RESULTS OF;THE ARCHBOLD EXPEDITIONS. NO. 35 A REVIEW OF THE GENUS WITH SPECIAL REFERENCE TO INDO-AUSTRALIAN SPECIES

- BY G. H. H. TATE

.; .-I

TtiE AMEMCAN mus HISTORY

I

BULLETIN OF.~~~~l THE AMERICAN MUSEUM OF NATURAL HISTORY. Vo. LXXVIIIj ART. V, pp. 353-393 New York Isued Augus8 19, 1941

I 'A'' &4'y4 4)74)4t4 '4> A44)/t''11A

(4'~~~~~~~~~~~~'41 44 ''4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I A4,)4(~~~~~~~~~~~~\ A7i "'

I''s 4> I

"'4 444

4WA YIs~~~~~~~ 'A9is,.'444 ~~' '' f's'/~zi>4 '~ 4> '' & 1 \ '4'41' ,~ I~~~~~~~~~~'1"44'½4 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~< /44> 47" '<44~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l'

'4 1' 9444 K,~~~~~~~~~~~A-C'1, 144

44 44,44, Article V.-RESULTS OF THE ARCHBOLD EXPEDITIONS. NO. 35

A REVIEW OF THE GENUS HIPPOSIDEROS WITH SPECIAL REFERENCE TO INDO-AUSTRALIAN SPECIES BY G. H. H. TATE

FIGURES 1 TO 5

PAGE INTRODUCTION.. 353 TAXONOMIC HISTORY...... 354 CRITERIA USED IN CLASSIFICATION ...... 355 Two NEW FORMS OF Hipposideros...... 357 LIST OF NAMED FORMS, UNDER GROUPS ...... 358 SUBGENERIC GROUPS OF Hipposideros ...... 360 ANNOTATED LIST OF NAMED FORMS OF Hipposideros ...... 380 Hipposideros MATERIAL IN THE ARCHBOLD COLLECTIONS ...... 391 Hipposideros OF THE NEW GUINEA REGION...... 392

INTRODUCTION Preparation of the reports on the mam- Coelops, and Anthops. Most of mals of the Archbold collections has shown those genera appear to represent specialized the need for a review of the relationships of types with many characters of Hippo- the of the genus Hipposideros Gray. sideros. In , Rhinonycteris and The material available for accomplishing Asellia the transverse noseleaf is sub- this undertaking, though ample, is not in- divided and the toothrows are subparallel; clusive enough to warrant a complete revi- in Cloeotis, Asellia and Rhinonycteris p2 iS sion of the genus. Besides the collections eliminated; in Anthops and Coelops the in the American Museum and the consider- length of the tail has been greatly reduced. able series assembled by Richard Archbold, But indications of the same kinds of speciali- the collections at Washington, Philadel- zation are observable also in groups still phia, Chicago, Pittsburgh and Cambridge currently placed within the genus Hippo- have been accessible through the generous sideros: The tail is much shortened in H. co-operation of their respective curators. commersonii; the upper premolarl is stated In addition, photographs of the skulls of to be absent in H. sabanus; nearly parallel the majority of the type specimens in toothrows occur in H. muscinus. The bats European museums and detailed notes on which students now separate generically the same, made in the summer of 1937, have from those still retained in Hipposideros been of great value to supplement the litera- seem mainly to be members of "groups" ture on many species not represented in the which began specializing fast and early. American collections. In spite of the large number of names, African species of Hipposideros have been approximately 126, which have been pro- studied only in order to learn their relation- posed for bats of this genus, close study of ships to Oriental species. No study has their structure has resulted in their tenta- been made of sub-relationships within tive classification into only eleven major African groups, such as caffer and its allies. groups totaling about 30 species, leaving a The Hipposiderinae are currently sepa- remnant of names of uncertain position, rated from the Rhinolophinae through re- due to inadequate description in the litera- duction by coalescence of the number of ture, inaccessibility of type specimens, or pedal phalanges from three to two. Ac- uncertain relationships morphologically. cording to that classification the subfamily 1 Miller, 1907, is followed for the terminology of Hipposiderinae includes, besides Hippo- the two upper and two lower premolars, namely, PA sideros, Triaenops, Rhinonycteris, Cloeotis, and p4. 353 354 Bulletin American Museum of Natural History [Vol. LXXVIII TAXONOMIC HISTORY Hipposideros has long been known to be fulva Gray divisible into a number of species-groups. murinus Gray atratus Kelaart Gray' in 1866 proposed a number of names fulva < cineraceus Blyth under brief diagnoses of possible subgeneric Chrysonycteris Gray: aruensis Gray value: examples, aurita Tomes Macronycteris: type, gigas bicolor I bicolor Temminck Gloionycteris: type, armiger antricola Peters Rhinophylla2: type, labuanensis amboinensis Peters4 Speorifera: type, vulgaris Chrysonycteria: type, fulva Sideroderma Peters: Phyllorhina: examples, nobilis and pygmaea type, fuliginosa Temminck Peters,3 five years later, grouped a num- ber of species and synonyms under Gray's Ptychorhina Peters: r caffer Sundevall names and himself set up six more sub- examples, caffra gracilis Peters genera, as follows: t bicornis Heuglin Cyclorhina Peters: Doryrhina Peters: type, obscura Peters type, cydops Temminck doriae Peters

( diadema Geoffroy Thyreorhina Peters: Phyllorhina Gray: nobilis Horsfield type, coronata Peters examples, griseus Meyen lankadiva Kelaart ? galerita Cantor Syndesmotis Peters: ? pygmaea Waterhouse type, megalotis Heuglin

Gloionycteris Gray: armiger Hodgson examples, swinhoei Peters Dobson' treated the subgeneric groups of Gray and Peters as synonyms of Macronycteris G commersonii Geoffroy Phyllo- -ray: rhina = examples, gigas Wagner Hipposideros. His work extended vittata Peters the revisional treatment by Peters. Ander- sen6 in 1905, 1906 and 1907, reviewed the f larvatus Horsfield diadema, armiger, commersonii and caffer larvata X deformisvulgaris HorsfieldHorsfield insignis Horsfield 4 New name for Temminck's bicolor from Amboina, thus restricting the type locality of bicolor to "Java crumeniferus Peron ? and Timor." Rhinophylla marsupialis Geoffroy? 5 Dobson, 1876, Monogr. Asiatic Chiroptera, Lon- Gray dukhunensis Sykes don, pp. 58-73; 1878, Cat. Chiropt. Brit. Mus., and speoris London, pp. 127-152. three apiculatus Gray 6 Anderson, 1905, Ann. Mag. Nat. Hist., (7) XVI, Speorifera pp. 497-507; op. cit., 1906, (7) XVII, pp. 35-48, leaflets pennicillatus Gray Ann. Gray: templetoni Kelaart 269-283; 1907, Mus. Civ. Storia Nat., Genova, (3) III, pp. 18-20. examples, aureus Kelaart taitiensis Zelebor labuanensis Tomes cervinus f cervinus Gould two leaflets I albanensis Gray longicauda Peters 1 Gray, 1866, Proc. Zool. Soc. London, p. 82. 2 Homonym of Rhinophylla Peters, 1865, Monats- ber. Akad. Wiss. Berlin, p. 520. 3 Peters, 1871, Monatsber. Akad. Wiss. Berlin, pp. 312-330. 19411] Tate, Results of the Archbold Expeditions. 35 355 groups. In 19181 a superficial excerpt from portant contributions in the shape of de- what was probably a detailed revision in scriptions of new forms and added descrip- manuscript was published under his name, tive matter relating to known species, none in which a number of new forms were has attem wo aimin synthe briefly sketched. That revision was never has attempted work aiming at synthetic printed. treatment of the genus as a unit or of more Though later writers have offered im- of its subgeneric divisions.

CRITERIA USED IN CLASSIFICATION The morphological evidence indicating height of the cusp from 2/3 or 3/4 (in bi- distinct evolutionary trends within the color and calcaratus) to 1/2 the height of p4, genus Hipposideros is perplexing and often while in cyclops, semoni and sabanus, in contradictory. Such characters as: pro- which p2 has become vestigial2 the height gressive simplification of the upper incisors of the cusp of P2 is less than 1/3 that of p4. or reduction of the anterior premolars, The toothrows become shorter in propor- widening of the rostral region or increase in tion to the palatal width. the number of leaflets lateral to the horse- The second character mentioned, modi- shoe, seldom combine to indicate clearly fication of the incisors, varies to only a recognizable evolutionary trends. In- limited degree. Typically no occlusion stead, those characters behave as inde- occurs between upper and lower incisors, pendent tendencies latent perhaps in all the mandible being strongly undershot. species, active in some species, quiescent in The single pair of upper incisors, their tips others. The student tries to discriminate convergent except in the commersonii between characters representing basic di- group, is formed of bilobate teeth usually vergence and those indicating only parallel- well separated, and with the outer lobe ism or convergence. obsolescent to a degree varying with the Retention of the anterior upper premolar species group. The two pairs of lower in the primitive dental alignment remains incisors are trilobate, the median teeth out- occasionally in several groups of species, wardly overlapping the lateral ones except particularly in the bicolor group (sensu in muscinus, by the extent of one lobe. stricto) and the calcaratus group. Even in The crown dimension of the outer tooth those groups the tendency for the teeth to varies from sub-equal (in muscinus, cyclops, be displaced laterally outward from the calcaratus, gigas, galeritus and bicolor toothrows can be discerned. In other groups) to twice the thickness and one and groups the anterior premolars usually be- a half times the width of the inner tooth (in come excluded and the posterior premolars speoris, pratti, armiger, commersonii, dia- and canines attain virtual or total contact. dema groups). In these latter groups the Concomitant with such changes in the teeth upper incisors retain most of the external of the upper jaw, the lower anterior pre- lobe, are stouter, and become nearly con- molars become modified. With movement tiguous (except commersonii); in the former of p2 out of the toothrow, P2 becomes pro- the outer lobe is obsolescent and the teeth gressively smaller, both by shortening of are weaker and spaced more widely. Pos- the cingulum and by reduction of the terior canine cusps are present in abae, commersonii, calcaratus and galeritus 1 Anderson, 1918, Ann. Mag. Nat. Hist., (9) II, pp. 379-383. 2 Thomas believed it absent in sabanus. 356 Bulletin American Museum of Natural History [Vol. LXXVIII groups. They are developed also in other the incisive foramina. About the center genera such as Triaenops and Coelops. of the anterior edge of each premaxilla is A tendency is evident for the W pattern the alveolus of the incisor tooth. But this of m3 (and the talonid of M3) to be incom- pattern is strikingly modified in certain plete; the posterior cusp and commissure groups: In some the posterior V is rounded fail to develop and the antero-posterior to a U. And the foraminal openings may length of the tooth is reduced. In the become narrowly slit-like as in cervinus or muscinus group m' remain virtually per- oval as in calcaratus, or the anterior enclos- fect. Elsewhere more or less reduction ing processes may fail to reach the maxil- appears. Partial reduction is seen in laries as in armiger and speoris; or again galeritus, caffer, bicolor, cyclops, larvatus, they may become enlarged, encircle the armiger and pratti. And marked reduction foramina, and unite with the posterior proc- shows in commersonii, diadema and cal- esses (pratti, cyclops, gigas), thus excluding caratus groups. the maxillary bones from margining any The rostral area exhibits fundamentally a part of the incisive foramina. In the double rounding or swelling, anterior to muscinus group a special condition is seen: the ante-orbital region, in combination the foramina are large, and oval; the pre- with a varying degree of lateral expansion; maxillae (paired) reach the front of the and, secondarily, in the armiger group, a palate by a narrow, spatulate process and marked flattening of the entire top of the the sides by two smaller lateral spatulate rostrum, coupled with a greater degree of processes. The position of the incisive ossification. In commersonii group the alveoli is at the middle of the anterior edge rostrum is very high; in pratti it is elon- of the premaxilla in most species. In gated. Minimal lateral expansion of the muscinus and gigas they are placed at the rostrum is seen in the calcaratus group; outer corners of the anterior edge. maximal expansion is reached in such The cochleae in the majority of groups species-groups as armiger, pratti, diadema, are small. They are wider than their cyclops, semoni, speoris. Intermediate con- distance apart in bicolor, muscinus (greatly ditions are seen in bicolor, cervinus, caffer, so), cyclops and gigas groups. In galeritus, sabanus. armiger and diadema groups they are sub- Normally the interorbital area shows a equal to their width apart, and in pratti decided constriction between the rostral they are distinctly smaller. expansion and the braincase. In the cal- Skin characters in most species have been caratus group and in coxi this condition is studied with care by many former students. less marked. Because of its scarcely ex- The structure of the noseleaves presents panded rostrum, the ratio great variation. The foremost character' to claim attention is the development of least interorbital width in calcaratus may two club-like processes one behind the width across rostrum other in the African species cyclops and the New Guinea-Australian muscinus groups, rise above 65 per cent. unique in Hipposideros, although homolo- The width across the squamosal roots of gous structures are seen in some other the zygomata exceeds the greatest width at Hipposiderine genera. the mastoid level in all groups except the The transverse leaf (posterior) is rela- bicolor group and part of calcaratus. In the tively simple and small in the bicolor and bicolor group the zygomata are peculiarly galeritus groups. It may have one or three narrow and the braincase is well inflated. vertical supporting ridges which divide it Important modifications appear in the into two or four shallow cells. In the delicate premaxillary bones. In the bicolor armiger and pratti groups it tends to be tri- group they are narrowly oblong, and, taken lobate, and to differ in size according to the together, form a wedge-shaped contact sex of the bearing it. A secondary with the palate. Their lateral edges are deeply notched to form, with the maxillae, 1 Except perhaps the united ears in megalotis. 1941 ] Tate, Results of the Archbold Expeditions. 35 357 transverse ridge is present in the caffer sub- f6rmigen Anhang" (Schneider) of speoris. group, rising behind from the base of the In the bicolor and calcaratus groups the primary leaf. Its crest is serrate. A same fold becomes a distinct pocket. In similar but probably not strictly homolo- cyclops, gigas and muscinus which have gous secondary transverse ridge was de- elongate, acutely pointed ears, it is ab- scribed by Thomas and Doria for wollastoni. sent. The number of leaflets lateral to the In speoris, galeritus, diadema, armiger and horseshoe is significant in certain groups: pratti groups the ears are large, broad, No lateral leaflets occur in the bicolor pointed, emarginate at outer edge near the group. In calcaratus none (to one, incipi- tip. The distal half of the pinna of all ent). In galeritus and pratti groups two.' except galeritus is naked. In sabanus the In speoris group three. In diadema three ear is similar, though proportionally or sometimes four. In armiger four. smaller. The galeritus group, with ears Among the African species caffer has two, also of similar general shape, can be further gigas and abae three. Cyclops and mus- recognized from the fact that body fur cinus have each two, the outer one extend- extends outward over more than three- ing back to the rear of the transverse leaf. quarters of the surface of the pinna. The frontal sac, which is usually de- H. (Syndesmotis) megalotis, in which the veloped, may be related to sex. Present in ears are united at the base, is still inade- males, it may be reduced or absent in fe- quately known. It may not be truly males. It is entirely absent in the diadema referable to Hipposideros. The united group and in lankadiva and allies and ap- condition of the ears suggests relationship proaches obsolescence in females of the to the Megadermidae. speoris group (including abae). Other A vestigial tragus can be seen in the case species whose females lack the frontal sac of muscinus. are papua, pygmaeus and coxi. Color has not been employed to any The ear in both the calcaratus and bicolor appreciable extent in this paper to aid in groups is quite large and obtuse to round- classification. Many of the species dealt pointed. Such bats exhibit what Tem- with appear in both red and gray phases. minck called "un pli interne," or internal In such cases color cannot aid the classifier. fold in the position of the absent anti- So far as is known bats without alternating tragus. This fold bears rarely a distinct color phases occur in the groups muscinus tubercle or papilla-the "kleinen kegel- and cyclops.

