Foraging ecology and organisation of a desert bat fauna

Abstract

Airframe design parameters related to flight performance, stability and control had tight, functionally appropriate relationships with the foraging niches and echolocation parameters of nine species comprising the bat fauna of the Little Sandy Desert, Australia. The airframe parameters segregated into two near-independent groups, one related to microhabitat use, the other to foraging strategy. The structure of the desert's bat fauna is displayed in these terms, and its organisation is compared with the faunas of surrounding regions. A diversity–productivity model of faunal structure is revealed, with an organisation that conforms with the 'specialisation' hypothesis. Clear family-level relationships between phylogeny and foraging ecology imply that ecological specialisations occurred early in the evolution of bats.

Full text

Publishing
Volume 50, 2002
© CSIRO 2002
AAuussttrraalliiaann JJoouurrnnaall
ooff ZZoooollooggyy
Australian Journal of Zoology
CSIRO Publishing
PO Box 1139 (150 Oxford St)
Collingwood, Vic. 3066, Australia
Telephone: +61 3 9662 7622
Fax: +61 3 9662 7611
Email: publishing.ajz@csiro.au
Published by CSIRO Publishing
for CSIRO and the Australian Academy of Science
www.publish.csiro.au/journals/ajz
All enquiries and manuscripts should be directed to:
A journal for the publication of the results of original
scientific research in all branches of zoology,
except the taxonomy of invertebrates

Australian Journal of Zoology, 2002, 50, 529–548
© CSIRO 2002 0004-959X/02/05052910.1071/ZO01029
Foraging ecology and organisation of a desert bat fauna
N. L. McKenzie
A
, A. N. Start
A
and R. D. Bullen
B
A
Department of Conservation and Land Management, PO Box 51,
Wanneroo, WA 6065, Australia.
B
43 Murray Drive, Hillarys, WA 6025, Australia.
Abstract
Airframe design parameters related to flight performance, stability and control had tight, functionally
appropriate relationships with the foraging niches and echolocation parameters of nine species comprising
the bat fauna of the Little Sandy Desert, Australia. The airframe parameters segregated into two near-
independent groups, one related to microhabitat use, the other to foraging strategy. The structure of the
desert’s bat fauna is displayed in these terms, and its organisation is compared with the faunas of
surrounding regions. A diversity–productivity model of faunal structure is revealed, with an organisation
that conforms with the ‘specialisation’ hypothesis. Clear family-level relationships between phylogeny and
foraging ecology imply that ecological specialisations occurred early in the evolution of bats.
ZO01029
Or ganisation of a desert bat faunaN. L. McKenzie
et
al .
Introduction
Various recent papers have identified relationships between wing shape and resource
partitioning in bats (e.g. Aldridge and Rautenbach 1987; Findley 1993). Two wing ratios
(aspect ratio and wing loading) provide robust predictors of microhabitat use (e.g. Findley
and Wilson 1982; Norberg and Rayner 1986; McKenzie and Rolfe 1986). Recently, Bullen
and McKenzie (2001) showed that tail and ear ratios indicate species’ foraging strategies.
Herein, we use data on bats from the Little Sandy Desert to further test these findings, and
determine whether relationships between airframe ratios, echolocation parameters,
microhabitat use and foraging strategy are functionally appropriate. We display the
structure of the desert’s fauna in these terms, and draw comparisons with faunas of
surrounding regions to assess its organisation.
Investigations of community structure, such as faunal studies, should emphasise the role
of species’ foraging method because of its potential importance as a mechanism of resource
allocation (Simberloff and Dyan 1991). Because it is difficult to observe, data on the
foraging microhabitats and strategies of most Australian bats are scarce. Thus, our field
program focussed on collecting these data, and we report them in detail.
Methods
Study area
The Little Sandy Desert covers 110 000 km
2
of Australia’s western interior (Fig. 1). One of Australia’s ‘red
centre’ deserts, it comprises sand-dune fields and interdune plains covered with Triodia hummock
grasslands with emergent shrubs or low ‘tree-savanna’ of mulga (Acacia aneura). Dune crests support open
shrublands with emergent bloodwood trees. Claypans and areas of fine-textured soils on interdune floors
and footslopes of ranges support low open woodlands of mulga over tussock grasses. Scattered through the
desert are abrupt ranges of Proterozoic sandstone covered in hummock grassland with sparse, emergent low
trees. Groves of river gums (to 15 m high) and swards of couch grass fringe watercourses and semi-
permanent pools in the ranges. Along with thickets of Melaleuca lasiandra to 5 m high, these river gum and
couch grass communities also fringe semi-active, internal drainage lines such as Savory Creek and Rudall
River, which traverse the dune-fields. Melaleuca thickets and swards of marine couch grass form shrubland
communities on floodplains and around saline playas such as Lake Disappointment. In western parts of the
desert there are ferricrete mesas covered with hummock grassland and sparse mulga. The climate is arid
with locally unreliable summer rainfall. Beard (1975) provides details of climate and vegetation.

530 N. L. McKenzie et al.
Field sampling
The results of 1975, 1976 and 1979 fieldwork in the Carnarvon Range, Durba Hills and Rudall River
headwaters (Fig. 1) are summarised in McKenzie and Burbidge (1979) and Burbidge and McKenzie (1983).
Our 1996 and 1999 field surveys were focussed within 15 km of three campsites: Beyondie Camp
(24°39′27″S, 120°08′08″E), Cooma Well (24°04′58″S, 120°20′28″E) and Savory Camp (23°51′10″S,
120°31′30″E). These three survey areas are all in western parts of the desert (Fig. 1). Using mist nets,
ultrasound detectors, spotlighting traverses and a portable floodlight, we surveyed the bats present in each
survey area, noting the location, date, flight behaviour, air speed and foraging microhabitat (in relation to
vegetation structure) of each record. Bat echolocation call sequences were stored on magnetic tape, and
annotated with these ecological data.
Further data were collected by P. Kendrick, J. Rolfe and A. H. Burbidge in the headwaters of the Rudall
River during November 1996, and by A. A. Burbidge in the Calvert Range (23°56′S, 122°41′E) during
August 2000.
Taxonomic voucher specimens were taken from all localities except the Calvert Range, and have since
been lodged in the West Australian Museum collection. Nomenclature follows Strahan (1995). The work
was conducted under CALM Animal Ethics Committee Permit No. 14/93.
Airframe characteristics
Airframe measurements were taken from a sample of live, non-pregnant adults belonging to each microbat
species that we captured during the field programme. These measures allowed us to calculate seven flight
Fig. 1. The Little Sandy Desert. B = Beyondie Camp, CO = Cooma Camp, S = Savory
Camp, CR = Carnarvon Range, R = Rudall Headwaters, D = Durba Hills, and CL = Calvert
Range.

