae located at the midline of a single undivided muscle sheet (0); or near the tip of the tongue (Park and Hall, two muscles present, m. cricothyroideus an- 1951). The horny papillae are all of roughly terior and m. cricothyroideus posterior (1). uniform size in all extant species of Pteron- M. cricothyroideus in most mammals con- otus. In contrast, the horny papillae are of sists of single muscle sheet that originates different sizes in both species of Mormoops. from the cricoid cartilage of the larynx and In these species, the posteromedial horny pa- inserts on the thyroid cartilage (Griffiths, pillae are much larger than others in the clus- 1978). In contrast, two cricothyroid muscles ter. Among the outgroups, a similar condition are present on each side of the larynx in is seen in Mystacina tuberculata, Artibeus 36 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 jamaicensis, and both species of Macrotus. ward the midline of the tongue, is seen in In contrast, the horny papillae are of uniform Saccopteryx bilineata and both species of size in Saccopteryx bilineata and both spe- Noctilio. Data were not available for Pter- cies of Noctilio. Data were not available for onotus pristinus and Mystacina robusta, so Pteronotus pristinus and Mystacina robusta, these taxa were scored ‘‘?’’ in our analysis. so these taxa were scored ‘‘?’’ in our analy- Character 75: Basketlike medial-posteri- sis. or mechanical papillae present (0); or absent Character 73: Lateral circumvallate pa- (1). ‘‘Basketlike papillae’’ are medial-poste- pillae absent (0); or present (1). Paired cir- rior mechanical papillae that have central cumvallate papillae (one lateral and one me- concavities and fleshy, cylindrical bases dial) are located near the proximal base of (Park and Hall, 1951; Wetterer et al., 2000). the tongue in many bats (Park and Hall, The rim of each basketlike papilla is uneven 1951; Greenbaum and Phillips, 1974; Grif- and may be adorned with from 1 to 10 point- fiths, 1982; Gimenez, 1993; Gimenez et al, ed projections that form a fringe along the 1996; Wetterer et al., 2000). Lateral vallate basket edge (Park and Hall, 1951; Wetterer papillae are present in all extant species of et al., 2000). Basketlike papillae are present Pteronotus and Mormoops. Among the out- in all extant species of Pteronotus and Mor- groups, they are also present in Mystacina moops. Among the outgroups, they are also tuberculata, Artibeus jamaicensis, and both present in Saccopteryx bilineata, Mystacina species of Macrotus. In contrast, lateral val- tuberculata, and both species of Macrotus. late papillae are absent in Saccopteryx bili- In contrast, basketlike papillae are absent in neata and both species of Noctilio. Data were both species of Noctilio and Artibeus jamai- not available for Pteronotus pristinus and censis. Data were not available for Pterono- Mystacina robusta, so these taxa were scored tus pristinus and Mystacina robusta, so these ‘‘?’’ in our analysis. taxa were scored ‘‘?’’ in our analysis. Character 74: Distal tips of medial-pos- Character 76: Pharyngeal region of terior mechanical papillae directed toward tongue totally bare, papillae absent along midline of tongue (0); or directed toward lateral edges (0); or with medial bare patch, pharyngeal region (1); or directed toward papillae present along lateral edges (1). The anterior tip of tongue (2); or papillae basin- pharyngeal region of the tongue is defined as shaped with no clear inclination (3). The me- that part of the tongue that lies proximal to dial-posterior mechanical papillae are those papillae located anterior to the circumvallate the circumvallate papillae (Sontag, 1925). papillae in roughly the middle one-third of The pharyngeal region of the tongue is com- the tongue (Gimenez, 1993; Wetterer et al., pletely bare of papillae in all extant species 2000). Most medial-posterior mechanical pa- of Pteronotus and Mormoops. A similar con- pillae are thick, fleshy, and terminate in one dition is seen in Saccopteryx bilineata. In or more rounded points (Wetterer et al., contrast, only the medial portion of the pha- 2000). The distal tips of the medial posterior ryngeal region of the tongue is bare in both mechanical papillae are directed toward the species of Noctilio, Mystacina tuberculata, pharyngeal region in both species of Mor- both species of Macrotus, Artibeus jamaicen- moops. In contrast, these papillae are some- sis, and both species of Noctilio. In these what basin-shaped and show no clear incli- forms, mechanical papillae are present along nation in any direction in extant species of the lateral edges of the tongue in the pharyn- Pteronotus. Among the outgroups, the tips of geal region. Data were not available for Pter- the medial-posterior mechanical papillae are onotus pristinus and Mystacina robusta, so directed toward the pharyngeal region in these taxa were scored ‘‘?’’ in our analysis. Mystacina tuberculata and both species of Macrotus, and in the opposite direction (to- FACE,EARS, AND VIBRISSAE ward the tip of the tongue) in Artibeus ja- maicensis (Wetterer et al., 2000). A fourth Character 77: Internal labial papillae ab- condition, in which the medial posterior me- sent (0); or present (1). Internal labial papil- chanical papillae are directed medially to- lae are small, fleshy projections that occur in 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 37

patches inside the lips and cheeks of some than the height of the nostril. In contrast, all bats (Silva-Taboada and Pine, 1969; Wetterer extant species of Pteronotus have a large, et al., 2000). Internal labial papillae are ab- platelike lower lip that has a height equal to sent in all extant species of Pteronotus and or greater than height of the nostrils. The out- Mormoops. A similar condition is seen groups all have a small or moderate lower among in Saccopteryx bilineata, Mystacina lip whose dorsoventral height is less than the tuberculata, both species of Noctilio, and height of the nostrils. Data were not available both species of Macrotus. In contrast, inter- for Pteronotus pristinus, which was scored nal labial papillae are present along the lin- ‘‘?’’ for this character. gual edge of the lower lip in Artibeus ja- Character 81: Lower lip without small maicensis. Data were not available for Pter- horseshoe-shaped pad at midline (0); or pad onotus pristinus and Mystacina robusta, so present (1). The lower lip of both species of these taxa were scored ‘‘?’’ in our analysis. Mormoops is characterized by a horseshoe- Character 78: Cheek pouches absent (0); shaped dermal pad that is centered above the or present (1). Cheek pouches are epitheli- chin and extends perpendicular to the nostrils um-lined pockets in the cheeks that open in- when the mouth is closed. This horseshoe- side the mouth. These are absent in most shaped pad is tiny; that is, the height of the bats, including all extant species of Pteron- pad is approximately one-fourth the width of otus and Mormoops. Among the outgroups, the lower lip. A similar horseshoe-shaped lip cheek pouches are absent in Saccopteryx bil- pad is absent in extant members of the genus ineata, both species of Mystacina, Artibeus Pteronotus and all of the outgroups. Data jamaicensis, and both species of Macrotus, were not available for Pteronotus pristinus, but are present in both species of Noctilio. which was scored ‘‘?’’ for this character. The cheek pouches in Noctilio are formed by Character 82: No transverse flap below posterolateral expansion of m. buccinator, lower lip (0); or one transverse flap present with the resulting pocket extending posteri- below lower lip (1); or two transverse rows orly to cover a large section of m. masseter of flaps present (2). Most bats do not have (Murray and Strickler, 1975). These pouches transverse flaps below the lower lip. In con- are apparently used primarily for food stor- trast, extant species of Pteronotus and Mor- age during foraging (Murray and Strickler, moops have one or more transverse flaps of 1975). Data were not available for Pterono- skin that project from the lower lip. Extant tus pristinus, which was scored ‘‘?’’ for this species of Pteronotus have a single trans- character. verse flap on the lower lip, while both spe- Character 79: Lower lip and chin without cies of Mormoops have two rows of trans- multiple dermal papillae (0); or with multi- verse flaps. Among the outgroups, both spe- ple dermal papillae (1). Multiple dermal pa- cies of Noctilio have two rows of transverse pillae are present on the lower lip and chin flaps below the lower lip. Saccopteryx bili- in all extant species of Pteronotus and Mor- neata, both species of Mystacina, Artibeus moops. Among the outgroups, similar mul- jamaicensis, and both species of Macrotus tiple papillae are present in Artibeus jamai- lack flaps transverse flaps entirely. Data were censis. Multiple papillae are absent in both not available for Pteronotus pristinus, which species of Macrotus, both species of Nocti- was scored ‘‘?’’ for this character. lio, both species of Mystacina, and Saccop- Character 83: Anterior edge of labio-na- teryx bilineata. Data were not available for sal region protrudes anteriorly beyond edge Pteronotus pristinus, which was scored ‘‘?’’ of lower lip, muzzle appears pointed (0); or for this character. edge does not protrude anteriorly, muzzle Character 80: Lower lip of small or mod- appears blunt (1). The anterior edge of the erate size, dorsoventral height less than labio-nasal region does not protrude anteri- height of nostrils (0); or lip large and plate- orly beyond the edge of the lower lip in any like, dorsoventral height equal or greater extant species of Pteronotus or Mormoops. than the nostrils (1). Both species of Mor- As a result, the muzzle in these taxa appears moops have a lower lip of moderate size, to be relatively blunt (i.e., not pointed). A characterized by dorsoventral height less similar condition is seen among the out- 38 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 groups in Artibeus jamaicensis and both spe- and flatter than that seen in Mormoops and cies of Macrotus. In contrast, the anterior it lacks a papillated surface, instead having a edge of the labio-nasal region does protrude smooth anterior face. Data were not available anteriorly beyond edge of lower lip in Sac- for Pteronotus pristinus, which was scored copteryx bilineata, both species of Noctilio, ‘‘?’’ for this character. and both species of Mystacina, giving the Character 86: Anterodorsal edge of nar- muzzle a pointed appearance. Data were not ial pad unadorned, without ridge of skin or available for Pteronotus pristinus, which was noseleaf (0); or with papillated ridge of skin scored ‘‘?’’ for this character. present (1); or with noseleaf present. Both Character 84: Nostrils tubelike, openings species of Mormoops lack a noseleaf or ridge face slightly laterally and are not adjacent of skin on the anterodorsal edge of the narial (0); or nostrils open directly adjacent to each pad. In contrast, all extant species of Pter- other and face anteriorly (1). The nostrils onotus have a ridge of skin on the antero- open adjacent to each other and face anteri- dorsal edge of the narial pad. This ridge has orly in all extant species of Pteronotus. In an uneven, papillated dorsal edge that gives contrast, both species of Mormoops have the impression of being composed of a series tubelike nostrils that open slightly laterally. of indistinct dermal papillae. Among the out- Saccopteryx bilineata, both species of Noc- groups, the anterodorsal edge of the narial tilio, and both species of Mystacina also have pad is unadorned and lacks a ridge or nose- tubelike nostrils, while both species of Ma- leaf in Saccopteryx bilineata, both species of crotus and Artibeus jamaicensis have nostrils Mystacina, and both species of Noctilio. Ar- that are adjacent and open anteriorly. Data tibeus jamaicensis and both species of Ma- were not available for Pteronotus pristinus, crotus have a well-developed noseleaf that which was scored ‘‘?’’ for this character. projects from the anterodorsal edge of the Character 85: Narial pad with v-shaped narial pad. Transformations in this character notch present between nostrils (0); or narial were ordered 0 ↔ 1 ↔ 2toreflect the hy- pad with narrow, papillated vertical ridge pothesis that presence of a ridge of skin on between nostrils (1); or narial pad with wide, the anterodorsal narial pad represents a nec- smooth-sided vertical ridge between nostrils essary intermediate condition between pres- (2); or narial pad flat, with no notch or ridge ence of a noseleaf and absence of all struc- between nostrils (3). tures on the edge of the narial pad. Data were The narial pad in bats is highly variable in not available for Pteronotus pristinus, which form and ornamentation. There is no notch was scored ‘‘?’’ for this character. or ridge between the nostrils on the narial Character 87: No dermal papillae on dor- pad in Pteronotus quadridens, P. personatus, sal surface of nostril (0); or one dermal pa- P. davyi, and P. gymnonotus. In these forms, pillae present on dorsal surface of each nos- the narial pad is flat and unornamented. An- tril (1). All extant members of the genus other condition is seen in Pteronotus parnel- Pteronotus lack dermal papillae on the dorsal lii and P. macleayii, which have a v-shaped surface of the nostrils. In contrast, both spe- notch in the narial pad between the nostrils. cies of Mormoops have one dermal papilla In contrast, both species of Mormoops have on the dorsal surface of each nostril. This a narrow, papillated vertical ridge between papilla is not located on the anterodorsal the nostrils. This structure is quite prominent edge of the narial pad (as are the ridges and and has smooth lateral edges, but the anterior noseleaves described in character 86), but the face of the ridge gives the impression of be- papilla is instead set back approximately ing composed of a series of indistinct dermal midway along the tubelike dorsal aspect of papillae. Among the outgroups, both species the nostril. All of the outgroups lack dermal of Mystacina have a flat narial pad, while papillae on the dorsal surface of the nostrils. there is a notch in the narial pad between the Data were not available for Pteronotus pris- nostrils in Saccopteryx bilineata and both tinus, which was scored ‘‘?’’ for this char- species of Noctilio. Artibeus jamaicensis and acter. both species of Macrotus and have a ridge Character 88: No dermal projection pre- between the nostrils, but this ridge is wider sent lateral to the nostrils (0); or dermal pro- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 39

jection present, rounded distally, lateral side ed by preservation in museum specimens of projection continuous with the lip (1); or (e.g., degree of desiccation in fluid-preserved projection triangular and pointed distally, material). Our investigations led us to con- lateral side not continuous with lip (2). Most clude that these ridges represent skin-covered bats lack dermal projections lateral to the auricular muscles. Due to difficulties asso- nostrils, but such structures are present in all ciated with interpreting the degree of devel- extant members of Pteronotus and Mor- opment of these low ridges in preserved moops. In these taxa, a single dermal projec- specimens (and our inability to clearly iden- tion of variable form extends anterodorsally tify any taxa that lack them entirely), we from a position lateral to the nostrils. This chose not to score their presence as a sepa- projection is rounded distally in Pteronotus rate character state. Among the outgroups, parnellii and P. personatus, and its lateral Artibeus jamaicensis, both species of Nocti- side is continuous with the lip. This arrange- lio, both species of Mystacina, and Saccop- ment makes the upper lip appear swollen on teryx bilineata lack an interauricular band, either side of the narial pad. In contrast, the while a prominent interauricular band is pre- dermal projection in both species of Mor- sent in both species of Macrotus. Data were moops, P. davyi, P. gymnonotus, P. macle- not available for Pteronotus pristinus, which ayii, and P. quadridens is triangular in shape was scored ‘‘?’’ for this character. and tapers distally to a point. The projections Character 91: Pinna not funnel-shaped, have well-defined lateral sides that are not base of pinna located well above the level of continuous with the lip. As a result, the lip the mouth (0); or pinna funnel-shaped, base does not appear swollen in these taxa. Der- of pinna extends to level or below the level mal projections lateral to the nostrils are ab- of the mouth (1). All extant species of Pter- sent in all of the outgroups. Data were not onotus and Mormoops have funnel-shaped available for Pteronotus pristinus, which was pinnae in which the base of the pinnae ex- scored ‘‘?’’ for this character. tends below the level of the mouth. Among Character 89: No dermal tubercle on the outgroups, both species of Noctilio have midline of muzzle posterodorsal to nostrils funnel-shaped pinnae, while the other species (0); or prominent dermal tubercle present do not. Data were not available for Pteron- (1). Pteronotus macleayii, P. quadridens, P. otus pristinus, which was scored ‘‘?’’ for this davyi, and P. gymnonotus lack a conspicuous character. dermal tubercle on the dorsal aspect of the Our scoring of this feature agrees with the muzzle. In contrast, Pteronotus parnellii, P. descriptions of Smith (1972) and Hall personatus, and both members of the genus (1981), but differs from some literature ac- Mormoops have a prominent dermal tubercle counts which have contrasted the ‘‘funnel- located on the midline of the muzzle above shaped’’ ears in Mormoops with ‘‘narrow’’ the nostrils. The dermal tubercle is absent in ears in Pteronotus (e.g., Koopman, 1984). all of the outgroups. Data were not available The latter distinction refers to the form of the for Pteronotus pristinus, which was scored distal pinna, which we treat as a separate ‘‘?’’ for this character. character below. Character 90: Interauricular band absent Character 92: Distal pinna lanceolate, ta- (0); or prominent interauricular band pre- pers to blunt point (0); or distal pinna sent (1). Both species of Mormoops have a squared, pinna height relatively uniform prominent, v-shaped interauricular band of across the entire width of the pinna (1); or skin that connects the pinnae across the top distal pinna rounded (2). All extant species of the head. In contrast, all extant species of of Pteronotus have pinnae that are lanceolate Pteronotus lack an interauricular band. In and taper to a blunt point. In contrast, both these taxa, a very low, inconspicuous ridge species of Mormoops have pinnae that are may be seen to extend medially from the squared distally with an edge that is of equal base of each pinna. The degree of promi- height across most of the breadth of the pin- nence of these ridges varies among individ- na. Among the outgroups, the distal edge of uals and perhaps among species (see Smith, the pinnae is rounded in both species of Ma- 1972), and appears to be significantly affect- crotus and lancolate in Artibeus jamaicensis, 40 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 both species of Noctilio, both species of Mys- pristinus, which was scored ‘‘?’’ for this tacina, and Saccopteryx bilineata. Data were character. not available for Pteronotus pristinus, which Character 96: Secondary fold on tragus was scored ‘‘?’’ for this character. absent (0); or fold present, smaller than body Character 93: Lateral edge of distal pin- of tragus (1); or fold present, equal to or na smooth (0); or serrated (1). As in most larger than body of tragus (2). The second- bats, Pteronotus parnellii, P. davyi, P. gym- ary fold is a distally rounded fold of skin that nonotus, and both species of Mormoops have extends from the ventromedial aspect of the pinnae with smooth edges. In contrast, P. tragus, lying roughly at a right angle to the macleayii, P. quadridens, and P. personatus main longitudinal axis of that structure have several serrations on the lateral edge of (Smith, 1972). The thickness of the fold is the distal pinna. The outgroups all have pin- related to its size; when the fold is approxi- nae with a smooth lateral edge. Data were mately the same size as the body of the tra- not available for Pteronotus pristinus, which gus, it is approximately the same thickness, was scored ‘‘?’’ for this character. while smaller folds are relatively thinner. Character 94: Distal tip of tragus round- Most bats lack a secondary fold on the tra- ed (0); or tapers to point (1); or bifurcate gus. In contrast, all extant species of the ge- (2). The distal tip of the tragus is rounded in nus Pteronotus have a secondary fold that is all extant species of Pteronotus and Mor- smaller than the body of the tragus, while moops. Among the outgroups, a similar con- both species of Mormoops have a secondary dition is seen in Saccopteryx bilineata. In fold that is equal to or larger than the body contrast, the distal tragus tapers to a point in of the tragus. The outgroups all lack a sec- ondary fold. Transformations in this charac- Artibeus jamaicensis, both species of Macro- ↔ ↔ tus, and both species of Mystacina. An alter- ter were ordered 0 1 2toreflect the native state is seen in both species of Noc- logical sequence of transformation from ab- tilio, which have a tragus with a bifurcate tip. sence of a secondary fold (‘‘0’’) to a second- ary fold equal to or larger than the body of Data were not available for Pteronotus pris- the tragus (‘‘2’’). Data were not available for tinus, which was scored ‘‘?’’ for this char- Pteronotus pristinus, which was scored ‘‘?’’ acter. for this character. Character 95: Tragus thin, of uniform Character 97: Two interramal vibrissae thickness throughout (0); or tragus with dis- present (0); or one interramal vibrissa pre- tinctly thickened area at base (1); or tragus sent (1); or interramal vibrissae absent (2). with a thickened central area that extends Vibrissae (‘‘whiskers’’) occur in several lo- from midpoint to tip of tragus; lateral and cations on the face of most mammals includ- medial edges of tragus thin (2); or tragus ing bats. Interramal vibrissae are found under uniformly thick (3). All extant members of the chin between the rami of the lower jaws, the genus Pteronotus have a tragus that is posterior to the mandibular symphysis (Po- relatively thin and of roughly uniform thick- cock, 1914; Brown, 1971). All extant species ness throughout. In contrast, the base of the of Pteronotus are characterized by presence tragus is thickened in both species of Mor- of a single interramal vibrissa. In contrast, moops. Among the outgroups, Saccopteryx both species of Mormoops have two inter- bilineata, both species of Noctilio, and both ramal vibrissae. Among the outgroups, Sac- species of Mystacina have a uniformly thin copteryx bilineata, both species of Noctilio, tragus that resembles that of Pteronotus. and Artibeus jamaicensis have two interra- Both species of Macrotus are characterized mal vibrissae. Mystacina tuberculata has one by a tragus with a vertically oriented, thick- interramal vibrissa, and both species of Ma- ened central area that extends up the poste- crotus lack interramal vibrissae entirely. rior face of the tragus from its midpoint to Transformations in this character were or- its tip. The lateral and medial edges of tragus dered 0 ↔ 1 ↔ 2toreflect our hypothesis are thin in this taxon. In contrast, Artibeus that the number of vibrissae in the interramal jamaicensis has a tragus that is uniformly cluster changes by sequential addition (or thick. Data were not available for Pteronotus loss) of individual vibrissae. Data were not 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 41

available for Pteronotus pristinus and Mys- sae. A similar condition is seen among the tacina robusta, which were scored ‘‘?’’ for outgroups in Mystacina tuberculata and Ar- this character. tibeus jamaicensis. In contrast, a single su- Character 98: Interramal tubercle well perciliary vibrissa is present above each eye developed (0); or absent (1). The interramal in Saccopteryx bilineata, both species of tubercle (ϭ ventral sinus-hair tubercle of Da- Noctilio, and both species of Macrotus. Data lquest and Werner, 1954) is a swelling as- were not available for Pteronotus pristinus sociated with the interramal vibrissae (see and Mystacina robusta, which were scored character 97) in some bats. The interramal ‘‘?’’ for this character. tubercle is well developed (width greater Character 101: Vibrissae on face poste- than one-half the distance between the man- rior and lateral to narial pad arranged in a dibular rami) in all extant species of Pter- single vertical column (0); or arranged in onotus and Mormoops. Among the out- two roughly parallel vertical columns (1). groups, this tubercle is similarly well devel- Wetterer et al. (2000) noted that there is var- oped in Saccopteryx bilineata, Mystacina iation in the arrangement of the vibrissae on tuberculata, both species of Noctilio, and the face posterior and lateral to the narial both species of Macrotus. In contrast, the in- pad. There are two vertical columns of vi- terramal tubercle is absent in Artibeus. The brissae in this region in all extant species of interramal vibrissae in this taxon originate Pteronotus and Mormoops: an anteromedial from a tiny pitlike perforation in the skin. column located adjacent to the narial pad, Data were not available for Pteronotus pris- and a posterolateral column located posterior tinus and Mystacina robusta, which were to the anteromedial column. These columns scored ‘‘?’’ for this character. are roughly parallel in orientation, and each Character 99: Two genal vibrissae pre- contains at least two vibrissae. A similar ar- sent on each cheek (0); or one genal vibrissa rangement is seen among the outgroups in present on each cheek (1); or genal vibrissae both species of Macrotus and both species of absent (2). Genal vibrissae are located on the Noctilio. In contrast, the posterolateral col- cheek ventral to the eye and subocular vi- umn is absent in Saccopteryx bilineata, Mys- brissae (Pocock, 1914; Brown, 1971). Both tacina tuberculata, and Artibeus jamaicensis. species of Mormoops lack genal vibrissae en- Data were not available for Pteronotus pris- tirely. In contrast, all extant species of Pter- tinus and Mystacina robusta, which were onotus have a single genal vibrissa on each scored ‘‘?’’ for this character. cheek. Among the outgroups, Saccopteryx Character 102: Four vibrissae present in bilineata, both species of Noctilio, and Mys- anteromedial column (0); or three vibrissae tacina tuberculata have a pair of genal vi- (1); or two vibrissae (2). The anteromedial brissae on each cheek. Both species of Ma- column of vibrissae on the face (which is lo- crotus have a single genal vibrissa on each cated adjacent to the narial pad; see character cheek, and Artibeus jamaicensis lacks genal 101) includes only two vibrissae in all extant vibrissae entirely. Transformations in this species of Pteronotus and Mormoops. character were ordered 0 ↔ 1 ↔ 2toreflect Among the outgroups, a similar condition our hypothesis that the number of vibrissae occurs in both species of Noctilio. In con- in the genal cluster changes by sequential ad- trast, this column consists of four vibrissae dition (or loss) of individual vibrissae. Data in Saccopteryx bilineata, and three vibrissae were not available for Pteronotus pristinus in Mystacina tuberculata, both species of and Mystacina robusta, which were scored Macrotus, and Artibeus jamaicensis. Trans- ‘‘?’’ for this character. formations in this character were ordered 0 Character 100: Superciliary vibrissae ↔ 1 ↔ 2toreflect our hypothesis that the present (0); or absent (1). Superciliary (ϭ number of vibrissae in this column changes supraorbital) vibrissae are located dorsal to by sequential addition (or loss) of individual the eye in many mammals (Pocock, 1914; vibrissae. Data were not available for Pter- Brown, 1971). All extant species of Pteron- onotus pristinus and Mystacina robusta, otus and Mormoops lack superciliary vibris- which were scored ‘‘?’’ for this character. 42 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

PELAGE AND PATAGIA colored or bicolored dorsal fur. Transforma- tions in this character were ordered 0 ↔ 1 Character 103: Facial stripes absent (0); ↔ 2 ↔ 3 ↔ 4 ↔ 5 to accommodate ob- or paired stripes present (1). Facial stripes served fur-color polymorphisms (which we are absent in all extant species of Pteronotus regard as intermediate conditions between and Mormoops. The fur appears to be of uni- fixed states) and to reflect our hypothesis that form color between the narial pad and ears changes in fur banding patterns occur in a in these taxa. A similar condition is seen sequential fashion by adding and subtracting among the outgroups in Saccopteryx bilinea- individual color bands. Data were not avail- ta, both species of Noctilio, both species of able for Pteronotus pristinus and Mystacina Mystacina, and both species of Macrotus. In robusta, which were scored ‘‘?’’ for this contrast, Artibeus is characterized by a pair character. of pale facial stripes that run from the ante- Character 105: Ventral fur unicolored rior rostrum to near the base of the ear. These (0); or polymorphic, either unicolored or bi- stripes are roughly parallel, with one dorsal colored (1); or bicolored (2); or polymor- to and one ventral to the eye. Brightness of phic, either bicolored or tricolored (3). The these stripes may vary among individuals ventral fur of bats exhibits a range of band- and populations, but facial stripes are present ing patterns similar to those seen in the dor- in all adults. Data were not available for sal fur. While dorsal and ventral banding pat- Pteronotus pristinus, which was scored ‘‘?’’ terns in some species are identical, they are for this character. often different so we scored these as separate Character 104: Dorsal fur unicolored characters. Pteronotus gymnonotus has uni- (0); or polymorphic, either unicolored or bi- colored ventral fur, while P. parnellii, P. ma- colored (1); or bicolored (2); or polymor- cleayii, P. quadridens, P. personatus, P. dav- phic, either bicolored or tricolored (3); or yi, and Mormoops blainvillii have bicolored tricolored (4); or with four bands of color ventral fur. Mormoops megalophylla is poly- (5). The fur of adult bats consists of hairs morphic for this feature, with either bicol- that may have alternating bands of dark and ored or tricolored ventral fur. Among the out- pale colors along the shaft. Hairs may be un- groups, both species of Noctilio have unico- icolored (of uniform color along the shaft), lored ventral fur, while bicolored fur occurs bicolored (two bands of color), tricolored in Saccopteryx bilineata, Mystacina tuber- (three bands), or have four distinct bands of culata (Dwyer, 1960), and both species of color. Because the pattern of color banding Macrotus. Artibeus jamaicensis is polymor- may vary over the surface of the body (e.g., phic, with either unicolored or bicolored ven- fur over the rump may have fewer color tral fur. As with character 104, transforma- bands than does fur over the shoulders), in tions in patterns of ventral fur banding were the present analysis we restricted our com- ordered 0 ↔ 1 ↔ 2 ↔ 3 to accommodate parisons to the dorsal fur on the upper back observed fur-color polymorphisms (which over the scapulae. Pteronotus davyi and P. we regard as intermediate conditions be- gymnonotus have unicolored fur in this re- tween fixed states) and to reflect our hypoth- gion. In contrast, P. personatus and Mor- esis that changes in fur banding patterns oc- moops blainvillii have bicolored dorsal fur, cur in a sequential fashion by adding and and Pteronotus macleayii and P. quadridens subtracting individual color bands. Data were have tricolored fur. P. parnellii is polymor- not available for Pteronotus pristinus and phic for this feature, with either bicolored or Mystacina robusta, which were scored ‘‘?’’ tricolored fur (see discussion in Smith, for this character. 1972). Mormoops megalophylla has fur with Character 106: Dorsal fur short, 4Ð7mm four color bands. Among the outgroups, Sac- in length (0); or long, 8Ð10 mm in length (1). copteryx bilineata and both species of Noc- All extant species of Pteronotus have rela- tilio have unicolored fur, and bicolored fur tively short hair, approximately 4Ð7mmin occurs in Mystacina tuberculata (Dwyer, length. In contrast, both species of Mor- 1960) and both species of Macrotus. Artibeus moops have longer hair, approximately 8Ð10 jamaicensis is polymorphic, with either uni- mm in length. Among the outgroups, both 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 43

