| THE ORIGIN OF FLIGHT IN BATS |
To go where no mammal has gone before
The origin and early evolution of bats is among the greatest controversies m modern organismal biology. Some workers believe that flying foxes, which are actually Old World fruit-eating bats, are more closely related to humans and other primates than to the small bats that Americans see flying overhead. In this article, we review the controversy surrounding the origin of bats, which make up the second-largest order of mammals (only rodents are more diverse), and highlight some of the formidable changes that an unknown ancestral four-footed mammal must have undergone to become a flying bat.
The fossil evidence for bat origins is extremely sparse. Like birds, bats have delicate bones and their skeletons usually do not stand up to the forces of fossilization. There are a few notable exceptions, including the hundreds of Eocene bat skeletons (Figure 1) known from a quarry near Darmstadt in Germany (Habersetzer and Storch 1987).
Unfortunately, the fossils available only complicate matters. They do not represent transitional morphologies between quadrupedal (four-footed) mammals and flying bats (Novacek 1985), and they represent animals nearly as specialized as their modern relatives. Little is known of the oldest bat fossils from the late Paleocene of Wyoming (Gingerich 1987), but nothing indicates that they were different from their Eocene descendants. There is no Archaeopteryx for bats. Rather, the fossil record indicates merely that bats are a group that diversified early in the evolution of placental mammals.
Today's bats share many specializations. They have elongate fingers with no nails on the third through fifth fingers (Figure 2). They also all have wing membranes between their fingers, in front of the arm, and between the fore and hind limbs. In addition, they often have wing membranes between hind limb and tail (or between the hind limb and buttocks if there is no tail). The hip joint of all bats is modified so that the knee points outward (lateral or even dorsal) instead of forward (cranial). All bats have enormous chest muscles (pectorals) to power the wing beat.
Many of these features are easily interpreted as being related to flight, but for some features flight is not an obvious explanation. For example, modern bats lack a muscle, the sartorius, that is found in most other mammals. The sartorius extends between the outside of the hip bone (ilium) and the inside of the knee (tibia) and flexes the hip in most mammals.
Traditionally, all known bats (order Chiroptera) were conveniently grouped into two suborders: Megachiroptera and Microchiroptera. The names of these groups are somewhat misleading. Megachiroptera, or megabats, are in general much larger than Microchiroptera, but there is a large area of overlap (megabats vary between 13 and 880 grams; Microchiroptera, or microbats, between 2 and 160 grams).
The wings of the two groups are different: megabats have a simple shoulder joint and a clawlike nail on thumb and index finger, whereas microbats have a complicated shoulder joint and a claw only on the thumb. Most megabats navigate using their large eyes, whereas microbats use reflections of their own ultrasonic cries to determine the shape of their surroundings. This system, called echolocation, is independently acquired in a few megabats.
Megabats usually eat fruit, flowers, or nectar, whereas microbats typically eat insects. There are many exceptions to this dietary distinction. Several members of the microbat family Phyllostomidae resemble megabats in diet. Interestingly, phyllostomids only occur in the New World tropics, where there are no megabats, suggesting that New World microbats radiated into niches filled by megabats in the Old World. Reflecting these diets, microbat dentitions often have sharp crests and pointed cusps to deal with struggling prey, whereas megabat molars show only blunt topography for crushing soft plant food. Physiological constraints restrict megabats to the tropics, whereas microbats inhabit temperate and tropical zones. The physiology of microbats that live in regions with severe winters allows them to hibernate or migrate.
All these differences have led several researchers to question whether microbats and megabats are really as closely related as has been traditionally assumed. Perhaps each of the two suborders is more closely related to a separate group of quadrupedal mammals than they are to each other. If that is true, then flight would have evolved twice in mammals, and this evolution would have led to surprisingly similar solutions to problems of wing formation. For instance, both megabats and microbats have elongate fingers with wing membranes stretched between them. The other groups of flying vertebrates, birds and pterosaurs, have evolved different ways to support their wings. Thus, if megabats and microbats evolved flight independently, then their evolution would be an important lesson of how similar selection pressures produce similar morphologies apparently because of historical constraints inherent to their common mammalian origin.
