AMERICAN MUSEUM NOVITATESI Published by Number 1140 THE AMERICAN MUSEUM OF NATURAL HISTORY August 20, 1941 New York City

RESULTS OF THE ARCHBOLD EXPEDITIONS. No. 36 REMARKS ON SOME OLD WORLD LEAF-NOSED BY G. H. H. TATE

When reviewing recently the genus the cochleae (not so large, however, as in ,I it became necessary to study H. muscinus). other available hipposiderine genera, to re- Many of the following notes are based examine Rhinolophus, and to some extent upon specimens kindly lent us by the Cu- to study the remaining leaf-nosed bats, the rators of at Washington, Chicago Megadermidae and Nyeteridae. and Cambridge. Material referable to , Anthops, Cloeotis, , , Rhinolophus, Anthops ornatus Thomas Megaderma, Lavia, Nycteris, Lyroderma U.S.N.M. 123441, Guadalcanar. was examined. (Rhinonycteris is appar- Ears much as the "emarginate ears" of ently unrepresented in American collec- members of Hipposideros, but with anti- tions.) Notes made upon their compara- tragal fold somewhat larger. Horseshoe tive structures are presented herewith. with two lateral leaflets, the inner quite The hipposiderine genera are considered small, the outer large. Transverse leaf first, then briefly the Nyeteridae and with three raised, rounded processes, each Megadermidae. The isolated position of hollowed out behind and each representing Coelops is pointed out. Only incidental the extension of the three thickened septa remarks are offered on the Rhinolophinae, which in front support the leaf (as in H. reviewed two years ago2 and now in course larvatus). The transverse leaf subtended of extensive revision by C. C. Sanborn. by two small lateral leaflets of its own, A list of materials belonging to these separate from those margining the horse- genera contained in the Archbold collec- shoe. Traces of a small frontal sac. tions is appended. Tail very short. Skull of the short, massive type seen in Hipposiderinae Hipposideros speoris and larvatus. Broken. The rostral portion triangular, the rostral ANTHOPS THOMAS swellings at the apices of the triangle not Anthops THOMAS, 1888, Ann. Mag. Nat. extremely prominent. Infraorbital fora- Hist., (6) I, p. 156; 1888, Proc. Zool. Soc. inclosed a bar. London, p. 477.-MILLER, 1906, Bull. U. S. Nat. men by Intertemporal Mus., LVII, p. 113. constriction pronounced. Cochlea and GENOTYPE.-A. ornatus Thomas from Aola ring not greatly enlarged (diameters 3.4 Guadalcanar. and 2.8). Back of palate level with M2. Thomas distinguished Anthops from Incisive foramina closed by maxilla. Back Hipposideros by its short tail, which Miller of premaxillae rounded. External lobe says is less than half the length of the fe- of incisors virtually obsolete; c with low mur and contains four caudal vertebrae, posterior cusp; p2 neither greatly reduced and by the projections of the transverse (cingulum length, 0.7) nor wholly out of noseleaf being "rounded and hollow be- toothrow; m3 with partly reduced W- hind." Miller indicated the larger size of pattern; rami of mandibles at symphysis in- I "Results of the Archbold Expeditions. No. 35," forming a narrow inverted V; outer 1941, Bull. Amer. Mus. Nat. Hist., LXXVIII, Art. 5. cisors slightly thickened, inner ones 2 Tate and Archbold, 1939, Amer. Mus. Novit., No. 1036, pp. 1-12. strongly overlapping; 2 AMERICAN MUSEUM NOVITATES [No. 1140

P2 unreduced cingulum length P2 1.10 Upper incisors clearly bilobate, but very ' cingulum length P4 1.15 small and with the roots exceedingly deli- Back of skull destroyed. cate, divergent, placed at the outer corners Sanborn records the species from Choi- of the premaxillae as in Asellia and H. seul, Troughton from Ysabel and Bougain- commersonii; p2 still with small cusp and ville Islands. retained in toothrow; m3 less reduced than Contrary to statements by authors, in Asellia tridens; toothrows almost par- Anthops is not very close to Asellia. It allel as in H. muscinuts. Lower incisors, differs in several features: its much heavier as in H. muscinus, scarcely at all imbri- incisors, the retention of p2, the form of cated; P2 unreduced, its cingulum length rostrum and sagittal crest, the less re- actually greater than that of P4 (0.9 : 0.8), duced m3 and P2- but markedly compressed laterally (width, The type of A. ornatuts, B.M. 88.1.5.16, 0.5 to 0.6). old adult female, two paratypes, B.M. Tricuspidatus, placed by Dobson and 88.1.5.17 and 88.1.5.28, female and male, by Peters in a special division of Hipposi- and a juvenal male topotype, B.M. deros, betrays affinities with Anthops, 89.4.3.5, all in alcohol, were examined by Asellia, Triaenops and Cloeotis by the divi- me. The forearms measuredl 50, 51, 48.5 sion of the upper edge of the noseleaf into and 37 mm., respectively. three points. It appears nevertheless to My measurements of the type skull are: represent a distinct line of development greatest length, 20.0 mm.; zygomatic which may be distinguished as breadth, 10.4; mastoid breadth, 9.0; breadth of braincase, 8.4; least inter- NEW GENUS orbital width, 2.0; cochlea, 3.0; c-m3, GENOTYPE. Rhinolophuts tricuspidatus Tem- minck from Amboina. 7.0. DESCRIPTION.