TWO NEW FORMS OF HIPPOSIDEROS Hipposideros bicolor macrobullatus, Skull with braincase slightly wider than in new subspecies other bicolor from Celebes, but chiefly note- worthy on account of the enlarged,bullae. TYPE.-No. 102367, Amer. Mus. Nat. Hist.; MEASUREMENTS.-Head and body, 52 mm.2; adult d; Talassa (Maros), South Celebes; 300 tail, 33; hind foot, 7; ear, 23. The ear, dry, meters; collector, G. Heinrich, Nov. 9, 1931. now measures 20 mm. from crown and 17 mm. Type, a skin with skull, in good condition. wide. Forearm, 42 mm.; tibia, 17; calcar, 8.5. GENERAL CHARACTERS.-A "bicolor" , dis- Skull: occipito-canine length, 17 mm.; zygo- tinguished by its proportionally larger ears and matic width, 8.2; mastoid width, 9.0; rostral horseshoe, wide internasal septum (1 mm.), very width, 4.8; least intertemporal. width, 2.4; large bullae, back of palate rounded and ex- width of cochlea, 2.5; width of bulla, 3.1; tended beyond the back of m3. length, c-mi3, 5.6; m'-3, 3.3. DESCRIPTION OF TYPE.-Skin brownish gray Unfortunately the type specimen is the above, the bases of the hairs whitish, as in all bi- only one in the collection. The color; underparts slightly paler; ears of "bicolor" type of type, quite large; tail about one and one-half as macrobullatus was compared with material long as tibia; thumb with the elongate meta- representing bicolor from Roeroekan, N. carpal and short basal phalanx of the group. Celebes, from Peleng Island, Halmahera, 1 Exceptions occur: see sabanus, dyacorumn. 2 Measured in the field by Heinrich. 358 Bulletin American Museum of Natural History [Vol. LXXVIII and another skull from Talassa. It di- hind, its posterior edge level with m3-3; cochleae verged in the same manner from all. moderate, larger than their distance apart. Canines with small cusp posteriorly, near the cingulum; p2 much reduced; W-pattern of m3 Hipposideros breviceps, new species incomplete; p2 very small, the length of its cin- TYPE.-No. 103335, Amer. Mus. Nat. Hist.; gulum to that of P4 as 0.5 mm.: 0.9 mm. adult d; North Pagi, Mentawi Islands, low- MEASUREMENTS.-Head and body, 42 mm.'; lands; collector, J. J. Menden, Jan. 31, 1935. tail, 23; ear, 15; hind foot, 7; forearm, 43. Type, skin with skull, in fair condition. Skull: occipito-canine length, 16.9; zygomatic GENERAL CHARACTERS.-Member of galeritus breadth, 9.5; mastoid width, 8.9; rostral width, group, externally very similar to schneideri but 5.2; least intertemporal width, 3.2; cochlea 2.5; smaller (forearm only 43 mm.); skull with much distance apart of cochlea, 1.9; c-m3, 5.8; cingu- shorter rostrum-resembling dyacorum, but lum length of p2, 0.3, of p4, 1.1, of P2, 0.45, of P4, horseshoe with 2 lateral leaflets instead of none. 0.9. DESCRIPTION OF TYPE.-Pelage with tips very This new species is represented in our col- dark, almost fuscous, the bases grayish white; lections by a series of 37 specimens, all underparts pale brownish gray, bases even paler. Membranes and ears blackish. Horseshoe paratypes. Breviceps appears to represent small, provided with two lateral leaflets, the a half-way stage in the line of evolution outer much the smaller. from an insolens-like form to the specialized Skull with short, broad outline much resem- dyacorum. The skull and teeth, although bling that of dyacorum of Borneo. Rostral swellings distinct, individually inflated; pre- not so small, have already attained the maxillae not exceeding fronts of canines; jugal condition of dyacorum; the lateral leaflets, eminence rising abruptly; palate rounded be- on the other hand, are unreduced.

LIST OF NAMED FORMS List of the named forms of Hipposideros saevus Andersen Kei Island, New and allies Guinea (tricuspidatus omitted), arranged aruensis Gray Aru Island, New geographically under groups. Guinea albanensis Gray Cape York, N. E. Australia TYPE REGION, OR TYPE NAMED FORMS LOCALITY calcaratus group Papuan (also Philip- pine ?) bicolor group Indo-Australian region calcaratus Dobson Duke-of-York Island micropus Peters Simla, India cupidu8 Andersen Papua cineraceus Blyth Punjab, India ? coronatus Peters Mindanao, Philip- murinus Gray India pines fulgens Elliot Mahratta, India auritus Tomes India commersonii group2 African and Mascarene fulvus Gray Madras, India marungensis Noack West Tanganyika pallidus Andersen Kathiawar, India synonym mostellum Kenya pomona Andersen Coorg, India Thomas atratus Kelaart Ceylon thomensi8 Bocage St. Thome Island nicobarulae Miller Nicobar Island gigas Wagner Angola gentilis Andersen Burma syn. vittata Peters Ibo Island, Mozam- sinensis Andersen Fukien, China bique nequam Andersen Selangor, Malaya gambiensis Andersen Gambia atrox Andersen Selangor, Malaya niangarae J. A. Allen Belgian Congo major Andersen Engano Island, Su- commersonii Geoffroy Madagascar matra ridleyi Robinson and Singapore, Malaya pratti group Indo-Chinese Kloss lylei Thomas Northern Siam bicolor Temminck N. W. Java pratti Thomas Szechwan, China javanwus Sody Java doriae Peters Sarawak, Borneo galeritu8 group Indo-Australian; antricola Peters Luzon, Philippines Oceanic wrighti Taylor Luzon, Philippines brachyotus Dobson Central India; Cey- erigens Lawrence Mindoro, Philippines lon (Phillips) toala Shamel Celebes aureus Kelaart Ceylon macrobullatus Tate S. Celebes 1 Field measurements by Menden. amboinensis Peters Amboina 2 Interrelationships not studied. 19411] Tate, Results of the Archbold Expeditions. 35 359

galeritus Cantor Penang, Malaya nobilis Horsfield Java labuanensis Tomes Labuan, Borneo speculator Andersen S. Celebes insolens Lyon S. E. Borneo pelingensis Shamel Peleng Island, Cele- celebensis Sody S. Celebes bes crumeniferus Peron Timor Island diadema Geoffroy Timor Island batchianus Matschie Batjan Island euotis Andersen Batchian Island cervinus Gould N. Queensland; its custos Andersen Kei Island range extending to pullatus Andersen Papua Santa Cruz Islands reginae Troughton N. Queensland trobrius Troughton Trobriand Islands, galeritus group deriva- Sunda to Philippines New Guinea tives mirandus Thomas Admiralty Islands sabanus Thomas N. Borneo dinops Andersen Rubiana, Solomon schneideri Thomas Sumatra Islands dyacorum Thomas Borneo oceanitis Andersen Guadalcanar, Solo- breviceps Tate N. Pagi Island, Su- mon Islands matra demissus Andersen San Christobal, Solo- longicauda Peters Java mon Islands pygmaeus Waterhouse Philippines Mainland offshoot of Indian; reported to galeritus group deriva- Wholly African diadema group reach Burma tives' lankadiva Kelaart Ceylon caffer Sundevall Natal indus Andersen Kanara, India gracilis Peters Mozambique mixtus Andersen E. Mysore, India bicornis Heuglin Eritrea unitus Andersen Saugore, India angolensis Sebra Angola schistaceus Andersen Bellary, India centralis Andersen Uganda guineensis Andersen Gaboon speoris group Chinese-Burmese-Sunda niapu J. A. Allen Belgian Congo larvatus Horsfield Java tephrus Cabrera Morocco insignis Horsfield Java beatus Andersen Cameroons deformis Horsfield Java fulginosus Temminck "Coast of Guinea" leptophylla Dobson Khasia Hills, Assam nanus J. A. Allen Belgian Congo grandis G. M. AllerI Chindwin, Burma ruber Noack Tanganyika neglectus Sody C. Borneo curtus G. M. Allen Cameroons vulgaris Horsfield Java barbensis Miller St. Barbe Island Derived from base of poutensis J. A. Alle:n Hainan Island,China speoris - galeritus stem. (?) Not closely related to speoris Schneider Peninsular India, Cey- each other lon and Africa coxi Shelford Sarawak, Borneo speoris Schneider Tranquebar, S. India obscurus Peters Luzon, Philippines dukhunensis Sykes Dekkan, India papua Thomas and Misori Island, Dutch templetoni Kelaart Ceylon Doria New Guinea aureus Kelaart Ceylon armiger group Indo-Chinese taitiensis Fitzinger Tahiti ? armiger Hodgson Nepal, India apiculatus Gray Madras, India debilis Andersen Malay Peninsula penicillatus Gray swinhoei Peters Amoy, China abae J. A. Allen Congo terasensis Kishida Formosa turpis Bangs Liu Kiu Islands muscinus group Torresian pendleburyi Chasen Peninsular Siam muscinus Thomas Fly River, Papua semoni Matchie Cooktown, N. Aus- diadema group Primarily Insular, to tralia Australia and Solo- stenotis Thomas' Mary R., N. Aus- mon Islands, but tralia also Burmese- wollastoni Thomas S. W. Dutch New Chinese Guinea nicobarensis Dobson Nicobar Island masoni Dobson Moulmein, Burma cyclops group2 Central African enganus Andersen Engano Island, Su- cyclops Temminck "Coast of Guinea" I matra synonym micaceus Gaboon vicarius Andersen Sarawak, Borneo de Winton natunensis Chasen Natuna Island synonym langi J. Belgian Congo griseus Meyen Luzon, Philippines A. Allen anderseni Taylor Philippines megaloti8 group2 East African megalotis Heuglin Eritrea 1 List of named forms. Interrelationships not studied. 2 Not studied. 360 Bulletin American Museum of Natural History [Vol. LXXV+III SUBGENERIC GROUPS OF HIPPOSIDEROS The eleven major species-groups which 4.-Transverse leaf greatly enlarged in males, in the genus are listed lateral leaflets 2, skull with anteriorly pro- can be recognized longed rostrum. Size large. . pratti group. synoptically: Transverse leaf not specially enlarged in 1.-bicolor group India to North Aus- males, rostrum otherwise...... 5. tralia 5.-Two lateral leaflets or less, size medium to 2.-calcaratus group New Guinea and Solo- small...... galeritus group. mon Islands Three or four lateral leaflets, size medium to 3.-commersonii group African and Madagas- large...... 6. car 6.-Frontal sac absent in both sexes; rostrum 4.-pratti group Southeastern Asia rounded, with paired inflations variously 5.-galeritus group India to Australia, developed. Size large.... diadema group. New Hebrides, Africa Frontal sac present in both sexes, or absent 6.-armiger group India to Formosa, Liu in 9S; rostrum rounded, size medium....

Kiu Islands ...... speoris group. 7.-diadema group Burma and Siam to 7.-Tail very short. Skull with toothrows an- Australia and Solo- teriorly convergent. M3 with W-pattern mon Islands reduced. Size large ..... cyclops group. 8.-speoris group India to China, Bor- Tail unshortened. Skull with toothrows al- neo, Java; and Cen- most parallel. M3 with W-pattern com- tral Africa (abae) plete. Size small ...... muscinus group. 9.-muscinus group New Guinea, North 8.-Ears united by frontal band. megalotis group. Australia 10.-cyclops group Africa 11.--megalotis group East Africa Hipposideros bicolor group Three of these groups, bicolor, galeritus Ears large, the antitragal lobe provided and diadema, are widespread and abundant with a distinct "internal fold" (Temminck), through the Indo-Australian region. Two, the outer edge of the ear not emarginate. calcaratus and muscinus, are Papuan and Horseshoe moderate (width about 15 per Australian; two, pratti, armiger, are conti- cent of forearm); transverse leaf thin, sup- nental (S. E. Asia); and three, commersonii, ported by three weak ribs which divide it cyclops, megalotis, are African. into four "cells"; no lateral leaflets. Fron- Muscinus and cyclops come from a com- tal sac present in both sexes. (Phillips mon but remote origin. Megalotis, known states it is obsolescent in atratus.) Joints only from the type, is perhaps not strictly of thumb of the southeastern forms pro- referable to Hipposideros. Triaenops portioned as in calcaratus group (in gentilis (tricuspidata, etc.) have been omitted. A and allies basal phalanx is longer). Length simple, partly artificial key to the main of tibia 40 per cent of forearm; calcar 20 groups is offered. per cent; tail 70 per cent of forearm; part 1.-Without two longitudinally placed club- of terminal joint often exceeding patagium. shaped processes on sella and transverse Skulls of species of this group are recog- 2. process ...... nizable by the enlarged braincase and nar- With two such processes ...... 7. rowed zygomata, combined with their 2.-Ears not united by frontal band ...... 3, Ears united by frontal band. 8. elongate, tapered outlines. The only other 3.-Ears large, the outer edge not emarginate, bats at all resembling them-bats of the skull with mastoid width exceeding zygo- calcaratus group-have similarly tapered matic width, size small ..... bicolor group. Ears large, sub-triangular, outer edge not skulls, but widened inter-orbital area, less emarginate, skull with zygomatic width broadened braincase, smaller cochlea. exceeding mastoid, intertemporal con- They are considerably larger. striction only slightly less than rostral Upper incisor with outer lobe obsoles- width. Size medium .... calcaratus group. Ears narrow, elongate, acute. Skull with cent; canine without posterior cusp; p2 rostrum almost as high as braincase, upper still usually included in toothrow; outer incisors placed at outer corners of pre- lower incisor unenlarged but overlapped by maxillae. Tail much shortened. Size il; crown of P2 1/2 to 3/4 of height of large ...... commersonii group. Ears medium to large, triangular, the outer crown p4 (reduced in atratus, nequam); the 4 m3 edge emarginate ...... W-pattern of reduced. 19411 Tate, Results of the Archbold Expeditions. 35 361

H. bicolor was described by Temminck as Lower p2 three-fourths of height of crown of coming from "Java, Amboina and Timor." p4. Significant measurements: Forearm, Temminck continued, "specimens from 47; zygomatic breadth, 9.1; greatest Amboina (subsequently named by Peters mastoid breadth, 9.4; least intero-rbital amboinensis) have dimensions a little less width, 3.0; c-m3, 6.5; ml-3,3.9. than those from Java." He had before From Temminck's description of the him "ten females and four males." Of the coloring "roux marron" tips and white length of the forearm he wrote, "anti- bases, the Javanese specimen "d" must brachium 1 inch 3 lines [32 mm. ].... The have represented the reddish-brown phase adult from Java ... antibrachium 1 inch 8 of bicolor, and the 47-mm. forearm indi- lines [42 mm.]." cates it was one of the larger species of its Fifty years later Jentinck listed': Skulls group. "a" and "b" adults. Amboina. Types of Further study of bats of the bicolor group Rhinolophus bicolor Temminck; and speci- from Java and surrounding territories sug- mens "a"i-"s," most of which bear dates gests that two, and possibly three, distinct showing they were collected later than species may be present on the island. Be- Temminck's type series. Of them, only sides the genotype (specimen "d" of Jen- "a," "b,"Y "ic," "ld,"Y "if,"~it, "2flY"1, cp," tinck from Anjer), with which javanicus is "q" could have been examined by Tem- very probably synonymous, I have ex- minck. Possibly Dobson's (1878, p. 151) amined M.C.Z. 12798, a, from Soeka- "specimen e," from Leyden Museum, repre- boemi, West Java, which has the forearm sents another, obtained by exchange. Of only 39 mm. and c-m3, 5.6 mm. This bat the foregoing, skins "a," "b" and "c" (from is structurally indistinguishable from our Sumatra) can be eliminated. Skins "d" series from Bali, the island immediately and alcoholics "p" and "q" as well as Dob- adjoining Java on the east, except that its son's specimens are from Java, and "f" and infra-orbital foramen is large and slit-like. "g" (from Amboina) were not found. No It appears to be very close to nicobarulae skulls had been extracted from others of the and antricola (of Lawrence), both of which series. Dr. Junge in 1937 very kindly had have the same forearm and toothrow meas- the skull of the Javanese "specimen d" urements. extracted and cleaned, so that I might The series from Bali -has forearm 39-41 study and photograph it. mm., c-im3, 5.3. With one ex6eption these The type locality (and consequently the bats have the pelage a lucent, golden color geographical race) of H. bicolor seems yet quite unlike the majority of species, which to be unrestricted. I now propose to re- have brown or fuscous-tipped hairs with strict the type locality of "c6te d'Anjer" whitish bases. in the extreme northwest of Java where Van This leaves in the Indo-Malaysian region Hasselt collected specimen "d," and to the fulvus-gentilis bats, to which perhaps make specimen "d" lectotype of Tem- erigens is related, cineraceus with forearm minck's type series (which included speci- only 34 and c-m3 but 5.1 mm., and atratus mens of amboinensis). From my notes I with forearm 36 and c-m3, 5.4. quote: H. gentilis sinensis has the forearm 41 and Specimen "d" (lectotype): sex ?; ear the toothrow 6.1 mm. It is approached broad, squarish in outline; dorsal color now quite closely by erigens of the Philippines light brown with whitish bases, underparts but has the subterminal joint of the thumb with chest and throat now white, posteri- decidedly longer (2.6:2.1). I suggest, too, orly brownish buff with paler bases. Skull that true gentilis may be very nearly related broken in occipital region, elongate; nasal to fulvus of peninsular India. eminence slight; rather pronounced pos- There remain to be considered the repre- terior interparietal swelling. Anterior pre- sentatives of the group in Celebes, New molar only slightly out of line with p4 and c. Guinea and adjoining areas. Toala Sha- whose 1 Jentinck, 1887, Cat. Osteologique, pp. 272-273; mel, type has been recently exam- 1888, Cat. Syst., pp. 168-169. ined, has the forearm 39 mm., c-im3, 5.5. 362 Bulletin American Museum of Natural History [Vol. LXXVIII

