Foraging Behaviour of the Australian Ghost Bat, Macroderma gigas (Microchiroptera: Megadermatidae)

704 J. M. P. Joss and J. A. Minard birth would not have sperm stored and would require a spring mating. Although there is no e idence from this study that L. delicata produces more than one clutch per season in the seasnnVT0"' f 'lsuevldence that youn8 f™ale L. delicata can produce a clutch in the ason following their birth, which explains why the males of both species are capable of Zdnfed m sPnng and late summer. The reduction of the number of clutches drofrht*2 h SeaSOn y gu,chenoti'from tw° in ihe years of good rainfall to one in the consfderaht'e TSS lts. reproductlve potentia' down to that of L. delicata. There is now StSSff '""d ramfan Ca" affeC' "Zard reproducti°p (« "'era.ure cited I pr°ducing cha"ges in productivity of annual vegetation and lizards fhe deZ r u" tUm de,ermine the availability of food resources for the ahe fclesht a cvh'" Tn TT'dUnng ,Ms Study appeared bp insufficient to alter the testicular cycles of the males, but follicular development in the females involving as "nr™ vitellogenesis, did show considerable variatLn irom me years ot good to that of poor rainfall. dryTseL'rofT979tSsnWhlCh Ca" m 7™ 35 a" indlcatio" of condition of the lizards over the by^decreased^ fZ mL"T*"" " 'S lhe fema,es which are affected oy decreased food supply. In mature females liver weights fell during the early serine as rapfdlyTlSraSnhbetweenWse'El:tS *°Se " Spring after thplr apiai> ieii again between September and November, presumably because the males were feedffig'IThise^^ fonowed'li0^UCI'0H a°d spcnd'ng more time in matin^behavimm than in reproduc "4ly .XeTThev 7 '" °n howi *•**=»"' and L. delicata remain investigation. References (Elsevier-North LLd BicZsterLT "" * PP" 223"34" "s7S: for communal eg8"'ayin8 by aad °r^h - -*** - <* «• Manuscript received 25 October 1984; accepted 9 April 1985 Aust. J. ZooL 1985. 33. 705-13 Forag g Behaviour of the Australian Ghost at, Macroderma gigas (Microehiroptera: Megadermatidae) C. R. Tidemann\ D. M. Pridde/B, J. E. Nelsonc and J. D. PettigrewD A Department of Zoology, Australian National University, GPO Box 4. Canberra City. A.C.T. 2601. B School of Biological Sciences, Zoology Building, University of Sydney, N.S.W. 2006. c Department of Zoology, Monash University, Clayton, Vic. 3168. D Department of Physiology and Pharmacology. University of Queensland. St Lucia. Qld 4067. Abstract The foraging behaviour of Macroderma gigas was investigated in the field by radio-tracking of tagged individuals and by direct observation of feeding bats in May 1983. Foraging areas were centred, on average, 1 -9 km from the day roost. The mean size of foraging areas was 61 ha and tagged bats generally returned to the same areas each night. Hunting behaviour of bats within foraging areas consisted of observation at vantage points, with brief sallies to capture prey, mostly insects on the ground, although hawking of flying insects was also observed. Vantage points were changed about every 15 min during foraging periods. The mean distance between them was 360 m. Foraging areas were not exclusive, there was overlap between the ranges of several tagged individuals, and in one case an area was used by 20 bats. Introduction The family Megadermatidae contains five extant species; Laviafrons and Cardioderma cor in central Africa, Megadenna spasma and Me. lyra in south-east Asia and Macroderma gigas in northern Australia (Corbet and Hill 1980). There is a considerable size range, from Lavia weighing 23 g to Macroderma weighing 120 g (Stephan et a/. 