More than a decade ago, J. D. Smith and his colleagues proposed that megabats and microbats evolved from different quadrupedal mammals (e.g., Smith and Madkour 1980). More recently, J. D. Pettigrew revived Smith's hypothesis and suggested that megabats are most closely related to primates. Pettigrew and co-workers (Pettigrew 1986, 1991, Pettigrew et al. 1989) were impressed by the detailed similarity of the brains of megabats and primates, whereas microbat brains display different morphologies.
Critics of Pettigrew's inferences have pointed out that many of the neural similarities of megabats and primates are in the visual system (Baker et al. 1991, Simmons et al. 1991, Wible and Novacek 1988). Megabats and primates both rely on their eyes as their primary sense organ, whereas most microbats use their ears in echolocation. Thus, megabats and primates could have independently evolved similar neural hardware as adaptations for enhanced vision. On the other hand, one could make the same argument for flight. Numerous similarities between microbats and megabats are related to active flight, and strong selection pressures for aerial locomotion could have produced these morphologies twice.
There are three reasonable evolutionary scenarios. There could be either rampant convergences, in a first scenario, between the visual systems of megabats and primates, or, in a second scenario, between the locomotor systems of megabats and microbats (Pettigrew 1991). Third, improved vision could have been present in the common ancestor of megabats, microbats, and primates, only to be lost secondarily in the ancestors of microbats (Simmons et al. 1991).
One approach to select among the scenarios would be to ignore characters that are related to visual and locomotor systems. But it is difficult to identify all of these characters. The remaining traits do not show a clear pattern: some of these characters link megabats to primates, whereas others unite bats.
If the problem of bat origins is ever solved, it will be after a careful analysis of all characters of interest in the bats and their potential relatives. The structure of characters needs to be understood to show that superficial similarities are really the result of shared ancestry (homology). The function of each individual trait also needs to be evaluated. Selection pressure for a new system of locomotion (such as flight) or a new sensory system (improved vision) could cause a host of seemingly unrelated changes that in fact go hand in hand with each other.
Our contribution to the question of the origin of bats is an analysis of one similarity between megabats and microbats that is potentially related to flight (Thewissen and Babcock 1991). This similarity concerns the flap of the wing anterior to the arm of bats. This flap is called the propatagium and consists of a layer of skin that is folded, with the two leaves of the fold touching and the fold being the leading edge of the wing (Figure 3). The skin enfolds two important structures: the largest vein of the forelimb (the cephalic vein) and part of a complicated system of muscles that we call the propatagial muscle complex.
Both microbats and megabats have a propatagial muscle complex, but it is surprisingly different in the two groups. In megabats (Figure 4), the complex has four separate attachments near the midline of the body: the back of the skull, the side of the face, the ventral side of the neck, and the midline of the chest (Strickler 1978). The muscle stretches between these attachments, called origins, and attachments further out on the forelimb, called insertions. In microbats, the propatagial muscle complex typically has just two origins (Strickler 1978): one from the back of the head and one from the chest (the latter is absent in the microbat shown in Figure 4).
From their origin, the muscle bellies extend toward the shoulder in all bats. Near the shoulder, the different muscle bellies join to form one collagenous tendon that extends toward the arm. The muscle complex extends in the leading edge of the wing distal to the shoulder. The muscle complex may include muscle fibers, collagenous tendons, and elastic tendons, and the pattern into which these are arranged varies so strongly between different members of each of the suborders that there are no consistent differences between the suborders as a whole. In its most simple form, the complex lacks any muscle bellies (e.g., the microbat Rhinopoma), but in other bats there may be one or more small muscle bellies that are separated by strands of collagenous tendons or by strands of elastic tissue.
In gross morphology then, the propatagial muscle complex of megabats differs from that of microbats. Its position in the leading edge of the wing suggests that it is important in flight. This observation led Pettigrew et al. (1989) to suggest that the complex probably evolved independently in megabats and microbats as a result of selection for active flight.
