AMERICAN MUSEUM Novitates PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

Home , Carolliinae, Erophylla, Lichonycteris, Phyllonycteris

AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2742, pp. 1-45, figs. 1-33, table 1 August 2, 1982 Systematics of the New World Nectar-Feeding Bats (Mammalia, Phyllostomidae), Based on the Morphology of the Hyoid and Lingual Regions

THOMAS ALAN GRIFFITHS'

ABSTRACT Dissection and histological examination of the vory observed. The other group, comprising the hyoid and lingual regions of the New World nec- remaining 10 glossophagine genera (Glossopha- tar-feeding bats reveal marked modification ofthe ginae, sensu stricto), plus Phyllonycteris, Ero- tongue retractor musculature (Mm. sternohyoi- phylla, and perhaps Brachyphylla form a mono- deus, geniohyoideus, hyoglossus, styloglossus, and phyletic group. Within the newly restricted genioglossus) and modification ofthe internal and subfamily Glossophaginae there are two major external tongue structure from the conditions clades. One clade is composed of Glossophaga, found in non-nectar-feeding bats. Use of these Monophyllus, and surprisingly, Lichonycteris. The derived characters in a cladistic analysis leads to other is composed of the more derived nectar- the phylogenetic hypothesis that nectivory evolved feeding genera: Leptonycteris, Anoura, Hylonyc- twice independently in the family Phyllostomidae. teris, Choeroniscus, Choeronycteris, and probably One group of nectar-feeding phyllostomids, com- Scleronycteris and Musonycteris. Interestingly, prising the genera Lonchophylla, Lionycteris, and both karyotypic evidence, and evidence from den- Platalina (traditionally considered glossopha- tal and basicranial studies, can be interpreted to gines) deserves separate subfamilial status based support the phylogeny presented here. on the markedly different adaptations for necti- INTRODUCTION For almost 15 years the systematic status 1973; Baker and Lopez, 1970; Gerber and of the New World nectar-feeding bats Leone, 1971; Phillips, 1971; Stock, 1975; (subfamily Glossophaginae) has been dis- Gardner, 1977; Baker and Bass, 1979; Baker puted. Baker (1967) first suggested that the et al., 1981) have continued to examine this subfamily Glossophaginae might not be a question, but while many investigators agree monophyletic group on the basis ofhis karyo- that the group is probably not monophyletic, typic studies of several genera of the group. there is little agreement as to exactly how the Since then Baker and others (see Baker, 1970, group should be properly divided.

' Assistant Professor of Biology, Illinois Wesleyan University, Bloomington, IL 61701.

In addition to the comparatively large Within the Phyllostominae, he recognized subfamily Glossophaginae, there is a smaller four "groups": the Vampyri, the Glossopha- group of endemic Antillean genera that are gae, the Stenodermata, and the Desmo- at least partly nectivorous. These genera have dontes. The Glossophagae contained 10 been traditionally placed in a separate species in seven genera. These genera were subfamily, the Brachyphyllinae (=Phyllo- Glossophaga, Phyllonycteris, Monophyllus, nycterinae, Miller, 1907). However, recent Ischnoglossus (=Leptonycteris), Loncho- karyological work (Baker and Bass, 1979) and glossa (=Anoura), Glossonycteris (=Anoura), immunological work (Baker et al., 1981) in- and Choeronycteris. dicated that the Brachyphyllinae may be The "group" Glossophagae was considered closely related to Glossophaga and Mono- distinct from all other phyllostomids on the phyllus (both glossophagines). This finding basis of: (1) the long, narrow muzzle; (2) the led Baker and Bass (1979) to question the long, extensible tongue "clothed with filiform validity of the subfamily Brachyphyllinae, papillae"; and (3) the deep groove in the and to once again suggest that the Glosso- lower lip. Except for the inclusion of Phyl- phaginae might not be a monophyletic group. lonycteris, today considered to be grouped The hyoid and lingual regions of the glos- with Brachyphylla and Erophylla in a sepa- sophagine bats are highly modified (Sprague, rate, endemic Antillean subfamily, the 1943; Wille, 1954; Winkelmann, 1971; "group" Glossophagae includes all the bats Greenbaum and Phillips, 1974; Howell and then known that are today considered to be- Hodgkin, 1976; Griffiths, 1978a) presumably long to the subfamily Glossophaginae. to permit the hyperextension of the tongue Miller (1907) reexamined the species required for nectar-feeding. These modifi- known to Dobson, plus specimens in the cations are complex and extensive, and thus United States National Museum and mu- are ideal for use as derived characters (apo- seums in Paris, Leiden, and Berlin. Basing morphies, Hennig, 1966) in a cladistic study. his classification on the structure ofthe wing, The main purpose ofthis paper is to attempt sternum, shoulder girdle, and tooth cusps, to resolve the question of the monophyly of Miller divided the Chiroptera into two sub- the subfamily Glossophaginae via dissection orders, 17 families, and 19 subfamilies. All of the hyoid region and histological exami- subsequent classifications of the Chiroptera nation of the tongue. Secondary purposes are based on Miller's work, including Simp- include resolving the exact systematic rela- son (1945), Hall and Kelson (1959), Koop- tionship of the three genera of the other man and Cockrum (1967), Koopman and nectar-feeding subfamily (Brachyphyllinae) Jones (1970), Smith (1976), and Hall (1981). to the Glossophaginae, and determining the Miller (1907) divided the family Phyllostom- relationships of the genera within each idae into seven subfamilies: the Chilonycter- subfamily to one another. To accomplish inae (=family Mormoopidae, Smith, 1972), these goals, representative species of each Phyllostominae, Stenoderminae, Phyllonyc- genus of glossophagine (except Musonycteris terinae (=Brachyphyllinae, Baker, 1979), and Scleronycteris, which were unavailable) Hemiderminae, Sturnirinae, and Glosso- were dissected and compared with represen- phaginae. Miller (1907) recognized that Phyl- tative species ofeach genus ofbrachyphylline lonycteris, "Reithronycteris" (=Phyllonyc- bat, and with a variety of species ofnon-nec- teris, Koopman, 1952), and Erophylla tar-feeding phyllostomid bats. deserved separate subfamilial status on the basis of the "peculiar" tooth structure and HISTORY OF THE PROBLEM the modified noseleaf. Miller also added the The first attempt at a unified classification genera Lonchophylla, Hylonycteris, and Li- of a large number of New World bats was chonycteris to the subfamily Glossophaginae, that of Dobson (1878). Dobson recognized and recognized Dobson's "Ischnoglossa" and two subfamilies within the family Phyllo- "Glossonycteris" as Leptonycteris and An- stomidae: the Lobostominae (=Mormoopi- oura, respectively. With the addition of five dae, Smith, 1972) and the Phyllostominae. more genera (Scleronycteris, Thomas, 1912; 1 982 GRIFFITHS: NECTAR-FEEDING BATS 3

Lionycteris, Thomas, 1913; Choeroniscus, immunologic reactions of sera of glossoph- Thomas, 1928; Platalina, Thomas, 1928; agine bats, also suggested that the Glosso- and Musonycteris, Schaldach and Mc- phaginae were an artificial grouping ofnectar- Laughlin, 1960), the subfamily Glossophag- feeders. They too suggested that there was a inae was generically complete as it is tradi- distinct Glossophaga group and a distinct tionally recognized today. Choeronycteris group. However, relation- However, recent investigations in the areas ships of these groups to non-glossophagines ofchromosome morphology, immunological were directly opposite to those suggested by reactions ofblood sera, and hard morphology Baker (1967). Glossophaga soricina and of the basicranial skull and teeth have given Glossophaga commissarisi were immunolog- rise to speculation that the classic "Glosso- ically more closely related to Carollia than phaginae" may not be a monophyletic group. to Choeronycteris; Choeronycteris mexicana Baker (1967, 1970), on the basis of chro- was most closely related to Phyllostomus, mosome morphology, suggested that Lepto- Chrotopterus (both phyllostomines), and sur- nycteris sanborni, Glossophaga soricina, G. prisingly, to Desmodus, the vampire bat. alticola, and G. commissarisi form a distinct Recent electrophoretic and albumin immu- group which may be more closely related to nological work by Baker et al. (1981) seem Phyllostomus hastatus, Macrotus water- to contradict the karyotypic studies further. housii, and Trachops cirrhosus (all subfamily Baker et al. (1981) presented evidence that Phyllostominae) than to two other glossoph- Anoura, Glossophaga, Monophyllus, Lepto- agines: Choeronycteris mexicana and Choe- nycteris, Hylonycteris, and Choeroniscus form roniscus godmani. Baker (1967) hypothe- a clade. This suggestion directly contradicts sized that the Glossophaginae may actually the karyotypic studies (see Baker, 1967; be an artificial grouping of nectar-feeders, Baker and Bass, 1979) that suggest the Glos- evolved from two or more independent lines. sophaginae are not monophyletic. Choeronycteris and Choeroniscus karyotypi- Stock (1975) contributed to the Baker/Ger- cally showed great similarity to Carollia sub- ber and Leone controversy by reexamining rufa and Carollia perspicillata (subfamily chromosomes of Carollia and Choeroniscus Carolliinae), whereas another glossophagine, using G and C banding techniques. Stock Anoura geoffroyi, showed similarities to both showed that although the gross chromosomal the Leptonycteris-Glossophaga group and the morphology of Carollia and Choeroniscus is Choeronycteris-Choeroniscus group, but was similar, banding patterns show that there is karyotypically distinct from each. Baker and no relationship between Carollia and Choe- Lopez (1970) added Monophyllus redmani roniscus [refuting half of Baker's (1967) hy- to the Leptonycteris-Glossophaga group. They pothesis]. also, however, examined the karyotypes of Slaughter (1970) examined dentitions of a bats of the other New World nectar-feeding number of bats, including a few glossopha- subfamily, the Brachyphyllinae (=Phyllo- gines. Although he did not divide the Glos- nycterinae) and found that the chromosomes sophaginae into two formal groups, Slaughter ofErophylla and Brachyphylla (both brachy- did observe that "primitive" glossophagines phyllines according to Silva Taboada and such as Glossophaga and phyllostomines Pine, 1969; and Nagorsen and Peterson, such as Macrotus may have shared a recent 1975) are similar to each other, and to those common ancestor. Slaughter (1970) also of the Leptonycteris-Glossophaga-Mono- noted the similarity between "advanced" phyllus group. Further work by Baker (1973), glossophagine (Choeroniscus-type) teeth, and Gardner (1977), and Baker and Bass (1979), the teeth of carolliines and desmodontines confirmed this finding, and both Gardner (vampires). Phillips (1971) examined the (1977) and Baker and Bass (1979) suggested dentitions ofthe glossophagines, although in that the Brachyphyllinae (=Phyllonycteri- a much more rigorous manner. Phillips also nae) be grouped with the Glossophaga-Mono- split the Glossophaginae into two groups (see phyllus group within the Glossophaginae (see fig. 2). In regard to the origins of these two fig. 1). groups, he suggested that the teeth of Choe- Gerber and Leone (1971), studying the ronycteris resemble those of Phyllostomus, 4 AMERICAN MUSEUM NOVITATES NO. 2742

FIG. 1. Relationships among the genera of the Glossophaginae. Redrawn from Gardner (1977).

Mimon, Chrotopterus (all phyllostomines), Baker (1967), Baker and Bass (1979), and and Carollia (Carolliinae), whereas the teeth Gardner (1977) to include the brachyphyl- of Glossophaga and Leptonycteris resemble lines with Glossophaga, Monophyllus, and those of Macrotus and Artibeus (a phyllo- Leptonycteris. stomine and a stenodermatine). Other workers have also supported a There are a number of other, less rigorous schism within the Glossophaginae, although works that have made reference either to the they have not always suggested that the glos- origin (or origins) of the Glossophaginae, or sophagines are a polyphyletic group. Forman, to the possibility of a polyphyletic, false Baker and Gerber (1968), Forman (1971) on grouping of the nectar-feeders into a single the stomach morphology of bats, and Wille subfamily. Walton (1967), on the basis of (1954) on glossophagine tongues, observed postcranial osteology, suggested that the glos- basic morphological differences between sophagines arose as a group from a Vampy- Glossophaga-like and Choeroniscus-like rops-like stenodermatine ancestor. Walton glossophagines. Forman, Baker and Gerber did not suggest a polyphyletic origin for the tentatively suggested that there was a rela- Glossophaginae, although he did mention tionship between Choeronycteris and several that Glossophaga-like bats resemble steno- non-glossophagines (Chrotopterus, Phyllo- dermatines much more than do Choero- stomus, and Desmodus). nycteris-like bats. Interestingly, Walton (1967) Only two works deal comprehensively with believed that the Brachyphyllinae arose from the relationships within the subfamily Glos- a primitive Macrotus-like ancestor. Walton, sophaginae. These are the work of Phillips in suggesting that the relationship between (1971) on tooth and basicranial morphology, the brachyphyllines and glossophagines is not and the work by Gardner (1977) on karyol- close, was at variance with the proposals of ogy. Both papers divide the Glossophaginae 1 982 GRIFFITHS: NECTAR-FEEDING BATS 5