species of Macrotus, Artibeus jamaicensis, scales exhibit a continuous range of varia- and Mystacina tuberculata and have long tion from entire coronal to repand coronal hair, while Saccopteryx bilineata and both (1); or alternating entire coronal and hastate species of Noctilio have short hair. Data were coronal scales (2); or all scales dentate cor- not available for Pteronotus pristinus and onal (3); or alternating dentate coronal and Mystacina robusta, which were scored ‘‘?’’ hastate coronal scales (4); or all scales den- for this character. ticulate coronal (5). Benedict (1957) de- Character 107: Dorsal stripes always ab- scribed the structure of hair filaments and sent (0); pair of pale dorsal stripes always scales in many bats including all of the taxa present (1); or single pale middorsal stripe in our study. Benedict (1957) recognized two often present (2). Contrasting dorsal stripes principal types of hair scales: ‘‘imbricate,’’ are present in some species of bats. All ex- which overlap laterally, and ‘‘coronal,’’ tant species of Pteronotus and Mormoops which encircle the filament like a ring. Seen lack dorsal stripes. A similar condition is in lateral view, coronal scales on a hair fila- seen in both species of Mystacina, both spe- ment look like cups. ‘‘Entire’’ coronal scales cies of Macrotus, and Artibeus jamaicensis. are rings with uniformly level edges, so that A single, dark middorsal stripe is often (but each individual scale appears approximately not always) present in both species of Noc- the same height in all views (Benedict, tilio. Saccopteryx bilineata is characterized 1957). ‘‘Repand’’ coronal scales are the same by a pair of pale dorsal stripes. Data were shape as entire coronal scales, but have a not available for Pteronotus pristinus, which slightly uneven edge on one side, a feature was scored ‘‘?’’ for this character. not seen in entire coronal scales (Benedict, Character 108: Underhair and overhair 1957). ‘‘Hastate’’ coronal scales are charac- not differentiated in dorsal pelage (0); or terized by having one part of the distal ring clearly differentiated (1). Benedict (1957) edge that is significantly lower than the rest described hair structure in 165 genera and of the edge (Benedict, 1957). The low point 243 species representing most families of on any given scale is not necessarily in the bats. He found that many bats have clearly same position as the low points on the sur- differentiated overhair and underhair in the rounding scales. ‘‘Dentate’’ and ‘‘denticu- dorsal pelage. In these forms, individual ov- late’’ coronal scales are characterized by erhairs are straighter (and may thus appear presence of serrations along the edges of the longer) and are coarser than individual un- scale (Benedict, 1957). These serrations are derhairs. In contrast, all hairs in the dorsal small and close together in denticulate scales, pelage are of similar waviness and coarse- and are somewhat larger and farther apart in ness in species that lack differentiated over- dentate scales (Benedict, 1957). Among mor- hair and underhair (Benedict, 1957). All ex- moopids, Pteronotus davyi and P. gymnon- tant species of Pteronotus and Mormoops otus have entire coronal scales (Benedict, have clearly differentiated underhair and ov- 1957). P. parnellii has scales that exhibit a erhair (Benedict, 1957). Among the out- continuous range of variation from entire groups, clearly differentiated overhair and coronal to repand coronal (Benedict, 1957). underhair is seen only in species of Noctilio Denticulate coronal scales are present in (Benedict, 1957). The dorsal pelage in Sac- Pteronotus macleayii, P. quadridens, and P. copteryx bilineata, Mystacina tuberculata, personatus (Benedict, 1957). Dentate coronal both species of Macrotus, and Artibeus ja- scales are present in both species of Mor- maicensis is not differentiated into overhair moops (Benedict, 1957). Among the out- and underhair (Benedict, 1957). Due to the groups, Saccopteryx bilineata is character- variation among outgroups, the primitive ized by denticulate coronal hair scales, and condition could not be determined a priori. both species of Noctilio have scales that Data were not available for Pteronotus pris- range from entire coronal to repand coronal tinus and Mystacina robusta, which were (Benedict, 1957). Mystacina tuberculata has scored ‘‘?’’ for this character. only entire coronal scales, and both species Character 109: Midshaft scales on dorsal of Macrotus have alternating entire coronal hair filaments all entire coronal (0); or and hastate coronal scales (Benedict, 1957). 44 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Artibeus jamaicensis has alternating dentate (1). The wing membrane of bats are divided coronal and hastate coronal scales (Benedict, into the plagiopatagium (which extends be- 1957). Data were not available for Pterono- tween the body and the fifth digit), dactylo- tus pristinus and Mystacina robusta, which patagium (which extends between the digits), were scored ‘‘?’’ for this character. propatagium (which extends between the up- Character 110: Midshaft scales on dorsal per arm and forearm anterior to the elbow), hair filaments divaricate from hair shaft (0); and uropatagium (which extends between the or scales divergent from hair shaft (1); or hindlimbs). The medial attachment of the appressed to hair shaft (2). Variation occurs plagiopatagium is on the side of the body in in the degree to which hair scales diverge most bats, with dorsal fur exposed between from the shaft of the filament (Benedict, the right and left membranes. This condition 1957). Scales that are ‘‘appressed’’ lie flat is seen in Pteronotus parnellii, P. macleayii, against the filament so that the sides of the P. quadridens, P. personatus, and both spe- scale are parallel with the long axis of the cies of Mormoops. In contrast, the plagio- hair shaft, forming a smooth surface with ad- patagium attaches at the mid-dorsal line in jacent scales up and down the shaft (Bene- Pteronotus davyi and P. gymnonotus. The dict, 1957). In contrast, ‘‘divergent scales’’ right and left membranes meet over the spine are flared distally. The proximal portion of and cover the dorsal fur, which is present but each scale has sides that are parallel with not visible in dorsal view. This arrangement each other and the long axis of the hair shaft, of the patagia gives these bats a ‘‘naked- much like those of appressed scales, but the backed’’ appearance. Functional morphology distal portion of each scale flares outward of this trait has not yet been fully investigat- away from the axis of the shaft (Benedict, ed, and implications for flight mechanics re- 1957). ‘‘Divaricate’’ scales flare even more main unknown. Bonaccorso et al. (1992) markedly, with no portion of the side of the studied thermoregulation in P. davyi and scale remaining parallel to the long axis of concluded that the arrangement of the wing the hair shaft (Benedict, 1957). Each divari- membranes does not appear to strongly in- cate scale increases in diameter continuously crease the rate of heat loss at relatively cool from its proximal border (where it is narrow- ambient temperatures, but may help facilitate est) to its distal edge (where it is widest). All heat dissipation at high ambient tempera- extant species of Pteronotus have midshaft tures. The plagiopatagium attaches on the hair scales that are appressed to the hair shaft side of the body in all of the outgroups. Data (Benedict, 1957). In contrast, both species of were not available for Pteronotus pristinus, Mormoops have divergent hair scales (Ben- which was scored ‘‘?’’ for this character. edict, 1957). Among the outgroups, Mysta- Character 112: Wings folded by flexing cina tuberculata and both species of Noctilio proximal phalanx of digits III and IV poste- have appressed hair scales, Artibeus jamai- riorly (toward dorsal surface of wing), distal censis has divergent scales, and Saccopteryx phalanges folded anteriorly (toward under- bilineata and both species of Macrotus have side of wing) (0); or distal phalanges of dig- divaricate scales (Benedict, 1957). Transfor- its III and IV folded anteriorly, proximal mations in this character were ordered 0 ↔ phalanges not folded (1); or all phalanges in 1 ↔ 2toreflect the seemingly progressive digits III, IV, and V folded anteriorly toward nature of degree of hair scale divergence the underside of the wing (0). Bats at rest from the shaft (see discussion in Benedict, fold their wings in a variety of different ways 1957). Data were not available for Pterono- that are constrained by anatomy of the joints tus pristinus and Mystacina robusta, which in the hand and fingers. Every species is were scored ‘‘?’’ for this character. characterized by a single folding pattern, and Character 111: Plagiopatagium attaches wings cannot be forced into different patterns to the side of the body, dorsal fur exposed without damaging the joints, ligaments, and on back between right and left wing mem- muscles of the limb. All extant species of branes (0); or plagiopatagium attaches at Pteronotus and Mormoops fold their wings the mid-dorsal line, right and left membranes by flexing all phalanges in digits III, IV, and meet over spine and cover dorsal fur entirely V anteriorly toward the underside of the 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 45

wing. A similar pattern is seen among the sive fashion. Data were not available for outgroups in both species of Macrotus and Pteronotus pristinus, which was scored ‘‘?’’ Artibeus jamaicensis. In contrast, both spe- for this character. cies of Noctilio fold their wings by flexing Character 115: Uropatagium attaches to the distal phalanges of digits III and IV an- calcar, propatagium does not (0); or both teriorly without folding the proximal phalan- plagiopatagium and uropatagium attach to ges of those digits. A third pattern is seen in calcar (1). In most bats, including all of the both species of Mystacina and Saccopteryx outgroups in our study, the uropatagium at- bilineata, which flex the proximal phalanx of taches to the calcar while the plagiopatagium digits III and IV posteriorly while folding the attaches directly to the lateral surface of the distal phalanges anteriorly. Data were not leg and sometimes the foot. In contrast, both available for Pteronotus pristinus, which was the plagiopatagium and the uropatagium at- scored ‘‘?’’ for this character. tach to the calcar in all extant species of Character 113: Antebrachial wing sac Pteronotus and Mormoops. Data were not absent (0); or present (1). All extant species available for Pteronotus pristinus, which was of Pteronotus and Mormoops lack an ante- scored ‘‘?’’ for this character. brachial wing sac. The same is true of both Character 116: Calcar not bound to tibia species of Noctilio, both species of Mysta- (0); or caudal half of calcar bound to tibia, cina, Artibeus jamaicensis, and both species calcar restricted to position parallel to tibia of Macrotus. In contrast, Saccopteryx bili- (1). Pteronotus davyi, P. gymnonotus, and neata is characterized by presence of an an- both species of Mormoops have a calcar that tebrachial wing sac. This structure, which is is not bound to the tibia. In contrast, the cau- sexually dimorphic (larger in males), is lo- dal half of the calcar in P. parnellii, P. ma- cated in the propatagium anterior to the el- cleayii, P. quadridens, and P. personatus is bow. Data were not available for Pteronotus bound to the tibia by connective tissue, thus pristinus, which was scored ‘‘?’’ for this restricting the possible range of motion of the character. calcar. All of the outgroups have a calcar that Character 114: Uropatagium broad (0); is not bound to the tibia. Data were not avail- or with V-shaped cutout (1); or rudimentary able for Pteronotus pristinus, which was (2). The extent of the uropatagium varies sig- scored ‘‘?’’ for this character. nificantly both among and within families of bats (Schutt and Simmons, 1998). In all ex- tant species of Pteronotus and Mormoops the POSTCRANIAL SKELETON uropatagium is broad, completely spanning the area between the hindlegs from the rear Character 117: Ventral processes of sixth end of the body to the calcars. The trailing cervical vertebra (C6) approximately the edge of the membrane extends between the same width as those of C2ÐC5 (0); or C6 tips of the calcars and is not marked by a ventral processes enlarged to twice the width notchlike cutout. A similar condition is seen of C2ÐC5 ventral processes (1). The cervical among the outgroups in Saccopteryx bilinea- vertebrae of most bats are characterized by a ta, both species of Noctilio, and both species pair of longitudinal, ventrally directed pro- of Macrotus. In contrast, the uropatagium in cesses located on the centrum. The ventral Artibeus jamaicensis is less extensive and is processes of the sixth cervical vertebra (C6) characterized by a large, anteriorly directed are twice as wide as those of the other cer- V-shaped notch in the trailing edge between vical vertebrae in all extant species of Pter- the calcars. Even more extreme reduction in onotus and Mormoops. A similar condition the uropatagium is seen in both species of is seen in Mystacina robusta and both spe- Mystacina, which have a rudimentary uro- cies of Noctilio. In contrast, the ventral pro- patagium that is restricted to narrow strips of cesses of C6 are similar in width to the ven- membranes along the legs. Transformations tral processes of C2ÐC5 in Saccopteryx bil- in this character were ordered 0 ↔ 1 ↔ 2to ineata, both species of Macrotus, and Arti- reflect our hypothesis that changes in the ex- beus jamaicensis. Data were not available for tent of the uropatagium occur in a progres- Pteronotus pristinus and Mystacina tuber- 46 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 culata, which were scored ‘‘?’’ for this char- on T1ÐT4 of equal height (1). The relative acter. height of the neural spines on T1ÐT4 varies Character 118: Seventh cervical vertebra among bats that have neural spines on these (C7) not fused to first thoracic vertebrae (T1) vertebrae. The height of the neural spines (0); or C7 fused to T1 (1). The centrum, zyg- progressively decreases from T1 through T4 apophyses, and transverse processes of the in Pteronotus macleayii, P. quadridens, P. seventh cervical vertebra (C7) are partially davyi, P. gymnonotus, and P. personatus. In to fully fused with those first thoracic ver- contrast, the neural spines on T1ÐT4 are of tebrae (T1) in members of several bat fami- equal height in P. parnellii. Among the out- lies (Simmons and Geisler, 1998). C7 and T1 groups, the height of the neural spines de- are fused in both species of Mormoops. In creases from T1 through T4 in Artibeus ja- contrast, there is no evidence of C7/T1 fu- maicensis. The remaining outgroups, which sion in any extant species of Pteronotus. C7 lack neural spines on T1ÐT4 (see character and T1 are not fused in any of the outgroups. 120), and were scored ‘‘–’’ for this character. Data were not available for Pteronotus pris- Data were not available for Pteronotus pris- tinus and Mystacina tuberculata, which were tinus and Mystacina tuberculata, which were scored ‘‘?’’ for this character. scored ‘‘?’’ for this character. Character 119: Thirteen thoracic verte- Character 122: Ridges on ventral surface brae present (0); or 12 present (1); or 11 of T4 absent (0); or two parallel ridges pre- present (2). All bats have between 11 and 14 sent on ventral surface of T4 (1). A pair of thoracic (rib-bearing) vertebrae (Walton and parallel ridges are present on the ventral sur- Walton, 1970). Both species of Mormoops face of T4 in all extant species of Pteronotus have 13 thoracic vertebrae, while all extant and Mormoops. A similar condition is seen species of Pteronotus have only 12 thoracic in Macrotus californicus and Mystacina ro- vertebrae. Among the outgroups, both spe- busta. In contrast, these ridges are absent in cies of Noctilio have 11 thoracic vertebrae, Macrotus waterhousii, Artibeus jamaicensis, Saccopteryx bilineata and both species of both species of Noctilio, and Saccopteryx bil- Macrotus have 12, and Mystacina robusta ineata. Data were not available for Pteron- and Artibeus jamaicensis have 13 thoracic otus pristinus and Mystacina tuberculata, vertebrae. This character was ordered 0 ↔ 1 which were scored ‘‘?’’ for this character. ↔ 2toreflect the presumed progressive se- Character 123: Neural spines absent quence of changes in vertebral counts. Data from posterior thoracic vertebrae (T8 and were not available for Pteronotus pristinus above) (0); or spines present on posterior and Mystacina tuberculata, which were thoracic vertebrae (1). Neural spines are ab- scored ‘‘?’’ for this character. sent from the posterior thoracic vertebrae (T8 Character 120: Neural spines absent and above) in Pteronotus personatus, P. from first four thoracic vertebrae (T1ÐT4) quadridens, and both species of Mormoops. (0); or spines present on T1ÐT4 (1). Neural In contrast, neural spines are present on the spines are absent from the first four thoracic posterior thoracic vertebrae in Pteronotus vertebrae (T1ÐT4) in both species of Mor- parnellii, P. macleayii, P. gymnonotus, and moops. In contrast, neural spines are present P. davyi. Among the outgroups, Artibeus ja- on T1ÐT4 in all extant species of Pteronotus. maicensis and Mystacina robusta each have Among the outgroups, Artibeus jamaicensis neural spines on the posterior thoracic ver- has neural spines on T1ÐT4, while neural tebrae. Neural spines are absent from the spines are absent from these vertebrae in posterior thoracic vertebrae in both species both species of Macrotus, Saccopteryx bili- of Macrotus, Saccopteryx bilineata, and both neata, both species of Noctilio, and Mysta- species of Noctilio. Data were not available cina robusta. Data were not available for for Pteronotus pristinus and Mystacina tub- Pteronotus pristinus and Mystacina tuber- erculata, which were scored ‘‘?’’ for this culata, which were scored ‘‘?’’ for this char- character. acter. Character 124: No fusion between pos- Character 121: Height of neural spines terior thoracic and anterior lumbar verte- decreases from T1ÐT4 (0); or neural spines brae (0); or two posteriormost thoracic ver- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 47 tebrae, L1, and L2 fused (1). Vertebral fusion dition is seen in both species of Macrotus, is absent in the lower back of most bats. both species of Noctilio, and Saccopteryx bil- However, the last two thoracic and first two ineata. In contrast, the anterior face of the lumbar vertebrae fused together in all extant manubrium is a broad, triangular surface that species of Pteronotus and both species of extends onto the lateral processes and is de- Mormoops. These vertebrae are not fused in fined by three elevated ridges in Artibeus ja- any of the outgroups. Data were not available maicensis and Mystacina robusta. Data were for Pteronotus pristinus and Mystacina tub- not available for Pteronotus pristinus and erculata, which were scored ‘‘?’’ for this Mystacina tuberculata, which were scored character. ‘‘?’’ for this character. Character 125: Ridges on ventral surface Character 128: Angle between axis of of third lumbar vertebra (L3) absent (0); or ventral process of manubrium and body of two parallel ridges present on ventral sur- manubrium approximately 90Њ (0); or great- face of L3 (1). The third lumbar vertebra (L3) er than 90Њ (1). The ventral process of the lacks ridges on its ventral surface in Mor- manubrium is a bony projection that provides moops megalophylla. In contrast, M. blain- the anterior attachment point for a series of villii and all extant species of Pteronotus ligamentous sheets that run down the midline have two parallel ridges present in this lo- of the sternum. Together with these ligamen- cation. Among the outgroups, Saccopteryx tous sheets, the ventral process of the ma- bilineata, Mystacina robusta, and Artibeus nubrium forms the origin for the m. pecto- jamaicensis exhibit similar ventral ridges on ralis complex, which provides most of the L3, but both species of Macrotus and both power for the downstroke during flight species of Noctilio lack these ridges. Data (Vaughan, 1959, 1970b; Norberg, 1970; were not available for Pteronotus pristinus Strickler, 1978; Hermanson and Altenbach, and Mystacina tuberculata, which were 1983, 1985). The angle between the axis of scored ‘‘?’’ for this character. the ventral process (defined as the long axis Character 126: Ridges on ventral surface of the thickened base and central body of the of fourth and fifth lumbar vertebra (L4 and process) and the body of the manubrium L5) absent (0); or two parallel ridges present varies among bats (Simmons and Geisler, on ventral surface of both L4 and L5 (1). The 1998). All extant species of Pteronotus have fourth and fifth lumbar vertebrae (L4 and L5) an angle of more than 90Њ between the axis lack ridges on the ventral surface in both spe- of ventral process and the body of the ma- cies of Mormoops. In contrast, all extant spe- nubrium. As a result, the ventral process ap- cies of Pteronotus have two parallel ventral pears to project anteroventrally in these spe- ridges on the ventral surface of both L4 and cies. In contrast, the angle between the axis L5. Among the outgroups, paired ventral of the ventral process and the body of the ridges are present on L4 and L5 in Artibeus manubrium is approximately 90Њ and the jamaicensis, both species of Macrotus, and ventral process projects ventrally in both spe- Mystacina robusta, but are absent in Saccop- cies of Mormoops. Among the outgroups, the teryx bilineata and both species of Noctilio. angle between the axis of the ventral process Data were not available for Pteronotus pris- and the body of the manubrium is approxi- tinus and Mystacina tuberculata, which were mately 90Њ in Saccopteryx bilineata, both scored ‘‘?’’ for this character. species of Macrotus, and Artibeus jamaicen- Character 127: Anterior face of manu- sis, but this angle is greater than 90Њ in both brium small (0); or broad, defined by ele- species of Noctilio and Mystacina robusta. vated ridges (1). The anterior face of the ma- The primitive condition could not be deter- nubrium of the sternum is the site of origin mined a priori due to variation among the for the anterior division of m. pectoralis outgroups. Data were not available for Pter- (Vaughan, 1959, 1970b; Strickler, 1978; Her- onotus pristinus and Mystacina tuberculata, manson and Altenbach, 1983, 1985). The an- which were scored ‘‘?’’ for this character. terior face of the manubrium is relatively Character 129: Distal tip of ventral pro- small and poorly defined in all extant species cess of manubrium laterally compressed, of Pteronotus and Mormoops. A similar con- keel-like (0); or blunt and rounded (1). The 48 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 distal tip of the ventral process of the ma- erculata, which were scored ‘‘?’’ for this nubrium is laterally compressed and keel-like character. in all extant species of Pteronotus and Mor- Character 133: Second costal cartilage moops. Among the outgroups, a laterally articulates with sternum at manubrium-me- compressed, keel-like ventral process is also sosternum joint (0); or second rib articulates seen in Artibeus jamaicensis, both species of with manubrium only (1). The second costal Macrotus, both species of Noctilio, and Sac- cartilage articulates with the sternum at the copteryx bilineata. In contrast, the distal tip manubrium-mesosternum joint in all extant of the ventral process is blunt and rounded species of Pteronotus and Mormoops. A sim- in Mystacina robusta. Data were not avail- ilar condition is seen among the outgroups in able for Pteronotus pristinus and Mystacina Saccopteryx bilineata and Mystacina robus- tuberculata, which were scored ‘‘?’’ for this ta. In contrast, the second rib articulates with character. the manubrium and there is no contact be- Character 130: Length of manubrium ap- tween the rib (or costal cartilage) and the me- proximately equal to maximum width (0); or sosternum in Artibeus jamaicensis, both spe- less than maximum width (1). The length of cies of Macrotus, and both species of Noc- the manubrium posterior to the lateral pro- tilio. Data were not available for Pteronotus cesses is less than the transverse width of this pristinus and Mystacina tuberculata, which portion of the manubrium in all extant spe- were scored ‘‘?’’ for this character. cies of Pteronotus and Mormoops. A similar Character 134: Mesosternum articulates condition is seen in Mystacina robusta, both with five costal cartilages posterior to second species of Noctilio, Artibeus jamaicensis, and rib (0); or with six costal cartilages (1). Var- both species of Macrotus. In contrast, the iable numbers of costal cartilages connect the length of the manubrium posterior to the lat- mesosternum with the bony ribs in bats eral processes is approximately equal to its (Simmons and Geisler, 1998). On each side width in Saccopteryx bilineata. Data were of the body, the mesosternum articulates with not available for Pteronotus pristinus and six costal cartilages posterior to the second Mystacina tuberculata, which were scored rib in all extant species of Pteronotus and ‘‘?’’ for this character. Mormoops. A similar condition is seen in Character 131: Mesosternum with medial both species of Macrotus. In contrast, the keel low or absent (0); or with high keel (1). mesosternum articulates with only five costal The mesosternum has a low median keel, or cartilages posterior to the second rib in Ar- lacks the keel altogether, in all extant species tibeus jamaicensis, both species of Noctilio, of Pteronotus and Mormoops. A similar con- Mystacina robusta, and Saccopteryx bilinea- dition is seen in Artibeus jamaicensis, Mys- ta. Data were not available for Pteronotus tacina robusta, both species of Noctilio, and pristinus and Mystacina tuberculata, which Saccopteryx bilineata. In contrast, a high were scored ‘‘?’’ for this character. keel (maximum keel height Ͼ width of me- Character 135: Ribs with no anterior sosternum) is present in both species of Ma- laminae (0); or narrow anterior laminae pre- crotus. Data were not available for Pteron- sent (1); or wide anterior laminae present otus pristinus and Mystacina tuberculata, (2). Anterior laminae are thin plates of bone which were scored ‘‘?’’ for this character. that run along the leading edges of ribs an- Character 132: Xiphisternum with prom- terior to the main body of the rib (Simmons inent median keel (0); or without keel (1). and Geisler, 1998). These structures, which The xiphisternum has a prominent median are often nearly transparent, appear to pro- keel in all extant species of Pteronotus and vide an increased area for muscle attachment Mormoops. Among the outgroups, this con- (Simmons and Geisler, 1998). The anterior dition is a also seen in Mystacina robusta laminae are relatively narrow (lamina width and Saccopteryx bilineata. In contrast, the xi- less than that of the main body of the rib) in phisternum lacks a keel in both species of Pteronotus macleayii, P. quadridens, P. per- Noctilio, both species of Macrotus, and Ar- sonatus, P. davyi, P. gymnonotus, and Mor- tibeus jamaicensis. Data were not available moops blainvillii. In contrast, the anterior for Pteronotus pristinus and Mystacina tub- laminae are relatively wide (lamina width 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 49

equal to or greater than that of the main body between these two processes by a series of of the rib) in Pteronotus parnellii and Mor- ligaments (Strickler, 1978; Simmons and moops megalophylla. Among the outgroups, Geisler, 1998). The clavicle articulates with wide anterior laminae are seen in both spe- or lies in contact with the coracoid process cies of Noctilio, narrow laminae occur in in all extant species of Pteronotus and Mor- Saccopteryx bilineata, and anterior laminae moops. A similar condition is seen among are entirely absent in Mystacina robusta, the outgroups in Saccopteryx bilineata, both both species of Macrotus, and Artibeus ja- species of Macrotus, and Artibeus jamaicen- maicensis. Transformations in this character sis. In contrast, the clavicle is suspended by were ordered 0 ↔ 1 ↔ 2toreflect our hy- ligaments between the acromion and cora- pothesis that changes in size of the laminae coid processes in both species of Noctilio and occur in a progressive fashion. Data were not Mystacina robusta. The primitive condition available for Pteronotus pristinus and Mys- could not be determined a priori due to var- tacina tuberculata, which were scored ‘‘?’’ iation among the outgroups. Data were not for this character. available for Pteronotus pristinus and Mys- Character 136: Ribs with no posterior tacina tuberculata, which were scored ‘‘?’’ laminae (0); or narrow posterior laminae for this character. present (1); or wide posterior laminae pre- Character 138: Pit for attachment of cla- sent (2). Posterior laminae are thin plates of vicular ligament absent from scapula (0); or bone that run along the leading edges of ribs present anterior and medial to glenoid fossa posterior to the main body of the rib (Sim- (1). The clavicular ligament extends between mons and Geisler, 1998). Wide posterior the base of the coracoid process and the clav- laminae (lamina width equal to or greater icle (Strickler, 1978). A distinct pit for the than that of the main body of the rib) are attachment of the clavicular ligament is pre- seen in all extant species of Pteronotus and sent anterior and medial to the glenoid fossa Mormoops megalophylla. In contrast, the in all extant species of Pteronotus and Mor- posterior laminae are relatively narrow (lam- moops. In contrast, there is no evidence of a ina width less than that of the main body of pit for attachment of the clavicular ligament the rib) in Mormoops blainvillii. Narrow pos- in any of the outgroups. Data were not avail- terior laminae are present among the out- able for Pteronotus pristinus, which was groups in Saccopteryx bilineata, both species scored ‘‘?’’ for this character. of Noctilio, Artibeus jamaicensis, and both Character 139: Coracoid process elon- species of Macrotus. The posterior laminae gate, tip approximately the same width as the are entirely absent in Mystacina robusta. shaft (0); or coracoid process elongate, tip Transformations in this character were or- distinctly flared (1); or coracoid process re- dered 0 ↔ 1 ↔ 2toreflect our hypothesis duced to a small triangular process (2). The that changes in size of the laminae occur in coracoid process in most bats is an elongate a progressive fashion. Data were not avail- bar of bone with a distinct shaft and tip (Sim- able for Pteronotus pristinus and Mystacina mons and Geisler, 1998). The tip of the cor- tuberculata, which were scored ‘‘?’’ for this acoid process is blunt and is approximately character. the same width as the shaft of the coracoid Character 137: Distal clavicle articulates in both species of Mormoops. In contrast, the with or lies in contact with coracoid process tip of the coracoid process is distinctly flared (0); or clavicle suspended by ligaments be- (wider than the shaft) in all extant species of tween acromion and coracoid processes of Pteronotus. Among the outgroups, Artibeus scapula (1). The distal (dorsolateral) clavicle jamaicensis and both species of Macrotus articulates with the scapula in the region be- have a coracoid process with a flared tip, tween the tip of the acromion process and the while the tip of the coracoid in Saccopteryx base of the coracoid process. The articulation bilineata and both species of Noctilio is not can be accomplished in several different flared. In contrast to these conditions (all of ways, depending on whether the clavicle is which involve a relatively elongate cora- associated principally with the acromion, coid), the coracoid process of both species of principally with the coracoid, or is suspended Mystacina is reduced to a small triangular 50 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 process that lacks a definable shaft. Data for Pteronotus pristinus, which was scored were not available for Pteronotus pristinus, ‘‘?’’ for this character. which was scored ‘‘?’’ for this character. Character 142: Suprascapular process Character 140: Distal acromion process absent (0); or present (1). The suprascapular without anteromedial projection (0); or with process is a projection that extends medially triangular anteromedial projection that does and somewhat anteriorly from the anterolat- not contact rim of scapula (1); with elongate eral border of the suprascapular notch at the projection that extends anteriorly and medi- base of the coracoid process, just medial to ally to fuse with anteromedial rim of scapula the anterior edge of the glenoid fossa and just (2). A triangular anteromedial projection is anterior to the pit for the clavicular ligament present just ventral and medial to the tip of (in those forms that have this pit; Simmons the acromion process in all extant species of and Geisler, 1998). A suprascapular process Pteronotus and Mormoops. This projection is is present in all extant species of Pteronotus directed toward (but does not contact) the an- and Mormoops. Among the outgroups, this teromedial rim of the scapula. Among the process is similarly present in both species outgroups, a similar process is present in of Mystacina and both species of Noctilio. In both species of Noctilio. A different condi- contrast, the suprascapular process is absent tion is seen in both species of Mystacina, in Saccopteryx bilineata, Artibeus jamaicen- which have a very long anterior projection sis, and both species of Macrotus. The prim- that extends anteriorly and medially to con- itive condition could not be determined a tact and fuse with the anteromedial rim of priori due to variation among the outgroups. the scapula medial to the suprascapular Data were not available for Pteronotus pris- notch. The resulting bar of bone forms a tinus, which was scored ‘‘?’’ for this char- complete bridge over the supraspinous fossa. acter. A secondary sutural contact is also present Character 143: Dorsal articular facet on between the lateral surface of the bony scapula absent (0); or dorsal articular facet bridge and the tip of the suprascapular pro- faces dorsolaterally, consists of a small cess (see character 142), enclosing the supra- groove on the anteromedial rim of the glen- scapular notch to form a small canal. In con- oid fossa (1); or faces dorsolaterally, con- trast, the distal acromion process lacks an an- sists of oval facet on anteromedial rim of terior projection entirely in Saccopteryx bil- glenoid fossa (2); or faces dorsally, consists ineata, Artibeus jamaicensis, and both of a large, flat facet clearly separated from species of Macrotus. Transformations in this the glenoid fossa (3). A secondary articula- character were ordered 0 ↔ 1 ↔ 2toreflect tion between the humerus and scapula occurs our hypothesis that changes in extent of the in many bats when the humerus is abducted anterior projection of the acromion process and the trochiter (ϭ greater tuberosity) con- occur in a progressive fashion. Data were not tacts a dorsal articular facet on the scapula available for Pteronotus pristinus, which was (Vaughan, 1959, 1970b; Hill, 1974; Strickler, scored ‘‘?’’ for this character. 1978; Altenbach and Hermanson, 1987; Character 141: Distal acromion process Schlosser-Strum and Schleimann, 1995; Sim- without posterolateral projection (0); or with mons and Geisler, 1998). The dorsal articular triangular posterolateral projection (1). A facet faces dorsolaterally and consists of an triangular projection from the distal acromi- oval facet on the anteromedial rim of the on process is present in Pteronotus davyi and glenoid fossa in all extant species of Pter- P. gymnonotus. In contrast, this process is onotus and Mormoops. The same condition absent in Pteronotus personatus, P. quadri- is seen in Artibeus jamaicensis and both spe- dens, P. macleayii, P. parnellii, and both cies of Macrotus. Dorsal articular facet in species of Mormoops. Among the outgroups, Saccopteryx bilineata also faces dorsolater- a posterolateral projection is present in both ally, but consists of a small groove on the species of Noctilio, but this projection is ab- anteromedial rim of the glenoid fossa. In sent in Saccopteryx bilineata, both species of contrast, the dorsal articular facet faces dor- Mystacina, both species of Macrotus, and sally and consists of a large, flat facet clearly Artibeus jamaicensis. Data were not available separated from the glenoid fossa in both spe- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 51