It appears thus to be related to our speci- lateral leaflets. Frontal sac present in both mens from Bali and M.C.Z. 12798 from sexes. Thumb strongly developed, the Java. A.M.N.H. 109104 from Peleng metacarpal equaling or slightly exceeding Island is not to be distinguished from Cele- the basal phalanx. Length of tibia 50 per bes material. A.M.N.H. 102364 from cent of forearm. Calcar elongate (30 per Halmahera is likewise identical to the cent of forearm). Tail 70 to 75 per cent of specimen from Java. Macrobullatus is forearm, its terminal joint not exceeding easily distinguished through its large bul- uropatagium. lae. Skull unspecialized; inter-orbital region Papuan material, of which we have am- unconstricted (attaining 70 per cent of ple, is a little smaller: forearm 37 mm., c- width of rostrum); rostrum scarcely ex- m3, 5.4. Specimens from the Fly River panded, skull tapering anteriorly. Pre- have the slit-like infra-orbital foramen men- maxillae (from below) elongate (3.2 mm.), tioned above, elongated to 1.6 mm.; those narrow (1.0 mm.), the eliptical openings of from Idenburg River less so. Both show the incisive foramina closed extero-pos- greater deepening of the posterior third of teriorly by the maxillae. Palate in front the zygoma than appears in material from with U-shaped opening for reception of any of the more westerly localities. Such maxillae, behind not extending beyond m3. posterior processes on the zygomata are Vomer projecting far back into mesoptery- present normally in most groups of Hippo- goid fossa, thickened at its lower margin. sideros. They are highly developed in Cochleae about as wide as their distance Asellia and other genera, and therefore apart. Mandible with strong coronoid characteristic of the family. They tend process and a heavy knob-like angular proc- toward obsolescence in all members of the ess. Upper incisors simplified, their bicolor group except those from New crowns with only trace of outer lobes, their Guinea. tips inclining inward but separated by the There is good reason to believe that thickness of one incisor. Canine proodont, micropus belongs with cineraceus, but the with high posterior cusp. Second pre- affinities of doriae must yet be proved. molar, though very much reduced, still re- H. nequam Andersen has not been allo- tained in toothrow. Main cusp of p4 higher cated in the group. Its greatly reduced P2 than molar cusp line. W-pattern of third separates it from true bicolor, with which molar reduced in conformity with the the size of its forearm would otherwise characters by which Miller separated Hip- place it. I have failed to observe more posideros and Rhinolophus. Outer lower than a moderate degree of reduction in P2 Of incisor slightly enlarged. Crown of P2 atratus (specimens in M.C.Z. obtained about 3/4 height of P4. from W. W. A. Phillips), which was so So far as known the calcaratus group is strongly emphasized by Andersen. In restricted to the New Guinea region (see M.C.Z. 27500 the cingulum length of p2: map, figure 5). Coronatus Peters from p, = 0.6:0.75. In a form in which P2 is Mindanao (type of Thyreorhina, subgenus) "comparatively large" (Andersen, 1918), had characters which agreed substantially the cingulum length of P2:P4 = 0.8:0.8. with those of calcaratus. Nicobarulae, atratus and antricola have Our calcaratus material from the island skulls which are virtually indistinguishable, of New Guinea and outliers, falls into two and they are almost identical in size. readily separable groups: Hipposideros calcaratus group 1.-Smaller bats with forearm 47-50 mm.; pos- terior canine cusp descending only 1/3 of Ear broad, triangular, obtuse, its outer length of tooth; c-m3, 7.0-7.5; openings, margin not concave. Horseshoe area sim- separating the alisphenoid areas from the ple: horseshoe medium in size (width posterior pterygoid area, relatively large (1 12 cent of mm. wide); the sphenoid "bridge' nar- about per forearm); posterior row (1.3 mm.) and the sphenoidal depres- leaf simple, unwidened, with one supporting sion weakly developed. ridge (two lateral ones absent or weak); no 2.-Larger bats with forearm 51-55 mm.; pos- . CD ri

.cvt b0 t

X

o t-

4 t 0a 364 Bulletin American Museum of Natural History [Vol. LXXVIII

.terior canine cusp descending nearly half- Hipposideros pratti group way down tooth, the tooth proportionately more massive; c-m3, 8.0 to 8.6; openings The forms pratti and lylei, though they adjoining pterygoid "bridge" small (0.7 resemble the armiger group superficially in mm.), partly concealed by "bridge," which color and size, have developed independ- is wide (1.8); sphenoidal depression rela- tively well defined. ently from well down the Hipposideros There is no character in Dobson's de- stem. scription of calcaratus to indicate to which They are characterized by the enormous of these groups it is related. But Ander- development of their lobate transverse sen's brief comparison made when he pro- noseleaves2 in males, by a supplementary posed cupidus would appear to settle the transverse structure, probably the homo- the logue of the sella in Rhinolophus, and by the matter-cupidus the smaller, calcaratus anterior edge of the larger. notching of the median Both species appear to be widely dis- horseshoe (as in larvatus). No supporting tributed. Probably they have been con- septa appear on the face of the transverse in literature. But leaf, which is quite thick. The horseshoe fused with each other is margined by only two lateral leaflets (as when the skulls are compared they are sac present. Ear easily distinguishable. The skins, apart in cervinus). Frontal from their decided differences in size broad, pointed, slightly emarginate. absolute sizes of Skull with rostral area low and markedly (which applies to the feet, widened, but differing from condition in phalanges and claws of thumbs, etc.), ap- by possession of well-de- pear not to differ. armiger group of calcaratus, veloped frontal depression and by greater The bright rufous phase forward extension of the canine-bearing represented by four bats from the upper maxillae. Premaxillae Fly River, is noteworthy. portions of the from Jobi is a wholly enclosing incisive foramina; roofs of U.S.N.M. 18480 specimen narial and mesopterygoid canals showing of H. cupidus. step-like discontinuity, described in armiger The commersonii group group. (Macronycteris, subgenus) Upper dentition much as in armiger: Large species of African and Madagas- outer lobe of i weak but distinct; outer car, first distinguished by the "longi- lower incisor half as thick again as inner. tudinal"' opening of the frontal sac (pres- Lower anterior p with crown height 1/2 ent in both sexes). Horseshoe entire in that of P4. Molars less heavy. Upper front, with three to four lateral leaflets; third molar with W-pattern reduced. See transverse leaf with median septum. Ear distribution map, figure 4. acute, emarginate, not broadened. Thumb with basal phalanx equaling or exceeding Hipposideros galeritus group the metacarpal. Tail shorter than tibia. Skull with moderately inflated and The range of this group of bats (s.s.) is broadened but high rostrum-almost as rather more restricted than that of the bi- high as braincase minus sagittal crest. color group. It extends from India (brach- Sagittal and lambdoidal crests strongly de- yotus) through the Malay Peninsula, Sunda veloped. Premaxillae enclosing incisive Islands, Borneo and Celebes to New foramina. Angular process of mandible Guinea, northern Australia and New but little flexed outward; coronoid high. Hebrides. In typical form it has not yet Symphysis strong and deep. been recorded from Indo-China and China. Upper incisors weakly bilobed, divergent. Pygmaeus represents it in the Philippine Canine with high though weak posterior Islands. Inclusion of the caffer subgroup cusp. Anterior premolar excluded, much extends the range to most of Africa. reduced. Outer lower incisors but slightly All typical species possess two lateral enlarged. Anterior lower premolar re- leaflets. Many-perhaps all-can further duced-its crown only 1/3 the height of p4. 2 See illustration in Osgood, 1932, Field Mus. Nat. 1 Nearly circular. Hist., Zool., XVIII, p. 223. No t s S o .tct >_eD t_ tX

tt C>1 >

eD . * *;> > rt O> <*s

> W te C) ..s t Ct ^ R .%X .t R .t A..4 Xt >0t X .20 Q Ct C) > s k X t X m3 _ t-s O Ct t H COo ° Ct t n X tD Cs st Ct O CO GO * CSD a4t *t A _ Ct *CZ > t . .L *ss cs sa _ 5 >-Cag X >.QXt tz

GD_ * ^ . CD . C,a, *s X t > Ctt t . 9

>_ t _ > .. > 00

s .. e Q CD s X . _ a Ct Ce Ct > . _

O t t

Ct t X = - Q

Ct

a Ct-CQ*C) 366 Bulletin American Museum of Natural History [Vol. LXXVIII be recognized by absence of pigmentation pear partly intermediate between the cer- of the narial portions of the horseshoe, in vinus and bicolor groups. Horseshoe small, galeritus and cervinus extending to almost its front edge notched, without lateral leaf- the whole horseshoe. lets; transverse leaf without vertical septa; Transverse leaf four-celled, simple: the frontal sac present in both sexes. Skull "sella" portion of the nasal region full and rather short. Zygomatic width slightly cushion-like; horseshoe proper simple, exceeding mastoid width. Rostral swelling with its anterior margin unnotched (except moderate, 57 per cent of width of brain- sabanus). case. Incisors weak, outer lobe obsoles- Ears not enlarged, broad, pointed, with cent. Canine without cusps, less pro6dont body fur on three-fourths of their outer sur- than in cervinus or bicolor groups. Lower face, somewhat emarginate along the outer anterior premolar less than half the height margin just below the tip. Tail moderate of crown of p4. (to very long, if longicauda can be referred H. sabanus from North Borneo appears here). to be allied to H. dyacorum from Sarawak. Skulls shorter and broader than those of Dyacorum has the frontal gland present in either the bicolor group or the calcaratus both sexes, lacks lateral leaflets, but has group. Rostral area moderately full, ears substantially like those of cervinus. rounded, with incipient post-orbital emi- The forearm is exceptionally strongly nences, its width about 50 per cent of zygo- arched. The skull is shortened as in matic breadth, which last is greater than sabanus, but p2 though extremely minute greatest mastoid width. Least inter-tempo- is still present. In the lower jaw P2 iS ral width about 50 per cent of rostralwidth. reduced in size as in sabanus. H. schneideri Palate broad and short. Incisive foramina of Sumatra appears also to be specialized by in premaxillae large, broadly oval, their considerable reduction of p2. Breviceps extero-posterior margins formed by the resembles dyacorum in skull and dentition maxillae. but retains the two pairs of lateral leaflets. Incisors with only a trace of the external Hipposideros caffer and allies (Ptycho- lobes. Canine width variously developed derma, subgenus) apparently constitute the posterior cusp. Anterior upper premolar African offshoot from the stem leading to partially excluded. Hypocone better de- galeritus group. Frontal sac usually pres- veloped in ml than in M2. Lower outer ent in both sexes.' Horseshoe entire in incisor scarcely or not at all thicker than front, provided with two lateral leaflets, as inner one. Crown of P2 typically 2/3 of in galeritus group. Transverse leaf with- height of crown of P4. out the three septa; at most two-celled; Here may be included as characteristic specialized by development of a serrated cervinus, batchianus, celebensis and insolens. secondary transverse leaf from its posterior But a more specialized series is found in the face. Malay area, and the Sunda Islands, in Skull much as in galeritus, slightly lower. which reduction of p' has been carried a Outer lobe of incisor similarly much re- stage farther, so that the crown of P2 re- duced; p2 only partly displaced. Slight mains only from 1/2 to 1/3 the height of the enlargement of outer lower incisor discerni- crown of P4. Such are schneideri and proba- ble; height of crown of P2, 1/3 to 2/3 that bly galeritus. Whether brachyotus (In- of P4- dia) and aureus (Ceylon) show this greater Allen2 has listed six geographical races dental specialization I do not yet know. In under caffer. It seems, too, that fuli- the Philippines area a specialized derivation ginosus (Sideroderma, subgenus) should be of the cervinus group is found: pygmaeus at least associated with the group. H. has become greatly reduced in size and its frontal sac is obsolete (? perhaps only in ' In females I have examined the frontal sac, females). though small, is present. In fuliginosus it is said to sabanus and others, appar- be absent. Certain bats, 2 Allen, G. M., 1939, Bull. Mus. Comp. Zool., ently derived from the galeritus group ap- XXXVIII, pp. 79-81. 19411 Tate, Results of the Archbold Expeditions. 35 367

9.0 mm. nanus is closely allied to caffer (I have ex- in the picture) are, respectively, amined the type); curtus and ruber also. and 8.2 mm. The rostral width is 4.9 and c-mr3, 5.6. The skull may therefore belong major, atrox and ridleyi of the bicolor of H. galeritus with Note on the type group. Dobson (1878, p. 142) wrote of galeritus This state of affairs, if true, might upset "the nasal membranes in the type speci- the whole classification of the group which in speci- status of the type of men . . . appear much larger than depends upon the mens from other localities." (With galeri- galeritus. To prevent such an occurrence tus he had synonymized labuanensis, longi- the body of the type, not the skull bearing a cauda and brachyotis.) His remarks on the similar number, must be regarded as Can- spacing out of c, p2 and p4, if observed from tor's type specimen. The forearm of the the type specimen, would be significant. type measures 45.5 mm. (1937). But one cannot be sure of this, as he had It is customary for authors to assume before him specimens from Singapore, that galeritus extends through Sumatra, Sarawak, Ceylon and Dekkan. The only Java, Borneo to Celebes. Without evi- forearm measurement given by him was dence to the contrary this assumption must 1.75 inches (44.4 mm.). He showed ex- continue, but then for the sake of consis- actly the same length of forearm in his next tency galeritus must be extended to include species-cervinus-from Australia and New the almost similar cervinus of New Guinea Guinea. and Australia.' Provisionally it is sug- The type of galeritus Cantor raises a diffi- gested that all members of the group having culty not often met with. There is some two lateral leaflets, the frontal sac present question in my mind whether the skin and in both sexes, and the anterior upper and skull are not mismatched. The specimen lower premolars not extremely reduced, be (in alcohol ?) was the true basis of Cantor's placed within a single species galeritus.2 description which follows, in abbreviated We shall then have, as more or less valid form: subspecific forms, brachyotus, labuanensis, simple cervinus. ". . . flesh-colored nasal appendage insolens, celebensis, batchianus, but large ... horseshoe ... covers the short H. sabanus, schneideri, longicauzla and two leaves on either respective specializa- ... muzzle, which has dyacorum, with their

