Elisabeth K. V. Kalko ? Hans-Ulrich Schnitzler
Ingrid Kaipf ? Alan D. Grinnell
Echolocation and foraging behavior of the lesser bulldog bat,
Noctilio albiventris : preadaptations for piscivory?
Received: 21 April 1997 /Accepted after revision: 12 January 1998
Abstract We studied variability in foraging behavior of
Noctilio albiventris (Chiroptera: Noctilionidae) in Costa
Rica and Panama? and related it to properties of its
echolocation behavior. N. albiventris searches for prey in
high (>20 cm) or low (<20 cm) search flight, mostly
over water. It captures insects in mid-air (aerial cap-
tures) and from the water surface (pointed dip). We once
observed an individual dragging its feet through the
water (directed random rake). In search flight, N. al-
biventris emits groups of echolocation signals (duration
10?11 ms) containing mixed signals with constant-fre-
quency (CF) and frequency-modulated (FM) compo-
nents, or pure CF signals. Sometimes, mostly over land,
it produces long FM signals (duration 15?21 ms). When
N. albiventris approaches prey in a pointed dip or in
aerial captures, pulse duration and pulse interval are
reduced, the CF component is eliminated, and a termi-
nal phase with short FM signals (duration 2 ms) at high
repetition rates (150?170 Hz) is emitted. Except for the
last pulses in the terminal phase N. albiventris avoids
overlap between emitted signals and echoes returning
from prey. During rakes, echolocation behavior is sim-
ilar to that in high search flight. We compare N. al-
biventris with its larger congener, N. leporinus, and dis-
cuss behavioral and morphological specializations that
can be interpreted as preadaptations favoring the evo-
lution of piscivory as seen in N. leporinus. Prominent
among these specializations are the CF components of
the echolocation signals which allow detection and
evaluation of fluttering prey amidst clutter-echoes, high
variability in foraging strategy and the associated
echolocation behavior, as well as morphological spe-
cializations such as enlarged feet for capturing prey from
the water surface.
Key words Bats ? Echolocation ? Foraging ?
Evolution ? Piscivory
Introduction
The development of flight and echolocation give bats
(Microchiroptera) access to a wide range of habitats and
foods. Echolocation allows orientation at night and to a
varying degree detection, classification, and localization
of food. Signal structure varies widely across micro-
chiropteran families and genera (e.g., Fenton 1995;
Neuweiler 1990; Simmons and Stein 1980). To under-
stand the adaptive value of variability in signal design,
comparative field studies of bats living under similar
ecological conditions are essential for linking use of
signal types to ecological constraints imposed by for-
aging habitat, feeding mode, and type of food, also
taking into consideration the phylogenetic relationships
of species (e.g., Fenton 1990; Neuweiler 1989, 1990;
Schnitzler and Kalko 1998; Simmons and Stein 1980).
Bats foraging over water are excellent models for such
studies. Members of several families capture prey from
or near water surfaces. These include evening bats
(Vespertilionidae: Myotis; subgenera Leuconoe,
Pizonyx), sheath-tailed bats (Emballonuridae: Rhynch-
onycteris naso), New World leaf-nosed bats (Phyllos-
tomidae: Macrophyllum macrophyllum), bulldog bats
(Noctilionidae: Noctilio albiventris and N. leporinus),
Old World false vampire bats (Megadermatidae:
Behav Ecol Sociobiol (1998) 42: 305?319
?
Springer-Verlag 1998
E.K.V. Kalko (&) ? H.-U. Schnitzler ? I. Kaipf
Animal Physiology, University of Tu
?
bingen,
Auf der Morgenstelle 28, D-72076 Tu
?
bingen, Germany
e-mail: Elisabeth.Kalko@uni-tuebingen.de,
Fax: (49) 7071-29-2618
E.K.V. Kalko ? H.-U Schnitzler
Smithsonian Tropical Research Institute,
P. O. Box 2072, Balboa, Panama?
E.K.V. Kalko
National Museum of Natural History, MRC 108,
Washington, DC 20560, USA
A.D. Grinnell
Department of Physiology and the Ahmanson Laboratory
of Neurobiology, School of Medicine, UCLA,
Los Angeles, CA 90024, USA
Megaderma lyra), and slit-nosed bats (Nycteridae: Ny-
cteris grandis) (e.g., Altenbach 1989; Audet et al. 1990;
Britton et al. 1997; Gudger 1943; Harrison 1975; Hood
and Jones 1984; Hood and Pitocchelli 1983; Jones and
Rayner 1988, 1991; Kalko 1995a; Kalko and Schnitzler
1989a; Marimuthu et al. 1995; Maya 1968; Novick and
Dale 1971; Patten and Findley 1970; Schnitzler et al.
1994; Suthers 1965, 1967). Although all of these bats
forage in similar habitats, signal design is highly diverse
ranging from short frequency-modulated (FM) signals
to relatively long constant-frequency (CF) and mixed
CF-FM signals.
Here we focus on the lesser bulldog bat, Noctilio al-
biventris, which feeds mainly on insects. Although this
bat will eat small pieces of fish in captivity and parts of
fish have been found occasionally in its feces (Howell
and Burch 1974), fish do not play an important role in
its diet (e.g., Hood and Pitocchelli 1983; Hooper and
Brown 1968). We compare N. albiventris with its con-
gener N. leporinus which is well-known for its fish-eating
habits but occasionally also eats insects. An insectivo-
rous ancestor resembling N. albiventris may have given
rise to N. leporinus (Novick and Dale 1971).
Field studies of N. leporinus have revealed several
hunting strategies: the bat takes prey from the water
surface in a pointed dip, rakes with its claws through the
water, and captures insects in the air (e.g., Hood and
Jones 1984; Novick and Dale 1971; Schnitzler et al.
1994). In contrast, few published studies focus on for-
aging behavior of N. albiventris (e.g., Fenton et al. 1993).
They indicate that it ga?s insects from the water surface
(e.g., Brown et al. 1983; Hood and Pitochelli 1983;
Howell and Burch 1974; Suthers and Fattu 1973).
In both noctilionids, general structure and pattern of
echolocation signals are similar (Roverud and Grinnell
1985a, b; Suthers 1965, 1967). They produce CF signals
interspersed with mixed signals consisting of a CF
component and steep FM components. This pattern is
unique among bats known to take food from the water
surface. CF components facilitate the detection of
moving prey amidst clutter-producing background (e.g.,
Neuweiler 1990; Roverud et al. 1991; Schnitzler and
Kalko 1998), a challenge confronting both Noctilio when
hunting over water or over land.