IE -Choeronycteris Musonycteris I Choeroniscus -Platalina *Lichonycteris -Hylonycteri s -Scleronycteris . Lionycteris -Loncho ph yl la *Anoura -Leptonycteris Glossophaga Monophyl lus FIG. 2. Relationships among the genera of the Glossophaginae. Redrawn from Phillips (1971). into essentially the same discrete groups (see genera of glossophagines. Second, and of figs. 1 and 2), but in the proposed relationship great interest, Platalina, Lichonycteris, Hy- of these groups to one another, these two lonycteris, Scleronycteris, Lionycteris, and papers differ considerably. Phillips (197 1) Lonchophylla were placed with the Choero- divided the Glossophaginae into two major nycteris-group, whereas Phillips (197 1) placed groups: (1) a group containing Choeroniscus, these bats with the Glossophaga-group. Both Choeronycteris, and Musonycteris which he researchers agreed that the subfamily Glos- called the "Choeronycteris-group"; and (2) a sophaginae might not be monophyletic. It is large group containing the remaining 10 glos- of further interest that Gardner (1977) ten- sophagine genera which he called the "Glos- tatively placed the Brachyphyllinae with the sophaga-group." On the basis of molar con- glossophagine genera Leptonycteris, Glos- figurations, he then divided the latter group sophaga, and Monophyllus in his phylogeny. into three subgroups: (1) Glossophaga, If the Glossophaginae are monophyletic, Monophyllus, and Leptonycteris; (2) Loncho- however, Gardner (1977) concluded that all phylla, Lionycteris, and Anoura; and (3) Scle- modern glossophagine karyotypes could have ronycteris, Lichonycteris, Hylonycteris, and been derived from a primitive karyotype Platalina. Of these three subgroups, the An- similar to the present day Lonchophylla tho- oura subgroup was considered most primi- masi karyotype. tive, or most like the hypothetical glos- The use of the hyoid and lingual regions: sophagine ancestor that Phillips proposed. It is obvious from the preceding discussion The Platalina subgroup was considered most that there are differences of opinion on the derived. phylogeny of the Glossophaginae. Much of Gardner (1977) generally followed Phillips the confusion has resulted from lack of rec- (1971) in his breakdown of the Glossophag- ognition of the fact that it is incorrect to use inae into subgroups (see fig. 1). There were shared primitive characters (symplesio- two basic differences. First, Anoura was morphies) to unite taxonomic groups. With placed in its own group, separate from other few exceptions (e.g., Baker and Bass, 1979), glossophagines. This placement reflected no effort has been made to differentiate de- Gardner's belief that Anoura has had a sep- rived from primitive character states. Un- arate evolutionary history from the other fortunately, the few studies that have done 6 AMERICAN MUSEUM NOVITATES NO. 2742 so have been hampered by a lack of suitable ease to the reader, but see the Systematic specimens ofsome ofthe rarer glossophagine Conclusions for my arrangement): genera. Previous work on the hyoid region of bats Family Phyllostomidae (Sprague, 1943; Wille, 1954; Winkelmann, Subfamily "Glossophaginae" 1971; and Griffiths, 1978a, 1978b) demon- Glossophaga soricina: Texas Tech Uni- strated that the hyoid muscles controlling the versity (TTU) Nos. 3326, 3327, 3328, tongue are markedly modified in the Glos- 3329, 3330. All from Mexico, Ta- sophaginae, apparently to permit the tongue maulipas, La Gruta de Quintero. Uni- to be extended and manipulated for nectar- versity of Michigan Museum of Zo- feeding. Work on the tongue structure of ology (UMMZ) 11 1278, 111280, the Leptonycteris-Glossophaga-Monophyl- 111294,111300,111301,111302.All lus group (Wille, 1954; Winkelmann, 1971; from Costa Rica, Cartago, 1 mi. S Car- Greenbaum and Phillips, 1974; Howell and tago, Agua Caliente. Hodgkin, 1976; and Griffiths, 1978a) and the Monophyllus redmani: University of Choeronycteris group (Wille, 1954; Winkel- Massachusetts (UMA) 2252, 2258. mann, 1971) has demonstrated that intrinsic Both from Haiti, Dept. du Sud, Reyn- and extrinsic tongue structure is distinctly olds Station above Miragoane, 2000 different in the two groups ofglossophagines, ft. and that in both groups these structures are Leptonycteris sanborni: UMMZ 122935, very different from those of non-glossopha- 122936,122937,122938,122939. All gine bats. Until the current study, however, from Arizona, Cochise Co., 9 mi. SW the majority of the genera of glossophagines San Simon. had not been examined. Lichonycteris obscura: National Mu- Modifications for nectar-feeding are so seum ofNatural History, Smithsonian marked (Griffiths, 1978a) that they appear Institution (NMNH) 519892. From to be irreversible without abandoning the Panama, Darien, Tacarcuna Village, nectar-feeding niche. Stated another way, all 1900 ft. NMNH 432194. Collection modifications for nectar-feeding are clearly locality unknown. apomorphies, and thus ideal for use in a taxo- Anoura geoffroyi: UMMZ 108640, nomic study. The sole problem, then, is to 108641, 108653, 108655, 108680, determine which are true shared, derived 108681. All from Chiapas, 1 mi. S characters (synapomorphies, and useful in Tuxtla Gutierrez. taxonomy) and which are convergent char- Hylonycteris underwoodi: American acters (misleading unless recognized as such). Museum ofNatural History (AMNH) Specialized mechanisms for tongue hyper- 238199. From Panama, Darien, Cerro extension have evolved independently in the Mali, 1400 m. NMNH 506578. From Mammalia several times (Doran and Baggett, Costa Rica, Heredia Prov., Finca La 1971), and although the overall mechanism Salva, Rio Puerto Viejo. in every known case is similar, there are ob- Choeroniscus godmani: NMNH 522934. servable morphological differences among From Venezuela, Yaracuy, San Felipe, them. Thus the question of the monophyly Puente Marroquina. of the taxon Glossophaginae should be re- Choeronycteris mexicana: UMMZ solvable by observations of the hyoid and 77750, 77751. From Mexico, Sonora, lingual regions to answer the question: did a Pilares. UMMZ 77755, 77756. From morphology facilitating nectar-feeding evolve Mexico, Sonora, El Tigre Mtns. more than once in the phyllostomid bats? UMMZ 77760. From Mexico, So- nora, St. Marie Mtns. Lonchophylla robusta: UMMZ 114923, MATERIALS AND METHODS 114927, 114928, 114929, 114930, Fluid preserved specimens ofthe following 114933. All from Costa Rica, Limon, species were dissected (species are listed here Los Diamantes, 4 mi. N Headquar- in the traditional taxonomic classification for ters, 250 m. 1 982 GRIFFITHS: NECTAR-FEEDING BATS 7

Lionycteris spurrelli: NMNH 499771. Additionally, the basicranial region and Colombia, Antioquia, Zaragoza. teeth of the following specimens were ex- Platalina genovensium: NMNH 268766. amined: Peru, Carivelli. SUBFAMILY Brachyphyllinae (=Phyllonyc- Glossophaga soricina: AMNH 92234, terinae) 92235, 92236, 92237, 92238, 92239, Brachyphylla cavernarum: Personal col- 92240, 92241, 92242, 92631, 92700. lection, T. Griffiths (TAG) 2, and G. longirostris: AMNH 130665. UMA 3086. Both from Puerto Rico. Monophyllus redmani: AMNH 19106, Erophylla sezekorni: AMNH 164255, 19107, 23782, 23783, 41157. 164281. From Bahamas, New Provi- Lichonycteris obscura: NMNH 331258, dence, Hunt's Cave. 335187, 362595, 364348, 483374. Phyllonycteris poeyi: AMNH 176023. Leptonycteris nivalis: NMNH 88017, From Cuba, Habana, 3 mi. E Tapaste, 88018, 88019, 88026. Cueva del Indio. TAG 1, 3. Haiti? Leptonycteris curasoae: NMNH 101850, Subfamily Phyllostominae 105130 (alcoholic with skull re- Macrotus waterhousii: TTU 12484, moved), 434424. 12485,12486, 12487, 12488. All from Anoura geoffroyi: NMNH 88022, 88023, Arizona, Pinal Co., 25 mi. S Casa 92260, 92263, 92420, 319248, Grande-Old Mammon Mine. Univer- 319250, 319251, 323182. sity of Vermont (UVM) 2372, 2374. Anoura caudifer: NMNH 483371, From Haiti, Dept. du Sud, 1 mi. SE 499308, 499309, 499310. Duchity, 2400'. UVM 2939, 2410. Hylonycteris underwoodi: NMNH Haiti, Dept. du Sud. AMNH 120972, 331260, 337984. 120977. From Dominican Rep., Cha- Scleronycteris ega: NMNH 407889. von, E of La Romana, caves near the Lonchophylla thomasi: AMNH 209358. river. NMNH 361570, 361571, 393013, Phyllostomus hastatus: AMNH 202308. 460097. Trinidad. Lonchophylla hesperia: NMNH 498827, Micronycteris nicefori: UMA 2819. From 498828, 498829, 498830, 498831. Panama. Lonchophylla handleyi: NMNH 507172. Subfamily Stenodermatinae Lionycteris spurrelli: AMNH 97220, Artibeusjamaicensis: UVM 1651, 1656. 97221, 97222, 97224, 97260, 97261, From Haiti, Dept. du Sud. UVM 97264. NMNH 499303, 499304, 3501, 3503. From Haiti, Dept. du 499305, 499306. Sud, 6 km. SW Miragoane, 580 m. Platalina genovensium: NMNH 268765 Phyllops haitiensis: UVM 2750. Haiti, (alcoholic with skull removed). Dept. du Sud. Because of small size, muscles of many Uroderma bilobatum: UMA 3034. Pan- specimens were exceedingly fragile. Forma- ama, Colon Prov., Santa Rosa. lin-preserved specimens were less prone to Vampyressa pusilla: UMA 3334. Pan- muscle breakage than those specimens which ama Canal Zone, Bohio Point Ridge. had been stored in isopropyl alcohol for a Vampyrops helleri: UMA 2695. Trini- long period. A binocular dissecting micro- dad, St. George Co., Simla, 5 mi. N scope was used for all dissections. Complete Arima. drawings ofall dissections and other anatom- Subfamily Carolliinae ical preparations were made at a scale of Carollia perspicillata: UMA 3060. Pan- either lOX or 5X the natural size. From these, ama Canal Zone, 2 km. N Frijoles. selected drawings showing specific anatomi- Subfamily Desmodontinae cal differences were inked for presentation. Desmodus rotundus: AMNH 208895, Tongue sectioning for all species examined 208899, 208902. All from Mexico, followed classic histological technique as de- Oaxaca, S Felipe del Agua (Cerro S. scribed in Humason (1972). Tongues were Felipe). excised just anterior to the hyoid bone (ba- 8 AMERICAN MUSEUM NOVITATES NO. 2742

sihyal), as close to the bone as possible. Each of Massachusetts Museum of Zoology; and tongue was then re-fixed in buffered formalin Dr. Charles A. Woods, University of Ver- solution, embedded in paraffin, and sec- mont Museum of Zoology. tioned. Staining was done with Mallory Tri- This work was supported in part by Uni- ple Connective Tissue Stain, and/or with he- versity Fellowship #87015 from the Uni- matoxylin and eosin. All measurements versity of Massachusetts. Illinois Wesleyan given to identify particular tongue segments University generously provided logistical are in microns from the tip of the tongue. support in the preparation ofthe manuscript, In the descriptive sections of this paper, plus technical support from Pat Meyers and the morphology of Glossophaga soricina is the Media Center in mounting the illustra- described in great detail. All other glos- tions. Special thanks to Judith Switzer for sophagines, and then all other bats dissected assistance. are described in less detail in separate sec- Finally, I will always be grateful to the late tions under each muscle. At the end of each Dr. Herbert Potswald for his friendship, sup- muscle description is a section entitled port, and interest in my work. "Comments." This section is designed to emphasize the functional and systematic high- DESCRIPTION OF THE MUSCLES lights of each muscle, points to be especially remembered when reading the Systematic BRANCHIOMERIC MUSCULATURE Conclusions. MYLOHYOID GROUP ACKNOWLEDGMENTS The muscles of this group are innervated by the mylohyoid nerve, a branch ofN. man- This work was part ofthe author's doctoral dibularis, which is a branch ofN. trigeminus dissertation at the Univeristy of Massachu- (V). The mandibular nerve enters both ofthe setts. I am indebted to the members of my muscles described on their deep surfaces. doctoral committee, Drs. David Klingener, Margery Coombs, Karl Koopman, Charles Pitrat, Dana Snyder, and Donald Kroodsma, M. MYLOHYOIDEUS for their thoughtful guidance during the Figures 3, 6, 9, 11, 14, 16, 18 course ofthis study and for their careful read- ORIGIN: In Glossophaga soricina, from the ing of the manuscript. I am grateful to Dr. medial surface of the mandible. The origin James Dale Smith for his rigorous review of extends for much of the length of the man- the manuscript. I thank my fellow graduate dibular ramus, from a point about 1 mm. students from the University of Massachu- posterior to the symphysis to the angle ofthe setts Museum ofZoology for their friendship jaw. and encouragement while this study was in INSERTION: Into its antimere, forming a progress, particularly Jeanne Bertonis, Car- raphe along the ventral midline of the jaw roll Schloyer, Doug Smith, and William Wall. region. I am grateful to my wife, Cara-Sue, and my OTHER GLOSSOPHAGINES: The muscle is the daughters Jennifer and Anne for their sup- same in all glossophagines dissected. In some port, and for putting up with me during the individuals, the mylohyoid is quite thin an- tough moments of preparing this work. teriorly, to the point of revealing the genio- I thank the following individuals and in- hyoid beneath. This anterior thinness is vari- stitutions for providing me with specimens: able within a species; however, in all cases Dr. Karl F. Koopman, American Museum the mylohyoid consists of a solid sheet of of Natural History; Dr. Charles 0. Handley, muscle. National Museum ofNatural History, Smith- OTHER BATS: In the three genera of brach- sonian Institution; Dr. Robert J. Baker, The yphyllines, the origin and insertion of the Museum, Texas Tech University; Dr. Philip mylohyoid are the same, but there is a slight Myers, University of Michigan Museum of break in the muscle anteriorly. This has the Zoology; Dr. David J. Klingener, University effect of dividing the mylohyoid into thick 1 982 GRIFFITHS: NECTAR-FEEDING BATS 9

-'--M. masseter sup.

