Morphology and homology of the chiropteran calcar, with comments on the phylogenetic relationships of _Archaeopteropus_

Morphology and Homology of the Chiropteran Calcar, with Comments on the Phylogenetic Relationships of Archaeopteropus William A. Schutt, Jr. 1–3 and Nancy B. Simmons 1 Most researchers have considered the calcar to be a unique and homologous structure within Chiroptera (e.g., the presence of this structure and its associated musculature has been cited as a synapomorphy supporting bat monophyly). However, we report that significant morphological variation exists between Microchiroptera and Megachiroptera. In microchiropterans, a calcified or cartilaginous element articulates directly with the calcaneal tuberosity of the ankle and projects into the uropatagium. In megachiropterans, a cartilaginous structure projects from the tendon of the gastrocnemius muscle into the uropatagium and has no articulation with the calcaneal tuberosity. Considerable variation also exists in the musculature associated with these structures. Phylogenetic interpretation of hindlimb morphology of extant and fossil taxa indicates that the calcar may not be homologous in all bats. We suggest retention of the term "calcar" for the microchiropteran structure and propose a new term, "uropatagial spur," for the megachiropteran structure. The fossil bat Archaeopteropus transiens (Oligocene) has long been presumed to be a megachiropteran; however, this form has a microchiropteran-type calcar. Reconsideration of morphological evidence from this and previous studies indicates that Archaeopteropus is not a megachiropteran but, rather, a basal member of the microchiropteran lineage. KEY WORDS: Archaeopteropus, calcar; Chiroptera; hindlimb morphology; Megachiroptera; Microchiroptera. The calcar is an element of doubtful homology. —Allen (1893) INTRODUCTION Many of the most obvious specializations for flight seen in bats (Chiroptera) involve the wing membranes and the structures that support them. The bat wing is composed of a flexible membrane (the patagium) that extends laterally from the body and is anchored anteriorly along the upper arm and forearm. Distally, this membrane spans the spaces between the elongated, bony elements of the hand. Posteriorly, the wing membrane inserts along the dorsolateral surface of the hindlimb and sometimes the foot. An asso- ' Department of Mammalogy, The American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024. 2 The Division of Natural Sciences and Math, Bloomfield College, Bloomfield, New Jersey 07003. 3To whom correspondence should be addressed at Department of Mammalogy, The American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024. e-mail: Wa- [email protected]g 1 1064-7554/98/0300-0001$15.00/0 © 1998 Plenum Publishing Corporation Journal of Mammalian Evolution, Vol. 5, No. 1, 1998 2 Schutt and Simmons ciated flight membrane, the uropatagium or interfemoral membrane, is present between the hindlimbs in both microchiropteran and megachiropteran bats. Most bats that have a uropatagium also possess an associated structure referred to as a calcar (derived from the Latin calx, heel). The first reference to a "calcar" is uncertain, but this term was employed by Allen (1893), Miller (1907), and Andersen (1912) in their taxonomic studies. Typically, the calcar has been described as a slender cartilaginous or bony spur that projects from the ankle region into the posterior margin of the uropatagium, where it terminates. In some bats this structure is extremely long and blade-like (e.g., Noctilio leporinus, Dicliduris scutatus), while it is minute or absent in other taxa (i.e., Craseonycteris, Rhinopoma, Diaemus, Mystacina, Syconycteris, Harpyionycteris, and Notopterus). The relative size of the calcar falls somewhere between these two extremes in most bat species. Researchers have employed variation in calcar size and form in taxonomic studies (e.g., Miller, 1907; Andersen, 1912; Findley, 1972; Pine, 1993), and the presence/absence of a calcar has been used as a character in phylogenetic analyses (e.g., Van Valen, 1979; Simmons, 1994, 1995, 1998; Simmons and Geisler, 1997, 1998). The calcar functions to help to spread the uropatagium and to adjust its camber during flight (Vaughan, 1959). Additionally, the calcar straightens and braces the posterior border of the uropatagium, preventing the trailing edge of the membrane from flapping (Vaughan, 1959, 1970a). The size of the uropatagium exhibits considerable variation among taxa, ranging from a broad sheet to narrow strips of skin that run along the medial borders of the legs. In many aerial insectivores (e.g., vespertilionids), the uropatagium is extensive and the membrane stretches between the hindlimbs and the tail. The uropatagium in these bats provides a substantial lift surface and functions during braking and turning maneuvers (Vaughan, 1959, 1970b; Norberg, 1990). The uropatagium, controlled in part by the position of the calcar, is also employed by some vespertilionid bats (i.e., Pipistrellus pipistrellus, P. nathusii, P. kuhlii, Myotis daubentoni) during aerial feeding to capture insects before transfer to the mouth (Webster and Griffin, 1962; Schnitzler et al., 1987; Kalko and Schnitzler, 1989; Kalko, 1995). The uropatagium may additionally serve as a "safety net" to cradle the newborn during parturition (Hill and Smith, 1984). In bats that trawl for prey using gaff-like hindfeet (i.e., Noctilio leporinus and Myotis vivesi), the calcar functions to draw the uropatagium away from the surface of the water (Hill and Smith, 1984). In contrast to bats with a well-developed uropatagium (e.g., aerial insectivores), the uropatagium is extremely narrow or rudimentary in some frugivorous, nectarivorous, and sanguinivorous bats (e.g., Syconycteris, Desmodus, Diaemus, Sturnira), extending like a fringe from the proximomedial surface of hindlimbs and tapering toward the foot. The calcar is typically very short or absent in taxa that have a narrow or rudimentary uropatagium. However, there are some exceptions. In Diphylla ecaudata (a desmodontine with a rudimentary uropatagium), a uniquely digitiform calcar extends some 3 mm past the trailing edge of the narrow uropatagium and is employed as a sixth digit to facilitate branch grasping during arboreal locomotion (Schutt and Altenbach, 1997). Craseonycteris thonglongyai, the smallest extant microchiropteran, is the only bat that has an extensive uropatagium but no calcar. Numerous studies have included partial or complete descriptions of the hindlimb Morphology and Homology of the Chiropteran