To link foraging with echolocation behavior and to
elucidate behavioral and sensory adaptations which may
have favored the evolution of the piscivory seen in N.
leporinus we studied N. albiventris in the field at study
sites in Costa Rica and Panama
?
.
Methods
Field sites and bats
From a boathouse (Fig. 1) and from a boat at mid-river we ob-
served N. albiventris for 20 nights (20 October?10 November 1990)
at the Tortuguero River, a tidal estuary about 60 m wide, at
Tortuga Lodge, Costa Rica. We made additional observations of
N. albiventris in Panama
?
on 35 nights (June 1991?September 1995)
around Barro Colorado Island (BCI), a field station of the
Smithsonian Tropical Research Institute in Lake Gatun (9 09?N,
79 51?W), Panama Canal, and at Gamboa on the Chagres River,
about 10 km southeast of BCI. Recording and observation ses-
sions started when the bats began to forage about half an hour
after dusk, and usually ended around 12:30 a.m. We observed
foraging bats with night vision goggles (type Wild) and/or head-
lights and followed them acoustically with a custom-made bat
detector. N. albiventris and N. leporinus are widespread through-
out tropical and subtropical South and Central America. N. al-
biventris is similar to N. leporinus but is readily identified in the
field by its smaller size (30?35 g) and higher frequency echoloca-
tion calls.
Photography of hunting bats and sound recordings
We photographed bats under stroboscopic illumination with a
custom-made multiflash array consisting of two 35-mm cameras
(Nikon 301, 35-mm lens) and a custom-made flash unit (Heinze,
Germany) with four flash tubes (guide number 50). We triggered
the flash unit up to four times with flash intervals set to 50 or
100 ms, which gave 16 images of an individual bat on a single
Fig. 1 a Typical flight paths of
hunting Noctilio albiventris in-
cluding b pointed dip, c raking
pass, d aerial capture and e a
combination of aerial capture
and pointed dip on the Tort-
uguero River near Tortuga
Lodge (A), in front of the boat
house (B) and the dock house
(C), and near an anchored boat
(D). The boat is somewhat
enlarged and not drawn to scale
306
photograph. Simultaneously, the echolocation signals of the bats
were picked up by ultrasound microphones of two bat detectors
(model QMC 100), amplified and recorded on two amplitude-
modulated channels of a high-speed tape recorder (Lennartz 6000/
607, ?-inch tape) at 76 cm/s. The frequency response of the system
was flat within about 10?15 dB between 20?120 kHz. Synchroni-
zation pulses between flashes and sound recordings and spoken
comments were recorded on the two frequency-modulated channels
of the recorder. N. albiventris showed no obvious changes in
echolocation or hunting behavior in response to the flashes.
For three-dimensional reconstruction we digitized the photo-
graphs (Bitpad One, Summagraphics, Fairfield, USA, maximum
resolution 0.1 mm) and used custom-made computer programs for
analyses. Sound recordings were analyzed at 1/16 speed with a
digital frequency-analyzer MOSIP 3000 (Modular Signal Proces-
sor; Medav, Germany) using a fast Fourier transformation (FFT).
A frequency range of 160 kHz was chosen for analysis with a
Hanning Window 256. The sequences were displayed in consecutive
20-ms segments. Sound parameters were measured with a cursor on
the screen. These settings resulted in a frequency resolution of
400 Hz, and time resolution of 40 ms (interpolated). The dynamic
range was restricted to 60 dB to eliminate background noise.
Measurement points were set 40 dB below maximum. Components
>1 ms duration and sweep rates <0.4 kHz/ms were defined as
constant frequency (CF). In case of mixed signals we define the
onset of the FM component as the point where the frequency
dropped 4 kHz below the highest frequency in the CF component.
The FM portions often showed a characteristic pattern of nulls
resulting from interference between the emitted signal and reflec-
tions from the water surface (Kalko and Schnitzler 1989a). Within
sequences, the frequency of the signals deviated due to angle-de-
pendent Doppler shifts. To minimize this error we used only pulses
recorded when the bats flew almost straight toward the micro-
phone. Quantitative data on echolocation behavior are given as
mean values of sample size ?1 SD. A general t-test was used for
comparisons.
The three-dimensional analysis of the photographs was merged
with the sound recordings using custom-made software. We took
23 films (784 photographs; Kodak Tmax 400 black-and-white print
film) and recorded 16 tapes at high speed (12 min real time each;
Afga PE 49 professional) and three audiocassettes (90 min each) of
echolocation sequences. For final analysis of echolocation we se-
lected 121 sequences (40 in high search flight; 34 while raking; 24 in
low search flight; 23 with pointed dips). Of our photographs 51%
showed foraging or transecting N. albiventris, including 55 captures
(10 aerial captures, 40 pointed dips, 5 captures from rakes). To
illustrate correlation of foraging behavior with echolocation we
selected 53 photographs (4 with high search flight; 2 with low
search flight; 10 aerial captures and 28 pointed dips from high
search flight; 5 rakes with captures; 4 rakes without captures)
synchronized with good sound recordings. We also included all
relevant information from the remaining photographs which were
either not suited for 3D reconstruction or lacked good sound
recordings. For more details on methods see Kalko (1995b), Kalko
and Schnitzler (1993), and Schnitzler et al. (1994).
Results
Hunting strategies and echolocation behavior
of N. albiventris foraging over water
At our study sites, N. albiventris arrived every evening
about half an hour after dusk. They either foraged close
to the shore, flying back and forth in a figure-eight
pattern (40?50 m long) or hunted 20?30 m or more from
the shore in large, meandering circles above the water
(Fig. 1a). Foraging bats caught insects in the air
(Fig. 1d), took prey from the water surface in pointed
dips (Fig. 1b), used a combination of aerial captures and
pointed dips (Fig. 1e), or raked with their claws through
the water (Fig. 1c). Typically, foraging activity was high
for the first 1?1.5 h, then tapered o? to 2?3 bats/10
minutes. The level of activity changed from night to
night. On the bright nights around the full moon, the
activity dropped to a very low level, with no more than
an occasional passing flight. We observed a similar de-
crease in activity in N. leporinus which foraged in the
same areas (Schnitzler et al. 1994). However, we cannot
rule out the possibility that the bats hunted elsewhere on
those nights.
High search flight
In high search flight (>20 cm above water or ground)
N. albiventris flapped its wings continuously, with about
4 full wingbeats/s at flight speeds of 7?7.5 m/s (Fig. 2a).
It typically emitted groups of mostly two or three CF or
CF-FM echolocation signals of medium duration
(10 ms) near the top of each wing beat (Fig. 2a,b,f). In
groups with two pulses, the first pulse was usually a CF
signal followed by a mixed CF-FM signal (Fig. 2b,c).