-A Hipposiderine genus of bats with upper margin of transverse noseleaf di- "Rhinolophus" tricuspidatus Temminck vided into three points, no papilla on the sella, two lateral leaflets margining horseshoe, iio A.M.N.H. 109987, Hollandia, Dutch frontal sac in either sex, tail extending con- New Guinea. siderably beyond membrane, as in Asellia. Skull with large separate rostal inflations, A quite small species with horseshoe premaxillary bones divergent ini front, cochlea unnotched in front, two lateral leaflets, rather large. the outer pair connecting with the trans- Teeth. Incisors bilobed, smnall, divergent; verse leaf above and uniting in front canine with low posterior cusp; p2 included in toothrow; r3 scarcely i'educed; P2 low, un- beneath the horseshoe; no lateral cells as crowded, laterally comnpressed. Toothrows in Triaenops; no papilla-like process on parallel. "sella;" transverse leaf divided into three It seems probable that stoliczkanus' and parts somewhat as in Asellia tridens, but 1 Since this manuscript was completed a small the lateral processes much less elevated. series of "Asellia" heeleri Osgood from Chipwi, Ears shaped as in H. galeritus, although northern Burma, has been examined. The species is a little smaller than tricuspidatus, with forearm, somewhat narrower and more acute and 49 mm., instead of 54; the ear less elongated and more like the car in the H. galeritus group; the tail thus approaching H. muscinus in form. elongate instead of short (40 mm.: 25 mm.) and its Tail considerably exserted beyond mem- tip extends two or three millimeters beyond the membrane. brane. The noseleaf, surrounded by rather dense facial hair, somewhat as irn Coelops, is very like that of Skull somewhat narrow, the zygomata tricuspidatus, with three small points on the trans- only slightly exceeding mastoid width, and verse leaf, which terminate three supportiing ribs or septa. There are two lateral leaflets. The front rostral expanse only slightly less. Rostral of the horseshoe shows indications of a cleft. No frontal sac is present in the male. Wheeleri and swellings large, high. Zygomatic emi- stoliczkanus appear closely related. nence rising gradually in front; high be- The skull is slightly reminiscent of the skull of Coelops, in that it is low, and the canine-bearing hind. Cochlea rather large and fairly portion and premaxillae are elongated; while the canines are strongly proodont, though devoid of the close together (2.1 : 1.0). Premaxillae strongly developed supplemental cusps of Coelops. weakly united, their contact with palate V- The skull represents a quite advanced Hipposider- ine near Aselliscus; the premaxillae though longer shaped. preserve the same relations with the upper incisors, 1941 ] RESULTS OF THE ARCHBOLD EXPEDITIONS. 36 trifida Peters from the Malay and Burma shoe is broadened to form two angled ex- region must also be referred to Aselliscus. pansions. The ears are simple, somewhat The new genus is distinguished from acute and appear to be slightly emarginate. Anthops and Triaenops by its very different Canine with large secondary cusp, and premaxillae, zygoma and upper incisors, peculiarities of premaxillae, rostrum, zy- from Cloeotis by its relatively low jugal emi- goma and sagittal crest (Miller, 1907). nence and much deeper skull, from Asellia R. aurantius is found chiefly in northern (tridens) by its lower sagittal crest, much and northwestern Australia. Wood-Jones inflated rostrum, pore-like infraorbital writes that it occurs rarely in South Aus- foramen and parallel toothrows, and from tralia. Rhiornycteris by the several characters given for the last by Dobson (1878). TRIAENOPS DoBSON The geographical range of A. tricus- Triaenops DOBSON, 1871, J. Asiatic Soc. Ben- includes Amboina, Buru (Dammer- gal, XL, 2, p. 455. pidatus GENOTYPE.-Triaenops persicus Dobson, from man), the north coast of Dutch New Shiraz, Persia. Guinea, Batjan (Peters), Duke-of-York Island (Dobson), Morty Island, New Additional forms named: afer Peters Mombasa Britain, and New Ireland (Dobson), Kei furcula Trouessart W. Madagascar (Peters and Doria), Aola, Guadalcanar humbloti Milne Edwards E. Madagascar (Thomas), Woodlark Island (Thomas), ruifus Milne Edwards E. Madagascar Trobriand Island (Heller), Solomon Island aurita Grandidier Diego Suarez furinea Trouessart Aldabra Islands and Santa Cruz Island (Sanborn). ?wheeleri Osgood Tonkin RHINONYCTERIS GRAY With the exception of wheeleri Osgood,t GRAY, 1847, Proc. Zool. Soc., already dealt with under Aselliscus, this London, p. 16. genus is indigenous to the part of the world Rhinonycteris GRAY, 1866, Proc. Zool. Soc., which includes Persia, Arabia, East Africa London, p. 81. and the Mascarene group of islands. GENOTYPE.