Behavior as a taxonomic clue: Relationships of Lissonycteris (Chiroptera)

BREVIORA Miiseiiiim of Comparative Zoology Cambridgk. Mass. Ai'Kil 18, IWA Numher 184 BEHAVIOR AS ATAXONOMIC CLl'E :KELATIONSIILPS OF LIS.SONYCTERIS (CHIROPTEKA) By Barbara Lawrence MiLseuni uf Cunipiiijitive Zoology, Harvard University and Alvin Novick^ Osborn Memorial Lalioratories and Peabody Museum of Natural Historv, Yale University Re-examination of the ^enerie status and i-elationships of Ro}(scffus, sensu stricfo, and Lissonyctrris, asupposed su])g'enus of Rouscttus, has been prompted by the great ditferenees observed between live specimens of the two genera. In 1955 and 1956, Novick was able to make extensive observations of living bats not only of these two genera but of six other pterojuds as well (Xovick, 1958). Of these, according to Andersen (1912), Eidolon helrioii and Ptcropus g. (jigantcus as well as Rouscttus belong in the Rouscttus section of the Pteropinae, Cynoptcrus brachyotis luzonicnsis, Cynoptcrus sphinx ccylonensis. and Ptcnochirus joyorii belong in the Cynoptcrus section of the Pteropinae, while Eonyctcris spclaca glandifcra, Eonyctcris rob}istn, and Macroglossus J. lagochilus are in tlie subfamily Macroglossinae. BEHAVIORAL COMPARISONS Asingle wild colony of about twenty Lissonyctcris angolcnsis was observed in the Belgian Congo and various wild colonies of Rouscttus amplcxicaudatus, R. scnii)iudus, and R. acgyptiacus were observed in the Philippines, Ceylon, and the Belgian Congo, respectively. Two Lissonyctcris angolcnsis were captured 1Fornicrly ;it tin' lliolojiical lialxtrMturics, Ilarvaril I'nivtTsit.v. 2BREVIORA No. 184 and one, amatnre female, lived in eaptivity for fifteen months. Many R. acgyptian(.s were also taken alive and observed for a period of some weeks in the Belgian Congo. One, amature male, was kept for over nine months concurrently with the Lissonycteris mentioned above. In addition, some six K. se)nini(di(s and adozen K. amplcricaudafus were observed in captivity for periods of six weeks to more than three months. From the very first, the differences in behavior of Lissonycteris and RoHsettus were grossly apparent. These can best be described as differences in limb use and in orientation. Limb Use Rousettiis ordinarily occupy diml}- lighted caves with large entrances and sheltered retreats. Here they commonly hang in large clusters along the walls or ceiling. On the wall, they hang by their hind feet with their legs so turned that their backs are to the wall, their wings folded at their sides. They take flight by swinging their body up and rotating around their feet Avhich remain temporarily in position here though the legs are now crossed. This position is well illustrated by Kulzer (1958), in an excellent paper on the biology of Roust ft us. Roxsctfus also roost in crevices. Here one cannot see their undisturbed resting position but they must use all four limbs in entering and leaving such crevices. Rousettiis can fold their wings considerably and walk awkwardly on their wrists and hind feet or they can climb vertically, head first, or move along the irregular surface of a cave or crevice ceiling using the claws of their thumbs and hind feet. When Rouseftus roost in trees, as they do when disturbed in their caves or at night when feeding, they almost always hang by both hind feet plus one or both thumbs in asloth-like position, one foot and one thumb on each side of the branch and their wings folded. They frecpiently move along the undersid(> of the branch in this position using all four limbs. Thus, wild Rousettus use their wrists for walking and their thumbs for climbing, roosting, and moving along the underside of branches or fruits. Captive Rouseftus would roost in all of the positions desr-ribed above. When they were exposed to light, they would fre(ni(Mitly crawl into acorner of the cage floor or into their feeding disli where they would ])r('ss tliciuselves as closely as possible against theiisui-roundings, avoiding th(> light on Iheir eyes. Crawling is anot uncommon mode of locomotion for Rousettus. whcthei' 1[)68 RELATIOXSIIIPS OK LISSON YCTERIS 3 rcstrictod to acajic oi' allowed to move freely about avooiii. While feedinji', foiexample, tliey would crawl between the liands of astem of bananas. In addition, they oeeasionally used their wrists and tliumbs as pushers to reailjust morsels of fruit in their mouths and R. avgyptiacus once was seen using the claws of one