Organisation of a desert bat fauna 531
performance, stability and control parameters: aspect ratio, wing loading, tail area ratio, tail length ratio,
ear area ratio, ear length ratio and leading-edge flap chord ratio. Their measurement protocols, relevant
formulae, and a discussion of aerodynamic mechanisms and implications are provided in Bullen and
McKenzie (2001).
• Aspect ratio (AR) and wing loading (WL) are primary determinants of longitudinal flight performance
because they summarise lift, thrust and drag relationships – a bat’s ability to generate lift across a wide
range of airspeeds, accelerate to speed and then maintain that speed.
• Tail area ratio (TAR) and tail length ratio (TLR) relate to the bat’s manoeuvrability (rate of turn in degrees
of arc per second), an important characteristic in airframe control and stability. TAR, in particular, is a
useful measure of the aerodynamic leverage that the tail can apply to the rest of the bat in flight.
• Ear area ratio (EAR) and ear length ratio (ELR) relate to a bat’s agility, its ability to generate a turn
acceleration (degrees of arc / s
–2
). E
bar
(= EAR*ELR) summarises the ear’s ability to sustain the consequent
aerodynamic and inertial loads (without deforming) while carrying out its aerodynamic and auditory
functions (i.e. shape and orientation).
• The coarse, ‘wing-twist’ mechanisms used by highly agile species to generate gross rolling moments
(eg ‘wing-fold’ and ‘asymmetric wing angle-of-attack’) induce high drag penalties at medium to high air
speeds. They are less efficient energetically than the finely controlled rolling moments generated by
leading-edge wing-flaps. We assessed species’ wing-flaps by calculating their ‘leading-edge flap chord
ratio’ (C
LEF
/C
TIP
) from McKenzie and Bullen (2001).
Echolocation calls
Data on bat echolocation calls were collected during the October 1996 field program for two reasons.
Firstly, because echolocation parameters are known to reflect differences in species’ foraging ecology
(e.g. Neuweiler 1984; Schnitzler et al. 1987; Fenton 1990). Secondly, so that we could identify the free-
foraging bats from which we recorded microhabitat, flight speed and foraging strategy data.
Call sequences were recorded using the frequency-division function of A
NABAT II (Titley Electronics,
Australia) or D940 (Pettersson Elektronik, Sweden) ultrasound detectors (divided by 16 or 10,
respectively). Output was stored directly onto Metal IV cassette tapes using a Sony Walkman Professional
(WMD6C) tape recorder.
Only consistently shaped, regularly spaced, sequential sets of calls emitted at a rate of no more than one
call per wing-beat, and with a clean, well defined, low-frequency component, were measured. These were
calls of bats in ‘search’ mode, rather than ‘approach’, ‘interception’ or ‘departure’ modes (Schnitzler et al.
1987; Kalko and Schnitzler 1993, 1998; McKenzie and Muir 2000).
Species calls were characterised in terms of three attributes using A
NABAT5 software (Titley Electronics,
Australia): the frequency of the shallowest section of the frequency-sweep near the end of the frequency-
modulated call (F
min
), the highest detected frequency (F
max
), and the ‘duration’ of the call.
C
OOL EDIT (Syntrillium Software, USA) was used to digitise the analogue tape recordings as ‘.wav’ files
using at a mono-sampling rate of 44 100 at 16-bit resolution, and display each call sequence in
frequency–time domain. A 2048-point Blackmann–Harris fast-fourier transformation of the fundamental
harmonic of each search-mode call provided a profile of the call’s frequency spectrum. The frequency
maintained for the greatest number of cycles (F
peak
) was used as a fourth call attribute. Each search-mode
call was scanned in its entirety to obtain this value.
Usual foraging microhabitat
Microbat records from our 1975, 1976, 1979 and 1996 fieldwork were assigned to one of four foraging
microhabitats, depending on differences in the degree of air-space clutter at the point where the free-flying
bat was first seen. Observations were made with a spotlight, in the glow of the floodlight or at dusk. They
were recorded if the bat was seen to be foraging or if its ‘search–approach–interception’ echolocation
sequences were detected. Microhabitat definitions follow McKenzie and Muir (2000), and were modified
from McKenzie and Rolfe (1986):
(1) cluttered airspace inside stands of vegetation (IS) or within 2 m of foliage, bark, ground or surface of
pools (BS/A) (IS and BS/A could not be discriminated in the open vegetation of the Little Sandy
Desert);
(2) semi-cluttered air spaces within a few metres (>2) of the sides and underside of the canopy or rock
faces (BS/O);

532 N. L. McKenzie et al.
(3) semi-open air spaces within a few metres (>2) of the top of the canopy or bare hills (AC); and
(4) open air spaces more than 5 m above or beside canopies, cliffs, hill-slopes or hilltops (OC).
The proportion of observations in each microhabitat was used as a measure of preferred foraging
habitat. Because relatively few bats were actually captured during the 1996 fieldwork, most of the
microhabitat observations came from ecological data we recorded with the ultrasound sequences that were
used to determine the species’ identities.
Usual foraging strategy
Species were assigned among three foraging strategies, depending on the manoeuvrability and agility of
turning manoeuvres observed during the 1996 fieldwork. We followed the foraging strategy definitions and
conventions detailed in Bullen and McKenzie (2001):
(1) air interceptor (fast, straight-line interception of prey – neither agile nor manoeuvrable),
(2) surface (intercept or glean prey in clutter at medium to low speeds – manoeuvrable but not agile), and
(3) air superiority (agile, to out-manoeuvre prey at medium to high speeds in open, semi-cluttered or
cluttered air spaces), with two classes being recognised – agile species that foraged in open airspaces,
and highly agile species of semi-cluttered to cluttered air spaces.
As mentioned above, manoeuvrability is the rate of turn (degrees of arc per second) while agility is the
rate of change of turn (degrees of arc per second per second) and relates to the load imposed on the wing-
spar, etc. A manoeuvrable bat has a small radius of turn while an agile bat can make these tight turns at
moderate to high speeds, and can abruptly tighten the radius of its turn. Therefore, by definition, agile bats
are inherently manoeuvrable, but manoeuvrable bats are not necessarily agile. Agility is most readily
recognised when bats perform horizontal turns at speed. By horizontal, we mean that the manoeuvre
doesn’t involve either a climb or a descent, so that we can estimate the aerodynamic loading (normal
acceleration in the saggital plane = cosφ
–1
, where φ is the bank-angle). During these turns, agile bats can
roll to bank-angles that exceed 70° (>2.9 g: Bullen and McKenzie 2001) so that turns appear abrupt rather
than radiused. Although agile turns are quick, in practice it is only necessary to assign bank-angle
observations among three categories: ‘shallow’ (<50°, not agile), ‘steep’ (60–70°, agile) and ‘near-vertical’
(>80°, highly agile).
We collected data for the foraging strategy assignments as follows:
• We measured flight speeds and made notes on manoeuvres used by bats as they foraged at dusk and at
night. After dark, the bats were observed in the glow of a powerful 12-V neon floodlight (‘Versa-Light’
by Burn-Brite Lights, Australia), supplemented by a wide-beam, 150-W spotlight for high-flying species.
They were identified later from their echolocation call sequences, although Saccolaimus flaviventris was
readily identified by its off-white undersurface, and Tadarida australis from its white under-wing stripes.
Speeds were measured using a K-band Doppler-shift meter (Municipal Electronics, UK, Model KR3).
All data were stored on the stereo cassette recorder, with the echolocation sequences immediately followed
by the corresponding speed and flight-behaviour observations.
• Captured bats were released at night with bioluminescent tags (‘Cyalume’) and their subsequent
microhabitat use and speeds recorded if bats returned to forage nearby.
• A tunnel-shaped enclosure was erected in the shade of a patch of river gums near Beyondie Soak. It was
17 m long, 2.5 m wide and 2.5 m high, and comprised three 20 m × 2.5 m mist nets stretched over a frame
of poles. Small species of bat were released into the tunnel one by one, so that their turning manoeuvres
could be observed in detail in daylight. Their minimum radius of turn and flight-speeds were estimated
using a stop watch, tape measure and pre-marked points along the side of the tunnel. Speeds over 4 m s
–1
were measured using the Doppler-shift meter.
All recognisable turn-manoeuvres were noted because tight wing-overs, stall-turns and flick rolls induce
airframe loadings equal to or greater than those generated during 70° or 80° ‘horizontal banking turns’,
implying agility (airframe loading of 2.9 or 5.8 g respectively: Bullen and McKenzie 2001). Jones and
Rayner (1988) measured airframe loading in Myotis daubentoni during an open wing-over at 23 m sec
–2
(2.34 g).
Analysis of ecological patterns
Rank Correlation Analysis (Kendall’s τ) was used to test for significant relationships between average
airframe parameters, echolocation call attributes and foraging microhabitats and strategies.

