-
Echolocation by Insect-Eating BatsAuthor(s): HANS-ULRICH
SCHNITZLER and ELISABETH K. V. KALKOReviewed work(s):Source:
BioScience, Vol. 51, No. 7 (July 2001), pp. 557-569Published by:
University of California Press on behalf of the American Institute
of Biological SciencesStable URL:
http://www.jstor.org/stable/10.1641/0006-3568%282001%29051%5B0557%3AEBIEB%5D2.0.CO%3B2
.Accessed: 10/02/2012 18:15
Your use of the JSTOR archive indicates your acceptance of the
Terms & Conditions of Use, available at
.http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars,
researchers, and students discover, use, and build upon a wide
range ofcontent in a trusted digital archive. We use information
technology and tools to increase productivity and facilitate new
formsof scholarship. For more information about JSTOR, please
contact [email protected].
University of California Press and American Institute of
Biological Sciences are collaborating with JSTOR todigitize,
preserve and extend access to BioScience.
http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=ucalhttp://www.jstor.org/action/showPublisher?publisherCode=aibshttp://www.jstor.org/stable/10.1641/0006-3568%282001%29051%5B0557%3AEBIEB%5D2.0.CO%3B2?origin=JSTOR-pdfhttp://www.jstor.org/stable/10.1641/0006-3568%282001%29051%5B0557%3AEBIEB%5D2.0.CO%3B2?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsp
-
July 2001 / Vol. 51 No. 7 BioScience 557
Articles
Bats (order Chiroptera) are ecologically more diversethan any
other group of mammals. Numerous mor-phological, physiological, and
behavioral adaptations of sen-sory and motor systems permit bats
access to a wide range ofhabitats and resources at night. The more
than 750 species ofthe suborder Microchiroptera occupy most
terrestrial habi-tats and climatic zones and exploit a great
variety of foods,ranging from insects and other arthropods, small
vertebrates,and blood to fruit, leaves, nectar, flowers, and
pollen. Echolo-cation is one of the adaptations that make bats so
successful.
Echolocating animals emit signals of high frequency
(mostlyultrasonic) and analyze the returning echoes to detect,
char-acterize, and localize the reflected objects.
Sophisticatedecholocation systems have evolved only in the bat
suborderMicrochiroptera and in dolphins. Less efficient systems
havebeen reported for a few species of the bat suborder
Megachi-roptera and for some birds (Henson and Schnitzler 1980).
Batsuse echolocation for orientation in space, that is, for
deter-mining their position relative to the echo-producing
envi-ronment. In addition, many bats, especially those that huntfor
flying insects, use echolocation to detect, identify, andlocalize
prey.
Bats use a wide variety of species-specific signal types
dif-fering in frequency structure, duration, and sound
pressurelevel (SPL). In addition, signal structure varies
dependingon the echolocation task confronting the bat. Search
signalsthat are emitted when bats search for prey differ from
approachsignals that are emitted when they approach prey.
The echolocation signals and hearing systems of bats arewell
adapted for gathering behaviorally relevant informa-tion (e.g.,
Schnitzler and Henson 1980, Neuweiler 1989, Fen-ton 1990, Denzinger
et al. forthcoming). In this article we de-scribe the echolocation
behavior of insect-eating bats and showhow differing circumstances
such as habitat type, foragingmode, and diet favor different signal
types. To demonstraterelationships between echolocation and
ecological condi-tions, we outline the perceptual tasks that must
be performedby foraging bats and discuss the suitability of typical
ele-ments of echolocation signals for solving such problems. Wethen
define habitat types according to the problems they im-pose on bats
and relate the observed variability in signalstructure to
ecological constraints set by habitat type andforaging mode.
Perceptual problems for foraging batsForaging bats confront a
multitude of problems when flyingto their hunting grounds and
searching for prey. These prob-lems differ depending on where bats
hunt, what they eat,and how they acquire their food. For example,
bats huntingfor insects in the open encounter conditions different
fromthose that search for prey near the edges of vegetation, in
veg-etation gaps, in dense forest, or near the ground. The
prob-lems also differ depending upon whether they capture mov-ing
prey in flight (aerial mode) or mostly stationary preyfrom surfaces
such as leaves or ground (gleaning mode) or wa-ter (trawling
mode).
Foraging bats must detect, classify, and localize an insect
anddiscriminate between echoes of prey and echoes of
unwantedtargets such as twigs, foliage, or the ground, referred to
as clut-ter echoes, or simply clutter.For many bats echolocation
de-livers all of the information they need to catch an insect.
Hans-Ulrich Schnitzler (e-mail:
Hans-Ulrich.Schnitzler@uni-tuebin-
gen.de) is professor and head of the Lehrstuhl Tierphysiologie
of the
University of Tbingen, Auf der Morgenstelle 28, D-72076
Tbingen,
Germany. Elisabeth Kalko is professor and head of the Abteilung
Ex-
perimentelle kologie of the University of Ulm, Albert Einstein
Allee
11, D89069 Ulm. E. Kalko is a staff member and H.-U.
Schnitzler
is a research associate of the Smithsonian Tropical Research
Insti-
tute, Panama. E. Kalko is also a research associate at the
National
Museum of Natural History in Washington, DC. 2001 American
In-
stitute of Biological Sciences.
Echolocation by Insect-Eating BatsHANS-ULRICH SCHNITZLER AND
ELISABETH K. V. KALKO
WE DEFINE FOUR DISTINCT FUNCTIONAL
GROUPS OF BATS AND FIND DIFFERENCES
IN SIGNAL STRUCTURE THAT
CORRELATE WITH THE TYPICAL ECHOLO-
CATION TASKS FACED BY EACH GROUP
-
558 BioScience July 2001 / Vol. 51 No. 7
Articles
However, some bats rely partly or entirely on other
sensorysystems, such as passive hearing, to detect
prey-generatedsignals (e.g., calls and rustling sounds of insects)
to find theirprey. Independent of specific foraging conditions, all
batsmust perform the following basic perceptual tasks:
Detection.A bat must decide whether or not it has receivedan
echo of its own echolocation signal or has heard, seen,smelled, or
felt something that indicates prey or othertargets of interest.
However, it is difficult to conceptualizedetection independent of
classification and localization.
Classification. Bats categorize targets by means of specificecho
information (Ostwald et al. 1988) or other featuresthat reveal
their nature. Target properties such as size,shape, material, and
texture are encoded in the complextemporal and spectral parameters
of an echo.Rhythmical amplitude and frequency modu-lations in the
echo reveal movements of preysuch as the beating wings
characteristic of afluttering insect (Schnitzler 1987).
Localization. Echolocation reveals the positionof a target by
its range and by its horizontal andvertical spatial angles. The
time delay betweenemitted signal and returning echo encodesrange.
Binaural echo cues describe the hori-zontal angle, and monaural
spectral cues thevertical angle. For moving bats, the flow fieldof
reflected sound delivers additional infor-mation that may be used
for target localization(Lee et al. 1992, Mller and Schnitzler
1999,2000). Bats using other sensory cues must lo-calize the actual
position of the source of aprey-generated sensory signal.
Interfering factors, such as internal and exter-nal noise,
clutter echoes, and signals from otherbats, set limits on the
echolocation processes in-volved in detecting, classifying, and
localizing apreferred target. Masking effects between targetecho
and clutter, and between target echo andemitted signal, notably
restrict the processing ofrelevant information.