. . pyg- side ... the ears . .. broader than long tions, appear to be derived species; pyriform, narrowing toward the apex ... maeus seems to be a derivative of more re- more than 2/3 of the back of the ear is mote origin; and papua, incertae sedis, but

covered with fur . . differs [from speoris, probably from the source stem of speoris which has the frontal sac present only in and galeritus. males] in the absence of the frontal pore . . . Separation of dyacorum and breviceps a solitary male was captured in the valley (and probably sabanus) from the races of Pinang." Consequently, if it can be with less reduced premolars is warranted shown that the skull now marked "type" is by actual measurement: cingulum length that of a different species, the specimen, not of P2 in all of the larger-toothed forms the skull, becomes the object to which the 0.4-0.5 MM. varies only from ; the same name galeritus will be restricted. 0.8-0.9 The specimen, as seen from Cantor's teeth in dyacorum from North Borneo and in writings, clearly had affinities with cervinus breviceps from N. Pagi (which however has and allied bats with which for years stu- 0.2-0.3 mm. dents have associated it. But the skull of the braincase longer) measures 0.2-0.5 mm. type label which I have photographs (with The reduced size of the lower tooth is es- attached and clearly legible) appears to be a of a member of the bicolor group. Its skull ' Thomas, 1894, Ann. Mus. Civ. Storia Nat., Genova, mastoid width and greatest zygo- XIV, p. 108. greatest (2) of the on and 2 It can only be determined by discovery matic width (measured the picture type or collecting of topotypes whether crumeniferus checked against a millimeter rule included should displace galeritus. 368 Bulletin American Museum of Natural History [Vol. LXXVIII pecially striking. In dyacorum there is and c-m3, 6.2. Further, the W-pattern of even indication of some displacement out- m3 is less reduced, and the dorsal promi- ward from the lower toothrow. nence of the zygoma rises abruptly instead From photographs it is possible to note of gradually, as in labuanensis. A few that P2 of labuanensis has a cingulum specimens of insolens were collected at length of approximately 1 mm. Schneideri, Perboewa as well as our series of labuanen- to which part of our material from Pagi is sis. referable, shows P2 only 0.5 mm. long, but Specimens from Banka (e.g., U.S.N.M. the occipital part of the braincase is con- 124834) agree wholly with labuanensis. A siderably more produced than is the case in large series in our collection from Talassa dyacorum. It is a larger animal. (Maros), South Celebes (celebensis), agrees The present group, like the bicolor group, in most of the characters with insolens, as has begun in certain species to reduce the do a few individuals from Bantimoerang, size of p2, particularly in the region of the South Celebes. In Celebes specimens, Greater Sunda Islands. Thus in the two which are rather smaller, m3 has been short- species labuanensis (forearm 48 mm.) and ened antero-posteriorly and its cusp-pat- .schneideri (forearm about 44-46 mm.) tern is incomplete, as in labuanensis. Indi- cingulum lengths of p2 measure longitudi- viduals in U.S.N.M. from Peleng Island ex- 0.4 0.2 actly resemble celebensis. Material from nally and respectively. And for the Jobi Island in Geelvink Bay; Weyland reasons given earlier I cannot decide Mountains, Dutch New Guinea; Hollandia, whether the type of galeritus is large- Dutch New Guinea; Aru Islands is still toothed or small-toothed. On geographi- much as insolens. But in material from New cal grounds based upon our specimens from Guinea mainland the enlarged, slit-like in- Pagi and upon borrowed material from fra-orbital foramina reappear and, in the Sumatra it should have small premolars. Fly River area at least, the are In fact schneideri may well be galeritus re- larger (forearm 50-52 mm.). New Hebri- named. It is desirable that topotypes of des specimens are like those of South New galeritus from Penang be secured and their Guinea. Both still have considerably teeth minutely described. All of our ma- smaller teeth and differently shaped P2 terial with long forearms from Sumatra from those of labuanensis. The zygomatic and Pagi appears to be referable to schnei- eminence in material from Fly River, Aru deri. and New Hebrides (cervinus) rises less From Borneo we have two members of abruptly than that of insolens but more the group in which p2 are unreduced: a suddenly than in the case of labuanensis. larger (labuanensis) and a smaller (in- A small but distinct depression (perhaps solens). Our labuanensis, from Landak in individual) between the rostral swellings is N. W. Borneo, match the photos of the developed in the three last mentioned. type specimen very closely, both in dimen- In none of the material reviewed (except sions and in structure; details include the schneideri of Sumatra) are p' much reduced. infra-orbital foramen, which is small but Of the more specialized offshoots of the elongate (0.8); the back of the palate level galeritus group the collection contains a with back of M2, provided with blunt post- small series of dyacorum from Perboewa, palatal spine; the form of P2, seen from Landak, North Borneo, thus nearly topo- above, rounded, not elongate; c-m3, about typical. Compared with the widely dis- 6.7 mm. Our specimens must be virtually tributed insolens-celebensis-cervinus type, topotypes of labuanensis. Insolens is repre- the skulls are at once much smaller and sented by Lyon's original material from the their palates are much shorter (c-m3 = 5.4 upper Pasir River. In it the foramen is mm.). The zygomatic eminence is small almost pore-like; the back of palate ex- and is placed far back. The infra-orbital tends a little behind m2 and lacks a spine; foramen is large and slit-like. The pre- P2 is more elongate than wide (1.0 X 0.8), maxillae scarcely extend beyond the ca- 19411 Tate, Results of the Archbold Expeditions. 35 369 Brachyotus is unrepresented in our ma- nines. The cingulum lengths P2 = 0.4 mm. terial. The forearm of specimens from the reduction of the lower tooth being Ceylon is given by Phillips as 48 to 51 mm. especially remarkable. The bats are no- Therefore, it is probably allied either to the table for possessing no lateral leaflets. small-toothed schneideri or the large- Pygmaeus, the only member of the entire toothed labuanensis. Longicauda, too, group thus far recorded from the Philippine which Sody discussed, requires for certain Islands (the status of the type of obscurus allocation in this survey re-study of the remains undetermined, in spite of Taylor's type. Our collections contain no material assumption that he had re-collected the from Java, but from the geographical dis- species), is also one of the most peculiar. position of the forms in surrounding terri- It possesses two lateral leaflets, the inner tory, no less than three forms must be ones almost continuous beneath the front looked for there. of the horseshoe, three septa down the H. coxi, obscurus, crumeniferus and papua front of the transverse leaf, emarginate ears, are discussed in the annotated list. They and a frontal sac in males only. It is al- must remain provisionally incertae sedis. most as small as cineraceus of the bicolor The systematic arrangement of the ori- group, even though the forearm is longer ental members of the group must remain (39-40 mm.). uncertain pending determination of the Skull with large paired rostral swellings exact status of the two old names, crumeni- (4.0 mm.); zygomatic and mastoid widths ferus and galeritus. There appears to be sub-equal (6.8); a weak jugal eminence; no doubt that cervinus, batchianus, celeben- cochleae 2.1 broad and 1.8 apart. Lower sis and insolens are conspecific; labuanen- incisors unthickened; c with strong pos- sis, schneideri, dyacorum, breviceps, sabanus terior cusp; p' unreduced; m3 with W- and pygmaeus are good species; but doubt pattern not greatly reduced; c-m3, 4.8; remains as to the relationship of the re- m-3, 2.8 mm. maining named forms. . e

OD -t t~

00

O~ o6lCc 0- O0 0 S s 0> t i S00 § e t w

0SX3

_ c12 0 . O . C000_0

00*U

. 00f 19411] Tate, Results of the Archbold Expeditions. 35 371

Hipposideros armiger group Upper incisors distinctly bilobate. Ca- (Gloionycteris, subgenus) nines strong with heavy cingula, but no pos- Large specialized bats derived from low terior cusps; p2 excluded, terete in section down the speoris stem. Transverse nose- through cingulum; p4 in contact with c. leaf complicated: somewhat enlarged, be- Hypocones of ml and m2 as in cervinus and coming lobate, the lobes slightly crenulate, speoris groups; parastyle of m3 as in pocketed, the degree of development partly speoris. Outer lower incisor thicker than dependent on sex; horseshoe without an- inner one, as in allied groups; crown height terior notch; lateral leaflets four. Ears of P2 from 1/3 to 1/2 that of P4. sub-acute, their outer edges emarginate The members of this group are restricted behind the tips. Ears and noseleaves to the Asiatic mainland, Formosa and the deeply pigmented. Liu Kiu Islands. H. turpis of Liu Kiu, as Skull heavily ossified, the rostral area pointed out by Andersen,' is a member not secondarily flattened, perhaps in accommo- of the speoris but of the armiger group. dation to enlargement of the nasal foliation Recently pendleburyi has been described and frontal sac (compare pratti). Sagittal from Siam. and temporal crests even more strongly de- Study of armiger bats from Sikkim (Field veloped than in diadema group. Palate Mus. N.H. 35421), near the type locality of broad (distance between ml-', approxi- armiger, North Burma (A.M.N.H.), Siam mately twice width of ml). Premaxillae (U.S.N.M.), Szechwan (A.M.N.H.), Fukien solidly fused together from their middle (A.M.N.H.), and a mounted specimen, backward, the resulting fused structure skull inside, from Amoy (M.C.Z. 6283), fan-shaped behind (not pointed, as in the type locality of swinhoei, offersno characters weakly fused diadema). Incisive foramina upon which to base any subspecies. U.S. not enclosed by premaxillae. In the pos- N.M. 239916 from Mokusaku, Formosa, terior narial region a distinct step down which represents terasensis is inseparable from the narial roof to the level of the roof from armiger from the mainland. The of the pterygoid fossa can be observed. validity of debilis from Malay Peninsula is (In other groups already described the questionable. Turpis, on the other hand, roofs of narial fossa and pterygoid fossa are of which a large series exists at M.C.Z. and approximately continuous on a single level.) one at A.M.N.H., is distinct chiefly be- Cochleae small; their diameters rather less cause so markedly small. Pendleburyi too than their distance apart. Angular process is perhaps distinct. of mandible slenderer than in diadema 1 Andersen, 1906, Ann. Mag. Nat. Hist., (7) XVII group. p. 38. .co1

,-~OD

c64 P

ez

t4

Cf P 19411] Tate, Results of the Archbold Expeditions. 35 373

Hipposideros diadema group' External lobe of incisors present. Ca- (Phyllorhina,2 subgenus) nine without posterior cusp. Anterior The type species of this group, diadema, premolar partly or wholly excluded from was recorded from Timor by Geoffroy St. toothrow. Parastyle of m3weak. Second Hilaire in 1813. It was the third species of lower incisor distinctly thicker than first Hipposideros to be made known to science. (as in speoris group); lower canines with About a score of additional forms have now enlarged cingula, approximated; crown of been named with type localities extending P2 about one-half height of crown of P4. from Burma and the Nicobar Islands to the Two moderately distinct divisions can Philippines, northern Australia and the be discerned in this group: the diadema eastern Solomon Islands. Osgood3 records division proper with numerous described it from Lao Bao, Annam. H. lankadiva insular races, distributed from Burma and and allies occupy the Indian Peninsula and Siam to Australia and the Solomon Islands, Ceylon. and the lankadiva division of Peninsular Size from medium-large to large (forearm India and Ceylon. Their characters were from 64 mm. to 96 mm.). Many forms of defined by Andersen, 1905. diadema are grayish, marked with white along the sides dorsal to the wings. Many With but a limited number of specimens show lack of pigment in the facial foliations. of diadema for study the following two pro- Frontal sac absent in both sexes. Horse- portional distinctions are seen: the ratio shoe without anterior notch (as in speoris rostral width to zygomatic width results in group), provided with three or four lateral 48 to 51 per cent in our material from Java, leaflets; an incipient "connecting process" Bali, Papua, Solomon Islands, and even (compare Rhinolophus) recalls the cyclops the large form pelingensis (forearm, 96 and muscinus groups; transverse noseleaf mm.) from South Celebes; and in the large, weakly lobate. Ears large, acute, smaller form (forearm 85 to 87 mm.), from their outer margins emarginate behind tip. both North and South Celebes, 57 to 58 per Skull with broad, well-inflated rostral cent-an expression of the extreme infla- region, well-broadened zygomata and mas- tion of the rostrum and relative narrowness Rostral width of the zygomata. The second ratio, of toid areas. Zomat width about 57 per mastoid width to zygomatic width, pro- Zygomatic width vides widely variable results; maximum inter-orbital width cent; about 38 per cent. mastoid combined with less zygomatic rostrala width expansion appears in the smaller North Sagittal crest high, reaching forward onto and South Celebes animals and in the compressed inter-orbital region. Cochleae Papua material-85 to 95 per cent; the as in speoris group, infra-orbital foramen reverse occurs in the giant South Celebes very large, closed by a slender bar which race, only 77 per cent; Balinese, 81 per in some forms reaches a length of 3 mm. cent; Solomon Islands, 83 per cent; but Incisive foramina in premaxillae quite nar- Javanese only 76 per cent. row, their posterior walls formed by maxil- Summarizing from our material, photo- lae, except in material from Bali, Java, graphs of and notes on types, and on speci- Borneo. Angular process of mandible mens loaned to us, three weakly marked well developed as in speoris group. divisions are indicated in diadema: 1 Andersen, 1905, Ann. Mag. Nat. Hist., (7) XVI, pp. 497-507; 1907, Ann. Mus. Civ. Storia Nat., 1.-Diadema proper, and numerous Genova, (3) III, pp. 6-11. 2 The genus Phyllorhina Bonapart, a synonym of named races inhabiting Celebes, Timor and Hipposideros, is employable in the subgeneric sense all of the regions of New Guinea, northern to designate the diadema group of Hipposideros. 3 1932, Field Mus. Nat. Hist., Zool., XVIII, p. 221. Australia and the Solomon Islands: 374 Bulletin American Museum of Natural History [Vol. LXXVIII

TYPE REGION FOREARM 2.-Larger bats of the Greater Sunda diadema Timor 77 mm. forearm 80 to proper' region: 90 mm.; rostrum diadema euotis Batchian Island "84-89" usually more flattened than in the diadema Andersen division; rostral width 8.6 to 10 mm.; mas- diadema speculator South Celebes "76-87" toid width proportionately less. Andersen diadema custos Kei Island 78 TYPE REGION FOREARM Andersen diadema nobilis Java 87 mm. diadema pullatus Papua "75-81" Horsfield Andersen diadema enganus Enganus Island 89 diadema trobrius Trobriand Island "73-76" Andersen Troughton diadema natunensis Natuna Island 88 diadema regitnae N. Queensland "79-82" Chasen Troughton2 diadema griseus Luzon Island "82-86" diadema demi8sus San Christobal 63.5 Meyen4 Andersen Island diadema vicarius Sarawak "80-86" diadema mirandus Admiralty Island 67 Andersen Thomas diadema oceanitis Guadalcanar Island 79 A specimen of nobilis A.M.N.H. 106736 Andersen from Cheribon, Java, and two, M.C.Z. 38416-38417 from Batavia, agree closely This species, though highly variable in with each size from place to place, is in general char- other. Specimens of griseus acterized borrowed from M.C.Z. are slightly smaller by moderately to much expanded than nobilis, particularly respecting their rostrum (49 to 57 per cent of zygomatic outer lower width), and widely mastoid incisors. Some suggestion ap- expanded pears in the Philippine bats of enclosure of regions (83 to 94 per cent of same); the premaxillae the incisive foramina in the premaxillae, as not enclosing the incisive is typically the condition in the Balinese foramina; the teeth rather small. Our form. collections include several distinct races: From South 3.-Very large species from Celebes, Celebes (speculator): fore- Peleng, and Rubiana, Solomon Islands: arm, ='=85; c-m3, 12.5; rostrum very wide (10 Forearm, 94-98 mm.; zygomata much mm.) and much inflated, longi- expanded; rostrum moderately enlarged; tudinally grooved; incisive foramina not premaxillae not wholly enclosing foramina, completely enclosed by premaxillae; upper 14.0. incisors almost touching; p2 included; heel c-im3, Teeth quite heavy, par- of ticularly p4. These bats appear to be P4 not very broad. Of the same general allied to dinops. And dinops itself, because type are the three specimens from North of the Celebes (vicarius ?) and specimens from great expansion of its zygomata, be- Java The of trays closer affinity with lankadiva than do (nobilis). rostrum nobilis is, the forms with inflated rostra however, narrower and propor- (9 mm.), not quite so tionally narrower zygomatic width. high, the forearm 85 mm., and p2 is ex- cluded. The South Celebes series (pelingensis) is The series from composed of bats almost as large as dinops Fly River, Papua (pul- of Rubiana. Forearm about 95 mm., latus), although agrees pel- smaller, closely also age with pale brownish cast (similar to with nobilis. It has forearm 74 c- that mm., mentioned for custos), and the white mark- m3, 11.8, and a narrower rostrum (8.5). The bats from Solomon ings present in most forms obsolescent and Islands, identi- buffy. Rostrum_much lower and flatter fied by Sanborn3 as diadema oceanitis, also than in those bats agree substantially with the described-more like Papuan series. the rostrum of lankadiva (but back of palate 1 No topotypical specimens seen. Information less deeply V-shaped than in based upon original description, in which the forearm lankadiva). in illustration, marked "natural size" is 77 mm.; Zygomata widely expanded (20.8 mm.) in not Andersen (1905) who gives forearm as 84-91 mm. 2 I cannot separate reginae (M.C.Z. 29092 from 4Anderseni Taylor from Philippines, considered a Lake Barrine, Queensland) from our Fly River pul- race distinct from griseus, was held by Lawrence to be latus. synonymous. But Taylor's animal was certainly 3 1931, Sanborn, Field Mus. Nat. Hist., Zool., much smaller and might possibly have represented XVIII, p. 24. something analogous to trobrius or pullatus. 1941] Tate, Results of the Archbold Expeditions. 35 375 comparison with rostral width (10.2 mm.), The lankadiva section contains only mastoid width slight. lankadiva, indus, mixtus, unitus, schistaceus. TYPE REGION FOREARM The second, third and fourth are scarcely, if at all, separable from lankadiva of ditops Andersen Rubiana Island 95 mm. Ceylon, pelingensig Shamel Peleng Island 93-97 as the photographs of their type specimens show. Schistaceus appears to be farther Although nicobarensis and masoni belong removed. All have the characteristic pal- to the diadema group no material is avail- able to assist in determining their relation- ate and very heavy incisors and canines of ships. Probably they, with specimens in lankadiva. the U.S.N.M. and F.M.N.H. from Siam, Only one specimen representing this represent the western and northern limits branch of Hipposideros seems to exist in of range of nobilis. an American museum-M.C.Z. 32972. . ct * c:>

. *t

, . * -c;

ga

ct s *t S ce,

o X o m -* -t * L - cs eD z .¢

* ,4

a cs .