M. digastricus

M. styloglossus M. mand.-hby.

M. omohyoideus M. hyoglossus

M. sternomastoideuss- XX M. sternohyoideus

FIG. 3. Ventral view of the superficial hyoid musculature of Glossophaga soricina. Bar = 1 mm. anterior and posterior parts, separated by a sected had the most robust mylohyoid mus- fleshy aponeurosis. In all three phyllosto- cle I have ever observed on a bat. mines, there is a pronounced break, which COMMENTS: Obviously, there are two quite results in the mylohyoid having a distinct different conditions for this muscle in the anterior and a distinct posterior part. This is Phyllostomidae: (1) a single, non-divided also the case in all five stenodermatine genera mylohyoid found in all glossophagines, Ca- dissected. In Carollia and Desmodus, the rollia, and Desmodus; and (2) a divided my- mylohyoid is a single sheet. This is particu- lohyoid found in all phyllostomines and larly true in Carollia; the specimen I dis- stenodermatines. In the Brachyphyllinae, 10 AMERICAN MUSEUM NOVITATES NO. 2742

M. I -M. genioglossus

M. S4-M. styloglossus

/M. digastricus M. I

M. M. geniohyoideus, deep insertion

- M. sternomastoid.

FIG. 4. Ventral view of the deep hyoid musculature of Glossophaga soricina. Bar = 1 mm. there is an aponeurosis anteriorly where the gines. The primitive condition is obviously muscle becomes quite thin, but the muscle the undivided one, found in most bats (see is essentially unbroken as in the glossopha- Sprague, 1943; Griffiths, 1978a, 1978b, and 1 982 GRIFFITHS: NECTAR-FEEDING BATS I1I

M. geniohyoideus,deep insert. M. ceratohyoideus M. styloglossus

auditory bulla .LV, fu

thyrohyal M.jugulohyoideus M. cricothyroideus M. stylopharyngeus M. thyrohyoideus tracheal_---- M. sternothyroideus

FIG. 5. Ventral view of the hyoid apparatus and larynx of Glossophaga soricina. Bar = 1 mm. in press). The divided condition could very longer directly act on the hyoid apparatus, its well be a synapomorphy shared by the Phyl- function of lateral pull was assumed by this lostominae and Sternodermatinae. small muscle that I suggest may have origi- nally been part of the posterior mylohyoid. M. MANDIBULO-HYOIDEUS The mandibulo-hyoid is found in all phyl- Figures 3, 4, 6, 7, 9, 10-12, 14, 16, 18, 19 lostomids, and thus is of no value for the M. mylohyoideus profundus (Griffiths, 1978a). taxonomic problem at hand. ORIGIN: In Glossophaga soricina, from the HYOID CONSTRICTOR GROUP medial surface of the mandible, just dorsal and slightly posterior to the origin ofthe pos- The muscles of this group are innervated terior edge of the mylohyoid. by branches of N. facialis (VII), some of INSERTION: This muscle passes deep to the which are extremely small and difficult to anterior digastric and hyoglossus muscles, to trace completely. insert on the ventrolateral surface of the basihyal bone, and into the lateral and dorsal M. STYLOHYOIDEUS fibers of the sternohyoid. ORIGIN: In Glossophaga soricina, from the OTHER GLOSSOPHAGINES: The muscle is the medial surface of the anterior '/2-1 mm. of same in all glossophagines dissected. In some the stylohyal bone. The origin ofthis muscle individuals, the insertion into the fibers of is very closely associated with the origin of the sternohyoid is reduced or absent, but this the stylopharyngeus. is variable within genera. INSERTION: On the posterolateral tip of the OTHER BATS: This muscle is the same in thyrohyal and on the lateral basihyal. all bats dissected. OTHER GLOSSOPHAGINES: This muscle is COMMENTS: The mandibulo-hyoid muscle similar in Monophyllus and Lichonycteris. In obviously developed in the phyllostomid Leptonycteris this muscle is variable. In bats in response to the freeing of the ster- about halfthe specimens dissected it was ab- nohyoid-hyoglossus complex from the basi- sent. In the other half, it was reduced to a few hyal. Since the mylohyoid proper could no fibers embedded in fascia. In Lonchophylla, 12 AMERICAN MUSEUM NOVITATES NO. 2742

M. myl ohyoideus l A\ M. geniohyoideus-/ I -LlL M. genioglossus

M. omohyoid. M. styloglossus M. hyoglossus

M.mond.-hyoid.

M.digastricus

M. sternohyoideus

M. sternomastoid.

FIG. 6. Ventral view of the superficial hyoid musculature of Lonchophylla robusta. Bar = 1 mm.

Lionycteris, Platalina, Anoura, Hylonycteris, OTHER BATS: This muscle is present in Choeroniscus, and Choeronycteris, this mus- Phyllonycteris and Erophylla, but completely cle was absent in all specimens dissected. absent in Brachyphylla. In Phyllonycteris, 1982 GRIFFITHS: NECTAR-FEEDING BATS 13

M. geniohyoideus L il---M. genioglossus M. styloglossus / N. hypoglossus M.hyoglossus Il \\M. geniohyoideus, M. hyoglossus_ deepIXtinsertio M. digastricus M. mand.-hyoid.

M. sternomastoideus M. sternohyoideus

FIG. 7. Ventral view of the deep hyoid musculature of Lonchophylla robusta. Bar = 1 mm. this is a well-defined muscle that originates In Erophylla this muscle is very reduced; it from the medial stylohyal and passes medi- inserts on the ceratohyal only. In all phyllos- ally and anteriorly to insert by tendon on the tomines and stenodermatines dissected, this hyoid bone at the ceratohyal/epihyal joint. muscle is absent. The muscle is absent in 14 AMERICAN MUSEUM NOVITATES NO. 2742

Desmodus, but present (though reduced) in at most of a few fibers that pass laterally in the specimen of Carollia dissected. In Ca- fascia, to insert weakly on the inner skin sur- rollia, the insertion is on the tip of the thy- face of the cervical region. rohyal. OTHER GLoSSOPHAGMIES: This muscle is COMMENTS: Sprague (1943) reported that similar in form in Monophyllus, Lichonyc- this muscle was absent in most Microchirop- teris, Leptonycteris, Anoura. In Hylonycteris tera he dissected, including all phyllostomids. reduction is almost complete, and in Choe- Obviously, the presence or absence of this roniscus and Choeronycteris no trace of this muscle is somewhat variable. Nevertheless, muscle could be observed. In Lonchophylla, Phyllonycteris, Erophylla, Glossophaga, Lionycteris, and Platalina this muscle was Monophyllus, Lichonycteris, Carollia and very different in form. It consisted of two some specimens of Leptonycteris share the distinct parts, both originating from the fas- plesiomorphic condition ofretaining the sty- cia of the posterior mylohyoid region as in lohyoid. All other phyllostomids are apo- Glossophaga. One part passed laterally to in- morphic in that they have lost the stylohyoid. sert on the inner surface of the cervical skin. The other part passed anterolaterally to insert M. JUGULOHYOIDEUS on the inner surface of the cervical skin at a Figures 5, 8, 13, 15, 17, 20 45-degree angle just lateral to the mandible. ORIGIN: In Glossophaga soricina, from the OTHER BATS: In Brachyphylla, the sphinc- basioccipital shelf, about 1 mm. posterior to ter colli profundus is very similar to the same the auditory bulla. muscle in Lonchophylla. This muscle is ab- INSERTION: This muscle passes ventrally sent in Erophylla and Phyllonycteris, except and anteriorly, curving around the posterior for fascial tracts that pass laterally where the surface of the auditory bulla to insert on the muscle used to be. In the phyllostomines dis- expanded lateral tip of the stylohyal bone. sected, this muscle is quite variable. In Mi- OTHER GLOSSOPHAGmIES: This muscle is cronycteris there are three distinct slips, all similar in Monophyllus, Lichonycteris, Lep- originating from the basihyal raphe region. tonycteris, Anoura, Hylonycteris, Choeronis- One slip passes anterolaterally, one passes cus, Choeronycteris, Lonchophylla, Lionyc- laterally, and one passes posterolaterally. teris, and Platalina. Macrotus and Phyllostomus both lack the slip OTHER BATS: This muscle is similar in all that passes posterolaterally, and thus have phyllostomids dissected. only two slips. All five genera of stenoder- COMMENTS: This muscle is comparatively matines have a sphincter colli profundus that more robust in all "glossophagine" genera is similar: it consists of two bellies that pass dissected. Functionally, this is to be expected, anteriorly, one anterior to the other. The an- because the jugulohyoid muscle must anchor terior slip is very similar to the anterior slip the distal end of the anterior hyoid comu in the phyllostomines. The posterior slip is against the pull of the styloglossus muscle similar to the slip in the phyllostomines that (see M. passes laterally, except that in the stenoder- styloglossus). matines it passes anteriorly as well as laterally. In Carollia this muscle is similar to the M. SPHINCTER COLLI PROFUNDUS condition observed in Macrotus and Phyl- This muscle is nominally a skin muscle, lostomus. In Desmodus, this muscle consists but it is included here because it takes origin only ofa single slip that passes anterolaterally from the basihyal region and is innervated from the basihyal region. by N. facialis. Sprague (1943) also included COMMENTS: The plesiomorphic condition this muscle in his work on the hyoid region is obviously having three distinct heads, of bats. passing anterolaterally, laterally, and pos- ORIGIN: In Glossophaga soricina, this mus- terolaterally (see Sprague, 1943). This is cle originates from the fascia of the posterior found only in Micronycteris of all the bats mylohyoid region. dissected here. In Macrotus, Phyllostomus, INSERTION: This muscle is extremely re- all the stenodermatines, Carollia, Brachy- duced in all specimens examined, consisting phylla, Lonchophylla, Lionycteris, and Pla- 1 982 GRIFFITHS: NECTAR-FEEDING BATS 15

M. geniohyoideus, deep insert.

M. sternothyroideus FIG. 8. Ventral view of the hyoid apparatus and larynx of Lonchophylla robusta. Bar = 1 mm. talina the posterior slip ofthis muscle shows between the hyopharyngeus and thyropha- a progression ofloss, from only a single weak ryngeus muscles. lateral slip to complete loss. In Erophylla and OTHER GLOSSOPHAGINES: This muscle is Phyllonycteris, this muscle is absent. Other similar in Monophyllus, Lichonycteris, Lep- than the distinctive "both bellies passing an- tonycteris, Anoura, Hylonycteris, Choeronis- teriorly" found in all the Stenodermatinae, cus, Choeronycteris, Lonchophylla, Lionyc- this muscle is so variable it seems to be of teris, and Platalina. minimal use in a systematic study. In the OTHER BATS: This muscle is similar in all glossophagines, it is interesting that more bats dissected. "advanced" nectar-feeders have lost this muscle, whereas Lonchophylla, Lionycteris, and Platalina have one ofthe more primitive M. CERATOHYOIDEUS conditions observed. This argues against the Figures 5, 8, 13, 15, 17, 20 placement of these three genera with the ORIGIN: In Glossophaga soricina, from more "advanced" glossophagines, as has the anterior edge ofthe thyrohyal bone (lesser been done in the past (see Systematic Con- hyoid cornu), and from the fibers of the thy- clusions). rohyoid muscle. INSERTION: This muscle passes anteriorly GLOSSOPHARYNGEAL GROUP to insert on the medial tip of the stylohyal, on the epihyal, and on the lateral tip of the The muscles of this group are innervated ceratohyal bones. by branches of N. glossopharyngeus (IX). OTHER GLOSSOPHAGINES: This muscle is similar in Monophyllus, Leptonycteris, and M. STYLOPHARYNGEUS Anoura. In Lichonycteris, Hylonycteris, Figures 5, 8, 13, 15, 17, 20 Choeroniscus, Choeronycteris, Loncho- ORIGIN: In Glossophaga soricina, from the phylla, Lionycteris, and Platalina it inserts posteromedial border of the stylohyal bone, on the entire posterior surface of the epihyal at a point about 2 mm. from the lateral tip. and the lateral tip of the ceratohyal (not on INSERTION: This muscle passes medially to the medial stylohyal at all). insert in the fibers of the pharyngeal wall, OTHER BATS: In all three brachyphylline 16 AMERICAN MUSEUM NOVITATES NO. 2742

M. mylohyoideus M. geniohyoideus

M. masseter suP.

M. mand.- hy.

M. stytoglossus

M. digastricus M. hyoglossus

M. sternomastoidGus M. sternohyoideus

FIG. 9. Ventral view of the superficial hyoid musculature of Choeronycteris mexicana. Bar = 1 mm. genera, this muscle inserts on the ceratohyal Glossophaga). In Macrotus and Phyllosto- and epihyal only. In the Phyllostominae and mus, the insertion is only on the epihyal and Stenodermatinae the muscle insertion varies stylohyal, with the reduced ceratohyal play- considerably among genera, with what ap- ing no part. In the Stenodermatinae, three pears to be great taxonomic significance. In different conditions are seen. In Vampyressa Micronycteris, the insertion is on the cera- pusilla and Vampyrops helleri, insertion is on tohyal, epihyal, and medial stylohyal (as in all three anterior cornu elements, as in Glos- 1 982 GRIFFITHS: NECTAR-FEEDING BATS 17

M. genioglossus- M. geniohyoideus

-' --M. masseter sup.

