AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2877, pp. 1-1.8, figs. 1-10 April 28, 1987

Auditory Features and Affinities of the Eocene Bats Icaronycteris and Palaeochiropteryx (Microchiroptera, incertae sedis)

MICHAEL J. NOVACEKI

ABSTRACT The earliest known bats are skeletons of Icaro- ly, there is no reason to recognize a "primitive- nycteris index from the early Eocene of western ancestral" group, Eochiroptera, that is excluded Wyoming and a few less well-represented species from Microchiroptera or Megachiroptera. The re- from the early Eocene of France. Also known are lationships ofIcaronycteris and Palaeochiropteryx Palaeochiropteryx tupaiodon and several other within Microchiroptera remain uncertain. Asso- species from the middle Eocene of western Ger- ciation ofthese taxa and several other Eocene forms many. These taxa have been regarded as primitive within the microchiropteran superfamily Palaeo- forms, either "ancestral" to echolocating micro- chiropterygoidea fails to clarify these relation- chiropterans or "ancestral" to both micro- and ships. Palaeochiropterygoidea has not been de- megachiropterans. Details ofbasicranial structure fined by derived characters, and Icaronycteris and suggest that these Eocene forms were, however, Palaeochiropteryx are more accurately designated specialized echolocators comparable to Recent Microchiroptera incertae sedis. Several primitive microchiropterans. Moreover, quantitative anal- features shown by Icaronycteris suggest that the ysis reveals that the Eocene bats have a more pro- development of a sophisticated system for echo- nounced expansion of the cochlea than many Re- 'location within Microchiroptera occurred earlier cent microchiropteran species. There is clear than certain modifications ofthe postcranial skel- justification for reference ofIcaronycteris and Pa- eton. laeochiropteryx to the Microchiroptera. Converse- INTRODUCTION The oldest bats in the fossil record are early scribed by Jepsen (1966) from an exquisite, Eocene in age. Icaronycteris index was de- nearly complete skeleton found in the Green

I Chairman and Associate Curator, Department ofVertebrate Paleontology, American Museum ofNatural History.

Copyright © American Museum of Natural History 1987 ISSN 0003-0082 / Price $2.00 2 AMERICAN MUSEUM NOVITATES NO. 2877

River Formation of Fossil Lake, Wyoming. The above classification is, however, in Later found were two other skeletons of this conflict with certain other studies. Segall genus (Novacek, 1985a). Also known from (1971) argued that Icaronycteris index shared early Eocene faunas are Icaronycteris? menui, a more general basicranial structure with er- Archaeonycteris brailloni, andAgeina tobieni, inaceid insectivorans and megachiropterans, all described by Russell et al. (1973) from the and lacked many of the morphological cor- Mutigny and Avenay quarries in northeast relates with ultrasonic echolocation found in Epemay, Maine, France. These latter three microchiropterans. Furthermore, Jepsen species are known only from partial denti- (1966, 1970) had also emphasized the large tions. number of primitive traits in the skeleton of The middle Eocene bat Palaeochiropteryx Icaronycteris. Accordingly, Van Valen (1979) tupaiodon was first described by Revilliod recognized Icaronycteris and other palaeo- (1917) from the famous "Grube Messel" near chiropterygoideans as members of the sub- Darmstadt, Hessen, West Germany. A large order Eochiroptera, a group either more re- number of skeletons (more than 54 individ- motely divergent than or "ancestral" to the uals) of this bat have been recovered from Microchiroptera and Megachiroptera (Van the site (see Smith and Storch, 1981). Other Valen, 1979, fig. 1). bats from the Messel pit include Archeonyc- Study of described and undescribed ma- teris trigonodon Revilliod, 1917; "Archaeo- terial representing Icaronycteris contradicts nycteris" revilliodi Russell and Sige, 1970; Segall's (1971) description of the auditory Hassianycteris messelensis Smith and Storch, system of this bat as "non-microchiropter- 1981; and Hassianycteris magna Smith and an." Icaronycteris and Palaeochiropteryx both Storch, 1981. All of these forms are repre- clearly show several features that correlate sented by nearly complete skeletons. In ad- with the presence of ultrasonic echolocation dition, several species of middle Eocene bats in Recent microchiropterans. This point was (Cecilionycteris prisca Heller, 1935; Mat- the subject ofa briefreport (Novacek, 1985a). thesia germanica Sige and Russell, 1980; and The purpose of this paper is to provide a Matthesia? insolita Sige and Russell, 1980) more thorough description of these traits, are known from the "Grube Cecilie," Gei- better reproductions of the relevant illustra- seltal, near Halle a. S., East Germany. tions, and a discussion on the systematic im- The affinities ofthese fossil forms are prob- plications of the auditory morphology and lematic. Winge (1923, p. 304) regarded Pa- other cranioskeletal features in these fossil laeochiropteryx as essentially a primitive ves- bats. pertilionid of the suborder Microchiroptera. In a recent review, Smith and Storch (1981) ACKNOWLEDGMENTS endorsed some ofWinge's views, but did not refer Palaeochiropteryx to the Vespertilion- I thank Karl Koopman, Malcolm C. idae. They grouped the above-cited species McKenna, Andre Wyss (AMNH), Elizabeth (except for Hassianycteris which they loosely Pierson (University of California, Berkeley), allied with emballonurids or rhinolophids and Thomas A. Griffiths (Illinois Wesleyan Uni- recognized as Microchiroptera incertae sedis) versity), and Gary S. Morgan (Florida State within the Palaeochiropterygoidea Revilliod Museum) for criticisms of the manuscript. I 1917, suborder Microchiroptera. From this am especially appreciative to Dr. Griffiths for group they removed Archaeopteropus tran- his detailed comments and for sharing with siens Meschinelli, 1903 (locality: Monteviele, me his insights on the hyoid apparatus in northeast Italy, early Oligocene), and referred bats. For loan of specimens I am grateful to this species to the Megachiroptera. In this J. A. Lillegraven (University of Wyoming), action they claimed that Palaeochiroptery- D. Baird (Princeton University collections goidea was ". . . less paraphyletic" (Smith and now transferred to Yale University), and J. Storch, 1981, p. 163). Their grouping also Franzen (Senckenberg Museum, Frankfurt, endorses Jepsen's (1966) original description West Germany). of Icaronycteris, wherein this taxon was as- Figures 1, 3, 7, 8, and 9 are stereophotos, signed in the Microchiroptera. radiographs, and drawings from G. L. Jep- 1987 NOVACEK: EOCENE BAT EARS 3