Calcar 3 anatomy of microchiropterans (e.g., MacAlister, 1872; Maisonneuve, 1878; de Fenis, 1919; Vaughan, 1959, 1970b–d; Grasse, 1971) and/or megachiropterans (e.g., Humphry, 1869; MacAlister, 1872; Andersen, 1912; de Fenis, 1919; Mori, 1961; Grasse, 1971) (see Historical Overview, below). Cartilaginous or bony structures projecting medially from the ankle region into the uropatagium have been variously referred to as "supplementary calcaneal bones" (Humphry, 1869), "styliform bones" (Humphry, 1869; MacAlister, 1872), "spurs" (MacAlister, 1872), "cartilaginosa del calcagno" (Morra, 1899), "stylets" (de Fenis, 1919), "calcars" (Andersen, 1912; Vaughan, 1959, 1970b, c), "Fusswurzelstachels" (Mori, 1961), and "les eperons" (Grasse, 1971). Despite the fact that these studies discussed and illustrated structures that varied in composition, anatomical position, and associated musculature, most authors have recognized these as representing a single, uniquely chiropteran structure: the calcar (e.g., Miller, 1907; Andersen, 1912; Vaughan, 1970c; Van Valen, 1979; Hill and Smith, 1984; Simmons, 1994, 1995, 1998; Schutt, 1998). Surprisingly, we have found but one reference suggesting that the calcar might not be homologous in all bats. Allen (1893, p. 27), in his monograph of the bats of North America, stated, "The calcar is an element of doubtful homology." Nevertheless, Allen's (1893) cautionary statement has been largely ignored and the presence of a calcar has often been considered a synapomorphy of Chiroptera (e.g., Baker et al., 1991; Simmons, 1994, 1995) or at least potentially homologous in Megachiroptera and Microchiroptera (Simmons and Geisler, 1998). Most recently, Simmons and Geisler (1998) used new evidence on the phylogenetic relationships of fossil and extant bats to hypothesize that the calcar may have evolved independently in Microchiroptera and Megachiroptera. However, they did not review the morphology of the calcar in these two groups. In this study, we examined hindlimb morphology in representative specimens from both chiropteran suborders. We also examined specimens (when available) and literature accounts of five fossil bat genera, including Archeopteropus (Early Oligocene; Italy), which is widely regarded as an early megachiropteran (e.g., Meschinelli, 1903; Andersen, 1912; Revilliod, 1922; Dal Piaz, 1937; Jepsen, 1970; Habersetzer and Storch, 1987; Norberg, 1989; Bergmans, 1997; McKenna and Bell, 1997; Simmons and Geisler, 1998). The goals of our study were (1) to determine if morphological evidence supports the assumption that the calcar is a homologous structure is all bats, (2) to interpret the results of our morphological comparisons in the context of previous hypotheses of bat relationships, and (3) to reevaluate the placement of Archeopteropus in Megachiroptera in the light of previous evidence and new data on chiropteran hindlimb morphology. HISTORICAL OVERVIEW: THE "CALCAR" AND ASSOCIATED MUSCULATURE Humphry (1869) studied the myology of Pteropus edwardsii [= P. giganteus giganteus (Andersen, 1912)] and presented a more accurate description of the megachiropteran "styliform bone" and its associated musculature than found in any subsequent studies (e.g., those of MacAlister, 1872; Mori, 1961; Grasse, 1971). Humphry (1869, p. 313) described two delicate heads of m. gastrocnemius (one from the lateral femoral condyle, the other from the medial femoral condyle) as follows: 4 Schutt and Simmons They unite and terminate in a thin tendon which runs down to be inserted into the os calcis (= calcaneum). A little above its (m. gastrocnemius) insertion, the base of the styliform bone is closely connected with it. The bone takes its course inward towards the opposite limb in the connecting cutaneous fold, having no direct bony connection. [emphasis added] Initiating what would become a confusing trend in future studies, Humphry (1869, p. 300) referred to this spur-like projection into the uropatagium by two names—"styliform bone" and "supplementary calcaneal bone." Humphry (1869, pp. 300, 313, 316) described three additional muscles associated with the "styliform bone": (1) m. ischiocutaneous, which "runs from the ischium, across the ischiopedal fold of skin (= uropatagium) and the supplementary calcaneal bone upon the dorsum of the foot"; (2) m. styliform, which is "a slip of muscular fibers, (that) passes from the proximal end of the fifth metacarpal (sic) and radiates out to be inserted along the lower margin of this (styliform) bone"; and (3) m. abductor minimi digiti, "a short, thick muscle extending from the os calcis, near the insertion of the tendo Achillis, to the fibular side of the fifth metatarsal." MacAlister (1872) examined specimens of both chriropteran suborders and extensively described the myology of the hindlimb. However, he did not discuss the morphology of the "styliform bone" or "spur" in detail, nor did he illustrate the structure in megachiropterans or attempt to distinguish it from the microchiropteran structure (e.g., he made no mention of the relationship between the megachiropteran styliform bone and m. gastrocnemius). MacAlister (1872, p. 128) described Humphry's ischiocutaneous muscle in the megachiropteran Eleutherura marginata [ = Cynopterus sphinx (Andersen, 1912)] as "thin and band-like. . . passing in the uropatagium from the ischium to the integument over the calcaneum and dorsum of the foot." MacAlister (1872) suggested that this muscle might be "the biceps flexor cruris diminished to a rudiment." M. ischiocutaneous was not found in any of the other species examined (6 megachiropterans and 12 microchiropterans) by MacAlister. He did not describe additional cutaneous muscles associated with the microchiropteran hindlimb (i.e., mm. tibiocutaneous interna and calcaneocutaneous). MacAlister (1872) renamed Humphry's "styliform muscle," calling it "m. depressor ossis styliformis," the name used by most subsequent authors. Additionally, he described a different origin for this muscle in Microchiroptera and Megachiroptera. In the microchiropteran Noctulina altivolans (=Nyctalus noctula), m. depressor ossis styliformis was described as originating from the plantar surface of the calcaneum, while in Pteropus and its allies it was described as originating from the fifth metatarsal bone. MacAlister (1872, p. 159) also described the "levator ossis styliformis, a slender muscle from the back of the lower part of the ankle to the upper surface of the styliform bone." Unfortunately, although the muscle was listed in his "Explanation of the Plates" (MacAlister, 1872, p. 168, Fig. 14. m.), it