-Rhinolophus aurantius Gray, from North Australia. Rhinonycteris is unrepresented in our Triaenops persicus Dobson collection. Study of the illustration of the U.S.N.M. 123439, Arabia. noseleaves shown by Dobson, 1878 (P1. Ears short, acute, very like the emargi- viii, fig. 2), shows a few cellular structures nate ear in the galeritus group, but with a lateral to the transverse leaf, development well-developed fold at anterior edge, also a of the anterior process of the internasal low tragus inside antitragal lobe. Horse- septum and a medial pit in front of the shoe not cleft, but interrupted by flat transverse leaf, much like those of the rectangular process from internasal sep- Triaenops. But the transverse leaf is tum. Supplemental leaflets two, the inner simple and rounded above, slightly pointed interrupted at each side near front, and at the middle. There appears to be a its anterior parts uniting smoothly in ad- median papilla on the "sella." The horse- vance of horseshoe. The posterior half divides and anastomoses to form a com- which are similarly delicate in structure; the molar plex system of six cells between the eye, series are parallel; p2 is similar in size, shape and position; m3 has the W-pattern similarly unreduced. the horseshoe and the transverse leaf. It differs in possessing a quite large posterior zygo- The outer leaflet small and broad. matic eminence somewhat as in Triaenops, Asellia and Cloeotis and in the reduced, compressed form of The transverse leaf modified into three p2. From Triaenops it is distinguished by its much high, pointed processes, much as in Cloeotis. narrower rostrum in proportion to braincase and in In front of the central one of the processes other characters (see T. persicus beyond). From Cloeotis and Asellia (both described beyond) it ex- a large median pit, and in front of the pit hibits a number of distinguishing characters. Per- haps it should be accorded generic separation. the "sella," upon which stands a high, The status of stoliczkcanus Dobson can be cleared pointed, laterally compressed papilla, the up only by restudy of the type. Dobson's illustra- tion (1878, pl. 8), of the face of stoliczkanus, which may be poorly drawn, seems to differ from wheeleri 11932, Mammals of the Kelly-Roosevelt Expedi- in several particulars. tion, p. 224. 4 AMERICAN MUSEUM NOVITATES [No. 1140 equivalent of similar organs in Cloeotis and and between center and outer margin a H. muscinus. small eminence (on each side). Thumbs Skull rather high and with high rostrum, quite small. Frontal sac? Tail not re- canine-bearing part extended forwards duced (7 joints). (as in H. pratti); rostrum broad, the swell- Skull short, low, broad (like Hipposideros ings individual and their inflations in the dyacorum). Rostral swellings moderate, frontal rather than in the maxillary portion rostrum low; intertemporal constriction of the rostrum; jugal eminence very large marked; premaxillae not clearly visible; and high, occupying almost the whole palate to half m2-2 only, with a broad length of the zygoma; squaniosal roots of spine; zygomata wide, jugal eminences, zygomata little expanded; premaxillae contrary to Thomas' description large, and indistinct (not described); roofs of ptery- unlike those of Asellia, inclined backwards, goid and narial canals with step-like discon- limited to portion of zygoma directly over tinuity as in H. armiger; cochlea rather glenoid; glenoid surfaces broad and flat; large; a foramen developed in the web Eustachian tubes well developed; cochlear of bone tying together the angular and ar- spirals visible (large); tympanic rings not ticular processes of mandible. enlarged; angular processes connected by Outer lobe of upper incisor well de- web of bone with articular; coronoid low. veloped, and teeth convergent, approxi- Upper incisors two-lobed but very small, mated, placed towards center of combined the roots above the crowns very slender as premaxillae; canine with definite posterior in Aselliscus; c with posterior cusp descend- cusp to half its length; P2 excluded, not ex- ing for half its length; p2 greatly reduced, tremely reduced; ml-' with W-pattern excluded; p4 well developed; hypocones only slightly reduced. Outer of three weak, internal interdental spaces large; lobes of outer lower incisor obsolescent; m3 with W-pattern almost complete. P2 P2 low but cingulum length not reduced. not much reduced, low, but its cingulum length not shortened. CLOEOTIs THOMAS Cloeotis THOMAS, 1901, Ann. Mag. Nat. ASELLIA GRAY Hist., (7) VIII, p. 28. Asellia GRAY, 1838, Mag. Zool. Bot., II, p. GENOTYPE.-ClOeOtiS percivali Thomas from 493; 1843, List Spec. Mamm. British Mus., British East Africa. p. 24; 1866, Proc. Zool. Soc. London, p. 81. Additional form since named: australis GENOTYPE.-RhinolophUs tridenr Geoffroy (by Roberts. monotypy) from Egypt. Additional forms since described: murraiana J. Anderson Karachi, India Cloeotis percivali Thomas patrizii de Beaux Ethiopia U.S.N.M. 154589, Rhodesia, South italosomalica de Beaux Italian Somaliland Africa. pallida Laurent French Morocco A very small with forearm 34 mm. diluta Andersen Algerian Sahara Ears short, low, rounded, only the suspicion Asellia, excluding tricuspidata and its of a "tip" perceivable. Horseshoe deeply allies for which the genus Aselliscus has cleft in front, the two sides of the cleft now been proposed, is seen to range across raised and produced as two small processes. arid northern Africa, Arabia and Iraq to Two lateral leaflets, the outer larger, coa- northwest India. lesce upwards to form a single fold which Below is a redescription of A. tridens reaches the side of the transverse leaf. murraiana J. Anderson, 1881, based upon Transverse leaf modified into three high material at the Field Museum of Natural pointed processes which rise from the three History. septa. Transverse leaf subtended on each side by large lateral papilla, which low Asellia tridens murriana J. Anderson cutaneous ridges connect with inner can- F.M.N.H. 48728, d , Baghdad. .thus of eye and with transverse leaf. Ears much like H. speoris, with anti- Sella with small clubbed process at center tragal notch and groove, no tragus, outer 1941] RESULTS OF THE ARCHBOLD EXPEDITIONS. 