hind foot to manipulate the food in its mouth. Lissonyctcris never use their win^s for locomotion other than flight. Tn the wild, and in captivity in alarge flight room, IAss<t)nfrf< ris always roosted hanging free from the ceiling or abranch or, occasionally, hanging from an irregularity of the wall, their ventral surface out as with Rousettns, but never using their wrists or thumbs for support and never entering a crevice. Liss(mycteris would hang from abranch or the ceiling by the claws of one or both hind feet. If they used both feet, characteristically both would be on the same side of the branch or irregularity. Lissouycicris seemed incapable of folding their wings tightly as Rouseffiis do; when roosting they hold their wings onlv slightlv folded enveloping their ventral surface (see Fig. 5). " Lissonifcfen's were incajjable (»f walking or climbing since both activities recjuire that the wrists or thumbs be used. They moved along their roost only by releasing their hold with one foot, moving this foot along to anew hold and then following with the other foot. Such movements, always slow and inefficient, were suitable only foi' slightly changing their roosting position. Lissonyctcris never landed on ahorizontal surface liy choice and, when forced to do so to take its fruit from adish, its movements were awkward and incompetent. After picking up a piece of fruit, it would flap otf without crawling. Lissonyctcris regularly use tlieiihind feet, but rarely their wrists or thumbs, for handling their food. When one Avas handed apiece of banana, it would grasj) this with its teeth. Then, having readjusted its roosting foothold, it would bring one foot dowai (either one with apparently e(|ual facility) to its mouth and hold the piece of fruit with its widely spread toes while biting or breaking off amanageable fragment (see Fig. 1). Then, while chewing this fragment, it would hold the bulk of the piece of fruit, with its toes, against its chest or abdomen, frequently partly or completely covered by its wings (see Fig. 4). IJarely, when through carelessness or fragmentation it was about to drop the piece of fruit Avhich it was chewing, the bat would regain its tooth-hold by using its wrists and thumbs as awkward pushers. Thus Rousettus, sensu stricto, differs from Lissonyctcris in its 4BREVIORA No. 184 use of wrists, tlnimbs, and hind feet in roostinp', (•limbin<r, moving along branches, walking, and handling its food. Among the other pteropids, Eidulon and Pfcropiis, though they never enter caves, roost in trees in much the same posture as Ronseitiis, holding on with both hind feet as Avell as with one or both thumbs. Eoiiycfcris, in caves and in captivity, roost and move in all ways like Roiisdfus. In all of these genera the wings can be considerably folded ;some also commonly roost Avith their wings enveloping their ventral surface. Neither Eidolon nor Pieropus can walk well on horizontal surfaces and both would probably be incapable of entering cave crevices as do Poiisdfus and Eonjfcfrris. Pteropiis, at least, use their Avrists and thumbs skillfully to handle fruit and to orient the morsels in their mouth. Cynopfcrns, Ptcnochirus and, to acertain extent, Macroglossus also use their limbs very much as does Lissonyctcris. All three hang from above, by their hind feet only, though, unlike Lissor}yctcris, they appear to use their thumbs for climbing up walls or along branches. The wings of Cyuopirrus and Macro(/lostsus do not seem to fold in life any more tightly than those of Lissonycteris. Judging from pictures of Epomophorus (Allen, Lang, and Chapin, 1917, pi. 46) and the position of Mifonyctcris' wings in alcoholic specimens, their posture in life is also like that of Lissonyctcris. Lissonyctcris (as also ('ynopt< rus and Ptcnochirus) tended to hold large amounts of food in their cheeks and to carry food to their roost, there to chew very slowly. AVhen recpiired to fly while eating, they would sometimes drop areally large piece. If Lissonyctcris Avas holding apiece of fruit with its hind foot, it would either transfer this to its mouth for flight or drop it. Rouscttus never seemed to store fruit in its cheeks or to fly with any in its mouth but, like Eonyctcris, Eidolon, and Ptcropus, generally stayed at the food source, biting and woi-ryinu off fragments and chewing these at once. Lissonyctcris would chew and crush fruit, such as melon and pineapple, until all of the juice had lieen expressed and swallowed and then push the remaining small bolus of fiber out of its mouth with its tongue. Banana, being non-fibrous, was swallowed completely after chewing. Rousettus usually ate banana and. like Lissonyctcris, swallowed the whole fruit hut when Rousettus ate mango, ])apaya, or melon, uidike LisstDiycto'is. they seemed to swallow the whole substance. ^\)C)^^ RELATIOXSIIIPS OF LISSOX Vf "PKKIS 5 ()Kli:.