The restricted range of echolocation sets spa-tial limits on
where bats can find their prey. Thesound pressure level, or SPL, of
echoes decreasessharply with increasing target distance because
ofgeometric and atmospheric attenuation of soundtraveling in air.
Additionally, the echo SPL is re-duced because of the target
strength, which de-pends on the size and the form of a target.
Batsreceive detectable echoes from flying insects onlyover rather
short distances. In an environmentwith a temperature of 20C, a
relative humidity of50%, and a realistic detection threshold of
about
15 dB, a bat with a signal SPL of 112 dB 40 cm in front of
itshead, with a signal frequency of 20 kHz, can detect a
flutter-ing insect with a wing length of 2.5 cm at a detection
rangeof not more than 10.5 m. Under similar conditions, a
spherewith a diameter of 2.5 cm cannot be detected beyond 7 m.
Themaximum detection distance decreases still more with in-creasing
signal frequency, humidity, temperature, and de-creasing prey size
(Kober and Schnitzler 1990). Thus, echolo-cation is a system that
works only over short distances. Forlong-distance orientation, bats
must use other sensory systems,such as vision.
Signals adapted for specific tasksEcological constraints exert
strong selection pressure on sig-nal structure, leading to
species-specific signals that areadapted for specific tasks. To
better understand these
Figure 1. Search and approach signals of foraging bats. (ae)
Signals ofbats that captured a flying insect at the end of the
sequence. (f, g) Signalsof bats that gleaned insects from a surface
at the end of the sequence, (f)out of continuous search flight and
(g) after the prey has been detectedfrom a perch. In all sequences
the increase in repetition rate and thereduction of sound duration
indicate the switching from search toapproach phase. Note the
distinct terminal phase in bats that capturedflying insects
(ae).
-
adaptations, we first discuss the kinds of information thatcan
be carried by individual elements in the echolocationsignals of
bats.
Most echolocation signals of microchiropterans consistof
narrowband or broadband components, or combinationsof these (Figure
1). Narrowband components comprise twosubtypes: quasi-constant
frequency (QCF) elements withfrequency changes of a few kHz between
the onset and the endof the component (shallow modulation), and
long constantfrequency (CF) elements with frequency changes of a
fewhundred Hz within the component. Broadband componentsnormally
consist of a downward frequency-modulated (FM)element of large
bandwidth (steep modulation). These signalelements differ in
absolute frequency, bandwidth, harmonicstructure, duration, and
SPL, creating the wide variety ofsignal types found in echolocating
bats (reviewed in Pye1980, Schnitzler and Henson 1980, Simmons and
Stein 1980,Neuweiler 1989, Fenton 1990). For the classification of
signalcomponents according to their bandwidth, it is necessary
todefine why a signal is classified as narrowband or broad-band.We
propose that in narrowband signals the most promi-nent harmonic
sweeps over less than half an octave (startingfrequency of the
sweep is less than 141% of the terminal fre-quency), whereas in
broadband signals it covers more thanhalf an octave.
The information that can be extracted from the echoes ofvarious
elements of a signal depends on their physical struc-ture and on
the performance of the bats auditory system. Asa rough
approximation, the input stage of the auditory sys-tem of most bats
(with the exception of bats with long CF sig-nals) can be described
as a bank of neuronal filters with sim-ilar Q values (the best
frequency, or the frequency at whichneural responses have the
lowest threshold, divided by band-width). Depending on its absolute
frequency, bandwidth,and SPL, a signal element evokes neuronal
activity in oneor more of the frequency-selective filters. The
longer a signalstays within the response range of a filter, the
higher theneuronal activity.
Narrowband signals. Signal elements such as CF or
shal-low-modulated QCF components are well suited for detec-tion of
echoes because they activate the neuronal filters tunedto the
corresponding frequency band during the entire echo.Narrowband
signals, especially those of long duration, also canbe used for
target classification if bats evaluate the amplitudeand frequency
modulations in the echoes arising from char-acteristic target
movements.When a signal hits a fluttering in-sect at the favorable
instant when the insects wings are per-pendicular to the impinging
sound wave, a short and veryprominent amplitude peak in the echo,
an acoustic glint, re-veals the fluttering insect target. This
glint, which can be upto 2030 dB stronger than an echo from the
body of the in-sect, also increases the probability of detection
(Kober andSchnitzler 1990, Moss and Zagneski 1994). The probability
ofreceiving such a glint depends on the duty cycle (the
percentageof time in which signals are emitted) of the bat and the
wing-
beat rate of the insect. For example, a duty cycle of 10% andan
insect wingbeat rate of 60 Hz produce an average perceivedglint
rate of 6 glints/s.
Narrowband signals are less suited for precisely localizinga
target when bats must accurately measure range as well ashorizontal
and vertical angles. Range is encoded in the timedelay between an
emitted signal and its returning echo. Foraccurate range
determination, bats must determine the exactinstant of sound
emission and echo reception. Narrowbandsignals are rather imprecise
time markers because they per-sist within the corresponding
neuronal filter for an extendedtime, thus diminishing range
accuracy. The horizontal angleis encoded in binaural echo cues, and
the vertical angle inmonaural echo cues. Narrowband signals, with
their small fre-quency range, activate only a few channels that
deliver suchcues, thus reducing a precise angle determination.
Broadband signals. Signals such as steep-modulated FMsignals are
less suited for the detection of weak echoes. Thesesignals sweep
rapidly through the tuning areas of the corre-sponding neuronal
filters so that each detector is activated onlyfor a very short
time. Frequency-modulated signals of broadbandwidth are well suited
for exact target localization whererange and angle must be measured
accurately. Steep-modu-lated broadband FM signals activate each
filter for only avery short instant, producing the precise time
markers neededfor an exact determination of the time delay that
encodes therange. Large signal bandwidths activate more neuronal
filters,improving the accuracy of range determination if, as is
indi-cated by behavioral experiments, the range information is
av-eraged over all activated channels (Moss and Schnitzler
1995).The activation of many channels also improves the accuracyof
angle determination with increasing bandwidth.
Frequency-modulated signals of large bandwidth also de-liver
spectral cues that can be used for target classification. Tar-get
features such as texture (which affects the absorption ofsound at
different frequencies) and depth structure (whichcauses an
interference pattern due to overlapping multi-wave-front echoes)
are somehow reflected in the echo spec-trum, thus encoding
information about the character of a tar-get (Ostwald et al. 1988).
In the laboratory, bats learn to usesuch spectral differences to
discriminate among targets, so itis assumed that broadband FM
signals allow spectral char-acterization of prey (Neuweiler 1989,
1990). However, this istrue only if the spectral signature of
echoes is so specificthat, independent of aspect angle, it is
possible for the echofrom an insect to be distinguished from
clutter echoes.
Long CF signals with Doppler-shift compensa-tion. Long CF
signals, like other narrowband signals, are wellsuited for the
detection of weak echoes and less well suited forthe exact
localization of targets. In combination with Doppler-shift
compensation (see below) and a specialized hearingsystem, the long
CF signals facilitate the detection and clas-sification of
fluttering insects in a cluttered environment.The beating wings of
insects produce a rhythmical pattern of
July 2001 / Vol. 51 No. 7 BioScience 559
Articles
-
amplitude and frequency modulations that encode wingbeatrate,
wing size, and other species-specific information. Themost
prominent flutter features are the very short and strongamplitude
peaks of acoustic glints, produced when the wingsare perpendicular
to the impinging sound waves (Kober andSchnitzler 1990).