Organisation of a desert bat fauna 533
Results
Extant fauna
Eight microbat species had already been collected from the desert at the outset of this study:
Tadarida australis, Saccolaimus flaviventris, Taphozous georgianus, Mormopterus
beccarii, Chalinolobus gouldii, Scotorepens greyi, Vespadelus finlaysoni and Nyctophilus
geoffroyi. The earliest specimen was a Nyctophilus geoffroyi (M1424) collected by
O. Lipfert in 1930. Most of the others were collected between 1975 and 1979 from
southern, central and northern parts of the desert (McKenzie and Burbidge 1979; Burbidge
and McKenzie 1983). We recorded all of these species during our 1996 fieldwork in
western parts of the desert, as well as a ninth (Chaerephon jobensis) that was found only at
the Beyondie and Savory campsites, near the desert’s western edge (Fig. 1).
Airframe and echolocation attributes
Table 1 summarises data on the airframe parameters of Little Sandy Desert bats. Summary
statistics for search-mode echolocation calls are provided in Table 2.
Foraging observations
The usual foraging microhabitats and strategies of Little Sandy Desert bats are listed in
Tables 3 and 4, respectively. Foraging strategy assignments were made on the basis of
differences in species’ foraging behaviours that we observed during this study, combined
with published observations from elsewhere in Australia (e.g. Churchill 1998). We include
a detailed account of our observations, in context, because data on foraging behaviour are
difficult to collect in the field, and published data are scarce for most Australian bats.
Surface foragers
Small bats, identified as Nyctophilus geoffroyi from their echolocation calls, were
observed for 2 h in October 1996 as they foraged over, through and along the side of mulga
woodland fringing the pool at Beyondie Soak (24°29′26″S, 120°08′08″E; see Fig. 1). They
were gleaning prey from surfaces such as the ground, bark and foliage (close to and in
clutter – IS). They were also intercepting airborne prey by straight passes, while
‘contouring’ within 2 m of foliage, bark, ground or surface of pools (BS/A: Table 3). Flight
speeds were the lowest of the bats observed during the field program. Turns were frequently
tight, but not abrupt, and performed at low speeds near vegetation and when climbing away
from the water’s surface (as low as walking pace, ~1.2 m sec
–1
).
Two male and three female N. geoffroyi were captured in mist nets at this site. Flight speeds
and details of turning manoeuvres were recorded from two individuals released inside the
observation tunnel. Turning manoeuvres with radii of less than 1 m had low bank-angles
(often as skidding turns) and were performed after slowing to speeds as low as 1.1, 1.7 and
2.2 m sec
–1
, although straight-line speeds of up to 4.7 m sec
–1
were measured with the Doppler-
shift meter. Level, straight-line flight speeds achieved when the five individuals were
subsequently released at night with bioluminescent markers were 5.8 m sec
–1
(s.d. = 0.8,
n = 7). However, speeds of 4.4 and 5.0 m sec
–1
measured when two of the released bats returned
for a second pass suggest that flight speeds are usually lower. Turns observed at these speeds
had radii estimated at 1–2 m, implying bank-angles of less than 45°. In the Carnarvon Range
a specimen was observed ‘close contouring’ over a pool and through surrounding vegetation.
These limited observations are consistent with observations from elsewhere in Australia
(Table 4), that indicate a ‘surface’ foraging strategy in BS/A and IS microhabitats. For

534 N. L. McKenzie et al.
Table 1. Mean values of the airframe design parameters estimated for the microbat species recorded in the Little Sandy Desert
N, no. of specimens measured; W, weight; AR, aspect ratio; WL, wing loading (N m
–2
); TAR, tail area ratio; TLR, tail length ratio; E-bar, ear volume ratio. Standard
deviations are shown in parentheses
Species N
A
W (g) AR WL TAR TLR E-bar C
LEF
/C
TIP
Tadarida australis 4 37.9 (3.8) 8.33 (0.06) 13.8 (1.2) 0.092 (0.004) 1.37 (0.06) 0.0156 (0.0015) 0.104 (0.014)
Saccolaimus flaviventris 4 45.6 (5.6) 8.27 (0.26) 11.4 (0.6) 0.088 (0.006) 1.33 (0.06) 0.0034 (0.0006) 0.159 (0.008)
Chaerephon jobensis 3 23.4 (2.6) 8.04 (0.02) 13.1 (1.6) 0.083 (0.009) 1.30 (0.03) 0.0169 (0.0015) 0.083 (0.010)
Taphozous georgianus 4 25.4 (2.2) 7.74 (0.24) 9.0 (0.8) 0.123 (0.006) 1.40 (0.04) 0.0035 (0.0002) 0.159 (0.032)
Mormopterus beccarii 3 14.6 (1.9) 7.27 (0.12) 12.0 (1.2) 0.095 (0.010) 1.28 (0.05) 0.0102 (0.0017) 0.092 (0.010)
Chalinolobus gouldii 3 10.4 (1.7) 6.65 (0.26) 6.4 (0.7) 0.141 (0.021) 1.28 (0.04) 0.0016 (0.0001) 0.080 (0.003)
Scotorepens greyi 5 6.8 (0.9) 6.23 (0.07) 6.2 (0.4) 0.109 (0.007) 1.23 (0.02) 0.0025 (0.0001) 0.081 (0.007)
Vespadalus finlaysoni 4 5.5 (0.4) 6.17 (0.12) 5.19 (0.4) 0.133 (0.012) 1.11 (0.06) 0.0027 (0.0008) 0.076 (0.005)
Nyctophilus geoffroyi 5 5.4 (0.9) 5.73 (0.11) 4.7 (0.9) 0.123 (0.004) 1.14 (0.02) 0.0153 (0.0011) 0.092 (0.011)
A
Data for T. georgianus and C. jobensis are from the Pilbara (21°S, 120°E); data for T. australis and two M. beccarii are from Carnarvon Basin (25°S, 115°E).

Organisation of a desert bat fauna 535
Table 2. Free-flight echolocation call attributes of species comprising the microbat fauna of the Little Sandy Desert
Attributes are defined in Methods. Values are means with standard deviations shown in parentheses. N = number of individual bats, ‘Seq’ = number of
call sequences, ‘Calls/seq’ = average number of calls measured per sequence
Species Sample size Call attributes
N Seq Calls/seq F
min
(kHz) F
max
(kHz) Dur (sec) Frequency with peak no. of cycles
F
peak
(kHz) N Calls/seq
Tadarida australis
A
7 9 4.7 (1.9) 11.4 (0.9) 16.4 (4.6) 17.5 (2.8) 12.2 (1.4) 5 14.8 (11.7)
Saccolaimus flaviventris 4 4 6.3 (2.5) 18.3 (0.3) 22.8 (0.6) 11.4 (1.1) 18.4 (1.4) 11 9.4 (10.0)
Chaerephon jobensis
A
3 4 3.8 (1.5) 18.6 (1.3) 28.5 (5.3) 11.9 (2.7) 19.3 (1.8) 8 7.6 (6.0)
Taphozous georgianus 4 5 8.6 (6.3) 24.2 (0.4) 27.9 (1.1) 11.6 (1.8) 24.6 (0.8) 511.2 (5.4)
Mormopterus beccarii 4 5 7.4 (2.9) 25.5 (0.4) 29.3 (0.9) 11.2 (1.6) 26.3 (1.2) 7 8.7 (3.4)
Chalinolobus gouldii 4 7 5.0 (2.0) 31.7 (0.7) 43.7 (3.3) 4.7 (0.7) 33.7 (1.2) 411.0 (7.8)
Scotorepens greyi 6 10 5.7 (4.0) 36.7 (1.0) 59.4 (10.9) 5.9 (2.1) 39.9 (1.9), 5 13.0 (9.2)
Vespadelus finlaysoni 4 9 5.4 (4.7) 51.8 (0.7) 73.8 (10.9) 4.6 (1.4) 54.4 (1.6) 3 9.7 (5.6)
Nyctophilus geoffroyi 4 6 4.7 (2.3) 43.6 (0.9) 69.0 (7.9) 3.7 (0.8) 48.5 (1.7) 4 8.3 (3.4)
A
Includes data from the Pilbara.