M. styloglossus M. mandibulo-hyoid.

M. geniohyoideus, deep M. digastricus insertion (cut) M. hyoglossus

M. sternomastoideus M. sternohyoideus

FIG. 10. Ventral view of the deep hyoid musculature of Choeronycteris mexicana. Bar = 1 mm. sophaga. In Artibeus and Phyllops, insertion is difficult to ascertain with certainty here due is on the ceratohyal and epihyal only. In Uro- to the wide variability of this muscle. It is derma, insertion is on the ceratohyal and lat- interesting to note that the three distinct con- eral basihyal. In Carollia, the insertion is as ditions observed in the stenodermatines dis- in Artibeus, and in Desmodus the insertion sected correspond very well to the branches is as in Macrotus. ofthe stenodermatine phyletic tree presented COMMENTS: The plesiomorphic condition by Baker (1973). 18 AMERICAN MUSEUM NOVITATES NO. 2742

PHARYNGEAL CONSTRICTOR GROUP M. THYROPHARYNGEUS The muscles of this group are innervated M. constrictor pharyngeus medius, Sprague, 1943; by branches of the N. vagus (X) as follows: middle constrictor, Novick and Griffin, 1961 N. laryngeus cranialis innervates M. cricothyroideus; N. recurrens innervates Mm. hy- ORIGIN: In Glossophaga soricina, from the opharyngeus, thyropharyngeus, and crico- dorsal surface ofthe tip ofthe thyrohyal bone pharyngeus. (posterior hyoid cornu). INSERTION: This muscle passes dorsally, M. CRICOTHYROIDEUS then medially to. insert into the dorsal pha- Figures 5, 8, 13, 15, 17, 20 ryngeal midline and into its antimere. ORIGIN: In Glossophaga soricina, from the OTHER GLOSSOPHAGINES: This muscle is entire ventral surface ofthe cricoid cartilage, similar in all glossophagines dissected. medial to the origin of the cricopharyngeus. OTHER BATS: This muscle is similar in all INSERTION: This muscle passes anteriorly bats dissected. and laterally fanning out to insert on the ventral and ventrolateral thyroid cartilage walls, M. CRICOPHARYNGEUS and on the ventral edge ofthe short posterior thyroid cornu. M. constrictor pharyngeus inferior, Sprague, 1943; OTHER GLOSSOPHAGINES: This muscle is inferior constrictor, Novick and Griffin, 1961 similar in Monophyllus, Lichonycteris, Lep- ORIGIN: In Glossophaga soricina, the mus- tonycteris, Anoura, Hylonycteris, Choeronis- cle originates in several slips from the lateral cus, Choeronycteris, Lonchophylla, Lionyc- border of the cricoid cartilage and from the teris, and Platalina. posterior thyroid cornu. OTHER BATS: This muscle is similar in all INSERTION: The posterior slips pass dor- bats dissected. sally to insert on the pharyngeal midline, and COMMENTS: In all phyllostomids, this mus- in their antimere. The anterior slips pass dor- cle is very simple structurally (unlike mor- sally and anteriorly to also insert on the pha- moopids and vespertilionids, see Griffiths, ryngeal midline and their antimeres. 1978b, in press). This is associated, no doubt, OTHER GLOSSOPHAGINES: This muscle is with the diminished role played by echolo- similar in Monophyllus, Lichonycteris, Lep- cation in food gathering. tonycteris, Anoura, Hylonycteris, Choeronis- M. HYOPHARYNGEUS cus, and Choeronycteris. In Lonchophylla, Lionycteris, and Platalina there were only M. constrictor pharyngeus superior, Sprague, two slips. 1943; superior constrictor, Novick and Griffin, OTHER BATS: In the Brachyphyllines the 1961 muscle is similar, though there are only two This muscle is almost completely absent slips present. In all the phyllostomines and in Glossophaga soricina; only a few muscle stenodermatines dissected, there are three fibers of what may be the remnants of this slips present in this muscle, as is the case in muscle were found attached to the dorsal Carollia. In Desmodus, this muscle consists buccopharyngeal fascia. This is also the case of a single large slip. in Monophyllus redmani, Lichonycteris ob- COMMENTS: Note that all "glossophagines" scura, Leptonycteris sanborni, Anoura geof- except Lonchophylla, Lionycteris, and Pla- froyi, Hylonycteris underwoodi, Choeronis- talina are similar. cus godmani, Choeronycteris mexicana, Lonchophylla robusta, Lionycteris spurrelli, and Platalina genovensium. MYOTOMIC MUSCULATURE OTHER BATS: This muscle is similar (al- LINGUAL GROUP most completely absent) in all brachyphyllines dissected, and in all other phyllostomids The muscles of this group are innervated dissected. by the N. hypoglossus (XII). 1982 GRIFFITHS: NECTAR-FEEDING BATS 19

M. mylohyoideuss

M. geniohyoideus

. styloglossus M. omohyoideus M. hyoglossus

M. digostricus. M. mand.-hyoid

M. sternom ostoid. M. sternohyoideus

FIG. 11. Ventral view of the superficial hyoid musculature of Brachyphylla cavernarum. Bar = 1 mm.

M. GENIOGLOSSUS icina this muscle curves laterally as it ap- Figures 4, 6, 7, 10, 12 proaches the basihyal bone. It inserts on the ORIGIN: In Glossophaga soricina, from the posterior 3 mm. of the venter of the tongue, medial surface of the anterior 1 mm. of the and does not insert on the basihyal. mandible, deep to the origin of the genio- OTHER GLoSSOPHAGINES: This muscle is hyoid. similar in Monophyllus, Lichonycteris, Lep- INSERTION: This muscle passes posteriorly tonycteris, Anoura, Hylonycteris, Choeronis- to insert into the ventral surface ofthe tongue cus, and Choeronycteris. In Lonchophylla just lateral to the tongue midline. In G. sor- and Lionycteris this muscle begins inserting 20 AMERICAN MUSEUM NOVITATES NO. 2742

M. geniohyoideus M. styloglossus N. hypoglossus_... M. genioglossus M. mand.-hyoid.. -M. hyoglossus

LM. digastricus

M. sternomostoideus"" ) 'M. sternohyoideus

FIG. 12. Ventral view of the deep hyoid musculature of Brachyphylla cavernarum. Bar = 1 mm. much farther anteriorly on the venter of the ventral tongue surface. It is interesting to note tongue. Instead of inserting in only the pos- that in all three brachyphyllines, this muscle terior 3 mm., this muscle inserts into the pos- seems to be in several parts, or slips, that are terior half of the ventral tongue surface, just easily separable. In all other phyllostomids, lateral to the midline. In Platalina, the ge- this muscle inserts over much of the ventral nioglossus inserts about halfway between tongue surface. these two extremes, inserting into the pos- CoMMENTs: Once more, all the "glos- terior 7 mm. of the tongue. sophagines" except Lonchophylla, Lionyc- OTHER BATS: In the Brachyphyllinae, this teris, and Platalina are united by a synapo- muscle inserts into much ofthe length of the morphous character. In this case, the 1982 GRIFFITHS: NECTAR-FEEDING BATS 21

basihyal M. ceratohyoideus

M. jugulohyoid. A. stylopharyngeus sternothyroideus

FIG. 13. Ventral view of the hyoid apparatus and larynx of Brachyphylla cavernarum. Bar = 1 mm. condition seen in Lonchophylla/Lionycteris bats, however, are covered completely by the could be "ancestral" to the condition in the expanded insertion of the geniohyoideus, other glossophagines, or it could be indepen- which forms a "tunnel" that encompasses the dently derived. The genioglossus ofPlatalina hyoglossus (see M. geniohyoideus). The ba- is most easily derived from the Loncho- sihyal raphe has almost completely disap- phylla/Lionycteris condition. There seems to peared, and the former break between the be an interesting synapomorphous condition hyoglossi and sternohyoidei is not visible in uniting the three brachyphyllines, too: the any of these six genera, except as the point separability of this muscle into slips. at which the geniohyoideus inserts. OTHER BATS: In the Brachyphyllinae, this muscle originates by tendon from the basi- M. HYOGLOSSUS hyal bone. In all other phyllostomids dis- Figures 3, 4, 6, 7, 9, 10-12, 14, 16, 18, 19 sected, this muscle originates by a very short ORIGIN: In Glossophaga soricina, from the tendon from the basihyal bone. former basihyal raphe (now disconnected COMMENTS: See M. geniohyoideus Com- from the basihyal bone); or more accurately, ments. from the insertion ofthe sternohyoideus. The hyoglossus and sternohyoideus have fused to M. STYLOGLOSSUS form a single muscle that passes unbroken Figures 3-20 from the sternum to the tongue. ORIGIN: In Glossophaga soricina, from the INSERTION: This muscle passes anteriorly expanded distal tip ofthe stylohyal, and from to insert in the posterior tongue surface, me- the posterior 4 mm. of the lateral surface of dial to the insertion of the styloglossus. the stylohyal bone. OTHER GLOSSOPHAGINES: The hyoglossus INSERTION: This muscle passes anteriorly is similar in Monophyllus, Lichonycteris, and somewhat medially, to insert in the pos- Lonchophylla, Lionycteris, and Platalina. In terolateral "corner" of the tongue. The fibers Leptonycteris, Anoura, Hylonycteris, Choe- of this muscle intermingle (at right angles) roniscus, and Choeronycteris the origin and with those of the hyoglossus. insertion are the same. The hyoglossi ofthese OTHER GLOSSOPHAGINES: This muscle is 22 AMERICAN MUSEUM NOVITATES NO. 2742

M. omohyoldeus > 3 M. styloglossus N. hypoglossus M. hyoglossus M. mand.-hyoid. M. digastricus

M. sternohyoidous

M. sternomostoideus

FIG. 14. Ventral view of the hyoid musculature of Phyllonycteris poeyi. Bar = 1 mm. similar in Monophyllus, Lichonycteris, Lep- or is composed of two distinct heads (Des- tonycteris, Anoura, Hylonycteris, Choeronis- modus AMNH 208902). These seem to be cus, and Choeronycteris. In Lonchophylla, peculiarities of individual specimens, how- Lionycteris and Platalina the insertion ofthis ever, and not entire species. The apomor- muscle is not in the posterior "corner" ofthe phous condition observed here (the posterior tongue, but rather into the lateral tongue sur- insertion of this muscle) is once again found face, as in the bats below. in all "glossophagines" other than Loncho- OTHER BATS: In the Brachyphyllinae, the phylla, Lionycteris, and Platalina. origin of this muscle is similar to that of the Glossophaginae. The insertion is into the lateral surface of the tongue for much of the MEDIAL VENTRAL CERVICAL GROUP tongue length. This is true of all phyllosto- The muscles of this group are innervated mids dissected. by an anastomosis of nerves made up of the COMMENTS: Occasionally, this muscle is anterior cervical nerves, except for M. ge- very reduced (Leptonycteris AMNH 122936), niohyoideus which is apparently innervated 1982 GRIFFITHS: NECTAR-FEEDING BATS 23

M. stylopharyngeus sternothyroideus

FIG. 15. Ventral view of the hyoid apparatus and larynx of Phyllonycteris poeyi. by N. hypoglossus (XII). Despite the appar- Choeronycteris, and Choeroniscus the ge- ent differing innervations, these muscles are niohyoideus is very different. The origin is treated as a group on the basis of similar essentially the same, though it is expanded embryonic differentiation (Edgeworth, 1916). laterally on the medial mandible. The muscle passes posteriorly in the same fashion as in Glossophaga, and splits into a superficial and M. GENIOHYOIDEUS a deep insertion. The deep insertion is on the Figures 3-12, 14, 16, 18, 19 anterior face of the basihyal, as in Glos- ORIGIN: In Glossophaga soricina, by ten- sophaga. However, the superifical insertion don from the medial surface ofthe mandible is not relatively weak, as in Glossophaga. at a point approximately 2 mm. lateral to the Rather it is very well developed and strong, symphysis. inserting in a loop around the ventral and INSERTION: There are actually two separate dorsal surfaces of the intersection of the hy- insertions for this muscle: a superficial in- oglossus and sternohyoideus muscles. To do sertion and a deep insertion. As this muscle this, the geniohyoideus literally forms a "tun- passes posteriorly, it splits. The superficial nel" around the hyoglossus, enclosing it al- fibers pass ventral to the basihyal and con- most completely (see figs. 9 and 10). tinue posteriorly to insert in the fibers of the OTHER BATS: In the Brachyphyllinae, this hyoglossus, and in the fibers of the sterno- muscle inserts by tendon to the basihyal. In hyoideus via the former basihyal raphe. This all other phyllostomids, this muscle inserts insertion is relatively weak. The deep fibers via short tendon to the basihyal. of this muscle insert directly on the anterior COMMENTS: There seem to be two apo- surface ofthe basihyal bone; this insertion is morphous states in the glossophagine bats: relatively strong. (1) where the geniohyoid has a strong, deep OTHER GLOSSOPHAGINES: The geniohyoi- basihyal bone insertion, and a weak, super- deus is similar in Monophyllus, Lichonyc- ficial basihyal raphe insertion; and (2) where teris, Lonchophylla, Lionycteris and Plata- the geniohyoid has a strong insertion in both lina. In Leptonycteris, Anoura, Hylonycteris, places. In no. 2, the geniohyoid accomplishes 24 AMERICAN MUSEUM NOVITATES NO. 2742

.|.s.'jUIUIIyuIUeu5 tM. styloglossus M. omohyoideus M. hyoglossus M. mand.-hyoid.