Transmitters and receivers of the echoloca-tion systems of bats
emitting CF signals are especially adaptedto process this kind of
information (Schnitzler and Ostwald1983, Neuweiler 1990). By
lowering emission frequency, thesebats compensate for Doppler
shifts caused by their own flightmovement. Thus, the frequency of
the CF component of in-sect echoes is kept within an expectation
window.A corre-sponding analysis windowis established in the
hearing sys-tem by a specialized cochlea with a highly expanded
frequencyrepresentation in the range of the insect echoes. This
acousticfovea leads to an over-representation of sharply tuned
neu-rons with special response characteristics throughout the
au-ditory pathway. With these specific adaptations, bats using
CFsignals can discriminate the modulated insect echo
fromoverlapping unmodulated clutter echoes and classify
insectsaccording to their prey-specific modulationpattern (vd Emde
and Schnitzler 1990).
Detection versus localization trade-off. Narrowbandsignals are
good for target detection but less well suited for tar-get
localization. Broadband FM signals, however, aregood for
localization but less well suited for detec-tion. This trade-off
between detectability and accuracyof localization is reflected in
the structure of search sig-nals in some bats. Bats that have to
perform several taskssimultaneously combine suitable signal
elements. Typ-ical combinations of broadband and
narrowbandcomponents are found in FMQCF, QCFFM, andCFFM signals,
and also in FM signals with changingsteepness. For instance, bats
of the genus Myotis (Ves-pertilionidae) flying in the open often
produce signalsin which the steep initial part of the signal is
followedby a shallower part that ends in a steeper segment.
Thesteeper FM components are well suited for localization,and the
shallow component improves detection byintroducing more signal
energy into the correspond-ing neuronal filters.
The masking problemThe separation of target echo from
interfering signalsis an important task facing echolocating bats.
Theevaluation of sonar echoes from a target is hamperedwhen the
neuronal activity evoked by clutter echoes andby the bats own
emitted signal interferes with theactivity evoked by the target
echo. Interfering signalsthat precede the target echo, such as the
emittedsignal, produce a forward-masking effect. Interferingsignals
that follow the target echo, such as clutterechoes, produce a
backward-masking effect. Depend-ing on the signal type, several
strategies are used toavoid masking.
Field and laboratory studies indicate that bats (with the
ex-ception of those using long CFFM signals) normally avoidan
overlap of the target echo with clutter echoes and also withtheir
own emitted signal (Kalko and Schnitzler 1989, 1993).This avoidance
of overlap suggests that all signal types exceptfor CF elements are
sensitive to overlap. Due to overlap in-terference, the ability of
bats to evaluate insect echoes dependson the position of an insect
relative to the bat and to clutterecho producing background
targets. When an insect flies soclose to a bat that the returning
echo overlaps the emitted sig-nal, forward-masking effects
interfere with the evaluation ofthe insect echo. Therefore forward
masking reduces the prob-ability of detection in the zone in front
of the bat where over-lap occurs. The width of this signal-overlap
zone depends onsignal duration (Figure 2). For example, with a
signal dura-tion of 10 ms the overlap zone is 1.70 m wide at a
speed ofsound of 340 m/s. If undisturbed detection is only
possiblebeyond this signal-overlap zone, signal duration sets a
min-imum detection distance. Each ms of signal duration adds 17cm
to this minimum detection distance.
When an insect flies so close to clutter that its echoes
over-lap the clutter echoes, backward masking reduces the
prob-ability of detection in this clutter-overlap zone. The width
ofthis clutter-overlap zone is determined also by signal
duration.Only insects flying far away from the bat and from the
clut-
560 BioScience July 2001 / Vol. 51 No. 7
Articles
Figure 2. Schematic diagram of the masking situation for a
batforaging near vegetation and emitting signals with a duration of
6 ms.The prey echo overlaps the emitted signal when the insect
flies in thesignal-overlap zone, and it overlaps the clutter echoes
when it flies inthe clutter-overlap zone. No overlap occurs when
the insect flies in theoverlap-free window. At a distance of 2 m
the overlap-free window isclosed, and for the given signal duration
the bat has reached theminimum gap size where overlap-free
echolocation is impossible.
-
ter-producing background where no overlap occurs can be
de-tected without interference. By our definition, these insects
flyin an overlap-free window (Figure 2). The forward-
andbackward-masking effects strongly depend on signal structure,on
the SPL of prey echo and masking signals, and on their tem-poral
relationship. Therefore the definition of an overlapzone as an area
where masking effects interfere with theevaluation of the echo is
only a rough approximation of theeffective masking zones. Depending
on the signal type usedby a bat, the masking zone may be smaller
than the overlapzone calculated from sound duration. For example, a
studyon Myotis nattereri, a bat that searches for prey using
wide-band FM signals very close to vegetation, indicates that
thesebats tolerate some overlap between prey and clutter
echoes(Siemers and Schnitzler 2000). Nevertheless, the
calculationof the overlap zone is a useful tool to judge the
dangerof masking.
In bats with long CF signals, the CF component of the emit-ted
signal and that of the returning echo often overlap (Fig-ure 3d).
This overlap produces no masking effect, becauseDoppler-shift
compensation keeps the target echo in therange of the extremely
sharply tuned neurons of the acousticfovea, whereas the emitted
signal is lower in frequency and fallsin a range where the auditory
threshold is high (Schnitzler andOstwald 1983, Neuweiler 1990).
Therefore, long CF compo-nents of Doppler-compensating bats are not
vulnerable tooverlap.
Foraging habitats defined by clutter conditionsComparative
studies reveal that for foraging bats, clutterconditions are the
most important ecological constraints.They can be used to define
various foraging habitats (Schnit-zler and Kalko 1998, Denzinger et
al. forthcoming). Clutterconditions are described by the proximity
of the desired preyitem to clutter such as vegetation or ground.
Such clutterrepresents perceptual as well as mechanical problems
for bats(Fenton 1990). Perceptually, bats are constrained by
theirsensory capacities (e.g., echolocation, vision, olfaction,
pas-sive listening) to detect, classify, and locate prey in the
vicin-ity of clutter-producing background targets. Mechanically,
batsare constrained by their motor capacities, such as flight
abil-ities (Norberg and Rayner 1987). For instance, bats that
for-age near clutter need special maneuverability (e.g.,
adaptationsin wing morphology) to intercept insects while also
avoidingcollisions. Here we discuss only the perceptual problem,
andwe define three habitat types according to clutter
conditions(Figures 3 and 4).
Uncluttered space. For bats that forage for insects
inuncluttered spacein open spaces, high above the ground andfar
from vegetationclutter echoes from the backgroundare so far from
the emitted signal and target echoes that theyplay no role in the
echolocation process (Figures 3a and 4).In these bats a returning
echo generally indicates a flyinginsect.
Background-cluttered space. For bats that forage forinsects in
background-cluttered spacenear the edges ofvegetation, in
vegetation gaps, or near ground or water sur-facesthe pulseinsect
echo pair is followed by clutter echoesfrom the background (Figures
3b and 4). These bats must solvetwo problems. First, they must
recognize the insect echo andseparate it from the echoes of
background clutter. Second, theymust characterize the
clutter-producing background targetsto identify landmarks for
navigation and to avoid collision.