\TA'l'I()N jAssonf/ctfris and Noiisrffus differ strikiiij>ly in their orientation. Lisson!irt( ris oi-ient entirely by vision (Novielv, 1958). The captive hat was helph'ss in the dark and refused to fly in the dark oiwhen hlindfoUh'd. When it was forced to fly witliout vision. I)y Ijein""' tlirown into the air. it always crashed into the first ol)staele it met. Roiisdfus anip1( .ricanddl IIS, R. sciniiiiidiis, and R. argyptiacus all orient visually and acoustically (Mohres and Kulzer, 1956; Kulzer, 1956, 1958; Novick, 1958; Griffin, Noviek, and Korntield, 1958). The sin^ile R. acgyptiacus ol)served for over nine months in captivity oriented largely visually in strong light but oriented acoustically in the dark, in dim light, when avoiding intricate obstacles, and when landing or taking off. Such acoustic orientation obviously demands specializations for sound ])roduction, emission, reception, and interpretation. The brain, in particular, must be highly specialized for handling acoustic information. In the pteropids, acoustic orientation has so far been found oidy in Rousettus, scusii stricto, and not in Eidolon, Pteropus, Cynoptcrus, PtcHocJiiriis, Eonycteris, Macroglossus, or Lissonycteris (Novick, 1958). The orientation of Ste^wnycteris, a subgenus of Roiiscttiis, has not been observed. The distinctness of Rouscitus, not only from Lissonycteris but from afair sample of other pteropids as well, is clearly establisiied by its acoustic orientation and the behavioral, physiological, and anatomical features associated with it. Summary To summarize, Lissoiiycfi ris differ from Rouscitus, scnsu stricto, in their roosting posture, in their non-flight locomotion, and in their iiuU)ility to orient acoustically. Comparing the roosting and locomotory behavior of those bats observed alive, we find that Rouscttus resembles Eidolon, Pteropus, and Eonycteris, while Lissonycteris resembles Cynopterus most closely. Brief ol)servati()n of live epomophorine bats and examination of pictures and preserved specimens of epomophorine bats and of Myonycteris suggest that in behavioiLisso7iycteris resembles these bats as well. 6BREVIORA No. 184 MORlMIOLOdiCAL C0.MPAK180NS The behavioral differenees between tlie live bats are reinforced by less speetaeular, l)iit equally definite, characters of the more conventional taxononiic sort. These make it clear that Lissoin/rteris is far closer to Myonycteris, asupposed intermediate genus between the rousettine and cynopterine "roups, than to Rousettus. They also show that Lissonycteris and Myonycteris form anatural gi-oup probably intermediate between the rousettine and epomophorine groups, and one whose resemblances to the cynopterines are more apparent than real. The rousettine and cynopterine sections of the Pteropinae are chiefly distinguished by the following characters possessed by the latter and not by the former: The rostrum is shortened and the facial axis is not deflected, Ms and usually Mare lost, the eyes are larger, and there is atendency to form tubular nostrils. The differences in the ridges of the soft palate described by Andersen (1912, pp. Ix, 485, 591, figs. 29A, 50) are also characteristic. On the basis, largely, of these characters, Andersen points out that Myonycteris is somewhat intermediate between the cynopterine and rousettine sections. He says (1912, pp. Ivi-lvii) : "Myonycteris . . . has in many respects remained on the Rousettine level of development, while in others it exhibits modifications approaching those of Cynopterus. The general appearance, the dental formula, and the palate ridges are qviite or nearly as in Botisettus, but the rostrum is conspicuously shortened, the facial axis less deflected, mn and m- (last lower and upper molar) reduced almost to rudiments, the orbits larger, the nostrils more prominent and the calcar weaker," and later (p. Ixi) :"The fact is that this genus has retained many characters of Rousettus, while in jn-actically all the features in which it differs from Rousdtus it more or less closely approaches to Cynopterus. Whether