cies of Mystacina. Both species of Noctilio facet in both species of Noctilio, both species lack a dorsal articular facet. Data were not of Mystacina, Artibeus jamaicensis, and both available for Pteronotus pristinus, which was species of Macrotus. Data were not available scored ‘‘?’’ for this character. for Pteronotus pristinus, which was scored Character 144: Infraspinous fossa of ‘‘?’’ for this character. scapula wide, maximum width greater than Character 146: Axillary border of scap- 50% of maximum length (0); or infraspinous ula flat, level within dorsoventral plane (0); fossa narrow, maximum width less than or or axillary border curved in dorsoventral equal to 50% of maximum length (1). The plane, concave ventrally (1). The axillary infraspinous fossa, located on the dorsal as- border of the scapula is that edge which faces pect of the scapula posterolateral to the scap- the axilla or armpit. The axillary border of ular spine, is the site of origin for m. infra- the scapula is flat and lies level within the spinatus and m. teres major, muscles that act dorsoventral plane in all extant species of to flex, rotate, and (in the case of m. infra- Pteronotus. In contrast, both species of Mor- spinatus) abduct the humerus (Vaughan, moops have a scapula with a curved axillary 1959, 1970b; Norberg, 1970; Strickler, 1978; border that is concave ventrally. All of the Hermanson and Altenbach, 1983, 1985). The outgroups have a scapula with a flat axillary infraspinous fossa is relatively wide in Mor- border that is level within the dorsoventral moops blainvillii and all extant species of plane. Data were not available for Pteronotus Pteronotus. In these taxa, the maximum pristinus and Mystacina tuberculata, which width of the fossa is greater than 50% of its were scored ‘‘?’’ for this character. maximum length. In contrast, M. megalo- Character 147: Anterior portion of axil- phylla has an infraspinous fossa that is rela- lary border of scapula rounded (0); or flat- tively narrow, with maximum width less than tened (1). The anterior portion of the axillary 50% of length. Among the outgroups, a wide border of the scapula is flattened to form a infraspinous fossa is seen in Saccopteryx bil- sharp edge in both species of Mormoops and ineata, Mystacina robusta, both species of all extant species of Pteronotus. The anterior Noctilio, and Artibeus jamaicensis. Both spe- portion is similarly flattened in both species cies of Macrotus have a narrow infraspinous of Macrotus. In contrast, the anterior portion fossa. Data were not available for Pteronotus is rounded with no definitive sharp edge in pristinus and Mystacina tuberculata, which Artibeus jamaicensis, both species of Nocti- were scored ‘‘?’’ for this character. lio, Mystacina robusta, and Saccopteryx bil- Character 145: Intermediate infraspinous ineata. Data were not available for Pteron- facet narrower than posterolateral facet (0); otus pristinus and Mystacina tuberculata, or facets subequal (1). The infraspinous fos- which were scored ‘‘?’’ for this character. sa of the scapula has three facets in all of the Character 148: Head of humerus round taxa in our study2: an anteromedial facet (lo- in outline in medial view (0); or oval or el- cated adjacent to the spine of the scapula), liptical (1). The head of the humerus, which an intermediate facet, and a posterolateral articulates with the glenoid fossa on the scap- facet (adjacent to the axillary border of the ula, is oval or elliptical in outline when seen scapula). The intermediate infraspinous facet in medial view3 in all extant species of Pter- is narrower than the posterolateral facet in onotus and Mormoops. A similar condition Pteronotus personatus and Mormoops me- is seen among the outgroups in both species galophylla. In contrast, these facets are sub- Noctilio. In contrast, Saccopteryx bilineata, equal in Pteronotus parnellii, P. macleayi, P. both species of Mystacina, both species of quadridens, P. davyi, P. gymnonotus, and Macrotus, and Artibeus jamaicensis have a Mormoops blainvillii. A similar condition is humeral head that is round in outline in when seen in Saccopteryx bilineata. The interme- seen in medial view. Data were not available diate facet is narrower than the posterolateral 3 Determination of anatomical directions for bat limb elements is complicated by the unusual orientation of 2 Our observations differ from those of Simmons and the limbs during flight. In specifying ‘‘medial view’’ in Geisler (1998), who erroneously reported that Mystacina this character, we have followed Smith (1972) and as- has only two facets in the infraspinous fossa. sumed that the wing is folded and the humerus adducted. 52 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Fig. 8. Anterior views of the distal end of the humerus of (A) Pteronotus parnellii, (B) Pteronotus personatus, and (C) Mormoops megalophylla (redrawn from Vaughan and Bateman, 1970: fig. 5). Note the differences between Pteronotus and Mormoops in degree of development of the capitulum, location of the trochlea relative the shaft, and presence/absence of a deep groove between the trochlea and distal spinous process. for Pteronotus pristinus, which was scored The combined articular surface typically ‘‘?’’ for this character. forms a spool-shaped structure, although the Character 149: Trochiter extends proxi- form of this structure varies considerably mally just to the level of the proximal edge among and within families of bats (Smith, of the humeral head (0); or extends proxi- 1972; Simmons and Geisler, 1998). The mally well beyond the level of the head (1). structural unit formed by the three distal ar- The trochiter (ϭ greater tuberosity) of the ticular surfaces lies in line with the shaft of humerus in bats extends varying distances the humerus in both species of Mormoops. proximally relative to the head of the hu- The medial edge of the trochlea minimally merus (Simmons and Geisler, 1998). The tro- lies in line with the medial surface of the chiter extends proximally just to the level of humeral shaft (e.g., fig. 8C) and may extend the proximal edge of the head in all extant medially beyond this level. In contrast, the species of Pteronotus and Mormoops. distal articular facet complex is laterally dis- Among the outgroups, a similar condition is placed from the line of the humeral shaft in seen in Saccopteryx bilineata and both spe- all extant species of Pteronotus. In these cies of Noctilio. In contrast, the trochiter ex- forms, the edge of the trochlea does not ex- tends proximally well beyond the level of the tend medially as far as the medial surface of humeral head in both species of Mystacina, the humeral shaft (fig. 8A, B). Among the both species of Macrotus, and Artibeus ja- outgroups, the distal articular facets are in maicensis. Data were not available for Pter- line with the shaft in Saccopteryx bilineata, onotus pristinus, which was scored ‘‘?’’ for both species of Noctilio, and both species of this character. Mystacina. These facets are displaced later- Character 150: Humerus with distal ar- ally in Artibeus jamaicensis and both species ticular facets in line with shaft, not displaced of Macrotus. Data were not available for laterally (0); or articular facets displaced Pteronotus pristinus, which was scored ‘‘?’’ laterally from line of shaft (1). The articular for this character. facets of the distal humerus include the Character 151: Central surface of capit- trochlea, capitulum, and lateral surface of the ulum of humerus not reduced, height equal capitulum (fig. 8; Vaughan and Bateman, to that of trochlea (0); or central surface of 1970; Smith, 1972). The articular surfaces lie the capitulum reduced, height less than that adjacent to one another and are distinguished of trochlea (1). The capitulum on the distal on the basis of a series of ridges and grooves. end of the humerus is an articular surface 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 53

comprised of two ridges with a groove be- condition is seen in Mystacina tuberculata. tween them (fig. 8). The central surface of In contrast, the height of the distal spinous the capitulum is more medial of these ridges, process is less than or equal to the height of lying adjacent to a third ridge, the trochlea the trochlea in Mystacina robusta, Artibeus (Vaughan, 1959; Vaughan and Bateman, jamaicensis, both species of Macrotus, both 1970; Smith, 1972). The central capitulum species of Noctilio, and Saccopteryx bilinea- surface is the same height as the trochlea in ta. Data were not available for Pteronotus all extant species of Pteronotus (fig. 8). In pristinus, which was scored ‘‘?’’ for this contrast, both species of Mormoops have a character. reduced central capitulum whose height is Character 154: Distal spinous process of less than that of the trochlea (fig. 8). All of humerus with no ventral projection (0); or the outgroups all have an unreduced central with ventral projection (1). The distal spi- capitulum surface that is equal in height to nous process of the humerus lacks a ventral the trochlea. Data were not available for projection in Mormoops megalophylla and Pteronotus pristinus, which was scored ‘‘?’’ all extant species of Pteronotus (fig. 8). In for this character. contrast, the distal spinous process of Mor- Character 152: Distal spinous process of moops blainvillii is characterized by a ventral humerus separated from trochlea by deep projection that extends perpendicular to the notch (0); or process located directly adja- shaft of the humerus. A similar ventral pro- cent to trochlea, notch absent (1). The distal jection is seen in both species of Macrotus. spinous process of the humerus is a bony Artibeus jamaicensis, both species of Mys- projection that extends distally from the me- tacina, both species of Noctilio, and Saccop- dial epicondyle of the humerus in bats (fig. teryx bilineata lack a ventral projection from 8; Vaughan and Bateman, 1970; Smith, the distal spinous process. Data were not 1972). This process is the site of origin for available for Pteronotus pristinus, which was m. flexor carpi ulnaris, which is a flexor of scored ‘‘?’’ for this character. the fifth metacarpal (Vaughan, 1959; Character 155: Olecranon process of Vaughan and Bateman, 1970; Smith, 1972). ulna well developed, extends proximally be- The distal spinous process is separated from yond end of radius (0); or reduced, does not the trochlea by a deep notch in all extant spe- extend beyond end of the radius (1). Both cies of Pteronotus (fig. 8). In contrast, this species of Mormoops have a well-developed process lies directly adjacent to the trochlea olecranon process that extends proximally (and the notch is absent) in both species of beyond the end of the radius. In contrast, all Mormoops (fig. 8). Among the outgroups, extant species of Pteronotus have a reduced the distal spinous process is separated from olecranon process that does not extend be- the trochlea by a deep notch in Saccopteryx yond the end of the radius. The olecranon bilineata, both species of Noctilio, Artibeus process is well -developed in all of the out- jamaicensis, and both species of Macrotus. group species. Data were not available for Both species of Mystacina have the distal Pteronotus pristinus, which was scored ‘‘?’’ spinous process located adjacent to the troch- for this character. lea and lack the notch. Data were not avail- Character 156: Pisiform bone rodlike, of able for Pteronotus pristinus, which was uniform width throughout (0); or dumbbell- scored ‘‘?’’ for this character. shaped, enlarged at both ends (1). The pisi- Character 153: Height of distal spinous form bone in bats lies on the ventral surface process of humerus less than or equal to of the carpus where it is attached by fascia height of trochlea (0); or greater than height to the trapezium and the metacarpal of the of trochlea (1). The height of the distal spi- fifth digit (Vaughan, 1959). All extant spe- nous process of the humerus (measured cies of Pteronotus have a pisiform bone that along the long axis of the spinous process, is rodlike in form, roughly uniform in width which runs roughly parallel to the long axis throughout its length. In contrast, both spe- of the humeral shaft) is greater than the cies of Mormoops have a pisiform that is en- height of the trochlea in all extant species of larged at both ends, giving it a dumbbell Pteronotus and Mormoops (fig. 8). A similar shape. A similar condition is seen among the 54 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 outgroups in Saccopteryx bilineata and both na, and both species of Macrotus. Data were species of Macrotus. Mystacina robusta, not available for Pteronotus pristinus, which both species of Noctilio, and Artibeus jamai- was scored ‘‘?’’ for this character. censis have rodlike pisiform bone. Data were Character 159: Wing digit II with long, not available for Pteronotus pristinus and ossified first phalanx (0); or first phalanx ab- Mystacina tuberculata, which were scored sent or tiny and unossified (1). The first pha- ‘‘?’’ for this character. lanx in wing digit II is relatively long (length Character 157: Metacarpal formula more than four times the diameter of the 5Ͻ4Ͻ3 (0); or 5Ͻ4ϭ3 (1); or 3ϭ4ϭ5 (2); shaft) and ossified in all extant species of or 3ϭ5Ͻ4 (3); or 3Ͻ4Ͻ5 (4). Metacarpal Pteronotus and Mormoops. A similar condi- formulae in bats describe the relative length tion is seen among the outgroups in both spe- of the third, fourth and fifth digits. For ex- cies of Noctilio, Artibeus jamaicensis, and ample, the formula 5Ͻ4Ͻ3 describes the both species of Macrotus. In contrast, the condition in which the third metacarpal is first phalanx is apparently absent in both spe- longest and the fifth is shortest. All extant cies of Mystacina and Saccopteryx bilineata. species of Pteronotus have a metacarpal for- It is possible that a minute, unossified rem- mula of 5Ͻ4Ͻ3, where metacarpals 4 and 3 nant of the first phalanx remains in these are only slightly longer than 5. In contrast, forms, but we were unable to detect it in both species of Mormoops have a metacarpal standard museum preparations. Data were formula of 5Ͻ4ϭ3, where metacarpals 4 and not available for Pteronotus pristinus, which 3 are much longer than 5. Among the out- was scored ‘‘?’’ for this character. groups, both species of Noctilio have a meta- Character 160: Wing digit III with long, carpal formula of 3ϭ5Ͻ4, where metacarpal completely ossified third phalanx (0); or with 4 is only slightly longer than 3 and 5. Both long third phalanx that is ossified at the base species of Macrotus have a metacarpal for- and unossified at the tip (1); or with long but mula of 3Ͻ4Ͻ5, where each is only slightly completely unossified third phalanx (2); or longer than the preceding, and both species with minute, unossified third phalanx (3). of Mystacina have a formula of 5Ͻ4ϭ3, with The third phalanx of wing digit III is long the relative lengths similar to Mormoops. (length more than four times the diameter of Saccopteryx bilineata has a metacarpal for- the shaft) and fully ossified in all extant spe- mula of 5Ͻ4Ͻ3, similar to Pteronotus. All cies of Pteronotus and Mormoops. A similar three metacarpals are subequal in length condition is seen among the outgroups in Ar- (3ϭ4ϭ5) in Artibeus jamaicensis. Data were tibeus jamaicensis and both species of Ma- not available for Pteronotus pristinus, which crotus. In contrast, the third phalanx is long was scored ‘‘?’’ for this character. but is ossified only at its base in both species Character 158: Wing digit I with first of Mystacina, and is completely unossified in phalanx longer than metacarpal (0); or first both species of Noctilio. The third phalanx phalanx and metacarpal subequal in length of wing digit III is a very tiny, unossified (1). Phalanges in both the manus and pes are element in Saccopteryx bilineata. Transfor- numbered from proximal to distal. The first mations in this character were ordered 0 ↔ phalanx of each wing digit is identified by 1 ↔ 2 ↔ 3toreflect our hypothesis that its articulation with the distal end of the cor- changes in phalangeal proportions and de- responding metacarpal. The first phalanx in gree of ossification occur in a progressive wing digit I is longer than the metacarpal of fashion. Data were not available for Pteron- that digit in both species of Mormoops. In otus pristinus, which was scored ‘‘?’’ for this contrast, all extant species of Pteronotus are character. characterized by a wing digit I in which the Character 161: Second phalanx of wing first phalanx and metacarpal are subequal in digit IV longer than first phalanx (0); or pha- length. These elements are similarly sube- langes subequal in length (1); or second qual in length in both species of Noctilio and phalanx shorter than first phalanx (2). The Artibeus jamaicensis, but the first phalanx is second phalanx of wing digit IV is longer markedly longer than the metacarpal in Sac- than the first phalanx of that digit in all ex- copteryx bilineata, both species of Mystaci- tant species of Pteronotus. In contrast, the 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 55

first and second phalanges of digit IV are dorsomedial edge of the ascending process of subequal in both species of Mormoops. ilium is strongly upturned and flares dorsally Among the outgroups, the second phalanx is above the level of the iliosacral articulation longer than the first phalanx in Artibeus ja- in Mormoops megalophylla, Pteronotus par- maicensis, both species of Noctilio, and Sac- nellii, and P. personatus. In these forms, the copteryx bilineata. These phalanges are sub- iliac fossa is large and well defined. In con- equal in both species of Mystacina and the trast, the dorsomedial edge of the ascending second phalanx is shorter than the first pha- process of ilium is not upturned, does not lanx in Macrotus. Transformations in this extend dorsally beyond the level of the ilios- character were ordered 0 ↔ 1 ↔ 2toreflect acral articulation, and the iliac fossa is not our hypothesis that changes in phalangeal large or well defined in Mormoops blainvillii, proportions occur in a progressive fashion. Pteronotus macleayii, P. quadridens, P. dav- Data were not available for Pteronotus pris- yi, and P. gymnonotus. This condition (low tinus, which was scored ‘‘?’’ for this char- ascending process and small, poorly defined acter. iliac fossa) is also seen in Saccopteryx bili- Character 162: Sacral vertebrae do not neata, both species of Mystacina, Artibeus contact ischium (0); or posterior sacral ver- jamaicensis, and both species of Macrotus. tebrae fused with posterior ischium (1). The An upturned, flared ascending process with principal articulation between the pelvis and a large, well-defined iliac fossa occurs in the vertebral column in bats occurs between both species of Noctilio. Data were not avail- the sacral vertebrae and the ilium. In some able for Pteronotus pristinus, which was species, a secondary articulation occurs be- scored ‘‘?’’ for this character. tween the posterior sacral vertebrae and the Character 164: Ischium with ischial tu- posterior portion of the ischium. This sec- berosity small or absent, does not project ondary articulation is separated from the dorsally beyond level of ramus (0); or with main iliosacral articulation by a gap across large ischial tuberosity that projects dorsally which there is no contact between the pelvis from posterior horizontal ramus (1). The and vertebral column. A secondary articula- dorsal ischial tuberosity is a projection that tion in which the posterior sacral vertebrae extends dorsally and/or posteromedially from are fused with the ischium occurs in all ex- the posterior end of the ischium, originating tant species of Pteronotus. In contrast, the from the ‘‘corner’’ of the pelvis at the junc- ischium does not contact the vertebral col- tion between the horizontal and vertical rami umn in either species of Mormoops. A sim- of the ischium (Vaughan, 1959, 1970a; Sim- ilar condition is seen among the outgroups in mons and Geisler, 1998). Two muscles may Saccopteryx bilineata, Mystacina robusta, originate from the ischial tuberosity: m. sem- Artibeus jamaicensis, and both species of itendinosus and m. semimembranosus, both Macrotus. Both species of Noctilio are char- of which act to extend the femur and flex the acterized by presence of contact and fusion lower leg (Vaughan, 1959, 1970b). A large, between the ischium and posterior sacral ver- dorsally projecting ischial tuberosity is ab- tebrae. Data were not available for Pterono- sent in all extant species of Pteronotus and tus pristinus and Mystacina tuberculata, Mormoops. A similar condition is seen which were scored ‘‘?’’ for this character. among the outgroups in Artibeus jamaicen- Character 163: Dorsomedial edge of as- sis, both species of Macrotus, and Saccop- cending process of ilium not upturned, does teryx bilineata. In contrast, a large, dorsally not extend dorsally beyond the level of the projecting ischial tuberosity is present in iliosacral articulation, iliac fossa not large both species of Mystacina and both species or well defined (0); or dorsomedial edge up- of Noctilio. Data were not available for Pter- turned, flares dorsally above the level of the onotus pristinus, which was scored ‘‘?’’ for iliosacral articulation, iliac fossa large and this character. well defined (1). The dorsomedial edge of the Character 165: Length of pubic spine less ascending process of the ilium in bats forms than or equal to one-third the length of the the origin for m. tensor fascia latae and m. ilium (0); or equal to or greater than one- gluteus medius (Vaughan, 1959, 1970b). The half the length of the ilium (1). The pubic 56 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 spine is an elongate projection from the an- the sides of the femoral head. All of the out- teroventral corner of the pubis (Vaughan, groups have well-developed trochanters that 1959; Walton and Walton, 1970). Presence of extend far beyond the femoral head. Data a pubic spine is a synapomorphy of bats were not available for Pteronotus pristinus, (Simmons, 1994; Simmons and Geisler, which was scored ‘‘?’’ for this character. 1998). All extant species of Pteronotus and Character 168: Femur with well-devel- Mormoops have a pubic spine whose length oped ridge present distal to lesser trochanter is equal to or greater than one-half the length (0); or with well-developed ridge present of the ilium. A similar condition is seen dorsal to lesser trochanter (1); or ridge ab- among the outgroups in Artibeus jamaicensis sent from femoral shaft (1). A high, longi- and both species of Macrotus. In contrast, the tudinally oriented ridge is present on the length of the pubic spine in Saccopteryx bil- shaft of the femur distal to the lesser tro- ineata, both species of Noctilio, and both chanter in all extant species of Mormoops species of Mystacina is less than or equal to and Pteronotus. Among the outgroups, a one-third the length of the ilium. Data were similar ridge is present in Artibeus jamaicen- not available for Pteronotus pristinus, which sis and Saccopteryx bilineata. In contrast, was scored ‘‘?’’ for this character. both species of Macrotus have a ridge pre- Character 166: Articulation between sent dorsal to the lesser trochanter. The fem- right and left pubes in male restricted to a oral shaft lacks any such ridges in both spe- small area, consists of an ossified interpubic cies of Noctilio and both species of Mysta- ligament (0); or articulation between pubes cina. Data were not available for Pteronotus broad, consists of a symphysis that is long in pristinus, which was scored ‘‘?’’ for this an anteroposterior dimension (1). The con- character. tact between the right and left pubes in male Character 169: Shaft of femur straight bats varies considerably. Pteronotus parnel- (0); or with bend that directs the distal shaft lii, P. personatus, P. davyi, P. gymnonotus, dorsally (1). The shaft of the femur is and both species of Mormoops are charac- straight in all extant species of Pteronotus terized by a broad symphysis that is long in and Mormoops. A similar state is seen in Ar- an anteroposterior dimension. A similar con- tibeus jamaicensis, both species of Mystaci- dition is seen in Artibeus jamaicensis, both na, both species of Noctilio, and Saccopteryx species of Macrotus, and both species of bilineata. In contrast, in both species of Ma- Noctilio. In contrast, the articulation between crotus the shaft of the femur is bent some- the pubes in male Saccopteryx bilineata is what so that the distal end is directed more restricted to a small area. The right and left dorsally (the equivalent of a lateral bend if pubes are not in direct contact, but are joined the femur were in the position typical of non- by a short, ossified interpubic ligament. Data volant mammals). Data were not available were not available for Pteronotus pristinus, for Pteronotus pristinus, which was scored P. macleayii, P. quadridens, and both species ‘‘?’’ for this character. of Mystacina, which were scored ‘‘?’’ for this Character 170: Fibula well developed character. and fully ossified (0); or thin and threadlike, Character 167: Greater and lesser tro- often only partly ossified (1); or absent or chanters of femur well developed, flanges ex- entirely unossified (2). The fibula in most tend well beyond sides of femoral head (0); mammals is well developed (relatively ro- or both trochanters reduced, flanges do not bust) and complete (ossified from knee to an- extend beyond sides of femoral head (1). kle). In contrast, the fibula in all extant spe- Most bats have well-developed greater and cies of Pteronotus and Mormoops is relative- lesser trochanters on the femur. In these ly much thinner, almost threadlike. This el- forms, the flanges that comprise the trochan- ement is often only partially ossified, which ters extend well beyond the sides of the head makes it appear incomplete in dried skeleton of the femur. In contrast, all extant species preparations. A similar condition is seen in of Pteronotus and Mormoops have reduced Artibeus jamaicensis, both species of Macro- trochanters that are more closely appressed tus, and both species of Noctilio. The fibula to the shaft and do not extend much beyond is either absent or entirely unossified (no 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 57

remnant present in dried skeletons) in Sac- (Schutt and Simmons, 1998). Although thin copteryx bilineata. In contrast, the fibula is and distally flexible, the calcar is apparently well developed and fully ossified in both spe- entirely calcified in all extant species of Pter- cies of Mystacina. The primitive condition onotus (Schutt and Simmons, 1998; Schutt, could not be determined a priori due to var- personal commun.). In contrast, the base of iation among outgroups. Data were not avail- the calcar is calcified but the distal portion able for Pteronotus pristinus, which was of this element is cartilaginous in both spe- scored ‘‘?’’ for this character. cies of Mormoops (Schutt and Simmons, Character 171: Tail absent (0); or tail 1998; Schutt, personal commun.). A transi- length approximately one-half the length of tional zone is clearly visible between the cal- the tibia (1); or approximately three-fourths cified and cartilaginous portions of the calcar the length of the tibia (2); or approximately in these species. Among the outgroups, both the same length as the tibia (3); or one and species of Macrotus and both species of Noc- one-half times the length of the tibia (4). tilio also have a calcar that is cartilaginous Tail length in bats can vary enormously distally with a clearly demarcated calcified among species. The tail of all extant species base (Schutt and Simmons, 1998). The calcar of Pteronotus and Mormoops is approxi- in Saccopteryx bilineata is entirely calcified mately the same length as the tibia. Among (Schutt and Simmons, 1998). Mystacina ro- the outgroups, a similar condition is seen in busta and Artibeus jamaicensis exhibit a Mystacina tuberculata. In contrast, the tail third condition in which the calcar is entirely is one and one-half times the length of the cartilaginous (Schutt and Simmons, 1998). tibia in both species of Macrotus, three- Transitions in this character were ordered 0 fourths the length of the tibia in Mystacina ↔ 1 ↔ 2toreflect our hypothesis that in- robusta and Saccopteryx bilineata, one-half creased or decreased calcification of the cal- the length of the tibia in both species of car occurs in a progressive fashion. Data Noctilio, and the tail is absent in Artibeus were not available for Pteronotus pristinus jamaicensis. Transformations in this char- and Mystacina tuberculata, which were acter were ordered 0 ↔ 1 ↔ 2 ↔ 3 ↔ 4to scored ‘‘?’’ for this character. reflect our hypothesis that reduction or in- Character 174: Calcar length less than or crease in length of the tail occurs in a pro- approximately equal to one-half the length of gressive fashion. Data were not available for the hind foot (0); or approximately the same Pteronotus pristinus, which was scored ‘‘?’’ length as hind foot (1); or one and one-half for this character. times the length of the hind foot (2); or twice Character 172: Calcaneum not expanded the length of the hind foot (3). The relatively or flattened (0); or expanded and flattened length of the calcar can be assessed by com- (1). The calcaneum of all extant species of paring its length (from base to tip) with the Pteronotus and Mormoops is a rectangular length of the hind foot (including claws). The bone that is not noticeably flattened or ex- calcar is approximately the same length as panded. A similar condition is seen among the hind foot in Pteronotus parnellii, P. the outgroups in Artibeus jamaicensis, both quadridens, and P. personatus. In contrast, species of Macrotus, both species of Mys- the calcar is approximately one and one-half tacina, and Saccopteryx bilineata. In con- times the length of the hind foot in P. ma- trast, the calcaneum in both species of Noc- cleayii, P. gymnonotus, and P. davyi. The tilio is flattened and expanded into a nearly calcar is even longer—approximately twice triangular bone when seen in anterior view. the length of the hind foot—in both species Data were not available for P.pristinus, of Mormoops. Among the outgroups, the cal- which was scored ‘‘?’’ for this character. car is less than or equal to one-half the length Character 173: Calcar entirely cartilagi- of the hind foot in both species of Mystacina nous (0); or cartilaginous distally, with and Artibeus jamaicensis, approximately one clearly demarcated calcified base (1); or cal- and one-half times the length of the hind foot car entirely calcified (2). The degree of cal- in Saccopteryx bilineata and both species of cification of the calcar varies among and Macrotus, and is twice the length of the hind within families of microchiropteran bats foot in both species of Noctilio. Data were 58 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 not available for Pteronotus pristinus, which so these taxa were scored ‘‘?’’ in our analy- was scored ‘‘?’’ for this character. sis. Character 175: Length of foot less than Character 178: M. occipitopollicalis with or equal to one-half the length of the tibia tendinous attachments to the anterior divi- (0); or greater than three-fourths the length sion of m. pectoralis profundus (0); or with of the tibia (1). The relative length of the no tendinous attachment to the anterior di- hind foot varies among bats. The length of vision of m. pectoralis profundus (1). M. oc- the foot (including claws) is less than or cipitopollicalis originates from the posterior equal to one-half the length of the tibia in all surface of the skull, runs along the leading extant species of Pteronotus and Mormoops. edge of the wing membrane, and inserts into Among the outgroups, a similar condition is structures in the region of the wrist (Strickler, seen in Saccopteryx bilineata and both spe- 1978). This muscle is variable in structure, cies of Macrotus. In contrast, the hind foot including different combinations of segments is at least three-fourths the length of the tibia of muscle, tendon, and elastic tissue, having in Artibeus jamaicensis, both species of Noc- variable insertion points, and exhibiting dif- tilio, and both species of Mystacina. Data ferent tendinous connections to other mus- were not available for Pteronotus pristinus, cles in different taxa (Strickler, 1978). M. oc- which was scored ‘‘?’’ for this character. cipitopollicalis has a tendinous attachment to Character 176: Claws without basal tal- the anterior division of m. pectoralis profun- ons (0); or with basal talons (1). The claws dus in Pteronotus parnellii, Pteronotus dav- on the thumb and hind feet of most bats, in- yi, and Mormoops megalophylla (Strickler, cluding extant species of Pteronotus and 1978). Among the outgroups, a similar ten- Mormoops, lack basal talons or spurs. A sim- dinous attachment is seen in Saccopteryx bil- ilar condition is seen among the outgroups in ineata and both species of Macrotus, but no both species of Macrotus, Artibeus jamaicen- such attachment is present in either species sis, both species of Noctilio, and Saccopteryx of Noctilio (Vaughan, 1959; Strickler, 1978). Data were not available for Pteronotus ma- bilineata. In contrast, the claws of both spe- cleayii, P. quadridens, P. personatus, P. cies of Mystacina are characterized by well- gymnonotus, P. pristinus, Mormoops blain- developed basal talons. Data were not avail- villii, Artibeus jamaicensis, and both species able for Pteronotus pristinus, which was of Mystacina, so these taxa were scored ‘‘?’’ scored ‘‘?’’ for this character. in our analysis. Character 179: M. occipitopollicalis in- POSTCRANIAL MYOLOGY sertional complex includes muscle fibers dis- tal to band of elastic tissue (0); or entirely Character 177: M. humeropatagialis pre- tendinous distal to band of elastic tissue (1). sent (0); or absent (1). Humeropatagialis is The insertional complex of m. occipitopolli- a muscle unique to bats that originates from calis is entirely tendinous distal to band of the humerus (and sometimes the ulna) and elastic tissue in Pteronotus parnellii, Pter- inserts into the plagiopatagium. It apparently onotus davyi, and Mormoops megalophylla functions to tense the plagiopatagium poste- (Strickler, 1978). Among the outgroups, a rior and lateral to the elbow (Strickler, 1978). similar arrangement is seen in both species M. humeropatagialis is present in extant spe- of Macrotus, but the insertional complex in- cies of Pteronotus and Mormoops. This mus- cludes muscle fibers distal to the band of cle is present among the outgroups in Sac- elastic tissue in Saccopteryx bilineata and copteryx bilineata and both species of Noc- both species of Noctilio (Vaughan, 1959; tilio (Strickler, 1978). In contrast, this muscle Strickler, 1978). Data were not available for is absent in both species of Macrotus Pteronotus macleayii, P. quadridens, P. per- (Vaughan, 1959). The primitive condition sonatus, P. gymnonotus, P. pristinus, Mor- could not be determined a priori due to var- moops blainvillii, Artibeus jamaicensis, and iation among the outgroups. Data were not both species of Mystacina, so these taxa were available for Pteronotus pristinus, both spe- scored ‘‘?’’ in our analysis. cies of Mystacina, and Artibeus jamaicensis, Character 180: M. spinodeltoideus orig- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 59