The border between background-cluttered and unclut-tered space
is indicated in many species of bats by theirecholocation behavior
(Kalko and Schnitzler 1993, Schnitzlerand Kalko 1998, Jensen and
Miller 1999). In uncluttered
July 2001 / Vol. 51 No. 7 BioScience 561
Articles
Figure 3. Schematic diagram of the input into theauditory system
of bats that forage in different cluttersituations. The emitted
pulse and the returning insectecho are depicted in black. (a) In
uncluttered space, thepulse echopair is far from clutter echoes.
(b) Inbackground-cluttered space, the pulseecho pair isfollowed by
clutter echoes (depicted in white). (c, d) Inhighly cluttered
space, the target echo is buried inoverlapping clutter echoes.
Sound duration and envelopeform correspond to search signals
typical for the differentspaces: (a) QCF signal of an open-space
forager; (b)broadband FMQCF signal of an edge and gap forager;(c)
broadband FM signal of a narrow-space FMforager; (d) long CFFM
signal of a narrow-space CFforager; the echo shows amplitude
modulations, or glints,created by the beating wings of an
insect.
-
space bats often use relatively long search signals of
narrowbandwidth and mostly emit a search signal only every secondor
third wingbeat. In background-cluttered space most batsemit one
search signal of larger bandwidth at every wingbeat,thus indicating
that echoes from the clutter-producing back-ground also guide bats
behavior. When crossing the borderbetween the two spaces, bats
change their echolocationbehavior (see signals of Pipistrellus in
Figures 4 and 5). Thebest indicator for the switch between the two
spaces is thechange in the rhythm of sound emission.
Highly cluttered space. For bats that forage for flyinginsects
in highly cluttered spacevery close to surfaces suchas leaves or
groundand for bats gleaning stationary food (sit-ting insects,
other animals, fruit, leaves, nectar, blood) fromsurfaces, two
situations occur. For gleaning bats that useshort broadband FM
signals, echoes from food items areburied in clutter (Figure 3c) so
that masking hampers theirevaluation. For bats using long CF
signals to detect flying in-sects, the overlapping pulseecho pair
also overlaps clutterechoes (Figure 3d). Bats foraging in highly
cluttered space have
the problem of discriminating between echoes from the fooditem
and overlapping clutter echoes. Moreover, they have tonavigate
along landmarks and to avoid collisions. The relationbetween insect
echoes and masking clutter echoes definesthe border between highly
cluttered and background-clutteredspace. A bat forages in highly
cluttered space when the preyis situated in the clutter-overlap
zone, where clutter echoesoverlap prey echoes. In the few species
that tolerate someoverlap between prey and clutter, highly
cluttered space cor-responds more precisely to the zone where the
prey echo ismasked by the clutter echoes. Furthermore, bats flying
par-allel to clutter may be able to reduce the masking effect of
clut-ter echoes by spatial clutter rejection. However, to
facilitatecomparison between species, we propose to define the
highlycluttered space as the area where insect echo and
clutterechoes overlap as a first approximation to describe the
mask-ing problem.
Guild structure of batsComparative studies have revealed that
bats foraging in sim-ilar habitats with similar foraging modes for
similar diets
562 BioScience July 2001 / Vol. 51 No. 7
Articles
Figure 4. Schematic diagram of the foraging habitats of bats
according to the clutter situation, with flight silhouettes
andsearch signals of representative species. In uncluttered space,
bats forage for prey far away from vegetation and the ground.In
background-cluttered space, bats hunt for insects flying near
obstacles (e.g., at edges of vegetation, near the ground orwater
surfaces, and in gaps between and in vegetation). In highly
cluttered space, bats forage for prey flying close to orsitting on
vegetation or the ground. The bats depicted are not sympatric. The
border between uncluttered and background-cluttered space is
defined by the echolocation behavior of bats. When entering the
uncluttered space from background-cluttered space, foragers switch
from broadband signals to narrowband signals, and vice versa (e.g.,
Pipistrellus kuhli). Inthis diagram this border is about 5 m away
from vegetation and the ground as described for pipistrelles. The
border betweenbackground and highly cluttered space is determined
by the beginning of the clutter-overlap zone in which insect
echoesoverlap clutter echoes.
-
encounter similar ecological constraints and share
similaradaptations of sensory and motor systems. In this
context,arranging assemblages of species into guilds, that is,
groupsof species that live under similar ecological conditions
(Root1967), has proven a useful approach toward
understandingadaptations of sensory and motor systems. Habitat
type, for-aging mode, and diet characterize our proposed guilds for
bats(Kalko et al. 1996, Schnitzler and Kalko 1998).
To categorize habitats we use the three clutter spaces as
de-fined above. For foraging modes we distinguish bats thatcapture
insects in the air (aerial mode) from those that gleaninsects and
other food from solid surfaces (gleaning mode)or from water
(trawling mode). According to their preferreddiet we categorize
bats as insectivores, carnivores, piscivores,sanguivores,
frugivores, nectarivores, or omnivores. Theseguilds characterize
ecological constraints as sets of tasks im-posed on sensory as well
as on motor systems. The catego-rization presented here differs
somewhat from that in formerpublications (Schnitzler and Kalko
1998). In light of newdata (Boonman et al. 1998, Rydell et al.
1999), we propose anadditional foraging mode (trawling) and
introduce a newguild background-cluttered space trawling
insectivore/pis-civore, which replaces the former guild highly
clutteredspace gleaning piscivore (Table 1).
Search signals of insect-eating batsIn describing the
echolocation behavior of insect-eating bats,we will show that bats
belonging to the same guild sharemany similarities in echolocation
behavior, especially in thestructure of search signals, which are
intimately linkedto habitat type and foraging mode. Search signals
areemitted when bats are searching for prey, or when theycommute
from one place to another and do not approach aspecific target.
Aerial insectivores in uncluttered space. Bats hunt-ing in open,
uncluttered space, high above the ground orcanopy and far from
obstacles, are found mainly in four fam-ilies: free-tailed bats
(Molossidae, e.g., Molossus, Promops,Tadarida), mouse-tailed bats
(Rhinopomatidae, Rhinopoma),sheath-tailed bats (Emballonuridae,
e.g., Diclidurus, Per-opteryx, Taphozous), and evening bats
(Vespertilionidae, e.g.,Lasiurus, Nyctalus; Figures 1a and 1b, 4,
and 5a).When search-ing for insects, these bats have no masking
problem as longas the emitted signal does not overlap the returning
insect echo.However, they often have the problem of rather small
prey be-ing sparsely distributed in a large space. In that case, a
bat mustcover a large search area to find an insect. This and the
ratherlow SPL of insect echoes make it difficult to detect
potentialprey. Thus, echolocation signals should be optimized
fordetection.
Typically, bats flying in the open emit overlap-sensitive,
nar-rowband search signals of ratherlong duration (approximetely825
ms) with a low frequency (half an octave) signals of intermediate
duration (approx-imetely 310 ms), with a medium frequency
narrowbandcomponent (approximately 3060 kHz). Signal emission
isusually correlated in a 1:1 ratio with wingbeat. Pulse intervalis
shorter (approximetely 70150 ms) than that produced byaerial
insectivores in uncluttered space.