agenus exhibiting characters of this description ought, in alinear arrangement, to be classed at or near the end of the Rousettine section oi' as the opening' genus of the Cynopterine section, must necessarily remain amatter of opinion." The resemblance of Lisso)iycieris to Myonycteris in all the above traits, except the shortening of the rosti'uiii and i-eduction of the dentition, was not noted by Andersen when he classified tile foniici-asan ahcrrjint form related to but {)ossibly generically distinct from /t<nis(ltHs {(ip. r//., 1!I12. p. 814). At tlie same time he does note (1912, p. xlix ):"the In-ain case [Lissonycteris] is |)eculiarly flattened postei'ioi-ly and tlie facial axis even less 1!)6."{ RELATIONSHIPS OF LISSONYCTERIS 7 deflected tlian in Rouscttus, both eliaraeters giving the skull, vi(»we(l in pi-otih', ai-itlieistriking i-esemblanee to Ejjoinops." External Characters Aetnally, in addition to the ratiier generalized niyonyeterine characters given above, both Lissonijcteris and Mijuinjcteris have anumber of traits in common which the small size of Myonycteris and the superficial i-esemblance of Lissonycteris to Rousettus have tended to mask. Chief among these, externally, is the development of the wing in these two genera Avhich greatly exceeds that of Rousettus. Though it is more similar in size to that in Cynoptcnis, it is ditt'erent in proportions. The size ditference shows most conspicuously in the greater length of the combined metacarpal and first phalanx of digits three to five as compared with length of forearm. The following comparison of finger lengths refers only to these two joints. In Lissonycteris, the fifth finger is conspicuously longer than the forearm; in Myonycteris, the slightly smaller wing has digit five subequal in length to the forearm. In Rousettus, the fingers are much shorter; usually only digit three has the upper joints as long as the forearm ; occasionally three is somewhat longer and four is subequal to the forearm. In all three genera, these two joints of digit five are shortest though the relative lengths vary specifically. Cynopterus, while longer fingered than Rousettus, is less extreme than the other two genera and differently proportioned. The metacarpal and first phalanx of digits four and five are subequal in length, with four being slightly shorter than five which is about the same length as the forearm, while digit three is conspicuously longer. Other differences in ratios of wing bones can be worked out, but the most significant fact is that in development of the wing Lissonycteris and Myonycteris are not intermediate between Rousettus and Cynopterus but differ equally from both. The same is true of the attachment of the wing wiiich in the two former genera is near the middle of the first phalanx of the second toe, in Cynopterus is near the distal end of the first phalanx of the first toe, and in Rousettus is usually between metatarsals one and two, sometimes near the bases of the first phalanges, often well proximal to this. Lissonycteris and Myonycteris are also equally distinct from tlie other two genera in shape and arrangement of the odontoid papillae. These papillae, which border the lips, inside, from the 8BREVIORA No. 184 angle of the mouth forward, are rather liigh and pointed and, in Lissontjctcris, form asingle row extending about to the eanines. In Myonycferis, the shape and arrangement of the papillae are about the same, but in some instances on the upper lip toward the angle of the mouth there is apoorly defined second row. In RoHscttKS, the single row of small papillae is much reduced in extent and size. In Cynoptcnis, the more numerous, larger I)apillae are irregularly arranged in abroad band two or three rows wide and somewhat better developed on the upper than the lower lip. The arrangement of the palatal ridges in Lissonijcten's is very different from that in Cynopicnis and, while more like that in RoiiscttKs, ditt'ers in certain details which again point up Lissonycfcris' close resemblance to Myonycieris. In the two latter, the three anteriormost ridges are not divided and scarcely, if at all, bowed forward; four and five are divided in the middle, slojic forward (as in Rousetfits), and have the inner ends recurved ;six and seven, which lie behind the tooth row, are similarly sliaped and these last four converge somewhat at their medial ends. Behind this, there are two or three other poorly defined ridges. Rousettus differs