With more pulses in the groups, we found all possible
arrangements of the two pulse types but usually the
group ended with a CF-FM signal. Sometimes a bat also
emitted long FM calls (Fig. 3a,b) with varying sweep
rates, a call type most often observed when N. albiventris
was foraging high (>2 m) over land.
Pulse duration of search signals varied. FM signals
were longest (17?21 ms), followed by CF-FM and CF
signals (10?11 ms) (Fig. 2c,g,i). In CF-FM signals, pulse
duration of the CF component was about equal to the
FM component (Fig. 2g,i). The interval between the last
signal of a group and the first signal of the next group
(intergroup interval) was 70?210 ms, approximately 2?4
times the interval (30?50 ms) between signals within a
group (intragroup interval) (Fig. 2d,h). The duty cycle
(amount of time filled with sound) in high search flight
(Fig. 2j), varied from 5?7% between signal groups to
more than 30% within signal groups (Fig. 2e).
CF signals typically began with a short (1?2 ms)
upward modulation of 0.4?1 kHz, followed by a con-
stant middle-portion where the frequency stayed within
0.4 kHz, and terminated in a short (2?3 ms) downward-
modulated portion of 2?4 kHz bandwidth (Fig. 2 l). The
frequency structure of the CF component of the CF-FM
signals was similar to that of CF signals. CF-FM signals
ended with a short (5 ms) downward FM sweep
(Fig. 2m) and a bandwidth of 32 kHz (Fig. 2k). The
energy of CF and CF-FM signals in high search flight
was concentrated mainly in the first harmonic
(Fig. 2l,m). Rarely, spectograms also revealed part of
the second or third harmonic. The frequency of the re-
corded signals is higher than the emitted frequency due
to Doppler shifts. We measured best frequencies of the
CF component between 67 and 72 kHz.
307
Fig. 2a?k (for legend see next page)
308
Aerial capture from high search flight
The predominant hunting strategy of N. albiventris in
Costa Rica was capture of insects in mid-air during high
search flight, mostly 2?5 m above the water. Prey ranged
from small mosquito-sized insects which mostly could
not be discerned on the photographs to medium-sized
moths with head and body lengths of 1?3 cm (estimated
from the photographs). N. albiventris captured prey with
its uropatagium alone (tail scoop) or reached out with a
wing to funnel the insect onto its interfemoral membrane
(wing capture with tail scoop) (Fig. 4a). There is no
evidence in our photographs with aerial captures where
the prey was clearly visible (n ? 4) that the bat seized
insects directly with its mouth. Flight speed was reduced
during captures to 1?2 m/s. Often the bat came almost
to a standstill. The bat immediately retrieved the prey
from its tail pouch (head down) and resumed search
flight. It presumably stored the prey temporarily in its
large cheek pouches (Murray and Strickler 1975). We
often observed that moths reacted to approaching bats
with evasive maneuvers that frequently led them onto
the water surface. Repeatedly, the bats then turned
around and scooped the still fluttering insects in a
pointed dip from the water surface.
With the beginning of the approach flight, N. al-
biventris switched in echolocation behavior from search
phase to the approach sequence, characterized by a
continuous reduction in pulse duration and pulse inter-
val (Figs. 4b?d, 5). Throughout the approach sequence
the CF component of the signals was reduced and finally
dropped. Prior to capture the bat produced a series of
short FM signals (2 ms) at a high repetition rate (up to
160?180 Hz), the terminal phase (Fig. 4b). Except for
the last signals of the terminal phase, the emitted signals
did not overlap echoes returning from the prey (Fig. 5).
Signal emission ceased when the bat was 15?20 cm from
the prey. The duty cycle rose from 19?30% in high or
low search flight to 40?45% in the terminal phase
(Fig. 4e). Preparations for the capture coincided with
the second half of the terminal phase in echolocation
behavior. After a capture maneuver N. albiventris swit-
ched back to search phase after a pause of 90.2 ?
64.4 ms (n ? 10, minimum 18.1 ms, maximum
180.3 ms) where no echolocation calls could be record-
ed. In two of the four photographs with aerial captures
where the prey could be clearly seen the bat went
through the whole capture maneuver but missed the
insect (Fig. 4a).
Pointed dip from high search flight
N. albiventris took prey also directly from the water
surface with its large feet in pointed dips that were di-
rected to a spot where a fluttering insect drifted on the
water surface (Fig. 6a). The bat slowed down from 7?
7.5 m/s to 5?6 m/s (n ? 23 sequences). The prey was
immediately retrieved and processed as in an aerial
captures. As in aerial captures the bat sometimes went
through the whole capture maneuver but missed or lost
the prey (Figs. 4a, 6a).
When the bat descended towards the target, it swit-
ched from search to approach phase (Fig. 6b, f). Echo-
location behavior (Fig. 6b) changed much as it did in
aerial captures. Shortly before dipping into the water the
bat emitted a terminal phase (Fig. 6c?e, g). The pause
after the terminal phase lasted 59.9 ? 49.8 ms (n ? 29,
minimum 11 ms, maximum 188 ms). We did not find a
Fig. 2 Flight and echolocation behavior of N. albiventris during a
pass in high search flight with a 13 images of the same bat (numbered
1?15). The dotted images are reflections of the bat on the water
surface. The water surface is about half-way between the image of the
bat and the image of its reflections. b Sonagrams (frequency versus
time) of echolocation pulses emitted by the bat during the flight in a
Numbers correspond to the photographic images. Plots of signal
parameters of the same sequence including: c pulse duration and
duration of FM component, d pulse interval, and e duty cycle.
Histograms of all analyzed pulses in high search flight, including
means ?x? and standard deviations (SD) for: f pulse duration, g
duration of FM components, h intra- and intergroup interval, i
duration of CF component in CF-FM signals and CF signals, j duty
cycle, and k FM bandwidth. Sonagrams of: l CF and m CF-FM
pulses of N. albiventris (marked * in b), with sonagrams, averaged
spectra, and oscillograms (amplitude)
309
clear correlation between capture success and duration
of the pause. Of 13 photographs with pointed dips where
the prey insect was visible to us, 11 captures were suc-
cessful and prey was missed in 2 cases. Pauses after
successful captures ranged from 22 to 188 ms; the pause
after the unsuccesful captures were 32 and 46 ms, re-
spectively. We measured time intervals of 300?600 ms
between detection and actual capture (n ? 3). This
translates at estimated flight speeds of about 5 m/s into
detection distances of 1.5?3 m. As in aerial captures,
N. albiventris avoided, except for a few signals at the end
of the terminal phase, echo-overlap between emitted
signal and returning echoes from prey.