The narrowband components of the mixed signals greatlyfacilitate
detection of prey, and the broadband FM compo-nents are well suited
for the localization and characterizationof extended background
targets necessary for recognizinglandmarks and avoiding collisions.
Because of the sensitivityto overlap, bats foraging near edges
mainly search for and de-tect prey within the overlap-free window,
where the preyecho is not masked by the emitted call
(forward-masking ef-fect) and where echoes from the background do
not interferewith the prey echo (backward-masking effect). Field
studiesshow that the shorter the signals, the closer to vegetation
a batcan hunt (e.g., Kalko and Schnitzler 1993).
Some bats are highly flexible in their foraging and
echolo-cation behavior and switch between uncluttered and
back-ground-cluttered space. In uncluttered space the
echolocationbehavior does not change depending on distance to
back-ground targets, whereas in background-cluttered space
batsreact to the background targets. The border between thosespaces
is defined by changes in the echolocation behavior ofthe bats. For
instance, pipistrelles hunting for insects in openspaces emit
rather long, shallow-modulated, narrowbandsignals (QCF), whereas
those hunting within less than 5 m ofclutter-producing background
switch to shorter signalsand add a steep-modulated, broadband
component(FMQCF). Some preliminary field data from the big brownbat
(Eptesicus fuscus) and the noctule bat (Nyctalus noctula)suggest
that the transition distance between spaces isspecies specific.
564 BioScience July 2001 / Vol. 51 No. 7
Articles
-
A special group within background-cluttered space insec-tivores
consists of bats that catch insects in the trawling modefrom calm
and uncluttered water surfaces (e.g., Myotis dauben-tonii). They do
not receive clutter echoes from the water (butfrom background
targets such as the shore) because a calmsurface is like a mirror
that reflects the emitted signals awayfrom the bat (Rydell et al.
1999). Therefore these trawling batscan be categorized as
background-cluttered space trawlinginsectivores(Table 1). When
foraging for insects sitting onor flying close to duckweed or
rippled water, bats have diffi-culty perceiving prey as the prey
echo is buried in clutterechoes (Boonman et al. 1998). The two
species of bulldog bats(Noctilionidae), one of which, Noctilio
leporinus, is wellknown for its fishing habits (Schnitzler et al.
1994, Kalko andSchnitzler 1998), also forage for insects close to
or drifting onthe water surface while emitting high-intensity CF
and CFFMsignals of medium duration that are rather similar to
theQCFFM signals of mormoopid bats.
Aerial insectivores in highly cluttered space. Batssearching for
and catching fluttering insects in highly clutteredspace close to
vegetation or the ground include all horse-shoe and Old World
leaf-nosed bats (Rhinolophidae, Hip-posideridae; Figures 1e, 4, and
5f) and the New World mus-tached bat, Pteronotus parnellii
(Mormoopidae). All of thesebats must cope with a situation in which
insect echoes areburied in background clutter (Figure 3d). Only if
the echoesare so unique that they can be distinguished from the
clutterechoes can these bats detect, classify, and localize prey
byecholocation alone. Moreover, these bats must also knowtheir
exact spatial position to navigate and to avoid collision.
Bats in the guild highly cluttered space aerial
insectivoressolve their problems by using long-duration
(approximately10100 ms), medium to high frequency (> 30 kHz) CF
orCFFM signals. They emit mostly one search signal per wing-beat.
The CF component is overlap insensitive, but the FMcomponent is
not. Groups of several search signals per wing-beat have also been
recorded.
Because of the long duration of signals and the proximityof
background targets, bats using CFFM signals almost ex-clusively
forage for insects flying in the clutter-overlap zone.Special
adaptations of the echolocation system, such asDoppler-shift
compensation, an auditory fovea in the cochlea,and a highly
modified processing area in the auditory cortex,enable bats using
CFFM signals to distinguish insect echoesmodulated in the rhythm of
the wingbeat from the overlap-ping, unmodulated emitted signals and
clutter echoes. As inother bats, the overlap-sensitive FM component
is used for pre-cise localization of targets. With these
adaptations, the signalsare especially adapted to forage for
insects in highly clut-tered environments.
The mustached bat (Pteronotus parnellii) forages contin-uously
on the wing, whereas others (Hipposideros, Rhino-lophus) sometimes
also hunt from perches in a flycatcherstyle. They capture
fluttering insects, mainly in the air, but theysometimes also glean
them from surfaces.
Gleaning insectivores in highly cluttered space.Bats gleaning
their prey from surfaces of vegetation or theground forage in
highly cluttered space. They are mainlyfound among ghost-faced bats
(Megadermatidae), slit-facedbats (Nycteridae), New World leaf-nosed
bats (Phyllostomi-dae), and evening bats (Vespertilionidae; Figures
1f and 1g,4, 5d and 5e). Like bats in the highly cluttered space
aerialinsectivore guild, they must cope with a situation in which
preyechoes are buried in background clutter (Figure 3c). Onlyif the
echoes have a prey-specific signature that canbe distinguished from
clutter echoes can the bats find theirprey by echolocation.
Moreover, they also face the problemthat they must know their exact
spatial position andthe position of landmarks for navigation and
collisionavoidance.
Bats of the highly cluttered space gleaning insectivoresandall
other gleaners use broadband, overlap-sensitive uni- or
mul-tiharmonic calls of short duration (approximately 13 ms),often
at a very low SPL (whispering bats). Depending on thedistance from
clutter targets, the bats emit either a single sig-nal or groups of
two or more signals per wingbeat. Theechoes of their mostly
stationary insect prey are buried in clut-ter, making the use of
echolocation for detection, localization,and classification
difficult if not impossible. In laboratory stud-ies, narrow-space,
gleaning, FM-emitting foragers have learnedto discriminate various
targets according to spectral differencesin the echoes (summarized
in Ostwald et al. 1988), and it hasbeen suggested that bats can use
spectral cues to distinguishstationary prey from clutter (Simmons
and Stein 1980,Neuweiler 1990). However, to date no one has been
able todemonstrate that spectral cues in echoes from nonflying
in-sects are specific enough for bats to recognize them
usingbroadband FM signals under normal field conditions in ahighly
cluttered situation.
Many gleaners that emit broadband FM signals use prey-generated
acoustic cues (e.g., calls of insects and frogs orrustling noises
of insects walking on leaf litter) for detec-tion, classification,
and localization of prey (Tuttle and Ryan1981, Belwood and Morris
1987, Faure and Barclay 1994).Many gleaners are characterized by
large ears that facilitate pas-sive acoustic localization of prey.
All gleaners always emitecholocation signals in flight to determine
the position of thesite with food, to navigate, and to avoid
collisions. The low SPLof the calls may prevent overloading of the
hearing system withloud clutter echoes.
The use of prey-generated acoustic cues for the
detection,localization, and classification of prey does not exclude
the pos-sibility of a gleaning bat using echolocation to find its
prey un-der favorable conditions. Learning in a context-specific
situ-ation may play an important role. For example, if we offer
aMyotis myotis individual a noisy insect sitting on a screen,
itfirst uses the prey-generated sounds for passive
localization,approach, and capture of the insect. After some
experience,the bat learns that echolocation cues from a protruding
tar-get indicate a sitting insect and thus also approaches silent
in-sects. Bats make such a transfer to echolocation only if, in
a
July 2001 / Vol. 51 No. 7 BioScience 565
Articles
-
specific place, some kind of echolocation cues can be
associ-ated with prey. In other places they do not react to
similarecholocation cues.