Fig. 1. Stereophotograph of ventral skull of Icaronycteris index; type, PU (Princeton University) 18150. Note this specimen is now deposited in the Yale Peabody Museum Collection. Figure from Jepsen (1970). Symbols are: An, angular process ofdentary; Ax, axis; C, upper right canine; Cp, coronoid process of left dentary; Fp, postglenoid foramen; GI, glenoid fossa; i3, third lower left incisor; L, locator bristles; Ld, labial surface of left dentary; M3, third upper right molar; m3, third lower left molar; Oc, occipital condyles; P4, fourth upper right premolar; p3, third lower left premolar; SI, left stylohyal; Vc, vertebroarterial canal of atlas; Vd, ventral border of right dentary; Z, zygomatic arch.

sen's collection of figures of Icaronycteris in- University] 18150) did not show marked en- dex. These were loaned by D. Baird at Prince- largement of the cochlea, although he ac- ton and are now maintained, along with the knowledged that only a small part of this type specimen, by the Yale Peabody Mu- structure was preserved. He based his as- seum. Dr. H. Tuengerthal made the radio- sessment on the observation that in Icaro- graphs shown in figure 4; Chester Tarka pre- nycteris the basioccipital is rather broad and pared figure 2; Barbara Pilarzk made figure straight along its boundaries with the tym- 5; and Lisa Lomauro prepared figures 6 and panic cavity and diverges only in its posterior 10 as well as the final layouts for all figures. region. Segall (1971) contrasted this condi- This paper was supported by the Frick Lab- tion, which is characteristic of many mam- oratory Endowment Fund in Vertebrate Pa- mals including megachiropterans, with the leontology at the American Museum ofNat- typically narrow, strongly biconcave basioc- ural History. cipital of microchiropterans. Whereas some microchiropterans do show a very reduced, AUDITORY STRUCTURES biconcave basioccipital (particularly natal- COCHLEA: Segall (1971) remarked that the ids, some rhinolophids, and vespertilionids), type of Icaronycteris index (PU [Princeton in several other microchiropterans (many 4 AMERICAN MUSEUM NOVITATES NO. 2877

Fig. 2. Ventral view ofbasicranium andjaws ofIcaronycteris cf. index; UW (University ofWyoming) 2244. Modified from Novacek (1985a). Symbols are: Co, cochlea; Ect, ectotympanic; Oa, orbicular apophysis of the malleus. For other symbols, see fig. 1. phyllostomids, some mormoopids [e.g., larger in modern microchiropterans than in Pteronotus macleayii] and some emballo- megachiropterans (Pye, 1966; Henson, 1970; nurids) the basioccipital is nearly as broad Novacek, 1980, 1 985a). In Icaronycteris only between the cochleae as is typically found in a rough estimate of cochlear dimensions can megachiropterans. Nevertheless, the basioc- be made by measuring distances between au- cipital is generally narrower and more strong- ditory structures adjacent to the cochlea. For ly biconcave in microchiropterans than in example, the location of the medial border megachiropterans. This condition is not only of the ectotympanic annulus (slightly shifted a function ofthe large cochlea, but is also due during preservation to lie in a more horizon- to the presence of a distinct basicochlear fis- tal position) gives an indication ofthe medial sure medial to the cochlea, an opening less edge of the promontorium cochlea. The re- developed in megachiropterans, and alto- cessed area (the facial canal and the epitym- gether absent in most mammal groups (No- panic recess) below the malleus give some vacek, 1980). Unfortunately the condition in indication of the lateral edge of the promon- Icaronycteris for the basioccipital is very dif- torium cochlea (figs. 1-3). These relation- ficult to determine. Given the distortion and ships can be appreciated in the better pre- damage to this element, which is even more served basicrania of Palaeochiropteryx (fig. severe in the type, PU 18150 (fig. 1) than in 4). UW (University of Wyoming) 2244 (fig. 2), The maximum width of the cochlea is a it is difficult to appreciate the basis for Segall's useful parameter because it is a function of observation on the original specimen. How- the expansion of the basal turn (fig. SB), the ever, a radiograph of the type specimen (fig. region ofthe cochlea particularly sensitive to 3) is more revealing. It suggests, contrary to high-frequency vibrations in adult mammals Segall's (1971) assessment, that Icaronycteris (Dallos, 1973; Henson, 1970; Bruns, 1979; does have a distinct biconcave basioccipital Bruns et al., 1983-84; Harris and Dallos, similar to that in many microchiropterans. 1984). The large cochlea in microchiropter- The problem remains to establish the de- ans is dominated by the basal turn which gree of the development of the cochlea in lends to the structure the appearance of a Icaronycteris, for this structure is clearly much snail shell strongly tapered toward its apex 1987 NOVACEK: EOCENE BAT EARS 5

Fig. 3. Stereophoto-radiographs of ventral skeleton of Icaronycteris index; PU 18150. From Jepsen (1970). Symbols are: Bo, basioccipital; Cf, fourth caudal vertebra; HI, left humerus; Tir, right tibia. 6 AMERICAN MUSEUM NOVITATES NO. 2877

A B 3mm 3mm 5 -

An ,Cp z

SI- / sCow LCo Ect Fig. 4. Palaeochiropteryx tupaiodon. Positive prints ofradiographs of(A) dorsoventral view ofSMF- ME (Senckenberg Naturmuseum und Forschungsinstitut) 788-B and (B) lateral view of SMF-ME 1 127. Modified from Novacek (1985a). For symbols see figs. 1 and 2.

(fig. 5B). By contrast, most other mammals megachiropterans show allometric relation- have cochleae that are not so strongly ta- ships typical of many groups of small to me- pered, due to the smaller basal turn and a dium-size mammals, a contrast with inter- significantly smaller width relative to other esting implications (Graybeal and Novacek, skull elements. in prep.). The fact that maximum cochlear width can Palaeochiropteryx is more accessible to only be estimated in Icaronycteris is irrele- quantitative analysis. Radiographs of speci- vant to broad-based comparisons among both mens from the Senckenberg Museum collec- megachiropteran and microchiropteran tions show beautifully preserved cochlea (fig. species. Figure 6, a plot ofmaximum cochlear 4A). These readily yield large cochlear widths width against skull length for 96 bat species, that lie well within the microchiropteran clearly shows that despite a very liberal range polygon (fig. 6). Note that both Icaronycteris of error in measurement, the estimated al- (by estimate) and Palaeochiropteryx show lometric width ofthe cochlea in Icaronycteris relatively more expanded cochlea than many is well within the polygon defining micro- microchiropteran species (explicit treatment chiropterans. In support of the above-cited of cochlear size variation within the bat sub- observations, the cochlear width is relatively orders is provided by Graybeal and Novacek, much greater in microchiropterans than in in prep.). megachiropterans. Note that the polygons for Although cochlear size might seem a direct the two suborders do not overlap. Also, the function of the degree of emphasis of echo- cochlear width in microchiropterans shows a location, the relationship is by.nomeans clear- weaker correlation and more positive allom- cut. This is due to the problem ofquantifying etry with skull length (r = 0.64, slope = 0.8) in a meaningful way the degree to which a than in the case of megachiropterans (r = species has refined echolocation pulses. It has 0.92, slope = 0.30). These parameters in mi- been noted that the cochlea, especially the crochiropterans are highly unusual, whereas basal turn, seems best developed in species 1987 NOVACEK: EOCENE BAT EARS 7 A B