1981), but all have a very similar appearance, typical of a number of species of bats which capture their food by gleaning: very large ears, large eyes and a simple but prominent nose-leaf. It is known, on the basis of direct observation or the analysis of diet, that all five species of megadermatids capture a large proportion of their prey by gleaning. Those which have been observed are unusual among gleaning bats in that they spend very little time in flight. Most of their foraging time is spent in carefully inspecting their surroundings from selected vantage points; flight is restricted to brief sallies to capture prey and short movements to new vantage points. Cardioderma hunts by hanging in the low branches of trees and then swooping to the ground to catch insects, which are carried back to the perch and eaten (Vaughan 1976). Lavia uses a similar foraging technique of short flights from a vantage point (Shortridge 1934; Wicklerand Uhrig 1969; Kingdon 1974). Megaderma lyra and Me. spasma eat many species of flightless animals (Brosset 1962; Advani 1981; Advani and Makwana 1981) and M. lyra locates live prey in the laboratory in the same way as Cardioderma and Lavia do under natural conditions (Fiedler 1979). Published information suggests that Macroderma also is likely to use gleaning as one means of obtaining food. It includes in its diet a wide variety of small animals, many of which are flightless (Douglas 1967; Vestjens and Hall 1977). The brain morphology of Macroderma suggests a combined substrate and foliage gleaner (Stephan and Nelson 1981) and in the laboratory it locates prey by the 'sit and inspect' technique used by Cardioderma and Lavia (Kulzer et al. 1984). However, little is known of its foraging behaviour in the field. 0004-959X/85/050705$0: 706 C. R. Tidemann el al. The purpose of this study was to investigate the foraging strategy employed by M. gigas under natural conditions. Materials and Methods The study was carried out in May 1983 on a colony of M. gigas day-roosting in Kohinoor Mine, about 1 km south of Pine Creek, N.T. (13°49'S„ 13T49'E.). This disused gold mine was originally excavated in the 1880s and used intermittently until the early 1900s. It consists of a 200-m horizontal tunnel with several lateral workings and a number of entrances and ventilation shafts. The colony of Macroderma, estimated conservatively to number 450 bats, roosted mainly along the largest lateral working. for which 10 or more triangulated locations were obtained, are shown, together with the estimated centre of each. The sex of these individuals and the size of their foraging ranges are given in Table 1. The area surrounding the mine includes a series of small hilly ridges rising up to 70 m above extensive flat plains. Vegetation is woodland or open woodland, dominated by Eucalyptus tectifica, E. alba, E. tetrodonta, E. latifolia and E. apodophylla. The understorey is predominantly grasses: Cymbopogon bombycinus, Aristida latifolia, Themeda australis, Plectrachnesp. and Eriachne sp. (Wood 1977). The climate of the region is characterized by distinct wet and dry seasons. Mean annual rainfall at Pine Creek is 1200 mm, of which 1100 mm falls in the wet season from December to March. Twelve adult M. gigas (seven male and five female) were caught within the mine with hand nets and mist nets. These individuals were fitted with radio transmitters (150 MHz) and immediately released. Foraging Behaviour of Ghost Bat 707 Their movements during the next hve nights were monitored from a network of three wo types of transmitter were used: two Austee KG80 and » * by Eveready CR2032 lithium cells and were htted w,20-cm* ere attached t0 the lower (1982) J nave tracked^everal species of small vespertihomds carrying 25% ofmeTrbody weight with no apparent ill effect^ h attached transmitters caused little Prevtous observations of two captiveM ndt^tM^ ^ the captive anunals after 4 and discomforiorlossofmanoeuvreabihty.Thetransmme ^ibSKbU^m2) on Rhnolophm-T^ Tdays, leaving bare patehesofsktosimdartothosedescnbe^ y^ ^ ^ ^ ^ ,racked from Austec transmitters had a maximum rang ,„tinn was distances over 8 km. -d nearby the mine (Fig. 1)- Eac s Three tracking stations were constructed o g , nt yagi antenna or a null-peak . equipped with a Telonics TR2 receiver a*^^ deduced from the tnangulation ofbeanngs of two four-element Yagts. The location of ^^^^guiation of three beanngs yrdds toe recorded concurrently from the three tracking ^"onsri % inger 1979). If the ana of thts erior discrete intercepts which form the vertices of:ani error ^ length of one side of to mror triangle was greater than 2 • 0 km the mad ^eanngsjas ^ ^ of ^ bat was uken as he mangle was less than half the length of was taken as