536 N. L. McKenzie et al.
instance, Churchill (1998, p. 157) described their flight as undulating, slow, fluttering and
highly manoeuvrable, and characterising them as gleaners and aerial foragers close to
vegetation.
Open air superiority
Saccolaimus flaviventris and Taphozous georgianus used fast, semi-agile flight to catch
insects in the open and semi-open airspaces more than 5 m above or away from the side of
tree-canopies (OC, AC & BS/O: Table 3). They frequently used tight lateral turns to catch
prey, indicating agility.
Table 3. Number of observations in each foraging microhabitat, and derived ‘clutter’ estimates
Microhabitats were discriminated in terms of air space clutter, i.e. their obstruction to straight-ahead flight
(see Methods for definitions)
OC AC BS/O BS/A–IS Clutter
A
Tadarida australis 230001.0
Chaerephon jobensis
B
373001.1
Taphozous georgianus 31001.3
Saccolaimus flaviventris 76201.7
Mormopterus beccarii 715 4 01.9
Chalinolobus gouldii 11019 22.7
Scotorepens greyi 01118 22.7
Vespadalus finlaysoni 004123.8
Nyctophilus geoffroyi 003183.9
A
Each microhabitat was assigned an ordinal clutter value from 1 (OC) to 4 (BS/A–IS). Thus the average
clutter value for Saccolaimus flaviventris = (7×1 + 6×2 + 2×3 + 0×4) / (7 + 6 + 2 + 0) = 1.9
B
Pilbara data
Table 4. Usual foraging strategy from field observations and confirmed by literature review
a, authors’ field observations (see text). Literature citations are as follows: b, Richards (1995a); c,
Thompson (1991); d, Richards (1995b); e, Churchill (1998); f, Fenton (1982); g, Bullen and McKenzie
(2001), h, Menkhorst (1995); i, McKenzie (1995); j, Bonaccorso (1998); k, Taylor et al. (1987); l, Grant
(1991); m, McKenzie and Rolfe (1995); n, McKenzie et al. (1993); o, McKenzie et al. (1994); p, Richards
(2001); q, McKean and Thomson (1995)
Species Foraging strategy
1. Interceptor 2. Surface Air superiority
3. Agile 4. Highly agile
Chaerephon jobensis a, b, c, e, j
Mormopterus beccarii a, e, i, j
Tadarida australis a, c, d, e, g, m
Nyctophilus geoffroyi a, c, e, g, h,
k, l, m, n, o, p
Taphozous georgianus a, e, q
Saccolaimus flaviventris p a, e
Scotorepens greyi a, f, p
Chalinolobus gouldii h a, e, g, m
Vespadalus finlaysoni a, e

Organisation of a desert bat fauna 537
T. georgianus foraged along escarpments and gullies in hill country. We measured level,
open-air speeds of 7.2 m sec
–1
(s.d. = 0.7, n = 6) at Red Hill (22°06′S, 116°15′E), up to
10.0 m sec
–1
in a descent after daytime release of captured individuals, and speeds as low
as 4.4 m sec
–1
during abrupt manoeuvres to catch insects attracted by our floodlight. They
rolled to steep bank-angles (more than 60°, not ‘near vertical’) during these turns.
S. flaviventris searched for prey using straight flight with measured level air speeds of
7.4 (s.d. = 0.8, n = 7) m sec
–1
in OC microhabitats at Cooma and above riparian woodland
at Nanyuarri Pool near Beyondie. It also often hawked into AC and BS/O microhabitats
(Table 3), slowing to less than 4.0 m sec
–1
to do tight 360° turns with ‘steep’ bank-angles.
Air speeds of 5.2 (s.d. = 0.8, n = 5) m sec
–1
were recorded from individuals manoeuvering
tightly around and between tree canopies illuminated by floodlight at Savory Camp.
On the basis of these data we assigned both species to a medium-fast air superiority
foraging strategy in un-cluttered air spaces. Our observations are consistent with the ‘rapid
flight in open air with abrupt changes in direction to capture prey’ reported for Taphozous
from elsewhere in Australia (Churchill 1998 for Taphozous australis, and McKean and
Thomson 1995 for T. kapalgensis). Also consistent are the fast, above-canopy flight, ‘aerial
dogfights’, ‘spiraling’ and abrupt ‘veering’ reported by Churchill (1998) for S. flaviventris
(Table 4).
Semi-clutter and clutter air superiority
At Beyondie Soak in October 1996, Chalinolobus gouldii and Scotorepens greyi were
observed free-foraging over a floodlit pool and in mulga woodland nearby. Individuals were
captured in mist nets, then released with bioluminescent tags on the following night, after
their flight capabilities had been assessed in the observation tunnel.
Free-foraging flight by both species in the diffuse glow of a floodlight was also observed
around a patch of flowering river gums near Beyondie Homestead (24°47′S, 120°02′E) and
in the riparian corridor along Savory Creek. They both used medium-fast, agile flight as
they pursued insects in the semi-cluttered air spaces more than 2 m from the sides,
underside or top of the vegetation canopy (Table 3). Both species frequently rolled to near-
vertical bank-angles during abrupt horizontal turns. C. gouldii also used abrupt longitudinal
(pitching) manoeuvres in the form of ‘stall turns’. Free-foraging C. gouldii entered these
agile manoeuvres at measured air speeds of 4.4–6.4 m sec
–1
(n = 11), and S. greyi entered
them at 4.7–6.7 m sec
–1
(n = 17). Straight-and-level speeds maintained by individuals
released at night with bioluminescent tags were higher: C. gouldii averaged 6.1 m sec
–1
(s.d. = 0.8, n = 9, 2 males, 1 female) and S. greyi averaged 5.8 m sec
–1
(s.d. = 0.7, n = 7,
1 male, 1 female).
Flight speeds of two C. gouldii released into the observation tunnel were measured at
4.7–5.8 m sec
–1
(n = 16). They were unable to turn in the width of the tunnel at these
speeds, even by rolling to near-vertical angles of bank. The three S. greyi tested in the
tunnel employed similarly agile turns in attempts to reverse direction at measured speeds
of 4.4–5.6 m sec
–1
(n = 15), but collided with the mist-net forming the tunnel’s walls on
each occasion. C. gouldii managed to turn using ‘stall turns’ in which all forward speed
was lost during the pitch-up entry to the turn. Both species made successful turns by
slowing to 2–2.5 m sec
–1
(estimated), then skidding around ‘wings-level’ by using extreme
wing-beat flapping rates and amplitude to maintain lift at what appeared to be ‘near-stall’
speed.
We class both species as having a medium-speed, ‘air superiority’ foraging strategy. This
assignment is consistent with observations by Churchill (1998) and others (Table 4). For

538 N. L. McKenzie et al.
instance, Richards (2001) characterised S. greyi’s hunting strategy in north-eastern
Australia as fast aerial pursuit with abrupt changes in direction.
Vespadelus finlaysoni often began to hunt at dusk. Individuals were observed as they
foraged along scree slopes and gullies in the Durba Hills and Carnarvon Range, in an open
Mulga woodland at Nooloo Soak (22°52′S, 121°57′E), and along a breakaway at Cooma.
They were hunting in the cluttered air spaces within vegetation stands or while close-
contouring vegetation canopies, breakaway scarps, scree slopes and other surfaces. Flight
was BS/A–IS, sometimes BS/O (Table 3).
Straight-and-level flight speeds of three specimens captured and released with Cyalume
tags at the Savory camp were measured at 5.4 (s.d. = 0.5, n = 7) m sec
–1
, nearly as slow as
the ‘surface’ species Nyctophilus geoffroyi. However, turns by this tiny bat were often too
abrupt to recognise in the diffuse light available. Along the breakaway at Cooma, ‘wings-
vertical’ turns and ‘abrupt steep climbs to stall’ (stall turns) were observed at measured
entry-speeds of 5.0–5.6 m sec
–1
(n = 4), as at least two individuals pursued insects at dusk.
The line of flight was often zig-zag, a rapid, alternating series of near-vertical rolls in
opposite directions.
These observations accord with Churchill (1998), who described them as aerial foragers,
and their flight as ‘fluttery and fast, with erratic changes of direction’. We concluded that
V. finlaysoni has an air superiority strategy (Table 4), but uses lower flight speeds than the
two other species in this category.
Air interceptors
Three species were assigned to this group. More than 20 Mormopterus beccarii were
observed in floodlights as they foraging above river gum trees along a gully in the
Carnarvon Range (McKenzie and Burbidge 1979). Futher observations were made using
spotlights as individuals foraged over riparian vegetation in the Beyondie and Savory
survey areas (search–approach–intercept echolocation sequences were recorded in both
places). In general, M. beccarii favoured semi-open conditions (AC: Table 3). Tadarida
australis microhabitat at the Savory, Cooma and Beyondie survey areas was identified from
their intense 12-kHz echolocation sequences, heard high overhead (even the ‘feeding
buzzes’ are audible). Direct observations of T. australis OC flight behaviour had previously
been made using a pair of 240-V floodlights to illuminate the sky over an open gully in the
Carnarvon Ranges. Flight behaviour of Chaerephon jobensis was not observed in the study
area. Elsewhere in arid Western Australia, C. jobensis is most frequently found foraging
along the narrow bands of river gum trees that fringe the dry beds of ephemeral river
systems. It frequents OC microhabitats (Table 3), but also uses AC and even BS/O
situations on occasion (McKenzie and Muir 2000). Throughout their geographical ranges,
both T. australis and C. jobensis hunt insects in the unobstructed air-spaces found well
above the canopy and well above the ground in large clearings (Thompson 1991;
Bonaccorso 1998; Churchill 1998).
We collected data on flight speeds and foraging behaviour of M. beccarii in the Beyondie
and Savory survey areas (see above). We have subsequently collected these data for
C. jobensis from Red Hill (22°06′S, 116°15′E), and published quantitative measurements
of the agility and air speeds of T. australis from Coolgardie (Bullen and McKenzie 2001).
Both Red Hill and Coolgardie are in other arid regions of Western Australia, but well within
the population range of these large, mobile molossids.
When foraging, the flight of all three molossids was fast to very fast: Mormopterus
beccarii averaged 7.9 m sec
–1
(s.d. = 1.2, n = 9, N = 3 individuals), Tadarida australis