PGF- PY- -FB SVII- -CO (BT)

Fig. 5. Right basicrania of (A) Pteropus policephalus, SDSU (San Diego State University) 1152 and (B) Rhinolophus sp., SDSU 1215. Both specimens with bulla and ectotympanic removed. Not to scale. Symbols are: AQ, aqueductus cochleae; CF, condyloid foramen; CO, cochlea; CO(BT) basal turn of cochlea; ETR, epitympanic recess; FB, basicochlear fissure; FER, fenestra rotundum; FOO, foramen ovale; H, hamular process; JF, jugular foramen; OC, occipital condyle; PGF, postglenoid foramen; PR, promontorium cochlea; PY, pyriform fenestra; SVII, sulcus for facial (VII) nerve; TE, flange for eustachian tube.

(e.g., Rhinolophus hipposideros, Pteronotus tween these taxa and the echolocating micro- parnellii) that emit a long, constant frequency chiropterans. (CE), pure tone terminated by a brief "flick" MALLEUS: The ear ossicles are not clearly ofan FM (frequency modulated) sweep (Hen- preserved in any ofthe specimens ofPalaeo- son, 1970). These long CF-short FM bats chiropteryx that I have studied. Likewise, the have a region ofthe basilar membrane within type (PU 1850) of Icaronycteris fails to show the basal turn of the cochlear labyrinth (an evidence ofthese elements. A partial malleus "acoustic fovea") that is highly sensitive to is, however, apparently preserved in the Uni- the long pure tone of the bat's vocalization versity of Wyoming specimen of Icaronyc- (Bruns, 1979). Unfortunately, such refined teris (fig. 2) and this bone shows at least one analyses do not pertain to most species. feature of interest. The orbicular apophysis Hence, it is virtually impossible to attribute, at the base of the manubrium of the malleus through analogy with living species, any par- is unusually large, a condition common to ticular "adaptive mode" to cochlear devel- microchiropterans, many small insectivo- opment in Icaronycteris or Palaeochiropte- rans, such as soricids, and many rodents ryx. There is, however, no indication that (Fleischer, 1978, p. 73). By contrast, this pro- these fossil bats were any more conservative cess is weakly developed in megachiropterans in refinement ofthe echolocation system than and most other mammals. are most living species of Microchiroptera. Because of the widespread occurrence of The apparent expansion of the cochlea in this structure it is difficult to attribute a par- Icaronycteris and Palaeochiropteryx is im- ticular mode of auditory sensitivity to the portant in suggesting a close relationship be- presence of the orbicular apophysis. Never- 8 AMERICAN MUSEUM NOVITATES NO. 2877