inates from vertebral border of scapula only total axillary border of the scapula in Pter- (0); or from vertebral border of scapula plus onotus parnellii and P. davyi (Strickler, one-half to two-thirds of the transverse scap- 1978). In contrast, the origin of this muscle ular ligament (1); or from vertebral border is restricted to just the tip of the axillary bor- of scapula plus more than three-fourths of der in M. megalophylla (Strickler, 1978). the transverse scapular ligament (1). M. spi- Among the outgroups, the origin of m. teres nodeltoideus in bats originates from the scap- major includes approximately one-fourth to ula and associated ligaments, and inserts on two-fifths of the axillary border in Artibeus the dorsal ridge of the humerus (Strickler, jamaicensis, both species of Macrotus, and 1978). Contraction of this muscle acts to flex, Saccopteryx bilineata, and it is even more abduct, and supinate the humerus (Strickler, extensiveÐ-extending from one-half to three- 1978). M. spinodeltoideus originates from fourths of the axillary borderÐ-in both spe- the vertebral border of scapula plus one-half cies of Noctilio (Vaughan, 1959; Strickler, to two-thirds of the transverse scapular lig- 1978; Hermanson and Altenbach, 1985). ament in Pteronotus parnellii and P. davyi Transformations in this character were or- (Strickler, 1978). The origin of this muscle is dered 0 ↔ 1 ↔ 2 based on our assumption even more extensive in M. megalophylla, that changes resulting in increased or de- where m. spinodeltoideus originates from the creased extent of origin occur in a progres- vertebral border of scapula plus more than sive fashion. Data were not available for three-fourths of the transverse scapular liga- Pteronotus macleayii, P. quadridens, P. per- ment (Strickler, 1978). Among the out- sonatus, P. gymnonotus, P. pristinus, Mor- groups, m. spinodeltoideus originates from moops blainvillii, and both species of Mys- the vertebral border of scapula plus one-half tacina, so these taxa were scored ‘‘?’’ in our to two-thirds of the transverse scapular lig- analysis. ament in Artibeus jamaicensis and both spe- Character 182: M. teres major inserts cies of Macrotus, but the origin is restricted into ventral ridge of humerus (0); or into to the vertebral border of the scapula in Sac- ventral base of pectoral crest (1). The inser- copteryx bilineata and both species of Noc- tion point of m. teres major onto the humerus tilio (Vaughan, 1959; Strickler, 1978; Her- varies among bats (Strickler, 1978). This manson and Altenbach, 1985). Transforma- muscle inserts into the ventral base of the tions in this character were ordered 0 ↔ 1 pectoral crest of the humerus in Pteronotus ↔ 2 based on our assumption that changes parnellii, P. davyi, and Mormoops megalo- resulting in increased or decreased extent of phylla (Strickler, 1978). A similar condition origin from the transverse scapular ligament is seen among the outgroups in Artibeus ja- occur in a progressive fashion. Data were not maicensis and both species of Macrotus available for Pteronotus macleayii, P. quad- (Vaughan, 1959; Hermanson and Altenbach, ridens, P. personatus, P. gymnonotus, P. 1985). In contrast, m. teres major inserts into pristinus, Mormoops blainvillii, and both the ventral ridge of the humerus in Saccop- species of Mystacina, so these taxa were teryx bilineata and both species of Noctilio scored ‘‘?’’ in our analysis. (Strickler, 1978). Data were not available for Character 181: M. teres major originates Pteronotus macleayii, P. quadridens, P. per- from one-half to three-fourths of the axillary sonatus, P. gymnonotus, P. pristinus, Mor- border of the scapula (0); or originates from moops blainvillii, and both species of Mys- one-fourth to two-fifths of the axillary border tacina, so these taxa were scored ‘‘?’’ in our (1); or origin restricted to the ventral tip of analysis. the axillary border (2). M. teres major in bats Character 183: M. triceps brachii caput originates from the scapula and associated medialis present (0); or absent (1). Vaughan fascia, and inserts onto the humerus and Bateman (1970) described the myology (Vaughan, 1959; Strickler, 1978). Contrac- of the forelimb of all species of Pteronotus tion of this muscle acts to flex and rotate the and Mormoops. M. triceps brachii is a tri- humerus (Vaughan, 1959; Strickler, 1978). partite muscle with three heads: caput later- The origin of m. teres major includes ap- alis, caput medialis, and caput longus. The proximately one-fourth to two-fifths of the caput medialis originates from the posterior 60 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 surface of the distal humerus and inserts into not available for P. pristinus, both species of the common tendon of the triceps; it may Mystacina, and Artibeus jamaicensis, so also be attached by connective tissue to the these taxa were scored ‘‘?’’ in our analysis. olecranon process of the ulna (Vaughan, Character 185: M. biceps brachii inser- 1959). M. triceps caput medialis is present in tional tendons not encased in sling (0); or Pteronotus, but is absent in both species of encased in sling (1). The insertional tendons Mormoops (Vaughan and Bateman, 1970). of m. biceps brachii are encased in a sling of This muscle is present in Saccopteryx bili- connective tissue that is attached to the dor- neata, both species of Noctilio, and both spe- sal surface of the humerus in both species of cies of Macrotus (Vaughan, 1959; Strickler, Mormoops (fig. 9; Vaughan and Bateman, 1978). Data were not available for P. pristin- 1970). In contrast, these tendons are not en- us, both species of Mystacina, and Artibeus cased in a sling in any extant species of Pter- jamaicensis, so these taxa were scored ‘‘?’’ onotus (Vaughan and Bateman, 1970). The in our analysis. sling is similarly lacking in Artibeus jamai- Character 184: M. coracobrachialis the censis, both species of Macrotus, and both same size or smaller than coracoid head of species of Noctilio (Vaughan, 1959; Strickler, m. biceps brachii, originates from tip of cor- 1978; Hermanson and Altenbach, 1985). acoid process (0); or approximately twice Data were not available for P. pristinus, Sac- the size of coracoid head of biceps, origi- copteryx bilineata, and both species of Mys- nates proximal to tip of coracoid process (1). tacina, so these taxa were scored ‘‘?’’ in our M. coracobrachialis and m. biceps brachii are analysis. muscles that originate from the coracoid pro- Character 186: M. brachialis present (0); cess and insert on the humerus and radius, or absent (1). M. brachialis in bats originates respectively (Vaughan, 1959). M. coracobra- on the medial portion of the anterior surface chialis acts to adduct and extend the humer- of the humerus and inserts on the radius us, while m. biceps brachii flexes and rotates (Vaughan, 1959). It acts to flex and rotate the the forearm and adducts the wing (Vaughan, radius, and may help to stabilize the wing 1959). M. coracobrachialis originates from during flight (Vaughan, 1959). M. brachialis the tip of the coracoid process and is roughly is absent in both species of Mormoops, but the same size or smaller than the coracoid this muscle is present in all species of Pter- head of the biceps in Pteronotus parnellii onotus (fig. 9; Vaughan and Bateman, 1970). and both species of Mormoops (Vaughan and M. brachialis is present in Saccopteryx bili- Bateman, 1970). In contrast, m. coracobra- neata, both species of Macrotus, and both chialis is twice the size of the coracoid head species of Noctilio (Vaughan, 1959; Strickler, of the biceps in P. macleayii, P. quadridens, 1978), suggesting that presence of m. bra- P. personatus, P. davyi, and P. gymnonotus, chialis may be the primitive condition. Data and its origin is proximal to the tip of the were not available for P. pristinus, both spe- coracoid process (Vaughan and Bateman, cies of Mystacina, and Artibeus jamaicensis, 1970). M. coracobrachialis originates at the so these taxa were scored ‘‘?’’ in our analy- tip of the coracoid process and is the same sis. size or smaller than the coracoid head of the Character 187: No sesamoid in tendon of biceps in Saccopteryx bilineata, both species m. extensor carpi radialis longus (0); or ses- of Macrotus, and both species of Noctilio amoid present (1). M. extensor carpi radialis (Vaughan, 1959; Strickler, 1978). Data were originates from the lateral epicondyle and

→ Fig. 9. Dorsal views of the arm of (A) Pteronotus personatus and (B) Mormoops megalophylla (redrawn from Vaughan and Bateman, 1970: figs. 1, 2). In each figure, the shoulder is shown at the top and the wrist (not included in this view) is to the left. Note the sling for the biceps tendons in Mormoops (a structure absent in Pteronotus) and presence of m. brachialis in Pteronotus (absent in Mormoops). M. triceps caput medialis, which is absent in Mormoops but present in Pteronotus, cannot be seen in this dorsal view. 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 61 62 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 distal shaft of the humerus, and inserts on the 1959). Data were not available for the other dorsal bases of the metacarpals of the first outgroups in our study, which were scored and second digits (Vaughan, 1959; Vaughan ‘‘?’’ for this character, as was Pteronotus and Bateman, 1970). This muscle acts to ex- pristinus. tend the metacarpals and may serve to spread Character 190: Tendon from m. extensor the distal part of the wing (Vaughan, 1959; digitorum communis inserts on second pha- Vaughan and Bateman, 1970). No sesamoid lanx of fifth digit (0); or on shaft of meta- is present in the tendon of m. extensor carpi carpal of fifth digit (1). The tendon of m. radialis longus in any extant species of Pter- extensor digitorum communis inserts on the onotus (Vaughan and Bateman, 1970). In second phalanx of the fifth digit in Pteron- contrast, a sesamoid is associated with the otus parnellii, P. davyi, P. gymnonotus, Mor- tendon of this muscle in both species of Mor- moops megalophylla, and M. blainvillii. In moops (Vaughan and Bateman, 1970). contrast, the tendon of m. extensor digitorum Among the outgroups, a sesamoid in this lo- communis inserts on the shaft of the meta- cation is absent in both species of Macrotus carpal of the fifth digit in P. macleayii, P. (Vaughan, 1959). Data were not available for quadridens, and P. personatus (Vaughan and the other outgroups in our study, which were Bateman, 1970). In both species of Macrotus scored ‘‘?’’ for this character, as was Pter- the tendon inserts on the second phalanx of onotus pristinus. the fifth digit (Vaughan, 1959). Data were Character 188: M. extensor policis brevis not available for the other outgroups in our originates from ulna (0); or from both ulna study, which were scored ‘‘?’’ for this char- and radius (1). M. extensor policis brevis is acter, as was Pteronotus pristinus. an extensor of the thumb that inserts on the Character 191: M. extensor digiti quinti metacarpophalangeal joint and second pha- proprius originates from lateral epicondyle lanx of the first digit (Vaughan, 1959; of humerus and dorsal base of ulna (0); or Vaughan and Bateman, 1970). M. extensor from lateral epicondyle only (1); or from sec- policis brevis originates from the dorsal sur- ond division of m. extensor digitorum com- face of the ulna in both species of Mormoops munis and posterodorsal base of radius (2). (Vaughan and Bateman, 1970). In contrast, M. extensor digiti quinti proprius is a digital this muscle originates from the dorsal surface extensor that inserts on the second phalanx of the ulna and the adjacent surface of the of the fifth digit of the wing. M. extensor radius in all extant species of Pteronotus digiti quinti proprius originates from the lat- (Vaughan and Bateman, 1970). In both spe- eral epicondyle of humerus and the dorsal cies of Macrotus the origin of this muscle is base of the ulna in Pteronotus macleayii, P. restricted to the dorsal surface of the ulna quadridens, P. davyi, and P. gymnonotus. In (Vaughan, 1959). Data were not available for contrast, this muscle originates from the lat- the other outgroups in our study, which were eral epicondyle of the humerus only in P. scored ‘‘?’’ for this character, as was Pter- personatus and both species of Mormoops; onotus pristinus. an ulnar origin is lacking in these taxa. A Character 189: M. extensor digitorum third condition is seen in P. parnellii, in communis originates from radius, ulna, and which m. extensor digiti quinti proprius orig- humerus (0); or from radius and ulna only inates from the second division of m. exten- (1). M. extensor digitorum communis is a sor digitorum communis and the posterodor- digital extensor that inserts onto the third, sal base of the radius (Vaughan and Bateman, fourth, and fifth digits of the wing. This mus- 1970). In both species of Macrotus the mus- cle originates from the proximal half of the cle originates from the lateral epicondyle and radius and ulna in all extant species of Pter- the dorsal base of the ulna (Vaughan, 1959). onotus and Mormoops (Vaughan and Bate- Data were not available for the other out- man, 1970). In contrast, in both species of groups in our study, which were scored ‘‘?’’ Macrotus, m. extensor digitorum communis for this character, as was Pteronotus pristin- has a more extensive origin that includes the us. proximal radius, proximal ulna, and the lat- Character 192: M. flexor digitorum pro- eral epicondyle of the humerus (Vaughan, fundus inserts via tendon onto second pha- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 63

Fig. 10. Ventral view of the hand of Mormoops megalophylla (redrawn from Vaughan and Bateman, 1970: fig. 6).

lanx of digit IV of wing (0); or does not in- dus inserts onto the second phalanx of digit sert onto digit IV (1). M. flexor digitorum V of the wing in all extant species of Pter- profundus is a digital flexor that originates onotus and Mormoops (fig. 10; Vaughan and from the distal humerus and inserts via ten- Bateman, 1970). In contrast, m. flexor digi- dons onto a variable number of wing digits torum profundus does not insert on digit V (fig. 10; Vaughan, 1959; Vaughan and Bate- in either species of Macrotus (Vaughan, man, 1970). This muscle inserts onto the sec- 1959). Data were not available for the other ond phalanx of digit IV of the wing in all outgroups in our study, which were scored extant species of Pteronotus and Mormoops ‘‘?’’ for this character, as was Pteronotus (fig. 10; Vaughan and Bateman, 1970). In pristinus. contrast, m. flexor digitorum profundus does It may be noted that the taxonomic distri- not insert on digit IV in either species of Ma- bution of states of this character is identical crotus (Vaughan, 1959). Data were not avail- to that of character 192, which also describes able for the other outgroups in our study, the insertion of m. flexor digitorum profun- which were scored ‘‘?’’ for this character, as dus. We opted to score these as separate char- was Pteronotus pristinus. acters because the taxonomic distribution of Character 193: M. flexor digitorum pro- these two features is not identical when ad- fundus inserts via tendon onto second pha- ditional taxa are considered. For example, m. lanx of digit V of wing (0); or does not insert flexor digitorum profundus inserts on digit onto digit V (1). M. flexor digitorum profun- IV, but not on digit V, in Phyllostomus has- 64 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 tatus (Vaughan and Bateman, 1970). These nellii, P. macleayi, P. gymnonotus, Mor- observations indicate that insertion on digit moops megalophylla, and M. blainvillii IV and digit V may vary independently. (Vaughan and Bateman, 1970). A similar in- Character 194: M. palmaris longus ro- sertion is seen among the outgroups in Ma- bust (0); or vestigial (1). M. palmaris longus crotus (Vaughan, 1959). In contrast, m. pal- originates on the spinous process of the me- maris longus lacks an insertion onto the dial epicondyle of the humerus and inserts metacarpal of digit III in Artibeus jamaicen- on the metacarpals of the second and/or third sis and both species of Noctilio (Vaughan digits of the wing (see characters 195 and and Bateman, 1970; Straney, 1980). Data 196; Vaughan, 1959; Vaughan and Bateman, were not available for Pteronotus pristinus, 1970). M. palmaris longus is a vestigial mus- Saccopteryx bilineata, and both species of cle that is much narrower than m. flexor dig- Mystacina, so these taxa were scored ‘‘?’’ in itorum profundus in all extant species of our analysis. As noted above, Vaughan and Pteronotus (Vaughan and Bateman, 1970). In Bateman (1970) attempted to trace the inser- contrast, this muscle is relatively robust (ap- tion of m. palmaris longus in Pteronotus per- proximately the same size as m. flexor digi- sonatus, P. quadridens, and P. davyi, but torum profundus) in both species of Mor- moops (fig. 10). A similar condition is seen found that the tendons were too small to fol- among the outgroups in both species of Ma- low. We accordingly scored these taxa ‘‘–’’ crotus, Artibeus jamaicensis, and both spe- for this character. cies of Noctilio (Vaughan, 1959; Vaughan Character 197: Digital tendon locking and Bateman, 1970). Data were not available mechanism absent from digits of hind feet for Pteronotus pristinus, Saccopteryx bili- (0); or ratchetlike lock present, consists of neata, and both species of Mystacina, so tubercles on the proximal flexor tendon and these taxa were scored ‘‘?’’ in our analysis. plicae on the adjacent tendon sheath (1); or Character 195: M. palmaris longus in- friction lock present, consists of retinaculum serts onto metacarpal of digit II (0); or does on flexor tendon sheath (2). The hind feet of not insert onto metacarpal of digit II (1). M. many bats are equipped with a digital tendon palmaris longus inserts onto the metacarpal locking mechanism (TLM) that apparently of digit II of the wing in Pteronotus parnellii, facilitates hindlimb hanging with minimal P. macleayi, P. gymnonotus, Mormoops me- muscular effort (Schaffer, 1905; Schutt, galophylla, and M. blainvillii (fig. 10; 1993; Bennett, 1993; Quinn and Baumel, Vaughan and Bateman, 1970). A similar in- 1993; Simmons and Quinn, 1994). All extant sertion is seen among the outgroups in Arti- species of Pteronotus and Mormoops lack a beus jamaicensis (Vaughan and Bateman, TLM (Simmons and Quinn, 1994; Schutt, 1970). In contrast, m. palmaris longus lacks personal commun.). Among the outgroups, an insertion onto the metacarpal of digit II in Mystacina tuberculata also lacks a TLM both species of Macrotus (Vaughan, 1959). (Simmons and Quinn, 1994). In contrast, a Data were not available for Pteronotus pris- ratchetlike TLM, consisting of tubercles on tinus, Saccopteryx bilineata, both species of the proximal flexor tendon and plicae on the Noctilio, and both species of Mystacina, so adjacent tendon sheath, is present in Saccop- these taxa were scored ‘‘?’’ in our analysis. Vaughan and Bateman (1970) attempted to teryx bilineata and both species of Noctilio trace the insertion of m. palmaris longus in (Simmons and Quinn, 1994; Schutt, personal Pteronotus personatus, P. quadridens, and P. commun.). Both species of Macrotus and Ar- davyi, but found that the tendons were too tibeus jamaicensis lack a ratchetlike TLM, small to follow. We accordingly scored these instead having a friction-lock on the digital taxa ‘‘–’’ for this character. tendons that consists of a bandlike retinacu- Character 196: M. palmaris longus in- lum in the tendon sheath (Schutt, 1993; Sim- serts onto metacarpal of digit III (0); or does mons and Quinn, 1994). Data were not avail- not insert onto metacarpal of digit III (1). M. able for Pteronotus pristinus and Mystacina palmaris longus inserts onto the metacarpal robusta, so these taxa were scored ‘‘?’’ for of digit III of the wing in Pteronotus par- this character. 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 65

REPRODUCTIVE TRACT found to be characterized by a similar uterine morphology in which the uterine horns are Character 198: Female external genitalia approximately one-half the length of the with transverse vulval opening (0); or vulval common uterine body (Hood and Smith, opening oriented anteroposteriorly (1). The 1982, 1983). Among the outgroups, Saccop- female external genitalia in bats consists of teryx bilineata and Mystacina robusta have a clitoris and vulval opening. The vulval uterine horns that are more than three-fourths opening is oriented anteroposteriorly in all the length of the common uterine body, both extant species of Pteronotus and Mormoops. species of Noctilio have uterine horns that A similar arrangement is seen among the out- are one-half the length of the common uter- groups in Artibeus jamaicensis, both species ine body, Macrotus californicus has uterine of Macrotus, and Mystacina tuberculata. In horns that are one-fourth the length of the contrast, the vulval opening is oriented trans- common uterine body, and Artibeus jamai- versely in Saccopteryx bilineata and both censis has a fully simplex uterus that lacks species of Noctilio. Data were not available free uterine horns (Hood and Smith, 1982, for Pteronotus pristinus and Mystacina ro- 1983). No data were available for Pteronotus busta, so these taxa were scored ‘‘?’’ for this personatus, P. davyi, P. gymnonotus, P. pris- character. tinus, Mormoops megalophylla, Mystacina Character 199: Clitoris of moderate size, tuberculata, and Macrotus waterhousii, so length approximately equal to width at base these taxa were scored ‘‘?’’ for this character. (0); or clitoris elongated, length approxi- Following Hood and Smith (1982, 1983), mately twice width at base (1). The clitoris we ordered this character 0 ↔ 1 ↔ 2 ↔ 3 of mormoopid bats varies in relative length. to reflect the presumed process by which a The clitoris is of moderate size (clitoris duplex uterus (such as that seen in ptero- length approximately equal to width at base) podids and some emballonurids) may be in Pteronotus parnellii, P. personatus, P. transformed into a simplex uterus by pro- quadridens, P. davyi, P. gymnonotus, and gressive fusion of the uterine horns begin- both species of Mormoops. In contrast, the ning at the cervical end of the uterus. As clitoris is elongated (clitoris length approxi- fusion progresses, the common uterine body mately twice width at base) in Pteronotus is thought to increase in length at the ex- macleayii. Among the outgroups, the clitoris pense of decreased length of the free uterine is of moderate size in Saccopteryx bilineata, horns. This process seems confirmed (at both species of Macrotus, and Artibeus ja- least in bats) by observations of an inverse maicensis, but it is elongated in both species relationship between length of the common of Noctilio. Data were not available for Pter- uterine body and length of the free uterine onotus pristinus and both species of Mysta- horns (Hood and Smith, 1982). Transfor- cina, so these taxa were scored ‘‘?’’ for this mations in this series may have worked both character. ways in at least some groups (e.g., within Character 200: External uterine horns Phyllostomidae), with reversals accom- more than three-fourths the length of the ex- plished by progressive splitting of the uter- ternal common uterine body (0); or one-half ine horns and reduction in relative size of the length of the common uterine body (1); the common uterine body (see discussion in or one-fourth the length of the common uter- Wetterer et al., 2000). ine body (2); or uterus fully simplex, uterine Character 201: Common uterine lumen horns not distinct from common uterine body relatively short, coronal lumina join well (3). Hood and Smith (1982, 1983) examined within common uterine body (0); or common morphology of the female reproductive tract uterine lumen large, coronal lumina join im- in representatives of most bat families and mediately within the common uterine body found considerable variation in the extent of (1); or common uterine lumen very large, external fusion of the uterine horns. Four coronal lumina either reduced to a tubular mormoopids were included in their study: intramural uterine cornua or entirely absent Mormoops blainvillii, Pteronotus parnellii, (2). Internal uterine fusion in bats is broadly P. macleayi, and P. quadridens. All were correlated with external uterine fusion, but 66 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 these two features are decoupled to some ex- megalophylla, Mystacina tuberculata, and tent because the degree of external fusion is Macrotus waterhousii, so these taxa were not always indicative of a similar degree of scored ‘‘?’’ for this character. fusion of the internal spaces (coronal lumina) Character 203: Oviductal mucosal folds within the common uterine body (Hood and occur throughout the oviduct (0); or folds re- Smith, 1982, 1983). In Mormoops blainvillii, stricted to extramural oviduct (1). Mormoops Pteronotus parnellii, P. macleayi, and P. blainvillii, Pteronotus parnellii, P. macleayi, quadridens, the common uterine lumen is and P. quadridens have oviductal mucosal relatively large and the coronal lumina join folds that occur throughout the oviduct, in- immediately upon their entry into the com- cluding the intramural portion that is en- mon uterine body (Hood and Smith, 1982, closed within the wall of the uterus (Hood 1983). Among the outgroups, a similar con- and Smith, 1982, 1983). This condition also dition is seen in Macrotus californicus. In occurs among the outgroups in Saccopteryx contrast, the common uterine lumen is rela- bilineata, both species of Noctilio, and Mys- tively short and the coronal lumina join well tacina robusta (Hood and Smith, 1982, within common uterine body in Saccopteryx 1983). In contrast, Macrotus californicus and bilineata, Mystacina robusta, and both spe- Artibeus jamaicensis are characterized by re- cies of Noctilio (Hood and Smith, 1982, striction of oviductal mucosal folds to the ex- 1983). Another distinct condition is seen in tramural oviduct; the intramural portion of Artibeus jamaicensis, which has a very large the oviduct is smooth and lacks any such common uterine lumen and lacks any rem- folds (Hood and Smith, 1982, 1983). No data nants of coronal lumina (Hood and Smith, were available for Pteronotus personatus, P. 1982, 1983). As with the preceding character, davyi, P. gymnonotus, P. pristinus, Mor- we followed Hood and Smith (1982, 1983) moops megalophylla, Mystacina tuberculata, in ordering this character 0 ↔ 1 ↔ 2 to re- and Macrotus waterhousii, so these taxa flect the presumed process by which a duplex were scored ‘‘?’’ for this character. uterus may be transformed into a simplex Character 204: Ovarian ligament extends uterus (or vice versa) by progressive fusion from ovary to external entry of oviduct (0); uterine horns and lumina. No data were or to lateral border of common uterine body available for Pteronotus personatus, P. dav- (1). The ovarian ligament extends from the yi, P. gymnonotus, P. pristinus, Mormoops ovary to the external entry of the oviduct megalophylla, Mystacina tuberculata, and into the uterus in Mormoops blainvillii, Pter- Macrotus waterhousii, so these taxa were onotus parnellii, P. macleayi, and P. quad- scored ‘‘?’’ for this character. ridens(Hood and Smith, 1982, 1983). This Character 202: Uterotubal junction with condition also occurs among the outgroups oviductal papilla (0); or oviductal papilla in Saccopteryx bilineata, Mystacina robusta, absent (1). Bats of many families have a both species of Noctilio, and Macrotus cali- uterotubal junction characterized by an ovi- fornicus (Hood and Smith, 1982, 1983). In ductal papilla that projects into the lumen of contrast, the ovarian ligament extends from the uterine horn (Hood and Smith, 1982, the ovary to the lateral border of the common 1983). In contrast, Mormoops blainvillii, uterine body in Artibeus jamaicensis (Hood Pteronotus parnellii, P. macleayi, and P. and Smith, 1982, 1983). No data were avail- quadridens have a simple uterotubal junction able for Pteronotus personatus, P. davyi, P. and lack oviductal papilla (Hood and Smith, gymnonotus, P. pristinus, Mormoops megal- 1982, 1983). Among the outgroups, Saccop- ophylla, Mystacina tuberculata, and Macro- teryx bilineata, both species of Noctilio, Ma- tus waterhousii, so these taxa were scored crotus californicus, and Artibeus jamaicensis ‘‘?’’ for this character. similarly have a simple uterotubal junction lacking an oviductal papilla, but an oviductal DIGESTIVE TRACT papilla is present in Mystacina robusta (Hood and Smith, 1982, 1983). No data were Character 205: Submaxillary glands of available for Pteronotus personatus, P. dav- mucous type (0); or of serous type (1); or yi, P. gymnonotus, P. pristinus, Mormoops absent (2). Submaxillary glands are salivary 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 67