* ^ .

e . ^

^b . <;G '.. t

. . 2 Q E-c; X = tH Xb:

co UD ob-. Uoo

X* AO 9,> *uSu 1941 ] Tate, Results of the Archbold Expeditions. 35 377

Hipposideros speoris group phylla, grandis, neglectus, barbensis, pouten- (Hipposideros, subgenus) sis and probably deformis. Medium sized, strongly built species H. abae, although it resembles speoris in with emarginate ears, a small process at the the obsolescence of the frontal sac in fe- antitragal fold, three lateral leaflets. males and in having the anterior margin of Skull with rostrum inflated, inter-orbital the horseshoe entire, is a much larger area short and constricted, a moderately species of bat, with forearm 57 mm. and developed sagittal crest. Zygomata well c-mi3, 8.3. The incisive foramina are al- expanded. Palate (excluding premaxillae) most wholly enclosed in the premaxillae. very slightly longer than broad. Incisive A posterior cusp is present on the canine. foramina partly enclosed by premaxillae. H. speoris, the first bat of the genus Premaxillae quite firmly fused together, Hipposideros to be recognized, was de- their united posterior process wedge- scribed by Schneider in Schreber's Sauge- shaped. Cochleae moderately large (width thiere, Vol. 5, in an unpaged separate slightly exceeding the distance between bound (in Amer. Mus. copy) opposite pages them). Angular processes of mandible 962 and 966, and illustrated in plate 59B. strong, prominent. Speoris, according to Schneider, had three Upper incisor weakly bilobed; c without lateral leaflets, a horseshoe wider above posterior cusp; p2 partly excluded from than below, a transverse leaf about as toothrow; outer lower incisor slightly wide as the horseshoe, a large transverse thicker than inner; crown of P2 low-from frontal sac, ears acute and emarginate, and 1/2 to 2/3 height of crown of p4. part of the tip of the tail extending beyond Two sharply distinct species, which em- the membrane. Schneider's title-line gave brace a number of races and synonyms, are the origin of speoris as "Ostindien." Geof- included under this heading; namely, froy and Peron believed the point of origin speoris proper of Peninsular India, and to be Timor but Peters considered that larvatus = vulgaris of the Malay-China- speoris came from Tranquebar, on the Sunda region. Abae of central Africa, southeast coast of India. Dobson listed though specifically separable, yet is much material from the Indian peninsula and closer to speoris than larvatus. Ceylon. Andersen gave "Ceylon, Kanara, Other characters shared in common are: Bombay, Khandeish, Mysore," and de- simple transverse leaf with three anterior scribed a smaller race pulchellus from Bel- septa; three lateral leaflets; short, rela- lary. Speoris has its headquarters in tively massive skull, with bullae only peninsular India and Ceylon and I now slightly wider than their distance apart; suggest restriction of its type locality to broad rostrum with well-developed supra- Tranquebar. temporal ridges and incipient post-orbital Bats referred by Andersen (1918) to processes; heavy dentition. speoris speoris are said to have the average Speoris is distinguished from larvatus by forearm 52 mm. (49.8-54) and the length obsolescence of the frontal sac in females of skull from the base of canines 19.7 (19- (it is represented merely by a tuft of modi- 20.3). The same measurements in speoris fied hairs), by its smaller size (forearm 54 pulchellus were 49.4 (45.8-51.5) and 18.8 mm. or less). To it are referable either as (18.0-19.8). The forearm in Schreber's races or as synonyms: pulchellus, temple- plate measured 47 mm. But the illustra- toni and apiculatus; also probably peni- tion is not necessarily of natural size and cillatus and taitiensis. the structure may also be slightly fore- Larvatus is considerably larger than shortened. An excellent illustration of speoris. The female possesses a frontal "speoris" was published by Phillips' re- sac. Both sexes are recognized by the cently. well-defined notch in the middle of the A series of undoubted speoris, one from front edge of the horseshoe. A quite long Trichinopoli and 18 from Ceylon, is avail- list of names is here referred as subspecies 1 Phillips, 1935,, Manual of of Ceylon, p. -or synonyms: vulgaris, insignis, lepto- 93. 378 Bulletin American Museum of Natural History [Vol. LXXVIII able at M.C.Z. for study. The characters Speorifera Gray, 1866, of which vulgaris is of the species, based upon M.C.Z. 33748, designated type, falls as a synonym of e, and 27508, 9 (both from Ceylon), may Hipposideros, subgenus, of which Selater' be defined as follows: has designated speoris the type. Ear slightly emarginate along outer edge, H. larvatus is recognizable by a number with distinct "kugelf6rmig Anhang" at of characters, one of the easiest to distin- antitragal fold, 1/2 to 3/4 of pinna naked, guish being the small but clearly defined 15 mm. from notch, 10 from crown. Fron- median cleft in the anterior margin of the tal sac large in males, in females obsolescent horseshoe. The forearm length varies but invariably marked by tuft of dark from 57 to 63 mm. Tibia, 21 to 24, or less brown hairs (hence, probably Gray's name than 50 per cent of the forearm. Calcar = of penicillatus). Supra-orbital tubercles 12 mm. Ear emarginate. Lateral leaf- developed, bearing several long setae and lets three. Transverse noseleaf with 3 probably provided with glandular openings. septa and 4 cells. Frontal sac present in Transverse leaf divided anteriorly by three both sexes. This last character, the cleft distinct septa into four cells. In front of it noseleaf and the much greater size readily the broad analogue of the "connecting distinguish larvatus from speoris. process" of Rhinolophus is angular, low, H. 1. neglectus Sody, of Borneo, which and bears two setae each side. In front of was reputed to have the foreariii longer the last the medium-sized horseshoe shows than the Javanese race, is only doubtfully only the faintest trace of emargination of valid. Study of large series from Java, the center of its front edge (wholly unlike Borneo, a number of the islands of the the deep notch observable in larvatus). South China Sea, Sumatra, Malay Penin- Three lateral leaflets. Upper lip with two sula, Siam, Indo-China and Hainan shows minute papillae, one above the other. chiefly a considerable degree of individual Lower lip with two pairs of warts near the variation in the length of the forearm (3-4 midline, the inner ones larger than the mm. is common). H. 1. grandis G. M. outer. Forearm, 51 mm. Thumb with Allen, whose type has been returned to the basal joint 3.5, median joint 3.0. Tibia, 20; Indian Museum, has not been seen by me. calcar, 9; tail, 21-22, five-jointed, the Our Chindwin River topotypes are all terminal point half exserted from uropata- strongly rufous and have forearms 62-64 gium. mm. The teeth appear very slightly heav- Skull rather short and massive. Zygo- ier than those of Javanese bats, but I am matic width exceeding mastoid width. unable to observe the differences in m3 of Premaxillae not exceeding canines an- which Allen wrote. teriorly. Inter-orbital constriction pro- As stated elsewhere, the types of larvatus, nounced (2.6-2.9); rostral width, 5.1-5.4. vulgaris, deformis, insignis, barbensis and Incisive foramina not wholly enclosed poutensis have been studied; also topo- by premaxillae. Cochlea (2.7) slightly typical material representing grandis, proba- broader than their distance apart (2.1). bly a synonym of leptophylla, and neglec- Upper incisor with external lobe rudi- tus. The conclusion seems inescapable mentary. Outer lower incisor not greatly that all represent a single species of exten- thicker than inner one. Canines less pro- sive range through Assam and Burma to odont than in cervinus. Upper p2 not China, Indo-China, Sumatra and adjoining wholly excluded from toothrow; lower p2 islands, Java and Borneo. It appears to be with crown 1/2 to 2/3 height of crown of p4. unrepresented in the Philippine Islands Condylo-canine length, 16.8; c-m3, 7.1 and Celebes, nor does it extend (?) into mm. peninsular India. If subgeneric names should be employed under Hipposideros, speoris must belong in Hipposideros muscinus group Hipposideros, subgenus. Speoris Schneider Ear acutely pointed as in the cyclops and vulgaris Horsfield (= larvatus) are group and, to a less degree, commersonii probably consubgeneric, in which event I Sclater, 1901, Mamm. S. Africa, II, pp. 116-118. 19411] Tate, Results of the Archbold Expeditions. 35 379

group. Nasal foliations bearing the dently had North Australian material, specialized club-shaped structures described which they alluded to as muscinus. for cyclops. Transverse leaf with three Our series from the Astrolabe Mountains, septa and four cells. Lateral leaflets, two. behind Port Moresby (one also recorded by Tail =i= 150 per cent of tibia (compare Thomas, 1897), is anatomically identical cyclops). Length of basal phalanx to to the specimens collected in the western metacarpus of thumb is as 1.5 to 4 mm. division of Papua, though in some of the Thus the metacarpus is much elongated, a Astrolabe animals the toothrow is a little condition even more pronounced in Coelops. longer. For distribution, see map, figure 3. Skull with rostral area greatly broadened The status of wollastoni is doubtful. Its and inflated. Mastoid portion of skull skull appears to be very close also to that of scarcely at all widened. Zygomata moder- muscinus but the special characters of the ately expanded. Sagittal crest high, fork- transverse noseleaf, described by Thomas, ing to form pair of temporal crests which distinguish it. enclose small frontal depression. Infra- orbital foramen minute, terete, sometimes Hipposideros cyclops group paired. Palate short. Premaxillae fused (Doryrhina, subgenus) throughout. Incisive foramina very large, Bats of this group are restricted to oval, enclosed by ligulate processes from Africa. They show the specialized club- sides of premaxillae which reach maxillae, like structures of the nasal region in com- and posteriorly by a single median, fan- mon with muscinus, structures so peculiar, shaped process formed by the fused pre- specialized and seemingly functionless maxillae. Postero-laterally the walls of that they are unlikely to have arisen inde- the foramina are formed by maxillae. pendently. Transverse noseleaf in cyclops Cochleae extremely large, the largest for with three well-developed septa and four the genus; width of single cochlea about 6 cells. Two accessory leaflets to horseshoe. to 8 times their distance apart (compare No tubercle-like process above eye. Tail cyclops). Angular process elongate, con- greatly shortened-only 70 per cent of nected by a thin web of bone to articular tibia. process. Skull with well-rounded rostral swell- Teeth arranged with molariform series ings, a frontal depression, braincase un- parallel. Upper incisors weakly bilobate. widened at mastoid, palate short. Pre- Canine without posterior cusp. Canine. maxillae wholly enclosing incisive foramina. and p4 in contact, excluding minute p2. Cochleae large-width of one approxi- No trace of thickening of i2; no overlap of mately four times their distance apart. il. Anterior lower p small, its cusp only Incisor retaining minute external lobe. 1/4 of height of crown of p4. The W-pat- Canine without cusp. Canine and p4 in tern of m3 unreduced. contact, totally excluding minute p2. No Although our material represents only thickening of outer lower incisors. Papuan muscinus, there can be almost no Allen (1939) admits only the one species doubt that muscinus, semoni and stenotis of cyclops, with micaceus De Winton and North Australia are conspecific. Dahl and langi J. A. Allen in synonymy. It appears Collett, both of whom wrote in 1897, evi- to be restricted to Central and West Africa. 380 Bulletin American Museum of Natural History [Vol. LXXVIII ANNOTATED LIST OF NAMED FORMS OF HIPPOSIDEROS' Alphabetically arranged abae J. A. ALLEN, 1917, Bull. Amer. Mus. Subspecies of caffer (Allen, 1939). Nat. Hist., XXXVII, p. 432. MATERIAL.-One specimen at TYPE REGION.-Belgian Congo. A.M.N.H. Related to speoris. antricola PETERS, 1861, Monatsber. Akad. MATERIAL.-The type and series of Wiss. Berlin, p. 709. paratypes at A.M.N.H.2; specimens at TYPE REGION.-Luzon, Philippine Is- M.C.Z. lands. Remarks.-Distinguished from speoris Related to bicolor. by its greater size; forearm of type, 58.5 MATERIAL.-Series at M.C.Z.; one at mm.; Allen gave a variation range of U.S.N.M. forearm from 54 to 60.5 (shortest 56 in REMARKS.-A smaller species of two males). Andersen (1918) gave the maxi- closely similar (the other erigens) present mum of speoris as 54 mm. and minimum in the Philippines. Forearm 39. The (sp. pulchellus) as 45.8. Both have type could not be found while I visited "dark" and "red" phases. Allen be- Berlin in 1937. lieved abae related to caffer. apiculatus GRAY, 1838, Mag. Zool. Bot., II, albanensis GRAY, 1866, Proc. Zool. Soc. p. 492. London, p. 220. TYPE LoCALITY.-Madras, India. TYPE REGION.-N. W. Queensland. Related to speoris; probably synony- Related to bicolor, probable synonym mous. of aruensis. MATERIAL.-Co-types: B.M. 19a, MATERIAL.-The type, B.M. 66.4.23.7, forearm 51 mm.; 19b (skull in skin; c; skin, with skull in fragments; photo. forearm 50); 19c photo. of skull; 19e of palate and jaws. (forearm 49; photo. of skull). REMARKS.-Forearm, measured in armiger HODGSON, 1935, Jour. Asiatic Soc. 1937, 39 mm.; by Gray, "1 1/2 inches." Bengal, IV, p. 699. amboinensis PETERS, 1871, Monatsber. TYPE REGION.-Nepal, India. Akad. Wiss. Berlin, p. 323. MATERIAL.-Non-topotypical mate- TYPE REGION.-Amboina Island. rial from Sikkim, Burma, China, etc., in Related to bicolor, probably synonym several museums in U.S.A. of aruensis. aruensis GRAY, 1858, Proc. Zool. Soc. Lon- MATERIAL.-The type, Berlin, No. 369; don, p. 107.-GRAY AND GRAY, 1859, photo. of skull (back of braincase broken). Cat. Mamm. New Guinea Brit. Mus., pp. anderseni TAYLOR, 1934, Philippine Land 1-2. Mammals, p. 246. TYPE REGION.-Aru Islands. TYPE REGION.-Luzon, Philippine Related to bicolor. Islands. MATERIAL.-The type: B.M. 58.2.- Related to diadema, probable synonym 20.1, ad. ci, with skull in fragments; of griseus. photo. of palate and toothrows. MATERIAL.-The type not seen. atratus KELAART, 1852, Prodr. Fauna Zey- angolensis SEABRA, 1898, Jour. Sci. Math. lanica, p. 16. Phys. Nat., Lisboa, (2) V, p. 256. TYPE REGION.-Ceylon. TYPE REGIoN.-Angola, West Africa. Related to aruensis, of bicolor group. 1 Tricuspidatus, stoliczkanus, and related forms are MATERIAL.-Topotypes at M.C.Z. omitted. (obtained from W. W. A. Phillips, Cey- 2 Abbreviations used in this list: A.M.N.H., American Museum of Natural History, New York; lon); one at U.S.N.M. M.C.Z., Museum of Comparative Anatomy, Cam- bridge, Mass.; F.M.N.H., Field Museum of Natural atrox ANDERSEN, 1918, Ann. Mag. Nat. History, Chicago; C.M., Carnegie Museum, Pitts- II, p. 380. burgh; U.S.N.M., United States National Museum, Hist., (9) Washington; B.M., British Museum, South Kensing- TYPE REGION.-Selangor, Malay Pen- ton, London; Berlin, Museum fiur Naturkunde, In- validenstrasse Berlin. insula. 1941]1 Tate, Results of the Archbold Expeditions. 35 381 Subspecies of gentilis, and related to TYPE REGION.-Central India. bicolor. Related to galeritus. MATERIAL.-The type: B.M. 1.3.9.4, MATERIAL.-Photo. of skull of type. 9Q, with photo. of skull. breviceps TATE, described in this paper. aureus KELAART, 1852, Prodr. Fauna Zey- caffer SUNDEWALL, 1846, Ofversigt K. lanica, p. 17. Svenska Vet. Akad. Forh., Stockholm, TYPE REGION.-Ceylon. III, Art. 4, p. 118. Probable synonym of speoris. TYPE LOCALITY.-Near Port Natal, MATERIAL.-Co-types: B.M. 52.5.9.3 East Africa. (9, forearm 51); 52.5.9.4 (forearm 48); MATERIAL. Skin, B.M. 49.11.22.11, 52.5.9.5 (forearm 51), all from Trinco- and skull, 48.6.2.16 marked "co-type" mali. (forearm 44). Large series in museums REMARKS.-No data regarding skulls. of U.S.A. auritus TOMES, 1859, Proc. Zool. Soc. -calcaratus DOBSON, 1877, Proc. Zool. Soc. London, p. 76. London, p. 122. TYPE REGION.-India (label of type TYPE LoCALITY.-Duke of York Is- specimen). land, between New Britain and New Probable relative of fulvus and bicolor. Ireland. MATERIAL.-The type: B.N. 7.1.1.303 MATERIAL.-The type: B.M. 77.7.- (forearm 38); photo. of skull. 18.13, adult 9, with photo. of skull; barbensis MILLER, 1900, Proc. Wash. Acad. topotype, A.M.N.H. Sci., II, pp. 233-234. cekebensis SODY, 1936, Natuur. Tijdschr. v. TYPE LOCALITY.-St. Barbe Island, Ned. Ind., XCVI, p. 47. between Linga and Borneo. TYPE REGION.-S. Celebes. Closely related to larvatus. Related to galeritus and cervinus. MATERIAL.-The type and paratype MATERIAL.-Sody coll. No. 4, adult 9 series at U.S.N.M. (the type), with photo. of skull; series in batchianus MATSCHIE, 1900, Sauget. Archbold collection. Kukenthal . . . Halmahera, Batjan u. Ann. Nat. Celebes, p. 273. centralis ANDERSEN, 1906, Mag. TYPE LoCALITY.-Batchian Island. Hist., (7) XVII, p. 275. Related to galeritus. TYPE REGION.-Uganda. No. Subspecies of caffer (Allen, 1939). MATERIAL.-The type: Berlin, MATERIAL.-The type: B.M. 99.8.- 15628, 9, in alcohol; photo. of skull. c no of skull; beatus ANDERSEN, 1906, Ann. Mag. Nat. 4.8, (forearm 52), picture Hist., (7) XVII, p. 279. small series at U.S.N.M. and M.C.Z.; TYPE REGION.-Cameroons. large series at A.M.N.H. Related to caffer. cervinus GOULD, 1863, Mamm. Australia, MATERIAL.-Series at M.C.Z.; one at III, P1. xxxiv. U.S.N.M. TYPE REG10N.-N. W. Australia. bicolor TEMMINCK, 1835, Monogr. Mamm., Related to galeritus. II, p. 18. MATERIAL.-The type: B.M. 55.11.- TYPE LoCALITY.-Anjer coast, N. W. 7.13, with photo. of skull (back of skull Java. destroyed); ample material from New MATERIAL.-Series of co-types studied; Guinea in Archbold collection; large photo. of skull of co-type "d." series from New Hebrides at U.S. bicornis HEUGLIN, 1861, Nova Acta Akad. museums. Caes. Leop. Carol., Halle, XXIX, No. 8, cineraceus BLYTH, 1853, Jour. Asiatic Soc. pp. 4, 7. Bengal, XXII, p. 410. TYPE REGION.-Eritrea. TYPE LOCALITY.-Punjab Salt Range Synonym of caffer (Allen, 1939). Survey, near Pind DAdan Khan. brachyotus DOBSON, 1874, Jour. Asiatic Soc. Related to bicolor; the smallest mem- .Bengal, (N.S.) XLIII, pt. 2, p. 237. ber of the group. 382 Bulletin American Museum of Natural History [VOl. LXXVIII