s.o1

E~~~~~~~ E

3.50

PAL*

w 3.0 CA0 8 0 0 , .0...... 0. 2.5

0

0 0

0 0 2.0

10 15 20 25 30 35 40 45 50 55 60 SKULL LENGTH (mm) Fig. 6. Plot of maximum width of cochlea (ordinate) against skull length (abscissa) for 96 species of Microchiroptera (0), Megachiroptera (0), Palaeochiropteryx (*), and Icaronycteris (vertical line represents range of error in estimate of maximum cochlear width). From Novacek (1985a). Points represent 17 of 19 Recent bat families (two monotypic families of microchiropterans, the Myzopodidae, and Craseo- nycteridae, are omitted). Convex polygons define the outer boundaries of values for each of the two suborders. Cochlear width has a weaker correlation and more positive allometry with skull length in microchiropterans (r = 0.64, slope = 0.80) than in megachiropterans (r = 0.92, slope = 0.30). theless, the presence of this feature in Icaro- because of the relatively enormous orbicular nycteris corroborates the argument that the apophysis. Fleischer (1978, p. 33) further ar- auditory system of this fossil bat shared the gues ". . . thus starting from the ancestral type, specialized auditory condition seen in micro- two radically different lines of evolutionary chiropterans. Moreover, it is of interest that adaptation evidently occurred: in one, the the orbicular apophysis is particularly prom- center of mass is shifted toward and finally inent in groups (e.g., soricids, certain rodents, into the rotational axis, resulting in the freely see Fleischer, 1973, 1978) that display some, mobile type. In the other, the center of mass albeit less refined, mode of high-frequency is shifted further away from the axis by the echolocation (Fenton, 1984). development of the orbicular apophysis, re- Taking these facts into account, Fleischer sulting in the microtype. That the microtype (1978) hypothesized that the malleus-incus is a highly functional ear is easily apparent complex in microchiropterans represents a because it is found in bats, which have a so- strong departure from the primitive condi- phisticated sonar system." tion. In the latter, the center of mass lies near This explanation encounters difficulty, as the axis of rotation of the ossicle system. In the large orbicular apophysis acts to increase the "microtype" exhibited by microchirop- the moment ofinertia ofthe system, lowering terans and the above noted groups, the center its natural resonant frequency. Hence, the of mass is shifted away from this axis, largely apophysis may seem an odd feature in ani- 1987 NOVACEK: EOCENE BAT EARS 9 mals particularly sensitive to ultra high-fre- emission of a series ofhigh-frequency pulses quency sounds. Addressing this apparent during vocalization (Suthers and Fattu, 1973). contradiction, Fleischer (1978, p. 34) consid- The well-developed stylohyal is then an in- ered echolocating bats with the following ar- trinsic component of the specialized laryn- gument: geal-hyoid system in echolocating bats. Its "But these animals also have great prob- similarly strong development in Icaronycte- lems perceiving low frequencies since their ris and Palaeochiropteryx offers additional middle ear is tuned to extremely high fre- support for the argument that these taxa as quencies. Predominately, this is the result of well were accomplished echolocators. It the minute size and the rigid fusion between should be noted that certain other mammal the gonial and tympanic. Without the orbic- groups (e.g., canid and felid carnivores) have ular apophysis, the natural frequency of the a well-developed styohyal and hyoid appa- malleus-incus complex might simply be too ratus, but these conditions seem to relate to high, even for the bats. These relations hold the rapid tongue movement (through action true for the other mammals with a microtype of the stylohyoideus muscle) associated with ear as well, because the middle ear in bats feeding. Presumably, this specialization was does not really differ from theirs." acquired independently of that in microchi- Fleischer's arguments seem reasonable, ropterans. though a bit ad hoc. I leave it to others to In reviewing this manuscript, T. A. Grif- scrutinize their validity with experimenta- fiths mentioned his own observations of fea- tion. In the context ofthis study, the presence tures of the hyoid apparatus in Icaronycteris of the large orbicular apophysis (1) is shared (based on Jepsen's 1970 stereophoto) that by Icaronycteris and microchiropterans, (2) corroborate those given above. In addition is probably a derived mammalian feature, to the stylohyal, he noted the robust propor- and (3) seems to be a consistent component tions of other elements of the hyoid appa- in forms that exhibit some mode ofultrasonic ratus. He also pointed out that the stylohyal's audition. posterolateral end is expanded into a paddle- STYLoHYAL: The stylohyal element is pre- shaped moiety, a feature found throughout served in both the Princeton University (figs. microchiropterans but not in megachiropter- 1, 3) and University ofWyoming (fig. 2) spec- ans (Sprague, 1943). I thank Dr. Griffiths for imens of Icaronycteris and in at least a few sharing this unpublished information with specimens (fig. 4B) ofPalaeochiropteryx. This me. observation by itself is hardly noteworthy. OTHER BAsIcRANLAL FEATURES: In his de- The stylohyal, a component ofthe laryngeal- scription ofthe basicranium ofIcaronycteris, hyoid apparatus, is common in mammals, Segall (1971) mentioned a few features in including megachiropterans, and is clearly a which this taxon resembled either microchi- primitive feature. Nevertheless, in megachi- ropterans or megachiropterans. These fea- ropterans, as in many mammals, the stylo- tures are not easily connected with a partic- hyal is usually a weak, filamentous structure ular function of the auditory system, but are easily lost from macerated skulls (Klaauw, of phylogenetic interest. 1931). That this element does remain in most Based on PU 1850, Segall (1971) observed macerated skulls ofmicrochiropterans is due that the ectotympanic in Icaronycteris was to its more robust development and its strong annular (a complete ring with a dorsal mal- attachment to other elements of the hyoid- leolar plate) rather than horseshoe-shaped laryngeal apparatus. (open dorsally without a well-developed mal- As is shown in the excellent study by Grif- leolar plate). However, the annular mor- fiths (1983), the stylohyal is the site of at- phology of this element, though probable, is tachment for muscles that constrict the phar- not absolutely certain in either the Princeton ynx and brace the hyoid apparatus. The hyoid specimen (figs. 1, 3) or the University ofWy- stabilizers allow the extrinsic laryngeal mus- oming specimen (fig. 2). Moreover, it is not cles to pull from the hyoid and move the clear in specimens of Palaeochiropteryx that larynx forward, an action that precedes the I have studied whether the ectotympanic is 10 AMERICAN MUSEUM NOVITATES NO. 2877 annular or horseshoe-shaped. This ambiguity riorly directed only in its most medial region is unfortunate because the "horseshoe- (fig. 1). By contrast, megachiropterans and shaped" ectotympanic is characteristic of insectivorans show an oblique orientation for megachiropterans, the annular form, of mi- most ofthe length ofthis border (Segall, 1971, crochiropterans (Segall, 1971). It is generally fig. 1G, H). Moreover, the oblique medial argued that the dorsally open ectotympanic portion of this border in Icaronycteris illus- is the primitive mammalian condition trated by Segall (1971, fig. 1J1) may simply (Klaauw, 1931; Segall, 1971; Archibald, 1977; be an artifact, due to the compression of the Novacek, 1980). alisphenoid against the anterior edge of the Perhaps another aspect ofthe ectotympan- glenoid (fig. 1). In Palaeochiropteryx the mor- ic is ofgreater interest. In microchiropterans phology in this region is not clearly preserved this element is anteromedially expanded to (fig. 4). form a large tapered styliform process whose The position of the postglenoid foramen dorsal surface is distinctly concave. Thus, the relative to the tympanic cavity in Icaronyc- styliform process forms a cradle for both the teris is also problematic. The type seems to auditory (eustachian) tube and the tensor preserve the relationship Segall (1971, fig. 1J1) tympani muscle (Stanek, 1933; Henson, reconstructed. However, comparison of this 1970). The process is present in the ectotym- specimen (figs. 1, 3) with UW 2244 (fig. 2) panics of many other mammals, including shows that the tympanic region has been some megachiropterans, but does not ap- shifted somewhat medially in the type, as proach the expanded form seen in microchi- there is clearly a gap between this area and ropterans, where the tensor tympani muscles the postglenoid region. Reconstruction would are unusually large (Stanek, 1933; Henson, yield a condition where the postglenoid fo- 1961, 1970; Novacek, 1980). In Icaronycteris ramen is like that in microchiropterans in the styliform process is not clearly preserved, being situated adjacent to the anterolateral but the ectotympanic does show marked ex- corner of the tympanic cavity. pansion in its anteromedial region (figs. 1- Segall's (1971) description of the lamellar 3). This suggests that the apex ofthis process suprameatal bridge in Icaronycteris seems may have been present but was damaged dur- correct. However, this lamella does not ap- ing fossilization. More definite identification pear so horizontal as it does dorsoventral in is not possible. Specimens of Palaeochirop- its orientation (note the perspective shown teryx do not clearly show such details of the by the stereophotograph in fig. 1). The large ectotympanic. superior meatal lamella in this form is prob- Segall (1971) noted several basicranial fea- ably, as Segall (1971) suggested, a more con- tures of Icaronycteris which resembled echi- servative condition typically modified to a nosoricine hedgehogs and megachiropterans very narrow bridge in most microchiropter- in their generalized, or primitive, status. These ans. I have not been able to identify the mor- are: (1) the more oblique orientation of the phology of this region in specimens of Pa- anterior edge of the glenoid fossa (in micro- laeochiropteryx available to me. chiropterans this border is usually oriented Finally, it should be noted that both Icaro- transversely or even recedes medially), (2) the nycteris (figs. 1-3) and Palaeochiropteryx (fig. relatively more lateral position of the post- 4) resemble microchiropterans in having a glenoid foramen with respect to the antero- very well-defined, hooklike angular process. lateral corner ofthe tympanic cavity, and (3) As Segall (1971) observed, the angular pro- the well-developed horizontal lamella of the cess is much less spurlike in megachiropter- superior external meatus. These features in ans. The significance ofthis difference is moot, Icaronycteris are not easily recognized. The as the hooklike angular process is found in anterior edge of the glenoid region is not so many other mammals, including lipotyphlan oblique in the University of Wyoming spec- insectivorans. However, it seems probable imen (fig. 2). In fact, the orientation of this that the angular process in megachiropterans, border seems more transverse (and more mi- which is of a form common in both euthe- crochiropteran-like) in the type specimen than rians and metatherians, represents the more Segall (1971) suggested. This edge is ante- conservative condition. 1987 NOVACEK: EOCENE BAT EARS I I