Other meaningful categorizations of batsSimilar ecological
constraints exert similar selective pres-sures on signal design,
resulting in a coherent and distinctiveclass of characters in the
search signals of bats that live undersimilar conditions. This
intimate linkage allows one to clas-sify bats into functional
groups such as guilds. The prerequisitefor the categorization of
bats into meaningful groups is thedefinition of habitats. Fenton
(1990) summarizes suchapproaches and distinguishes three main
habitat types: open,edge, and closed habitat. We use a rather
similar classificationand define explicit borders between the
habitat types(Schnitzler and Kalko 1998). We propose that bats
foragingmainly in one of the three defined habitat types,
unclutteredspace, background-cluttered space, or highly cluttered
space,can be categorized as open-space foragers, edge and gap
foragers,and narrow-space foragers, respectively. With
open-spaceforagers and with edge and gap foragers,the grouping
matches the correspondingguilds uncluttered aerial or trawling
in-sectivoresand background-cluttered aer-ial or trawling
insectivores. The group ofnarrow-space foragers covers all the
guildsof bats that forage in highly cluttered space,gleaning their
prey from surfaces or cap-turing aerial prey close by. In both
situa-tions, the prey echo overlaps with the clut-ter echo from the
substratum.
This overlap is likely to mask importantinformation, and to
solve this problemtwo behavioral strategies have evolved.The
preferred foraging habitats, behav-ioral strategies, and the
associated signaltypes are used to categorize two subgroupsof
narrow-space foragers. To avoid misin-terpretation (Fenton 1999)
and to makeclear that the cited signal types indicate aspecific
behavioral strategy, the abbrevia-tions of these signal types
appear in quo-tation marks (Schnitzler et al. forthcom-ing). The
group of narrow-spaceflutter-detecting CFforagers correspondsto the
guild highly cluttered space aerial in-sectivores. By Doppler-shift
compensa-tion and specialized hearing systems, theyseparate the
long CFFM emitted signalfrom the overlapping returning echoes inthe
frequency domain, and they evaluateflutter information. The group
of narrow-space gleaning FM foragers represents allguilds that
include gleaning bats (with theexception of the trawling bats),
includingthe guild highly cluttered space gleaning in-
sectivores described above. It makes sense to categorize
allgleaners into one group, as they have to solve rather
similarproblems when searching for food. While flying in
narrowspaces, these bats mainly use prey-generated cues to
detect,localize, and classify their prey, and they use their
broadbanduni- or multiharmonic FM signals of short duration andlow
SPL mainly for orientation.
Fenton proposes another categorization to describe
theecholocation behavior of bats and the approach an animaltakes to
foraging (summarized in Fenton 1999). He distin-guishes between
high and lowduty cycle bats based on therelative amount of time
signals are emitted. The highdutycycle bats correspond to the
narrow-space flutter-detectingCF foragers, whereas the lowduty
cycle bats cover all theremaining foragers and unite groups as
different as open-space,edge and gap, and narrow-space gleaning FM
foragers. Ad-ditionally, Fenton differentiates between low and
highsig-nal intensity bats under the premise that this
classification dis-tinguishes aerial insectivores from gleaning
bats. We prefer tocategorize bats into meaningful groups according
to their
566 BioScience July 2001 / Vol. 51 No. 7
Articles
Figure 6. Approach flight and capture maneuver, with the
corresponding searchand approach signal sequence, of a Pipistrellus
pipistrellus capturing an insect.The numbers allow a correlation
between hunting and echolocation behavior(small numbers indicate
the position of the insect). At points 14 the bat is inapproach
flight, at 5 it extends one wing toward the insect and moves its
tail intoa capture position (tail down), at 6 the bat bends its
head into a pouch formed bythe tail membrane to retrieve the insect
(head down), and at 7 the bat straightensits body and resumes its
search flight. In the sequence of echolocation signals,arrows
indicate the following phases in echolocation behavior: SP, search
phase;AP, approach phase; BI, buzz I; BII, buzz II; P, pause.
-
main foraging habitat and behavioral strategy, not accordingto
less distinctive signal parameters such as duty cycle and thehighly
variable signal intensity.
Approach signals of insect-eating batsAfter insect-eating bats
have detected prey, they approachand capture it (Figure 6).
Depending on how these bats for-age, they face echolocation
problems of varying difficulty. Batsthat approach a target or a
landing site emit a sequence of ap-proach signals. These signals
have the function of guiding thebat to the chosen target or site.
The sequence of approach sig-nals is less dominated by the habitat
type than by the move-ment of the chosen target (see Kalko and
Schnitzler 1998).
Insectivorous bats that capture prey in the air need to
con-tinuously determine the position of prey while closing in onit.
Because these bats rely on echolocation to locate and trackmoving
insects, we describe this foraging mode as the activemode. By
comparison, gleaning bats that take prey from sur-faces or the
ground mostly evaluate prey-generated acousticcues to detect,
classify, and localize their usually stationary prey.Because
gleaners largely do not use echolocation to find theirprey, we
describe this foraging mode as the passive mode.
In a few cases, bats do not use echolocation or other sen-sory
cues directly to find distant prey but screen known or pre-sumed
feeding sites based on previous experience.Accordingly,these bats
forage in a random mode (Schnitzler et al. 1994,Schnitzler and
Kalko 1998).
Aerial or trawling insectivores. From photographic se-quences of
aerial captures by bats in the field and in the lab-oratory, and
from synchronized sound recordings, we foundthat when aerial or
trawling insectivorous bats detect an in-sect, they immediately
switch from search flight to a target-oriented approach flight,
with head and ears pointing to-ward the insect. Simultaneously, the
bats change theirecholocation behavior from the search phase, with
corre-sponding search signals, to the approach phase or
approachsequence, with distinctive approach signals (Figures 1aeand
6; Griffin et al. 1960).
In the approach phase, bats emit signals in groups corre-lated
with wingbeat and respiratory cycles. The start of the ap-proach is
characterized by a change in signal parameters.With a few
exceptions, signal duration and pulse interval arereduced
throughout the approach phase, which ends in theterminal phase, or
buzz, prior to capturing an insect. The ter-minal phase is
characterized by a series of short signals at ahigh repetition rate
(up to 180200 Hz), usually in one butsometimes in two or more
groups. Typically, the terminalphase of pipistrelles and some
vespertilionids (Pipistrellus,Myotis) consists of two parts: buzz I
(BI) and buzz II (BII;Kalko and Schnitzler 1989). Buzz II signals
are reduced inbandwidth and lower in frequency than buzz I
signals(Figure 6).
In contrast to the various types of search signals foundamong
insectivorous bats, approach signals are remarkablysimilar. Evening
bats (Vespertilionidae) and free-tailed bats
(Molossidae) emit broadband FM signals. Narrowband com-ponents
are eliminated. Similarity in call design reflects asimilar
challenge: the exact localization and tracking of amoving target in
space. Short FM signals are well suited forthis task. Moreover, the
high repetition rate enhances the in-formation flow needed to
control last-instant changes in theinsects position in space.