glands located adjacent to the upper jaw of cous-producing pyloric glands are distributed bats (Dalquest and Warner, 1954). Two types in an extremely narrow zone immediately ad- of submaxillary glands are found in bats: (1) jacent to pyloric sphincter in Pteronotus par- mucous glands, which are nongranular, have nellii and Mormoops megalophylla (Fore- an epithelial membrane, and secrete mucous, man, 1971, 1972, 1973). Among the out- and (2) serous glands (ϭ albuminous glands), groups, a similar condition is seen in Sac- which are granular, have a layer of mesothe- copteryx bilineata and both species of lium, and secrete a watery fluid. Mucous- Noctilio (Foreman, 1971, 1972, 1973). In type submaxillary glands are found in Pter- contrast, mucous-producing pyloric glands onotus parnellii and P. davyi (Dalquest and are distributed in a much broader zone of the Warner, 1954). In contrast, Mormoops me- stomach in Macrotus waterhousii and Arti- galophylla is characterized by serous-type beus jamaicensis (Foreman, 1979). Data submaxillary glands (Dalquest and Warner, were not available for Pteronotus quadri- 1954). Among the outgroups, mucous-type dens, P. macleayii, P. personatus, P. davyi, submaxillary glands are present in Saccop- P. gymnonotus, P. pristinus, Mormoops teryx bilineata, and submaxillary glands are blainvillii, both species of Mystacina, and absent in Artibeus jamaicensis (Dalquest and Macrotus californicus, so these taxa were Warner, 1954). No data were available for scored ‘‘?’’ in our study. Pteronotus personatus, P. quadridens, P. Character 208: Pyloric sphincter asym- macleayi, P. gymnonotus, P. pristinus, Mor- metrical, valve in lesser curvature smaller moops blainvillii, both species of Noctilio, than valve in greater curvature (0); or both species of Mystacina, and both species sphincter symmetrical, valves of similar size of Macrotus, so these taxa were scored ‘‘?’’ (1). The pyloric sphincter separates the stom- for this character. ach from the duodenum (Foreman, 1971, Character 206: Cardiac vestibule of 1972, 1973, 1979). The pyloric sphincter is stomach very small or absent (0); or large asymmetrical, with the valve in the lesser (1). Foreman (1971, 1972, 1973, 1979) de- curvature of the stomach smaller than the scribed the structure and histology of the valve in the greater curvature, in Pteronotus stomach and duodenum of a wide variety of parnellii and Mormoops megalophylla (Fore- bats, and our observations of character vari- man, 1971, 1972, 1973). A similar condition ation in these structures are based entirely on is seen among the outgroups in both species his work. Foreman (1971, 1972, 1973, 1979) of Noctilio and Macrotus waterhousii (Fore- defined the cardiac vestibule as that part of man, 1971, 1972, 1973, 1979). In contrast, the stomach which is immediately adjacent the pyloric sphincter is symmetrical, with the to the esophageal opening. The cardiac ves- valves of similar size, in Saccopteryx bili- tibule is very small or absent in Pteronotus neata and Artibeus jamaicensis (Foreman, parnellii and Mormoops megalophylla (Fore- 1973, 1979). Data were not available for man, 1971, 1972, 1973). A similar condition Pteronotus quadridens, P. macleayii, P. per- is seen among the outgroups in Saccopteryx sonatus, P. davyi, P. gymnonotus, P. pristin- bilineata, both species of Noctilio, and Ma- us, Mormoops blainvillii, both species of crotus waterhousii (Foreman, 1971, 1972, Mystacina, and Macrotus californicus, so 1973, 1979). In contrast, the cardiac vesti- these taxa were scored ‘‘?’’ in our study. bule is large in Artibeus jamaicensis (Fore- Character 209: Brunner’s glands at gas- man, 1979). Data were not available for Pter- troduodenal junction present in small to onotus quadridens, P. macleayii, P. person- moderate numbers, with broad tubules and atus, P. davyi, P. gymnonotus, P. pristinus, moderate to large cells with large, spherical Mormoops blainvillii, both species of Mys- nuclei that are juxtaposed to the basement tacina, and Macrotus californicus, so these membrane (0); or glands abundant, with tu- taxa were scored ‘‘?’’ in our study. bules of moderate breadth and small cells Character 207: Mucous-producing pylo- with extremely small, spherical nuclei that ric glands distributed in extremely narrow are juxtaposed to the basement membrane zone immediately adjacent to pyloric sphinc- (0); or glands unusually abundant, with nar- ter (0); or distributed in broad zone (1). Mu- row tubules and small cells with extremely 68 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

small, flattened nuclei that are not juxta- ANALYSIS 1: EXTANT TAXA ONLY,ORDERED posed to the basement membrane (1). Brun- CHARACTERS ner’s glands occur in most bats in the gastro- duodenal junction just distal to the pyloric Our first analysis included all of the extant taxa in our study (Pteronotus pristinus was sphincter (Foreman, 1971, 1972, 1973, excluded) and used the ordered character 1979). These glands are unusually abundant transformations described above under indi- in Pteronotus parnellii and Mormoops me- vidual character descriptions. Parsimony galophylla (Foreman, 1971, 1972, 1973). analysis of these data resulted in a single Brunner’s glands in these taxa are character- most-parsimonious tree (fig. 11) of 438 steps ized by narrow tubules and small cells with (CI ϭ 0.653; CI excluding uninformative extremely small, flattened nuclei that are not characters ϭ 0.619; RI ϭ 0.750). Monophyly juxtaposed to the basement membrane (Fore- of Mormoopidae was very well supported man, 1971, 1972, 1973). Among the out- (fig. 11; tables 2, 3). Within the family, very groups, a similar condition is seen in both strong support was also found for monophyly species of Noctilio (Foreman, 1971, 1972, of Pteronotus and Mormoops. Within Pter- 1973). In contrast, in Macrotus waterhousii onotus, monophyly of a clade consisting of and Artibeus jamaicensis have Brunner’s P. davyi ϩ P. gymnonotus was similarly well glands present in small to moderate numbers, supported. The sister group of this lineage with broad tubules and moderate to large was a larger clade consisting of P. parnellii, cells with large circular nuclei that are jux- P. personatus, P. quadridens, and P. macle- taposed to the basement membrane (Fore- ayi, but monophyly of this group was only man, 1979). Saccopteryx bilineata is char- weakly supported. Within this group, P. par- acterized by an intermediate condition in nellii occupied the basal branch, with the re- which the Brunner’s glands are abundant, maining species forming a poorly supported with tubules of moderate breadth and small clade. Within the latter group, weak support cells with extremely small, spherical nuclei was found for a clade consisting of P. quad- that are juxtaposed to the basement mem- ridens ϩ P. macleayi (fig. 11; tables 2, 3). brane (Foreman, 1972, 1973). Transforma- With the exception of the P. davyi ϩ P. tions in this character were ordered 0 ↔ 1 gymnonotus clade, support values for most ↔ 2toreflect our hypothesis that transfor- groupings within Pteronotus were disap- mations in structure and abundance of Brun- pointingly low (e.g., bootstrap and jackknife ϭ ϭ ner’s glands occurred in a progressive fash- values 58Ð66%, decay values 1; tables ion. Data were not available for Pteronotus 2, 3). However, support values for alternative quadridens, P. macleayii, P. personatus, P. arrangements of Pteronotus species were davyi, P. gymnonotus, P. pristinus, Mor- considerably worse in the bootstrap and jack- moops blainvillii, both species of Mystacina, knife analyses. Other than those clades that and Macrotus californicus, so these taxa appeared in the most-parsimonious tree (fig. 11), only six clades received support in more were scored ‘‘?’’ in our study. than 5% of the bootstrap and jackknife iter- ations: P. parnellii ϩ P. personatus (boot- RESULTS strap and jackknife support values ϭ 25Ð 26%), P. davyi ϩ P. gymnonotus ϩ P. per- As noted under ‘‘Materials and Methods,’’ sonatus ϩ P. macleayii ϩ P. quadridens we conducted several separate phylogenetic (support values ϭ 23%), P. personatus ϩ P. analyses of our data to investigate the effects quadridens (support values ϭ 20%), P. par- of different assumptions and the inclusion of nellii ϩ P. davyi ϩ P. gymnonotus (support the fossil species Pteronotus pristinus. Re- values ϭ 11%), P. parnellii ϩ P. personatus sults of each of these analyses are described ϩ P. quadridens (support values ϭ 9Ð10%), individually below. Decay values, branch and P. parnellii ϩ P. macleayii ϩ P. quad- lengths, bootstrap values, and jackkinfe val- ridens (support values ϭ 5Ð6%). These re- ues for clades recovered in these analyses are sults indicate that while the hypothesis of summarized in tables 2 and 3. Pteronotus relationships shown in figure 11 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 69

Fig. 11. Tree showing the results of our Analysis 1, which used ordered characters and did not include the fossil Pteronotus pristinus. This tree shows the topology of the single most-parsimonious tree recovered (438 steps; CI ϭ 0.653; CI excluding uninformative characters ϭ 0.619; RI ϭ .750). Decay values are given above each branch; bootstrap values are below (see tables 2 and 3 for associated branch length values and results of jackknife analysis).

is not strongly supported, it nevertheless ap- moopidae and Phyllostomidae, with Nocti- pears much more likely than any of the al- lionidae and Mystacinidae falling outside this ternatives. clade. Virtually no support was found for al- Among the outgroups, very strong support ternative arrangements grouping Mormoopi- was found for monophyly of Noctilionidae dae and Noctilionidae as sister taxa (boot- and Mystacinidae. Although our taxonomic strap and jackknife support values ϭ 3%) or sampling within Phyllostomidae was admit- Noctilionidae and Phyllostomidae as sister tedly very poor, monophyly of this family taxa (support values Ͻ1%). Similarly, there (represented by Artibeus jamaicensis and two was no support for monophyly of Noctilion- species of Macrotus) was also well supported oidea as traditionally recognized (e.g., ex- in our analysis. Moderate support was found cluding Mystacinidae; support values Յ1%). for a sister-group relationship between Mor- Indeed, such relationships were contraindi- 70 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Fig. 12. Tree showing the results of our Analysis 2, which used ordered characters and included the fossil Pteronotus pristinus. This tree shows the topology of the single most-parimonious tree recovered (439 steps; CI ϭ 0.652; CI excluding uninformative characters ϭ 0.618; RI ϭ .749). Decay values are given above each branch; bootstrap values are below (see tables 2 and 3 for associated branch length values and results of jackknife analysis). Results of our Analysis 3, which included the same taxa but treated all characters as unordered, produced an identical tree topology; see text for discussion. cated by the strong support found for a Mor- character transformations described above moopidae ϩ Phyllostomidae ϩ Mystacinidae under individual character descriptions. Par- clade that excludes Noctilionidae. simony analysis of these data resulted in a single most-parsimonious tree of 439 steps ϭ ANALYSIS 2: ALL TAXA,ORDERED CHARAC- (fig. 12; CI 0.652; CI excluding uninfor- ϭ ϭ TERS mative characters 0.618; RI 0.749). The topology of the branching pattern among the Our second analysis included all of the extant taxa was identical to that found in our taxa in our study (including the fossil species Analysis 1. Pteronotus pristinus was unam- Pteronotus pristinus) and used the ordered biguously placed as the sister group of P. 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 71

TABLE 2 Bootstrap and Jacknife Supporta for Identified Clades

parnellii. As in Analysis 1, no support was 1. Within Pteronotus, decay values for the P. found for alternative arrangements placing davyi ϩ P. gymnonotus clade (which was Mormoopidae as the sister group of families very strongly supported by bootstrap and other than Phyllostomidae (bootstrap and jackknife results) were reduced 50% with the jackkinfe values Յ3% for all such arrange- inclusion of P. pristinus. Although no de- ments), and little support was found for other crease in support for monophyly of Noctilion- arrangements of taxa within Pteronotus idae, Mystacinidae, Phyllostomidae, Macro- (bootstrap and jackknife values Յ22% for all tus, and Mormoops was detected in the boot- alternative clades) strap and jackknife analyses, the decay values The only substantive differences between for each of these clades were reduced by 25Ð the results of Analysis 1 and Analysis 2 were 65% when Pteronotus pristinus was included in perceived support for several groupings, in the analysis (tables 2, 3). mostly along the backbone of the tree (tables 2, 3). Bootstrap and jackknife support for ANALYSIS 3: ALL TAXA,UNORDERED CHAR- monophyly of Mormoopidae and a Mor- ACTERS moopidae ϩ Phyllostomidae clade decreased approximately 10% when the incomplete fos- Our third analysis included all of the taxa in sil P. pristinus was included, and support for our study, but differed from the previous anal- monophyly of Pteronotus was reduced by yses in treating all character transformations as 15% (table 2). Support for these clades in de- unordered. Parsimony analysis of this data set cay analyses was even more strongly affected resulted in a single most-parsimonious tree of (table 3), with Analysis 2 decay values re- 416 steps (CI ϭ 0.688; CI excluding uninfor- duced by 65Ð85% in comparison to Analysis mative characters ϭ 0.655; RI ϭ 0.767). To- 72 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

TABLE 3 Decay Values and Branch Lengths for Identified Clades

pology of this tree (not figured) was identical ous phylogenetic arrangements by searching to that found in Analysis 2. Results of Analysis for the shortest trees compatible with those 3 differed from those of Analysis 2 only in the hypotheses. The shortest tree compatible degree of perceived support for various clades with Smith’s (1972) phylogeny (fig. 1) was (see tables 2 and 3). Decay values for most found to be 443 steps long, 4 steps longer clades were not significantly changed when the than the most-parsimonious tree (see results characters were treated as unordered (table 3); of Analysis 2 above). The shortest fully re- however, bootstrap and jackknife support for solved tree compatible with the phylogeny of some groupings were somewhat reduced. Arnold et al. (1982; fig. 2) was 449 steps Bootstrap and jackknife support for the Mys- long, 10 steps longer than the most-parsi- tacinidae ϩ Mormoopidae ϩ Phyllostomidae monious tree. Finally, the shortest tree com- clade decreased approximately 10%, and sup- patible with the trees published by Kennedy port for Mormoopidae ϩ Phyllostomidae de- et al. (1999) was 468 steps long, 29 steps creased 15% (table 2). Values for many clades longer than the most-parsimonious tree. To- within Pteronotus remained relatively constant, pology of our most-parsimonious tree is en- although support for the P. parnellii ϩ P. pris- tirely congruent with Van Den Bussche and tinus ϩ P. personatus ϩ P. macleayii ϩ P. Hoofer’s (2000) phylogeny of noctilionoid quadridens clade, as well as the P. macleayii relationships. ϩ P. quadridens clade, was reduced by 6Ð7% (table 2). DISCUSSION AND CONCLUSIONS INTRAFAMILIAL RELATIONSHIPS OF ALTERNATIVE HYPOTHESES MORMOOPIDS In addition to the analyses described The results of our phylogenetic analyses above, we also attempted to evaluate previ- were remarkably congruent with one another. 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 73

Monophyly of Mormoopidae, Mormoops, were undoubtedly excluded during the re- and Pteronotus was confirmed in all analyses sampling process in many replicates. regardless of taxonomic sampling or char- Comparison of the results of Analyses 2 acter ordering. Within Pteronotus, several and 3 indicates that ordering character trans- clades were recovered: (1) P. davyi ϩ P. formations did not bias tree topology in our gymnonotus; (2) P. macleayi ϩ P. quadri- study, as identical trees were recovered in dens; (3) P. personatus ϩ P. macleayi ϩ P. both analyses. The only effect of ordering quadridens; (4) P. parnellii ϩ P. pristinus ϩ transformations in this study was to slightly P. personatus ϩ P. macleayi ϩ P. quadri- increase perceived levels of support at most dens; and (5) P. parnellii ϩ P. pristinus. nodes, an expected outcome given the extra These results support monophyly of all of the steps required by ordered characters under subgenera recognized by Smith (1972), and some circumstances. Overall length of the additionally indicate that the subgenera Pter- shortest tree was increased by 23 steps (5%) onotus and Chilonycteris are sister taxa. when we included ordered characters. Be- Our experiments with including and ex- cause the transformations implied by our or- cluding the fossil taxon Pteronotus pristinus dering of character states are logical from an from otherwise identical analyses (Analyses evolutionary and constructional standpoint 1 and 2) indicated that this species, known (see justifications under individual character only from cranial material, can be placed un- descriptions above), it is our view that these ambiguously relative to extant species of data are phylogenetically informative. We mormoopids. The levels of support for its therefore regard the results (including the placement as the sister group of P. parnellii support values) from Analyses 1 and 2 as equalled values associated with several best representing the outcome of our study. clades of extant taxa (tables 2, 3). It is clear Comparisons of our tree with the shortest from these results that the information pre- trees compatible with phylogenies of other authors indicated that our hypothesis is only served in fossils of P. pristinus was phylo- slightly better supported by morphological genetically informative despite the fact that data than that of Smith (1972). This is not we could score this species for less than 16% surprising, given the general congruence of of the characters in our study (table 1). How- these hypotheses (figs. 1, 11, 12). Smith’s ever, there is a downside to including incom- (1972) hypothesis was based on morphology pletely known taxa in a phylogenetic study. (indeed, we considered many of the same Inclusion of P. pristinus in the analysis did characters), and our tree differs from his only not affect tree topology, but it did negatively in placement of two taxa, Pteronotus par- influence the decay, bootstrap, and jackknife nellii and P. personatus. Greater incongru- values at many nodes, particularly those de- ence exists between our tree and that of Ar- fining the family Mormoopidae and the ge- nold et al. (1982). Arnold et al.’s (1982) phy- nus Pteronotus. Decay values were most logeny, which was based on morphology, al- strongly influenced. These effects are likely lozymes, immunology, and chromosome due to the large amounts of missing data and data, differs in the position of P. parnellii correspondingly small numbers of informa- within Pteronotus and, more significantly, tive characters associated with P. pristinus. placement of Noctilionidae as the sister Placement of P. pristinus in many alternative group of Mormoopidae (figs. 2, 11, 12). As positions in the tree (e.g., as the sister group noted above under ‘‘Results’’, we found no to all extant species of Pteronotus, to all ex- support for the latter hypothesis in our data tant mormoopids, or to Mystacinidae or Noc- set. tilionidae) requires only a few additional The greatest incongruence between our re- steps. As a result, inclusion of this taxon in sults and those of other authors involves trees our analyses had the effect of reducing decay based on cytochrome-b sequence data re- values at many nodes in these regions of the cently published by Kennedy et al. (1999). tree. Bootstrap and jackknife analyses were As noted in the ‘‘Introduction’’, those au- also affected by the relatively small numbers thors found that Mormoopidae was paraphy- of informative characters, some of which letic, with Noctilio placed as the sister group 74 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 of Pteronotus (i.e., Noctilionidae nesting TABLE 4 within Mormoopidae). This novel hypothesis Proposed Classification of Mormoopid Bats was refuted in our analyses, which found strong support for monophyly of Mormoop- idae under all analysis parameters. Given the sparse sampling of Mormoopidae in the Ken- nedy et al. (1999) study (which included only one species of Mormoops and two species of Pteronotus) and the heavy character weight- ing employed by those authors (see ‘‘Intro- duction’’), we believe that it is too early to assess the real level of incongruence between the phylogenetic signal preserved in cyto- chrome-b and morphological data. Interestingly, there does not appear to be any significant incongruence between our morphological data and the 12S rRNA—t- RNAValϪ 16S rRNA data of Van Den Bus- sche and Hoofer (2000). Eight species from their study were included in our analyses species Pteronotus pristinus (which was de- (Saccopteryx bilineata, Noctilio albiventris, scribed after Smith’s revision) in the subge- Noctilio leporinus, Mystacina tuberculata, nus Phyllodia to reflect its close relationships Artibeus jamaicensis, Macrotus waterhousii, to Pteronotus parnellii. The classification we Mormoops megalophylla, and Pteronotus propose is given in table 4. parnellii), and the recovered phylogenetic re- Below we provide taxonomic diagnoses of lationships among these taxa were identical all mormoopid clades based on DELTRAN in our trees and theirs. This degree of con- and ACTRAN optimizations of our data gruence between morphological and molec- (with characters ordered) on our shortest tree ular data is remarkable, and suggests that fu- (including Pteronotus pristinus; fig. 12). Un- ture combined-data analyses will be very ambiguously derived conditions that repre- profitable. In our opinion, Kennedy et al.’s sent synapomorphies are shown in italic type. (1999) finding of mormoopid paraphyly is Synapomorphies that are unique and unrev- unlikely to be substantiated in future analy- ersed in the context of our data set are ad- ses. At least three separate data sets support ditionally highlighted in boldface type. Ple- monophyly of Mormoopidae: morphology siomorphic states, derived conditions that ap- (Arnold et al., 1982; this study), karyotype ply at higher taxonomic levels (e.g., derived data (Arnold et al., 1982), and 12S rRNA— features shared between mormoopids and tRNAValϪ 16S rRNA gene sequences (Van phyllostomids), and states whose optimiza- Den Bussche and Hoofer, 2000). In this con- tions are ambiguous are listed in standard text, mormoopid monophyly clearly repre- (Roman) type. Many of the latter features sents the best-supported working hypothesis. may represent synapomorphies of the taxa under which they are listed, but their status CLASSIFICATION AND TAXONOMIC DIAGNOSES remains unclear in the absence of additional OF MORMOOPID CLADES polarity information. In cases where missing data make the level of transformation of a Results of our study indicate that the clas- character uncertain, we list each derived state sification of Mormoopidae developed by (putative synapomorphy) under the smallest Smith (1972) is entirely appropriate from a group in which it can be demonstrated to oc- phylogenetic standpoint. All of the higher- cur and note the other levels at which the level taxa recognized in Smith’s classifica- transformation might apply. tion (i.e., families, genera, subgenera) are ap- The taxonomic diagnoses presented here parently monophyletic. The only alteration should be interpreted as hypotheses subject that we suggest is placement of the extinct to further testing. These hypotheses are ob- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 75

viously directly dependent on tree topology crown height of P3 subequal to that of P4; and taxonomic sampling. If future studies in- three lower premolars (p1, p3, and p4) pre- dicate phylogenetic relationships incongruent sent; three upper molars (M1, M2, M3) pre- with topology of our shortest tree, then these sent; three lower molars (m1, m2, m3) pre- diagnoses will require modification. Even if sent; cristid obliqua on m1 and m2 extends the topology remains stable, it is possible from hypoconid to middle of posterior wall some features interpreted as unique synapo- of protocristid; inferior colliculi partially ex- morphies in the context of the present anal- posed on dorsal surface of brain; first 5Ð8 ysis may have evolved convergently in tracheal rings enlarged to form tracheal ex- groups not sampled in this study (e.g., in oth- pansion posterior to larynx; basihyal with er phyllostomid genera). Conversely, addi- entoglossal process present; ceratohyal re- tional polarity information may in the future duced to half the length of epihyal; m. my- make it possible to identify other unique, un- lohyoideus inserts on basihyal and basihyal reversed synapomorphies from among those raphe only; m. mandibulo-hyoideus absent; characters whose polarity is unclear at pre- m. stylohyoideus absent or reduced to a few sent. In the mean time, the diagnoses pre- fibers fused with m. stylopharyngeus, deep sented below offer working hypotheses that to the digastric muscles; m. geniohyoideus may be used for identification and other pur- originates by long tendon from the mandible; poses. paired m. geniohyoideus muscles partially or completely fused across midline; midline hy- FAMILY MORMOOPIDAE oid strap muscles with m. genioglossus and m. hyoglossus attached directly to basihyal EMENDED DIAGNOSIS: Size small to medi- via fleshy fibers; m. styloglossus with one um (forearm length ϭ 35Ð66 mm, condylo- belly; m. ceratohyoideus insertion included basal length ϭ 12Ð22 mm; 3Ð28 g); premax- ceratohyal but does not include stylohyal; m. illa fused to maxilla ; palatal branch of pre- sternothyroideus originates from clavicle; m. maxilla well developed, right and left pre- omohyoideus originates from scapula; m. cri- maxillae fused across midline of palate; pair copharyngeus consists of a single large slip of incisive foramina present; hard palate ex- of muscle; m. cricothyroideus complex con- tends posteriorly into interorbital region; na- sists of two separate muscles, m. cricothy- sals concave upward in lateral view; nasal roideus anterior and m. cricothyroideus foramina small or absent; postorbital process posterior; lateral circumvallate papillae pre- absent; epitympanic recess deep and con- sent; basketlike medial-posterior mechanical stricted in area; fossa for m. stapedius indis- papillae present; pharyngeal region of tongue tinct; fenestra cochleae (ϭ fenestra rotun- totally bare of papillae; internal labial papil- dum) enlarged, maximum diameter Ͼ25% of lae absent; cheek pouches absent; lower lip the external width of the first half-turn of the with multiple dermal papillae; one or more cochlea; pars cochlearis of petrosal loosely transverse flaps present below lower lip; an- attached to basisphenoid via ligaments and/ terior edge of labio-nasal region does not or thin splints of bone; cochlea greatly en- protrude anteriorly beyond edge of lower lip; larged; tympanic annulus semivertical, lies at one dermal projection present lateral to each angle of 75Ð90Њ relative to basicranial plane; nostril; pinnae funnel-shaped; distal tip of angular process of lower jaw projects above tragus rounded; secondary fold present on the level of the occlusal plane of cheekteeth; tragus; interramal vibrissae present; interra- coronoid process of lower jaw does not ex- mal tubercle well developed; superciliary vi- tend dorsally above level of condyloid pro- brissae absent; vibrissae on face posterior cess; two pairs of upper incisors (I1 and I2) and lateral to narial pad arranged in two present; inner upper incisors (I1) bilobed; roughly parallel vertical columns; two vibris- height of outer upper incisors (I2) less than sae present in anteromedial column; facial one-half that of inner incisors (I1); two pairs stripes absent; dorsal stripes absent; under- of lower incisors (i1 and i2) present; inner hair and overhair clearly differentiated; lower incisors (i1) trilobed; two upper pre- scales on dorsal hairs divergent or appressed, molars (P3 and P4) present, (P1) absent; not divaricate; wings folded by flexing all 76 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 phalanges in digits III, IV, and V anteriorly occipitopollicalis with a tendinous attach- toward the underside of the wing; antebra- ment to the anterior division of m. pectoralis chial wing sac absent; uropatagium broad; profundus; m. occipitopollicalis insertional both uropatagium and plagiopatagium at- complex entirely tendinous distal to band of tach to calcar; C6 with enlarged ventral pro- elastic tissue; m. teres major inserts into ven- cesses; 12 or more thoracic vertebrae present; tral ridge of humerus; m. extensor digitorum pair of parallel ridges present on ventral sur- communis originates from radius and ulna face of T4; last two thoracic and first two only, no humeral origin; m. flexor digitorum lumbar vertebrae fused; L3 with pair of par- profundus inserts onto second phalanges of allel ridges on ventral surface; anterior face digits I, III, IV, and V; m. palmaris longus of manubrium of sternum small, not defined inserts onto metacarpal of digits II and III; by elevated ridges; distal tip of ventral pro- digital tendon locking mechanism absent cess of manubrium laterally compressed and from digits of hind feet; female external gen- keel-like; length of manubrium less than italia with vulval opening oriented antero- width; mesosternum without large keel; sec- posteriorly; external uterine horns one-half ond rib contacts sternum at manubriumÐme- length of common uterine body; common sosternum joint; mesosternum articulates uterine body large, coronal lumina join im- with six costal cartilages posterior to second mediately within the common uterine body; rib; xiphisternum with prominent median oviductal papilla absent; oviductal mucosal keel; clavicle articulates with or lies in con- folds occur throughout oviduct; ovarian lig- tact with coracoid process of scapula; pit for ament extends from ovary to external entry attachment of clavicular ligament present of oviduct; cardiac vestibule of stomach very on scapula; tip of acromion process with tri- small or absent; mucous-producing pyloric angular anterior projection; suprascapular glands distributed in an extremely narrow process present; dorsal articular facet con- zone immediately adjacent to pyloric sphinc- sists of an oval facet on the anteromedial rim ter; pyloric sphincter asymmetrical, valve in of the glenoid fossa, faces dorsolaterally; an- lesser curvature smaller than valve in greater terior portion of axillary border of scapula curvature; Brunner’s glands at gastroduode- flattened; head of humerus oval or elliptical nal junction unusually abundant, with narrow in medial view (not round); trochiter extends tubules and small cells with extremely small, proximally just to the level of the proximal flattened nuclei that are not juxtaposed to the edge of the humeral head; height of distal basement membrane. spinous process of humerus greater than height of trochlea; metacarpal of digit V of Genus Mormoops wing shorter than metacarpals of digits III and IV; first phalanx in wing digit II long, EMENDED DIAGNOSIS: Size small to medi- length more than four times shaft diameter; um (forearm length ϭ 43Ð61 mm; condylo- third phalanx of wing digit III long and com- basal length ϭ 12Ð14 mm; 6Ð20 g); rostrum pletely ossified; ischial tuberosity small or strongly upturned, angle between long axis absent; length of public spine equal to or of anterior half of zygomatic arch and oc- ;greater than one-half length of ilium; pubic clusal surface of molar toothrow 28Ð34؇ symphysis in male relatively long; greater maximum rostral breadth less than or equal and lesser trochanters of femur reduced, to length of maxillary toothrow; rostral flanges do not extend beyond sides of fem- length less than one-half of total length of oral head; well-developed ridge present on skull; infraorbital foramen not enlarged, di- femoral shaft distal to lesser trochanter; distal ameter less than one-eighth the height of ros- femur straight; fibula thin and threadlike, of- trum; infraorbital foramen located above ten only partly ossified; tail length approxi- posterior half of P4; anterior rim of orbit mately the same as length of tibia; calcaneum terminates above M1; zygomatic breadth not expanded or flattened; calcar at least part- greater than mastoid breadth; parietals in- ly ossified; length of hind foot less than or flated; cochlea phanerocochlear; ectotympan- equal to one-half tibia length; claws without ic bulla extends medially across two-thirds or basal talons; m. humeropatagialis present; m. more of cochlea; longitudinal furrows in ba- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 77 sisphenoid wide and shallow; basioccipital cess without a posterolateral projection; ax- not constricted between cochlea; lingual cin- illary border of scapula curved in dorsoven- gulae absent on upper incisors (I1 and I2); tral plane, concave ventrally; humerus with diastema present between outer upper incisor distal articular facets in line with shaft, not (I2) and canine; outer lower incisors (i2) tri- displaced laterally; central surface of capit- lobed; diastema present between upper P3 ulum of humerus reduced, height less than and P4; lower p3 double-rooted; crown that of trochlea; distal spinous process of length of lower p3 and p4 subequal; no cus- humerus located directly adjacent to troch- pule present on premetacrista on M1 and M2; lea, no notch present between process and lower first and second molars (m1 and m2) trochlea; olecranon process of ulna well de- nyctalodont; horny papillae on tongue of dif- veloped, extends beyond proximal end of ra- ferent sizes, largest near midline; distal tips dius; pisiform dumbbell-shaped, enlarged at of medial-posterior mechanical papillae di- both ends; metacarpal formula 5 Ͻ 4 ϭ 3; rected toward pharyngeal region; lower lip of wing digit I with first phalanx longer than moderate size, dorsoventral height less than metacarpal; first and second phalanges of height of nostrils; lower lip with horseshoe- wing digit IV subequal in length; sacral ver- shaped dermal pad; two transverse flaps pre- tebrae not fused to ischium posteriorly; cal- sent below lower lip; nostrils tubelike, open- car cartilaginous distally, with clearly de- ings face slightly laterally and are not adja- marcated calcified base; calcar length ap- cent; narial pad with narrow, papillated ver- proximately twice length of hind foot; m. tical ridge between nostrils; anterodorsal triceps caput medialis absent; m. coraco- edge of narial pad unadorned, without ridge brachialis smaller than coracoid head of m. of skin or noseleaf; one dermal papilla pre- biceps brachii, originates from tip of cora- sent on dorsal surface of each nostril; der- coid process of scapula; m. biceps brachii mal projection lateral to nostril triangular and insertional tendons encased in sling of con- pointed distally, lateral side not continuous nective tissue; m. brachialis absent; sesa- with lip; prominent dermal tubercle present moid present in tendon of m. extensor carpi on midline of muzzle posterodorsal to nos- radialis longus; m. extensor policis brevis trils; prominent interauricular band present; originates from ulna; tendon from m. exten- distal pinna squared, pinna height relatively sor digitorum communis inserts on second uniform across the entire width of the pin- phalanx of digit V; m. extensor digiti quinti na; lateral edge of distal pinna smooth; tra- proprius originates from lateral epicondyle gus with distinctly thickened area at base; of humerus only; m. palmaris longus robust; secondary fold equal to or larger than body clitoris of moderate size. of tragus; two interramal vibrissae present; genal vibrissae absent; dorsal and ventral fur Mormoops blainvillii with two or more bands of color; dorsal fur long, 8Ð10 mm; dorsal hairs with dentate EMENDED DIAGNOSIS: Size medium-small coronal scales; dorsal hair scales divergent; (forearm length ϭ 43Ð50 mm; condylobasal plagiopatagium attaches to the side of the length ϭ 12Ð14 mm; 6Ð11 g); dorsal fur bi- body, dorsal fur exposed on back between colored; ventral fur bicolored; pair of parallel right and left membranes; calcar not bound ridges present on ventral surface of third to tibia; seventh cervical vertebra fused to lumbar vertebra; ribs with narrow anterior first thoracic vertebra; 13 thoracic vertebrae laminae; ribs with narrow posterior laminae; present; neural spines absent from all thorac- infraspinous fossa of scapula wide, maxi- ic vertebrae; pair of parallel ridges absent mum width greater than 50% of maximum from ventral surface of fourth and fifth lum- length; intermediate infraspinous facet and bar vertebra; angle between axis of ventral posterolateral facet subequal in width; distal process of manubrium and body of manubri- spinous process of humerus with ventral pro- um approximately 90Њ; anterior and posterior jection that extends perpendicular to shaft; laminae present on ribs; tip of coracoid pro- ascending process of ilium without flared or cess not flared, approximately the same upturned dorsomedial edge upturned, iliac width as coracoid shaft; tip of acromion pro- fossa small and poorly defined. 78 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