Organisation of a desert bat fauna 539
averaged 8.3 m sec
–1
(s.d. = 1.3, n = 531, N = 51), and Chaerephon jobensis averaged
6.7 m sec
–1
(s.d. = 1.2, n = 9, N = 4).
When hawking prey illuminated in the floodlight, and even when evading our spotlight
beam, all three species used straight or gently curved flight rather than tight or abrupt turns.
Horizontal turns were frequent, but neither T. australis nor C. jobensis were seen to exceed
a 45° bank-angle, and the bank-angles of Mormopterus beccarii were only marginally
steeper as they foraged along the illuminated canopy-edge of a patch of flowering river
gums near Beyondie Homestead. All three species used only gentle pitching manoeuvres,
although M. beccarii performed open wing-overs (a turn that imposes loads of ~2.3 g: Jones
and Rayner 1988).
Thompson (1991), Bonaccorso (1998) and Churchill (1998) characterised these three
species in similar terms: as fast-flying species with limited ability to turn. Thus, we have
classed all three species as air interceptors (Table 4), T. australis and C. jobensis in open
air-spaces and M. beccarii in semi-open spaces.
Analysis of ecological patterns
Table 5 shows the strong, significant correlations between aspects of airframe design,
echolocation, and foraging ecology of the microbats. Microbats that foraged in cluttered
air-spaces had higher echolocation call frequencies, but lower call durations, aspect ratios,
wing loadings and tail length ratios than did the bats of more open microhabitats (Table 5,
Fig. 2). In relation to foraging strategy, air-superiority bats (strategy = 3 and 4) had lower
E
bar
values than the ‘interceptors’ (strategy = 1), while the ‘surface’ species (strategy = 2,
N. geoffroyi) had high tail area ratios and high E
bar
(Fig. 3). Table 5 also shows that the
correlations fall into two sets, one related to microhabitat partitioning (= clutter), the other
to the different way species catch prey within their foraging microhabitats (= strategy).
There was virtually no overlap between the two sets (one weakly significant correlation
would be expected by chance alone in a matrix of this size).
Faunal organisation
The two perspectives on foraging ecology provide a basis for comparing the organisation of
the Little Sandy Desert fauna with faunas in surrounding arid regions (Table 6). The more
Table 5. Correlations between echolocation call, flight performance and flight control
parameters and foraging microhabitats or foraging strategies of Little Sandy Desert microbats
Kendall’s values are listed. NS, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001. See Table 3,
Table 4 and Methods for definitions of terms
Clutter F
peak
Strategy
Clutter – 0.82
**
0.47
NS
F
peak
0.82
**
–0.46
NS
F
min
0.82
**
0.99
***
0.46
NS
Call duration –0.99
***
–0.78
**
–0.46
NS
Aspect ratio –0.87
***
–0.94
***
–0.40
NS
Wing loading –0.82
**
–0.78
**
–0.58
*
Tail length ratio –0.74
**
–0.82
**
–0.31
NS
Ear area ratio –0.34
NS
–0.25
NS
–0.72
**
Ear length ratio –0.29
NS
–0.31
NS
–0.82
**
E
bar
–0.37
NS
–0.28
NS
–0.84
**
Tail area ratio 0.51
NS
0.53
*
0.56
*

540 N. L. McKenzie et al.
arid the region, the poorer the fauna, especially in bats of ‘cluttered’ microhabitats
(BS/A–IS) and those with ‘surface’ and ‘air superiority’ foraging strategies.
Discussion
Foraging ecology
The Little Sandy Desert’s extant bat fauna comprises nine microbat species, although
Chaerephon jobensis records are restricted to the desert’s western periphery. In general, our
observations on the foraging microhabitats, foraging strategies and flight capabilities of
Little Sandy Desert species are consistent with data from elsewhere in Australia. Such data
are summarised in Menkhorst (1995), Churchill (1998), Strahan (1995) and, for Western
Australian regions, in Bullen and McKenzie (2001), McKenzie and Muir (2000) and
McKenzie and Start (1989). There are three exceptions:
(1) Menkhorst’s observation that the flight of Chalinolobus gouldii is ‘not very
manoeuvrable’ (see Table 4), is at odds with Churchill’s observations (‘sudden zig-zag
changes in direction’), with video analyses by Bullen and McKenzie (2001), and with
observations made during this study.
(2) Richards (2001) classed Saccolaimus flaviventris as an aerial interceptor, whereas our
Little Sandy Desert observations indicate air superiority and accord with Churchill
(1998).
Fig. 2. Three-dimensional scatterplot of foraging microhabitat (‘clutter’) against average
aspect ratio and average wing loading (N m
–2
). Standard deviation values are provided in
Table 1. Species codes comprise the first letter of the genus and species names.

Organisation of a desert bat fauna 541
(3) In contrast to observations herein and from elsewhere in Australia (e.g. Dwyer 1965;
Fullard et al. 1991; Grant 1991; Richards 2001), south-eastern Australian populations
of Nyctophilus geoffroyi studied by Brigham et al. (1997) and O’Neill and Taylor (1986)
were seldom observed to glean from the ground or vegetation. However, the ‘slow,
manoeuvrable flight near and amongst vegetation’ they reported is consistent with Little
Sandy Desert observations. Both foraging modes are included in our ‘surface’ strategy
(see ‘Methods’).
Echolocation call parameters herein are consistent with data from free-flying bats
belonging to the relevant species from elsewhere in Australia (Jones and Corben 1993;
McKenzie et al. 1995b; McKenzie and Muir 2000). While variation in the search mode
calls among bat populations from different geographical regions has been demonstrated
(Parsons 1997; Barclay et al. 1999), it is usually small (Murray et al. 2001).
Previous studies have identified relationships between flight indices, echolocation
characteristics and resource partitioning in bats (eg McKenzie and Rolfe 1986; Aldridge
and Rautenbach 1987; McKenzie and Start 1989; Fenton 1990; Findley 1993). Thus, we
expected that clutter (as a measure of microhabitat use) would be closely related to various
air-frame ratios and echolocation parameters. Its relationship with aspect ratio, wing
loading and echolocation parameters is shown in Table 5. Tail length ratio, on its own,
correlated with clutter rather than strategy because of its second-order relationship to aspect
ratio – wing-chord is an important factor in both of these air-frame parameters.
Fig. 3. Three-dimensional scatterplot of foraging strategy against TAR and E
bar
(average
values). Log
10
(E
bar
) is used in the plot. Standard deviation values are provided in Table 1.
Species codes comprise the first letter of the genus and species names.