the arithmetic cm to* midpoint of the shortest side. , ™\se thmocationvn during selected periods between 900 error triangle. Each tagged bat was radiotracked every k„„ations one day and 0700 h of the next. radio-tracking was used to guide visua o , Information gamed about foraging ™ scopes (Javelin: Smith: and Wesmnk of foraging behaviour. Bats were observed throughni^ q ^ male No. 712. Much of the ground Most of these observations were made inAJ 0' and the ground was bare except or em Results Fate of Tagged Bats Five transmitters were retneved Transmitters remained attached to bats for up ' after attachment (mean from within the mine after they tQ J* seen in captive animals 5 days). Three bats with bare patches in their , ^ ^ ^ ^ Kohmoor Mine which had shed transmitters, were seen flying. PP d Tw0 others were flying with One bat was recaptured with the not recaptured, transmitters still in place 7 days after at , ransmitter. Despite a brief post-mortem One bat died a few hours after attachment It is p0Ssible that the bat sustained examination, the cause of death could notbe> ^ used t0 attach the transmitter. The injury during capture or ingested some of th d bim through overloading was fact that the animal died so soon after capture suggests an unlikely cause of death. Timing of Activity £t approximately During the study period sunrise was and returned shortly before 1825 h. Although most bats left the mine sh ^ ^ between individuals, sunrise, the times of exit and particuiarly - £ was 1951 h ± 11 mm (SE) (range 708 C. R. Tidemann el al. early as 0330 h and did not re-emerge. The latest return of any tagged bat was 0617 h, but untagged bats were seen to return to the mine up until 15 min before sunrise. Repeated forays from the mine during a single night were not observed. o* * A O 790 Fig. 2. Successive triangulated locations of two M. gigas, 790 (3) and 712 (3). over three successive nights. • Locations visited on night 1. o Night 2. • Night 3. • Kohinoor Mine. i 1 1 Km o °* . • • • • jp • * • 712 A A A o° • • Foraging Areas Upon emergence tagged bats dispersed rapidly from the mine. Following these direct commuting flights bats generally remained within relatively small individual 'foraging areas' (Fig. 1) to which they returned on consecutive nights (Fig. 2). However, there were exceptions to this. For example, a female M. gigas (No. 976) foraged in the same area on two consecutive nights, but on the two subsequent nights it was located in another area 10-15 km to the north-west of Kohinoor Mine. On these latter two nights this bat did not return to Kohinoor, and presumably roosted in another mine. Just before sunrise on the fourth night it entered the North Enterprise Mine 1 km north of Kohinoor. Over the five nights of radio-tracking, three bats each foraged in two separate areas and on two occasions bats visited two foraging areas during a single night (Fig. 3). Foraging areas were inspected during the day, and a number of trees used as feeding roosts were found by the evidence of bat faeces and the discarded remnants of prey beneath. Food debris was almost exclusively remains of the yellow-winged locust, Gastrimargus musicus, which was extremely common in the area throughout the study, particularly in burnt areas where grass shoots were emerging. Other species recovered from beneath feeding roosts were katydids, scarabaeid beetles and large moths, and in the mine itself some feathers were found. However, it is not known if these were of recent origin. Plots of all foraging areas were compared with aerial photographs and topographic maps. There was no apparent preference for different types of woodland or topographic features; Foraging Behaviour of Ghost Bat 709 tagged bats appeared to be rando^^V e/a/t°984) made by this to detect the presence of untagged ant ^ also indicated that M. gigas foraged species while foraging. The distriu vjcjnity of the three tracking stations. Plots o throughout the study area' incl^d.1hna8t these couid overlap (for example, bats No. 730 and individual foraging areas showedd 2Q M gigas in close proximity to 811; Fig. 1) and sight observations at one locamy each other. 