M. digasmn.hosidLs

\\\a M. sternohyoideus M. sternomostoid.

FIG. 16. The hyoid musculature of a phyllostomine bat, Macrotus waterhousii. Bar = 1 mm. this strong insertion by forming a tunnel nycteris, Anoura, Hylonycteris, Choero- which envelops the hyoglossus. Condition niscus, and Choeronycteris. no. 2 could be derived from condition no. 1, or the two could have independently evolved. In any event, the tunnel insertion (condition M. STERNOHYOIDEUS no. 2) is definitely an apomorphous condition Figures 3, 4, 6, 7, 9-12, 14, 16, 18 that has never been reported for any other ORIGIN: In Glossophaga soricina, from the mammal group. It therefore most strongly dorsal surface of the xiphoid process of the links the genera that share this trait: Lepto- sternum. 1982 GRIFFITHS: NECTAR-FEEDING BATS 25

stylohyaolN 1. styloglossus

auditory bull M. jugulohyoid. M. stylopharyngeus sternothyroi deus

FIG. 17. Ventral view ofthe hyoid apparatus and larynx ofa phyllostomine bat, Macrotus waterhousii. Bar = 1 mm.

INSERTION: This muscle passes anteriorly from the lateral manubrium and proximal to insert by raphe in the fibers ofthe hyoglos- head ofthe clavicle. In Erophylla, the medial sus and in the superficial fibers of the ge- fibers alone have a shifted origin. In both niohyoideus muscle. There is no connection Brachyphylla and Phyllonycteris, the medial to the hyoid apparatus. fibers take origin from the anterior body of OTHER GLOSSOPHAGINES: The sternohyoi- the sternum as well, but also the more lateral deus is similar in Monophyllus, Lichonyc- fibers have a shifted origin too. They take teris, Lonchophylla, Lionycteris, and Plata- origin from the proximal head ofthe first rib, lina. In Leptonycteris, Anoura, Hylonycteris, instead of from the lateral manubrium. In Choeroniscus, and Choeronycteris the origin both Brachyphylla and Phyllonycteris, there of the sternohyoid is similar to that in Glos- are extremely weak, lateralmost fibers that sophaga. The insertion is in the fibers of the retain the origin from the medial head ofthe hyoglossus, and in the highly modified cir- clavicle. In all three genera, the insertion of cular insertion of the geniohyoideus (see M. the sternohyoid is via tendon to the basihyal geniohyoideus). In one individual of Lon- bone. chophylla robusta and Lionycteris spurrellii In all other phyllostomids except Desmo- there was a remnant of the basihyal tendon dus, this muscle originates from the dorsal present. surface of the manubrium of the sternum, OTHER BATS: In the Brachyphyllinae, this and from the medial head ofthe clavicle and muscle originates from a point slightly pos- sterno-clavicular articulation. In Desmodus, terior to the "normal" origin, although not the origin is similar to that of Phyllonycteris. shifted as far posteriorly as in the glos- COMMENTS: Obviously, the posterior shift sophagines. In Erophylla, medial fibers of of the origin of the sternohyoid muscle is an this muscle originate from the anterior body apomorphous condition, as in the insertion ofthe sternum (not the manubrium), whereas into the fibers of the hyoglossus, rather than more lateral fibers of this muscle originate the basihyal bone. All three brachyphyllines, 26 AMERICAN MUSEUM NOVITATES NO. 2742

M.my . geniohyoideus -M. styloglossus

M.dii N. hypoglossus

M. sternohyoi hyoideus

sternomastoideus

FIG. 18. The superifical hyoid musculature of a stenodermatine bat, Vampyressa pusilla. Bar = I mm. including Brachyphylla, share a slightly pos- this muscle is the same in the Brachyphylli- terad shifted origin of this muscle. The nae and in all phyllostomids dissected. question of whether the shifts in the glossophagines are synapomorphies or due to con- M. OMOHYOIDEUS vergence is discussed below. Figures 3, 6, 11, 14, 16 in M. STERNOTHYROIDEUS This muscle is reduced all glossophagines and absent in many specimens. Figures 5, 8, 13, 15, 17 ORIGIN: In Glossophaga soricina, from the ORIGIN: In Glossophaga soricina, from the deep surface of the bone of the scapula im- dorsal surface ofthe medial clavicle, just lat- mediately surrounding the scapular notch. eral to the sterno-clavicular articulation. INSERTION: This muscle passes ventrally INSERTION: This muscle runs anteriorly to and medially, curving around the cervical insert on the lateral surface of the thyroid area deep to both the sternomastoid and clei- cartilage, just posterior to the origin of the domastoid muscles. It inserts weakly in the thyrohyoideus muscle. fibers ofthe hyoglossus and mylohyoid mus- OTHER GLOSSOPHAGINES: The sternothy- cles. roideus is similar in Monophyllus, Lich- OTHER GLOSSOPHAGINES: This muscle is onycteris, Leptonycteris, Anoura, Hylo- similar in Monophyllus (though slightly more nycteris, Choeroniscus, Choeronycteris, Lon- robust), Lichonycteris, Leptonycteris, Anoura chophylla, Lionycteris, and Platalina. (more reduced), Hylonycteris (more re- OTHER BATS: The origin and insertion of duced), Choeroniscus, and Choeronycteris 1 982 GRIFFITHS: NECTAR-FEEDING BATS 27

1. geniohyoldeus

M. hyogil M. digasl

M. thyrohyoid hyoideus

M. cricothyroideus

FIG. 19. The deep hyoid musculature of a stenodermatine bat, Vampyressa pusilla. Bar = 1 mm.

(extremely reduced). In Lonchophylla and INSERTION: This muscle passes anteriorly, Platalina this muscle is much more robust. fanning out as it does so, to insert on the This muscle was not observed in the single thyrohyal bone (posterior hyoid cornu) and specimen of Lionycteris dissected, but this partially in the fibers of the ceratohyoideus was probably due more to the poor state of muscle. preservation of the hyoid region. OTHER GLOSSOPHAGINES: The thyrohyoi- OTHER BATS: In the Brachyphyllinae, the deus is similar in all glossophagine bats dis- origin and insertion of this muscle are into sected. the stemohyoid and mylohyoid muscles, OTHER BATS: The thyrohyoideus is similar though this muscle appeared to be absent in in all bats dissected. one specimen ofErophylla (AMNH 164281). In the remaining phyllostomids this muscle TONGUE MORPHOLOGY AND was variable within species. When present, HISTOLOGY the origin and insertion were the same as in Brachyphylla. EXTERNAL MORPHOLOGY COMMENTS: Because of the extreme vari- GLOSSOPHAGINAE: To avoid confusion and ation observed, this muscle is unreliable as facilitate comparison, I follow the terminol- a taxonomic indicator. ogy used by Greenbaum and Phillips (1974) in their work on the tongues of Leptonycteris M. THYROHYOIDEUS sanborni and L. nivalis. The tongues of all Figures 5, 8, 13, 15, 17 "glossophagines" except Lonchophylla, ORIGIN: In Glossophaga soricina, from the Lionycteris, and Platalina are similar (see lateral surface of the thyroid cartilage, just figs. 21, 27, 28, and 29). Greenbaum and anterior to the insertion of the sternothy- Phillips (1974) identified seven types of pa- roideus muscle. pillae on the tongues of Leptonycteris: hair- 28 AMERICAN MUSEUM NOVITATES NO. 2742

M. ceratohyoideus ceratohyal M. stylopharyngeus epihyal M. styloglossus hyroh yl

bulla M. jugulohyoideus stylohyal

FIG. 20. Ventral view ofthe hyoid apparatus and larynx of a stenodermatine bat, Vampyressa pusilla. Bar = 1 mm. like papillae, horny papillae, bifid papillae, low). The circumvallate papillae are present singly-pointed papillae (two types, fleshy and in all "glossophagines" except Lonchophylla, hooklike), fungiform papillae, and circum- Lionycteris, and Platalina, where they are vallate papillae. I agree completely with the absent. In Glossophaga, Monophyllus, Li- classification of papilla types as proposed by chonycteris, and usually Leptonycteris, there Greenbaum and Phillips (1974). See figure are four circumvallate papillae: two lateral 21 for relative placement of the various pa- and two medial. In every specimen dissected, pilla types on the tongue. the lateral papillae were larger than the me- The hairlike papillae are distributed over dial ones. In one specimen of Leptonycteris the anterolateral surface of the tongue in all sanborni, one medial circumvallate papilla "glossophagines" except Lonchophylla, was missing. In every specimen of Anoura, Lionycteris, and Platalina. These papillae Hylonycteris, Choeroniscus, and Choeronyc- form an anterior "brush," presumed by teris dissected only the two large lateral cir- Greenbaum and Phillips (1974) and Howell cumvallate papillae were present. and Hodgkin (1976) to be used in nectar- In Lonchophylla, Lionycteris, and Plata- feeding (see fig. 21). The horny papillae are lina the external tongue morphology is com- found in all "glossophagines" except Lon- pletely different. There are only two papilla chophylla, Lionycteris, and Platalina on types present: short hairlike papillae and either side ofthe dorsal midline ofthe tongue. fleshy monofid papillae. Additionally, there Their arrangement, shown in figure 22, is of is a deep, longitudinal groove in the lateral some interest. Generally, they consist of two surface of the tongue from a point just pos- or more large, posteriorly-directed papillae, terior to the tip to the base ofthe tongue. The surrounded by several smaller papillae. The short hairlike papillae do not form a brush shape and number of horny papillae seems tip, but rather form a line just dorsal to and constant within a genus, but varies between just ventral to the longitudinal groove. The genera except in the cases of Monophyllus! fleshy monofid papillae coat the dorsal Glossophaga and Hylonycteris/Choeronis- tongue surface. All other papilla types are cus/Choeronycteris. The bifid papillae, sin- absent. gly-pointed papillae, and fungiform papillae BRACHYPHYLLINAE: The tongues of Ero- are distributed as shown in figure 21, and are phylla and Phyllonycteris are similar to one common to all "glossophagines" except Lon- another, and to the tongues of the Glosso- chophylla, Lionycteris, and Platalina (see be- phaga-type glossophagines. The tongue of 1982 GRIFFITHS: NECTAR-FEEDING BATS 29

HLP HP

MCP

H P

H L P

FIG. 21. Upper left: the tongue of Macrotus waterhousii, dorsal surface. Upper right: the tongue of Lionycteris spurrelli, dorsal surface. Lower left: the tongue of Glossophaga soricina, dorsal surface. Lower right: the tongue ofChoeronycteris mexicana, dorsal surface. Bar = 1 mm. MCP = median circumvallate papilla, LG = longitudinal groove, HP = horny papillae, HLP = hairlike papillae. 30 AMERICAN MUSEUM NOVITATES NO. 2742

oQ3o °oo o

A B C D

°qUo0 000 00000

E F G H

I J FIG. 22. The horny papillae of the bats examined in this paper. A. Brachyphylla, B. Phyllonycteris, C. Glossophaga, D. Monophyllus, E. Lichonycteris, F. Leptonycteris, G. Anoura, H. Hylonycteris, I. Choeroniscus, J. Choeronycteris. Bar = 1 mm.