Low search flight
Occasionally N. albiventris switched from high search
flight to short segments of low search flight (<20% of
photographs) when it flew within 20 cm of and parallel
to the water surface (Fig. 7a). It continued to beat its
wings (Fig. 7a) and employed a full wing beat amplitude
during the upstroke. The amplitude of the downstroke
was somewhat reduced compared with that of high
search flight. Flight speed decreased from 6.5?7 m/s to
4?5 m/s at the end of the segment (n ? 24 sequences).
The bat stayed in low search flight for only 2?5 m before
returning to high search flight.
Signal pattern di?ered strikingly from that of high
search flight. Long series (50?200 ms) of short CF-FM
pulses were interspersed with groups of longer signals,
mostly one CF signal followed by 2?4 CF-FM signals
(Fig. 7b,c). Overall, pulse duration in low search flight
was much lower than in high search flight (6.5 vs.
10.9 ms) (Fig. 2f, 7f). Duration of pure CF signals and
FM components, as well as bandwidth of the FM
components resembled those of high search flight
(Fig. 7g,i,k). Structure of short CF-FM signals within
signal groups resembled approach phase signals. Pulse
interval varied from 10 to 100 ms or more (Fig. 7d,h).
Both intragroup and intergroup intervals were shorter in
low search flight than they were in high search flight.
Within groups of short CF-FM signals, intragroup in-
tervals ranged mostly between 10 and 30 ms (Fig. 7h).
Between groups of short CF-FM signals or longer CF
and CF-FM signals, intergroup interval mostly lasted
40?70 ms (Fig. 7d,h). The duty cycle in low search flight
(28%) (Fig. 7j) was much higher than it was in high
search flight (Fig. 2j). It oscillated between 10 and 40%,
or, rarely, more (Fig. 7e,j). Signal emission was corre-
lated with wing beat. Groups of short CF-FM signals
and groups of longer CF and CF-FM signals corre-
sponded to one wing beat cycle.
Pointed dip from low search flight
This behavior appears to be very rare in N. albiventris.
Behavioral patterns in two pointed dips from low search
flight were similar to those in pointed dips from high
search flight.
Directed random rake
On one occasion at our observation site in Costa Rica,
thousands of small shrimp (estimated body length 0.5?
1 cm from captured individuals) were frequently jump-
ing out of the water. N. albiventris approached this site
in low search flight and raked its claws through the
Fig. 3 a Echolocation sequence of N. albiventris flying in high search
flight over land, including long FM signals and a terminal phase at the
end; b sonagram and averaged spectrum of FM pulse marked with an
arrow in a
310
Fig. 4a?e Flight and echolocation behavior ofN. albiventris during an
attempted aerial capture of an insect. a The bat is numbered with large
numerals, the insect with small numerals. At images 10 + 11 the bat
attempts to capture the insect by reaching out with its wing. In 13?18
the bat has bent its head into its tail membrane. Although the bat goes
through the whole capture maneuver it has lost the insect (see small
numerals 13?18). b Corresponding echolocation sequence. Plots of c
pulse duration, d pulse interval, and e duty cycle
311
water for distances up to 2?5 m (Fig. 8a). During the
rakes the large interfemoral membrane was folded for-
ward and upward by the strong calcars so the bat?s legs
were unobstructed. Wing beat amplitude, particularly
the downstroke, was reduced. Wing beat rate increased
from about 5 beats/s in high search flight to 6?8 beats/s.
Flight speed dropped from 6?7 m/s to 4?5 m/s. Echo-
location pattern and signal structure resembled those in
high search flight (Fig. 8b?k).
Hunting strategy and echolocation behavior
of N. albiventris foraging over land
We observed N. albiventris foraging over land in Pan-
ama
?
. The bats hawked insects for 10?15 min every eve-
ning at heights of 15?20 cm above a clear-cut meadow
adjacent to the Chagres River. On two nights they for-
aged at heights of 3?5 m around street lights in housing
areas 300?500 m from the river. On one occasion we
recorded foraging N. albiventris flying back and forth
between buildings on BCI at flight heights of 3?5 m.
The pattern of the echolocation behavior was similar
to that of aerial captures from high search flight over
water except that the bats frequently emitted long (17?
21 ms) FM signals (Fig. 3a). The long FM signals
started with a short (3?4 ms), steep (7?10 kHz/ms)
downward FM sweep that leveled out in a shallow-
modulated (about 1 kHz/ms) component of 10?12 ms
duration (Fig. 3b). The signal ended with another steep
(10?12 kHz/ms), downward FM sweep of about 1 ms
duration. In contrast to CF and CF-FM signals, spec-
trograms of long FM signals revealed also part of the
second and third harmonic (Fig. 3b).
Discussion
Hunting strategies and capture techniques
We found high flexibility in hunting strategies in N. al-
biventris, ranging from aerial captures over water and
land, pointed dips from high and low search flight to
directed random rakes. Captures of prey by pointed dips
have been reported previously (e.g., Hooper and Brown
1968; Novick and Dale 1971; Suthers and Fattu 1973),
but aerial captures and the rare raking passes have not
been described in published accounts. The observed
flexibility in hunting strategies possibly allows N. al-
biventris to quickly adjust to prey availability and thus
may enhance foraging e?ciency.
N. albiventris is well adapted morphologically to
grasp objects from the water surface in flight. With
pointed dips it takes prey with its large feet and claws in
a manner similar to other dipping bats, particularly
Myotis of the subgenera Leuconoe and Pizonyx (Ves-
pertilionidae), Macrophyllum macrophyllum (Phyllos-
tomidae), and N. leporinus (e.g., Dalquest 1957; Findley
1972; Jones and Rayner 1988, 1991; Kalko and Schnit-
zler 1989b; Schnitzler et al. 1994; authors unpublished
observations). In aerial captures over water or over land,
N. albiventris like other aerial insectivores, uses its in-
terfemoral membrane and/or a wing to snare its prey
(e.g., Gri?n et al. 1960; Kalko 1995b; Webster and
Gri?n 1962).
In contrast to N. leporinus, which we observed fre-
quently raking for distances up to 10 m (Schnitzler et al.
1994), we documented raking passes only once in
N. albiventris. Probably the limited use of raking by this
bat is imposed by energetic and morphological con-
straints. Raking is energetically costly since the bat has
to overcome the hydrodynamic drag created by dragging
its immersed claws through water. Morphological ad-
aptations for drag reduction including lateral compres-
sion of claws are not as well developed in N. albiventris
as in N. leporinus (Fish et al. 1991). Furthermore, small
body size limits its carrying capacity. Hence, the size of
prey that it could catch would be much less than in
N. leporinus, which frequently catches fish (e.g., Alten-
bach 1989; Brooke 1994; Schnitzler et al. 1994). Studies
of N. leporinus have show that capture success is often
very low in raking passes (Bloedel 1955; Schnitzler et al.