in general in having the anterior ridges more bowed forward, number four usually not divided, and the posterior ridges more nearly parallel, with each other. Other external characters which set Lissonycteris and Myonycferis apart from Ro)isettns are the shorter, less robust tail, the longer, denser fur on the notopatagium and on the proximal, dorsal surface of the til)ia, the smaller foot with the basal quarter to third of the first j)lialanges wel)bed (see Pig. 1), the more slendeiclaws, the moi'e delicate calcar, and the extensive patch of ulandular fur on the throat of adult males. o' Hair Stri'cture Additional evidence of tlie close relationship of Lissonyctcris and Myoiiffcicris comes from Benedict (1})57) who says that they have essentially the same liair structure: the pigment distribution, absence of amedulla, and form, an-angement and dimensions of the scales being alike. The slight difference she describes in the more distal maximum diameter of the hair in Myonyctcris seems scarcely imj)or1ant. The hairs of Cynopicvus are so different as not to need comparison here. Those of Ii(/its( ft its are less compellingly so, being similar in pigment dist i-itmt ion and in68 RELATIONSI Ill's oi' I,ISSONYCTERIS 9 at ISC Ilee of ;i iiicdiil l;i. Tlicy appjirciit ly ditVciliowcvfT in arraiigeiiu'iil aiul size of s(';;|cs ;iii(l. 1(» a(•ci'tiiiii cxtciil in lliciiform. In ucnt'i'al, Ih.rsc (lirfci-cnccs seem more of dcHi-cc ilmn kind. Cranial Characters C'r;inially, as cxtcrnjilly, Lissoiijich iis ;ind Ihioinich )is i)Ossess ainnnl)('iof fcatui-cs in coininon wliicli dist inuuisli tlicni ('({iially from R<jus(llus and (' iinopivnis. In addition to the jjroportions of the skull as awhole, llic shape and spaeiii<i' of the teeth, the occlusion i)a1tcrn, and the structure of the interorbital retiion are particularly iinpoi-tant. In <i'cneral jiroportions. tlie doiniuant featui-e of the small Mi/oinich ris skull, and one ill which it aui-ees with the much lar<i'(M' Lissonyctcris. is the relatively ^reat leni>th of tlie anterior part of the skull (measured Irom behind the postorbital processes to th(^ tips of the premaxillaries) as compared with both leniith of the brain ease (measured from behind the postorlatal jiroeesses) and its bulk. Typically, also, the bi-aiii case in Mi/oiii/ch rIs is elon<iatcd behind the temi)oral root of the zyo'omatie arch, ami the rostrum is slender with the posterior ends of the nasals depressed giving the skull in profile adish-faced ajipeai'ance, an appearance that is accentuati'd by tlie great develoi)ment of the lateral of the two pairs of frontal sinnses. Compared with Iti)iis( ff us. Llss<)inict( ris and Mijoinjctcris have amore slender rostrum. In the two latter, aprojection of aline along the top of the nasals ])asses ventral to the top of the orbit and the brain case. Tn Roiisctfus. sciis}/ <iiricio, snch aline passes dorsal to both of tlK^se points. In the myonycterines, the brain case is elongated and flattened so that the distance from the bottom of the occipital coiulyles to the top of the occipital crest is less than the distance fi-om the postglenoid jirocess to the back of the comlyle. whereas in Ixousciius the first distance is ecjual to (ugreater than the s(M'ond. The largeiorbit, though difficult to measure, is easy to sec. In Lissoiiijcfifis and Mnonijctcris the diameter of the orbit taken approximately parallel to its anterorostral margin is e(pial to oigreater than the laci-ymal width; in Nousdtiis it is less. Further, the antero-venti-al border of the oi'bit foi-med by the i-oot of the zygomatic arch is abetter develojx'd, sharper edged rim; it is thin and i)late-like Avhere the infraorbital canal pierces it. The thicker root in Roiisctfus results in anoticeablv longer canal. 16 BREVIORA No. 184 J^^ Figures '2, '.'>, niid 4. Lismui i/cl cris ainioh iisis li.-iinlliii^' iiicccs (if li;iii;iii,i. The use (if tlic fddt ill Ji.-iiiijiiiii;' fddii .iml tlic piistnri' of tlic wiii^is wiiiic doing so is t-lciirly siiown. Figure ~). Li.s.son!jcteri.'< (tiifiolnisis resting, Imiigiiig on ;i (;iii\;is \v;ill .-it its junction witli tlie ceiling. Xote tiie position of the feet wiiicli c;in hold onto ;i I'oost equally well in this or the oiijiosite orientation. The wings are being held well o\'erla|iiieii, sulistnnt ially enwiapping tlie body. If the bat had not liecn disturbed by the photographer, its chiii would luive heen tucked against its chest and its eyes closed.