No information on structure of the vom- present on upper incisors (I1 and I2); diaste- eronasal complex, hyoid apparatus, shoulder ma absent between upper P3 and P4; lower myology, or submaxillary glands is available p3 single-rooted; crown length of lower p3 for this species. less than or equal to one-half crown length of p4; horny papillae on tongue of uniform Mormoops megalophylla size; medial-posterior mechanical papillae basin-shaped with no clear inclination; low- EMENDED DIAGNOSIS: Size medium-large er lip large and platelike, dorsoventral (forearm length ϭ 49Ð61 mm; condylobasal height equal to or greater than height of length ϭ 12Ð22 mm; 12Ð20 g); vomeronasal nostrils; lower lip without horseshoe-shaped epithelial tube rudimentary, lacking a neu- dermal pad; one transverse flap present be- roepithelial medial lining (may apply at level low lower lip; nostrils open directly adjacent of Mormoops); vomeronasal cartilage curved to each other and face anteriorly; anterodor- in cross section, ‘‘J’’, ‘‘C’’, ‘‘U’’,or‘‘O’’ sal edge of narial pad with papillated ridge shaped (may apply at level of Mormoops); of skin present; no dermal papillae on dorsal nasopalatine duct present (may apply at level surface of nostril; interauricular band absent; of Mormoops); accessory olfactory bulb ab- distal pinna lancolate, tapers to blunt point; sent (may apply at level of Mormoops); m. tragus thin, of uniform thickness throughout; sternohyoideus originates from entire surface secondary fold smaller than body of tragus; of manubrium, no clavicular origin (may ap- one interramal vibrissa present; one genal vi- ply at level of Mormoops or Mormoopidae); brissa present on each cheek; dorsal fur dorsal fur with four bands of color; ventral short, 4Ð7mm; dorsal hairs with appressed fur polymorphic, either bicolored or tricol- scales; seventh cervical vertebra not fused to ored; pair of parallel ridges absent from ven- first thoracic vertebra; 12 thoracic vertebrae tral surface of third lumbar vertebra; ribs present; neural spines present on first four with wide anterior laminae; ribs with wide thoracic vertebrae; pair of parallel ridges posterior laminae; infraspinous fossa of scap- present on ventral surface of third, fourth, ular narrow, maximum width less than or and fifth lumbar vertebrae; angle between equal to 50% of maximum length; interme- axis of ventral process of manubrium and diate infraspinous facet narrower than pos- body of manubrium greater than 90Њ; ribs terolateral facet; distal spinous process of hu- with wide posterior laminae; tip or coracoid merus without ventral projection; ascending process distinctly flared; infraspinous fossa process of ilium with dorsomedial edge up- of scapula wide, maximum width greater turned, flares dorsally above the level of the than 50% of maximum length; intermediate iliosacral articulation, iliac fossa large and and posterolateral infraspinous facets sube- well defined; m. spinodeltoideus originates qual (reversed in P. personatus); axillary from vertebral border of scapula plus more border of scapula flat, level within dorsoven- than three-fourths of the transverse scapular tral plane; humerus with distal articular fac- ligament (may apply at level of Mormoops); ets displaced laterally from line of shaft; cen- m. teres major origin restricted to tip of ax- tral surface of capitulum of humerus not re- illary border of scapula (may apply at level duced, height equal to that of trochlea; distal of Mormoops); submaxillary glands of se- spinous process of humerus separated from rous type (may apply at level of Mormoops). trochlea by a deep notch; distal spinous pro- cess of humerus without ventral projection; Genus Pteronotus olecranon process of ulna reduced, does not EMENDED DIAGNOSIS: Size small to medi- extend beyond the proximal end of radius; um (forearm length ϭ 35Ð66 mm; condylo- pisiform rodlike, of uniform width through- basal length ϭ 12Ð22 mm; 3Ð28 g); rostrum out; metacarpal formula 5 Ͻ 4 Ͻ 3; first pha- slightly upturned, angle between long axis of lanx of wing digit I longer than metacarpal; anterior half of zygomatic arch and occlusal second phalanx of wing digit IV longer than surface of molar toothrow 13Ð26Њ; anterior first phalanx; posterior sacral vertebrae rim of orbit terminates above posterior half fused to posterior ischium; calcar entirely of M2; parietals not inflated; lingual cingulae calcified; m. spinodeltoideus originates from 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 79

vertebral border of scapula and one-half to fossa small and poorly defined; calcar length two-thirds of transverse scapular ligament; approximately one and one-half times length m. teres major originates from one-fourth to of hind foot; m. coracobrachialis approxi- two-fifths of axillary border of scapula; m. mately twice the size of coracoid head of m. triceps caput medialis present; m. biceps bra- biceps brachii, originates proximal to tip of chii insertional tendons not encased in sling; coracoid process of scapula; tendon from m. m. brachialis present; no sesamoid present in extensor digitorum communis inserts on sec- tendon of m. extensor carpi radialis longus; ond phalanx of digit V; m. extensor digiti m. extensor policis brevis originates from quinti proprius originates from lateral epi- both ulna and radius; m. palmaris longus condyle of humerus and dorsal base of ulna; vestigial; submaxillary glands of mucous clitoris of moderate size. type. Pteronotus davyi

Subgenus Pteronotus EMENDED DIAGNOSIS: Size medium-small (forearm length ϭ 40Ð50 mm; condylobasal EMENDED DIAGNOSIS: Size small to medi- ϭ um (forearm length ϭ 40Ð56 mm; condylo- length 13Ð16 mm; 5Ð11 g); m. sternohy- basal length ϭ 13Ð17 mm; 5Ð18 g); maxi- oideus originates from entire surface of ma- mum rostral breadth greater than length of nubrium, no clavicular origin (may apply at maxillary toothrow; rostral length less than level of subgenus Pteronotus or Mormoopi- one-half of total length of skull; infraorbital dae); ventral fur bicolored. foramen not enlarged, diameter less than one-eighth height of rostrum; infraorbital fo- Pteronotus gymnonotus ramen located above posterior half of M1; EMENDED DIAGNOSIS: Size medium-large zygomatic breadth greater than mastoid (forearm length ϭ 49Ð56 mm; condylobasal breadth; cochlea phanerocochlear; ectotym- length ϭ 15Ð17 mm; 11Ð18 g); ventral fur panic bulla extends medially across two- unicolored. thirds or more of cochlea; longitudinal fur- rows in basisphenoid wide and shallow; ba- Unnamed Clade 1 (Phyllodia ϩ Chilonyc- sioccipital not constricted between cochlea; teris) diastema present between outer upper incisor DIAGNOSIS: Size small to medium (forearm (I2) and canine; outer lower incisors (i2) tri- ϭ ϭ lobed; no cuspule present on premetacrista length 35Ð66 mm; condylobasal length on M1 and M2; lower first and second mo- 12Ð22 mm; 3Ð28 g); maximum rostral lars (m1 and m2) myotodont; narial pad flat, breadth less than or equal to length of max- with no notch or ridge between nostrils; der- illary toothrow; vomeronasal cartilage bar- mal projection lateral to nostril triangular and shaped in cross section; dorsal fur bicolored pointed distally, lateral side not continuous or tricolored; ventral fur bicolored; plagio- with lip; no dermal tubercle on midline of patagium attaches to the side of the body, muzzle posterodorsal to nostrils; lateral edge dorsal fur exposed on back between right and of distal pinna smooth; dorsal fur unico- left membranes; caudal half of calcar bound lored; midshaft scales on dorsal hair fila- to tibia, calcar restricted to position parallel ments all entire coronal; plagiopatagium at- to tibia; tip of acromion process without a taches at mid-dorsal line, right and left posterolateral projection; length of calcar membranes meet over spine and cover dor- and hindfoot subequal (reversed in P. macle- ayi). sal fur; calcar not bound to tibia; neural spines decrease in height from first thoracic through fourth thoracic vertebra; neural Subgenus Phyllodia spines present on T8ÐT12; ribs with narrow EMENDED DIAGNOSIS: Size medium (fore- anterior laminae; distal acromion process arm length ϭ 48Ð66 mm; condylobasal with triangular posterolateral projection; as- length ϭ 16Ð22 mm; 10Ð28 g); rostral length cending process of ilium without flared or less than one-half of total length of skull; in- upturned dorsomedial edge upturned, iliac fraorbital foramen located above anterior half 80 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 of M1; zygomatic breadth greater than mas- midline of muzzle posterodorsal to nostrils; toid breadth; longitudinal furrows in basi- lateral edge of distal pinna smooth; dorsal sphenoid narrow and deep; no cuspule pre- fur polymorphic, either bicolored or tricol- sent on premetacrista on M1 and M2; lower ored (may apply at level of Phyllodia); dor- first and second molars (m1 and m2) myoto- sal hairs with alternating entire coronal and dont. hastate coronal scales; height of neural Because one member of this clade (Pter- spines subequal on first four thoracic ver- onotus pristinus) is a fossil species known tebrae (may apply at level of Phyllodia); only from cranial material, we have chosen neural spines present on T8ÐT12; ribs with not to include postcranial and soft-tissue wide anterior laminae (may apply at level of characters in the diagnosis of Phyllodia. Phyllodia); ascending process of ilium with However, numerous derived states in these dorsomedial edge upturned, flares dorsally character systems are found in Pteronotus above the level of the iliosacral articulation, parnellii (see below), and many of these may iliac fossa large and well defined; m. cora- actually apply at the level of Phyllodia. cobrachialis smaller than coracoid head of m. biceps brachii, originates from tip of cora- Pteronotus parnellii coid process of scapula; tendon from m. ex- tensor digitorum communis inserts on second EMENDED DIAGNOSIS: Size medium (fore- phalanx of digit V; m. extensor digiti quinti arm length ϭ 35Ð66 mm; condylobasal proprius originates from second division of length ϭ 16Ð22 mm; 10Ð28 g); infraorbital m. extensor digitorum communis and pos- foramen not enlarged, diameter less than terodorsal base of radius (may apply at level one-eighth height of rostrum; cochlea cryp- of Phyllodia); clitoris of moderate size tocochlear (may apply to all Phyllodia); ec- totympanic bulla extends medially across no Pteronotus pristinus (extinct) more than one-third of cochlea (may apply at level of Phyllodia); basioccipital markedly EMENDED DIAGNOSIS: Size medium (great- constricted between cochlea (may apply at est length of skull 17Ð18 mm [condylobasal level of Phyllodia); diastema present be- length not available]); infraorbital foramen tween outer upper incisor (I2) and canine enlarged, diameter one-quarter to one-third (C); outer lower incisors (i2) bilobed; vom- height of rostrum; diastema absent between eronasal epithelial well developed, neuro- outer upper incisor (I2) and canine (C); outer epithelium medial lining present (may apply lower incisors (i2) trilobed. at level of Phyllodia); nasopalatine duct ab- No information on postcranial skeletal sent; accessory olfactory bulb present (may morphology or soft-tissue characters is avail- apply at level of Phyllodia); ceratohyal re- able for this species. duced to approximately one-half length of epihyal; m. mylohyoideus occurs as an un- Subgenus Chilonycteris divided sheet of muscle with an anterior apo- EMENDED DIAGNOSIS: Size small to medi- neurosis (may apply at level of Phyllodia, um (forearm length ϭ 35Ð49 mm; condylo- Phyllodia ϩ Chilonycteris, genus Pterono- basal length ϭ 13Ð16 mm; 3Ð11 g); infra- tus, or Mormoopidae); m. sternohyoideus or- orbital foramen not enlarged, diameter less igin includes clavicle; m. styloglossus with than one-eighth the height of rostrum; zygo- two bellies; m. sphincter coli profundus ab- matic breadth less than mastoid breadth; sent (may apply at level of Phyllodia, Phyl- cochlea phanerocochlear; ectotympanic bulla lodia ϩ Chilonycteris, genus Pteronotus, or extends medially across two-thirds or more Mormoopidae); narial pad with V-shaped of cochlea; longitudinal furrows in basisphe- notch present between nostrils (may apply at noid wide and shallow; basioccipital not con- level of Phyllodia); dermal projection lateral stricted between cochlea; diastema present to nostrils rounded distally, lateral side of between outer upper incisor (I2) and canine; projection continuous with lip (may apply at outer lower incisors (i2) trilobed; lower first level of Phyllodia or Phyllodia ϩ Chilonyc- and second molars (m1 and m2) nyctalodont; teris); prominent dermal tubercle present on lateral edge of distal pinna serrated; dorsal 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 81

fur either bicolored or tricolored; dorsal ascending process of ilium without flared or hairs with denticulate coronal scales; neural upturned dorsomedial edge, iliac fossa small spines decrease in height from first thoracic and poorly defined; m. extensor digiti quinti through fourth thoracic vertebra; ribs with proprius originates from lateral epicondyle of narrow anterior laminae; m. coracobrachialis humerus and dorsal base of ulna. approximately twice the size of coracoid head of m. biceps brachii, originates proxi- Pteronotus macleayii mal to tip of coracoid process of scapula; tendon from m. extensor digitorum commu- EMENDED DIAGNOSIS: Size small (forearm ϭ ϭ nis inserts on shaft of metacarpal of digit V. length 41Ð46 mm; condylobasal length 14Ð16 mm; 4Ð8 g); narial pad with V- Pteronotus personatus shaped notch present between nostrils; neu- ral spines present on T8ÐT12; calcar length EMENDED DIAGNOSIS: Size medium-small approximately one and one-half times length (forearm length ϭ 40Ð49 mm; condylobasal of hind foot; clitoris elongated. length ϭ 13Ð16 mm; 5Ð11 g); rostral length less than one-half of total length of skull; in- Pteronotus quadridens fraorbital foramen located above anterior half of M2; cuspule present on premeta- EMENDED DIAGNOSIS: Size small (forearm crista on M1 and M2; vomeronasal epithe- length ϭ 35Ð40 mm; condylobasal length ϭ lial tube absent (may apply at level of Chi- 12Ð14 mm; 3Ð6 g); narial pad flat, with no lonycteris); nasopalatine duct present; acces- notch or ridge between nostrils; neural spines sory olfactory bulb absent; ceratohyal unred- absent on T8ÐT12; length of calcar and hind uced, approximately equal in length to foot subequal; clitoris of moderate size. epihyal (may apply at level of Chilonycteris); narial pad flat, with no notch or ridge be- MONOPHYLY AND DIAGNOSES OF tween nostrils; dermal projection lateral to NOCTILIONOID FAMILIES nostrils rounded distally, lateral side of pro- In addition to the diagnoses of mormoopid jection continuous with lip; prominent der- taxa presented above, our study also has im- mal tubercle present on midline of muzzle plications for the diagnoses of Phyllostomi- posterodorsal to nostrils; dorsal fur bicol- dae, Noctilionidae, and Mystacinidae, all of ored; neural spines absent on T8ÐT12; inter- which appear monophyletic in our study. As mediate infraspinous facet narrower than noted previously, our sampling within Phyl- posterolateral facet; ascending process of il- lostomidae was poor given the diversity of ium with dorsomedial edge upturned, flares that family (3 species sampled out of more dorsally above level of iliosacral articulation, than 140), but our sample of Noctilionidae iliac fossa large and well defined; m. extensor and Mystacinidae was complete (2 out of 2 digiti quinti proprius originates from lateral species in each family). Lack of data for epicondyle of humerus only. many characters in Mystacina robusta and M. tuberculata (73% and 62% complete in Unnamed Clade 2 (P. macleayi ϩ P. quad- the context of our character set; see table 1) ridens) and absence of additional outgroups to Noc- EMENDED DIAGNOSIS: Size small (forearm tilionoidea (e.g., vespertilionids, nataloids, length ϭ 35Ð46 mm; condylobasal length ϭ molossoids) rendered polarity of many char- 13Ð16 mm; 3Ð8 g); rostral length equal to acters uncertain at the base of our tree. In or greater than one-half of total length of recognition of these multiple sources of un- skull; infraorbital foramen located above an- certainty, we have not attempted to produce terior half of M1; no cuspule present on pre- complete diagnoses of Phyllostomidae, Mys- metacrista on M1 and M2; dermal projection tacinidae, and Noctilionidae. Instead, we pro- lateral to nostril triangular and pointed dis- vide here lists of synapomorphies of these tally, lateral side not continuous with lip; no families identified in our study. Features that dermal tubercle on midline of muzzle pos- are unique and unreversed in the context of terodorsal to nostrils; dorsal fur tricolored; our data set and tree are shown in bold. 82 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Unambiguous synapomorphies supporting ula faces dorsally, consists of a large, flat monophyly of Phyllostomidae include: (1) facet clearly separated from the glenoid epitympanic recess shallow and broad; (2) fossa; (8) distal spinous fossa of humerus lo- diastema absent between outer upper incisor cated directly adjacent to trochlea, no notch (I2) and canine; (3) inner lower incisor (i1) present between these structures; (9) calcar bilobed; (4) outer lower incisor (i2) bilobed; length equal to approximately one-half hind (5) vomeronasal epithelial tube well devel- foot length; and (10) claws of thumb and oped, neuroepithelium medial lining present; hindfoot with basal talons. Several other (6) m. cricopharyngeus consists of 3 or derived conditions that we could score in more slips; (7) narial pad with wide, only one species of Mystacina (but which smooth-sided vertical ridge between nos- may occur in both taxa) would represent ad- trils; (8) noseleaf present; (9) ventral pro- ditional synapomorphies of Mystacinidae if cesses of C6 not enlarged, similar in width they can be shown to occur in both species: to those of C2ÐC5; (10) xiphisternum with- (11) vomeronasal cartilage absent; (12) m. out median keel; (11) second costal cartilage mandibulo-hyoideus present; (13) origin articulates with manubrium only (no contact of m. sternohyoideus restricted to medial with mesosternum); (12) tip of acromion pro- manubrium; (14) insertion of m. ceratoh- cess without triangular anterior projection; yoideus does not include ceratohyal; (15) m. (13) suprascapular process absent; (14) fric- sternothyroideus originates from lateral tion tendon locking mechanism present on manubrium (no clavicular origin); (16) m. digits of hind feet; (15) length of external omohyoideus absent; (17) anterior face of uterine horns less than or equal to one- manubrium broad, defined by elevated ridg- fourth length of common uterine body; es; (18) distal tip of ventral process of ma- (16) oviductal mucosal folds restricted to nubrium blunt and rounded (not keel- extramural oviduct; (17) mucous-produc- like); (19) ribs with no posterior laminae; ing pyloric glands distributed in a broad (20) calcar entirely cartilaginous; and (21) zone in stomach; and (18) Brunner’s uterotubal junction with oviductal papilla glands at gastroduodenal junction present present. in small to moderate numbers, with broad Unambiguous synapomorphies supporting tubules and moderate to large cells with monophyly of Noctilionidae include: (1) in- large spherical nuclei that are juxtaposed cisive foramina absent; (2) pair of large to the basement membrane. Other derived nasal foramina present in posterior nasal conditions that were incompletely sampled in region below forehead; (3) maximum ros- our study might represent additional syna- tral breadth greater than length of maxillary pomorphies of Phyllostomidae if future stud- toothrow; (4) cochlea cryptocochlear; (5) ies demonstrate that they apply at the base of cristid obliqua on lower first and second the phyllostomid tree. These include: (19) ac- molars (m1 and m2) extends from hypo- cessory olfactory bulb present; (20) m. hu- conid to metaconid; (6) basketlike medial- meropatagialis absent; and (21) submaxil- posterior mechanical papillae absent; (7) lary glands absent. cheek pouches present; (8) two rows of Unambiguous synapomorphies supporting transverse flaps present below lower lip; (9) monophyly of Mystacinidae include: (1) in- pinna funnel-shaped; (10) distal tip of tra- fraorbital foramen enlarged, diameter one- gus bifurcate; (11) two vibrissae present in quarter to one-half height of rostrum; (2) in- anteromedial column (adjacent to narial pad); fraorbital foramen located above posterior (12) ventral fur unicolored; (13) single pale half of M1; (3); anterior rim of orbit termi- middorsal stripe often present; (14) over- nates above posterior half of M2; (4) uro- hair and underhair clearly differentiated in patagium rudimentary; (5) coracoid pro- dorsal pelage; (15) scales on dorsal hairs cess reduced to a small triangular process; exhibit continuous range of variation from (6) distal acromion process with elongate entire coronal to repand coronal; (16) projection that extends anteriorly and me- wings folded by flexing the distal phalan- dially to fuse with anteromedial rim of ges of digits III and IV anteriorly, proxi- scapula; (7) dorsal articular facet on scap- mal phalanges not folded; (17) 11 thoracic 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 83