1994). Possibly, N. albiventris rakes only when it is as-
sured of a high capture rate in exceptionally rich patches
of small prey such as the mass of leaping crustaceans at
our study site in Costa Rica.
Fig. 5 Pulse duration of the FM component in relation to the bat?s
distance from the insect prey during search, approach, and terminal
phases of an aerial capture by N. albiventris. The prey, a moth, was
successfully caught in this sequence. The solid line marks the range of
echo-overlap
Fig. 6a?g Flight and echolocation behavior of N. albiventris during a
pointed dip from high search flight. The bat goes through the capture
maneuver but misses the prey. a Multiflash sequence and b
corresponding echolocation sequence. Plots of c pulse duration, d
pulse interval, and e duty cycle. Sonagram, averaged spectra, and
oscillogram of f one signal of the approach and g two short FM
signals of the terminal phase marked * in b
c
312
Fig. 6a?g
313
Fig. 7 Flight and echolocation behavior of N. albiventris during a, b
low search flight with c?e plots of pulse duration, pulse interval, and
duty cycle. f?k Histograms of all analyzed sequences from low search
flight
314
Fig. 8 Flight and echolocation behavior ofN. albiventris during a, b a
raking pass, with c?e plots of pulse duration, pulse interval, and duty
cycle. f?k Histograms of all analyzed sequences of raking passes
315
Functional significance of call design
in foraging N. albiventris
Signal design is highly variable in N. albiventris, ranging
from long CF, long and short CF-FM, and long FM
signals in search phase, to short FM signals in approach
and terminal phase. Variability in signal design is linked
to the perceptual tasks that have to be performed by
N. albiventris employing a variety of hunting strategies.
Hunting for fluttering insects in the air
and from the water surface
N. albiventris caught fluttering insects in the air or
gleaned them from the water surface. While searching for
prey, the bat emitted a mixure of CF and CF-FM search
signals. Laboratory studies have shown that echoes of
CF signals from flying prey such as fluttering insects are
characterized by ??acoustical glints??, patterns of prey-
specific amplitude and frequency modulations that en-
code wing beat rate, wing size, and other species-specific
information (e.g., Schnitzler 1987; Kober and Schnitzler
1990; Schnitzler et al. 1983). Chances of perceiving such
temporary glints depend on the duty cycle of the bat?s
echolocation sequence and the wing beat rate of the in-
sects (Kober and Schnitzler 1990). With an average duty
cycle of 19% in high search flight and 30% in low search
flight, and estimated wing beat rates of 50?100 Hz in the
prey, N. albiventris could perceive 9.5?30 glints (duty
cycle ? wing beat rate) per second. Since N. albiventris
has a repetition rate of about 11 signals per second in
high search flight and about 27 Hz in low search flight, it
could perceive up to 1?3 glints in every signal.
We propose that N. albiventris, like other bats with
CF signals (rhinolophids, hipposiderids, Pternonotus
parnellii (Mormoopidae), and N. leporinus) use echolo-
cation information for prey selection by evaluating the
characteristic patterns of temporary acoustic glints in
echoes from fluttering insects (e.g., vd Emde and
Schnitzler 1986, 1990; Jones 1990; Kober and Schnitzler
1990; Schnitzler et al. 1994). In the case of N. albiventris
hunting for fluttering insects drifting on the water sur-
face, most unwanted targets (e.g., debris) float on the
water surface and hence generate fairly stationary glints
easily to be discriminated from the rhythmical pattern of
temporary glints in echoes from fluttering insects (Kober
and Schnitzler 1990; Schnitzler et al. 1994). From dif-
ferences in their search patterns for food we infer that
the two noctilionids evalute glint pattern with di?erent
search images. For instance, we never observed the in-
sectivorous N. albiventris attacking jumping fish, which
inevitably triggered foraging behavior in N. leporinus
(Schnitzler et al. 1994), and when we used an underwater
pump to produce a water jet that broke the water surface
and mimicked a jumping fish, it elicited no response in
N. albiventris, but N. leporinus attacked it. We conclude
that this type of glint pattern is not part of the search
image of N. albiventris.
A number of CF bats are known to compensate for
the Doppler shift caused by their own flight movement
relative to the surroundings. By lowering the emitted
frequency, they keep the frequency of the CF component
of the echo constant at the so-called reference frequency.
Adaptations in the hearing system are characterized by a
specialized cochlea (auditory fovea at the reference fre-
quency) and an overrepresentation throughout the sys-
tem of sharply tuned neurons in the frequency range of
the prey echoes. These enable the bats to discriminate
the characteristic echoes of fluttering insects from the
overlapping, unmodulated emitted signal and from
overlapping clutter (e.g., Neuweiler 1989; Schnitzler and
Henson 1980). Laboratory studies have shown that
N. albiventris also has the capacity for Doppler shift
compensation (Roverud and Grinnell 1985c) although it
is still unclear whether this compensation is as complete
as in horseshoe bats (e.g., Schnitzler and Henson 1980)
or incomplete as in N. leporinus (Wenstrup and Suthers
1984).
We propose that the CF components in the search
signals of N. albiventris adapts it to forage for insects in
??highly cluttered space?? (Schnitzler and Kalko 1998)
where prey echoes overlap clutter echoes from other
objects. This is the case when N. albiventris takes prey in
pointed dips or searches for prey flying lower than 2 m
above the water or ground. At search signal durations of
6.5?10.9 ms, echoes from a target overlap with clutter
echoes at distances of 1.1?1.9 m and less. Conversly,
when the bat hunts for insects higher than the overlap-
zone, echoes indicating potential prey in mid-air are
followed by, but do not overlap, clutter echoes from the
water surface. Thus, N. albiventris according to our
definition is in ??background cluttered space?? (Schnitzler
and Kalko 1998). In both cases, glint detection facilitates
prey detection.
Behavioral and electrophysiological studies of N. al-
biventris suggest that CF signal components may also
play a role in estimating distance. The onset of the CF
component activates a gating mechanism which creates a
time window during which distance information can be
extracted by comparing the FM components of pulse-
echo pairs of the CF-FM search signals (Roverud 1988;
Roverud and Grinnell 1985a,b). However, both nocti-
lionids drop the CF component of their signals during
approach and terminal phase when distance information
is also crucial. Thus, the mechanism for extracting dis-
tance information must di?er in search and approach.