On the other hand, the propatagial complex is present in all megabats and microbats and has never been found in any terrestrial (nongliding) mammal. Supporters of the single-origin theory for the evolution of bats interpreted this observation as evidence in favor of their theory (Baker et al. 1991, Wible and Novacek 1988). To further investigate these contradictory views concerning the homology of the propatagial muscles of megabats and microbats, we address two issues: Are the propatagial muscles necessary for aerial locomotion, and does the gross structure of the complex preclude its homology in microbats and megabats?
No experimental data has been brought to bear on the first question. Theoretical considerations suggest that a bat wing can only function if its leading edge is stiff (Norberg 1969), but they do not dictate that a muscle is necessary for this. We can only make inferences concerning the function of the propatagial muscles on the basis of comparison with other vertebrates.
If the propatagial muscles are essential for aerial locomotion, then one would expect to find similar muscles in other aerial vertebrates. Active flight among mammals only occurs in bats, but gliding flight has evolved independently six times in mammals varying from mouse- to cat-size. The phylogenetic history of each of these taxa is distinct, and the taxa have little in common beyond the gliding membranes.
Placental gliders include the flying squirrels from the Northern Hemisphere (Petauristinae), African scaly tailed squirrels (Anomaluridae), and East Asian flying femurs (Cynocephalus, not a primate but the only genus of the order Dermoptera). Marsupial gliders include feathertail gliders (Acrobates, Burramyidae), sugar gliders (Petaurus, Petauridae), and greater gliding phalangers (Petauroides, Pseudocheiridae).
Of these taxa, a well-developed propatagium is lacking only in Petaurus and Petauroides. The morphology of the muscle complex present in the leading edge of the propatagium of all other gliders varies greatly. The propatagial complex of Cynocephalus, for instance, is not cordlike as in bats, but rather it consists of two sheets of muscle, one of which extends over the entire width of the propatagium.
In addition, a propatagial muscle complex is present in birds, and well-presented fossil impressions suggest that it also occurred in pterosaurs (Wellnhofer 1975). Given that a propatagial muscle is present in all vertebrate flyers and gliders with a propatagium, it follows that the complex is probably essential in aerial locomotion. As such, we cannot rule out that the muscle evolved separately under the influence of selection pressures related to flight in the ancestors of megabats and microbats.
Does the morphology of the propatagial complex preclude its homology in microbats and megabats? Differences in gross morphology have been interpreted by different investigators to both confirm and refute this statement, making closer evaluation necessary. One important realization is that several parts of the propatagial muscle complex are extremely variable. Within the megabat genus Pteropus, for instance, the segment of the propatagial complex in front of the arm may consist of one long muscle belly or a short belly with two long tendons. We have not been able to draw evolutionary conclusions from the variation of these parts.
Regions of the muscle complex that are more conservative within lower taxonomic units are probably more reliable guides to the homology of the propatagial muscles of microbats and megabats. Two characteristics of the propatagial muscle complex occur in all microbats and megabats: the origin from the back of the skull and the dose association of the cephalic vein with the muscle. Apart from the propatagial position of the cephalic vein in sugar gliders, these particular aspects of the propatagial muscle complex do not occur in mammalian gliders or in birds. Therefore, they are probably unrelated to flight. Because they are unique to microbats and megabats, these characters suggest that microbats and megabats originated from a single flying ancestor.
In addition to these two gross morphological characters, the innervation of the propatagial muscles suggests even more strongly that all bats share a single ancestor. Each muscle of an animal's body is connected by a nerve to the small section of the central nervous system that controls the contraction of that muscle. Occasionally, different parts of a single muscle may receive nerves from different parts of the central nervous system. The propatagial muscles of bats are an extreme example of such multiple innervation.
The propatagial muscles of bats receive one nerve from the brain and several nerves from adjacent parts of the spinal cord. The nerve that comes from the brain is a branch of the facial nerve. This combined innervation of a single muscle by branches of the facial nerve and nerves from the spinal cord is unique among mammals. It occurs only in the propatagial complex of microbats and megabats (Thewissen and Babcock 1991). This type of innervation is not necessary for flight because it is not known to occur in any mammalian glider or in birds. Because such multiple innervation does not occur in any other mammalian muscle, it suggests that the situation in bats evolved only once, in the common ancestor of microbats and megabats.