was not illustrated in any plate. The identity of m. levator ossis styliformis has remained a mystery and there has been no reference to this muscle in subsequent studies (e.g., Vaughan, 1959, 1970d; Mori, 1961; Grasse, 1971). Although MacAlister (1872) vividly described the unique structure of the bat hindlimb and its components (e.g., a knee joint rotated by up to 180° from the typical mammalian condition), few researchers followed up his examination of chiropteran hindlimb morphology and apparently none questioned the homology of the styliform Morphology and Homology of the Chiropteran Calcar 5 bone. Studying adaptations for hanging, de Fenis (1919) noted differences in the presence and absence of muscles involved in movement of the microchiropteran "stylet" but did not elaborate on the condition in megachiropterans. Vaughan (1959) examined relationships between interspecific morphological differences (e.g., the shape and extent of the uropatagium) and differing functional requirements (e.g., degree of flight maneuverability) in his classic work on the functional morphology of the hindlimb skeleton and muscles in Eumops, Myotis, and Macrotus. In subsequent studies, he described the chiropteran skeletal (Vaughan, 1970c) and muscular systems (Vaughan, 1970d) and added another microchiropteran taxon (Hipposideros) to his analysis. Vaughan (1959, 1970d) described m. depressor ossis styliformis as originating from the dorsolateral surface of the calcaneus and the dorsal surface of the base of the fifth metatarsal bone and inserting along the anterior surface of the calcar in the microchiropterans examined. The inferred function of this muscle was to spread the uropatagium by swinging the calcar laterad, away from the shank (Vaughan, 1959, 1970d). Neither m. uropatagialis nor m. levator ossis styliformis was discussed by Vaughan (1959, 1970c, d), and he did not mention cutaneous muscles associated with the hindlimb (mm. tibeocutaneous and calcaneocutaneous). Homology of the calcar in all bats was implied throughout Vaughan's (1959, 1970c, d) studies—e.g., "the greatest specialization of the (bat) foot is the development of the calcar, a slender bone that articulates proximally with the posterolateral surface of the calcaneus and projects into the adjacent border of the uropatagium" (Vaughan, 1970c, pp. 124–125). In his work on the musculature of Pteropus edulis, Mori (1961) illustrated a "Fusswurzelstachel" (tarsal spine), the equivalent of Humphry's (1869) styliform bone. This structure was described as projecting from m. gastrocnemius at its point of insertion on the os calcis. Unfortunately, the actual position of the Fusswurzelstachel on the hindlimb and its relationship to associated muscles was not illustrated with consistency in Mori's (1961) figures. In one of his illustrations (Fig. 25), the Fusswurzelstachel appears to originate from the tarsal region, while in another (Fig. 26) it protrudes from an area proximal to the tarsal region, from a large unlabeled muscle or tendon (which is not m. gastrocnemius). Mori (1961) indicated that m. uropatagialis ( = Humphry's "ischiocutaneous muscle") originated on the coccygeal bone before splitting into two heads: one spanning the uropatagium before inserting at the proximal base of the Fusswurzelstachel and the other running along the proximal half of the free edge of the uropatagium. Additionally, a short, narrow m. adductor digiti minimi was described as originating on the os calcis, near the base of the Fusswurzelstachel, and insereting at the base of the fifth metatarsal bone. An unidentified muscle (Mori, 1961, Figs. 25–27) was shown originating from either the area adjacent to the insertion of m. gastrocnemius or near the base of metatarsal five before inserting onto the distal terminus of the Fusswurzelstachel. In all likelihood, this unlabeled muscle was MacAlister's (1872) m. depressor ossis styliformis, but Mori's figures do not lend themselves to positive identification of these delicate muscles. Grasse (1971) provided a summary of hindlimb osteomyology of bats, but unfortunately, his figures of the limb of Pteropus were adapted from those of Mori (1961). Since Mori's (1961) figures were incomplete (with regard to myology) and inconsistent (regarding the position of the styliform bone), this served to confuse rather than elucidate relationships of hindlimb structures. For example, a muscle in Pteropus [unidentified in 6 Schutt and Simmons Mori's (1961) figures] that was shown inserting onto the distal portion of "l'eperon" was positively identified by Grasse (1971, p. 388) as m. depressor ossis styliformis. This previously unidentified muscle was regarded by Grasse (1971, p. 388, Fig. 351, p. 389, Fig. 352) as the same muscle seen in two companion figures that showed similar regions of microchiropteran hindlimbs. That this muscle had different insertions in microchiropterans and megachiropterans was not mentioned. Rather, Grasse (1971) considered it to be unique to bats and homologous in microchiropterans and megachiropterans. Grasse (1971) described two additional muscles associated with the uropatagium (mm. calcaneocutaneous and tibiocutaneous intema) whose presence and absence varied among microchiropteran species. These muscles had been initially described in Myotis by Morra (1899). M. calcaneocutaneous was illustrated as a series of narrow bundles that extended medially from the "cartilaginosa del calcagno" into the uropatagium where they ended (Morra, 1899, p. 4). M. tibiocutaneous interna originated as a series of thin muscle bands from the dorsolateral surface of the tibia that extended into the uropatagium and terminated near the tail. Grasse (1971, p. 388) regarded "le developpement d'un eperon osseux ou cartilagineaux articule avec le bord externe du calcaneum" as a homologous characteristic of the chiropteran hindlimb. In summary, (1) spur-like projections from the ankle and their associated muscles have been described for both Microchiroptera and Megachiroptera, but descriptions have been inconsistent and different terminology has been applied by different authors; (2) there appears to be real variation in these structures between Microchiroptera and Megachiroptera, but this information cannot be extracted from the available literature; and (3) despite possible differences, the general trend has been to regard the "calcar" as a homologous structure in all bats. We propose to test this hypothesis. MATERIALS AND METHODS We examined hindlimb morphology in adult specimens and/or photographs and literature accounts from species representing all