AFFINITIES OF ICARONYCTERIS form of the cheek teeth (2 above) in Icaro- AND PALAEOCHIROPTERYX nycteris is either a primitive bat trait over- whelmingly prevalent in Microchiroptera or In the introduction ofthis paper I outlined a derived trait that distinguishes this subor- some ofthe contradictory views regarding the der (see below). The long tail (6) seen in Icaro- affinities ofearliest bats. Two hypotheses seem nycteris is present in a few other groups (e.g., likely: either these bats are bona fide micro- vespertilionids, natalids, rhinopomatids). The chiropterans or they are members of a group cervical vertebrae (4) may actually indicate that if not "ancestral," is more remotely di- a special feature shared with microchiropter- vergent than the modern bat suborders. (The ans (see below). The index claw is undoubt- third possibility, that the fossil taxa are most edly a primitive bat character retained in most closely related to megachiropterans is un- megachiropterans but lost in all living mi- popular and unlikely, but will be given its crochiropterans and Palaeochiropteryx. Al- due in the analysis to follow.) though in most living bats the digits of the In this context, Icaronycteris poses the most hindfoot are of relatively equal length, digit interesting problem. The genus was described 1 (12) is relatively shorter in various species and illustrated in some detail by Jepsen (1966, of pteropodids, emballonurids, and vesper- 1970), who showed that it retained a number tilionids, as well as in Icaronycteris. Finally, of very primitive features either modified or the calcar, while present in most groups, is poorly documented in Palaeochiropteryx. It also absent (13) in most Phyllostomidae and stands to reason then that the latter will cer- some Pteropodidae and in Rhinopomatidae tainly fall in with the Microchiroptera if and Craseonycteridae (Hill, 1974). Icaronycteris does. The converse may not be Remaining then are a few traits which po- true, because Palaeoahiropteryx appears to tentially document a more primitive condi- have a slightly "more derived skeleton" than tion in Icaronycteris than in Recent bat sub- Icaronycteris -cf. Revilliod (1917) and orders. These are: Richter and Storch (1980) with Jepsen (1966, Uncoalescedribs, vertebrae, andsternalele- 1970). ments (fig. 7): In many bats, particularly in Jepsen (1966, 1970) noted the following older individuals, there is a tendency toward characters in which Icaronycteris would be ankylosed fusion of the elements of the cos- regarded as "primitive" or "generalized" or tal-sternal thoracic region (Walton and Wal- as lacking specializations seen in other bats: ton, 1970). Variation in this condition is not (1) large number ofteeth, (2) "insectivorous" well documented, but the sternal elements in shapes of teeth, (3) uncoalesced ribs, verte- modern bats are, according to Vaughan brae, and sternal elements, (4) shapes of cen- (1970a, p. 110), "invariably fused." Hence, tra and neural arches on cervical vertebrae, Icaronycteris appears to retain a more prim- (5) lack of a prominent keel on the meso- itive condition for the sternal elements than sternum, (6) long tail, (7) shape ofthe scapula, in living bat suborders. It should be added (8) relatively short radius, (9) index claw (i.e., that the degree of fusion of these elements is clawed terminal phalanx on digit 2 of wing), very difficult to determine in the Princeton (10) complete phalangeal formula, (11) head specimen. There also remains the possibility and neck of the femur at angle to shaft, (12) that this specimen represents a juvenile in- big toe (digit 1) shorter than others, (13) no dividual where fusion of bony elements is calcar, and (14) low aspect ratio ofwings (es- incipient. timated to have been between 2.75 and 2.84). Lack ofa prominent keel on the mesoster- It is clear that several of these features, num (fig. 7): The mesosternal keel is well de- though primitive, are not confined to Icaro- veloped in many bats (Vaughan, 1970a; Wal- nycteris. Many bats have the dental count of ton and Walton, 1970) and, superficially, 38 (character 1, above) seen in Icaronycteris Icaronycteris appears to be more conserva- (Natalidae, Thyropteridae, Myzopodidae, and tive in lacking this feature. However, the Vespertilionidae including the nearly world- mesosternum in Icaronycteris hardly differs wide genus Myotis-see Vaughan, 1970a). from those ofseveral species (Taphozouspeli, The "insectivorous" (i.e., dilambdodont) Tadarida media, Myotis lucifugus, Rousettus 12 AMERICAN MUSEUM NOVITATES NO. 2877

-4_

-14 Fig. 7. Lateral view ofright side ofskeleton ofIcaronycteris index (PU 18150) x 1.4. Reconstruction originally published as figure 2 in Jepsen, 1970. Numbers correspond to characters listed and explained in figure 10. A minus sign indicates that the corresponding character is lacking in Icaronycteris but is present in the Recent Mega- and Microchiroptera. aegyptiacus, and others -see Walton and the metacarpals III-V seems to differentiate Walton, 1968, 1970; Vaughan, 1970a; Nor- Icaronycteris from most living species, al- berg, 1972; Kingdon, 1974) where the me- though quantitative data for this comparison dian ridge of the mesosternum fails to de- have not been reviewed. Based on this and velop into a prominent keel. Icaronycteris, lengths ofother elements, the wings ofIcaro- like other bats, has a well-developed ventral nycteris were thought by Jepsen (1970) to be keel on the manubrium, a feature which var- very short and broad with a low aspect ratio, ies considerably in form and size within the although his estimate of wing proportions order (Vaughan, 1970a). seems doubtful (see below). Shape ofthe scapula (fig. 7): Modem bats Complete phalangeal formula (fig. 7): In are distinctive in having a scapula with a much Icaronycteris the terminal phalanges ofdigits larger infraspinous fossa than supraspinous III, IV, and V in the wing are nubbins ofsmall fossa. The infraspinous fossa is the site of bones, and there is even a tiny ossified clawlet attachment for important flight muscles on the terminal phalanx of digit V (Jepsen (Vaughan, 1970a, p. 128). In Icaronycteris 1970, fig. 14). In Rhinopoma the distal pha- the supraspinous fossa is relatively larger, but langes of digits II, III, IV, and V carry small hardly more so than in pteropodids (cf. Jep- claw-shaped cartilages that fail to ossify in sen, 1970, figs. 1,9; Vaughan, 1970a, fig. 22a). the adult (Wassif and Madkour, 1963). The It is therefore highly uncertain that the scap- usual number of ossified phalanges in digits ula has a more derived form in all living chi- III, IV, and V is only two in Megachiroptera, ropterans relative to that in Icaronycteris. Rhinopomatidae, Emballonuridae, Noctilio- Relatively short radius (fig. 7): The short nidae, Nycteridae, Megadermatidae, Rhi- length ofthis element in relation to total wing nolophidae, Hipposideridae, Natalidae, and span, length of the humerus, and length of Furipteridae (Walton and Walton, 1970). 1987 NOVACEK: EOCENE BAT EARS 13

Fig. 8. Stereophotographs of ventral view of pelvic region of Icaronycteris index PU 18150. From Jepsen (1970). Symbols are: A, acetabulum; Ap, anterolateral process of seventh lumbar vertebra; Cf, fourth caudal vertebra; F, fibula; Fl, lesser trochanter offemur; G, greater trochanter of femur; Hf, head of femur; Ls, seventh lumbar vertebra; 0, obturator foramen; Ps, pubic symphysis; Sa, sacrum; Sp, spine of pubis.