Sheath-tailed bats (Emballonuridae) and mouse-tailedbats
(Rhinopomatidae) enlarge the bandwidth of the domi-nant harmonic
less strongly than seen in the FM approach sig-nals of
vespertilionids and molossids. They maintain the ba-sic structure
of their search signals and increase the amplitudeof other
harmonics (Figure 1d). This change may increase theoverall
bandwidth of calls of these species.
With the exception of narrow-space flutter-detecting CFforagers,
aerial insectivorous bats avoid an overlap betweenapproach signals
and prey echoes by reducing sound durationand pulse interval to
keep the insect in the overlap-free win-dow. It has been shown for
pipistrelles that sound emissionactually stops several centimeters
before the bat reaches theprey (Kalko 1995).
Narrow-space flutter-detecting CF foragers reduce thesound
duration of the FM portion of their calls but maintaina rather long
CF component throughout the approach se-quence. As in search phase,
the CF component of the echooverlaps with the emitted signal. This
has no masking effect,because the emitted CF component does not
mask theDoppler-shifted CF component of the echo due to the
acousticfovea and sharply tuned neurons in this frequency
range.Presumably, most narrow-space flutter-detecting
CFforagersneed the CF component also during approach to
discriminatethe fluttering prey from the background. The FM
component,however, is reduced and is used for localization.
Gleaning insectivores. Echolocation is used by
gleaninginsectivorous bats for orientation in space and to guide
thebat to the site with prey. After the detection of
prey-generatedsignals, narrow-space gleaning FM foragers fly toward
thesound source and, nearing the site with food, switch to an
ap-proach sequence in which they increase the repetition rate
andreduce the signal duration (Figures 1f and g). However, in
con-trast to aerial insectivores, gleaning bats do not produce a
dis-tinct terminal phase when they are close to the site with
food.By comparison, the repetition rate remains much lower.
Thisbehavioral discrepancy reflects the major difference
betweengleaners, which approach mostly stationary prey, and
aerialinsectivores, which track moving prey and constantly need
toupdate the information about an insects position to suc-cessfully
intercept it.
Flexibility in foraging and echolocation behaviorBats are highly
variable in their foraging and echolocation be-havior (Fenton
1990). Some bats hunt in more than onehabitat, or they may use both
aerial and gleaning modes.Some species that mainly glean insects
from surfaces in highly
July 2001 / Vol. 51 No. 7 BioScience 567
Articles
-
cluttered space in the passive mode also catch insects in
theaerial mode in background-cluttered space or even in
un-cluttered space (e.g., Plecotus). Increasingly, field studies
showthat many bats that forage mainly in background-clutteredspace
also search for insects in uncluttered space (e.g.,
Pip-istrellus).
Flexibility in foraging behavior correlates with plasticity
inecholocation behavior, such as switching from mixed signalswith a
broadband FM component in background-clutteredspace to narrowband
signals in uncluttered space (e.g., Pip-istrellus). Another
strategy is switching from short multihar-monic FM signals of low
intensity when gleaning in highlycluttered space to much louder and
longer uniharmonic FMsignals with a distinct shallow-modulated
component whenforaging in background-cluttered space (e.g.,
Plecotus). Manygleaners use louder signals when they fly more in
the open.Therefore signal intensity is not a good parameter to
catego-rize bats into meaningful groups. However, there are limits
tosuch behavioral plasticity. Bats that are especially adaptedfor
hunting in uncluttered space (e.g., Tadarida) are usuallyrestricted
to this habitat and cannot search for insects
inbackground-cluttered and highly cluttered space. Bats that
aremainly adapted for background-cluttered space do not
exploithighly cluttered space. Thus, the access of bats from
theirspecific habitat to a less-cluttered space is possible, but
the re-verse is not. Fenton (1990) explained this restriction based
onperceptual problems (ability to detect prey in clutter) and
me-chanical problems (ability to fly close to clutter).
Limitationsof the motor system, particularly flight performance
deter-mined by wing shape, largely prevent access to habitats witha
more difficult clutter situation. Sensory abilities, or at leastthe
ability to produce the suitable echolocation signals, shouldbe less
restrictive, as all bats can produce the short FM signalsnecessary
in cluttered situations.
Flexibility creates exceptions and can make it difficult
toclearly define boundaries between guilds and to assign
indi-vidual species to a particular guild. However, in such cases
wewould classify bats into guilds according to their dominant
sen-sory and motor adaptation and assign them to the habitat
typewhere they face the more difficult clutter situation. For
in-stance, bats that hunt in background-cluttered and
unclutteredspace would be assigned to background-cluttered
space.
The comparison of the echolocation behavior of insecteating bats
reveals that bats foraging in similar habitats withsimilar foraging
modes for similar diets share similar adap-tations in their
echolocation systems.When moving from onehabitat to another, bats
change their echolocation behavior andemit the habitat-specific
search calls (Figure 5). Therefore, theecholocation behavior of
bats and especially the structure ofsearch signals are good
indicators of the ecological constraintsunder which bats search for
food.
AcknowledgmentsSome of the ideas of this paper have been
stimulated by thework of people who cannot all be cited due to
limited space.We thank them all, in particular Don Griffin, Brock
Fenton,
Gerhard Neuweiler, and their coworkers.We also thank I. Kaipffor
technical assistance in the laboratory and in the field.This
article has benefited greatly from the comments of An-nette
Denzinger, Charles Handley, Jo Ostwald, Cindy Moss,and Allen Herre.
Our field work has been supported by grantsfrom the Deutsche
Forschungsgemeinschaft (SFB 307, SPPMechanismen zur
Aufrechterhaltung tropischer Diversitt;SFB 550) and Smithsonian
trust funds. We also acknowl-edge the excellent working conditions
of the SmithsonianTropical Research Institute in Panama.
References citedBarclay RMR. 1985. Long- versus short-range
foraging strategies of hoary (La-
siurus cinereus) and silver-haired (Lasionycteris noctivagans)
bats andthe consequences for prey selection. Canadian Journal of
Zoology 63:2507251.
. 1986. The echolocation calls of hoary (Lasiurus cinereus) and
silver-haired (Lasionycteris noctivagans) bats and the consequences
for prey se-lection. Canadian Journal of Zoology 64: 27002705.
Barclay RMR, Brigham RM. 1991. Prey detection, dietary niche
breadth, andbody size in bats: Why are aerial insectivorous bats so
small? American Nat-uralist 137: 693703.
Belwood JJ, Morris GK. 1987. Bat predation and its influence on
calling be-havior in neotropical katydid. Science 238: 6467.
Boonman AM, Boonman M, Bretschneider F, Grind vd WA. 1998. Prey
de-tection in trawling insectivorous bats: Duckweed affects hunting
be-havior in Daubentons bat, Myotis daubentonii. Behavioral Ecology
andSociobiology 44: 99107.
Denzinger A, Kalko EKV, Jones G. Forthcoming. Ecological and
evolution-ary aspects of echolocation in bats. In Thomas JT, Moss
CF,Vater M, eds.Advances in the Study of Echolocation in Bats and
Dolphins. Chicago:University of Chicago Press.
Emde vd G, Schnitzler HU. 1990. Classification of insects by
echolocatinggreater horseshoe bats. Journal of Comparative
Physiology A 167:423430.