family-level lineages of extant bats, two ordinal-level extant outgroup taxa (Scandentia, Dermoptera), four Eocene fossil bat genera (Icaronycteris, Archaeonycteris, Palaeochiropteryx, and Hassianycteris), and one Oligocene fossil bat (Archaeopteropus); see the Appendix for specimens examined. Only adult individuals [as judged by complete epiphyseal fusion in the long bones (Kunz and Anthony, 1982)] were included in this sample. Alcohol-preserved specimens as well as skeletons were examined and/or dissected with a stereo-zoom microscope (Nikon SMZ10) equipped with a camera lucida. Gross examination of these specimens provided information on the morphology of the uropatagium. Microscopic examination and physical manipulation were used to determine whether anatomical structures (i.e., calcar, uropatagial spur) were calcified or cartilaginous (see Table I). Cartilaginous structures were identified by their appearance (clear or opaque when viewed against a background light source) and relatively high flexibility and elasticity (i.e., easily bent when physically manipulated). In contrast, calcified structures were identified by an alternative appearance (white, pale yellow, or gray, not clear or opaque) and lack of flexibility (i.e., structure brittle and not easily bent when manipulated). These visual and physical characteristics can be best contrasted by examining taxa in which the base of the calcar is calcified and the distal portion of the structure is cartilaginous (e.g., see Noctilio albi- Morphology and Homology of the Chiropteran Calcar 7 ventris). Microdissection was also used to gather data on the myological features of extant bats (see Table II). Archaeonycteris, Palaeochriopteryx, and Hassianycteris were examined by Johnathan Geisler (see Table III); the specimens he examined are listed by Simmons and Geisler (1998). Since the only known specimen of Archaeopteropus was destroyed during World War II, we examined a photograph of the fossil in Meschinelli's (1903) original publication. RESULTS Morphology of Extant Microchiroptera Although there is variation in the size, shape, and composition of the calcar among species, several morphological components were observed in all microchiropteran specimens except those that lack a calcar (Craseonycteris, Rhinopoma, and Diaemus). These features are summarized in Tables I and II. The microchiropteran calcar projects from the distal end of the calcaneal tuberosity, a bony extension of the calcaneum that projects anteromedially on the plantar surface of the ankle (Fig. 1). A small facet is present on the calcaneal tuberosity at the point of articulation between the calcaneum and the proximal base of the calcar (Fig. 2A). The calcar itself may appear entirely cartilaginous (e.g., Vampyrum, Phyllostomus) or calcified (e.g., Saccopteryx, Pteronotus, Molossus). In some taxa (e.g., Noctilio, Trachops), the base and proximal portion of the calcar are calcified, while the distal portion of the structure is cartilaginous. Calcars vary in length from tiny nubs of less than 1 mm (Desmodus and Mystacina) to greater than 30 mm (e.g., Noctilio leporinus). Utility of variation in calcar structure for addressing phylogenetic relationships within Microchiroptera will be investigated elsewhere (Schutt and Simmons, in preparation). Movement of the microchiropteran calcar is apparently controlled by three muscles: (1) m. depressor ossis styliformis ( = Humphry's "styloid muscle"), (2) m. calcaneocutaneous, and (3) indirectly, m. gastrocnemius. M. depressor ossis styliformis originates from the medial surface of the calcaneum and the dorsomedial surface of the base of the fifth metatarsal. The muscle inserts along the posterior border of the calcar. M. depressor ossis styliformis presumably functions to abduct the calcar away from the shank (i.e., toward the foot), thus helping to spread the uropatragium (Vaughan, 1959). In most microchiropteran taxa we examined, the delicate m. calcaneocutaneous is composed of a series of fine fibers that appear to originate along the anterior surface of the calcar, run anteromedially within the uropatagium, and insert along the midline of the membrane posterior to the distal end of the tail (Fig. 1). With its insertion along the shaft of the calcar opposite the insertion of m. depressor ossis styliformis, m. calcaneocutaneous presumably plays a role in stabilizing the calcar and as an antagonist of m. depressor ossis styliformis. Variation from this condition was noted in three microchiropteran clades: Rhinopomatidae, Megadermatidae + Rhinolophidae, and within Phyllostomidae (see Table II). In these forms we found that, in addition to the morphology described above, a portion of the calcaneocutaneous is band-like, forming a narrow but thickened strip of muscle that extends from the caudal ligament of the tail to the base of the calcar. M. gastrocnemius has two heads. The medial head of m. gastrocnemius originates 8 Schutt and Simmons Table I. Hindlimb Structures of Extant Batsa Taxon Pteropodidae Acerodon celebensis Cynopterus brachyotis Cynopterus sphinx Dobsonia praedatrix Eidolon dupreanum Eidolon helvum Eonycteris spelaea Epomophorus haldemani Epomops buettikoferi Harpyionycteris whiteheadi Lissonycteris angolensis Macroglossus minimus Megaloglossus woermanni Micropteropus pusillus Notopteris macdonaldi Nyctimene aello Nyctimene albiventor Nyctimene cephalotes Nyctimene draconilla Nyctimene major geminus Nyctimene robinsoni Paranyctimene raptor Pteropus hypomelanus Pteropus tonganus Rousettus aegyptiacus Rousettus amplexicaudatus Scotonycteris zenkeri Styloctenium wallacei Syconycteris australis Thoopterus nigrescens Emballonuridae Saccopteryx bilineata Taphozous melanopogon Rhinopomatidae Rhinopoma muscatellum Craseonycteridae Craseonycteris thonglongyai Nycteridae Nycteris thebaica Megadermatidae Lavia frons Macroderma gigas Megaderma spasma Rhinolophidae Rhinolophinae Rhinolophus arcuatus Hipposiderinae Hipposideros ruber Phyllostomidae Desmodus rotundus Diaemus youngi Diphylla ecaudata Vampyrum spectrum Phyllostomus hastatus Glossophaga soricina Sturnira lilium Carollia perspicillata Artibeus literatus Uropatagiumb V-shaped V-shaped V-shaped Y-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped Rudimentary V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped V-shaped Absent V-shaped Broad Broad V-shaped Broad Broad Broad Broad Broad Broad Broad U-shaped Rudimentary Rudimentary Broad Broad Broad Rudimentary Broad V-shaped Calcarc Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Calcified Calcified Absent Absent Calcified Cartilaginous