Three bony phalanges are retained, however, in most modem bat groups except Hipposi- in at least the third digit of the Phyllostom- deridae, Thyropteridae, and Myzopodidae idae, Thyropteridae, Myzopodidae, Mor- (Miller, 1907). moopidae, and Mystacinidae (Hill, 1974). Headand neck offemur at angle to theshaft Miller (1907) presented evidence that the (fig. 8): The head of the femur is offset cra- third phalanx of digit III might be repre- niomedial to the shaft in Icaronycteris but sented by cartilage in those microchiropter- not to the extent seen in terrestrial mammals. ans that have only two phalanges. The ho- In this trait, Icaronycteris is clearly more con- mology of these cartilaginous elements with servative than living mega- and microchi- the terminal phalanges has been questioned ropterans, where the head is nearly aligned (see discussion in Walton and Walton, 1970), with the shaft and the neck is very short or although it is generally accepted that in bats absent (Vaughan, 1970; Walton and Walton, the distal phalanx ofthe third and fourth dig- 1970). its, if present, is always cartilaginous. Thus, Low aspect ratio: The relatively short length the significance of the condition in Icaro- ofthe radius indicates that Icaronycteris had nycteris is that the complete phalangeal for- rather short and broad wings. However, Jep- mula of the wing (2-3-3-3-3) is represented sen's (1970) estimate ofaspect ratio (the ratio entirely by ossified elements. By contrast, of wing length over wing width) of between modem bats either variously reduce this pha- 2.75 and 2.84 seems extraordinarily low, as langeal formula in one or more of digits II- these ratios generally range between 5.6 and V or have only cartilaginous terminal pha- 10 (Vaughan, 1970b; Norberg, 1972). The langes. The complete phalangeal formula (2-3- standard formula of calculating aspect ratios 3-3-3) ofthe pes in Icaronycteris is also found is wing span2/wing area, due to the irregular 14 AMERICAN MUSEUM NOVITATES NO. 2877

shape of the wing. A precise calculation is talpid insectivorans, didelphid marsupials, virtually impossible in Icaronycteris, al- and a variety ofother extant and fossil groups. though from reconstruction of approximate The origin of the peculiar molar crown pat- wing shape (Jepsen, 1966, cover illustration) tern in megachiropterans is highly uncertain. and measurements oflimb elements (Jepsen, There is no evidence for or against the sup- 1970), I estimate an approximate aspect ratio position that this pattern derived from di- of 5.0. Hence, the aspect ratio for the wings lambdodonty (Koopman and MacIntyre, in Icaronycteris may have been low but not 1980). Within Microchiroptera, however, the significantly lower than in certain living bats highly specialized molars of desmodontines (e.g., Artibeus, Rousettus, see Vaughan, 1 970b; and nectar-feeding bats are probably a de- Norberg, 1972). rivitive of the dilambdodont condition These considerations shorten the list of (Phillips, 1971; Griffiths, 1985). The latter primitive features found in Icaronycteris that construction, particularly well suited for mas- are lost or modified in the Recent mega- and tication of hard insects parts, is undoubtedly microchiropterans. I would only recognize the primitive crown morphology for micro- four such features: the unfused sternal ele- chiropterans, yet it cannot be taken as the ments, the relatively short radius (although primitive state for all bats and the morpho- this character has not been quantitatively as- typical condition for the order remains am- sessed), the complete phalangeal formula for biguous. An additional feature shared by the wing (2-3-3-3-3) represented entirely by Icaronycteris and microchiropterans is nyc- ossified elements, and the canted head and talodonty (where the hypoconulid ofthe low- neck of the femur. er molars is situated far lingual and a well- By themselves, then, these features suggest defined postcristid runs to the hypoconulid- that all modern bats form a group that ex- see Menu and Sige, 1971). It is not clear that cludes Icaronycteris in fusion of the sternal nyctalodonty is primitive for Microchirop- elements, lengthening of the radius, loss of tera as it is absent in some bats (Van Valen, phalanges or retention of (embryonic) car- 1979). tilaginous phalanges in some digits of the Interlocking accessoryprocesses on the me- wing, and reduction of the femoral neck and dial-ventral section ofcervical vertebrae: These alignment of the head near the axis of the processes are developed on cervical vertebrae shaft. 2 an 3 in Icaronycteris, but in some living Contradicting this argument are a number microchiropterans they are found on verte- ofspecializations shared by Icaronycteris and brae 2-5. These structures appear to confer microchiropterans, but lacking in megachi- stability to the vertebral column during strong ropterans. These include the basicranial fea- backward flexion of the neck. As a result, tures noted above, namely (1) enlarged coch- microchiropterans are able to thrust their lea, (2) the biconcave basioccipital ?, (3) the heads down and backwards (away from the large orbicular apophysis of the malleus, (4) chest) during roosting so that the head is ac- the well-developed, ossified stylohyal, (5) the tually right side up. Such modifications in annular ectotympanic ?, (6) the hooklike an- structure and behavior are not shared by gular process of the mandible, and (7) the megachiropterans (see discussion in Jepsen, development of ultrasonic echolocation that 1970). Apparently, Icaronycteris shared the is inferred from the presence offeatures 1, 2, unusual cervical condition with living mi- 3, and 4. crochiropterans. In addition to these, Icaronycteris shares Narrow neural arches on cervical vertebrae with microchiropterans several other poten- separated by gaps equal to or greater than the tial synapomorphies. width of the arches: This modification also Dilambdodont upper molars (fig. 1): The promotes extreme flexion of the neck. In dilambdodont crown pattern-wherein the megachiropterans the cervical neural arches labial cusps are linked by a high, sharp are ofa more general design, being developed W-shaped ectoloph-is probably a special- as broader lamellae. ization that evolved several times within Eu- Pectoral and deltoid ridges of humerus theria. Dilambdodonty is found in soricid and united as a singleflange (fig. 9): This feature 1987 NOVACEK:NEOCENE BAT EARS 15