0045 0445 0515 Fig 3 Foraging behaviour of two M. gigas on single nights. ®°*^^yUa^r.7Times at the beginning two'discrete foraging areas. The day roost (Kohinoor Mme)«in^cat«iby£u at whjch a bat was located and end of sequences of behaviour are shown. Each spot represents during one or more 15-min periods. Where 10 or more locations of a bat could be Theoretical the individual was calculated by the mini™a"\ 8 m9) were estimated, and the discentres of activity within these foraging a ( (Kohinoor Mine) were calculated tance between these centres of activity and the day distance of their (Table 1). The mean size of the ^separate foraging areas, the centres from the day roost was 1 - 9 km. For bats whicn use estimates for the two areas were averaged. Foraging Behaviour chnrtlv after All tagged bats showed similar Here they 710 C. R. Tidemann el al. more, before they resumed foraging in the early morning. It was observed that the resumption of feeding activity by tagged bats after inactive periods was often synchronous with a renewed chirping by other bats in the vicinity. The movements of two individuals during one night are shown in Fig. 3. Each bat used two discrete foraging areas. Bat No. 976 (female) was in continuous radio contact from the time of emergence from the mine at 2015 h until 0130 h. Between 2030 h and 2200 h it made short flights about every 15 min, usually for less than 1 min. During these flightless periods radio contact was often lost or the signal strength diminished. Each time radio contact was re-established, the bat was still in the same location. Such fluctuations in signal strength indicated that perching bats made regular forays onto the ground, presumably to capture prey. After foraging for about 90 min, No. 976 remained stationary at a night-roost until 2314 h, at which time it resumed the foraging behaviour observed earlier in the night. Later, at 0045 h, the bat moved directly 3-5 km to a second area where it recommenced short foraging flights. Contact was discontinued at 0130 h. Table 1. Sizes of foraging areas and distances of theoretical centres of activity from day roost of five M. gigas The number of locations on which each calculation was based are shown; values are given only for those individuals for which 10 or more locations were obtained Individual Sex Size of Distance from day No. of No. foraging roost of centre locations area (ha) of activity (km) 712 6 28-47 0-45 26 790 S 84-00 2-95 22 811 6 33-84 1-32 16 730 9 37-04 2-66 10 976 9 120-80 2-09 41 Mean ± SE 60-83 ± 18 0 1-89 ±0-45 Bat No. 811 (male) emerged from the mine shortly before 2130 h. It remained within a few hundred metres of the mine, making periodic short foraging flights, until contact was discontinued at 2245 h. When radio-tracking resumed at 0230 h, the bat was still close to the mine. A few minutes later it flew rapidly to a second foraging area 2 km away and recommenced foraging until 0445 h. By 0515 h it was back near the mine and eventually entered it about an hour later. To estimate the average length of flights within foraging areas, distances between consecutive triangulated locations were calculated. The mean distance of 79 such flights was 360 m. The longest timed flight of any bat was 3-5 km in 10 min—an average flight speed of 20-7 km h-1. Thus, most flights between vantage points would take less than 1 min. Observations of Foraging Observations through night vision scopes showed that foraging bats spent most of the time hanging from small branches or the main trunk in the mid-to-upper canopy of eucalypts at heights up to 3 m from the ground. From these vantage points they appeared to inspect their surroundings with frequent head and ear movements. Chirp calls were produced frequently in association with this behaviour, although it was difficult to ascertain which animal had made a particular call and no measures of repetition rate were recorded. Fifteen instances of prey capture were observed: six of these involved flying insects: on the other nine, insects were caught on the ground. Prey capture was rapid and was never far from the vantage point. Insects were seized swiftly and carried back to the vantage point Foraging Behaviour of Ghost Bat 711 where they were consumed. Portions, such as wings, were discarded and were allowed to the ground. Discussion . ,, This study shows that M. gtgas uses a foraging Zardioderma and Lavia: it locates its prey y r®™ai 'capture prey and brief estricts its flight activity during foraging to ^.