Cartilaginous Cartilaginous Cartilaginous Cartilaginous Tab-like,e cartilaginous Absent Digitiform,e cartilaginous Cartilaginous Cartilaginous Cartilaginous Absent Cartilaginous Cartilaginous Uropatagial spurd Present Present Present Present Present Present Present Present Present Absent Present Present but minute Present Present Absent Present Present Present Present Present Present w/fork Present Present Present Present Present Present Present Absent Present Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Morphology and Homology of the Chiropteran Calcar 9 Table I. Continued Taxon Phyllostomidae Continued Trachops cirrhosus Noctilionidae Noctilio albiventris Mormoopidae Mormoops megalophylla Pteronotus parnellii Mystacinidae Mystacina robusta Myzopodidae Myzopoda aurita Thyropteridae Thyroptera tricolor Furipteridae Furipterus horrens Natalidae Natalus stramineus Antrozoidae Antrozous pallidus Molossidae Molossinae Eumops perotis Molossus rufus Molossus molossus Tomopeatinae Tomopeas ravus Vespertilionidae Vespertilioninae Eptesicus furinalis Miniopterinae Miniopterus schreibersi Myotinae Myotis nigricans Myotis riparius Murininae Murina cyclotis Kerivoulinae Kerivoula papillosa Uropatagiumb Broad Broad Broad Broad Rudimentary Broad Broad Broad Broad Broad Broad Broad Broad Broad Broad Broad Broad Broad Broad Broad Calcarc Cartilaginous w/ calcified base Cartilaginous w/ calcified base Cartilaginous w/ calcified base Calcified Cartilaginous Cartilaginous Cartilaginous Cartilaginous w/ calcified base Calcified Cartilaginous w/ calcified base Cartilaginous w/ calcified base Calcified Calcified Cartilaginous w/ calcified base Cartilaginous w/ calcified base Cartilaginous w/ calcified base Cartilaginous w/ calcified base Cartilaginous w/ calcified base Calcified Cartilaginous w/ calcified base Uropatagial spurd Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent Absent aSee Appendix for specimens examined. Classification follows Simmons (1998) and Simmons and Geisler (1998). bDescriptions of uropatagial membrane: V-shaped, posterior uropatagial border with V-shaped cutout; U-shaped, posterior uropatagial border with U-shaped cutout; rudimentary, fringe-like or greatly reduced; broad, uropatagium spans the region between the calcars, with no V-shaped or U-shaped cutout in posterior border. cThe calcar is defined as a cartilaginous or calcified element that articulates directly with the calcaneal tuberosity and projects into the uropatagial region. dUropatagial spur is defined as a cartilaginous element that projects from the tendon of m. gastrocnemius into the uropatagial region and has no direct articulation with the calcaneal tuberosity. eSee Schutt and Altenbach (1997, Fig. 1). 16 Schutt and Simmons Fig. 4. Photograph of the holotype of Archaeopteropus transiens reproduced from a plate by Meschenelli (1903). Note the well-developed calcar indicated by the white arrow. Morphology and Homology of the Chiropteran Calcar 17 DISCUSSION The results of our study suggest that the calcar may not be a homologous structure in microchiropteran and megachiropteran bats. In no megachiropteran did we find a structure projecting from the ankle that articulated directly with the calcaneum as in the microchiropteran condition. Conversely, we never found a microchiropteran that had a spur that emerges from the tendon of m. gastrocnemius as in the megachiropteran condition. In all microchiropterans examined, structures projecting from the ankle into the uropatagium articulated directly with the calcaneum. Mapping the distribution of the uropatagial spur on a phylogeny of Megachiroptera recently proposed by Springer et al. (1995) indicates that the presence of a uropatagial spur is primitive for Megachiroptera (Fig. 5). Reduction or loss of the uropatagial spur in some taxa (e.g., Syconycteris, Macroglossus, Notopterus) appears to be secondarily derived. Similarly, mapping the distribution of the calcar on a phylogeny of bats recently proposed by Simmons and Geisler (1998) indicates that the presence of a calcar is primitive for extant microchiropterans (Fig. 6). Loss of the calcar in a few taxa (e.g., Rhinolophidae, Craseonycteridae) appears to be secondarily derived. We found differences between Microchiroptera and Megachiroptera in the musculature associated with the calcar and uropatagial spur. Our results suggest that megachiropterans possess a m. uropatagialis, while microchiropterans do not. In megachiropterans this muscle is apparently responsible for adducting the uropatagial spur (or drawing it toward the shank). Originating near the base of the tail as two separate heads, m. uropatagialis spans a wide portion of the uropatagium. One portion of the muscle inserts at the base of the uropatagial spur, while the other inserts along its anterior shaft. ConFig. 5. Presence of the uropatagial spur mapped on the phylogeny of Megachiroptera proposed by Springer et al. (1995). Character optimization suggests that the presence of the uropatagial spur (shown in white) is primitive for Megachiroptera. 18 Schutt and Simmons Fig. 6. Presence/absence of the calcar and uropatagial spur mapped on the phylogenetic tree of bats proposed by Simmons and Geisler (1998). (A) Optimization assuming that Icaronycteris and Archaeonycteris lacked a uropatagial spur. Under this hypothesis, the spur and calcar evolved independently and are not homologous. (B) An alternative optimization assuming that Icaronycteris and Archaeonycteris has a uropatagial spur but that it was not preserved. Under this hypothesis, the uropatagial spur may have evolved into the calcar through modification of the distal end of the tendon of m. gastrocnemius and its connection to the calcaneum. In light of the current evidence, we consider hypothesis A to be more likely than B (see text for discussion). Note that both optimizations indicate that the presence of a calcar is primitive for extant Microchiroptera. Morphology and Homology of the Chiropteran Calcar 19 versely, two muscles associated with the calcar (m. calcaneocutaneous) and uropatagium (m. tibiocutaneous) were identified in microchiropterans but not in megachiropterans. We also found differences in m. depressor ossis styliformis in the two bat suborders. This muscle inserts along most of the posterior border of the calcar in microchiropterans, while in megachiropterans it runs along the posterior border of the uropatagium and inserts onto the distal tip of the uropatagial spur. Given the confusing nature of information in the literature (e.g., Mori, 1961; Grasse, 1971) and the observed variation in