If the evidence from the propatagial muscles suggests that all bats originated from a common ancestor with wings, what did this ancestor look like? In 1910, the American Museum of Natural History paleontologist William King Gregory discussed the relations between the orders of mammals in a monograph that still influences ideas today. Gregory believed that bats, flying femurs, tree shrews, elephant shrews, and primates were closely related. He called this cluster Archonta (which means "ruling beings," because it includes primates).
Several modifications to mammalian systematics have altered Gregory's archontan concept slightly. Elephant shrews, for example, are now generally excluded. Another change in the archontan concept concerns a group of Paleocene and Eocene mammals called Plesiadapiformes. These animals range from shrew- to marmot-size and are somewhat reminiscent of rodents in morphology. Until recently (including in Gregory's monograph), Plesiadapiformes were thought to be closely related to primates, but now most workers treat them as a separate order, although still including it in Archonta. Character evidence supporting Archonta is scanty, but many systematists do think that Archonta is a homogeneous group consisting of at least Primates, Chiroptera (bats), Dermoptera (flying femurs), Scandentia (tree shrews), and Plesiadapiformes.
Among Archonta, flying femurs have been proposed as close relatives to bats. The presence of gliding membranes as well as several other traits led Leche (1886) to suggest such ties, and the idea is still current (Novacek and Wyss 1986). Evidence from the propatagial muscles is certainly consistent with this theory. The propatagial complex of the flying femur consists of two layers of muscle. The dorsal layer is sheetlike, originating from the side of the face and extending below the skin over the whole breadth of the propatagium and inserting on the forearm (Figure 4). The second muscle sheet of the propatagial complex consists of a few diffuse fibers that originate from the dorsal midline of the neck and extend toward the shoulder and the leading edge of the wing perpendicular to the dorsal sheet. The dorsal sheet is innervated by the facial nerve, whereas the ventral sheet is innervated by cervical spinal nerves (Thewissen and Babcock 1991). The nerves involved in innervating the propatagial muscles in flying femurs are thus the same as in bats. But there is a difference: facial and cervical spinal nerves go to different muscles of the propatagial complex in flying femurs, whereas in bats these nerves together innervate the single propatagial muscle complex.
One could, however, easily conceive of an evolutionary event that would rearrange the muscles of flying femurs to look like those of bats. In ontogeny, the two sheetlike muscles would need to fuse into a single cord-shaped muscle, which could retain the original innervation of the two muscles. This fusion would result in the dually innervated muscles of bats. This scenario is straightforward and, although still speculative, is consistent with what previous workers have said about the relations between bats and flying femurs.
Not every mammalogist believes in close ties between bats and flying lemurs. Flying femurs are in the middle of another controversy in mammalian phylogeny. Based on independent evidence, Beard (1990) and Kay et al. (1990) proposed that flying femurs might be related to the diverse group of fossil mammals called Plesiadapiformes. Beard (1990) took it one step further. He discussed two families of plesiadapiforms, Plesiadapidae and Paromomyidae, and appointed the latter as the closest relative of flying femurs.
The best-described plesiadapids have limb proportions that are common for quadrupeds. According to Beard (1990), paromomyids have different limb proportions: their intermediate phalanx (the middle bony segment of a finger) is longer than their proximal phalanx (the bony segment of a finger closest to the palm). The intermediate phalanx of flying femurs is also longer than their proximal phalanx, and Beard (1990) made this distinction his main argument for proposing that paromomyids had wing membranes and were closely related to flying femurs.
Beard's theory remains controversial and his inferences are doubted by Krause (1991). But assume for the moment that paromomyids and flying femurs are closely related. Does this preclude a special relationship between the flying femurs and bats?
Figure 5 is a plot of lengths of proximal and intermediate phalanges of bats and flying femurs. It shows that many bats (including Eocene forms), as well as flying femurs, have intermediate phalanges that are longer than their proximal phalanges. Bats obviously have wing membranes between the fingers, and the two characters are thus shared by bats, flying femurs, and possibly paromomyids. Therefore, a special relationship between bats and flying femurs is not at odds with potential dose relationships between paromomyids and flying femurs.