MATERIAL.-Scattered specimens from vinus; oldest name for that group if rela- Burma, Siam, Indo-China in U.S. mu- tionship can be proved. seums. MATERIAL.-Peron's plate xxxv only. REMARKS.-Blyth writes "smaller REMARKS.-It is difficult to determine than . . . murinus (Elliot)," then gives from Peron's plate the number of lateral the forearm as 1 3/16 inches. Murinus leaflets. The 4-celled transverse leaf, was proposed first by Gray, 1838. A the forearm length 53 mm. (on the plate), forearm of only 30 mm. is smaller than the frontal sac-all are characters com- any I have seen in the genus. mon to males of speoris and cervinus. commersonii E. GEOFFROY, 1813, Ann. Mus. Geoffroy's (1813) picture of "crumeni- Hist. Nat. Paris, XX, p. 263. ferus" which clearly shows three lateral TYPE REGION.-Madagascar. leaflets may have been drawn from dif- MATERIAL.-One specimen from ferent material. Peron's figures show a Madagascar at U.S.N.M.; African ma- more lightly built bat than speoris. Also terial at U.S.N.M. and M.C.Z. Timor is well within the extensive distri- coronatus PETERS, 1871, Monatsber. Akad. butional range of the galeritus group, Wiss. Berlin, p. 327. whereas, excepting the doubtfully identi- TYPE REGION.-Mindanao, Philippine fied taitiensis, speoris is unknown beyond Islands. India and Ceylon, some 2000 miles from This may represent the calcaratus Timor. (New Guinea) group in Philippines. cupidus ANDERSEN, 1918, Ann. Mag. Nat. MATERIAL.-None. Hist., (9) II, p. 383. REMARKS.-Peters erected the sub- TYPE LoCALITY.-Eaga, Central Div., genus Thyreorhina ("upper noseleaf Papua. with thickened margin") to contain this Related to calcaratus. species. It is possibly allied to calcara- MATERIAL.-The type: B.M. 97.12.- tus, which was still undescribed when 6.4, y. ad. e in alcohol, and photo. of Peters wrote. Neither Taylor nor Law- skull; series from New Guinea in Arch- rence re-collected it during their work in bold collection. the Philippines. curtus G. M. ALLEN, 1921, Revue Zool. coxi SHELFORD, 1901, Ann. Mag. Nat. Afrique, IX, p. 194. Hist., (7) VIII, p. 113. TYPE LoCALITY.-Sakbayeme, Came- TYPE REGION.-Sarawak, Borneo. roons. MATERIAL.-Photo. of skull only. Related to caffer group. REMARKS.-An anomalous species MATERIAL.-The type seen at M.C.Z. with exceptionally broad intertemporal custos ANDERSEN, 1918, Ann. Mag. Nat. and mastoid region, and high rostral Hist., (9) II, p. 381. swellings. Forearm, 53 mm. Shelford TYPE LoCALITY.-Kei Island. describes two lateral leaflets, character- Related to diadema. istic of the galeritus group, but says "no MATERIAL.-The type: B.M. 10.3.- frontal gland in 9." No skull charac- 1.27, ad. ci, and photo. of skull. ters were given and unfortunately I did cyclops TEMMINCK, 1853, Esquisses Zool. not find the type in London. Guin6e, p. 75. Coxi may well belong with other spe- TYPE REGION.-Gold Coast, West cialized members of the galeritus group of Africa. Borneo and Philippines. Pygmaeus also MATERIAL.-Ample series in most lacks the frontal sac in the 9 . U.S. museums. crumeniferus PERON, 1807,1 Voyage De- REMARKS.-Cyclops (= micaceus = couv. Terres Australes, Atlas, P1. xxxv. langi), which Peters (1871) separated TYPE REGION.-Timor. from Hipposideros under the subgeneric Probably related to galeritus and cer- name Doryrhina, finds its nearest rela- tives in the Torresian region (muscinus, 1 Date from Sherborn. semoni, stenotis). The very peculiar pair 1941 ] Tate, Results of the Archbold Expeditions. 35 383

of club-like processes, placed one behind Derivative of galeritus group. the other on the "sella" and posterior MATERIAL.-Photo. of type-skull only; leaf, respectively, are unlikely to have series from Borneo in Archbold collec- originated independently. In the skulls tion; specimens in U.S.N.M. of the African and Australian bats, REMARKS.-Thomas assumed dya- quite profound differences are apparent, corum to be "allied to bicolor" because of which shows that separation took place its lack of lateral leaflets, which are in remote times. represented on each side by two small debilis ANDERSEN, 1906, Ann. Mag. Nat. papillae. But the skull is that of a mem- Hist., (7) XVII, p. 37. ber of the galeritus group, specialized by TYPE REGION.-Malay Peninsula. shortening of the palate, toothrow and Subspecies of armiger. premaxillae, and reduction of pt. See MATERIAL.-The type: B.M. 79.11.- also under sabanus. 21.80, e (forearm 88), and photo. of enganus ANDERSEN, 1836, Ann. Mus. Civ. skull; few in U.S.N.M. from Siam. Storia Nat., Genova, (3) III, p. 8. deformis HORSFIELD, 1824, Zool. Res. Java. TYPE LoCALITY.-Engano Island. TYPE REGION.-Java. Related to diadema. Probable synonym of larvatus. MATERIAL.-Photo. of skull of type MATERIAL.-The type: B.M. 79.11.- (ad. 9 ) only. 21.94 (forearm 53), no photo. of skull. erigens LAWRENCE, 1939, Bull. Mus. Comp. demissus ANDERSEN, 1909, Ann. Mag. Nat. Zool., LXXXVI, p. 56. Hist., (8) III, p. 268. TYPE LOCALITY.-Mt. Halcon, near TYPE REGION.-San Christobal, Solo- Calapan, Mindoro, Philippine Islands. mon Islands. Related to bicolor, distinct from antri- Related to diadema. cola. MATERIAL.-The type: B.M. 13.11.- MATERIAL.-Type and paratypes at 7.1, ad. e, photo. of skull (back of skull M.C.Z. broken). euotis ANDERSEN, 1905, Ann. Mag. Nat. diadema E. GEOFFROY, 1813, Ann. Mus. Hist., (7) XVI, p. 502. Nat. Hist. Paris, XX, p. 263. TYPE LoCALITY.-Batchian Island. TYPE REG10N.-Timor. Related to diadema. MATERIAL.-None from Timor. MATERIAL.-The type: B.M. 7.1.1.- dinops ANDERSEN, 1905, Ann. Mag. Nat. 312, ad. e, and photo. of skull (braincase Hist., (7) XVI, p. 502. destroyed). TYPE REGION.-Rubiana Island, Solo- fulgens ELLIOT, 1840, Cat. Mamm. S. mon Islands. Maharatta Country, p. 8. Related to diadema. TYPE REGION.-S. Maharatta Coun- MATERIAL.-The type: B.M. 88.1.5.- try, India. 22, ad. 9 in alcohol, with photo. of skull. Probably related to bicolor. doriae PETERS, 1871, Monatsber. Akad. MATERIAL.-None. Wiss. Berlin, p. 326. fuliginosus TEMMINCK, 1853, Esquisses TYPE REGION.-Sarawak, Borneo. Zool. Guin6e, p. 77. Probably member of bicolor group, TYPE REGION.-"Coast of Guinea." perhaps related to aruensis. MATERIAL.-None. Related to caffer. dukhunensis SYKES, 1831, Cat. Mamm. MATERIAL.-Small series at M.C.Z. Dukhun, p. 4. fulvus GRAY, 1838, Mag. Zool. Bot., II, p. TYPE REGION.-Dekkan, India. 492. Probable synonym of speoris. TYPE LoCAL1TY.-Madras, India. MATERIAL.-None. Related to bicolor. dyacorum THOMAS, 1902, Ann. Mag. Nat. MATERIAL.-A specimen from Ma- Hist., (7) IX, p. 291. dras, B.M. 22a (also marked "m.") with TYPE REGION.-Sarawak, Borneo. forearm 38.5; no photo. of skull; also 384 Bulletin American Museum of Natural History [Vol. LXXVIII

material from Burma at U.S.N.M. and Subspecies of larvatus: not improba- from India at M.C.Z. bly a synonym of leptophylla. galeritus CANTOR, 1846, Jour. Asiatic Soc. MATERIAL.-Type in India; series of Bengal, XV, p. 183. three from Siam at U.S.N.M. TYPE LoCALITY.-Penang, Malay Pen- griseus MEYEN, 1833, Nova Acta Ac. Nat. insula. C., XVI, No. 2, p. 608. First-described species of its group, TYPE REGION.-Luzon, Philippine Is- unless crumeniferus, 1807, can be defi- lands. nitely so assigned. Other members are Related to diadema. cervinus, celebensis, insolens, labuanensis. MATERIAL.-Ample material at MATERIAL.-The type: B.M. 79.11.- M.C.Z. and U.S.N.M. 21.85, c (forearm 45.5), with photo. of REMARKS.-A small-toothed form skull. No topotypical material in U.S. (thickness of outer lower incisor .7-.8 Supposed "galeritus" from Borneo, Batu mm.). Tendency for foramina to be- Island, Banka Island at M.C.Z. Sody' come enclosed by premaxillae. Upper believed he had true galeritus in his large incisors much smaller than in nobilis- series from Celebes (later renamed crown width, .7-.8 mm.: .9-1.0 mm. celebensis). We have an equally large guineensis ANDERSEN, 1906, Ann. Mag. series from the same island. The photo- Nat. Hist., (7) XVII, p. 275; 1907, Ann. graph of the type shows an elongate Mus. Civ. Storia Nat., Genova, (3) III, braincase, zygomata not exceeding mas- pp. 17-20. toid width, narrow rostrum, and p2! too TYPE REGION.-Gaboon. indistinct for interpretation. Subspecies of caffer (Allen, 1939). gambiensis ANDERSEN, 1906, Ann. Mag. MATERIAL.-Large series at M.C.Z.; Nat. Hist., (7) XVII, p. 42. few at U.S.N.M. and A.M.N.H. TYPE REGION.-Gambia, West Africa. indus ANDERSEN, 1918, Ann. Mag. Nat. Subspecies of gigas (Allen, 1939). Hist., (9) II, p. 382. MATERIAL.-One from Congo at TYPE REGION.-Kanara, India. M.C.Z. (?) Subspecies of lankadiva. gentilis ANDERSEN, 1918, Ann. Mag. Nat. MATERIAL.-The type: B.M. 12.11.- Hist., (9) II, p. 380. 28.20, 9 (forearm 75), and photo. of TYPE REGION.-Burma. skull. Related to bicolor. insignis HORSFIELD, 1824, Zool. Res. Java. MATERIAL.-Ample specimens from TYPE REGION.-Java. China at A.M.N.H. and M.C.Z. Probable synonym of larvatus. gigas WAGNER, 1845, Archiv f. Naturges., MATERIAL.-Photographs oftwoskulls, XI, pt. 1, p. 148. "co-types": B.M. 60.5.4.16 (in frag- TYPE REGION.-Angola, West Africa. ments, toothrows only); B.M. 79.11.21.- Related to commersonii. 94 (basal part of braincase destroyed). MATERIAL.-Two from S. W. Africa insolens LYON, 1911, Proc. U. S. Nat. Mus., at M.C.Z. XL, p. 129. gracilis PETERS, 1852, Reise nach Moss- TYPE REGION.-S. E. Borneo. ambique, Sauigeth., p. 36. Related to galeritus. TYPE REGION.-Zambesi River, East MATERIAL.-Type and few paratypes Africa. at U.S.N.M. Synonym of caffer (Allen, 1939). This, like labuanensis, is a member of MATERIAL.-Ample. the galeritus group in which p- are rela- grandis G. M. ALLEN, 1936, Records Indian tively reduced. Mus., XXXVIII, pt. 3, p. 345. javanicus SODY, 1937, Temminckia, II, p. TYPE LOCALITY.-Chindwin, Burma. 215. TYPE LoCALITY.-Tjilitjap, C. Java. I 1930, Natuurk. Tijdschr., XC, pp. 267-270. Related to bicolor. 1941 ] Tate, Results of the Archbold Expeditions. 35 385