insertion, position, and shape between microchiropterans and megachiropterans, it is possible that the megachiropteran m. depressor ossis styliformis and the microchiropteran m. depressor ossis styliformis are not homologous muscles. Mapping the distribution of the calcar and uropatagial spur on the phylogeny presented by Simmons and Geisler (1998) suggests two possible evolutionary interpretations of our combined data for megachiropterans, microchiropterans, and the fossil taxa (Fig. 6). If the states in the fossil taxa Icaronycteris and Archaeonycteris have been correctly assessed (i.e., no calcar or uropatagial spur present), the interpretation shown in Fig. 6A is supported: the calcar and uropatagial spur are not homologous, having evolved independently in Megachiroptera and in the lineage leading to Microchiroptera. However, if Icaronycteris and Archaeonycteris had a uropatagial spur that was not fossilized, then mapping the distribution of these structures suggests a different interpretation (Fig. 6B). Under this scenario, the uropatagial spur may have evolved into the calcar. This may have resulted from a progressive shortening in length of the tendon of m. gastrocnemius between the uropatagial spur and calcaneum. We consider the latter hypothesis (Fig. 6B) to be somewhat less likely for several reasons. First, available fossilized skeletons of Icaronycteris and Archaeonycteris are relatively well preserved and include elements that were probably cartilaginous in life [e.g., costal and hyoid cartilages (see Simmons and Geisler, 1998)]. This argues against the idea that these taxa had cartilaginous uropatagial spurs that were not preserved in the fossils. Second, myological differences between megachiropterans and microchiropterans suggest that the uropatagial spur and calcar may have had different evolutionary histories. Accordingly, we prefer the interpretation shown in Fig. 6A: that the calcar and uropatagial spur are nonhomologous structures that evolved independently in the two bat suborders. We note, however, that this represents a working hypothesis that may require modification if new fossil finds or developmental data refute any of our observations. Archaeopteropus: A Megachiropteran? Our interest in Archaeopteropus stems from the fact that it has a well-developed calcar (Fig. 4). Archaeopteropus transiens (Lower Stampian of Northern Italy) was described by Meschinelli (1903), from a single, incompletely preserved specimen. Unfortunately, this specimen was destroyed during World War II, but casts and photographs have been used by researchers for the past 50 years. Although he was not explicit, Meschinelli's (1903) choice of Archaeopteropus for a generic name clearly indicates that he regarded this bat as an early megachiropteran. In his review of Megachiroptera, Andersen (1912, p. xxxix) concluded that Archaeopteropus had a "genuine Megachiropteran hand," although in some features "it rather approached the hand of some Microchiroptera." Although the skull was poorly preserved, Revilliod (1922) and Dal 20 Schutt and Simmons Paiz (1937) agreed that Archaeopteropus was a megachiropteran based primarily upon similarities in wing morphology. Revilliod (1922) did note, however, that the molars of Archaeopteropus resembled those of a microchiropteran bat. Winge (1941), Simpson (1945), and Romer (1966) placed Archaeopteropus in Pteropodidae. Simpson (1945) erected a subfamily within Pteropodidae—Archaeopteropodinae—to accommodate Archaeopteropus. This arrangement was followed by Koopman and Jones (1970). Russell and Sige (1970) placed several early Tertiary bats into a new microchiropteran superfamily Palaeochiropterygoidea. Two families (Palaeochiropterygidae and Icaronycteridae) were recognized, with Archaeopteropus tentatively placed in Icaronycteridae (Russell and Sige, 1970). Smith (1977) disagreed, arguing that no evidence supported this placement of Archaeopteropus. Rather, he recommended that Archaeopteropus be removed from the microchiropteran superfamily Palaeochiropterygoidea and transfered to Pteropodidae. Hill and Smith (1984) placed Archaeopteropus in Archaeopteropodinae within Pteropodidae. In the only detailed review of Archaeopteropus presented since Andersen (1912), Habersetzer and Storch (1987) asserted that Archaeopteropus was probably a megachiropteran based on the following characters: (1) considerable body size (extended wingspan exactly 90 cm); (2) very broad plagiopatagium; (3) high wing tip index (=1.68); (4) long and strong thumb with thick claw phalanx; (5) strong second finger with three phalanges, terminal phalanx with thick claw; (6) greater tubercle (=trochiter) of humerus poorly delineated from head and relatively low; (7) deltopectoral crest of humerus not flange-like; (8) epitrochlea of humerus very broad; and (9) relatively long hindlimbs (femur:humerus ratio = 0.78). Based on their analysis, Habersetzer and Storch (1987) placed Archaeopteropus in Archaeopteropodinae within Pteropodidae. Carroll (1988), Bergnans (1997), and McKenna and Bell (1997) followed Habersetzer and Storch (1987) and retained Archaeopteropus in Megachiroptera: Pteropodidae. Simmons and Geisler (1998) transferred Archaeopteropus to Megachiropteramorpha incertae sedis to reflect its uncertain affinities to extant pteropodids. Despite the growing consensus that Archaeopteropus should be classified as a megachiropteran, many authors have noted that the available data are not entirely consistent with this hypothesis of relationships. Studying the sole specimen of Archaeopteropus, Andersen (1912) determined that Meschinelli (1903) had mistakenly confused digit III and digit V. Correcting the identification of wing elements, Andersen (1912) noted a number of differences between Archaeopteropus and recent megachiropterans. He found that Archaeopteropus had metacarpal proportions not found in any recent megachiropteran (metacarpal III being conspicuously shorter than IV, and IV somewhat shorter than V). However, these proportions have close parallels in Microchiroptera. In Archaeopteropus, the fifth finger (including phalanges) is slightly longer than the fourth (as in some microchiropterans), while, as a rule, the fifth finger of recent megachiropterans is distinctly shorter than the fourth (Andersen, 1912). There are also three ossified Morphology and Homology of the Chiropteran Calcar 21 phalanges in digit III of the wing in Archaeopteropus, not two as in recent megachiropterans (Andersen, 1912; Simmons and Geisler, 1998). Finally, the tail of Archaeopteropus is long, a condition seen in most microchiropterans but