In contrast to other CF bats such as rhinolophids,
hipposiderids, and Pteronotus parnellii which maintain a
CF signal component throughout approach and termi-
nal phase (e.g., Schnitzler et al. 1985), N. albiventris
drops the CF component of its signals during approach
phase and emits pure FM calls in the terminal phase. We
conclude that in the final part of an approach the FM
component alone delivers the necessary information for
the pursuit of an insect prey after it has been detected
and classified with flutter cues. Behavioral experiments
suggest that for an echo of a FM signal in the terminal
316
phase 100 or more discrete filters are present for about
10 msec and that each filter analyzes information from a
band of a few hundred Hertz (Roverud 1993).
Short FM signals or FM components are overlap-
sensitive (Schnitzler and Kalko 1998). Consequently,
bats with overlap-sensitive signals reduce signal duration
and pulse interval during approach and terminal phase
(Kalko and Schnitzler 1989b, 1993). Presumably, over-
lap would mask important information from the bat. At
the end of the terminal phase, the duration of FM pulses
in N. albiventris is longer (2 ms) than FM signals re-
corded in terminal phases of other bats (0.3?0.5 ms)
when capturing insects from the water or in the air (e.g.,
Kalko and Schnitzler 1989b, 1993). Hence, in N. al-
biventris signals overlap the prey echo at distances of
about 26 cm or less. However, like aerial insectivorous
pipistrelle bats (Vespertilionidae) (Kalko 1995b), N. al-
biventris often stops signal emission 10?20 cm before the
capture maneuver, thereby preventing potential overlap.
Gleaning prey from water surface
with directed random rake
The pattern of echolocation signals produced during
raking passes through patches with many jumping
shrimps resembles that in high search flight, except ap-
proach and terminal phases are absent. This suggests
that similar to raking N. leporinus (Schnitzler et al. 1994)
raking N. albiventris do not localize single targets but
search for prey in a random manner in a area where
many temporary glints indicate prey. Possibly optical
cues or the splashing sounds of potential prey also might
assist N. albiventris in prey detection. Since we saw
N. albiventris raking only once, we cannot say whether it
would rake also in places where no target cues are
present but where it had hunted successfully before, as
we found it in the memory directed random rakes of
N. leporinus (Schnitzler et al. 1994).
Comparison of foraging and echolocation behavior
of N. albiventris and N. leporinus
Both noctilionids employed similar hunting strategies
but di?ered in the frequency of the use of various types.
Overall, aerial captures and pointed dips from high
search flight were most used by N. albiventris, whereas
pointed dips from high and low search flight and raking
phases predominanted in N. leporinus.
Echolocation behavior was similar in the two species.
There were few species-specific di?erences (e.g., Gri?n
and Novick 1955; Hartley et al. 1989; Schnitzler et al.
1994; Suthers 1965, 1967; Suthers and Fattu 1973). We
found the CF part of search signals (66?72 kHz) in
N. albiventris at least 10 kHz higher than in N. leporinus
(53?56 kHz). We attribute the di?erences in call fre-
quencies in the two species to di?erences in body mass, a
relationship well documented for a variety of bats (e.g.,
Barclay and Brigham 1991). The smaller N. albiventris
(30?35 g) produces higher frequency calls than the larger
N. leporinus (65?75 g). Although duration of search
signals was significantly shorter in N. albiventris than in
N. leporinus (t
168
? 7.29, P < 0.001), its FM compo-
nent was significantly longer (t
95
? 4.54, P < 0.001),
5.3 ms versus 3.9 ms. In N. albiventris, the FM compo-
nent encompassed about half (50%) of a CF-FM signal
but only 30% in N. leporinus. CF components were
significantly shorter (t
95
? 12.05, P < 0.001) in N. al-
biventris (4.4?5.5 ms) than in N. leporinus (5.4?8.9 ms).
Furthermore, the bandwidth of the FM component was
significantly higher (t
172
? 12.37, P < 0.001) in N. al-
biventris (31?33 kHz) than in N. leporinus (24?26 kHz).
In comparisons of the relative proportions of FM and
CF components in long CF-FM signals, the FM part
was more accentuated in N. albiventris and the CF part
in N. leporinus.
Possibly a shorter CF component delivers adequate
information to N. albiventris since it hunts mainly for
fluttering insects, constantly producing acoustical glints
in the rhythm of their wing beats. The preponderance of
FM components may enhance localization accuracy,
which is of particular importance to a bat capturing
insects in three-dimensional space. Conversely, the
longer CF components in N. leporinus enhance its
chances of detecting the brief glint patterns produced by
fish which jump out of the water for only 50?100 ms.
A possible scenario for the evolution of piscivory
The many similarities in hunting and echolocation be-
haviors of the two noctilionids provide clues to how
hunting for fish may have evolved (Schnitzler et al.
1992). Like other CF bats noctilionids can detect flut-
tering insects by the glint pattern in the echoes. We
suppose that the ancestor of N. leporinus was hunting for
insects including moths much as N. albiventris does
today (Novick and Dale 1971). Since many moths can
hear, they often react to ultrasound with evasive move-
ments and frequently drop. If a moth accidentially
landed in the water, the ancestral bat may have wheeled
and scooped up the still fluttering and thus glint-pro-
ducing insect. Simultaneously, fish may attack, even
jump for insects fluttering on the water surface. Since
jumping fish that break the water surface also produce
characteristic temporary glints, it seems probable that
the insect-hunting ancestor of N. leporinus sometimes
accidentially detected, localized, and caught a jumping
fish instead of a fluttering moth. Since fish have higher
nutritional value than moths, we assume that in areas
where jumping fish were abundant a bat specialized to
hunt fish could evolve. This scenario is supported by our
observation that the behavioral separation of the two
species is not complete. N. leporinus retains its ability to
hunt for insects, and N. albiventris has evolved hunting
modes such as low search flight and directed random
rake to hunt for prey other than insects which produce
317
specific glint patterns. Morphological adaptations in-
cluding the enlarged feet with laterally compressed claws
further enhanced the ability of Noctilio to catch fish.
Signal design of phylogenetcially closely related species
Although there is some disagreement about the degree of
relatedness of noctilionids on the superfamily level, there
is broad consensus that the Noctilionidae and the
Mormoopidae (leaf-chinned bats) form a monophyletic
group (Hood and Smith 1982; Novacek 1991; Smith
1972, 1976; Simmons 1995) and that noctilionids could
be particularly close to the genus Pteronotus (Mor-
moopidae) (Baker and Bickham 1980; Sites et al. 1981).