Another controversy over the homology of propatagial muscles raged more than a century ago. At that time, US Army Captain R. W. Shufeldt described a small section of a muscle that entered the propatagium of some, but not all, birds. Shufeldt (1887a-d) held this slip to be homologous among the different birds having it and thus of interest in determining phylogenetic relations. His colleague Stejneger (1887,1888,1889) disputed this claim in harsh terms. Their controversy centered around the same points as the current debate about the propatagial complex of bats: function, variability in structure, and their bearing on determining homology.
We believe that the evidence from the propatagial muscle complex of bats supports the idea that all bats share a single ancestor with wings. This idea is consistent with bats going through a flying femur-like stage before acquiring active flight. Of course, the propatagial evidence is not conclusive. Our evidence only forms a piece of the puzzle of bat ancestry.
We thank J. Habersetzer and G. Storch (Senckenberg Museum, Frankfurt) for the opportunity to study the fossil Messel bats, and B. D. Patterson (Field Museum, Chicago) and D. E. Wilson (Smithsonian Institution, Washington, DC) for access to recent specimens. We thank N. B. Cant and K. K. Smith for the use of their laboratories. We also thank M. C. Maas, P. W. Freeman, and E. E. Culotta for reading and commenting on the manuscript.
PHOTO (BLACK & WHITE): Figure 1. Fossil microbat (Archaeonycteris) from the Eocene (49 million years ago) locality of Messel in Germany. This locality has yielded hundreds of fossil bats; it was once an anoxic lake with poisonous volcanic fumes hanging over it. Bats that flew over the lake became unconscious, fell in the water, and drowned. Preservation of soft tissues shows that propatagial muscles were already present in these bats (arrow).
PHOTO (BLACK & WHITE): Figure 2. Skeleton and wing outline of representative microbat (left) and megabat (right). Notice the large orbit in the megabat, indicating a large eye and well-developed vision. The eye of a microbat is much smaller, and there are fewer bony crests surrounding its orbit. Nails are absent on four fingers in the microbat, but only on three in the megabat. (Not to scale.)
PHOTO (BLACK & WHITE): Figure 3. Cross-section through the arm of a microbat (Tadarida brasiliensis). The bone seen in section is the humerus (H). The flap of wing on top is the propatagium, which envelops two structures at its leading edge: the cephalic vein (C) and the propatagial muscle complex (M). Detailed similarities in the muscle complex lead us to believe that flight evolved only once in mammals.
PHOTO (BLACK & WHITE): Figure 4. Homology and homoplasy: the muscles of the leading edge of the wing in a microbat (Myotis lucifugus), a megabat (Pteropus sp.), a flying femur (Cynocephalus volans), and an American robin (Turdus migratorius). Muscle tissue is stippled; tendons and elastic tissue are white. The propatagial muscle complex of microbats and megabats shows special similarities and we believe is homologous. Less detailed similarities are present between flying femurs and bats; the complex probably occurred in a simplified form in the last common ancestor of these groups. The muscle complex is independently evolved in birds and bats. Modified after Thewissen and Babcock (1991), Corpyright 1991 American Association for the Advancement of Science.
GRAPH: Figure 5. Proportions of the fingers of microbats (open symbols), megabats (filled symbols), and flying femur Cynocephalus (Dermoptera; half-filled symbols). Digital proportions of specimens of 28 microbats, 8 megabats, and 1 flying femur are plotted. The microbats represent nine families, whereas the megabats exemplify the only extant family, and the flying femur typifies the only extant genus. Some members of both suborders of bats have intermediate phalanges longer than proximal phalanges as in flying femurs (points above the diagonal). Other bats have typical mammalian proportions (below the diagonal).
Baker, R. J., M. J. Novacek, and N. B. Simmons. 1991. On the monophyly of bats. Syst. Zool. 40: 216-231.
Beard, K. C. 1990. Gliding behaviour and palaeoecology of the alleged primate family Paromomyidae (Mammalia, Dermoptera). Nature 345: 340-341.
Gingerich, P. D. 1987. Early Eocene bats (Mammalia, Chiroptera) and other vertebrates in freshwater limestones of the Willwood Formation, Clarks Fork Basin, Wyoming. Contrib. Mus. Paleontol. Univ. Mich. 27: 275-320.