in only one megachiropteran, Notopteris. Because Notopteris nests within Megachiroptera in recent phylogenies (e.g., Springer et al., 1995), it seems clear that a long tail is secondarily derived in this taxon. A short tail seems to be primitive for extant Pteropodidae (Springer et al., 1995). Andersen's (1912) somewhat weak support for a megachiropteran affiliation for Archaeopteropus was based upon one character of digit II (the index finger): the presence of a third phalanx (=ungual or terminal phalanx) equipped with a claw. Andersen (1912) considered this feature to be a diagnostic character for Megachiroptera. However, presence of a third phalanx equipped with a terminal claw on digit II is now recognized to be primitive for Chiroptera, because most mammals exhibit this condition and it is also present in Icaronycteris and Archeonycteris, two Eocene bats that are more closely related to extant Microchiroptera than Megachiroptera (Habersetzer and Storch, 1987; Simmons and Geisler, 1998). The absence of an index claw in two additional fossil taxa (Hasslanycteris and Palaeochiropteryx) and all extant microchiropterans is apparently derived (Simmons and Geisler, 1998) (Fig. 7). The same is apparently true for absence of the index claw in a few megachiropterans [Dobsonia, Eonycteris, Melonycteris, and Notopteris (Springer et al., 1995)] (Fig. 8). Although Smith (1976) placed Archaeopteropus in Pteropodidae, he did appear to question the megachiropteran affinities of this taxon. As did Revilliod (1922) and Russell and Sige (1970), Smith (1976, pp. 53–54) noted that the badly fragmented dentition of Archaeopteropus "more closely resembles that of the Microchiroptera in appearance." He suggested that Archaeopteropus may have been derived from "Paleochiroptera," a Fig. 7. Presence of an index claw (ossified third phalanx on wing digit II) mapped on the phylogenetic tree of bats proposed by Simmons and Geisler (1998). Note that an ossified index claw was apparently lost independently in the lineage leading to Microchiroptera and in some megachiropterans (see Fig. 8). 22 Schutt and Simmons Fig. 8. Presence of an index claw (ossified third phalanx on wing digit II) mapped on the phylogenetic tree of Megachiroptera proposed by Springer et al. (1995). Note that presence of an ossified index claw is apparently primitive for Megachiroptera. primitive and generalized Early to Middle Eocene grade thought to be ancestral to all extant Chiroptera. Norberg (1989) slightly modified the wing reconstructions used in Habersetzer and Storch's (1987) study of fossil bat species (including Archaeopteropus) (Figs. 9A and B). On the basis of the revised wing reconstructions, she recalculated wing measurements, body mass, and wing loading according to formulae derived by Norberg and Rayner (1987) based on large samples of extant bats. Contrary to Habersetzer and Storch (1987), who used a high wing tip index as evidence that Archaeopteropus was a megachiropteran, Norberg's data indicated that wing shape indices in Archaeopteropus (e.g., aspect ratio, size and shape of the wing tip) were very similar to those in many microchiropteran bats and, also, in the fossil relatives of Microchiroptera (Icaronycteris, Archaeonycteris, and Paleochiropteryx). Norberg (1989) also determined that Habersetzer and Storch (1987) may have overestimated body mass and wing span for Archaeopteropus. Her revised wing span estimate of 82 cm is indeed very large for a typical microchiropteran, but it still falls within the range of extant microchiropterans including Vampyrum spectrum [70–100 cm (Navarro and Wilson, 1982)]. The preservation of the distal portion of the hindlimb in Archaeopteropus is excellent and Habersetzer and Storch (1987) were correct in ascribing a long, bony calcar to Archaeopteropus (Fig. 4). The base of the calcar in Archaeopteropus clearly articulates with the calcaneum, which is a feature not present in any extant megachiropteran. We propose that the morphology and position of the calcar in Archaeopteropus are typical of a microchiropteran bat and therefore offer strong evidence that this bat was not a megachiropteran. Fig. 9. (A) Habersetzer and Storch's (1987) reconstruction of the wing and associated membranes in Archaeopteropus. This figxure depicts Archaeopteropus with a typical megachiropteran uropatagium. (B) Norberg's (1989) modification of the Habersetzer and Storch's (1987) reconstruction of Archaeopteropus. The shape of the wing tip has been modified to reflect the form in most extant bats, but the reconstruction of the uropatagium follows Habersetzer and Storch (1987). (C) Reconstruction of Archaeopteropus wing and associated membranes based on results of the present study. In developing this hypothesis we have retained Norberg's (1989) modified wing tips but have inferred that the presence of a calcar and tail suggests a more extensive uropatagium than included in previous reconstructions. 24 Schutt and Simmons Some of the other features cited by Habersetzer and Storch (1987) as indicating the megachiropteran affinities of Archaeopteropus are also problematic. A low, poorly defined trochiter of the humerus is probably a plesiomorphic condition for bats, as is a low deltopectoral crest that is not flange-like and the presence of a long, strong pollex (Simmons and Geisler, 1998; personal observation). While a broad epitrochlea on the humerus is indeed a derived characteristic of most megachiropterans, it also appears in many microchiropterans [e.g., noctilionids, phyllostomids, natalids, thyropterids (Smith, 1972)]. Similarly, relatively long hindlimbs occur in some microchiropterans (e.g., natalids, some vespertilionids). Even when taken together, we do not consider the characters cited by Habersetzer and Storch (1987) to provide conclusive evidence for the affinities of Archaeopteropus. Because of the incomplete nature of the sole fossil specimen of Archaeopteropus (see Fig. 4), we did not attempt a formal phylogenetic analysis including this genus. In our view, two characters of Archaeopteropus provide evidence useful for placing it in a phylogeny with other fossil and extant bats—the presence of a calcar and presence of a clawed third phalanx on the index finger. Based on these characters, Archaeopteropus appears to nest among the fossil relatives of Microchiroptera as the sister group of the clade comprising Hassianycteris, Palaeochiropteryx, and Microchiroptera (Fig. 10). Although this hypothesis is clearly based on very limited evidence, we consider it to represent a better working hypothesis than placement of Archaeopteropus