simm FT, -

Fig. 9. Stereophotographs ofright shouldergirdle ofIcaronycteris index. From Jepsen (1970). Symbols are: Ac, acromion process; Cv, clavicle; Gt, greater tuberosity (trochiter) of humerus; Hh, head of humerus; Hp, pectoral ridge of humerus. is found in microchiropterans, Icaronycteris, midline, long nasal bones, and coronoid pro- Palaeochiropteryx, and other Eocene chirop- cess of mandible with rounded superior bor- terans known from skeletal features, but is der and claw on index finger (digit II). The lacking in Archaeopteropus and Recent presence ofunited or separate premaxillaries Megachiroptera (Van Valen, 1979). The varies within both bat suborders, and other "deltoid-pectoral crest" is also lacking in characters are doubtlessly primitive. Hence, Craseonycteridae but the condition in that none ofthese features provide clear evidence taxon seems divergently specialized (Hill, for the close affinities of Icaronycteris with 1974). The crest is the site of insertion for megachiropterans, and this hypothesis can be the deltoid and pectoralis muscles, which are regarded as unsubstantiated. extremely important for control of the wing- The argument favored here, then, is that beat cycle (Vaughan, 1970a). both Icaronycteris and Palaeochiropteryx are The addition of these features to the basi- close relatives of living microchiropterans, cranial traits noted above muster a strong and (at least the former genus) retain(s) a few case for the close relationship ofIcaronycteris primitive skeletal traits independently mod- with microchiropterans (fig. 10). If this hy- ified in megachiropterans and other micro- pothesis is accepted, the question ofrelation- chiropterans, as shown in figure 10. This ships of Palaeochiropteryx is easily settled, scheme favors the idea that the Eocene bats for it shares with Icaronycteris and micro- (at least Icaronycteris) while a member ofMi- chiropterans the specializations noted above crochiroptera, are an outgroup to all of the (where known) and does not retain some of Recent families of this suborder. However, the primitive traits (e.g., index claw, rela- the possibility that these Eocene taxa have tively short radius) found in the North Amer- special affinities within Microchiroptera (i.e., ican fossil (see Revilliod, 1917; Richter and are related to one or another microchirop- Storch, 1980). teran family) cannot, at present, be dis- Jepsen (1966, 1970) noted several char- missed. In the meantime, the Palaeochirop- acters in which Icaronycteris resembled terygoidea seems merely an artificial megachiropterans but not microchiropter- convention to group several early microchi- ans. These are: premaxillaries not united at ropterans whose relationships with modem 16 AMERICAN MUSEUM NOVITATES NO. 2877

12,13?,14,15 123,137,14,15

11,2?,3,4,5?,9 6,7,8,9,10,11?

Fig. 10. Cladogram favoring close relationship between Icaronycteris and Recent Microchiroptera showing both supportive and contradictory characters. ? indicates an incompletely documented character or a condition inferred from the morphological evidence. Numbers refer to the following characters: (1) enlarged cochlea, (2) biconcave basioccipital, (3) large orbicular apophysis ofmalleus, (4) well-developed ossified stylohyal, (5) annular ectotympanic, (6) hooklike angular process of the mandible, (7) dilamb- dodont upper molars, (8) interlocking accessory process on cervical vertebrae, (9) narrow, biconcave neural arches on cervical vertebrae, (10) pectoral and deltoid ridges of humerus united as single flange, (1 1) ultrasonic echolocation as inferred from characters 1, 2, 3, and 4, (12) fusion of sternal elements, (13) lengthening of radius, (14) complete phalangeal formula (2-3-3-3-3) reduced or terminal phalanges unossified, (15) neck of femur reduced or lost and head aligned with long axis of shaft.

families ofthis suborder remain poorly known chiropteran (fig. 10) are insignificant in com- (see Smith and Storch, 1981). Hence, Icaro- parison to the complex of skeletal features nycteris and Palaeochiropteryx are more ac- (primarily relating to wing structure) shared curately designated as Microchiroptera in- by these groups and found in no other mam- certae sedis. Moreover, there seems no mals. Neither this paper nor any other pre- justification for a formal designation of an sents crucial evidence that refutes the concept "ancestral" group (e.g., Eochiroptera sensu of bat monophyly. Van Valen, 1979) to distinguish Icaronycter- In conclusion, the foregoing analysis is is, Palaeochiropterxy, and other Eocene bats consistent with the view that a sophisticated from the Recent chiropteran suborders. echolocating system was present in the ear- It is noted in passing that the scheme shown liest known bats, and was acquired before in figure 10 might be construed by advocates some modifications of the postcranial skel- of the diphyletic origin of bats (e.g., Petti- eton. That the earliest bats can be allied with grew, 1986) as supportive evidence. It is im- a particular bat suborder suggests that the portant, however, to remind such advocates order arose much earlier than their first ap- that the postcranial features thought to evolve pearance in the fossil record. How much ear- independently in modem mega- and micro- lier is anybody's guess. 1987 NOVACEK: EOCENE BAT EARS 17