MATERIAL.-The type: Sody coll. TYPE LOCALITY.-Khasia Hills, As-

No. 3, ad. 6, and photo. of skull. sam. REMARKS.-One of the larger members Related to larvatus. of the group (forearm, 45.5; c-M3, 6.6). MATERIAL.-None. It is probably specifically identical to longicauda PETERS, 1861, Monatsber. "specimen d" of the Jentink catalogue Akad. Wiss. Berlin, p. 708; 1871, op. cit., which was one of Temminck's co-typical pp. 321-322. series of bicolor. TYPE LoCALITY.-Surakarta, Java. labuanensis TOMES, 1858, Proc. Zool. Soc. MATERIAL.-None. London, p. 537. REMARKS.-In all respects mentioned TYPE REGION.-Labuan, Borneo. by Peters, except the extraordinary Related to galeritus. length of the tail, Peters' description of MATERIAL.-The type: B.M. 7.1.1.- this species agrees with the galeritus 305, and photo. of skull (the base of group. It will doubtless prove to be braincase broken). Two "labuanensis" atypical, however. Sodyl believed he from Sumatra at U.S.N.M. had obtained a series of Peters' species, langi J. A. ALLEN, 1917, Bull. Amer. Mus. even though the tails reached only 75 per Nat. Hist., XXXVII, p. 434. cent of the length given for the type TYPE REGION.-N. E. Belgian Congo. longicauda. Synonym of cyclops (Allen, 1939). lylei THOMAS, 1913, Ann. Mag. Nat. Hist., MATERIAL.-The type and series of (8) XII, p. 88. paratypes in A.M.N.H. TYPE REGION.-N. Siam. REMARKS.-Allen described and fig- Related to pratti. ured special structures to which writers MATERIAL.-None. on cyclops had not drawn attention. macrobullatus TATE, described in this KELAART, 1852, Prodr. Fauna paper. lankadiva major ANDERSEN, 1918, Ann. Mag. Nat. Zeylanica, p. 19. TYPE REGION.-Ceylon. (Andersen Hist., (9) II, p. 380. records it from Burma.) TYPE LoCALITY.-Engano Island. of the chiefly Subspecies of gentilis, of bicolor group. The continental offshoot MATERIAL.-None. insular diadema group. marungensis NOACK, 1887, Zool. Jahrb., MATERIAL.-One co-type: B.M. 7.- Syst.,II,p.272. 1.1.311, e (forearm 82), and photo. of TYPE REGION.-West Tanganyika skull. (Collected by Kelaart. Marked Territory. "lectotype.") One specimen from Bom- Subspecies of commersonii (Allen, bay at M.C.Z. Lankadiva, with the con- 1939). tinental forms unitus, indus, mixtus, etc., MATERIAL.-One specimen at M.C.Z. is separable from diadema and allies by masoni DOBSON, 1872, Jour. Asiatic Soc. characters discussed by Andersen (1905). Bengal, p. 338. larvatus HORSFIELD, 1823, Zool. Res. Java, TYPE LocALITY.-Moulmein, Burma. No. 6, P1. ix. Related to diadema. TYPE REGION.-Java. MATERIAL.-None. Earliest known member of its group. megalotis HEUGLIN, 1861, Nova Acta Akad. Other forms are neglectus, leptophilIa. Caes. Leop. Carol., Halle, XXIX, No. 8, Numerous synonyms. p. 4, 8; 1877, "Reise in Nordost Africa," MATERIAL.-The type: B.M. 79.11.- II, pp. 20-21. 21.93 (forearm 53) no photo. of skull; TYPE REGION.-Eritrea. large series from Java at U.S.N.M. and MATERIAL.-None. M.C.Z.; also from Burma, Siam, Kari- REMARKS.-This unique specimen was mata, Natuna, Nias, etc. separated by Peters (1871) under the leptophylla DOBSON, 1874, Jour. Asiatic Soc. subgenus Syndesmotis, on account of the Bengal, LXIII, p. 234. ' Natuurk. Tijdschr., XC, p. 270. 386 Bulletin American Museum of Natural History [Vol. LXXVIII

band described by Heuglin as uniting the vinus], with inner margin convex from ears. Examination of the skin and skull base to apex, . . . terminates in a very will probably show that megalotis is acute point . . ." Of the noseleaf the aberrant in other characters also. authors wrote "the sella has in the cen- micaceus DE WINTON, 1897, Ann. Mag. ter ... a fleshy tubercle . . . the erect pos- Nat. Hist., (6) XX, p. 524. terior process [i.e., transverse leaf] has a TYPE REGION.-Gaboon. small tubercle which projects forward. ." Synonym of cyclops (Allen, 1939). Those are the longitudinally placed, club- micropus PETERS, 1872, Monatsber. Akad. like processes to be seen in the two species Wiss. Berlin, p. 256. named and in cyclops of Africa. "No TYPE LoCALITY.-Dehra Doon, near frontal sac in 9." The illustration of Simla, N. India. the noseleaf is very poor. No descrip- Related to bicolor, probably the north- tion of the skull was published. But the western representative of cineraceus. skull of the 9 co-type mentioned as MATERIAL.-None. having been ceded to the British Mu- mirandus THOMAS, 1914, Ann. Mag. Nat. seum has since been cleaned. It shows Hist., (8) XIII, p. 437. the characteristics of semoni and agrees TYPE LoCALITY.-Manus Island, Ad- with my topotypical series of muscinus: miralty Islands. the parallel toothrows, the greatly en- Related to diadema. larged rostrum, the closely approximated MATERIAL.-The type: B.M. 14.4.1.- cochleae, etc. 8, ad. 9, with photo. of skull. nanus J. A. ALLEN, 1917, Bull. Amer. Mus. mixtus ANDERSEN, 1918, Ann. Mag. Nat. Nat. Hist., XXXVII, p. 434. Hist., (9) II, p. 382. TYPE REGION.-Uele distr., Belgian TYPE REGION.-E. Mysore, India. Congo. Subspecies of lankadiva. Related to caffer. MATERIAL.-The type: B.M. 13.4.- MATERIAL.-The type (in alcohol, 11.19, e (forearm 81.5), with photo. skull cleaned) at A.M.N.H. of skull. natunensis CHASEN, 1940, Bull. Raffles mostellum THOMAS, 1904, Ann. Mag. Nat. Mus., No. 15, p. 43. Hist., (7) XIII, p. 385. TYPE LOCALITY.-Natuna Island. TYPE REGION.-Kenya Colony. Related to diadema. Synonym of marungensis (Allen, 1939). MATERIAL.-None. MATERIAL.-The type: B.M. 89.3.8.- neglectus SODY, 1936, Natuurk. Tijdschr. 3, e (forearm 89.3), no photo. of skull. Ned. Ind., XCVI, p. 46. murinus GRAY, 1838, Mag. Zool. Bot., II, TYPE REGION.-Central Borneo. p. 492. Subspecies of larvatus. TYPE REGION.-Peninsular India, MATERIAL.-Large series from Borneo probably. at U.S.N.M. and one from Borneo at Related to bicolor, probably synonym M.C.Z. of fulvus. nequam ANDERSEN, 1918, Ann. Mag. Nat. muscinus THOMAS AND DORIA, 1886, Ann. Hist., (9) II, p. 381. Mus. Civ. Storia Nat., Genova, (2) IV, TYPE REGION.-Selangor, Malay Pen- p. 203. insula. TYPE LOCALITY.-Fly River, Papua. Related to bicolor. MATERIAL.-One co-type: B.M. 86.- MATERIAL.-Photo. of skull of type 11.3.10, ad. 9 in alcohol, and photo. of only (B.M. 85.8.1.369. Braincase de- skull. One in M.C.Z., from Mt. Misim, stroyed). Dutch New Guinea. Series at A.M.N.H. niangarae J. A. ALLEN, 1917, Bull. Amer. from Fly River and Astrolabe Mountains. Mus. Nat. Hist., XXXVII, p. 438. REMARKS.-There is no doubt that TYPE REGION.-Uele distr., Belgian muscinus is very closely related to semoni Congo. and stenotis. "Ear r larger [than cer- Subspecies of gigas (Allen, 1939). 1941 ] Tate, Results of the Archbold Expeditions. 35 387

MATERIAL.-The type in A.M.N.H. TYPE REGION.-Guadalcanar Island, niapu J. A. ALLEN, 1917, Bull. Amer. Mus. Solomon Islands. Nat. Hist., XXXVII, p. 431. Related to diadema. TYPE REGION.-N. E. Belgian Congo. MATERIAL.-The type, B.M. 88.1.5.23 Subspecies of caffer (Milen, 1939). ad. 9 in alcohol, and photo. of skull; MATERIAL.-The type and small series also one from Isabel Island, Solomons, at A.M.N.H.; one at M.C.Z. at M.C.Z. nicobarensis DOBSON, 1871, Jour. Asiatic pallidus ANDERSEN, 1918, Ann. Mag. Nat. Soc. Bengal, XL, p. 262. Hist., (9) III p. 381. TYPE LOCALITY.-Nicobar Islands. TYPE REGION.-Kathiawar, India. Related to diadema. Subspecies of fulvus, related to bicolor. MATERIAL.-None. MATERIAL.-The type, B.M. 18.8.3.5, nicobarulae MILLER, 1902, Proc. U.S. Nat. ci" (forearm 38); no photo. of skull. Mus., XXIV, p. 781. papua THOMAS AND DORIA, 1886, Ann. TYPE LoCALITY.-Nicobar Islands. Mus. Civ. Storia Nat., Genova, (2) IV, Related to bicolor. p. 204. MATERIAL.-The type and a large se- TYPE LoCALITY.-Misori Island, ries of paratypes at U.S.N.M. Dutch New Guinea. nobilis HORSFIELD, 1823, Zool. Res. Java, MATERIAL.-One co-type examined: No. 6, PI. vii. B.M. 86.11.3.9, ad. 9 in alcohol, and TYPE REGION.-Java. photo. of skull. Related to diadema. REMARKS.-A perplexing species with MATERIAL.-Two "co-types" (skins), rostrum greatly widened (6 mm.) in pro- B.M. 79.11.21.83 (forearm 87), photo. of portion to mastoid width (10 mm.); skull, basal part of braincase destroyed; c-in3 7.6 mm.; while the forearm is only B.M. 79.11.21.84 (forearm 82, no photo. 50.5 mm. Lateral leaflets only two. of skull); also ample series from Java at Pouch absent in female. The rostral M.C.Z. and U.S.N.M. character resembles muscinus but the REMARKS.-Teeth larger than in toothrows are convergent, the ears broad griseus, smaller than material from Bali. and not emarginate, and the bullae un- This is particularly noticeable in the in- enlarged and not approximated. The cisors. Thickness of outer lower in- toothrow exceeds that of any known cisor, .9-1.0 mm. member of the galeritus group by at least obscurus PETERS, 1861, Monatsber. Akad. 1 mm. It even exceeds the toothrow of Wiss. Berlin, p. 709. speoris, which papua resembles in lacking TYPE REGION.-Luzon, Philippine Is- the frontal sac in females. If this species lands. exists in Timor it may be the form which Peron (1807) named crumeniferus. The MATERIAL.-None. width of rostrum is similar REMARKS.-Another difficult to that of species, wollastoni, which, however, has the en- with emarginate ear, no "cells" in trans- larged cochleae and parallel toothrows of verse noseleaf, frontal sac in "( ?), muscinus. forearm 46. Later combined by Peters H. papua may have been derived from with doriae in Cyclorhina ("no lateral near the base of the speoris and galeritus leaflets; transverse leaf with thin mar- (cervinus) stems. gin"). Forearm 46 mm. pelingensis SHAMEL, 1940, J. Mamm., XXI, It may belong with sabanus or dya- No. 3, p. 353. corum, aberrant Bornean members of the TYPE LocALITY.-Peleng Island, east galeritus group. Taylor (1934) believed of Celebes. he had re-collected it. Forearm of Tay- A member of diaderma group, related to lor's animals, 41-44 mm. dinops. oceanitis ANDERSEN, 1905, Ann. Mag. Nat. MATERIAL.-The type at U.S.N.M.; a Hist., (7) XVI, p. 497. series from S. Celebes at A.M.N.H. 388 Bulletin American Museum of Natural History [Vol. LXXVIII penicillatus GRAY, 1838, Mag. Zool. Bot., MATERIAL.-Series in U.S.N.M. from II, p. 493. Luzon. TYPE REGION.-Somewhere in India. REMARKS.-In spite of the very small Probably synonymous with speoris. size of this bat, the forearm, as pointed MATERIAL.-None. out by Lawrence,' is proportionately pendleburyi CHASEN, 1936, Bull. Raffles ForearM= 39 or Mus., XII, p. 133. elongate. Thus, Tibia= j-3 280%, TYPE REGION.-Peninsular Siam. whereas in the smallest member of the MATERIAL.-None. bicolor group, cineraceus, the same lengths "Appearing to differ [from H. a. de- 34 0 bilis] in no essential except size." Fore- are j-, or 260%. arm, 79-79.5 mm.; c-m3, 10.2. reginae TROUGHTON, 1937, Australian Zool., pomona ANDERSEN, 1918, Ann. Mag. Nat. VIII, pt. 4, p. 275. Hist., (9) II, p. 380. TYPE REGION.-N. Queensland. TYPE REGION.-Coorg, India. Related to diadema; synonym of Related to bicolor. pullatus? MATERIAL.-Photo. of type skull only MATERIAL.-One specimen at M.C.Z. (B.M. 18.8.3.4). from Australia. I find no character by poutensis J. A. ALLEN, 1906, Bull. Amer. which this specimen can be separated Mus. Nat. Hist., XXII, p. 483. from my series from the Fly River. TYPE REGION.-Hainan Island. ridleyi ROBINSON AND KLOSS, 1911, J. Fed. Related to larvatus. Malay States Mus., IV, p. 241. MATERIAL.-Type and paratypes in TYPE LOCALITY.-Singapore, Malay A.M.N.H. Region. pratti THOMAS, 1891, Ann. Mag. Nat. Large member of bicolor group. Hist., (6) VII, p. 527. MATERIAL.-None. TYPE REGION.-Szechwan, China. REMARKS.-Forearm, 47.2 mm.; mas- Represents a specialized group to toid width, 10.8; zygomatic width, 9.6; which lylei also belongs. c-m3, 6.8. MATERIAL.-Photo. of type skull only ruber NOACK, 1893, Zoo]. Jahrb., Syst., VII, (B.M. 91.5.11.1, 9); large series at 586. A.M.N.H., U.S.N.M. and M.C.Z. TYPE REGION.-Tanganyika Terr. pulchellus ANDERSEN, 1918, Ann. Mag. Nat. Related to caffer. Hist., (9) II, p. 383. MATERIAL.-Large series at U.S.N.M. TYPE LOCALITY.-Bellary, India. and M.C.Z. Subspecies of speoris. sabanus THOMAS, 1898, Ann. Mag. Nat. MATERIAL.-Photo. of skull of type Hist., (7) I, p. 243. only (B.M. 13.4.10.13). TYPE REGION.-N. Borneo. pullatus ANDERSEN, 1905, Ann. Mag. Nat. MATERIAL.-Photo. of skull only. Hist., (7) XVI, p. 498. REMARKS.-The peculiarities of sa- TYPE LoCALITY.-Haveri, Papua. banus emphasize the need of more col- Related to diadema. lections to assist in establishing the rela- MATERIAL.-Four co-types: B.M. tionships of such anomalous forms. Re- 97.8.7.6-9 (6-8, d; 9, 9 ). Skulls of all duced noseleaf, absence of secondary extracted and cleaned. Photo. of skull leaflets, shortened palate, obsolescence of of 97.8.7.9; series at A.M.N.H. from p2; also (Thomas, 1902, Ann. Mag. Nat. Papua. Hist., (7) IX, p. 272) a notched front pygmaeus WATERHOUSE, 1843, Proc. Zool. edge of the horseshoe and the absence of Soc. London, p. 67. vertical septa in the posterior noseleaf. TYPE REGION.-Philippine Islands. Thomas regarded it as a member of the An isolated species probably derived bicolor group, probably because it lacks from the line leading to the galeritus group. 1 1939, Bull. Mus. Comp. Zool., LXXXVI, p. 57. 19411] Tate, Results of the Archbold Expeditions. 35 389

lateral leaflets, but the relationship of cludes stenotis, whose nearest relative ap- zygomatic width to mastoid width and pears to be cyclops of Central Africa. the broadening of the zygomata at their CHARACTERIsTIcs.-Horseshoe with maxillary roots are more like that of the two club-like processes, one behind the galeritus group than the bicolor group. other; elongate, acute, pointed ears; The ears, at least in dyacorum, are emar- toothrows parallel; p- markedly reduced; ginate and haired for most of their cochleae much enlarged and very close length, wholly unlike bicolor ears. together. See also stenotis, muscinus, saevus ANDERSEN, 1918, Ann. Mag. Nat. papua and wollastoni. Hist., (9) II, p. 380. sinensis ANDERSEN, 1918, Ann. Mag. Nat. TYPE LOCALITY.-Kei Islands. Hist., (9) II, p. 380. Related to bicolor. TYPE REGION.-Fukien. MATERIAL.-None. Subspecies of gentilis, member of bi- REMARKS.-The bicolor bats of Queens- color group. land, Aru Island and South New Guinea MATERIAL.-The type: B.M. 92.2.1.- (forearm 37-39) are definitely smaller 3 (forearm 38.5), and photo. of skull than those of Ceram, Celebes, Halma- (braincase destroyed). Many specimens hera, and also Kei (forearm 40-42). from China at U.S.N.M., A.M.N.H. and This difference appears also in the teeth: M.C.Z. c-m3, 5.1-5.3 mm. and 5.6-6.1, respec- REMARKS.-Gentilis itself would seem tively. The former are referable to to be conspecific withfulvus of Peninsular aruensis, the latter to saevus (= toala). India. schistaceus ANDERSEN, 1918, Ann. Mag. speculator ANDERSEN, 1918, Ann. Mag. Nat. Hist., (9) II, p. 382. Nat. Hist., (9) II, p. 381. TYPE REGION.-Bellary, India. TYPE REGION.-S. Celebes. MATERIAL.-The type: B.M. 13.4.- Related to diadema. 10.3, e, and photo. of skull. MATERIAL.-Photo. of skull of type According to Andersen, an offshoot of only; series at A.M.N.H. from Celebes. lankadiva. speoris SCHNEIDER, in Schreber's Sauige- schneideri THOMAS, 1904, Zool. Anz., thiere, V, and plate. XXVII, p. 722. RANGE.-India and Ceylon.2 TYPE REGION.-Sumatra. GENOTYPE.-Nearest allied species are MATERIAL.-Photo. of skull only. abae in Africa and larvatus of the Burma- REMARKS.-The picture of the skull of Sunda area. the type indicates close relationship to MATERIAL.-Series from Ceylon at the galeritus group; zygomatic width ex- M.C.Z.; 3 from India at U.S.N.M. ceeds mastoid width; palate and pre- REMARKS.-Characters given else- maxillae rather short. p2 is absent and where, under discussion of the genotype P2 much reduced. The ear is emarginate of the genus Hipposideros. and frontal gland well developed. Fur- stenotis THOMAS, 1913, Ann. Mag. Nat. ther details were given by Andersen1 in Hist., (8) XII, p. 206. 1907. Forearm, 48 mm.; c-m3, 6.7. TYPE LoCALITY.-Mary River, North- semoni MATSCHIE, 1903, Zool. Forschungen, ern Territory, North Australia. Related Austral. u. Malay. Archip., V, Art. 6, p. to muscinus. 774. MATERIAL.-The type: B.M. 97.4.- TYPE LOCALITY.-Cooktown, Queens- 12.7, 9 in alcohol, and photo. of skull; land. series from Papua in A.M.N.H., 3 from MATERIAL.-None. North Australia at U.S.N.M. REMARKS.-Member of muscinus, a REMARKS.-Closely allied, if not equal, specialized Torresian group, which in- to semoni. swinhoei PETERS, 1870, Proc. Zool. Soc. 1 Ann. Mus. Civ. Storia Nat., Genova, (3) III, pp. London, p. 616. 1-22. 2 Various authors cite it from Java and Borneo. 390 Bulletin American Museum of Natural History [VOl. LXXVIII