36 S edge emarginate, no costal fold as Tri- 1878, DOBSON, Cat. Mamm. Brit. Mus., pp. aenops. 152-153, reviewed Coelops. 1907, MILLER, Bull. U. S. Nat. Mus., LVII, Horseshoe weakly notched in front, no pp. 113-114, reviewed Coelops. trace of process from internarial region; 1908, BONHOTE, J. Fed. Malay States Mus., lateral leaflets two, the inner smaller; III, p. 4, described C. robinsoni from transverse leaf with three Pahang. processes, the 1911, MILLER, 10c. cit., described Chilophylla median high and narrow, the lateral ones hirsuta from Philippines. nearly as high but much broader; no 1912, MILLER, P. Biol. Soc. Wash., XXV, p. 117, papilla on sella (see Triaenops, Cloeotis, remarked on dentition and skull of "Chilo- and H. muscalis and cyclops), no complex phylla hirsutum," not from Philippines but from Federated Malay States (thus of cells lateral to noseleaves (Triaenops). probably C. robinsoni). Tail extending considerably (one joint) 1928, HORIKAWA, Trans. Nat. Hist. Soc. For- beyond uropatagium. mosa, XVIII, No. 98, described C. for- Skull with Mosanus. well-expanded zygomata, 1928, G. M. ALLEN, Amer. Mus. Novit., No. jugal eminence not enlarged as in Tri- 317, p. 4, described C. sinicus from aenops, rostrum wide and almost as high Szechwan. as braincase without sagittal crest; pre- 1928, MILLER, P. Biol. Soc. Wash., XLI, p. 85, maxillae strongly fused, rounded pos- synonymized his Chilophylla with Coelops (in footnote) and described as new C. teriorly at contact with palate, their inflata from Fukien. lateral processes nearly contiguous with 1929, OSGOOD, Field Mus. N. H., Zool., XVIII, maxilla and not enclosing incisive fo- pp. 226-228, discussed the genus. He ramina; back of palate level with backs of indicated the supplementary canine cusps. m3-3; cochlea more than twice as wide as The named forms of Coelops may be their distance apart (2.8: 1.0). listed as follows: Upper incisors peg-like (outer lobe ob- frithii Blyth Bengal solete), placed at bernsteini Peters Java divergent, outer corners robinsoni Bonhote Pahang, Malay of premaxillae; canine with posterior inflata Miller Fukien cusps descending below middle of tooth, sinicus G. M. Allen Szechwan much as Triaenops; absolutely no indica- formosanus Horikawa Formosa tion of p2; p4 large; interdental spaces hirsuta Miller Philippines somewhat reduced; m3 with much re- The following remarks on Coelops are duced W-pattern; outer lower incisor based upon examination of Coelops "robin- considerably thicker than the inner, the soni," U.S.N.M. 175000, Malay States. outer lobe unreduced (compare Triaenops). Coelops represents a type of bat very Anterior lower premolar considerably re- different from the Hipposiderine genera duced, and moved slightly outwards in previously described. Horseshoe, if it the toothrow. may be so called, with deep anterior median notch reaching back to the inter- nasal septum, each lateral portion of COELOPS BLYTH horseshoe parted into two greatly over- Coelops BLYTH, 1848, J. Asiatic Soc. Bengal, XVII, 1, p. 251. lapped lobes, the outer over the inner, the Chilophylla MILLER, 1911, Proc. U. S. Nat. inner extended farther forwards than the Mus., XXXVIII, p. 395. outer one. Lateral leaflets apparently ab- GENOTYPES.-Coelops, frithii Blyth from sent, or else obscured by a dense develop- "Soonderbuns" = Sandarbans (Dobson), Ben- ment of stiff hairs which margin the nasal gal; Chilophylla, hirsuta Miller from Mindoro. area. Transverse leaf rather low, pos- teriorly pocketed, with, however, a small The following further systematic work median eminence. "Sella" with slight on Coelops has been done: median raised ridge connecting internasal 1862, PETERS, Archiv f. Naturg., II, p. 117. area with middle of transverse leaf. No 1865, PETERS, Monatsber. Akad. Berlin, pp. 641-648, described C. bernsteini. visible frontal sac (Dobson says it is pres- 1871, PETERS, op. cit., pp. 330-332, reviewed the ent). Ear short, rounded, with antitragal genus, synonymizing his own species. lobe, devoid of strengthening "ribs" as in 6 AM1ERICAN MUSEUM1 NOVITATES [No. 1140