542 N. L. McKenzie et al.
While foraging microhabitats could be ordered in terms of differences in air-space
clutter, foraging strategies are based on divergent sets of behavioural attributes because they
categorise the different ways in which species hunt their prey. Consequently, they are not
easily arrayed along a single gradient in two-dimensional space. This is the reason Bullen
and McKenzie (2001) used Discriminant Function Analysis to investigate relationships
between foraging strategy and bat air-frame variables. However, the simple structure of the
Little Sandy Desert’s fauna allowed us to use the manoeuvrability and agility of the turning
manoeuvres we observed to roughly order species in terms of their foraging strategy. Thus,
we could apply non-parametric correlation analyses to untangle relationships between air-
frame/echolocation variables and foraging strategy/microhabitat classes. Among Little
Sandy Desert bats, foraging strategy was correlated with three tail and ear ratios, none of
which were correlated with foraging microhabitat (‘Clutter’, see Table 5). The same
relationships emerged from our previous study of Coolgardie bats (Bullen and McKenzie
2001).
The regular, near-linear relationships overt in Figs 4 and 5 indicate that both foraging
strategy and microhabitat use need to be taken into account if we are to untangle factors
determining the organisation of bat communities – the Little Sandy Desert’s fauna is not
Table 6. Microbat faunas of tropical arid regions of Western Australia
Updated from Strahan (1995) and records of the West Australian Museum
Foraging strategy and species Foraging
micro-
habitat
Pilbara Gascoyne Carnarvon Southern
Great
Sandy
Desert
Little
Sandy
Desert
Gibson
Desert
Interceptor
Tadarida australis OC ×× × ×××
Chaerephon jobensis OC ×× × ×
A
×
A
Mormopterus beccarii AC ×× × ×××
Mormopterus loriae AC ××
Air superiority – Agile
Saccolaimus flaviventris AC ×× × ××
Taphozous hilli OC ×× × ×
Taphozous georgianus OC ×××
Air superiority – Highly agile
Chalinolobus gouldii BS/O ×× × ×××
Scotorepens greyi BS/O ××××
Scotorepens balstoni BS/O ××
Chalinolobus. morio BS/A–IS ×
Rhinonicteris aurantius BS/A–IS ××
Ve s p a d e l u s finlaysoni BS/A–IS ×× × ×××
Surface
Nyctophilus geoffroyi BS/A–IS ×× × ×××
N. arnhemensis BS/A–IS ××
N. bifax BS/A–IS ××
Macroderma gigas BS/A–IS ××
Richness 16 12 11 9 9 7
Rainfall
B
(mm) 350 240 250 240 230 220
A
Peripheral records only.
B
Long-term annual average rainfall at geographic centroid of region.

Organisation of a desert bat fauna 543
dominated by ‘closely packed, overlapping species’ (cf. Findley and Black 1983; McKenzie
et al. 1989; Findley 1993). We suggest that both sets of eco-morphological surrogates are
needed to characterise differences in bat foraging niches. This could explain why sympatric
species of Myotis can have similar aspect ratios and wing loadings, but different diets and
‘primary habitats’ (e.g. Saunders and Barclay 1992; Arlettaz 1999).
Another airframe control parameter emerged as being important in differentiating the
foraging strategies of these desert bats. Three of the five air-superiority species in the Little
Sandy Desert fauna forage in at least partially cluttered air spaces, where extreme turns are
required (Ve s p a d e l u s finlaysoni, Chalinolobus gouldii and Scotorepens greyi). The gross
wing-twist mechanisms used to generate these turns create turbulent and/or separated air-
flows that increase drag (Bullen and McKenzie 2001). The two other air-superiority species
favour more open microhabitats (the emballonurids, Taphozous georgianus and
Saccolaimus flaviventris, in Tables 3 and 4). Both species have significantly wider leading-
edge wing-flaps (C
LEF
/C
TIP
) than their semi-clutter counterparts (Table 1). Wide leading-
edge flaps provide sensitive roll-control for manoeuvering at medium to high speeds, with
only a minimal drag-increment.
There are few published measurements of straight-line, free-flight speeds in Australian
bats. Mode values from Coolgardie, further south in Western Australia’s arid zone, provide
a useful cross-check, given that (1) data herein are based on small sample-sizes, (2) three
of the Little Sandy Desert species also occur in Coolgardie, and (3) three others have
closely related congenerics in the Desert. Table 7 summarises these data, and shows that
species’ straight-and-level flight-speed values from the two regions are comparable, even
for congenerics. Given that speed is tightly coupled to wing loading (e.g. Bullen and
McKenzie 2001), the higher flight speed of Coolgardie Chalinolobus gouldii, compared
with Little Sandy Desert C. gouldii, is consistent with their significantly higher wing
loading: 7.47 (s.d. = 0.73, n = 23) versus 6.38 (s.d. = 0.68, n = 3) N m
–2
, respectively.
Because they result from ecological processes, rather than being mere artifacts of
phylogeny, the relationships between echolocation call frequencies, airframe ratios,
foraging microhabitats and foraging strategies are functionally appropriate. For instance,
high-frequency sounds are quickly attenuated by the atmosphere. While they provide more
detailed imagery than low frequencies, it is at the expense of range. Thus, fast-flying
species that hunt in clutter (such as Vespadalus regulus) or semi-clutter (Chalinolobus
Table 7. Straight and level, free-flight speeds of Little Sandy Desert bats (mean) compared with
Coolgardie populations (mode, from Bullen and McKenzie 2001)
N = number of specimens; n = number of readings
Flight speeds (m sec
–1
)
Desert (s.d., n) Coolgardie (n, N)
Nyctophilus geoffroyi 5.3 (0.8, 7) 4.4 (213, 19) N. geoffroyi
Vespadalus finlaysoni
Scotorepens greyi
Chalinolobus gouldii
5.4 (0.5, 7)
5.8 (0.7, 7)
6.1 (0.8, 9)
5.3 (498, 35)
5.8 (262, 5)
7.5 (839, 69)
V. regulus
S. balstoni
C. gouldii
Taphozous georgianus
Saccolaimus flaviventris
7.2 (0.7, 6)
7.4 (0.8, 7)
–
–
–
–
Chaerephon jobensis
Mormopterus beccarii
Tadarida australis
6.7 (1.2, 9)
7.9 (1.2, 9)
8.3 (1.3, 51)
–
7.8 (509, 11)
8.6 (531, 51)
–
M. planiceps
T. australis

544 N. L. McKenzie et al.
gouldii and Scotorepens greyi) have search-mode calls with a wide band-width (Kingston
et al. 1999). They need the high-frequency component for detailed imagery to discriminate
prey, as well as an intense peak as low in the band as possible for range (Table 8). They also
need to have an agile airframe (low E
bar
) that can sustain the forces imposed by tight
manoeuvres at high speed (Figs 2, 3).
In contrast, species with a ‘surface’ foraging strategy, such as Nyctophilus geoffroyi, fly
slowly as they seek prey in and against clutter. Their requirement for manoeuvrability (and
high surplus lift) at low speeds implies a large ‘tail area ratio’, large E
bar
and minimal wing
loading, which make commuting any distance energetically expensive (Figs 2, 3). Their
need for high-resolution imagery implies high-frequency calls and short echolocation
distances (e.g. see Kingston et al. 1999). While a frequency-agile call improves their prey-
discrimination/identification (Table 8), a long-range component (a low-frequency peak of
high-intensity) would only alarm prey (Grant 1991).
Nevertheless, phylogenetic patterns are strong: clear family-level relationships to
foraging strategy and microhabitat are overt in Tables 7 and 8, implying that these
ecological specialisations ocurred early in the evolution of bats (cf. McKenzie et al. 1995a).
Faunal organisation
Investigations of community structure, such as guild studies, should emphasise the role of
species’ foraging method because of its potential importance in effecting differences in
resource-use (Hespenheide 1975; Simberloff and Dyan 1991). Eco-morphological
characters related to foraging method (food resource partitioning) can provide useful tools
for understanding the organisation of bat faunas because they average the effects of
dynamic processes over evolutionary time and allow variation and complexity to be
quantified (Findley 1993; McKenzie et al. 1995a). Given that species define niches
(Rosenzweig 1992), and that there are few foraging microhabitats for bats to partition in
regions with low vegetational complexity (McKenzie and Muir 2000), a species-poor bat
Table 8. Relationship between phylogeny, foraging ecology and search-mode echolocation
Standard deviations are shown in parentheses. Values for total bandwidth and average peak frquency are
derived from Table 2
Family and species Strategy and microhabitat Total bandwidth
(kHz) (F
max
– F
min
)
Average peak
frequency (kHz)
Molossidae I
NTERCEPTOR
Tadarida australis Open 5.0 (3.8) 12.2 (1.4)
Chaerephon jobensis Open 9.9 (4.3) 19.3 (1.8)
Mormopterus beccarii Semi-open 3.8 (0.8) 26.3 (1.2)
Emballonuridae
AIR SUPERIORITY – AGILE
Saccolaimus flaviventris Open 4.5 (0.4) 18.4 (1.4)
Taphozous georgianus Open 3.4 (1.6) 24.6 (0.8)
Vespertilionidae
AIR SUPERIORITY – HIGHLY AGILE
Chalinolobus gouldii Semi-clutter 12.0 (2.9) 33.7 (1.2)
Scotorepens greyi Semi-clutter 22.6 (10.8) 39.9 (1.9)
Vespadalus finlaysoni Clutter 22.0 (10.5) 54.4 (1.6)
Nyctophilinae
SURFACE
Nyctophilus geoffroyi Clutter 25.4 (8.2) 48.5 (1.7)
A
A
Broad-band call (from 43 to 70 kHz) with an ill-defined peak.