Brachyphylla is morphologically different. OTHER BATS: The three phyllostomine gen- The Brachyphylla tongue has horny papillae, era are similar. All three have horny papillae, singly-pointed papillae (mostly fleshy), fun- singly-pointed (mostly fleshy monofid) pa- giform papillae, and circumvallate papillae. pillae, fungiform papillae, and circumvallate It does not have hairlike papillae at the tip, papillae (two lateral and two medial). Figure and no bifid papillae were observed. Figure 21 shows the placement of these papillae on 23 shows the arrangement of papillae on the the tongue surface. The horny papillae and tongue of Brachyphylla. There are only two the circumvallate papillae are virtually iden- medial (and no lateral) circumvallate papillae tical in Phyllostomus and Micronycteris. The present. horny papillae in these two genera form a In Erophylla and Phylionycteris, the tongues distinctive 3/4 row (three anterior and four are similar to that of Brachyphylla in some posterior, see fig. 24A) and the circumvallate respects, and to those ofthe Glossophaga-type papillae are all large and well developed. In "glossophagines" in others (see fig. 23). No Macrotus, the horny papillae show a less or- bifid papillae are present, but there are hair- derly appearance (fig. 24B), and the two lat- like papillae forming a brush tip similar to eral circumvallate papillae are very reduced. that of Glossophaga. Four circumvallate pa- All five stenodermatine genera have a gen- pillae are present in all specimens dissected, erally similar tongue morphology. All have though the medial two are always larger. In horny papillae, singly-pointed (mostly fleshy) all other respects, the tongues of Erophylla papillae, fungiform papillae, and four cir- and Phyllonycteris are similar to that of cumvallate papillae. They also all share a Brachyphylla. distinctive cluster ofbifid papillae located in 1982 GRIFFITHS: NECTAR-FEEDING BATS 31

FIG. 23. Dorsal surfaces of the tongues of Brachyphylla cavernarum (left) and Phyllonycteris poeyi (right). Bar = 1 mm. MCP = median circumvallate papilla, FP = fungiform papilla, HP = horny papilla, HLP = hairlike papillae. the posteromedian region of the dorsal riorly directed papillae, surrounded by sev- tongue surface (though this cluster is reduced eral smaller papillae (fig. 26). There are no in Vampyressa). The horny papillae of Uro- bifid papillae, but the remaining papilla types derma resemble those of Phyllostomus/Mi- of Glossophaga are present. There are four cronycteris in having the 3/4 pattern (fig. 24); circumvallate papillae. all other stenodermatines have a "less or- The tongue of Desmodus is, in some re- derly" Macrotus-type appearance (see fig. spects, most similar to the tongues of Lon- 25). Interestingly, the three tongue types ob- chophylla, Lionycteris, and Platalina. There served in the Stenodermatinae correspond are very few papilla types on the tongue sur- well to the three divisions of the Stenoder- face: essentially there are only fleshy monofid matinae proposed by Gardner (1977). papillae and perhaps a few fungiform pa- The tongue of Carollia is most similar to pillae. No other type is present. Additionally, the Glossophaga-type tongue in some char- there is a longitudinal groove along the lateral acters. There are no hairlike papillae, but the tongue surface, though this groove is not horny papillae consist of two large, poste- nearly as deep as the one in the Lonchophylla 32 AMERICAN MUSEUM NOVITATES NO. 2742

0 0000

A B

FIG. 24. The horny papillae of Phyllostomus hastatus (left) and Macrotus waterhousii (right). The major papillae are filled in to distinguish them from the minor papillae (ones that do not open posteriorly, see text). Bar = 1 mm. group of bats. There are no hairlike papillae U1\____ lining the groove in Desmodus, as there are in the Lonchophylla group. FIG. 26. Dorsal surface of the tongue of Ca- rollia perspicillata and an inset of the horny pa- INTERNAL HISTOLOGY pillae. Bars = 1 mm. MCP = median circumvallate papilla, MP = monofid papillae, HP = horny GLOSSOPHAGINAE: All glossophagines ex- papilla. amined except Lonchophylla, Lionycteris, and Platalina have a similar internal tongue morphology (see figs. 27 and 28). Anteriorly teries as in most mammals. To either side of in the tongue, there is a single, midline lingual the artery are two large lingual veins, con- artery, rather than left and right lingual ar- nected to the artery by artero-venous shunts. The horizontal lingual musculature passes around each of these lingual veins (fig. 27), enclosing them. Toward the tip ofthe tongue, branches of the lingual veins pass laterally and dorsally into the interior ofeach hairlike papilla. MCP The intrinsic tongue structure of the remaining three genera is shown in figure 29. B~~~~BP Most prominent are the deep longitudinal grooves on each side ofthe tongue. Undoubt- edly, the shape of these grooves is controlled by the complex bundles of skeletal muscle that pass in all directions within the tongue. There are left and right lingual arteries and veins, plus accessory arteries and veins in the HP dorsal part of the tongue. BRACHYPHYLLINAE: The intrinsic tongue structure of Phyllonycteris and Erophylla is very similar to that of the Glossophaga-like glossophagines. There is a single, midline lingual artery connected by shunts to two slightly enlarged lingual veins (fig. 31). The FIG. 25. Dorsal surface of the tongue of the intrinsic muscle bundles are also similar. In stenodermatine bat, Artibeusjamaicensis, with an Brachyphylla, however, the tongue structure inset of the horny papillae. Bars = 1 mm. is more similar to the non-nectar-feeder MCP = median circumvallate papilla, BP = bifid tongues described below (fig. 30). There are papillae, HP = horny papilla. two lingual arteries and veins, with no sign 1 982 GRIFHITHS: NECTAR-FEEDING BATS 33 LP

FIG. 27. Cross section of the tongue of Glossophaga soricina. Bar = 1 mm. LP = lingual papilla, LA = lingual artery, LV = lingual vein, SM = skeletal muscle. of the large shunts found in most glos- regions. Altthough they share certain hyoid sophagines. There is not a single, midline lin- characters Nwith the "other glossophagines" gual artery present. (posterior slhift ofthe sternohyoid origin and OTHER BATS: In all phyllostomines and developmerit of a sternoglossus complex by stenodermatines studied, the intrinsic lingual loss of the cconnection to the basihyal), I re- structure is the same. There are two lingual gard these ccharacters as convergent charac- arteries and veins, with predominantly hor- ters rather than as synapomorphies. The izontal skeletal muscle bundles passing same two clharacter states have occurred in- around them (fig. 32). In Carollia, these hor- dependentl)y in at least three other groups izontal muscle bundles are extraordinarily within the IMammalia (see Doran and Bag- well developed. In Desmodus, the tongue has gett, 1971, or Griffiths, 1978a) in animals predominantly horizontal muscle bundles. that must Ehyperextend the tongue to feed. There is a groove in the lateral tongue surface, The strongemst arguments for Lonchophylla, quite dissimilar to the groove of Loncho- Lionycteris, and Platalina having become in- phylla, Lionycteris, and Platalina in that it dependentlyy nectivorous can be seen in the is not so deep, and there are no complex tongue morlphology (see figs. 21, 27, and 29). muscle bundles inserting on it. The tongue,s of these three genera have lost most ofthe Ipapilla types found on the tongues SYSTEMATIC CONCLUSIONS of other phhyllostomids, most notably the horny and circumvallate types. Although The systematic relationships ofthe 16 gen- there are haairlike papillae present, they are era studied are presented in the cladogram, very differently arranged than in the Glos- figure 33. Tongue and hyoid apomorphies are sophaga-lik;e glossophagines. Rather than summarized in table 1. The traditional being conce,ntrated on the anterodorsal sur- subfamily "Glossophaginae" is clearly a di- face to form a brush tip, the hairlike papillae phyletic group, and the new subfamily are arrangedI in two lines, one line just dorsal "Brachyphyllinae" (proposed by Baker, 1979) and one line just ventral to a deep longitu- may not be a valid group. The three genera dinal groovee (see fig. 21) that runs the length Lonchophylla, Lionycteris, and Platalina, ofthe tonguie in all three genera. The internal heretofore regarded as glossophagines, have tongue struccture is also very different (com- morphologically distinctive tongue and hyoid pare figs. 277 and 29). In Lonchophylla, Lio- 34 AMERICAN MUSEUM NOVITATES NO. 2742

LA FIG. 28. Cross section ofthe tongue ofHylonycteris underwoodi. Bar = 0.5 mm. LP = lingual papilla, LA = lingual artery, LV = lingual vein, SM = skeletal muscle. nycteris, and Platalina, there are complex (already described), and by: (1) M. crico- bundles of muscle running in many direc- pharyngeus reduced to two, rather than three tions within the tongue, probably to control slips; and (2) the posterior shift of the origin the shape ofthe groove during nectar-feeding. of the sternohyoid and formation of a "ster- These complex bundles are not present in noglossus" complex. Note that certain other other phyllostomids. muscle modifications such as posterior shift It is most unlikely that a bat line that had of the insertion of the styloglossus, are pres- developed this morphologically complex ent in the other nectar-feeding line, but not tongue would lose these adaptations, and in this one. None of the characters I have then develop the equally complex tongue examined distinguish Lonchophylla from found in Phyllonycteris, Erophylla, and the Lionycteris, but Platalina has a slightly remaining glossophagines. It is therefore shifted insertion ofthe genioglossus, and thus likely that the common ancestor shared by I show it as distinct on the cladogram. Two Lonchophylla/Lionycteris/Platalina and the final points should be made about this line. other glossophagines was not a nectar-feeder, First, because I have demonstrated that there and that the two lines developed nectar-feed- are non-glossophagines that are more closely ing independently. I thus propose that the related to the Glossophaga-Choeronycteris genera Lonchophylla, Lionycteris, and Pla- group than are Lonchophylla, Lionycteris, talina are an independently derived line, and Platalina the last three genera must be united by the synapomorphies of the tongue placed in a separate taxon of equal status to 1982 GRIFFITHS: NECTAR-FEEDING BATS 35

FIG. 29. Cross section of the tongue of Lonchophylla robusta. Bar = 1 mm. DA = dorsal artery, LA = lingual artery, LP = lingual papillae, LV = lingual vein, LG = longitudinal groove, SM = skeletal muscle. the Glossophaga-Choeronycteris group. I classification below). Second, there may be thus must raise a new subfamily (see formal other phyllostomids besides Brachyphylla, 36 AMERICAN MUSEUM NOVITATES NO. 2742

FIG. 30. Cross section of the tongue ofBrachyphylla cavernarum. Bar = 1 mm. LA = lingual artery, LP = lingual papilla, LV = lingual vein, SM = skeletal muscle.

Phyllonycteris, and Erophylla that are also ronycteris group. Brachyphylla does not pos- more closely related either to my new sess any of the complex internal or external subfamily, or the subfamily Glossophaginae tongue modifications observed in the other (sensu stricto). I have left the base of the 12 genera. However, ofall the bats examined cladogram open to signify this very likely in this study, Brachyphylla is the most similar possibility. to what the primitive, non-nectivorous The remaining 13 genera form what ap- ancestor of the other 12 might have looked pears to be a monophyletic group. The genus like (for example, it has the proper hyoid Brachyphylla is the only possible exception bone morphology). For this reason, and be- to this statement. Other than the small pos- cause Nagorsen and Peterson (1975), Silva terior shift of the origin of the sternohyoid, Taboada and Pine (1969), and Baker and there is no synapomorphy clearly uniting Bass (1979) have all suggested that there may Brachyphylla with the Phyllonycteris-Choe- be a close relationship between Brachyphylla 1982 GRIFFITHS: NECTAR-FEEDING BATS 37 LP

FIG. 31. Cross section of the tongue of Erophylla sezekorni. Bar = 1 mm. LA = lingual artery, LP = lingual papilla, LV = lingual vein, SM = skeletal muscle. and Phyllonycteris/Erophylla, I tentatively xiphisternum, and there is the elongation of place Brachyphylla at the base of the clado- the hyoglossus with the freeing of the ster- gram. The only other possible placement nohyoid-hyoglossus from the basihyal, form- would be on the same line with Phyllonyc- ing a new "sternoglossus" complex. Both of teris/Erophylla (it could be a non-nectivore these adaptations have been paralleled in the derived from a nectivorous ancestor), but I Lonchophylla line of nectar-feeders. How- have no evidence to support this view. I must ever, the following derived character states conclude that Brachyphylla is on a mono- have no parallels: (1) the styloglossus has an generic line that is a sister-group to the Phyl- insertion far posterior, on the posterior cor- lonycteris-Choeronycteris group. Brachy- ner of the tongue; and (2) the genioglossus phylla cannot be placed in the same subfamily inserts only into the posterior few mm. ofthe as Phyllonycteris/Erophylla, as proposed by tongue venter, rather than over the entire Baker (1979). For my arrangement, see my ventral surface. formal classification below, and see figure 33. After Branching Point 3, there is a single After Branching Point 1 on the cladogram, character uniting the group Leptonycteris- the remaining 12 genera are united by a Choeronycteris. Despite it being only a single strong set of synapomorphies. These are: (1) character, it is a good synapomorphy that development of a single, midline lingual ar- clearly unites Leptonycteris-Choeronycteris tery, large lingual veins, and shunts connect- as a holophyletic group. This character is the ing the two; (2) development of hairlike pa- tunnel insertion of M. geniohyoideus, where pillae in a particular pattern on the the geniohyoideii completely envelop the anterodorsal tongue surface (a "brush-tip," hyoglossi (fig. 9). This derived condition see figs. 21 and 23); and (3) a peculiar 2/1 could easily have developed from the genio- pattern of the major horny papillae (two pa- hyoid found in Glossophaga/Monophyllus! pillae anterior and one posterior) shared by Lichonycteris (or for that matter, from the Phyllonycteris, Erophylla, and the more geniohyoideus arrangement found in Lon- primitive Glossophaginae (sensu stricto). chophylla/Lionycteris/Platalina, though the After Branching Point 2, there is a very tongue morphology is very different). I sug- strong set of synapomorphies uniting Glos- gest that the expanded tunnel insertion ofthe sophaga-Choeronycteris. There is the poste- geniohyoid in the Leptonycteris-Choeronyc- rior origin of the sternohyoideus from the teris group would facilitate elongation of the 38 AMERICAN MUSEUM NOVITATES NO. 2742

FIG. 32. Cross section of the tongue of Micronycteris nicefori. Bar = 0.5 mm. LA = lingual artery, LP = lingual papilla, LV = lingual vein, SM = skeletal muscle.