Hipposideros, etc. Tlhuml) with metacar- C. robinsoni with forearln 37 mnm. was pal extraordinarily lengthened, its length state(l by Bonhote to be the srnallest of to that of basal phalanx = 5.7: 1.2. Tail the genus. The forearm of U.S.N.M. extremely short, if not absent. Calcar 175000 is saidl by Taylor to measure only much shortened, its length to that of 33.5. The skull of this specimnen differs tibia = 3: 15 mm. sharply from all others seen not only in The skull is correspondingly peculiar: its small size but in the obsolescence of the brainicase short and roun(led; intertem- customary large pocket or recess inside the poral region narrow; rostral area with in- maxillary root of the zygoma, anterior to flations confined to maxillary portion, the the temporal fossa an(l directly above the nasal portion uninflated, flat; zygomata roots of m4 an(I the posterior half of M3. short, little expanded, without jugal emi- This pocket in U.S.N.M. 175000 (robin- nence; infraorbital foramen a pore, not en- soni ?) is reduced to a small pore. Further, closed by a bar, as in most Hipposideros; the upper ed(ge of the zygoma terminates tooth-bearing portion of maxilla greatly abruptly on meeting the rostrum level extended forwards, carrying the toothrow with the back of m2, while in other species with it so that the back of the premaxillae examined the zygomatic edge runs up- is level with the middle of M2; premaxillae wards and forwards onto the rostrum to correspondingly extended to allow the in- terminate almost level with the back of cisors to be in line with the fronts of the mi. proO(lont canines; back of palate slightly Again, the pterygoids are placed much behind m3; cochlea and tympanics not nearer the back of the palate than in any modified; but basicranial foramina enor- other form examined: in "robinsoni" the mously enlarged. Mandibles, with tooth- distance from back of palate to back of rows parallel, U-shaped at symphysis. hamulus of pterygoid = 1 mm., and in Upper incisors minute, convergent, the others the distance is 1.5 mm. or greater. outer lobe obsolescent; canine highly pro- C. inrflata was compared chiefly with o(lont, provided with two accessory cin- robinsoni. It and bernsteini of Java appear gulum cusps, each of which exten(ds more to be related; the forearm length in each than halfway (lown the body of the tooth; is 35.5. The toothrow of bernsteini, the posterior cusp, common to many Hip- measured against a millimeter rule photo- posiderinae and Megadermidae, and an graphed with the type skull, amounts to internal cusp; p2 small but still with cusp, 6.1 mm.; that of robinsoni, obtained in the retained in toothrow; p4 rather weaker same way, to 4.9 mm. Also it can be seen than in most Hipposiderinae; m3 with W- clearly that the zygoma (loes not merge pattern virtually complete; lower incisors with the rostrum as (lescribe(l above for not imbricated, retaining all three cusps; robinsoni. Our series from Bali is ap- lower canines small, vertical; P2 unreduced, parently near bernsteini: toothrow, 5.6- compressed, somewhat as in tricuspidata, 5.8 mm.; forearm, 38, and greatest wiclth moved slightly outwards in toothrow. of braincase, 7.6. In sinicuts the toothrow, The skull of this specimen was use(l in c-m3 attains 6.2 mm.; the forearm, 36; 1912 by Miller to describe the "cranial an(d width of braincase, 8.0. dlental characters of Chilophylla" and again The relationship of frithii from Bengal by Taylor' for the same purpose. and hirsuta from the Philippines to the Besides the skull of "robinsoni" just Indo-China and Javanese forms can be mentioned, I have before me for stu(ly the cleared up only when the type of the type skull of sinicus, two specimens from former has been re-examined and topo- Tonkin marked inflata (F.M. 32232-3), types of the latter secure(d and studied. and five from Bali (A.M.N.H. 107508-12); Formosanus is probably close to injiata or and in addition photographs of the skulls sinicits. of the types of bernsteini an(d robinsoni. It has long been customary to place C. bernsteini of Java in tile synonymy of C. 1 1934, Taylor, "Philippine Land AMammals," pp. 247-248. frithii of Bengal. But considering the RESULTS OF THE ARCHBOLD EXPEDITIONS. 36 7 new forms which have been distinguished in the association of grandis from Guinea in the Malay region, China and Philip- (Wrest Africa) and javanica. pines, I doubt whether the practice should Dobson made Iiis first division on char- be longer heldI to, at least until the sy- acters of the tragus and his subsidiary nonymy can be re-established by (lirect ones on the structure of incisors, ears and comparison of topotypes of each. There- lower premolars. His plan brought Jav- fore the name bernsteini is here revivedl anica closest to aethiopica and next to and applied to the Archboldl series from grandis. Bali, just east of Java. N. tragatats Andersen, 1912, (liffers from javanica chiefly by characters of the tragus NYCTERIDAE an(l greater size of the teeth. The fourth premolar is similarly unreduced. The NYCTERIS GEOFFROY East Indian species compared with most Nycteris GEOFFROY, 1803, Cat. Maim-ii. Mus. African are unspecialized with Nat. Hist., p. 64. Nycteris Petalia GRAY, 1838, Mag. Zool. Bot., II, p. scarcely reduced p4, an(l unenlargedl ears. 484. They are specialized, in that the outer Petalia GR.AYv, 1866, Proc. Zool. Soc. Lonldoni, lobes of the originally trifidl lower incisors p. 83. Nycterops GR.AY, 