LITERATURE CITED land. Bull. Brit. Mus. Nat. Hist., Zool., 27:301-336. Archibald, J. D. Jepsen, G. L. 1977. Ectotympanic bone and internal carotid 1966. Early Eocene bat from Wyoming. Sci- circulation ofeutherians in reference to ence, 154:1333-1339. anthropoid origin. J. Human Evol., 6: 1970. Bat origin and evolution. In W. A. 609-622. Wimsatt (ed.), Biology ofthe bats, vol. Bruns, V. 2, pp. 1-65. New York: Academic Press. 1979. Functional anatomy as an approach to Kingdon, J. frequency analysis in the mammalian 1974. East African mammals. Vol. 2A (Insec- cochlea. Verhandl. Deutch. Zool. Ge- tivores and bats). Chicago: Univ. Chi- sellsch., 1979:141-154. cago Press, 341 pp. Bruns, V., J. Fiedler, and H.-J. Kraus Klaauw, C. J., van der 1983-84. Structural diversity of the inner ear 1931. The auditory bulla in some fossil mam- of bats. Myotis, 21-22:52-60. mals. Bull. Am. Mus. Nat. Hist., 62:1- Dallos, P. 352. 1973. The auditory periphery: biophysics and Koopman, K. F., and G. T. Maclntyre physiology. New York: Academic Press, 1980. Phylogenetic analysis of chiropteran 548 pp. dentition. In D. E. Wilson and A. L. Fenton, M. B. Gardner (eds.), Proceedings Fifth Inter- 1984. Echolocation: implications for ecology national Bat Research Conference, pp. and evolution ofbats. Q. Rev. Biol., 59: 279-288. Lubbock: Texas Tech Press. 33-53. Menu, H., and B. Sige Fleischer, G. V. 1971. Nyctalodontie et myotodontie: impor- 1973. Studien am Skelett des Gehororgans der tants caracteres des grades evolutifs chez Saiugetiere, einschliesslich des Men- les Chiropteres entomophages. Comp. schen. Sdugetierk. Mitt., 21:131-239. Rend. Seances Acad. Sci., ser. D, 272: 1978. Evolutionary principles of the mam- 1735-1738. malian middle ear. Adv. Anat. Embry. Cell Biol., 55:1-70. Meschinelli, L. Griffiths, T. A. 1903. Un nuovo chiroterro fossile (Archaeop- 1983. Comparative laryngeal anatomy of the teropus transiens Mesch.) delle ligniti Big Brown bat, Eptesicusfuscus, and the de Monteviale. Atti lst. Veneto Sci., 62: Mustached bat, Pteronotus parnellii. 1329-1344. Mammalia, 47:377-394. Miller, G. S. 1985. Molar cusp patterns in the bat genus 1907. The families and genera of bats. Bull. Brachyphylla: some functional and sys- U.S. Natl. Mus., 57: 1-282. tematic observations. J. Mammalogy, Norberg, U. M. 66:544-549. 1972. Functional osteology and myology ofthe Harris, D. M., and P. Dallos wing ofthe dog-faced bat Rousettus ae- 1984. Ontogenetic changes and frequency gyptiacus (E. Geoffroy) (Mammalia, mapping of the mammalian ear. Sci- Chiroptera). Zeit. Morphol. Tiere, 73: ence, 225:741-743. 1-44. Heller, F. Novacek, M. J. 1935. Fledermiiuse aus der eozanen Braun- 1980. Phylogenetic analysis ofthe chiropteran kohle des Geiseltales by Halle a.S. Nova auditory region. In D. E. Wilson and A. Acta Leopoldina, 2:301-314. L. Gardner (eds.), Proceedings Fifth In- Henson, 0. W. ternationalBat Research Conference, pp. 1961. Some morphological and functional as- 317-330. Lubbock: Texas Tech Press. pects ofcertain structures ofthe middle 1985a. Evidence for echolocation in the oldest ear in bats and insectivores. Univ. Kan- known bats. Nature, 315:140-141. sas Sci. Bull., 42:151-255. 1985b. Comparative morphology ofthe bat au- 1970. The ear and audition. In W. A. Wimsatt ditory region. Fortschr. Zool., 30:149- (ed.), Biology ofthe bats, vol. 2, pp. 181- 151. 263. New York: Academic Press. Pettigrew, J. D. Hill, J. E. 1986. Flying primates? Megabats have the ad- 1974. A new family, genus, and species of bat vanced pathway from eye to midbrain. (Mammalia, Chiroptera) from Thai- Science, 231:1304-1306. 18 AMERICAN MUSEUM NOVITATES NO. 2877

Phillips, C. J. with especial reference to the bats. Am. 1971. The dentition ofglossophagine bats: de- J. Anat., 72:385-472. velopment, morphological characteris- Stanek, V. J. tics, variation, pathology, and evolu- 1933. K. Topograficke a Strovnavaci Anato- tion. Misc. Publ. Mus. Nat. Hist., Univ. mii Sluchoveho Organu Nasich Chirop- Kansas, 54:1-138. ter.NdklademCeske6Akad., Vjdaume'ni, Pye, A. Prague, 67 pp. 1966. The structure of the cochlea in Chirop- Suthers, R. A., and J. M. Fattu tera. II. Megachiroptera and Vesper- 1973. Mechanisms of sound production by tilionoidea of the Microchiroptera. J. echolocating bats. Am. Zool., 13:1215- Morphol., 119:101-120. 1226. Revilliod, P. Van Valen, L. 1917. Fledermause aus der Braunkohle von 1979. The evolution of bats. Evol. Theory, 4: Messel bei Darmstadt. Abh. Hessischen 103-121. Geol. Landesanst., 7:157-201. Vaughan, T. A. Richter, G., and G. Storch 1970a. The skeletal system. In W. A. Wimsatt 1980. Beitrage zur Erniihrungsbiologie eo- (ed.), Biology ofbats, pp. 97-138. New ziiner Fledermiiuse aus der Grube Mes- York: Academic Press. sel. Natur und Museum, 110:353-367. 1970b. Flight patterns and aerodynamics. In W. Russell, D. E., P. Louis, and D. E. Savage A. Wimsatt (ed.), Biology of bats, pp. 1973. Chiroptera and Dermoptera of the 195-216. New York: Academic Press. French early Eocene. U. Calif. Publ. Walton, D. W., and G. M. Walton Geol. Sci., 95:1-57. 1968. Comparative osteology ofthe pelvic and Russell, D. E., and B. Sige pectoral girdles ofthe Phyllostomatidae 1970. Revision des chiropteres lutetiens de (Chiroptera, Mammalia). J. Grad. Res. Messel (Hesse, Allemagne). Palaeover- Center, So. Methodist Univ., 37:1-35. tebrata, 3:83-182. 1970. Postcranial osteology of bats. In B. H. Segall, W. Slaughter and D. W. Walton (eds.), 1971. Auditory region in bats including Icaro- About bats, pp. 93-126. Dallas: So. nycteris index. Fieldiana, Zool., 58:103- Methodist Univ. Press. 108. Wassif, K., and G. Madkour Sige, B., and D. E. Russell 1963. Studies on the osteology ofthe rat-tailed 1980. Complements sur les chiropteres de bats of the genus Rhinopoma found in l'Eocene moyen d'Europe. Les genres Egypt. Zool. Soc. Egypt, Bull., 18:56- Palaeochiropteryx et Cecilionycteris. 80. Palaeovertebrata, Mem. Jubil. R. Lav- Winge, H. ocat, 91-126. 1923. Pattedyr-Slaegter. 1. Monotremata, Smith, J. D., and G. Storch Marsupialia, Insectivora, Chiroptera, 1981. New middle Eocene bats from "Grube Edentata. Copenhagen, 360 pp. [Trans- Messel," near Darmstadt, W-Germany. lated by E. Deichmann and G. M. Allen, Senckenbergiana Biol., 61:153-167. 1941, The interrelationships of the Sprague, J. M. mammalian genera, vol. 1. C. A. Reit- 1943. The hyoid region ofplacental mammals zels, ed., Copenhagen, 418 pp.]

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates, Bulletin, and Anthropological Papers published during the last five years are available l free of charge. Address orders to: American Museum of Natural History Library, Department D, Central Park West at 79th St., New York, New York 10024.