contracted sternohyoid, thus aiding in ex- respect and possibly should be combined in tending the tongue maximally (see fig. 10). the genus Glossophaga (Varona, 1974, does Regardless of the functional reason for de- this). It is not surprising that the characters veloping the tunnel insertion, the condition of the hyoid and tongue regions fail to sep- is apparently unique to this group; it has not arate them. The placement of Lichonycteris even been reported in the various other is more ofa problem. Lichonycteris has been mammal groups that hyperextend the tongue placed with Platalina/Hylonycteris/Sclero- (Doran and Baggett, 1971). nycteris (Gardner, 1977; and see Smith, The other line to the left at Branching Point 1976); and with Hylonycteris alone (Phillips, 3 must also be justified since there is more 1971). Additionally, there are certain dental than one genus on the line. All three genera, similarities to Leptonycteris, and Thomas Glossophaga, Monophyllus, and Lichonyc- (1895) suggested in the original generic de- teris possess the apomorphies at Branching scription that Lichonycteris resembled both Point 2, but do not have the tunnel insertion Glossophaga and Choeronycteris, especially synapomorphy of Branching Point 3. It is the latter. The characters I have used in this clear that all three genera are placed at the study suggest that Lichonycteris should not proper grade ofglossophagine evolution, but be placed after Leptonycteris or before Phyl- unfortunately, the character states of the lonycteris/Erophylla. I have placed it as an hyoid and tongue regions do not help here early offshoot of the Glossophaga/Mono- to sort out the genera cladisticly. Glossophaga phyllus line, rather than choosing the other and Monophyllus are very similar in every option of placing it on its own monogeneric 1982 GRIFFITHS: NECTAR-FEEDING BATS 39

TABLE 1 Summary of the Apomorphies Used in Constructing the Nectar-Feeding Bat Cladogram (fig. 33) + = presence of the character; 0 = absence of the character; +/0 = variable. Derived Character State Br Ph Er GI Mo Li Le An Hy Ch Cht Lo Lio P1 1) post. shift of sternohyoid origin + + + + + + + + + + + + + + 2) xiphoid origin of sternohyoid 0 0 0 + + + + + + + + + + + 3) loss of sternohyoid's connection to hyoid bone 0 0 0 + + + + + + + + +/0 +/0 + 4) hyoglossus elongated and loses connection to hyoidbone 0 0 0 + + + + + + + + + + + 5) double insertion of geniohyoid 0 0 0 + + + + + + + + + + + 6) tunnel insertion of geniohyoid 0 0 0 0 0 0 + + + + + 0 0 0 7) post. shift of styloglossus insertion 0 0 0 + + + + + + + + 0 0 0 8) post. shift of see genioglossus insertion 0 0 0 + + + + + + + + + + text 9) loss ofthe stylohyoid + 0 0 0 0 0 +/0 + + + + + + + 10) reduction of sphc. col. prof. 0 0 0 0 0 0 0 0 + + + 0 0 0 11) cricopharyn. reduced to two bellies 0 0 0 0 0 0 0 0 0 0 0 + + + 12) groove in tongue lined w/ hairlike papillae 0 0 0 0 0 0 0 0 0 0 0 + + + 13) almost complete loss of ling. papillae 0 0 0 0 0 0 0 0 0 0 0 + + + 14) brush tip formed by hairlike papillae 0 + + + + + + + + + + 0 0 0 15) two medial CV papillae absent 0 0 0 0 0 0 +/0 + + + + see text 16) horny papillae in 3/4 pattern 0 0 0 0 0 0 0 0 + + + 0 0 0 17) single midline ling. art. 0 + + + + + + + + + + 0 0 0 18) enlarged ling. veins 0 + + + + + + + + + + 0 0 0 line either after or before the Glossophagal suggest that the placement of Anoura in this Monophyllus line. This decision is based not position should be regarded as tentative). The on lingual morphology, but rather for karyo- synapomorphies after Branching Point 6 are: typic and dental reasons discussed in the next (1) complete loss of the two medial circum- section. vallate papillae; and (2) complete loss of M. After Branching Point 6, there are only stylohyoideus. Ashlock (1974) quite correctly minor characters uniting the Anoura-Choe- pointed out that generally one should be sus- ronycteris group (and there are some dental picious of "loss" characters on a cladogram. characters discussed in the next section that There is little question that Anoura belongs 40 AMERICAN MUSEUM NOVITATES NO. 2742

Cl) l) Cl) __ = Cl cl 0) 0)

.() m 01 C (.* CU mC C- >% 0 _ ,. Q CL C -;u > C.)% 0 Cl) a C) *_ ._CO- > 0 -c Q 0Cl0 co m X 0 2 0 C L- 0 0-3 -Ji a: CI 0. w aY 2

FIG. 33. Cladogram showing the relationships of the nectar-feeding genera discussed herein. See text for characters used.

with Leptonycteris, Hylonycteris, Choero- auditory bullae posteriorly. There are other niscus, and Choeronycteris but its placement dental and karyotypic data uniting this group, on the cladogram merely shows the point it discussed below. seems to fit best and, as with Lichonycteris, The remaining three genera are all closely should be regarded as tentative. related, probably having become generically After Branching Point 7, there are once distinct comparatively recently. The char- again several strong synapomorphies that in- acters of the hyoid and lingual regions are dicate that the Hylonycteris-Choeronycteris insufficient to distinguish between them, and group is monophyletic. Scleronycteris and so I must rely on dental and karyotypic ob- Musonycteris exist only as study skins in col- servations, discussed below. lections, so obviously I have not dissected them. I place them near what I believe are SUPPORTING EVIDENCE FROM their closest relatives, based on my examination of their teeth and basicranial regions. OTHER DISCIPLINES Synapomorphies of the entire group include: The analyses of the karyotypic and dental (1) a strong reduction of the sphincter colli data have produced apparently conflicting profundus muscle, and (2) the development phylogenies for the subfamily Glossophag- ofa complex 4/3 pattern ofthe horny papillae inae (see figs. 1 and 2). The purpose of this (see fig. 22) that is unique to this group of section is to reexamine these data, and show glossophagines. Additionally, Phillips (1971) that they support the cladogram presented reported, and I confirm here, that the genera here as well as they support the phylogenies of the Hylonycteris-Choeronycteris group constructed by Phillips (1971) and Gardner share an unusual and highly derived basi- (1977). cranial morphology, where the pterygoid pro- Upper Incisors: Evidence from the exam- cesses are strongly inflected and contact the ination of the glossophagine upper incisors 1982 GRIFFITHS: NECTAR-FEEDING BATS 41

(Winkelmann, 1971; Phillips, 1971) and my teris, and the Hylonycteris-Choeronyc- own personal observations support the clado- teris group. If my cladogram is correct, the gram I present here. The incisors of Loncho- upper canines would have had to become re- phylla, Lionycteris, Platalina, and Glos- duced independently three times: once in sophaga are the least reduced, or most Platalina, once in Lichonycteris, and once in primitive. They resemble the upper incisors the Hylonycteris-Choeronycteris group. There of Brachyphylla, Phyllonycteris, and Ero- is no evidence for or against this view. phylla in that the inner incisors are notably Upper Premolars: The upper premolars are larger than the outer ones, and have broad all simple, primitive teeth in Lonchophylla, cutting edges (Phillips, 1971), whereas the Lionycteris, Glossophaga, Monophyllus, and outer incisors are more pointed. Monophyllus Leptonycteris (Phillips, 1971). Anoura has and Leptonycteris also have "relatively un- similar upper premolars, but possesses a per- reduced" upper incisors (Phillips, 1971), manent P2, which is probably a duplicated though the inner pair is reduced to the same P3 (Koopman, personal commun.), and thus size as the outer pair. Apparently because of not of great taxonomic significance. The re- this reduction, a small gap has developed maining genera (Platalina, Lichonycteris, between the two inner incisors of these two Hylonycteris, Scleronycteris, and the Choe- genera (see Winkelmann, 1971). The trend ronycteris/Choeroniscus group) all have spe- for reduction ofthe incisors continues through cialized upper premolars (Phillips, 1971). Anoura and Hylonycteris to Choeroniscus, The Choeronycteris/Choeroniscus line has Choeronycteris, and Musonycteris where the become extremely specialized, with the pre- inner incisors are small, peglike teeth sepa- molars reduced to long, thin teeth. Hylo- rated by a wide gap, and the outer incisors nycteris and Lichonycteris show similar re- are reduced as well. Interestingly, the upper ductions, with the loss ofvarious shelves and incisor configuration of Lichonycteris is an stylar elements, which caused Phillips (197 1) additional strong argument against placing to group the two genera together (see fig. 2). this genus with Hylonycteris, Choeroniscus, This would not be permitted in a cladistic or Choeronycteris. The upper incisors of Li- analysis, however, because as Phillips (1971) chonycteris are reduced, but in a different himself pointed out, the second upper pre- manner. The inner incisors are broad and the molar in Lichonycteris retains the primitive outer incisors are pointed, as in Glossophaga. "postero-lingual shelf' and the "small ante- There is almost no gap between the inner rior and posterior styles" as in the less de- incisors (though they are very reduced), as rived, Glossophaga-like teeth. I believe that there is in all the more advanced glos- the more primitive tooth morphology of Li- sophagines. The incisors of Lichonycteris chonycteris argues against including Licho- could be most easily derived from those of nycteris with the more advanced glossopha- Phyllonycteris, Erophylla, or Glossophaga gines. though, ofcourse, the condition in these bats Upper Molars: Phillips (1971) used differ- is primitive for the subfamily as a whole. The ences he observed in the configuration ofthe differently reduced incisors in Lichonycteris upper molars as a primary means of classi- can only be used to emphasize the lack of fication of the glossophagines. I suggest that relationship to the Hylonycteris-Choeronyc- the upper molar configurations support my teris group. Unfortunately, there is no good cladogram (fig. 33) at least as well as they incisor synapomorphy uniting the Glos- support his phylogenetic tree (fig. 2). Plata- sophaga/Monophyllus line and Lichonyc- lina, Hylonycteris, Scleronycteris, Choero- teris. niscus, and Choeronycteris all share the de- Upper Canines: In Lonchophylla, Lionyc- rived character ofobliteration ofthe anterior teris, Brachyphylla, Phyllonycteris, Ero- half of the W-shaped ectoloph (found in the phylla, Glossophaga, Monophyllus, Lepto- more "primitive" glossophagine genera). Re- nycteris, and Anoura, the canines are moval of Platalina from the group (which I unreduced, with a prominent cingular style have shown elsewhere to be not closely re- (Phillips, 1971, and personal observ.). The lated to the others) allows the use ofthis char- canines are reduced in Platalina, Lichonyc- acter as a synapomorphy uniting the genera 42 AMERICAN MUSEUM NOVITATES NO. 2742 after Branching Point 7 on my cladogram. ronycteris the molars are long and thin, with The primitive W-shaped ectoloph is promi- a "flange" on the postcristid (Phillips, 1971) nent in Glossophaga, Monophyllus, and An- which is present, though underdeveloped in oura but in Leptonycteris the anterior half is Choeroniscus. This character can be used to considerably straighter than in the Glos- unite Choeronycteris and Musonycteris above sophaga ectoloph. This straightening could Branching Point 9 on the cladogram. be considered a precursor to the condition in Basicranial Region: Phillips (1971) cor- the Hylonycteris-Choeronycteris group, rectly reported that the pterygoid processes though Anoura, which retains the prominent of Choeroniscus, Choeronycteris, and Mu- W-shaped ectoloph, would have to be moved sonycteris have an inflected tip that comes in out ofthe Leptonycteris-Choeronycteris group contact with the auditory bullae. Addition- if this character were found to be significant. ally, Hylonycteris has a slightly inflected re- Once again, Lichonycteris retains the primi- gion that could be considered a precursor to tive state of a character, which suggests that that above. Phillips (1971) used the pterygoid it does not belong among the more advanced character (and several upper molar charac- genera. ters) to divide the glossophagines into a Lower Incisors: The lower incisors are "Choeronycteris group" and a "Glossophaga completely absent in Lichonycteris, Anoura, group." The condition noted in the Choe- Hylonycteris, Scleronycteris, Choeroniscus, ronycteris group is clearly a synapomorphy Choeronycteris, and Musonycteris (Phillips, that unites Choeroniscus-Musonycteris, and 1971, and personal observ.). In Loncho- Hylonycteris-Musonycteris (if the condition phylla, Lionycteris, and Platalina, the lower in Hylonycteris is a valid precursor). Phil- incisors are very large, flat teeth (Phillips, lips's "Glossophaga group" is not a valid 1971; Winkelmann, 1971) with distinctive group because the pterygoids ofall these bats "trifid" crowns (Phillips, 1971). These dis- are primitive (except for Anoura, which in a tinctive incisors can be used as a synapo- few cases shows some pterygoid inflection). morphy uniting the above three genera. The Phillips (1971) was thus using a symplesio- lower incisors are present (primitive condi- morphic condition to unite a group. tion) in Glossophaga, Monophyllus, and Lep- Karyology: Unfortunately, because of the tonycteris. Except for the condition in Li- rarity of many of the species of the Glosso- chonycteris (which would have to be the phaginae, some genera have not been karyo- result of convergence), these data could be typically examined. Additionally, only a few interpreted as a synapomorphous character have been examined with the G and C band- (loss oflower incisors) uniting the genera after ing techniques now available (see Baker and Branching Point 6 on my cladogram. Bass, 1979). The karyotypic data available Lower Canines: The lower canines ofevery can be interpreted to support the cladogram genus to the right of Branching Point 7, plus presented here. Lichonycteris, are reduced. This reduction Baker (1967, 1973, 1979) proposed that could be considered a synapomorphy uniting the primitive karyotype for phyllostomid the Hyloncteris-Choeronycteris group. Again, bats is: diploid number (2n) = 30-32, fun- Lichonycteris would have had to develop the damental number (FN) = 56-60. If this be condition independently. the case, then Brachyphylla, Phyllonycteris, Lower Premolars: There is considerable Erophylla, Glossophaga, Monophyllus, and variation in premolar morphology among the Leptonycteris would share the least derived various genera (Phillips, 1971). In my opin- (32/60) karyotype within the nectar-feeder ion, the only good synapomorphy uniting a complex. The Lichonycteris karyotype (either group is the character oflong, thin premolars 28/50 or 24/44, Baker, 1979) is most similar found in Choeroniscus, Choeronycteris, and to, and most easily derivable from, the karyo- Musonycteris (see Phillips, 1971); this would type ofone ofthe above bats (I have suggested unite the group above Branching Point 8. from the Monophyllus/Glossophaga line), as Lower Molars: The lower molars of all Gardner (1977) illustrated. The remaining glossophagine genera are remarkably similar genera (Anoura-Choeronycteris), with succes- (Phillips, 1971). In Musonycteris and Choe- sively decreasing 2n and FN's (see Baker, 1982 GRIFFITHS: NECTAR-FEEDING BATS 43