vertebrae present; (18) third lumbar verte- and Noctilionidae) in our shortest trees (figs. bra without pair of ridges on ventral surface; 11, 12), and support was found for this (19) xiphisternum without prominent median grouping in all analyses (tables 2, 3). Boot- keel; (20) second costal cartilage articulates strap and jackknife support for the alternative with manubrium only (no contact with me- grouping (Mormoopidae ϩ Noctilionidae) sosternum); (21) ribs with wide anterior lam- was much lower in all analyses (bootstrap inae; (22) distal acromion process with tri- and decay values Յ3). No support was found angular posterolateral projection; (23) dorsal for a Phyllostomidae ϩ Noctilionidae clade articular facet on scapula absent; (24) head in any analysis (bootstrap and decay values of humerus oval or elliptical, not round; (25) Յ1). Ͻ 4; (26) first Echoing the results of Van Den Bussche 5 ؍ metacarpal formula 3 phalanx and metacarpal of wing digit I sub- and Hoofer’s (2000) study of mitochondrial equal in length; (27) posterior sacral verte- gene sequences, results of our analyses in- brae fused with posterior ischium; (28) as- dicated that Mystacinidae is the sister group cending process of ilium with dorsomedial of the Mormoopidae ϩ Phyllostomidae edge upturned, flares dorsally above the level clade. These three families formed a clade of the iliosacral articulation, iliac fossa large (excluding Noctilionidae) in our shortest and well defined; (29) tail length approxi- trees (figs. 11, 12), and strong support for mately equal to one-half of tibia length; (30) was found for this grouping in all analyses calcaneum expanded and flattened; (31) (tables 2, 3). calcar length approximately twice length of As with the family diagnoses presented hind foot; (32) m. occipitopollicalis with no above, we provide lists below of unambigu- tendinous attachment to anterior division ous synapomorphies of the larger noctilion- of m. pectoralis profundus; (33) m. teres oid clades identified in our study. Features major originates from one-half to three- that are unique and unreversed in the context fourths of axillary border of scapula; and of our data set and tree are shown in bold. (34) clitoris elongated, length approximately Our data cannot provide a diagnosis for Noc- twice width at base. Two derived conditions tilionoidea because our study did not include that we could score in only one species of enough outgroups (e.g., nataloids, molos- Noctilio would represent additional synapo- soids, or vespertilionoids). morphies of Noctilionidae if they can be Unambiguous synapomorphies supporting shown to occur in both species: (35) cera- monophyly of a Phyllostomidae ϩ Mor- tohyal reduced to a tiny cartilaginous el- moopidae clade include: (1) tympanic an- ement; and (36) right and left m. genio- nulus semivertical, lies at angle of 75Ð90؇ hyoideus not fused across midline. to basicranial plane; (2) vomeronasal epi- thelial tube present; (3) anterior edge of la- INTERFAMILIAL RELATIONSHIPS bio-nasal region does not protrude ante- riorly beyond edge of lower lip, muzzle ap- Phylogenetic relationships among nocti- pears blunt; (4) narial pad flat, with no lionoid families have been the source of con- notch or ridge between nostrils; (5) genial siderable debate in recent years. As noted in vibrissae reduced to one on each cheek or the ‘‘Introduction,’’ previous phylogenetic absent entirely; (6) wings folded by flexing analyses have placed Mormoopidae either all phalanges in digits III, IV, and V an- with Noctilionidae (e.g., Patton and Baker, teriorly toward the underside of the wing; 1978; Arnold et al., 1982; Simmons, 1998; (7) third phalanx of wing digit III long and Simmons and Geisler, 1998; Kennedy et al., fully ossified; (8) pubic spine length equal 1999) or as the sister group of Phyllostomi- to or greater than one-half length of ilium; dae (Van Valen, 1979; Novacek, 1991; (9) m. occipitopollicalis insertional com- Kirsch et al., 1998; Van Den Bussche and plex entirely tendinous distal to band of Hoofer, 2000). Results of our study support elastic tissue (this feature could not be the latter hypothesis. The phyllostomid and scored in either species of Mystacina, so it mormoopid taxa in our study together may apply at the next higher level); (10) m. formed a clade (excluding both Mystacinidae spinodeltoideus originates from vertebral 84 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 border of scapula plus one-half to two-thirds complete sampling of many soft-tissue char- of the transverse scapular ligament (this fea- acter systems), as well as data analysis in the ture could not be scored in either species of context of more outgroups. New data sets Mystacina, so it may apply at the next higher should also be explored, particularly DNA level); (11) m. teres major inserts into ven- sequence data. Some of this work is now un- tral base of pectoral crest (this feature der way (Van Den Bussche, personal com- could not be scored in either species of Mys- mun.), and such studies should provide ad- tacina, so it may apply at the next higher ditional tests of our phylogeny in the future. level); and (12) common uterine lumen With a relatively well-resolved phylogeny large or very large, coronal lumina either of Mormoopidae, it will be possible to ad- join immediately within the common uter- dress a number of questions concerning the ine body, are reduced to tubular intra- evolutionary and ecological history of this mural cornua, or are absent entirely. group. Detailed analyses of morphological Unambiguous synapomorphies of a Mys- evolution, historical biogeography, and the tacinidae ϩ Mormoopidae ϩ Phyllostomidae evolution of echolocation call structure and clade include: (1) anterior rim of orbit ter- foraging strategies—topics beyond the scope minates at or posterior to the anterior half of of the present study—can in the future be M2; (2) ectotympanic bulla extends medially conducted in a phylogenetic context. Intrigu- across two-thirds or more of the cochlea; (3) ing areas for future research include the evo- upper P3 present; (4) horny papillae on lution of mormoopid wing morphology, tongue of different sizes, largest near mid- which is characterized by phylogenetically line; (5) lateral vallate papillae present; (6) structured variation in membrane attachment distal tips of medial-posterior mechanical pa- sites, osteology, and myology, evolution of pillae directed toward pharyngeal region; (7) mormoopid craniodental morphology and superciliary vibrissae absent; (8) dorsal fur diet, patterns of change in brain structure, bicolored; (9) dorsal fur 8Ð10 mm in length; evolution of high-duty-cycle CF echoloca- (10) pair of ridges present on ventral surface tion, and biogeographic history of mormoop- of L4 and L5; (11) third phalanx of wing ids in the Caribbean. It is our hope that the digit III long and at least partially ossified; present study will facilitate studies such as (12) ratchet-like digital tendon locking these and thereby contribute to a better un- mechanism absent for digits of foot; and derstanding of the biology and evolution of (13) female external genitalia with vulval opening oriented anteroposteriorly. mormoopid bats. Our phylogeny may also have interesting implications for understanding the biogeo- DIRECTIONS FOR FUTURE RESEARCH graphic history of the major noctilionoid The results of our phylogenetic analyses clades. Noctilionidae, Phyllostomidae, and allowed us to test monophyly of the family Mormoopidae are all entirely Neotropical in Mormoopidae, to test previous hypotheses of distribution. In contrast, Mystacinidae is an interrelationships among mormoopid species, Old World family with Recent species re- and to evaluate the affinities of the fossil stricted to New Zealand, and Miocene fossil Pteronotus pristinus to extant species. Based representatives in Australia (Hand et al., on character optimizations, we were also able 1998; Kirsch et al., 1998). Kirsch et al. to provide revised diagnoses for Mormoopi- (1998) proposed a complex historical bio- dae and all superspecific clades (e.g., genera geographic hypothesis to explain the geo- and subgenera) within the family. However, graphic and temporal distribution of nocti- our results were not entirely satisfactory. lionoid taxa, but this hypothesis was devel- Levels of support for several clades within oped under the assumption that Mystacinidae Pteronotus were relatively weak, and prob- was the sister group to a Noctilionidae ϩ lems with polarizing many characters made Mormoopidae ϩ Phyllostomidae clade. Our their interpretation difficult. Addressing these results, which agree with those of Van Den issues will require further analyses including Bussche and Hoofer (2000) in placing Mys- additional morphological data (e.g., more tacinidae within the Neotropical radiation, 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 85 suggest that Kirsch et al.’s (1998) biogeo- REFERENCES graphic hypothesis should be reevaluated. Recognizing that all phylogenies are them- Adams, J. K. selves hypotheses that require continued test- 1989. Pteronotus davyi. Mamm. Species 346: 1Ð5. ing, the present study provides an important Albuja, V. L. starting point for more comprehensive phy- 1982. Murcie«lagos del Ecuador. Quito: Es- logenetic analyses of bats using a dense-sam- cuela Polite«cnica Nacional. pling approach. Future work on noctilionoid Allen, J. A. relationships will require more comprehen- 1904. New bats from tropical America, with sive sampling within Phyllostomidae as well notes on Otopterus. Bull. Am. Mus. as inclusion of additional putative outgroups. Nat. Hist. 6: 247Ð248. Further testing of monophyly of Noctiliono- Altenbach, J. S., and J. W. Hermanson idea as defined here (i.e., including Mysta- 1987. Bat flight muscle function and the sca- cinidae) will minimally require addition of pulo-humeral lock. In M. B. Fenton, P. Racey, and J. M. Rayner (eds.), Recent representatives of other yangochiropteran advances in the study of bats: 100Ð118. families to the data set (e.g., Myzopodidae, Cambridge: Cambridge Univ. Press. Thyropteridae, Furipteridae, Natalidae, Mo- Andersen, K. lossidae, Vespertilionidae). Expansion of tax- 1912. Catalogue of the Chiroptera in the col- onomic sampling will necessitate integration lection of the British Museum, 2nd ed. of many new morphological characters rele- Vol. 1: Megachiroptera. London: Brit- vant to relationships among these taxa. Some ish Museum. [Reprinted by Johnson such characters have been described previ- Reprint Corporation, New York] ously (e.g., in Simmons and Geisler [1998] Arnold, M. L., R. L. Honeycutt, R. J. Baker, V. and Wetterer et al. [2000]), but many new M. Sarich, and J. K. Jones 1982. Resolving a phylogeny with multiple characters will doubtless prove to be relevant data sets: a systematic study of phyl- in the context of improved species-level sam- lostomoid bats. Occas. Pap. Mus. Texas pling. Finally, we believe that combined-data Tech Univ. 77: 1Ð15. analyses including morphological and molec- Arroyo-Cabrales, J. ular data will ultimately provide the most ro- 1992. Sinopsis de los murcie«lagos fo«siles de bust tests of current hypotheses of bat rela- Me«xico. Rev. Soc. Mex. Paleontol. 5: tionships, including those presented in this 1Ð14. contribution. Arroyo-Cabrales, J., and T. Alvarez 1990. Restoso«seos de murcie«lagos proceden- tes de las excavaciones de las Grutas ACKNOWLEDGMENTS de Loltu«n. Me«xico, Distrito Federal: In- stituto Nacional de Antropolog«õa e His- Special thanks go to Tom Griffiths, Matt toria, 103 pp. Rockman, Bill Schutt, and Andrea Wetterer, Baker, R. J., and J. W. Bickham who were kind enough to share unpublished 1980. Karyotypic evolution in bats: evidence data with us. Two anonymous reviewers read of extensive and conservative chromo- earlier verstions of the manuscript, and we somal evolution in closely related taxa. thank them for their comments. We thank Syst. Zool. 29: 239Ð253. Michael Carleton, Charles Handley, and Lin- Barbour, R. W., and W. H. Davis da Gordon (USNM), Mark Engstrom 1969. Bats of America. Lexington: Univ. (ROM), Jim Patton (MVZ), and Michel Kentucky Press, 286 pp. Tranier (MNHN) for access to specimens in Baron, G., H. Stephan, and H. D. Frahm their care. We also thank Chris Vogel for pre- 1996. Comparative neurobiology in Chirop- paring skeletons for use in this project, Pa- tera, vol. 2. Basel: Birkha¬user Verlag. Bateman, G. C., and T. A. Vaughan tricia Wynne for the excellent illustrations in 1974. Nightly activities of mormoopid bats. J. figures 3Ð10, Eric Stiner for illustrations 11Ð Mammal. 55: 45Ð65. 12, and Pat Brunauer for many hours of li- Benedict, F. A. brary work. This study was supported by 1957. Hair structure as a generic character in NSF Research Grants DEB-9106868 and bats. Univ. California Publ. Zool. 59: DEB-9873663 to NBS. 285Ð548. 86 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Bennett, M. B. Czaplewski, N. J. 1993. Structural modifications involved in the 1997. Chiroptera. In R. F. Kay, R. H. Madden, fore- and hind limb grip of some flying R. L. Cifelli, and J. J. Flynn (eds.). Ver- foxes (Chiroptera: Pteropodidae). J. tebrate paleontology in the Neotropics: Zool. (London) 229: 237Ð248. the Miocene fauna of La Venta, Colum- Bhatnager, K. P. and F. C. Kallen bia: 410Ð431. Washington D C: Smith- 1974. Cribiform plate of ethmoid, olfactory sonian Institution Press, 592 pp. bulb and olfactory acuity in forty spe- Czaplewski, N. J., and C. Cartelle cies of bats. J. Morphol. 142: 71Ð90. 1998. Pleistocene bats from cave deposits in Birney, E. C., J. B. Bowles, R. M. Timm, and S. Bahia, Brazil. J. Mammal. 79: 784Ð L. Williams 803. 1974. Mammalian distributional records in da Fonseca, G. A. B., G. Herrmann, Y. L. R. Lei- Yucata«n and Quintana Roo, with com- te, R. A. Mittermeier, A. B. Rylands, and J. L. ments on reproduction, structure, and Patton status of peninsular populations. Occas. 1996. Lista anotada dos mam«õferos do Brasil. Pap. Bell Mus. Nat. Hist. Univ. Min- Occas. Pap. Conserv. Biol. 4: 1Ð38. nesota 13: 1Ð25. Dalquest, W. W., and E. R. Hall Bonaccorso, F. J., A. Arends, M. Genoud, D. Can- 1949. Five bats new to the known fauna of toni, and T. Morton Mexico. J. Mammal. 30: 424Ð427. 1992. Thermal ecology of mustached and Dalquest, W. W., and H. J. Werner ghost-faced bats (Mormoopidae) in 1954. Histological aspects of the faces of Venezuela. J. Mammal. 73: 365Ð378. North American bats. J. Mammal. 35: Bowles, J. B., J. P. Cope, and E. A. Cope 147Ð160. 1979. Biological studies of selected Peruvian Dobson, G. E. bats of Tingo Mar«õa, Departmento de 1875. Conspectus of the suborders, families, Hua«nuco. Trans. Kansas Acad. Sci. 82: and genera of Chiroptera arranged ac- 1Ð10. cording to their natural affinities. Ann. Bremer, K. Mag. Nat. Hist. (4th ser.) 95: 345Ð357. 1988. The limits of amino acid sequence data Donoghue, M. J., J. A. Doyle, J. Gauthier, A. G. in angiosperm phylogenetic reconstruc- Kluge, and T. Rowe tions. Evolution 42: 795Ð803. 1989. The importance of fossils in phylogeny Brosset, A., and P. Charles-Dominique reconstruction. Annu. Rev. Ecol. Syst. 1990. The bats from French Guiana: a taxo- 20: 431Ð446. nomic, faunistic and ecological ap- Dwyer, P. D. proach. Mammalia 54: 509Ð559. 1960. New Zealand bats. Tuatara 8: 61Ð71. Brosset, A., P. Charles-Dominique, A. Cockle, J. Easterla, D. A. F. Cosson, and D. Masson 1973. Ecology of the 18 species of Chiroptera 1996. Bat communities and deforestation in at Big Bend National Park, Texas. Part French Guiana. Can. J. Zool. 74: 1974Ð II. Northwest Missouri State Univ. 1982. Stud. 34: 54Ð165. Brown, J. C. Easterla, D. A., and J. O. Whitaker 1971. The description of mammals. 1. The 1972. Food habits of some bats from Big external characters of the head. Mam- Bend National Park, Texas. J. Mammal. mal Rev. 1: 151Ð168. 53: 887Ð890. Carter, D. C., and P. G. Dolan Eisenberg, J. F. 1978. Catalogue of type specimens of Neo- 1989. Mammals of the Neotropics. The north- tropical bats in selected european mu- ern Neotropics. Vol. 1. Panama, Co- seums. Spec. Publ. Museum Texas lombia, Venezuela, Guyana, Suriname, Tech Univ. 15: 1Ð136. French Guiana. Chicago: Univ. of Chi- Ceballos, G., and C. Galindo L. cago Press. 1984. Mam«õferos silvestres de la cuenca de Emmons, L. H. Me«xico. Me«xico: Editorial Limusa, S. 1997. Neotropical rainforest mammals: a field A.: 300 pp. guide, 2nd ed. Chicago: Univ. Chicago Corbet, G. B., and J. E. Hill Press. 1980. A world list of mammalian species. Eshelman, R. E., and G. S. Morgan London: British Museum (Natural His- 1985. Tobagan recent mammals, fossil verte- tory), 226 pp. brates, and their zoogeographical im- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 87

plications. Nat. Geog. Soc. Res. Rep. Goodwin, G. G., and A. M. Greenhall 21: 137Ð143. 1961. A review of the bats of Trinidad and Fenton, M. B. Tobago: descriptions, rabies infection, 1980. Adaptiveness and ecology of echolo- and ecology. Bull. Am. Mus. Nat. Hist. cation in terrestrial (aerial) systems. In 122: 187Ð302. R. -G. Busnel and J. F. Fish (eds.), An- Goodwin, R. E. imal sonar systems, NATO Adv. Study 1970. The ecology of Jamaican bats. J. Mam- Inst. Ser. A 28: 427Ð448. mal. 51: 571Ð579. 1982. Echolocation, insect hearing, and the Graham, G. L., and L. J. Barkley feeding ecology of insectivorous bats. 1984. Noteworthy records of bats from Peru. In T. H. Kunz (ed.), Ecology of bats: J. Mammal. 65: 709Ð711. 261Ð285. New York: Plenum Press. Gray, H. 1984. Echolocation: implications for the ecol- 1913. Anatomy, descriptive and surgical. ogy and evolution of bats. Q. Rev. Biol. New York: Lea and Febiger. 59: 33Ð53. Gray, J. E. 1994. Echolocation: its impact on the behav- 1838. A revision of the gerna of bats (Veper- ior and ecology of bats. E« coscience 1: tilionidae), and the description of some 21Ð30. new genera and species. Mag. Zool. 1995. Natural history and biosonar signals. In Bot. 2: 483Ð505. A. N. Popper and R. R. Fay (eds.), 1839. Descriptions of some Mammalia dis- Hearing in bats: 37Ð86. New York: covered in Cuba by W. S. MacLeay, Springer-. Esq. Ann. Nat. Hist. 4: 1Ð7. Foreman, G. L. 1843. [Letter addressed to the Curator.] Proc. 1971. Gastric morphology in selected mor- Zool. Soc. London 1843: 50. moopid and glossophagine bats as re- Greenbaum, I. F., and C. J. Phillips lated to systematic problems. Trans. Il- 1974. Comparative anatomy and general his- linois State Acad. Sci. 64: 273Ð282. tology of tongues of long-nosed bats 1972. Comparative morphological and histo- (Leptonycteris sanborni and L. nivalis) chemical studies of stomachs of select- with reference to infestations of oral ed American Univ. Kansas Sci. Bull. mites. J. Mammal. 55: 489Ð504. 49: 591Ð729. Griffiths, T. A. 1973. Studies of gastric morphology in North 1978. Modification of m. cricothyroideus and American Chiroptera (Emballonuridae, the larynx in the Mormoopidae, with Noctilionidae, Phyllostomatidae). J. reference to amplification of high-fre- Mammal. 54: 909Ð923. quency pulses. J. Mammal. 59: 724Ð 1979. Alimentary tract. In R. J. Baker, J. K. 730. 1982. Systematics of the New World nectar- Jones, and D. C. Carter (eds.), Biology feeding bats (Mammalia, Phyllostomi- of bats of the New World family Phyl- dae), based on the morphology of the lostomatidae. Part III. Spec. Publ. Mus. hyoid and lingual regions. Am. Mus. Texas Tech Univ. 16: 205Ð227. Novitates 2742: 45 pp. Gauthier, J., A. G. Kluge, and T. Rowe 1983. Comparative laryngeal anatomy of the 1988. Amniote phylogeny and the importance Big brown bat, Eptesicus fuscus, and of fossils. Cladistics 4: 105Ð209. the Mustached bat, Pteronotus parnel- Gill, T. lii. Mammalia 47: 377Ð394. 1901. The bat genus Pteronotus renamed 1994. Phylogenetic systematics of Slit-faced Dermonotus. Proc. Biol. Soc. Washing- bats (Chiroptera, Nycteridae), based on ton 14: 177. hyoid and other morphology. Am. Mus. Gimenez, E. A. Novitates 3090: 17 pp. 1993. Morfologia lingual comparada, filogen- Griffiths, T. A., and A. L. Smith ia e evoluc¸a˜o dos ha«bitos alimentares 1991. Systematics of emballonuroid bats na superfam«õla Phyllostomoidea (Mam- (Chiroptera: Emballonuridae and Rhin- malia: Chiroptera). M.S. Thesis, Insti- opomatidae), based on hyoid morphol- tuto de Biocieˆncias, Botucatu, UNESP. ogy. Bull. Am. Mus. Nat. Hist. 206: Gimenez, E. A., H. Ferrarezzi, and V. A. Taddei 62Ð83. 1996. Lingual morphology and cladistic anal- Griffiths, T. A., A. Truckenbrod, and P. J. Spon- ysis of the New World nectar-feeding holtz bats (Chiroptera: Phyllostomidae). J. 1992. Systematics of megadermatid bats Comp. Biol. 1: 41Ð64. (Chiroptera, Megadermatidae), based 88 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

on hyoid morphology. Am. Mus. Nov- Honeycutt, R. L. itates 3041: 21 pp. 1981. Molecular evolution in New World Gundlach, J. leaf-nosed bats of the family Phyllos- 1840. Beschreibung von vier auf Cuba ge- tomidae with comments on the super- fangnen Fledermauden. Arch. Natur- family Noctilionoidea. Ph.D. diss, Tex- gesch. 6: 356Ð358. as Tech Univ., Lubbock. Habersetzer, J., and G. Storch Hood, C. S., and J. D. Smith 1992. Cochlea size in extant Chiroptera and 1982. Cladistical analysis of female reproduc- Middle Eocene Microchiroptera from tive histomorphology in Phyllostoma- Messel. Naturwissenchaften 79: 462Ð toid bats. Syst. Zool. 31: 241Ð251. 466. 1983. Histomorphology of the female repro- Hall, E. R. ductive tract in phyllostomoid bats. Oc- cas. Pap. Mus. Texas Tech Univ. 86: 37 1981. Mammals of North America, 2nd ed. pp. New York: Wiley. Howell, D. J., and D. Burch Hand, S. J., P. Murray, D. Megirian, M. Archer, 1974. Food habits of some Costa Rican bats. and H. Godthelp Rev. Biol. Trop. 21: 281Ð294. 1998. Mystacinid bats (Microchiroptera) from Jones, J. K. the Australian Tertiary. J. Paleontol. 72: 1989. Distribution and systematics of bats of 538Ð545. the Lesser Antilles. In C. A. Woods Handley, C. O. (ed.), Biogeography of the West Indes: 1959. A revision of American bats of the gen- past, present, and future: 645Ð660. era Euderma and Plecotus. Proc. U.S. Gainesville, FL: Sandhill Crane Press. Nat. Mus. 110: 95Ð246. Kalko, K. V. and H.-U. Schnitzler 1966. Checklist of the mammals of Panama. 1998. How echolocating bats approach and P.p. In R. L. Wetzel and V. J. Tipton acquire food. In T.H. Kunz and P.A. Ra- (eds.), 753Ð795. Ectoparasites of Pan- cey (eds.), Bat biology and conserva- ama. Chicago: Field Museum of Natu- tion: 197Ð204. Washington DC: Smith- ral History, 861 pp. sonian Institution Press. 1976. Mammals of the Smithsonian Venezue- Kennedy, M., A. M. Patterson, J. C. Morales, S. lan Project. Brigham Young Univ. Sci. Parsons, A. P. Winnington, and H. G. Spencer Bull. Biol. Ser. 10(5): 991 pp. 1999. The long and short of it: branch lengths Henson, O. W. and the problem of placing the New 1970. The central nervous system. In W.A. Zealand Short-tailed bat, Mystacina. Wimsatt (ed.), Biology of bats, vol 2: Mol. Phylogenet. Evol. 13: 405Ð416. 58Ð152. New York: Academic Press. Kirsch, J. A. W., J. M. Hutcheon, D. G.P. Byrnes, Herd, R. M. and B. D. Lloyd 1983. Pteronotus parnellii. Mamm. Species 1998. Affinities and historical zoogeography 209: 1Ð5. of the New Zealand Short-Tailed bat, Hermanson, J. W., and J. S. Altenbach Mystacina tuberculata Gray 1843, in- ferred from DNA hybridization com- 1983. The functional anatomy of the shoulder parisons. J. Mammal. Evol. 5: 33Ð64. of the Pallid bat, Antrozous pallidus. J. Koopman, K. F. Mammal. 64: 62Ð75. 1951. Fossil bats from the Bahamas. J. Mam- 1985. Functional anatomy of the shoulder and mal. 32: 229. arm of the fruit eating bat Artibeus ja- 1955. A new subspecies of Chilonycteris maicensis. J. Zool. (London A) 205: from the West Indes and a discussion 157Ð177. of the mammals of La Gonave. J. Hill, J. E. Mammal. 36: 109Ð113. 1974. A new family, genus, and species of bat 1958. Land bridges and ecology in bat distri- (Mammalia: Chiroptera) from Thai- bution on islands of the northern coast land. Bull. Brit. Mus. (Nat. Hist.) Zool. of South America. Evolution 12: 429Ð 27: 301Ð336. 439. Hill, J. E., and M. J. Daniel 1978. Zoogeography of Peruvian bats with 1985. Systematics of the New Zealand short- special emphasis on the role of the An- tailed bat Mystacina Gray, 1843 (Chi- des. Am. Mus. Novitates 2651: 1Ð33. roptera: Mystacinidae). Bull. Br. Mus. 1982. Biogeography of the bats of South (Nat. Hist.) Zool. 48: 279Ð300. America. In M. A. Mares and H. H. 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 89

Genoways (eds.), Mammalian biology Mares, M. A., M. R. Willig, K. E.Streilein, and T. in South America: 272Ð302. Pymatun- E. Lacher, Jr. ing Lab. Ecol. Spec. Publ. Ser. 6: 539 1981. The mammals of northeastern Brazil: a pp. preliminary assessment. Ann. Carnegie 1984. Bats. In S. Anderson and J. K. Jones Mus. 50(4): 81Ð137. (eds.), Orders and families of Recent Martin, R. A. mammals of the world: 145Ð186. New 1972. Synopsis of late Pliocene and Pleisto- York: Wiley. cene bats of North America and the 1989. A review and analysis of the bats of the Antilles. Am. Midl. Nat. 87: 326Ð335. West Indes. In C. A. Woods (ed.), Bio- McDaniel, V. R. geography of the West Indes: past, pre- 1976. Brain anatomy. In R. J. Baker, J. K. sent, and future: 635Ð644. Gainesville, Jones, and D. C. Carter (eds.), Biology FL: Sandhill Crane Press. of bats of the New World Family Phyl- 1993. Order Chiroptera. In D. E. Wilson and lostomidae: 147Ð200. Spec. Publ. Mus. D. M. Reeder (eds.), Mammal species Texas Tech Univ. 10: 218 ppÐ218. of the world, a taxonomic and geo- McKenna, M. C., and S. K. Bell graphic reference, 2nd ed.: 137Ð241. Washington, D C: Smithsonian Institu- 1997. Classification of mammals above the tion Press. species level. New York: Columbia 1994. Chiroptera: systematics. Handbook of Univ. Press. Zoology, vol. VIII, pt. 60, Mammalia: Menu, H., and B. Sige« 1Ð217. Berlin: de Gruyter. 1971. Nyctalodontie et myotodontie, impor- Lancaster, W. C., and E. K. V. Kalko tants caracte`res de grades e«volutifs 1996. Mormoops blainvillii. Mamm. Species chez les chiropte`res entomophages. C. 544: 1Ð5. R. Acad. Sci. Paris 272: 1735Ð1738. Leach, R. P. Miller, G. S., Jr. 1821a. The characters of three new genera of 1907. The families and genera of bats. Bull. bats without foliaceous appendages to U.S. Nat. Mus. 57: 282 pp. the nose. Trans. Linn. Soc. 13: 69Ð72. Morgan, G. S. 1821b. The characters of seven genera of bats 1989. Fossil Chiroptera and Rodentia from with foliaceous appendages to the nose. the Bahamas, and the historical bioge- Trans. Linn. Soc. 13: 73Ð82. ography of the Bahamian mammal fau- Linares, O. J., and J. Ohasti na. In C. A. Woods (ed.), Biogeography 1974. Una nueva subespecie del murcielago of the West Indes: past, present, and fu- Pteronotus parnellii, en las cuevas de ture: 685Ð740. Gainesville, FL: Sand- la Peninsula de Paraguana, Venezuela hill Crane Press. (Chiroptera: Mormoopidae). Bol. Soc. 1991. Neotropical Chiroptera from the Plio- Venez. Espeleol. 5: 73Ð78. cene and Pleistocene of Florida. Bull. Livezey, B. C. Am. Mus. Nat. Hist. 206: 176Ð213. 1998. A phylogenetic analysis of the Gruifor- Morgan, G. S., and C. A. Woods mes (Aves) based on morphological 1986. Extinction and zoogeography of West characters, with an emphasis on the Indian land mammals. Biol. J. Linn. rails (Rallidae). Philos. Trans. R. Soc. Soc. 28: 167Ð203. London B Biol. Sci. 353: 2077Ð2151. Murray, P. F., and T. Strickler Luckett, W. P. 1975. Notes on the structure and function of 1980. The use of fetal membrane characters cheek pouches within the Chiroptera. J. in assessing chiropteran phylogeny. In Mammal. 56: 673Ð676. D. E. Wilson and A. L. Gardner (eds.), Neuweiler, G. Proceedings Fifth International Bat Re- search Conference, Pp. 245Ð266. Lub- 1984. Foraging, echolocation and audition in bock: Texas Tech Press. bats. Naturwissenschaften 71: 446Ð Maddison, W. P., and D. R. Maddison 455. 1992. MacClade, version 3.0. Sunderland, 1989. Foraging ecology and audition in ech- MA: Sinauer. olocating bats. Trends Ecol. Evol. 4: Maddison, W. P., M. J. Donoghue, and D. R. Mad- 160Ð166. dison 1990. Auditory adaptations for prey capture 1984. Outgroup analysis and parsimony. Syst. in echolocating bats. Physiol. Rev. 70: Zool. 33: 83Ð103. 615Ð641. 90 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

Neuweiler, G., and M. B. Fenton Patton, J. C., and R. J. Baker 1988. Behavior and foraging ecology of ech- 1978. Chromosomal homology and evolution olocating bats. In: P. E. Nachtigall and of phyllostomatoid bats. Syst. Zool. 27: P.W.B. Moore (eds.), Animal sonar: 449Ð462. processes and performance: 535Ð549. Peters, W. New York: Plenum Press. 1864. U¬ ber einige neue Sa¬ugethiere (Mor- Nixon, K. C., and J. M. Carpenter moops, Macrotus, Vesperus, Molossus, 1993. On outgroups. Cladistics 9: 413Ð426. Caprimys), Amphibien (Platydacylus, Norberg, U. M. Otocryptis, Euprepes, Ungalia, Drom- 1970. Functional osteology and myology of icus, Tropidonotus, Xenodon, Hylodes) the wing of Plecotus auritus Linnaeus und Fishe (Sillago, Sebastes, Channa, (Chiroptera). Ark. Zool. 22: 483Ð543. Myctophum, Carassius, Barbus, Ca- Novacek, M. J. poe¬ta, Poecilia, Saurechelys, Lepto- 1985a. Evidence for echolocation in the oldest cephalus). Monatsber. K. Preuss. Akad. known bats. Nature 315: 140Ð141. Wiss. 1865: 381Ð399. 1985b. Comparative morphology of the bat au- Pierson, E. D. ditory region. Fortschr. Zool. 30: 149Ð 1986. Molecular systematics of the Microchi- 151. roptera: higher taxon relationships and 1987. Auditory features and affinities of the biogeography. Ph.D. diss., Univ. Cali- Eocene bats Icaronycteris and Palaeo- fornia, Berkeley. chiropteryx (Microchiroptera, incertae Pierson, E. D., V. M. Sarich, J. M. Lowenstein, sedis). Am. Mus. Novitates 2877: 18 M. J. Daniel, and W. E. Rainey pp. 1986. A molecular link between the bats of 1991. Aspects of morphology of the cochlea New Zealand and South America. Na- in microchiropteran bats: an investiga- ture 6083: 60Ð63. tion of character transformation. Bull. Pocock, R. I. Am. Mus. Nat. Hist. 206: 84Ð100. 1914. On the facial vibrissae of Mammalia. 1992. Fossils as critical data for phylogeny. Proc. Zool. Society London 1914: 889Ð In M. J. Novacek and Q. D. Wheeler 912. (eds.), Extinction and phylogeny: 46Ð Quinn, T. H., and J. J. Baumel 88. New York: Columbia Univ. Press. 1993. Chiropteran tendon locking mecha- 1994. Morphological and molecular inroads nism. J. Morphol. 216: 197Ð208. to phylogeny. In L. Grande and O. Ray, C. E., S. J. Olsen, and H. J. Gut Rieppel (eds.), Interpreting the hierar- 1963. Three mammals new to the Pleistocene chy of nature: from systematic patterns fauna of Florida, and a reconsideration to evolutionary process theories: 85Ð of five earlier records. J. Mammal. 44: 131. New York: Academic Press. 373Ð395. Novick, A. Reid, F. A. 1963. Pulse duration in the echolocation of 1997. A field guide to the mammals of Cen- insects by the bat, Pteronotus. Ergeb. tral America and southeast Mexico. Biol. 26: 21Ð26. New York: Oxford Univ. Press. 1965. Echolocation of flying insects by the bat, Chilonycteris psilotis. Biol. Bull. Rezsutek, M., and G. M. Cameron 128: 297Ð314. 1993. Mormoops megalophylla. Mamm. Spe- 1977. Acoustic orientation. In W. A. Wimsatt cies 448: 1Ð5. (ed.), Biology of bats, vol. 3, pp. 73Ð Rodr«õguez-Dura«n, A. 289. New York: Academic Press. 1995. Metabolic rates and thermal conduc- Novick, A., and J. R. Vaisnys tance in four species of Neotropical 1964. Echolocation of flying insects by the bats roosting in hot caves. Comp. bat, Chilonycteris parnellii. Biol. Bull. Biochem. Physiol. 110A: 347Ð355. 127: 478Ð488. 1998. Nonrandom aggregations and distribu- Omland, K. E. tion of cave-dwelling bats in Puerto 1997. Correlated rates of molecular and mor- Rico. J. Mammal. 79: 141Ð146. phological evolution. Evolution. 51: Rodr«õguez-Dura«n, A., and T. H. Kunz 1381Ð1393. 1992. Pteronotus quadridens. Mamm. Spe- Park, H., and E. R. Hall cies 395: 1Ð4. 1951. The gross anatomy of the tongues and Rodr«õguez-Dura«n, A., and A. R. Lewis stomachs of eight New World bats. 1987. Patterns of population size, diet, and Trans. Kansas Acad. Sci. 54: 64Ð72. activity time for a multispecies assem- 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 91

blage of bats at a cave in Puerto Rico. Schutt, W. A., Jr. and N. B. Simmons Caribb. J. Sci. 23: 352Ð360. 1998. Morphology and homology of the chi- Rodr«õguez-Dura«n, A., A. R. Lewis, and Y. Montes ropteran calcar, with comments on the 1993. Skull morphology and diet of Antillean phylogenetic relationships of Ar- bat species. Caribb. J. Sci. 29: 258Ð chaeopteropus. J. Mamm. Evol. 5(1): 261. 1Ð32. Sanchez-Herrera, O., G. Tellez-Giron, R. A. Med- Silva-Taboada, G. ellin, and G. Urbano-Vidales 1974. Fossil Chiroptera from cave deposits in 1986. New records of mammals from Quin- central Cuba, with descriptions of two tano Roo, Mexico. Mammalia 50: 275Ð new species (genera Pteronotus and 278. Mormoops) and the first West Indian Schaffer, J. record of Mormoops megalophylla. 1905. Anatomisch-histologische Untersuchun- Acta Zool. Cracov. 19: 34Ð73. gen u¬ber den Bau der Zehen bei Fleder- 1976. Historia y actualizacio«n taxonomica de mau¬sen und einigen kletternden Sa¬uge- algunas especies antillanas de murc«õe- tieren. Z. Wiss. Zool. 83: 231Ð284. lagos de los ge«neros Pteronotus, Bra- Schlosser-Strum, E., and H. Schliemann chyphylla, Lasiurus, y Antrozous 1995. Morphology and function of the shoul- (Mammalia: Chiroptera). Poeyana 153: der joint of bats (Mammalia: Chirop- 1Ð24. tera). J. Zool. Syst. Evol. Res. 33: 88Ð 1979. Los murc«õelagos de Cuba. Havana: Ed- 98. itorial Academia, Academia de Cien- Schneider, R. cias Cuba, 423 pp. 1957. Metrische Untersuchungen am Gro§hirn Silva-Taboada, G., and R. H. Pine der Chiroptera. Anat. Anz. 104: 142Ð 1969. Morphological and behavioral evidence 149. for the relationship between the bat ge- 1972. Zur quantitativen Morphologie und nus Brachyphylla and the Phyllonycter- Evolution des Fledermausgehirns. inae. Biotropica 1: 10Ð19. Anat. Anz. 130: 332Ð346. Simmons, J. A. Schnitzler, H.-U. 1980. Phylogenetic adaptations and the evo- 1970a. Echoortung bei der Fledermaus Chilon- lution of echolocation in bats. In D. E. ycteris rubiginosa. Z. Vgl. Physiol. 68: Wilson and A. L. Gardner (eds.), Pro- 25Ð39. ceedings Fifth International Bat Re- 1970b. Comparison of the echolocation behav- search Conference: 267Ð278. Lubbock: ior in Rhinolophus ferrumequinum and Texas Tech Press. Chilonycteris rubiginosa. Bijdr. Dierkd. Simmons, J. A., and R. A. Stein 40: 77Ð80. 1980. Acoustic imaging in bat sonar: echo- 1987. Echoes of fluttering insects: informa- location signals and the evolution of tion for echolocating bats. In M. B. echolocation. J. Comp. Physiol. A 135: Fenton, P. A. Racey, and J. M. V. Ray- 61Ð84. ner (eds.), Recent advances in the study Simmons, N. B. of bats: 226Ð243. Cambridge: Cam- 1993. The importance of methods: archontan bridge Univ. Press. phylogeny and cladistic analysis of Schnitzler, H.-U., and O. W. Henson morphological data. In R. D. E. 1980. Performance of airborne animal sonar MacPhee (ed.), Primates and their rel- systems. I. Microchiroptera. In R. G. atives in phylogenetic perspective: 1Ð Busnel, and J. F. Fish (eds.), Animal so- 61. New York: Plenum. nar systems: 109Ð181. New York: Ple- 1994. The case for chiropteran monophyly. num. Am. Mus. Novitates 3103: 54 pp. Schnitzler, H.-U. and E. K. V. Kalko 1998. A reappraisal of interfamilial relation- 1998. How echolocating bats search and find ships of bats. In T. H. Kunz and P. A. food. In T. H. Kunz and P. A. Racey Racey (eds.), Bat biology and conser- (eds.), Bat biology and conservation: vation: 3Ð26. Washington, D. C.: 183Ð196. Washington, D.C.: Smithson- Smithsonian Institution Press, 365 pp. ian Institution Press. Simmons, N. B., and J. H. Geisler Schutt, W. A. 1998. Phylogenetic relationships of Icaron- 1993. Digital morphology in the Chiroptera: ycteris, Archaeonycteris, Hassianycter- the passive digital lock. Acta Anat. is, and Palaeochiropteryx to extant bat 148: 219Ð227. lineages, with comments on the evolu- 92 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