in Megachiroptera. No unambiguous synapomorphies link Archaeopteropus with Megachiroptera; one such synapomorphy (presence of a calcar) clearly links Archaeopteropus with derived members of the microchiropteran lineage. Norberg (1989, p. 205) determined that Archaeopteropus "had a very low relative wing loading, and was similar in wing-shape characters to Rousettus." It is important to note, however, that different regression equations were used to estimate the body weight of Archaeopteropus because Norberg (1987) assumed that it was a megachiropteran. In addition, wing shape measurements were based on Habersetzer and Storch's Fig. 10. Proposed placement of Archaeopteropus in the phylogenetic tree presented by Simmons and Geisler (1998). This tentative hypothesis is based on the two characters shown on the tree. See text for discussion. Morphology and Homology of the Chiropteran Calcar 25 (1987) diagrammatic reconstruction of a narrow uropatagium in Archaeopteropus (Figs. 9A and B). This feature, which is not preserved in the fossil record, gives Archaeopteropus a uropatagium morphologically similar to that of typical extant megachiropterans (e.g., Pteropus). With its long, well-developed calcar and tail, we hypothesize that Archaeopteropus may have had a much broader uropatagium (Fig. 9C) then depicted by Habersetzer and Storch (1987) or Norberg (1989). With a greater wing area due to an increase in the calculated area of the uropatagium, we suspect that wing loading and relative wing loading in Archaeopteropus may have been considerably lower than published estimates. These observations are significant in light of correlations between dietary habits and body mass, aspect ratio, and relative wing loading in extant microchiropterans. Norberg and Fenton (1988) demonstrate that a combination of high body mass (>17 g), low aspect ratio (<6.3), and low relative wing loading (<36) significantly identifies carnivorous microchiropteran species from other animal-eating forms. In other words, this combination of traits distinguishes bats that include terrestrial vertebrates in their diet from those that are strictly insectivorous. Norberg (1989) estimated that Archaeopteropus had a body mass of well over 100 g, an aspect ratio of 6.1, and a relative wing loading of 43. Our new hypothesis of the phylogenetic affinities of Archaeopteropus, taken together with our revised reconstruction of the uropatagium, suggests a new dietary reconstruction for this taxon. The presence of a calcar and broad uropatagium, interpreted in light of Simmons and Geisler's (1998) hypotheses concerning the evolution of foraging strategies, suggests that Archaeopteropus was probably capable of aerial insectivory. However, if revised estimates of patagial area significantly lower estimates of wing loading, Archaeopteropus may be found to fall within the morphological range that Norberg and Fenton (1988) identified as characteristic of carnivorous bats. Though testing this hypothesis is beyond the scope of our study, it is intriguing to note that rather than being an early member of a frugivorous lineage, Archaeopteropus might represent the earliest known carnivorous bat. CONCLUSIONS This study illustrates the problems that can arise when assumptions about morphology and homology are perpetuated in the literature without adequate testing. Far from being an unambiguous synapomorphy of bats (as suggested by Baker et al., 1991; Simmons, 1994, 1995), the chiropteran "calcar" is not the same in all bats. New evidence from comparative morphology and a phylogenetic analysis of character transformations indicates that the calcar of microchiropteran bats and the uropatagial spur of megachiropterans may have evolved independently in these two groups and may not be homologous structures. Similarly, differences in the presence or absence, structure, and origin/insertion points of mm. uropatagialis, depressor ossis styliformis, calcaneocutaneous, and tibiocutaneous interna suggest that these muscles may have had different evolutionary histories in the two bat suborders and may not be homologous (although further work will be required to settle the latter question). Analysis of characters preserved in the Oligocene fossil bat Archaeopteropus, particularly the presence of a welldeveloped calcar, suggests that this taxon is an early member of the microchiropteran lineage rather than a megachiropteran bat. Recognition of this relationship not only 32 Schutt and Simmons Smith, J. D. (1976). Chiropteran evolution: In: Biology of the New World Family Phyllostomldae, R. J. Baker, J. K. Jones, and J. D. Smith, eds., pp. 49–69, Special Publications, No. 10, The Museum, Texas Tech University, Lubbock. Smith, J. D. (1977). Comments on flight and the evolution of bats. In: Major Patterns in Vertebrate Evolution, M. K. Hecht, P. C. Goody, and B. M. Hecht, eds., pp. 427–438, Plenum Press, New York. Smith, J. D., and Storch, G. (1981). New Middle Eocene bats from the "Grube Messel" near Darmstadt, W. Germany (Mammalia: Chiroptera). Senkenberglana Biol. 61: 153–168. Springer, M. S., Hollar, L. J., and Kirsch, J. A. W. (1995). Phylogeny, molecules versus morphology, and rates of character evolution among fruit bats (Chiroptera: Megachiroptera). Aust. J. Zool. 43: 557–581. VanValen, L. (1979). The evolution of bats. Evol. Theory 4: 104–121. Vaughan, T. A. (1959). Functional morphology of three bats: Eumops, Myotis, Macrotus. Publ. Mus. Nat. Hist. Univ. Kans. 12: 1–153. Vaughan, T. A. (1970a). Flight patterns and aerodynamics. In: Biology of Bats, W. A. Wimsatt, ed., pp. 195–216, Academic Press, New York. Vaughan, T. A. (1970b). Adaptations for flight in bats. In: About Bats, R. Slaughter and D. Walton, eds., Vol. 1, pp. 127143, Southern Methodist University Press, Dallas. Vaughan, T. A. (1970c). The skeletal system. In: Biology of Bats, W. A. Wimsatt, ed., Vol. 1, pp. 97–138, Academic Press, New York. Vaughan, T. A. (1970d). The muscular system. In: Biology of Bats, W. A. Wimsatt, ed., Vol. 1, pp. 139– 194, Academic Press, New York. Webster, F. A., and Griffin, D. R. (1962). The role of the flight membranes in insect capture by bats. Anim. Behav. 10: 332–340. Winge, H. (1941). The Interrelationships of Mammalian Genera, Vol. I. Monotremata, Marsupialia, Insectlvora, Chiroptera, Edentata, C. A. Reitzels, Copenhagen. [English translation by E. Deichmann and G. M. Allen of H. Winge (1923). Pattedyr-Slaegter. 1. Monotremata, Marsupialia, Insectivora, Chiroptera, Edentata. Copenhagen. This volume was a collection of papers written between 1887 and 1918, with footnotes dating up to 1922.]