Organisation of a desert bat fauna 545
fauna might be expected in the Little Sandy Desert. In terms of environmental productivity,
regions such as this provide scant resources for slow-flying bats, whether manoeuvrable or
agile. These are the species that can forage in cluttered microhabitats. Their low aspect ratio
and low wing loading airframes with large tail area ratios induce high drag coefficients and
make long flights energetically expensive.
These deductions are consistent with the observed structure of bat faunas in Western
Australia’s tropical deserts. The Gibson, southern Great Sandy and Little Sandy Deserts all
have species-poor bat faunas, comprising 7–9 species (Table 6). Their faunas are dominated
by fast-flying species with low drag coefficients (high aspect ratios and wing loadings),
capable of efficiently foraging over large distances. For instance, seven of the nine species
in the Little Sandy Desert forage in open and semi-cluttered microhabitats using
‘interceptor’ or ‘air superiority’ strategies (Table 8). Chaerephon jobensis, the slowest of
these interceptors, is found only around the wetter periphery of the deserts. These deserts
are all poor in ‘surface’ and ‘air superiority’ species compared with more productive
regions of Western Australia. For instance, Nyctophilus geoffroyi is the only ‘surface’
species found in these deserts, and is confined to productive sites such as Dragon Tree Soak
in the Great Sandy Desert (Burbidge and McKenzie 1983), and riparian fringes along
valleys in ranges country. Similarly, only one ‘air superiority’ species hunts in the cluttered
microhabitats in these deserts, Vespadalus finlaysoni.
Less arid regions further west have richer microbat faunas than the deserts (Table 7),
with a higher proportion of ‘surface’ and ‘air superiority’ species that hunt in cluttered
microhabitats. Most of the additional species are confined to productive habitat mozaics
and/or depend on physiologically benign day-roosts in caverns. Nyctophilus bifax,
N. arnhemensis, Rhinonicteris aurantius, Chalinolobus morio and Macroderma gigas are
examples of these specialists. The Pilbara, for instance, receives higher rainfall than deserts
such as the Little Sandy (350 vs 230 mm annually), includes extensive areas of rugged,
cavernous uplands, and is traversed by major river systems with permanent pools fringed
by structurally complex riparian woodlands as well as extensive floodplains with fertile
soils and a variety of complex vegetations (Beard 1975). Unlike the deserts, the Pilbara,
Gascoyne and Carnarvon regions support an open-range pastoral industry that depends on
the productive soil mosaics.
Positive, exponential relationships between productivity and community diversity
(richness, evenness and degree of difference between species) are well recognised
(e.g. Rosenzweig and Abramsky 1993; Srivastrava and Lawton 1998; Nijs and Roy 2000),
particularly in the context of strong inter-specific competition. As well as the array of
‘specialist’ surface and air-superiority bats listed above, the Pilbara fauna includes an extra
interceptor (Mormopterus loriae). Like the other specialists, it is confined to a productive
landscape mosaic, in this case to mangroves (McKenzie and Start 1989). The absence of
these specialist components in faunas of the least productive regions of Western Australia’s
tropical arid-zone suggests a diversity–productivity model of faunal organisation that is
consistent with the ‘specialisation hypothesis’ (Srivastrava and Lawton 1998; Troumbis
et al. 2000).
Acknowledgments
A. A. Burbidge, A. H. Burbidge, P. G. Kendrick, W. P. Muir and J. K. Rolfe assisted in the
bat sampling program. N. K. Cooper facilitated our access to the bat collections held by the
Western Australian Museum. The authors thank Ian Abbott for helpful discussion, and
G. Fuller for lending his Doppler-shift meter.

546 N. L. McKenzie et al.
Funding for the 1996 fieldwork was provided by the Commonwealth through the
National Reserve System Cooperative Program of the Australian Nature Conservation
Agency (now Environment Australia), together with funds provided by the Western
Australian Department of Conservation.
References
Aldridge, H. D. J. N., and Rautenbach, I. L. (1987). Morphology, echolocation and resource partitioning in
insectivorous bats. Journal of Animal Ecology 56, 763–778.
Arlettaz, R. (1999). Habitat selection as a major resource partitioning mechanism between two sympatric
sibling species, Myotis myotis and Myotis blythii. Journal of Animal Ecology 68, 460–471.
Beard, J. S. (1975). ‘Pilbara. Explanatory Notes to Sheet 5, 1 : 1 000 000 Series. Vegetation Survey of
Western Australia.’ (University of Western Australia Press: Nedlands.)
Bonaccorso, F. J. (1998). ‘Bats of Papua New Guinea.’ (Conservation International: Washington, DC.)
Brigham, R. M., Francis, R. L., and Hamdorf, S. (1997). Microhabitat use by two species of Nyctophilus
bats: a test of ecomorphology theory. Australian Journal of Zoolgy 45, 553–560.
Bullen, R., and McKenzie, N. L. (2001). Bat airframe design – flight performance, stability and control in
relation to foraging ecology. Australian Journal of Zoology 49, 235–261.
Burbidge, A. A., and McKenzie, N. L. (Eds) (1983). Wildlife of the Great Sandy Desert. Wildlife Research
Bulletin Western Australia No. 12 (127 pp.). Department of Fisheries and Wildlife, Perth.
Barclay, R. M. R., Fullard, J. H., and Jacobs, D. S. (1999). Variation in the echolocation calls of the hoary
bat (Lasiurus cinereus): influence of body size, habitat structure, and geographic location. Canadian
Journal of Zoology 77, 530–534.
Churchill, S. (1998). ‘Australian Bats.’ (Reed New Holland: Sydney.)
Dwyer, P. D. (1965). Flight patterns of some eastern Australian bats. Victorian Naturalist 82, 36–41.
Fenton, M. B. (1982). Echolocation calls and patterns of hunting and habitat use of bats (Microchiroptera)
from Chillagoe, North Queensland. Australian Journal of Zoology 30, 417–425.
Fenton, M. B. (1990). The foraging ecology and behaviour of animal-eating bats. Canadian Journal of
Zoology 68, 411–422.
Findley, J. S. (1993). ‘Bats: A Community Perspective.’ (Cambridge University Press: Cambridge.)
Findley, J. S., and Black, H. L. (1983). Morphological and dietary structuring of a Zambian insectivorous
bat community. Ecology 64, 625–630.
Findley, J. S., and Wilson, D. E. (1982). Ecological significance of chiropteran morphology. In ‘Ecology of
Bats’. (Ed. T. H. Kunz.) pp. 243–260. (Plenum Press: New York.)
Fullard, J. H., Koehler, C., Surlykke, A., and McKenzie, N. L. (1991). Echolocation, ecology and flight
morphology of insectivorous bats (Chiroptera) in south-western Australia. Australian Journal of
Zoology 39, 427–238.
Grant, J. D. A. (1991). Prey location by two Australian long-eared bats, Nyctophilus gouldi and N. geoffroyi.
Australian Journal of Zoology 39, 45–56.
Hespenheide, H. A. (1975). Prey characteristics and predator niche width. In ‘Ecology and Evolution of
Communities’. (Eds M. L. Cody and J. M. Diamond.) pp. 158–180. (Belknap Press: Cambridge, MS.)
Jones, G., and Corben, C. (1993). Echolocation calls from six species of microchiropteran bats in south-
eastern Queensland. Australian Mammalogy 16, 35–38.
Jones, G., and Rayner, J. M. V. (1988). Flight performance, foraging tactics and echolocation in free-living
Daubenton’s bats Myotis daubentoni (Chiroptera: Vespertilionidae). Journal of the Zoological Society
of London 215, 113–132.
Kalko, E. K. V., and Schnitzler, H. U. (1993). Plasticity in echolocation signals of European pipistrelle bats
in search flight: implications for habitat-use and prey detection. Behavioural Ecology and Sociobiology
33, 415–428.
Kalko, E. K. V., and Schnitzler, H. U. (1998). How echolocating bats approach and acquire food. In ‘Bat
Biology and Conservation’. (Eds T. H. Kunz and P. A. Racey.) pp. 197–204. (Smithsonian Institution
Press: Washington.)
Kingston, T. G., Jones, G., Akbar, Z., and Kunz, T. H. (1999). Echolocation signal call design in
Kerivoulinae and Murininae (Chiroptera: Vespertilionidae) from Malaysia. Journal of Zoology
(London) 249, 359–374.
McKean, J., and Thomson, B. (1995). Arnhem sheathtail-bat Taphozous kapalgenis. In ‘The Mammals of
Australia’. (Ed. R. Strahan.) pp. 476–477. (Reed Books: Sydney.)