tion of echolocation and foraging strat- Sprague, J. M. egies in Microchiroptera. Bull. Am. 1943. The hyoid region of placental mammals Mus. Nat. Hist. 235: 182 pp. with especial reference to the bats. Am. Simmons, N. B., and C. O. Handley J. Anat. 72(3): 385Ð472. 1998. A revision of Centronycteris Gray Steadman, D. W., G. K. Pregill, and S. Olson (Chiroptera: Emballonuridae) with 1984. Fossil vertebrates from Antigua, Lesser notes on natural history. Am. Mus. Antilles: evidence for Late Holocene Novitates 3239: 28 pp. human- caused extinctions in the West Simmons, N. B., and T. H. Quinn Indes. Proc. Natl. Acad. Sci. U.S.A. 81: 1994. Evolution of the digital tendon locking 4448Ð4451. mechanism in bats and dermopterans: a Stephan, H. phylogenetic perspective. J. Mammal. 1977. Encephalisationsgrad su¬damerikanisch- Evol. 2: 231Ð254. er Flederma¬use und Makromorphologie Simmons, N. B., and R. S. Voss ihrer Gehirne. Gegenbaurs Morphol. 1998. The mammals of Paracou, French Gui- Jahrb. 123: 151Ð179. ana: a Neotropical rainforest fauna. Straney, D. O. Part 1. Bats. Bull. Am. Mus. Nat. Hist. 1980. Relationships of phyllostomatine bats: 237: 219 pp. evaluation of phylogenetic hypotheses. Sites, J. W., J. W. Bickham, and M. W. Haiduk Ph. D. diss., Univ. California, Berkeley. 1981. Conservative chromosome change in Strickler, T. L. the bat family Mormoopidae. Can. J. 1978. Functional osteology and myology of Genet. Cytol. 23: 459Ð467. the shoulder in the Chiroptera. Contrib. Slaughter, B. H. Vertebr. Evol. 4: 1Ð198. 1970. Evolutionary trends of chiropteran den- Swofford, D. L. titions. In B. H. Slaughter and W. D. 2000. Phylogenetic Analysis Using Parsimo- Walton (eds.), About bats, Fondren Sci. ny (PAUP*), version 4.0b3a. Sunder- Ser. 11: 51Ð83. Dallas, TX: Southern land MA: Sinauer. Methodist Univ. Press. Thomas, O. Smith, J. D. 1908. Missing premolar of Chiroptera. Ann. 1972. Systematics of the chiropteran family Mag. Nat. Hist. ser. 8, 1: 346Ð348. Mormoopidae. Univ. Kansas Mus. Nat. Van Den Bussche, R., and S. R. Hoofer Hist. Misc. Pub. 56. 2000. Further evidence for the inclusion of 1976. Chiropteran evolution. In R. J. Baker, the New Zealand short-tailed bat (Mys- J. K. Jones, and D. C. Carter (eds.), Bi- tacina tuberculata) within Noctiliono- ology of bats of the New World family idea. J. Mammal. 81: 865Ð874. Phyllostomatidae. Pt. 1: 49Ð69. Spec. Van Valen, L. Publ. Mus. Texas Tech Univ. vol. 10. 1979. The evolution of bats. Evol. Theory 4: Lubbock: Texas Tech Univ. 104Ð121. 1977. On the nomenclatural status of Chilon- Vaughan, N., and J. E. Hill ycteris gymnonotus Natterer, 1843. J. 1996. Bat (Chiroptera) diversity and abun- Mammal. 58: 245Ð246. dance in banana plantations and rain 1980. Chiropteran phylogenetics: introduc- forest, and three new records for St. tion. In D. E. Wilson and A. L. Gardner Vincent, Lesser Antilles. Mammalia (eds.), Proceedings Fifth International 60: 441Ð447. Bat Research Conference: 233Ð244. Vaughan, T. A. Lubbock: Texas Tech Press. 1959. Functional morphology of three bats: Smith, J. D., and G. Madkour Eumops, Myotis, Macrotus. Univ. Kan- 1980. Penial morphology and the question of sas Publ. Mus. Nat. Hist. 12: 153 pp. chiropteran phylogeny. In D. E. Wilson 1970a. The skeletal system. In W. A. Wimsatt and A. L. Gardner (eds.), Proceedings (ed.), Biology of bats, Vol. 1: 98Ð139. Fifth International Bat Research Con- New York: Academic Press. ference: 347Ð365. Lubbock: Texas 1970b. The muscular system. In W. A. Wimsatt Tech Press. (ed.), Biology of bats, Vol. 1: 140Ð194. Sonntag, C. F. New York: Academic Press. 1925. The comparative anatomy of the Vaughan, T. A. and G. C. Bateman tongues of the Mammalia. XII. Sum- 1970. Functional morphology of the forelimb mary, classification and phylogeny. of Mormoopid bats. J. Mammal. 51: Proc. Zool. Soc. London 1: 701Ð762. 217Ð235. 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 93

Vieira, C. O. C. Whitaker, J. O., and J. S. Findley 1955. Lista remissivia dos mam«õferos do Bras- 1980. Foods eaten by some bats from Costa il. Arq. Zool. (Sa˜o Paulo) 8: 354Ð374. Rica and Panama. J. Mammal. 61: Villa-R., B. 540Ð543. 1967. Los murcie«lagos de Me«xico. Anal. Inst. Wible, J. R., and K. P. Bhatnagar Biol. Univ. Nac. Auto«n. Me«x. 491 pp. 1996. Chiropteran vomeronasal complex and Vizotto, L. D., V. Rodriguez, and A. J. Dumbra the interfamilial relationships of bats. J. 1980. Sobre ocorre«ncia e dados biome«tricos de Mammal. Evol. 3: 285Ð314. Pteronotus (Pteronotus) gymnonotus Willig, M. R. (Natterer, in Wagner, 1843), no Estado do 1983. Composition, microgeographic varia- Piaui (ChiropteraÐ-Mormoopidae). Rev. tion, and sexual dimorphism in Caatin- Nordestina Biol. 3(especial): 246Ð247. gas and Cerrado bat communities from Wagner, J. A. northeast Brazil. Bull. Carnegie Mus. 1843. Diagnosen neuer Arten brasilischer Nat. Hist. 23: 131 pp. handflunger. Arch. Naturgesch. 9: 365Ð 1985. Reproductive patterns of bats from 368. Caatingas and Cerrado biomes in north- Walton, D. W., and G. M. Walton east Brazil. J. Mammal. 66(4): 668Ð 1970. Post-cranial osteology of bats. In B.H. 681. Slaughter and D. W. Walton (eds.), Willig, M. R., and M. A. Mares About bats: 93Ð126. Dallas TX: South- 1989. Mammals from the Caatinga: an updat- ern Methodist Univ. Press. ed list and summary of recent research. Webb, R. G., and R. H. Baker Rev. Bras. Biol. 49: 361Ð367. 1962. Terrestrial vertebrates of the Pueblo Wilkins, K. T. Nuevo area of southwestern Durango, 1983. Pleistocene mammals from the Rock Mexico. Am. Midl. Nat. 68: 325Ð333. Springs local fauna, central Florida. Wetterer, A. L., M. V. Rockman, and N. B. Sim- Brimleyana 9: 69Ð82. mons Woloszyn, B. W., and G. Silva-Taboada 2000. Phylogeny of phyllostomid bats: data 1977. Nueva especie fo«sil de Artibeus (Mam- from diverse morphological systems, malia: Chiroptera) de Cuba, y tipifica- sex chromosomes, and restricted cion preliminar de los depo«sitos fosil«õ- sites. Bull. Am. Mus. Nat. Hist. 248: feros cubanos contentiovos de mam«õ- 200 pp. feros terrestres. Poeyana 161: 1Ð17.

APPENDIX 1: SPECIMENS EXAMINED Pteronotus parnelli parnelli: Cuba: Oriente; 188314Ð19, 188321Ð24); Oaxaca; Juchitan, 3 mi Siboney, Cueva de Los Majaes (AMNH 176182); N El Jicaro (AMNH 185771); Oaxaca; Juchitan, Pinar del Rio, La Mulata, Pan de Guajaibon Tapanatepec (AMNH 177738); Tamaulipas; Acu- (AMNH 176181); Jamaica: St. Catherine; St. n˜a, Sierra de Tamaulipas (AMNH 249063Ð64). Clair (AMNH 247594); St. Elizabeth; Balaclava, Pteronotus parnelli mesoamericanus: Mexico: Oxford Cave (AMNH 45206Ð17); Hanover; Lu- Chiapas; 1.3 mi SE Zapaluta (AMNH 252516); sea (AMNH 12382Ð83, 12385); Trelawny; 5 mi Honduras: Lempira; Las Flores (AMNH N and 2.5 mi WNW Quick Step (USNM 545139Ð 129009Ð34, 129741, 129743Ð50); Guatemala: 40). Sta. Rosa; 2 km NE Chiquimilla (AMNH Pteronotus parnelli pusillus: Dominican Re- 243727); Sta. Rosa; 13 km N Avellana (AMNH public: Santiago, 4 km SW Moncio«n, Cueva Dur- 243728); Nicaragua: Leo«n; 50 mi N Se«baco a«n (AMNH 212996); San Rafael; Rancho La (AMNH 177740); Costa Rico: San Jose«; Las Pa- Guardia (AMNH 213893); Haiti: Sud; Sapoti (19 vas (AMNH 135243, 135245, 135249). km SW Beaumont) (AMNH 236655). Pteronotus parnelli rubiginosus: Panama: Ca- Pteronotus parnelli portoricensis: United nal Zone; Chilobrillo (AMNH 143797); Venezue- States: Puerto Rico; Trujillo Alto, Cueva de Tru- la: Araqua; Rancho Grande (AMNH 144389Ð90); jillo Alto (AMNH 39369Ð70). Trinidad and Tobago: St. Andrew; Tamana«, Mt. Pteronotus parnelli mexicanus: Mexico: Coli- Tamana« (AMNH 175540, 246152Ð53, 246155); ma; Pueblo Juarez (AMNH 188747Ð48); Jalisco; St. George; Blanchisseuse, Las Cueves (AMNH 2.3 mi SW La Cuesta (AMNH 254606); Jalisco; 182693); Brazil: Amazonas; Rio Bronco, Serra de 7.5 km W Jama, Cueva D.C.V. (AMNH 204958); Cuchy (AMNH 78121Ð25, 78127Ð31, 78133Ð34, Jalisco; Los Masos (AMNH 27293); Mexico; 6 78173). km SEE Tunatico; Cueva del Coyote (188311Ð12, Pteronotus parnelli fuscus: Colombia: Mag- 94 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258 dalena; Cacagualito (AMNH 23656); Venezuela: co; Los Masos (AMNH 27294Ð96); Morelos; Aragua; Rancho Grande (AMNH 144842, Cueva de Tequesquitengo (AMNH 204454Ð57, 144845); Carabobo; Las Quiguas (AMNH 31563, 204967Ð75); Sonora; 5 mi W Alamos (Mine Tun- 31578Ð79); Carabobo; San Esteban (AMNH nel at La Adurna) (AMNH 180312); Ocozocoau- 31564Ð31570, 31574Ð31576); Yaracui; Aroa, Bo- tla; Ocozocoautla (AMNH 171652); San Luis Po- liviar R.R. (AMNH 32099Ð103, 32126). tos; El Salto (AMNH 177588Ð611); El Salvador: Pteronotus macleayii macleayii: Cuba: Ha- La Union; Tabanco (AMNH 239229). bana; Bucaranao (AMNH 174250); Oriente; Ba- Pteronotus davyi incae: Peru: Cajamarca; Jaen racoa, Cueva de Mojana (AMNH 61151Ð53); Pi- (AMNH 69233); Piura; Huancabamba (AMNH nar del Rio; San Vicente (AMNH 17618Ð88). 63721, 64082). Pteronotus macleayii grisea: Jamaica: Hano- Pteronotus gymnonotus: Panama: Canal Zone; ver; Lucea (AMNH 12359); St. Elizabeth; Bala- Barro Colorada Island, Allee Creek (USNM clava, Oxford Cave (AMNH 45256, 45268Ð75); 498707); Canal Zone; Madden Dam, Chillibrillo St. James; Montego Bay (AMNH 12390, 60917, Cave (AMNH 89998, USMN 306526, 306529Ð 60896, 60899, 60902, USNH 260730); St. James, 30); Chilibre (USNH 305177); Colombia: Boli- Monetgo Bay, Providence Cave (34884Ð86, var; Cartagena (USNM 431544); Venezuela: Yar- 45257Ð67); Trelawny, Windsor (ANMH 61022). acui; Aroa; Bolivar R.R. (AMNH 32083Ð89, Pteronotus quadridens fuliginosus: United 32091Ð94). States: Puerto Rico; San Juan, Pueblo Viejo, Cue- Pteronotus pristinus: Cuba: Las Villas; Cueva va de Fari (AMNH 39353Ð59, 39363Ð67); Truji- de Jaguey, near Trinidad (ROM 59132). llo Alto (39396Ð03); Dominican Republic: be- Mormoops blainvillii: [No data] (AMNH tween Cabra and Nagua, Cueva de Muricelagos 213889); United States: Puerto Rico; Mona Is- (AMNH 238138); La Altagracia; Boca de Yuma land (AMNH 72547); Puerto Rico; Trujillo Alto, (AMNH 238139); Santo Domingo; Cana Honda Cueva de Trujillo Alto (AMNH 72370); Domin- (AMNH 25760); Haiti: Cerca La Source (USNH ican Republic: Barahano; Pedernales, Cabo Rio 253631Ð32); Sud; Sapoti (19 km SW Beaumont) (AMNH 238144); San Cristobal; Cueva La Che- (AMNH 236654); Jamiaca: St. Elizabeth; Bala- pa, 10 km N Bayaguana (USNH 535251); San clava, Oxford Cave (AMNH 45248Ð55). Rafael; Rancho la Guardia (AMNH 213897); Ja- Pteronotus quadridens torrei: Cuba: [No ad- maica: St. Catherine; St. Clair (AMNH 246123); ditional information] (AMNH 60637Ð38); 8 mi E St. Elizabeth; Balaclava, Oxford Cave (AMNH Baracoa (USNH 121061); La Habana; Bararanao, 45197Ð205); Trelawny Parish; Quick Step (USNH Cueva de El Fantasma (AMNH 176061); Las Vil- 511225Ð27, 545146); Cuba: [No additional in- lo; Punta Caquanes, Yaguajay ‘‘Cueva de Colon’’ formation] (AMNH 213889); Oriente; Siboney, (AMNH 186970Ð71). Cueva de Los Majaes (AMNH 176149Ð50); Pinar Pteronotus personatus psilotis: Mexico: Mo- del Rio; San Vincente, Cueva del Indio (AMNH relos; Tequesquitengo (AMNH 180443Ð44, 176139). 180446); Nayarit; NayaritÐ-Sinaloa state line, Rio Mormoops megalophylla megalophylla: Mexi- Canas, near la Concha (USNH 508415); Oaxaca; co: Jalisco; Los Masos (AMNH 27298, 27300Ð Juchitan, Tapanatpec (175018Ð19, 177672Ð93, 03, 27305); Morelos; Teco«men (AMNH 206692Ð 177695Ð96, 178468). 93); Nayarit; Rancho Palo Amarillo (AMNH Pteronotus personatus personatus: Colombia: 25070Ð71); Oaxaca; 5 mi W Chilltepec (AMNH Bolivar; Cartagena (USNH 433517, 433531); 190138Ð39); Oaxaca; Tapamatepec (AMNH Panama: San Blas; Armila (Quebrada Venado) 177423Ð24); Sonora (USNH 314677); Sonora; 15 (USNH 335085Ð86); Venezuela: Sucre; la Brea mi NW Guaumas (AMNH 239230, 239232); Ta- (AMNH 33300); Yaracui; Aroa, Bolivar R.R. maulipas; Acun˜a; Sierra de Tamaulipas (AMNH (33112, 32123Ð24, 32127Ð32133, 32136Ð37); 249066). Trinidad and Tobago: St. George; Tacariquas; Mormoops megalophylla intermedia: Nether- St. Augustine (AMNH 175560). land Antilles: Aruba (AMNH 185613, 185615Ð Pteronotus davyi davyi: Dominica: Roseau 17); Curacao (USNH 102225); Curacao; Round (AMNH 35923); Trinidad and Tobago: St. Cliff (AMNH 149395Ð96). George: Arima Heights of Aripo (AMNH 29694); Mormoops megalophylla tumidiceps: Colom- St. George; Diego Martin, Carenage (AMNH bia: Bolivar; Cartagena (USNM 431641); Santan- 29703Ð04); St. George; Port of Spain (AMNH der; San Gil (AMNH 149257, 183315Ð16); Trin- 7445, 19122); Nariva (AMNH 207060); St. Pat- idad and Tobago: [No additional information] rick, Siparia (AMNH 179954). (AMNH 3173); St. Andrew; Tamana«, Mt. Tamana« Pteronotus davyi fulvus: Mexico: Chiapas; 18 (AMNH 175566, 175576); St. George; Port of mi W Ixhuatan (AMNH 203563Ð64); Jalisco; 7.5 Spain, Goodwood Park (AMNH 183845). km W Jamai, Cueva ‘‘Deb’’ (204959Ð65); Jalis- Macrotus waterhousii: Mexico: Jalisco; Bolan˜o 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 95

(AMNH 2003); Morelos; Lago de Tequesquiten- Madiera, Auara Igarape« (AMNH 91960Ð75); go, Cueva del Cerro (USNM 559490Ð91); Nay- Para«; Fero (AMNH 93803); Para«: R. Tocantins, arit; Islas Tres Marias; Maria Madre (AMNH Ilha do Taiuna (AMNH 96930); Peru: Amazonas; 180523); Oaxaca; Tehuantepec; Arroyo Ocotita Rio Amazonas, Pt. Indiana (AMNH 93932, (AMNH 182158); Dominican Republic: Caves 93970); Pacso; Oxapampa, San Juan (AMNH near River Charon; E La Romana (AMNH 230105, 230108Ð09); Pasco; Oxapampa, San Pab- 120975); near San Juan Bay, Rolla Seco (AMNH lo (AMNH 230110Ð12); Bolivia: Beni; ca. 23 km 91314); Haiti: Sud; 1 mi S Duchity (7 km S W San Javier (AMNH 210598); Cochabamba; Jo- Beaumont) (AMNH 236657Ð58); Cuba: Maran- doe Santoe (AMNH 38562); Paraguay: Boquer- zos, near Cidra (AMNH 61161); Jamaica: Clar- on; Pozo Colorado (trans-chaco route, km 250) endon; Mahoe Gardens, Portland Cave (USMN (AMNH 234277); Trinidad and Tobago: North 511232Ð33); St. Catherine; Healthshire Hills Manzanilla (AMNH 175534); United States: (AMNH 268068); St. Elizabeth; Balaclava, Ox- Puerto Rico; Humacao, Old Loiza (AMNH ford Cave (AMNH 45229); St. Elizabeth; Houn- 39574, 39587); St. Kitts-Nevis: Nelson Spring slow; Money Cave (AMNH 268071); Trelawny; (AMNH 246993). 1 km E Quick Step (USMN 511230); Trelawny; Noctilio albiventris: Colombia: Amazonas; Le- Windsor (AMNH 45230Ð31); Cayman Islands: ticia (AMNH 207772, 207774); Brazil: Amazon- Grand Cayman Island; 1 km E Spots (USNM as; Rio Madeira, Auarer Igarape« (AMNH 92019Ð 538120). 30); Amazonas; Rio Negro, near Manaus (AMNH Macrotus californicus: United States: Arizona 79626, 79632); Amazonas; Rio Amazonas, Villa (USNH 314679); California; Imperial, 3 mi N Bella Imperatriz (AMNH 92998); Peru: Oxapam- Potholes (AMNH 139575Ð78); Mexico: Sonora; pa, San Pablo (AMNH 23102); Bolivia: Beni Isla Tiburon (USNH 513814, 513816). (AMNH 210624); Beni; 5 km S Guyarmarin Artibeus jamaicensis: Mexico: Chiapas; 15 mi (AMNH 209254); Beni; ca. 23 km W San Javier SE Mapastepec (by road), 150’ (AMNH 251657Ð (AMNH 210595); Paraguay: Olempo XVI Feurte 58); Jalisco; 2.3 mi SW (by road) Le Cuesta Olempo (AMNH 234256Ð66). (AMNH 254662); Oaxaca; Tapanatepec (AMNH Mystacina robusta: New Zealand: Stewart Is- 177744); Vera Cruz; 1 mi SE Oja de Agua land (AMNH 160269); Island off SW of Stewart (AMNH 254669); Yucantan; Chichen Itza Isl. (AMNH 214243). (AMNH 91236); Yucatan; Sobre La Cairetera a Mystacina tubuerculata: New Zealand: [No Tixkokof (AMNH 204983); Panama: Canal additional data] (MNHN 1983.1464; MVZ Zone; rodman (AMNH 213402); Venezuela: El 174825); North Island; Omahuta Kauri Forest Enoatado (AMNH 135159); Brazil: Rio Tocan- (MVZ 173919; ROM 80308). tins; Iles do Tavina (AMNH 97036Ð37); Bolivia: Saccopteryx bilineata: Mexico: Ejido Tepextles La Paz; 6.6 road km downstream Caranaui, Rio (AMNH 172063); Oaxaca; Sacatepec; Cotzocon; Coraica Valley (AMNH 246641, 246644); Bar- Mazatlan (AMNH 175300Ð01, 178776Ð77); Cos- bados: 10 mi E Bridge Town, Chancery Lane ta Rico: Palmer (Pacific) (AMNH 139449); Pan- (AMNH 239927) United States: Puerto Rico; ama: Canal Zone; Gatun Lake (USNM 514986); Utuado (AMNH 39117Ð18); St. Croix; Cane Bay Chilibrillo Caves (USNM 306471); Guyana: Ru- (AMNH 214228); Jamaica: Cambridge (AMNH pununi Dist.; 15 mi E Dadanawa (USNM 45170); St. Mary, 2 mi W Port Maria (AMNH 338917); French Guiana: Paracou; near Sinna- 219170). mary (AMNH 267060, 267064, 267377); Brazil: Noctilio leporinus: Panama: Canal Zone; Cha- Para; Altamira, 52 km SSW, E Bank Rio Xingu gres River (AMNH 173912); Brazil: Amazonas; (USNM 549311); Bolivia: Beni; 5 km NW Rio Madiera (AMHH 91937); Amazonas; Rio Grande river mouth (AMNH 210469). 96 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 258

APPENDIX 2: TAXON-CHARACTER MATRIX The taxon-character matrix presented below corresponds to the character descriptions and coding described in the text. An electronic version of this data matrix (in MacClade format, including character-state labels) is available over the internet at ftp://ftp.amnh.org/pub/mammalogy.

Saccopteryx bilineata 11-10 00000 00000 12000 01021 1210- 1-020 01--0 1--00 10??? ??001 01100 ?0000 021?0 00000 00000 00000 00000 00010 00000 10002 01060 00100 00020 -0101 00000 01001 10000 00101 10000 00001 10013 00000 00202 20220 00000 1000? 0???? ????? ?1000 01000 0011 Noctilio albiventris 00201 01100 11100 12111 01000 12000 0122- 11--1 01100 21??? ??0?? ????? ????? ????? ?0001 10100 02000 00000 10030 00000 02000 02112 01000 01030 -0100 00101 00102 11001 11000 10100 00001 03102 01110 10001 11131 00100 00000 0???? ???0? 01011 0100? 0002 Noctilio leporinus 00201 01100 11100 12111 01000 12000 0121- 11--1 01100 21010 11012 00113 10100 02000 ?0001 10100 02000 00000 10030 00000 02000 02112 01000 01030 -0100 00101 00102 11001 11000 10100 00001 03102 01110 10001 11131 00100 00000 0???? ???0? 01011 0100? 0002 Mystacina robusta 00000 00001 23100 10110 10000 0110- 0-22- 10001 01000 20??? ??011 00013 ?2111 002?? ????? ??000 00003 00000 00000 0???? ??0?? ?0??? 00020 01010 -1001 11111 01000 01022 01300 10010 01001 01011 10010 ?0000 20001 1???? ????? ????? ????? ????0 0000? ???? Mystacina tuberculata 00000 00001 23100 10110 00000 0110- 0-22- 10001 01000 20020 ????? ????? ????? ????? ?1110 10000 00003 00000 00000 01001 11022 10002 00020 0???? ????? ????? ????? ??022 013?0 ??010 01101 ?1011 1?010 ?0000 30?01 1???? ????? ????? ????? ?01?? ????? ???? Macrotus waterhousii 00001 00000 12100 02110 10110 00010 00111 10100 00000 20??? ??011 20113 ?0101 12?01 01110 10000 00112 20001 02002 02010 01022 10020 02000 00020 -0100 10001 10110 10010 00210 11011 00011 14000 20001 10111 40120 01011 11000 00010 01100 1210? ????? 0100 Macrotus californicus 00001 00000 12100 02110 10110 00010 00111 10100 00000 20200 ?00?? ????? ????? ????? ?1110 10000 00112 20001 02002 02010 01022 10020 02000 00020 -1100 10001 10110 10010 00210 11011 00011 14000 20001 10111 40120 01011 11000 00010 01100 12102 1110? ???? Artibeus jamaicensis 00001 00100 02100 02110 00100 01001 00111 10001 010-1 3-201 0201? 20113 00100 02001 01121 11010 00112 20000 00003 00121 11111 10051 02010 00011 00001 11001 00100 10010 00200 10011 00001 02100 00001 10201 00001 0???1 11??0 ????? ???01 02103 21112 1110 Pteronotus parnellii 00001 10000 13100 10111 02111 12010 00121 10010 01100 20210 01211 10110 00110 02010 10130 00011 01110 10110 10010 11011 02032 00112 02001 11021 11011 10101 01012 20111 01201 11101 00100 00100 01101 11201 30210 00011 11000 00100 20011 10101 11000 0002 Pteronotus pristinus 00001 10001 13100 10??? 0??1? ??0?? ?11?? 100?0 01100 20??? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ???? 2001 SIMMONS AND CONWAY: PHYLOGENY OF MORMOOPID BATS 97

APPENDIX 2: Continued.

Pteronotus personotus 00001 10000 33110 10110 00100 12010 01120 10010 01110 10011 1?210 ????? ????? ????? ?0130 00011 01113 10110 10110 11011 02022 00142 02001 11021 01111 10101 01011 20111 01200 11101 00100 00100 01101 11201 30210 00??? ??010 00101 1001- -010? ????? ???? Pteronotus macleayii 00001 10010 13110 10110 00100 12010 01120 10010 01100 10??? ??2?? ????? ????? ????? ?0130 00011 01110 10200 10110 11011 02042 00142 02001 11021 01011 10101 01011 20111 01201 11101 00100 00100 01001 ?1201 30220 00??? ??010 00101 00011 10111 1100? ???? Pteronotus quadridens 00001 10010 13110 10110 00100 12010 01120 10010 01100 10??? ??2?? ????? ????? ????? ?0130 00011 01113 10200 10110 11011 02042 00142 02001 11021 01111 10101 01011 20111 01201 11101 00100 00100 01001 ?1201 30210 00??? ??010 00101 0001- -0101 1100? ???? Pteronotus davyi 00001 10100 23100 10110 00100 12010 01120 10010 01100 20??? ??2?? ????? ?1??? ????0 10130 00011 01113 10200 10010 11011 02002 00102 12001 01021 01011 10101 01011 20111 11201 11101 00100 00100 01001 11201 30220 00011 11010 00100 0001- -010? ????0 ???? Pteronotus gymnonotus 00001 10100 23100 10110 00100 12010 01120 10010 01100 20??? ?12?? ????? ????? ????? ?0130 00011 01113 10200 10010 11011 02000 00102 12001 01021 01011 10101 01011 20111 11201 11101 00100 00100 01001 11201 30220 00??? ??010 00100 00011 1010? ????? ???? Mormoops blainvillii 00002 10000 01101 10110 00100 12010 11120 10110 00000 10??? ??2?? ????? ????? ????? ?1110 00010 12101 01211 11011 20021 02022 10131 02001 01110 -1111 00001 01011 10101 01201 21100 11111 11000 10001 11201 30130 00??? ??101 11000 10001 10101 1100? ???? Mormoops megalophylla 00002 10000 01101 10110 00100 12010 11120 10110 00000 10101 112?? ????? ?1??? ????0 11110 00010 12101 01211 11011 20021 02053 10131 02001 01110 -1110 00001 01012 20101 01210 21100 11101 11000 10101 11201 30130 00012 21101 11000 10001 1010? ????1 0002