Mormoopids, particularly Pteronotus, share many sim-
ilarities in their echolocation behavior and ecology with
noctilionids. All mormoopid bats feed on insects. The
mustached bat, Pteronotus parnellii forages in highly
cluttered space in the forest and produces rather long
(20?30 ms) CF signals. The other Pteronotus hunt in
background cluttered space such as forest edges and
gaps and emit shorter signals (8?12 ms) composed of a
short CF component followed by a downward FM
component (Pye 1980; Schnitzler et al. 1991; authors,
unpublished observations).
Due to similarities in signal structure we tentatively
propose that noctilionids and mormoopids evolved from
a common ancestor which used signals with CF com-
ponents to improve detection of fluttering prey. Subse-
quent morphological modifications such as the enlarged
feet in noctilionids allowed them to grasp objects from
surfaces.
Acknowledgements We thank Richard LaVal for advice and help in
the field, the sta? of Tortuga Lodge for making our stay at the
lodge comfortable, and the Smithsonian Tropical Research Insti-
tute (Panama? ) for logistical support. We are most grateful to
Charles Handley, Allen Herre, and three referees for their valuable
comments and suggestions on the manuscript. Our research was
supported by grants from the Deutsche Forschungsgemeinschaft to
Schnitzler (SFB 307) and to Schnitzler and Kalko (SPP ??Mecha-
nismen der Aufrechterhaltung tropischer Diversita
?
t??) and by a
NATO postdoctoral fellowships to Kalko.
References
Altenbach JS (1989) Prey capture by the fishing bats Noctilio le-
porinus and Myotis vivesi. J Mammal 70:421?424
Audet D, Krull D, Marimuthu G, Sumithran S, Singh JB (1990)
Foraging behavior of the Indian false Vampire bat, Megaderma
lyra (Chiroptera: Megadermatidae). Biotropica 23:63?67
Baker RJ, Bickham JW (1980) Karyotypic evolution in bats: evi-
dence of extensive and conservative chromosomal evolution in
closely related taxa. Syst Zool 29:239?253
Barclay RMR, Brigham RM (1991) Prey detection, dietary niche
breadth, and body size in bats: why are aerial insectivorous bats
so small? Am Nat 137:693?700
Bloedel P (1955) Hunting methods of fish-eating bats, particularly
Noctilio leporinus. J Mammal 36:390?399
Britton ARC, Jones G, Rayner JMV (1997) Flight performance,
echolocation and foraging behavior in the pond bat, Myotis
dasycneme (Chiroptera: Vespertilionidae). J Zool Lond 241:
503?522
Brooke A (1994) Diet of the fishing bat, Noctilio leporinus (Chir-
optera: Noctilionidae). J Mammal 75:212?218
Brown PE, Brown TW, Grinnell AD (1983) Echolocation, devel-
opment, and vocal communication in the lesser bulldog bat,
Noctilio albiventris. Behav Ecol Sociobiol 13:287?298
Dalquest WW (1957) Observations on the Sharp-nosed bat,
Rhynchiscus naso (Maximilian). Texas J Sci 9:219?226
Emde G vd, Schnitzler HU (1986) Fluttering target detection in
hipposiderid bats. J Comp Physiol A 159:767?772
Emde G vd, Schnitzler HU (1990) Classification of insects by
echolocating greater horseshoe bats. J Comp Physiol A 167:
423?430
Fenton MB (1990) The foraging behaviour and ecology of animal-
eating bats. Can J Zool 86:411?422
Fenton MB (1995) Natural history and biosonar signals. In: Fay
RR, Popper AN (eds) Hearing by bats (Springer handbook of
auditory research). Springer, Berlin Heidelberg New York, pp
17?86
Fenton MB, Audet D, Dunning DC, Long J, Merriman CB, Pearl
D, Syme DM, Adkins B, Pedersen S, Wohlgenant T (1993)
Activity patterns and roost selection by Noctilio albiventris
(Chiroptera: Noctilionidae) in Costa Rica. J Mammal 74:607?
613
Findley JS (1972) Phenetic relationships among bats of the genus
Myotis. Syst Zool 21:31?52
Fish FE, Blood BR, Clark BD (1991) Hydrodynamics of the feet of
fish-catching bats: influence of the water surface on drag and
morphological design. J Exp Zool 258:164?173
Gri?n DR, Novick A (1955) Acoustic orientation of neotropical
bats. J Exp Zool 130:251?299
Gri?n DR, Webster FA, Michael CR (1960) The echolocation of
flying insects by bats. Anim Behav 8:141?154
Gudger EW (1943) The fish-eating bats of India and Burma.
J Bombay Nat Hist Soc 43:635?640
Harrison DL (1975) Macrophyllum macrophyllum. Mamm Sp 62:
1?3
Hartley DJ, Campbell KA, Suthers RA (1989) The acoustic be-
havior of the fish-eating bat, Noctilio leporinus, during prey
capture. J Acoust Soc Am 86:8?27
Hood CS, Jones JK Jr (1984) Noctilio leporinus. Mammal Sp
216:1?7
Hood CS, Pitocchelli J (1983) Noctilio albiventris. Mammal Sp
197:1?5
Hood CS, Smith JD (1982) Cladistic analysis of female reproduc-
tive histomorphology in phyllostomatoid bats. Syst Zool
31:241?251
Hooper ET, Brown JH (1968) Foraging and breeding in two
sympatric species of neotropical bats, genus Noctilio. J Mam-
mal 49:310?312
Howell DJ, Burch D (1974) Food habits of some Costa Rican bats.