TYPE LOCALITY.-Amoy, China. REMARKS.-The describer of toala Subspecies of armiger. compared it with gentilis and sinensis MATERIAL.-Photo. of type-skull; one rather than with the geographically (and topotype (from Amoy) at M.C.Z.; others genetically) nearer aruensis, saevus, etc. from China at U.S.N.M., M.C.Z. and We have other specimens from Celebes A.M.N.H. which I believe to be closest to saevus. taitiensis FITZINGER, 1869, Sitzungsber. trobrius TROUGHTON, 1937, Australian Akad. Wiss. Wien, LX, p. 865. Zool., VIII, pt. 4, p. 276. TYPE LoCALITY.-Tahiti ? TYPE LoCALITY.-Trobriand Islands. MATERIAL.-None. Related to diadema. REMARKS.-Taitiensis appears to be MATERIAL.-None. referable either to the speoris group or to turpis BANGS, 1901, Amer. Nat., XXXV, p. the galeritus group. Speoris, otherwise 561. restricted to Peninsular India and Cey- TYPE REGION.-Liu Kiu Islands. lon, agrees closely with Fitzinger's de- Related to armiger. scription of 3 lateral leaflets subtending MATERIAL.-The type and paratypes the horseshoe in taitiensis. On the other at M.C.Z.; one topotype at A.M.N.H. hand cervinus, of the galeritus group, with REMARKS.-Turpis, like pendleburyi, but two leaflets, extends at least as far seems to differ from armiger, swinhoei, eastwards as New Hebrides Islands. debilis and terasensis chiefly by its smaller Dobson (P.Z.S., 1877, p. 121) wrote, "we size. are not absolutely certain that the speci- unitus ANDERSEN, 1918, Ann. Mag. Nat. men ... was really obtained at the Island Hist., (9) II, p. 382. of Tahiti." TYPE REGION.-Saugor, C. P., India. templetoni KELAART, 1849-50, Jour. Ceylon Subspecies of lankadiva. Branch, Royal Asiatic Soc., (2) II, p. 322. MATERIAL.-The type: B.M. 12.11.- TYPE REGION.-Ceylon. 29.20, Q (forearm 87.5), and photo. of Probable synonym of speoris. skull. MATERIAL.-Two co-types: B.M. vicarius ANDERSEN, 1905, Ann. Mag. Nat. 52.1.24.2 and 4, and photos. of skulls Hist., (7) XVI, p. 499. (base of braincase in each destroyed). TYPE REGION.-Sarawak. terasensis KISHIDA, 1924, Zool. Mag. Related to diadema. Tokyo, XXXVI, p. 42. MATERIAL.-Photo. of skull of type TYPE REGION.-Formosa. only; two from Sarawak at M.C.Z. Subspecies of armiger. These disagree with griseus, and agree MATERIAL.-One specimen from For- with nobilis in the size of the incisors: mosa at U.S.N.M. thickness of i2, .9 mm., width of crown of REMARKS.-This form and swinhoei i', 1.0 mm. Mastoid width : zygomatic seem to be indistinguishable from armi- width = 15:19. Enclosure of foramina ger. by premaxillae incipient. thomensis BOCAGE, 1891, J. Sci. Math. vittata PETERS, 1852, Reise nach Mossam- Phys. Nat. Lisboa, (2) II, p. 88; 1904 (7) bique, Sauigeth., p. 32. VI, p. 67. TYPE LoCALITY.-Ibo Island, Cap TYPE LOCALITY.-San Thom6 Island. Delgado Group, East Africa. Subspecies of commersonii (Allen, Synonym of gigas (Allen, 1939). 1939). MATERIAL.-Three from Zanzibar at MATERIAL.-None. M.C.Z. toala SHAMEL, 1940, J. Mamm., XXI, No. vulgaris HORSFIELD, 1824, Zool. Res. Java. 3, p. 352. TYPE REGION.-Java. TYPE LoCALITY.-Toeare, Celebes. Probable synonym of larvatus. "Subsp. of gentilis," member of bicolor MATERIAL.-The type: B.M. 79.11.- group. Probable synonym of saevus. 21.575 (forearm 56), and photo. of skull MATERIAL.-The type at U.S.N.M. (back of braincase broken). 1941 ] Tate, Results of the Archbold Expeditions. 35 391 woUlastoni THOMAS, 1913, Ann. Mag. Nat. primary crest, is serrate. But in wol- Hist., (8) XII, p. 205. lastoni the two are separated by the in- TYPE LoCALITY.-Utakwa River, S. terrupted groove described by Thomas. W. New Guinea. The skull resembles that of muscinus in MATERIAL.-The type: B.M. 13.6.- all essential characters: greatly en- 18.4, ad. 9 in alcohol, and photo. of larged rostrum and cochleae, parallel skull. toothrows, etc. REMARKS.-Thomas compared this wrighti TAYLOR, 1934, Philippine Bur. Sci., bat with muscinus but not with semoni Monogr. 30, p. 237. and stenotis, to which it is obviously also TYPE LoCALITY.-Baguio, Benguet, closely related. It appears to reach the Philippine Islands. very apex of specialization in this little Related to bicolor. group of Torresian bats by the develop- MATERIAL.-None. ment of the remarkable secondary trans- REMARKS.-If we accept Lawrence's' verse leaf behind the primary transverse interpretation, Taylor has redescribed leaf. In the African caffer (mentioned antricola Peters. See also antricola and by Thomas) the secondary development erigens. arises directly from the base of the pri- l 1939, Bull. Mus. Comp. Zool., LXXXVI, pp. 54- mary one. Its crest, lower than the .56. HIPPOSIDEROS MATERIAL IN THE ARCHBOLD COLLECTIONS Skin and Alcoholic Skull Hipposideros larvatus Horsfield 6 N. W. Borneo: Perboewa (Landak) 6 Sumatra: Bukit Panggoel (Bankoda) 3 South Sumatra: Macarah Doewa (Palembang) 18 Mentawi Islands: N. Pagi 4 Hipposideros galeritus labuanensis Tomes N. W. Borneo: Perboewa (Landak) 21 Hipposideros galeritus galeritus Cantor (?) Mentawi Islands: N. Pagi 8 Hipposideros galeritus celebensis Sody South Celebes: Talassa (Maros) 29 is Is : Banti-moerang 3 Hipposideros galeritus cervinus Gould Dutch New Guinea: Hollandia 20 25 "6 "6 It : Weyland Mountains 2 Papua: Oriomo River 72 7 Hipposideros insolens Lyon N. W. Borneo: Perboewa 7 Hipposideros schneideri Thomas Mentawi Islands: N. Pagi 7 Hipposideros dyacorum Thomas N. W. Borneo: Perboewa (Landak) 9 Hipposideros breviceps Tate Mentawi Island: N. Pagi 38 Hipposideros diadema pelingensis Shamel South Celebes: Talassa (Maros) 17 Hipposideros diadema nobilis Horsfield Java: Cheribon 1 Bali: Oboed 3 Bali: N. W. Nossa Tenida 5 Hipposideros diadema euotis Andersen N. Celebes: Koemersot 3 S. Celebes: Talassa (Maros) 3 Hipposideros diadema pullatus Andersen Papua: Upper Fly River 4 12 Hipposideros bicolor major Andersen Mentawi Islands: N. Pagi 3 392 Bulletin American Museum of Natural History [Vol. LXXVIII

Hipposideros bicolor aruensis Gray Papua: Fly River 9 " : Wassi Kussa River 6 Dutch New Guinea: Idenburg River 1 9 S. of Geelvink Bay 1 Hipposideros bicolor aruen8is, yellow phase Bali: Saugsit 9 Hipposideros bicolor saevus Andersen Celebes: Peleng Island 6 2 " : Roeroekan 1 Halmahera 1 Hipposideros cupidus Papua: Javereri, Musgrave River 4 Dutch New Guinea: Hollandia 5 Hipposideros calcaratus Papua: Javereri, Musgrave River 7 " : Middle Fly River 3 6 " : Upper Fly River (Red phase) 4 Dutch New Guinea: Hollandia 1 1 Hipposideros muscinus muscinus Papua: Upper Fly River 1 2 " : Middle Fly River 3 " : Wassi Kussa 2 River 2 2 Sogeri, East of Port Moresby 6

HIPPOSIDEROS OF THE NEW GUINEA AREA New Guinea and nearby territories-the different times. I collected both species in Solomon Islands and tropical Australia- the limestone caverns at Javereri, upper contain representatives of three widely dis- Kemp Welch River basin, but in different tributed groups, namely, bicolor, galeritus caves. Specimens of calcaratus were se- and diadema. In addition they are the cured in hollow trees along the middle Fly home of two autochthonous groups, mus- River and four bats, representing the cinus and calcaratus. The former appears bright rufous stage of the same species, to be distantly related to cyclops of Central were shot and brought into camp by the Africa. The latter may possibly be allied bushmen at Palmer River. At Hollandia, to coronatus of the Philippine Islands. on the north coast of Dutch New Guinea, Two anomalous species are wollastoni, several cupidus and a single calcaratus were apparently a specialized offshoot of the collected. Both species are present at muscinus group, and papua (of which we Duke of York Island, the type locality of have no specimens) springing possibly from calcaratus. Our bats from the Tabar the basal stem from which the speoris and Islands and U.S.N.M. 18480 from Jobi galeritus groups are derived. are all referable to cupidus. H. bicolor is generally found in Southern H. galeritus cervinus is in the same case as Papua in hollow trees growing in the aruensis, a widely dispersed form showing patches of gallery woods that margin the occasional slight geographical variations. streams of the plains country. But it also The bats found in the caverns at Dogwa, occurs in densely forested areas. The Oriomo River, include two color phases, red forms that have received names are and gray, and are slightly larger than amboinensis (Amboina), saevus (Kei), aru- specimens from Hollandia and Weyland ensis (Aru) and albanensis (Cape York, N. Mountains, but these differences appear to Queensland). They are almost indis- be too inconstant to merit a distinct sub- tinguishable, differing only very slightly in specific name. size. The oldest name is aruensis. Galeritus bats of the region already H. calcaratus and H. cupidus, related but named are cervinus (Cape York, N. Queens- perfectly distinct members of the same land), batchianus (Batjan) and crumen- group, have seemingly been confused at iferus (Timor), the last still uncertainly 19411] Tate, Results of the Archbold Expeditions. 35 393 identifiable, so that the name employed pro- of sticks, whipped rapidly to and fro across visionally should be cervinus. The species fly-ways. A specimen (M.C.Z. 29927) is widely distributed, reaching 4000 feet was taken by Stevens at Morobe, Mt. (Shaw-Meyer in the Weyland Mountains). Misim, 5700 feet. H. diadema has had ten forms named in H. wollastoni must be left provisionally as the New Guinea region. It seems to be a a distinct species belonging to the muscinus highly plastic species. A series brought in group. The skull, judging from its photo- by natives of the upper Fly River is re- graph, is identical to those of muscinus but ferred to pullatus. It is open to question Thomas's description of the "double" whether true diadema of Timor, custos, transverse leaf is so precise that one cannot pullatus, oceanitis, reginae, speculator should doubt an important difference exists. not be synonymized. Euotis, rather larger, Wollastoni is known only from the type and demissus and mirandus, much smaller, from the Utakwa River, Dutch New may be separable. To determine the mat- Guinea. ter a much larger series than I have with H. papua, another dubious species, must topotypes is required. Dinops is very likewise be left undisturbed. The greater much larger and has the rostrum lower and size of skull, teeth and forearm precludes flatter, and the zygomata more expanded. any very close relationship to the smaller H. muscinus, whose very near relatives Papuan species, while, on the other hand, are semoni and stenotis (both perhaps syn- it is much smaller than any variety of dia- onymous), was found along the middle and demra. Unlike wollastoni it is founded upon upper Fly River in hollow trees. Not two specimens, both from Misori Island, of many specimens, however, were taken, the which the female was sent to the British bats appearing to be solitary. They were Museum and the male was kept at Genoa. collected again at Baruari on the Astrolabe Our photograph shows the skull of the Range, Central Division, Papua, by means female.

'N'~'7~k>' '~ ;' V'N " C ( N (A NN N 'N Aq~~~~~~~~~~~~~~~~~~~~~N N'~~~~~N-N)N A~~~~~~~~~ 'N'NNN4 K~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ k~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~j 'N'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~3

4N'N NJJA''Nt' N 'N "V~~~~~~~~~~~~~~~4 JN-

zi4~~~~~~~~~~N N N' "NNA N(NN4 ~ ~'~iNjpN. 'N 4"N' N.N N N

'Q '" NN' A C ~ 'NN NN'NN '2N NN\ NNAJ

N 'N NN "4 ""! ~ 'N "4

"N 'N~4~A ~~~~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ 'Nf''ANN''NN'NN A NA . NN A V 0~~~~~JJ~N N ¾ '' '~~~~N'~~~~~N'NNNN)N'/ K ~~~~~~~~~~~~~~~~~~~~ N 'N N'~~~~~~~i N'N~~~~~~NN'4, ~~~~~~~~~ 4' 4'~~~~~~~~~ 'N "'N NAN "'4 'N~~~~~~~~~~~~~~~~ N~ FNA'~~~~~~ ''N~~~J~ ~~N / '''""' 'NN 'NA (A N'N N'N N NNN4N

'N'$~~~~~N ")-1N')Nt 4NN NN'" N" N'' N~~~~~~~~~~~~~~~V

N)'A',N 4'N N N''' NN'~~~~~~ NN'~' 'N'A'N"NN' 'N NNNN 'N~~~'N 'N~~~~' NAN N~NN~~~~~ " JNfA J ' N' 'N a~~~~~~~~~~~~~~~~~~~'''N'~ 4V~~N 'N\N NN ''NN'N' ' ~ ' 'N''N 4~ '' 'N 'N '' ' N" (.A'~N 4N,A"

'N 4 ,'N' N) " ''

4NNN N) 4AN~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~41 ~ .NJA'4N4~~~~~~~~~~~~~~~'NA ~ ~ ANN'N?'N'N' '/~~~~~"'N > -~~~~~(~~~~~~~~~~~~~~~~~~131

4/42~~~~~~~~~~~~~J14

V.~~~~~~~~~~~~~~~~~~~~- 74,~~~~~~~~~~~~~4

1. ~~~~~~~ ~~~r L XI

-~>IV> *

w-~ ~ ~ k -

`~~~~~~~~~~~~~~4 O

kt-~~~~ ~ ~ ~ ~ ''/4'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~//~~~~~~~~~~~~~~~~~~~~7 Y, ~ 2KFIy~4 V4' ''1 F~ ~~~~~~~~~~~ ~~~~~~~'- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.4 A,)~~~7F4'V

'N~~~~~~~~~~~~~~~~~~~~~-W'~~~~~~~~~~~~~~~~~4 iA~~~~~~~'I " Y,J

4v~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'~~~~~~~~~~~~~~~~

2 1

P -V~~~~~~~~~~~~~~~~~~~~~~'

A~~~~~~~~N

F 4,V- 2- '~~~~~~~~~~~~~~~~~~~~~~~~~ F,~,

A I'M~~~~~~~~~~~~~~~~~~~~~~~~~~~~~A

4'~~~~~~~~~~~~~~~~~~~~~~~4