Faure PA, Barclay RMR. 1994. Substrate-gleaning versus
aerial-hawking:Plasticity in the foraging and echolocation
behaviour of the long-earedbat, Myotis evotis. Journal of
Comparative Physiology A 174: 651660.
Fenton MB. 1990. The foraging behavior and ecology of animal
eating bats.Canadian Journal of Zoology 68: 411422.
. 1999. Describing the echolocation calls and behaviour of bats.
ActaChiropterologica 1: 127136.
Griffin DR, Webster FA, Michael CR. 1960. The echolocation of
flying insectsby bats. Animal Behaviour 3: 141154.
Henson OW Jr, Schnitzler HU. 1980. Performance of airborne
animal sonarsystems, II:Vertebrates other than Microchiroptera.
Pages 183195 in Bus-nel RG, Fish JF, eds. Animal Sonar Systems. New
York: Plenum Press.
Jensen ME, Miller LA. 1999. Echolocation signals of the bat
Eptesicus serot-inus recorded using a vertical microphone array:
Effect of flight altitudeon searching signals. Behavioral Ecology
and Sociobiology 47: 6069.
Jones G. 1994. Scaling of wingbeat and echolocation pulse
emission rates inbats: Why are aerial insectivorous bats so small?
Functional Ecology 8:450457.
Kalko EKV. 1995. Foraging behavior, capture techniques, and
echolocationin European pipistrelle bats (Microchiroptera). Animal
Behaviour 50:861880.
Kalko EKV, Schnitzler HU. 1989. The echolocation and hunting
behavior ofDaubentons bat, Myotis daubentoni. Behavioral Ecology
and Sociobiology24: 225238.
. 1993. Plasticity in echolocation signals of European
pipistrelle batsin search flight: Implications for habitat use and
prey detection. Behav-ioral Ecology and Sociobiology 33:
415428.
. 1998. How echolocating bats approach and acquire food.
Pages197204 in Kunz TH, Racey PA, eds. Bat Biology and
Conservation.Wash-ington (DC): Smithsonian Institution Press.
568 BioScience July 2001 / Vol. 51 No. 7
Articles
-
Kalko EKV, Handley CO Jr, Handley D. 1996. Organization,
diversity and long-term dynamics of a neotropical bat community.
Pages 503553 in CodyM, Smallwood J, eds. Long-term Studies in
Vertebrate Communities. LosAngeles: Academic Press.
Kober R, Schnitzler HU. 1990. Information in sonar echoes of
fluttering in-sects available for echolocating bats. Journal of the
Acoustical Society ofAmerica 87: 882896.
Lee DN, van der Weel FR, Hitchcock T, Matejowsky E, Pettigrew
JD. 1992.Common principle of guidance by echolocation and vision.
Journal ofComparative Physiology A 171: 563571.
Moss CF, Schnitzler HU. 1995. Behavioral studies of auditory
informationprocessing. Pages 87145 in Popper AN, Fay RR, eds.
Springer Handbookof Auditory Research: Hearing by Bats. New York:
Springer-Verlag.
Moss CF, Zagaeski M. 1994. Acoustic information available to
bats using fre-quency modulated echolocation sounds for the
perception of insectprey. Journal of the Acoustical Society of
America 95: 27452756.
Mller R, Schnitzler HU. 1999. Acoustic flow perception in
cf-bats: Proper-ties of the available cues. Journal of the
Acoustical Society of America 105:29582966.
. 2000. Acoustic flow perception in cf-bats: Extraction of
parameters.Journal of the Acoustical Society of America 108:
12981307.
Neuweiler G. 1989. Foraging ecology and audition in echolocating
bats.Trends in Ecology and Evolution 6: 16066.
. 1990. Auditory adaptations for prey capture in echolocating
bats.Physiological Review 70: 615641.
Norberg UM, Rayner JMV. 1987. Echological morphology and flight
in bats(Mammalia, Chiroptera): Wing adaptations, flight
performance, forag-ing strategy and echolocation. Philosophical
Transactions of the RoyalSociety of London B 316: 335427.
Ostwald J, Schnitzler HU, Schuller G. 1988. Target
discrimination and tar-get classification in echolocating bats.
Pages 413434 in Nachtigall P, ed.Animal Sonar Systems. New York:
Plenum Press.
Pye JD. 1980. Adaptiveness of echolocation signals in bats:
Flexibility in be-haviour and in evolution. Trends in NeuroScience
3: 232235.
Root RB. 1967. The niche exploitation pattern of the blue-gray
gnatcatcher.Ecological Monographs 37: 31750.
Rydell J, Miller LA, Jensen ME. 1999. Echolocation constraints
of Dauben-tons bat foraging over water. Functional Ecology 13:
247255.
Schnitzler HU. 1987. Echoes of fluttering insects: Information
for echolocatingbats. Pages 226243 in Fenton MB, Racey PA, Rayner
JMV, eds. Ad-vances in the Study of Bats. Cambridge (UK): Cambridge
University Press.
Schnitzler HU, Henson OW Jr. 1980. Performance of airborne
animal sonarsystems, I: Microchiroptera. Pages 109181 in Busnel RG,
Fish JF, eds.An-imal Sonar Systems. New York: Plenum Press.
Schnitzler HU, Kalko EKV. 1998. How echolocating bats search and
find food.Pages 183196 in Kunz TH, Racey PA, eds. Bat Biology and
Conserva-tion. Washington (DC): Smithsonian Institution Press.
Schnitzler HU, Ostwald J. 1983. Adaptations for the detection of
flutteringinsects by echolocating bats. Pages 801827 in Ewert JP,
Capranica RR,Ingle DJ, eds. Advances in Vertebrate Neuroethology.
New York: PlenumPress.
Schnitzler HU, Kalko EKV, Kaipf I, Grinnell AD. 1994. Fishing
and echolca-tion behavior of the greater bulldog bat, Noctilio
leporinus, in the field.Behavioral Ecology and Sociobiology 35:
327345.
Schnitzler HU, Kalko EKV, Denzinger A. Forthcoming. Evolution of
echolo-cation and foraging behavior in bats. In Thomas JT, Moss CF,
Vater M,eds. Advances in the Study of Echolocation in Bats and
Dolphins.Chicago: University of Chicago Press.
Siemers BM, Schnitzler HU. 2000. Natterers bat (Myotis nattereri
Kuhl,1818) hawks for prey close to vegetation using echolocation
signals of verybroad bandwidth. Behavioral Ecology and Sociobiology
47: 400412.
Simmons JA, Stein RA. 1980. Acoustic imaging in bat sonar:
Echolocationsignals and the evolution of echolocation. Journal of
Comparative Phys-iology A 135: 6184.
Tuttle MD, Ryan MJ. 1981. Bat predation and the evolution of
frog vocal-izations in the neotropics. Science 214: 677678.
Waters DA, Rydell J, Jones G. 1995. Echolocation call design and
limits onprey size: A case study using the aerial hawking bat
Nyctalus leisleri. Be-havioral Ecology and Sociobiology 37:
321328.
July 2001 / Vol. 51 No. 7 BioScience 569
Articles
Individual members of AIBS enjoy
FREE FUNDING INFORMATION
online viaCommunity of Science Funding Opportunities
tracking more than 20,000 publicand private funding sources
worldwidefor research, education, and training.
Updated daily.
Access via www.aibs.org