Gregory, W. K. 1910. The orders of mammals. Bull. Am. Mus. Nat. Hist. 27: 1-524.
Habersetzer, J., and G. Storch. 1987. Ecology and echolocation of Eocene Messel bats. Pages 213-233 in V. Hanak, I. Horacek, and J. Gaisler, eds. European Bat Research. Charles University Press, Praha.
Kay, R. F., R. W. Thorington Jr., and P. Houde. 1990. Eocene plesiadapiform shows affinities with dying femurs not primates. Nature 345: 342-344.
Krause, D. W. 1991. Were paromomyids gliders? Maybe, maybe not. J. Human Evol. 21: 177-188.
Leche, W. 1886. Uber die Sauegethiergattung Galeopithecus: eine morphologische Untersuchung. Kungl. Svenska Vetenskap. Akad. 21: 1-92.
Norberg, U. M. 1969. An arrangement giving a stiff leading edge to the hand wing in bats. J. Mammal. 50: 766-770.
Novacek, M. J. 1985. Evidence for echolocation in the oldest known bats. Nature 315: 140-141.
Novacek, M. J., and A. R. Wyss. 1986. Higher-level relationships of the Recent eutherian orders: morphological evidence. Cladistics 2: 257-287.
Petigrew, J. D. 1986. Flying primates? Megabats have the advanced pathway from eye to midbrain. Science 231: 1304-1306.
------. 1991. Wings or brain? Convergent evolution in the origins of bats. Syst. Zool. 40: 199-216.
Petigrew, J. D., B. G. M. Jamieson, S. K. Robson, L. S. Hall, K. I. McNally, and H. M. Cooper. 1989. Phylogeneic relations between microbats, megabats and primates (Mammalia: Chiroptera and Primates). Philos. Trans. R. Soc. Lond B Biol. Sci. 325: 489-559.
Schufeldt, R. W. 1887a. The dermo-tensor patagii muscle. Auk 4: 353-356.
------. 1887b. Another muscle in birds of taxonomic value. Science 9: 623-624.
------. 1887c. The dermo-tensor patagii muscle. Science 10: 57.
------. 1887d. A review of the muscles used in the classification of birds. J. Comp. Med. Surg. 8: 321-344.
Simmons, N. B., M. J. Novacek, and R. J. Baker. 1991. Approaches, methods, and the future of the chiropteran monophyly controversy: a reply to J. D. Pettigrew. Syst. Zool. 40: 239-243.
Smith, J. D., and G. Madkour. 1980. Penial morphology and the question of chiropteran phylogeny. Pages 347-365 in D. E. Wilson and A. L. Gardner, eds. Proceedings of the 5th International Bat Research Conference. Texas Tech Press, Lubbock.
Stejneger, L. 1887. Pars propatagialis musculi cucullaris. Science 10: 70-71.
------. 1888. Propatagialis cucullaris. Auk 5: 120-123.
------. 1889. Cucullaris propatagialis in oscinine birds. Science 13: 16.
Strickler, T. L. 1978. Functional osteology and myology of the shoulder in the Chiroptera. Contr. Vert. Evol. 4: 1-198.
Thewissen, J. G. M., and S. K. Babcock. 1991. Distinctive cranial and cervical innervation of wing muscles: new evidence for bat monophyly. Science 251: 934-936.
Wellnhofer, P. 1975. Die Rhamphorhynchoidea (Pterosauria) der Oberjura Plattenkalke Suddeutschlands. I. Allgemeine Skeletemorphologie. Palaontographica A148: 1-33.
Wible, J. R., and M. J. Novacek. 1988. Cranial evidence for the monophyletic origin of bats. Am. Mus. Novit. 2911: 1-19.
~~~~~~~~
By J. G. M. Thewissen and S. K. Babcock
J. G. M. Thewissen is a postdoctoral fellow and S. K Babcock is a graduate student in the Department of Biological Anthropology and Anatomy, Duke University Medical center, Durham NC 27710. Copyright 1992 American Institute of Biological Sciences.