Organisation of a desert bat fauna 547
McKenzie, N. L. (1995). Beccari’s freetail-bat Mormopterus beccarii Peters 1881. In ‘The Mammals of
Australia’. (Ed. R. Strahan.) pp. 480–481. (Reed Books: Sydney.)
McKenzie, N. L., and Bullen, R. D. (2001). Bat airframe design: flight performance, stability and control
in relation to foraging ecology. Australian Journal of Zoology 49, 235–261.
McKenzie, N. L., and Burbidge, A. A. (Eds) (1979). The wildlife of some existing and proposed nature
reserves in the Gibson, Little Sandy and Great Victoria Deserts, Western Australia. Wildlife Research
Bulletin Western Australia No. 8 (36 pp.). Department of Fisheries and Wildlife, Perth.
McKenzie, N. L., and Muir, W. P. (2000). Bats of the southern Carnarvon Basin, Western Australia. Records
of the We s t e r n Australian Museum Supplement 61, 465–477.
McKenzie, N. L., and Rolfe, J. K. (1986). Structure of bat guilds in the Kimberley mangroves, Australia.
Journal of Animal Ecology 55, 401–420.
McKenzie, N. L., and Rolfe, J. K. (1995). Vertebrate fauna. Records of the We s te r n Australian Museum
Supplement 49, 31–65.
McKenzie, N. L., and Start, A. N. (1989). Structure of bat guilds in mangroves: disturbance and
determinism. In ‘Patterns in the Structure of Mammalian Communities’. (Eds D. W. Morris,
Z. Abramski, B. J. Fox, and R. Willig.) pp. 167–178. (Texas Tech University: Lubbock, TX.)
McKenzie, N. L., Rolfe, J. K., Hall, N. J., and Youngson, W. K. (1993). Vertebrate fauna. Records of the
We s t er n Australian Museum Supplement 42, 33–55.
McKenzie, N. L., Rolfe, J. K., and Youngson, W. K. (1994). Vertebrate fauna. Records of the Wes te rn
Australian Museum Supplement 47, 51–85.
McKenzie, N. L., Gunnell, A. C., Yani, M., and Williams, M. R. (1995a). Correspondence between flight
morphology and foraging ecology in some Palaeotropical bats. Australian Journal of Zoology 43,
241–257.
McKenzie, N. L., Fontanini, L., Lindus, N. V., and Williams, M. R. (1995b). Biological inventory of Koolan
Island, Western Australia. 2. Zoological notes. Records of the Wes ter n Australian Museum 17, 249–266.
Menkhorst, P. W. (1995). ‘Mammals of Victoria.’ (Oxford University Press: Melbourne.)
Murray, K. L., Britzke, E. R., and Robbins, L. W. (2001). Variation in the search-phase calls of bats. Journal
of Mammalogy 82, 728–737.
Neuweiler, G. (1984). Foraging echolocation and audition in bats. Naturwissenschaften 71, 446–455.
Nijs, I., and Roy, J. (2000). How important are species richness, species evenness and interspecific
differences to productivity? A mathematical model. Oikos 88, 57–66.
Norberg, U. M., and Rayner, J. M. V. (1987). Ecological morphology and flight in bats (Mammalia:
Chiroptera): wing adaptations, flight performance, foraging strategy and echolocation. Philosophical
Transactions of the Zoological Society of London B 316, 335–427.
O’Neill, M. G., and Taylor, R. J. (1986). Observations on the flight patterns and foraging behaviour of
Tasmanian bats. Australian Wildlife Research 13, 427–432.
Parsons, S. (1997). Search-phase echolocation calls of the New Zealand short-tailed bat (Mystacina
tuberculata) and long-tailed bat (Chalinolobus tuberculatus). Canadian Journal of Zoology 75,
1487–1494.
Richards, G. C. (1995a). Northern mastiff-bat Chaerephon jobensis (Miller 1902). In ‘The Mammals of
Australia’. (Ed. R. Strahan.) pp. 479–480. (Reed Books: Sydney.)
Richards, G. C. (1995b). White-striped freetail-bat Nyctinomus australis (Gray 1838). In ‘The Mammals of
Australia’. (Ed. R. Strahan.) pp. 487–488. (Reed Books: Sydney.)
Richards, G. C. (2001). Ecological and evolutionary determinants of bat community structure in south-
eastern Australian forests. Ph.D. Thesis, University of New South Wales.
Rosenzweig, M. L. (1992). Species diversity gradients: we know more and less than we thought. Journal of
Mammalogy 73, 715–730.
Rosenzweig, M. L., and Abramsky, Z. (1993). How are diversity and productivity related? In ‘Species
Diversity in Ecological Communities: Historical and Geographical Perspectives’. (Eds R. E. Ricklefs
and D. Schluter.) pp. 52–65. (University of Chicago Press: Chicago.)
Saunders, M. B., and Barclay, R. M. R. (1992). Ecomorphology of insectivorous bats: a test of predictions
using two morphologically similar species. Ecology 73, 1335–1345.
Schnitzler, H. U., Kalko, E., and Surlykke, A. (1987). The echolocation and hunting behaviour of the bat,
Pipistrellus kuhli. Journal of Comparative Physiology A 161, 267–274.
Simberloff, D., and Dyan, T. (1991). The guild concept and the structure of ecological communities. Annual
Review of Ecology and Systematics 22, 115–143.
Srivastrava, D. S., and Lawton, J. H. (1998). Why more productive sites have more species: an experimental
test of theory using tree-hole communities. The American Naturalist 152, 510–529.

548 N. L. McKenzie et al.
http://www.publish.csiro.au/journals/ajz
Strahan, R. (Ed.) (1995). ‘The Mammals of Australia.’ (Reed Books: Sydney.)
Taylor, R. J., O’Neill, M. G., and Reardon, T. (1987). Tasmanian bats: identification, distribution and
natural history. Papers and Proceedings of the Royal Society of Tasmania 121, 109–119.
Thompson, B. G. (1991). ‘A Field Guide to the Bats of the NorthernTerritory.’ (Conservation Commission
of the Northern Territory: Darwin.)
Troumbis, A. Y., Dimitrakopoulos, P. G., Siamantzioras, A. D., and Memtsas, D. (2000). Hidden diversity
patterns in mixed Mediterranean grasslands. Oikos 90, 549–559.
Manuscript received 23 April 2001; accepted 27 August 2002