^^ally not far apart and it movements to change vantage points, an ag During Mav (end of the wet rrobably takes less than 1 min for a bat » ^"he gmund surface, although h season) M. gigas was behaving primarily as g restricted range of prey ilso caught some prey in mid-air. However, it was teeding on at this time. Pntirelv remains of the yellow-winged Food debris beneath feeding roosts was almost entiy in Qctober 1983 locust G. musicus. Debris examined beneath a major role (end of the dry season) showed that vertebrates, p consumed. Food remnants collected in the diet, although some invertebrates were s 1 11 . daSyurids, some skinks in October contained numerous feathers, a ew sm^ Gastrimargus) moths, beetles and and several bat wings. Remains of)Large gnasshoppfrSn(^^cation) It is evident also from cockroaches were also found (M. Schulz persona animals (Jones 1925: Douglas other studies that M. gigas feeds on a wide variety of small animals (Jones 1967; Vestjens and Hall 1977) mpgaderrnatids. Advani and Makwana Similar diversity in diet has been shownim other ^ yeaf They (1981) found substantial variation in the diet ; g w bm that vertebrates were concluded that insects were eaten whenever th insects were uncommon. relied upon to a large extent during the ry s . '... » -s ^Hed at making aerial Cardioderma modifies its diet according to pre> a^ . D rtsof-thewetseason; at other captures ofinsects and uses this technique extensive y season it feeds almost times it captures prey by gleaning from an nhrig (1969) observed Lavia to feed entirely from the ground (Vaughan 1976). Wrfteiffld °bse found ^ fed exclusively on flying insects, whereas Shortndge (1934) and Kmgao t Smagadermatids have a they can employ to capture many nQt known. Vaughan (1976) thought sit-and-wait method of prey location in all sit Pthnd but Brosset (1962) recorded that Cardioderma located its prey univer^Hy„ the walls of that Me. lyra and Me. spasma hunted by flying frpnuentlv by a sit-and-wait the cliffs'. It is very unlikely that diurnal1 bir s wou e c ^ ^ ^ (1977) speCulated approach, unless they were captured at dus or a - Ja numbers of diurnal that Vampyrum, a carnivorous phyllostomatid, which laboratory, M. birds year-round, actively sought out its prey by (Kulzcr gigas did not detect silent and motionless mice, even detection by this species. 1984) seems to preclude the use of smell as a L its Irrespective of the method by which M. gigas is substantially different from that foraging behaviour when it is feeding mainly on birds is substan iall> dittere observed during this study, when it was feeding main \ romoared to those covered 1, or,1... 0.2 », -o,h .h.. = - 'J« 712 C. R. Tidemann el al. to suggest from the present study that M. gigas is territorial in its foraging behaviour. On the contrary, the feeding ranges of several tagged bats overlapped and a loose group of 0 individuals was observed in the feeding area of one tagged male^ A calculationhe mean size of foraging areas (61 ha) and the mean distance of the roost (1-9 km) suggests also that it is unlikely that this species defends feeding areas (450 individuals in n>£). Similarly, Brosset (1962) estimated from the distribution of feeding roosts that foraging individuals from a colony of about 1500 Me. lyra spread out only abo Tte significant of the chirp call of M. gigas is unclear. It is loud and has audible component that can be detected by humans over 100 ^ away _ The ^ong o Cardioderma is similar in these respects, but in that species it serves a tern tonal functionm the establishment of exclusive foraging areas (Vaughan 1976). Quite dearly, not exclude others from its foraging areas. However, it was apparent from_ observa ions made during this study that foraging activity in the field was almost invariably associated wfth frequent chirping. In the laboratory Kulzer « al. (1984) recorded " emitted chirps at regular intervals, and on several occasions one bat reacted to nearby flight of another with a series of chirps. Whatever the function of the call, it seems> one of its effects would be to enable foraging individuals to maintain ^ other, and possibly thereby to achieve spacing appropnate to particular prey species densities. Acknowledgments We are indebted to the many people who provided assistance 'with the field work: B Baker D Baker-Gabb, G. Baverstock, R. Coles, L. Conole, S. Churchill P. Dabd, C Ellis, K. Fitzherbert, E. Gano, A. Guppy, L. Hall, P. Helman L. Hodges, IPulsford G Richards and M. Schulz; the study could not have been accomplished without them The Conseiwation Commiss.on of the Northern Territory provided ^ project and logistical support in the field. To Russell Dann of GoldfieW Explorabon Enterprises we are grateful for information on the location of mine entrances. Purchase of telemetry equipment was supported by an A.R.G.S. grant to J.D.P. and a Gran from the Conservation Fund of the South Australian Department of Environment and pla™in§ C R T We thank the New South Wales National Parks and Wildlife Service for the loan of telemeU equipment. The constructive criticisms of two referees are gratefully acknowledged. References Advani. R. (1981). Seasonal fluctuations in the feeding ecology of the Indian faBevampire Megaderma lyra lyra (Chiroptera: Megadermatidae) in Rajasthan, India. Z. Saeugetierkd. , • Advani R„ and Makwana, S. C. (1981). Composition and seasonal occurrence ofammal remams roosting habitat of Megaderma lyra lyra in Rajasthan India. Angew. Zool. 68 75 Brander R B and Cochran. W. W. (1969). Radio-location telemetry. In Wildlife Management Tprh'ninues' (Ed. R. H. Giles.) pp. 95-103. (Wildlife Society: Washington, D.C.) Brosset, A. (1962). The bats of central and western India. Part 2. J. Bombay Nat. Hist. Soc. 59, Corbet. G.V and Hill, J. E. (1980). 'A World List of Mammalian Species.' (Comstock Publishing Dougtas.C A^M967)! The natural history of the Ghost Bat (Macroderma gigas) (Microchiroptera: Megadermatidae) in Western Australia. West. Aust. Nat. 10, 125-38. v*m«irp Fiedler. J. (1979). Prey catching with and wihtout echolocation in the Indian False \ ampire (Megaderma lyra). Behav. Ecol. Sociobiol. 6, 155-60. Hayne D W (1949). Calculation of size of home range. J. Mammal. 30, 1 18Jones, F. W. (1925). 'The Mammals of South Australia.' (Government Printer: Adelaide.) Foraging Behaviour of Ghost Bat 713 lc volume IT Part A Insectivores and Bats.' (Academic Kingdon, J. (1974). 'East African Mammals. Volume , Press: London.) F p (1984). Prev-catching behaviour and Kulzer, E„ Nelson, J. E„ McKean, J. L'^Macroderma gigas (Microchiroptera:Megadermatidae). echolocation in the Australian ghost bat, Macroderma gigus Aust. Mammal. 7, 37-50. annlications and limitations. In 'Animal Marking: Macdonald, D. W. (1978). ^1°-t^in|S°®rchPP(Ed B. Stonehouse.) pp. 194-204. (MacmiUan: Recognition Marking of Animals in Researcn. ( London.) . n(.Wftr,L American small mammals. Am. Midi. Mohr, C. O. (1947). Table of equivalent populations of North American sma Nat. 25, 196-233. f - , Vol j (Heinemann: London.) Shortridge, G. C. (1934). 'The Mammals of S°uth"^st ^ . triangulation. J. Wildl. Manage. Springer J. T. (1979). Some sources of bias and sampling error in radio tnanguia JS£ZS. (1982). Radio-tracking greater horsehoe bats with prehminary observattons on fiight SteptrH Brains of Australian Chiroptera. I. Encephalization and Step™~ Evolutionsforsch. 19, 195-222. , , African faise vampire bat (Cardioderma cor). Vaughan. T. A. (1976). Nocturnal behaviour of the African laise P J. Mammal. 57, 227-48. observations on the foraging behaviour Vehrencamp, S. L., Stiles, F. G„ and Bradbury. J. W. (1977k Observat MammaL 58. and avian prey of the neotropical camtvorous bat Vampyrum spectrum. VeSw. JM., and Hall, L. S. ,1977,. Stomach contents of forty two species of bats from the Wi^er^^andTJhri&Ch (19^9)^Vadialtenund skologishche Nwch^der Gelbflugelfledermaus, Lavia frons (Geoflroy) (Chiroptera, fforestry,' Fisheries and Land Manuscript received 14 November 1984; accepted 10 April 1985