Rev Biol Trop 21:281?294
Jones G (1990) Prey selection by the Greater horseshoe bat (Rhi-
nolophus ferrumequinum): optimal foraging by echolocation? J
Anim Ecol 59:587?602
Jones G, Rayner JMV (1988) Flight performance, foraging tactics
and echolocation in free-living Daubenton?s bats Myotis
daubentoni (Chiroptera: Vespertilionidae). J Zool Lond
215:113?132
Jones J, Rayner JMV (1991) Flight performance, foraging tactics
and echolocation in the trawling insectivorous bat, Myotis ad-
versus (Chiroptera: Vespertilionidae). J Zool Lond 225:393?412
Kalko EKV (1995a) Echolocation signal design, foraging habitats,
and guild structure in six Neotropical sheath-tailed bats (Em-
ballonuridae). Symp Zool Soc Lond 67:259?273
Kalko EKV (1995b) Prey capture and echolocation in foraging
European pipistrelle bats (Microchiroptera). Anim Behav
50:861?880
Kalko EKV, Schnitzler HU (1989a) Two-wave-front interference
patterns in frequency-modulated echolocation signals of bats
flying low over water. J Acoust Soc Am 85:961?962
318
Kalko EKV, Schnitzler HU (1989b) The echolocation and hunting
behavior of Daubenton?s bat, Myotis daubentoni. Behav Ecol
Sociobiol 24:225?238
Kalko EKV, Schnitzler HU (1993) Plasticity in echolocation sig-
nals of European pipistrelle bats in search flight: implications
for habitat use and prey detection. Behav Ecol Sociobiol
33:415?428
Kober R, Schnitzler HU (1990) Information in sonar echoes of
fluttering insects available for echolocating bats. J Acoust Soc
Am 87:882?896
Marimuthu G, Habersetzer J, Leippert D (1995) Gleaning from the
water surface by the false vampire bat. Ethology 99:61?74
Maya JA (1968) The natural history of the fish-eating bat, Pizonyx
vivesi. PhD thesis, University of Arizona
Murray PF, Strickler T (1975) Notes on the structure and function
of cheek pouches within the Chiroptera. J Mammal 56:637?676
Neuweiler G (1989) Foraging ecology and audition in echolocating
bats. Trends Ecol Evol 4:160?166
Neuweiler G (1990) Auditory adaptations for prey capture in
echolocating bats. Physiol Rev 70:615?641
Novacek MJ (1991) Aspects of the morphology of the cochlea in
microchiropteran bats: an investigation of character transfor-
mation. Bull Am Mus Nat Hist 206:84?100
Novick A, Dale BA (1971) Foraging behavior in fishing bats and
their insectivorous relatives. J Mammal 52:817?818
Patten DR, Findley LT (1970) Observations and records of Myotis
(Pizonyx) vivesi Menegaux (Chiroptera: Vespertilionidae)
(Contributions to science 183). Los Angeles County Museum,
Los Angeles
Pye JD (1980) Adaptivness of echolocation signals in bats. Trends
Neuro sci 3:232?235
Roverud RC (1988) A time window for distance information pro-
cessing in the bats, Noctilio albiventris and Rhinolophus rouxi.
In: Nachtigall PE, Moore PWB (eds) Animal sonar processes
and performance. Plenum, New York, pp 513?517
Roverud RC (1993) Neural computations for sound pattern rec-
ognition: evidence for summation of an array of frequency fil-
ters in an echolocating bat. J Neurosci 13: 2306?2312
Roverud RC, Grinnell AD (1985a) Discrimination performance
and echolocation signal integration requirements for target
detection and distance determination in the CF/FM bat, Noc-
tilio albiventris. J Comp Physiol A 156:447?456
Roverud RC, Grinnell AD (1985b) Echolocation sound features
processed to provide distance information in the CF/FM bat,
Noctilio albiventris: evidence for a gated time window utilizing
both CF and FM components. J Comp Physiol A 156:457?469
Roverud RC, Grinnell AD (1985c) Frequency tracking and
Doppler shift compensation in response to an artificial CF/FM
echolocation sound in the CF/FM bat, Noctilio albiventris.
J Comp Physiol A 156:471?475
Roverud RC, Nitsche V, Neuweiler G (1991) Disrimination of wing
beat motion by bats, correlated with echolocation sound pat-
tern. J Comp Physiol A 168: 259?263
Schnitzler HU (1987) Echoes of fluttering insects: information for
echolocating bats. In: Fenton MB, Racey PA, Rayner JMV
(eds) Advances in the study of bats. Cambridge University
Press, Cambridge, pp 226?243
Schnitzler HU, Henson OW Jr (1980) Performance of airborne
animal sonar systems. I. Microchiroptera. In: Busnel RG, Fish
RF (eds) Animal sonar systems. Plenum, New York, pp 109?
181
Schnitzler HU, Kalko EKV (1998) How bats search for food. In:
Kunz TH, Racey PA (eds) Bats: phylogeny, morphology, and
conservation biology. Smithsonian Institution Press, Washing-
ton, in press
Schnitzler HU, Menne D, Kober R, Heblich K (1983) The acous-
tical image of fluttering insects in echolocating bats. In: Huber
F, Markl H (eds) Neuroethology and behavioral physiology:
roots and growing points. Springer, Berlin Heidelberg New
York, pp 235?250
Schnitzler HU, Hackbarth H, Heilmann U, Herbert H (1985)
Echolocation behavior of rufous horseshoe bats hunting for
insects in the flycatcher style. J Comp Physiol 157:39?46
Schnitzler HU, Kalko EKV, Kaipf I, Mogdans J (1991) Compar-
ative studies of echolocation and hunting behavior in the four
species of mormoopid bats of Jamaica. Bat Res News 32:22?23
Schnitzler HU, Kalko EKV, Kaipf I, Grinnell AD (1992) A plau-
sible hypothesis for the evolution of fish-catching behavior in
noctilionid bats. In: Elsner N, Richer DW (eds) Rhythmogen-
esis in neurons and networks. Thieme, Stuttgart, p 211
Schnitzler HU, Kalko EKV, Kaipf I, Grinnell AD (1994) Fishing
and echolocation behavior of the greater bulldog bat, Noctilio
leporinus, in the field. Behav Ecol Sociobiol 35:327?345
Simmons NB (1995) Bat relationships and the origin of flight. In:
Racey PA, Swift SM (eds) Ecology, evolution and behavior of
bats. Symp Zool Soc Lond 67:27?43
Simmons JA, Stein RA (1980) Acoustic imaging in bat sonar:
echolocation signals and the evolution of echolocation. J Comp
Physiol 135:61?84
Sites JW, Bickham W, Haiduk MW (1981) Conservative chromo-
somal change in the bat family Mormoopidae. Can J Genet
Cytol 23:459?467
Smith JD (1972) Systematics of the chiropteran family Mor-
moopidae. Misc Publ Mus Nat Hist Univ Kansas 56:1?132
Smith JD (1976) Chiropteran evolution. In: Baker RJ, Jones JK,
Carter DC (eds) Biology of bats of the New World family
Phyllostomatidae, part I. Spec Publ Mus Tex Tech Univ 10:49?
69
Suthers RA (1965) Acoustic orientation by fish-catching bats. J
Exp Zool 158:319?348
Suthers RA (1967) Comparative echolocation by fishing bats. J
Mammal 48:79?87
Suthers RA, Fattu JM (1973) Fishing behaviour and acoustic
orientation by the bat (Noctilio labialis). Anim Behav 21:61?66
Wenstrup JJ, Suthers RA (1984) Echolocation of moving targets by
the fish-catching bat, Noctilio leporinus. J Comp Physiol A 155:
75?89
Webster FA, Gri?n RD (1962) The role of flight membranes in
insect captures by bats. Anim Behav 10:332?342
Communicated by G. Wilkinson
319