1866, Proc. Zool. Soc. Loni- are obsolete. don1, p. 83. Our collections contain both javanica GENOTYPES. -Nycteris, Vespertilio hispiduis an(l tragata, the latter represented by two Schirebei, Africa. Petalia (Pelatia), Nycteris specimens from Macarah Doewa, Palem- javanica Geoffroy, Java. Nyeterops, Nycterops pilosa Gr-ay, Africa. bang, Sumatra. The conspicuously larger teeth (c-m3, 8.9 and 8.8 mm.) and P4 When Gray set up Nycterops andl further even less reduce(d in proportion to P2, and characterized "Pelatia," he exemplified the slightly larger palate and rostrum dis- Nyeteris Geoffroy by N. thebaica instead of tinguish tragata from jqvanica, in addition N. hispidus. Allen' shows hispidus from to other features published by Andersen. Senegal an(l thebaica from Egypt as dis- We have photographs of the skull of N. tinct species. He lists a total of 22 full tragata. species from Africa. Only javanica and tragata are recordled from the Orient. So MEGADERMIDAE Nycteris is to be regardled as primarily MEGADERMA GEOFFROY African with a subgeneric offshoot (Pe- Mlegaderma GEOFFROY, 1810, AInn1. Mus. talia) in the East Indies. Hist. Nat. Paris, XV, p. 197. Although in some African species, for Spasma GRAY, 1860, Proc. Zool. Soc. Lonidonl, p. 83. example, arge, hispida, major, considerable GENOTYPES.-Megaderma, Vespertilio spasma further reduction of the posterior lower Lininaeus. Spasmna, Vespertilio spasma Linnaeus. premolar is attained, and in major it be- Megaderma and allie(d genera have been comes vestigial, neither of the Javanese reviewed repeatedly. Probably the clear- species shows this modification. est way to present the taxonomic work Nycteris was reviewed by 1'eters2 andl done on the genera is to trace it chronologi- by Dobson.3 Peters appended an excel- cally: lent plate to his paper, showing lower jaws 1838, GRAkY, Mag. Zool. Bot., II, pp. 490-491, and the condition of the lower teeth in proposed Lavia, with type M. frons eight species. Reduction and displace- Geoff roy. ment of the posterior lower molar are clearly 1847, HODGSON, S. Asiatic Soc. Benigal, XVI, p. 891, proposed Eucheira for Al. lyra seen in many of them. He classifie(d the Geoffroy. As shown later by Miller species by their ears, the cusps of the in- (1907) this is a horrmoniymii of Eucheira cisors, the state of P4. His system resulted Westwood, 1836, Tranis. Entom. Soc. LondoI, I, p. 44. 1 1939, Bull. Mus. Comp. Zodi., LXXXIII, pp. 1872, PETERS, Monatsber. Akad. Wiss. Berlin, 67-71.2 1870, Monatsber. Akad. Wissen. Berlin, pp. pp. 192-196, Ireviewed Megaderma, de- 900-914. scribed Ml. cor. and erected Lyroderma for 3 1878, Cat. Chir. Brit.'AIus., PP. 161-16;7. 11. lyra. 8 AMERICAN MUSEUM NOVITATES [No. 1140

1873, PETERS, op. cit., p. 488, proposed Cardio- medius Andersen Singapore derma for cor. majus Andersen Lower Chindwin 1878, DOBSON, Coll. Chirop. Brit. Mus., pp. minus Andersen Cambodia 154-160, treated all foregoing names as ceylonense Andersen Trincomalee, Ceylon subgenera of Megaderma. Lyroderma 1880, DOBSON, Proc. Zool. Soc. London, p. 461, lyra Geoffroy "probably E. coast of described M. gigas. Madras" 1906, MILLER, Proc. Biol. Soc. Wash., XIX, caurina Andersen Surat, India p. 84, proposed Macroderma for Megaderma and Wroughton gigas. sinense Andersen and Amoy, China 1907, ANDERSEN AND WROUGHTON, Ann. Mag. Wroughton Nat. Hist., (7) XIX, pp. 129-145, carnatica Elliot S. Martha Country reviewed the family Megadermidae, pro- schistacea Hodgson N. E. Bengal posing several new species and races. spectrum Wagner ? Kashmir They attempted to revive Eucheira and discussed the wing structure and general We have photographs of the type skulls relationships of the genera. of Megaderma natuna, majus, medium 1907, MILLER, Bull. U. S. Nat. Mus., Wash., minus, pangandarana and of Lyroderma LVII, pp. 101-106, recognized the genera Megaderma, Lyroderma, Macroderma, sinense and caurina. Lavia and Cardioderma. He showed Our material from Celebes appears in- Eucheira Hodgson to be a homonym. separable from our series from Cheribon, Since that time the generic situation has Java. In pangandarana Sody seems to remained static. have named a series of bats which average The named forms of the Oriental sub- very slightly larger than usual. And in genera Megaderma and Lyroderma are spite of the extraordinary differences shown in the accompanying list: Geoffroyl has shown in their nasal leaves, Megaderma it remains questionable whether trifolium spasma Linnaeus Ternate IslELand of Java is significantly different from trifolium Geoffroy Java spasma. Topotypes from Ternate, type carimatae Miller Karimata Iigland locality of spasma, are needed. Spasma natunae Andersen Natuna Island and Wroughton was founded upon a figure in Seba. philippinense Water- Philippines On the other hand, a specimen from town Burma, A.M.N.H. 54815, referable to horsfieldi Blyth "Continentsil India" majus, has decidedly larger molar teeth lasiae Lyon Lasia Islanc niasense Lyon Nias Island (c-m3, 10.3; ml-3, 5.9 mm.), compared siumatis Lyon Siumat Islaiad with 9.1 and 5.2 in Cheribon (Java) ma- pangandarana Sody S. Central JJava terial.