1979) can be derived from this Glossophaga- Subfamily Phyllonycterinae like karyotype (Baker, personal commun.). Phyllonycteris The arrangement I present (fig. 33) also ex- Erophylla plains why the karyotypes of Lonchophylla Subfamily Glossophaginae and Lionycteris are so different from those Glossophaga of the "other glossophagines" (Platalina has Monophyllus not yet been karyotyped), as Gardner (1977) Lichonycteris showed. Baker (personal commun.) has Leptonycteris stated that it would take a large number of Anoura chromosomal rearrangements to derive the Hylonycteris karyotypes of Lonchophylla and Lionycteris Scleronycteris from the "other glossophagines"; this is eas- Choeroniscus ily explained by my suggestion that the "lon- Choeronycteris chophylline" bats are not closely related to Musonycteris the glossophagines (sensu stricto). NEW SUBFAMILY Summary and Formal Classification: Ex- Subfamily Lonchophyllinae cept for some problems in the dentition of Lonchophylla Lichonycteris, the karyotypic, dental, and Lionycteris hyoid/lingual data can be interpreted to sup- Platalina port the cladogram presented here (fig. 33). Because ofthe distinct differences in tooth I stress that although there may be problems morphology between the phyllonycterines with the exact placement ofLichonycteris and and glossophagines, I am not now in favor possibly Anoura, the available evidence of including Brachyphylla, Phyllonycteris, strongly supports the monophyly ofthe Glos- and Erophylla in the subfamily Glossophag- sophaginae (minus Lonchophylla, Lionyc- inae, though this may eventually be done teris, and Platalina), the distinctiveness and (perhaps at the Tribe level). For now, I have monophyly of the three lonchophylline gen- applied the traditional classification system era, the monophyly of the Phyllonycteris- to my results recognizing that Brachyphylla Choeronycteris group, the monophyly of the cannot be considered a member of the Phyl- Leptonycteris-Choeronycteris group, and the lonycterinae under the rules of cladistic clas- monophyly of the Hylonycteris-Choeronyc- sification. The only option available until teris group. The overall form of the clado- such time as it might be included in the Glos- gram is probably correct, though Lichonyc- is teris and Anoura may be slightly misplaced. sophaginae to relegate it to its own mono- I am opposed to the hierarchical classifi- generic subfamily Brachyphyllinae (name cation presented by many cladists, where proposed by Baker, 1979). each branch of the cladogram must be des- LITERATURE CITED ignated as a higher taxonomic group. Using Ashlock, Peter D. this method, making the slightest change on 1974. The uses of cladistics. Ann. Rev. Ecol. any level disturbs the stability of the classi- Syst., vol. 5, pp. 81-99. fication. I am also in favor of the principle Baker, Robert J. of disturbing the current classification the 1967. Karyotypes of bats of the family Phyl- least when making changes. The simplest lostomidae and their taxonomic impli- method of following both precepts is pre- cations. Southwestern Nat., vol. 12, pp. sented below. 407-428. 1970. The role of karyotypes in phylogenetic Family Phyllostomidae studies of bats. In B. H. Slaughter and D. W. Walton (eds.), About bats, South- Subfamily Phyllostominae ern Methodist Univ. Press, Dallas, pp. Subfamily Stenodermatinae 303-3 12. Subfamily Carolliinae 1973. Comparative cytogenetics of the New Subfamily Desmodontinae World leaf-nosed bats (Phyllostomati- Subfamily Brachyphyllinae dae). Periodicum Biologicum, vol. 75, Brachphylla pp. 37-45. 44 AMERICAN MUSEUM NOVITATES NO. 2742

1979. Karyology. In R. J. Baker, J. K. Jones, tology of tongues of long-nosed bats Jr., and D. C. Carter (eds.), Biology of (Leptonycteris sanborni and L. nivalis) bats of the New World family Phyllo- with reference to infestation of oral stomatidae. Part III. Spec. Publ. 16, The mites. Jour. Mammal., vol. 55, pp. Museum, Texas Tech Univ., Lubbock, 489-504. pp. 107-155. Griffiths, Thomas A. Baker, Robert J., and Rebecca A. Bass 1978a. Muscular and vascular adaptations for 1979. Evolutionary relationship ofthe Brachy- nectar-feeding in the glossophagine bats phyllinae to the glossophagine genera Monophyllus and Glossophaga. Jour. Glossophaga and Monophyllus. Jour. Mammal., vol. 59, pp. 414-418. Mammal., vol. 60, pp. 364-372. 1978b. Modification of M. cricothyroideus and Baker, Robert J., Rodney L. Honeycutt, Michael the larynx in the Mormoopidae, with L. Arnold, Vincent M. Sarich, and Hugh H. reference to amplification of high-fre- Genoways quency pulses. Ibid., vol. 59, pp. 1981. Electrophoretic and immunological 724-730. studies on the relationship of the [In press] Comparative laryngeal anatomy of Brachyphyllinae and the Glossophag- the big brown bat, Eptesicusfuscus, and inae. Jour. Mammal., vol. 62, pp. the mustached bat, Pteronotus parnellii. 665-672. Mammalia. Baker, Robert J., and G. G. Lopez Hall, E. Raymond 1970. Karyotypic studies of the insular popu- 1981. The Mammals of North America. vol. lations of bats on Puerto Rico. Caryo- 1, New York, John Wiley and Sons, logia, vol. 23, pp. 465-472. xv + 600 pp. + 90 pp. Dobson, G. E. Hall, E. Raymond, and Keith R. Kelson 1878. Catalogue of the Chiroptera in the col- 1959. The Mammals of North America. vol. lection of the British Museum. British 1, New York, Ronald Press, xxx + 546 Museum, London, xlii + 567 pp. [plates pp. + 79 pp. I-XXX]. Hennig, W. Doran, G. A., and H. Baggett 1966. Phylogenetic systematics. Univ. Illinois 1971. A structural and functional classification Press, Urbana, 263 pp. ofmammalian tongues. Jour. Mammal., Howell, Donna J., and Norman Hodgkin vol. 52, pp. 427-429. 1976. Feeding adaptations in the hairs and Edgeworth, Francis H. tongues of nectar-feeding bats. Jour. 1916. On the development and morphology of Morphol., vol. 148, pp. 329-336. the pharyngeal, laryngeal, and hypo- Humason, Gretchen L. branchial muscles of mammals. Quart. 1972. Animal Tissue Techniques. San Fran- Jour. Micro. Sci., London, n. s., vol. 61, cisco, W. H. Freeman and Co., xiv + 641 pp. 383-432. pp- Forman, G. L. Koopman, Karl F. 1971. Gastric morphology in selected mor- 1952. The status of the bat genus Reithronyc- moopid and glossophagine bats as re- teris. Jour. Mammal., vol. 33, pp. lated to systematic problems. Trans. Il- 255-258. linois Acad. Sci., vol. 64, pp. 273-282. Koopman, Karl F., and E. Lendell Cockrum Forman, G. L., R. J. Baker, and J. D. Gerber 1967. Bats. In S. Anderson and J. K. Jones, Jr. 1968. Comments on the systematic status of (eds.), Recent mammals of the world- vampire bats (family Desmodontidae). A synopsis offamilies. New York, Ron- Syst. Zool., vol. 17, pp. 417-425. ald Press, pp. 109-150. Gardner, Alfred L. Koopman, Karl F., and J. Knox Jones, Jr. 1977. Chromosomal variation in Vampyressa 1970. Classification ofbats. In B. H. Slaughter and a review ofchromosomal evolution and D. W. Walton (eds.), About bats. in the Phyllostomidae (Chiroptera). Syst. Southern Methodist Univ. Press, Dallas, Zool., vol. 26, pp. 300-318. pp. 22-28. Gerber, J. D., and C. A. Leone Miller, Gerrit S. 1971. Immunologic comparisons of the sera 1907. The families and genera ofbats. Bull. U. of certain phyllostomatid bats. Syst. S. Natl. Mus., vol. 57, xvii + 282 pp. Zool., vol. 20, pp. 160-166. Nagorsen, D. W., and R. L. Peterson Greenbaum, Ira F., and Carleton J. Phillips 1975. Karyotypes of six species of bats (Chi- 1974. Comparative anatomy and general his- roptera) from the Dominican Republic. 1982 GRIFFITHS: NECTAR-FEEDING BATS 45

Royal Ontario Mus., Life Sci. Occas. with especial reference to the bats. Papers, vol. 28, pp. 1-8. Amer. Jour. Anat., vol. 72, pp. 385-472. Novick, Alvin, and Donald R. Griffin Stock, A. Dean 1961. Laryngeal mechanisms in bats for the 1975. Chromosome banding pattern homol- production of orientation sounds. Jour. ogy and its phylogenetic implications in Exper. Zool., vol. 148, pp. 125-145. the bat genera Carollia and Choeronis- Phillips, Carleton J. cus. Cytogenet. Cell Genet., vol. 14, pp. 1971. The dentition of glossophagine bats: 34-41. Development, morphological character- Thomas, Oldfield istics, variation, pathology, and evolu- 1895. On small mammals from Nicaragua and tion. Univ. Kansas Mus. Nat. Hist., Bogota. Ann. Mag. Nat. Hist., ser. 6, vol. Misc. Publ., vol. 54, pp. 1-138. 16, pp. 55-60. Schaldach, W. J., and C. A. McLaughlin 1912. New bats and rodents from S. America. 1960. A new genus and species of glossopha- Ibid., ser. 8, vol. 10, pp. 403-411. gine bat from Colima, Mexico. Mus. 1913. A new genus of glossophagine bat from Nat. Hist., Los Angeles County, Con- Colombia. Ibid., ser. 8, vol. 12, pp. trib. Sci., vol. 37, pp. 1-8. 270-271. Silva Taboada, G., and Ronald H. Pine 1928. A new genus and species of glossopha- 1969. Morphological and behavioral evidence gine, with a subdivision of the genus for the relationship between the bat ge- Choeronycteris. Ibid., ser. 10, vol. 1, pp. nus Brachyphylla and the Phyllonycter- 120-123. inae. Biotropica, vol. 1, pp. 10-19. Varona, Luis S. Simpson, George Gaylord 1974. Catalogo de los Mamiferos Vivientes y 1945. The principles of classification and a Extinguidos de las Antillas. Academia classification of mammals. Bull. Amer. de Ciencias de Cuba, Habana, viii + 139 Mus. Nat. Hist., vol. 85, xvi + 350 pp. pp- Slaughter, Bob H. Walton, Dan Wells 1970. Evolutionary trends ofchiropteran den- 1967. Subfamilial relationships in the family titions. In B. H. Slaughter and D. W. Phyllostomatidae (Mammalia: Chirop- Walton (eds.), About bats, Southern tera) based on post-cranial osteology and Methodist Univ. Press, Dallas, pp. myology. Unpubl. Ph.D. Dissertation, 51-83. Tulane Univ. Smith, James Dale Wille, Alvaro 1972. Systematics of the chiropteran family 1954. Muscular adaptation of the nectar-eat- Mormoopidae. Univ. Kansas Mus. Nat. ing bats (subfamily Glossophaginae). Hist., Misc. Publ., vol. 56, pp. 1-132. Trans. Kansas Acad. Sci., vol. 57, pp. 1976. Chiropteran evolution. In R. J. Baker, 315-325. J. K. Jones, Jr., and D. C. Carter (eds.), Winkelmann, John R. Biology ofbats ofthe New World family 1971. Adaptations for nectar-feeding in glos- Phyllostomatidae. Part I. Spec. Publ., sophagine bats. Unpubl. Ph.D. Disser- 10, The Museum, Texas Tech Univ., tation, Univ. Michigan, Ann Arbor. Lubbock, pp. 49-69. Sprague, James M. 1943. The hyoid region ofplacental mammals