ON THE INTER-RELATIONSHIPS OF THE LEAF-NOSED BATS The affinity of Anthops, Aselliscus, Nycteris the posterior leaf is so much re- Rhinonycteris, Triaenops, Cloeotis, Asellia duced as to be rudimentary. A true horse- and Coelops to Hipposideros rather than shoe, with broadened, shield-like inter- to Rhinolophus is indicated chiefly by re- narial septum appears in Lyroderma and duction of their pedal phalanges to two Megaderma spasma, but in Lavia there is and by lack of p3. Further they are well no internarial development. The number differentiated from the megadermid and of phalanges of the toes is complete in all. nyeterid bats. None shows the special modifications of In the Megadermidae and Nyeteridae a the thumb to be seen in Coelops. tendency for the ears to become enlarged In its skull, Nycteris alone has developed and united is developed. Various types the extreme shortening of the palate com- of nasal appendages are seen. The tragus mon in Rhinolophus, but this condition is present, bifid. Megaderma, Lavia and cannot show relationship, since the pre- Lyroderma resemble Rhinolophus in pos- 1 1810, Ann. Mus. Hist. Nat. Paris, XV, pp. sessing high posterior noseleaves, but in 187-198. RESULTS OF THE ARCHBOLD EXPEDITIONS. 36 9 maxillae and rostrum are formed in wholly ing "septum." The noseleaf of Asellia is different ways. Two pairs of bifid upper somewhat similar. incisors and three of trifid lower incisors At this stage one of two courses is can be seen. In the mandible of N. re- followed: the three lobed or raised parts voilii p3, very minute, is found beneath of the leaf become higher and acutely the outside of the adjoining cingula of P2 pointed as with Cloeotis and Triaenops; and p4, but p3 is absent from N. javanica or else, the median one only of the three and arge. The angular process of the becomes enormously enlarged while the mandible is much reduced, a unique char- lateral ones degenerate, as in Rhinolophus acter among the bats being discussed. and the Megadermid bats. Contrary to Megaderma and Lyroderma should per- all of the foregoing, the transverse nose- haps not be separated more than sub- leaf of Coelops develops twin points in- generically. Tragus very large (as in stead of one or three. Lavia; in Nycteris quite small); tail ob- In conjunction with the transverse nose- solescent (also in Lavia, but well developed leaf special cells or pockets develop: the in Nycteris). Skull in many ways un- median, anterior pouch and the honeycomb specialized-e.g., palate, basis cranii, ros- of lateral pits found in Rhinonycteris and tral region-but premaxillae apparently Triaenops; the frontal sac in Hipposideros. never ossified and upper incisors entirely The ear develops along two wholly dif- wanting. Canine possessing the large, ferent lines of specialization. In the posterior cusp, so frequent among Hip- Megadermidae and in Hipposideros mega- posiderinae; p2 minute, inside c and p4; lotis the ears become enlarged and united W-pattern of m3 much reduced; lower in- across the forehead by a common fold of cisors (3-lobed) two; p3 absent, p2 unre- skin. But also in the Megadermidae the duced. large bifid tragus is retained, whereas in Lavia has developed expanded post the Hipposiderinae it becomes obsolete. and anteorbital processes and a low rostral Folds strengthening the pinna of the ear, region, similar to the depressed frontal which appear in all genera with large ears, area of Nycteris. But, like Megaderma, are non-existent in Coelops. In it the ears it has lost the premaxillae and upper in- are small, rounded and uniform in struc- cisors. Posterior canine cusp large; p2 ture. They are commonly described as obsolete; m3 much reduced; lower incisors "funnel-shaped." two; p obsolete. The angular processes The rostral region of the skull remains are unreduced in Lavia, Megaderma and broad, flat and relatively unspecialized in Lyroderma. Megaderma and in the Hipposideros armi- As was pointed out under Hipposideros ger group. In most of Hipposideros and the lines of development of the several allied genera and in Rhinolophus the ros- structures usually employed in classifica- trum is raised and variously inflated, the tion of the leaf-nosed bats seem to function swelling affecting usually both nasal and and change independently. Some of these maxillary portions. On the other hand, a structures are traced below, as they ap- marked dorsal depression or pit occupies pear in some genera of the Nyeteridae, the whole frontal area in Nycteris and Megadermidae and Rhinolophidae. Lavia. Coelops manifests a similarly de- The transverse noseleaf is seen in very pressed rostrum. simple form in Nycteris and in the bicolor The premaxillae are in course of reduc- and calcaratus groups of Hipposideros. tion by three methods. In Megaderma In Hipposideros it is primarily a flat, the premaxillary bones remain cartilagi- rounded structure with three "septa" sup- nous. In Nycteris the outer portions, each porting it in front. In the H. diadema, carrying two incisors, become ossified and armiger and pratti groups it becomes lobed more or less completely fused to the max- and enlarged. In Anthops the lobes are illae laterally. In all Rhinolophidae it is further developed, the apex of each lobe the center of the combined premaxillae being also the termination of a correspond- which ossifies (bearing one incisor only on 10 AMERICAN MUSEUM NOVITATES [No. 1140 each side), while the space laterally between premaxillae and premolars, nor the Mega- maxillae and the palatal branch of pre- dermidae because of their exaggerated maxillae tends to remain open. transverse noseleaves and total obsoles- Of the upper premolars p2 is rudimentary cence of premaxillary bones can show di- in all families except Nycteridae. It is ob- rect relationship to the hipposiderine solete in Nycteris, rudimentary and dis- genera. But they still exhibit a number placed inwards from the toothrow in of those character trends which were con- Hipposideros and Rhinolophus. In the stantly in operation during the evolution lower jaw of Rhinolophus p3, greatly re- of the superfamily taken as a whole. En- duced, is seen on the outer side of the largement of the rostrum,2 non-ossifica- ramus, displaced by the contiguous P2 and tion of the lateral connections of the pre- p4. The tooth is absent in all other genera. maxillae to the maxillae, the tendency for In most genera P2, although wholly dif- the ears to remain free, obsolescence of the ferent in shape from p4, due to its being tragus-these are characters of the Rhino- single-rooted instead of double-rooted, is lophidae which the Megadermidae and about as large as p4. It becomes slightly Nycteridae lack. reduced in Lavia, Rhinolophus, Triaenops, The single large genus Rhinolophus, Asellia and Hipposideros, and quite mark- commonly given subfamily rank, differs edly so in H. muscinus and in Cloeotis. from the remainder chiefly by non-coa- But in the exceptional case of Nycteris p4 lescence of the toes of the feet to make two is absent, while P2 takes its place as the joints instead of three (all Hipposiderinae); large functioning premolar, with p3 much by the exaggerated development of the or slightly reduced behind it (N. javanica posterior noseleaf into a tall blade and the and arge), or even vestigial and concealed development of corresponding modifica- beneath the cingulum of P2 (N. thebaica). tions of the "sella" region; and by the The W-pattern of m3 remains fairly pronounced shortening of the palatal complete in Cloeotis, Coelops, Triaenops, bridge. Further characters of Rhinolo- Rhinolophus and in Hipposideros muscinus. phus, which, however, are duplicated by It is more or less reduced in the majority other genera and by one or another of the of Hipposideros, and in Nycteris, Lavia, species groups of Hipposideros, are: linear Megaderma, Asellia. arrangement of the lower incisors (also in Obviously most modifications in the six H. muscinus and Coelops), wide separation sets of structures just discussed are of an- of the upper incisors (H. commersonii, cient origin and have operated with vary- Asellia, Aselliscus), large size and approxi- ing intensities upon the species of the three mation of cochlea (H. muscinus). Thus families just discussed, which Weber' some of the conspicuous features which unites as the superfamily Rhinolophoidea. distinguish certain Hipposiderine bats When character changes turn up indis- are "standard" characters in Rhinolophus. criminately through the three families Out of the large number of characters and their genera, those same characters usable for defining a species only a very are unlikely to prove universally valuable few are absolute characters of kind. The as indicators of relationships. For ex- preponderance represent merely differ- ample, because progressive reduction of ences of degree which frequently fluctuate the pattern of m3 shows in all groups, the among the species composing the genera condition of the tooth is significant chiefly whose phylogeny one seeks to trace out. when the W-pattern is unreduced. But Although the fluctuating characters ought characters marking a distinct, irretraceable to carry much weight in estimating phylo- departure from type, which are confined to genetic relationships it is the few absolute a small group of species, such as shows in ones which in practice are seized upon. the premolars of Nycteris, are of value. Because Coelops has but two joints to the Neither the Nycteridae because of their toes of its hind feet it is placed with the

2 In 1 1927, "Die Saugetiere," I, p. xiv. Coelops, incipient only. 19411 RESULTS OF THE ARCHBOLD EXPEDITIONS. 36 11

Hipposiderinae. But apart from its spe- to grant it subfamily rank, the Coelopsinae. cialized feet, it shows several quantitative Its unique characters are listed: characters and at least two qualitative- 1.-Structureless condition of the ear pinna. 2.-Strongly differentiated "horseshoe." the depressed frons and the highly indivi- 3.-Enormously elongated first metacarpal. dual type of noseleaves-which suggest it 4.-Absence of tail. may belong elsewhere and that the fusion 5.-Greatly extended canine-bearing portion of maxilla. of the tarsals indicates a case of conver- 6.-Much fenestrated ba8is ranii. gence. 7.-Internal as well as posterior supplemental canine cusps. Coelops is, in fact, so remote from other 8.-U-shaped symphysial portion of combined hipposiderine bats that it may be desirable mandibles.

LEAF-NOSED BATS, OTHER THAN RHINOLOPHUS AND HIPPOSIDEROS, IN THE ARCHBOLD COLLECTIONS NYCTERIDAE Nyeteris javanica Cheribon, Java 41 Nycteri8 javanica Noesa Penida, Bali 18 Nycteris tragata Ma carah Doewa, Sumatra 2 MEGADERMIDAE Megaderma spasma Cheribon, Java 6 Megaderma spasma Ta]liaboe Island, Moluccas 2 Megaderma 8pasma To,gian Island, Celebes 1 Megaderma spasma Waawo, S. Celebes 1 HIPPOSIDERINAE Asellicus tricuspidata Hollandia, Dutch New Guinea 44 COELOPINAE Coelops Oelboed, Bali 5 Species occurring in the New Guinea- natus, Rhinonycteris aurantia and Macro- Australian region but not obtained by the derma gigas. Archbold expeditions are: Anthops or-